ee a Munhae Veeck abe as-% plang ‘ ‘ I y » Peer ae ead . . ¥: i on Seri ¥ . . ‘ : . "1 : prebs “dia ek ete te : ; : : ‘ RS bd Ne Aveaea’s ada : ; i ee Matatuant tant ast, as : ' : vse tie pare 7 ‘ 5 vie o ue. PU ate u in senna oH ‘ rest : Nye f y ; : Pied 4 : on ‘ Soar ie 4 5 : yalhsve ee Aaee . vat . A fi . ft 4 See ‘ ms . { 5 H 4 Pansat teas } 4 ¥ , Sati ‘ ‘ ‘ fon late ‘ cacy) hata va dna tems . Wienges a cd ae ' Pais : : 4 : : : Fos ‘ a . * : : ands Gta tea gh AASAL EASA ee Se Radin nana ie anoget® - i ; Sintesate REIT Benda a2 teense, by Ue btn are tude eran cs A ¢ f é i SCS oa, ae Nae it : TN ae . : : ris . we z vena in stadt A 3) < ' eae 4 . > ‘wins a tm i ; Searae noes Sat Sees ; i: ‘ Baath gabe ve Hy : : singe ey : . : ui-95 . ; ‘ : 4 z 4 5 { 4 z agtsaece wade . teed ? if rn Pevrenrercer as 4 . t we ; : ea aT Areetnuc ed a wee htasass Saleen # 2 ae TRAE i ig : 5 ae errer tests age ha é . 5 . * x RS ae ts Nae ; if TEA DARR Lathe sapere NAF n sasnaition Witty 2 wah d Pa ! ing - 3 “ : seeitiee engine ay ee epeLhty ince 2800 $ gebeaty Serie ect g f ea oe Tees , mee vendpepetie sees cayay suerte” . Sec raceenet “el Statelscnie ; oy ; i : f veer Way puree ganas ged é is ars Steet seedy te ¥ a acta ore eur raw fh are : cee te wine orrerrt es Son 4 ge 0 peasgae arywrety autre ‘ y iret « * en pg wiweeneee LE ee Sure PLONE err tied J Rerren tice ye aay Picatora A etites MBL/WHOI JOURNAL OF SHELLFISH RESEARCH VOLUME 9, NUMBER 1 JUNE 1990 The Journal of Shellfish Research (formerly Proceedings of the National Shellfisheries Association) is the official publication of the National Shellfisheries Association Dr. Neil Bourne (1992) Fisheries and Oceans Pacific Biological Station Nanaimo, British Columbia VOR 5K6 Canada Dr. Monica Bricelj (1991) Marine Sciences Research Center State University of New York Stony Brook, New York 11794-5000 Dr. Anthony Calabrese (1991) National Marine Fisheries Service Milford, Connecticut 06460 Dr. Kenneth K. Chew (1991) College of Fisheries University of Washington Seattle, Washington 98195 Dr. Peter Cook (1992) Department of Zoology University of Cape Town Rondebosch 7700 Cape Town, South Africa Dr. Charles Epifanio (1991) College of Marine Studies University of Delaware Lewes, Delaware 19958 Editor Dr. Sandra E. Shumway Department of Marine Resources and Bigelow Laboratory for Ocean Science West Boothbay Harbor Maine 04575 EDITORIAL BOARD Dr. Jonathan Grant (1992) Department of Oceanography Dalhousie University Halifax, Nova Scotia B3H 4J1 Canada Dr. Paul A. Haefner, Jr. (1991) Rochester Institute of Technology Rochester, New York 14623 Dr. Robert E. Hillman (1991) Battelle Ocean Sciences New England Marine Research Laboratory Duxbury, Massachusetts 02332 Dr. Herbert Hidu (1991) Ira C. Darling Center University of Maine Walpole, Maine 04573 Dr. Lew Incze (1991) Bigelow Laboratory for Ocean Sciences McKown Point West Boothbay Harbor, Maine 04575 Dr. Louis Leibovitz (1991) Marine Biological Laboratory Woods Hole, Massachusetts 02543 Journal of Shellfish Research Volume 9, Number 1 ISSN: 00775711 June 1990 Dr. Roger Mann (1991) Virginia Institute of Marine Science Gloucester Point, Virginia 23062 Dr. Islay D. Marsden (1992) Department of Zoology Canterbury University Christchurch, New Zealand Dr. James Mason (1992) 1 Airyhall Terrace Aberdeen AB1 7QN Scotland, United Kingdom Dr. A. J. Paul (1992) Institute of Marine Science University of Alaska Seward Marine Center P.O. Box 730 Seward, Alaska 99664 Dr. Gilbert Pauley (1991) College of Fisheries University of Washington Seattle, Washington 98195 Dr. Les Watling (1991) Ira C. Darling Center University of Maine Walpole, Maine 04573 DENNIS J. CRISP, C.B.E., F.R.S. Professor Emeritus University of Wales, Bangor On January 18", 1990 Dennis Crisp, one of Great Britain’s best known marine biologists, died peacefully while under treatment. His remarkable tenacity and courage allowed him since 1968 to continue leading a productive life as a scientist despite an ever pervasive and unrelenting cancer of the lymphatic system, an affliction from which he had suffered for more than 20 years and which required repeated and frequent chemotherapy. Dennis’s attitude toward his research was dedicated and uncompromising and he expected the same high standards from those around him. Competence and imagination in research by his students and staff were invariably rewarded by his unqualified support. Dennis was a man of extraordinary intelligence and had a childlike inquisitiveness; every trip to a tide pool was like his first and he was genuinely excited at his finds. That excitement stayed with him throughout his career and he was on a constant quest to know and understand more. His contributions to comparative invertebrate physiology and marine ecology, in particular to the larval ecology of barnacles, are enduring contributions. His academic reputation was global and he followed in the footsteps of no less scientist than Charles Darwin. A part of Dennis’s success in his research could be attributed to the rigor imposed by his early training in physics and mathematics. He took his first degree in Zoology at Cambridge and then for, ‘consorting with the physical chemists’ was turned out of the Zoology Department by Sir James Gray and subsequently took his Ph.D. in the Department of Colloid Science. A complete chronicle of Dennis’ activities was written by his close friend and long-time colleague, Professor E. W. Knight- Jones on the occasion of Dennis’ 71st birthday (Knight Jones, 1987 in: Barnacle Biology, A. J. Southward (ed). Bal- kema, Rotterdam). He became a member of the National Shellfisheries Association in 1959 and became an honored life member in 1973 along with Drs. Daniel Quayle, J. C. Medcof, Gordon Gunter and Lyle St. Amant. In 1968 he was elected a Fellow of the Royal Society and in 1978 was awarded Commander of the British Empire (CBE) for services to marine biology. He was also Honorary Fellow of the Indian Academy of Sciences (1984) and Honorary Member of the i =P IN MEMORIAM: DENNIS J. CRISP American Society of Zoologist: | . 987). Of all these honors, Dennis held one particularly dear. It was his MBE— Marine Biologist Extraordinaire—p:~ nted to him by his students and colleagues at Menai Bridge. His vast number of publica- tions attests to his versatilii ad breadth of knowledge. The accompanying list is, of necessity, incomplete as Dennis was still actively engaged in varch and writing at the time of his death. Dennis had a large © slerance for incompetence and administrators. Rarely, if indeed ever, did he give way to what he regarded as stupidit’ officialdom. He disallowed interference from university administrators and seemingly was en- gaged endlessly in putes with accountants and taxmen. He had no use for laziness or bureaucratic intervention and was disinclined to svc fools gladly. His quixotic encounters with unprepared authority provided much affectionate amuse- ment to his stv. ats and colleagues. On the other hand, he took seriously his own administrative duties as Director of the Marine Biol y Station at the Univesity in Bangor and he did not use his position for his personal advantage among his colleagues id students. Denn: had an impish wit! He has been variously described as pugnacious, irascible, stubborn, energetic, and a deman- ‘ag, if not sometimes uncomfortable, colleague. He probably would have enjoyed such a characterization; one can imagine him grinning broadly upon hearing himself described thus. But he was also many other things—courageous, stimulating, highly intelligent, clever, and sometimes in retrospect very funny. Dennis received from his former students around the world their continued respect and admiration. One telling tribute was a retirement dinner attended by over 100 of his past and present students and colleagues, many from distant parts of the world. After his retirement he was often inclined to visit colleagues and former students. His students held for him not only gratitude but affection (often tinged with a little exasperation); one student has noted that, though his visits were often trying, they always left one with a sense of enlightenment and invigoration. Dennis seemed the epitome of the absent-minded professor. At his retirement dinner several years ago, he was con- fronted with a long list of items collected from all over the world, allegedly left behind or forgotten by him during his visit. To order his life he was known to write itemized lists which were known to include such items as—‘‘lunch, glasses, teeth, calculator, keys, comb hair, money, kiss wife.’’ Whether this seeming incompetence in the practical matters of daily. life was real or feigned is not entirely clear (one expects a little of each), but in any event there almost always seemed to be someone to the rescue. More often than not it was his wife Ella, who stood by him and kept him on the straight and narrow throughout their 45 years of married life. Anyone who has known Dennis will have their own stories. It happened that one of us contributed to a symposium volume for which Dennis was editor. It was with some trepidation that we coined the new term ‘‘teleplanic’’ larvae, which comes' from the Greek teleplanos meaning far-wandering. The manuscript came back with the usual editorial marks and queries and next to “‘teleplanic’’ Dennis had inserted in parentheses for his classically less literate colleague, the notation *‘(Aesculus, Prometheus bound, 1. 576).”’ On his visits to Woods Hole, Dennis would often grace our round kitchen table for a meal and hold forth with outrageous stories and opinions, some of which I reckon were calculated to arouse our indignation. He would on these occasions wear his special tie given to him by his daughter, dark blue and adorned with small, red pig heads and the monogram MCP (male-chauvinist pig)! On other occasions he would treat us to stories of scientists and laboratories known to most of us only through the literature. Dennis was indeed a remarkable man and a distinguished scientist. The world has lost a brilliant mind and many of us have lost a cherished friend. Rudolf S. Scheltema Woods Hole and Sandra E. Shumway Boothbay Harbor May 1990 il IN MEMORIAM: DENNIS J. CRISP PUBLICATIONS BY PROFESSOR D. J. CRISP 1946 Crisp, D. J., The orientation of macromolecules by interfaces. In J. B. Speakman & G. H. Elliott (eds.), Fibrous Proteins: 100-104. Leeds: Society of Dyers and Colourists. Crisp, D. J. Surface films of polymers. I. Films of the fluid type. J. Colloid Sci. 1:49—70. Crisp, J. D. Surface films of polymers. II. Films of the coherent and semi-crystalline type. J. Colloid Sci. 1:161—184. Crisp, J. D. Two-dimensional transport at fluid interfaces. Trans. Faraday Soc. 42:619-635. Forman, J. & D. J. Crisp. Radio-frequency absorption spectra of solutions of electrolytes. Trans. Faraday Soc. 42A:186—193. 1947 Crisp, D. J. Some applications of the surface film technique in polymer research. Proc. Eleventh Internat. Congr. Pure and Appl. Chem. London. Crisp, D. J. & W. H. Thorpe. A metal micro-respirometer of the Barcroft type suitable for small insects and other animals. J. Exp. Biol. 24:304—309. Thorpe, W. H. & D. J. Crisp. Studies on plastron respiration. I. The biology of Aphelocheirus (Hemiptera, Aphelocheiridae (Naucoridae)) and the mechanism of plastron respiration. J. Exp. Biol. 24:227—269. Thorpe, W. H. & D. J. Crisp. Studies on plastron respiration. II. The respiratory efficiency of the plastron in Aphelocheirus. J. Exp. Biol. 24:270—303. Thorpe, W. H. & D. J. Crisp. Studies on plastron respiration. III. The orientation response of Aphelocheirus (Hemiptera, Aphelocheiridae (Naucor- idae)) in relation to plastron respiration, together with an account of specialised pressure receptors in aquatic insects. J. Exp. Biol. 24:310—328. 1948 Crisp, D. J. Invasion of Eliminius modestus. Discovery July 1948:229. Crisp, D. J. & P. N. J. Chipperfield. Occurrence of Elminius modestus (Darwin) in British waters. Nature 161:64. Crisp, D. J. & W. H. Thorpe. The water-protecting properties of insect hairs. Discussions Faraday Soc. 3:210—220. 1949 Crisp, D. J. Adsorption of amines on negatively charged monolayers. Research, Suppl. on Surface Chemistry: 65-78. Crisp, D. J. A two-dimensional phase rule. I. Derivation of a two-dimensional phase rule for plane interfaces. Internat. Congr. Surface Chemistry, Bordeaux 1949:17—22. Crisp, D. J. A two-dimensional phase rule. II. Some applications of a two-dimensional phase rule for a single interface ongr. Surface Chemistry, Bordeaux 1949:23-35. | Marine Bic 1950 Crisp, D. J. Breeding and distribution of Chthamalus stellatus. Nature 166:311. Cnsp, D. J. The stability of structures at a fluid interface. Trans. Faraday Soc. 46:228-235. Crisp, D. J. & A. H. N. Molesworth. Habitat of Balanus amphitrite var. denticulata in Britain. Nature 167:489—490. Crisp, D. J. & W. H. Thorpe. A simpel replica technique suitable for the study of surface structures. Nature 165:273. 1951 Crisp, D. J. Antifouling. Pain, Oil and Colour J. Nov. 30 1951:1271. Barnes, H., D. J. Crisp. & H. T. Powell. Observations on the orientation of some species of barnacles. J. Anim. Ecol. 20:227—241. Norms, E., L. W. G. Jones, T. Lovegrove & D. J. Crisp. Variability in larval stages of cirripedes. Nature 167:444. 1952 Southward, A. J. & D. J. Crisp. Changes in the distribution of the intertidal barnacles in relation to the environment. Nature 170:416—417. 1953 Crisp, D. J. Changes in the orientation of barnacles of certain species in relation to water currents. J. Anim. Ecol. 22:331-343. Crisp, D. J. Marine biology in North Wales. Proc. Llandudno, Colwyn Bay and District Field Club, 1953:15—22. Crisp, D. J. Selection of site and position by some marine larvae. Br. J. Anim. Behav. 1:80-81. Crisp, D. J., L. W. G. Jones & W. Watson. Use of X-ray stereoscopy for examining shipworm infestation in vivo. Nature 172:408—409. Crisp, D. J. & E. W. Knight-Jones. The mechanism of aggregation in barnacle populations. A note on the recent contribution by Dr. H. Barnes. J. Anim. Ecol. 22:360—362. Crisp, D. J. & A. J. Southward. Isolation of intertidal animals by sea barriers. Nature 172:208—209. Knight-Jones, E. W. & D. J. Crisp. Gregariousness in barnacles in relation to the fouling of ships and to antifouling research. Nature 171:1109—1110. Norris, E. & D. J. Crisp. The distribution and planktonic stages of the cirripede Balanus perforatus Bruguiére. Proc. Zool. Soc. Lond. 123:393—409. 1954 Crisp, D. J. The breeding of Balanus porcatus (Da Costa) in the Irish Sea. J. Mar. Biol. Ass. U.K. 33:473—496. Crisp, D. J. & H. Bares. The orientation and distribution of barnacles at settlement with particular reference to surface contour. J. Anim. Ecol. 23:142-162. Crisp, D. J. & J. Hobart. A note on the habitat of the marine tardigrade Echiniscoides sigismundi (Schultze). Ann. Mag. Nat. Hist. Ser. 12, 7:554—560. Crisp, D. J. & E. W. Knight-Jones. Discontinuities in the distribution of shore animals in North Wales. Rept. Bardsey Observatory 1954:29—34. Jones, L. W. G. & D. J. Crisp. The larval stages of the barnacle Balanus improvisus Darwin. Proc. Zool. Soc. Lond. 123:765—780. Southward, A. J. & D. J. Crisp. The distribution of certain intertidal animals around the Irish coast. Proc. R. Irish Acad. 57B:1—29. Southward, A. J. & D. J. Crisp. Recent changes in the distribution of the intertidal barnacles Chthamalus stellatus Poli and Balanus balanoides L. in the British Isles. J. Anim. Ecol. 23:163-177. 1955 Crisp, D. J. The behaviour of barnacle cyprids in relation to water movement over a surface. J. Exp. Biol. 32:569—590. ili IN MEMORIAM: DENNIS J. CRISP Crisp, D. J. Factors governing the di ution of the littoral fauna. Challenger Soc. Ann. Rep. 3, 7:28. Crisp, D. J. & P. A. Davies. Obse: cons in vivo on the breeding of Elminius modestus grown on glass slides. J. Mar. Biol. Ass. U.K. 34:357—380. Crisp, D. J. & L. W. G. Jones omparison between the north and south coasts of the English Channel in regard to the distribution of the littoral fauna. Challenger Soc. Any — zp. 3, 7:27-28. Crisp, D. J. & A. J. Southw The distribution of certain intertidal animals on the coasts of Britain. Challenger Soc. Ann. Rep. 3, 7:26. 1956 Crisp, D. J. The adsorr 1 of alcohols and phenols from non-polar solvents on to alumina. J. Colloid Sci. 11:356—376. Crisp, D. J. The inte’ 1 zoology of Rockall. In J. Fisher (ed.), Rockall: 177-179. London: Geoffrey Bles. Crisp, D. J. A sub: ce promoting hatching and liberation of young in cirripedes. Nature 178:263. Crisp, D. J. & A. Southward. Demonstration of small scale water currents by means of milk. Nature 178:1076. Barnes, H. & T .. Crisp. Evidence of self-fertilization in certain species of barnacles. J. Mar. Biol. Ass. U.K. 35:631—369. Southward, 4 4. & D. J. Crisp. Fluctuations in the distribution and abundance of intertidal barnacles. J. Mar. Biol. Ass. U.K. 35:211—229. 1957 Crisp, Ps. Effect of low temperature on the breeding of marine animals. Nature 179:1138—1139. Crisp. ». J. Liberation and hatching of barnacle nauplii. Challenger Soc. Ann. Rep. 3, 9:19. Criss, D. J. & D. H. A. Marr. Energy relationships in physical toxicity. Proc. 2nd Internat. Congr. Surface Activity: 310—320. London: Butterworths. Crisp, D. J., C. P. Spencer & D. H. A. Marr. Toxicity of copper compounds in the sea. Challenger Soc. Ann. Rep. 3, 9:22. Cusp, D. J. & H. G. Stubbings. The orientation of barnacles to water currents. J. Anim. Ecol. 26:179—196. Bishop, M. W. H. & D. J. Crisp. The Australian barnacle in France. Nature 179:482—483. Bishop, M. W. H., D. J. Crisp, E. Fisher-Piette & M. Prenant. Sur lecologie des cirripédes de la cOte atlantique frangaise. Bull. Inst. Océanogr. Monaco 54(1099):1—12. De Bruin, G. P. H. & D. J. Crisp. The influence of pigment migration on vision in higher Crustacea. J. Exp. Biol. 34:447—463. Southward, A. J. & D. J. Crisp. Behaviour and feeding of cirripedes. Challenger Soc. Ann. Rep. 3, 9:20. 1958 Crisp, D. J. The spread of Elminius modestus Darwin in northwest Europe. J. Mar. Biol. Ass. U.K. 37:483—520. Crisp, D. J. Surface films of polymers. In J. F. Danielli, K. G. A. Pankhurst & A. C. Riddiford (eds.), Surface phenomena in chemistry and biology: 23-54. London: Pergamon Press. Crisp, D. J. & B. S. Pastel. The relationship between breeding and ecdysis in cirripedes. Nature 181:1078—1079. Crisp, D. J. & A. J. Southward. The distribution of intertidal organisms along the coasts of the English Channel. J. Mar. Biol. Ass. U.K. 37:157—208. Crisp, D. J. & C. P. Spencer. The control of the hatching process in barnacles. Proc. R. Soc. Lond. B 149:278—299. Austin, A. P., D. J. Crisp & A. M. Patel. The chromosome numbers of certain barnacles in British waters. Quart. J. Micr. Sci. 99:497—504. Bishop, M. W. H. & D. J. Crisp. The distribution of the barnacle Elminius modestus in France. Proc. Zool. Soc. Lond. 131:109—134. 1959 Crisp, D. J. Breeding and exuviation in Balanus balanoides. Proc. 15th Internat. Congr. Zool. Lond.: 298-300. Crisp, D. J. Factors influencing the time of breeding of Balanus balanoides. Oikos 10:275—289. Crisp, D. J. A further extension of Elminius modestus Darwin on the west coast of France. Beaufortia 7:37—39. Crisp, D. J. The influence of climatic changes on animals and plants. Geogr. J. 125:1-19. Crisp, D. J. The rate of development of Balanus balanoides (L.) embryos in vitro. J. Anim. Ecol. 28:119—132. Crisp, D. J. & E. Fischer-Piette. Répartition des principales espéces intercotidales de la céte atlantique frangaise en 1954—1955. Ann. Inst. Océanogr. Monaco 36:275—287. Crisp, D. J. & A. J. Southward. The further spread of Elminius modestus in the British Isles to 1959. J. Mar. Biol. Ass. U.K. 38:429—437. Crisp, D. J. & A. J. Southward. Recent changes in the distribution of marine organisms in northwest Europe. Proc. 1st Internat. Congr. Oceanography, New York 1959:148-151. Southward, A. J. & D. J. Crisp. Modes of cirral activity in barnacles. Proc. 15th Internat. Congr. Zool. London.: 295—296. 1960 Crisp, D. J. Factors influencing growth rate in Balanus balanoides. J. Anim. Ecol. 29:95—116. Crisp, D. J. Mobility of barnacles. Nature 188:1208—1209. Crisp, D. J. Northern limits of Elminius modestus in Britain. Nature 188:681. Crisp, D. J. & A. P. Austin. The action of copper in antifouling paints. Ann. Appl. Biol. 48:787—799. Crisp, D. J. & D. J. Clegg. The induction of the breeding condition in Balanus balanoides (L.). Oikos 11:265—275. Crisp, D. J. & B. S. Patel. The moulting cycle in Balanus balanoides L. Biol. Bull. 118:31—47. J Crisp, D. J. & J. S. Ryland. The influencing of filming and of surface texture on the settlement of marine organisms. Nature 185:119. Crisp, D. J. & G. B. Williams. The effect of extracts from fucoids in promoting settlement of epiphytic Polyzoa. Nature 188:1206—1207. Patel, B. & D. J. Crisp. The influence of temperature on the breeding and the moulting activities of some warm water species of operculate barnacles. J. Mar. Biol. Ass. U.K. 39:667—680. Patel, B. & D. J. Crisp. Rates of development of the embryos of several species of barnacles. Physiol. Zool. 33:104—119. 1961 Crisp, D. J. Territorial behaviour in barnacle settlement. J. Exp. Biol. 38:429—446. Crisp, D. J. & P. J. S. Boaden. Hypogean faunas. Review of Biologie des eaux souterraines, littorales et continentales. Nature 190:50. Crisp, D. J. & B. Patel. The interaction between breeding and growth rate in the barnacle Elminius modestus Darwin. Limnol. Oceanogr. 6:105—115. Crisp, D. J. & A. J. Southward. Different types of cirral activity of barnacles. Phil. Trans. R. Soc. B 243:271—308. Patel, B. & D. J. Crisp. Relation between the breeding and moulting cycles in cirripedes. Crustaceana 2:89—107. iv IN MEMORIAM: DENNIS J. CRISP 1962 Crisp, D. J. Grazing in terrestrial, freshwater and marine environments. Nature 194:1028—1029. Crisp, D. J. The larval stages of Balanus hameri (Ascanius, 1767) Crustaceana 4:123—130. Crisp, D. J. Marine Science Laboratories, Menai Bridge. Nature 195:549—551. Crisp, D. J. The planktonic stages of the Cirripedia Balanus balanoides and Balanus balanus (L.) from north temperate waters. Crustaceana 3:207—221. Crisp, D. J. Release of larvae by barnacles in response to the available food supply. Anim. Behav. 10:382—383. Crisp, D. J. Swarming of planktonic organisms. Nature 193:597—598. Crisp, D. J. & P. S. Meadows. The chemical basis of gregariousness in cirripedes. Proc. R. Soc. Lond. B 156:500—520. 1963 Crisp, D. J. Waterproofing mechanisms in animals and plants. In J. L. Moilliett (ed.), Waterproofing and water repellancy: 416-481. Amsterdam: Elsevier. Crisp, D. J. & J. D. Costlow. The tolerance of developing cirripede embryos to salinity and temperature. Oikos 14:22—34. Crisp, D. J. & P. S. Meadows. Adsorbed layers: the stimulus to settlement in barnacles. Proc. R. Soc. Lond. B 158:364—387. Crisp, D. J. & P. S. Meadows. The chemical basis of gregariousness in cirripedes. Anim. Behav. 11:2-3. Christie, A. O. & D. J. Crisp. A diffusion technique for assessing antifouling activity. Ann. Appl. Biol. 51:361—366. Southward, A. J. & D. J. Crisp. Catalogue of main marine fouling organisms 1: Barnacles. Paris, O.E.C.D. 1964 Crisp, D. J. An assessment of plankton grazing by barnacles. In D. J. Crisp (ed.), Grazing in terrestrial and marine environments: 251—264. Oxford: Blackwell. Crisp, D. J. The chemical basis of substrate selection by certain marine invertebrate larvae. Proc. 16th Internat. Congr. Zool. Washington D.C. 1963 1:58. Crisp, D. J. The effect of winter of 1962—63 on the British marine fauna. Helgoldnder wiss. Meeresunters. 10:313-—327. Crisp, D. J. (ed.) The effects of the severe winter of 1962/63 on marine life in Britain. J. Anim. Ecol. 33:165—210. Crisp, D. J. Introduction. In D. J. Crisp (ed.), Grazing in terrestrial and marine environments: XI—-XVI. Oxford: Blackwell. Crisp, D. J. Plastron respiration. In J. F. Danielli, K. G. A. Pankhurst & A. C. Riddiford (eds.), Recent progress in surface science 2:377—425. New York & London: Academic Press. Crisp, D. J. Racial differences between North American and European forms of Balanus balanoides. J. Mar. Biol. Ass. U.K. 44:33—45. 1965 Crisp, D. J. The ecology of marine fouling. In G. Goodman, R. W. Edwards & J. M. Lambert (eds.), Ecology and the industrial society: 99-117. Oxford: Blackwell. Crisp, D. J. The influence of adsorbed layers on larval settlement. Challenger Soc. Ann. Rep. 3, 17:30-31. Crisp, D. J. Observations on the effects of climate and weather on marine communities. In C. G. Johnson & L. P. Smith (eds.), The biological significance of climatic changes in Britain: 63—77. London & New York: Academic Press. Crisp, D. J. Surface chemistry, a factory in the settlement of marine invertebrate larvae. Botanica Gothoburgensia 3:51—65. Bhatnagar, K. M. & D. J. Crisp. The salinity tolerance of nauplius larvae of cirripedes. J. Anim. Ecol. 34:419—428. Christie, A. O. & D. J. Crisp. Toxicity of aliphatic amines to barnacle larvae. Comp. Biochem. Physiol. 18:59-69. Gray, J. S. & D. J. Crisp. Substrate selection in Protodrilus symbioticus. Challenger Soc. Ann. Rep. 3, 17:32. Kensler, C. B., K. M. Bhatnagar & D. J. Crisp. Distribution and ecological variation of Chthamalus species in the Mediterranean area. Vie et Milieu 16:271—293. Kensler, C. B. & D. J. Crisp. The colonization of artificial crevices by marine invertebrates. J. Anim. Ecol. 34:507-—516. Southward, A. J. & D. J. Crisp. Activity rhythms in barnacles in relation to respiration and feeding. J. Mar. Biol. Ass. U.K. 45:161—185. 1967 Crisp, D. J. Barnacles. Sci. J. 3:69-73. Crisp, D. J. Chemical factors inducing settlement in Crassostrea virginica Gmelin. J. Anim. Ecol. 36:329-335. Crisp, D. J. Chemoreception in cirripedes. Biol. Bull. 133:128—140. Crisp, D. J., J. H. Bailey & E. W. Knight-Jones. The tubeworm Spirorbis vitreus and its distribution in Britain. J. Mar. Biol. Ass. U.K. 47:511-521. Crisp, D. J., A. O. Christie & A. F. A. Ghobashy. Narcotic and toxic action of organic compounds on barnacle larvae. Comp. Biochem. Physiol. 22:629-649. Crisp, D. J. & B. Patel. The influence of the surface contour of the substratum on the shapes of barnacles. Proceedings of the symposium on Crustacea, Ernakulam 2:612—629. Bangalore: Marine Biological Association of India. Crisp, D. J. & D. A. Ritz. Changes in temperature tolerance of Balanus balanoides during its life-cycle. Helgoldnder wiss. Meeresunters. 15:98—115. Buchan, S., G. D. Floodgate & D. J. Crisp. Studies on the seasonal variation of the suspended matter in the Menai Straits. 1. The inorganic fraction. Limnol. Oceanogr. 12:419—431. Christie, A. O. & D. J. Crisp. Activity coefficients of the n-primary, secondary and tertiary aliphatic amines in aqueous solution. J. Appl. Chem. 17:11-14. 1968 Crisp, D. J. Differences between North American and European populations of Balanus balanoides revealed by transplantation. J. Fish. Res. Board Canada 25:2633-2641. ? Crisp, D. J. Distribution of the parasitic isopod Hemioniscus balani with special reference to the east coast of North America. J. Fish. Res. Board Canada 25:1161—1167. Crisp, D. J. & D. A. Ritz. Temperature acclimation in barnacles. J. Exp. Mar. Biol. Ecol. 1:236—256. v IN MEMORIAM: DENNIS J. CRISP 1969 Crisp, D. J. Studies of barnacles h: ug substance Comp. Biochem. Physiol. 30:1037—1048. Crisp, D. J. & B. Patel. Environ’ .al control of the breeding of three boreo-arctic cirripedes. Mar. Biol. 2:283—295. 1970 Castilla, J. C. & D. J. Cris .esponses of Asterias rubens to olfactory stimuli. J. Mar. Biol. Ass. U.K. 50:829—847. Ritz, D. A. & D. J. Crisr _easonal changes in feeding rate in Balanus balanoides. J. Mar. Biol. Ass. U.K. 50:223—240. 1971 Crisp, D. J. Energy .2w measurements. In N. A. Holme & A. D. Mclntyre (eds.), Methods for the study of marine benthos: 197—279. Oxford: Blackwell. Crisp, D. J. (ed — ourth European Marine Biology Symposium. Cambridge: University Press. Crisp, D. J. & .. F. A. A. Ghobashy. Responses of the larvae of Diplosoma listerianum to light and gravity. In D. J. Crisp (ed.), Fourth European Marine ' Jlogy Symposium: 443—465. Cambridge: University Press. Crisp, D.. & R. Williams. Direct measurement of pore-size distribution on artificial and natural deposits and prediction of pore space accessible to inter .dal organisms. Mar. Biol. 10:214—226. Forbes L., M. J. B. Seward & D. J. Crisp. Orientation to light and the shading response in barnacles. In D. J. Crisp (ed.), Fourth European Marine FP ology Symposium: 539-558. Cambridge: University Press. 1973 Crisp, D. J. Mechanism of adhesion of fouling organisms. In R. F. Acker, B. Floyd Brown, J. R. de Palma & W. P. Iverson (eds.), Proc. 3rd Internat. Congr. Mar. Corrosion & Fouling: 691—709. Gaithersburg: U.S. National Bureau of Standards. Crisp, D. J. The role of the biologist in anti-fouling research. In R. F. Acker, B. Floyd Brown, J. R. de Palma & W. P. Iverson (eds.), Proc. 3rd Internat. Congr. Mar. Corrosion & Fouling: 88—93. Gaithersburg: U.S. National Bureau of Standards. Crisp, D. J. & V. L. M. Klein. Contributions to the knowledge of Philometra lateolabracis Yamaguti, 1935 (Nematoda, Filarioidea). Mem. Inst. Oswaldo Cruz, Rio de Janeiro 71:481—484. Crisp, D. J. & D. A. Ritz. Responses of cirriped larvae to light. 1. Experiments with white light. Mar. Biol. 23:327—335. Buchan, S., G. D. Floodgate & D. J. Crisp. Studies of the seasonal variation of the suspended matter of the Menai Straits. 2. Midstream data. Dt. hydr. Z. 26:74-83. Castilla, J. C. & D. J. Crisp. Responses of Asterias rubens L. to water currents and their modification by certain environmental factors. Neth. J. Sea Res. 7:171-190. 1974 Crisp, D. J. Energy relations of marine invertebrate larvae. Thalassia Jugoslavica 10:103—120. Crisp, D. J. Factors influencing the settlement of marine invertebrate larvae. In P. T. Grant & A. M. Mackie (eds), Chemoreception in marine organisms: 177—265. London: Academic Press. 1975 Crisp, D. J. Secondary productivity in the sea. Productivity of world ecosystems: 71-89. Seattle: U.S. National Academy of Sciences. Crisp, D. J. Surface chemistry and life in the sea. Chemistry and Industry 5:187—193. Crisp, D. J. & C. A. Richardson. Tidally produced internal bands in the shell of Elminius modestus. Mar. Biol. 33:155—160. Bourget, E. & D. J. Crisp. An analysis of the growth bands and ridges of barnacle shell plates. J. Mar. Biol. Ass. U.K. 55:439—461. Bourget, E. & D. J. Crisp. Early changes in the shell form of Balanus balanoides (L.). J. Exp. Mar. Biol. Ecol. 17:221—237. Bourget, E. & D. J. Crisp. Factors affecting deposition of the shell in Balanus balanoides (L.). J. Mar. Biol. Ass. U.K. 55:231—248. Flowerdew, M. W. & D. J. Crisp. Esterase heterogeneity and an investigation into racial differences in the cirripede Balanus balanoides (L.). using acrylamide gel electrophoresis. Mar. Biol. 33:33—39. Walker, G., P. S. Rainbow, P. Foster & D. J. Crisp. Barnacles: possible indicators of zinc pollution? Mar. Biol. 30:57—65. 1976 Crisp, D. J. The British contribution to the IBP programme on marine productivity. Phil. Trans. R. Soc. B. 274:393—399. Crisp, D. J. Prospects of marine science in the Gulf area—the background paper UNESCO Tech. Pap. Mar. Sci. 26:19-38. Crisp, D. J. The role of the pelagic larva. In P. Spencer Davies (ed.), Perspectives in experimental biology 1:145—155. Oxford & New York: Pergamon Press. Crisp, D. J. Settlement responses in marine organisms. In R. C. Newell (ed.), Adaptations to the environment: 83—124. London: Butterworth. Flowerdew, M. W. & D. J. Crisp. Allelic esterase isozymes, their variation with season, position on the shore and stage of development in the cirripede Balanus balanoides. Mar. Biol. 35:319—325. Hildreth, D. I. & D. J. Crisp. A corrected formula of calculation of filtration rate of bivalve molluscs in an experimental flowing system. J. Mar. Biol. Ass. U.K. 56:111—120. Hughes, R. N. & D. J. Crisp. A further description of the echiuran Prashadus pirotansis. J. Zool. Lond. 180:233—242. 1977 Crisp, D. J., J. Davenport & P. A. Gabbott. Freezing tolerance in Balanus balanoides. Comp. Biochem. Physiol. 57A:359—361. 1978 Crisp, D. J. Genetic consequences of difference reproductive strategies in marine invertebrates. In B. Battaglia & J. A. Beardmore (eds.), Marine organisms: genetics, ecology and evolution: 257-273. New York: Plenum Press. Crisp, D. J., A. R. Beaumont, M. W. Flowerdew & A. Vardy. The Hardy-Weinberg Test—a correction. Mar. Biol. 46:181—183. vi IN MEMORIAM: DENNIS J. CRISP 1979 Crisp, D. J. Dispersal and reaggregation in sessile marine invertebrates, particularly barnacles. Syst. Ass. Spec. Vel. 11:319—328. Barnett, B. E. & D. J. Crisp. Laboratory studies of gregarious settlement in Balanus balanoides and Elminius modestus in relation to competition between these species. J. Mar. Biol. Ass. U.K. 59:581—590. Barnett, B. E., S. C. Edwards & D. J. Crisp. A field study of settlement behaviour in Balanus balanoides and Elminius modestus (Cirripedia, Crus- tacea) in relation to competition between them. J. Mar. Biol. Ass. U.K. 59:575—580. Dando, P. R., A. J. Southward & D. J. Crisp. Enzyme variation in Chthamalus stellatus and Chthamalus montagui (Crustacea: Cirripedia): Evidence for the presence of C. montagui in the Adnatic. J. Mar. Biol. Ass. U.K. 59:307—320. Grenon, J. F., J. Elias, J. Moorcroft & D. J. Crisp. A new apparatus for force measurements in marine bioadhesion. Mar. Biol. 53:381—388. Lucas, M. I., G. Walker, D. L. Holland & D. J. Crisp. An energy budget for the free-swimming and metamorphosing larva of Balanus balanoides (Crustacea: Cirmipedia). Mar. Biol. 55:221—229. Richardson, C. A., N. W. Runham & D. J. Crisp. Tidally deposited growth bands in the shell of the common cockle, Cerastoderma edule (L.). Malacologia 18:277—290. 1980 Crisp, D. J. Extending conservation seawards. In B. Patel (ed.), Management of the environment: 96—103. New Delhi: Wiley (Eastern) Publications. Crisp, D. J. How marine organisms deal with oil contamination. Oceanology international 80, Tech. Sess. M: 34—41 Brighton: BPS Exhibitions Ltd. Crisp, D. J. Management of the environment. In J. Sharma (ed.), Nuclear India: 1—8. Indian Dept. Atomic Energy Publ. 18/7. Bombay: Tata Press. Richardson, C. A., D. J. Crisp & N. W. Runham. An endogenous rhythm in shell deposition in Cerastoderma edule. J. Mar. Biol. Ass. U.K. 60:991—1004. Richardson, C. A., D. J. Crisp & N. W. Runham. Factors influencing shell growth in Cerastoderma edule. Proc. R. Soc. Lond. B 210:513-531. Richardson, C. A., D. J. Crisp, N. W. Runham & L. D. Gruffydd. The use of tidal growth bands in the shell of Cerastoderma edule to measure seasonal growth rates under cool temperate and sub-arctic conditions. J. Mar. Biol. Ass. U.K. 60:977—989. 1981 Crisp, D. J. General considerations. Marine fouling of offshore structures 1(1):1—3 Society for Underwater Technology. Crisp, D. J., A. J. Southward & E. C. Southward. On the distribution of the intertidal barnacles Chthamalus stellatus, Chthamalus montagui and Euraphia depressa. J. Mar. Biol. Ass. U.K. 61:359—380. Dalley, R. & D. J. Crisp. Conchoderma: a fouling hazard to ships under way. Mar. Biol. Letts. 2:141—152. Richardson, C. A., D. J. Crisp & N. W. Runham. Factors influencing shell deposition during a tidal cycle in the intertidal bivalve Cerastoderma edule. J. Mar. Biol. Ass. U.K. 61:465—476. Richardson, C. A., N. W. Runnam & D. J. Crisp. A histological and ultrastructural study of the cells of the mantle edge of a marine bivalve, Cerastoderma edule. Tissue & Cell 13:715—730. Young, G. A. & D. J. Crisp. Marine animals and adhesion. In K. W. Allen (ed.), Adhesion: 19-39. London: Applied Science Publishers. 1982 Clare, A. S., G. Walker, D. L. Holland & D. J. Crisp. Barnacle egg hatching: a novel role for a prostaglandin-like compound. Mar. Biol. Letts. 3:113-120. Crisp, D. J. Cold tolerance in Balanus balanoides. In: Symposium on Invertebrate Cold Hardiness, Oslo, Norway, 1982. Summaries of Papers. Crisp, D. J. Freezing tolerance in the intertidal invertebrates. In: Symposium on Invertebrates Cold Hardiness, Oslo, Norway, 1982. Summaries and Papers. Ekaratne, K., A. H. Burfitt, M. W. Flowerdew & D. J. Crisp. Separation of two Atlantic species of Pomatoceros, P. lamarkii and P. triqueter (Annelida: Serpulidea) by menas of biochemical genetics. Mar. Biol. 71:257—264. Ekaratne, S. U. K. & D. J. Crisp. Tidal micro-growth bands in intertidal gastropod shells, with an evaluation of band-dating techniques. Proc. R. Soc. Lond. B 214:305—323. Lane, D. J. W.,J. A. Nott & D. J. Crisp. Enlarged stem glands in foot of the post-larval mussel, Mytilus edulis: adaptation for bysso-pelagic migration. J. Mar. Biol. Ass. U.K. 62:809-818. Manahan, D. T. & D. J. Crisp. The role of dissolved organic material in the nutrition of pelagic larvae: amino acid uptake by bivalve veligers. Amer. Zool. 22:635—646. Yule, A. B., D. J. Crisp & I. H. Cotton. The action of acetozolamide on calcification in Balanus balanoides. Mar. Biol. Letts. 3:273—278. 1983 Crisp, D. J. Chelonibia patula (Ranznai), a pointer to the evolution of the complemental male. Mar. Biol. Letts. 4:281—294. Crisp, D. J. Extending Darwin’s investigations on the barnacle life-history. Biol. J. Linn. Soc. 20:73-83. Crisp, D. J., A. Burfitt, K. Rodrigues & M. D. Budd. Lasaea rubra—a apomictic bivalve. Mar. Biol. Letts. 4:127—136. Ekaratne, S. U. K. & D. J. Crisp. A geometric analysis of growth in gastropod shells with particular reference to turbinate forms. J. Mar. Biol. Ass. U.K. 63:777-797. Holland, D. L., D. J. Crisp & J. East. Changes in the fatty-acid composition of the Ocean-strider, Halobates fijiensis (Heteroptera, Gerridae) after starvation. Mar. Biol. Letts. 4:259—265. Manahan, D. T. & D. J. Crisp. Autoradiographic studies on the uptake of dissolved amino acids from sea water by bivalve larvae. J. Mar. Biol. Ass. U.K. 63:673—-682. Yule, A. B. & D. J. Crisp. Adhesion of cypris larvae of Balanus balanoides to clean and arthropodin treated surfaces. J. Mar. Biol. Ass. U.K. 63:261—271. Yule, A. B. & D. J. Crisp. A study of feeding behaviour in Temora longicornis (Muller) (Crustacea: copepoda). J. Exp. Mar. Biol. Ecol. 71:271—282. 1984 Crisp, D. J. Energy flow measurements. In N. A. Holme & A. D. McIntyre (eds.). Methods for the study of the marine benthos 2nd edition: 284—373. Oxford: Blackwell. Vii IN MEMORIAM: DENNIS J. CRISP Crisp, D. J. Overview of research on ne invertebrate larvae, 1940-1980. In J. D. Costlow & R. C. Tipper (eds.), Marine biodeterioration: an interdisciplinary study: 103—126 —~adon: Spon. Crisp, D. J. & K. Ekaratne. Polyr = aism in Pomatoceros. Zool. J. Linn. Soc. 80:157—175. Crisp, D. J. & A. H. Lewis. Fa’; controlling cold tolerance in and breeding in Balanus balanoides. J. Mar. Biol. Ass. U.K. 64:125—145. Crisp, D. J.,G. Walker,G. A vung & A. B. Yule. Adhesion and substrate choice in mussels and barnacles. J. Colloid and Interface Sci. 104:40—50. Ekaratne, S. U.K. & D. J. p. Seasonal growth studies of intertidal gastropods from shell micro-growth band measurements, including a comparison with alternative metho’ /. Mar. Biol. Ass. U.K. 64:183—210. Gruffydd, L. D.,R. Hux &D. J. Crisp. The reduction in growth of Mytilus edulis in fluctuating salinity regimes and the exaggeration of the effect by using tap water as diluting medium. J. Mar. Biol. Ass. U.K. 64:401—409. Holland, D.C., D. Crisp, R. Huxley & J. Sisson. Influence of oil shale on intertidal organisms: effect of oil shale extract on settlement of the barnacle Balan balanoides (L.). J. Exp. Mar. Biol. Ecol. 75:245-—255. Huxley, R., D. —_ Holland, D. J. Crisp & R. S. L. Smith. Influence of oil shale on intertidal organisms: effect of oil shale surface roughness on settlement the barnacle Balanus balanoides. J. Exp. Mar. Biol. Ecol. 82:231—238. 1985 Clare, A >»., G. W. Walker, D. L. Holland & D. J. Crisp. The hatching substance of the barnacle Balanus balanoides. Proc. R. Soc. Lond. B. Biol. Sci -24:131-148. Crisp D. J., A. B. Yule & K. N. White. Feeding by oyster larvae: the functional response, energy budget and a comparison with mussel larvae. J. Mar. iol. Ass. U.K. 65:759-783. Cxisp, D. J. Recruitment of barnacle larvae from the plankton. Bull. Marine Science (International Symposium sponsored by Scripps Inst. Oceanog- raphy, Plankton Soc. Japan and Western Soc. of Naturalists, Shimizu, Japan 1984). 37:478—486. Crisp, D. J. & E. Bourget. Growth in barnacles. Advances in Marine Biology. 22:199-—244. Crisp, D. J., G. Walker, G. A. Young & A. B. Yule. Adhesion and substrate choice in mussels and barnacles. J. Colloid and Interface Sci. 104:40—S0. Gill, C. W. & D. J. Crisp. The effect of size and temperature on the frequency of limb beat of Temora longicornis (Crustacea: copepoda). J. Exp. Mar. Biol. Ecol. 86:185—196. Gill, C. W. & D. J. Crisp. Sensitivity of intact and antennule amputated copepods Temora longicornis to water disturbance. Mar. Ecol. Prog. Ser. 21:221-228. 1986 Crisp, D. J. A comparison between the reproduction of high and low-latitude barnacles, including Balanus balanoides and Tetraclita (Tesseropora) pacifica. In. M.-F. Thompson, R. Sarojini & R. Nagabhushanam (eds.), Biology of Benthic Marine Organisms, pp. 69-84. Rotterdam: A. A. Balkema. Holland, D. L., R. Huxley, E. M. Hill & D. J. Crisp. The effect of the blackstone oil shale exposure on intertidal organisms at Clavell’s Hard, Kimmeridge, Dorset, UK: A review of an ecological and experimental study. Proc. Dorset Nat. Hist. Archaeol. Soc. 107:135—139. 1987 Crisp, D. J. Reduced discrimination of laboratory-reared cyprids of the barnacle Balanus amphitrite amphitrite Darwin, Crustacea Cirripedia, with a description of a common abnormality. In M.-F. Thompson, R. Sarojini & R. Nagabhushanam (eds.), Marine Biodeterioration. New Delhi: Oxford. Crisp, D. J. On the sizes and shapes of barnacle eggs. Contributions in Marine Sciences, Dr. S. Z. Qasim Sastyabdapurti felicitation volume, pp. 1—26. Crisp, D. J. The effects of tidal fronts on intertidal faunas. In: 22nd European Marine Biology Symposium— Abstracts (full paper in press), Barcelona 17-22 August, 1987. Crisp, D. J. & B. Mwaiseje. Diversity in marine ecosystems, with special reference to the corallina community. In: 22nd European Marine Biology Symposium— Abstracts (full paper in press). Barcelona 17—22 August, 1987. Carvalho, G. R. & D. J. Crisp. The clonal ecology of Daphnia magna (Crustacea: Cladocera). 1. Temporal changes in the clonal structure of a natural population. J. Animal Ecol. 56:453—468. Furman, E. R., D. J. Crisp & A. Yule. Gene flow between Balanus improvisus (Cirripedia) populations in British estuaries. In: 22nd European Marine Biology Symposium— Abstracts (full paper in press). Barcelona 17—22 August, 1987. Lucas, M. I. & D. J. Crisp. Energy metabolism of eggs during embryogenesis in Balanus balanoides. J. Mar. Biol. Ass. U.K. 67:27—54. 1988 Crisp, D. J. Reduced discrimination of laboratory reared cyprids of the barnacle Balanus amphitrite with a description of a common abnormality. In ‘Marine Biodeterioration’’ Ed. Mary-Frances Thompson, Rachakonda Sarojini & Rachakonda Nagabhushanam, pp. 409-432. Oxford & IBH Publ. Co. New Delhi. Crisp, D. J. Tidally deposited bands in bivalves and barnacles. In R. Crick (ed.), Origin Evolution and Modern Aspects of Biomineralisation in Plants and Animals, 562 pp. New York: Plenum. Crisp, D. J. & A. Standen. Lasaea rubra (Montagu) (Bivalvia: Erycinacea), an apomictic crevice-liking bivalve with clones separated by tidal level preference. J. Exp. Mar. Biol. Ecol. 117:27—45. Crisp, D. J. & G. E. Fogg. Taxonomic instability continues to irritate. Nature 335:120—121. 1989 Furman, E. R. & D. J. Crisp. Biometrical changes during growth of isolated individuals of Balanus improvisus. J. Mar. Biol. Ass. U.K. 69:511—522. Tyler-Walters, H. & D. J. Crisp. The modes of reproduction in Lasaea rubra (Monatgu) and L. australis (Lamarck): (Erycinidea; Bivalvia). In R. S. Ryland & P. A. Tyler (eds.), Reproduction, Genetics and Distributions of Marine Organisms, pp. 299-308. Denmark: Olsen & Olsen. 1990 Foltz, D. W., S. E. Shumway & D. J. Crisp. Relationships among metabolic rate, weight and heterozygosity in the marine snail, Littorina littorea. Submitted. Viil IN MEMORIAM: DENNIS J. CRISP Shumway, S. E. & D. J. Crisp. **Specific dynamic action’’ demonstrated in herbivorous marine periwinkles, Littorina littorea and Littorina obtusata (Mollusca, Gastropoda) with a description of food preferences in New England populations. Submitted. Shumway, S. E., R. C. Newell, D. J. Crisp & T. L. Cucci. Particle selection in filter-feeding bivalve molluscs: A new technique on an old theme. In B. Morton (ed.), The Bivalvia. Proceedings of a Symposium in Memory of Sir Charles Maurice Yonge, Edinburgh, 1986. Hong Kong: Hong Kong University Press. Crisp, D. J. Gregariousness and systematic affinity in some North Carolinian barnacles. Bulletin of Marine Science (in press). Crisp, D. J. Field experiments in the settlement, orientation and habitat choice of Chthamalus fragilis (Darwin). Biofouling 2 (in press). Crisp, D. J. & F. J. Maclean. The relationship between the dimensions of the cirral net, the beat frequency and the size and age of the animal in Balanus balanoides and Elminius modestus. J.M.B.A. 70(3) (in press). Crisp, D. J. Marine Biology Today. Collagium in Japan (in press). Crisp, D. J., E. M. Hill & D. L. Holland. The hatching mechanism in barnacles. Crustacea Issues 7, Wenner A. & A. Kuris (eds.) (in press). Crisp, D. J., J. Weighall & C. A. Richardson. Tidal band in Siphoneria gigas. Submitted Crisp, D. J. & B. Mwaiseje. The ecology of tide pools in North Wales, South East Scotland & South West Ireland. Crisp, D. J. The influence of tidal fronts on the distribution of intertidal fauna and flora. Submitted or fo to 7 = a _ > : : : =a om *4 a oe 7 we , a oe, j ot nh rma. | ae tt Meee? Oe 23} a: a ee estan V0 pam ah ae ert : ny e ee x : 7 = % 7 ay a i ey © =m Ls é = al - iRPee |) es r= cs it a ; F mI >» > . if = ; ie ~ x Me 4 2 ¥F ’ a > —. a : : , tratece = & Pr Journal of Shellfish Research, Vol. 9, No. 1, 1-28, 1990. THE ROLE OF NUTRITION IN MATURATION, REPRODUCTION AND EMBRYONIC DEVELOPMENT OF DECAPOD CRUSTACEANS: A REVIEW KIM E. HARRISON Department of Biology Dalhousie University Halifax, Nova Scotia Canada B3H 4J1 ABSTRACT Reproduction in crustaceans entails maternal mobilization, biosynthesis and bioaccumulation of materials for export as self-sufficient capsules, the unfertilized eggs (from fewer than 50 to as many as 3 million per spawn). Once fertilized with genetic information from the sperm, the contents of each egg must wholly support the development of the embryo and the first larval stages until molting or metamorphosis into a ‘‘first-feeding’’ larva. This lecithotrophic strategy (reliance, by the embryo and newly-hatched larvae, on the egg yolk for complete nutrition) is fundamental to understanding the nutrient requirements of broodstock. Maternal diets must be augmented to provide sufficient energy and appropriate nutrients to meet the metabolic costs of reproduction (e.g. for the manufacture of gonadal tissue), as well as to provide and replace all essential nutrients and energy lost to the eggs (reproductive output). Despite the commercial value of providing optimal nutrition to crustacean broodstock, there is limited understanding of nutrition-reproduction interactions or knowledge of specific nutrient requirements for reproduction. The primary objectives of a commercial crustacean broodstock diet should include a capacity to support: 1) promotion (or induction) of sexual maturation, 2) enhancement of fertility and promotion of mating, and 3) increased fecundity by improving egg quality, egg quantity, and viability of offspring. The primary objectives of this review are: 1) to provide a background on crustacean reproductive biology, from a nutrient metabolism perspective, 2) to summarize published data pertaining to specific nutrient metabolism and the possible nutrient require- ments of reproduction in crustaceans, 3) to present a summary model of the nutrient metabolism during ovarian maturation in crustaceans, 4) to give a historical perspective and summarize current diets and nutritional practices for crustacean broodstock, and 5) to identify gaps in our knowledge of specific nutrient requirements and to make research recommendations. KEY WORDS: crustacean nutrition, reproduction, broodstock, vitellogenesis, maturation, embryogenesis, lecithotrophy, eyestalk ablation, EFA, PUFA; Note: Abbreviations are listed at end of references. 1.0 INTRODUCTION Nutrition is an important factor influencing maturation! in wild and captive crustacean broodstock. Despite ad- vances in commercial formulations and the availability of broodstock-specific diets that began in the mid 1980's, ade- quate, reliable maturation diets are not yet available. There is still a reliance upon diet supplementation with fresh or fresh-frozen natural foods (e.g. squid, marine worms, bi- valves, etc.) to provide adequate broodstock nutrition and to promote maturation (pers. comm. R. P. McIntosh 1989 and J. Ogle 1989). Research on crustacean broodstock diets began con- certedly during the past decade, with the escalating demand for controlled reproduction in commercial facilities. Pre- viously, shrimp production was based entirely on wild seed stock (postlarvae captured from the wild) or ‘‘sourcing”’ impregnated, ready-to-spawn females from the wild to stock hatcheries with larvae. Once techniques for inducing maturation became identified, diet composition became a critical factor in maintaining broodstock from onset of mat- uration to spawning. These techniques include eyestalk ab- lation? (Kulkarni and Nagabhushanam 1980; Quacken- ‘In this review, maturation refers to sexual development and not general (somatic) growth and development to adulthood. ?Ablation entails removal of one (unilateral) or both (bilateral) eye- stalk(s), or contents therein; more specifically, enucleation refers to re- 1 bush and Herrnkind 1981; Sandifer 1986 for review; Ma- kinouchi and Primavera 1987), hormone injection (summary by Yano 1985, 1987), hormone application by thoracic ganglion implant from maturing female crusta- ceans (Yano et al. 1988), and control of environmental cues such as photoperiod (Palaemon, Ryckaert et al. 1974; Penaeus, Lumare 1979; Beard and Wickins 1980; Hor- marus, Nelson et al. 1983, 1988a,b), light quality (Em- merson et al. 1983), light intensity (Chamberlain and Lawrence 1981a; Wurts and Stickney 1984), water temper- ature (Lumare 1981; Lee and Fielder 1982), and tempera- ture cycling (Nelson et al. 1988a,b). Other exogenous factors include water depth and tank size or adequate space (Yano 1985), and tank color (Brown et al. 1979), all of which may relate to ambient light intensity (Wurts and Stickney 1984), and appropriate substratum (eg. earth, sand or mud to accommodate burrowing shrimp such as Penaeus monodon) (AQUACOP 1977). Understanding nu- trition-reproduction interactions and determining the nu- trient requirements for successful maturation and spawning are needed to enable year-round, large volume hatchery production of postlarval crustaceans for grow-out opera- tions world-wide. Additionally, determining the role of specific nutrients (and natural dietary sources) in reproduc- tion, and the shifts in the dynamics of deposition, mobili- moval of the eyestalk contents (the sinus gland complex) while extirpation is the removal of the entire eyestalk. i) , should advance the un- eproductive physiology and 28. Moore (1985) briefly ad- . broodstock nutrition within the context of an overview the role of feeds and feeding in aquatic animal produ on. Reliable data o che nutrient requirements specific to maturation, repre .ction and embryogenesis in crustaceans are scant and fr ,mentary. Holland (1978) reviewed energy reserves and .ochemical changes in egg and larval devel- opment of } athic marine invertebrates, and Clarke and co- workers ( 385) examined lipid composition of the eggs of amphip ds and patterns of lipid utilization during embryo- genes s. The few published studies concerning decapod broc dstock nutrition address only natural or practical diets. Experiments based on semipurified, nutrient-defined diets have not been reported. A summary of sources of important circumstantial data concerning nutrition-biochemistry of reproduction in decapods is presented in this review. Since the 1970’s, several studies on inducing maturation in decapod crustaceans have been conducted and provide some information on diet composition and the testing of natural feed ingredients (Table 1). Information on brood- stock nutrient requirements can also be deduced from re- search on the compositional changes as associated with maturation of the parental organs, especially the gonads and hepatopancreas?, and from the composition of the spermatophores, eggs, and larvae from wild caught animals (Pillay and Nair 1973; Read and Caulton 1980; Teshima and Kanazawa 1983; Clarke et al. 1985; Teshima et al. 1989; Jeckel et al. 1989a,b; Castille and Lawrence 1989). Other researchers have documented the changes in compo- sition with induced maturation (Teshima et al. 1988b). In most of these studies, however, nutrition has not been con- trolled. Other clues to nutrient requirements of broodstock are provided by research on nutrient metabolism during maturation (e.g. lipid metabolism) (Kanazawa et al. 1988; Teshima et al. 1988a), and on the endocrinology of repro- duction and vitellogenesis (Charniaux-Cotton 1980, 1985; Quackenbush 1989). The overall knowledge of crustacean nutrient metabo- lism is limited. Nutrient requirements have most recently been published by the National Research Council (1983) for warmwater crustaceans; Conklin (1982) and D’ Abramo and Conklin (1985), for Homarus; Sandifer and Smith (1985) for Macrobrachium; and Akiyama and Dominy (1989) for Penaeus. Nevertheless, an incomplete founda- tion for understanding maternal needs independent of re- production remains. This is the first review of crustacean broodstock nutritional biochemistry. The primary objectives of this review are: 1) to provide zation, and utilization of nutrie derstanding of both crustacee crustacean life cycle strate dressed the ‘*bottleneck’’ *The terms hepatopancreas, digestive gland and midgut gland are used interchangeably in the literature. Hepatopancreas is most commonly used in the papers reviewed on this topic, and has been selected for this paper according to the arguments of Gibson and Barker (1979). HARRISON a background on crustacean reproductive biology, from a nutrition-biochemistry perspective, 2) to summarize pub- lished data pertaining to specific nutrient requirements for reproduction in crustaceans, 3) to present a model of the nutrient metabolism during ovarian maturation in crusta- ceans, 4) to provide a historical perspective and summarize current nutritional practices and commercial diet formula- tions for crustacean broodstock, and 5) to identify gaps in our knowledge of specific nutrient requirements and ac- cordingly recommend areas for research. 2.0 THE CRUSTACEAN REPRODUCTIVE CYCLE: A NUTRITIONAL PERSPECTIVE In decapod crustaceans, the natural age at first (sexual) maturity ranges from 4 months to 6 years (AQUACOP 1975, 1979; Wickins 1982). Although successful ovarian maturation can be induced in prepubescent shrimp, puberty may be delayed due to nutrient deficiencies in the early life stages, as has been described for fish (Luquet and Wa- tanabe 1986). Additionally, nutritional deficiency during prepubescence and puberty may impair health and fertility of broodstock and result in reduced egg number, smaller egg size, altered egg composition and reduced egg hatch- ability. There are no available data, however, on the impact of nutritional deficiencies prior to onset of maturity on re- productive success in crustaceans, although it is known that animals fed poor quality compound diets do not reproduce (pers. comm. Ceccaldi). This is an especially important factor in the maintenance and culture of animals for long term genetics and breeding programs. As discussion of broodstock nutrition requires an under- standing of reproductive physiology and endocrinology, a synopsis of the tissues and processes involved in the repro- ductive cycle of crustaceans follows. Specific nutrient interactions are discussed more fully in Section 3. 2.1 Reproductive Endocrinology In Decapods, the sexes are usually separate and the paired gonads are located dorsally and laterally to the gut. They are often surrounded by midgut cecae or lobes of the hepatopancreas. Each ovary or testis has an oviduct or vas deferens leading to separate gonopores. Crustacean reproduction and ovarian maturation are reg- ulated by steroid, peptide, and terpenoid hormones (Quackenbush 1986, for review; Van Beek and De Loof 1988; Tobe et al. 1989). The endocrine mechanisms of crustaceans, including endocrine structures and roles of hormones in reproduction and molting, are reviewed by Fingerman (1987). Endocrine control of vitellogenesis is reviewed by Meusy and Charniaux-Cotton (1984) and Charniaux-Cotton (1980, 1985). Circumstantial evidence *Ecdysteroids are a class of compounds derived from ecdysone (a polyhy- droxylated steroid derivative of cholesterol) which has hormonal activity (especially in the activation of molting) in arthropods (for review, see Chang, 1989). CRUSTACEAN BROODSTOCK NUTRITION: A REVIEW TABLE 1. A summary of food items used for broodstock of various species of crustaceans (alphabetical listing by species). Species Macrobrachium australiense M. rosenbergii M. rosenbergii Panulirus argus Parapenaeopsis hardwickii Penaeus indicus P. indicus P. japonicus ablated > unablated > P. japonicus P. japonicus P. japonicus P. kerathurus P. merguiensis P. monodon P.monodon P.monodon P. orientalis P. plebejus P. semisulcatus P. semisulcatus P. setiferus P. setiferus P. stylirostris & P. vannamei P. yvannamei Maturation Feedstuffs Diets: fish (Priacanthus macracanthus) vs. squid (Sepioteuthis, sp.) fresh mussels (Mytilus edulis) > + frozen shrimps (Crangon crangon) > natural aquatic biota is maintained in holding pools; mixed commercial feeds + vegetable matter + natural food organisms (eg. live snails; mosquito fish, Gambusia) live fiddler crabs (Uca pugilator) + frozen fish no food was provided post ablation fresh frozen prawns @ 3.5% body wt. + formulated pellet @ 2.5% body wt. squid, marine annelids (Perinereis nuntia vallata), & green mussel (Perna viridis) > commercial pellets (45% protein) > short-necked clam (Tapes philippinarium) pelleted diets > + clam & krill meat > pelleted diets + krill flesh pelleted diets mussels fresh mussels (Mytilus edulis) + frozen whole shrimp (Crangon crangon) salted mussels (Modiolus metcalfei) fresh mussels (Mytilus edulis) + frozen whole shrimp (Crangon crangon) fresh frozen prawns @ 3.5% body wt. formulated pellet @ 2.5% body wt. fresh mussel flesh (Mytilus edulis) + frozen shrimp (Crangon crangon) for 10 days pre-ablation: frozen blue mussels (Mytilus edulis planulatus) post-ablation: fresh mussel + occasional fresh or frozen prawn tails frozen adult Artemia salina > frozen fish + frozen shimp — frozen adult Artemia salina > frozen fish + frozen squid > blood worms (Glycera dibranchiata), sand worms (Nereis viridens), squid (Loligo sp.) oysters (Crassostrea sp.) & mussels (Mytilus edulis) oyster (Crassostrea sp.), or prepared dried feed > squid (Loligo sp.) > sandworm (Nereis viridens) > Diets: 1) clams only (Mercenaria mercenaria) 2) shrimp only (P. aztecus) 3) squid only (Loligo & Lolliguncula) 4) blood worms (Glycera dibranchiata) 5) composite pelleted diet, frozen squid, + blood worms Regimen daily 1x weekly varies, depending on whether pools are indoors or outdoors combined, daily 10 days combined, 2 x daily ad libitum: 8-9 AM daily 5—6 PM daily 4% body wt./day 2% body wt./day 1x daily every 2 days @ sunset 2x daily @ 0900h & 1800h 1x daily @ sunset 30% body wt./day 1 x daily, in excess 15% body wt./day combined, 1 x daily, in excess combined, 2 x daily combined, | x daily 10% body wt. day except Saturdays: @ 1 hr before dark morning feeding late afternoon morning feeding late afternoon 3% body wt./day 4x daily @ 0800 (worms), 1200 (squid), 1600 (worms), 2000 (bivalves) 3% body wt./day @ 0800h 1130h 1500h daily, @ 10% of estimated biomass feedings @ 0800, 1000, 1400, 1700h combined, 2 x daily Reference Lee & Fielder (1982) Wickins & Beard (1974) Corbin et al. (1983) Quackenbush & Herrnkind (1981) Kulkarni & Nagabhushanam (1980) Emmerson (1980) Makinouchi & Primavera (1987) Teshima et al. (1988a) Yano (1984) Yano (1985) Yano (1987) Lumare (1979) Beard et al. (1977) Primavera (1978) Beard & Wickens (1980) Emmerson (1983) Arnstein & Beard (1975) Kelemec & Smith (1980) Browdy & Samocha (1985a, b) Browdy et al. (1986) Brown et al. (1979) Lawrence et al. (1980) Chamberlain & Lawrence (198 1a) Yano et al. (1988) 4 HARRISON has implicated ecdysteroids* ir .zg and embryonic devel- opment (Chang 1984, 1989. Manufacture and release of crustacean reproductive hc ones are in response to both exogenous factors (e.g iemperature and light (Aiken 1969a,b); see Introduc .n), and endogenous factors such as developmental sta‘ (Sochasky et al. 1973; Kulkarni and Nagabhushanam 1° J; Quackenbush and Herrnkind 1981), molt stage (Blan .et-Tournier 1982), and energy and nu- trient reserves t support the cost of reproduction (Clarke et al. 1985). In ae wild, molt cycles and maturation cycles are likely ynchronized with seasonal food quality and availabili’, (Clarke 1982). This synchrony has been postu- lated to ve in response to annual light cycles, which would accovu .t for the importance of light on maturation in crusta- cears, and for the presence of “‘eyestalk mediated repro- ductive mechanisms’’ (Wurts and Stickney 1984). Both nutrient and energy status are important considerations for captive reproduction particularly when precocious matura- tion is rapidly induced by eyestalk ablation (within three or four days, AQUACOP 1979) (see also Section 3). A particular consideration in steroidogenesis is metabo- lism of the steroid precursor, cholesterol. Crustaceans are not capable of de novo cholesterol synthesis (Teshima and Kanazawa 1971a), so immediate dietary intake of choles- terol or the mobilization of previously consumed choles- terol reserves is required to support production of steroid hormones (see Section 3.2.3). Evidence demonstrates that several tissues are active in steroid metabolism and pos- sibly steroidogenesis, and should be examined in experi- ments on broodstock cholesterol requirements and metabo- lism: 1) In some species the Y-organ is the most active site of synthesis of the molting hormones, a- and B-ecdysone. However, in other species the situation is not clear, for ex- ample, there is evidence that synthesis or activation of B- ecdysone may also occur in other tissues such as the hepa- topancreas. Their release and proportional titres vary over the molt cycle. Little is known about the complex endo- crine regulatory mechanisms which affect the relationship between molting and reproduction. 2) The androgenic gland, exclusive to males, has a role in the development and maintenance of male sexual characteristics, and in spermatogenic activity (Charniaux-Cotton 1962; Nagabhu- shanam and Kulkarni 1981). Steroidogenesis in the andro- genic gland has been demonstrated in the prawn, Macro- brachium (Veith and Malecha 1983) and in the blue crab, Callinectes (Teshima and Kanazawa 1971la). The andro- genic gland is also implicated in the production of the ter- penoid farnesylacetone (Quackenbush 1986) and polypep- tide or protein endocrines (Fingerman 1987), both pre- sumed to stimulate growth of gonads. 3) Ecdysteroids have been found in gonadal tissue, but their role in reproduction is not yet known (Quackenbush 1986). Steroid metabolism has been demonstrated in the ovaries of the crab Portunus trituberculatus (Teshima and Kanazawa 1971b), and there is speculation that ovarian steroids may be at least partially responsible for the mobilization of nutrient reserves from the hepatopancreas. The most widely used technique to induce maturation in prepubescent crustaceans is eyestalk ablation. The sinus glands at the base of the eyestalks are the source of gonad inhibiting hormone (GIH), molt inhibiting hormone (MIH), and several other neurohemal endocrines. In females, GIH inhibits both secondary vitellogenesis and the synthesis of vitellogenin. Vitellogenin synthesis has been described in the oocytes and hepatopancreas and is implicated in other tissues such as subepidermal tissues in the lateral and dorsal hypoderm of the cephalothorax and the abdomen of fe- males (Vazquez-Boucard et al. 1986), subepidermal con- nective tissue (Paulus and Laufer 1987), or subepidermal adipose tissue (Tom et al. 1987). Whether the hepatopan- creas and other tissues as well as the ovary are the target organs of GIH has been a topic of active research (Char- niaux-Cotton 1985; Quackenbush 1989). Ablation of an eyestalk results in an increase in total ovarian mass due to the acceleration of primary vitellogenesis and the onset of secondary vitellogenesis (Kulkarni and Nagabhushanam 1980; Kanazawa et al. 1988). In males, ablation induces precocious spermatogenesis, enlargement of the vas def- erens and hypersecretion and hypertrophy in the androgenic gland (summarized by Fyffe and O’Connor 1974) and in- creases gonadal development in the lobster, Panulirus argus, (Quackenbush and Herrnkind 1981). Eyestalk abla- tion also removes the source of MIH, HGH (hypoglycemic hormone), and other neurohemal compounds which trigger ecdysis and affect hemolymph glucose regulation and other hormonally-mediated processes. The affects on metabolism and mobilization of nutrients are complex and may be syn- ergistic or antagonistic, depending on the molt stage. Re- production in decapods is characterized by a ‘‘pre- spawning molt’’ but no data are available on the relation- ship between vitellogenesis and the molt hormone (Charniaux-Cotton 1985). The interactions and balance of these factors merit a comprehensive and in-depth review but are outside the framework of this paper. Although the importance of prostaglandins in reproduc- tion has been demonstrated in many species, apparently no studies on the role of prostaglandins during reproduction in crustaceans have been published. The possible dietary re- quirement of omega-6 and/or omega-3 polyunsaturated fatty acids (PUFAs) to provide prostaglandin precursors is discussed in Section 3.2. 2.2 Maturation Sexual maturation entails several processes including gonadogenesis (the development of gonadal tissue), game- togenesis (the production of oocytes or spermatocytes) and in females, primary and seconary vitellogenesis (the endog- enous and exogenous production of egg yolk proteins). Kanazawa et al. (1988) have described three stages of ovarian maturation in penaeid shrimp based on histological CRUSTACEAN BROODSTOCK NUTRITION: A REVIEW 5 examinations of developing ovaries. The juvenile phase corresponds to the onset of ovarian tissue development, prepubescence is characterized by oogenesis and primary vitellogenesis, and puberty is distinguished by secondary vitellogenesis. At sexual maturity, females have fully de- veloped ovaries and oocytes, and are ready for oviposition (ovulation, or egg laying) and fertilization. This is sup- ported by the histological and histochemical descriptions of ovarian maturation in Penaeus monodon by Tan-Fermin and Pudadera (1989). The Gonado-Somatic Index (GSI), also called the gonad index, or the ovarian, ovary or testis index, is a standard measure of degree of gonad maturation in crustaceans (Pillay and Nair 1973; Kulkarni and Nagabhushanam 1979; Lawrence et al. 1979; Jeckel et al. 1989a). GSI = (gonad wet weight/body wet weight) x 100. In female crusta- ceans, there is a significant increase in GSI as maturation approaches and a significant post-spawning drop. This re- lationship has not been demonstrated in naturally maturing male shrimp (Lawrence et al. 1979; Middleditch et al. 1980a; Jeckel et al. 1989b). However, GSI was reported to be an effective measure of testicular development in hor- mone-injected Parapenaeopsis hardwickii, as compared to controls (Nagabhushanam and Kulkarni 1981). The GSI of female Penaeus vannamei was significantly different among treatments of different natural diets (Chamberlain and Lawrence 198 1a). 2.2.1 Gonadogenesis: (development of gonads) Ovogenesis should be technically defined as develop- ment of ovarian (somatic) tissue. In the scientific literature changes in weight and composition of whole ovaries are usually reported and thus the contribution of ovarian tissue distinct from that of intraovarian processes such as 00- genesis and vitellogenesis is not always clear. The termi- nology should also be distinct regarding the developing ‘oocytes’ or ‘ova’ which are contained within the ovaries and which can take up nutrients through absorption and pinocytosis, vs. the spawned ‘eggs’ which are in general impermeable to organic nutrients (but may exchange respi- ratory gases by simple diffusion through the egg mem- brane, Adiyodi and Subramoniam 1983). Gonadogenesis in males refers to development of the testes and vas deferens. In penaeids, tissue development of the vas deferens includes synthesis of a large ampoule for sperm storage, and manufacture of spermatophores (Talbot et al. 1989). The testis is comprised of germinal cells and somatic cells, which are rich in organic reserves (Pochon- Masson 1983). Nagabhushanam and Kulkarni (1981) re- ported an increase in testicular content of glycogen and fat levels and a depletion of hepatopancreatic levels of these energy reserves, concomitant with testicular development and spermatogenesis in hormone-injected penaeid shrimp. Maturation of the testes (and spermatogenesis) in the shrimp Pleoticus muelleri is marked by a greater than 20% increase in nitrogen content, attributed to the biosynthesis of nucleic acids (Jeckel et al. 1989b). 2.2.2 Gametogenesis: (gamete development) Oogenesis is a continuous process from oogonia to the end of primary vitellogenesis. Stages in oogenesis are de- scribed by Meusy and Charniaux-Cotton (1984) (Table 2), Anderson et al. (1984), and Charniaux-Cotton (1985). In some species (e.g. the crabs Cambarus and Libinia), ovaries may simultaneously contain oocytes in several stages of development and vitellogenesis, while in others TABLE 2. Sites of nutrient processing for accumulation by developing oocytes in decapod crustaceans. Stages and features of oocyte development as described by Meusy and Charniaux-Cotton (1984). Stage of Oocyte Development Developmental Features Meiotic prophase (up to Sites of Nutrient Manufacturing & Processing 1) Primary follicular envelope (FE) around the oocytes diakinesis) Pre-Vitellogenesis 2) Visible chromosomes 1) Primary FE around the oocytes 2) Chromosome disappearance 3) Ist enlargement in diameter (slow) 4) Storage of free ribosomes Nutrient storage in the hepatopancreas General protein synthesis & accumulation by the oocytes 5) Development of rough endoplasmic reticulum (RER) & fragmentation (with formation of RER vesicles) Primary Vitellogenesis 1) Primary FE around the oocytes 2) Continued Ist enlargement (slow) Secondary Vitellogenesis microvilli 3) 2nd enlargement (rapid) 4) Color development 1) Secondary follicle cells around the oocytes 2) Formation of vitellin envelope & development of Glycoprotein synthesis by oocyte RER (future cortical rod proteins) Lipovitellin synthesis by the oocytes (endogenous vitellogenesis) Exogenous vitellogenesis (by the hepatopancreas) Pinocytotic uptake of vitellogenin & accumulation of associated carotenoids Continued endogenous vitellogenesis 6 HARRISON (e.g. Uca), oocyte developme. is synchronous (Wolin et al. 1973). Although, in ger al, decapods produce a large number of heavily yolkec ggs, between species, there is an apparent correlation + .ween the amount of yolk and the duration of embryoge’ sis as well as with the level of de- velopment of the hy aed larvae. The numbers of oocytes developed may al be a direct effect of maternal nutrient status. Descriptior of spermatogenesis in penaeid shrimp are reported by Nagabhushanam and Kulkarni (1981) ac- cording te ae criteria of Rao (1978), and in lobsters sper- matoger sis is summarized by Aiken and Waddy (1980). The nv aber of spermatozoids in penaeids varies in quantity and juality during the spermatophore life (Goguenheim et al. 1987, unpublished). In a review by Pochon-Masson (.983) on spermatogenesis in crustaceans, it is reported that decapod spermatozoa contain a “‘remarkable abun- dance of polysaccharides, total proteins and basic pro- teins.”’ 2.2.3. Vitellogenesis: (egg yolk production and accumulation) Vitellogenesis includes production of egg yolk precursor (vitellogenin, VG), the major egg yolk lipoprotein (lipovi- tellin, LV), and accumulation of both organic and inor- ganic constituents of yolk by the oocytes (Adiyodi and Su- bramoniam 1983). Since crustacean VG was first discov- ered in the late 1960’s there has been much controversy regarding the site of biosynthesis of egg yolk proteins; more recently, the origin of the egg lipids is also under investigation. Two stages of vitellogenesis (primary and secondary) corresponding to endogenous (intraooplasmic) and exogenous (extraovarian) sources of yolk protein, re- spectively, have been identified (Dehn et al., 1983; Char- niaux-Cotton 1985 and Quackenbush 1986 for reviews; Vogt et al. 1987). The debate on the site(s) of vitellogenin synthesis continues, fueled partly by the vague use of the term egg yolk “‘protein’’ and partly by the lack of speci- ficity of the assay methods. The recent emergence of radio- immunoassay techniques should obviate the latter short- coming. Using immunoassay techniques, Yano and Chinzei (1987) demonstrated synthesis of vitellogenin in the ovary and lack of synthesis in the hepatopancreas of Penaeus japonicus. However, tissues were examined from only two early vitellogenic females. Using similar proce- dures, Quackenbush (1989) investigated, in Penaeus van- namei, synthesis of proteins which were immunoreactive with anti-yolk antibody (indicating them to be the principal egg yolk protein, vitellin, VP). VP synthesis was measured in the ovary and the hepatopancreas, contrary to the results of Yano and Chinzei (1987). Quackenbush attributed this disparity to differences in the specificity of antibodies used in the techniques, and to possible differences among pen- aeid species. Stage of maturity may also have been a con- tributing factor. In order to clarify the impact of maternal nutrient status on egg yolk accumulation, in particular the contribution of Ovarian vs. extraovarian nutrient manufacture and reserves, the terms describing oocyte and egg composition as used by Adiyodi and Subramoniam (1983) will be adopted for this review. In summary, the oviposited egg has two main com- partments, the ooplasm and the vitellus (yolk), which are enclosed by an egg membrane, the chorion (Fig. 1). The chorion is secreted by the ovary (Neville, 1975) or by the oviduct (Yonge, 1938). The ooplasm contains mitochon- dria, cortical granules (or rods), which are peripherally lo- cated glycoproteins manufactured within the oocyte by golgi and rough endoplasmic reticulum (RER), and lipid globules. These spheres of neutral lipid may be produced by the golgi and smooth endoplasmic reticulum (SER) or may possibly be accumulated from hemolymph lipids car- ried from the hepatopancreas as high density lipoproteins (HDLs). The yolk is surrounded by a vitelline membrane and is comprised of water and water soluble fractions (in- cluding glycoproteins) and both inorganic and organic solids. The organic constituents are primarily proteins and lipids (approximately 24 and 22% of the egg wet weight, respectively), and as lesser fractions, carotenoid pigments and carbohydrates (conjugated to the LV) and free amino acids, free sugars, nucleotides, and nucleic acids. The magnitude of the build-up during vitellogenesis of the major organic components can be seen in Table 3 from An- ilkumar (1980) cited in Adiyodi and Subramoniam (1983). In addition to the principal egg yolk protein, VP, (packaged as the HDL, LV, in discrete membrane-bound yolk platelets), the yolk is also comprised of non-conjugated simple proteins, and glycoprotein vesicles, which are man- ufactured by the golgi and RER, and may be engulfed by yolk platelets. Durliat (1984) investigated the occurrence of hemocyanin and other blood proteins in the ovaries and eggs of the crayfish, Astacus leptodactylus. In nature, primary vitellogenesis may take several months, and as yolk glycoproteins are manufactured within the RER and golgi apparatus of the ooplasm (Lui et al. 1974) (Fig. 1A), there is a slow increase in oocyte size (Table 2). According to Meusy and Charniaux-Cotton (1984) these proteins were incorrectly identified as vitellins in the earlier literature. This misidentification has contrib- uted to the confusion regarding LV accumulation. Secondary vitellogenesis is marked by rapid accumula- tion of LV. Lui and O’Connor (1976, 1977) demonstrated that the oocytes are capable of producing at least 3 of the 5 subunits of the LV molecule, as well as assembling the composite LV molecule (Fig. 1B, C). The lipid fraction includes phosphatidylcholine, phosphatidylethanolamine and may be conjugated with carotenoids and carbohydrate moieties. Also during secondary vitellogenesis, the oocytes de- velop villi and exhibit micropinocytotic (also called endo- cytotic) uptake of large quantities of VG from the hemo- lymph (Zerbib and Mustel 1984; Paulus and Laufer 1987), KEY TO ABBREVIATIONS LIPIDS: Pie 1G. ERO TG - TRIACYLGLYCERIDES Cron - DIACYLGLYCERIDES - MONOACYLGLYCERIDES FREE FATTY ACIDS PHOSPHOGLYCERIDES LYSOPHOSPHOGLYCERIDES PROTEINS: (PROT.) - PROTEINS OF DIETARY ORIGIN - DIETARY AMINO ACIDS - FREE AMINO ACIDS - PROT. FROM HEPATOPANCREAS STORES - VITELLIN PROTEINS - PROT. FROM OVARIAN STORES - LIPOVITELLINS HER: - CAROTENOIDS - CARBOHYDRATES STORAGE RE SOP) FAA G Figure 1. Model of the dynamics of lipid and protein intake, absorption, storage, synthesis, and mobilization during ovarian maturation, and concomitant accumulation of nutrients by the oocytes, in decapod crustaceans. Compartments represented are diet, midgut (lumen and midgut cell), hepatopancreas (lumen and R-cell), hemolymph (showing high density lipoproteins, HDLs), ovary, and oocyte. Pathways of absorption of nutrients into the hemolymph are shown by numbers 1-4. Absorption requires conversion of lipids to PLs, which are conjugated with proteins synthesized by the rough endoplasmic reticulum: 1) absorption of dietary lipids and proteins across the lining of the midgut (1-6 hours post prandial), 2) absorption of dietary lipids and proteins across the hepatopancreatic R-cells (12-24 hours post prandial), 3) mobilization and absorption of nutrients stored in the hepatopancreas, and 4) absorption of vitellogenin across the R-cells which specialize as vitellogenocytes (during vitellogenesis only). Modes of accumulation of oocyte nutrients are shown as A-F. A) glycoprotein synthesis, B) endogenous yolk synthesis, C) endocytosis of HDL, D) endocytosis of vitellogenin, E) reesterification of lipids, and storage of TG, F) de novo TG synthesis. The timing and other details of these events are discussed in section 4.0. EFA and NEFA are essential and non-essential fatty acids. 7 8 HARRISON (Fig. 1C) which results in a - and leads to oviposition. T the oocyte after passing th ugh the intercellular spaces be- tween the ovarian foll’ e cells (Kessel 1968, cited in Wolin et al. 1973). V adarajan and Subramoniam (1980) obtained histochemi . evidence that the follicle cells of the anomuran crab, C /anarius clibanarius have a high AMP content, and sug -st the AMP may serve a similar function as in vertebra’ cells, in facilitating the diffusion of dis- solved mole .ies across cell membranes. They further sug- gest that tt follicle cells may assist in endocytosis of mac- romolec’ .es by the oocytes. The selectivity of the endocy- tosis h s not been determined (Charniaux-Cotton 1985) and is an .mportant consideration when examining nutrient mo- bil’-ation during maturation and possible dietary contribu- tion (particularly of omega-3 and omega-6 PUFAs) to the accumulation of yolk reserves for the embryo. Vitellogenin, a carotenoglycolipoprotein (Zagalsky et al. 1967) is unique to maturing females and is synonymous with the term Female-Specific Protein (FSP) identified in other arthropods and used in the earlier crustacean litera- ture. During ovarian maturation, it is a major circulating HDL in the hemolymph. VG of has been shown to be com- prised of 5 subunits (Lui and O’Connor 1976, 1977; Vaz- quez-Boucard et al. 1986) and to be immunologically iden- tical to ovarian LV (Vazquez-Boucard et al. 1986). It is considered to be an extra-ovarian egg yolk precursor. The protein component of both VG and LV is VP; this and the components of it are discussed in Section 3. Quinitio et al. (1989) identified and characterized VPs in the hemolymph and ovaries of the protandrous hermaphrodite shrimp, Pan- dalus kessleri. In decapod crustaceans, VG is produced in vitellogeno- cytes of the hepatopancreas (Vogt et al. 1985; Paulus and Laufer 1987), (Fig. 1, No. 4). There is, however, circum- stantial evidence that, in addition to the oocytes, the hemo- cytes and subepidermal connective tissue (Aiken and Waddy 1980) or subepidermal adipose tissue (Tom et al. 1987) also produce exogenous VG. Paulus and Laufer (1987) demonstrated in Carcinus that maximal synthesis of VG by the hepatopancreas is concurrent with the period of micropinocytosis by the oocyte, and that the subsequent sharp decline in synthesis is synchronous with reduced “‘ovarian receptivity to VG uptake’’ and reduced rate of pinocytosis as the oocyte matures (Wolin et al. 1973). Based partly on this synchrony, Paulus and Laufer (1987) speculate that the hepatopancreas as well as the oocytes are “targets for hormonal factors regulating reproduction.”’ Circulating levels of VG have also been attributed to ovarian resorption, i.e. “‘leaking,’’ of LV from degener- ating oocytes, possibly en route to the hepatopancreas for reabsorption and elimination (Lui and O’Connor 1976, 1977) or for storage and recycling, ready to be mobilized upon initiation of rematuration. id increase in oocyte size hemolymph proteins reach 2.3. Mating, Ovulation and Fertilization Courtship and mating may occur before or after ovarian maturation, depending on the crustacean species. Courtship activities such as stroking, dancing and behavioral protec- tion of newly molted females by males are believed to be hormonally induced. Mating is by copulation, whereby the male inserts a spermatophore into the thelyca of the fe- male; fertilization occurs externally, upon ovulation and passage of the oocytes through the gonopore. Spermato- phores may remain implanted through several ovarian mat- uration cycles, and fertilize as many as six spawns within one molt cycle (Beard and Wickins 1980). Browdy and Sa- mocha (1985a) reported a single implant fertilizing four spawns ‘‘with no significant decrease in the percent fer- tility.’” Ecdysis, however, usually results in the loss of at- tached spermatophores (Kelemec and Smith 1980; Beard and Wickins 1980; Yano 1985) and therefore new matings are required for fertilization to occur. Artificial insemina- tion has been successful in caridean shrimp (Sandifer and Lynn 1982) and is being attempted with lobsters (Ho- marus) (Talbot 1984). 2.4 Spawning (Egg Laying), Embryogenesis, and Eclosion (Hatching) Once spawned and fertilized, the eggs are either released (e.g. penaeids) or are oviposited on special pleopods and are brooded, without any cytoplasmic connections, beneath the female’s abdomen (e.g. carideans, lobsters, brachy- urans). The reported impermeability of crustacean eggs to organic substances (Claybrook 1983) may be attributed to chitin in the one or more serosal cuticles (Neville 1975) of the external membrane, which are secreted within the chorion by the developing embryo (Goudeau 1976). The components of the developing embryo are contained en- tirely within the egg; development is controlled by the ge- netic information and supported by the available energy and biochemical precursors. Detailed aspects of crustacean embryogenesis are reviewed by Anderson (1982). In penaeids, embryonic development is relatively short and therefore extremely rapid; fertilized eggs hatch into nauplii within approximately 12 to 24 hours, depending on temperature and condition of the embryo (Primavera and Posadas 1981). The newly hatched nauplii do not feed and are nourished by the remaining yolk which must sustain them through several (5 or 6) molts and through metamor- phosis into protozoea larvae, within about 48 hours. The embryonic period is protracted in some species, and thus the nutrient and energy reserves of individual eggs are even more critical. For example, Macrobrachium may brood their eggs (with developing embryos) for 3 weeks, after which the embryos hatch into free-swimming larvae, and lobsters may brood their eggs up to 9 months (Wickins 1982). Biesiot (1982, 1986) found lipid derived from yolk metabolism remaining among the hepatopancreatic tubules (as lipid vacuoles in the resorptive, or R-cells) of late em- CRUSTACEAN BROODSTOCK NUTRITION: A REVIEW 9 bryos and first stage larvae of the lobster, Homarus ameri- canus; it was depleted during early larval development, and lipid storage was not again observed until stage IV. Sasaki (1984) and Sasaki et al. (1986) examined the quantity and quality of nutrient reserves in eggs, embryos, and freshly hatched larvae of the lobster (H. americaus). The research revealed that reserves of the egg yolk, especially EFAs and other nutrients which can not be synthesized de novo, may be depleted during embryogenesis and are thus insufficient to support larval development; they further demonstrated the importance of first feeding. Anger et al. (1985) exam- ined the effects of starvation on the ultrastructure of the hepatopancreas R-cells of stage I lobster larvae (H. ameri- canus), and reviewed the importance of lecithotrophy, as compared to feeding, on early larval development. Nutrient uptake, bioenergetics, and secondary lecithotrophy during larval development are beyond the scope of this review, but research papers by Anger (1986, 1988, 1989), Clarke et al. (1985), Dawirs et al. (1986), Frank et al. (1975) and the citations within, provide a foundation for further inquiry. For example, topographical distribution of remaining vi- tellic reserves in the first free larvae in cells directly linked with the digestive tract should also be studied in detail (pers. comm. Ceccaldi). 3.0 NUTRITION-REPRODUCTION INTERACTIONS: THE ROLES OF SPECIFIC NUTRIENTS The embryo and pre-feeding larvae of crustaceans are lecithotrophic>, as their nutrition is solely supplied by egg yolk reserves. The quality and quantity of nutrients in egg yolk is dependent on maternal body reserves, capacity for biosynthesis, and dietary intake during maturation. The contribution of each of these (and species differences in relative contribution from each source) must be understood in order to formulate effective broodstock diets and feeding regimes. Goguenheim et al. (1987, unpublished) report that quality of spawned eggs is highly correlated with fresh food given during the very rapid secondary ovogenesis period. Maternal nutrition must be augmented to provide sufficient energy and appropriate nutrients to meet the metabolic costs of biosynthesis and of mobilization of nutrients for the manufacture of gonads, oocytes and egg yolk. Dietary intake must also provide and replace all essential® nu- trients lost to the eggs in support of embryogenesis and early larval development. When maturation is induced by eyestalk ablation, ‘The term lecithotrophic (reliance on egg yolk nutrition) is substituted by the term vitellogenic in some citations. However, vitellogenic more accu- rately describes an animal or organ in the state of vitellogenesis (egg yolk production or accumulation). °In nutrition science, an ‘essential’ nutrient is one which cannot be manu- factured by the body (or manufactured fast enough) to satisfy the or- ganisms’s requirement for that nutrient. drastic, accelerated, hormonal and metabolic changes occur, and ovarian development may be stimulated, de- pending partly on the stage of ecdysis and the predisposi- tion or readiness of the female (i.e. in a pubescent, or pre- vitellogenic state). The immediate shift into biosynthetic processes and in the sequestering and bioaccumulation of energy reserves for the oocytes requires rapid mobilization of endogenous nutrient reserves, and replenishment of those which can not be synthesized de novo. Under natural conditions, however, either a female may not reach full maturity without adequate nutrition and sufficient build-up of reserves, or the metabolic shift and rate of maturation may be more gradual. In contrast, an ablated animal shifts (metabolically) into maturation due to induced hormonal changes irrespective of nutritional status. Post-ablation di- etary intake may not fully compensate for inadequate nutri- tional status and therefore a broodstock diet may be re- quired in advance of ablation to build up nutrient reserves. As suggested by Teshima et al. (1988a), the nutritional status of crustaceans before ablation (and reproductive mat- uration) is likely to be of critical importance for successful maturation and spawning. An additional complication of eyestalk ablation is the non-specific affects on the endo- crine system. In a histological comparison of the ovaries and oocytes of ablated and nonablated Penaeus monodon, Vogt et al. (1987) found that while no changes in hepato- pancreatic cells were induced by eyestalk ablation, exami- nation of vitellogenic features in post-spawn ovaries indi- cated a possible negative influence of ablation on the mobi- lization of hepatopancreatic reserves to the ovaries. 3.1 Energy and Bioenergetics Dietary intake of energy substrates in addition to energy reserves of the body, must exceed the maintenance and ac- tivity costs of an animal in order to have sufficient energy to invest in somatic growth, molting or reproduction. Go- nadal maturation and growth (including molting) can occur simultaneously in some crustaceans, such as Penaeus in- dicus, as long as energy and nutrient requirements are met and environmental conditions are appropriate (Emmerson 1983). In other crustaceans, such as the caridean prawn Macrobrachium, growth is sacrificed at the expense of re- production as can be seen by a decline in growth rate of females entering sexual maturity (Wickins and Beard 1974; Emmerson 1983; Ra’Anan 1987). Dietary carbohydrates, lipids and proteins can all be me- tabolized by crustaceans for energy (Claybrook 1983; Akiyama and Dominy 1989; Lim and Persyn 1989), but the efficiency of utilization of each of these substrates (espe- cially carbohydrates), and the optimal dietary balance, seems to vary among crustacean species. During embryonic development of Palaemon serratus, yolk proteins are ap- parently oxidized for energy as well as reincorporated into tissues of the embryo, as indicated by an approximately 10 HARRISON 25% decrease in total protein ntent of the egg (Richard and Ceccaldi 1977, cited in © iybrook 1983). Capuzzo and Lancaster (1979a,b) and C .ford and Brick (1983) review energy partitioning and t’ hierarchy of substrate utilization in larval and post-me .norphic crustaceans, but there is apparently no infor «tion specific to crustacean females during reproductiv maturation. Reproduction .ad gonadal maturation have enormous associated ene .y costs due to the increase in biosynthetic work, such? the manufacture of lipids, proteins, carbohy- drates and .ucleic acids, for the production of genetic ma- terial (D? A), gonads, oocytes, and egg yolk. The increased metabe ic energy demands of maturing females and males have .ot been reported and are an important area for inves- tig.ion. In addition, the most efficient energy sources for promoting vitellogenesis need to be determined and such information initially requires an understanding of energy sources preferred in the non-reproductive state. 3.2 Lipids Lipids play several important roles in the biochemistry, metabolism and reproduction of decapod crustaceans. Neu- tral lipids, particularly triacylglycerides (TG), are a major energy source, and the predominant form of energy storage in the adult, egg, and pre-feeding larva (Ward et al. 1979; Middleditch et al. 1979; Teshima and Kanazawa 1983; Clarke 1982). To provide energy when required, the fatty acids (acyl chain) are cleaved from the glycerol backbone by enzymes (in vertebrates, by hormonally controlled li- pases). The acyl chains are transported to their destination via the blood, by protein carriers (in crustaceans, primarily as high density lipoproteins), activated, and shuttled into the inner mitochondria where they are catabolized to acetyl CoA by B-oxidation and thus provide acetyl CoA to the TCA cycle, for energy production. In decapods, the hepatopancreas is the major lipid storage and processing organ for postembryonic stages (Vogt et al. 1985). During maturation the ovaries become an additional center for lipid metabolism, including lipo- genesis (TG synthesis) by the SER. There is increasing evi- dence, however, that hepatopancreatic lipids are the major source of lipid accumulation in maturing ovaries (Teshima et al. 1988b). Ovarian lipids (and carbohydrates) provide fuel for the biosynthetic processes of oogenesis and vitello- genesis and are apparently taken up and accumulated by the developing oocytes. Neutral lipids are accumulated in globules comprised primarily of non-essential fatty acids, NEFA, (such as 16:0 and omega-9 family fatty acids) (Te- shima et al. 1988b) (Fig. 1E and F). Phospholipids (PLs, also called phosphoglycerides, are polar lipids) and sterols have important functions as cyto- plasm and membrane constituents of cells, affecting struc- tural and physiological properties. The importance of sterols as precursors to steroidal hormones is discussed in Section 2.1. Phospholipids are also the major transport form of lipids in the hemolymph, accounting for as much as 87 to 88% of the lipid composition of circulating HDL, generally in a protein to lipid ratio of 1:1 to 1.2:1 (Teshima and Kanazawa 1978, 1979, 1980a,b; Lee and Puppione 1978, 1988). Dietary TGs and PLs are digested by the ac- tion of lipases secreted by the hepatopancreas. The cleaved free fatty acids (FAA), monoacylglycerides (MG) and pos- sibly lysophosphoglycerides (LP) are absorbed by the epi- thelial cells lining the midgut (Gibson 1982). Although some authors cited in this review report absorption across the hindgut, this is likely due to confusion in defining the gut regions or the distal end of the midgut, which may ex- tend into the abdomen; the hindgut is lined with chitin and is thus impermeable to organic compounds (McLaughlin 1983, and Dall and Moriarty 1983 reviewed internal anatomy and structural and functional aspects of crustacean digestive systems). The absorbed glycerides and fatty acids are converted to PLs, combined with proteins, then ab- sorbed into the hemolymph as HDLs, within | to 3 hours after ingestion; otherwise, they are taken up by the hepato- pancreas. During ovarian maturation, it appears that PLs and TGs mobilized from the hepatopancreas are also trans- ported to the ovaries as PLs in HDLs. Regarding ovarian lipids, answers to two key questions are necessary to achieve an understanding of maturation: (I) What is the contribution of each of the following to the increase in neutral and polar lipids in the developing ovary? 1) mobilized lipids, previously stored in the hepato- pancreas: TGs (including reesterified dietary NEFAs), and EFAs (principally ‘stored’ as omega-3 and omega-6 PLs) 2) de novo synthesis and export of NEFAs by the hepa- topancreas during maturation 3) de novo synthesis of TGs (ie. NEFAs) in the ovary 4) direct uptake and utilization by the oocytes of di- etary fatty acids (ie. without prior deposition in the hepatopancreas and ovaries, and subsequent remo- bilization) desaturation and elongation of precursors of desired EFA products (ie. which dietary PUFAs can satisfy EFA requirements, and where does the processing occur?) (II) What is the effect of rapidly induced maturation on each of the above? The dynamics of lipid mobilization and metabolism are summarized in a schematic model (Fig. 1) and in Section 4.0. The model is based on evidence presented in the re- mainder of this section. 5 Clarke (1982) addressed the first question in his studies of the polar shrimp, Chorismus antarcticus. In vitro lipid syn- thesis in the ovary and hepatopancreas was examined by measuring uptake and incorporation of tritiated saline media, and desaturase activity was detected by measuring in vitro desaturation of [1-!*C]-palmitic acid. Clarke found that there were insufficient lipid reserves in the hepatopan- creas and too low a level of ovarian de novo lipid synthesis to account for the substantial increases in ovarian lipids during vitellogenesis. He concluded that the accumulation of ovocyte lipids depended on maternal food intake during CRUSTACEAN BROODSTOCK NUTRITION: A REVIEW 11 vitellogenesis and that the hepatopancreas acted primarily to modify incoming lipids for export to the ovaries. How- ever, the life cycle and energetic strategy of this organism differ significantly from the penaeid and caridean decapods primarily addressed in this review, and therefore his findings may not be generally applicable. Direct dietary input to oocyte lipid accumulation is sup- ported by Galois (1984) for the shrimp, Penaeus indicus, and Goguenheim et al. (1987, unpublished) who report that lipids ingested during vitellogenesis influence the fatty acid composition of hemolymph and egg lipids. Radiotracer studies have demonstrated that dietary TGs appear to be sequestered to the developing oocytes within 24 hours of ingestion (Teshima et al. 1988b). Jeckel et al. (1989a) ana- lyzed neutral and polar lipid fractions of immature, slightly mature and very mature ovaries of the shrimp Pleoticus muelleri and found that the neutral lipid composition ap- peared to be influenced by seasonal changes in available foods high in omega-3 PUFAs, while the polar lipid com- position remained unchanged. Zerbib and Mustel (1984) examined uptake of radiola- beled VG by the oocytes of the amphipod Orchestia gam- mareullus. The tritiated VG, injected into the hemolymph, was taken up by the oocytes and accumulated as yolk platelets (ie. as lipoproteins). Additionally they were able to distinguish between the labeled lipoproteins and the en- dogenously manufactured glycoproteins. Apparently, label was not evident in the oocyte lipid globules (i.e. as TG). Transfer of hepatopancreas lipid reserves to the ovary during induced maturation of female Penaeus japonicus was further investigated by Teshima et al. (1988b) using a double tracer experiment. A single diet application con- taining labeled palmitic acid, [*H]-16:0, and linolenic acid, [}4C]-18:3w3, was fed to 12 intermolt prawns. Tissue lipids of 3 animals were analyzed 24 hours after feeding to deter- mine immediate fate of dietary fatty acids. Label was pri- marily distributed in the phosphatidylcholines and FFAs of the hepatopancreas, and in the muscle phosphatidlycho- lines. The remaining shrimp were equally divided into a control group and a group that underwent unilateral eye- stalk ablation. The distribution of labeled fatty acids 5 days after induced ovarian maturation was compared with the distribution in non-maturing adults. Concomitantly, there was a reduction in the proportion of label within the hepa- topancreatic lipids and an increase in label in the ovarian fatty acids. Partitioning of the two labels among the ovarian lipid fractions apparently occurred, with labeled omega-3 fatty acids primarily in PLs (primarily phosphatidylcholine) and the non-essential, labeled palmitic acid and metabo- lites, distributed primarily within the TGs. Although the fate of labeled fatty acids during maturation can be seen from this study, the fraction of labeled to non-labeled fatty acids in each tissue compartment is not reported and conse- quently the mobilization of fatty acids relative to tissue stores (e.g. in the hepatopancreas) cannot be determined. Additionally, prawns were maintained on short necked clam (Tapes philippinarium), presumably rich in polyunsa- tured omega-3 fatty acids, thereby preventing the discrimi- nation of the relative contributions of the diet and hepato- pancreas reserves. Also, the level of PUFA in food or- ganisms has been shown to vary seasonally. It would be interesting to repeat this study: 1) eliminating the clam diet, 2) repeating the double tracer diet, and 3) administering a second diet containing a different label (i.e. a “chaser’) to compare the simultaneous relative incorporation of dietary fatty acids with hepatopancreas fatty acids into the neutral and polar lipid fractions of mature ovaries. 3.2.1 Total Lipid and Lipid Classes The increase in total ovarian lipids with maturation in wild-caught crustaceans has been documented by Pillay and Nair (1973), Guary et al. (1974), Gehring (1974), Kulkarni and Nagabhushanam (1979), Read and Caulton (1980) and Galois (1984). Jeckel et al. (1989a) report that ovary weight increased from 0.63 to 3.47 g during maturation in wild-caught Pleoticus muelleri, with a 4-fold increase in the gonadosomatic index over approximately 10 weeks (rate of 0.10 units per day). Total ovarian lipids doubled from 10.2% to 19.3% of dry weight during that time. Teshima et al. (1988a) investigated lipid metabolism in ablated prawn (Penaeus japonicus), specifically relating the induction of maturation to the accumulation of lipid in the ovary. They demonstrated a ten-fold increase in ovarian lipids (mg/prawn) in ablated prawns compared to unablated prawns and a concomitant decrease in hepatopancreatic lipids. Several of the cited authors describe a decrease in total lipid of the hepatopancreas that is accompanied by an in- crease in ovarian lipids during maturation, and attribute this observation to mobilization of lipids. While evidence of mobilization exists, depletion of the hepatopancreatic total lipid could also be explained by: a decrease in dietary lipid or total dietary calories (Armitage et al. 1972; Clarke 1982), a possible increase in energy consumption attributed to increased biosynthetic activity during maturation, or an increase in metabolic activity induced by eyestalk ablation, but independent of maturation. For example, Teshima et al. (1989) report a two-fold increase in diet consumption by ablated prawns (Penaeus japonicus) as compared to unab- lated controls. Galois (1984) investigated changes in the lipid composi- tion of the ovary, hepatopancreas, hemolymph and muscle of the shrimp, Penaeus indicus, during vitellogenesis. He concluded that the hepatopancreatic lipid reserves only par- tially contributed to vitellogenesis and that dietary lipids must be processed rapidly through the hepatopancreas and exported to the ovary. Depletion of hepatopancreas neutral and polar lipid fractions could not fully account for the in- crease in hemolymph PLs and the dramatic rise in ovarian total lipids (Fig. 2), both neutral and polar fractions. 4 Hepato .ncreas A Hemr ,;mph @ Ov y TOTAL LIPID (mg/ g animal) GSI Figure 2. Variations in the total lipid content of the hepatopancreas, hemolymph, and ovary of the shrimp, Penaeus indicus, during ovarian maturation. Degree of maturation is indicated as the gonado- somatic Index, GSI. Data presented are the means and standard de- viations of six samples. [Translated from Galois (1984).] Lipid export from the hepatopancreas and transport into the hemolymph were investigated by Teshima and Kana- zawa (1980a) who found that TG reserves are converted to PLs and transported as HDLs. It would be interesting to investigate the relative fraction of non-VP HDLs which may simply shuttle dietary lipids to the ovary, and the con- tribution of VG, manufactured from hepatopancreas stores. Lee and Puppione (1988) compared the lipid composi- tion of hemolymph HDLs from male, female, and vitello- genic female blue crabs (Callinectes sapidus). Lipid classes were similar among HDL types (each expressed as a per- centage of total lipid), with phosphatidylcholine as the pre- dominant lipid (80—85%), phosphatidylethanolamine (3%), lysophosphatidylcholine (~1%), TGs (S—8%), cho- lesterol (3—4%), and sphingomyelin (2—3%). The weight percentage of omega-3 PUFAs were higher in the PLs than in the TGs (20:5w3, 10-12% vs. ~3%; 22:6w3, 12-15% vs. ~3%), while the TGs were higher in saturated and monounsaturated fatty acids (16:0 and 16:1w7, 27-29% 2 HARRISON and 10-15% respectively, compared to 13-16% and 5—9% in PLs). 3.2.2 Essential Fatty Acids While there has been substantial circumstantial informa- tion since the mid 1970's indicating the importance of lipids, especially omega-3 PUFAs, for maturation in crus- taceans (Middleditch et al. 1979; Lawrence et al. 1979; Brown et al. 1980; see also Section 5.0) there are few data on the metabolism and dietary requirements for fatty acids (including omega-3 and omega-6 PUFAs). For a review of fatty acid metabolism in crustaceans, see Castell (1982). Several studies which document fatty acid composition of the organs of different shrimp species have demonstrated that, throughout ovarian maturation, ovarian lipids con- tained higher proportions of 20:5w3 and 22:6w3 than the hepatopancreas (Teshima and Kanazawa 1983; Jeckel et al. (1989a,b). The gonads and muscles of maturing Pleoticus muelleri contained 20:5w3 and 22:6w3 at 9-14% and 7—12% of total fatty acids, respectively, and 20:4w6 ranged from 4 to 8% in female gonadal tissues, and 7 to 9.5% in the male reproductive organs. From the combined evidence of high proportions of omega-3 and omega-6 PUFAs in gonadal tissues, the es- tablished essentiality of these fatty acids to crustaceans (Zandee 1967; Kanazawa and Teshima 1977), and indica- tions of the direct influence of dietary fatty acids on the fatty acid composition of the gonads and eggs, the impor- tance of a maturation diet high in PUFAs, especially omega-3 and omega-6 fatty acids, cannot be overempha- sized. However, research is needed to determine not only the optimal levels of inclusion, but also the correct propor- tion of dietary fatty acids of the omega-3 and omega-6 fam- ilies. In the rush to acknowledge the importance of omega- 3 fatty acids and to fortify broodstock diets with sources of these, the possible negative consequences of an imbalance in the ratio of dietary omega-3 to omega-6 ratio imbalance seems to have been overlooked (pers. comm. H. W. Cook). Conversion of 18:2w6 to 20 and 22 carbon chain length metabolites requires desaturation and elongation. The required desaturase enzymes act on and have a higher affinity for 18:3w3 and its derivatives than for equivalent chain length omega-6 family fatty acids. Excess 18:3w3 may compete for available enzyme and result in competa- tive inhibition of the conversion of 18:2w6 to omega-6 family fatty acid metabolites. This may particularly reduce the level of arachidonic acid (20:4w6) which is an impor- tant precursor of prostaglandins in insects and vertebrates. It is clear then, that although the level of 18:2w6 in the diet may satisfy a minimum requirement, indiscriminate fortifi- cation of dietary 18:3w3 may result in its ‘relative over- abundance’, which could be deleterious. Prostaglandins, synthesized from specific PUFAs of membrane-bound PLs, are autocoids (locally acting hor- mones) and have many functions in the reproductive pro- ——— CRUSTACEAN BROODSTOCK NUTRITION: A REVIEW 13 cesses of vertebrates and in arthropods other than crusta- ceans. For example, Brenner and Bernasconi (1989) dem- onstrated that prostaglandin PGE, is synthesized from 20:4w6 in testicles and spermatophores of the insect Tria- toma infestans. During copulation, the male insects transfer arachidonic acid ‘‘and the prostaglandin synthe- sizing machinery’’ along with the spermatophores. Brenner and Bernasconi (1989) concur with Stanley-Samuelson and Loher (1983, 1986) that the prostaglandins possibly play a role in egg-laying behavior. It is likely that part of the dietary requirement of omega- 6 and omega-3 essential fatty acids may be to provide pre- cursors of prostaglandins. Middleditch et al. (1979, 1980b) reported that there were no published studies on endoge- nous prostaglandins in crustaceans. The occurrence and functions of prostaglandins in crustacean reproduction is an important area of investigation and will likely lead to a better understanding of the role of omega-6 fatty acids in crustacean diets, as well as support earlier work on the gen- eral benefits of omega-3 fatty acids to reproductive suc- cess. Several issues related to lipid metabolism and dietary fatty acid requirements in crustaceans, in general, and broodstock, in particular, need to be resolved. To establish the required levels of inclusion of specific PUFAs and the appropriate dietary balance of these, research should be di- rected to determine the location and identification of key enzymes involved in lipid metabolism and mobilization, 2) the role of specific fatty acids as precursors to metabolites with known reproductive functions, such as prostaglandins, 3) the capacity of each species to desaturate and elongate 18 carbon omega-3 and omega-6 fatty acids to essential 20 and 22 carbon chain length metabolites, and 4) the compet- itive interaction between dietary omega-3 and omega-6 fatty acids. Knowledge of the hormonal effectors on lipid storage and mobilization, and changes associated with ec- dysis and maturation are also essential to a comprehensive explanation of the requirements of specific dietary fatty acids. 3.2.3. Cholesterol and other sterols Cholesterol is an important component of all cell mem- branes and the precursor of many bioactive molecules such as steroid hormones. Kanazawa and Teshima (1971) dem- onstrated the in vivo conversion of cholesterol to steroid hormones in ovaries of the spiny lobster, Panulirus ja- ponicus. Cholesterol is an essential nutrient for crustaceans because they are incapable of de novo synthesis of the ste- roid ring (Van den Oord 1966; Zandee 1967; Teshima and Kanazawa 1971a), and therefore cholesterol stores within the muscle, hepatopancreas and gonads are derived from the diet (Middleditch et al. 1980b). Kean et al. (1985) re- viewed the function and dietary requirements of cholesterol for lobsters and other crustaceans, and Akiyama and Do- miny (1989) recommend inclusion levels and dietary sources for crustacean feeds. Although several studies have demonstrated the ability of crustaceans to utilize other di- etary sterols, including phytosterols in place of cholesterol (for review of crustacean sterol metabolism, see Teshima 1982), nutrition studies have indicated that they are inferior to cholesterol in promoting growth in larval and postlarval prawns (Teshima et al. 1983; Kanazawa et al. 1971b) or in juvenile lobsters (D’Abramo et al. 1984). Additionally, Middleditch et al. (1980b) report that 24-methylcholesterol and sitosterol are found in the hepatopancreas but not in the gonads or tail muscle and attribute the absence to a lack of processing in those tissues. Nutrition studies which specifi- cally examine the requirement of dietary cholesterol, or the substitution of cholesterol with other dietary sterols, in pro- moting maturation or successful reproduction in crusta- ceans have yet to be conducted. This lack of information warrants caution in the substitution of cholesterol by other sterols in diets for crustacean broodstock. It is possible that while other dietary sterols may partly spare the requirement for cholesterol, the bioconversion may not be sufficiently rapid to ensure sufficient cholesterol delivery to the devel- oping oocytes. Cholesterol serves many functions in crustacean repro- duction. Free sterols are one of the major lipid classes in the ovaries of shrimp (Teshima and Kanazawa 1983). Total ovarian sterols increase with maturation (Gehring 1974; Lautier and Lagarrigue 1988) and may comprise 6.4 to 22% of the total ovarian lipids (Teshima and Kanazawa 1983). A concomitant decrease in hepatopancreatic choles- terol suggests that mobilization of hepatopancreatic choles- terol stores may contribute to the build-up of ovarian cho- lesterol. Lautier and Lagarrigue, (1988) report a similar de- crease in hepatopancreatic cholesterol during vitellogenesis in the crab, Pachygrapsus marmoratus, and cite earlier papers which also report this observed relationship in other crustaceans (Whitney 1969; Adiyodi and Adiyodi 1971; Zandee and Kruitwagen 1975). Experiments by Teshima et al. (1988a), however, indicate that cholesterol is seques- tered to the ovaries from the muscle stores. Using injected radioactive [!4C]-cholesterol, Kanazawa et al. (1988) mea- sured tissue uptake in Penaeus japonicus during ovarian maturation (induced by eyestalk ablation). Surprisingly, there was no difference in tissue-specific incorporation of ['4C]-cholesterol between ablated and nonablated shrimp. Initial incorporation of label was highest in the carcass, and decreased inversely with the ‘‘slow progressive accumula- tion’’ in the muscle. The largest total incorporation of label, 192 hours post injection, was in the muscle, further implicating this tissue as a major storage site of cholesterol or its metabolites. The ovaries exhibited the highest con- centration of label, implicating them as a major site of cho- lesterol metabolism (or storage) during maturation. From this study, Teshima and coworkers conclude that while the hepatopancreas may be a major site of cholesterol metabo- lism, cholesterol that is mobilized and sequestered to the 14 ovaries during maturation o- 51/ 6.1 4.5 Leucine* 9.0 6.2 6.3 6.6 10.9 8.0 7.1 Methionine* 2.6 Del 2.3 2.3 0.0 2.4 3.6 Phenylalanine* 3.9 3.8 3:5 3.2 4.3 4.7 S57 Proline 4.5 3.5 3:3 333 6.3 4.5 N/A Serine 6.2 3.6 3.4 3.0 10.6 8.8 6.8 Threonine* 5.0 3.3 3.3 3.2 6.3 Sy-7/ 5.9 Tryptophan* N/A 0.7 0.9 1.1 N/A N/A N/A Tyrosine 2.9 3.8 3.2 3.2 2) 3.8 2.0 Valine* 7.4 4.3 4.2 4.2 8.0 ell 6.3 Author Richard Penaflorida Lui & O'Connor Fyffe & O’Connor Vazquez-Boucard (1982)*? (1989) (1977) (1974) (1986) ' See also Zagalsky (1972) for amino acid profiles of carotenoproteins of several species of crustaceans. ? Cited in Vazquez-Boucard et al. (1986) 16 HARRISON of dietary carbohydrates appe ceans. Dietary monosacchari s are rapidly absorbed, but are poorly utilized and hig levels (210%) may actually suppress growth (Lim an’ versyn 1989). Carbohydrase ac- tivity (eg. amylase and -llulase) has been reported for the hepatopancreas and greater in adults. Although further research is on crus .cean carbohydrases is required, it is known that an e -ess of carbohydrates in compounded diets inhibits c¢ .bohydrase activities (pers. comm. Cec- caldi). Carbohy’ -ates are stored in the muscle and hepatopan- creas as g’ ycogen, and are mobilized to serve as precursors of meta’ olic intermediates in the production of energy and non-e‘ sential amino acids. Their specific roles in the pro- duct.on of nucleic acids and as a component in ovarian pigments (carotenoglycolipoproteins) may be especially important for maturation processes such as oogenesis, vitellogenesis, and for embryogenesis. Kulkarni and Nag- abhushanam (1979) and Nagabhushanam and Kulkarni (1981) report a significant depletion in glycogen (and fat) in the hepatopancreas of Parapenaeopsis hardwickii during both ovarian maturation and spermatogenesis, and a simul- taneous increase in these in the ovaries and testes, respec- tively. Carbohydrates also have a critical role in the pro- duction of glucosamine, the precursor in the synthesis of chitin, the principal constituent of the exoskeleton. The usual strategy in formulating feeds is to reduce costs by maximizing the inclusion of carbohydrate while simulta- neously sparing expensive protein (because excess protein will otherwise be used for energy). This strategy of diet formulation must be tested for crustacean broodstock diets. The intense biosynthetic activities of maturation may war- rant the provision of a higher requirement for dietary pro- tein and a supplement of essential amino acids. 5 to vary among crusta- 3.5 Vitamins and Minerals The limited data on crustacean vitamin and mineral re- quirements is briefly summarized in several reviews on crustacean nutrition (Conklin 1980, and D’Abramo and Conklin 1985, for Homarus; Sandifer and Smith 1985, for Macrobrachium; and Akiyama and Dominy 1989, for Penaeus). Conklin (1982) reviewed the role of micronu- trients in the biosynthesis of crustacean exoskeleton; the re- view provides a foundation for future investigations of mi- cronutrient metabolism during embryonic and early larval development. Most of the information on functions and re- quirements of vitamins and minerals is not derived from diet trials or nutrient metabolism research on crustaceans, but is adopted from the literature on finfish and other verte- brates. The recommended level of inclusion for most vi- tamins and minerals in premixes for commercial diets is based primarily on practical experience and observation, rather than on test results. The strategy, for ensuring de- livery to crustaceans, is to overfortify commercial diets to compensate for losses due to heat processing and to leaching into the aquatic environment (Akiyama and Do- miny 1989). Although no studies were found which address vitamin and mineral requirements or functions specific to crustacean reproduction or embryogenesis, several of these nutrients have been demonstrated to fulfill specific roles in the reproduction of other species and are noted in this re- view as a suggestion for further investigation. It might be expected that the requirement level of several vitamins and minerals would be elevated during maturation due to both the intense biosynthetic activities and to possible “seques- tering’ of nutrients by the developing oocytes. Several general questions which need to be addressed regarding vitamin and mineral nutrition specific to crusta- cean broodstock include the following: What are the addi- tional requirements, if any, of the female and male during maturation? What vitamin and mineral levels must be present in each egg to fully support embryogenesis and early larval development? What are the delivery routes and mechanisms for packaging vitamins and minerals into the developing oocyte? How do these differ for fat soluble and water soluble vitamins? If the levels of vitamin and mineral fortification in broodstock diets are increased, is there a resultant higher concentration of each in the eggs? Is this favorable? (i.e. are there toxic effects in the embryo due to hypervitaminosis, particularly of fat soluble vitamins?) Are there vitamin-mineral interactions affecting absorption and metabolism of each, and their optimal balance in diets for broodstock, affecting gonadal and embryonic develop- ment? 3.5.1 Fat-soluble Vitamins (Vitamins A, D, E, and K) Among the essential nutrients vitamin E is claimed to be the most important for gonadogenesis and fecundity in fish (Kanazawa 1986). Vitamin E acts primarily as an antioxi- dant, and its prevention of free-radical attack of cell and organelle membranes could be especially critical during embryogenesis. Vitamin E deficiency can slow ovarian de- velopment in carp (Watanabe and Takashima 1977) and can result in decreased hatching rate and survival of hatched ayu larvae (Takeuchi et al. 1981). Vitamin D is important in calcium and phosphorus me- tabolism in crustaceans as well as in vertebrates. Its role in the absorption and deposition of calcium into the exoskel- eton would suggest that it be a vital component of the egg contents, to support embryogenesis, hatching and molting in pre-feeding larvae. In vertebrates, vitamin A functions in several reproduc- tive processes such as spermatogenesis, oogenesis and em- bryonic growth as well as functioning in somatic processes such as growth and differentiation. Fisher and Kon (1958) review vitamin A in the invertebrates and cite evidence that crustaceans build up vitamin A reserves during maturation, which are then transferred to the oocytes. Specific func- tions of vitamin A in crustacean eggs and embryos have not been reported. That the variation in concentration of vi- CRUSTACEAN BROO™STOCK NUTRITION: A REVIEW 17 tamin A in crustaceans appears to reflect the seasonal avail- ability in the diet (Fisher and Kon 1958), may indicate the importance of vitamin A in broodstock diets if it is deter- mined to be important to embryogenesis and overall repro- ductive success. Carotenoid pigments (e.g. B-carotene and astaxanthin) may serve as vitamin A precursors, and thus help to satisfy a possible vitamin A requirement. 3.5.2 Water-soluble Vitamins The following water soluble vitamins are included in vi- tamin premixes for formulated diets: thiamin (B,), panto- thenic acid, riboflavin (B,), inositol, pyridoxine (Bg), cho- line, folic acid, niacin, B, (cyanocobalamine), biotin, and ascorbic acid (vitamin C). Levels of inclusion are listed in several of the above citations and do not necessarily reflect levels required by the body. Little is known of their specific roles and importance in most metabolic processes in crusta- ceans, and accordingly their functions specific to reprodu- tive processes have not been reported. However, their roles and essentiality is indicated from research on other species, particularly fish. The importance of vitamin C in broodstock diets for fish is reviewed by Waagbo et al. (1989). In addition to the general roles of ascorbic acid as an antioxidant and as an enzyme cofactor in the formation of collagen (which may be especially important during embryonic and early larval development), it is also implicated in the regulation or bio- synthesis of reprodutive hormones in ovarian follicle cells (Levine and Morita 1985). Several studies report the effects of ascorbic acid deficiency or supplementation in fish broodstock diets on egg hatchability and survival of fry (Sandnes 1984; Sandnes et al. 1984a,b; Soliman et al. 1986). A new form of vitamin C, MAP (Mg-L-ascorbyl-2- phosphate), which is more stable during processing, storage and in water, has been used effectively in place of L-ascorbic acid in diets for Penaeus japonicus (Shigueno and Itoh 1988). Its efficacy in broodstock diets, including the accumulation of it by the oocytes and its affect on em- bryogenesis and hatchability should be investigated. 3.5.3 Minerals Mineral nutrition in aquatic animals, including crusta- ceans, is complicated by their ability to absorb waterborn minerals (including calcium, phosphorus, magnesium, so- dium potassium, chloride, and trace elements) across the gill epithelial cells and across the intestinal mucosal cells. In addition, there are possible interactions between vitamin levels (particularly vitamin C) and mineral absorption and metabolism, as has been reported in fish (Hilton 1984), and which can impact gonadal development (Sandnes et al. 1984a). Mineral exchange between the egg and the medium may be expected to occur as reported for fish (Zeitoun et al. 1976) and is likely affected by the mineral concentration in the environment. Mineral deficiencies or imbalances, particularly of ma- croelements, could affect crustacean reproduction in two ways. First, the following physiological stresses could trigger oocyte resorption or otherwise reduce reproductive fitness of the broodstock: by causing electrolyte imbalance, inducing osmoregulatory stress and concomitant dehydra- tion, loss of appetite, altered metabolism and impaired re- spiratory and excretory functions. A second, direct effect of mineral malnutrition is altered composition and quality of the eggs, and as a result, can indirectly impact embryo- genesis, egg hatchability, and viability of the larvae. Deficiencies of micronutrients (trace minerals) could in- duce metabolic disorders due especially to their roles as enzyme cofactors. Rapid proliferation of cellular material during gonadogenesis, gametogenesis, and embryogenesis could be particularly effected. Lall and Hines (1985) re- ported high embryo mortalities and low hatching rate in brook trout (Salvelinus fontinalis) fed diets deficient in manganese. Although Read and Caulton (1980) reported no change in total inorganic body constituents with maturation in Penaeus indicus, it is possible that certain minerals have a higher level of requirement during maturation to support increased metabolic activity and to be sequestered to the developing oocytes for storage and subsequent use by the developing embryos and pre-feeding larvae. Although minerals can be absorbed by crustaceans from the aquatic environment, and fish and shrimp meals in practical diets are believed to contain sufficient trace min- erals, the following minerals are included in mineral pre- mixes for crustacean diets: macrominerals (calcium, Ca; phosphorus, P; magnesium, Mg; sodium, Na; potassium, K; chloride, Cl); trace elements (iron, Fe; copper, Cu; zinc, Zn; manganese, Mn; selenium, Se; cobalt, Co). The trace element lithium is not normally included in mineral premixes for formulated diets, however, Spaar- garen (1988) from a study on brown shrimp (Crangon crangon), has speculated a role of lithium in regulating en- zyme activity in reproduction. The findings showed that body levels are regulated with temperature and salinity, in- dicating a role in metabolic processes. Furthermore, it was concluded that higher levels in females than in males sug- gested a possible link to female-specific reproductive me- tabolism. This, however, remains circumstantial evidence. Total mineral analysis of broodstock diets and analysis of waterborn minerals should be reported (in addition to mineral mix composition) in order to evaluate the impact of total mineral availability in feeding trials with crustacean broodstock, and to determine the proper levels of supple- mentation of specific minerals. Lall (1990, pers. comm.) noted that the reported ash levels in commercial broodstock diets (converted to dry weight basis, Table 4) seemed high considering the possible competitive interactions and resul- tant negative consequences on the bioavailability and me- tabolism of the minerals. These reported ash (total mineral) levels may be attributed to the inclusion of fish, shrimp and 18 HARRISON crab meals in the diets. The m eral content of these meals can vary widely and shor . be monitored. Dean and Akiyama (1989) reported .at calcium and phosphorus in diets for penaeid shrir » should not exceed 2.8% and ~1.8% of feed, respe ively and noted the importance of maintaining a calciu’ .phosphorus ratio of 1:1 to 1.5:1. El- emental analyses * the gonads, eggs, and newly hatched larvae would pre ide valuable information in assessing the impact of high ash diets on reproductive performance and determining .ppropriate levels of inclusion in crustacean broodstock diets. 3.6 Cc otenoids Carotenoid pigments are accumulated in crustaceans in several forms (as free pigments, esterified to fatty acids, covalently bound to macromolecules (e.g. chitin), non-co- valently associated with proteins and carbohydrates in complexes such as carotenoglycolipoproteins, or simply linked with proteins as in crustacyanins). Crustaceans, like all animals so far investigated, are not capable of de novo carotenoid synthesis. Their carotenoids, which are ab- sorbed from dietary sources and directly deposited or first metabolized to other forms, originate from plant materials. The nutritional value of carotenoids to crustaceans and di- etary requirements have yet to be determined, but their con- spicuous accumulation in the ovaries during sexual matura- tion, and subsequent pigmentation of the eggs, has lead many researchers to speculate on their significance and their role in reproduction, in the eggs, and during embry- onic and larval development (Cheesman et al. 1967; Her- ring 1968; Gilchrist and Lee 1972; Gilchrist and Zagalsky 1983; Vincent et al. 1988). Nelis et al. (1989) prepared a comprehensive review of carotenoids in relation to repro- duction and development in Artemia (and other crusta- ceans); they cite many previous reviews and papers on the distribution, physicochemical properties and biochemistry of carotenoids. Information regarding carotenoid nutrition for crustaceans is limited. During early maturation free and esterified carotenoids may accumulate in the hepatopancreas. The quantity and types of carotenoids which accumulate depend on several factors, including seasonal changes in the composition and availability of carotenoids in the diet (Jeckel et al. 1989a), and temperature (Vincent 1989). Selectivity of intestinal absorption and the ability to metabolize different dietary carotenoids may vary among species. Otazu-Abrill and Ceccaldi (1984) examined the influence of Otazu-Abrill and Ceccaldi (1984) examined the influence of cantha- xanthin and Carophyll red® (10% canthaxanthin) added to compounded diets on pigmentation in Penaeus japonicus. Eyestalk ablation has been shown to affect the metabolism and deposition of carotenoids in the prawn Macrobrachium rosenbergii (Maugle et al. 1980). They demonstrated that, in eyestalk ablated prawns, B-carotene is more effectively converted to astaxanthin and deposited than is cantha- xanthin. During secondary vitellogenesis, it appears that carote- noids are mobilized from the hepatopancreas to the ovaries (Gilchrist and Lee 1972; Anikulmar 1980, Table 3; Vincent et al. 1988, Table 6). They are transported in the hemo- lymph, complexed to HDLs (Fig. 1, Hepatopancreas R- cell, No. 4). The carotenoglycolipoproteins appear to be taken up directly by the oocytes, by pinocytosis (Fig. 1D, Ovary). Color development of the ovaries becomes visible and is used as an indicator of the degree of maturation. The range of colors varies with which carotenoids are accumu- lated and with the type of glycoprotein or lipoprotein com- plexes which are formed (Vincent et al. 1988). Astaxanthin is the predominant pigment in the eggs and ovaries of decapod crustaceans; other pigments include B-carotene, lutein, and in homarids and the prawn, Penaeus monodon, ovoverdin imparts a green color. Vincent and Ceccaldi (1988) investigated the relationship between carotenoids and their associated fatty acids in the copepod crustacean, Calanipeda aquae-dulcis and reported that the linkage of fatty acids classes (eg. PUFAs vs. monoenes) depends on the degree of oxygenation of the pigments. Several roles and functions of carotenoids and caroteno- protein complexes, with regards to accumulation in the ovaries and eggs, have been postulated. A brief summary follows, since these are comprehensively reviewed by Za- galsky et al. (1967) and by Nelis et al. (1989). Carotenoids may function during hemolymph transport, and in the eggs, as antioxidants or light shields, protecting nutrient reserves and embryonic tissues from oxidative damage or solar radi- ation. They may also confer structural stability to lipopro- tein reserves by forming nonstoichiometric complexes, and thus protecting the nutrients until required by the embryos or larvae. The carotenoids of eggs may also serve as pig- ment reserves, used by the embryos and larvae in the for- mation of chromatophores and eyespots, or as precursors to vitamin A. Further research is required to more clearly determine the functions, specificity, and quantities of carotenoids used during development of oocytes and embryos of dif- TABLE 6. Variation of total carotenoid pigment in the hepatopancreas and ovaries of the shrimp, Penaeus schmitti during sexual maturation. Adapted from data of Vincent et al. (1988). Values reported on wet weight basis. Earl iddl t Ovarian Development Retin Re Mice ieee Gonadosomatic Index 0.5 0.8 2.2 7.5 7.8 Organ Carotenoids (g/g organ) Hepatopancreas 237 — 588 _— 391 Ovaries - 39 — 463*! _ ' Dry weight value approximately 1630 jg/g organ. CRUSTACEAN BROODSTOCK NUTRITION: A REVIEW 19 ferent crustacean species. Vincent et al. (1988) emphasized that since the diet is the only source of carotenoid pigments for crustaceans, carotenoid enrichment of broodstock diets may influence egg quality, and the viability and vitality of the larvae produced. The effectiveness of different dietary carotenoid sources will also need to be determined. 4.0 INTAKE, MOBILIZATION AND SYNTHESIS OF LIPIDS AND PROTEINS DURING MATURATION, AND THE ORIGIN OF EGG YOLK PROTEINS: A SUMMARY MODEL A schematic diagram of the dynamics of lipid and pro- tein metabolism during late maturation is presented in Figure 1. It summarizes current information regarding the origin of proteins and lipids in crustacean eggs. The contri- bution of each of the following to the increase in TGs and PLs in the developing oocyte is depicted: a) lipids and pro- teins mobilized from hepatopancreatic reserves, including i) NEFAs (from TG stores and reesterified dietary fatty acids and from de novo synthesized fatty acids from non- lipid precursors) and ii) omega-3 and omega-6 PUFAs (es- terified primarily with PLs), b) de novo synthesis of NEFAs in the hepatopancreas concomitant with matura- tion, and exported to the ovaries/oocytes, c) de novo syn- thesis of TGs in the ovaries and oocytes, and d) ingestion and immediate transport of dietary lipids (and carotenoids, etc.) to the ovaries. There is yet no available information on the capacity and location(s), during maturation, for de- saturation and elongation of fatty acid precursors of desired EFA products. Dietary TGs and PLs (each may be comprised of EFAs and/or NEFAs) are hydrolyzed in the foregut by lipases, assisted by non-cholesterol emulsifiers, both secreted by the hepatopancreas. Dietary proteins are broken down into smaller peptides and free amino acids (DPs and DAAs, re- spectively, denoting dietary origin) by proteases, also manufactured and secreted by the hepatopancreas. The products of hydrolysis (mono- and diacylglycerides, lyso- phosphoglycerides (LPs), and FFAs, as well as DPs and DAAs) may be immediately absorbed by cells lining the midgut. The absorbed FFAs and glycerides are converted to PLs for export into the hemolymph as components of HDLs (which have a protein to lipid ratio of approximately 1.2 to 1.0). The peripherally located proteins and phos- phate moieties facilitate solubilization of the lipids in the aqueous hemolymph. The protein fraction of these HDLs is likely resynthesized from DPs and DAAs to form new DPs. There may be some reesterification of dietary glycerides to form TGs, which are also exported as HDLs. Absorption of these HDLs may occur within the first one to six hours after ingestion (Fig. 1, M-Cell, No. 1, HDLs: DP/PL and DP/TG). Most of the MGs, LPs, FFAs, carotenoid pigments (C), carbohydrates (G), etc. are passed with the fluid digesta into the lumen of the hepatopancreas, and taken up by R- cells, which serve to resorb (or absorb) and store lipids and carbohydrates. Some DPs and DAAs may also be taken up by R-cells, although the majority may be absorbed by the F-cells which are specialized for protein synthesis (pri- marily of digestive enzymes). Export of fatty acids from the hepatopancreas to the hemolymph also occurs in the form of PLs as the major lipid component of HDLs. Typi- cally, there are two compartmental sources of fatty acids and proteins which are exported (Fig. 1, R-cell, No. 2 and 3). An additional compartment appears to operate during secondary vitellogenesis (Fig. 1, R-Cell, No. 4). Normal, post prandial absorption from the hepatopan- creas of dietary fatty acids and amino acids occurs within 12 to 24 hours (Fig. 1, No. 2, HDL: DP/PL). Mobilization of dietary fatty acids which had been reesterified and con- verted to TGs for storage in the hepatopancreas, or from TG stores accumulated by de novo synthesis from non-lipid precursors is depicted in Figure 1, No. 3 (HDL: HP/PL). Absorption into the hemolymph requires breakdown of stored TGs to DGs + FFAs, subsequent conversion to PLs, synthesis of carrier proteins (possibly resynthesized from hepatopancreatic protein and amino acid stores, HP and FAA), and export as HDLs. During maturation, certain R-cells of the hepatopancreas specialize as vitellogenocytes and synthesize the principal egg yolk protein, vitellin (Fig. 1, VP). The VP is appar- ently combined with a polyunsaturated PL, which may originate from the hepatopancreatic lipid reserves, or may be directly available from dietary fatty acids. The com- bined VP and PL is the HDL, vitellogenin, which may be complexed with glycogen and/or dietary carotenoid pig- ments and exported to the hemolymph (Fig. 1, No. 4, HDL: VP/PL + C + G). This HDL appears to have an important role in carrying specific lipids, proteins (and pos- sibly other nutrients) to the oocytes. The net result of lipid mobilization and export as HDLs from the hepatopancreas is a decrease in the total lipid, protein, and glycogen in the hepatopancreas, and an in- crease in hemolymph PLs. The magnitude of increase of hemolymph PL levels varies among species, but in general is not large, due to the assumed rapid uptake of the PLs and proteins by the ovaries. There are several mechanisms that lead to oocyte accu- mulation of egg yolk and lipid energy reserves (Fig. 1, Ovary A, B, C, D, E and F). Glycoprotein synthesis by the golgi apparatus and RER (Fig. 1A) begins during primary vitellogenesis and requires amino acids from ovarian stores (OP), presumably resynthesized from circulating HDLs (DP and HP). As a result, glycoprotein vesicles accumulate in the yolk and, by late vitellogenesis, appear as cortical bodies around the periphery of the oocyte. Debate con- cerning the contribution of these glycoproteins to the yolk mass continues. Although they contribute to the total yolk protein, they are not the principal yolk proteins, vitellins, which, combined with PLs, form LVs. Intraoocytic VP synthesis occurs during secondary vitel- 20 HARRISON logenesis (Fig. 1B). As with the glycoproteins, this LV fraction may be synthesized by the RER from amino acids and fatty acids accumulated in the ovaries during primary vitellogenesis, or may be immediately resynthesized from HPs and DPs taken up by the ovarian tissues (including follicle cells) during secondary vitellogenesis. Concomitant with the onset of secondary vitellogenesis, the oocytes develop microvilli, and micropinocytotic up- take of hemolymph “‘particles’’ (including HDLs) is initi- ated (Fig. 1C). The selectivity of uptake is unknown. The particles pass from the hemolymph to the oocytes through ‘“‘channels’’ between the follicle cells. This facilitates di- rect uptake of nutrients by the oocytes, bypassing absorp- tion into ovarian tissues. The HPs, DPs, PLs, etc. accumu- lated in this way may contribute to the general nutrient re- serves in the yolk or may be resynthesized within the oocyte to form LVs. During secondary vitellogenesis extraovarian VPs (eg. produced by the vitellogenocytes of the hepatopancreas, Fig. 1, No. 4), are also taken up by the oocytes, by selec- tive micropinocytosis (Fig. 1D). These vitellogenins not only contribute to the total LV accumulation, but also to the build-up of carotenoid pigments. Lipid globules are also formed in the oocytes during vi- tellogenesis. The results of radioisotope experiments sug- gest that these are TGs and are comprised of NEFAs. They originate from de novo synthesis by the SER (Fig. 1F), and possibly from reesterification of ovarian lipids derived from circulating NEFAs (Fig. 1B). Due to these processes, the ovaries undergo a 4-6 fold increase in size, and increase in total lipid and lipid con- centration (both neutral and polar lipids), and in total pro- tein, carbohydrate, and carotenoid content. Uptake of vi- tamins, minerals, and other specific compounds has not been addressed in the literature. Of fundamental importance to understanding the role of dietary manipulation on oocyte composition is determining the temporal accumulation of oocyte organic (and inor- ganic) constituents, from each of the contributing sources, and the compartmentalization or differentiation of specific components of each fraction. For example, the radioisotope experiments of Teshima and coworkers (1988b) suggests that dietary essential PUFAs are selectively sequestered within the oocyte as a component of yolk PLs (and are thereby available to satisfy EFA requirements during 00- genesis and embryogenesis), while NEFAs become a com- ponent of the oocyte TG stores (for subsequent use as fuel by the embryos and pre-feeding larvae). Lee and Puppione (1988) isolated and characterized two different HDLs in the hemolymph of female blue crabs (Callinectes sapidus) during vitellogenesis, and suggested that each plays a sepa- rate role in transporting specific nutrients to specific tissues, however, no major difference was found between the fatty acid compositions of the two HDLs. An understanding of the timing and degree of dietary influence requires further research. The importance of identifying the HDL components which are formed in the R-cells of the hepatopancreas (ie. the fate of dietary amino acids and fatty acids) can be seen from the HDL exporting compartments depicted in the model. The relative contribu- tions of each of the compartments during rapid oocyte pro- duction relative to eyestalk ablation may be determined using radiotracers in pulse and pulse-chase experiments, along with isolation and possibly specific antibodies to pro- teins of the different classes of lipoproteins (pers. comm. H. W. Cook). In addition, detailed analyses of fatty acids, amino acids, and other organic and inorganic constituents of healthy eggs, embryos, and larvae are needed to formu- late maternal diets so that the appropriate balance of re- quired nutrients is supplied to the developing oocytes. 5.0 THE HISTORY OF CRUSTACEAN BROODSTOCK DIETS Techniques for reproduction, in captivity, of penaeid shrimp were developed by Japanese researchers in the early 1960’s (Laubier-Bonichon and Laubier 1976). Elsewhere however, prior to the mid-1970’s, hatchery production of postlarvae was based on capture (by trawler fishing) of ‘“‘spawners.’’ Females fertilized in the wild and in ad- vanced stages of ovarian maturation were captured. Spawning usually occurred within a day of capture. As a result, diet history and nutritional status were undefined and their role in maturation, spawning and reproductive success could not be determined. In the mid-1970’s, several techniques were developed to induce maturation in wild-caught and pond-reared females. Sandifer (1986) reviewed developments in crustacean re- production and methods used for successfully induced mat- uration, in captivity, of at least 16 penaeid species. Induced maturation naturally demanded care and maintenance of broodstock shrimp, and consequently natural feeds which supported survival and reproduction were identified. The natural feeds included mussels, squid, bloodworms, and shrimp (Laubier-Bonichon and Laubier 1976; Beard et al. 1977; Lumare 1979) (Table 1). Marine organisms with ma- ture gonads were also recommended for broodstock diets as they may contain factors or nutrients essential to gonad maturation in shrimp (AQUACOP 1977). Two important studies conducted by Middleditch and co-workers (1979, 1980b) investigated the contribution of dietary lipids to successful maturation of penaeid shrimp in captivity. In the 1979 study, natural diets were evaluated as a variable affecting ovarian maturation and spawning with animals held in captivity, and component lipids in those diets were analyzed. In the sequel, lipid profiles of gonads, hepatopancreas and muscle of male and female shrimp cap- tured from the wild during the natural spawning season were analyzed. Profiles of females in four stages of matura- tion, from immature to ripe, were reported. Although they did not quantitate the levels of sterols and fatty acids in the tissues, they assessed the relative concentrations of sterol CRUSTACEAN BROODSTOCK NUTRITION: A REVIEW 21 and fatty acid fractions. Cholesterol was the predominant sterol in all tissues and PUFAs comprised a significant por- tion of total fatty acids (see Section 3.3 for details). The effects of manipulation of lipid profiles of test diets through the inclusion of appropriate natural organisms were mea- sured by the reproductive performance of the diet. They attributed the high performance associated with dietary squid to the high percentage of cholesterol in squid sterol (98%) and the success of dietary polychaete worms to the high proportion of long-chain fatty acids. The importance of these lipids for vitellogenesis was also suggested. Other studies devoted to testing the performance of various natural foods on improving broodstock survival and accelerating maturation and enhancing reproductive perfor- mance followed (Chamberlain and Lawrence 1981a). There are, however, many problems concomitant with natural foods. Availability and nutritional quality of diets vary with source and season (Middleditch et al. 1980a) and quality can deteriorate rapidly with storage and handling. Thus, control of the levels of essential nutrients delivered to the crustaceans is difficult and may be especially critical in broodstock nutrition. Additional drawbacks include the greater probability of water fouling and high cost. By the late-1970’s, research designed to develop formu- lated feeds which could partially replace natural foods in broodstock diets was initiated. Advantages of formulated feeds include their reproducible quality, ease of applica- tion, reduction of water fouling, and stability under storage. Although designated as broodstock diets, these were formulated to provide a healthy base diet; the uniden- tified ‘‘essential’’ ingredients for reproduction were still supplied by fresh and frozen natural foods. Commercial crustacean feeds, formulated and marketed for broodstock nutrition, became available in the mid- 1980’s (several manufacturers are listed in Table 4). The formulations only approximate broodstock nutrient require- ments, because to date there has been limited research or published findings. Formulations are based on general pro- duction diets and fortified with several nutrients presumed important for maturation. Unfortunately, results from field trials of commercial maturation diets remain proprietary. Personal communication with repesentatives from several commercial feed manufacturers (1989) revealed that none of these companies advocate replacing more than 50% of live or fresh-frozen feed with their maturation diet. As an example of the primitive state of maturation diets, one commercial hatchery in Ecuador with a production of 36 million postlarvae per month in November and De- cember 1986, relied on a local supply of squid, oysters and clams for their maturation “‘diet.’’ Blood worms were im- ported from Panama, Central America and Maine, USA, because marine polychaetes were not available locally. A reliance on fresh ingredients was common for all hatch- eries. Marine worms have been hailed for several years as an essential component of maturation diets. The high concentration of PUFAs, especially omega-3 fatty acids, are believed to contribute to their efficacy in promoting maturation. For this they are marketed as “‘omega worms.” As of January 1990, commercial maturation diets still required supplementation with natural foods because essen- tial nutrients and optimal balance of nutrients (e.g. protein to calorie ratio, amino acid profile) required for successful reproduction are still undetermined. Research is urgently needed to identify essential nutrients for broodstock, to quantify minimum requirements, and to test their bioavail- ability from different feed ingredient sources. In one of the few published diet studies on crustacean broodstock, Galgani et al. (1989a) report success in partly replacing natural feedstuffs with formulated pelleted feeds in maturation diets for Penaeus vannamei and Penaeus sty- lirostris. They included an ‘‘arbitrary’? minimum of 12% (dry matter basis) of natural food (fresh mussels, two species of fresh fish, and frozen squid; Table 1) in all test diets to ensure induction of maturation. Performance on the artificial diet alone was unfortunately not tested for com- parison and the three formulated diets varied in 23 ingre- dients and in the levels of protein, lipid, and energy. As in previous studies, this approach precludes any conclusions regarding the benefit or requirement of particular ingre- dients or even the optimal proximate composition. How- ever, practical diet formulations are identified which sup- port penaeid broodstock. A similar approach was used in a second study by Galgani and coworkers (1989b) in which they tested the effects of 10 practical diets on reproduction of Penaeus indicus. 6.0 OBJECTIVES AND PERFORMANCE CRITERIA OF BROODSTOCK DIETS Crustacean broodstock diets have been targets to meet maternal nutritional requirements. In addition, the primary objectives of a commercial crustacean broodstock diet are: 1) to promote (or induce) maturation, 2) to enhance fertility and promote mating, and 3) to increase fecundity by im- proving egg quality, egg quantity, and viability of off- spring. The criteria in Table 7, summarized from published lit- erature, have been used for evaluating reproductive perfor- mance in crustacea in many types of studies. These criteria can be applied in assessing the efficacy of crustacean broodstock diets. Since total control of reproduction includes selection of healthy fertile males or the maturation of males in cap- tivity, the nutrition-reproduction interactions in male crus- taceans should also be investigated. Apparently no studies which investigate the impact of nutrition on sperm quality, quantity, or reproductive success of male crustaceans have been published, but poor nutrition is one of several factors which may contribute to degeneration of the male repro- ductive tract and infertility in captive male shrimp (Talbot 22 HARRISON TABLE 7. Criteria for evaluating reproductive performance in crustaceans. A. Criteria for assessing egg quality B. Criteria for assessing sperm quality*! C. Criteria for assessing D. Overall reproductive 1. Eggs per spawn; fertile eggs 1. Spermatophore weight per spawn 2. External condition of the . Egg size or weight spermatophore . Embryo development rate 3. Sperm count . Time to eclosion . Percent hatch rate (hatchability) 4. Percent live sperm ne WP ' From Leung-Trujillo and Lawrence (1987) offspring performance 1. Nauplii (or larvae) per spawn 1. Sexual precocity (age to first 2. Number of potential nauplii maturation) (or larvae) per female per 2. Successful mating month*? (fertilization rate) 3. Healthy nauplii 3. Spawning index (number of a) growth or developmental spawns per female per month) rate 4. Fertile spawns b) molt increment 5. Rematuration and repeat c) molt interval 4. Naupliar development rate 5. Protozoea | length 6. Percent metamorphosis to zoea | performance ? Galgani et al. (1989)—this corresponds in theory to the number of fertile and viable eggs (Primavera and Posadas, 1981) likely to hatch into nauplii “‘in practice, the actual number of nauplii is less than the theoretical’. et al. 1989). Criteria used by Leung-Trujillo and Lawrence (1987) in evaluating sperm quality in the shrimp, Penaeus setiferus, could be used to evaluate maturation diets for male shrimp (Table 7B). Sperm motility is not a valid crite- rion because penaeid sperm are non-motile. 7.0 SUMMARY AND CONCLUSIONS Despite the importance of proper nutrition to maturation and reproductive success of crustaceans, the formulations of broodstock diets are still at the trial and error stage, and fresh (or frozen) natural ingredients are required to supple- ment even the best available diets. Research is needed on several fronts to combat the inadequacies and inefficiencies of current broodstock feeding practices. To improve the overall quality of practical diet formula- tions, a systematic approach and controlled experiments are necessary. The environmental conditions during experi- ments must be reported, as they can influence reproductive maturation and mask the differences attributed solely to diet composition. It is equally important that researchers use standardized terms, clarify the stage of maturation or development, and specify tissues or components under in- vestigation. To test the effects of practical diet ingredients on reproductive performance, the number of variables in each test diet should be limited and control diets must be included to compare each variable tested. Feeding studies using defined, semi-purified diets are necessary to identify specific nutrient requirements for broodstock of each species, and the biochemical composition of the gonads, oocytes, eggs, sperm, and newly hatched larvae should be examined to identify how they differ from the composition of adult muscle tissue, on which many diet formulations are based. In addition, reporting of the compositional changes during maturation of the hepatopancreas and the hemo- lymph will help to determine the importance of nutrient re- serves prior to maturation and of nutrient mobilization during maturation. Research on the energy partitioning and energy require- ments of broodstock females, in particular, on the effect of induced maturation on metabolism should provide a basis for determining caloric density and optimal energy sub- strates in broodstock diets. Formulation of a high energy, high nutrient “‘booster’’ diet may be indicated to compen- sate for the immediate high demands imposed by induced maturation through eyestalk ablation. Radiotracer and radioimmunoassay studies will further our understanding of several key issues: nutrient storage and mobilization, the ‘‘sequestering’’ or valorization of nu- trients and energy to the gonads during vitellogenesis, the source of each constituent of the oocytes, and in particular, the extent and timing of dietary contributions to oocyte composition. Enzyme research will advance our knowledge of the shifts in intermediary metabolism that occur with primary and secondary vitellogenesis. The 1990’s offer exciting challenges in several research disciplines which investigate crustacean reproduction and nutrition. Coordinating these diverse research efforts and integrating the findings will enable us to clearly define broodstock nutrient requirements, formulate practical diets, and achieve highly controlled, efficient, and cost effective reproduction of crustaceans in captivity. In addition, it will further our understanding of the environmental, physiolog- ical and biochemical factors which influence reproductive success of crustaceans in the wild. CRUSTACEAN BROODSTOCK NUTRITION: A REVIEW 23 ACKNOWLEDGEMENTS This review is dedicated to the late Dr. Shao-wen Ling, for his lifetime contributions to aquaculture and especially for his famous ‘‘soy sauce solution’ and pioneering work in closing the life cycle of the freshwater prawn, Macro- brachium rosenbergii. He passed away in July, 1990. The author wishes to thank Mr. R. P. McIntosh and Mr. J. Ogle for discussions on maturation diets, Ms. Joan Kean-Howie, for reviewing early drafts of the manuscript, and members of my examining committee, Dr. R. G. Ackman, Dr. J. D. Castell, and Dr. G. F. Newkirk. The detailed critical reviews and recommendations on an earlier draft by Dr. Harold W. Cook, Dr. Santosh P. Lall, Dr. Louis D’Abramo and Dr. Judith McD.-Capuzzo, were greatly appreciated. In addition, the editorial revisions and rearrangement of sections suggested by Dr. D’Abramo im- proved the effectiveness of the presentation. The author is thankful to Dr. H. J. Ceccaldi whose comments assisted in clarifying key issues in this manuscript. Thanks also to Mr. Terry Collins for preparing the final draft of the nutrient mobilization model. I am grateful to the editor, Dr. Sandra Shumway, for accepting a review of this scope, for rapid publication, for financial support, and for encouragement throughout this endeavor. Support for reprint costs, pro- vided by Mr. Monty Little, president, Syndel Laboratories Ltd., is gratefully acknowledged. A synopsis of this review was presented by invitation in a special session on Feeds and Feeding Strategies at the Annual Conference of the American Fisheries Society, joint with the World Aquaculture Society, LA, Calif., Feb- ruary 1989. The author is grateful to Dr. Tom Morse for support in attending the conference. Travel funds were pro- vided by Clearwater Fine Foods, Ltd. This paper was sub- mitted in partial fulfillment of the requirements for the Ph.D., Department of Biology, Dalhousie University, Feb- ruary 1990. REFERENCES CITED* Adiyodi, R. G. & K. G. Adiyodi. 1971. Lipid metabolism in relation to reproduction and molting in the crab, Paratelphusa hydrodromous (Herbst): cholesterol and unsaturated fatty acids. Indian J. Exp. Biol. 9:514—S15. & T. Subramoniam. 1983. Arthropoda-Crustacea. Adiyodi, K. G. and R. G. Adiyodi, eds. Reproductive Biology of Invertebrates. Vol 1: Oogenesis, Oviposition, and Oosorption. New York: John Wiley & Sons Ltd. Ch. 18:443—495. Aiken, D. E. 1969a. Photoperiod, endocrinology and the crustacean molt cycle. Science 164:149-—155. 1969b. Ovarian maturation and egg laying in the crayfish Orco- nectes virilis: Influence of temperature and photoperiod. Can. J. Zool. 48:931—935. & S. L. Waddy. 1980. Reproductive biology. Cobb, J. S. and B. F. Phillips, eds. The Biology and Management of Lobsters. Vol. I. Physiology and Behaviour. New York: Academic Press. p. 215-276. Akiyama, D. M. & W. G. Dominy. 1989. Penaeid shrimp nutrition for the commercial feed industry. Texas Shrimp Farming Manual, Volume 1: Growout Technology, Technical Report of the Agricultural Extension Service and Texas A&M University Sea Grant College Pro- gram. Anderson, D. T. 1982. Embryology. Abele, L. G., ed. The Biology of Crustacea, Vol. 2: Embryology, Morphology, and Genetics. New York: Academic Press. p: 1—41. Anderson, S. L., E. S. Chang and W. H. Clark, Jr. 1984. Timing of postvitellogenic ovarian changes in the ridgeback prawn Sicyonia in- gentis (Penaeidae) determined by ovarian biopsy. Aquaculture 42:257-271. Anger, K. 1986. Changes of respiration and biomass of spider crab (Hyas araneus) larvae during starvation. Mar. Biol. 90:261—269. 1988. Growth and elemental composition (C, N, H) in /nachus dorsettensis (Decapoda: Majidae) larvae reared in the laboratory. Mar. Biol. 99:255—260. 1989. Growth and exuvial loss during larval and early juvenile development of the hermit crab Pagurus bernhardus reared in the lab- oratory. Mar. Biol. 103:503—511. *Note: Nippon Suisan Gakkaishi was formerly cited Bull. Jap. Soc. Sci. Fish. , V. Storch, V. Anger, & J. M. Capuzzo. 1985. Effects of starva- tion on moult cycle and hepatopancreas of stage I lobster (Homarus americanus) larvae. Helg. Meers. 39:107—116. Anilkumar, G. 1980. Reproductive physiology of female crustaceans. Ph.D. Thesis, Calicut University. AQUACOP. 1975. Maturation and spawning in captivity of penaeid shrimp: Penaeus merguiensis de Man, Penaeus japonicus Bate, Penaeus aztecus Ives, Metapenaeus ensis de Haan, and Penaeus se- misulcatus de Haan. Proc. World Maricult. Soc. 6:123-132. 1977. Reproduction in captivity and growth of Panaeus monodon, Fabricus in Polynesia. Proc. World Maricult. Soc. 8:927—94S. 1979. Penaeid reared brood stock: closing the life cycle. Proc. World Maricult. Soc. 10:445—452. Armitage, K. B., A. L. Buikema, Jr. & N. J. Willems. 1972. Organic constituents in the annual cycle of the crayfish Orconectes nais (Faxon). Comp. Biochem. Physiol. 41A:825—842. Arnstein, D. R. & T. W. Beard. 1975. Induced maturation of the prawn Penaeus orientalis Kishinouye in the laboratory by means of eyestalk removal. Aquaculture 5:411—412. Beard, T. W. & J. F. Wickins. 1980. The breeding of Penaeus monodon Fabricus in laboratory recirculating systems. Aquaculture 20:79-89. , J. F. Wickins & D. R. Amstein. 1977. The breeding and growth of Penaeus merguiensis De Man in laboratory recirculation systems. Aquaculture 10:275—289. Biesiot, P. 1982. Histological changes associated with development of the hepatopancreas in larvae of the American lobster (Homarus ameri- canus). Pruder, G. D., C. J. Langdon & D. E. Conklin, eds. Pro- ceedings of the Second International Conference in Aquaculture Nu- trition: Biochemical and Physiological Approaches to Shellfish Nutri- tion. 1981 October 27—28. Rehoboth Beach, DE. World Maricult. Soc. Spec. Publ. p. 413. 1986. Changes in midgut gland morphology and digestive enzyme activities associated with development in early stages of the American lobster Homarus americanus. Ph.D. Thesis. Massachusetts Institute of Technology/Woods Hole Oceanographic Institution, WHOI-86-20. Blanchet-Tournier, M. F. 1982. Quelques aspects des interactions hor- monales entre la mue et la vitellogenése chez le Crustacé Amphipode Orchestia gammarellus (Pallas). Reprod. Nutr. Dév. 22:325—344. Brenner, R. R. & A. Bernasconi. 1989. Prostaglandin biosynthesis in the 24 HARRISON gonads of the hematophagus insect Triatoma infestans. Comp. Bio- chem. Physiol. 93B(1):1—4. Browdy, C. L. & T. M. Samocha. 1985a. Maturation and spawning of ablated and nonablated Penaeus semisulcatus de Haan, 1844. J. World Maricult. Soc. 16:236—249. & T.M. Samocha. 1985b. The effect of eyestalk ablation of spawning, molting, and mating of Penaeus semisulcatus de Haan. Aquaculture 48:19—29. , T. M. Samocha, A. Hadani, T. M. Samocha & Y. Loya. 1986. The reproductive performance of wild and pond-reared Penaeus se- misulcatus De Haan. Aquaculture 59:251—258. Brown, A., Jr., J. O. McVey, B. S. Middleditch, & A. L. Lawrence. 1979. Maturation of white shrimp (Penaeus setiferus) in captivity. Proc. World Maricult. Soc. 10:435—444. , J. P. McVey, B. M. Scott, T. D. Williams, B. S. Middleditch & A. L. Lawrence. 1980. The maturation and spawning of Penaeus sty- lirostris under controlled laboratory conditions. Proc. World Maricult. Soc. 11:488—499. Capuzzo, J. M. & B. A. Lancaster. 1979a. Some physiological and bio- chemical considerations of larval development in the American lob- ster, Homarus americanus Milne Edwards. J. exp. mar. Biol. Ecol. 40:53-62. 1979b. The effects of diet on the growth energetics of postlarval lobsters (Homarus americanus). Woods Hole Oceanographic Institu- tion Technical Report WHOI-79-55. 23 p. Castell, J. D. 1982. Fatty acid metabolism in crustaceans. Pruder, G. D., C. J. Langdon & D. E. Conklin, eds. Proceedings of the Second In- ternational Conference in Aquaculture Nutrition: Biochemical and Physiological Approaches to Shellfish Nutrition. 1981 October 27—28. Rehoboth Beach, DE. World Maricult. Soc. Spec. Publ. 2:124—145. Castille, F. C. & A. L. Lawrence. 1989. The relationship between matu- ration and biochemical composition of the gonads and digestive glands of the shrimps Penaeus aztecus Ives and Penaeus setiferus (L.). J. Crust. Biol. 9(2):202—211. Chamberlain, G. W. & A. L. Lawrence. 1981a. Maturation, reproduc- tion, and growth of Penaeus vannamei and P. stylirostris fed natural diets. J. World Maricult Soc. 12:209—224. & A. L. Lawrence. 1981b. Effect of light intensity and male and female eyestalk ablation on reproduction of Penaeus stylirostris and P.. vannamei. J. World Maricult. Soc. 12:357—372. Chang, E. S. 1984. Ecdysteroids in crustacea: role in reproduction, molting, and larval development. Engles, W., W. H. Clark, Jr., A. Fischer, P. J. W. Olive and D. F. Went, eds. Advances in Inverte- brate Reproduction 3. New York: Elsevier Science Pub. p. 223-230. 1989. Endocrine regulation of molting in crustacea. Rev. Aquatic Set. 1(1):131-—157. Charniaux-Cotton, H. 1962. Androgenic gland of crustaceans. Gen. Comp. Endocrinol. Supplement 1:241—247. 1980. Experimental studies of reproduction in malocostraca crus- taceans. Description of vitellogenesis and of its endocrine control. Clark, W. H. Jr. and T. S. Adams, eds. Advances in Invertebrate Re- production, Vol. II. North Holland, Amsterdam: Elsevier. p. 177-186. 1985. Vitellogenesis and its control in malacostracan crustacea. Am. Zool. 25:197—206. Cheesman, D. F., W. L. Lee, & P. F. Zagalsky. 1967. Carotenoproteins in invertebrates. Biol. Rev. 42:131—160. Clarke, A. 1982. Lipid synthesis and reproduction in the polar shrimp Chorismus antarcticus. Mar. Ecol. Prog. Ser. 9(1):81—90. , A. Skadsheim & L. J. Holmes. 1985. Lipid biochemistry and reproductive biology in two species of Gammaridae (Crustacea: Am- phipoda). Mar. Biol. 88:247—263. Claybrook, D. L. 1983. Nitrogen metabolism. Mantel, L. H., ed. The Biology of Crustacea, Vol 5. New York, N.Y.: Academic Press. Ch. 3:163—213 Clifford, H. C., Ill. & R. W. Brick. 1983. Nutritional physiology of the freshwater shrimp Macrobrachium rosenbergii (De Man)—I. Sub- strate metabolism in fasting juvenile shrimp. Comp. Biochem. Physiol. 74A(3):561—568. Conklin, D. E. 1980. Nutrition. Cobb, J. S. and B. F. Phillips, eds. The Biology and Management of Lobsters, Vol. 1. Physiology and Behav- iour. New York: Academic Press. p. 277—300. 1982. The role of macronutrients in the biosynthesis of the crusta- cean exoskeleton. Pruder, G. D., C. J. Langdon & D. E. Conklin, eds. Proceedings of the Second International Conference in Aquacul- ture Nutrition: Biochemical and Physiological Approaches to Shellfish Nutrition. 1981 October 27—28. Rehoboth Beach, DE. World Mari- cult. Soc. Spec. Publ. 2:146—165. Cook, H. W. Department of Biochemistry, Dalhousie University, Halifax, NS, Canada. Corbin, J. S., M. M. Fujimoto and T. Y. Iwai, Jr. 1983. Feeding prac- tices and nutritional considerations for Macrobrachium rosenbergii culture in Hawaii. McVey, J. P., ed. Handbook of Mariculture Vol 1. Crustacean Aquaculture. Boca Raton, Florida: CRC Press. p. 391-412. D’Abramo, L. R., C. E. Bordner, D. E. Conklin & N. E. Baum. 1984. Sterol requirement of juvenile lobsters, Homarus sp. Aquaculture 42:13-25. & D. E. Conklin. 1985. Lobster Aquaculture. Huner, J. V. and E. E. Brown, eds. Crustacean and Mollusk Aquaculture in the United States. Westport, Conn.: AVI Publishing Co. p. 159-201. Dall, W. & D. J. W. Monatry. 1983. Functional aspects of nutrition and digestion. Mantel, L. H., ed. The Biology of Crustacea, Vol 5. New York, N.Y.: Academic Press, Inc. Ch. 4:215—261. Dawirs, R. R., C. Puschel & F. Schorn. 1986. Temperature and growth in Carcinus maenas L. (Decapoda: Portunidae) larvae reared in the labo- ratory from hatching through metamorphosis. J. Exp. Mar. Biol. Ecol. 100:47-—74. Dehn, P. F., D. E. Aiken & S. L. Waddy. 1983. Aspects of vitello- genesis in the lobster Homarus americanus. Can. Tech. Rep. Fish. Aquat. Sci. No. 1161. 24 p. Deshimaru, O. 1982. Protein and amino acid nutrition of the prawn, Penaeus japonicus. Pruder, G. D., C. J. Langdon & D. E. Conklin, eds. Proceedings of the Second International Conference in Aquacul- ture Nutrition: Biochemical and Physiological Approaches to Shellfish Nutrition. 1981 October 27—28. Rehoboth Beach, DE. World Mari- cult. Soc. Spec. Publ. 2:106—122. Durliat, M. 1984. Occurrence of plasma proteins in ovary and egg extracts from Astacus leptodactylus. Comp. Biochem. Physiol. 78B:745—753. & K. Kuroki. 1974. Studies on a purified diet for prawn. I. Basal composition of diet. Bull. Jap. Soc. Sci. Fish. 40(4):413-419. Emmerson, W. D. 1980. Induced maturation of prawn Penaeus indicus. Mar. Ecol. Prog. Ser. 2:121—131. 1983. Maturation and growth of ablated and unablated Penaeus monodon Fabricus. Aquaculture 32:235—241. , D. P. Hayes & M. Ngonyama. 1983. Growth and maturation of Penaeus indicus under blue and green light. S. Afr. J. Zool. 18:71—75. Fingerman, M. 1987. The endocrine mechanisms in crustaceans. J. Crust. Biol. 7(1):1—24. Fisher, L. R. & S. K. Kon. 1958. Vitamin A in the invertebrates. Biol. Rev. 34:1—36. Fyffe, W. E. & J. D. O'Connor. 1974. Characterization and quantifica- tion of a crustacean lipovitellin. Comp. Biochem. Physiol. 47B:851— 867. Frank, J. R., S. D. Sulkin & R. P. Morgan II. 1975. Biochemical changes during larval development of the xanthid crab Rhithropa- nopeus harrisii. 1. Protein, total lipid, alkaline phosphatase, and glu- tamic oxaloacetic transaminase. Mar. Biol. 32:105—111. Galgani, M.-L. & AQUACOP (J. Goguenheim, F. Galgani, G. Cuzon) 1989a. Diets in relation to reproduction of Penaeus vannamei and Penaeus stylirostris in captivity. Aquaculture 80(1—2):97—109. CRUSTACEAN BROODSTOCK NUTRITION: A REVIEW , G. Cuzon, F. Galgani & J. Goguenheim. 1989b. Influence de régime alimentaire sur la reproduction en captivité de Penaeus indicus. Aquaculture 81(3—4):337—350. Galois, R. G. 1984. Variations de la composition lipidique tissulaire au cours de la vitellogenése chez la crevette Penaeus indicus Milne Ed- wards. J. Exp. Mar. Biol. Ecol. 84:155—166. Gehring, W. R. 1974. Maturational changes in the ovarian lipid spectrum of the pink shrimp Penaeus duorarum duorarum Burkenroad. Comp. Biochem. Physiol. 49A:511—524. Gibson, R. 1982. Feeding and digestion in decapod crustaceans. Pruder, G. D., C. J. Langdon & D. E. Conklin, eds. Proceedings of the Second International Conference in Aquaculture Nutrition: Biochem- ical and Physiological Approaches to Shellfish Nutrition. 1981 Oc- tober 27—28. Rehoboth Beach, DE. World Maricult. Soc. Spec. Publ. 2:59-85. & P. L. Barker. 1979. The Decapod hepatopancreas. Oceanogr. Mar. Biol. Ann. Rev. 17:285—346. Gilchrist, B. M. & W. L. Lee. 1972. Carotenoid pigments and their pos- sible role in reproduction in the sand crab, Emerita anologa (Stimp- son, 1857). Comp. Biochem. Physiol. 42B:263—294. Gilchrist, B. M. & P. F. Zagalsky. 1983. Isolation of a blue canthax- anthin-protein from connective tissue storage cells in Branchinecta packardi Pearse (Crustacea: Anostraca) and its possible role in vitello- genesis. Comp. Biochem. Physiol. 76B(4):885—893. Goguenheim, J., J. Barret, J. Patrois, J. Cahu, C. Fauvel, AQUACOP. 1987. Penaeus vannamei: Broodstock constitution, maturation and ar- tificial insemination. Abstr. 18th Ann. Mtg. World Aquaculture So- ciety, Jan. 18—23, 1987, Guayaquil, Ecuador. Goudeau, M. 1976. Secretion of embryonic envelopes and embryonic molting cycles in Hemioniscus balanai Isopoda Epicaridea. J. Mor- phol. 148:427—451. Guary, J. C., M. Kayama & Y. Murakami. 1974. Lipid class distribution and fatty acid composition of prawn, Penaeus japonicus Bate. Bull. Jap. Soc. Sci. Fish. 40:1027—1032. Herring, P. J. 1968. The carotenoid pigments of Daphnia magna Straus —Il. Aspects of pigmentary metabolism. Comp. Biochem. Physiol. 24:205-221. Hilton, J. W. 1984. Ascorbic acid-mineral interactions in fish. Wegger, I., F. J. Tagwerker & J. Moustgaard, eds. Proceedings of workshop on Ascorbic Acid in Domestic Animals. Copenhagen: Royal Danish Agricultural Society: p. 218-224. Holland, D. L. 1978. Lipid reserves and energy metabolism in the larvae of benthic marine invertebrates. Malins, D. C. and S. R. Sargent, eds. Biochemical and Biophysical Perspectives in Marine Biology, Vol III. New York: Academic Press: p. 85-123. Jeckel, W. H., J. E. A. de Moreno & V. J. Moreno. 1989a. Biochemical composition, lipid classes and fatty acids in the ovary of the shrimp Pleoticus muelleri Bate. Comp. Biochem. Physiol. 92B:(2):271—276. , J. E. A. de Moreno & V. J. Moreno. 1989b. Biochemical com- position, lipid classes and fatty acids in the male reproductive system of the shrimp Pleoticus muelleri Bate. Comp. Biochem. Physiol. 93B(4):807—-811. Kanazawa, A. 1986. New developments in fish nutrition. MacLean, J. L., L. B. Dizon and L. V. Hosillos, eds. The First Asian Fisheries Forum. Asian Fisheries Society, Manilla, Philippines. Philippines p. 9-14. , Liet-Chim & A. Laubier. 1988. Tissue uptake of radioactive cholesterol in the prawn Penaeus japonicus Bate during induced ovarian maturation. Aquat. Living Resour. 1:85—91. , A. N. Tanaka, S. Teshima & K. Kashiwada. 1971b. Nutritional requirements of prawn. III. Utilization of the dietary sterols. Bull. Jap. Soc. Sci. Fish. 37:1015—1019. & S. Teshima. 1971. In vivo conversion of cholesterol to steroid hormones in the spiny lobster, Panulirus japonicus. Bull. Jap. Soc. Sci. Fish. 37:891—897. i) Nn & S. Teshima. 1981: Essential amino acids of the prawn. Bull. Jap. Soc. Sci. Fish. 47:1375—1377. Kean, J. C., J. D. Castell, A. G. Boghen, L. R. D’Abramo, & D. E. Conklin. 1985. A re-evaluation of the lecithin and cholesterol require- ments of juvenile lobster (Homarus americanus) using crab protein- based diets. Aquaculture 47:143—149. Kelemec, J. A. & I. R. Smith. 1980. Induced ovarian development and spawning of Penaeus plebejus in a recirculating laboratory tank after unilateral eyestalk enucleation. Aquaculture 21:55—62. Kessel, R. G. 1968. Mechanisms of protein yolk synthesis and deposition in crustacean oocytes. Z. Zellforsch. Microsk. Anat. 89:17—38. [refer- ence 1968b, cited in Wolin et al., 1973] Kulkarni, G. K. & R. Nagabhushanam. 1979. Mobilization of organic reserves during ovarian development in a marine penaeid prawn, Parapenaeopsis hardwickii (Miers) (Crustacea, Decapoda, Pe- naeidae). Aquaculture 18:373—377. & R. Nagabhushanam. 1980. Role of ovary-inhibiting hormone from eyestalks of marine penaeid prawns (Parapenaeopsis hardwickit) during ovarian development cycle. Aquaculture 19(1):13-19. Lall, S. P. & J. A. Hines. 1985. Manganese bioavailability and require- ments of Atlantic salmon (Salmo salar) and brook trout (Salvelinus fontinalis) Abstr. Symp. on assessment by physiological and biochem- ical methods of nutrient availability in fish, Brighton, U.K. Laubier-Bonichon, A. & L. Laubier. 1976. Reproduction contrélée chez la crevette Penaeus japonicus. FAO Technical Conference on Aqua- culture, Kyoto, Japan 26 May—2 June 1976, FAO-FIR: AQ/Conf/76/ E.38. p. 1-6. Lautier, J. & J.-G. Lagarrigue. 1988. Lipid metabolism of the crab Pa- chygrapsus marmoratus during vitellogenesis. Biochemical System- atics and Ecology 16(2):203—212. Lawrence, A. L., D. Ward, S. Missler, A. Brown, J. McVey, & B. S. Middleditch. 1979. Organ indices and biochemical levels of ova from penaeid shrimp maintained in captivity versus those captured in the wild. Proc. World Maricult. Soc. 10:453—463. , Y. Akamine, B. S. Middleditch, G. Chamberlain, & D. Hut- chins. 1980. Maturation and reproduction of Penaeus setiferus in cap- tivity. Proc. World Maricult. Soc. 11:481—487. Lee, C. L. & D. R. Fielder. 1982. Induced spawning in the freshwater prawn, Macrobrachium australiense Holthuis, 1950 (Crustacea: Deca- poda: Palaemonidae). Aquaculture 29:45—52. Lee, R. F. & D. L. Puppione. 1978. Serum lipoproteins in the spiny lob- ster, Panulirus interruptus. Comp. Biochem. Physiol. 59B:239—243. Lee, R. F. & D. L. Puppione. 1988. Lipoproteins I and II from the hemo- lymph of the blue crab Callinectes sapidus: Lipoprotein II associated with vitellogenesis. J. Exp. Zool. 248:278—289. Leung-Trujillo, J. R. & A. L. Lawerence. 1987. Observations on the de- cline of sperm quality of Penaeus setiferus under laboratory condi- tions. Aquaculture 65:363—370. Levine, M. & K. Morita. 1985. Ascorbic acid in endocrine systems. Vitam. Horm. 42:1—64. Lim, C. & A. Persyn. 1989. Practical feeding—penaeid shrimps. Lovell, T., ed. Nutrition and Feeding of Fish. New York: Van Nostrand Reinhold. p. 205—222. Lui, C. W. & J. D. O’Connor. 1976. Biosynthesis of lipovitellin by the crustacean ovary. II. Characterization of and in vitro incorporation of amino acids into purified subunits. J. Exp. Zool. 155:41—52. & J. D. O'Connor. 1977. Biosynthesis of crustacean lipovitellin. III. The incorporation of labeled amino acids into the purified lipovi- tellin of the crab Pachygrapsus crassipes. J. Exp. Zool. 199:105—108. , B. A. Sage & J. D. O'Connor. 1974. Biosynthesis of lipovitellin by the crustacean ovary. J. Exp. Zool. 188:289—296. Lumare, F. 1979. Reproduction of Penaeus kerathurus using eyestalk ab- lation. Aquaculture 18(3):203—214. 1986. Artificial reproduction of Penaeus japonicus Bate as basis for commercial production of eggs and larvae. Proceedings of the 26 HARRISON EMS/WMS Conference on Aquaculture, Venice, Italy, 21—25 Sep- tember, 1981, pp. 1—20. Luquet, P. & T. Watanabe. 1986. Interaction ‘‘nutrition-reproduction’’ in fish. Fish Physiology and Biochemistry. 2(1—4):121—129. Makinouchi, S. & J. H. Primavera. 1987. Maturation and spawning of Penaeus indicus using different ablation methods. Aquaculture 62:73— 81. Maugle, P., T. Kamata, S. McLean, K. L. Simpson, & T. Katayama. 1980. The influence of eyestalk ablation on the carotenoid composition of juvenile Macrobrachium rosenbergii. Bull. Jap. Soc. Sci. Fish. 46(7):901—904. McIntosh, R. P. 1989. Aquatic Farms Hawaii Ltd. (On assignment in Bangkok, Thailand). McLaughlin, P. 1983. Internal Anatomy. Mantel, L. H., ed. The Biology of Crustacea, Vol 5. New York: N.Y.: Academic Press, Inc. Ch. 1, p. 1-52. Meusy, J. J. & H. Charniaux-Cotton. 1984. Endocrine control of vitello- genesis in Malocostraca crustaceans. Engles, W. H. Clark, Jr., A. Fischer, P. J. W. Olive and D. F. Went, eds. Advances in Inverte- brate Reproduction 3. New York: Elsevier Science Pub. p. 231-241. Middleditch, B. S., S. R. Missler, H. B. Hines, E. S. Chang, J. P. McVey, A. Brown & A. L. Lawrence. 1980a. Maturation of penaeid shrimp: lipids in the marine food web. Proc. World Maricult. Soc. 11:463—470. , S. R. Missler, H. B. Hines, J. P. McVey, A. Brown, D. G. Ward & A. L. Lawrence. 1980b. Metabolic profiles of penaeid shrimp: dietary lipids and ovarian maturation. J. Chromatogr. 195:359—368. , 5. R. Missler, D. G. Ward, J. P. McVey, A. Brown & A. L. Lawrence. 1979. Maturation of penaeid shrimp: dietary fatty acids. Proc. World Maricult. Soc. 10:472—476. Millamena, O. M., R. Pudadera & M. R. Catacutan. 1984. Variation in tissue lipid content during ovarian maturation of unablated and ablated Penaeus monodon broodstock. Proc. First International Conference on the Culture of Penaeid Prawns/Shrimps, Iloila City, Philippines, 4-7 Dec., 1984. p. 166 (Abstract). Moore, L. 1985. The role of feeds and feeding in aquatic animal produc- tion. GeoJournal 10(3):245—251. National Research Council. 1983. Nutrient Requirements of Warmwater Fishes and Shellfishes, Revised Edition. National Academy Press, Wash. D.C. 102 pp. Nagabhushanam, R. & G. K. Kulkarni. 1981. Effect of exogenous testos- terone on the androgenic gland and testis of a marine penaeid prawn, Parapenaeopsis hardwickii (Miers) (Crustacea, Decapoda, Pe- naeidae). Aquaculture 23:19—27. Nelis, H. J., P. Lavens, L. Moens, P. Sorgeloos, & A. P. De Leenheer. 1989. Carotenoids in relation to Artemia development. MacRae, T. H., J. C. Bagshaw and A. H. Warner, eds. Biochemistry and Cell Biology of Artemia. CRC Press, Boca Raton, Florida. 7:160—190. Nelson, K., D. Hedgecock & W. Borgeson. 1983. Photoperiodic and ec- dysial control of vitellogenesis in lobsters (Homarus) (Decapoda, Nephropidae). Can. J. Fish. Aquat. Sci. 40:940—947. , D. Hedgecock & W. Borgeson. 1988a. Factors influencing egg extrusion in the American lobster (Homarus americanus). Can. J. Fish. Aquat. Sci. 45:797—804. , D. Hedgecock & W. Borgeson. 1988b. Effects of reproduction upon molting and growth in female American lobsters (Homarus americanus). Can. J. Fish. Aquat. Sci. 45:805—821. Neville, A. C. 1975. Biology of the Anthropod Cuticle. Berlin and New York: Springer-Verlag. Ogle, J. T. 1989. Gulf Coast Research Laboratory, Ocean Springs, MS, USA. Otazu-Abnill, M. & H. J. Ceccaldi. 1984. Influence of purified carote- noida added to compounded diets on pigmentation of Penaeus ja- ponicus (Crustacea, Decapoda). Aquaculture 36:217—228. Paulus, J. E. & H. Laufer. 1987. Vitellogenocytes in the hepatopancreas of Carcinus maenas and Libinia emarginata (Decapoda, Brachyura). Int. J. Invert. Reprod. Dev. 11:29—44. Penaflorida, V. Dy. 1989. An evaluation of indigenous protein sources as potential component in the diet formulation for tiger prawn, Penaeus monodon, using essential amino acid index (EAAI). Aquaculture 83: 319-330. Pillay, K. K. & N. B. Nair. 1973. Observations on the biochemical changes in gonads and other organs of Uca annulipes, Portunus pela- gicus, and Matapenaeus affinis (Decapoda: Crustacea) during the re- productive cycle. Mar. Biol. 18(3):167—198. Pochon-Masson, J. 1983. Arthropoda-Crustacea. Adiyodi, K. G. and R. G. Adiyodi, eds. Reproductive Biology of Invertebrates. Vol II: Spermatogenesis and Sperm Function. New York: John Wiley & Sons Ltd. Ch 21:407—449. , J., G. G. Payen, C. Portemer and F. Chatagner. 1984. Variations of taurine concentrations in the ovaries and hepatopancreas of the fe- male crab Carcinus maenas L. (Decapoda Brachyura) during the dif- ferent phases of its genital activity. nt. J. Invert Reprod. Dev. 7:127— 133. Primavera, J. H. 1978. Induced maturation and spawning in five-month- old Penaeus monodon Fabricus by eyestalk ablation. Aquaculture 13:355-359. & R. A. Posadas. 1981. Studies on the egg quality of Penaeus monodon Fabricus based on morphology and hatching rates. Aquacul- ture 22:269—277. Quackenbush, L. S. 1986. Crustacean endocrinology: a review. Can. J. Fish. Aquat. Sci. 43:2271—2282. 1989. Vitellogenesis in the shrimp, Penaeus vannamei: in vitro studies of the isolated hepatopancreas and ovary. Comp. Biochem. Physiol. 94B(2):253—261. & W. F. Hermkind. 1981. Regulation of molt and gonadal devel- opment in the spiny lobster, Panulirus argus (Crustaceana: Palinu- ridae): effect of eyestalk ablation. Comp. Biochem. Physiol. 69A:523-—527. Quinito, E. T., A. Hara, K. Yamauchi, T. Mizushima, & A. Fuji. 1989. Identification and characterization of vitellin in a hermaphrodite shrimp, Pandalus kessleri. Comp. Biochem. Physiol. 94B(3):445— 451. Ra’Anan, Z. 1987. Maturation vs. growth in males and females of the fresh water prawn Macrobrachium rosenbergii—description of pat- tern and implications to aquacultural practices. (Abstract) 18th Ann. Mtg. World Aquaculture Society, Jan. 18—23, 1987, Guayaquil, Ecuador. Rao, P. V. 1978. Maturation and spawning of cultivable marine penaeid prawns. CMFRI Special Publication on Summer Institute of Breeding and Rearing of Marine Prawns. Cochin. No. 3:57—67. Read, G. H. L. & M. S. Caulton. 1980. Changes in mass and chemical composition during the molt cycle and ovarian development in imma- ture and mature Penaeus indicus Milne Edwards. Comp. Biochem. Physiol. 66A:431—437. Richard, P. & H. J. Ceccaldi. 1977. Variations des caractéristiques pon- dérales et des compositions amino-acide et protéique pendant le déve- loppement embryonnaire de Palaemon serratus. 3rd Meeting ICES Work. Group. Maric. Brest 10-13 Mai 1977. Actes Coll. CNEXO. 4:203—212. Ryckaert, M., N. Vial & H. J. Ceccaldi. 1974. Effect de l’ablation des pédoncles oculaires en longue et courte photophase sur le dévelopment des gonades chez Palaemon serratus. C. R. Soc. Biol. 168:1364— 1368. Sandifer, P. 1986. Recent advances in the culture of crustaceans. Pro- ceedings of the EMS/WMS World Conference on Aquaculture, Venice, Italy, 21-25 September, 1981. 47 p. & J. Lynn. 1982. Artificial insemination of caridean shrimp. Clark, W. H., Jr. and T. S. Adams, eds. Advances in Invertebrate Reproduction. North Holland, Amsterdam: Elsevier, p. 271—288. & T. 1. J. Smith. 1985. Freshwater Prawns. Huner, J. V. and CRUSTACEAN BROODSTOCK NUTRITION: A REVIEW 27 E. E. Brown, eds. Crustacean and Mollusk Aquaculture in the United States. Westport, Conn.: AVI Publishing Co. p. 63-125. Sandnes, K. 1984. Some aspects of ascorbic acid and reproduction in fish. Wegger, I., F. J. Tagwerker & J. Moustgaard, eds. Proceedings of Workshop on Ascorbic Acid in Domestic Animals. Copenhagen: Royal Danish Agricultural Society: p. 206—212. . K. Julshamn & O. R. Braekkan. 