Fishery Bulletin Vol. 82, No. 1 January 1984 ROPES, JOHN W., DOUGLAS S. JONES, STEVEN A. MURAWSKI, FREDRIC M. SERCHUK, and AMBROSE JERALD, JR. Documentation of annual growth lines in ocean quahogs, Arctica islandico Linne 1 BOWMAN, RAY E. Food of silver hake, Merluccius bilinearis 21 LARSON, RALPH J., and EDWARD E. DeMARTINI. Abundance and vertical distribu- tion of fishes in a cobble-bottom kelp forest off San Onofre, California 37 COYER, JAMES A. The invertebrate assemblage associated with the giant kelp, Mac- rocystis pyrifera, at Santa Catalina Island, California: A general description with emphasis on amphipods, copepods, mysids, and shrimps 55 ANTONELIS, GEORGE A., JR., CLIFFORD H. FISCUS, and ROBERT L. DeLONG. Spring and summer prey of California sea lions, Zalophus californianus, at San Miguel Island, California, 1978-79 67 GRIS WOLD, CAROLYN A., and THOMAS W. McKENNE Y. Larval development of the scup, Stenotomus chrysops (Pisces: Sparidae) 77 HETTLER, WILLIAM F. Description of eggs, larvae, and early juveniles of gulf menhaden, Brevoortia patronus , and comparisons with Atlantic menhaden,B. tyrannus, and yellowfin menhaden, B. smithi 85 BARNETT, ARTHUR M., ANDREW E. JAHN, PETER D. SERTIC, and WILLIAM WATSON. Distribution of ichthyoplankton off San Onofre, California, and methods for sampling very shallow coastal waters 97 McGURK, MICHAEL D. Ring deposition in the otoliths of larval Pacific herring, Clupea harengus pallasi 113 MACDONALD, J. STEVENSON, MICHAEL J. DADSWELL, RALPH G. APPY, GARY D. MELVTN, and DAVID A. METHVEN. Fishes, fish assemblages, and their seasonal movements in the lower Bay of Fundy and Passamaquoddy Bay, Canada 121 TILSETH, S., and B. ELLERTSEN. The detection and distribution of larval Arcto- Norwegian cod, Gadus morhua, food organisms by an in situ particle counter 141 EWING, R. D., C. E. HART, C. A. FUSTICH, and GREG CONCANNON. Effects of size and time of release on seaward migrations of spring chinook salmon, Oncorhynchus tshawytscha 157 CAMPANA, STEVEN E. Interactive effects of age and environmental modifiers on the production of daily growth increments in otoliths of plainfin midshipman, Porichthys notatus 165 V (Continued on hack cover) Seattle, Washington Fishery Bulletin CONTENTS 1985 Vol. 82, No. 1 January 1 984 ROPES, JOHN W., DOUGLAS S. JONES, STEVEN A. MURAWSKI, FREDRIC M. SERCHUK, and AMBROSE JERALD, JR. Documentation of annual growth lines in ocean quahogs, Arctica islandico Linne 1 BOWMAN, RAY E. Food of silver hake, Merluccius bilinearis 21 LARSON, RALPH J., and EDWARD E. DeMARTINI. Abundance and vertical distribu- tion of fishes in a cobble-bottom kelp forest off San Onofre, California 37 COYER, JAMES A. The invertebrate assemblage associated with the giant kelp, Mac- rocystis pyrifera, at Santa Catalina Island, California: A general description with emphasis on amphipods, copepods, mysids, and shrimps 55 ANTONELIS, GEORGE A., JR., CLIFFORD H. FISCUS, and ROBERT L. DeLONG. Spring and summer prey of California sea lions, Zalophus californianus, at San Miguel Island, California, 1978-79 67 GRISWOLD, CAROLYN A., and THOMAS W. McKENNE Y. Larval development of the scup, Stenotomus chrysops (Pisces: Sparidae) 77 HETTLER, WILLIAM F. Description of eggs, larvae, and early juveniles of gulf menhaden, Brevoortia patronus , and comparisons with Atlantic menhaden, B. tyrannus , and yellowfin menhaden, B. smithi 85 BARNETT, ARTHUR M., ANDREW E. JAHN, PETER D. SERTIC, and WILLIAM WATSON. Distribution of ichthyoplankton off San Onofre, California, and methods for sampling very shallow coastal waters 97 McGURK, MICHAEL D. Ring deposition in the otoliths of larval Pacific herring, Clupea harengus pallasi 113 MACDONALD, J. STEVENSON, MICHAEL J. DADSWELL, RALPH G. APPY, GARY D. MELVIN, and DAVID A. METHVEN. Fishes, fish assemblages, and their seasonal movements in the lower Bay of Fundy and Passamaquoddy Bay, Canada 121 TILSETH, S., and B. ELLERTSEN. The detection and distribution of larval Arcto- Norwegian cod, Gadus morhua, food organisms by an in situ particle counter 141 E WING, R. D., C. E. HART, C. A. FUSTICH, and GREG CONCANNON. Effects of size and time of release on seaward migrations of spring chinook salmon, Oncorhynchus tshawytscha 157 CAMPANA, STEVEN E. Interactive effects of age and environmental modifiers on the production of daily growth increments in otoliths of plainfin midshipman, Porichthys notatus 165 (Continued on next page) Seattle, Washington 1984 For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402 - Subscription price per year $21.00 domestic and $26.25 foreign. Cost per single issue: $6.50 domestic and $8.15 foreign. ( 'ontents — continued LOVE, MILTON S., GERALD E. McGOWEN, WILLIAM WESTPHAL, ROBERT J. LAVENBERG, and LINDA MARTIN. Aspects of the life history and fishery of the white croaker, Genyonemus lineatus (Sciaenidae), off California 179 MORRIS, PAMELA A. Feeding habits of blacksmith, Chromis punctipinnis , associated with a thermal outfall 199 MYRICK, ALBERT C., JR., EDWARD W. SHALLENBERGER, INGRID KANG, and DAVID B. MacKAY. Calibration of dental layers in seven captive Hawaiian spinner dolphins, Stenella longirostris, based on tetracycline labeling 207 ROSS, STEVE W. Reproduction of the banded drum, Larimus fasciatus, in North Carolina 227 Notes SCHMITT, P. D. Marking growth increments in otoliths of larval and juvenile fish by immersion in tetracycline to examine the rate of increment formation 237 ENNIS, G. P. Tag-recapture validation of molt and egg extrusion predictions based upon pleopod examination in the American lobster, Homarus americanus 242 ENNIS, G. P. Comparison of physiological and functional size-maturity relationships in two Newfoundland populations of lobsters Homarus americanus 244 ECHEVERRIA, TINA, and WILLIAM H. LENARZ. Conversions between total, fork, and standard lengths in 35 species of Sebastes from California 249 The National Marine Fisheries Service (NMFS) does not approve, recommend or en- dorse am proprietary product or proprietary material mentioned in this publication. No reference shall be made to NMFS. or to this publication furnished by NMFS, in any ad- vertising oi sales promotion which would indicate or imply that NMFS approves, recom- mends or endorses any proprietary product or proprietary material mentioned herein, or which has as its purpose an intent to cause directly or indirectly the advertised pro- duct to be used or purchased because of this NMFS publication. \ \ \ A A ^ ^ \ I <® ^ % Best NMFS Publications for 1982 The Publications Advisory Committee of the National Marine Fisheries Service has announced the best publica- tions authored by the NMFS scientists and published in the Fishery Bulletin and the Marine Fisheries Review for 1982. Only effective and interpretive articles which sig- nificantly contribute to the understanding and knowledge of NMFS mission-related studies are eligible, and the following papers were judged as the best in meeting this requirement. "Development of the vertebral column, fins and fin supports, branch iostegal rays, and squamation in the swordfish, Xiphias gladius" by Thomas Potthoff and Sharon Kelley appears in Fishery Bulletin 80(2): 161-186. Thomas Potthoff, fishery biologist, and Sharon Kelley, research assistant, are from the Southeast Fisheries Center's Miami Laboratory, Miami, Fla. "A review of the offshore shrimp fishery and the 1981 Texas closure" by Edward F. Klima, Kenneth N. Baxter, and Frank J. Patella, Jr. appears in Marine Fisheries Review 44(9- 10): 16-30. Edward F Klima, Director of the Galveston Laboratory, Kenneth N. Baxter, supervisory fishery biologist, and Frank J. Patella, Jr. , fishery biologist, are also from the Southeast Fisheries Center but from the Galveston Laboratory, Galveston, Tex. Bo r$ £U DOCUMENTATION OF ANNUAL GROWTH LINES IN OCEAN QUAHOGS, ARCTICA ISLANDICA LINNE John W. Ropes,1 Douglas S. Jones,2 Steven A. Murawski,1 Fredric M. Serchuk,1 and Ambrose Jearld, Jr.1 ABSTRACT About 42,000 ocean quahogs, .Arcif'ca islandica Linne, were marked and released at a deep (53 m) oceanic site off Long Island, New York, in 1978. Shells of live specimens recovered 1 and 2 years later were radially sec- tioned, polished, and etched for preparation of acetate peels and examination by optical microscopy or micro- projection; selected specimens were similarly prepared for examination by scanning electron microscopy. Specific growth line and growth increment microstructures are described and photographed. An annual periodicity of microstructure is documented, providing a basis for accurate age analyses of this commercially important species. Numerous bivalve species form periodic growth lines in their shells (Rhoads and Lutz 1980). Internal growth lines found in the shells of ocean quahogs, Arctica islandica Linne, have stimulated interest in using these markings to determine age and growth (Thompson et al. 1980a, b), since fishery exploitation has increased significantly within the past decade (Serchuk and Murawski 19801). Documentation of age and growth of ocean quahogs has been incomplete. Some studies included no account of aging methodologies (Thorson in Turner 1949; Jaeckel 1952; Loosanoff 1953; Skuladottir 1967); in others, concentric "rings" or "bands" formed in the periostracum of small quahogs (ggild (1930:286) reported that he was unable to confirm Sorby's observations. Instead he stated that Arctica islandica belongs to a group of species within the Arcticidae (= Cyprinidae) having the least visible structure among all the bivalves. He terms this structure homogeneous but suggests there are small traces of other structures in the shell. Boggild (1930) goes on to point out that the lower 16 ROPES ET AL.: GROWTH LINES OF OCEAN OIAHOGS part of the shell (inner layer) is perhaps more "... representative of the common, complex struc- ture . . . and . . . there are alternating layers of more transparent layers and finely grained ones." More recently Taylor etal. (1969, 1973) examined the shell microstructure ofArctica islandica, which they adopt- ed as their "type species" to illustrate homogeneous shell microstructure. Basically, the general picture by Btfggild (1930) agrees with that of Taylor et al. (1969), who used electron microscopy in their inves- tigation. However, they disagreed sharply with B^ggild that the inner shell layer was "representative of the common complex structure." After examining unetched fractured sections and polished and etched sections of both shell layers, Taylor et al. (1969, 1973) concluded that both shell layers in Arctica islandica are built of minute, irregular rounded granules, quite variable in size (1.5-3 fim across), having highly irregular contacts with their neighbors and being poorly stacked. Taylor et al. (1969:51) further reported: "In peels and sections of the inner layer, within the pallial line there is a marked colour banding, in greys and browns. The only fine structure that can be resolved is a suggestion of minute grains, which are most conspicuous in the translucent, grey-colourless parts of the shell. These grains are arranged in sheets parallel to the shell interior. In the outer layer grains can also be resolved, but are arranged in sheets parallel to the margin of the shell and growth lines." They also noted that these features are more clearly seen in the umbonal region where the orientation of grains normal to layering is very conspicuous. Taylor et al. (1969) suggested that the layering is a reflection of repeated (?diurnal) deposition of carbonate (a prospect deemed very unlikely by Thompson et al. 1980a). Also in the umbonal region are thin (2-3 ju.m) prismatic bands which parallel the layering. Outside the pallial line, Taylor et al. (1969) reported the outer shell layer to be very dense and opaque, with the most obvious structural features being fine grains arranged in sheets giving the layer a finely banded appearance. Analyses under SEM of oriented fractured, and polished and etched sections of ocean quahog shells revealed that microstructural variation is more com- plex than had been proposed by Btfggild (1930) or Taylor et al. (1969, 1973). Thin sections of isolated periostracal fragments examined under crossed nicols confirmed the presence of embedded aragonite granules in the periostracum of ocean quahogs reported for other recent bivalves (Carter and Aller 1975). These granules probably form a layer like that described for the blue mussel, Mytilus edulis, by Carriker (1979). After special treatment of the valves for examination by SEM, he found "a thin discrete calcareous layer continuous over the outer surface of the valves between the periostracum and the outermost shell layer." The layer is called mosaio- stracum. The shell microstructure in the growth incre- ment sublayer beneath the periostracum is HOM, as B0ggild (1930) and Taylor et al. (1969, 1973) re- ported. The "... minute, irregular, rounded gran- ules . . . have highly irregular contacts ..." (Taylor et al. 1969:51) that are particularly well exposed in fracture sections. An abundant transitional CA- CL microstructure was found in the middle portion of the outer shell layer and growth increment sub- layer. This study confirmed its presence in ocean quahogs as reported by Carter (1980). The growth line sublayer of the outer shell layer had four grades of prismatic structure (FP, SphP, ComP, and ISP). Lutz and Rhoads (1977) examined the inner shell layer near the umbo of ocean quahogs and found bands of simple aragonitic prisms alternating with complex-crossed lamellar and homogeneous struc- tures. We found similar microstructures in the inner shell layer of the valve of ocean quahogs. Our analyses identified distinct microstructures, not unlike those found in the valve for the growth line and growth increment layers in the hinge plate. Growth line deposition more nearly approximates an annual event than any shorter or longer interval. Marked clams recovered in late August 1979 had formed only one growth line other than the mark- induced check soon after the notching operation in 1978. They had been free about 22 d longer than a calendar year. Those recovered in early September 1980 all had formed the growth line soon after the notching operation, like those recovered in 1979, and a second line appeared midway to the ventral valve edge, which in all probability had been formed after the late August 1979 recovery effort. These clams were free about 33 d more than 2 calendaryears since the notching operation. A feature of the specimens recovered in 1980 was that about half had formed a third line very near the ventral valve edge and along the inner margin. All of the narrow growth lines were separated by relatively even, broad areas of growth increment deposits suggestive of no more or less than an annual interval for the deposition of growth lines, even though the time of formation of such lines may not correspond to an exact number of calendar days. These observations confirm similar conclusions of an annual periodicity of growth line formation by Thompson and Jones (1977), Thompson et al. (1980a, b), and Jones (1980). Radiometric techniques for aging bivalve shells have recently been applied to ocean quahogs. 17 FISHERY BULLETIN: VOL. 82, NO. 1 Thompson et al. (1980a) reported that the predicted radiometric age of an ocean quahog having 22 bands corresponded exactly to 22 yr when aged using 228Ra. Turekian et al. (1982) concluded that age deter- minations of ocean quahogs from radiometric analyses are compatible with counts of bands formed annually. Thus, radiometric studies support the con- tention of an annual periodicity of growth lines in ocean quahogs. Various environmental disturbances have been implicated in the formation of shell abnormalities and atypical growth lines in other bivalve species (Weymouth et al. 1925; Shuster 1957; Merrill et al. 1966; Clark 1968; Palmer 1980). It is therefore, con- ceivable that the stress imposed by dredging, mark- ing, and returning the ocean quahogs to the ocean floor and their burrowing activities hastened the for- mation of a growth line in 1978. Thereafter, natural events affecting the metabolism of shell deposition are more likely stimuli. Such events apparently did not occur during the period after the formation of the growth line in 1978 and recovery of clams in late August 1979. Instead a growth line that in all prob- ability had formed in 1979 was found in the shells of clams recovered on 9 September 1980. Its formation may have occurred in late August 1979, but the third line found in half of the clams recovered on 9 Septem- ber 1980 suggests the possibility of its formation in early September 1980. By inference, then, growth line formation in 1979 and 1980 occurred in September. The reported life span (150 yr, Thompson et al. 1980a) of ocean quahogs surpasses similar estimates for other bivalves. Age and growth of the far east mussel, Crenomytilus grayanus, have been deter- mined from examinations of shell structure, an oxygen-isotope method, and notching experiments (Zolotarev 1974; Zolotarev and Ignat'ev 1977; Zolotarev and Selin 1979). These investigations indicated that longevity of the mussel may exceed 100 yr. Turekian et al. (1975) proposed a longevity of about 100 yr for a deep-sea nucoloid, Tindaria callis- tiformis, after determining ages by radiometric means and counting regularly spaced bands in the shell of one of the largest (8.4 mm in shell length). It seems likely that longevity of ocean quahogs may exceed 150 yr. Murawski and Serchuk (1979) report- ed a maximum shell length of 131 mm for ocean quahogs in extensive collections taken from the Mid- dle Atlantic Bight. A specimen of this size is half again as large as the 88 mm example of a 149-yr-old ocean quahog reported by Thompson et al. (1980a). In conclusion, the foregoing description of annual growth line formation in marked ocean quahogs and analyses of growth in the same specimens by Murawski et al. (1982) present significant supporting evidence for the hypothesis of slow growth and a long life span in the species. Ocean quahogs apparently live longer than any other bivalve known to man. ACKNOWLEDGMENTS We thank Brenda Figuerido and John Lamont for their assistance in preparing the art work and photographs, and Ida Thompson, University of Edin- burgh, Department of Geology, King's Building, Edinburgh EH 9 3JW, Scotland, for encouragement in undertaking the study and helpful comments on the manuscript. LITERATURE CITED B0GGILD, O. B. 1930. The shell structure of mollusks. K. Dan. Vidensk. Selsk. SK (Copenhagen) 2:231-325. Carriker, M. R. 1979. Ultrastructure of the mosaicostracal layer in the shell of the bivalve Mytilus cdulis. Veliger 21:41 1-414. Caddy, J. F., R. A. Chandler, and D. G. Wilder. 1974. Biology and commercial potential of several underex- ploited molluscs and crustaceans on the Atlantic coast of Canada. Proceeding of a Symposium on the Industrial Development Branch of Environmental Canada, Mon- treal, Feb. 5-7, 1974, 111 p. (Prepared at Fish. Res. Board St. Andrews Biol. Stn., N.B.) Carter, J. G. 1980. Guide to bivalve shell microstructures. In D. C. Rhoads and R. A. Lutz (editors), Skeletal growth of aquatic organisms, p. 645-674. Plenum Press, N.Y. Carter, J. G., and R. C. Aller. 1975. Calcification in the bivalve periostracum. Lethaia 8:315-320. Chandler, R. A. 1965. Ocean quahog resources of Southeastern Northum- berland Strait. Fish Res. Board Can., Manuscr. Rep. (Biol.) 828, 9 p. Clark, G. R., II. 1968. Mollusk shell: Daily growth lines. Science (Wash., D.C.) 161:800-802. 1974a. Growth lines in invertebrate skeletons. Annu. Rev. Earth Planet. Sci. 2:77-99. 1974b. The Paleoperiodicity Newsletter. Vol. 1, p. 1-2. Gordon, J., and M. R. Carriker. 1978. Growth lines in a bivalve mollusk: Subdaily patterns and dissolution of the shell. Science (Wash., D.C.) 202:519-521. Jaeckel, S., Jr. 1952. Zur Oekologie der Molluskenfauna in der westlichen Ostsee. Schr. Naturwiss. Ver. Schleswig-Holstein 26:18- 50. Jones, D. S. 1980. Annual cycle of shell growth increment formation in two continental shelf bivalves and its paleoecologic significance Paleobiology 6:331-340. 18 ROPES ET AI. .: GROWTH LINES OF OCEAN QUAHOGS LOOSANOFF, V. L. 1953. Reproductive cycle in Cyprina islandica. Biol. Bull. (Woods Hole) 104:146-155. LOVEN, P. M. 1929 Bietrae zur Kenntnis der Cyprina islandica L. in Ore- sund (Contributions to the knowledge of ( 'yprina islandica L. in the Oresund). K. Fysiogr. Sallsk. Lund Handl. N.F. 41:1-38. [Transl. by Lang. Serv. Div., Off. Int. Fish., NMFS, U.S. Dep. Commer., Wash., D.C.] LUTZ, R. A., and D. C. Rhoads. 1977. Anaerobiosis and a theory of growth line for- mation. Science (Wash., D.C.) 198:1222-1227. 1980. Chapter 6. Growth patterns within the molluscan shell. An overview In D. C. Rhoads and R. A. Lutz (editors), Skeletal growth of aquatic organisms, p. 203-254. Plenum Press, N.Y. Merrill, A. S., J. A. Posgay, and F. E. Nichy. 1966. Annual marks on shell and ligament of sea scallop Placopecten magellanicus. U.S. Fish Wildl. Serv., Fish. Bull. 65:299-311. Mi rawski, S. A., J. W. Ropes, and F. M. Serchuk. 1982. Growth of the ocean quahog, Arctica islandica, in the Middle Atlantic Bight. Fish. Bull., U.S. 80:21-34. Ml IRAWSKl, S. A„ AND F. M. SERCHl'K. 1979. Shell length-meat weight relationships of ocean quahogs, Arctica islandica, from the Middle Atlantic Shelf. Proc. Natl. Shellfish. Assoc. 69:40-46. Palmer, R. E. 1980. Observations on shell deformities, ultrastructure, and increment formation in the bay scallop Argopecten irradians. Mar. Biol. (Berl.) 58:15-23. Rhoads, D. C, and R. A. Lutz (editors). 1980. Skeletal growth of aquatic organisms. Plenum Press, N.Y., 750 p. Ropes, J. W., and A. S. Merrill. 1970. Marking surf clams. Proc. Natl. Shellfish. Assoc. 60:99-106. Shuster, C. N., Jr. 1957. On the shell of bivalve mollusks. Proc. Natl. Shellfish. Assoc. 47:34-42. Skuladottir, U. 1967. Krabbadyr og skeldyr (Crustaceans and mollusks). Radstefna Isl. Verkfraedinga. 52:13-23. Proceedings of the Conference of Islandic Professional Engineers. [Transl: Transl. Bur., Fish. Res. Board Can., Biol. Stn., St. Andrews, N.B., No. 1206.] SORBY, H. C. 1879. Address on the structure and origins of limestones. Q. J. Geol. Soc. Lond. 35:56-95. Taylor, J. D., W. J. Kennedy, and A. Hall. 1969. The shell structure and mineralogy of the Bivalvia. Introduction. Nuculacea-Trigonacea. Bull. Br. Mus. (Nat. Hist.) Zool. Suppl. 3:4-125. 1973. The shell structure and mineralogy of the Bivalvia. II. Lucinacea - Clavagellacea. Conclusions. Bull. Br. Mus. (Nat, Hist.) Zool. 22:255-294. Thompson, I., and D. S. Jones. 1977. The ocean quahog, Arctica islandica, "tree" of the North Atlantic shelf. [Abstr.| Geol. Soc. Am. 9:1 199. Thompson, I., D. S. Jones, and D. Dreibelbis. 1980a. Annual internal growth banding and life history of the ocean quahog Arctica islandica (Mollusca: Bivalvia). Mar. Biol. (Berl.) 57:25-34. Thompson, I., D. S. Jones, and J. W. Ropes. 1980b. Advanced age for sexual maturity in the ocean quahog Arctica islandica (Mollusca: Bivalvia). Mar. Biol. (Berl.) 57:35-39. Turekian, K. K., J. K. Cochran, D. P. Kharkar, R. M. Cerrato, J. R. Valsnys, H. L. Sanders, J. F. Grassle, and J. A. Allen. 1975. Slow growth rate of a deep-sea clam determined by 228Ra chronology. Proc. Natl. Acad. Sci. 72:2829-2832. Turekian, K. K., J. K. Cochran, Y. Nozaki, I. Thompson, and D. S. Jones. 1982: Determination of shell deposition rates of Arctica islan- dica from the New York Bight using natural "8RA and 228Th and bomb-produced 14C. Limnol. Oceanogr. 27: 737-741. Turner, H. J., Jr. 1949. The mahogany quahog resource of Massachusetts. In Report on investigations of improving the shellfish re- sources of Massachusetts. Commonw. Mass. Dep. Con- serv., Div. Mar. Fish., p. 12-16. Weymouth, F. W., H. C. Mcmillin, and H. B. Holmes. 1925. Growth and age at maturity of the Pacific razor clam, Siliqua patula (Dixon). U.S. Dep. Commer., Bur. Fish. Doc. 984, p. 201-236. ZOLOTAREV, V. N. 1974. Determination of age and growth rate of the far east mussel, Crenomytilus grayanus (Dunker), from its shell structure. Dokl. Akad. Nauk SSSR 216(5): 1 195-1 197. ITransl. by Plenum Publ. Corp., N.Y., 1974, p. 308- 309.] Zolotarev, V. N., and A. V. Ignatev. 1977. Seasonal changes in the thickness of the main layers and temperature growth of the marine molluscan shells. Biol. Morya 5:40-47. |Transl. by Pelnum Publ. Corp., N.Y., 1978, p. 352-358.] Zolotarev, V. N., and N. I. Selin. 1979. The use of tags of shells to determine growth of the mussel Crenomytilus grayanus. Biol. Morya 1:77-79. (Transl. by Plenum Publ. Corp., N.Y.. 1979, p. 58-59.) 19 FOOD OF SILVER HAKE, MERLUCCIUS BILINEARIS Ray E. Bowman1 ABSTRACT Stomach contents of 2,622 silver hake collected in the Northwest Atlantic have been analyzed. Fish were collected on bottom trawl surveys conducted from 1973 to 1976. The mean fish fork length (FL) was 20 cm and the average stomach content weight was 1.5 g. Silver hake <20 cm FL prey mostly on amphipods, decapod shrimp, and euphausiids. Fish 20 cm FL and longer take increasing proportions of fish and squid as part of their diet. Stomach contents of male and female fish of similar size indicate that females eat larger quantities of food (particularly more fish) than the males. The females are also, on the average, longer than the males. Silver hake feed primarily at night. Feeding begins near dusk and continues until just after mid- night. In the spring a second feeding period seems to occur near noon. Silver hake feed intensively during spring. Their stomachs contain almost twice as much food in spring as they do in autumn. Significant dif- ferences were noted in the intensity of feeding between areas. Stomachs of fish, caught in the Middle Atlantic, contain the largest quantities of food. The species of prey taken by silver hake are highly variable and likely reflect prey availability during different years and seasons in various areas. When silver hake spawn, their dietary intake is reduced. The diet of fish taken in deep water (> 150 m) is mostly euphausiids and squid, and the quantity of food found in their stomachs is less than that in stomachs taken from fish collected at depths <150m. Silver hake, Merluccius bilinearis (Mitchill 1814), is a Northwest Atlantic gadiform fish whose range ex- tends from continental shelf waters off South Caro- lina to the Newfoundland Banks. It is most abundant in offshore waters extending from New York to Cape Sable, Nova Scotia (Bigelow and Schroeder 1953). Previous investigations have shown that large silver hake eat mostly fish and/or squid, while smaller silver hake feed on euphausiids, amphipods, and decapod shrimp. Among the first to report these findings were Nichols and Breder (1927), who noted 75 herring about 7 cm long in the stomach of a 59 cm fish. Bigelow and Schroeder (1953) reported that silver hake are extremely voracious and will prey on smaller silver hake or any other of the schooling fishes such as young herring, mackerel, menhaden, alewives, or silver- sides. Evaluation of other studies on the diet of silver hake caught in various areas and during different years establishes that the prey of silver hake is very predictable in that it is usually comprised of a variety of fish, squid, and crustaceans (Jensen and Fritz 1960; Schaefer 1960; Vinogradov 1972; Noskov and Vinogradov 1977; Bowman and Langton 1978; Lang- ton and Bowman 1980). Investigations by Swan and Clay (1979), Edwards and Bowman (1979), and Bow- man and Bowman (1980) have shown that silver hake feed mostly at night. Until recently the potential impact of silver hake on 'Northeast Fisheries Center Woods Hole Laboratory, National Marine Fisheries Service, NOAA, Woods Hole, MA 02543. Manuscript accepted July 1983. FISHERY BULLETIN: VOL. 82, NO. 1, 1984. the Northwest Atlantic ecosystem had not been de- termined. Edwards and Bowman (1979) estimated the annual consumption of the principal predators in the Northwest Atlantic. They concluded that silver hake alone could potentially consume almost 10% of the standing crop of all fish within the study area an- nually, the bulk of which would be small or juvenile fish. They suggested that silver hake, more than any other species, plays the principal predatory role in regulating the Northwest Atlantic ecosystem. The purpose of this report is to document the quantities and types of food eaten by silver hake during the years 1973-76, and further, to identify feeding trends which may be of consequence when attempting to precisely determine silver hake's impact on other fish populations. METHODS AND MATERIALS A total of 325 samples from 2,622 silver hake stomachs was collected during eight MARMAP (Mar- ine Resources Monitoring, Assessment, and Predic- tion) bottom trawl survey cruises conducted by the National Marine Fisheries Service during spring and fall 1973-76 (Table 1). The cruise periods were as follows: 16 March-15 May 1973; 26 September-20 November 1973; 12 March-4 May 1974; 20 Sep- tember-14 November 1974; 4 March-12 May 1975; 15 October- 18 November 1975; 4 March-8 May 1976; 20 October-23 November 1976. On spring cruises a two-seam modified Yankee No. 4 1 trawl was 21 FISHERY BULLETIN: VOL. 82. NO. 1 TABLE 1.— Number of silver hake stomachs examined from each geographic area by year and season. Nurr ber examined Southern Year Season Middle Atlantic New England Georges Bank 1973 Spring Fall 39 144 105 129 48 191 1974 Spring Fall 189 S4 93 117 103 157 1975 Spring Fall b8 100 120 92 146 1976 Spring Fall 1 1 1 93 125 129 63 115 Totals 789 918 915 recorded. A stomach was considered empty when no food items could be identified and the material found in the stomach weighed <0.001 g. Data were ana- lyzed with FORTRAN IV programs written for use on a Honeywell SIGMA 73 computer system located in Woods Hole, Mass. Food data are presented in terms of the mean stomach content weight, adjusted stomach content weight (discussed below), and the percentage weight 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. fished, and during fall cruises a standard Yankee No. 36 was used. The cod end and upper belly of both trawls were lined with 13 mm mesh netting to retain smaller fish. A scheme of stratified random trawling was conducted within the study area (Fig. 1), and fishing continued over 24 h/d2. All tows were 30 min in duration at a vessel speed of 3.5 kn in the direction of the next station. Sampling of stomachs was concentrated in three areas: Middle Atlantic, Southern New England, and Georges Bank (Fig. 1). Fish within two length groups (>20 cm and <20 cm) were randomly selected (50 fish/group) during each cruise from the bottom trawl survey catches in each area. At each station within a particular area no more than 10 fish were taken for each of the two length groups, and fish were not sam- pled at two consecutive stations. The only exception to this collection method occurred when it appeared (during the cruise) that 50 large or 50 small fish would not be collected within a particular area. In this case, all fish caught were collected in an attempt to obtain the minimum sample size. Stomachs of large fish were excised aboard ship; individually wrapped in gauze with a label denoting vessel, cruise, species, fork length (FL), sex, and maturity; and preserved in 3.7% formaldehyde (small fish were preserved whole). In the laboratory the preserved stomachs were in- dividually opened, and their contents emptied onto a 0.25 mm mesh opening screen sieve to permit wash- ing without loss of any food items. The stomach con- tents were sorted, identified, counted, and damp dried on absorbent paper. Major prey items and com- monly occurring but relatively minor prey, in terms of weight, were identified to species whenever possible. The wet weight of all stomach content groups was determined to the nearest 0.00 1 g and all information 'Further details of the bottom trawling techniques may be obtained from the Resource Surveys Investigation, Northeast Fisheries Cen- ter Woods Hole Laboratory, National Marine Fisheries Service, NOAA, Woods Hole, MA 02543. \ S % \ ■h. cO /' V Portland ,'/T| f GULF OF •\ f MAINE ♦o IS \ FIGURE 1.— Offshore areas sampled during bottom trawl surveys conducted by the Northeast Fisheries Center between the years of 1973 and 1976, inclusive. 22 BOWMAN: FOOD OF SILVER HAKE each prey group made up of the total stomach con- tents weight. All tables follow a standard format to aid in making comparisons. In the tables, subtotals of the percentage weight of major stomach content groups are offset to the left. The minor prey groups are discussed in further taxonomic detail in the text. Adjusted stomach content weights are weights ad- justed by a correction factor which allows direct com- parison of the stomach content weights of different- sized fish. Adjustment of the stomach content weights was necessary, before any quantitative com- parisons could be made between variables such as sex or area. Observations on stomach tissue weight (excluding contents), mean stomach content weight, and whole fish weight (Fig. 2) revealed that neither the mean stomach content weight nor the stomach tissue weight is proportional to the body weight of different-sized fish. Stomach tissue weights of 526 silver hake were gathered during a study jointly con- ducted by American and Soviet scientists on Georges Bank, September 1978, aboard the Soviet RV Belo- gorsk (operated by the Atlantic Research Institute of Marine Fisheries and Oceanography, Kaliningrad, USSR). Mean stomach content weight data were derived from the 1973-76 food data given in this report, and the fish body weights were calculated us- ing the silver hake length-weight equation described bv Wilk et al. (1978). Silver hake weighing < 100 g, or >300 g, have larger stomachs (stomach tissue weight being an indication of stomach size), and stomachs which contain on the average more food in terms of percentage body weight, than fish weighing between 100 and 300 g. Since both the stomach tissue weight and the mean stomach content weight were dis- proportionate when presented as percentage body weight for different-sized fish (but were generally proportionate relative to each other), and because the mean stomach content weight data was much more variable than the stomach tissue weight data, the data adjustment was based on stomach tissue weight rather than on body weight or mean stomach content weight. The following equation was used to adjust the stomach content weights: AL = xl wl where A= Adjusted stomach content value. The adjusted stomach content value was converted to grams by multiply- ing it by the stomach tissue weight of a 30 cm FL fish. xl = Mean stomach content weight of all fish at a given length. wl = Mean stomach tissue weight of silver hake at a given length. The adjusted stomach content data for fish 4 (0.3 g) to 15 (21 g) cm FL and 24 (90 g) to 35 (292 g) cm FL are presented separately in forthcoming sections. 3 0 2 5 O i STOMACH TISSUE WEIGHT/ BODY WEIGHT EXPONENTIAL CURVE FIT r2 = 091, 0 =0 107, b *0092 x STOMACH CONTENT WEIGHT/ BODY WEIGHT ,2 ., EXPONENTIAL CURVE FIT r^ = O 94, a = 0 006, b = 0 170 200 300 BODY WEIGHT (G) 400 500 FIGURE 2. — Percentage body weight made up by the stomach tissue weight and the stomach content weight of different size silver hake. Area enclosed by solid lines represents more than 80% (excluding juveniles) of the silver hake population (fish 2-7 yr old), based on survey data. Stomach tissue weight/fish length and stomach content weigh t/fish length data were fit to an exponential curve (formjy = aebx). The data are presented in terms of body weight for illustrative purposes. 23 FISHERY BULLETIN: VOL. 82. NO. 1 These two length groups were chosen because the food consumption of fish < 1 yr old (4- 1 5 cm FL) dif- fers substantially from the food consumption of older fish (evident from Figure 2). In addition, too few fish outside these length ranges were sampled to warrant inclusion in any of the calculations dealing with com- parisons between data sets. An analysis of variance (one way) was used to test the observed differences among sample means (e.g., between geographic areas). RESULTS The contents of 2,622 silver hake stomachs, of which 803 (30.4%) were empty, were analyzed. Fish sampled averaged 20 cm FL and had, including the empty ones, a mean stomach content weight of 1.5 g. Sources of potential variation in the data presented below include size, sex, and maturity stage offish, as well as the time of day, area, year, season, bottom depth, and temperature when or where the fish were caught. Each variable considered in this analysis is treated separately, i.e., the data were pooled over other variables with no attempt to determine the possible confounding effects of different variables on the results. Dietary trends noted within each par- ticular variable examined should be considered only as preliminary observations. Composition of the Diet Overall, in terms of percentage weight, the diet of silver hake consists almost entirely of fish (80.0%), crustaceans (10.2%), and squid (9.2%), as can be seen in Table 2 . The importance of crustaceans to the diet is overshadowed by the fish portion because large silver hake eat heavier meals consisting pri- marily of fish. However, Table 2 is useful because it serves as a composite list of the prey types commonly found in the stomachs of silver hake. Fish such as silver hake, Merluccius bilinearis; Atlantic mackerel, Scomber scombrus; butterfish, Peprilus triacanthus; herring (Clupeidae); American sand lance, Am- modytes americanus; scup, Stenotomus chrysops; At- lantic saury, Scomberesox saurus; and longfin hake, Phycis chesteri, each make up >0.1% of the stomach contents. The "Other Pisces" category, most of which could not be identified, accounts for a substan- tial portion (52.07c) of the "Pisces" group. Fishes which could be identified within this category (all contributed <0.1% to the diet) include summer flounder, Paralichthys dentatus; redfish, Sebastes marinus; codfishes (Gadidae); and flatfishes (Pleuro- nectiformes). Crustacea in the diet is represented principally by euphausiids (mostly Meganyctiphanes norvegica, 3.7%, and Euphausia, <0.1%) and decapods such as the Crangonidae (mainly Crangon septemspinosa, 1.4%, and Sclerocrangon boreas, <0.1%), Pandalidae (al- most exclusively Dichelopandalus leptocerus, 2.0%, although some Pandalus borealis, <0.1%, was also found), Pasiphaeidae (only Pasiphaea multidentata, 0.1%), and other unidentified decapods (0.4%) which were mostly shrimp (0.3%). Amphipods found in the stomachs consist primarily of the families Ampe- liscidae (<0.1% each of Ampelisca agaxxizi, A. spinipes, A uadorum, and Byblis serrata), Oedicerotidae (<0.1% of Monoculodes edwardsi and M. intermedius), and Hyperiidae (exclusively the genus Parathemisto, 0.1%). The remaining crusta- cean groups are the Mysidacea (comprised of Neomysis americana, 0.7%, and Erythrops, <0.1%), Cumacea (mostly Leptocuma, <0.1%, and some un- identified diastylids, <0.1%), Copepoda (almost all identified as calanoids, <0.1%), and "Other Crus- tacea" (all of which was well-digested crustacean remains, 0.3%). The only other stomach contents identified were the cephalopods (Loligo pealei, 4.17c, and Rossia, Table 2.— Dietary composition of 2,622 silver hake caught in the Northwest Atlantic during the years 1973-76. (+ indicates <0.1%.) Percentage Prey weight Polychaeta 0.1 Crustacea 10 2 Amphipoda 1.3 Ampeliscidae 1.0 Oedicerotidae 0.1 Hyperiidae 0.1 Other Amphipoda 0.1 Decapoda 39 Crangonidae 1.4 Pandalidae 2.0 Pasiphaeidae 0.1 Other Decapoda 0.4 Euphausiacea 40 Mysidacea 0.7 Cumacea + Copepoda + Other Crustacea 0.3 Cephalopoda 9.2 Loligo 76 Other Cephalopoda 1.6 Pisces 80.0 Scomberesox saurus 1.5 Clupeidae 2.7 Merluccius bilinearis 9.2 Phycis chesteri 02 Ammodytes americanus 1.8 Scomber scombrus 7.5 Stenotomus chrysops 1.6 Pepnlus triacanthus 3.5 Other Pisces 52.0 Miscellaneous 0.5 No. of stomachs examined 2.622 No. of empty stomachs 803 Mean stomach content weight (g) 1.477 Mean fish FL (cm) 20.3 24 BOWMAN: FOOD OF SILVER HAKE <0.1%), Polychaeta, and the "Miscallaneous" category, which consisted of small amounts (<0.1%) of Echinodermata, Chaetognatha, unrecognizable digested matter, and sand. The percentage weights of various prey of silver hake within specified length groups are listed in Ta- ble 3. Silver hake <20 cm FL eat mostly crustaceans (>80% on the average), whereas the food of in- dividuals >20 cm FL is mostly fish and squid (average over 50%). Stomachs of silver hake 3-5 cm FL contain the largest percentages of smaller crusta- cean forms, such as amphipods and copepods. Decapods, euphausiids, and mysids, which are generally larger organisms (see Gosner 1971), make up the largest percentage of the diet of fish 6-20 cmFL. Diet Differences Between Males and Females The diet of male and female silver hake differs in both quality and quantity of food (Table 4). The stomachs of males have the largest percentage of crustaceans, while those of females have the largest percentage offish and squid. The mean stomach con- tent weight of the males is only about one-fifth that of the females. Males also occur less frequently in the samples (42% of the fish collected were males) and are generally smaller than the females (mean FL males, 28.4 cm; females, 32.1 cm). Since female fish are, on the average, longer than the males, the dif- ferences noted above had to be dealt with in con- siderably more detail. A comparison of the data in Tables 5 (food of males) and 6 (food of females) indicates that males and females within the same size groupings consume dif- ferent types and amounts of food. The same dietary patterns noted for male and female fish in the preced- ing paragraph can be seen within most of the in- dividual length groups in these two tables (e.g., when males and females within the same size group are compared, the stomachs of the females contain larger quantities of food and higher percentages offish and squid). The number of males sampled generally ex- ceeds the number of females for length groups <30 cm, while females dominate the length groups >30 cm. A subset of the data were analyzed separately using only fish lengths for which 20 or more individuals each of males and females were sampled (Fig. 3). This group offish (ranging in FL from 24 to 34 cm) is fairly representative of the adult silver hake population sampled. The mean stomach content weight (Fig. 3A), percentage crustaceans (Fig. 3B), and per- centage fish and squid (Fig. 3C) data presented graphically illustrate the differences between the diet of male and female silver hake of the same length. The stomachs of females contain more food, on the average, than those of males; the stomachs of males contain higher percentages of crustaceans than females; and the stomachs of females contain more fish and squid than those of males. Adjustment (by stomach tissue weight) of the mean stomach con- tent weights given in Figure 3A revealed that the stomachs of females contain, on the average, 1.5 times the quantity of food found in the stomachs of males. o O u I o < O 30 2.0 10 100 80 60 40 20 100 80 60 40 20 1 1 I I 1 1 MEAN TOTAL CONTENTS ( G ) 1 T CRUSTACEANS L 24 25 26 27 28 29 30 31 32 3 3 34 FISH LENGTH ( FL IN CM ) FIGURE 3. — A) Mean stomach content weight of male and female silver hake versus fish length, B) percentage of total stomach con- tent weight made up by crustaceans for male and female silver hake, C) percentage of total stomach content weight made up by fish and squid for male and female silver hake. Diurnal Variation in Feeding Intensity The adjusted mean stomach content weight data presented in Figures 4 and 5 indicate the feeding periods of silver hake vary by season and size of fish. In autumn, the stomachs of larger fish (24-35 cm FL) are fullest just after midnight, while smaller fish (4-15 cm FL) have the fullest stomachs in late afternoon 25 FISHERY BULLETIN: VOL. 82, NO. 1 so 3! J3 5i 3 O c < s E o .S 3j J4 03 -C •-- o V 3 .5 en •- 4- -C -5 2P S is C c o o, S o CO C CD U tt CD Oh I w J X < E- III +6 | | II I I I I I I o I I + I I I I 111° |||+ oo | | I I I 00 o 111 + o cn w- o cn r~ r- if) r~ d | cb ir> | r-~ cb d | d + 11 + + + m cn d- I o n 6 t CO ID CN id co *? O co | H r^ CD 6 + CM CD CN ^ CN CN CNO^t «- CM If) CO o o> CD in »- cn r*- r- CO ^ CN CO CO CO CD O CM CM «- ** CO CM *- CO *- d d d d CO CO CO i- ^ ^ 6 — ID CD CM «- CO cb cb ^ co i-^ CO (O cb d + I o CN CO O CN in CN T CN ,- O O O o o CD CM CO o — o 1 CN to I - I ri I I «i I + I I I « CO ID d in — o d od CD + CO ID d do ID «- - r- CO cn cb CN d + ID CD O r- 00 cn S 3 ■o TO - "D hi , 9 5 e-S y-tcD^^cjirraro*- : E ro a to T3 5 2 -i ** CD cn E 1 ~ g|i ° E ^ -=■ U) O c t % > ° -J e £E "• 2 £°f u ^ - - z z 5 5 26 BOWMAN: FOOD OF SILVER HAKE and just after midnight (Fig. 4). During springtime, large silver hake have substantial quantities of food in their stomachs (almost twice as much as during autumn) for two time periods, one near dusk and the other just before noon. Smaller fish have the most food in their stomachs just after midnight during spring (Fig. 5). No indication of a particular prey being eaten at a particular time of day was noted. silver hake within all geographic areas. Silver hake caught in the Middle Atlantic have the highest per- centage offish in their diet (Middle Atlantic, 87.5%; Southern New England, 78.4%; Georges Bank, 76.4%), but most was unidentified (60.4%). Silver hake (20.8%) and herring (Clupeidae, 3.2%) make up Diet Within Geographic Areas Stomach content data for silver hake collected in various geographic areas (i.e., Middle Atlantic, Southern New England, and Georges Bank) are pre- sented in Table 7. Fish is by far the dominant prey of 10 i a I <_) I 9 I o (/I O 0 2 V- V) \ 0 1 0 AUTUMN 108 61 LARGE FISH - 64 87 98 49 (24-35cmFL) 22 72 88 SMALL FISH (4-15cmFL) 71 208 194 110 " 12 1 NOON 5 18 DUSK 1 24 0 MIDNIGHT 3 06 09 12 DAWN NOON FIGURE 4. — Adjusted mean stomach content weight of large (24-35 cm FL) and small (4-15 cm FL) silver hake collected in the autumn versus time of day. The number of fish sampled in each time period is given just above the histogram. TABLE 4.— Stomach contents of male and female silver hake collected in the Northwest Atlantic during 1973-76. Data are expressed as a percentage weight. (+ indicates <0.1%.) Prey Male Female Polychaeta 0.2 + Crustacea 35.0 4.5 Amphipoda 06 0.2 Ampeliscidae 0.2 0.1 Oedicerotidae 0.1 + Hyperudae 0.2 0.1 Other Amphipoda 0.1 + Decapoda 11.9 23 Crangonidae 5.1 0.6 Pandalidae 55 1 5 Pasiphaeidae — + Other Decapoda 1.3 02 Euphausiacea 18.8 1.7 Mysidacea 2.7 02 Cumacea + + Copepoda + — Other Crustacea 1.0 0.1 Cephalopoda 4.3 10.4 Loli go 3.4 8.6 Other Cephalopoda 0.9 18 Pisces 59.1 84.6 Scomberesox saurus — 18 Clupeidae — 32 Merluccius bttinearts 22.6 76 Phycis chesten — 0.2 Ammodytes amencanus 1.4 20 Scomber scombrus 3.8 8.4 Stenotomus chrysops — 1.9 Pepnlus tnacanthus 3 3 3.7 Other Pisces 280 55.8 Miscellaneous 1.4 05 No examined 613 842 No. of empty stomachs 252 354 Mean stom. cont. wt. (g) 0.85: 4204 Mean fish FL (cm) 28.4 32 1 Length range (cm) 6-59 7-64 o 20 t~ l o UJ fe 1.0 UJ Z O o X u < o 0.5 o UJ h- in § o SPRING LARGE FISH 49 (24-35cmFL) 53 101 59 34 37 I SMALL FISH - (4-15cmFL) 83 83 23 90 10 Lj 1 NO 1 b, 79 2 15 18 2 ON DUSK 1 24 C MIDNIGHT 3 06 0 DAWN 9 12 NOON Figure 5. -Adjusted mean stomach content weight of large (24-35 cm FL) and small (4-15 cm FL) silver hake collected in springtime versus time of day. The number of fish sampled in each time period is given just above the histogram. 27 FISHERY BULLETIN: VOL. 82, NO. 1 TABLE 5.— Composition of the diet of male silver hake in terms of percentage weight versus fish length. (+ indicates <0.1 %.) Length group (cm) Prey 5-10 11-15 16-20 21-25 26-30 31-35 36-40 >41 — — 0.3 + 03 — — 64 .1 97 2 29.3 73. 1 32.7 38 1.9 - 2 7 1.9 0.8 1.4 0.6 0.6 02 1.0 04 0 1 0.4 0.1 0 2 + 0.1 0.1 + 0.1 0.1 + + + + 1.9 1.9 1.1 0.6 0.5 10.7 1 5 H 3 0.9 19.1 7.4 9.8 1.9 15.0 7 7 5 7 1.6 2.6 0.9 1.7 1.5 0.1 1.4 50.3 92 7 14.2 41.4 15 9 + — 11 9 — — 10.4 0.8 0.2 0.4 — — 0.4 + + — — — 0.7 + — — — — — — 2 6 12 0.8 09 — Polychaeta — Crustacea 19.2 Amphipoda — Ampeliscidae — Oedicerotidae — Hyperudae — Other Amphipoda — Decapoda 0.3 Crangonidae — Pandalidae — Pasiphaeidae — Other Decapoda 0.3 Euphausiacea — Mysidacea 11.3 Cumacea — Copepoda — Other Crustacea 7,6 Cephalopoda — 4.4 02 83 2.5 Lohgo — — — — — 8.1 — — Other Cephalopoda — — + 4 4 0.2 0.2 2 5 — Pisces 714 216 — 64 1 23.6 57 2 93 7 98.1 Scomberesox saurus — — — — — — — — Clupeidae — — — — — — — — Merluccius bilmeans — — — 10.0 5 0 7.7 70.0 66.2 Phycis chesten — — — — — — — — Ammodytes amencanus Scomber scombrus Stenotomus chrysops Pepnlus tnacanthus Other Pisces Miscellaneous 94 No. examined No. empty Mean stom. cont. wt. (g) Mean fish FL (cm) 50.8 21.6 — — — 3.1 9 .' — 20.6 4 14.3 + 2.8 + 54 1 1.9 18 6 3.1 1.5 80 29.2 23.7 3 12 4 0030 84 5 0 0.435 13.4 20 4 0.414 19.1 119 50 0400 23,7 248 109 0.456 28 5 178 73 1215 32 .2 21 9 3 565 37.1 8 3 7.282 509 Table 6.- — Composition of the diet of female silver hake in terms of percentage weight versus fish length. (+ indicates <0.1%.) Length group (cm) Prey 5-10 11-15 16-20 21-25 26-30 31-35 36-40 >41 Polychaeta — — — — 0.4 0 1 + + Crustacea 8.7 100 0 75 2 27.9 39.9 13.0 2 0 0.2 Amphipoda 0.3 — 0.3 1.8 1.3 0,8 + + Ampeliscidae — — — 07 0.5 02 + — Oedicerotidae — — — + 0.3 + + — Hyperudae 03 — 01 0.5 0.3 0.4 — — Other Amphipoda — — 0.2 0 6 0.2 0.2 + + Decapoda — 95.4 7 2 21.1 20.3 5 9 14 0.1 Crangonidae — 95.4 18 6 6 5.1 19 0.2 + Pandalidae — — 4.7 12.9 13.3 3.1 1.1 0.1 Pasiphaeidae — — — — — + — Other Decapoda — — 0 7 16 19 0 9 0.1 — Euphausiacea 7.5 4 0 66 8 3.3 13 5 5 2 0.6 0 1 Mysidacea 0.9 — — 0.3 3.8 05 + + Cumacea — — — 0.1 + + Copepoda — — — — — — Other Crustacea — 0 6 0.9 1.3 1.0 0.6 + + Cephalopoda — — — 28.4 61 18.7 15.1 5.9 Lohgo — — — 27 2 — 16.6 107 5.8 Other Cephalopoda — — — 1.2 6.1 2.1 4.4 0.1 Pisces 819 — 22.0 42.9 518 66.7 82 7 93.6 Scomberesox saurus — — — — — — 6.1 — Clupeidae — — — — — 5.4 3 8 2 8 Merluccius bilmeans — — — 31 9 5 0 6.6 20 8 — Phycis chesten — — — — — — — — Ammodytes amencanus 81.9 — — — 0 1 3 2 0.5 2 7 Scomber scombrus — — — — — 7.3 9.5 9.3 Stenotomus chrysops — — — — 16 — — 3.7 Pepnlus tnacanthus — — — — — — 3.6 5.0 Other Pisces — — 22,0 11.0 45 1 44,2 38.4 70.1 Miscellaneous 9.4 — 2.8 0.8 18 1.5 0.2 0.3 No. examined 9 3 22 113 202 259 126 103 No. empty 2 0 3 45 83 1 20 54 47 Mean stom. cont. wt. (g) 0 099 0 152 0 670 0.571 0.673 1.597 8.185 17 826 Mean fish FL (cm) f^0 12.0 18.5 23.4 28 0 32.9 37.7 46 0 28 BOWMAN: FOOD OF SILVER HAKE TABLE 7. — Geographic breakdown of the prey found in the stomachs of silver hake caught in the Northwest Atlantic during the years 1973-76. Data are expressed as a percentage weight. (+ indicates <0.1%). Middli 5 Southern Georges Prey Atlanti C New England Bank Polychaeta 0.1 0.1 0.1 Crustacea 73 7 3 16.4 Amphipoda 0.5 0.2 0.4 Ampehscidae 0.1 0.1 0.1 Oedicerotidae 02 + 0.1 Hypenidae 0.1 0.1 0.1 Other Amphipoda 0 1 + 0.1 Decapoda 49 26 6.5 Crangonidae 2.4 1 0 1.3 Pandalidae 1.8 1.2 4.4 Pasiphaeidae 0.4 — + Other Decapoda 0.3 04 0.8 Euphausiacea 1.2 3.4 7.9 Mysidacea 0.3 0.7 1.2 Cumacea — 0.1 + Copepoda + + + Other Crustacea 0.4 0.3 0.4 Cephalopoda 4.3 13.7 6.7 Loligo 2 9 . 13.0 6.7 Other Cephalopoda 14 0 7 + Pisces 87 5 78.4 76.4 Scomberesox saurus — — 6.1 Clupeidae 32 1.3 5 0 Merluccius btlineans : 208 7.9 0.4 Phycis Chester/ — — 0.8 Ammodytes amencanus 1.7 0.4 4.8 Scomber scombrus — 6.0 21.1 Stenotomus chrysops — 4.1 — Pepnlus triacanthus 1.4 2.2 89 Other Pisces i 30.4 565 29.3 Miscellaneous 0.8 0.5 0.4 No. of stomach examined 789 918 915 No. of empty stomachs 180 357 268 Mean stom. cont. wt. (g) 1.544 1.815 1.080 Mean fish FL (cm) 17.5 22.5 20.8 Length range (cm) 3-57 3-59 3-64 phausiids (3.4%) and decapods (2.6%) account for most of the Crustacea. The Cephalopoda was the only other prey group recognized as an important food of silver hake. Fish in Southern New England eat the largest quantities of squid (13.7%). Silver hake sampled on Georges Bank and in the Middle Atlantic also take fairly large amounts of squid as prey (6.7% and 4.3%, respectively). A comparison between the quantities of food in the stomachs of fish from each area revealed that Middle Atlantic silver hake have about two to three times more food in their stomachs (on the average) than fish from Southern New England or Georges Bank. Stomach content data for fish 24-35 cm FL from each area were adjusted for fish length; the adjusted mean stomach content weights were Middle Atlantic, 1.328 g; Southern New England, 0.593 g; and Georges Bank, 0.707 g. The quantity of food in the stomachs of Middle Atlantic silver hake is significantly different (with 95% confidence) from the quantity in the stomachs of fish from Southern New England (F = 6.862 exceeds F005 lj21 = 4.32). The adjusted mean stomach content weights of small (4-15 cm FL) silver hake from each area were Middle Atlantic, 0.149 g; Southern New England, 0.198 g; and Georges Bank, 0.214 g. Yearly and Seasonal Differences the majority of the identified fish prey. The stomachs of silver hake caught in Southern New England con- tain fairly high percentages of silver hake (7.97o), Atlantic mackerel (6.0%), and scup (4.1%). Silver hake caught on Georges Bank eat mostly Atlantic mackerel (21.1%), butterfish (8.97o), Atlantic saury (6.1%), herring (Clupeidae, 5.0%), and American sand lance (4.8%). Evidence of the cannibalistic na- ture of silver hake is seen in all three areas. In addi- tion, silver hake taken as prey comprise the highest percentage of identified fish in both the Middle Atlantic and Southern New England (Table 7). Crustaceans are most important in the diet of silver hake collected from Georges Bank (16.4%). Eu- phausiids (7.9%), decapods (mostly pandalid shrimp, 4.4%, and crangonid shrimp, 1.3%), and mysids (1.2%) account for the majority of crustacean prey consumed on Georges Bank. In the Middle Atlantic and Southern New England, Crustacea is of equal importance (7.3%) as a food. For Middle Atlan- tic fish, decapods (4.9%) and euphausiids (1.2%) make up the majority of crustacean prey identified in the stomachs. In Southern New England, eu- Percentages of various prey categories in the silver hake diet between years, seasons, and geographic areas indicate the stomach contents are quite vari- able (Table 8). For example, in the Middle Atlantic, the Crustacea portion of the diet of silver hake varies from 3.1% (spring 1973) to 70.0% (fall 1976). Similar variability can be seen in the percentages listed for most of the prey categories. Much of the observed variation is probably due to differences in predator lengths (note mean fish FL's given at the bottom of Table 8). Only one prey, the American sand lance, was noted as being unique in the diet of silver hake. American sand lance was only found in the stomachs of silver hake collected in the spring during 1975 and 1976. The largest percentage weights of American sand lance were derived from samples collected only during the spring of 1976 in all three areas. Another observation is that fish sampled in the spring tend to be larger (see mean lengths at bottom of Table 8) than those collected in the autumn. The adjusted stomach content data for large and small silver hake from all areas and years combined indicate that about twice as much food is found in the stomachs during spring than in autumn. The adjust- 29 FISHERY BULLETIN: VOL. 82, NO. 1 Table 8. — Annual and seasonal breakdown of the stomach contents for silver hake collected in the Middle Atlantic, Southern New England, and Georges Bank. Data are expressed as a percentage weight for fish collected during the spring (S) and autumn (F) of 1 973-76. (+ indicates present but <0.l7r.) 1973 1974 1975 1976 Prey S F S F S F S F MIDDLE ATLANTIC Polychaeta — — 0.1 — 05 — 16 — Crustacea 3.1 4,2 9.6 6.5 24.7 4.7 34.0 70.0 Amphipoda + 0.4 1.2 1.2 1.3 2 7 2.1 15.2 Ampeliscidae — 0.2 + 0.7 — 0.3 0.6 1.3 Oediceroiidae — + 1.1 — 04 — 0 9 — Hyperndae — 0.1 — 0.5 + 2.1 — 12.1 Other Amphipoda + 0.1 0.1 + 0.9 03 0.6 1.8 Decapoda 3 1 3.3 3.7 5.1 89 0.4 22 9 46.6 Crangonidae 1 4 0 5 2.0 44 59 03 11.1 258 Pandalidae 1.0 2.4 — 0 7 26 — 11.7 13.3 Pasiphaeidae 0.6 — — — — — — — Other Decapoda 0 1 0.4 1.7 + 04 0.1 0.1 75 Euphausiacea — 0.2 4.4 — 14.4 0 3 + — Mysidacea — + — — — — 5.4 — Cumacea — + + + 0.1 — + — Copepoda — — + + — + — — Other Crustacea + 03 0 3 0 2 + 1 3 3.6 82 Cephalopoda — 14.9 9 7 — 25.2 — 6.3 — Loligo — 12.4 — — 24.9 — — — Other Cephalopoda — 2.5 9.7 — 0.3 — 6.3 — Pisces 96 5 809 79.5 93.0 46.6 93.7 548 52 Scomberesox saurus — — — — — — — — Clupeidae — — — 91.5 — — — — Merluccius bihneans 233 490 — — 4.0 — — — Phycis chesten — — — — — — — — Ammodytes amencanus — — — — 10.7 — 19.8 — Scomber scombrus — — — — — — — — Stenotomus chrysops — — — — — — — — Pepnlus tnacanthus — — — — 24 4 — — — Other Pisces 73.2 31.9 79.5 l 5 7 5 93.7 35.0 5 2 Miscellaneous 0.4 + 1.1 0.5 3.0 1.6 3 3 24.8 No. examined 39 144 193 54 67 91 1 1 1 93 No empty 1 1 52 26 10 7 23 22 29 Mean stom. cont wl (g) 19 960 0982 0466 0 793 1 057 0.243 0606 0075 Mean fish FL {cm) 33.9 180 14.1 12 9 198 13 5 21.7 16.9 Length range (cm) 20-53 4-45 3-46 4-37 5 -44 3-40 8-57 3-35 SOUTHERN NEW ENGLAND Polychaeta 0.1 — + — + + 02 + Crustacea 2 8 12 5 3 3 46 .1 7 9 17.0 19.8 2.2 Amphipoda + 1.7 + 4.0 0 1 0.8 02 0.5 Ampeliscidae — 1 6 + 1.5 0.1 0.2 + + Oedicerotidae — — — — + — + + Hyperndae — 0 1 — i 9 + 0.5 0.1 0 5 Other Amphipoda + + + 0 6 + 0 1 0.1 + Decapoda 1 8 8 4 0 1 13 7 6 9 9.7 5.5 1.2 Crangonidae 02 0.9 + 4 5 2.0 04 4 7 0 2 Pandalidae 0.9 7 3 — 7.0 1.8 9 1 0 8 1.0 Pasiphaeidae — — — — — — — — Other Decapoda 0.7 0 2 0 1 2.2 3.1 0.2 — + Euphausiacea 0.5 0.9 3 2 23 5 0.8 49 99 + Mysidacea 0.3 + — — 0.1 0.9 3.8 — Cumacea + — + 1.7 + — + + Copepoda — + — + — + — — Other Crustacea 0.2 1.5 — 3 2 + 0.7 0.4 05 Cephalopoda 789 1 6 03 — 20.2 — — 2 8 Loligo 78.2 — — — 20.2 — — — Other Cephalopoda 0.7 1 6 03 — — — — 28 Pisces 18 2 859 95 6 45.2 70.1 829 79.8 94.5 Scomberesox saurus — — — — — — — — Clupeidae — — — — — 31.8 — — Merluccius bilineans 0.2 07 — 2.3 55 16 — 44.9 Phycis Chester/ — — — — — — — — Ammodytes amencanus — — — — 16 — 1.8 — Scomber scombrus — — 15.7 — — — — — Stenotomus chrysops — — — — — — — 24.7 Pepnlus tnacanthus 14.7 — — — — — — — Other Pisces 33 85.2 79.9 42.9 63.0 . 