1984a. Interrelationship be- tween ascorbic acid and trace elements on ovarian development in fish. Wegger, I., F. J. Tagwerker & J. Moustgaard, eds. Proceedings of Workshop on Ascorbic Acid in Domestic Animals. Copenhagen: Royal Danish Agricultural Society: p. 213-217. , Y. Ulgenes, O. L. Braekkan, & F. Utne. 1984b. The effect of ascorbic acid supplementation in broodstock feed on reproduction of rainbow trout (Salmo gairdneri). Aquaculture 43:167—177. Sasaki, G. C. 1984. Biochemical changes associated with embryonic de- velopment and larval development in the American lobster Homarus americanus Milne Edwards. Ph.D. Thesis, Woods Hole Oceano- graphic Institution, Woods Hole, Mass., Massachusetts Institute of Technology, Cambridge, Mass., 458 pp. , J. McD. Capuzzo & P. Biesiot. 1986. Nutritional and bioener- getic considerations in the development of the American lobster Ho- marus americanus. Can. J. Fish. Aquat. Sci. 43:2311—2319. Shigueno, K. & S. Itoh. 1988. Use of Mg-L-ascorbyl-2-phosphate as a vitamin C source in shrimp diets. J. World Aquacult. Soc. 19(4):168— 174. Sochasky, W. H., D. E. Aiken & D. W. McLeese. 1973. Does eyestalk ablation accelerate molting in the lobster Homarus americanus? J. Fish. Res. Bd. Can. 30:1600—1603. Soliman, A. K. K. Jauncey & R. J. Roberts. 1986. The effect of dietary ascorbic acid supplementation on hatchability, survival rate and fry performance in Oreochromis mossambicus (Peters). Aquaculture 59:197—208. Spaargaren, D. H. 1988. Presence and significance of lithium in the brown shrimp, Crangon crangon. Comp. Biochem. Physiol. 91A(3): 457-461. Stanley-Samuelson, D. W. & W. Loher. 1983. Arachidonic acid and other long chain polyunsaturated fatty acids in spermatophores and spermathecae of Teleogryllus commodus: significance in prosta- glandin-mediated reproductive behaviours. J. /nsect. Physiol. 29:41— 45. & W. Loher. 1986. Prostaglandins in insect reproduction. Forum Ann. Entomol. Soc. Am. 79:841—853. Takeuchi, M., S. Ishi & T. Ogino. 1981. Effect of dietary vitamin E on growth, vitamin E distribution and mortalities of the fertilized eggs and fry in ayu Plecoglossus altivelis. Bull. Tokai Reg. Fish. Res. Lab. 104:111—122. Talbot, P. 1984. Problems and progress in controlling reproductive bi- ology in the American lobster (Homarus). Clark, W. H., Jr. and T. S. Adams, eds. Advances in Invertebrate Reproduction. North Holland, Amsterdam: Elsevier, p. 473-480. , D. Howard, J. Leung-Trujillo, T. W. Lee, W.-Y. Li, H. Ro & A. L. Lawrence. 1989. Characterization of male reproductive tract degenerative syndrome in captive penaeid shrimp (Penaeus setiferus). Aquaculture 78(3—4):365—377. Tan-Fermin, J. D. & R. A. Pudadera. 1989. Ovarian maturation stages of the wild giant tiger prawn, Penaeus monodon Fabricus. Aquaculture 77(2—3):229-242. Teshima, S. 1982. Sterol metabolism. Pruder, G. D., C. J. Langdon & D. E. Conklin, eds. Proceedings of the Second International Confer- ence in Aquaculture Nutrition: Biochemical and Physiological Ap- proaches to Shellfish Nutrition. 1981 October 27—28. Rehoboth Beach, DE. World Maricult. Soc. Spec. Publ. 2:205—216. & A. Kanazawa. 1971a. Biosynthesis of sterols in the lobster, Panulirus japonica, the prawn, Penaeus japonicus, and crab, Por- tunus trituberculatus. Comp. Biochem. Physiol. 38B:597—602. & A. Kanazawa. 1971b. Bioconversion of progesterone by the ovaries of crab, Portunus trituberculatus. Gen. Comp. Endocrinol. 17:152—157. & A. Kanazawa. 1978. Release and transport of lipids in the prawn. Bull. Jap. Soc. Sci. Fish. 44:1269—1274. & A. Kanazawa. 1979. Lipid transport mechanism in the prawn. Bull. Jap. Soc. Sci. Fish. 45(10):1341—1346. & A. Kanazawa. 1980a. Transport of dietary lipids and role of serum lipoproteins in the prawn. Bull. Jap. Soc. Sci. Fish. 46(1):51— 55. & A. Kanazawa. 1983. Variation in lipid compositions during the ovarian maturation of the prawn. Bull. Jap. Soc. Sci. Fish. 49(6): 957-962. , A. Kanazawa, S. Koshio & K. Horinouchi. 1988a. Lipid metab- olism in destalked prawn Penaeus japonicus: Induced maturation and accumulation of lipids in the ovaries. Nippon Suisan Gakkaishi. 54(7):1115—1122. , A. Kanazawa, K. Horinouchi & S. Koshio. 1988b. Lipid metab- olism in destalked prawn Penaeus japonicus: Induced maturation and transfer of lipid reserves to the ovaries. Nippon Suisan Gakkaishi. 54(7):1123-1129. , A. Kanazawa, S. Koshio & K. Horinouchi. 1989. Lipid metabo- lism of the prawn Penaeus japonicus during maturation: variation in lipid profiles of the ovary and hepatopancreas. Comp. Biochem. Physiol. 92B(1):45—49. , A. Kanazawa & H. Sasada. 1983. Nutritional value of dietary cholesterol and other sterols to larval prawn, Penaeus japonicus Bate. Aquaculture. 31:159—167. Tobe, S. S., D. A. Young, H. W. Khoo, and F. C. Baker. 1989. Farne- soic acid as a major product of release from crustacean mandibular organs in vitro. J. Exp. Zool. 249:165—171. Tom, M., M. Goren & M. Ovadia. 1987. Localization of the vitellin and its possible precursors in various organs of Parapenaeus longirostris (Crustacea, Decapoda, Penaeidae). Int. J. Invert. Reprod. Dev. 12:1— 122 Van Beek, E. & A. De Loof. 1988. Radioimmunological determinations of concentrations of six C,,, Cy) and Cj, steroids during the reproduc- tive cycle of female Artemia sp. (Crustacea: Anostraca). Comp. Bio- chem. Physiol. 89A(4):S95—599. Van den Oord, A. 1966. The biosynthesis of the emulsifiers of the crab Cancer pagurus L. Comp. Biochem. Physiol. 17:715—718. Varadarajan, S. & T. Subramonium. 1980. Histochemical investigations on vitellogenesis of an anomuran crab, Clibanarius clibanarius. Int. J. Invert. Reprod. 2:47-58. Vazquez-Boucard, C., H. J. Ceccaldi, Y. Benyamin & C. Roustan. 1986. Identification, purification, et caractérisation de la lipovitelline chez un crustacé décapode natantia Penaeus japonicus (Bate). J. Exp. Mar. Biol. Ecol. 97:37-—S0. Veith, W. J. & S. R. Malecha. 1983. Histochemical study of the distribu- tion of lipids, 3a- and 3R-hydroxysteroid dehydrogenase in the andro- genic gland of the cultured prawn, Macrobrachium rosenbergii (de Man) (Crustacea; Decapoda). S. Afr. J. of Sci. 79(3):83—85. Vensel, W. H., E. Spaziani & L. S. Ostegaard. 1984. Cholesterol turn- over and ecdysone content in tissues of normal and de-stalked crabs (Cancer antennarius). J. Exp. Zool. 229:383—392. Vincent, M. 1989. Influence of water temperature on carotenoids and ca- rotenoid metabolism in Palaemon serratus (Pennant) (Crustacea: De- capoda). Biochem. System. Ecol. 17(4):319—322. Vincent, M. & H. J. Ceccaldi. 1988. Relations entre acides gras et pig- ments caroténoides chez un crustacé copépode, Calanipeda aquae- bulcis. Biochem. System. Ecol. 16(3):317—324. Vincent, M., L. Ramos & M. Oliva. 1988. Variations in the total carote- noid pigments in the ovary and hepatopancreas of Penaeus schmitti during ovarian maturation. Biochem. Syst. Ecol. 16(4):431—436. Vogt, G., V. Storch, E. T. Quinito & F. P. Pascual. 1985. Midgut gland as monitor organ for the nutritional value of diets in Penaeus monodon (Decapoda). Aquaculture 48:1—12. 28 HARRISON , E. T. Quinito & F. P. Pascual. 1987. Interaction of the midgut and the ovary in vitellogenesis and consequences for the breeding suc- cess: A comparison of unablated and ablated spawners of Penaeus monodon. Preprint. Waagbo, R., T. Thorsen & K. Sandness. 1989. Role of dietary ascorbic acid in vitellogenesis in rainbow trout (Salmon gairdneri). Aquacul- ture 80:301—314. Ward, D. G., B. S. Middleditch, S. R. Missler, and A. L. Lawrence. 1979. Fatty acid changes during larval development of Penaeus seti- ferus. Proc. World Maricult. Soc. 10:464—471. Watanabe, T. & F. Takeshima. 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. Watson, R. D. & E. Spaziani. 1985. Biosynthesis of ecdysteroids by crab Y-organs, and eyestalk suppression of cholesterol uptake and secretory activity, in vitro. Gen. Comp. Endocrinol. 59:140—148. Whitney, J. O. 1969. Acta Embryol. Exp. 111. Wickins, J. F. 1982. Opportunities for farming crustaceans in western temperate regions. Muir, J. F. & R. J. Roberts, eds. Recent Advances in Aquaculture. Colorado: Westview Press. 2:87-—176. & T. W. Beard. 1974. Observations on the breeding and growth of the giant freshwater prawn Macrobrachium rosenbergii (De Man) in the laboratory. Aquaculture. 3:159-174. Wolin, E. M., H. Laufer & D. F. Albertini. 1973. Uptake of the yolk protein, lipovitellin, by developing crustacean oocytes. Dev. Biol. 35:160—170. Wurts, W. A. & R. R. Stickney. 1984. An hypothesis on the light re- quirements for spawning penaeid shrimp, with emphasis on Penaeus setiferus. Aquaculture 41(1):93—98. Yano, I. 1984. Induction of rapid spawning in kuruma prawn, Penaeus japonicus, through unilateral eyestalk enucleation. Aquaculture 40: 265-268. Yano, I. 1985. Induced ovarian maturation and spawning in greasyback ABBREVIATIONS DG, diacylglyceride; EFA, essential fatty acid; FFA, free fatty acid; GIH, gonad inhibiting hormone; HDL, high density lipoprotein; LV, lipovitellin, MG, monoaclygly- ceride; MIH, molt inhibiting hormone; NEFA, non-essen- tial fatty acid; PL, phosphoglyceride, PUFA, polyunsatu- rated fatty acid; RER, rough endoplasmic reticulum; SER, smooth endoplasmic reticulum; TG, triaclyglyceride; VP, Vitellin, VG, vitellogenin. Addendum: The author would like to thank Dr. George Chamberlain for reading the galley proofs and bringing to my attention re- cent work by Dr. L. Keeley’s group in the Entomology Department at Texas A&M University, not reviewed in this shrimp, Metapenaeus ensis, by progesterone. Aquaculture 47:223- 229. 1987. Effect of 17a-hydroxy-progesterone on vitellogenin secre- tion in kuruma prawn, Penaeus japonicus. Aquaculture 61:49—57. & Y. Chinzei. 1987. Ovary is the site of vitellogenin synthesis in kuruma prawn, Penaeus japonicus. Comp. Biochem. Physiol. 86B(2):213—218. , B. Tsukimura, J. N. Sweeney & J. A. Wyban. 1988. Induced Ovarian maturation of Penaeus vannamei by implantation of lobster ganglion. J. World Aquacult. Soc. 19(4):204—209. Yonge, C. M. 1938. The nature and significance of the membranes sur- rounding the developing eggs of Homarus vulgaris and other Deca- poda. J. Zool. 107:499-517. Zagalsky, P. F. 1972. Comparative studies on the amino acid composi- tions of some carotenoid containing lipoglycoproteins and glycopro- teins from the eggs and ovaries of certain aquatic invertebrates. Comp. Biochem. Physiol. 41B(2):385—395. Zagalsky, P. F., D. F. Cheesman & H. J. Ceccaldi. 1967. Studies on carotenoid-containing lipoproteins isolated from eggs and ovaries of certain marine invertebrates. Comp. Biochem. Physiol. 22:851—871. Zandee, D. I. 1967. Absence of cholesterol synthesis as contrasted with the presence of fatty acid synthesis in some anthropods. Comp. Bio- chem. Physiol. 20:811—822. & E.C. Kruitwagen. 1975. Depot sterols in comparison with structural sterols in Cancer pagurus and Eriocheir sinensis. Neth J. Sea Res. 9:214—221. Zeitoun, I. H., D. E. Ullrey, W. G. Bergen, & W. T. Magee. 1976. Mineral metabolism during ontogenesis of rainbow trout (Salmo gairdneri). J. Fish. Res. Board Can. 33:2587-—2591. Zerbib, C. & J. J. Mustel. 1984. Incorporation de la vitellogénine tritée dans les ovocytes du Crustacé Amphipode Orchestia gammarellus (Pallas). Int. J. Invert. Reprod. Dev. 7:63-68. manuscript. Two points from their work with Penaeus van- namel should be noted: 1) They identified the ovary as the primary source of protein and found little evidence of syn- thesis of vitellogenin in the hepatopancreas. This is con- trary to their findings with Uca, where the hepatopancreas is a primary source. 2) They have identified two major pro- teins in the egg (cortical gel protein, and yolk protein) and have antibodies for these. Additional references include: Quackenbush, L. S. 1989. Yolk protein production in the marine shrimp Penaeus vannamei. J. Crust. Biol. 9(4):509—516. Rankin, S. M., J. Y. Bradfield and L. L. Keeley. 1990. Ovarian develop- ment in the South American white shrimp, Penaeus vannamei. N.O.A.A. Tech. Rep. N.M.F.S. Series [in Press]. Rankin, S. M., J. Y. Bradfield and L. L. Keeley. 1989. Ovarian protein synthesis in the South American white shrimp, Penaeus vannamei, during the reproductive cycle. Invert. Reprod. and Dev. 15:27—33. Journal of Shellfish Research, Vol. 9, No. 1, 29-32, 1990. BREEDING SUCCESS OF SUBLEGAL SIZE MALE RED KING CRAB PARALITHODES CAMTSCHATICA (TILESIUS, 1815) (DECAPODA, LITHODIDAE) J. M. PAUL AND A. J. PAUL Institute of Marine Science University of Alaska Seward Marine Center P.O. Box 730 Seward, Alaska 99664 ABSTRACT The reproductive potential of red king crab (Paralithodes camtschatica) (Tilesius) males 80 to 139 mm carapace length (CL) was examined by placing individual males with four females and noting breeding behavior, ovulation and percentage of eggs showing cleavage in clutches. A mating was considered successful if a male induced a female to ovulate and eggs initiated division. Males 80—89 mm CL were successful in inducing ovulation with 75, 38, 12 and 12% of their Ist, 2nd, 3rd and 4th potential mates respectively. An average of 68% of the eggs initiated division in clutches of their first mate. Corresponding values for their 2nd, 3rd and 4th consecutive mates were 18, 12 and 12% respectively. As male size increased so did the ability to mate with successive mates. Males in the 130-139 mm group induced an average of 88, 78, 100 and 44% of their four successive potential mates to ovulate. Clutches of the first through fourth females bred by 130-139 mm males had 87, 76, 95, and 38% of the eggs initiate division on the average. KEY WORDS: King crab, reproduction, maturation, ovulation INTRODUCTION Male red king crab, Paralithodes camtschatica, pre- viously supported an important commercial fishery in Alaska. For example the Alaskan harvest in 1980 was ~188 million pounds and by 1985 it had dropped to 16.5 million pounds. Currently several fishing areas are closed to harvest because of low crab abundance. The reasons for the large scale population decreases are unknown, but their occurrence has underscored the importance of under- standing the reproductive biology of the species. The cur- rent minimum size limit for males is around 145 mm cara- pace length. With the fishery removing the largest males it is important to know the reproductive capacity of sublegal size males. The objective of this study was to examine the ability of small red king crab to mate repeatedly and to quantify the percentage of eggs initiating division in clutches of their mates. MATERIALS AND METHODS All female specimens used in these experiments were collected in lower Cook Inlet or near Kodiak Island. Males over 100 mm CL also came from these areas, while smaller males were collected near Juneau, Alaska. In all observations one male was placed with four multi- parous females, all having eggs in the process of hatching. Following a females molt, breeding activity was moni- tored. A mating was considered successful if a male in- duced a female to ovulate and eggs initiated division. If ovulation was not observed within six days of their molt, females were placed with another male. All females placed with other males mated and extruded viable clutches. After ovulation females were isolated and held for one week or until eggs developed to the four to 64 cell stage. 29 Then a group of at least 100 eggs from each pleopod was randomly selected and examined under a microscope for cell division. Values from pleopod subsamples were aver- aged to estimate the percentage of dividing eggs in clutches. Previous work demonstrated that for 10 days after molting males are incapable of mating (Powell et al. 1972). Males used in breeding experiments were either hard shell captives or new shell males held for a minimum of two weeks after molting. No apparent differences in male re- productive ability of new or hard shell specimens have been observed (Powell et al. 1974). Because the timing of fe- male molt was not controllable intervals between matings varied from 1 to 71 days. The number of days between individual matings and their success is recorded in a data report (Paul and Paul in press). The interval between mating for males did not markedly affect reproductive suc- cess. The tank size used in experiments was 800 | and the water temperature was between 4—6°C. Salinity ranged from 31—32 ppt. For data presentation males are grouped into 10 mm size groups based on carapace length (CL). This measurement is taken from the right eye notch to the central portion of the rear margin of the carapace. Each group contained at least five males. The size range of males used included 80 to 139 mm CL. Female sizes ranged from 102 to 154 mm CL. RESULTS The results of the breeding experiments are summarized in Table 1. Both the percentage of females induced to ovu- late (75%) and the percentage of dividing eggs (68%) in clutches of their first mates suggest that not all 80-89 mm 30 PAUL AND PAUL males are mature. Most of these males failed to induce more than one mate to ovulate. Only one male in this size range was able to fertilize all four females. One male killed three of his potential mates after they molted even though all had ripe ovaries. He molted and was cannibalized so it was not possible to check for sperm presence. Eighty-eight percent of 90-99 mm males induced their first mate to ovulate, but only 66, 55 and 0% were suc- cessful at breeding females 2 through 4 respectively. The percentage of dividing eggs in clutches of females, bred by 90-99 mm males, decreased with each successive mating. Their first mates’ clutches contained an average of 86% dividing eggs. Females that were 2nd, 3rd, and 4th in mating chronology had averages of 63, 41 and 0% of their eggs dividing. Two males in this size class killed some newly molted ripe females. In both cases males fertilized their first mates but did not induce their second mate to ovulate. One male killed females 3 and 4; the other killed female 4. Males in the 100—109 mm group induced an average of 86% of their first potential mates to ovulate. Less than 42% of these males induced a second female to ovulate and only 14% bred a fourth female. An average of 83% of the eggs of first mated females initiated division while the 2nd, 3rd and 4th females in mating chronology had only 39, 26, and 13% of their eggs cleaving respectively. All males in the 110—119 mm size group induced one female to ovulate and 90, 72, and 54% of them bred a 2nd, 3rd, and 4th mate respectively. The percentage of eggs initiating division in clutches of female mates | through 4 was 98, 74, 62 and 44% respectively. One male killed the second female after she molted but successfully bred females 1, 3 and 4. Males in the 120-129 mm group mated with all four females available to them in all but one case. The per- centage of cleaving eggs in clutches of females that ovu- lated was typically high, 87—100%. All but two of the 130—139 mm males bred three fe- males successfully but 5 of 9 males did not induce a fourth available female to ovulate. Egg division rates averaged 87 and 76% for females Ist and 2nd in chronology, and 95 and 38% for the 3rd and 4th female mated. One male killed two of his potential mates after they molted, but females | and 3 ovulated and over 92% of their eggs were developing. The percentage of females induced to ovulate by the dif- ferent size classes of males is summarized in Figure 1. Qualitatively all clutches were of normal size and egg count. Even clutches with low percentages of dividing eggs did not exhibit gross abnormalities in regards to clutch size. Evidently even non fertilized eggs may attach normally. Egg counts for individual clutches in these experiments are available in a data report (Paul and Paul, in press). DISCUSSION Some reports suggest that both male and female red king crab have attained maturity around 100 mm CL (Powell TABLE 1. Percentage of eggs dividing in clutches of Paralithodes camtschatica mated successively by a single male. A 0 indicates male did not induce female to ovulate, * means male killed that female. (Data listed as increasing male size) Male % Eggs % Eggs % Eggs % Eggs Carapace Dividing Dividing Dividing Dividing length (mm) Mate 1 Mate 2 Mate 3 Mate 4 80-89 59 0 0 0 0 0 0 0 94 15 0 0 92 40 0 0 99 0 0 0 0) * * * 100 96 98 99 99 0 0 0 Mean 68 18 12 12 90-99 99 99 95 0 98 0 “2 - 99 98 0 0 0 0 0 0 99 86 59 0 86 0 0 s 99 88 80 0 99 97 41 0 97 99 98 0 Mean 86 63 41 0 100-109 89 0 (0) 0 98 99 99 0 99 0 0 0 100 0 0 99 96 81 0 0 0 0 0 99 80 0 92 Mean 83 39 26 13 110-119 100 92 99 89 99 83 86 0 99 100 99 99 100 99 100 99 98 99 0 0 90 99 40 0 99 = 96 20 100 1 67 75 99 97 0 99 99 62 100 0 92 88 0 0 Mean 98 74 62 43 120-129 99 99 92 97 0 99 98 87 98 98 98 98 100 99 97 99 99 99 99 98 98 99 99 100 Mean 82 98 97 96 130-139 99 99 99 96 99 98 99 87 99 99 96 0 99 0 66 62 0 98 100 0 92 * 98 = 99 92 99 95 100 99 100 0 99 98 99 0 Mean 87 76 95 38 KING CRAB 31 100 [oe Sgn tS Paes oo eraameeai ec an LU ee on \ L066 ™ —| 80 ING “206, ne =) SS SOG. = > My oy oe (@) S. S205 ™ Ss © Ne do ovale eecores eae SSN 00 A anal creel sa = (a) MALE BN oS LU LENGTH Se : O (mm) Al . =) 40 dees ° Q —— 80-89 =< : Fa eeccce 90-99 a . a —---— 100-109 ee oe XS ——— 110-119 iy See 2.0 rae aaa 120-129 inssee —— 130-139 oats 0 ‘ 6) 1 2 3 4 SUCCESSIVE FEMALE MATES Figure 1. The percentage of successive females induced to ovulate by male Paralithodes camtschatica of a given carapace length (mm). Each male had access to four potential mates. and Nickerson 1965, Gray and Powell 1966, Somerton 1980) while others suggest that smaller males are mature (Powell et al. 1972). Powell et al. (1972) reported that 50% of six 84-89 mm males mated a single female successfully but egg viability was not estimated. In our observations with 80—89 mm males, their first mate extruded normal appearing clutches but only an average of 68% of these eggs initiated division. The majority of males in this size class are not capable of fertilizing a second mate, only an average of 18% of eggs in clutches of their second mates were viable (Table 1). In nature males in grasping pairs are typically larger than 120 mm (Powell and Nickerson 1965, Powell et al. 1972, 1974). Perhaps larger males exclude smaller ones from breeding (Powell et al. 1972). Several studies report multiple matings with king crabs. In an early report 11 new shell males, 120-144 mm, bred 51 females held with them in a boats live well for ten days, and they all extruded full clutches (Powell and Nickerson 1965). Males near legal size (140 mm CL) have been re- ported to mate as many as 13 successive times, but their mating ability decreased after the sixth or seventh mating (Powell et al. 1972, 1974). None of those reports quanti- fied egg viability and it is possible that the reproductive capacity of those red king crab males was overestimated since females can have clutches that appear normal to the unaided eye but contain only unfertilized eggs (Table 1). Our contrasting results indicate that mating experiments with legal size males should be redone and egg viability monitored to verify the existing observations. Results of this study indicated that smaller sublegal male king crab can not be counted on to breed more than one or two females without reduced reproductive output. But, sev- eral authors (Gray and Powell 1966, Powell et al. 1972, 1974) have noted that male size is not the only factor modi- fying breeding success. Geographic sex segregation, male molting during the mating season, inability of males to mate for approximately 10 days after molting, and naturally occurring and fishery caused differences in sex ratios have been identified as variables that might affect the reproduc- tive success of king crab. While laboratory studies will continue to provide insight into the reproductive process for king crab, future studies should emphasize intensive in situ observations of reproductive success. Perhaps some of Ww i) Alaska’s isolated bays could be used as experimental pre- serves where sex and size ratios could be modified for ob- servation of reproductive success. ACKNOWLEDGMENTS This work is a result of research sponsored by the Alaska Sea Grant College Program, cooperatively sup- ported by NOAA, Office of Sea Grant and Extra Mural programs Department of Commerce, under grant number KING CRAB NA86AA-D-SG041, project R/06-27; by the University of Alaska with funds appropriated by the state; and by the Alaska Department of Fish and Game. Facilities were pro- vided by the Seward Marine Center, University of Alaska. Specimens for study were provided by Kodiak and Homer offices of Alaska Department of Fish and Game and Na- tional Marine Fisheries Service Auke Bay Laboratory staff. This is contribution 807 from the Institute of Marine Science, University of Alaska. LITERATURE CITED Gray, G. W. & G.C. Powell. 1966. Sex ratios and distribution of spawning king crabs in Alitak Bay, Kodiak Island, Alaska (Decapoda Anomura, Lithodidae). Crustaceana 10:303—309. Paul, J. M. & A. J. Paul. Reproductive success of sublegal size male red king crab with access to multiple mates. In Melteff, B. (ed). Interna- tional Symposium on King and Tanner Crabs. Lowell Wakefield Fish- eries Symposia Series, University of Alaska, Fairbanks, Sea Grant Report. In press Powell, G. C. & R. B. Nickerson. 1965. Reproduction of king crabs Pa- ralitnodes camtschatica (Tilesius). J. Fish. Res. Bd. Canada 22:101- ua Powell, G. C., B. Shafford & M. Jones. 1972. Reproductive biology of young adult king crabs Paralithodes camtschatica (Tilesius) at Kodiak Island, Alaska. Proc. Natl. Shellfish Assoc. 63:77-87. Powell, G. C., K. E. James & C. L. Hurd. 1974. Ability of male king crab, Paralithodes camtschatica, to mate repeatedly, Kodiak, Alaska, 1973. Fish. Bull. 72:171-179. Somerton, D. A. 1980. A computer technique for estimating the size of sexual maturity in crabs. Can. J. Fish. Aquat. Sci. 37:1488-—1494. Journal of Shellfish Research, Vol. 9, No. 1, 33-40, 1990. PUTATIVE BACTERIAL AND VIRAL INFECTIONS IN BLUE CRABS, CALLINECTES SAPIDUS RATHBUN, 1896 HELD IN A FLOW-THROUGH OR A RECIRCULATION SYSTEM GRETCHEN A. MESSICK! AND VICTOR S. KENNEDY? 'Department of Commerce, NOAA, NMFS Northeast Fisheries Center Oxford Laboratory, Oxford, Maryland 21654 ?Horn Point Environmental Laboratories University of Maryland Cambridge, Maryland 21613 ABSTRACT The blue crab (Callinectes sapidus) industry of Chesapeake Bay uses flow-through and recirculation shedding systems to produce soft-shell crabs. Three replicate experiments were performed in summer of 1986 (June, July, August) to compare mortality and crab disease at four different holding densities in the two systems. Replicate sets of crabs (75-80% intermolt) were held for 22 days in each system, with dead crabs and those appearing moribund removed daily. Selected tissues were removed from moribund crabs and those surviving to the end of the experiments, and were examined histologically. Early mortalities were attributed to bacterial infections, which were first noted between days 2—3 (recirculation) or days 2—20 (flow-through). In five of six instances, infections that we attributed to viruses were first detected from 2 to 18 days later than bacterial infections. As summer progressed, mortalities decreased and fewer crabs had bacterial or viral infections. No statistically significant differences were detected in mortali- ties or in occurrence of bacterial or viral infections between recirculation or flow-through systems, or among the four different crab densities within each system. KEY WORDS: INTRODUCTION Two types of shedding systems are used in Chesapeake Bay to produce soft-shelled blue crabs, Callinectes sa- pidus, for market. In the traditional flow-through system, estuarine water is constantly pumped through the shedding tank, and back to the source. A newer, more innovative recirculation system (Manthe et al. 1983; Malone and Manthe 1985) can be operated far from an estuary to allow water quality to be better controlled. Artificial or natural estuarine water is recirculated through this system; reoxy- genation and a biological filter are required to maintain high water quality. Mortalities of crabs held in shedding systems vary from 5—40% on the east coast, and water chemistry, flow rates, and stocking densities are critical factors requiring control (Manthe et al. 1983; Hochheimer 1986). Normal stress from molting and concurrent changes in physiology may also cause mortalities (Johnson 1976a). Bacterial and viral infections can be a major cause of mortality in captive crabs (Johnson and Bodammer 1975; Johnson 1976a, 1977, 1978, 1984a). Stress from holding crabs in confined or crowded conditions can affect the prevalence and intensity of disease (Couch 1974; Lightner et al. 1983; Johnson 1984a). Our study compared mortality and disease in crabs held in these two types of shedding systems at four different holding densities. Pathology was based solely on histologic and gross observations. Bacteria were not isolated nor was electron microscopy used to identify viruses. Although we used a small number of pre- 33 blue crabs, shedding systems, bacteria, viruses, infections, pathology. molt and postmolt crabs, the study was performed princi- pally with intermolt crabs. This was because intermolt crabs dominated the fishery. One advantage of using inter- molt crabs was that they could be held in captivity for a prolonged experimental period. Thus our study was a pre- liminary evaluation of disease in the different holding systems, rather than an evaluation of the suitability of one over the other for the shedding industry. MATERIALS AND METHODS In summer 1986 we performed three experiments (one each in June, July, and August) to examine pathological responses in tissues of crabs held captive at four holding densities in the two types of shedding systems. Small male and juvenile female crabs (6.0 to 13.5 cm) were caught on baited trot lines deployed in the Miles River (Experiment 1) and Tred Avon River (Experiments 2 & 3). These rivers are adjacent tributaries of Chesapeake Bay, separated by an isthmus and located in a predominantly agricultural region. For each of the three experiments, 85 apparently healthy crabs with no external physical damage were placed in a flow-through system and 85 more in a recirculation system, both at Horn Point Environmental Laboratories. Both tanks were subdivided into 0.75 m* quarters by plastic-coated wire mesh that allowed water but not crabs to pass among subsections. To test the incidence of disease at difference crab densities, the number of crabs per subsection was varied from 35 to 25 to 15 to 10; crabs were randomly as- signed to each subsection. 34 MESSICK AND KENNEDY Tanks had approximately equal flow rates of about 40 L/min, surface area of about 3 m*, and water volume of 300 L (J. Hochheimer, Horn Point Environmental Labora- tories, personal communication 1988). Water entered both systems forcefully from a single delivery point and, when necessary, was added to the recirculation system. Tempera- ture and salinity were monitored daily in each system with a thermometer and hand-held refractometer. Ammonia and nitrite levels in the recirculation system were measured at 2-3 d intervals throughout the summer, but pH and dis- solved oxygen were monitored only in August, the warmest month. Each experiment lasted 22 d, a period we consid- ered adequate to observe effects of stress. Between 0700 and 0900 h daily, crabs were fed equal amounts of thawed fish, and dead crabs were removed and not replaced. Those with any odor were not examined his- tologically because of rapid bacterial degradation of tissues which would mask the effects of other possible pathogens. Moribund and slow-to-react crabs were removed and trans- ported chilled to the National Marine Fisheries Service Laboratory at Oxford, Maryland, as were all surviving crabs at the end of each experiment. Organs or tissue fixed and processed for histologic observation included epi- dermis, gill, gut, antennal gland, hepatopancreas, brain, thoracic ganglion, heart, and hemopoietic tissue. Tissues were fixed in Helly’s fixative for about 18 h. Embedded tissues were cut at 5 ym and stained with three different stains; 1: alcian blue counterstained with nuclear fast red, 2: Periodic Acid Schiff’s counterstained with Weigert’s iron chloride hematoxylin, and 3: Feulgen reaction coun- terstained with picromethy! blue (Howard and Smith 1983; Johnson 1980). Prepared slides were examined microscopically and data recorded on tissue conditions, molt stage, and pathological and abnormal histologic characteristics (Johnson 1983a). Although we did not attempt to culture or isolate bacterial species from infected tissue, we believe it is reasonable to attribute the cause of disease to bacteria, based on histo- logic characterization of tissue responses to bacteria as re- ported by Johnson (1976a) who studied the progression of bacterial infections. We looked for evidence of this pro- gression, which begins initially in heart tissue with hemo- cyte groups that contain a small center of pyknotic or karyorrhectic nuclei surrounded by normal hemocytes. These groups may be enclosed by flattened granular or hyaline hemocytes. The centers of the small cell aggrega- tions become hyaline with pyknotic nuclei or chromatin fragments; these groups (nodules) may be encapsulated by hyaline hemocytes (Figure 1A). As the infection pro- gresses, nodules and cell aggregations appear in blood sinuses and arteries throughout the crab, the antennal gland may have general or multifocal degeneration and aggrega- tions of hemocytes, and hemopoietic tissue may have pyk- notic nuclei. Other tissue and organs such as brain and tho- racic ganglion are visibly affected when hemocyte aggrega- tions are present in surrounding hemal spaces. In heavy infections, bacteria may be seen in nodules and fixed phagocytes of the hepatopancreas. We performed no biochemical or electron microscope analysis of tissues taken from crabs in this experiment, so no definitive diagnosis of virus species was accomplished. The criteria for tentative diagnosis of viral infections were based solely on histologic light microscope analysis of tissues (e.g., see Figure 1B for normal hemopoietic tissue) with characteristics consistent with those reported by Johnson (1977, 1978, 1983b, 1984a, 1985). Lethargy, pa- ralysis, or trembling of legs are gross symptoms of sus- pected viral infections in blue crabs. Histologically, virus may be present when necrosis of hemocytes, hemopoietic, or nervous tissue occurs and when tissues contain cyto- plasmic inclusions that produce Feulgen negative, non-al- cianophilic, and Periodic Acid-Schiff negative reactions. These inclusions are often angulate but may be round or amorphous (Johnson 1985; see Figure 1C). Hemopoietic tissue (Figure 1D) may have increased cytoplasmic volume (appear hypertrophied), and karyolysis, karyorrhexis, and pyknosis may be present; young hemocytes may also have increased cytoplasmic volume, viral inclusions, and nuclei of various sizes (Figure 1E). We used some or all of these characteristics in tentatively assigning the cause of disease to viruses. We also noted hemocyte reactions (groups of hemocytes or tissue with basophilic vacuoles in the cytoplasm) be- cause they may represent an alteration in tissue morphology due to known or unknown pathogens. Statistical Analysis To examine effects of (A) crabs being held in different systems and (B) density, a separate split-plot 2 « 4 factorial Analysis of Variance was peformed on the arc sin transfor- mations of each of the following variables (Steel and Torrie 1960): 1) Percent mortality, 2) Percent bacterial infection, 3) Percent viral infection. Percent bacterial and viral infections were calculated for each of the three experiments from the combined number of crabs that either survived 22 d or became moribund during that period. As noted earlier, dead crabs were not included in these calculations unless they were judged to have only recently died and had no odor; thus our estimates of disease and mortality are conservative. Percent variation due to month of experiment, system, and crab density was calculated by dividing the sum of squares for these factors by the total sum of squares for all sources of variation derived by the ANOVA (Steel and Torrie 1960). This was calculated with ANOVA results for mortality, bacterial infections, and viral infections. Subse- quently, the average variation for these three variables due to month of experiment, system, and crab density was de- termined. Ge, BLE OF ‘ft Oe q h . ; ~ Pes var ‘ nis : ne wie Rey £7 SO Gy 0): pal C t Figure 1A. Nodule (n) in heart tissue with groups of aggregated hemocytes (h). Note hyaline center of nodule with pyknotic nuclei and thin layer of flattened hyaline hemocytes (arrow) surrounding nodule. Line = 12 pm. Figure 1B. Normal hemopoietic tissue in intermolt crab. (n) = nuclei. Line = 7 pm. Figure 1C. Intermolt hemopoietic tissue packed with putative virus infected cells. Note angulate and amorphous cytoplasmic inclusions (arrows) and pyknotic nuclei (n). Line = 7 pm. Figure 1D. Putative viral infection in hemopoietic tissue of intermolt crab. Note increased cytoplasmic volume (v), inclusions (arrows), and normal cells (c). Line = 7 pm. Figure 1E. Putative viral infection in young hemocytes released from hemopoietic tissue. Note increased cytoplasmic volume (v), possible viral inclusions (arrows), nuclei of various sizes (n), and normal young hemocytes (y). Line = 7 pm. 36 MESSICK AND KENNEDY RESULTS Temperatures and salinities in both systems ranged from 20° to 28°C and 10 to 14 ppt, with means increasing about 1°C or | ppt per month. In August, dissolved oxygen values in the recirculation system ranged from 6.6 ppm to 7.6 ppm, and pH ranged from 6.81 to 7.79. Dissolved ox- ygen values in the flow-through system ranged from 4.7 ppm to 7.2 ppm. From mid-June to mid-August in the re- circulation system, ammonia remained below 0.02 ppm, except one day when it was 0.20 ppm (causing no apparent increase in mortality or disease); water was changed imme- diately and ammonia levels dropped to below 0.02 ppm. Nitrite ranged from 0.1 ppm to 3.2 ppm during the sum- mer. Pathological responses were associated predominantly with putative bacterial and viral infections (Tables 1, 2), with other histologic responses evoked by microsporidia, gregarines, fungi, and rickettsia-like organisms. Diseases occurred either singly or in conjunction with any one or more of the other diseases. Analysis of variance revealed that there was no signifi- cant difference (P > 0.05) in mortality (F = 10.9, df = 1, 2) nor in bacterial (F = 0.18, df = 1, 2) or viral (F = 0.82, df = 1, 2) infection observed between different systems (Factor A) or among different crab densities (Factor B; Mortality F = 0.76, df = 3, 12; Bacteria F = 2.71, df = 3, 12; Virus F = 1.5, df = 3, 12). There was also no interaction effect in infections between systems and crab densities (P > 0.05) for either mortality (F = 0.50, df = 3, 12), bacteria (F = 0.37, df = 3, 12), or virus (F = 2.4, df = 3, 12) between systems (Table 1). Experimental variation due to month of experiment (24%) was greater than the variation due either to type of system (11.1%) or densities (15.6%). Because the incidences of disease from different holding densities were not statistically different, we combined the data on surviving crabs (excludes moribund crabs) at these densities for each of the three experiments (Table 2). In the flow-through system, 51 crabs that had survived in the three experiments for 22 d were dissected; an average of 20% had bacterial infections, 6% had viral infections, and 47% had hemocyte reactions. Small percentages had rick- ettsia or microsporidia. In the recirculation system, 23 sur- viving crabs were dissected; an average of 4% had bacterial or viral infections, and 48% showed hemocyte reactions. About | in 5 (range = 20-23%) of the surviving crabs held in the flow-through system showed no evidence of disease compared to about | in 3 (range = 10—67%) of the surviving crabs held in the recirculation system. To determine if individual stages of the molt cycle or one sex or the other was more prone to pathological re- sponses, data were combined for surviving and moribund crabs from the recirculation and flow-through systems (Table 3). An average of 22% of the premolt crabs and TABLE 1. Percent of crabs which died (mortality) and percent of moribund and surviving crabs (plus a few non-odiferous dead crabs) with putative bacterial and viral infections at the end of three experiments. In each experiment, four different crab densities were held in a recirculation (R) or a flow-through (F-T) shedding system. For each crab density, values of dead crabs calculated from percent mortality plus the number examined may not sum to that density because, at the end of an experiment, some crabs were healthy or had other infections not listed; some crabs had both bacterial and viral infections; in experiment three, not all surviving crabs were dissected; and a few non-odiferous dead crabs were included in the number examined for presence of bacteria or viruses. Percent Number Percent Percent Mortality Examined*® with Bacteria with Virus Crab Experiment Density R F-T R F-T R F-T R F-T 10 100 70 0 3 0 0 0 100 1 15 93 73 1 0 0 0 75 (June) 25 88 56 3 11 67 9 67 45 35 83 83 afl 6 43 33 43 67 Total 88 72 11 2 45 12 45 63 10 80 70 2 3 0 0 0 33 2 15 67 67 5 5 40 0 20 20 (July) 25 88 68 7 8 14 0 43 38 35 14 14 2 9 8 u 58 2 Total 78 71 26 25 15 4 42 28 10 50 40 2 3 0 33 50 33 3 15 47 IH} 4 6 25 33 0 0 (August) 25 76 40 5 7 20 14 20 14 35 86 43 zl 10 1 40 l4 10 Total 72 39 18 2 39 31 17 12 Average 79 60 29 16 35 33 * These values may include moribund crabs, surviving crabs, and a few non-odiferous dead crabs. BACTERIA AND VIRUSES IN BLUE CRABS 37 TABLE 2. Numbers of crabs surviving (does not include moribund crabs) at the end of three experiments that, upon being dissected, were either apparently disease free, or that showed evidence of infection. Values in parentheses are percentages. R = recirculation system; F-T = flow-through. Total Crabs No Hemocyte Survivors Dissected Disease Bacteria Virus* Reaction Experiment R F-T R F-T R F-T R F-T R F-T R F-T 1 3 9 3 9 2 (67) 2 (22) 0 1 (11) 0 2 (22) 0 3 (33) 2 10 17 10 ig} 5 (50) 4 (24) 0 1 (6) 1 (10) 0 4 (40) 10 (59) 3 20 51 10 25¢ 1 (10) 5 (20) 1 (10) 8 (32) 0 1 (4) 7 (70) 11 (44) Average? (35) (22) (4) (20) (4) (6) (48) (47) In experiment 1, there was one crab in each system for which there was inconclusive evidence for presence of virus. > One crab contained rickettsia and a second contained microsporidians. © Microsporidians were found in one crab with virus and in one crab with hemocyte reaction. 4 Calculated by dividing number of survivors in a disease category in each system by the total number of crabs dissected for that system. 15% of intermolt crabs showed no evidence of disease. No postmolt crabs were disease free. Bacterial or viral infec- tions were found in 17 to 36% of blue crabs, with hemocyte reactions occurring in 28 to 50%. Microsporidian infections were uncommon in all instances. About the same percent (15—16%) of both sexes were apparently healthy, and there was little difference between sexes in proportion of individuals with bacteria, or viruses, or microsporidians (Table 3). Females exhibited a greater incidence of hemocyte reactions than did males. Putative viruses encountered in this experiment were strongly suggestive of virus RLV (reolike virus) and RhVA (rhabdolike virus) in that they demonstrated gross behav- ioral and histological characteristics closely similar to those described by Johnson (1977, 1978, 1983b, 1984a, 1985). Infected crabs displayed a delayed reaction to stimulation and often appeared partially paralyzed or had tremors of appendages. Hemolymph removed from moribund animals clotted incompletely or not at all, but hemocyte aggrega- tions were still present. Hemopoietic tissue and hemocytes of these crabs were usually necrotic hypertrophied cells with increased cytoplasmic volume and karyorrhectic and pyknotic nuclei. Cells contained opaque, basophilic, amor- phous cytoplasmic inclusions that were Feulgen-negative; this is consistent with descriptions of RLV (Johnson 1977). Similar viral inclusions were also observed in glial cells of the ganglia, nerves, hemocytes, connective tissue, and various epithelia. In experiments 1 and 3, bacterial infections were noted first (flow-through, days 2 or 4; recirculation, days 2 or 3), followed by viral infections (flow-through, days 4 or 22; recirculation, days 5 or 11) (Figure 2). In experiment 2, bacteria appeared first in the recirculation system (day 3) followed by viruses on day 8. In the flow-through system, viruses appeared on day 9 compared with day 20 for bac- teria. In half of the treatments, the cumulative numbers of viral infections exceeded the cumulative number of bacte- rial infections; the reverse occurred twice, and bacterial and viral infections were equally abundant once. For the three experiments, the percentage of survivors gradually in- creased each month, with the flow-through system always yielding higher numbers of survivors (Table 2; Figure 2). Aside from bacteria and viruses, gregarines (Sprague 1970) were the most common pathogen; they were found in a larger percentage of moribund (29-62%) than surviving (11-18%) crabs. Microsporidian infections (Overstreet TABLE 3. Numbers and average percentages (in parentheses) of surviving and moribund crabs displaying evidence (or its lack) of disease in relation to molt condition and sex. No Hemocyte Category Numbers* Disease Bacteria Virus Reactions Microsporidians Premolt 23 (18) 5 (22) 4 (17) 5 (22) 8 (35) 1 (4) Intermolt 101 (77) 15 (15) 22 (22) 36 (36) 28 (28) 6 (6) Postmolt 6 (5) (0) 2 (33) 1 (17) 3 (50) (0) Males 96 (74) 15 (16) 22 (23) 32 (33) 25 (26) 6 (6) Females 34 (26) 5 (15) 6 (18) 9 (26) 14 (41) 1 (3) @ Numbers in this column may not agree with the sum across each row because some crabs had more than one disease. Values in parentheses in this column represent the percentage of each molt type or each sex. MESSICK AND KENNEDY RECIRCULATION 0 FLOW-THROUGH JUNE JUNE SURVIVORS + BACTERIA * VIRUS NUMBERS OF SURYIVORS 100 AUGUST 100 AUGUST DAYS FROM START OF EXPERIMENT Figure 2. Declining number of survivors of blue crabs in three recirculating or flow-through experiments in 1986 and cumulative numbers of moribund individuals with bacterial or viral infections. -O- = survivors; -@ = bacteria; -@ = virus. BACTERIA AND VIRUSES IN BLUE CRABS 39 1978) occurred infrequently (July, August) in both recircu- lation and flow-through systems, and in both moribund crabs and healthy survivors. Fungal infections (Johnson 1983b) occurred only in the recirculation system (July) as did rickettsia-like infections (Johnson 1984b) (only 3 of 130 crabs dissected). Trematode metacercariae (Sprague 1970) were seen in one moribund and one surviving crab in the recirculation system in August. An unidentified type of strand-like bacterial infection (Johnson 1976b) was de- tected (31% recirculation; 12% flow-through), but occurred only in the lumen of the hepatopancreas of crabs observed histologically. Of those crabs with this bacterium-like or- ganism, 4% had no other pathogen in conjunction, 12% also had hemocoelic bacterial infections, 65% had viral in- fections, 11% had both circulating bacterial and viral infec- tions, and 8% had microsporidians and either fungus or rickettsia-like organisms (Roe 1988). DISCUSSION Crabs that became moribund and were dissected during the period of high mortalities (except flow-through, July) were diagnosed with putative bacterial infections. We pro- pose that crabs we were unable to dissect during high initial mortalities (Figure 2) because they had begun to decom- pose, died due to similar bacterial infections caused by stress from capture and captivity (Johnson 1976a). Thus crab shedders should emphasize to their premolt crab sup- pliers the importance of careful handling to help reduce the possible stress that may cause high initial mortalities in shedding systems. Tubiash et al. (1975), Colwell et al. (1975), and Welsh and Sizemore (1985) found that captured crabs had low- level bacterial infections, even when they were collected by presumably nonstressful means. However, Johnson (1976a) has argued that ‘‘naturally’’ infected crabs acquired bacte- rial infections during the stress of capture and transport, not before. She suggested that crabs sampled by Tubiash et al. (1975) were stressed and traumatized during collection. Bang (1970) stated that the blood of normal invertebrates is sterile but animals with transient infections may be misin- terpreted as being “‘normal’’ carriers of bacteria. This sub- ject requires additional research on the interaction between defense mechanisms in blue crabs and ubiquitous bacteria found in seawater. Some crabs in this study may have survived initial bac- terial infections, although histologically they demonstrated hemocyte reactions. Tissues of obviously healthy crabs often exhibited groups of hemocytes with basophilic vac- uoles in the cytoplasm. These hemocytes may have been altered by their action in attacking and removing pathogens from hemolymph and tissues. Although microbial agents may not be visible microscopically under such conditions, pathological changes due to their presence can be observed in histologic preparations of tissues (Johnson 1983b). We found no obvious signs of lysed or degenerating bacteria or other pathogens in tissues with hemocyte reactions. Most crabs with hemocyte reactions were not moribund but behaved normally. Perhaps the foreign organisms re- sponsible for the reactions had invaded earlier and were then removed from the tissues, leaving only the altered he- mocytes as a sign of their invasion. There is evidence that certain crustaceans can recover from disease (Bang 1974). As noted, we believe the viruses in our experimental crabs were RLV and RhVA. The RhVA occurs in crabs in nature and in stressed crabs. Johnson (1978, 1983b) sug- gested that if RLV is present in crab tissue, RhVA will also be present and that these two viruses act synergistically to cause glial necroses, ultimately resulting in death. As with bacterial infections, viral infections may become patent due to stress from captivity and crowding. Some viral diseases may be latent in blue crabs with the virus being enzootic under natural conditions and not prone to becoming epi- zootic until the crab is exposed to stresses (Couch 1974; Yudin and Clark 1979; Johnson 1977, 1978, 1984a, 1985; Lightner et al. 1983). The gradual increase in blue crab survival from June through July and August may have been due to seasonal physiological changes. As summer progresses, crabs ma- ture, molt less, and will soon migrate to saltier or deeper water for the winter. There is less stress from frequent ec- dysis and crabs have increased body weight and are perhaps better equipped to fight infection and pathogens than in early June, when they may be recovering from winter con- ditions in the Chesapeake Bay. Finally, our experiments should be extended to consider primarily premolt crab pop- ulations, because this is the molt stage customarily held in various types of shedding systems. ACKNOWLEDGMENTS We thank Phyllis T. Johnson for her guidance during this study, and for her assistance in the characterization of bacterial and viral infections (nevertheless, we alone are responsible for any errors in these diagnoses). We also thank J. Bodammer, A. Farley, F. Kern, J. Mihursky, M. Newman, A. Rosenfield, C. Sindermann, and D. Wright for their comments and review of this manuscript. We are grateful to J. Hochheimer for the use of his shedding systems, B. Colburn and G. Schnaitman for allowing us to collect crabs while they worked, A. Rosenfield for getting us started, and the staff of Oxford Laboratory for support. Financial support was provided by the National Marine Fisheries Service, National Oceanic and Atmospheric Ad- ministration, U.S. Department of Commerce. This paper is based on an M.S. thesis submitted by G. A. Roe to the Marine Estuarine Environmental Sciences program at Uni- versity of Maryland College Park. Contribution number 2022 of the Center for Environmental and Estuarine Studies. 40 MESSICK AND KENNEDY LITERATURE CITED Bang, F. B. 1970. Disease mechanisms in crustacean and marine arthropods. Jn: S. F. Sniesko (ed), A Symposium on Diseases of Fishes and Shellfishes. Am. Fish. Soc. Spec. Pub. 5:383—404. Bang, F. B. 1974. Pathogenesis and autointerference in a virus disease of crabs. Infect. Immu. 9:1057—1061. Colwell R. R., T. C. Wicks & H. S. Tubiash. 1975. A comparative study of the bacterial flora of the hemolymph of Callinectes sapidus. Mar. Fish. Rev. 37(5—6):29-33. Couch, J. A. 1974. An enzootic nuclear polyhedrosis virus of pink shrimp: ultrastructure, prevalence, and enhancement. J. /nvertebr. Pathol. 24:311-331. Hochheimer, J. 1986. Water quality in crab shedding. The Shedder, Volume 2. University of Maryland Sea Grant Extension Program, Cambridge, Maryland. Howard, D. W. & C. S. Smith. 1983. Histologic Techniques for Marine Bivalve Mollusks. NOAA Tech. Memor. NMFS-F/NEC-25. 97 pp. Johnson, P. T. 1976a. Bacterial infection in the blue crab, Callinectes sapidus: course of infection and histopathology. J. Invertebr. Pathol. 28:25-36. Johnson, P. T. 1976b. An unusual microorganism from the blue crab, Callinectes sapidus. Proceedings of the First International Colloquium on Invertebrate Pathology, Kingston, Ontario, p. 316. Johnson, P. T. 1977. A viral disease of the blue crab, Callinectes sapidus: histopathology and differential diagnosis. J. Invertebr. Pathol. 29:201-209. Johnson, P. T. 1977. A viral disease of the blue crab, Callinectes sapidus: histopathology and differential diagnoses. J. Invertebr. Pathol. 29:201—209. Johnson, P. T. 1978. Viral diseases of the blue crab, Callinectes sapidus. Mar. Fish. Rev. 40(10):13-15. Johnson, P. T. 1980. Histology of the Blue Crab, Callinectes sapidus: A Model for the Decapoda. Praeger, New York, 440 pp. Johnson, P. T. 1983a. Diagnosis of crustacean diseases. Rapp. P.-V. Reun. Comm. Int. I’ Exploration Scientifique de la Mer Mediterranee Monaco 182:54—57. Johnson, P. T. 1983b. Diseases caused by virus, rickettsiae, bacteria and fungi. Jn: A. J. Provenzano Jr. (ed.), The Biology of Crustacea. Volume 6:1—78. Academic Press, New York. Johnson, P. T. 1984a. Viral diseases of marine invertebrates. Helgol. Meeresunters. 37:65-—98. Johnson, P. T. 1984b. A rickettsia of the blue king crab, Paralithodes platypus. J. Invertebr. Pathol. 44:112—113. Johnson, P. T. 1985. Blue crab (Callinectes sapidus Rathbun) viruses and the diseases they cause. Jn: H. M. Perry and R. F. Malone (eds.), National Symposium on the Soft-Shelled Blue Crab Fishery. Gulf Coast Research Laboratory, Biloxi, Mississippi, pp. 13-18. Johnson, P. T. & J. E. Bodammer. 1975. A disease of the blue crab, Callinectes sapidus, of possible viral etiology. J. Invertebr. Pathol. 26:141—143. Lightner, D. V., R. M. Redman & T. A. Bell. 1983. Observation on the geographic distribution, pathogenesis and morphology of the baculo- virus from Panaeus monodon Fabricius. Aquaculture 2:209—233. Malone, R. F. & D. P. Manthe. 1985. Chemical addition for accelerated nitrification of biological filters in closed blue crab shedding systems. In: H. M. Perry and R. F. Malone (eds.), National Symposium on the Soft-Shelled Blue Crab Fishery. Gulf Coast Research Laboratory, Bi- loxi, Mississippi, pp. 41—47. Manthe, D. P., R. F. Malone & H. M. Perry. 1983. Water quality fluctu- ations in response to variable loading in a commercial, closed shed- ding facility for blue crabs. J. Shellfish Res. 3:175—182. Overstreet, R. M. 1978. Marine maladies? Worms, germs and other sym- bionts from the northern Gulf of Mexico. Mississippi-Alabama Sea Grant Consortium, MASGP-78-021. Ocean Springs, Mississippi. Roe, G. A. 1988. A comparison of health factors of blue crabs, Calli- nectes sapidus, held in two shedding systems. M.S. thesis. University of Maryland, College Park, Maryland. 81 pp. Sprague, V. 1970. Some protozoan parasites and hyperparasites in marine decapod Crustacea. /n: S. F. Sniesko (ed.), A Symposium on Diseases of Fishes and Shellfishes. Am. Fish. Soc. Spec. Pub. 5:416—430. Steel, R. G. D. & J. H. Torme. 1960. Principles and Procedures of Sta- tistics. McGraw Hill, New York, 481 pp. Tubiash, H. S., R. K. Sizemore & R. R. Colwell. 1975. Bacterial flora of the hemolymph of the blue crab, Callinectes sapidus: most probable numbers. Appl. Microbiol. 29:388—392. Welsh, P. C. & R. K. Sizemore. 1985. Incidence of bacteremia in stressed and unstressed populations of the blue crab, Callinectes sa- pidus. Appl. Environ. Microbiol. 50:420—425. Yudin, A. I. & W. H. Clark Jr. 1979. A description of rhabdovirus-like particles in the mandibular gland of the blue crab, Callinectes sapidus. J. Invertebr. Pathol. 33:133-—147. Journal of Shellfish Research, Vol. 9, No. 1, 41—43, 1990. WINTER TEMPERATURE AND SPRING PHOTOPERIOD REQUIREMENTS FOR SPAWNING IN THE AMERICAN LOBSTER, HOMARUS AMERICANUS, H. MILNE EDWARDS, 1837 D. E. AIKEN AND S. L. WADDY Department of Fisheries and Oceans Invertebrate Fisheries Section Biological Station, St. Andrews, New Brunswick Canada E0OG 2X0 ABSTRACT Ovarian maturation and spawning in feral female American lobsters (Homarus americanus) from nearshore stocks (minimum winter seawater temperature 0°C) is normally regulated by the seasonal change in seawater temperature and is not affected by length or direction of change in photophase. However, when lobsters are held at elevated temperatures during the winter (=9°C after the winter solstice), photoperiod becomes influential. The minimum effective photophase is 12 h (12% of females spawn) and an increase from 8 h to either 14 or 16 h will cause 50-80% of preovigerous females to spawn. This suggests that the critical photophase for spawning is between 12 and 13 h. The critical temperature which induces photoperiodism following the winter solstice is between 6 and 8°C. Preovigerous lobsters held at a winter temperature of 0.001). DISCUSSION Previous studies have shown the importance of long day photoperiod (LD 16:8) in inducing spawning in lobsters that had overwintered at 13—14°C (Aiken & Waddy 1989, Nelson et al. 1983, Nelson 1986). Experiment A of this study shows that the threshold photoperiod for this re- sponse is close to LD 12:12—50% spawned at LD 14:10, only 12% spawned in at LD 12:12, and none spawned at LD 10:14 (Table 1). Photoperiod plays no role in spawning induction when lobsters are overwintered at temperatures close to 0°C (Aiken and Waddy 1985, 1986). Expeiment B shows that the threshold winter temperature for the switch to photo- period control of spawning is between 6 and 8°C (Table 2). On the basis of this study alone it would not be possible to state that photoperiod played no role in Group | of Ex- periment B (overwintered at 5°C). The difference of roughly 70 days in mean spawning time of Group | vs Groups 2 and 3 could be due to a low temperature enhance- TABLE 1. Effect of length of long-day photophase on spawning incidence of preovigerous females maintained at 13-14°C from the winter solstice. Initial daylength was 8 h (LD8:16) and change to long days occurred on 12 June. Length of Mean Spawning Photophase (h) Spawning Date Incidence (%) 10 _ 0 12 4 October 12 14 2 October 50 ENVIRONMENTAL REQUIREMENTS FOR LOBSTER SPAWNING 43 TABLE 2. Effect of water temperature after the winter solstice (21 December) on response to daylength. Photoperiod changed on 12 June from LD 8:16 to LD 16:8. Winter Mean Spawning Temperature (°C) Spawning Date Incidence (%) 5 10 July 96 9 15 September 96 13 19 September 82 ment of the photoperiodic response. However, we know from earlier studies with this same stock that lobsters over- wintered at S°C or less will spawn at approximately that same time, even on short (1 h) or declining daylengths (Aiken and Waddy 1985, 1986). Therefore the difference in spawning times in Experiment B is a reflection of the switch from temperature control (Group 1) to photoperiod control (Groups 2, 3). Furthermore, we know from studies on both coasts that photoperiodically-induced spawning in Gulf of St. Lawrence lobster occurs about 100 d after long- day onset (Aiken and Waddy 1989, Nelson 1986) as in Groups 2 and 3 of Experiment B (but not Group 1). In short, results from Experiment B are consistent with all previous studies: Groups 2 and 3 were photoperiodically- regulated, Group | was not. Thus in this study, we have shown that female lobsters held at winter seawater temperatures of 9°C or higher starting at the winter solstice require the stimulus of long days in order to spawn the following summer, but those held at winter seawater temperatures of 5°C or lower spawn in response to increasing spring temperatures and have no requirement for specific daylengths. Although local winter seawater temperature in most Canadian nearshore areas ap- proaches 0°C, 5°C allows spawning at the normal time (late June to early August in our laboratory). Interestingly, Nelson et al. (1988) reported that exposure to 6°C for varying lengths of time had no effect on the photoperiodic response. The critical temperature therefore lies between 6 and 8°C, slightly below the minimum winter temperature that occurs on much of the offshore lobster grounds. This supports Nelson’s (1986) suggestion that offshore lobsters utilize daylength to synchronize spawning. Spawning incidence was higher at 9°C (96%) than at 13—14°C (82%), but time of spawning was no different, suggesting that the lower temperature is more conducive to successful ovary maturation. Actually, the 82% spawning incidence in the 13—14°C group was higher than in any previous experiment at this temperature and is considered atypical. Comparable studies at this temperature normally produce a spawning incidence of only 50-60% (Aiken and Waddy 1985, 1986, Nelson et al. 1983, Nelson 1986). In summary, the threshold winter seawater temperature is between 6—8°C; below that spawning is controlled by temperature, above that it is controlled by photoperiod. The critical photoperiod for induction of ovary maturation and spawning is approximately LD 12:12. In shorter daylengths spawning is unlikely. In longer daylengths the incidence of spawning increases to 50—80%, significantly less than is obtained with optimum temperature conditions. REFERENCES CITED Aiken, D. E. 1969. Ovarian maturation and egg laying in the crayfish Orconectes virilis: influence of temperature and photoperiod. Can. J. Zool. 48:931—935. Aiken, D. E. 1973. Proecdysis, setal development, and molt prediction in the American lobster (Homarus americanus). J. Fish. Res. Board Can. 30:1337—1344. Aiken, D. E. & S. L. Waddy. 1982. Cement gland development, ovary maturation and reproductive cycles of female American lobsters, Ho- marus americanus. J. Crustacean Biol. 2:315—327. Aiken, D. E. & S. L. Waddy. 1985. The uncertain influence of spring photoperiod on spawning in the American lobster, Homarus ameri- canus. Can. J. Fish. Aquat. Sci. 42:194—197. Aiken, D. E. & S. L. Waddy. 1986. Oocyte maturation and spawning in wild American lobsters (Homarus americanus): lack of evidence for significant regulation by photoperiod. Can. J. Fish. Aquat. Sci. 43:1451—1453. Aiken, D. E. & S. L. Waddy. 1989. The interaction of temperature and photoperiod in the regulation of spawning by American lobsters (Ho- marus americanus). Can. J. Fish. Aquat. Sci. 46:145—148. Giese, A. C. & J. S. Pearse. 1974. Introduction: General principles. In Reproduction of marine invertebrates (A. C. Giese & J. S. Pearse, eds.). New York: Academic Press. Vol. 1, pp. 1—49. Lowe, M. E. 1961. The female reproductive cycle of the crayfish Cam- barellus shufeldti: the influence of environmental factors. Tulane Stud. Zool. 8:157—176. Nelson, K. 1986. Photoperiod and reproduction in lobsters (Homarus). Amer. Zool. 26:447—457. Nelson, K., D. Hedgecock & D. W. Borgeson. 1983. Photoperiodic and ecdysial control of vitellogenesis in lobsters (Homarus) (Decapoda, Nephropidae). Can. J. Fish. Aquat. Sci. 40:940—947. Nelson, K., D. Hedgecock & W. Borgeson. 1988. Factors influencing egg extrusion in the American lobster (Homarus americanus). Can. J. Fish. Aquat. Sci. 45:797—804. Perryman, E. K. 1969. Procambarus simulans: light-induced changes in neurosecretory cells and in ovarian cycle. Trans. Amer. Microsc. Soc. 88:514—-524. Rice, P. R. & K. B. Armitage. 1974. The influence of photoperiod on processes associated with molting and reproduction in the crayfish Or- conectes nais (Faxon). Comp. Biochem. Physiol. 47A:243—259. Stephens, G. J. 1952. Mechanisms regulating the reproductive cycle in the crayfish Cambarus I. The female cycle. Physiol. Zool. 25:70—83. Waddy, S. L. & D. E. Aiken. 1990. Manipulation of spawning in the American lobster: environmental requirements vary with season. Bull., Aquacul. Assoc. Canada (In Press). sD Cages: (NE? 0 ae OR awh 42.00 Gee ees a. a | 7 i tT ad a) 2 rae » » aij i=l x 400 A percentage in points for prey 1 x 100 percentage of occurrence for prey 1 = n aij being the number of points of prey i in stomach j, n the number of crabs examined having food, A the total number of points for all crabs and prey, and bi the number of crabs containing prey 7. Data Analysis Non-parametric statistical tests were used to analyze dif- ferences in main component consumption in the diet, be- tween the groups of individuals pertaining to different sta- tions, sizes, sex or molt stage. Seasonal variations were not studied due to possible interannual variability; tests were done separately for the different samples and grouped for all the stations when comparing diet by sizes (total column in Table 5). When comparing stations, samples were ana- lyzed by months and total stomachs (total column in Table 2). When there were two groups we used the Mann- Whitney test, and we obtained values of statistic t, distrib- uted as a ta,©; when there were more than two groups, the Kruskal-Wallis test was used, with statistic H, distributed as a X*a,d.f.. Diet diversity was obtained using the Shannon-Weaver index (1963), H’, for each station, month sampled, and size group. We carried out association analyses for different station/ date groups and for diet components, using percent data by points, for the twelve main components selected, which made up 80% of the diet. The percent similarity index, PSC (Sanders, 1960), was used as an association index and the average linkage between groups (UPGMA) as the fusion strategy. A factorial analysis of principal components (PCA) was carried out using the same data as in the associ- ation analyses. RESULTS We were able to identify 53 different components in the diet of L. arcuatus (Fig. 2; Table 1). The diet of this species is composed mainly of crustaceans, molluscs and ulvaceans, with polychaetes, fishes, Zostera, ophiuroids, and foraminiferans having secondary importance. Spatial Variations The results of stomach analyses show pronounced dif- ferences in the three habitats studied (Fig. 2; Table 1). In B1, a typical raft station located in the inner ria, the domi- MUSSEL RAFT CULTURE AND DIET OF L. ARCUATUS 47 i Pisinia OTHER DECAPODS AMPHIPODS 3 OTHER CRUSTACEANS [LJ myTILus OTHER BIVALVES IH =| POLYCHAETES ZZ SEAWEEDS ZOSTERA WY UNIDENTIFIED 7 fa OTHERS Figure 2. Relative importance of the main components (average % points determined for all crabs caught at each station) in the diet of L. arcuatus and percent similarity index (PSC) among the stations sampled. Others = fishes, gastropods, ophiuroids, etc. nant component is the mussel, Mytilus edulis (L.) (43.5% of the points), along with Pisidia longicornis (L.) (10.0%), a dominant decapod species in the raft epifauna (Gonzalez- Sanjurjo, 1982; Roman & Pérez, 1982), and plants, among which Zostera (15.8%) predominates over the seaweeds (10.2%). Station P3, a beach zone, is clearly differentiated from the two raft stations, as P. longicornis does not constitute part of the diet, and M. edulis appears only occasionally. In this area, crustaceans are dominant (40.0%), the decapods having 7.2% of the points, and predominating amphipods (7.9%) and other undetermined crustaceans (23.6%), gen- erally small in size. Molluscs make up only 9.4%, com- posed equally of bivalves, gastropods, opistobranchs. The chief plant component, making up nearly one-third of the diet of L. arcuatus in this area is seaweed (30.9%). Zostera also appears in small quantities. Station B6 represents an intermediate habitat between the beach and raft zones in the diet of L. arcuatus, showing a particularly high presence of polychaetes (15.5% of points). In this area the diet is composed of molluscs (25.1%), mainly bivalves. M. edulis comprises 12.3% of the points, seaweeds (23.4%), crustaceans (17.4%), among which P. longicornis makes up 5.5% and Zostera is less important with 4.9%. The latter is another indication of the intermediate characteristics this area displays. Although the percentage of the diet composed of plant elements is high in all three stations, Zostera is progressively substituted by seaweeds as we go from B1 to B6 to P3. Similarly, the consumption of M. edulis and P. longicornis decreases. Due to its intermediate position, B6 shows a high similarity percentage with BI and P3 (48% and 53% respectively), whereas the value obtained between the two latter stations is substantially lower (27%) (Fig. 2). The use of non-parametric statistical tests allows us to determine the main components, which clearly differentiate the three stations, as well as the degree of significance these differences have (Table 2). We notice the low number of individuals for the Bl sample in February (n = 19), which causes the practical absence of significant differ- ences when stations B1/B6 and B1/P3 are compared in this month. Crustaceans as a whole separate the raft areas with P. longicornis and other decapods dominating, from the beach zone, where amphipods and other non-decapod species are abundant. We observed highly significant t values, P < 0.01, when consumption of P. longicornis and amphipods is compared between each raft zone and the beach area. 48 FREIRE ET AL. TABLE 1. Diet components of L. arcuatus. Values in % points (% frequency of occurrence) for the different stations and months sampled (— = absent). February-83 July-81 Bl Bo P3 Bl B6 P3 TOTAL Number examined 26 112 128 77 158 144 645 Number with food (19) (56) (62) (51) (93) (71) (352) Fishes 8.9 (3.8) = 6.5 (8.1) 2.7 (2.6) iC) (67) 1.4 (1.4) 2.6 (3.1) Gobiidae 8.9 (3.8) — 3.1 (1.6) — 1.0 (1.1) 1.4 (1.4) Iesy (lei) Gobius sp. — -- — — 1.0 (1.1) = 0.3 (0.3) Pomatoschistus minutus — —_— _ — _ 1.4 (1.4) 0.3 (0.3) Gobiidae, undet. 8.9 (3.8) — SE (EG) — = — 0.9 (0.5) Fishes, undet. _— a 3.4 (4.8) at} (PL) 0.9 (2.2) — 12 GES) Crustaceans 21.0 (15.4) 14.2 (10.7) 20.3 (19.4) 10.0 (10.4) 19.0 (25.8) 55.3 (57.7) 21.1 (25.9) Decapods 20.4 (7.7) 8.0 (4.5) 7.0 (4.8) 10.0 (10.4) 13.6 (19.4) 7.4 (7.0) 10.1 (8.5) Natantia 3.8 (3.8) — 1.4 (1.6) — _ Sal! (O40) 1.6 (1.6) Crangon crangon 3.8 (3.8) — - — a 0.7 (1.4) 0.3 (0.5) Natantia, undet. — — 1.4 (1.6) — — 5.0 (4.2) Nes} (al5i) Brachyurans — — 3:6) (6:2) = 4.3 (6.5) 1.7 (1.4) 2.5 (2.6) Pilumnus hirtellus _- oe Shall (Gles)) -- a a 0.5 (0.3) Brachyurans, undet. — — 2.5 (1.6) — 4.3 (6.5) 1.7 (1.4) 2:0) (233) Decapods, undet. 6.4 (3.8) 6.4 (3.6) — — 7 (2-2) — 1.8 (1.3) Pisidia longicornis 10.2 (3.8) 1.6 (0.9) — 10.0 (10.4) 7.6 (10.8) —_ 4.1 (4.7) Amphipods — — 2-8) \(B°2) —_ 0.5 (1.1) 11.9 (14.1) 32) (7) Mysids -- 3525 (OF9) — — — —- 0.5 (0.3) Copepods — — — a — 2.2 (1.4) 0.5 (0.3) Crustacean eggs — — — — 17 (eA) — 0.5 (0.3) Crustaceans, undet. 0.6 (3.8) 3.0 (5.4) 10.6 (11.3) — 3.2 (6.5) 33.8 (35.2) 10.5 (11.9) Molluscs 32.5 (23.1) 19.2 (17.0) 17.3 (25.8) 53.7 (44.2) 28.2 (47.3) 333) (12257)) 23.0 (29.9) Bivalves 20.4 (15.4) 13.2 (11.6) 4.9 (4.8) 51.0 (44.2) 25.5 (43.0) 3.1 (11.3) 18.3 (23.6) Mytilus edulis 20.4 (15.4) 10.7. (8.9) — 51.0 (44.2) 13821235) 1.0 (4.2) 13.0 (15.8) Musculus marmoratus — 0.8 (1.8) - — — — 0.1 (0.3) Nucula sp. — — — = 120) 7G:2) — 0.3 (0.9) Cardiacea — — _— Oo 1.0 (2.8) 0.2 (0.6) Lucinacea — 1.8 (1.8) — a — —_ 0.3 (0.3) Bivalves, undet. — — 4.9 (4.8) = 11.3 (18.3) Isl) (26) 4.4 (6.8) Gastropods —_— 4.5 (4.5) 4.2 (14.5) Dali (123) 0.3 (2.2) 0.2 (1.4) 1.8 (4.3) Nassa sp. — 3.7 (3.6) 1.4 (1.6) PP (GIES) — — Lee) Gibbula tumida — 0.8 (0.9) i — — — 0.1 (0.3) Gastropods, undet. a _— 2.9 (12.9) — 0.3 (2.2) 0.2 (1.4) 0.6 (3.1) Opistobranchs — — 5.7 (3.2) = — —_— 0.9 (0.6) Aplysia sp. — — 2.9 (1.6) = = — 0.5 (0.3) Opistobranchs, undet. — —_ 2.8 (1.6) — — — 0:5), 1(023) Molluscs, undet. P2217) 1S) G7) 2.4 (3.2) 0.6 (1.3) Dsy (BP) a 1.9 (2.4) Ophiuroids — OMG) — —_— a 2.4 (1.4) 1.6 (0.7) Ascidians — 0.3 (0.9) = — — — 0.1 (0.3) Hydrozoans — — 0.0 (1.6) 0.2 (1.3) = — 0.0 (0.5) Foraminiferans — 0.7 (8.0) 2.7 (29.0) 0.0 (1.3) 1.4 (18.3) 1.2 (14.1) 122)(1453) Polychaetes — 22.2 (14.3) 0.3 (1.6) — 11.8 (12.9) 2.8 (2.8) el! \(G:5) Aphroditidae —- — — _ DET \(222) — 0.8 (0.6) Arenicola sp. — — — — 5.0 (3.2) — 1.5 (0.9) Capitellidae — _ a = ONCE) = 0.2 (0.3) Chaetopteridae — — — _ OS ae) 0.8 (1.4) 0.3 (0.6) Eunicidae — 3.2 (0.9) — — — —_— 0.5 (0.3) Nereidae — 1.1 (0.9) — a 1.4 (1.1) — 0.6 (0.4) Platynereis dumerilii — a — — 1.4 (1.1) — 0.4 (0.3) Nereidae, undet. — 1.1 (0.9) —_— _ — — 0.2 (0.3) Phyllodocidae — 2.6 (0.9) — = — — 0.4 (0.3) Terebellidae — = —_ — OFS G11) — 0.2 (0.3) Polychaetes, undet. — 15.3 (11.6) 0.3 (1.6) _ 0.8 (3.2) 149)(1:4) 3:26:33) MUSSEL RAFT CULTURE AND DIET OF L. ARCUATUS 49 TABLE 1. continued February-83 July-81 BI Bo P3 Bl B6 P3 TOTAL Number examined 26 112 128 77 158 144 645 Number with food (19) (56) (62) (51) (93) (71) (352) Seaweeds 21.0 (15.4) 15.0 (10.7) 38.0 (43.5) 6.7 (5.2) 28.0 (38.7) 25.3 (29.6) 24.0 (27.2) Ulvaceans LOR) 15.0 (10.7) 31.9 (27.4) 5:9) 1829) 23.3 (35.5) 22.0 (25.4) 20.8 (22.0) Ulvaceans, undet. 19.1 (7.7) 15.0 (10.7) 31.9 (27.4) 5.9 (3.9) 23.0 (34.4) 22.0 (25.4) 20.7 (21.7) Enteromorpha ramulosa — — = —_— O:35 (en) = 0.1 (0.3) Brown seaweeds — _— 3:3)9(8:2) 0.4 (1.3) 8p (3¥2) 2.6 (2.8) Pian (22) Ectocarpaceans — — Vy C126) 0.4 (1.3) 1.8 (3.2) 2.4 (1.4) 1.3 (1.6) Gifordia sp. — — — = les} (C74) 1.2 (1.4) 0.6 (0.9) Ectocarpus sp. — —— = = OF Gen) 1.2 (1.4) 0.3 (0.3) Ectocarpaceans, undet. _ — 1.1 (1.6) 0.4 (1.3) 0:3) Gen) — 0.3 (0.8) Fucus sp. — — 1:0: (126) — a — 0.2 (0.3) Sphacellaria sp. =. — 1.2 (1.6) — 0.2 (1.4) 0.3 (0.6) Red seaweeds — — — — 15) (2:2) — 0.4 (0.6) Polysiphonia sp. -- a — — (252 -- 0.3 (0.6) Audovinella sp. — a — a OFS) 11) — 0.1 (0.3) Seaweeds, undet. EO Ast) — 2.8 (16.1) 0.4 (1.3) 1.4 (3.2) 0.7 (2.8) 1.2 (4.9) Zostera sp. a 9.9 (8.0) 1.9 (6.5) 20.9 (18.2) 2.2 - (4.3) = 5.2 (6.2) Animal components, undet. 16.6 (23.1) 11.4 (13.4) 12.8 (27.4) 5.7 (7.8) 7.4 (34.4) 8.4 (18.3) 9.3 (22.1) Molluscs, especially M. edulis, as well as Zostera de- crease gradually from B1 to P3 with an intermediate per- centage shown in B6. Thus we can find highly significant differences (P < 0.01) for the mussel in the three pairs of stations, with the maximum value being t = 7.41 between B1 and P3 in July. The same occurs with Zostera, but the significance level reaches 99% only between B1 and P3 and between B1 and B6, both in July. The presence of polychaetes in the diet of L. arcuatus separates station B6 from the others, and as a result, this station presents its own special features, and not only as a intermediate zone. However, seaweeds differentiate Bl from the other zones where they are much more abundant. Differences be- tween B6 and P3 are not significant (a difference is only observed in July for total seaweeds but with P < 0.05, al- though this difference does not exist in the case of ulva- ceans). Comparing B1 with the other stations using July samples, where number of individuals is higher, clear sig- nificant differences were obtained. The same pattern noted for the seaweeds is also followed by the foraminiferans, which despite their minor quantitative value, show a very high value in terms of frequency of occurrence (a total of 14.3% and up to 29.0% in P3 in the February sample). Therefore they could serve as an indication of the environ- ment in which L. arcuatus feeds. The presence of foramin- iferans means that sediment and existing infauna is being used. Although each station was sampled in July 1981 and February 1983 (Table 1), differences between both samples are less important than the differences between stations. The only notable change corresponds to M. edulis con- sumption in B1, although the small size of the sample in February does not allow us to evaluate it correctly. We may also note the increase of non-decapod crustaceans in the diet of individuals found in P3 in July. The latter is due to the fact that the number of small size individuals in this sample is greater than in February (crustacean consumption is higher in these sizes—see next section). However it is not correct to talk about seasonal differences when there may be interannual variations. If we consider the number of components and diversity values of the diets corresponding to the different zones (Table 3), we can confirm what we mentioned earlier, up to a certain point. B1 shows the lowest value (H’ = 2.65), due the dominance of a few components, particularly M. edulis. In P3 we find an increase in diversity values (3.59) as the dominance of certain components is smaller. There- fore B6 is the station where the diet has the largest number of components, showing the highest diversity (4.24). This confirms its intermediate nature in that the great variety of existing resources are used in a more homogeneous manner. Diet Analysis by Size Group Through the analysis of diet composition by size, sex, molt stage and maturity, we found that the major changes in the diet were due to growth. If we consider the overall data for all stations and months (Table 4) and the Kruskal-Wallis test data (Table 5), an increase in size means a decrease in non-decapod 50 FREIRE ET AL. TABLE 2. Comparison of main component consumption in the diet of L. arcuatus (using absolute data in points) among stations, using the Kruskal- Wallis test (H values are distributed as a X*a, 2); and an a posteriori comparison of stations in pairs using the Mann-Whitney test (t values are distributed as a ta, ~). Mann-Whitney B1/B6 B1/P3 B6/P3 Kruskal-Wallis FEB-83 JUL-81 TOTAL FEB-83 JUL-81 TOTAL FEB-83 JUL-81 TOTAL Crustaceans L82534** 0.00 ns 1.28 ns 1.03 ns 0.06 ns 4.91*** 3.86*** 0.12 ns 4.47*** 356% Pisidia 1S24 k** 0.83 ns 0.79 ns 1.31 ns 1.81* B144e** 4.22*** 1.05 ns 2.84*** 3.17*** Other decapods 8.09** 0.47 ns 1.85* 2.40** 0.84 ns 1.93 ns 1.02 ns 0.05 ns 0.16 ns ee i Amphipods 17S OF* 0.00 ns 0.65 ns 0.68 ns 0.79 ns DI Sees DSSEE* 1.35 ns Se2055% 3.375% Molluscs S7ALO** 0.01 ns OME 22655 0.64 ns 6.46*** 550*** 0.99 ns Seliee* 4.53*** Mytilus 76.98*** 0.39 ns Si 9FE* 4.84*** S1685** TEAL O76*** 3465** 3.845** So! A Other bivalves 15.04*** 1.19 ns 3.34*** Brag tex 0.97 ns 2a2** DeDDe* 0.46 ns 2.13** 223255 Foraminiferans LAS 2 *x* 1.85* 2.84*** 3 34Exx 216245* DeS2DEs 6.50*** 1.73* 0.70 ns 0.84 ns Polychaetes 33.35 54% DiGSt** DIBES* 3:91 *2* 0.11 ns 0.26 ns 1.26 ns 4.35*** 2.44** 4.51*** Seaweeds 14.59*** 0.06 ns 3.94*** 3.08*** NEAT. 3.09*** Sol ees DES 4 Ses 0.91 ns 0.80 ns Ulvaceans S229555 0.89 ns SiS an ee 3:97 E** 1.38 ns DO355* 3530245 1.06 ns 0.97 ns 0.71 ns Zostera ig/eiee2 1.85* 4.04*** 2.42** 1.13 ns 4.66*** S:695** 75s* 1.76* 2.03** *** P< 0/01 H > 9.21 t > 2.58 ** P< 0.05 H > 5.99 t> 1.96 *P<0.1 H > 4.61 t> 1.65 crustaceans in the diet, with an abundance of decapods in intermediate size individuals (between 20 and 39 mm). However, neither P. Jongicornis nor the other decapods in the various size classes show significant differences whereas the amphipods do (P < 0.01). Larger sized individuals of L. arcuatus consume a higher quantity of molluscs (P < 0.05), both bivalves and gastropods. M. edulis in particular appears in the diet of larger individuals more frequently, but to a lesser extent and it is not significant statistically. The ophiuroids follow the same pattern as the molluscs, reaching 6.0% of the points in individuals larger than 40 mm, but they are not as important in the diet. Seaweeds and polychaetes do not show clear variations among the different size groups. We first analyzed diet by size group separately for both samples taken from each station. The Kruskal-Wallis test does not show significant differences in most cases (except for polychaetes in B6 February —H = 10.18, P <0.01—, and crustaceans in P3 July —H = 6.88, P < 0.05 —). This may be attributed to the small number of individuals for each group size when considering each sample separately. Therefore as spatial variations are much greater than the variations between the different samples from each station, the test was applied by size group for each station (Ta- ble 5). Diet composition and available prey are very different in each zone and the same component may be used in varying amounts, depending on the abundance of other available prey in each area (Fig. 3). In station B1 where all individuals examined had a cara- pace width measuring at least 20 mm, and a larger average size than in the other stations (unpublished data), no signif- icant statistical differences were found among the different size groups. This is due to the overwhelming dominance of M. edulis, which does not show great variations, although it does increase slightly in larger sizes. Molluscs actually comprise 84.5% of the points in the 40—49 mm class diet, whereas plants and Zostera in particular are more abundant in individuals measuring 20—29 mm, representing nearly 30% of the diet. Individuals captured in station P3 have a smaller average size than in raft stations, the maximum being 39 mm. In this zone, the diet of L. arcuatus analyzed by size group TABLE 3. Number of diet components (N) and diet diversity (H’) in L. arcuatus for each station, sample, size group and total. Carapace width mm 10-19 20-29 30-39 40-49 TOTAL N H N H N /H! oN SHOR Ne Bl a 1 2°31 13) 2162. 5) GG R2765 B6 13° 2:91 25 4:04 26 3:94 18 3°26) 41" 4i24 P3 Te 1291) 25) 93635 18st 5 — 320 e3t59 BIFEB — Si 227 47 S410 2296 BIJUL _ 8) 2227) 92102) (O87 2a B6FEB = 14° 3:19) 14: 3:28) 13) 3015 20ReSeit B6JUL 13, 2:91) 19) 93276) 920)-93'33) 9) 32312 noe P3FEB 5 54 16) 3109) 15) 3'43 — 23) 3160 P3JUL Sy dls7ey ils Sesh 1K) 22-777 7 21 3:07 Total 14 2:57 40 (4°35) 33) 4:00) 919) 9)3'425 s53eN4226 MUSSEL RAFT CULTURE AND DIET OF L. ARCUATUS 51 TABLE 4. Diet components of L. arcuatus by size group (maximum carapace width in mm). Values in % points (% frequency of occurrence) (— = absent). Carapace width mm 10-19 20-29 30-39 40-49 Number examined 48 295 258 1 Number with food (38) (139) (145) (28) Fishes — ZO (222) 3.9 (4.8) 0.6 (3.6) Gobiidae _ 1.5 (1.4) 2.3 (1.4) = Gobius sp. = 0.8 (0.7) — _ Pomatoschistus minutus — 0.8 (0.7) = — Gobiidae, undet. —_ oa Psys { (la) — Fishes, undet. a 13087) 1.6 (3.4) 0.6 (3.6) Crustaceans 50.2 (47.4) 27.6 (30.9) 19.0 (20.7) 1.5 (10.7) Decapods 14971(5:3) 12.4 (13.7) 11.5 (11.0) PD (75310) Natantia — 25 7e (229) 1.3 (1.4) = Crangon crangon — 0.4 (0.7) 0.4 (0.7) == Natantia, undet. — 2-3 (222) 0.9 (0.7) = Brachyurans _ 32271222) 353)1(251)) 0.3 (3.6) Pilumnus hirtellus — 123) (O37) — — Brachyurans, undet. — 1.9 (1.4) Eyer (PAID) 0.3 (3.6) Decapods, undet. — 2.4 (2.2) 2.0 (2.1) 0.9 (3.6) Pisidia longicornis 1.4 (5.3) 5.2 (6.5) 4.9 (6.2) — Amphipods 11.6 (7.9) 4.2 (5.8) — = Mysids _ — 1.4 (0.7) = Copepods — 1.1 (0.7) — — Crustacean eggs — ites) (7) — — Crustaceans, undet. 37.2 (34.2) 8.6 (11.5) 6.2 (8.3) 0.3 (3.6) Molluscs 9.8 (21.1) 18.7 (30.9) 26.8 (40.0) 43.5 (53.6) Bivalves 9.0 (18.4) 15.2 (24.5) 20.5 (31.7) 33.7 (39.3) Mytilus edulis 5.2 (7.9) 11.3 (19.4) 15.9 (24.1) 17.9 (25.0) Musculus marmoratus — — os ies G10) Nucula sp. 1k6" (6:3) = 0.3 (2.1) — Cardiacea — 0.5 (1.7) — —_— Lucinacea — OS O87) 0.2 (0.7) — Bivalves, undet. 2322) 2.9 (5.8) 4.2 (5.5) 14:3: (en) Gastropods 0.4 (2.6) 1.7 (5.8) 1.9 (3.4) 4.0 (14.3) Nassa sp. os 0.7 (1.4) 1.6 (1.4) 2.4 (7.1) Gibbula tumida — — — 1.5 (3.6) Gastropods, undet. 0.4 (2.6) 1.0 (4.3) 0.4 (2.1) — Opistobranchs — 1.1 (0.7) ESI (O27) — Aplysia sp. — — 123) (0:7) — Opistobranchs, undet. — Nl (32) — _— Molluscs, undet. 0.4 (2.6) 0.6 (2.2) 3.1 (4.1) 3:8) (Gel) Ophiuroids _ 0.6 (0.7) 2.3 (1.4) 6.0 (3.6) Ascidians a — = 0.6 (3.6) Hydrozoans — 0.0 (0.7) 0.1 (0.7) — Foraminiferans 0.1 (10.5) 1.8 (16.5) 1.1 (12.4) 0.7 (14.3) Polychaetes 4.4 (2.6) 12.2 (10.8) 3.8 (6.2) 10.9 (17.9) Aphroditidae _ 1.9 (0.7) = = Arenicola sp. — 2-5) (14) 1e2(OW7)) — Capitellidae _— 0.8 (0.7) — — Chaetopteridae 1.2 (2.6) — 0.5 (1.4) —_— Eunicidae — 1.3 (0.7) a —_— Nereidae 3.3 (2.6) 0.4 (0.7) — — Platynereis dumerilii 3.3 (2.6) —_— = = Nereidae, undet. — 0.4 (0.7) — — Phyllodocidae — — — 4.9 (3.6) Terebellidae — 0.6 (0.7) _ —_— Polychaetes, undet. — ATE)! 2.1 (2.8) 6.1 (10.7) continued on next page 52 FREIRE ET AL. TABLE 4. continued Carapace width mm 10-19 20-29 30-39 40-49 Number examined 48 295 258 44 Number with food (38) (139) (145) (28) Seaweeds 24.5 (44.7) 23.8 (28.1) 25.2 (23.4) 21.9 (21.4) Ulvaceans 23.7 (42.1) 20.1 (22.3) 21.1 (17.9) 20.7 (14.3) Ulvaceans, undet. 23.7 (42.1) 20.1 (21.6) 20.9 (17.9) 20.7 (14.3) Enteromorpha ramulosa — — 0.2 (0.7) — Brown seaweeds — 2.3 (3.6) 1.9 (2.1) — Ectocarpaceans _— 1.8 (2.9) 1.4 (1.4) — Gifordia sp. — 0.6 (1.4) 1.0 (0.7) — Ectocarpus sp. — 0.7 (0.7) a — Ectocarpaceans, undet. _ 0.4 (1.4) 0.4 (0.7) — Fucus sp. — 0.4 (0.7) — — Sphacellaria sp. — 0.1 (0.7) 0.5 (0.7) = Red seaweeds — — 1.3 (1.4) — Polysiphonia sp. — _ 0.9 (1.4) — Audovinella sp. — — 0.3 (0.7) — Seaweeds, undet. 0.8 (7.9) 1.5 (4.3) 0.9 (4.8) PA (7/21) Zostera sp. _ 4.3 (4.3) 8.3 (14.5) 3.6 (10.7) Animal components, undet. 11.2 (18.4) 7.9 (29.5) 9.8 (20.0) 10.6 (21.4) has the same characteristics as seen in the total data. We can see a decrease in the non-decapod crustaceans (P < 0.01 for amphipods), and an increase in decapods and mol- luscs (P < 0.01 for the two components) with growth. Sea- weeds do not undergo any major changes, although in the February 1983 sample, their consumption by individuals in the 30—39 mm size class (24.7% of the diet) is reduced, while in the smaller class size, they make up 50% of the diet. However in the July 1981 sample, when seaweeds are less abundant, we found no appreciable variations. In station B6 decapod crustaceans are consumed to a lesser extent by extreme sized individuals, large and small alike (P < 0.01 for P. longicornis and other decapods), whereas consumption of non-decapods, which are of minor importance in this zone, decreases in accordance with size. Predation of molluscs increases with size, especially in in- dividuals over 40 mm. This holds true for bivalves (P < 0.01) as well as gastropods, although consumption of M. edulis remains constant. Seaweeds and, to a lesser extent, ophiuroids, also become more important in larger sized in- dividuals. Seaweeds stand out in individuals over 30 mm, with highly significant H values (P < 0.01), whereas poly- chaetes decrease as of the 30 mm size (P < 0.01). For all samples, highest diversity values in the diet cor- respond to the intermediate sizes (20-39 mm), while in extreme size groups values decrease, owing to a more spe- cialized diet based on few components. Analysis by Sex The Mann-Whitney test was applied to compare the consumption of different components between males and females. Total data as well as information from each station and sample show non-significant t values, except in the case of Zostera in B6 total data (t = 2.68, P < 0.01), while in B1, although differences were also observed, they did not reach significant values (t = 1.84, P < 0.1). In both stations Zostera is consumed in greater quantity by males. Although total data did not show significant differences between ovigerous and non-ovigerous females, in station B1 ovigerous females did consume more molluscs (t = 3.27, P < 0.01) and in particular, mussels (t = 2.97, P< 0.01), possibly due to their availability and easy accessi- bility. In P3 non-ovigerous females consume more crusta- ceans (t = 2.24, P < 0.05). This is perhaps because of their size (the ovigerous females are larger), which stands out over the ability to capture different prey. In this station the ovigerous females also consume slightly more seaweed, which again is more accessible, although on a less signifi- cant level (P < 0.01). Diet Analysis During the Intermolt Cycle Of the individuals analyzed, 18 were in the pre-ecdysis stage, and in all of the cases, their stomachs were empty of food. This indicates that during the pre-molt period this species stops feeding, as do other decapods. Thirty-four in- dividuals in the early stages (periods A and B, Drach & Tchernigovtzeff, 1967) of the intermolt cycle were ana- lyzed. Although no significant changes in the diet were seen with respect to individuals in the intermolt stage, in general a higher consumption of P. longicornis was ob- served, although with a low significant level (Mann- MUSSEL RAFT CULTURE AND DIET OF L. ARCUATUS 53 TABLE 5. Comparison of main component consumption in the diet of L. arcuatus (using absolute data points) between size groups, for total data as well as for each station, using the Kruskal-Wallis test (H values are distributed as X?a, d.f.). Complementary analyses were carried out separately for each station and date (referred to in the text). Carapace width TOTAL BI B6 P3 10-19mm 38 — 14 24 20—29mm 140 12 60 68 30—39mm 148 48 61 39 40-49mm 28 10 16 2 3 df. 2 df. 3 df. 3 df. Crustaceans 22NISEES 1.05 ns Uae S291 Pisidia 1.91 ns 0.59 ns 19.98*** a Other decapods 4.27 ns — 11.34*** ZOOS Amphipods 14.85*** —_— — LOGE Molluscs 11.00** 3.49 ns Telit. 11293 54% Mytilus 6.04 ns 1.77 ns 6.88 ns _- Other bivalves 2.17 ns — 15.89*** 13 .16*** Foraminiferans 5.48 ns — 853% 1299 e* Polychaetes O25 se — 135 24*5% — Seaweeds Lai 2eee 0.35 ns Bes 5.49 ns Ulvaceans 18.92*** 0.43 ns 16.60*** 7.90** Zostera LOMS** 0.48 ns EOF iT tts — 2 g.l. Sirals *** P < 0.01 H > 9.21 H > 11.35

5.99 7582 *P<0.1 H > 4.61 Hi 6:25 Whitney test t = 1.66, P < 0.1). In B6 fewer molluscs are consumed (t = 2.23, P < 0.05), specifically the mussel (t = 1.68, P < 0.1). We can see a greater tendency towards the consumption of smaller sized prey which are more @ OTHER DECAPODS PISIDIA AMPHIPODS OTHER CRUSTACEANS MYTILUS OTHER BIVALVES POLYCHAETES @ QO 8 0 HH} =| VA SEAWEEDS ZOSTERA ANY OTHERS UNIDENTIFIED & 8 40-49 CARAPACE WIDTH mm 30-39 20-29 Figure 3. Distribution of main components in the diet of L. arcuatus in each station by size group. easily accessible. In P3 differences were not analyzed, due to the low number of individuals in state B. Association Analysis In the association analysis carried out between station- date samples (Fig. 4), different groups are formed. The pair having the greatest similarity is made up of the two samples from station B6, followed by the pair in P3. The sample taken in July from the inner raft station B1 is clearly sepa- rated from the others due to the high consumption of mussel and Zostera, and the minor importance attributed to seaweed. However in February B1 shows a greater simi- larity to B6 than to P3, forming a raft group opposite the beach group. The association analysis done comparing the main diet components (Fig. 5) gives rise to the formation of two large groups. One is made up of typical beach area components, although some of these also appear in raft zones, such as the ulvaceans, other bivalves, etc. Within this first group we can separate amphipods, undetermined crustaceans and natantia, all more abundant in July, from the rest of the components, which are more important in February (prob- ably due to the higher contribution of small size individuals in the July sample). The second group is composed of diet components consumed mainly in the raft areas. In this second group polychaetes, typical in station B6, and unde- termined decapods, more abundant in the same zone, can be distinguished, and are separated from P. longicornis, M. edulis and Zostera, appearing in both raft stations, but in greater abundance in B1. 40 50 60 INDEX 70 PSC 80 90 100 Bay IU B1 FEB B6 FEB B6 JUL P3FEB P3 JUL Figure 4. Dendrogram of station-date groups based on point per- centage data of main diet components. 54 FREIRE ET AL. PSC INDEX ° © Ca ~ a wo cs w ne = tS) ° ° } ° ° re) ro) 5 3 ° ESS Ss SS SS SS ee AMPHIPODS C RUSTACEANS NATANTIA BRACHYURA BROWN SEAWEEDS OTHER BIVALVES ULVACEANS PISIDIA MYTILUS ZOSTERA OTHER DECAPODS POLYCHAETES Figure 5. Dendrogram of main diet components using point per- centage data for each station-date. Principal Component Analysis This analysis presents four axes; the first three explain 85% of the variance (Fig. 6, 7, 8 and 9). Axis I, which makes up 47.6% of the variance, sepa- rates the beach samples as well as the typical components of this area, such as the ulvaceans, amphipods, etc, that show a high positive correlation, from the typical raft sta- tion B1 which presents a negative correlation, as do its typ- ical components, M. edulis, Zostera, and P. longicornis. Station B6 is situated in an intermediate area and has very little correlation with axis I (Fig. 7 and 9). This axis ap- pears to represent the contribution of food from the rafts, separating diets and typical components from bottoms which have been transformed by detritus with an abun- dance of organisms originating from the epifauna (raft zones), from non-transformed bottoms, the beach areas. Station B6 benefits from the food contribution from the rafts, even though they are primarily devoted to oyster cul- ture. These bottoms have also undergone very little trans- formation with abundance of seaweed, showing a greater similarity to those of beach zones. This is why B6 is con- sidered to be an intermediate station. Axis Il, which makes up 21.4% of the variance, sepa- rates the B6 samples, which are positively correlated, from the other two stations. Polychaetes and other bivalves, which are very abundant in B6, appear to be positively cor- related on this axis. Axis II also shows the variance be- tween both beach station samples; amphipods, natantia and undetermined crustaceans, more abundant in the July sample, are negatively correlated. II BIVALVES » OTHER BRACHYURA © POLYCHAETES « ULVACEANS © BROWN SEAWEEDSe | * OTHER DECAYPODS PISIDIA © ZOSTERA ° MYTILUS OTHER CRUSTACEANS » AMPHIPODS © NATANTIA © Figure 6. PCA. Factorial loads of diet components on axes I and II. Axis III, with a lower variance percentage, 16.1%, may correspond to the sample variance within the raft zones, in particular Bl, as the July sample and M. edulis and Zos- tera, which are very abundant in the diet in July 1981, show positive correlations, whereas the undetermined decapods, only present in February, have a high negative value (Fig. 8 and 9). The negative correlation of the poly- chaetes and undetermined decapods is possibly due some- what to its variation in B1, this as well as the positive cor- relation of P. longicornis can be attributed to the higher abundance in February and July respectively in B6. How- ever, because of the great diversity in the diet, there are fewer variations in this station. DISCUSSION The trophic web in the Ria de Arousa has been altered considerably by the introduction of intense mussel culture Mytilus edulis on some 2000 rafts, grouped in polygons, comprising 10% of the surface of the ria (Tenore & Gon- zalez, 1975; Tenore et al., 1982, 1985). The ‘‘upper part’’ of this web (Arntz, 1978) made up chiefly of megabenthos, has undergone a lot of changes as a result. In general, they efficiently use the new resources at their disposal, so their biomass reaches much higher values in raft areas than in zones where there are none. In the case of crustaceans these values are as much as 6 times higher (Chesney & Iglesias, 1979; Gonzalez-Gurriaran, 1982; Olaso, 1982; Romero et al., 1982). Within this group of organisms, certain species have benefitted, dominating the raft area communities; demersal fishes, such as Lesueurigobius friesii (Malm), Gobius niger L., Trisopterus luscus (L.) and Pomatoschistus minutus MUSSEL RAFT CULTURE AND DIET OF L. ARCUATUS 55 Figure 7. PCA. Distribution of station-date groups on axes I and II. (Pallas), echinoderms such as Asterias rubens L. and Aslia lefevrei (Barrois), and the decapods Liocarcinus depurator and Necora puber. Although L. arcuatus is typically found in beach areas, it is fairly abundant in raft zones, especially if the bottoms have not undergone much transformation or if they are shallow river runoff areas (Gonzalez-Gurriaran, 1982). We can see similarities in the diet of L. arcuatus in the beach zone of the Ria de Arousa and in the Adriatic Sea (Stevcic, 1987), where there is an abundance of shallow waters, having a high presence of seagrass and seaweeds. In the Adriatic Sea the diet consists mainly of plants, chiefly Zostera and Zosterella, while in Arousa ulvaceans (26.3%) predominate. However the largest percent of the diet is made up of crustaceans (40.0%), which are also very important in the Adriatic, and to a lesser extent we find the molluscs. We wish to emphasize the major role that plants play in the diet of L. arcuatus in both areas. The differ- ences can probably be attributed to the greater importance of the seagrass in the bottom areas studied in the Adriatic and to the ulvaceans in Arousa. Brachyurans do not frequently consume plants, but some species have been known to do so, especially in shallow estuary type zones. Carcinus maenas (L.), which has a distribution similar to that of L. arcuatus in Arousa (Gonzalez-Gurriaran, 1982) has a variable diet, based on the area studied. Green seaweeds and polychaetes are dom- inant and molluscs appear in smaller proportions (Le Calvez, 1987), which is largely similar to the diet of L. arcuatus in B6. In other zones the main component is made up of bivalves and crustaceans, and seaweeds, although in a smaller percentage (Ropes, 1969; Elner, 1981). In an- MI « MYTILUS BROWN SEAWEEDS © 1OSTERA Siigts : BRACHYURA PISIDIA OTHER} BIVALVES . AMPHIPODS | CRUSTACEANS © ° e _ULVACEANS NATANTIA OTHER POLYCHAETES e OTHER DECAYPODS Figure 8. PCA. Factorial loads of diet components on axes II and III. other portunid Callinectes sapidus Rathbun, seaweeds may be a major part of the diet if they are abundant in the zone (Alexander, 1986) or they hardly appear at all in the stomach contents if they are not common to the habitat studied (McLaughlin, 1979). In most of the studies that were carried out, the diet is shown to be conditioned largely by the abundance of the different potential prey, as well as by species preference. Thus L. arcuatus living in raft areas shows a change in Bl JUL @ P3 FEB Figure 9. PCA. Distribution of station-date groups on axes II and III. 56 FREIRE ET AL. feeding habits. This also happens with other epibenthic species in the Ria de Arousa, which start consuming more abundant prey in these areas. In a typical raft area like B1, L. arcuatus feeds mainly on the mussel M. edulis (43.5% of the diet), as well as epifauna components such as P. longicornis (10.0%). The infauna, however, composed chiefly of polychaetes in this area (90 to 100% of the bio- mass) (Lopez-Jamar, 1982), are not found in stomach con- tents. In spite of the lack of studies on oyster raft epifauna and infauna in the other raft zone B6, we assume that L. arcuatus probably takes advantage of the infauna in this station to a greater extent. Therefore, this station’s biomass should be greater, since bottoms have not been transformed (Lopez-Jamar, 1982), even though this species still utilizes resources offered by the rafts. This is proven by the abun- dance in the diet of polychaetes (15.5%), seaweed (23.4%) or molluscs (25.1% —of this M. edulis comprises only half). Diet analysis in raft zones must be related to culture dy- namics and to the evolution of the communities settled on ropes. The mussel as well as the epifauna growing on the ropes, reach high biomass levels (Gonzalez-Sanjurjo, 1982; Roman & Pérez, 1982), and culture dynamics involve gath- ering activities, cleaning and dividing up the ropes, con- tributing more food to the bottom that can be consumed by the crabs. In July there is more work to be done on the rafts and the mussel reaches a larger mean size, perhaps making it more appropriate for consumption by L. arcuatus than in February when there are more ropes with spat of a very small size (Marino et al., 1982). This could explain the higher abundance of mussels in the stomach contents in July. The same holds true for L. depurator (Gonzalez-Gur- riaran et al., in press). However, the development of the communities settled on ropes (Gonzalez-Sanjurjo, 1982; Roman & Pérez, 1982; Lapointe et al., 1984) involves a variation in the avail- ability of the different prey throughout the year. We ob- served that seaweed consumption in the inner raft zone, Bl, is much higher in February than in July, when it is replaced by Zostera. This coincides with the fact that the seaweed community settled on raft ropes show a large pro- portion of ulvaceans in winter, whereas in summer brown seaweeds predominate (Lapointe et al., 1984). The latter may not be appropriate for consumption by L. arcuatus due to their size and consistency. In any case, we must bear in mind the possibility of interannual variations that may overlap with seasonal variations. Diet variations in L. arcuatus within a given area, are chiefly related to size although, to a certain extent they may also be attributed to the intermolt stage. Growth is accom- panied by a decrease in crustacean consumption and an in- crease in mollusc consumption. These changes have also been observed in other brachyurans such as Callinectes ar- cuatus Ordway (Paul, 1981), C. sapidus (McLaughlin, 1979) and C. maenas (Ropes, 1969). This could be caused by the close link between predator and prey size; the energy obtained can be maximized and/or the time spent in con- suming the prey can be reduced depending on the predator size (Elner & Hughes, 1978; Hughes & Seed, 1981). This is probably the reason why L. depurator with carapace widths of over 40 mm. (similar to the maximum size of L. arcuatus), increases its consumption of mussels in the raft areas (Gonzalez-Gurriaran et al., in press). L. arcuatus may migrate within the Ria de Arousa in relation to growth, and motivated by food resources in the raft areas; during which time this species utilizes the abun- dant resources found in the raft areas. In fact, the average individual size in these areas is larger than in beach zones (Gonzalez-Gurriaran, unpublished data). In these zones both L. arcuatus and L. depurator have a diet dominated by the mussel, and to a lesser extent by P. longicornis, al- though practically no plant components appear in the diet of L. depurator. However, another decapod species studied, Necora puber (Gonzalez-Gurriaran, 1978) con- sumes a larger amount of P. longicornis than mussel, prob- ably due to the fact that the mechanical design of its chela is more appropriate for the capture of moving prey than for breaking bivalve shells (ap Rheinallt & Hughes, 1985). The species of demersal fishes whose diet has been studied in raft areas in the Ria de Arousa, consumed mainly P. longicornis (as much as 67% of the diet of Lesueuri- gobius friesii) and also amphipods from the epifauna and infaunal polychaetes. However, mussel was only consumed in small quantities, making up less than 2% of the diet (Chesney & Iglesias, 1979; Lopez-Jamar et al., 1984). Thus decapods and fishes share the food resources coming largely from the rafts. This is correlated with an increase in the megafauna biomass in the raft polygon areas. In any case, food cannot be considered a limiting factor, due to its great abundance and accessibility, but competition between species that use similar resources and co-exist in the same area, decreases. Dominant species in raft zones must be limited by non-trophic causes (territorialism in fish, larval predation, competition for the substratum, etc.), given the fact that there is an excess of food that can also be con- sumed by species not as well adapted to this habitat, such as L. arcuatus. Studies on the megafauna diet in ecosystems manipu- lated and changed by man, as is the case of the Ria de Arousa, are necessary in order to complete our knowledge of the energy flow between different trophic levels (Penas, 1984). Studies are also needed to help us develop and vali- date multispecific models that show predation as the main interaction, and to allow us to explain interactions within the ecosystem (Livingston, 1985); as for example mussel mortality caused by predation, or to determine the effects an increase in mussel culture would have on the system (Penas, 1984). These studies should go into depth on diet MUSSEL RAFT CULTURE AND DIET OF L. ARCUATUS 57 variations caused by different factors, relating predatory population dynamics to abundance cycles and prey dy- namics. ACKNOWLEDGMENTS We would like to thank E. Poza for his help with the stomach contents analyses. Drs. J. Mora and V. Urgormi, J. Barbara and F. Corcobado for their collaboration in deter- mining some of the food components. Part of this research was carried out with funds from the Cooperative Program No. 0020 ‘‘Investigaciones Biold- gicas en las Rias de Galicia,’’ of the Instituto Espanol de Oceanografia in collaboration with the Skidaway Institute of Oceanography, and with funds of the Conselleria de Pesca of the Xunta de Galicia in agreement with FEUGA. LITERATURE CITED Alexander, S. K. 1986. Diet of the blue crab, Callinectes sapidus Rathbun, from nearshore habitats of Galveston Island, Texas. Texas J. Sci. 38:85-89. ap Rheinallt, T. & R. N. Hughes. 1985. Handling methods used by the velvet swimming crab Liocarcinus puber when feeding on molluscs and shore crabs. Mar. Ecol. Prog. Ser. 25:63—70. Amtz, W. E. 1978. The ‘upper part’’ of the benthic food web: the role of macrobenthos in the western Baltic. Rapp. P.-v. Réun. Cons. int. Ex- plor. Mer 173:85—100. Chesney, E. J. & J. Iglesias. 1979. Seasonal distribution, abundance and diversity of demersal fishes in the inner Ria de Arosa, Northwest Spain. Est. Coast. Mar. Sci. 8:227—239. Drach, P. & C. Tchernigovtzeff. 1967. Sur le méthode de détermination des estades d’intermue et son application générale aux crustacés. Vie et Milieu 18:595—609. Elner, R. W. 1981. Diet of green crab Carcinus maenas (L.) from Port Hebert, Southwestern Nova Scotia. J. Shell. Res. 1:89-94. Elner, R. W. & R. N. Hughes. 1978. Energy maximization in the diet of the shore crab, Carcinus maenas. J. Anim. Ecol. 47:103-116. Gonzalez-Gurriaran, E. 1978. Introduccion al estudio de la alimentacion en la nécora, Macropipus puber L. (Decapoda-Brachyura). Bol. Inst. Esp. Oceanog. 4(242):81—93. Gonzalez-Gumaran, E. 1982. Estudio de la comunidad de crustaceos de- capodos (Brachyura) en la Ria de Arousa (Galicia-NW Espana) y su relacion con el cultivo de mejillon en batea. Bol. Inst. Esp. Oceanog. 7(2):223-254. Gonzalez-Gurriaran, E. 1986. Seasonal changes of benthic megafauna in the Ria de Muros e Noia (Galicia, North-West Spain). II. Decapod crustaceans (Brachyura). Mar. Biol. 92:201—210. Gonzalez-Gurmiaran, E., J. Freire, L. Fernandez & E. Poza (in press). Incidencia del cultivo de mejillon en la dieta de Liocarcinus depurator (L.) (Brachyura:Portunidae) en la Ria de Arousa (Galicia, NW Espana). Cah. Biol. Mar. Gonzalez-Sanjurjo, R. 1982. Estudio de la epifauna de la semilla de me- jillon en la Ria de Arosa. Bol. Inst. Esp. Oceanog. 7(336):49—71. Hughes, R. N. & R. Seed. 1981. Size selection of mussel by the blue crab Callinectes sapidus: energy maximizer or time minimizer? Mar. Ecol. Prog. Ser. 6:83-89. Iglesias, J. 1981. Spatial and temporal changes in the demersal fish com- munity of the Ria de Arosa (NW Spain). Mar. Biol. 65:199—208. Iglesias, J. 1982. Ecologia de la comunidad de peces demersales de la Ria de Arosa, con especial referencia a la familia Gobiidae. Fac. Bio- logia, Univ. Santiago. Resumen de Tesis Doctoral, 54 pp. Lapointe, B. E., F. X. Niell & J. M. Fuentes. 1984. Community struc- ture, succession, and production of seaweeds associated with mussel- rafts in the Ria de Arosa, N.W. Spain. Mar. Ecol. Prog. Ser. 5:243- 253. Le Calvez, I. Ch. 1987. Location of the shore crab Carcinus maenas L., in the food web of a managed estuary ecosystem: the Rance basin (Brittany, France). Inv. Pesq. 51(supl.1):431—442. Livingston, P. A. 1985. An ecosystem model evaluation: the importance of fish food habits data. Mar. Fish. Rev. 47:9-12. Lopez-Jamar, E. 1982. Distribucion espacial de las comunidades ben- tonicas infaunales de la Ria de Arosa. Bol. Inst. Esp. Oceanog. 7(347):255—268. Lopez-Jamar, E., J. Iglesias & J. J. Otero. 1984. Contribution of infauna and mussel-raft epifauna to demersal fish diets. Mar. Ecol. Prog. Ser. 15:13-18. Marino, J., A. Pérez & G. Roman. 1982. El cultivo del mejillon (Mytilus edulis L.) en la Ria de Arosa. Bol. Inst. Esp. Oceanog. 7(350):297- 308. McLaughlin, R. A. 1979. Trophic ecology and population distribution of the blue crab, Callinectes sapidus Rathbun, in the Apalachicola Es- tuary (North Florida, U.S.A.). Ph.D. Thesis, Florida State University, 143 pp. Olaso, I. 1982. Ecologia de los equinodermos de la Ria de Arosa. Bol. Inst. Esp. Oceanog. 7(334):3-29. Paul, R. K. G. 1981. Natural diet, feeding and predatory activity of the crabs Callinectes arcuatus and C. toxotes (Decapoda, Brachyura, Por- tunidae). Mar. Ecol. Prog. Ser. 6:91—99. Penas, E. 1984. Modelos de simulacion de ecosistemas: el caso de la Ria de Arosa. Inf. Téc. Inst. Esp. Oceanog. 10, 43 pp. Roman, G. & A. Pérez. 1982. Estudio del mejillon y de su epifauna en los cultivos flotantes de la Ria de Arosa. IV. Evolucion de la comunidad. Bol. Inst. Esp. Oceanog. 7(349):279—296. Romero, P., E. Gonzalez-Gurriaran & E. Penas. 1982. Influence of mussel rafts on spatial and seasonal abundance of crabs in the Ria de Arousa, North-West Spain. Mar. Biol. 72:201—210. Ropes, J. W. 1969. The feeding habits of the green crab, Carcinus maenas (L.). Fish. Bull. Fish. Wildl. Serv. U.S. 67:183—203. Sanders, H. T. 1960. Benthic studies in Buzzards Bay. II. The structure of the soft-bottom community. Limnol. Oceanogr. 5:138—153. Shannon, C. E. & W. Weaver. 1963. The mathematical theory of commu- nication. Urbana: University of Illinois Press, 111 p. Stevcic, Z. 1987. Autoecological investigations of the crab Liocarcinus arcuatus. Inv. Pesq. (supl.1):375—387. Tenore, K. R. & N. Gonzalez. 1975. Food chain patterns in the Ria de Arosa, Spain: an area of intense mussel aquaculture. In: Persoone, G., Jaspers, E. (ed.) 10th Eur. Mar. Biol. Symp., Vol. 2. Universa Press, Wetteren, p. 601-619. Tenore, K. R., L. F. Boyer, R. M. Cal, J. Corral, C. Garcia-Fernandez, N. Gonzalez, E. Gonzalez-Gurriaran, R. B. Hanson, J. Iglesias, M. Krom, E. Lopez-Jamar, J. McClain, M. M. Pamatmat, A. Pérez, D. C. Rhoads, G. Santiago, J. Tietjen, J. Westrich, H. L. Windom. 1982. Coastal upwelling in the Rias Bajas, NW Spain: Contrasting the benthic regimes of the Rias de Arosa and de Muros. J. Mar. Res. 40:701-772. Tenore, K. R., J. Corral, N. Gonzalez & E. Lopez-Jamar. 1985. Effects of intense mussel culture on food chain patterns and production in coastal Galicia, NW Spain. In: Chao, N. L., Kirley Smith, W. (ed.) Proc. Siuec. Brazil, Vol. 1, p. 321-328. Williams, M. J. 1981. Methods for analysis of natural diet in portunid crabs (Crustacea:Decapoda:Portunidae). J. exp. mar. Biol. Ecol. §2:103—113. i = in . Journal of Shellfish Research, Vol. 9, No. 1, 59-62, 1990. GROWTH OF JUVENILE QUEEN CONCH, STROMBUS GIGAS LINNAEUS, 1758 OFF LA PARGUERA, PUERTO RICO RICHARD S. APPELDOORN Department of Marine Sciences University of Puerto Rico Mayagiiez, Puerto Rico 00709 ABSTRACT Queen conch grow in shell length until the onset of sexual maturation. Growth in shell length of juvenile queen conch was studied over a two-year period in a population at 17 m depth off southwest Puerto Rico. Parameters of the von Bertalanffy growth model were determined using two types of data: length-at-age obtained from length-frequency analysis, and growth increments. Resulting parameters were, respectively, L., = 340 mm, k = 0.437 yr~!, t9 = 0.462 yr and L,, = 460 mm, k = 0.250 yr~!", tf) = 0.244 yr. Model parameters from the two types of data are significantly different and are felt to reflect differences in the nature of the respective data. Both models give good fit to the age-length data and can be used for predictive purposes. Predicted ages for the onset of sexual maturation (length = 240 mm) are 3.19 yr and 3.28 yr for the growth-increment and age-length models, respectively. The high values of L., relative to mean adult length are reflective of the fact that growth at maturation does not stop, but only changes in form, with subsequent shell growth resulting in a thickening of the adult’s flared shell-lip. KEY WORDS: INTRODUCTION The queen conch is one of the most prized fishery re- sources in the Caribbean. As a consequence, much atten- tion has been given to its biology, ecology and fisheries potential. Growth has been the subject of various studies. Conch grow in shell length only until maturation. At this time the flared shell-lip, characteristic of the species, is formed. Subsequent shell growth occurs as a progressive thickening of the shell-lip (Appeldoorn 1988). Because of this change in mode of shell growth, most studies have concentrated on growth of juveniles. Randall (1964) pre- sented data on juvenile growth in shell length from tagging studies in St. Johns, U.S. Virgin Islands. Berg (1976) used this and other data to develop von Bertalanffy growth models of juvenile shell growth. Hesse (1976) developed a growth trajectory for conch at Turks and Caicos. Alcolado (1976) used length-frequency analysis and tagging studies to model shell growth in a number of populations from Cuba. Strasdine (1988) reported similar work in Belize. Iversen et al. (1987) calculated von Bertalanffy parameters for conch in the Bahamas. Weil and Laughlin (1984) gave growth trajectories of marked individuals from Los Roques, Venezuela, but did not determine model param- eters. Comparison between areas has shown growth to be markedly variable, both in rate of growth and in time to, and size at, maturation (Tables | and 2). Such variability is due, in large part, to local environmental factors (Alcolado 1976). Growth and population dynamics were studied in a Strombus gigas population off southwest Puerto Rico over a 2-year period (Appeldoorn 1987a, 1988). This paper re- ports on juvenile growth for this population. sy) queen conch, Strombus gigas, growth, Puerto Rico METHODS The study site was located 7 km south of La Parguera, P.R. and consisted of a broad, patchy sand and macroalgal plain with occasional patch reefs. Depth was approximately 17 m, and temperature ranged from 25.5°C to 29.5°C. Sampling ran from August 1983 to August 1985 and was conducted quarterly, generally in the latter half of August, November, February, and May, resulting in a total of nine samples. Attempts were made to locate a minimum of 200 individuals (juveniles and adults) for each quarterly sample. Further details on sampling are given in Appel- doorn (1987a). All individuals, when initially encountered, were tagged using 4.5-cm strips of numbered Dymo label tape tied to the shell spire with nylon line. Upon each encounter, numbers were recorded and individuals measured in situ for shell length (tip of the spire to end of the siphonal canal) to the nearest 1 mm using calipers. Data thus consisted of growth increment information from recaptured individuals and of length-frequency distri- butions for each sampling period. Only data from juveniles are considered here. Growth was modelled using the von Bertalanffy growth function, Le vLe(Ieem 5) () where |, is length (mm) at time t (years), L,. is asymptotic length (mm), and k is the growth coefficient. The param- eter fg is a location parameter and is defined as the hypo- thetical age at which length equals zero assuming that ex- trapolated early growth follows the von Bertalanffy model. This parameter does not convey information on growth rate, but is essential for estimating size at age. 60 APPELDOORN TABLE 1. Reported estimates of von Bertalanffy parameters for growth in shell length of juvenile Strombus gigas from the literature. Lengths are in millimeters, time is in years. Location ES Boca Chica, Belize 268. Tres Cocos, Belize 332. Water Caye, Belize 269. St. John, U.S.V.I. 260.4 St. Croix, U.S.V.I. 241.7 Cabo Cruz, Zone A, Cuba 383.4 Cabo Cruz, Zone B, Cuba 380.6 Diego Perez, Zone A, Cuba 232.7 Diego Perez, Zone B, Cuba 207.6 Cayo Anclitas, Cuba 259.8 Rada Inst. Oceanol., Cuba 334.0 Six Hill Cay, Turks & Caicos 256.0 Berry Islands, Bahamas 300. * From Hesse (1976). For growth increment data, the von Bertalanffy equation is recast in the following form (Fabens 1965): 1) d where i is the growth increment, |, is the length at release, and d is the time between length measurements. Estimation of model parameters was made by nonlinear least-squares regression of Equation 2 using SYSTAT (Wilkinson 1987), which also gives standard errors of the estimates. This method cannot estimate f9 without specific size-at-age in- formation (see Results). The analysis treated multiple re- capture measurements made on individuals as if they were independent. In these cases, the increment used for each recapture was over the time period from initial capture, as opposed to most recent previous recapture. Multiple recap- tures made up 10% of the data. To compensate for small sample sizes, length-frequency data were pooled by season, resulting in four samples. However, only samples from August, November and Feb- i =F Gs e 4) (2) TABLE 2. For Strombus gigas, reported age (years) and mean shell length (mm) at the onset of maturation, defined as when growth in length ceases and the flared-lip begins to form. Location Age Length Source Bermuda 4 —_ Wefer & Killingley 1980 Bahamas 4+ 193 Iversen et al. 1987 Turks & Caicos Islands 2.8 212 Hesse 1976 Cuba 3-4 173-234 Alcolado 1976 Belize 3 204 Strasdine 1988 Puerto Rico 372 240 Appeldoorn 1988 St. John, U.S.V.I. 3 204 Berg 1976, Randall 1964 St. Kitts/Nevis 2.3-2.8 — Wilkins et al. 1987 k to Source —0.05 Strasdine 1988 —0.33 Strasdine 1988 0.290 — Strasdine 1988 0.516 — Berg 1976 0.420 — Berg 1976 —0.05 Alcolado 1976 —0.12 Alcolado 1976 —0.09 Alcolado 1976 —0.09 Alcolado 1976 0.09 Alcolado 1976 0.13 Alcolado 1976 —0.16 Appeldoorn et al. 1987* —0.65 Iversen et al. 1987 ruary were used. Data from May were felt to be unsatisfac- tory because significant partial recruitment and partial mat- uration within year classes affected the frequency distribu- tion (Appeldoorn 1987b). Underlying distributions were determined using Akamine’s (1985) method, a nonlinear, maximum-likelihood technique. In total, 1124 measure- ments were included in the analysis. This yielded a total of 8 estimates of mean length at time. Coupled with resulting standard deviations and number of individuals per year class, these were used to generate 1124 observations of length at time. Corresponding ages were calculated as- suming a birth date of July 1, the approximate midpoint of the spawning season. Resulting length-at-age data were used to calculate von Bertalanffy parameters by nonlinear regression of Equation | using SYSTAT. This procedure allowed all the variability contained in the data to be incor- porated into parameter estimation. RESULTS A total of 187 growth increments from 168 individuals were recorded and used for growth parameter estimation. A total of 1416 measures of shell length of juveniles were taken. Resulting estimates of length-at-age are given in Table 3 and plotted in Figure |. Also plotted are the re- sulting von Bertalanffy curves for the two methods. Param- eter values for the models are given in Table 4. The esti- mate of tf) for tagging data was approximated by substi- tuting the growth-increment derived model parameters (L.., k) into Equation | and fitting the model to the length-at-age data obtained from length-frequency analysis. The non- linear regression was solved for fp using SYSTAT. DISCUSSION No attempt was made to account for seasonality in growth. It is known from other studies (Alcolado 1976, GROWTH OF JUVENILE QUEEN CONCH 61 TABLE 3. Results of length-frequency analysis for juvenile Strombus gigas. Mean lengths and standard deviations are in millimeters, ages are in years, N is the number of individuals, and Date is the midpoint of each sampling period. TABLE 4. Estimates of von Bertalanffy parameters for growth in shell length of juvenile Strombus gigas from La Parguera, Puerto Rico. Lengths are in millimeters, time is in years. Values in parentheses are standard deviations. Value of ¢, for the growth-increment model was estimated by fitting the model to age-length data (see text). Mean Standard Date Age Length Deviation N Source I BE k to XI-21 1.40 117.9 12.36 47 Growth-Increment II-21 1.65 133.5 13.08 108 Data 460 (67.2) 0.250 (0.061) (0.244 (0.025)] VIll-21 1.90 163.7 16.34 231 Age-Length Data 340 (22.9) 0.437 (0.062) 0.462 (0.068) XI-21 2.15 168.2 20.51 131 I-21 2.40 191.2 21.73 134 VIll-21 2.65 212.2 15.64 301 XI-21 2.90 223.4 16.29 112 Using the calculated confidence limits, it is evident that I-21 3.15 230.7 10.89 60 the differences in parameters between the growth-incre- Weil and Laughlin 1984) that juveniles do show seasonal growth, but that seasonal variation in growth is not large. As such, it was felt the use of more complicated models was unwarranted. Nevertheless, a seasonal pattern in growth is indicated in the age-length data. Data points from February (minimum water temperature) are low, while points from August and November (maximum water tem- perature) are high, relative to the predicted growth curve. Both models predict an L,, substantially higher than the mean or largest adult sizes observed from the La Parguera population, 240 mm and 283 mm, respectively. A general rule of thumb is that L,, should be roughly 105% of the maximum observed size (Beverton 1954, in Sundberg 1984, Pauly 1980). The discrepancy, here, lies in the fact that conch continue to grow after maturation, but not in shell length (Appeldoorn 1988). Thus, the extrapolation of the curves toward L,, represents potential growth in length had shell gowth continued in the same manner and had en- ergy not been utilized for reproduction. SHELL LENGTH |mm| 1 2 3 4 AGE lyears] Figure 1. Length at age and von Bertalanffy growth curves for juve- nile Strombus gigas from Puerto Rico. A: growth curve derived from growth-increment data, B: growth curve derived from age-length data. ment and age-length models are statistically different. This is felt to be due to differences in the nature of the data used for the two models, particularly at lengths near maturation. Being at one extreme of the data, these points exert greater influence during regression than points within the midrange of data. An effort was made to eliminate data from the length-frequency analysis obviously subject to partial year- class maturation (Appeldoorn 1987b), and the largest year- class mode used in the analysis was at 230 mm, 10 mm below mean adult length. However, all data for tagged ju- veniles were used; growth-increment data included juvenile growth up to 254 mm. Furthermore, Alcolado (1976) pre- sented data indicating that populations of large individuals at maturation are large because they grow faster, and not because they grow for a longer period of time. Thus, on the one hand, growth-increment data tend to show continued growth near maturation, resulting in a higher L,, and lower k, while on the other hand age-length data did not show this, and, in addition, they may have been affected in an opposite manner by the fact that the last (largest length) data point came from February, and its mean length may be suppressed due to low winter growth. This would tend to lower L,, and increase k. In the most practical sense, both models give a good fit to the known age-at-length data. Thus, they can be used equally well for predictive purposes, either for growth rate or length-at-age, within the range of data. Predicted ages at the mean adult size of 240 mm were similar, at 3.19 yr and 3.28 yr for the growth-increment and age-length models, respectively. Note should be taken of the magnitude of f. This parameter is necessary if the growth-increment model is to be used to predict length-at-age. Some previous studies have not been able to incorporate fg into their pre- dictions (e.g. Berg 1976, Wood and Olsen 1983). The fact that fp here is positive indicates the presence of an early inflection point in growth. Relationships useful for converting length to shell weight, wet tissue weight, and wet meat weight (after re- moval of the visceral mass) for juveniles in the La Parguera population were given by Appeldoorn (1988). Using these, 62 APPELDOORN the von Bertalanffy models can be further used to investi- gate potential fisheries yield. However, it is clear that ex- trapolated L,, values should not be used, via conversion equations, to generate values of asymptotic weight, W.., for yield-per-recruit calculations as has been done in the past (Wood and Olsen 1983, Berg and Olsen, 1989). At a min- imum, such an extrapolation would not account for the lost proportion of energy channelled to reproduction. It is also possible that decreases in adult shell-volume (Randall, 1964) could adversely affect adult tissue growth. A more appropriate approach would be to model adult growth di- rectly (Appeldoorn 1988). ACKNOWLEDGEMENTS I wish to thank all those who aided in data collection, particularly D. L. Ballantine, A. T. Bardales, J. Colley, G. Gonzalez, I. M. Sanders, and Z. A. Torres. Bonnie Bower-Dennis drew the figure. This work was supported by the U.S. National Marine Fisheries Service (NA81-GA- C-00015). LITERATURE CITED Akamine, T. 1985. Consideration of the BASIC programs to analyse the polymodal frequency distribution into normal distributions. Bull. Jpn. Sea Reg. Fish. Res. Lab. 35:129-160. Appeldoorn, R. S. 1987a. Assessment of mortality in an offshore popula- tion of queen conch, Strombus gigas L., in southwest Puerto Rico. U.S. Fish. Bull. 85:797—804. Appeldoom, R. S. 1987b. Practical considerations in the assessment of queen conch fisheries and population dynamics. Proc. Gulf. Carib. Fish. Inst. 38:307—324. Appeldoom, R. S. 1988. Age determination, growth, mortality and age of first reproduction in adult queen conch, Strombus gigas L., off Puerto Rico. Fish. Res. 6:363—378. Appeldoorn, R. S., G. D. Dennis & O. Monterrosa Lopez. 1987. Review of shared demersal resources of Puerto Rico and the Lesser Antilles region. FAO Fish. Tech. Rept. 383:36—106. Berg, C.J., Jr. 1976. Growth of the queen conch Strombus gigas, with a discussion of the practicality of its mariculture. Mar. Biol. 34:191— 199. Berg, C. J., Jr., & D. A. Olsen. 1989. Conservation and management of queen conch (Strombus gigas) fisheries in the Caribbean. Jn: J. F. Caddy (ed.). The scientific basis of shellfish management. John Wiley & Sons, New York. 752 pp. Fabens, A. J. 1965. Properties and fitting of the von Bertalanffy growth curve. Growth 29:265—289. Hesse, K. O. 1976. An ecological study of the queen conch, Strombus gigas. M.S. Thesis, Univ. Connecticut, Storrs, Conn,. 107 pp. Iversen, E. S., E. S. Rutherford, S. P. Bannerot & D. E. Jory. 1987. Biological data on Berry Islands (Bahamas) queen conchs, Strombus gigas, with mariculture and fisheries management implications. U.S. Fish Bull. 85:299—310. Pauly, D. 1980. A new methodology for rapidly acquiring basic informa- tion on tropical fish stocks: growth, mortality and stock recruitment relationships, p. 154—172. Jn: S. B. Saila & P. Roedel (eds.). Stock assessment for tropical small-scale fisheries. /nternatl. Cent. Mar. Re- source Devel., Univ. Rhode Island, Kingston, R.I. Randall, J. 1964. Contributions to the biology of the queen conch Strombus gigas. Bull. Mar. Sci. 14:246—295. Sundberg, P. 1984. A Monte Carlo study of three models for estimating the parameters of the von Bertalanffy growth equation. J. Cons. Int. Explor. Mer. 41:248—258. Strasdine, S. A. 1988. The queen conch fishery of Belize: an assessment of the resource, harvest sector and management. M.S. Thesis. Univ. British Columbia, Vancouver, B.C. 216 pp. Wefer, G., & J. S. Killingley. 1980. Growth histories of strombid snails from Bermuda recorded in their 0-18 and C-13 profiles. Mar. Biol. 60:129—135. Weil M., E., & R. Laughlin G. 1984. Biology, population dynamics, and reproduction of the queen conch Strombus gigas Linné in the Archi- pielago de los Roques National Park. J. Shellfish Res. 4:45—62. Wilkins, R. M., M. H. Goodwin, & D. M. Reed. 1987. Research applied to conch resource management in St. Kitts/Nevis. Proc. Gulf Carib. Fish. Inst. 38:370—375. Wilkinson, L. 1987. SYSTAT: the system for statistics. SYSTAT, Inc., Evanston, Ill. Wood, R., & D. A. Olsen. 1983. Application of biological knowledge to the management of the Virgin Islands conch fishery. Proc. Gulf Carib. Fish. Inst. 35:112—121. Journal of Shellfish Research, Vol. 9, No. 1, 63—65, 1990. SEASONAL CHANGES IN OXYGEN CONSUMPTION OF THE WEST INDIAN FIGHTING CONCH, STROMBUS PUGILIS LINNAEUS, 1758 ILSE M. SANDERS* Department of Marine Sciences University of Puerto Rico Mayaguez, Puerto Rico 00708 ABSTRACT The relation between body weight, temperature and respiratory rates of Strombus pugilis L. from La Parquera, Puerto Rico, was investigated. Respiratory rates were measured at ambient temperatures over the course of one year. Total specific oxygen consumption ranged from 0.19 to 0.35 mgQ,j shell free dry weight (g)~' hr~! and appeared to follow monthly temperature changes (26.0 to 29.4°C). Multiple regressions of oxygen uptake on temperature and weight were significant for temperature by sex and for all individuals, and for weight for all individuals. The Qj for a conch of average weight (6.27 g SFDW) between 26 to 29°C was 2.28, but not significantly different from 2.0. Seasonal changes in weight-specific rate of oxygen consumption appear to be a response to seasonal changes in temperature. KEY WORDS: INTRODUCTION Respiratory rates have been recorded for a large number of marine molluscs of different trophic levels and at dif- ferent temperatures (Bayne and Newell, 1983). Most of these studies have been of temperate species. In the tropics, respiration rates have been measured mostly for intertidal gastropods (Hughes, 1971a,b; Brown and DaSilva, 1979). In temperate areas, temperature-dependency of respiratory rates has been reported for subtidal gastropods in areas of narrow temperature fluctuations, whereas some degree of temperature independence has been found in widely fluc- tuating environments (Hughes, 1986). The West Indian fighting conch, Strombus pugilis L., is a tropical subtidal gastropod found on sandy-mud areas at depths of two to 20 m. Whether S. pugilis has a respiratory rate dependent upon the annual change in ambient tempera- ture, is presently unknown. Rate of oxygen uptake for S. pugilis living in Barbados was measured only over a narrow range in weight and at one ambient temperature, 27°C (Sander and Moore, 1978). The present study presents respiration rates of adult S. pugilis measured at ambient temperatures over the course of one year. Several in situ oxygen uptake experiments were performed at the sampling site in order to compare measurements with those taken at the laboratory. The data obtained were used to de- termine the relationships of respiratory rate to body weight and temperature. These relationships were used to further investigate the possibility of regulation of respiratory rate of S. pugilis in response to seasonal temperature fluctua- tions. METHODS Adult Strombus pugilis were collected monthly (De- cember 1984 to October 1985, except May 1985) from a shallow mangrove inlet, 10 km southwest of La Parguera, * Present address: Department of Biology, Interamerican University, San German, Puerto Rico, 00753 63 conch, respiratory rates, temperature, Strombus pugilis L. Puerto Rico. This inlet has a depth of 1—3 m, and the bottom consists of sandy mud with patches of macroalgae. Water temperature and oxygen concentration were recorded on each collection. Conch were transported to the laboratory and held in flowing seawater at ambient temperature for 24 hr, without feeding, prior to initiating experiments. Shell length and lip-thickness were measured and sex determined for an average of 12 conch/experiment (Table 1). Following com- pletion of experiments, each animal was removed from its shell and its shell-free dry weight determined, after drying to constant weight at 60°C. Ambient experimental temperatures were maintained by means of a flow-through seawater bath. Three experimental airtight, rectangular plexiglass chambers were immersed in the bath and maintained at ambient temperature. Conch were placed singly in the three-liter chambers filled with aerated Millipore-filtered (0.45 mw) seawater. A magnetic stirrer placed beneath the double floor of each chamber caused continuous water movement. Oxygen uptake was measured by using a YSI oxygen polarographic, tempera- ture-compensated, probe held within the chamber during the one to two hours of the experiment. Oxygen concentra- tions ranged from 4—6 ppm. The probe was calibrated by using a modified Winkler technique (Parsons et al., 1984). Oxygen concentration and temperature within each chamber were recorded at approximately 30 minute in- tervals; the former was not allowed to drop below four ppm. Controls, chambers without conch added, were run to determine the rate of oxygen depletion due to the electrode. Temperatures fluctuated according to the ambient daily cycle, but changes did not exceed 0.3°C within an indi- vidual experimental chamber. Changes in oxygen concen- tration per time during the course of each experiment were averaged, and taken as the rate of oxygen consumption by the conch. Oxygen uptake in situ was measured once on each of three days during Cctober, 1986, using a 7.2-liter belljar 64 SANDERS TABLE 1. Size and respiration rates of female and male Strombus pugilis between January and October, 1985. Mean and S.D. of length in mm; mean and S.D. of specific oxygen consumption in mg O, SFDW (g)~! hr-!; N = number of conch. Females Males Month N Size Respiration N Size Respiration January 4 84.4 + 1.1 0.25 + 0.06 5 81.1 + 1.4 0.25 + 0.06 February 6 89.7 + 6.6 0.18 + 0.03 1 80.2 0.28 March 10 86.2 + 8.5 0.21 + 0.06 4 TAN = 4:2 0.28 + 0.12 April 10 87.2 + 4.2 0.26 + 0.09 1 83.0 0.32 June 8 84.3 + 4.6 0.29 + 0.07 6 81°35 -= 9228 0.37 + 0.10 July 9 8722) = 3:5 0.25 + 0.06 4 83.6 + 4.3 0.26 + 0.04 August 10 85.9 + 3.1 0.26 + 0.07 6 78.4 + 3.9 0.43 + 0.07 September 10 87.0 + 4.2 0.30 + 0.07 3 VP se Ne! 0.41 + 0.14 October 7 86.3 + 3.4 0.27 + 0.07 7 81.9 + 3.4 0.34 + 0.09 with a magnetic stirrer and polarographic probe. Three conch, taken from the area adjacent to the experimental ap- paratus, were placed simultaneously under the belljar during the two to three hours of each experiment. For con- trols, oxygen uptake was measured at the sandy-mud sur- face under the bell-jar, with the same instrument and by the same procedure as in the laboratory. These were run for one hour prior to the beginning of each experiment. RESULTS During the year, specific total oxygen consumption ranged from 0.19 to 0.35 mgO, shell free dry weight (g)~! hr! (S.D. 0.04 to 0.10), and temperature ranged from 26.0 to 29.4°C (Figure 1). Because of difficulty in ob- taining conch during November, it was impossible to com- plete an annual cycle. Multiple linear regressions of the log of oxygen uptake per individual on temperature and the log of weight were run for 122 conch, both pooled and by sex. Sex was not determined for nine conch in the total regression. Although the total regression for oxygen uptake on weight was signif- icant, those for males and females separately were not (Table 2). All multiple regressions of oxygen uptake on temperature were significant. Variability in oxygen con- sumption is only partially explained by weight and temper- ature (multiple r? = 0.400), although rate of oxygen con- sumption does seem to follow the monthly trend in temper- ature fluctuations (Figure 1). The weight exponent (b) of less than one (0.229; S.E. 0.090) does indicate that weight- specific oxygen consumption decreases as body size in- creases (since the specific rate of oxygen consumption is proportional to body weight~°; Hughes, 1986). At mean temperature of 27.8°C a one gram conch would consume oxygen at a rate of 1.08 mg hr~! (0.75 ml hr~!). In situ mean oxygen uptake at 28.8°C (S.D. 0.7) was 1.62 mgO, ind! hr~! (S.D. 0.56) for the nine conch whose mean siphonal length was 90.08 mm (S.D. 5.58). This siphonal shell length was converted to shell-free dry weight by the regression: log (shell free dry weight g) = —4.749 + 2.881 (log length mm), (n = 122; 1? = 0.523), and was equal to 7.61 g. Laboratory mean oxygen uptake for a conch of equivalent weight at 28.8°C was 1.78 mgO, ind~! hr~! (using the regression in Table 2). The in situ rate is well within the confidence limits of the regression prediction. DISCUSSION Seasonal fluctuations in weight-specific rate of oxygen consumption are apparently a response to seasonal changes in temperature (Figure 1). However, the large unexplained variability in the rates of oxygen consumption within each month suggests that the coupling between the two is not strong. Factors that could account for this variability are as follows: 1) variability in the condition of the conch, such as weight and gamete condition, 2) differences in the level of activity while in the respiratory chamber, 3) differences in response to temperature versus response to patterns of go- nadal growth (Parry, 1978), and 4) physiological acclima- tion, allowing S. pugilis to maintain relative constant respi- ratory rate over seasonal changes in temperature. The results reported here were assumed to be of routine oxygen uptake, representing the energy used in non-loco- motory activity plus maintenance. After an initial period of TABLE 2. Parameters of the multiple regression, log R = log a + b, log W + b,T, for respiration of females, males and total sample. R, oxygen uptake (mgO, ind~! hr~!); W, shell-free dry body weight (g); T, temperature; r?, multiple regression coefficient; a, a constant; b,, the weight exponent; b,, the temperature exponent; N, number of conch; NS, not significant; *P, <0.05; **P, <.01. Weight Temperature Component a b, b, r? N Females —1.081 0.253 NS 0.396** 0.411 76 Males —1.467 0.435 NS 0.049** 0.492 37 Total —0.951 0.229% 2.036** 0.400 122 SEASONAL CHANGES IN OXYGEN CONSUMPTION OF STROMBUS PUGILIS LINNAEUS 65 db 24 ] / © lies . 4 = 28 | i, © o 27 o 4 E y | o 264 y 25+ T T T T T T aT 7 T T 1 oxygen consumption mgO, SFDW (g)"' hr-! ero Rua e Ss s1Jt—t} 4 | $ 0154 O14 0.05 4 o-+ T el T =T 1 DEC JAN FEB MAR APR MAY JUN JUL AUG SEP OCT 1984 1985 months Figure 1. Seasonal changes in specific oxygen consumption rates of Strombus pugilis with environmental temperatures, between De- cember, 1984 and October, 1985. Mean and S.D. of oxygen consump- tion in mgO, SFDW(g)~! hr—!; mean and S.D. monthly temperature in °C; N = 121. movement lasting less than eight minutes, conch remained quiescent, and therefore oxygen uptake could have in- creased only slightly. Qo calculated from the regression of log respiration on log SFDW and temperature (Table 1) for a conch of average weight (6.27 g SFDW) at 26 to 29°C, is 2.28. This value is not significantly different from a Qj, of 2.0, sug- gesting that S. pugilis increases oxygen uptake as a passive response to increase in temperature and, therefore, is not regulating its respiratory rate. Most gastropods show posi- tive temperature dependence (Q)9 of 2—3) of metabolic rate (Hughes, 1986). However, some species subject to wide fluctuations in temperature, have metabolic rates indepen- dent of temperature. Bullia digitalis (Dillwyn), living in the temperate zone and subject to a widely fluctuating tempera- ture regime, showed metabolic independence of tempera- ture, while the related B. rhodostoma Reeve living in a moderately fluctuating environment, had a metabolic rate dependent upon temperature (Brown and DaSilva, 1984). The independence of temperature was associated with the conservation of energy during periods of starvation. Gener- ally, arctic aquatic molluscs show considerable acclimation when compared to tropical aquatic species (Vahl, 1978). Based on intrageneric comparisons of tropical and tem- perate species, Bayne and Newell (1983) concluded that the higher respiration showed by tropical species suggested non-acclimation. Many marine gastropods have a rate of oxygen con- sumption significantly correlated with body weight (W,), but for some species b can range from 0.03 to unity (Hughes, 1986: Table 4). Rate of oxygen consumption of adult S. pugilis does appear to be related to weight, but not strongly. Considering the large degree of variability that characterizes this relationship in other species, this result is not surprising. It appears that oxygen consumption of adult S. pugilis follows seasonal changes in temperature, but that in addition, its physiological state, related to trophic or re- productive condition, probably also influences its respira- tory rate. ACKNOWLEDGMENTS The author thanks Dr. Richard S. Appeldoorn for his guidance during the course of this research and for criti- cally reviewing the manuscript, and Dr. David Ballantine for providing laboratory space. Thanks are also due to Ivan Lopez and Victor Rosado for all their help in collecting conch. REFERENCES CITED Bayne, B. L. & R. C. Newell. 1983. Physiological energetics of marine molluscs. In: A. S. M. Saleuddin and K. M. Wilbur, (eds.), The Mollusca, Vol. 4, Academic Press, New York, pp. 407-523. Brown, A. C. & F. M. DaSilva. 1979. The effects of temperature on oxygen consumption in Bullia digitalis Meuschen (Gastropoda, Nas- saridae). Comp. Biochem. Physiol. 62A:573—576. Brown, A. C. & F. M. DaSilva. 1984. Effects of temperature on oxygen consumption in two closely-related whelks from different temperature regimes. J. Exp. Mar. Biol. Ecol. 84:145—153. Hughes, R. N. 1971la. Ecological energetics of the keyhole limpet, Fis- surella barbadensis Gmelin. J. Exp. Mar. Biol. Ecol. 6:167—178. Hughes, R. N. 1971b. Ecological energetics of Nerita (Archeogastro- poda, Neritacea) populations on Barbados, West Indies. Mar. Biol. 11:12-24. Hughes, R. N. 1986. A functional biology of marine gastropods. Johns Hopkins University Press, Maryland, 245 pp. Parry, G. D. 1978. Effects of growth and temperature acclimation on met- abolic rate in the limpet, Cellana tramoserica (Gastropoda: Patel- lidae). J. Anim. Ecol. 47:351—368. Parsons, T. R., Y. Maita & C. M. Lalli, 1984. A manual of chemical and biological methods for seawater analysis. Pergamon Press, Oxford, England. 173 pp. Sander, F. & E. A. Moore. 1978. Comparative respiration in the gas- tropods Murex pomum and Strombus pugilis at different temperatures and salinities. Comp. Biochem. Physiol. 60A:99- 105. Vahl, O. 1978. Seasonal changes in oxygen consumption of the iceland scallop (Chlamys islandica (O. F. Muller)) from 70°N. Ophelia. 17:143-154. Journal of Shellfish Research, Vol. 9, No. 1, 67—73, 1990. A COMPARISON OF TWO INDUCERS, KCI AND LAURENCIA EXTRACTS, AND TECHNIQUES FOR THE COMMERCIAL SCALE INDUCTION OF METAMORPHOSIS IN QUEEN CONCH STROMBUS GIGAS LINNAEUS, 1758 LARVAE MEGAN DAVIS!3, WILLIAM D. HEYMAN?, WILBERT HARVEY}, CHRIS A. WITHSTANDLEY! 1Caicos Conch Farm 7600 S.W. 87th Avenue Miami, FL 33173 2P.O. Box 325 Garrett Park, MD 20896 ABSTRACT Intensive Queen Conch Strombus gigas (L.) mariculture requires a routine, inexpensive and effective technique for the induction of metamorphosis of veligers. Extract of the red marine macroalga Laurencia poitei has been successfully used for the past five years as a metamorphic inducer at the Caicos Conch Farm, Turks and Caicos Islands, but there are inherent variations and high costs involved with using the extract. The goal of this research is to determine if potassium chloride (KCl) can replace Laurencia extracts as a metamorphic inducer at the commercial scale. Results indicate that Laurencia extracts produced an average 80.6 + 11.1% recovery (minimum % metamorphosis) of juveniles from induced larvae, with an average growth rate of 0.17 + 0.02 mm/day, 10—16 days after metamorphosis. Similarly, 6 h, 8h, 10 h, and 16 h exposures to seawater with KCI concentration increased by 15 mM over ambient, produced 72.7 + 14.5% recovery of juveniles from induced larvae, with an average growth rate of 0.18 + 0.04 mm/day post induction. Since KCl is more convenient and less expensive to use than Laurencia extracts, and does not appear to adversely effect conch postlarvae, KCI could be adopted as the metamorphic inducer on the commercial scale. KEY WORDS: INTRODUCTION The Caicos Conch Farm, Turks and Caicos Islands is developing a commercially viable way to culture Queen Conch Strombus gigas (Linne). The conch egg laying season in the Turks and Caicos Islands is from March to October (Davis et al. 1984). Since it is not yet possible to induce spawning outside of this breeding season, the hatchery must be efficient during the natural spawning pe- riod. The intensive hatchery at the Conch Farm (Davis et al. 1986) is capable of producing over 100,000 competent veligers during each week of the breeding season. The hatchery production far exceeds the production capacity of the remainder of the facility, so only a portion of these larvae are then carried through metamorphosis. The com- mercial scale induction of metamorphosis in Queen Conch must be efficient and cost effective. While the Conch Farm is commercially oriented, this paper describes the specific research behind the development of our large-scale meta- morphosis system. At the onset of metamorphic competence, many inverte- brate larvae may encounter environmental cues indicating suitable juvenile habitat (Thorson 1950, Crisp 1974, and Hadfield 1978). For natural metamorphosis to occur, a competent larva receives a cue through a receptor and can thus be induced to settle and metamorphose (Scheltema 1974, Burke 1983). The affected receptor presumably ini- tiates a series of neuronal and/or hormonal events that Author to whom all correspondence should be addressed 67 conch, Strombus gigas, metamorphosis, Laurencia extract, KCl eventually culminate in the morphological and physiolog- ical changes associated with metamorphosis (Hadfield 1978, 1984, Morse et al. 1979, Highnam 1981, Burke 1983, and Trapido-Rosenthal and Morse 1986). The red marine macroalga Laurencia poitei (La- mouroux) contains phycoerythrins and related protein con- jugants that initiate the onset of metamorphosis in Strombus (Siddall, 1983). L. poitei is found on reefs and rocks in shallow waters from Bermuda to Brazil (Taylor 1972). Ex- tracts of L. poitei have been used successfully as a meta- morphic inducer for the past five years at the Caicos Conch Farm (Davis and Dalton 1987). The time consuming col- lection and processing of the algal extract makes the Lau- rencia process very labor intensive, as is often the case when using partially purified natural extracts for the induc- tion of metamorphosis in invertebrate larvae (Veitch and Hidu 1971, Hadfield 1977, and Heslinga 1981). Because of the high costs and intensive labor involved with inducing metamorphosis of conch larvae with Laurencia extracts, other means of inducing metamorphosis were investigated. Metamorphosis has been induced with a wide variety of neuroactive compounds, consistent with neuronal involve- ment along the hypothetical pathway between receptor and the process of metamorphosis (Hadfield 1978, Morse et al. 1979, and Coon et al. 1985). Also consistent with nervous involvement along the pathway, molluscan metamorphosis has been induced with micromolar additions of potassium chloride (KCI) (Rumrill and Cameron 1983, Baloun and Morse 1984, Nell and Holliday 1986, Yool et al. 1986, Pechenik and Heyman 1987, Heyman et al. 1989). KCl 68 DAVIS ET AL. may act by depolarizing a membrane, somewhere along the pathway, and consequently induce metamorphosis without the natural reception of a metamorphic cue (Baloun and Morse 1984). The goal of this study is to determine if po- tassium chloride could replace Laurencia extracts as the metamorphic inducer for Queen Conch larvae. To be an effective replacement of Laurencia extracts, KCl should produce an equal number of healthy postlarvae from in- duced veligers, and be more cost effective on the commer- cial scale. MATERIALS AND METHODS Strombus gigas veligers were cultured to metamorphic competence in the intensive hatchery system described in Davis et al. (1986). Veligers were cultured to an average final density of 30 + 10 larvae/] and were competent for metamorphosis in 21 + 2 days at a mean shell length of 1.2 + 0.1 mm. Competency in conch veligers is recog- nized morphologically when the eyes have migrated out- ward, the tentacles are the same length, the foot has dark green pigmentation, and the buccal mass has developed (Brownell et al. 1977). Two metamorphic inducers, Lau- rencia poitei and potassium chloride (KCl) were compared with experiments conducted between July 19 and October 24, 1988 at the Caicos Conch Farm, Turks and Caicos Is- lands. Using snorkel equipment, L. poitei were hand collected from shallow, sandy-grass flats on the windward side of the Caicos Islands in the Turks and Caicos, and transported to shore in mesh bags. Onshore, the algae were rinsed and sorted to remove coral pieces, sponges, and excess sand. In a Waring industrial blender, 500 g of Laurencia and 250 ml of seawater (2 g:1 ml) were blended for 2 min. This solu- tion was frozen for a minimum of 2 days to lyse the cells, in order to release the phycoerythrins. The frozen solution was then allowed to thaw overnight. The algae solution was then filtered through a 75 ppm polyethylene screen and the resulting extract was refrozen in 10 | containers until needed. Before any given Laurencia extract was used at the large scale, the appropriate concentration of the extract was de- termined with a small scale dosage test. Dosage tests in- clude three concentrations of the extract; 7, 10, and 15 ml extract/] seawater, with 25 larvae in each. After 3.5 hours of exposure, the veligers were observed for metamorphosis with a Swift Stereo 80 Wide Field Microscope (40 x ). Metamorphosis in conch is recognized when the velar lobes are lost, the conch are crawling with propodium, and are searching for food with their proboscis. Percent metamor- phosis was calculated for each dosage. A minimum of 60% metamorphosis was considered effective to select a given dosage for use at the large scale. Similar small scale tests were performed on each hatchery tank, to determine when the entire culture was ready for metamorphosis. Potassium chloride (KCl) (AR granular Mallinckrodt) was used to raise the natural KCI concentration in seawater (9 mM) (Cavanaugh 1956) to create the artificial metamor- phic inducer. Preliminary experiments revealed that a solu- tion of seawater with the KCI concentration raised by 15 mM produced the highest number of metamorphosed conch; therefore, 15 mM was utilized for all subsequent, large scale experiments. Using seawater with the KCl con- centration increased by 15 mM, a series of large scale ex- periments were conducted to determine the most effective exposure duration. For all metamorphosis experiments, tanks were clorox- cleaned and set up one hour before use. The fiberglass tanks each contain eight polyethylene molded trays (60 cm?) that fit tightly in two rows and are supported by the tank edge and a central wooden brace (Davis and Dalton 1987, in press). Polyethylene screen (275 2m) was rubber cemented onto the open bottom of each tray. A standpipe was placed in the central drain which limited the tank’s capacity to 200 |. Each replicate of the experiment con- sisted of one control tank, with Laurencia extract, and ei- ther one or two experimental tanks with 15 mM added KCl as the inducer. The Laurencia extract was stirred well and then introduced to the tank to make a final extract concen- tration of 7—12.5 ml/L (depending on the dosage test per- formed earlier). KCl was weighed, then dissolved in a | | pitcher with seawater, and poured into another tank to raise the final KCl concentration by 15 mM. The inducing agent was always added while the tanks were being filled with sand-filtered, ambient temperature seawater. After the in- ducing agents were added, the screen trays were put in place and the tanks were filled to capacity. Veligers were siphoned out of the hatchery tanks into a screen sieve, hosed into a 20 | bucket, and suspended in 18 1 of seawater. Initial larval density in the bucket was esti- mated by counting subsamples of known volumes. The ve- ligers were kept in suspension with gentle aeration, and with a hand stirring disk. Ten aliquot samples were taken successively, with a 60 ml beaker. The volume of each sample was measured in a 100 ml graduated cylinder, while larvae were counted with a hand counter, and subsequently returned to the bucket. Volumes and counts were recorded to determine larval density in veligers/ml. The highest and lowest larval density estimates were discarded. The re- maining eight were averaged to determine veligers/ml and multiplied by 1,800 mls to determine the total number of veligers in the bucket. Using a 500 ml pitcher, to collect the appropriate volume of water and suspended larvae, ap- proximately 1,600 veligers were distributed into each tray. For each experiment, veligers from a single egg mass were randomly divided between an experimental tank with KCl, and a control tank with Laurencia. The exposure time for Laurencia extract was 5 h, while the exposure times tested for KCl were 6, 8, 10, and 16 hours. At the end of these exposure times, the conch were examined for the onset of metamorphosis and were fed a COMMERCIAL SCALE METAMORPHOSIS IN CONCH 69 flocculated diatom, Chaetoceros gracile. Then a flow of water was introduced to the tank via holes in a PVC pipe, that created a vortex in each tray and downwelled contin- uously at a rate of 0.5 I/min/tray. Day of metamorphosis completion was considered Day 0. The newly metamor- phosed conch were maintained on the screen trays for 10 to 16 days. The screens were flushed daily by first draining and then spraying the tank with a high pressure spray wand, pushing uneaten food and waste down the drain. The tanks were then refilled with seawater and the juveniles fed the daily ration of food. In normal operations, postlarvae are first sorted and re- distributed 10—16 days after metamorphosis, so these ex- periments were terminated at this first tank exchange. All conch were hosed into an 875 ym sorting screen, and the live conch sorted from dead by gently agitating the sorting screen, allowing the smaller, dead animals to go through. The number of live juveniles remaining on the screen was recorded. To compare the effectiveness of the two meta- morphic inducers, and to determine the best exposure time for KCl, the average growth rate and percent recovery of juveniles were calculated from each replicate of each treat- ment. To calculate growth rate at the completion of each ex- periment, 20 juveniles from each tank were randomly se- lected and their shell lengths measured. Shell length was measured from siphonal canal to apex, with a compound dissecting microscope (40 x ), equipped with an ocular mi- crometer. Average daily growth rate was calculated by subtracting the initial shell length (1.2 + 0.1 mm) from the final shell length, and dividing by the number of days after metamorphosis was induced. Percent recovery was defined as the proportion of juve- niles remaining (large enough to be caught on the sorting screen, 10 to 16 days after metamorphosis) to the total number of larvae induced. Recovery is therefore a min- imum estimate of the percent of larvae which successfully completed metamorphosis. Recovery, however, also ac- counts for some postlarval mortality. Recovery was mea- sured with the aid of a predetermined table that presents the relationship between shell length (mm) and volume (ml) for cultured juvenile conch (Heyman unpublished data 1988). The table can be used to determine the number of conch/ml at any given average shell length (2—16 mm) with an accu- racy of within 10%. Recovery is determined by measuring the total volume of the juveniles in a 100 ml graduated cyl- inder. Total volume is multiplied by the number of conch/ ml to determine final count. Percent recovery is calculated by dividing the final, juvenile count by the initial, larval count. The effect of exposure times of 6, 8, 10, and 16 h to KCl on percent recovery and growth rate, were analyzed with a one-way ANOVA (Sokal and Rohlf 1969). To sat- isfy the assumptions of homogeneity of variance, the ANOVAs were also calculated with transformed data. The log (x + 1) transformation was used to test the effect of exposure time on growth rate while the arcsine transforma- tion was used to test the effect of exposure time on percent recovery. The effect of inducing agent (KCI and Lau- rencia) on growth and percent recovery was assessed using a two-tailed t-test (Snedecor and Cochran, 1980). Finally, the cost of using the two inducers on a commercial scale was closely examined. RESULTS The first goal of this experiment was to determine the optimum duration of exposure to 15 mM KCl, by testing the parameters of postlarval growth rate and percent re- covery of juveniles. Juvenile recovery varied widely within and among treatments. For instance within the ten repli- cates of the 16 h exposure to KCl, recovery varied from 52 to 95% (Table 1) and may have been affected by external factors. Variations inherent in the large scale experimental design may have included larval rearing temperature, larval and juvenile counting techniques, water quality variations, and seasonal differences. These variations may have masked the differences caused by the different exposure pe- riods. Nonetheless, of the durations tested (6, 8, 10, and 16 h) there was no significant difference (one-way ANOVA; p > 0.05) in the percent recovery or growth rate of the juve- niles, 10 to 16 days after metamorphosis (Table 2). This holds true when ANOVAs were performed with original and transformed data. The 6 h and 8 h exposures used in this ANOVA only had two replicates so the data from each of these treatments were kept separate, but were pooled rather than averaged. Since no optimum KCI exposure duration was deter- mined, exposure times will be selected that best coincide with the eight hour work day routine of the Caicos Conch Farm. Six hours is convenient within a single day, and 16h allows metamorphosis to be set up in the late afternoon, and finish the following morning. During the 1989 season, the effects of 6 and 16 h exposures will be thoroughly in- vestigated. The main objective of this experiment was to determine if there was any significant difference in the juvenile re- covery (minimum estimate of percent metamorphosis) or growth of Strombus gigas when using Laurencia extract or 15 mM KCl as the metamorphosis inducing agent for com- petent veligers. Since it was determined previously that the exposure periods tested for KCl did not produce signifi- cantly different results, these data were combined for com- parison to the combined results of Laurencia extracts. To further support combining these data, the effects of external variations were examined. There were no significant corre- lations (r? < 0.15) between veliger age, temperature during metamorphosis, or time on screen trays, and growth rate or percent recovery (Table 3). These experimental variations, therefore, do not appear to affect the comparison between the effects of the two inducers. Laurencia extracts pro- 70 DAVIS ET AL. TABLE 1. The effect of 15 mM KCI exposure duration on the growth rate (mm/day) and percent recovery of induced Strombus gigas postlarvae. Temperature In Veliger Initial Metamorphosis Age Veliger Date (°C) (Days) Count Exposure Time 6 h 08/03 27.8 24 6,750 08/03 27.8 24 10,800 Exposure Time 8 h 08/03 27.8 24 10,800 08/10 28.5 24 12,928 Exposure Time 10 h 07/19 28.0 23 12,800 08/27 27.3 20 12,800 08/27 27.3 20 12,800 Exposure Time 16 h 09/03 28.3 20 12,800 09/12 27.8 22 12,800 09/12 27.8 22 12,800 10/06 28.8 20 9,000 10/08 28.8 19 16,000 10/11 295 19 12,800 10/14 28.3 18 14,560 10/14 28.0 22 21,240 10/18 27.0 22 10,800 10/24 28.8 22 15,000 duced a mean juvenile recovery of 80.6 + 11.1% and a mean growth rate of 0.17 + 0.02 mm/day; KCl produced a mean recovery of 72.7 + 14.5% and a mean growth rate of 0.18 + 0.04 mm/day (Table 4). There was no significant difference (t-test; p > 0.01) between the percent recovery or early juvenile growth rate of Strombus juveniles induced with Laurencia extracts or 15 mM KCl. Juveniles exposed to the different inducers showed similar growth and sur- vival up to 35 mm when they were moved to sea-based culture, and data collection became impossible. Time on Final Final Screen Shell Growth Juvenile Percent Trays Length Rate Count Recovery (Days) (mm) (mm/day) 4,792 71% 13 4.1 0.22 7,720 71% 13 4.1 0.22 8,649 80% 14 3.9 0.19 6,890 53% 14 4.1 0.21 10,755 84% 13 35 0.18 9,300 73% 15 3.1 0.13 10,500 82% 12 2.9 0.14 9,783 76% 11 3.2 0.18 12,198 95% 14 3.6 0.17 10,914 85% 13 3e7/ 0.19 5,064 56% 11 4.1 0.26 10,276 64% 10 33 0.21 10,284 80% 15 3.5 0.15 11,133 76% 16 S15) 0.14 15,246 72% 15 S15) 0.15 5,720 53% 13 4.0 0.22 7,803 52% 16 4.1 0.18 DISCUSSION During the normal metamorphosis of many invertebrate species, the physical changes of metamorphosis are usually preceded by a period of substrate examination and selection (Crisp 1974, Coon et al. 1985). When Crepidula larvae are induced to metamorphose with KCl, the larvae metamor- phose without searching the substrate (Pechenik and Heyman 1987). Strombus gigas larvae also do not exhibit the normal behavioral pattern of settlkement when induced to metamorphose with KCl. In spite of this behavioral dif- TABLE 2. The effect of exposure duration on the average growth rate (mm/day) and average percent recovery of Strombus gigas post-larvae. Totals and percent recovery are pooled from the replicates of each treatment; averages are weighted. Exposure time to 15 mM KCI did not significantly affect (one-way ANOVA; 0.05 [3,13] = 3.41) growth rate (F = 2.35)* or percent recovery (F = 0.40).* Average Average Average Average Total Total Time on Final Average Exposure Number Veliger Temp. in Initial Final Screen Pooled Shell Growth Duration of Age Metamorphosis (Larval) (Juvenile) Trays Percent Length Rate (hours) Replicates (days) (°C) Count Count (days) Recovery (mm) (mm/day) 6 2 24 27.8 17,550 12,512 13.0 71% 4.1 0.22 8 2 24 28.1 23,728 15,539 14.0 65% 4.0 0.20 10 3 21 DES 38,400 30,555 13.3 80% 3.2 0.15 16 10 21 28.3 137,800 98,421 13.6 11% 3.6 0.19 * F statistics are calculated with growth rate data log (x + 1) transformed and recovery data arcsine transformed. COMMERCIAL SCALE METAMORPHOSIS IN CONCH 71 TABLE 3. A comparison of 15 mM KCI and Laurencia extracts as metamorphic inducers for competent Strombus gigas veligers. There were no significant correlations (r? < 0.15) between veliger age, average temperature during metamorphosis, or time on screen trays, and percent recovery or growth rate of juveniles. Time on Final Veliger Exposure Initial Final Initial Final Screen Shell Growth Exp. Age Time Temperature Temperature (Larval) (Juvenile) Percent Trays Length Rate # Date (days) Inducer Dosage (hours) CC) (CC) Count Count Recovery (days) (mm) (mm/day) 1A 07/19 23 LAUR 12.5 ml/l 5 28.0 28.0 12,800 10,857 85% 14 4.0 0.20 1B 07/19 23 KCL 15 mM 10 28.0 28.0 12,800 10,755 84% 13 3.6 0.18 2A 08/10 24 LAUR 12.5 m/l 5 28.0 29.0 12,928 10,122 18% 14 3:9 0.19 2B 08/10 =. 24 KCL 15 mM 8 28.0 29.0 12,928 6,890 53% 14 4.1 0.21 3A 08/27 = 20 LAUR _ 7.0 ml/l 5 27.5 27.0 12,800 9,240 72% 13 3.0 0.14 3B 08/27 20 KCL 15 mM 10 Aes) 27.0 12,800 10,500 82% 12 2.9 0.14 3C 08/27 20 KCL 15 mM 10 27.5 27.0 12,800 9,300 73% 15 3 0.13 4A 09/03 20 LAUR 7.0 ml/I 5 29.0 28.0 12,800 12,822 100% 12 3.0 0.15 4B 09/03 20 KCL 15 mM 16 29.0 27.5 12,800 9,783 76% 11 BZ 0.18 SAN 09/12" 22 LAUR _ 7.0 ml/l 5 28.0 27.0 12,800 11,172 87% 14 3.6 0.17 5B 09/12 22 KCL 15 mM 16 28.5 27.0 12,800 12,198 95% 14 3.6 0.17 SC w09/12) , 22 KCL 15 mM 16 28.5 27.0 12,800 10,914 85% 13 Si 0.19 6A 10/06 20 LAUR _ 7.0 ml/I 5.5 28.5 29.0 12,800 11,646 91% 12 3.3 0.18 6B 10/06 20 KCL 15mM 16 29.5 28.0 9,000 5,064 56% 11 4.1 0.26 7A 10/08 19 LAUR _ 7.0 ml/l Sy) 28.5 29.5 12,800 9,024 1% 13 3.9 0.21 7B 10/08 19 KCL 15 mM 16 29.5 28.0 16,000 10,276 64% 10 3.3 0.21 8A 10/11 19 LAUR 7.0 ml/l 5 30.0 29.0 12,800 8,256 65% 12 3.4 0.18 8B 10/11 19 KCL 15 mM 16 31.0 28.0 12,800 10,284 80% 15 355 0.15 9A 10/24 22 LAUR 7.0 ml/I 7 28.5 29.0 15,000 = 11,466 716% 15 3.5 0.15 9B 10/24 22 KCL 15 mM 16 29.5 28.0 15,000 7,803 52% 16 4.1 0.18 ference observed in both species, in neither case did the KCl appear to affect the early growth of young postlarvae (Pechenik and Heyman 1987, and this study). Larvae of many invertebrate species have been induced to metamorphose with micromolar additions of KCI in- cluding Katherina tunicata, Saccostrea commercialis, Ha- liotis rufescens, Phestilla sibogae, Astraea undosa, Phrag- matopoma californica, and Crepidula fornicata (Rumrill and Cameron 1983, Nell and Holliday 1986, Yool et al. 1986, Pechenik and Heyman 1987). The effects of in- ducing agents on juvenile fitness has only been undertaken with Crepidula fornicata (Eyster and Pechenik 1988) and with Strombus gigas, in this study. It is hypothesized that the K+ ions depolarize externally accessible membranes, bypassing the normal receptor-stimulus interaction, and then activating the normal neuronal and/or hormonal pathway, resulting in metamorphosis, as proposed for Ha- liotis (Yool et al. 1986). The results reported by Eyster and Pechenik (1988) and in this study contribute to the growing body of evidence that KCl can safely be used to induce metamorphosis in some invertebrate species. The intensive hatchery at the Conch Farm can produce an overabundance of larvae at no additional cost and yet there are significant costs associated with inducing batches TABLE 4. The effect of metamorphic inducing agents on the growth rate (mm/day) and percent recovery of induced Strombus gigas postlarvae. There is no significant difference (t-test; 0.01 [1, 18] = 2.88) between using KCI or Laurencia extracts as a metamorphic inducer, in terms of percent recovery (t = 3.4) or growth rate (t = 1,429). Average Average Average Average Time Average Growth Number Veliger Temp. in Initial Final on Screen Percent Average Final Rate Metamorphic of Age + S.E. Metamorphosis (Larval) (Juvenile) Trays + S.E. Recovery Shell Length (mm/day) Inducer Replicates (days) + S.E. (°C) Count Count (days) + S.E. + S.E. (mm) + S.E. Laurencia 9 2180) ls = 28:48 s0570 117,528 94,605 132250 80/6) 11) 35S =H 10%4 | OF 10:02 15 mM KCl 11 210+1.6 28.3+0.69 142,528 103,767 13.141.9 72.74 145% 36+0.4 0.18 + 0.04 72 DAVIS ET AL. of larvae to metamorphose. Having indicated that there is not a significant difference in the growth or percent re- covery of Strombus postlarvae, induced with KCl as op- posed to the traditional Laurencia extracts, differences in cost and ease of use for the two inducers were closely ex- amined. Processing costs for the Laurencia extracts include collecting the algae from the sea, sorting, weighing and blending the algae, filtering the final produce, and per- forming dosage tests. These techniques total 2 person- hours/l of extract. Electricity costs for the extract include operation of the blender and freezer. Material costs for the extract include collection bags, scale, blender, storage con- tainers, freezer, and filtering apparatus. Potassium chloride on the other hand only requires purchasing of the product and scale, and weighing and mixing the chemical. KCI has no variation in the dosage, 15 mM, so does not require dosage testing with each batch. Laurencia extract costs $15.44/batch metamorphosis while KCl costs $4.92/batch metamorphosis; the extract costing more than 3 times as much as the salt. Since KCl is not significantly different than Laurencia extract in percent recovery and growth rate of induced postlarvae, its other desirable features stand out. Potassium chloride is a readily available and inexpensive chemical that can be used to induce metamorphosis with a consistent dosage and with a savings of labor (Pechenik and Heyman 1987). These factors make KCl an economical metamor- phic inducer for Strombus gigas veligers on a commercial scale. KCI induced metamorphosis may be useful for a va- riety of practical purposes including species control in mariculture, physiological studies on newly metamor- phosed juveniles, and biological toxicity tests on affected surface waters. ACKNOWLEDGMENTS We thank the Caicos Conch Farm for the use of the fa- cility. We thank Dr. Maureen Kaplan for patient and knowledgeable statistical advice. We thank Eastern Re- search Group, Inc., Arlington, MA for computer time and work space during the final preparation of this manuscript. We also gratefully acknowledge Ross Dobberteen for lo- gistical support, helpful conversations, and for critically re- viewing this manuscript. REFERENCES CITED Baloun, A. J. & D. E. Morse. 1984. Ionic control of settlement and meta- morphosis in larval Haliotis rufescens (Gastropoda). Biol. Bull. 167:124-138. Brownell, W. N. 1977. Reproduction, laboratory culture and growth of Strombus gigas, S. costatus and S. pugilis in Los Roques, Venezuela. Bull. Mar. Sci., 27(4):688—780. Burke, R. D. 1983. The induction of metamorphosis of marine inverte- brate larvae: stimulus and response. Can. J. Zool. 61:1701—1719. Cavanaugh, G. M. 1956. Formulae and Methods IV of the Marine Biolog- ical Laboratory Chemical Room, Woods Hole, MA; 62-69. Coon, S. L., D. B. Bonar & R. M. Weiner. 1985. Induction of settlement and metamorphosis of the Pacific Oyster, Crassostrea gigas (Thun- berg), by L-DOPA and catecholamines. J. Exp. Mar. Biol. Ecol. 94:211-221. Crisp, D. J. 1974. Factors influencing the settlement of marine inverte- brate larvae. Grant, P. T. & A. M. Mackie, eds. Chemoreception in marine organisms. New York, New York: Academic Press; p. 177-265. Davis, M., B. A. Mitchell & J. Brown. 1984. Breeding behavior of the queen conch Strombus gigas held in a natural enclosed habitat. J. Shellfish Res. 4(1):17—21. Davis, M., C. R. Hesse, & G. Hodgkins. 1986. Commercial hatchery produced queen conch, Strombus gigas, seed for the research and growout market. Proc. Gulf Caribb. Fish. Inst. 38:326—335. Davis, M. & A. Dalton. New large-scale culturing techniques for Strombus gigas post larvae, in the Turks and Caicos Islands. Proc. Gulf Caribb. Fish. Inst. (in press). Eyster, L. S. & J. A. Pechenik. 1988. Comparison of growth, respiration, and feeding of juvenile Crepidula fornicata (L.) following natural or KCl-triggered metamorphosis. J. Exp. Mar. Biol. Ecol. 118:269—279. Hadfield, M. G. 1977. Chemical interactions in larval settling of a marine gastropod. Faulkner, D. J. & W.H. Fenical, eds. Marine Natural Products Chemistry. New York: Plenum Publishing Company; p. 403-413. Hadfield, M. G. 1978. Metamorphosis in marine molluscan larvae: an analysis of stimulus and response. Chia, F. S. & M. E. Rice, eds. Settlement and Metamorphosis of Marine Invertebrate Larvae. Am- sterdam: Elsevier/North-Holland; p. 165—175. Hadfield, M. G. 1984. Settlement requirements of molluscan larvae: new data on chemical and genetic roles. Aquaculture 39:283—298. Heslinga, G. 1981. Larval development, settlement and metamorphosis of the tropical gastropod Trochus niloticus. Malacologia 20:349—357. Heyman, W. D., R. A. Dobberteen, L. A. Urry, & A. M. Heyman. 1989. Pilot hatchery for the Queen Conch Strombus gigas (L.), shows potential for inexpensive and appropriate technology for larval aqua- culture in the Bahamas. Aquaculture 77:277—285. Highnam, K. G. 1981. A survey of invertebrate metamorphosis. Gilbert, L.1., & E. Freiden, eds. Metamorphosis: A Problem in Develop- mental Biology. New York: Plenum Press; p. 43-73. Morse, D. E., N. Hooker, H. Duncan, & L. Jensen. 1979. y-aminobu- tyric acid, a neurotransmitter, induces planktonic abalone to settle and begin metamorphosis. Science. 204:407—410. Nell, J. A. & J. E. Holliday. 1986. Effects of potassium and copper on the settling rate of Sydney rock oyster (Saccostrea commercialis) larvae. Aquaculture. 58:263—267. Pechenik, J. A. & W. D. Heyman. 1987. Using KCl to determine size at competence for larvae of the marine gastropod Crepidula fornicata (L.) J. Exp. Mar. Biol. Ecol. 112:27-38. Rummill, S. S. & R. A. Cameron. 1983. Effects of gamma-aminobutyric acid on the settlement of larvae of the black chiton Katherina tunicata. Mar. Biol, 72:243—247. Scheltema, R. S. 1974. Biological interactions determining larval settle- ment of marine invertebrates. Thalassia Jugosl. 10:263—296. Siddall, S. E. 1982. Biological and economic outlook for hatchery pro- duction of juvenile queen conch. Proc. Gulf Caribb. Fish. Inst. 35:46-52. Snedecor, G. W. & W. G. Cochran. 1980. Statistical Methods: Seventh Edition. Ames, lowa: Iowa State University Press. COMMERCIAL SCALE METAMORPHOSIS IN CONCH 73 Sokal, R. R. & F. J. Rohlf. 1969. Biometry. San Francisco, California: W. H. Freeman & Company. Taylor, W. R. 1972. Marine Algae of the Eastern Tropical and Sub- tropical Coasts of the Americas. Ann Arbor, Michigan: University of Michigan Press. Thorson, G. 1950. Reproductive and larval ecology of marine bottom in- vertebrates. Biol. Rev. 25:1—45. Trapido-Rosenthal, H. G. and D. E. Morse. 1986. Availability of chemo- sensory receptors is down-regulated by habituation of larvae to a mor- phogenetic signal. Proc. Nat. Acad. Sci. 83:7658—7662. Veitch, F. P. & H. Hidu. 1971. Gregarious setting in the American Oyster Crassostrea virginica Gmelin: 1. Properties of a partially purified “‘setting factor.’’ Chesapeake Sci. 12(3):173—178. Yook, A. J., S. M. Grau, M. G. Hadfield, R. A. Jensen, D. A. Markell, & D.E. Morse. 1986. Excess potassium induces larval metamor- phosis in four marine invertebrate species. Biol. Bull. 170:255—266. Journal of Shellfish Research, Vol. 9, No. 1, 75—85, 1990. MARINE MUSSELS OF SOUTHERN AFRICA—THEIR DISTRIBUTION PATTERNS, STANDING STOCKS, EXPLOITATION AND CULTURE C. VAN ERKOM SCHURINK & C. L. GRIFFITHS Marine Biology Research Institute & Zoology Department, University of Cape Town, Rondebosch 7700, South Africa ABSTRACT Four species of mussel are abundant in southern Africa. The cool, upwelled waters of the west coast support the major populations of three of these, the ribbed mussel Aulacomya ater, the black mussel Choromytilus meridionalis and the introduced Mediterranean species Mytilus galloprovincialis, which has recently become the dominant intertidal species in the area. The warmer south and east coasts are colonized mainly by the brown mussel Perna perna, although all of the other three species penetrate along the south coast and Mytilus appears to have the potential to dominate this region in the future, at the expense of Perna. The overall standing stock of intertidal mussels is estimated at 114 x 10 metric tons whole wet mass. Of this 69% is found west of Cape Agulhas, 20% along the south coast, 4% in Transkei and of 7% in Natal. Mytilus galloprovincialis comprises 44% of total standing stock, while Perna perna contributes 39%. Despite the abundance of mussels along both west and south coasts, exploitation is minimal in these regions. The reasons for this are that shellfish are not traditional food resources in these areas and, in the case of the west coast, because of low human population density and the perceived risk of paralytic shellfish poisoning. The much smaller resources along the Transkei and Natal coasts are heavily exploited, with annual rates of removal approaching or even exceeding standing stock in some areas. Overall crop is probably less than 2000 tons, of which an estimated 316 tons is taken from the Transkei and 347 tons from Natal annually. Although mussel aquaculture is a recent development in southern Africa, the first farm having opened in 1984, current production of ca. 800 tons p.a. is already approaching the entire wild crop. Aquacultural output is, however, directed at the luxury market and is unlikely to play any role in reducing the pressure on wild stocks by subsistence gatherers. KEY WORDS: INTRODUCTION South Africa has a long, wave-exposed coastline which extends for some 2570 km from the Namibian border (28°S 16°E) in the west to that with Mocambique (26°S 32°E) in the east (Fig. 1). Sea temperatures within this coastal strip are determined largely by the two major current systems. To the east the southerly-flowing Agulhas Current trans- ports warm (21—26°C) water close inshore along the Natal coast, but south of East London this is deflected offshore by the progressively widening Agulhas Bank. The Ben- guela System of the west coast is characterized by frequent upwelling events, resulting in cooler conditions, with min- imum temperatures of 9—10°C being experienced during summer, when offshore winds predominate, while maxima of 15—16°C occur in winter, when sun-warmed surface waters advect onshore (Branch & Griffiths 1988). Based on these current systems the southern African coastline can be divided into three major biogeographical provinces (Brown & Jarman 1978): a sub-tropical east coast region extending southwards to about East London; a warm-temperate south coast reaching from there to Cape Agulhas and a cold-tem- perate west coast region extending northwards into Na- mibia (Fig. 2). Some 27 species of mytilid mussels have been recorded in southern African waters (Kilburn & Rippey 1982), four of which attain sufficient size and density to form extensive beds on rocky intertidal and subtidal reefs. Of these one, the brown mussel Perna perna, is found predominantly in the sub-tropical east and warm-temperate south coast re- gions. The other three, the indigenous black mussel Choro- 18) mussels, Aulacomya, Choromytilus, Mytilus, Perna, distribution, exploitation, culture mytilus meridionalis, the ribbed mussel Aulacomya ater, and the introduced European mussel Mytilus galloprovin- cialis (Grant et al. 1984; Grant & Cherry 1985), all attain their greatest densities in the cooler waters of the west coast (Fig. 2). In addition the smaller Semimytilus algosus has been re- corded as abundant in northern Namibia. Penrith & Kensley (1970) give the southern limit of this species as Swakopmund (22°S), although recent data show it as oc- curing as far south as Elizabeth Bay (27°S) (C. Beyers, Sea Fisheries Research Institute, Swakopmund, pers. comm.). Numerous publications have addressed aspects of the ecology or physiology of the various southern African mussel species. Most of these have been reviewed by Grif- fiths & Griffiths (1987) and Branch & Griffiths (1988). The rapidly increasing exploitation of Perna perna by indige- nous peoples in the Transkei, and the effects of this on community structure have also received considerable atten- tion (Bigalke 1973; Siegfried et al. 1985; Hockey & Bosman 1986; Hockey et al. 1988; Lasiak & Dye 1989). In this paper we take a more comprehensive view of the mussel resources of the region by making a first estimate of the stocks of littoral mussel species and of their geograph- ical distribution. This is then related to the exploitation pressure on such stocks and the recent development and rapid expansion of a mussel culture industry in the region. SPECIES CHARACTERISTICS The features characterizing the four common mussel species found along the southern African coastline are sum- 76 VAN ERKOM SCHURINK AND GRIFFITHS NAMIBIA 2 Marcus Island &@ Dassen Island Robben Island Port Bloubergstrand Elizabeth Cape Peninsula Cape Point False Bay Tsitsikamma Coastal National Park Cape Agulhas Hermanus Betty's Bay Plettenberg Bay Figure 1. The southern African subregion showing sites referred to in the text. marized in Table 1. The ribbed mussel Aulacomya ater is the most easily recognized species, by virtue of the strong external radial sculpturing on its shell. Perna perna is most reliably distinguished by the divided posterior rectractor muscle scar on the interior surface of the shell, but in the field the brown shell color provides a good identification feature, especially for specimens collected on the south and east coasts. On the west coast P. perna is uncommon and tends to occur singly amongst Mytilus galloprovincialis, which can also be brown. Confusion can thus occur, but after examination of the adductor scars of a few specimens “mmm High population density along the coast = over 20 hab/km2 (Reader's digest of South Africa) South western AULACOMYA| South coast o Mussel farms Figure 2. Distribution patterns of major mussel species around the southern African coastline—the thickness of lines indicates relative abundance. Locations of the four mussel farms currently in produc- tion are also shown. identifications can usually still be made on the basis of color and shape, the shells of Perna tending to be more elongate and slender than those of Mytilus. Choromytilus meridionalis can be reliably distinguished from P. perna by shell color, but is easily confused with M. galloprovin- cialis, especially when small. The dark brown color of the female gonad in Choromytilus, the absence of pits in the resilial ridge and an anterior adductor muscle are the best distinguishing features, but specimens can be also distin- guished by shape in the field. Choromytilus is much thinner and more symmetrical in cross-section than Mytilus, large specimens of which can be so broad and ventrally flattened that they stand un-supported on the ventral shell surface. DISTRIBUTION PATTERNS Distribution patterns were determined from museum collections, published distribution records and from direct field observations at sites in the Durban region and at 12 sampled, and numerous other visually examined, locations between East London and Groen River in the Cape Prov- ince, in addition to visual inspection of numerous other sites. Detailed vertical zonation patterns were also exam- ined at three locations in the south western Cape, where the ranges of all four mussel species overlap. These were car- ried out by removing a series of 0.0625 m? (25 cm x 25 cm) quadrats along a belt across the intertidal from high to low water of spring tides and counting the mussels of each species contained within each sample. The ribbed mussel, Aulacomya ater, which is wide- spread on both Atlantic and Pacific coasts of South America (Suchanek 1986) has a southern African distribu- tion extending from approximately Rocky Point (18°S) in northern Namibia (Penrith & Kensley 1970), to Port Alfred (26°S) on the south-east coast (Fig. 2) (Day 1974; Kilburn & Rippey 1982). Aulacomya reaches its maximum abun- dance in the sublittoral and is the dominant species found in the kelp beds and sublittoral reefs along the Cape west coast (Pollock 1979; Field et al. 1980). Over the past de- cade there has been a marked decline in the numbers of Aulacomya in the intertidal, as they have become replaced by the invasive Mytilus, which is a superior competitor for intertidal space (Hockey & van Erkom Schurink 1990). Choromytilus meridionalis is restricted to areas of cool upwelled water on the west and south coast of southern Africa. There is some doubt as to whether the species is taxonomically distinct from the South American C. chorus (Kilburn & Rippey 1982). The distribution ranges from about Walvis Bay (22°S) to northern Transkei (30°E), but there is a marked reduction in abundance east of Cape Agulhas. Extensive beds of C. meridionalis were long re- garded as a characteristic feature of west coast rocky shores (e.g. Brown & Jarman 1978), but Mytilus galloprovincialis now dominates most intertidal wave-exposed sites in this region. Observations by A. du Plessis (Sea Harvest Corpora- tion, Saldanha Bay, unpublished) and Hockey & van 77 MARINE MUSSELS IN SOUTHERN AFRICA WL SZ} WW 06 JEJONIIGNS P JEp!va}u| SWYM-HO yjoows ‘suiBsew uo Ajjeloadsa N33Y»D ulm pabuly Sawjawos NMOUE ww Op} Wu Sg JEJONIGQNS ve MO}|9A 0} SYM-}3O Jepiieiu| yjoows ‘(noyBnosy) UMOIG YH) Ajases) AjjesjuUaA UMOIG O} Bulpeys yO AijeaidA 1 sijerosulaosdoyjeb sniAW ww OS} WWW O06 JEJONNGNS 2 6 MO||9A 0} OVYM-}}JO jepiiayuy “d0yo yWeG YJOOWS ‘UMOIG yep Ajjeuoiseo00 yovid Ajrensy yjoows ea SIJPEUOIPlaw snjyAwoioyug WW G6 wu OZ JEJONIGNS ve MO||9A 0} SWYM-HO JEpija}u| squ |eipes Anem paysew YIM MOV1d 0 NMOY yjoows BY, 4 Jaye eAwoorjny *QUIP}SEOD UBIIY UJIYINOS ay) Suoye JURpUNge satdads jassnul INO ay} JO YORI BuIZII1a}IeIeY sainjeIay ‘TT aATav azis wnuwyjxew ayewlsxoiddy Inojoo ysaj4 ainyxe} FP 4nojoo ays jeusayxy eBpis jeyisey aytyoud uol}aeaS SSO1D ¥B (y9e1q ul sueos ajosnw Buimoys) |jays JO M@AIA [eUusa}U] 78 VAN ERKOM SCHURINK AND GRIFFITHS Erkom Schurink (1990) suggest that this change has re- sulted from the recent introduction and rapid spread of My- tilus galloprovincialis at the expense of C. meridionalis, as postulated by Grant & Cherry (1985). Intertidal populations of C. meridionalis are currently found on low-shore rock surfaces subject to sand cover or abrasion (Branch & Griffiths 1988). Extensive sublittoral beds still occur at selected sites throughout the region, as is evidenced by the abundance of large, fresh shells cast ashore following winter storms, and by direct diving obser- vations, such as those at Marcus Island, Saldanha Bay, un- dertaken by Barkai & Branch (1988). The brown mussel Perna perna is the dominant mussel species in northern Namibia, but is virtually absent on the central and southern parts of the west coast. It becomes dominant again along the south and east coasts of southern Africa into Mocambique (Fig. 2). Isolated individuals occur throughout the west coast, but do not form beds in the area between False Bay and central Namibia. Outside the southern African region P. perna is widely distributed in the tropical and subtropical regions of the Indian and Atlantic oceans (Berry 1978). Populations in the Mediterra- nean are considered by some recent authors to represent a distinct species, P. picta (Shafee 1989). The Mediterranean mussel, Mytilus galloprovincialis, is widespread in the Mediterranean, along the Atlantic coast of Europe and in north west Africa, where it may hybridize with M. edulis (Beaumont et al. 1989). It has also colon- ized Japan (Wilkins et al 1983; Hosomi 1984). The pres- ence of this species along the west coast of South Africa was first reported by Grant et al. (1984), although the iden- tification was only confirmed the following year (Grant & Cherry 1985). M. galloprovincialis is currently the domi- nant intertidal mussel species over an extensive area stretching between Luderitz (26°S) and Cape Point (34°S), but also occurs in lower density on the east coast up to East London (27°E). This is despite the fact that it is thought to have been introduced very recently, perhaps only over the past 20 years (Grant & Cherry 1985). M. galloprovincialis is a fast-growing species with a high reproductive output and a considerable tolerance for desiccation. It outcompetes the indigenous west coast species, at least in silt-free intertidal sites (Hockey & van Erkom Schurink 1990). For some unknown reason M. gal- loprovincialis is uncommon in the sublittoral, although it flourishes when permanently submerged in suspended rope culture (see below). Experimental studies (van Erkom Schurink & Griffiths, in prep.) have shown that M. gallo- provincialis grows rapidly in the warmer conditions at Port Elizabeth. This suggests that it has the potential to invade the south and east coasts in the future, in competition with Perna perna. VERTICAL ZONATION PATTERNS Three mussel species coexist at sites on the west coast (Fig. 3a,b). Mytilus is usually the most abundant of these MYTILUS MARCUS ISLAND ea 09 GREE CHOROMYTILUS a) o6 0.3 0 MYTILUS BLOUBERGSTRAND AULACOMYA Leét L292 €€P f0o> ZS€ OSt E8t NVSW Apnys siyt LS2 \> 681 ep GZ - OOlP OO€ + 006 ae) 9p £6 8z puesjsbiaqnojg Apnjs siyt 60€ \> LE% Gz Ly - 0062 006 » 002 G ae OF g 6 pue|s| snoieyy Apmis siyt _Se% : _SSI ra ss = 000 002 | 00e > (@ 74 8k ___J8ANY UBOID d Ww ©) V d Ww ©) V d Ww 6) Vv aoinos AIS 10} as0ys A901 |W} SUO} aioys AyO01 Jeplajul z-w by ayIS , Wy SUO} [BIO] yoo}s Buipueys jeuoibay ,-W4 paq jassnu zw SSBWWOIG Jam UBAY\ *JSBOD UBILIZY UJIYINOS dy} PUNOIE SUOISIA SNOLIBA 1OJ SUO} ILIVIU UL S[assNU JO SyI0}s SuIpUR)S [EPIAIIUL JO SaVUIS| ‘CT ATAVL (uopuo7 jseq 0} (seyjnBy edeg (julog adeg o} seyjnby ades) 0} JUulog odes) LSVOO HLNOS Jaaty aBueso) LSVOO LSAM dadVO “MS MARINE MUSSELS IN SOUTHERN AFRICA 81 TABLE 3. Intertidal mussel stocks (in metric tons) and their rates of exploitation in each region. Sources of data see Table 2 and text. Stock size (t) West S. West South Transkei Natal Total % Aulacomya 10 609 535 —_ _ = 11 242 9.8 Choromytilus 6 542 697 = _— 7 7 141 6.4 Mytilus 47 457 2 641 222 _ -- 50 335 44.2 Perna <1 9 882 21 905 <5 000 8 400 ca.45 006 39.6 Total 64 608 13 770 22 127 <5 000 8 400 ca.113 905 Region as % total stock 56.7 12.1 19.4 ca.4.4 7.4 Crop (ty~') <100 <100 ? 316 347 Crop as % standing stock <1% <1% 2 ca.6.3 4.1 with less than 1% of the biomass comprising Mollusca, most of which are gastropods. Perna perna is the only abundant mussel in Transkei, but, as a result of intense exploitation pressure (see below) extensive intertidal beds are found only within nature re- serves and on inaccessible rock faces (Lasiak & Dye 1989). Elsewhere intertidal populations comprise either newly set- tled cohorts, or have been reduced to small groups of indi- viduals scattered from the lower shore into the sublittoral fringe. Although Hockey & Bosman (1986) report P. perna densities ranging from 0—300 individuals m~? at four sites in the area, no biomass estimates are available for the Transkei. Some idea of overall standing stock can, how- ever, be obtained from the rate of shellfish removal. Hockey et al. (1988) estimate that tribal peoples remove a total of 555 tons of shellfish from the Transkei coast per annum (see below). This rate of exploitation is severely depleting the resource (Siegfried et al. 1985; Hockey & Bosman 1986; Lasiak & Dye 1988), despite the fact that P. perna has a high production to biomass (P/B) ratio of about 4 on the Natal coast (Berry 1978). At some sites, the an- nual rate of removal exceeds the standing stock (Siegfried et al. 1985) indicating that overall stock is unlikely to sub- stantially exceed the crop in these areas. Sections of the coast are, however, inaccessible and hence poorly ex- ploited. These factors make it difficult to make a reliable estimate of standing stock, although it appears unlikely this could exceed 5000 tons. Largely because of extreme expo- sure to wave action nothing is known of sublittoral mussel stocks in this area. These are protected from exploitation by virtue of their inaccessibility and are likely to exceed the intertidal stock. In Natal, as in Transkei, P. perna is the only common mussel species. Some standing stock estimates have been published by Berry (1978), but the reef on which he worked has subsequently been covered by sand, elimi- nating the population. There are indications that population densities have declined throughout the region over the past decade, partly as a result of increased exploitation pressure (Martin & de Freitas 1987; de Freitas pers. comm.). Standing stock estimates are thus based on unpublished data provided by de Freitas (Oceanographic Research Insti- tute, Durban, pers. comm.). Mussel standing stock within the mussel zone at two sites, Umhlodti and Cape Rock, are 2.1 kg m~? and 7.9 kg m~? (whole wet weight) respec- tively, giving an average of 5 kg m~?. The average width of the mussel zone is estimated at 10 m, and there are 168 km of rocky shore in Natal. This gives a total potential area of mussel habitat of 1.68 x 10° m? and a standing stock estimate of 8400 tons of mussels for the region as a whole. SIZE-FREQUENCY DISTRIBUTIONS The size-frequency distributions of mussels were deter- mined from a series of sites around the coast. Size-distribu- tions are dependent upon recent settlement success, and hence vary with season and over time, even within a fixed site (e.g. Griffiths 1981). Aulacomya ater is consistently smaller than other species. This can be attributed to the lower tolerance of this species to aerial exposure (Hockey & van Erkom Schurink 1990), its smaller terminal size in this area (Table 1) and its low rate of growth (Wickens & Field 1988). Intertidal mussels of all species attain a much smaller terminal size than do sublittoral populations (Table 1), largely because of the reduced time available for feeding. For this reason few intertidal mussels exceed 75 mm, even at sites which are unexploited because they are remote from centers of population (e.g. Groenrivier) or closed to the public (e.g. Marcus Island). The largest mussels on west coast shores can either be Choromytilus or Mytilus, de- pending on the topography of the site and the population age structure. Very large individuals of both species (> 100 mm) are sometimes encountered in the intertidal, but these are usually confined to lowshore gullies or have been dis- placed from sublittoral beds and become reattached in shel- tered intertidal pools. The size distributions of P. perna 82 VAN ERKOM SCHURINK AND GRIFFITHS populations are extremely variable. This can be attributed to unpredictable settlement success in this species, com- bined with its rapid growth rate and short life span of only 2—3 years (Berry 1978). Intense human exploitation pres- sures nevertheless have had a marked impact on popula- tions of P. perna, as is evidenced by the much higher pro- portion of larger (<50 mm) individuals at unexploited, rel- ative to exploited sites (Fig. 4, see also Lasiak & Dye 1989 and Crawford & Bower 1983). The generally smaller average size of individuals in our more northerly samples is all the more notable given that P. perna grow more rapidly in the warmer waters of this region. Thus in Durban, inter- tidal populations attain 50—60 mm in their first year (Berry 1978), whereas this declines to 30—40 mm in Transkei (Lasiak & Dye 1989) and Tsitsikamma (Crawford & Bower 1983). Growth rates in the south western Cape are even slower (van Erkom Schurink, unpublished data). EXPLOITATION The north western Cape coastal belt is a semi-desert re- gion of poor agricultural potential which supports a sparse human population. This, combined with the restricted number of access points to the coast, has constrained ex- ploitation of intertidal food resources. However, mussels are becoming more popular amongst urban residents, par- ticularly immigrants from those European countries with a tradition of seafood consumption. This has led to some in- crease in consumption, particularly in the population centers around Cape Town. Many would-be collectors are, however, discouraged from removing mussels by the risk of paralytic shellfish poisoning, which is a fairly common occurrence along the Cape west coast (Grindley & Nel 1970), caused by blooms of the dinoflagellates Gonyaulax grindleyi & G. catanella (Horstman 1981). The net result of these factors is that, despite the abun- dance and ready availability of mussels in this region, they are subject to very low levels of human exploitation. While we are not in a position to quantify the amounts taken, these probably total <100 tons and are insignificant com- pared with the high rates of mortality resulting from intra- specific competition for space and consumption by natural predators (Griffiths & Hockey 1987). A similar situation prevails in the south-western Cape region, although anglers in this area can inflict substantial damage to intertidal mussel beds while collecting mussel worms, Pseudonereis variegata, for bait (van Herwerden 1990). Here again the total biomass removed is probably less than 100 tons y~!. The southern Cape coast is a popular tourist area, and holidaymakers frequently collect mussels for food, particu- larly during the peak summer season. As a result, marked differences between the size-frequency distributions and biomass of Perna inside and outside of conservation areas may occur, as shown in Fig. 4 (Crawford & Bower 1983). This effect is, however, probably confined to the vicinity of resorts or residential areas. More sustained and intense ex- R ) ( | | : Hid 1D | p Size frequency Figure 4. Size-frequency distributions of mussel species at a variety of sites arranged from west (top left) to east (bottom right) around the southern African coastline. Data for Tsitsikamma after Crawford & Bower (1983), for Transkei after Hockey (unpublished) and for Durban after Berry (1978). ploitation pressure occur to the east of this region, around East London, by virtue of increased densities of tribal peoples, who exploit the shore for their domestic needs (Branch & Shackleton 1988). Unfortunately no quantifica- tion of the crop removal rate is available for this area. Far better data are available for the Transkei, where Hockey et al. (1988) have estimated the annual rate of re- moval by rural shellfish collectors using aerial survey tech- niques. Exploitation pressure varies widely along the coast, depending upon human population density, rock type and geographical location. In general, subsistence exploitation is far more intensive in the southern sector (shown in Fig. 5), where it totals some 5573 kg km“! of rocky shore y~!, than in the more sparcely populated northern region (206 kg km~! y~!). Taking the relative lengths of coastline into consider- ation the total shellfish crop removed from the entire region has been estimated at 555 tons y~!, of which 316 tons is P. perna (Hockey et al. 1980). While exploitation by rural peoples in the Transkei goes largely unregulated, the harvesting of intertidal and subtidal invertebrates in Natal is controlled by provincial ordinance, collectors being obliged to purchase a license in order to collect from the shore. They are also required to submit catch returns, the analysis of which give an indication of total annual catch. The total number of mussel and general bait licenses (which also permit collection of mussels) in- MARINE MUSSELS IN SOUTHERN AFRICA 83 Figure 5. Intense exploitation of Perna perna in the southern Transkei by subsistence gatherers (photo: C. L. Griffiths). creased from 1012 in 1974 to 10,606 in 1986 and the mussel catch by licenced collectors rose from 105 tons y~! to 317 tons y_! over the same period (de Freitas & Martin 1988). Since only those catches reported by license holders are incorporated into this figure, it clearly represents a min- imum estimate. Not only are license holders likely to un- derestimate their take, but subsistence harvesting by tribal peoples is seldom included in catch returns. This can be significant in certain areas. Van der Elst & Tregoning (in press) have, for example, documented exploitation by rural people along the Maputoland coast, adjacent to the Mo- cambique border. Mussels are by far the most important marine crop in this area, each collector removing an average of 4—6 kg shelled mussel flesh per collecting trip —this amount being determined simply by the maximum whole unshelled mussel weight the women can comfortably carry up from the shore in a single trip. The total whole wet mass of mussels removed during 1981 from this 83 km of shore—of which only 17 km are rocky —was estimated by van der Elst & Tregoning (in press) at 30 tons y~!. This amount can be added to that for the rest of Natal, calculated above. The total crop collected in Natal is thus at least 347 tons (1986 figures). The exploitation pressure on the shore both by rural peoples and license holders, continues to increase, as indicated by the demand for licenses in Natal, which is increasing at a rate of some 23% p.a (de Freitas & Martin 1988). Collected crops estimated above have thus probably already been considerably exceeded. CULTURE Mussel culture is not a traditional activity in southern Africa. The first commercial culture facility, Seafarm, was established in 1984 and makes use of an enclosed **dam’’ cut off from Saldanha Bay by a causeway. Two additional operations, Atlas Sea Farms and Sea Harvest Corporation, began production in open water in the same area in 1987, while Atlas Sea Farms also operates a satellite facility at Port Elizabeth. All three companies use the Spanish method of rope culture, the ropes being suspended either from longlines or from rafts. The three farms in Saldanha Bay primarily market the introduced Mytilus galloprovincialis. Wild seed, usually at a shell length of 20—40 mm are collected either from caissons or other structures in the harbour, or settle natu- rally on the culture ropes. Mussels are marketed at a shell length of 55—90 mm. Growth to this size is usually accom- plished within 4—5 months, largely independent of time of year. Some Choromytilus are found amongst the seed stock, but although these grow as fast and are as palatable as Mytilus, they are not favoured for marketing, largely be- cause of the dark brown color of the female gonad. The Atlas Sea Farms operation in Port Elizabeth grows mainly Perna perna, although some Mytilus have been translo- cated from Saldanha and grow well in the warmer water conditions at Port Elizabeth. There has been a rapid increase in production of farmed mussels since 1984 (Fig. 6). Output for 1989 is estimated to be 800 tons whole wet mass. Most of this is sold live in the shell at a wholesale price of R3—4 kg~! (about US$1—1.4 at current exchange rates) giving the industry as a whole a gross annual income of the order of R3m (US$Im). Other processed products such as frozen seafood mixes, fresh and canned meats and smoked mussels under development are likely to absorb much of the future growth in output. Export potential is also being investigated. CONCLUSIONS This paper presents a first attempt to quantify the marine mussel resources of the southern Africa region and to pro- vide an overview of their distribution patterns, exploitation and management status. Three indigenous and one introduced species of mussel occur in the region in sufficient densities to be considered exploitable. All four differ in their geographical ranges, vertical distribution patterns and in such biologically im- portant parameters as growth rate and terminal size. These differences are reflected in the stock size, accessibility and vulnerability to exploitation, which show marked regional as well as species-specific variation. 1000 900 800 700 (tons) 600. Production 1987 1984 1985 1986 1988 1989 Figure 6. Annual production of cultured mussels in the southern Af- rican region 1984-1989. 84 VAN ERKOM SCHURINK AND GRIFFITHS The cooler west-coast region supports the major portions of the stocks of three of the four species. Of these the indig- enous Aulacomya ater is primarily sublittoral and has a rel- atively slow growth rate and small terminal size. Thus, al- though it attains a significant intertidal (and very large sub- littoral) biomass, and is a crucial food source for some benthic predators, it is the least suitable of the species in terms of human exploitation and culture. Choromytilus meridionalis is a faster-growing, larger species and has hitherto been regarded as the major intertidal mussel species in this region. It has, however, recently been largely displaced by the introduced Mytilus galloprovin- cialis, although it remains abundant at some sites, notably those subject to sanding. Because of its high fecundity, rapid growth, considerable tolerance to aerial exposure and ability to grow in dense beds, M. galloprovincialis is now the dominant intertidal mussel along the coastline, com- prising some 74% of the total littoral stock of 64,608 metric tons. The establishment of M. galloprovincialis has almost certainly increased total mussel biomass, since it both oc- cupies a large vertical range and attains a higher biomass per m? than do the species it has displaced. Moving from the west coast eastwards there is a pro- gressive decline in the abundance of all three of the above species and they are replaced by a single, warm-water form, Perna perna. Mussel standing stock on the east coast is relatively low due in part to the narrower niche breadth of the single east coast species, relative to those of the three west coast forms combined. Other factors, such as human predation pressure and competition between space-occu- pying species on the shore in the different regions, also play some role, since there are few sites at which P. perna form the extensive beds typical of west coast shores, even in the absence of human exploitation pressure. The more intense exploitation pressure on the east rela- tive to the west coast makes it difficult to speculate as to the size of the original unexploited stock. Paradoxically, the regions subject to most intense ex- ploitation are those that probably always supported the smallest stocks, while the extensive resources along the west coast remain virtually pristine. In theory the annual removal of ca 1,000 tons could easily be sustained by the total stock of 114,000 tons without any adverse effects, were it to be evenly spread geographically. Indeed, under a properly managed regional management policy the resource could probably support a crop at least an order of magni- tude larger than that taken at present. However, this is im- practical since the demand originates from subsistence- gatherers who reside in areas distant from the centers of mussel abundance. Within the commercial sector potential might exist for the exploitation of wild mussel stocks. This largely luxury market, however, demands a high quality product which is clean, free of sand and of regular size. Such a market is best supplied by mussels grown under controlled culture conditions. Such operations have shown an exponential growth in South Africa over the past few years. Indeed, the existing local market may already be ap- proaching saturation and further growth may depend upon either development of an export market or an increasing shellfish consumption by the local population. Neither of these factors is likely to reduce pressure on wild resources by local people on the east coast, which, in the light of increasing population pressure, remains a serious manage- ment problem. ACKNOWLEDGEMENTS We are indebted to Prof. A. J. de Freitas for providing us with unpublished data on mussel stocks in Natal and to the managers of Sea farm, Sea Harvest and Atlas Sea Farms for allowing us access to their production figures. Our thanks to Sandy Tolosana for help with preparation of this manuscript and to Prof. G. M. Branch and Dr. P. A. R. Hockey for their constructive comments on earlier drafts. REFERENCES Bally, R., C. D. McQuaid & A. C. Brown. 1984. Shores of mixed sand and rock: an unexplored marine ecosystem. S. Afr. J. Sci. 80:500— 503. Barkai, A. & G. M. Branch. 1988. Contrasts between the benthic com- munities of subtidal hard substrata at Marcus and Malgas Islands: a case of alternative stable states? §. Afr. J. mar. Sci. 7:117—137. Beaumont, A. R., R. Seed & P. Garcia-Martinez. 1989. Electrophoretic and morphometric criteria for the identification of the mussels Mytilus edulis and M. galloprovincialis. In: Ryland. J. S. & P. A. Tyler. Re- production, Genetics and Distributions of Marine Organisms. Olsen & Olsen, Fredensborg, Denmark. pp 251—258. Berry, P. F. 1978. Reproduction, growth and production in the mussel, Perna perna (Linnaeus). Investl. Rep. oceanogr. Res. Inst. S. Afr. 48:1—28. Bigalke, E. H. 1973. The exploitation of shellfish by coastal tribesmen of Transkei. Ann. Cape Prov. Mus. 9:159—175. Branch, G. M. & C. L. Griffiths. 1988. The Benguela ecosystem part V. The coastal zone. Oceanogr. Mar. Biol. Annu. Rev. 26:395—486. Brown, A. C. & N. Jarman. 1978. Coastal marine habitats. In: Biogeog- raphy and Ecology of Southern Africa: M. J. A. Werger & A. C. van Bruggen (Eds). W. Junk, The Hague 1241-1277 p. Connell, A. D. 1988. A photographic survey of reefs in the vicinity of the SAICCOR pipeline at Umkomaas, Natal. CSIR, Durban. Crawford, R. J. M. & D. F. Bower. 1983. Aspects of growth, recruit- ment and conservation of the brown mussel Perna perna along the Tsitsikamma coast. Koedoe 26:123—133. Day, J. H. 1974. A Guide to Marine Life on South African Shores. A. A. Balkema, Cape Town. 300p. De Villiers, G. 1977. The feasibility of commercial exploitation of a black mussel population near Hermanus, South Africa. Fish. Bull. S. Afr. 9:1-5S. Duvenage, I. R. 1988. Biological structure of an Agulhas Bank subtidal reef community and consumption estimation for a dominant species. CSIR Research Report 638:1—88. Field, J. G., C. L. Griffiths, R. J. Griffiths, N. Jarman, P. Zoutendyk, B. Velimirov, & A. Bowes. 1980. Variation in structure and biomass of MARINE MUSSELS IN SOUTHERN AFRICA 85 kelp communities along the south-west Cape coast. Trans. Roy. Soc. S. Afr. 44:145—203. Grant, W. S. & M. I. Cherry. 1985. Mytilus galloprovincialis Lmk. in southern Africa. J. Exp. Mar. Biol. Ecol. 90:179-191. Grant, W. S., M. I. Cherry & A. T. Lombard. 1984. A cryptic species of Mytilus (Mollusca:Bivalvia) on the west coast of South Africa. S. Afr. J. mar. Sci. 2:149—162. Griffiths, R. J. 1981. Population dynamics and growth of the bivalve Choromytilus meridionalis (Kr.) at different tidal levels. Estuar. Coastal Shelf Sci. 12:101-118. Griffiths, C. L. & R. J. Griffiths. 1987. Bivalvia. In: Animal Energetics. T. J. Pandian & F. J. Vernberg (Eds). Academic Press, New York 1-88 p. Griffiths, C. L. & P. A. R. Hockey. 1987. A model describing the inter- active roles of predation, competition and tidal elevation in structuring mussel populations. S. Afr. J. mar. Sci. 5:547—556. Grindley, J. R. & E. A. Nel. 1970. Red water and mussel poisoning at Elands Bay, December 1966. Fish. Bull. S. Afr. 6:36—55. Hockey, P. A. R. & C. van Erkom Schurink. 1990. The invasive biology of the mussel Myrilus galloprovincialis on the southern African coast. Trans. Roy. Soc. S. Afr. (in press). Hockey, P. A. R. & A. L. Bosman. 1986. Man as an intertidal predator in Transkei: disturbance, community convergence and management of a natural food resource. Oikos 46:3—14. Hockey, P. A. R., A. L. Bosman & W. R. Siegfried. 1988. Patterns and correlates of shellfish exploitation by coastal people in Transkei: an enigma of protein production. J. Appl. Ecol. 25:353—363. Horstman, P. A. 1981. Reported red-water outbreaks and their effects on fauna of the west and south coasts of South Africa, 1959-1980. Fish. Bull. S. Afr. 15:71-88. Hosomi, A. 1984. Ecological studies on the mussel Mytilus galloprovin- cialis (Lamarck)—III. Rise and fall of recruited cohorts in adult mussel beds. Venus 43:157—171. Jackson, L. F. 1976. Aspects of the intertidal ecology of the east coast of South Africa. Invest. Rep. oceanogr. Res. Inst. S. Afr. 46:1—72. Kilburn, R. & E. Rippey. 1982. Sea Shells of Southern Africa. Macmillan South Africa, Johannesburg 249 p. Lasiak, T. & A. Dye. 1989. The ecology of the brown mussel Perna perna in Transkei, southern Africa: Implications for the management of a traditional food source. Biol. Cons. 47:245—257. Martin, L. K. & A. de Freitas. 1987. Mussels—an exploited resource. Poster presentation: Sixth National Oceanographic Symposium, Stel- lenbosch, July 1987. Penrith, M. L. & B. F. Kensley. 1970. The constitution of the intertidal fauna of rocky shores of South West Africa. Part 2. Rocky Point. Cimbebasia 1:243—268. Pollock, D. E. 1979. Predator-prey relationships between the rock lobster Jasus lalandii and the mussel Aulacomya ater at Robben Island on the Cape west coast of Africa. Mar. Biol. 52:347—356. Seigfried, W. R., P. A. R. Hockey & A. A. Crowe. 1985. Exploitation and conservation of brown mussel stocks by coastal people of Transkei. Environ. Conserv. 12:303—307. Shafee, M. S. 1989. Reproduction of Perna picta (Mollusca: Bivalvia) from the Atlantic coast of Morocco. Mar. Ecol. Prog. Ser. 53:235— 245. Suchanek, T. H. 1986. Mussels and their role in structuring rocky shore communities. In: The Ecology of Rocky Coasts—P. G. Moore, & R. Seed (eds) Columbia University Press, New York 70—96p. Van Herwerden, L. 1989. Collection of mussel worms (Pseudonereis va- riegata) for bait—a legislative anachronism. S. Afr. J. mar. Sct. 8:363—366. Wickens, P. A. & J. G. Field. 1988. Rock lobsters Jasus lalandiit and mussels Aulacomya ater: demographic parameters and simulation models of their populations and interaction. S. Afr. J. mar. Sci. 7:193-217. Wilkins, N. P., Fujino, K., & E. M. Gosling. 1983. The Mediterranean mussel, Mytilus galloprovincialis Lmk. in Japan. Biol. J. Linn. Soc. 20:365—374. Wooldridge, T. 1988. Proposed marine development in the southwest sector of Plettenberg Bay—an ecological assessment. Institute for Coastal Research, University of Port Elizabeth. Report No 18:1—62. Journal of Shellfish Research, Vol. 9, No. 1, 87—93, 1990. EFFECTS OF EXPOSURE TO WAVE ACTION ON ALLOCATION OF RESOURCES TO SHELL AND MEAT GROWTH BY THE SUBTIDAL MUSSEL, MYTILUS GALLOPROVINCIALIS DAVID RAUBENHEIMER* AND PETER COOK Department of Zoology University of Cape Town Rondebosch, 7700, South Africa ABSTRACT Populations of the subtidal mussel, Mytilus galloprovincialis (Lmk.), growing in a rocky habitat exposed to direct wave action had thicker shells and a lower meat weight/shell weight ratio than both wild and cultured populations sheltered from wave action. Young mussels transferred from the thick-shelled population and grown in a sheltered habitat developed thin shells, suggesting that shell thickening in this population was a phenotypic response rather than a genetically fixed attribute. The mussels growing in the open ocean but in a position sheltered from direct wave action had a higher meat weight/shell weight ratio than those grown in a sheltered dam. Both the exposed and the sheltered populations growing in the open ocean had a greater slope for the regression of shell weight on shell surface area than the mussels grown in the sea water dam. It is concluded that any growth benefits due to growing in exposed habitats may be channelled into increasing shell thickness rather than increasing shell length or body weight. KEY WORDS: Mytilus galloprovincialis, mussels, resource allocation, wave action, phenotype, aquaculture INTRODUCTION The energy ingested as food by bivalve molluscs is par- titioned into metabolic maintenance and growth, the latter comprising growth of the flesh (including gonads) and of the shell (Griffiths and Griffiths, 1987). In commercial aquaculture it is desirable to manipulate energy parti- tioning, not only to maximise the energy allocated to growth, but also to influence the proportions of this con- verted to meat and shell growth. Meat is the primary mar- ketable component and any aquaculture operation should therefore aim to maximise the energy allocated to growth of the flesh. At the same time, where mussels are marketed in their shells, these must be thick enough not to break at the processing stage and must meet the standards in appearance set by the consumer. It is therefore important in the design and management of aquaculture operations to understand the factors in- fluencing resource allocation in bivalve molluscs. Such factors may be genotypic (resulting from genetic differ- ences between the populations) or phenotypic (due to dif- ferent environmental factors acting on genetically similar populations). While there is evidence for several bivalve species that growth-related attributes may be determined by environmental factors (Seed, 1968, 1981; Eager, 1978; Al- drich and Crowley, 1986; Griffiths and Griffiths, 1987), the importance of genetically-determined differences cannot be ruled out. Haley et al. (1975), for example, stated that the growth rate of bivalves is determined geneti- cally and that, in some cases, genetic differences could lead to differences in growth rates within or between popula- tions. Similarly, Hasuo (1958) concluded, from a series of experiments using oysters, that ecological factors alone may not always explain the observed differences in shell *Corresponding author; current address: Department of Zoology, Univer- sity of Oxford, South Parks Road, OX1 3PS, England. shape. Several studies have found correlations between he- terozygosity and various fitness-related parameters in bi- valve molluscs, including growth rates (Koehn and Gaffney, 1984; Hawkins et al., 1986; Skibinski and Ro- derick, 1989), fecundity (Rodhouse et al., 1986), survival (Zouros et al., 1983; Diehl and Koehn, 1985) and meta- bolic efficiency (Koehn and Shumway, 1982; Diehl et al., 1985; Hawkins et al., 1986). Knowledge of the relative roles of genotypic and pheno- typic effects is of potential importance in aquaculture, since it indicates the extent to which the desired morphological attributes can be attained through manipulating the condi- tions under which mussels are grown. Likewise, where en- vironmental effects are demonstrated, it is important to identify the major causal factor or factors impinging on the phenotype (Seed, 1981) so that, where possible, these can be manipulated to meet management objectives. An oppor- tunity to study these relationships for the mussel Mytilus galloprovincialis (Lmk.) was recognised when it was no- ticed that a commercially cultured population grown on ropes in an enclosed sea water dam had thinner shells than those of a naturally occurring population in the subtidal zone outside of the dam. The objectives of this study were, firstly, to determine whether any differences between these populations in rela- tive allocation of resources to meat and shell growth were genotypic or phenotypic. Secondly, we wished to deter- mine whether wave action (the major apparent difference) was an important factor determining the phenotypic effects in the two habitats. This was done by including in the anal- ysis mussels sampled from an additional population growing outside the dam, in a position close to the first but sheltered from direct wave action. MATERIALS AND METHODS To investigate differences in allocation of resources to shell and meat growth between populations growing under 88 RAUBENHEIMER AND COOK different environmental conditions, 30 adult M. gallopro- vincialis (45—95mm in length) were collected from each of three subtidal populations at Saldanha Bay, on the south- western coast of South Africa. The first was a cultured pop- ulation growing on ropes inside a dam (approx. 500m by 700m, with average depth 4m) sheltered from wave action and with limited tidal exchange. A second, wild population grew on rocks outside the dam and was exposed to frequent wave action. The third population also grew outside the dam, approximately 10m from the exposed population, on the sheltered side of a concrete pillar and was thus pro- tected from direct wave action. Samples from both wild populations were taken from approximately 2 to 3m below the low tide mark. The cultured mussels were sampled from a depth of approximately 2m below the surface. Age is a potentially important factor in between-popula- tion comparisons of shell morphology (e.g. Rodhouse et al., 1984), since it could be that beyond a certain age shell thickness increased at the expense of increment in surface area. In such a case thicker shells in one population could be due simply to a greater mean age of the individuals rather than to some population-specific genetic or pheno- typic factor. Acetate peels may be used successfully to esti- mate the age of intertidal mussels which are periodically emersed, but the periodicity of growth bands in subtidal mussels which are permanently immersed may not always be a reliable indicator of age (Seed, 1969; Richardson, 1989). The age of mussels in the wild populations in the present study was therefore not measured, but an attempt was made to take account of this factor by sampling a second population from inside the dam, which was older than the first; the second population grew lower on the same culture-rope as the first (at approx. 3m). This additional, control sample served a dual purpose. Firstly, their larger shell size relative to the exposed popu- lation (Table 1) increased the possibility that the cultured mussels were at least as old as the wild, exposed popula- tion. Secondly, the shells of the older cultured population were no thicker than those of the younger cultured popula- tion (Table 3). This indicates that shell thickening was not a function of age alone, but more likely due to some habitat- specific phenotypic or genotypic factor. Following collection, the mussels were cleaned of all fouling organisms and the separated shell and meat of each was dried in an oven to constant weight (70°C for 24h), and a final weight was taken. The byssus threads were removed before the mussels were weighed, since these are known to vary in weight with aspects of the environment, including the nature of the surface to which they attach (Young, 1983a; 1983b) and degree of exposure to wave action (Price, 1982). To obtain an estimate of shell surface area and shell thickness, the maximum length and height (along the dorsal-ventral axis) of the shells were measured to the nearest 0.1mm using vernier callipers. Surface area was TABLE 1. Multiple comparison of means (+SE) for shell surface area of five populations of Mytilus galloprovincialis. Means followed by different letters are significantly different (P < 0.05; Tukey’s HSD test). Shell surface area Population* (mean + SE cm?) Cy 24.3 + 0.99 Co 29.7 + 1.0° We 23.6 + 1.18 Ws 27.7 + 1.2 T 19.1 + 0.6¢ * Cy = younger cultured population Co = older cultured population We = exposed wild population Ws = sheltered wild population T = translocated population taken to be proportional to the area of the rectangle formed by the product of the maximum height and length of shells. Shell thickness was estimated from the relationship be- tween shell surface area and shell weight. Regression slopes of shell weight on shell surface area were therefore compared for all between-population comparisons of in- terest. Where the null hypothesis of similar slopes was ac- cepted, ANCOVA was used to compare means for shell weight corrected for shell surface area. Where regression slopes differed, the population with the steeper slope was taken to have thicker shells, provided that the regression lines did not cross within the positive range of the co- variate. This prerequisite was confirmed if the y-intercept for the population with the steeper slope was not signifi- cantly lower than that for the population with the smaller slope. This somewhat more complicated analysis was chosen in favour of the comparison of the shell weight/surface area ratios, since for M. galloprovincialis there is an allometric relationship between shell weight and various size param- eters (Hosomi, 1985). Where ratios are used to standardise a variable and the denominator is allometrically related to the numerator, the hypothesis tested may not be the one of interest (e.g., see Packard and Boardman, 1988). How- ever, the statistical outcome of ANOVA on the shell weight/shell surface area ratios did not differ qualitatively from regression analysis reported here. ANOVA was used for statistical comparisons of the meat weight/shell weight ratios. In this case the caveats of Packard and Boardman (1988) do not apply, since rather than standardising the comparisons of meat weight using shell weight, the test was intended to investigate whether the relative resource allocation to meat and shell weight differed between the populations. The ratios were found to be normally distributed using the Kolmogorov-Smirnov test, and were therefore not transformed. Aldrich and Crowley (1986) have recommended the use of this ratio as WAVE ACTION AND GROWTH IN MUSSELS 89 a sensitive condition index for ‘‘studies of local popula- tions, or the reactions of similar stocks of mussels to a va- riety of growth conditions’. To distinguish the purely genotypic and phenotypic con- tributions to between-population differences in shell thick- ness and body size, 100 juvenile mussels (mean length 2.7mm) were collected from the population growing on the exposed rocks outside the dam. They were then transferred to the inside of the dam and grown in nylon net bags, under similar conditions to those in which the mussels in the cul- tured population in the dam were grown. A reciprocal transfer from inside to outside the dam was not made, since it was not possible for technical reasons to establish young mussels on the exposed rocky habitat outside the dam. After a 90-day period, thirty of the translocated mussels were randomly selected, oven-dried and their physical di- mensions determined as described above. These were com- pared with the physical dimensions of the mother popula- tion and of the cultured population in the dam. RESULTS Comparisons between populations for mean shell size are shown in Table 1. Population was significant as a main effect in the ANOVA for shell size (P < 0.001; Fog 145) = 506.7). While the older cultured population had larger shells than the younger, there was no significant difference in shell thickness between these populations (Table 3). This suggests that for this species, at least within the measured size range, increase in shell thickness at the expense of growth in surface area was not a function of age alone. Any differences in shell thickness were therefore most likely a result of habitat-specific phenotypic effects or genetic dif- ferences between populations. The sheltered and exposed populations growing in the open ocean did not differ in the regression slopes of shell weight on shell surface area, but both slopes were signifi- cantly greater than those for populations grown in the sea water dam (Table 2). However, the population exposed to direct wave action had a significantly greater mean shell weight corrected for shell surface area than the sheltered wild population (Table 3). While the regression coefficient for the mussels exposed to wave action was greater than that for the cultured mussels, the y-intercepts did not differ significantly (Table 2). Therefore, increment in shell weight per unit surface area was greater for the exposed population and the regression line for this population was at no point lower than that for the cultured mussels. This indi- cates that mussels exposed to wave action had thicker shells than those cultured in the sheltered dam, but the effect was dependent on shell size. The regression slopes for the younger and older cultured populations did not differ (Table 2), and these populations had similar means for shell weight corrected for shell surface area (Table 3). The sample of mussels transplanted from the exposed popula- TABLE 2. a) Slopes and y-intercepts for the regression of shell weight on shell surface area for populations of M. galloprovincialis growing in different habitats and b) orthogonal pairwise comparisons of populations. Codes used for populations are as in Table 1. a) Population Regression slope Intercept Cy 0.404 —1.13 Co 0.449 Sey We 0.736 — 3:05 Ws 0.640 — 6.70 I 0.285 0.21 b) Slope Intercept prob. prob. Contrast S.E. t of t S.E. t of t Cy vs. Co 0.080 0.55 0.578 2.20 0.36 0.721 Cy vs. We 0.079 4.20 <0.001 1.94 0.99 0.323 We vs. Ws 0.067 1.41 0.158 1.75 2.10 0.380 We vs. T 0.103 4.36 <0.001 2.12 1.54 0.127 tion to inside the dam had a significantly smaller regression slope than the mother population, but the y-intercepts did not differ (Table 2). This demonstrates that the transplanted mussels developed thin shells, similar to those of the cul- tured mussels in the dam. Since these mussels were of the same genetic stock as the thick-shelled population exposed to wave action, it must therefore be concluded that shell thickness in this population was a phenotypic rather than a genetically-fixed attribute. Population was significant as a main effect in the ANOVA of meat weight/shell weight ratios (P < 0.001; TABLE 3. ANCOVA table for comparison of shell weight means corrected for shell surface area. Pairwise comparisons are only made for populations with equal slopes (see Table 2). Source Sum of Mean prob. Contrast squares d.f. square F of F 1. Cy vs Co Regression 281.54 1 281.54 174.01 <0.001 Population 1.98 1 1.98 1223) 0.273 Error 92.34 57 1.62 population: Cy Co adjusted mean: 9.84g 10.26g 2. We vs Ws Regression 1038.76 1 1038.76 235.5 <0.001 Population 503.36 1 503.36 114.14 <0.001 Error 251.37 57 4.41 population: We Ws adjusted mean: 15.71g 9.592 90 RAUBENHEIMER AND COOK TABLE 4. Contrasts for the shell weight/meat weight ratios of M. galloprovincialis sampled from five populations. Significance levels are adjusted for experimentwise error rate using Tukey’s HSD-test. Codes used for populations are as in Table 1. Contrasts Means t-value Sign. population population (a) (b) (a) (b) Cy vs. Ws 0.16 0.26 9.84 <0.05 Cy vs. Co 0.16 0.15 133) ns We vs. Ws 0.12 0.26 11.80 <0.05 We vs. Cy 0.12 0.16 3.70 <0.05 We vs. T 0.12 0.14 2.03 <0.05 Fo4.145) = 61.5). Table 4 shows the pairwise contrasts of means for this analysis. There was no statistical difference between the meat/shell weight ratios for the older and the younger cultured populations. Mussels growing outside the dam and subject to wave action had significantly lower meat/shell weight ratios than both the wild sheltered popu- lation and those growing inside the dam. Similarly, the mussels translocated from the thick-shelled population into the dam had significantly higher meat/shell weight ratios than the parent population. This suggests that in terms of energy partitioning, the mussels growing in the habitat ex- posed to frequent wave action had a higher investment in shell growth than those growing in sheltered habitats. How- ever, the sheltered population growing outside the dam had greater meat/shell weight ratios than the cultured popula- tions. DISCUSSION Our results confirmed that Mytilus galloprovincialis mussels cultured in the sheltered habitat of a sea water dam had thinner shells than those sampled from a subtidal hab- itat exposed to frequent wave action. There were, besides degree of wave action, several possible environmental dif- ferences between the dam and the exposed habitat outside the dam. The dam had limited tidal exchange, and it may therefore be that there were chemical and temperature dif- ferences with the open ocean. Any chemical differences in sea water may result in considerable differences in nutrient availability between the habitats (e.g. Allen, 1934). Nu- trient availability (Page and Hubbard, 1987; His et al., 1989), degree of pollution (Stromgren et al., 1986; Mag- nusson et al., 1988), temperature and to a lesser extent sa- linity (His et al., 1989) may all influence growth patterns in mussels. Some of these factors might account for the dif- ference in regression slopes of shell weight on shell size between the mussels grown in the dam and those in the open ocean, although the mechanisms or reasons for such slope differences are not immediately apparent. Addition- ally, such differences between the water in the dam and the )pen ocean may explain the greater meat/shell weight ra- tios in the wild sheltered mussels than in the cultured popu- lation. However, these factors aloné cannot account for the thicker shells of the population exposed to wave action, since the second wild population, which had thin shells, grew in the same water body as the exposed population and thus served as a control for water quality. There may have been differences in density between the rope-cultured and the free-living populations. These were unavoidable, and difficult to measure, owing to the consid- erable difference in substrate provided by culture ropes as compared to an extended surface. While population den- sities may affect growth rates and shell shape of Mytilus edulis (Seed, 1968), Okamura (1984; 1986a) found that the mussels in her study did not invest differentially in shell material when growing in groups of different density. Fur- thermore, the exposed and sheltered populations judged to be growing outside the dam in the present study were of similar densities. The accumulated evidence therefore sug- gests that degree of exposure to wave action was the major environmental correlate with shell thickness in populations of M. galloprovincialis in the present study. There are sev- eral possibilities regarding the mechanism underlying this response. One possibility is that thick shells are an evolutionary response to natural selection in the exposed population. Such selection may, inter alia, be due to the destructive effects of wave action or the effects of differential preda- tion on thin-shelled mussels (Okamura 1986b; Barkai and Branch, 1989). It has in recent years become clear that there may be appreciable within- and between-population genotypic differences in mussels (Mallet et al., 1986; Ski- binski and Roderick, 1989). However, the mussels in the present study transplanted from the exposed population to the sheltered habitat inside the dam produced thin shells, indicating that the production of thick shells was not a ge- netically fixed trait in the parent population. Furthermore, being broadcast spawners, it seems highly unlikely that sufficient isolation could exist between the two outside populations to bring about the genetic differentiation re- quired for such an evolutionary response (Mayr, 1963). Wilson (1975) has estimated that as little as 10% gene flow per generation is sufficient to counteract “‘fairly intensive natural pressures that tend to differentiate populations’’. A second possibility is selection without evolutionary change (Sober, 1984). In this case, the thick and thin- shelled populations would form a common genetic stock, but in the exposed habitat wave action would eliminate the thin-shelled individuals from each successive cohort of spat settlement, leaving a biased sample of thick-shelled mussels. Williams et al. (1973) suggested that such a phe- nomenon may explain how the American eel, Anquila ro- strata, which spawn in a common area, diverge in different areas of the dispersal range. If this were the case for the mussels in the present study, it could be expected that there would be a narrower variation in shell thickness in the WAVE ACTION AND GROWTH IN MUSSELS 91 thick-shelled populations, since these would be held under strong directional selection for thick shells. This was not the case however, since the widest variation in shell thick- ness existed for the exposed population (shell weight corre- lated more strongly with surface area in the wild sheltered population (r = 0.94) than the exposed population (r = 0.86)). Thirdly, thick shells could be a phenotypic response to some factor associated with the habitat exposed to wave action. This is strongly suggested by the results of the translocation experiment, since the mussels translocated from a thick-shelled population and grown in an area shel- tered from wave action developed thin shells. This explana- tion would predict a wider variation in shell thickness in the population exposed to wave action, since individuals would have experienced different degrees of exposure depending on the micro-habitat in which they were attached. One possibility to account for any phenotypic effects on growth in areas exposed to wave action is that high turbu- lence increased the rate of water flow over the gills thus providing a greater opportunity for filter feeding and gas- eous exchange (Rosenberg and Loo, 1983; Loo and Rosen- berg, 1983). Turbulence in shoreline habitats would also support a greater load of organic particulate matter, and food availability is known to be one of the major factors influencing mussel growth (Widdows et al., 1979; Page and Hubbard, 1987; His et al., 1989). Since the sheltered and exposed wild populations were only 10m apart, both would to some extent benefit from increased oxygenation and suspended organic solids, and this may be an important factor accounting for the high meat/shell weight ratio in the sheltered wild population. However, the low meat/shell weight ratio of the mussels in the exposed population indi- cates that direct exposure to wave action may result in any such advantages due to increased turbulence being chan- neled into increasing shell thickness rather than shell size or meat weight. That differences in shell thickness between the exposed population and the thin-shelled mussels grown in the dam were expressed as a difference in regression slopes, is con- sistent with an environmental effect on phenotype. Seed (1968) found that, despite considerable divergence in the shell shape of older M. edulis taken from different sites, there was little or no difference in form between young mussels (under 2.5mm). From this Seed concluded that “‘the greater differences in form in larger animals is due to the environmental conditions to which the animals have been subjected, and the degree of divergence will be di- rectly proportional to the time spent under those condi- tions’’. It is therefore interesting that the thin-shelled wild population had a similar regression slope to the mussels exposed to direct wave action, with a constant difference in shell thickness. This suggests that there may have been se- lection in the thin-shelled wild population against small mussels with thin shells. A plausible explanation for this may be differential predation on such mussels (Seed, 1969; Worrall and Widdows, 1984; Okamura, 1986b; Barkai and Branch, 1989). There have been several studies in which mussels growing in shoreline habitats have been found to have ro- bust shells relative to those growing suspended in the open ocean. Barkati and Choudhry (1988) found that Perna viridus growing in the intertidal developed heavier shells than those growing on buoys. Coe and Fox (1942) and Fox and Coe (1943) found that M. californianus collected from sheltered pier supports had thinner shells than those col- lected from shores experiencing heavy seas. Aldrich and Crowley (1986) found that the shell weights of M. edulis increased and the meat weights decreased in samples taken from commercial rafts, subtidal and intertidal habitats in that order. Intertidal Perna canaliculus in New Zealand were found by Hickman and Illingworth (1980) to have heavier and thicker shells than those grown on rafts, and Brown and Seed (1977) found that intertidal horse mussels (Modiolus modiolus) had heavier shells than subtidal ones. By contrast, subtidal M. edulis and M. californianus were found by Rao (1953) to have heavier shells than intertidal populations. However, few studies have specifically investigated the effects of wave action. Perhaps this is partially due to the practical difficulties of quantifying wave action, and of performing successful transplant experiments along high- energy coastlines. An exception is to be found in the work of Harger (1970) who developed a simple instrument to measure the strength of wave action, and performed a series of caged transplant experiments using M. edulis and M. californianus. Harger found that growth was slowed and the frequency of growth bands increased in mussels exposed to wave action. From an evolutionary viewpoint, it is not surprising that sessile organisms with a free-floating larval stage should be capable of adapting shell structure to wave action through a facultative, phenotypic response; Smith (1980) has made a similar point about plants. Mussel larvae have relatively little control over the nature of the habitat in which they become established and may thus benefit from the ability to monitor important aspects of the environment, such as de- gree of wave action, and respond in an adaptive way. Larvae having thin shells as a fixed genetic attribute would be disadvantaged in a habitat where protection from the de- structive effects of wave action is required. Similarly, those with a fixed genetic trait for thick shells that settled in a sheltered habitat might invest unnecessary resources in shell structure at the expense of reproductive output. The same is true for the investment made in attachment to the substrate, and Price (1982) has found that the most impor- tant factor determining byssus production by M. edulis is wave action. Indeed, such phenotypic flexibility may be a general characteristic of some mussel species. Mallet et al. (1987), in a series of wide-ranging transplant experiments, 92 RAUBENHEIMER AND COOK found that mortality rates among some stocks of M. edulis did not change when transplanted among different sites, and concluded: ‘‘This implies physiological flexibility or the ability to tolerate a wide range of environmental condi- tions’’. It is interesting to speculate on the mechanisms whereby wave action may affect shell growth. Cycles of emersion and submersion are known to produce growth bands in mussel shells (Seed, 1968, 1969; Harger, 1970; Richard- son, 1989), and it has commonly been observed that mussels that are periodically emersed develop thick shells (Coe and Fox, 1942; Fox and Coe, 1943; Seed, 1968, 1969; Barkati and Choudhry, 1988). Seed (1969) provides the following description of the formation of growth rings: ‘‘During unfavourable external conditions, the mantle edge is slightly withdrawn into the shell, causing the cessation of accretionary growth at the shell margin. Even so, increase in shell thickness continues, and when growth in length is resumed, old and new parts are no longer continuous, and a disturbance ring is formed’’. It might therefore be that bouts of strong wave action cause periods of cessation of feeding, and where these bouts are frequent, thickened growth rings are closely spaced increasing the overall thickness of shells. Furthermore, the availability of calcium for shell growth is governed by the concentration of cal- cium ions in the water and the rate that water flows over the gill surface (Rao, 1953; Wilbur and Jodrey, 1952). Areas of strong wave action may be particularly rich in dissolved calcium due to the abrasion of calcium-rich marine debris, and high turbulence may increase the rate of water flow over the gills. In conclusion, it appears that M. galloprovincialis grown in the open ocean may benefit from the increased flow rate relative to those grown in a sheltered dam with limited tidal exchange. However, for mussels grown in areas subject to frequent wave action, this benefit may not be channelled into increased meat production. Rather, it is predicted that most of the increased scope for growth would be used to increase the thickness of the shell, a trait which beyond some lower limit is unlikely to be desirable to the aquaculture operator. ACKNOWLEDGMENTS We thank Jacky Tonin, Beth Okamura, Sandy Shumway and an anonymous referee for many helpful comments on the manuscript. Many thanks to Phillip Steyn and West- coast Maricult for allowing us use of their facilities and premises. REFERENCES Aldrich, J. C. & Crowley, M. 1986. Condition and variability in Myrilus edulis (L.) from different habitats in Ireland. Aquaculture 52:273- 286. Allen, W. E. 1934. The problem of methods in marine plankton investi- gations. Internationale Revue der gesamten Hydrobiologie und Hy- drographie 31:40—65. Barkai, A. & Branch, G. M. 1989. Growth and mortality of the mussels Choromytilus meridionalis (Krauss) and Aulocomaya ater (Molina) as indicators of biotic conditions. J. Molluscan Stud. 55(3):329-342. Barkati, S. & Choudhry, Y. 1988. Effect of tidal height on growth of mussels. Pakistan J. Sci. Ind. Res. 31(6):415—422. Brown, R. A. & Seed, R. 1977. Modiolus modiolus—an autecological study. Proceedings of the 11th European Marine Biology Symposium, pp 93-100. Coe, W. R. & Fox, D. L. 1942. Biology of the Californian sea mussel Mytilus californianus. 1. Influence of temperature, food supply, sex and age on the rate of growth. J. Exp. Zool. 90:1—30. Diehl, W. J. & Koehn, R. K. 1985. Multiple-locus heterozygosity, mor- tality and growth in a cohort of Mytilus edulis. Mar. Biol. 88:265— ZTAe Diehl, W. J., Gaffney, P. M., McDonald, J. H. & Koehn, R. K. 1985. Relationship between weight-standardised oxygen consumption and multiple-locus heterozygosity in the marine mussel, Mytilus edulis L. (Mollusca). Proceedings of the 19th European Marine Biological Symposium pp 531-536. Eager, M. C. 1978. Shape and function of the shell: a comparison of some living and fossil bivalve molluscs. Biol. Rev. 53:169—210. Fox, D. L. & Coe, W. R. 1943. Biology of the Californian sea-mussel (Mytilus californianus). 11. Nutrition, metabolism, growth and calcium deposition. J. Exp. Zool. 93:205—249. Gniffiths, C. L. & Griffiths, R. J. 1987. Bivalvia. In: T. J. Pandain and F. J. Vernberg (eds.) Animal Energetics, Vol 2 pp 1-89. Academic Press, N.Y. Haley, L. E., Newkirk, G. F., Waugh, D. W. & Doyle, R. W. 1975. A report on the quantitative genetics of growth and survivorship of the American oyster, Crassostrea virginica under laboratory conditions. Proceedings of the 10th European Marine Biological Symposium, pp 221-228. Harger, J. R. E. 1970. The effect of wave impact on some aspects of the biology of sea mussels. Veliger 12:401—414. Hasuo, M. 1958. The shell variation during growth in the Mie pearl oysters and those transplanted from Nagasaki Prefecture, Bull. Natl. Pearl Res. Lab. 4:318—324. Hawkins, A. J. S., Bayne, B. L. & Day, A. J. 1986. Protein turnover, physiological energetics and heterozygosity in the blue mussel Mytilus edulis: basis of variable age-specific growth. Proc. Roy. Soc. (Ser. B) 229:161-176. Hickman, R. W. & Illingworth, J. 1980. Condition cycle of the green- lipped mussel Perna canaliculus in New Zealand. Mar. Biol. 60:27— 38. His, E., Robert, R. & Dinet, A. 1989. Combined effects of temperature and salinity on fed and starved larvae of the Mediterranean mussel Mytilus galloprovincialis and the Japanese oyster Crassostrea gigas. Mar. Biol. 100:455—463. Hosomi, A. 1985. On several fundamental allometries of the mussel My- tilus galloprovincialis. Venus (The Jap. Jour. Malac.) 44(3):172- 182. Koehn, R. K. & Gaffney, P. M. 1984. Genetic heterozygosity and growth rate in Mytilus edulis. Mar. Biol. 81:1—7. Koehn, R. K. & Shumway, S. E. 1982. A genetic/physiological explana- tion for differential growth rate among individuals of the oyster Cras- sostrea gigas (Gmelin.). Mar. Biol. Lett. 3:35—42. Loo, L. O. & Rosenberg, R. 1983. Mytilus edulis culture: growth and production in western Sweden. Aquaculture 35:137—150. Magnusson, K., Berggren, M. & Granmo, A. 1988. Energy budget of caged common mussels (Mytilus edulis) in the vicinity of an industrial centre on the Swedish West Coast. Vatten 44(1):59—64. Mallet, A. L., Zouros, E., Gartner-Kepay, K. E. & Freeman, K. R. WAVE ACTION AND GROWTH IN MUSSELS 93 1986. Genetics of growth in blue mussels: family and enzyme hetero- zygosity effects. Mar. Biol. 92:475—482. Mallet, A. L., Carver, C. E. A., Coffen, S. S. & Freeman, K. R. 1987. Mortality variations in natural populations of the blue mussel. Can. J. Fish. Aquat. Sci. 44(9):1589-1594. Mayr, E. 1963. Animal Species and Evolution. Harvard University Press, Cambridge, Mass. Okamura, B. 1984. Functional and ecological consequences of aggrega- tion and coloniality in benthic, marine suspension feeders. PhD. dis- sertation, University of California, Berkeley, CA. Okamura, B. 1986a. Group living and the effects of spatial position in aggregations of Mytilus edulis. Oecologia 69:341—347. Okamura, B. 1986b. Formation and disruption of aggregations of Myrilus edulis in the fouling community of San Francisco Bay, California. Mar. Ecol. Prog. Ser. 30:275—282. Packard, G. C. & Boardman, T. J. 1988. The misuse of ratios, indices and percentages in ecophysiological research. Physiol. Zool. 61(1): 1-9. Page, H. M. & Hubbard, D. M. 1987. Temporal and spatial patterns of growth in mussels Mytilus edulis on an offshore platform: relationships to water temperature and food availability. J. Exp. Mar. Biol. Ecol. 111:159-179. Price, H. A. 1982. An analysis of factors determining seasonal variation in the byssal attachment strength of Mytilus edulis. J. Mar. Biol. Ass. U.K. 62:147-155. Rao, K. P. 1953. Shell weight as a function of intertidal height in a littoral population of pelecypods. Experientia 9:465—466. Richardson, C. A. 1989. An analysis of microgrowth bands in the shell of the common mussel Mytilus edulis. J. Mar. Biol. Ass. U.K. 69:477— 491. Rodhouse, P. G., Roden, C. M., Burnell, G. M., Hensey, M. P., McMahon, T., Otway, B. & Ryan, T. H. 1984. Food resource, ga- metogenesis and growth of Mytilus edulis on the shore and in sus- pended culture: Killary Harbour, Ireland. J. Mar. Biol. Ass. U.K. 64:513-529. Rodhouse, P. G., McDonald, J. H., Newell, R. I. E. & Koehn, R. K. 1986. Gamete production, somatic growth and multiple-locus enzyme heterozygosity in Mytilus edulis. Mar. Biol. 90:209—214. Rosenberg, R. & Loo, L. O. 1983. Energy-flow in a Mytilus edulis cul- ture in western Sweden. Aquaculture 35:151-161.; Seed, R. 1968. Factors influencing shell shape in Mytilus edulis. L. J. Mar. Biol. Ass. U.K. 48:561—584. Seed, R. 1969. The ecology of Mytilus edulis L. (Lamellibranchiata) on exposed rocky shores. II. Growth and mortality. Oecologia 3:317— 350. Seed, R. 1981. Shell growth and form in the Bivalvia. In: D. C. Rhoads & R. A. Lutz (eds). Skeletal Growth of Aquatic Organisms. Biolog- ical Records of Environmental Change. Plenum Press, N.Y. Skibinski, D. O. F. & Roderick, E. E. 1989. Heterozygosity and growth in transplanted mussels. Mar. Biol. 102:73—84. Smith, R. L. 1980. Ecology and Field Biology. Harper and Row, New York. Sober, E. 1984. The Nature of Selection—Evolutionary Theory in Philo- sophical Focus. MIT Press, Cambridge. Stromgren, T. Nielsen, M. V., & Ueland, K. 1986. Short-term effects of microencapsulated hydrocarbons on shell growth of Mytilus edulis. Mar. Biol. 91:33-39. Widdows, J., Fieth. P., & Worrall, C. M. 1979. Relationship between seston, available food and feeding activity in the common mussel My- tilus edulis. Mar. Biol. 50:195—207. Wilbur, K. M. & Jodrey, L. H. 1952. Studies on shell formation. I. Mea- surement of the rate of shell formation using Ca**. Biol. Bull. 103:269—276. Williams, G. C., Koehn, R. K. & Mitton, J. B. 1973. Genetic differen- tiation without isolation in the American eel, Anquila rostrata. Evolu- tion 27:192—204. Wilson, E. O. 1975. Sociobiology: The New Synthesis. Cambridge: Har- vard University Press. Worrall, C. M. & Widdows, J. 1984. Investigation of factors influencing mortality in Mytilus edulis L. Mar. Biol. Lett. 5:85—97. Young, G. A. 1983a. The effect of sediment type upon the position and depth at which attachment occurs in Mytilus edulis. J. Mar. Biol. Ass. U.K. 63:641-651. Young, G. A. 1983b. Response to, and selection between, firm substrata by Mytilus edulis. J. Mar. Biol. Ass. U.K. 63:653—659. Zouros, E., Singh, S. M., Foltz, D. W. & Mallet, A. L. 1983. Post-set- tlement viability in the American oyster (Crassostrea virginica): an overdominant phenotype. Gen. Res. 41:259—270. Journal of Shellfish Research, Vol. 9, No. 1, 95-101, 1990. CONTAMINATION OF THE MUSSEL, MYTILUS EDULIS LINNAEUS, 1758, BY ENTERIC BACTERIA A. PLUSQUELLEC*, M. BEUCHER*, D. PRIEUR**, Y. LE GAL*** *Institut Universitaire de Technologie—B.P. 319 29191 QUIMPER (France) ** Station Biologique—29211 ROSCOFF (France) ***aboratoire de Biologie Marine, Collége de France—29110 CONCARNEAU (France). ABSTRACT Contamination of the mussel Mytilus edulis by sewage pollution bacteria was followed both in a laboratory system and in natural conditions. When mussels are exposed to a bacterial contamination, a maximal level in mussel tissue contamination is rapidly achieved. The bacterial enrichment in mussel tissue versus seawater contamination differs according to the bacterial group considered, and is higher with Enterococci group. This factor is not dependent on the bacterial density in the water, but in contrast it is directly influenced by the particle density in the water. In situ observations confirmed that the enrichment in mussel tissue differs between bacterial groups. Seasonal differences between contamination levels and enrichment factors in the mussel have been pointed out. KEY WORDS: Among sanitary aspects of bacterial pollution in the ma- rine environment the most clearly established hazards are those which result from the consumption of contaminated shellfish. Bacteria generally involved in epidemic out- breaks are enteric bacteria such as Salmonella, Vibrio cho- lerae, Escherichia coli, and for some geographic areas an halophilic bacterium Vibrio parahaemolyticus (Bryan 1980). The most serious shellfish poisoning events are imputed to bivalves and especially to filter feeders such as oysters or mussels. Because of their capacity to filter large volumes of water, these shellfish are able to concentrate a large number of particles and bacteria from environmental water (Fleet 1978). So, in a polluted area, enteric bacteria may be con- centrated in a viable state in the bivalve flesh. This concen- tration of enteric pathogens or indicators is a general obser- vation described for different bivalves, different bacteria, and several geographic areas (Trollope 1984). Neverthe- less, most of the data have been obtained in situ in non- controlled conditions and have only compared pathogen or indicator levels in mollusc meat and in seawater without any indication about the accumulation process. Alterna- tively, some works concern bivalve and bacteria relation- ships but without reference to enteric bacteria (Prieur 1981). However, essential data concerning kinetics of ac- cumulation, of depuration and incidence of environmental parameters have been obtained by Cabelli and Heffernan (1970a, 1970b) in experimental conditions with Merce- naria mercenaria and Mya arenaria. The present study aims to state more precisely the mech- anisms of ingestion, retention and elimination of fecal bac- teria by marine bivalves. To this end accumulation and elimination of enteric bacteria by the mussel Mytilus edulis have been studied both in a laboratory microcosm and in an area exposed to domestic sewage. 95 Mussel, Mytilus edulis, sewage bacteria, Coliforms, Enterococci, bacterial enrichment. In standard laboratory conditions, it is possible to con- trol the variation and thus to estimate the effect of some parameters such as the nature of bacterial contamination, initial bacterial concentration in seawater, particle load. Study was performed in situ, to attempt to verify labora- tory data in natural conditions and to estimate incidence of seasonal factors. EXPERIMENTAL PROCEDURES The mussel (Mytilus edulis) was used for experimental assays, both in laboratory and in situ studies. Mussels came from long line cultures and were depurated by immersion in running pure seawater during five days before experiments. Laboratory Studies Mussels were set over a grid into a 100 litre aquarium aerated and stabilized at 19°C. This tank and a control tank without mussels were simultaneously inoculated with a bacterial suspension. Bacteria used for inoculations included cultures of the most usual indicators of fecal pollution: Escherichia coli and Streptococcus faecalis and of an enteric pathogen Salmonella anatum (These three strains had been pre- viously isolated from indigenous mussels). Bacteria cultured on Nutrient Agar slants were sus- pended in sterile seawater. Optical density (650nm) of the suspensions was adjusted at 0.12 corresponding to a den- sity of 1 to 3 108 CFU per millilitre, and then adequately diluted. For inoculation bacteria were added to a suspension of domestic sewage sterilized by autoclaving, which is the source of particles (final concentration of sewage in the tank was 1%c). 96 PLUSQUELLEC ET AL. In situ Studies Bacteriological survey in natural conditions was per- formed at Concarneau (France) in an area subjected to urban pollution. Mussels were placed in nylon nets and in- troduced 48 hours before the beginning of the survey on the station at two meters depth. Every day a net of mussels and a one litre flask of water were sampled. Contamination of mussels and water were followed in parallel over four seasons, May—June 1984, October—No- vember 1984, February 1985, May—June 1985, through the enumeration of fecal coliforms and Enterococci in sea- water and mussel tissue. Bacteriological Methods For both in situ and laboratory measurements mussels were washed under running seawater, briefly flamed and opened. The liquid contained in the mussel shell was sepa- rately collected by dripping. The flesh was then collected, diluted in sterile saline solution (dilution = 4), and sus- pended by “‘Stomacher’’ blendor for bacteriological anal- ysis. The same homogenization was applied to other samples: seawater and mussel liquid. Each sample of mussel liquid and mussel flesh was resulting from the opening of ten mussels. Enumeration of E. coli and Streptococcus faecalis in bi- valve samples was performed by the poured plate method which provides a sensitivity higher than the spread plate technique (Al Jebouri & Trollope 1981). Salmonella were enumerated by the spread plate method according to the procedure recommended by Hawa et al. (1984). Enumeration in seawater was performed by the poured plate method or the membrane filtration technique ac- cording to the bacterial concentration expected. Media used for enumeration were the following: —E. coli or fecal coliforms: MacConkey agar (oxoid CM7) incubated at 44°C 24 hours; — Streptococcus faecalis or Enterococci group: Slanetz Bartley agar (oxoid CM377) incubated at 37°C 24 hours; — Salmonella anatum: Dulcitol Bile Novobiocin agar (Hawa et al. 1984) inoculated by spread plate method and incubated at 37°C 24 hours. LABORATORY RESULTS Mussel Contamination Pattern After similar contamination in the tank with mussels and the control tank by a mixture of E. coli and sewage, bacte- rial concentration was followed during 46 hours in mussel flesh, in mussel liquid, in experimental tank water and in control tank water. A typical contamination profile may be defined (Figure la). When mussels are exposed to a sudden bacterial input, the bacterial concentration in the flesh (Ny) increases rap- idly and remains at a maximum level for a period before « Mussel flesh(NMF) LogN mt" > Mussel Liquid(Nm_ ) > Water (Nw) Enrichment (Log E ) e—Water T-0 4 T(mn) Figure 1. A) Mussel contamination pattern (experimental contamina- tion by E. coli), B) Contamination of mussel tissue by E. coli during the starting contamination period. 100 200 decreasing with a rate similar to the bacterial reduction in the water of the tank (Ny). After the first accumulating phase, the contamination in mussel flesh remains higher than in the surrounding water (Nw). The difference between these values represents the bacterial enrichment factor. This factor corresponds to the accumulation factor defined by Cabelli and Heffernan (1970a) as the ratio E = Ba/Bw where Ba is the bacterial level per gram of shellfish tissue and Bw the bacterial level per millilitre of water. Thus log E is the difference between the log value of counts in bivalve meat and seawater. This factor remains relatively constant at about 10 (log E about 1) after the up- take is established. Contamination of mussel liquid (Ny) is generally within the limits of concentration in the mussel flesh and water concentration, but quite irregular. The kinetics of bacterial accumulation in mussel flesh has been more accurately determined by counts between 0 and 200 minutes. Maximal concentration in the flesh is rap- idly obtained within 30 minutes (Figure 1b). Comparison between Bacterial Strains Mussels have been contaminated by simultaneously adding individual bacterial suspensions of E. coli, Strepto- CONTAMINATION OF THE MUSSEL BY ENTERIC BACTERIA 97 coccus faecalis, and Salmonella anatum, which were at similar initial concentrations. Then a selective enumeration of these bacteria was performed during four hours. For those three strains contamination in mussel flesh follows the same profile but the enrichment factor differs between the two enteric gram negative rods (E = 10) and Strepto- coccus faecalis (E = 38) (Figure 2). Other studies in this laboratory using identical condi- tions (Plusquellec et al. 1986) have enabled to obtain a mean value of the enrichment factor (E) for each of these three strains. They are given in Table 1. Influence of Bacterial Concentration in Seawater on Mussel Flesh Contamination To determine the role of the bacterial load in the water, four lots of mussels (A to D) were exposed simultaneously to various levels of contamination by E. coli. Initial bacte- rial densities of seawater in tanks were ranging from 4.10! ml~! (A) to 3.107 ml~! (D) with a constant particle den- sity, obtained by addition of sterilized sewage (one part of sewage per 1000 parts of seawater). Figure 3a displays that in tank A, where bacterial density is low, the contamination in mussel is not according with the pattern described pre- viously: after an accumulation stage the contamination in mussel tissue decreases progressively until less than the bacterial level in the water. Above a density of 10% E. coli ml~! in the water, the profiles described in Figure la are obtained (Figure 3a, B, C, D). The calculated enrichment factors obtained are respectively: tank A: 1.2; tank B: 7.0; tank C: 4.92; tank D: 6.0. Thus above a minimal level of contamination in the water, the enrichment in mussel tissue is not clearly influenced by the bacterial concentration in the water. Even a massive bacterial concentration of 3.107 ml~! is not a limit for accumulation process by mussels. Influence of Particle Concentration in Seawater The previous experiment was accomplished with a con- stant particle density (0.1% of sterilized sewage). In con- LogN mt"! & a Escherichia coli © # Salmonella anatum o e Streptococcus faecalis SS UY Safes T(mn) 100 200 Figure 2. Comparison between bacterial strains (Water T = 0 indi- cates the bacterial densities at T = 0). TABLE 1. Mean values of enrichment factors observed during laboratory contamination of mussels by bacterial suspensions. Bacterial Escherichia Streptococcus Salmonella Strain coli faecalis anatum E 9.8 33.1 13.8 S 1.9 1.6 2.3 n 11 8 5 s = standard deviation E = Mean enrichment factor S n = number of experiments trast a second experiment was performed with a range of sewage particle densities. The mixture bacteria plus sterilized sewage was added to three tanks respectively in the proportion of 0.001% (tank A’), 1% (tank B’), 10% (tank C’). Results presented in Figure 3b show that the concentration process is directly dependent on particle load in the water. The maximal en- richment in mussel flesh (E = 9) is obtained for the lowest particle density (tank A’). In contrast, no concentration ef- fect in mussel flesh is observed in presence of water with Log N [ag a ae D Water D Oo, Mussels C 4 “Mussels B 3 WaterB : as Mussels A T(mn) 100 200 4 er Cee eee / Mussels C / ee a ea Musselsis 4 ater B / Mussels A 100 200 T(mn) Figure 3. A) Influence of E. coli density in seawater on mussel flesh contamination. (A, B, C, D: see text). B) Influence of particle density in seawater on mussel flesh contamination. (A’, B’, C’: see text). 98 PLUSQUELLEC ET AL. high particle density (tank C’: E = 0.7). For a middle par- ticle density (tank B’), E value is 4.9. Elimination of Bacteria by Mussels The previous experiments show that a rapid uptake of bacteria is operated by mussels exposed to a sewage input. What happens, in contrast, when contaminated mussels are placed in pure running water? Following a contamination period of three hours, mussels were transfered to a tank with pure running sea- water and decontamination in the mussel flesh was fol- lowed. A progressive elimination of bacteria was observed (Figure 4) with closely related rates for E. coli and Salmo- nella anatum. For Streptococcus faecalis, a similar disap- pearance was observed in the first stage, but was followed by a persistence stage. A four day period is necessary to obtain a depuration down to undetectable levels in the case of E. coli and Salmonella anatum. This period is not suffi- cient for a complete elimination of Streptococcus faecalis. IN SITU RESULTS The quantitative evaluation of the concentration of the most usual indicator groups (Fecal coliforms and Entero- cocci) has been conducted in parallel in the meat of captive mussels and in the surrounding seawater. The results were expressed for 100 ml of seawater or 100 ml of mussel tissue. The results of this daily survey are represented in Figure 5. Enumeration of Fecal coliforms and Enterococci in water and mussels as functions of time shows that the counts in mussel flesh are constantly higher than in sur- rounding water. There is obvious parallelism between the simultaneous bacterial counts in seawater and in mussel flesh. It was confirmed by regression analysis of mussel contamination versus seawater contamination. The relation was clear LogN m-" LogN 100 mi FECAL COLIFORMS e mussels e seawater « seawater Figure 5. Results of daily counts of fecal coliforms and Enterococci in seawater and mussel flesh. (A: May-June 1984; B: October—No- vember 1984; C: February 1985; D: May-June 1985). especially in the case of fecal coliforms (Plusquellec et al. 1986). After logarithmic transformation it was confirmed that each series of samples was following a normal distribution. Thus, each group of samples can be characterized by its mean (x) and standard deviation (s): these values and the coefficients of variation (CV) are presented in table 2. Student’s t test applied on these values established clearly that in each case the mean obtained in mussels is significantly higher than the mean in seawater. The more marked differences correspond to the enumeration of En- terococci. For an analysis of seasonal variations autumn and winter results have been grouped and compared to summer results. The contamination levels, the variability of the results and the value of enrichment factor obtained in mussel flesh have been considered in this inter-seasonal comparison. The results are presented in Table 3. seawater mussels Escherichia coli a ce 6 Salmonella anatum . Bae 5 vy Streptococcus faecalis e O--O dl 0 10 20 30 Figure 4. Elimination of bacteria from contaminated mussel tissue. 40 | indicates the transfer to pure running seawater. CONTAMINATION OF THE MUSSEL BY ENTERIC BACTERIA 99 TABLE 2. Mean values of daily counts in seawater and mussel flesh over four campaigns. Seawater Mussels Fecal Fecal coliforms Enterococci coliforms Enterococci A (n = 24) x 2.65 (4,4.107) 1.61 (4,0.10') 4.10 (1,2.10*) 4.65 (3,5.104) s 0.57 1.18 0.94 0.57 CV 21.5% 73.3% 22.9% 12.5% B (n = 7) x 3.25 (1,7.103) 3.11 (1,3.103) 4.51 (3,2.104) 4.50 (3,1.104) s 0.64 0.61 0.49 0.51 CV 19.7% 19.8% 11.0% 11.4% C (n = 21) x 3.31 (2,0.103) 2.75 (5,6.107) 4.58 (3,8.104) 4.48 (3,0.10*) s 0.29 0.33 0.33 0.23 CV 8.7% 12.0% 7.2% 5.1% D (n = 17) x 2.73 (5,3.102) 2.10 (1,2. 102) 3.94 (8,7.10) 4.09 (1,2.104) s 0.43 0.42 0.49 0.39 CV 15.7% 20% 12.4% 9.5% All samples grouped (n = 69) x 2.92 (8,3.107) 2.23 (1,7.102) 4.25 (1,7.104) 4.38 (2,4.104) S 0.66 0.97 0.67 0.45 CV 22.6% 43.5% 15.7% 10.2% A: May—June 1984; B: October-November 1984; C: February 1985; D: May—June 1985. xX: average of logarithmic counts in 100 ml of seawater or 100 ml of mussel tissue. In brackets geometric average of these counts. s: standard deviation n: number of samples CV: coefficient of variation = s - 100/x Some observations result from this comparative survey: — Fecal coliform levels are significantly higher in au- tumn—winter than in summer both in mussels and water. This seasonal difference does not appear to occur in the case of Enterococci. — Daily variations are more pronounced in summer than in autumn—winter. This observation is general for the two sampling materials and the two indicator groups. — In contrast, these two bacterial groups differ with re- gard to the enrichment factor in mussels. This factor is much higher for Enterococci than for fecal coli- forms. In addition, these two groups differ in the re- sponse to seasonal influence. A constant enrichment factor is observed in the case of fecal coliforms but, on the contrary, this factor varies highly with the season in the case of Enterococci. DISCUSSION The accumulation of enteric bacteria by the mussel My- tilus edulis appears through the present study as a regular phenomenon, which has been observed in a laboratory system as well as in natural conditions. Accumulation in the shellfish tissue was demonstrated with different batches of mussels, different bacterial groups, whatever the season. The contamination pattern of mussels exposed to bacte- rial contamination in standard laboratory conditions is uni- form with Escherichia coli, Streptococcus faecalis, and Salmonella anatum. It is characterized by a quite rapid ac- cumulation of bacteria in mussel flesh. Trollope and Al Sa- lihi (1984) estimated that an immersion period of two hours was satisfactory to obtain a steady contamination in mussels exposed at a coastal station. It is demonstrated here that under optimal conditions (of temperature and aeration), the maximal concentration may be achieved in a period as short as thirty minutes. This rapid increase to a maximal contamination sup- poses that mussels are homogeneous in their activity to concentrate the bacterial contamination. A variation in the response of individual bivalves has been described by Hef- fernan and Cabelli (1971) in the case of Mercenaria mer- cenaria. In the present study, where mussel flesh samples were resulting from ten mussels, an individual variation would have been indicated by a contamination under the maximal level. Thus it can be considered that, at the time corresponding to the maximal contamination in mussels, the totality of the animals have been contaminated. A difference in the enrichment in mussels, according to the bacterial group is pointed out by the laboratory experi- ments. This difference is clear between enteric rods and 100 PLUSQUELLEC ET AL. TABLE 3. Seasonal comparison of seawater and mussel contamination, and of enrichment in mussels. Seawater Mussels Enrichment Fecal Fecal coliforms Enterococci coliforms Enterococci FCS ES FCM EM FCM — FCS EM — ES Summer earl x 1.84 x 4.03 x 4.33 1.33 2.49 n= 41 s 0.50 s 0.93 s 0.76 s 0.53 0.57 0.89 E = 21,3 E = 309 Autumn winter x 3.29 x 2.84 x 4.56 x 4.49 1.26 1.64 n = 28 s 0.39 s 0.43 s 0.37 s 0.31 0.39 0.33 E = 18,2 E = 43,6 F test 1.66 4.77 4.38 3.11 P3113} 7.18 NS EK **KX ** * EK t test 5.05 0.91 3.55 1.68 1.05 5.02 **X NS **K NS NS KX = number of samples = average of logarithmic counts in 100 ml = standard deviation E = enrichment in mussel tissue. n x n F test and t test: NS = non significative; * = significative (p = 0.05); ** = very significative (p = 0.01); *** = very highly significative (p = 0.001). Enterococci, and may result from a more efficient trapping of the chains of cocci on account of their greater length. This hypothesis is supported by experiments of contami- nation as a function of the bacterial density, and the particle density in seawater. It follows from these experiments that the bacterial accumulation is more influenced by the par- ticle density than by the bacterial density. These results are consistent with Cabelli and Heffernan, who concluded (1970a) that in Mercenaria mercenaria the accumulation of E. coli is dependent not on the bacterial concentration in the water, but on the ratio E£. coli to the total ingestible particles. Following this hypothesis, one can assume that Enterococci chains are assimilated as ingestible particles more than the smaller size rods (E. coli, Salmonella). Furthermore Enterobacteriaceae and Enterococci showed differences in their depuration rates. The elimina- tion of E. coli and Salmonella anatum is quite similar, as previously described in the Sydney rock oyster (Rowse and Fleet 1984). In contrast, elimination of Enterococci is achieved more slowly. In this case, a longer persistence of Enterococci in the mussel may likely be due to a higher resistance of these bacteria to the selective conditions en- countered in the bivalve. Decrease rates observed in these laboratory experiments were about 100 times within 24 hours, and are consistent with many data concerning self cleansing by bivalves (Hef- fernan and Cabelli 1970; Perkins et al. 1980; Timoney and Abston 1984). The comparative survey of the bacterial contamination in seawater and in captive mussels under natural conditions emphasizes the reality of accumulation of enteric bacteria in mussel tissue. Bacterial counts in mussel flesh are con- stantly higher than the counts in surrounding seawater. When all paired samples are grouped (n = 69) the geo- metric mean factor of enrichment in mussels and their stan- dard deviation are respectively: E = 20.3 + 3.1 for fecal coliforms E = 125 + 6.5 for Enterococci. These values are similar to the enrichments obtained previously in another site in Concarneau Bay with indige- nous mussels (Plusquellec et al. 1983), and the values cited by Delattre and Delesmont (1981), concerning captive mussels and cockles in North Sea. A comparison with the values obtained with bacterial cultures in a closed laboratory system (Table 1) shows that enrichment factors observed in situ are superior to enrich- ment factors obtained in these experimental conditions. Furthermore the difference between Enterobacteriaceae and Enterococci are more marked in situ. This should be the consequence of both the higher uptake and the slower elim- ination observed for Enterococci in laboratory experiments. Seasonal variations in the coliform enrichment in oysters have been reported (Hussong et al. 1981); they were not observed in the present study. In contrast, the accumulation of Enterococci is much more pronounced in summer than in autumn—winter. This may be related to differences in the mussel activity or differences in particle density in the sea- water. During the winter campaign, low temperatures were ob- served in seawater (from 4.5°C to 9°C). The results shown in Figure 5 and Table 2 do not indicate that accumulation in mussels was affected by those low temperatures. Accumulation of bacteria within bivalve tissues may be CONTAMINATION OF THE MUSSEL BY ENTERIC BACTERIA 101 considered in another aspect, namely, the value of bivalves as sampling material in monitoring seawater bacterial con- tamination. As a result of their ability to concentrate and retain bacteria, filtering molluscs can be used as indicators for the detection of pathogens or for quantitative evaluation of water pollution (Trollope 1984, Plusquellec et al. 1983). Some data of the present study support this potential use. The monitoring of water pollution by the means of mussels offers an increased sensitivity due to the enrichment in the bivalve tissue and a decrease in variability (Table 2). Espe- cially in the case of fecal coliforms this enrichment is con- stant and not affected by seasonal variations. These results emphasize the potentialities of bivalves as bioindicators in seawater pollution surveys. LITERATURE CITED Al Jebouri M. M. & Trollope D. R., 1981. The Escherichia coli content of Mytilus edulis from analysis of whole tissue or digestive tract. J. of Appl. Bacteriol. 5:135-142. Bryan F. L. 1980. Epidemiology of foodborne diseases transmitted by fish, shellfish and marine crustaceans in the USA. 1970-1978 Journ. of Food Protect. 43(11):859-876. Cabelli V. J. & Heffernan W. P. 1970a. Accumulation of Escherichia coli by the nothern quahaug. Appl. Microbiol. 19(2):239—244. Cabelli V. J. & Heffernan W. P. 1970b. Elimination of bacteria by the soft shell clam Mya arenaria. Journ. Fish Res. Board Canada 27:1579—1587. Delattre J. M. & Delesmont R. 1981. L’analyse des coquillages peut-elle servir au controle microbiologique du littoral. Rev. Int. Océanog. Medic. 43-44, 11-16. Fleet G. H. 1978. Oyster depuration: a review. Food Technol. Austr. 30:444—454. Hawa S. G.; Morrison G. J. & Fleet G. H., 1984. Method to rapidly enu- merate Salmonella of chicken carcasses. Journ. of Food Protect 47(12):932-936. Heffernan W. P. & Cabelli V. J. 1970. Elimination of bacteria by the nothern quahaug Mercenaria mercenaria; environmental parameters significant to the process. J. Fish Res. Board Canada 27:1569— 1577. Heffernan W. P. & Cabelli V. J. 1971. The elimination of bacteria by the nothern quahaug: variability in the response of individual animals and the development of criteria. Proceed of the Nat. Shellfish Assoc. 61:102—108. Hussong D.; Colwell R. R. & Weiner R. M. 1981. Seasonal concentra- tion of Coliform bacteria by Crassostrea virginica, the eastern oyster in Chesapeake Bay. Journ. of Food Protect 44:201—203. Perkins F. O.; Haven D. S.; Alamo R. M. & Rhodes M. W. 1980. Up- take and elimination of bacteria in Shellfish. Journ. of Food Protect. 43(2):124-126. Plusquellec A.; Beucher M. & Le Gal Y. 1983. Enumeration of the bacte- rial contamination of bivalves in monitoring the marine bacteria pollu- tion. Marine Poll. Bull. Vol. 14(7):260-263. Plusquellec A.; Beucher M.; Quemere M.; Le Gall M. & Petillon C. 1986. Etude des modalités de contamination bactérienne des moules. Contrat Sretie N°84 298. Ministére de |’ Environnement (France). Prieur D. 1981. Experimental studies of trophic relationships between ma- rine bacteria and bivalve molluscs. Kieler Meeres. Sonderh. 5:376- 383. Rowse A. J. & Fleet G. H. 1984. Effects of water temperature and sa- linity on elimination of Salmonella charity and E. coli from Sydney Rock oysters (Crassostrea commercialis). Appl. and Environ. Micro- biol. 48(5):1061—1063. Timoney J. F. & Abston A. 1984. Accumulation and elimination of E. coli and Salmonella typhimurium by hard clams in an in vitro system. Appl. and Environ. Microbiol. 47(5):986—988. Trollope D. R. & Al Salihi S. B. S. 1984. Sewage derived bacteria moni- tored in a marine water column by means of captive mussels. Marine Environ. Research 12:311—322. Trollope D. R. 1984. Use of molluscs to monitor bacteria in water. Micro- biological Methods for Environmental Biotechnology ISBN 012.29.5040.2. Society for Applied Bacteriology. Journal of Shellfish Research, Vol. 9, No. 1, 103-108, 1990. GEOGRAPHIC AND SEASONAL VARIATION OF OKADAIC ACID CONTENT IN FARMED MUSSELS, MYTILUS EDULIS LINNAEUS, 1758, ALONG THE SWEDISH WEST COAST J. HAAMER!, P.-O. ANDERSSON?, O. LINDAHL?, S. LANGE‘, X. P. LI*, AND L. LEDEBO4 ‘Department of Physical Oceanography Univ. of Géteborg Box 4038 S-40040 Goteborg 2P| 6147 Kémpersvik S-45081 Grebbestad 3Kristineberg Marine Biological Station S-45034 Fiskebdckskil 4Department of Clinical Bacteriology University of Goteborg Guldhedsgatan 10 S-41346 Goteborg, Sweden ABSTRACT Similar to the observation period August 1987—April 1988 a continuous period of accumulation of okadaic acid (OA) in the mussels in the north (N 58°37’ and further north) started in the beginning of October 1988, and the OA concentrations remained elevated until February 1989. OA levels up to 218 jg per 100 g mussel meat were recorded. During this period mussels further south in the Orust-Tj6m-Lyr6n area (N 58°04—06’) consistently showed lower levels, usually below 20 xg OA per 100 g mussel meat. In May, June and July 1988 the levels were below 10 wg OA per 100 g mussel meat at all sampling sites. Large differences were observed between closely located sites. In late November 1988, when one site located far out in the archipelago showed 218 yg OA per 100 g mussel meat, a satellite farm situated ca one nautical mile further towards the mainland showed 63 pg. In January 1989, when the OA-levels were fairly stable, mussels were collected all along the coast and analyzed. Low OA levels (<10 yg OA per 100 g mussel meat) were found in mussels grown close to the mainland, particularly where fresh water was discharged into the sea. Moderately high levels (ca. 50 wg OA per 100 g) were found quite south (N 57°35’) in mussels from the outer archipelago. The highest levels (134 zg OA per 100 g) were observed in mussels from the open archipelago at N 58°18’. The observations indicate that toxic Dinophysis sp. carry the OA to the mussels from the open sea. Coastal water with lower salinity may antagonize the OA contamination of the mussels by producing unfavourable conditions for OA synthesis and by sup- porting the growth of non-toxic algae. KEY WORDS: okadaic acid, mussels, Mytilus edulis, Sweden INTRODUCTION The Swedish mussel farming area is situated where water-masses of different origin mix. On the latitude of N58° highly saline water from the Atlantic-North Sea-Ska- gerrak area, which is transported eastwards towards the Swedish coast by the Jutland Current, meets low-saline water of the Baltic current (Fig. 1) which is transported northwards as a surface current along the coast. The mean fresh water supply to the Baltic is ca 15000 m/s which means that seasonal salinity variations of the Baltic Current have strong influences on the salinity of the mussel farming waters. Furthermore, weather conditions can totally change the directions of the currents. In the autumn 1983 diarrheic shellfish poisoning (DSP) was recognized in Sweden in people who had consumed blue mussels (Mytilus edulis). Before this fresh mussels harvested from the wild as well as from several mussel 'To whom correspondence is to be addressed. growing farms had been consumed without any certain DSP-cases being reported to the health authorities. How- ever, the DSP-condition was not recognized in Sweden be- fore 1983 and might have been misdiagnosed as bacterial diarrhoea. Since 1983 the mussels along the Swedish west coast have been periodically contaminated with diarrheic shellfish toxin (DST) each year. The DST found in Sweden is almost exclusively okadaic acid (OA) (Kumagai et al. 1986). Its presence in mussels coincides with the presence of Dinophysis spp. in the sea water (Dahl and Yndestad 1985, Yasumoto et al. 1980). In 1984, when there was a protracted suspension in the harvesting of mussels due to the occurrence of DSP, a pro- gram for analysis of the toxic condition of the mussels was started for DSP and PSP. Even if the awareness of toxic algal blooms has increased in recent years, the toxic species Gymnodinium catenatum has been found in sediments 400 years old (Nordberg and Bergsten 1988). Thus, modern human effects on the environment is not necessarily the cause of newly recognized toxic algal blooms. It is, how- 103 104 HAAMER ET AL. LI son Jyltland- current tenet Baltic current | Figure 1. A simplified map of the surface currents affecting the mussel-growing area. The Jutland Current which carries water with high salinity hits the Swedish coast close to N58° where it mixes with the Baltic Current and turns north. The water of the Baltic Current is of lower salinity, the lowest values seen in the spring and early summer. Strong west-northwest winds may turn the currents south which raises the salinity in the mussel-growing area (map from Artur Svansson 1975, with permission). ever, important to realize that, the pursuit of fishing and aquaculture requires management of the marine biotoxins. MATERIALS AND METHODS Cultivation of mussels. The mussels are cultivated ac- cording to the long-line method, a system of steel wires kept suspended by buoys. From the wires 5 cm wide and ca 7 m long plastic ribbons are hanging down. Mussel larvae settle spontaneously on the ribbons and grow until harvest. On the ribbons mussels usually grow in between depths of 1.5 and 12 m. This implies that mussels on one ribbon may originate from water masses with large differences in sa- linity, temperature and plankton content. Salinity, temperature and currents. Great temporary variations in salinity are encountered in the mussel growing area (Fig. 2). Seasonal variations are caused by variations in the flow of brackish water originating from the Baltic and streaming northwards along the Swedish west coast in the Baltic Current. From the melting of snow and ice during the spring large quantities of fresh water are poured into the Baltic which then affect the mussel growing area. Thus, the lowest salinities are usually seen in June (Svansson 1975). The water temperatures in the mussel growing area vary between sub-zero and 20°C tempera- tures. The Swedish mussel growing area is affected also by the Jutland Current which is streaming along the Danish west coast and reaching the Swedish coast just south of 58°N, where it usually turns northwards. The Jutland Current is bringing a mixture of Skagerrak and North Sea water to the farming area. This current as well as the Baltic Current may occasionally turn southwards (Fig. 1). Regular control of toxin content. Weekly sampling is performed from five mussel growing plants by the mussel growers (Fig. 3). Each sample contains 12 mussels, four mussels taken each from the surface, middle and bottom portion of a ribbon. The results of the toxin analysis is usually obtained by the mussel growers within 24 hours. Surveying the coast January 17-21, 1989. In order to obtain a more complete picture of the toxin contamination of the Swedish west coat, mussels were collected from sev- eral additional sites (Fig. 3). Mostly wild mussels were taken from the bottom of the sea and from anchorings for sailing-marks. These mussels were sampled during four days, when the OA-concentration of the mussels seemed to remain stable. OA determination. The analysis of the OA concentration was performed after extraction, derivatization with 9- anthryldiazomethane (ADAM), clean-up and HPLC anal- ysis essentially as described previously (Lee et al. 1987) with minor modifications (Edebo et al. 1988b). RESULTS The weekly sampling yielded low OA-concentrations during the spring and summer 1988 with a moderate rise in the latter half of August starting in the south (Fig. 4). In all sites but one, the latter located in the south, the OA-con- centrations receded during September. In early October, however, there was a strong OA rise in the mussel plants located at 58°37’ and 58°47’. With the exception of tempo- rary recessions the OA-concentrations continued to in- crease until late November when a peak value of 218 pg OA per 100 g mussel meat was reached at 48°47’ on No- vember 26. From then on, when the water temperature was ca 5°C, the OA-concentrations generally declined. Smaller increases were, however, noted. In the middle of February nov| o€c| | [san | Fee| mar| APR | MAY une|yucy| AUG] SEP | OCT 35 \ Figure 2. Long-term monthly means of salinities measured at Borné Hydrographical Stations 1931-1960 situated far into the Gullmar fiord. During the months April—-September, when the salinities have been low, <26%c, the OA concentrations of the mussels have been low. OKADAIC ACID IN MUSSELS 105 . Nyckelbyviken . Stridsfjorden . Kulefjorden . Bratté . Abyfjorden . N6sund . Moll6sund . Stigfiorden . Bornéstation OON DOA BwWD = CEXIILEDEZEI EEA (Cea 58°30’ S ? f ‘ eee ° = aM ‘ ° S> y : r?) 3 S04 a a aha e Kungshamn Gs 6h Ari (fux0 UDDEVALLA ; 58° 20' 7 Ef = 45 Ces WY LYSEKI yt Skarhamn(sy 57°50 GOTEBORG Figure 3. Map showing some major mussel-growing sites (1-7 within circles) along the Swedish west coast. During May 1988—April 1989 weekly sampling was done from the sites 1, 2, 3, 6 and 7. Within the squares the OA concentrations of mussels (wg per 100 g mussel meat) are shown for the period January 17-21, 1989. The highest OA concentrations were found in mussels collected from the area where the Jutland Current hits the Swedish coast. The lowest OA concentrations seem to exist in areas with low salinity and adjacent land-farming. 106 HAAMER ET AL. 2 e/a & ee Bak 150 \ a lg ; 2 seseeeees Nyckelbyviken 58° 54° here mules we oF nM te \ | ea \ gS Stridsfjorden 58 47 Ie te Nel Ss —-— Kulefjorden 58° 37° i | \\ S 100 === Nosund 58° 06" NA a SAUIEE EINE % —— Mollosund 58°04 Ln ue = iki! , z \ 6 aC a (Pr a| ph ne ee a et ee ena Bi eee ee sears : pie Ze a : oN ere KKK T T aan [ SSSI SSS May June July Aug Feb March April 1988 1989 Figure 4. Weekly monitoring of the concentration of OA in mussels from mussel-growing sites along the Swedish west coast (map, see Fig. 3). The mussel-growing sites at 58°04’ and 58°06’ are the ones most influenced by Baltic water. Furthermore they are close to land-farming areas. The mussel-growing sites at 58°37’ and 58°47’ are located in rather open fiords with large interchange with Skagerrak and the Jutland Current from the North Sea. The site at 58°54’ and the satellite of 58°47’ are located in the inner archipelago closer to the mainland. *upper limit for sale in Sweden (60 pg OA/100 g mussel meat). **upper limit for sale in most countries (20 1g OA/100 g mussel meat). ***details not always shown below 5 pg OA/100 g mussel meat. all mussel-growing plants showed OA-concentrations below 20 jg per 100 g mussel meat. As a consequence of earlier experiences (Edebo et al. 1988b) a satellite farm was placed ca one nautical mile east of the mother farm at 58°47’ in a strait with outward currents dominating. When the peak value 218 wg per 100 g mussel meat was reached on November 26, the satellite contained only 63 pg. The satellite remained on lower levels until the mother farm had reached lower levels at the end of January. Daily sampling was performed from three plants located closely at 58°04’—58°06' (Fig. 5) October 24 to November 3, 1988. During this period only low levels were noted. In the earlier part of the observation period one site (Ndsund) was free whereas another (Bockholmen) contained ca 10 wg OA per 100 g mussel meat. Later an almost opposite situation was seen. A third site (Gulskar) showed more sluggish OA concentrations. The survey along the Swedish coast which was per- formed during January 17-21, 1989 showed the highest OA concentrations, 92—134 wg OA per 100 g mussel meat in an area between 58°14’ and 58°18’ (Fig. 3). The lowest concentrations 4.2—7.3 zg OA per 100 g were found at the regular sampling sites in the south (Fig. 3: 6,7) but also at an old mussel growing plant located far into a bay (58°48’). Low concentrations were also found in the estuary of river Joralven at 58°34’. Moderately high concentrations 53 and 53.6 wg per 100 g were recorded in the outer archipelago of the south, at 57°43’ and 57°48’, respectively. DISCUSSION Weekly sampling from mussel growing plants for the determination of DST have proceeded for three years now (1986-1989). During the first year the i.p. mouse test Ya- sumoto et al. 1980) and the rat intestinal ligated loop (Edebo et al. 1988a) were used. Since August 1987 HPLC- determination of OA (Lee et al. 1987, Edebo et al. 1988) has been employed allowing more samples to be analyzed at lower cost and higher precision and with faster results. No DSP cases have been encountered in people con- suming batches of mussels containing less than 20 pg OA per 100 g mussel meat. During all the years the DST-levels have been lower from March to August and highest in Oc- tober—December. During the peak period October—No- vember 1988 the peak values were reached within six OKADAIC ACID IN MUSSELS 107 10 Nosund Bockholmarna Stockholmen Gulskar 0 881024—= 881103 Hg okadaic acid/100g mussel meat Oo Figure 5. Close-up map of the area 58°04’-58°06' where daily mussel samples from closely located mussel-growing sites were analyzed for OA during the period October 24 to November 3. weeks. Individual samples showed, however, decreases. Since increases and decreases seemed to occur in concert at the different mussel growing sites, they probably reflected alterations in the mussels rather than sampling and analysis errors and were consequences of the patchiness of the dis- tribution of toxic algae. Patchiness of the distribution of Dinophysis spp. with regard to site as well as depth seen in the Skagerrak area (Bohle et al. 1987) may account for the differences (Fig. 4) and time lags (Fig. 5) seen between closely located mussel growing sites belonging to the same water system. We also want to call attention to the fact that most of the clearing of OA from the mussels occurred in December to January. The winter 1988—89 was unusually mild and the water temperature was ca 5°C. The photosyn- thesis was, however, small and only low concentrations of plankton were found in the water. During the years 1986-1989 the mussel growing plants in the south have nearly always been less contaminated with DST than those in the north. The plant located furthest north (58°54'), however, showed lower OA concentrations than those at 58°37’ and 58°47’. This plant is situated in a bay. When the whole coast-line was surveyed in the latter part of January (Fig. 3) the highest OA concentrations were recorded from 58°14’ to 58°18’, where the Jutland Current is considered to hit the Swedish coast generally. In con- trast, it seems to be common to all sites with lower concen- trations of OA, viz. 58°04’—58°06', 58°34’ and 58°48’ (in- nermost site), that they belong to water systems greatly in- fluenced by fresh or brackish water. We hypothesize, therefore, that most of the toxic plankton is transported to the Swedish coastal waters by the Jutland Current. In some coastal areas the toxic organisms may be antagonized by unfavourable conditions such as reduced salinity, improper nutrient balances or competing non-toxic organisms prolif- erating on the richer nutrient concentrations in the coastal waters. The sites with lower concentrations of OA men- tioned above are situated, where water from land-farming districts is discharged into the sea. These observations sug- gest an extension of our hypothesis implicating that the phosphate and nitrogen pollution of the coastal waters may favour non-toxic plankton which are filtered off by mussels and transferred into mussel meat, rich in essential amino acids and marine-type lipids and relatively low in OA, whereas mussels growing in less nutrient-rich waters may be more exposed to Dinophysis toxins. ACKNOWLEDGMENT This work was supported by grant 934/88 from the Swedish Council for Forestry and Agricultural Research. The skilful technical assistance of Mrs. Hu Yao Juan and Mr Szlama Wysoki is greatfully acknowledged as is the co-operation of the persistent mussel growers. 108 HAAMER ET AL. REFERENCES Bohle, B., E. Dahl, M. Yndestad, & G. Langeland. 1987. Avoiding shellfish toxicity by lowering mussel plant below the pycnocline. Flo- devigen Meldinger Nr 2, 1987. ISSN 0800-7667. Dahl, E. & M. Yndestad. 1985. Diarrhetic Shellfish Poisoning (DSP) in Norway in the autumn 1984 related to the occurrence of Dinophysis spp. 3rd Internat. Congr. on Toxic Dinoflagellates, St. Andrews, New Brunswick, Canada D. M. Anderson, A. W. White and D. G. Baden, eds. (Elsevier, New York 1985) pp. 495-500. Edebo, L., S. Lange, X. P. Li & S. Allenmark. 1988a. Toxic mussels and okadaic acid induce rapid hypersecretion in the rat small intestine. APMIS 96:1029-1035. Edebo, L., S. Lange, X. P. Li, S. Allenmark, K. Lindgren & R. Thompson. 1988b. Seasonal, geographic and individual variation of okadaic acid content in cultivated mussels in Sweden. APMIS 96:1036— 1042. Kumagai, M., T. Yanagi, M. Murata, T. Yasumoto, M. Kat, P. Lassus & J. A. Rodriguez-Vazquez. 1986. Okadaic acid as the causative toxin of diarrhetic shellfish poisoning in Europe. Agric. Biol. Chem. 50:2853—2857. Lee, J. S., T. Yanagi, R. Kenma & T. Yasumoto. 1987. Fluorometric determination of diarrhetic shellfish toxins by high-performance liquid chromatography. Agric. Biol. Chem. 51:877—881. Nordberg, K. & H. Bergsten. 1988. Biostratigraphic and sedimentological evidence of hydrographic changes in the Kattegat during the later part of Holocene. Marine Geology 83:135—158. Svansson, A. 1975. Physical and chemical oceanography of the Skagerrak and Kattegat. Fishery Board of Sweden. Yasumoto, T., Y. Oshima, W. Sugawara, Y. Fukuyo, H. Ogun, T. Iga- rashi & N. Fujita. 1980. Identification of Dinophysis fortii as the caus- ative organism of diarthetic shellfish poisoning. Bull. Japan Soc. Sci. Fish. 46:1405—1411. Journal of Shellfish Research, Vol. 9, No. 1, 109-112, 1990. EFFECTS OF TRANSPLANTATION AND REIMMERSION OF MUSSELS MYTILUS EDULIS LINNAEUS, 1728, ON THEIR CONTENTS OF OKADAIC ACID J. HAAMER', P.-O. ANDERSSON?, S. LANGE?, X. P. LI, AND L. EDEBO? ‘Department of Physical Oceanography University of Goteborg Box 4038 S-400 40 Goteborg 2Pl. 6147, Kdémpersvik S-450 81 Grebbestad 3Department of Clin. Bacteriology University of G6teborg Guldhedsgatan 10 S-413 46 Goteborg, Sweden ABSTRACT Transplantation of mussels (Mytilus edulis) contaminated by okadaic acid (OA) from a more toxic environment in the northern part of the Swedish west-coast to a less toxic environment in the southern part showed a decrease in OA content of 12 pg OA/100 g mussel meat per day. Transplantation of less toxic mussels from the south to the north did not show a rapid uptake of OA. Toxic mussels from the north were reimmersed into two basins. One of them contained ordinary sea water, to the other one boiled baker’s yeast was added. Decreases of 4—5 xg OA/100 g mussel meat per day were observed. The OA-data showed a more consistent behaviour when boiled yeast was added. Without yeast decreases alternated with increases. The peak of the OA concentration in plankton particles 25-100 j2m coincided with that of Dinophysis cells in the sea water, D. acuta and D. acuminata being the dominating species. However, the toxicity of the mussels was less than anticipated if the ordinary filtering capacity of mussels is considered. KEY WORDS: INTRODUCTION The uptake of okadaic acid (OA) by mussels (Mytilus edulis) along the west coast of Sweden causes a severe im- pact on the Swedish mussel cultivation. OA causes intes- tinal disturbances such as diarrhea, nausea, vomiting and abdominal pain (Yasumoto et al. 1978, 1984, 1985). The condition is called diarrheic shellfish poisoning (DSP) after the dominating symptom diarrhea. There is evidence (Yasumoto et al. 1978, Edebo et al. 1988) of rapid uptake and also rapid depuration of OA by mussels in their natural environment. This points to the possibility of de-toxifying mussels by immersing into con- trolled environment in basins or by moving to a toxin-free part of the sea. However, the mechanisms of uptake, storage, and depuration are poorly understood. In order to gather some basic data simple transplantation experiments of mussels from one environment to another and also of reimmersion into basins have been performed. MATERIALS AND METHODS Sample preparation. Mussels (Mytilus edulis) were taken at 2 m depth from the collector bands in a mussel farm at Kulefjorden (site 3, see p 105, Fig. 3 this issue) situated in the Northern part of 'To whom correspondence is to be addressed. 109 mussels, transplantation, depuration, Mytilus edulis, okadaic acid. the Swedish west coast (N58°38") and from a similar farm at Molldsund (site 7, see p 105, Fig 3 this issue) in the Southern part (N58°4’). The mussels of each site were mixed in order to get a random distribution before putting them into plastic net bags with about 20 mussels in each. Half of the bags from each site were reimmersed into the sea at the same location at 2 m depth, and the remaining bags were transferred within 6 hr to the other site and im- mersed into the sea-water at 2 m depth. For reimmersion into basins toxic mussels from the northern site were put into bags in the same way and brought to the basins located at the southern site. OA-Analysis of Mussels Upon sampling twelve mussels from the net bag were chucked. After dewatering for 5 min. the meat was weighed. Thereafter the hepatopancreas (digestive glands) were removed and weighed. This gave the percentage of hepatopancreas related to the total weight of meat. The glands were put into a plastic tube and mailed to the De- partment of Clinical Bacteriology, University of Goteborg, where they arrived in the next morning and were analyzed for the OA-content according to Lee et al. (1987) as modi- fied by Edebo et al. (1988). It has been shown that the diarrheic shellfish toxins (DST) accumulate in the hepato- pancreas (Yasumoto et al. 1978) and that the main compo- nent in Europe is OA (Kumagai et al. 1986). 110 HAAMER ET AL. OA-Analysis of Plankton In order to examine the concentration of OA in plankton, water from 2 m depth was filtered through a 25 jz£m mesh conical net. The fraction collected on the net was dispersed and filtered through an ordinary cellulose filter which was kept in the deep-freeze until it was extracted as for mussels (vide supra). Initial experiments showed that, when the same amount of sea-water was filtered through a corresponding 100 1m mesh net, no toxicity was retained. Characteristics of the Sea Water At the same time as the net bags with mussels were taken out of the sea for analysis, the water was measured for temperature and salinity at the actual depth of 2 m. This was performed at the northern site (Kulefjorden) on every sampling occasion. At the same time samples for identifi- cation and quantification of algae were taken from the same depth. After sedimentation of 25 ml of the sea-water, the total content of algae was counted and the species identi- fied. Basins for Reimmersion of Mussels Two basins were each filled with 4 m? of local sea-water (no OA found). The water was recirculated in a closed loop, i.e. the same water was used throughout the whole experiment. The water was aerated by pumping water from the lower part of the basin and allowing it to fall through the air back into the basin. Because of the relatively small amount of mussels in the basin this aeration was enough. The faeces of the mussels sedimented to the bottom. To one basin 0.5 kg of boiled baker’s yeast was added, and to the other one no addition was made. RESULTS Transplantation Two mussel growing sites were chosen for transplanta- tion, one with high concentrations of OA in the native mussels (northern site, Kulefjorden N58°38’) the other one essentially non-toxic at the time preceding transplantation (southern site, Moll6sund N 58°04’). The transplantation was performed on October 24, 1988 from north to south and vice versa and the OA concentration of transplanted as well as native mussels analysed daily for ten days (Fig. 1). Already from the reference values taken on October 24— the native mussels in the north (72 wg OA/100 g mussel meat) and the transplanted mussels in the south (101 pg OA/100 g mussel meat) which belonged to the same batch of mussels—it appears that the difference between dif- ferent samples of mussels from the same lot may be great although 12 mussels for each sample were taken. Northern Site (toxic) The native mussels of the northern site remained toxic throughout the observation period (Oct. 24—Nov. 3) at TOXIC ENVIRONMENT (north) 150 - 100-- St to oO pe okadaic acid | 100 l sea water LESS TOXIC ENVIRONMENT 100 south He okadaic acid / 100g mussel meat 50 24 25 26 27 28 29 30 31 1 2 3 October 1988 November © NATIVE MUSSELS (0 TRANSPLANTED MUSSELS Figure 1. Transplantation of mussels from a toxic environment (north) to a less toxic one (south) and vise versa. levels 72-135 pg OA/100 g (Fig. 1). Also the sea water showed high concentrations of OA (10—40 pg OA/100 | sea water) from the start of the sampling (Oct. 26) until Nov. 1, whereas the last two days (Nov. 2—3) showed lower values (0.9—1.4 wg OA/100 1). In contrast, non- toxic mussels transplanted from the southern to the northern environment showed only minor increases in the OA concentration, the highest being 15 pg OA/100 g. Mi- croscopic examination of sea water, sampled when water was taken for OA analysis, showed presence of Dinophysis spp. mainly D. acuta and D. acuminata (Fig. 2). The peak 4000 3000 - rotula norwegica - acuminata . acuta Cells per litre 25=25 26 2¢ 28°29 S0NSiy ll) (2) 3 1988 November Figure 2. Presence of Dinophysis sp. in the water (cells/l) of the toxic environment (north). October TRANSPLANTATION OF MUSSELS 111 concentration of D. acuta 2500 per | and of D. acuminata 1200 per | were observed on October 29, when also the OA concentration of the sea water peaked. However, there were days, when the Dinophysis concentration was high in relation to the OA concentration (Oct. 28), and when it was low (Oct. 26 and 31). Further observations during this period showed that ini- tially there was a Gyrodinium aureolum bloom with 3 x 10° cells per 1 and a rather low Seechi-depth (Fig. 3). At the end of the period the concentration of G. aureolum was low and the Secchi-depth high. During the last two days (Nov. 2—3) the salinity increased indicating influx of sea water. Southern Site (Less Toxic) The native mussels at the southern site which had shown less than 2 wg OA per 100 g during the preceding weeks showed a moderate increase starting right after the trans- plantation and remained elevated until October 30 (Fig. 1). During this period the transplanted mussels showed in- creases and decreases which largely coincided with the na- tive ones. From October 31 and on the native mussels were free of OA. During this period the depuration of OA from the transplanted mussels averaged 12 wg per 100 g mussel meat per day. Reimmersion into Basins Several experiments have been conducted in which net bags with toxic mussels were immersed into the 4 m? tanks containing sea water modified with different additions. Usually the mussels were kept for observation periods of one week. No gross adverse effects on the mussels were observed. In a preliminary experiment one kg of fresh baker’s yeast was added to 4 m3 of the basin water and mussels immersed into the yeast suspension. The hepatopancreas homogenate from these mussels showed conspicuous pro- duction of gas and smelled of ethanol indicating that the fermentative capacity of the yeast had been transferred to the mussel hepatopancreas. Therefore, in future experi- ments yeast boiled for 10 min. was employed. One repre- sentative experiment is shown in Fig. 4. When plain sea Gyrodinium aureolum Secchi-depth ~ Salinity Temperature Cait Fast) Far aA easy PY zie) ESl hh 2 3 October 1988 November Figure 3. Presence of Gyrodinium aureolum (X10° cells/l), Secchi- depth (m), salinity (%c), and temperature (°C) in the water of the toxic environment (north). ib | 0 Sea water op © Sea water plus yeast 5 a So Rg Ss = 3 > Sim ] - A bo aS 8 ab aS is} 8 i od ° bo Re Ib | 0 1 2 3 4 5 6 i Days Figure 4. Reimmersion of toxic mussels into basins containing sea water or sea water plus yeast. water was used, there was considerable fluctuation of the OA concentration of the mussels from day to day. There were also considerable differences in between the duplicate samples taken on the same day, the largest difference being recorded on the last day of the experiment, when one sample showed 0.84 and the other one 2.95 jg OA per g hepatopancreas (Fig. 4) i.e. 9.5 and 38.8 wg OA per 100 g mussel meat, respectively. On the third day both samples showed increases compared to the initial sample. Toxic mussels immersed into sea water with yeast showed similar results (Fig. 4). In experiments not presented here also sub- stantial increases were recorded. All experiments resulted in lower concentrations of OA at the end of the reimmer- sion period than at the start the average decrease being 4—5 wg per 100 g mussel meat per day. DISCUSSION The recently published method for HPLC-analysis of the OA concentration in mussels (Lee et al. 1987) has shown good reproducibility (CV 4.9%) (Edebo et al. 1988) and thus offers a tool to control the OA content in mussels. However, the OA concentration in individual mussels 112 HAAMER ET AL. growing at the same site differs considerably showing dif- ferences as great as 0.63 and 10.0 mg/g heaptopancreas between mussels growing on the same collector bands (Edebo et al. 1988). We have tried to compensate for this variation by sampling 12 mussels from representative parts. Nevertheless the toxic reference samples (Fig. 1, October 24) labelled native mussels in the north and transplanted mussels in the south showed 72 and 101 pg OA/100 g re- spectively, although they were taken from the same lot. This may be due to the variation but may also be caused by drying during transport of the mussels to become trans- planted. Thus 72 pg OA/100 g may be closer to the true value. If so native and transplanted mussels at the southern site showed more similar variations of OA-concentrations both increasing from the first day on. During the later part of the transplantation period, when the native mussels were free of OA (Oct. 31—Nov. 3), there was an average de- crease of the transplanted mussels of 12 wg OA per 100 g per day. The real depuration capacity may be greater than this since the native mussels may handle minor quantities of OA without showing any content on analysis. The fact that OA was retained on 25 pm but not on 100 zm nets supports earlier observations (Dahl and Yndestad 1985, Andersson 1987) indicating that Dinophysis spp. carry the OA. When the different data of October 29 are compared 0.4 zg OA per | sea water corresponds to 2500 D. acuta and altogether 3800 Dinophysis cells. If all the OA is carried by these cells it would mean 1 — 1.6 X 10-4 wg OA/scell. Uptake of 12,500—20,000 Dinophysis cells per 10 g mussel may make them hazardous for con- sumption (20 jg per 100 g mussel meat). The peak con- centration of OA in the sea water amounted to 40 pg/100 1. The filtration rate of mussels with shell lengths of 43 mm has been estimated to 100 ml/min (Riisgard and Moehlen- berg 1979). Since the present mussels were close to this size, and if we assume a constant filtration rate, one single mussel should filtrate 144 1 per day and might accumulate 58 pg OA. If the weight of one mussel is taken to be 10 g an increase of 580 yg OA per 100 g mussel meat per day should be possible. Increases and even concentrations of this magnitude have never been observed by us. Different mechanisms may contribute to the lower levels seen in mussels viz. (1) mussels may close themselves, particularly on encounter of a toxic substance as OA; (2) the absorptive capacity of the mussels may become impeded; and (3) the depuration mechanism may increase. Since nontoxic mussels transplanted into the heavily toxic northern envi- ronment did not accumulate conspicuous quantities of OA a mechanism blocking the accumulation of OA in mussels seems to operate. It is generally believed and often stated in the literature that the toxic dinoflagellates have little effects on their molluscan hosts (see Gainey and Shumway 1988). However, in a recent series of publications concerning the PST-producing dinoflagellate Protogonyaulax tamarensis Shumway and associates (Gainey and Shumway 1988) have elegantly demonstrated profound biological effects and protection mechanisms in the mussels such as decrease in heart rate and cardiac arythmias, reduced oxygen con- sumption, inhibition of byssus production, shell valve clo- sure, siphon constriction and reduced filtration rates. An- imals from areas previously exposed to the toxic dinofla- gellates were generally less affected. Since, in our study, the blocking of OA uptake was so efficient for the mussels transplanted from OA free to OA containing waters, we propose that also DST-producing dinoflagellates induce be- havioural responses and protection mechanisms in the mussels such as shell valve closure and/or reduced filtration rates. These mechanisms may be activated particularly, when the mussels encounter great relative OA increments in the water. ACKNOWLEDGMENT This work was supported by grant 934/88 from the Swedish Council for Forestry and Agricultural Research. The skilful technical assistance of Mrs Hu Yao Juan and Mr Szlama Wysoki is greatfully acknowledged as is the co- operation of the persistent mussel growers. REFERENCES CITED Andersson, P.-O.: Unpublished results. Dahl, E. & M. Yndestad. 1985. Diarrhetic shellfish poisoning (DSP) in Norway in the autumn 1984 related to the occurrence of Dinophysis spp. 3rd Internat. Congr. on Toxic Dinoflagellates, St. Andrews, New Brunswick, Canada. D. M. Anderson, A. W. White and D. G. Boden, eds. (Elsevier, New York 1985) pp. 495-500. Edebo, L., S. Lange, X. P. Li, S. Allenmark, K. Lindgren, & R. Thompson. 1988. Seasonal, geographic and individual variation of okadaic acid content in cultivated mussels in Sweden. APMIS 96:1036—1042. Gainey, L. F., Jr. & S. E. Shumway. 1988. A compendium of the re- sponses of bivalve molluscs to dinoflagellates. J. Shellfish Res. 7:623-628. Kumagai, M., T. Yanagi, M. Murata, T. Yasumoto, M. Kat, P. Lassus & J. A. Rodriguez-Vasquez 1986. Okadaic acid as the causative toxin of diarrhetic shellfish poisoning in Europe. Agric. Biol. Chem. 50:2853-— 2857. Lee, J. S., T. Yanagi, R. Kenma, & T. Yasumoto. 1987. Fluorometric determination of diarrhetic shellfish toxins by high pressure liquid chromatography. Agric. Biol. Chem. 51:877—881. Riisgard, H. U. & F. Moehlenberg, 1979. An improved automatic re- cording apparatus for determining the filtration rate of Mytilus edulis as a function of size and algal concentration. Marine Biology 52:61— 67. Yasumoto, T., Y. Oshima & M. Yamaguchi. 1978. Occurrence of a new type of shellfish poisoning in the Tohoku district. Bull. Jpn. Soc. Sci. Fish. 44:1249-1255. Yasumoto, T., M. Murata, Y. Oshima, G. K. Matsumoto & J. Clardy. 1984. Diarrhetic shellfish poisoning. In: Ragelis, E. P. (Ed): Seafood toxins, ACS Symp. Ser. 262, ACS, Washington, D.C. pp 207-214. Yasumoto, T., M. Murata, Y. Oshima, M. Sano, G. K. Matsumoto & J. Clardy. 1985. Diarrhetic shellfish toxins. Tetrahedron 41:1019—1025. Journal of Shellfish Research, Vol. 9, No. 1, 113-118, 1990. THE EFFECTS OF MUSSEL (MYTILUS EDULIS, LINNAEUS, 1758) POSITION IN SEEDED BOTTOM PATCHES ON GROWTH AT SUBTIDAL LEASE SITES IN MAINE CARTER R. NEWELL Great Eastern Mussel Farms, Inc. Tenants Harbor, Maine 04860 ABSTRACT The effects of mussel (Mytilus edulis) position in dense aggregations (bottom patches) on growth (shell length and meat yield) were tested in experimentally seeded plots on 6 different subtidal mussel lease sites in Maine. In large patches (10 meters diameter), mussel final size was always significantly larger at the edge of patches than in the middle of the patches at all sites (20 replicates). In smaller patches (2—5 meters diameter), growth was significantly faster at the edge only at low-current sites (19 replicates). At any site, patches under 2 meters diameter resulted in the largest final mean shell lengths and meat weights. At one site, mussel growth on the edges of large patches was dependent on direction of tidal flows. These results present new field evidence which supports earlier hypotheses of the effects of mussel density on growth (Wildish and Kristmanson, 1985). In order to optimize the growth of mussels spread on bottom lease sites, careful consideration of mussel patch size, mussel patch position, current speed and direction and upstream depletion effects may significantly improve seed to harvest yields. The possibilities of density-dependent food limitation (Wildish, 1977) for even small patches at low current sites indicate a minimum current speed (about 3 cm per second) below which bottom culture of mussels may not be cost effective. INTRODUCTION The growth responses of mussels (Mytilus edulis) to their density and position in mussel beds are of interest in light of recent evidence concerning the role of hydrody- namic factors in limiting production of suspension-feeding bivalves (Wildish and Kristmanson, 1979; Fréchette et al., 1989). In dense assemblages of mussels, growth may be greater on the edge of large mussel beds and lesser in the middle when there is depletion of seston by adjacent mussels (Wildish and Kristmanson, 1984). Since at the edge of patches mussels have no local neighbors clearing food particles from the surrounding area, food resources may be higher, especially during periods of low current. Reduction in meat size of mussels at the sediment-water interface as compared with one meter off the bottom lends support to the hypothesis that mussel beds are food limited (Fréchette and Bourget, 1985). In a laboratory experiment using natural seston, mussel growth was dependent upon mussel density and current speed (Wildish and Krist- manson, 1985). Thus, seeding density should be more crit- ical at low-current sites, and high-current sites may support a higher biomass of mussels at bottom culture sites. Current direction may also be important. In tidally re- versing flows, conditions of phytoplankton abundance may be tidal direction dependent (Carlson et al., 1984). Mussels living at ‘“‘upstream’’ (relative to flood tide) and *‘down- stream’’ edges in the mussel patch would experience dif- ferences in food availability. This edge effect results in non-uniform growth of mussels in bottom culture. When mussels are seeded for bottom cultivation, seed distribution by the seeding vessel may result in bottom patches of different diameter. If the mussels are distributed in numerous small patches or fewer larger patches, the number of mussels on the edge will vary. If growth varies significantly in relation to mussel patch position, mussel seed distribution is critical to final meat and volume yield in the mussel bottom culture in- dustry. The responses of mussel growth to mussel patch size and mussel position within bottom patches allows us to determine under which conditions of mussel density, cur- rent flow, and food availability the growth of mussel meat and shell are inhibited. Since the goal of mussel culture is to produce uniform, fast growing market sized mussels from seed, the elimination of factors which limit mussel growth are desirable. Lower current sites can be farmed with higher yields through careful spreading of seed. In addition to the seston depletion effect, physical com- petition for space may also play a role in the observed ef- fects of mussel density and position in bottom patches on growth. An apparent reduction of mussel growth due to high density has been observed by some authors (Harger, 1967 and 1972; Bertness and Grosholz, 1985) in which mussels from the centers of clumps were smaller and had evidence of physical disturbance (thick, deformed and twisted shells). The effects of spatial position in aggrega- tions of mussels suspended on panels on growth and repro- ductive output were demonstrated by Okamura (1986) such that mussels in the centers of groups of 21—28 individuals grew more slowly than those on the edges or in smaller groups. Unfortunately, current speeds were not recorded at those sites. The phenomenon of self-thinning (the reduction in population density caused by competitively induced losses within a cohort of growing organisms) has been ex- amined in mussels (Hughes and Griffiths, 1988; Hosomi, 1985). In studies of individual mussel weight in relation to density, they observed the development of asymptotic mass per unit area. In the absence of mortality, growth of dense aggregations of mussels beyond the available space be- comes possible with the formation of hummocks (multi- layered packing). Since both seston depletion and space limitation have accounted for density-dependent growth differences, it is difficult to determine which factors are the most important 113 114 at a given site. Since both space and food availability are less limiting on the edge of mussel patches, the results pre- sented here cannot be used to distinguish the relative im- portance of each factor. Nonetheless, there is little field evidence which examines the growth of bivalves in relation to bottom patch position. The purpose of this paper is to demonstrate, at 6 different commercial mussel culture sites, that mussel seed distribution is a significant factor in seed to harvest yields due to intraspecific competition for food and space, resulting in the “‘edge effect’’. NEWELL METHODS Mussel seed was spread at 6 commercial mussel bottom culture operations and allowed to grow for approximately 12-18 months. Spreading was accomplished using a 400 kg capacity hydraulic bucket from a 20 m vessel travelling at 25-75 m sec! or 50 kg capacity fish trays from a 12 m vessel travelling at variable speed. Mussel seed lots varied from 5,000 to 30,000 kg and were obtained from wild in- tertidal seed beds. After seeding, mussels formed patches of varying size. When continuous patches were formed, TABLE 1. Sample dates, locations, patch size, patch position, and final mean size and of mussels seeded at 6 bottom culture sites in Maine. All tissue weights are dry tissue weight except Wohoa Bay, April 4, 1986 which is steamed meat weight. At Mud Cove, October 20, 1987, up designates upstream edge in relation to flood tide direction and down designates downstream edge (see methods). Patch Patch Date Site Size Position 4/15/87 Webb 2m Edge 2m Middle $/27/87 Webb Clumps NA 20m Edge 20m 1.5 M In 6/3/87 Ray 3m Edge 3m Middle 5m Edge Sm Middle 10/20/87 Mud 10m Up Edge 10m Middle 10m Down Edge 10/21/86 Wohoa 10m Edge 10m Middle Mud 10m Edge 10m Middle 4/4/86 Wohoa 11 10m Edge 10m Middle Wohoa 12 10m Edge 10m Middle 10/6/87 Webb 10m Edge 10m Edge 10m Middle 10m Middle 10/20/87 Mud 10m Edge 10m Edge 10m Middle 10m Middle 9/18/89 Roque 5m Edge 5m Edge 5m 2.0 M In Sm 2.0 M In 7/25/89 Stave Sm Edge 5m Edge 5m Edge 5m 1.0 M In 5m 1.0 M In 5m 1.0 M In 5m Middle Sm Middle 5m Middle Shell Tissue Length Std. Weight Std. (mm) Error N (g) Error 46.5 1.41 30 0.81 0.077 43.3 1.21 30 0.48 0.040 50.4 1.29 40 1.21 0.100 47.0 1.21 40 0.82 0.080 45.4 0.97 40 0.60 0.040 54.7 0.65 55.4 0.65 55u 1.03 48.3 0.69 48.4 0.96 41.0 0.71 43.6 0.63 57.8 0.99 51 1.20 0.06 50.5 0.95 51 0.62 0.04 60.8 1.28 50 1.59 0.10 54.3 1.26 50 0.90 0.07 42.0 0.85 48 1.59 0.09 S72: 1.02 48 0.78 0.06 ai le7/ 0.93 48 2.15 0.13 43.0 ilest7) 48 1.26 0.09 51.7 0.88 52.1 0.88 44.9 0.72 45.7 0.62 43.3 0.64 43.6 0.82 41.9 0.58 40.9 0.55 54.5 0.93 30 0.99 0.06 50.0 0.94 30 0.83 0.06 39.7 0.95 30 0.30 0.03 38.0 0.83 30 0.27 0.02 65.2 0.45 30 1.43 0.10 58.8 1.04 30 1.43 0.09 61.7 0.84 30 1.16 0.06 51.2 1.04 30 0.77 0.09 51.9 0.69 30 0.63 0.04 50.8 0.79 30 0.69 0.05 51.9 0.81 30 0.66 0.04 51.6 0.69 30 0.77 0.04 48.3 0.81 30 0.57 0.08 EFFECTS OF MUSSEL POSITION ON GROWTH AT SUBTIDAL SITES 115 they often took a circular shape and patch diameter was measured directly by diver using a meter stick. Where mussel patch diameter was less than 0.5 meters, the mussel patch was classified as “‘small clumps’”’ Density estimates were made by diver cores (0.613 square meters) which were taken at the edge and varying locations within mussel patches. Mussel samples were ana- lyzed for density and individual shell length (to the nearest 0.1 mm), and dry tissue weight after drying to constant weight at 80 C (to the nearest .001 g) was determined from a subsample of each grab sample. Since all mussels from the seed lot were of the same approximate initial mean size, mussel size at sampling was considered an indication of the growth rate of the seed in relation to bottom patch position. On one occasion, mussel meat size was determined by cooking a 2 liter sample for determination of steamed meat size (Newell, 1990). In order to investigate possible conditions which re- sulted in differences in mussel growth, grow-out sites were listed according to measurements of current speed (to the nearest 0.1 cm sec~!) by an S4 solid state electromagnetic current meter moored 0.5 m from the bottom. Sites were characterized as: 1. Low current (mean under 3 cm sec~!): Webb Cove, Roque Island. 2. Moderate current (mean between 3 and 10 cm sec~'): Mud Cove, Ray Point, Stave Island. 3. High current (mean over 10 cm sec~!): Wohoa Bay. Water depth at the 6 sites ranged between 1—5 m at low tide, and tidal range at the study sites averaged 3—5 m. Since flux of sestonic food has been shown to be more sig- nificant to the growth of mollusks than food or current alone, and since food availability was not monitored, this study assumes that increased current is an indication of an increase in the flux of food to mussels. At one site (Mud Cove), samples were taken in relation to tidal direction. At this site, the ‘‘upstream’’ position of the patch was designated as the position which was first exposed to water entering the lease site on the flood tide. The “‘downstream”’ position was designated as the position which was first exposed to water leaving the lease site on the ebb tide. Previous studies at Mud Cove (unpublished) have shown significant differences in chlorophyll a fluores- cence in relation to the tide, with greater food availability on the flood tide. This is in agreement with previous pub- lished studies at other shallow sites in Maine (Carlson et al., 1984). RESULTS In a series of growth samples taken from the low current sites (Webb Cove and Roque Island), mussels achieved a higher final meat weight and shell length on the edges of bottom patches 3 to 5 m diameter than in the middle of the patches (Tables 1 and 2, Figures 1a and 4a). In comparison to the patches, mussels grew faster at Webb Cove when seeded in small clumps than at the edge of larger bottom patches (Figure 1b). When a Student Newman Keuls Mul- tiple Range Test (SNK) was performed on the data, small TABLE 2. Results of statistical tests on the significance of differences in final shell length and meat weight of mussels samples from different positions in seeded bottom patches. Patch positions are as described in Table 1. All meat weights are dry tissue weight except April 4, 1986 which is steamed meat weight. For October 21, 1986 and April 4, 1986, 2-way bivariate ANOVA’s were performed. For October 6 and 20, 1987, nested ANOVA’s with 2 replicates for each mussel patch position were performed. Patch Significance Date Site Size Comparison Attribute Test Level 4/15/87 Webb 2m Edge vs Middle Shell t-test 0.05 Webb 2m Edge vs Middle Meat t-test 0.001 5/27/87 Webb 20m Edge vs Middle Shell ANOVA 0.01 vs. 1.5M In Meat ANOVA 0.001 6/3/87 Ray 3m Edge vs Middle Shell t-test NS Ray 5m Edge vs Middle Shell t-test 0.05 10/20/87 Mud 10m Up Edge vs Down Shell ANOVA 0.01 Edge vs Middle 10/21/86 Mud 10m Edge vs Middle Shell ANOVA .001 and Wohoa 10m Edge vs Middle Meat ANOVA .001 Lease site Shell ANOVA .003 Lease site Meat ANOVA 001 4/4/86 Wohoa 11 10m Edge vs Middle Shell ANOVA .001 and Wohoa 12 10m Edge vs Middle Meat ANOVA .001 Lease zone Shell ANOVA -001 Lease zone Meat ANOVA .001 10/20/87 Webb 10m Edge vs Middle Shell ANOVA .001 10/6/87 Mud 10m Edge vs Middle Shell ANOVA -001 116 WEBB COVE April 15, 1987 50 ee 1 = ° Ee re ~ > = = aD S o & 2D = (72) 2 - 7) pa S Q SS (= = \\ 2 = \ Edge Patch Middle = Patch 2m 2m WEBB COVE May 27, 1987 55 18 a aD E pl = ‘= f= = > e dD 5 = & 2 = a) [ ) Mean Shell Length ( mm ) Mean Steamed Meat Weight ( g 11 Edge 12 Middle 11 Middle 12 Edge 3b FIGURE 3. Final shell length and meat weight of mussels sampled from bottom patches at a moderate (Mud Cove) and high (Wohoa Bay) current site. a.) Growth in relation to patch position and culture site. b.) Growth in relation to patch position and lease zone. DISCUSSION The results presented in this study indicate that mussel growth at subtidal areas is significantly affected by site, position in bottom patches, and mussel patch size. The data provide new field evidence to support earlier hypotheses on the effects of mussel density on growth (Wildish and Krist- manson, 1985). The possibilities of density dependent food limitation (Wildish, 1977) are supported by the data, since the ‘‘edge effect’’ increased as measured current speeds decreased. Thus, at low current sites (mean <3 cm sec~'), mussel growth was significantly less at the middle of bottom patches than at the edge of patches just 2 m in diam- eter. At one moderate current site (mean 3—10 cm sec™'), growth was faster at the upstream edge of the patch than at the downstream edge. Thus, at some sites with reversing tidal flows, mussel position in bottom patches in relation to tidal direction will result in varying growth rates. These results suggest that the seston depletion effect may be more ROGUE ISLAND 5m PATCH September 18, 1989 0 ——— ===> 1.2 = oS Ee = E = — (ey = ® S Ss $ 2 = A) 2 = ” > [= Q 3 5 = ® = Edge Edge 2m in 2min 4a STAVE ISLAND 5m PATCH July 25, 1989 SS a a Ly oO : ae = l15 © fe ® oD s (= + 1.2 © so =} 4 a Ft +09 5 > c F06 & (= : Los § = en 1 0 Edge Edge Edge imin 1min 1min Mid Mid Mid 4b FIGURE 4. Final shell length and dry tissue weight of mussels seeded at a low (Roque Island) and moderate (Stave Island) current site. a.) Growth in relation to patch position at a 5 m patch at Roque Island. b.) Growth in relation to patch position at a 5 m patch at Stave Island. Transect across the patch at the edge, 1 m from the edge, and mid- point (2.5 m from the edge). TABLE 3. Densities and biomass of mussels recorded in this study in relation to previous published studies. All biomass values in this study are mean dry tissue weight per square meter at the edge or at varying positions within a continuous mussel patch. Density Biomass Area nm-? g ADW m~? Reference Roque Island Edge 5 M Patch 669 594 This study Middle 5 M Patch 1610 497 Stave Island Edge 5 M Patch 877 1158 This study 1 M from Edge 1683 989 Middle 5 M Patch 1408 1178 Other sites 800-5400 450-1900 This study Rhode Island 2140 1370 ~=Nixon et al, 1971 Wadden Sea, FRG 1382-1820 883-1445 Asmus, 1987 Wadden Sea, NL 4919 818 | Dame and Dankers, 1988 118 NEWELL important than competition for space per se within bottom patches of mussels. When density and biomass of mussels were examined (Table 3), mussel biomass reached a constant value within a given site regardless of position within bottom patches. Since mussels at the edge of the patches were at a lower density, biomass was distributed over fewer individuals, resulting in a higher mean size at the edge. Thus, by im- proved spreading of mussel seed at bottom culture sites re- sulting in less mussels in the middle of large bottom patches and more at the edge of smaller patches, growth rates will improve significantly. These results support ear- lier observations of the development of asymptotic mass of mussels per unit area at any given site (Hughes and Grif- fiths, 1988; Hosomi, 1985). In addition, low current sites (Roque Island) had lower biomass values within the mussel patches than sites with higher currents (Stave Island). This supports earlier work (Wildish and Peer, 1983) which found the highest biomass values of horse mussels (Mo- diolus modiolus) in the Bay of Fundy at the highest current areas. ACKNOWLEDGEMENTS This research was supported by the National Science Foundation SBIR Awards ISI8660201 and ISI8809760. I thank Deb Murphy for help with the growth samples and D. J. Wildish, M. Frechette, A. Smaal and an anonymous reviewer for helpful comments. REFERENCES Asmus, R. 1986. Secondary production of an intertidal mussel bed com- munity related to its storage and turnover compartments. Mar. Ecol. Prog. Ser., 39:251—266. Bertness, M. D. & E. Grosholz. 1985. Population dynamics of the ribbed mussel, Geukensia demissa: the costs and benefits of an aggregated distribution. Oecologia 67:192—204. Carlson, D. J., D. W. Townsend, A. L. Hyliard & J. F. Eaton. 1984. Effects of an intertidal mudflat on plankton of the overlying water column. Can. J. Fish. Aquat. Sci. 41:1523—1528. Dame, R. F. & N. Dankers. 1988. Uptake and release of materials by a Wadden Sea mussel bed. J. Exp. Mar. Biol. Ecol. 118:207—216. Fréchette, M. & E. Bourget. 1985. Food-limited growth of Mytilus edulis L. in relation to the benthic boundary layer. Can. J. Fish. Aquat. Sci. 42:1166—1170. Fréchette, M., C. A. Butman & W. R. Geyer. 1989. The importance of boundary-layer flows in supplying phytoplankton to the benthic sus- pension feeder, Mytilus edulis L. Limnol. Oceanogr. 34:19-36. Harger, J. R. 1967. Population studies on Mytilus communities. Ph.D. diss., Univ. Calif. Santa Barbara, Univ. Microfilms No. 69-1719. Harger, J. R. 1972. Competitive coexistence: Maintenance of interacting associations of the sea mussels Mytilus edulis and Mytilus califor- nianus. Veliger 14:387—410. Hosomi, A. 1985. On the persistent trend of constant biomass and the constant total occupation area of the mussel Mytilus galloprovincialis (Lamark). Venus Jpn. J. Malacol. 44:33—38. Hughes, R.N. & C. L. Griffiths. 1988. Self-thinning in barnacles and mussels: the geometry of packing. Am. Nat. 132:484—491. Newell, C. R. 1990. A guide to mussel quality control. Maine Sea Grant E-MSG-90-1, 20 pp. Nixon, S. W., C. A. Oviatt, C. Rogers & K. Taylor. 1971. Mass and metabolism of a mussel bed. Oecologia 8:21—30. Okamura, B. 1986. Group living and the effects of spatial position in aggregations of Mytilus edulis. Oecologia 69:341—347. Wildish, D. J. 1977. Factors controlling marine and estuarine sublittoral macrofauna. Helg. Wiss. Meers. 30:445—454. Wildish, D. J. & D. D. Kristmanson. 1979. Tidal energy and sublittoral macrobenthic animals in estuaries. J. Fish. Res. Bd. Can. 36:1197— 1206. Wildish, D. J. & D. Peer. 1983. Tidal current speed and production of benthic macrofauna in the lower Bay of Fundy. Can. J. Fish. Aquat. Sci. 40 (Suppl. 1):309-321. Wildish, D. J. & D. D. Kristmanson. 1984. Importance to mussels of the benthic boundary layer. Can. J. Fish. Aquat. Sci. 41:1618—1625. Wildish, D. J. & D. D. Kristmanson. 1985. Control of suspension feeding bivalve production by current speed. Helg. Wiss. Meers. 39:237-243. Journal of Shellfish Research, Vol. 9, No. 1, 119-124, 1990. INDUCTION OF METAMORPHOSIS OF THE JAPANESE SCALLOP PATINOPECTEN YESSOENSIS JAY BRIAN C. KINGZETT,! NEIL BOURNE,! KAREN LEASK? ‘Department of Fisheries and Oceans Biological Sciences Branch Pacific Biological Station Nanaimo, British Columbia, Canada. V9R 5K6 2University of Victoria Department of Biology Victoria, B.C., Canada. V8W 2Y2 ABSTRACT Hatchery reared larvae of the Japanese scallop, Patinopecten yessoensis, were treated with different levels of neuro- transmitters including, norepinephrine, epinephrine, L-DOPA, serotonin to test the ability of these compounds to increase percent metamorphosis in the absence of a suitable substrate. Thermal shock and the addition of ammonia were also tested for their effect on mature larvae. Norepinephrine, epinephrine and L-DOPA produced significant increases in percent metamorphosis. Results with ammonia were variable and significant increases in percent metamorphosis depended on concentration and exposure time. No consis- tent significant increase in percent metamorphosis was observed when mature larvae were treated with serotonin or subjected to cold temperature shock. KEY WORDS: Japanese scallop, Patinopecten yessoensis, metamorphosis, neurotransmitters, ammonia, temperature shock INTRODUCTION The process by which many marine invertebrate plank- tonic larvae transform to become bottom dwelling juveniles can be divided into two stages; settlement, a repeatable be- haviourial stage and metamorphosis, an irreversible physio- logical stage. Many mature invertebrate larvae have been induced to metamorphose in response to specific environ- mental cues (Crisp 1974; Hadfield 1977; Burke 1983). Chemical, photic and tactile cues may indicate the presence of a substratum or habitat suitable for juvenile life. Present methods for inducing metamorphosis of com- mercially important bivalve larvae are not reliable and the fundamental biological processes underlying their activity are poorly understood. Neuroactive compounds are effec- tive in inducing many species of molluscan larvae to settle and/or metamorphose under laboratory conditions (Coon et al. 1985; Cooper 1982; Hadfield 1977; Levantine and Bonar 1986; Morse 1985; and Morse et al. 1979). Epineph- rine has been used to produce cultchless Pacific oyster, Crassostrea gigas, spat (Coon et al. 1986). The biological mechanism for induced metamorphosis caused by exoge- nously applied neuroactive compounds has been investi- gated in a number of species of molluscs (Coon and Bonar, 1987a and 1987b; Coon et al. 1988; Cooper 1982; Hadfield and Scheuer 1985; Trapido-Rosenthal and Morse 1985a and 1985b). Ammonia also induces settlement behaviour in larvae of the eastern oyster, Crassostrea virginica, and the Pacific oyster, C. gigas, but it is thought to act by in- creasing the intracellular pH, a mechanism of action quite separate from that of neuroactive compounds (Coon et al. 1988). Recent studies have shown the potential for scallop cul- ture in British Columbia using hatchery methods to produce juveniles (Bourne et al. 1989). An integral part of these studies has been the development of an efficient nursery system to set mature larvae quickly and raise the juveniles to an outplanting size rapidly to make efficient use of cul- ture tanks (Bourne and Hodgson In press). Preliminary in- vestigations indicated that the exogenous application of cat- echolamine solutions and exposure to cold water tempera- tures had some positive, although variable, effects on the induction of metamorphosis of mature scallop larvae (Bourne and Hodgson In press). In the present study the effect of the exogenous applica- tion of neuroactive compounds, L-DOPA, epinephrine (EPI), norepinephrine (NOREP), and serotonin (SER), as well as the use of ammonia and cold temperature shock on the metamorphosis of mature larvae of the Japanese scallop, Patinopecten yessoensis, was investigated. The purpose was to develop a reliable method to enhance the rate of metamorphosis of mature Japanese scallop larvae. MATERIALS AND METHODS Different groups of Japanese scallop larvae raised at the Pacific Biological Station from March to July, 1989 were used. Larvae grown at 15° C in 2,500 or 5,400 | fibreglass tanks were fed a mixture of three species of cultured phyto- plankton, Thalassiosira pseudonana, Chaetoceros calci- trans and Isochrysis galbana (Tahitian variety) (Bourne et al. 1989). Larvae, 20—26 days old, were collected and those retained on a 200 ym Nitex screen were used in the experiments. Criteria used to determine whether the larvae were ma- ture and competent to metamorphose included the presence 119 120 KINGZETT ET AL. of a well developed foot, developing gill bars, eyespot, and a mean shell length in excess of 260 ym (SD less than 10 42m) (Hodgson and Bourne 1988, Bourne et al. 1989). Methods used to investigate the effect of neurotrans- mitters on metamorphosis were those used for other bivalve larvae (Coon et al. 1985). Assays were conducted in 24 (2.5 ml) well tissue culture plates to insure adequate experi- mental replication. No setting substrate material or food was added to the wells in any experiment. Mature larvae when selected were resuspended in sea- water filtered to | wm and counted. Approximately 70—150 larvae were pipetted onto 1 cm diameter 200 wm Nitex mesh screens made from automatic pipette tips. Screens with larvae were added to solutions being tested. In the first set of experiments, larvae in screens were placed in beakers containing the test compound at 15°. After the desired exposure to test chemicals, the larvae on the screens were rinsed three times with filtered seawater, then pipetted into 2 ml of fresh filtered seawater in the wells of the tissue culture plates and placed in an incubator atalsnC: After the larvae were held in an incubator for 48 hours they were examined under a dissecting microscope and a 7% METAMORPHOSIS Be 30 + FZ a 20 4 FV 7 7 10 4 ~ 7 count made of the number of metamorphosed, unmetamor- phosed and dead larvae. Metamorphosed larvae (spat) were identified by the lack of a velum, presence of gill bars with elongated primary filaments and the orientation of the foot in a ventral or anterior position rather than in the posterior position found in larvae (Hodgson and Bourne 1989). Rep- licates of each well were made and all assays were repeated at least three times to standardize for the effect of treatment between groups of larvae. Stimuli tested in this work were chemical compounds and temperature shock. The first group of chemical stimuli tested were neuroac- tive compounds. Those selected were norepinephrine, epi- nephrine, L-DOPA, and serotonin. These compounds have been shown to either stimulate metamorphosis in other bi- valve larvae or were neuroactive compounds known to be present in molluscs (Coon et al. 1985; Coon and Bonar 1986; Walker 1986). These chemicals can be obtained in the (+), (+,—) or (—) form. The (—) or L form of each chemical was used whenever possible since these are the forms that occur naturally in bivalves (Walker 1986). These chemicals oxidize rapidly in seawater and activity is lost. Therefore, solutions were used within 24 hours after ASSAY Figure 1. Average percent metamorphosis of mature Japanese scallop, Patinopecten yessoensis, larvae in controls by assay from March—July, 1989. Bars indicate 95% confidence levels. INDUCTION OF METAMORPHOSIS OF PATINOPECTEN YESSOENSIS JAY 121 preparation. Since 10 Molar stock solutions in distilled water are acidic and toxic to bivalve larvae, buffered solu- tions were used (10 mM TRIZMA-HCL in distilled water). To test for possible effects of buffered distilled water and the resulting 10% decrease in salinity, controls for each ex- periment were performed in 10% buffered distilled water and in filtered seawater. The effect of each chemical on mature scallop larvae was tested at 10-4, 10-5, and 10~® M, concentrations which have been found to be effective for stimulating oyster larvae to metamorphose (Coon et al. 1985). Larvae were exposed to all neuroactive chemicals for 60 minutes. Ammonia was tested at 10-7, 0.5 * 10~? and 10-3 M, based on concentrations found to be effective for stimu- lating oyster larvae to metamorphose (Bonar pers comm 1989). Scallop larvae were exposed for 10, 30 and 60 minutes at each concentration. For the second set of experiments cold temperature shock was tested. Larvae raised at 15° C were placed in beakers with seawater at 4, 8 and 11° C for 10, 30 and 60 minutes. Control larvae were held in beakers with seawater at 15° C. The effects on metamorphosis were determined as described previously. Since the number of larvae in each well was not con- stant, counts of the number of metamorphosed larvae in each experiment were converted into percentages of the total number of larvae in each well. For statistical analysis, percent metamorphosis rates were normalized using an arc- sine transformation. The effect of each compound or thermal shock on the rate of metamorphosis was calculated by determining the increase in metamorphosis for each treatment replicate over the mean rate of metamorphosis in the control for each assay. Results were pooled for all assays and replicates, means and variance for the effect of each treatment were determined. Control variance was expressed as a 95% con- fidence interval about a zero deviation. A two way ANOVA using SAS statistical software (PROC GLM) was performed on each data set using the calculated increase in metamorphosis as the dependent vari- TABLE 1. The percent increase in metamorphosis of mature Patinopecten yessoensis, larvae over controls when treated with (—) norepinephrine (NOREP), (—) epinephrine (EPI), L-DOPA, serotonin (SER), ammonia and cool temperature shock. Significant results at the p < 0.05 level are denoted by ****. Conc. Exposure Treatment (molar) (min) NOREP Ome 60 10m 60 1Ome 60 EPI 105 60 10m 60 10° 60 L-DOPA 1054 60 10-5 60 10-6 60 SER 10-4 60 10-4 60 10m° 60 Ammonia 10m 10 30 60 0.5 x 10-2 10 30 60 10m 10 30 60 Temp. Shock ay 10 30 60 8° 10 30 60 1S 10 30 60 Increases Signif. (%) (P < 0.05) # Assays 18.0 + 8.2 BEAT? 7 19.0 + 7.1 tte 7 I7/es}) == 7/53} AAS 7 1529865519 AES 7 12235522 ao 7 9.6 + 6.1 cee 7 16.0 + 8.4 oF Fa: 3 7.9 + 4.1 Fee 3 ES) 41 3 Sr4ee el 3 Bh) ae oh) 3 Pee 285 3 627) 6:5 3 sisi == 7s) the 3 61471 3 WS} S= S¢/ oa 3 16:3) 7/22 aA 3 ils) Ss TRS cette 3 8.2 + 6.9 3 8.5 + 15.9 3 7.1 + 8.8 3 6.7 + 6.5 3 9.0 + 10.9 3 OMS 56:55 3 11.0 + 10.0 3 3.6 + 6.7 3 1.69 761 3 6.7 + 4.6 3 pd, a= (oy) 3 6h(0) B= S165) 3 122 22 KINGZETT ET AL. able, and the assay and treatments (chemical or physical shock) as fixed effects. A Least Squares Design multiple range test (SAS Institute Inc. 1982) was performed to de- termine which treatments increased significantly from the variance about the controls. RESULTS Controls The percent metamorphosis in controls varied from 10—40% between assays (Fig. 1). This reflected the com- petency of different groups of larvae to metamorphose since they were obtained from different adults and raised at different times of the year. A Students T-test showed no significant difference be- tween the percent metamorphosis of larvae held in controls with normal seawater and larvae held in seawater plus buff- ered distilled water. Consequently, results from both con- trol groups were pooled for the analysis of the effect of the neuroactive compounds. Neuroactive Compounds The largest increase in percent metamorphosis occurred when mature larvae were treated with norepinephrine NEUROQACTIV 7 INCREASE (Table 1, Fig. 2.). The mean increase in percent metamor- phosis over controls between the three concentrations of norepinephrine ranged from 17—19% which was not signif- icantly different (p < 0.05). When mature larvae were treated with epinephrine there was a mean increase of 10—16% in metamorphosis over the controls (Fig. 2). As with norepinephrine, there was no sig- nificant difference in the rate of metamorphosis when dif- ferent concentrations were used. The effects of L-DOPA on mature larvae were more variable than those with norepinephrine or epinephrine and more dependent on concentration (Fig. 2.). The mean in- crease in the rate of metamorphosis over controls was only significant at concentrations of 10~* and 10~5 M. A signif- icant difference in the rate of metamorphosis was observed with decreasing concentrations of the compound (p = 0.05). No significant difference in the percent metamorphosis of mature larvae over controls was observed when they were treated with serotonin (Fig. 2.). Ammonia Exposures at an ammonia ion concentration of 10~? for 30 minutes and at 0.5 * 10~? at all exposures caused a E COMPOUNDS 30 Ze A A AU AI \ \\\I CON NOREP elP| 10710°10° SS eel 10°10” ) T T 10° ae 1071071 CONCENTRATION Figure 2. Increase in percent metamorphosis over controls of mature Japanese scallop, Patinopecten yessoensis, larvae when treated with 10-4, 10~5, and 10~¢ M solutions of norepinephrine (NOREP), epinephrine (EP), L-DOPA and serotonin (SER). + indicates 95% confidence levels. INDUCTION OF METAMORPHOSIS OF PATINOPECTEN YESSOENSIS JAY 12 significant increase in the percent metamorphosis (Table 1, Fig. 3.). High variation in response to ammonia treatment was observed. Temperature Shock Temperature shock did not cause an overall significant increase in the percent metamorphosis of mature scallop larvae over controls (Table 1, Fig. 4). High variance was observed between experiments. Significant increases in metamorphosis were observed at some temperatures and at some exposure times but they were not reproducible be- tween experiments. DISCUSSION The large variation in rate of metamorphosis in groups of control larvae was not unexpected and has been observed in previous work at the Pacific Biological Station (Bourne et al. 1989). The difference in percent metamorphosis of different groups of mature larvae is thought to be due to several factors including differences in growth rates of larvae within groups, condition of the larvae at maturity and the quality of the eggs in female broodstock which may affect larval condition (Gallager and Mann 1986). The lowest percentages of metamorphosis were observed in the AMM 7 INCREASE Ww later experiments and these larvae came from broodstock that had been held in conditioning tanks for as long as seven months. Partial resorption of gonadal material may have begun in some of these animals and the condition of the eggs would have been much less than those in brood- stock that were used in the earlier experiments. The greatest increase in percent metamorphosis occurred when mature larvae were treated with norepinephrine; the mean increase was 17—19% over controls. Increases in percent metamorphosis with other treatments was less than 16%. Increases in the percent metamorphosis were consis- tent with results reported for other bivalve larvae (Bonar et al. 1989). No explanation can be given for the greater effect of norepinephrine on metamorphosis of Japanese scallop larvae over the other catecholamines and L-Dopa. It is not known if the addition of norepinephrine or epinephrine will induce metamorphosis without settlement behaviour as has been reported for oyster larvae (Bonar et al. 1989). This warrants further investigation. Ammonia had a lesser and more highly variable effect on increasing the rate of metamorphosis than the neuroac- tive compounds. It is believed the effect of ammonia on mature larvae is transitory as the larvae acclimate to the ONIA 30 a AAO) = fers VO = FAB Za Z ZZ Sita ZZ Za ga (6) | LEA AAA AA ~ | con 110 oe Vem =10 ria ) 10° 3@ 6@ 10 30 60 HOM SOMASO, Figure 3. Increase in percent metamorphosis over controls of mature Japanese scallop, Patinopecten yessoensis, larvae when treated with 10-2, 5 * 10-2 and 10-3 M solutions of ammonia for 10, 30 and 60 minutes. + indicates 95% confidence levels. 124 KINGZETT ET AL. 7 INCREASE TEMPERATURE Sy Clee 0) = DS = 205 15 - 10 4 iz = ZZ ae ZA 5 i AA ZLEGZ Z ZZ Z AF 5 ZomgZ —-5 4 CON on ae foe’ 0 SSS Te 10% 230. | 60 10> 30> 3660 EXPOSURE Figure 4. Increase in percent metamorphosis over controls of mature Japanese scallop, Patinopecten yessoensis, larvae when exposed to 5°, 8° and 11° C for 10, 30 and 60 minutes. + indicates 95% confidence levels. increased presence of the ammonia ion. Bonar et al. (1989) postulated that increased ammonia levels may represent a non-specific settlement cue for bivalve larvae. Ammonia levels may be a cue in the natural environment that aid in gregarious larval settlement (Coon et al. 1988, Bonar et al. 1989). Increased levels of ammonia have been associated with bacterial fouling which has been shown to increase metamorphosis of oyster larvae (Wiener et al. 1989). Higher rates of settlement and metamorphosis of Japanese scallop larvae have been observed on fouled substrates compared to clean substrates (D. O’Foighl pers. comm.). The non appreciable effect of increasing metamorphosis by exposing larvae to cool water temperatures was sur- prising since other work showed it had a positive effect in increasing metamorphosis (Bourne and Hodgson In Press). In routine experimental hatchery work at the Pacific Bio- logical Station, larvae are customarily chilled before they are placed in setting tanks (Bourne et al. 1989). A possible explanation is that temperature shock may stimulate a set- ting behaviour and that metamorphosis is only increased when a suitable substrate is present. It is postulated that L-Dopa and ammonia stimulate reversible settlement be- haviour in oyster larvae which then leads to increased meta- morphosis if suitable conditions are encountered (Bonar et al. 1989) and this may be similar for Japanese scallop larvae. Further experiments should be carried out to determine if the addition of food and/or a suitable substrate would increase the rate of metamorphosis further than observed in the present work. Larvae treated with ammonia or cold temperature shock may show more consistent increased rates of metamorphosis if a suitable substrate was present. Results of the present work showed that metamorphosis was consistently increased when mature larvae were treated with (— ) norepinephrine and to lesser degrees with L-Dopa and (—) epinephrine. Treatment with norepinephrine is in- expensive, easily carried out and has an application in aquaculture operations. In a commercial hatchery an average increase of 18% in metamorphosis would be eco- nomically significant and would warrant its use. Further experiments are also required to determine if spat of larvae induced to metamorphose by chemicals or physical shock grow and survive as well as non treated spat. Previous work producing cultchless oysters with epi- INDUCTION OF METAMORPHOSIS OF PATINOPECTEN YESSOENSIS JAY 12 nephrine and norepinephrine showed that post-set survival and growth was normal (Coon et al. 1986) but this should be verified for scallop larvae. ACKNOWLEDGMENTS Support for B. Kingzett was provided by the Ministry of n Agriculture and Fisheries of British Columbia and for K. Leask in part from a grant to R. D. Burke from the Natural Sciences and Engineering Research Council of Canada. We sincerely thank Drs. J. N. C. Whyte, C. Clarke and R. D. Burke for review of the manuscript and S. Cross and M. Saunders for assistance with the statistical analysis. LITERATURE CITED Bonar, D. B., S. L. Coon, M. W. Walch, R. M. Wiener, and W. Fitt (1989) Control of oyster settlement by endogenous and exogenous chemical cues. Bull. Mar. Sci. In Press. Bourne, N., C. A. Hodgson and J. N. C. Whyte. 1989. A manual for scallop culture in British Columbia. Can. Tech. Rep. Fish. Aquat. Sci. 1694:215p. Bourne, N. and C. A. Hodgson. In Press. Development of a viable nursery system for scallop culture. J. World Aqua. Soc. Burke, R. D. 1983. The induction of metamorphosis of marine inverte- brate larvae: stimulus and response. Can. J. Zool. 61:1701—1719. Coon, S. L., D. B. Bonar and R. M. Wiener. 1985. Induction of settle- ment and metamorphosis of the Pacific oyster Crassostrea gigas (Thunberg) by L-Dopa and catecholmines. J. Exp. Mar. Biol. Ecol. 9:211—221. Coon, S. L., D. B. Bonar, and R. M. Weiner. 1986. Production of cultchless oyster seed by epinephrine and norepinephrine. Aquaculture 58:255-—262. Coon, S. L., and D. B. Bonar. 1986. Norepinephrine and Dopamine con- tent of Larvae and spat of the pacific oyster, Crassostrea gigas. Biol. Bull. 171:632—639. Coon, S. L., D. B. Bonar. 1987a. Pharmacological evidence that alpha-1 adrenoreceptors mediate metamorphosis of the Pacific oyster Crasso- strea gigas. Neurosci. 23:1169—1174. Coon, S. L., D. B. Bonar. 1987b. The role of DOPA and dopamine in oyster settlement behaviour. Amer. Zool. 27(4):128a. Coon, S. L., M. Walch, W. K. Fitt, D. B. Bonar and R. M. Wiener. 1988. Induction of settlement behaviour in oyster larvae by ammonia. Amer. Zool. 28(4):70a. Cooper, K. 1982. A model to explain the induction of settlement and metamorphosis of the planktonic eyed-pediveligers of the blue mussel Mytilus edulis L. by chemical and tactile clues. J. Shellfish Res. 2(1):117. Crisp, D. J. 1974. Factors influencing the settlement of marine inverte- brate larvae. In. Chemoreception in marine organisms. (P. T. Grant and A. M. Mackie Ed.) Academic Press, London. p 177—277. Gallager, S. M. and Mann, R. 1986. Growth and survival of larvae of Mercenaria mercenaria (L.) and Crassostrea virginica (Gmelin) rela- tive to broodstock conditioning and lipid content of eggs. Aquaculture, 56:105—121. Hadfield, M. G. 1977. Chemical induction in larval settling of a marine gastropod. Marine Natural Products Chemistry. (D. J. Faulkner and W. H. Fenical Ed.) Plenum Press, New York. p 403-413. Hadfield, M. G. and D. Scheuer. 1985. Evidence for a soluble metamor- phic inducer in Phestilla: ecological, chemical and biological data. Bull. Mar. Sci. 37:697—706. Hodgson, C. A. and N. Bourne. 1988. Effect of temperature on larval development of the spiny scallop, Chlamys hastata Sowerby, with a note on metamorphosis. J. Shellfish Res. 7:3 pp 349-357. Levantine, P. L. and D. B. Bonar. 1986. Metamorphosis of //yanassa obsoleta: natural and artificial inducers. Amer. Zool. 26:14a. Morse, D. F. 1985. Neurotransmitter-mimetic inducers of larval settle- ment and metamorphosis. Bull. Mar. Sci. 37:697—706. Morse, D. P., N. Hooker, H. Duncan and L. Jensen. 1979. Gamma- amino butyric acid—a neurotransmitter— induces planktonic larvae to settle and begin metamorphosis. Science 204:407—410. Sas Institute Inc. 1982. SAS User’s Guide: Statistics, 1982 Edition. Cary, NC: SAS Institute Inc. Trapido-Rosenthal, H. G. and D. E. Morse. 1985a. Regulation of re- ceptors controlling settlement and metamorphosis in larvae of a gas- tropod mollusc (Haliotis rufescens). Bull. Mar. Sci. 39:283—292. Trapido-Rosenthal, H. G. and D. E. Morse. 1985b. L- ,w-diamino acids facilitate GABA induction of larval metamorphosis in larvae of a gas- tropod mollusc (Haliotis rufescens). J. Comp. Physiol. 155b:403- 414. Walker, R. J. 1986. Transmitters and modulators. /n The mollusca. vol 9. Neurobiology and behaviour, Part 2. A. O. D. Williams editor. Aca- demic Press Inc. pp 279-486. Wiener, R. M., M. Walch, M. Labare, D. B. Bonar, R. R. Colwell. 1989. Effect of biofilms of the marine bacterium Alteromonas colwel- liana (LST) on the set of the oysters Crassostrea gigas and C. Vir- ginica. J. Shellfish Res. 8:117—123. Journal of Shellfish Research, Vol. 9, No. 1, 127—134, 1990. BURIAL OF TRANSPLANTED BAY SCALLOPS ARGOPECTEN IRRADIANS IRRADIANS (LAMARCK, 1819) IN WINTER STEPHEN T. TETTELBACH!, CHRISTOPHER F. SMITH?, JAMES E. KALDY III’, THOMAS W. ARROLL!, AND MICHAEL R. DENSON! Natural Science Division Southampton Campus of Long Island University Southampton, New York 11968 2Cornell Cooperative Extension of Suffolk County Marine Program Riverhead, New York 11901 ABSTRACT To better understand some of the factors which affect the success of Argopecten irradians irradians reseeding pro- grams, the progression and prevalence of scallop burial and mortality due to shifting sediments were examined. Ten days after their release in Northwest Creek, NY during December, 14% of all observed scallops were completely buried and 27% were partially buried. Mortality among observed individuals increased to 70-90% by late March, but overall recovery was only 23-39%. In laboratory studies, all scallops buried by 1 cm of sediment died within | day; almost all partially buried scallops survived for 1 month. Mortality was greater on muddy-sand than sand. Movements which resulted in partially buried scallops unburying themselves were much less frequent at water temperatures below 3°C. Burial in winter is seen as a potentially significant cause of bay scallop mortality which should be considered when implementing reseeding programs. KEY WORDS: INTRODUCTION During 1985, 1986, and 1987, extensive blooms of Au- reococcus anophagefferens (Sieburth, Johnson and Har- graves) (“brown tides’’) in eastern Long Island, New York waters decimated most of the commercially valuable stocks of bay scallops, Argopecten irradians irradians (Lamarck) (Anomymous, 1985; Siddall, 1986). Recruitment of larval scallops in these areas also appeared to be very poor during both 1985 and 1986 (Siddall and Nelson, 1986; Bricelj et al., 1987a). These events have virtually eliminated the bay scallop fishery on Long Island, an industry which in 1984 produced approximately one-fourth of the total United States harvest and employed from 400—600 full-time baymen (Anomymous, 1985). Bay scallop reseeding efforts were initiated by the Long Island Green Seal Committee in 1986 (Wenczel et al., 1986). Selection of areas for reseeding scallops has largely been based on the premise of reducing potential losses to starfish (Asterias forbesi (Desor)) and crab predation (Wenczel et al., 1986), factors which have been implicated as major sources of juvenile bay scallop mortality (Morgan et al., 1980; Tettelbach, 1986). Other considerations in site selection (Wenczel et al., 1986) have included historical scallop productivity, likelihood of larval settlement fol- lowing spawning at the planting location (Siddall et al., 1986), and protection from stranding of scallops on beaches following winter storms (see Kelley, 1981). Burial of bay scallops by shifting sediments was not ad- dressed in the site selection process in 1986 and 1987. Tet- telbach (1986) observed this phenomenon in the Poquonock scallop, Argopecten, mortality, burial, sediment, transplant, seasonality River, Connecticut and concluded that mortality due to burial probably did not exceed 1% during 1983-1984. Burial was only observed during December—March, and not during the rest of the year. Stewart et al. (1981) also suggested that bay scallop mortality resulted from burial in Connecticut, and during the winter of 1986-1987 Smith (unpubl. data) observed extensive burial of transplanted scallops in Orient Harbor, N.Y. In light of recent bay scallop reseeding in Connecticut, Rhode Island, and Massachusetts (Capuzzo and Taylor, 1981; E. Rhodes, pers. comm.) and continuing efforts in New York (Wenczel, 1988), the present study was under- taken to better understand the potential importance of burial to the success of reseeding operations. The major objec- tives were to quantify the prevalence of scallop burial and resultant mortality at a selected site, to observe the progres- sion of burial of individual scallops, and to examine the effect of different sediment types (bare sand, sand and eel- grass, muddy-sand) on rates of burial and mortality. MATERIALS AND METHODS This study included both field and laboratory experi- ments. Field experiments utilized tagged bay scallops to study the progression and prevalence of burial between De- cember 1987—March 1988. Laboratory experiments exam- ined the survival of scallops placed in various states of par- tial and complete burial. Data from both sets of experi- ments were combined to estimate the contribution of burial to mortality in the field. 127 128 Field Studies The study site was located off the western shore of Northwest Creek, East Hampton, New York (Fig. 1). The area is situated within several hundred meters of two active reseeding sites for bay scallops. Tagged bay scallops were released into three replicate areas on each of three bottom types (MLW depth = 1.5—2 m) which were judged to have qualitative differences in sediment composition: sand and eelgrass (<200 stalks/m?), sand, and muddy-sand (substrates 1, 2, and 3, respec- tively). The release areas were circles of 2 m in diameter. Replicate areas for a given bottom type were placed 5 m apart in a N-S direction. Substrates 1 and 3 were located about 50 m from shore, with the latter areas approximately 25 m to the south of substrate 1. Substrate 2 was located about 10 m farther offshore of substrate 1. A bed of dense eelgrass (>200 stalks/m*) was located 1—2 m inshore of the release areas for substrates 1 and 3. The sediment sur- faces of all release areas were cleared of stones, shell hash and eelgrass debris immediately prior to the release of scallops. Samples of surficial sediment to a depth of 5 cm were taken from each of the nine release circles at the start of the field studies. Grain size fractions, at 0.25 phi intervals, of comeceur ), LI SHELTER ISLAND NORTHWEST \ HARBOR NORTHWEST CREE rs 1000 m om 12a 5 Figure 1. Map of Northwest Creek, showing its location in Long Is- land, New York, and the site of the field studies (@). TETTELBACH ET AL. these samples were examined in the laboratory using a gravimetric settling tube (Gibbs, 1974). An analysis of variance of mean grain size (Folk and Ward, 1957) yielded a significant result (F = 2.24, p < .001), but none of the field samples were found to be significantly different in a Scheffé test (p > .05). Because the sediments were found to be poorly sorted, an additional ANOVA was done to examine whether the percentage (by weight) of grains smaller than a phi size of 3.0 (very fine sand, silt, clay) differed among the release circles (see Eckman, 1987). After this second ANOVA (F = 7.12, p < .001) a Scheffé test revealed no significant differences (p > .05) between the field sediments. All scallops used in field and laboratory experiments were obtained from a commercial growout facility in New York, and ranged from 19.5—25.4 mm in height (as mea- sured from the umbo to the ventral margin of the shell). The mean height of scallops used in field or laboratory ex- periments was 21.8 mm (SE = 0.14, n = 115). Prior to their release in the field, all scallops were tagged with a strip of orange, plastic surveyors tape (8—11 cm long and 0.5 cm wide) to which a small piece of cork (0.5—1.0 cm?, 0.3 cm thick) was attached to one end. The other end of the tag was then attached to the upper valve of the scallop. A marine epoxy (Polypoxy underwater patching compound®, Pettit Paint Co., Inc., Rockaway, N.J.) was used to attach the corks and tags. A unique number code was printed on both sides of the tags. Tagged scallops were released in Northwest Creek on 21 December 1987. Between 95—115 scallops were released underwater into each of the nine circles, for a total of ap- proximately 1000. Six 3-mm mesh pearl nets, each con- taining 7—12 identically-tagged scallops, were also de- ployed on the above date. These served as a control to esti- mate sources of mortality other than burial and predation. Resampling was initially done on 31 December 1987, and then semimonthly between 3 February and 31 March 1988. Divers recorded in situ observations of scallops within release circles and at distances up to 100 m to the ENE and WSW of outlying release areas. All scallops ob- served were classified as live unburied, 3 buried (4 of the shell covered by sediment) (Fig. 2), 74 buried, totally buried (Fig. 3), or dead unburied. Counts also were made of scallops which showed visible signs of predatory attack in the form of crushed or drilled shells, or chipped shell margins (Carriker, 1951; Boulding and Hay, 1984; Tettel- bach, 1986), or which were being actively attacked by star- fish (which leave no traces of shell damage). If scallops were found on a bottom type other than that on which they were initially released, this was also noted. Counts of live and dead scallops in pearl nets were made on each sampling date, and nets were cleaned when necessary. Bottom water temperature (at a depth of 0.5 m) was recorded with a diving thermometer, and surface salinity was measured with an optical refractometer. Current velocity was mea- WINTER BURIAL OF BAY SCALLOPS 129 Figure 2. Tagged bay scallops which were live unburied (left) and partially (/) buried (right) on 1 March 1988 at Northwest Creek, N.Y. The epoxy which bonds the tag to the scallop shell can be seen on the individual to the right. Tag numbers can be seen at the top. sured 10 cm above the bottom (substrate 2) with a rotary current meter (General Oceanics, Inc.) on 18 March. Since one of the objectives of the study was to follow changes in the state of burial in which scallops were found, individuals were not handled prior to the last sampling data. Rather, assessments of whether they were alive or dead were made visually. On 31 March 1988, buried scallops were exhumed and then examined to see if they were alive. Statistical analyses of differences in the observed numbers of totally buried scallops on the three bottom types were conducted for each sampling date. Data for replicate release circles were pooled and subjected to G-tests (Sokal and Rohlf, 1969). Unpooled data, expressed as arcsine- transformed proportions (Zar, 1984) of scallops which were totally buried, were examined via the more conservative analysis of variance (see Peterson, 1982). Because of the propensity of bay scallops for dispersal (Moore and Mar- shall, 1967; Morgan et al., 1980; Tettelbach, 1986), pro- Figure 3. Tagged bay scallops which were completely buried on 31 December 1987 at Northwest Creek, N.Y. Note the pieces of cork which buoy the distal ends of the tags. portions are based on only those scallops which were actu- ally observed. Laboratory Studies Laboratory experiments to examine survival rates of par- tially and completely buried scallops in two types of sedi- ment were conducted in running ambient seawater tables (247 cm long X 64 cm long x 30 cm deep). One table contained a 4 cm layer of sand obtained from the study site, which was sieved to remove pebbles, shell hash, and or- ganic detritus. The other table utilized a 6 cm layer of muddy-sand which was sieved after removal from Old Fort Pond in Southampton, N.Y. The mean grain sizes of the two sediments were not significantly different (Scheffé test, p > .05) from the field sediments or each other. The per- centage of grains larger than a phi size of 3.0 was signifi- cantly greater (Scheffé test, p < .05) in the muddy-sand (mean = 19.0%) than in the laboratory sand (mean = 0.1%) and field substrates 1 and 2 (mean = 0.4, 0.4%), but not significantly different (p > .05) from field substrate 3 (mean = 1.7%). In each table, 36 scallops were used for each of seven different treatments: unmarked unburied (UU), marked un- buried (MU), unmarked 4 buried (14), unmarked 7 buried (74), unmarked and buried by 1 cm of sediment (UB1), marked and buried by | cm of sediment (MB1), and un- marked and buried by 3 cm of sediment (UB3). Tags were identical to those used in the field. The area of each of the two tables, except along the walls, was partitioned into a grid system such that each scallop was placed into a 7 cm x 7 cm area and buried to the appropriate degree. Grid positions were randomly selected. Scallop activity and water temperature were monitored daily from 14 January—24 March 1988. Water temperature ranged from 0.3—9.0°C, salinity remained near 26.5 ppt, and flow rate into the tables was 9 l/min. Any scallop that moved and/or unburied itself was replaced to its proper po- sition. The frequency of movement of UU, MU, and par- tially buried scallops out of their respective grid positions was recorded daily from 19 January—13 February (T = 0.3—4.8°C). Data were analyzed by G-test to determine if tags affected scallop movement and thus the ability of par- tially buried scallops to unbury themselves. During the 10-week survival experiments, sampling was conducted every 7 days. At this time, three scallops of each of the seven burial treatments were removed from each table. These were classified as dead or alive on the basis of whether they showed obvious movement or responded to tactile probing of the tissues. G-tests were performed to examine whether survival differed with respect to substrate type, presence of tags, or burial treatment. Survival of marked and unmarked scallops buried under 1 cm of sand or muddy-sand was examined daily in a sepa- rate experiment. Twenty-one scallops of each of the two groups were introduced to each water table on 18 March. 