49.5 78.0 24.9 Miscellaneous + + 08 8 7 1.8 0.1 02 05 No. examined 105 119 93 117 100 1 20 125 140 No. empty 33 86 40 38 41 31 43 45 Mean stom. cont. wt. (g) 2406 0.401 6 902 0 107 0 952 0581 2.181 1 970 Mean fish FL (cm) 15.9 27.5 31 2 16.8 24 4 18 1 23.0 22.9 Length range (cm) 6-47 4-49 9-59 4-37 6-55 4-55 3-53 4-54 30 BOWMAN: FOOD OF SILVER HAKE Table 8. -Continued 1973 1974 1975 1976 Prey S F S F S F S F GEORGES BANK Polychaeta — — — — + + + — Crustacea 70.8 15.0 41.8 18.2 10.9 5.9 18 7 60 Amphipoda 1.4 04 0.3 13 02 04 09 0.1 Ampehscidae 0.1 03 + 07 — 0.2 — 0.1 Oedicerotidae — + + 0.3 0.1 + — Hypemdae — — — — — 0.1 0.8 + Other Amphipoda 1.3 0.1 03 03 0.1 0.1 0.1 + Decapoda 60.7 13.9 2 5 12 6 1.0 3.2 2.8 45 Crangomdae 1 9 2.0 1.3 2.1 0 5 1.6 0.6 2.1 Pandahdae 44 5 11.6 — 8 3 — 1 1 20 2.2 Pasiphaeidae — — 0.1 — — — — — Other Decapoda 14.3 0.3 1.1 2.2 05 0.5 02 0.2 Euphausiacea 2.3 0.2 31.2 26 94 0.5 14.8 + Mysidacea — 0.1 7 8 16 0.2 1.7 — 0.1 Cumacea + + — + — + — — Copepoda — + — + — + — — Other Crustacea 59 04 + 0 1 0.1 0.1 02 1.3 Cephalopoda — — — — — — 1 2.8 56.4 Loll go — — — — — — 12.8 56.2 Other Cephalopoda — — — — — — — 02 Pisces 23.7 84.9 57 9 81 8 88.1 94.1 68.5 35.8 Scomberesox saurus — — — 1 38.8 — — — — Clupeidae — — — — — 39.2 — — Merluccius bilmeans — — — 4.1 — — — — Phycis Chester/ — — — — 3.2 — — — Ammodytes amencanus — — — — — — 31.6 — Scomber scombrus — 31.0 — — S3 7 — — — Stenotomus chrysops — — — — — — — — Peprilus tnacanthus — 45.1 — — — — — — Other Pisces 23.7 88 57.9 89 21.2 549 36.9 35.8 Miscellaneous 5 5 0.1 0.3 + 1.0 + — 1.8 No. examined 48 198 103 157 92 146 63 115 No. empty 24 39 39 27 18 39 34 48 Mean stom. cont. wt. (g) 0340 1 029 0996 0.577 2 629 0906 2 478 0.767 Mean fish FL (cm) 31.4 16.6 24.2 16.0 24.5 18.1 29.7 22.3 Length range (cm) 27-42 4-54 8-49 4-40 1 1-54 4-48 10-64 3-55 ed mean stomach content weights are presented in Table 9 for each season, year, and geographic area. In almost every year, in all areas, the stomachs of similar- sized fish contain larger quantities of food in the spring than in the fall. Only two exceptions were noted to this trend (for which there is no ready explanation): Large fish collected on Georges Bank in 1973 and small fish collected on Georges Bank in 1974. Table 9. — Annual and seasonal brea ikdown of the adju sted mean i stomach content' weight data of large (24- 35 cm FL) and small (4- 15 cmFL) silver hake gathered from three geographical areas in the Northwest Atlan- tic during 1973-76. (S = spring, F = autumn.) 1973 ■ 974 1975 1976 Averages Area S F S F S F S F S F Middle Atlantic Large fish Adjusted weight (g) 5.545 1.081 0995 0.325 2.203 0.912 0.936 0.149 2.420 0.617 Number in sample 26 68 44 9 26 29 38 43 Small fish Ad|usted weight (g) — 0.108 0 180 0.096 0.148 0.142 0.207 0.155 0.178 0.131 Number in sample — 61 136 33 31 45 47 42 Southern New England Large fish Adjusted weight (g) 0.242 0.122 0488 0.303 0694 0.657 0.987 0.976 0 603 0.515 Number in sample 17 67 51 33 47 49 63 58 Small fish Adjusted weight (g) 0256 0.036 0 200 0.074 0.414 0 184 0205 0.149 0.269 0 111 Number in sample 73 15 4 49 35 62 39 58 Georges Bank Large fish Adjusted weight (g) 0400 0 743 0.916 0.576 1 239 0506 0.735 0.734 0823 0 640 Number in sample 43 58 50 53 32 57 27 51 Small fish Adjusted weight (g) — 0 140 0.321 0.325 0566 0.106 0473 0 117 0453 0 183 Number in sample — 119 36 95 16 80 9 50 Ave large fish adj. wt. 1.282 0.591 Ave. small fish adj. wt 0.300 0.142 31 FISHERY BULLETIN: VOL. 82, NO. 1 Maturity Stage Versus Diet Information on maturity was gathered in conjunc- tion with food data for 759 adult silver hake (Table 10). Gonads were classified as 1) resting - gonad small in size and relatively translucent, 2) developing - gonad enlarged and either cream (males) or yellow- orange (females) colored, 3) ripe - gonad fills most of gut cavity, reproductive material either runs freely from an incision in the gonad or is extruded with pres- sure on abdomen of fish, 4) spent - gonad is flaccid, hemorrhaging is often evident. depth range (0.1 g). The quantity of food found in stomachs of large fish is variable; it steadily de- creases between the 27-37 m and 74-110 m depth ranges; increases at the 111-146 m range; and from 1 1 1-146 m to 257-293 m continues to decrease (Ta- ble 12). Overall, the trend is for fish sampled at deeper depths to have less food, on the average, in their stomachs. It should be mentioned here that silver hake are known to regurgitate part or all of their stomach contents when they are retrieved from deep water depths (pers. obs.). Although fish which show obvious signs of regurtitation (e.g., everted stomach) TABLE 10. — Relationship between the adjusted stomach content weight and maturity stage of silver hake. Fish were caught on spring and autumn bottom trawl survey cruises conducted in the Northwest Atlantic from 1973 to 1976. Stomach content data Maturity stage Resting Ad], weight (g): 0 826 Developing 1.004 Ripe 0 122 Spent 1 292 No. of fish examined Mean fish FL (cm) Length range (cm) 379 286 24-35 297 30.6 24-35 29 31 3 27-34 54 31.2 25-35 No particular prey type is found in the stomachs of fish in specific maturity stages; all mature silver hake eat mostly fish. However, the stomachs of spawning (ripe) silver hake contain an average of about nine times less food than the stomachs of fish otherwise classified (Table 10). During pre- and postspawning periods, stomachs contain the largest quantities of food (1.0 and 1.3 g, respectively). Influence of Depth Analysis of samples from silver hake caught at dif- ferent bottom water depth ranges (27->365 m) revealed that the average length of fish, food type consumed, and quantity of food in the stomachs, varies with depth (Table 1 1). The majority (69.47c) of silver hake were caught at depths between 38 and 110 m. Considering only the depth ranges where more than 50 fish were sampled (i.e., 27-220 m, and representing 95.6% of all silver hake collected) the mean FL offish increases with an increase in depth. Also, the percentage weight of euphausiids and squid in the stomachs tends to increase at deeper bottom depths, while the percentage weight offish in the diet shows a corresponding decrease. The adjusted mean stomach content data for both small and large fish are given in Table 12. The data are from only those depth ranges from which more than 20 fish (within a size group) were collected. The adjusted stomach content weight of small silver hake steadily decreases from the 27-37 m depth range (0.3 g) to the 111-146 m are not sampled on survey cruises, some fish may regurgitate and not be discernable from those which did not, This phenomenon, in part (other factors such as the decrease in abundance of typical prey of silver hake with an increase in depth or decrease in bottom water temperature may also be important in this regard, see Williams and Wigley 1977) could explain the decrease noted in stomach content weights with an increase in water depth. DISCUSSION The diet of silver hake consists almost exclusively of a combination of fish, crustaceans, and squid. The relative importance of each particular prey group as a food of silver hake is, for the most part, dependent on the size of the predator and/or the availability of the prey (Bigelow and Schroeder 1953; Jensen and Fritz 1960; Fritz 1962; Dexter 1969; Vinogradov 1972). The composition of the diet of male and female silver hake is known to differ (Vinogradov 1972; Bow- man 1975). The present investigation confirms earlier reports that females feed predominantly on fish and that males eat mostly crustaceans. In addi- tion, it has been established that the stomachs of females contain larger quantities of food than the amounts in the stomachs of males of similar size. Since the rate of growth in fishes is directly related to their dietary intake, it is not surprising that females grow faster than males (Schaefer 1960). Bowman and Bowman (1980) studied diurnal varia- 32 BOWMAN: FOOD OF SILVER HAKE CD an CD a co a X eo a CO CO o CN C '3b c to a. 1 « - S 2g -* V eo C -C '" a, i, > — — <— .5f O a> C CD c o u re E o I I o | I III MM I I II I m co rid I M M || || M II M M II I II I d cd I I *- in d d | | do I + II I I I I II II II II II II -c Cm O 0) b£ c CO -a >> O -a CO a> 3D W -3 X < I r. + | O O | + + i- i- CN .- d d d d co co r- difi I - I I o I I I I I I CO CO ".' «) o -V <* CN CO ■" M II II I I i"*- co oo CO O | CN | | CN CN | oS iri I I o 1- CN O 6- I + d en co d cm I 6 M I I I I -' I I | I <* + d | co | CN O CM CO O IT) co »— d lo r^- 01 CO CN if) CO CN ** O CO CO CH CN CN «* CN «- CN LD m CN CO CN If) CO O — - O + + O oo in oS in <* CN co ■<* cm r- to r- CN CM ^ O CO m to cm "-; + CN — + «-; CM CO O CN O in r- .- d d II S I I I M 00 CO CO co d + + d d ** CD - — re "O O a. (o re fl) re c E S E < o < i o ° It - v OCJ0.0.O 3>-5 o £ ^ o ,r c i E * c B to S TO ^ "* a> re O ^ Qj <* Co 2.-B * 2 § 5! ' 5 ?' •~ co , CO 0. ' ID cn o rn oa CO — r-» CO E E o o 1*1 LU OJ o-S c c re re Z Z IV — - OD1- 55J id -a Ef — co o. a; c a8 S-c CO C IB 00 CO CD """ co X! •2 > c a ° CD ^. IB " > . <~ c -J CO s c c x CD CD p CO is E S o X! i " CC ■ — 6 5 TJ E CD CB CB T3 3 C CB on J M E cj - < re CN C '- § -2 3 CD .2 .£ CO *-* > -C e O —I o Li. E o i' -C E Sfl 3 4 Z ■C a^ 03 H -C re D ~ E 6 53 CO < 5 OO ^t if) . cn o co cn ^ tn m CM CO tO CO in oo co i- E «= I I I m 'j m co ^r co co cn cn cn co 20 cm FL collected during late summer-early autumn have small quantities of food (mean stomach content weight of 0.2 g) in their stomachs (Bowman and Bowman 1980). The stomach contents of silver hake collected on Georges Bank during the winter (December-January) of 1976-77 were analyzed by Bowman and Langton (1978). They found the mean stomach content weight offish 20 cm FL and larger to be 0.4 g. The stomachs of silver hake (all >29 cm FL) collected in February (late winter) of 1977 on Georges Bank, by American and Polish scientists aboard the Polish RV Wieczno (conducting research in conjunction with the Woods Hole Laboratory), contained an average of 0.1 g of food (unpublished data available from the author). The pattern of feeding intensity for silver hake throughout the year, based on the above information, is intensive feeding in the spring and early summer; curtailment of feeding in summer and early autumn (during spawning); resumption of feeding in the autumn, but to a lesser degree than in the spring; and finally a reduction in feeding throughout the winter. Somewhat similar feeding patterns have been es- tablished for other species of marine fish (Tyler 1971). Grosslein et al. (1980) reported an increase in bot- tom trawl survey catches of American sand lance in 1976 in the Northwest Atlantic. The population up- surge of American sand lance combined with the high percentage weights of American sand lance found in silver hake stomach contents during 1976 is an in- dication of silver hake's opportunistic predatory behavior. Availability of prey is probably one of the most important factors in determining what types and how much food silver hake eat. ACKNOWLEDGMENTS I thank M. Grosslein for his critical review of the manuscript; J. Towns, J. Murray, and others for their help in analyzing the fish stomach contents and in tabulating the data; and especially G. Kelley, laboratory typist, for her patience. LITERATURE CITED BlGELOW, H. B., AND W. C. SCHROEDER. 1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 53, 577 p. Bowman, R. E. 1975. Food habits of Atlantic cod, haddock, and silver hake in the Northwest Atlantic, 1969-1972. U.S. Natl. Mar. Fish. Serv., Northeast Fish. Cent., Woods Hole Lab. Ref. 75-01, 53 p. Bowman, R. E., and E. W. Bowman. 1980. Diurnal variation in the feeding intensity and catch- ability of silver hake (Merluccius bilinearis). Can. J. Fish. Aquat. Sci. 37:1565-1572. Bowman, R. E., and R. W. Langton. 1978. Fish predation on oil-contaminated prey from the re- gion of the ARGO MERCHANT oil spill. In In the wake of the ARGO MERCHANT, p. 137-141. Univ. R.I. Cent. Ocean Manage. Stud. 34 BOWMAN: FOOD OF SILVER HAKE Dexter, R. W. 1969. Studies on the food habits of whiting, redfish, and pollock in the Gulf of Maine. J. Mar. Biol. Assoc. India ll(l&2):288-294. Edwards, R. L., and R. E. Bowman. 1979. Food consumed by continental shelf fishes. In H. Clepper (editor). Predator-prey systems in fisheries man- agement, p. 387-406. Sport Fish Inst., Wash., D.C. Fritz, R. L. 1962. Silver hake. U.S. Fish Wildl. Serv., Fish. Leafl. 538, 7 p. GOSNER, K. L. 197 1. Guide to identification of marine and estuarine inver- tebrates. Cape Hatteras to the Bay of Fundy. Wiley, N. Y., 693 p. Grosslein, M. D., R. W. Langton, and M. P. SlSSENWINE. 1980. Recent fluctuations in pelagic fish stocks of the North- west Atlantic, Georges Bank region, in relation to species interactions. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 177:374-404. Jensen, A. C, and R. L. Fritz. 1960. Observations on the stomach contents of the silver hake. Trans. Am. Fish. Soc. 89:239-240. Langton, R. W., and R. E. Bowman. 1980. Food of fifteen northwest Atlantic gadiform fishes. U.S.Dep. Commer.,NOAATech. Rep. NMFS SSRF-740, 23 p. Nichols, J. T., and C. M. Breder, Jr. 1927. The marine fishes of New York and southern New Eng- land. Zoologica (N.Y.) 9:1-192. Noskov, A. S., and V. I. Vinogradov. 1977. Feeding and food chains of the fish of Georges Bank. Rbyn. Khoz. 53:19-20. (Can. Fish. Mar. Serv., Transl. Ser. 4540. D.O.E. St. John's, Nfld. 1979.) SCHAEFER, R. H. 1 960. Growth and feeding habits of the whiting or silver hake in the New York Bight. N.Y. Fish Game J. 7:85-98. Swan, B. K., and D. Clay. 1979. Feeding study on silver hake (Merluccius bilinearis) taken from the Scotian Shelf and ICNAF Subarea 5. ICNAF Res. Doc. 79/VI/49, 14 p. Tyler, A. V. 1971. Monthly changes in stomach contents of demersal fishes in Passamaquoddy Bay, N.B. Fish. Res. Board Can., Tech. Rep. 288, 103 p. Vinogradov, V. I. 1972. Studies of the food habits of silver and red hake in the Northwest Atlantic area, 1965-1967. ICNAF Res. Bull. 9:41-50. Wilk, S. J., W. W. Morse, and D. E. Ralph. 1978. Length-weight relationships of fishes collected in the New York Bight. Bull. N.J. Acad. Sci. 23(2):58-64. Williams, A. B„ and R. L. Wigley. 1977. Distribution of decapod Crustacea off northeastern United States based on specimens at the Northeast Fish- eries Center, Woods Hole, Massachusetts. U.S. Dep. Commer., NOAA Tech. Rep. NMFS Circ. 407, 44 p. 35 ABUNDANCE AND VERTICAL DISTRIBUTION OF FISHES IN A COBBLE-BOTTOM KELP FOREST OFF SAN ONOFRE, CALIFORNIA Ralph J. Larson1 and Edward E. DeMartini2 ABSTRACT Using visual belt transects on the bottom and vertically stratified belt transects taken with movie cameras in the water column, we assessed the species composition, vertical distribution, and standing stock of fishes in a forest of giant kelp and a nearby kelp-depauperate area off San Onofre, California. The volume of water- column "cinetransects" was calibrated for water clarity. Species such as garibaldi, blacksmith, and various rockfishes, which depend on high-relief rocky substrates, were rare or absent in these low-relief, cobble- bottom habitats. The species present in the kelp forest apparently did not depend on high-relief rock, at least in the presence of kelp. These species fell into three groups, based upon their vertical distributions: "canopy" species (kelp perch, giant kelpfish, and halfmoon), which occurred mainly in the upper water column; "cos- mopolites" (kelp bass, white seaperch, and senorita) .which occurred throughout the water column; and "bot- tom" species (California sheephead and various seaperches), which occurred mainly near the bottom. Despite the absence of reef-dependent species, estimated standing stocks of 388-653 kg/ha in the San Onofre kelp forest were as large or larger than estimates made by others in kelp forests located on higher relief bottoms. The kelp-forest areas at San Onofre also supported a larger standing stock of fishes (other than barred sand bass) than the adjacent area with little kelp. The relatively large standing stock of fishes in the kelp forest can be attributed to the presence of kelp and to the depth of the kelp forest. Located in relatively deep water (15m), this kelp forest possessed an extensive midwater zone. The attraction of fish in moderate densities to the midwater zone of this kelp forest contributed substantially to overall biomass. We conclude that kelp per se can enhance the standing stock of fishes on a temperate reef, at least in areas of low bottom relief. Rocky reef and giant kelp, Macrocystis pyrifera, habitats off the coast of southern California support a diverse and abundant assemblage of fishes (Lim- baugh 1955; Quast 1968 a, b; Feder et al. 1974; Ebel- ing et al. 1980 a, b). Much of the richness of this ichthyofauna has been attributed to the rocky sub- strate; areas with a rugose, rocky bottom and little kelp seem to support more fish than areas with a flat bottom and dense kelp (Quast 1968 a, b, Ebeling et al. 1980a). However, kelp itself also provides a uni- que habitat for some fishes (Coyer 1979; Ebeling et al. 1980a) and a point of orientation in the water column for others (Quast 1968 a, b; Bray 1981). The kelp canopy may also serve as a nursery area for some species of fish (Miller and Geibel 1973; Feder et al. 1974; M. Carr3 Unpubl. data). Several approaches have been used to assess the influence of habitat on the abundance and composi- 1 Marine Science Institute, University of California, Santa Barbara, Calif.; present address: Department of Biological Sciences, San Francisco State University, San Francisco, CA 94132. -Marine Science Institute, University of California, Santa Barbara, Calif.; present address: Marine Review Committee Research Cen- ter, 531 Encinitas Boulevard, Encinitas, CA 92024. 3M. Carr, Moss Landing Marine Laboratories, Moss Landing, CA 95039. tion of fish assemblages in nearshore kelp and rock habitats off California. Perhaps the best analytical approach is experimental, as employed by Miller and Geibel (1973), Bray (1981), and Carr (footnote 3) ; however, the comparative approach of Limbaugh (1955; also reported in Feder et al. 1974), Quast (1968 a, b), and Ebeling et al. (1980a) is also of value. Based on observations in a variety of areas, Lim- baugh described the habits and habitats of many nearshore fishes. Quast and Ebeling et al. employed broad-scale quantitative sampling of fish assem- blages in different areas. Quast's interpretation of data extended Limbaugh's natural history approach, and added to it the actual comparison of abundances in different habitats. Ebeling etal. (1980a) employed a multivariate analysis of habitat characteristics and relative abundances of species to define subassem- blages of fishes, and also compared abundances in areas of different habitat characteristics. In this paper we examine the abundance, vertical distribution, and species composition of noncryptic fishes in a forest of giant kelp near San Onofre, Calif. We also report the abundance and species composi- tion of fishes in a nearby area with little kelp. This study, undertaken initially to predict the effects of a Manuscript accepted June 1983. FISHERY BULLETIN: VOL. 82, NO. 1, 1984. 37 FISHERY BULLETIN: VOL. 82. NO. 1 possible loss of kelp (Dean4) on the indigenous fish fauna, also allowed us to extend the comparative approach of Quast and E beling to assess two features of kelp-forest fish faunas and to further evaluate a sampling technique. The portion of the kelp forest we examined was located in relatively deep water (15 m) and was anchored on a low-relief cobble bottom. Since it lacked a highly heterogeneous substrate, we were able, by comparison, to further evaluate the effects of kelp per se on nearshore fishes. Because the kelp forest was in deep water, we also had the opportunity to examine the vertical distribution of fishes in greater detail than other workers, by sampling four vertical strata, rather than the two strata (canopy and bottom) sampled by Quast (1968b) and E beling etal. (1980a, b). Besides visual transects to sample fish on or near the bottom, we used underwater movies ("cinetran- sects") to estimate the abundance of fishes in the water column above the bottom. Alevizon and Brooks (1975) and Ebeling et al. (1980b) discussed the advantages and disadvantages of cinetransects, but provided only rough estimates of the area sampled in a cinetransect. In this paper we more carefully evaluate cinetransect volume, emphasizing the effect of underwater visibility on cinetransect width. Our objectives in this paper are 1) to estimate cine- transect volume as a function of underwater visibility; 2) to examine the vertical distribution of fishes in a deep-water kelp forest; 3) to estimate the overall abundance and biomass of fishes, integrated over depth, in this kelp forest; and 4) to evaluate the importance of kelp to nearshore fishes, by comparing our data from the San Onofre kelp forest with that from an adjacent kelp-depauperate area and with other published data from kelp forests located on more rugose substrates. MATERIALS AND METHODS Study Areas This study was conducted in and near the offshore portion of a giant kelp, Macrocystis pyrifera, forest near the San Onofre Nuclear Generating Station, between San Clemente and Oceanside, Calif. (Fig. 1). 4T. A. Dean. 1980. The effects of San Onofre Nuclear Generating Station on the giant kelp, Macrocystis pyrifera. Annual report of the Kelp Ecology Project, January-December 1979, to the Marine Re- view Committee of the California Coastal Commission. Unpubl. rep., 189 p. Kelp Ecology Project, Marine Science Institute, University of California, Santa Barbara, CA 93106. San Onofre kelp (SOK) varied in areal extent from <5 to 95 ha during the mid- to late 1970's, and covered about 75 ha during the fall of 1979 (Dean footnote 4). SOK occupied a shallowly sloping, low- relief (< 1 m) cobble and sand substrate between the depths of about 10 and 15 m. Two relatively perma- nent, offshore portions of SOK, and an area with little kelp located <100 m upcoast from SOK, served as our study areas. The upcoast (SOK-U) and downcoast (SOK-D) areas within SOK, and the kelp- depauperate area ("kelpless" cobble), were all about 15 m deep and 2-3 km from shore. Because of its depth, low relief, and periodic inundation by sand, the cobble substrate in all areas was relatively bare of understory algae and sessile invertebrates. However, some stands of the 1 m tall laminarian kelp Pterygophora californica were present, especially along the fringes of the Macrocystis forest and throughout the kelpless cobble area. Sampling Methods Our general sampling plan was to stratify fish cen- suses by depth and to replicate these samples over several dates. In the two kelp-forest areas, we cen- sused each of three, equally spaced strata in the water column, plus a bottom stratum. Only the bot- tom stratum was censused at the kelp-depauperate area, since few kelp-associated fishes were observed above the bottom in this area. Sampling at each stratum was replicated hierarchically: A number of replicate transects were made within an area on a given sampling day, and counts from these transects were averaged. This was repeated on 4 or 5 d at each site. The daily averages at each stratum and area were themselves used as replicates that provided reasonably precise estimates of means per stratum and that allowed estimates of variability due to sam- pling error. Because of time and manpower con- straints, the various study areas were usually sampled on different dates. All three water-column strata in a given area were sampled on the same day; the bottom stratum, however, was usually sampled on a different day. All sampling took place from October through December 1979. This time of year offers the most consistently clear and calm water conditions. Since most migratory and transient species were excluded from analysis (see below), our fall study should reasonably characterize the general distribution and abundance of "resident", kelp-associated fishes at SOK. Within this period, sampling was generally limited to dates when horizontal visibility exceeded 3 m. 38 LARSON and DeMARTINI: DISTRIBUTION OF FISHES IN KELP FOREST 120' 119' 118' 117 < Santa D . Barbara $an Santa Mi^el Is. ^gp |s h34° Santa Rosa Is. Los Angeles San Santa Catalina Is. Nicholas Is. San Clemente JOceanside 1-33° San V> Clemente Is. 9.1 nMk 18.1 m x5.5 FIGURE 1. — Location of areas sampled during fall 1979, near San Onofre, Calif.: Upcoast (U) and downcoast (D) portions of the San Onofre kelp bed, and nearby kelp-depauperate area (kelpless cobble). San Mateo Creek San Onofre Creek San Kelp Bed kelpless cobble San Onofre ? Kelp Bed d 0 12 3 4 5 6 kilometers 39 FISHERY BULLETIN: VOL. 82, NO. 1 In each area, two permanently buoyed stations served as foci for sampling. At each station, we deter- mined a range of suitable compass headings for tran- sects. To assure complete coverage of the area, we divided each range of suitable headings into five equal subarcs and randomly chose transect headings from each subarc. Headings were selected separately for each sampling stratum. One transect per subarc was made on each sampling day for bottom sampling. In the water-column strata, where fish patchiness necessitated more samples, we made one transect in each subarc and added another transect from one of the subarcs (randomly chosen). Thus, five transects were usually made from each station per date on the bottom, and six at each station and depth stratum in the water column. Regardless of sampling method, transects began 7-10 m from the station hub. Tran- sects were taken from both sampling stations on a sampling day. Data from the two stations at an area were pooled, since the abundances of major species were generally indistinguishable between stations in an area on a given date. On the bottom, fish sampling was conducted visually in 75 m long strip transects. Divers (one per station) counted fish in bands estimated to be 3 m wide and 1.5 m high, while reeling out 75 m long lines along the predetermined compass headings. All non- cryptic fishes within this band were identified and counted, with separate tallies kept for juvenile, sub- adult, and adult members of each species (Table 1). All subadult and adult Mac rocystis plants >1 m tall (Dean footnote 4) were counted in the same 3 m wide band while reeling in the transect line on the return trip. Transects in the water column at the two kelp-forest areas were made with underwater movie strips, using Elmo Super 311 Low Light5 movie cameras (F/l.l), Giddings Cine-Mar housings, and Kodak Ekta- chrome 164 super-8 film cartridges. At 18 frames/s, the transects lasted about 3 min. Divers swam pre- determined compass headings and photographed fish occurring in a 120° horizontal arc about the tran- sect axis and 1.5 m above and below the diver's depth. The transect ended when the film cartridge was exhausted. Water-column transects were made in three depth strata: 3 m, 7.6 m, and 12 m (Table 2). Horizontal visibility was measured with each set of transects (at a depth on a sampling date), as the dis- tance at which an olive-tan colored, 10 cm long float ("fish mimic") became indistinct. Films were later viewed in slow motion by at least two observers, at 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Table 1.— Common and scientific names of fishes observed at the San Onofre kelp bed and adjacent kelpless cobble area during fall 1979 with the estimated weight of juveniles, subadults, and adults. Body weights for teleosts were estimated from average observed lengths, converted to weights using the length-weight regressions of Quast (1968a: Appendix B), after adjusting for the bias (underestimate) from the use of average body length to predict average body weight (see Pienaar and Ricker 1 968). Weights of elasmobranchs were estimated from fishes trapped in the intakes of the San Onofre Nuclear Generating Station, Unit 1, during 1976-79. 1 Asterisks indicate species not included among kelp-bed "residents." Common names after Robins et al. (1980). Weight (g) Weight (g) Family and species Juvenile Subadult Adult Family and Species Juvenile Subadult Adult Serranidae Scorpaenidae Paralabrax clathratus, kelp bass 7 200 1,050 Scorpaena guttata, California scorpionfish — — 550 Paralabrax nebulifer. barred sand bass 20 300 1,500 Sebastes rastrelliger, grass rockfish2 — — 400 Embiotocidae Sebastes serranoides, olive rockfish2 4 175 — Brachyistius frenatus, kelp perch — — 25 Sebastes spp,, juvenile rockfish2 1 — — Embiotoca /acksom . black perch 10 75 350 Sciaenidae Phanerodon furcatus , white seaperch 10 50 175 'Cheilotrema saturnum. black croaker — — 225 Damalichthys vacca . pile perch 15 175 500 Pristopomatidae Rbacochilus toxotes. rubberlip seaperch 15 150 700 'Xentstius californiensis. salema — — 75 Hypsurus caryi. rainbow seaperch 10 60 150 Athennidae Labndae *silversides spp. — — 20 Oxy/ulis califomica . sehonta 0.5 5 55 Carangidae Semicossyphus pulcher. California sheephead 50 250 875 *Trachurus symmetricus. jack mackerel — 115 — Halichoeres semicmctus. rock wrasse 25 100 250 Sphyraenidae Girellidae 'Sphyraena argentea. Pacific barracuda — 150 — Giretla nigricans, opaleye — — 950 Carcharhinidae Scorpididae 'Tnakis semifasciata. leopard shark — — 2,000 Medialuna califormensis, halfmoon — — 250 Rhinobatidae Pomacentridae 'Ptatyrhmotdes tnsenata . thornback — — 240 Chromis punctipinms. blacksmith 2 — — Myliobatidae Hypsypops rubicundus. garibaldi 25 120 500 *Myliobatis califomica. bat ray — — 6,700 Clinidae Torpedmidae Heterostichus rostratus, giant kelpfish 3 30 175 'Torpedo califomica. Pacific electric ray — — 9,450 Cottidae Scorpaenichthys marmoratus, cabezon — — 1,500 'E. DeMartini and R Larson. 1980. Predicted effects of the operations of San Onofre Nucler Generating Station Units 1, 2, and 3 on the fish fauna of the San Onofre region. Report submitted to the Marine Review Committee of the California Coastal Commission. Unpubl. rep., 27 p. Marine Science Institute, University of California, Santa Barbara, CA 93106. 2Members of the genus Sebastes will be grouped under rockfish spp. in subsequent tables. 40 LARSON and DeMARTINI: DISTRIBUTION OF FISHES IN KELP FOREST TABLE 2.— Bathymetric sampling strata at the San Onofre kelp bed. Weighting factors (WJ are shown for the above-bottom strata and for the above-bottom versus bottom strata. Sampling Depth Range Extent of W^ (above- Wh (all depth (m) represented (m) range (m| bottom only) strata) 3 0-5.3 5.3 03926 7.6 5.3-9.8 4.5 03333 '0.9 12 9.8-13.5 3.7 0.2741 15 (bottoml 13.5-15.0 1.5 — 0.1 0-15 1.0 1.0 'Weighting factor for above-bottom strata combined. which time fish that were distinguishable on film were identified, counted, and assigned to maturity classes as above. Transect Volume The volume of visual bottom transects was con- sidered to be fixed, and the volume of water-column cinetransects to be dependent on underwater visibility. The volume of bottom transects was fixed at75mX 3mX 1.5 m = 337.5 m\ since the length of transects was measured, and the height and width of transects were fixed at values less than horizontal visibility. Cinetransect length was taken as the average distance covered in simulated, 3-min cine- transects swum by three divers over a metered line. Each diver swam two simulations against the current, and two with the current. The cross-sectional area of a cinetransect was treated as an ellipse with a minor (vertical) axis of 1.5 m, the distance above and below the diver that fish were photographed. The major axis of the ellipse was a function of camera range, the distance at which fish could be distinguished on film. The particular function was cos 30° X camera range, since divers photographed fish within a 120° arc (60° on each side of the transect axis) (Fig. 2). Thus, the volume of cinetransects at a given depth on a given day was calculated as V= 1.5 ttL (cos 30° X CR), where V was cinetransect volume in cubic meters; 1.5, the minor axis of the ellipse; L, the cinetransect length as determined above; and CR, the camera range at that depth on that day. Camera range itself was estimated as a function of the horizontal visibility at a depth on a sampling date. The relationship between camera range and horizontal visibility was estimated empirically under different conditions. The main "other condition" that we evaluated was the orientation of the camera to the sun. In trials run at different visibilities, two fish of similar appearance (usually a kelp perch, Brachyistius frenatus, and a white seaperch, Phanerodon furcatus) were held on a spear by one diver and photographed with our usual equipment by another diver at distances decremented from the limits of horizontal visibility (measured as described above). At each visibility, trials were run with the camera facing into the sun and with the camera facing away from the sun. Two observers viewed the film from each trial and determined camera range as the greatest distance at which the two fish could be dis- tinguished on film. The criteria for distinguishability were the same as those used in evaluating whether or not to count a fish when we viewed regular cinetransects. Data for camera range versus horizontal visibility were fit to several asymptotic functions. The fitting CINETRANSECT VOLUME A. CINETRANSECT SHAPE B. CINETRANSECT CROSS SECTION Camera Range FIGURE 2.— A. Estimated shape of area sampled in under- water transects taken with motion pictures (cinetransects). The length of 76 m was estimated from simulated tran- sects. B. Elliptical cross section of a cinetransect, with minor axis (a) of 1.5 m and major axis (b) calculated from camera range when divers surveyed a 120° horizontal arc about the central axis of the transect. FISHERY BULLETIN: VOL. 82, NO. 1 routine was BMDP program P3R, nonlinear regres- sion (Dixon and Brown 1979). The function with the smallest residual mean square was selected to repre- sent the relation between camera range and horizon- tal visibility, and was employed in estimating camera range at a depth on a sampling date. Data Analysis We reduced data into two general forms: densities (number or biomass per unit volume) in different strata, and abundances integrated throughout the entire water column. The first was used to examine the vertical distribution of individual species or of the entire assemblage and to compare the relative abun- dances of species in a stratum. The second was used to estimate the overall abundance of the assemblage and to compare the overall abundances of different species. In both cases, the final point and interval estimates were based on the means and variances, over dates, of daily means. The daily estimate of density (per 1 ,000 m3) for each species in a depth stratum was estimated as the mean number or biomass per transect on that day, times the ratio (1,000/transect volume), where transect volume was estimated as above. Biomass of a species on a given transect was estimated by counts of individuals in different maturity classes, converted to wet weights by the key in Table 1. Our estimate of a species' density in a depth stratum was calculated as the mean of the daily den- sity estimates in that stratum. Similar estimates were made for the sum of all "resident" teleosts. Excluded from the analysis of total fish density and abundance were elasmobranchs and certain teleosts (silver- sides, jack mackerel, Pacific barracuda, black croaker, and salema) that were rare at SOK, are seasonal visitors to kelp beds, or are not primarily associated with rock reefs and kelp forests (Feder et al. 1974). Species such as white seaperch and barred sand bass often occur in other habitats, but were included in our analysis because they may have at least a marginal association with kelp-rock habitats and were frequently encountered and abundant in our samples. By weighting the average density of a species (or the assemblage) in a stratum by the volume of water rep- resented by samples in that stratum, we were able to obtain estimates of abundance integrated from sur- face to bottom (Snedecor and Cochran 1980:444). The sampling day was an integral component of our analysis, but only the above-bottom strata were sampled on the same day at a given site. To obtain accurate estimates of variance for integrated abun- dances, then, we assembled our integrated estimates in two stages. We first estimated stratified mean den- sity for the above-bottom strata on each day and averaged these values over days. We also computed mean density (over days) in the bottom stratum. Secondly, we computed stratified mean density (and its standard error) for the above-bottom and bottom strata, using the means and variances calculated above. The stratified mean density estimates for the entire water column were then scaled to represent abundances over 100 m2 of bottom. Samples in each stratum were assumed to represent a range of depths extending to the midpoints be- tween strata, with the 3 m stratum also extending to the surface (Table 2). Weighting factors for the strata were determined from the relative extents of the depth ranges represented. Among the above-bottom strata, relative weighting factors were the vertical ranges of these strata divided by 13. 5 m. For the bot- tom versus above-bottom strata the depth ranges were divided by 15 m. Daily estimates of stratified mean density in the above-bottom strata were calculated as Dm = 2 WhDh, where Dm was the estimate of stratified mean density in the 3 m, 7.6 m, and 12 m strata; W,„ the weighting factor; and Dh, the mean density on that day in stratum h (Snedecor and Cochran 1980). The mean (Dwc) and variance (S2UJ of these daily estimates were then computed. The mean {Db) and variance (S26) of estimated daily densities on the bottom were also calculated. Stratified mean abundance throughout the entire water column was estimated as A .-( 1,500 1,000 XWhD n /!> where Ast was the stratified mean estimate of integrated abundance over 100 m2 of bottom, Wh was the weighting factor, and Dh was the mean density in either the above-bottom strata {Dwc) or in the bottom stratum (Dh). The term in the summation is the estimate of stratified mean density (per 1,000 m3) over all strata, and the ratio (1,500/1,000) converts this value to abundance over 100 m2 of bottom. The standard error of Ast was calculated as ■4" v v 1,000 ' h h '' h where S2h was the variance of daily density estimates in either the above-bottom (S2,,,.) or bottom (S2b) 42 LARSON and DeMARTINI: DISTRIBUTION OF FISHES IN KELP FOREST strata; Wh, the weighting factor; and nh, the number of days sampled in stratum h. The portion of the for- mula included in the summation is the usual estimate of variance for stratified means (Snedecor and Cochran 1980), and the root of this sum is the stan- dard error of mean density (per 1,000 m3) throughout the water column. Multiplying by (1,500/1,000)2 adjusts the standard error for the larger volume of water in the column over 100 m2. Estimates of integrated abundance at the kelp- depauperate site were obtained by converting mean density on the bottom to mean density over 100 m . Arithmetic means (of untransformed data) were used for all estimates of density and abundance. Geometric means (obtained by back-transforming the means of log-transformed data) underestimate absolute densities in a manner proportional to their variances. Adjustments for this underestimation (Elliott 1971) are usually based on the assumption of log-normal distributions, and we could not make such an assumption. However, some statistical com- parisons were made with log-transformed data to avoid the problem of heterogeneous variances. These were comparisons of mean numbers and biomass on the bottom, where varying transect volume did not confound the calculation of variance. Other comparisons, however, were made with untransformed data. These included tests for dif- ferences in numbers or biomass in the above-bottom strata and in the entire water column. When all three areas were compared, a one-way ANO VA was used if variances were not heterogeneous. T'-tests for un- equal variances (Bailey 1959) were used for pairwise comparisons of areas when variances were un- equal. RESULTS Cinetransect Calibration We estimated cinetransect length to be about 76 m. Six down-current trials averaged 78.3 m in length (standard error (SE) = 1.5 m, range = 74-82 m), 6 upcurrent trials averaged 72.8 m in length (SE = 2.3 m, range = 67-82 m), and the overall average was 75.6 m (SE = 1.5 m). Camera range was an asymptotic function of horizontal visibility, with little increase in camera range at visibilities beyond 7-9 m (Fig. 3). Camera range was appreciably lower when the camera was facing the sun than vice versa, particularly at greater visibilities. This was reflected in each of the curves fit (Table 3). Since divers did not record whether actual transects faced into or away from the sun, we used the curve fit to all camera range-horizontal visibility values to calibrate cinetransect volume. The logistic equation provided, by slight margin, the best fit to O 3 < 2 . < 1 0 284 + 1 893(0.582 * <8> * -*e e- • INTO SUN : AWAY FROM SUN 0 1 2 3 4 5 6 7 8 9 10 11 12 13 HORIZONTAL VISIBILITY (m) Figure 3.— Relation of camera range (the distance at which fish could be distinguished on film) and horizontal visibility. Points are observations of maximum camera range at different visi- bilities with the camera facing into and away from the sun. The equation and line show the logistic function fit to these points. 43 FISHERY BULLETIN: VOL. 82, NO. 1 these data (Table 3) and was the one employed in calculating cinetransect volume. Distribution and Abundance of Fishes Five sets of bottom transects were made in each study area. Water-column samples were taken on five dates at SOK-U and on four at SOK-D. Transect Table 3. — Functions fit to camera range (Y) versus horizontal visi- bility (X) relationship, and the best fit parameters as determined by BMDP program P3R (Dixon and Brown 1979). Also noted are the asymptotes calculated for each equation and data set, and the resid- ual mean squares. Into= trials made with the camera facing into the sun; Away = trials made with the camera facing away from the sun; All = curves fit to all data. Pj, P2, and P3 are arbitrary symbols for the parameters of each function; there is no implied correspondence between the numbered parameters of different functions. Asymp- Residual Function name Set of tote mean and formula trials P, ?7l P3 (m) square Logistic All 0 284 1.89 0.582 3.52 0369 Y Away 0.259 2.63 0 560 3.86 0 250 Y= 1/(P, + P3) Into 0 317 1.20 0618 3.15 0355 Gompertz All 1.27 -3 19 0647 ' 3.56 0370 Y=e|P. + P.pJ Away 1 37 -3.88 0 648 3 94 0255 Into 1 15 -2.35 0.656 3.16 0.354 Von Benalanffy All 360 0334 1.43 3.60 0.372 . p , Away 4.03 0.301 1.62 4.03 0.261 Y= P, (1 - e 2 3 ) Into 3 17 0361 1.07 3 17 0353 Michaelis-Menton All 4.21 1.92 203 421 0.377 P, I* - PJ Away 4 94 1.91 2.79 4.94 0269 Y- ' 2 Into 3.51 2.01 1 28 3 51 0354 P +X- P 3 2 AM 0.194 1.06 — 5.15 0388 Beverton-Holt Away 0.158 1.17 — 633 0284 Into 0241 092 — 4.15 0352 Y= 1/|P, + P2/X) number and visibility at depth on each date are shown in Table 4. Of the 28 species recorded in this study, 19 were "resident" teleosts. Of these, 13 species were record- ed on more than two transects in the two kelp-forest areas (Table 5). These 13 common species could be assigned to bathymetric categories, based on their vertical patterns of frequency of occurrence (Table 5) and density (Tables 6, 7) within SOK. Kelp perch, halfmoon, and giant kelpfish were most common in the upper strata and are designated "canopy" species. While halfmoon and giant kelpfish were observed in all strata, all three species were most abundant in the 3 m stratum. Only halfmoon reached moderate abundances at 7.6 m in the SOK-D area (Tables 5, 6, 7). Sehorita, white seaperch, and kelp bass were com- mon throughout the water column (Tables 5, 6, 7) and are designated "cosmopolites". These three species were among the most common and abundant fishes in all strata. The white seaperch was the most cos- mopolitan of the three in 1979, its density and fre- quency of occurrence on transects varying little with depth. The sehorita was the most abundant species in nearly all strata. The kelp bass was also abundant at all depths. Its numerical density varied little among the water-column strata, but was generally greater on the bottom. Its biomass was greater in the lower strata (Tables 6, 7). Young kelp bass concen- trated in the upper water column (Table 8), con- tributing to the relatively low biomass per fish for kelp bass in the 3 and 7.6 m strata. Our data indicate Table 4. — Sampling dates, number of transects, and visibilities measured during fall 1979 sampling in two areas within the kelp bed at San Onofre (SOK-U and SOK-D) and in a nearby cobble-bottom area with little kelp (Cobble). Horizontal visibility (vis.) measured in meters SOK-U SOK-D Cobble 3 m 7.6 m 12 m Bottom 3 m 7.6 m 12 m Bottom Bottom Date V VIS. V vis. V vis V VIS. V VIS V VIS. V VIS. V vis. V VIS. 10 Oct 10 2 95 15 Oct. 9 2.14 17 Oct. 9 2.89 7 300 22 Oct 10 2.75 9 3.42 24 Oct. 10 2.60 26 Oct 1 1 1400 12 8.50 1 1 3.50 31 Oct. 10 3.85 10 5.00 7 Nov. 10 3.90 12 Nov. 12 7.30 12 5 10 12 4.75 14 Nov 10 5.50 16 Nov. 10 4.50 10 4.85 21 Nov. 10 8.75 26 Nov. 12 10.25 12 700 12 4.00 28 Nov. 10 4.00 30 Nov 12 12.55 12 7.05 12 3 15 5 Dec. 12 16.00 12 13.75 12 7.25 7 Dec 12 10.50 12 585 12 5.10 10 Dec. 12 8.25 12 7.80 12 6.90 12 Dec. 12 9.45 12 6.95 12 850 19 Dec. 13 10.50 12 8 50 12 5.25 Total 61 60 60 49 47 48 47 48 47 Mean 10 06 6.69 5 35 4.24 12.13 926 541 3.59 4.19 44 LARSON and DeMARTINI: DISTRIBUTION OF FISHES IN KELP FOREST that the upper kelp canopy serves as a nursery for young-of-the-year kelp bass, and these cryptic fish were probably much more abundant there than shown by our counts. We examined vertical segrega- tion of size classes only for kelp bass. This is because our 1979 data were too few to evaluate vertical segregation by size that has since been noted for two other species (senorita and blacksmith) in several Table 5.— Percent of transects on which species were observed during fall 1979, in two portions of a kelp forest near San Onofre, Calif. (SOK-U and SOK-D) and in a nearby kelpless cobble area (Cobble). Species' ranks are shown in parentheses. Number of transects is noted in the column heading. SOK-U SOK ■D Cobble 3 m 7 6 m 12 m Bottom 3 m 7.6 m 12 m Bottom Bottom Species n=61 n=60 n=60 n=49 n=47 n=48 n=47 r?=48 n=47 kelp bass 52(3) 50(3) 60(1.5) 61(2.5) 74(2) 77(2) 81(1) 81(2) 26(4.5) barred sand bass 2(10) 8(4.5) 59(4) 9(8) 58(5) 53(1) kelp perch 59(2) 13(4) 2(13) 49(3) 10(6.5) 9(8) black perch 8(4.5) 41(5) 9(8) 65(3) 26(4.5) white seaperch 41(4) 58(2) 60(1.5) 39(6.5) 40(4) 56(3) 62(3) 44(7.5) 15(8) pile perch 2(10) 3(10) 20(9) 4(9) 17(5) 42(9) 11(9) rubberlip seaperch 3(10) 16(10) 19(10) 4(12.5) rainbow seaperch 39(6.5) 44(7.5) 19(6.5) senorita 93(1) 87(1) 58(3) 61(2.5) 96(1) 94(1) 66(2) 63(4) 43(2) California sheephead 5(7.5) 58(1) 2(10.5) 19(5) 36(4) 90(1) 36(3) rock wrasse 29(8) 2(9.5) 4(10) 46(6) 6(105) opaleye 3(8) 2(15) halfmoon 16(6.5) 7(6) 2(13) 2(14) 36(5) 38(4) 11(6) 4(12.5) 19(6.5) blacksmith 2(10) 2(11.5) garibaldi 4(12.5) giant kelpfish 24(5) 8(5) 7(6) 3(11) 21(6) 10(6.5) 4(12.5) cabezon 2(14) 2(16.5) California scorpionfish 2(16.5) 4(12.5) rockfish spp. 3(10) 2(14) 6(10.5) black croaker 2(16.5) salema 4(12.5) silversides 16(6.5) 19(7) jack mackerel 2(9.5) 3(7.5) 5(7.5) 17(8) 8(8) 2(11.5) Pacific barracuda 2(10.5) 2(9.5) leopard shark 2(14) thornback 2(15) bat ray 2(14) 2(16.5) Pacific electric ray 2(9.5) 3(7.5) 2(13) 2(15) TABLE 6. — Mean numerical and biomass densities (per 1,000 m3) of fishes observed in n daily samples per depth stratum at the SOK-U area in the San Onofre kelp bed during fall 1979. Values are the grand means (± 1 standard error) of the daily means (adjusted for transect volume) over transects taken each sampling day. SOK-U Numerical dens ity (no/1. 000 m3) Biomass densi ty (kg/1.000 m3) 3 m 7.6 m 12 m Bottom 3 m 7 6 m 12 m Bottom l«=5) C=5) (n=5) (n=5) (n=5) (n=5) 0=5) (r>=5) Species x SE x SE x SE x SE x SE x SE x SE x SE kelp bass 1.57 0.87 2 67 1 19 2 48 093 4.76 1.20 0.091 0071 0416 0.171 0664 0270 1 372 0 372 barred sand bass 0 0.02 0.02 0.13 0.04 3.30 0.70 0 0024 0.024 0.173 0 046 4.434 0930 kelp perch 1 39 0.26 0.23 0 13 0.02 0.02 0 0.035 0.007 0006 0.003 neg. 0 black perch 0 0 0 12 0.07 2.25 065 0 0 0046 0.028 0.717 0 209 white seaperch 1.91 1.21 3.16 1.20 2.33 0.86 3.07 0 59 0.319 0210 0491 0.209 0 287 0.105 0.376 0108 pile perch 0 002 0.02 0.08 0.05 0.66 0.11 0 0 009 0.009 0039 0.025 0263 0.079 rubberlip seaperch 0 0 0.04 0.03 1.08 0 35 0 0 0 028 0.017 0634 0.265 rainbow seaperch 0 0 0 2.02 0.92 0 0 0 0.167 0 068 senorita 2695 6.53 2445 5.78 4.66 2 22 14.16 5.95 0950 0.223 1.103 0225 0.241 0 110 0566 0.237 California sheephead 0 0 013 0.06 4.87 1 16 0 0 0058 0.040 1.561 0338 rock wrasse 0 0 0 1.20 1.24 0 0 0 0.237 0022 opaleye 0.03 0.03 0 0 0 0033 0.033 0 0 0 halfmoon 0.27 0.20 0.08 0.05 0.02 0 02 006 006 0068 0050 0 020 0 012 0006 0.006 0.015 0.015 blacksmith 0 002 0.02 0 0 0 neg. 0 0 garibaldi 0 0 0 0 0 0 0 0 giant kelpfish 0.35 0.08 009 0.04 0.08 0.04 0.18 0.12 0 018 0.007 0004 0003 0.015 0 008 0.014 0.012 cabezon 0 0 0 0.06 0.06 0 0 0 0089 0089 Calif, scorpionfish 0 0 0 0 0 0 0 0 rockfish spp. 0 0 0.04 0.02 0.06 0.06 0 0 0.003 0.003 0.024 0.024 black croaker 0 0 0 0 0 0 0 0 salema 0 0 0 0 0 0 0 0 silversides 4.21 1.54 0 0 0 0092 0.029 0 0 0 jack mackerel 0.09 0.90 8 77 8.74 0.50 0.36 0 0010 0.010 1 008 1.005 0057 0.041 0 Pacific barracuda 0 0 0 0 0 0 0 0 leopard shark 0 0 0 0.06 0.06 0 0 0 0 119 0.119 thornback 0 0 0 0 0 0 0 0 bat ray 0 0 0 006 0.06 0 0 0 0.397 0.397 Pacific electric ray 001 0.01 003 002 0.02 0.02 0 0136 0136 0 320 0 196 0.154 0.154 0 45 FISHERY BULLETIN: VOL. 82, NO. 1 TABLE 7.— Mean numerical and biomass densities (per 1,000 m3) of fishes observed inn daily samples per depth stratum at the SOK-D area in the San Onofre kelp bed during fall 1 97 9. Values are the grand means (± 1 standard error) of the daily means (adjusted for transect volume) over transects taken each sampling day. SOK-D Numerical densi ty (no/1, 000 m3) Biomass density (kg/1 .000 m3) 3 m 7.6 m 12 m Bottom 3 m 7.6 m 12 m Bottom (n=4) (1=4) (n = =4) (n=5) (n=4) (n=4) (1=4) (1=5) Species x SE x SE * SE x SE . SE x SE x SE x SE kelp bass 4 23 0.63 4 61 1.09 4.84 1.07 12.87 395 0 726 0 162 1.101 0.440 1621 0.672 2 363 0675 barred sand bass kelp perch 0 0.83 0.20 0 0.19 0.09 0.12 0.11 0.02 0.08 3 14 029 0 0 0.021 0 005 0 0005 0002 0.178 0.029 0003 0.002 3446 0.577 0 black perch white seaperch pile perch 0 3.50 2.38 0.04 0.04 0 415 1 52 0 0 18 4.83 0.23 0.11 0 94 0.13 4 77 0.63 3.64 148 1.74 0.18 0 0.582 0 407 0 013 0.013 0 0681 0269 0 0 040 0.017 0 693 0 180 0.105 0.068 1.401 0 089 0 399 0.137 0.682 0.056 njbberlip seaperch rainbow seaperch sehonta California sheephead rock wrasse 0 0 19 46 2.82 0.02 0 02 0 0 0 2104 3 57 0.60 0.23 0.02 0.02 568 1.52 0.06 0 0 1.82 0.42 0.03 064 0.24 249 0.71 13.31 7.77 13.66 1.29 1.86 0.49 0 0 0569 0.078 0.017 0.017 0 0 0 1.039 0.158 0.181 0.119 0005 0.005 0 0 0.312 0.100 0.770 0.386 0028 0004 0.447 0.165 0.238 0.053 0.435 0.205 4990 0.322 0.405 0.110 opaleye halfmoon 0 1.09 0.44 0 2.92 1.83 0.35 0 0.19 0 0.12 0.12 0 0.237 0 1 10 0 0730 0.457 0 0087 0047 0 0.030 0.030 blacksmith 0 0 0.03 0.04 0 0 0 neg. 0 garibaldi giant kelpfish cabezon 0 0.28 0.06 0 0 0 10 0.02 0 0 0 0 0.12 0.07 012 0.07 007 0.06 0 0024 0007 0 0 0008 0004 0 0 0 0 0.014 0.009 0.012 0.010 0099 0099 Calif, scorpionfish rockfish spp. black croaker 0 0 0 <) 0 0 0 0 0 006 006 0 11.85 11.85 0 0 0 0 0 0 0 0 0 0.033 0033 0 2.667 2.667 salema silversides jack mackerel Pacific barracuda 0 5 99 3.96 2096 9.05 0.13 0.13 0 0 19.34 17 69 0.61 0.61 3.32 0 0 3.32 0 889 5.93 0 0 0 0 0.120 0.079 2.410 1.040 0.019 0.019 0 0 2.224 2.035 0.092 0092 0 0 0.381 0.381 0 0.667 0.444 0 0 0 leopard shark thornback bat ray Pacific electric ray 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.12 0.12 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.794 0.794 0 TABLE 8. — Mean numerical densities (per 1,000 m3) of young-of- the-year (yoy), all juveniles (including yoy), subadult, and adult kelp bass inn daily samples per depth stratum at SOK-U and SOK-D during fall 1979. Grand means calculated as in Tables 6 and 7. Numerical dens ity (no./1,000 m3) 3 m (i n = 5) 7,6 m (n = 5) 12 m |i = 5) Bottom (n = 5) SOK-U X SE X SE X SE < SE yoy all juvs. subadults adults 065 1 23 0.32 002 0.20 062 0.25 0.01 036 085 1.76 0.05 0.12 0.34 0.94 0 04 0.10 0.90 1.18 0.40 0.05 0.38 0.52 0.24 024 1 36 2.59 0.80 0.24 0.76 069 026 3 m (, n = 4| 7,6 m |i = 4) 12 m (1 = 4) Bottom (1 = 5) SOK-D X SE X SE X SE X SE yoy all juvs. subadults adults 0.88 1.50 2 52 0.20 0.42 0 62 0.98 0.12 033 1.24 2.88 0.49 0 13 0.19 0.87 028 020 1.37 2 39 1.08 0.09 0.50 0.84 0.49 0.12 5.21 6 72 0.94 0.12 3.26 3.05 0 18 kelp beds off northern San Diego County (DeMartini et al.6). Seven of the 13 common species were most abun- dant near the bottom (Tables 5, 6, 7). Rainbow seaperch and rock wrasse rarely, if ever, strayed above the bottom. Black perch and rubberlip seaperch were recorded occasionally at 12 m, but 6E. DeMartini, F. Koehrn, D. Roberts, R. Fountain, and K. Plum- mer. Variations in the abundances of fishes within and between stands of giant kelp {Macrocystis pyrifera) during successive years. Manuscr. in prep. Marine Science Institute, University of Califor- nia, Santa Barbara, CA 93106. were much more abundant on the bottom. Pile perch were seen, at one site or the other, in all strata, but were most abundant on the bottom and at 12 m. Barred sand bass also concentrated on the bottom and, to a lesser degree, at 1 2 m. California sheephead were observed as shallow as 3 m at SOK-D, but no shallower than 12 m at SOK-U. Species composition and relative abundance in each stratum reflected the distributional patterns of the species (Tables 9, 10). The three cosmopolitan species were among the three to five most abundant species in every stratum, particularly above the bot- tom. At 3 and 7.6 m, they made up 89-99% of total numerical density. The remaining fish in these strata were mainly upper water-column species, with a few of the more errant bottom species (such as California sheephead and pile perch) entering at 7.6 m. The three cosmopolites again dominated the assemblage at 12 m, forming 86-94% of fish numbers. A few individuals of canopy species were present at 12 m, however, and a greater number of bottom species were observed. The bottom stratum contained the greatest number of recorded species, and individuals were distributed more evenly among these species. The cosmopolites were still among the most abun- dant species on the bottom, but several of the bottom-zone species (such as California sheephead, black perch, and barred sand bass) were also abun- 46 LARSON and DeMARTINI: DISTRIBUTION OF FISHES IN KELP FOREST Table 9.— Percent contribution of species to total numerical and biomass density at the SOK-U area of the San Onofre kelp bed during fall 1979. Percentages are given by stratum and for abundance integrated throughout the water column. Only those species contributing 1% or more are listed. Stratum values are based on data in Tables 6 and 7; integrated abundances on Table 1 1. 3 m 7.6 m 12 m Bottom Integrated Species % Species % Species % Species % Species % SOK-U Numbers senorita 83.0 senonta 79.5 senorita 46 0 senorita 37.5 senorita 72.0 white seaperch 59 white seaperch 10.3 kelp bass 24 5 Calif, sheephead 12.9 white seaperch 9.3 kelp bass 4 8 kelp bass 8 7 white seaperch 230 kelp bass 12.6 kelp bass 9.1 kelp perch 43 barred sand bass 1.3 barred sand bass 8 7 kelp perch 2.1 giant kelpfish 1.1 Calif, sheephead 1.3 white seaperch 8 1 Calif, sheephead 2.0 black perch 1.2 black perch rainbow seaperch rock wrasse rubberlip seaperch pile perch 6.0 54 3.2 2.9 1.7 barred sand bass 1.4 SOK-U Biomass senonta 62.7 senorita 53.2 kelp bass 42.6 barred sand bass 42.4 senonta 30.2 white seaperch 21.1 white seaperch 23 7 white seaperch 18 4 Calif, sheephead 149 barred sand bass 19.1 kelp bass 6.0 kelp bass 20 1 senonta 15.4 kelp bass 13.1 kelp bass 1 7.7 halfmoon 45 barred sand bass 1.2 barred sand bass 1 1.1 black perch 6.8 white seaperch 14.2 kelp perch 2.3 halfmoon 1.0 Calif, sheephead 3.7 rubberlip seaperch 6.1 Calif, sheephead 6.6 opal 2.2 black perch 2.9 senorita 5.4 black perch 3.2 giant kelpfish 1.2 pile perch 2.5 white seaperch 36 rubberlip seaperch 2 7 rubberlip seaperch 18 pile perch 2 5 pile perch 1.5 giant kelpfish 1.0 rock wrasse rainbow seaperch 2.3 1.6 halfmoon 1.2 Table 10. — Percent contribution of species to total numerical and biomass density at the SOK-D area of the San Onofre kelp bed during fall 1979. Percentages are given by stratum and for abundance integrated throughout the water column. Only those species contributing 1% or more are listed. Stratum values are based on data in Tables 6 and 7; integrated abundances on Table 1 1. 3 m 7 6m 12 m Bottom Integrated Species % Species % Species % Species 'V. Species % SOK-D Numbers senorita 66.1 senorita 62.6 senorita 31.6 Calif, sheephead 23.3 senorita 51 7 kelp bass 14 4 kelp bass 13.7 kelp bass 27.0 senorita 22 7 kelp bass 174 white seaperch 119 white seaperch 12 3 white seaperch 27.0 kelp bass 22.0 white seaperch 13.1 halfmoon 3.7 halfmoon 8 7 Calif sheephead 8 5 black perch 8.1 Calif, sheephead 63 kelp perch 2 8 Calif, sheephead 1 8 halfmoon 1.9 white seaperch 6.2 halfmoon 4,4 pile perch 1.3 barred sand bass 5.4 black perch 1.7 black perch 1 0 rainbow seaperch rock wrasse pile perch rubberlip seaperch 4.2 3.2 3.0 1.1 kelp perch barred sand bass 1.2 1.1 SOK-D Biomass kelp bass 32.6 kelp bass 29.4 kelp bass 42 2 Calif, sheephead 33 3 kelp bass 28.2 white seaperch 26.2 senorita 27.7 Calif, sheephead 20.1 barred sand bass 23.0 Calif, sheephead 17.2 senonta 256 halfmoon 19.5 white seaperch 1 8.1 kelp bass 15 8 senonta 14.5 halfmoon 12 3 white seaperch 18 2 senonta 8.1 black perch 9.3 white seaperch 14.3 giant kelpfish 1.1 Calif, sheephead 4.8 barred sand bass 4.6 pile perch 4,5 barred sand bass 89 pile perch 2 7 rubberlip seaperch 30 halfmoon 7 8 halfmoon 2.3 senorita 2 9 black perch 3.4 black perch 1.0 rock wrasse white seaperch rainbow seaperch 2.7 2.7 1.6 pile perch rock wrasse rubberlip seaperch 2.3 1 1 1.0 dant. The gradual change in species composition that occurred between the water-column strata became more abrupt at the bottom. The vertical profile of total numerical density reflected changes in the abundance of the most numerous species, senorita, and the increase in species number on the bottom. Numerical density was about the same at 3 and 7.6 m, dropped at 12 m, and peaked on the bottom (Fig. 4). Small differences in species composition at 3 and 7.6 m led to only small differences in the abundances of noncosmopolites, and the cosmopolites (particularly senorita) had similar densities in these strata (Tables 6, 7). Despite increased abundances of bottom species at 12 m, the loss of upper water-column species and the decline in abundance of senorita led to low overall numerical densities in this stratum (Tables 6, 7). Senorita became more abundant again in the bottom stratum, kelp bass reached peak density, and the bottom species became abundant (Tables 6, 7), leading to high numerical densities on the bottom (Fig. 4). Biomass density did not differ among the water- column strata, but reached an exaggerated peak on the bottom (Fig. 5). At 12 m, the increase in size of kelp bass, and the addition of large-bodied species like California sheephead, barred sand bass, and 47 FISHERY BULLETIN: VOL. 82, NO. 1 X (- Q. u - t 5- \ 1 \l \l l\ 1 \ • SOK U SOK D I . // / / / / / / / / io- / / / / / / / / / / ^ ■ J 1 1 1 1 — i r — i 1 10 20 30 40 50 60 70 80 MEAN NUMERICAL DENSITY (Number/lOOOm3) ','« Q. LU Q 0- 1 \ 1 \ 1 \ • SOK U SOK D \ \ I \ 1 \ \ \ 111 J 1 1 1 1 1 1 1 "***-.«. — ^~_ IS- 5 10 MEAN BIOMASS DENSITY (kg/1000m3) FIGURE 4. — Vertical distribution of the numerical densities of all resident teleosts in two areas within the San Onofre kelp bed during fall 1979. Points are mean densities over sampling dates at each site and stratum, and bars represent one standard error of the mean. FIGURE 5. — Vertical distributions of the biomass density of all resi- dent teleosts in two areas within the San Onofre kelp forest during fall 1979. Points are mean densities over sampling dates at each site and stratum, and bars represent one standard error of the mean. various embiotocids compensated for the decline in abundance of senorita (Tables 6, 7). The higher numerical densities of these large fishes on the bot- tom contributed most to the peak biomass densities in this stratum. Weighting densities for the size of stratum, we estimated that on average about 40 and 46 fish occurred over 100 m2 at SOK-U and SOK-D, respec- tively, with corresponding biomass values of 3.9 and 6.5 kg/100 m2 (Table 11). About 66% (SOK-D) to 77% (SOK-U) of all individuals occurred in the upper two strata, 9% (SOK-U) to 14% (SOK-D) at 12 m, and 14% (SOK-U) to 19% (SOK-D) on the bottom. The small vertical extent of the bottom stratum diminished its contribution to the abundance offish integrated over the entire water column. About 44- 45% offish biomass occurred in the two upper strata, 15% (SOK-U) to 22% (SOK-D) occurred at 12 m, and 34% (SOK-D) to 40% (SOK-U) on the bottom. Thus much of biomass was near the bottom, but because of Table 1 1. — Abundance of resident teleosts, based on densities integrated through the water column over 100 m2 of bottom. The standing stock in numbers and biomass is given for each of two areas (SOK-U and SOK-D) within the San Onofre kelp bed, and for an adjacent area of cobble bottom with little kelp (Cobble), for samples taken in fall 1979. Numbers 1... , U)l) m2 B omass (kg) per 100 m* SOK-U SOK-D C Dbble SOK-U SOK-D Cobble Species X SE X SE X SE X SE X SE X SE kelp bass 3.66 1.02 8.04 0 80 0.25 0 14 0.67 0 15 1.83 041 0 12 004 barred sand bass 0.55 0.1 1 0.52 0.04 116 0.37 0.74 0 14 0.58 0.09 1 69 0.55 kelp perch 085 0.20 0.57 0 05 0 0.02 0.01 0.01 0.01 0 black perch 038 0.10 0.78 0 10 0 54 032 0.13 003 023 0.02 0.19 0 1 1 white seaperch 3 76 1.43 6.05 1 54 046 0 31 0.55 0.24 0 93 030 007 0.06 pile perch 0 14 0.03 0.37 007 005 002 0.06 0.02 0.15 003 002 001 rubberlip seaperch 0 18 0.05 0.10 0.04 0 03 002 0.11 0.04 0.07 0 02 0.02 0.01 rainbow seaperch 030 0 14 037 0 1 1 001 0 12 0.03 0.01 0 04 001 001 001 senorita 28 86 4.63 2388 2.06 2 16 0 77 1.17 0.21 0.95 0 05 0.06 004 Calif, sheephead 0 78 0.18 2.89 0.30 0.61 0.20 0.26 005 1.12 0 17 0 18 006 rock wrasse 0 18 0.04 0.31 0.08 0.03 0.01 004 0.01 0.07 0.02 0.01 0.01 opaleye 0.02 0.02 0 001 0.01 0.02 0.02 0 0.01 001 halfmoon 0.20 0 13 2 04 1 03 0 11 005 0.05 003 051 026 0.03 0.01 blacksmith 001 0.01 001 0.04 0 neg neq 0 garibaldi 0 0.02 0.01 0 0 neq 0 giant kelpfish 028 007 021 0.04 0 0.02 001 0.02 001 0 cabezon 0.02 001 0.01 0.01 0 0.01 0.01 0.02 0.02 0 Calif, scorpionfish 0 001 001 0.02 001 0 neq 0.01 001 rockfish spp 0.02 0.01 0 0.04 0.03 0.01 001 0 0.01 0.01 All residents 40.4 6.0 46.2 4.1 5.6 0.94 3 '1 0.5 65 0.7 2.4 0.6 48 LARSON and DeMARTINI: DISTRIBUTION OF FISHES IN KELP FOREST their more extensive bathymetric ranges, the low biomass- density upper strata still contributed nearly one-half of total biomass. The most abundant species at SOK were the cos- mopolites (Tables 9, 10, 1 1). Sehorita, kelp bass, and white seaperch comprised 82 and 90% of all individuals in the kelp forests at SOK-D and SOK-U, respectively. These species also contributed strongly to overall integrated biomass, although large species like California sheephead, barred sand bass, and halfmoon were also important. As a result, the dis- tribution of biomass among species was more even than the distribution of numbers (Tables 9, 10, 11). Two relatively large fishes were more abundant at SOK-D than SOK-U during fall of 1 979, contributing to the differences (see below) in our estimates of total biomass at each site (Table 1 1). The integrated abun- dance of kelp bass was significantly higher, or nearly so, at SOK-D (Numbers: t = 3.37, df = 7, 0.0K P < 0.02; Biomass: t = 2.65, df = 4, 0.05 < P < 0.1). California sheephead were also more abundant at SOK-D, as tested with log-transformed bottom data (Numbers: t = 4.81, df = 6, P < 0.01; Biomass: t = 3.35, df = 5, 0.02 < P < 0.05) and with integrated abundances (Numbers: t = 6.03, df = 5, P < 0.01; Biomass: t = 4.92, df = 4, P < 0.01). Halfmoon seemed to be more abundant at SOK-D, but the dif- ference was not significant (Numbers: t = 1.78, df = 3, P > 0.1; Biomass: t = 1.78, df = 3, P > 0.1). At the kelpless cobble site, most fish were bottom species and cosmopolites (Tables 5, 11). While barred sand bass, black perch, and California sheephead were fairly abundant in this area, the average abundances of other species were less than in the kelp-bed areas. The integrated numerical abundance of all fishes was significantly lower in the kelpless cobble area (cobble vs. SOK-U: t = 5.71, df = 4,P< 0.01; cobble vs. SOK-D: t = 9.42, df = 3,P < 0.01; SOK-U vs. SOK-D: f = 0.79, df= 7, P> 0.4). A one-way ANOVA of log-transformed counts on the bottom showed significant differences among the three areas (F2 12 = 9.42, P < 0.01), but an a priori comparison of SOK-U and SOK-D versus the cobble area was not significant (Fx 12 = 1.207, P > 0.25). Thus, the lower overall numerical abundance at the kelpless cobble area was due largely to the presence offish above the bottom at SOK. The integrated total biomass of fish did not differ significantly among the three areas (F2>11 = 0.25, P > 0.75), even though the point estimate of 2.4 kg/100 m2 at the cobble area was lower than both values at SOK. However, barred sand bass made up over 70% of fish biomass in the cobble area, so most other species were much less abundant there. We estimated the density of Macrocystis plants >1 m tall to be 7.51 ± 0.71 (1 SE) plants/ 100 m2 at the "kelpless" cobble area, 23.11 ± 1.47 plants/100 m2 at SOK-U, and 30.18 ± 1.69 plants/100 m2at SOK-D. Thus, some kelp was present at the cobble area, but the density of subadult-adult plants there was 25- 32% of density in our kelp-bed areas. DISCUSSION Sampling Regardless of water clarity, our camera and film were unable to resolve fish beyond 3-4 m; this set an upper limit of just over 1,000 m3 to cinetransect volume. Alevizon and Brooks (1975) noted that in very clear, shallow waters, fish seemed difficult to distinguish on film beyond 5 m. Ebeling et al. ( 1 980b) found camera range to be 3-3.5 m at horizontal visibilities of 4 and 15 m, and concluded that there was essentially no relation between camera range and horizontal visibility. Our data show this to be true at visibilities >7-9 m. The fixed focal length of the camera, shallow depth of field at maximum aperture, and quality of film account for the limited camera range, as discussed by Ebeling et al. (1980b). However, our data show that camera range decreases when visibility decreases to values that approach maximum camera range. Corrections for visibility are common in terrestrial line transects, whether the area of a given transect is taken as fixed throughout or as variable (Caughley 1977; Burnham et al. 1980). We regarded the volume of a given cinetransect to be fixed, its width determined by visibility. The relatively low upper limit to camera range may help to make cinetransects in the water column more accurate than visual censuses. Searching efficiency would likely be poorer for broad visual transects made to the limits of visibility. Furthermore, it is dif- ficult to judge arbitrary smaller distances in open water, unless they are only a meter or two on either side of the diver. Cinetransects provide an almost automatic upper limit to transect width, and this limit is wide enough (about 3 m to either side in mod- erately clear water) that a substantial volume of water is censused. We have not verified the exact volume sampled in each of our cinetransects, nor are we able to compare densities measured in cinetransects with actual den- sities (Brock 1982), since the latter have not been measured by any method. To our knowledge, only Keast and Harker (1977) have actually marked the outside boundaries of visual underwater transects. However, Terry and Stephens (1976) and Stephens 49 FISHERY BULLETIN: VOL. 82, NO. 1 and Zerba (1981) utilized two divers, swimming parallel, unmarked courses and counting fish be- tween each other, to sample rocky-reef fishes. Perhaps such a method could be used to evaluate densities estimated in cinetransects. Species Composition, Distribution, and Abundance The species observed in the San Onofre kelp forest were a subset of the species found in other nearshore areas of hard substrate and vegetation off southern California. Many reef-dependent fishes that are very common in other kelp forests were either rare or unrecorded at San Onofre. Species such as black- smith and opaleye (Ebeling and Bray 1976; Hobson and Chess 1976), garibaldi (Clarke 1970), painted greenling (DeMartini and Anderson 1979), and some species of Sebastes (Larson 1980) depend on rugose reefs for shelter or spawning sites. Some turf-grazing and otherwise bottom-feeding species of embi- otocids also appeared to be less abundant at San Onofre than in other areas. Our estimates of 14-37 kg/ha of pile perch, 38-78 kg/ha of black perch, and 10-18 kg/ha of rubberlip seaperch were mostly smaller than the estimates of Ebeling et al. (1980b) off Santa Barbara and Santa Cruz Island. The rarity and low abundance of all these species markedly alters the character of the fish assemblage at San Onofre. The abundant species at San Onofre kelp forest either are less dependent on rock reefs (at least, if kelp is present) or associate preferentially with low- relief substrates. The former group might include the canopy species, the cosmopolitan kelp bass and sehorita, and perhaps the epibenthic California sheephead. The latter group might include barred sand bass and white seaperch. These two species (and perhaps sehorita) were more common at San Onofre than others (Ebeling et al. 1980a, b) have reported in kelp forest anchored on high-relief sub- strates. Barred sand bass occurred in over half of the bottom transects at SOK, but in no more than 12% of bottom transects near Santa Barbara (Ebeling et al. 1980a). We found white seaperch in 40-60% of our transects, while Ebeling etal. (1980a) saw them on 7- 42% of all transects (but 20-42% of "sandy margin" transects). Both of these species have been reported as associating with sand or the sand-rock interface (Quast 1968a; Feder et al. 1974; Ebeling et al. 1980a). Moreover, barred sand bass have a warmwater affinity (Frey 1971) and on average should be more abundant farther south in the Southern California Bight. The abundance of white seaperch at SOK may be unusually high during the fall. At this time, white seaperch appear to use the SOK habitat for mating as well as feeding. While some individuals of white seaperch are found in kelp forests all year, much of their populations in kelp beds off northern San Diego County move offshore after fall (authors' observations). The vertical distributions of species present at the San Onofre kelp bed were similar to patterns de- scribed in other kelp forests. Kelp perch, giant kelpfish, and, to a lesser extent, halfmoon have been recognized as water-column and canopy species (Quast 1968a; Feder et al. 1974; Bray and Ebeling 1975; Ebeling and Bray 1976; Hobson and Chess 1976; Coyer 1979; Ebeling et al. 1980a, b). Kelp bass and white seaperch have been described as members of a vertical "commuter" group of fishes in kelp forests near Santa Barbara (Ebeling et al. 1980a). The term "cosmopolite" better describes the habits of these two fishes. Sehorita also fell into Ebeling et al.'s "canopy" group, but its occurrence throughout the water column was recognized by Hobson (1971), Ebeling and Bray (1975), Bernstein and Jung (1979), and others. We feel that it too should be considered a cosmopolite. Pile perch and rubberlip seaperch were also assigned to the commuter group of Ebeling et al. (1980a) and did appear above the bottom at San Onofre. However, the dense midwater aggregations of these species observed elsewhere were not present at San Onofre. Perhaps the relatively low density of these species at San Onofre was responsible for the absence of these aggregations. On the other hand, our fairly frequent observation of California sheephead well above the bottom is apparently new. Quast (1968a), in fact, noted that sheephead seem "reluctant" to leave the bottom. Barred sand bass, black perch, rainbow seaperch, and rock wrasse occurred almost exclusively on the bottom, and have been generally recognized as bottom dwellers. Our estimates of vertically integrated standing stock were surprisingly high. Most estimates of fish biomass on tropical and temperate reefs fall into the range of a few to several hundred kg/ha (Brock 1954; Bardach 1959; Randall 1963; Quast 1968b; Talbot and Goldman 1972; Miller and Geibel 1973; Jones and Chase 1975; Russell 1977). It is encouraging that our estimates of 3.88-6.53 kg/100 m2 (388-653 kg/ ha) fell within this range. Furthermore, our density estimates for fall 1979 are generally similar to subse- quent estimates made for canopy and bottom strata during the fall periods of 1980 and 1981 (E. DeMar- tini7 Unpubl. data). In particular, the densities of resi- 7E. DeMartini, Marine Science Institute, University of California, Santa Barbara, CA 93106. 50 LARSON and DeMARTINI: DISTRIBUTION OF FISHES IN KELP FOREST dent species (kelp bass and California sheephead) that contributed most to biomass estimates for fall 1979 were not consistently larger or smaller, if dif- ferent at all, at SOK during fall 1 980 and 1 98 1 . Hence we feel that our estimates for fall 1979 are typical for SOK during this season. Furthermore, while species such as kelp bass and sheephead were most abun- dant at SOK-D during fall 1979, this was not always true in 1980 and 1981; the site of greater abundance switched between SOK-U and SOK-D for many species over the period of 1979-81 (DeMartini et al. footnote 6). Thus we also conclude that apparent dif- ferences between SOK-U and SOK-D during fall 1979, although perhaps statistically real, are not meaningful for our general characterization of stand- ing stock at SOK. For this reason, we have provided data for the areas separately as brackets for our estimates of conditions in the San Onofre kelp bed in general, and do not specifically attribute the greater abundance of fishes at SOK-D to greater numerical density of giant kelp plants > 1 m tall. The surprising aspect of our standing-stock estimates is that they are as large or larger than those of Quast (1968b) in nearshore areas of greater bot- tom relief. Subtracting elasmobranchs, "nonresi- dent" teleosts, and cryptic bottom species, his estimates of standing stock at two sites near San Diego were about 366 kg/ha for Del Mar and 299 kg/ ha for Bathtub Rock. Thus, even though our areas at San Onofre lacked many individuals of such great contributors to biomass at Quast's sites as opaleye, blacksmith, kelp rockfish, and garibaldi, our brack- eted values of biomass were of the same order to nearly twice Quast's estimates. Below, we examine three possible reasons for this perceived disparity: Bias due to sampling methods, bias due to the times and places sampled, and the possibility that there really was a relatively large standing stock of fishes at San Onofre. Our sampling methods may have led to over- estimates, or Quast's (1968b) to underestimates, of standing stock. Quast's quantitative collection at Del Mar lacked a wall net, so some fish may have escaped. Although he used transect densities for three of the abundant species in his corrected estimates, his tran- sect method of counting fish to the limits of visibility may have led to reduced searching efficiency (as dis- cussed above). It is less likely that we counted fish in a larger volume than we think. We may have inflated our estimates of integrated abundance by sampling the bottom stratum on different days than the water- column strata, so that the same individuals could have figured into average density in more than one stratum as distributions changed from day to day. Such errors would have been most serious in the cos- mopolitan species, and perhaps in large bottom species (like California sheephead) that also occurred in the water column. However, even in our 3 m stratum, the average numbers of senorita and white seaperch per transect (uncorrected for visibility) were greater than similar averages obtained by Ebelingetal. (1980a, b) in cinetransects off Santa Barbara, implying that these species really were abundant during the fall at San Onofre. For kelp bass, the average standing stock above the bottom was 48 ± 13 (SE) kg/ha at SOK-U and 148+40 at SOK-D. These values are large fractions of our total respective estimates of about 69 and 183 kg/ha. Similarly, our estimates of sheephead biomass on the bottom alone were 23 ± 5 kg/ha at SOK-U and 75 ± 5 kg/ha at SOK-D, compared with our total estimates of about 26 and 112 kg/ha at the respective areas. We conclude that, while sampling problems may have contributed some bias to both our estimates and those of Quast's, much of the difference between Quast's estimates and ours is real, and fish really were relatively more abundant in the areas we sam- pled at SOK during the fall. Our selection of sampling times and places could have led to estimates that are somewhat unrep- resentative of conditions in general at San Onofre. Seasonal factors might be involved for some of our "resident" species. Dense concentrations of some fishes (notably white seaperch) may be atypically high at SOK and perhaps other kelp beds during the fall, when these areas are used for breeding. Many species of fish can be found in kelp beds all year, but their abundances might nevertheless fluctuate greatly as individuals move among areas within kelp beds, between different kelp beds, and perhaps be- tween different nearshore habitats. We feel that our samples accurately characterize the standing stock of fishes at San Onofre kelp in the fall, but cannot extend our observations to other seasons. Horizontal patchiness in the distribution of fish may also have affected our estimates. Our kelp-forest sampling areas were near the offshore edge of a large area of surface canopy, and fish often were quite dense at the actual edge of the kelp forest. Limbaugh (1955), Quast (1968a), Feder et al. (1974), Hobson and Chess (1976), Bray (1981), and others have dis- cussed this "edge effect". Although many of our tran- sects did not (by chance) sample the edge of the bed, the averages we calculated nonetheless may have overestimated the density of some species through- out the entire bed. However, our estimates of fish density at the particular study areas should be relatively unbiased. Quast's (1968b) Del Mar collec- 51 FISHERY BULLETIN: VOL. 82. NO. 1 tion was also made at the edge of a kelp forest, so comparison with our areas is warranted. The comparatively large standing stock of fishes at SOK in part reflects the nature of the kelp forest off San Onofre. This kelp forest was located in relatively deep (15 m) water, and was of moderate (0.1 adult plant/m2; Dean footnote 4) kelp density, with a sur- face canopy. Both of Quast's (1968b) sites were located in relatively shallow (7.6-10.7 m) water. Furthermore, Quast's Bathtub Rock site lacked a surface kelp canopy. A substantial part of the fish biomass we observed at San Onofre was in the exten- sive canopy and midwater zones. Nearly half of the biomass occurred in the upper two strata at each site, and about one-quarter occurred in the midwater (7.6 m) stratum alone. The contribution of the upper water column to overall standing stock is also illus- trated by the relative importanceof the cosmopolitan species. Ranging throughout the water column, kelp bass, white seaperch, and sehorita comprised about 60% of total biomass at the San Onofre kelp bed. The relative contribution of water-column species to overall standing stock would be lower in kelp forests anchored on high-relief rock, because reef-de- pendent species would be more abundant than at San Onofre. However, the presence of an extensive bathymetric zone from the canopy into midwaters provided space, forage, and orientation for a substan- tial standing stock of fishes in the San Onofre kelp bed. The lack of such an extensive midwater zone may have limited the abundance of canopy and cos- mopolitan species at Bathtub Rock and Del Mar, accounting, in part, for the relatively low estimates of standing stock in these areas. Our study, then, suggests that kelp per se can enhance the potential standing stock of fishes in an area. Our kelp-forest areas lacked a high-relief bot- tom and the species of fish that depend on it. The remaining fish were those that either tolerate or are not influenced by a cobble bottom, and those that depend intimately on kelp. Yet the standing stock of fishes at the San Onofre kelp bed was substantial. The reduced numerical abundance of fishes and smaller biomass (excluding barred sand bass) in our kelp-depauperate area further indicates the impor- tance of kelp at San Onofre. Experimental manipula- tion of kelp density is probably the best test of the influence of kelp on fish abundance (Miller and Geibel 1973; Bray 198 1; M. Carr footnote 3). We also recognize that large-scale oceanographic factors may strongly affect survivorship of planktonic larvae and the subsequent abundance of juvenile and adult fishes (Stephens and Zerba 1981; Parrish et al. 1981). However, our comparisons indicate that giant kelp, even in only moderate density, was necessary for the existence of a large standing stock of diverse fishes in cobble-bottom areas. We conclude that, while rock reefs enhance the fish fuana of an area whether or not there is kelp, the presence of kelp in an area of low-relief bottom also augments the abun- dance of juvenile and adult fish on a local scale. Kelp may also contribute strongly to the standing stock of fish in areas of high-relief bottom, but no one to date has adequately evaluated this hypothesis. We pre- dict that the densities of canopy species and cos- mopolites like kelp bass and sehorita will also prove to be related to the density of giant kelp on high- relief bottoms. ACKNOWLEDGMENTS We thank Ken Plummer and Mark Wilson for assis- tance with filming cinetransects. Sandy Larson and Jan Fox typed versions of the manuscript. Diane Fenster drafted Figures 2-4. This paper is a result of research funded by the Marine Review Committee (MRC), Encinitas, Calif. The MRC does not necessarily accept the results, findings, or con- clusions stated herein. A. W. Ebeling kindly loaned the cameras and housings used in the study. LITERATURE CITED Alevizon, W. S., and M. G. Brooks. 1975. The comparative structure of two western Atlantic reef-fish assemblages. Bull. Mar. Sci. 25:482-490. Bailey, N. T. J. 1959. Statistical methods in biology. English Univ. Press Ltd., Lond., 200 p. Bardach, J. E. 1959. The summer standing crop of fish on a shallow Ber- muda reef. Limnol. Oceanogr. 4:77-85. Bernstein, B. B., and N. Jung. 1979. Selective pressures and coevolution in a kelp canopy community in southern California. Ecol. Monogr. 49:335-355. Bray, R. N. 1981. Influence of water currents and zooplankton densities on daily foraging movements of blacksmith, Chromis punctipmnis, a planktivorous reef fish. Fish. Bull., U.S. 78:829-841. Bray, R. N., and A. W. Ebeling. 1975. Food, activity, and habitat of three "picker-type" mi- crocarnivorous fishes in the kelp forests off Santa Barbara, California. Fish. Bull., U.S. 73:815-829. Brock, R. E. 1982. A critique of the visual census method for assessing coral reef fish populations. Bull. Mar. Sci. 32:269-276. Brock, V. E. 1954. A preliminary report on a method of estimating reef fish populations. J. Wildl. Manage. 18:297-308. Burnham, K. P., D. R. Anderson, and J. L. Laake. 1980. Estimation of density from line transect sampling of 52 LARSON and DeMARTINI: DISTRIBUTION OF FISHES IN KELP FOREST biological populations. Wildl. Monogr. 72, 202 p. Caughley, G. 1977. Analysis of vertebrate populations. John Wiley & Sons, N.Y., 234 p. Clarke, T. A. 1970. Territorial behavior and population dynamics of a pomacentrid fish, the garibaldi, Hypsypops rubicun- da. Ecol. Monogr. 40:189-212. Coyer, J. A. 1979. The invertebrate assemblage associated with Mac- rocystis pyrifera and its utilization as a food resource by kelp forest fishes. Ph.D. Thesis, Univ. Southern Califor- nia, Los Angeles, 364 p. DeMartini, E. E., and M. E. Anderson. 1979. Comparative survivorship and life history of painted greenling (Oxylebius pictus) in Puget Sound, Washington and Monterey Bay, California. Environ. Biol. Fish. 5:33- 47. Dixon, W. J., and M. B. Brown (editors). 1979. BMDP-79. Biomedical computer programs, P- series. Univ. Calif. Press, Berkeley, 880 p. Ebeling, A. W., and R. N. Bray. 1976. Day versus night activity of reef fishes in a kelp forest off Santa Barbara, California. Fish Bull., U.S. 74:703- 717. Ebeling, A. W., R. J. Larson, and W. S. Alevizon. 1980a. Habitat groups and island-mainland distribution of kelp-bed fishes off Santa Barbara, California. In D. M. Power (editor), The California Islands: Proceedings of a Multidisciplinary Symposium, p. 403-431. Santa Bar- bara Museum of Natural History. Ebeling, A. W., R. J. Larson, W. S. Alevizon, and R. N. Bray. 1980b. Annual variability of reef-fish assemblages in kelp forests off Santa Barbara, California. Fish. Bull., U.S. 78:361-377. Elliott, J. M. 197 1. Some methods for the statistical analysis of samples of benthic invertebrates. Freshw. Biol. Assoc, Sci. Publ. 25, 144 p. Feder, H. M., C. H. Turner, and C. Limbaugh. 1974. Observations on fishes associated with kelp beds in southern California. Calif. Dep. Fish Game, Fish Bull. 160, 144 p. FREY, H. W. (editor). 1971. California's living marine resources and their uti- lization. Calif. Dep. Fish Game, 148 p. Hobson, E. S. 1971. Cleaning symbiosis among California inshore fishes. Fish. Bull., U.S. 69:491-523. Hobson, E. S., and J. R. Chess. 1976. Trophic interactions among fishes and zooplankters near shore at Santa Catalina Island, California. Fish. Bull., U.S. 74:567-598. Jones, R. S., and J. A. Chase. 1975. Community structure and distribution of fishes in an enclosed high island lagoon in Guam. Micronesica 11:127-148. Keast, A., AND J. HARKER. 1977. Strip counts as a means of determining densities and habitat utilization patterns in lake fishes. Environ. Biol. Fish. 1:181-188. Larson, R. J. 1980. Competition, habitat selection, and the bathymetric segregation of two rockfish {Sebastes) species. Ecol. Monogr. 50:221-239. Miller, D. J., and J. J. Geibel. 1973. Summary of blue rockfish and lingcod life histories; a reef ecology study; and giant kelp, Macrocystis pyrifera, experiments in Monterey Bay, California. Calif. Dep. Fish Game, Fish Bull. 158, 137 p. Parrish, R. H., C. S. Nelson, and A. Bakun. 1981. Transport mechanisms and the reproductive success of fishes in the California current. Biol. Oceanogr. 1:175- 203. Pienaar, L. V., and W. E. Ricker. 1968. Estimating mean weight from length statistics. J. Fish. Res. Board Can. 25:2743-2747. Quast, J. C. 1968a. Fish fauna of the rocky inshore zone. In W. J. North and C. L. Hubbs (editors), Utilization of kelp-bed re- sources in southern California, p. 35-55. Calif. Dep. Fish Game, Fish Bull. 139. 1968b. Estimates of the populations and standing crop of fishes. In W. J. North and C. L. Hubbs (editors), Utiliza- tion of kelp-bed resources in southern California, p. 57- 79. Calif. Dep. Fish Game, Fish Bull. 139. Randall, J. E. 1963. An analysis of the fish populations of artificial and natural reefs in the Virgin Islands. Caribb. J. Sci. 3:31- 47. Robins, C. R., R. M. Bailey, C. E. Bond, J. R. Brooker, E. A. Lachner, R. N. Lea, and W. B. Scott. 1980. A list of common and scientific names of fishes from the United States and Canada. 4th ed. Am. Fish. Soc, Spec. Publ. 12, 174 p. Russell, B. C. 1977. Population and standing crop estimates for rocky reef fishes of north-eastern New Zealand. N.Z. J. Mar. Freshw. Res. 11:23-36. Snedecor, G. W., and W. G. Cochran. 1980. Statistical methods. 7th ed. Iowa State Univ. Press, Ames, 507 p. Stephens, J. S., Jr., and K. E. Zerba. 1981. Factors affecting fish diversity on a temperate reef. Environ. Biol. Fish. 6:111-121. Talbot, F. H., and B. Goldman. 1972. A preliminary report on the diversity and feeding relationships of reef fishes of One Tree Island, Great Barrier Reef system. Proceedings, Symposium on Corals and Coral Reefs, 1969, p. 425-444. Terry, C. B., and J. S. Stephens, Jr. 1976. A study of the orientation of selected embiotocid fishes to depth and shifting seasonal vertical temperature gradients. Bull. South. Calif. Acad. Sci. 75:170-183. 53 THE INVERTEBRATE ASSEMBLAGE ASSOCIATED WITH THE GIANT KELP, MACROCYSTIS PYRIFERA, AT SANTA CATALINA ISLAND, CALIFORNIA: A GENERAL DESCRIPTION WITH EMPHASIS ON AMPHIPODS, COPEPODS, MYSIDS, AND SHRIMPS1 James A. Coyer2 ABSTRACT The motile invertebrate assemblage associated with the giant kelp, Macrocystis pyrifera, fronds was examined monthly from June 1975 through December 1976, at Santa Catalina Island, California. Replicate samples were collected from each of three vertical zones (canopy |C|, middle [M|, bottom [B]). The number of species collected from all zones was 114 and ranged from 51 to 75 for any given month. Amphipods, copepods, mysids, and shrimps comprised the majority of invertebrate abundance (86 [C], 92 [M], 93f; |B|) and biomass (90 [C|, 89 [M|, 86fi [B]). Gammarid amphipods dominated the assemblage in numbers (34 [C|, 60 [M], 519? |B]), biomass (34 [C|, 68 [M], 67% [B]), and number of species (20). The assemblage displayed three patterns of vertical stratification within the Macrocystis forest: 1) The mean number of species progressively decreased from the bottom to the canopy (several species displayed zone preferences); 2) more individuals and a greater total biomass were present in the lower zones than in the canopy; and 3) the mean lengths of gammarids, mysids, and shrimps were significantly larger and propor- tionately greater numbers of large individuals were present in the canopy than in either of the lower zones. Subtidal forests of giant kelp have long attracted the interest of biologists, beginning with Darwin's (1860: 240) description of the organisms associated with the giant kelp forests off Tierra del Fuego. Since the advent of scuba techniques in the mid-1950's, several studies have examined in detail the attached and/or motile species of invertebrates associated with surfaces of the giant kelp, Macrocystis pyrifera (Limbaugh 1955; Clarke 1971; Ghelardi 1971; Jones 1971; Wing and Clendenning 1971; Miller and Geibel 1973; Lowry et al. 1974; Bernstein and Jung 1979; Yoshioka 1982 a, b). Few, however, have attempted a long-term and comprehensive examina- tion of the entire assemblage of small and motile invertebrates found with the giant kelp. The as- semblage is important for several reasons, notably as the major source of food for most fishes residing within the kelp forests (see fish diet studies by Quast 1968; Hobson 197 1; Bray and Ebeling 1975; Hobson and Chess 1976). The present report examines the composition, pat- terns of vertical stratification, and seasonal dynamics of the small and motile invertebrate assemblage 'Contribution No. 37, from the Catalina Marine Science Center. :Catalina Marine Science Center (University of Southern Califor- nia), Avalon, Calif.; present address: Division of Science and Mathematics, Marymount Palos Yerdes College, Rancho Palos Ver- des, CA 90274. associated with the fronds of M. pyrifera. A general overview of the assemblage and a detailed examina- tion of the amphipods, copepods, mysids, and shrimps are presented. STUDY AREA The study area was Habitat Reef, located in Big Fisherman Cove, Santa Catalina Island, Calif, (lat. 33°28'N, long. 118°29'W). Habitat Reef is a fingerlike extension of bedrock ranging in depth from 2 to 18 m and is bounded on the three outer margins by an expansive area of shelly debris substrate. The western and northern sides of the reef slope sharply to a depth of 20-25 m, whereas the eastern edge slopes gradually to a shallower area ranging from 8 to 19 m. Water temperatures at Habitat Reef ranged from 13.6° to 21.2°C during the study, warmest dur- ing July through September and coolest from December to February. The algal community of the shoreward portion (<3 m depth) of Habitat Reef was dominated by Phyllo- spadix torreyi, Eisenia arborea, Cystoseira neglecta, and Sargassum muticum (seasonally) . The outermost portion (>3 m depth) was dominated by Macrocystis and the understory algae in this area was sparse, although small patches of Dictyopteris zonarioides and C. neglecta were present in some areas. Manuscript accepted August 1983. FISHERY BULLETIN: VOL. 82, NO. 1, 1984. 55-^ FISHKRY BULLETIN: VOL. 82, NO. 1 MATERIALS AND METHODS Zonation and Kelp Density The kelp forest at Habitat Reef was divided into three vertical zones: Canopy (C), middle (M), and bottom (B). The canopy extended from the water sur- face to a depth of 1 m, the bottom ranged from just above the kelp holdfasts to 2 m above the substrate, and the middle included the area between the canopy and the bottom. Holdfasts were not examined. Kelp density was measured by randomly establishing 25 circular 1 m: plots within the study area during November 1975 and October and December 1976. The number of enclosed plants and the number of fronds/plant were determined. Sampling Procedure Samples were collected monthly from plants in the central portion of the kelp forest (7-9 m depth) during tidal heights ranging from +1.0 to +1.3 m mean lower low water. From June through September 1975, three replicate samples were collected from each zone; from October 1975 through December 1976, five replicates were collected. Only one sample was collected from any plant, and this sample con- sisted of the entire plant portion within the desired zone. The middle and bottom zones were collected by carefully severing the upper portions and allow- ing them to drift away. Disturbance to the lower zones during this procedure was negligible. Similar amounts of kelp were collected from each zone throughout the study (n = 19; kg = 2.5[C] 2.1[M], 2.3 [B]). The kelp-associated invertebrates were collected by scuba divers maneuvering a plankton net (1 m diameter, 3 m long, 0.33 mm mesh) over the desired portion of the plant. This procedure captured most motile invertebrates on the kelp, as well as within the surrounding water column (1 m diameter). The enclosed sample was placed in a large container filled with warm freshwater (providing a thermal and salinity shock), vigorously agitated, and removed. The remaining water was filtered through a 0.25 mm sieve and the residue preserved. Thus, the term "in- vertebrate" in this investigation refers to all motile individuals larger than 0.33 mm (excluding pro- tozoans, cnidarians, and nematodes). The efficiency of the agitation-freshwater method was tested by placing the processed kelp into another container of warm freshwater and allowing it to stand for 4 h. Subsequent agitation and filtering indicated that 96% of all motile invertebrates in each zone were removed by the initial agitation-freshwater treat- ment. Organisms were identified to species (except for some juveniles). The wet weight of kelp from each sample was measured, and abundances of all taxa were expressed as the number of individuals per kilogram (wet weight) of kelp. The somewhat uncon- ventional normalization of species abundance to unit biomass was selected for three reasons. First, struc- tural complexity within the kelp forest habitat is created by interdigitating kelp blades and stipes and is a function (in part) of both kelp surface area and biomass. Many kelp-associated species, particularly the swarming mysids, may respond primarily to structural complexity of the habitat when seeking shelter and/or food. Secondly, biomass is much easier and faster to measure than is surface area (con- version ratios of kelp wet weight to surface area [both sides of blades + stipes] and kelp dry weight to wet weight are presented in Table 1). Thirdly, unit biomass will facilitate comparisons with invertebrate associations of other species of marine algae for which it is difficult to compute a unit area (i.e., bushy reds and browns). TABLE 1. — Ratios of kelp wet weight (kg) to kelp surface area (rrr) and dry weight (kg) to wet weight (kg). Wet weight/area Dry we ight/Wet weight Zone < SD n SD n Canopy Middle Bottom 021 0 19 0 42 0 002 0 002 0 040 10 10 10 0 16 0 15 0 13 0010 6 0025 6 0.042 6 Determination of Invertebrate Lengths and Biomass Growth series within the principal taxa were established. Individuals (n = 30-94) were measured to the nearest 0.04 mm, using a dissecting micro- scope and occular micrometer, blotted dry, and weighed using an analytical balance to determine length-weight relationships. Smaller and/or minor taxa (copepods, ostracods, caprellids, molluscs, etc.) were assigned constant weights based on the mean weight of 20 individuals. Vertical patterns of size-stratification were ex- amined by measuring the lengths of principal taxa within each zone for each quarter from January 1975 through October 1976. Single samples were collected in January and April 1975; subsequent samples were replicated (3 or 5). For shrimps and mysids, all (January through July 1975) or up to 75 individuals of each major species were measured 56 COVER: INVERTEBRATE ASSEMBLAGE WITH GIANT KELP from each replicate of each zone; for gammarid amphipods, at least 50 individuals (comprising all species) were measured from one randomly selected replicate of each zone. Replicates were pooled and size-frequency distributions were determined for each taxon within each of the zones. The non- parametric Kolmogorov-Smirnov (K-S) two-sample test (one-tailed) was used to test whether the values from one distribution were stochastically larger than the values from another distribution (Siegel 1956). The mean weight of an individual within a major taxon (shrimps, mysids, gammarids) was determined from the mean length and the appropriate length- weight formula. The mean weight then was mul- tiplied by the mean monthly abundance of the taxon to determine the taxon biomass. Quarterly length measurements were applied to the month preceding and following the measuring month (i.e., April measurements were assigned to March and May) for biomass measurements. Monthly abundance values of the smaller taxa were multiplied by the assigned weight to estimate the biomass. RESULTS Kelp Density Macrocystis density at Habitat Reef was high (4.7 plants/m2) from November 1975 through August 1976 (Table 2). In late September 1976, density and canopy cover were reduced ( 1 .5/nr) and continued to decline over the next 4 mo. TABLE 2. — Macrocystis density and the number of fronds/plant (± width of 959? C.I./2) at Habitat Reef. Sample size in parentheses. Date Density m- No. of fronds/plant General Taxonomic Composition of the Invertebrate Assemblage The invertebrate assemblage associated with the fronds of Macrocystis was composed primarily of amphipods, copepods, mysids, and shrimps (Tables 3, 4). Mysids and shrimps were among the largest TABLE 4.— Mean abundance (no./kg kelp ± width of 95', C.I./2) for the major invertebrate taxa within each zone. Parenthetical values are the mean length and weight (mm, mg) of each taxon; an asterisk indicates that a constant length and weight was used for all zones. All values are averaged over the entire 19-mo study. Nov. 1975 Oct 1976 Dec. 1976 4.7+2.3 (25) 1.5±0.7 (29) 0.7±0.3 (25) 3.4±1.0(1 18) 6.7+1.6 (46) 4.7+2.6 (17) Taxon Canopy Middle Bottom Gammarid amphipods 882.4 ±267.0 4,1 23.0 ±890.2 3,1 1 7 8 ± 715.3 (2.8.0.6) (1 8,0.4) (20.04) Copepods 1,1 28.0 ±370.2 1 .977.0 ± 540.8 '(0.8,0.1) 2,453.1 ±441.0 Ostracods 188.2 ± 76.7 108.8 ± 51.1 '(0.9.0.1) 65 7 ± 300 Echinoids (juv.) 13 9±25.5 260.5 ±375 4 •(0.5,0.1) 83 0± 119.2 Mysids 91 .4 ± 57.8 151 8±389 108 0 ±50. 5 (6 2.3.5) 14.7.1.3) (4.4.1 2) Molluscs 1 shelled ) 14.8 ±9.0 98.0 ±21.8 •(1.3.0.7) 1684 ±44 9 Candean shrimps 136.5 ±48 4 65 2 ±28.4 51 4 ± 16 0 (7.1,3.8) (6.0,2.7) (5 4.2 3) Platyhelminthes 31 7 ± 17 1 36.8 ± 16 2 '(-.38) 34.0 ± 17 0 Cladocerans 72.2 ±93.4 9.2 ±5 .1 '(0.7,0.1) 9.3 ±6.9 Polychaetes 88± 11 3 28.0 ±8.0 "(3 3.0.5) 17 4 ± 7.2 Cypris (barnacle) 13.4 ± 16.6 24.0 ±22.1 14. 3± 9 4 larvae '(0.7,0.1) Molluscs (nudibranchs) 10 9 ± 11.3 13. 4± 11 8 •|1 3.1.1) 8 1+75 Sphaeromatid isopods 0.1 ±0.1 0.2 ±0 1 •(2.4.1.1) 19 7 ±23.0 Caprellid amphipods 4 1 ±20 2.7 ± 1.0 '169.0.8) 1.8 ±1.3 Idoteid isopods 3.1 ± 3.2 0.1 ±0.1 •(7.2,4,0) 0 1 ±<0.1 Asteroids (juv.) 0.2 ±0.1 1.1 ±1.3 •(2.7.2.0) 0 6±08 Jaeropsid isopods 0.1 ±0.1 0.2 ±0.3 "(2.3.0.3) 1 4 + 09 Cumaceans 0 — 02±0 2 0 3 ±0.3 Brachyurans (zoea) 0 — 0 — 0.1 ± <0.1 Ophiuroids (juv.) 0 — <0.1 ± <0 1 <0 1 ± <0 1 Tanaids 0 — <0 1 ±0.1 0.1 ±0.1 TABLE 3. — The mean (± width of 95' i C.I./2) monthly abundance (no. organisms/kg kelp) and biomass (mg organisms/kg kelp) for each major invertebrate group associated with the giant kelp. Data are averaged over the entire 19-mo study; proportions of total numbers or biomass (all species) are presented in parentheses. Zone Gammarids Copepods Mysi ds Shrim 3S Total Canopy Numbers 882 ±267 (33 9) 1.128±370 (43.4) 91 ±58 (3.5) 136 ±48 (5.2) 2.599 ±580 Biomass 589 ±236 (33.8) 56 ± 18 (3.2) 336±255 (19.3) 583 ± 300 (33 4) 1,743 ±765 Middle Numbers 4.123 ±890 (59.8) 1.977 ±541 (28 7) 152 ±39 (2.2) 65 ± 28 (0.9) 6,900 ± 1,382 Biomass 1,634 ±359 (68.4) 99 ±27 (4.1) 218±68 (9.1) 1 74 ± 71 (7.3) 2,387 ± 493 Bottom Numbers 3.1 18 ± 71 5 (50.8) 2,453 ±441 (39.9) 1 08 ± 5 1 (1 7) 51 ± 16 (0.8) 6.153 ±937 Biomass 1.388 ±337 (67.4) 1 23 + 22 i6.0) 143 ±83 (6.9) 116± 37 (5.6) 2.061 ±454 57 FISHERY BULLETIN: VOL. 82, NO. 1 species present, copepods among the smallest (Table 4). The number of species collected from all zones totaled 114, but ranged from 51 to 75 for any given month (Fig 1). When ranked by the mean monthly abundance, 7 species were dominant (>100/kg), 23 were common (10-100/kg), 24 were uncommon (1- 10/kg), and 60 were rare ( E o -o 2 c -o t: o - o o ^ m □ E3 D € 1 ^ IW^:.:.x.',;,:',:.:j -S -^^£^2S -F??F -E D u I/) -i J I L JE "1 o o Q. COI -ExTi J I L i -E f^rrr ^E & lvoS Q O t" H ~ ,',,,',',',' 4— J 1 L O O O m O O o 8 di3M 6>i/a3aiAinN a H 01 C o N (0 ~ OJ > J3 oa cs E a. ?. W, cu &3 L* " CC r Oi 8 t- CO CD — J3 r — ~ u c CD 0! Q C lO ~ r~- *-. C35 ;- I— < 5 ft CU CI -O o . ~ o _ -r • CD a > w — a O y. lO r cn p i — i c CD "C: — c ~ ^ -. — ^~ " CD — X >, >> - - >-3 43 O CD J3 — ■ : o <_> c 01 b*> CD 05 £ — CD n J3 3) ■•^ c CO CO C o> CD fci CO CD 111 u — CD en CD — - = - C3 > — eg >> c c o 2 1 :•: -coi lE M J.ZZ tmmmi™ -t mm :U Zi f , ,", ' M< $mz$ M -^M:;:fr£l !■"■€ crm J I zv jzz OIC £ 'SIC t,. .,.,.. ^ S.i£ r.-x-x /s o o o +^x'^> ::*t»i ■/Z2 -^,,.,..,,^, J ;:G12 -fcv:-.-:-::i/ . . '.. . *• j i J [62 -p ::■■■: 9zz fv '^ I •l££ m\\y^ 6Z£ ■ ;:•■.•'.•:•:• 9cc -Ry^^Bze * ■teg;"':""' i'i i 1 1 in 06l s ss. ■KKVt^^tM^ s^ ^Satt^^^^*^*^ to o ^ Q. in jE .^ Q. O ■D -D o 5 U o i ^:'x*:-: S¥S¥9S|: ^^Ms -EZ s5 £S ^^13 ■n1:1:1:1.". ::9s; ^\\\\\\\^\\\\\\^^ -E ::£S:: 4^k\\\\\\\\\\\\\\\^^ ^2 o O Z3 Q. < o O Q. < O Is- m rO CM — (wuj) H19N31 AQOa ■M n CO t3 O > -C - 1> 3 — E - CO CC u "3 c C B, ; < a^ cc / 3 —> a> -e ~ a i> «2 31 t- tj ~~ — H^» ~. u CL -~ B j u < 5 JZ ■J: — CO a: £ CO a CO CO — £ Is- ■a J5 > ~ *— o O 09 c £ o ■J. c c £ c — l~ ~ i> -w > > CO 3 C 3 - — X 1) > £ u — 01 d B ^ o it-* ~- 11 O c co c CO H 0) = .- ts X £ n £ | > 3 B ■^, 1/ C CO "x o 1 {£ N a £ S 4J CJ a 11 Pu CO ♦J 62 COYER: INVERTEBRATE ASSEMBLAGE WITH GIANT KELP LENGTH (mm) Figure 5. — Combined size-frequency distributions of copepods, gammarid amphipods, mysids, and shrimps measured quarterly from July 1975 through October 1976 for each of the three vertical zones. Copepods were measured during 1 mo only because of their small size and variability. After normalization (%), the distributions of each taxon were weighted according to mean monthly abundance to create the combined distributions. The numbers of each taxon measured before weighting are (C, M, B): copepods (54, 54, 55), gammarids (308, 323, 317), mysids (2,037, 2,625, 2,500), and shrimps 1 1,896. 1,776, 1,561). Statistics determined after weighting are displayed in the figure. Mysids are remarkably specific in habitat prefer- ences. Clarke (1971) found 12-14 species of mysids cooccurring in the kelp forests off San Diego and Baja California, but only A. sculpta and S.pacifica were 10 CANOPY / •o-a V HjfG J — I — I — I — I — I — I — 1 — I I I 1 I I I I ■ I I C7< o o o 10 BOTTOM _ ~&J /V<' \ W-s-- \ x><\ va i i i i i ' ' i i ' i ' I 1 I I I I L 7 6 5 4 3 2 1 JJASONDJF MAMJ J AS0ND 1975 1976 FIGURE 6. — Monthly variation in numbers and biomass of all inver- tebrate taxa (combined) within each vertical zone. Each monthly value for the canopy, middle, and bottom represents a mean of three (June-September 1975) or five (October 1975-December 1976) replicate samples. associated with the kelp fronds. Similar patterns were observed at Habitat Reef, as both A. sculpta and S.pacifica were present in large numbers within the kelp fronds, but were rarely observed in Mac- rocystis holdfasts or other algal habitats within or near Habitat Reef (Hammer and Zimmerman 1979). Hobson and Chess (1976) found a few individuals of A. sculpta in the water column at night, but most remained closely associated with the kelp which was utilized as food. In contrast, S.pacifica migrated from kelp fronds into the surrounding open water at night to capture small plankton (Hobson and Chess 1976). 63 FISHERY BULLETIN: VOL. 82. NO. 1 Vertical Patterns of Species, Abundances, and Sizes Several of the commonly occurring species within the Habitat Reef kelp forest were far more abundant in the canopy than in the lower zones. Ampithoe plea, Hyale frequens,Acanthomysis sculpta, and Hippolyte clarki all displayed this type of distribution, and other investigators have noted the canopy prefer- ences of these species. Limbaugh (1955) described a large canopy-dwelling amphipod {Ampithoe sp.) that formed a tube by rolling and "stitching" the edge of a Macrocystis blade. Several investigators working in kelp forests off San Diego and at Habitat Reef have noted the canopy occurrence of Aconthomysis sculpta (Limbaugh 1955; Clutter 1967; Clarke 1971; Hobson and Chess 1976) and H. clarki (Hobson and Chess 1976). Lowry (unpubl., cited in Lowry et al. 1974) observed large numbers off/, californiensis, a close relative of//, clarki, in the canopy of kelp forests off central California. The canopy contained larger gammarids, mysids, and shrimps as well as proportionately greater num- bers of large individuals of these groups than in either of the lower zones. Size-selective predation by fishes frequently has been documented to be a major factor in structuring aquatic communities (Brooks and Dodson 1965; Archibald 1975; Vince et al. 1976; Macan 1977; Nelson 1979) and may account for the size distributions of invertebrates observed at Habitat Reef. The interdigitating fronds of the canopy greatly increase the structural complexity in this zone and may offer more spatial refuge for motile invertebrates than provided by the middle and bot- tom zones. As increased structural complexity has been demonstrated to decrease effectiveness of prey capture by fishes, particularly larger prey (Vince et al. 1976; Brock 1979; Coen et al. 1981; Heck and Tho- man 1981; Savino and Stein 1982), the canopy com- plexity may discourage extensive foraging by fishes. Relatively few fishes forage within the kelp canopies off southern California. The most abundant fish is the kelp perch, Brachyistius frenatus, a small diurnal species that forages preferentially in the canopy and preys extensively on small gammarids and copepods (Hobson 1971; Bray and Ebeling 1975; Hobson and Chess 1976). Other fishes are observed in the kelp canopy, but the large-mouthed species are much less abundant than the kelp perch and forage more often in other areas of the kelp forest, and the small-mouthed species capture small planktonic prey or utilize small invertebrates attached directly to the kelp surfaces (Bray and E bel- ing 1975; Hobson and Chess 1976; Bernstein and Jung 1979). Consequently, predation pressure on larger individuals of motile prey in the canopy may be reduced relative to the lower zones, resulting in a pro- portionately greater abundance of larger individuals. For example, the mysid S. pacifica was much more abundant in the lower zones than in the canopy, yet the largest individuals consistently were present in the canopy. Alternate hypotheses may explain the size stratification of some species. Intraspecific behav- ioral interactions may confine certain size classes to specific zones, as demonstrated experimentally for an amphipod (Van Dolah 1978). Larger individuals may be more abundant in the canopy simply in re- sponse to the presence of preferred food types and/ or sizes, although this hypothesis has not been examined. The size distribution of invertebrates in the lower zones resembled the size distribution of insects in temperate terrestrial forests (Schoener 197 1), in that both areas supported large numbers of small, and few large, individuals. The size distribution in the canopy, however, was somewhat similar to the insect size distribution of tropical terrestrial forests where there are proportionately greater numbers of large insects (Schoener and Janzen 1968; Schoener 1971). The presence of larger insects in the tropical forests effectively expands the food size dimension relative to the temperate forests (assuming equal abundance). The expansion has been hypothesized to account for some of the increased diversity of bird species in the tropics, as much of this increase is due to the addition of insectivorous birds adapted to capture large insects (Schoener 1971). In contrast to the tropical forests, the higher propor- tion of large prey items in the Habitat Reef kelp canopy apparently did not attract additional species of fish predators. Nevertheless, it may be useful to examine the size distributions of important prey items in other kelp forests to determine whether a relationship exists between prey size distributions and fish species diversity. Seasonal Patterns of Species, Abundances, and Sizes The kelp-associated invertebrates as a group did not exhibit seasonal cycles. Numbers and biomass generally were highest during winter 1975, with the marked increase in biomass due primarily to increased abundances of the relatively large canopy mysid A. sculpta and shrimp H. clarki. Gammarid amphipods, particularly M. litotes, were largely re- 64 COYER: INVERTEBRATE ASSEMBLAGE WITH CIANT KELP sponsible for the increased abundances in the lower zones during this period. Fluctuations in the population size of several species may have been associated with changes in kelp biomass, particularly the general decline of kelp biomass beginning in fall 1976. The canopy mysid probably attains its greatest population size during winter; however, the canopy was markedly reduced in area by winter 1976-77 and the mysid was rare. Copepods and gammarids displayed decreased canopy abundances during late 1976, and in the lower zones, abundances of the gammarid M. litotes began to decline as kelp biomass was reduced. As the canopy mysid andM. litotes were major components of the general invertebrate peak observed during winter 1975-76, their reduced abundances in late 1976 undoubtedly were a major reason for the absence of a general invertebrate peak in late 1976. Reduction in kelp biomass, however, did not affect H. clorki. Even though the shrimp was most numerous in the canopy, its abundance in the reduced canopy of late 1976 was similar to levels recorded in the larger canopy of late 1975. Although the amount of kelp biomass ultimately must determine the abundance and occurrence of kelp-associated invertebrates, the importance of proximal factors remains to be determined. Proximal factors may be particularly important in many areas of southern California, where the kelp forests are characterized by relatively long-term cycles of loss and renewal (Rosenthal et al. 1974). In such con- ditions of relative biomass constancy, abundances of some species may not be correlated with seasonal changes (i.e., temperature, day length, nutrients, etc.). Additional research is necessary to determine the importance of proximal factors such as kelp quality (healthy vs. decomposing), inter- and intra- specific competition for space and food, and preda- tion by fishes and/or motile invertebrates, in determining the abundance and occurrence of kelp- associated invertebrates. ACKNOWLEDGMENTS The manuscript was adopted from a portion of a doctoral dissertation completed at the University of Southern California. I thank my committee, chaired by J.N. Kremer, and am grateful to R. L. Zimmerand R. R. Given for their support and cooperation at the Catalina Marine Science Center. The substantial field assistance of J. R. Chess, J. F. Pilger, C. S. Shoemaker, and T. E. Audesirk is sincerely appre- ciated. Special thanks to D. Cadien, J. R. Chess, G. Kramer, B. Myers, J. Soo-Hoo, J. Word, and R. C. Zimmerman for assistance with species identifica- tion and to G. S. Hageman for help in sorting samples. The valuable suggestions of R. J. Schmitt, R. F. Ambrose, and two anonymous reviewers improved earlier drafts of the manuscript. The research was supported in part by the NOAA Office of Sea Grant under Grant No. USDC 04-158- 44881 to the University of Southern California and by Sea Grant Traineeships. LITERATURE CITED Archibald, C. P. 1975. Experimental observations of the effects of predation by goldfish (Carrassius auratus) on the zooplankton of a small saline lake. J. Fish. Res. Board Can. 32:1589- 1594. Bernstein, B. B.. and N. Jim; 1979. Selective pressures and coevolution in a kelp canopy community in southern California. Ecol. Monogr. 49:335-355. Bray, R. N.. and A. W. Ebeling. 