130 TETTELBACH ET AL. On each of the next seven days, three marked and three unmarked individuals were exhumed and examined as above. Water temperature ranged from 2.8—9.9°C. Estimates of Field Mortality Results of field and laboratory studies were used to esti- mate mortality of transplanted scallops in Northwest Creek. Field mortality on a given sampling date was calculated as the # of dead scallops on that day (D;) + total number of dead and live scallops observed on that day (D; + Aj), where D; = # being eaten by predators + # unburied dead + (# completely buried) x (m,) + (# partially buried) x (m,). The quantities m, and m, represent the proportions of completely and partially buried scallops, re- spectively, which were dead at a given sampling time. The latter proportion was estimated from survival rates of par- tially buried scallops in the laboratory. Estimates of m, were derived from: (1) the proportion of completely buried scallops in the field which were observed alive on a subse- quent sampling date, and (2) the proportion of completely buried scallops which were alive after they were exhumed on the last field sampling day. To examine whether rates of mortality differed for scallops held in pearl nets and those released onto the bottom, the following equation was utilized: D; 3 Dj-; fe D, + A; DR ae AR GS Gea where: M; = proportion of scallops of a given group dying per day during the interval between sampling dates i andi — 1; = as above; = # dead and live scallops, respectively, at sampling date i — 1; # days since the initial release, at sampling dates iandi — 1, respectively. At tp, 100% of the scallops were alive. Five values of M; were calculated over the period 21 De- cember 1987 to 18 March 1988 for scallops held in pearl nets and for those released onto the bottom (first using esti- mate #1 for m,, and then estimate #2). After an arcsine transformation of the values of M,, two t-tests were per- formed to compare means for caged vs. free scallops. If one or both of these tests was significant, an additional t- test was conducted to examine whether the difference in mortality rates could be due to a greater probability of ob- serving dead scallops than live ones. This test was done by making the extreme assumption that all released scallops which were not observed on a given sampling date, except those that were recorded as dead on a previous date, were alive. Thus D; + A; = D;_, + A,;_,; = 1000. RESULTS Performance of Tags Performance of tags during the study was satisfactory. No tag loss was observed in the field or laboratory. Corks detached from the plastic tape in only six of 772 (0.8%) different individuals which were observed in the field during the three months following transplantation. Field Studies Temperature ranged from — 1.4 to 4.4°C, while salinity remained at 26—27 ppt throughout the study. A 3 cm sheet of ice covered the study area on 31 December 1987, but ice was not encountered on any other sampling dates. Current velocity was 0.04 m/sec at mid-ebb tide on 18 March 1988. Percent recovery of marked scallops summed over all nine release areas (using an initial n = 1000) ranged from a high of 39.4% on 31 December 1987 to a low of 22.5% on 18 March 1988. Scallops were found as far as 75 m to the north and south of the study area. On 31 December, 11 scallops were found in the inshore eelgrass bed; on 31 March, 69 scallops were seen here. After 21 December, cobble, gravel, shell hash, and large Mercenaria shells were noted in and around substrates | and 2; eelgrass de- tritus and shell hash was seen on substrate 3. A combined total of 14.4% (57/395) of the scallops ob- served on 31 December was completely buried, with a range from 0% (0/24) in release area 4 to 29.3% (22/75) in areas 8 + 9. Scallops which were partially buried on the first sampling date accounted for 26.8% (106/395) of all observed individuals; no unburied dead individuals were found on this day (Fig. 4). Totally buried scallops were significantly more abundant on substrate 3 than the other bottom types on 31 December (G = 14.4, p < .001) and 18 March (G = 31.6, p < .001). No significant differences (p < .20) on any of the sampling dates were revealed in the ANOVA except on 18 March, when burial was significantly lower (p < .O1) in substrate 2 than substrates | and 3. While the observed proportion of live unburied indi- viduals from all bottom types declined from 58.9% (232/395) on 31 December to 10.9% (28/258) on 31 March, the overall proportion of unburied dead scallops in- creased from 0% (0/395) to 59.3% (153/258) during this period (Fig. 4). Seven of 92 (7.9%) and two of 80 (2.6%) dead unburied scallops observed on | and 18 March, re- spectively, contained meats. Observed levels of predation during the study period were low. On 3 and 17 February, 0.4% (1/277) and 1.0% (3/301) of the observed scallops, respectively, were ac- tively being preyed upon by Asterias forbesi. No starfish predation was observed on other dates. A total of 11 As- terias was observed from December—March. No visible evidence of crab or gastropod predation on marked scallops was noted, although such species as Callinectes sapidus, Neopanope sayi, Pagurus pollicaris, Ovalipes ocellatus, e @ COMPLETELY BURIED WINTER BURIAL OF BAY SCALLOPS @—e DEAD UNBURIED o—oO LIVE UNBURIED v vy PARTIALLY BURIED 131 100, 0 SUBSTRATE 1 80 | 604 : 404 Tat | ee 4 e = SCT 204 / a= Se " | ] bie ele A Ae cent eine | ye ? 123 108 60 48 40 100, 02 SUBSTRATE 2 Se 804 \ a Sy | bike of 5 604 cae FZ 7 NN > 404 \ | (Ej 1 + LJ J R or on Se a aN | . im Pe a Se Se 100, 0 SUBSTRATE 3 \ e A 1988 4.1 1.0 -1.4 1.4 -0.6 sa 4.4 Figure 4. Temporal changes in burial of tagged bay scallops at the Northwest Creek study site from December 1987—March 1988. Data points represent mean percentages of observed scallops for the three replicate releases on each of the three substrates; vertical lines reflect ranges. When fewer than four scallops from a release circle were found on a given date, counts were pooled with those for the adjacent circle. Numbers of scallops recovered on each sampling date are given at the bottom of the three plots, except for 21 December when n = 333 scallops per substrate type. Water temperatures for each sampling date are given below the third plot. Libinia spp., Urosalpinx cinerea, Eupleura caudata, Busy- cotypus canaliculatus, and Busycon carica were observed. A total of 33 crabs and 12 gastropods were recorded during the study. The states of burial in which individual scallops were found often changed during the study period, such that to- tally buried scallops were sometimes found later on the sed- iment surface. Of 200 completely buried scallops which were observed again, 48 (24%) were unburied and appar- ently alive, 73 (36.5%) remained buried, and 79 (39.5%) were unburied and dead. Of the 109 dead unburied scallops which were observed again, 39 (35.8%) were later found completely reburied. The greatest depth of interment for any individual scallop was 6 cm. No mortality of scallops was observed in pearl nets until 1 March. On this date, mortality in nets averaged 24% (range = 17-29%); on 18 March, mortality averaged 28% (range = 22-33%). Data for 31 March are excluded be- cause mud crabs (Neopanopi sayi) had entered nets and preyed on some of the scallops. Laboratory Studies No differences were observed in the frequency with which MU, UU, and partially buried scallops moved out of 132 TETTELBACH ET AL. their respective grid positions on the sand table (G = 0.76, p > .25). The number of recorded movements out of the possible total from 19 January—13 February 1988 was 33/786 for MU scallops and 39/786 for each of the UU, %, and *4 groups. Movement on the sand table was signifi- cantly greater (G = 174.14, p < .001) than on the muddy- sand table, where only five movements were noted: UU (2/786), MU (1/786), ¥% (0/786), 7 (2/786). The frequency of movement on the sand table at temperatures between 3.6—4.8°C (71/480) was significantly greater (G = 91.70, p < .001) than that at temperatures between 0.3—2.8°C (79/2664). In addition to movements out of their grid position, scallops were seen to clap their valves together while re- maining in a given position—a process which sometimes resulted in the removal of the sediment covering the shell. This was particularly prevalent during the first five days of the experiments. This activity was exhibited by scallops which were partially buried or buried under 1 cm of sedi- ment, but not by UB3 individuals. Unburied scallops were often observed in shallow depressions within their specified grid; this was observed as early as four days after the ex- periment was initiated. Survival of unburied and partially buried scallops was significantly lower (G = 19.38, p < .001) on the muddy- sand than the sand table (Table 1). While there were no differences in the survival of marked vs. unmarked scallops which were unburied (G = 1.86, p > .05), or in the sur- vival of ¥%3 vs. % scallops (G = 0.05, p > .05), the sur- vival of partially buried (% and 7%) scallops was signifi- cantly lower (G = 8.14, p < .005) than unburied (UU and MU) scallops (Table 1). Most of the mortality of unburied (11 of 12 dead) and partially buried (22 of 28 dead) scallops occurred during weeks 7—10. None of the marked or unmarked scallops buried by 1 or 3 cm of sediment were alive after one week. In the 18—25 March experiment, all scallops buried under | cm of either substrate died within | d. All of the scallops examined were gaping and had sediment in the gills and mantle cavity. Estimates of Field Mortality The two estimates of m,, the proportion of completely buried scallops which suffered mortality, were 0.76 and 0.92. The former estimate reflects the observation of 48 live scallops out of 200 completely buried individuals which were seen at least once after they were recorded as buried. The higher value of m, is based on data from 31 March, when 59 of 64 completely buried scallops were found to be dead after they were exhumed. Of the five live individuals, four were covered by eelgrass detritus but were resting on the sediment surface. For mortality calculations, all partially buried scallops were considered to be alive because levels of mortality were low during the first 7 weeks of laboratory experi- ments. Temporal trends in estimated mortality (%) of observed scallops (Fig. 5) suggest that individuals released on the bottom died at a faster rate than those in pearl nets. This difference was barely confirmed when a m, of .92 was used (T = 2.33, p < .05) and was not seen when a m, of .76 was used (T = 2.19, p > .05). When all unobserved scallops on a given sampling date (except those previously seen dead) were assumed to be alive, there was no differ- ence between mortality of caged and free scallops (T = 0.56, p > .50). DISCUSSION The lack of a significant difference between the fre- quencies of movement of marked and unmarked scallops in the laboratory suggests that tags did not affect this be- havior. The question arises, however, as to whether tags affected the rate of burial in the field. If tags acted as baffles, much as eelgrass blades do (Eckman, 1987), sus- pended sediments would be more likely to settle out in the TABLE 1. Survival of scallops held in the laboratory under different burial treatments on two substrates. Cell blocks are total number of scallops recorded during the course of the 10-week experiment for a given treatment. Treatments: UU = unmarked unburied, MU = marked unburied, 3 = % of shell buried, 74 = % of shell buried. G-tests were performed on untransformed data; *** = p < .001, ** = p< .01, * = p< .05, NS = p> .05. Sand UU MU yA) x dead 1 0 5 2 alive 26 30 23 26 Total 27 30 28 28 Hypothesis survival independent of substrate survival independent of treatment survival: UU vs. MU survival: 4 vs. “ survival: UU + MU vs. 4% + Muddy-Sand UU MU Ys % 7 4 9 12 40 20 24 17 17 183 Total 27 28 26 29 223 df G 19.38*** 10.06* 1.86 NS 0.05 NS 8.14** woe —a— yi x WINTER BURIAL OF BAY SCALLOPS 133 100 o——o SUBSTRATE 1(A) e e- — -e SUBSTRATE 2(A) aleas _. 8074. « SUBSTRATE 3(A) u 3 ¥ -¥ COMBINED(B) ped Wolpe 0 — 6047 o—o PEARL(A) - ap ee ‘Z ed =] / 40 + Oy e aes : 7 e -- Oo OQ 20 oS ze rr a = LEE aA ale v ¢) eer ok Se o D J F M A 1988 Figure 5. Comparison of temporal changes in average percent mor- tality for scallops released on the bottom and for those held in pearl nets. Average percent mortality = # of dead scallops + # dead + live scallops: (A) Mortality estimates for scallops in pearl nets and re- leased scallops based only on those individuals which were observed on a given date with 92% of completely buried scallops presumed dead; 3 bottom types considered separately. (B) Mortality estimate for released scallops based on the assumption that all unobserved indi- viduals (except those previously observed to be dead) were alive (total n = 1000); data for 3 bottom types combined. vicinity of the tagged scallops. However, the sediment cov- ering buried scallops instead appeared to be similar to the substrate upon which scallops were released. Various lines of evidence suggest that burial resulted from shifting sediments, rather than particle settlement. The fact that many scallops were buried and later became uncovered strongly supports this contention. Cobble and large shells also were found on the sediment surface in re- lease areas after these types of objects had been removed in December. Because no major storm events occurred during this study (see Engle, 1948; Peterson, 1985), it appears, therefore, that scallop burial resulted from the shifting of sediments due to tidal and wind-driven currents (in spite of the low (0.04 m/sec) current velocity which was recorded). Shulenberger (1970) suggested that wind-driven currents may be responsible for burial and mortality of Gemma gemma. The flexible tags (which averaged 0.38 g in weight) probably neither enhanced nor mitigated the pro- cess of burial. In contrast to the techniques of vertical digging with the foot and extension of the siphons which many infaunal bi- valves use to escape burial (Shulenberger, 1970; Kranz, 1974; Peterson, 1985), epifaunal bay scallops must rely on a completely different behavior because of their greatly re- duced foot and lack of siphons. Kranz (1974) found that bay scallops had limited ability to escape burial, but could remove | cm of sand by clapping their valves. This was confirmed in the present study. Laboratory results from this study suggested that move- ments which resulted in scallops becoming unburied were less frequent at temperatures lower than 2.9°C. Thus, bay scallops appear more susceptible to burial by shifting sedi- ments at low water temperatures. Burial may also be en- hanced by the propensity of scallops to be found in shallow depressions, or potholes, in winter (Stewart et al., 1981; Tettelbach, 1986; this study). G-tests revealed that totally buried scallops sometimes occurred in higher proportions in substrate 3 than substrates 1 and 2. One factor which may affect these results is the relative prevalence of surfaces to which scallops may bys- sally attach. In Northwest Creek, cobble and pieces of shell to which scallops were attached sometimes showed clear indications of scouring around their edges. Thus, attached scallops may suffer less burial compared to individuals lying freely on the bottom. Another consideration which is suggested by the present study is that smaller scallops may be more susceptible to burial than larger individuals. Kranz (1974) did not observe any difference in the ability of small and large bay scallops to escape burial, but details of the sizes used were not re- ported. However, the prevalence of complete burial in this study (14.5—45.3% of observed 20—25 mm individuals) appears to be higher than that in the Poquonock River,Con- necticut, where about 1% of the natural population of scallops was buried (mean height = 40—45 mm) (Tettel- bach, 1986). Since larger scallops have deeper shells (Bri- celj, et al., 1987b) more sediment would likely be neces- sary for complete burial. Other possible reasons for the apparent difference in burial prevalence in the Poquonock River and Northwest Creek may reflect differences in the study sites or history of the scallops. The Poquonock River site was shallower (about | m at MLW) and had a more extensive eelgrass bed than the Northwest Creek study area. Burial in the Po- quonock River was primarily evident in the sandy portions where current speed appeared to be higher than in the eel- grass areas (Tettelbach, 1986). Since scallops in the Po- quonock River were from an extant population, they may have been less stressed and more likely to move to avoid burial than the individuals transplanted into Northwest Creek on 21 December 1988 (initial T = 4.1°C). However, some of the scallops which were moved from other areas of the laboratory into tables used for burial experiments (ini- tial T = 1.3°C) were observed to move and unbury them- selves on the day they were deployed. Given the number of unburied dead and completely buried scallops which were seen in Northwest Creek, it is not surprising that the observed mortality rate of released scallops (when m, = .92) was higher than that of scallops held in pearl nets. The cumulative mortality in nets, how- ever, was unexpectedly high in comparison to that of first year bay scallops held in identical nets in the Poquonock River, CT (<3%) (Tettelbach, 1986) and in suspended cages in Flax Pond (<5%) and Southhold, N.Y. (approxi- mately 8%) (Bricelj, et al., 1987b) during the months of December—March. Conceivably, scallops were predated by mud crabs (Neopanope sayi) before they were first ob- served in nets on 31 March. 134 TETTELBACH ET AL. The relatively low recovery of tagged scallops in North- west Creek (23—39%) confounds a determination of the ac- tual level of mortality due to burial by sediments. Since completely buried scallops are more likely than live un- buried scallops to remain in a given location and be ob- served, estimates of mortality which are based solely on observed individuals are likely to be overestimated. How- ever, unburied dead scallops may be less likely to remain at a specific location than live individuals which are byssally attached. Without knowing the probabilities of recovering live vs. dead or buried vs. unburied scallops, one probably could conclude safely that levels of mortality due to burial are higher than those estimates which are based on the as- sumption that all unrecovered scallops are alive. Even if the latter, highly conservative approach is adopted, the cumu- lative mortality of scallops is still seen to be 30.4% on 31 March and is 13.8% by 17 February (before any mortality was observed in pearl nets). The phenomenon of burial by shifting sediments is seen as a potentially significant cause of bay scallop mortality from December—March. Winter burial clearly should be considered when implementing programs to reseed bay scallop populations. Results of laboratory studies suggest that muddier sites should probably be avoided when planting scallops. The effects of both scallop size and the prevalence of gravel and shell in the transplanting area on rates of burial and mortality should be investigated further. ACKNOWLEDGMENTS We thank Peter Auster, V. Monica Bricelj and three an- onymous reviewers for their comments on the manuscript, and Linda Kallansrude for typing the paper. We also appre- ciate the assistance of James Peterson and Seth Magot of the LIU-Southampton Academic Computing Center, the help of Larry McCormick with the sediment analyzer, and valuable discussions of statistics with Russell Myers and Eric Posmentier. This work is the result of research spon- sored by the NOAA Office of Sea Grant, U.S. Department of Commerce, under Grant #NA86AA-D-SG045. The U.S. Government is authorized to produce and distribute reprints for governmental purposes notwithstanding any copyright notation that may appear hereon. REFERENCES Anonymous. 1985. Algae destroy Eastern L.I.’s scallop harvest. The New York Times, September 14, 1985, pp. 1,28. Boulding, E. G. & T. K. Hay. 1984. Crab response to prey density can result in density-dependent mortality of clams. Can. J. Fish. Aquat. Sci. 41(3):521—525. Bricelj, V. M., J. Epp & R. E. Malouf. 1987a. Intraspecific variation in reproductive and somatic growth cycles of bay scallops Argopecten irradians. Mar. Ecol. Prog. Ser. 36:123—137. Bricelj, V. M., J. Epp & R. E. Malouf. 1987b. Comparative physiology of young and old cohorts of bay scallop Argopecten irradians irra- dians (Lamarck): mortality, growth, and oxygen consumption. J. Exp. Mar. Biol. Ecol. 112:73-91. Capuzzo, J. M. & R. E. Taylor, Jr. 1981. Seeding program for the bay scallop: comparison of local bays, Falmouth, MA. 1980-81 Annual Sea Grant Report, WHOI, Woods Hole, MA. Carriker, M. R. 1951. Observations on the penetration of tightly closing bivalves by Busycon and other predators. Ecology 32(2):73—83. Eckman, J. E. 1987. The role of hydrodynamics in recruitment, growth, and survival of Argopecten irradians (L.) and Anomia simplex (D’Or- bigny) within eelgrass meadows. J. Exp. Mar. Biol. Ecol. 106:165— 191. Engle, J. B. 1948. Investigation of the oyster reefs of Mississippi, Loui- siana, and Alabama following the hurricane of September, 1947. U.S. Fish & Wildl. Serv., Spec. Sci. Rept. 53, 70 pp. Folk, R. L. & W. C. Ward. 1957. Brazos River bar: a study in the signifi- cance of grain size parameters. J. Sed. Petrol. 27:3—26. Gibbs, R. J. 1974. A settling tube system for sand-size analysis. J. Sed. Petrol. 44(2):583—S88. Kelley, K. M. 1981. The Nantucket bay scallop fishery: the resource and its management. Shellfish and Marine Dept., Nantucket, MA, 108 pp. Kranz, P. M. 1974. The anastrophic burial of bivalves and its paleoecolo- gical significance. J. Geol. 82:237—265. Moore, J. K. & N. Marshall. 1967. An analysis of the movements of the bay scallop, Aequipecten irradians, in a shallow estuary. Proc. Natl. Shellfish. Assoc. 57:77-82. Morgan, D. E., J. Goodsell, G. C. Matthiessen, J. Garey & P. Jacobson. 1980. Release of hatchery-reared bay scallops (Argopecten irradians) onto a shallow coastal bottom in Waterford, Connecticut. Proc. World Maricul. Soc. 11:247—261. Peterson, C. H. 1982. Clam predation by whelks (Busycon spp.): experi- mental tests of the importance of prey size, prey density, and seagrass cover. Mar. Biol. 56:159—170. Peterson, C. H. 1985. Patterns of lagoonal bivalve mortality after heavy sedimentation and their paleoecological significance. Paleobiology 11(2):139-153. Shulenberger, E. 1970. Response of Gemma gemma to a catastrophic burial. Veliger 13(2):163—170. Siddall, S. E. 1986. *‘Brown Tide’’ algal blooms of 1985 and 1986. On the Water (Sept./Oct. 1986), Cornell Cooperative Extension Sea Grant Program: pp. 4—S. Siddall, S. E. & C. L. Nelson. 1986. Failure of bay scallop larval recruit- ment during Long Island’s ‘brown tide’’ of 1985. Paper presented at the 6th Annual Shellfish Biology Seminar, March 4, 1986, Milford, Cr: Siddall, S. E., M. E. Vieira, E. Gomez-Reyes & D. W. Pritchard. 1986. Numerical model of larval dispersion—Phase I of the East End Algal Bloom Program. Marine Sciences Research Center Special Report #71. 28 pp. Sokal, R. R. & F. J. Rohlf. 1969. Biometry. W. H. Freeman & Co., San Francisco, 776 pp. Stewart, L. L., P. J. Auster & R. Zajak. 1981. Investigation on the bay scallop, Argopecten irradians, in three eastern Connecticut estuaries, June 1980—May 1981. Final report submitted to U.S. Dept. of Comm., NOAA, NMFS, Milford, CT. Tettelbach, S. T. 1986. Dynamics of crustacean predation on the northern bay scallop, Argopecten irradians irradians. Ph.D. Dissertation, Uni- versity of Connecticut, Storrs, CT, 229 pp. Wenczel, P. 1988. Green Seal bay scallop reseeding activities for 1988-1989. Proposal submitted to the Urban Development Corp., New York, N.Y. Wenczel, P., J. Scotti & C. Smith. 1986. Long Island Green Seal Com- mittee bay scallop rehabilitation program: plan of work. Report sub- mitted to the Urban Development Corp., New York, N.Y. Zar, J. H. 1984. Biostatistical Analysis, 2nd Ed. Prentice-Hall, Inc., En- glewood Cliffs, N.J., 718 pp. Journal of Shellfish Research, Vol. 9, No. 1, 135-144, 1990. GROWTH AND SURVIVAL OF JUVENILE JAPANESE SCALLOPS, PATINOPECTEN YESSOENSIS, INNURSERY CULTURE D. O FOIGHIL', B. KINGZETT?, G. 0 FOIGHIL? AND N. BOURNE? 'Bamfield Marine Station* Bamfield British Columbia Canada VOR 1BO 2Department of Fisheries and Oceans Biological Sciences Branch Pacific Biological Station Nanaimo, British Columbia Canada V9R 5K6 3Department of Zoology University of Aberdeen Aberdeen AB9 2TN U.K. ABSTRACT Growth and mortality rates of newly metamorphosed Japanese scallops, Patinopecten yessoensis, were analyzed in nursery systems. Higher numbers of juveniles attached to lightly fouled cultch and the presence of an epifloral film significantly enhanced early post-metamorphic growth. The nutritional importance of epifloral films for juvenile growth was, however, much less than that of suspended microalgae. Post-larval performance was monitored by regular subsampling of nursery cultures and by exam- ining death assamblages collected from the bottom of setting tanks. Two periods of heightened juvenile mortalities were discovered. Heavy losses occurred in most nursery sets before the juveniles attained 400 jm of dissoconch shell growth, especially in sets that experienced the greatest metamorphic success. Surviving juveniles encountered a second critical period, characterized by reduced growth rates and heavy mortality rates, by the time they attained 600 zm in dissoconch height. Diet supplementation with wild microalgae enhanced juvenile performance during this latter period, indicating that the cultured microalgal diet employed may have been qualitatively inadequate. A thorough investigation of the ontogeny of suspended particle capture ability in juvenile P. yessoensis is needed to allow the rational design of an optimized nursery system for this species. KEY WORDS: INTRODUCTION The feasibility of culturing the Japanese scallop, Patino- pecten yessoensis, is being examined in British Columbia (B.C.) because of its proven commercial success in Japan (Taguchi, 1978; Ventilla, 1982; Kafuku and Ikenoue, 1983; Ito, 1990). However, it is not native to the north- eastern Pacific and its commercial future there hinges on developing reliable hatchery technology to produce ade- quate quantities of juveniles. Pilot culture of this species has been performed by the Department of Fisheries and Ocean Pacific Biological Station (PBS) in conjunction with the B.C. Ministry of Agriculture and Fisheries. Consider- able progress has been made, especially in larval produc- tion, and results from grow-out sites have been encour- aging (Bourne et al., 1989). Post-metamorphic juvenile survival is frequently, however, very poor, often much less than 5% (Bourne and Hodgson, in press). This is a *Present Address: Department of Biological Sciences Simon Fraser Uni- versity Burnaby, British Columbia Canada V5SA 1S6 135 Japanese Scallop; Patinopecten yessoensis; juveniles; nursery culture; growth; survival common, but poorly understood, problem in bivalve cul- ture (Loosanoff et al., 1966; Castanga and Duggan, 1971; Tremblay, 1988) and it presently represents one of the main technical obstacles to the development of a viable hatchery-based scallop culture industry in B.C. (Bourne and Hodgson, in press). Two approaches have been taken in an attempt to solve this problem. Because post-metamorphic development is initially fueled by energy reserves sequestered in the plank- totrophic veliger larval stage, the scallop culture research group at PBS has focused on optimizing larval diets of cul- tured microalgae (Whyte et al., 1987, 1989). This present study aimed to complement that research by concentrating on the relative growth and survival rates of juvenile P. yes- soensis under different nursery conditions. In most bivalve species, including scallops, the foot per- sists after metamorphosis and reaches its greatest allometric importance during the juvenile stage of development (Sastry, 1965; Hodgson and Burke, 1988). In addition to its obvious role in locomotion and attachment, the relatively large foot also helps create the anterior inhalent mantle 136 FOIGHIL ET AL. cavity circulation typical of juveniles and has been impli- cated in a form of indirect deposit feeding described for a variety of bivalve species (Caddy, 1969; Bayne, 1971; Aabel, 1983; King, 1987) termed ‘‘pedal-palp feeding by King (1987). This little-studied feeding mode may play an important role in some species by filling the nutritional gap between the loss of the velum at metamorphosis and the development of effective suspension feeding juvenile gills. Behavioural studies of newly metamorphosed P. yessoensis have established that a low incidence of pedal-palp feeding on deposited material occurs during early benthic ontogeny (D. O Foighil, unpubl.). This raised the possibility that the presence of an epiflora or organic film on cultch may be of nutritional significance to this species during the first weeks of benthic existence. Post-metamorphic scallops differ from other cultured bi- valve species in that pectinid gills typically lack eulatero- frontal cirri and laterofrontal tracts are composed solely of unbranched cilia (Owen and McCrae, 1976). Adult scallops capture relatively large suspended particles (SO—200 p.m) with high efficiency (Chipman and Hopkins, 1954; Jgrgensen, 1960) and are capable of ingesting particles up to 950 wm in diameter (Mikulich and Tsikhon-Lukanina, 1981; Shumway et al., 1987). With one exception (Vahl, 1973), studies have consistently found pectinid gills to be relatively inefficient at capturing smaller suspended par- ticles (<5 ym), achieving maximum filtration efficiency for particles >5—7 pm in diameter (Vahl, 1972; Mghlen- berg and Riisgard, 1978; Palmer and Williams, 1980; Cranford and Grant, 1990; Lesser et al., 1990). In contrast, maximum filtration efficiency is attained by oyster species when particles exceed 3 wm in diameter (Mghlenberg and Ruisgard, 1978; Palmer and Williams, 1980). The different suspension-feeding characteristics of scallop gills have practical implications for pectinid nursery culture. Leighton and Phleger (1981) found that juvenile rock scallops, Cras- sadoma gigantea, (previously known as Hinnites multiru- gosus) fed cultured microalgal species experienced poorer growth than juveniles raised in field nurseries feeding on wild microalgae, with the exception of individuals fed a relatively large species of cultured microalga Rhodomonas lens (8—14 jm in cell diameter). Juvenile scallops fed wild microalgae encounter a large variety of potential food species of varying cell sizes. A much lower variety of cul- tured microalgal species is available and it is possible that some of the smaller microalgal species, which are suitable food for larval pectinids, are inadequate sources of nutri- tion for juvenile scallops. Our research aims in this study were to 1) assess the importance of cultch epiflora in the diet of early juvenile P. yessoensis; 2) compare the effect of a cultured microalgal diet and a wild microalgal diet supplemented with cultured microalgae on growth and mortality rates of juveniles and 3) pinpoint the onset of pronounced juvenile mortality in nursery systems. MATERIALS AND METHODS All specimens of P. yessoensis employed in this study were reared at the experimental scallop hatchery of the Pa- cific Biological Station. Competent larvae were set on con- ditioned, artificial ‘‘Kinran’’ fibre cultch for 24 hours at 15°C in | pm cartridge-filtered seawater while being fed cultured microalgae as detailed in Bourne et al. (1989), (equal mixture of the TX strain of /sochrysis galbana, Chaetoceros calcitrans and the 3H strain of Thalassiosira pseudonana to give a final concentration of 15,000 cells/ ml). Newly metamorphosed spat (260—280 jm in length) were removed by agitating the cultch. A choice test was conducted to test the attractiveness of diatom mats to newly metamorphosed juveniles. Two thou- sand juveniles were placed in a finger bowl with 30, 10 cm lengths of conditioned Kinran that were lightly fouled with one week of diatom growth. An equal number of juveniles were added to a second finger bowl containing control lengths of Kinran that were first soaked in a 1% solution of sodium hypochlorite for 20 minutes to remove adhering or- ganic material, then thoroughly washed in distilled water. Juveniles were incubated with the test substrates at 15°C overnight in 0.45 jm filtered seawater without the addition of cultured microalgae. The Kinran strands were then re- moved and the numbers of juveniles that had adhered to them by byssal threads enumerated. Results were statisti- cally assayed using a Student T-test. A series of experiments were performed in order to eval- uate the importance of an epifloral film on cultch for juve- nile P. yessoensis growth and survival. Newly metamor- phosed juveniles were cultured for a week in 1 liter jars containing 0.45 jm filtered seawater under the following conditions: A) attached to clean kinran (sodium hypochlo- rite treatment) without added microalgae; B) same as A but with added microalgae (1.5 x 10* cells/ml/day on an equal mixture of cultured TX strain of /sochrysis galbana and the 3H strain of Thalassiosira pseudonana); C) kinran condi- tioned for one week in running seawater tables without added microalgae and D) same as C but with added mi- croalgae (identical ration to B). Water temperatures were maintained at 15°C throughout the experiment, daily water changes were made and treatments A and B were main- tained in subdued light to prevent epifloral diatom growth on the cleaned cultch. After one week, the attached juve- niles were fixed and their dissoconch shell heights were measured to the nearest 4 wm. Growth in shell height was compared among the different treatments using a Tukey Studentized Range Test. Juvenile growth and survival were analysed in three 2,500 liter setting tanks representing different treatment combinations and inoculated with 1.4 x 10° pediveliger P. yessoensis larvae in mid May 1989. In Tank S2 the larvae were set on Kinran cultch that had been conditioned in the dark and the tank was covered to minimize the develop- GROWTH OF JUVENILE JAPANESE SCALLOPS IN CULTURE ment of an epifloral film. Cultch in Tank S3 was condi- tioned in the light and the tank was uncovered to promote epifloral growth. Juveniles in both $2 and S3 were fed only with cultured microalgae. Both tanks were equipped with a circulating pump and underwent water changes twice weekly with filtered seawater. They were fed cultured mi- croalgae (initially 1.5 x 10* cells/ml/day of an equal mix- ture of cultured TX strain of /sochrysis galbana and the 3H strain of Thalassiosira pseudonana, rising to 5 x 104 cells/ml/day after 2 weeks of nursery culture) as detailed by Bourne et al. (1989). In the third tank, B1, this diet of cultured microalgae was supplemented with a constant flow of coarsely-filtered (50 jm) Departure Bay surface sea- water taken from a 3 meter depth that contained wild mi- croalgae. At weekly intervals subsamples consisting of three 15 cm lengths of Kinran were taken from each tank. The numbers of juveniles on each subsample were deter- mined to estimate the flux in juvenile numbers over time. A random sample of 50 juveniles per length of cultch was measured using an ocular micrometer, yielding a total sample of 150 individuals per tank per weekly interval. Data on juvenile size with time provided an estimate of growth rates in the nursery systems. Death assemblages (dead valves) were siphoned from the bottom of the nursery tanks. Total numbers of the re- covered valves were estimated volumetrically and sub- sample dissoconch heights were measured to construct size-frequency histograms. This recovery method was not developed in time to sample the setting runs examined in detail but nevertheless yielded valuable supplementary data on larval and juvenile mortalities in 3 later sets. In one of these sets (Tank S1, 1.4 x 10° larvae set on June 17th) 6% of the larvae added grew to an advanced stage of juvenile development with a mean size of 1.3 mm (84,000 out of 1.4 x 10°). However, both other sets (Tank $2, 1.5 x 10° larvae set on June 30th; Tank B2, 1.5 x 10® larvae set on September | 1th) suffered much higher mortalities during the nursery phase. Juveniles in both of the *‘S’’ tanks were held in filtered seawater and fed only cultured microalgae. Juveniles in the B2 tank had their diet supplemented with coarsely-filtered (SO zm) surface seawater containing wild microalgae. Death assemblage size distributions were com- pared using a Tukey’s Studentized Range Test. RESULTS Importance of Epiflora Results of the juvenile choice experiment are given in Figure 1. A significantly higher number of newly meta- morphosed juveniles attached to the fouled cultch relative to the cleaned kinran (P < 0.05). Figures 2a—2d show how the presence of an epifloral cover on kinran, and the addi- tion of cultured microalgae, affect dissoconch growth rates in the first week after metamorphosis. All treatments dif- fered significantly from each other (P < 0.05). As ex- 137 NUMBER PER STRAND 100 60 4 poner eanrmed metre np 40 4 20 4 ee = CLEAN CULTCH FOULED CULTCH Figure 1. Results of the choice experiment using newly-metamor- phosed P. yessoensis placed with 30 equal lengths of clean (mean = 12.97 + 6.93 S.E.; N = 30) and fouled (mean = 59.93 + 28.03 S.E.; N = 30) kinran cultch (P < 0.05). pected, the addition of cultured, suspended microalgae (Figs. 2c, 2d) dramatically increased growth rates over starved controls (2a, 2b). However, presence of a light epi- floral cover on cultch (Figs. 2b, 2d) resulted in a small, but significant (P < 0.05), enhancement of juvenile growth during the first week of benthic life relative to controls grown on clean cultch (Figs. 2a, 2c). Treatments in which a light epifloral cover was present grew to a greater mean size and growth enhancement was particularly apparent for smaller individuals. Analysis of the death assemblages of these experiments confirmed that the absence of individuals with dissoconch heights <30 ym from treatments with fouled cultch (Figs. 2b, 2d) was due to differential growth, not differential mortality. There is no evidence that the presence of an epifloral film enhances juvenile growth after one week post metamorphosis (Figs. 3a, 3b). Growth and Mortality in Nursery Tanks Figures 3a, 3b and 3c respectively summarize the re- spective changes over time in numbers and growth rates of juvenile P. yessoensis juveniles in settlement tanks S2, S3 and B1. Settlement density of juveniles on the cultch was much higher in Tanks $2, S3 than in Tank B1. However, there was a marked reduction in juvenile numbers in the former two tanks during 2—6 weeks post-settlement, espe- cially in Tank S3, where the population experienced pro- 138 FOIGHIL ET AL. 2A) CLEAN\STARVED > PERCENT Be et Lo} 40 80 120 160 200 DISSOCONCH HEIGHT (um) 2B) -OULED\STARVED PERCENT 160 200- 10) 40 80 120 DISSOCONCH HEIGHT (um) 2C) CLEAN\FED PERCENT | all fe) 160 200 DISSOCONCH HEIGHT (um) 2D) FOULED\FED PERCENT 0) 40 80 120 160 200 DISSOCONCH HEIGHT (um) Figure 2a. Size frequency histograph of newly metamorphosed P. yessoensis that were held for one week on clean Kinran cultch without feeding. Mean dissoconch height = 55 y»m. N = 136. 2b. Size frequency histograph of newly metamorphosed P. yessoensis that were held for one week on fouled Kinran cultch without feeding. Mean dissoconch height = 79 um. N = 174. 2c. Size frequency histograph of newly metamorphosed P. yessoensis that were held for one week on clean Kinran Cultch while fed cultured microalgae. Mean dissoconch height = 126 pum. N = 225. 2d. Size frequency histograph of newly metamorphosed P. yessoensis that were held for one week on fouled Kinran Cultch while fed cultured microalgae. Mean dissoconch height = 160 um. N = 209. nounced mortalities during week 4 (Fig. 3b). The relative loss of juveniles during this period in Tank B1 is much less pronounced (Fig. 3c). Growth rate data have been calculated from weekly size frequency histograms for each tank. Growth rates for all tanks are similar at the beginning and end of the nursery phase but show crucial differences in weeks 4—6 of nursery culture. For the first 3 weeks, juveniles in all tanks had relatively modest growth rates. The negative growth in Tank S3 (Fig. 3b) in the fourth week after metamorphosis is an artifact due to massive losses, predominently of larger juveniles. Tanks S2 (Fig. 3a) and B1 (Fig. 3c) both show accelerated growth in this period, B1 maintaining this rate for the subsequent week, but experiencing little growth in week 6. Tanks S2 and S3 show little growth in either week 5 or 6. This result is especially noteworthy in Tank 3 where juvenile numbers were by then greatly depressed and the available microalgal ration per surviving juvenile was markedly increased. Juveniles in all three tanks exhibited the greatest growth rates in week 7. Death Assemblages Figures 4a, 4b and 4c show the results of the death as- semblages in three separate nursery tanks (B2, S2 and SI respectively). Recovery rate estimates of the efficiency of this method can be calculated by dividing the estimated losses (difference between the number of larvae added to the tank from the number of live juveniles recovered) by the number of dead valves retrieved. For Tank S1 (Fig. 4c) the estimated recovery of the dead valves was 60% (1.4 xX 10° — 84 x 107/79 x 104). For tanks B2 (Fig. 4a) and S2 (Fig. 4b) the respective estimates were 11.1% and 46.1%. 3A) NUMBER PER DISSOCONCH HEIGHT (um) SAMPLE (x1000) 6 4 2 O oO L) O ZZ 7 O EZZZZZIZZZZAA \\ qi A Lo Ny sn 2500 2000 — 1500 5 1000 ~ 500 5 3B) 2500 2000 — Figure 3. (See legend on following page.) 139 140 FOIGHIL ET AL. 3C) 2500 > 10 2000 - ae ee 1500 - -6 1000 500 0 ZZ NUMBER/SAMPLE —=~ DISSOCONCH HEIGHT Figure 3a. Change in numbers and in height of new shell (dissoconch) of juvenile P. yessoensis grown in the dark and fed cultured microalgae in Tank S2. Mean figures and standard errors are presented (+’s for number S.E.; squares for dissoconch height S.E.). Figure 3b. Change in numbers and in height of new shell (dissoconch) of juvenile P. yessoensis grown in the light and fed cultured microalgae in Tank S3. Mean figures and standard errors are presented (+’s for number S.E.; squares for dissoconch height S.E.). Figure 3c. Change in numbers and in height of new shell (dissoconch) of juvenile P. yessoensis fed cultured microalgae, supplemented with wild microalgae after two weeks, and raised in an uncovered tank (B1). Mean figures and standard errors are presented (+’s for number S.E.; squares for dissoconch height S.E.). The former low number may result from the increased debris loading associated with the coarsely-filtered sea- water intake in this particular tank. The pie charts accompanying Figs. 4a—4c show the pro- portion of larvae that died prior to completing metamor- phosis and initiating dissoconch shell formation. Larval mortality rates give a valuable indication of larval quality and sets in Tanks B2, S2 and S1 respectively experienced 39.6%, 48.5% and 68.3% mortalities. Interestingly, larval performance at metamorphosis did not accurately predict juvenile vigour in the three tanks. All three tanks had sig- nificantly different death assemblage size distributions (P < 0.05), however, there was a dichotomy in the results. Juveniles in Tanks B2 and S2 suffered very heavy losses by the time they reached a mean dissoconch height of 400 wm (Figs 4a, 4b). Tank S1, which had the heaviest loss of larvae, resulting in a 6% survival to an advanced juvenile stage and produced a markedly different death assemblage profile (Fig. 10c). After substantial mortalities in the first 60 wm of dissoconch growth, the juveniles showed good survival until they exceeded 600 zm of post-metamorphic growth. The most striking feature of this death assem- blage is the dramatic increase in mortality rates in the 600—1,000 .m size range, forming a normal curve. Inter- estingly, this corresponds with the period of reduced growth rates in the previous nursery runs (see Figs. 3a—3c). DISCUSSION Early juvenile P. yessoensis exist in a transitional tro- phic stage of development between the loss of the velum at metamorphosis and the morphogenesis of effective suspen- sion feeding juvenile gills. At the commencement of the study it was hypothesized that pedal-palp feeding on depos- ited material might be nutritionally important for early ju- veniles. Results obtained in this study show that more indi- 4A TOTAL SHELLS RECOVERED 305 240 420 600 780 960 1140 1320 60 TOTAL SHELLS RECOVERED 4B The 1140 1320 t 960 Figure 4. (See legend on following page.) a imeaaliia 780 240 420 600 60 30 5 25 4 T T T je) wo (e) N ve — (o) o~ 141 142 FOIGHIL ET AL. TOTAL SHELLS RECOVERED DISSOCONCH HEIGHT (um) Figure 4a. Size distribution of the death assemblage of a P. yessoensis set (September 11th 1989) recovered from the B2 nursery tank. Dark section of the pie chart shows the proportion of larvae that died prior to completing metamorphosis (39.6%). Histograms show size distribution of the remaining individuals that died after metamorphosis (mean dissoconch height = 226 1m; N = 303). Figure 4b. Size distribution of the death assemblage of a P. yessoensis set (June 30th 1989) recovered from the S2 nursery tank. Dark section of the pie chart shows the proportion of larvae that died prior to completing metamorphosis (48.5%). Histograms show size distribution of the remaining individuals that died after metamorphosis (mean dissoconch height = 274 um; N = 266). Figure 4c. Size distribution of the P. yessoensis death assemblage recovered from the S1 nursery tank (June 17th 1989 set) in which a 6% recovery of live juveniles was recovered. Dark section of the pie chart shows the proportion of larvae that died prior to completing metamor- phosis (68.3%). Histograms show size distribution of the remaining individuals that died after metamorphosis (mean dissoconch height = 757 pm; N = 334). viduals attached to lightly fouled cultch and that the pres- ence of an epifloral film does significantly enhance early juvenile growth. However, it is obvious that the nutritional importance of epifloral films is much less than that of sus- pended microalgae for early juvenile growth (Figs. 2a—2d, 3a, 3b) and that pedal-palp feeding is a vestigial behaviour in this species. Available data on juvenile P. yessoensis growth and mortality in nursery culture present a complex and some- times contradictory picture. In many respects the most useful data were obtained from examining death assem- blages from the three nursery tanks. Larval quality was generally adequate to yield a satisfactory rate of metamor- phosis and two distinct patterns of juvenile mortality were observed. Juveniles in some sets experienced high mortality rates prior to achieving 400 ym of dissoconch growth (Figs. 10a, 10b). This corresponds to the first three weeks of nursery culture (Figs. 3a—3c) and is probably the period of major juvenile losses in most sets. P. yessoensis juveniles reach their lowest nutrient reserves by approximately 2—3 weeks after metamorphosis (lan Whyte, pers. comm.). This implies that juvenile feeding during this period is inad- equate to prevent the gradual exhaustion of energy reserves accumulated during the late larval stages (Whyte et al., 1987). The modest growth rates in dissoconch heights re- corded in the first weeks after metamorphosis, may there- fore not accurately reflect the nutritional status of juveniles. One approach to minimize early juvenile mortality is to maximize larval energy reserves prior to metamorphosis by optimizing larval diets. This has been pursued with consid- erable success by the PBS scallop research group, how- ever, as Figs. 3a—3c and 4a—4c demonstrate, much prog- ress still remains to be made before consistently high sur- vival of early juveniles can be achieved. Larvae set in Tank S1 on June 17th produced a 6% yield of >1 mm juveniles. Examination of the death assemblage for this exceptional setting run revealed that early juveniles (<400 jm in dissoconch height) did not experience high mortality rates (Fig. 10a). The reason(s) for this are not GROWTH OF JUVENILE JAPANESE SCALLOPS IN CULTURE 143 readily apparent because the early juveniles were treated exactly the same as those in many less successful setting runs, including that depicted in Fig. 4b. One possibility is that the larvae set in SI had exceptional nutrient reserves, however, the greater proportion of larvae that failed to complete metamorphosis in this set (Fig. 4c), relative to other sets (Fig. 4a, b), indicates that this may not have been the case. Data presented in Figs. 4a—4c reveal that the onset of significant juvenile mortality was earliest in sets that experienced the greatest metamorphic success and, consequently, the lowest ration per juvenile. Presently em- ployed microalgal rations during this critical period may simply be too low to promote high survival in most sets. Juvenile growth rate data (Figs. 3a, 3c) and the death assemblage profile of the most successful set (Fig. 4c) both indicate that a second period of juvenile P. yessoensis vul- nerability occurred in the nursery systems at 4—6 weeks post-metamorphosis, when the juveniles attained 600—1,000 zm of dissoconch growth. Depressed growth rates were experienced by juveniles in this size range (Figs. 3a, 3c) indicating that nutritional deficiencies may exist at this stage. It is interesting to note that both the mortality rates and the reduction of growth rates during this second critical period was least when the cultured microalgal diet was supplemented with wild microalgae after the second week, as was the case in Tank B1 (Fig. 3c). Juveniles in this tank could feed on a much greater variety of microalgal species and sizes than could sibling juveniles in Tanks S2 and S3 (Figs. 3a, 3b). Two possibilities exist: 1) cultured microalgal diets were quantitatively insufficient and the ad- ditional biomass of wild microalgae enhanced survival; 2) cultured microalgal diets were qualitatively inadequate and the greater diversity of wild microalgae enhanced juvenile survival. The depressed growth rates of surviving juveniles in Tank 3 in weeks 5 and 6, despite a greatly enhanced ration per juvenile, provides preliminary support for the latter hypothesis. As reviewed in the Introduction, adult scallops are typi- cally much less efficient than other cultured bivalves at fil- tering particles <5—7 wm in diameter (Vahl, 1972; M@hlenberg and Riisgard, 1978; Palmer and Williams, 1980; Cranford and Grant, 1990; Lesser et al., 1990). The filtering capability of the developing gills of early juvenile P. yessoensis are inferior to that of later developmental stages but nothing is known about their particle size capture efficiency profiles. Juvenile gills may prove to be even poorer at small particle capture than are those of adult scallops and a thorough investigation of the ontogeny of suspended particle capture ability in juvenile pectinids is needed. Leighton and Phleger (1981) found that the large microalgal species Rhodomonas lens (8—14 .m in cell di- ameter) was the best cultured algal food for juvenile rock scallops, Crassadoma gigantea. Obviously the nutritional content of the cultured microalgae is very important, how- ever, high nutritional status of microalgae is of limited ben- efit to the juveniles if inappropriate cell dimensions hinder efficient particle capture. ACKNOWLEDGEMENTS Our thanks to Linda Townsend for her assistance with this project. Bamfield Marine Station generously made re- search facilities available for much of the work. Financial support was provided by Science Council of British Co- lumbia Grant #81 (RC-18) and a Bamfield Marine Station Research Fellowship to D. O Foighil. Support for B. King- zett was provided by the British Columbia Ministry of Agriculture and Fisheries. BIBLIOGRAPHY Alabel, J. P. 1983. Morphology and function in postmetamorphal Abra alba (Bivalvia: Tellinacea). Sarsia 68:213—219. Bayne, B. L. 1971. Some morphological changes that occur at the meta- morphosis of the larva of Mytilus edulis. In: D. J. Crisp (ed.), Pro- ceedings of the 4th European Marine Biology Symposium, Cambridge University Press. pp. 259—280. Bourne, N., C. A. Hodgson & J. N. C. Whyte. 1989. A manual for scallop culture in British Columbia. Can. Tech. Rep. Fish. Aquat. Sci. #1694, 215 pp. Bourne, N. & C. A. Hodgson. In Press. Development of a viable nursery system for scallop culture. Proceedings of the 7th International Pec- tinid Workshop, World Aquaculture Society Special Publ. Caddy, J. F. 1969. Development of mantle organs, feeding and locomo- tion in postlarval Macoma balthica (L.) (Lamellibranchiata). Can. J. Zool. 47:609-617. Castanga, M. & W. Duggan. 1971. Rearing the bay scallop Aequipecten irradians. Proc. Nat. Shellfish. Assoc. 61:80—85. Chipman, W. A. & J. G. Hopkins. 1954. Water filtration by the bay scallop, Pecten irradians as observed with the use of radioactive algae. Biol. Bull. 107:80—91. Cranford, P. J. & J. Grant. 1990. Particle clearance and absorption of phytoplankton and detritus by the sea scallop Placopecten magel- lanicus. J. Exp. Mar. Biol. Ecol. In Press. Hodgson, C. A. & R. D. Burke. 1988. Development and larval mor- phology of the spiny scallop, Chlamys hastata. Biol. Bull. 174:303— 318. Ito, H. 1990. Scallop Fisheries and Aquaculture in Japan. Jn: Shumway, S. E. (ed.), Scallops: Biology, Ecology and Aquaculture. Elsevier Scientific, In Press. Jgrgensen, C. B. 1960. Efficiency of particle retention and rate of water transport in undisturbed lamellibranchs. J. Con. Int. Explor. Mer. 26:94—116. Kafuku, T. & H. Ikenoue. 1983. Modern methods of aquaculture in Japan. Vol. 11. Developments in Aquaculture and Fisheries Science. Elsevier Press, Amsterdam. King, J. 1987. Ontogeny of feeding in juvenile Panope abrupta. M.Sc. Thesis. University of Victoria, 227 pp. La Borgne, Y. 1981. Nursery culturing of post-larvae. Key to further de- velopment for bivalve mollusc hatcheries. pp. 141-149 In C. Claus, N De Pauw and E. Jaspers (eds) Nursery Culturing of Bivalve Mol- luscs. Proceedings of the International Workshop on Nursery Cul- turing of Bivalve Molluscs, Ghent, Belgium, 24—26 February, 1981. European Mariculture Society Special Publication #7. Loosanoff, V. L., H. C. Davies & P. Chanley. 1966. Dimensions and shapes of larvae of some marine bivalve molluscs. Malacologia 4:351-435. 144 FOIGHIL ET AL. Leighton, D. L. & C. F. Phleger. 1981. The suitability of the purple- hinge rock scallop to marine aquaculture. Calif. Sea Grant College Prog. Sea Grant Publ. Tech. T-CSGCP-001, 85 pp. Lesser, M. P., S. E. Shumway, T. Cucci, J. Barter & J. Edwards. 1990. Size-specific selection of phytoplankton in juvenile filter-feeding bi- valves: Comparison of the sea scallop Placopecten magellanicus with Mya arenaria and Mytilus edulis. In Press. In: Shumway, S. E. (ed) An International Compendium of Scallop Biology and Culture. World Aquaculture Society. Mikulich, L. V. & Y. A. Tsikhon-Lukania. 1981. Food of the scallop. Oceanol., 21:633—635. M@hlenberg, F. & H. U. Riisgard. 1978. Efficiency of particle retention in 13 species of suspension feeding bivalves. Ophelia 17:239—246. Owen, G. & J. M. McCrae. 1976. Further studies on latero-frontal tracts of bivalves. Proc. Roy. Soc. Ser. B 194:527—544. Palmer, R. E. & L. G. Williams. 1980. Effect of particle concentration on filtration effeciency of the bay scallop Argopecten irradians and the oyster Crassostrea virginica. Ophelia 19:163—174. Sastry, A. N. 1965. The development and external morphology of pelagic larvae and post-larval stages of the bay scallop Aequipecten irradians concentricus Say, reared in the laboratory. Bull. Mar. Sci. 15:417— 435. Shumway, S. E., R. Selvin & D. F. Schick. 1987. Food resources related to habitat in the scallop Placopecten magellanicus (Gmelin, 1791): a qualitative study. J. Shellfish Res., 6:89—95. Taguchi, K. 1978. A manual of scallop culture methodology and manage- ment. Fish. Mar. Serv. (Can.) Transl. Ser. 4198:146 pp. Tremblay, M. J. 1988. A summary of the proceedings of the Halifax sea scallop workshop, August 13—14th 1987. Can. Tech. Rep. Fish. Aquat. Sci., #1605, 12 pp. Vahl, O. 1972. Particle retention and relation between water transport and oxygen uptake in Chlamys opercularis (L.) (Bivalvia). Ophelia 10:67—74. Vahl, O. 1973. Efficiency of particle retention in Chlamys islandica (O. F. Miiller). Astarte 6:21—25. Ventilla, R. F. 1982. The scallop industry in Japan. Adv. Mar. Biol., 20:309—382. Whyte, J. N. C., N. Bourne & C. A. Hodgson, 1987. Assessment of bio- chemical composition and energy reserves in larvae of the scallop Pa- tinopecten yessoensis. J. Exp. Mar. Biol. Ecol., 113:113—124. Whyte, J. N. C., N. Bourne & C. A. Hodgson. 1989. Influence of algal diets on biochemical composition and energy levels in Patinopecten yessoensis larvae. Aquaculture 78:333—347. Journal of Shellfish Research, Vol. 9, No. 1, 145—148, 1990. TOLERANCES OF PERKINSUS SPP. (PROTOZOA, APICOMPLEXA) TO TEMPERATURE, CHLORINE AND SALINITY C. LOUISE GOGGIN, KIM B. SEWELL AND ROBERT J. G. LESTER Department of Parasitology, University of Queensland, St Lucia, Brisbane, Australia 4067. ABSTRACT The tolerances of the molluscan parasites Perkinsus spp. to temperature, chlorine and salinity were investigated to determine ways to eliminate the parasites from infected meats during processing. Trophozoites in the tissues of naturally infected blood cockles, Anadara trapezia, survived at least one day at 4°C and 0°C, and for 197 days at —60°C. In all cases, the parasite successfully developed after subsequent culture in thioglycollate medium and seawater. Free prezoosporangia from A. trapezia and Haliotis laevigata were killed within 30 minutes in 6 ppm chlorine while those enclosed in pieces of host tissue survived longer than 2 hours. Free prezoosporangia survived over 6 hr in 7%c seawater. All were killed within 6 hr in distilled water and within 10 minutes in 120%c NaCl solution at 50°C. KEY WORDS: Perkinsus, Apicomplexa, tolerance, abalone, parasite eradication INTRODUCTION Parasites of the genus Perkinsus (Mackin, Owen & Col- lier, 1950) are widespread in marine molluscs (Lauckner, 1983; Goggin & Lester, 1987) and have been associated with extensive mollusc mortalities both in culture and in the wild. Perkinsus marinus, the cause of “‘Dermo’’ disease, kills cultured oysters, Crassostrea virginica, on the east coast of the U.S.A. each summer (Mackin, 1962; Lauck- ner, 1983). Heavy infections of P. olseni appear to have killed the abalone, Haliotis laevigata, off the coast of South Australia (Lester, 1986); P. atlanticus has been as- sociated with mortality in Ruditapes decussatus from Por- tugal (Azevedo, 1989) and an unknown species of Per- kinsus has been implicated in the mortalities of giant clams, Tridacna gigas, on the Great Barrier Reef (Alder & Braley, 1988). Perkinsus marinus can survive in a range of environ- mental conditions. It is found in oysters in waters between 3.5%e and 50%o, can overwinter as subclinical infections in oysters at S5°C, and some prezoosporangia in “gapers’ can survive freezing (Ray, 1954; Andrews, 1965). Chu & Greene (1989) found prezoosporangia could withstand tem- peratures of 4°C for up to 4 days and a salinity of 6%c at 28°C. Zoospores tolerated 4 to 32%c at 20 to 28°C for up to 28 days. Our work on the tolerance of Australian Perkinsus spp. to temperature, salinity and chlorine was undertaken to see if there was a simple means of killing the parasite during processing. Abalone, Haliotis spp., infected with Per- kinsus olseni frequently show no clinical signs of disease, so infected tissue can easily reach processing plants. In South Australia, abalone are usually shucked (shell, gonad and viscera removed) at sea and the muscle and shell only returned to the processing plants. Any trophozoites of P. olseni present in the tissues are likely to swell and undergo 145 development prior to processing in the same manner as when infected tissues are cultured in thioglycollate me- dium. We therefore investigated the effects not only on tro- phozoites in tissue but also on free prezoosporangia. Few reports have been published on the tolerance of Perkinsus species in vitro to salinity and chlorine (Chu & Greene, 1989). This work supplements current knowledge on survival of Perkinsus spp. prezoosporangia from tissues cultured in fluid thioglycollate medium which were ex- posed to various salinity and chlorine levels. A range of salinities likely to be encountered during processing of meats was tested to determine what treatments could kill the parasites and thus be used to prevent their spread from plants processing infected shellfish. In Australia, abalone meats are washed in heated brine (120% NaCl, 50°C) be- fore canning, and in fresh water prior to freezing. The sur- vival of Perkinsus sp. in frozen meats was investigated, and as chlorine is routinely used as a Sterilizing agent (Herwig, 1979), the levels of chlorine needed to kill Per- kinsus were evaluated. MATERIALS AND METHODS Perkinsus parasites were obtained from two sources: naturally infected blood cockles, Anadara trapezia (De- shayes, 1840), from Moreton Bay, Queensland; and greenlip abalone, Haliotis laevigata Donovan, 1808, from South Australia, which had been experimentally infected with P. olseni from South Australia. We used Perkinsus from A. trapezia in most of the tests as we were unable to consistently obtain P. olseni and comparative data (Table 2) suggested their tolerances were similar. Temperature Tolerance Tissues from the foot, gill, digestive gland, mantle and muscle of infected A. trapezia were stored in 28 ml 146 GOGGIN ET AL. McCartney bottles moist (drained of seawater) and sub- merged in seawater at 4°C (cold room), 0°C (ice bath) and — 60°C (ultra cold freezer) for | day, and at — 60°C for up to 197 days. After storage, frozen tissues were thawed for | hour at room temperature (approximately 25°C) and cul- tured in fluid thioglycollate medium (Ray, 1966) at 27°C for 5 days. Infected tissue was removed to Petri dishes filled with seawater (35%c) at 27°C and teased apart to re- lease prezoosporangia. Prezoosporangia that adhered to the dish were allowed to develop for 3 days. To calculate the percent survival, the numbers of live and dead parasites were counted using a compound microscope at 35 x mag- nification in 30 fields of view or, if few parasites were present, over the whole dish. Perkinsus sp. were deemed to be alive only if motile zoospores were visible within spo- rangia. In Experiment 1, tissue from 5 A. trapezia was stored moist and in seawater at 4°C and 0°C for 24 hours. This was done twice under moist conditions and four times in seawater. Table | lists the combined totals from all repli- cates. In Experiment 2, tissue from | A. trapezia was drained and stored at — 60°C for 1, 3, 5, 7, 9, 11, 13, and 28 days. Tissue from a second A. trapezia was stored in seawater at — 60°C for 1, 5 and 197 days (Table 1). Chlorine and Salinity Tolerance To obtain prezoosporangia for use in the chlorine and salinity experiments, tissues from H. laevigata infected with P. olseni and A. trapezia infected with Perkinsus sp. were cultured in fluid thioglycollate medium at 27°C for at least seven days. Tissue was removed to Petri dishes at 27°C filled with seawater and left for 1 hour for the prezoo- sporangia to adhere to the dish. After treatment with chlorine or salinity, solutions were replaced with seawater and the dishes left for 42 hours, with one change of seawater, to allow surviving parasites to develop. The proportion of live and dead parasites was as- sessed as above, except that prezoosporangia that had un- dergone division were also considered viable. Chlorine Tolerance In preliminary trials, chlorine levels remained stable in the dark but dropped from 6 ppm to 1.5 ppm in 2 hours in dishes exposed to light. All experiments with chlorine, therefore, were done in the dark. Chlorine solutions were prepared from commercial grade sodium hypochlorite (125 g/l) and free chlorine levels monitored with an ‘R70’ chlo- rine test kit (AJAX Chemicals, Searle Australia Pty Ltd, Bankstown, Australia 2200). Chlorine solutions were tested before and after use to ensure concentrations were maintained throughout the experiments. Six dishes of prezoosporangia from A. trapezia were ex- posed to 6 ppm chlorine for either 2, | or 2 hours. Eight dishes of prezoosporangia from A. trapezia and H. laevi- gata were immersed in either 6, 12, 24, or 48 ppm chlorine for 2 hours in translucent, plastic containers, sealed with lids. Six dishes of prezoosporangia were immersed in 0 ppm chlorine for 2 hours as controls (Table 2, A). Two infected gill fragments of A. trapezia were exposed after thioglycollate culture to 6 ppm chlorine for 2 hours and 2 gill fragments were held in seawater as controls. After exposure to chlorine, Perkinsus sp. from gill frag- ments were allowed to adhere to the bottom of 6 Petri dishes. Tissue was removed and the plates rinsed with sea- water (Table 2, B). Salinity Tolerance Perkinsus sp. prezoosporangia from A. trapezia were isolated on Petri dishes. Twenty dishes in groups of four were exposed for 6 hours to one of: distilled water, 7%o, 14%c, 21%c or 35%c; four dishes were exposed to 120%o TABLE 1. The numbers of dead and live prezoosporangia of Perkinsus sp. recovered from infected tissues of Anadara trapezia held at 4°C, 0°C and — 60°C, either moist or in seawater, for various times. Moist Seawater Temp Expt (°C) Days Dead Live % Live Dead Live % Live 1 4°C 1 126 36 22 69 158 70 1 0°c 1 167 61 27 78 106 58 2 — 60°C 1 10 25 71 121 84 41 2 — 60°C 3 12 29 71 — — — 2 — 60°C 5 3 27 90 67 37 36 2 — 60°C 7 12 23 66 _ 7 7 2 — 60°C 9 27 40 60 — — — 2 — 60°C 11 21 51 71 - — — 2 — 60°C 13 55 0 0 — — -- 2 — 60°C 28 183 42 19 — a — 2 — 60°C 197 — _ -- 130 75 37 — no data available TOLERANCE OF PERKINSUS SPP. 147 TABLE 2. The numbers of dead and live prezoosporangia of Perkinsus sp. after exposure to different treatments of chlorine in seawater (35%c). Number and Percent of Prezoosporangia Chlorine Conc. Dead Live % Live A: Free prezoosporangia used a) Haliotis laevigata Control 1 (Oppm) 5 86 94 6ppm for 2 hours 365 0 0 12ppm for 2 hours 93 0 0 24ppm for 2 hours 117 0 0 48ppm for 2 hours 14 0 0 Control 2 (Oppm) 34 28 45 (b) Anadara trapezia Control 1 (Oppm) 28 566 95 6ppm for 2 hour 510 0 0 6ppm for | hour 470 0 0 6ppm for 2 hours 514 0 0 12ppm for 2 hours 307 0 0 24ppm for 2 hours 327 0 0 48ppm for 2 hours 9 0 0 Control 2 (Oppm) 55 446 89 B: Whole gill immersed Control (Oppm) 80 375 83 6ppm for 2 hours 268 570 68 NaCl at 50°C for either 2 minutes or 5 minutes, and 2 dishes were exposed to 120%o NaCl at 50°C for 10 minutes. Six dishes were exposed to 35%o seawater at 27°C as con- trols (Table 3). The dishes were submerged completely in a saline bath to ensure that all parasites were exposed to the hypersaline and high temperature conditions. RESULTS Temperature Tolerance Many trophozoites of Perkinsus sp. survived for 24 hours in drained tissue at 4°C (22% survival) and at 0°C (27%), and in tissue in seawater at 4°C (70%) and 0°C (58%). Many trophozoites survived for 28 days in drained tissue at — 60°C (19% survival) and for 197 days in tissue in seawater at — 60°C (37% survival) (Table 1). Chlorine Tolerance Prezoosporangia free of host tissues died within 30 minutes in chlorine solutions of 6 ppm and higher (Table II, A). Prezoosporangia surrounded by host tissue survived for at least 2 hours in 6 ppm chlorine (68% survival, Table 2, B). Salinity Tolerance Many prezoosporangia from A. trapezia survived for 2 minutes in a 120%o NaCl solution at 50°C (21%), a few survived for 5 minutes (2%) and all were dead after 10 minutes (Table 3). One prezoosporangium survived for 6 hours in distilled water and 17% survived for 6 hours in 7%o seawater. DISCUSSION Infected tissues of A. trapezia that had been chilled for 24 hours at 0°C or 4°C released live prezoosporangia. Aba- lone frequently are transported chilled prior to processing and thus it is likely that the parasite also can be transported in this way. Parasites within tissues of A. trapezia that were submerged in seawater survived —60°C for 197 days, de- veloped in subsequent thioglycollate culture and released motile zoospores. Andrews & Hewatt (1957) found tropho- zoites of Perkinsus marinus in oyster tissue, frozen for up to 8 days, swelled in thioglycollate media and possibly were still alive though further development was not fol- lowed. Chu & Greene (1989) observed that prezoospo- rangia of P. marinus survived at 4°C for up to 4 days but did not survive below 0°C for | day. Thus it appears that trophozoites within host tissue are more tolerant to freezing than prezoosporangia. Chilled or frozen meats infected with Perkinsus sp., therefore, could contain live parasites that might initiate infections when thawed tissues are dis- charged into the sea. Parasites surrounded by host tissue also were less sus- ceptible to chlorine than parasites free of tissue. Within tissue Perkinsus sp. survived for at least 2 hours in chlorine concentrations of 6 ppm (approximately 40 mg/l) though all free prezoosporangia (from cultured tissue) were killed within this time. Though it is possible that prezoosporangia are more susceptible to chlorine than trophozoites, it is more likely that the increased susceptibility of the parasite after culture is due simply to the lack of physical protection afforded by the tissues. We have initiated infections in naive hosts with zoospores from prezoosporangia cultured TABLE 3. The numbers of dead and live prezoosporangia of Perkinsus sp. from Anadara trapezia after exposure to various salinity regimes. Number and Percent of Prezoosporangia Regime Dead Live % Live 120%c NaCl, 50°C (2 mins) 342 88 21 120%c NaCl, 50°C (5 mins) 78 2 2 120% NaCl, 50°C (10 mins) 106 0 0 Control 1 (35%c seawater, 27°C) 219 427 66 Control 2 (35%e seawater, 27°C) 47 112 70 Distilled Water for 6 hours (27°C) 617 1 0 7%0c for 6 hours (27°C) 601 124 17 14%c for 6 hours (27°C) 342 688 67 21%0c for 6 hours (27°C) 240 877 79 Control 1 (35%c for 6 hours, 27°C) 226 969 81 Control 2 (35%e for 6 hours, 27°C) 154 1543 91 148 GOGGIN ET AL. in thioglycollate (Goggin, Sewell & Lester, 1989) which shows zoospores were not adversely effected by thioglycol- late culture, despite it being an artificial situation. Herwig (1979) recommended 50 mg/l of chlorine for sterilization of seawater, however, Bower (1989) found 25 mg/l] chlorine killed the protozoan parasite, Labyrinthuloides haliotidis, isolated from abalone within 20 minutes. Chlorine may not be a useful method of killing Perkinsus sp. because many parasites found in a processing plant would be enclosed in host tissue and therefore protected. In some processing plants, shucked abalone are washed in freshwater for 5 to 10 minutes in a mechanical ‘tumbler’ prior to freezing. We found that immersion in fresh water was an effective way of killing free prezoosporangia, most dying within minutes of exposure. However, washing meats for 5 to 10 minutes is unlikely to kill P. olseni as in our experiments, one prezoosporangium from A. trapezia, probably surrounded by host tissue, survived in distilled water for 2 hours. A single sporangium of P. marinus is capable of releasing approximately 354,700 zoospores (Chu & Greene, 1989) which clearly indicates a need to kill all parasites. Perkinsus sp. prezoosporangia from A. tra- pezia survived 6 hours in 7%c salinity which is similar to presporangia from P. marinus which sporulated in 6%c sa- linity at 28°C (Chu & Greene, 1989). Thus, low salinity cannot be relied upon to kill all parasites in a reasonable time. Prior to canning, shucked abalone frequently are washed in a tumbler with a 120%c salt solution at 50°C for approxi- mately 10 minutes. As we found that prezoosporangia of Perkinsus sp. from A. trapezia could not survive this treat- ment it is likely that Perkinsus sp. on the surface of abalone would be killed, although parasites deep in the tissues prob- ably survive. Abalone are canned at 115.6°C for 1 hour and this certainly would kill the parasite. Waste from some processing plants drains into the sea and is eaten by fish. We fed infected tissue of A. trapezia, cultured in thioglycollate, to toadfish, Tetractenos hamil- toni (Gray & Richardson, 1843), recovered faeces, and ob- served motile zoospores after development in seawater (un- published data). Hoese (1967) found that trophozoites of P. marinus survived passage through the intestine of fish. Therefore, fish could disperse Perkinsus spp. when fed in- fected tissues which drain from the tumbler. In conclusion, the prezoosporangia and trophozoites of Perkinsus spp. were resistant to a range of salinities and temperatures and, when enclosed in tissue, also were resis- tant to chlorine. Though many Perkinsus infections seem to have no detectable effect on their host population (Goggin & Lester, 1987), if it is considered necessary to prevent the spread of Perkinsus sp. from processing plants which operate as we have outlined, we recommend mollusc tissues are not returned to the sea. ACKNOWLEDGMENTS We thank Mr A. Hanson of Tas Seafoods, Smithton, Tasmania, Mr J. George of the Western Abalone Pro- cessors Cooperative and the staff of Australian Bight Fish- ermen Pty. Ltd. in Port Lincoln, South Australia, for their help and cooperation. The work was sponsored by Dover Fisheries, Dover, Tasmania, and supported in part by grants from the Fishing Industry Research Trust Account (FIRC 87/9) and the Australian Research Council (ARC 785). The senior author was supported by a Common- wealth Postgraduate Research Award. REFERENCES CITED Alder, J. & R. D. Braley. 1988. Mass mortalities of giant clams on the Great Barrier Reef. In: Copland, J. W. & Lucas, J. S. Giant Clams in Asia and the Pacific. ACIAR, Canberra. Andrews, J. D. & W. G. Hewatt. 1957. Oyster mortality studies in Vir- ginia Il The fungus disease caused by Dermocystidium marinum in oysters in Chesapeake Bay. Ecol. Monogr. 27:1—25. Azevedo, C. 1989. Fine structure of Perkinsus atlanticus (Apicomplexa, Perkinsea) parasite of the clam Ruditapes decussatus from Portugal. J. Parasitol. 75:627—635. Bower, S. 1989. Disinfectants and therapeutic agents for controlling La- byrinthuloides haliotidis (Protozoa: Labyrinthomorpha), an abalone pathogen. Aquaculture 78:207—215. Chu, F. E. & K. H. Greene. 1989. Effect of temperature and salinity on in vitro culture of the oyster pathogen Perkinsus marinus (Apicom- plexa: Perkinsea). J. Invert. Pathol. 53:260—268. Da Ros, L. & W. J. Canzonier. 1985. Perkinsus, a protistan threat to bivalve culture in the Mediterranean basin. Bull. Eur. Ass. Fish. Pathol. 5:23-25. Goggin, C. L. & R. J. G. Lester. 1987. Occurrence of Perkinsus species (Protozoa, Apicomplexa) in bivalves from the Great Barrier Reef. Dis aquat. Org. 3:113-117. Goggin, C. L., Sewell, K. B. & R. J. G. Lester. 1989. Cross infection experiments with Australian species of Perkinsus. Dis. aquat. Org. 7:55-59. Herwig, N. 1979. Handbook of Drugs and Chemicals used in the Treat- ment of Fish Diseases. Charles C. Thomas, Springfield, Illinois. Hoese, H. D. 1964. Studies on oyster scavengers and their relation to the fungus Dermocystidium marinum. Proc. Natl. Shellfish. Assoc. 53:161-173. Lauckner, G. 1983. Diseases of Mollusca: Bivalvia. In: Kinne, O. (ed) Diseases of Marine Animals Vol. 2. Biologische Anstalt Helgoland, Hamburg. Lester, R. J. G. 1986. Abalone die-back caused by protozoan infection? Australian Fisheries 45:26—27. Lester, R. J. G. & G. H. G. Davis 1981. A new Perkinsus species (Api- complexa, Perkinsea) from the abalone Haliotis ruber. J. Invert. Pathol. 37:181—187. Mackin, J. G. 1962. Oyster disease caused by Dermocystidium marinum and other microorganisms in Louisiana. Publ. Inst. Inst. Mar. Sci. Univ. Texas 7(1961):132—229. Ray, S. M. 1954. Biological studies of Dermocystidium marinum, a fungus parasite of oysters. Rice Institute Pamphlet, Special Issue 113 p. Ray, S. M. 1966. A review of the method for detecting Dermocystidium marinum with suggested modifications and precautions. Proc. Natl. Shellfish. Assoc. 54:55—69. Journal of Shellfish Research, Vol. 9, No. 1, 149-158, 1990. BIOLOGICAL AND ECONOMICAL ASSESSMENT OF AN OYSTER RESOURCE DEVELOPMENT PROJECT IN APALACHICOLA BAY, FLORIDA MARK E. BERRIGAN Florida Department of Natural Resources Division of Marine Resources Tallahassee, Florida 32399 ABSTRACT During 1986 and 1987, the Florida Department of Natural Resources restored 155.8 hectares (385 acres) of oyster reefs in Apalachicola Bay, Franklin County, Florida. Oyster reefs, including substrate and standing stocks were severely depleted during passage of Hurricane Elena in September 1985. Special management practices were applied and recovery processes were monitored on restored reefs. Field surveys were conducted on Bulkhead Bar, a 20.2 hectare (SO acre) restored reef. Standing stocks increased to 160 oysters/m? one year after setting and reached 578 oysters/m? after 18 months. Controlled harvesting was applied when standing stocks reached densities of 22.7 marketable oysters/m?; estimated yields ranged from 15,648 to 18,373 bags. Controlled harvesting accounted for 6,083 bags valued at $135,020 in seven days; added yields during the summer harvesting season were estimated at 11,734 bags. Combined yields were predicted to provide $395,515 in dockside revenues. Actual and estimated revenues from Bulk- head Bar demonstrated that restoration costs were recovered after the first harvesting season. Cost:benefit ratios ranged from 1:2.3 after two years to 1:20.7 after ten years. Economic analyses indicated that restoration efforts ultimately benefit levels of the industry from harvest to retail sales; added value benefits from Bulkhead Bar were predicted to reach $1,575,000 after two years. KEY WORDS: INTRODUCTION Extreme environmental and meteorological conditions associated with the passage of Hurricane Elena from 29 August through 2 September 1985 resulted in severe devas- tation of oyster reefs in Apalachicola Bay, Franklin County, Florida (Berrigan 1988). Apalachicola Bay con- tains Florida’s most commercially productive reefs and was the source of 92% of the oysters, Crassostrea virginica (Gmelin 1791), landed in the State in 1984. Reported landings for Franklin County in 1984 were 6.2 million pounds (2,820,000 kg) of oyster meats valued in excess of $6.8 million. From January through August 1985, 3.8 mil- lion pounds (1,720,000 kg) were landed and less than 0.5 million pounds (230,000 kg) were landed in 1986. De- clining landings were a direct result of resource losses as- sociated with Hurricane Elena (Berrigan 1989). Following Hurricane Elena, assessments of oyster re- sources throughout the Apalachicola Bay system identified oyster reefs located in western St. George Sound and eastern Apalachicola Bay as the most severely damaged (Berrigan 1988). Reefs classified as severely impacted were historically productive reefs which sustained exten- sive losses of live oysters and cultch or were subjected to extensive sedimentation. Progressive recovery of severely impacted reefs was expected to be slow and largely depen- dent upon reconstructive efforts. In response to debilitating losses of oyster resources, the Florida Department of Natural Resources and the Florida Marine Fisheries Commission implemented comprehensive management programs and regulatory restrictions to foster resource recovery and facilitate restoration. An important component of the recovery plan was the large scale restora- tion of severely damaged reefs. Oyster resource restoration, oysters, Apalachicola Bay, resource management, population dynamics, economics based on the concept of providing accessible and suitable substrate for oyster larvae to attach, has long been an ac- cepted practice of fisheries managers and shellfish lease- holders along the Gulf Coast (May 1971, Whitfield 1973, Dugas 1977, MacKenzie 1977, and Hoffstetter 1981). Al- though, the Florida Department of Natural Resources has managed a program to construct and rehabilitate oyster reefs for 40 years (Whitfield and Beaumariage 1977, Futch 1983), the extent of damage resulting from Hurri- cane Elena necessitated expanding the scope of restoration efforts (Berrigan 1988). Expanded resource restoration efforts were implemented using emergency assistance funding through the Commer- cial Fisheries Research and Development Act, Public Law 88-309(4B). In 1985, $1,570,000 were released from con- gressional appropriations to restore damaged reefs in Apa- lachicola Bay. In 1986, $918,000 were used to restore se- verely damaged reefs and $553,960 were released to com- plete the project in 1987. Approximately 385 acres (155.8 ha) were restored using 96,230 yd? (73,578 m3) of clam shell. Reefs restored during this project were protected as Spe- cial Resource Recovery Areas and special management was applied while recovery progressed. Restored reefs were monitored throughout the recovery process, and emergency measures to regulate oyster harvesting provided a unique opportunity to focus on the value of fisheries information. Stock assessments, yield estimates, harvesting pressure, and landing statistics were used to determine the success of resource rehabilitation programs. Damaged reefs were ex- pected to produce marketable oysters as early as two years after restoration, demonstrating the cost effectiveness of oyster resource development programs. 149 150 BERRIGAN A special harvesting season on Bulkhead Bar was im- plemented in May 1989 based on the availability of stocks sufficient to support limited harvesting. Controlled har- vesting during the special season was designed to provide a mechanism to evaluate fisheries management practices and reduce harvesting pressure on reefs in the winter harvesting area, while providing timely economic benefits to the de- pressed oyster industry. MATERIALS AND METHODS Bulkhead Bar, a restored reef presented the most advan- tageous experimental protocol to evaluate reef restoration and shellfish management practices. Successful spatfall during September 1987 was verified in May 1988, subse- quent sampling indicated growth among juvenile oysters, and in April 1989 field surveys demonstrated that market- able oyster stocks were sufficient to support commercial harvesting activity. Additionally, the reef’s location within the summer harvesting area provided protection from har- vesting throughout the recovery process. Initial harvesting was conducted by commercial harvesters under controlled conditions to maintain the ‘“‘natural laboratory’’ experi- mental protocol. Controlled harvesting efforts applied to a special harvesting season provided the mechanism to 1) evaluate stock assessment techniques, 2) validate popula- tion estimates as a predictive index for potential landings, 3) monitor oyster population dynamics in response to har- vesting pressure, and 4) accurately project cost:benefit ratios for resource development projects. Description of Study Area Bulkhead Bar is an oyster reef located in eastern Apa- lachicola Bay, approximately 10 km southeast of the mouth of the Apalachicola River. Historically, Bulkhead Bar was part of a large consolidated shoal orientated perpendicular to the mainland and St. George Island forming a geo- graphic demarcation between Apalachicola Bay to the west and St. George Sound to the east. More recently, construc- tion of the Bryant Patton Bridge and Causeway along the long axis of the shoal and dredging the Gulf Intracoastal Waterway across the shoal, have obscured the geographical distinction of the reef (Fig. 1). Extreme hydrologic activity associated with Hurricane Elena reduced the southwestern portion of Bulkhead Bar to a relatively barren and compacted shoal. The surface of the shoal was covered by sand and shell rubble; live oysters and shell suitable for supporting oysters were conspicu- ously absent. Affected areas exhibited moderate elevation surrounded by areas of soft sediments; water depths ranged from | to 2 meters (MLW). Shellfish growing waters in the Apalachicola Bay system are classified as Conditionally Approved based on bacteriological water quality and seasonality. Oyster har- vesting in Conditionally Approved shellfish growing waters is regulated on a seasonal basis, including summer and winter harvesting seasons. Harvesting during the summer season is permitted from July 1 through September 30; the winter harvesting season extends from October | through June 30. Bulkhead Bar is located in the summer harvesting area in Apalachicola Bay near the boundary between the winter harvesting area in St. George Sound. Bacteriological water quality at Bulkhead Bar satisfies criteria for Condi- tionally Approved shellfish growing waters during the summer months, but does not satisfy criteria necessary for Conditionally Approved shellfish growing waters during all the winter months. However, under favorable meteorolog- ical and hydrographic conditions, water quality satisfies criteria for harvesting during the warmer and drier months of the winter harvesting season. Resource Restoration Methods for applying cultch to improve substrate char- acteristics have become common practice for increasing oyster productivity by promoting larval attachment and sur- vival. Cultch planted in areas where natural reproduction occurs stimulates larval setting and establishment of new oyster populations. Reef restoration activities were initiated in May to take advantage of potential summer through fall spatfall peaks. Earlier research had shown that spawning in Apalachicola Bay is seasonally regular and of long dura- tion, but that setting intensity is variable (Ingle and Dawson 1953). More recent investigations indicated that successful spatfall occurred in eastern portions of the bay during late summer and in the fall (Berrigan 1989). Oyster reef restoration was divided into two phases; Phase I was completed in May 1986 and Phase II was com- pleted in June 1987 (Table 1). During Phase I, 225 acres (91.1 hectares) of oyster reefs were restored using 56,470 yd? (43,177 m3) of clam shell; including acreage on Hotel Bar in Apalachicola Bay and Peanut Ridge Bar in St. in St. George Sound. During Phase II, 160 acres (64.7 hectares) were restored using 39,760 yd? (30,401 m3) of clam shell; including acreage on Cat Point and Bulkhead Bar in Apalachicola Bay and Peanut Patch Bar in St. George Sound (Figure 1). Restoration of Bulkhead Bar was completed on 3 June 1987 and required 12,500 yd? (9,558 m2?) of shell at a cost of $174,265. The amount of reef area restored was based on application rates of approximately 250 yd? of cultch/acre (472 m?/ha). Clam shells, Rangia spp., dredged from Lake Pontchar- train, Louisiana, were transported by barge to planting sites where they were transferred to shallow draft barges. Cultch was washed overboard using high pressure water cannons and pumps aboard a second barge. Both units were posi- tioned and moved across reefs by a tug boat. Perimeter boundaries were marked before restoration efforts were ini- tiated. Initially, boundary markers were positioned and reef areas calculated using LORAN C. Bulkhead Bar was resur- veyed after construction and the delineated area was calcu- lated using coordinate geometry to 49.9 acres (20.2 ha). ASSESSMENT OF AN OYSTER RESOURCE DEVELOPMENT PROJECT 151 APALACHICOLA BAY ‘S MIKES CUT NORNAN'S @ LUMPS PLATFORM BAR PEANUT op LIGHTHOUSE BAR BULKHEAD BAR = Figure 1. Location of resource restoration sites in Apalachicola Bay, Florida from 1986 through 1987. Approximately 45 acres (18.2 ha) of the delineated 50 acres (20.2 ha) of Bulkhead Bar were improved during the cultch planting project. Maneuvering barges under various hydrographic and meteorological conditions made uniform application difficult, especially along peripheral areas. Es- timates of restored area based on effective cultch dispersal alone, may represent overestimates of improved substrate. Field surveys conducted on transects across Bulkhead Bar on 21 April 1989 demonstrated density differences across the delineated reef area. Differences between quadrats and transects reflected levels of substrate improvement re- sulting from varying cultch application rates and indicated that restoration was not uniform over the entire reef. The abundance of cultch and oysters was used to define reef areas as fully restored, improved, and unimproved. Esti- mates of fully restored substrate included 25 acres (10.1 TABLE 1. Oyster reefs in Apalachicola Bay and St. George Sound restored in May 1986 and June 1987 Cultch Area ———————— Cost Reef Phase Date (m?) (yd?) (ha) (ac) ($) Platform Bar I 5/86 12,593 16,470 26.3 65 267,638 East Hole Bar I 5/86 12,425 16,250 24.3 60 264,262 Hotel Bar I 5/68 18,159 23,750 40.5 100 386,097 Phase I Totals 43,177 56,470 91.1 225 917,997 Cat Point II 6/87 13,197 17,260 28.3 70 240,301 Bulkhead Bar I 6/87 9,558 12,500 20.2 50 174,265 Peanut Ridge Il 6/87 7,646 10,000 16.2 _40 139,380 Phase II Totals 30,401 39,760 64.7 160 $53,946 Project Totals 73,578 96,230 155.8 385 1,471,943 152 BERRIGAN hectares) in the center of the reef, improved areas included 20 acres (8.1 hectares) in marginal areas, and about 5 acres (2.0 hectares) in peripheral areas were not improved (Table 2). Subsequent field surveys were confined to transects in areas where substrate had been improved and restored. Resource Assessment Field surveys were conducted on Bulkhead Bar using sampling protocol and statistical analyses designed to vali- date sampling and analytical techniques used in a compre- hensive oyster resource assessment program in Apalachi- cola Bay which began in 1982 (Berrigan 1989). Oyster populations were compared using the Kruskal-Wallis test to determine significant differences between length-frequency distributions among replicate samples and between succes- sive sampling intervals (Statistical Analysis System 1985). Sampling on Bulkhead Bar was conducted on 29 Sep- tember 1988, 21 April 1989, 23 May 1989, and 1 No- vember 1989. These dates were approximately 16 and 23 months after restoration, following controlled harvesting, and following the summer harvesting season, respectively. Transects were established across the restored reef and quadrats were selected along transects by tossing a PVC grid from the survey vessel. The number of samples ranged from 5 to 20 quadrats per sampling period (Table 2). Tran- sects and quadrats were considered unbiased, since the reef was subtidal and oyster distributions and densities could not be determined from surface observations. A weighted 0.25 m* PVC grid was used to delineate sample quadrats. Samples were collected by divers; live oysters, shell, and associated fauna were removed to a depth of 15 cm, placed in mesh collecting bags, and deliv- ered to the survey vessel. Live oysters were measured to the nearest lower 0.5 cm length (longest dimension). All live oysters were measured from samples collected on 29 September 1988, 23 May 1989, and 1 November 1989. During the 21 April 1989 sampling period, only oysters greater than 75 mm (3 in) were measured from all 20 quadrats; all live oysters were measured from 10 quadrats. Substrate characteristics, competitors, and freshly dead oysters (boxes) were noted. Standing stocks were estimated from oysters collected from 0.25 m? quadrats. Length-frequency distributions were developed for each sampling period (Figure 2). Oysters equal to or greater than 25 mm (1 in) in length were used in population estimates. Oysters between 50 mm to 70 mm in length were used to predict growth rates, mortality rates, and recruitment into marketable size classes (greater than or equal to 75 mm). Legal-sized oysters equal to or greater than 75 mm in length provided estimates of market- able oysters/m? and densities were extrapolated to calculate potential production levels. Estimated yield was defined as bags/acre, where the capacity of a 60 lb bag (27.2 kg) was 225 oysters (Berrigan 1989). Production estimates used in data analyses were extrapolated for 45 (18.2 ha) and 50 (20.2 ha) acres. Estimated standing stocks and yields were TABLE 2. Population parameters for oysters sampled during field surveys of Bulkhead Bar. No. Range Transect Oysters Mean Stand. Min. Max. No. (0.25 m?) (mm) Dev. (mm) September 1988 1 31 41.9 11.4 25 65 1 66 47.7 8.6 30 80 1 31 47.4 9.7 30 70 1 27 46.1 9.8 30 70 1 46 45.9 9.6 30 65 1 (Total) 201 46.1 9.7 25 80 April 1989 1 278 36.6 13.6 25 135 1 74 40.8 19.3 25 95 1 132 35.6 14.6 25 90 1 120 46.1 19.1 25 90 1 161 8255 10.5 25 75 2 71 B32 13.2 25 80 2 83 33.8 14.7 25 80 2 295 32.2 11.8 25 110 2 86 35.3 15.9 25 90 2 142 40.9 21.8 25 110 2 (Total) 1,142 36.2 15.6 25 135 May 1989 1 63 30.5 11.0 25 70 1 42 30.2 11.2 25 85 1 96 33.0 14.2 25 70 1 99 32.8 13.8 25 90 1 150 30.9 9.5 25 80 2 127 36.9 15.5 25 95 2 90 3773, 15.6 25 95 2 74 36.6 17.5 25 90 2 129 33.6 12.9 25 70 2 146 35.3 15.6 25 110 3 223 33.3 15.6 25 100 3 255 34.1 12.8 25 80 3 116 34.3 16.1 25 95 3 84 30.7 14.2 25 95 3 43 34.3 13.4 25 70 4 21 45.0 13.3 25 80 4 42 34.4 10.7 25 60 4 21 30.2 11.1 25 65 4 109 35.0 14.8 25 95 4 89 35.6 14.1 25 90 4 (Total) 2,019 34.0 14.2 25 110 November 1989 1 70 40.4 16.0 25 85 1 58 32.8 10.0 25 70 1 33 33.5 8.4 25 60 1 31 39.8 14.7 25 75 1 120 39.0 11.6 25 80 2 44 34.3 13.8 25 80 2 47 33.3 12.3 25 80 2 40 37.8 9.5 25 65 2 55 36.6 12.7 25 75 2 51 35:3 11.6 25 70 3 18 32.2 8.4 25 60 3 88 34.0 9.9 25 70 3 74 39.3 11.8 25 70 3 25 32.4 9.7 25 60 3 40 39.9 11.7 25 75 3 (Total) 794 36.6 12.1 25 85 ASSESSMENT OF AN OYSTER RESOURCE DEVELOPMENT PROJECT 153 expressed as oysters per acre and bags per acre, conforming to common terminology. Natural Mortality Estimates of population losses were important in pre- dicting yields based on standing stock assessments. Mor- tality rates for the populations surveyed were expressed as: Crea GF ye = where Y, and Y, are densities in oysters/m? and t, and t, are the corresponding times (weeks) at which the densities were observed (Tyler and Gallucci 1980). Thus, if Y,, the time period, and r, are known, densities (Y,) can be pre- dicted. Estimated mortality rates were calculated using di- rect correlations between predicted densities over selected time intervals. These calculations do not account for nu- merous factors affecting natural mortality, but provide an estimate of mortality among similar populations. Harvesting Field surveys on 21 April 1989 indicated that standing oyster stocks on Bulkhead Bar were sufficient to support commercial harvesting. Concurrently, water quality data and bacteriological analyses demonstrated that water quality parameters satisfied Conditionally Approved shell- fish growing water criteria and did not threaten public health. In accordance with the management plan to foster resource and economic recovery of Apalachicola Bay’s oyster industry, the Department of Natural Resources rec- ommended a special harvesting season. Oyster harvesting was permitted on Bulkhead Bar from 8 May through 18 May 1989. Oystermen were asked to participate in controlled harvesting under current manage- ment provisions for commercial harvesting in Apalachicola Bay. Harvesting from the specified reef was permitted from sunrise until 4 p.m. on Monday through Thursday, and all oysters harvested were reported and tagged at the on-site check station before they were delivered to certified shell- fish dealers. Check stations, established as part of the re- covery management plan following Hurricane Elena, moni- tored daily harvests, number of vessels engaged in har- vesting, catch per vessel, and number of oysters per bag. Provisions for commercial harvesting also included size limits of 3 inches (75 mm), daily bag limits of fifteen 60-lb bags (408 kg), and gear was restricted to hand tongs only. Before oysters were tagged at the check station, selected bags were routinely sampled to determine whether they met legal size requirements. Additionally, oysters from selected bags were counted to determine the number of oysters/bag. Counts during routine inspections ranged from 255 to 423 oysters/bag and averaged 360 oysters/bag. Since yield esti- mates were based on bags containing 225 oysters, the ac- tual number of ‘‘tagged bags’’ was converted to “‘adjusted bags’’ also containing 225 oysters. Controlled harvesting was terminated at 4:00 pm on 18 BULKHEAD OYSTER SURVEY SEPTEMBER 1988 70 60 APRIL 1989 MAY 1989 NOVEMBER 1989 OYSTER LENGTH (mm) Figure 2. Length-frequency histograms for oyster populations on Bulkhead Bar between September 1988 and November 1989. May 1989 due to declining landings and reduced harvesting effort. Harvesting from Bulkhead Bar was prohibited until 3 July 1989, when harvesting was permitted under condi- tions of the summer harvesting season. During the summer harvesting season, harvesting was permitted Monday through Thursday from July 1 through September 30. Ex- clusive monitoring of Bulkhead Bar, including harvesting effort and landings was not part of the monitoring program during the summer harvesting season. RESULTS Successful spatfall occurred on Bulkhead Bar four times during the survey period (June 1987 through October 1989). Initially, spatfall occurred in fall 1987, followed by intensive spatfall in fall 1988 and moderate spatfall events in spring and fall 1989. Field surveys indicated standing 154 BERRIGAN stocks increased from essentially zero to 160 oysters/m? in one year. Standing stocks increased to 578 oysters/m? in April 1989 and declined to 212 oysters/m? in November 1989 (Figure 2). Field surveys of Bulkhead Bar on 21 April 1989 indi- cated that standing stocks of marketable size oysters (3 inches) exceeded 17.4 oysters/m?; standing stocks of mar- ketable size oysters from more productive reef areas ranged from 22.7 to 25.6 oysters/m?. Standing stocks, defined as marketable oysters/m*, were converted to estimated yields, defined as bags/acre. Estimated yields ranged from 313 bags/acre over the entire reef to 460 bags/acre over the most productive reef areas (Table 3). Standing stocks of marketable oysters were calculated by extrapolating the number of oysters equal to or greater than 75 mm/0.25 m?. Estimated yields and economic ben- efits were calculated using the following assumptions; a) delineated area of Bulkhead Bar was 50 acres (20.25 hectares), b) improved reef area was estimated at 45 acres (18.23 hectares), c) a 60-lb bag contains 225 oysters, and d) shellstock was valued at $0.37/lb ($22.20/bag). Differ- ences in standing stocks and estimated yields reflected areas where substrate differences were demonstrated. Ac- cordingly, densities of 17.4 oysters/m* were extrapolated to 3,520,900 marketable oysters or 15,648 bags for 50 acres. Similarly, densities of 22.7 oysters, representing only the improved substrate, were extrapolated to 4,134,015 mar- ketable oysters or 18,373 bags for 45 acres. Standing TABLE 3. Population estimates for oysters on a restored reef, Bulkhead Bar, and two undamaged natural reefs, Lighthouse Bar and Norman’s Lumps, before and after the 1989 summer harvesting season. Field Surveys Oysters 1 ———— Bas Date Quad. /m? >50mm/m? >75mm/m? >75mm/ac /acre Bulkhead Bar 9/29/88 5* 161 73 0.8 3,561 16 4/21/89 NO Sy 104 25.6 103,603 460 1525 NS NS 22h 91,867 408 20° NS NS 17.4 70,418 313 5/23/89 20 404 61 10.6 42,898 191 11/1/894 15°) 212 29 3h) 14,165 63 Lighthouse Bar 6/13/89 5 300 127 16.8 67,990 302 10/31/89 5 163 83 6.4 25,901 115 Norman’s Lumps 6/13/89 10 157 82 18.4 74,465 331 10/31/894 10 146 42 5-2 21,044 94 2 includes quadrats from restored areas (25 acres); > includes quadrats from improved areas (20 acres); © includes quadrats from unimproved areas (5 acres); 4 samples collected after 1989 summer harvesting season; NS indicates that only oysters >75 mm were measured. stocks represented potential dockside revenues ranging from $347,385 to $407,880, if marketable stocks could be completely harvested. The special harvesting season in May 1989 marked the first time that the Bulkhead Bar Resource Recovery Area had been commercially harvested since it was restored. Fisheries statistics collected at the on-site check station in- dicated that 3,802 bags of oysters were harvested by 561 vessel-trips during seven harvesting days. Actual bags counts (3,802 tagged bags) were recalculated to provide an adjusted value (adjusted count) which more accurately rep- resented total landings. Bags measured at check stations contained an average of 360 oysters; adjusted landings were converted to 1,368,720 oysters, or 6,082 bags when adjusted to 225 oysters/bag (Table 4). Dockside value of oysters landed during the controlled harvesting period was estimated at $135,020 (6,082 bags @ $22.20/60-lb bag). Daily landings ranged from 1,944 bags (tagged) on the first day to 33 bags (tagged) on the final day. Similarly, the number of vessels engaged in harvesting declined from 208 to 10 vessels during the same period. Concomitant effi- ciency declined from 9.3 to 3.3 bags/vessel/day. Initial plans were to continue controlled harvesting throughout May, discontinue harvesting during June, and continue harvesting during the summer harvesting season (July through September). However, concentrated harvesting pressure rapidly depleted standing stocks and substantially shortened the planned harvesting period. Harvesting during the first two days accounted for 77.5% of the total landings. The marked decrease in harvesting effort after the second day indicated that stocks were rapidly exploited to a level where harvesting efficiency was no longer more ad- vantageous on Bulkhead Bar than on reefs in the winter harvesting area. Continued declines in harvesting effort and landings prompted the termination of the special season after seven harvesting days. Field surveys following con- trolled harvesting indicated that yield estimates had been reduced to a level where harvesting effort was predicted to slow. Field surveys of Bulkhead Bar on 23 May 1989 fol- lowing the controlled harvesting period indicated that standing stocks of marketable size oysters had been re- duced to densities of 10.6 oysters/m?; estimated yield was reduced to 191 bags/acre (Table 3). Potential production from the remaining population extrapolated over 45 acres was 1,930,410 marketable oysters, indicating a reduction of 2,203,605 oysters between sampling periods. Landings during controlled harvesting accounted for 1,368,720 oysters, or 62% of the estimated reduction in standing stocks, assuming that standing stocks and harvesting effort were concentrated on improved areas only. Following the summer harvesting season, field surveys on | November 1989 indicated that densities were reduced to 3.5 marketable oysters/m?, or 63 bags/acre on Bulkhead Bar. Potential yields from 45 acres, based on projected ASSESSMENT OF AN OYSTER RESOURCE DEVELOPMENT PROJECT 155 TABLE 4. Tagged and adjusted oyster landings from Bulkhead Bar reported to monitoring stations between 8 May and 18 May 1989. Tagged Adjusted Tagged Adjusted Bags/ Bags/ Date Bags Bags Vessels Vessel Vessel May 8 1,944 3,110 208 93 15.0 May 9 1,001 1,602 188 5:3 8.8 May 11 338 540 58 5.8 9.3 May 15 293 469 56 5.2 8.4 May 16 104 166 21 5.0 7.9 May 17 89 142 20 4.5 7.1 May 18 33 53 10 3,3} 522 Total 3,802 6,082 561 6.8 10.8 growth and mortality rates, ranged from 11,734 bags to 21,447 bags with estimated values of $260,495 and $476,123, respectively. However, because landings from individual reefs were not recorded at check stations during the summer harvesting season, yields and values for Bulk- head Bar were not determined. Check stations reported 16,001 bags landed in July, 9,947 bags landed in August, and 3,871 bags landed in September, totaling 29,819 (tagged) bags for the summer harvesting season from all reefs in the summer harvesting area. The number of bags tagged was converted to 47,710 bags landed. Field surveys of oyster reefs in the summer harvesting area, including Lighthouse Bar and Norman’s Lumps, indi- cated that standing stocks were reduced to levels of approx- imately 100 bags/acre after the summer harvesting season (Table 3). Standing stock estimates suggested that har- vesting pressure may have been concentrated on Bulkhead Bar at the end of the summer harvesting season when other reefs were also depleted. Estimated yields were reduced to 63 bags/acre on Bulkhead Bar compared to 115 bags/acre on Lighthouse Bar and 94 bags/acre on Norman’s Lumps. DISCUSSION Production Ingle and Whitfield (1968) and Whitfield (1973) esti- mated that about 400 bu/acre could be harvested from pro- ductive artificially constructed reefs within two years of planting cultch. During field surveys of natural and con- structed oyster reefs in Apalachicola Bay, Berrigan (1989) developed a scale using defined sampling protocol to deter- mine the relative condition of oyster resources based on production estimates. Estimated production exceeding 400 bags/acre was applied as an indicator of healthy oyster reefs capable of sustaining commercial harvesting. Accordingly, oyster populations were 1) capable of supporting limited commercial harvesting when stocks exceeded 200 bags/ acre, 2) below levels necessary to support commercial har- vesting when stocks fell below 200 bags/acre, and 3) con- sidered depleted when marketable stocks were below 100 bags/acre. Estimated yields for Bulkhead Bar ranged from 313 bags/acre for the delineated 50 acres to 408 bags/acre for the improved acreage (45 acres) by 21 April 1989. Esti- mated yields reached 460 bags/acre in 18.5 months (80 weeks) on highly productive reef areas (25 acres). Esti- mated production from Bulkhead Bar ranged from 15,648 to 18,373 bags. Two critical factors influencing estimated yields and landings were identified during data analyses. First, esti- mated yield (bags available for harvest) was a function of the area where substrate was improved and level of im- provement. Over estimates of the productive area may have contributed to disparity between estimated stock reductions and stock reductions accounted for in landings. Secondly, harvesting success (bags landed) was a function of the number of oysters contained in each bag landed. To com- pensate for areal dependent production estimates (acres) and numerical dependent landings (bags), both yield esti- mates and landings were recalculated using adjusted values (45 acres and 225 oysters/bag) to evaluate resource assess- ment techniques. Adjusted values provided a standard unit to compare yield estimates with harvesting success and landings. Popu- lation estimates used in data analyses were extrapolated for 45 acres to represent productive acreage and the area where harvesting effort was concentrated. Standing stocks and es- timated yields were adjusted to represent fully restored, improved, and unimproved reef areas (Table 3). Yield esti- mates ranged from 313 to 460 bags/ac; lowest levels re- flected standing stocks over the entire 50 acres while highest levels were confined to the most productive 25 acres. Sampling was subsequently confined to improved areas to reduce variations in population estimates. The number of oysters in each bag harvested was also adjusted to compensate for oystermen overfilling their bags. It is common practice among oystermen to overfill bags since neither volumetric measure (ten gallons) nor weight (60-Ib/bag) are strictly monitored when there is no obvious intent to circumvent daily bag limits. However, adjusted landings suggested that some harvesters exceeded 15 bag limits (900 Ibs) on the first day of controlled har- vesting (Table 4). To evaluate yield estimates made from standing stock assessments on Bulkhead Bar, declines in predicted yields were compared to adjusted landings. Predicted yield was 18,373 bags for 45 acres before harvesting was initiated. Following controlled harvesting, remaining yields of mar- ketable oysters were estimated at 8,580 bags indicating a decrease of 9,793 bags. Landings accounted for 6,082 bags (adjusted) or 62% of the estimated yield. Declines in esti- mated yields from 408 bags/acre to 191 bags/acre indicated reductions of 217 bags/acre; landings accounted for 135 bags/acre. Natural mortality was calculated to account for 156 BERRIGAN approximately 15% of population losses between sampling intervals. When losses to natural mortality prior to har- vesting were calculated, landings accounted for 86% of population reductions. Population Structure Analyses of oyster populations on Bulkhead Bar were aided by the fact that the date when the population under surveillance was established could be determined. Intensive spatfall occurred during a single event of relatively short duration making population analyses relatively straightfor- ward compared to analyzing populations in which recruit- ment occurs over an extended period. When intense spatfall occurs in the spring, continuous low intensity spatfall throughout the summer followed by peaks in the fall tend to obfuscate population trends. In this instance, a single event occurred following restoration of the substrate. Rapid growth during the fall and winter further distinguished this cohort from oysters recruited during the following spring. Intensive spatfall did not occur until the fall of 1988 and again during the spring of 1989 (Figure 2). Population parameters at Bulkhead Bar, including re- cruitment, standing stocks, growth rates, mortality rates, and harvesting pressure, were developed from field surveys for the population established in September 1987. Analyses of these parameters provided data for developing assump- tions to provide population dynamics for subsequently re- cruited stocks. Estimated growth rates and mortality rates were particularly important to predicting potential yields based on standing stock assessments. Estimates of recruit- ment to marketable stocks were critical to developing long term cost:benefit ratios when landings from Bulkhead Bar were not exclusively monitored. Juvenile oysters observed on Bulkhead Bar in May 1988 indicated that successful spatfall had occurred during the fall of 1987. Length-frequency distributions developed from field surveys in September 1988 confirmed that juve- nile stocks were the result of setting in September or Oc- tober 1987. Median lengths (44 mm) of oysters sampled 29 September 1988 indicated growth rates of 0.85 mm/wk, as- suming that spatfall occurred during the same week in Sep- tember 1987 (0.85 mm X 52 weeks = 44.2 mm). Approx- imately 25% of the sample population was greater than 50 mm in length, suggesting that growth rates among rapidly growing oysters may have exceeded 0.96 mm/wk. Growth rates were expressed as the mean length (mm) increase over time (52 weeks), and did not account for variability among individuals, size dependent growth, density dependence, and environmental factors. Growth rates of 0.9 mm/wk (Ingle and Dawson 1952) and 0.85 mm/wk (Berrigan 1988) have been reported for Apalachicola Bay. Assuming growth rates of 0.96 mm/wk, fastest growing oysters in the population established in September 1987 on Bulkhead Bar should have reached marketable size by the end of April 1989 (80 weeks post set). On 29 September 1988 (52 weeks post set), oysters greater than or equal to 50 mm numbered 72 oysters/m?. Assuming growth rates of 0.90 mm/wk and no losses to mortality, standing stocks should have reached 73 marketable oysters/m? after an ad- ditional 28 weeks. However, surveys in April 1989 from the same transects indicated standing stocks of 25.6 mar- ketable oysters/m* (Table 3). Declines in standing stocks from densities of 73 to 25.6 marketable oysters/m? sug- gested that losses to natural mortality had been substantial or growth rates were lower than predicted. Significantly slower growth rates between the sampling periods were discounted since growth rates were expected to be highest during periods when water temperatures are cooler (Ber- rigan 1989, Ingle and Dawson 1952). Therefore, natural mortality was considered the most probable causative factor reducing standing stocks of adult and subadult oysters within the sample population. Differences in standing stocks between sample intervals were compared to determine the effects of natural mortality on predicted yields. A weekly mortality rate, expressed as the average number of oysters/m? lost each week during the 28 week sampling interval, was used to represent reduc- tions in extant populations. For example, during the first week of the sampling interval (week 52), 1.7 oysters/m? were lost from the initial population of 73 oysters/m?; by the last week of the sampling interval (week 80) 1.7 oysters/m? were lost from the surviving population of 27.3 oysters/m?. Natural mortality, expressed as percent reduc- tion, ranged from 2.3% to 6.2% per week, respectively. Population losses attributed to natural mortality accounted for a 65% reduction during 28 weeks, or a mean reduction of 3.34% per week. Extending intervals to 30 weeks, ac- counting for additional losses during the two weeks be- tween the previous sampling period and when harvesting was initiated, increased cumulative population losses to 70%. Similarly, growth and mortality rates were included in calculations based on length-frequency distributions of oysters between 60 and 75 mm to predict yields during the next summer harvesting season. Standing stocks of market- able oysters during the summer harvesting season were es- timated from 1) marketable stocks remaining after the spe- cial harvesting season (9 oysters/m?), 2) recruitment of sublegal-sized oysters to marketable stocks (21 oysters/m?), and 3) population losses due to natural mortality. Com- bined standing stocks, including recruitment and mortality, were expected to produce approximately 18 marketable oysters/m? at the beginning of the summer harvesting season. Throughout the three month harvesting season, more than 30 marketable oysters/m? were expected to be available for harvest. Cumulative production from standing stocks during the summer harvesting season was estimated at 121,410 oysters/acre (5,463,450 oysters/45 ac), or 24,282 bags. Estimated value of standing stocks, assuming $22.20/bag, was $539,060. ASSESSMENT OF AN OYSTER RESOURCE DEVELOPMENT PROJECT 157 However, during the warm summer months, mortality may be expected to increase over levels projected during the cooler months. Increased mortalities may be associated with the pathogenic protozoan, Perkinsus marinus and in- creased stress (Berrigan 1989, Quick and Mackin 1971). Furthermore, slower growth may also be expected during spawning peaks when oysters expend greater metabolic en- ergy on reproduction than on growth. The combined effect of these factors may reduce actual standing stocks when compared to predicted standing stocks. Theoretically, growth rates and mortality rates can be adjusted to account for variability between populations and to predict a range for potential yields. As an example, reducing growth rates by 25% and increasing mortality rates by 25% in the standing stocks previously discussed would reduce pre- dicted densities from 30 to 18 marketable oysters/m? avail- able during the summer harvesting season. This more con- servative estimate, representing the lower range of potential yields expected during the summer months, would produce 72,846 marketable oysters/acre (3,278,070 oysters/45 ac) or 14,569 bags during the summer season. Estimated value of these stocks would be $323,432, assuming $22.20/bag. Based on projected recruitment and field surveys following the summer harvesting season, potential landings from 45 acres ranged from 11,734 bags to 21,447 bags with esti- mated values of $260,495 and $476,123, respectively. Reported landings from all reefs in the summer har- vesting area suggested that conservative yield estimates for Bulkhead Bar may more closely reflect standing stocks and landings during the summer season than estimates based on more optimal conditions. Additionally, projections based on recruitment, growth, and mortality during optimal pe- riods (October through April) may not accurately estimate population levels during the warmer months (May through September). Economic Benefits Success of resource restoration programs is difficult to evaluate within an environmental and economical context. The environmental value of oysters resources, although clearly identifiable as a critical element in the Apalachicola Bay ecosystem, can not easily be expressed in economical terms. Within an economical framework, the most practical alternative is to identify economic contributions of restored resources to commercial fisheries and dependent industries. Economic value to harvesting and marketing sectors con- sists of revenues from landings and revenues generated by added value through wholesale and retail sales. In this con- text, the present evaluation of an oyster resource restoration project provides an analytical framework based on revenues from commercial landings, predictions of revenues based on stock assessments, and predictions of added value rev- enues. Restoration of Bulkhead Bar was completed at a cost of $174,265. The project was accomplished by contract and included all costs except costs of contract administration. Economic benefits were determined by monitoring com- mercial landings during controlled harvesting and by pre- dicting yields during the regular summer harvesting season. During controlled harvesting, 6,082 bags (adjusted) of oysters were landed and valued at $135,020. Dockside value accounted for 77% of restoration costs during initial harvesting efforts. Furthermore, estimated yields during the summer harvesting season resulted in additional values ranging from $260,495 to $476,123. Combining the more conservative value for estimated yields during the summer harvesting season, accounting for approximately 25% of dockside value of landings during the summer harvesting season, and dockside value of landings during the con- trolled harvesting period, would produce revenues of $395,515 after two years. Actual and estimated revenues from restored resources on Bulkhead Bar indicated that restoration costs were re- covered after the first harvesting season or within two years of restoration. Cost:benefit ratios based on predicted dock- side values ranged from 1:2.3 to 1:3.5 after two years. Ex- perience has shown that restored reefs remain productive for ten or more years (Whitfield 1973). With no further costs to maintain reefs over this period, conservative cost:benefit ratios may reach 1:9.2 after five years and 1:20.7 after ten years, assuming continued productivity. Cost:benefit ratios of 1:20 have been reported for suc- cessful shell plants in Louisiana (Dugas 1988). Considering restoration costs of $3,873/acre, benefits would exceed $8 ,790/acre annually and range from $2.30 after the second year to $20.70 over ten years for each $1.00 expended. Apalachicola Bay oysters are sold throughout the United States. Much of the nationwide distribution is concentrated in sales for the half-shell market. Prochaska and Keithly (1984) reported that 78% of sales were generated through sales of unshucked oysters. This trend has continued since 1985, and sales of shellstock remain the primary marketing channel. Because of this marketing strategy, added value to the product is primarily in distribution rather than pro- cessing. Currently, added value from processing may con- sist simply of washing, grading, and packaging. In this marketing strategy, the majority of added value is received by distributors and retailers who may be outside the local industry. Colberg and Windam (1965) indicated that oyster tongers share about 14%, packers about 7%, and truckers and retailers about 79% of retail value of half-shell oysters. In a review of the U.S. oyster industry, Dressel et al. (1983) reported that, oyster harvesters receive 33.3% of retail dollar sales, while the remaining 66.7% goes to pro- cessors, distributors, wholesalers, and retailers. Estimated values for Texas’s oyster industry were recently generated using an economic multiplier of $3.12 dollars for each $1.00 of direct input (Quast et al. 1988). Roberts (1988) used an economic multiplier of 6.9 to determine retail values from dockside values for domestic oyster sales na- 158 BERRIGAN tionwide. In the absence of better indicators, a multiplier of $4.00 for each $1.00 in dockside value may be appropriate to express the economic impact of Apalachicola Bay oysters at the retail level (Whitfield 1973). Thus, economic contributions from resource restoration efforts ultimately benefit levels from harvest to retail sales. Added value revenues exceeding $35,000 per acre annually represent returns ranging from $9.20 after the second year to $82.80 after ten years for each $1.00 expended. Cost:benefit ratios increase to 1:36.8 after five years and to 1:82.8 after ten years. Estimated economic benefits from Bulkhead Bar after two years would reach $1,575,000, ex- ceeding costs for the entire program and restoration of 385 acres in Apalachicola Bay. The economic estimates developed for Bulkhead Bar are expected to be representative for reef restoration efforts. However, accurate fisheries information is lacking for all other reefs restored during the program, hence there is no assurance that oyster populations on Bulkhead Bar are representative of oyster populations on other reefs. The vagaries of environmental conditions throughout Apa- lachicola Bay strongly influenced population dynamics on individual reefs throughout the recovery phase. While production may be variable between reefs, the magnitude of returns from restored productive acreage probably more than compensates for periods of low productivity. At a time when oyster resources are increasingly stressed by a barrage of factors and resource managers face tightening fiscal constraints to rehabilitating depleted re- sources, proven resource restoration and development prac- tices remain viable and economical alternatives for oyster fisheries management. Restoring suitable habitat provides shellfish resource managers the almost singular opportunity to mitigate resource losses, enhance productivity, and con- tribute direct economic benefit to the fishery industry and its dependent economy. REFERENCES CITED Bermigan, M. E. 1988. Management of oyster resources in Apalachicola Bay following Hurricane Elena. J. Shellfish Res. 7(2):281—288. Berrigan, M. E. 1989. Oyster resources in Apalachicola Bay. (Unpub- lished Manuscript) Fla. Dept. Nat. Res. Tallahassee, Florida. 93 p. Colberg, M. R. & D. M. Windam. 1965. The oyster-based economy of Franklin County, Florida. U.S. Pub. Health Serv., Washington D.C. 23 p. Dressel, D. M., D. Whitaker & T. Hu. 1983. The U.S. oyster industry: an economic profile for policy and regulatory analysts. Natl. Mar. Fish. Serv., Washington, D.C. 44 p. Dugas, R. J. 1977. Oyster distribution and density on the production por- tion of state seed grounds in southeastern Louisiana. La. Dept. Wildl. Fish. Tech. Bull. No. 1. 27 p. Dugas, R. J. 1988. Administering the Louisiana oyster industry. J. Shell- fish Res. 7(3):493—499. Futch, C. R. 1983. Oyster reef construction and relaying programs. An- dree, S. ed. Apalachicola oyster industry: conference proceedings. Florida Sea Grant College Rept. No. 57:34—38. Hofstetter, R. P. 1981. Rehabilitation of public oyster reefs damaged or destroyed by a natural disaster. Management Data Series No. 21. Texas Pks. & Wildl. Dept. Austin, Texas. 9 p. Ingle, B. M. & C. E. Dawson. 1952. Growth of the American oyster, Crassostrea virginica (Gmelin), in Florida waters. Bull. Mar. Sci. Gulf Carib. 2(2):393—404. Ingle, B. M. & W. K. Whitfield, Jr. 1968. Oyster culture in Florida. Fla. Board. Conser. Mar. Res. Lab., Ed. Ser. No. 5. 25 p. Mackenzie, C. L., Jr. 1977. Development of an aquacultural program for rehabilitation of damaged oyster reefs in Mississippi. Mar. Fish. Rev. 39(8):1—13. May, E. B. 1971. A survey of the oyster and oyster shell resources in Alabama. Ala. Mar. Res. Bull. 4:1—53. Prochaska, F. J. & W. R. Keithly. 1985. Market structure and channels for Florida processed and marketed oysters. Ward, D. R. & G. Treece. eds. Proceedings Tenth Annual Tropical and Subtropical Fish- eries Conference of the Americas. Texas A&M Univ. Sea Grant. No. 86-102. p 23-31. Quast, W. D., M. A. Johns, D. E. Pitts, Jr., G. C. Matlock, & J. E. Clark. 1988. Texas Oyster Fishery Management Plan. Fishery Man- agement Plan Series No. 1. Texas Pks. & Wildl. Dept. Austin, Texas. 178 p. Quick, J. A., Jr., & J. G. Mackin. 1971. Oyster parasitism by Labyrinth- omyxa marina in Florida. Fla. Dept. Nat. Resour. Mar. Res. Lab., Prof. Pap. Ser. No. 13. 55 p. Roberts, K. 1988. Economic Profile of the U.S. Oyster Industry. Burrage, D. ed. The Mississippi Oyster Industry: Past Present and Future. Mis- sissipp/ Alabama Sea Grant Consortium, MASGP 88-048. p 4-7. SAS Institute, Inc. 1985. SAS User’s Guide: Statistics. Cary, North Caro- lina. 956 p. Tyler, A. V. & V. F. Gallucci. 1980. Dynamics of fished stocks. Lackey R. T. & L. A. Nielsen. eds. Fisheries Management. Oxford, England: Blackwell Scientific Publications. p 111-147. Whitfield, W. K., Jr. 1973. Construction and rehabilitation of commercial oyster reefs in Florida from 1949 through 1971 with emphasis on eco- nomic impact in Franklin County. Fla. Dept. Nat. Resour. Mar. Res. Lab., Spec. Sci. Rept. No. 38. 42 p. Whitfield, W. K., Jr. & D. S. Beaumariage. 1977. Shellfish management in Apalachicola Bay. Past-present-future. Livingston, R. J., & E. A. Joyce, Jr. eds. Proceedings of the Conference on the Apalachicola Drainage System. Fla. Mar. Res. Lab. Publ. No. 26:130—140. Journal of Shellfish Research, Vol. 9, No. 1, 159-163, 1990. PHYSIOLOGICAL RESPONSES TO ACUTE TEMPERATURE ELEVATION IN OYSTERS, CRASSOSTREA VIRGINICA (GMELIN, 1791), PARASITIZED BY HAPLOSPORIDIUM NELSONI (MSX) (HASKIN, STAUBER, AND MACKIN, 1966)! D. T. J. LITTLEWOOD AND S. E. FORD Rutgers University Cook College/New Jersey Agricultural Experiment Station Shellfish Research Laboratory Port Norris, New Jersey 08349 ABSTRACT We investigated the effects of acute temperature elevation (20°C to 30°C), as an external source of stress, on the metabolism of parasitized and unparasitized oysters, Crassostrea virginica (Gmelin, 1791). Oysters infected with the protozoan parasite Haplosporidium nelsoni (MSX) (Haskin, Stauber and Mackin, 1966), displayed similar rates of oxygen uptake as uninfected animals at ambient temperature (20°C). In response to rapid temperature elevation, however, the infected animals consumed nearly 70% more oxygen than did uninfected individuals. The immediate cause of the observed response may have been due directly to the parasite, or to the fact that parasitism is associated with a loss of condition. Thus, dry weights (but not total weights or shell dimensions) of infected oysters were lower than those of uninfected animals, resulting in possible size-related differences in metabolic rates. Nevertheless, the ultimate correlate was parasitism. Our results indicate the need to investigate the effects of disease on physiological rates under changing, as well as ambient or static conditions. We conclude that the effect of external sources of stress, natural or anthropogenic, on C. virginica may compound the damage caused by parasitism and further compromise the ecological fitness of the host. KEY WORDS: Crassostrea virginica, Haplosporidium nelsoni, MSX, oyster, temperature stress, physiology INTRODUCTION Stress as applied to bivalve molluscs has been defined as ‘‘the environmental stimulus which, by exceeding a threshold value, disturbs normal animal function’’ (Bayne, 1985). The response of molluscs to many sublethal envi- ronmental stressors has received considerable attention over the past two decades, but the presence of parasites as a source of stress has been conspicuously absent (Bayne, 1985; Newell & Barber, 1988). Still fewer studies address combined effects of sublethal environmental stressors and parasitism on hosts, despite the obvious value in under- standing the etiology of parasite-induced diseases prevalent in changeable environments. Recent evidence has shown that infections of the proto- zoan parasite Haplosporidium nelsoni (MSX) (Haskin et al., 1966) reduce clearance rates, condition index (Newell, 1985), gamete development, relative fecundity and tissue lipid, protein and glycogen (Ford & Figueras, 1988; Barber et al. 1988 a, b) in its host Crassostrea virginica (Gmelin, 1791). The parasite is therefore a source of stress that should compromise the ‘‘ecological fitness’’ of the oyster (Newell & Barber, 1988). However, Newell (1985) found no significant differences between the oxygen consumption of infected and uninfected oysters. This is an unexpected result since the presence of most parasites contributes to the energetic requirements of their hosts (Thompson, 1983; see also Barber et al., 1988 a, b; Ford & Figueras, 1988), al- though Koehn & Bayne (1989) argue that a potential source 1Contribution #90-02 of the Institute of Marine and Coastal Sciences, Rutgers University of stress may be neutralized by homeostatic physiological compensation. Newell (1985) made his measurements at ambient temperatures of 10, 15 and 24°C on acclimated animals, which may not provide as sensitive an index of the parasite-induced stress response as those subjected to a sudden environmental change. We wished to determine whether differences in meta- bolic rate associated with MSX infection, which are not apparent under relatively stable conditions, might occur when oysters experience a rapid but realistic change in am- bient conditions. This would indicate parasite-related dif- ferences in the ability of oysters to maintain homeostasis through physiological compensation. Temperature is the major determinant of metabolic rate in poikilotherms and an acute short-term temperature increase that induces thermal stress is an easily manipulated variable, which occurs naturally within the intertidal and shallow subtidal zones. We report a preliminary study designed to investi- gate the interactions of parasitism and thermal stress on two physiological processes: oxygen consumption and ammonia production, and an integral of the two (O:N ratio). These measures of metabolism were determined for uninfected and MSX-infected oysters at ambient temperature (20°C) and at an elevated temperature (30°C) that is within the zone of thermal tolerance for this species (Shumway & Koehn, 1982). MATERIALS AND METHODS Oysters (n = 44) with shell heights between 55—90 mm were selected from the survivors of stocks that had been exposed to infections in lower Delaware Bay since June 159 160 LITTLEWOOD AND FoRD 1988. Ambient temperature at the time of collection, in June 1989, was between 20 and 22°C. Oysters were scrubbed and soaked for 20 minutes in a 0.1% solution of hypochlorite (Clorox Bleach) to remove epibionts such as mudworms (Newell, 1985) and were maintained in a holding tank supplied with natural bay water (20°C and 18 ppt) pumped directly from the Delaware Bay and changed twice daily. For each oyster, shell height, shell length, and whole weight were recorded, and whole volume was deter- mined by displacement. Four cylindrical respiratory chambers (165 mm radius x 87 mm height) were used to determine oxygen con- sumption and ammonia excretion. Each chamber rested on a magnetic stirrer. Single animals were placed on perfo- rated shelves over magnetic stir bars in each of three chambers, which had been filled with air-saturated water, at ambient salinity (18 ppt), previously warmed to the trial temperature. The fourth chamber served as a control (without an oyster) to determine background ammonia pro- duction and oxygen consumption. Temperature was main- tained by recirculating water from a temperature controlled bath (+0.1°C) around each chamber. After animals were in place, the chambers were sealed, air bubbles removed, and pO, monitored continuously with polarographic ox- ygen electrodes (Mickel et al., 1983) for 1 to 1.5 hr. The rate of oxygen consumption (VO,) was calculated for each oyster from a linear regression of oxygen concentration against time. Weight specific oxygen consumption (QO,) was subsequently calculated by dividing VO, by the weight of dry meat weight and was expressed as ml O,- g~! dry meat weight - h~! (Bayne et al., 1976). Respiration rates were calculated only for the period when the oyster was actually consuming oxygen. Oysters usually opened and began consuming oxygen within 10—15 mins from the be- ginning of the run. Between runs, oxygen electrodes were held in filtered sea water at the trial temperature. Temperatures of 20 and 30°C were used sequentially on each oyster. We chose to subject the same animals to both temperatures in an attempt to minimize the individual variability typical within populations of C. virginica (Shumway, 1982). After measurements at ambient temper- ature, oysters were returned to the holding tank for 24 hr before determining oxygen consumption and ammonia pro- duction at the elevated temperature. At the end of each run, three 10-ml samples of water were removed from each experimental and control chamber. Ammonia concentration was determined ac- cording to the method of Solorzano (1969) as outlined by Widdows (1985b); values were corrected for background (control) ammonia and expressed as a function of dry meat weight of the oyster (ug - g~! dry meat weight - h~!). Ox- ygen-to-nitrogen ratios (O:N) were calculated using atomic equivalents (Widdows, 1985b) based on total oxygen con- sumed and total ammonia produced during each run. At the end of each experiment, oysters were fixed in Davidson’s fixative. Cross-sections of tissue were subse- quently removed, weighed, and prepared for histology. The remaining tissue was weighed, then dried at 80°C for total dry weight estimation according to methods described in Barber et al. (1988b). Presence or absence of MSX infec- tion was scored from examination of histological sections (Ford & Haskin, 1982). Because some oysters died before physiological measurements could be completed, actual sample sizes from which data were taken are given in the results. To determine whether there was a difference in response to the temperature change associated with infection, we employed a paired comparison test, using animals that had been measured at both trial temperatures. To test for differ- ences between infected and uninfected oysters at the two temperatures, we performed a separate analysis of variance for data collected at each temperature. Because the dry weight (but not whole weight or shell dimensions) of in- fected oysters was less than that of uninfected animals, we further analysed data that showed significant differences by using analysis of covariance, with both size and infection state as covariates. RESULTS Histology revealed that 40% of the oysters were infected with Haplosporidium nelsoni. Of these MSX infected indi- viduals, 14% showed light epithelial gill infections, 57% showed light systemic infections and the remaining 29% showed advanced systemic infections. Shell dimensions, whole animal weights, and displacement volumes were the same for both infected and uninfected oysters, but dry meat weights of infected animals were only two-thirds that of uninfected individuals (Table 1). Dry meat weights ranged from 1.18—3.14 g in uninfected oysters and 0.45—2.89 g in infected oysters. However, there were no statistically sig- nificant differences between the two groups with respect to any size related measurement (p > 0.05 in one-way ANOVAs). Negligible volumes of oxygen (less than 1% total initial oxygen per hour) were consumed in the control chambers. Oxygen uptake rates (QO,) of infected and uninfected oysters were approximately equal at ambient temperatures (Fri36, = 9-002, P = 0.964; Fig. 1). Paired comparisons TABLE 1. Mean + 1 SD of oyster size variables for uninfected and MSX infected oysters; sample size (N). Uninfected Infected shell height (mm) 75 + 8 (23) 72 + 9 (15) shell length (mm) STEERS SOE 7 whole weight (g) 63.9 + 15.5 69.1 + 21.5 volume (ml) 32:5) 59:0 esis ss 1K0k7/ dry meat weight (g) 2.01 + 0.64 1.33 + 0.76 THERMAL STRESS RESPONSE IN OYSTERS WITH MSX 161 Oxygen consumption (ml/g/h) 150 100 50 Ammonia production (ug/g/h) O:N Figure 1. (a). oxygen consumption, (b). ammonia production and (c). O:N ratios of Crassostrea virginica uninfected and infected with Ha- plosporidium nelsoni at ambient (20°C) and elevated (30°C) tempera- tures. Bars represent mean values + 1 SE; sample size is indicated within open boxes; in (a) and (b) data taken only from oysters moni- tored at both temperatures. tests showed that temperature elevation had no significant effect on QO, within the uninfected group (P = 0.291) where mean QO, increased by 30%. Temperature caused a much greater increase (P = 0.054) amongst infected an- imals (Table 2; Fig. 1) with mean oxygen consumption doubling between 20 and 30°C. At 30°C, QO, was signifi- cantly greater amongst infected compared with uninfected oysters (F,, 33; = 5.764, P = 0.021); however, when size was introduced as a covariate, the effect of infection be- came statistically insignificant (P > 0.373). Mean Q,,9'70-30°C] was 2.77 and 5.14 for uninfected and infected oysters respectively. Q;) values were not signifi- cantly different between these groups (Fj,33; = 1.596, P = 0.215) as there was a wide range of respiratory re- sponses amongst both infected and uninfected oysters. Temperature elevation caused an increase in the rate of ammonia production in infected (P = 0.047) and unin- fected (P = 0.078) oysters (Table 2). Although only statis- tically significant amongst infected oysters the increase was about 130% in each group. There was no significant differ- ence in ammonia production between infection categories at ambient (F,; 36, = 1.279, P = 0.265) or elevated tem- perature (Fj, 33; = 2.326, P = 0.133). Ratios of O:N were arcsine transformed prior to statis- tical analysis. Mean O:N ratios fell in response to the ele- vated temperature in both groups of oysters, but a paired comparison test indicated that this change was not signifi- cant (Fig. 1, Table 2) and analysis of variance indicated no difference between infection categories at ambient (Fi, 6) = 0.799, P = 0.617) or elevated temperature (F;; 30) = 0.210, P = 0.654). DISCUSSION Previous studies of the effects of MSX infection on en- ergy metabolism of C. virginica suggested that the parasite posed an energetic burden on its host, but detected no dif- ferences in metabolic rate between infected and uninfected animals (Newell, 1985; Barber et al., 1988 a, b). At am- bient temperature, we also found no effect of infection on the rate of metabolism or excretion. Even so, Newell (1985) pointed out that the reduced feeding rates resulting from infection, without compensatory lowering of energy TABLE 2. Results of paired comparison tests investigating the differences in oxygen consumption (QO,), ammonia production (NH,-N) and O:N ratios between ambient (20°C) and elevated (30°C) temperature conditions; probabilities (italicized) and F ratios [df]. See Fig. 1 for plot of means + 1 SE for each infection category. Source of Variable variation Uninfected Infected QO, temperature 0.291 0.054 1.17714 20) 4.380, 13) individuals 0.105 0.187 2.82411 20) 1.9165 13) NH,-N temperature 0.078 0.047 3.37141 20) 4.715113 individuals 0.223 0.259 1.567, 20) 1.388), 13) O:N temperature 0.598 0.343 0.76144 14) 1.023, 8) individuals 0.617 0.576 0.8226) 14) 0.723118) 162 LITTLEWOOD AND FORD demand (QO,), would cause the decreased physiological condition observed in infected oysters. The present study demonstrates an additional mechanism likely to be of im- portance in natural populations. Under conditions of rapid temperature elevation, the metabolic rate of oysters in- fected by MSX was nearly 70% higher than that of oysters not infected by the parasite. Furthermore, although mean Qjo values fell within the range of a typical biochemical response in uninfected oysters (Dejours, 1975; Shumway & Koehn, 1982), high Q,) amongst infected oysters indicated severe metabolic stress associated with environmental change. This situation, if it persisted, would exaggerate the disparity between energy acquisition and expenditure of in- fected animals. The above argument is based on differences in measure- ments not standardized for animal size. Although there were no statistically significant differences in dry weight related to infection, the infected oysters were smaller than the uninfected ones, a finding consistent with previous re- ports of reduction in condition index with increasing MSX infection (Newell, 1985; Barber et al., 1988). Thus, infec- tion-associated “‘size’’ (i.e., dry weight) differences may scale physiological rates in a manner similar to size scaling due to age and growth rate differences. At the elevated tem- perature, the significant influence of infection was, in fact, removed when size was used as a covariate in analysis of variance comparing infected (smaller) and uninfected (larger) animals. Statistically, size was not a factor in the paired comparison test because the comparison, between temperatures, involved repeated measures on the same an- imals. However, we cannot exclude the possibility that the immediate physiological basis of the observed differences in response was because smaller (rather than infected) an- imals may show a greater metabolic response after acute temperature change than larger (rather than uninfected) in- dividuals. Fortunately, we can resolve this uncertainty in the case of MSX-infected oysters. Whether the observed differences were related directly to parasite activities or in- directly to parasite-induced dry weight loss, the ultimate correlate was with parasitism. The atomic O:N ratio is recognized as a good indicator of overall metabolic state in many molluscs. In general, a reduction in O:N values is consistent with increasing stress (e.g., Widdows et al., 1981) reflecting increased protein catabolism as food reserves are depleted (Mann, 1978). Ratios of less than 30 are taken to suggest protein catabo- lism in bivalve molluscs (Widdows, 1978; Widdows, 1985a). Most oysters in our study had ratios of less than 30 and we found no significant differences related to infection, although ratios fell in response to acute temperature eleva- tion. Overall, O:N ratios in this study were low. Mayzaud (1973) suggests that short measurement periods for trials of this nature may lead to a greater excretion or lower respira- tion rate without comparative change in the other variable. One such mechanism by which this may have occurred in our short-term experiments is that whilst oyster valves were adducted, during which time aquatic oxygen consumption is zero, oysters may still have been generating and building up nitrogenous waste, from the digestion and assimilation of food retained in the gut, at a greater rate than the oxygen debt incurred. In our system, the O:N ratio appeared to be a less sensitive measure of physiological response to pre- sumed stressors than oxygen uptake or ammonia excretion alone. The metabolic response to acute temperature elevation (within minutes to hours) is ‘‘an intrinsic property of the system’’ (Prosser, 1973), a measure of physiological ho- meostasis, and the capacity to buffer against environmental change (Koehn & Shumway, 1982). Our results suggest that in MSX infected oysters, energy-demanding processes may be more sensitive to rapid temperature change, perhaps through an inability to buffer against these changes, than in uninfected individuals. Further investiga- tion is required to determine the extent to which the in- crease in metabolic rate resulted directly from the meta- bolic requirements of the parasite or from the host re- sponse, such as the proliferation and increased activity of host hemocytes (Ford, 1985; Fisher et al., 1987), or indi- rectly from the decrease in dry weight of infected animals. Our observations indicate the potential risk, if only soft- tissue dry weight is considered the basis for standardizing sizes of invertebrates, of confounding “‘normal’’ growth- associated size scaling with disease-related loss of tissue. In the latter case, the disease state itself may also influence physiological rates. Further, our results demonstrate the need to investigate the effects of disease on physiological rates under changing, as well as ambient or static condi- tions. Environmental fluctuations are characteristic qualities of estuaries in response to which oysters have evolved adap- tive eurythermal, euryoxic, and euryhaline traits (Galtsoff, 1964). Nevertheless, the effect of thermal shock on metab- olism at the sublethal level demonstrates that environmental change may further reduce the ‘‘ecological fitness’’ of the infected oyster. This may be particularly true of ‘‘resis- tant’? oysters, which carry low-level infections (Ford & Haskin, 1987). Whereas highly susceptible oysters are readily killed by MSX under virtually any condition in which they become infected, the ability of more resistant oysters to tolerate parasitism is likely to be influenced by other environmental factors. In the light of the present study we suggest that other natural and/or anthropogenic stressors, characteristic of coastal environments, may compound the damage already caused by chronic, sublethal infections, which may persist in the survivors of epizootics and in their more resistant offspring (e.g., Ford, 1985; Ford & Haskin, 1987). Dis- ease may also reduce the ‘‘zone of tolerance’’ (Fry, 1947) and restrict capacity adaptations (Prosser, 1973) towards THERMAL STRESS RESPONSE IN OYSTERS WITH MSX 163 other stressors, underscoring the need to consider it as a contributory and interactive source of stress on estuarine animals (see also Couch, 1988). ACKNOWLEDGMENTS We thank S. R. Fegley for advice on statistical analyses, B. J. Barber and J. R. Healey for critically reviewing an earlier draft of this work, and R. D. Barber for determining MSX infection levels. This work is the result of research sponsored by the USDA Northeast Regional Aquaculture Center and by NOAA, Office of Sea Grant, Department of Commerce, under grant No. NA85AA-D-SG084, (Project No. R/F-24). The US Government is authorized to produce and distribute reprints for governmental purpose notwith- standing any copyright notation that may appear hereon. This is New Jersey Agricultural Experiment Station Publi- cation No. D 32405-1-90 and New Jersey Sea Grant Publi- cation No. NJSG-89-201, supported by state funds. REFERENCES Barber, B. J., Ford, S. E. & H. H. Haskin. 1988a. Effects of the parasite MSX (Haplosporidium nelsoni) on oyster (Crassostrea virginica) en- ergy metabolism. I. Condition index and relative fecundity. J. Shell- fish Res. 7:25-31. Barber, B. J., Ford, S. E. & H. H. Haskin. 1988b. Effects of the parasite MSX (Haplosporidium nelsoni) on oyster (Crassostrea virginica) en- ergy metabolism. II. Tissue biochemical composition. Comp. Bio- chem. Physiol. 91A:603—608. Bayne, B. L. 1985. General introduction. Bayne, B. L. ed. The effects of stress and pollution on marine animals. New York, NY: Praeger, xi— XVi. Bayne, B. L., R. J. Thompson & J. Widdows. Physiology: I. Bayne, B. L. ed. Marine mussels, their ecology and physiology. International Biological Programme. Cambridge: Cambridge Univ. Press, 121— 206. Couch, J. A. 1988. Role of pathobiology in experimental marine biology and ecology. J. Exp. Mar. Biol. Ecol. 118:1-6. Dejours, P. 1975. Principles of comparative respiratory physiology. New York, NY: American Elsevier. 253 pp. Fisher, W. S., M. Auffret & G. Balouet. 1987. Response of European flat oyster (Ostrea edulis) hemocytes to acute salinity and temperature changes. Aquaculture 67:179—190. Ford, S. E. 1985. Chronic infections of Haplosporidium nelsoni (MSX) in the oyster Crassostrea virginica. J. Invert. Pathol. 45:94—107. Ford, S. E. & A. J. Figueras. 1988. Effects of sublethal infection by the parasite Haplosporidium nelsoni (MSX) on gametogenesis, spawning, and sex ratios of oysters in Delaware Bay, USA. Dis. Aquat. Org. 4:121-133. Ford, S. E. & H. H. Haskin. 1982. History and epizootiology of Haplo- sporidium nelsoni (MSX), an oyster pathogen, in Delaware Bay, 1957—1980. J. 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. Parasitol. 73:368—376. Fry, F. E. J. 1947. Effects of the environment on animal activity. Univ. Toronto Stud. Biol., Ser. 55:1—62. Galtsoff, P. S. 1964. The American oyster Crassostrea virginica Gmelin. U.S. Fish and Wildlife Service, Fish. Bull. 64. Washington, D.C. 480 pp- Haskin, H. H., L. A. Stauber, & J. G. Mackin. 1966. Minchinia nelsoni n.sp. (Haplosporida, Haplosporidiidae): causative agent of the Dela- ware Bay oyster epizootic. Science 153:1414—1416. Koehn, R. K. & B. L. Bayne. 1989. Towards a physiological and ge- netical understanding of the energetics of the stress response. Biol. J. Linn. Soc., 37:157-171. Koehn, R. K. & S. E. Shumway. 1982. A genetic/physiological explana- tion for differential growth rate among individuals of the American oyster, Crassostrea virginica (Gmelin). Mar. Biol. Lett. 3:35—42. Mann, R. 1978. A comparison of morphometric, biochemical and physio- logical indexes of condition in marine bivalve molluscs. In: Energy and environmental stress in aquatic systems. DOE Symp. Ser. No. 48, 484-497. Mayzaud, P. 1973. Respiration and nitrogen excretion of zooplankton. II. Studies of the metabolic characteristics of starved animals. Mar. Biol. 21:19-28. Mickel, T. J., L. B. Quetin, & J. J. Childress. 1983. Construction of a polarographic oxygen sensor in the laboratory. Gnaiger, E. & Forstner, H. eds. Polarographic oxygen sensors; aquatic and physio- logical applications. Berlin: Springer-Verlag, 81—85. Newell, R. I. E. 1985. Physiological effects of the MSX parasite Haplo- sporidium nelsoni (Haskin, Stauber & Mackin) on the American oyster Crassostrea virginica (Gmelin). J. Shellfish Res. 5:91-9S. Newell, R. I. E. & B. J. Barber. 1988. A physiological approach to the study of bivalve molluscan diseases. American Fisheries Society Spe- cial Publication. 18:269—280. Prosser, C. L. 1975. Physiological adaptations in animals. Vernberg, F. J. ed. Physiological adaptation to the environment. New York, NY: Intext Educational Publishers, 3-18. Shumway, S. E. 1982. Oxygen consumption in oysters: an overview. Mar. Biol. Lett. 3:1—23. Shumway, S. E. & R. K. Koehn. 1982. Oxygen consumption in the American oyster Crassostrea virginica. Mar. Ecol. Prog. Ser. 9: 59-68. Sokal, R. R. & F. J. Rohlf. 1981. Biometry. 2nd ed. San Francisco, CA: W. H. Freeman & Co., 859 pp. Solorzano, L. 1969. Determination of ammonia in natural waters by the phenolhypochlorite method. Limnol. Oceanogr. 14:799—801. Thompson, S. N. 1983. Biochemical and physiological effects of meta- zoan endoparasites on their host species. Comp. Biochem. Physiol. 74B:183-211. Widdows, J. 1978. Physiological indices of stress in Mytilus edulis. J. mar. biol. Ass. UK. 58:125—142. Widdows, J. 1985a. Physiological measurements. Bayne, B. L. ed. The effects of stress and pollution on marine animals. New York, NY: Praeger, 3—45. Widdows, J. 1985b. Physiological procedures. Bayne, B. L. ed. The ef- fects of stress and pollution on marine animals. New York, NY: Praeger, 161—178. Widdows, J., D. K. Phelps, & W. Galloway. 1981. Measurement of physiological condition of mussels transplanted along a pollution gra- dient in Narragansett Bay. Mar. Environ. Res. 4:181—194. Journal of Shellfish Research, Vol. 9, No. 1, 165-172, 1990. RECRUITMENT AND GROWTH OF OYSTERS ON SHELL PLANTED AT FOUR MONTHLY INTERVALS IN THE LOWER POTOMAC RIVER, MARYLAND! REINALDO MORALES-ALAMO AND ROGER MANN Virginia. Institute of Marine Science School of Marine Science College of William and Mary Gloucester Point, VA 23062 ABSTRACT Oyster shells were planted on four successive months (May to August 1986) in contiguous plots at Jones Shore Bar in the Potomac River, Maryland, to study the effect of differences in time of cultch planting on settlement and survival of oyster spat. The plots were usually sampled at two-week intervals from time of planting through November, 1986, and once in June, 1987. A massive concentration of the tunicate Molgula manhattensis covered the bottom in all plots within four to six or eight weeks following shell planting. A commercially acceptable number of spat per shell, between 1.8 and 2.2 (approximately equivalent to 900-1200 spat per bu), was recorded at three of the plots on June 26, 1987, in spite of the heavy tunicate fouling of 1986. Recruitment of oyster spat was lower in the plot on which cultch was planted earliest, on May 13, than in the other three plots on which cultch was planted 1—3 months later. Number of spat was highest in the plot on which shells were planted on July 14; accidental planting of cultch into two elongated mounds on that plot may have contributed to the high recruitment of spat observed. Mean spat height was lowest in the plot on which cultch was planted on August 12 and highest in the plots on which shell was planted on May 13 and June 16. The lower number of spat found on shells planted on May 13 was probably associated with the early planting date. The data suggest that combined maximum recruitment and growth of oyster spat is most likely to occur at Jones Shore on cultch planted between late June and mid-July, although plantings as early as mid-June and as late as early August may also produce commercially-acceptable results. KEY WORDS: Crassostrea virginica, oysters, fouling, recruitment, growth INTRODUCTION Oyster shells from shucking houses are planted on public and private estuarine bottoms in Virginia and Mary- land to provide new clean substrate on which larvae of the oyster Crassostrea virginica (Gmelin 1791) can set. The time selected for planting shell cultch has always been con- sidered critical to successful recruitment of oyster spat be- cause fouling by organisms and sedimentation reduce the amount of space readily available for settlement of oyster larvae (Manning 1952; Shaw 1967; Abbe 1988). Shells planted too early in the year may become heavily fouled prior to the beginning of the oyster settlement season; how- ever, if shell cultch is planted too late in the season, the peak oyster settlement period could be missed. The objective of this study was to investigate the effect of cultch planting time on recruitment and growth of oysters at Jones Shore Bar, in the lower Potomac River, under conditions similar to the usual cultch planting prac- tices of the oyster industry in that region. Jones Shore Bar was selected as the experimental site because oyster settle- ment in the Maryland shore of the lower Potomac River has usually been higher than on bars further upriver or on the Virginia shore (Davis et al. 1976; Krantz and Davis 1983; Whitcomb 1985). MATERIALS AND METHODS The study site at Jones Shore was located on the north side of the Potomac River, approximately 6.5 km upriver ‘Contribution No. 1606 of the Virginia Institute of Marine Science, School of Marine Science, College of William and Mary, Gloucester Point, VA 23062. from Point Lookout and | km from the shoreline (Fig. 1); water depth at that location is approximately 3.6 m at mean low water. The river bottom at the site had a muddy sand texture with scattered clumps of oysters and shells prior to introduction of the experimental shell cultch. The experimental area was a square approximately 20 m on each side aligned parallel to the shoreline. The area was divided into four square plots (labelled A, B, C and D), each approximately 100 m*. The central juncture of the four plots was defined by an existing cylindrical steel marker; this marker was also the structure from which shellstrings were suspended in spatfall-monitoring studies of the Virginia Institute of Marine Science. The boundary between adjoining plots was marked on the outside edge by a wooden pole. Oyster shells were broadcast from a barge over each plot by a private contractor in the manner em- ployed by commercial oyster growers. Plantings were made at monthly intervals in 1986: plot A on May 13 (361 bu), plot B on June 16 (380 bu), plot C on July 14 (418 bu) and plot D on August 12 (361 bu). Divers’ observations of the bottom in each of the plots following planting of cultch indicated that shell distribution over plots A and B was uneven, with scattered areas in which no new cultch was found. Shell distribution over plot D was more even than in plots A and B. Shells in plot C were accidentally concentrated into two elongated mounds approximately 5 m long, 2—3 m wide and 1.5 m high, joined at one end to form a V with an angle of approxi- mately 45 degrees and the apex pointing in a N-NE direc- tion toward the central cylindrical marker. Shell samples were collected at 2-week intervals be- tween June 3 and November 4, 1986; except that no collec- 165 166 MORALES-ALAMO AND MANN OYSTER BARS IN LOWER POTOMAC RIVER Wy OYSTER GROUND BONUMS BAR Voy! Figure 1. Chart of the lower Potomac River showing location of Jones Shore oyster bar. Approximate location of experimental station is marked by an X. Modified from Haven (1976). RECRUITMENT AND GROWTH OF POTOMAC RIVER OYSTERS tions were made on August 12 and on October 7 and 21 because of inclement weather or other unavoidable circum- stances. No sample collections were made between No- vember 1986 and June 1987. Samples were collected by SCUBA-equipped divers. Marked floating lines guided the divers to the approximate location of three randomly-selected quadrats in each plot. A 0.25-m? square frame was dropped over the bottom at each of the selected quadrats and two plastic 4-liter bags were filled with shells from the area within the frame. In instances when the frame landed in an area devoid of new cultch, it was moved to the nearest shell concentration. Lo- cations sampled on plot C were selected differently because of the aggregation of shells into two mounds; there, the square frame was placed on the side of the mound closest to the location of the selected quadrat. The height on the mound from which shells were collected was arbitrarily chosen by the divers. Shell samples were transported to the laboratory in large plastic buckets filled with river water where they were placed in a 4% solution of ethanol in river water for 2 hr prior to preservation in a 70% solution of 167 ethanol. Temperature measurements were made at the sta- tion and water samples collected for salinity determina- tions. Oyster spat on the shells were counted and measured after the shells were air-dried. An oyster spat is defined here as the attached post-larval form that shows evidence of shell growth beyond the margin of the larval shell. Spat were also counted and measured on other shells selected at random from the three subsamples when needed to increase the number of shells examined to 20. Height of each spat was measured as the distance from the umbo to the farthest point on the opposite edge of the shell. Measurements were grouped into height class intervals of 4 mm. Analysis of variance and Scheffe’s multiple contrast test (Zar 1984) were used to compare means when variances were homogeneous. In cases where the variances were het- erogeneous, the nonparametric Mann-Whitney test (Olson 1988) was applied for mean comparisons. A significance probability level of 0.15 was used for rejection of the null hypothesis in comparisons of mean number of spat and mean spat height between plots and dates to enhance per- as) 1.0 0.5 72 || 104 |) 106 lh 85 ‘a 114 PLOTD o 0 = 1.0 iu = hi 0.5 r < 80 149 148 200 215 | i 137 PLOTC a. 0 7) 0.5 Vs 1 30 45 94 105 99 109 121 PLOTB a alas le = ina a is” = 2 4 ds} Zz z 4 = af oO a a > z < =) =) > =) > Ww Ww (e) =) WwW =) = =) 7 10 S = ne) a © o 2 +t Sc 0.5 1 1 145 25 | 31 US 69 | 49 | 38 PLOTA 0 40 80 0 40 80 0 40 80 0 40 80 0 40 80 0 40 80 1986 1987 SPAT HEIGHT (mm) Figure 2. Mean number of spat per shell in different shell height classes for groups of 60 oyster shells collected on different dates from cultch planted at four experimental plots on Jones Shore Bar in the Potomac River. Shell height intervals of 4 mm. Value for single bar on July 15 in plot A was 2.4. Shell cultch was planted at monthly intervals in 1986: plot A on May 13; plot B on June 16; plot C on July 14; plot D on August 12. Total number of spat given by each histogram. 168 ception of the probable relationship among means while maintaining a low probability of committing a Type I error. The coefficient of variation (Sokal and Rohlf 1981) was computed as a measure of the relative variability of the data on number of spat per shell. RESULTS Visual observation of the bottom by divers indicated that the tunicate Molgula manhattensis appeared to cover com- pletely, or almost completely, the experimental shell sub- strate within 4—6 weeks after the shells were planted (8 weeks in plot A). A heavy tunicate cover persisted through the last sampling date in 1986 (November 4). Diver obser- vations indicated that tunicate coverage was considerably lower on June 26, 1987, than was found during most of the summer in 1986. Many tunicate clusters were lost during collection and handling of shells because the strength of their attachment to the shells was easily overcome by the weight of the clusters. Those losses prevented accurate quantification of fouling; however, the presence of other fouling organisms, predominantly barnacles and encrusting bryozoa, was evident on most of the shells. Spat were first found in plot A on June 17, 1986, ap- proximately one month after shells were planted (Fig. 2). At the other plots, spat were first found on the first sam- pling date, two weeks after shell planting. The first sub- stantial number of spat (15 or more) was not found in plots A and B until July 15, eight and four weeks after planting, respectively; substantial numbers, however, were found in plots C and D only two weeks after planting. Spat < 8.0 mm were presumed to have set in the two weeks preceding the sampling date because almost all spat in samples collected two weeks after shells were planted were 8.0 mm or smaller. This assumption was supported by the bimodal size frequency distribution of spat in later samples, which could be separated into two distinct size groups, one composed of spat < 8.0 mm and the other one made up of spat > 8.0 mm (Fig. 2). After July 15, spat < 8.0 mm were found at all plots in substantial numbers on every sampling date through Sep- tember 23 and in reduced numbers on November 4 (Figs. 2 and 3). They were also present in plots C and D on June 26, 1987, but in very low numbers. According to data col- lected by the Virginia Institute of Marine Science, using shellstrings suspended over the bottom and exposed for one-week intervals, oyster settlement at the experimental site in Jones Shore extended from the week of July 7—14 to the week of September 1—8 in 1986 (Whitcomb 1986). Thus, the number of spat < 8.0 mm on September 23 may represent settlement after September 8 that was not ob- served on the suspended shellstrings. Water temperature was 24°C on September 23 (Table 1); this was sufficiently high to permit continued spawning by oysters. The pres- ence of spat < 8.0 mm on November 4 was probably the MORALES-ALAMO AND MANN NUMBER SPAT SHELL! zaOoOaz FZ — | (Ojifet =) =)'s) (iS) >>) 10S) Sey ie) =) ® Bae wo nw oO rHeea ~~ San PLOT B PLOTC PLOT D Figure 3. Mean number and 95% confidence interval of spat per shell in groups of 60 oyster shells collected on different dates from cultch planted at four experimental plots on Jones Shore Bar in the Potomac River. Shell cultch was planted at monthly intervals in 1986 as indi- cated in legend for Figure 2. result of lag in growth of the spat, rather than of new re- cruitment, because water temperature had declined to 15.5°C between September 23 and November 4. The low numbers recorded on June 26, 1987, most likely represent early spat set on that summer. No significant difference (P < 0.15) could be detected in mean number of spat < 8.0 mm between plots A and B TABLE 1. Water temperature and salinity at Jones Shore, Potomac River, Maryland, on sampling dates at experimental area on which shell cultch was planted. Temperature (°C) Salinity (%c) Date Surface Bottom Surface Bottom 1986 June 3 21.5 20.0 13.98 14.12 17 26.5 25.5 13.64 14.57 July 1 24.5 24.0 15.51 15.06 15 28.0 27.4 14.82 14.87 29 30.0 28.6 14.87 16.02 Aug 26 24.8 24.5 16.87 17.00 Sept 9 22.5 23.0 17.02 17.14 23 23.5 24.0 17.93 18.14 Nov 4 15.9 15.5 18.55 18.64 1987 June 26 26.0 25.8 14.10 13.97 RECRUITMENT AND GROWTH OF POTOMAC RIVER OYSTERS 169 on any of the sampling dates except one, primarily because of the high variation among samples in each plot; the ex- ception was found on July 15 when the highest number of spat in that size group recorded during the study occurred in plot A (Fig. 3, Table 2). Mean number was significantly higher in plot C than in plots A and B on every sampling date but one (August 26), suggesting that recruitment of newly-set spat was greater in plot C than in A and B. No difference was evident, however, between plots C and D, probably because cultch was planted during peak spatfall periods in those two plots. Mean number of spat > 8.0 mm increased significantly (P < 0.15) with time in all plots as a result of the contin- uous recruitment through the settlement season (Figure 3). Mean number of spat per shell was significantly higher in plot C than in the other plots on most dates (Fig. 3, Table 2). Likewise, on most dates, mean number of spat was sig- nificantly lower in plot A than in the other plots. On Sep- tember 23, however, there was no evidence of a difference in mean number of spat > 8.0 mm between plot A and plots B and D, the probable result of better than usual re- cruitment in plot A during the preceding weeks. The coefficient of variation (CV) for mean number of spat < 8.0 mm shell was considerably lower in plot C than in the other plots on all but one of the sampling dates, (Table 3); the exception was September 23, when CV was also lower in plots A, B and D than on any of the other sampling dates (with the exception of July 15 in plot A) indicating a reduction in variability among samples col- lected on that date. We cannot suggest an explanation for the lower CV values on September 23. CV for mean number of spat > 8.0 mm was relatively high on all sam- pling dates. Size frequency distribution in all four plots was approxi- mately bell-shaped on June 26, 1987, although numbers were low in plot A (Figure 2). In plot B the frequency dis- tribution was slightly skewed towards the larger sizes and in plots C and D it was slightly skewed towards the smaller sizes, which reflects the presence of older (thus, larger) spat in plot B. Height differences between plots among spat > 8.0 mm were closely related to the time of shell planting except that mean height of spat > 8.0 mm was similar in plots A and B on most dates (Fig. 4, Table 4). On most dates, mean height was significantly higher (P < 0.15) in plots A and B than in plots C and D and on all dates mean height was significantly lower in plot D than in the other three plots. Differences in mean height could not be detected between plots A and B on most dates, probably due to a scattered distribution of spat over the size range in plot A (Figure 2). There were, however, more spat in the larger size classes in plot B than in plot A on all sampling dates (Figure 2) indi- cating better survival and growth in B than in A. DISCUSSION The complete or nearly complete cover of the bottom substrate by the tunicate Molgula manhattensis observed by divers early in our study indicated a dominance of fouling TABLE 2. Probability values for Mann-Whitney tests between mean number of spat per shell in paired experimental plots at Jones Shore, Potomac River, Maryland, on sampling dates following planting of clean shell cultch. Cultch planted on staggered dates in 1986 at four plots: plot A on May 13, plot B on June 16, plot C on July 14 and plot D on August 12. Probabilities <0.15 underlined. Superscripts identify plots with higher mean. Date July 1 July 15 July 29 Size Group: <8.0 mm Plot A vs. Plot B 1.00 0.054 0.70 vs. Plot C 0.04¢ vs. Plot D Plot B vs. Plot C BIBS vs. Plot D Plot C vs. Plot D Size Group: >8.0 mm Plot A vs. Plot B 1.00 vs. Plot C 0.80 vs. Plot D Plot B vs. Plot C 1.00 vs. Plot D Plot C vs. Plot D 1986 1987 Aug 26 Sept 9 Sept 23 Nov 4 June 26 0.18 0.70 0.39 0.39 0.06° 0.06° 0.01° 0.07 0.18 0.39 0.02 0.59 0.31 0.09° 0.00° 0.03° 0.69 0.39 0.03? 0.24 0.69 0.82 0.70 0.48 0.59 0.058 0.158 0.22 0.008 0.008 01° 01° 0.01° 0. 0.00° 0.32 0.98 BLOP 0.00? 0.08° 0.50 0.08° 08° 0. 0.138 0.138 0.118 0.89 o1c 0.03 0.01° 0.43 170 MORALES-ALAMO AND MANN TABLE 3. Coefficient of variation (Std. Dev./Mean x 100) for number of spat per shell on sampling dates at four experimental plots planted with clean shell cultch at Jones Shore, Potomac River, Maryland. Values <75 underlined. Shell Height <8.0 mm Date Plot A Plot B Plot C 1986 June 3 June 17 173 July 1 173 73 15 49 65 29 122 114 49 Aug 26 126 81 54 Sept 9 114 98 41 23 66 40 47 Nov 4 127 245 72 1987 June 26 155 Shell Height >8.0 mm Plot D Plot A Plot B Plot C Plot D 173 173 173 92 127 86 77 86 119 118 89 159 34 117 87 95 100 123 112 81 93 88 245 146 91 82 73 by that species in the experimental plots at Jones Shore in 1986. M. manhattensis can cover cultch surfaces com- pletely in a very short time and can reach maximum size in lower Chesapeake Bay in less than two weeks, quickly dominating new or established fouling communities (An- drews 1953 and Otsuka and Dauer 1982). Distribution and abundance of other fouling species on 40 30 20 10 MEAN SPAT HEIGHT (mm) 1986 POTOMAC RIVER: JONES SHORE shell cultch in our experimental plots was probably affected by the high density of tunicates. Those species, however, as well as oyster spat, were still able to settle and survive under the tunicate cover throughout the study. This is in agreement with Sutherland and Karlson (1977) who inter- preted results presented by Boyd (1972) as indicating that resident adults inhibit subsequent larval recruitment into a 1987. Figure 4. Mean shell height and 95% confidence interval of spat on shells collected on different dates from cultch planted at four experimental plots on Jones Shore Bar in the Potomac River. Mean height computed for spat > 8.0 mm only. RECRUITMENT AND GROWTH OF POTOMAC RIVER OYSTERS 171 TABLE 4. Probability values for Mann-Whitney tests between mean spat height of spat >8.0 mm in paired experimental plots at Jones Shore, Potomac River, Maryland, on sampling dates following planting of clean shell cultch. Probabilities <0.15 underlined. Superscripts identify plots with higher mean. Date July 1 July 15 July 29 Size Group: >8.0 mm PLot A ys. Plot B 1.00 vs. Plot C 0.50 vs. Plot D Plot B vs. Plot C 0.50 vs. Plot D Plot C vs. Plot D 1986 1987 Aug 26 Sept 9 Sept 23 Nov 4 June 26 0.018 0.94 0.37 0.128 0.36 0.06° 0.31 0.024 0.004 0.014 0.004 0.004 0.004 0.004 0.068 Ons 0.068 0.008 0.008 0.008 0.008 0.008 0.008 0.00° 0.00° 0.04¢ 0.00° fouling assemblage but do not stop it entirely. It is also partially in agreement with Young (1989), whose experi- ments with the tunicate Molgula occidentalis in Florida suggested that larval predation by tunicates may not be im- portant in determining community composition or settle- ment density of fouling assemblages. Higher recruitment of spat < 8.0 mm in plots C and D than in plots A and B may be attributed primarily to planting of shells in C and D having coincided in time with the most intense period of spat settlement at Jones Shore, thus giving oyster larvae the opportunity to settle and grow before fouling could become the potentially negative factor it is presumed to be in oyster settlement. Higher numbers of newly-set spat, as well as smaller variances among samples, in plot C may have been associated with a greater uniformity in distribution of spat over the cultch on that plot, as evidenced by the lower coefficients of variation (CV) computed for those data. This is a significant depar- ture from what appears to be the norm; Sutherland and Karlson (1977) concluded that recruitment into fouling communities appears to be a universally variable process after examining data from four different studies in which CV values for all species were extremely high, usually ex- ceeding 100. Greater uniformity in distribution of newly- set spat in plot C may have resulted from concentration of the volume of shells planted in that plot over a smaller area of bottom than in the other plots. The higher numbers of spat found in plot C may have also been related to the high elevation of the mounds and the concomitant increase in quantity of exposed surface shells and of interstitial spaces, factors which are characteristic of highly productive oyster bottoms (Haven and Whitcomb 1983, and DeAlteris 1988). The effect of time of cultch planting on oyster recruit- ment could not be correlated clearly with fouling coverage because of the massive unquantified coverage by tunicates; aggregation of cultch into mounds in plot C also interfered with interpretation of the results obtained. Consequently, definitive conclusions about the relationship between time of cultch planting, fouling, and oyster recruitment and growth cannot be advanced. Nevertheless, the lower number of spat recorded in plot A suggests that the reduced recruitment observed in that plot was most likely associated with the early planting date (mid-May) because, except for the aggregation of cultch into mounds in plot C, time of planting was the most outstanding difference between plots. Combined maximum recruitment and growth of oyster spat appears most likely to be attained at Jones Shore on cultch planted between late June and mid-July, as indicated by the absence of substantial numbers of spat before July 1 and the lag in growth of spat on shell planted in mid-Au- gust (plot D). Shell plantings as early as mid-June and as late as early August, however, may also produce commer- cially acceptable recruitment, especially in view of re- corded annual variations in spatfall peaks (Kennedy 1980). The number of spat found in plots B, C and D in June 1987, between 1.8 and 2.3 spat per shell, which translates into between 900 and 1200 spat per bushel (based on an estimated 500 shells in one bushel), support that conclu- sion. MacKenzie (1981) used a criterion of 2.5 spat per shell to define a commercially successful oyster set on shells in Long Island Sound. These suggestions may apply to most of the oyster-producing areas of the Chesapeake Bay because onset of spatfall does not vary greatly throughout the bay, as is shown by the data in Shaw (1967), Kennedy (1980) and Whitcomb (1986). ACKNOWLEDGMENTS This study was sponsored jointly by the Potomac River Fisheries Commission and the Virginia Institute of Marine Science. We are grateful to K. A. Carpenter, PRFC execu- tive secretary, for his cooperation in development and con- duct of this study; to the Maryland Department of Natural Resources and the Virginia Marine Resources Commission and their respective marine patrols and personnel for pro- viding the boats used as work platforms during sample col- 172 MORALES-ALAMO AND MANN lections; to Rick Rheinhardt for assistance in planning and implementation of the study procedures; to Bernardita Campos, Lyn Cox, David Eggleston, Kevin McCarthy, Rick Rheinhardt, Curtis Roegner, Ya-Ke Shu and Kenneth Walker for their assistance in collection and examination of the shell samples; to Joseph T. DeAlteris, Robert A. Blay- lock, Rick Rheinhardt and Curtis Roegner for valuable comments and suggestions in reviews of the manuscript; to Diane Bowers for the art work; and to Valise Jackson for transcription of the manuscript and tables. REFERENCES CITED Abbe, G. R. 1988. Population structure of the American oyster, Crassos- trea virginica, on an oyster bar in central Chesapeake Bay: changes associated with shell planting and increased recruitment. J. Shellfish Res. 7:33—40. Andrews, J. D. 1953. Fouling organisms of Chesapeake Bay. Inshore Survey Program: Interim Report 17. Chesapeake Bay Institute. So- lomons, MD. 16 pp. Boyd, M. J. 1972. Fouling community structure and development in Bo- dega Harbor, California. Ph.D. Dissertation, University of California, Davis, CA. 201 pp. Davis, J., D. Haven, K. G. Drobeck & E. A. Dunnington. 1976. Plans for management of the fisheries of the tidal Potomac River. SRAMSOE No. 117, Virginia Institute of Marine Science, Gloucester Point, VA. 38 pp. DeAlteris, J. T. 1988. The geomorphic development of Wreck Shoal, a subtidal oyster reef of the James River, Virginia. Estuaries 11:240— 249. Haven, D. S. 1976. The shellfish fisheries of the Potomac River. In: The Potomac estuary biological resources: Trends and options. Pro- ceedings of a symposium at Alexandria, VA, June 1975. W. T. Mason and K. C. Flynn (Eds.), pp. 88—94. Publ. 76-2, Interstate Commision on the Potomac River Basin, Bethesda, MD. Haven, D. S. & J. P. Whitcomb. 1983. The origin and extent of oyster reefs in the James River, Virginia. J. Shellfish Res. 3:141—151. Kennedy, V. S. 1980. Comparison of recent and past patterns of oyster settlement and seasonal fouling in Broad Creek and Tred Avon River, Maryland. Proc. Natl. Shellfish. Assoc. 70:36—46. Krantz, G. E. & H. A. Davis. 1983. Maryland oyster spat survey: Fall 1982. Tech. Rept., Publ. No. UM-SG-TS-83-01, Maryland Sea Grant Program, College Park, MD. 14 pp. MacKenzie, C. L., Jr. 1981. Biotic potential and environmental resistance in the American oyster (Crassostrea virginica) in Long Island Sound. Manning, J. H. 1952. Setting of oyster larvae and survival of spat in the St. Mary’s River, Maryland, in relation to fouling of cultch. Ad- dresses, Convention Natl. Shellfish. Assoc. pp. 74-89. Olson, C. L. 1988. Statistics: Making sense of data. Wm. C. Brown Pub- lishers, Dubuque, IA. 809 + 103 Append. pp. Otsuka, C. M. & D. M. Dauer. 1982. Fouling community dynamics in Lynnhaven Bay, Virginia. Estuaries 5:10—22. Shaw, W. M. 1967. Seasonal fouling and oyster setting on asbestos plates in Broad Creek, Talbot County, Maryland, 1963—65. Chesapeake Sci. 8:228-236. Sokal, R. R. & F. J. Rohlf. 1981. Biometry. W. H. Freeman & Co., San Francisco, CA. 859 pp. Sutherland, J. P. & R. H. Karlson. 1977. Development and stability of the fouling community at Beaufort, NC. Ecol. Monogr. 47:425—446. Whitcomb, J. 1985. 1985 Annual Summary: Oyster spatfall in Virginia rivers. Marine Resource Special Report, Virginia Sea Grant Program, Virginia Institute of Marine Science, Gloucester Point, VA. 19 un- numbered pp. Whitcomb, J. 1986. 1986 Annual Summary: Oyster spatfall in Virginia rivers. Marine Resource Special Report, Virginia Sea Grant Program, Virginia Institute of Marine Science, Gloucester Point, VA. 19 un- numbered pp. Young, C. M. 1989. Larval depletion by ascidians has little effect on settlement of epifauna. Mar. Biol. 102:481—489. Zar, J. H. 1984. Biostatistical Analysis. Prentice-Hall, Englewood Cliffs, NJ. 718 pp. Journal of Shellfish Research, Vol. 9, No. 1, 173-175, 1990. SETTLEMENT OF OYSTERS, CRASSOSTREA VIRGINICA (GMELIN, 1791), ON OYSTER SHELL, EXPANDED SHALE AND TIRE CHIPS IN THE JAMES RIVER, VIRGINIA! ROGER MANN, BRUCE J. BARBER, JAMES P. WHITCOMB, AND KENNETH S. WALKER Virginia Institute of Marine Science School of Marine Science College of William and Mary Gloucester Point, VA 23062 ABSTRACT The effectiveness of oyster shell, expanded shale, and tire chips as substrates for settlement of oysters, Crassostrea virginica (Gmelin), was compared at four locations in the James River, Virginia, over three two-week time intervals in August and September, 1988. Only differences between substrate were significant (P < 0.001). Over all locations and time intervals, a signifi- cantly higher (P < 0.001) proportion of total oyster settlement occurred on oyster shell (63.8%) than on either tire chips (22.1%) or expanded shale (14.2%). KEY WORDS: Crassostrea virginica, settlement, cultch, spat INTRODUCTION The provision of substrate (cultch) to enhance settlement of oysters can be traced to Roman times and the writings of Plinius. Widespread use of planted shell was first used in the United States by oyster growers in New York waters in the mid-nineteenth century. The practice became common- place within several decades along the entire east coast and remains a continuing activity to this day. Longterm re- moval of shells associated with prevailing harvesting prac- tices, without subsequent replacement, has resulted in a de- pletion of shells and has prompted a search for alternative cultch materials. This communication describes a comparison of ex- panded shale (natural shale heated to induce cavitation and expansion) and tire chips (shredded tire casings) to oyster shell for suitability as cultch material in an area noted for consistently high levels of natural oyster settlement. MATERIALS AND METHODS Mesh bags (1.25 cm) containing 0.1 bu of substrate were deployed at four sites in the James River, Virginia during three consecutive two-week time intervals during August and September, 1988, a period of generally high oyster settlement (Haven and Fritz, 1985). A comparison of substrates based on volume compares favorably with typical mass plantings of substrate, where thickness of ap- plication is far from constant. The sites of deployment, chosen to provide good spatial coverage and variability in settlement activity, were Naseway Shoal, Rock Wharf, Wreck Shoal, and Point of Shoals (Fig. 1). One bag of each substrate was hung 0.5 m off the bottom from wooden stakes at each of the locations on August 5, 1988. These ‘Contribution No. 1605 from the Virginia Institute of Marine Science, College of William and Mary. 173 were recovered and replaced on August 19. The second set of bags were recovered and replaced on September 2, and the third set of bags was recovered on September 16. A two-week exposure period was chosen to maximize settle- ment and allow sufficient time for metamorphosed oysters (spat) to grow to observable size, without being overgrown by fouling organisms. After being retrieved, bags of sub- strate were returned to the laboratory, air-dried, and the number of spat contained in each bag was counted with the aid of a dissecting microscope. Oyster settlement as a function of substrate over both location and time interval was examined using analysis of variance (ANOVA). RESULTS AND DISCUSSION There were obvious temporal and spatial differences in the number of spat found on each of the substrate types (Table 1). Nonetheless, in 11 out of the 12 cases, oyster shell had far greater numbers of spat than either expanded shale or tire chips. To compare substrate types between locations and expo- sure periods, settlement was expressed as a percentage of total spat count (Table 1). A three-way ANOVA on these percentages (arc sin transformed) demonstrated that only differences due to substrate were significant (P < 0.001). Data were thus pooled and a one-way ANOVA and SNK multiple comparison were used to compare settlement over all locations and exposure periods. This revealed that oyster shell had a significantly higher (P < 0.001) per- centage of spat than both expanded shale and tire chips (ex- panded shale and tire chips were not statistically different, P > 0.05). Per unit volume, oyster shell is substantially preferable as a cultch material than either expanded shale or tire chips. This argues for stringent conservation of shell resources. 