1975. Food, activity, and habitat of three "picker-type" mi- crocarnivorous fishes in the kelp forests off Santa Bar- bara, California. Fish. Bull., U.S. 73:815-829. Brock. R. E. 1979. An experimental study on the effects of grazing by parrotfishes and role of refuges in benthic community structure. Mar. Biol. (Berl.) 51:381-388. Brooks. J. L., and S. I. Dodson. 1965. Predation, body size, and composition of plankton. Science (Wash., D.C.) 150:28-35. Clarke, W. D. 1971. Mysids of the southern kelp region. In W. J. North (editor), The biology of the giant kelp beds (Macrocystis I in California, p. 369-380. Beih. Nova Hedwigia 32. Clutter, R. I. 1967. Zonation of nearshore mysids. Ecology 48:200-208. Coen, L. D.. K. L. Heck, Jr.. and L. G. Abki.e. 1981. Experiments on competition and predation among shrimps of seagrass meadows. Ecology 62:1484-1493. Darwin, C. 1860. The voyage of the Beagle. Anchor Books, Doubleday and Co., Garden City, N.Y., 524 p. (1962) Ghelardi, R. J. 1971. Species structure of the animal community that lives in Macrocystis pyrifera holdfasts. In W. J. North (editor). The biology of giant kelp beds (Macrocystis ) in California, p. 381-420. Beih. Nova Hedwigia 32. Hammer, R. M., andR. C. Zimmerman. 1979. Species of demersal zooplankton inhabiting a kelp forest ecosystem off Santa Catalina Island, Califor- nia. Bull. South. Calif. Acad. Sci. 78:199-206. Heck, K. L., Jr., andT. A. Thoman. 1981. Experiments on predator-prey interactions in vegetated aquatic habitats. J. Exp. Mar. Biol. Ecol. 53:125-134. HOBSON. E. S. 1971. Cleaning symbiosis among California inshore fishes. Fish. Bull, U.S. 69:491-523. HoBsn\. e. S., and. J. R. Chess. 1976. Trophic interactions among fishes and zooplankters 65 FISHERY BULLETIN: VOL. 82, NO. 1 near shore at Santa Catalina Island, California Fish. Bull., U.S. 74:567-598. Jackson, c, a. 1977. Nutrients and production of giant kelp, Macrocystis pyrifera, off southern California. Limnol. Oceanogr. 22:979-995. Jones, l. G. 1971. Studies on selected small herbivorous invertebrates habiting Macrocystis canopies and holdfasts in southern California kelp beds. In W. J. North (editor), The biology of giant kelp beds (Macrocystis) in California, p. 343- 367. Beih. Nova Hedwigia 32. LlMBAI GH, C. 1955. Fish life in the kelp beds and the effects of kelp harvest- ing. Univ. Calif. Inst. Mar. Resour. Ref. 55-9, 158 p. Lowry, L. F., A. J. McElroy, and .J. S. Pearse. 1974. The distribution of six species of gastropod molluscs in a California kelp forest. Biol. Bull. (Woods Hole) 147:386-396. Macan.T.T. 1977. The influence of predation on the composition of fresh- water animal communities. Biol. Rev. 52:45-70. Miller, D. J., and J. J. Geibel. 1973. Summary of blue rockfish and lingcod life histories; a reef ecology study; and giant kelp, Macrocystis pyrifera experiments in Monterey Bay, California. Calif. Dep. Fish Game, Fish Bull. 158, 137 p. Nelson, W. G. 1979. Experimental studies of selective predation on amphipods: Consequences for amphipod distribution and abundance. J. Exp. Mar. Biol. Ecol. 38:225-245. North, W. J. 1971. Introduction and background. In W. J. North (editor), The biology of giant kelp beds [macrocystis) in California, p. 1-96. Beih. Nova Hedwigia 32. QUAST, J. C. 1968. Observations on the food of the kelp-bed fishes. In W. J. North and C. L. Hubbs (compilers and editors), Utiliza- tion of kelp-bed resources in southern California, p. 109- 142. Calif. Dep. Fish Game, Fish Bull. 139. Rosenthal, R. J., W. D., Clarke, and P. K. Dayton. 1974. Ecology and natural history of a stand of giant kelp, Macrocystis pyrifera, off Del Mar, California. Fish. Bull., U.S. 72:670-684. Sayino, J. F., and R. A. Stein. 1982. Predator-prey interaction between largemouth bass and bluegills as influenced by simulated, submerged vegetation. Trans. Am. Fish. Soc. 111:255-266. SCHOENER, T. W. 1971. Theory of feeding strategies. Annu. Rev. Ecol. Syst. 2:369-404. SCHOENER, T. W., AND D. H. JANZEN. 1968. Notes on environmental determinants of tropical ver- sus temperate insect size patterns. Am. Nat. 102:207- 224. SlEGEL, S. 1956. Nonparametric statistics for the behavioral sciences. McGraw-Hill, N.Y., 312 p. Van Dolah, R. F. 1978. Factors regulating the distribution and population dynamics of the amphipod (lammaruspalustris in an inter- tidal salt marsh community. Ecol. Monogr. 48:191-217. Vince, S., I. Valiela, N. Backus, and J. M. Teal. 1976. Predation by the salt marsh killifish Fundulus heternclitus (L.) in relation to prey size and habitat struc- ture: Consequences for prey distribution and abundance. J. Exp. Mar. Biol. Ecol. 23:255-266. Wing, B. L., and K. A. Clendenning. 1971. Kelp surfaces and associated invertebrates. In W. J. North (editor), The biology of giant kelp beds (Macrocys- tis) in California, p. 319-341. Beih. Nova Hedwigia 32. YOSHIOKA, P. M. 1982a. Predator-induced polymophism in the bryozoan Membranipora membranacea (L.). J. Exp. Mar. Biol. Ecol. 61.233-242. 1982b. Role of planktonic and benthic factors in the popula- tion dynamics of the bryozoan Membranipora mem- branacea. Ecology 63:457-468. 66 SPRING AND SUMMER PREY OF CALIFORNIA SEA LIONS, ZALOPHUS CALIFORNIANUS, AT SAN MIGUEL ISLAND, CALIFORNIA, 1978-79. George A. Antonelis, Jr., Clifford H. Fiscus, and Robert L. DeLong1 ABSTRACT During the late spring and summer of 1978 and 1979, 224 scats were collected from rookeries of the Cali- fornia sea lion, Zalophus californianus , at San Miguel Island for the purpose of identifying prey species. A total of 2,629 otoliths and 2,06 1 cephalopod beaks were recovered. The frequency of occurrence for the four most commonly identified prey species was 48.7% Pacific whiting, Merluccius productus; 46.7% market squid, Loligo opalescens; 35.9% rockfish, Sebastes spp.; and 20.0% northern anchovy, Engraulis mordax. Seasonal variability in the frequency of occurrence of these four prey species from late spring to summer indicates that California sea lions feed opportunistically on seasonally abundant schooling fishes and squids. Five species of fish (California smoothtongue, Bathylagus stilbius; northern lampfish, Stenobrachius leucop- sarus; chub mackerel, Scomber japonicus; medusafish, Icichthys lockingtoni; sablefish, Anoplopoma fimbria) and one cephalopod (two-spotted octopus, Octopus bimaculatus) were identified as previously unreported prey of the California sea lion. The California sea lion, Zalophus californianus, is the most abundant pinniped inhabiting the coastal waters off California (Le Boeuf and Bonnell 1980). During the summer most California sea lions are on or near their breeding sites which are located on islands south of Point Conception, along the coast of southern California, Baja California, and into the Gulf of California. After the breeding season in the summer, a portion of the subadult and adult male sea lion populations migrates north of Point Conception as far as British Columbia, while the rest of the pop- ulation remains off the coasts of southern California and Baja California, Mexico (Peterson and Bartho- lomew 1967). Numerous studies of the food of migrant male California sea lions have been con- ducted in the areas north of their traditional breeding islands (Briggs and Davis 1972; Jameson and Kenyon 1977; Morejohn et al. 1978; Bowlby 1981; Everitt et al. 1981; Jones 1981; Ainley et al. 1982; Bailey and Ainley 1982), while comparatively little information has been reported on the feeding behavior of sea lions in areas off the coast of Cali- fornia south of Point Conception (Rutter et al. 1904; Scheffer and Neff 1948; Fiscus and Baines 1966). From the information presented in all of these studies, it has been suggested that California sea lions feed opportunistically on a variety of prey species (Antonelis and Fiscus 1 980) and that "switch feeding" is probably an important component of their feeding behavior (Bailey and Ainley 1982). However, since most of the information on sea lion feeding behavior is based on observations north of their breeding islands, additional information from within their breeding range would allow us to deter- mine if similar feeding characteristics can be expect- ed in other geographical areas. Studies conducted before 1970 usually obtained stomach contents for feeding information by killing sea lions, while most post- 1970 feeding studies have used nonlethal techniques including examination of scats and oral rejecta (spewings) and direct behavioral observations. Another method was the examination of gastrointestinal tracts from animals found dead. In this study, prey-species classification is based on the identification of fish otoliths and cephalopod beaks found in scats collected during the spring and summer for two consecutive years on the California sea lion rookeries of San Miguel Island, Calif. In addition to the identification of prey, we calculated the percent frequency of occurrence of each prey, compared annual and seasonal differ- ences in prey selection, and estimated the lengths and weights of the most frequently occurring prey species. 'Northwest and Alaska Fisheries Center National Marine Mammal Laboratory, National Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115. Manuscript accepted July 1983. FISHERY BULLETIN: VOL. 82, NO. 1, 1984. MATERIALS AND METHODS Scats were collected from areas utilized exclusively 67 FISHERY BULLETIN: VOL. 82, NO. 1 by California sea lions on the west end of San Miguel Island, Calif., during spring (2-3 May 1978; 2 and 16 May 1979) and summer (3-4 August 1978; 30-3 1 July 1979). During both sample periods, scats were collected from areas where mostly females and juveniles of both sexes occurred and relatively few (<12% of the total animals censused) adult and sub- adult males were present. In order to document the occurrence of prey species which were consumed at or close to the time of collections, only recent scats, which showed no obvious signs of desiccation, were collected. Each scat was placed in a plastic bag, where it was later soaked in water or a solution of about 1 part liquid detergent to 100 parts water for about 24 h. Each bag was shaken occasionally to facilitate emulsification of the digested organic material, and then rinsed with water through three nested sieves with screen mesh sizes of 3.35 mm, 2.00 mm, and 1.00 mm from top to bottom. After most of the soft digested organic material was washed away, fish otoliths and cephalopod beaks were removed and stored in a solution of 70% ethanol. Prey totals were determined by using the higher number of left or right otoliths and upper or lower squid beaks. The otoliths were identified by the late J. Fitch, California Department of Fish and Game, Long Beach, Calif., the octopus beaks by E. Hochberg, Santa Barbara Museum of Natural His- tory, Santa Barbara, Calif., and the squid beaks by the second author. The data for each of the four major prey species were summarized by a three-way (2X2X2) con- tingency table and tested for independence of occurrence by season, year, and both season and year (Fienberg 1977). Length measurements of these otoliths and squid beaks were used to estimate the body lengths or ages of the most frequently occurring prey species. Although many otoliths and beaks of all sizes were recovered from the scats in good condition, some were not measured because they were broken or showed obvious signs of damage from digestion. We assumed that damage to the otoliths and squid beaks collected in this study was not dependent on size. Lengths of northern anchovy, Engraulis mordax, were estimated from a regression equation of fish lengths on otolith lengths (Spratt 1975). Length information for rockfish, Sebastes spp., was obtained from previously reported data (Phillips 1964) for specimens (bocaccio, Sebastes paucispinis) of the same age as most of the rockfish reported in this study. Bocaccio was chosen as the representative rockfish because it has been reported as the most abundant rockfish in the waters near San Miguel Island (Best and Oliphant 1965). The regression equation used to estimate the length of Pacific whit- ing, Merluccius productus, was derived in this study from specimens collected off the coast of southern California by the National Marine Fisheries Service (NMFS). The Pacific whiting otoliths and the corre- sponding length information were provided by K. Bailey of the NMFS Northwest and Alaska Fisheries Center, Seattle, Wash. Market squid, Loligo opales- cens, lengths were estimated from a regression equa- tion of dorsal mantle length on upper hood length of the beak. Upper hood measurements were chosen for the estimation of squid lengths because they were reported as having the highest correlation to dorsal mantle length (Kashiwada et al. 1979). In order to detect changes in the diet which would reflect apparent yearly changes in the age and size composition of a specific prey-species population, we compared the lengths of otoliths for 1978 and 1979 using the Wilcoxon rank sum test (Hollander and Wolfe 1973). Weight estimates of the most frequently occurring prey species were obtained by using the prey length estimate (described above) in regression equations of length and weight measurements or by obtaining weight data from fish which were the same age as those identified in the scats (Phillips 1964; Fields 1965; Dark 1975; Pacific Fishery Management Council 1978). The total estimated weight for each of the four major prey species was obtained by mul- tiplying the weight of the average-sized prey by the number of individuals represented in the scat collec- tion. Differences between these estimates could not be statistically analyzed because the raw data for the growth curves of each species were not available. The names of fishes follow Fitch and Lavenberg (1968) and Robins (1980), and those of cephalopods follow Fields (1965) and Young (1972). RESULTS We collected 224 California sea lion scats on San Miguel Island during the spring and summer of 1978 and 1979. From 195 (87%) of those scats, we recovered 2,629 otoliths and 2,061 cephalopod beaks. Twenty-nine (13%) scats did not contain otoliths or cephalopod beaks. The prey species iden- tified in the scats are shown in Table 1 by their per- centage of occurrence. The four most frequently occurring prey in scats containing otoliths and/or cephalopod beaks were Pacific whiting (48.7%), market squid (46.7%), juvenile rockfish from the Sebastes paucispinis -goodei-jordani complex 68 ANTONELIS ET AL.: SPRING AND SUMMER PREY OF CALIFORNIA SEA LIONS TABLE 1. — Percentage occurrence of all prey species identified from 195 California sea lion scats collected on San Miguel Island, Calif., spring and summer, 1978-79. Prey Oc currence Scientific name Common name No. % Merluccius producws Pacific whiting 95 48.7 Loligo opalescens market squid 91 46 7 Sebastes spp. rockfish (juvenile) 70 359 Engrauhs mordax northern anchovy 39 20.0 Octopus rubescens red octopus1 19 9.7 Trachurus symmetncus jack mackerel 9 46 Onychoteuthis nail squid 9 46 boreahjapomcus Gonatidae (other than squid 8 4.1 Gonatus sp.) Scomber japonicus2 chub mackerel 7 3.6 Pepnlus similhmus Pacific pompano 5 2.5 Symbolophorus California lantern- 5 25 californiensis fish Gonatus sp. squid 2 1.0 Microstomus pacificus Dover sole 2 1.0 Bathylagus stilbius2 California smooth- tongue 2 1.0 Senphus pohtus queenfish 2 10 Zalembtus rosaceus pink surf perch 0.5 Anoplopoma fimbria2 sablefish 05 Poncbthys notatus plainfin midshipman 0.5 Ictchthys lockingtoni2 medusafish 0 5 Stenobrachius leucopsarus2 northern lampfish 05 Octopus bimacu/atus2 two-spotted octopus 0.5 1 Pelagic life stage 2Not previously reported as prey of the California sea lion. (35.99c)2, and northern anchovy (20.0%). All other prey species occurred in <10.0% of the scats. Relative length and weight estimates of the four major prey species and the information used to calcu- late these estimates are shown in Figure 1 and Table 2, respectively. The length and weight information for rockfish is from data reported by Philips (1964) for one of the three species (5. paucispinis) repre- sented in this study. Measurements of otoliths from Pacific whiting and northern anchovy provided sufficient information to compare changes in the size and age of each prey group from 1978 to 1979. For Pacific whiting the lengths of otoliths were significantly greater (W* = •'About 95Tc of the juvenile rockfish were yearlings and were in- cluded in this three-species complex because their otoliths are too similar to differentiate. 5.82, P< 0.0001) in 1979 (x = 7.71 mm,n = 90) than in 1978 (x = 6.71 mm, n = 132). From these otolith measurements, we estimated the mean length of Pacific whiting at 156 mm in 1978 and 176 mm in 1979. All of the Pacific whiting otoliths were obtained from 1- and 2-yr-old fish. The occurrence of 1-yr-old fish in the sea lion diet was estimated at 98.5% in 1978 and 70% in 1979. For northern anchovy, the lengths of otoliths were significantly greater (W* = 4.36,P < 0.0001) in 1978 (j = 3.58 mm, n= 19) than in 1979 (x = 3.01 mm, n = 75). For these otolith measurements we estimated the mean length of northern anchovy at 111 mm in 1978 and 92 mm in 1979. Although all age classes of northern anchovy were recovered from the scats, there was a notable change in the percent occurrence of yearling fish from 1978 (42%) to 1979 (81%). The percentage of occurrence in the four major prey species is shown for the spring and summer of 1978 and 1979 in Figure 2. From the three-way con- tingency table analysis, it was determined that Pacific whiting occurred significantly more frequent- ly in 1978 than in 1979 (P < 0.01), and there was a greater percentage of occurrence in spring than in summer (P < 0.01). For rockfish, there was no signifi- cant difference in occurrence between years; however, there was a greater percentage of occurrence in the summer than in spring (P < 0.01). The percentage occurrence of northern anchovy was not significantly different between season, but there was a significantly greater occurrence in 1979 than in 1978 (P < 0.01). The relative proportion of oc- currence for the two seasons for each year was significantly different (P < 0.01) for Pacific whiting, rockfish, and northern anchovy. Tests of significance could not be done for market squid because of the strong three-way interaction between occurrence, season, and year. It is apparent, however, that the percent occurrence of market squid did increase from spring to summer during both years of the study (Fig. 2). Table 2.— Information used in estimating the length of the four major prey species identified from the scats of California sea lions, on San Miguel Island, Calif., 1978-79. Prey species Regression equation R2 Reference Market squid Y = 0.243 + 0.0481X 60 0.974 upper hood dorsal mantle Kashiwada length (mm) length (mm) etal. 1979 Pacific whiting Y = 26 2 + 19.38X 84 0 977 fork otolith This study length (mm) length (mm) Juvenile rockfish' (') 155 (') (') Phillips 1964 Northern anchovy Y = -8 4946 + 33 216X 677 0.774 standard otolith Spratt length (mm) length (mm) 1975 'Length measurements are from yearling bocaccio. Sebastes paucispinis. 69 FISHERY BULLETIN: VOL. 82. NO. 1 MARKET SQUID x=127mm (Weight = 47.0 g) SD = 17 mm Range = 62-185 mm n = 76 PACIFIC WHITING x= 166 mm (Weight = 42.6 g) SD = 60 mm Range = 89-261 mm n = 222 BOCACCIO (Rockfish) x=171mm (Weight = 45.4 g) SD = 22 mm Range = 129-227 mm n = 155 ^7*V 4tZ NORTHERN ANCHOVY x = 95 mm (Weight = 10.8 g) SD = 8 mm Range = 55-141 mm n = 94 Fir.i re 1 .—Relative length and weight estimates of the four major prey species identified in California sea lion scats collected on San Miguel Island, Calif., spring and summer, 1978-79. Methods used to calculate these estimates are shown in Table 2. The number of prey species occurring in individual scats changed from spring to summer. For combined years, the percentages of scats containing single or multiple prey are shown in Figure 3. In the spring, the percentage of singly occurring prey species in scats was 59.7%; in the summer the percentage dropped to 34.6%. Scats containing more than one prey increased from 17 species combinations occurring in 40.3% of the scats in the spring to 23 in 65.4% during the summer. The percentages of the total estimated weight of the four major prey species for spring and summer are 70 ANTONELIS ET AL.: SPRING AND SUMMER PREY OF CALIFORNIA SEA LIONS Number of scats Spring 1978 n=21 1979 n=46 Summer 1978 n=43 1979 n=85 1978 1979 Northern anchovy Juvenile Rockfish Market squid Pacific whiting Spring Summer n = 39 n = 18 n = 26 -i i i l i i i i l i i ■ ■ ■ i ■ ■ ' ■ ■ ■ 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Occurrence (percent) Occurrence (percent) FIGURE 2. — Spring and summer occurrence (percentage) of the four major prey species identified in California sea lions scats collected on San Miguel Island, Calif., 1978-79. Spring, n = 67 Summer, n = 128 c u v- k_ L. D U O O 2 3 4 5 Number of different species FIGURE 3. — Occurrence of single and multiple prey species in in- dividual sea lion scats collected on San Miguel Island, Calif., 1978- 79. shown in Figure 4. The seasonal changes in the per- cent of weight for Pacific whiting showed a decrease from spring to summer in 1978 and 1979, while an increase occurred from spring to summer for market squid in 1978 and rockfish in 1979. There was relatively little change in the percentage of weight between the two seasons for market squid in 1979 and rockfish in 1978. The northern anchovy also showed little difference between the two seasons during both years. Additionally, the results from this analysis show that market squid made the greatest contribution to the total estimated weight of prey in the summer of 1978 (71.2%) and for both spring (53.9%) and summer (48.7%) of 1979, while Pacific whiting made the greatest contribution to the total estimated weight only in the spring of 1978 (87.3%). DISCUSSION Pacific whiting, market squid, juvenile rockfish, and northern anchovy were the four most important prey of California sea lions at San Miguel Island during the spring and summer of 1978 and 1979. These four prey species have also been reported as common prey of California sea lions in areas north of Point Conception (Morejohn et al. 1978; Everitt et al. 1981; Jones 1981; Ainley etal. 1982) and exemplify 71 FISHERY BULLETIN: VOL. 82, NO. 1 Estimated weigh in kg Spring 1978 1979 10.7 kg 35.0 kg Summe • 1978 1979 29.7 kg 29.3 kg 1978 Northern anchovy 0 y Spring Summer Juvenile Rockfish - "•"""" * ■n mam ■ Pacific whiting - i i i i i 1 1 I l ( ) 10 20 30 40 50 60 70 80 90 100 Percentage 1979 0 10 20 30 40 50 60 70 80 90 100 Percentage FIGURE 4.— Percentages of the total estimated weight for the four major prey species in spring and summer, 1978-79. the type of large, dense schooling prey which are commonly fed upon by many of the pinnipeds in the coastal waters off California (Antonelis and Fiscus 1980). Furthermore, the variety of food items re- ported in this and other studies (Jameson and Kenyon 1977; Morejohn et al. 1978; Bowlby 1981; Jones 1981; Ainley et al. 1982) indicates that California sea lions are capable of foraging on a wide range of fish and cephalopods. The range in the average length estimates of the four major prey species (95-171 mm) does not exhibit a great diversity in size, and may reflect a size pref- erence for sea lions feeding in the waters near San Miguel Island. Both Pacific whiting and rockfish attain a much larger size as adults (Phillips 1964; Dark 1974), while the length estimates of northern anchovy and market squid are within the size range of juveniles and adults (Fields 1965; Spratt 1975). As more information is obtained on the prey and the foraging behavior of California sea lions, researchers will attempt to evaluate the biomass of each prey species consumed (Bailey and Ainley 1982). These types of studies require information on the variations in the diet of California sea lions throughout their range. For this reason, we compare the estimated length data of market squid and Pacific whiting from this study with similar information reported in areas north of Point Conception. The estimated lengths were similar for market squid which were preyed upon by California sea lions in Monterey Bay, Calif. (Morejohn et al. 1978) and in the waters near San Miguel, with mean values of 130 mm (Morejohn et al. 1978, estimated from figure 27) and 127 mm, respect- ively. California sea lions foraged on all age classes of market squid in both areas. For Pacific whiting, however, differences between the northern and southern range of the California sea lion were apparent, with estimated length averages ranging from 250 to 360 mm at Southeast Farallon Island, Calif. (Bailey and Ainley 1982) compared with an average of 166 mm at San Miguel Island. Primarily 1- and 2-yr-old fish were preyed upon near San Miguel, while 2- and 3-yr-old fish were reported as prey at Southeast Farallon. From these comparisons, we assume that squid of all sizes and age classes will be preyed upon by California sea lions, in both their breeding and non- breeding ranges. For Pacific whiting, however, there are apparent differences in the size and age classes consumed by California sea lions in the two areas. These differences could be related to three possible factors: 1) There could be differential feeding according to various age and/or sex classes of sea lions which occur in the two areas. When present, there are mostly subadult and adult males at Southeast Farallon Island, and at San Miguel Island there are comparatively fewer subadult and adult males and many more females and juveniles of both sexes (Peterson and Bartholomew 1967; Le Bouef and Bonnell 1980; Ainley et al. 1982). 2) Differences 72 ANTONELIS ET AL.: SPRING AND SUMMER PREY OF CALIFORNIA SEA LIONS between the two areas may be an artifact of the dif- ferent methods used for estimating fish length. 3) What appears most probable to us, is the differential geographical distribution of Pacific whiting accord- ing to age. Generally, the younger fish occur in the southern portion of their range, and, although there is some overlap in age groups, the age and size of the fish increase in a northward direction (Bailey et al. 1982). In cases where sufficient life history information is available, seasonal or annual changes in the occurrence of the four major prey (Fig. 2) can be related to known changes in the prey's relative abun- dance and availability to California sea lions. During both years of this study, the decrease in the occurrence of Pacific whiting in the scats from spring to summer appears to reflect known changes in the migration pattern of the species when adults and a portion of the juvenile population migrate toward shore and north of Point Conception (T. Dark'). For market squid and juvenile rockfish, however, the movement patterns off the coast of California are conspicuously different than Pacific whiting. Gener- ally, market squid increase in abundance in shallow waters (5-50 m depth) near the northern California Channel Islands in late spring, and peak numbers occur in the early summer during spawning (S.Kato4). Inspection of the unpublished data from the 1970-75 commercial catches of market squid within 30 nmi of Point Bennett, San Miguel Island, also indicated that peak abundance occurs during the summer months.5 Similarly, in spring through summer, juvenile rock- fish (S. paucispinis and 5. jordani) from the three- species complex identified in this study begin to move into more shallow waters (5-50 m depth) as they com- plete the pelagic stage of their life cycles (E. Hob- son6). In these three instances, seasonal changes in the relative availability of Pacific whiting, market squid, and juvenile rockfish are reflected in the fre- quency of their occurrence in sea lion scats. A similar relationship was also suggested by Bailey and Ainley (1982), when they observed a seasonal change in the prey consumed by California sea lions near the Farallon Islands. Although the percentage of occurrence of northern 'T. Dark, Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, Seattle, WA 98112, pers. commun. 1982. 4S. Kato, Southwest Fisheries Center Tiburon Laboratory, Na- tional Marine Fisheries Service, NOAA, Tiburon, CA 94920, pers. commun. 1981. 'Data provided bv E. Knaggs, Calif. Dep. Fish and Game, Long Beach, Calif., 1982". fiE. Hobson, Southwest Fisheries Center Tiburon Laboratory, Na- tional Marine Fisheries Service, NOAA, Tiburon, CA 94920, pers. commun. 1981. anchovy in the scats showed no significant seasonal changes from spring to summer, the annual occurrence of otoliths from northern anchovies in the sea lion scats was significantly greater in 1979 than in 1978. Their low numbers in the 1978 scats could be related to a decline in the northern anchovy population resulting from poor recruitment of the 1974-77 year classes (Mais 1981). In 1978, however, the year-class recruitment was strong (Mais 1981), and the increased abundance appears to be reflected in an increased percentage of occurrence in the 1979 collection. This explanation is corroborated by our comparison of the northern anchovy otoliths collect- ed during the 2 years, where we found that the 1979 scats contained significantly smaller fish which were mostly (81%) yearlings from the 1978 year class. Differences in the annual occurrence of Pacific whiting and market squid were also noted in this study. For market squid, there was no fishery infor- mation available during the time of this study which would provide us with a possible explanation for these differences. With Pacific whiting, however, the decrease in occurrence in the scats from 1978 to 1979 appears to be related to exceptionally high re- cruitment of the 1977 year class which was followed by an average, or possibly somewhat less-than- average, recruitment in 1978 (T. Dark footnote 3). This information is corroborated by a comparison of the Pacific whiting otoliths collected during the 2 years of our study. In 1978, sea lions preyed upon significantly smaller fish which were mostly (98.5%) yearlings from the 1977 year class. Our analysis of the frequency occurrence of prey species per individual scat (Fig. 3) suggests that California sea lions commonly feed on single prey species during the spring and feed more frequently on multiple prey species in the summer. This shift from single to multiple occurrence of prey species in scats could reflect a decrease in the overall availability of the potential prey species in the sum- mer which may necessitate foraging on a greater variety of food items for survival (Morse 1980). Alter- natively, numerous potential prey species may become more available (Morse 1980) during the summer; thus, California sea lions could forage opportunistically on a greater variety of schooling fishes or squids which concentrate in a comparatively small area of high productivity. There are, however, a variety of factors which could affect prey-species availability. Seasonal migration, diel vertical migration, variability in schooling behavior, or physiological changes associated with spawning (Moyle and Cech 1982) are probably some of the more important factors related to prey selec- 73 FISHERY BULLETIN: VOL. 82, NO. 1 tion and preference of California sea lions which necessitate additional research. Unfortunately, virtually no information has been reported on the digestive rates or retention time of the prey species' hard parts in California sea lions. Therefore, it is presently impossible to ascertain how many meals, or portions thereof, are represented in a single scat. There is some evidence, however, from feeding studies (Pitcher 1980) of harbor seals, Phoca uitulina, and (Miller 1978) northern fur seals, Callorhinus ursinus , which indicates that cephalopod beaks are not readily passed through the intestinal tract and are regurgitated. This would result in an underrepresentation of cephalopod beak percent- age-of-occurrence data from scats as suggested by Morejohn et al. (1978). Furthermore, the possible occurrence and identification of hard parts of second- ary prey (from the stomach of the prey of the marine mammal) could bias the results of scat or stomach analysis (Perrin et al. 1973). Additional information on the feeding habits of California sea lions can also be obtained from the weight estimates of the four major prey species iden- tified in this study. The 1978 and 1979 percentages of total weight estimates (Fig. 4) for each major species showed seasonal changes that are similar to the analysis of percentage of occurrence (Fig. 2), although there are a few exceptions. In 1979 the market squid weight estimate showed a slight de- crease, instead of an increase, from spring to sum- mer, however, of more importance, is its relationship to Pacific whiting. The estimated weight of market squid from the scats clearly exceeded the relative weight of Pacific whiting and other prey species con- sumed during the spring and summer of 1979. These results suggest market squid may be a more impor- tant food item than was predicted from the analysis of their percent of occurrence. The importance of the squid in the diet of the California sea lion during the summer months near the northern California Chan- nel Islands was also documented by Rutter et al. (1904), when they found that 84.6% (n = 13) of the animals examined had squid in their stomachs. Bailey and Ainley (1982) estimated the spring and summer percent (weight) of Pacific whiting in the California sea lion diet in the southern region to be within a range of 50 to 90%. Yet our estimates fell below 40% in the spring of 1979 and below 20% in the summer of both 1978 and 1979, and only one instance (spring 1978) did our estimates fall within the range suggested by Bailey and Ainley (1982). Since Bailey and Ainley (1982) based their estimates on data from California sea lions in the northern region, we assume our estimates more accurately represent the percent (weight) of Pacific whiting in the diet of California sea lions south of Point Concep- tion, and we recommend that additional feeding studies of California sea lions be conducted throughout their range. The percentage of estimated weight results also suggests that Pacific whiting was preyed upon more heavily in the spring of 1978 than in the spring of 1979. This is consistent with the exceptionally high recruitment of the 1977 year class of Pacific whiting (discussed above) which was available as yearlings to California sea lions in 1978. Although these weight (biomass) estimates are only approximate measurements, they appear instructive when used in conjunction with percentage-of- occurrence data. Unfortunately, there is some uncer- tainty as to the accuracy of using estimates of weight to estimate consumption. Our ability to make con- sumption estimates awaits the resolution of several questions: 1 ) What proportion of a given meal is repre- sented in a single scat? 2) Are there differential digestive rates of fish and squid? 3) Do sea lions of different ages and sexes digest food differently? The results of this study suggest that the California sea lions found on San Miguel Island feed oppor- tunistically on prey species of changing availability, and we agree with Bailey and Ainley (1982) that they are behaviorally flexible enough to switch from one major prey species to another, both seasonally and annually. This type of flexibility in foraging appears to be adaptive and may be a major factor contributing to the success of the California sea lion population off the coasts of California and Baja California. ACKNOWLEDGMENTS Permission to work on San Miguel Island was grant- ed by the National Park Service in conjunction with the U.S. Navy. Logistical assistance was provided by the U.S. Coast Guard, U.S. Navy, and the Channel Islands National Park. Superintendent W. Ehorn and his staff of the Chan- nel Islands National Park frequently assisted us dur- ing our research activities. Others who volunteered their time and assistance during our research included L. Antonelis, T. Antonelis, and K. Antonelis of Seattle, Wash.; P. Collins of Santa Barbara Museum of Natural History, Santa Barbara, Calif.; E. Jameyson of MARIS, Seattle, Wash.; R. Morrow of Oregon State University, Corvallis, Oreg.; D. Seagars of NMFS Southwest Region, Terminal Island, Calif.; and B. Steward of Hubbs-Sea World Research In- stitute of San Diego, Calif. M. Weber of California Marine Mammal Center, Fort Cronkhite, Calif., pro- 74 ANTONELIS ET AL.: SPRING AND SUMMER PREY OF CALIFORNIA SEA LIONS vided valuable assistance during the entire 1978 field season. Advice and aid during statistical analysis of the data were given by J. Breiwick, R. Kappenman, R. Ryel, and A. York of NMFS Northwest and Alaska Fisheries Center, Seattle, Wash. The otoliths were identified by the late J. Fitch of Long Beach, Calif., and the octopus beaks by F. E. Hochberg of Santa Barbara Museum of Natural His- tory, Santa Barbara, Calif. Additional consultation and assistance in our efforts to understand the forag- ing ecology of the California sea lion were given by K. Bailey of NMFS Northwest and Alaska Fisheries Center, Seattle, Wash; E. Knaggs, and K. Mais of California Fish and Game, Long Beach, Calif; A. Mearns and M. Perez of NMFS Northwest and Alaska Fisheries Center, Seattle, Wash.; and P. Smith and G. Stauffer of NMFS Southwest Fisheries Center, La Jolla, Calif. The senior author (Antonelis) presented a portion of this paper at the 4th Biennual Conference on the Biology of Marine Mammals in San Francisco, December 1981. LITERATURE CITED AlNLEY, D. G., H. R. HUBER, AND K. M. Bailey. 1982. Population fluctuations of California sea lions and the Pacific whiting fishery off central California. Fish. Bull., U.S. 80:253-258. Antonelis, G. A., Jr. and C. H. Fiscus. 1980. The pinnipeds of the California current. Calif. Coop. Oceanic Fish. Invest. Rep. 21:68-78. Bailey, K. M., and D. G. Ainley. 1982. The dynamics of California sea lion predation on Pacific whiting. Fish. Res. 1:163-176. Bailey, K. M., R. C. Francis, and P. R. Stevens. 1982. The life history and fishery of Pacific whiting, Merluc- cius productus. Calif. Coop. Oceanic Fish. Invest. Rep. 23:81-98. Best, E. A., and M. S. Oliphant. 1965. Evaluation of the fish resources of the Pt. Arguello Area. Marine Resources of the Pt. Arguello Area, Part II. Bur. Commer. Fish. Contract 14-17-001-995, as part of agreement AT (49-7)-2428 for the Div. Biol, and Med.. U.S. At. Energy Comm., 129 p Bowlby, C. E. 1981. Feeding behavior of pinnipeds in the Klamath River, northern California. M. A. Thesis, Humboldt State Univ., Eureka, Calif., 74 p. Briggs, K . T., and C. W. Davis. 1972. A study of predation by sea lions on salmon in Mon- terey Bay. Calif. Fish Game 58:37-43. Dark, T. A. 1975. Age and growth of Pacific hake, Merluccius produc- tus. Fish. Bull., U.S. 73:336-355. Everitt, R. D., P. J. Gearin, J. S. Skidmore, and R. L. DeLong. 1981. Prey items of harbor seals and California sea lions in Puget Sound, Washignton. Murrelet 62:83-86. Fields, W. G. 1965. The structure, development, food relations, reproduc- tion, and life history of the squid Loligo opalescens Berry. Calif. Dep. Fish Game, Fish Bull. 131:1- 108. Fienberg, S. E. 1977. The analysis of cross-classified categorical data. Mass. Inst. Technol. Press, Cambridge, Mass., 151 p. Fiscus, C. H., and G. A. Baines. 1966. Food and feeding behavior of Steller and California sea lions. J. Mammal. 47:195-200. FITCH, J. E., AND R. J. LAVENBERG. 1968. Deep-water teleostean fishes of California. Univ. Calif. Press, Berkeley, 155 p. Hollander, M., and D. A. Wolfe. 1973. Nonparametric Statistical Methods. John Wiley and Sons, Inc., N.Y., 503 p. Jameson, R. J., and K. W. Kenyon. 1977. Prey of sea lions in the Rogue River, Oregon. J. Mam- mal. 58:672. Jones, R. E. 1981. Food habits of smaller marine mammals from northern California. Calif. Acad. Sci. Proc. 42:409-433. Kashiwada, J., C. W. Recksiek, and K. A. Karpov. 1979. Beaks of the market squid, Loligo opalescens, as tools for predator studies. Calif. Coop. Oceanic Fish. Invest. Prog. Rep. 20:65-69. LE BOEUF, B. J., AND M. L. BONNELL. 1980. Pinnipeds of the California Islands: Abundance and Distribution. In D. M. Power (editor), The California Islands: Proceedings of a Multidisciplinary Symposium, p. 475-493. Santa Barbara Museum of Natural History, Santa Barbara, Calif. Mais. K. F. 1981. Age-composition changes in the anchovy, Engraulis mordax, central population. Calif. Coop. Oceanic Fish. Invest. Prog. Rep. 22:82-87. Miller, L .K. 1978. Energetics of the northern fur seal in relation to climatic and food resources of the Bering Sea. Mar. Mam- mal Comm., Wash., D.C., 32 p. (Available U.S. Dep. Com- mer., Natl. Tech. Inf. Serv., as PB-275296.) Morejohn, G. V., J. T. Harvey, and L. T Krasnow. 1978. The importance of Loligo opalescens in the food web of marine vertebrates in Monterey Bay, California. In C. W. Recksiek and H. W. Frey (editors), Biological, oceanographic and acoustic aspects of the market squid, Loligo opalescens. Berry. Calif. Dep. Fish Game, Fish Bull. 169:67-97. Morse, D. H. 1980. Behavioral mechanisms in ecology. Harvard Univ. Press, Cambridge, Mass., 383 p. MOYLE, P. B., AND J. J. CECH, Jr. 1982. Fishes: an introduction .to ichthyology. Prentice-Hall, Englewood Cliffs, N.J., 593 p. Pacific Fishery Management Council. 1978. Implementation of northern anchovy fishery manage- ment plan, Book 2. Fed. Reg. 43(141):31651-31879. Perrin, W. F., R. R. Warner, C. H. Fiscus, and D. B. Holts. 1973. Stomach contents of porpoise, Stenella spp., and yellowfin tuna, Thunnus albacares, in mixed-species aggregations. Fish. Bull., U.S. 71:1077-1092. Peterson, R. S., and G. A. Bartholomew. 1967. The natural history and behavior of the California sea lion. Am. Soc. Mammal., Spec. Publ. 1, 79 p. 75 FISHERY BULLETIN: VOL. 82, NO. 1 Phillips, -J. B. 1964. Life history studies on ten species of rockfish (genus Sebastodes). Calif. Dep. Fish Game, Fish Bull. 126, 70 P- Pitcher, k. W. 1980. Stomach contents and feces as indicators of harbor seal, Pkoca i itulina, foods in the Gulf of Alaska. Fish. Bull., U.S. 78:797-798. ROBINS, C. R. (chairman). i:i,sn. A list of common and scientific names of fishes from the United States and Canada. 4th ed. Am. Fish. Soc, Spec. Publ. 12, 174 p. Ri iter, ('., R. E. Snodgrass, and E. C. Starrs. 1904. Report on the sea lion investigation, 1901. In H. M. Smith, Report on the inquiry respecting food-fishes and the fishing-grounds, p. 116-119. U.S. Comm. Fish Fish., Part XXVin, Rep. Comm. 1902. SCHEFFER, V. B., AND J. A. NEFF. 1948. Food of California sea-lions. J. Mammal. 29:67-68. Spratt, J. D. 1975. Growth rate of the northern anchovy, Engraulis mor- dax, in southern California waters, calculated from otoliths. Calif. Fish Game 61:116-126. Young, R. E. 1972. The systematics and areal distribution of pelagic cephalopods from the seas off Southern Califor- nia. Smithson. Contrib. Zool. 97,159 p. 76 LARVAL DEVELOPMENT OF THE SCUP, STENOTOMUS CHRYSOPS (PISCES: SPARIDAE)1 Carolyn A. Griswold2 and Thomas W. McKenney3 ABSTRACT Larval scup, Stenotomus chrysops (Linnaeus 1766), were reared from eggs hatched in an aquarium. Measurements of morphological features for 88 specimens from 2.0 to 16.9 mm SL indicate that growth is gradual and continual with no well-defined changes in relative body proportions. Twenty-four myomeres are present in larvae, agreeing with published reports of vertebrae numbers in adult scup. Ossification begins first in the skulls of 6.1 mm SL larvae, and by 7.0 mm SL the vertebrae, neural spines, and fin rays are begin- ning to ossify. Ossification is nearly complete in 18.7 mm SL juveniles. Three preopercular spines are present in 4.1 mm SL specimens; the numbers of spines increase and by 16.9 mm SL the preopercular margin is serrate. Median fin development occurs at 4.1 mm SL, all fins are present in 8.8 mm SL larvae, and a full com- plement of rays are observed by 12.8 mm SL. Larvae are completely scaled by 13.0 mm SL. Scup, Stenotomus chrysops (Linnaeus 1766), the only common sparid in southern New England waters, is a popular sport and commercial fish in spring and sum- mer. Their range is from South Carolina to Sable Island, Nova Scotia, although they are uncommon north of Cape Cod (Breder 1948; Bigelow and Schroeder 1953; Leim and Scott 1966). Scup move inshore in schools in early April in the Chesapeake Bay area and in May north to Cape Cod. Most scup spend the summer in bays or within 8-10 km of the coast where they spawn from May to August with a peak in June in Narragansett Bay (Perlmutter 1939; Bigelow and Schroeder 1953; Wheatland 1956; Her- man 1963). In late October scup begin to move offshore to depths of 40-100 m. Commercial catches between January and April indicate that many scup winter off Virginia and North Carolina (Neville and Talbot 1964; Smith and Norcross 1968). Despite the commercial importance and abundance of this species, only one description of the eggs and larvae exists (Kuntz and Radcliffe 1917). This de- scription, which has been paraphrased several times, and the accompanying illustrations, which have been reprinted several times, provide no information on osteological development nor do they present meris- tic and morphometric data. Consequently we under- took to rear larvae from laboratory-spawned eggs to provide specimens for a more complete description which would be useful for identification of wild larvae. 'MARMAP Contribution MED/NEFC 81-03. 2Northeast Fisheries Center Narragansett Laboratory, National Marine Fisheries Serivce, NOAA, Narragansett, RI 02882. 'Northeast Fisheries Center Sandy Hook Laboratory, National Marine Fisheries Service, NOAA, Highlands, NJ 07732. METHODS Adult fish captured by trawl in Narragansett Bay, R.I., were held in a 58 m1 aquarium until they spawned naturally. Fertilized eggs were collected from the aquarium with plankton nets and incubated in 40 1 aquaria at 18° and 21C in 3 l%o salinity. Thepostin- cubation series for this study was reared at 18°C. After hatching, the larvae and juveniles were fed zooplankton and brine shrimp nauplii. Larvae were removed regularly for our studies and preserved in 4% buffered Formalin4 and Formalin with Ionol added as a color preservative. Specimens up to 19.5 mm standard length (SL) are included in this descrip- tion, but scup were reared to >40 mm in some of our experiments. Eighty-eight larvae from 2.0 to 16.9 mm SL were measured with an ocular micrometer. The data were pooled for all fish of the same SL, and all measurements converted into percentages of SL and summarized in Table 1. The following measurements were made: Total length (TL) : Tip of snout to end of caudal fin or finfold. Standard length (SL): Tip of snout to end of notochord in larvae prior to and during notochord flexure; tip of snout to base of hypural plate once it is formed. All references to length or size in the text refer to SL unless otherwise noted. Postanal length: Anus to end of notochord mea- sured along midline of body. Manuscript accepted June 1983. FISHERY BULLETIN: VOL. 82, NO. 1. 1984. 4Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 77 FISHERY BULLETIN: VOL. 82. NO. 1 TABLE 1.— Summary of nine morphological features of specimens of Stenotomus chrysops as shown by their percentages of standard length. Postanal Preanal Head Snout Prepectoral Prepelvic Eve Body SL TL length: length: length: length: length: length: diam. depth: (mm) (mm) SL:TL SL SL SL SL SL SL SL SL 2.0 2.1 95.2 450 40.0 10 0 2.5 — — 7.5 300 2 1 2.3 91.3 46.7 40.5 11.0 3 8 — — 7.7 29.0 2 2 : 4 91.7 466 42.7 1 2 5 3.4 — — 7.6 26.0 2 3 2 & 92.0 43.5 38.0 10.9 4.3 — — 7 6 25.0 2 8 92.9 42 .3 36 5 17.3 38 19.2 — 7.5 240 2.9 93.1 41.0 36.6 18.2 5.4 21.5 — 7.7 210 2 a 3.0 93 3 40.8 36.0 18.6 6 5 20.8 — 6.2 24.0 2 y 3 l 93 5 414 36.6 18.5 4.1 21.6 — 6.9 20.7 3 0 3 2 93.8 42.8 37 8 17 8 50 21.1 — 7.6 224 3 2 3.4 94 1 43.8 37 5 18 8 4.7 21.9 — 8 1 22 6 :•; 4 3.6 94.4 42.6 38.2 19.1 5 1 23 5 — 8 1 22.6 3 5 3 7 946 41 4 37.1 18.1 43 22.4 — 8 8 28.1 i 6 3.8 94.7 40 7 37 0 18.5 56 21 8 — 8.8 20.6 3 7 1 ■< 94 9 41 9 37 2 18.9 5 4 21.6 — 8 8 229 3 9 4 i 95 1 436 38.5 17.9 5 1 20.5 — 8 1 284 4.1 4 3 95 3 439 39.0 17 1 4 9 22.0 — 8 3 24.4 4 < 4.6 93 5 46.5 41.9 209 4 7 22.1 — 8 2 29 5 4 6 49 939 45.7 41.3 21 2 4.9 23.9 — 8 2 295 48 5 1 94 1 45.7 41.7 20.8 6 3 250 8.9 30.0 4 9 5.2 942 44 9 40.8 219 5.6 25 5 — 8.7 30.4 ci 4 5 6 964 51.9 48 1 22 2 56 25.0 — 9.5 23.8 5 & 59 932 51 9 46.7 23.9 7 3 27,6 — 10.9 382 5 6 62 90.3 48 2 44 6 232 5 4 268 — 9.5 224 6 1 69 884 53 0 48.1 25.1 7.7 30.1 9.8 29 5 >.4 7 4 86.5 56 3 500 25.0 9 4 29.7 — 10.0 25.8 66 6 6 868 54.5 50.0 25 8 7.6 31.8 — 9.8 30.3 7 1 8 0 88 8 54.9 507 25.4 7 0 31.0 — 99 21 1 ; q 9.2 859 53.2 494 24.7 7 6 29 7 — 9 6 25 3 8 6 10.3 82.5 57.1 52 9 253 7 1 28 8 35.3 9.4 27.1 9 2 10.9 84.4 55.4 51 1 26.1 7 6 29.3 380 92 27 2 9.9 11.8 839 54 5 50.5 242 9 1 32.3 33 3 9.1 27.3 10.0 12.0 83 3 58 5 52.5 27 5 9 0 30.5 37.0 9 5 260 12.0 14.5 828 56.7 508 258 9 2 30.0 35.8 9 2 26.7 12.6 154 81.8 55 6 500 262 8.7 31.0 34.1 8.7 25.4 13 1 156 84.0 565 51.9 28.2 8 4 32.1 35.9 10.7 30.5 135 16.3 82 8 563 52 6 21.5 8.1 304 34.1 89 296 14.6 17.0 85 9 54.8 50.0 25.3 8.9 32.2 35 6 9 6 26.0 14.9 17.1 87.1 54.4 48 3 25.5 8 1 30.9 34.2 8.7 26.8 15.9 18.8 84.6 59.1 54.7 25.2 8.2 31.4 384 7.5 32.7 169 20.6 82.0 62.1 59.2 30.2 11.2 35.5 42.6 10.7 33.7 'Notochord flexion Preanal length: Tip of snout to anus measured along midline of body. Head length: Tip of snout to posterior margin of otic capsules in young larvae; tip of snout to cleithrum once it is apparent. Eye diameter: Horizontal distance between anterior and posterior edges of orbit. Snout length: Tip of snout to anterior margin of eye. Body depth: Vertical height of body at pectoral axis. Prepectoral length: Tip of snout to axil of pectoral fin, or its anlage, measured along midline of body. Prepelvic length: Tip of snout to axil of pelvic fins, measured along midline of body. Meristics: Fin rays and spines were counted as they became apparent. Myomeres (total, precaudal, and caudal) were counted. Seventeen specimens were cleared and stained by Hollister's method (Hollister 1934) to determine the ossification sequence of 78 developing skeletal elements to verify counts of bony structures. DESCRIPTION Eggs The scup egg is spherical, buoyant, and transparent. The shell is unsculptured and the yolk unsegmented. It has one gold-colored oil globule that is posterior in the yolk sac and bears black pigment. The yolk is about 187( and the oil globule about 21% of the egg diameter. The average diameter of the 97 eggs we measured was 0.93 mm (range 0.81-1.00 mm). They hatched in 70-75 h at 18°C and in 44-54 h at 21°C. These measurements and the incubation time are similar to those found by others for this species (Kuntz and Radcliffe 1917; Perlmutter 1939; Bigelow and Schroeder 1953; Wheatland 1956). GRISWOLD and McKENNEY: LARVAL DEVELOPMENT OF SCUP Larvae Newly hatched larvae average 2.0 mm SL. The eyes are not pigmented and the mouth is not functional. The head is bent slightly over the elliptical yolk sac. Yolk sac and oil globule are absorbed and gut dif- ferentiation occurs between 48 and 72 h after hatch- ing at 18°C. During this period the eyes become pigmented, the mouth functional, and the larvae begin to feed. Larvae ranging in size from 2.0 to 18.7 mm are shown in Figure 1. Yolk Absorption and Gut Differentiation At hatching the gut is a tube with a constriction at its posterior end that extends to the ventral edge of the finfold, but by 48 h (2.7 mm) a foregut and hindgut A DAY I 2.0 B DAY 4 2.8 C DAY 5 3.0 F DAY 13 5.7 G DAY 15 7.3 H DAY 17 9.4 D DAY 6 3.4 I DAY 21 14.9 E DAY 9 4.2 J DAY 24 18.7 FIGURE l.— Development of Stenotomus chrysops. Lengths (SL) are in millimeters. 79 FISHERY BULLETIN: VOL. 82, NO. 1 can be distinguished. The hindgut appears to be muscular and remains a tube until between day 7 and day 9 (ca. 4.0 mm), when a well-defined stomach becomes apparent and the hindgut is relatively shorter. Total Length and Standard Length Larval growth appears to be gradual and continuous with no well-defined changes in relative body propor- tions. Apparent slight changes which are noticeable after notochord flexion relate to a change in measure- ment from an SL which is actually notochord length to one which is a true SL. Snout Length As with eye diameter, there is considerable varia- tion among individuals of the same size. At hatching the snout length is 2.5% of SL, but this increases gradually to 9.4% of SL at 15.9 mm and 11.2% of SL in the juveniles. Body Depth Body depth ranges from 25 to 30% in newly hatched larvae, but once the yolk is absorbed it decreases to between 21 and 24.4% of SL (with one exception) up to 3.9 mm, and then increases to 22.4 to 33.7% of SL. Postanal Length Postanal length remains about 45% from 2.0 to 4.9 mm SL, when notochord flexion is occurring. A gradual increase to 62.1% in juveniles longer than 16.9 mm SL is concurrent with development of ver- tebrae and overall growth of the larvae. Preanal Length Preanal length increases relative to SL from 36.0f/f at 2.0 mm to 41.9% at 4.9 mm to >59% for juveniles longer than 16.9 mm. The lengthening of the body cavity during growth accounts for the increase in pre- anal length. Head Length Head length increases relative to SL from an average of 11.1% (10-12.5%) in newly hatched larvae (2.0-2.3 mm) to 17.3% in 2.6 mm larvae, then gradually increases to 30.2% in the largest juvenile specimen. In very young larvae the otic capsules are the reference structure for head measurements. However, once the cleithrum develops it is used as the reference structure for subsequent head measurements and an increase in head length per- centage is observed. Eye Diameter The ratio of eye diameter to SL in our series is 6.2- 10.9% of SL. It averages 7.5% SL in 2.0-3.0 mm lar- vae, and 8.5% SL (range 8.1-8.9%) in 3.2-4.9 mm larvae. In larvae >4.9 mm the average is 9.5% of SL (range 7.5-10.9%). Variation in individuals of the same size is considerable. Prepectoral Length Anlagen are present at hatching. Initially prepec- toral length is about 19.2% of SL. This increases gradually during the larval and postlarval period to 35.5% of SL in the juvenile. Prepelvic Length Pelvic fin buds do not appear until the larvae are 8.0-8.5 mm long. Prepelvic to SL ratio is about 35.6% (range 33.3-38.4%) for larvae from 8.5 to 15.9 mm SL, but increases to 42.6% of SL in the juvenile. MERISTICS Scup, being typical of most perciform fish, have 24 myomeres. This agrees with Miller and Jorgenson's (1973) vertebrae numbers for adult fish. FIN DEVELOPMENT At hatching a finfold extends from the top of the head to the visceral sac interrupted only by the anus. There are no fin rays. A remnant of this persists be- tween the anus and the first anal fin ray in a larva 9.1 mm long. Fin sequence development is given in Table 2. Anlagen of the pectoral fins are present in most, if not all, hatchlings. These are low buds at first, but by the time the larvae are about 2.5 mm these fins have bases and blades. By removing pectoral fins from one side of some of our larvae and flattening them out, we could see 13 rays in one 4.9 mm larva and 10 rays in a 5.7 mm larva. Aside from these two, however, we could not see pectoral fin rays, even on cleared and 80 GRISWOLD and McKENNEY: LARVAL DEVELOPMENT OK SCUP TABLE 2. — Summary of fin development sequence in larvae of Stenotomus chrysops. Notochord or standard length (mm) Buds Rays Full Number of rays first first complement in fully Fin appear appear of rays developed fin Dorsal 5.5-6.0 10.8 XII + 12 Caudal 10.4-108 32-34 Principal 43 5.3 Dorsal 9 Ventral 8 Secondary 53 Dorsal 7-8 Ventral 8-9 Anal 5.5-6.0 108 III + 1 1 Pelvic 5 7 8 8-10.0 12.8-13.2 I + 5 Pectoral 2 3 2 9-3.0 10.4-10.8 16 stained specimens until the larvae were about 8.0 mm, when the larvae had nearly the full complement of 15-16 pectoral rays. An anlage of the caudal base can be seen in larvae as small as 3.4 mm. Some of the principal caudal rays are detectable in the finfold of larvae as small as 4.3 mm, and are the first rays of any fins to appear. Notochord flexion in our series begins at 4.7 mm. By 5.3 mm all of the principal caudal rays are present as are some of the secondary ones. Flexion is complete at about 8.0 mm and the caudal fin begins to fork at about 10 mm. Full complements of caudal rays (9-10+9+8+8-10) are present in larvae 14 mm or longer. Secondary rays develop in a posterior to anterior direction. The soft-ray parts of the median fins first develop beginning at 5.3 mm in our series. Anal and dorsal rays develop together. In both fins, the central soft rays develop first. The development of anterior and posterior rays follows rapidly so that when the larvae in our series are >6.0 mm, full complements of 11-12 soft rays are present in these fins. Development of the spiny rays in these fins is from posterior to anterior and follows the soft-ray development. An exception is the posteriormost spiny rays in both fins that appear first as soft rays. The last fins to appear are the pelvics. Anlagen are first seen in our series in some larvae at 5.7 mm. Other larvae are >7.0 mm long before these anlagen are visible. Development thereafter is from the distal edge medially. Full arrays of 1 spine and 5 soft rays are present in larvae 8.5 mm or longer. Adult scup have six pairs of branchiostegals. Five pairs of these are present in a 4.2 mm larva of our series. They were visible in all of our series that were 5.0 mm or larger. The sixth pair, the median one, is not visible in some of our larvae even at 16.5 mm. The first five pairs usually appear simultaneously, but the sixth appears later. PIGMENT Although scup have chromatophores other than melanophores, these faded rapidly after preserva- tion in Formalin. This account is confined to melanophore pigmentation (Fig. 1). Pigmentation other than that by melanophores is extensive on embryos and early larvae and is described and illus- trated by Kuntz and Radcliffe (1917). Head Region Newly hatched scup have unpigmented eyes. Two rows of stellate melanophores, one on either side, extend from the snout back over the eyes and con- tinue as part of a lateral series on the trunk. At a length of about 2.5 mm there is a hiatus in this series that extends from mideye level to over the visceral sac. At 4.0-5.0 mm length, the pattern that will culminate in that of the juvenile has begun to appear. There are few, usually no, melanophores on the dor- sal and lateral parts of the head anterior to the middle of the eyes. However, there are several prominent melanophores on the posterior midbrain and several on the hindbrain. Ventrally there is usually no pig- ment on the head. A few of our specimens have one or two small melanophores. Development beyond this stage consists of a gradual increase of pigment on the dorsal and dor- solateral parts of the head. Most of it occurs above mideye level. A few melanophores appear on the snout and below the eye. There is a prominent melanophore, sometimes accompanied by one or two small ones as well, at the articulation of the lower jaw with the quadrate bone. Between the head and the trunk pigmentation in the occipital region there is a gap in the dorsal pigment with relatively little pigment in it . This gap is part of the barred pattern of the juvenile. Trunk and Tail Region At hatching there are two dorsal rows of stellate melanophores extending from the head to beyond myomere 20. They appear to be between myomeres on the myosepta. Occasionally these rows are interrupted by "missing" melanophores. When this is so, the melanophore is usally lower down on the side of the body. A few, usually three or four, melanophores occur at various places on the anterior part of the yolk sac, and there are one or two on the oil globule. Some specimens have widely spaced melanophores on the ventral margin of the tail. 81 FISHERY BULLETIN: VOL. 82, NO. 1 This pattern persists until the larvae are about 2.5 mm long with a gradual increase in the number of melanophores along the ventral margin of the tail. The melanophores on the yolk sac and oil globule dis- appear with the exception of one or two on the mid- ventral line of the anterior part of the yolk sac. In larvae 4.0-5.0 mm long, most of the pigment is on the peritoneum dorsal to the viscera and along the midventral line. Anteriorly there is a melanophore at the cleithral symphysis and posterior to it a large one midventrally on the anterior belly and a smaller one on the posterior belly. There is a prominent melanophore on the hindgut just anterior to the anus. Posterior to this there is a melanophore on most of the anal pterygiophores; this pattern is continued externally on the ventral myosepta. Dorsally there are several melanophores on the posterior pterygiophores of the dorsal fin. There are usually a few scattered spots on the finfold and, on some specimens, a few on the sides. The peritoneal pigment becomes denser and more prominent in 9-10 mm larvae, however, it is often obscured because of the opacity of the thickening body musculature in preserved specimens. The hindgut is nearly covered by large melanophores. The melanophores on the midventral line are still present, usually accompanied by two or three smaller ones. The melanophore just posterior to the anus is still present, but less prominent. The trunk and tail pigment is more extensive at 9- 1 0 mm. There are many more pigment spots along the sides, but these are still widely spaced, especially anteriorly. There is pigment along the bases of the dorsal and anal fins that continues posterior to them to the procurrent caudal rays. A line of melanophores runs dorsoventrally at about the juncture of the caudal fin rays and caudal bones. Internally there are melanophores near the bases of the haemal and neural arches. These become increasingly obscure as the body musculature thickens. Pigment development beyond this size is charac- terized by the development of the barred pattern of the juvenile accompanied by a general increase in pigment everywhere, especially above the mid- lateral line. OSSIFICATION A total of 17 fish were stained with Alizarin Red to determine where ossification began and the sequence in which the bones ossified. A summary of osteological development is presented in Table 3. There is no dye uptake in 5.2 or 6.0 mm larvae, although the cartilaginous skeleton is easily dis- tinguished. Cartilaginous hypural plates are present in larvae undergoing notochord flexure (5.4-5.6 mm). The first ossification occurs in skulls of 6.1 mm lar- vae. The premaxillary, maxillary, dentary, articular, and quadrate bones associated with the jaws, the pre- operculum, hyomandibular, branchiostegal rays, and cleithrum showed varying degrees of dye uptake. There is no ossification posterior to the cleithrum. By 7.0 mm more ossification of the skull occurs, notably the pterygoid, metapterygoid, opercular series, supracleithrum, and frontal bones. The cir- cumorbitals, as well as the parasphenoid and the scapula, show the beginning of dye uptake. Ossifica- tion has begun in the first 10 vertebrae, the neural spines of the first 4 vertebrae, and the pectoral and caudal fin rays. In 9.3 mm specimens the skull is further developed; teeth are visible and the lachrymal, dermethmoid, nasal, prefrontal, and urohyal bones show varying degrees of ossification. The postcleithrum is well developed and the radials, scapula, and coracoid are ossifying. The entire vertebral column is ossified with the exception of the ultral centrum and penultimate vertebrae. Both haemal and neural spines are ossified. The pleural ribs, and the dorsal and anal fin rays are beginning to ossify; hypural plates and caudal fin rays are partially ossified. By 10.8 mm the distal vertebrae (caudal complex) have ossified and scales are present. The pelvic fin supports and rays show some dye uptake. Pterygio- phores are present as cartilage. All the dorsal and anal fin rays and spines have ossified. Skull development and ossification of most of the bones of the skull is complete by 14.5 mm. The radials and scapula which were just beginning to ossify in the 10.8 mm fish are now complete. Pelvic fin supports are complete. The pleural ribs are TABLE 3. — Summary of osteological development in laboratory reared larvae of Stenotomus chryxops. Notochord or standard length (i nm) First First evidence appearance of ossification All in cartilage (stain uptake) ossifying Cartilaginous skeleton 5.2 Hypural plates 5.4 93 14.5 Vertebrae and neural spines 7.0 10.8 Pectoral girdle 7,0 108 14 5 Pectoral and caudal fin rays 7.0 9.3 Haemal spines 93 93 Pleural ribs 93 14.5 Dorsal and anal fin rays 93 10.0 Caudal complex 108 10.8 Pelvic fin supports 108 14.5 Pelvic girdle 108 18.7 82 GRISWOLD and McKENNEY: LARVAL DEVELOPMENT OF SCUP stained as are the pterygiophores. The hypural plates are all present and completely ossified; a few dorsal plates are still partially cartilaginous. By the time scup are 18-19 mm long, they are juveniles. Ossification continues in the skull with the bones being joined at suture points; the pelvic girdle is complete. The pterygiophores and ribs have com- pleted ossification. PREOPERCULAR SPINES Figure 2 shows the development of preopercular spines. We saw them first on a 4.1 mm specimen, which has three spines on the preopercular margin. Thereafter their number increases until there are so many on a 16.9 mm specimen that the margin is serrate. Specimens larger than about 25 mm have nearly smooth preopercular margins. and Schroeder 1928) placed S. aculeatus in the synonomy of S. chrysops; Robins et al. (1980) did not list S. aculeatus. Dahlberg (1975) mentioned young stages of S. chrysops with crossbars (i.e., juveniles) in his account of Georgia coastal fishes, although it is not clear whether he had taken such specimens in his collections. This issue is further complicated by lack of informa- tion about the northern extent of spawning of other sparid fishes. If their spawning ranges overlap with that of S. chrysops, then the younger larvae of some species will probably be confused with scup larvae, at least until the dorsal, anal, and pectoral fin rays can be counted. Except for the reference to juvenile scup on the Georgia coast by Dahlberg, the authors can find no references to such an overlap. We have seen larval scup in collections misiden- tified as Scomber scombrus, the Atlantic mackerel, from which they can be separated at all stages by the numbers of myomeres (24 in scup and 31 in mack- erel). We have also seen larval gerreid fishes misiden- tified as scup. Among other characters, scup differ from gerreid fishes in lacking the long premaxillary spines that extend up between the eyes in gerreids. Figure 2.— Development of the preopercular spines of Stenotomus chrysops. Standard lengths in millimeters of the specimens are A) 4.1, B) 5.6, C) 8.3, D) 9.8, and E) 16.9. ACKNOWLEDGMENTS We are grateful to John Colton, Bernard Skud, Donna Busch, Wallace Smith, and Michael Fahay for their reviews of the manuscript. We thank Jennie Dunnington and Maureen Montone for typing and retyping the manuscript. We are particularly indebt- ed to Lianne Armstrong who prepared Figure 1 and Alyce Wells for help with Figure 2. SCALES The first scales are seen between 9.9 and 10.8 mm. At 12.3-13.0 mm the larvae are completely scaled. COMPARISONS The geographical extent of spawning of S. chrysops is not known. The authors can find no record of it spawning south of the New York Bight. At least one other species of Stenotomus, S. caprinus (Bean 1882), occurs in the western North Atlantic. Accord- ing to Geohagen and Chittenden (1982), the major population of this species is in the northern Gulf of Mexico and it occurs only rarely along the east coast to North Carolina. A third nominal species, S. aculeatus (Valenciennes 1830), said to replace S. chrysops south of Cape Hatteras, is of doubtful validity. Birdsong and Musik (in 1977 reprint of Hildebrand LITERATURE CITED BlGELOW, H. B., AND W. C. SCHROEDER. 1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 53, 577 p. Bremer, C. ML, Jr 1948. Field book of marine fishes of the Atlantic coast from Labrador to Texas. G. P. Putnam's Sons, N.Y., 349 p. Dahlberg, M. D. 1975. Guide to coastal fishes of Georgia and nearby states. Univ. Georgia Press, Athens, 186 p. Geohagen, P., and M. E. Chittenden, Jr. 1982. Reproduction, movements, and population dynamics of the longspine progy, Stenotomus caprinus. Fish. Bull., U.S. 80:523-540. Herman, S. S. 1963. Planktonic fish eggs and larvae of Narragansett Bay. Limnol. Oceanogr. 8:103-109. Hildebrand, S. F., and W. C. Schroeder. 1928. Fishes of Chesapeake Bay. T. F. H. Publications, Inc., Neptune, N.J., 388 p. (1 972 reprint with comments by R. S. Birdson and J. S. Musik.] S3 FISHERY BULLETIN: VOL. 82, NO. 1 HOLLISTER, G. 1934. Clearing and dyeing fish for bone study. Zoologica (N.Y.) 12:89-101. KlNTZ, A.. AND L. RADCLIFFE. 1917. Notes on the embryology and larval development of twelve teleostean fishes. Bull. U.S. Bur. Fish. 35:89- 134. Leim, a. H.. am) W. B. Scott. 1966. Fishes of the Atlantic coast of Canada. Fish. Res. Hoard Can. Bull. 155.485 p. Miller. G. L., and s C. Jorgenson. 1973. Meristic characters of some marine fishes of the west- ern Atlantic Ocean. Fish. Bull., U.S. 71:301-312. Neville, W. C, and G. B. Talbot. 1964. The fishery for scup with special reference to fluc- tuations in yield and their causes. LIS. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 459, 61 p. Perlmitter, A. 1939. A biological survey of the salt waters of Long Island. Section I. An ecological survey of young fish and eggs iden- tified from tow-net collections. Suppl. 28th Annu. Rep. N.Y. Cons. Dep., Par! 11:1 1-71. Robins, C. R., R. M. Bailey, C. E. Bond, J. R. Brooker, E. A. Lachner, R. N. Lea, and W. B. Scott. 1980. A list of common and scientific names of fishes from the United States and Canada. 4th ed. Am. Fish. Soc. Spec. Publ. 12, 174 p. Smith. W. G., and J. J. Norcross. 1968. The status of the scup (Stenotomus chrysops) in winter trawl fishery. Chesapeake Sci. 9:207-216. Wheatland, S. 1956. Oceanography of Long Island Sound, 1952-1954. VII. Pelagic fish eggs and larvae. Bull. Bingham Oceanogr. Collect., Yale Univ. 15:234-314. 84 DESCRIPTION OF EGGS, LARVAE, AND EARLY JUVENILES OF GULF MENHADEN, BREVOORTIA PATRONUS, AND COMPARISONS WITH ATLANTIC MENHADEN, B. TYRANNUS, AND YELLOWFIN MENHADEN, B. SMITHI ] William F. Hettler2 ABSTRACT Morphometric, merist ic, and pigmentation descriptions of laboratory-reared gulf menhaden, Brevoortia pa- tronus, and Atlantic menhaden.fi. tyrannus, indicate that larvae of these species can be distinguished from each other by the number of myomeres and vertebrae; that Atlantic menhaden can be distinguished from yellowfin menhaden, B. smithi, by the number of myomeres and vertebrae, by pigmentation, and by morphometries; and that gulf menhaden can be separated from yellowfin menhaden by pigmentation and morphometries. Unlike yellowfin menhaden, gulf and Atlantic menhaden lacked paired melanophores along the dorsal midline forward of the dorsal fin and along the ventral midline between the paired fins. Compared with yellowfin menhaden larvae of equal lengths, gulf menhaden had less body depth, shorter heads and snouts, smaller eyes, and longer prepelvic and predorsal distances. Gulf menhaden eggs averaged 1.29 mm in total diameter, 0.95 mm in yolk diameter, and 0.20 mm in oil droplet diameter. Twelve-hour-old larvae had a snout-notochord tip length of 3.3 mm. Their growth rate averaged 0.30 mm/day through 90 days of rearing at 20°C. On specimens 6-17 mm the mean number of myomeres was 44.6; on specimens >15 mm the mean number of vertebrae was 45.3. Postdorsal-preanal myomeres decreased from 5.3 to 1.8 as the dorsal fin grew and the gut shortened during development. Transformation from larva to juvenile in laboratory-reared gulf menhaden was completed at a smaller size than reported for field-caught fish (25 vs. 28 mm SL). Eggs and larvae of gulf menhaden, Brevoortia pa- tronus Goode, have not been described, even though this species is the most economically important clupeid in the United States. The gulf menhaden purse seine fishery landed an average of 660,368 t annually from 1977 to 1981, making it the largest fishery in the United States (U.S. National Marine Fisheries Service 1982). Gulf menhaden, one of three species of Brevoortia in the Gulf of Mexico, are found from Florida Bay to the Gulf of Campeche, Mexico. They spawn in the northern gulf at least as far offshore as the 80 m isobath between mid-October and late March, with a peak in December (Christmas and Waller 1975'); juveniles are estuarine depen- dent. Yellowfin menhaden, B. smithi, and finescale menhaden, B. gunteri, co-occur with gulf menhaden, but contribute <1% to the landings. The Atlantic menhaden, B . tyrannus , which supports a large purse seine fishery along the U.S. Atlantic coast, is a large- ■Contribution No. 83-33B of the Southeast Fisheries Center, Beaufort Laboratory, National Marine Fisheries Service, NOAA. -Southeast Fisheries Center Beaufort Laboratory, National Marine Fisheries Service, NOAA, Beaufort, NC 28516-9722. 'Christmas, J. Y., and R.S. Waller. 1975. Location and time of menhaden spawning in the Gulf of Mexico. Unpubl. manuscr., 20 p. Gulf Coast Research Laboratory, Ocean Springs, MS 39564. Manuscript accepted July 1983. FISHERY BULLETIN: VOL. 82, NO. 1, 1984. scaled cognate of the gulf menhaden, but does not occur in the Gulf of Mexico (Hildebrand 1963). Distribution of yellowfin menhaden is contin- uous around Florida to as far north as North Caro- lina. Menhaden larvae superfically resemble the larvae of other clupeids with which they co-occur and can be distinguished from them (Houde and Fore 1973; Houde and Swanson 1975), but current descriptions (Suttkus 1956; Houde and Fore 1973; Houde and Swanson 1975; Jones et al. 1978) are not adequate to separate sympatric Brevoortia larvae. Eggs, larvae, and juveniles of yellowfin menhaden have been de- scribed (Houde and Swanson 1975), whereas the early development of finescale menhaden has not. Gulf and yellowfin menhaden hybrids in the eastern Gulf of Mexico (Hettler 1968; Turner 1969; Dahlberg 1970) further complicate separation by species. Although gulf and Atlantic menhaden larvae cannot be confused in ichthyoplankton collections because of their allopatric separation by the Florida Peninsula, Atlantic and yellowfin menhaden larvae may be confused in collections from the east coast of Florida, where both species are known to spawn dur- ing the winter (Dahlberg 1970). In this paper, I describe the eggs, larvae, and early 85 FISHERY BULLETIN: VOL. 82. NO. 1 juveniles of gulf menhaden spawned and reared in the laboratory using morphometries, meristics, and pigmentation features, and I compare gulf menhaden larvae with yellowfin menhaden larvae described by Houde and Swanson (1975). Morphometric and meristic data on laboratory-spawned and reared Atlantic menhaden are also presented to supplement the composite description of this species by -Jones et al. (1978) and to aid in the separation of Atlantic menhaden and yellowfin menhaden larvae. Charac- ters for separating Brevoortia from other clupeids are reviewed. cleithra, exclusive of the finfold. Dorsal and anal fin base lengths — distance from anterior to posterior edges of fin base; in larvae with incomplete fins, distance from origin of first ray to the insertion of the last ray. Head length— tip of snout to posterior margin of otic capsules in yolk-sac larvae; tip of snout to opercular margin in older larvae and juveniles. Snout length — tip of snout to anterior margin of eye. Eye diameter — horizontal distance between anterior and posterior edges of fleshy orbit. METHODS Gulf menhaden were collected as mature adults in September 1981 near Gulf Breeze, Fla., tranported to the Beaufort Laboratory, and induced to spawn with human chorionic gonadotropin (HCG) and carp pituitary (Hettler 1983). Spawnings that occurred in November 1981 and February 1982 provided a developmental series of eggs, larvae, and juveniles up to 90 d old, reared at a temperature of 20° ± 2°C and a salinity of 30%o. One hundred eggs, preserved during the early embryo stage, and 100 live eggs were measured. Atlantic menhaden were captured as juveniles in September 1978 near Beaufort, N.C., and reared to sexual maturity in the laboratory for 19 mo. They were induced to spawn in April 1980, and the larvae were reared at temperatures that began at 15°C and increased to 25°C during development (Hettler 1981). This spawning resulted in a developmental series of larvae and juveniles up to 130 d old. All specimens were preserved in 2% buffered for- maldehyde in seawater before being measured. The following morphometic measurements were taken with an ocular micrometer in a dissecting microscope on 123 gulf menhaden and 196 Atlantic menha- den. Standard length (SL) — tip of snout to tip of notochord before and during notchord flexion; in postflexion larvae, tip of snout to posterior margin of hypural bones. All references to length in this paper are standard length unless otherwise stated. Preanus length — tip of snout to posterior end of anus, measured along midline. Predorsal length — tip of snout to anterior edge of dorsal fin base, measured along midline. Prepelvic length — tip of snout to anterior insertion of pelvic fin, measured along midline. Body depth — vertical depth at symphysis of the Myomeres were counted on semidry specimens (not completely immersed) up to 1 7 mm with transmitted unpolarized light by adjusting the microscope mirror to give maximum contrast between myosepta and myomeres. Myomeres were classified as follows: Total myomeres — all myomeres between the most anterior myoseptum and the most posterior myoseptum. Preanal myomeres — number anterior to the myo- mere in which the anterior ray of the anal fin is inserted or to the myomere in contact with the downward curve of the dorsal margin of the anus in larvae without anal fin rays. Postanal myomeres — number posterior to the anterior insertion of the anal fin. Predorsal myomeres — number anterior to the myomere containing the origin of the first dorsal fin ray. Postdorsal-preanal myomeres — number between the myomere connected to the last dorsal fin ray and the most posterior preanal myomere. Following morphometric measurements on all specimens and myomere counts on specimens with visible myomeres, the pigment pattern was recorded and specimens of gulf menhaden were illustrated with a camera lucida. Atlantic menhaden were not illustrated as the figures in Jones et al. (1978) are adequate. Specimens were then used for counts of fin rays, pterygiophores, predorsal bones, vertebrae, and scutes. Specimens were transferred to 95 /f ethanol, stained with alcian blue for cartilage, cleared with trypsin, stained with alizarin red S for bone, and stored in 100% glycerin4. "Taylor, W. R., and G. C. Van Dyke. 1978. Staining and clearing small vertebrates for hone and cartilage study. Unpubl. manuscr., 19 p. National Museum of Natural History, Washington, DC 20560. 86 HETTLER: DESCRIPTION OF Ol'LF MENHADEN DESCRIPTION Embryos Gulf menhaden eggs were spherical, and had an unsculptured chorion, a faintly segmented yolk, and a single oil droplet. Living eggs were buoyant in salinities >26%o. Twenty-seven percent had both an outer and inner chorionic membrane. This has not been reported in wild-caught Brevoortia eggs. This inner chorion was not an artifact of preservation, since live eggs also contained a double chorion, but may have been a result of induced ovulation by HCG and carp pituitary. Dimensions of preserved and live eggs were the same as maximum sizes given by Houde and Fore (1973) for gulf menhaden eggs taken in plankton collections (Table 1). At its widest point the perivitelline space was 24-28% of the egg diameter. Eggs produced during December 1982 by another spawning group of gulf menhaden were smaller than gulf menhaden eggs produced the year TABLEl. Mean diameter (mm) of j-cull menhaden, Hrevnnrtia pa- tronus, eggs. Numbers in parentheses are equal to one standard de\ iation oi the mean. Eggs Total diameter Inner chorion diameter (if present) Yolk diameter (along axis) Oil droplet diameter Preserved 100 1.29(0.04) Live 100 1.30(0.05) 1.23 (0.04) 1.25 (0.03) 0.95 10.05) 0 97 (0.04) 0.20(0.02) 0.19 (0.01) before; total diameter was 1.18-1.22 mm; the yolk diameter was 0.66-0.79 mm; the oil droplet was 0.16 mm. The adults producing these eggs were smaller (17.8 cm mean length, 90 g mean weight) than the spawners that produced the larger eggs (20 cm, 135 g) (Table 1). Small adult size may be responsible for the small eggs as well as the reduced fecundity. Only a few hundred fertilized eggs were collected from the December 1982 group of 20 fish. Advanced embryos had 30-40 small melanophores on each side along the dorsal surface from the pos- terior end of the head to the notochord tip (Fig. 1 A). 1 mm V^frA FIGURE 1.— Early stages oiBrevoortia patronus. A. Embryo 40 h after fertilization. B. 2.6 mm larva, 5 min after hatching, e. 3.5 mm larva, 1 d after hatching. D. 3.9 mm larva, 2 d after hatching. 87 FISHERY BULLETIN: VOL. 82, NO. 1 About 15-20 myomeres were visible in the caudal region. The yolk was faintly segmented into irregular globules. E ggs hatched in 40-42 h at a water tempera- ture of 19°-20°C. Atlantic menhaden eggs spawned in the laboratory were larger than gulf menhaden eggs in total diameter (1.54- 1.64 mm) but similar in yolk diameter (0.82-0.95 mm) and oil droplet diameter (0.20- 0.23). Larvae Growth Gulf menhaden larvae were 2.6-3.0 mm SL immediately after hatching (Fig. IB), but within 6 h had a mean length of 3.3 mm. The yolk and oil droplet were absorbed, the eyes were pigmented, and the mouth was functional at a length of 4.5 mm, 4 d after hatching. The growth rate of larvae at 20° ± 2°C averaged 0.30 ± 0.03 mm/d through 90 d of rearing (Fig. 2). Yellowfin menhaden reared for 32 d at 20°C grew 0.36 mm/d (Hettler 1970). Yellowfin menhaden reared at 26°C grew 0.45 mm/d until the 20th day (Houde and Swanson 1975). Body Proportions For 123 gulf menhaden, 3.1-34.9 mm, body depth, head length, prepelvic length, dorsal fin base length, anal fin base length, snout length, and eye diameter all increased relative to standard length as larvae grew, while preanus length and predorsal length de- 36r 32 28 e I | 20 LLI _l D 16 DC < D I12 co 8 ./: 0 10 20 30 40 50 60 70 80 90 DAYS AFTER HATCHING FlOlRE 2. — Growth of laboratory-reared larvae of Hrcvaartki pa- t ran us. Lines connect means of each age group. creased (Table 2). The decrease in predorsal length resulted from the forward movement of the dorsal fin, and the decrease in preanus length reflected the transformation from an elongate clupeiform larva shape to the laterally flattened fusiform shape of the juvenile. Transformation from the larval to the juvenile form in gulf menhaden began at about 19 mm (Fig. 3C) and was completed at about 25 mm. Atlantic menhaden larvae completed transformation at about 27 mm. TABLE 2. — Proportions of head and body parts of gulf menhaden, Brevoortia patronus, expressed as a percent of stan- dard length. Characters were not developed at lengths marked with a dash. Length class Number of Preanus Predorsal Prepelvic Body Dorsal fin Anal fin Head Snout Eye (mm, SL) specimens length length length depth base length base length length length diameter 30-3.9 ( 840 _ — 14 1 2 3 5,2 4.0-4.9 19 8<> 2 — — 9.7 — — 13,5 1.7 5,4 5.0-5.9 12 81.4 — — 9 6 — — 15.8 3.0 5,2 6.0-6.9 B 82.4 69.3 — 8 4 4 4 — 155 3.1 5,0 7.0-7.9 7 83 0 70.2 — 8 2 5 0 — 15.4 2 9 4,9 8.0-8.9 4 832 67 6 — 7 9 8 2 _ 15 5 3 1 4 8 9.0-9.9 5 839 658 — 8 3 10.0 3 8 16.2 3 6 50 10.0-10.9 6 85 5 656 — 8 3 1 15 4 3 169 3 7 5.0 11.0-11.9 1 85 5 652 — 8 9 13 2 5.6 17 7 3 9 5 2 120-12 9 2 83 .1 63.0 — 8 6 13 3 6.0 16.9 3,6 4 8 130-13.9 3 84 2 62 8 — 9,9 15 1 68 17.8 3.7 4.9 14.0-14 9 i 81.0 620 41.5 100 14 5 7 5 17 0 3 5 50 15,0-15.9 i 82.2 61 2 — 10 7 15 4 7,5 18.2 3.7 5.1 160-169 4 79.8 60.8 44 2 108 15 3 94 18.9 40 5 5 17 0-17 9 3 79.0 61 0 44 3 12 4 14 8 11.0 19.4 4 1 5.5 180-189 2 760 570 47 6 18.1 16,8 12,3 24 1 5.0 7 3 19 0-19 9 4 762 56.4 469 17 8 16,6 12 8 240 5.1 6 8 200-21 9 B 71 4 51.8 50.0 25.8 186 16,0 28 1 60 8.1 22.0-23.9 8 70.2 48.6 49.2 280 18 4 15.7 28 9 6,1 84 240-25.9 6 70 7 47,9 49,6 29,1 193 16,0 29 3 6.6 8 5 26.0-279 3 70.6 44,7 50.0 31.6 19 2 17 0 29 7 6 9 8 6 280-29.9 1 70.2 43.5 49.4 30 1 20 1 16,4 27 7 64 7.7 30.0-34 9 7 72 7 47.7 51 3 36.0 19 4 17,1 31 5 7.2 7 8 88 HETTLER: DESCRIPTION OF GULF MENHADEN 2mm FIGURE 3.— Larval Brevoortia patronus: (A) 13.0 mm (28 d after hatching). (B) 16.5 mm (44 d after hatching). (C) 18.9 mm (53 d after hatching). Gulf menhaden larvae and Atlantic menhaden lar- vae could not be separated morphometrically (Table 3, Fig. 4), but both could be separated from yellowfin menhaden larvae between 10 and 20 mm (Houde and Swanson 1975) by body depth, prepelvic length, and head length. Snout length and eye diameter may be useful to distinguish 15-25 mm specimens; snouts >7% of SL and eye diameter >9% of SL probably identify yellowfin menhaden. Myomeres The total number of myomeres could be counted only on specimens under 17 mm in length. Although the preanal myomeres could be easily counted on larger specimens, the last few postanal myomeres on the peduncle became indistinguishable. The number of myomeres (mean = 44.6) did not change significantly with length in gulf menhaden and cor- responds with the number of adult vertebrae (44-46; mean = 44.7 not counting the hypural bones) report- ed by Dahlberg (1970). Radiographs of 20 adult gulf menhaden spawners used in my study showed that all fish had either 45 or 46 vertebrae (counting hypurals), with a mean of 45.6. During development the dorsal and anal fins moved in relation to the myomeres (Table 4). The anterior end of the dorsal fin moved from myomere 30 forward to myomere 23, numbered from head to tail. The posterior end of the dorsal fin remained fixed at myomere 32. The anus and the anterior end of the anal fin moved forward from myomere 37 to myomere 34. The postdorsal- preanal myomere count of 2 or 3 is diagnostic for Brevoortia at lengths >14 mm. Atlantic menhaden larvae 6-16 mm SL had a mean of 47.2 myomeres, with about two more predorsal myomeres and one more postanal myomere than gulf menhaden. Myomere number and distribution for gulf men- haden and yellowfin menhaden (Houde and Swanson 1975) were so similar that neither were useful for 89 Table 3. FISHERY BULLETIN: VOL. 82, NO. 1 -Proportions of head and body parts of Atlantic menhaden, Brevoortia tyrdnnus, expressed as a percent of standard length. Characters were not developed at lengths marked with a dash. Length class Number of Preanus Predorsal Prepelvic Body Dorsal fin Anal fin Head Snout Eye (mm. SLJ specimens length length length depth base length base length length length diameter 30-3.9 4 85.4 — — — — — 14 5 1 9 6.7 4.0-4.9 10 82 8 — — 8 3 — — 12 0 1.9 54 5.0-5.9 15 81 0 — — 84 — — 12.1 2 2 50 6.0-6.9 7 81 4 — — 8 4 — — 13.3 2 5 4.8 7.0-7.9 18 82 3 71.0 — 8 0 2 6 — 13.7 2 6 4 8 8.0-8 9 12 82 7 696 — 7.9 4 3 — 13 9 2 6 4 8 9.0-99 13 82 8 67.3 — 8 3 6 5 24 15.0 3 0 5.2 10.0-10.9 i 3 85 6 669 — 8.6 9.5 4.2 16.4 3.4 5.3 11.0-11.9 8 85.9 66.4 — 3 7 10.1 5 0 16.6 3 5 5.4 120-12 9 10 84.7 64 6 — 9 1 1 1 5 5 5 17 6 3 7 5.6 13.0-13 9 10 83.2 63.6 — 9 4 13.0 6 / 18 2 4 0 6.0 14.0-14.9 7 82.9 62 7 45.8 9 8 13 6 7.1 18.3 40 62 150-15.9 7 81 7 61 9 45 3 10 0 14.0 7 8 18.3 40 6 2 16.0-16.9 9 80.8 62 5 45 5 11.4 14.0 8.8 20 2 4 1 6 8 170-17.9 3 79 9 60 2 47 6 12 8 15 2 10.0 22.9 4 5 7.3 180-18.9 >■, 77.9 586 47 0 14 2 15 7 106 23.2 4 4 7.4 19 0-199 9 76 9 57 3 48 0 16 1 16 0 11.8 23 8 4 6 7.8 20.0-21 9 7 74 2 538 486 19.8 17 1 142 27.2 48 8.0 220-239 3 73.4 504 509 24 7 17.7 16 1 29 8 5.5 80 240-25.9 2 72.7 51.4 51.3 25.4 17 6 15 9 31 0 5.7 7 6 26.0-27 9 3 74.7 49 6 52 8 29 1 19.6 18.0 31.1 7.0 8 3 280-29.9 1 72.6 48 9 51 5 290 17.3 16 3 31 3 b 8 7.8 30 0-349 4 75.4 49.6 52.4 32.6 20.0 15.6 330 8.2 88 350-39.9 3 76.0 49.9 53 2 36.5 20 4 16.6 33.6 7.6 7.8 40 0-49.9 4 74.9 49.6 52 2 33.5 20 5 17.1 32.2 1 6 8 3 60.0-69.9 3 74.8 48.9 52 2 334 19 5 16 8 324 7.0 5.3 TABLE 4. — Number of myomeres relative to dorsal fin and anal locations on gulf menhaden, Brevoortia patronus, larvae. Length class (mm. SL) Preana Postana Predorsa 1 Postc ursal- Preanal N Range Mean N Range Mean N Range Mean N Range Mean <60 4 36-37 36 7 4 3 8.0 — — — — — — 6 1-80 16 36-37 36.7 3 7-9 7 7 9 28-30 28.9 9 4-6 5 3 8.1-10.0 9 35-38 36 3 9 8-10 8 6 9 26-28 27.3 9 8 5 4 4 10 1-120 10 33-37 354 10 8-10 9.1 10 23-27 252 10 3-4 3.3 12 1-140 4 33-35 34.0 4 8-10 9 5 4 23-25 23.7 4 2-3 2 2 14 1-17 0 4 32-33 32.5 — — — -1 22-23 22 2 9 1-2 1.8 separating small larvae of these species. Yellowfin menhaden had a mean of 45.7, about one less pre- dorsal myomere, and about one to two more postanal myomeres than gulf menhaden. Atlantic menhaden had about two more preanal myomeres and about one more postanal myomere than gulf menhaden at each size class (Table 5). Meristics In gulf menhaden the caudal and dorsal fins were the first fins to initiate development and the pectoral fins were the last fins to complete development, even though they were the first fins to form as nonrayed buds (Table 6, Fig. 1C). Two specimens had an extra principal ray in both the upper and lower group of caudal rays. Vertebrae centra did not first stain with alcian blue as did other bony structures. At 13 mm, vertebrae first stained with alizarin red S, with the staining reaction progressing from the middle of the column towards each end as length increased. The neural and haemel spines initially stained blue, beginning at each end of the column and progressing towards the middle. The mean number of vertebrae, TABLE 5. — Number of myomeres relative to dorsal fin and anal locations on Atlantic men- haden larvae, Brevoortia tyrannus. Myomeres on specimens <6 mm could not be accurately counted. Length class (mm. SLl Preana Postana Predors, I Postdorsal- Preanal /V Range Mean N Range Mean N Range Mean N Rang ; Mean 6.1-8.0 13 38-40 38 7 13 8-10 9.0 10 30-31 30 7 10 5-6 5 7 8 1-10.0 16 37-40 384 16 8-11 99 16 27-30 29.0 10 4-6 5 2 10 1-120 16 36-37 36 1 16 10-11 108 16 25-28 262 16 3-5 4 II 12.1-140 14 35-37 35.6 10 10-11 10 7 14 24-26 25.1 14 3-4 3 2 14 1-16.0 2 35-36 35 5 — — — 2 24-25 24 5 3 3 3.0 90 HETTLER: DESCRIPTION OF GULF MENHADEN 40r -I CO O z UJ -I co Z < UJ a CO I »- o z UJ < CO a O o UJ a o. 90 80 70- 7 Or 60- 50- 55 O Z UJ > CO _l UJ a. UJ a a. * -j CO I H o z UJ 45 10 5- o z CO x .-«r> •v N •T • P s. • p , . -7 T"—' / ■yp M* # CO a UJ Q > Q O CO 30 20 10 40r CO O z UJ _l Q < UJ I 30- 20- CO a. UJ < Q UJ > 10 15 10 - ». .4 - '■m ■••• •■•'' £— // --v ^ g^t ^c- -- -£-' ,V 10 20 30 10 20 30 SIZE CLASS (mm) SIZE CLASS (mm) FIGURE 4. — Morphometric comparisons as a percentage of standard length of laboratory-reared Brevoortiapatronus (P).ff tyrannus (T),andR smithi (S). Yellowfin menhaden data from Houde and Swanson (1975). 91 FISHERY BULLETIN: VOL. 82, NO. 1 TaBI I 6. Meristics in gulf menhaden, Hrcvtmrtia patronus, (35 specimens) and in Atlantic menhaden, li tyrannus, (3 I specimens). Size (mm SL) when Size (mm SL) when first stained all are stained Number in fu II complement Meristic B patronus B. tyrannus B. patronus B tyrannus B patronus B tyrannus Caudal fin rays Principal 8 3 9 12 10-11 (dorsal) 10 (dorsal) 11. I i 18 20 9-10 (ventral) 9 (ventral) Procurrent 8-9 7-8 (dorsal) (ventral) 7-8 (dorsal) 6-7 (ventral) Diiis.iI tm Pterygiophores B B 16 n, 19-21 18-19 Rays 8 '< 19 17 21-23 20-22 Anal fin Pterygiophores 9 10 16 15 17-20 17-20 Rays 10 12 17 1!, 18-22 19-21 Pelvin fin rays li, !'■ 18 18 7 7 Pectoral fin rays 18 IK 21 .'1 13-15 15-17 Predorsal bones pi 1 J 21 21 9-1 1 10-12 Vertebrae 13 14 16 15 45-46 48-49 Ventral scutes 21 21 31 27 29-31 32-33 including the hypural bones, was 45.3 counted in 21 specimens longer than 16 mm SL. The first bones to stain with alizarin red S were the dentaries, the max- illaries, and the cleithra which occurred in 9 mm specimens. Only vertebrae and ventral scute counts were useful in separating gulf menhaden and Atlantic menhaden; other meristics overlapped (Table 6). Yellowfin menhaden larvae could not be separated from the two large-scaled menhaden by meristics, with the possible exception of Atlantic menhaden that had 47-48 vertebrae and yellowfin menhaden that had 45-47 (including the hypural bones) (Dahlberg 1970). Pigmentation Pigmentation of gulf menhaden larvae (Figs. 1,3,5, 6) was similar, but not identical, to the pigmentation described for yellowfin menhaden (Houde and Swan- son 1975) and Atlantic menhaden (Jones etal. 1978). Gulf menhaden up to 8 mm had 1 melanophore on the dorsal side of the notochord tip and 1 or 2 mela- nophores on the ventral side of the notochord tip, which is diagnostic for the genus Rrevoortia (Figs. 1C, D, 5A). Lateral pigmentation, although found on the trunk of specimens as small as 4.9 mm, was not found on all small specimens. At 10 mm, all specimens had 5-20 melanophores scattered the length of the trunk. Larvae 4-5 mm had 10-20 tiny melanophores on top of the head. One 7.8 mm larva had a single stellate melanophore on top of the head behind the eyes. One single medial melanophore, which enlarged into additional melanophores as larvae grew, was present along the isthmus (ventral midline forward of the cleithrum) on 6 mm and larger larvae. On 8-20 mm larvae, 1 or more melanophores occurred along the Vrftl 1mm Fit, i RE 5. Larval Brei oortla patronus: (A) 7.2 mm (12 d alter hatching). (B) 9.2 mm (20 d after hatching). 92 HETTLER: DESCRIPTION OF GULF MENHADEN 5mm FIGURE 6.— Juvenile Brevoortia patronus 33.8 mm (90 d after hatching). cleithrum axis on each side. Along the surface, lateral and parallel with the dorsal surface of the foregut, there were usually 6-10, but sometimes up to 20, rec- tangular melanophores on each side. These paired melanophores were positioned anteriad to 2 or 3 stellate melanophores covering the dorsal surface of the gas bladder. A series of 10-18 medial, unpaired melanophores occurred between the trunk muscula- ture and the dorsal surface of the gut. This series merged into 1-3 stellate melanophores projecting ventrally over the end of the gut towards the anus. A medial string of nearly continuous, thin mela- nophores traced the junction of the finfold along the ventral surface of the hindgut. Dorsal to the base of the anal fin 2 or more melanophores were always pre- sent in larvae >5 mm. The caudal fin was pigmented by 10 mm, whereas the medial fins, lower jaw tip, snout, and nape acquired pigment by 18 mm (Fig. 3C). Pigment was absent on the surface lateral to the ventral portion of the foregut between the distal end of the pectoral fin rays and the pelvic fin. Melanophores were present on specimens >17 mm along the base of the dorsal fin and along the dorsal midline between the dorsal and caudal fins. Paired melanophores were absent between the head and dorsal fin. For pigment descriptions of gulf menhaden larvae and juveniles >19 mm, see Suttkus (1956). Other Structures By 4.5 mm, the dentaries, maxillaries, branchial arches, cleithra, and hypurals were stained with alcian blue, but the first bones to accept alizarin red S stain, and thus indicate ossification, were the cleithra in 8.5 mm specimens. Flexion of the notochord upward to initiate caudal fin development began at 7 mm. Ossification of the hypural bones began at 10 mm and was completed at 15 mm. Eight maxillary teeth and three dentary teeth on each side were observed on 10 mm larvae. Fourteen teeth on each maxillary and three teeth on each dentary were still visible on 25 mm juveniles. In the oral cavity of 16-24 mm larvae, one or two teeth projected downward from each endopterygoid bone and one or two teeth projected upward from the second basibranchial car- tilage. These teeth were absent in fully transformed juveniles. Scales were first visible along the dor- solateral margin of the caudal peduncle and along the midline on each side of the trunk at the beginning of transformation, which occurred at 19 mm. COMPARISON AMONG BREVOORTIA AND WITH OTHER CLUPEIDS Of the Brevoortia species, eggs and larvae of gulf menhaden were the most difficult to distinguish from yellowfin menhaden. Gulf menhaden had 44-46 myomeres, whereas yellowfin menhaden had 45-47 (Houde and Swanson 1975). Morphometries may be useful to distinguish 10-25 mm specimens of gulf menhaden from yellowfin menhaden. At equal lengths, gulf menhaden had less body depth, a short- er head length, a longer prepelvic distance, a longer predorsal distance, a shorter snout, and a smaller eye. Yellowfin menhaden >17 mm had paired melanophores between the head and the dorsal fin (Houde and Swanson 1975), whereas gulf menhaden did not. Wild specimens of yellowfin menhaden from southern Florida also had a double row of melanophores along the ventral midline between the pectoral and pelvic fins, but neither laboratory- 93 FISHERY BULLETIN: VOL. 82, NO. 1 reared gulf menhaden or wild specimens of gulf menhaden collected from four locations along the northern Gulf of Mexico had ventral midline pig- ment. Gulf menhaden had more dorsal fin rays, but both species had an equal number of anal rays. Fer- tilized eggs of the two species had the same diameter, but gulf menhaden had a larger oil droplet (0.20 vs. 0.15 mm) than yellowfin menhaden. No description of finescale menhaden larvae exists, but presumably they have 42-43 myomeres, based on the number of vertebrae reported for this species (Dahlberg 1970). Although gulf menhaden larvae are geographically separated from Atlantic menhaden larvae, they can be separated by counting myomeres or vertabrae; gulf menhaden, 44-46; and Atlantic menhaden, 47- 48. Atlantic menhaden and yellowfin menhaden had nearly equal dorsal and anal fin ray numbers, but Atlantic menhaden had one to four more myomeres and lacked dorsal and ventral midline paired melanophores anterior to the dorsal and pelvic fins. Mophometric differences between Atlantic men- haden and yellowfin menhaden are similar to dif- ferences between gulf menhaden and yellowfin menhaden. There are some differences in egg and larval meris- tics and morphology data between my study and the literature, which may be due to differences between laboratory-reared and wild specimens. Houde and Fore (1973) reported that gulf menhaden had 45-48 myomeres (vs. 44-46 that I found for gulf menhaden), 20-23 anal rays (vs. 19-21), 17-21 dorsal rays (vs. 20- 22), and reported that pelvic fins in northern gulf specimens were not developed until 20 mm (vs. 18 mm). They also reported that gulf menhaden eggs had a diameter of 1.04-1.30 mm (vs. 1.18-1.34 mm), an oil droplet of 0.08-0.20 mm (vs. 0.16-0.22 mm), and a wide perivitelline space of about 33% (vs. 24- 28';). Jones et al (1978) reported that Atlantic menhaden egg diameter was 1.30-1.95 mm (vs. 1.54- 1.64 mm that I found for Atlantic menhaden), that yolk diameter was 0.90-1.20 (vs. 0.82-.095 mm), and that the oil droplet diameter was 0.11-0.17 (vs. 0.20-0.23). For Atlantic menhaden larvae of unspecified lengths they reported 16-18 dorsal rays (vs. 20-22), 18-20 anal rays (vs. 19-21), and a body depth:standard length ratio of about 0.05 at 23 mm total length (vs. about 0.20 I found at the same length); however, the body depth ratio is undoubt- edly a typographical error. Laboratory-reared gulf menhaden and Atlantic menhaden both appeared to transform into juveniles at a smaller size than wild fish. Morphometric data and photographs of specimens of gulf menhaden from Louisiana indicated that the juvenile form was not reached until about 30 mm SL (Suttkus 1956). Lewis et al. (1972) indicated that Atlantic menhaden from North Carolina did not complete "prejuvenile" growth until about 33 mm SL. Houde and Swanson (1975) suggested that tank-reared yellowfin men- haden transformed at smaller sizes than did wild fish, and I concur. Characters useful for separating eggs and larvae of Brevoortia from other clupeids have been identified (Houde and Fore 1973; Richards et al. 1974; Houde and Swanson 1975; Powles 1977). Sardinella and Opisthonema have about the same total myomere counts as Brevoortia, but usually have 6-9 post- dorsal-preanal myomeres. Ktrumcus has the same or more total myomeres than Brevoortia, but about 10 fewer anal rays. The smaller larvae of Sardinella, Opisthonema, and Etrumeus have no pigment on the dorsal side of the notochord tip, whereas Brevoortia, Harengula, and Jenkinsia have this pigment. However, Jenkinsia and Harengula have 42 or fewer myomeres. The spawning seasons of all these genera overlap with the spawning season of Brevoortia species (Houde and Fore 1973; Powles 1977; Jones et al. 1978). Larvae of Dorosoma and Alosa are not normally found in marine waters with Brevoortia. ACKNOWLEDGMENTS I thank John J. Govoni, William R. Nichols, and Allyn and B. Powell of the Beaufort Laboratory for reviewing the early drafts of the manuscript; Ed Houde of the University of Maryland for his review of a later draft; and Thomas Potthoff of the Southeast Fisheries Center, NMFS, and G. David Johnson of the South Carolina Wildlife and Marine Resources Department for their comments on terminology. This research was supported by a contract from the Ocean Assessments Division, National Ocean Services, NOAA. LITERATURE CITED Dahlberg, M. D. 1970. Atlantic and Gulf of Mexico menhadens, Genus Brevoortia (Pisces: Clupidae). Bull. Fla. State Mus., Biol. Sci. 15:91-162. Hettler, W. F., Jr. 1968. Artificial fertilization among yellowfin and Gulf menhaden [Brevoortia) and their hybrid. Trans. Am. Fish. Soc. 97:119-123. 1970. Rearing larvae of yellowfin menhaden, Brevoortia smithi. Copeia 1 970:775-776. 1981. Spawning and rearing Atlantic menhaden. Prog. Fish- Cult. 43:80-84. 1983. Transporting adult and larval gulf menhaden and techniques for spawning in the laboratory. Prog. Fish- 94 HETTLER: DESCRIPTION OF GULF MENHADEN Cult. 45:45-48. HlLDEBRAND, S. F. 1963. Family Clupeidae. In H. B. Bigelow (editor), Fishes of the western North AtlantT. Par) Three, p. 257-454. Mem. Sears found. Mar. Res. Yale Univ. 1. HOUDE, E. D., and P. L. Fore. 1973. Guide to identity of eggs and larvae of some Gulf of Mexico clupeid fishes. Fla. Dep. Nat. Resour., Mar. Res. Lab., Leafl. Ser. Vol. IV, Pt. 1, No. 23, 14 p. Hoi UK, E. D., AND L. J. SWANSON, JR. 197"). Description of eggs and larvae of yellowfin menhaden. Brcvnortia smithi. Fish. Bull.. U.S. 73:660-673. Jones, P. W. . F. D. Martin, and J. D. Hardy, Jr. 1978. Development of fishes of the Mid-Atlantic Bight. Vol. 1, Acipenseridae through Ictaluridae. U.S. Fish Wildl. Serv., Biol. Serv. Program FWS/OBS-78/12, 314 p. Lewis, R. M., E. P. H. Wilkens, and H. R. Gordy. 1972. A description of young Atlantic menhaden, Brevoortia tyrannus, in the White Oak River estuary. North Car- olina. Fish. Bull., U.S. 70:115-118. POWLES, H. 1977. Description of larval Jenkinsia lamprotaenin (Clupeidae, Dussumieriinae) and their distribution off Barbados, West Indies. Bull. Mar. Sci. 27:788-801. Richards, W. J., R. V. Miller, andE. D. Houde. 1974. Egg and larval development of the Atlantic thread her- ring. Opisthnncma nglinum. Fish. Bull., U.S. 72:1123- 1 136. SUTTKUS, R. D. 1956. Early life history of the largescale menhaden, Brevoor- tia patronis, in Louisiana. Trans. North Am. Wildl. Conf. 21:390-407. Turner, W. R. 1969. Life history of menhadens in the eastern Gulf of Mex- ico. Trans. Am. Fish. Soc. 98:216-224. U.S. National Marine Fisheries Service. 1982. Fisheries of the United States, 1981. U.S. Dep. Com- mer., NOAA. Natl. Mar. Fish. Serv., Curr. Fish. Stat. 8200, 131 p. 95 DISTRIBUTION OF ICHTHYOPLANKTON OFF SAN ONOFRE, CALIFORNIA, AND METHODS FOR SAMPLING VERY SHALLOW COASTAL WATERS Arthur M. Barnett,1 Andrew E. Jahn,2 Peter D. Sertic,1 and William Watson1 ABSTRACT Spatial abundance patterns of inshore marine fish larvae, together with day-night and ontogenetic changes in these patterns, were investigated at a single site off the southern California coast using neustonic, midwater, and epibenthic samplers. Fifteen of the nineteen most abundant taxa showed statistically significant abun- dance patterns: Five taxa were principally in the inshore (<2 km from shore) epibenthos, one in the inshore neuston, two in the neuston and midwater less than about 5 km from shore, three to midwater 2-5 km from shore, and four in midwater offshore of about 3.5 km. Abundance patterns for the three most common taxa, Engraulis mordax, Genyonemus lineatus, and Seriphus politus, shifted toward shore and toward the bottom with increasing larval size. Comparison of E. mordax egg and larval abundances indicated a large excess of larvae over eggs nearshore. Only two taxa showed statistically significant day-night pattern changes; both were lower in the water column during the day. The existence of inshore abundance maxima implies significant survival value in occupying the nearshore zone. The shallow waters of the southern California coast may represent a nursery area comparable in impor- tance to the estuarine nurseries of the Atlantic coast of North America. Through the pioneering California Cooperative Oceanic Fish Investigation (CalCOFI) work of the late E. H. Ahlstrom and co-workers (Ahlstrom 1959, 1965), ichthyoplankton of the Southern California Bight are generally well known. However, the CalCOFI effort was concentrated on species found principally offshore of the 100 m isobath, and the lar- vae of most inshore fishes are rare or missing in the published CalCOFI data. Recent studies of ichthyoplankton in the Southern California Bight inshore of the 100 m isobath (Brewer et al. 1981; Gruber et al. 1982; Brewer and Smith 1982) have indicated that many of these larvae are found in the relatively shallow waters. In this paper we present methods for sampling quantitatively the entire water column in shallow waters (6-75 m) and describe the spatial abundance patterns of the most commonly occurring larval fishes. Of particular interest was the distribution of larvae in the onshore-offshore vertical plane. Ontogenetic pattern changes were investigated for three abundant species: Engraulis mordax, Geny- onemus lineatus, and Seriphus politus. 'Marine Ecological Consultants of Southern California, 531 Encinitas Boulevard, Suite 110, Encinitas, CA 92024. 2Marine Ecological Consultants of Southern California, 531 Encinitas Boulevard, Suite 110, Encinitas, Calif.; present address: Los Angeles County Museum of Natural Histoiy, 900 Exposition Boulevard, Los Angeles, CA 90007. The study was done off San Onofre, Calif., (Fig. 1) from September 1977 to September 1979. Unit 1 of the San Onofre Nuclear Generating Station, a 500- megawatt plant located 1.5 km northwest of the sampling area, was operating continuously through- out the course of the study. However, this plant has been shown to have only very localized effects which have not interfered measurably with the results reported herein (Marine Review Committee 19793; Bartlet et al. 198 14). This study was completed prior to the beginning of operation of Units 2 and 3 of the San Onofre Nuclear Generating Station. Our sampling methodology resulted from a pre- liminary study in which we found that a combination of sampling gear was necessary to estimate nearshore larval abundance. The chief purpose of this paper is to present these sampling methods. Results are shown which verify the effectiveness of these methods and further suggest some peculiarities of the nearshore habitat. 'Marine Review Committee. 1979. Interim report of the Marine Review Committee to the California Coastal Commission. Part 1: General summary of findings, predictions, and recommendations concerning the cooling system of the San Onofre Nuclear Generating Station. In Marine Review Committee Document 79-02, p. 1- 20. Marine Review Committee of the California Coastal Commis- sion, 631 Howard Street, San Francisco, CA 94105. 4Barnett,A.M.,P.D. Sertic, and S.D. Watts. 1981. Final report: Ichthyoplankton preoperational monitoring program. Marine Ecological Consultants of Southern California, 531 Encinitas Boulevard, Encinitas, CA 92024, 8 p. Manuscript accepted June 1983. FISHERY BULLETIN: VOL. 82, NO. 1. 1984. 97 FISHERY BULLETIN: VOL. 82, NO. 1 San Onofre Nuclear Generating Station 18 m 37 m 55m 74 m 13 m it- A '"oV-iT- San Onofre Kelp c D Long Beoch^ N 16 km Pacific Ocean FIGURE 1.— Chart of the sampling area and its position off the southern California coast. The one- and two-dimensional pattern analyses were based on samples taken at a randomly selected isobath in each of the five sampling blocks (A-E) on each sampling date. The study of daily vertical migration was based on samples taken along the 8 and 13 m isobaths (dotted lines). size, and sampling time for the ensuing full-scale program. The results of this brief study indicated that 1. Filtration efficiency was at least 85% for all nets and lengths of tow. 2. Samples of 400 m3 were adequate to attain asymptotes of numbers of taxa per tow. A sampled volume of 400m3 from the epibenthos was the max- imum that could be handled economically. 3. The 12 most abundant larval fish taxa were neither randomly nor evenly distributed with respect to the three vertical strata. Half the taxa were prin- cipally epibenthic, while 25% were neustonic and 25% were most abundant in midwater. 4. Only one of these taxa showed a daily vertical migration; Paraclinus integripinnis, not a top-ranking species in the ensuing study, tended to descend from midwater to the epibenthic layer at night. 5. Size of individuals and apparent abundance of most taxa increased at night, probably because of visual avoidance during the day. 6. Nitex netting of 0.333 mm mesh retained more fish eggs and smaller anchovy larvae than did 0.505 mm mesh. From the preliminary results, it was clear that the bongo net alone would undersample significant frac- tions of many larval populations. Since our goal was to estimate the density and distribution of nearshore ichthyoplankton, we decided to use all three types of gear with 0.333 mm mesh and to filter a target volume of 400 m3. METHODS Preliminary Study In shallow depths, interfaces at the sea surface and seabed comprise a substantial portion of the water column. In addition, concentration of a species at either interface would necessitate sampling the epibenthic and neustonic layers as well as the mid- water column to obtain quantitative abundance estimates. Neustonic, midwater, and epibenthic samplers were used in a preliminary study5 between Septem- ber and November 1977, to verify their effectiveness and to select mesh size, net design, standard sample !Barnett, A. M.,J. M. Leis, and P. D. Sertic. 1978. Report to the Marine Review Committee on the preliminary ichthyoplankton studies. Marine Ecological Consultants of Southern California, 53 1 Encinitas Boulevard, Encinitas, CA 92024. Sampling Gear A bongo net was selected for sampling the midwaters, as recommended by Smith and Richardson (1977). An opening-closing 71 cm Brown-McGowan bongo net (total mouth area = 0.79 m2) was used. A General Oceanics6 (GO) flowmeter was mounted in the star- board frame. The bongo net, as conventionally used, is placed on the wire some distance above a weight and towed astern. The geometry of this arrangement and the circular net mouths make the gear ill-suited for sampling the plankton in the neustonic and epi- benthic strata near the sea surface and seabed, re- spectively. Therefore, specially designed samplers, described below, were used to sample these layers. We chose the brown manta net (Brown and Cheng 6Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 98 BARNETT ET AL.: DISTRIBUTION OF ICHTHYOPLANKTON OFF SAN ONOFRE, CALIF 1981) as our neustonic sampler. This net had an 88 cm wide mouth and fished to a depth of 16 cm. Fiberglass-covered styrofoam floats kept the top of the net out of water, and a 3 m spar and asymmetrical bridle kept the gear outboard of the bow wave. A weight suspended from the end of the wire held the bridle well below the surface, out of the path of the net. The sampler was launched and recovered off the quarter by means of a tag line. Both a Tsurumi-Seiki (TSK) flowmeter and a GO flowmeter were mounted in the mouth of the net. The GO meter served as a back-up for the TSK, which sometimes fouled with kelp and eelgrass. The Auriga net,7 used to sample the epibenthic layer, consisted of a rectangular net frame (0.5 m high X 2 m wide) attached to a chassis equipped with a pair of side-mounted, 2 m diameter wheels. The device rolled on the bottom so that the mouth of the net was 10 cm (original design) or 17 cm (later ver- sions) above the bottom of the wheels. A series of 12 cm diameter plastic rollers below the mouth of the net helped prevent the sampler from digging into the bottom and presumably minimized escapement below the net. Both GO and TSK flowmeters were mounted within the mouth of the Auriga net. The Auriga net was towed off the stern. Divers have observed (M. Sowby8) that the mouth of the Auriga assumes a horizontal attitude when the wheels are off the bottom. We therefore believe that contamination of the epibenthic samples by midwater plankton was minimal during launch and recovery, when the main component of (relative) water movement was across, rather than through, the mouth. Any contamination that did occur should have been a function of depth, which was always <209r of the length of an epibenthic tow (this potential source of error has been ignored in the density calculations). Although serious clogging was not apparent in the preliminary study, denser plankton concentrations at other times of the year might clog the nets before 400 m1 of water could be filtered. Clogging would be most serious for oblique bongo tows, because it would result in undersampling of the upper part of the water column. In anticipation of this possibility, the area of mesh in all nets was increased according to the criteria suggested by Smith et al. (1968, equation 5) in order to sample 500 m3 (bongo), 400 m3 (Auriga), and 200 m3 (Manta) for "green" coastal waters. The filtering ratios (R = mesh pore area/net mouth area) of bongo, Auriga, and Manta nets were increased to 7.8, 6.6, and 1 0.7, respectively, by adding mesh cylin- ders ahead of the conical portions of the nets. Exter- nal flowmeters were not used in the subsequent surveys, but tows were carefully timed. Internal flow- meter readings were checked upon recovery, and samples were repeated if the readings differed by more than 20% from expected values. Except for the limited study of daily vertical migra- tion, all sampling was done at night. The deck lights were always off during the neuston tows. All samplers were launched, towed, and recovered with the vessel underway at about 1 m/s. For bongo tows, wire was paid out (scope about 2:1) until the weight, located 1.5 below the center of the net frame, bumped the bottom. Then the nets were opened, and a stepped oblique tow was made consisting of 18 30-s steps. The Auriga sampler was towed with a scope of 3:1 and recovered after 6.5 min on the bottom. With the small-mouthed Manta net, the volume of 400 m3 was achieved by towing two nets simultaneously, off port and starboard, for 20 min (about 1.4 km). Samples were preserved in 5-10% seawater- Formalin. Sampling Locations and Frequency Since we eventually wanted to assess the effects of a power plant cooling system, it was necessary to con- centrate much of our sampling effort within the depth contours encompassing the cooling structures. At the same time, in order to estimate the abundance of nearshore species, we needed to sample far enough from shore to delimit their centers of abundance. We decided upon a stratified random sampling design (Snedecor and Cochran 1967) wherein, on each sam- pling date, the neustonic, midwater, and epibenthic layers were sampled along a randomly chosen depth contour in each of five blocks (Figs. 1, 2). The five blocks were defined by depth contours: A) 6-9 m, cor- responding to cooling water intake locations; B) 9- 1 2 m and C) 12-22 m, both corresponding to future dif- fuser discharge locations; D) 22-45 m, corresponding to a faunal break between inshore and coast- al zooplankton assemblages (Barnett and Sertic9); and E) 45-75 m, chosen a priori as the likely offshore limit of most nearshore larval fishes. The sampling transect thus consisted of 15 strata: Three depth layers in each of five blocks (Fig. 2). To "Marine Biological Consultants, Inc., 947 Newhall Street, Costa Mesa, CA 92627. "M. L. Sowby. Marine Biological Consultants, Inc., 947 Newhall Street, Costa Mesa, CA 92627, pers. commun. 