174 MANN ET AL. HOG ISLAND Paga" River Naseway Shoal 3 HAMPTON ———— ; NAUTICAL WILES ROADS MILOMETERS Stee Figure 1. Sites of substrate deployment in the James River, Virginia. OYSTER SETTLEMENT ON SHELL, SHALE, AND TIRE CHIPS 175 TABLE 1. Comparison of oyster spat settlement on three substrates at four sites in the James River, Virginia, for three exposure periods during 1988. Settlement expressed as number of spat and as a percentage of total spat for that location and exposure period. Exposure Period 8/5—8/19 8/19—9/2 9/2-9/16 Tire chips, besides being less effective as cultch material, are prone to being dispersed by currents and wave action (Gibbons et al., 1989). Expanded shale, the least effective cultch material, has potential value in stabilizing marginal oyster bottom, prior to shell planting. Haven, D. S. & L. W. Fritz. 1985. Setting of the American oyster Cras- sostrea virginica in the James River, Virginia, USA: temporal and spatial distribution. Mar. Biol. 86:271—282. Gibbons, M. C., R. Mann & L. D. Wright. 1989. Laboratory and field Location Naseway Shoal Wreck Shoal Rock Wharf Pt. of Shoals Naseway Shoal Wreck Shoal Rock Wharf Pt. of Shoals Naseway Shoal Wreck Shoal Rock Wharf Pt. of Shoals $+ Oo nN ~ lalla b JIL ye mMmm oo ne Shale % 20.1 18.7 11.6 24.6 27.6 26.9 9.4 16.3 1.7 7.9 Qe) 2.6 Substrate Tire Shell Spat % Spat % 77 17.8 269 62.1 194 26.3 406 55.0 187 23.6 514 64.8 32 17.1 109 58.3 23 23.5 48 48.9 34 43.6 23 29.5 44 34.6 71 55.9 22 25.6 50 58.1 45 9.4 426 88.9 11 17.5 47 74.6 47 10.4 393 86.9 12 15.4 64 82.0 ACKNOWLEDGMENTS This work was supported by a grant from the Virginia Marine Resources Commission. Thanks to D. Eggleston and M. Lynch for reviewing an earlier version of the manu- script. REFERENCES studies of oyster larvae settlement on three substrates, oyster shell, tire chips, and expanded shale, and the relative mobility of the three sub- strates. Marine Resources Report No. 89-3, Virginia Institute of Ma- rine Science, Gloucester Point, VA. Journal of Shellfish Research, Vol. 9, No. 1, 177—185, 1990. AGE AND GROWTH RATE DETERMINATIONS FOR THE ATLANTIC SURF CLAM, SPISULA SOLIDISSIMA (DILLWYN, 1817), IN PRINCE EDWARD ISLAND, CANADA THOMAS W. SEPHTON AND CLAIR F. BRYAN Canada Department of Fisheries & Oceans Science Branch, Gulf Region P.O. Box 5030 Moncton, New Brunswick E1C 9B6 ABSTRACT The age and growth rates of surf clams, Spisula solidissima (Dillwyn) collected from the south (Northumberland Strait), north (Gulf of St. Lawrence) and east (Cardigan Bay) shores of Prince Edward Island, Canada, were determined by analysis of thin sections of the shell chondrophore. Growth increment | included both a broad, diffused band formed during the spawning season (late July to October) and a narrow distinct band formed during the winter, which was used as an age marker. The site specific nature of the growth curves may be the result of differences in water temperature. The divergence of the curves at the age of sexual maturation (4 years) suggested that temperature differentially stimulated somatogenesis and gametogenesis among study sites. The higher growth rate of surf clams from Cardigan Bay compared with those from more southern latitudes may be due to the absence of high water temperatures during the summer. Compared with the growth curve estimates presented here for the southern Gulf of St. Lawrence, and based on the thin section technique, previous growth curve estimates based on external shell growth marks overesti- mate age and underestimate growth rates of young clams, while underestimating age and overestimating growth rates of older clams. KEY WORDS: growth rate, aging, chondrophore thin sections, surf clam, Spisula solidissima, sexual maturation INTRODUCTION Knowledge of population age distributions and growth rates is essential for the application of resource yield models used in fisheries management (Beverton and Holt 1957, Ricker 1975). There are few published studies on the age and growth rates of surf (or bar) clams, Spisula solidis- sima (Dillwyn) in the southern Gulf of St. Lawrence (Kers- will 1944, Caddy and Billard 1976, Robert 1981) and all have estimated age from the examination of annual growth rings (or checks) on the exterior of the shell. Jones et al. (1978) concluded, however, that this method was mis- leading because it overestimated the age of young clams, due to the presence of disturbance rings (false growth rings), and underestimated the age of older clams because of the erosion of early rings on the umbo and crowding of later rings at the ventral margins. Ropes (1980) reviewed the different methods used to determine the age of surf clams and concluded that the only accurate and reliable method was to examine the internal growth lines of a pol- ished section of the whole shell (Jones et al. 1978, Am- brose et al. 1980, Jones 1980, Jones et al. 1983) or a pol- ished thin section of the shell chondrophore (Ropes and O’Brien 1979). The objectives of the present study were to determine the age of surf clams collected from locations on the south (Northumberland Strait), north (Gulf of St. Lawrence) and east (Cardigan Bay) shores of Prince Edward Island (PEI) using the chondrophore technique of Ropes and O’Brien (1979) and to estimate their growth rates. A field experi- ment was conducted to examine the timing of growth incre- ment formation for clams from Northumberland Strait. The age of sexual maturity was estimated from the shells of specimens used previously to examine the reproductive cycle of surf clams in PEI (Sephton 1987). MATERIALS AND METHODS Age Determinations Sampling sites were located on the eastern (Cardigan Bay) (latitude 46°10’ N), southern (Northumberland Strait) (latitude 46°09’ N) and northern (Gulf of St. Lawrence) (latitude 46°30’ N) shores of PEI as shown in Fig. 1. All areas have been commercially fished to varying degrees (Sephton and Bryan 1987). Sediment samples were taken at each location with a box sampler and substratum composi- tion was determined using the dry sieve method of Akagi and Wildish (1975). Surf clams were collected with a towed hydraulic clam dredge from the near shore areas at depths of water ranging from 1.75 to 7.75 m in 1984. The dredge had a fishing width of 68 cm, sampled to a depth of 25 cm and retained clams longer than 30 mm shell length. Clams representative of most size classes (range: 40—180 mm) were selected from each study site for further analysis and maximum anterior-posterior length (to the nearest mm) was measured with vernier callipers. Thin sections were made from the shell chondrophores using the technique of Ropes and O’Brien (1979). Briefly, the right hand valve was secured to the manipulative arm of an Isomet low speed geological saw by gluing a wooden dowel to the shell. A 3 mm section was cut through the chondrophore using a pair of diamond wafer saw blades, with one cut passing just anterior to the umbo. The umbo side of the section was hand polished and glued (5 min epoxy) to a glass slide. The section was re-cut with a single blade to produce a thin section (0.1 mm) of chondrophore 177 178 Gulf of St. Lawrence Foy Pan, ‘4 CE é MOIS / (2) NEW BRUNSWICK Gulf of St. Lawrence Borden Northumberland s Strait % %, @) Study Site Figure 1. Location of the Gulf of St. Lawrence, Northumberland Strait and Cardigan Bay study sites in Prince Edward Island, Canada. 30 km attached to the slide. The thin section was then hand pol- ished and examined under a compound microscope at 250 x. The growth intervals (GI) or bands, as described by Ropes and O’Brien (1979) and Jones (1980), were enumer- ated. GI 1 is composed of cris-crossed lamellar structures while GI 2 is an irregular, complex, crossed lamellar, fi- brous and spherulitic prismatic structure (Jones 1980) (Fig. 2). GI 1 appears as a light increment in photomicrographs while GI 2 is dark coloured (Fig. 2). Field Experiment A field experiment was conducted in Hillsborough Bay on the Northumberland Strait shore of PEI to document the approximate date when growth bands were formed. The water depth at the site of the plot was 5m and the sediment, typical of surf clam habitat in PEI, was composed of fine and medium grain sand. In the autumn of 1985, all natu- rally occurring clams were removed from the plot and others collected from close proximity (size range: 40—130 mm) and the ventral margin of shells marked with a super- ficial cut made with a jeweller’s saw. Clams were returned the following day and placed at a density of 5 m~? on the experimental plots by divers. Those which did not rebury themselves within 12 hours were removed. After overwin- tering on the plot, a random sample of clams (25 clams), SEPHTON AND BRYAN ranging in size from 40 to 130 mm, was collected by divers 4 times per year at regular seasonal intervals and processed for chondrophore analysis. (Growth rate estimates from this mark-recapture experiment are not available at this time.) Statistical Analysis The parameters of the von Bertalanffy growth function (VBGF): eh = =) where; Lt = shell length (mm) at time t; Le = maximum asymptotic length; K = growth parameter and ty = time when Lt = 0, were derived for graphical presentation in the present study using nonlinear regression analysis (Mar- quardt algorithm) (SAS NLIN, SAS Institute 1985). Methods and programs for the statistical comparison of VBGF’s are available (Allen 1976, Misra 1986) but inter- pretation of the results is problematic because of the physi- ological derivation of the formula, the nonlinear nature of the model and the high degree of correlation among param- eters (Roff 1980, Moreau 1987). MacDonald and Thompson (1985) showed that analysis of covariance tech- niques could be used to compare growth profiles of scallops (Placopecten magellanicus) from Newfoundland. More re- cently, Chouinard and Mladenov (in press) compared the VBGF growth profiles of P. magellanicus from the southern Gulf of St. Lawrence using cubic polynomial re- gression analysis and the pairwise analysis of covariance technique of Rao (1973). This latter technique was used to compare the growth profiles of Spisula solidissima from the three study sites. The independent variable (age) of the cubic polynomial regression (Y = By + B,X! + BX? + 83X37 + e, where Y = shell length (mm) and X = age (yr)) was centred about the mean age of all clams used in the study (total average age = 10 yr) to reduce the correla- tions between odd and even powers to zero (p 667, Sokal and Rohlf 1981). It was assumed that the measurement error was small and constant and had limited effect on the regression analysis. Pair-wise comparisons are made of the growth profiles by testing the equality of the regression co- efficients of the site specific data and using the combined data sets to estimate the common regression and residual sum of squares. Rao (p 281—283, 1973) stated that the sig- nificance of the ratio of mean squares due to deviation from hypothesis to residual due to separate regressions 1s tested and the results are summarized as analysis of variance tables. The F test statistic compared to F table values of a = 0.05 and df = (4,n — 8). Age of Sexual Maturity The age of sexual maturity was estimated from spec- imens used in the study of the reproductive cycle of bar clams (Sephton 1987), and from additional samples of smaller, and presumably younger, clams which were not included in the earlier study. Ages were estimated from GROWTH RATE OF SURF CLAMS IN PEI, CANADA 179 WINTER Gi Git SUMMER AUTUMN GI2 Gi SUMMER Figure 5 WINTER Git Gl2 SPRING Figure 2. Photomicrograph of a thin section of the shell chondrophore of the surf clam, Spisula solidissima, showing GI 1 winter growth interval followed by the formation of the pre-spawning growth interval (GI) 2 during the spring and early summer. Also shown is the initial formation of GI 1 associated with spawning. Specimen was collected in July. Figure 3. Photomicrograph of a thin section of the shell chondrophore of the surf clam, Spisula solidissima, showing the completed formation of GI 1, which occurred during the late summer during the spawning period, and the recommencement of GI 2. Specimen was collected in September. Figure 4. Photomicrograph of a thin section of the shell chondrophore of the surf clam, Spisula solidissima, showing the formation of the post spawning GI 2 during autumn and prior to winter. Specimen was collected in December. Figure 5. Photomicrograph of a thin section of the shell chondrophore of the surf clam, Spisula solidissima, showing the formation of age marker used in the present study, GI 1, during the winter. GI 2 formed in the autumn and the spring. Specimen was collected in May. thin sections of the chondrophores using the technique de- scribed above. The sex and state of gonad activity were assessed from histological gonad preparations as described by Sephton (1987). RESULTS Formation of Growth Interval The alternating pairs of growth increments were distinct and appeared as light and dark bands in the chondrophore (Fig. 2—5). GI 2 formed a broad dark band during the spring and early summer prior to spawning (Fig. 2). A broad light band, GI 1, interspersed with narrow opaque bands (Fig. 3), formed during the spawning season (late July to October) (Sephton 1987). The narrow opaque bands within the GI 1 were darker, resembling GI 2. GI 2 com- menced after spawning and continued until late December as water temperatures fell to 0°C (Fig. 4). Over the winter, GI 1 formed a narrow distinct band consisting of an accu- mulation of growth lines (Fig. 5). Gl 2 resumed in mid March with the spring thaw (Fig. 5). The distinct, thin winter GI 1 between the GI 2 bands was used as the age marker in the present study. Ageing and Growth Rates The non-linear estimated parameters of the VBGF, the growth profiles and the age/length data of surf clams from the three study sites are shown in Fig. 6. Fewer animals were sampled from Cardigan Bay (n = 47) (Northumber- 180 SEPHTON AND BRYAN 160 140 120 = o0 oO © oOo Oo © Cardigan Bay Shell Length (mm) O Gulf of St. Lawrence 0 10 oe +0 0-0) A Northumberland Strait 141.0:5.97 G4 + O-¢-0-4 + $44-4 Oo = Oo L=+95%C!l K+95%Cl t#95%Cl iis n 166.510.03 0.250.07 0.70.25 0.99 47 0.27%0.06 0.040.19 099 118 142,124.12 0.240.04 0.14015 0.99 108 20 30 40 Age (yr) Figure 6. Summary of the growth profiles, the non-linear estimated parameters of the von Bertalanffy growth function and the age/length data of surf clams, Spisula solidissima, for the Cardigan Bay, Northumberland Strait and Gulf of St. Lawrence study sites in PEI, Canada. The 95% confidence limits for Lx, K, ty and the coefficient of determination (r?) are also shown. land Strait n = 118, Gulf St. Lawrence n = 108) but there was a similar distribution of sizes and ages as in the other samples for the analysis. The average age (and range) for Cardigan Bay, Northumberland Strait and Gulf of St. Lawrence was 8.4 years (2—23 yrs), 7.9 years (2—25 yrs), and, 13.5 years (2—37 yrs), respectively. Fig. 6 shows that growth was very rapid until age 5—7 years for all study sites and that growth slowed and was negligible after age 15 years. The growth profile for Cardigan Bay differed from the other two areas, diverging at age 5 years with consistently greater size at age up to the L~ of 165 mm. The Gulf of St. Lawrence and Northumberland Strait growth profiles were similar although clams from North- umberland Strait were somewhat larger at age than those from the Gulf shore. Clams from these two areas grew to a L of about 140 mm. The results of the analysis of vari- ance comparing (pair-wise) the cubic polynomial regres- sions of the growth profiles are summarized in Tables | and 2 and showed that all of the regressions were significantly different (p < 0.001) from one another. This result was not as evident for the Northumberland Strait and Gulf of St. Lawrence profiles shown in Fig. 6 as it was for Cardigan Bay. Age of Sexual Maturity A re-examination of the histological preparations of gonad material from a study of reproduction (Sephton 1987) and an examination of juvenile surf clams, showed that sex determination is possible with clams from all loca- tions as young as 4 years. A 4 year old clam ranges in length from about 80 mm (Gulf St. Lawrence) to 95 mm (Cardigan Bay). There was some cellular differentiation in the gonadal area of 3 year olds but sexes were not distin- guishable. Environmental Data Although long-term temperature records were not avail- able for the study areas, data for two of the regions were available from the literature. Data representative of the Northumberland Strait are surface water temperature data from 1983 to 1987 for Borden, PEI, located in the central part of the Northumberland Strait and 50 km east of the GROWTH RATE OF SURF CLAMS IN PEI, CANADA 181 TABLE 1. Summary of the individual and combined site estimates of the regression coefficients of the cubic polynomial regression (Y = BO + BIX + B2X? + B3X°, where Y = shell length (mm) and X = age (yr)) used to compare the growth profiles of surf clams from the three study sites using the method of Rao (1973) as described in the text. The standard errors (SE) of the estimates, coefficient of determination (r2), Durbin-Watson D statistic (DW) and the number of observations (n) are also shown. Regression n pO SE Bl SE B2 SE B3 SE im DW Northumberland Strait 118 131.75 1.3819 2.8615 0.4465 — 0.4812 0.0511 0.0262 0.0054 0.7562 1.381 Gulf of St. Lawrence 108 125.25 2.1722 4.9149 0.2716 — 0.3997 0.0458 0.0092 0.0017 0.8000 1.055 Cardigan Bay 47 153.13 2.5348 3.9088 0.8151 — 0.7544 0.0922 0.0428 0.0119 0.8421 1.988 Gulf + Northumb. 226 130.05 1.3096 4.4282 0.1924 —0.4017 0.0309 0.0099 0.0013 0.7462 1.151 Northumb. + Cardigan 165 138.06 1.4653 3.6575 0.4577 —0.5177 0.0533 0.0251 0.0059 0.7239 1.082 Gulf + Cardigan 155 136.78 2.119 5.0731 0.2837 —0.5203 0.0474 0.0131 0.0018 0.7066 0.955 Hillsborough Bay study site, from Dobson and Petrie (1984, 1985), Walker et al. (1986, 1987) and Gregory et al. (1988). Data representative of Cardigan Bay are surface water temperature data from 1983 to 1987 for St. Mary’s Bay, PEI, located 10 km from the study site and within the Cardigan Bay estuary, from Drinkwater and Petrie (1988). Drinkwater and Petrie (1988) also reviewed the physical oceanography of the Cardigan Bay area and reported that there was no evidence of thermal stratification because of wind and tidal mixing of the water. Fig. 7 shows the surface water temperature from May to November, 1983 to 1987, for the locations representative of the Northumberland Strait and Cardigan Bay study sites (Borden and St. Mary’s Bay, respectively). In most cases, surface water in Cardigan Bay warmed more quickly in the spring and cooled more quickly in the autumn, than in Northumberland Strait. The annual summer peak in tem- perature usually occurred earlier and was an average of 1.3°C (SD = 1.1°C) higher in Cardigan Bay than in North- umberland Strait. In both locations winter freeze-up occurs by late December or early January and both are ice free by April. Dry sieve analysis of sediment samples from each of the study sites showed that they were very similar. All sedi- ments were composed of fine and medium grained sand (99.6% —99.9% sand by weight) with a median particle size range from 0.17 mm (Northumberland Strait) to 0.34 mm (Cardigan Bay). DISCUSSION The correct ageing of bivalves is dependent upon knowl- edge of the periodicity of growth ring formation (Rhoads and Lutz 1980). The results of the present study show that the winter GI 1 ring formed in Spisula solidissima from the southern Gulf of St. Lawrence is a good age marker and is consistent with that observed by others using observations over longer time periods (Jones et al. 1978, Jones 1980, Jones et al. 1983). The use of winter growth rings (GI 1) TABLE 2. Summary of the results of analysis of covariance comparing the cubic polynomial regressions of the growth profiles of Spisula solidissima from the three study areas in Prince Edward Island. The pair-wise analyses are based on the technique of Rao (1973) as explained in the text. Comparison of Gulf of St. Lawrence & Northumberland Strait df SS MS F Significance Deviation 4 $904.21 1476.05 11.18 p < 0.0001 Separate Regressions 218 28771.95 131.98 Common Regression 222 34676.16 Comparison of Northumberland Strait & Cardigan Bay df SS MS F Significance Deviation 4 2012.0 503.0 4.26 p < 0.001 Separate Regressions 157 17722.16 112.88 Common Regression 161 19734.16 Comparison of Gulf of St. Lawrence & Cardigan Bay df SS MS F Significance Deviation 4 6852.03 1713.01 10.97 p < 0.0001 Separate Regressions 147 22956.53 156.17 Common Regression 151 29808 .56 182 SEPHTON AND BRYAN Se NS) © NS) aS (7p) feos) (Sf) Surface Water Temperature (°C) Oo NM 1983 1984 @——#_ Borden, PEI Northumberland St. +---++ Cardigan Bay MJJASON MJJASON MJJASON MJJASON MJJASON 1985 1986 1987 Figure 7. Summary of the surface water temperature data from 1983 to 1987 for Borden, PEI, located in the central part of the Northumber- land Strait and 50 km east of the Hillsborough Bay study site, (data from Dobson and Petrie (1984, 1985), Walker et al. (1986, 1987) and Gregory et al. (1988), respectively) and for St. Mary’s Bay, located 10 km from the Cardigan Bay study site and within the Cardigan Bay estuary (data from Drinkwater and Petrie (1988)). for ageing and their formation in response to cold tempera- tures is well documented for many species of bivalves (Ste- venson and Dickie 1954, Barker 1964, Feder and Paul 1974, Kennish and Olsson 1975), including S. solidissima as reported by Jones et al. (1983) and further substantiated by the present study. However, the conflicting data of Jones et al. (1978), Jones (1980) and Jones et al. (1983) show that the time of GI 1 formation in surf clams within their main geographic range (offshore areas from Cape Hatteras to Cape Cod) is variable and dependent upon the natural climatic processes of ocean warming and cooling. Mercenaria mercenaria also shows variability in the timing of growth ring formation; quahaugs from different areas of their geographic distribution form annual winter marks over part of their range and summer marks elsewhere (Kennish and Olsson 1975). Jones et al. (1983) retracted an original conclusion (Jones et al. 1978, Jones 1980) that GI 1 was formed during and associated with the spawning season of surf clams because of conclusive evidence obtained from longer term data based on the analyses of oxygen and carbon stable isotopes. Interestingly, the results of the present study substantiated Jones’s original conclusions that GI 1 is formed during the period of spawning but that it is different than the winter GI 1 age marker. This vari- ability in Spisula solidissima, in that the ring can form be- tween autumn and late winter, remains a concern and a source of error when attempting to compare and interpret growth profiles from throughout the geographic range of this species. The significant difference (p < 0.001) among the growth profiles is apparent in Fig. 6 as all populations di- verge at age 4 years, with clams from Cardigan Bay growing to the largest size. The similarity of the bottom sediments, the depth of water and the low densities (0.1 to 1.2 clams m~?) (Sephton and Bryan 1985) suggest that these factors are unlikely to have a significant influence on the growth rates observed. Growth rates of bivalves are in- fluenced to varying degrees by the interaction of abiotic and biotic environmental factors such as: sediment type, water depth, population density and food availability (Pratt and Campbell 1956, Loosanoff 1958, Galtsoff 1964, Am- brose et al. 1980, Rhoads and Lutz 1980, Ropes 1980, Bri- GROWTH RATE OF SURF CLAMS IN PEI, CANADA celj and Malouf 1984, Fréchette and Bourget 1985, Héral et al. 1987). A prominent factor for many species, in- cluding Spisula solidissima, is water temperature (Belding 1910, Ambrose et al. 1980, Jones 1981, Jones et al. 1983). Ambrose et al. (1980) showed that the growth rate of surf clams from near shore (<5 km) areas of the mid Atlantic Bight was slower than that of offshore (>5 km) areas be- cause of the cooler minimum temperatures of near shore water. We, however, hypothesize that mid-summer high temperatures are probably a principal controlling factor of growth rates in the present study and, possibly, in other geographic areas. Information concerning the distribution of Spisula solidissima around PEI indicates that all popula- tions are influenced by similar cold water temperatures as- sociated with the winter ice cover (Sephton and Bryan 1985). The fastest and largest growing surf clams in Car- digan Bay (Fig. 6) are usually exposed to warmer water (Fig. 7) during the early part of growing season than those found in the Northumberland Strait, and presumably the Gulf shore of PEI. It appears that summer temperatures of Cardigan Bay may not have the same negative effect on growth as hypothesized by Menesguen and Dreves (1987) for the mid Atlantic Bight. Jones (1981) found that the greatest mean annual growth of S. solidissima in near shore and offshore study sites was strongly correlated with lower annual temperatures and this implied to Menesguen and Dreves (1987) that the summer temperatures were too warm for this species and had a negative effect on growth. As mentioned above, growth is influenced by an interaction of the various factors and populations at different latitudes and water depths may not be affected to the same degree by the same factors. The divergence of the growth profiles at about the age of sexual maturation (Fig. 7) may be due to the effect of tem- perature on the differential stimulation of somatogenesis and gametogenesis (Lubet 1976, Menesguen and Dreves 1987). Menesguen and Dreves (1987) contend that the in- terference of somatic growth by gametogenesis is related to a threshold temperature stimulation and may explain the temperature and growth anomalies of the three bivalve species (which included Spisula ovalis) in their study. This also suggests that the timing and duration of reproduction may be different in Cardigan Bay than that observed pre- viously for Northumberland Strait and Gulf of St. Lawrence shores of PEI (Sephton 1987). The later matura- tion of surf clams in PEI (age 4 years, size range 80—95 mm), compared with other estimates of 2 years (Ropes 1979b) suggests that young surf clams can enter the com- mercial fishery before they have had the opportunity of contributing to the natural reproductive and recruitment processes of the population. The present legal size limit for commercial fishing of surf clams in PEI is 76 mm (3 inches) which corresponds to an age of 3 years in all loca- tions. It may be a prudent conservation measure to increase the commercial fishing size in the southern Gulf of St. 183 Lawrence to a size equal to or greater than the size range at sexual maturation until such time as other biological infor- mation (fecundity, recruitment and population dynamics) become available. A comparison of our growth curves with those selected from the literature is shown in Fig. 8. Growth curve | is from Ropes (1979a) and integrates data collected throughout the mid Atlantic Bight into a generalized growth curve. Curves 2 and 3 are from Jones (1980) for the near shore and offshore areas adjacent to Pt. Pleasant, New Jersey, respectively, in the mid Atlantic Bight. Curves 1—3 were determined from the examination of internal growth increments. Curves 4 and 5 are for areas in Northumber- land Strait west of our study site; Buctouche, N.B. (Caddy and Billard 1976) and Mt. Carmel, PEI (Robert 1981), re- spectively. These growth curves were determined from the 1 4 Ropes 1979, General USA A 2 Oo Jones 1980, Offshore # 3 A Jones 1980, Nearshore 60 4 * Caddy & Billard 1976 5 + Robert 1981 40 6 # Present Study, Cardigan Bay 7 A Present Study, Northumberland St. 20 8 H Present Study, Gulf St Lawrence 6558 10 12 14 16 18 20 Age (yr) Figure 8. Comparison of growth curves of surf clams, Spisula solidis- sima, from the present study (curves 6 to 8) with those selected from the literature. Curve 1: Ropes (1979a), generalized curve for the mid Atlantic Bight, based on internal growth indicators (IGI); Curves 2 and 3: Jones (1980), near shore and offshore areas adjacent to Pt. Pleasant, New Jersey, respectively, IGI; Curve 4: Caddy and Billard (1976), Northumberland Strait at Buctouche, New Brunswick, ex- ternal growth indicators (EGI); Curve 5: Robert (1981), Northum- berland Strait at Mt. Carmel, PEI (EGI). 184 examination of external shell growth increments. Curves 6, 7 and 8 represent Cardigan Bay, Northumberland Strait and Gulf of St. Lawrence areas, respectively, of the present study. The Northumberland Strait growth profiles (curves 4 and 5) are very similar. Clams grew to a final size that is intermediate between the Cardigan Bay and the Northum- berland Strait locations in the present study. These growth profiles estimated a smaller length at age until age 7 years from those observed in the present study (curves 6—8) (Fig. 8) and then estimated a larger length at age. The differences among the growth profiles (4 and 5 with 6, 7 and 8), is likely associated with the use of external growth marks on the shell to determine age and growth rates. Jones et al. (1978) described this phenomena from the comparison of growth curves from mid Atlantic Bight surf clams and showed that curves determined from external shell markings overestimated age and underestimated growth rates in the first few years of life (3 to 6 years) while un- derestimating age and overestimating growth rates in later years. Although the largest size attained by surf clams ap- pears variable in Northumberland Strait (Fig. 8), we con- tend that our growth profiles are a better estimate of the growth of surf clams in the southern Gulf of St. Lawrence than those reported previously. The offshore growth curve of Jones (1980) (curve 2) (Fig. 8) and that for Cardigan Bay (curve 6) and Northum- berland Strait (curve 7) of the present study are remarkably similar for the first 3 to 4 years. Similarly, Jones’s (1980) near shore growth profile (curve 3) and that for the Gulf of St. Lawrence (curve 8) follow the same trajectory in early life. The dissimilarity of these growth profiles from curve | (Fig. 8) for surf clams in the mid Atlantic Bight suggests that this general curve, which incorporates data from a va- SEPHTON AND BRYAN riety of sources (Ropes 1979a), is not representative of the site specific growth which has been observed. Upon diverging, the surf clams from the Northumber- land Strait and Gulf of St. Lawrence grow to a size that is intermediate between Jones’s (1980) near shore and off- shore area populations while those from Cardigan Bay grow to a size larger than these areas but smaller than that projected for the overall mid Atlantic Bight (curve 1, Fig. 8). This suggests that the conditions for growth in Cardigan Bay may be superior to those in the near shore and offshore areas of Pt. Pleasant, New Jersey (Jones 1980). This may be due to the absence of high summer water temperatures, which may inhibit growth (Menesguen and Dreves 1987) and greater food abundance, as a consequence of living in a shallow coastal embayment (Ambrose et al. 1980). The area specific differences in growth observed in the present study for surf clams within the same general geographic zone (southern Gulf of St. Lawrence), as well as those in the literature, have implications for bivalve fisheries man- agement and regulations. It is rare to find management reg- ulations that are site specific, other than large groupings of smaller areas (example: Georges Bank vs Gulf of Maine) and our results demonstrate that valuable information would be lost on the site specificity of age and growth rela- tionships if they were pooled over the entire geographic zone. ACKNOWLEDGEMENTS This paper is dedicated to the memory of Mr. John W. Ropes, whose passion for biology influenced all those who knew him. We gratefully acknowledge the technical and field assistance provided by the following: Jody Gamouf, Cheryl Coulson, Paul Baker and Rick MacKenzie. We thank T. Rowell, S. McGladdery and D. Sephton for criti- cally reviewing early drafts of the manuscript. REFERENCES Akagi, H. & D. J. Wildish. 1975. Determination of the sorting character- istics and organic carbon content of estuarine sediments. Fish. Res. Bd. Can. Manuscr. Rep. 1370:15 pp. Allen, R. L. 1976. Method for comparing fish growth curves. N.Z. J. Mar. Freshw. Res. 10:687—692. Ambrose, W. G. Jr., D. S. Jones & I. Thompson. 1980. Distance from shore and growth rate of the suspension feeding bivalve, Spisula soli- dissima. Proc. Natl. Shellfish. Assoc. 70:207—215. Barker, R. M. 1964. Microtextural variation in the pelecypod shell. Ma- lacologia 2:69-86. Belding, D. L. 1910. The growth and habits of the sea clam (Mactra solidissima). Rep. Comm. Fish. Game 1909, Comm. Mass. Public Doc. 25:26-41. Beverton, R. J. H. & S. J. Holt. 1957. On the dynamics of exploited fish populations. Fish. Invest. Minist. Agric. Fish. Food (G.B.), Ser. II, Salmon Freshw. Fish. 19:533 pp. Bricelj, V. M. & R. E. Malouf. 1984. Influence of algal and suspended sediment concentrations on the feeding physiology of the hard clam Mercenaria mercenaria. Mar. Biol. 84:155—165. Caddy, J. F. & A. R. Billard. 1976. A first estimate of production from an exploited population of the bar clam Spisula solidissima. Fish. Mar. Ser. Tech. Rep. 648:13 pp. Chouinard, G. A. & P. V. Mladenov. 1990. Comparative growth and morphology of the sea scallop (Placopecten magellanicus Gmelin) in the southern Gulf of St. Lawrence. Can. Bull. Fish. Aquat. Sci. (in press). Dobson, D. & B. Petrie. 1984. Long-term temperature monitoring pro- gram 1983: Scotia-Fundy, Gulf Regions. Can. Data Rep. Hydrogr. Ocean Sci. 22:406 pp. Dobson, D. & B. Petrie. 1985. Long-term temperature monitoring pro- gram 1984: Scotia-Fundy, Gulf Regions. Can. Data Rep. Hydrogr. Ocean Sci. 35:691 pp. Drinkwater, K. & B. Petrie. 1988. Physical oceanographic observations in the Cardigan Bay region of Prince Edward Island, 1982-1987. Can. Tech. Rep. Hydrogr. Ocean Sci. 110:37 pp. Feder, H. M. & A. J. Paul. 1974. Age, growth, and size-weight relation- ships of the soft shell clam, Mya arenaria, in Prince William Sound, Alaska. Proc. Natl. Shellfish. Assoc. 64:45—52. Fréchette, M. & E. E. Bourget. 1985. Food-limited growth of Mytilus edulis L. in relation to the benthic boundary layer. Can. J. Fish. Aquat. Sci. 42:1166—1170. Galtsoff, P. S. 1964. The American oyster Crassostrea virginica Gmelin. Fish. Bull. U.S. Fish. Wildl. Serv. 64:480 pp. Gregory, D. N., E. Verge, D. Dobson & C. Smith. 1988. Long-term GROWTH RATE OF SURF CLAMS IN PEI, CANADA temperature monitoring program 1987: Scotia-Fundy, Gulf of St. Lawrence and Newfoundland. Can. Data Rep. Hydrogr. Ocean Sci. 65:497 pp. Héral, M., J. M. Deslous-Paoli, J. Prou & D. Razet. 1987. Relation entre la nourriture disponible et la production de mollusque en milieu es- tuarien: variabilité temporelle de la colonne d'eau. Haliotis 16:149— 190. Jones, D. S. 1980. Annual cycle of shell growth increment formation in two continental shelf bivalves and its paleoecologic significance. Pa- leobiol. 6:331—340. Jones, D. S. 1981. Annual growth increments in shells of Spisula solidis- sima record marine temperature variability. Science 211:165—167. Jones, D. S., I. Thompson & W. Ambrose. 1978. Age and growth rate determinations for the Atlantic surf clam, Spisula solidissima (Bi- valvia: Mactracea), based on internal growth lines in shell cross-sec- tions. Mar. Biol. 47:63—70. Jones, D. S., D. F. Williams & M. A. Arthur. 1983. Growth history and ecology of the Atlantic surf clam, Spisula solidissima (Dillwyn), as revealed by stable isotopes and annual shell increments. J. Exp. Mar. Biol. Ecol. 73:225—242. Kennish, M. J. & R. K. Olsson. 1975. Effects of thermal discharges on the microstructural growth of Mercenaria mercenaria. Environ. Geol. 1:41-64. Kerswill, C. J. 1944. The growth rate of bar clams. Fish. Res. Bd. Can. Prog. Rep., Atl. Coast Sta. 35:18—20. Loosanoff, V. G. 1958. Some aspects of behaviour of oysters at different temperatures. Biol. Bull. 114:57—70. Lubet, P. 1976. Ecophysiologie de la reproduction chez les mollusques lamellibranches. Haliotis 7:49—55. MacDonald, B. A. & R. J. Thompson. 1985. Influence of temperature and food availability on the ecological energetics of the giant scallop Placopecten magellanicus. 1. Growth rates of shell and somatic tissue. Mar. Ecol. Prog. Ser. 25:279—294. Menesguen, A. & L. Dreves. 1987. Sea-temperature anomalies and popu- lation dynamic variations: effects on growth and density of three bi- valves. Mar. Ecol. Prog. Ser. 36:11—21. Misra, R. H. 1986. Fitting and comparing several growth curves of the generalized von Bertalanffy type. Can. J. Fish. Aquat. Sci. 43:1656— 1659. Moreau, J. 1987. Mathematical and biological expression of growth in fish: recent trends and further developments. /n: R. C. Sumerfelt and G. E. Hall (eds.). The age and growth of fish. lowa State Univ. Press, Ames, p 81-113. Pratt, D. M. & D. C. Campbell. 1956. Environmental factors affecting growth in Venus mercenaria. Limnol. Oceanogr. 1:2-17. 185 Rao, C. R. 1973. Linear statistical inference and its applications. Second edition. John Wiley and Sons, New York, 625 pp. Rhoads, D.C. & R. A. Lutz. 1980. Skeletal growth of aquatic or- ganisms: biological records of environmental change. Plenum, New York, 750 pp. Ricker, W. E. 1975. Computation and interpretation of biological sta- tistics of fish populations. Bull. Fish. Res. Bd. Can. 191:382 pp. Robert, G. 1981. Dynamics of an unexploited population of bar clam, Spisula solidissima. Can. Manuscr. Rep. Fish. Aquat. Sci. 1607. 12 PP. Roff, D. A. 1980. A motion for the retirement of the von Bertalanffy function. Can. J. Fish. Aquat. Sci. 37:127—129. Ropes, J. W. 1979a. Biology and distribution of surf clams (Spisula soli- dissima) and ocean quahaugs (Arctica islandica) off the Northeast coast of the United States. /n: Proccedings of Northeast Clam Indus- tries: management for the future. Exten. Sea Grant Prog., Univ. Mass. and Mass. Inst. Tech. SP-112:47—66. Ropes, J. W. 1979b. Shell length at sexual maturity of surf clams, Spisula solidissima, from an inshore habitat. Proc. Natl. Shellfish. Assoc. 69:85-91. Ropes, J. W. 1980. Biological and fisheries data on the Atlantic surf clam, Spisula solidissima (Dillwyn). North. Fish. Ctr. U.S. Natl. Mar. Fish. Serv. Tech. Rep. Ser. 24:88 pp. Ropes, J. W. & L. O’Brien. 1979. A unique method of aging surf clams. Bull. Am. Malacol. Union Inc., p 58-61. Sephton, T. W. 1987. The reproductive strategy of the Atlantic surf clam, Spisula solidissima, in Prince Edward Island, Canada. J. Shellfish Res. 6:97-102. Sephton, T. W. & C. F. Bryan. 1985. A preliminary assessment of the American bar/surf clam, Spisula solidissima, in Prince Edward Island, 1984. Can. Atlan. Fish. Sci. Adv. Comm. Res. Doc. 85/33. 19 p. Sephton, T. W. & C. F. Bryan. 1987. A survey of commercial catch rate data for the 1986 Prince Edward Island bar clam (Spisula solidissima) Fishery. Can. Atlan. Fish. Sci. Adv. Comm. Res. Doc. 87/32. 13 p. Sokal, R. R. & F. J. Rohlf. 1981. Biometry. Second edition. Freeman, San Francisco, 859 pp. Statistical Analysis Institute. 1985. SAS/STAT guide for personal com- puters. Cary, N.C., 1028 pp. Stevenson, J. A. & L. M. Dickie. 1954. Annual growth rings and rate of growth of the giant scallop, Placopecten magellanicus (Gmelin), in the Digby area of the Bay of Fundy. J. Fish. Res. Bd. Can. 11:660—671. Walker, R. E., D. Dobson & P. Stead. 1986. Long-term temperature monitoring program 1985: Scotia-Fundy, Gulf of St. Lawrence and Newfoundland. Can. Data Rep. Hydrogr. Ocean Sci. 49:521 pp. Walker, R. E., D. Dobson & P. Stead. 1987. Long-term temperature monitoring program 1986: Scotia-Fundy, Gulf of St. Lawrence and Newfoundland. Can. Data Rep. Hydrogr. Ocean Sci. 53:529 pp. - =: wo” _ aan, ar Journal of Shellfish Research, Vol. 9, No. 1, 187-193, 1990. CAGE CULTURE OF YEARLING SURF CLAMS, SPISULA SOLIDISSIMA (DILLWYN, 1817), IN COASTAL GEORGIA RONALD GOLDBERG! AND RANDAL L. WALKER? National Marine Fisheries Service Northeast Fisheries Center Milford Laboratory Milford, Connecticut 06460-6499 2Marine Extension Service Shellfish Research Laboratory University of Georgia P.O. Box 13687 Savannah, Georgia 31416-0687 ABSTRACT The surf clam, Spisula solidissima, found from Labrador to Cape Hatteras, North Carolina, was reared in coastal Georgia in different substrates to test the feasibility of mariculture beyond its natural geographic range. Spisula was reared from a seed size of 12—21 mm length to a marketable size of 50 mm during the months of October to May. Clams were reared in partially-buried cages to exclude macro-predators. Laboratory and field experiments indicated that different substrate types of mud, sand, or a mixture of sandy-mud had no effect on growth rate. Clams maintained in a mud substrate repeatedly forced themselves out of the sediment and did not remain burrowed. Nested ANOVA indicated significant variability in growth and survival of clams in different cages within several meters distance. Clams reared in cages within 3 km of the mouth of Wassaw Sound grew at a faster rate than those 8 and 1|1 km thalweg upstream along the Wilmington River. These results indicate large areas of coastal marsh are suitable for growth and establishes the biological feasibility of mariculture of this species in southern coastal waters beyond its natural range. KEY WORDS: Spisula solidissima, growth, survival, substrate, Georgia INTRODUCTION Mariculture of bivalves in the estuaries of the south- eastern United States has a potential to supplement dwin- dling natural fisheries (Walker, 1983). Increasing numbers of commercial hatcheries producing seed clams has pro- vided incentive to explore new possibilities for field grow- out. Hatchery reared surf clam seed, Spisula solidissima (Dillwyn), has been grown-out to a marketable 50 mm length during May to October in Connecticut (Goldberg, 1980). Consumer evaluation studies have determined young clams to be an acceptable steamed, raw, or fried product (Krzynowek and Wiggin, 1982). Recently, surf clams have been successfully raised and marketed commer- cially as a fresh seafood product (Monte, pers. comm.). Surf clams are well suited for culture because of their rapid growth and their excellent production potential. Well es- tablished hatchery and nursery methods can be employed to raise seed clams. Grow-out of seed clams to a marketable size has been accomplished by planting caged clams in the natural environment. Protection of seed clams from predators is the key to successful grow-out in the field. Surf clam seed, 12—21 mm in length, has been reared to 50 mm length with high survival (~85%) in partially buried cages in Long Island Sound, Ct (Goldberg, 1989). Cage culture offers an advan- tage of maintaining clams in a substrate, while excluding most macro-predators. Clams can be reared at high den- sities and their harvest is a simple process. Although Cape Hatteras, North Carolina has been de- scribed as the southern limit of the geographic range of surf clams (Merrill and Ropes, 1969), the climate and marine environment of coastal Georgia provide an opportunity for mariculture. In the summer, common inshore seawater temperatures in Georgia exceed 30°C, a level that has been shown in laboratory experiments (Savage, 1976), to be lethal to the surf clam. Seawater temperatures during fall to spring, however, are suitable for active growth of this species. Winter grow-out in the south and summer grow- out in the north might enable production of 2 annual crops. In natural populations, surf clams inhabit medium to coarse sand to gravel bottoms (Yancey and Welch, 1968; Fay et al., 1983). Evidence has been presented that the hard clam, Mercenaria mercenaria, grows more slowly in mud substrates (Pratt, 1953, Pratt and Campbell, 1956). Am- brose et al. (1980), indicate a positive partial correlation between growth of Spisula, and mean sediment grain size when variables of temperature, distance from shore, and depth were controlled in regression analysis. Most impor- tant for mariculture, the estuarine marsh systems of coastal Georgia offer a wide selection of naturally occurring sub- strate types ranging from fine silt to coarse sand. The goal of this study is to determine the biological fea- sibility of growing caged surf clam seed to market size be- tween October and May in coastal Georgia. Laboratory and 187 188 field studies were designed to investigate the effects of sub- strate type on growth and survival. Surf clams were also reared in different locations within the estuarine system to assess growth variability along the gradient between ocean and land. METHODOLOGY To determine the effects of substrate type on growth and survival of the surf clam within the laboratory, groups of 62 clams (8.6 mm mean length), spawned and reared at the National Marine Fisheries Service, Milford Laboratory, were maintained in 0.3 m? trays containing either mud (particle size < 64 2), sand (64 p to 2 mm), or an equal mixture of sand and mud. Three trays (n = 3) of each treatment were randomly placed in a large 24002 tank receiving filtered seawater and cultured algae. To determine the effects of substrate type on growth and survival of clams in the field, duplicate cages (n = 2) were placed at 3 locations with different substrates within a | km? area in Wassaw Sound, Georgia from November 1984 to May 1985 (Figure 1). The cages were within 0.1 km of shore and located within 3—4 km thalweg distance from the pee Wilmington BSS slan 4 Skidaway { Institute ORS A wo = & Cc 2 D 1S € 2 GOLDBERG AND WALKER mouth of the Wilmington River. Each location had a pre- dominant substrate type of either mud (<64 w), sand (64. to 2 mm), or a sandy-mud mixture. Dimensions of the cages were 0.6m X 0.6m X 0.25 mand they were fabri- cated from vinyl-coated wire mesh with 8 mm xX 8 mm square openings. All cages were deployed at a depth of | m below mean low water and partially buried about 20 cm in the substrate. The seabed was partly excavated and the cages pushed down in to the sediment. The cage was then filled with natural sediment through the cage mesh to screen out large debris and possible predators. One-meter long PVC poles were pushed into the sediment at each cage corner down to the top of the cage and attached to the cage with cable ties. On November 19, 1984, 200 clams (21.6 mm mean shell length) were planted in each of the 6 cages. Clams were sampled bimonthly and the length and number of live clams were recorded. Two additional field experiments were conducted be- tween 1985 and 1987 to assess variability of growth and survival of caged clams within the Wassaw Sound estuary. A common substrate type of sandy-mud mixture was chosen for all sites based on the consistently high survival Kilometers Figure 1. Site locations for the growth and survival of surf clams, Spisula solidissima, in Wassaw Sound, Georgia. CULTURE OF SPISULA SOLIDISSIMA IN GEORGIA 189 TABLE 1. Growth and survival of surf clams, Spisula solidissima, planted in different substrates within the laboratory. November 21, 1985 Mean Shell Length January 10, 1986 Mean Shell Length Substrate mm + SE (n) mm + SE (n) Mud 8.6 + 0.2 (62) 14.2 + 0.2 (51) Mud 8.6 + 0.2 (62) 14.2 + 0.2 (49) Mud 8.6 + 0.2 (62) 13.5 + 0.3 (54) Sand 8.6 + 0.2 (62) 14.6 + 0.2 (59) Sand 8.6 + 0.2 (62) 1457 = 0220 (77) Sand 8.6 + 0.2 (62) 14.8 + 0.2 (57) Sandy-mud 8.6 + 0.2 (62) 14.7 + 0.2 (58) Sandy-mud 8.6 + 0.2 (62) 15.0 + 0.2 (59) Sandy-mud 8.6 + 0.2 (62) 14.7 + 0.2 (60) March 20, 1986 Mean Shell Length June 10, 1986 Mean Shell Length % mm + SE (n) mm + SE (n) Surv. 24.2 + .07 (23) 3156) = EI) 8.1 DAD Len) 29.2 + 1.2 (6) 9.7 22.8 + 0.4 (24) 29.0 + 0.8 (18) 29.0 24.6 + 0.3 (61) 32.4 + 0.5 (37) 59.7 24.4 + 0.3 (51) 30.7 + 0.5 (33) 53.2 24.5 + 0.3 (46) 32.2 + 0.5 (44) 71.0 25.6 + 0.3 (61) 31.4 + 0.6 (37) 59.7 23.5 + 0.4 (47) 31.1 + 0.6 (43) 69.4 2513) =10'31(52) 31.0 + 0.8 (38) 61.3 and growth observed in the previous experiments. Cages were fabricated and deployed as described. In 1985, 4 sites were chosen within a5 km x 5 km area near Cabbage Island in Wassaw Sound (Fig. 1). Only minor differences in temperatures and salinity were ex- pected among sites, although site 4 was located several km closer to the Atlantic Ocean than the other 3 sites. On No- vember 25, 1985, clams (n = 65) of a mean initial shell length of 12.4 mm, were placed in triplicate cages (n = 3) at each site. In 1987, we tested clam growth in cages (n = 2) at five sites differing in relative position in the estuary, since growth at site 4 during the 1985-1986 experiments ex- ceeded growth observed at all other sites. All sites were 0.1 km from shore. Sites 2, 3, 4, and 5 were 1,5, 8, and 11 km thalweg distance from the mouth of Wassaw Sound along the Wilmington River. Site 1 was 3 km thalweg distance from the mouth of Wassaw Sound along the Bull River. Sites | and 2 are more influenced by oceanic waters while the other sites are more estuarine. RESULTS Growth and survival data of surf clams placed in dif- ferent substrates within the laboratory are given in Table 1. ANOVA was performed on the final mean lengths of clams in each treatment (n = 3), which indicated no significant difference (P > 0.05). Means of final lengths and their standard errors for the mud, sand, and sandy-mud treat- ments were 29.9 mm + 0.84, 31.7 mm + 0.54, and 31.2 mm + 0.12, respectively. Cages maintained in the area with sand substrate in the field experiment in 1984—85, were found to have had the sand washed-out when they were sampled in March, prob- ably because of a storm event. Growth of these clams (Fig. 2) during the period between January and March was min- imal, most likely caused by this disturbance. These cages were reburied and rapid growth resumed during March to May. ANOVA revealed no difference (P > 0.05) in final mean lengths (Table 2) of clams in cages (n = 2) main- tained in different areas with either mud, sand, or sandy- mud substrate. Means and their standard errors for the final lengths of clams in mud, sand, or sandy-mud treatments were 44.1 mm + 1.00, 43.8 mm + 0.60, 48.6 mm + 2.40, respectively. Survival in all cages ranged between 65.0 and 79.0% with no significant difference between sites. Differences were observed in the behavior of clams in the substrate treatments. Clams in mud repeatedly forced themselves out of the sediment both in the laboratory and in the field. Lower survival (Table 1) of clams held in mud within the laboratory was not caused by the treatment. Clams in mud simply escaped their treatments and fell into TABLE 2. Growth and survival of surf clams, Spisula solidissima, planted in different substrates in Wassaw Sound, Georgia (1984-1985). November 19, 1984 January 19, 1985 March 19, 1985 May 6, 1985 Mean Shell Length Mean Shell Length Mean Shell Length Mean Shell Length % Substrate mm + SE (n) mm + SE (n) mm + SE (n) mm + SE (n) Surv Mud 21.6 + 0.2 (200) 33.4 + 0.3 (ND)* 39.2 + 0.5 (158) 45.2 + 0.4 (134) 67 Mud 21.6 + 0.2 (200) 33.4 + 0.4 (ND) 41.7 + 0.5 (159) 43.0 + 0.4 (145) 72 Sand 21.6 + 0.2 (200) 33.6 + 0.4 (ND) 36.5 + 0.3 (146) 44.4 + 0.3 (157) 78 Sand 21.6 + 0.2 (200) 33.3 + 0.3 (ND) 34.8 + 0.3 (161) 43.2 + 0.4 (143) 71 Sandy-mud 21.6 + 0.2 (200) 32.7 + 0.3 (ND) 45.0 + 0.5 (132) 46.2 + 0.4 (138) 69 Sandy-mud 21.6 + 0.2 (200) 33.2 + 0.4 (ND) 43.5 + 0.5 (141) 51.0 + 0.6 (129) 64 * Not Determined 190 GOLDBERG AND WALKER TABLE 3. Growth and survival of surf clams, Spisula solidissima, reared at four locations within Wassaw Sound, Georgia (1985-1986). Nov. 25, 26, 1985 Jan. 9, 1986 May 1, 1986 Mean Shell Length Mean Shell Length Mean Shell Length % Area (mm) + SE (n) (mm) + SE (n) (mm) + SE (n) Surv. Site 1 Cage 1 12.4 + 0.1 (65) 19.7 + 0.3 (ND)* 3527-1 0101@) IES Cage 2 12.4 + 0.1 (65) 15.8 + 0.4 (ND) 33.4 + 0.6 (26) 40.0 Cage 3 12.4 + 0.1 (65) 16.7 + 0.4 (ND) 38.0 + 0.3 (56) 86.2 Site 2 Cage | 12.4 + 0.1 (65) 20.1 + 0.3 (ND) 35.6 + 0.7 (32) 49.2 Cage 2 12.4 + 0.1 (65) 19.6 + 0.3 (ND) 34.2 + 0.6 (37) 56.9 Cage 3 2.4 + 0.1 (65) 20.2 + 0.3 (ND) 40.5 + 0.5 (51) 78.5 Site 3 Cage | + 0.1 (65) 18.8 + 0.3 (ND) 34.9 + 0.3 (15) 23.1 Cage 2 12.4 + 0.1 (65) 17.2 + 0.2 (ND) 30.4 + 0.5 (44) 67.7 Cage 3 12.4 + 0.1 (65) 18.7 + 0.3 (ND) 35.6 + 0.7 (28) 43.1 Site 4 Cage 1 4 + 0.1 (65) 21.2 + 0.3 (ND) 50.1 + 0.5 (55) 84.6 Cage 2 .4 + 0.1 (65) 20.7 + 0.4 (ND) 47.9 + 0.4 (50) 76.9 Cage 3 12.4 + 0.1 (65) 21.2 + 0.3 (ND) 47.2 + 0.5 (55) 84.6 * ND—Not Determined the outer holding tank. At the termination of the field ex- periment, clams held in cages in areas with mud substrate were found on the surface of the sediment within the cage. Data on growth and survival of clams planted at 4 dif- ferent locations in Wassaw Sound in sandy-mud substrate in 1985 are given in Table 3. By May 1986, significant differences in surf clam growth among sites occurred as determined by a nested ANOVA (Table 4). Mean lengths were significantly (P < 0.05) different among clams in cages within a location, as well as among different loca- tions. The results of the Scheffé multiple comparison test (P < 0.05) show that clam growth was the same at site 1, site 2, and site 3. Clams at site 4 averaged at least 12 mm TABLE 4. Final mean length measurements, nested ANOVA, and Scheffé results of surf clams, Spisula solidissima, reared at four locations within Wassaw Sound, Georgia (1985-1986). Number Mean Standard Site of Cages Length (mm) Error 1 3 35.50 1.34 2 3 36.76 1.90 3 3 33.63 1.63 4 3 48.40 0.87 NESTED ANOVA Source of Variation DF SS MS F 12 Cage within location 8 2086.82 260.8 21.69 0.0001* Location 3 17702.26 5900.7 22.60 0.0003* Scheffé Test* Site 3 = Site 1 = Site 2 < Site 4 *P<0.05 larger than those at other sites. Survival at all sites ranged from 2% to 86%. Growth and survival data of clams planted in 1987 within sandy-mud substrate are given in Table 5. Since one cage was lost and the other disturbed at site 3, these were removed from statistical treatment. The results of a nested ANOVA (P < 0.05) reveal that differences in clam growth occurred at different sites (Table 6). Since we were testing an a priori hypothesis that growth at the ocean sites would be greater than at the upriver sites, a T-test was then used to TABLE 5. Growth and survival of surf clams, Spisula solidissima, planted in sandy-mud substrate at five areas in Wassaw Sound, Georgia (1987). January 26, 1987 April 27, 1987 Mean Shell Length Mean Shell Length % Site mm + SE (n) mm + SE (n) Surv. 1 Cage | 21.4 + 0.3 (100) 35.8 + 0.5 (56) 56 Cage 2 21.4 + 0.3 (100) 35.6 + 0.5 (40) 40 2 Cage 1 21.4 + 0.3 (100) 36.6 + 0.4 (64) 64 Cage 2 21.4 + 0.3 (100) 32.1 + 0.4 (49) 49 3 Cage | 21.4 + 0.3 (100) 26.0 + 0.4 (51)* 51 Cage 2 21.4 + 0.3 (100) OF 0 4 Cage 1 1.4 + 0.3 (100) 30.0 + 0.4 (79) 79 Cage 2 21.4 + 0.3 (100) 30.2 + 0.4 (73) 73 5 Cage 1 21.4 + 0.3 (100) 31.8 + 0.4 (82) 58 Cage 2 21.4 + 0.3 (100) 31.9 + 0.3 (82) 58 * Cage had no sediment within it, but was held in place by stakes. ** Cage was washed away. CULTURE OF SPISULA SOLIDISSIMA IN GEORGIA 191 TABLE 6. Final mean length measurements and nested ANOVA of surf clams, Spisula solidissima, reared in four locations (one site eliminated, see text) within Wassaw Sound, Georgia (1987). Number Mean Standard Stie of Cages Length (mm) Error 1 2 35.70 0.10 2 2 34.35 2325 4 2 30.10 0.10 5 2 31.85 0.24 NESTED ANOVA Source of Variation DF SS MS F P Cage within location 4 478.32 119.58 74.02 0.0001* Location 3 2517.49 839.16 7.02 0.0452* *P < 0.05 compare the final mean lengths of clams in each cage at sites 1 and 2 to the final mean lengths of clams in each cage at sites 4 and S. This test indicated a significant difference (P < 0.05) in the final lengths of clams between ocean and upriver sites. DISCUSSION Experiments in the laboratory and the field have indi- cated no direct influence of sediment type on the growth rate of Spisula. Efforts by clams to escape and the lack of burrowing in mud substrate, suggest the unsuitability of this sediment type although, growth was not affected. Interpretation of growth and survival data of clams in the field is clearly not dependent on substrate alone. Es- tuarine substrate type in nature may be in fact, largely the result of prevailing current speed and direction. Since surf clams occur naturally in areas with sand to gravel substrate (Yancey and Welch, 1968; Fay et al., 1983), one might speculate that this environment would be optimal for growth. As indicated, caged clams in sand substrate (1984-5) were disturbed and their growth rate declined (Fig. 2) after the sediment in their cages was washed out. 50 Length (mm) p ie) = 4——-4 sandy—mud O—O mud @—-®@ sond Jan 1985 Dec 1984 Date Figure 2. Growth of Spisula solidissima in 3 substrate types in coastal Georgia (1984-1985). @—-@ temperature a a 2° an xe ee iS eee TF fo) YM 25 eg Co ° 2 20 f 2 =} 7. SS 5 15 e O—O salinity o = (oe) Oct Feb Date Figure 3. Salinity and water temperature data for Wassaw Sound, Georgia. Data from Winker et al. (1985). Dec Apr Had this not occurred, it is likely that the final mean lengths would have been even closer and again no signifi- cant difference would have been detected. Our conclusion is that there is no appreciable difference in growth of clams, maintained in either sandy-mud or sand areas of the estuary. It is not evident, why growth of surf clams was higher at sites in closer proximity to the ocean. It was beyond the scope of this study to identify the specific causes and ef- fects that promote rapid growth. We can, however, suggest the ecological factors at work. Current speed, tidal influ- ence, temperature, phytoplankton species and abundance are among the major variables that influence physiology and growth on a large spatial scale. Additionally, local ef- fects, such as siltation within a particular cage or bottom topography, may account for differences among clams in nearby cages on a smaller spatial scale. Recently, Grizzle and Lutz (1989), have experimentally investigated relation- ships among growth of Mercenaria mercenaria, horizontal seston flux, and bottom sediment type and have developed a descriptive statistical model. The length of the active growth season in Georgia for Spisula corresponds to the period when seawater tempera- tures are below 24°C. Growth continues throughout the TABLE 7. Annual growth rates of surf clams at different locations along the Atlantic Coast of the United States. Total Hengthi(mim), Increase Location Initial Final (mm) Source Milford, CT (Long Island Sound) 15.7 = 47.3 31.6 Goldberg, 1989 Pt. Pleasant, NJ 38.0 62.0 24.0 Jones et al., 1978 Barnegat Bay, NJ 34.0 56.0 22.0 Chang et al., 1976 Ocean City, MD 39:0 57.0 18.0 Chang et al., 1976 Chincoteague Bay, VA 42.2 ~— 68.6 26.4 Ropes, 1969 Wassaw Sound, GA 216) 50 29.4 Present study 192 GOLDBERG AND WALKER 50 + >| E s S 30+ D c o = 20+ O—oO 1984-5 @—e 1985-6 4—A 1987 ti acts al a a eee rer | Nov Jan Mar May Date Figure 4. Fastest growth rates measured for surf clam, Spisula soli- dissima, in Wassaw Sound, Georgia (1984-1987). winter months when typical seawater temperatures drop to approximately 13°C. Our observations indicate that growth rate is most rapid in late fall and early spring and slower in mid-winter. Temperature and salinity data were not col- lected regularly during the course of the present study, however, a synoptic curve (Fig. 3) was derived from an extensive compilation of environmental data sets for this area (Winker et al., 1985). Temperatures at sites | and 2 (1987) are generally about 2°C lower and salinity about 2 ppt higher than at the other sites upriver. Slightly lower temperatures near the mouth of Wassaw Sound may enable growth earlier and later in the grow-out season by pro- viding more time at temperatures below 24°C. This might explain the exceptional growth at site 4 in 1985—6 and that the final mean lengths of clams at sites | and 2 in 1987 were about 3 mm larger than those of clams at the upriver sites. The growth rates of surf clams observed in this study from fall to spring are comparable to annual growth rates reported for natural surf clam populations in more northern areas (Table 7). The best average growth of clams placed in sandy-mud in Georgia during 1984 to 1985 was from 21.6 mm to 51.0 mm, an increase in shell length of 28.4 mm. In all experiments, the least full season increase in growth was 21.4 mm for clams held in cages placed in mud. Even this growth rate is still within the range of values reported in Table 7. The fastest growth rates from selected individual cages in the three field experiments (Fig. 4) were substan- tial and hold promise for future mariculture efforts. Before mariculture of Spisula is attempted in the South- east, state laws pertaining to importation of shellfish stock must be addressed. Legislation concerning species or stock importation is not consistent among coastal states. Consid- eration of possible negative impacts of transplantation on the local ecology should be made by resource managers. A practical manual of guidelines to evaluate potential risk/ benefit of stock movement has been developed by the Inter- national Council for the Exploration of the Sea (Turner (ed- itor), 1988). It is unlikely Spisula would compete with native or- ganisms in Georgia, since it is not able to survive common summer temperatures in excess of 30°C. Inadvertent intro- duction of disease organisms or algal cysts could be pre- vented by examination, depuration, and/or short-term quar- antine of the young seed clams. Many commercial hatcheries make considerable effort to certify their seed to be ‘‘disease-free’’. It is also possible that southern hatch- eries could produce their own seed, after producing sev- eral generations of quarantined broodstock. Rapid growth and high survival rates, make high-density cage culture an attractive possibility. The design of cages in the present study is successful because of its effectiveness in excluding predators. The ability of the clams to burrow in sediment within the cage promotes an ideal environment for growth. Clams are harvested by simply pulling the cage out of the substrate, eliminating the need for digging or dredging. For commercial culture, the dimensions of these units could be scaled up or another unit devised, as long as it adequately excludes predators. Efficiency in handling should be a key factor in selecting an optimal device. To ensure success and predictability of clam production, mariculture efforts should include field growth trials to re- fine site-selection criteria. Many highly productive es- tuarine areas along the coast of the southeastern United States have potential for farming marine species. Maricul- ture of surf clams during the winter months, beyond their natural range in southern waters, is an innovative approach to using our natural resources. ACKNOWLEDGMENTS The authors wish to thank Mrs. J. Haley, Mrs. G. Willis, and Mrs. D. Burull for typing the manuscript and Ms. S. McIntosh and Ms. A. Boyette for the graphics herein. This work was partially supported by Georgia Sea Grant (NA-84AA-D00072). LITERATURE CITED Ambrose, W. G., D. S. Jones & I. Thompson. Distance from shore and growth rate of the suspension feeding bivalve, Spisula solidissima. Proc. Natl. Shellfish. Assoc. 70:207—215. Chang, S., J. W. Ropes & A. S. Merrill. 1976. An evaluation of the surf clam population in the middle Atlantic Bight. U.S. Natl. Mar. Fish. Serv., Northeast Fisheries Center, Sandy Hook, NJ, Lab. Ref. No. 76-1. 43 pp. Fay, C. W., R. J. Neves & G. B. Pardue. 1983. Species profiles: life histories and environmental requirements of coastal fishes and inverte- brates (Mid-Atlantic) . . . surf clam. U.S. Fish Wildlife Service, Di- vision of Biological Services, FWS/OBS-82/11.13. U.S. Army Corps of Engineers, TR EL-82-4. 23 pp. Goldberg, R. 1980. Biological and technological studies on the aquacul- ture of yearling Atlantic surf clams. Part Il: Aquaculture production. Proc. Natl. Shellfish. Assoc. 70:55—60. Goldberg, R. 1989. Biology and culture of the surf clam. In: J. J. Manzi CULTURE OF SPISULA SOLIDISSIMA IN GEORGIA 193 and M. Castagna (eds.) Clam Mariculture in North America. Elsevier Press, Amsterdam, Netherlands. Grizzle, R. E. & R. A. Lutz. 1989. A statistical model relating horizontal seston fluxes and bottom sediment characteristics to growth of Mer- cenaria mercenaria. Marine Biology. 102(1):95—106. 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. Marine Biology. 47:63—70. Krzynowek, J. & K. Wiggin. 1982. Commercial potential of cultured At- lantic surf clams, Spisula solidissima (Dillwyn). Journal of Shellfish Research. 2:173-175. Merrill, A. S., and J. W. Ropes. 1969. The general distribution of the surf clam and ocean quahog. Proc. Natl. Shellfish. Association. 59:40—45. Monte, David. 1986. Mercenaria Manufacturing, Inc. Millsboro, Dela- ware (personal communication). Pratt, D. M. 1953. Abundance and growth of Venus mercenaria and Cal- locardia morrhuana in relation to the character of the bottom sedi- ments. J. Mar. Research. 12:60—74. Pratt, D. M. & D. C. Campbell. 1956. Environmental factors affecting growth in Venus mercenaria. Limnol. Oceanogr. 1:2—17. Ropes, J. W., J. L. Chamberlin & A. S. Merrill. 1969. Surf clam fishery. Pages 119-125. In: R. E. Firth (ed). Encyclopedia of Marine Re- sources. Van Nostrand Reinhold, Co., New York. Savage, N. B. 1976. Burrowing activity in Mercenaria mercenaria (L.) and Spisula solidissima (Dillwyn) as a function of temperature and dissolved oxygen. Mar. Behav. Physiol. 3:221—234. Turner, G. E. (editor). 1988. Codes of practice and manual of procedures for consideration of introductions and transfers of marine and fresh- water organisms. ICES, Cooperative Research Report, No. 159. Walker, R. L. 1983. Feasibility of mariculture of the hard clam, Merce- naria mercenaria, (Linne) in coastal Georgia. Journal of Shellfish Re- search. 3(2):169—174. Winker, C. D., L. C. Jaffe & J. D. Howard. 1985. Georgia Estuarine Data 1961-1977. Georgia Marine Science Center Technical Report Series 85-7, Vol. 1, Savannah, Georgia. 420 pp. Yancey, R. M. & W. R. Welch. 1968. The Atlantic coast surf clam— with a partial bibliography. U.S. Fish. Wild. Serv. Circ. 288. 13 pp. Journal of Shellfish Research, Vol. 9, No. 1, 195—203, 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 T. W. ROWELL,! D. R. CHAISSON,? AND J. T. MCLANE? 'Department of Fisheries and Oceans Biological Sciences Branch, Habitat Ecology Division Bedford Institute of Oceanography P.O. Box 1006, Dartmouth, Nova Scotia, Canada B2Y 4A2 ?Bio-Atlantech Limited Dartmouth, Nova Scotia, Canada ABSTRACT Ocean quahaugs, Arctica islandica, were collected from near shore populations in south western Nova Scotia from May 1982 to June 1984, and examined histologically for a study of their sexual maturation and gametogenic cycle. One hundred and sixty-eight quahaugs, 16—65 mm in shell length, were sampled for developmental stage relative to size, age, and sex. Of these, 47 were undifferentiated, 84 were intermediate, and 37 mature. Sex could be determined in some males and females by age 3, but ages of undifferentiated quahaugs ranged from 3—12 years. Males appear to differentiate at a smaller size and younger age than females. Both males and females matured as early as age 7, the mean ages of mature specimens being 13.1 and 12.5, respectively. Females, however, matured in a ‘‘knife-edge’’ manner, over a generally much shorter age span than males. To determine the gametogenic cycle, a total of 894 quahaugs were sampled, over a 25 month period. The timing and duration of gametogenesis varied between years. Spawning activity, as inferred from histological examination, appears to have continued from the July-September period of 1982 through to late 1983, after which a large percentage of both males and females were found to be in the spent and early active phases. The February-May periods of 1983 and 1984 were particularly different from each other, all animals being partially spawned in 1983 and virtually all phases being well represented in 1984. Sex ratios between males and females were examined relative to size for a total of 1029 quahaugs, and relative to age for a sample of 346. The male to female (M:F) ratios varied considerably between size-classes, with males consistently dominating, and an overall M:F ratio of 2:1. In the youngest and smallest animals, the predomi- nance of males is even greater. This early male dominance may be in part related to their earlier development of germinal cells and may also reflect a compensatory strategy to ensure that sufficient numbers of males reach the size and age of reproduction. KEY WORDS: INTRODUCTION The ocean quahaug, Arctica islandica, is widely distrib- uted over the continental shelves of both Europe and North America. It occurs, generally in sandy mud and mud bottoms, along the east coast of North America from New- foundland to Cape Hatteras, on the coasts of Iceland, the Faroes, the Shetlands, the British Isles, and along the Euro- pean coast from the White and Barents Seas to the Bay of Cadiz in Spain (Nicol 1951; Merril and Ropes 1969; Ropes 1979b). In North American waters it has been reported from depths as shallow as 4 M (Rowell and Chaisson 1983) to a maximum recorded depth of 256 M (Merril and Ropes 1969). In Canadian waters, the ocean quahaug is found over most areas of the Scotian Shelf, with the greatest den- sities occuring in the harbours and bays of southwestern Nova Scotia at depths of 4-18 M. In offshore Canadian waters, densities are lower, but major concentrations, with commercial potential, are located on Sable Island Bank in depths of 36-54 M (Rowell and Chaisson 1983), a depth range similar to that of concentrations off the U.S. coast (Merril and Ropes 1969). The ocean quahaug is the focus of an important com- mercial fishery in the U.S., with 1988 landings of 44.2 sexual maturity, gametogenic cycle, Arctica islandica million lbs. (20,045 metric tons) of meat (Mid-Atlantic Fishery Management Council, Preliminary catch data'). In Canada, the quahaug has to date supported only a minor inshore fishery, with highest landings, in the years 1970-71, on the order of 900—1300 tons? (Caddy et al. 1974, Rowell and Chaisson 1983). With the growing development and utilization of qua- haug products in the U.S. throughout the 1960’s and 70’s, considerable interest was aroused as to the possibility of developing an expanded Canadian fishery. At the time, knowledge of the quahaug resource in Canadian waters was restricted to a few inshore areas. Hiltz (1977) reviews the state of knowledge up to that time. As a result of this interest, a study was undertaken be- tween 1980 and 1984 to assess the distribution, abundance, and biology of the ocean quahaug and other possible com- mercially valuable bivalve molluscs on the Scotian Shelf, as a basis for both development and management (Rowell and Chaisson 1983; Chaisson and Rowell 1985; Rowell and 'Mid-Atl. Fish Management Counc., Rm 2115, Federal Bldg., 300 South New St., Dover, Delaware, U.S.A. ?Round weight 195 196 Amaratunga 1986; Amaratunga and Rowell 1986). An ex- amination of the gametogenic cycle and the size and age of sexual maturity was included in the 1980—1984 study, but has not till now been reported on. Previous studies on size and age at sexual maturity have been reviewed by Thompson et al. (1980a) and Ropes et al. (1984a) for the ocean quahaug in the Mid-Atlantic Bight area of the U.S. The gametogenic cycle of this species has also been reported on for eastern U.S. waters by Loosanoff (1953), Jones (1981), and Mann (1982) and for European waters by von Oertzen (1972). This study represents the first documentation of these aspects of the biology of the ocean quahaug for more northerly areas in western Atlantic waters. MATERIALS AND METHODS Size and Age of Sexual Maturity Quahaugs for the size and age at maturity study were collected by hydraulic dredge, in July 1982, from a depth of 34 M in St. Marys Bay, N.S. (Fig. la). Quahaugs were sorted from the catch and returned directly to the labora- tory. The entire visceral mass of each quahaug was re- moved and preserved in Bouin’s fixative for 48—72 hr, be- fore storage in 70% ethanol. Shells were measured for length to the nearest mm, and aged by internal shell band 66° 20° 44° 50! BOW @ SAMPLE SITE ——— FATHOMS 44° 10'f 0 1 a Nm 44° 710° -2 oon @ noe % ( | 43° 55° @ SAMPLE SITE ——— FATHOMS i) 1 eee Nm 55° 64° 50° 45° Figure 1. a) Sampling sites of 168 small Arctica islandica used for determination of size and age at sexual maturity. b) Sampling site for Arctica islandica used for reproductive cycle study, between Sep- tember 1982 and June 1984. ROWELL ET AL. counts, as described by Thompson et al. (1980b) and Ropes et al. (1984b). A representative section of gonadal tissue was obtained for histological study by cutting dorso-ventrally from the hinge region to the ventral region of the mantle edge and removing a 2—3 mm thick section of gonad. In small an- imals sections included the whole gonad, whereas in larger animals they consisted of a planar section of up to 15 mm? of gonad from the mid-ventral area which was considered representative of the entire gonad. Histological preparation of the tissue was according to Humason (1972) with em- bedding in paraffin, sectioning at 8 w, and staining with Harris Hematoxylin and Eosin Y counterstain. After preparation, the sections were examined micro- scopically for the presence of differentiated gonads. Those specimens having little or no tubule development, no cel- lular structures defineable as male or female, and much of the gonad area filled with connective tissue were desig- nated undifferentiated. Those with sufficient development to be differentiable as males and females were further clas- sified as intermediate or mature in their gonadal develop- ment according to the criteria described in Ropes et al. (1984a). These stages equate with the early and late devel- opmental stages of Thompson et al. (1980a). Intermediate specimens were typified by reduced to sparse tubule devel- opment with tubules widely spaced and separated by vesic- ular connective tissue. The tubules themselves displayed varying degrees of development; from those of small diam- eter and lacking germinal cells in portions of the epithelium to those typifying the mature condition. Mature quahaugs had very little connective tissue and larger, more clearly defined, tubules which generally completely filled the go- nadal area. Gametogenic Cycle Quahaugs collected for study of the gametogenic cycle were taken, by SCUBA, from a depth of 13 M in Port Mouton, N.S., at roughly two week intervals between Sep- tember 1982 and June 1984 (Fig. 1b). Three additional samples, covering the June-August period of 1982, were also available from nearby sites, at approximately the same depths, in Rose and Jordon Bays and Shelburne Harbour. Over the 25 month period of field sampling, with 36 sampling dates, a total of 894 quahaugs, with shell lengths ranging from 43-103 mm were collected and examined. After collection, they were treated and prepared for histo- logical examination in the manner described above. Gonadal sections were classified according to the cri- teria of Ropes (1968), and as used by Thompson et al. (1980a) and Jones (1981), into one of five phases: 1) early active; 2) late active; 3) ripe; 4) partially spent; and 5) spent. Ropes (1968) notes that the gametogenic process in the ocean quahaug is more or less continuous and that de- marcation between phases is not sharp. Mann (1982), while using the similar criteria of Holland and Chew (1974) in GAMETOGENESIS IN THE QUAHAUG ARTICA ISLANDICA 197 describing the quahaug’s reproductive cycle, also com- ments on the qualitative nature of this classification. Al- though convenient, the divisions between phases remain subjective. Bottom temperature data were collected for the Port Mouton sampling area between September and December 1982, using a Ryan Thermograph, but, following the loss of two recorders, direct on-site temperature monitoring was discontinued. Additional Ryan temperature data were ob- tained, for the period November 1982 to May 1983, from an array of long-term temperature monitoring sites, estab- lished by the Bedford Institute of Oceanography, at 18—25 M depth in Port Mouton, very near the quahaug sampling area and, for the entire period of the study, from a similar array of temperature sites at 13—15 M depth in Cape Sable. Sex Ratio Sex ratio was examined, relative to size, in 10 mm size classes, for a total of 1029 animals having shell-lengths ranging from 21—103 mm. Sex ratio relative to age (yr) was determined for a sample of 346 animals ranging from 3-144 yr. Additionally, the intermediate and mature spec- imens from the 168 animals used in the size and age at maturity study were examined in an attempt to explain the shift in sex ratio observed with increasing size and age. RESULTS Size and Age of Sexual Maturity Of the 168 quahaugs sampled, 47, ranging in length from 16—45 mm, were found to be sexually undifferen- tiated (Table 1, Fig. 2). Of these, 25 animals, 21—42 mm in length, could be aged. Ages ranged from 3—12 yr, with a mean of 4.6 yr. Sex could be determined in 121 specimens; 83 males and 38 females. The size and age of sexually undifferen- tiated clams and of males and females in intermediate and mature stages are plotted in Fig. 2. The data indicate that males generally differentiate at a smaller size and younger age than females. Differentiable males ranged in length 60+ ° e e ° 50 ee 8 . AL) © e 40+ 8 g ° 30+ ° 8 FEMALES 207 @ MATURE © INTERMEDIATE 10+ | 0 =n 1 1 4 e 60f O e = e e E 50 : Ss 8 = ° ° 7 ° t } - 40 oO e oO ° a 30 a8 ° ° = 8 MALES = ° um 20t @ MATURE BS © INTERMEDIATE (V>1 ANIMAL) 30F ‘ : 3 UNDIFFERENTIATED 3 (¥>1 ANIMAL) 207 10F 0 rn n 1 1 am 4 0 3 10 15 20 25 30 AGE (YEARS) Figure 2. Maturity stages relative to shell length and age for Arctica islandica from St. Mary’s Bay, N.S. from 21—64 mm (x = 34.3 mm). Fifty-three aged males, 24—65 mm in length, ranged from 3—20 yr in age (x = 7.5 yr). Females ranged in length from 25—64 mm (x = 42.7 mm). Twenty-six females, 27-65 mm in length, had a mean age of 9.8 yr while ranging between 3—24 yr. Eighty-four quahaugs, 65 males and 19 females were determined to be intermediate in their gonadal develop- ment. Intermediate males ranged in length from 21—48 mm. Forty males between 24—48 mm in length were aged. They ranged in age from 3—20 yr, with a mean age of 5.6 yr. Intermediate females ranged from 25—52 mm in length. TABLE 1. Stages of gonadal development relative to size, age, and sex. Intermediate Mature Undifferentiated Males Females Males Females Total Aged Total Aged Total Aged Total Aged Total Aged Number 47 25 65 40 19 11 18 13 19 15 Length (mm) 16-45 21-42 21-48 24-48 25-52 27-52 27-64 40-64 30-65 40-65 X Length 27.90 28.12 30.10 33.20 34.10 33.18 47.10 49.90 49.20 50.80 SD Length 4.90 4.60 7.01 6.00 6.89 6.54 8.41 TNT. 8.39 7.38 Age (yr.) a 3-12 — 3-20 — 3-24 — 7-20 = 7-28 X Age 4.60 — 5.60 — 6.20 — 13.10 — 12.50 SD Age — 5.24 — 3.44 — 5.68 — 4.70 — 5.52 198 ROWELL ET AL. Of these, 11 females between 27—52 mm in length were aged. They ranged from 3—24 yr and had a mean age of 6.2 yr. Intermediate males displayed all phases of the gameto- genic cycle between early active and partially spawned, although no partially spawned males were less than 26 mm in length. Females displayed only the early gametogenic stages of early and late active phases. In the size at maturity sample, 37 quahaugs were found to be mature; 18 males and 19 females. Males ranged in length from 27—64 mm, with a mean of 47.1 mm, while females ranged from 30—65 mm with a mean of 49.2 mm. In both cases, only one mature animal was <40 mm in length. Mature males displayed late active through partially spawned phases, while females were in the early active through ripe phases. In the age at maturity sample, 28 ma- ture specimens were found; 13 males and 15 females. Ma- ture males (lengths 40—64 mm) ranged from 7—20 yr, with a mean of 13.1 yr. Mature females (length 40-65 mm) ranged from 7—28 yr and had a mean age of 12.5 yr. When Fig. 2 is examined, it is apparent that both males and females begin to mature at around 40 mm shell length and have very similar patterns of maturation relative to size. Relative to age, however, the sexes have very dif- ferent patterns of maturation. Females appear to mature in a “‘knife-edge’” manner at age 7, whereas, males, while commencing maturation at the same age, still have 41% of those age 7 and older in an intermediate stage. Maturation among males, relative to age, appears to be a much more extended process. Gametogenic Cycle Changes in the % of males and females in the various stages of the gametogenic cycle throughout the sampling period are presented, in a manner similar to that used by Mann (1982), in Fig. 3. Commencing in the May-July period of 1982, males were predominantly in the late active to ripe phases, al- though all stages, except spent, were present. During the same period, females were primarily in the early active to late active gametogenic phases. In late July through Au- gust, 100% of the females sampled were ripe. A majority of males (63—100%) were in the partially spawned phase from August 1982 through May 1983, with 8—13% in the ripe phase and 4—37% in the spent phase between De- cember 1982 and January 1983. In excess of 80% of fe- males remained in the ripe phase until late September 1982. Partially spawned (73%) and spent (16%) females com- prised the bulk of the population in the October through December period, with 87% being partially spawned during the January through June 1983 period. Late active (39%) and ripe (56%) phases were predomi- nant in females during July 1983, with 65% becoming par- tially spawned between August and October. Males in the June to November 1983 period were largely in the partially spawned phase, with the remainder (10—32%) being ripe. During the winter months of November 1983 through Feb- ruary 1984, 67% of males and 47% of females were in the spent phase. From April through June 1984, both males and females had high percentages in the early gametogenic stages with 85% of females and 44% of males in the early to late active phases. The timing and duration of gametogenic events varied between years. In 1982, spawning, as indicated by a high percentage of partially spawned individuals, began during July for males and in September—October for females. In 1983, the second year of the study, 100% of the animals were in a partially spent phase from February through to May, with a few spent females seen in June. This was fol- lowed by a period in which late active and ripe females and ripe males were again observed. Spawning appeared to be continuous until November. Spawning continued into the winter months in both 1982—83 and 1983-84, although only a few spent individuals were found in samples from the first winter. This contrasts with the second winter, where there were high percentages of spent males (67%) and females (47%). Generally, later stages of gametogenisis appear to domi- nate during periods when sea temperatures are rising or are at their highest levels, while early stages are more prevalent during periods of lower water temperatures (Fig. 3); how- ever, during the period of July—November 1982, males did not appear to follow this pattern, although females did. Sex Ratio Sex ratios relative to size groups and age in years are presented in Tables 2 and 3, respectively. The overall M:F ratio is 2:1. The changes in sex ratio observed with in- creasing size and age immediately suggest the possibility of protandry, however, only one hermaphrodite (75 mm) was observed among all animals sampled. Although the sex ratio varies considerably between size- classes (Table 2), males are consistently dominant throughout. There appears to be no general pattern of change with size; however, the high ratio of males to fe- males (6.5:1) seen in the smallest size-class suggests that males predominate greatly during the earliest period of sexual differentiation and that this predominance then de- clines rapidly. Sex ratio relative to age (Table 3) showed a similar vari- ability to that seen with size-classes, but the data do not as strongly suggest a heightened preponderance of males among the youngest sexually differentiated animals; how- ever, the data in Table | indicate a shift in M:F ratio from approximately 3.4:1 in the intermediate stage to 1:1 in the mature stage. Additionally, as noted in the section on age and size of maturity, the data presented in Figure 2 suggest a distinct difference in the rate at which male and female quahaugs mature relative to age, but not relative to size. This may, in some manner, result in the very high ratio of males to females in the youngest and smallest animals and GAMETOGENESIS IN THE QUAHAUG ARTICA ISLANDICA 199 1982 1983 1984 inh apa Tals To dw loo le itelaA lw lala is To Ups plo We ie a a it A 1028 19111022 19 14 8 24 2011 1216 2017 15211916 16172013212015 11 16 16 15 24 19 100 me ES oo leas 80 fF N % 6ot N : \ Z N U = 40} N e NN xe N aa aot N | N N | ot S ThA tot . \ ; \ S N 20 F N S 6s N a WwW 40 + z N Z \ Z cot |, N Ww 60 N x N al 80 + \ is NS Sas bis Bik oon & i AS 7A OOo Me PPE G ihe & GF A @ 12 5 10 EARLY J LATE RIPE PARTIALLY SPENT ACTIVE WS ACTIVE bo SPAWNED Ocean temperature, Cape Sable, N.S Degrees Celsius Figure 3. Percent changes in gametogenic stages between May 1982 and June 1984 and ocean temperature profile for the same period. Numbers of males and females sampled at each interval are shown at ends of bars. Short-term sample site temperature data between September 23 through December 22, 1989, are shown as a dashed line. in the tendency of the ratio to decrease among older larger 1979a). The size and age of maturity results are very sim- animals. ilar to those reported in the literature for quahaugs from the DISCUSSION more southern waters of the eastern U.S. Undifferentiated or immature quahaugs were observed to a maximum length Size and Age of Sexual Maturity Sexual maturity is attained by an animal when gametes eee: are produced for the first time in its life-history (Ropes Male:female sex ratios relative to age. TABLE 2. KecGroun Numbers MF Male:female sex ratios relative to size. (yrs.) Males Females ratio 37) 41 17 2.41:1 SeeGronn Numbers M:F 8-12 21 15 1.40:1 (mm) Males Females ratio Hild ae 2 oe 18-22 8 8 1.00:1 20-29 26 = 6.50:1 23-27 11 13 0.85:1 30-39 35 16 DAG a 28-32 13 9 1.44:1 40-49 38 23 1.65:1 33-37 15 8 1.88:1 50-59 36 33 1.09:1 38-42 9 5 1.80:1 60-69 99 1p2 1.38:1 43-47 5 4 E2531 70-79 145 77 1.88:1 48-52 3 0 — 80-89 163 50 3.26:1 53-57 4 2 3.50:1 90-99 132 61 2.16:1 58-100 59 29 2.03:1 >100 12 7 7g > 100 7 2.43:1 1 Total 686 343 2.00:1 Total 221 125 200. of 45 mm and an age of 12 yr. This compares to 47 mm and 14 yr for quahaugs off Rhode Island (Thompson et al. 1980a) and to 46 mm and 8 yr for quahaugs off Long Island (Ropes et al. 1984a). In this study, degrees of maturity, among differentiated animals, were apparent, as reported in Thompson et al. (1980a) and Ropes et al. (1984a). Thompson et al. (1980a) reported 77% of differentiated quahaugs <51 mm in length being in the early developmental stage [intermediate stage of Ropes et al. (1984a)]; they ranged from 26—51 mm for males and 39—50 mm for females. We found 69% of dif- ferentiated quahaugs to be in the intermediate stage, ranging between 21—48 mm for males and 25—52 mm for females. Ropes et al. (1984a), although further defining in- termediate quahaugs into those with sparse and those with moderate tubule development, found an overall size range of 20—48 mm. Minimum age to attain the intermediate stage of matu- rity was 3 yr for both males and females, the same as re- ported for males by Ropes et al. (1984a). They, however, reported a minimum age of 5 for intermediate stage fe- males. Our intermediate males and females averaged 5.6 and 6.2 yr of age, respectively, similar to means of 5.7 and 6.3 yr for intermediate quahaugs reported by the same au- thors (their sparse and moderate tubule development classes combined). Thompson et al. (1980a) also found interme- diate stage individuals at 6—7 yr. They also report imma- ture specimens ranging in age from 4—14 yr, with a mean of 9.4 yr at a shell length of approximately 39 mm. They suggest this wide range may be related to growth rate and locality; their samples having been collected over a rela- tively large area between Long Island and New Jersey (a distance of 2° of latitude). Our data, with immature spec- imens ranging from 3—12 years of age at one location, sug- gest that this variability in age of maturity is less influenced by environmental location than they suggest. The smaller size and apparently younger age at which males reach the intermediate stage may, as suggested by Ropes et al. (1984a), explain the high ratio of males to females seen in smaller and younger quahaugs. They con- cluded that the disparity in age for the initiation of gameto- genesis, with males producing germinal cells at a smaller size and younger age than females, was probably respon- sible for the highly imbalanced sex ratio seen among inter- mediate stage quahaugs. Our results showed evidence of spawning in interme- diate males as small as 26 mm, but females in the same stage did not proceed beyond the late active phase. This is in agreement with Ropes et al. (1984a) and Thompson et al. (1980a) who reported morphologically ripe sperm in males while in females oogenesis never progressed beyond an early developmental stage. One 27 mm male and one 30 mm female were classified as fully mature; smaller than previously reported for this stage of development. These two specimens were the only ROWELL ET AL. fully mature individuals <40 mm in length. Thompson et al. (1980a) reported fully mature quahaugs as small as 42 mm (aged 11 yr), while Ropes et al. (1984a) found males as small as 36 mm and as young as 5 yr. Ropes et al. (1984a) found means for this stage of 49.7 mm and 10.9 yr. This compares to a mean length of 49.9 mm and 13.1 yr for males and 49.2 mm and 12.5 yr for females in our study. When the size and age of maturity data for the smaller size and age groups are examined (Fig. 2), the pattern of maturation (from intermediate to mature stage) observed is somewhat different than that described above from Table 1. Relative to age, the data show a ‘‘knife-edge’’ pattern of maturation for females and a gradual one for males. Addi- tionally, the data suggest that the rate of maturation by size in males and females is very similar. It logically follows that, in general, females must grow more rapidly to the sizes at which they mature. Such a pattern could have sig- nificant advantages for the survivorship of young female quahaugs and possibly provide a further basis for the much higher ratio of males to females among smaller and younger animals and the subsequent decline in this imbal- ance with increasing size and age. For example, by achieving a larger size more rapidly, females would acquire the advantages of thicker shells and deeper burrowing and hence be less subject to predation. Males, on the other hand being generally slower to grow to the size of maturity, would require greater numbers of individuals in order to balance off their lower survivorship prior to maturation. Unfortunately, our length-at-age data for the lower age and size classes were insufficient to examine the growth rates of intermediate stage males and females. Grouping the two intermediate categories from Table 1 of Ropes et al. (1984a), allows a direct comparison with the data in Figure 2 of this study. They found no intermediate males or fe- males beyond 48 and 45 mm, respectively, however, be- cause of the manner in which their data is presented (10 mm size classes), it is impossible to determine whether there is any difference in the maturation of the two sexes relative to size. Their age data suggest that maturation in males takes place between 5 and 8 yr, and in females some- where between 6 and 10 yr. Unfortunately, they have data for only two females in the critical ages from 8 to 11, one intermediate at 8 yr and one mature at 11 yr. Despite these limitations, their data appears sufficient to indicate a younger maturation of males than observed in this study, and that the maturation of males is protracted over a range of ages, as in this study. Growth rates during the intermediate stage, and the pos- sible link between maturation and the widely observed im- balance and shift in sex ratios, deserve further study. Gametogenic Cycle Both Jones (1981) and Mann (1982) have shown the timing and duration of events in the quahaug gametogenic GAMETOGENESIS IN THE QUAHAUG ARTICA ISLANDICA 201 cycle to be highly variable between years. A similar vari- ability is observed in our results, with the 1983—84 pattern bearing the greatest similarity to the cycle observed by others. The great year to year variability within the data undoubtedly reflects many factors, both environmental and endogenous. Some of the apparent variability may be arti- factual, resulting from small sample size and, possibly, sampling technique. For example, it is known that sequen- tial development occurs within the gonads of bivalve mol- luscs, including the ocean quahaug. Jones (1981) and Mann (1982) both found sequential development in ocean quahaugs, but took the mid-ventral area, as in this study, to be representative of the overall reproductive state. Mann (1982) suggested that his results indicated a shorter period to attain ripeness in females than in males. Our results indicate that males do tend to achieve ripeness prior to females, with ripe males apparently first present in April, May, and June, while ripe females are first seen in June and July. This, with the need for synchrony in spawning in a dioecious species such as the quahaug, would require that females ripen more quickly. An unusual aspect of the 1982—83 period was that, fol- lowing the initiation of spawning in July 1982 for males and September for females, the partially spawned condition represented approximately 70—100% of both males and fe- males through to July 1983; essentially a full year. A sim- ilar pattern can be seen in the population studied by Mann (1982), where approximately 60—100% of the males sam- pled over two years were in the partially spawned or spent phases. Both our results and those of Mann (1982) were much more variable than those reported by Jones (1981). The exact combination of environmental factors which influence spawning in quahaugs is not understood. Sastry (1975, 1979) reviews the physiology and ecology of repro- duction in pelecypod molluscs and generally concludes that the diverse patterns of gametogenic cycles observed are in- fluenced by a number of environmental and endogenous factors; temperature being one of the important environ- mental factors. Landers (1976) found that gametogenesis in A. islandica could be accelerated in the laboratory by simu- lating summer temperatures; but only at certain times of the year. Reduced temperatures have been shown to delay ga- mete development and maturation in the bay quahaug, Mercenaria mercenaria Linné (Loosanoff and Davis 1951). Ropes (1968) stated that temperature may delay or hasten both the gametogenic cycle and spawning of ripe Spisula solidissima Dillwyn. Newell et al. (1982), in a study on temporal variability in the cycle of Mytilus edulis, suggest that variations in food availability and nutrient re- serves may have a strong influence on reproductive condi- tioning, while Griffiths (1977), working with another species of mussel in South African waters, concluded that food availability may be of greater importance than temper- ature. Although we were unable to obtain a complete water temperature record from the Port Mouton sampling area, the longer time series available for Cape Sable matches well for the periods of overlap and may, for our purpose, be considered representative of the annual cycle for the study area. While the observed pattern of gonadal develop- ment and spawning does generally follow the sea tempera- ture cycle, the variability seen suggests that other factors may also have had a variable, yet important, influence in the cycle for different years. Although some patterns can be discerned for the dif- ferent phases of the cycle, they are so highly variable as to suggest that year round spawning likely occurs, possibly with one or two peak periods which vary in timing and intensity from year to year. The extensive periods during which partially spawned but few spent individuals were en- countered can be explained if one assumes continual condi- tioning of the gonads coupled with continual spawning. Newell et al. (1982) suggested such a possibility for M. edulis and Rowell (1967) reported a similar phenomenon in the horse-mussel Modiolus modiolus Linné. The lack of ripe and the few spent specimens during these same periods may reflect a low incidence of animals which have com- pletely spawned out and subsequently reconditioned. Al- though larger sample sizes would clearly have enhanced our understanding relative to the timing and variability of peak spawning periods, the current data are sufficient to document continuous release of gametes throughout the year. In our results, virtually 100% of males were in either the ripe or the partially spawned stages for the entire 15 months between July 1982 and October 1983 (one sample, in January 1983, having 5% spent). Females showed more variability during this same period, having a few spent indi- viduals in November and December 1982 and in May, June and October 1983 as well as some late active individuals in June and July 1983. von Oertzen (1972) reported spawning commencing in late April-early May and extending to early October for Baltic Sea quahaugs. In the same study, ripe eggs and sperm are reported from February through No- vember. Mann (1982) reported spawning for quahaugs off Rhode Island beginning in May, but with the heaviest ac- tivity between August and November. Jones (1981) sug- gested that-spawning off New Jersey is a fall and early winter event, although it may be delayed and extend into winter. The dominance of early active through ripe phases in our Spring 1982 data (May-June) suggests a major spawning out of the population may have occured in the late winter prior to commencement of this study. The ab- sence of ripe, and the high percentage of spent specimens, seen in the November through January samples also sup- ports the occurence of peak spawning periods during the winter months. Our results then, indicate continuous year-round spawning with one annual peak of heavier spawning in the winter period between November and January or February. Mann (1982) stated that gonadal maturation occurs only 202 once per year. This is indirectly supported by Thompson et al. (1980b) who determined that annual shell bands are de- posited at the time of spawning, and, that in the quahaug, the number of bands could only reasonably be explained if deposition was on an annual basis. Ropes et al. (1984b) provide further confirmation that the growth lines are an- nual and that they are laid down in the early fall. The results of this study and previous studies of the qua- haug gametogenic cycle indicate that on a population level, spawning may be very protracted. It is also clear from our results that spawning on the individual level may be very protracted as well; partially as a result of sequential devel- opment within the gonads and possibly as a response to environmental factors. Sex Ratio The overall predominance of males within the popula- tion is clear. Previous studies of sex ratios in quahaug pop- ulations have generally indicated ratios in favour of males, although results have been quite variable, and the ratios less imbalanced. In some studies, females have been found to be dominant among larger animals. The imbalance in favour of males (2:1), found in this study for quahaugs of 21-103 mm, is higher than the 1.39:1, for quahaugs 58-125 mm in length from depths of 25—32 M off New Jersey, reported by Jones (1981). Our results are very dif- ferent from those of Mann (1982) and Ropes et al. (1984a) who reported sex ratios not significantly different from parity for quahaugs in the ranges of 70-110 and 57-103 ROWELL ET AL. mm, respectively, from depths of 27—50 M off Rhode Is- land and 53 M off New Jersey. When Ropes et al. (1984a) separated their data into 10 mm size groups, they found a significant difference (P < 0.05) in favour of males for the 80—89.9 mm size group, and a highly significant differ- ence (P < 0.01) in favour of females in the 100-110 mm size class. Consideration of the apparent shift in sex ratios, as seen in our samples and reported in the literature, from male to female predominance as size increases, hints strongly at protandry; however, our finding of only one hermaphrodite eliminates this possibility. Mann (1982) also observed hermaphrodism, but only in two individuals. Other researchers have also concluded that the ocean qua- haug is strictly dioecious (Loosanoff 1953, von Oertzen 1972, Thompson et al. 1980a, Jones 1981, Mann 1982, and Ropes et al. 1984a). As discussed above, under size and age of maturity, more rapid growth by females to a size at which maturity is achieved may allow them greater survivorship and provide an explanation for the shift in sex ratio with size and age. ACKNOWLEDGMENTS The authors wish to thank Dr. Brian Petrie and Ms. Helen Hayden for their assistance in obtaining the required sea temperature data, Mr. Steve Smith for his advice on the value of a statistical analysis of the sex ratio data, and Drs. Shawn Robinson and Thomas Sephton for their construc- tive reviews of the manuscript. REFERENCES Amaratunga, T. & T. W. Rowell. 1986. New estimates of commercially harvestable biomass of Stimpson’s surf clam, Spisula polynyma, on the Scotian Shelf based on the January through April 1986 Test Fishery and new age data. Can. Atl. Fish. Sci. Adv. Com. Res. Doc. 86/112:24 p. Caddy, J. F., R. A. Chandler & D. G. Wilder. 1974. Biology and com- mercial potential of several underexploited molluscs and crustaceans on the Atlantic coast of Canada. Report presented to the Federal-Pro- vincial Fisheries Committee Meeting on Utilization of Atlantic Re- sources, Montreal, February 5S—7, 30 p. Chaisson, D. R. & T. W. Rowell. 1985. Distribution, abundance, popu- lation structure, and meat yield of the ocean quahaug (Arctica islan- dica) and Stimpson’s surf clam (Spisula polynyma) on the Scotian Shelf and Georges Bank. Can. Ind. Rep. Fish. Aquat. Sci. 155:ix + 125 p. Chandler, R. A. 1983. Ocean quahaug survey, south shore of Nova Scotia, 1971-1972. Can. Man. Rep. Fish. Aquat. Sci. 1726:1v + 28 p. Griffiths, R. J. 1977. Reproductive cycles in littoral populations of Choro- mytilus meridionalis (Kr.) and Aulacomya ater (Molina) with a quanti- tative assessment of gamete production in the former. J. Exp. Mar. Biol. Ecol. 30:53-71. Hiltz, L. L. 1977. The ocean clam (Arctica islandica). A literature re- view. G. E. Mack, Ed. Can. Fish. Mar. Serv. Tech. Rept. 720:117 p. Holland, D. A. & Chew, K. K. 1974. Reproductive cycle of the manila clam (Venerupis japonica), from Hood Canal, Washington. Proc. Natl. Shellfish. Assoc. 64:53—58. Humason, G. L. 1972. Animal tissue techniques, third edition. W. H. Freeman. San Francisco. 569 p. Jones, D. S. 1981. Reproductive cycles of the Atlantic surf clam, (Spisula solidissima), and the ocean quahaug, (Arctica islandica), off New Jersey. J. Shellfish Res. 1(1):23-32. Landers, W. S. 1976. Reproduction and early development of the ocean quahaug, Arctica islandica, in the laboratory. Nautilus 90(2):88—92. Loosanoff, V. L. 1953. Reproductive cycle in Cyprina islandica. Biol. Bull. (Woods Hole) 104(2):146—155. Loosanoff, V. L. & H. C. Davis. 1951. Delaying spawning of lamelli- branchs by low temperature. J. Mar. Res. 10:197—202. Mann, R. 1982. The seasonal cycle of gonadal development in Arctica islandica from the southern New England shelf. Fish. Bull. 80(2): 315-326. Merrill, A. S. & J. W. Ropes. 1969. The general distribution of the surf clam and ocean quahaug. Proc. Natl. Shellfish. Assoc. 59:40—45. Newell, R. I. E., T. J. Hilbish, R. K. Koehn & C. J. Newell. 1982. Temporal variation in the reproductive cycle of Mytilus edulis L. (Bi- valvia: Mytilidae) from localities on the east coast of the United States. Biol. Bull. (Woods Hole) 162:299—310. Nicol, D. 1951. Malacology. Recent species of the veneroid pelecypod Arctica. J. Wash. Acad. Sci. 41:102—106. Ropes, J. W. 1968. Reproductive cycle of the surf clam, Spisula solidis- sima, in offshore New Jersey. Biol. Bull. (Woods Hole) 135:349—365. 1979a. Shell length at sexual maturity of surf clams, Spisula soli- dissima, from an inshore habitat. Proc. Natl. Shellfish. Assoc. 69:85— 91. GAMETOGENESIS IN THE QUAHAUG ARTICA ISLANDICA 20 1979b. Biology and distribution of surf clams (Spisula solidissima) and ocean quahaugs (Arctica islandica) off the northeast coast of the United States. Proc. Northeast Clam Ind. Manag. Future Ext. Sea Grant Prog. Univ. Mass. Inst. Technol. SP-112:46—66. Ropes, J. W. & S. A. Murawski. 1980. Size and age at sexual maturity of ocean quahaugs, Arctica islandica Linné, from a deep oceanic site. I.C.E.S. Shellfish Com. K: 26, 7 p. Ropes, J. W., S. A. Murawski & F. M. Serchuk. 1984a. Size, age, sexual maturity, and sex ratio in ocean quahaugs, Arctica islandica Linné, off Long Island, New York. Fish. Bull. 82(2):253—267. Ropes, J. W., D. S. Jones, S. A. Murawski, F. M. Serchuk & A. Jearld Jr. 1984b. Documentation of annual growth lines in ocean quahaugs, Arctica islandica Linné. Fish Bull. 82(1):1-19. Rowell, T. W. 1967. Some aspects of the ecology, growth, and reproduc- tion of the horse-mussel, Modiolus modiolus L., M.Sc. Thesis, Queen’s University, Kingston, Ont. 136 p. Rowell, T. W. & Chaisson, D. R. 1983. Distribution and abundance of the ocean quahaug (Arctica islandica) and Stimpson’s surf clam (Spi- sula polynyma) resource on the Scotian Shelf. Can. Ind. Rep. Fish. Aquat. Sci. 142:v + 75 p. Rowell, T. W. & T. Amaratunga. 1986. Distribution, abundance, and preliminary estimates for production potential of the ocean quahaug Ww (Arctica islandica) and Stimpson’s surf clam (Spisula polynyma) on the Scotian Shelf. Can. Atl. Fish. Sci. Adv. Com. Res. Doc. 86/56: 21 p. Sastry, A. N. 1970. Reproductive physiological variation in latitudinally separated populations of the bay scallop, Aequipecten irradians La- marck. Biol. Bull. (Woods Hole) 138:56—65. Sastry, A. N. 1975. Physiology and ecology of reproduction in marine invertebrates. In Physiological Ecology of Estuarine Organisms F. J. Vernberg, Ed. Univ. South Carolina Press. p. 279-299. Sastry, A. N. 1979. Pelecypoda (excluding ostreidae). In Reproduction of Marine Invertebrates. A. C. Giese and J. S. Pearse, Ed. Acad. Press. Vol. V:113—292. Thompson, I., D. S. Jones & J. W. Ropes. 1980a. Advanced age for sexual maturity in the ocean quahaug, Arctica islandica (Mollusca:Bi- valvia). Mar. Biol. (Berl.) 57:35—39. Thompson, I., D. S. Jones & D. Dreibelbis. 1980b. Annual internal growth banding and life history of the ocean quahaug, Arctica islan- dica (Mollusca:Bivalvia). Mar. Biol. (Berl.) 57:25—34. von Oertzen, J. A. 1972. Cycles and rates of reproduction of six Baltic Sea bivalves of different zoogeographical origin. Mar. Biol. (Berl.) 14:143-149. Journal of Shellfish Research, Vol. 9, No. 1, 205-213, 1990. MICROSTRUCTURE OF THE OUTER SHELL LAYER OF RANGIA CUNEATA (SOWERBY, 1831) FROM THE DELAWARE RIVER: APPLICATIONS IN STUDIES OF POPULATION DYNAMICS LOWELL W. FRITZ, LISA M. RAGONE,! AND RICHARD A. LUTZ New Jersey Agricultural Experiment Station Rutgers University Shellfish Research Laboratory P.O. Box 687 Port Norris, NJ 08349 ABSTRACT The outer shell layer of Rangia cuneata (Sowerby) is composed of two types of crossed-lamellar (CL) microstructure: rapid-growth, deposited in spring and fall, and slow-growth, deposited in summer, which correspond to seasons of relatively rapid and slow shell growth. The two types of CL microstructure were distinguishable on the basis of the: (1) width of 1° lamellae; (2) angle of deposition of 1° lamellae with respect to the inner shell surface; and (3) amount of interdigitation of 2° lamellae between adjacent 1° lamellae. In winter, outer layer CL microstructure is replaced by one or a series of prismatic sublayers, counts of which were used to determine age. Shell length at each winter prismatic band (annulus) was measured directly on intact valves and used to determine size at age. The bimodal distribution of shell length at the first annulus and peaks in size-specific dry-tissue weight in spring and fall suggest that the sampled population in the Delaware River spawns twice each year. At any sampling time, fall recruits were approxi- mately S—10 mm smaller in shell length than spring recruits of the same year-class. Von Bertalanffy growth equations were calculated separately for spring and fall recruits and revealed that age-specific growth rates of the two groups were almost identical. KEY WORDS: Rangia cuneata, shell, microstructure, age determination, growth INTRODUCTION The wedge clam, Rangia cuneata (Sowerby), is a common inhabitant of brackish waters in coastal areas of the Gulf of Mexico and the Atlantic coast of the United States (Hopkins and Andrews 1970, Abbott 1974). An ex- tension of its range north from Chesapeake Bay to Dela- ware Bay was recently reported (Counts 1980). Mass mor- talities of R. cuneata have been attributed to cold winter temperatures in the northern portion of its range which may limit its distribution further north (Gallagher and Wells 1969). The species has moderate commercial importance in the southern portions of its range (Wolfe and Petteway 1968), but little is known of its growth and reproductive biology north of Chesapeake Bay. Studies of populations in Potomac River, Maryland (Pfitzenmeyer and Drobeck 1964) suggest a late summer-fall spawning period and growth to between 35—45 mm shell length in approxi- mately 4 years. A similar size/age relationship was reported by Wolfe and Petteway (1968) for populations of R. cu- neata in the Neuse River, North Carolina. Fairbanks (1963) reported that Rangia spawns twice each year, in spring and fall, in Louisiana and grows to a length of approximately 30 mm in 3 years. However, all previous reports of R. cu- neata size at age have been based on analyses of growth lines on the shell exterior, which can be unreliable guides to individual age (see Lutz and Rhoads 1980). In this study, specimens of Rangia cuneata were collected from a site in the Delaware River to determine the seasonal growth pat- ‘Present Address: Virginia Institute of Marine Science, Gloucester Point, VA 23062. tern in shell microstructure for use in age determination and, potentially, environmental impact assessment. The shell of Rangia cuneata, like all species in the family Mactridae, is composed of three primary carbonate layers (Fig. 1), which are: (1) an inner complex crossed-la- mellar layer; (2) a thin pallial myostracum, to which the mantle is attached; and (3) an outer crossed-lamellar layer (Taylor et al. 1973). Growth patterns in the outer layer proved to be more valuable than those in the inner layer for reconstruction of growth history. The outer shell layer, de- posited by regions of the mantle ventral to the pallial myo- stracum, is composed of first-order (1°), second-order (2°), and third-order (3°) lamellae. First-order lamellae are de- posited concentrically, approximately parallel to the ventral shell margin. Second-order lamellae within each 1° lamella are arranged much like shingles on a roof, with their angle of deposition (with respect to the inner shell surface) in one 1° lamella alternating with respect to those in adjacent 1° lamellae. Second-order lamellae are themselves composed of 3° lamellae which are generally poorly defined, small, lath- or rod-like crystalline units joined at their sides (see Taylor et al. 1969, 1973, Carter 1980 and Dieth 1985 for further discussion of shell microstructural terminology). The outer shell layer is also elaborated into a series of mi- crogrowth increments (visible in polished and etched sec- tions), which parallel the inner shell surface and may repre- sent alternating periods of shell deposition and dissolution (Lutz and Rhoads 1977). MATERIALS AND METHODS Specimens of Rangia cuneata were collected by hand and with a clam rake from one location on the Delaware 205 206 u Figure 1. Line drawing of radial section of a Rangia cuneata shell valve showing the shell macrostructure (ol = outer crossed-lamellar layer; pm = pallial myostracum; il = inner complex crossed-lamellar layer). Growth is to the right and the shell interior is toward the bottom. Approximate distance from the umbo (u) to the ventral margin (vm) is 30 mm. River south of New Castle, DE (39°37'N; 75°36'W). Be- tween 8 and 38 clams were collected on each of 12 dates between November 1985 and March 1987 (total number collected = 248). The population of R. cuneata was lo- cated at a depth of approximately 1.0—1.5 m at mean low water in muddy-sand offshore from submerged rhizome- peat mats. A marsh composed primarily of Phragmites spp. was located immediately onshore from the sampling site. A water sample for salinity determination (with a Guildline Autosal model 8400 salinometer) was collected on each sampling date. Salinity ranged from 0.6 to 5.7 ppt from November 1985 to March 1987, with the highest salinities recorded in Jate summer and fall. Specimens were kept cold in transit to the lab where animals were sacrificed, usually within 4 hours of collec- tion. Soft tissues were removed from the shell, and both the drained wet and freeze-dried total meat weight were mea- sured. Shell morphometric measurements (length— greatest antero-posterior dimension; and height— greatest distance from the umbo to the ventral margin) were ob- tained on each specimen. Shell valves were thoroughly cleaned, individually numbered, rinsed well with fresh water, and air-dried. One of the two valves of between 8—15 specimens from each collection (total sample size = 203) was imbedded in liquid casting plastic and radially sectioned along the height axis using a Raytech 10-inch cir- cular rock saw. Acetate peels of the radial surfaces, pre- pared according to the methods outlined in Fritz and Haven (1983), were analyzed at 40 and 100 x magnification on a compound microscope. Shell growth in specified periods was measured (using a calibrated ocular micrometer) along the surface of maximum growth on the acetate peels, which in Rangia cuneata is the external surface of the outer shell layer (Pannella and MacClintock 1968). Radial fracture shell sections for examination in a scan- ning electron microscope (SEM) were prepared by the gen- eral methods outlined by Kennish et al. (1980). Shell frag- ments from 10 specimens (two each from the collections in November 1985 and March, May, July and September 1986) were glued to aluminum stubs with cyano-acrylate cement and carbon paint, coated with gold-palladium in a FRITZ ET AL. sputter coater, and analyzed at 20 kV accelerating voltage in an Hitachi S-450 SEM. To prepare polished and etched sections for SEM anal- ysis (6 specimens total), imbedded and sectioned spec- imens were polished with diamond grits down to 6 ym grit size and etched in a 0.1 M EDTA solution for between 0.75 and 1.5 minutes. Sections were thoroughly dried (in air), sputter-coated with gold-palladium and analyzed at 20 kV in the SEM. Through analyses of acetate peel replicas of polished and etched shell sections, annually-formed microstructures (formed each winter) were used to locate specific growth rings on the shell exterior surface; shell length at each ring, or annulus, was measured directly off the unimbedded valve of each specimen. Because of the bimodal distribu- tion of shell length at the first annulus across all year- classes, growth rates of the two groups were analyzed sepa- rately. Individual specimens were arbitrarily assigned to ei- ther spring or fall cohorts of each yearclass (YC) if the shell length at the first annulus was greater or less than 16 mm, respectively. Von Bertalanffy growth equations [fit ac- cording to the methods of Walford (1946) and Beverton (1954) as modified by Ricker (1975)] were developed sepa- rately for spring and fall recruits. Von Bertalanffy growth equations have the form: l, = L.[1 = e—K(t—to)) | where |, is the computed shell length (in mm) at time t (in years), L,, is the asymptotic shell length, e is the base of the natural logarithm, K is the growth constant, and tp is the hypothetical age when shell length is equal to zero. Fall recruits were assigned an age of 0.5 years at the first an- nulus, while spring recruits were assigned an age of | year. The von Bertalanffy growth equation was chosen because it has been shown to not only accurately model bivalve growth (e.g. Bachelet 1980; Brousseau 1984), but was used previously to calculate size-at-age data for Rangia cuneata (Wolfe and Petteway 1968). RESULTS Seasonal Microstructure of the Outer Crossed-Lamellar Shell Layer and Age Determination Seasonal differences in the microstructure of the outer crossed-lamellar (CL) layer of Rangia cuneata are shown in Fig. 2. In winter (March 1986), a thin prismatic sublayer (labelled p in Fig. 2A) had replaced the CL microstruc- ture along the inner surface of the outer layer (top portion of Fig. 2A). Along its inner surface, the composition of the outer layer varied between slightly etched prism tips (left portion of Fig. 2B) and prism tips with a thin organic (?) coating (lower right portion of Fig. 2B). Second-order la- mellae immediately above the prismatic sublayer were de- posited almost parallel to the inner surface of the outer layer (defined as a small angle of deposition), and 1° la- SHELL MICROSTRUCTURE OF RANGIA CUNEATA Figure 2. Scanning electron micrographs of the radial fracture (A, C & E) and inner surfaces (B, D & F) of the outer crossed lamellar shell layer of Rangia cuneata collected on 18 March (A & B; p—prismatic sublayer along the inner shell surface), 27 May (C & D) and 31 July 1986 (E & F). In C-F, 1 and 2 denote two adjacent 1° lamellae. In the micrographs of fracture surfaces, the inner shell surface is at the bottom; growth in all micrographs is to the right. Scale bars = 3 ym. mellae are not readily distinguishable (Fig. 2A). This re- gion of the shell was deposited in late fall/early winter and its poor organization and the small deposition angle of 2° lamellae could be a result of cold water temperatures during this period. By contrast, in spring (May 1986), 2° lamellae were an- gled steeply with respect to the inner shell surface resulting in the exposure of broad faces of some 2° lamellae (Fig. 2C). Second-order lamellae terminated along the inner shell surface (Fig. 2D) in a characteristic shingled pattern of lo- bate 2° lamellae tips with their angle of deposition alter- nating in adjacent 1° lamellae. This pattern was also ob- served in specimens younger than 3 years of age collected in fall (September and October 1986), and herein will be termed rapid-growth CL microstructure. First-order lamellae within regions of the outer layer de- posited in summer were approximately twice as wide (dorso-ventrally) as those deposited in spring and fall (Fig. 2C). Summer-deposited 1° lamellae were approximately 10 44m wide, while those deposited in spring and fall ranged between 3 and 6 ym wide. First-order lamellae deposited in summer had more discrete dorsal and ventral edges than those deposited in spring or fall, with less interdigitation of 2° lamellae between adjacent 1° lamellae. Second-order la- mellae in summer-deposited 1° lamellae intersected the inner shell surface at smaller angles than those formed in spring or fall, resulting in a relatively smooth surface com- posed of sharply angled tips of 2° lamellae (Fig. 2F). The pattern observed in summer will herein be termed the slow- growth CL microstructure. Differences in the seasonal microstructure of the outer layer in spring/fall and summer were also evident in SEM and light microscopic analyses of polished and etched ra- dial sections. The etching process revealed both the broad faces of 2° lamellae in rapid-growth CL microstructure de- posited in spring as well as the boundaries of microgrowth increments, which remained as continuous ridges across the outer layer running perpendicular to each 1° lamella (Fig. 3A). Dorsal and ventral edges of 1° lamellae deposited in spring were not readily distinguishable, but were distinct in regions deposited in summer (Fig. 3B). The etchant also more clearly revealed 2° lamella in the slow-growth than in Figure 3. Scanning electron micrographs of polished and etched radial sections of the outer crossed lamellar (CL) shell layer of Rangia cuneata. The inner shell surface is beyond the bottom and growth is to the right in each micrograph. Micrographs were taken approximately midway between the inner and outer surfaces in regions of the outer layer deposited in: (A) spring, rapid-growth CL microstructure; (B) late summer, slow-growth CL microstructure, diagonal lines in both A & B (from upper right to lower left) are microgrowth increment boundaries; (C) fall, rapid- (upper left) and slow- (lower right) growth CL microstructure separated by a growth cessation mark; and (D) winter, prismatic sublayer surrounded by regions of slow-growth CL microstructure. See Figure 4 for representative locations of each micrograph within the outer shell layer. Scale bars = 5 pm. SHELL MICROSTRUCTURE OF RANGIA CUNEATA the rapid-growth CL microstructure, indicating that their angle of deposition with respect to the inner layer was less in the former than in the latter. Periods of little or no growth (possibly on the order of days) were represented in polished and etched radial sec- tions of the outer layer as thick microgrowth increment boundaries, or growth cessation marks (gem; Fig. 3C and 4). The composition of the outer layer changed abruptly at the gem in Fig. 3C, from rapid-growth CL microstructure preceding it (left portion of Fig. 3C) to one resembling slow-growth CL microstructure following it (right portion of Fig. 3C). Prismatic sublayers formed in winter were composed of thin, rod-like columnar sub-units, or prisms, which often had a granular texture (evidence of dissolution; Lutz and Rhoads 1977) on their depositional faces (Fig. 3D). More than one prismatic sublayer could be formed in each of the first two or three winters, with the annulus consisting of a series of closely-spaced prismatic sublayers, each between 3—8 pm thick (Fig. 4). With increasing age, winter was represented most often by a single prismatic sublayer (Fig. 209 3D). The micrograph in Figure 2A was taken on a spec- imen collected in March 1986 and shows a prismatic sub- layer in the midst of formation. Annual outer shell layer increments deposited in 1985 by clams of three different ages are shown in Fig. 4. Winter prismatic sublayers appeared as thick, distinct dark lines surrounded on both sides (dorsal and ventral) by rapid- growth CL microstructure in specimens younger than 3 years. In older specimens (or in annual increments depos- ited when the clam was 3 years or older), prismatic sub- layers were usually preceded (dorsal) by slow-growth CL microstructure formed in fall and followed (ventral) by rapid-growth CL microstructure formed in spring. Regions of rapid and slow-growth CL microstructure appear as light and dark bands, respectively, in acetate peels. Rapid- growth CL microstructure formed a larger percentage of each annual increment in younger (Fig. 4A) than in older (Fig. 4B & C) clams. Furthermore, bands of rapid-growth CL microstructure formed in fall became increasingly smaller with age. Microgrowth increment boundaries are thin dark lines which parallel the inner shell surface. ie 1984 _ MM aa es 1985 TS a el WINTER SPRING SUMMER = S| oe eee B SUMMER FALL Cc SUMMER WINTER SPRING SUMMER = FALL WINTER { sPRING PALL SUMMER Figure 4. Light micrographs of acetate peel replicas of polished and etched radial sections of the outer crossed-lamellar (CL) shell layer of Rangia cuneata. Shell growth is to the right and the outer shell surface is at the top of each section. Regions of the outer layer are labelled by season of deposition in 1984 and 1985. Rapid-growth CL microstructure appears light, while slow-growth CL microstructure appears dark in acetate peels. Microgrowth increment boundaries in the outer layer appear as thin dark diagonal lines, while winter prismatic sublayers are thicker and more distinct. Letters A-D on micrographs indicate representative locations of micrographs in Figure 3 A—D, respectively. Scale bar in A = 500 ym; scale is the same in all three micrographs. A) 1983 yearclass (YC) specimen collected 3 February 1986. The ventral shell margin forms the right side of the micrograph. B) 1982 YC specimen collected 29 April 1986. The ventral shell margin forms the right side of the micrograph. C) 1981 YC specimen collected 26 March 1987. The ventral shell margin is beyond the right side of the micrograph. 210 Growth cessation marks (Fig. 3C & 4) appeared as distinct microgrowth increment boundaries in both rapid- and slow- growth CL microstructure. Winter prismatic sublayers (Fig. 3D) were distinguished from gcm in spring or fall by: (1) the intensity and defini- tion of the line(s) in acetate peels: prismatic sublayers were darker and had more discrete dorsal and ventral boundaries than gcm; and (2) the presence of vertical lamellae sepa- rating each columnar prism, which were often discernible (even at 100 x magnification of acetate peels) within pris- matic sublayers and were rarely present in gem (Fig. 3C & D). Along the exterior shell surface, however, gem and winter prismatic sublayers were very difficult to distinguish from each other. In most cases, but not all, the darkest rings on the shell exterior were associated with gcm’s and not winter prismatic sublayers, and were accompanied by an indentation in the shell exterior surface. Of the 203 specimens sectioned for growth pattern anal- ysis, we were able to determine the age of 185 individuals, or 91% of the total sample. The most difficult annulus to locate in shell section was the first: of the 185 individuals whose age was determined, a specific growth ring on the shell exterior formed during the first winter was measured on 158, or in 85% of the aged sample. Population Studies and Seasonal Growth Rates Five YC’s (1981-85) were represented in the collec- tions from November 1985 through March 1987, with the two most numerous being those recruited in 1982 and 1984. Shell length at the first annulus ranged from 6—25 mm in all specimens analyzed (Table 1). However, the modal shell length at the first annulus of each YC was ei- ther between 8—12 mm, or between 20—23 mm, as shown for the 1982 and 1984 YC’s in Fig. 5. As will be discussed in more detail below in relation to seasonal changes in dry tissue weight, the bimodal length frequency at the first an- nulus strongly suggests that the species has two spawning periods each year. Individuals with larger shell lengths at their first annuli were spawned earlier in the year (i.e. spring) than those with smaller, first annular shell lengths TABLE 1. Mean and range in shell length (mm) of spring and fall recruits of Rangia cuneata at annuli (winter prismatic bands) 1-5. N = number of each annulus identified. Spring and fall recruits were defined as those with shell lengths greater or less than 16 mm, respectively, at the first annulus (see Figure 4). Arianalns Spring Fall Number Mean N Range Mean N Range 1 20.4 55 16.0—25.0 10.6 112 6.0—15.7 2 38.2 41 30.0-47.3 32.6 107. 21.9—-41.0 3 51.8 10 46.3—58.5 47.2 88 35.9-54.9 4 55.8 8 49.6-61.0 52.4 69 41.0-61.9 5 — — _- 54.7 9 44.7-58.0 FRITZ ET AL. Es] 1982, Ye Mm 1984 YC PERCENT S 7 © hh i 1 iW 1 21 2 25 SHELL LENGTH (mm) Figure 5. Percent length-frequency at the first annulus (winter pris- matic sublayer) of the 1982 (N = 93) and 1984 (N = 54) year-classes (YC’s) of Rangia cuneata collected from November 1985 through March 1987. (spawned in fall). The data in Fig. 5 suggest that the fall- spawn resulted in more successful recruitment to the popu- lation in 1982 than the spring-spawn, and vice-versa for 1984. Growth rates of spring and fall recruits were almost identical in each YC, but mean shell lengths of the fall re- cruits through the fifth annulus were always less than those of the spring recruits (Table 1; Fig. 6A). However, there was considerable over-lap in the ranges in shell length at all annuli after the first (Table 1). In the separately derived von Bertalanffy growth equations, L,. was larger and K was smaller for spring than for fall recruits (Table 2). However, with the 0.5 year difference in age, there was little differ- ence in the calculated size-at-age between the two groups (Fig. 6B), suggesting that size- and age-specific growth rates were independent of recruitment time. At any point in time that a sample is obtained, however, the 0.5 year dif- ference in age between spring and fall recruits of the same YC would result in as much as a 5—10 mm difference in shell length. In Delaware River populations of Rangia cuneata, shell growth resumed in mid-May 1986, after a period of little or no growth in winter, and continued through mid-December 1986 (Fig. 7). Growth rates of the 1984 YC were greater in spring (from late May through June) and fall (from mid- September through October) than in summer (July through August). This pattern was also reflected in the microstruc- ture of the outer layer, as described in the previous section, with rapid-growth CL microstructure being formed in spring and fall and slow-growth CL microstructure in summer. With increasing age, not only did total annual growth rates decline (Fig. 6), but growth rates tended to remain low from summer through winter (Fig. 7). In spec- imens older than three years, spring was the only season when shell growth rates were relatively rapid. Spring and fall were also periods of relatively high dry tissue weight compared with winter and summer (Fig. 8). SHELL MICROSTRUCTURE OF RANGIA CUNEATA O----O SPRING DN IFA LE SHELL LENGTH (mm) fo) 1 2 3 4 5 6 ANNULUS NUMBER O SPRING A FALL SHELL LENGTH (mm) AGE (years) Figure 6. A) Mean shell length at each annulus (winter prismatic sublayer) of spring and fall recruits of all Rangia cuneata collected. Data in Table 1. B) Calculated shell lengths of spring and fall recruits of Rangia cuneata according to the von Bertalanffy growth equations described in Table 2. Spring and fall recruits were assigned ages of 1 and 0.5 years, respectively, at their first annuli (winter prismatic sub- layers). Mean dry tissue weights (calculated for a 30 mm SL spec- imen at each sampling date) declined from late 1985 through March 1986, but increased slightly through April 1986. The increase in mean dry tissue weight by late April occurred prior to any detectable shell growth by this time (Fig. 7). Mean dry tissue weight declined to its lowest cal- culated value in late July, which coincided with the ob- served decline in shell growth rates. Through late summer and fall, however, mean dry tissue weights increased rap- idly to the highest calculated value in late October. Peaks in mean dry tissue weight in both spring and fall suggest that this population spawned twice in 1986. DISCUSSION Replacement of crossed-lamellar (CL) by prismatic mi- crostructures during periods of little or no shell growth (dormancy) has been observed previously in other bivalves [in Arctica islandica (L.) (Lutz and Rhoads 1980, Ropes et al. 1984); in Corbicula fluminea (Miller) (Fritz et al. 1987); and in Astarte elliptica (Brown) (Trutschler and Samtleben 1988)]. Prismatic microstructures in shells are TABLE 2. Calculated parameters of von Bertalanffy growth equations and calculated shell lengths (SL; mm) at age (years) for spring and fall recruits of Rangia cuneata. L,, = asymptotic shell length; K = growth constant; ty) = hypothetical age at which clam had zero length. Spring Fall L. = 66.2 60.6 K = 0.519 0.604 to = 0.273 0.189 Age SL 0.5 10.4 1.0 20.8 1.5 33.1 2.0 39.2 2.5 45.6 3.0 50.1 3.5 52.4 4.0 56.6 4.5 56.1 5.0 60.5 5.5 58.1 located at sites of attachment of the soft tissues to the shell (pallial and adductor myostraca) or may also result, within other shell layers, from periods of reduced oxygen tensions within the tissues (Lutz and Rhoads 1980). Prolonged pe- riods of valve closure (which are thought to accompany pe- riods of dormancy) during adverse environmental condi- tions can result in the use of anaerobic metabolic pathways by bivalve molluscs (Crenshaw and Neff 1969, Lutz and Rhoads 1977). In A. islandica, for instance, prismatic bands within the inner shell layer may result from extended periods of anaerobiosis associated with valve closure during its burrowing activities (Taylor 1976, Lutz and Rhoads 1980). In Delaware River populations of Rangia cuneata, no measurable shell growth occurred in winter, ] 204 fe 164 O 1984 YC E o~ = a = 124 d S | ro) ] (oy: = 84 Be = O | eg | 1982 YC | n | ' i 4 ’ fe —f—-o + x I II file Go o-@9 a -— ) FMA TM J dA Ss © Figure 7. Shell growth (mean and range) by members of the 1982 and 1984 yearclasses (YC’s) of Rangia cuneata from the winter 1985-86 prismatic sublayer to the ventral shell margin (= the date of collec- tion). Specimens were collected on ten dates between January 1986 and March 1987. 212 0.5 0.44 = 0.34 q if eae payee : pe Ala i | kewl [Th / 0.0 aewats.! ——1— ONDJFMAMJJIASONOD J FMAM 1986 1987 Figure 8. Calculated total dry tissue weight (mean + 95% confidence limits) of a 30 mm (shell length) specimen of Rangia cuneata on twelve dates between November 1985 and March 1987. Calculated dry tissue weights were obtained from a series (one for each collection date) of least-squares linear regression equations of log-transformed dry tissue weight on log-transformed shell length. which was also the period during which one or more pris- matic sublayers replaced the CL microstructure in the outer shell layer. At no other time of the year were prismatic sublayers observed in the outer shell layer, which enabled the use of counts of prismatic sublayers to determine age and reconstruct growth history. Wada (1972) and Dieth (1985) reported that crystalline units formed during periods of slow shell growth tended to be larger and be deposited with greater order (in which the edges of units are more clearly defined) than those formed during periods of rapid shell growth. Results of the present study, in which slow-growth CL microstructure was com- posed of larger, more ordered 1° lamellae than rapid- growth CL microstructure, support their conclusions. Measured and computed shell lengths at each age were between 10—25 mm greater in the present study of Dela- ware River populations than in studies of southern popula- tions of Rangia cuneata. Fairbanks (1963) used external growth lines to determine shell length at age in populations of R. cuneata in Louisiana, while Wolfe and Petteway (1968) determined growth rates and size-at-age from changes in modal size frequency in sequential samples from Neuse River, North Carolina populations. Both methods could have underestimated growth rate due to in- accuracies in interpreting: (1) external growth lines (distin- guishing growth cessation marks from annually-deposited FRITZ ET AL. lines); and (2) changes in length-frequency distributions over time (given the size differences in spring- and fall-re- cruits of the same yearclass at any point in time). The pos- sibility that growth rates of R. cuneata are greater in the Delaware River than in southern population is unlikely since this is the most northern populations on the east coast of North America and shell growth is limited to only the period from May through mid-December each year. Analyses of the microstructure of the outer shell layer of Rangia cuneata revealed the growth history of individual specimens. In the event of an environmental disturbance, such growth histories could provide some of the *‘after-the- fact’’ data (pre- and post-disturbance seasonal and annual growth rates, time of growth cessation mark formation, size-at-age, changes in outer shell layer microstructure, for example) necessary to quantify the sub-lethal effects of the perturbation. Furthermore, growth rates can be standard- ized with respect to differences associated with age and/or season of recruitment for unbiased estimates of the distur- bance’s effects on the growth of a resident bivalve species. The shell growth patterns of three other bivalve species, one living in freshwater (Corbicula fluminea; Fritz and Lutz 1986) and the other two from estuarine habitats (Mya arenaria (L.); MacDonald and Thomas 1982) and Merce- naria mercenaria (L.); Kennish and Olsson 1975) have been used previously to document the effects of chronic and episodic environmental disturbances. Analyses of the microstructural shell growth patterns of R. cuneata could provide a powerful tool for monitoring environmental change and assessing the impact of perturbations in the oli- gohaline portion of estuaries. ACKNOWLEDGMENTS We thank Lisa Wargo for preparing the plates of micro- graphs; Tim Jacobsen, Margaret Schenk, Ya-Ping Hu, and Ray Grizzle for assistance in collecting and working-up specimens; John Grazul and Richard Triemer, Bureau of Biological Research, Rutgers University, for assistance in scanning electron microscopy; and two anonymous re- viewers for their contributions to this work. New Jersey Agricultural Experiment Station Publication No. D-27204- 1-89 supported by state funds and grants from the New Jersey Department of Environmental Protection, Division of Science and Research. LITERATURE CITED Abbott, R. T. 1974. American Seashells: The Marine Mollusca of the Atlantic and Pacific Coasts of North America. (2nd ed.) New York, NY: Van Nostrand Reinhold Co. 663 p. Bachelet, G. 1980. Growth and recruitment of the tellinid bivalve Ma- coma balthica at the southern limit of its geographical distribution, the Gironde Estuary (SW France). Mar. Biol. 59:105—117. Beverton, R. J. H. 1954. Notes on the use of theoretical models in the study of the dynamics of exploited fish populations. U.S. Fish. Lab., Beaufort, N.C., Misc. Contrib. 2:1—159. Brousseau, D. J. 1984. Age and growth rate determinations for the At- lantic ribbed mussel, Geukensia demissa Dillwyn (Bivalvia: Myti- lidae). Estuaries 7:233-241. Carter, J. G. 1980. Environmental and biological controls of bivalve shell mineralogy and microstructure. In Rhoads, D. C. & R. A. Lutz, eds., Skeletal Growth of Aquatic Organisms. New York and London: Plenum Press, pp. 69-113. Counts, C. L. III. 1980. Rangia cuneata in an industrial water system (Bivalvia: Mactridae). The Nautilus 94:1—2. SHELL MICROSTRUCTURE OF RANGIA CUNEATA 21 Crenshaw, M. A. & J. M. Neff. 1969. Decalcification at the mantle-shell interface in molluscs. Am. Zool. 9:881—885. Deith, M. R. 1985. The composition of tidally deposited growth lines in the shell of the edible cockle, Cerastoderma edule. J. Mar. Biol. Ass. U.K. 65:573-S81. Fairbanks, L. D. 1963. Biodemographic studies on the clam Rangia cu- neata Gray. Tulane Stud. Zool. 10(1):3—47. Fritz, L. W. & D. S. Haven. 1983. Hard clam, Mercenaria mercenaria: Shell growth patterns in Chesapeake Bay. Fishery Bulletin 81:697— 708. Fritz, L. W. & R. A. Lutz. 1986. Environmental perturbations reflected in internal shell growth patterns of Corbicula fluminea (Mollusca: Bi- valvia). The Veliger 28(4):401—417. Fritz, L. W., L. M. Ragone, & R. A. Lutz. 1987. Utilization of bivalve shells for assessment of environmental stress. I. Corbicula fluminea. Rept. to NJ Dept. Env. Prot., Div. Science Research Contract No. C-29526. 63 p. Gallagher, J. L. & H. W. Wells. 1969. Northern range extension and winter mortality of Rangia cuneata. The Nautilus 83(1):22—25. Hopkins, S. H. & J. D. Andrews. 1970. Rangia cuneata on the East coast: thousand mile range extension or resurgence? Science 167:686. Kennish, M. J.,R. A. Lutz & D. C. Rhoads. 1980. Preparation of acetate peels and fractured sections for observations of growth patterns within the bivalve shell. In Rhoads, D.C. & R. A. Lutz, eds., Skeletal Growth of Aquatic Organisms. New York and London: Plenum Press, pp. 597-601. Kennish, M. J. & R. K. Olsson. 1975. Effects of thermal discharges on the microstructural growth of Mercenaria mercenaria. Environ. Geol. 1:41-64. Lutz, R. A. & D. C. Rhoads. 1977. Anaerobiosis and a theory of growth line formation. Science 198:1222—1227. Lutz, R. A. & D. C. Rhoads. 1980. Growth patterns within the molluscan shell: an overview. In Rhoads, D.C. & R. A. Lutz, eds., Skeletal Ww Growth of Aquatic Organisms. New York and London: Plenum Press, pp. 203-254. MacDonald, B. A. & M. L. H. Thomas. 1982. Growth reduction in the soft-shell clam Mya arenaria from a heavily oiled lagoon in Cheda- bucto Bay, Nova Scotia. Marine Environmental Research 6:145—156. Pannella, G. & C. MacClintock. 1968. Biological and environmental rhythms reflected in molluscan shell growth. J. Paleontol. 42:64—80. Pfitzenmeyer, H. T. & K. G. Drobeck. 1964. The occurrence of the brackish water clam, Rangia cuneata in the Potomac River, Maryland. Chesapeake Science 5(4):209—212. Ricker, W. E. 1975. Computation and interpretation of biological sta- tistics of fish populations. Bull. Fish. Res. Board Canada No. 191. 382 p. Ropes, J. W., D. S. Jones, S. A. Murawski, F. M. Serchuk & A. Jerald, Jr. 1984. Documentation of annual growth lines in ocean quahogs, Arctica islandica Linne. Fishery Bulletin 82(1):1—19. Taylor A. C. 1976. Burrowing behavior and anaerobiosis in the bivalve Arctica islandica (L.). J. Mar. Biol. Assoc. U.K. 56:95—109. Taylor, J. D., W. J. Kennedy & A. Hall. 1969. The shell structure and mineralogy of the Bivalvia: Introduction, Nuculacea-Trigoacea. Bull. Br. Mus. (Nat. Hist.) Zool., Suppl. 3, 125 pp. Taylor, J. D., W. J. Kennedy & A. Hall. 1973. The shell structure and mineralogy of the Bivalvia: II. Lucinacea—Clavagellacea. Conclu- sions. Bull. Br. Mus. (Nat. Hist.) Zool. 22(9):253—294. Trutschler, K. & C. Samtleben. 1988. Shell growth of Astarte elliptica (Bivalvia) from Kiel Bay (Western Baltic Sea). Mar. Ecol. Prog. Ser. 42:155—162. Wada, K. 1972. Nucleation and growth of aragonite crystals in the nacre of some bivalve molluscs. Biomineralisation 6:141—159. Walford, L. A. 1946. A new graphic method of describing the growth of animals. Biol. Bull. 90(2):141—147. Wolfe, D. A. & E. N. Petteway. 1968. Growth of Rangia cuneata Gray. Chesapeake Science 9(2):99- 102. Journal of Shellfish Research, Vol. 9, No. 1, 215—225, 1990. ANNUAL SHELL BANDING, AGE, AND GROWTH RATE OF HARD CLAMS (MERCENARIA SPP.) FROM FLORIDA DOUGLAS S. JONES,! IRVY R. QUITMYER,! WILLIAM S. ARNOLD? AND DAN C. MARELLI2 ‘Florida Museum of Natural History University of Florida Gainesville, FL 32611 Florida Marine Research Institute 100 8th Avenue, S.E. St. Petersburg, FL 33701-5095 ABSTRACT Year-round collection of hard clams (Mercenaria mercenaria and M. campechiensis) from three coastal sites in Florida permits documentation of the annual cycle of shell growth increment formation in these bivalve species in the southern portion of their tange. This cycle consists of alternating, macroscopic light increments (opaque white to beige) and dark increments (translucent grey to blue to purple) that are visible in the inner, middle, and outer shell layers of radially sectioned valves. These increments form seasonally; the light increment, characteristic of rapid shell growth, forms primarily during the spring in M. campechiensis from the Gulf Coast and during the winter in M. mercenaria from the northeastern Atlantic Coast of Florida. Over the remainder of the year, and particularly during the late summer and fall, the dark, slow-growth increment is added. The annual shell increments were used to determine age and growth rates in 10 clam populations from both the Atlantic and Gulf coasts of the State of Florida. Hard clam growth was then modeled using the von Bertalanffy growth function which facilitated the comparison of growth between the Florida sites as well as with other, more northerly populations. The greatest growth was observed in populations of Mercenaria campechiensis from Boca Ciega Bay, where the largest known specimens have been reported. The oldest specimens (28 years old) came from the Cedar Key area, which also yielded large clams. Gulf Coast populations of M. campechiensis did not always exhibit greater growth than Atlantic Coast M. mercenaria, and clear latitudinal gradients in shell growth were not evident on either coast. In fact, growth variations between populations from collection sites within a single estuary occasion- ally exceeded those among sites on separate coasts. The growth rates measured for Florida hard clams are typically greater than those reported for clams from New England, the Middle Atlantic Bight, and the southeastern United States, although the life span of the Florida clams is apparently shorter. Our results provide an assessment of hard clam growth parameters in Florida where relatively little data were heretofore available for comparison and where commercial exploitation of the resource has increased dramatically in recent years. KEY WORDS: Mercenaria, hard clam, growth, shell banding, Florida INTRODUCTION Possibly no other bivalve mollusk has been the focus of as much research into shell growth increment formation as Mercenaria mercenaria (Linné), the northern quahog or hard clam. Throughout the hard clam’s geographic range in eastern North America, from the Gulf of Mexico to Cape Cod with isolated populations extending into Maine and Canada (Ansell 1968), its shell increments have been studied. Cyclical microgrowth patterns ranging in scale from subdaily tidal, through daily, fortnightly tidal, lunar monthly, and annual have been described in this species (e.g., Barker 1964; Pannella and MacClintock 1968; Rhoads and Pannella 1970; Cunliffe 1974; Kennish and Olsson 1975; Thompson 1975; Gordon and Carriker 1978; Kennish 1980, 1984). In addition to these cyclical patterns, an entire spectrum of growth breaks, some periodic (e.g., annual spawning) and some stochastic (e.g., storms), is potentially interpretable from hard clam shell records (Ken- nish and Olsson 1975). Among this plethora of shell increments and growth breaks, annual, incremental banding patterns have proven to be the most consistent and readily interpretable (Lutz and 21 Nn Rhoads 1980). Annual increments have received wide ap- plication in marine ecological contexts, particularly in de- termining age and growth rates. Annual increments have been identified in shells of Mercenaria mercenaria throughout most of its geographic range (e.g., New En- gland—Rhoads and Pannella 1970; Jones et al. 1989; New York/New Jersey—Kennish and Olsson 1975; Ropes 1987; Grizzle and Lutz 1988; Maryland/Virginia—Fritz and Haven 1983; North Carolina—Peterson et al. 1983, 1985; Georgia—Clark 1979; Quitmyer et al. 1985). The annual pattern of hard clam shell growth consists of macroscopic, alternating light and dark increments in the inner, middle, and outer shell layers, best viewed in radial sections of either valve (Fig. 1). Recent work on the sea- sonal timing of formation of these increments (e.g., Clark and Lutz 1982; Peterson et al. 1983; Grizzle and Lutz 1988) suggests there are geographic differences in annual shell growth patterns, reflective of seasonal differences in ambient water temperature fluctuations. Of particular im- portance are summer and winter extremes (Lutz and Rhoads 1980). In comparing hard clams from North Caro- lina and New Jersey, Clark and Lutz (1982) observed that, , Translucent increment 4 Annual growth cycle Opaque increment \OQuter shell layer \Middle shell layer ‘Inner shell layer Ventral r f margin 5cm Figure 1. Reflected-light photograph of radial shell cross-sections of M. campechiensis (top), shell height = 104.66 mm, collected 10/14/88 at Cedar Key, Florida showing translucent (dark) growth increment forming at ventral margin and M. mercenaria (bottom), shell height = 67.38 mn, collected 5/23/82 at Kings Bay, Georgia showing opaque (light) growth increment forming at ventral margin. M. campechiensis specimen shows 10 annual shell growth increment cycles whereas M. mercenaria specimen shows 7+ years of growth. Lower schematic of radial shell section indicates annual growth cycle and structural ele- ments of shell. ‘*_ . features characteristic of winter in one locality can occur in summer in the other.’’ Recognizing that the varia- tions may be considerable and that a latitudinal gradient in season of growth increment formation may exist, Grizzle and Lutz (1988) have emphasized the need for documenta- tion and description of such patterns. The purpose of this investigation was two-fold. First, we wanted to examine the seasonal cycle and verify the annual periodicity of macroscopic growth increment formation in hard clam populations from Florida, including both the northern quahog, Mercenaria mercenaria, and the southern quahog, Mercenaria campechiensis (Gmelin). Compar- isons of these patterns with those described for populations to the north should verify the existence of latitudinal differ- ences. Relative to M. mercenaria, much less is known about seasonal shell increment patterns, age, and growth rate in M. campechiensis, making the need for such infor- mation even more acute. Our second objective was to de- termine, using annual growth increments, the age and growth relationships for populations of both species from around the state. Such information is significant for several reasons: 1) growth rate and longevity estimates are gener- ally lacking for both coasts of Florida; therefore, intra-re- JONES ET AL. gional as well as inter-regional comparisons with other hard clam populations are not presently possible; 2) appreciation of growth rates near the southern limit of M. mercenaria will further elucidate the often cited influence of tempera- ture upon the growth of these animals (Ansell 1968); and 3) potential growth differences and similarities between M. mercenaria and M. campechiensis may be assessed. The recent and dramatic rise in shellfishing pressure upon these species in Florida, combined with the insufficient baseline growth data presently available and the burgeoning hard clam aquaculture industry in Florida, make these growth studies all the more important and timely. MATERIALS AND METHODS Hard clams were collected from nine sites located on both the Atlantic and Gulf coasts of Florida (Fig. 2). A tenth site in Boca Ciega Bay was analyzed from published data (Saloman and Taylor 1969). Hard clam collections (total N = 1,578) were assembled and analyzed by either the Gainesville authors (Kings Bay, N = 451; Matanzas River, N = 60; Bokeelia and Catfish Creek in Charlotte Harbor, N = 399; and Cedar Key, N = 259) or the St. Petersburg authors (Indian River Body C, N = 153; and Body F, N = 149; Tampa Bay, N = 50; and St. Joseph Bay, N = 57). Most specimens were located by treading in shallow water and were collected manually. However, those from the Indian River lagoon were hand-raked in Au- gust 1986, and those from St. Joseph Bay were mechani- cally harvested by Mr. Harry Lawder during March 1987. At three of these sites (Kings Bay, Charlotte Harbor, and Cedar Key), year-round hard clam collections were made at FLORIDA ‘Alligator Harbor Atlantic COLLECTION SITES Ocean Kings Bay, Georgia Matanzas River (Crescent Beach) Indian River (Body C) Indian River (Body F) Bokeelia (Charlotte Harbor) . Catfish Creek (Charlotte Harbor) Boca Ciega Bay Tampa Bay (Skyway Sand Flat) Cedar Key (Suwannee Reef) St. Joseph Bay St. Petersburg SCHOMNAH EWN = Gulf of Mexico 2° Figure 2. Map of Florida indicating hard clam collection localities. Site numbers are also identified in Table 1. GROWTH OF FLORIDA HARD CLAMS monthly intervals to investigate the annual cycle of growth increment formation. Specimens from the remaining sites were collected after one (Matanzas River—June 1988) or repeated visits to the site (Skyway Bridge —October 1986, June 1987; March 1988). Following transport back to the laboratory, each clam was eviscerated (usually with the aid of a microwave oven, which prevents shell damage), and the valves were washed, dried, numbered, and stored for later analysis. All hard clams were prepared using similar techniques (see Rhoads and Lutz 1980). Each shell pair was disarticu- lated and one valve, usually the left, was appropriated. Clams from St. Joseph Bay, Tampa Bay, and the Indian River were embedded in Epon 815 epoxy resin prior to sec- tioning. When fully cured, the embedded valve was ra- dially sectioned along the axis of maximum growth (max- imum shell height, greatest distance from umbo to ventral margin) to reveal the growth increment record. This cut was made using a Highland Park Model 20SSP lapidary saw equipped with a diamond saw blade. The Gainesville group did not embed their shells. Sectioning was accom- plished on a Lortone Model FS lapidary trim saw outfitted with an 8-inch diamond saw blade. In either case, the smooth cuts did not normally require further polishing or the preparation of acetate peels (Rhoads and Lutz 1980) in order to distinguish and measure the annual increments. In those cases where further polishing was required, 240-, 400-, and 600-grit emery papers were used in succession, followed by 1.0-micron alumina microgrit applied with a high-speed rotating lapidary wheel. There was little difficulty identifying the first 5 to 10 annual shell increments in most specimens. Thereafter, crowding together of increments as ontogenetic shell growth rates declined occasionally resulted in the inability to uniquely distinguish between successive years, particu- larly in older specimens. If polishing and preparation of acetate peels could not resolve the situation, the shell was not used. Approximately 5% of the sectioned specimens were discarded. To determine the season of increment formation, a total of 451 hard clams were analyzed from Kings Bay, Georgia, near the northeasternmost corner of Florida. These clams were collected at monthly intervals (approximately 20 per month) for two year-long periods, from August 1981 to July 1982 and again from November 1983 to November 1984. The results of the 1981—1982 study have been pub- lished as a modern comparative data set for an archaeolog- ical investigation of aboriginal seasonal clam harvesting in this region (Quitmyer et al. 1985). The second year of col- lection was undertaken to gauge interannual variability. Two analogous collections were recently completed on the Gulf Coast of Florida in areas were large coastal shell middens and seasonal aboriginal habitation are of archaeo- logical interest. Monthly collections were made at two lo- calities within Charlotte Harbor (near Bokeelia and at the 217 mouth of Catfish Creek) from March 1986 to February 1987. Bokeelia and Catfish Creek samples were pooled for analysis (N = 399). Similarly, monthly collections were made at Cedar Key (Suwannee Reef) from December 1987 to November 1988 (N = 259). In contrast to the Kings Bay study, southern quahogs, Mercenaria campechiensis, were recovered from these sites. Specimens were collected during ebb tide and generally were from subtidal habitats, except for a few individuals from exposed sand bars at Bokeelia and Kings Bay. The substrates at each site were variable. Typically, muddy sand was dominant with shell hash present at all sites and seagrass beds common in Charlotte Harbor and at Cedar Key. Temperature and salinity were recorded concurrently with the monthly clam collections at all three sites. The ventral shell margins of specimens collected monthly at the three sites involved in the seasonal banding study (Kings Bay, Charlotte Harbor, and Cedar Key) were examined in detail. In order to assess the annual pattern of growth increment formation, each sectioned shell was as- signed to one of two categories, T or O. This assignment depended on whether the translucent (dark) or opaque (white) increment (as viewed in reflected light on the shell cross-section) was forming at the edge when the clam was captured (Fig. 1). By noting the percentages of clams in both of these categories on a monthly basis throughout the year, an annual pattern was described. For the growth comparison study, a random subsample of 50 clams was selected from each population for analysis. This procedure served to standardize the sample sizes of all populations. To the nine stations actually visited in this study, a tenth was added based upon age and growth-rate data for 93 large specimens of Mercenaria campechiensis from Boca Ciega Bay reported by Saloman and Taylor (1969). These authors also interpreted age by counting the internal shell growth increments, considered to form an- nually. With the naked eye or occasionally with the aid of a low-power binocular microscope, the ventral (distal) edge of each annual growth increment was marked with a sharp pencil at the point where it intersected the outer shell sur- face. Counting the total number of annual increments (pairs of light and dark bands) yielded the age of each specimen. The shell height at each age within a given shell was deter- mined by measuring from the umbo to each successive pencil mark, so that a complete shell height versus age record (growth curve) was produced for every specimen. For consistency, we assumed that the first dark band in each shell represented an age of one year, and constructed our age profiles accordingly. In actuality, the time repre- sented by the first dark band is almost always less than one year; it can vary from place to place as well as from year to year and depends upon the season in which the clams were spawned. Measurements were performed with electronic, digital 218 calipers (MAX-CAL, Fred V. Fowler Co., Newton, Mas- sachusetts) that read to 0.01 mm and were configured to an IBM-PC microcomputer. INCAL, a general purpose data entry and digital caliper drive program for the IBM-PC, was used to create and store data files which were later transferred to LOTUS 1-2-3 for analysis. To facilitate comparison between regions, hard clam growth was modeled by fitting a von Bertalanffy growth function to the shell height-age data. This function is de- scribed by the following equation: SH, = SH,[1 — e~K¢-%] where t = time (or age in years), SH, = shell height at t, SH,, = maximum asymptotic shell height, k = growth constant, and ty) = time when SH, = 0. The von Berta- lanffy function was fit to the data using the NLIN proce- dure of SAS (1985). This iterative curve-fitting procedure employs nonlinear least-squares regression (multivariate secant method) and yields parameter values, estimates of their asymptotic standard errors, and an asymptotic correla- tion matrix of the parameters. The von Bertalanffy growth function has traditionally received wide application in the analysis of bivalve growth (e.g., Brousseau 1979; Gallucci and Quinn 1979; Apple- doorn 1983; Schick et al. 1988; Tanabe 1988; Jones et al. 1989). However, it is not the only function used to describe molluscan growth (e.g., Peterson and Black 1987). In fact, Kennish and Loveland (1980) reported that ontogenetic growth in Mercenaria mercenaria from Barnegat Bay, New Jersey, was best described by the Gompertz equation. In coastal Georgia, Walker and colleagues have used the power function to describe hard clam growth (Walker 1984; Walker and Humphrey 1984; Walker and Tenore 1984). To insure that the von Bertalanffy growth curve was the most appropriate for our study, four other commonly used functions (see Kaufmann 1981), including the Gom- pertz, logistic, exponential, and power curve (with and without intercept), were also fit to the data. As in an analo- gous study of hard clam growth in Rhode Island (Jones et al. 1989), the von Bertalanffy curve provided the best fit (highest R* values) and was used throughout the remainder of the investigation. The single growth parameter, w (= k X SH..), along with its variance, was calculated from the von Bertalanffy parameter estimates for each sample site according to the methods of Gallucci and Quinn (1979). Since w is more robust than either k or SH., this parameter offers a pow- erful and straightforward way of comparing organism growth curves between regions (Gallucci and Quinn 1979; Appledoorn 1980, 1983). The tg variable is basically a po- sition parameter. It does not affect the growth rate compar- isons and is not considered here. The w parameter corre- sponds to the growth rate near ty and is suitable for compar- isons of the compound null hypothesis Hp:k; = k, =... JONES ET AL. = k, and SH.., = SH... =. . . = SH. forn regions. A w value and its 95% confidence interval were calculated ac- cording to the method of Appledoorn (1980) for each sample site. It was then possible to rank the w values and ascertain which of the sample sites were statistically dif- ferent (P < 0.05) from one another by noting whether or not their confidence intervals overlapped. This procedure is straightforward, easily interpretable, and conservative. RESULTS The seasonal cycle of annual shell growth increment formation in Mercenaria mercenaria from Kings Bay, Georgia, and in M. campechiensis from Charlotte Harbor and Cedar Key, Florida, shows the same general pattern at all three sites, with some apparent differences in timing (Fig. 3). Monthly examinations of growing shell margins indicate that at Kings Bay, the episode of most rapid shell growth occurs throughout the winter, when the opaque (light) growth increment is added. At any give time within this period (December— March), 70—80% of the population is forming an opaque increment. A transition occurs during April/May when the percentages are reversed and a greater proportion of the hard clam population is characterized by having the translucent (dark) increment at the shell margin. Following this transition, from June through October, vir- tually 100% of the specimens were forming translucent in- crements. The annual shell growth pattern then comes full cycle in November/December when another transition occurs and clams with opaque marginal increments pre- dominate. The analysis of Mercenaria campechiensis from Gulf Coast localities indicates a similar pattern with a temporal shift. In Charlotte Harbor the opaque increment is added primarily during the spring; this period of rapid shell growth is centered around the month of April and is appar- ently briefer here than it is at Kings Bay. During the re- mainder of the year, and particularly from July through De- cember, nearly 100% of the clam population is forming the translucent increment. At Cedar Key the pattern is very similar, but the distinc- tions between periods of opaque and translucent increment formation are not as sharp. As illustrated in Figure 3, the episode of opaque growth increment formation is of longer duration but is still principally a spring phenomenon. It ap- pears to begin slightly earlier (November/December) and extend slightly longer (into July). As the proportion of specimens with opaque marginal increments declines in the early summer, the percentage of clams forming translucent increments increases. This percentage reaches a maximum during the months of August, September, and October, when nearly 100% of the specimens collected were charac- terized by having the translucent marginal increment. During the succeeding months, from November through February, approximately 60-75% of the population was GROWTH OF FLORIDA HARD CLAMS 219 N: 37 29 27 30 47 30 46 39 56 33 46 31 / 451 100 50 T ~ oO ~ oF ~ a oO a = % O KINGS BAY S Ea o ic £ : = o 50 O = 100 JFMAMJJASOND JFMAMJJASOND MONTH MONTH N: 25 18 20 14 18 25 20 27 21 20 22 22 / 259 100 50 as T a ya ° ~ aed a oO a S % 0 CEDAR KEY 34 B F3 c 2 = = oO o (ep) 50 Oo F 100 JFMAMJJASOND MONTH MONTH N: 38 27 47 39 30 22 29 41 33 43 18 32 / 399 100 50 es T 5 T ° ~ = a @o a CHARLOTTE 5 is % O © = &S HARBOR 2 = @o = 2 = oO E o 50 fe) re 100 JFMAMJJASOND LeMAMWd Jd AB OW Lf MONTH MONTH Figure 3. Annual cycle of growth increment formation in populations of M. campechiensis from Charlotte Harbor and Cedar Key, Florida and M. mercenaria from Kings Bay, Georgia, based on monthly collections of specimens. Vertical bars represent the percentage of the monthly sample forming either the translucent (T, dark) or opaque (O, light) growth increment. Accompanying temperature and salinity values were measured concurrently with hard clam collections. 220 still forming the translucent increment, whereas the re- mainder, mostly in the smaller, faster-growing size classes, had begun to form the opaque increment. Comparison of the yearly shell cycle at any of these three sites where the temperature data were collected simul- taneously (Fig. 3) reveals that the translucent increments are added during the warmest portion of the year, whereas the opaque increment is more characteristically formed during cooler conditions. This pattern is particularly evi- dent in the Kings Bay and Charlotte Harbor data but is somewhat less clear-cut in the Cedar Key data. At none of the sites is the relationship so precise as to suggest that the changeover occurs at a particular temperature. Comparison of the shell cycle with the salinity data does not indicate a direct relationship, except possibly in the case of Kings Bay, where the temperature and salinity profiles are posi- tively correlated. The results indicate that annual growth increment pat- terns are present in these two species in Florida and that they are valid age indicators in hard clams living in this portion of their range. Using these increments, the shell height for each year of growth was determined for SO hard clams for each sample site. A von Bertalanffy growth curve was then fit to the data for each site according to the proce- dures outlined earlier. A curve was also fit to mean shell height versus age data for 93 hard clams reported by Sal- oman and Taylor (1969). These curves are shown in Fig. 4 for all ten sample sites, including both the Atlantic (Mercenaria mercenaria) and Gulf (Mercenaria campe- chiensis) coasts. The von Bertalanffy parameter estimates and their asymptotic standard errors for each site are listed in Table 1. Also listed in Table 1 are the ranked values of w for each sample site and a 95% confidence interval about each w. These w values and confidence intervals are graphically displayed (Fig. 5) to enhance appreciation of variability. Inspection of the w values (Table 1, Fig. 5), as well as the growth curves shown in Figure 4, indicates that the null hypothesis of identical growth properties for hard clams at each sample station should be rejected. The greatest w value was associated with the hard clams from Boca Ciega Bay reported by Saloman and Taylor (1969). This site has produced the largest recorded specimens of Mercenaria campechiensis (Sims 1964). The growth curve for this site plots well above all the others, indicating gigantic final sizes (SH..) produced by rapid initial growth rates that re- main high throughout the first decade or so of life. After Boca Ciega Bay, the next highest w values were from Cat- fish Creek in Charlotte Harbor and Tampa Bay, respec- tively. The k value or growth constant determined for the former site was the largest encountered in this study, whereas both the k and SH,, parameters at the latter site were relatively large. A 95% confidence interval could not be calculated about the w value for Boca Ciega Bay be- cause of the nature of the data reported by Saloman and Taylor (1969); however, this highest w value falls within JONES ET AL. the 95% confidence interval for Catfish Creek, indicating statistically inseparable (P < 0.05) w values. In order of descending rank, the next four w values are associated with populations of Mercenaria mercenaria from the Atlantic Coast of Florida: Matanzas River, Indian River Body C and Body F, and Kings Bay. The values of the Matanzas River population and both populations from the Indian River are not statistically different at the 0.05 level. The w value for the Kings Bay population, while separable at the 0.05 level, is nevertheless quite similar to the other east coast stations. Clearly, there is a greater de- gree of homogeneity among the von Bertalanffy growth pa- rameters associated with these hard clam populations than there is between the populations on the west coast of Florida (Fig. 5). The lowest three w values were determined for Merce- naria campechiensis populations at Bokeelia, situated at the northern end of Pine Island in Charlotte Harbor; for St. Joseph Bay on the Florida Panhandle; and for Suwannee Reef, located just north of Cedar Key. As indicated in Table 1, hard clams for Bokeelia achieve final sizes that are similar to those attained by clams from Catfish Creek, also located within Charlotte Harbor. The difference in the w values between these two sites results from the higher k at the latter site. This difference within one estuary exceeds that observed between clams from either Bokeelia or Cat- fish Creek and clams on the Atlantic Coast (Table 1). The SH,, determined for the St. Joseph Bay clams was the lowest for any of the Gulf Coast M. campechiensis popula- tions and is largely responsible for the low w. In fact, the growth curve for this area (Fig. 4) illustrates that the size versus age relationship for these clams was much more similar to that of Atlantic Coast M. mercenaria than to their Gulf Coast counterparts. Finally, the lowest w was ob- tained for the Cedar Key population. Though this locality produced large clams (second only to Boca Ciega Bay), the growth constant k was clearly the lowest encountered in this study. Because k is largely responsible for the curva- ture of the von Bertalanffy function (Gallucci and Quinn 1979), the Cedar Key growth curve (Fig. 4) appears much straighter than the other nine curves. Whereas their growth rates were the lowest, the oldest measurable clams in this study (28 years) came from the Cedar Key locality, followed by St. Joseph Bay (23 years), Boca Ciega Bay (22 years), and Bokeelia, Tampa Bay and Indian River Body F (20 years). The oldest specimens at the remaining stations were 13 years old at capture (Figure 4). For several reasons discussed in the following section, these age determinations necessarily represent minimum estimates of maximum age and should be carefully inter- preted in the context of longevity. DISCUSSION Annual shell growth increment patterns have been re- ported in Mercenaria from New England to the south- GROWTH OF FLORIDA HARD CLAMS 221 200 190 180 170 160 150 140 130 120 110 100 SHELL HEIGHT (mm) rT 12°35 4 5 6 7 8) 9 1011 12 13 14°15 16 17 18 19) 20°21) 22°23: 24 25:26:27 28 29/30 AGE (years) Figure 4. Best fit von Bertalanffy growth curves relating shell height and age for each of the 10 sites investigated in this study. Site numbers are identified in Fig. 2 and Table 1, as are the von Bertalanffy parameter estimates. Both M. mercenaria (1-4) and M. campechiensis (5-10) are represented here. eastern United States. The ‘‘classic’’ interpretation of the dark increments as winter, slow-growth phenomena arose from studies of northern populations (e.g., Kerswill 1941; Pannella and MacClintock 1968; Rhoads and Pannella 1970). However, Kennish and Olsson (1975) showed that high as well as low temperatures can cause growth-rate in- hibition and induce shell increment formation. Recent studies have emphasized geographic differences in this basic pattern that appear to have a latitudinal component. Therefore, it is important to compare the shell increment patterns described herein for Florida hard clams with those described elsewhere for other localities. Clark and Lutz (1982) described incremental growth patterns for Mercenaria mercenaria from sites extending from Maine to Georgia. These authors stated that features characteristic of winter in one locality can occur in summer in the other. Clark (1979) reported that hard clams from coastal Georgia formed the translucent increment during times of warm water temperatures; growth slowed during the summer and fall and was fastest during winter and spring. He also observed that the summer growth halt, “*. . . fits the data reported by Ansell (1968), who shows that Mercenaria mercenaria has a winter growth halt in waters from Virginia to Canada, and a summer growth halt from North Carolina to Florida.’’ Peterson et al. (1983) suggested the dark increment formed between May and October, whereas the light increment formed from No- vember to April. This pattern is very similar to that ob- served here for the Florida hard clams (Fig. 3), except that in Florida the interval of translucent increment formation is extended by a month or two. This may be a real difference or it may reflect the particular year in which sampling oc- curred. In the present study, temporal trends in annual shell increment formation were assessed for one-year intervals (except for Kings Bay, where two years’ data were com- bined) so that the average natural cycle might be shifted to the extent that the particular year of observation was atyp- ical. A similar pattern was reported by Fritz and Haven (1983) for clams from Chesapeake Bay, Virginia, and more recently by Grizzle and Lutz (1988) for those in New Jersey, except that in both of these cases, a second dark increment often formed in the winter. Thus, dark incre- ments have been reported from northern specimens as oc- curring both in summer and winter. The summer dark in- crement is apparently wider and bounded in spring and fall by white increments reflecting rapid growth phases. The winter dark increment, when present, is apparently much narrower and is perhaps better described as a dark ‘break’ (Grizzle and Lutz 1988). No such winter breaks were re- 919)9) JONES ET AL. TABLE 1. Best fit von Bertalanffy growth curve parameter estimates and asymptotic standard errors (parentheses) for each sample location. Sample site numbers refer to Fig. 2. Ranked values of w (= SH.. x k, see Gallucci and Quinn, 1979) with 95% confidence intervals indicate spectrum of growth variation and possible statistical differences (p < 0.05) between sites. Sample Sample Site # Location n ty Mercenaria mercenaria 1 Kings Bay (Southern Georgia) 50 — 0.45 (0.16) 2 Matanzas River (Crescent Beach) 50 0.07 (0.09) 3 Indian River (Body C) 50 0.15 (0.09) 4 Indian River (Body F) _50 0.28 (0.07) Total/Mean 200 0.05 Mercenaria campechiensis 5 Bokeelia (Charlotte Harbor) 50 — 1.15 (0.21) 6 Catfish Creek (Charlotte Harbor) 50 —0.13 (0.11) dis Boca Ciega Bay (near Tampa Bay) 93 0.09 (0.08) 8 Tampa Bay (Skyway Sand Flat) 50 — 0.23 (0.07) 9 Cedar Key (Suwannee Reef) 50 — 1.05 (0.13) 10 St. Joseph Bay (Panhandle) _50 — 1.04 (0.14) Total/Mean 343 =0'59 Rank k SH. ro) (oy 95% CI 0.38 (0.04) 72.38 (1.75) 27.50 7 26.85—28.15 0.43 (0.03) 71.83 (1.42) 30.89 4 30.21—31.57 0.38 (0.04) 80.51 (3.08) 30.59 5 29.99—31.19 0.35 (0.03) 85.86 (1.92) 30.05 6 29.48—30.62 0.39 77.65 29.76 0.25 (0.02) 90.19 (1.96) 22.55 8 22.16-22.94 0.49 (0.05) 89.77 (1.74) 43.99 2 42.93—45.05 0.25 (0.01) 177.80 (1.03) 44.45 1 — 0.39 (0.02) 105.19 (0.69) 41.02 3 40.49—41.55 0.11 (0.01) 134.98 (2.44) 14.85 10 14.55-15.15 0.28 (0.02) 77.36 (0.70) 21.66 9 21.27—22.05 0.30 112.55 31.42 * Curve fit to mean size-age data from Saloman and Taylor (1969), confidence intervals not calculable. ported for North Carolina clams by Peterson et al. (1983), nor were they encountered in the Florida samples. The seasonality of shell growth observed in this study matches that determined by R. W. Menzel (1963, 1964; 45 40 35 25 20 15 1 2 3 4 5 6 vy Gh ©) 19) SITE NUMBER Figure 5. Plot of » values and 95% confidence intervals (vertical bars) for M. mercenaria (M) and M. campechiensis (C) from each col- lection site in this study, contrasting intermediate and tightly-clus- tered values for Atlantic Coast M. mercenaria with highly variable values for M. campechiensis from the Gulf Coast. Site numbers are identified in Table 1 and Fig. 2. Menzel and Sims 1964) nearly 25 years ago for hard clams from Alligator Harbor, Florida (Fig. 2). In the late 1950s and early 1960s, growth of transplanted Mercenaria mer- cenaria and local M. campechiensis, as well as their hy- brids, was investigated by experimental plantings and peri- odic measurement of clams at this Gulf Coast site. Apart from the growth curve presented by Saloman and Taylor (1969) for the unusually large clams from Boca Ciega Bay, these early studies represent the only previously published data on hard clam growth in Florida. The results suggest that M. mercenaria grew fastest in spring and late fall, less in winter, and slowest in summer. M. campechiensis exhib- ited a similar pattern but apparently grew slowest in winter. Menzel (1963) felt that Mercenaria spp. from Florida ex- hibited greater annual growth than at other sites from Maine to the southeastern U.S. because of continued growth throughout winter. He also speculated that M. cam- pechiensis might be more vigorous with a faster innate growth rate. The data of Ansell (1968) and others (e.g., Loosanoff 1939; Pratt and Campbell 1956; Hamwi 1969; Walne 1972) argue forcibly for the role of temperature as a major influ- ence on growth and physiological activity of Mercenaria. This idea has found recent support in an investigation of long-term hard clam growth in Narragansett Bay (Jones et al. 1989). Ansell (1968) reported an optimum growth tem- perature for M. mercenaria of 20°C. Growth declined sym- metrically above and below this value and ceased below 9°C and above 31°C. No evidence of a geographical trend was reported. Within temperature limits, other factors such as food availability and substrate character may often determine the actual rate of growth of hard clams (Ansell 1968; Pratt GROWTH OF FLORIDA HARD CLAMS 223 1953; Pratt and Campbell 1956; Eversole 1987). Recent data by Peterson and Fegley (1986) suggest that juvenile and adult hard clams may grow at different size-adjusted rates throughout portions of the year. Anomalously low adult relative to juvenile growth rates during the winter may not result solely from cold temperatures, but from dif- ferences in resource partitioning related to adult gameto- genic activity in spring. Clearly, temperature is not the only factor involved. Nevertheless, given these caveats, it is in- teresting to note (Fig. 3) that the lowest temperature mea- sured during our year-round monitoring study was 10°C (January, Kings Bay and Cedar Key), whereas the highest was 32°C (September, Cedar Key). The bulk of our tem- perature data and a perusal of unpublished, longer-term water temperature variations for Florida indicate that the temperature limits for growth in Mercenaria mercenaria cited by Ansell (1968) are seldom and only briefly ex- ceeded. Thus, the generally high growth potential for these populations at the southern extreme of their range is not totally unexpected. Ansell (1968) used Menzel’s Florida data in his geo- graphic survey of hard clam growth. Ansell concluded, however, that there were no regional trends in the values of shell growth parameters and that the growth/age curves for sites within the continuous range from the Gulf Coast of Florida to Massachusetts were similar. A standardized, four centimeter clam, whether from Florida or New England, had approximately the same maximum annual growth in- crement. In Figure 5 of Ansell (1968), length versus age relationships are plotted for Mercenaria mercenaria from many localities. The curve for Florida indicates that most rapid size increase with age although it is not appreciably higher than the curves for other sites. To better compare this growth relationship of Ansell (1968) with the ones de- termined here for the ten Florida sites, we converted the shell length data from Ansell’s Figure 5 to shell height based upon the average height:length ratio (0.91) for our samples and then fit a von Bertalanffy curve to the data. The resulting parameters, k = 0.34 and SH, = 82.94, yielded a w = 28.41. This value is almost identical to the w values reported for the Atlantic Coast Mercenaria mer- cenaria populations from Kings Bay, Matanzas River, and Mercenaria spp. from the Indian River (Body C and Body F) and is intermediate with respect to the w values for the M. campechiensis stations (Table 1; Fig. 5). Despite living for two or three decades, most of the sig- nificant size increase in the Florida hard clams occurs during the initial several years of life (Fig. 4). Thereafter, the rate of increase declines progressively with age. This relationship, modeled here and in Rhode Island (Jones et al., in press) by the von Bertalanffy growth function and elsewhere by the Gompertz equation (Kennish and Love- land 1980) or a power function (e.g., Walker 1984), has been observed in hard clams throughout their distribution. When the growth curves for Mercenaria mercenaria shown in Figure 4 are compared to those assembled by Ansell (1968, Fig. 5) for his ‘*best growth’’ North American sites (and shell lengths are converted to heights), the Florida sites all plot above the others. This suggests that growth of Florida hard clams is indeed more rapid, as Menzel (1963) hypothesized. It would facilitate inter-regional comparison of growth curves if modeled growth parameters were available for each region. In a recent study of hard clam growth in Nar- ragansett Bay, Rhode Island, the w parameter of Gallucci and Quinn (1979) was used to expedite growth compar- isons around the Bay (Jones et al. 1989). The w values determined for Narragansett Bay ranged between 11.33 and 21.87, with a mean of 15.40. The hard clam growth in Narragansett Bay was judged comparable to that recently reported by Ropes (1987) for northern New Jersey. Both of these sites, as the w values indicate, fall near the bottom of the spectrum of growth determined for the Florida clams (Table 1; Fig. 5). A great deal of variability surrounds growth estimates in Florida as well as in Rhode Island, but the idea that real geographic differences in growth of Mer- cenaria mercenaria exist and that growth in Florida is among the highest appears well supported. The growth of M. campechiensis can exceed that of M. mercenaria but it apparently does not in all cases (Fig. 5). The maximum age encountered in the Florida hard clam samples is significantly less than that reported elsewhere. For example, Jones et al. (1989) reported two specimens from Narragansett Bay that were 40 years old at the time of capture. Lutz and Haskin (1985) described two marked- and-recaptured specimens from New Jersey that were 36 and 33 years old. Peterson (1983, 1986) reported old indi- viduals, living up to 46 years, from North Carolina and Hopkins (1941) as well as Ropes (personal communication) felt that a 75-year longevity may be attained in hard clams. In contrast, the oldest specimen measured in this study came from Cedar Key and was 28 years old when captured. The next oldest specimen also came from Cedar Key and was 25 years old. Overall, less than 1% of all clams in this study attained an age of 25, only 4% lived to age 20, and 12% had survived to age 15 when captured. Longevity de- terminations for the Florida clams should be considered minimum estimates for two reasons: 1) the animals were still alive at the time of capture; and 2) the +5% of the clams that were discarded from analysis because of diffi- culty uniquely interpreting their shell records were invari- ably old individuals with growth increments crowded at their margins and were potentially the oldest specimens. Nevertheless, none appeared to approach the ages of the oldest specimens reported from northern poopulations. CONCLUSIONS Both Mercenaria mercenaria and M. campechiensis from coastal Florida form annual shell increments which can be utilized in age and growth rate determinations. The yearly incremental growth pattern consists of macroscopic, alternating dark (translucent) and light (opaque) increments 224 best viewed in radial shell cross-sections. In M. mercenaria from Florida’s Atlantic coast, the light increment, repre- senting episodes of rapid shell growth, forms primarily during the winter. In M. campechiensis from the Florida Gulf Coast, the light increment is added principally during the spring. The translucent or dark, slow-growth increment forms over the remainder of the year, especially the late summer and fall. No winter dark increments were observed in the Florida specimens and the annual shell cycle is sim- ilar to that described for hard clams from North Carolina and Georgia. Using annual shell increments as age markers, shell height versus age relationships were investigated for ten sites around Florida, including both coasts. Measured shell growth was modeled with the von Bertalanffy equation and the w parameter of Gallucci and Quinn (1979). Growth was found to be highly variable both within and between local- ities. The greatest growth was observed in populations of Mercenaria campechiensis from Boca Ciega Bay where the largest known living specimens of hard clams occur. En- hanced growth also was encountered at Catfish Creek in Charlotte Harbor on the Gulf Coast. Intermediate growth characterized populations of M. mercenaria from the At- lantic Coast whereas lower w values were associated with M. campechiensis from Bokeelia (Charlotte Harbor), St. Joseph Bay, and Cedar Kay. While a wide spectrum of hard clam growth was observed, comparison with growth data for populations from New England and the Middle At- JONES ET AL. lantic Bight confirmed earlier suggestions of overall higher annual growth rates in Florida. This may be related to fa- vorable temperature regimes which permit growth throughout the winter months. The use of annual shell growth increment patterns in age determination has generally had the effect of increasing es- timates of hard clam longevity. The oldest specimen en- countered in this investigation was 28 years old at the time of capture, far younger than the oldest reported elsewhere. Only 4% of the Florida specimens had attained the age of 20, even though they were often of considerable size. It appears that both species of Mercenaria from Florida grow faster than their counterparts to the north but do not seem to live as long. ACKNOWLEDGMENTS This paper is dedicated to the memory of a good friend and scientist, John W. Ropes. We thank B. J. MacFadden for help with computing and data manipulation and R. L. Walker and F. J. Maturo for helpful discussions. L. Newsom, M. Russo, N. Borremans, C. Romanek, D. Haveard, B. Knight, H. Coulter, C. Haven, P. Gill, D. Hesselman, and H. Lawder assisted in collecting hard clams. Parts of this work were supported by the Associates of The Florida Museum of Natural History, the U.S. De- partment of the Navy (Contract #N00025-79-C-0013), the National Science Foundation (BNS-8519814), and funds generated from hard clam license fees. REFERENCES CITED Ansell, A. D. 1968. The rate of growth of the hard clam Mercenaria mercenaria (L) throughout the geographical range. J. Cons. Perm. Int. Explor. Mer. 31:364—409. Appledoorm, R. S. 1980. The growth and life-history strategy of the soft- shell clam, Mya arenaria L. Ph.D. dissertation, University of Rhode Island. 