1979. 'Barnett, A.M., and P. D. Sertic. 1979. Spatial and temporal pat- terns of temperature, nutrients, seston, chlorophyll-a and plankton off San Onofre from August 1976 - September 1978, and the relationships of these patterns to the SONGS cooling system. In Marine Review Committee Document 79-01, p. vii through 9- 89. Marine Review Committee of the California Coastal Commis- sion, 631 Howard Street, San Francisco, CA 94105. 99 FISHERY BULLETIN: VOL. 82, NO. 1 FIGURE 2. — Diagrammatic profile of the study transect showing the 15 strata sam- pled off San Onofre, Calif. Neustonic and epibenthic layers are vertically ex- aggerated. Distance from Shore (km) 2 3 4 5 6 7 8 _1 I I I I 1 I 0 - 10 20 30 o ft) 40 ? - 50 3 - 60 70 80 avoid the San Onofre kelp bed, some of the tows in the B and C blocks were offset by about 1 km. VVilcoxon signed rank tests of samples taken from B block and B offset (Fig. 1) showed no significant dif- ferences in species abundances (P > 0.05) between the main block and the offset which could not be related to the inshore-offshore patterns discussed below. The transect was sampled monthly in January and February 1978, fortnightly from March through August 1978, and again monthly through September 1979. During each of these 28 sampling periods, the five blocks were surveyed once each night for 1-3 nights, giving a total of 57 sampling dates for the 21- mo study. As noted above, we chose a standard sampled volume of 400 m3 based on the preliminary study. This volume was large enough to assure a representa- tion of all abundant species throughout the year. Volume was used as the sampling unit, although an argument based on the scale of patchiness could be made for length of tow (i.e., 400 m in each water layer) as the criterion, rather than volume filtered (P. Smith10). Most tows were at least 400 m long. Laboratory Procedures Samples were sorted for fish eggs and larvae under dissecting microscopes at 10X magnification. The choice of 400 m3 as the sampled volume was made at a time of year when ichthyoplankton abundance was low (Walker et al.11); consequently the samples from other times of year were larger than necessary to rep- ,0P. E. Smith, La Jolla Laboratory, Southwest Fisheries Center, National Marine Fisheries Service, NOAA, La Jolla, CA 92038, pers. commun. 1979. "Walker, H. J., A. M. Barnett, and P. D. Sertic. 1980. Seasonal patterns and abundance of larval fishes in the nearshore Southern California Bight off San Onofre, California. Marine Ecological Con- sultants of Southern California, 53 1 Encinitas Boulevard, Encinitas, CA 92024. resent the nearshore assemblage. Samples with large plankton volumes were subsampled, using a Folsom plankton splitter before sorting. The size of the sub- sample was set to include at least 100 non-engraulid larvae (the mean number of larvae counted per sub- sample was 277, of which 62.8% was is. mordax). This fraction was usually one-fourth and was seldom smaller than one-eighth. Eggs were sorted from 1%, 5%, or 10f/r (to get at least 100 eggs) of the residue of the fraction sorted for larvae. Sorting efficiency was maintained above 90%. Some epibenthic samples contained so much sand and detritus that it was necessary to clean them before sorting, using a flotation technique adapted from Ladell (1936). After removal of large fish and debris, such a sample was mixed with a 40% MgS04 solution (specific gravity = 1.2) in a large separator fashioned from a 19 1 (5-gal) plastic carboy with the bottom cut off and the neck fitted with a rubber hose and ball valve. Most detritus sank, while plankton floated to the top. The heavy material was drained off and processed once or twice more to ensure separa- tion of the plankton. Checks of the heavy residue of three such samples showed that more than 99% of the larvae were separated by flotation. All larvae were identified to the lowest taxonomic category currently possible. Eggs were identified as Engraulis mordax or "other". In some larval cate- gories (e.g., Atherinidae, Goby Type A), our ability to discriminate among species or larval types (sensu Richardson and Pearcy 1977) improved as the study progressed. However, not all of the old collections were reprocessed. When mixed taxa showed seasonal and spatial coherence, they were retained for the analyses presented here. Pattern Analysis All counts of eggs and larvae were standardized to number/400m-\ Thus the standardized numbers 100 BARNETT ET AL.: DISTRIBUTION OF ICHTHYOPLANKTON OFF SAN ONOFRE, CALIF. were roughly the same as the actual numbers of eggs and larvae caught, a desirable situation for analysis with transformed data (Murphy and Clutter 1972). These values were transformed by log (X + 1) before analysis for offshore and vertical pattern. The results were back-transformed, resulting in geometric means with asymmetric confidence bounds, and pre- sented as number/1 00m1. To describe the cross-shelf abundance patterns of ichthyoplankton, a procedure was adopted involving Hotelling's T2 test and a series of a posteriori '(-tests (Morrison 1976) to divide the 15 strata into groups. These parametric methods allowed us to detect significant differences in mean abundance among components of a pattern and to determine con- fidence bounds on the means. Hotelling's T l test was selected over an analysis of variance (ANOVA) because the covariance struc- tures in the data tended not to meet the assumptions of standard ANOVA models (i.e., errors were not independent; the abundances of neighbor strata were likely to be correlated). The T 2-test allows this correlation by using the sample covariance matrix, rather than (as in ANOVA) assuming a specified covariance pattern (Winer 1971; Morrison 1976). With a significant T1 test result obtained (P< 0.05), a posteriori multiple (-tests were used to separate strata into groups having significantly different abundances. The strata were contrasted in a series of (-tests using the Bonferroni statistic, ((0.05),,, where k — potential number of contrasts, s = number of sampling periods — 1, and 0.05 = overall type / (a) error. The value of k was set as the number of all poss- ible contrasts among m strata plus 5, for further tests employing combinations of the initial strata: i.e., (m)(m-l) k + 5. Bonferroni (-values were taken After the initial series of (-tests of all possible com- parisons, strata found not to differ significantly were pooled into initial groups. The time-averaged abun- dance of each stratum was used to calculate the initial groups' mean abundance Zj = z Z,/n where Z, is the initial group mean, n is the number of strata in the initial group, and Z, are the means of individual strata. Further (-tests (the total of all tests 3 individuals/400 m3 (0.75/100 m3). N = Numbers of surveys used in analysis. Geometric mean abundances with 957c confidence bounds for each of these groups are given in Table 1. dates when a taxon was present. The latter method was used only to obtain cross-shelf patterns; in these cases, mean abundances in the various parts of the pattern are relative numbers, and confidence bounds were not calculated (Table 1). All testing was done on the basis of abundance alone, without regard to the strata being grouped. Final groupings of strata are shown in diagrams of the cross-shelf transect (Fig. 3). Occasionally, non- abutting strata were members of the same statistical group. These are depicted as being physically con- nected when such an interpretation is reasonable. In all cases, shading is used to indicate groups of strata which differ significantly. RESULTS Cross-Shelf Patterns The 19 larval taxa analyzed were those which rank- 102 BARNETT ET AL.: DISTRIBUTION OF ICHTHYOPLANKTON OFF SAN ONOFRE, CALIF. ed among the 10 most abundant in any of the 5 sam- pling blocks. Fourteen taxa showed significant differences among the strata which were resolved into spatial patterns (Table 1, Fig. 3). Taxa with cen- ters of abundance nearest shore tended to be concen- trated in either the epibenthic or the neustonic layer. Of the five epibenthic taxa, four (Gibbonsia Type A, Seriphuspolitus,Gobiesox rhessodon, and Goby Type A [consisting of Ilypus gilberti and Quietulay-cauda]; Fig. 3A-D) had centers of abundance within 2 km of shore. The fifth, Genyonemus lineotus, was most abundant out to about 4 km (Fig. 3E). Atherinidae (Fig. 3F) were neustonic and most abundant within 2 km of shore. Hypsopsetta guttulata (Fig. 3G) was abundant in the neustonic and midwater layers out to 2 km. It had the most nearshore pattern of any mid- water taxon. Hypsoblennius spp. were concentrated in the neustonic and midwater layers out to about 5 km and in the neustonic layer beyond 5 km from shore (Fig. 3H). The remaining six taxa with discernible patterns were all most concentrated in midwater. The centers of abundance of Engraulis mordax and Paralichthys californicus (Fig. 31, J) extended from 2 to ~5 km from shore, while those of Pleuronichthys verticalis, Citharichthys spp., Sebostes spp., and Stenobra- chius leucopsarus appeared to extend seaward of the sampling area (Fig. 3K-N). Five taxa (Chromis punctipinnis, Parolobrax spp., Porophtys vetulus, Peprilus simillimus, and Pleuro- nichthys ritteri) were not shown to have patterns by this analysis. Vertical Migration Because the basic study plan called for nighttime sampling, the patterns described would pertain to nighttime distributions. The preliminary study found little evidence of daily vertical migration; neverthe- less, we conducted a further small study of vertical migration to test whether the vertical component of the patterns remained the same during daylight hours. The study was conducted at two inshore locations (Fig. 1). A description of the vertical study is given in the Appendix. There was no indication of vertical migration at the 8 m station, but at the 13 m station two taxa, Hyp- soblennius spp. and Paralichthys californicus, showed significant (P < 0.05) vertical shifts downward in the water column during the day (Fig. 4). The low probability (0.055) of the F value for Gobiesox rhcssodon (App. Table 2), though higher than the customary rejection level of 0.05, suggests a daily change in vertical distribution. The data indi- cate this species may, like Paraclinus integripinnis in the preliminary study, tend to migrate or settle from midwater into the epibenthic layer at night. Onshore-Offshore Abundance The analysis of cross-shelf pattern assumes that lar- vae are uniformly distributed throughout each mid water stratum, an assumption that becomes in- creasingly untenable with depth of stratum. Layering of ichthyoplankton within the midwater zone will cause an apparent decrease in density in the seaward blocks, as more of the volume used in the density calculations comes from deeper waters where a species may be rare. To eliminate bias in the cross- shelf patterns due to inclusion of noncontributing depths in the density calculations, one-dimensional abundances were calculated based on the estimated number of larvae under a unit ( 1 00 m2) of sea surface in each offshore block jV 3 z rtjdi where n = larvae/ 1 00 m-1 in stratum i and d — vertical thickness of stratum i in meters (0.16 m, neustonic; 0.50 m, epibenthic; depth of water column — 1 m, midwater). The one-dimensional patterns, which emphasize numbers of larvae (Table 2), provide a useful com- parison to the two-dimensional patterns which emphasize larval density (Table 1, Fig. 3). All epibenthic and neustonic taxa had similar onshore- offshore centers of abundance as determined by both methods. This was expected, since their cross-shelf abundance patterns were essentially one- dimensional. Gibbonsia Type A, Seriphus politus, Gobiesox rhessodon, Goby Type A, and Atherinidae, all with abundance centers within 2 km of shore in the two-dimensional analysis (Fig. 3), likewise had one- dimensional maxima shoreward of 2 km. With the exception of S. politus, these taxa were less than half as abundant beyond 2 km. Genyonemus lineatus, most concentrated in the epibenthic layer within about 4 km of shore, had a one-dimensional max- imum at 2-4 km but remained abundant (>V£ max- imum) out to ~5 km. Of the eight midwater taxa, only two had one- dimensional patterns which differed from their two- dimensional patterns Engraulis mordax appeared more abundant farther offshore in one dimension (cf. Table 2 and Fig. 31). The steady increase in abun- dance of E. mordax with distance from shore is at odds with its two-dimensional pattern (Fig. 31) and 103 Hypsoblennius spp. 24 July 1978 Mean Number/100 m 65432 10 I 2345 Hypsoblennius spp. 30 August 1978 Mean Number/ 100 m 10 5 0 5 10 15 25 20 FISHERY BULLETIN: VOL. 82, NO. 1 Hypsoblennius spp. 22 September 1978 Mean Number/ 100 m 15 10 5 0 I i i l T~ V 1 5 ~r 10 0 2 4 a 10 12 ' Day Night Paralichthys californicus 24 July 1978 Mean Number/100 m 2 10 12 I 1 1 1 1 Day Night Paralichthys californicus 30 August 1978 Mean Number/IOOm '10 12 3 Day Night 50 Paralichthys californicus 22 September 1978 Mean Number/IOOm 40 30 20 10 0 10 20 0 2 4 E. 6 .c QJ 8 Q 10 ILT Day Night Day Night Day Night FIGURE 4. —Average vertical abundance profiles of Hypsoblennius spp. andParalichthys californicus during the study of daily vertical migration off San Onofre, Calif. The depth ranges of the five sampling strata are the averages (based on four to six profiles) for each sampling period. Note that the horizontal (abundance) scale varies. TABLE 2.— Numbers of larvae under 100 m2 of sea surface in the five sampling blocks, averaged over 57 cruises, off San Onofre, Calif. Sampling block: A B c D E Offshore limits (km): 0.5-1.1 1.1-1.9 1.9-3.7 3.7-54 5 4-7 2 Gibbonsia Type A 6.4 103 1.5 03 1.1 Senphus pohtus 273.9 103 9 217.9 118.9 93 7 Gobiesox rhessodon 46 12.1 5.3 1.1 30 Goby Type A 24.5 17.5 3 5 2 9 1.1 Genyonemus /meatus 1 32.7 312.4 623 3 5665 221.1 Athennidae 35.7 28 1 11.7 89 49 Hypsopsetta guttulata 3.1 3 2 39 06 0 7 Hypsoblennius spp. 27.5 26.9 48.1 63.0 369 Engrauhs mordax 9700 1,833.4 6.454 4 9.2502 10.263.5 Paralichthys californicus 4 3 1 14 90 0 103 2 42 4 Pleuronichthys vertical's 04 2 3 13 4 36.4 11 7 Pleuromchthys ntten <0.1 02 56 30 9 13 9 Cithanchthys spp. 2 9 3.5 99 17.9 31 0- Sebastes spp. <0 .1 <0 .1 18 2 77 7 5186 Stenobrachius leucopsarus 0.1 04 4 4 29 1 106.1 Chromis punctipmnis 0 0 08 66 53 3 Paralabrax spp. 0.1 08 34 3 97.8 84 .1 Parophrys vetulus 0 5 03 0.1 7.3 33.6 Peprilus simillimus 20 4 1 3 6 10.0 17.4 104 BARNETT ET AL.: DISTRIBUTION OF ICHTHYOPLANKTON OFF SAN ONOFRE. CALIF. indicates that this species must be vertically stratified beyond the 45 m contour. This agrees with the findings of Ahlstrom (1959) in which the majority of E. mordax larvae occurred above 50 m. In contrast, Pleuronichthys veHicalis peaked in abundance at 4-5 km rather than extending offshore as in the two- dimensional analysis (cf. Table 2 and Fig. 3K). This result may have occurred because the tests used in the two-dimensional analyses failed to distinguish between offshore blocks due to the small number (27) of non-zero observations for this species. Four of the five taxa lacking statistically significant two-dimensional patterns (Chromis punctipinnis, Paralabrax spp., Parophrys vetulus, Peprilus siml- imus) appeared to be most abundant beyond 4-5 km when considered in one dimension (Table 2). The fifth, Pleuronichthys ritteri, peaked in abundance at 4-5 km from shore. Ontogenetic Pattern Changes Larvae of the three most abundant species were divided into size groups, which were analyzed separately for spatial pattern. To prevent temporal bias in the patterns, only 1978 data were used since they covered a full year. Larvae of two sciaenids, Genyonemus lineatus and Seriphus politus, were each divided into groups corresponding to developmental stages. Preflexion larvae, with straight notochords and no hypural development, were analyzed separately from more fully developed, and pre- sumably more mobile, flexion and postflexion larvae. Hypural development was found to begin at 3.8 mm for G. lineatus and at 4.1 mm for S. politus. Similarly, Engroulis mordax larvae were divided into early and late developmental stages, but this was done on the basis of size alone and did not correspond to flexion of the notochord. Early preflexion larvae (<6 mm), termed "early stage", were analyzed separately from other larvae, termed "late stage". One hundred lar- vae or all specimens, whichever was less, were mea- sured for each species in each collection. When only the first LOO larvae were measured, the proportions of the various size classes were applied to the total. To examine the ratio of older to younger larvae, the total number in each sampling block (Fig. 1) was calculated, using a longshore dimension of 1 m, i.e., number in block, Nb = N ■ L, where N is number under 1 00 m2 of sea surface in the block, and L is the onshore-offshore extent of the block in hundreds of meters. The patterns of all three species were more nearshore and epibenthic for older larvae (Table 3, Fig. 5). The ratio of older to younger larvae was about 1:2 for all three species (transect totals, Table 4). This ratio increased in the shoreward blocks for G. lineatus and 8. politus, reaching maxima in blocks A and B. The ratio of older to younger E. mordax larvae was maximum in blocks C and D. The remarkable aspect of the E. mordax data is that there were far too few eggs in the nearshore zone to account for the numbers of larvae. The ratio of total E. mordax lar- vae to eggs was about 28:1. The median size of the larvae was about 6 mm, corresponding to an average age of roughly 10 d (Methot and Kramer 1979). Zweifel and Lasker (1976) found a time to hatching of 2.5 d (at about 16°C). The ratio of 10-d-old larvae to eggs thus has an upper limit of the order 4:1 in the absence of mortality, implying at least a sevenfold excess of larvae in these nearshore samples. The minimum diameter of E. mordax eggs during the months of maximum egg abundance is about twice the mesh opening of the plankton nets used, so that sampling deficiencies for these immobile objects should be negligible. Table 3.— Geometric mean abundance (no. 100m3) with 95% confidence bounds (C.B.) for younger and older age groups of larvae of EngrauHsm&rdax, Genyonemus lineatus, and Seripkus politus, showing statistically signifi- cant cross-shelf patterns off San Onofre, Calif. Groups of strata which differ significantly in mean abundance are ranked from highest to lowest. Refer to Figure 5 for locations of these groups. Mean abundance" Strata groups: 1 2 3 4 95% C.B.: Lower Mean Upper Lower Mean Upper Lower Mean Upper Lower Mean Upper Engraults mordax early stage larvae 2 33 13.21 7006 0 52 3.31 16.24 022 1.10 3.63 late stage larvae 23.43 62 42 165.65 5.53 14.34 36.60 0.92 2.99 8.72 Genyonemus lineatus Preflexion stage larvae 1 55 7 42 3240 0 73 3 15 11 52 0 33 1.11 2.93 004 0 26 0.64 flexion and postflexion stage larvae 7.46 3088 125.51 0 53 1 56 3 97 0 10 062 1.90 0 02 0.08 0.14 Senphus politus preflexion stage larvae 0 58 1.37 2.90 0 15 0.49 1.12 004 0 16 0.33 flexion and postflexion stage larvae 4 31 2064 95 57 0.50 1 90 5.86 0 0.10 0.22 105 FISHERY BULLETIN: VOL. 82, NO. 1 DISTANCE FROM SHORE (km) 12 3 4 5 6 7 Engrauhs mordax eggs 0 10 20 30 -|40 50 60 70 J80 DISTANCE FROM SHORE (km) 0 12 3 4 5 6 Engrauhs mordax early stage larvae DISTANCE FROM SHORE (km) 12 3 4 5 6 7 — i 1 1 1 1 1 1 Engrauhs mordax late stage larvae 0 10 20 £ 30" - 40 fE ^°£ 60 o 70 J80 Genyonemus hneatus\ preflexion stage larvae 0 10 20 30 40 50 60 - 70 J 80 Genyonemus hneatus% flexion and postflexion stage larvae 0 10 20 £ 30- - 40 f -50 a- 60 £ H70 80 Senphus pohtus preflexion stage larvae o 10 20 30 40 - 50 - 60 - 70 J80 Senphus politus flexion and postflexion stage larvae 0 10 20 E 30 — 40 f 50 Si 60 q 70 80 FIGURE 5. — Changes with development stage in the cross-shelf abundance patterns of Engraulis mordax, Genyonemus lineotus, and Seriphus politus off San Onofre, Calif. Shading indicates relative abundance in groups of strata differing significantly in mean abundance. Heavier shad- ing indicates higher abundance; the darkest shading (black) is reserved for densities >3 individuals/400 m5 (0.75/100 m3). Geometric mean abundance and 95' i confidence bounds for each group are given in Table 3. Table 4.— Early life stages of Engraulis mordax, Genyonemus lineatus, and Seriphus politus, for 1978 off San Onofre, Calif. See Figure 1 for description of sampling blocks. Sampl ng block (avg. no./m of coastline) Total Species A B C D E no Engraulis mordax eggs 3,100 25.334 85,372 75.238 95.782 284,826 larvae <6 mm 85,302 363.002 1,387,638 1,770,549 1,770.939 5,377,430 larvae >6 mm 42,970 86,977 816.941 1,164,41 7 607,805 2,719,1 10 Genyonemus /meatus preflexion larvae 463 688 7.440 9.290 3.724 21,605 flexion and post- flexion larvae 464 2,969 4,198 2.699 107 10,437 Senphus politus preflexion larvae 592 490 4.200 2.137 2,103 9,522 flexion and post- flexion larvae 2.214 809 779 197 96 4,095 DISCUSSION The methods we have employed for sampling very shallow inshore waters, though not without short- comings, have proven satisfactory in that they clearly emphasize the degree to which many larval fishes are concentrated in different layers, especially near bot- tom. Any quantitative sampling of nearshore fish lar- vae over soft bottom (at least) in the Southern California Bight must clearly include the epibenthic layer. However, our method of doing so may leave room for improvement. The Auriga net probably does not sample the 17 cm immediately above the substrate, unless the rollers induce an avoidance re- sponse such that larvae swim upward and into the mouth. Moreover, we have not determined the thick- ness of the epibenthic microhabitat or whether it is the same for all species. The sharpness of some abun- dance patterns suggests this layer may be no more than 1 m thick (the bongo net tows began about 1 m above the bottom), but small errors in this deter- mination, and failure to sample obliquely from the 106 BARNETT ET AL.: DISTRIBUTION OF ICHTHYOPLANKTON OFF SAN ONOFRE, CALIF. top of the range of the epibenthic gear, could make large differences (by a factor of 2) in the abundance estimates of some taxa. Other studies from the Southern California Bight have shown cross-shelf patterns similar to those which we describe. For example, Gruber et al. (1982; sampling neuston and midwater) and Brewer et al. (1981; sampling the entire water column) both showed vertical and cross-shelf changes in species composition. In both studies, atherinid larvae were principally neustonic. Brewer et al. (1982) took 69% of all larvae on their surveys from the epibenthic stratum. Both studies showed that clinids, most gobiids, sciaenids, and atherinids were most pre- valent nearer shore. Such inshore-offshore patterns have also been shown further north along the west coast (Pearcy and Meyers 1974; Richardson and Pearcy 1977). Icanberry et al. (1978) conducted a distributional study of ichthyoplankton above the epibenthic stratum at two nearshore stations off Diablo Canyon, about 100 km northwest of the Southern California Bight. Though there is taxonomic overlap between their study and ours, their sampling was too nearshore to delimit the offshore extent of any species in our study. Published data on widely (offshore) ranging species are contained in the CalCOFI atlas series (Kramer and Ahlstrom 1968; Ahlstrom 1969, 1972; Ahlstrom and Moser 1975) and complement some of the offshore patterns report- ed here. Engraulis mordax, one of these widely ranging species, spawns principally offshore (Richardson 1981; Brewer and Smith 1982). The number of excess E. mordax larvae (over those which can be accounted for by eggs) in the nearshore zone must come from outside the sampling area, and these lar- vae must begin moving shoreward at an early age. Richardson (1981) suggested that currents might be a mechanism through which larvae of the northern subpopulation of E. mordax are redistributed. We presently cannot identify a mechanism for the redis- tribution off San Onofre. However, if one assumes it involves some behavioral response to environmental cues, it is worth considering just how far a larval anchovy might swim. Hunter (1972) estimated cruis- ing speed on the order of one-half body length/s. At this speed, a 6 mm larva would swim about 250 m/d, far enough to move several kilometers along an environmental gradient during the larval period. Any behavior allowing larvae to remain in the nearshore zone (e.g., orientation toward the bottom), once encountered, could help explain their observed concentration. The increased concentration of older larvae of E. mordax, Genyonemus lineatus, and Seriphus politus nearshore and near the bottom is reminiscent of the invasion and retention of larval and postlarval fishes in estuaries and tidal creeks of the Atlantic coast (cf. Chao and Musick 1977; Weinstein et al. 1980). Older larvae of Paralichthys californicus, although too rare for statistical analysis, also appeared more concen- trated nearshore than did the younger larvae. Whatever the mechanisms for such ontogenetic redistribution, they must be at least partly behavioral. Weinstein et al. (1980) found vertical movements in response to tides, whereby postlarvae became more concentrated near the bottom during ebb flows, thus taking advantage of the slower seaward current in the boundary layer. In the Southern California Bight the mean nearshore flow is alongshore, with relatively weak cross-shelf com- ponents (Hendricks 1977; Reitzel 197911; Parrish et al. 1981; Winant and Bratkovich 1981). The major source of cross-shelf water motion is internal waves of tidal frequency (Winant and Olson 1976) which propagate toward shore. For these waves to pro- pagate, the water column must be stratified. It is not- able that larval S. politus, which displayed the most intense ontogenetic redistribution, is most abundant during late summer-early fall (Walker et al. foot- note 11), the season of maximum thermal stratification in the Bight (Cairns and Nelson 1970). Thus it may be that S. politus and other semi- planktonic organisms of the shallow shelf waters take advantage of internal tides in somewhat the same way that the estuarine fauna use the surface tide to regu- late position. It is conceivable that due to dissipation of energy, seaward motions in the boundary layer are slower than shoreward motions. A similar internal wave mechanism for shoreward migration has been suggested by Norris (1963). He hypothesized that postlarval Girella nigricans might swim ahead of the cold waters of the incoming inter- nal wave fronts, thus producing the observed early shoreward migration of that species. Brewer and Smith (1982) estimated that the num- bers of E. mordax larvae spawned in the nearshore waters were approximately proportional to the area the nearshore waters represented in the total waters inhabited by the central subpopulation. They con- cluded that the nearshore region off southern "Reitzel, J. 1979. Physical/chemical oceanography. In Interim report of the Marine Review Committee to the California Coastal Commission. Part II: Appendix of technical evidence in support of the general summary. MRC Document 79-02(11), p. 6-23. Marine Review Committee of the California Coastal Commission, 631 Howard Street, San Francisco, CA 94105. 107 FISHERY BULLETIN: VOL. 82, NO. 1 California was not a preferred habitat for adult spawning during 1978-80. Our ratios of E. mordax eggs to early larvae support this conclusion. On the other hand, larval survivorship may be enhanced in these nearshore waters. Hjort (1914), Lasker (1975), and Brewer and Smith (1982) pointed out that the number of eggs and larvae surviving to recruitment may vary independently of spawning stock size. Brewer and Smith (1982) indicated that the shallow coastal region's importance as a nur- seryground for E. mordax is not yet clear. Their pre- liminary length-frequency data show relatively high numbers of large size classes nearshore, which are rare further offshore. Our preliminary length- frequency data corroborate this. The onshore ontogenetic shift of these larvae is a conspicuous and persistent feature of our data set (fig. 5). Thus nearshore environmental conditions may enhance growth or survivorship or both fori?, mordax larvae as well as for other larvae with typically inshore patterns. The larval taxa discussed in this paper represent some 12' ? of the types identified in the course of this study. Less common taxa were omitted for statistical reasons, but inspection of the data suggests that the patterns of abundance shown here are typical. Lar- vae of many species found in our study are most abundant in shallow water within a few kilometers from shore. Laroche and Holton (1979), noting the inshore abundance of 0-age Parophrys vetulus off the Oregon coast, suggested a nusery function for those open, nearshore areas. Concentration of juvenile fishes well inshore of adult depth ranges is also well known along the southern California coast (Lim- baugh 1961; Feder et al. 1974). Whether such patterns result from behavioral mechanisms leading to nearshore concentration, from differential onshore-offshore mortality, or sim- ply from random movements away from very localized spawning sites, their evolution and main- tenance imply significant value in occupying nearshore waters. Eppley et al. (1978) found higher concentrations of phytoplankton inshore of the 50- 100 m depth contours, and Lasker (1975, 1978) showed that nearshore abundance of suitable-sized phytoplankton can be an important determinant of year-class strength in E. mordax. Gruber et al. (1982) noted that Pacific sardine, Sardinops caeruleus, once spawned over wide areas of the California Current region, but the reduced stock now concentrates its spawning effort nearshore. They suggested the pro- ductive nearshore zone may be especially important to recovering fish stocks, a situation which might apply to northern anchovy at some future date. Pearcy and Myers (1974) noted that a number of studies found estuaries of northern California and Oregon to be important nurseries. However, estuaries in the Southern California Bight, as along much of the Pacific coast of North America, are small and far between. Enhanced productivity in the shallow waters of the open coast seems to provide a nursery area for many Southern California fishes analogous to the estuarine nurseries of other regions. ACKNOWLEDGMENTS This paper is a result of research funded by the Marine Review Committee (MRC), Encinitas, Calif. The MRC does not necessarily accept the results, findings, or conclusions stated herein. We are indebted to Jeffrey M. Leis for his important contributions to all parts of the preliminary study and to the field and laboratory aspects of the main study. Susan Watts provided invaluable assistance in the computer analysis of the data. Keith Parker and Allen Oaten assisted with the statistical problems encountered. Paul Smith offered many helpful sug- gestions on a manuscript dealing with the pre- liminary study. Edward DeMartini, H. J. Walker, Jr., and Robert J. Lavenberg read earlier versions of this manuscript and offered useful comments. The paper has also benefitted from the comments of an anonymous reviewer. Judy Sabins, Carolyn Davis, and Karen Lee typed the various versions of the manuscript. We especially wish to thank the many technicians who spent long hours in the collection and processing of samples. LITERATURE CITED Ahlstrom, E. H. 1959. Vertical distribution of pelagic fish eggs and larvae off California and Baja California. Fish Wildl. Serv., Fish. Bull, U.S. 60:107-146. 1965. Kinds and abundance of fishes in the California Current region based on egg and larval surveys. Calif. Coop. Oceanic Fish. Invest. Rep. 10:31-52. 1969. Distributional atlas of fish larvae in the California Current region: jack mackerel, Trachurus symmetricus, and Pacific hake, Merluccius productus, 1951 through 1966. Calif. Coop. Oceanic Fish. Invest. Atlas 11, xi + 187 charts. 1972. Distributional atlas of fish larvae in the California Current region: six common mesopelagic fishes — Vin- ciguerria lucetia, Triphoturus mexicanus, Stcnobrachius leueopsarus, Leuroglossus stilbius, Bathylagus wesethi, and Bathylaxus miuilrnsis. 1955 through 1960. Calif. Coop. Oceanic Fish. Invest. Atlas 17, xv + 306 charts. Ahlstrom, E. H., and h. G. Moser. 1975. Distributional atlas of fish larvae in the California 108 BARNETT ET AL.: DISTRIBUTION OF ICHTHYOPLANKTON OFF SAN ONOFRE, CALIF. Current region: flatfishes, 1955 through 1960. Calif. Coop. Oceanic Fish. Invest. Atlas 23, xix + 207 charts. Brewer, G. D., R. J. Lavenberg, and G. E. McGowen. 1981. Abundance and vertical distribution of fish eggs and larvae in the Southern California Bight: June and October 1978. In R. Laskerand K. Sherman (editors), Symposium on the early life history of fish. Introduction and back- ground, Woods Hole, April 1979, Vol. 178, p. 165- 167. Rapp. P.-V. Reun. Cons. Int. Explor Mer. Brewer, G. D., and P. E. Smith. 1982. Northern anchovy and Pacific sardine spawning off southern California during 1978-1980: preliminary' obser- vations on the importance of the nearshore coastal re- gion. Calif. Coop. Oceanic Fish. Invest. Rep. 23:160-171. Brown, D. M., and L. Cheng. 1981. New net for sampling the ocean surface. Mar. Ecol. Prog. Ser. 5:225-227. Cairns. J. L., and K. W. Nelson. 1970. A description of the seasonal thermocline cycle in shallow coastal water. J. Geophys. Res. 75:1 127-1131. CHAO, L. N., AND J. A. MUSICK. 1977. Life history, feeding habits, and functional morphology of juvenile sciaenid fishes in the York River estuary, Virginia. Fish. Bull.. U.S. 75:657-702. Eppley, R. W., C. Sapienza, and E. H. Renger. 1978. Gradients in phytoplankton stocks and nutrients off southern California in 1974-76. Estuarine Coastal Mar. Sci. 7:291-301. Feder, H. M.. C. H. Turner, and C. Limbaugh. 1974. Observations on fishes associated with kelp beds in southern California. Calif. Dep. Fish Game, Fish Bull. 160, 144 p. Gruber, D., E. H. Ahlstrom, and M. M. Mullin. 1982. Distribution of ichthyoplankton in the Southern California Bight. Calif. Coop. Oceanic Fish. Invest. Rep. 23:172-179. Hendricks, T. J. 1977. Coastal currents. In Southern California Coastal Water Research Project, Annual Report, p. 53-62. El Segundo, Calif. H.JORT, J. 1914. Fluctuations in the great fisheries of northern Europe viewed in the light of biological research. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 20:1-228. Hunter, J. R. 1972. Swimming and feeding behavior of larval anchovy Engraulis mordax Fish. Bull., U.S. 70:821-838. ICANBERRY. J. W., J. W. WARRICK, AND D. W. RlCE, JR. 1978. Seasonal larval fish abundance in waters off Diablo Canyon, California. Trans. Am. Fish. Soc. 107:225-233. Kramer, D., and E. H. Ahlstrom. 1968. Distributional atlas of fish larvae in the California Current region: northern anchovy, Engraulis mordax Girard, 1951 through 1965. Calif. Coop. Oceanic Fish. Invest. Atlas 9, xi + 268 charts. Ladell, W. R. S. 1936. A new apparatus for separating insects and other arthropods from the soil. Ann. Appl. Biol. 23:862-879. Laroche, W. A., and R. L. Holton. 1979. Occurrence of 0-age English sole, Parophrys vetulus, along the Oregon coast: an open coast nursery area? Northwest Sci. 53:94-96. LASKER, R. 1975. Field criteria for survival of anchovy larvae: The rela- tion between inshore chlorophyll maximum layers and successful first feeding. Fish. Bull., U.S. 73:453-462. 1978. The relation between oceanographic conditions and lar- val anchovy food in the California Current: Identification of factors contributing to recruitment failure. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 173:212-230. Limbaugh, C. 1961. Life-history and ecologic notes on the black croaker. Calif. Dep. Fish Game 47:163-174. Methot, R. D., and D. Kramer. 1979. Growth of northern anchovy, Engraulis mordax, larvae in the sea. Fish. Bull., U.S. 77:413-423. Morrison, D. F. 1976. Multivariate statistical methods. 2d ed. McGraw- Hill, N. Y., 415 p. Mcrphy, G. I., and R. I. Clutter. 1972. Sampling anchovy larvae with a plankton purse seine. Fish. Bull., U.S. 70:789-798. Myers, J. L. 1972. Fundamentals of experimental design. 2d ed. Allyn and Bacon, Boston, 465 p. Norris, K. S. 1963. The functions of temperature in the ecology of the per- coid fish Girella nigricans (Ayres). Ecol. Monogr. 33:23- 62. Parrish, R. H., C. S. Nelson, and A. Baki n. 1981. Transport mechanisms and reproductive success of fishes in the California Current. Biol. Oceanogr. 1:175- 203. Pearcy, W. G., and S. S. Myers. 1974. Larval fishes of Yaquina Bay, Oregon: A nursery ground for marine fishes? Fish. Bull., U.S. 72:201-213. Richardson, S. L. 1981. Spawning biomass and early life of northern anchovy, Engraulis mordax, in the northern subpopulation off Oregon and Washington. Fish. Bull., U.S. 78:855-876. Richardson, S. L., and W. G. Pearcy. 1977. Coastal and oceanic fish larvae in an area of upwelling off Yaquina Bay, Oregon. Fish. Bull., U.S. 75:125-145. Smith, P. E., R. C. Counts, and R. I. Clutter. 1968. Changes in filtering efficiency of plankton nets due to clogging under tow. J. Cons. Int. Explor. Mer 32:232-248. Smith, P. E., and S. L. Richardson. 1977. Standard techniques for pelagic fish egg and larva sur- veys. FAO Fish. Tech. Pap. 175, 100 p. Snedecor, G. W., and W. G. Cochran. 1967. Statistical methods. 6th ed. Iowa State Univ. Press, Ames, 593 p. SOKAL, R. R., AND F. J. ROHLF. 1969. Biometry. The principles and practice of statistics in biological research. W. H. Freeman, San Franc, 776 p. Weinstein, M. P., S. L. Weiss, R. G. Hodson, and L. R. Gerry. 1980. Retention of three taxa of postlarval fishes in an inten- sively flushed tidal estuary. Cape Fear River, North Car- olina. Fish. Bull., U.S. 78:419-436. WlNANT, C. D., AND J. R. OLSON. 1976. The vertical structure of coastal currents. Deep-Sea Res. 23:925-936. Winant, C, D., and A. W. Bratkovich. 1981. Temperature and currents on the Southern California shelf: a description of the variability. J. Phys. Oceanogr. 11:71-86. Winer, B. J. 1971. Statistical principles in experimental design. 2d ed. McGraw-Hill, N.Y., 907 p. Zweifel, J. R., AND R. Lasker. 1976. Prehatch and posthatch growth of fishes— a general model. Fish. Bull., U.S. 74:609-621. 109 APPENDIX 1 FISHERY BULLETIN: VOL. 82, NO. 1 On 24 July, 30 August, and 22 September 1978, vertically stratified samples were taken at one station along the 8 m isobath and at another along the 13 m isobath. A sample set, or profile, consisting of five strata was sampled at each station: Neustonic, three midwater strata, and the epibenthic layer (the mid- water strata were chosen with regard to the depths of power plant cooling structures). At the 8 m station, the midwater strata were 1) the lower 3 m of the water column, 2) 3 m above the bottom, and 3) the water column above stratum 2. At the 13 m station, the lower midwater stratum was the lower 2 m of the water column, while the upper two depended on the vertical thermal structure. When a thermocline was present, as during the September cruise and inter- mittently during the August cruise, the middle stratum extended from 2 m above the bottom to the base of the thermocline, and the upper stratum from the top of the thermocline to just below the surface. In the absence of a well-defined thermocline, the water column above 2 m from the bottom was divided into two equal parts. Sample sets were replicated four to six times in the day and again at night, result- ing in 325 samples in the vertical migration study. Data from the two stations were analyzed separately, since all sampling depths (except the neustonic layer) differed between stations. No analysis was done of the effects of the thermocline, since its extreme movements with respect to the ver- tical scale of interest would require a more intensive sampling program. In this analysis nominal sampling depths were treated as constants. Because of patchy distributions of ichthyoplankton and movements of the thermocline (August and Sep- tember), inherent variability was expected among the sets of profiles taken on a given date. In order to separate this variability from variability due to sam- pling date (cruise), time of day, and "error", we analyzed the data in a repeated-measures type analysis of variance design (App. Table 1). In this design, the depth effect was contained within the fixed-effect time of day and the random-effect cruise. The questions addressed were 1) whether there was a depth effect, i.e., significant differences among strata, within cruise X time-of-day blocks, and 2) if a depth effect did exist, whether there was a significant depth X time-of-day interaction. This interaction, interpreted (when significant) as daily vertical migra- tion, was evaluated as the F-ratio of the depth X time of day to the depth X time of day X cruise mean square errors. When the three-way term was insignificant (in this case, P > 0.75), the error sums of squares and the three-way sum of squares were pooled, and this pooled term was used as the denominator in the F-ratio (Sokal and Rohlf 1969: 266). The 10 most frequently occurring taxa were analyzed (App. Table 2). (A high frequency of occurrence was important to keep cell variances relatively homogeneous.) To reduce the effect of day- night differences in apparent abundance (most likely from visual net avoidance), we reduced each profile to a set of differences, or A's between adjacent strata, e.g. A, = (abundance at depth 1) — (abundance at depth 2). Abundance was expressed as log,,, (A' +1), where X = larvae/1 00m'. Any daily change in the relative abundance in two strata would thus be manifest in a change in sign and/or magnitude of the correspond- ing A. Appendix Table 1 .— ANO VA model applied in the analysis of daily vertical migration. The last two terms can form the error estimate (e) in Appendix Table 2. i/ km ' where V, + CTM + CD I'* + TDm + DP[mlllJk] + CTD(l/k) + e,lkm i /km M c, oT OPtrnt,,lk) crbm i/km Density Mean effect Sampling date (cruise) effect Time-of-day effect (day-night) Depth profile within cruise and time-of-day Depth effect Interaction, cruise X day-night period Interaction, cruise X depth Interaction, day-night period X depth Depth k for profile m within cruise and time-of-day Interaction, cruise X day-night period X depth Residual error 110 BARNETT ET AL.: DISTRIBUTION OF ICHTHYOPLANKTON OFF SAN ONOFRE. CALIF. Appendix Table 2. — F-table for the 1 0 most frequently occurring larval fish taxa off San Onofre, Calif.: repeated-measures ANOVA. D = depth, TD = day-night period X depth, CTD = cruise X day/night period X depth. When the CTD mean square error (MSE) was insignificant (P > 0.75), the CTD and Error (e) sums of squares were pooled. The TD interaction term, when significant (*P < 0.05; **P < 0.0 1 ). is interpreted as daily vertical migration. Frequency refers to the number of samples in which a taxon occurred out of 325 total samples. Results are presented for the 13 m station only. Taxa Freq. Source df MSE Engraulis mordax Senphus polttus Hypsoblenmus spp 251 232 206 Genyonemus hneatus^ 148 Cheilotrema salurnum 1 44 Menticirrhus undulatus 125 Paralabrax spp. 122 Paratichthys califormcus 1 1 9 Gibbonsta Type A Gobiesox rbessodon 114 113 D 3 0.79065 4.101 0.010 TD 3 0 18989 0335 0.801 CTD 6 056687 2 940 0.013 f 69 0.19281 D 3 1.721 10 8 943 0.9). An examina- tion of the individual percent shrinkages showed no trend with live standard length. Frozen lengths were corrected to live lengths by multiplying by the factor 114 McGURK: PACIFIC HERRING OTOLITH RINGS 1.063. Alcohol-preserved lengths did not require correction. Ring Counting After extraction from the skull the sagittae were placed on a glass slide under immersion oil; their diameters were measured with an ocular micrometer. Sagittae are slightly flattened spheroids in young lar- vae and tend to become more oval in shape as the fish grows. The diameter measured was always the long- est axis of the otolith. The sagittae were photo- graphed at 400-1, 000X, the developed film was projected on a screen, and the rings were counted. A single ring consisted of a dark band and an adjacent light band. All rings, no matter how faint, were coun- ted in order to avoid observer bias towards a daily ring pattern. Two classes of rings were observed: 1) A group of 1-5 thin, faint rings clustered about the nucleus surrounded by 2) wider, darker rings that composed the majority of the rings in most larvae. In some sagittae the second class of rings were separated from the first by a distinct ring which may have been a check deposited in response to the exhaustion of the yolk. The two classes could not always be clearly distinguished, particularly in slow- growing fish. The first class corresponds to Geffen's (1982) "yolk sac" rings and the second to her "nor- mal" or "regular" rings. In this paper the first class will be unnamed for two reasons: 1) Most of the rings were found in the larvae that had completely absorbed their yolk, so they were not exclusively yolk-sac rings, and 2) it has not been established that the two classes of rings are fundamentally different from each other, so the introduction of new terminol- ogy is premature. Geffen (1982) defined a "first heavy ring" that was found between the outer margin of the nucleus and the first normal ring. This term has not been used because the first normal ring was not always distinguishable from subsequent normal rings on the basis of width or darkness. Each sagitta was counted three times, and the mean of the three counts was taken as the final count of that sagitta. The ring count of a fish was the mean of the final counts of its two sagittae. The mean (± 1 SD) dif- ference in final counts between sagittae from the same fish was 1.3 ± 1.4 which was not significantly different from zero (t = 0.9028, df = 148, 0.4 > P > 0.2). The sagittae of 21 large larvae (live length range = 14-29 mm, age range = 20-54 d posthatch) se- lected at random from several groups were photo- graphed and then fixed to a glass slide with cyanoacrylate glue and ground to the midplane with metallic lapping paper. They were rephotographed and recounted. The mean (± 1 SD) difference was 1.1 ± 2.0 which was not significantly different from zero (t = 0.5273, df = 20, 0.5 > P > 0.9). Inspection of the data revealed no trend of the difference with age or with the ring count of the nonground sagittae. Data Analysis The average rates of ring deposition and of growth in length were calculated as the slopes of linear pre- dictive regressions of mean ring number and mean length on age posthatch. The homogeneity of the variances of the means of a group was tested with Bartlett's test (Sokal and Rohlf 1969), and, if they were found to be heterogenous, each mean was weighted with its sample size divided by its variance. T-tests were used to test the significance of differ- ences between the slope of a regression of mean ring number on age and 1 ring/d and 0 ring/d. F-tests were used in covariance analyses to test for significant dif- ferences between two slopes. RESULTS Growth in live standard length was positive in all groups except 1980C and 1980B, in which the starv- ing larvae shrank (Fig. 1). There are indications that growth was curvilinear, especially in 1980A and 198 IB where the growth rates between the two last sampling dates in each group were much less than the previous growth rates. However, linear growth was assumed for the purpose of obtaining average growth rates to compare with the average ring deposition rates (Table 2). Growth rate was highest in the 2,000 1 culture chamber and lowest in the 25 1 aquarium, and there was a positive but nonsignificant correlation between growth rate and container size in the four fed groups (n = 4, r = 0.90, 0.05 > P > 0.10). Thin, faint rings of the first class were found in the otoliths of most of the 1980 fish that were < 14 mm long, but were not found in the otoliths of any 1981 and 1982 fish (Fig. 2). These rings may have been deposited at any time between the late embryo and the postyolk-sac stage. The only sample of otoliths TABLE 2. — Linear regressions of mean standard length on age in ' groups of Pacific herring larvae. /-intercept Slope SE of No. of Group (mm) (mm/d) slope r means n df 1980A 10 4 0 180 0030 097 4 36 1.2 1980B 13.1 -0.004 0.019 0 19 3 20 1,1 1980C 11.2 -0.107 0.031 0.90 5 50 1.3 1981 A 82 0.231 0.011 0.99 6 57 1.4 1981B 8.4 0290 0.049 0.96 5 60 1.3 1982A 10.6 0.090 0.047 0.89 3 38 1.1 1982B 1 1.4 0.100 0.035 0.89 4 39 1,2 115 FISHERY BULLETIN: VOL. 82, NO. 1 30 26 22 ia 14 - — « 10 b ■» — ' 3 0 T H 2h (3 z 22 HI 1 ia Q CC 14 < a 10 z < b \- rn in 1980B 1980C 1981B 40 60 0 20 AGE (DAYS) Fic.i'RE 1.— Mean (± 1 SD) live standard length at age posthatch for seven groups of Pacific herring larvae. See Table 2 for the regres- sion equations. from yolk-sac larvae was a single sample from 1980D that had a mean (± 1 SD) ring count of 5.2 ± 0.8 (n = 9) on day 1 posthatch. The rings were not observed in older, larger larvae; they may have been present but obscured by overburden over the nucleus. This phenomenon has been observed in the otoliths of lar- val largemouth bass, Micropterus salmoides, (Miller and Storck 1982). A group of 7-8 "prolarval rings" that were clustered about the nucleus at swim-up were visible for only 10-15 d afterward, because the nucleus became more opaque with age. The first normal ring was deposited in all groups including 1980C by day 6 posthatch, the day after complete yolk absorption. This agrees well with the age at first increment of 4.5 (range = 0-9 d) found for Atlantic herring by Lough et al. (1982) and with the age of 6 d found for the same species by Geffen (1982). This indicates that herring larvae of both species do have a fixed age at first increment deposi- tion and that it coincides with the age at complete yolk absorption. Rates of subsequent ring deposition for the four fed 40 60 AGE (DAYS) FIGURE 2.— Mean (±1 SD) ring count at age posthatch for seven groups of Pacific herring larvae. Open circles are total rings and closed circles are normal rings only. See Table 3 for the regres- sion equations. groups were not all daily, and they ranged from 0.12 to 0.96 rings/d (Table 3); only two groups, 1980A and 198 IB, had rates that were not significantly different from 1 ring/d (t = 0.5772, df = 3, 0.5 > P> 0.9 andt = 2.0142, df = 4, 0.10 > P > 0.20, respectively). The 1981A group had a rate that was significantly <1 ring/d (t = 6.3465, df=5, 0.01 >P> 0.001) butalso significantly > 0 (t = 10.8062, df = 5,P< 0.001) and the 1982 A group had a rate that was significantly < 1 ring/d {t = 10.0228, df = 2, 0.01 > P> 0.001) andnot significantly >0 (t = 1.3667, df = 2, 0.20 > P > 0.40). The rate of ring deposition in 1980C, the group that was starved from hatch, was —0.05 ring/d, which was Table 3. — Linear regressions of mean normal ring number on age in 7 groups of Pacific herring larvae. ^-intercept Slope SE of No. of Group (mm) (nng/d) slope T means n df 1980A -4 12 096 0.06 0 99 4 36 1.2 1980B 2.06 023 0.28 063 3 20 1.1 1980C 2 12 -0.05 002 0.83 5 50 1.3 1981A -931 0.63 0.05 0.99 6 57 1.4 1981B -5 60 0.83 0 08 099 5 60 1.3 1982A 1.45 0 12 0.08 0.83 3 38 1.1 1982B 4.90 0 10 0 11 053 4 39 1.2 116 McGURK: PACIFIC HERRING OTOLITH RINGS not significantly different from 0 (t = 2.2831, df = 4, 0. 10 > P > 0.20). This indicates that the starvation of first-feeding larvae stopped ring production. The 1980B group had a rate which was not significantly different from one of 1 ring/d (t = 2.3397, df = 2, 0.10 > P > 0.20) and not significantly different from a rate of 0 (t = 0.6989, df = 2, 0.50 > P > 0.90) or from the rate of its parent feeding group, 1980A (F = 5.9185, df = 1.3, 0.25 > P > 0.50). One reason for these results is that the 1980B group had only three data points for the regression, and the standard error of the slope was therefore relatively high: 122% of the value of the slope (Table 3). I conclude that starva- tion for 5 d after a feeding period of 6-7 d has no effect on the rate of ring deposition. The 1982B group had a ring depoistion rate that was not significantly dif- ferent from 0 (t = 0.7843, df = 3, 0.40 > P> 0.50) and which was not significantly different from the rate of its parent feeding group, 1982A (F = 0.1352, df = 1, 3, P > 0.75). I conclude that starvation for 8 d after a feeding period of about 25 d has no effect on the rate of ring deposition, at least not in 25 1 enclosures. The average ring deposition rates were significantly positively correlated with the average growth rates (n = l,r = 0.83, 0.01 > P > 0.05) (Fig. 3). The regres- sion of ring rate on growth rate was: Ring rate — 0.14 + 2.40 (growth rate). The residuals of this regression were not correlated with container size, and there was no obvious relationship with prey type. However, there was a significiant positive correlation between the residuals and the mean rearing temperature (n = 7,r = 0.83,0.01 >P> 0.02). The midpoints of the tem- perature range were used as an estimate of the mean temperature (Table 1). A regression of ring deposi- tion rate on growth rate and temperature increase the multiple r to 0.99: Ring rate = -1.39 + 3.36 (growth rate) + 0.14 (temperature). These results confirm the correlation between ring deposition rate and growth rate found for Atlantic herring larvae by Geffen (1982), who interpreted the relationship as being curvilinear and linearized it by transforming both variables with logarithms. In order to compare the two sets of data the relationship be- tween ring deposition rate and growth rate was assumed to be linear. A covariance analysis of the slopes of the two linear regressions indicated that there was no significant difference between them at the 0.05 probability level. Data from this study and from Geffen's were pooled and a single linear regres- sion was calculated (n = 12, r = 0.85, P < 0.001): Ring rate = 0.17 + 2.12 (growth rate). < Q U z cr LU i- < DC O z DC 1.2 1.0 0.8 0.6 0.4 0.2 0.0 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4 0.5 GROWTH RATE (MM/DAY) FIGURE 3. — Relationship between the average ring deposition rates and the average growth rates of seven groups of Pacific herring lar- vae. See text for regression equation. The influence of temperature on ring deposition rate could not be compared between the two data sets because the rearing temperature for Geffen's fish was not constant over the rearing period. Plots of fish length on otolith diameter for the seven populations were curvilinear, and the rate of growth of fish length decreased with increasing otolith diameter. Transforming otolith diameter with logarithms best linearized the data, transforming both variable with logarithms produced lower cor- relation coefficients in all groups. Thus length was regressed on log (otolith diameter) (Table 4, Fig. 4). An analysis of covariance that included all seven groups indicated that the slopes of the regres- sions were significantly different from each other at the 0.05 probability level. Inspection of the slopes and their standard errors indicated that the fed groups and 1980B had slopes of a similar value and that 1980C and 1982B had slopes of a similar value but that they were much lower than those of the fed groups. The two groups were subjected to separate covariance analyses, and in each group the slopes were found to be not significantly different from each other at the 0.05 probability level. The 117 FISHERY BULLETIN: VOL. 82, NO. 1 Table 4.— Linear regressions of fish length on log (oto- lith diameter). y-intercept Slope SE of Group (mm) |mm fxm) slope r n df 1980A -5.76 1 1 57 0.49 097 36 1.34 1980B -4 54 10 77 3 17 0.73 1.' 1,10 1980C 2 73 4.40 2.10 0.28 52 1.50 1981A -7.50 13.36 043 0.97 57 1.55 1981B -5.50 .' 14 0.46 096 il() 1.58 1982A -7 24 12.73 1 45 083 5-8 d is necessary in order to demonstrate a statistically significant effect. Larger rearing con- tainers are also recommended to produce greater contrast in growth rates between feeding and starv- ing fishes. Container size, temperature, or prey size may possibly have additional effects on the rate of ring deposition apart from that which is explained by growth rate. Temperature does explain some of the residual variance of the ring deposition rate-growth rate regression. However, published evidence on effect of constant temperature on ring deposition does not support the hypothesis that higher tem- peratures produce more increments. For example, Neilson and Geen (1982) found no difference be- tween the number of increments produced by juvenile chinook salmon, Oncorhynchus tshawytscha, reared at 5.2°C and at 11.0°C. The effects of such environmental factors as light, temperature, and prey type on the ring pattern of herring sagittae can only be determined with a well-controlled ex- perimental study. ACKNOWLEDGMENTS I gratefully acknowledge the assistance given by the staff of the Bamfield Marine Station. I also thank Gary Kingston for assistance in rearing fish in 1980, Jeff Marliave for advice on rearing healthy marine fish larvae, and Steve Campana for discussions on grinding and reading fish otoliths. The paper has benefited substantially from reviews by two anonymous reviewers. This study was funded by a G.R.E.A.T. Award from the Science Council of British Columbia and by a grant to N.J. Wilimovsky from the National Sciences and Engineering Research Council of Canada. LITERATURE CITED Brothers, E. B. 1978. Exogenous factors and the formation of daily and sub- daily growth increments in fish otoliths. Am. Zool. 18:631. Brothers, E. B., C. P. Mathews, and R. Lasker. 1976. Daily growth increments in ototliths from larval and adult fishes. Fish. Bull., U.S. 74:1-8. Geffen, A. J. 1982. Otolith ring deposition in relation to growth rate in her- ring (Clupea harengus) and turbot (Scophthalmus max- ima*) larvae. Mar. Biol. (Berl.) 71:317-326. GJ0SAETER, H., AND S. TlLSETH. 1982. Primary growth increments in otoliths of cod larvae (Gadus morhua L.) of the Arcto-Norwegian cod stock. Fiskeridir. Skr. Ser. Havunders. 17(7):287-295. Haegele, C. W., and D. N. Outram. 1978. The effects of diet and ration on the growth and survival of Pacific herring (Clupea harengus pallasi) lar- vae. Fish. Mar. Serv. Can. Tech. Rep. 767, 41 p. Kendall, A. W., Jr., and D. Gordon. 1978. Growth rate of Atlantic mackerel (Scomber scombrus) larvae in the middle Atlantic Bight. Rapp. P.-V*. Reun. Cons. Int. Explor. Mer 178:337-341. Kramer, D., and J. R. Zweifel. 1970. Growth of anchovy (Engraulitt mordax Girard) in the laboratory as influenced by temperature. Calif. Coop. Oceanic Fish. Invest. Rep. 14:84-87. Laroche, J. L., S. L. Richardson, and A. A. Rosenberg. 1982 Age and growth of a pleuronectid, Parophrys vetulus, during the pelagic larval period in Oregon coastal waters Fish. Bull. U.S. 80:93-104. Lough, R. G., M. Pennington, G. R. Bolz, and A. A. Rosenberg. 1982. Age and growth of larval Atlantic herring, Clupea 119 FISHERY BULLETIN: VOL. 82, NO. 1 harengus L., in the Gulf of Maine-Georges Bank region based on otolith growth increments. Fish. Bull., U.S. 80:187-199. Marliave, J. B. 1981. Use of tidal power for culture of marine animals. Proc. 1981 Conf. Am. Assoc. Zool. Parks and Aquariums, New Orleans, La.. Sept. 13-17, 1981, p. 103-105. Marshall, S. L„ and S. S. Parkkr. 1982. Pattern identification in the microstructure of sockeye salmon {Oncorhynchus nerka) otoliths. Can. J. Fish. Aquat. Sci. 39:542-547. Methot, R. D., and D. Kramer. 1979. Growth of northern anchovy, Engraulis mordax, larvae in the sea. Fish. Bull, U.S. 77:413-423. Miller, S. J., and T. Storck. 1982. Daily growth rings in otoliths of young-of-the-year largemouth bass. Trans. Am. Fish. Soc. 1 1 1:527-530. NEILSON, J. D., AND G. H. GEEN. 1982. Otoliths of chinook salmon (Oncorhynchus tsha- wytscha): Daily growth increments and factors influencing their production. Can. J. Fish. Aquat. Sci. 39:1340- 1347. Radtke, R. L., and J. M. Dean. 1982. Increment formation in the otoliths of embryos, larvae, and juveniles of the mummichog, Fundulus hetero- clitus. Fish. Bull., U.S. 80:201-215. Ralston, S. 1976. Age determination of a tropical reef butterflyfish utiliz- ing daily growth rings of otoliths. Fish. Bull., U.S. 74:990- 994. SOKAL, R. R., AND F. J. ROHLF. 1969. Biometry. The principles and practice of statistics in biological research. W. H. Freeman, San Franc, 776 p. Taubert, B. D., and D. W. Coble. 1977. Daily rings in otoliths of three speices of Lepomls and Tilapia mossambica. J. Fish. Res. Board Can. 34:332- 341). Taylor, F. H. C. 1964. Life history and present status of British Columbia her- ring stocks. Fish. Res. Board Can., Bull. 143, 81 p. Theilacker, G. H. 1980. Rearing container size affects morphology and nutri- tional condition of larval jack mackerel, Trachurus sym- metricus. Fish. Bull., U.S. 78:789-790. TOWNSEND, D. W., AND J. J. GRAHAM. 1981. Growth and age structure of larval Atlantic herring, Clupea harengus harengus, in the Sheepscot River estuary, Maine, as determined by daily growth increments in otoliths. Fish. Bull., U.S. 79:123-130. UCHIYAMA, J. H., AND P. STRUHSAKER. 1981. Age and growth of skipjack tuna, Katsuwonnus pelamis, and yellowfin tuna, Thunnus albacares, as indicated by daily growth increments of sagittae. Fish. Bull., U.S. 79:151-162. Victor, B. C. 1982. Daily growth increments and recruitment in two coral- reef wrasses, Thalassoma bifasciatum, and Halichoeres bu-ittatus. Mar. Biol. (Berl.) 71:203-208. 120 FISHES, FISH ASSEMBLAGES, AND THEIR SEASONAL MOVEMENTS IN THE LOWER BAY OF FUNDY AND PASSAMAQUODDY BAY, CANADA J. Stevenson Macdonald,1, Michael J. Dadswell,2 Ralph G. Appy,3 Gary D. Melvin,2 and David A. Methven4 ABSTRACT Five fish assemblages, dominated by pleuronectids, cottids, gadids, clupeids, and rajids, were identified from collections taken during a 5-year survey in the lower Bay of Fundy region, Canada. Individual assem- blages occurred in each of estuarine, beach, pelagic, and offshore hard- and soft-bottom habitats. Species and/or age-class components within assemblages varied seasonally but, in general, each assemblage was dis- tinct. There was a progressive seaward displacement of these assemblages from shallow, inshore to deeper, offshore habitats in winter followed by a reversal during summer. Yearly changes in species occurrence and abundance during the study period were predominantly attributable to variation in ocean climate. Long-term changes in abundance of two commercial species at one of the sampling sites, since a similar study there in 1965, appear related to population fluctuations in the Bay of Fundy and the Gulf of Maine. The beach habitat apparently served as a major nursery area for juvenile gadids, pleuronectids, and clupeids. Although the fish fauna of the Bay of Fundy-Gulf of Maine system is well documented (Bigelow and Schroeder 1953; Leim and Scott 1966), few studies have examined long-term spatial and temporal changes or interrelationship among the fish assem- blages. Previous studies in this region were con- cerned with the biology and seasonal movements of a single species (McCracken 1959, 1963; McKenzie and Tibbo 1961; Wise 1962) or the occurrence and composition of communities at a single site (Bigelow and Schroeder 1939; Tyler 1971). Moore (1977) and Quinn (1980) have emphasized the need for long-term research to establish baseline information and estimates of natural variability for fisheries assessments and pollution impact studies. This is particularly true for inshore regions because of their importance as nurseries and feeding grounds (Warfel and Merriman 1944; Rauck and Zijlstra 1978). The increasing interest in trophic rela- 'Department of Zoology, University of Western Ontario, London, Ontario. Canada; present address: Department of Fisheries and Oceans, West Vancouver Laboratory, Vancouver, British Columbia, Canada V7V 1N6. 'Fisheries and Environmental Sciences, Department of Fisheries and Oceans, Biological Station, St. Andrews, New Brunswick, Canada EOG 2X0. 'Fisheries and Environmental Sciences, Department of Fisheries and Oceans, Biological Station, St. Andrews, New Brunswick, Canada; present address: Department of Zoology, College of Biological Sciences, University of Guelph, Ontario, Canada NIG 2W1. 'Fisheries and Environmental Sciences, Department of Fisheries and Oceans, Biological Station, St. Andrews, New Brunswick, Canada; present address: Department of Biology, Memorial Univer- sity, St. John's, Newfoundland, Canada A1B 3X9. tionships among entire communities of fishes is further reason to document movement, abundance, and co-existence of fishes potentially utilizing the same food resource (Richards 1963; Keast 1970; Tyler 1972; Steiner 1976; Hacunda 1981). Long-term changes in fish assemblages have been attributed to overexploitation of one or more of the species within the assemblage (Brown et al. 1973; Burd 1978; Sherman et al. 1981) and climatic variations (Dow 1964; Sutcliffe et al. 1977). However, it is usually difficult to separate natural fluctuations from those caused by imbalance in com- petitive and predator-prey relationships due to exploitation (Cushing 1980; Daan 1980; Sissenwine et al. 1982). With the view in mind of assessing these long-term changes to properly assign cause and effect, repetitive, in-depth studies of well-known or type localities are needed. This study examines spatial and temporal variation in fish diveristy and abundance over a 5-yr period at two offshore stations within Passamaquoddy Bay, one offshore station in the Bay of Fundy, and at inshore and beach stations in Passamaquoddy Bay. One offshore station was the same station sampled by Tyler (1971) during 1965-66, allowing documen- tation of changes that have occurred over the inter- vening 10-15 yr. METHODS Three offshore stations in the Bay of Fundy (B) and in Passamaquoddy Bay (A, C) (Fig. 1) were sampled Manuscript accepted August 1983. FISHERY BULLETIN: VOL. 82, NO. 1, 1984. 121 FISHERY BULLETIN: VOL. 82. NO 1 FIGURE 1.— Passamaquoddy Bay and the adjacent Bay of Fundy indicating sampling stations occupied during the study. at approximately monthly intervals over a 5-yr period, 1976-81 (Table 1). Station A was the same site sampled by Tyler (1971) during 1965-66. Fish were collected using a %-35 shrimp trawl (3.8 cm stretch mesh nylon; 15.5 m foot rope), similar to the %-35 Yankee trawl used by Tyler (1971), towed by the 1 50-hp, 14m stern trawler, Fisheries and Oceans' RV Pandalus II. Tows at each station were along a 1 .6 km transect at about 4 km/h. Stations A and B were sampled once per trip between 1976 and 1979, and station C was sampled sporadically. From 1979 to 1981, tows at stations A and B were replicated and station C was sampled regularly. Captured fishes were identified to species, and adults and juveniles were categorized by size and enumerated separately. During the final year of collecting, fork length of all fishes was recorded to the nearest centimeter and otoliths were collected from Atlantic cod, ocean pout, American plaice, winter flounder, and witch flounder for age determination. Atlantic cod otoliths were sec- tioned for aging, other species were aged using the whole otolith. Results reported are the empirical length at age. Between June and September 1976, 12 estuarine, intertidal, and inshore marine stations were sampled within Passamaquoddy Bay and Head Harbour Passage (Fig. 1). In addition, station 3 was sampled monthly during the period May 1976-November 1977, station 8 was sampled at approximately weekly intervals from May to September 1981, and stations 1 and 10 were sampled in December 1980 (Table 1). Fish were collected using a 9 m, 1.3 cm mesh beach seine, a 3.7 m shrimp trawl with a 3 mm cod end towed behind a 5 m Boston whaler, or bottom-set gill nets with stretched mesh sizes ranging from 7.6 to 17.8 cm. Standard fishing efforts employed with each gear type were shore seine hauls of 5 min during the 2-h period before and after low water, trawl tows of 10 min, and overnight gill net sets of 16 h. Temperature, salinity, and substrate type were 122 MACDONALD ET AL.: FISH ASSEMBLAGES AND THEIR SEASONAL MOVEMENTS Table 1.— Physical and chemical characteristics and sampling history of stations in the Bay of Fundy and Passamaquoddy Bay. Gear: ST = shrimp trawl; S = seine; GN = gill net. Bottom type: M = mud; Sa = sand; Rk = gravel or rock. Sampling Sampling Maximum temp. salinity depth Bottom range range Collection Sampling Station Gear (m) type (C) (%o) period trips A ST 80 M-Rk 0-15 29 5-32.5 1976-81 39 B ST 80 M 1-12 31 0-32 5 1976-81 37 C ST 20 M 0-15 — 1978-81 15 1 S 1.5 M-Rk 14.5-20.0 22.1-260 06-08/76. 12/80 3 2 S 1.5 M-Rk 15.5-22.5 26.0-29.5 05-08/76 4 3 S 1.5 Sa-Rk 0 0-16.0 21 0-30.0 05/76-11/77 16 4a S 1.5 Rk 12 5 29.0 06. 07/76 2 4b ST 7 5 Rk 12.5 — 07/76 1 5a S 1.5 Sa 14 5 30.0 07. 08/76 2 5b GN 33 Sa-M — — 08/76 1 6 S 1.5 Sa-M 14.0 28.0-300 08-09/76 2 7 GN 30 M 13.5 28.0 06/76 1 8a S 1.5 M-Sa 11.0-18.5 28.7-30.7 06. 07/76 05-09/81 2 23 8b ST 12 Rk-Sa — — 06, 07/76 2 9 S 1.5 Sa 140 29.5 06.08/76 2 10 GN 3 M-Rk 13.0 280 06. 09/76. 1 2/80 3 1 1 S 1 5 M — — 07/76 1 12a s 1.5 Sa-Rk 150 280 07, 09/76 2 12b ST 15 Sa-Rk — — 07/76 1 recorded for most sampling sites (Table 1). Bottom temperature and salinity data inside and outside Passamaquoddy Bay came from routine monthly sampling by the Department of Fisheries and Oceans at a site opposite the Biological Station (near Station A) and at "Prince 5" 3.2 km south of Bliss Islands in the Bay of Fundy (near station B). Temperatures at deep stations were taken with a reversing ther- mometer attached to a Nansen bottle and at shallow stations with a hand thermometer. Salinities were determined with a laboratory salinometer from sam- ples collected in the field. Substrate samples at deep stations were obtained with a PONAR grab. At shallow stations, substrate type was assessed visually. Fishes were identified using Leim and Scott (1966) with the exception of red and white hake and redfish, which were determined by using Musick (1973) and Ni (1982), respectively. Because we were unaware of the problem of distinguishing between young Raja ocellata and R. erinacea (McEachran and Musick 1973), these determinations may be incorrect. Coefficients of community were calculated using the formula: X 100 A + B-C where C = number of common species, A = number in assemblage 1, and B = number in assemblage 2 (Jaccard 1932; Kontkanen 1957). An index that com- pared presence and absence of species at each sta- tion (binary data) was used because species abundances among stations were not comparable due to different gear used. RESULTS AND DISCUSSION Station Environmental Characteristics Temperature and salinity at stations A and B (Fig. 2) followed the typical, yearly cycle of a cold tem- perate sea (Fig. 3). Annual temperature range in the Bay of Fundy was less than in Passamaquoddy Bay. Summer temperatures at inshore sites were simlar to offshore sites with the exception of higher tem- peratures at some estuary stations (i.e., 1 and 2) (Table 1). Two notable variations occurred: The early months of 1977 and August 1978 were abnor- mally warm, particularly at station A ( J. Hull5); and throughout the study period there was a generalized cooling trend. Salinities were highest in late summer through the fall and lowest in spring at both sites. At all times of year, salinities were higher in Bay of Fundy (station B) than at station A (Fig. 2). Inshore sites had salinities of 1-2 ppt less than station B, and salinities at estuarine sites were as low as 2 1 .0 ppt during sum- mer (Table 1). Substrates of most sites were composed of sand and/or mud (Table 1). Station A had the steepest slope, about 2: 100 m. Slopes at stations B and C were 0.4:100 and 0.6:100 m, respectively. Slopes at coastal intertidal sites were gradual, about 1:100 m. Estuarine stations (1,2, and 10) had extremely soft mud bot- toms and station 2 had extensive eel grass beds. 'Fisheries and Environmental Sciences, Department of Fisheries and Oceans, Biological Station, St. Andrews, New Brunswick, Canada EOG 2X0. 123 FISHERY BULLETIN: VOL. 82, NO. 1 Figure 2.— Bottom tem- perature and salinities at station A in Passama- quoddy Bay and station B in the Bay of Fundy during 1976-81. ° a 6+ age group), but catches from November to May were 6-25 cm juveniles (0-6 yr) (Figs. 6, 7). Adult witch flounder on the Scotian Shelf also move from intermediate depths (100 m) in summer to deeper water in winter (Powles and Kohler 1970). Both Powles and Kohler (1970) and Markle (1975) reported juvenile witch flounder from deep water (150-1,000 m) over hard bottom, quite unlike the situation we encountered except for similar tempera- ture regimes. Also, replacement of adults by juveniles during winter seems peculiar to our study, but may have been observed because of year- round sampling. Juvenile American plaice, Hippoglossoides pla- tessoides, were a major summer component of station C and a regular component of the Bay of Fundy sta- tion (Fig. 4), both soft-bottom habitats, but was only occasional at the hard-bottom station (A). Age-2 plaice (6-14 cm; Fig. 6) were first captured with our shrimp net in April. By the following year, recruit- ment to the gear appears complete at an average size for the age-class of 17 cm (Fig. 7). Juvenile plaice are 127 FISHERY BULLETIN: VOL. 82, NO. 1 60i- Glyptocephalus cynoglossus Pseudopl»uronect«s smerlcanus Hlppoglossoldas platessoldus Figure 6.— Fish length versus age for five fish species caught at stations A and B; December 1980-June 1981. Lines are fitted by eye. I^U ** 100 " Gadus morhua s / / / 80 A / 60 ./ ^^^ 40 — / / / / • ,jr Macrozoarcas amerlcanus 20 n / / i 1 ,i,i i,l 6 8 10 Age (Years) 12 14 16 20r 10 Psaudoplauronectat amerlcanus rffTl 1 1 1 rn rhi a n= 83 20 1 ■ r> = 2 . d n = 128 10 0 — .ccaJllL I Glyptocaphalus cynoglossus Jun* 23, 1980 -i 1 1 1 — I i Aug. 13, 1980 ■ n= 32 r Oct. 9, 1980 ■ Dsll " T i , i Dec. 13, 1980 ■ n = 17 W — ¥— "I 1 — *i Jan. 15, 1981 - ■ n - 16 ^ 1 1 1 1 April 29, 1981 ■ n = 2 Hlppoglossoldas platassoidas |-| ■ n - 60 i 1 — ^^* ^*T 1 1 X ■ n = 44 o n = 2 ■ n= 27 o n = 1 T 1 1 n= 77 ■ n = 30 a n = 3 n= 76 n= 7 ^Afc. — i r i 1 1 1 20 40 60 0 20 40 60 0 20 40 60 Length (cm) Figure 7.— Seasonal size distributions of flat- fishes (Pseudopleuronectes americanus, Glyp- tocephalus cynoglossus, and Hippoglossoides platessoides) from offshore stations in the Bay of Fundy and Passamaquoddy Bay, 1980 and 1981. Shaded area is captures at station B; un- shaded, station A inside Passamaquoddy Bay. 128 MACDONALD ET AL.: FISH ASSEMBLAGES AND THEIR SEASONAL MOVEMENTS sedentary, soft-bottom dwellers, that exhibit little seasonal movement, and migration from nursery ground to adult stock is diffusive (Bigelow and Schroeder 1953; Leim and Scott 1966). However, some seasonal movement does occur when plaice leave soft-bottom, middepth habitat (30 m) for win- ter and return in summer (present study). Plaice were a regular, low-abundance component at station A in 1965 (Tyler 1971), but we found they were virtually absent between 1976 and 1981. The difference may be attributable to the general decline of groundfish abundance in the Bay of Fundy after 1970 (Hare 1977). Among other flatfishes, windowpane, Scoph- thalmus aquosus, was a regular component at station C and the smooth flounder, Liopsetta putnami, was common among the inshore-estuarine communities during summer (Tables 2, 3). Yellowtail flounder, Limanda ferruginea, was a rare member (4-5/tow) of the summer assemblage at station A and occasional at the other two deep stations. Juvenile Atlantic halibut, Hippoglossus hippoglossus, was a low- abundance member (2-3/tow) of the winter assem- blage at station A. The fourspot flounder, Paralichthys oblongus, was captured once at station A during the abnormally warm fall of 1978. Eight species of gadoid fishes were captured during the study (Tables 2, 3). Adult Atlantic cod, Gadus morhua, was an abundant member of the summer component at offshore sites in Passamaquoddy Bay, particularly station A, but was absent from there in winter. It was a common member of the early winter assemblage in the Bay of Fundy but rare thereafter (Figs. 8, 9). During summer, juvenile Atlantic cod (10-20 cm) were captured occasionally while seining beach sites, but were more common in gill net catches at intermediate depth (30 m) inshore (stations 5 and 7; Table 3). The shallow water abundance maxima of 400 300 200 100 Gadus morhua ■ B J-r4, AM X i i^Avflj pM A 1 R5Q0 ' ■"■i-A+C adult means G5SS3-A + C juvenile means I i-B means * -A + C no catch A -B no catch Pp,n,n^, , Jj , ^m^rn* pa, p A , ' A E*4 | | | I 200r f-i*V- rAi— I6 ^ I I I 0 100 50 Melanogrammus aegleflnus -rt, , f,JUq ^rrfWYf^ r'AM i ■ | ■*■ i i I'l'V^Vt i [M. iM '*' »'■ ■** iAT*P-r-rAr*i ULf i»4*+V-t* Urophycis tenuis - Urophycls chuss 50 r -Ar-r-AV -r4* i'i P'i'i i ■ | i*T-r-r-r* Enchelyopus cimbrius J A O 1976 ^Tf A J A O W F A J A O D F A J A O D F A J A O D^ F A J A 1 1977 ' 1978 ' 1979 ' 1980 1981 Winter FIGURE 8.— Seasonal occurrence and abundance of gadoids at offshore stations in the Bay of Fundy and Passamaquoddy Bay, 1976-81. 129 FISHERY BULLETIN: VOL. 82. NO. 1 10r Gadus morhua June 23, 1980 ■ n - 0 Bay of Fundy - n -. 6 Passamaquoddy Bay ,n -i 1 M " l" . 1 1" 1 1 1 Aug. 13, 1980 10r ■ n= 3 a n = 4 -p — i ■ r"-i 1 — tP-°- 10 Oct. 9, 1980 ■ n = 2 □ n = 4 -i r r 20 Dae. 13, 1980 10 ■1 10=35 0 30 1 r l i i i Jan. 15, 1981 1 1 20 - JL ,„•:!? 10 - 0 60 #1 i ' i ' I i i ' I SO ~ 40 _ April 29, 1981 30 - ■ n ; 3 , on: 163 20 - 10 - 0 ^ <"i 1 i r ■ - i ■ i 0 20 40 60 80 100 Length (cm) FIGURE 9. — Seasonal size distributions of Gadus morhua at station B in the Bay of Fundy and station A in Passamaquoddy Bay, 1980 and 1981. young cod (0+, 1 + , <17 cm) has been previously reported in the western North Atlantic (Schroeder 1930) but is not well documented. On the other hand, this occurrence of young cod in the North Sea is well known (Daan 1978). During winter, juvenile cod were abundant at station A or in colder winters at station B (Fig. 8, 1980 and 1981). Both juvenile and adult cod were more abundant at station A during our study than during 1965 (20-70/tow, Tyler (1971); 1976-81, 50-400/tow). Haddock, Melanogrammus aeglefinus, were never abundant during our study. Adults were captured only at the hard-bottom station A during summer (Fig. 8) and juvenile haddock (1+) were occasionally captured at the same site in winter. Catches of had- dock declined from a maximum of 25/tow to <5/tow during the study period (Fig. 8). However, up to 260 haddock/tow were caught at station A during 1965 (Tyler 197 1). Decline in abundance after 1965 might be the cause for the collapse of the Gulf of Maine had- dock stock in 1970 (Hare 1977; Clark et al. 1982). Only juvenile pollock, Pollachius virens, were cap- tured during the study. Pollock of the annual year class (0+) were either rare or extremely abundant at beach sites (100+/seine haul) in a given year, depending perhaps, on the size of the annual year class. Pollock dominated beach catches during early summer but disappeared from this region by Sep- tember (Fig. 5). In years when 0+ pollock were abun- dant along the beach in summer, members of the same year class were also abundant the following winter at station A (1976-77, 1981) and, in summers of low abundance on the beach, they were correspond- ingly rare offshore in winter (1977-78; Fig. 8). Large numbers of pollock larvae were present in the plankton during March 1979 (Scott 1980), and we again encountered large number of 0+ juveniles at station A in the winter of 1979-80. Present findings suggest there may have been three large year classes produced during our study period, 1976, 1979, and 1981. Adult white, Urophycis tenuis, and red, U. chuss, hakes were common summer components at offshore stations A and B (Markle et al. 1982). Juvenile white hake (<15 cm) were a summer component at beach stations (Fig. 5), but were rarely captured thereafter and only then at offshore sites in winter. Also in 1965 few small hake were captured after December (Tyler 1971). Apparently hake leave Passamaquoddy Bay in winter (Markle et al. 1982). In the present study, the one time hake were observed during winter was at station B in the Bay of Fundy (Fig. 8). The fourbeard rockling, Enchelyopus cimbrius, was a regular component at station B in the Bay of Fundy and occasional in summer at station A (Fig. 8). The mesh size of our gear was just small enough to cap- ture large individuals of this species, and it was prob- ably more abundant than indicated. Larval rockling were a rare summer component of inshore sites (Table 4). Battle (1930) and Tyler (1971) both con- sidered rockling a summer occasional in Passama- quoddy Bay, occurring there during spawning migration. Tyler's catch rate at station A (2-3/tow) was similar to ours at that site. Larger catch rates at station B (10-50/tow) may be due to rocklings pref- erence for soft-bottom habitat (Bigelow and Schroeder 1939). Silver hake, Merluccius bilinearis, was often the most abundant gadoid found at offshore stations dur- ing summer, and juveniles were a regular component at station B year round (Fig. 10). Large numbers of adult silver hake were present during fall (Fig. 10) in company with other migratory summer occasionals, including American shad, Alosa sapidissima; spiny dogfish, Squalus acanthias; and butterfish, Porono- tus triacanthus . All these fishes may carry out coun- terclockwise spring to fall migrations around the Bay of Fundy similar to the shad (Dadswell et al. 1983). 130 MACDONALD ET AL.: FISH ASSEMBLAGES AND THEIR SEASONAL MOVEMENTS The Atlantic tomcod, Microgadus tomcod, was a summer (Fig. 5) and in estuaries in early winter regular component of the inshore assemblage and (Table 3). was particularly abundant at beach sites during early Clupeids and osmerids made up a major portion of Table 4.— Catch of fishes at intertidal seining station 3 (Brandy Cove) during period May 1976-November 1977. Fish captured during three 5-min seine hauls (100 X 15 m) (|j] = juvenile; [1| = larvae). Species 1976 1977 15/05 14/06 13/07 18/08 15/09 10/10 08/12 15/02 20/03 10/04 30/05 29/06 15/07 18/09 10/10 17/11 fla/a radiata [j] R. ennacea [j] Alosa pseudoharengus Clupea harengus Salmo salar [j] Osmerus mordax Fundulus heteroclftus Gasterosteus aculeatus G wheatlandi Enchelyopus cimbnus [I] Gadus morhua [j] Microgadus tomcod Pollachius virens [|| Urophycis tenuis [)) Scomber scombrus Pho/is gunnel/us Cyctopterus lumpus Myoxocephalus aeneus M scorpius [|] M octodecemspinosus [j] Hemitnpterus amencanus Pseudopleuronectes amencanus Liopsetta putnami Memdia menidia 25 — 5 51 1 1 115 2 3 — 132 15 12 1 2 — - - -observed never captured - — 1 2 — — 3 1 10 2 — 2 — — — — 15 3 1 4 1 1 1 I 1 2 — — 2 I 15 4 10 — — 4 - 3 - 6 4 5 26 32 3 5 1 27 3 2 3 8 2 — 3 2 — — 3 6 1 1 — — - 4 1 2 1 12 2 1 1 3 2 1 3 2 1 - 22 8 3 2 1 1 2 — 2 — — 2 4 300 200 100 .2 o I 150 ■o 100 1000 ^.eso Clup»a harengus harengus ~^*J — I — l^^n L i-A +C means 3-B means -A + C no catch -B no catch -P-P-r-i'M-P-^T* ' i i | i^ I i i*t*P^*i»* \ j* i^ 600r> g600 < 50 Merlucclus bilinear!* , , | ,lif , a-a^A-^+ I I I I'M I I l*f 1200 700 I I t-|"i ■"■ M< A rt l[*4- 40 20 Squalus acanthlas ♦ i ■ ■ J AODFAJ A0DFAJ AOD|FAJ AOD|FAJ A0D|FAJ 1976 1977 1978 1979 1980 1981 Winter FIGURE 10.— Seasonal occurrence and abundance of pelagic fishes and dogfish at offshore stations in the Bay of Fundy and Passamaquoddy Bay, 1976-81. 131 FISHERY BULLETIN: VOL. 82, NO. 1 the fishes caught at inshore sites (Table 2). At beach station 8, alewives, Alosa pseudoharengus; Atlantic herring, Clupea harengus harengus; and American smelt, Osmerus mordax; appeared in mid-July and increased in abundance during August (Fig. 5). Her- ring were abundant in estuaries during summer and were replaced there by smelt in winter (Table 3). Large American smelt were present at offshore sites in Passamaquoddy Bay in mid-summer as observed by Tyler (1971). During most winters, juvenile Atlan- tic herring (10-20 cm) were abundant at offshore sites, particularly inside Passamaquoddy Bay at intermediate depths (station C; Fig. 10). Catches were variable, possibly because of schooling behavior (Brawn 1960). Tagging experiments indi- cate herring move from inshore during summer to deeper water in winter (McKenzie and Tibbo 1961). Six species of sculpin (Table 2) were commonly encountered at offshore station of which two — longhorn sculpin, Myoxocephalus octodecemspino- sus, and sea raven, Hemitripterus americanus — were abundant, regular components (Fig. 1 1). Juveniles of most species were common at beach sites in summer (Table 4) and at station A in winter (Table 2). Increases in abundance of longhorn sculpins at sta- tion B during winter were observed (Fig. 1 1) and may be the result of migration out of Passamaquoddy Bay. Two small species, Arctic hookear sculpin, Artediellus uncinatus, and mailed sculpin, Triglops murrayi, were winter occasionals at station A. They were perhaps more abundant than catch rates indicated (2-5/tow) because their maximum size range was at the lower limit of catchability for our trawl. The blennioid-like fishes were represented by seven species (Tables 2, 3) of which ocean pout, Macrozoarces americanus, was regular at offshore stations in Passamaquoddy Bay (Fig. 11), and rock gunnel, Pholis gunnellus, was a regular component at beach sites (Table 4). Ocean pout abundance in Passamaquoddy Bay was generally highest in early summer and declined thereafter (Tyler 1971; Fig. 11). Abundance of ocean pout usually increased at station B in late summer and fall, suggesting move- ment from Passamaquoddy Bay to the Bay of Fundy. 200f: 150 100 o 50 a. u c CO ■o 50 a < O^T 100r 500 I -i — i — i — ►" i i P"A 360 I u n f Myoxocephalus octodecemspinosus 430 h-A + C means z=i-B means * -A + C no catch * -B no catch ul ,I»I|A + f\- •^4 i*l ^ ivU- j i fn i i 200 i Hemitripterus americanus A-t-i-jlVvP < Pi i ■aM*i,i .^ni^i>i ^T ■ i [ P. i i .'i*4 «i AM i j «P i r*P Mi 1M-1M -P^r-r-rV 50- J A o d| f a j JAODFAJA 1976 1977 Mai ," O D| F Macrozoarces americanus 4-lM, ^ffl A J A '6 'd| 'f' A J A 1978 1979 ODJFAJ A O D| F A J 1980 1981 Winter FIGURE ll.— Seasonal occurrence and abundance of sculpins and ocean pout at offshore stations in the Bay of Fundy and Passama- quoddy Bay, 1976-81. 132 MACDONALD ET AL.: FISH ASSEMBLAGES AND THEIR SEASONAL MOVEMENTS It may be a response to avoid warm temperatures (Olsen and Merriman 1946). Movement of ocean pout is generally thought to cover only short dis- tances (Orach-Maza 1975; Sheehy et al. 1977). Other blennioids occurred infrequently at station A (Table 2). Selectivity of our shrimp trawl may have been a factor in these low catches. One species, radiated shanny, Ulvaria subbifurcata, which was thought to be rare in Passamaquoddy Bay (Leim and Scott 1966), was often captured (5/tow) at station A during winter. Scuba searches during summer revealed radiated shanny were abundant inshore, under rocks in 6-9 m of water (Dadswell and Melvin, pers. obs.). Five species of skate were captured during the study (Table 2): Two species, thorny skate, Raja radiata, and smooth skate, R. senta, were common, regular components of the offshore site in the Bay of Fundy; two little skate, R. erinacea, and winter skate, R. ocellata, were regular components of station A in Passamaquoddy Bay; and one species, the barndoor skate, R. laeuis, was encountered occasionally at sta- tion B. The species cooccurrences of skates and their habitat selection are as described by McEachran and Musick (1975). Some seasonal movement into Passamaquoddy Bay was exhibited. Abundance of smooth and thorny skates at station A increased dur- ing summer and declined after late fall. Juveniles of thorny, little, and winter skates were often captured at beach sites during summer (Table 3). Several smaller fishes were captured at inshore sites only, but again this may be an artifact of sam- pling gear. Threespine stickleback, Gasterosteus aculeatus, was a regular component at most beach sites (Table 4). Other sticklebacks were more or less confined to estuarine areas (Table 3). Mummichog, Fundulus heteroclitus, and Atlantic silversides, Menidia menidia, occurred mainly in estuaries during summer but were part of the winter community at beach sites (Table 4). Assemblages and Diversity Species assemblages in the study area varied according to site and season. If juveniles and adults of some dominant species are considered as separate taxonomic units (Table 2), calculated coefficients of community show similarity between similar habitat types (e.g., soft bottom) at a given season, and be- tween the summer assemblage of one habitat and the winter assemblage of the next seaward habitat (Table 5). In general, movement of assemblages was from inshore in summer to offshore in winter with some return movement in spring (Fig. 12). Some species, however, exhibited a partial reverse of this pattern (Atlantic tomcod, ocean pout). Specific groupings of fish were segregated among the available habitats according to season. The "es- tuarine" assemblage was dominated by warmwater, euryhaline species, including sticklebacks, Atlantic silversides, mummichogs, and juvenile clupeids. Most of this group moved to adjacent, inshroe marine habitat in winter (Tables 3,4), but Atlantic tomcod and American smelt moved in the reverse direction to form a winter estuarine group (Table 3). The summer "beach" assemblage consisted of regulars such as threespine stickleback and rock gun- nel and a summer component including juvenile gadids, juvenile sculpins, flounders, and juvenile alosids. Juvenile gadids (pollock, white hake, and Atlantic tomcod) were most abundant in early sum- mer but were replaced by steadily increasing num- bers of clupeids in late summer (Fig. 5). Numerous other postlarval and juvenile fishes, including four- beard rockling and lumpfish, Cyclopterus lumpus, appeared in the beach zone during the summer (Table 3). In late fall, most of this assemblage left the beaches and occupied offshore sites in Passama- quoddy Bay. Atlantic herring concentrated at the soft-bottom station C and the gadids, sculpins, and winter flounder (juveniles) at the hard-bottom sta- tion A. Threespine stickleback and rock gunnel Table 5. — Coefficients of community among seasonal fish assemblages in the lower Bay of Fundy. Se award Estuarine Estuarine Beach Beach C C A A B B winter summer winter summer winter summer winter summer winter summer Estuarine winter — 10 0 200 70 00 12.5 00 2.0 0.0 00 Estuarine summer — — 50.0 12.5 20.0 0.0 0.0 00 00 0.0 Beach winter — — — 66 14 3 3 8 42 0.0 0.0 0.0 Beach summer — — — — 6 6 33.3 36.1 17 3 21 2 0.0 C winter — — — — — 12.5 4 2 5 7 6.6 00 C summer — — — — — — 48 400 40.0 47.0 A winter — — — — — — — 20.9 43.0 26.3 A summer — — — — — — — — 36.4 42 8 B winter — — — — — — — — — 258 133 FISHERY BULLETIN: VOL. 82, NO. 1 Estuarine Station A Winter Community Pollock (juvenile) Cod (juvenile) Haddock (juvenile) Winter flounder (juvenile) Herring (adult) Regular Community Sea raven Little skate Longhorn sculpin Ocean pout Summer Community Cod (adult) Haddock (adult) Winter flounder (adult) Thorny skate Silver hake White hake Fourbeard rockling 11 Gulf of Maine- Scotian Shelf Winter Community Witch (adult) Cod (adult) Haddock (adult) Silver hake Dogfish White hake Beach Winter Community Silversides Mummichog Regular Community 3-spine stickleback Tomcod Rock gunnel Summer Community Pollock (juvenile) Cod (juvenile) White hake (juvenile) Winter flounder (juvenile, adult) Herring (juvenile) Sea raven (juvenile) Winter Community Tomcod Smelt Summer Community Herring (juvenile) Sticklebacks Mummichog Silversides Smooth flounder American eel Station C Winter Community Herring (juvenile, adult) Summer Community Plaice Silver hake Winter flounder Ocean pout Station B Winter Community Winter flounder (adult) Witch (juvenile) Longhorn sculpin Herring Atlantic sturgeon Regular Community Plaice Sea raven Thorny skate Smooth skate Silver hake Fourbeard rockling Summer Community White hake Witch (adult) Dogfish Ocean pout American shad FIGURE 12. — Communities of fishes occurring at each site divided into summer component (SC), winter component (WC), and regular component (RC). Arrows indicate direction of seasonal movement. remained at beach sites over winter and were joined by Atlantic silversides and mummichog to form a winter assemblage (Table 4). During summer an "offshore, hard-bottom" assemblage consisting of adult gadids (Atlantic cod, haddock, white and red hake), adult flounders (win- ter yellowtail), ocean pout, adult sculpins, and skates assembled inside Passamaquoddy Bay. Sea raven, longhorn sculpin, ocean pout, and little skate remained at this site over winter and were joined by juvenile fishes from the beach zone. The other species apparently move to offshore sites in the Bay of Fundy and/or to the Scotian Shelf (McCracken 1959; Wise 1962; Edwards 1965; Kulka and Stobo 1981). The "offshore, soft-bottom" assemblage consisted of American plaice, witch flounder, white hake, four- beard rockling, and skates as described by Bigelow and Schroeder (1939). This group at station B was the most stable assemblage studied and had the largest regular component. Conversely, similar assemblages which occurred at the shallower, soft- 134 MACDONALD ET AL.: FISH ASSEMBLAGES AND THEIR SEASONAL MOVEMENTS bottom station C were the most seasonally dynamic (Fig. 12). Adult witch flounder and most hakes left station B in winter for grounds further offshore in the Gulf of Maine (Powles and Kohler 1970; Kulka and Stobo 1 98 1), and this site was occupied by adult win- ter flounder and longhorn sculpin, perhaps from inside Passamaquoddy Bay or other adjacent inshore sites (McCracken 1963). Superimposed on the two offshore, essentially benthic fish assemblages was a seasonal semipelagic component. In summer, silver hake was the numerically dominant species. During fall, diversity increased with the arrival of spiny dogfish, butterfish, and American shad. In winter, Atlantic herring numerically dominated the pelagic component at all offshore sites (Fig. 12). Diversity, expressed simply as number of species captured, varied appreciably at beach sites during the year. Diversity was 2-5 species in winter-spring, 9-13 species in summer, and 4-6 species in fall-winter (Fig. 13). Total number of species captured at inshore sites was 35, compared with 51 species cap- tured at offshore sites. Diveristy of assemblages at deep offshore sites (80+ m) was more stable on an annual basis because of the seasonal influx and departure of species from and to adjacent habitats (Fig. 14). Species number varied between 7 and 1 7 fishes at station B and 7 and 20 fishes at station A, fluctuating about a mean of 1 2/ sampling trip. During 1965, Tyler (1971) observed a higher mean diversity of 17 species/ trip at station A with a maximum occurrence of 24. The difference between his observations and ours may be accounted for partially by the decline in haddock abundance o 5 14 12 - 10 - 8- Q. CO ~ 6 2 - A l\ / 1 - 1 \ / 1 : A "*\ Station ' i \ Station 8 — A i II 1 1 1 1 J J A Month FIGURE 13.— Monthly diversity of fishes at intertidal stations 3 and 8 in Passamaquoddy Bay. Species/month for station 3 is mean of 1976 and 1977 samples. 20 16 12 2 0 a CO Station A \ Station B ' ' i i ' i i i i i i . i i i i i i i i i i i i i i i i ■■■ i F A J A O D| F A J A O D| F A J A 1976 1977 1978 Figure 14.— Seasonal diversity of fishes at station A (Passamaquoddy Bay) and station B (Bay of Fundy). Ver- tical bars represent the range among replicated collections. 01 — I — 1—1 — I — I — 1 — I — I I 1—1 ,_l I I 1 L_l I I I i i i 1 I I I I I l_l F A J A O D] F A J A O D| F A J A 1979 1980 1981 135 FISHERY BULLETIN: VOL. 82, NO. 1 since 1965 and the recent absence of American plaice from this site, and partially by his use of a 0.6 cm cod end liner, which would have retained small, occasional species more often than our 2.5 cm cod end. Highest diversities occurred during winter at sta- tion B and during summer at station A (Fig. 14) as a result of seasonal exchange between these sites and the arrival of periodics. The highest diversities record- ed during the study period occurred at station A dur- ing the fall, coinciding with maximum annual temperatures (Fig. 2). Diversity at station C, the mid- depth site, decreased from 13 species in May 1978 to 4 species in May 1980, perhaps in response to a general decline in lower Bay of Fundy temperatures during the study period (Fig. 3). GENERAL DISCUSSION Most authors have related the occurrence and dis- tribution of adult benthic fishes in the North Atlantic to substrate type and temperature (Edwards 1965; Colton 1972; McEachran and Musick 1975; Scott 1976) and have shown that there is a marked seasonal variation (Lux and Nichy 1971; Jeffries and Johnson 1974). Our findings agree and suggest yearly dif- ferences at the same site for a given time may be influenced mainly by annual ocean climate pertuba- tion. Species occurrence and abundance appeared to change in response to seemingly small changes in temperature. Jeffries and Johnson (1974) reported a similar observation concerning winter flounder abundance over a 7-yr period in Narragansett Bay. Pelagic and semipelagic species (Atlantic herring, silver hake) demonstrated little or no substrate pref- erence. Occurrence was apparently related to annual migratory behavior. Seasonal movements of the various species was largely from an inshore, shallow-water locality in summer to an offshore, deepwater locality in winter with a reverse movement occurring in spring. Cause of this movement may have a large physiological com- ponent related to temperature effects on the osmoregulation of marine fishes (Potts and Parry 1964). In the southern part of their range, fish such as winter flounder migrate onshore in winter (Bigelow and Schroeder 1953) in response to availability of preferred temperature but never encounter the low temperatures found at northern latitudes. Atlantic tomcod, a species known to produce an antifreeze in its blood (Fletcher et al. 1982), was one of the few fishes exhibiting onshore migration to lower salinities during winter in this area. For many species (pollock, Atlantic herring, white hake), migration from inshore habitat to offshore is unidirectional for the individual, since each year the beach community consists of the new 0+ year class. For other species (winter flounder, juvenile sculpins, radiated shanny), the return inshore is an annual occurrence, triggered perhaps as much by resource availability and pre- dator avoidance as by physiology. Tyler (1971) concluded that in Passamaquoddy Bay movements of large fish independent of the small individuals of a species were not evident for fishes other than hake, but we found obvious dif- ferences in size-class distributions and abundance between summer and winter populations of winter flounder, witch flounder, Atlantic cod, and pollock at offshore sites and a complete lack of most fish inshore. This suggests marked segregation between juveniles (at least 0+ age group) and adults for these species. The use of shallow water habitat as nursery area by fishes of commercial important in the Cana- dian North Atlantic has received little attention. In Europe, this fact has been amply demonstrated for many fish species, including Atlantic cod and pollock (Zijlstra 1972; Daan 1978; Burd 1978; Rauck and Zijlstra 1978). The use of beach habitat as nursery by these fishes makes them susceptible to coastal pollu- tion impacts and puts their adult fisheries at risk to coastal degradation and development, Decline in haddock abundance in Passamaquoddy Bay since 1965 coincides with increased numbers of Atlantic cod. However, previous studies indicate lit- tle interaction between these two species (Tyler 1972; Jones 1978). Catches in 1965 (Tyler 1971) coincided with the largest haddock abundance on record (Clark et al. 1981). Fishermen in Passamaquoddy Bay may only catch haddock consistently during years preceded by large recruitment on Georges Bank, the Scotian Shelf, and the Gulf of Maine. In the Bay of Fundy region, fish assemblages are segregated according to habitat and, although fish movement is influenced by seasonal climatic regime, assemblages appear cohesive through time. In sum- mer, fishes assembled and exploited the available resources as members of 1) estuarine, 2) beach, 3) offshore, hard-bottom, 4) offshore, soft-bottom, and 5) migratory-pelagic assemblages. With winter, movement of species and/or age groups resulted in different seasonal assemblages in each habitat, but major groupings remained essentially intact and replaced each other seaward. The reverse movement occurred in spring. A large portion of benthic and pelagic components occurring at the offshore, hard- bottom habitat were migratory. In contrast, the offshore, soft-bottom assemblage was more senden- tary. Smaller seasonal variation in the water tem- 136 MACDONALD ET AL.: FISH ASSEMBLAGES AND THEIR SEASONAL MOVEMENTS perature at the Bay of Fundy, soft-bottom site, and the greater seasonal stability of invertebrate food resource production in this type of habitat ( Wildish and Dadswell in press) may also be important. The dynamic nature of the hard-bottom community, par- ticularly among commercially valuable species, emphasizes the need for well-designed, seasonal sampling programs in order to properly assess the occurrence of species and abundance offish stocks in a local area. Long-term changes are apparent from annual assessment data (Brown et al. 1973), but higher resolution surveys at "type" localities are needed to properly determine causative factors, whether physical or biological. ACKNOWLEDGMENTS We thank Captain Tom Allen and Floyd Johnson, crew of the Pandalus for their help. Bill McMullon and Frank Cunningham prepared the figures and Brenda Fawkes and Jeanine Hurley typed the manuscript. J. S. Scott, W. B. Scott, D. Markle, and D. J. Scarratt reviewed the manuscript. Work by J. S. Macdonald was in partial fulfillment for Ph.D. requirements supported by NSERC Grants to R. H. Green. This work was supported in part by a grant from the Canadian National Sportsman's Fund (4- R88) to D. Methven. LITERATURE CITED Battle, H. I. 1930. Spawning periodicity and embryonic death rate of Enchelyopus cimbrius (L.) in Passamaquoddy Bay. Con- trib. Can. Biol. Fish., New Ser. 5:361-380. BlGELOW, H. B., AND W. C. SCHROEDER. 1939. Notes on the fauna above mud bottoms in deep water in the Gulf of Maine. Biol. Bull. (Woods Hole) 76:305- 324. 1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 53, 57 p. Brawn, V. M. 1960. Seasonal and diurnal vertical distribution of herring (Clupea harengus L.) in Passamaquoddy Bay, N.B. J. Fish. Res. Board Can. 17:699-711. Brown, B. E„ J. A. Brennan, E. G. Heyerdahl, M. D. Grosslein, AND R. C. HENNEMUTH. 1973. An evaluation of the effect of fishing on the total finfish biomass in ICNAF Sub-area 5 and statistical area 6. Int. Comm. Northw. Atl. Fish. Res. Doc. 73/8, 32 p. Burd, A. C. 1978. Long-term changes in North Sea herring stocks. In G. Hempel (editor), North Sea fish stocks - recent changes and their causes, p. 137-153. Rapp. P. — V. Reun. Cons. Perm. Int. explor. Mer 172. Clark, S. H., W. J. Overholtz, and R. C. Hennemuth. 1982. Reveiw and assessment of the Georges Bank and Gulf of Maine haddock fishery. J. Northw. Atl. Fish. Sci. 3:1- 27. Colton, J. B., Jr. 1972. Temperature trends and the distribution of groundfish in continental shelf waters, Nova Scotia to Long Island. Fish. Bull., U.S. 70:637-657. Cushing, D. H. 1980. The decline of the herring stocks and the gadoid out- burst. J. Cons. Int. Explor. Mer 39:70-81. Daan, N. 1978. Changes in cod stock and cod fisheries in the North Sea. In G. Hempel (editor), North Sea fish stocks - recent changes and their causes, p. 39-57. Rapp. P.— V. Reun. Cons. Perm. Int. Explor. Mer 172. 1980. A review of replacement of depleted stocks by other species and the mechanisms underlying such re- placement. In A. Saville (editor), The assessment and management of pelagic fish stocks, p. 405-421. Rapp. P.- V. Reun. Cons. Perm. Int. Explor. Mer 177. Dadswell, M. J., G. D. Melvin, and P. J. Williams. 1983. Effect of turbidity on the temporal and spatial utiliza- tion of the inner Bay of Fundy by American shad (Alosa sapidissima) (Pisces: Clupeidae) and its relationship to local fisheries. Can. J. Fish. Aquat. Sci. 40 (Suppl. 1):322- 330. Dow, R. L. 1964. A comparison among selected marine species of an assocation between sea water temperature and relative abundance. Extr. J. Cons., Cons. Int. Explor. Mer 28:425-431. Edwards, R. L. 1965. Relation of temperature to fish abundance and dis- tribution in the southern New England area. Int. Comm. Northw. Atl. Fish., Spec. Publ. 6:95-110. Fletcher, G. L., C. L. Hew, and S. B. Joshi. 1982. Isolation and characterization of antifreeze glyco- proteins from the frostfish, Microgadus tomcod. Can. J. Zool. 60:348-355. HACUNDA, J. S. 198 1. Trophic relationships among demersal fishes in a coast- al area of the Gulf of Maine. Fish. Bull., U.S. 79:775- 788. Hare, G. M. 1977. Atlas of the major Atlantic coast fish and invertebrate resources adjacent to the Canada-United States boun- dary areas. Can. Fish. Mar. Serv. Tech. Rep. 681, 97 p. Jaccard, P. 1932. Die statistisch-florestische Methode als Grundlage der Pflanzen-soziologie. Handb. Biol. Arbeitsmed. 5:165- 202. Jeffries, H. P., and W. C. Johnson. 1974. Seasonal distributions of bottom fishes in the Narragansett Bay area: Seven-year variations in the abun- dance of winter flounder {Pscudopleuronectes americanus). J. Fish. Res. Board Can. 31:1057-1066. Jones, R. 1978. Competition and co-existance with particular reference to gadoid fish species. In G. Hempel (editor), North Sea fish stocks - recent changes and their causes, p. 292-300. Rapp. P.-V. Reun. Cons. Perm. Int. Explor. Mer 172. Keast, A. 1970. Food, specializations and bioenergetic interrelations in the fish faunas of some small Ontario waterways. In J. H. Steel (editor), Marine food chains, p. 377-411. Univ. Calif. Press. Kontkanen, P. 1957. On the delimitation of communities in research on 137 FISHERY BULLETIN: VOL. 82, NO. 1 animal biocoenotics. Cold Spring Harbour Symp. Quant. Biol. 22:373-378. KULKA, D. W., AND W. T. STOBO. 1981. Winter distributions and feeding ofm a ckerel on the Scotian Shelf and outer Georges Bank with reference to the winter distribution of other finfish species. Can. Tech. Rep. Fish. Aquat. Sci. 1038, 38 p. Leim, A. H., and W. B. Scott. 1966. Fishes of the Atlantic coast of Canada. Fish. Res. Board Can., Bull. 155, 485 p. Lux, F. E., and F. E. NlCHY. 1971. Number and lengths, by season, of fishes caught with an otter trawl near Woods Hole, Massachusetts, Septem- ber 1961 to December 1962. U.S. Dep. Commer., Natl. Mar. Fish. Ser., Spec. Sci. Rep. Fish. 622, 15 p. Markle, D. F. 1975. Young witch flounder, Glyptoeephalus cynoglossus , on the slope off Virginia. J. Fish. Res. Board Can. 32:1447- 1450. Markle, D. F., D. A. Methven, and L. J. Coates-Markle. 1982. Aspects of spatial and temporal cooccurrence in the life history stages of the sibling hakes, Urophycis chuss (Walbaum 1792) and Urophycis tenuis (Mitchill 1815) (Pisces: Gadidae). Can. J. Zool. 60:2057-2078. McCracken, F. D. 1959. Studies of haddock in the Passamaquoddy Bay region. J. Fish. Res. Board Can. 17:175-180. 1963. Seasonal movements of the winter flounder, Pseudopleuronectes americanus (Walbaum), on the Atlan- tic coast. J. Fish. Res. Board Can. 20:551-586. McEachran, J. D., and J. A. Musick. 1973. Characters for distinguishing between immature specimens of the sibling species, Raja erinacea and Raja evellata (Pisces: Rajidae). Copeia 1973:238-250. 1975. Distribution and relative abundance of seven species of skates (Pisces: Rajidae) which occur between Nova Scotia and Cape Hatteras. Fish. Bull., U.S. 73:110-136. McKenzie, R. A., and S. N. Tibbo. 1961. Herring movements in the Bay of Fundy and Gulf of Maine, 1957 and 1958. J. Fish. Res. Board Can. 18:221- 252. Moore, R. H. 1977. Variations in the diversity of summer estuarine fish populations in Aransas Bay, Texas, 1966-1973. Estuarine Coastal Mar. Sci. 6:495-501. Musick, J. A. 1973. A meristic and morphometric comparison of the hakes, Urophycis chuss and U. tenuis (Pisces, Gadidae). Fish. Bull., U.S. 71:479-488. Nl.I-H. 1982. Meristic variation in beaked redfishes, Sebastes men- tella and S. fasciatus, in the Northwest Atlantic. Can. J. Fish. Aquat. Sci. 39:1664-1685. Olsen, Y. H., and D. Merriman. 1946. Studies on the marine resources of southern New England. IV. The biology and economic importance of the ocean pout, Macrooarces americanus (Block and Schneider). Bull. Bingham Oceanogr. Collect. Yale Univ. 9(4):1-184. Orach Meza, F. L. 1975. Distribution and abundance of the ocean pout, Mac- rozoarces americanus (Block and Schneider, 1801) in the Western North Atlantic ocean. M.S. Thesis, Univ. Rhode Island, Kingston, 108 p. Potts, W. T. W., and G. Parry. 1964. Osmotic and ionic regulation in animals. Pergammon, N.Y., 423 p. Powles, P. M., and A. C. Kohler. 1 970. Depth distributions of various stages of witch flounder (Glyptoeephalus cynoglossus) off Nova Scotia and in the Gulf of St. Lawrence. J. Fish. Res. Board Can. 27:2053- 2062. QuiNN, N. J. 1980. Analysis of temporal changes in fish assemblages in Serpentine Creek, Queensland. Environ. Biol. Fish. 5:117-133. RAUCK, G., AND J. J. ZlJLSTRA. 1978. On the nursery-aspects of the Waddensea for some commercial fish species and possible long-term changes. In G. Hempel (editor), North Sea fish stocks - recent changes and their causes, p. 266-275. Rapp. P.-V. Reun. Cons. Perm. Int. Explor. Mer 172. Richards, S. W. 1963. The demersal fish population of Long Island Sound. Bull. Bingham Oceanogr. Collect,, Yale Univ. 18:1-101. Saila, S. B. 1961. A study of winter flounder movements. Limnol. Oceanogr. 6:292-298. Schroeder, W. C. 1930. Migrations and other phases in the life history of cod off southern New England. Bull. U.S. Bur. Fish. 46:1-136. SCOTT, J. S. 1976. Summer distribution of groundfish on the Scotian Shelf, 1970-74. Can. Fish. Mar. Serv. Tech. Rep. 635, 13 p. 1980. Occurrence of pollock, Pollachius virens, and sand lance, Ammodytes sp., larvae in the Bay of Fundy. J. Northw. Atl. Fish. Sci. 1:45-48. Sheehy, P. J., S. Y. K. Sheriouda, A. J. Alton, S. B. Saila, and S. M. Constanttnides. 1977. The ocean pout: An example of underexploited fisheries resource development. Mar. Fish. Rev. 39 (6):5- 15. Sherman, K., C. Jones, L. Sullivan, W. Smith, P. Berrien, and L. Ejsymont. 1981. Congruent shifts in sand eel abundance in western and eastern North Atlantic ecosystems. Nature (Lond.) 291:486-489. Sissenwtne, M. P., E. B. Brown, J. E. Palmer, and R. J. Esig. 1982. Empirical examination of population interactions for the fishery resources off the northeastern USA. In M. C. Mercer (editor), Multispecies approaches to fisheries management advice, p. 82-94. Can. Spec. Publ. Fish. Aquat. Sci. 59. Steiner, R. G. 1976. Neritic reef fish communities - some problems and a possible solution. In C. A. Simenstad and S. J. Lipovsky (editors), Fish food habit studies. First Pacific Northw. Tech. Workshop, 193 p. Sutcliffe, W. H., Jr., K. Drinkwater, andB. S. Muir. 1977. Correlations offish catch and environmental factors in the Gulf of Maine. J. Fish. Res. Board Can. 34:19-30. Tyler, A. V. 1971. Periodic and resident components in communities of Atlantic fishes. J. Fish. Res. Board Can. 28:935-946. 1972. Food resource division among northern, marine, demersal fishes. J. Fish. Res. Board Can. 29:997-1003. Van Guelpen, L., and C. C. Davis. 1979. Seasonal movements of the winter flounder, Pseudopleuronectes americanus, in two contrasting 138 MACDONALD ET AL.: FISH ASSEMBLAGES AND THEIR SEASONAL MOVEMENTS inshore locations in Newfoundland. Trans. Am. Fish. Soc. Bay of Fundy. Proc. N.S. Acad. Sci. 108:26-37. Wise, J. P. WARFEL, H. E., AND D. MERRIMAN. 1962. Cod groups in the New England area. U.S. Fish Wildl. 1944. Studies on the marine resources of southern New Serv., Fish. Bull. 63:189-203. England. I. An analysis of the fish population of the shore ZlJLSTRA, J. J zone. Bull. Bingham Oceanogr. Collect., Yale Univ. 1972. On the importance of the Wadden Sea as a nursery area 9(2): 1-91. in relation to the conservation of the southern North Sea Wildlish, D. J., AND M. J. Dadswell. fishery resources. Symp. Zool. Soc. (Lond.) 29:233- In press. Sublittoral Gammaridae of soft sediments in the 258. 