115 pp. Appledoom, R. S. 1983. Variation in the growth rate of Mya arenaria and its relationship to the environment as analyzed through principal com- ponents analysis and the w parameter of the von Bertalanffy equation. Fish. Bull. 81:75-84. Barker, R. M. 1964. Microtextural variation in pelecypod shells. Malaco- logia 2:69-86. Brousseau, D. J. 1979. Analysis of growth rate in Mya arenaria using the von Bertalanffy equation. Mar. Biol. 51:221—227. Clark, G. C., II. 1979. Seasonal growth variations in shells of Recent and prehistoric specimens of Mercenaria mercenaria from St. Catherines Island, Georgia. Anthropol. Pap. Am. Mus. Nat. Hist. 56:161—172. Clark, G. C., I] & R. A. Lutz. 1982. Seasonal patterns in shell micro- structure of Mercenaria mercenaria along the U.S. Atlantic coast. Abstr. Prog., Geol. Soc. America 14(7):464. Cunliffe, J. E. 1974. Description, interpretation, and preservation of growth increment patterns in shells of Cenozoic bivalves. Ph.D. dis- sertation, Rutgers University. 171 pp. Eversole, A. G. 1987. Species profiles: life histories and environmental requirements of coastal fishes and invertebrates (South Atlantic)— hard clam. U.S. Fish and Wildlife Service Biological Report 82(11.75), U.S. Army Corps of Engineers, TR EL-82-4:1—33. Fritz, L. W. & D. S. Haven. 1983. Hard clam, Mercenaria mercenaria: shell growth patterns in Chesapeake Bay. Fish. Bull. 81:697—708. Gallucci, V. F. & T. J. Quinn, II. 1979. Reparameterizing, fitting, and testing a simple growth model. Trans. Am. Fish. Soc. 108:14—25. Gordon, J. & M. R. Carriker. 1978. Growth lines in a bivalve mollusk: subdaily patterns and dissolution of the shell. Science 202:519—521. Grizzle, R. E. & R. A. Lutz. 1988. Descriptions of macroscopic banding patterns in sectioned polished shells of Mercenaria mercenaria from southern New Jersey. J. Shellfish Res. 7:367—370. Hamwi, A. 1969. Oxygen consumption and pumping rate of the hard clam Mercenaria mercenaria L. Ph.D. dissertation, Rutgers University. 185 pp. Hopkins, H. S. 1941. Growth rings as an index of age in Venus merce- naria. Ant. Rec. 8\(Suppl.):53—54. Jones, D. S., M. A. Arthur & D. J. Allard. 1989. Sclerochronological records of temperature and growth from shells of Mercenaria merce- naria, Narragansett Bay, Rhode Island. Mar. Biol. 102:225—234. Kaufmann, K. W. 1981. Fitting and using growth curves. Oecologia 49:293-299. Kennish, M. J. 1980. Shell microgrowth analysis: Mercenaria merce- naria as a type example for research in population dynamics, p. 255-294. In: Skeletal growth of aquatic organisms. D. C. Rhoads & R. A. Lutz, eds. Plenum Press, New York. Kennish, M. J. 1984. The use of shell microgrowth patterns in age deter- minations of the hard clam, Mercenaria mercenaria (Linné), p. 143-164. In: Invertebrate models in aging research. D. H. Mitchell & T. E. Johnson, eds. CRC Press, Boca Raton, Florida. Kennish, M. J. & R. E. Loveland. 1980. Growth models of the northern quahog, Mercenaria mercenaria (Linné). Proc. Natl. Shellfish. Assoc. 70:230—239. Kennish, M. J. & R. K. Olsson. 1975. Effects of thermal discharges on GROWTH OF FLORIDA HARD CLAMS the microstructural growth of Mercenaria mercenaria. Envir. Geol. 1:41-64. Kerswill, C. J. 1941. Some environmental factors limiting growth and distribution of the quahog, Venus mercenaria L. Ph.D. dissertation, University of Toronto. 104 pp. Loosanoff, V. L. 1939. Effect of temperature upon shell movements of clams, Venus mercenaria (L.). Biol. Bull. (Woods Hole) 76:171—182. Lutz, R. A. & H. H. Haskin. 1985. Some observations on the longevity of the hard clam Mercenaria mercenaria (Linné). (Abstract of tech- nical paper presented at National Shellfisheries Association Annual Meeting, June 25—28, 1984, Tampa, FL) J. Shellfish Res. S(1):39. Lutz, R. A. & D. C. Rhoads. 1980. Growth patterns within the molluscan shell: an overview, p. 203-254. In: Skeletal growth of aquatic or- ganisms. D. C. Rhoads & R. A. Lutz, eds. Plenum Press, New York. Menzel, R. W. 1963. Seasonal growth of the northern quahog, Merce- naria mercenaria and the southern quahog, M. campechiensis, in Alli- gator Harbor, Florida. Proc. Natl. Shellfish. Assoc. 52:37—46. Menzel, R. W. 1964. Seasonal growth of northern and southern quahogs, Mercenaria mercenaria and M. campechiensis, and their hybrids in Florida. Proc. Natl. Shellfish. Assoc. 53:111—119. Menzel, R. W. & H. W. Sims. 1964. Experimental farming of hard clams, Mercenaria mercenaria, in Florida. Proc. Natl. Shellfish. Assoc. 53:103—109. Pannella, G. & C. MacClintock. 1968. Biological and environmental rhythms reflected in molluscan shell growth. J. Paleontol. Mem. 42(Suppl. 5):64—80. Peterson, C. H. 1983. A concept of quantitative reproductive senility: ap- plication to the hard clam, Mercenaria mercenaria (L.)? Oecologia 58:164—168. Peterson, C. H. 1986. Quantitative allometry of gamete production by Mercenaria mercenaria into old age. Mar. Ecol. Prog. Ser. 29:93- Mle Peterson, C. H. & R. Black. 1987. Resource depletion by active suspen- sion feeders on tidal flats: Influence of local density and tidal eleva- tion. Limnol. Oceanogr. 32:143—166. Peterson, C. H., P. B. Duncan, H. C. Summerson & B. F. Beal. 1985. Annual band deposition within shells of the hard clam, Mercenaria mercenaria: consistency across habitat near Cape Lookout, North Carolina. Fish. Bull. 83:671—677. Peterson, C. H., P. B. Duncan, H. C. Summerson & G. W. Safrit, Jr. 1983. A mark-recapture test of annual periodicity of internal growth band deposition in shells of hard clams, Mercenaria mercenaria, from a population along the southeastern United States. Fish. Bull. 81:765— 779. Peterson, C. H. & S. R. Fegley. 1986. Seasonal allocation of resources to growth of shell, soma, and gonads in Mercenaria mercenaria. Biol. Bull. (Woods Hole) 171:597—610. nN Nm nN Pratt, D. M. 1953. Abundance and growth of Venus mercenaria and Cal- locardia morrhuana in relation to the character of bottom sediments. J. Mar. Res. 12:60—74. Pratt, D. M. & D. A. Campbell. 1956. Environmental factors affecting growth in Venus mercenaria. Limnol. Oceanogr. 1:2—17. Quitmyer, I. R., H. S. Hale & D. S. Jones. 1985. Paleoseasonality deter- mination based on incremental shell growth in the hard clam, Merce- naria mercenaria, and its implications for the analysis of three south- east Georgia coastal shell middens. Southeastern Archaeol. 4:27—40. Rhoads, D. C. & R. A. Lutz. 1980. Skeletal growth of aquatic or- ganisms. Plenum Press, New York. 750 pp. Rhoads, D. C. & G. Pannella. 1970. The use of molluscan shell growth patterns in ecology and paleoecology. Lethaia 3:143—161. Ropes, J. W. 1987. Age and growth, reproductive cycle, and histochem- ical tests for heavy metals in hard clams, Mercenaria mercenaria, from Raritan Bay, 1974—75. Fish. Bull. 85:653—662. Saloman, C. H. & J. L. Taylor. 1969. Age and growth of large southern quahogs from a Florida estuary. Proc. Natl. Shellfish. Assoc. 59:46— SI: SAS Institute Inc. 1985. SAS user’s guide: statistics, version 5 edition. SAS Institute Inc. Cary, North Carolina. 956 pp. Schick, D. F., S. E. Shumway & M. A. Hunter. 1988. A comparison of growth rate between shallow water and deep water populations of scallops, Placopecten magellanicus (Gmelin, 1791), in the Gulf of Maine. Am. Malacolog. Bull. 6:1-8. Sims, H. W., Jr. 1964. Large quahog clams from Boca Ciega Bay. Quart. J. Florida Acad. Sci. 27:348. Tanabe, K. 1988. Age and growth rate determinations of an intertidal bivalve, Phacosoma japonicum, using internal shell increments. Lethaia 21:231—241. Thompson, I. 1975. Biological clocks and shell growth in bivalves, p. 149-162. In: Growth rhythms and the history of the Earth's rotation. G. D. Rosenberg & S. K. Runcorn, eds. John Wiley and Sons. London. Walker, R. L. 1984. Effects of density and sampling time on the growth of the hard clam, Mercenaria mercenaria, planted in predator-free cages in coastal Georgia. The Nautilis 98:114—119. Walker, R. L. & C. M. Humphrey. 1984. Growth and survival of the northern hard clam Mercenaria mercenaria (Linné) from Georgia, Virginia, and Massachusetts in coastal waters of Georgia. J. Shellfish. Res. 4:125—129. Walker, R. L. & K. R. Tenore. 1984. The distribution and production of the hard clam, Mercenaria mercenaria, in Wassaw Sound, Georgia. Estuaries 7:19—27. 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. Journal of Shellfish Research, Vol. 9, No. 1, 227-231, 1990. SETTLEMENT OF BIVALVE SPAT ON ARTIFICIAL COLLECTORS IN A SUBTROPICAL EMBAYMENT IN QUEENSLAND, AUSTRALIA WAYNE D. SUMPTON, IAN W. BROWN AND MICHAEL C. L. DREDGE Queensland Department of Primary Industries Southern Fisheries Centre P.O. Box 76, Deception Bay Queensland, Australia, 4508 ABSTRACT Two types of spat collectors, one containing open weave polypropylene bags and the other, monofilament gill netting, were used to collect bivalve spat in Hervey Bay, Queensland. Mimachlamys gloriosa Reeve, Pinctada fucata (Gould) and Pinna bicolor Gmelin were caught in large numbers, often in excess of 200 per collector. During the time of deployment neither collector was effective at collecting spat of Amusium balloti (Bernardi) and only eight spat of that species were collected. Other pectinids collected included Mimachlamys leopardus (Reeve) and Decatopecten decatopecten strangei (Reeve). Settlement of all spat was greatest in collectors 6—12 m off the sea bed, although the size of spat collected was homogenous with respect to depth. KEY WORDS: Mimachlamys, Pinctada, Pinna, spat INTRODUCTION Collection of naturally produced pectinid spat on artifi- cial collectors has been successfully undertaken for nu- merous species in various locations around the world and forms a part of scallop cultivation and growth programmes in many countries (Ito and Byakuo 1989). The settlement of temperate species such as Chlamys opercularis (L.), Pecten maximus (L.), Placopecten magellanicus (Gmelin) on artificial collectors has been extensively studied in the northern hemisphere (Naidu 1970, Brand et al. 1980, Paul 1981). These species showed seasonal and depth related variations in their settlement patterns and maximum settle- ment occurred on collectors 4—10 m off the sea bed. In Australian waters research has also focussed predomi- nantly on temperate species such as Pecten alba Tate (Sause et al. 1987) and Pecten meridionalis Tate (Dix and Sjardin 1975). Little is known about the suitability of artifi- cial collectors for collecting more tropical species such as the saucer scallop Amusium balloti. This latter species in- habits coastal waters of northern Australia and supports valuable trawl fisheries in Western Australia (Heald and Caputi 1981, Joll 1989) and Queensland. Recent declining catch rates (Dredge 1988) in the Queensland saucer scallop fishery has led to increased interest in methods of en- hancing the wild fishery and developing as yet unexploited scallop resources. This paper reports the results of trials designed to deter- mine the suitability of artificial collectors for collecting subtropical pectinid species. Information on the settlement of other bivalve species is also presented. MATERIALS AND METHODS Two types of spat collectors were used in this study. Type 1: two polypropylene open weave bags (5 mm mesh onion bags one inside the other) containing 300—400 g of tN weathered nylon monofilament gill netting. Type 2: two polypropylene onion bags filled with eight other loosely packed onion bags. All the bags were 70 x 35 cm and 5 mm mesh size. Sets of collectors were deployed at four stations in Hervey Bay at sites where scallops had previously been fished commercially (Fig. 1). Collectors were deployed during the period August 1987 to January 1988, which cor- responds with the spawning and settlement season of A. balloti (Dredge, 1981). Monofilament collectors (Type 1) were deployed at Stations 1 and 3 and collectors containing onion bags (Type 2) at Stations 2 and 4. At each station, six collectors were each placed 3 m apart on a buoyed and an- chored line with the first collector 3 m from the sea bottom. Water depth at Stations 1—3 was 20 m while there was 33 m of water at Station 4. The first set of collectors were deployed on 17 August 1987 and recovered on 16 Sep- tember 1987. A further two sets were deployed on 16 Sep- tember and 19 November 1987 and were subsequently re- covered on 19 November and 18 January 1988 respec- tively. These three periods are hereafter referred to as late winter, spring and early summer respectively. Spat collecting bags were frozen on board the recovery vessel and returned to the laboratory. After thawing, the majority of spat were removed by agitating the bags and netting in large tubs of fresh water. The remaining (mainly attached) spat were subsequently removed by checking the washed bags under a binocular microscope. Shell heights of all spat were then recorded. RESULTS Six species of Pectinidae, namely Mimachlamys glo- riosa, Mimachlamys leopardus, Decatopecten decatopecten strangei, Amusium balloti, Pecten fumatus (Reeve) and Complicachlamys dringi wariana Iredale, were found on Di 228 Australia Figure 1. Hervey Bay showing sites where collectors were deployed. spat collectors. Identification of these species has been based upon Iredale (1939) and Woodburn (1989). The latter two species were relatively rare (<5 per collector) and will not be discussed further. Two other bivalves namely Pinna bicolor Gmelin and Pinctada fucata (Gould) which were abundant in the collectors were identified and measured but other bivalves were not identified. Comparatively few bivalve spat settled during late winter (August-September) and over 90% of those recov- ered were less than 2 mm in length. Both Mimachlamys TABLE 1. Average number of spat collected in Hervey Bay during three collection periods. Mean Number of Spat Per Collector Late Early Species Winter Spring Summer Amusium balloti 0 0.4 0.1 Mimachlamys gloriosa 31.4 598.5 40.5 Mimachlamys leopardus 0.2 34.9 6.5 Decatopecten decatopecten strangei 0.3 ZO 39.5 Pinctada fucata N/R N/R 128.1 Pinna bicolor N/R N/R 142.2 Late winter: 17 August—16 September Spring: 16 September— 19 November Early summer: 19 November—18 January N/R: Not recorded SUMPTON ET AL. gloriosa and Mimachlamys leopardus were more abundant on collectors during spring (September—November) with an average of 599 and 35 per collector respectively (Table 1). Decatopecten decatopecten strangei were more com- mon during early summer (November—January). Mean 20 Mimachlamys leopardus > a JO ©. 5 z Ww > CG 10 Ww x LL R 5 fe) 5 10 15 20 25 SIZE (mm) 20 Decatopecten d. strangei > oO z Ww > ie} Ww x Ww R 5 10 15 20 25 SIZE (mm) 20 Mimachlamys gloriosa > oO z Ww =) ie} Ww x Ww RX 5 10 15 SIZE (mm) Figure 2. Shell height frequency of (A) M. leopardus, (B) D.d. strangei and (C) M. gloriosa collected on artificial collectors in Hervey Bay during 16 September-19 November 1987 (@) and 19 November 1987-18 January 1988 (Mf). SETTLEMENT OF BIVALVE SPAT ON COLLECTORS 229 lengths of D.d. strangei and M. gloriosa were significantly (P < 0.05) greater in summer than in spring (Fig. 2). This was most evident in D. strangei, the mean size of which increased from 5.9 to 10.7 mm. The difference in mean size between spring (8.0 mm) and summer (8.9 mm) col- lections of M. leopardus was not significant (P < 0.05). For each of the three species of scallop, collectors set near the surface and close to the sea floor trapped less than half the number of spat as did those in mid-water (9—12 m above the bottom) (Fig. 3). Throughout the study only Mimachlamys leopardus oO ate re eh vA 8 ao Seat PROPORTION OF SPAT 3 6 9 12 15 18 DISTANCE FROM SEA BOTTOM (m) Figure 3. Proportion of M. leopardus, D.d. strangei and M. gloriosa collected on artificial collectors set at various depths during Sep- tember—November at sites 1-3. eight Amusium balloti spat (size range 4—11 mm) were found in collectors. All of these were unattached and taken from collectors in the middle depth range. The mean size of M. leopardus ranged from 7.8 mm in collectors closest to the sea surface to a maximum of 8.9 mm in collectors 12 m off the sea bottom (Table 2), but analysis of variance showed that differences in mean size of scallops with depth were not significant (P > 0.05). Absolute catches of spat on the two types of collectors are not strictly comparable since collector types differed between sites, however spat of all species examined set- tled in large numbers on both types of collectors. Mann- Whitney U tests also failed to show significant differences in the size of spat collected on either type of collector. Spat of Pinctada fucata and Pinna bicolor were not counted during the first two collection periods, but in the final period, counts in excess of 100 individuals of each species per collector were recorded (Table 1). While the shell heights of most P. fucata were less than 15 mm (Fig. 4a), a few specimens had grown to more than 30 mm in the two month period. By comparison the modal size class of P. bicolor was 10—14 mm, although some had reached shell heights in excess of 60 mm during the same time (Fig. 4b). DISCUSSION The insignificant settlement of Amusium balloti spat on collectors during the known settlement period for the species (Dredge 1981, Williams and Dredge 1981) indi- cates that the collectors used were unsuitable for that partic- ular species or that there was limited spatfall in the areas chosen. The latter seems unlikely since commercial vessels working in these areas during 1989 have reported signifi- cant A. balloti catches. Recent work by Rose et al. (1988) has shown that A. balloti has a relatively short byssal stage and attachment of spat by byssal threads was never perma- nent. This suggests that the species may not be suitable for collection on artificial collectors. On the other hand, both TABLE 2. Mean shell height of three species of pectinid spat from collectors set at different depths in Hervey Bay during 16 September-19 November 1987. Mean Spat Shell Height (mm) Decatopecten Distance From Mimachlamys decatopecten Sea Bottom (m) M. leopardus gloriosa strangei 18 7.8 8.2 6.5 15 8.0 8.1 6.5 12 8.9 8.7 7.8 9 8.7 Ie) 6.3 6 8.7 7.0 Tal 3 8.7 7.0 6.6 Statistical significance NS NS NS 230 (A) 40 > oO z 5 @ 20 Ww c uw & 0 SO i ON St Ost: Ui rere ON CN ieee CD) orf) Th Doo 4 (oy) fo) %) fo) Sam ON 1CNiic) SIZE CLASS (mm) (B) % FREQUENCY 40-44 45-49 50-54 55-59 60-64 65-69 70-74 oa ! wo - vw vt = N ew | ' (oy {py Xe) - N 25=2/9 30-34 35-39 SIZE CLASS (mm) Figure 4. Shell Height frequency of (A) Pinctada fucata and (B) Pinna bicolor collected on artificial collectors in Hervey Bay during 19 No- vember 1987-18 January, 1988 types of collectors caught large numbers of potentially commercially important species such as Pinctada fucata and M. leopardus with some spat of the former growing to more than 30 mm in two months. The larger average size of Mimachlamys gloriosa and Decatopecten decatopecten strangei collected during early summer suggests that settlement was greatest during the early part of that collection period or that warmer sea tem- peratures at that time resulted in faster growth. This, and the relatively large numbers of smaller sized spat found in collectors during spring suggests that maximum settlement was occurring around October—November. This conclusion is consistent with the settlement of Pecten alba in southern Australia, which has a single annual settlement during Oc- tober—December (Sause et al. 1987). However, the fact that small (S—10 mm shell height) M. gloriosa, D.d. strangei and M. leopardus were taken throughout the study indicates that spawning and settlement may be prolonged in these species. Sause et al. (1987) found that Pecten alba spat which SUMPTON ET AL. had settled on collectors in southern Australia during a one month period, could not readily be distinguished from the spat of other bivalves. The faster growth rate of the tropical and subtropical species collected during the present study enabled their identification on collectors which had only been immersed for 30 days (first collection period). The poor spat settlement during this trial may reflect the shorter time available for spat settlement in comparison to the 64 and 60 days for Trials 2 and 3 respectively. Alternatively, spawning and settlement may have been limited during the early part of the first collecting period. The greater settlement of spat on collectors located midway in the water column has been noted in other studies (Sause et al. 1987, Brand et al. 1980). Brand et al. (1980) believed that either algal fouling or the effects of wave ac- tion and turbulence were responsible for the poor settle- ment of Chlamys opercularis and Pecten maximus on col- lectors near the water surface. During the present study algal fouling was also a problem on some collectors al- though this was not only restricted to collectors close to the water surface. The decrease in numbers of spat near the sea bed may have been due to the effects of silting, since Naidu and Scaplen (1976) have suggested a detrimental effect of silting on the respiratory apparatus of Placopecten magel- lanicus. Predation may also be involved since small fish, crabs and other predators were more commonly found in collectors close to the sea bottom. Many of the collectors deployed for 60 days were heavily fouled with algae, which presumably reduced their spat collecting efficiency. Any extension of the immersion period would increase fouling by algae, ascidians and other organisms and thus be counter productive. In conclusion this study has demonstrated that artificial spat collectors made of 5 mm mesh may be an appropriate method of obtaining relatively large numbers of tropical bi- valve spat for either an aquaculture industry or for stock enhancement programmes (providing settlement patterns are comparable in subsequent years). The spat collectors used in this study failed to attract or retain meaningful numbers of A. balloti spat and it appears that finer mesh bags or a completely different spat collecting device will be needed to collect spat of this species. ACKNOWLEDGMENTS We gratefully acknowledge the assistance of P. Smith, C. Lupton and G. Lowe in fabricating, deploying and re- trieving the collectors. Thanks also to G. Davidson and E. Donovan who typed the manuscript. Dr T. Hailstone and K. Lamprel assisted with bivalve identification. REFERENCES CITED Brand, A. R., J. D. Paul & J. N. Hoogesteger (1980) Spat settlement of the scallop Chlamys opercularis (L.) and Pecten maximus (L.) on arti- ficial collectors. J. Mar. Biol. Assoc. UK 60:379-390. Dix, G. D. & M. J. Sjardin (1975) Larvae of the commercial scallop, Pecten meridionalis from Tasmania, Australia. Aust. J. Mar. Freshw. Res. 26:109—112. Dredge, M. C. L. (1981) Reproductive biology of the saucer scallop Amusium japonicum balloti (Bernardi) in central Australian waters. Aust. J. Mar. Freshw. Res. 32:775—787. Dredge, M. C. L. (1988) Recruitment overfishing in a tropical scallop fishery? J. Shellfish Res. 7:233—239. Heald, C. I. & N. Caputi. (1981) Some aspects of the growth, recruitment SETTLEMENT OF BIVALVE SPAT ON COLLECTORS 231 and reproduction in the southern saucer scallop Amusium balloti (Ber- nardi 1861) in Shark Bay, Western Australia. Fish. Res. Bull. West. Aust. 25:1—33. Iredale, T. (1939) Mollusca, Part I. Great Barrier Reef Expedition 1928-1929 Scientific Reports V.5 No 6. Ito, S. and A. Byakuno (1989) The history of scallop culture in Japan. In Proceedings of the Australasian Scallop Workshop. (eds) M. Dredge, W. Zacharin and L. Joll., Tasmanian Government Printer, Hobart. Joll, L. (1989) History, biology and management of Western Australian Stocks of the saucer scallop Amusium balloti. In Proceedings of the Australasian Scallop Workshop. (eds) M. Dredge, W. Zacharin and L. Joll., Tasmanian Government Printer, Hobart. Naidu, K. S. (1970) Reproduction and breeding cycle of the giant scallop, Placopecten magellanicus (Gmelin) in Port au Port Bay, Newfound- land. Can. J. Zool. 48:1003—1012. Naidu, K. S. & A. Scoplen (1976) Settlement and survival of the giant scallop, Placopecten magellanicus larvae on enclosed polyethylene film collectors. F.A.O. Technical Conference on Aquaculture, Kyoto, Japan, 1976. FIR: AQ/Conf/76/E.7, 5 pp. Paul, J. D. (1981) Natural settlement and early growth of spat of the queen scallop Chlamys opercularis (L.) with reference to the forma- tion of the first growth ring. J. Moll. Stud. 47:53—-58. Sause, B. L., D. Gwyther & D. Burgess (1987) Larval settlement, juve- nile growth and the potential use of spatfall indices to predict recruit- ment of the scallop Pecten alba Tate in Port Phillip Bay, Victoria, Australia. Fish. Res. 6:81—92. Williams, M. J. & M. C. L. Dredge (1981) Growth of the saucer scallop Amusium japonicum balloti, Bernardi in central Queensland waters. Aust. J. Mar. Freshwater Res. 32:657—666. Woodburn, L. (1989) Genetic variation in southern Australian Pecten. In Proceedings of the Australasian Scallop Workshop. (eds) M. Dredge, W. Zacharin and L. Joll., Tasmanian Government Printer, Hobart. Journal of Shellfish Research, Vol. 9, No. 1, 233-237, 1990. A REVIEW AND EVALUATION OF BIVALVE CONDITION INDEX METHODOLOGIES WITH A SUGGESTED STANDARD METHOD* MICHAEL P. CROSBY! AND LAURENCE D. GALE? "Belle W. Baruch Institute for Coastal Research and Marine Biology Baruch Marine Field Laboratory University of South Carolina P.O. Box 1630 Georgetown, SC 29442 College of Charleston Grice Marine Laboratory 205 Fort Johnson Charleston, SC 29412 ABSTRACT Problems in comparing bivalve condition indices between various studies exist primarily due to the lack of an accepted standard condition index formula. A review of twentieth century literature yields at least six different condition index formulae currently in use. We have statistically compared the three primary formulae from which all others are derived. We conclude that the following gravimetric formula has less measuring errors, lower coefficient of variation, is the easiest and fastest to use, and is most meaningful as an index of current bivalve nutritive status and recent stress: Condition Index = [dry soft tissue wt (g) x 1000]/internal shell cavity capacity (g). We recommend that it be accepted as the future standard method for determining bivalve condition index. We also suggest that further investigations into the seasonal and age-dependant relationships between the tissue:shell body component and gravimetric condition indices are needed and may provide valuable new insights into bivalve energetics. KEY WORDS: INTRODUCTION Condition indices are generally regarded as useful mea- surements of the nutritive status of bivalves. These indices may be used to follow seasonal changes in gross nutrient reserves or indicate differences in the commercial quality (meat yield) of bivalve populations. Condition index may also be employed as an assay for monitoring various pol- lutants and disease. In the early 1900's, a qualitative index of bivalve condition referred to as the degree of ‘‘fat- tening’’ was employed (i.e., Moore, 1908). Savage (1925) attributed the ‘‘fattening’’ of an oyster to the accumulation of glycogen reserves. Galtsoff (1964) credits the first quan- tification of condition index to Grave (1912). This index was based on the percentage of the internal shell volume occupied by the oyster soft body tissue. Walne (1970), however, attributes the first quantitative index to Milroy (1909) who used an index based on wet soft body tissue. A. E. Hopkins 1s given credit for the first definable quanti- tative condition index equation, as described by Higgins (1938) and given as follows: dry soft tissue wt (g) x 100 condition index (CI) = - > internal shell cavity vol (ml) (1) *Belle W. Baruch Institute Contribution #794. Submitted to the Journal of Shellfish Research, 12/89; Revised version, 2/90 ‘Author to whom all correspondence should be sent bivalve, body component index, clam, condition index, oyster, mussel, nutritive status, stress The internal shell cavity volume in equation | was the dif- ference between the volume of water displaced by a whole live oyster (or other bivalve) less the volume of water dis- placed by the shell alone. The above equation has served as the blueprint for a multitude of modern condition index for- mulae. Quayle (1950) suggested using Archimedes’ Prin- ciple to determine internal shell cavity volume via gravi- metric techniques. This method calculates internal shell cavity volume (SC) as in equation 2. SC = [emersed whole live wt (g) — immersed whole live wt (g)] — [emersed shell wt (g) — immersed shell wt (g)] (2) Baird (1958) concluded that, from his statistical evalua- tions, no differences were to be found between using dry versus wet weight of the soft body tissues in equation 1. Thus began the plethora of different variations of equation 1 in use today, although the use of the wet tissue weight in condition index formulae seems to be of little use to most investigators (i.e., Millar, 1961). Westley (1959) used a combination of equations 1 and 2 with pooled samples of Crassostrea gigas and Ostrea lurida to obtain seasonal changes in condition indices for these two species. Haven (1960) also pooled samples of Crassostrea virginica to ascertian seasonal and spatial variability in condition in- dices, but used equation | with its original volumetric methods for determining shell cavity volume. Walne (1970) followed Haven’s technique of pooling samples for 233 234 dry weight and volumetric assessment of internal shell cavity volume but increased the condition index by an order of magnitude by altering the formula as follows: _ dry soft tissue wt (g) x 1000 3 internal shell cavity vol (ml) 2 Walne and Mann (1975) further modified this formula such that dry soft tissue weight was a function of dry shell weight as in equation 4. _ dry soft tissue wt (g) x 1000 Cl dry shell wt (g) (4) Lawrence and Scott (1982) followed several years later with yet another revision of the condition index formula as presented in equation 5. ee dry soft tissue wt (g) x 100 5 ~~ internal shell cavity capacity (g) ©) The shell cavity capacity of a bivalve is determined by sub- tracting dry shell weight (g), in air, from the total whole live weight (g), in air, of a cleaned animal. By 1982, then, there were no less than five different condition index for- mulae using unstandardized methods and reporting index ranges that varied by an order of magnitude. The commu- nity of bivalve scientists apparently could reach no con- sensus of which formula to use. Soniat and Ray (1985) pooled samples of C. virginica and used equation | with it’s volumetric displacement methodology. Newell (1985) and Emmett et al. (1987) also employed equation 1, but with individual C. virginica and Mytilus edulis, respec- tively. Hawkins et al. (1987) calculated Mya arenaria con- dition index with a modification of equation 5 as given below: dry soft tissue wt (g) x 1000 = 6 internal shell cavity capacity (g) o Although Hawkins et al. (1987) derive internal shell cavity capacity via the gravimetric techniques of Lawrence and Scott (1982), the former authors were apparently unaware of the latter’s publication and credit the origin of this gravi- metric technique to an unpublished manuscript by Drinnan and Henderson (1959). Littlewood and Gordon (1988) cal- culated condition indices for Crassostrea rhizophorae with equation 4. The lack of uniformity for condition index for- mulae should now be readily apparent. This lack of unifor- mity is illustrated in the same recent volume of the Journal of Shellfish Research where Barber et al. (1988) calculated condition index for C. virginica with Walne’s (1970) for- mula (equation 3) while Abbe and Sanders (1988) utilized Lawrence and Scott’s (1982) formula (equation 5) for the same species. In and of themselves, none of these formulae are “‘right’’ nor ‘‘wrong’’. However, this situation has precipi- tated a high degree of difficulty for inter- and intra- CROSBY AND GALE specie(s) comparison of condition indices between the mul- titude of published and future studies. It would seem that some form of a standardized condition index formula for future studies is warranted. The purpose of this manuscript is to: 1) present a case for the standardized condition index formula for bivalves, 2) evaluate the three primary methods for calculation of bivalve condition indices and 3) suggest a standard formula for future use. MATERIALS AND METHODS Samples of the oyster, Crassostrea virginica were col- lected from three different sites in South Carolina, USA between May and June, 1989. The first two sites were Clambank and Bread and Butter creeks in the North Inlet estuary/salt marsh. Oysters were collected from the inter- tidal zone at nine stations in Clambank creek and two sta- tions in Bread and Butter creek. The third collection site was a biculture shrimp pond at the Waddell Mariculture Center. These oysters were originally set on PVC pipes in the Charleston estuary, then transported to the ponds for growout. One hundred eighty-three oysters were collected from the Clambank creek, 48 from the Bread and Butter and 120 from the Waddell Mariculture Center sites for a total of 351 oysters. The oysters were cleaned of all epi- bionts, and whole live animal volume (ml) and weight (mg), shucked dry shell volume (ml) and weight (mg), and dry (80°C, 48 h) soft tissue weight (mg) determined for each individual oyster. Condition indices based on shell cavity volume (eq-3), shell cavity gravimetric capacity (eq-6) and shell weight (eq-4) were calculated for each oyster using the previously described methods for deter- mining volumes and weights. The condition indices calculated using all three equa- tions were first compared using the Kruskal-Wallis one- way analysis of variance by ranks (Sokal and Rohlf, 1981). The Mann-Whitney U test (Sokal and Rohlf, 1981) was used to compare the volumetric condition index (CI-vol) versus both the gravimetric condition index (CI-grav) and shell condition index (ClI-shell), as well as the Cl-grav versus Cl-shell. The Kendall rank correlation analysis (Sokal and Rohlf, 1981) was also used to determine whether a significant relationship existed between any of the three condition indices. Level of significance was p S 0.05 for all statistical analyses. RESULTS The graphical presentations of the three different condi- tion indices are given in Fig. 1 a, b, and c. Measuring errors are always a potential problem when calculating condition indices. In reviewing the plot of raw data for the Cl-vol (fig. la) it is apparent that the two negative values must be due to measuring errors and should be considered outliers. There also seem to be a number of unusually high Cl-vol values. The great amount of scatter of the Cl-vol data, however, makes it rather difficult to determine which STANDARD CONDITION INDEX FORMULA FOR BIVALVES Condition Index [dry tissue wt (g) X 1000/cavity vol (mi)]} Individual Observation Condition Index [dry tissue wt (g) X 1000/cavity capacity (g)} 2 ope. 3 [} i oO Ll = Lh re: ee rice ote Piet atari 7 h, SATPLU Sas at CAAT aR nT sk LTO ag i en Mee TLL a i Individual Observations Condition Index [dry tissue wt (g) X 1000/dry shell wt (g)] Individual Observations Figure 1. a) Scattergram graphical illustration of 351 individual oyster condition indices derived from volumetric methods (see text for details). Observations range from —210.5 to 1200. b) Scattergram graphical illustration of 351 individual oyster condition indices de- rived from gravimetric methods (see text for details). Observations range from 7.1 to 1910. c) Scattergram graphical illustration of 351 individual oyster condition indices derived from shell weight methods (see text for details). Observations range from 2.1 to 59.6. values are outliers. By contrast, the scattergram of raw CI- grav values (fig. 1b) exhibits no negative values and a very tight grouping of the data points at values of <200 with only three very obvious outliers. The Cl-shell values (fig. 1c) seem to be grouped fairly well at values of <60. It was, therefore, concluded that any condition index values <0 or >200 were outliers and not considered in any further anal- i) Ww n yses. This range restriction resulted in a loss of ~11% of the original Cl-vol observations, <1% of the Cl-grav ob- servations and none of the Cl-shell observations. Means, standard errors (SE), coefficients of variation (CV) and ranges of values for original and range modified condition indices are given in table 1. The Kruskal-Wallis one-way analyses of variance by ranks indicated a significant difference (H corrected for ties = 662.22, p < 0.005) between the mean ranks of the three types of condition indices. The Mann-Whitney U test indi- cated no significant difference between Cl-vol and Cl-grav. There were, however, significant differences between CI- vol and Cl-shell (Z corrected for ties = —21.781, p < 0.001), as well as between Cl-grav and Cl-shell (Z cor- rected for ties = —22.531, p < 0.001). This is to be ex- pected since Cl-shell is based on dry shell weight while ClI-vol and Cl-grav are based on the volume and capacity of the shell cavity. The Kendall rank correlation analyses in- dicated significant positive correlations for Cl-shell with Cl-vol (Z corrected for ties = 6.797, p < 0.001), Cl-shell with Cl-grav (Z corrected for ties = 8.178, p < 0.001) and Cl-vol with Cl-grav (Z corrected for ties = 9.236, p < 0.001). DISCUSSION Although investigators may be quite meticulous in the way in which they make condition index measurements measuring errors are possible. Minute errors in measuring and recording volume displacement, for example, may be amplified into large errors in Cl-vol when one considers that an accuracy of +1 ml translates into an error of +1 g dry tissue weight. The Kendall rank correlation analyses demonstrated a significant positive relationship between the three different condition index techniques, with the strongest correlation between CI-vol and Cl-grav. If all three methods of mea- suring condition index yield the same results, comparison between studies using different methods would be valid. TABLE 1. Means, standard errors (SE), coefficients of variation (CV), minimum value (MIN), maximum value (MAX) and number of osbervations (N) are given for original and range restricted values of volumetric condition indices (CI-vol), gravimetric condition indices (Ci-grav) and shell condition indices (CI-shell). Range Restricted Original Observations Observations CI-vol ClI-gray ClI-shell ClI-vol ClI-gray Cl-shell Mean 116.27 87.98 22.03 82.32 79.42 22.03 SE 7232 5.87 0.40 1.89 1.28 0.004 CV 117.92 124.99 33.90 40.62 30.05 33.90 MIN — 201.50 7.10 2.10 8.90 7.10 2.10 MAX 1200.00 1910.00 59.60 200.00 189.90 59.60 N 351 351 351 313 348 351 236 The results of the Kruskal-Wallis analysis, however, dem- onstrate that this is not the case. Analyses via the Mann- Whitney U test indicate that results from studies utilizing the Cl-vol and Cl-grav techniques are comparable. How- ever, it would not be valid to compare results from studies utilizing the Cl-shell technique with studies using the Cl- vol or Cl-grav techniques. The Cl-vol and Cl-grav tech- niques yield indices which assess the proportion of avail- able internal cavity capacity utilized by a bivalve’s soft body tissues. The Cl-shell technique is not a measure of how much available space is utilized and does not account for possible variations of internal cavity capacity due to overall shape and shell thickness variability in bivalves (Mann, 1978). It is, instead, a body component index (sim- ilar to that described by Bayne et al., 1985) which com- pares the proportions that soft body tissue and shell weight compose of the total dry bivalve weight. The value of a tissue:shell body component index is that it is an “‘abso- lute’’ index (as opposed to a “‘relative’’ index such as CI- vol and Cl-grav) comparing metabolism directed towards calcification processes and metabolism focused towards so- matic and gametic processes of glycogen storage, protein synthesis and vitellogenesis. Cl-shell is not, then, an index of nutritive status and should not be used as being indica- tive of recent catabolic or anabolic activity within a bi- valve. Borrero and Hilbish (1988) found that soft body tissue in the mussel Geukensia demissa was variable be- tween January and May, but consistently high between May and October. Shell growth increased steadily throughout the year to a maximum between June and July. The rate of shell growth decreased after spawning in Au- gust, but tissue continued at a high rate. Cl-shell may, then, be quite misleading if employed as an index of recent CROSBY AND GALE stress in bivalves. Future investigations of the seasonal and age-dependant relationships between Cl-grav and tissue:shell body component indices may yield valuable new insights into bivalve energetics. It is apparent that either ClI-vol or Cl-grav should be used if one wishes to ascertain nutritive status of bivalves or determine whether the animals are under stressful condi- tions. Although the Mann-Whitney U test deomonstrated no significant difference between Cl-vol and Cl-grav, sev- eral rather important differences do exist between these two techniques. Removal of outliers from each data set resulted in a loss of ~11% of the Cl-vol observations, but <1% of the Cl-grav observations. The mean Cl-vol and Cl-grav were not significantly different. However, the Cl-vol coef- ficient of variation was ~25% greater than for Cl-grav. Due to the lower rate of measuring error (as seen in <1% outliers), less variability surrounding the mean (as seen in a lower coefficient of variation), the ease of technique and savings in time, it is recommended that all future studies measuring condition index in bivalves use the ClI-grav methods in combination with equation 6. The use of such a uniform method for calculating bivalve condition would allow valid and meaningful comparisons to be made be- tween results of these studies. ACKNOWLEDGMENTS The authors would like to thank Angela Smith and Paul Kenny for their assistance in collecting and measuring some of the samples. Ms. Smith was supported by a summer intern grant from the E. W. Moore Foundation. One of the authors (MPC) would also like to thank both the Belle W. Baruch Foundation and the Belle W. Baruch In- stitute for their financial support of this study. REFERENCES CITED Abbe, G. R. & J. G. Sanders. 1988. Rapid decline in oyster condition in the Patuxent River, Maryland. J. Shell. Res. 7:57—60. Baird, R. H. 1958. Measurement of condition in mussels and oysters. J. Cons., Cons. Int. Explor. Mer. 23:249—257. Barber, B. J., S. E. Ford & H. H. Haskin. 1988. Effects of the parasite MSX (Haplosporidium nelsoni) and oyster (Crassostrea virginica) en- ergy metabolism. I. Condition index and relative fecundity. J. Shell. Res. 7:25-32. 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. Stebbing & J. Widdows. 1985. The effects of stress and pollution on marine animals. Praeger Sci., N.Y. 384 pp. Borrero, F. J. & T. J. Hilbish. 1988. Temporal variation in shell and soft tissue growth of the mussel Geukensia demissa. Mar. Ecol. Prog. Ser. 42:9-15. Drinnan, R. E. & E. B. Henderson. 1959. An index of condition in oysters. Its determination and significance. Fish. Res. Board of Can., Biol. Sub-sta., Ellerslie, P.E.I. 11 pp. Emmett, B., K. Thompson & J. D. Popham. 1987. The reproductive and energy storage cycles of two populations of Mytilus edulis (Linne) from British Columbia. J. Shell. Res. 6:29—36. Galtsoff, P. S. 1964. The American Oyster Crassostrea virginica Gmelin. Fish. Bull. Fish Wild. Ser., 64. Grave, C. 1912. A manual of oyster culture in Maryland. 4th Rep. Shell Fish Comm. MD. 279-348. Haven, D. 1960. Seasonal cycle of condition index of oysters in the York and Rappahannock Rivers. Proc. Nat. Shell. Ass. 51:42—66. Hawkins, C. M., T. W. Rowell & P. Woo. 1987. The importance of cleansing in the calculation of condition index in the soft-shell clam, Mya arenaria (L.). J. Shell. Res. 6:29—36. Higgins, E. 1938. Progress in biological inquiries, 1937. Bull. U.S. Bur. Fish., Admin. Rep. No. 30. 1-70. Lawrence, D. R. & G. I. Scott. 1982. The determination and use of con- dition index of oysters. Estuaries 5:23—27. Littlewood, D. T. & C. M. Gordan. 1988. Sex ratio, condition and gly- cogen content of raft cultivated mangrove oysters Crassostrea rhizo- phorae. J. Shell. Res. 7:395—400. Mann, R. 1978. A comparison of morphometric, biochemical, and physi- ological indexes of condition in marine bivalve molluscs. In: Thorp, J. H. & J. W. Gibbons (eds.) Energy and environmental stress in aquatic systems. Tech. Info. Center, USDOE. p. 484—497. Millar, R. H. 1961. Scottish Oyster Investigations 1946-1958. Mar. Res. (3). Milroy, J. A. 1909. Seasonal variations in the quantity of glycogen present in samples of oysters. Scient. Invest., Lond., Ser. 2. 17(6). Moore, H. F. 1908. Volumetric studies of the food and feeding of oysters. Bull. U.S. Bur. Fish. 28:1297-1308. STANDARD CONDITION INDEX FORMULA FOR BIVALVES 237 Newell, R. I. E. 1985. Physiological effects of the MSX parasite Haplo- sporidium nelsoni (Haskin, Stauber and Mackin) in oysters. J. Shell. Res. 5:91—96. Quayle D. B. 1950. The seasonal growth of the Pacific oysters (Ostrea gigas) in Ladysmith Harbour. British Columbia Dept. Fish. Rep. 1950. 85-90. Savage, R. E. 1925. The food of the oyster. Min. Agric. Fish., Fish. Invest., Ser. 2. 7(1). Sokal, R. R. & F. J. Rohlf. 1981. Biometry (2nd ed.). W. H. Freeman and Comp., San Francisco. 859 pp. Soniat, T. M. & S. M. Ray. 1985. Relationships between possible avail- able food and the composition, condition and reproductive state of oysters from Glaveston Bay, Texas. Cont. Mar. Sci. 28:109-121. Walne, P. R. 1970. The seasonal variation of meat and glycogen content of seven populations of oysters Ostrea edulis L. and a review of the literature. Min. Agric. Fish. Food. Fish. Invest., Ser. 2. 26(3). Walne, P. R. & R. Mann. 1975. Growth and biochemical composition in Ostrea edulis and Crassostrea gigas. In: Barnes, H. (ed.), Proc. 9th Europ. mar. biol. Symp. Aberdeen Univ. Press. pp. 587-607. Westley, R. E. 1959. Selection and evaluation of a method for quantita- tive measurement of oyster condition. Proc. Nat. Shell. Ass. 50:145— 150. Journal of Shellfish Research, Vol. 9, No. 1, 239-255, 1990. COMMERCIAL SHELLFISH FINISHING WITHIN AN INLAND, CLOSED SYSTEM! LAWRENCE P. RAYMOND International Bio-Resources, Inc. 1667 Cole Boulevard, Suite 400 Golden, Colorado 80401 ABSTRACT This paper reports the design and operation of an inland, closed-system, oyster finishing process conducted for com- mercial purposes in Rifle, Colorado. Findings regarding coliform and vibrio removal are presented, illustrating some concerns with existing depuration practices. An operating case history is presented, identifying and discussing cost sensitivities associated with an integrated holding, flavoring, and depurating facility. Conclusions are drawn from this experience regarding the commercial viability of shellfish finishing. KEY WORDS: INTRODUCTION Six cases of toxigenic cholera Ol were reported within the United States during the summer of 1988, linked di- rectly to oysters originating in the Gulf of Mexico (Doran et al. 1989). Three cases occurred within inland markets; one was derived from repackaged oyster, and two were re- corded within Gulf states (Pavis et al. 1987). The first case resulted from certified oysters transported to the Inland Mariculture facility in Rifle, Colorado, designed to hold and improve the quality of shellfish destined for inland markets. Inland mariculture was fully licensed and per- mitted for wet-holding, and was operated to meet and ex- ceed regulatory requirements for shellfish depuration. These criteria included strict control of total and fecal coli- forms within holding waters as well as within shellfish tissue samples, and operations never waivered from this compliance. One principal objective was to provide the best possible protection against the increasing hazards as- sociated with eating raw bivalves. The Inland Mariculture Corporation (IMC) was estab- lished to provide fresh, full flavored, clean molluscan shellfish to inland markets. As a private corporation, IMC designed and operated an integrated, closed-system process for shellfish finishing, conducted for commercial purposes. Its products were designated for the high-end of the market, meeting an unsatisfied demand for high quality, well fla- vored, firmly textured, and sanitary shellfish that would satisfy the most particular customers. The intent was to provide a structured marketing organi- zation that would make available year-round supplies of pure and optimally flavored shellfish to quality restaurants and grocery stores in Colorado, Utah, New Mexico, Ari- zona, and Nevada. This was seen as an option for the shell- fish industry, providing price stabilization, quality stan- dards, and reliability for a variety of shellfish now grown and harvested along the American Coastline. The ocean is the pasture for product grow-out; the facility was the ‘Presented at the First International Conference on Molluscan Shellfish Depuration, November 5—8, 1989, Orlando, Florida, U.S.A. shellfish, Crassostrea virginica, depuration, Vibrio, environmental control feedlot—converting animals into uniformly high quality products. This paper documents the operations, engineering, and control of this facility, presents the measures developed to control Vibrio accumulations, and draws from this com- mercial experience to develop recommendations for techni- cally valid and cost-effective systems. Its intent is to smooth the way for other ventures endeavoring to help pre- serve traditional molluscan markets. THE FEEDLOT PROCESS The IMC process consists of three major components: 1) receiving, 2) preparation & holding, and 3) shipping. The preparation & holding operations prepare sanitary, FDA- approved bivalves within low temperature saline water (im- proves product texture) while being fed high quality mi- croalgae (restores shelf life and enhances flavor), enabling distributors to supply fresh product to their clients ac- cording to customer needs. Extended preparation and in- ventory management operations provide further improved flavor and texture, yielding shellfish of very high quality. Fig. 1 provides a schematic representation of the pro- cess, indicating timing, capacity, and flow paths in the flow of material inventories. Incoming FDA Certified shellfish are placed in temperature controlled recirculating saltwater tanks sized for holding approximately 70 bushels of oyster stock, turning over every four days for a weekly capacity of approximately 1000 bushels. Operations within these tanks are staged and adjusted to provide the appro- priate conditions, feeding, and water management neces- sary to accomplish each process step, performed in such a way as to physically handle the shellfish inventory only during receiving and shipping operations. The rest of the operations are automated, controlled in accordance with the rates at which the facility is able to provide its treatment and enrichment operations; output rates are therefore con- trolled by product quality. Treatment progress is monitored continuously via selected sensors (pressure, temperature, pH, redox potential, conductivity, oxygen, and flow rate). Adjustments are made entirely by valving and mixing of process flow streams, making this a fully hydraulic, low- skill labor operation. 239 240 COUNT AND | ___ ARRIVAL RECORD ' o 500 bu/3.5 days COLD STORAGE o 750 bu capacity RAYMOND Legend Process Flow MRP Inventory Control Quality Control o Must Meet Sanitation Standards pe | HARVEST 96 hours Required PLACEIN PURIFICATION aS AND FEEDING o 1000 bu capacity COUNT AND RECORD 250 bu/day 4 days/week DISTRIBUTE © 1000 bu/week 0 250 bu/trip i} i i} 1 if 1 ! ! i} i | i i 1 i i i i 1 i i i i} i i 1 i i i i | Y / Y Samples for Microblological Testing TIME (DAYS) 1 2 3 4 5 6 U Figure 1. Schematic of the IMC process, showing the process flow, points of inventory control, and timing of samples for quality control purposes. Fig. 2 is the flow diagram of the shellfish holding system process, designed in compliance with the funda- mental guidelines of Furfari (1966, 1976, 1987) and meeting the operating and maintenance standards of the ISSC, Parts I and II, 1988. The holding process had five operational modes: 1) sterilization and recirculating water treatment, consisting of filtration through a rapid sand filter, sterilization via UV light (120,000 microwatts-sec/ cm?), and protein/organics removal via foam fractionation (microspargers with air); 2) temperature control, accom- plished as a 30-50% shunt off the main flow, passing seri- ally through heat exchangers for heating or chilling; 3) feeding, accomplished as a bypass of the filtration loop via the injection of harvested microalgae into the recirculation water stream; 4) seawater recharging and exchange, pro- vided as a shunt off the recirculation line or as an addition to the return flows, and 5) tank cleaning and maintenance, provided through independent drains in each tank section or via a vacuum system, both connected to a sludge/waste re- moval loop. Fig. 3 provides a general schematic of the IMC facility. The shellfish holding tank shown in Fig. 2 represents one module out of eight in the total system; each module was divided into three equivolume sections. Water turnover rates were 32 minutes minimum and 60 minutes maximum, depending on the treatment cycle. At the minimum, this rate equated to one gallon per minute per bushel capacity. The saltwater is synthetic, made by adding Marine Envi- ronment Salts mixture to municipally treated, dechlorinated freshwater, held in the seawater storage tanks. Included in the facility were areas for feed production, shellfish holding, cold storage, washing & packing, a microbiolog- ical laboratory, dry storage, and administrative offices. All components of the system were organized to mini- mize system downtime and maximize utilization of design capacity. Computerization was required to accomplish this. FEED Traditionally, the cost of feed is among the largest ex- pense items associated with aquaculture operations (Ray- mond et al. 1974; Raymond 1982; Urban and Langdon, 1984). Microalgae, at best, convert through photosynthesis less than 8% of the total solar isolation into the stored chemical energy of feeds. Since microalgal productivity is an area function, the greater the amount of light, the less the area required to produce the feed required. Both facility capitalization and operating efficiencies are determined in INLAND COMMERCIAL SHELLFISH FINISHING 241 Feed Production at T — Unit Shellfish Heat Exchangers Holding Tank \ b= ——— agg —— — = S95 Temperature ¢ | Seawater | Control Loop Hot Cold | Storage L WS Et A ae ee | Teal A Foam | Fractlonator al aaa al I | | | i Fitration voop gl | UY Unit ! — -— Sludge/Waste Removal Unit Legend Outflow Inflow =) = RS Filters | Micro Bubble -i|/----- Aeration Figure 2. Process flow diagram for the IMC seafood finishing process, illustrating one of the ten shellfish holding modules and one of the four feed production units. large part by the area occupied; capital costs go up and efficiencies go down as photo-productivity declines. A second major cost factor in aquaculture is contaminant control, particularly when the quality and consistency of the feed is an important product issue (Raymond et al. 1974). Molluscan shellfish take on the flavor of their envi- ronment, making the quality of their feed particularly crit- ical to their consistency and character as food products. Contaminants in the system reduce productivity, modify feed composition, affect flavor, and decrease product quality. In short, they spell disaster, and their control is of paramount importance. Microalgae are the mainstay of the molluscan shellfish diet, and no source has ever been found better (Raymond et al. 1974; Urban et al. 1983; Pruder and Bolton 1976). Re- cent work at the Solar Energy Research Institute and through the Aquatic Species Program of the U.S. Depart- ment of Energy initiated by Raymond has disclosed new opportunities for production cost savings, contaminant con- trol, and control over feed flavor and chemistry. Photosyn- thetic efficiencies now are consistently high, utilizing ap- proaches initiated by Kok (1953), Phillips and Myers (1954), Raymond (1979), Laws et al. (1983), and Terry and Hock (1985) utilizing shallow cultures, controlled peri- odicity of light, and a combination of species selection and cultivation techniques (see Terry and Raymond, 1985). The latter factor has become of critical importance in flavor control. The system utilized at IMC is illustrated schematically in Fig. 4. Its central unit is a shallow, recirculating raceway driven by an airlift pump. Mechanisms are in place for nu- trient, salinity, flow rate, and temperature control, as well as for control of patterns of water motion. Operated in a semi-continuous mode under computer control and under the lighting conditions of Rifle, Colorado (elevation 5,500 feet), productivities were consistently maintained at 45 grams, ash-free dry weight, per m* per day from June through September. DATA ACQUISITION AND CONTROL Initial computerization of IMC utilized conventional process control system logic and components provided by Keithly Instruments, coupled with sensors and signal con- ditioners distributed by Omega Engineering. Use of this in- strumentation clearly demonstrated its utilities and limita- tions. Advantages were apparent in terms of labor savings and the depth of understanding of process interactions made possible by the system. Disadvantages were in cost of programming, responding in time to unknowns, attempting 242 Greenhouse Algal Flumes Seawater Steel Warehouse RAYMOND Second Floor 8x21 Mech Room Reception 12x21 17.5x10 Lab/Comp 17.5x10 Conference 8x10 Sales 8x10 Books 90’ 17.5x12 Admin/Finance Storage Shipping Receiving Figure 3. General design layout for the IMC facility. to sort out complex algorithms, and updating the system as demands changed and knowledge improved. Additionally, signal to noise ratios associated with sensors and signal processors became so low as to make data difficult to inter- pret; this was found to result from a fundamental property in the system. Salt water flowing over a static surface at relatively high velocities (>0.3 ft/sec) produces a voltage potential similar to a sodium carbide battery. The algal flumes, for example, developed an 80 volt ground potential at a flow rate of 1 ft/sec. This noise level thoroughly confuses the millivolt outputs of standard sensors. Alterations in signal pro- cessing circuitry and switching from voltage to current sensing permitted circumventing this problem; however, this had to be developed, engineered, and manufactured before it could be put to use. Patterns in data relationships associated with various steps in the process began to emerge with time. It became apparent that the control system needed to interpret and re- spond to a complex series of algorithms being created by a multitude of biological interactions within the chemical and physical environment. This was an ideal application for ar- tificial intelligence-based languages, an area pursued inten- sively over the last year. Artificial intelligence operates on a concept of pattern development and recognition, creating and retrieving pat- terns specific to physiological condition. The patterns are centralized with other process data, providing comprehen- sive information on a real-time basis that can be compared against a developing intelligent and historical pattern refer- ence. It functions on a higher plane of analysis than con- ventional logic systems. INLAND COMMERCIAL SHELLFISH FINISHING 243 V1 co2 con roy C2 C Mixa )— 2 — V3 C MixB V4 Mix C poe Cmixp >} [Eos] Cmixe )—f=}—¥® V7 Salt H20 ss Bee] A 9 ie) Vtg ia 2 F Relations Ambient Out Flume PHS Efficiency Air Temp Overall Efficiency Humidity Total Yield Sunlight Substrate 1 Efficiency Wind Speed Flow Rate Wind Speed Substrate 2 Efficiency 1 Power Consumption |_| Pressure Total Power Conductivity | Total H20 Use Figure 4. Flow diagram of one feed production unit, showing quantified inputs and outputs, as well as the relations deemed critical to process performance. In the context of shellfish finishing operations, instru- and pattern coded to display rapidly any deviations in mentation has been configured for ease of use. Fig. 5 system operations. Each of the process operations is shows the screen constantly updating the operational status shown, with indicators identifying on/off status, direction of the entire system. This is a schematic screen that is color of flow, the active process, and its stage of treatment. Si- : 1 2 3 O RAYMOND SHELLFISH HOLDING SYSTEM: OVERVIEW STATUS 10 Pe) CS 6 7 STATUS ON Ready OFF fj Stage 4 rai stees Empty ({] Stage 1 Filling Draining Closed Maintenance In Progress ACTION Filter Recycle Feeding Harvesting Recharging Tank Oxygen UV Sterilizer Filter Unit Vacuum Unit Ammonia | _ Pressure |) Salinity Decant Sterilizer Figure 5. The first of three monitor screens developed to display the status of real time operations within the IMC process. This screen provides a comprehensive overview of operations and conditions for the full facility. multaneously, real-time data regarding the current status of critical parameters is shown, including oxygen, pH, ORP, temperature, ammonia concentrations, operating pressures, and salinity. Should a deviation occur from expected operating per- formance, its process location alarms, flashing red on the screen. A first level analytical screen, as shown in Fig. 6, was created to permit identification of the parameters in- volved in the process change, so that they can be isolated for more detailed study and full pattern definition. This screen identifies process relationals, current building con- ditions, detail on tank inputs and outputs, and their status. Patterns defining the event are studied on a strip chart recorder, built on the monitor screen, for analysis over time. An example screen is given in Fig. 7, showing a series of complex pattern changes. The process operator studies these changes, analyzes operational changes likely to resolve the problem, and effectuates those changes through control relays. The computer records the patterns and the responses, tracking the impact of operational changes on the process. It remembers specific patterns and successful responses for future use, whenever the pattern redevelops. When it does redevelop, the computer automat- ically imposes the operational adjustments that were suc- cessful in the past. If these changes are not successful, the computer alarms the operator, who analyzes the process, and the educational process is continued. MICROBIOLOGY IMC worked closely with the U.S.F.D.A., the Colorado State Department of Health, and the Centers for Disease Control, U.S. Department of Health and Human Services following the infection of a local customer with toxigenic Vibrio cholerae Ol in August, 1988 in order to determine the true source of the causative organism. This was a totally unexpected occurrence as this was only the third case to be documented from oysters in recorded history (Doran et al. 1989), and the first ever in Colorado (Pavis et al. 1987). INLAND COMMERCIAL SHELLFISH FINISHING 245 SHELLFISH HOLDING SYSTEM OPERATIONS: UNIT ONE Relations F.] PHS Efficiency Overall Efficiency Total Yield Substrate 1 Efficiency Substrate 2 Efficiency Power Consumption Total Power = Total H20 Use Flume Alr Temp Humidity Sunlight Wind Speed Flow Rate | Conductivity ui] Pressure || Wind Speed Ready Stage 4 Stage 3 Stage 2 Stage 1 | | Alr Temp || Humidity Ammonla Pressure Salinity Figure 6. The second screen of the IMC artificial intelligence data acquisition and control system. This screen provides greater detail of system operations, particularly those parameters critical to performance. U.S.F.D.A. regulations provide no recommendations for monitoring Vibrio concentrations in shellfish, nor do they provide or recommend any procedures for doing so. The causative organism was shown by VcA-3 phage typing to have been acquired in the Gulf of Mexico. Five other cases have been reported in other states since this oc- currence, with no additional cases resulting from oysters processed by IMC. These cases had some significant char- acteristics in common. All occurred during the period of August through early October when few shellfish sources can be relied upon to serve the market. This is also a time when shellfish quality is lowest— meats are small and soft, shellfish are spawing, and flavor is poor. All occurred from the consumption of raw oysters harvested from the Gulf during a period of exceptionally warm waters with low sa- linity (<10 ppt). During the summer of 1989, when the Gulf waters were not as warm, no cholera cases were re- ported. IMC initiated a strong, concerted effort under the direc- tion of microbiologist Carol Richards, having four objec- tives. First was to develop a routine procedure for selec- tively detecting Vibrio presence at the species level. Our interest was not necessarily distinguishing concretely be- tween all the species, strains, and types of vibrio; rather, it was important that their concentrations be detected and monitored in the course of treatment. Dressel (1988), for example, concisely showed the public health problems as- sociated with vibrioses in general, related to the consump- tion of raw shellfish, and it is precisely these problems that IMC was created to solve. Second was to apply this tech- nique to examining the extent of contamination within in- coming lots of transported oysters, as well as within all facets of the IMC process. Third was to explore several techniques for removing Vibrio contamination by modifica- tion of the IMC process. Last was to examine the impact of Vibrio removal on the overall performance of the process. Each of these was done, but not to conclusive levels due to a lack of both funds and time. Selective Detection The routine procedure developed is shown in Fig. 8, combining procedures of Collier, et al. (1950) using vi- 246 RAYMOND Screen Status AQUATRACKER Clock Status Press Space Bar to Continue 02-07-89 Start 17:36:52 Stop 17:38:52 Page #1 Cycle 5 Days DEMOTRACK M1 ET 08:23:11 M2 ET 20:58:21 FlumORP FlumCond Orp 254.6 Cond 40.74 Temp 22.5 FlumTemp Cumltco02 SunLite gCO2 96.0 uE 1098 Figure 7. The most critical monitor screen developed to track and record performance patterns within operational systems. This screen allowed parameter relationals to be developed and quantified, adding rapidly to an evolving knowledge of treatment and process interactions. briostatic agents, and Twedt (1984) entailing enrichment procedures followed by plating on TCBS. Several modifi- cations and substitutions in procedure were made to in- crease the effectiveness of this approach. One substitution was the use of Lauryl Tryptose Broth (LTB) in place of the more standard alkaline peptone broth; this avoids a duplication of effort in laboratories conducting routine coliform MPN tests. Growth of vibrios appeared unaffected in our studies by this approach. Vibrio colonies on TCBS plates can be distinguished by color and character; Vibrio cholerae (Ol and non-O1) are yellow, V. parahaemolyticus, vulnificus, and fluvialis are blue-to-bluegreen. With practice, V. parahaemolyticus can be distinguished as a darker, waxy appearing colony that is sticky, compared with either vulnificus or fluvialis. Black colonies may also appear with a strong H,S odor; these were found to be Pseudomonas putrefaciens. Confirmation of Vibrio is based on halotolerance and sensitivity to vibriostatic compounds. Therefore, positive colonies from TCBS plates were tested for growth on nu- trient agar with 1.5% NaCl, inhibiting the growth of non- halotolerant forms, leaving only firm candidate colonies for confirmation. This was done by replica plating, a technique that permits plates to be cloned using a sterile adsorptive surface—the area and shape of the agar plate surface— that is touched to the source plate, attaching bacteria at each location where colonies exist. The surface of this ad- sorptive plug is touched first with the clean surface of a water agar plate, and subsequently to the surface of the target agar, replicating the number and patterns of colonies on the source plate with high precision. Up to eight replicas can be prepared from a single adsorptive plug, saving enor- mous amounts of time in transferring colonies for confir- mation tests. Collier et al. (1950) demonstrated that pteridine com- pounds inhibited vibrio growth selectively; however, Merkel (1972) later showed that salt inhibited the vibrio- static activity of 2,4-diamino-6,7-Diisopropyl Pteridine used to confirm Vibrio identifications after positive colony colors on TCBS plates. This problem was circumvented through the use of agar overlays. BHI agar was found supe- rior to plate count agar for enumerating Vibrio in pure cul- ture and was used as the agar of choice for overlays of positive saline nutrient agar plates. Zones of inhibition were taken as confirmation of Vibrio presence. Table 1 shows the results of an experiment run to test the effectiveness of this method in distinguishing between coli- forms (E. coli) and various vibrio species. This demon- INLAND COMMERCIAL SHELLFISH FINISHING FLOW CHART FOR DETECTION OF VIBRIO "Sample (Oyster Tissue/Inlet H O,or shell broth, etc.) [ 4omIlTube - Alkaline Peptone (1) ——— Discard 3 ia = ~ growth incubate 24hrs y growth at35C aan 4 Streak | TCBS Agar Plate (2) | incubate 24hrs _—~others at 35C Se Yellow or Discard Y Bue Colonies | Replica Plate = a —— aa incubate 24hrs as. Nutrient Agar SESSIC | Nutrient Agar ] SS [withNacl | growth y Srowth 4 Prepare Overlay Plate with | Discard BHI, Place Disc Prepared with | | u} F-159 (3) and Check for Inhibition Zones (1) Alkaline Peptone 10g Peptone, 10g NaCl, 1L water, pH 8.5 (2) TCBS: Thiosulfate-citrate-bile-salt-sucrose agar (3) F- 159: 2,4 diamino-6,7-diisopropyl pteridine 1mg/ml dissolved in acetone. Pipette 0.2 ml portions onto 1/4 blank disc. Place disc on surface of media overlay after acetone is evaporated. Figure 8. Flow chart developed by Richards for detecting and distin- guishing Vibrio spp in shellfish tissues samples, tank waters, or shell broth. strated that the detection of Vibrio cholerae appears pos- sible by this method. This technique is relatively straight forward, but suffers from the time required to obtain positive results; 48 hours is required to obtain first cut verification of vibrio presence, and 96 hours is required for full confirmation. It was neces- sary to develop a method that could at least serve as a early warning indicator, that could be relied upon to protect the 247 consumer from contaminated shipments. Richards was on the path to validating such an approach when her efforts were cut short. However, what she had accomplished is reported here. Richards reasoned that TCBS could be used as a screening tool for vibrios through color development; she also believed that, if made into a broth, initial enrichment cultures might be eliminated, reducing first cut screening to a 24 hour test. This hypothesis was tested through the use of a TCBS broth. Tubes were then arranged as appropriate for a standard three tube MPN series, with appropriate di- lutions. These tubes were incubated along with coliform MPN tubes at 35 degrees for 24 hours. Development of a yellow coloration was taken as a positive reading. This procedure was not run long enough to draw any solid conclusions. However, it was utilized on two separate lots of incoming shellfish, on the waters in the shellfish tanks, and on the extratissue fluid within the shell, along with the plating technique and standard coliform water and tissue tests. The results appeared promising, paralleling plate results, but consistently yielding a higher estimate of vibrio populations. Technique Application Investigation of incoming lots of shellfish over a three week period (August 9 through September 6) consistently displayed high levels of coliforms and vibrio within tissue samples (these oysters were obtained originally from certi- fied sources). Total plate counts were in excess of 1.2 X 10°, MPN in excess of 1600 and vibrio total counts at 1000 per 100 grams of tissue. Over a period of 96 hours treat- ment within the IMC holding tanks at 21 degrees C, salin- ities of 14 ppt and consistent feeding (required for uniform coliform reduction), MPN was reduced by 2 orders of mag- nitude, and total plate counts dropped 3—4 orders of mag- nitude into legally allowable ranges, meeting all federally approved criteria for shellfish safety. However, as Fig. 9 shows, vibrio concentrations increased nearly 5 fold over TABLE 1. Test of the sensitivity of Richards’ method to detect and distinguish between coliforms and vibrios. Test Culture Escherichia Vibrio Vibrio Vibrio coli cholerae 01 cholerae non-01 from tissues Alkaline Peptone + 3h + +r Lauryl Tryptone Broth Growth sf + + ar Gas Production of: = — = TCBS Growth ~_ + + + Color N vic Y. ay Nutrient Agar # = = = Nutrient Agar + NaCl = + + + F-159 Inhibition N Y PY. aYZ 248 Concentration Change, Log10 Nn °o °o -5.00 tt) 24 96 120 144 Time, hours & Plate Count, Total Vibrio MPN, Coliform Figure 9. Graphic representation of the impact standard depuration procedures were found to have on total bacterial counts, coliforms, and vibrios. As other bacteria were purged from oyster tissue, vibrios were found to accumulate; modifications in holding procedures were required to reduce vibrio concentrations. the same period, and continued to increase over the entire 144 hour period of each experiment. Based on this infor- mation, and our experience with cholera, it may have been hazardous to public health should IMC have elected to dis- tribute this product, even though the product complied with all U.S.F.D.A. guidelines for assuring shellfish sanitation. By their standards, these oysters were ready for market. Techniques for In-tank Treatment of Shellfish-borne Vibrio Singleton et al. (1982a, b), Kelly (1982), Kelly and Stroh (1988), and others have demonstrated the sensitivity of the Vibrionaceae to temperature, salinity, and dissolved organics, showing that, in general, moderate salinities (7—25 ppt) and high temperatures (28—37 degrees C) fa- vored vibrio growth. Singleton, et al. (1982a) further showed that salinity had a marked adverse effect on V. cholerae growth when dissolved organic concentrations were less than | mg/l; at higher organic loadings, the sa- linity optima broadened. A subsequent study (Singleton et al. 1982b) showed V. cholerae were able to survive less than 4 days at a salinity of 25 ppt when temperature was comparatively low (10°C). These findings led Richards to hypothesize that low temperature and high salinity could be utilized advantageously in the IMC facility to inhibit vibrio growth. Richards prepared BHI agar plates containing NaCl at concentrations ranging from zero to ten percent, inoculated with V. cholerae non-O1 using the hockey stick method. These plates were incubated at 35°C for 24 hours to con- firm first the inhibitory effect of salinity on Non-Ol. The results are shown in Figure 10, comparing colony growth against a zero salt reference. Full seawater salinities (~35 ppt) reduced growth by approximately 5 orders of magni- tude. This experiment showed that salinity in excess of 30 RAYMOND 0 *#—.- “2 a ~ 1 = a . ” ° 5 -2 N 2 Sy ee Bo a cm -44 oo gs —— Q =5 - ® g = 6 = 6 —— 3 — (c} 7- al -8 - ) 2 4 6 8 10 Salt (NaCI) Concentration (%) Figure 10. Illustration of the inhibitory effect of salinity, measured as the quantity of NaCl added, on the growth of Vibrio cholerae non-O1. Cultures were grown on BHI agar for 24 hours at 35°C. ppt is detrimental to vibrio growth. A repeat of this experi- ment at —5, 4, and 10 degrees C showed no vibrio growth at salinities greater than 25 ppt, with salinity tolerance falling rapidly with temperature. The opportunity to con- firm this finding in the laboratory was unavailable. Impact on Process Performance Mortality of oysters within the IMC process was a major factor from the onset. A high mortality had been expected on the first load of oysters simply because of process shake down, and seeing that level cut in half by end of the second load was viewed to be in the right direction. However, be- ginning in late June, mortalities again rose and remained high. The circumstances surrounding these mortalities were not apparent to us. IMC was able to prepare firmly textured oysters with excellent flavor, purified to meet State and Federal health standards. Beginning in late June, survival in the holding tanks remained good for 48—60 hours, after which a large TABLE 2. Facility component sizes and costs. Relative Cost per Weighted Size to Square Cost per Activity Total Foot Ft? Feed Production 30 $15 $4 Shellfish Holding 40 $35 $14 Prep & Packing .20 $40 $8 Cold Storage 08 $40 $2 Dry Storage .02 $18 $1 Total $29 Fixed Components Offices 2000 ft? at $50 $100,000 Laboratory 600 ft? at $132 $ 79,200 Thus, Facility Cost is $180,000 + $29 per square foot of facility. INLAND COMMERCIAL SHELLFISH FINISHING 249 Historical Mortalities Records By Lots 100% 90% 80% 70% 60% 50% 40% Percent Mortality 30% 20% 10% 0% 26-Apr 24-May 24-Jun 02-Jul 2 1-Jul 30-Jul O09-Aug 20-Aug O06-Sep 20-Sep 28-Sep Date Received Figure 11. Historical mortalities of oysters received by IMC from April through September. An ability to control vibrio levels proved critical to the survival of shellfish subjected to the IMC process. number of the oysters would die, coinciding with the times able market response. Post mortem examination of the dead at which coliform counts were negligible in the tissue. showed large gas bubbles formed inside the cavity near the Oysters that survived this period of heavy mortality did heart. well, had reasonable shelf lives, and received a very favor- Several factors were surfacing simultaneously. First, TABLE 3. Operations analysis, comparing trade-offs in operating choices, assuming mortalities are kept to less than three percent. Base Certified Noncertified Noncertified Certified Value Cull/Wash Cull/Wash Certified C. gigas Unit Revenues $0.27 $0.27 $0.27 $0.27 $0.20 Variable Costs* Freight In $0.008 $0.008 $0.008 $0.008 $0.008 Factory $0.003 $0.003 $0.003 $0.003 $0.003 Raw Material $0.070 $0.145 $0.090 $0.030 $0.016 Supplies $0.030 $0.015 $0.015 $0.030 $0.030 Labor $0.028 $0.014 $0.014 $0.028 $0.028 Overhead $0.033 $0.016 $0.016 $0.033 $0.033 $0.172 $0.201 $0.146 $0.132 $0.118 Gross Margin $0.098 $0.069 $0.124 $0.138 $0.082 Percent 36.2% 25.6% 46.0% 51.0% 41.1% * Does not include cost of distribution, except for packing and package. 250 RAYMOND Sensitivity to Operational Alternatives 140.0% 130.0% 120.0% = > iv) fe) s 110.0% a a FS 100.0% [= ° oO 90.0% © 2. £ ia 80.0% 70.0% 60.0% 50.0% 70.0% 90.0% 110.0% 130.0% 150.0% Change from Base Case Value =» Freight + Factory ° Raw Mats 4 Supplies x Labor v Overhead Figure 12. Analysis of the impact changes in the cost of specific operating costs could have upon the total cost of operations. The base case condition is described in Table 3. When Crassostrea virginica is utilized, changes in the cost of raw materials (shellfish) are likely to have the greatest impact on process profitability. water temperatures in the Gulf increased, salinities dropped, and spawing began. Second, purification efforts were completed in shorter periods as process operations improved, but the more rapidly these improvements oc- curred, the higher the mortalities became. Thirdly, similar experiences were found with feeding. As shown earlier, feeding the oysters was found essential to uniform depura- tion. However, the more the oysters were fed, the more rapidly they seemed to die. Experience suggested that am- monia toxicity to the oyster gills might be the cause. Ammonia concentrations had been holding between 10 and 20 mg/l, below the limits found stress producing by Pruder’s group at Delaware. Regardless, a biologically- based nutrient stripping system, coupled with increased oxygenation of storage and recirculating waters, was put in place. Ammonia levels were maintained at less than | mg/1, and mortalities were reduced by about 20% from previous rates; however, they were still in excess of 30%. Following Richards’ suggestion of raising salinity and reducing temperature, mortalities were virtually eliminated overnight. As illustrated in Figure 11, mortalities dropped to less than 5%, and gas pockets within the coelomic cavity became rare. One interpretation of these data is that vibrio are not purged from the oyster gut by flushing with clean water and microalgae as coliforms and other non-attaching bacteria area. It might be conjectured that classical depuration re- leases vibrios from competitive pressures, and that they in- crease within the gut, feeding on organics excreted through the gut lining. It may be that oysters have an upper toler- ance threshold, such that, when surpassed, they experience gastroenteritis leading to death. Creating an environment that is adverse to vibrios may eliminate this sequence. Not creating this environment may be dangerous to public health. ECONOMICS A critical issue is whether a venture such as this can be made economically practical. IMC, as the initial pioneer in this effort, has provided sufficient hard data to solidify INLAND COMMERCIAL SHELLFISH FINISHING 251 some of the initial critical assumptions regarding items such as facility sizing and proportions, mass balances and mass flows, and actual direct costs associated with produc- tion operations. It has also shown that such an undertaking is technically feasible. What remains are some of the basic decisions fundamental to business operations, such as 1) what species, source, and form of supply is likely to pro- vide the highest returns on investment, 2) what are the most important costs to manage, 3) what are the greatest risks related to operations, 4) what remains to be done, and 5) what are the long term prospects. Species, Source & Form The following data are taken directly from IMC oper- ating and capitalization records, assessed for realistic con- ditions on the basis of lessons learned at the time IMC sus- pended operations. These data have been converted to unit costs for analytical purposes. The bases for this analysis are as follows: 1. Unit cost for operating a shellfish finishing facility are taken as the sum of the costs for transportation, Ww facility charges directly linked with product manu- facture, raw material costs, costs of supplies, direct labor charges, and direct overhead. These are the manageable cost items. Facility costs are derived on the basis of a capitaliza- tion of $55.10 per square foot, amortized at 12% over a fifteen year period. This results from the ac- tual square foot costs for each operational part of the IMC facility, corrected in size to efficient relative proportions, as follows: A. Building: The sum of manufacturing units plus fixed components B. Tanks and Equipment $283 per bushel weekly output ca- pacity 27 oysters per ft? holding area c. Shellfish Density 200 oysters per bushel. Revenues are based on current wholesale price re- ceived for the product F.O.B. Rifle, averaging $0.27 each. Sensitivity to Operational Alternatives 115.0% 110.0% 105.0% 100.0% 95.0% Impact on Gross Margin 50.0% 70.0% — Freight + Factory ° Raw Mats 90.0% A 130.0% 150.0% 110.0% Change from C. gigas Case Value Supplies X Labor Vv Overhead Figure 13. An analysis of the sensitivity of total operating costs to changes in the costs of specific operating parameters when Crassostrea gigas is utilized as the raw material within the IMC process. Under this scenario, management directed towards improvements in the cost of raw materials, labor, and overhead can produce significant positive changes in profit margins. 252 RAYMOND 5. Costs of raw materials are dependent upon source and condition. Louisiana oysters can be obtained certified, sorted, washed, and culled for $0.145 each, or they can be obtained in the sack and un- washed for $0.07 each. A quotation was received for uncertified oysters, dockside, culled and washed for $0.09 each, and for uncertified, unwashed, sacked for $0.03 each, as they would be if a depuration fa- cility were placed on the Gulf shores. The last source option considered was for certified C. gigas obtained in the Pacific Northwest for $0.016 each. An operations analysis for IMC, based upon the above information, is shown in Table 3. The principal factors affecting the cost of supply and form are the raw materials, supplies, labor, and overhead. Raw Material Only three options are open to inland facilities regarding raw materials, and those are the ones dealing with ship- ments of certified shellfish. Interstate transport of uncerti- BASE CASE: CERTIFIED & SACKED C. virginica Cold Store (1.6%) Unload (4.4%) b) Count/Pack (23.9%) Harvest (13.1%) Maintain (2.6%) Wash (6.5%) Load Tank (2.2%) 5 Pack (4.4%) Feed (8.7%) Impact of Wasteage on Percentage Cost Cold Store (1.3%) Count/Pack ae Harvest (10.6%) Maintain (2.1%) Li Load Tank (1.8%) Pack (3.5%) Feed (7.1%) Wash (8.1%) Unload (5.4%) fied stock is illegal, based on a concern by public officials that they may lose control over product quality to the en- dangerment of the consumer. This option has been included however to provide insight as to plant siting. Supplies Fewer supplies, in terms of gloves, aprons, benches, washers, boots, etc. are necessary of the oysters obtained are already culled and washed. However, washing can not be eliminated entirely since oysters leak and shed during transport; this would unnecessarily foul holding tank waters, as well as overload water treatment equipment. Labor and Overhead These are considered as one unit since labor is by far the major contributor to overhead costs. There are significant labor savings in obtaining culled and washed materials; however, they do not offset the added costs of raw mate- rial. Impact of Mortality on Percentage Cost Cold Store (1.1%) Unload (2.9%) Cull (22.1%) ount/Pack (48.6%) Wash (4.4%, Feed (5.9%) Pack (2.9%) Load Tank (1.5%) Maintain (1.8%) Harvest (8.8%) LEGEND Circle Diameter >» Diameter => Cost Figure 14. Illustration of the partial costs associated with shellfish feedlot processing, as effected by material wastage and mortality. In the absence of these losses, the costs associated with culling and packing contribute most to the total cost of operations. Culling becomes more significant when wastage is considered; however, the costs of counting and packing become overwhelming when mortalities are high. INLAND COMMERCIAL SHELLFISH FINISHING The highest value operational choice is the use of non- certified shellfish purchased from shellfisherman at the dock. From the standpoint of administrative burden, how- ever, there may be advantages to purchasing noncertified stock from a packer that already has the facilities and per- sonnel to do the job, although there will certainly be issues associated with the handling of certified and uncertified stock within the same packer house. Probably the cleanest option in terms of business and regulatory issues is the use of certified C. gigas which would entail a Pacific Coast location. Cost Sensitivities An analysis of the degree to which gross margin is sen- sitive to percentage change in each variable cost item is shown in Fig. 12. This shows impacts upon the profitability of the base case, which is equivalent to the initial process obligations at Rifle. This entailed the receipt of sacked and certified oysters from either Texas or Louisiana that were culled, washed, and sorted by IMC. Results show raw ma- terial cost to be the single greatest determinant of profit- ability. Labor, overhead, and supplies all rank about equally at about half the impact of raw material. However, PROCESS MANAGEMENT Establish Time/ Treatment Relations Develop Inventory Controls PROCESS ENGINEERING Document Holding System Validate Feed Production System PROCESS CONTROL Identify Control Develop DAC Points System Establish Key Parameters Develop Nolse- Free Signals TASKS COMPLETED Provide Quality Assurances Establish Mass Balances and Flows Define Domain Relations and Goals 253 since labor and overhead are tightly linked, when summed, they are as important as raw material. This leads to a conclusion that the most cost efficient facilities will be those with the lowest raw material cost and the best access to effective and manageable labor. A ten percent improvement in each of these factors against the base case could lead to as much as a twenty-five percent increase in gross margin. This analysis was repeated for the case utilizing C. gigas as the raw material source. Under this circumstance, Figure 13 shows raw material costs have about half the impact of either supplies, labor or overhead, reducing manageable costs to the most manageable factors. However, a ten per- cent improvement in each of these factors against the base indicates only as much as a 12% increase in gross margin. Primary Risks Primary risks in terms of operations focus on the proba- bility of product inventory loss, since unit costs will be in- curred regardless of whether the product reaches market or not. IMC was able to analyze its records with respect to the impact on labor associated with inventory wastage resulting from the shipment of unculled, unwashed oyster as com- Validate Vibrio Treatment Scheme Improve Cost Effectiveness Ammonia Removal Optimize Hydrodynamics Cleaning Train Pattern Recognition Validate Cost Savings TASKS REMAINING Figure 15. Illustration of the principal tasks associated with realizing profitable shellfish finishing under commercial operations. IMC has moved this concept significantly closer to reality, leaving the completion of tasks remaining to the next venture into this arena. 254 RAYMOND pared to clean. It was also able to look with some detail at the impact of mortalities on both cost and the labor pool. The man-hours required to process 500 bushels of oysters per week were partitioned among the labor activities asso- ciated with the process. These activities were unloading, culling, washing, feeding, packing, loading tanks, main- taining tanks and oysters, harvesting, counting and packing, and movement to/from cold storage. The results of these analyses are shown in Fig. 14, com- paring the relative costs of wastage and mortality to the base case. The partial costs of each labor activity are shown as sections of each pie; the size of the pie represents the relative magnitude of total labor cost. The major cost con- tributor is mortality, primarily as a result of the very large increase in effort to count and pack product after harvest, as well as to dispose of waste. The actual cost of wastage was not as large an increase as expected, adding primarily to the task of culling. The impact on unit cost associated with shipping was not a major factor considering the whole. This leads to a conclusion that control of Vibrio is abso- lutely essential to the economic finishing of C. virginica obtained from the Gulf. Without this control, inventory losses will occur and process economics are likely to be prohibitive. SEAFOOD FINISHING PROCESS DECISION TREE Quantity Adequate? Natural Production Controlled Production Tasks Remaining Fig. 15 shows the tasks and task structure IMC had pre- pared to validate the commercialization of its process. Many of the process management, engineering, and control tasks have been completed. However, additional work re- mains before the process is ready for optimization, in both technical and business areas. Technical tasks must focus on validating the Vibrio treatment work begun here. This work is crucial for the process to be practical. Opportunities exist for improving ammonia control, for providing improved water flows in tanks, and for providing more efficient tank maintenance —each affecting either inventory survival or labor. Efforts also need to focus on building historical pat- tern bases and defining the domains of operation necessary for artificial intelligence-based data acquisition and control systems to maximize their value. The capacity of this in- strument has to effectuate cost savings is very large, but— like all technical advances of this magnitude —this must be validated. Business efforts need to focus on marketing. The public needs to be made more aware of the advantages finishing processes have for making the food they eat both safer and better tasting. Flavors thought lost forever can be restored Direct Market Finishing Process Finishing Process |. Adequate with Respect to Capacity ll. Adequate with Respect to Consistency, Flavor, Texture, and Sanitary Quality Figure 16. Illustration of the logic sequence in deciding whether seafood finishing would be beneficial to shellfish companies and/or the in- dustry. Finishing processes have the potential to provide a higher percentage of available product to the market, consistently, and at higher quality. INLAND COMMERCIAL SHELLFISH FINISHING 255 within inland markets. Good marketing will lead con- sumers to this expectation more rapidly. Future Potential Fig. 16 illustrates the questioning process necessary to decide whether seafood finishing is needed or not. First, are the unit profits being obtained from natural supplies, in their current form, adequate and are they likely to stay that way? Second, is the quality of the product so good that improvements will not lead to increases in market share, price, or both? If the answer to the first question is no, aquaculture can help by augmenting natural supply. If the answer to the second question is no, whether from aquacul- ture or natural sources, seafood finishing is desirable. REFERENCES CITED Bacteriological Analytical Manual of the Division of Microbiology, Center for Food Safety & Applied Nutrition, U.S.F.D.A. 6th Ed. 1984. Collier, H. O. J., N. R. Campbell, & M. E. H. Fitzgerald. 1950. Vi- briostatic activity in certain species of pteridines. Nature (Lond.) 165:1004—1005. Doran, M., P. Shillam, R. E. Hoffman, & L. M. McFarland. 1989. Tox- igenic Vibrio cholerae O1 infection acquired in Colorado. MMWR 38(2):19—20. Dressel, D. 1988. Vibrio vulnificus in raw oysters. Memo to ISSC Execu- tive Board. Furfari, S. A. 1966. Depuration Plant Design. U.S. Department of Health, Education and Welfare. Public Health Service, Division of Environmental Engineering and Food Protection. Furfan, S. A. 1976. Shellfish Purification: A review of current tech- nology, FAO Technical Conference on Aquaculture. 16pp. Furfari, S. A. 1987. Current shellfish purification practices. personal communication. Kelly, M. T. 1982. Effect of temperature and salinity on Vibrio vulnificus occurrence in a Gulf Coast environment. Appl. Environ. Microbiol. 44(4):820-824. Kelly, M. T. & E. M. Stroh. Occurrence of Vibrionaceae in natural and cultivated oyster populations in the Pacific Northwest. Diagn. Micro- biol. Infect. Dis. 9(1):1—5, 1988. Kok, B. 1953. Experiments on photosynthesis by Chlorella in flashing light. In: J. B. Burlew (ed.) Algal Culture: From Laboratory to Pilot Plant, Carnegie Inst. Wash. Publ. 600:673-75. Laws, E. A., K. L. Terry, J. Wickman, & M. S. Chalup. 1983. A simple algal production system designed to utilize the flashing light effect. Biotech and Bioeng XXV. p. 221. Merkel, J. R. 1972. Influence of salts on vibriostatic action of 2,4-Dia- mino-6,7-Diisopropy! pteridine. Arch. Mikrobiol. 81:379—382. Pavis, A. T., J. F. Campbell, P. A. Blake, J.D. L. Smith, T. W. McKinley, & D. L. Martin. 1987. Cholera from raw oysters shipped interstate (Letter) JAMA 258:2374. Phillips, J. N. & J. Myers. 1954. Growth rate of Chlorella in flashing light. Plant Physiol. 29:152-161. Pruder, G. D. & E. T. Bolton. 1976. System configuration and perfor- mance: Bivalve Molluscan Mariculture. Proc. World Maricul. Soc. pp. 747-759. Raymond, L. P., P. K. Bienfang & J. A. Hanson. 1974. Nutritional con- siderations of open sea mariculture. In: J. A. Hanson (ed.) Open Sea Mariculture: Perspective, Problems and Prospects. pp. 129-182. Raymond, L. P. 1979. Initial investigations of a shallow-layer algal pro- duction system. Contribution to ASME Solar Energy Conference, San Diego, Calif., March 12—15. ASME Publication 79-Sol-34. Raymond, L. P. 1982. Aquatic biomass as a source of fuels and chem- icals. In: S. W. Yuan (ed.) Energy, Resources and Environment: Pro- ceeding of the first U.S.-China Conference Pergamon Press. pp. 75-82. Scura, E. D., A. M. Kuljis, R. H. York, & R. G. LeGoff. 1979. The commercial production of oysters in an intensive raceway system. Proc. World Maricul. Soc. 10:624—630. Singleton, F. L., R. W. Attwell, M. S. Jangi, & R. R. Colwell. 1982. Influence of salinity and organic nutrient concentration on survival and growth of Vibrio cholerae in aquatic microcosms. Appl. Environ. Mi- crobiol. 43(5):1080—5. Singleton, F. L., R. W. Attwell, M. S. Jangi, & R. R. Colwell. 1982. Effects of temperature and salinity on Vibrio cholerae growth. Appl. Environ. Microbiol. 44(5):1047—1058. Terry, K. & S. Hock. 1985. Photosynthetic efficiency enhancement in modulated light: Dependence on the frequency of modulation. In: Solar Energy Research Institute Aquatic Species Program Review, March 20-21. SERI Publications SERI/CP-231-2700. Terry, K. L. & L. P. Raymond. 1985. System design for autotrophic pro- duction of microalgae. Enzyme Microb. Technol. 7:474—487. Twedt, R. M. Recovery of Vibrio parahaemolyticus and related halophilic vibrios. pl2.01—12.08. in Bacteriological analytical manual, U.S.F.D.A. Urban, E. R., G. D. Pruder, & C. J. Langdon. 1983. Effect of ration on growth and growth efficiency of juveniles of Crassostrea virginica. Jour. Shellfish Res. 3(1):51—57. Urban, E. R. & C. J. Langdon. 1984. Reduction in costs of diets for the american oyster, Crassostrea virginica, by the use of non-algal supple- ments. Aquaculture 38:277—291. y <= tH * 7 = POT ee) Bi) ol P's hea = a ‘ = ; ; . in@ _ ni) guy. tue ’ Sig, bail el - ahi Sp Pla " - sO ues FE <4 Cm é Ae = ‘. i Pe, 1). 6a - : * ws an == he oo tee Ld 7 ve se ‘inne ip a ERRATUM Food value of eurytopic microalgae to bivalve larvae of Cyrtopleura costata (Linnaeus, 1758), Crassostrea virginica (Gmelin, 1791) and Mercenaria mercenaria (Linnaeus, 1758) ANTONIETO TAN TIU, DAVID VAUGHAN, THOMAS CHILES and KIMON BIRD Journal of Shellfish Research 8(2):399—405, 1989. Table 1 should read as follows: TABLE 1. Average shell lengths (wm) and arcsine transformed survival percentages (p’) of bivalve larvae fed on different algal diets on subsequent times (days) after fertilization. p’ significantly different according to Student-Newman-Keuls test (a = 0.05). Feeding ration was based initially on 25,000 cells/larva, unless Diet otherwise indicated. md = missing data, nd = no data, n Mercenaria mercenaria Length n C. muelleri Ellipsoidon sp. I. aff. galbana Nannochloris sp. Not Feed C. muelleri Ellipsoidon sp. I. aff. galbana C. muelleri Ellipsoidon sp. Nannochloris sp. C. muelleri I. aff. galbana Nannochloris sp. Ellipsoidon sp. I. aff. galbana Nannochloris sp. C. muelleri Ellipsoidon sp. I. aff. galbana Nannochloris sp. Error Mean Square I. aff. galbana Not Feed C. muelleri Ellipsoidon sp. C. muelleri I. aff. galbana C. muelleri Nannochloris sp. Ellipsoidon sp. I. aff. galbana Ellipsoidon sp. Nannochloris sp. I. aff. galbana Nannochloris sp. I. aff. galbana 50,000 cells/ml C. muelleri Ellipsoidon sp. I. aff. galbana Ellipsoidon sp. 50,000 cells/mL Error Mean Square Survival Experiment MM1 (Day 9) 155? 160° 160 99F 95® 1984 1708S 149? 1778 1798 38.9 6048 794 6548 6648 103 6748 764 620.8 Experiment MM2 (Day 9) 1864 117° 147° 17548 133° 1854 145¢ 166® 1954 1834 nd 79.9 nd 6148 694 4gaBc nd 112.9 = nd = number of replicates (beakers). Crassostrea virginica Length n Experiment CV1 (Day 17) 928 4 243A 4 2364 4 96¢ 4 120° 4 247A 3 229A 4 1808 4 1898 4 238A 3 413.4 Experiment CV2 (Day 17) 19848 4 md md 1638 4 20148 4 135° 4 20348 4 1648C 4 133° 4 2144 4 2114 4 nd nd 397.6 Survival 464 684 STA 434 424 684 664 504 604 107.2 704 OB 70A 784 604 664 704 744 754 784 nd HH HHH Ww = nd Length Cyrtopleura costata n Survival Experiment CCI (Day 17) 120° 20548 2324 go? 77? 240A 1828 351.7 1614 md 1514 1914 888 Sb) oe 1644 1574 nd 718 491.1 4 58° 4 56cP 4 7648 4 42> 4 54cD 4 82A 4 678¢ 4 41> 4 69BC 4 60° 54.6 Experiment CC2 (Day 9) 4 295 md oc 3 644 4 5848 4 444 4 344 4 S1AS 4 352 4 Tiles nd nd 3 614 210.8 = arcsin Vp, where p = proportion of surviving larvae. Averages with superscripts of similar letters are not HH HHH as nd iilewwtern vluherers Ter eultiee Gee a3 ; nie ~ | , Only through the peer review system can we maintain high standards for contributions published in the Journal of Shellfish Research. It is a pleasure to thank those listed below who gave of their time and expertise to review manuscripts over the past two years. George Abbe S. K. Alexander William G. Ambrose William S. Arnold Peter J. Auster Herbert M. Austin Bruce Barber Robert C. Bayer Peter S. Beninger Norman J. Blake Walter Blogoslawski Mark L. Botton V. Monica Bricelj Diane J. Brousseau Anthony Calabrese Judith McDowell Capuzzo Melbourne R. Carriker Michael Castagna Kenneth Chew Fu Lin Chu Harvey Cook Cyr Coutourier Louis R. D’Abramo Michael Dadswell Richard F. Dame Joseph DeAlteris William D. DuPaul Ralph Elston Charles Epifanio Arnold G. Eversole Dave Feigenbaum Graham Fenwick William S. Fisher David W. Foltz Susan E. Ford Louis Gainey Rodman G. Getchell Ronald Goldberg Edith Gould Charles Griffiths Raymond E. Grizzle Robert R. L. Guillard Nancy Hadley Harold H. Haskins Dexter Haven Christopher M. Hawkins Peter B. Heffernan Herbert Hidu Thomas T. Hilbish Robert Hillman David L. Holland Roger Hughes John W. Hurst Lewis S. Incze Douglas S. James Douglas S. Jones Victor S. Kennedy James E. Kirkley John N. Kraeuter Jay S. Krouse Christopher J. Langdon Richard W. Langton Peter Lawton Louis Leibovitz Michael P. Lesser D. T. J. Littlewood Mark Luckenbach Richard A. Lutz Roger Mann John J. Manzi David Marino Islay D. Marsden R. Morales- Alamo Michael A. Moyer Steve Murawski Carter R. Newell Richard C. Newell Roger I. E. Newell Gilbert B. Pauley Charles Peterson Eric Powell Frank Robb Ginnette Robert Shawn Robinson Robert P. Romaire Terence W. Rowell Timothy M. Scott Lynda Shapiro Fredric M. Sherchuk Scott E. Siddall David Somerton Thomas Soniat Rob Stephenson John Supan Steven Tettelbach John Trembley D. R. Trollope Randal L. Walker Clement J. Walton Les Watling Earl Weidner Gary Wikfors Clarice M. Yentsch INFORMATION FOR CONTRIBUTORS TO THE JOURNAL OF SHELLFISH RESEARCH Original papers 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 paper should be carefully prepared in the style followed in Volume 8 Number 2, of the Journal of Shellfish Research (1989) be- fore submission to the Editor. Papers published or to be published in other journals are not acceptable. Title, Short Title, Key Words, and Abstract: The title of the paper should be kept as short as possible. Please include a ‘‘short running title’’ of not more than 48 char- acters including space between words, and approximately seven (7) key words or less. Each manuscript must be ac- companied by a concise, informative abstract, giving the main results of the research reported. The abstract will be published at the beginning of the paper. No separate sum- mary should be included. Text: Manuscripts must be typed double-spaced throughout one side of the paper, leaving ample margins, with the pages numbered consecutively. Scientific names of species should be underlined and, when first mentioned in the text, should be followed by the authority. Abbreviations, Style, Numbers: Authors should follow the style recommended by the fourth edition (1978) of the Council of Biology Editors [CBE] Style Manual, dis- tributed by the American Institute of Biological Sciences. All linear measurements, weights, and volumes should be given in metric units. Tables: Tables, numbered in Arabic, should be on sepa- rate pages with a concise title at the top. Illustrations: Line drawing should be in black ink and planned so that important details will be clear after reduc- tion 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 preferably should be pre- pared 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 re- duction. Originals of graphic materials (i.e., line drawings) are preferred and will be returned to the author. Each illus- tration 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. No color illustrations will be accepted unless the author is prepared to cover the cost of associated reproduction and printing. References Cited: References should be listed alphabet- ically at the end of the paper. 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 at the end of papers in Volume 8, Number 2, of the Journal of Shellfish Research or refer to Chapter 3, pages 51—60 of the CBE Style Manual. Page Charges: Authors or their institutions will be charged $50.00 per printed page. If illustrations and/or tables make up more than one third of the total number of pages, there will be a charge of $30.00 for each page of this material (calculated on the actual amount of page space taken up), regardless of the total length of the article. All page charges are subject to change without notice. Proofs: Page proofs are sent to the corresponding author and must be corrected and returned within seven days. Al- terations other than corrections of printer’s errors will 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: Particularly appropriate photo- graphs may be submitted for consideration for use on the cover of the Journal of Shellfish Research. Black and white photographs, if utilized, are printed at no cost. Color illus- trations may be submitted but all costs associated with re- production and printing of such illustrations must be cov- ered by the submitter. Corresponding: An original and two copies of each manuscript submitted for publication consideration should be sent to the Editor, Dr. Sandra E. Shumway. Department of Marine Resources, and Bigelow Laboratory for Ocean Science, West Boothbay Harbor, Maine 04575. Phone: 207-633-5572 FAX: 207-633-7109 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 resourses. The Association has a membership of more than 900 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. —lIf you are young, you will benefit from the experience of your elders. —lIf you are an elder, you will be rejuvenated by the fresh ideas of youth. —lIf you are a student, you will make most useful contacts for your job search. —lIf 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 30 US $ per year ($20 for students) and that includes the Journal and the Newsletter! NATIONAL SHELLFISHERIES ASSOCIATION—APPLICATION FOR MEMBERSHIP (NEW MEMBERS ONLY) Name: For the calendar year: Date: Mailing address: Institutionallafiiliation.if «) ae ri 4 7 e~. & : a = On ee poe a iy y @ a : 4 @e8 SA (a 7 He ee | ey “ - 2 cat. \4 a : we : * ae ss — 1 * i . \ \ a “= Pa \ zs S ( [ = - a = - hoe 7 0 eh? aX i : “a iy Journal of Shellfish Research, Vol. 9, No. 2, 261—272, 1990. REPRODUCTIVE CYCLE OF THE WESTERN AUSTRALIAN SILVERLIP PEARL OYSTER, PINCTADA MAXIMA (JAMESON) (MOLLUSCA:PTERIIDAE) R. A. ROSE,!? R. E. DYBDAHL! AND S. HARDERS! 'Fisheries Department Western Australian Marine Research Laboratories P.O. Box 20 North Beach, W.A. 6020, Australia ?Present address: Pearl Oyster Propagators Pty. Ltd. 7 Tabard St. Greenwood, W.A. 6024, Australia ABSTRACT The seasonal gonad development of the commercially important Indo-Pacific, silverlip pearl oyster, Pinctada maxima, was investigated as part of a mariculture program. Histological preparations from 1,328 adults from populations off the northwest coast of Western Australia were collected approximately twice monthly over a six year period (1982—1988) to examine the pattern of gametogenesis. Histological findings were further supported by visually scoring the gonads of 2,588 broodstock and by observing hatchery and field spawnings of approximately 10,000 oysters. Possible exogenous reproductive stimuli were also investigated. P. maxima was confirmed to be a protandrous hermaphrodite which matured as a male during year one at a shell height greater than 110 mm. Bisexuality was uncommon. The pattern of gametogenesis was shown to be similar in both sexes with the mean percentage of mature gametes being highest during the warmer austral months. Maturity indices also showed that both sexes followed a similar annual cyclical pattern in which maturity index was highest during months of warmer seawater temperatures and least during the cooler months. The breeding season extended from September/October to March/April with a primary spawning peak at the beginning of the season and a secondary one at the end. Both sexes were multiple spawners and females during hatchery spawnings released between 0.5 x 10° and 12 x 10° ova per spawning. Except for water temperature and possibly chlorophyll-a, the exogenous factors measured during this study did not provide any practical indication for predicting the onset or duration of the reproductive cycle in P. maxima. KEY WORDS: INTRODUCTION The Western Australia cultured pearl industry is worth in excess of $A 80 million ($US 65) per annum, making it Australia’s third most lucrative commercial fishing activity after rock lobsters and prawns. Currently the industry relies almost entirely on wild stock of the tropical, Indo-Pacific, silverlip (or goldlip) pearl oyster, Pinctada maxima (Jameson) for pearl culture. Reliance on wild stock is bio- economically risky for an industry based on large capital investments for pearl production (Dybdahl and Rose 1986). The extent of this risk is compounded if the industry fishes more wild stock to increase production or experiences un- expectedly high levels of mortality amongst oysters after collection, as occurred in Western Australia during the last decade (Pass et al. 1987). In addition, operating costs asso- ciated with the collection and transportation of wild oysters for pearl production are likely to continue to increase above the current value of $A 12 to $A 16 ($US 10 to 13) per oyster. For these reasons artificial propagation is being es- tablished as an alternative source of pearl oysters. Qualitative aspects of the reproductive biology of P. maxima previously described by Wada (1942, 1953a, b, and c) and Tranter (1958a), do not appear to vary greatly from those of P. margaritifera and P. albina (Tranter 1958b, c, d, e). Field observations on northern Australian fishing grounds by Wada (1953a and 1953c) found P. 261 reproductive cycle, gametogenesis, maturity index, spawning, mariculture, pearl oyster, Pinctada maxima maxima to be a protandrous hermaphrodite, reaching matu- rity as a male in the first year of its life (110-120 mm in shell height, SH) with the incidence of female sexuality increasing with age or size. The relationship between shell height and sex-ratio indicated that at least 30%—40% of individuals which survive to larger sizes change sex from male to female. Oysters collected from the wild had a sex ratio approaching 1:1 when their shell height was 2200 mm. Similarly, Tranter (1958a) confirmed the presence of a relationship between sex change and shell size/age, and conducted preliminary spawning trials with oysters from Queensland. For mariculture purposes and a better knowledge of the population dynamics of the wild stock, it is essential to un- derstand the reproductive cycle of P. maxima from Western Australia. Prior to this study, the only data on spawning seasons were those of Wada (1953a) who noted that the presence of juveniles on the fishing grounds in Torres Strait, Queensland coincided with a marked reduction in the gonadal development of pearl oysters. From this he sur- mised that the spawning period for this population was from October/November to February/March. This paper details the annual reproductive cycle of pearl oysters from populations off the coast of northwest Aus- tralia over a six year period. The results presented were determined from histologically sectioned gonads and from 262 ROSE ET AL. visually scored gonads of live oysters. These data were supported by observed hatchery and field spawnings. Envi- ronmental conditions possibly influencing the reproductive cycle were also investigated. MATERIALS AND METHODS Gonad Collection During the normal fishing period from March to No- vember, pearl oysters were collected mainly from the tradi- tional fishing grounds off Eighty-Mile Beach (between Lat. 18°30’ S; Long. 120°41’ E and Lat. 19°50’ S; Long. 120°51’ E) and occasionally from grounds near Broome and Onslow (Fig. 1). During the cyclonic period (De- cember—February), systematic sampling from the fishing grounds was not always possible. At these times, gonadal samples were obtained from large oysters no longer suit- able for culture which had been living for at least one year on pearl culture lease sites near Broome. Gonad samples were collected twice monthly for ap- proximately six years (September 1982—June 1988). A total of 1,328 P. maxima (>120 mm SH) were examined histologically to determine the gametogenic cycle of both sexes. Gonads of a 12 smaller oysters ranging from 95 to 115 mm SH were also sectioned to confirm the size at which maturity was reached. 18°S 26 WESTERN AUSTRALIA N 4 200 km ° ——s 32 116°E 122% 129 Figure 1. Location diagram of Western Australia showing the collec- tion area for pearl oysters off Eighty-Mile Beach. The gonads of P. maxima are not discrete organs. Re- productive follicles originate near the urogenital papilla proximal to the retractor muscle and proliferate within the connective tissue between the epithelium and viscerum. Sections taken from different regions of the gonad indi- cated that germ cell development through the gonad is rea- sonably uniform. Gonad tissue between the proximal end of the gut loop and base of the foot was excised to obtain the largest sections possible for quantative assessment of gametogenesis. Samples were placed in Davidson’s fixa- tive, preserved in 70% alcohol, dehydrated with serial dilu- tions of alcohol, embedded in paraffin and sectioned and stained with Harris’s haemotoxylin and eosin. Stages of gametogenesis were photographed at magnifications of 100 and 200 with a camera attached to a compound micro- scope. Gonad Developmental Stages Excluding the bisexual phases, the reproductive cycle of P. maxima was simplified into five broad gametogenic stages following a scheme developed by Tranter (1958b and c) for P. albina: Stage O: Indeterminate or inactive. No evidence of go- nadal development, except empty, collapsed follicles and connective tissue containing different types of granulocytes and phagocytes (Fig. 2A). Stage 1: Early gametogenesis. Testis: Follicles initially small and lined with stem cells and spermatogonia (Fig. 2B). As spermatogenesis proceeds, primary and secondary spermatocytes rapidly proliferate, filling-up the follicular lumen (Fig. 2C). Ovary: Follicles initially small, poorly formed and empty, with walls lined with stem cells and developing oocytes (Fig. 3A). Oogonia and early (or pri- mary) oocytes have little or no yolk, each with a large (blue-stained) nucleus, and often adhere to the follicular wall in clusters. As oogenesis proceeds, oogonia and young oocytes proliferate along the inside walls with a few larger oocytes beginning to elongate (Fig. 3B). Stage 2: Actively developing to near-ripe gameto- genesis. Testis: Follicles begin to enlarge with spermato- gonia and spermatocytes proliferating along the periphery of the lumen and with spermatids and some spermatozoa filling the center, their acidophilic tails appear as pink lines radiating from the center of the lumen (Fig. 2D). Near-ripe follicles have enlarged greatly with developing sperm ap- pearing as a dark blue band (several cells deep) around the periphery of the follicular wall which has decreased in thickness (Fig. 2E). Except for isolated pockets of sper- matocytes and spermatids, the follicular lumen is packed with spermatozoa. Ovary: Oocytes connected to the follic- ular wall have begun to accumulate yolk and expand into the lumen, with a few free oocytes appearing in the center (Fig. 3C). Near-ripe follicles are densely packed with mainly large elongated oocytes (with some showing both a nucleus and nucleolus) which are still connected to the fol- REPRODUCTIVE CYCLE OF Pinctada maxima 263 ES ge eee A cade —-” B Figure 2. Indeterminate sexual phase and various stages of male gametogenesis in P. maxima. (A) Indeterminate phase (Stage 0). (B) Beginning of gametogenesis (Stage 1) with stem cells (s tc) and spermatogonia (sg) proliferating along inside of follicular walls. (C) Advanced stage of early gametogenesis (Stage 1) with lumen of follicles willed with spermatogonia, spermatocytes (sc) and a few spermatids (st). (D) Actively developing testis (stage 2) showing stem cells embedded along the inside wall of the follicles, spermatogonia and spermatocytes along the periphery of the lumen, and spermatids and spermatozoa (sz) in the center. Note tails of sperm (arrows). (E) Near-ripe testis (Stage 2) with developing sperm shown as a dark band, several cells wide, around periphery of follicles and with spermatozoa occupying the centers. (F) Spawning-ripe testis (Stage 3) with confluent follicles almost entirely filled with spermatozoa. (G) Partially spawned testis (Stage 4). (H) Spent testis (Stage 4) showing residual spermatozoa (r sz), phagocytes (pc) and re-development occurring along inside walls of follicles (Stage 1). 264 ROSE ET AL. sy og Sxa § Figure 3. Various stages of female gametogenesis in P. maxima. (A) Beginning of gametogenesis (Stage 1) with stem cells (st c) and oogonia (og) appearing along the inside wall of follicles. (B) Advanced stage of early gametogenesis (Stage 1) with follicles showing oogonia and young oocytes (y oc) proliferating along inside wall and the presence of a few older, larger oocytes (0 oc). (C) Actively developing ovary (Stage 2) with lumen of follicles beginning to fill with connected oocytes (c oc) and a few free oocytes (f oc). (D) Near-ripe ovary (Stage 2) with lumen of follicles occupied with both free and connected oocytes. (E) Spawning-ripe ovary (Stage 3) with confluent follicles almost entirely filled with free oocytes. (F) Partially spawned ovary (Stage 4) with the appearance of small amounts of resorptive tissue (rt) within follicular lumen. (G) Partially spawned ovary (Stage 4) slightly more advanced with follicular lumen occupied by residual oocytes (r oc) surrounded by large amounts of resorptive tissue and with re-development occurring along inside walls (Stage 1). (H) Spent ovary (Stage 4) showing practically empty follicles filled with resorptive tissue and a few residual, free oocytes. REPRODUCTIVE CYCLE OF Pinctada maxima 265 licular wall by a long, narrow stem of yolk material (Fig. 3D). Stage 3: Spawning-ripe. Testis: Follicles distended, confluent and almost entirely filled with spermatozoa. Spermatocytes and spermatids are restricted to lining the follicular walls which have become increasingly thinner with maturation (Fig. 2F). Ovary: Confluent follicles packed with almost entirely free, polygonal-shaped oocytes displaying both a nucleolus and nucleus (Fig. 3E). Stage 4: Partially spawned to spent. Testis: Gonad con- tains follicles with partially empty lumen. Those which are still full have a gap between the follicular wall and mass of spermatozoa remaining in the lumen (Fig. 2G). Partially spawned follicles contain phagocytes amongst sperma- tozoa. Spent follicles are empty except for small pockets of residual sperm and phagocytes inhabiting the lumen. Rede- velopment can be seen along the walls of some follicles (Fig. 2H). Ovary: Follicles are partially empty, with small amounts of resorptive material occurring in the space be- tween free oocytes which have become rounded or pear- shaped (Fig. 3F). Follicles which are almost completely spent have extensive redevelopment occurring along the in- side follicular wall, with large amounts of resorptive mate- rial surrounding free oocytes undergoing cytolysis (Fig. 3G). Spent follicles are almost entirely empty with no sign of gametogenesis except for isolated regressing oocytes surrounded by resorptive tissue, phagocytes and interstitial connective tissue (Fig. 3H). Bisexual phase: Although this sexual condition was rarely observed, two forms of bisexuality could be distin- guished. The less frequent showed both sexes developing concomitantly in the same follicle (Fig. 4A). The second form exhibited one sexual phase overlapping with the other; for example, when oogenesis had commenced before residual sperm had been completely removed from the fol- licular lumen (Fig. 4B). Gametogenesis was a continuous process in both sexes and a distinction between the various stages was not always possible, with some stages overlapping within the same gonad. Therefore, actively developing and near-ripe ga- metes were combined in Stage 2 and partially spawned and spent gametes in stage 4. Quantitative Analysis of Histological Data The process of gametogenesis was described on a monthly basis by viewing a histological section of each go- nadal sample at low magnifications (40 x and 200) to classify it as indeterminate, male or female. These sections were then scored at 400 x magnification by counting dif- ferent types of easily distinguishable gametes (described below) from three replicate positions randomly selected over each section. Replicate counts were then averaged to give the mean proportion of each type of gamete present per individual per month. Three types of gamete stages were readily distinguishable for males and four types for % * Figure 4. Bisexual phase of P. maxima. (A) Both sexes actvely devel- oping in the same follicle (Stage 2) with sperm surrounding oocytes (oc). (B) Ovary beginning gametogenesis with residual spermatozoa (r sz) occupying lumen of follicle (f). females. Results presented were derived from a sample size of 613 males (x = 51 oysters/mo.), 496 females (x = 42 oysters/mo.) and 219 indeterminates (x = 18 oysters/mo.). Spermatogenesis was assessed by calculating the per- centage of early (immature, tail-less), free (mature, tailed) and residual (regressing or spent) sperm in one cm? of fol- licle. An ocular graticule, subdivided into 100 sq mm grids, was superimposed over a follicle to count the number of grids occupied by each type of sperm. A grid was counted if at least 75% of its area was filled by one of the three types. Oogenesis was measured by counting the number of early (immature), connected (growing to near-ripe), free 266 ROSE ET AL. (spawning-ripe) and residual oocytes within an entire oc- ular field. Only gametes completely within view were counted. Free oocytes were defined as only those gametes occupying the lumen and displaying both a nucleus and nu- cleolus. In addition, for females collected from October 1982 to December 1985, the diameter of 30—50 free oocytes (or the most mature oocytes present within a follicle) was mea- sured along the longest axis. A total of 216 oysters (x = 7 individuals/mo.) were sampled, involving 2,125 oocytes. A monthly maturity index (MI) was also calculated for individuals of both sexes using a modified version of a for- mula developed by Seed (1969): MI = mean proportion of gametes at a given developmental stage for three replicates per individual x numerical ranking for that stage. For males: MI = [mean proportion of residual sperm x 1] + [mean proportion of early sperm Xx 2] + [mean proportion of free sperm X 3]. For females: MI = [mean proportion of residual oocytes < 1] + [mean proportion of early oo- cytes X 2] + [mean proportion of connected oocytes x 3] + [mean proportion of free oocytes x 4]. A cost-benefit analysis of preliminary results indicated that to obtain an accuracy within 95% confidence limits a monthly sample size of 15 specimens of each sex was re- quired for measurement. Except for two of the six years, it was not possible to collect this number of each sex during the cyclonic months (December to February). During the winter months (June to August), it was not always practical to collect this many of each sex because almost all indi- viduals were sexually indeterminate. As a result of these limitations, monthly sample sizes over the years ranged from 5 to 23 specimens per sex. Monthly mean egg diameters were compared over three years using a one-factor analysis of variance (ANOVA). Monthly mean proportion of the various gametogenic stages for each sex were compared over the six years using a two-factor ANOVA. Heterogeneity of variances among stages required arc sinV (proportion) transformation of monthly data. Significant differences between means were analyzed with a Student Newman Keuls Multiple Range Test (SNK, P < 0.05). Gonadal Development of Live Oysters The gonads of 2,588 oysters (greater than 150 mm SH) used for broodstock in a mariculture program were visually scored for development every month from January 1987 to April 1989. In addition, over the six years of this investiga- tion approximately 10,000 oysters, which were used as broodstock for either spawning or gonad-conditioning ex- periments, were inspected. Sexual development of gonads of wild and farm-held pearl oysters were staged, as follows: Stage 0: Gonad tissue flaccid or invisible, sex indeter- minate. Stage 1: Gonad visible but proliferation to gut loop was slight and proximal, gonad appeared granular and difficult to sex by color (male-white and female-yellow). Stage 2: Sex easily determined by color, tissue had pro- liferated distally along lateral walls of gut loop and ap- peared semi-confluent (at this stage spawning could occur but gametes were usually immature or non-viable). Stage 3: Gonad ripe and bulging, gonad tissue extended over the surface of stomach, gut loop and digestive gland; gonad appeared confluent and when pierced diffused pro- fusely. Each developmental stage was routinely confirmed mi- croscopically by withdrawing a sample of gametes with a syringe inserted into the gonad of live male and female oysters. Hatchery and Field Spawnings Oysters at Stages 2 and 3 were used in over 40 hatchery spawning trials in which either or both sexes participated. During these trials oysters were induced to spawn by tem- perature manipulation or in combination with serotonin in- jections, ultra-violet irradiated sea-water and sperm suspen- sions (Rose et al. 1986). Records were also kept of oysters observed spawning naturally in tanks on board transporta- tion vessels and on pearl culture lease sites; when possible their gametes were collected for rearing. This information was used to augment findings from visual and histological gonad data. Possible exogenous reproductive stimuli were investi- gated during this study by recording the following param- eters: water temperature, salinity, turbidity and chloro- phyll-a (as a measure of the amount of phytoplankton). The level of the following nutrients in seawater were also mea- sured: orthophosphate phosphorous (PO,_P), total phos- phorous (Total-P), ammonium nitrogen (NH,_N), nitrite plus nitrite nitrogen (NO, + NO _N) and total Kjeldahl nitrogen (Total-N). For comparative purposes the above water quality parameters were measured by standard tech- niques described in earlier studies (Pass et al. 1987). RESULTS Sexuality Histological and visual examination of small pearl oysters indicated that the male sex matured first, at a shell size of 110 mm SH or larger. Female development usually occurred when shells were larger than 135 mm SH. Histo- logical examination of 395 oysters (>150 mm SH), col- lected off-shore from Eighty-Mile Beach during October and November 1984, revealed that 49% were male, 38% female, 1% bisexual and 12% indeterminate. Bisexuality was uncommon and numerous field and hatchery spawnings provided no evidence that P. maxima was a functional hermaphrodite. REPRODUCTIVE CYCLE OF Pinctada maxima Pearl oyster broodstock monitored for gonadal develop- ment took at least five weeks to mature from the indeter- minate/early development stages to the spawning-ripe stage regardless of sex. Under hatchery conditions both sexes were multiple spawners, with females releasing between 0.5 x 10° and 12 x 10° yellow-colored ova per indi- vidual. On at least three occasions, oysters changed sex between seasons. Under intensive husbandry conditions, the male sexual phase was observed more frequently amongst broodstock and females rarely developed to the near-ripe stage. Reproductive Cycle The mean percentage of mature gametes in either sex was greater during the austral spring and summer (Oc- tober—February/March) and less during the late autumn and winter (May—August). Conversely, the incidence of inde- terminate gonads was greater during the cooler months and less during the warmer months. The mean percentage of the area in monthly gonadal sections occupied by each of the three spermatogenic stages varied significantly among the six years (two-factor ANOVA, P < 0.001). Differences were largely due to nat- ural variations in the length and intensity of each annual breeding period. When the samples over the years were pooled, the percentage of spermatozoa (tailed free sperm) in monthly samples was greater than 60% from October to March and less than 40% from April to September (Fig. 5). In contrast, the percentages of residual sperm were greater from April to September and less from October to March. Monthly percentages of immature, tail-less (early) sperm followed a similar but weaker pattern to that observed with free sperm. Immature sperm occurred more frequently from 100; 90: 80: 70: 60: 50: 40: 30: HL, Vi REF REF REF REF REF REF REF Jan Feb Mar Apr May Jun Jul Figure 5. Annual spermatogenic cycle of P. maxima presented as the mean percentage of residual (R), early (E) and free (F) sperm occu- pying monthly gonad sections involving 613 oysters collected from October 1982 to June 1988. Percentages plotted were derived from montly samples pooled over six years, (x = 51 oysters/mo.). Mean Percentage Le) 102: WUCAaaaaaaaaaaaaaaaaaaaaaay AAeVVVAAAeeeeaaaacaaaaaaagunal SNSSSSSSSSSSSS JAXX AAAAACACLLLAaaaaaaaSSaaaaaaaaaaay Sa eS ak } i. REF REF REF REF REF Aug Sep Oct Nov Dec NI ASS 267 100 90 80 | 70 | 60 | 50 | 40 | 30 | Mean Percentage xxxxx> 5 x x x x lel lel lel 1M illo} om 4 ha AM ol] RECF RECF RECF RECF RECF RECF RECF RECF RECF RECF RECF RECF Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Figure 6. Annual oogenic cycle of P. maxima presented as the mean percentage of residual (R), early (E), connected (C) and free (F) 00- cytes counted in monthly gonad sections involving 496 oysters col- lected from October 1982 to June 1988). Percentages plotted were derived from pooled monthly samples (x = 42 oysters/mo.). April to June than during any other time of the year (SNK, P5<10!05): Like spermatogenesis, the mean percentage of each of the four oogenic stages counted in monthly gonadal sec- tions varied significantly over the six years (two-factor ANOVA, P < 0.001). When data for all years were pooled, the percentage of free oocytes was highest from September to March and lowest from May to August (Fig. 6). The monthly percentages of connected oocytes followed a similar pattern to that observed for free oocytes except during July and August. During this period, they were not significantly different from September through February/ March (SNK, P < 0.05). Immature (early) oocytes oc- curred more frequently from April to August and less fre- quently from September to March (SNK, P < 0.05). The mean diameter of oocytes provided another indica- tion that female gonads were not full with mature gametes during the austral winter (Fig. 7). The diameter of free and/ or connected oocytes from monthly samples was signifi- cantly smaller (less mature) during July and August and larger from September to May (ANOVA, P < 0.001). The relationship between the size of oocytes and time was con- sistent with changes observed histologically in ovarian tissue and macroscopic development. The maturity indices calculated for both sexes show that both the testis and ovary followed similar cyclical patterns in which maturity was highest during the warmer months and least during the cooler months (Fig. 8). Variations in overall amplitude of the curves for each sex were due to different scales of measurement adopted for each index (i.e., three developmental stages for testes and four for ovaries). Visual inspection (along with the occasional monthly i) fon) oo nm ow b a ro POTS E OUR UO, et OR: nO = oO ONDJ FMAMJJASONDJ FMAMJJASONDJ FMAMJJASOND 1982 1983 1984 1985 Figure 7. Mean oocyte diameter (+ standard error) of monthly gonad samples of P. maxima collected 1982 through 1985, involving a total of 216 females (x = 7 individuals/mo.) and 2,125 oocytes. =p Mean (4s.e.) Oocyte Diameter (ym) biopsy) of the gonads of 2,588 oysters from January 1987 to April 1989 revealed that the combined percentages of indeterminate (Stage 0) and early developers (Stage 1, both sexes) were greatest during autumn and winter and least during spring and summer (Fig. 9a, b). Conversely, per- centages of near-ripe oysters (Stage 2, both sexes) were least during the colder months and greatest during the warmer months. Spawning-ripe oysters (Stage 3, both sexes) were rarely observed and represented less than 8% of the monthly samples taken. Except for temperature and possibly chlorophyll-a, the exogenous factors measured during this study did not pro- vide any practical or obvious cue for predicting the onset or duration of the breeding cycle in P. maxima. A composite breeding season derived from records of field and hatchery spawnings taken over six years was superimposed onto a three-year time series graph of surface sea water tempera- 3.54 2.0} 1.54 Maturity Index 1.04 - 82 83 84 85 86 87 88 89 Year Figure 8. Maturity index of P. maxima females (dash) and males (solid) from monthly samples collected from October 1982 to June 1988. The index was based on 496 females and 613 males. ROSE ET AL. 100: 90: 80: 70: 601. 50; (a) * os * * Percentage * 10; ° : Q — — -0 ee -o eo» +o —_-__--¢ Oct OL oe Jan Apr Jul Oct Jan Apr Jul TPAC), 7 T 90: 80: 70: 60: 50: 40: n | & 30| a “Ns 20: : 10} s ~ Percentage ae a a~ — Sa Jan Apr Jul Oct Jan Apr 1988 1989 MONTH-YEAR Figure 9a and b. Monthly gonad development of P. maxima brood- stock from January 1987 to April 1989. Percentages of different devel- opmental stages for each month and sex are shown: *—* indeter- minate (Stage 0) and early developers (Stage 1); 0- -0 male Stage 2; @—@ male Stage 3; A— —A female Stage 2; A—A female Stage 3. ol oa ee =— Jan Apr Jul Oct 1987 & a Se ee tures (Fig. 10). The results indicated that P. maxima begins to breed during the annual, rapid rise in water temperature in September/October and continues over the summer months before ending when the temperature drops in March/April. Broodstock or wild oysters could not be in- duced to spawn later than April when water temperatures had fallen. A similar relationship was observed with chlorophyll-a levels except that seasonal peaks were less pronounced during periods of highest water temperature (Fig. 10). Sea- sonal variability in phytoplankton productivity was prob- ably dampened by low nutrient availability throughout the year in tropical coastal waters off Broome (Table 1). DISCUSSION Gametogenesis The overall pattern of gametogenesis in P. maxima is similar to that described for P. albina and P. margaritifera from the Torres Strait, Queensland (Tranter 1958b to e). REPRODUCTIVE CYCLE OF Pinctada maxima 269 34) i 45 32) S a a =I © 30 ON 13 ® r * > = 28) g [ \ y a 2 2G) Gy f J 28 5 3, ; } ©) E 24) 4 § o i ~ - cok = s \ i: p 12 = DOW. 8 pb ome ‘ , = 20) ¢ € | SS eee a ee ————— “98 AMJJSASONODO FMAMJJASONDJFMAMJJASONDJFMA 1986 1987 1988 1989 Month-Year Figure 10. The annual breeding season of P. maxima derived from six years of recorded field and hatchery spawnings (horizontal bars). Also plotted are the surface sea-water temperatures (°C) (0—0) and chloro- phyll-a levels (— —) at the Broome Jetty, Roebuck Bay from April 1986 to April 1989. Cyclical temperature curve plotted was determined by the following equation: °C = 26.95 + 2.46 sin g + 4.09 cos g — 0.78 sin (26) — 0.96 cos (29), when ¢ = 2 (day of year)/365, r? = 0.93 and n = 150. The rudimentary follicles and stem cells form in a similar manner, as do the appearance and distribution of the germ cell stages. In all three species the course and rate of game- togenesis is similar. Histological evidence indicates that spawning in P. maxima is not complete with some overlap occuring be- tween early, ripe and spent stages similar to P. albina. The follicles of recently spawned individuals of both P. maxima and P. albina possess more residual products and phago- cytes than P. margaritifera. Moreover, in both species re- gression is incomplete before commencement of the next developmental cycle. The connected or free oocytes in P. maxima, however, have one or two ‘‘yolk nuclei’ (7-8 zm diameter) embedded in their cytoplasm like those in P. margaritifera (Tranter 1958e). These organelles which ap- pear as circles, crescents or saucers do not occur in P. al- bina (Tranter 1958e) nor is their function understood. The breeding seasons of each of these species also differ. P. maxima breeds annually between September and April with peaks at either end. P. albina breeds contin- uously but with a peak in activity during April and May while P. margaritifera has two distinct cycles from March to August and from September to February (Tranter 1958d and e). The findings of this study support those of Wada (1953a) and Tranter (1958a) who both considered P. maxima to be a protandric hermaphrodite. The male and female phases are typically separated by time but occasion- ally they occur together in the same gonadal follicle as de- scribed for P. (fucata) martensii (Ojima and Maeki 1955) and P. albina (Tranter 1958d). Evidence from hatchery spawnings suggest that the bisexual phase is nonfunctional in P. maxima and it occurs more frequently when a new sexual phase is beginning and less frequently when both sexes are actively developing concomitantly. The ability of P. maxima to change sex after a certain size appears to be typical for members of the genus, for example: P. (fucata) martensii (Ojima and Maeki 1955); P. albina (Tranter 1958d); and P. margaritifera (Tranter 1958e). This phenomenon which also occurs in Ostreidae, Teredinidae and Pectinidae (Tranter 1958d) may be related to a ‘‘weak hereditary sex-determining mechanism”’ as hy- pothesized for P. albina (Tranter 1958b). Similarly, P. maxima may also inherit the ability to develop either sex and that the sexual phase of rudimentary germ cells is de- termined physiologically by variations in levels of the body’s food reserves. Cells would then differentiate into female gametes when food levels were high and into male gametes when low. A physiological basis for explaining the onset of gametogenesis has been shown to be related to the TABLE 1. Surface water temperature, phytoplankton nutrients and chlorophyll-a values (mean + standard deviations and range) at the Broome Jetty sampling site for the period April 1986 to April 1989. Data are based on approximately monthly water sample collections over the three years (n = 39). For comparison, values from surface water samples taken from further offshore on the main pearl oyster collection grounds are reproduced from Pass et al. (1987, Table 2, p 157). Parameter Water Temp PO,-P Total-P NH,-N NO, + NO,-N Total-N Chlorophyll-a Location (CC) (pg/l) (pg/l) (pe/l) (pe/)) (peg/)) (pg/l) Broome Jetty Surface 26.9 + 3.8 6.2 + 5.5 16.2 + 6.8 10.3 + 9.3 2.8 + 1.4 181.3 + 72.4 0.7 + 0.4 (21.0—32.3) (1-24) (2-35) (2-52) (1-6) (48-314) (0.01—1.78) Bottom 26.8 + 3.6 54 4 16.9 + 7.9 10.2 + 8.2 238) = 14 160.1 + 58.7 0.9 + 0.5 (19.8—32.1) (1-19) (4-43) (1-33) (1-7) (37-258) (0.19—2.13) Collection Grounds Surface 24.0 + 2.8 4.9 + 1.3 30.4 + 13.5 5.9 + 4.8 Saif 25 WS) Sil se Abily 0.3 + 0.4 (20.0—26.8) (3-7) (17-62) (1-13) (3-6) (74-645) (O—1.35) Bottom 2378) 2-7) 5.2 + 0.8 26.9 + 5.8 4s5e3)0) 4) = el 331.9 + 179.7 (Osh aa (07) (20.0—26.7) (4-6) (19-35) (1-9) (3-6) (107-563) (0.1—0.6) 270 ROSE ET AL. storage of glycogen in the digestive gland of the bay scallop, Aequipectin irradians (Sastry and Blake 1971). More recently, the extent of sexual differentiation in the tropical mussel, Perna perna, has been shown to be posi- tively correlated with food availability at favorable temper- atures (Velez and Epifanio 1981). The mechanism for egg maturation in P. maxima is more similar to that of P. albina (Tranter 1958c) than to that of P. margaritifera (Tranter 1958e). Meiosis in the ovarian oocytes of P. maxima and P. albina never pro- gresses beyond prophase (as indicated by the presence of an intact germinical vesicle). Subsequent maturation of free oocytes occurs outside the follicle at the onset of spawning. By contrast, disappearance of the germinal vesicle and completion of the first division of meiosis in free oocytes of P. margaritifera occurs within the ovary before the onset of spawning. Excitation of the ovarian oocytes can be induced chemi- cally by penetration of the membrane with weak solutions of cations with high permeability (e.g., K*, NH,* Ba?*) or physically by thermal and electrical energy during hatchery spawnings (Iwata 1952). This in turn, stimulates the discharge of gametes from gonadal follicles. Excised oocytes in P. maxima do not become activated either in the presence of seawater or mature spermatozoa. If exposed to low concentrations of ammonium hydroxide (Wada 1942, 1953b; personal observations), the germinical vesicle is broken down and fertilization occurs. Similarly, sperma- tozoa excised from the testis can only be activated in the presence of ammoniated seawater. As found with P. fucata from India (Alagarswami et al. 1983), P. maxima exposed to low concentrations of ammonium hydroxide spawn ga- metes. However, unless the gonads of P. maxima are suit- ably ripe the gametes released are invariably less viable and the percentage of fertilization is low (<30%) and subse- quent abnormal larval development is high (>85%). Reproductive Cycle Histological examination of the ovarian and testicular tissues of P. maxima from Western Australia show a defi- nite annual reproductive cycle with maximal and minimal developmental periods consistent with correspondingly high and low water temperatures (Fig. 8). Wada (1953a) found that the macroscopic gonadal development in P. maxima from the Arafura Sea, northern Australia, follows a similar pattern but that mature oysters could be observed outside the main breeding period, during the cooler months. He surmised that P. maxima was potentially ca- pable of spawning throughout the year. Histological evi- dence presented indicates that the gonads of either sex from Western Australian populations are unlikely to be suffi- ciently ripe to successfully spawn during the colder months (males contained <25% free sperm (Fig. 5) and females <2% free oocytes (Fig. 6). The presence of mature ga- metes during the colder months is not evidence of potential, spawning-ripe gonads but rather an indication that mature and residual developmental stages overlap. Visual monthly monitoring and biopsies of broodstock and wild oysters also suggest that spawnings are unlikely during the colder months (Fig. 9a b). Differences observed between popula- tions are most likely due to differences in latitude. Western Australian populations studied here are from 7 to 11 de- grees south of those from the Northern Territory and Queensland. Thus, P. maxima experiences a colder (min- imum 18°C) and wider range (18°C—32°C) of water tem- peratures which may markedly reduce the incidence of ma- ture oysters during the colder months. Histological data for P. maxima indicate that during the cooler months the sequence of spermatogenic development is different to that of oogenesis. In the testis, residual sperm are the dominant gametes; while, in the ovaries, early 00- cytes are dominant. This apparent sequential difference is likely to be associated with the gametogenetic classification scheme used (i.e., three stages for males and four stages for females). Although histological data for P. maxima from Western Australia did not reveal any obvious bimodal reproductive cycle, field and hatchery spawnings indicated two main spawning periods: a primary one associated with a rise in water temperature during October-December and a sec- ondary one with a fall in water temperature during Feb- ruary—April (Fig. 10). These spawning periods are in agreement with those found for P. maxima populations from the Torres Strait, northern Queensland (Wada 1953a; Minaur 1969; Tanaka and Kumeta 1981). Although other factors could be used to explain the reduced occurrence of observed spawnings between the main spawning periods (e.g., not enough time or energetic resources for the oysters to initiate further sexual activities), the discrepancy be- tween histological data and observed spawnings may be re- lated to high water temperatures during January and Feb- ruary. Within this period, hatchery spawnings were often weak producing a high proportion of non-viable gametes. Temperatures used to induce broodstock were two to four degrees above ambient seawater temperature which was al- ready near the upper limit experienced by wild oysters (Fig. 10). Thermal stress has been shown to lower fecundity and reduced egg size in Mytilus edulis (Bayne et al. 1978) and inhibit gonadal development in Perna perna (Velez and Epifanio 1981). P. maxima may be somewhere in between these two species: high temperatures may not prevent ga- metes from fully developing within the gonad but they may prevent or inhibit their release. As the temperature drops and becomes more favorable, the large reserve of con- nected oocytes present in the gonads of females (Fig. 6) quickly matures and spawning resumes. As electrophoretic studies to date have not detected any genetically different stocks within Western Australia (Dyb- REPRODUCTIVE CYCLE OF Pinctada maxima 271 dahl, unpublished data), a lack of a distinct bimodal game- togenic cycle may be the result of the method of oyster collecton. Oysters sampled during this study inhabited a wide zoogeographic range (Fig. 1) and variety of different microhabitats (e.g., deep water (30—35 m), shallow water (10-15 m), pearl culture farms (2—12 m). This would ef- fectively confound any detectable differences between sub- populations. If separate subpopulations of P. maxima are examined, gametogenic cycles may be detected. Spawning and gonad monitoring records suggest that pearl oysters from deep water fishing grounds off Broome consistently produces the most intense spawnings and the greatest per- centage of viable gametes. Oysters from pearl culture farms generally release fewer and/or less viable gametes per spawning. Pearl oysters from southern populations near Onslow (Fig. 1) generally mature more slowly than those from populations north of Broome. The onset of gameto- genesis in populations of the oyster, Crassostrea virginica, from different geographic areas has been found to be tem- perature related (Loosanoff 1969). Various stocks of the clam, Mercenaria mercenaria, have separate spawning and recovery peaks and differences in the mean level of gonad development (Knaub and Eversole 1988). Seasonal varia- tions in the combined, macroscopic gonad index of both sexes of the abalone, Haliotis rubra, have been found to exist for three of the four populations from Victoria, Aus- tralia (McShane et al. 1986). Field observations on the gonadal development of oysters from the Torres Strait suggest that the main spawning period occurs just before the cyclone season, during November/December (Wada 1953). The weather at this time is changing from the dry to the wet season, the seas are markedly calm, and the water temperature is rising rapidly. The Broome region is also subjected to a similar weather pattern during these months and at this time P. maxima was found easier to spawn in the hatchery, pro- ducing a greater number of viable gametes per individual. The larvae were reared to settlement more successfully during November/December when the ambient water tem- perature ranged between 26° and 30°C (Fig. 10). Records of field and hatchery spawnings suggested that the gonadal development varied locally, with oysters from populations south of Eighty-Mile Beach, near Onslow sometimes several weeks behind those near or north of Broome (Fig. 1). As shown statistically with histological data, the breeding season varied annually. Records of the number and intensity of viable gametes per spawning indi- cated that there were generally two peaks, a primary one between November/December and a secondary peak be- tween late February/March. On the basis of successful hatchery spawnings, the months of the breeding season could be ranked in decreasing order, as follows: No- vember, December, October = March, February, January and April. As demonstrated by Lannan et al. (1980) for C. gigas, effective broodstock management should be concerned with identifying the ‘‘optimum window’’ of gametogenesis which in turn will maximize larval survival per spawning. Thus, mariculture spawning trials are recommended during November/December for P. maxima. Artifical propagation of P. maxima outside these months is neither necessary nor cost-effective, given the pearl culture industry’s concern to prevent over-production of pearls. Any future domestica- tion of this species should concentrate on selectively breeding oysters whose progeny will produce high quality pearls. ACKNOWLEDGMENTS We appreciated the efforts of Nicole Roberts, Michael Mannion, Serena Sanders and Shayne Baker for their assis- tance with the preservation and microscopic examination of gonadal material. Histological sections were prepared at the School of Veterinary Studies, Murdoch University, Western Australia. Photographs of gonadal tissue were taken and printed by Christopher Surman. Dr. Nick Caputi and Kevin Donohue assisted with the statistical analyses, and Christine Nicholson assisted with collation of data and typing of the manuscript. This study was supported by the Australian Commonwealth Fishing Industries Research and Developmental Council. REFERENCES Alagarswami, K., S. Dharmaraj, T. S. Velayudhan, A. Chellam & A. C. C. Victor. 1983. On controlled spawning of Indian pearl oyster Pinctada fucata (Gould). Symp. Ser. Mar. Biol. Assoc. India. 6:590— SeHk Bayne, B. L., D. L. Holland, M. N. Moore, D. M. Lowe & J. Widdows. 1978. Further studies on the effects of stress in the adult on the eggs of Mytilis edulis. J. Mar. Biol. Assoc. U.K. 58(4):825—-841. Dybdahl, R. & R. A. Rose. 1986. The pearl oyster fishery in Western Australia. In A. K. Haines, C. G. Williams & C. Coates (eds.) Torres Strait fisheries seminar, Port Moresby, 11-14 Feb. 1985. Aust. Gov. Publ. Service. Canberra: 122-132. Iwata, K. S. 1952. Mechanism of egg maturation in Mytilus edulis. Bio- logical Journal of Okayama University. 4:1—11. Knaub, R. S. & A. G. Eversole. 1988. Reproduction in different stocks of Mercenaria Mercineria. J. Shellfish Res. 7(3):371—376. Lannan, J. E., A. Robinson & W. P. Breese. 1980. Broodstock manage- ment of Crassostrea gigas. 11. broodstock conditioning to maximize larval survival. Aquaculture. 21:337-345. Loosanoff, V. L. 1969. Maturation of gonads of oysters, Crassostrea vir- ginica, of different geographical areas subjected to relatively low tem- peratures. Veliger. 11(3):153—163. McShane, P. E., K. H. H. Beinssen, M. G. Smith, S. O’connor & N. J. Hickman. 1986. Reproductive biology of the blacklip abalone Haliotis Ruber Leach from four Victorian populations. Department of Conser- vation Forests and Lands, Fisheries and Wildlife Service. Marine Sci- ence Laboratories. Technical Report No. 55: 13 pp. 272 ROSE ET AL. Minaur, J. 1969. Experiments on the artificial rearing of the larvae of Pinctada maxima (Jameson) (Lamellibranchia). Aust. J. Mar. Freshw. Res. 20:175—187. Ojima, Y. & K. Maeki. 1955. Some cytological account of the maturation of the gonad in the pearl oyster, Pinctada martensi Dunker. Kwansei Gakuin University Annual Studies. 3:(41). Kwansei Gakuin Univer- sity, Nishinomiya, Japan. March 1955. Pass, D. A., R. Dybdahl & M. M. Mannion. 1987. Investigations into the causes of mortality of the pearl oyster, Pinctada maxima (Jameson). Aquaculture. 65:149—169. Rose, R. A., R. Dybdahl, S. Sanders & S. Baker. 1986. Studies on artifi- cially propagating the gold or silver-lipped pearl oyster, Pinctada maxima (Jameson). In R. E. Pyne (ed.) Darwin Aquaculture Work- shop. Technical Report No. 3, Fisheries Div. Dept. Primary Indust. and Fish. Aust. Gov. Publ. Service, Canberra: 60-67. Sastry, A. N. & N. J. Blake. 1971. Regulation of gonad development in the bay scallop, Aequipecten irradians (Lamarck). Biol. Bull. Mar. Biol. Lab. Woods Hole. 140:274-83. Seed, R. 1969. The ecology of Mytilis edulis L. (Lamellibranchia) on exposed rocky shores. 1. breeding and settlement. Ocecologia (Berl.). 3:277-316. Tanaka, Y. & M. Kumeta. 1981 (In Japanese). Successful artificial breeding of the silver-lip pearl oysters, Pinctada maxima (Jameson). Bull. Natl. Res. Inst. Aquaculture. 2:21—28. Tranter, D. J. 1958a. Reproduction in Australian pearl oysters. Univ. of Queensland, Australia: MS Thesis. 143 p. Tranter, D. D. 1958b. Reproduction in Australian pearl oysters (Lamelli- branchia). 1. Pinctada albina (Lamarack): primary gonad develop- ment. Aust. J. Mar. Freshw. Res. 9:135—143. Tranter, D. J. 1958c. Reproduction in Australian pearl oysters (Lamelli- branchia). 11. Pinctada albina (Lamarack): gametogenesis. Aust. J. Mar. Freshw. Res. 9:144-158. Tranter, D. J. 1958d. Reproduction in Australian pearl oysters (Lamelli- branchia). 111. Pinctada albina (Lamarack): breeding season and sex- uality. Aust. J. Mar. Freshw. Res. 9:191—216. Tranter, D. J. 1958e. Reproduction in Australian pearl oysters (Lamelli- branchia). 1V. Pinctada magaritifera (Linnaeus). Aust. J. Mar. Freshw. Res. 9:509-525. Velez, A. & C. E. Epifanio. 1981. Effects of temperature and ration on gametogenesis and growth in the tropical mussel Perna perna (L.). Aquaculture 22:21—26. Wada, S. K. 1942. Artificial fertilization and development of the silver- lip pearl oyster (P. maxima). Science of the South Seas. 4:202—208. Wada, S. 1953a. Biology and fisheries of the silver-lip pearl oyster. Un- published report in the library CSIRO Marine Laboratories, Hobart, Tasmania: 1-86. Wada, S. K. 1953b. Biology of the silverlip pearl oyster, Pinctada maxima (Jameson). |. Artificial fertilization and development. Mar- garita, 1:3—-15. Wada, S. K. 1953c. Biology of the silver-lip pearl oyster Pinctada maxima (Jameson). 2. breeding season. Margarita. 1:17—28. > Journal of Shellfish Research, Vol. 9, No. 2, 273-278, 1990. GROWTH AND SIZE AT MATURITY OF THE PACIFIC GAPER TRESUS NUTTALLII (CONRAD 1837) IN SOUTHERN BRITISH COLUMBIA A. CAMPBELL, N. BOURNE AND W. CAROLSFELD Department of Fisheries and Oceans Biological Sciences Branch Pacific Biological Station Nanaimo, British Columbia V9R 5K6 ABSTRACT Measurements were made to determine growth rates and size at maturity of Pacific gaper (horse clams), Tresus nuttallii (Conrad) (family: Mactridae) from two areas in southern British Columbia. Growth was determined by measuring shell length at annuli. Growth rates of T. nuttallii were slower from Newcastle Island than from Lemmens Inlet. Shells became heavier than the soft body parts at a faster rate for horse clams from Newcastle Island compared with those from Lemmens Inlet. Examination of histolog- ical sections of gonads indicated that size at maturity occurred at about 68 mm shell length or at 3 y of age for T. nuttallii from Lemmens Inlet. Although samples were taken only in the May— August period, the gonad histological sections indicate that 7. nuttallit is a summer (April— August) spawner. KEY WORDS: Pacific gaper, horse clam, Tresus nuttallii, size at maturity, growth, reproduction INTRODUCTION The horse clam, Tresus nuttallii (Conrad 1837) (Bi- valvia: Mactridae), occurs commonly along the west coast of North America from California to Alaska (Latitude 28° to 58°N) (Bernard 1983). The clam is found in coastal waters of British Columbia (B.C.) in mud-sand substrates from low intertidal beach levels to subtidal depths of at least SO m. A small subtidal commercial dive fishery has recently developed for horse clams (includes two species) which was worth about Can. $300,000 for 325 t during 1988 in B.C. Although the other horse clam species, T. capax (Gould 1850), is more common intertidally (Bourne & Smith 1972b), T. nuttallii is the more abundant (>75%) subtidally of the two horse clam species found in B.C. (A. Campbell, unpubl. data). No detailed biological information has been published on T. nuttallii from B.C. (Quayle 1960, Quayle & Bourne 1972, Bourne & Smith 1972b). Most biological data on T. nuttallii is from intertidal populations in Washington (Swan & Finucane 1952, Pearce 1965, Goodwin & Shaul 1978) and California (MacGinitie 1935, Fitch 1953, Armstrong 1965, Stout 1967, 1970, Laurent 1971, Clark 1973, Clark et al. 1975, Kvitek et al. 1988). Studies on inter and sub- tidal T. capax biology extend throughout the clam’s range (Lat. 28° to 58°N) (Pearce 1965, Reid 1969, Machell & DeMartini 1971, Bourne & Smith 1972a, b, Wendell et al. 1976, Goodwin & Shaul 1978, Breed-Willeke & Hancock 1980, Robinson & Breese 1982). A third species T. paja- roana (Conrad 1857) has a limited subtidal distribution from California to Washington (Dinnell & DeMartini 1974). This paper presents information on the growth and sexual maturity of 7. nuttallii from two subtidal areas in B.C. which will be needed for fishery management of the species. 273 MATERIALS AND METHODS Samples from as wide a size range as possible of 7. nut- tallii were obtained from Lemmens Inlet, near Tofino (Lat. 49°12.2' Long. 125°52.3') during 25 May and 10 August, 1989, and Newcastle Island, near Nanaimo (Lat. 49°12.2' Long. 123°56.5’) during 11 July, 1989, at depths between 4—8 m for both areas. The clams were transported to the laboratory in coolers (2°C) and kept in running sea water (ambient temperature) until processed within 48 h of cap- ture. For each clam, the shell length (SL) was measured as the straight line distance between anterior and posterior margin of the shell to the nearest mm with vernier calipers, wet weights of drained total body and shell, shell only, whole soft body and siphon (neck) only (cut at base of si- phon) were recorded to the nearest 0.01 g. Growth of 7. nuttallii was determined by measuring shell length at each annulus after Weymouth et al. (1925) and discussed by Bourne and Smith (1972b). Horse clams from both areas had pronounced annuli; clams with indis- tinct annuli (<1%) or with broken shells were discarded. The reproductive condition of 7. nuttallii was deter- mined by removing the central portion of the gonad and preserving the tissue in Davidson’s Solution. Histological slides, stained with haematoxylin-eosin were prepared from sections of the gonad. Five stages (1. inactive, 2. active, 3. ripe, 4. partially spent, and 5. spent) according to Machell and DeMartini (1971) and Laurent (1971) were used to classify the histological sections of the gonads of each horse clam. Sexual maturity was determined from the his- tological sections by categorizing them as either (1) imma- ture (no differentiation in gonadal tissue; loose vesicular connective tissue in gonad), or (2) mature (connective tissue well developed, primary germ cells evident on fol- licle walls on eggs or sperm development evident). Allometric relationships between body, shell and neck 274 weights (Y) and shell length (X) were estimated using the exponential equation of the form log, Y = log, A + B log, X, where A and B are constants calculated using the least squares method. The relationships between the ratios, shell weight/body weight and neck weight/body weight (Y) and shell length or age (X) were estimated using the linear equation Y = A + BX. Comparison between the two sample areas for each relationship was accomplished by testing comparable size ranges of 100-202 mm SL and ages 7-16 yr for homogeneity between slopes and subse- quently comparing intercepts of lines by adjusting the Y variables and testing for differences by analysis of covari- ance (ANCOVA) using shell lengths or age as covariates (Snedecor & Cochran 1967) using SYSTAT computer pro- grams (Wilkinson 1989). Average von Bertalanffy growth curves were fitted to all data points of size at age using the equation: fh. = Ibe (dl = e— kit to)) where t = age in years, /, = shell length at t, L, = theo- retical maximum size, k = constant, determining rate of increase or decrease in length increments, and t, = hypo- thetical age at which the organism would have been at zero length. The parameters L,., k and t, were estimated using a nonlinear least squares method (Wilkinson 1989). Growth rates between males and females between May and August 1989, from Lemmens Inlet were similar when compared graphically and therefore the data were combined for each area. Only growth rates from the annuli data are presented since growth rates from total shell length and annuli pro- duced similar curves for horse clams from each area. Lee (1912) suggested that older individuals may exhibit slower growth rates than smaller individuals due to differential mortality rates but this was not the case in this study. The proportion of mature clams (¢) at shell length (1) was estimated using the equation: 1 P(/) = jee e(a— bl) where a and b are constants determined using maximum likelihood methods (e.g., Welch & Foucher 1988). Male and female data were combined since the curves for each sex were similar. There were no immature horse clams ob- tained from Newcastle Island so size at maturity ogives were not attempted for this area. RESULTS Growth A) Shell Length-Age The oldest T. nuttallii collected was 16 y from both study areas. Growth rates of T. nuttallii from Newcastle were slower than those from Lemmens Inlet (Fig. 1). CAMPBELL ET AL. 2005 = 9-1 eae ry slh vee al a 3 come: it == ISOs Die il e a ee E a ana ac Ly = Yon Oo jal z 1004 Cy sr wy / fe} = a ay wx Ww for ae fi vp) 50-7 f! = ie h = O- T T al y (@) 5 10 15 20 AGE (years) Figure 1. Growth curves for 7. nuttallii collected from Lemmens Inlet (solid line and closed dots) and Newcastle Island (broken line and open circles). Curves from von Bertalanfly growth parameters, mean (dots) and 95% confidence intervals (vertical lines) from raw data. Equations in Table 1. Within 5 y the horse clams had reached a mean of 106 mm SL from Lemmens Inlet and 97 mm SL from Newcastle, at 10 y 161 mm and 145 mm SL, and at 16 y 187 mm and 169 mm SL, respectively. The von Bertalanffy growth param- eters were similar for T. nuttallii from both areas except for L,, which was estimated to be higher for those from Lemmens Inlet than those from Newcastle Island (Table 1). The largest horse clam was 195 mm SL collected at Lemmens Inlet and 202 mm SL from Newcastle Island in this study. The largest T. nuttallii specimen from an unsub- stantiated report was as long as 250 mm (Nicol 1964). B) Length-Weight The size range of horse clams collected from Lemmens Inlet was 36-195 mm SL (N = 146) and from Newcastle Island was 107—202 mm SL (N = 52). All the length- weight relationships were highly positively correlated indi- cating that shell and body weights increased with in- creasing SL (Table 2, Figs. 2, 3). There were no differ- ences (ANCOVA, p > 0.05) in all pair comparisons between the two sample areas in slopes of the power re- gressions of weight-length and linear regressions of ratios- length for comparable size ranges of 100—202 mm SL which allowed subsequent comparison of intercepts (Tables 2, 3). For the power regressions, although there were no differences in intercepts of body and neck (siphon) weights, there were significant differences (ANCOVA, p < 0.01) for total weights and shell weights at the equiva- HorRSE CLAM GROWTH AND MATURITY TABLE 1. Von Bertalanfly growth parameters for 7. nuttallii from Lemmens Inlet and Newcastle Island. Values in brackets are approximate 95% confidence intervals. Area L. K ty Lemmens Inlet 202 0.167 0.50 (+3) (+0.006) (+0.05) Newcastle Island 183 0.168 0.51 (+5) (+0.012) (+0.10) lent SL range of horse clams between the two areas. For the ratio-length equations, there were no differences in inter- cepts for the neck/body and neck/total ratios (ANCOVA, p > 0.05), but significant differences (ANCOVA, p < 0.01) for the shell/body ratios between the two areas. Conse- quently the rate of increase in weight was greater for the shells than for the soft body parts with increases of SL of horse clams from both areas (Tables 2, 3, Fig. 3). The shell weights were heavier for horse clams of the equivalent SL from Newcastle than from Lemmens Inlet (Fig. 3, Tables 2, 3). There were no significant differences in weight (ANCOVA, p > 0.05) of soft body parts at equivalent SL of horse clams collected from both areas. Neck weights in- creased at a slower rate than shell or the whole soft body weights (Fig. 3, Table 2) which 1s indicated also by a gen- TABLE 2. Regression coefficients for various morphological relationships of T. nuttallii from (1) Lemmens Inlet and (2) Newcastle Island for equation Log, Y = log, A + Blog, X, where X = shell length (SL in mm) or age (in years) and Y variables are weights in g. All R? values are significant at p < 0.01. Body includes all soft body parts. All horse clam data used Lemmens Inlet 36-195 mm SL (N = 146) and Newcastle 107-202 mm SL (N = 52) for equations. Regression Variables Coefficients SE of ¥ x Area A B Estimate R? Total SL ] — 10.029 3.219 0.118 0.993 2 —9.310 3.087 0.116 0.914 Body SL 1 — 93325 2.969 0.130 0.991 2 — 8.442 2.786 0.139 0.857 Neck SL 1 —9.751 2.883 0.165 0.984 2 —7.608 2.456 0.185 0.725 Shell SL 1 — 13.059 3.642 0.152 0.991 2 — 12.159 3.499 0.171 0.863 Total Age 1 1.550 2.039 0.392 0.932 2 2.592 1.489 0.739 0.739 Body Age l 1.370 1.874 0.378 0.922 2 2.539 1.247 0.234 0.596 Neck Age 1 0.620 1.825 0.372 0.920 2 2.200 1.048 0.260 0.450 Shell Age 1 0.046 2.306 0.412 0.937 2 1.044 1.807 0.207 0.799 275 1500 4 1000 5004 oO be ate © W ) = (e) 50 100 150 200 =I aq 1500 bk eo) oo 1000 500 O 50 100 SHELL LENGTH (mm) Figure 2. Total weight and shell length relationship for T. nuttallii collected from (A) Lemmens Inlet and (B) Newcastle Island. Equa- tions in Table 2. 150 200 eral decline in ratios of neck/body weights with increase in SL (Table 3). C) Weight-age The total age range of horse clams collected from Lemmens Inlet was 2—16 y and Newcastle Island was 7-16 y. Weight increases by age were all greater for horse clams from Lemmens Inlet than Newcastle Island (Ta- ble 4). Although there were no significant differences 276 600 5 BODY D ad O 50 100 150 200 KE aE oO uj 6005 = B 4 4004 20074 Se 100 150 SHELL LENGTH (mm) Figure 3. Body, shell and neck weight and shell length relationships for T. nuttallii collected from (A) Lemmens Inlet and (B) Newcastle Island. Equations in Table 2. T 200 (ANCOVA, p > 0.05) between areas for all slopes of power and linear regressions (Tables 2, 3) for comparable ages (7—16 y), intercepts were significantly different (AN- COVA, p < 0.01) for all weight-age relations (Table 2) and shell/body ratios, except for neck/body ratios and neck/total (%)-age relations (Table 3). Shell/body ratios were higher for the equivalent age for horse clams from Newcastle than from Lemmens Inlet. In contrast, neck/ body ratios and neck/total (%) were similar for horse clams from both areas (Table 3). Size at Maturity Size at 50% maturity was 68 mm SL for horse clams from Lemmens Inlet (Fig. 4). The largest immature clam CAMPBELL ET AL. TABLE 3. Relationships between ratios of shell weight/body weight, neck weight/body weight (Y), and neck weight/total wt (%), with shell length (SL) or Age (years) (X) for 7. nuttallii from (1) Lemmens Inlet and (2) Newcastle Island using equation Y = A + BX. Body includes all soft body parts. R? are all significant at p < 0.01 except where indicated with * P < 0.05. All data used as per Table 2. Regression Variables Coefficients SE of VG x Area A B Estimate R? Shell/Body = SL 1 0.133 0.004 0.100 0.762 2 0.285 0.004 0.182 0.139* Neck/Body SL 1 0.482 —0.0004 0.057 0.082 2 0.582 —0.0009 0.055 0.085* Neck/Total SL 1 39.109 —0.093 4.227 0.518 (%) 2 38.799 —0.099 3.984 0.174 Shell/Body Age 1 0.298 0.041 0.093 0.795 2 0.421 0.041 0.167 0.271 Neck/Body Age 1 0.460 —0.003 0.057 0.056 2 0.526 —0.007 0.054 0.099* Neck/Total Age 1 34.417 —0.925 4.527 0.451 (%) 2 34.206 —0.918 3.730 0.276 was 86 mm SL and smallest mature clam was 51 mm SL. All horse clams collected from Newcastle Island were ma- ture. Sex ratio was 1:1 for all horse clams that had mature gonads in which sex was discernable. There were insufficient data to determine the exact spawning period(s) because seasonal monthly samples were not collected. However, from the reproductive phase exam- ined of horse clams >100 mm SL, spawning started in Lemmens Inlet just prior to the 25 May 1989 sample (42% in stage 2-active, 2% in stage 3-ripe, and 56% in stage 4- > eee coos © cn © fg = : 2 ie 2 +e 11-846 — 0.165 2 = Zz S | a o N S os] & omar a Ne 3 40 60 80 100 120 140 SHELL LENGTH (mm) Figure 4. Size at maturity ogive for T. nuttallii (sexes combined) col- lected from Lemmens Inlet. N = number of individuals. Equation for the predicted curve is shown in graph. HORSE CLAM GROWTH AND MATURITY 277 partially spent, N = 41, and was nearly complete by 10 August 1989 (47% stage 4-partially spent and 53% stage 5-spent, N = 57). For Newcastle Island horse clam spawning was nearly complete by 11 July (3% stage 3, 5% stage 4, and 92% stage 5, N = 40). DISCUSSION Growth of 7. nuttallii from Newcastle Island was slower than those from Lemmens Inlet. Shells became heavier than the soft body parts at a faster rate for horse clams from Newcastle Island compared with Lemmens Inlet. Growth for juvenile 7. nuttallii from Elkhorn Slough, California, was about 50 mm SL in their first year (Laurent 1971, Clark 1973) which was faster than recorded from either of the two B.C. study areas. Slower growth rates of other species of bivalves have been reported with northward dis- tribution; razor clams (Bourne & Quayle 1970, Weymouth & McMillin 1930), butter clams (Quayle & Bourne 1972) and littleneck clams (Quayle & Bourne 1972). The reasons for the differences in 7. nuttallii growth rates is not known, but could be attributed to a variety of environmental factors associated with different habitats, e.g. substrate type, food availability, temperature. Both B.C. study areas had sim- ilar mud sand substrate and temperature regimes, however. Growth rates directly associated with food availability and length of feeding period in various clam species have been documented (Smith 1928, Coe & Fitch 1950, Stickney 1964). Our limited seasonal data indicate that spawning 7. nut- tallii in B.C. probably occurs during April—August. This adds support to the suggestion by Quayle (1960), Quayle and Bourne (1972) and Bourne and Harbo (1987) that 7. nuttallii spawns during summer in B.C. In contrast, 7. nut- tallii may spawn continuously throughout the year in Elk- horn Slough, California, with bimodal spawning peaks during April-June and November—February (Laurent 1971, Clark 1973, Clark et al. 1975). Tresus capax spawning occurs during one annual period from mid Feb- tuary to May in B.C. (Bourne & Smith 1972b). Bimodal spawning for T. capax for Humboldt Bay, California has been reported by Wendell et al. (1976). Breed-Willeke and Hancock (1980) suggested that T. capax populations from southern latitudes have slightly earlier spawning periods than horse clams from more northern latitudes along western North America. Size at maturity of T. nuttallii from Lemmens Inlet was estimated at approximately 68 mm SL or at about 3 y of age. Clarke (1973) suggested that 7. nuttallii females ma- ture at about 70 mm SL in Elkhorn Slough. Bourne and Smith (1972b) found that 7. capax at Seal Island, B.C. became sexually mature at about 70 mm SL. As juvenile horse clams mature they loose their ability to dig into the substrate. Pohlo (1964) found there was a change in burrowing ability and shell morphology with growth of 7. nuttallii juveniles, with the capability to re- burying lost at about 60 mm SL (probably near maturity). Armstrong (1965) found that 7. nuttallii juvenile survival was sensitive to reburying positions. No pinnotherid pea crabs (e.g., Pinnixa faba (Dana 1851)) were found in 7. nuttallii although they were found in the T. capax also collected from Lemmens Inlet in this study. Although pea crabs are found to live commensally in the mantle cavity of 7. nuttallii in southern California where 7. capax does not occur (MacGinitie 1935, Laurent 1971), in areas further north pea crabs are found only in 7. capax, but not in T. nuttallii (Pearce 1965, Stout 1967, N. Bourne & A. Campbell, personal observation in B.C.). The genus Tresus also serves as host to a variety of para- sitic invertebrates (MacGinitie 1935, Ricketts et al. 1968) and commensals (Stout 1970). Besides man, the most im- portant natural predators of horse clams are probably sea stars (e.g., Pisaster brevispinus (Stimpson 1857)), moon snails (Polinices lewisii (Gould 1847)), crabs (e.g., Cancer magister Dana 1852) (Bernard 1967, Wendell et al. 1976, Sloan & Robinson 1983), elasmobranchs (e.g., skates and rays) (Stout 1967) and sharks (e.g., Triakis semifasciatus Girard 1854) that feed off clam siphons (Laurent 1971) and sea otters (Enhydra lutris (Merriam 1923)) that dig for whole horse clams (Kvitek et al. 1988). Results of the life history information in this study and the recent developing subtidal horse clam fishery suggests that conservative harvest levels should be maintained for B.C. horse clam stocks to monitor changes in biological parameters that may result from fishing pressures. Long term studies are required to determine the effect of the fishery on horse clam densities, mortality and recruitment rates. ACKNOWLEDGMENTS We thank J. Bagshaw for preparing the histological slides of horse clams, J. Baxter for technical assistance, R. Harbo for collecting horse clams, D. Pierce for typing, D. Welch for calculating the size at maturity ogive and W. Shaw for reviewing an earlier draft of this paper. REFERENCES Ammstrong, L. R. 1965. Burrowing limitations in Pelecypoda. Veliger 7:195—200. Bernard, F. R. 1967. Studies on the biology of the naticid clam drill Po- linices lewisti (Gould) (Gastropoda, Prosobranchia). Fish. Res. Board Can. Tech. Rep. 42:41 pp. Bernard, F. R. 1983. Catalogue of the living Bivalvia of the eastern Pa- cific Ocean: Bering Strait to Cape Horn. Can. Special Publ. Fish. Aquat. Sci. 61:vii + 102 pp. Breed-Willeke, G. M. & D. R. Hancock. 1980. Growth and reproduction of subtidal and intertidal populations of the gaper clam Tresus capax (Gould) from Yaquina Bay, Oregon. Proc. Natl. Shellfish Assoc. 70:1—13. 278 Bourne, N. & R. M. Harbo. 1987. Horse clams, in Harbo, R. M., and G. S. Jamieson (ed.). Status of invertebrate fisheries off the pacific coast of Canada (1985/86). Can. Tech. Rep. Fish. Aquat. Sci. No. 1576:89—94. Bourne, N. & D. B. Quayle. 1970. Breeding and growth of razor clams in British Columbia. Fish. Res. Board Can. Tech. Rep. 232:1 + 42 pp. Bourne, N. & D. W. Smith. 1972a. The effect of temperature on the larval development of the horse clam, Tresus capax (Gould). Proc. Nat. Shellfish Assoc. 62:35—37. Bourne, N. & D. W. Smith. 1972b. Breeding and growth of the horse clam, Tresus capax (Gould), in southern British Columbia. Proc. Nat. Shellfish Assoc. 62:38—46. Clark, P. C. 1973. Aspects of the life history of Tresus nuttallii in Elkhorn Slough. M.A. thesis Calif. State Univ., Hayward. iti + 46 pp. Clark, P. C., J. Nybakken & L. Laurent. 1975. Aspects of the life history of Tresus nuttallii in Elkhorn Slough. Calif. Fish. Game. 61:215—227. Coe, W. R. & J. E. Fitch. 1950. Populations studies, local growth rates, and reproduction of the Pismo clam (Tivela stultorum). J. Mar. Res. 9:188—210. Dinnel, P. A. & J. D. DeMartini. 1974. A supposedly extinct bivalve species found living off California. Veliger 17:44—47. Fitch, J. E. 1953. Common marine bivalves of California. State of Calif. Rept. Fish. Game Mar. Fish. Branch. Fish. Bull. No. 90. 102 pp. Goodwin, L. & W. Shaul. 1978. Puget Sound subtidal hardshell clam survey data. State of Washington Rept. Fish. Progress Rep. 44. 92 pp. Kvitek, R. G., A. K. Fukayama, B. C. Anderson & B. K. Grimm. 1988. Sea otter foraging on deep-burrowing bivalves in a California coastal lagoon. Mar. Biol. 98:157—167. Laurent, L. L. 1971. The spawning cycle and juvenile growth rate of the gaper clam, Tresus nuttallii, of Elkhorn Slough, California. M.A. thesis. San Francisco State College, ix + 56 pp. Lee, R. M. 1912. An investigation into the methods of growth determina- tion in fishes. Cons. Explor. Mer, Publ. de Circonstance 63:35 pp. MacGinitie, G. E. 1935. Ecological aspects of a California marine es- tuary. Amer. Midland Naturalist 16:629—765. Machell, J. R. & J. D. DeMartini. 1971. An annual reproductive cycle of the gaper clam, Tresus capax (Gould), in south Humboldt Bay, Cali- fornia. Calif. Fish. Game 57:274-282. Nicol, D. 1964. An essay on size of marine pelecypods. J. Paleontol. 38:968—974. Pearce, J. B. 1965. On the distribution of Tresus nuttallii and Tresus capax (Pelecypoda: Mactridae) in the waters of Puget Sound and the San Juan Archipelago. Veliger 7:166—170. Pohlo, R. H. 1964. Ontogenetic changes of form and mode of life in Tresus nuttallii (Bivalvia: Mactridae). Malacologia 1:321—330. CAMPBELL ET AL. Quayle, D. B. 1960. The intertidal bivalves of British Columbia. B.C. Proc. Museum Handbook 17, 104 pp. Quayle, D. B. & N. Bourne. 1972. The clam fisheries of British Co- lumbia. Bull. of Fish. Res. Board Can. 179, vii + 70 pp. Reid, R. G. B. 1969. Seasonal observations on diet, and stored glycogen and lipids in the horse clam, Tresus capax (Gould, 1850). Veliger 11:378-381. Ricketts, E., J. Calvin & J. Hedgpeth. 1968. Between Pacific tides. Fourth edition. Stanford Amer. Press, Stanford, CA, 614 pp. Robinson, A. M. & W. P. Breese. 1982. The spawning season of four species of clams in Oregon. J. Shellfish Res. 2:55—57. Sloan, N. A. & S. M. C. Robinson. 1983. Winter feeding by asteroids on a subtidal sandbed in British Columbia. Ophelia 22:125—140. Smith, G. M. 1928. Food material as a factor in growth rate of some Pacific clams. Trans. Roy. Soc. Can. 22:287—291. Snedecor, G. W. & W.G. Cochran. 1967. Statistical methods. /owa State Univ. Prep, Ames, I. A. 593 pp. Stickney, A. P. 1964. Feeding and growth of juvenile soft shell clams, Mya arenaria. Fish. Bull. Fish. Wild. Serv. U.S. 63:635—642. Stout, W. E. 1967. A study of the autecology of the horseneck clams, Tresus nuttallii and Tresus capax, in south Humboldt Bay, California. M.Sc. thesis, Humboldt State College, Arcata, CA, 42 pp. Stout, W. E. 1970. Some associates of Tresus nuttallii (Conrad, 1837) (Pelecypoda: Mactridae). Veliger 13:67—70. Swan, E. F. & J. H. Finucane. 1952. Observations on the genus Schi- zothaerus. Nautilus, 66:19—26. Welch, D. W. & R. P. Foucher. 1988. A maximum likelihood method- ology for estimating length-at-maturity with application to Pacific cod (Gadus macrocephalus) population dynamics. Can. J. Fish. Aquat. Sci. 45:333-343. Wendell, F., J. D. DeMartini, P. Dinnel & J. Siecke. 1976. The ecology of the gaper or horse clam, Tresus capax (Gould 1850) (Bivalvia: Mactridae), in Humboldt Bay, California. Calif. Fish and Game, 62:41—64. Weymouth, F. W. & H. C. McMillin. 1930. Relative growths and mor- tality of the Pacific razor clam (Siliqua patula, Dixon) and their bearing on the commercial fishery. U.S. Bureau Fish. Bull. 46:543— 567. Weymouth, F. W., H. C. McMillin & H. B. Holmes. 1925. Growth and age at maturity of the Pacific razor clam, Siliqua patula (Dixon). Bull. U.S. Bur. Fish. 41:201—236. Wilkinson, L. 1989. SYSTAT: The system for statistics. Evanston, IL: Systat, Inc., xviii + 638 pp. Journal of Shellfish Research, Vol. 9, No. 2, 279-281, 1990. KARYOTYPE OF THE DWARF SURFCLAM MULINIA LATERALIS (SAY, 1822) (MACTRIDAE, BIVALVIA) KATSUHIKO T. WADA,! JOHN SCARPA,? AND STANDISH K. ALLEN, JR.? ‘National Research Institute of Aquaculture Nansei, Mie 516-01 Japan *Rutgers Shellfish Research Laboratory P.O. Box 687 Port Norris, N.J. 08349 ABSTRACT Chromosomal analysis of dwarf surfclam, Mulinia lateralis, larvae revealed that the diploid number is 38, different from that reported in the literature (2n = 36). Karyology of eggs (n = 19) and induced triploid larvae (3n = 57) corroborates the diploid count. All chromosomes in the complement were classified as acrocentric. KEY WORDS: INTRODUCTION The dwarf surfclam, Mulinia lateralis (Say 1822), is ubiquitously dispersed on the east coast of the United States (Calabrese & Rhodes 1974). Adults generally live for only two years and rarely grow to 30 mm. A number of charac- teristics make this species an ideal model for genetic studies in bivalves (Calabrese 1969a, b, Kidder 1972). The species is euryhaline and eurythermal. Fecundity is high and generation time may be as short as six to eight weeks. The sex of mature individuals is easily determined by ex- amination of gonad color through the thin shell: females are pink to orange, males are milky white. The larval cycle is brief (8—10 days) and juveniles are easy to culture. Karyotype analysis has not been described for this species. Menzel (1968) examined unfertilized eggs and zy- gotes and reported a diploid chromosome number of 36. As part of ongoing cytogenetic investigations of M. lateralis in the laboratory of SKA, we observed chromosomes in eggs and trochophore larvae cells. Our findings are in disagree- ment with those in the literature. MATERIALS AND METHODS Specimens of M. lateralis were obtained from a natural set at Massey’s Landing, Delaware, U.S.A. and from Vir- ginia Institute of Marine Sciences, Wachapreague, Vir- ginia, U.S.A. For examination of meiotic chromosomes, 25 females were induced to spawn (13 individuals and 3 group spawns) by thermal shock in glass bowls. Eggs were fixed in 1% (v/v) phosphate buffered formalin and stained with the fluorochrome 4’, 6-diamidino-2-phenylindole (DAPI, Sigma Chem.) (Scarpa 1985). Chromosomes were examined with the aid of a Nikon Optiphot epifluorescent microscope using a UV-1 (Nikon) filter block. For mitotic chromosomes, trochophore larvae were ob- tained from separate matings of five females with several males. Ten to sixteen hour old larvae were treated with 0.02% colchicine in sea water for 20 minutes, transferred karyotype, bivalve, Mulinia, aneuploid, triploid to 0.075 M KCI solution for 20—30 minutes and then fixed in ice cold methanol-acetic acid (3:1). Larvae were washed three to five times with fixative. Chromosome preparations were made on glass slides by placing fixed larvae in a drop of 50% acetic acid and chopping with a scalpel to create a cell suspension (Komaru & Wada 1985). After drying, cells were stained with 0.2% Giemsa or 0.03% Leishman’s solution made in phosphate buffer. Photographs of well- spread metaphase chromosomes were taken with a Nikon model UFX-IIA photomicroscopic system. Triploid larvae were produced from fertilized eggs incu- bated at 21 C treated with cytochalasin B (0.5 mg/L) at 22 to 37 minutes post-fertilization (Scarpa 1985). Metaphase spreads were obtained from trochophores and observed as described above. RESULTS Meiotic chromosome counts were obtained from two populations: Delaware and Virginia. For Delaware, the haploid chromosome number was determined in 580 eggs from 22 females. Chromosome number ranged from 16 to 21 with the majority (83.1%) of eggs containing 19 (Fig. 1). Chromosome counts other than 19 were distributed as follows: 16—0.2%, 17—0.5%, 18—13.4%, 20—1.9%, and 21—0.9%. For the Virginia population, haploid chro- mosome number was determined in 225 eggs from three individual females. Chromosome number ranged from 16 to 20 with the majority (88.0%) of eggs containing 19. Chromosome counts other than 19 were found in this popu- lation, also: 16—0.9%, 17—0.4%, 18—8.0%, and 20—2.7%. Populations of eggs from different females varied widely in the proportion of eggs that contained 19 chromosomes (30—100%, Table 1). All but one of the fe- males had aneuploid eggs. Six of the ten Delaware females and two of the three Virginia females produced eggs with greater than +1 chromosome. The modal chromosome number observed in 70 mitotic 279 280 WADA ET AL. Figure 1. Metaphase I chromosomes of Mulinia lateralis egg (n = 19). Scale bar: 5 um. metaphase spreads from larvae was 38 (Fig. 2). Thirty-six and 37 chromosomes were counted in two and four meta- phase spreads, respectively. Only one mitotic figure had 40 chromosomes. Actual and relative lengths ({chromosome length/total diploid length] Chromosome Number 1 2.79 0.55 3.70 0.21 2 2.62 0.51 3.47 0.14 Female 16 17 18 19 20 21 n 3 ae Oy 328 O11 1DE? 3 31 60 6 35 4 2.38 0.46 3.15 0.12 2DE 5 25 65 3) 20 5 2.30 0.46 3.03 0.11 3DE 5 30 60 5 20 6 222 0.43 2.93 0.10 4DE 3 21 76 29 7 2.14 0.41 2.85 0.12 SDE 70 30 10 8 2.10 0.41 2.79 0.12 6DE 100 20 9 2.05 0.38 pip) 0.10 7DE 12 86 2 50 10 1.99 0.37 2.64 0.08 8DE 14 84 2 50 11 1.95 0.37 2.58 0.07 9DE 7 92 1 95 12 1.91 0.36 2.53 0.06 10DE 10 89 I 71 13 1.83 0.33 2.43 0.08 11—15DE 5 89 4 2 80 14 1.78 0.32 2.36 0.12 16—19DE 12 85 2 80 15 1.70 0.31 2.26 0.16 20—22DE 10 85 5 20 16 1.59 0.35 2.10 0.15 1VA> 3 7 90 30 17 1.45 0.32 1.91 0.13 2VA 11 85 4 110 18 1.33 0.32 E75) 0.15 3VA 1 l 5 91 2 85 19 1.16 0.28 153 0.15 * DE = Delaware population. @ Mean difference between pairs = 0.114%. > VA = Virginia population. > Mean standard deviation between pairs = 0.120%. DWARF SURF CLAM KARYOTYPE 281 (1d cM elk 1G 4K CCC (le als 11¢ cXe aif man ald 060 Pea 11 12 13 14 15 2@a 20a ala ave 16 alia 18 19 Figure 3. The triploid karyotype of Mulinia lateralis (3n = 57). Scale bar: 2 um. DISCUSSION Menzel (1968) reported that M. lateralis eggs and larvae had a haploid chromosome number of 18 and diploid chro- mosome number of 36, respectively. In the present study, the modal chromosome number for eggs was n = 19 and for trochophore larvae was 2n = 38. Menzel’s samples were prepared from a Chesapeake Bay population, most likely the York or James River (Mike Castagna, VIMS, pers. comm., 1990). We have examined the meiotic chro- mosomes of unfertilized and fertilized eggs obtained from broodstock originating from another Virginia population. The modal haploid number was 19. Further studies would be needed to confirm if there is variation in karyotype among populations or localities. Diagnostic uncertainties and preparative artifacts (e.g., misdiagnosis of overlapping chromosomes and chromo- some breakage, respectively) (Vaas & Pesch 1984) may have led to incorrect counts, but probably of only + 1 chro- mosome. However, chromosome numbers other than 18—20 probably occur naturally from a disruption of the germ-cell proliferation process. For all eggs examined, 1.4% were missing greater than + 1 chromosome which is about one-half that reported for the American oyster Cras- sostrea virginica (Stiles et al. 1983). Because of the low occurrence of aneuploid eggs, chromosome counts of 100 or more eggs per individual female would be necessary for accurate quantification. According to Nakamura (1985), 2n = 38 is the most frequent chromosome number in the class Bivalvia (40% of the reported 125 species). M. lateralis can be added to this list. It appears that the all acrocentric karyotype is unique. No bivalve mollusc species karyotype has been reported to contain only acrocentric chromosomes (Nakamura 1985). ACKNOWLEDGMENTS Travel and tenure in the United States for KTW was supported by the Science and Technology Agency of Japan. NJAES publication No. D-32100-1-90 was sup- ported by NJ state and NJ Marine Sciences Consortium funding. Contribution #90-18 of the Institute of Marine and Coastal Sciences, Rutgers University. REFERENCES CITED Calabrese, A. 1969a. Reproductive cycle of the coot clam, Mulinia la- teralis (Say), in Long Island Sound. Veliger 12:265—269. Calabrese, A. 1969b. Mulinia lateralis: molluscan fruit fly? Proc. Natl. Shellfish. Assoc. 59:65—66. Calabrese, A. & E. W. Rhodes. 1974. Culture of Mulinia lateralis and Crepidula fornicata embryos and larvae for studies of pollution ef- fects. Thalassia Jugoslavia 10:89-102. Kidder, G. M. 1972. Gene transcription in mosaic embryos. I. The pat- tern of RNA synthesis in early development of the coot clam Mulinia lateralis. J. Exp. Zool. 180:55—74. Komaru, A. & K. T. Wada. 1985. Karyotype of the Japanese pearl oyster, Pinctada fucata martensii, observed in the trochophore larvae. Bull. Natl. Res. Inst. Aquaculture (Japan) 7:105—107. Levan, A., K. Fredga & A. A. Sandberg. 1964. Nomenclature for centro- meric position on chromosomes. Hereditas 52:201—220. Menzel, R. W. 1968. Chromosome number in nine families of marine pelecypod mollusks. Nautilus 82:45—58. Nakamura, H. 1985. A review of molluscan cytogenetic information based on the CISMOCH—Computerized Index System for Molluscan Chromosomes. Bivalvia, Polyplacophora and Cephalopoda. Venus (Jap. J. Malacology) 44:193—225. Scarpa, J. 1985. Experimental production of gynogenetic and partheno- genetic Mulinia lateralis (Say). Thesis for Master of Science, Univer- sity of Delaware, Newark, Delaware, U.S.A. 79 pp. Stiles, S., J. Choromanski & A. Longwell. 1983. Cytological appraisal of prospects for successful gynogenesis, parthenogenesis and andro- genesis in the oyster. Intl. Council Explor. Sea, Mariculture Comm. Paper F:10 1-14. Vaas, P. & G. G. Pesch. 1984. A karyological study of the calenoid co- pepod Eurytemora affinis. J. Crust. Biol. 4:248—251. Journal of Shellfish Research, Vol. 9, No. 2, 283-289, 1990. A RE-EXAMINATION OF THE INCIDENTAL FISHING MORTALITY OF THE TRADITIONAL CLAM HACK ON THE SOFT-SHELL CLAM, MYA ARENARIA LINNAEUS, 1758 S. M. C. ROBINSON! AND T. W. ROWELL? 'Department of Fisheries and Oceans Biological Sciences Branch Aquaculture and Invertebrate Fisheries Division St. Andrews Biological Station St. Andrews, New Brunswick, Canada, EOG 2X0 2Department of Fisheries and Oceans Biological Sciences Branch, Habitat Ecology Division Bedford Institute of Oceanography P.O. Box 1006 Dartmouth, Nova Scotia, Canada, B2Y 4A2 ABSTRACT A simple model was constructed, using values of growth taken from the literature and making some assumptions about natural mortality, to examine the theoretical effects of incidental fishing mortality! on the yield of a cohort of the soft-shell clam, Mya arenaria Linné. Results indicate that high rates of incidental fishing mortality could significantly affect the yield coming from a clam flat and have important implications for management. Based on these results, a study was conducted in the Scotia-Fundy Region of the Canadian Maritimes to examine the effect of the traditional clam hack (fork) on mortality rates in the soft-shell clam. At each experimental site, clams of various size ranges were stained with Alizarin Red dye and uniformly planted in four 1 m? plots. Two of the four plots were dug in a commercial fashion with a clam hack and two left intact as controls. All four plots were harvested two weeks later. This experimental protocol was followed at six different sites, at three different times (seasons) of the year, and repeated, in one of these seasons, for four different size ranges of clams. Results indicated that hack induced mortality rates are substantially lower than previously reported, ranging from 2—48% with an overall mean of 16.8% + 13.7 (S.D.) for all seasons and sizes combined. Both sediment type and time of year influenced mortality rates. KEY WORDS: mortality, fishing, hack, Mya arenaria INTRODUCTION The fishery for the soft-shell clam, Mya arenaria Linné, in the Canadian Maritimes is an old and traditional one. Harvesting is still carried out by hand, with individual diggers turning over chunks of sediment on clam flats, using modified garden or manure forks called ‘‘clam hacks”’ (Fig. 1). The chunks are completely inverted and the larger exposed clams are then removed by hand and put in buckets, clam hods, or similar containers. Observations on digging rates of individual clam fishermen indicate, on average, each is capable of turning over approximately 80 m? + 34.2 (S.D.) per tide (Robinson & Rowell, unpub- lished data). The disadvantages of using clam hacks to harvest soft- shell clams have been previously described. Needler and Ingalls (1944) and Medcof and MacPhail (1964) reported that approximately 50% of the undersized clams that re- mained after harvesting died as a result of the harvesting process either by breakage or smothering. Based on this, Medcof and MacPhail (1964) concluded that frequent dig- ging on clam flats was probably responsible for the decline in clam landings in the 1950’s. Glude (1954) demonstrated there was an increased size-specific mortality on soft-shell 'The mortality caused by direct physical damage or by smothering to those clams not harvested from the area of clam digging. 283 clams that were buried in the sediment and also that it de- pended on burial orientation. Emerson et al. (1990) con- ducted a series of laboratory experiments examining the ef- fects of burial and exposure on several size-classes of clams in different sediment types. They found sediment type sig- nificantly affected the rate and success of both reburial and burrowing to regain siphonal contact with the surface. Implications of incidental digging mortality (hack mor- tality) to proper management of clam harvesting are, there- fore, potentially great. At present, the clam hack is the pri- mary harvesting tool used in the Scotia-Fundy Region of the Maritime Provinces and more automated forms of har- vesting, such as hydraulic dredges or rakes (Medcof & MacPhail 1962), are regulated and are not currently in use, mainly due to resistance by the industry. In addition, there is local interest in having size limits raised in order to in- crease the yield from clam flats. The possibility exists that an increase in the legal size limit may lead to a decrease in yield if the incidental harvest mortality is as great as pre- viously believed. The objectives of this study were: 1) to model the effects of a change in the incidental hack mortality on the yield of a clam flat and 2) to empirically test the assumption of a 50% incidental hack mortality rate on the soft-shell clam in different locations, at different times of the year, and with different size classes. 284 ROBINSON AND ROWELL Fa Mae > % Figure 1. Photograph of a typical clam hack (fork) used to harvest clams in Canadian maritimes. MATERIALS AND METHODS Modelling A simple yield model was devised to estimate produc- tion of a single year-class (10 million individuals) of soft- shell clams over the duration of their lifespan which was assumed to be 10 years. Growth was modelled with a von Bertalanffy growth curve from data on clams from the Scotia-Fundy area (Angus et al. 1985, Angus & Woo 1985, Mullen & Woo 1985). The von Bertalanffy growth equa- tion used was: 101.44 (1 — e(—0.093 (t — 0.21))) (1) Length (mm) = Annual natural mortality rates were assumed to be 70% in the first 6 months, 35% in the second 6 months and 10% for the duration of the 10 year life span. Fishing mortality (clams harvested) was assumed to be 60% for legal sized clams (Robinson & Rowell, unpublished data). The hack mortality rate was varied from 0 to 60% and the legal size limit was varied from 19 to 57 mm in approximately 6 mm increments. Growth and mortality were calculated at 3 month intervals (i.e. quarterly). Field Studies Seasonal Mortality Six sites were selected to test for hack mortality on soft- shell clams; three in New Brunswick and three in Nova Scotia (Fig. 2). These sites were all in areas that are com- mercially harvested. Clams (25—32 mm shell length) used in the study were collected from the Block House, St. Andrews, for experi- ments in New Brunswick and from Clam Harbour, near Halifax, for those in Nova Scotia (Fig. 2). They were pas- sively stained (pink), by holding for S—7 days in sea water containing a solution of Alizarin Red sodium monosul- fonate, following the technique of Hidu and Hanks (1968). This enabled us to readily identify our clams and to know how many clams should be recovered from the experiment. At each site, the clams were planted individually by hand, in a horizontal row of four plots (each 1 m X | m) near the mid-tide level at a uniform density of 100 m~? (Fig. 3), and allowed to burrow and acclimate for three days. The marked clams were planted among naturally occurring clams. The two experimental plots were then dug in an ex- perimentally standardized ‘‘commercial’’ fashion using a clam hack. After two weeks, which was considered suffi- cient time for any mortality to manifest itself, both experi- mental and control plots were excavated and sieved (4 mm mesh) to remove all marked clams present. The experi- ments were carried out in three seasons; spring (April 1989), late summer (August 1989), and late winter (Feb- ruary 1990). Mortality rates were calculated in two ways. The rela- tive mortality rate between the control and experimental plots was calculated as: M = (Emean(D/T) — Crean(D/T)) X 100 2) mean where M = mortality rate (%), D = number of dead clams, T = total number of recovered clams, Enea, = mean of experimental plots (D/T), and C,,.,, = mean of control plots (D/T). We also calculated the absolute min- imum and maximum mortality rates to bracket the relative value. The minimum mortality was defined as the total number of dead clams recovered in the experimentally dug plot. The maximum mortality was defined as the difference between the total number of live clams found in the experi- mental plot and the number initially planted (100). Size Related Mortality Results of the initial experiments indicated that late summer (August) was the season with the highest hack in- duced mortality. August 1990 was consequently chosen for experiments on size related mortality. Two sites, one which had shown generally high hack mortality levels, Poco- logan, and another which had demonstrated relatively low mortality levels, Thornes Cove, were chosen for this ele- ment of the study. Four size ranges representative of sizes present in local population were selected”: 13-19 mm, 25-32 mm, 38—44 mm, and 51—57 mm. Staining, planting, harvesting, and subsequent analyses were iden- tical to that described above. Sediment Analysis Two replicate core samples were taken to a depth of 20 cm from each study site and were analyzed using a standard sieve series in the soft sediment laboratory of the Atlantic Geoscience Centre at the Bedford Institute of Oceanog- ?These size-classes correspond to: 0.50"—0.75", 1.00"—1.25", 1.50"— 1.75", and 2.00’—2.25". The two larger size-classes bracket minimum size limits under management consideration, while the two smaller size- classes are representative of clams about to recruit to the fishery. INCIDENTAL FISHING MORTALITY ON MYA ARENARIA [ (New Brunswick yy ery2 - Z New Brunswick 285 Pocologen % Herbour Upper Economy Nova Scotia Maces Bay Thornes Cove Figure 2. Map showing the location of the six experimental sites in New Brunswick and Nova Scotia used in this study. raphy. Results from the grain size analysis of the sediment were categorized as mud (grain size <0.067 mm), sand (grain size =0.067 mm and <2.4 mm), and gravel (grain size =2.4 mm). The means of the replicates were used to calculate the percentage substrate composition at each site (Fig. 4). RESULTS Modelling Modelling results indicated there was a dramatic de- crease in the yield from the year-class as the incidental hack mortality increased from 0—60% (Fig. 5). For clam hack mortalities that were 20% or less, raising the size limit to a commercially acceptable size (38-50 mm) generally in- creased yield. As hack mortality increased, maximum yields from the year-class were obtained from smaller clams. Field Studies Seasonal Mortality Mean recovery rates of marked clams from the plots ranged from 52—99.5% with an overall mean of 86% (Table 1). There were no significant differences among sites with respect to recovery rates (ANOVA, P > 0.05), DUG CONTROL SHOREWARD pl ena | | | | | SEAWARD 9M v Figure 3. Experimental design of plot layout used in this study. 286 Mud 3.47% Oak Bay Block House Gravel 41.43% Sand 55.1% 9.24% 0.12% 24 88% Cole Harbour Economy 90.64% ROBINSON AND ROWELL 6.33% 16.57% Pocologan 37.64% 52.74% 6.67% 23.94% Thornes Cove 74.86% 51.18% Figure 4. Sediment profiles of the six sites used in this study broken down into mud (grain size <0.067 mm), sand (grain size =>0.067 mm and <2.4 mm), and gravel (grain size >2.4 mm). however there was a significant difference between experi- mental and control plots (ANOVA, P < 0.001). In addi- tion, recovery rates from experiments conducted during April and February were not significantly different from each other (ANOVA, P > 0.05) but, both were signifi- cantly higher than August (ANOVA, P < 0.001). Percent mortality due to the clam hack ranged from 2—48% with a mean of 17.6% + 15.3 (S.D.). The highest mortality rates were found in Procologan and Economy and approached 47%, however, most of the other areas were much lower (Fig. 6). Except for the Block House site, mor- tality rates were higher in summer than in either spring or winter. Absolute minimum and maximum mortality rates TABLE 1. Mean percent recovery rates of planted clams from experimental and control plots from the six different study sites at three different times of year. Oak Bay, Economy, and Cole Harbour were not sampled during February due to adverse weather conditions. Numbers in parentheses indicate standard deviation of the mean. Percent Recovery Rates April August February Site Exp. Cont Exp. Cont. Exp. Cont Oak Bay 94.0 99.5 72.5 83.0 — — (4.2) (0.7) (2.1) (0.0) Block House 95.0 94.5 65.0 79.0 87.5 99.0 (0.0) (2.1) (14.1) (0.0) (2.1) (0.0) Pocologan 73.0 93.5 57.0 83.0 93.5 92.5 (0.0) (9.2) (2.8) (2.8) (2.1) (6.4) Economy 74.5 96.5 52.0 86.0 — = (3.5) (0.7) (9.9) (4.2) Cole Harbour 88.5 99.5 85.5 98.0 _ = (2.1) (0.7) (4.9) (2.8) Thornes Cove 81.5 97.5 87.5 98.0 83.5 96.0 (6.4) (0.7) (0.7) (2.8) (2.1) (0.0) Site Mean 84.4 96.8 69.9 87.8 88.2 95.8 (9.6) (2.5) (14.6) (8.2) (5.0) (3.3) INCIDENTAL FISHING MORTALITY ON MYA ARENARIA 10 Year Yield from 1 Year-class (10 million spat) Yield (mt) (31.8) Size Figure 5. Results of modelling trials on the relationship between yield and minimum size limits with varying incidental hack mortalities. showed the same general pattern as relative mortality rates with the majority of the mortality ranges falling substan- tially below the 50% level (Fig. 7). The range tended to be greater in summer months. The ratio of unbroken to broken shells of dead clams recovered from the experimental plots was highest in the summer period (Table 2), with an overall ratio of 4.0. During August, the Oak Bay site had a very high number of unbroken shells (57 unbroken shells out of 58 dead ones recovered). The overall ratio for all three seasons and six sites was 2.0. Size Related Mortality Relative mortality rates for all sizes of marked clams at both sites were all substantially less than 50% with an overall mean of 15.4% + 9.0 (S.D.) (Fig. 8). As the sea- sonal study showed, Pocologan had a higher overall mor- tality rate (23% + 3.3 (S.D.)) than Thornes Cove (7.8% + 4.5 (S.D.)). There was no clear trend with mortality rate and size of clam (Fig. 7). The ratio of unbroken to broken dead clams recovered from the experimental plots in Pocologan (3.0) was three times higher than that observed from Thornes Cove (1.0) 50 April = 40 B August = OO February is s 30 = c oaez0 c a a Block House Oak Bay Pocologan Economy Cole Harbour Thornes Cove Figure 6. Relative mortality rates (see text) of clams due to the clam hack at six different sites and at three different times of the year. 287 TABLE 2. The ratio of unbroken to broken dead clams recovered from experimental plots at the six study sites over three seasons. Oak Bay, Economy, and Cole Harbour were not sampled during February due to adverse weather conditions. Season Site April August February Total Oak Bay 125 57.0 _ 4.6 Block House 0.9 2.0 0.2 0.8 Pocologan 2.0 7.8 0.8 2.6 Economy 0.0 2.6 — le Cole Harbour 0.3 1:3 — 1.0 Thornes Cove 0.3 1:3 2.0 0.9 Total 1.2 4.0 0.7 2.0 (Table 3). The overall ratio from both sites was 1.9. There was an inverse relationship between clam size and the ratio at Thornes Cove, however, the ratio stayed relatively con- stant at Pocologan except for a high value of 27.0 (27 un- broken:1 broken) for the 38—44 mm size group. DISCUSSION Modelling results indicated that incidental hack mor- tality could significantly affect clam stocks if it was as high as previously predicted by Medcof and MacPhail (1964). If their estimates are correct (50%), and our assumptions on growth and natural mortality are valid, then high fishing effort would substantially decrease the yield from the fishery. At present, the commercially acceptable (opera- tional) size limits in the industry are 38-60 mm. In this size range, the yield drops off with increasing size for hack mortalities of 30% and greater. Therefore, the implications of this model warranted a re-testing of the hypothesis that clam hacks inflict an incidental mortality rate of 50%. The high recovery rates found in these experiments indi- cate that the techniques used for recapturing the marked clams were effective and that any differences between treatments were real. Experimental plots were found to sail April a — — August a | | | February Gale | > = | eee oe B= | | | | 3 | | | | o [| | | | | a © 20 | | oO r | 1 ied | : | + Li | [ | ' v ' IP ocologan Economy Cole Thornes Harbour Cove ie) Oak Bay Block House Figure 7. Absolute mortality rates (see text) of clams due to the clam hack at six different sites and at three different times of the year. 288 ROBINSON AND ROWELL TABLE 3. Ratio of unbroken to broken dead clams recovered from the experimental plots at Pocologan and Thornes Cove for soft-shell clams of different sizes in August 1990. Size Class (mm) Site 13-19 25-32 38-44 51-57 Total Pocologan 2.6 2.2 27.0 De 3.0 Thornes Cove 11.0 it57/ 0.8 0.5 1.0 Total 3.4 2.0 3:2 1.0 1.9 have less clams than control plots upon harvesting after the two-week period. This was probably due to increased pre- dation pressure on the plots that were dug as the distur- bance of the sediment would undoubtedly attract predators such as herring gulls, Larus argentatus Pontoppidan, on exposed clam flats (Robinson, pers. obs.). It is also pos- sible that displacement from normal living depth and ef- forts of buried clams to reposition themselves could make them more susceptible to predation by other infaunal pred- ators such as the green crab, Carcinus maenas Linné, and the nemertean ribbon-worm, Cerebratulus lacteus Verrill, (Rowell & Woo 1990). Emerson et al. (1990) found that clams exposed on a mud sediment reburrowed to abnor- mally shallow depths. Recovery rates also differed between summmer and the other two seasons (spring and winter). This is consistent with the argument that predation pressure would tend to increase in summer due to higher tempera- tures which would promote more predator activity. Differ- ences in recovery rates between sites were probably due to different predator complexes in these areas. The mean seasonal hack mortality of 17.6% was sub- stantially less than the 50% originally suggested by Medcof and MacPhail (1964) for clams in the 25—32 mm size cate- gory at some of the same flats used in their study. The absolute ranges of mortality also indicated that mortality rates were generally below 50%. In all cases but four, the absolute maxima were well below the 50% level. The same BH Pocologan @ 3 Thornes Cove = - = ° = 4 c o ° re o a 13-19 mm 25-32 mm 38-44 mm 51-57 mm Size Class Figure 8. Relative mortality rates of the clam hack on different size ranges of clams at Pocologan and Thornes Cove. result was found with clams of differing sizes. In these ex- periments the highest mortality rates found were about 25%. The reason for the discrepancy between our results and those of earlier studies may be a function of the re- covery rates achieved. Medcof and MacPhail (1964) as- sumed equal densities of clams in their plots and also that recovery rates would be equal. This may not have been the case as we have observed clam densities to be quite patchy. The ratio of unbroken to broken dead clams recovered in the experimental plots may give an indication of the causes of mortality. For example, an unbroken shell would sug- gest that the animal smothered while a broken one might indicate hack-induced damage perhaps in conjunction with smothering. During the seasonal mortality study for clams 25—32 mm, the overall ratio in April was 1.2 which sug- gests that slightly more were suffocated than broken. In August, the ratio increased to 4.0 suggesting that suffoca- tion was much more prevalent. The ratio in February dropped back to 0.7 suggesting more hack damage than suffocation. The ratio for different sizes of clams in summer were also generally greater than 2.0 except for the larger sizes at Thornes Cove. The pattern emerging from these data appears to relate to temperature. Clams, being poikilothermic animals, metabolize at a rate governed by the external water temperatures. Although M. arenaria has been shown to respire anaerobically (van Dam 1935, Ricketts & Calvin 1968), their ability to withstand oxygen deprivation is obviously limited. If the clam’s ventilation is shut off by being buried during the digging operation, especially in the black anoxic sediments of muddy areas, the time the animal has to right itself and reach the surface is inversely proportional to the in situ temperature. When the flat is exposed in summer and warms up from the sun’s irradiation, the buried clams have less time to recover and therefore suffer higher mortality through suffocation. Highest mortality rates encountered in this study tended to come from the Pocologan and Economy areas. The sedi- ment at these sites was composed of a higher percentage of mud compared to the other four sites whose sediment char- acteristics were more sandy and gravelly in nature. Emerson et al. (1990) have shown that clams exposed on medium-fine sands can reburrow rapidly and re-establish their normal living depths, whereas those exposed on mud require considerably more time to reburrow and they re- burrow to shallower depths. They suggest that, although the mud may be easier to penetrate, it is much more diffi- cult for a surface exposed clam to gain sufficient pedal an- chorage for reburial. They also indicate that upward bur- rowing was more difficult in muddy sediments. In conclusion, it appears that the original estimate of a 50% incidental hack mortality rate is too high for clam stocks in the areas studied. Fisheries managers are cur- rently proposing to raise the minimum size-limit from 38 mm (1.5 inches) to 45 mm (1.75 inches). The overall mean mortality of <20% found in our experiments suggests that, INCIDENTAL FISHING MORTALITY ON MYA ARENARIA 289 according to the above model, there would be a small net gain (about | percent) in yield to the fishery if this size increase was introduced. ACKNOWLEDGMENTS The authors would like to thank R. Chandler, P. Woo, J. Martin, T. McLane, Q. Currie, T. McLeod, B. Thorpe, J. Wildish, D. Robichaud, and W. Young-Lai for their help in the field, both with planting and retrieving the clams. K. Asprey and D. Clattenburg of the Atlantic Geoscience Centre kindly processed the sediment samples. Thanks to B. McMullon and F. Cunningham for photography and drafting. This study was part of the Canada Department of Fisheries and Oceans Clam Enhancement Program. REFERENCES Angus, R. B., C. M. Hawkins, P. Woo & B. Mullen. 1985. Soft-shell clam survey of the Annapolis Basin, Nova Scotia— 1983. Can. MS Rep. Fish. Aquat. Sci. No. 1807: 130 pp. Angus, R. B. & P. Woo. 1985. Soft-shell clam (Mya arenaria) resource inventory—Buckman’s Creek, Charlotte County, New Brunswick— 1984. Can. MS Rep. Fish. Aquat. Sci. No. 1842: 17 pp. Dam, L. van. 1935. On the utilization of oxygen by Mya arenaria. J. Exp. Biol. 12:86—94. 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. 26:(in press). Glude, J. B. 1954. Survival of soft-shell clams, Mya arenaria, buried at various depths. Maine Dept. Sea and Shore Fisheries, Res. Bull. No. 22, 26 pp. Hidu, H. & J. E. Hanks. 1968. Vital staining of bivalve mollusk shells with alizarin sodium monosulfonate. Proc. Nat. Shellfisheries Assoc. 58:37-41. Medcof, J. C. & J. S. MacPhail. 1962. A new hydraulic rake for soft- shell clams. Proc. Nat. Shellfish Assoc. 53:11—32. Medcof, J. C. & J. S. MacPhail. 1964. Fishing efficiency of clam hacks and mortalities incidental to fishing. Proc. Nat. Shellfisheries Assoc. 55:53-72. Mullen, B. & P. Woo. 1985. Soft-shell clam resource survey of Three Fathom Harbour and Clam Harbour, Nova Scotia—1985. Can. MS Rep. Fish. Aquat. Sci. No. 1877: 39 pp. Needler, A. W. H. & R. A. Ingalls. 1944. Experiments in the production of soft-shelled clams (Mya). Fish. Res.. Bd. Canada, Atlantic Prog. Rep., No. 35:3-8. Ricketts, E. F. & J. Calvin. 1968. Between Pacific Tides. (4th ed.) Stan- ford University Press, California. 614 p. Rowell, T. W. & P. Woo. 1990. Predation by the nemertean worm, Cer- ebratulus lacteus Verrill, on the soft-shell clam, Mya arenaria Lin- naeus, and its apparent role in the destruction of a clam flat. J. Shell- fish Res. 9:291—297. 7 one es +a =; ae * Journal of Shellfish Research, Vol. 9, No. 2, 291—297, 1990. PREDATION BY THE NEMERTEAN WORM, CEREBRATULUS LACTEUS VERRILL, ON THE SOFT-SHELL CLAM, MYA ARENARIA LINNAEUS, 1758, AND ITS APPARENT ROLE IN THE DESTRUCTION OF A CLAM FLAT T. W. ROWELL AND P. WOO Department of Fisheries and Oceans Biological Sciences Branch, Habitat Ecology Division Bedford Institute of Oceanography P.O. Box 1006 Dartmouth, Nova Scotia, Canada B2Y 4A2 ABSTRACT Populations of the soft-shell clam, Mya arenaria Linné, in the Annapolis Basin, Nova Scotia, declined throughout the mid to late 1980’s. In the upper reaches of the Basin there are large areas of once-productive clam flats where no live clams could be found between 1987 and 1989. It appears that environmental and biological factors combined to push the population within these areas to collapse. In the areas affected, there was no settlement and (or) survival of soft-shell clams for several years and, in consequence, no recruitment to the stock. A surface sediment layer of watery silt (1.5—2.0 cm thick) was characteristic of these areas. Other areas in close proximity to, or interspersed with, those affected, but lacking the watery surficial sediment layer, continued to support small populations of clams. In the affected areas, where no recruitment was occurring, the nemertean worm, Cerebratulus lacteus Verrill, preyed on, and ultimately eliminated, the already existing clams. Although the literature contains several papers suggesting C. lacteus to be a predator of M. arenaria, no experimental studies have been reported. Field and laboratory studies reported here confirm that C. /acteus is a predator on M. arenaria, and not simply a scavenger, and that it may, in consequence of its partially free-swimming life-style and its ability to seek out its prey, have a devastating impact on clam stocks that have been reduced by overfishing and (or) environmental change. KEY WORDS: INTRODUCTION Population levels of the soft-shell clam (Mya arenaria) in the Annapolis Basin, Nova Scotia, Canada, declined throughout the mid to late 1980’s. The decline may, in part, be attributed to overfishing during the late 1970's and throughout the early and mid 1980’s when reported catches approached historic highs.' Overfishing cannot, however, explain the complete disappearance of clam populations from some areas of the upper Basin. On both shores, for roughly 9 km, there are large areas (>1 km7?), of once-pro- ductive clam flat where no live clams could be found from 1987 to 1989 (Fig. 1). Population surveys had been con- ducted in these areas in 1983 (Angus et al. 1985) and com- plaints by fishermen that environmental changes, sedimen- tation in particular, were responsible for the decline in the fishery had prompted a resurvey of one of these areas, Oak Point, in July 1986. The resurvey showed that a dramatic decline in population density had taken place in the inter- vening three years; from a mean density of 781 clams/m? in 1983 to 30 clams/m? in 1986 (unpublished data). The flat was again visited in April 1987 and an extensive area was examined by digging with shovels and clam hacks (in- cluding sieving the substrate through a 1.18 mm? mesh screen at a number of locations). In the 9 months that had elapsed since the 1986 survey, the entire clam population, \Statistics Division, Fisheries and Oceans, Scotia-Fundy Region, P.O. Box 550, Halifax, N.S., B3J 2S7. predation, worm, Cerebratulus lacteus, clam, Mya arenaria except for remnants in areas where rivulets crossed the flat and in the highest levels of the beach, had disappeared (Prouse et al. 1988). In the highest ~20 m of flat, where live clams could still be found, the substrate graded from a sandy mud to gravelly sand as one proceeded shoreward. The substrate over the remainder of the flat, where no live clams could be found, was a very firm blue marine clay, covered by a 1.5—2.0 cm thick layer of fine watery silt. This layer was absent in the rivulets running down over the flat. Throughout the area, the clay layer was filled with the still articulated empty shells of clams of all sizes. Other benthic organisms were found throughout the flat, appar- ently unaffected by whatever had affected the clam popula- tion. There being no obvious reason for such an apparently catastrophic mortality, a number of field and laboratory in- vestigations were undertaken. Initially these were directed at determining if there had been some epizootic event or a catastrophic environmental change, such as a chemical spill, which was somehow clam specific (Prouse et al. 1988). Here we are reporting the results of a series of observa- tions, field samplings, and experiments which, taken to- gether, strongly support the hypothesis that environmental change coupled with predation by the nemertean worm Cerebratulus lacteus, led to the complete collapse of this clam population. C. lacteus is known as a predator, its main prey believed to be polychaetes (Coe 1943, Wilson 1900). Field observa- 292 ROWELL AND Woo 65°40" BAY OF FUNDY iG PORT, MARSH 4” Ge ( S fRoyau re 4 fy \ a LLWS 4 44 )THORNESH / ex 44° 42 AGE goat Ly Fs ) alas i . y \ Lox ‘ e y / POINT / g x \ Be eSe~ | | y 7 ae } y 0 3000 | a y = “METRES a ANNAPOLIS y 4 —, AREAS SAMPLED BASIN ( | FOR CLAM SPAT | page ~ AND JUVENILES | S | 2 oye 65°40' 65°34" Figure 1. Annapolis Basin, Nova Scotia, showing Oak Point and the location of the Port Royal to Queen Anne Marsh and Thornes Cove flats. LLW indicates lowest low water and the extent of the mudflats. Darkly stippled area identifies approximate intertidal zone. tions have suggested C. lacteus to be a predator on the molluscs Ensis directus Conrad (McDermott 1976, McDer- mott & Roe 1985, Schneider 1982) and M. arenaria (Kalin 1984). The present study provides the first experimental evidence that C. /acteus is a true predator of M. arenaria, demonstrates the nemertean’s ability to seek out its prey, and suggests the clam may be a preferred prey. MATERIALS AND METHODS Transplant Experiment Initially, on discovery of the mortality in April 1987, it was decided to determine if it had been an isolated event or if the cause of the mortality was still acting. Experimental plots of live clams were establilshed on areas of the Oak Point flat between May 27 and June 4 of that year (Fig. 1). Three plots (5 m X 7 m) were each divided into 35 sub- plots; two plots (A and B) were set up near the mid-tide level of the flat, and a control (C) in the gravelly sand of the higher level of the flat near the donor site from which the clams for the experiment were taken. Except for loca- tion, clams used in the control plot were treated identically to those in experimental plots A and B. Corners of the plots were marked by capped iron survey stakes driven into, and almost flush with, the surface of the flat. Each plot was divided, using marked ropes and the comer stakes, into 35 sub-plots of 1 m?; either 6 or 12 sub- plots for experimental plantings of clams, and the re- maining 23 or 29 sub-plots serving as spacings to allow access for sampling of the planted clams (Fig. 2a). In plot A, 12 sub-plots were randomly assigned to be planted with either 50 or 100 clams. The clams, of 14—32 mm shell length, were marked on one valve with red nail polish. They were planted, evenly spaced, at a depth of roughly 30—60 mm in the centre 0.25 m? of the selected sub-plots. In plots B and C, 50 clams were planted in each of 6 ran- domly selected sub-plots. Clams being left over after the (A) 1 metre PLANTED AREA 7 metres 1 metre 5 metres Th 8 = 7 Es ] re] i | 5] = —— f Lo See E rae I s G H | J 2 PLANTED = EXPERIMENTAL PLOTS UNPLANTED Saree cie PLOTS ee ~0.75m ~. ae E 3\o 3 t —s Ke 5 SAMPLED / AREA ~ Figure 2a. Plot and sub-plot lay-out for the Preliminary Transplant Experiment at Oak Point. Figure 2b. Plot lay-out for Predation Experiment II on the Oak Point Clam flat, showing plots planted with marked clams and areas sam- pled for general density of Cerebratulus lacteus (unplanted plots). planting of plot A, it was decided to establish a small sepa- rate 0.25 m? control plot (C’) close to the original control plot (C). The intention was to sample selected sub-plots on a monthly basis for the first three months and subsequently at two month intervals. On each sampling date, four sub-plots were to be randomly selected for sampling; one of 50 and one of 100 clams/0.25 m? from plot A and one of 50 clams/0.25 m? from plots B and C. However, after ob- serving mortality patterns up to the second sampling date (July 29), the schedule was abandoned and final sampling carried out on August 5. On this date, one 100 clam/0.25 m? sample of plot A and one 50 clam/0.25 m? sample of plot B were taken. Plot C’ was also sampled on this date, as were two additional 0.1 m? areas of the donor site. For sampling, the centre 0.25 m? of each sub-plot was first marked using a grid. This central area, plus an addi- EEE OO PREDATION OF CEREBRATULUS LACTEUS ON MYA ARENARIA 293 tional 5—10 cm border on each side, was then dug out to a depth of approximately 15 cm. Samples were washed through a | mm? mesh screen to ensure collection of both planted clams and any small clams or recently set clam spat which might have been present. Predation Experiments Observations made during the transplant experiment suggested that C. /acteus might be preying on clams and led to further experiments aimed at determining if the worm was a predator or simply a scavenger of already dead or stressed clams. Experiment I An initial experiment, to examine the possibility that predation was a factor in the observed mortalities and that stress might have made the Oak Point clams more suscep- tible to predation, was carried out between August 19 and October 5, 1987. Four replicate groups of 14 Oak Point and 10 Cole Harbour clams (from a population showing no evi- dence of unusual mortalities) along with 4 C. lacteus were placed in buckets filled with Oak Point sediment. A sepa- rate tank, with buckets of Cole Harbour sediment con- taining several hundred Cole Harbour clams but no C. lacteus, was used as a control. All buckets were held in tanks of filtered, running seawater. Oak Point clams ranged in shell length from 21—71 mm (x = 32 mm, s.d. = 15) while those from Cole Harbour ranged from 15—74 mm (x = 37 mm, s.d. = 20). Sampling of individual buckets (groups) took place on August 24, August 31, September 23, and October 5, and all live and dead clams as well as worms counted. The control was sampled only at the con- clusion of the experiment. Experiment II A second transplant experiment was set up on the Oak Point flat in the spring of 1989 in order to determine if, after two years, clams could survive and grow there, and if predation by C. lacteus had caused eradication of the popu- lation. Clams, 25—32 mm in shell length, were obtained from Clam Harbour, N.S. and their shells stained for identifica- tion purposes by holding them approximately one week in a tank of sea water containing a 25 mg/I solution of Alizarin Red. On May 8, ten 0.25 m? plots, each separated by 1.5 m, were established in a block along three lines (Fig. 2b) on the mid-tide level of the Oak Point flat very close to the position of plot A of the original 1987 transplant experi- ment. Each plot was planted with 50 stained clams. The experimental design called for 2 randomly selected plots to be sampled on each of 5 sampling dates at intervals of 30 days. This schedule was adhered to for the first 60 days, when results indicated a need to compress the sampling in- tervals in order to maximize information gain. Further sampling was carried out at days 64, 71, and 90. Sampling consisted of digging up the area of each plot, including an additional buffer zone around its margin (0.75 m X 0.75 m, depth = 15 cm). Numbers of live and dead clams and of C. lacteus were recorded and the shells of all clams were retained for measurement of growth. The con- dition of all dead clams was also noted relative to the pres- ence of remaining body parts such as the periostracum bor- dering the valve edge and the siphon. On the last three sampling dates, total live weight of C. lacteus within the plots was also recorded to 0.01 g accuracy. Four additional areas of equal size, located roughly 3 m outside and distrib- uted equally around the plots, were then also sampled to determine general densities of the worm on the flat (Fig. 2b). Spat Settlement and Survival From random sampling in April 1987, over large areas of the Oak Point flat and much of both the north and south shores of the Basin upstream of Goat Island (Fig. 1), it became apparent that clam spat were either not settling on large areas of the flats or, if they were, they were not sur- viving to the juvenile and adult stages. The only areas still showing evidence of recruitment to the population were where rivulets cut down across the flats and in the high levels of the intertidal. In 1988, a series of transects was set up on some of the main clam flats in the Basin to monitor settlement and early survival of juvenile clams. Transects were again sampled on several of the flats in 1989 and 1990. The flats where transects were sampled in all three years are shown in Figure |. Queen Anne Marsh, on the shore opposite Oak Point, had similar appearing sediments to Oak Point and differed only in that it still supported a clam population. Many areas further upstream on the Queen Anne Marsh shore did not; apparently affected in the same way as Oak Point. The third location, Thornes Cove, is several km downstream of the others and has suffered neither discernible siltation or unusual decline in clam pop- ulation. In 1988, at Oak Point and Thornes Cove, starting in the high intertidal, samples were taken at every 20 m to the lowest level of the intertidal accessible on the day of sam- pling. Because the intertidal was narrower at Queen Anne Marsh, samples were taken every 10 m. In 1989 and 1990, sampling stations were spaced at 50 m intervals along the transects and marked with steel survey stakes. Sampling consisted of marking a 0.10 m? area with a grid, digging the sediment out to a depth of approximately 5 cm, and bagging the removed sediment for later sorting. Samples were sorted through sieves at the laboratory and the shell length of clams were measured to the nearest mm. RESULTS Transplant Experiment Examination of the plots during the first week after planting indicated that the clams were well established and 294 ROWELL AND WOO TABLE 1. Mortalities of planted Mya arenaria observed at Oak Point during May—August 1987. Elapsed Planting Sampling Days Since Density Marked! Clans Mortality* Date Plot Planting (no./0.25 m2) Alive Dead (%) June 29 A 34 50 10 15 60-80 100 64 31 33-36 B 32 50 24 24 50-52 (e 25 50 9 34 79-82 July 29 A 64 50 0 17 100 100 0 86 100 B 62 50 0 47 100 G 55 50 0 43 100 Aug. 5 A 70 100 0 65 100 B 68 50 0 35 100 (Ge: 69 100 0 66 100 * Minimum of range based on number of dead clams (generally, shell only) as a percentage of total live and dead recovered, while maximum assumes that all live clams will have been recovered. The assumption is that any missing clams are dead and have either been displaced from the plots or have lost the shell marking and therefore could not be identified even when recovered. ** C’ was an individual 100 clams/0.25 m? plot very close to Plot C. actively pumping. After 25—34 days from planting, 33—80% were dead in experimental plots A and B and 79-82% in the control plot C (Table 1).2 A second sam- pling of the plots 55—64 days after planting showed com- plete mortality in all four samples. Further sampling, on August 5, of plots A and B, as well as the small plot C’ located near the control plot C, also indicated the complete mortality of all planted clams. On the same date, two 0.1 m? samples were taken from the donor area, only 10-15 m distant from control plot C, to determine if clams there were now being affected by the mortality. The first yielded the articulated shells of 29 recently dead clams and no live clams, while the second yielded 16 dead and 17 live. These results indicated that the donor area was now being affected by whatever was causing the mortality, but also that some clams were still surviving in the higher reaches of the beach. At this time, some C. /acteus were observed inshore of control plot C but none were noticed in the experimental plots. Any C. lacteus which might have been encountered during the sampling prior to August 5 had gone unnoticed, suggesting that, if present, they were not abundant enough to arouse suspicion as a factor in the mortalities. A week later, however, when several other small areas of the upper flat 30—50 m from the donor area were dug, the association between C. lacteus and M. arenaria became much more evident. In some areas, there was no evidence of recent clam mortality and there were no C. lacteus found, in other areas recently dead clams were found in close association with C. lacteus. The abundance of C. lacteus and their ?The minimum in these ranges is based on the number of dead clams as a percentage of the total of live and dead clams recovered, while the max- imum assumes that all live clams have been recovered and that all missing clams are dead. proximity to the clams suggested strongly that they were either preying upon or scavenging on the clams. Several clams were observed with only remnants of their soft parts and the periostracal covering of the mantle edge and sipon remaining. Some of the clams had a thick clear mucous glob in the mantle cavity; indicative of nemertean feeding as described by McDermott (1976) and Kalin (1984). Although there was some evidence of predation by the spotted northern moon-shell, Lunatia triseriata Say, in all of the plots examined, the incidence was generally quite low (2% or less drilled). Predation Experiments Experiment I Results of the preliminary trial to determine the relative susceptibility of Oak Point and Cole Harbour clams to pre- dation by C. lacteus are presented in Table 2. When ex- posed to predation by C. lacteus, the pattern of mortality for clams from both sources was very similar, suggesting TABLE 2. Survivorship of Oak Point and Cole Harbour clams when held with Cerebratulus lacteaus. Percent Surviving Combined Elapsed Oak Point and Days Oak Point Cole Harbor Cole Harbour Control 5 93 90 92 12 79 50 67 35 14 20 17 47 0 0 0 100 PREDATION OF CEREBRATULUS LACTEUS ON MYA ARENARIA 295 that any possible environmental stress that the clams from Oak Point might have been under was not making them more susceptible to predation. Mortality through predation was complete in a period of 36—47 days. There was no mortality among the controls. Experiment IT The results of the experiment are presented in Figure 3, the data for each sampling date having been combined and the means plotted. On the first sampling, at day 30, mortalities of 12 and 20% were observed. The two plots contained 7 and 3 C. lacteus, respectively. At 60 days the clams had suffered 92 and 96% mortalities and the numbers of worms increased to 15 and 10. At 64 days, only one plot was sampled. Mor- tality in this plot was 82% and 15 worms were found. Sam- pling of the 4 areas outside the plots yielded 3 worms. Sampling on day 71 indicated mortalities of 86 and 98%, with 11 and 5 worms in the plots. Sampling outside the plots produced no worms. On day 90, there was 100% mor- tality in the three plots sampled. One worm was found in () NUMBER OF SAMPLES 44 ¢ MEAN AND RANGE 40 60} a S ° se, 744 “+y,,42(2) 40. *.. SURVIVING MYA NUMBERS OF SURVIVING MYA 8 8 tS} oS e INSIDE PLOTS NUMBERS OF CEREBRATULUS / 0.56 m2 ELAPSED DAYS Figure 3. Percent survival of Mya arenaria over time when planted on the Oak Point clam flat in Predation Experiment II. Numbers of Cer- ebratulus lacteus present within the experimental plots and in the sur- rounding area of the flat are also shown. one of the plots. The four samples from outside the plots produced 3 worms. Mean numbers and weights/m? for C. lacteus captured inside and outside the experimental plots are given in Table 3. When the data for all days with sampling both inside and outside of the plots, including day 90 when no live clams remained, are examined, worms were 15.5 times by number and 23 times by weight more abundant inside the plots. When only the data for days 64 and 71, when clams were still available in the plots, are considered, worms were 30.5 times by number and 62.5 times by weight more abundant inside the plots. The differences between number and weight ratios result from higher average weights of the worms inside the plots. Although the snails Lunatia heros Say, L. triseriata, and Nucella lapillus Linné as well as the green crab Carcinus maenas Linné, known predators of the soft-shell clam, were all present on the flat, none of the 426 clams recov- ered, either alive or dead, showed evidence of predation by these species (drilled, crushed, or chipped shells). All clams exhibited excellent growth; those still alive at day 71 having a mean shell length increase of 34.4% (s.d. = 8.4). Spat Settlement and Survival Random sampling on the Oak Point flat, during the 1987 to 1989 field experiments, and over both the north and south shores of the Basin upstream of Goat Island, indi- cated that clams were either not settling or not surviving on large areas of previously productive flat. The single evident difference in these areas was a thin watery surface sedi- ment. In the four years of sampling at Oak Point, no evi- dence of clam settlement and, or juvenile survival of the 1986, 1987, and 1988 year-classes was found over the main area of the flat except in the sandy bottoms of rivulets flowing down over it. In these rivulets, the watery sediment layer which covered the rest of the flat was absent. During these years, clam settlement also occurred in the highest reaches of the intertidal, where the gravelly mud substrate was again not covered by the watery sediment layer. Overall, clam settlement or survival was observed to be restricted to those areas where the watery surface sediment was absent. Prior to the 1990 sampling of the established transects, a visit to the Oak Point flat in mid-April revealed that a major settlement of clams had taken place in 1989. The entire flat, including the main mid-tide area where the watery sed- iment layer had for a minimum of three years apparently blocked such settlement, was peppered with 2—10 mm ju- venile clams. Any changes which may have occurred in the watery silt layer between the period of no recruitment and the present were not visually discernible. Results of 1988, 1989, and 1990 sampling of juvenile clams over the three flats monitored are given in Table 4. 296 ROWELL AND WOO TABLE 3. Mean number and mean live weight (g) of Cerebratulus lacteus per square meter inside and outside of the experimental plots. Ratios of inside to outside are also given. Weight Inside Numbers All days when both inside and outside sampled 13.8 Peak days when both sampled and while clams still available as prey 20.4 Ratio* Outside Inside/Outside Numbers Weight Numbers Weight 41.5 0.9 1.8 15.5 23.0 61.8 0.7 1.0 30.5 62.6 * Apparent discrepancies in ratios are due to rounding to one decimal place for the presentation of numbers and weights per square meter. The data confirm the lack of clam settlement and survival on the main areas of the Oak Point flat in 1987 and 1988, followed by a very large settlement and survival in 1989. The data also show that the Queen Anne Marsh flat, on the opposite shore from Oak Point, had very good settlement and survival, relative even to Thornes Cove, in all three years. DISCUSSION The results of these field and laboratory experiments demonstrate clearly that C. lacteus is a predator of clams and that it is highly effective in seeking out its prey. The high worm biomass found predating within the plots of Ex- periment II, in the middle of a flat otherwise devoid of clams, suggests that M. arenaria may be a “‘preferred”’ food of this species. Blue mussels (Mytilus edulis Linné), which were disbursed throughout the flat and readily avail- able as potential prey, had suffered no obviously discern- ible mortalities. This may reflect a preference by C. lacteus for clams, or it may suggest that the nemertean is primarily an infaunal feeder. The latter is supported by its well docu- mented predation on polychaetes (Wilson 1990, Coe 1943, McDermott & Roe 1985). Wilson (1900) notes the close association of Nereis virens Sars, M. arenaria, and C. lacteus and remarks on C. lacteus being known locally to fishermen as a ‘‘clam-worm’’. C. lacteus has also been rec- ognized as a principal constituent of the Mya-Nereis virens biome within the Bay of Fundy (Newcombe 1935); again indicating the close ecological relationship between the three species. In the initial and subsequent field investigations of the Oak Point clam mortalities, it was observed that M. aren- aria could still be found in the highest areas of the beach. Although C. lacteus was found preying on M. arenaria even at this level, it is likely that here the worm is on the fringe of its intertidal distribution. Wilson (1900) states that C. lacteus occurs most plentifully just above low-water mark, but that it may be found, in sheltered positions (ap- parently meaning unexposed shores), nearly up to the high- water mark of medium tides. Repeated sampling on the flat revealed that, in addition to the complete mortality of adult clams, no settlement or survival of small clams was occurring for a minimum of three years. The cause for lack of settlement of clam larvae and (or) the survival of juvenile clams during these years appears to have been the thin watery silt layer covering the clay substrate. The growth observed in the clams used in Experiment II clearly indicates that clams, once estab- lished, could do well if not destroyed by predators. The finding of juvenile clams in a few areas where rivulets had washed away the watery silt layer, exposing the firmer sub- strate below, suggests this layer to have been the sole cause of recruitment failure to this and other affected areas of the upper Basin. Unfortunately there does not appear to be any TABLE 4. Percentage of stations having juvenile clams and mean number per 0.1 m? of juveniles in May 1988 (1987 spat), 1989 (1988 spat), and 1990 (1989 spat) from three areas of the Annapolis Basin. Standard deviations of the mean are given in brackets. Percentage of Stations with Mean Number of Stations Sampled Juveniles Juveniles/0.10 m? Location 1988 1989 1990 1988 1989 1990 1988 1989 1990 Oak Point 40 15 21 0 Ose 95.2 O (N/A) 0.4* (N/A) 69.3 (124.9) Queen Anne Marsh 6 20 24 66.7 90.0 100.0 12.5 (18.0) 34.8 (59.2) 69.4 (131.4) Thornes Cove 93 33 27 8.6 30.3 85.2 0.5 (2.1) 1.7 (4.4) 18.9 (24.0) * Oak Point juveniles (6) were all from Station 1, the most shoreward station in the sand/gravel area of the high beach. PREDATION OF CEREBRATULUS LACTEUS ON MYA ARENARIA 297 way of confirming, either quantitatively or qualitatively, how the sediment layer may have acted to block larval set- tlement and what about this layer has now changed to again permit settlement. One possibility is that sufficient com- paction of the layer may have occurred with the passage of time and rendered the sediment surface again suitable for the settlement and (or) for the survival of larval clams. Such compaction of intertidal sediments is the current focus of a major study in both the Annapolis and Minas Basins of the Bay of Fundy. Based on the combined experimental and observational evidence, it appears that environmental change made the Oak Point flat unsuitable, betwen 1986 (possibly earlier) and 1988, for the settlement and survival of small Mya, thus cutting off recruitment to the population. At the same time, the clam population, its biomass already depressed by 1986 to a critical level by overfishing and, possibly, by the onset of reduced recruitment, was being heavily predated upon by Cerebratulus. The good survival and growth of the 1989 year-class demonstrates that clams can survive on the Oak Point flat provided recruitment keeps the population level high enough to absorb normal levels of predation by C. lacteus. Although earlier observations by McDermott (1976) and Kalin (1984) strongly suggested the role of C. /acteus as a predator of bivalve molluscs, the full importance of this nemertean in the ecology of clam flats was not recognized. Wilson (1900) described in great detail the predatory be- havior of C. lacteus on N. virens, even describing it as a preferred prey. He mentions other worms in general as prey items, but makes no mention of predation on molluscs. As previously noted, he does remark on their occurrence in association with clams and their being referred to as a ‘“‘clam-worm’’. He further noted their ‘‘gregarious”’ be- havior. Considering the apparently high densities observed in the plots of Experiment II, this behavior may in fact be a feeding induced response. The patchy or aggregated distri- bution of C. /acteus observed in the high intertidal area where the clams were obtained for the Transplant Experi- ment, also supports the suggestion of “‘gregarious’’ be- havior and the possible link of this behavior to feeding. Coe (1943) also described predation on Nereis but made no mention of molluscan prey. Referring to nemerteans in general, Coe suggested that their abundance may fluctuate greatly from year to year. If this is indeed the case, their impact on clam abundance must be recognized as a major influence in the population dynamics of clam stocks. The impact of gastropods and crustaceans as predators of clams on the Oak Point flat was, at least during the pe- riod of study, very limited. While this may in part reflect the situation when clams were generally absent as prey, and can not be used to evaluate the efficiency of these predators relative to C. lacteus on a well populated clam flat, it does suggest that C. lacteus may be more effective than these species in searching out isolated or remote prey. The worm’s apparent ability to home in on and destroy ‘‘the last clam’’ has not previously been recognized. The potential impact of C. lacteus in situations where overfishing or en- vironmental change have severely reduced a clam popula- tion deserves consideration and study by fisheries man- agers. ACKNOWLEDGMENTS The authors wish to thank Drs. K. H. Mann and S. M. Robinson and Mr. D. L. Peer for their constructive reviews of an earlier draft. REFERENCES Angus, R. B., C. M. Hawkins, P. Woo & B. Mullen. 1985. Soft-shell clam survey of the Annapolis Basin, Nova Scotia, 1983. Can. Man. Rep. Fish. Aquat. Sci. No. 1807: viii + 133 p. Coe, W. R. 1943. Biology of the nemerteans of the Atlantic coast of North America. Trans. Conn. Acad. Arts Sci. 35:129-317. Kalin, R. J. 1984. Observations of a feeding method of the Atlantic ribbon worm. Cerebratulus lacteus. Estuaries 7:179—180. McDermott, J. J. 1976. Predation of the razor clam Ensis directus by the nemertean worm Cerebratulus lacteus. Chesapeake Sci. 17:299—301. McDermott, J. J. & P. Roe. 1985. Food, feeding behavior and feeding ecology of nemerteans. Am. Zool. 25:113—125. Newcombe, C. L. 1935. Certain environmental factors of a sand beach in the St. Andrews region, New Brunswick, with a preliminary designa- tion of the intertidal communities. J. Ecol. 23:334—355. Prouse, N. J., T. W. Rowell, P. Woo, J. F. Uthe, R. F. Addison, D. H. Loring, R. T. T. Rantala, M. E. Zinck & D. Peer. 1988. Annapolis Basin soft-shell clam (Mya arenaria) mortality study: a summary of field and laboratory investigations. Can. Man. Rep. Fish. Aquat. Sci. No. 1987: vii + 19 pp. Schneider, D. 1982. Escape response of an infaunal clam Ensis directus Conrad 1843, to a predatory snail, Polinices duplicatus Say 1822. Veliger 24:371-—372. Wilson, C. B. 1900. The habits and early development of Cerebratulus lacteus (Verrill). A contribution to physiological morphology. Q. J. Microsc. Sci. 43:97-198. Journal of Shellfish Research, Vol. 9, No. 2, 299-307, 1990 GROWTH OF NORTHERN QUAHOGS (MERCENARIA MERCENARIA (LINNAEUS, 1758)) FED ON PICOPLANKTON ANN E. BASS,!:4 ROBERT E. MALOUF,? AND SANDRA E. SHUMWAY? 'Marine Sciences Research Center State University of New York Stony Brook, New York, 11794-5000 2New York Sea Grant Institute State University of New York Stony Brook, New York, 11794-5001 3Department of Marine Resources West Boothbay Harbor, Maine, 04575 ABSTRACT The growth of hard clams, Mercenaria mercenaria, feeding on chlorophyte and cyanobacterial picoplankton (<1—4 um in diameter) was investigated to determine if these small algae and cyanobacteria, and indirectly nitrogenous wastes from Long Island, NY duck farms, are responsible for poor growth of M. mercenaria in certain locations of Great South Bay and Moriches Bay, NY. Preliminary experiments verified that hard clams were capable of clearing “‘small forms’’ from suspension. In a six-week growth experiment, clams fed Nannochloris atomus, a common ‘‘small form’’ chlorophyte showed no tissue growth, while clams fed another alga, Pseudoisochrysis paradoxa, known to support growth in bivalve molluscs, grew well. In subsequent experiments, absorption efficiencies of ‘‘small form’’ chlorophytes and cyanobacteria by clams ranged from 17.6% to 31.1%, in contrast to 86.5% for algal species normally used for clam culture. KEY WORDS: Mercenaria mercenaria, quahog, feeding, growth, picoplankton INTRODUCTION The natural phytoplankton population of coastal waters is typically a mixed composition of diatoms, green flagel- lates, and dinoflagellates (Ryther 1954). Size fractionation studies in North American marine ecosystems on the east coast (Yentsch & Ryther 1959, Bruno et al. 1983) and the west coast (Malone 1971) indicate that, whereas nano- plankton (<20 ym) comprise the most abundant size frac- tions in terms of measured chlorophyll a during most of the year, net plankton (>20 1m) may become dominant during seasonal, winter-spring bloom periods (Bruno et al. 1983). Picoplankton cells (0.2—2.0 zm) have been the major com- ponents of many recent blooms from the 1950’s to the present and topic of a recent symposium (Cosper et al. 1989). In addition to blooms, they have been found to be present continually in northeastern coastal waters at con- centrations of 10°—10° cells 1~!, with cyanobacteria (pri- marily Synechococcus) at concentrations of 10*—10® (Har- grave et al. 1989, Tracey et al. 1988). Field studies in the 1950’s of the “‘small form’’ picoplankton in Great South Bay, Long Island, NY (Ryther 1954) and recent field studies in Narragansett Bay, R.I. (Tracey et al. 1988) have shown both beneficial and detrimental effects of pico- plankton on nutrition of bivalves. The studies suggest that 4Current address: Penobscot Valley Council of Governments, One Cum- berland Place, Suite 300, P.O. Box 2579, Bangor, Maine, 04401-8520, USA. picoplankton and its species composition have high poten- tial to influence the nutrition of bivalve molluscs. Ryther (1954) defined the locally named ‘small forms’’ in Great South Bay, Long Island, NY, as “‘small, unicel- lular, green organisms 2—4 w in diameter.’’ Dense blooms of the ‘‘small form’’ chlorophyte species Nannochloris and Stichococcus induced by the flow of duck wastes into the bay, along with salinity and circulation changes, coincided with the failure of the bay’s oyster industry around 1950. Data collected between 1933 and 1950 during summer months show a negative correlation between ‘‘small form’’ densities and oyster meats (Redfield 1951). Although these results are derived from Redfield’s 1952 work, the *‘small form’’ population was as high in 1981 (E. Carpenter, pers. com.) with peaks of ‘‘small form’’ concentrations during the summer months of 10° cells 1~!. Approximately 77% of the phytoplankton biomass in Great South Bay through the year consists of the <10 jm fraction as determined by chlorophyl a concentration (Lively 1981). In 1985 and 1986, there was a “‘brown tide’’ in Peconic Bay, NY. The effect of the small, ““brown tide”’ alga, Au- reococcus anophagefferens, (about 2 jm in diameter) on scallops, scallop larvae, and mussels has been described by Tracey (1988), Bricelj and Kuenstner (1989), Gallager et al. (1989), and Ward and Targett (1989). A. anophagef- ferens is retained with low efficiency by the bivalves’ gills during particle capture (Cosper et al. 1987, Bricelj & Kuenstner 1989), and feeding rates of scallops and mussels are depressed when fed A. anophagefferens at bloom den- 299 300 BASS ET AL. sities (Tracey 1988, Bricelj & Kuenstner 1989). Both bi- valves can, however, absorb the ‘“‘brown tide’’ alga with a maximum efficiency of about 90% (Bricelj & Kuenstner 1989). Oceanic seawater normally has a nitrogen to phosphorus ratio of 8—17:1 (Redfield 1952). Coastal eutrophication can lower the N:P ratio significantly, to values of 4—6:1 (Red- field 1952) and subsequently alter the phytoplankton as- semblage from mixtures containing diatoms such as Nitz- schia closterium, (40 4m, Newell & Newell 1977) to pri- marily the smaller cyanobacteria and chlorophytes (2—4 jm) (Ryther 1954). The low N:P ratio in pollutants cou- pled with the presence of organic nitrogen compounds favor the growth of Nannochloris and Stichococcus over the more typical estuarine phytoplankton (Ryther 1954). In a critique of the dominance of nanoplankton (2—20 jrm) as an indicator of marine pollution, Eppley and Weiler (1979) discussed selective effects of certain pollutants on phyto- plankton species assemblages that suggest the smaller forms, (nanoplankton), may persist where larger-celled and chain-forming phytoplankton have been lost as a result of pollution. In the examples they examined, dominance of nanoplankton appeared to be related to habitat features, food web interactions, and eutrophication rather than selec- tive toxicity of pollutants to larger phytoplankton (Eppley & Weiler 1979). Some ‘‘small-form’’ species, e.g., Chlorella and Nan- nochloris, have been recommended for oyster culture (Dupuy et al. 1977); however, these same species are also considered unacceptable food for many species of bivalve molluscs, including oysters (Redfield 1951, Ryther 1954), and larvae of angel wing clams, American oysters, and hard clams (Tiu et al. 1989). Nannochloris sp. did not sup- port growth in these three larval bivalve species under trop- ical experimental conditions (30°C) (Tiu et al. 1989). Stichococcus sp., one of the “‘small forms’’ in Great South Bay, caused cultures of Venus (=Mercenaria) mer- cenaria larvae to grow more slowly than those in unfed control cultures and Davis and Guillard (1958) suggested that Stichococcus may produce metabolites that are toxic to bivalve larvae. Walne (1973) showed that differences in food supply can alter the protein:carbohydrate ratio in clam tissues and that this ratio is affected by the algal species, the concentrations at which the algae are fed to the clam, and possibly the physiological state of the algal species. The rate of filtration by adult hard clams, Mercenaria mer- cenaria, on ‘small form’’ algae (Nannochloris atomus, 2 ym; Chlorella, 4 4m) is much lower than when the same clam filters water containing a diatom species (Nitzschia sp., 19 pm xX 5 wm; Nitzschia closterium, 43 wm Xx 4 zm) (Rice & Smith 1958). The average filtering rate by the hard clam was higher when diatoms and Nannochloris were in mixed suspensions than when the clam was in unialgal suspensions of Nannochloris (Rice & Smith 1958). Also, Chlorella cells appeared to have an unfavorable effect upon the filtering rate of the clam (Rice & Smith 1958). Thus, although the retention efficiency of cells declines with par- ticle size, these results suggest that there may be more than just size affecting the filtration rate. There is, therefore, some indirect evidence that “‘small forms’’ may inhibit growth of the hard clam but the question has never been rigorously studied. The primary objective of this study was to determine whether or not the ‘‘small form’’ picoplankton sustain growth in juvenile Mercenaria mercenaria. This paper ex- amines the effect of bivalve growth and feeding on Ryther’s “‘small forms’’, specifically the chlorophytes Nannochloris atomus GSB and Stichococcus sp. (now Nannochloropsis salina GSB ~ 17 »m?; Hibberd 1981), two clones: Say 2 and Say 3 (now both Nannochloropsis sp.), and the cyanobacteria Synechococcus bacillarius (Syna), and ASN C-3 (Synechococcus sp.). The hypothesis tested is that unialgal cultures of these particular *‘small forms’’ do not support growth in Mercenaria mercenaria. To test this, three contingent questions were addressed. First, preliminary experiments were run to verify a pre- vious study (Rice & Smith 1958) which indicated that hard clams are capable of clearing *‘small form’’ cells from sus- pension. Additional experiments were designed to answer two questions: can ‘small forms’’ support growth in Mer- cenaria mercenaria and can the clams efficiently absorb the organic material from the “‘small form’’ cell. MATERIALS AND METHODS Algae were cultured using standard methods (Guillard, 1975) at 17—20°C. A. Preliminary Filtration Experiments The objective of these experiments was to determine if the northern quahog, Mercenaria mercenaria, removes ‘small form’’ chlorophyte and cyanobacteria cells from suspension. In each of the three experiments, at least two containers per algal species were examined. One contained approximately 20 juvenile quahog (hard clams) while the other container held no quahogs (control). Experiment I ex- amined hard clam filtration of Nannochloris atomus (GSB), a chlorophyte of about 3 jzm diameter isolated by Ryther in 1952. The experiment was run in triplicate. Experiment II measured the grazing rate by hard clams of three other green ‘‘small form’’ species also about 2—3 jm: Sticho- coccus sp., isolated by Ryther in 1952, (now Nannochlo- ropsis salina GSB; Hibberd 1981) and two clones; Say 2 and Say 3, isolated by Guillard in 1965, (now both Nanno- chloropsis sp.). Experiment III examined hard clam filtra- tion of two species of cyanobacteria: Syna (Synechococcus bacillaris) less than 1 j.m in length and isolated from Long Island Sound by Guillard and Ryther, and ASN C-3 (Syne- chococcus sp.), approximately | jm in length, isolated from Great South Bay by Sarokin (1981). Experiments II and III were not replicated. GROWTH OF QUAHOGS FED PICOPLANKTON 301 The hard clams were approximately 30 mm in length and had been conditioned with a diet of mixed *‘small form’’ algal species for 4—5 days previous to the experi- ment. The quahogs were placed on screens approximately 2 cm off the bottom in aerated basins holding 7-liters of 0.22 ym filtered seawater. The basin and screen assemblies were soaked in filtered seawater before use. The initial cell concentrations in Experiment I were chosen to provide equal particulate organic carbon concen- trations to the quahogs. Pseudoisochrysis paradoxa (Va 12, 4 «.m in equivalent spherical diameter) and GSB Nanno- chloris atomus were sampled during logarithmic growth and analyzed on a CHN analyzer (Hewlett Packard model #185). The carbon in 10° cells - ml~! P. paradoxa is equivalent to that in 1.62 x 10° cells ml~! N. atomus. The ratio for equivalent carbon is 1.00:1.62. The initial cell concentrations in Experiments II and Ill were determined by calculating equal volumes (jm?) of the ‘‘small form’’ species. The cell concentration (cells - ml~!) during logarithmic growth of each algal cul- ture was multiplied by the equivalent spherical volume of the cells. The ratio of equivalent volumes in Exp. II was 0.13 Syn a: 0.58 ASN C-3: 1.00 P. paradoxa and in Exp. Ill the ratio was for all Stichococcus clones 0.35:1.00 P. paradoxa. Initial and final concentrations in each container were determined microscopically with a haemocytometer to esti- mate filtering activity. The experimental temperature was 21 + 3°C; the salinity was 27 + 2%c. B. Growth Experiment The objective of this experiment was to determine if the ‘*small form’’ Nannochloris atomus can support growth in hard clams. Of the available species, N. atomus was se- lected because it was shown by Ryther (1954) to be one of the dominant ‘‘small form’’ types in Great South Bay. Unfed animals and a group fed on the chrysophyte Pseu- doisochrysis paradoxa (Va 12) provided a growth compar- ison to animals fed N. atomus. P. paradoxa has been shown to be a relatively good food for hard clams (Epifanio et al. 1975). The algal combinations used in the growth experiments were as follows: 1) 100% P. paradoxa 0% N. atomus; 2) 75% P. paradoxa 25% N. atomus; 3) 50% P. paradoxa 50% N. atomus; 4) 25% P. paradoxa 715% N. atomus; 5) 0% P. paradoxa 100% N. atomus; and 6) Unfed. There were three replicates of each treatment. The basin and screen assemblies were the same as in the filtration experiments. In addition, the containers were rinsed with a mild chlorine wash and tap water every other day to control bacterial growth. Juvenile hard clams were randomly divided and distrib- uted into 18 growth containers of 100 clams each. Forty randomly selected samples of 10 clams each were frozen to provide an estimate of initial ash-free dry weights. The basins were placed in a circulating water bath at 18—20°C. The 7 L of filtered seawater in the containers were changed daily. Temperature ranged from 19°—21°C. The salinity was 26—27%o. In determining the food concentration needed per con- tainer per day, two boundary conditions were considered. The lower cell concentration boundary was based on the minimum number of cells required per clam for growth. The upper boundary was determined by the cell concentra- tion at which pseudofeces appeared to form. The food concentrations needed per container were based on information from the literature on consumption rates of Pseudoisochrysis paradoxa by juvenile Pacific oysters at temperatures comparable to those of the present study (Malouf & Breese 1978). Crassostrea gigas (mean length 3.2 mm; AFDW ~ 100 yg) filtered between 1 x 103 and 5.5 x 103 cells min~! oyster~! (Malouf & Breese 1978). At 1 xX 103 cells min~! and 100 animals, 100 x 103 cells min~! or 14.4 x 107 cells day~! per 7 L con- tainer are required as a minimum. The lower boundary cell concentration is 2 x 104 cells ml~!. As only one feeding per day was feasible during the growth experiment, a range of food concentrations, all greater than the minimum required for growth, were em- ployed: 3 x 103 to 5 x 10 cells ml~!. The initial cell concentration was calculated to be 6.2 x 10* to 10.3 x 104 cells ml~! basin~! day~!. Although no cell concentration for pseudofeces produc- tion in quahogs, M. mercenaria, has been determined, our preliminary observations show that little or no production occurred at cell concentrations less than 10° cells ml~!. To restrict pseudofeces production, no more than 10° cells - ml~! were administered to each container. These es- timates were based on experiments using P. paradoxa. To supply equivalent organic carbon content to all treatments, the carbon ratio of 1.00 P. paradoxa:1.62 N. atomus deter- mined in Filtration Exp. I was used. For each algal species, the required cell concentration « container~! was divided by the daily count of cultures (cells ml~'), and the food aliquot (mls) was calculated. The experiment was carried out over a period of 37 days. Growth was determined by the change in ash-free dry weight of tissue between initial and final samples in each treatment following standard methods (Gabbott & Walker 1971). The clams were dried at 95°C, weighed on a Cahn 26 automatic electrobalance, combusted in a muffle furnace at 450°C, and reweighed. C. Absorption Efficiency Experiments The objective of these experiments was to determine how efficiently juvenile quahogs absorb organic material from the ‘‘small form’’ algae tested. A dual '4C:°!Cr radio- 302 BASS ET AL. tracer technique (Calow & Fletcher 1972, Wrightman 1975, Cammen 1977, 1980, Lopez & Cheng 1982, 1983) was adapted for estimating absorption efficiencies of sus- pension feeders (Bricelj et al. 1984). The technique com- pared the >'Cr:'4C ratio of the food with that of the feces to calculate the absorption efficiencies (AE). Calow and Fletcher (1972) listed four conditions that must be met before employing the radiotracer technique: 1) that the '4C and 5!Cr be evenly distributed throughout the food material, 2) that '4C and *!Cr move along the gut at similar rates, 3) that *!Cr cannot be absorbed to any great extent, and 4) that the non-absorbed indicator is all present in feces (i.e., not readily leached out). Conditions 1, 3, and 4 were shown to be met under the circumstances employed in this study. Preliminary experiments showed, however, that '4C and *!Cr move along the gut at different rates and therefore that quantitative recovery of feces is necessary to obtain accurate estimates of >'Cr (see Bricelj et al. 1984). Three species of algae and cyanobacteria were used in each experiment: a ‘‘small form’’ chlorophyte, a cyanobac- teria, and Pseudoisochrysis paradoxa (Va 12). The chloro- phyte species used in absorption Experiments I and II were Nannochloris atomus (GSB) and Stichococcus clone Say II, (now Nannochloropsis sp.). Cyanobacterial clones used in Experiments I and II were Syn a and ASN C-3, both Syne- chococcus species. The quahogs, 29-35 mm in length, were conditioned for five days at warmer seawater temperatures (about 15°C) and fed an increased ration of (Pseudoisochrysis para- doxa). The experiments were run at room temperature, ap- proximately 26°C. Twelve clams were initially fed with food suspensions of 5 x 10* to 1 x 10° cells ml~! in 200 ml of 0.22 pm filtered seawater. The labeled algal volumes delivered were determined by cell counts. The labeled algal aliquots were centrifuged twice at about 8000 rpm for 10—15 min. They were rinsed with filtered seawater and brought up to volume to ensure only cell-incorporated '4C was fed to the clams. Algae were resuspended in filtered seawater, and a 5 ml aliquot was removed and filtered onto a 0.6 wm Nu- cleopore filter, which was stored in a glass scintillation vial for later analysis. The conditioned quahogs were then intro- duced and monitored individually, and the time at which they began to filter was recorded. The clams were allowed to feed for 30—45 min, because preliminary work indicated pseudofeces production began after about 45 minutes of ac- tive feeding (Bass, 1983). Seven of the twelve clams were transferred to filtered seawater containing unlabeled algae. Unlabeled algae were added at 3 x 10% to 4 x 10* cells ml~! every 1—2 hr for the first 12 hr, the critical period of time for metabolism of the labeled algae. Preliminary work showed that 82% of the '4C passed through the clam within 12 hr of ingestion. The clams were fed periodically for an- other 36 hr. The clams fed labeled “‘small forms’’ did not filter as readily as those fed labeled Pseudoisochrysis para- doxa. Therefore, clams initially fed labeled ‘‘small form’’ species were fed an equal, unlabeled mixture of their re- spective ‘‘small form’’ species and P. paradoxa. In this mixture, the clams filtered normally. Feces collection began 2 hr after the initial feeding. Samples of the feces were pipetted onto 0.6 4m Nucleopore filters. The filters were stored individually in glass scintillation vials for later analyses. Collection occurred every 4—5 hr for the first 12 hr and less regularly thereafter until 48 hr after the initial feeding. The samples were analyzed for '*C and >'!Cr dpm and the *'Cr to '4C ratio and absorption efficiency were calculated using the following expression (Calow & Fletcher 1972, Bricelj et al. 1984): AE = 100(1 — [dpm *!Cr/dpm '4C (susp.)]/ [dpm *!Cr/dpm !4C (feces)]) (1) The 95% confidence intervals were calculated for the mean absorption efficiency of each algal species (Sokal & Rohlf 1969). A t-test of difference between means (Sokal & Rohlf 1969) was computed for each combination of means of absorption efficiencies for all species used in both experiments. RESULTS Preliminary Filtration Experiments The results of the preliminary filtration experiments in- dicate that the hard clam, M. mercenaria, is capable of re- moving ‘‘small form’’ chlorophyte and cyanobacteria cells from suspension (Table 1). Clams fed a diet of the chlorophyte, Nannochloris atomus, removed 56% of the cells in suspension over 7 h. The clams fed diets of the Stichococcus clones removed a mean value of 61% of the cells from suspension over 5 h. The clams fed diets of cyanobacteria over 5 h removed TABLE 1. FILTRATION EXPERIMENTS. Data Summary. Filtration rates of ‘“‘small form’? cells by 30 mm quahogs at 26° C. Duration of Experiment I: 7 h; Experiment II and III: 5 h each. Controls With Clams Without Clams Initial Final Initial Final ‘small form’’ Conc. Conc. % Conc. Conc. Species (x 104 cells/ml) Removal ( x 104 cells/ml) Nannochloris 25 11 56 23 25 Syna 220 139 27 175 169 ASN C-3 109 22 80 101 96 Mean 53.5 Stichococcus 46 6 87 41 32 Say 2 44 26 41 48 47 Say 3 44 20 55 41 52 Mean 61 GROWTH OF QUAHOGS FED PICOPLANKTON 30 TABLE 2. Results of t-test of the differences between mean weights of M. mercenaria juveniles fed different diets of P. paradoxa and N. atomus, plus a no food control (n = 3) (Sokal & Rohlf 1969). * = p < 0.05; ** = p < 0.01; ***p < 0.001; ns = not significant. Comparison Significance (% P. paradoxa/% N. atomus) ts Level 100/0 xX initial 5.46 ** 100/0 x 0/100 8.09 eX 100/0 x unfed 10.75 ee* 0/100 x initial 1.38 ns 0/100 x unfed 6.01 he initial X unfed 48 ns 75/25 xX initial 4.67 er 100/0 x 75/25 PPG} ns 50/50 x 75/25 .67 ns 75/25 X 25/75 5.34 +t 75/25 x 0/100 9.59 ae 75/25 X unfed 13.23 oor 50/50 x initial 3.96 * 100/0 x 50/50 61 ns 50/50 x 25/75 251) ns 50/50 x 0/100 4.13 + 50/50 x unfed 5.91 +* 25/75 & initial 2.96 * 100/0 x 25/75 55 xx 25/75 x 0/100 6.53 + 25/75 x unfed 11.46 pvt fewer of the smaller Synechococcus species Syn a (<1 fm) from suspension, 27%, than of the larger species ASN C-3 (1 pm), 80%. In comparison with the control treatments of no clams, where the algal concentrations remained approxi- mately constant, between 78% and 127% of the initial con- centrations, results showed that Mercenaria mercenaria is capable of removing ‘‘small form’’ cells from suspension. Only experiment I was run in triplicate, with the means of the replicates given in the summary data table (Table 1). The experiment was analyzed by t-test for statistical differ- ences in the mean concentrations before and after filtration by 30 mm hard clams (n = 3) at 26°C (Sokal & Rohlf 1969). The t, for the clam filtration experiments was 17.06, a very significant correlation (p < 0.01). The t, for the controls (no clams) was 2.24 and not significant. B. Growth Experiment The results of the growth experiments showed that juve- nile hard clams, M. mercenaria, do not grow significantly on a sole diet of Nannochloris atomus (Table 2). The clams fed on intermediate concentrations of the ‘‘small form’’ grew at a slower rate than did those fed the control diet of 100% Pseudoisochrysis paradoxa (Fig. 1). The final mean individual weight of the group of juve- nile hard clams fed the 100% ‘‘small form’”’ diet (1.37 mg) did not differ from that of the clams in the initial samples (1.23 mg), whereas the clams fed the 100%, 75%, and 50% Ww P. paradoxa diet did differ significantly from the initial samples or the unfed control. Absorption Efficiency Experiments The absorption efficiencies of chlorophyte and cyano- bacterial picoplankton by M. mercenaria are relatively low, 17.6% to 31.1%, compared to the absorption efficiencies of P. paradoxa by hard clams, 80.3% and 86.5% (Ta- ble 3). Within groups, the absorption efficiencies did not differ; however, between groups the absorption efficiencies of 2.2 2.2 2.0 2.0 1.8 1.8 L yy F Y GF 1.6 1.6 GJ 1.4 S WY) on S 3 Y y 2 Pe LZ. Wh See INO. eat? Oo e ~ 1.0 = 10 = ~~ ‘ a 3 aE & > 0.8 g 0.8 al ie > 2 06 3 06 o o E Z s 0.4 £ 0.4 !Cr were found to have a significantly different gut passage time when hard clams, M. mercenaria, were fed labeled P. paradoxa, a highly digestible food source (Bricelj et al. 1984). A pos- sible explanation is that the two isotopes follow different pathways through the gut after initial breakdown of cells in the stomach (Bricelj et al. 1984). The *!Cr, absorbed onto the fragmented cell wall which may be less digestible than the cell contents, may be shunted directly to the intestine; whereas the '4C, which is incorporated intracellularly by TABLE 4. Results of t-test of the differences in the mean absorption efficiencies of hard clams, M. mercenaria (28-35 mm in shell length) fed diets of P. paradoxa (Va 12) and several ‘‘small form’’ species: chlorophytes, N. atomus (GSB), Stichococcus clones, Say If and Say II, and cyanobacteria Synechococcus clones, Syn a and ASN C-3, n = 7 (t-test: Sokal and Rohlf, 1979). Symbols as in Table 2. Test (5 Significance Vail2x Vail 1.45 ns GSB x Va 12 7.85 SA ASN C-3 x Va 12 9.46 4% Say 2 x Va 12 5.84 Schaes Syna X Va 12 5.47 << GSB x Va 12 9.35 zx ASN C-3 X Va 12 12.40 see Say 2 x Va 12 6.49 pats Syna X Va 12 6.29 nats Say 2 X Syna 1.09 ns GSB x ASN C-3 1.00 ns Say 2 x GSB itil? ns GSB x Syna 0.13 ns Say 2 x ASN C-3 1.81 ns ASN C-3 x Syna 0.58 ns GROWTH OF QUAHOGS FED PICOPLANKTON the algae, would pass directly into the digestive gland, be- fore egestion or absorption (Bricelj et al. 1984). On the other hand, the ability of the bivalve to digest specific proteins may determine its growth (Walne 1973). A low level of a good food such as /sochrysis galbana pro- duced the same accumulation of carbohydrate and high ni- trogen:glucose ratio in tissue as did all levels of a relatively poor diet (Walne 1973). Walne suggested that the assimila- tion of nitrogen (protein) may regulate growth, whereas there was much variation in the accumulation of carbohy- drate. The proteins differ between species and vary with culture conditions (Walne 1973). Walne (1974) also sug- gested that bivalves may be unable to digest the cyto- plasmic boundaries of some algae. The assimilation of P. suecica by juvenile oysters, Crassostrea virginica has been shown to be relatively low, (A.E. = 6.5%) (Romberger & Epifanio 1980). This sug- gests that the alga was relatively indigestible and inhibited growth (Romberger & Epifanio 1980). The assimilation of a mixed diet of P. suecica and J. galbana by the oysters appeared to be additive while growth appeared to be non- additive. Ingestion of the combined algal diet was higher than ingestion of either algal species singly, so that al- though P. suecica still presented digestive problems, the increased ingestion of the diet apparently resulted in non- additive growth (Romberger & Epifanio 1980). The cell wall of some ‘‘small form’’ species contains sporopellenin, an indigestible, highly resistant, polymer- ized carotenoid present in pollen grains and spores (Faegri & Iverson 1964). A strain of Chlorella, an organism related to Nannochloris (Hargraves et al. 1989), an alga described in Great South Bay and Moriches Bay, a strain of Scene- desmus, and two of three strains of Prototheca contain sporopollenin in their trilaminar wall component (Atkinson et al. 1972). Sporopollenin has been found in strains of Nannochloris (Sarokin 1981). There may be a correlation between the morphological attribute of trilaminar compo- nents and the presence of sporopollenin (Atkinson et al. 1972). The indigestibility and possibly the harmful effects of Chlorella as food may be attributed to the presence of sporopollenin (Schwimmer & Schwimmer 1964). Chlorella cells (clone 211/8) pass through the digestive system of a snail unharmed (Atkinson et al. 1972). Observations during the present study indicate that some of the ‘‘small form’’ species found in Great South Bay and Moriches Bay, NY pass through the digestive system of the hard clam intact. Other occurrences of ‘‘small form’’ species passing through digestive systems intact have been observed in na- ture and in the laboratory (Shumway, unpubl: hed data). In coastal regions, chroococcoid cyanobacteria (genus Syne- chococcus) have been found intact in both the gut and fecal pellets of calanoid copepods, Calanus finmarchicus, without any apparent ultrastructure degradation (Johnson et al. 1983). Cyanobacteria and a Chlorella-like cell have also been 305 found intact in fecal pellets of salps and pteropods and in marine snow (Silver & Bruland 1981, Silver & Alldredge 1981). Some of these small algae contain sporopollenin in their walls and the presence of this inert material is sug- gested as the main reason these algae are resistant to diges- tion (Silver & Bruland 1981). This suggestion is in agree- ment with the hypothesis of the present paper: that sporo- pollenin in the walls of some cyanobacteria and chlorophyte cells prevent the hard clam, M. mercenaria, from efficiently digesting the ‘‘small form’ cells and uti- lizing the absorbed carbon for growth. Several phytoplankton species, besides ‘‘small form”’ species can cause nuisance “‘tides’’. Nuisance bloom species can affect growth in bivalves and bivalve larvae in several ways, as described above. Among these is the “brown tide’? alga, Aureococcus anophagefferens. Re- cently, dense blooms of this 2 zm chrysophyte occurred in Narragansett Bay, Rhode Island (Sieburth et al. 1988) and Long Island embayments causing extensive damage to the scallop industry in Peconic Bay (Bricelj et al. 1987). This alga appears to be deleterious to growth in bivalves and bivalve larvae through chronic toxicity at high cell densities (Tracey 1988, Bricel} & Kuenstner 1989, Gal- lager et al. 1989, Bricelj et al. 1989b, Shumway 1990, Draper et al. 1990). The toxic effects of A. anophagef- ferens on scallop larvae (Gallager et al. 1989) at high cell densities appear to be caused by low capture efficiency of the cells. Results suggest that a cell surface property may interfere with the capture mechanism. In addition, the pres- ence of A. anophagefferens cells in a mixed species me- dium cause the larvae subsequently to reject most cells after capture regardless of nutritional value (Gallager et al. 1989). Draper et al. (1990) suggest that the toxic effects of high cell densities of A. anophagefferens on M. mercenaria and M. edulis are caused by inhibition of the gill cilia re- sulting in cessation of feeding. Small cell size and high density of A. anophagefferens during “‘brown tides’’ may have some detrimental effect on bivalve growth, but it does not appear to be enough to ac- count for the starvation by scallops in the field (Draper et al. 1990). Indigestibility apparently plays no role in the de- leterious effects of the ‘‘brown tide’’ alga on mussels and scallops, as both bivalves can absorb the alga with a max- imum efficiency of about 90% (Bricelj & Kuenstner 1989). The causes of starvation by scallops in the field exposed to the small, brown tide alga, then, are due to cell contents or cell wall composition, not cell size. The ‘‘small form’’ algae in the present study, in con- trast, do not appear to be toxic. The cells (2—4 jum) are captured, with a range of efficiencies (Table 1). The ‘*small form’’ cells can be assimilated to some extent: the absorption efficiencies (AE) of the ‘‘small forms’’ by clams ranged from 17.6%—31.1%. This AE level, does not however support growth in hard clams. Results of this study suggest that hard clams are inhibited from using the 306 carbon in the “‘small form’’ cells for growth. This inhibi- tion could be caused by the indigestible polymerized carot- enoid, sporopollenin, found in the walls of some cyanobac- teria and chlorophyte cells. The implications of these results extend to laboratory rearing of bivalves as well as the growth of the naturally occurring commercial species. Nannochloris and Syno- coccus do not support growth in M. mercenaria and should not be used in laboratory rearing of the bivalve. Further, pollution sources which alter the ambient nutrient concen- trations and N:P ratio and directly or indirectly promote dominance of the faster-growing cyanobacteria and chloro- phytes, should be controlled in areas that support commer- cially important bivalve fisheries. BASS ET AL. ACKNOWLEDGMENTS We thank G. Lopez for his guidance and assistance with radio-tracer experiments and L. Campbell and R. R. L. Guillard for supplying cultures. We also thank V. M. Bri- celj for her continued interest in the research and for criti- cally reading the manuscript and R. R. L. Guillard for thought-provoking discussions. This work is the result of research sponsored by the NOAA National Sea Grant Col- lege Program, U.S. Department of Commerce, under Grant NA81AA-d-00027 to the New York Sea Grant Institute. The U.S. Government is authorized to produce and dis- tribute reprints for governmental purposes notwithstanding any copyright notation that may appear hereon. LITERATURE CITED Atkinson, A. W., Jr., B. E. S. Gunning & P. C. L. John. 1972. Sporo- pollenin in the cell wall of Chlorella and other algae: ultrastructure, chemistry, and incorporation of '*C-Acetate, studied in synchronous cultures. Planta (Berl.) 107:1—32. Bass, A. E. 1983. Growth of the hard clams, Mercenaria mercenaria, feeding on chlorophyte and cyanobacterial picoplankton. M.S. thesis, Marine Sciences Research Center, State University of New York at Stony Brook, Stony Brook, New York, 66 pp. Bayne, B. L. 1983. Physiological ecology of marine molluscan larvae. The Mollusca, Vol. 3. In Wilbur, Karl M. (ed.). Academic Press, New York, pp. 299-343. Bricelj, V. Monica & Susan H. Kuenstner. 1989. Effects of the *‘Brown Tide’’ on the Feeding Physiology and Growth of Bay Scallops and Mussels. In “‘Novel Phytoplankton Blooms: Causes and Impacts of Recurrent Brown Tides and Other Unusual Blooms.’’ Cosper, E. M., V. M. Bricelj & E. J. Carpenter (eds.). Coastal and Estuarine Studies. Springer-Verlag. 1989. pp. 491—S09. Bricelj, V. M., A. E. Bass & G. R. Lopez. 1984. Absorption and gut passage time of microalgae in a suspension feeder: an evaluation of the 51Cr:!4C twin tracer technique. Mar. Ecol. Prog. Ser. 17:57-63. Bricelj, V. M., J. Epp & R. E. Malouf. 1987. Intraspecific variation in reproductive and somatic growth cycles of the bay scallop, Argopecten irradians. Mar. Ecol. Prog. Ser. 36:123—137. Bricelj, V. M., N. S. Fisher, G. B. Guckert & F.-L. Chu. 1989. Lipid composition and nutritional value of the brown tide alga, Aureococcus anophagefferens. In ‘‘Novel Phytoplankton Blooms: Causes and Im- pacts of Recurrent Brown Tides and Other Unusual Blooms.’’ Cosper, E. M., V. M. Bricelj & E. J. Carpenter (eds.). Coastal and Estuarine Studies. Springer-Verlag, Berlin, pp. 85—100. Bruno, S. F., R. D. Staker, G. D. Sharma & J. T. Turner. 1983. Primary productivity and phytoplankton size fraction dominance in a temperate north Atlantic estuary. Estuaries 6:200—211. Calow, P. & C. R. Fletcher. 1972. A new radiotracer technique involving '4C and *!Cr for estimating the assimilation efficiency of aquatic pri- mary producers. Oecologia (Berl.) 9:155—170. Cammen, L. M. 1977. On the use of liquid scintillation counting of *!Cr and '4C in the twin tracer method of measuring assimilation efficiency. Oecologia (Berl.) 30:249—251. Cammen, L. M. 1980. The significance of microbial carbon in the nutri- tion of the deposit feeding polychaete Neireis succi. Mar. Biol. 61:9—20. Claus, G. 1969. Life and pollution in Great South Bay. Underwater Natu- ralist 7(1):11—16. Cosper, E. M., W. C. Dennison, E. J. Carpenter, V. M. Bricelj, J. G. Mitchell, S. H. Kuenstner, D. Colflesh & M. Dewey. 1987. Recurrent and persistent “‘brown tide’’ blooms perturb coastal marine eco- systems. Estuaries 10:284—290. Cosper, E. M., V. M. Bricelj & E. J. Carpenter (eds.). 1989. Novel Phy- toplankton Blooms: Causes and Impacts of Recurrent Brown Tides and Other Unusual Blooms. Coastal and Estuarine Studies. Springer- Verlag, Berlin. 800 pp. Dahl, E., O. Lindahl, E. Paashe & J. Throndsen. 1989. The Crysochro- mulina polylepis bloom in Scandanavian waters during spring 1988. In “‘Novel Phytoplankton Blooms: Causes and Impacts of Recurrent Brown Tides and Other Unusual Blooms.’’ Cosper, E. M., V. M. Bricelj & E. J. Carpenter (eds.). Coastal and Estuarine Studies. Springer-Verlag. 1989. pp. 381-405. Draper,C., L. F. Gainey, S. E. Shumway & L. Shapiro. 1990. Effects of Aureococcus anophagefferens (**Brown Tide’’) on the lateral cilia of 5 species of bivalve molluscs. In **Toxic Marine Phytoplankton.”’ E. Graneli, B. Sundstrom, L. Edler & D. M. Anderson (eds.). Elsevier Science Publishing Co., Inc. pp. 128-131. Davis, H. C. & R. R. L. Guillard. 1958. Relative value of ten genera of micro-organisms as food for oyster and clam larvae. U.S. Fish. Wildl. Serv. Fish. Bull. 58:293—304. Dupuy, J. L., N. T. Windsor & C. E. Sutton. 1977. Manual for design and operation of an oyster seed hatchery for the American oyster, Crassostrea virginica. Special Report No. 142 in Applied Marine Sci- ence and Ocean Engineering of The Virginia Institute of Marine Sci- ence, Cloucester Point, Virginia. June, 1977. 111 pp. Epifanio, C. E. 1976. Culture and bivalve molluscs in recirculating systems, nutrition. In Proceedings First International Conference on Aquaculture Nutrition (K. S. Price, W. N. Shaw & K. S. Danberg, eds.) pp. 173-194. Epifanio, C. E. 1979. Growth in bivalve molluscs: Nutritional effects of two or more species of algae in diets fed to the American oyster Cras- sostrea virginica (Gmelin) and the hard clam Mercenaria mercenaria (L.). Aquaculture 18:1—12. Epifanio, C. E., C. M. Logan & C. Turk. 1975. Culture of six species of bivalves in a re-circulating seawater system. Proc. Euro. Mar. Biol. Symp. 10:97—108. Eppley, R. W. & C. S. Weiler. 1979. The dominance of nanoplankton as an indicator of marine pollution: a critique. Oceanol. Acta. 2(2):241- 245. Faegri, K. & J. Iverson. 1964. Textbook of Pollen Analysis. Copenhagen: Munksgaard. Gabbott, P. A. & A. J. M. Walker. 1971. Changes in the condition index and biochemical content of adult oysters (Ostrea edulis (L.)) main- tained under hatchery conditions. J. Int. Explor. Mer. 34(1):99-106. Gallager, S. M., D. K. Stoecker & V. M. Bricelj. 1989. Effects of the brown tide alga on growth, feeding physiology and locomotory be- havior of scallop larvae (Argopecten irrradians). In ‘‘Novel Phyto- plankton Blooms: Causes and Impacts of Recurrent Brown Tides and Other Unusual Blooms.’’ Cosper, E. M., V. M. Bricelj & E. J. Car- penter (eds.). Coastal and Estuarine Studies. Springer-Verlag. 1989. pp. 510-541. GROWTH OF QUAHOGS FED PICOPLANKTON Graneli, E., P. Carlsson, P. Olsson, B. Sundstr6m, W. Graneli & O. Lindahl. 1989. From anoxia to fish poisoning: The last ten years of phytoplankton blooms in swedish marine waters. In *‘Novel Phyto- plankton Blooms: Causes and Impacts of Recurrent Brown Tides and Other Unusual Blooms.’’ Cosper, E. M., V. M. Bricelj & E. J. Car- penter (eds.). Coastal and Estuarine Studies. Springer-Verlag, 1989. pp. 407-427. Guillard, R. R. L. 1975. Culture of phytoplankton for feeding marine in- vertebrates. In Culture of Marine Invertebrate Animals. Ed. W. L. Smith & M. H. Chanley, Plenum Pub. Corp. NY. pp. 29-59. Hargrave, P. E., R. D. Vaillancourt & G. A. Jolly. 1989. Autotrophic picoplankton in Narragansett Bay and their interaction with micro- plankton. In ‘‘Novel Phytoplankton Blooms: Causes and Impacts of Recurrent Brown Tides and Other Unusual Blooms.’’ Cosper, E. M., V. M. Bricelj & E. J. Carpenter (eds.). Coastal and Estuarine Studies. Springer-Verlag, 1989. pp. 23-38. Hibberd, D. J. 1981. Notes on the taxomony and nomenclature of the algal classes Eustigmatrophyceae and Tribophyceae (synonym Xan- thophyceae). Bot. J. Lin. Soc. 82:93-119. Johnson, P. W., Huai-Shu Xu & J. McN. Sieburth. 1982. The utilization of chroococcoid cyanobacteria by marine phytoplankters but not by calanoid copepods. Annales de L' Institut Oceanographique. 58:297—388. Lackey, J. B. 1951. The rehabilitation of Great South Bay. Consultant. Science Ed., The Blackiston Co., Philadelphia, PA. 25 pp. Langdon, C. J. & M. J. Waldock. 1981. The algal and artificial diets on growth and fatty acid composition of Crassostrea gigas spat. J. Mar. Biol. Ass. U.K. 61:431—448. Lewin, R. A. 1958. The cell wall of Platymonas. J. Gen. Microbiol. 19:87—90. Lively, J. S. 1981. Primary production and abundance of phytoplankton in a barrier island estuary. M.S. Thesis. State University of New York at Stony Brook. 63 pp. Lopez, G. R. & I-J. Cheng. 1982. Ingestion selectivity of sedimentary organic matter by the deposit-feeder Nucula annulata (Bivalva: Nucu- lidae). Mar. Ecol. Prog. Ser. 8:279—282. Lopez, G. R. & I-J Cheng. 1983. Synoptic measurements of ingestion selectivity, and absorption efficiency of natural foods in the deposit- feeding molluscs Nucula annulata (Vivalvia) and Hydrobia totteni (Gastropoda). Mar. Ecol. 2:55-62. Malone, T. C. 1971. The relative importance of nanoplankton and net- plankton as primary producers in the California Current System. Fish. Bull. 69:799—820. Malouf, R. E. & W. P. Breese. 1978. Intensive culture of the Pacific oyster, Crassostrea gigas (Thunberg), in heated effluents. Oregon State Univ. Sea Grant College Program. Publication no. ORESU-T- 78-803. 41 pp. Nielsen, M. V. & Strémgren. 1989. Shell length growth of mussels (My- tilus edulis) as a bioassay for toxic algae. Abstract only. Fourth Inter- national Conference on Toxic Marine Phytoplankton, Lund, Sweden, June 26—30, 1989. Nelson, C. L. & S. E. Siddall. 1988. Effects of an algal bloom isolate on growth and survival of bay scallop (Argopecten irradians) larvae. J. Shellfish Res. 7:683—694. Newell, G. E. & R. C. Newell. 1977. Marine Plankton, A Practical Guide (Sth ed.). Hutchinson of London. 244 pp. Owen, G. 1974. Feeding and digestion in the bivalvia. In O. Lowenstein (ed.). Advances in comparative physiology and biochemistry. V. 5 Academic Press, New York, NY pp 1—35. Redfield, A. C. 1951. Report on a survey of the hydrography of Great South Bay made during the summer of 1950 for the town of Islip, NY. Technical Report from Woods Hole Oceanographic Institution. 30 pp. Unpublished manuscript. Redfield, A. C. 1952. Report to the towns of Brookhaven and Islip, NY on the hydrography of Great South Bay and Moriches Bay. Unpub- lished manuscript. Reference no. 52-26. Oceanographic Institution, Woods Hole, MA. 16 pp. Rice, T. R. & R. J. Smith. 1958. Filtering rates of the hard clam, Venus 307 mercenaria determined with radioactive phytoplankton. Fish. Bull. 58:73-82. Roberts, K., M. Gurney-Smith & G. L. Hills. 1972. Structure, composi- tion, and morphogenesis of the cell wall of Chlamydomanas rein- hardii. 1. Ultrastructure and preliminary chemical analyses. J. Ultra- struct. Res. 40:599-613. Romberger, H. P. & C. E. Epifanio. 1981. Comparative effects of diets consisting of one or two algal species upon assimilation efficiencies and growth of juvenile oysters, Crassostrea virginica (Gmelin). Aqua- culture 25:77—87. Ryther, J. H. 1954. The ecology of phytoplankton blooms in Moriches Bay and Great South Bay, Long Island, New York. Biol. Bull. 106:198—209. Sarokin, D. J. 1981. Ultrastructure and taxomony of the genus Nanno- chloris (Chlorophyceae). M.S. Thesis. State Univ. of NY at Stony Brook, 53 pp. Schwimmer, D. & M. Schwimmer. 1964. Algae and medicine. In Algae and Man. D. F. Jackson, ed. p. 368-412. New York: Plenum Press. Shumway, S. E. 1990. A Review of the Effects of Algal Blooms on Shellfish and Aquaculture. J. World Aquac. Soc. 21:65—104. Sieburth, J. McN., P. W. Johnson & P. E. Hargraves. 1988. Ultrastruc- ture and ecology of Aureococcos anophagefferens gen. et. sp. nov. (Chrysophyceae): the dominant picoplankter during a bloom in Narra- gansett Bay, Rhode Island, Summer 1985. J. Phycol. 24:416—425. Silver, M. W. & A.L. Alldredge. 1981. Bathypelagic marine snow: deep-sea algal and detrital community. J. Mar. Res. 39:501—530. Silver, N. W. & K. W. Bruland. 1981. Differential feeding and fecal pellet composition of salps and pteropods, and the possible origin of deep-water flora and olive-green ‘‘cells’’. Mar. Biol. 62:263—273. Sokal, R. R. & F. J. Rohlf. 1969. Biometry. W. H. Freeman and Com- pany, San Francisco. 776 pp. Tester, P. A., P. K. Fowler & J. T. Turner. 1989. Gulf Stream transport of the toxic red tide dinoflagellate Ptychodiscus brevis from Florida to North Carolina. In ‘‘Novel Phytoplankton Blooms: Causes and Im- pacts of Recurrent Brown Tides and Other Unusual Blooms. Cosper, E. M., V. M. Bricelj & E. J. Carpenter (eds.). Coastal and Estuarine Studies. Springer-Verlag. 1989. pp. 349-358. Tiu, T. A., D. Vaughan, T. Chiles & K. Bird. 1989. Food value of eu- ryptopic microalgae to larvae of Cyrtopleura costata (Linnaeus, 1758), Crassostrea virginica (Gmelin, 1791) and Mercenaria merce- naria (Linnaeus, 1758). J. Shellfish Res. 8:399—40S. Tracey, G. A., P. W. Johnson, R. L. Steele, P. E. Hargraves & J. McN. Sieburth. 1988. A shift in photosynthetic picoplankton composition and its effect on bivalve mollusk nutrition: the 1958 ‘‘brown tide’’ in Narragansett Bay, Rhode Island. J. Shellfish. Res. 7:671—675. Tracey, G. A. 1988. Feeding reduction, reproductive failure, and mor- tality in Mytilus edulis during the 1985 ‘‘brown tide’’ in Narragansett Bay, Rhode Island. Mar. Ecol. Prog. Ser. 50:73-81. Walne, P. R. 1973. Growth rates and nitrogen and carbohydrate contents of juvenile clams, Saxidomus giganteus, fed three species of algae. J. Fish. Res. Board Can. 30:1825—1830. Walne, P. R. 1974. Culture of bivalve molluscs 50 years experience at Conwy. Fishing News (Books) Ltd., Surrey. 137 pp. Ward, J. E. & N. M. Targett. 1989. Are metabolites from the brown tide alga, Aureococcus anophagefferens, deleterious to mussel feeding be- havior? In ‘‘Novel Phytoplankton Blooms: Causes and Impacts of Re- current Brown Tides and Other Unusual Blooms.’’ Cosper, E. M., V. M. Bricelj & E. J. Carpenter (eds.). Coastal and Estuarine Studies. Springer-Verlag. 1989. pp. 543-556. Wrightman, J. A. 1975. An improved technique for measuring assimila- tion efficiency by the *'Cr and '4C twin tracer method. Oecologia (Berl.) 19:273—284. Winter, J. E. 1974. Growth in Mytilus edulis using different types of food. Ber. Dtsch. Wiss. Komm. Meeresforsch. 23:360—375. Yentsch, C. S. & J. H. Ryther. 1959. Relative significance of the net phytoplankton and nanoplankton in the waters of Vineyard Sound. J. Cons. Perm. Int. Explor. Mer. 24:231—238. a - & col hee poe FRY = 0.94 vs r? = 0.79) than TDW-SL relationship. For these reasons and simplicity, TWW-‘SL relationship was selected to calculate B, P and P/B. Net Production 450 570 750 Mortality Standing Crop Biomass Figure 1. Standing crop biomass (above the line) and cumulative biomass lost to mortality (below the line) in g/m? of a representative tray planted at 869 clams/m? in a subtidal location in South Carolina. A sum of the histograms above and below the line represents the production for that time (days) in the grow-out period (see text for further details). CLAM PRODUCTION 311 TABLE la. Results of the analysis of variance to test the influence of tidal locations (intertidal and subtidal) and population densities (290 clams/m?, 869 clams/m?2 and 1,159 clams/m2)on standing crop biomass, production and turnover ratio of tissue wet weight of clams after 3 years’ growth. Degrees of freedom (df) were the same for each analysis, and F = calculated F statistic, P = probability level and NS = not significant. Biomass Production Turnover Ratio Source of 8 Variation df F P F P F P Tidal location 1 19.33 <0.001 15.20 <0.01 3.16 NS Density 2 93.58 <0.001 112.95 <0.001 12.53 <0.001 Interaction 2 5.61 <0.05 $.22 <0.05 Danii NS Experimental Period (3 yr) Average standing crop biomass (B) for all treatments after a 3-year grow-out period was 1,834 g/m? (Table 1b) and approximately 6 times that of a value calculated from the data of Hibbert (1976, 1977b) for a wild population of M. mercenaria in Hambit Spit Estuary, Southhampton, En- gland. A number of factors may account for the vast differ- ence between the B estimates for the two sites. For ex- ample, predation related mortality was considerable at Hambit Spit (Hibbert 1976) and somewhat controlled in the present study (Eldridge et al. 1979). Recruitment of M. mercenaria was characterized as sporadic in Southhampton waters (Mitchell 1974) and the resultant population den- sities would be expected to be less uniform and dense as in this research aquaculture situation. Also, Hambit Spit pop- ulation structure of clams consisted of several cohorts (Hibbert 1976, 1977b) whereas only one cohort was planted in trays (Eldridge et al. 1979). Another factor re- sponsible for the B difference between the two sites was that clam growth occurred throughout the year in South Carolina (Eldridge et al. 1979) compared to only 6 months in England (Hibbert 1977a). Trays planted at the higher densities (1,159 and 869 clams/m*) had significantly more B than trays planted at 290 clams/m? (Table la, b). It is important to note that the average growth rates were significantly slower in the higher density trays (Eldridge et al. 1979). When the B estimates of tidal locations (with equal densities) were compared, significantly more B was observed in the subtidal location (Table la, b). Although not significant (P = 0.585), growth was faster in the subtidal trays (Eldridge et al. 1979). Production (P) was significantly greater at the higher densities and in the subtidal location (Table la, b). These treatments also had significantly higher survival than clams planted in the lowest density and intertidal location (El- dridge et al. 1979). Production estimates are a reflection of both growth and survival differences among the treatments. Growth and survival of clams were better in the subtidal trays perhaps because environmental conditions were less harsh, and clams had more time to feed than clams in the intertidal location (Belding 1912, Eldridge et al. 1979, Eversole et al. 1984). Turnover (P/B) is the number of times the clams add, in tissue growth, during a specified time period (e.g., life span, culture period) the equivalent of the B. Robertson (1979) observed a relationship for marine bivalves (logy P/B = 0.62065 — 0.78261 logy, L; n = 19, r? = 0.79) between annual P/B and life span (L). In our study, the life span consisted of a 3-year grow-out period and a 6-month hatchery-nursery phase for a total of 3.5 years. Using a 3.5-year life span and Robertson’s (1979) relationship for marine bivalves, we arrived at a calculated P/B value of 1.41. The observed overall average P/B from our study of 2.57 (Table 1b) is relatively high in comparison to the cal- culated P/B value and most values reported for natural pop- ulations of marine bivalves, including estimates for M. mercenaria (Table 2). The observed P/B value approxi- mates those values for species with 1—2 year life spans (Table 2). As would be expected, P/B values for intensely- managed environments, such as aquaculture systems, would be higher than that observed for natural populations. Walker and Tenore (1984) recorded population den- sities, biomass and production for natural populations of TABLE Ib. Mean standing crop biomass (B in g/m?), production (P in g/m?/day) and turnover ratio (P/B) of tissue wet weight of clams grown at different tidal locations and population densities after 3 years. Significant differences (P < 0.05) in mean B, P and P/B among main effects (tidal locations and densities) are indicated with alphabetical superscripts and interaction terms with numerical superscripts. 290 Clams/m? 869 Clams/m? Location B P P/B B P Intertidal 459! 0.46! 3.12 2,180? 2.252 Subtidal 787! 0.78! 2.92 2,375? 2.332 Mean 6234 0.628 3.028 2,277° 2.29» Dei 2.66 2.61 1,159 Clams/m? Mean P/B B P P/B B P P/B 1,9892 PMNS 2.43 1,542 1.623 Path | BP2i182 3.183 1.70 2125S” ALI? 2.42 2,601> 2.66° 2.07° 1,834 1.86 2.57 312 EVERSOLE ET AL. M. mercenaria in three intertidal areas in Wassaw Sound, Georgia. The P/B ratios were 0.23, 0.14 and 0.05 for popu- lations with 12, 18 and 49 clams/m?, respectively (Walker and Tenore 1984). A similar density-dependent relationship was observed in our study, mean P/B ratios were 3.02, 2.61 and 2.07 for the trays planted at 290, 869 and 1,159 clams/m?, respectively (Table 1b). Also the variation in P/B among density treatments was significant (Table la). Waters (1977) suggested that overcrowded populations ex- hibit lower P/B ratios, principally because growth is slower and a longer life span results. Walker and Tenore (1984) noticed that the population with the highest density was dominated (~64%) by old clams (>7 years) whereas the lowest density population contained only 12% old clams. Walker and Tenore (1984) attributed differences in P/B among the populations to the differing age structures. In our case, the age structure was the same among treatments and the lower P/B was probably more a function of slower growth (Eldridge et al. 1979) and the greater B in the high density trays. This greater B resulted jointly from a stocking rate 3 times that of the lowest density trays and a significantly higher survival rate (Eldridge et al. 1979). Recruitment probably plays an important role, similar to stocking density, in determining P/B values in natural pop- ulations (Robertson 1979). Calculated P/B ratio for clams in the subtidal trays was lower than the intertidal clams, but this difference was not significant (Table la, b). Apparently, the difference in tidal exposure between the locations was insufficient to affect a significant change in P/B. Wolff and de Wolf (1977) also observed small decrease in the P/B ratios of three mollusc species from the high down to the low water mark of the intertidal zone. Unfortunately, their data represent popula- tions from widely different sampling locations in Greve- lingen Estuary, Netherlands. A clarification of the relation- ship between P/B and tidal exposure will require empirical evidence from a wider range of tidal exposures than that used in the current study. Market Size Clams planted at 290/m? reached market size (45 mm SL) in 649 days, approximately 100 and 285 days faster than those clams at 869 and 1,159/m?, respectively (Table 3b) Clams in the intertidal location required about 40 addi- tional days to achieve market size. Adding a 6-month hatchery-nursery phase to these estimates to market size, commercial-sized clams would be available in 2.0—3.0 years. Seed clams (6 mm SL) planted at maintained den- sities of 509 and 1,009/m? reached 44—45 mm SL in ap- proximately 2 years in Georgia (Walker 1984). Although there are differences between the studies, clams (6-13 mm SL) planted at 500—1,000/m? consistently reached market TABLE 2. Estimates of annual tissue production (dry weight) and annual P/B ratios for selected bivalve species.? Production Life Span Taxa (gm/m2/yr) (years) P/B Location Mercenaria mercenaria 94.37? 355 2.57 South Carolina, USA M. mercenaria 6.043 8-9 0.14 Georgia, USA M. mercenaria 8.96 7-9 0.32 Southhampton, England Chione cancellata 17.8 6+ 0.83 Florida, USA Venerupis aurea 1.08 5 1.11 Southhampton, England V. pullastra 22.22 10 0.15 Western Norway Mya arenaria 2.66 8+ 0.48 Comwall, England M. arenaria 11.60 2+ 2.54 Peteswick Inlet, Nova Scotia Macoma balthica 1.93 Disks 1.53 Peteswick Inlet, Nova Scotia M. balthica 10.07 5-6 2.07 Ythan Estuary, Scotland M. balthica 0.31 Ur 0.91 Cormwall, England M. balthica 0.99 8-10 0.81 Grevelingen Estuary, Netherlands Dosinia elegans 239 2 2.81 Florida, USA Theora lubrica 4.30 1+ 4.08 Island Sea, Japan Verenolpa micra So] IES 3.20 Island Sea, Japan Nucula paulula 1.05 1 3.51 Island Sea, Japan Crassostrea virginica 826.40 2 2.01 South Carolina, USA ! Condensed from Robertson (1979) and references within. 2 Calculated using 13.9% TWW equals TDW. The life span was assumed to be the culture period of 3.5 years. 3 Calculated from the data of Walker and Tenore (1984) using 0.9 AFDW equals TDW (Waters 1977). CLAM PRODUCTION 31 Ww TABLE 3a. Results of the analysis of variance to test the influence of tidal locations (intertidal and subtidal) and population densities (290 clams/m?, 869 clams/m?2 and 1,159 clams/m?) on the time to harvest, standing crop biomass and production of tissue wet weight when clams reached market size (average size >45 mm SL). Degrees of freedom (df) were the same for each analysis, and F = calculated F statistic, P = probability level and NS = not significant. Time to Harvest Biomass Production Source of SS so Variation df F P F P F P Tidal location | 0.76 NS 15.65 <0.01 15.31 <0.01 Density 2 15.48 <0.001 170.75 <0.001 111.52 <0.001 Interaction 2 0.37 NS 7.64 <0.01 7.19 <0.01 size in under 3 years in both locations. According to Walker (1989), less dense natural populations of clams also routinely reach marketable size in 2—3 years in Georgia. Three years would appear to be a conservative but reason- able estimate in evaluating the economical feasibility of hard clam operations in the southeastern United States. After reaching market size, clams in the subtidal loca- tion had significantly greater B and P values than clams grown in the intertidal location (Table 3a, b). Clams planted at 1,159/m? yielded significantly more B than trays at 290 clams/m?. Assuming similar conditions and produc- tion, clams at 290/m* would require nearly 18 years to yield the same B that clams planted at 1,159/m? produced in approximately 3.5 years. Obviously, the highest density (1,159/m?) was the most productive treatment in this study, and it is quite possible that a higher density treatment could have been cultured without a negative impact at this site. Aquaculture Implications Stocking rates for M. mercenaria are site specific, so there is need to evaluate planting densities at proposed cul- ture sites. It is also our contention that energetic measures such as B and P allow for a sounder ecological evaluation of stocking density than the more traditional methods (e.g., growth and time to market). Case in point, Eldridge et al. (1979) suggested a stocking density of 300 clams/m? based on the observation of density-dependent growth among the densities (290, 869 and 1,159 clams/m?) planted at Clark Sound. Evidence from our B and P estimates indicate that an optimum density was not reached with these test den- sities and that a stocking density many times that rate of 300 clams/m? could have been recommended for this site. Once the appropriate ecological/biological consider- ations of stocking density have been determined for a par- ticular site, an aquaculturist should be able to create an op- timum production scenario. Linear programming tech- niques could be employed to obtain optimal mixes of planting densities to maximize production schedules and market potentials. For instance, if market prices varied sea- sonally, production could be manipulated by altering planting densities and/or planting times to produce crops in the season which has the highest price. The aquaculturist should be able to take advantage of stocking density infor- mation to help manage the problems facing clam operations such as cash flow, seed availability and market constraints. ACKNOWLEDGMENTS The authors thank all those fellows for helping with field work and laboratory analysis (C. A. Aas, R. S. Bisker, W. K. Michener, G. Steele, S. A. Summer, W. Waltz and J. M. Whetstone). Dr. L. W. Grimes assisted with statis- tical analysis. Thanks are also due to J. Richardson for dealing with many revisions. Finally, we are indebted to Dr. P. B. Heffernan, Dr. J. J. Manzi and Mr. R. L. Walker for critically reading an earlier draft of this manu- script. Financial support was provided by S.C. Agricultural Experiment Station, S.C. Wildlife and Marine Resources Department and S.C. Sea Grant Consortium. TABLE 3b. Mean time (days), standing crop biomass (B in g/m?) and production (P in g/m?/day) of tissue wet weight of clams grown at different tidal locations and population densities to market size. Market size was set at an average size of >45 mm SL. Significant differences (P < 0.05) in mean time to harvest, B and P among main effects (tidal locations and densities) are indicated with alphabetical superscripts and interaction terms with numerical superscripts. 290 Clams/m? 869 Clams/m? 1,159 Clams/m? Mean Location Time B |e Time B P Time B P Time B P Intertidal 660. 259! 0.46! 750 1,950? 2.832 980 2,133? 2.38 797 1,447 1.898 Subtidal 638 410! 0.81! 750 2,0492 2.852 890 3,1943 3.843 759 1,884> 2.50° Mean 6492 335? 0.63? 750? 2,000° 2.84> 935° 2,664° Seu 7718 1,666 2.19 314 EVERSOLE ET AL. LITERATURE CITED Belding, D. L. 1912. The quahaug fishery of Massachusetts. Commission of Massachusetts Department of Conservation, Marine Fisheries Series 2:1—41. Eldridge, P. J. & A. G. Eversole. 1982. Compensatory growth and mor- tality of the hard clam, Mercenaria mercenaria (Linnaeus, 1758). Ve- liger 24:276-278. Eldridge, P. J., A. G. 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. Proceedings of National Shellfisheries Associ- ation 69:30—39. Eversole, A. G., W. K. Michener & P. J. Eldridge. 1984. Gonadal con- dition of hard clams in a South Carolina estuary. Proceedings of An- nual Conference Southeastern Association of Fish and Wildlife Agencies 38:495—505. Hibbert, C. J. 1976. Biomass and production of a bivalve community of an intertidal mud-flat. Journal of Experimental Marine Biology and Ecology 25:249-261. Hibbert, C. J. 1977a. Growth and survivorship in a tidal-flat population of the bivalve Mercenaria mercenaria from Southhampton water. Marine Biology 44:71—76. Hibbert,C. J. 1977b. Energy relations of the bivalve Mercenaria mercen- aria on an intertidal mudflat. Marine Biology 44:77-84. Mitchell, R. 1974. Studies on the population dynamics and some aspects of the biology of Mercenaria mercenaria (Linné). Ph.D. Dissertation, University of Southhampton, England. Robertson, A. I. 1979. The relationship between annual production:bio- mass ratios and life spans for marine macrobenthos. Oecologia 38: 193-202. Russell-Hunter, W. D. 1970. Aquatic productivity. MacMillan Co., New York, USA. SAS Institute Inc. 1985. SAS User Guide: Statistics, Version 5 Edition. SAS Institute Inc., Cary, North Carolina. Walker, R. L. 1984a. Effects of density and sampling time on the growth of the hard clam, Mercenaria mercenaria, planted in predator-free cages in coastal Georgia. Nautilus 98:114—119. Walker, R. L. 1984b. Population dynamics of the hard clam, Mercenaria mercenaria (Linné), and its relation to the Georgia hard clam fishery. M.Sc. Thesis, Georgia Institute of Technology, Atlanta. Walker, R. L. 1989. Exploited and unexploited hard clam, Mercenaria mercenaria (L.), populations in coastal Georgia. Contributions in Ma- rine Science 31:61—75. Walker, R. L. & K. R. Tenore. 1984. The distribution and production of the hard clam, Mercenaria mercenaria, in Wassaw Sound, Georgia. Estuaries 7:19—27. Waters, J. F. 1977. Secondary production in inland waters. Advances in Ecological Research 10:91—164. Wolff, W. J. & L. de Wolf. 1977. Biomass and production of zoobenthos in the Grevelingen, The Netherlands. Estuarine and Coastal Marine Science 5:1—24. 2, 315-321, 1990. Journal of Shellfish Research, Vol. 9, No. ESTIMATES OF LOSSES ASSOCIATED WITH FIELD DEPURATION (RELAYING) OF MERCENARIA SPP. IN THE INDIAN RIVER LAGOON, FLORIDA DAN C. MARELLI AND WILLIAM S. ARNOLD Florida Marine Research Institute Department of Natural Resources 100 8th Avenue SE St. Petersburg, FL 33701-5095 ABSTRACT Experiments were performed during June 1987 and February 1988 to test the effects of field depuration, or “‘relaying,”’ on the mortality rates of Mercenaria spp. in the Indian River lagoon, Florida (USA). Mortality rates of clams in open 1.0 m? plots were compared with those of clams in plots protected by predator-exclusion cages over periods of two and four weeks. An additional experiment examined the effect of fences on rates of clam loss and demonstrated that loss from open plots could be attributed to predation. Mortality due to predation was significantly higher in open plots, and brachyuran crabs appeared to be an important cause of this mortality. Mortality was not significantly different between summer and winter, and was related to the elapsed time between the initiation of relaying and reharvest. Because of mortality and reharvest losses in harvest efficiency, relaying as practiced in the Indian River lagoon may be of minimal economic benefit to clam fishermen. KEY WORDS: Mercenaria, hard clam, relaying, mortality, predation INTRODUCTION Field depuration is commonly practiced by clammers and oystermen in the United States. Field depuration in- volves gathering bivalves from waters conditionally ap- proved, restricted, or prohibited for shellfishing and moving them to approved waters for a minimum of 14 days (USPHS 1965) to allow the bivalves to purge their gut of enteric bacteria. This process is generally termed ‘‘re- laying,’’ although relaying may also mean moving bivalves between open areas to promote growth. Relayed clams are generally held on submerged bottom lands leased to private shellfish operations. Alternatively, clams may be moved ashore to an approved closed-cycle depuration plant. Re- laying of hard clams, Mercenaria spp.,' is practiced in the Indian River lagoon (Fig. 1), on the east central coast of Florida, under permit from the Florida Department of Nat- ural Resources. In recent years, a sudden and unanticipated increase in Mercenaria landings from the Indian River lagoon has pro- vided a substantial commercial supply of hard clams. Al- most 46% of the commercially exploitable area of the la- goon is composed of waters unapproved for harvesting, and depuration has allowed for the harvest of many clams from these areas. However, because relaying involves distur- bance and concentration of the clams, losses due to stress and predation during the relay process may occur. The present study was designed to estimate the short-term sur- vivorship of relayed clams. These estimates are then used in a bioeconomic model to determine the most economical The Indian River lagoon contains two sympatric species of Mercenaria, M. mercenaria and M. campechiensis, plus hybrids (Dillon & Manzi 1989). These are very difficult to separate morphologically, and we have chosen to use the generic name Mercenaria herein. 315 method for independent Indian River clammers to deliver clams to seafood wholesalers. METHODS The study site, a shellfish lease north of Grant, Florida (Fig. 1), had a depth of approximately 1.5 m, a muddy sand substrate, and no submerged aquatic vegetation. The site is representative of benthic habitats in this region of the Indian River lagoon (Arnold and Marelli, unpubl. data). Experiments were conducted during June 1987 (summer) and February 1988 (winter). Each experiment involved twelve 1.0 m? plots; six plots were open, marked only with stakes and twine, and six plots were enclosed by predator- exclusion cages. Each cage consisted of a 1.0-m x 1.0-m x 0.5-m aluminum frame covered with I-cm x l-cm polyethylene mesh. The bottom edges of the cages pene- trated the substrate to a depth of 1—2 cm. Clams in the size range 50 to 75 mm shell length (SL = maximum anterior posterior distance), obtained from commercial clammers, were labeled by painting an orange spot on one valve, and were held overnight? out of water. One hundred clams were placed in each plot on the following day and allowed to rebury themselves. This density was determined by per- sonal observation to be equal to or lower than densities used in Indian River relay operations. Three caged plots and three open plots were harvested two weeks after relaying, and the remaining plots were re- harvested after four weeks. Harvest consisted of hand- raking as many clams as possible from each plot and then clearing the plot with a venturi-driven suction dredge to ?Clams being relayed generally are not held out of water overnight. Since Mercenaria spp. can survive in air for weeks, however, it was reasoned that 12 hours out of water would not significantly increase mortality. 316 MARELLI AND ARNOLD An additional experiment was performed in March and April, 1989 to determine the effect of lateral movement on 81°W iA \ f ; the unexplained loss of clams (number missing) from open \ TITUSVILLE plots. This experiment used 3 open plots, as described pre- viously, and 3 fenced plots. Fences were constructed of 15 cm high, | m long strips of polyethylene mesh supported by pressure-treated stakes and pressed 4 to 5 cm into the AEA substrate. Ninety-five marked clams were placed in each OE plot on March 2 and were reharvested on April 6. Data from this experiment were analyzed using an un- balanced one-way ANOVA (SAS Institute, Cary, NC, GLM procedure) since cell sizes were unequal. Differences in numbers of missing clams among treatments were tested for significance using a Student-Newman-Keuls mean pro- cedure. INDIAN RIVER INDIAN RIVER MELBOURNE " RESULTS STUDY SITE Data from the experiment estimating the effect of fencing on the percentage of missing clams (Table 1) were compared with data collected during the winter of 1988 on the percentage of clams missing from 4-week caged plots. Plot condition (caged, fenced, or open) had a significant effect on clam losses (Table 2). Fenced and open plots were statistically indistinguishable with respect to the percentage of missing clams, and many fewer clams were recorded as missing from caged plots (Table 3). Following the proce- dure of Underwood (1981), which was based on Winer (1971), we calculated the probability of a type II error at 0.09, making it highly unlikely that a lack of significant difference between the cage and fence treatments was due collect any clams missed by raking. All labeled clams that to a lack of replication. These results demonstrate that were recovered were counted and recorded. Shells of dead missing clams can be considered to be mortalities. clams were examined for indications of causes of mortality In the comparisons of open and caged plots, mortality (Magalhaes 1948, Carriker 1951, Landers 1954, Paine was similar in summer and winter experiments (Fig. 2). 1962, Vermeij 1978, Peterson 1982). Mortality was calcu- Mean mortality was greater in open plots than in caged lated as follows: [number dead + number relayed] x 100. plots during both summer and winter experiments (Tables 4 Total loss was calculated as follows: [(number dead + and 5). Examination of the shells of confirmed mortalities number missing) + number relayed] < 100. The results of indicated that brachyuran crabs were an important cause of this experiment were analyzed using a three-way com- mortality in both seasons. No evidence of predation by pletely randomized analysis of variance (ANOVA) on un- whelks was observed, and 2.3% of clams in all treatments transformed data. died from unspecified causes. Figure 1. The Indian River lagoon. The clam fishery is located prin- cipally between Titusville and Sebastian. TABLE 1. Survivorship, percentage missing, and mortality caused by crabs on Mercenaria spp. transplanted to open and fenced plots and harvested after 5 weeks, March-April 1989. Mortality represents confirmed deaths, and is subdivided by specific cause: crab mortality or undetermined reasons. All values are percentages. Plot Crab Other Condition Replicate Survivorship Mortality % Missing Mortality Mortality Open 1 78.9 1.1 20.0 ed 0 2 86.3 1.1 12.6 ile 0 3 78.9 1.1 20.0 1.1 0 Fenced 1 78.9 0 21.1 0 0 2 86.3 3:2 10.5 2.1 1.1 3 86.3 1.1 12.6 0 1.1 MERCENARIA FIELD DEPURATION LOSSES 317 TABLE 2. Analysis of variance for effects of fencing, caging, and open plots on percentage of Mercenaria spp. lost following transplantation and relaying. Source df SS MS F p-value Plot Condition 2 485.73 242.87 15.21 p < 0.005 Residual 9 143.74 15.97 Total 11 629.47 Analysis of the results of the experiment examining mortality in open and caged plots indicated that predator- exclusion cages significantly reduced mortality, and that mortality rates of Mercenaria individuals were not signifi- cantly different either between seasons or across harvest times (Table 6). DISCUSSION The results of this study establish that relayed hard clams suffered losses of 14% in unprotected plots during the course of this experiment. Clams located within pred- ator-exclusion cages experienced reduced mortality, aver- aging 3.5% over both the two and four week time periods. The significantly higher mortality in open plots can be di- rectly related to predation. Lateral movement by clams from open plots is apparently not responsible for the occa- sionally large number of missing individuals, because losses of clams from fenced control plots were similar to losses from open plots. Clams that were missing may have been missed during sampling or may have been carried away by predators. Brachyuran crabs often move clams to their shelters for subsequent consumption (Boulding & Labarbera 1986), scavengers may carry clam remains away (Peterson 1982), and black drum swallow clams whole, crushing them with pharyngeal teeth (Simmons & Breuer 1962). The results indicate that missing clams almost cer- tainly represent mortalities and that even though the experi- mental plots could potentially experience a large edge ef- fect, manifested as emigration, this was not seen. If a large edge effect had been identified, our experiments would over-estimate mortality when compared with large com- mercial leases. The fencing study indicates that predators, rather than emigration, are responsible for missing clams. TABLE 3. Student-Newman-Keuls procedure for the effect of fencing and caging condition on percentage of Mercenaria lost following transplantation and relaying. Means with the same grouping are not significantly different. Fence Condition Mean % Missing Grouping Open 13.85 A Fenced 12.63 A Caged 1.83 B MORTALITY (%) WINTER 4 WKS SUMMER WINTER 2 WKS SUMMER Figure 2. Mortality of Mercenaria spp. transplanted to caged plots and open plots during summer 1987 and winter 1988 and harvested after 2-week and 4-week intervals. Plots show means, ranges, and + standard error (n = 3 for all treatments). Although limited predation also occurred within the pred- ator-exclusion plots, losses were minimal when compared with total mortality in open plots, indicating that the exclu- sion devices were effective barriers to predation. Considerable information is available concerning losses of juvenile clams in field culture plots, but little informa- tion exists on mortality of clams of the sizes used in our experiments. For example, during a ten-month period in Alligator Harbor, Florida, Menzel et al. (1976) reported losses of 42.3% for clams in caged plots and 100% for clams in open plots for 7-10 mm SL Mercenaria spp. In an earlier seven-month study, Menzel and Sims (1962) re- ported mortalities of S—18% in both open and caged plots containing clams 33—44 mm SL. These losses were attrib- uted to predation by blue crabs, Callinectes sapidus, and whelks Busycon sinistrum (=B. contrarium) (Menzel & Sims 1962). The results of the present study indicate that loss rates of clams in the size range SO—75 mm SL are much lower than those observed for clams in smaller size classes. This agrees with previous observations, such as that of Arnold (1983), that predation rate of blue crabs gen- erally decreases with increasing clam size. 318 MARELLI AND ARNOLD TABLE 4. Survivorship and associated sources of mortality among replicates of 100 Mercenaria spp. transplanted to caged and open plots, June 1987. Total mortality is divided into ‘‘Confirmed Dead”’ and ‘‘Missing’’ categories. Confirmed deaths are subdivided by specific cause: crab predation or undetermined reasons. All values are percentages. Harvest Cage Replicate Crab Undet. Total Mean Time Condition No. Survivorship Dead Mortality Mortality Missing Mortality Mortality 2 Wk Open 1 90 5 2 3 5 10 10.67 2 94 3 0 3 3 6 3 84 8 l 7 8 16 2 Wk Caged 1 97 2 2 0 1 3 2.33 2 97 3 I 2 0 3 3 99 1 0 1 0 1 4 Wk Open 1 88 4 2 2 8 12 13.67 2 82 8 0 8 10 18 3 89 3 1 2 8 11 4 Wk Caged 1 98 2 1 1 0 2 5.00 2 93 5 1 4 2 7 3 94 2 1 1 4 6 Our experiments were performed during both summer and winter to examine the seasonal variation in mortality due to the stress of relaying and to the activity of predators. These data do not indicate seasonal differences in the re- sponse time of predators (as judged by two-week mortality rates). Indeterminate mortality, however, which may be stress related, appears to be higher in summer. Symptoms of stress, including reduced growth, reduced reproductive activity and increased incidence of neoplasia, have been re- ported during the summer for hard clams from the Indian River lagoon (Arnold & Marelli, unpubl. data) (Hesselman et al. 1988). Relay mortality is an important consideration in deter- mining the relative economic efficiency of field- versus shore-based depuration. Clam losses during relaying are in- evitable and acceptable as long as the losses, combined with the other associated costs of relaying (Holmsen & Stanislao 1966), do not outweigh the savings that would be realized by avoiding closed-cycle depuration. Estimates of the costs associated with relaying were obtained from five shellfish wholesalers and, using a modification of Holmsen and Stanislao’s equation quantifying the cost of relaying, a range of estimates was produced for a variety of conditions (see Appendix). Under the most favorable conditions, re- laying is economical only when reharvesting is very effi- cient (10% or less additional labor required to reharvest re- layed clams), loss rates caused by relaying are very low (5% or less), and the dockside price of clams is low. Re- TABLE 5. Survivorship and associated sources of mortality among replicates of 100 Mercenaria spp. transplanted to caged and open plots, February 1987. Total mortality is divided into ‘‘Confirmed Dead”’ and ‘‘Missing’’ categories. Confirmed deaths are subdivided by specific cause: crab predation or undetermined reasons. All values are percentages. Harvest Cage Replicate Time Condition No. Survivorship Dead 2 Wk Open 1 90 I 2 87 0 3 86 3 2 Wk Caged 1 97 2 2 93 1 3 96 l 4 Wk Open 1 77 8 2 83 7 3 82 1 4 Wk Caged 1 97 3 2 97 2 3 100 0 Crab Undet. Total Mean Mortality Mortality Missing Mortality Mortality 0 4 10 12.33 0 0 13 13 2 1 11 14 0 2 1 3 4.67 1 0 FT 0 1 3 4 4 4 15 23 19.33 2 5 10 17 1 0 17 18 1 2 0 3 2.00 1 1 1 3 0 0 (0) 0 MERCENARIA FIELD DEPURATION LOSSES 319 TABLE 6. Analysis of variance for effects of cage conditions, time of harvest, and season on mortality of transplanted Mercenaria spp. Source df SS MS F p-value Cage 1 661.50 661.50 75.60 p< 0.0001 Harvest Time 1 37.50 37.50 4.29 p< 0.055 Season 1 16.67 16.67 1.90 p< 0.187 Cage X Harvest Time 1 37.50 37.50 4.29 p<0.055 Cage X Season 1 24.00 24.00 2.74 p<0.117 Harvest Time X Season | 0.67 0.67 0.08 p< 0.786 Cage X Harvest Time x Season 1 32/67) 932567 93573, pi 0!071 Residual 16 140.00 8.75 Total 23 950.50 laying, as practiced in the Indian River lagoon, involves a team of 8 to 10 clammers moving clams to a shellfish lease, where they are simply dumped onto the lease bottom. After a minimum of 15 days the clams are reharvested by raking. With this method, neither efficiency level nor loss rate can be strictly controlled, although some shellfish dealers at- tempt to reduce losses by placing predator-exclusion mesh over the clams. The use of containerized relaying systems (e.g., Supan & Cake, 1982) would increase reharvest effi- ciency and decrease losses, but there are at least two problems associated with implementing containerized re- laying in the Indian River lagoon. First, the additional cost of using predator-exclusion devices would initially increase the cost of field-based depuration relative to the shore- based alternative. Second, Florida law currently prohibits the use of any structure which extends more than 6” above the bottom, which would make it very difficult to use con- tainerized relay systems. Despite the legal aspects, if this technology can be demonstrated to be both efficient and compatible with use plans for the Indian River lagoon, then serious consideration should be given to using some form of containerized relay system. One example of this tech- nology is currently under development by the Harbor Branch Oceanographic Institution, Division of Applied Bi- ology. In addition to the added cost of relaying, leaseholders must consider the costs of maintaining their leases; these costs include lease rental and the maintenance of lease markers. Other factors that argue against relaying are the two-week delay between the clam harvest and the receipt of payment, and the possibility of additional delays when rainfall events cause temporary closure of the lagoon area in which the lease is located. Additional delays not only reduce cash flow to the clammers but, according to our re- search, also increase clam mortality. Concentrations of unprotected hard clams in the Indian River lagoon are strongly and negatively affected by large, mobile predators such as Callinectes spp.,> Menippe mer- cenaria, and the black drum Pogonias cromis. Because of losses from predation and other costs associated with re- laying, this practice in the Indian River lagoon is currently economically advantageous only under a very restricted set of circumstances. These circumstances will rarely, if ever, be achieved in the lagoon, and additional complications are posed by fluctuating and generally declining water quality. Predator exclusion containers may make relaying more fea- sible, but costs of such containers must be balanced against any economic gain. Conditions that exist in other geo- graphic areas throughout the range of the hard clam may make relaying a more attractive alternative to closed-cycle depuration. ACKNOWLEDGMENTS Paige Gill and Donald Hesselman assisted with field sampling. Clayton’s Crab Company of Rockledge, Florida, provided clams used in the experiments. The New York Clam Farm of Florida, Central Seafood Co., Clayton’s Crab Co., Dyer Shellfish, Inc., and Carlisle Shellfish Co., provided information about and cost estimates of relay and depuration operations. Kenneth Kasweck granted permis- sion to use the Florida Institute of Technology shellfish lease. Thomas Perkins, William Lyons, David Forcucci, and Theresa Bert provided comments on the manuscript. This study was funded by the Florida Department of Nat- ural Resources Hard Clam License Trust Fund. 3Three species of Callinectes occur in the Indian River lagoon (Gore 1977), and we did not distinguish among them. REFERENCES CITED Amold, W. S. 1984. The effects of prey size, predator size, and sediment composition on the rate of predation of the blue crab, Callinectes sapidus Rathbun, on the hard clam, Mercenaria mercenaria (Linné). J. Exp. Mar. Biol. Ecol. 80:207—219. Boulding, E.G. & M. Labarbera. 1986. Fatigue damage: repeated loading enables crabs to open larger bivalves. Biol. Bull. 171:538— 547. Carriker, M. R. 1951. Observations on the penetration of tightly closing bivalves by Busycon and other predators. Ecology 32:73-83. Dillon, R. T. & J. J. Manzi. 1989. Genetics and shell morphology in a hybrid zone between the hard clams Mercenaria mercenaria and M. campechiensis. Mar. Biol. 100:217—222. Gore, R. H. 1977. Studies on decapod Crustacea from the Indian River region of Florida. VII. A field character for rapid identification of the swimming crabs Callinectes ornatus Ordway, 1863 and C. similis Williams, 1966 (Brachyura: Portunidae). Northeast Gulf Sci. 1:119- 123. Hesselman, D. M., N. J. Blake & E. C. Peters. 1988. Gonadal neo- plasms in hard shell clams Mercenaria spp., from the Indian River, Florida: occurrence, prevalence, and histopathology. J. Invertebr. Pathol. 52:436—446. Holmsen, A. & J. Stanislao. 1966. The economy of quahog depuration. Agricultural experiment station, Department of Food and Resource Economics, Univ. of Rhode Island, Bulletin 384: 36 p. Landers, W. S. 1954. Notes on the predation of the hard clam Venus mercenaria by the mud crab Neopanope texana. Ecology 35:422. Magalhaes, H. 1948. An ecological study of snails of the genus Busycon at Beaufort, North Carolina. Ecol. Monogr. 18:377—409. Menzel, R. W., E. W. Cake, M. L. Haines, R. E. Martin & L. A. Olsen. 1976. Clam mariculture in northwest Florida: field study on predation. Proc. Natl. Shellfish. Assoc. 65:59-62. Menzel, R. W. & H. W. Sims. 1962. Experimental farming of hard clams, Mercenaria mercenaria, in Florida. Proc. Natl. Shellfish. Assoc. 53:103—109. Paine, R. T. 1962. Ecological diversification in sympatric gastropods of the genus Busycon. Evolution 16:515—523. Peterson, C. H. 1982. Clam predation by whelks (Busycon spp.): experi- mental tests of the importance of prey size, prey density, and seagrass cover. Mar. Biol. 66:159—170. SAS Institution, Inc. 1985. SAS user’s guide: statistics, version 5 edition. SAS Institute Inc., Cary, NC. 956 p. Simmons, E. G. & J. P. Breuer. 1962. A study of redfish, Sciaenops 320 MARELLI AND ARNOLD ocellata Linnaeus and black drum, Pogonias cromis Linnaeus. Publ. Inst. Mar. Sct. Univ. Texas 8:189—211. Supan, J. E. & E. W. Cake, Jr. 1982. Containerized-relaying of polluted oysters (Crassostrea virginica [Gmelin]) in Mississippi sound using suspension, rack, and onbottom-longline techniques. J. Shellfish Res. 2:141-151. Underwood, A. J. 1981. Techniques of analysis of variance in experi- mental marine biology and ecology. Oceanogr. Mar. Biol. Annu. Rev. 19:513—605. U.S. Public Health Service. 1965. U.S. Department of Health, Education, and Welfare, Public Health Service, National Shellfish Sanitation Pro- gram manual of operations. Part I. U.S. Public Health Service Publ. No. 33, Washington, DC: 32 p. Vermeij, G. J. 1978. Biogeography and Adaptation: Patterns of Marine Life. Harvard Univ. Press. Cambridge, MA. 332 p. Winer, B. J. 1971. Statistical principles in experimental design. Second edition. McGraw-Hill, New York. 907 p. APPENDIX Estimates of the added costs of field depuration and closed-cycle depuration. Original equation from Holmsen and Stanislao (1966). Variable Description Estimate ny: Added cost of relaying per bushel Q No. bushels harvested per day (8-person relay team) 34—40 (c Cost of boat operation per day (relay team) $160 P Dockside price of bushel $30—75 D Efficiency of relaying as opposed to harvesting (=time spent reharvesting clams as a % of original harvest time) 30% N % loss due to relaying 11.5-16.5% O Cost of bonded observer $60 The equation for the added cost of depuration is TOOW ey O Or, OD Q Holmsen and Stanislao (1966) did not include the (Q x P x D) and the (O x D) terms. The (Q x P X D) term is included because time spent reharvesting field-depurated clams is time taken away from harvesting open-water clams and thus represents a loss of efficiency. The (O x D) term is added because an observer is needed when relaying to a lease or to closed-cycle depuration, and an observer must also be hired when clams are reharvested from the shellfish lease. Closed-cycle depuration costs can be estimated with Ys knowledge of the following two terms: CC = cost per clam for depuration (varies between $0.02 and $0.03), and B = number of clams per bushel. The number of clams per bushel (B) varies inversely with the size of the clams being harvested, and averages 200 to 350 for Indian River clams. Closed-cycle depuration costs using the range of variables are as follows: Cost of Closed-Cycle CC B Depuration (per bushel) $0.02 200 $4.00 0.02 250 5.00 0.02 300 6.00 0.02 350 7.00 0.025 200 5.00 01025" 250 6.25 0.025 300 7.50 0.025 350 8.75 0.03 200 6.00 0.03 250 7.50 0.03 300 9.00 0.03 350 10.50 Comparing these estimates to estimates of Y based on a range of levels of the variables indicates that relaying is theoretically not profitable as practiced in the Indian River unless certain variables are held at what we consider to be unreasonably low values. Cost of relaying declines when Q (number of bushels harvested) increases and declines when all other variables decrease. The most critical variables in determining the estimated costs are N (% loss due to re- laying), D (efficiency of relaying), and P (dockside price per bushel). In the following table, the variables C, O, and Q are set at conservative and realistic levels and the vari- ables P, D, and N are varied. MERCENARIA FIELD DEPURATION LOSSES Depuration —Savings/ (e P by (0) (0) N vy Cost Loss cc B 120) 50) O'S 60) 37:5 ‘OIS 33:15 6.25 — 26.9 0.025 250 120 37.5 0.5 60 37.5 0.115 25.46 6.25 —19.212 0.025 250 120) 25°) 10:5) (60) 37:5) (05115) 17-77, 6.25 —11.525 0.025 250 120 50 0.4 60 37.5 0.115 27.67 6.25 — 21.42 0.025 250 120 37.5 0.4 60 37.5 0.115 21.23 6.25 —14.982 0.025 250 120 25 0.4 60 37.5 0.115 14.79 6.25 —8.545 0.025 250 120 50 0.3 60 37.5 0.115 22.19 6.25 — 15.94 0.025 250 120 37.5 0.3 60 37.5 0.115 17.00 6.25 —10.752 0.025 250 1207-25) | 0:3 60° 37:8) 0/115! 11°81 6.25 —5.565 0.025 250 1200500 022) 560) 3725) ONIS 5 16271 6.25 —10.46 0.025 250 120n 37a) 0260) S725) 0.15" 12577 6.25 — 6.5225 0.025 250 120 25 0.2 60 37.5 0.115 8.835 6.25 —2.585 0.025 250 120 50 0.1 60 37.5 0.115 11.23 6.25 — 4,98 0.025 250 120 37.5 0.1 60 37.5 0.115 8.542 6.25 —2.2925 0.025 250 B20 257) 05) (600.3725) O'N1S' S:855 6.25 0.395 0.025 250 P2050" 0!59 G0 S755) 10105, 29:9) 6.25 — 23.65 0.025 250 1203 7-5110'5! 160) 37-9) (0105) 23102 6.25 —16.775 0.025 250 120/725) ~10:5" 60) 37.5: O0105\ 16.15 6.25 —9.9 0.025 250 120 50 04 60 37.5 0.05 24.42 6.25 =138:17: 0.025 250 120 37.5 0.4 60 37.5 0.05 18.79 6.25 —12.545 0.025 250 120 25 0.4 60 37.5 0.05 13.17 6.25 = 6192 0.025 250 120 SO 0.3 60 37.5 0.05 18.94 6.25 — 12.69 0.025 250 120 37.5 0.3 60 37.5 0.05 14.56 6.25 —8.315 0.025 250 120 25 0.3 60 37.5 0.05 10.19 6.25 —3.94 0.025 250 120 50 0.2 60 37.5 0.05 13.46 6.25 SPX 0.025 250 120 37.5 0.2 60 37.5 0.05 10.33 6.25 —4.085 0.025 250 120 25 0.2 60 37.5 0.05 7.21 6.25 — 0.96 0.025 250 120 50 0.1 60 37.5 0.05 7.98 6.25 — 1:73 0.025 250 120 37.5 0.1 60 37.5 0.05 6.105 6.25 0.145 0.025 250 120 25 0.1 60 37.5 0.05 4.23 6.25 2.02 0.025 250 This table clearly shows that when comparing costs of relaying with the costs of land-based depuration, relaying is only rarely profitable. In fact, there are only three cases in which relaying would be profitable, and the maximum ben- efit per 8-person relay team would total $82.50, or $10.31 per team member. We feel that in this scenario, an effi- ciency value of 10% and a loss value of 5% are unrealisti- cally low, and we conclude that relaying as currently prac- ticed in the Indian River lagoon is not profitable. ( Journal of Shellfish Research, Vol. 9, No. 2, 323-327, 1990. EFFECT OF DECREASING OXYGEN TENSION ON SWIMMING RATE OF CRASSOSTREA VIRGINICA (GMELIN, 1791) LARVAE ROGER MANN AND JULIA S. RAINER School of Marine Science Virginia Institute of Marine Science College of William and Mary Gloucester Point, VA 23062 ABSTRACT Four sizes of Crassostrea virginica Gmelin larvae (mean lengths 76.8, 118.1, 139.7 and 290.2 jm) were exposed to stepwise decreases in oxygen concentration from 100% saturation (5.38 ml/I at 22°C and 22 ppt salinity) to as low as 10% saturation and their swimming rates (net vertical movement per unit time) were recorded at each oxygen concentration. No cessation of swim- ming was observed and in only two conditions, that of 76.8 jm larvae at 10% saturation and 290.2 jm larvae at 21% saturation, was swimming rate significantly lower than that of the same size larvae at full saturation. KEY WORDS: INTRODUCTION The Chesapeake Bay and its tributary subestuaries expe- rience seasonal stratification in terms of density, salinity, temperature and dissolved oxygen content. Examination of seasonal hypoxia and anoxia has been the focus of much recent and continuing work (Mackiernan 1987). The sea- sonal occurrence of stratification coincides with or partially overlaps the period of spawning and settlement of the oyster Crassostrea virginica Gmelin. Although the spatial occurrence of hypoxia or anoxia is usually restricted to deeper waters, the seiching of deeper waters due to wind stress periodically results in irrigation of the shallower areas, where oyster reefs abound, with hypoxic or anoxic water. The present consensus is that bivalve larvae employ depth regulation to effect their retention in shallow, strati- fied estuaries (Mann 1986). The possibility therefore exists that larval stages of the oyster are subjected to stress of hypoxia or anoxia in the Chesapeake Bay during their planktonic existence if hypoxic conditions prevail in the deeper, more saline, upstream-flowing waters that are con- sidered integral to the mechanism of larval retention. What, then, is the behavioural response of oyster larvae to de- creasing oxygen tensions similar to that experienced in de- scending from surface waters to deeper, hypoxic strata? Would swimming behaviour result in avoidance of all but near-saturated water with resultant isolation of larvae in surface, seaward flowing water and their eventual loss from the estuarine system, or would hypoxia result in valve clo- sure and loss to the benthos due to sinking, or would some intermediary response be evident? With these options in mind, the following study examined the swimming re- sponse of various developmental stages of oyster larvae to stepwise decreases in oxygen tension. MATERIALS AND METHODS Oyster, Crassostrea virginica Gmelin, larvae at various stages of development were obtained from the Virginia In- 323 swimming, Crassostrea virginica, larvae, oyster, oxygen stitute of Marine Science (VIMS) oyster hatchery. Details of oyster spawning procedure and larval culture were sim- ilar to the techniques previously described for the hard clam, Mercenaria mercenaria L., by Castagna and Kraeuter (1981). Ripe oysters were spawned by thermal stimulation, the resultant eggs fertilized and the cultures maintained in water originating from the York River at Gloucester Point. Experimental larvae therefore originated from several parents rather than a single male-female cross. Larvae were cultured in 10001 tanks and the water changed at intervals of two days. At each water change larvae were fed with additions of the flagellate /sochrysis galbana Parke. No attempt was made to control the salinity of the culture water, which typically varies in the range 15—22 ppt at the hatchery site. When culture salinity differed from the desired experimental salinity larvae were acclimated to the latter by daily water changes with salinity adjustment not exceeding 2 ppt/day. Larvae of first-shelled (straight hinge), mid development (umbo) and competent-to-meta- morphose (pediveliger) stages were used in experiments. Appropriate size ranges of larvae were obtained by selec- tive sieving on nylon mesh screens. All larvae of one size class were from the same culture. Two cultures from the same parental broodstock were used to provide the four size classes examined. All experiments were effected at 22 ppt salinity and 22°C, water temperature being maintained by control of the laboratory air temperature. All observations of larval swimming were made in vertically oriented, square cross- section, borosilicate tubing (Wale Apparatus, Hellertown, PA) measuring 30 cm H X 6 mmL X 6 mmW internal dimensions (approximately 10.8 ml volume). The tube walls were optically flat and allowed both direct observa- tion of larval swimming and video recording. Both upper and lower ends of the tube were covered with 20 zm nylon mesh to retain larvae. Over each mesh was placed butyl rubber or Fisher brand C-flex tubing (of low porosity to 324 MANN AND RAINER oxygen) attached to a valve. The glass tube thus formed a chamber that could be sealed at both the top and bottom. The top valve was connected to a variable speed peristaltic pump (Buchler). The bottom valve was connected to a con- ical flask in which sea water, at the experimental tempera- ture and salinity, was bubbled with an appropriate mixture of nitrogen and air to obtain stable oxygen concentrations of less than 100% of saturation. Oxygen tension in the con- ical flask was measured with either a Radiometer or Strath- kelvin oxygen electrode connected to a Strathkelvin 781b amplifier/meter. The electrode was calibrated daily. The experimental procedure started with closing the bottom valve, temporary removal of the top valve and as- sociated mesh, partial filling of the glass tube with air-satu- rated sea water, gentle addition of larvae and air-saturated sea water using a Pasteur type pipette to fill the tube and replacement of the top mesh and valve. Between 30 and 150 larvae, equivalent to concentrations of 3 and 15 larvae/ ml when uniformly dispersed, were used with numbers in- creasing at smaller sizes. Larvae were allowed to recover for approximately 30 minutes, recovery being recognized by continuous active swimming. Throughout this period larvae were continually observed. Following acclimation video recordings were made of swimming activity for later determination of individual larval swimming rate. After a period of exposure at saturation, larvae were exposed to stepwise decreases in oxygen tension. These were accom- plished by opening both the bottom and top valves and gently aspirating water through the tube from the conical flask using the peristaltic pump. Larvae were retained by the mesh, but a complete flushing (checked by a previous dye study) of water was effected in a 2—3 minute period. The valves were then closed, the pump turned off and the larvae allowed to re-equilibrate for approximately 5 minutes, still under constant observation, before further measurements of swimming rate was made. The experiment employed stepwise decreases in oxygen tension from saturation to lower values (given in Table 1). A typical experimental protocol required 20—25 minutes at each oxygen tension before a subsequent further decrease. An experiment examining five different oxygen tensions required approximately two hours to complete. On termina- tion of the experiment, both valves were opened, the larvae were drained through the bottom valve, retained on a 53 jm mesh, transferred to a glass shell vial, fixed in 5% v/v buffered formalin and subsequently measured to obtain a mean individual length (maximum dimension parallel to the hinge line) using a compound microscope equipped with a calibrated ocular micrometer. Recordings for estimation of swimming rate were made with the system described in Mann (1988). A high resolu- tion, IR sensitive video camera (Dage-MTI SC65S with Ultricon phototube: Eastern Microscope Co., Raleigh, NC) was mounted on a vertically travelling stage (Velmex, Inc., E. Bloomfield, NY). The stage was driven by a Bodine S41 motor and Minarik SL-15 speed control allowing a variable speed traverse from 0.1—10 mm/sec—encompassing larval swimming speeds as recorded in the literature (see review by Mann 1986). General observation of all larvae in the tube was facilitated by movement of the camera on the stage; however, measurements of swimming rate were made with the camera fixed. Illumination for macro-video recording was facilitated by attaching a fiber optic ring light (Fiber Optic Specialties, Inc., Peabody, MA, model LS81A fiber optic with FA-83 filter holder) to the camera lens (50 mm Series E Nikon attached to a Nikon PB-6 bellows and video C-mount). Even though earlier experi- ments (Mann, unpublished data) had failed to demonstrate a response by larvae to intense, orientated white light, this potential artifact was eliminated by inserting a 695 or 850 nm long pass filter (Oriel Corp., Stratford, CT) in the filter holder and recording under low intensities of essentially IR light. Room lighting was maintained at minimal levels throughout the experiment. Video recording (Panasonic NV-8950 recorder and Panasonic WV-5410 monitor) at such low light levels is not problematic in that the Dage SC65S camera has high sensitivity to light wavelengths up to 1200 nm and operates optimally at an intensity of 3.7 x 10-3 ~W.cm~? (approximately 0.1 foot candles). Prior to each experiment a calibrated (in mm) plexiglass ruler was suspended in the borosilicate tube, equidistant be- tween its front and back (with respect to the “‘view’’ of the camera) walls. The camera was focussed on the ruler and its magnification adjusted, using the bellows, until a dis- tance of approximately 4 mm on the ruler filled the vertical displacement on the monitor screen. The camera focus and magnification were then fixed and a recording made of the tuler scale on the video tape together with a time and date overlay (Panasonic WJ-810 time-date generator). An audio commentary describing larvae to be used, proposed oxygen exposure regime and other relevant experimental details were also included on the videotape. The ruler scale pro- vided the basis for all subsequent measurements of larval swimming rate from that experiment’s recordings. Mea- surements of individual larval swimming speed were not made during the experiment, but were recorded from replay of the videotapes. A grid, corresponding to the aforemen- tioned ruler calibration, was temporarily fixed to the video monitor and the video tape replayed at reduced speed. The field of observation corresponds to a volume of approxi- mately 0.13 ml which typically contained |—4 larvae at any one time during recording. The vertical movement of indi- vidual larvae across fixed intervals of the grid was timed using the time elapsed recording on the video tape. From individual rates a measurement of mean rate of net vertical movement for each size of larva examined was thus ob- tained. Only one videotape was used per experiment to eliminate possible confusion in subsequent data analysis. OYSTER LARVAL SBWIMMING Ww iw) n TABLE 1. Swimming rate (net vertical movement per unit time in mm/sec) of oyster larvae at various concentrations of dissolved oxygen. ml/I values calculated from % saturation using Table 4 of Carpenter (1966) assuming salinity = 0.03 + 1.805 x chlorinity (Sverdrup et al. 1942). Age Length S.D. D.O. Swimming Rate mm/sec Days pm pm n % sat ml/l Mean S.D. Min Max 95% Interval n 2 76.8 2.1 30 100 5.38 0.98 0.25 0.63 1.43 0.80—1.16 10 77 4.14 0.85 0.51 0.27 Dele 0.48-1.21 10 45 2.42 0.99 0.36 0.45 1.47 0.73-1.25 10 6 118.1 10.6 32 100 5.38 1.48 0.67 0.27 2.27 1.01-1.96 10 56 3.01 1.12 0.44 0.45 2.00 0.80—1.44 10 36 1.94 1.21 0.73 0.35 2.50 0.69-1.73 10 28 1.50 Les 0.69 0.35 2.50 0.66—1.65 10 10 0.54 0.64 0.30 0.25 1.10 0.42—0.85 10 10 139.7 21.7 30 100 5.38 1.79 0.79 0.63 2.94 1.23-2.36 10 50 2.69 1.94 0.88 0.50 3.13 1.31-2.57 10 22 1.18 EDD 1.18 0.94 4.17 1.37—30.6 10 13 290.2 28.2 30 100 5.38 3.10 135) 1.32 5.00 2.14-4.07 10 45 2.45 1.93 0.75 1.14 3n13 1.40—2.47 10 21 1.13 1.36 0.42 0.86 2.38 1.06—1.66 10 RESULTS Table 1 summarizes data on age and length of larvae examined and their respective swimming rates under various concentrations of dissolved oxygen. Mean net ver- tical swimming rates vary in the range of 0.64—3.10 mm/ sec. A series of one-way analyses of variance were per- formed comparing swimming rates at each oxygen concen- tration at each size. Significant differences with decreasing oxygen were observed at 118.1 jzm length, where the swimming rate at 10% of saturation was lower (P < 0.05) than at saturation. At 290.2 pm length (pediveliger larvae) the swimming rate at 21% of saturation was lower (P < 0.01) than at saturation. With these exceptions decreasing oxygen concentration was not accompanied by a significant decrease in mean swimming rate within the time course of the experiment. Comparisons of mean swimming rate of different sizes of larvae are complicated by the fact that, with the exception of values recorded at saturation, oxygen concentrations and immediately prior oxygen environment are not identical. First-shelled veliger or D larvae at 76.8 um have statistically significant lower swimming rates at saturation than larvae with shell lengths geater than 139.7 um. Pediveliger larvae of 290.2 um length swim at signifi- cantly faster rates than either 76.8 or 118.1 jm length larvae at saturation. DISCUSSION The most significant findings of this study are that Cras- sostrea virginica larvae do not cease swimming as oxygen concentration decreases and that a statistically significant decrease in swimming rate is not observed until larvae are exposed to the lowest oxygen concentrations examined, even when exposure periods approach 20—25 minutes at each concentration and cumulative exposure to increasing levels of hypoxia approaches two hours. Previous studies with C. virginica larvae suggest a predominantly lipid-pro- tein based, aerobic energy metabolism (Gallager et al. 1986) similar to that of shipworm larvae at normoxia (Mann & Gallager 1985). The present observations suggest that aerobic metabolism can be maintained at hypoxia due to the large surface to volume ratio of all stages of veliger larvae examined, that oxygen requirements of the velar cells responsible for swimming can be satisfied in that dif- fusion pathways to them are short, and/or that some limited capability for anaerobiosis is present. Recently Widdows et al. (1989) examined heat production, oxygen consumption and feeding of C. virginica larvae under prolonged hypoxia and anoxia. They concluded that such larvae have limited capability to function anaerobically under hypoxia, as indi- cated by both feeding and activity observation. They also recorded a notable difference in response to prolonged an- oxia and hypoxia exposure stress by first-shelled and pedi- veliger larvae in that the former maintain activity under stress (essentially an avoidance response), but eventually succumb within a few hours, whereas the latter decrease activity within a short period but survive for considerably longer. Given the increase in specific gravity accompa- nying development from first-shelled to pediveliger larvae, and the contrasting roles in development (dispersal versus seeking metamorphic substrate) of these larval stages these responses are expected. In the present study a similar re- sponse is observed for the pediveliger stage, that is a marked reduction in swimming rate at 21% of saturation. With the exception of the lowest oxygen concentration ex- 326 MANN AND RAINER amined for 118.1 um larvae the maintenance of activity by larvae in the length range 76.8—139.7 jm under short term hypoxia stress is consistent with the aforementioned obser- vations of maintained activity by Widdows et al. (1989). The exception is notable in that it was recorded at the end of a cumulative hypoxia stress approaching two hours in duration, more consistent with the observation of eventual submission as recorded by Widdows et al. (1989). Morrison (1971) examined the influence of variable pe- riods of exposure to low oxygen environments on the em- bryonic and larval development of the hard shell clam Mer- cenaria mercenaria. Eggs developed normally at oxygen concentrations of 0.5 mg/l (7% of saturation at the experi- mental conditions of 28—30 ppt salinity and 25°C). Larval growth was curtailed at or below 2.4 mg/1 (34% of satura- tion) but proceeded normally above 4.2 mg/] (60% satura- tion). Larvae were capable of recovering from periods of growth inhibiting hypoxic conditions when subsequently transferred to normoxic conditions. The ability of C. vir- ginica larvae to grow under hypoxic stress in a manner comparable to M. mercenaria larvae has not been exam- ined but is clearly worthy of study. If such growth capabili- ties are present then the observation of sustained swimming activity at moderate hypoxia (over 60% of saturation) could be considered normal rather than an avoidance response as suggested earlier. In such an instance the description of an avoidance response should be restricted to sustained swim- ming activity under hypoxic conditions associated with growth inhibition or cessation. The methods of estimating swimming rate in the present study represent a significant advance over most previous efforts. The studies of Cragg (1980) and Mann and Wolf (1983) both used travelling microscopes to observe larval swimming. In the latter case, rate and magnitude of vertical movement was recorded, via a ten-turn potentiometer, on a strip chart recorder. A manually operated travelling micro- scope suffers from a prerequisite for considerable operator dexterity to obtain smooth output traces of larval move- ment, a need for the operator to estimate (from an eyepiece graticule) the horizontal component of an observed swim- ming pattern during recording, and a lack of production of a hard record of the observation. Visible light, a prerequi- site of direct observation, has been shown to influence the larval swimming of some bivalve species; however, pre- vious experiments described in Mann (1988) in both light proof boxes and under dim laboratory lighting failed to demonstrate any phototactic response in C. virginica. In these experiments the travelling microscope was replaced with a fixed video camera operating under appropriate light conditions to eliminate operator variability and provide high quality recordings. It is important to note that the re- ported values of net vertical movement per unit time differ from absolute swimming speed because the larva swims in a helical pattern; however, this is the ecologically relevant value in terms of rate of depth regulation. Further, the step- wise decreases in oxygen tension without periodic increases were chosen to simulate conditions of larvae gradually sinking from surface waters to deeper, hypoxic water. This was considered to be ecologically more realistic than expo- sure to a series of randomly chosen concentrations. Swimming rates reported in Table 1 are comparable to previously reported values for Crassostrea virginica larvae by Hidu and Haskin (1978, Fig. 2; 0.83, 1.0 and 1.83—2.33 mm/sec for 80, 160 and 230—270 pm larvae respectively at 25°C and 15—25 ppt salinity) and Mann (1988; 0.37 and 1.02 mm/sec for 75 wm and 157.5 pm larvae respectively at 22°C and 19-22 ppt salinity), for other bivalve veliger larvae including Ostrea edulis L. (Cragg & Gruffydd 1975; 1.23 mm/sec for 200—250 pm larvae at 20—21°C and 32-33 ppt salinity), Teredo bartschi! (Isham & Tierney 1953; 7.7 mm/sec at 20—28°C and unspecified salinity), and a variety of marine inverte- brate larvae as reviewed by Mileikovsky (1973). The range of mean swimming rates recorded in Table | (0.64—3.10 mm/sec) correspond to changes in absolute depth of 2.3 and 11.2 meters per hour for continuously swimming larvae. Given the bathymetric range of oyster reefs in the Chesapeake Bay, generally less than six meters in depth, and the shallow nature of the subestuaries of the bay, it is evident that larvae can, through active swimming alone, depth regulate and ensure retention in the proximity of suitable substrate through exploitation of salinity-driven, depth-specific circulation. If oyster larvae sank into deeper hypoxic zones in the Chesapeake Bay then sustained swim- ming at the aforementioned rates would only be required for intervals of one or two hours to return larvae to nor- moxic surface waters. This time interval is less than that required to reach a point of submission to hypoxic stress by the first-shelled larvae as reported by Widdows et al. (1989). The demonstration of unexpectedly high tolerance of oyster larvae to short term hypoxic stress prompts the ques- tion of whether larvae are the most susceptible stage of the oyster life cycle to this environmental stress. This may not be so in that small larval stages appear to be able to fulfill aerobic requirements by simple diffusive processes. The same may also be true of a wide variety of marine inverte- brate larvae with predominantly lipid-protein based energy metabolism. It is only as size increases, as impermeable external layers (such as shell) develop and the adoption of the sessile benthic form (which cannot escape from hypoxic events) occurs that the ability to supply a major proportion of the metabolic energy from sustained anaerobic activity becomes critical in surviving hypoxic or anoxic stress. Limited data on post settlement changes in gross biochem- ‘Although described as Teredo (Lyrodus) pedicellata by Isham and Tierney (1953) this species was later shown to be Teredo bartschi by Turner and Johnson (1971). OYSTER LARVAL SBWIMMING 327 ical composition of the oyster Ostrea edulis (see Holland & Spencer 1973) suggest that transition to typically adult an- aerobic capabilities, as indicated by an abundance of car- bohydrate reserves, may require as long as thirty days. If comparable periods apply to post settlement Crassostrea virginica in the Chesapeake Bay then periodic irrigation of shallow oyster reefs by hypoxic water caused by wind driven seiching may be a significant source of stress and mortality, indeed more so than the influence of such events on oyster larvae in the same location. ACKNOWLEDGMENTS This study was supported by funds from the National Oceanic and Atmospheric Administration, Office of Sea Grant, and the Commonwealth of Virginia Council on the Environment. We thank Kenneth Kurkowski and the staff of the VIMS oyster hatchery for the provision of larvae. The manuscript was improved by constructive criticism from George Grant, Morris Roberts Jr. and an anonymous reviewer. This is Contribution Number 1614 from the Vir- ginia Institute of Marine Science. LITERATURE CITED Carpenter, J. H. 1966. New measurements of oxygen solubility in pure and natural sea water. Limn. Oceanogr. 11(2):264—277. Castagna, M. & J. N. Kraeuter. 1981. Manual for growing the hard clam Mercenaria mercenaria. Va. Inst. Mar. Sci. Spec. Rep. Appl. Mar. Sci. Ocean Eng. No. 249. Cragg, S. M. 1980. Swimming behaviour of the larvae of Pecten maximus (L.) (Bivalvia). J. Mar. Biol. Ass. U.K. 60:551—564. Cragg, S. M. & L. D. Gruffydd. 1975. The swimming behaviour and the pressure responses of the veliconcha larvae of Ostrea edulis (L.). In: H. Barnes (ed.) Proceedings of the Ninth European Marine Biology Symposium, Oban, Scotland, 1974. Aberdeen University Press, Aber- deen, pp. 43-57. Gallager, S. M., R. Mann & G. C. Sasaki. 1986. Lipids as an index of growth and viability in three species of bivalve larvae. Aquaculture 56(2):81—104. Hidu, H. & H. H. Haskin. 1978. Swimming speeds of oyster larvae Crassostrea virginica in different salinities and temperatures. Es- tuaries 1:252—255. Holland, D. L. & B. E. Spencer. 1973. Biochemical changes in fed and starved oysters, Ostrea edulis L. during larval development, metamor- phosis and early spat growth. J. Mar. Biol. Ass. U.K. 53:287—298. Isham, L. B. & J. Q. Tierney. 1953. Some aspects of the larval develop- ment and metamorphosis of Teredo (Lyrodus) pedicellata De Quatre- fages. Bull. Mar. Sci. Gulf Carib. 2:574—589. Mackiernan, G. B. 1987. ed. Dissolved oxygen in the Chesapeake Bay: Processes and effects. Maryland Sea Grant publication number UM-SG-TS-87-03. 177 pp. Mann, R. 1986. Sampling of bivalve larvae. In Jamieson, G. S., N. Bourne, (eds.) North Pacific workshop on stock assessment and man- agement of invertebrates. Can. Spec. Publ. Fish. Aquat. Sci. 92:107— 116. Mann, R. (1988). Distribution of bivalve larvae at a frontal system in the James River, Virginia. Mar. Ecol. Prog. Ser. 50:29—44. Mann, R. & S. M. Gallager. 1985. Physiological and biochemical ener- getics of the larvae of Teredo navalis L. and Bankia gouldi (Bartsch) (Bivalvia: Teredinidae). J. Exp. Mar. Biol. Ecol, 85:211—228. Mann, R. & C. C. Wolf. 1983. Swimming behaviour of larvae of the ocean quahog Arctica islandica in response to pressure and tempera- ture. Mar. Ecol. Prog. Series. 13:211—218. Mileikovsky, S. A. 1973. Speed of active movement of pelagic larvae of marine bottom invertebrates and their ability to regulate their vertical position. Marine Biology 23:11—17. Morrison, G. 1971. Dissolved oxygen requirements for embryonic and larval development of the hardshell clam, Mercenaria mercenaria. J. Fish. Res. Bd. Canada 28:379-381. Sverdrup, H. U., M. W. Johnson & R. H. Fleming. 1942. The oceans: their physics, chemistry and general biology. Prentice Hall. 1060 pp. Turner, R. D. & A. C. Johnson. 1971. Biology of marine wood boring molluscs. In Jones, E. B. G., S. K. Eltringham, (eds.). Marine borers, fungi and fouling organisms of wood. Organization for Eco- nomic Co-operation and Development, Paris, pp. 259—301. Widdows, J., R. I. E. Newell & R. Mann. 1989. Effects of hypoxia and anoxia on survival, energy metabolism and feeding by oyster larvae (Crassostrea virginica Gmelin). Biol. Bull. 177:154—166. ; ; Journal of Shellfish Research, Vol. 9, No. 2, 329-339, 1990. SPATIAL AND TEMPORAL PATTERNS OF OYSTER SETTLEMENT IN A HIGH SALINITY ESTUARY* PAUL D. KENNY, WILLIAM K. MICHENER, AND DENNIS M. ALLEN Belle W. Baruch Institute for Marine Biology and Coastal Research University of South Carolina P.O. Box 1630 Georgetown, SC 29442 ABSTRACT Settlement patterns for the eastern oyster, Crassostrea virginica were studied during a five year period (1981—1986) in a high salinity southeastern estuary where oysters form densely populated intertidal reefs. Vertical arrays (four levels) of collecting plates at three locations were analyzed every two weeks to determine spatial and temporal patterns of distribution. Level within the intertidal zone and date accounted for the largest proportion of the variability in a nested ANOVA which also included location (site), side of plate (exposed or shaded), harness location (proximity to creek bank), and year as variables. Spat settlement occurred during the same 180 to 200 day period each year, and, although within year fluctuations in abundance were large, an early and late season peak usually occurred each year. Within years, the timing and abundance patterns of spat were similar at all three sites. Interannual variations in abundance patterns were also similar at all sites. Within and among year differences in settlement intensity were generally not related to changes in water temperature or salinity, but the lowest spatfall coincided with unusually high temperatures and salinity in the summer of 1986. Consistently higher numbers of spat on plates near mean low water and low numbers in the high intertidal zone were not solely related to duration of submergence. Significantly lower spatfall on tops (exposed side) of plates in the high intertidal and higher spatfall on bottoms (shaded side) of plates lower in the tidal zone indicated different responses of larvae to light, dessication, and/or sedimentation at different tidal levels. The importance of plate sides and proximity of the harness to the bank differed among sites. Similarities in temporal patterns among sites suggest that factors controlling oyster spat settlement are operating at the ecosystem or broader spatial level. Since the only relationship between spatfall intensity and physical factors (water temperature and salinity) was during extreme conditions, variations in other system-wide factors affecting behavior and survival of larvae and newly settled spat are probably more important in controlling intra- and interannual patterns of oyster settlement. The gregarious settlement of spat on replicated collection plates and competition with other invertebrates for space suggest that biological controls play an important role. KEY WORDS: Crassostrea virginica, settlement, spatial variability, temporal variability, oyster INTRODUCTION The eastern oyster, Crassostrea virginica (Gmelin 1791), occurs in estuaries along most of the Atlantic and Gulf of Mexico coasts of North America. Oyster reefs de- velop in a variety of estuarine habitats where they are ex- posed to broad ranges of conditions on many different tem- poral and spatial scales. Because of the economic impor- tance of the eastern oyster, much research has been directed toward understanding factors which limit their development and survival (e.g., Ingle 1951, Chestnut & Fahy 1953, Galtsoff 1964, Loosanoff 1966, MacKenzie 1970, Haven & Fritz 1985, Abbe 1986). Oyster reefs are primarily subtidal structures at the northern and southern ends of the species’ range; however, in southeastern coastal areas, most are intertidal. In South Carolina, approximately 95% of the oyster beds occur above mean low water (Gracy & Keith 1972). Dame et al. (1984) determined that oyster reefs reach their greatest density and biomass in the southeast, but, because of the difficulty in harvesting and shucking clusters of relatively *This paper represents Contribution Number 822 from the Belle W. Baruch Institute for Marine Biology and Coastal Research. 829 small oysters, these intertidal populations are not as impor- tant commercially as those in northern and Gulf coast es- tuaries (Burrell 1985). Regardless of their value as a fishery resource, intertidal oysters are dominant filter feeders which comprise an important link in the cycling of nutrients and energy within southeastern salt marsh ecosystems (Dame et al. 1984). The transformation from planktonic larvae to sessile spat and the survival of young oysters for up to two weeks rep- resents a period or event that, in this study, is referred to as settlement. We investigated patterns of settlement at sev- eral sites and levels within the intertidal zone of a well mixed southeastern estuary for a period of five years. We specifically addressed the following questions: (1) How much do the timing and intensity of oyster settlement vary among years and sites, and are these variations related to measurable environmental factors? (2) Are seasonal settle- ment dynamics modified by extreme climatic events such as the 1986 drought? (3) Is differential settlement within the intertidal zone mainly related to submergence (or expo- sure) time? (4) Does settlement occur differentially be- tween shaded and exposed surfaces? (5) Can horizontal set- tlement patterns within a site be attributed to random pro- cesses? 330 KENNY ET AL. STUDY AREA North Inlet Estuary (Fig. 1) is a tidally dominated salt marsh ecosystem located 5 km northeast of Georgetown, South Carolina. Semi-diurnal tides have an average range of 1.5 m, and typical peak current velocities can be greater than 2.3 m/sec (Kjerfve 1986). The system is well mixed and salinities are usually greater than 32 ppt. Water temper- ature ranges annually from 0 to 31°C (mean 19°C). Because of shallow water depths (average 3 m) and intense tidal flushing, North Inlet creeks are well mixed and almost ver- tically homogeneous. Three stations, Town Creek, Old Man Creek, and Oyster Landing were selected as spat settlement monitoring sites from 1982 to 1986 (Fig. 1). The Town Creek and Old Man Creek sites are located in tidal creeks with similar flow patterns and physical characteristics. Salinity remains high unless tide and weather conditions allow intrusion of brackish water from Winyah Bay into the North Inlet system or there is an extended period of significant precipi- tation input. Oyster Landing is located in a tidal creek adja- cent to the forest and differs from Old Man Creek and Town Creek in two conspicuous ways: (1) it receives fresh water runoff from a large portion of the adjacent uplands which may cause salinities to be lower than at the other } Winyah Ca Bay Figure 1. The North Inlet Estuary, showing the location of the three sampling sites, Town Creek (TC), Old Man (OM), and Oyster Landing (OL). Shaded regions represent marsh (light) and upland (dark). Insert indicates position of estuary along South Carolina coast. sites for extended periods, and (2) average current velo- cities are lower. MATERIALS AND METHODS At each station two (1982—1984) or four (1985-1986) vertically oriented rope harnesses were used to maintain 225 cm?, 5 mm thick asbestos cement plates in a horizontal plane at four different levels (Fig. 2). The harnesses were suspended from PVC pipes which were located approxi- mately 3 m from the creek bank; and concrete blocks held the arrays in place. Level | (LV1), the highest level, was 120 cm above mean low tide with Level 2 (LV2) half way between mean high and mean low at 70 cm above mean low. Spat collectors at Level 3 (LV3) were in the low inter- tidal zone, 30 cm above mean low, whereas Level 4 (LV4) was subtidal and 30 cm below mean low tide. Level 1 was discontinued in 1985 and, at the same time, the number of harnesses was increased to 4 at each station, thus increasing the number of plates from 2 to 4 at each level. One side of each plate was slightly textured and was used as the top side throughout the study. The possible effect of this tex- tural difference on settlhement was examined by Hidu (1978) using identical plates. In his study, all plates were inverted during two measurement periods and no effect was noted. Plates were replaced every 2 weeks (14 + 2 days) in the settlement season and every four weeks in winter. They were transported to the lab in upright slotted trays to pre- vent abrasion. If necessary, plates were gently rinsed with water to remove excess sediment. Counts of spat were made with a dissecting microscope (125 x ) on the entire top and bottom sides, excluding edges, of all plates. Spat ranged in size from 350—500 pm. Before reuse, plates were wire brushed under running water until all organisms were removed. In 1986, two sequential three minute oblique zoo- plankton tows were made every three days adjacent to spat collectors at the TC site. A 153 wm, 30 cm diameter con- ical Nitex plankton net fitted with a flowmeter was used. Estimates of late stage larval oyster densities are based on the mean of the two tows. Daily water temperatures and salinities at 1 m depth were recorded throughout the 5 year period at three nearby sites in North Inlet. All measurements represent observa- tions at 1000 hr. In addition, surface and bottom water tem- peratures at Town Creek were continuously recorded during 1986 with an in situ thermograph. Harmonic regres- sion analysis (Chatfield 1984) of existing tidal data (NOAA 1985) was used to calculate the percentage of time that plates at each level were exposed or submerged. STATISTICAL ANALYSIS Nested analysis of variance was employed as an explor- atory method to quantify the magnitude of variability in spat settlement associated with differences between site, OYSTER SETTLEMENT IN A HIGH SALINITY ESTUARY 331 Lp | a FLOODIN € INTERTIDAL REEF ooo - A. TOP VIEW MEAN HIGH RN i eal 120 cm ie a6 SPAT PLATE es a 70 cm == ee) = icamnl 30 a | - _MEAN LOW 0 [= S ANCHOR BOTTOM B. SIDE VIEW Figure 2. Top view (A) and side view (B) of an intertidal spat collec- tion apparatus. Positions of the plates relative to the direction of flooding tide, an intertidal reef, and mean low tide are shown. year, date, level in water column, side of plate, and harness location (SAS 1985, Michener et al. 1987). Standard anal- ysis of variance techniques (ANOVA) (SAS 1985) were performed on log;9(x+ 1) transformed data utilizing a com- plete block design (Sokal & Rohlf 1981). Tukey’s Studen- tized Range Test was used for all multiple comparison tests (Mize & Schultz 1985). Initial ANOVAs demonstrated sig- nificant interaction among two or more of the six factors. For further analyses, several reduced models were neces- sary in order to discern effects of various factors. Harness location was treated as a blocking factor for each ANOVA. The final models are discussed below: 1. Site and year analyses—Level 1 data were dropped from the data set to standardize comparisons and simplify interpretation. Numbers of spat were summed for both sides of each plate on each harness across all sampling dates within a given settlement season. Effects of year and site were considered as a single main effect. Side comparisons— All levels were retained for this series of analyses. Data from sampling dates within a settlement season were summed to obtain total an- nual spat settlement per side of each individual plate. An ANOVA was performed for each combination of site and year. The main effect tested was the effect of level in water column and side of plate. tO 3. Level in water column comparisons—Numbers of spat counted for the top and bottom of a plate were summed across all sampling dates within a settlement season to obtain total annual spat settlement per level in water column. An ANOVA was performed for each combination of site and year (15 total) to test for the effect of level in water column on oyster set. 4. Harness position—Spat counts were summed to ob- tain an estimate of total settkement per harness. The total harness count was divided by the total number of spat setting on all harnesses within an individual site to obtain percentage of spat setting on each har- ness. An ANOVA was performed on the percentage data for each site and year combination. Chi-square analysis (Steel & Torrie 1980) was used to test the hypothesis that biweekly settlement was equally distributed among the four plates at a single site and level in water column. Sides of plates were treated separately in the analysis. Only samples where total settlement on all four plates (single side; top or bottom) was greater than or equal to 20 were included in the analysis, following guide- lines recommended by Siegel (1956) and Dixon and Massey (1969). A total of 171 instances where at least 20 spat settled on the same side of the four plates at any site and level in water column was included in the analysis. These data were pooled for a single chi-square analysis. The chi-square test statistic was calculated using SAS (SAS 1983). RESULTS Variability Within the Long-Term Data Set Results from the nested analysis of variance indicate rel- ative magnitudes of variability associated with differences in oyster spat settlement among sites, years, dates, level in water column, side of plate, and harness location (Table 1). The dominant sources of variability were level in water column and date within settlement season. Harness location and year (settlement season) each accounted for about 15% of the total variation. Side of plate accounted for the smallest portion (8%). Location within the estuary was an insignificant factor in explaining the total variation within the data set. Temporal Pattern The duration of the setting season was between 180 and 200 days. Settlement began between late April and mid May and ended between late October and mid November during each of the five years (Fig. 3). Less than 1% of the total annual set occurred at the beginning and end of the season (before June and after October). Settlement first oc- curred each year when mean temperature during the 14 day sampling period was between 21.6 and 23.2°C. Maximum and minimum temperatures based on daily measurements during the same period were between 18 and 25°C (Fig. 3). 332 KENNY ET AL. TABLE 1. Components of variation for oyster spat settlement at three sites within North Inlet Estuary. The top and bottom sides of 15 x 15 cm settlement plates were sampled at three or four discrete levels within the water column on a biweekly basis throughout 5 sequential settlement seasons (1982-1986). Two or four harness locations were sampled within each site. Each source of variation is considered to be nested within preceding sources of variation (e.g., total variation due to date differences represents variation due to differences among dates within years nested within sites). Percentage of Total Variation Due to Differences Among Variance Source df Variance Sources Site 2 0 Year 12 14 Date 150 27 Level in water column 416 36 Side of plate 560 8 Harness location 1854 16 Mean spat counts were variable among sampling dates and ranged from 0 to 485 per side of plate, but most values were between 10 and 70 spat per side from June through August (Fig. 3). Mean water temperatures and salinities for biweekly collection periods ranged from 15—28°C and 29-36 ppt, but most values were between 23—27°C and 30—34 ppt (Fig. 3). Settlement was continuous during most summers, but two or more peaks occurred; the first was in early June and the second was in late July or early August (Fig. 3). During 1983, 1984, and 1985 the second peak was more intense and accounted for more than 30% of annual set, but in 1982 and 1986 the first peak was more intense. A third period of high settlement occurred in early fall 1982 and 1984. The first summer peak occurred each year when mean temperatures for the sampling period were between 23 and 25°C and generally rising. The second peak usually oc- curred during a period of very stable water temperature; mean temperature at the time of the second peak ranged between 24.6 and 28.7°C. In 1982 and 1984 a third peak occurred when mean water temperature was 22.9°C and decreasing. There appeared to be no consistent relationship between pulse settlement events and water temperature or salinity. The timing of each of the two main peaks each summer was predictable, but they occurred during periods of both in- creasing and decreasing temperatures and salinities (Fig. 3). Continuous water temperature recordings at the surface and 0.3 m below mean low tide at Town Creek in 1986 showed that there was no vertical stratification. Water tem- perature fluctuations during a 24 hr period were usually small (2 or 3°C). Although some daily fluctuations as large as 4°C were recorded, they did not coincide with any pulse event. Spat settlement was significantly higher in 1984 (p < 0.05) at both Town Creek and Oyster Landing Creek (Fig. 4). Settlement at Old Man Creek was also highest in 1984, but total annual settlement was not significantly higher than that observed in 1983. Settlement was lowest during 1986 at all three sites and could be related to the absence of a late summer pulse in late July and early August (Fig. 5). Daily observations revealed that water temperatures were higher than average and salinities were high and relatively stable during this period (Fig. 6). Comparisons Among Sites No significant differences (p < 0.05) in annual spat set- tlement were detected among sites. However, annual spat settlement was highest at Town Creek for 3 of the 4 com- plete seasons observed (Fig. 4). Highest settlement was ob- served at Old Man Creek in 1986 which was also the year of lowest total settlement. There were no consistent differ- ences in spat settlement between Old Man Creek and Oyster Landing. Temporal trends among and within sites were similar each year, but there were minor differences in the duration of setting events. The number of settlement peaks varied among years, but within each year all sites usually had the same number of peaks (Fig. 5). Despite the similarity in the timing of peak settlement, the intensity often varied mark- edly among sites. Comparisons Among Levels The percentage of time that plates were submerged dif- fered significantly among levels (p < 0.001). During the settlement season, the highest intertidal plates (LV1) were submerged 28% of the average tidal cycle whereas the sub- tidal plates (LV4) remained submerged throughout the average tidal cycle. Intertidal plates at LV2 and LV3 were submerged 52% and 72% of the tidal cycle, respectively. From 1982 to 1984, when all 4 levels of the water column were monitored, spat settlement was always lowest at the high intertidal LV1 (Fig. 7) and the average settle- ment at this level for all years combined was 2.5% of the total spat settlement. Annual average settlement was lower (p < 0.05) at LV1 for most combinations of site and year. Average settlement was similar at the other three levels during this period and averaged 27% at LV2, 38% at LV3, and 32% at LV4. The amount of sediment and the extent to which algae and barnacles covered collection plate surfaces were variable in time and space, but LV3 and especially LV4 were most affected. Assuming that settlement on plates is solely a function of submergence time, it is possible to calculate an expected proportion of settlement occurring at each level based on the percentage of the tidal cycle that each level is sub- merged. For the period 1982 through 1984, we would pre- dict that 11.1% of the total settlement would occur at LV1, 20.6% at LV2, 28.6% at LV3, and 39.7% at LV4. Corre- OYSTER SETTLEMENT IN A HIGH SALINITY ESTUARY 333 4 HAE HT! : vegpagaah ec ae — < ip) 10 oe oO > 30 Ww c — = a = Ww = 13 (oon Sn SE Sn LE Se SS en Sn ae oe oe oe oe oe ee on on oe oe oe ee oe oe oe eo ee oe ee oe oe el 400 pl 485 + 68 ; Zt es 70 uw e = & 200 WwW o 77) _ 150 « ” c Ww $ 100 5 z z a 4 Nha Ae Tee See TEI te PRE ae ON mw J Uw AS ON MJ SUA SO 1982 1983 1984 1985 1986 Figure 3. Mean spat density (+1 S.E.) for all sites and levels combined and corresponding surface water temperature and salinity values for each of the five settlement seasons. Water temperature and salinity values indicate the mean and range based on daily measurements made at 1000 hours during the biweekly collection period. OYSTER LANDING OLD MAN TOWN CREEK 12000 12000 12000 a 10000 _| 10000 10000 =z c = = go00 8000 8000 - z lw = 6000 4 6000 6000 | F Ww wn ° 4000 4000 4000 20 ppt (Bahr & Lanier 1981, Burrell 1986 for reviews for C. vir- ginica; Stanley & DeWitt 1983, Mulholland 1984 for Mer- cenaria). Frey and Howard (1969) and Wiedemann (1972) report M. mercenaria as the most characteristic member of the shallow subtidal to lower intertidal zone from quater- nary and tertiary estuarine deposits in Georgia. They de- scribe the ‘oyster biocoenosis,’’ which is dominated by C. virginica, as intertidal but extending somewhat into the subtidal. Hence, the local distributions of these species typ- 347 ically overlap, but to my knowledge there have been no studies on living populations where their distributions have been concurrently described for the same estuary. Local distributions of both species can be affected by numerous factors. Predation, which can be mediated by physical en- vironmental factors is probably of major importance. Tidal range and salinity mediate predation on C. virginica (Lunz 1943, Marshall 1954, Menzel & Nichy 1958, Menzel et al. 1966). Sediment type affects predation success on Merce- naria (Castagna & Kraeuter 1977, 1981, Arnold 1984). However, there is no good understanding of the relative importance of the environmental factors involved. Crassostrea virginica reefs predominantly occur only in the intertidal zone along the South Atlantic US coast (Bur- rell 1986). And as mentioned above, tidal range/predation interactions may largely explain this distribution. However, the relationships (if any) between areal coverage of, and variations in abundances on, these reefs relative to tidal range is unknown (Bahr & Lanier 1981). Such knowledge would be a necessary step in determining how tidal range and associated tidal currents are related to reef production on a regional scale. Along the South Atlantic US coast, oyster reefs have been most-studied in Georgia and South Carolina. There is very little quantitative information, especially on areal coverage, from areas north and south (see reviews by Bahr & Lanier 1981, Burrell 1986). Hence, regional comparisons of extant (or historical; e.g., Harris 1980) reef variations are not possible. 348 GRIZZLE Mercenaria abundances have been correlated with sedi- ment type, but there is no consistently reported relation be- tween the two (Stanley & DeWitt 1983, Mulholland 1984, Walker & Tenore 1984). This suggests that other factors such as water currents (Wells 1957) which are related to sediments, but typically not in a simple fashion in coastal waters (Ashley & Grizzle 1988), generally may be more important. It is also possible that sediment characteristics other than the gross ones (e.g., grain size distribution) usually measured may be most important in affecting abun- dances. Nonetheless, the effects of sediments on some predators have been experimentally demonstrated (e.g., Arnold 1984), and sediment alterations to control predation are a common practice in Mercenaria culture (Castagna & Kraeuter 1981, Manzi 1985). Mercenaria abundances also have been correlated to macrophyte cover, but there is no apparent consistent relationship (cf. Allee 1923, Peterson 1982, 1986, Peterson et al. 1984). It will probably require experimental investigations of many combinations of envi- ronmental factors and predators to obtain a detailed under- standing of the relationships involved. Hence, there is no overall understanding of how sediment characteristics, macrophytes, predators and associated factors such as tidal currents and tidal range, affect distribution and abundance. The objectives of the present report are: (1) to describe the distribution and abundance patterns of C. virginica and Mercenaria in a coastal lagoon in northeastern Florida; and (2) to compare these patterns to variations in environmental factors known to affect them. STUDY AREA AND METHODS Sampling sites were in the northern Mosquito Lagoon within the boundaries of the Canaveral National Seashore park (Figs. 1, 2 and 3), and all field work was done in July and August 1987. The Mosquito Lagoon is a shallow coastal lagoon in northeastern Florida, and it is part of the Indian River lagoonal system that extends about 220 km from the Mosquito Lagoon south (see special issue of the Florida Scientist [Vol. 46, No. 3/4] for studies on the la- goonal system). The study area contains about 100 man- grove (Rhizophora mangle and Avicennia germinans)-dom- inated islands, some with low-elevation uplands (Figs. 1, 2 and 3). Water depths are <1 m in most areas, and there are expansive seagrass flats dominated by Halodule wrightii. Water >1 m deep is largely restricted to the vicinity of the dredged intracoastal waterway along the western side of the lagoon, and a combined dredged and natural channel along the eastern side (Fig. 3). There are few surface freshwater discharges to the study area so salinities average near oce- anic concentrations, and at times are hypersaline (see Re- sults section). The nearest oceanic inlet is Ponce De Leon inlet 15 km north of the study area; there is no direct oce- anic connection to the south in the Lagoon proper. Thus, solar/lunar tidal influences decrease from north to south, and are negligible in the extreme south end of the study area (see Results sections). The mean tidal range at Ponce de Leon inlet is about 0.75 m (NOAA 1986). Field Methods Oyster reefs were censused using low-altitude color in- frared photographs (scale 1:12,000) taken in 1984 (avail- able from the Biomedical Office, Kennedy Space Center, FL 32899). I visited nearly all individual reefs identified from the aerials, and on some made visual estimates of the areal coverage of live vs. dead portions. Only reefs with living oysters were mapped. I sampled 45 reefs by counting all live oysters >5 cm length within each of two or three replicate 0.25 m? quadrats tossed onto the reef in an area where approximately maximal densities of live oysters oc- curred. Hard clams were sampled with rakes using a stratified random approach. The study area was systematically di- vided into 86 grid blocks with their boundaries drawn at 0.5-minute intervals of latitude and longitude. I randomly determined sampling locations within each block by navi- gating the boat to a haphazardly chosen area and blindly tossing a quadrat from the boat. This approach combines systematic and random sampling, and is a special case of stratified random sampling where each block is a stratum (Cochran 1977, p. 89, 205). Sampling locations were de- termined using Loran-C which had been calibrated to a known fixed point. I occasionally checked the Loran readings against coordinates of prominent map features, and in all cases they agreed to within a few hundreths of a minute. In waters <1 m deep, I thoroughly raked the sediment in two to six 1.0 m* quadrats within each block using a scratch rake, and removed all clams encountered. In deeper waters, I used a Shinnecock rake and estimated the area raked from the linear distance moved by a float attached by line to the head of the rake. At each of the deep-water sites I raked a minimum of | to 2 m? of sediment area. A total of 262 quadrats was sampled; of these 239 were sampled using the scratch rake. The teeth on both rakes were about 2 cm apart, so the smallest clams probably adequately sam- pled were 30 to 40 mm length; hence, only clams >35 mm length were recorded. For each sampling unit (quadrat), I characterized the gross sediment type in the field by touch and visual inspec- tion, and assigned it to one of four categories: sand (firm deposits with very little shell or silt-clay), sand/mud (soft deposits consisting of various mixtures of sand and silt-clay and little or no shell), shell/sand (*‘sand’’ with substantial amounts of shell), or shell/sand/mud (‘‘sand/mud’’ with substantial amounts of shell). Water depth was measured to the nearest 0.1 m, and macrophyte cover within each quadrat and the immediate area was noted. No quadrats were raked that were entirely on a seagrass-covered bottom because State of Florida regulations existing at the time of the field sampling prohibited raking in seagrasses. In all CRASSOSTREA AND MERCENARIA DISTRIBUTIONS 349 cases where the quadrat landed in such an area, it was moved to the nearest edge of the seagrass bed, or the nearest bare spot within the bed. If no macrophytes were within the quadrat or immediately adjacent to it, a value of “‘none’’ was recorded. Otherwise, either ““seagrasses’” or “*macroalgae’’ was recorded. Relative differences in tidal range were estimated at ten sites (Fig. 1). I pushed a PVC pipe marked at l-cm in- tervals into the sediment at each site, setting the initial water level at “‘zero.’’ I recorded water level changes at | to 2.5-hour intervals for 8 to 11 hours at these sites on | day. Water temperature and salinity data reported herein are from an ongoing monitoring program of Florida’s Depart- ment of Natural Resources. Eighteen sites scattered throughout the study area have been sampled since January 1985, with near-surface and near-bottom samples taken on most occasions. For the present study, data from January 1985 to August 1987 (n = 24 to 28 per site) were ana- lyzed. Data Analysis Methods Areal coverages for C. virginica reefs (Fig. 1) were de- termined by two methods: digitization of a 1:24,000 scale map into an ERDAS geographic information system with 0.05-ha resolution; and polar planimetry in combination with direct measurements using calipers for those reefs with rectangular or circular shapes, of photographic en- largements of a 1:24,000 map. Planimetry was also used to determine areal coverages for “‘open waters,”’ (= intertidal and subtidal areas not covered by mangroves) and islands (=intertidal and upland areas covered primarily by man- groves or upland vegetation). These areal coverages only were determined for those areas within a 0.75-km radius of each of the 10 tidal elevation sites shown on Figure | so quantitative comparisons could be made between tidal range and oyster reef coverage (Fig. 4). I analyzed spatial patterns in Mercenaria abundances by combining individual sampling blocks (or “‘strata,’’ see above) into larger strata based on location within the overall study area. The study area was divided into strata of approximately equal size along its length (N—S) and across its width (E—W). Five N-—S strata were defined, each con- taining 43 to 64 sampling quadrats, and three E—W strata were defined, each containing 80 to 92 quadrats. Fre- quency distribution plots of the untransformed clam data showed strong departures from normality because of the large number of quadrats where no individuals were col- lected. No transformations alleviated this problem, so the nonparametric Kruskal-Wallis test (PROC NPARIWAY, SAS 1982) was used to examine the clam density data for main effects (i.e., differences among strata or environ- mental parameters). Where significant main effects were found, nonparametric multiple comparisons were made using the ‘‘Q test statistic’’ (Zar 1984, p. 200—201). Because an inspection of Figure 3 indicated that Mer- cenaria densities might be related to distance from channels (i.e., deep-water areas where most water transport occurs), I produced a data set consisting of hard clam den- sity vs. distance to the nearest channel to which there was direct water access. These data were analyzed using the Kruskal-Wallis test. RESULTS Crassostrea virginica and Mercenaria Distributions and Abundances There was a total of 31.7 ha of living oyster reefs in the study area. The total study area was about 5,200 ha, hence oyster reefs occupied 0.61% of the lagoonal system. The amount of bottom area covered by C. virginica reefs de- creased from north to south (Fig. 1). These oyster reefs only occurred intertidally (Fig. 2), and they were primarily in the vicinity of waters >1 m deep where tidal currents are presumably greatest (compare oyster data on Fig. 1 with dashed lines showing channels on Fig. 3). Thus, oyster reefs were absent from the shallow waters of the north-cen- tral part of the study area, and in the southern half. Individual reefs ranged in size from a few square meters bottom area coverage, to ~6,000 m?. These ‘“‘living’’ reefs ranged from ~100% coverage by live oysters to those with <50% live oysters. Forty-five 0.25 m? quadrats from areas with live oysters present, and on 20 different reefs, had densities of 12 oysters (>5.0 cm shell length) m~? to 132 m~*, with a mean of 57.6 m~2. Plots of oyster densities on the reefs by location relative to N—S and E~W showed no trends. Figure 3 shows the abundances of Mercenaria in most of the 262 quadrats raked; in some cases data from two quadrats were combined because they were located very close to one another. However, in all statistical analyses data from individual quadrats were used. Generally, clams were only found in the vicinity of waters >1 m deep, ex- cept in the deeper waters in the extreme south end of the study area where none was collected. Kruskal-Wallis tests showed no significant differences (P = 0.2976) among the five N-—S strata, but significant differences (P = 0.0002) among the three E—W strata (Table 2). The densities in strata W (0.96 clam m~?) and E (0.56 clam m~2) were significantly higher than in stratum C (0.22 clam m~?). Environmental Factors Tidal ranges decreased from north to south (Table 1). Site 1 at the north end of the study area had a range of 50 cm, contrasted to a range of 2 cm at site 10 in the south end. Sites between these two had intermediate ranges and reflected the overall trend. The tidal range measured at site 1 was about two-thirds the mean tidal range at Ponce de Leon Inlet 15 km to the north. The 1-day measurements presented herein are meant primarily to indicate relative differences between sites. My observations over several weeks in the study area corroborated the north-south de- OYSTER REEFS Atlantic Ocean water elevation A creasing trend and the relative magnitudes indicated by these data. Temperature and salinity data showed very little spatial variations in the study area. Mean salinities from eighteen sampling sites during 1985—1987 ranged from 30.7 to 31.3 ppt, with a range of individual measurements from all sites of 22.1 to 40.0 ppt. All sites had very similar minimum and maximum values, and the measurements on any given day never varied among the sites by more than 3 ppt. Water temperatures ranged from 12.3°C to 32.6°C, with little among-site differences. Sediment types ranged from mixtures of soft sand/mud to firm sand/shell. Most recorded water depths were <0.5 m, and the maximum depth sampled was 2.5 m. The seagrass Halodule wrightii was abundant throughout the study area. Other seagrass species including Ruppia maritima and Syr- ingodium filiforme were also observed, but they only oc- curred in a few areas and H. wrightii was typically much more abundant in these areas. Bivalve Distributions Relative to Environmental Factors Because salinity and temperature variations in the study area showed only small spatial differences, they were not GRIZZLE = vt aS SPSS : fax meet Sage = ttter Pr considered to be important in explaining distribution and abundance patterns of C. virginica and Mercenaria. They are not discussed further in this section. Field observations at low tide of nearly all the reefs shown in Fig. | indicated that C. virginica is restricted to the intertidal zone (Fig. 2). Bottom areal coverage by C. virginica reefs was significantly correlated positively with tidal range (Fig. 4). In the vicinity of the site with maximal tidal range (50 cm), reef bottom areal coverage was 1373 m? ha~! (= 13.73% coverage; Table 3) of open water area, and there were no oysters in the vicinity of the three sites in the south end where the tidal range was <5 cm. However, there was no geographical trends in oyster densities on these reefs, so there was apparently no relationship be- tween oyster abundances on each reef and tidal range. As mentioned above, Mercenaria abundances showed no significant differences among the five N—S strata, and there were no trends evident. So there was no relationship between tidal range and overall clam abundance patterns. Kruskal-Wallis tests on the total data set showed significant differences relative to water depth (P = 0.0109). Q-tests showed the significance relations indicated in Table 2; the greatest abundances (1.71 clams m~?) were collected in ES INTRA RaO Tag CRASSOSTREA AND MERCENARIA DISTRIBUTIONS 351 Spe ; * ‘ = \ Sines Slough c ° } ° \ o o 4 ; ° = a oo ” °o = oc & NYRACOASTA WATERWAY = 3. e ee ° 4 35 Figure 1. Crassostrea virginica reefs. Hatched lines delimit boundaries of area where oyster reefs were censused. Sites numbered 1 to 10 are where tidal ranges were measured. Lettered (A—E) arrows indicate position and direction of view for photographs in Figure 2. water depths from 1.0 to 1.5 m. There were no differences relative to macrophyte cover (P = 0.6834). There was no significant difference (P = 0.0521) among sediment types, but trends were evident. The highest abundances occurred in sediments with substantial amounts of shell (shell/sand had 1.0 clam m~?) and the lowest (0.24 clam m~?) were in the soft mud/sand (Table 2). Plots of the total data set on Mercenaria abundances showed a general decreasing trend with increasing distance to a channel (Fig. 5). After grouping the quadrats into seven distance intervals, a Kruskal-Wallis test showed that there were significant differences (P = 0.006) in clam abundances among intervals (see inset of Fig. 5 for plots of the means from these distance intervals). Because of this relationship, subsets of the total data set consisting of only those quadrats within various distances of a channel were analyzed for differences among environ- mental variables. This was done in order to eliminate those sites distant from a channel that would have generally had very low numbers of hard clams present regardless of sedi- ment type, water depth, or macrophyte cover. In a sense, this is a recognition of the different spatial scales appro- priate for each environmental variable. The variable *‘dis- tance to the nearest channel’’ can be a mesoscale (distances of a kilometer or more) factor, while the other variables (sediment type, water depth, macrophytes) act at smaller spatial scales (see below for more explanation and discus- sion of possible collinearity). Another mesoscale factor is location within the lagoon relative to N—S or E—W. As indicated above, there were significant differences among the three E—W strata. So the effects of the three environ- mental factors on clam abundances were also analyzed after sorting the data by E—W strata. I “‘combined’’ the effects of the two mesoscale variables (distance to nearest channel, and E—W strata) by analyzing clam abundances (relative to the environmental variables) of only those quadrats within various distances from a channel, and by E—W strata. Kruskal-Wallis tests on the total data set after deleting quadrats from various distances to the nearest channel, showed similar significance relationships for each of the three environmental variables to those in part 2A of Table 2. Analysis after sorting by E—W strata, however, showed 352 GRIZZLE Figure 2. Crassostrea virginica reefs. Locations of reefs are keyed by letter (A—E) to Figure 1. All photographs were taken on 31 July 1987 and within about 1.5 hours of low tide, except E. A. Oysters among mangroves; note reef in background. B. Extensive reefs extending southeasterly at the outside bend of a major tidal channel and south of its confluence with a man-made canal. C. Reefs extending southerly from the hydraulic connection to a major tidal channel. D. Large arching reef on edge of a major tidal channel; note area of sediment and dead oysters in center. E. Extensive reefs extending northerly from near a major hydraulic connection to a tidal channel; this photograph was taken at about mid-tide. significant differences for water depth and sediment type in stratum W, and macrophyte cover in stratum E (results not shown herein). The lowest P values and greatest differences among means resulted from an analysis of a data set consisting of only those quadrats within 850 m of the nearest channel, and sorted by E—W strata (part B, Table 2). In this data set there were significant differences in clam abundances at different water depths and in different sediment types in stratum W, the stratum with the highest overall mean abun- dance of hard clams (0.96 clam m~?; part 1, Table 2). And there were significant differences in clam abundances at different macrophyte coverages in stratum E, the stratum with the second highest overall clam abundance (0.56 clam m*; part 1, Table 2). The environmental variables showed no significant effects in stratum C, which had the lowest overall mean abundance of hard clams (0.22 clam m~?; part 1, Table 2) and had the most quadrats distant from a channel. These findings suggest that the relationship be- tween environmental variables and clam abundances is cer- tainly not simple, with location affecting the effect of each variable. Further, plots (not presented herein) of all pair- wise combinations of the environmental variables showed some linear trends (collinearity), which would make a straightforward interpretation of these results impossible. Nonetheless, some possible explanations can be offered. DISCUSSION Both oysters and clams were rare in the shallow, sea- grass-covered areas in the north-central part of the lagoon, and both were largely restricted to the vicinity of channels and other deep-water areas where most water transport Mosquito Lagoon DNUDIIYW ° ° @ ° > 353 occurs and water currents are probably strongest. Oysters were only common in the northern half of the study area. Clams showed no north-south trends, except they were not collected in the deeper waters in the extreme south end. The distribution and abundance of both were related to sev- eral environmental factors. Crassostrea virginica and Environmental Factors As in other areas of the South Atlantic US, Crassostrea virginica reefs in the northern Mosquito Lagoon were re- stricted to the intertidal zone. Reefs occupied about 0.61% of the total lagoonal system. Previous studies in other areas showed that oyster reefs occupy from <1.0% (Bahr 1976) to 9% (McKenzie & Badger 1969) of an entire estuarine system (see Bahr and Lanier 1981 for review). Such vari- ability probably reflects the variability in environmental characteristics among estuarine systems that determine oyster abundance and distribution patterns, as well as dif- ferences in methods used to measure reef coverage. For among-system comparisons it might generally be more useful only to consider those parts of the lagoon or estuary potentially inhabitable by oysters. Based on data from the vicinity of the seven water elevation sites where oysters occurred (Fig. 1; Table 3), and when considering only ‘‘open waters’’ (i.e., all water-covered areas not occupied by mangroves) as potential oyster habitat, reefs occupied an average of 5.6% of the bottom area. This figure is in the range of previously reported reef coverages (see above), and indicates the potential ecological importance of oyster reefs in estuarine systems generally (Bahr and Lanier 1981). Several environmental factors affect the extent of oyster reef coverage in an estuary. The Habitat Suitability Index (HSI) model developed by the US Fish and Wildlife Ser- vice for Crassostrea virginica along the Gulf Coast of the US includes eight variables that probably cause changes in the distribution and abundance of this species (Cake 1983, Soniat & Brody 1988; also see Sellers & Stanley 1984, Burrell 1986, Stanley & Sellers 1986a, b for reviews of the environmental requirements of C. virginica). In the south- eastern US tidal range is clearly of major importance be- cause oysters are largely restricted to the intertidal zone (Bahr & Lanier 1981, Burrell 1986). In the present study the percent areal coverage of oyster reefs was significantly correlated linearly with tidal range (Fig. 4). To my knowl- edge, no previous studies have attempted to correlate tidal range with oyster reef coverage (also see Bahr & Lanier 1981, p. 62). Obviously vertical limits of such reefs are Figure 3. Mercenaria abundances at all sites sampled, except some data combined (see text). Dashed lines show major channels where most water transport occurs. Waters south of oblique dashed line at southern end of study area and east of dashed line to the west are >1.5 m deep. os) n rs 6 y = 0.30X - 1.43 Zs r = 0.9739 P<0.001 2 n=10 Oyster Reef Coverage (% of area) 10 20 30 40 50 Tidal Range (cm) Figure 4. Relationship between tidal range at ten sites shown on Fig. 1 (Table 1) and percent coverage of open water area by Crassostrea virginica reefs within a 0.75-km radius of each site (Table 3). affected by tidal range, and horizontal aspects such as areal coverage may simply be a reflection of the horizontal areal extent of the intertidal zone being related to tidal range. Hence it is possible that the relation shown in Figure 4 is trivial in the sense that it only indicates that tidal range and intertidal bottom area are correlated. For the present study area, there is no quantitative information on the extent of intertidal bottom areas. It is also possible that the generally stronger tidal currents in areas with greater tidal ranges contribute to greater reef coverages via their positive ef- fects on various pre- and post-settlement processes com- pared to areas with smaller tidal ranges (Lund 1957, Bahr & Lanier 1981, Bushek 1988). In any case, the relationship indicated by Figure 4 suggests that tidal range could gener- ally be a useful predictor of areal coverages of oyster reefs, even though its mechanistic role in this respect is not un- derstood. As mentioned in the Introduction section, tidal range probably acts as a mediator in controlling predation on the oyster, thereby limiting its occurrence in high-salinity waters of the southeastern US to the intertidal zone. Lunz (1943) stated that damage from the boring sponge, Cliona celata, is mainly restricted to subtidal areas. Marshall (1954) showed that in Alligator Harbor (on the Gulf Coast of Florida) where oysters were restricted to the intertidal zone, unprotected oysters placed subtidally had 91% mor- tality compared to 15% for caged individuals over a 2- month period. Menzel and Nichy (1958) and Menzel et al. (1966) also provide data supporting the importance of pre- dation/tidal range/salinity interactions in restricting C. vir- ginica reefs to the intertidal zone in their study areas. The relationship may be simple, but it may also involve many factors (e.g., disease, competition, food availability) yet to be adequately studied (cf. the complexity of interactions GRIZZLE between environmental factors involved in community zon- ation patterns in rocky intertidal areas; see review by Un- derwood and Denley 1984). Burrell (1986) provides a brief review of several hypotheses that need to be tested. Mercenaria and Environmental Factors In contrast to Crassostrea virginica, Mercenaria distri- bution and abundance patterns showed no north-south trends in the study area, except they were absent from the deeper waters in the extreme south end (Fig. 3; Table 2). Thus, there was no correlation between hard clam abun- dances and tidal range. Mercenaria abundances decreased with increasing distance to the nearest channel (Fig. 5), and there were consistent trends as well as instances of signifi- cant differences in clam abundances at different water depths and in different sediment types. These results sug- gest a complex situation with respect to distribution pat- terns. Factors affecting settlement (e.g., hydrodynamical conditions), and post-settlement factors (e.g., predation) are probably important (Peterson 1986, Bushek, 1988). The significant differences among Mercenaria abun- dances at different distances to the nearest channel seems most reasonably explained as a result of hydrodynamical influences, although alternative explanations can not be ruled out. It is not uncommon to find pronounced differ- ences in hard clam abundances related to variations in water flows on small spatial scales; e.g., high abundances in different sections of tidal creeks (Walker & Tenore 1984), or along the edges of small peninsulas (Wells 1957). However, I am not aware of reports showing a relationship on the scale reported in the present study. In an extensive review of the hydrodynamical processes pertinent to spatial patterns for benthic invertebrates, Butman (1987) empha- sized that different spatial scales are appropriate for dif- ferent processes. With respect to passive deposition of larvae, spatial scales of tens of meters to tens of kilometers are probably most appropriate, whereas scales of centimeters to meters pertain to active habitat selection (which may in- TABLE 1. Tidal ranges measured over one tidal cycle at ten sites (see Fig. 1). ‘Total Hours’’ gives total number of hours during which water levels were monitored. Site Total Hours Tidal Range (cm) 1 10.2 50 2 10.1 37 3 10.5 25 4 11.0 19 5 10.2 14 6 10.0 19 7 8.0 4 8 8.0 4 9 8.0 3 10 8.1 2 CRASSOSTREA AND MERCENARIA DISTRIBUTIONS 355 TABLE 2. Mercenaria abundances (Number m~?) by geographic location and environmental variable. N-S Strata = sampling strata from north to south: 1 is northernmost part of study area, 5 is southernmost. E-W Strata = sampling strata from east to west: E is easternmost, C is central area of lagoon, W is westernmost. P values are shown for Kruskal-Wallis tests on each main effect. The ‘‘Q test statistic’’ (Zar 1984) was used for multiple comparisons if there was a significant (P < 0.05) main effect, and the significance relations are shown by lower case letters following each mean; those with same letter were not significantly (P > 0.05) different, those with different letters were different (P < 0.05). Mean Clam Abundances by Location Var:N-S Strata! n Mean Var:E-W Strata? n Mean 4 43 0.79 WwW 90 0.96a I 64 0.73 E 82 0.56a 3 51 0.55 c 90 0.22b 2 53 0.42 5 5 0.41 1P = 0.2976 2P = 0.0002 Mean Clam Abundances by Environmental Variables A. Total Data Set Var: Water Depth? n Mean’ Var:Sediment* n Mean 1.lto1.5m 7 1.7la shell/sand 20 ~—-1.00 1.5m 18 0.00 b mud/sand SI 024 3P = 0.0109 *P = 0.0521 Var:Macrophyte Covers n Mean none 60 0.67 seagrasses 166 0.58 macroalgae 33 0.42 5P = 0.6834 B. By E-W Strata and Without Quadrats >850 m From Channel Easternmost Stratum (E) Var: Water Depth® n Mean’ Var:Sediment? mn Mean 1.5m 1 0.00 shell/sand g) 0.22 © P = 0.5412 7P < 0.4999 Var:Macrophyte Cover® n Mean macroalgae U 1.00 a seagrasses 47 0.70 a none 22 0.23 a 8 P = 0.0312; 0.05 < P < 0.10 for Q-test of 1.00 vs. 0.23 Central Stratum (C) Var:Water Depth? n Mean’ Var:Sediment!? n Mean 1.1 to 1.5m 2 1.00 mud/sand 15 0.40 0.6 to 1.0m 17 0.29 sand 34 0.29 1.5m 6 0.00 shell/mud/sand 7 0.00 9 P = 0.3884 10 Pp = 0.5098 TABLE 2. Continued Var:Macrophyte Cover!! n Mean none Il 0.36 seagrasses 42 0.31 macroalgae 8 0.00 1 P = 0.4139 Westernmost Stratum (W) Var:Water Depth’? n Mean Var:Sediment® n Mean 1.1 to 1.5m 3 3.00 a shell/sand 5 3.40a 1.5m ll 0.00 a mud/sand 24 0.13 be 12PF—10 10227: 5 P = 0.0002 0.05 < P < 0.10 for 3.00 vs. 0.00, and 1.23 vs. 0.00. Var:Macrophyte Cover'* n Mean none 21 1.43 seagrasses 43 IV) macroalgae 16 0.44 4p = 0.1115 volve bottom sediment characteristics, bottom roughness elements such as macrophytes, etc.). In the present study area, most water transport probably occurs in the channel areas. Processes such as growth and production rates (Wil- dish & Kristmanson 1979, 1984, Grizzle & Lutz 1989), and larval dispersal rates may be expected to be increased in these areas. Hence, it may be expected that water move- ments will have a generally positive (to some point) effect on clam densities. Wells (1957) reported a positive correla- tion between tidal current speed and Mercenaria densities. I suggest that the negative relationship between Mercenaria densities and distance to the nearest channel (Fig. 5), which is on a scale of kilometer(s), can best be explained as a hydrodynamical influence on larval dispersal, settlement, and perhaps post-settlement effects including predation and disturbance. The effects of sediment type, water depth, and macrophyte cover (which may affect larval dispersal and predation on recruited clams; see discussion below) prob- ably act on smaller spatial scales. Thus, their relationship to clam densities should, in addition to a lagoon-wide anal- ysis, be examined after considering the effects of “‘distance to the nearest channel’’ or geographic trends, as mentioned in the Results section. The highest (though non-significant, P = 0.0521; part 2A, Table 2) average abundance of Mercenaria when con- sidering the total data set, occurred in shell/sand sediments (1.00 m~). Also, when considering only those sampling sites within 850 m of a channel, in sampling stratum W there were significantly higher densities of clams (3.40 m?; part 2B, Table 2) in shell/sand. It is well established that sediment characteristics affect predation success on the hard clam, and thus potentially affect distribution and 356 GRIZZLE Mercenaria (No. m~2) O 250 500 CO) Distance to Nearest Channel 500 1000 1500 ® @=27105 @=6T010 @=>10 1000 1250 Cm) 1500 Figure 5. Relationship between distance to nearest channel (see dashed lines on Fig. 2) and Mercenaria abundances for total data set. The size of the circle is relative to the number of observations it represents, as indicated. Inset shows means with 95% confidence intervals for data in larger figure combined into seven distance intervals. Legends for both axes are as in larger figure, but y-axis scale is expanded in inset. abundance patterns. Arnold (1984) showed experimentally that Callinectes sapidus (blue crab), a major predator on hard clams (MacKenzie 1977) and abundant in the northern Mosquito Lagoon (pers. obs.), consumed significantly TABLE 3. Areal coverages for Crassostrea virginica reefs, ‘‘open water’’ (no mangroves), and islands within 0.75-km radius of each of the ten tidal range sites shown on Figure 1. Percent of Open Water Islands Open Water Reefs Covered by Site (ha) (ha) (ha) Oyster Reefs l 33.74 33.51 4.60 13.73 2 $2.63 52.07 5.68 10.91 3 52.78 35.88 2.26 6.30 4 23.54 76.20 3.68 4.83 5 63.13 113.58 1.35 1.19 6 32.31 72.22 1.34 1.86 7 30.12 104.82 0.28 0.27 8 43.41 133.30 0 0 9 5.16 121.12 0 0 10 4.06 109.57 0 0 lower amounts of hard clams in shell or gravel substrates compared to sand/mud and sand. Several descriptive/corre- lative studies have shown high densities of hard clams in shelly deposits (Pratt 1953, Wells 1957, Godwin 1968, Walker & Tenore 1984). The use of gravel or small stone aggregates in Mercenaria culture to inhibit predation is a standard practice (Castagna & Kraeuter 1981). However, recent unpublished work in the Indian River region of Florida does not show enhanced abundances of Mercenaria in shelly deposits (Dan Marelli, pers. comm.). Nonethe- less, although other factors (e.g., settlement success) may contribute to the increased abundances of hard clams in shelly or gravel sediments (see discussion in Arnold 1984), avoidance of predators is a reasonable explanation. With respect to water depth, the significantly highest densities of clams occurred at depths of 1.1 to 1.5 m, when considering both the total data set (part 2A, Table 2), and after eliminating quadrats distant from a channel and sorted by E—W strata (part 2B). Because water depths of >1 m generally occurred near channels, this finding could be ex- plained by the clam abundance/distance relationship shown in Figure 5 and discussed above. Nonetheless, human pre- dation effects could also contribute to the increased abun- CRASSOSTREA AND MERCENARIA DISTRIBUTIONS 357 dances at 1.1 to 1.5 m water depths. Nearly all clam har- vesting in the Mosquito Lagoon is done using scratch rakes (pers. obs.), which are most effective in waters less than about 1.0 m deep (see Wells 1957 for similar observation in Chincoteague Bay, Maryland). Thus, if human predation affects hard clam densities, clams should be most abundant in waters >1 m deep. However, I harvested no clams from the waters in the southern end of the study area which were generally >1.5 m deep (Fig. 3). Personal communications with fishermen in the area indicated that clams have not occurred in these deeper waters for many years. Hence, human predation probably does not contribute to their ab- sence from waters depths >1.5 m, and there is no informa- tion to explain their absence from such waters. The final environmental variable measured, macrophyte cover, did not show as consistent an effect on clam den- sities as the above two variables. There were no significant differences (P = 0.6834, part 2A, Table 2) in hard clam abundances among the three classes of macrophyte cover when considering the total data set. However, there were significantly (P = 0.0312) higher abundances in macroal- gal-covered areas in sampling stratum E (part 2B, Table 2). Generally, macrophytes can be expected to enhance benthic recruitment due to their effects on near-bottom flows (e.g., Eckman 1983, Peterson et al. 1984, but see Olafsson 1988), and/or post-settlement processes such as enhanced survival (Peterson 1986). The absence of a general trend in the present study may reflect the fact that sampling within sea- grass beds proper was not accomplished. In conclusion, area-wide (mesoscale) Mercenaria distri- bution and abundance patterns in the northern Mosquito la- goon are perhaps largely determined by hydrodynamical factors which affect various pre- and post-settlement pro- cesses, and/or by other factors related to geographic loca- tion in the lagoon. Local (less than mesoscale) distribution patterns are probably influenced by sediment- and depth- mediated predation pressures. Sediment characteristics af- fect invertebrate and fish predators, and water depth affects human predation. Such an explanation is obviously specu- lative, and as with most descriptive/correlative studies, re- quires experimental testing. ACKNOWLEDGMENTS I thank Richard Hanks, Art Graham, Linwood Jackson, and Norbert Psuty for their assistance in various adminis- trative aspects of the study. Linda Harnden drafted Figures 1 and 3, and assisted in various data analysis and report preparation tasks. John Breen and Steve Kloster were very helpful on a day-to-day basis during the field work. Bud Dewees kindly provided dockage for my boat, as well as much information on the Mosquito Lagoon. I am indebted to him and others at LeFils Fish Camp for their help and hospitality. Jane and Mark Provancha made available the aerial photographs used in mapping the oyster reefs, and they reviewed the manuscript. They also provided useful information concerning the study area as well as a conge- nial working environment. Jane digitized the oyster reef map and determined areal coverages. Chris Adamus pro- vided salinity and temperature data. Bill Arnold and Dan Marelli assisted in design of the sampling program; Dan, and Roger Newell, provided a detailed review of the manu- script. Fred Short suggested that the data set on clam abun- dance vs. distance to channels be produced. Two anony- mous reviewers helped improve the manuscript. Finally, I thank Nathan and Darren Walsh and Shelley Grizzle for their help. The study was funded by the US National Park Service, the Jackson Estuarine Laboratory of the University of New Hampshire, and Livingston University. REFERENCES Abbott, R. T. 1974. American Seashells. Van Nostrand Reinhold Co., New York. 663 p. Allee, W. C. 1923. Studies in marine ecology: I. The distribution of common littoral invertebrates of the Woods Hole region. Biol. Bull. 44:167-191. Andrews, J. 1971. Sea Shells of the Texas Coast. University of Texas Press, Austin and London. 298 p. Amold, W. S. 1984. The effects of prey size, predator size, and sediment composition on the rate of predation of the blue crab, Callinectes sa- pidus Rathbun, on the hard clam, Mercenaria mercenaria (Linne). J. Exp. Mar. Biol. Ecol. 80:207—219. Ashley, G. M. & R. E. Grizzle. 1988. Interactions between hydrody- namics, benthos, and sedimentation in a tide-dominated coastal la- goon. Mar. Geol. 82:61—81. Bahr, L. M., Jr. 1976. Energetic aspects of the intertidal oyster reef com- munity at Sapelo Island, Georgia (USA). Ecology 57:121—131. Bahr, L. M., Jr. & 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. FWS/OBS-81/15. 105 p. Bousfield, E. L. 1960. Canadian Atlantic Sea Shells. Canadian Dept. Northern Affairs and National Resources, National Museum of Canada. 72 p. Burrell, V. G., Jr. 1986. Species profiles: life histories and environmental requirements of coastal fishes and invertebrates (South Atlantic) — American oyster, U.S. Fish Wildl. Serv. Biol. Rep. 82(11.57). U.S. Army Corps of Engineers TR EL-82-4. 17 p. Bushek, D. 1988. Settlement as a major determinant of intertidal oyster and barnacle distributions along a horizontal gradient. J. Exp. Mar. Biol. Ecol. 122:1—18. Butman, C. A. 1987. Larval settlement of soft-sediment invertebrates: the spatial scales of pattern explained by active habitat selection and the emerging role of hydrodynamical processes. Oceanogr. Mar. Biol. Ann. Rev. 25:113—165. Cake, E. W., Jr. 1983. Habitat suitability index models: Gulf of Mexico American oyster. U.S. Fish Wildl. Serv. FWS/OBS-82/10.57. 37 p. Castagna, M. & J. N. Kraeuter. 1977. Mercenaria culture using stone aggregate for predator protection. Proc. Natl. Shellfish. Assoc. 67:1-6. Castagna, M. & J. N. Kraeuter. 1981. Manual for growing the hard clam Mercenaria. Virginia Inst. Marine Sci., Spec. Rept. No. 249. Cochran, W. G. 1977. Sampling Techniques. John Wiley & Sons. New York. 428 p. Dillon, R. T. & J. J. Manzi. 1989. Genetics and shell morphology in a 358 GRIZZLE hybrid zone between the hard clams Mercenaria mercenaria and M. campechiensis. Mar. Biol. 100:217—222. Eckman, J. E. 1983. Hydrodynamic processes affecting benthic recruit- ment. Limnol. Oceanogr. 28:241—257. Frey, R. W. & J. D. Howard. 1969. A profile of biogenic sedimentary structures in a Holocene barmier island-salt marsh complex, Georgia. Trans. Gulf Coast Assoc. Geol. Soc. 19:427—444. Godcharles, M. F. & W. C. Jaap. 173. Exploratory clam survey of Florida nearshore and estuarine waters with commercial hydraulic dredging gear. Fla. Dep. Nat. Res., Mar. Research lab., Profess. Pap. Ser: 215 77 p: Godwin, W. F. 1968. The distribution and density of the hard-clam, Mer- cenaria mercenaria, on the Georgia coast. Ga. Game & Fish Comm., Mar. Fish. Div., Contrib. Ser. 10. 30 p. Grizzle, R. E. & R. A. Lutz. 1989. A statistical model relating horizontal seston fluxes and sediment characteristics to growth of Mercenaria mercenaria. Mar. Biol. 102:95—105. Harris, C. D. 1980. Survey of the intertidal and subtidal oyster resources of the Georgia coast. Georgia Dept. Natural Res., Brunswick, GA. 44 p. Lund, E. J. 1957. Self-silting by the oyster and its significance for sedi- mentation in geology. Pub. Inst. Mar. Sci. Texas 4:320—327. Lunz, R. G., Jr. 1943. The yield of certain oyster lands in South Carolina. Am. Mid. Nat. 30:806—808. MacKenzie, C. L., Jr. 1977. Predation on hard clam (Mercenaria mer- cenaria) populations. Trans. Am. Fish. Soc. 106:530—537. Manzi, J. J. 1985. Clam aquaculture. In: J. V. Huner and E. E. Brown (Eds). Crustacean and mollusk aquaculture in the United States. AVI Publishing Co., Westport, CT. pp. 275-310. McKenzie, M. D. & A. C. Badger. 1969. A systematic survey of inter- tidal oysters in the Savannah River basin area of South Carolina. Con- trib. Bears Bluff Lab., SC. 50:3-15. Marshall, N. 1954. Factors controlling the distribution of oysters in a neu- tral estuary. Ecology 35:322—327. Menzel, R. W. & M. Y. Menzel. 1965. Chromosomes of two species of quahog clams and their hybrids. Biol. Bull. 129:181—188. Menzel, R. W. & F. E. Nichy. 1958. Studies of the distribution and feeding habits of some oyster predators in Alligator Harbor, Florida. Bull. Mar. Sci. Gulf Caribb. 8:125—145. Menzel, R. W., N. C. Hulings & R. R. Hathaway. 1966. Oyster abun- dance in Apalachicola Bay, Florida in relation to biotic associations influenced by salinity and other factors. Gulf Res. Rept. 2:73—96. Mermill, A. S. & J. W. Ropes. 1967. Distribution of southern quahogs off the Middle Atlantic coast. Comm. Fish. Rev. 29:62—64. Mulholland, R. 1984. Habitat suitability index models: hard clam. U.S. Fish Wildl. Serv. FWS/OBS-82/10.77. 21 p. NOAA. 1986. Tide Tables 1987, East Coast of North and South America. U.S. Dept. Commerce, NOAA, NOS. Rockville, MD. 289 p. Olafsson, E. B. 1988. Inhibition of larval settlement to a soft bottom benthic community by drifting algal mats: an experimental test. Mar. Biol. 97:571—574. Peterson, C. H. 1982. Clam predation by whelks (Busycon spp.): experi- mental tests of the importance of prey size, prey density, and seagrass cover. Mar. Biol. 66:159—170. Peterson, C. H. 1986. Enhancement of Mercenaria mercenaria densities in seagrass beds: is pattern fixed during settlement season or altered by subsequent differential survival? Limnol. Oceanogr. 31:200—205. Peterson, C. H., H. C. Summerson & P. B. Duncan. 1984. The influence of seagrass cover on population structure and individual growth rate of a suspension-feeding bivalve, Mercenaria mercenaria. J. Mar. Res. 42:123-138. Pratt, D. M. 1953. Abundance and growth of Venus mercenaria and Cal- locardia morhuana in relation to the character of bottom sediments. J. Mar. Res. 12:60—74. SAS. 1982. SAS User’s Guide: Statistics. SAS Institute Inc. Cary, NC 1290 p. Sellers, M. A. & J. G. Stanley. 1984. Species profiles: life histories and environmental requirements of coastal fishes and invertebrates (North Atlantic)—American oyster. U.S. Fish Wildl. Serv. FWS/OBS-82/ 11.23. U.S. Army Corps of Engineers, TR EL-82-4. 15 p. Soniat, T. M. & M. S. Brody. 1988. Field validation of a habitat suit- ability index model for the American oyster. Estuaries 11:87—95. Stanley, J. G. & R. DeWitt. 1983. Species profiles: life histories and environmental requirements of coastal fishes and invertebrates (North Atlantic)—hard clam. U.S. Fish Wild. Serv. FWS/OBS-82/11.18. U.S. Army Corps of Engineers, TR EL-82-4. 19 p. Stanley, J. G. & M. A. Sellers. 1986a. Species profiles: life histories and environmental requirements of coastal fishes and invertebrates (Gulf of Mexico)—American oyster. U.S. Fish Wildl. Serv. Biol. Rep. 82(11.64). U.S. Army Corps of Engineers, TR EL-82-4. 25 p. Stanley, J. G. & M. A. Sellers. 1986b. Species profiles: life histories and environmental requirements of coastal fishes and invertebrates (Mid- Atlantic)—American oyster. U.S. Fish Wildl. Serv. Biol. Rep. 82(11.65). U.S. Army Corps of Engineers, TR EL-82-4. 25 p. 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, Jr., D. Simberloff, L. G. Abele & A. B. Thistle (Eds) Ecological communities: conceptual issues and the evidence. Princeton University Press, Princeton. pp. 151-180. Walker, R. L. & K. R. Tenore. 1984. The distribution and production of the hard clam, Mercenaria mercenaria, in Wassaw Sound, Georgia. Estuaries 7:19—27. Wells, H. W. 1957. Abundance of the hard clam Mercenaria mercenaria in relation to environmental factors. Ecology 38:123—128. Wiedemann, H. U. 1972. Shell deposits and shell preservation in Quater- nary and Tertiary estuarine sediments in Georgia, U.S.A. Sediment. Geol. 7:103—125. Wildish, D. J. & D. D. Kristmanson. 1979. Tidal energy and sublittoral macrobenthic animals in estuaries. J. Fish. Res. Bd. Can. 36:1197— 1206. Wildish, D. J. & D. D. Kristmanson. 1984. Importance to mussels of the benthic boundary layer. Can. J. Fish. Aquat. Sci. 41:1618—1625. Zar, J. H. 1984. Biostatistical Analysis. Prentice-Hall, Inc. Englewood Cliffs, NJ. 718 p. Journal of Shellfish Research, Vol. 9, No. 2, 359-365, 1990. PATTERNS OF MSX (HAPLOSPORIDIUM NELSONI) INFECTION AND SUBSEQUENT MORTALITY IN RESISTANT AND SUSCEPTIBLE STRAINS OF THE EASTERN OYSTER, CRASSOSTREA VIRGINICA (GMELIN, 1791), INNEW ENGLAND GEORGE C. MATTHIESSEN,! SUNG Y. FENG,?”” AND LOUIS LEIBOVITZ? 'Ocean Pond Corporation Fishers Island, NY 06390 2Department of Marine Science University of Connecticut Groton, CT 06340 3Marine Biological Laboratory Woods Hole, MA 02543 ABSTRACT The seasonal pattern of infection and subsequent mortality among eastern oysters exposed to the pathogen MSX in New England waters has not been studied intensively. In recent years, this parasite has become endemic to many oyster-producing areas in this region, frequently resulting in heavy mortalities. The levels of resistance to MSX infection between two different oyster strains, one with established genetic resistance and the other with no history of resistance, were compared on infected oyster beds in southern Massachusetts. Although MSX infections occurred among members of the relatively resistant strain, mortality rates were appreciably lower than those observed in the more susceptible group. These observations indicate the advantages of utilizing MSX-resistant strains in areas exposed to infection. KEY WORDS: Haplosporidium, MSX, Crassostrea virginica INTRODUCTION In the spring of 1985, heavy mortalities were observed among oysters that had been planted the previous year in Cotuit Bay on the south shore of Cape Cod. Histological examination of surviving oysters by the National Marine Fisheries Service Pathology Laboratory in Oxford, Mary- land established that this population was heavily infected by the protozoan parasite MSX (Haplosporidium nelsoni) (Haskin et al. 1966). It was concluded on the basis of histo- logical examinations of recently imported oysters the pre- vious fall that this pathogen probably had been introduced with oysters from the Hammonassett River near Clinton, Connecticut. By the end of the summer of 1985, total losses on the Cotuit beds were estimated at 85% (Nelson 1987). The occurrence of MSX on these or any other oyster beds along the south shore of Cape Cod had never been recorded prior to this time. In areas of New England where MSX had been reported to occur, e.g., Wellfleet Harbor on the north shore of Cape Cod and parts of Long Island Sound (Haskin 1987), mortalities had not been well docu- mented. Therefore virtually nothing was known concerning the seasonality of infection and resulting mortality in this region. *Deceased 12 December, 1989. The pathological examinations discussed in this paper were performed by Dr. Feng, whose interest and expertise contributed greatly to this investigation. His enthusiasm and thoughtful suggestions will be greatly missed by those fortunate enough to have worked with him. 359 During 1986, investigations were initiated to determine the seasonal pattern of MSX infection and subsequent oyster mortality in Cotuit Bay (Leibovitz et al. 1987). Monthly sampling from an experimental group of unin- fected oysters obtained from Fishers Island, New York and planted on the Cotuit beds in April of 1986 revealed initial infections beginning in July, followed by heavy mortalities during August and September. Separate groups of oysters from the same origin as those described above were also planted in Cotuit Bay each month during the April—October period in 1986. Those planted during the months of April—July experienced very heavy mortalities; by October of 1987, the average survival among these four groups was only 6%. Over the same pe- riod, survival in the August group was 58% and averaged over 90% in the September and October groups. These observations indicated, for 1986 at least, a rela- tively short period of infection during mid-summer, 1.e., July—August, suggesting that oysters introduced to the area after August might not be subject to infection until at least the following summer. A partial solution to the problem of introducing oysters susceptible to MSX to enzootic areas such as Cotuit Bay might therefore be to transplant during early fall, when likelihood of infection appears small, with the intention of harvesting before the end of the following summer when mortalities from a new infection cycle would be expected to begin again (see also Ford & Haskin 1988). Attempting to ‘“‘work around’’ the disease in this way, however, requires that the seed oysters to be planted are sufficiently large to be harvested within a year’s time. 360 MATTHIESSEN ET AL. Oysters that failed to reach marketable size within this time, however, and that remained on the beds would risk exposure to infection. A more satisfactory approach, there- fore, would be the production of oysters that were geneti- cally resistant to MSX (Haskin & Ford 1979, Ford & Haskin 1987). To explore the feasibility of using resistant oysters in this region, the growth and survival of one genetically se- lected strain were compared with an unselected control group. Both groups were exposed to the pathogen in Cotuit Bay from April, 1988 to October, 1989. METHODS Broodstock having an established record of resistance to MSX was obtained from the Frank M. Flower Oyster Com- pany in Bayville, New York. These oysters were members of the sixth generation of a strain originally obtained from Long Island Sound in 1965 and bred for resistance at the Rutgers University Shellfish Research Laboratory on Dela- ware Bay (Haskin & Ford 1979). (The Rutgers code desig- nation for this strain is R-BLA; in this paper, the strain is referred to as LISR (=Long Island Sound Resistant). ) The strain selected as a control has been maintained by L Y = -886.5+192.3X-9.4x2 R= 0.99 % PREVALENCE AUG Ocean Pond Corporation on Fishers Island, New York, and has traditionally been used as a primary source of seed for the Cotuit Oyster Company. A large percentage of these oysters died in Cotuit Bay during 1985 and 1986 as a result of MSX infections. (These are designated FIS (= Fishers Island Susceptible). ) The FIS and LISR oysters were induced to spawn at the Ocean Pond Corporation hatchery facility on 6/12/87 and 7/9/87, respectively, and the larvae were reared to setting size by the methods described by Matthiessen (1983). The spat were transferred from the setting trays on shore to floating plastic mesh trays in an adjacent brackish-water pond, where they remained throughout the remainder of the summer and early fall. Both groups were transferred from the surface trays to pearl nets in late fall to avoid possible problems with winter storms and ice conditions. In mid-April of 1988, duplicate samples of 200 oysters from each group were transferred to the Cotuit oyster beds. Each sample was held in 6-mm mesh plastic bags, approxi- mately 0.6 =< 0.9 m in dimension, which in turn were en- closed in 25-mm mesh trays constructed of vinyl-coated wire. These were then placed on the bottom in Cotuit Bay. At the same time, an additional 30 oysters from each Y = -509.2+108.6X-5.2X2 R?= 0.95 SEP OCT NOV 1988 Figure 1. MSX prevalence in oysters, Crassostrea virginica. LISR (=Long Island Sound Resistant); FIS (= Fishers Island Susceptible). MSX INFECTION OF EASTERN OYSTERS IN NEW ENGLAND group were measured for shell height to the nearest mm by means of vernier calipers and weighed to the nearest 0.1 gm on a triple-beam balance. The oysters were then shucked; the meats were fixed in Davidson’s solution for 24 hours and then preserved in 70% ethyl alcohol for histo- logical examination. At monthly intervals, from mid-May to mid-November, 1988, the trays were retrieved, the plastic mesh bags hosed clean of silt and fouling organisms, and the oysters in each bag examined. Dead oysters were removed and recorded. Random samples of 30 live oysters were selected from one tray of each group for determination of shell height and total weight. At the same time, a random sample of 25 oysters from each group was collected from the replicate tray for histological examination. Because of time constraints, this study was designed to N LOCALIZED N SYSTEMIC IN 100 90 80 70 60 50 PERCENT 40 30 20 FIS 10 LISR Jul Au: Sep io} 361 focus on the results of the first season of exposure to MSX, and oyster measurements and histological examinations of the two groups were discontinued after November, 1988. However, at approximately two-month intervals between November, 1988 and October, 1989, mortality rates in both groups were recorded. RESULTS During 1988, the seasonal pattern of MSX infection and subsequent oyster mortality on the Cotuit beds was similar to that observed by Leibovitz et al. in 1986. Parasites were first detected in the samples collected in early August (Fig. 1), with highest prevalence in the control (FIS) group. Prevalence increased in September and by October ex- ceeded 90% in the FIS group, or nearly double that for the LISR group (56%). In all three categories—localized (gill ADVANCED \ N N N NA | N Oct Nov 1988 Figure 2. Comparative intensity of infection in Long Island Sound Resistant (LISR) and Fishers Island Susceptible (FIS) oysters. 362 100 90 80 70 fo7) Oo SURVIVAL (%) o1 oO 40 30 20 10 MATTHIESSEN ET AL. A M JJ AS O N DJ F M A M J J A 1988 1989 Figure 3. Percent survival of Long Island Sound Resistant (LISR) and Fishers Island Susceptible (FIS) oysters. S O MSX INFECTION OF EASTERN OYSTERS IN NEW ENGLAND 363 or epithelial infection); systemic; and advanced systemic — intensity of infection was greater in the FIS group than in the LISR group (Fig. 2). In Figure 3, mortality rates for the two groups, based upon the average mortality in the duplicate trays, are com- pared. From April through August, 1988, light mortality (<5%) was attributed to oyster drill (Urosalpinx cinerea) predation and to occasional breakage of oysters as clusters were separated. Mortality rates increased sharply in Sep- tember, a month after the first MSX infections were identi- fied, and survivorship declined steadily throughout the fall and winter. By April of 1989, nearly 50% of the FIS group had died, as compared with less than 25% for the LISR oysters. In October, 1989, total mortality among the FIS oysters were nearly 85%; in the LISR group, total mortality approximated 50%. The slopes of the survival curves for the two groups be- came more similar after November, 1988. However, the slopes of the regression lines for the two curves computed from the sample data for the period November, 1988—Oc- tober, 1989, were nevertheless found to be significantly different (P < 0.01, Student’s t-test), the mortality rate among the FIS group being higher than that for the LISR group. Growth rates for the FIS oysters began to decrease during August (Figs. 4 and 5), the month in which MSX infections first became evident. This is a distinctly atypical growth pattern for this strain, which normally grows 90 | = STANDARD DEVIATION 76 ——S ISR == FIs 62 48 SHELL HEIGHT (mm) 34 20 MA M J J 1988 Figure 4. Comparative increase in mean shell height of Long Island Sound Resistant (LISR) and Fishers Island Susceptible (FIS) oysters. i Ss © Wl 1 60 if = STANDARD DEVIATION 48 = iSR = AS 36 24 WHOLE WEIGHT (gm) 12 0 MA M J J 1988 Figure 5. Comparative increase in mean whole weight of Long Island Sound Resistant (LISR) and Fishers Island Susceptible (FIS) oysters. A S$ ON D steadily through the summer and fall. In contrast to the sus- ceptible group, the resistant oysters continued to grow into November. By this time, a significant percentage of the LISR group had attained a marketable size of 75 mm or more in shell height and 50 gm or more in total weight (Fig. 6). For samples collected in September, October and November, differences between the two groups with re- spect to shell height and total weight were statistically sig- nificant (P < 0.01, Student’s t-test). DISCUSSION These results indicate a clear superiority of the LISR oysters over the FIS oysters with respect to growth and sur- vival under MSX pressure. It seems evident that the LISR strain is more resistant than the FIS strain to initial infec- tion. In the August (1988) samples, for example, 13% of the LISR group were found to be lightly infected, with in- fections localized in the gills. In contrast, 42% of the FIS strain were infected at this time, and, of these, 17% of the infections, although predominantly in the gills also, were comparatively heavy in other tissues as well. The data suggest that no further infections occurred after disease prevalence reached its peak in early fall of 1988 (Fig. 1). In the case of the LISR group, disease prevalence reached a maximum of 56% in October. We conclude that the majority of oysters infected at that time died during the next 12 months. This would fit well with the observed sur- vival of 46% in October, 1989 (Fig. 3). Oysters that did not 364 MATTHIESSEN ET AL. 80 LISR GROUP 60 | | | 70 | | | GRAMS : 50 40 30 80 70 60 GRAMS 50 40 30 FIS GROUP 20 10 50 60 70 80 90 SHELL HEIGHT (mm) Figure 6. Comparative shell height and whole weight relationships of Long Island Sound Resistant (LISR) and Fishers Island Susceptible (FIS) oysters in November, 1988. MSX INFECTION OF EASTERN OYSTERS IN NEW ENGLAND 365 become infected in 1988, on the other hand, appear to have sustained very few losses during 1989, partially because they were now larger and less subject to predation, but pri- marily because there was no evidence of new MSX infec- tions in Cotuit Bay during 1989 (Susan Ford, personal communication). Members of the FIS group followed a similar pattern in that maximum prevalence (92%) also occurred in October, 1988. The 12% survival estimated in October, 1989, for this group would again be in good agreement with a sce- nario involving a) a brief period of infection limited to July and/or August of 1988, b) virtually total mortality among the infected oysters during the next 12 months, and c) a very high survival rate among members of the population that did not become patently infected in 1988. Newell and Barber (1988) have noted that oysters with even light infections of MSX have significantly reduced feeding rates as compared with noninfected individuals, and Ford (1988) has described the results of experiments in which susceptible strains under MSX pressure grew at only half the rate as that for selected strains. Certainly the im- pact of infection on the Cotuit beds appears evident in the growth curves for the two groups (Figs. 4 and 5). The flat- tening of the growth curves for the FIS group becomes ap- parent in August, at the time infections were first detected. These results have important implications for growers in regions where MSX is enzootic. As indicated earlier, oysters having a shell height of 75 mm (the minimum legal size in most New England states) and a total weight of 50 grams are large enough to be acceptable for the half-shell market. By early November of 1988, nearly 30% of the LISR group had reached the marketable category; none of the FIS oysters, on the other hand, could satisfy the size criteria given above (Fig. 6). Since the LISR oysters had been spawned in July of 1987, they were only 16 months old by November, 1988, which is a young age for market in this region. Neverthe- less, had these oysters been spawned in early spring—as is the procedure in commercial hatcheries—it is likely that a much higher percentage would have been marketable, even though less than two years old. The evident superiority of the LISR group was not as marked during the second season. Although its losses during the April, 1988—October, 1989 period might still be regarded as unacceptably high, it should be noted that, in the Cotuit Bay to date and since MSX observations were first undertaken here, a year of heavy MSX infection has been followed by a year with light or no infection-related mortalities. It seems likely that the degree of resistance demonstrated by the LISR strain would be sufficient to maintain an acceptable level of survival under light or in- termittent MSX pressure. ACKNOWLEDGMENTS This material is based upon work supported by the Na- tional Science Foundation under award number ISI- 8610397. The many helpful suggestions by Dr. Susan Ford and Dr. H. H. Haskin, who reviewed this manuscript, are gratefully acknowledged. REFERENCES 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). Jour. Parasit. 73:368—376. Ford, S. E. & H. H. Haskin. 1988. Management strategies for MSX (Ha- plosporidium nelsoni) disease in Eastern oysters, p. 249-256. In W. S. Fisher (ed.), Disease processes in marine bivalve mollusks. Amer. Fish. Soc., Spec. Pub. 18: 315 p. Haskin, H. H., L. A. Stauber & J. A. Mackin. 1966. Minchinia nelsoni n.sp. (Haplosporida, Haplosporidiidae): causative agent of the Dela- ware Bay oyster epizootic. Science (Washington, D.C.) 153:1414— 1416. Haskin, H. H. 1987. The history of MSX oyster disease, p. 1—3. In A. W. White (ed.), Shellfish diseases: current concerns in the North- east. Woods Hole Oceanographic Institution, Tech. Rep. WHOI-87- 13: 38 p. Haskin, H. H. & S. E. Ford. 1979. Development of resistance to Min- chinia nelsoni (MSX) mortality in laboratory-reared and native oyster stocks in Delaware Bay. U.S. National Mar. Fish. Serv., Marine Fish. Rev. 41(1—2):54-63. Leibovitz, L., G. C. Matthiessen & R. C. Nelson. 1987. A preliminary study of diseases of cultured American oysters (Crassostrea virginica) during an annual growing cycle at the Cotuit Oyster Company, p. 7-11. In A. W. White (ed.), Shellfish diseases: current concerns in the Northeast. Woods Hole Oceanographic Institution, Technical Re- port WHOI-87-13: 38 p. Matthiessen, G. C. 1983. Utilization of a brackishwater pond for the pro- duction of seed oysters (Crassostrea virginica). Aquaculture 31:319— 327. Nelson, R. C. 1987. Account of a recent, severe incidence of MSX oyster disease on Cape Cod, p. 4—6. In A. W. White (ed.), Shellfish dis- eases: current concerns in the Northeast. Woods Hole Oceanographic Institution, Technical Report WHOI-87-13: 38 p. Newell, R. I. E. & B. J. Barber. 1988. A physiological approach to the study of bivalve molluscan diseases, p. 269-280. In W. S. Fisher (ed.), Disease processes in marine bivalve mollusks. Amer. Fish. Soc. Pub. 18: 315 p. Journal of Shellfish Research, Vol. 9, No. 2, 367—371, 1990. HEMOLYMPH ASSAY FOR DIAGNOSIS OF PERKINSUS MARINUS IN OYSTERS CRASSOSTREA VIRGINICA (GMELIN, 1791) JULIE D. GAUTHIER! AND WILLIAM S. FISHER University of Texas Medical Branch Marine Biomedical Institute Galveston, Texas 77550 ABSTRACT Current techniques for assessing intensity of Perkinsus marinus disease of eastern oysters Crassostrea virginica rely on the enlargement of P. marinus trophozoites in oyster tissues after incubation in fluid thioglycollate medium (FTM). The enlarged parasites, or hypnospores, can be stained with iodine and observed in a tissue smear (usually mantle or rectal tissue) with low-power magnification. A semi-quantitative rating of disease intensity based on the apparent percentage of the tissue smear containing stained P. marinus hypnospores has been continuously applied as a useful and valid means to evaluate infections since it was introduced in the early 1950s. Oyster hemolymph is also infected by P. marinus and may be assayed in a similar manner, as described here, to evaluate disease intensity. Results from a completely quantitative hemolymph assay exhibited a linear relationship with the total oyster body burden of parasites and an exponential relationship with the traditional semiquantitative assay using a mantle tissue smear. The hemolymph assay also detected many infections that were misdiagnosed as negative by the tissue smear assay. Since hemolymph samples can be collected from oysters without serious damage, the hemolymph assay does not require that oysters be sacrificed. KEY WORDS: Crassostrea virginica, Perkinsus marinus, oyster, invertebrate pathology, disease diagnosis INTRODUCTION A fluid thioglycollate culture method (Ray 1952, 1966) is currently employed for diagnosis of disease caused by Perkinsus marinus (Mackin, Owen and Collier). The fluid thioglycollate medium (FTM) causes P. marinus tropho- zoites in infected oyster tissue to enlarge to sizes that are easily observed by light microscopy after staining with io- dine. Whole animals or pieces of excised rectal or mantle tissue are incubated in FTM for several days, the.. a tissue smear is prepared on a glass slide and examined for these enlarged parasites, or hypnospores. A scale of infection in- tensity was developed based on the estimated percentage of the tissue smear with hypnospores present (Mackin 1951, Craig et al. 1989). This procedure is routinely used in dis- ease studies even though there is substantial variation of infection intensity within and between tissues (Ray 1952, 1966, Choi et al. 1989). Due to the potential inaccuracies and the semi-quantitative nature of this methodology, an alternate approach was investigated using hemolymph as the assay tissue. Hemolymph analysis for Haplosporidium nelsoni (MSX) was recently found to detect 90—98% of the infec- tions in oysters diagnosed by standard histological methods (Burreson et al. 1988, Ford & Kanaley 1988). Oyster he- molymph is also known to be infected with P. marinus and may play a critical role in the spread and pathogenicity of the disease. Hemolymph becomes infected with P. marinus during the earliest stages of disease (Perkins 1976) and the parasites are believed to grow and multiply within the he- mocytes (Mackin 1951, Ray 1952, Perkins 1976). Infected hemocytes move throughout the oyster and can spread the 'Corresponding author. 367 infection to other tissues (Mackin 1951). Moreover, heavy parasite burdens in the hemolymph may eventually cause the occlusion of blood sinuses and lead to atrophy of vital organs (Mackin 1951, Perkins 1976). A standardized hemolymph assay could offer several advantages over current diagnostic techniques: (1) true quantification of infection intensity on a known volume of hemolymph, (2) measurement of systemic levels rather than localized foci, (3) detection of early infections, and (4) diagnosis without sacrificing the oyster. Our present moni- toring of Gulf Coast oysters requires that disease intensity be measured concurrently with host defense activities. We have developed a hemolymph assay patterned to fit within the framework of those procedures. A similar technique has been used by the National Marine Fisheries Service, (Ox- ford, Maryland) to diagnose Chesapeake Bay oysters for the past two years (A. Farley, personal communication). Following is a description of our quantitative assay that uti- lizes the FTM culture method for diagnosis of P. marinus in oyster hemolymph and a comparison of the hemolymph assay with current diagnostic methods and total oyster body burden. MATERIALS AND METHODS Oyster Collection and Hemolymph Assay Ninety-five oysters were collected during spring and summer 1990 from three sites in Texas (West Galveston Bay, South Padre Island and San Antonio Bay) and one site in Louisiana (Lake Borgne) and refrigerated during ship- ment for 24 hours. Oysters were notched, drained of mantle cavity fluid and a hemolymph sample (1 mL) was drawn from the adductor muscle (Fisher & Newell 1986) and added to 1.5 mL Eppendorf microcentrifuge tubes. Hemo- 368 GAUTHIER AND FISHER lymph was centrifuged at 265 x g for 5 min and cell-free serum was decanted. The pellets (containing hemocytes and P. marinus trophozoites) were resuspended in | mL fluid thioglycollate medium (FTM) containing 5 wl Myco- statin and 5 pl Chloromycetin (Ray 1966). Cultures were kept in the dark at room temperature for 5 days, then cen- trifuged at 265 x g for 5 min to remove FTM. Pellets were resuspended in | mL 2M NaOH which reduced interference caused by bacteria and hemocytes without destroying P. marinus hypnospores (Choi et al. 1989). Hypnospores were washed twice in deionized water, then two drops of Lugol’s Iodine working solution (Ray 1952; a 1:6 dilution of Lugol’s stock solution) were added to each pellet. After gently mixing, the volume was raised to | mL in deionized water. Each | mL suspension was placed in a separate well of a 24-well tissue culture plate and ten-fold serially diluted to an easily counted number (30—300) of hypnospores using an inverted microscope at 100 X magnification. Counting was more easily performed by adding 100 pL subsamples into a 96-well microtiter assay plate, but exam- ination of the entire | mL sample insured better detection of low hypnospore numbers. Comparison of Hemolymph Assay with Tissue Smear Technique After hemolymph was collected as described above, oysters were placed in ambient salinity (15—23 ppt for West Galveston Bay, 30—36 ppt South Padre Island, 18—22 ppt San Antonio Bay, and 2—8 ppt Lake Borgne) at ~25°C temperature for 24 hours then shucked and pro- cessed for mantle tissue analysis. A piece of mantle tissue (~4 mm?) immediately posterior to the labial palps was removed and placed in 10 mL FIM plus Chloromycetin and Mycostatin and incubated at room temperature for 5 days. Tissue pieces were stained with Lugol’s Iodine solu- tion and slide smears were examined for hypnospores and rated according to a modified version of Mackin’s scale (Craig et al. 1989). Comparison of Hemolymph Assay with Total Body Burden Systemic infection levels (hypnospores/mL hemolymph) were compared to total body parasite burdens (hypno- spores/g wet tissue) in 24 oysters from West Galveston Bay. Twelve oysters were processed for hemolymph and body burden within 24 hrs of collection and 12 were held at 34°C and 30 ppt salinity for 10 days to increase infection intensity. Oysters were notched and hemolymph assayed, then whole oysters were minced with a razor, weighed and placed in tubes containing FTM and antibiotics for total body burden assay. All FTM cultures were kept in the dark at room temperature for 5 days, then emptied into separate beakers containing 20 mL of 2M NaOH/g wet tissue and incubated at 50°C for 1—3 hrs (Choi et al. 1989). Sub- samples from each beaker were washed 3 x in deionized water and examined for hypnospores as in the hemolymph assay. Assay Variability Within-test variability of the hemolymph assay was de- termined by withdrawing a 2 mL hemolymph sample from each of 4 oysters and dividing the sample equally into four lots of 0.5 mL each. Within-animal variability was deter- mined by withdrawing 0.5 mL of hemolymph from four areas of the adductor muscle (anterior, posterior, left, right) for each of four different oysters. All hemolymph samples were assayed for hypnospores as previously described. RESULTS Hemolymph Assay P. marinus hypnospores were clearly distinguishable under 100 x magnification after staining with Lugol’s Io- dine solution. The addition of 2M NaOH eliminated back- ground tissue and created a watery medium containing only hypnospores, which settled easily to the bottom of the plate for counting. Differential centrifugation tests showed a 95% decrease in supernatant trophozoites (measured as hypnospores) when centrifugation speed was 265 x g or greater for 5 min. Comparison of Hemolymph Assay with Tissue Smear Technique Systemic infection levels measured by the hemolymph assay increased exponentially in relation to Mackin’s semi- quantitative scale based on tissue smears from the same oysters (Fig. 1). This relationship is described by the fol- lowing equation: hypnospores/mL hemolymph = 42.5(10°-68468x) (r2 = 0.71) where x is the numerical value assigned to Mackin’s scale by Craig et al. (1988) which ranged from 0 (negative) to 5 (heavy). Of the 95 oysters examined in this study, 73 were diag- nosed positive for P. marinus by the hemolymph assay (Fig. 1) whereas only 58 were diagnosed positive by mantle tissue analysis. Of the 37 negatives diagnosed by the mantle tissue assay, 15 were found positive by the hemo- lymph assay; this represents a 40% false negative diagnosis rate for the mantle tissue method. Most false negative diag- noses by the mantle tissue technique occurred in the sample from Lake Borgne, LA; all 12 oysters in this sample were negative by the tissue smear assay yet 11 of these had P. marinus hypnospores in their hemolymph. In all cases, positive infections determined by mantle tissue smears were also detected by the hemolymph assay. Comparison of Hemolymph Assay with Total Body Burden Two groups of oysters from different environmental conditions demonstrated linear relationships between the oysters’ total body parasite burden and systemic infection as determined by the hemolymph assay (Fig. 2). The re- OYSTER HEMOLYMPH ASSAY FOR PERKINSUS MARINUS OG, 0 HYPNOSPORES/ML HEMOLYMPH Le D. LM S) 4 M MH THE MACKIN SCALE Figure 1. Comparison of hypnospore numbers (log,)) per mL determined by the hemolymph assay with the numeric ratings (Mackin scale, see Craig et al. 1988) from the mantle tissue smear assay. Each circle (n = 73) represents the values for one oyster. Note that hypnospores were detected by hemolymph assay in 15 oysters that were diagnosed negative by the tissue smear assay. Oysters that were diagnosed negative by both assays (n = 22) are not shown. Mackin scale notations are NEG = negative, VL = very light, L = light, LM = light/moderate, M = moderate, MH = moderate/heavy, H = heavy and the equation for the regression is described in the text. gression for the group maintained 10 days at high tempera- ture and high salinity is hypnospores/g wet tissue = 312.0(10!-989x) (r? = 0.89) and for the group analyzed within 24 hrs of collection is 78 .6( 1Q1.0484x) hypnospores/g wet tissue (r? = 0.53) where x is the log;) number hypnospores/mL hemolymph. Assay Variability Early tests for variability within a single hemolymph sample sometimes revealed an order of magnitude differ- ence due to aggregates of hypnospores during dilution. Variability was substantially reduced when the following techniques were employed: Hemocytes (initial pellets) were thoroughly resuspended in FIM before incubation; after centrifugation to remove FTM, pellets were completely dissolved in NaOH; during washes pellets were resus- pended by gentle pipetting or tapping the tubes; and in cases of high disease intensity, aggregation of hypnospores was reduced by performing dilutions in separate tubes, which allowed for more thorough mixing. There was less variability between aliquots of the same hemolymph withdrawal than there was between separate hemolymph withdrawals from the same animal. The means + standard errors of hypnospore number per mL hemo- lymph (log-transformed) for four aliquots from the same hemolymph withdrawal from each of eight oysters were 328500094106) = 2O7E S205St == 2065) 35142 20653268 a= gll35 CHO} se 0}, Sty se (OME eral Sts10) a= 409), ine means + standard errors for four separate withdrawals from each of four oysters were 5.31 + .07, 2.95 + .16, 5.44 + .17, and 3.73 + .26. DISCUSSION Hypnospores of P. marinus were easily detected in the hemolymph of infected oysters using the FTM culture LOG, 0 HY RPNOSPORES/G Wer TISSUE = 1 Z 5 LOGig GAUTHIER AND FISHER + © 6 HYPNOSPORES/ML HEMOLYMPH Figure 2. Comparison of total body hypnospore numbers (log,9) per gram wet tissue with hemolymph hypnospore numbers (log,)) per mL from the hemolymph assay. One group of 12 oysters (circles) were assayed within 24 hrs of collection (bottom line, y = 78.6(10!-%*)). The other group of 12 oysters (triangles) were collected at the same time and assayed after 10 d at 30 ppt salinity and 34°C temperature (top line, y = 312.0(10!-°*)). method of Ray (1952, 1966). Quantitation of hemolymph hypnospores has been accomplished in a simple procedure that is more sensitive than current methodology (mantle tissue smears) and allows diagnosis of living oysters. Re- sults from the hemolymph assay are highly compatible with the traditional method (Fig. 1) except at low intensity in- fections where hypnospores were detected in the hemo- lymph but not in mantle tissue smears. Comparison of he- molymph hypnospore counts with total body burden of par- asites from the same oysters exhibited a linear relationship (Fig. 2). Hemolymph hypnospore counts are probably a more consistent measure of disease intensity than analysis of any other tissue because they are relatively independent of lo- calized infections. There is significant variation in the inci- dence of P. marinus in different areas of the mantle (Mackin 1951, Ray 1952, Choi et al. 1989) and the poten- tial for inaccurate diagnosis due to localized infections is greater with light infections or during early stages of the disease. The accuracy of tissue smears probably improves at higher disease intensities when infections are more evenly distributed. Localized infections in the mantle could also account for the many false negative diagnoses by the mantle assay. Ray (1952) reported that false negatives could be reported when infections occurred in some tissues but not others. Of the 37 samples diagnosed as negative by the mantle assay in this study, 15 were positive by the hemolymph technique (sometimes with over 1000 hypnospores/mL). For oysters from the low salinity Lake Borgne site, the mantle assay diagnosis was incorrect in 92% of the cases. Rectal tissue (used by some researchers) may become infected before mantle tissue but no comparison was made between rectal tissue smears and the hemolymph assay. Considering the many positive hemolymph infections that were not detected in the mantle and the histological findings of other investigators (Mackin 1951, Ray 1952, Perkins 1976), we speculate that relatively early stages of infection occur in the hemolymph. Although the primary site of infection may be the digestive gland (Andrews 1988) or epithelial layers of the gill, labial palps and mantle (Perkins 1988), once parasites breech the basement mem- brane they can be engulfed by phagocytes and carried to other tissues by the hemolymph sinuses (Mackin 1951, Perkins 1976). It is even possible that the pathogens are engulfed in the lumen of the stomach by phagocytes which OYSTER HEMOLYMPH ASSAY FOR PERKINSUS MARINUS 371 return to the hemolymph and other tissues. In any case, the parasites are believed to grow and multiply in the hemo- cytes (Mackin 1951, Ray 1952, Perkins 1976) and may lyse the hemocytes to spread infective particles to new tissues. In freshly-stained hemolymph samples, small (2—4 wm) P. marinus trophozoites (Perkins 1988) can be observed within hemocytes. Mature trophozoites with protoplasmic partitioning can be found outside the hemocytes and rup- turing through them. Each of the uninucleate cells enlarge in FTM to produce 15—100 jz£m diameter hypnospores that can be stained and counted with 100 magnification. Mackin (1951) reported that virtually all hemocytes found in highly infected tissues (observed in histology sections) can carry the pathogen. P. marinus presence in hemocytes may be a critical pathogenic factor since the parasite proliferates rapidly and eventually destroys or obstructs hemolymph sinuses leading to vital organs. Atrophy of digestive tubules has been correlated with disease progression (Mackin 1951, Gauthier et al. 1990) and it has been suggested that this is due to poor hemolymph circulation (Mackin 1951, Perkins 1976). It is also significant that hemocytes infected with P. marinus are probably less able to function as a principle factor in internal defense. Suppression of defensive poten- tial could leave the oyster more susceptible to other para- sites and pathogens. Hypnospore counts from the hemolymph assay related exponentially to mantle assays from the same oysters rated on Mackin’s semi-quantitative scale (Fig. 1). Applying numbers to the Mackin scale (ranging from negative = 0 to heavy = 5; see Craig et al. 1989), the regression for this relationship was hypnospores/mL hemolymph = 42.5(10°-68468*) (r2 = 0.71) where x = numerical value on the Mackin scale. Choi et al. (1989) also found an exponential relationship between hypnospore counts in mantle, gill and digestive gland with values assigned by the Mackin scale using the smear proce- dure on the same tissue pieces. The relationship reported by those authors was hypnospores/g wet tissue = 1409.9 (10° ©4296) (r? = 0.91). In the present study, samples were taken from two different tissues (hemolymph and mantle), yet a relatively strong correlation between Mackin’s scale and hypnospore number was found. There was an approximate 30-fold de- crease in hypnospores per mL hemolymph in this study as compared with the per gram wet tissue of Choi et al. (1989); this may be a result of fewer host cells in the hemo- lymph than in the same volume of mantle tissue. The Ray (1966) technique of incubating oyster tissue in a thioglycollate medium for hypnospore enlargement can be used on hemolymph tissue for a quantitative assessment of P. marinus disease. The close correlation between Mackin’s scale and systemic (hemolymph) levels indicates that either assay can be used for diagnosis of heavy infec- tions, however the hemolymph assay is more accurate for diagnosis of light infections. This test is a valid alternative to the currently used method and may be especially useful where repeated samples from the same oyster are needed over time. ACKNOWLEDGMENTS We wish to thank A. Farley for his early suggestions on this project and F. L. Chu, T. M. Soniat and M. Chintala for their constructive criticism of the manuscript. This work was funded by the Texas A&M Sea Grant College Program, National Oceanic and Atmospheric Administra- tion, U.S. Department of Commerce (Grant NA89AA-D- SG139). The U.S. government is authorized to produce and distribute reprints for governmental purposes, not withstanding any copyright notation that may appear hereon. REFERENCES Andrews, J. D. 1988. Epizootiology of the disease caused by the oyster pathogen Perkinsus marinus and its effects on the oyster industry. Amer. Fisher. Soc. Sp. Publ. 18:47—63. Burreson, E. M., M. Robinson & A. Villalba. 1988. A comparison of paraffin histology and hemolymph analysis for the diagnosis of Haplo- sporidium nelsoni (MSX) in Crassostrea virginica (Gmelin). J. Shell- fish Res. 7:1923. Choi, K. S., E. A. Wilson, D. H. Lewis, E. N. Powell & S. M. Ray. 1989. The energetic cost of Perkinsus marinus parasitism in oysters: Quantification of the thioglycollate method. J. Shellfish Res. 8:125— 131. Craig, A., E. N. Powell, R. R. Fay & J. M. Brooks. 1989. Distribution of Perkinsus marinus in Gulf Coast oyster populations. Estuaries 12:82-91. Gauthier, J. D., T. M. Soniat & J. S. Rogers. 1990. A parasitological survey of oysters along salinity gradients in coastal Louisiana. World Aqua. 21:168—200. Fisher, W. S. & R. I. E. Newell. 1986. Salinity effects on the activity of granular hemocytes of the American oyster Crassostres virginica. Bio- logical Bulletin (Woods Hole) 170:122—134. Ford, S. E. & S. A. Kanaley. 1988. An evaluation of hemolymph diag- nosis for the detection of the oyster parasite Haplosporidium nelsoni (MSX). J. Shellfish Res. 7:11-18. Mackin, J. G. 1951. Histopathology of infection of Crassostrea virginica (Gmelin) by Dermocystidium marinum Mackin, Owen and Collier. Bull. Mar. Sci. Gulf and Carrib. 1:72-87. Perkins, F. O. 1976. Dermocystidium marinum infection in oysters. U.S. National Marine Fisheries Service Marine Fisheries Review 38:19— 21. Perkins, F. O. 1988. Structure of protistan parasites found in bivalve mol- luscs. Amer. Fish. Soc. Spec. Publ. 18:93-111. Ray, S. M. 1952. A culture technique for the diagnosis of infection with Dermocystidium marinum Mackin, Owen and Collier in oysters. Sci- ence (Wash D.C.) 116:360—361. Ray, S. M. 1966. A review of the culture method for detectiing Dermo- cystidium marinum, with suggested modifications and precautions. Proc. Natl. Shellfish. Assoc. 54:55—69. Journal of Shellfish Research, Vol. 9, No. 2, 373-381, 1990 YIELD ESTIMATES FOR THE VIRILE CRAYFISH, ORCONECTES VIRILIS (HAGEN, 1870), EMPLOYING THE SCHAEFER LOGISTIC MODEL WALTER T. MOMOT, P. LYNN HAUTA, AND JAMES A. SCHAEFER Department of Biology Lakehead University Thunder Bay, Ontario P7B SEI ABSTRACT Estimates of surplus production employing the Schaefer model for a population of the crayfish, Orconectes virilis, in Dock Lake, Ontario suggests that the optimal surplus production for this stock is approximately 20 kgs for females and 30 kgs for males and is obtained at an effort of approximately 4,000 trap days per season. Calculation of parameter estimates for MSY and optimal fishing effort employing a modified Gulland method corroborated these empirical values. The calculated values were 17.0 to 17.4 for females and 30.1 to 33.1 for males at an optimal effort between 4360 and 4540 trap days depending on the exploitation rate employed. Despite marked changes in fishing effort, the crayfish were harvested in proportion to their age specific abundance. Therefore the fishery did not distort the age structure of the population. Earlier empirical evidence that efforts greater than 6000 trap days produced lower yields was supported by the use of the Schaefer model which suggested that high efforts lower ‘surplus production.”’ The confirmatory results from employment of the Schaefer model are probably due to the fact that (1) catchability is constant for males, (2) the crayfish are harvested across all age groups and (3) crayfish are well regulated with both excellent responses to harvest and short biomass turnover times. Whether such models could be used in a predictive manner, however, requires additional data from other stocks and fisheries KEY WORDS: crayfish, Orconectes virilis, yield INTRODUCTION Exploitation of wild stocks of freshwater crayfish in North America has increased the need for economical, simple, quick and effective methods for stock assessment and management purposes. This paper illustrates how lo- gistic models might be employed to enable us to manage crayfish fisheries. METHODS For the past 12 years, we exploited a crayfish population in Dock Lake, Ontario (Momot 1986). We carried out se- quential Schumacher-Eschemyer mark and recapture esti- mates for each sex of both young-of-the-year, yearling and adult crayfish. In most years, we also recorded recaptures of previously marked animals during the annual harvest. This allowed us to measure directly the exploitation rate (u) (Ricker 1975), from the ratio of marked to unmarked adult crayfish captured for each year. [Since this was a type I fishery, the values were converted to the fishing from the formula u = (1 —e~F) Ricker (1975)]. In addition, the availability of annual recapture data allowed us to make a separate Petersen population estimate in most years. By comparing the actual catch with both the Schumacher and Petersen estimates, we could calculate (F) the instantaneous rate of fishing mortality for most years of the fishery. This provided three estimates of the fishing rate during each of the years 1979 and 1981 to 1987. The mark recapture tech- nique, used for determining population density prior to the harvest, provided the opportunity to assess the precision 373 and bias in estimating exploitation rates. The number of marked recaptured crayfish in the harvest relative to the total number of marked individuals in the population pro- vide an unbiased exploitation rate estimate (Ricker 1975). This assumes that marked and unmarked individuals ran- domly mix. The 1979 and 1981—87 harvests have complete records of marked crayfish caught. The proportion of the exploitable stock caught each year thus provided another possible estimate of exploitation rate. Fishing rate estimates derived from the proportion of the exploitable stock harvested proved to be consistently higher than those derived from recapture rates (Tables 1-3). This between estimate disparity increased at efforts of 4000 trap days or greater, especially for males. Throughout this period the number of sampling units had remained constant (200 traps) while the duration of harvest increased. Since increases in sampling duration decrease the relative proportion of age II males in the catch, this alters age and sex specific trapabilities. Apparently smaller crayfish were discouraged by these males from entering the traps during the harvest, and logically during the preceding population estimate (when crayfish were marked). This ef- fect was apparent in the harvest data where the mean size of marked crayfish was larger than for unmarked crayfish for all years. Consequently the daily proportion of harvested small crayfish increased as sampling duration increased. The result was a slight underestimation of exploitation rates at higher efforts. During the period of the study, the effort (f), measured in trap days (T.D.), (1 trap fished for 1 day), increased from 2600 T.D. to 6000 T.D. Effort in 1979 and 374 MOMOT ET AL. TABLE 1. Computation of surplus production for the crayfish, Orconectes virilis, in Dock Lake, Ontario employing a rate of fishing calculated directly from the recapture of previously marked crayfish in 1979 and 1981-87 of each sex. Rate of Yield Effort Fishing Year Sex (kg) (trap days) (F) 79 M 5.02 2600 390 F 3.66 300 81 M 6.25 2600 450 F 330 82 M 17.09 4000 1.000 F 10.35 610 83 M 23532 4000 1.040 F 13.82 570 84 M 22.44 4000 .670 F 9.95 350 85 M B25 6000. .750 F 17.05 480 86 M 18.47 6000 830 F 10.47 430 87 M 14.40 6000 .680 F 10.97 430 1981 was at 2600 T.D., in 1982-84 at 4000 T.D. and in 1985—87 it was increased to 6000 T.D. We, thus, have data for eight years on effort (f), rate of fishing (u) and yield in kgs. These were employed to construct a simple Schaefer model of surplus yield (Ricker, 1975) (Tables 1—3). The Schaefer model estimates surplus production for each individual year by dividing each year’s catch in kgs by its rate of fishing in order to estimate the mean stock (W) present during the year. The initial stock W at the begin- ning of the year is the average of the mean stock (W) for the two adjacent years. The difference between any two adjacent initial stocks is the decrease for the year in ques- tion and when added to the yield it gives surplus produc- tion. If surplus production is plotted against stock density and describes a well defined curve, such a curve can empir- ically define the position of maximum yield for any stock with a short life history and quick reaction to density changes (Ricker 1975) as exemplified by this species of Mean Initial Change of Surplus Stock Stock Stock Production (W) (W) (AW) (Y’) 12.98 12.18 13.44 13.06 13.89 2.05 8.3 13.95 2.40 7.0 15.49 15.46 4.29 21.4 17.09 5.14 1525 16.96 19.78 20.60 22.42 8.18 31.5 24.24 5.74 19.6 27.96 26.34 33.49 10.50 3249 28.43 5.64 38.46 31.98 43.43 — 5.62 26.9 35.52 = 2105 15.0 32.84 29.93 22.25 = ileal} 133 24.34 = 5,0) 5.4 Diver 24.92 21.17 25.51 crayfish. In addition, calculations of parameter estimates for maximum sustained yield (MSY) and optimal fishing effort (fs) were carried out using a modification of the Gul- land method (in Ricker 1975) from the data in Tables 1 to 3. Least squares regression lines (NoruSis 1986) were fitted to parabolic functions of surplus production (Y,) against effort (f) in the form: VG] ar= be = ¢€ (1) In contrast, the Gulland method specifies fitting a linear function of Y,/f against f. The modification, we believe, improves the model statistically and biologically. First, as Ricker (1975) has noted, the appearance of f on both sides of Gulland’s equation is statistically dubious. Furthermore, allowing a constant c to enter the equation effectively re- moves the requirement that the regression line pass through the origin and allows the parabola to shift to the right along the x-axis. In effect, this relaxes the assumption that at any YIELD ESTIMATES OF CRAYFISH TABLE 2. Computation of surplus production for the crayfish, Orconectes virilis, in Dock Lake, Ontario employing a rate of fishing calculated by dividing the number caught by the number estimated by the Schumacher-Eschemeyer methods for the years 1979, 1981-87. Rate of Mean Initial Change of Surplus Yield Effort Fishing Stock Stock Stock Production Year Sex (kg) (trap days) (F) (W) (W) (AW) (Y’) 79 M 5.02 2600 363 13.94 E 3.66 334 10.94 13.06 13.04 81 M 6.25 2600 13 12.18 10.83 13 4.60 304 15.14 4.58 9.34 17.64 4.74 17.78 82 M 17.09 4000 .740 23.09 26.11 18 10.35 507 20.41 9.02 17.76 26.66 7.41 25.19 83 M 23.32 4000 Stil 30.24 31.40 F 13.82 461 29.97 8.08 19.46 34.74 5.64 30.83 84 M 22.44 4000 39.23 29.48 F 9.95 3 31.69 7.04 W222 41.78 3h 77/ 34.10 85 M B2ESi/ 6000 s135 44.32 26.81 F 17.05 .467 36.51 —5.76 15.71 36.02 34. 32.76 86 M 18.47 6000 666 27.73 12.00 F 10.47 361 29.00 —6.47 5.79 29.55 —4.68 28.08 87 M 14.40 6000 459 S137 F 10.97 404 27.16 f > 0 there will be some surplus to be cropped. From equa- tion (1), therefore: a MS Ya 2b (2) Anes — NY es wb (3) RESULTS Examination of Figures 1—3 derived from Tables 1—3 graphically provide maximum surplus yield estimates for each of the three rates of fishing employed. Table | and Figure | empirically define the maximum surplus yield for each direct estimate of fishing based on the recapture of previously marked crayfish. For males, maximum surplus production is attained at an fishing rate of 0.67 while for females this value is 0.57. Table 2 and Figure 2 empirically define the maximum yield to be attained at an exploitation rate of 0.77 for males and 0.46 for females. The fishing rate for Fig. 2 and Table 2 was obtained by dividing the number of crayfish har- vested by the number estimated to be present using the Schumacher-Eschemeyer mark and recapture method. Table 3 and Figure 3 define maximum yield to occur at an exploitation rate of 0.49 for males and 0.46 for females. The exploitation rate for Figure 3 and Table 3 was obtained by dividing the number caught by the numbers estimated to be present by the Petersen method. Obviously, males produce maximum values at higher rates of fishing than females. Values for all four curves suggest a fit to a parabolic function for all years except 1986 (Figs. 1—3). In the summer of 1985, effort was in- creased to 6000 trap days. The values for 1986 were de- rived from data obtained at this increased level of effort and are considerably below those for other years. All three curves, therefore, suggest that 4000 trap days appears op- timal for exploitation of both sexes of this species in this 376 MOMOT ET AL. TABLE 3. Computation of surplus production for the crayfish, Orconectes virilis, in Dock Lake, Ontario employing a rate of fishing calculated by dividing the number caught by the number estimated by the Petersen method for the years 1979, 1981-87. Rate of Mean Initial Change of Surplus Yield Effort Fishing Stock Stock Stock Production Year Sex (kg) (trap days) (F) (W) (W) (AW) (Y’) 79 M 5.02 2600 330 15.34 F 3.66 .269 13.59 14.87 15.02 81 M 6.25 2600 434 14.40 B72 LIES i, F 4.60 280 16.44 4.5 9.10 19.99 19FS2 82 M 17.09 4000 .668 25.58 10.17 27.26 F 10.35 458 22.59 = 1558 17.93 30.16 27.10 83 M 23°32 4000 671 34.75 9:92 33.24 F 13.82 437 31.61 2.84 16.66 40.08 29.94 84 M 22.44 4000 494 45.42 13.42 35.86 Ie 9.95 352 28.27 6.57 16.52 53.50 36.51 85 M 32.57 6000 2529) 61.57 — 6.28 26.29 F 17.05 381 44.75 0.77 17.82 47.22 37.28 86 M 18.47 6000 562 32.86 = 16.19 2.28 F 10.47 351 29.82 — 6.70 Bef/7/ 31.03 30.58 87 M 14.40 6000 493 29.20 F 10.97 .350 31.34 lake. All three curves suggest optimal surplus production for this stock to be approximately 20 kgs for females and 30 kgs for males (Figs. 1-3). The calculated values for MSY and optimal fishing effort using the modified Gulland method are in good agreement with the empirical estimates derived from Figures 1—3 (Table 4). The estimates for MSY and optimal f are similar regardless of whether the recapture, Schumacher or Petersen derived exploitation rates are employed (Table 4). The equations are given in Table 4 and the values for MSY for males range from 30.1 to 33.1 while the values for optimal f fall between 4360 and 4540 trap days (Table 4). Values for females were calcu- lated as follows: for MSY values ranged from 17.0 to 17.4 kgs, while for optimal f values ranged from 4380 to 4540 trap days. These values will agree with the empirical esti- mates in Figures 1—3 and the calculated Schaefer estimates provided in Tables 1—3. DISCUSSION The greatest difficulties in applying surplus production models come from their stringent assumptions about the re- lation between stock size and observed abundance indices, such as catch per unit effort (CPUE) (Walters 1986). In the case of O. virilis, Morgan and Momot (1988) showed that the percentage of males harvested increased linearly with nominal fishing effort at least within what normally consti- tutes economically determined limits. Because no sex or size limits were imposed in this fishery, crayfish were gen- erally harvested in proportion to their age-specific abun- dance, despite marked changes in fishing effort. The fishery did not, therefore, distort the age structure of the population between years. Therefore, population processes were not directly altered by fishing. One criticism of surplus yield models is that they tend to ignore age struc- ture and thus the “‘real’’ biological processes which gen- erate the biomass. But since males are more vulnerable to fishing than females as the data in Tables |—3 indicate; catchability may therefore be an important consideration. Catchability (q) varies with sampling intensity, duration of harvest and population density. Increasing fishing effort changed susceptibility of crayfish to the gear. Gear satura- tion took place when we used 200 traps. The differences YIELD ESTIMATES OF CRAYFISH 377 RATEA 30 20 — (2) FEMALES on SURPLUS PRODUCTION IN Kg Ww (eo) 20 10 5 5 10 20 30 40 50 60 @ Males @ Females MEAN STOCK DENSITY IN Kg Figure 2. Surplus production in kgs of the crayfish, Orconectes virilis, in Dock Lake, Ontario plotted against the mean stock density in kgs given by the data in Table 2. 378 MOMOT ET AL. RATE B FEMALES SURPLUS PRODUCTION IN Kg 5 10 20 30 40 50 60 @ Males MEAN STOCK DENSITY IN Kg @ Females Figure 2. Surplus production in kgs of the crayfish, Orconectes virilis, in Dock Lake, Ontario plotted against mean stock density in kgs given by the data in Table 2. YIELD ESTIMATES OF CRAYFISH 379 RATE C 30 20 —% (2) FEMALES on SURPLUS PRODUCTION IN Kg w (2) 20 10 5 iO 20 30 40 50 60 @ Males fe Romales MEAN STOCK DENSITY IN Kg Figure 3. Surplus production in kgs of the crayfish, Orconectes virilis, in Dock Lake, Ontario plotted against mean stock density given by the data in Table 3. 380 MOMOT ET AL. TABLE 4. Calculation of MSY and optimal fishing effort (f) for crayfish, Orconectes virilis, in Dock Lake, Ontario using modified Gulland method. Method Sex Equation Optimal f MSY Recapture M Y. = —91.3 + .0538f — .00000596f? 4500 30.1 F Y. = —43.3 + .0273f — .00000307f? 4450 17.4 Schumacher M Y. = —77.4 + .0476f — .00000524f? 4540 30.7 F Y. = —30.5 + .0217f — .00000248f? 4380 17.0 Petersen M Y. = —99.7 + .0609f — .00000698f? 4350 33.1 F Y, = —32.5 + .0227f — .00000258f2 4400 17.4 ° between the 1979-81, 1982—84 and 1985-87 periods in- dicate decreased vulnerabilities as harvest duration in- creased (i.e., from 1979-87 the number of sampling units remained the same; only the number of days increased). In 1984 use of 250 traps slightly decreased total harvest and (CPUE) (Tables 1—3). This resulted from gear competition whereby the fraction of the population taken by a single unit of nominal effort decreased which in turn reduced total harvest. Daily age and sex specific harvests varied little over nine years. Generally 50% of the estimated population of males were caught midway through the harvest. The fe- males reached 50% harvest levels one day later. At 6000 trap-days effort, daily CPUE declined and lowered the 50% harvest midpoint. At 6000 trap-days the catch exhibited the following characteristics (1) smallest average size of cray- fish (2) high percentage of small crayfish (3) daily de- creasing male size (4) decreasing daily CPUE and (5) lower CPUE than expected from total harvest (Morgan 1987). Small crayfish <30 mm C.L. constituted a variable pro- portion of the catch between sexes and between years. Re- gardless of nominal effort females always constituted a greater percentage of these smaller individuals while in- creasing effort from 2600 to 4000 trap-days stabilized the percent harvest of small crayfish, at 6000 trap-days small crayfish dominated the catch. CPUE expressed as the number or weight of crayfish per trap-day remained con- stant from 1979 to 1981. CPUE increased two to three fold at nominal efforts of 4000 trap-days from 1982 to 1984. A similar concordant CPUE occurred in 1985 (i.e., similar to CPUE at 4000 trap days) although nominal effort increased to 6000 trap-days. After 1985, CPUE decreased at 6000 trap-days. The trend in CPUE was similar between the sexes. The annual harvest was proportional to nominal effort. However, the interaction between fishing effort and sam- pling intensity profoundly changed CPUE and total catch relationships between years. While CPUE at 2600 and 4000 trap-days was proportional to effort after 1985 the CPUE at 6000 trap-days was not. Exploitation visibly affected the percent harvest of age II male crayfish. During the first few days age II males com- prised 20-30% of the catch. However, by the end of the harvest of percentage of age II males declined to less than 10% (Morgan 1987). The proportion of age I males and females and age II females stabilized over the harvest pe- riod. Therefore declining daily catch of age II males al- lowed increased numbers of age I male and female and age II females to enter the traps. This kept daily CPUE con- stant. Crayfish were harvested in proportion to their age-spe- cific abundance despite marked changes in fishing effort. Combined male and female frequencies differed slightly from population age structure in 1979, 1981, and 1985. However, these differences were caused by disparity be- tween low age III abundance and their high vulnerability to passive fishing gear compared to that of younger groups (1.e., nearly all age III crayfish which make up a very small percentage of the total population (0.3—3.9%) (Momot & Hauta 1990) were harvested annually). Through removal of crayfish in proportion to their abundance, exploitation sta- bilized the male and female age structure. Changes in age and sex specific harvest parameters sug- gest that catchability varies with sampling intensity, dura- tion of harvest and population density. Interpretation of CPUE was made more difficult because of these interrelated variables. At harvest rates of 2600 to 4000 trap-days though male catchability remained relatively constant, female catchability varied inversely. Females experience signifi- cantly higher fishing mortality per unit of effort at lower densities i.e., catchability was density-dependent. This in- validates the assumption of a proportional relationship be- tween female CPUE and abundance. The non-linear relationship between female catchability and abundance could be the result of either gear saturation or a functional response. Except for 1984 sampling inten- sity was constant from 1979 to 1987; however, changes in CPUE and relative catch composition indicate a functional response. As harvest duration increased (without changing the number of traps) female vulnerability decreased. At 6000 trap-days effort, catchability decreased for both males and females, even though the percent harvest re- mained constant when compared to 4000 trap-days of effort (Morgan 1987). Apparently continuously employing the YIELD ESTIMATES OF CRAYFISH 381 same sampling intensity (i.e., 200 traps) led to time satura- tion between harvests. We never harvested more than 60% of the exploitable stock (70% of the males; 50% of the fe- males) regardless of sampling intensity or duration. The constant catchability of males is reflected in the better ‘fit’? of the data from males to a parabolic function (Figs. 1—3). At higher harvest levels (6000 trap days) there was no increase in exploitation rates of the stock. However, the CPUE at 6000 trap days was so low as to make overharvest in the economic sense very unlikely. The population is therefore protected from biological overfishing regardless of effort but the Schaefer model in addition suggest that such high efforts also lower “‘surplus production.’’ Thus it is both biologically and economically unsound to fish at such high rates. The Schaefer model and Gulland method confirm our earlier empirical results (Morgan & Momot 1988) in suggesting 4000 trap days to be optimal for har- vest of this particular population. Logistic models might help us to evaluate and manage crayfish fisheries. Several models are available; e.g., Schaefer (1954); Fox (1970); Pella and Tomlinson (1969). Therefore the choice depends on the exact form of the rate of change of biomass. It becomes important to choose the one which best resembles the growth of the stock in ques- tion (Pitcher & Hart 1982). Whether such logistic models could actually be used for exact yield prediction is an open question. We used surplus yield model as a confirmatory tool rather than as a direct yield predictor. The fit of our data was probably enhanced by the fact that catchability was constant for males, that the fishery was fished across all age groups and that these crayfish populations were well regulated density dependent populations with excellent re- sponses to harvest with short biomass turnover times (Momot 1986). This lessens problems arising because of time delays in component production responses (particu- larly recruitment) and persistent or cyclic disequilibrium in internal stock structure (Walters 1986). ACKNOWLEDGMENTS This work was supported by a long-term Natural Sci- ences and Engineering Research Council Grant. We thank Dr. N. Caputi, Western Australian Marine Research Labo- ratories for his comments. Many students from Lakehead University contributed their time over the years in the col- lection of these data. Miss L. Scarcello typed the manu- script. REFERENCES Fox, W. W. 1970. An exponential yield model for optimizing exploited fish populations. Transaction American Fisheries Society 99:80—88. Momot, W. T. 1986. Production and exploitation of the crayfish, Orco- nectes virilis, in the northern climates. Canadian Special Publication Fisheries and Aquatic Science 92:154—167. Momot, W. T. & P. L. Hauta. 1990. Effects of growth and mortality phenology on the cohort P/B of the crayfish Orconectes virilis. Fresh- water Crayfish 8:In press. Morgan, G. W. 1987. Population dynamics of Orconectes virilis in North- western Ontario. M.Sc. Thesis, Lakehead University, Thunder Bay, Ontario, p. 202. Morgan, G. & W. T. Momot. 1988. Exploitation, of Orconectes virilis, in northern climates: complementarity of management options with self-regulatory life history strategies. Freshwater Crayfish 7:69—80. NoruSis, M. J. 1986. SPSS/PC + Advanced Statistics. SPSS Inc. Chi- cago, Ill. 392 pp. Pella, J. J. & P. K. Tomlinson. 1969. A generalized stock production model. Bulletin Inter-American Tropical Tuna Commission 12:421- 96. Pitcher, T. S. & P. J. B. Hart. 1982. Fisheries Ecology, AVI Publishing, Connecticut. 414 p. Ricker, W. E. 1975. Computation and interpretation of biological sta- tistics of fish populations. Bulletin Fisheries Research Board Canada 191:1—382. Schaefer, M. B. 1954. Some aspects of the dynamics of populations im- portant to the management of commercial marine fisheries. Bulletin Inter-American Tropical Tuna Commission 1:27—56. Walters, C. 1986. Adaptive management of renewable resources. Mac- millan Publishing Company, New York. 374 p. ~*~ = Journal of Shellfish Research, Vol. 9, No. 2, 383—387, 1990. PURGING CRAWFISH IN A WATER SPRAY SYSTEM! THOMAS B. LAWSON, HARNARINE LALLA, AND ROBERT P. ROMAIRE? Agricultural Experiment Station Louisiana State University Agricultural Center Baton Rouge, LA 70803 ABSTRACT Crawfish purging systems include flow-through, batch and spray systems. Spray purging offers advantages over other system types. A spray purging system was evaluated using four crawfish loading densities (4.9, 14.6, 24.4 and 34.1 kg/m?) and four water application rates (3.6, 5.4, 10.7, and 12.5 L/min/m?). Water was sprayed over the animals with plastic agricultural spray nozzles. A fifth box receiving no spray served as a control. Crawfish purged within 40 hours after placement into the system. With a water spray to keep them moist, crawfish mortality ranged from 0 to 24.1% over 40 hours. Mortality significantly increased with an increase in loading density (P < 0.05). Mortality in the control ranged from 2.3 to 79.2% and decreased as loading density increased. Crawfish mortality did not differ with an increase in water spray application from 3.6 to 12.5 L/min/m?. An acceptable crawfish mortality of less than 5% per day was attained at a spray rate of 3.6 L/min/m? and crawfish loading density of 24.4 kg/m? or less. KEY WORDS: crawfish, crustaceans, purging INTRODUCTION Crawfish farming is the largest freshwater crustacean aquaculture industry in the United States. The red swamp crawfish, Procambarus clarkii and white river crawfish, P. zonangulus (formerly P. a. acutus), are widely distributed species of commercial importance in several states in the U.S. although Louisiana is the largest producer with annual production of 35,000—50,000 metric tons (Huner 1989). P. clarkii has been widely introduced outside its natural range (Avault & Huner 1985). Approximately 40 to 50% of the commercial harvest in Louisiana is processed in the form of peeled tail meat, and the remainder is marketed live or frozen in the round (Dellenbarger et al. 1986). As late as 1988, 80% of all crawfish produced in Loui- siana were consumed locally. Presently, 30% of locally produced crawfish is now exported to other states and Eu- ropean markets (Dellenbarger et al. 1990). Several barriers inhibit marketing efforts of Louisiana produced crawfish, particularly in European markets. These barriers target the aesthetic appearance and quality of the live and whole frozen animals. Crawfish often have ex- ternal epizooic growths attached to the exoskeleton; the in- testine contains partly digest food; the brachial chamber re- tains water and silt particles, often causing “‘swampy”’ or fishy odors; and debris may adhere to the appendages. Pro- ‘Approved for publication by the Director of the Louisiana Agricultural Experiment Station as manuscript number 90-07-4372. Trade names are used solely to provide specific information. Mention of a trade name does not constitute a warranty by the Louisiana Agricultural Experiment Station of the LSU Agricultural Center of the product nor an endorsement to the exclusion of other products not mentioned. Associate Professor, Agricultural Engineering Department; Former Grad- uate Assistant, Agricultural Engineering Department; and Associate Pro- fessor, School of Forestry, Wildlife and Fisheries, respectively. 383 cessors preparing crawfish for the tail meat market are not impacted by these factors, and local consumers in Loui- siana are not overly concerned about the presence of epi- zooic growths or the dark, digested food material found in the intestinal tract of unpurged animals. However, distrib- utors of live and whole frozen animals must deliver a high quality product, requiring factors causing quality loss to be addressed. Quality of the live animal may be improved by ‘“‘purging.’’ An analogous process called ‘‘controlled puri- fication’’ or ‘‘depuration’’ is often used in the oyster in- dustry to reduce the concentration of bacteria and viruses in the animals. Crawfish are purged, not to reduce pathogenic organisms, but primarily to empty the stomach and intes- tinal tract of digested matter. In unpurged crawfish the in- testinal tract is darkly colored, due to the presence of di- gested food. This material is unsightly when the tail muscle is removed, and the material must be washed from the meat during processing. When crawfish are purged the intestine is empty and thus clear. Additional quality factors are im- proved by purging: stale water in the brachial chamber is replaced with fresh water, reducing off-flavors and odors; and externally attached algae and microorganisms are abraded off by the constantly moving mass of animals. It normally requires from 24 to 48 hours for complete purging to occur (Lawson & Baskin 1985). Purging systems are of three basic types: batch, flow- through and spray. Flow-through systems have been the in- dustry mainstay but are rapidly being replaced with spray systems, which have been used successfully in Spain (Gaude 1983). Energy costs associated with pumping and aeration in flow-through systems are high compared with the lower water application rates in spray systems. In spray systems, crawfish are not completely immersed but merely 384 LAWSON ET AL. have water sprayed over them. Oxygen depletion, often a serious problem in batch and flow-through systems, does not occur. Of particular importance is that, during system down times, when pumps are non-operational, oxygen de- pletion is of little concern in spray systems. Purging system design and operational procedures are discussed in greater detail elsewhere (Lawson & Baskin 1985, Lalla & Lawson 1987, Lawson & Drapcho 1989). This study targeted unknown design parameters for crawfish spray purging systems. More specifically, re- search objectives were to determine: (1) if crawfish will purge satisfactorily in a spray system; (2) the effects of water application (spray) rate on purging and crawfish mor- tality; and (3) the relationship of animal loading density on purging and mortality. EQUIPMENT AND EXPERIMENTAL DESIGN Purging Box Construction Five crawfish purging boxes were used in the project. The 1.2 m X 0.9m X 0.5 m boxes were constructed from 50 mm X 50 mm treated wooden members (Fig. 1). The bottoms were 1.2 m X 0.9 m xX 0.013 m sheets of ply- wood with 0.63-cm diameter holes drilled 0.20-cm on center to facilitate drainage. Plastic mesh material (6-mm mesh size) was fastened to the sides and ends of the boxes. Each box was partitioned into four-0.20 m* inner compart- ments with wooden members and plastic mesh. The openings at the tops of the compartments were covered with 1.9-cm hexagonal mesh PVC-coated poultry netting to prevent the animals from escaping. The netting could be easily folded aside for loading and unloading the crawfish between replicate tests and for routine maintenance. A sheet of plywood was placed over each box during the tests to protect crawfish from the drying effects of the sun. Water Supply The research purging system is detailed in Figure 2. The purging boxes were arranged in a straight line within the 7 m X 2m X | mraceway constructed from concrete cinder blocks and placed outdoors. Dechlorinated city tap water was routed to four of the boxes with 12.5-mm diameter rigid plastic drip irrigation tubing. The fifth box, the con- trol, received no spray. A main water line ran the length of the row of boxes, and two laterals served each of the four boxes to be sprayed. The two laterals were attached to the top inner edge on either side of each box. Three 3-mm diameter holes were drilled into each lateral line in each compartment of the four boxes receiving a water spray. Maxijet plastic spray nozzles (Thayer Indus- tries, Inc., Dundee, Florida) with a 90° angle spray pattern were threaded into the holes and angled so that the bands of spray evenly covered the bottom of each compartment. The PLASTIC MESH SO-mm x SO0-na WOODEN MEMBERS Figure 1. Schematic of experimental crawfish purging boxes. nozzle orifice diameter varied in order to obtain a different spray rate for each purging box. The water application rate was the same for each of the four compartments in a given box. The nozzles were color-coded for identification ac- cording to orifice diameter as follows: blue, 1.03 mm; green, 1.28 mm; red, 1.54 mm and white, 1.79 mm. A pressure gauge was installed at the end of the main water line to monitor system water pressure. A maximum pressure of 34.2 kPa (5 psi) was available for the study. The nozzle discharge spray rates from manufacturer-sup- plied literature for nozzles operating at 102.7 kPa (15 psi) are presented in Table 1. Manufacturer’s data for nozzles operating below this pressure were not available; therefore, we experimentally determined the discharge spray rates from | liter grab samples for each size nozzle at 34.2 kPa (Table 1). Water temperature during the study ranged from 22 to 26°C. PROCEDURE The research was conducted at Louisiana Agricultural Experiment Station’s aquaculture research facility at Ben Hur Research Farm in Baton Rouge. Crawfish used in the purging studies were about 90% Procambarus clarkii and 10% P. zonangulus. No studies were conducted to deter- mine purging differences between species or between as- sorted sizes among species. The animals were harvested from six 2-ha ponds and mixed before being placed into the purging systems. Within three hours of harvest the animals were weighed and loaded into the purging boxes. Crawfish were loaded into the four compartments of each box at 4.9 (1.0), 14.6 (3.0), 24.4 (5.0), and 34.1 kg/m? (7.0 Ib/ft?), respectively (Fig. 3). Four of the boxes received water spray rates of 3.6, 5.4, 10.7 and 12.5 L/min/m’, respec- tively for a combination of four water spray rates at each of four loading densities. A fifth box, used as a control, re- ceived 0 L/min/m?. The spray rates and crawfish loading densities were selected based on discussions with indi- viduals who have operated commercial crawfish spray purging systems. PURGING CRAWFISH IN A WATER SPRAY SYSTEM 385 ie WATER FLOW 3.57 L/min/m® 5.36 L/min/m* 10.71 L/min/m 12.50 L/min/m 0 L/min/n® Figure 2. Arrangement of purging boxes and water supply. All purging tests were conducted for 40 hours, based on data from preliminary trials that demonstrated crawfish completely **purge’’ in 30—40 hours. After 40 hours, that test was terminated and dead crawfish were counted and mortality expressed as percent of weight. Ten randomly chosen crawfish from each compartment in each box were dissected and visually examined. Crawfish were considered completely purged when the intestinal tract was clear of fecal matter. Four replicate tests were conducted, and the boxes were rinsed with clean tap water between replicates. The experimental design was a randomized complete block design in a 4 X 5 factorial arrangement of treat- ments. The block effect was replicate test, and the two main effects were loading density and water spray rate. Data were analyzed with the analysis of variance with the 0 L/min/m? water spray rate data included in the analysis (4 x 5 factorial analysis) and with the 0 L/min/m* removed from the data base (4 x 4 factorial analysis). The response variable analyzed was percent mortality, and statistical dif- ferences between main effect (loading and spray rate) means were declared significant at a = 0.05. Statistical TABLE 1. Color codes, orifice diameters and discharge rates for maxijet spray nozzles at 102.7 and 34.2 kPa pressure. Discharge @ Discharge @ Orifice Diameter 102.7 kPa! 34.2 kPa? Nozzle —————“~— ———— —— Color mm in Ipm gpm Ipm gpm Blue 1.02 0.04 0.57 0.15 0.34 0.09 Green 1.27 0.05 0.87 0.23 0.49 0.13 Red 1.52 0.06 1.29 0.34 1.02 0.27 White 1.78 0.07 2.00 0.53 1.17 0.31 ' Data from Thayer Industries, Inc., Dundee, Florida. ? Experimentally determined in this study. analyses were conducted with Statistical Analysis Systems software using the General Linear Models Test (GLM) (SAS Institute, Inc. 1982). RESULTS Effects of Water Spray and Crawfish Loading on Purging The crawfish were observed to be completely purged (the intestine free of fecal matter) after 40 hours of spray purging (Fig. 4). Crawfish in the 0 L/min/m? control were also purged, but strong noxious odors were present. The animals in the control were moribund, and many died shortly after removal from the system. Crawfish in the con- trol were covered with slime and feces (Fig. 5), because DECHLORINATED CITY WATER MAINLINE WATER SPRAY NOZZLE LATERAL LINE Figure 3. Schematic of water spray system for each purging box. 386 TABLE 2. Percent crawfish mortality for four 40-hour purging tests in an experimental water spray purging system. Crawfish Water Spray Rate Loading Rate (L/min/m?) Replicate (kg/m?) 0 3.6 5.4 10.7 12.5 l 4.9 79.2 0 SI) 0 0 14.6 122 4.6 6.3 5.9 4.2 24.4 63.2 6.7 5.4 Shi7/ 10.0 34.1 45.0 11.5 9.2 12.1 12.5 2 4.9 70.4 She7/ 0 1.9 0 14.6 22.8 Shof/ Sho/ 4.9 4.1 24.4 20.4 4.1 5.9 5.9 5.9 34.1 21.4 6.4 ei 6.1 13.5 3 4.9 52.4 3.7, 7.4 Sel si 14.6 27.0 6.1 6.2 7.4 6.2 24.4 i723 6.3 7.4 ILS 5.6 34.1 15.8 9.2 16.7 14.6 8.2 4 4.9 5.9 Sf) She7/ 1.9 4.9 14.6 2.6 4.9 3.1 3.1 3.8 24.4 2.3 5.6 9.2 10.7 1.6 34.1 3.6 7.1 8.8 el 3)55) there was no water spray to wash away this material (Fig. 6). Quality-wise, these animals were considered unaccept- able. Replicate tests | and 3 had higher mortality than the other tests (Table 2). Larger crawfish (40 animals per kg) were used in tests | and 3, and the higher mortality ap- peared to be related to animals being crushed by the claws of the larger, more aggressive animals rather than from dessication or other causes. In replicate tests 2 and 4, the crawfish were smaller (50—60 per kg), aggressive behavior was less and mean mortality was lower. The highest mor- tality occurred in the control where crawfish were not sprayed with water. It was evident that some water spray is necessary to prevent massive mortality. Figure 4. Tail meats from purged (left) and un-purged (right) craw- fish. LAWSON ET AL. Figure 5. An un-purged crawfish showing characteristically unclean external surfaces. Effects of Crawfish Loading and Water Spray Rates Including 0 L/min/m? spray data, mean mortality for crawfish loadings of 4.9, 14.6, 24.4 and 34.1 kg/m? were 12.6, 10.1, 10.4 and 12.1%, respectively (P > 0.05) (Fig. 7). Mean mortality increased significantly with an increase in loading (P < 0.05), excluding data from the ‘‘no spray”’ control. Mean mortality was 2.8, 4.9, 6.6 and 9.8%, at loadings of 4.9, 14.6, 24.4 and 34.1 kg/m?. At the 0 L/min/m? spray rate crawfish mortality de- creased as loading increased. Crawfish were observed to cluster at the center of the compartments as loading in- creased, and they remained relatively inactive. The ‘‘clus- tering’’ of crawfish at higher densities increased survival by minimizing dehydration and reducing aggressive be- havior. Crawfish used in replicate 4 were loaded into the boxes immediately after harvest, thus retaining significant water in their brachial chambers and reducing the effects of dehydration during the test. Consequently, low crawfish mortality was observed in the 0 L/min/m? treatment relative to mortality in tests | through 3. The relationship between crawfish mortality and water , Figure 6. A purged crawfish showing clean external surfaces. PURGING CRAWFISH IN A WATER SPRAY SYSTEM Water Spray Rate — 2 [3.6 L/min/m 12.6 L/min/m> AM) 0 L/min/m? SH) 6.4 t/minsm. 10.7 Lymin/m> % Mortality 4.9 14.6 Crawfish Loading (kg/m ) 24.4 Figure 7. Relationship between crawfish loading rate and percent mortality at water spray rates of 0, 3.6, 5.4, 10.7 and 12.5 L/min/m?. spray rate at all loading densities, exclusive of 0 L/min/m? data are illustrated in Figure 8. At water spray rates of 3.6, 5.4, 10.7 and 12.5 L/min/m? mean mortality was 5.5, 6.5, 6.5 and 5.6%, respectively, and there was no difference in mean crawfish mortality between water spray rates (P > 0.05). DISCUSSION A minimal spray rate of 3.6 L/min/m? was adequate to provide sufficient moisture to maintain acceptable crawfish mortality while removing externally-adhered foreign matter from the animals and excreted fecal matter from the system. The lowest water spray rate of 3.6 L/min/m?, cor- responded to an application rate of 61 mm/hour (2.4 in/ hour) during the 40-hour purging period. This water appli- cation rate represents a relatively large volume of water ap- plied over a small area. Although water spray rates less than 3.6 L/min/m? were not evaluated, a lower water spray rate than 3.8 L/min/m? might be acceptable for use in com- mercial systems with significant savings in water and pumping costs and should be evaluated. At the lowest water spray of 3.6 L/min/m? and a craw- fish loading of 24.4 kg/m* mean mortality was less than Crawfish Loading 4.9 kg/m (0114.6 kg/m 24.4 kg/m. 34.1 kg/m % Mortality 3.6 6.4 Water Spray Rate (L/min/m ) Figure 8. Relationship between water spray rate and mortality for crawfish loading rates of 4.9, 5.4, 10.7 and 12.5 kg/m?. This figure does not include control data. 10.7 12.6 10% for the 40-hour period (less than 5% per day), and crawfish were satisfactorily purged. In the commercial sector, mortalities exceeding 5% per day in purging systems are not acceptable. Therefore, from this study we recommend that crawfish loading be kept below 24.4 kg/m? at a recommended water spray rate of 3.6 L/min/m?. Crawfish in this study remained in the boxes for 40 hours to insure complete purging. In this study and others (Lawson & Baskin 1985, Lawson & Drapcho 1989) it was determined that crawfish will purge satisfactorily in about 30 hours at water temperatures ranging from 22 to 28°C. Presumably, mortality would be less if crawfish were kept in purging systems for less time. However, this may not always be practical, since purging systems often provide a means to keep crawfish alive for several days without re- frigeration. Research should continue to observe if crawfish will require less time to purge at other water temperatures. Another point meriting attention is that related to size of crawfish in purging systems. It has been suggested that, if size graded before entry into purging systems, larger craw- fish will not be able to prey upon smaller animals, and mor- tality losses will be less. This question should also be ad- dressed by additional research. REFERENCES Avault, J. W. & J. V. Huner. 1985. Crawfish culture in the United States. In Huner J. V. & E. E. Brown (eds.), Crustacean and Mollusk Aqua- culture in the United States, AVI Publishing Company, Inc., West- port, CT, pp. 1-54. Dellenbarger, L. E., K. J. Roberts, S. S. Kelley & P. K. Pawlyk. 1986. An analysis of the crawfish processing industry and potential market outlets. Research Report No. 64, Department of Agricultural Eco- nomics and Agribusiness, Louisiana State University, Baton Rouge, LA. Dellenbarger, L. E., R. Hinson & A. Schupp. 1990. Agricultural mar- keting for 1990. Louisiana Rural Economist 52(1):2—5. Gaude, A. 1983. Procambarus clarkii in Spain. Crawfish Tales 2:15—17. Huner, J. V. 1989. Overview of international and domestic freshwater crawfish production. J. Shellfish Research 8:259—265. Lalla, H. & T. B. Lawson. 1987. Depuration of crawfish with a water spray. Paper No. 87-5034. American Society of Agricultural Engi- neers, St. Joseph, MI. Lawson, T. B. & G.R. Baskin. 1985. Crawfish holding and purging systems. Paper No. 85-5008. American Society of Agricultural Engi- neers, St. Joseph, MI. Lawson, T. B. & C. M. Drapcho. 1989. A comparison of three crawfish purging treatments. J. Aquacultural Engineering 8(5):339—347. SAS Institute, Inc. 1982. SAS User’s Guide: Statistics. SAS Institute, Inc., Gary, NC. 584 p. Journal of Shellfish Research, Vol. 9, No. 2, 389-393, 1990. EVALUATION OF ALTERNATIVE COOKING SCHEMES FOR CRAWFISH PROCESSING! GEORGE R. BASKIN? AND JOHN HENRY WELLS Department of Agricultural Engineering Louisiana State University Agricultural Center Baton Rouge, LA 70803-4505 ABSTRACT Commercial extraction of tail meat from red swamp crawfish, Procambarus clarkii (Girard, 1852) is a manual proce- dure with associated cost, productivity, and contamination problems. A commercial vegetable/fruit steam peeler was modified and evaluated for crawfish processing application. A decrease in cooking time and an increase in meat yield was observed with steam processing. Physical and microbial characteristics of steam-cooked meat compared favorably with meat that had been cooked using conventional boiling methods. Computer simulation was used to compare two processing schemes for cooking crawfish—boiling and steam processing. Simulation allowed comparison of the relative performance of each processing scheme with respect to time in process and process throughput. The simulation model for steam cooking exhibited a 350% throughput increase compared with traditional boiling methods. Steam processing of crawfish promises to increase processing productivity without compromising product quality. KEY WORDS: crawfish, cooking, processing INTRODUCTION Crawfish are fresh water crustaceans and the largest freshwater aquaculture in Louisiana with more than 132,000 acres of ponds and abundant wild stock. Total har- vest in 1988 was over 106 million pounds (Louisiana Coop- erative Extension Service, 1989). Procambarus clarkii and Procambarus acutus acutus are the two commercially im- portant species processed for food. Approximately 40% of the crawfish produced are processed into cooked whole crawfish or cooked extracted tail meat. While most of the crawfish are consumed in the state, international and do- mestic markets for processed whole crawfish and meat is increasing and is important to the state’s economy. Consis- tent product quality and the cost of processing are two im- portant concerns of the industry (Huner & Barr 1984). With conventional cooking methods, crawfish are placed in metal mesh baskets and boiled in large steam- jacketed or gas-fired kettles. Subsequently, the highly valued tail meat is manually extracted in one piece (or “‘peeled’’) from the exoskeleton. Such processing opera- tions are labor intensive and generally incorporate low levels of processing technology. Process automation and optimal management strategies could greatly benefit the in- dustry. Since manual peeling of tail meat constitutes the largest portion of processing cost, the aim of this investiga- tion was to evaluate alternative processing methods for pro- cessing efficiency and product consistency. Steam infusion is a widely used method of removing skin from fruits and vegetables. Steam peelers can improve yield recovery for certain products peeled by mechanical or lye peelers. The items to be “‘peeled’’ are placed in a pres- ‘Approved for publication by the Director of the Louisiana Agricultural Experiment Station as Manuscript No. 89-07-3489. ?Current address Sverdrup Corporation, 801 North Eleventh, St. Louis, MO 63101. 389 surized vessel and subjected to superheated steam at 517 to 1379 kPa (75 to 200 psi). Steaming destroys the outer pa- renchymal cells, to a depth controlled by a combination of pressure and heating exposure time. At the end of a prese- lected exposure time, the pressure chamber is rapidly vented and the interaction of steam expansion and destruc- tion of cells causes the skin to separate from the flesh. Steam infusion has been proposed as an alternative pro- cessing technology for crawfish cooking and tail meat ex- traction. Proposed changes in processing schemes can be evalu- ated effectively with mathematical modelling and computer simulation. Alternative systems can be simulated at various static and dynamic operating levels and overall perfor- mance compared at each level of operation. Rumsey (1986) modeled the operation of a fluid food evaporator with the simulation language ACSL (Advanced Control Simulation Language) and Shah et al. (1985a and 1985b) used the lan- guage SLAM (Simulation Language for Alternative Mod- eling) to model the operations and production scheduling of a meat processing plant. These studies demonstrated that alternative processing configurations can be evaluated through numerical investigation prior to in-plant installa- tion. The broad aim of this study was to evaluate an alterna- tive crawfish processing method. Specific objectives were to: 1) Evaluate a steam peeling application to extract craw- fish tail meat; 2) Evaluate the texture and microbial charac- teristics of steam processed crawfish tail meat; and 3) Eval- uate, with computer simulation, potential efficiency gains in steam processing vs. conventional boiling methods. MATERIALS AND METHODS Steam Processing Alternative A schematic diagram of the steps in the steam peeling process is shown in Figure 1. Compared to fruits and vege- 390 BASKIN AND WELLS 2. Steam In > 1. Product In 3. Kettle Closed Kettle Closed Steam Pressurized 5. Steam Out 6. Rotation 4. Kettle Closed 7. Kettle Open Steam Heating Product Out Figure 1. Schematic diagram for processing crawfish with steam in- fusion. Processing steps include: 1) Product is placed into processing kettle and the kettle is sealed; 2) Kettle is rapidly pressurized by steam infusion; 3) Pressurized steam permeates the skin; 4) Steam heating of product in relation to process time; 5) Pressure is rapidly released allowed steam in the skin to expand at the lower pressure; 6) Kettle is rotated to agitate product and loosen skin; and 7) Kettle is opened, product and condensed steam is discharged. tables, the investigators believed crawfish would require a much more aggressive peeling action. After preliminary in- vestigation, peeling action was found to be limited by the speed at which the exhaust valve could be opened to begin decompression and the speed at which the steam could be released after the valve was opened. The standard exhaust valve opening time was decreased by 75% and the exhaust steam flow was increased by 400%. Higher steam pressures were also found to increase destruction of the exoskeleton. To evaluate the process as a peeling aid, the steam pres- sure was maintained at its maximum value, 758 kPa (110 psi). The steam processing times investigated ranged from 10 to 45 sec. Experimentation with a 200 liter vegetable/ fruit steam peeler (modified as described above) demon- strated that the brittle exoskeleton of the crawfish could not be completely separated from the tail meat. However, it was determined that crawfish could be cooked with steam infusion. To evaluate steam infusion as a cooking method for whole crawfish, tail meat yield and ease of peeling were compared with crawfish cooked in boiling water. Tail meat yield and ease of peeling were evaluated for conventional boiling and steam cooking methods by a team of unskilled workers. Washed, live crawfish were cooked in 2500 g sample sizes for each steam cooking time. Each batch was weighed and allowed to cool for several minutes to facilitate handling. Tail meat was manually extracted, weighed and the total time to peel the sample was recorded. Yield was calculated as the percent of tail meat extracted from the live weight sample and the ease of peeling as amount of tail meat peeled per unit time. The peeled tail meat was placed on ice in preparation for product evalua- tion. Product Evaluation Texture evaluations were performed on steam cooked crawfish tail meat using an Instron Universal Testing Ma- chine (Model-1122) fitted with a Kramer shear cell. Set- tings of 100-g full scale, 100 mm/min crosshead speed and 100 mm/min chart speed were used. Sample sizes were 20- to 24-g and peak force values were used to determine the shear force (kg/g) for each sample. Microbiological profiles of the steam processed meat ‘were determined for selected processing trials that were perceived to be optimally processed for peeling and cooking. A subsample of the peeled tail meat was placed in a sterile bag, placed on ice, and delivered to the laboratory for microbial analysis. Microbial evaluations were con- ducted by an independent laboratory and included aerobic plant count (APC), enumeration of total and fecal coli- forms, and a test for E. coli. Coliforms analyses were con- ducted by the three-tube, most probable number (MPN) procedure (AOAC, 1987). All other procedures conformed to standard laboratory methods and procedures for micro- bial evaluations. Simulation of Alternative Processing Schemes The simulation language SLAM II is a combined dis- crete and continuous simulation language often used to ex- amine service related problems in operations management (e.g., Customer waiting time, optimal number of service resources, etc.) Discrete events are represented in SLAM as a connected network though which simulated entities flow. Entities may be created and terminated, delayed, combined, and placed into a state of waiting (queued), and/ or assigned selected attributes. The cookroom layout use for developing the simulation model for steam processing alternative is shown in Figure 2 and the network representa- tion of the SLAM model for used for simulation of the con- ventional boiling process is shown in Figure 3. An explana- tion of the network representation of a SLAM model is available in the simulation reference by Pritsker (1986). Alternative methods for cooking crawfish were simulated using the microcomputer implementation of SLAM (SLAM II/PC, Pritsker & Associates, Inc., West Lafayette, In- diana) on a MS-DOS compatible computer. RESULTS AND DISCUSSION Quality of Steam Processed Crawfish The results of product quality evaluation for crawfish processed with steam infusion are shown in Table 1. Peak force values were used to determine the shear force (kg/g) for each sample. Average Instron shear values observed for CRAWFISH PROCESSING 39] RECEIVING DOCK =x m ba PERSON RAW 5 @ PRODUCT PERSON 3 3 COOLER @ | WASH / GRADE 2 2 5 = STAGING > AREA i= COOKED COOKROOM LAYOUT WASTE PRODUCT CONVEYOR CHUTE 44 FT. X 16 FT. THREE EMPLOYEES Figure 2. Cookroom layout for steam cooking scheme. Steps in raw product flow follows include: 1) receiving dock; 2) scale; 3) raw product staging area (or raw product cooler); 4) wash/grade con- veyor; 5) steam peeler; and 6) cooked product chute. the steam processed meat ranged from 1.06 to 2.05 kg/g. These results compare favorably with the average Instron shear value ranging from 1.04 to 1.69 kg/g that were re- ported by Marshall et al. (1987) for conventionally boiled crawfish. Ease of peeling, measured in quantity of meat peeled per hour, was increase by 22% with steam pro- cessing. There was no indication that steam process time influenced ease of peeling. Meat yield from steam pro- cessed crawfish averaged 20% of live weight across all test. Moody (1980) reported an average yield of 15% from boiled crawfish. Bacteriological profiles did not detect fecal coliform and E. coli in the processed samples. Fecal coliform and E. coli were not detectable in all steamed samples. Total coliform was maintained at 2.3 MPN/g or less. Total aerobic plate count (APC) for the selected samples (Table 1) was main- tained below 100,000 CFU/g, the level considered to be acceptable within the industry (Moertle et al. 1985). Fur- ther evaluation of the steam processing method is needed to determine if proteinases are inactivated by the cooking method. Undercooking of whole crawfish can lead to mushiness of peeled tail meat if packaged with adhering hepatopancreas (Marshall et al. 1987). Comparison of Cooking Alternatives The simulation study compared two crawfish processing schemes—cooking in boiling water and cooking with steam infusion. SLAM simulation models were developed for boiling and steam processing schemes each having identical product input and output constraints. The number of batches processed and the time in process served as per- formance criteria to compare processing alternatives. The simulation model tracked the flow of entities through the processing activities within the cookroom (e.g, raw material handling, product washing/grading, and cooking). Raw material handling and washing/grading were modeled in the same manner for both processing schemes, but each method of cooking dictated different pro- cessing times and batch sizes as well as dissimilar operating constraints. Steam processing, for instance, necessitated cooking crawfish in smaller batch sizes because of the lim- ited volume of the steam peeler. The batch sizes for steaming and boiling processes were 60 lb. and 200 lb., respectively. Additionally, the boiling process required specialized resources (e.g., an overhead crane to handle baskets of crawfish), periodic disposal of cook water to maintain sanitary conditions, and time to refill and reheat the cook kettles between batches. Because of the differ- ences inherent to each system, product throughput and time in system were different for each processing alternative. Each processing scheme was evaluated for the industry standard 40-lb. sack of live crawfish (process input) and one pound of crawfish tail meat (process output) based on a 15% yield in both cases. The simulation results for an 8-hr working day comparing boiling and steam processing are given in Table 2. Processing crawfish by steam infusion resulted in a 350% increase in throughput and a 92% de- crease in time spent in the system. In addition to the sub- stantial increase in product throughput, the decrease in time spent in the system could have an impact on product quality. It is expected that a shorter length of time in pro- cess would reduce the extent of product deterioration and Overall Process Simulation Figure 3. Network representation of SLAM model for cooking craw- fish with boiling water (Overall Process Simulation) and for periodic changing of the cook water (Simulation of Cookpot Cleaning). 392 BASKIN AND WELLS TABLE 1. Effect of steam processing time on selected quality characteristics of peeled crawfish tail meat. Process Time Tail Meat Yield! Shear Force? APC? (sec) (% live weight) (kg/g) (CFU/g) 10 22.0 1.06 1S 21.3 1.60 20 20.5 1.73 4500 25 20.0 1.58 3700 30 19.3 t/t) 3900 35 18.9 1.85 40 18.2 2.05 45 17.7 ' Average of replications at each process time. ? Kramer shear cell peak force per unit weight of crawfish tail meat. 3 Total aerobic plate count. microbial contamination associated with conventional methods of crawfish cooking. The simulation detailed only the activities of the cookroom and assumed that other departments within the processing plant were sized to accommodate the product throughput processed by a particular cooking scheme. While cooking is generally not the processing bottle neck in a plant operation, the simulation results could be used to identify the potential impact of changes in cooking tech- nology on other plant operations and set design criteria for new processing operations (Escobar & Wells 1990). In the example indicated, a change in cookroom processing oper- ation results in a greater than three-fold increase in product throughput. This would indicate that additional capacities in product delivery, secondary processes, packaging, and storage would be needed in order to take full advantage of the newly introduced cooking technology. Implications for Processing Facilities Design Typically, new equipment within an existing processing plant will be implemented with a minor case evolution in the facility layout (Vollmann & Buffa 1966). Minor case evolution is predicated on near-term objectives and usually does not justify complete redesign of the facility. In minor case evolution the layout is generally viewed as a group of interactive subsystems with placement of a new equipment being dictated by existing layout and operating constraints. However, in the case where introducing equipment with processing capacities dramatically different from that of the substitution equipment, a minor case evolution will not ac- commodate process changes as the new technology cannot be introduced within the existing constraints. Such would be the case when introducing the steam cooking of crawfish into an existing processing plant. Introduction of new technology greatly exceeding the existing capacities of the remaining plant operations may require extensive retrofitting throughout the plant and may justify a major case evolution. In contrast to a minor case evolution, a major case evolution deals with process layout changes that occur at a single point in time and must reflect the long-term objectives of the processing facility. For the major case evolution, a process layout design is viewed as a single monolithic system. Generally, the boundary con- straints encountered in minor case evolution are not present in the major case evolution, and various computer-assisted techniques and design algorithms based on material han- dling cost and capital resource allocation have been devel- oped (Baskin 1989). The use of simulation to evaluate the impact of processing technology (illustrated by the evalua- tion of steam cooking of crawfish) serves as an example to introduce and highlight the study of systems-planning re- search. In the context of a minor facilities evolution arising from the introduction of new technology, the goal of a systems- planning methodology should be to maintain efficient utili- zation of all resources, both equipment and personnel, even in the event that technology advances. As suggested by the simulation study of steam cooking crawfish introduction of new technology can have a dramatic impact on processing capacity that in the absence of changes in interacting activi- ties (e.g., receiving, peeling, packaging, etc.) would not necessarily decrease overall efficiency. From a systems- planning prospective, the simulation techniques demon- strated in this study can be used to evaluate the impact of process mechanization and/or automation on existing oper- ations. Such techniques can be used in planning the addi- tional resources needed to accommodate implementation of the new technology (e.g., expansion of shipping and re- ceiving departments, storage capacity, labor requirement, etc.). Traditionally, process simulation models used to select processing equipment capacities represent the predominant computer-assisted technique employed by food engineers in TABLE 2. Comparison of conventional crawfish processing methods (BOIL) and processing with steam infusion (STEAM) for an 8-hr work day. Time in Processing Batches! Input? System? Output* Method Processed (sacks) (minutes) (Ibs) BOIL 30 152 77.1 900 STEAM 348 523 5.0 3,132 1 Batch size was 200 Ibs live crawfish for BOIL and 60 Ibs live crawfish for STEAM. 2 Input entities were based on 40 Ibs average weight of a standard sack of live crawfish. 3 Includes time to change and heat water for conventinal boiling method. 4 Output in pounds of cooked and peeled crawfish tail meat based on 157% yield of live weight input. CRAWFISH PROCESSING 39 the design of a food processing plant (Havlik et al. 1987). Layout design algorithms, used to arrange processing de- partments (including equipment) into a feasible layout, are utilized to a much lesser extent. An accurate approximation of the amounts of product flow between the separate unit operations within the facility is required for proper design with the layout algorithms. Estimation of product flow is often difficult and cumbersome to obtain since each layout problem is made unique by its particular constraints and assumptions. Logically, process design models could be used to estimate materials flows and should provide a nat- ural precursor for layout algorithms. An algorithm com- bining the two concepts would provide food plant designers with a useful tool with which to conduct sensitivity anal- yses, thereby exploring the full implications of changes in critical unit operations. However, no technique integrating these two concepts has been developed. Techniques that are available in related disciplines are frequently inappropriate for use in the food industry be- cause of the stringent regulatory environment. The food in- dustry is in need of computer-aided methods for food pro- cessing plant design and operations management. These methods must incorporate both process simulation and fa- cilities layout design considerations if a comprehensive ap- Ww proach to food processing plant design is to be realized. Such research will become increasingly important as man- datory inspection based on the hazard analysis of critical control points (HACCP) concept are imposed nationwide on seafood and shellfish processors by federal regulatory agencies. CONCLUSIONS The specific conclusions of this research are enumerated below: 1. A commercial vegetable/fruit steam peeler can be modi- fied to cook crawfish with no indicated loss of product yield. 2. The textual quality, as measured by Kramer shear force, of crawfish cooked with steam infusion compares favor- ably to conventionally boiled product. 3. Further evaluation of the steam cooking method is needed to determine minimum process times that over- come mushiness of peeled tail meat packaged with ad- hering hepatopancreas. 4. Simulation of alternative processing schemes can pro- vide insight into plant operations and assist in evalu- ating the impact of processing technology on plant operations. REFERENCES AOAC. 1987. Official Methods of Analysis. Association of Official Agri- cultural Chemists, Washington, D.C. Baskin, G. R. 1989. ARCH: A Robust Construction Heuristic for the layout design of food processing facilities. Ph.D. Dissertation. Loui- siana State University. Escobar, F. A. & J. H. Wells. 1990. Simulation based performance anal- ysis of crawfish processing operations. Paper No. 615, presented at the 1990 Annual Meeting of the Institute of Food Technologist., Ana- heim, CA. June 16—20. Havlik, S., P. Moyer & M. R. Okos. 1987. Computed-aided food process design. ASAE Paper No. 87-6567. Presented at the 1987 Winter Meeting of the American Society of Agricultural Engineers. Chicago, IL. December. Huner, J. V. & J. E. Barr. 1984. Red Swamp Crawfish: Biology and Ex- ploitation.Center for Wetland Resources, Louisiana State University. p.75. Louisiana Cooperative Extension Service. 1989. Louisiana Summary — Agricultural and Natural Resources. Louisiana Cooperative Extension Service, Louisiana State University. April. Marshall, G. A., M. W. Moody, C. R. Hackney & J. S. Godber. 1987. Effect of blanch time on the development of mushiness in ice-stored crawfish meat packed with adhering hepatopancreas. J. Food Sci. 52(6):1504— 1506. Moody, M. W. 1980. Louisiana seafood delight—the crawfish. Loui- siana Cooperative Extension Service, Louisiana State University. Bul- letin LSU-TL-80-002. Moertle, G. M., M. W. Moody & C. R. Hackney. 1985. Processing time effects on the texture of fresh packed crawfish meat. Paper No. 252, presented at 45th Annual Meeting of Inst. of Food Technologists, At- lanta, GA, June 9-12. Pritsker, A. B. A. 1986. Introduction to Simulation and SLAM II, 3rd Edition. Systems Publishing Corp. West Lafayette, Indiana. Rumsey, T. R. 1986. Dynamic analysis of a double effect evaporator using advanced continuous simulation language (ACSL). ASAE Paper No. 86-6537. Presented at the 1986 Winter Meeting of the American Society of Agricultural Engineers. Chicago, IL. December. Shah, S. A., M. R. Okos & G. V. Reklaitis. 1985a. A SLAM based computer simulation model of a meat processing plant. Transactions of the ASAE 28(5):1698—1703. Shah, S. A., M. R. Okos & G. V. Reklaitis. 1985b. Production sched- uling in food processing plants. Transactions of the ASAE 28(5): 2078-2082. Vollmann, T. E. & E. S. Buffa. 1966. The facility layout problem in perspective. Management Science 12(10):B450—B468. 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 resourses. The Association has a membership of more than 900 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. —lIf you are young, you will benefit from the experience of your elders. —lIf you are an elder, you will be rejuvenated by the fresh ideas of youth. —lIf you are a student, you will make most 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 30 US $ per year ($20 for students) and that includes the Journal and the Newsletter! NATIONAL SHELLFISHERIES ASSOCIATION—APPLICATION FOR MEMBERSHIP (NEW MEMBERS ONLY) Nae ee OT the calendar) car Dates Mailing address: Institutional affiliation, if any: Shellfishery interests: Regular or student membership: ————___ Student members only—advisor’s signature REQUIRED: Make cheques (MUST be drawn on a US bank) or international postal money orders for $30 ($20 for students with advisor’s signature) payable to the National Shellfisheries Association and send to Dr. Tom Soniat, Dept of Biology, University of New Orleans, New Orleans, Louisiana 70148 USA. aa (Gish. INFORMATION FOR CONTRIBUTORS TO THE JOURNAL OF SHELLFISH RESEARCH Oniginal papers 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 paper should be carefully prepared in the style followed in Volume 8 Number 2, of the Journal of Shellfish Research (1989) be- fore submission to the Editor. Papers published or to be published in other journals are not acceptable. Title, Short Title, Key Words, and Abstract: The title of the paper should be kept as short as possible. Please include a ‘‘short running title’’ of not more than 48 char- acters including space between words, and approximately seven (7) key words or less. Each manuscript must be ac- companied by a concise, informative abstract, giving the main results of the research reported. The abstract will be published at the beginning of the paper. No separate sum- mary should be included. Text: Manuscripts must be typed double-spaced throughout one side of the paper, leaving ample margins, with the pages numbered consecutively. Scientific names of species should be underlined and, when first mentioned in the text, should be followed by the authority. Abbreviations, Style, Numbers: Authors should follow the style recommended by the fourth edition (1978) of the Council of Biology Editors [CBE] Style Manual, dis- tributed by the American Institute of Biological Sciences. All linear measurements, weights, and volumes should be given in metric units. Tables: Tables, numbered in Arabic, should be on sepa- rate pages with a concise title at the top. Illustrations: Line drawing should be in black ink and planned so that important details will be clear after reduc- tion 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 preferably should be pre- pared 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 re- duction. Originals of graphic materials (i.e., line drawings) are preferred and will be returned to the author. Each illus- tration 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. No color illustrations will be accepted unless the author is prepared to cover the cost of associated reproduction and printing. References Cited: References should be listed alphabet- ically at the end of the paper. 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 at the end of papers in Volume 8, Number 2, of the Journal of Shellfish Research or refer to Chapter 3, pages 51—60 of the CBE Style Manual. Page Charges: Authors or their institutions will be charged $50.00 per printed page. If illustrations and/or tables make up more than one third of the total number of pages, there will be a charge of $30.00 for each page of this material (calculated on the actual amount of page space taken up), regardless of the total length of the article. All page charges are subject to change without notice. Proofs: Page proofs are sent to the corresponding author and must be corrected and returned within seven days. Al- terations other than corrections of printer’s errors will 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: Particularly appropriate photo- graphs may be submitted for consideration for use on the cover of the Journal of Shellfish Research. Black and white photographs, if utilized, are printed at no cost. Color illus- trations may be submitted but all costs associated with re- production and printing of such illustrations must be cov- ered by the submitter. Corresponding: An original and two copies of each manuscript submitted for publication consideration should be sent to the Editor, Dr. Sandra E. Shumway, Department of Marine Resources, and Bigelow Laboratory for Ocean Science, West Boothbay Harbor, Maine 04575. Phone: 207-633-5572 FAX: 207-633-7109 JOURNAL OF SHELLFISH RESEARCH Vol. 9, No. 2 December 1990 CONTENTS RADI GES? ganas nbusos oocedG boone dno nUrB od OU DED OU TOUT AUS OG GERD OOD DOO Amn DC ORE tS OOx: R. A. Rose, R. E. Dybdahl and S. Harders Reproductive cycle of the Western Australian silverlip pearl oyster, Pinctada maxima (Jameson) (Mollusca:Pteriidae) A. Campbell, N. Bourne and W. Carolfeld Growth and size at maturity of the Pacific gaper Tresus nuttallii (Conrad, 1837) in Southern British Columbia Katsuhiko T. Wada, John Scarpa and Standish K. Allen Karyotype of the dwarf surfclam Mulinia lateralis (Say, 11822) (Mactridae) Bivalvia) jeje. ee ee ee erate cele oel eared S. M. C. Robinson and T. W. Rowell A re-examination of the incidental fishing mortality of the traditional clam hack on the softshell, Mya arenaria [eintaATen nT wa obs LoS nde ben doo nc dno douoodes Gonoador Men eon s0 data d chommy OO NomEe peo do Odo bas ¢ T. W. Rowell and P. Woo Predation by the nemertean worm, Cerebratulus lacteus Verrill, on the softshell, Mya arenaria Linnaeus, 1758, and its apparent role in the destruction of a clam flat ...........- 0.0002 se este tenets t erent teeters see e ces Ann E. Bass, Robert E. Malouf and Sandra E. Shumway Growth of northern quahogs (Mercenaria mercenaria (Linnaeus, 1758)) fed on picoplankton .........--....----- Arnold G. Eversole, Joy G. Goodsell and Peter J. Eldridge Biomass production and turnover of northern quahogs, Mercenaria mercenaria (Linnaeus, 1758), at different densities ALGIERS, ooognooacccssoomppoohmundededgonsbe ac nos coUNGoUsSOOnta cocgmun cn DO por oR oR OOD OORT Dan C. Marelli and William S. Arnold Estimates of losses associated with field depuration (relaying) of Mercenaria spp. in the Indian River Lagoon, Florida Roger Mann and Julia S. Rainer Effect of decreasing oxygen tension on swimming rate of Crassostrea virginica (Gmelin, 1791) larvae ...........- Paul D. Kenny, William K. Minchener and Dennis M. Allen Spatial and temporal patterns of oyster settlement in a high salinity estuary .........----+-2se eee eee e sete eee G. Curtis Roegner and Roger Mann Settlement patterns of Crassostrea virginica (Gmelin, 1791) larvae in relation to tidal zonation .............--.--- Raymond E. Grizzle Distribution and abundance of Crassostrea virginica (Gmelin, 1791) (eastern oyster) and Mercenaria spp. (quahogs) in AGUS IFEOON 2cpcecodqune peseo 060 CoamMOsoumanoeae sbarQh er SaqnenDOGbTS oKRS Rael bar co poR eco George C. Matthiesssen, Sung Y. Feng and Louis Leibovitz Patterns of MSX (Haplosporidium nelsoni) infection and subsequent mortality in resistant and susceptible strains of the eastern oyster Crassostrea virginica (Gmelin, 1791), TINEA) Sacgcepasbocpotonpecsocecuspccaucsd oot Julie D. Gauthier and William S. Fisher Hemolymph assay for diagnosis of Perkinsus marinus in oysters Crassostrea virginica (Gmelin, 1791) .....--++--: Walter T. Momot, P. Lynn Hauta and James A. Schaefer Yield estimates for the virile crayfish Orconectes virilis (Hagen, 1870), employing the Schaefer logistic model ...... Thomas B. Lawson, Harnarine Lalla and Robert P. Romaire Purging Crawfish ina water spray System) 22-2. 50 062i nee eines ei es oie siaaas George R. Baskin and John Henry Wells Evaluation of alternative cooking schemes for crawfish processing ......--- 600+ + esses eee reser eet tresses COVER PHOTO: Cerebratulus lacteus Verrill on mudflat. Photo courtesy of T. W. Rowell (see p. 291). 283 341 347 359 367 373 383 389 MBL/WHOI LIBRARY AIM WH LAAT - Santa nin gs th def ride 34k: yrues peace cere ets iin teaNbend it. ey a thse BSice aaah SSNS ok ah as SO rrrer cer ine on b ava ikes sth Iasdumed ae 8! | ay ee dea Satie! Brera yemaneerey Abin sabe a yet 2 nlnthamaietay om est Cd AA ete te ete sees 9 aaa") dhene sil teas . SySenmaboaleapipodar one afi ap aAND OBE SUA ER sae eres SVs enn it . a eae hte, igeia ege nab dan s8 p> Tales whee os eas fat Be eaten ie Beate rere patra Pa ean Serr Serer eres ue he 7 rarer Tetris rare ose mbes eusee cast a te ee Daratars pptbatigeas in SPs, ee sry Pre rt as DGGE he Tialebatatcieteth oath seeder PUL Saa gers eh vam entsges Bescreiet ita eet sbeaaat seth dare tsyrere Se Done wo venegreeeraieeh Titieh i Cocantatmate ecard ry ernest Ram MWe Epon pee TE TS Se A Venps S04 eam shpat Omienyry EPH FAH EE pea True emmy fe eee ewe tence rece me yet fob plontane tp teteakd och eponebtatdl lh atir a prummeeoteenese riety re on SPW agy Net ne Pesta corre s we ey Pye yeyeviy 4 /mebers ¥/W 4 * ied Lies nearmeamante acs = teens ep tataties Hien ihhetae Td iatese iy Arid Serle rH ea ery ree. SenP hve wey gee owe TAverbe sway oleh Taare” thine s trkeerette hina epee hee? Veh ig eee wap eeeeens oman tate ro Pepys -: a. eee elevate pes Yr oe Sipe er! eae rf Tere tweuies eae Bhs AS Cy De ? irs ioe) Peer taret vie ar aa Se rus fe er Se ee Tere tens ys Nae Whe sey a a Uric Went 8 aU Be “a