139 THE DETECTION AND DISTRIBUTION OF LARVAL ARCTO-NORWEGIAN COD, GADUS MORHUA, FOOD ORGANISMS BY AN IN SITU PARTICLE COUNTER S. TlLSETH AND B. ELLERTSEN1 ABSTRACT An in situ particle counter system was developed to count measure food particles in numbers per liter within the size range 150-600 /urn, the sizes of copepod nauplii captured by first feeding cod larvae. Patches of particles/nauplii of 50- 1 00 per liter were found in the spawning and larval first feeding area. Different sizes of copepod nauplii showed diel vertical migration, and this influenced the formation of patches. Mixing of the water column by wind forces created a homogeneous vertical distribution of particles. Gut content analysis of cod larvae during these hydrographical conditions indicated reduced accessibility of food organisms to larvae. During the last few years fisheries scientists have done a great deal of laboratory work on the behavior of fish larvae and their energy requirements for growth and survival (Hunter 1972; Laurence 1974; Lasker and Zweifel 1978; Houde 1978; Werner and Blaxter 1980). A review of these data (Hunter 1981) shows that differences exist between the required density of prey particles for first feeding larvae to survive and the densities found in the sea. Since pelagic fish larvae are successful in their environ- ment, it is recognized that there must be patches of suitable concentrations of food organisms for first feeding larvae (Lasker and Zweifel 1978). This has been demonstrated for the northern anchovy, Engraulis mordax, in laboratory experiments by Hunter and Thomas (1974) and in a series of field investigations by Lasker (1978). Houde and Schek- ter ( 1 978) have shown increased survival of larval bay anchovy, Anchoa mitchilli, and sea bream, Archosargus rhomboidalis, when exposed to sim- ulated food patches in a laboratory experiment. This work has been stimulated by Hjort's (1914) hypothesis which simply stated that larval mortality rates may be due to variable feeding conditions at a critical stage, which in turn causes variations in year- class strength. It has been difficult to test this simple hypothesis in field surveys because of the inade- quacy of the sampling gear used (May 1974). To obtain a better understanding of the relationship be- tween estimates of food densities required by fish larvae in the laboratory and densities found in the 'Institute of Marine Research, Directorate of Fisheries, 5011 Bergen-Nordnes, Norway. open sea, samples should be taken which are relevant to larval searching behavior. This would require an enormous number of plankton samples. It would be time-consuming to obtain these samples with con- ventional plankton gear. Furthermore, water move- ment and dispersion would make it difficult to obtain time and space relationships for studying the forma- tion and dynamics of plankton patches (Steele 1978). One way of studying these relationships is by using in situ instruments (Boyd 1973; Pugh 1978; Tungate and Reynolds 1980). In this study an instrument designed to count and measure particles in situ in the size range of food organisms most frequently captured by cod larvae was used. Investigations were made on the spawning and first feeding grounds of the Arcto-Norwegian cod, Gadus morhua Linnaeus, during two successive years (1980-81). During the first survey, inves- tigations were made in a sheltered fjord where cod larvae are known to appear in high numbers (Ellertsen et al. 1977) and where the current system has been described (Furnes and Sundby 1981). The objective was to find and study the formation of mi- crozooplankton patches and to study larval cod feed- ing under different environmental conditions with regard to food density, water turbulence, etc. In the following year, the main first feeding area, an open ocean bay, was surveyed in order to find and study the vertical and horizontal distribution of micro- zooplankton patches in this exposed area. The present study is part of a project, started in 1975, dealing with growth, mortality, and drift of cod larvae in the Lofoten area (Ellertsen et al. 1976). Manuscript accepted June 1983. FISHERY BULLETIN: VOL. 82. NO. 1, 1984. 141 FISHERY BULLETIN: VOL. 82, NO. 1 MATERIALS AND METHODS The Particle Counter The in situ particle counter system was built and described by Mohus (1981), Eriksen (1981), and Eriksen and Mohus (1981). It is presented schematically in Figure 1. The system is based on a Hiac PC-320 Particle Counter2 which works on the principle of light blockage. The sensor (E-2500, dynamic range 80-2500 ^.m) is installed in a pressure-proof box together with a depth detector. A pump is connected to the sensor, and the sensor and pump are mounted to a rig which is lowered into the sea by winch. Seawater is pumped through a 60 cm long by 2.5 cm diameter hose through the sensor orifice (3 mm), at a flow rate of 6.15 1/min. Particles are counted by the Hiac PC-320 Particle Counter and depth is monitored by the depth detector unit. The "Micro -count" datalogger unit contains an Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. input-output interface to accomodate incoming data, a large internal data storage area, operator com- munication via a small CRT display, a keyboard, and a microprocessor with program to control the system. The microcomputer samples data from the Hiac PC- 320 Particle Counter and the data sample time can be selected from 1 to 99 s. Finally, a Silent 733 ter- minal is connected to the microcomputer. This ter- minal contains a full text keyboard and a page printer used for initial operator communication and printout of data tables. Two cassette tape stations are included in the terminal. The system operates from the surface to 50 m depth, and the registration of particles is presented on the TV monitor as the sensors are lowered into the sea. The vertical distribution of particles can be pre- sented on the monitor at 1, 2 , or 5 m depth, depend- ing on the selected depth intervals. Data are, however, printed out in 1 m depth intervals from the surface to 50 m depth as concentration of particles per liter in six different size groups (150-600 ju.m) on the Silent 733 terminal immediately after the sam- ples have been made. An in situ particle profile is I! TEST 11 BOX Cr - i DEPTH DETECTOR i V I I HIAC PC-320 PARTICLE COUNTER " ^ JL ^ - SENSORS INPUT-OUTPUT MICRO PROCE- SSOR DATA STOR- AGE COMMUNICATION Figure 1.— The particle counter system. ■ l_ SILENT-733 CASSETTE TAPE KEYBOARD-PRINTER MICRO-COUNT 142 TILSETH and ELLERTSEN: FOOD ORGANISMS OF LARVAL COD defined in the present paper as the concentration of particles within the size range of 150-600 /xm from the surface to 50 m depth in 1 m depth intervals. An object found in the Hiac sensor was measured so that the largest projected area was converted to a cir- cle of the same area. By calibration, the object was given a length similar to the diameter of this circle. The contours of Artemio nauplii were drawn by using a microscope drawing tube. Their areas were estimated by planimeter and converted to areas of circles and their diameters calculated. Their size dis- tribution was then divided into four 50 jum length groups of 200 to 400 /xm. Four of the Hiac Particle Counter channels were set according to the sensor calibration diagram to the corresponding size groups. The instrument system was tested and calibrated in the laboratory by comparing microscope and Hiac measurements of the size-frequency distribution of a sample of laboratory hatched Artemia nauplii. Tests were also made at sea when the research vessel was anchored. The in situ instrument data were com- pared with plankton pump samples taken simul- taneously. These samples were taken by a submer- sible electric pump (Flygt 2051, 250 1/min) which pumped samples on deck through a 50 m long by 5 cm diameter hose. Samples were taken at 0, 2.5, 5, 7.5, 10, 12.5, 15, 20, 25, 30, and 40 m depths. This is defined as a zooplankton pump profile. Seawater was collected in calibrated tanks (23.7 1), and zoo- plankton were filtered through 90 itm mesh plankton nets. Zooplankton were identified and counted by microscope, the whole sample (23.7 1) was counted. Results of the samples from these 11 depths were statistically compared with the in situ counts from corresponding depths by paired Mests. Field Investigations The main objectives of field investigations were to use the in situ instrument system to find particle patches and to identify larval cod food organisms and study their vertical distribution. Observations were made in the Lofoten area (Fig. 2). The effect of wind driven turbulence on the distribution of particles and the consequences on larval cod feeding incidence were studied in the Austnesfjord (Fig. 3), which is in the main spawning area of the Arcto-Norwegian cod. Stations and sections in the Austnesfjord are shown in Figure 3. A section is a transect with a series of stations. Austnesfjord was chosen because cod lar- vae are known to appear in high numbers (Ellertsen et al. 1977), and the dynamics of the current system are known (Fumes and Sundby 1981). During the 1980 cruise, a Wolfe wind recorder was placed on land in the fjord to continuously measure wind velocity and direction. In 1981, observations were also made in the main first feeding area, an open ocean bay (Fig. 2), for cod larvae. The objectives were to find these food parti- cle patches for cod larvae and to investigate the extent and densities of these patches in this exposed area. Distribution of cod larvae in the first feeding areas was studied from the Juday net (80 cm, 375 jum mesh) samples taken in vertical hauls from 30 to 0 m. In the Austnesfjord, three stations were taken on eight sections (Fig. 3). The vertical distribution of cod larvae in the Austnesfjord was investigated only when the ship was anchored. A total of 42 samples were taken by a submersible electric pump (Flygt B2125, 3.4 mVmin) at 5, 10, 15, 20, 25, 30, and 35 m depths every 3 h from 1600 h 13 May to 1000 h 14 May 1980. Fifteen cubic meters of seawater was sam- pled at each depth. Seawater was pumped through a 40 m long by 15 cm diameter hose and filtered through a Juday net (40 cm, 180 /xm mesh) into a large tank on deck. Cod larvae were preserved in 4% Formalin in 10%o seawater solution. Gut contents of FIGURE 2.— Map of the Lofoten area with stations and sections 21 April-8 May 1 98 1. The figures on the stations refer to number of cod larvae/m2 surface. 143 FISHERY BULLETIN: VOL. 82, NO. 1 TABLE 1. — Size frequency distribution of Artemia nauplii measured by the Hiac Particle Counter (n = 1542) and by microscope (n = 45). FIGURE 3. — Map of the Austnesfjord with stations. • Juday net and particle/zooplankton stations, position of the 24 h station -k, and the Wolfe wind recorder i£r. about 20 larvae from each depth were examined by dissecting the larval gut under the microscope. During 24-h stations in situ particle profiles, CTD profiles, and zooplankton pump profiles were made simultaneously every 2 h. On sections, zooplankton pump profiles were made on every second station. RESULTS In Situ Instrument Tests Results of the comparison between microscope and particle counter measurements is presented in Table 144 Size M Part o. of Anemia nai pi, counted by Iflm) cle counter Microscope 200-249 101 5 250-299 416 14 300-349 848 23 350-399 1 77 3 1. A chi-square test for independence in the 4X2 table (3 df) showed no significant difference (P < 0.05) between the two methods of measuringArterata nauplii. Paired tests between microscope and in situ parti- cle counts were done on data from two different 24-h stations in the Austnesfjord (Figs. 4, 5). Plankton pump samples were taken from 11 different depths on each profile, and the mean counts from these depths were tested against the mean in situ counts from the same depths. A comparison was also made between the mean of all plankton pump counts from each profile, and the mean of all in situ counts from the corresponding profile. During the first 24-h station, 19 vertical profiles were made. No significant differences (P < 0.05) was found when the mean counts (n = 19) from each of 1 1 different depths were compared, nor when the mean counts from the different profiles were compared. The same statistical test was made on data from 14 vertical profiles on the second 24-h station. There had been an increase in the variability of mi- crozooplankton both horizontally and vertically dur- ing this 24-h station (Fig. 5A, B). No significant differences (P < 0.05) was found between the mean in situ counts and the mean plankton pump counts when the different profiles were tested. We found, however, a significant difference (P < 0.05) when the mean counts from corresponding depths were tested. This difference was found between in situ and plankton pump counts both from 30 and 40 m depths. No significant difference (P < 0.05) was found between counts from 0, 0.5, 7.5, 10, 12.5, 15, 20, and 25 m depths. This difference may have resulted from samples having been taken at different depths. The in situ instrument was equipped with a depth detector, but the depth of the submersible pump was controlled only by the meter wheel on the winch. Distribution of Particle/Nauplii in the Fjord The vertical distribution of particles/nauplii for a 24-h station made during 22-24 April 1981 in the TILSETH and ELLERTSEN: FOOD ORGANISMS OF LARVAL COD A 21 22-24 Apr. 81 Austnesf jorden 01 05 09 13 17 21 01 FIGURE 4. — Isopleth diagrams of the particle concen- trations (per liter) (A), and nauplii (per liter) (B), center station, section 5 in Austnesfjord, 22-24 April 1981. 21 22 -24 Apr. 81 Austnesf jorden 01 05 09 13 17 21 01 05 09 H Austnesfjord is presented in Figure 4A and B. The maximum observed particle concentration was a small patch of 50 particles/1 at about 15 m (Fig. 4A). A patch of 40 nauplii/1 at the same depth was iden- tified from pump samples (Fig. 4B). The particle/ nauplii isolines in the upper 20 m show a tendency of ascending towards the surface at midnight, indicat- ing their diel vertical migration. This observation was repeated on another 24-h station made 6 d later at the same position (Fig. 5A, B). Particle concentration had increased markedly during this period; more than 50 particles/1 were found at 25-35 m depth on every profile. A very dense surface patch was found at midnight with more than 500 particles/1. Figure 5B shows a similar distribution of nauplii during the same 24-h station. Since there was no wind in the fjord and consequently little or no vertical turbulence, the hydrographic conditions during this 24-h station were perfect for this type of observation. This is shown in Figure 6 where the hydrographic conditions is presented by the temperature distribution in the upper 60 m. Figure 7AandB presents the particle (1 50-600 jiim) distribution from 0 to 40 m depth through a section of the Austnesfjord made at night on 27-28 April 1981 from 2130 to 0420 h. There was little or no wind in the fjord when the section was made. Patches of more than 100 particles/1 were found in the surface water of the outer parts of the fjord. A particle minimum layer (<10/1) was observed at 10 m in the middle of the fjord. In the bottom of the fjord three patches of more than 50 particles/1 were found at different depths. Figure 7B shows the naupliar distribution on the same section. Highest concentrations (> 100/1) were observed in the bottom of the fjord, at inter- mediate depths and in the surface water of the outer parts of the fjord. The same section made through the fjord the next day from 0950 to 1610 h (Fig. 8A, B) showed that the particle/nauplii distribution in the fjord had changed completely. A particle/nauplii minimum layer (< 10/ 1) was found from the surface down to about 20 m through most of the fjord length. The surface patches in the outer parts of the fjord had disappeared. Only 145 FISHERY BULLETIN: VOL. 82, NO. 1 28 -29 Apr. 81 Austnesfjorden 07 09 11 13 15 17 19 21 01 03 05 07 09 H B a » 28 - 29 Apr. 81 11 13 15 Austnesfjorden 17 19 21 23 01 03 05 07 09 H FIGURE 5.— Isopleth diagrams of the particle concentrations (per liter) (A), and nauplii (per liter) (B), cen- ter station, section 5 in Austnesfjord, 28-29 April 1981. 07 09 28 -29 Apr.81 11 13 15 Austnesfjorden 17 19 21 23 01 03 05 07 09 H 146 FIGURE 6. — Isopleth diagram of the temperature distribution, middle station, section 5 in Austnesfjord, 28-29 April 1981. TILSETH and ELLERTSEN: FOOD ORGANISMS OF LARVAL COD Austnesfjorden Hella 27 - 28 Apr. 81 A 0420 2 n. miles 21 30 H Austnesfjorden Hella 27 -28 Apr. 81 2 n. mile s 21JUH FIGURE 7.— Particle (A) and nauplii (B) distributions (per liter) in the upper 40 m of Austnesfjord, 27-28 April 1981, at 21.30 to 0420 h. (Particle size range 150-600 (im, nauplii all sizes.) one patch with >50 particles/nauplii per 1 was ob- served between 20 and 40 m at the bottom of the fjord. Effect of wind driven turbulence on vertically migrating particles is presented in Figure 9A, B, and C. The figure presents data collected continuously from 9 to 15 May 1980, on wind velocity and direc- tion, temperature, and particle distribution in the water column. Due to technical problems, only par- ticles within the size range 300-500 ju.m were measured by the particle counter in 1980. From 9 to 12 May the wind was blowing downfjord with varying velocity. On 1 2 May the wind changed direction 180° and blew upfjord with a velocity of 5- 1 0 m/s (Fig. 9A). Unfortunately, observations of temperature and par- ticle distribution were not made from 10 to 11 May. However, one 24-h station was made on 9 May during the period when the wind was blowing downfjord. At this time, the upper 10 m of the water column showed tendencies of mixing, and colder intermediate water masses were observed from 15 to 55 m above the transition layer. Within the cold intermediate water masses a particle maximum layer was found (Fig. 9C). It is believed that the wind was blowing the sur- face water downfjord and this was compensated for by intermediate water masses moving in the opposite direction. On 9 May we observed a patch of particle- rich intermediate water moving in from the outer part of the fjord. The particle isolines in the upper 10 m followed the isotherms (Fig. 9B, C). When the wind direction reversed and increased in velocity on 12 May (Fig. 9A), the fjord became more exposed to the wind force and the wave action from the open ocean outside the fjord. Under this condition the current system will reverse (Furnes and Sundby 1981). The surface water became completely mixed within about 24 h (Fig 9B), and no particle diel vertical migration was observed during this condition (Fig. 9C). The particle concentration decreased and became almost homogeneous from the surface to 40 m. 147 FISHERY BULLETIN: VOL. 82, NO. 1 Austnesfjorden -Holla 29 Apr. 81 a a 2 n. miles 1610 H B 0950 Austnesfjorden Holla 29 Apr. d1 2 n. miles 1610 H a Q 10H 20 30J 40 FIGURE 8.— Particle (A) and nauplii (B) distributions (per liter) in the upper 40 m of Austnesfjord, 29 April 1981, at 0950 to 1610 h. Distribution of Cod Larvae The highest concentration of cod larvae (140-290 larvae/m2) was observed in the middle of May at the bottom of the Austnesfjord both in 1980 and 1981 (Fig. 10). This has also been observed on previous cruises (Ellertsen et al. 1977). The research vessel was therefore anchored at the middle station on sec- tion 5, where 24-h stations were made. In 1 98 1 , the study of the distribution of cod larvae in the exposed open ocean bay of Vesteralsfjorden showed that larvae were only found on the innermost stations with a maximum of 4 larvae/m2 (Fig. 2), e.g., only two cod larvae in vertical Juday net hauls from 30 m depth. Gut contents of 738 cod larvae were examined from 39 pump samples. Fewer than 10 larvae were found in pump samples from 30 and 35 m depths from the 01-02 h pump profile and from 35 m depth from the 04-05 h pump profile. These larvae have not been included in the analysis (Fig. 11B). A total of 1,204 prey organisms were found, out of which 96.5% were identified as copepod nauplii. Only 1.7% of the prey organisms could not be identified. About 0.5% of the larval cod gut content was bivalve veliger larvae, copepod eggs, and phytoplankton (Peridinium sp.), and 1.3% was identified as copepod fecal pellets. The size distribution of the main prey organisms (e.g., copepod nauplii) ranged from 140 to 520 jiim with a mean size of 224 /xm (all measurements as carapace length). Gut content analysis of cod larvae is presented in Figure 1 IB as feeding incidence (percent larvae with gut content) and larval feeding ratio (number of prey organisms per larval gut). The feeding incidence varied between 73 and 100% in samples from the three pump profiles taken before midnight. In 6 1 % of these samples the feeding incidence was as high as 148 TILSETH and ELLERTSEN: FOOD ORGANISMS OF LARVAL COD 80- a. 0 q 50 >150 20- 40 r~ft~^~ Particle density \C-150 nos. per liter ' 100 40 fe\ 40 30 l^^x 207rO C/ 20 60 FIGURE 9. — Wind velocity (length of vector, see m/s scale) and direction from the abcissa (A), isopleth diagrams of temperature (B), and particle concentration (300-500 yum) distribution (C), at the middle station on section 5 in Austnesfjord, 9-15 May 1980. 90-100% .The larval feeding ratio was >1 prey/larval gut in all samples taken before midnight. In 71% of these samples the feeding ratio was >2 prey/larval gut and in 14% of the samples >3 prey/larval gut. In samples taken after midnight, however, the feeding incidence varied between 4 and 92%. The lowest level was found in pump samples from 25 m depth from the 01-02 h profile. In 38% of the samples taken after midnight the feeding incidence was <50%. Only in the last pump profile made at 09-10 h the larval feeding incidence was more than 50% in all samples. The feeding ratio was < 1 prey/larval gut in all sam- ples from 01-02 h profile, and <1 prey/larval gut in 61% of all samples taken after midnight. A feeding ratio level <1 prey/larval gut was not observed in samples taken before midnight. The highest feeding ratio observed in samples taken after midnight was 1.65 prey/larval gut from the 25 m depth samples taken from the 09-10 h pump profile. Distribution of Particles/Nauplii in Open Ocean Waters The main first feeding area of the Arcto-Norwegian cod is thought to be the waters outside the Lofoten islands and in the open ocean bay of the Vesterals- fjord (unpubl. data). Figure 12A and B shows the particle and nauplii distributions in the northeast section in the Vesteralsfjord. Plankton pump samples were only taken at every second station on the sec- tion. The figure shows a similar distribution pattern. However, due to the more frequent samples taken by the particle counter, a more accurate distribution picture of the particles on the section was achieved. 149 FISHERY BULLETIN: VOL. 82, NO. 1 LARVAE /M^ 80 • 60 40 20 H 0 SURFACE 29 APRIL 1980 1000M 7 6 3 2 1 LARVAE /M< STATIONS 140- 120- 1 DO 80 60 iO 20 SURFACE 5 MAY 1980 100OM 7 6 LARVAE/M2 4 3 2 1 STATIONS IX K 300- SURFACE • 12 MAY 1980 280 \ 1000M \ 1 ' 100 \ • 80- \ • 60 \y • \ 4 n \ 20 n • • 7 6 5 4 3 2 1 STATIONS IX LARVAE/M2 SURFACE 24 APRIL 1981 1000M LARVAE/M2 SURFACE 120 100 80 60 H 40 20 0 LARVAE /M2 SURFACE HO- 120 - 100 8 0 60H 40 20 H 0 4 2 STATIONS 29 APRIL 1981 1000M 7 6 5 4 3 2 1 STATIONS IX 8 MAY 1981 1000M 8 7 6 5 4 3 2 1 STATIONS IX FIGURE 10.— The average number of cod larvae/m2 surface on sections 1-8 and stations X and IX in the Austnesfjord, April-May 1980 and 1981. Sections were also made at four locations in the open water off the Lofoten islands. On three of these sections (Eggum, Myrland, and Fuglehuk), patches with high particle concentrations (>50/l) were ob- served about 1 1 km (8 n mi) off shore. All sections had low particle concentrations (10-30/1) in the surround- ing water masses (Figs. 13-15). The similarity of the positions of these three patches suggests that they are components of the same water mass with higher particle concentrations than the surrounding water masses. On the Skiva section (Fig. 16A-D) the parti- cle distribution patterns were more complicated. The section was surveyed during daytime and two patches were observed, one at about 5-10 m (>100 particles/1) and another 20-25 m (>50 particles/1). Particle concentration decreased further offshore. The same section was surveyed at night (Fig. 16C), and two surface patches were found. 150 TILSETH and ELLERTSEN: FOOD ORGANISMS OK LARVAL COD Q. LU o 16-17 HOURS 0 5 10 0-«. 10- """^-° °c 20- 30- ./" 19-20 HOURS 0 5 10 15 0 22-23 HOURS 01-02 HOURS 0^-05 HOURS 09-10 HOURS / 10 15 0 \ O / 10 0 10 0 a \ O J \ \ JOLarvae/rrr 10 Q- 20- LU Q 3 0 50 100 "I 2 50 100 v ? 4 1 \ \ • 3 a 50 100 J L. \ j a \ 50 ipo q_ 1 2 0 \\ / /I 50 ipo g_ 1 2 0 50 100 4 I / V »/oFI Naupl/larv K FIGURE 1 1 .—Distribution of first feeding cod larvae (per m3) (A), and the larval feeding incidence (% larvae with gut content) V and lar- val feeding ratio (nauplii/larval gut) O (B), during the 24 h sampling station, 13-14 May 1980, at middle station, section 5 in Austnesfjord. A Vesteralsf jorden 30 Apr 1 May 81 2000 ,2n.miles 5«H Q. a 20 30 0213 h B Vesteralsf jorden 30 Apr. - 1 May 2000 40-1 . 2 n.miles 02«H Figure 12. -Particle (A) and nauplii (B) distributions (per liter) in the upper 4 m on the section in Vesteralsfjord, 30 April- 1 May 1981. 151 FISHERY BULLETIN: VOL. 82, NO. 1 £ 0 £ a. 01 Q 10 20 1033 Eggum 26Apr.81 2 n. miles 1535 H 30 ^ 40 10 30 Eggum 26 Apr. 81 2 n. miles 1540H g-IOH Q 20 30- 40J FIGURE 13.— Particle (A) and nauplii (B) distributions (per liter) in the upper 40 m on the Eggum section, 26 April 1981. DISCUSSION Food particles found in the alimentary tract of larval cod consist, with few exceptions, of copepod nauplii in the size range of 140-520 tun. This observation did not differ significantly from that of Ellertsen et al. (1977), who found the size variation to be within 140- 600 /Am. The in situ instrument was set to detect par- ticles in this size range. Investigations have shown that in May copepod nauplii outnumber all other par- ticles in this size range in the Lofoten area (Ellertsen et al. 1977; Wiborg 1948a, b). The main objective when designing this instrument was to obtain a quick, reliable impression of naupliar distributions without laborious, time-consuming countings by microscope. The tests performed to compare the in situ instru- ment system and the plankton pump samples showed good agreement between the two methods. The critical food concentrations for first feeding cod larvae are not precisely known. They are thought to be on the order of 40-200 nauplii/1 based on studies of swimming activity, larval search volume, and oxygen requirements of first feeding cod larvae (Solberg and Tilseth 1984). Patches of particles/ nauplii with the required densities for first feeding cod larvae to survive were found in the spawning and first feeding area by these methods. The results presented in this paper show some of the dynamics in the formation and distribution in time and space of microzooplankton patches. The vertical distribution and density of nauplii changes due to the diel vertical migration of these organisms (Figs. 5, 6). The concentration of particles/nauplii in a patch was dependent on the hydrographic situation and on the distribution and concentration of micro- zooplankton in the water column (Figs. 5, 6). Conse- quently the vertical distribution of particles and nauplii will be dependent on factors such as hydro- graphic conditions and time of day when the ob- servations are made. Increased wind force caused mixing of the surface layers and led to a homogeneous vertical particle dis- tribution. No surface patch was observed at night during windy conditions, and the mean particle con- centration in the water column dropped steadily dur- 152 TILSETH and ELLERTSEN: FOOD ORGANISMS OF LARVAL COD A Myrland 26 Apr. 81 2126 Q. 41 ° 10 20 30- 40-1 2 n. miles 1714 H Myrland 26 Apr. 81 21°° 2 n. miles 1750H • 0- Q. 2? >20 ^~""— 10 Q 10- ] J / 20 -^ 20- V ■ A 10 <10 30- 10 z10^ > 10 c -20^p 10- 10 <10 E t 20~ a. Q 30 40- <10 Skiva 29 -30 Apr. 81 C 22*7 _i i_ 2 n.miles 0410H 10- Q. o 20- 30 4a q 2247 Skiva 29-30Apr. 81 0 2 n. miles 40J 041°H FIGURE 16.— The particle and nauplii distribution (per liter) in the upper 40 m on the Skiva section 27 April 1981 (A, B) and the particle and nauplii distribution 29-30 April 1981 (C, D). 154 TILSETH and ELLERTSEN: FOOD ORGANISMS OF LARVAL COD after this hydrographic condition had occurred. Dur- ing the following hours the larval feeding incidence increased again, most rapidly in larvae sampled at 15-30 m, indicating that food particle concentration did not become critical. (Note that the particle con- centration in Figure IOC only represents particles within 300-500 /Am size range.) However, the feeding ratio did not increase significantly, indicating a more difficult accessibility of food particles to the larvae. Similar observations were made by Lasker (1975, 1978), where stability of the water column in the upper 30 m was necessary for food organisms to aggregate in concentrations high enough to exceed the threshold for feeding stimulus of first feeding northern anchovy larvae. This observed reduced feeding in cod larvae cannot be explained by a diel feeding rhythm. Cod larvae are visual feeders; the light intensity threshold for feeding is 0.1 lx (Ellertsen et al. 1980). The light intensity in the upper 40 m does not drop below this level in Lofoten in May, and cod larvae are found with newly captured nauplii in the gut at all hours (Gj0saeter and Tilseth 1981). The number of cod larvae found in the main first feeding area was too small to do a comparison on lar- val feeding conditions. However, patches with particle/nauplii concentrations of more than 50/1 were observed on every section made in this area. Sizes of these patches were, on the other hand, small compared with the volume of water surveyed. The life span of these patches is probably very short because of the influence of biological and physical factors, especially when the upper 50 m of the water column is unstable. This is the normal situation in the Lofoten area in May (Furnes and Sundby 1981). Therefore, prey organism patches with concen- trations above the critical level for first feeding cod larvae would probably be broken down, due to increased water turbulence when the wind forces increase. A series of storms during the larval cod first feeding period could thereby have serious effects on larval feeding conditions and consequently on sur- vival and recruitment. ACKNOWLEDGMENTS The in situ instrument system was developed in collaboration with The Foundation of Scientific and Industrial Research at the Norwegian Institute of Technology. We thank I. Mohus, B. Holand, and I. O. Eriksen who were responsible for this work. We also thank 0. Ulltang at the Institute of Marine Research for his advice and help on statistics. LITERATURE CITED Boyd, C. N. 1973. Small scale spatial patterns of marine zooplankton examined by an electronic in situ zooplankton detecting device. Neth. J. Sea Res. 7:103-11 1. Ellertsen, B., E. Moksness, P. Solemdal, T. Str0mme, S. Tilseth, and V. 0iestad. 1976. The influence of light and food density on the feeding success in larvae of cod (Gadus morhua L.); field and laboratory observations. ICES, C. M. 1976/F:34, 31 p. [Processed.] Ellertsen, B., E. Moksness, P. Solemdal, T. Str0mme, S. Tilseth, T.Westgard, E. Moksness, and V. 0iestad. 1977. Vertical distribution and feeding of cod larvae in rela- tion to occurrence and size of prey organisms. ICES, C. M. 1977/L:33, 31 p. [Processed.) Ellertsen, B., P. Solemdal, T. Str0mme, S. Tilseth, T. Westgard, E. Moksness, and V. 0iestad. 1980. Some biological aspects of cod larvae (Gadus morhua L.). Fiskeridir. Skr. Ser. Havunders. 17:29-47. Eriksen, J. O. 1981. "Micro-count". Particle datalogger. Program man- ual. Sintefrep. STF 48 F 81019, 203 p. [Processed.] Eriksen, J. O., and I. Mohus. 1981. "Micro-count". Particle datalogger. User's man- ual. Sintefrep. STF 48 F 81017, 54 p. [Processed.] Furnes, G.K., and S. Sundby. 1981. Upwelling and wind induced circulation in Vestfjor- den. In R. Saetre and M. Mork (editors), Proceedings from the Norwegian Coastal Current Symposium, Geilo, Norway, 9-12 Sept. 1980, Vol. I, p. 152-177. Univ. Bergen, Norway. GJ0SAETER, H., AND S. TlLSETH. 1981. Primary growth increments in otoliths of cod larvae (Gadus morhua L.) of the Arcto-Norwegian cod stock. Fiskeridir. Skr. Ser. Havunders. 17:287-295. H.JORT, J. 1914. Fluctuations in the great fisheries of northern Europe viewed in the light of biological research. Rapp. P.— V. Reun. Cons. Perm. Int. Explor. Mer 20:1-228. HOUDE, E. D. 1978. Critical food concentrations for larvae of three species of subtropical marine fishes. Bull. Mar. Sci. 28:395- 411. HOUDE, E. D., AND R. C. SCHEKTER. 1978. Simulated food patches and survival of larval bay anchovy, Anchoa mitchilli, and sea bream, Archosargus rhomboidalis. Fish. Bull., U.S. 76:483-487. Hunter, J. R. 1972. Swimming and feeding behavior of larval anchovy Engraulis mordax. Fish. Bull., U.S. 70:821-838. 1981. Feeding ecology and predation of marine fish lar- vae. In R. Lasker (editor), Marine fish larvae, morphol- ogy, ecology, and relation to fisheries, p. 33-77. Univ. Wash. Press, Seattle. Hunter, J. R., and G. L. Thomas. 1974. Effect of prey distribution and density on the searching and feeding behaviour of larval anchovy Engraulis mordax Girard. In J. H. S. Blaxter (editor), The early life history of fish, p. 559-574. Springer- Verlag, Berl. Lasker, R. 1975. Field criteria for survival of anchovy larvae: The rela- tion between inshore chlorophyll maximum layers and successful first feeding. Fish. Bull., U.S. 73:453-462. 1978. The relation between oceanographic conditions and 155 FISHERY BULLETIN: VOL. 82, NO. 1 larval anchovy food in the California Current: identifica- tion of factors contrihuting to recruitment failure. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 173:212-230. LASKER, R.. AND J. R. ZWEIFEL. 1978. Growth and survival of first-feeding Northern anchovy larvae {Engraulis mordax) in patches containing different proportions of large and small prey. In J. H. Steele (editor). Spatial patterns in plankton communities, p. 329- 354. Plenum Press, N.Y. Laurence, G. C. 1974. Growth and survival of haddock (Melanogrammus aeglefinus) larvae in relation to planktonic prey concen- tration. J. Fish. Res. Board Can. 31:1415-1419. May, R. C. 1974. Larval mortality in marine fishes and I he critical period concept. //; J. H. S. Blaxter (editor). The early life history offish, p. 3-19. Springer-Verlag, Berl. Moms, I. 1981. "Micro-count". Particle datalogger. Equipment man- ual. Sintef rep. STF 48 F 81018, 90 p. [Processed.] PUGH, P. R. 1978. The application of particle counting to an understand- ing of the small-scale distribution of plankton. In J. H. Steele (editor), Spatial patterns in plankton communities, p. 111-129. Plenum Press, N.Y. SOLBERT, T., AND S. TlLSETH. 1984. Growth, energy consumption and prey density requirements in first feeding larvae of cod ((Indus morhua L.). In E. Dahl, D. S. Danielssen, E. Moksness, and P. Solemdal (editors). The propagation of cod Gadus morhua L., p. 145-166. F10devigen Rapp. Ser. 1. Steele, J. H. 1978. Some comments on plankton patches. In J. H. Steele (editor). Spatial patterns in plankton communities, p. 1- 20. Plenum Press, N.Y. TUNGATE, D. S., AND E. REYNOLDS. 1980. The MAFF on-line counting system. Fish. Res. Tech. Rep., MAFF Direct. Fish. Res., Lowestoft, (58), 11 p. Werner, R. G., and J. H. S. Blaxter. 1980. Growth and survival of larval herring (Clupea harengus) in relation to prey density. Can. J. Fish. Aquat. Sci. 37:1063-1069. WlBORG, K. F. 1948a. Experiments with the Clarke-Bumpus plankton sam- pler and with a plankton pump in the Lofoten area in Northern Norway. Fiskeridir. Skr. Ser. Havunders. 9(2):l-32. 1948b. Investigations on cod larvae in the coastal waters of Northern Norway. Fiskeridir. Skr. Ser. Havunders 9(3):l-27. 156 EFFECTS OF SIZE AND TIME OF RELEASE ON SEAWARD MIGRATION OF SPRING CHINOOK SALMON, ONCORHYNCHUS TSHAWYTSCHA R. D. Ewing,' C. E. Hart,2 C. A. Fustish,3 and Greg Concannon' ABSTRACT Juvenile spring chinook salmon, Oncorhynchus tshawytscha, from Round Butte Hatchery on the Deschutes River, Oregon, were released monthly into a 3.7 km fish ladder. Fish released into the ladder from February to May migrated through the ladder in mid-May in both 1977 and 1978. Fish released after mid-May migrated through the ladder within 2 weeks after release. The extent of migration decreased progressively in fish released after 15 June. The migration was presumably photoperiod dependent, although temperature may have acted both as a releasing factor for migration and as a stimulus for growth. In the fish ladder, size of the fish remained constant over a 3-week migration period, suggesting that larger fish migrated before smaller fish. After a migration of 213 km, fish captured at the Dalles Dam had very large apparent growth rates, suggesting that larger fish were faster migrants. Maximum survival of juvenile salmonids after release from hatcheries is dependent upon their rapid migration to the sea (Raymond 1979). Delays in this seaward migration may subject the juveniles to starvation and stress which rapidly deplete their numbers (Miller 1952, 1958). Residual hatchery juveniles in a river often have an impact on wild stocks of fish through piscivory (Sholes and Hallock 1979) and competition for food (Chapman 1966). Rapid migration of hatchery juveniles ensures max- imum survival to adulthood with minimal interaction with wild stocks. Timing and duration of the physiological conditions which result in migratory behavior are still relatively unknown. Timing of seaward migration in juvenile salmonids depends upon a number of environmental factors, including photoperiod (Wagner 1974), tem- perature (Solomon 1978), water flow (Mains and Smith 1964), and fish size (Wagner 1974). The interrelationships between these are not well understood, but the available data suggest that these relationships may be complex. Hoar (1958) and Baggerman (1960) have postulated that these environmental factors act as "releasers" which, in conjunction with a physiological readiness to migrate, trigger overt migrational behavior. 'Corvallis Fish Research Laboratory, Oregon State University, Corvallis, OR 97331. department of Zoology, Oregon State University, Corvallis, OR 97331. 'Oregon Department of Fish and Wildlife, Research and Develop- ment Section, Corvallis, OR 97331. Manuscript accepted August 198:!. FISHERY BCLLETIN: VOL. 82. NO. 1. 1984. In most river systems, the relative influence of such factors is estimated by extensive sampling programs which use multivariate analysis of the data. Control of environmental variables in such a system is not possible. Furthermore, the size of many river sys- tems prevents an unbiased sampling of juveniles dur- ing migration. It is difficult, therefore, to obtain reliable estimates of the size of fish at migration, the timing of migration, and the influence of the environ- ment on that timing. In the present study, an unused fish ladder provided a relatively constant environment for migration of juvenile spring chinook salmon, Oncorhynchus tshawytscha, over a 3.7 km distance. Serial releases of hatchery-reared juveniles into this system permit- ted an investigation of the timing of seaward migra- tion, the duration of the migration tendency of the juveniles, and the relationship of several environ- mental variables to seaward migration. METHODS Study Area The study area included the lower 175 km of the Deschutes River, Oreg., and the lower Columbia River from its confluence with the Deschutes River to the Dalles Dam (Fig. 1). Rearing Conditions Progeny from spring chinook salmon spawned at 157 FISHERY BULLETIN: VOL. 82, NO. 1 PELTON REGULATION DAM< P€LTOn\ PELTON DAM<^~ - LADOER c^.,k.^ D.,-r-r,- ^... jUrOUND BUTTE HATCHERY ROUND BUTTE DAM^SS -V-I80 FIGURE 1.— Map of the lower 175 km of the Deschutes River and its confluence with the Columbia River. Numbers refer to kilometers from the mouth of the Deschutes River. Round Butte Hatchery (river km 175 from the Columbia River) in 1976 and 1977 were used for experiments in 1977 and 1978, respectively. Eggs from 1976 brood fish were incubated in Heath4 incubators in 10UC spring water, and the resulting fry were reared in raceways using the same water source. Eggs from 1977 brood fish were divided into two groups. One group was reared under conditions as described above and referred to as "fast-reared". The second group of eggs was incubated in Heath incubators in spring water chilled to 5°-6°C. The resulting fry were transferred to raceways and reared in 7°-8°C tail-race water from Round Butte Dam. After 2 mo, the group was transferred to 10°C spring water and reared there until release. This group was referred to as "slow-reared" and was released in March 1979 as yearlings. ■"Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. In May and June, production lots of fast-reared spring chinook juveniles were released into the Deschutes River below Pelton Regulation Dam. At this time, experimental groups were transferred to oval fiber glass ponds supplied with 10UC spring wa- ter at 9.5 1/s. In May 1977, 5,600 fast-reared spring chinook juveniles (average fork length 10.0 cm) were transferred to a fiber glass pond and reared there through June 1978. In late March 1978, 2,500 fast- reared fish (average fork length 8.5 cm) were transferred to a fiber glass pond and reared there through August. All fish were reared under a natural photoperiod and fed to repletion daily with Oregon Moist Pellet. Seaward Migration Migratory behavior of the spring chinook salmon was assessed by the release and recapture of hatchery-reared juveniles from two groups. Migra- tion tendency of the experimental groups was assessed by monthly release of about 200 fish into the upper end of Pelton ladder during 1977 and 1978 (Fig. 1). The ladder is 3.7 km long and is constructed with concrete walls and bottom except for a 1.1 km central section which is a natural stream channel. It is supplied with water from Lake Simtustus (directly above Pelton Dam) at a constant flow rate of 1,130 1/s. Maximum depth of the ladder is 2. 1 m. The ladder is closed by revolving screens at both the upper and lower ends. A trap located at the lower end of the lad- der was used to capture migrants. Temperature of the water at the lower end of the ladder was measured by a thermograph placed near the trap. Fish from the various experimental groups were identified upon recapture in the trap at the lower end of the ladder by unique combinations of polystyrene dye (Phinney et al. 1967) and fin clips. The trap was checked 5 d a week during May and June and 2 d a week during the remainder of the year. Fish captured in the trap were considered migrants while those remaining in the ladder following the date of peak recapture were assumed to be residuals. Fork lengths and marks of each migrant were recorded upon cap- ture. In January 1978, the ladder was drained and all residual fish from the 1977 studies were removed before the 1978 releases. The second group of hatchery-reared fish used for assessment of migration were production lots of fast- reared juvenile chinook released into the Deschutes River immediately below Pelton Regulation Dam (river km 161). These fish were marked with coded wire tags (Jefferts et al. 1963). In 1977, 62,000 fast- reared fish were released on 2 May and 73,000 fast- 158 EWING ET AL.: EFFECTS OF SIZE AND RELEASE TIME ON SALMON reared fish were released on 3 June. These fish averaged 9.7 cm and 11.2 cm FL, respectively. On 31 May 1978, 121,000 fast-reared fish, which had been graded according to fork length, were released in two groups of 95,000 and 26,000 fish to test the effects of size on migration and survival to adulthood. These fish averaged 10.9 and 11.8 cm FL, respectively. Downstream movement in both years was monitored in the Columbia River at the Dalles Dam (52 km downstream from the mouth of the Deschutes River) by gatewell sampling conducted by the National Marine Fisheries Service and the Oregon Depart- ment of Fish and Wildlife. Sampling was conducted 5 d a week throughout May and June. Juveniles originating at Round Butte Hatchery were identified by analysis of coded wire tags. Apparent Growth Rates Apparent growth rates in Pelton ladder and in the Deschutes River were calculated from the size of the juveniles released into the ladder or the river and the size and time at which they were recaptured. Actual growth rates could not be measured, because selec- tive mortality of small fish or migration of larger ones could not be estimated. Differences in fork lengths were tested for significance at the 95% confidence level using Student's t test. RESULTS Timing of Migration Maximum migration of chinook salmon juveniles released in February and March into Pelton ladder occurred between mid-May and the first of June in both 1977 and 1978. There was little migration in these groups before or after this 4-wk period (Tables 1, 2). Fish released in April showed two peaks in migration. A large percentage of the fish moved through the ladder within 2 wk after release, while a second peak of migration occurred during the last 2 wk of May. Fish released in early May also had a large percent migration within 2 wk after release, but the greatest percent migration occurred during the first 2 wk in June. When chinook salmon juveniles were released from June to November, most of the fish moved through the ladder within 7 d after release. The maximum percent migration within 7 d after release occurred in fish released in early June 1977 (Fig. 2) and in mid-June 1978 (Fig. 3). Fish released in August and at later times had reduced migration and had a higher tendency to become residual (Tables 1, 2). Migration of slow-reared fish released into Pelton ladder from May to August 1978 was less than half that of fast-reared fish released at the same time (Fig. 3B). Daily migrations of two groups released in February and March 1978 were compared with those from 8 May to 8 June. Movement of both groups was coin- cidental throughout this period (Fig. 4), suggesting that environmental factors such as temperature influenced migration tendency. Temperatures in the ladder varied seasonally due to solar warming (Fig. 5). Maximum temperatures of 17°C were attained in August 1977 and in July and August 1978. Tem- peratures in both years exceeded 13 °C by June, sug- gesting a possible temperature threshold for migration. While the relationship between migration and temperature was very poor (correlation coeffi- cient, R2 = 0.074), there may have been a tendency for peaks in seaward migration to occur 1-2 d after transient increases in temperature (Fig. 4). IOO z o 5 80-1 a: o 5 60 S 40 20 IOO RELEASE M DATE 8 M 31 A 12 IOO H 200 M 2 M II J 14 J 12 A 9 0 14 N 16 X LENGTH 7 3 8 5 91 97 IOI 112 12 0 134 14 9 168 18 0 19 1 FIGURE 2. — Percentage seaward migration within 7 d following release for each group of fast-reared spring chinook salmon released into Pelton ladder in 1977. Above each bar is the number of fish released. Lengths are means of samples of 30 fish taken from the population at the time of release. IOO 80 60 40- 20 I, OOP RELEASE DATE F 14 M 15 A 15 M 15 J 15 J 14 A 15 XLENGTH 63 80 89 99 116 129 149 M 15 n j 15 J 14 83 98 114 A 15 127 FIGURE 3. — Percentage seaward migration within 7 d following release for each group of spring chinook salmon released into Pelton ladder in 1978. A) Fast-reared chinook salmon. B) Slow-reared chinook salmon. Above each bar is the number of fish released. Lengths are means of samples of 30 fish taken from the population at the time of release. 159 FISHERY BULLETIN: VOL. 82, NO. 1 Table 1.— Percentage downstream migration for fast-reared spring chinook salmon released into the Pelton ladder in 1977. Release date: 8 Mar. 31 Mar 12 Apr, 2 May 11 May 3 June 14 June 12 July 9 Aug 9 Sept. 15 Oct. 16 Nov. Capture X length (cm): 7.2 8 5 9.1 9 7 10 2 11.2 120 13.4 14 9 16.8 18.0 19.1 dates n: 200 99 194 100 200 100 198 198 200 199 200 175 3/1-3/15 3 5 3/16-3/31 1.0 4 1-415 10 1.0 17 5 4/16-4/30 0 0 0 0 0 5 5/1-5/15 1 5 1.0 3 0 S 'i 370 5/16-5/31 34.5 420 27 0 19.0 5 5 6/1-6/15 8.0 16 0 180 37.0 355 78 0 7.0 6/16-6/30 0.0 1.0 0.5 6.0 3.0 40 35 0 7/1-7/15 0 5 0.0 0.0 1)0 0 5 1 0 20 40.0 7/16-7/31 0 5 0 0 0 0 1.0 0.0 0.0 o 5 18.5 8/1-8/15 00 0 0 00 0 0 o o 0.0 0 0 05 260 8/16-8/31 00 0.0 0 5 0 0 0 0 1 0 1.0 0 5 7 5 9/1-9/15 0.0 0.0 0.0 1.0 o o 1.0 3 5 0 b 2.0 29 0 9/16-9/30 0 0 0 0 o o 0.0 0.0 0 0 00 0 0 1 0 3 0 10/1-10/15 0.0 0.0 i) 5 0 0 0 0 0 0 0 o 0.0 0.0 00 10/16-10/31 0.0 0 0 0 0 0.0 0.0 0 0 0.0 0.0 0.0 0.0 7.5 11/1-11/15 00 0 0 0.0 0.0 0 0 0 0 0.0 0.0 0.0 00 0.0 11/16-11/30 0.0 0.0 0 0 0 0 oo 0 0 0 0 00 05 0.5 2 5 50 12/1-12 15 0.0 0 0 0.0 00 00 0 0 0.0 0.0 00 1.5 1.5 0.5 12/16-12 31 00 0 0 on 0 0 I 1 0 (III 0.0 0.0 0.0 00 00 00 1/1-1/15 0.0 0 0 0 0 0 0 00 0 0 0.0 0 0 0.0 2 0 1.5 1 0 Total percentage migration 505 61 0 62 0 73.0 81.5 85.0 490 60.0 37 0 36 0 13.0 6 5 Percent residuals 0 0 0.0 0 0 2 0 0 0 00 0 0 2.0 14.5 31.5 43 5 55 0 Total percentage recovered 50.5 61 0 62 0 750 81.5 850 49.0 620 51 5 67 5 56.5 61.5 TABLE 2. — Percentage downstream migration over semimonthly intervals for fast-reared spring chinook salmon released into the Pelton ladder in 1978. R, lease daie 14 Feb. 15 Mar, 1 5 Apr. 15 May 1 5 June 14 July 1 5 Aug. Capture .X ength (cm): 6 3 80 89 99 116 12.9 14.9 dates n: 1,000 199 198 192 96 192 200 2/15-2/28 0 1 3/1-3/15 0 0 3/16-3/31 0 0 3 0 4/1-4/15 0 0 0 0 4/16-4/30 0.1 1 0 37 0 5/1-5/15 17 3 80 2.5 5/16-5/31 648 62 8 328 41 7 6/1-6/15 3 8 8.5 12 0 33 8 6/1 6-6/30 1.7 0 0 1.0 3.1 77 0 7/1-7/15 0 5 0 0 0 0 0 0 00 7/16-7/31 1 3 1.0 0.0 0 0 2.0 55.0 8/1-8/15 Oil 0.0 0.0 0.0 0 o 0 5 8/16-8/31 0 7 0 0 0.0 0 5 1.0 0 5 54 5 9/1-9/15 2.2 ..I \, 0 0 2 5 1 0 2 5 50 Total percen m gration 92 5 84.8 85 3 81 6 81 0 58 5 59 5 Recovery of Released Fish In 1978, the greatest recovery of fish liberated into Pelton ladder (92.5%) was from the large group of 1,000 fish released on 14 February (Table 2). From 81.0 to 85.3f/(: of the fish released from 15 March through 15 June were recovered. Only 58.5 and 59.5f7f of the fish released on 14 July and 15 August, respectively, were recovered in the trap as migrants. Presumably the remainder were residuals in the ladder. In 1 97 7, recovery of both migrants and nonmigrants from all groups was lower than in 1978 (Table 1), although the extent of migration of fish released near the time of maximum migration tendency on 11 May and 3 June was 8 1.5 and 85 c/'c, respectively, similar to that observed for most release groups in 1978. Few residual chinook salmon from releases before August 1977 were found when the ladder was drained in January 1978. Nonmigrant fish were recaptured in increasing numbers from releases from 12 July on. Size and Growth Relationships Growth rates of juvenile chinook salmon reared at Round Butte Hatchery were 0.046 and 0.058 cm/d 160 EWING ET AL.: EFFECTS OF SIZE AND RELEASE TIME ON SALMON rl6 TABLE 3. — Apparent growth rates of juvenile chinook salmon re- leased into Pelton ladder, 1977 and 1978. ■Z. 16-, o -j Release Av srage reca pture Apparent growth i= '«" A 2 m m date date rate (cm/d) or 12-1 1977 3/8 5/24 0078 - 10 J . . t / ■ V | v // \ 3) • 5j 4/12 5/2 6/1 5/28 0081 0.124 y \ /v\Js\ / ; 33 12 m Hatchery 5/11 6/4 0.120 0.048 V b- \ / A i V < i*f V o 1978 2/14 5/27 0034 < "- O 3/15 5/20 0071 Q 2- vj-' ' 4/15 5/26 0 097 -*~»— • -^ti -10 3 Hatchery 5/15 6/7 0 097 8 10 12 14 16 18 20 22 24 26 28 30 13 5 7 JUN 0 046 MAY I978 Fic.l'RE 4. — Daily percent seaward migration from 8 May to 8 June for groups released 14 February (solid circles) and 15 March (open circles). Temperature (triangles) is the average daily temperature. 20 or or UJ o S UJ < or UJ 4 - J I L JAN MAR MAY JUL SEP NOV FEB APR JUN AUG OCT DEC FIGURE 5. — Average monthly temperature in Pelton ladder in 1977 (solid circles) and 1978 (open circles). TABLE 4. — Fork lengths of juvenile chinook salmon at time of release into Pelton ladder and at time of recapture, 1977 and 1978. Values are means ± standard errors. Number of samples is given in parentheses. Date Mean fork 1 Bngth Date Mean fork I sngth of at release of at recapture release (cm) recapture (cm) 1978: 2/14 6.5±0 1 60) 5/15-6/9 13 4+0.1 245) 3/15 8 2±0.1 30) 5/16-5/24 13.0+.0 1 81) 4/15 9.0±0.1 100) 5/22-5/31 13.0+0 2 42) 5/15 10.1 ±0.1 29) 6/5-6/9 12 3±0 1 54) 6/15 1 1 3±0 1 30) 6/16 1 1.9±0 1 30) 7/14 13 7±0 .1 30) 7/17 13 2±0.2 30) 8/15 15 1±03 30) 8/16 15.5+0 4 9) 9/15 17.1±0.3 28) 9/18 17.2+.0.2 30) 1977 3/8 7.5±0 1 30) 5/24 13 5±0 1 25) 3/31 8.5±0.1 30) 5/24-5/28 13.5±0.1 26) 4/12 9 4+0 1 30) 6,1 13.4±0 1 26) 5/2 9 7±0 1 30) 5/28 12 9±0 1 7) 5/11 9.9±0 .1 29) 5/12-6/4 11 7±0 1 61) 6/3 1 1.2±0 1 30) 6/3 11.4±0 1 25) 6/14 12.0+.0.1 30) 6/17 12 3±0 .1 30) 7/12 1 3 5±0 1 60) 7/15 1 3 9±0 1 28) 8/9 15.1±02 88) 8/15 16 1+02 20) 9/9 16 7±0 2 88) 9/9-9/13 16.8±0 2 42) 10/15 17 5±0.5 30) 10/17 18 4+0 4 15) for fast-reared fish in 1977 and 1978, respectively. Slow-reared fish in 1978 grew at 0.043 cm/d. Apparent growth rates offish placed in Pelton ladder varied from 0.034 to 0.124 cm/d (Table 3). These apparent growth rates increased in later introduc- tions, reflecting the increasing water temperature of the ladder (Fig. 5). There was no evidence for differences in migration timing by fish of different sizes. Fork lengths of fish recaptured in the trap within a few days of release were usually not significantly different (P > 0.05) from those of fish at release (Table 4). However, fish recaptured from the large group of juveniles released on 1 4 February 1978 were similar over a 3-wk period (Table 5), suggesting that faster growing fish were migrating more rapidly that slower growing fish. Apparent growth rates of marked spring chinook juveniles released below Pelton Regulation Dam in 1977 were calculated from fork lengths of recaptured fish at the Dalles Dam, after a migration distance of TABLE 5.— Mean fork lengths of juvenile spring chinook salmon recovered in 1978 after release into Pelton ladder on 14 Feb- ruary 1978. Values are means ± standard errors. Number of samples is given in parentheses. Date of recovery Fork length (cm) 2/17 5/15 5/16 5/18 5/22 5/24 5/30 6/1 6/5 6/9 67+01 (30) 12.8±0.1 (30) 13 2±0.1 (30) 13 1±0.1 (30) 13.4±0.1 (30) 13.2±0.1 (30) 13.9±0.1 (30) 13 5±0.2 (19) 13.6±0.1 (21) 11.2±0.1 (25) 213 km (Table 6). This apparent growth rate is nearly twice that of fish reared at Round Butte Hatchery. 181 FISHERY BULLETIN: VOL. 82. NO. 1 TABLE 6.— Fork lengths and apparent growth rates of juvenile spring chinook salmon recaptured at the Dalles Dam after release into the Deschutes River. 1977. Fork lengths are means ± standard errors for the number of samples shown in parentheses. Apparent growth rate Recapture date Fork length (cm) (cm/d) 2 May release (9 7±0.1 cm fork length) 5/27 11.6±0 2 (12) 0075 6/3 12.2±0 1 (19) 0078 6 7 12 2±0 1 (22) 0.077 6/8 12.2±0.1 (21) 0076 3 June release (1 1 2±0 1 cm fork length) 6 7 11.9±0.1 (23) 0.168 6/8 11.8±0.1 (30) 0 120 DISCUSSION Determination of the migratory characteristics of juvenile chinook salmon during smolting has been complicated by the variety of migratory behaviors displayed by the juveniles. Some fry migrate from tributaries shortly after emergence from the gravel (Reimers 1973; Ewing et al. 1980), but there is little evidence that the fry move into the estuary at that time (Schluchter and Lichatowich 1977). In some stocks, a general movement of fish through the river occurs during the fall of the first year (Reimers 1973) with a majority of the fish entering the ocean during the fall of the first year (Reimers 1973; Schluchter and Lichatowich 1977; Buckman and Ewing 1982). In other stocks, seaward movement occurs primarily in the following spring when the fish are more than 1 yr old (Mains and Smith 1964; Diamond and Pribble 1978: Raymond 1979). Krcma and Raleigh (1970) reported migration of juvenile chinook salmon into Brownlee Reservoir (Snake River, Idaho) in fall and spring for 2 consecutive years. The migration pattern seems to depend upon stock, size, and rearing con- ditions and may be highly variable. It is therefore important in the culture of various stocks of juvenile chinook salmon to determine the timing of maximum migration tendency. In the present study, the major migration of fish released early into Pelton ladder occurred in mid- May. Fish from the same brood released into the Deschutes River at about this time were found to migrate 213 km to the Dalles Dam within 7 d, sug- gesting that the migrational behavior was seaward directed (Hart et al. 1981). It is difficult to confirm in the Deschutes River that the release of fish into Pelton ladder 1 mo before the time of maximal migra- tion tends to increase the time during which the fish will migrate. Release of the fish 1 mo later than the time of maximal migration tends to decrease the time for migration. It is important to note that it is not necessary to release the fish early to insure that all migrate to sea. Releases late in the migration period were recovered to the same extent as those released earlier. Migration tendency seems to be retained for some time, even though the fish are not permitted to begin migration. These results suggest that late re- leases hasten the seaward migration, thus removing the populations of hatchery fish quickly from the river system and affording maximum protection to the wild stocks. Those groups released later than July were recap- tured in the trap in decreasing numbers (Tables 1,2). In 1977, nonmigrant fish were recaptured in increas- ing numbers from releases after 12 July (Table 1). This result indicates that the decrease in numbers of fish recaptured at the trap was due to decreased migration tendency and not due to increased mor- talities at the higher water temperatures. A major advantage of utilizing a closed system such as the Pelton ladder for studies of migration was that fish populations and flows could be effectively con- trolled. Variables which remained uncontrolled in- cluded photoperiod, lunar periodicity, temperature, and food supply. Of these, photoperiod seems the most important in stimulating seaward migration. Previous studies utilizing a closed system for study- ing seaward migration of steelhead trout, Salmo gairdnvri, (Zaugg and Wagner 1973; Wagner 1974) and coho salmon, Ocorhynchus kisutch, (Lorz and McPherson 1976) also concluded that photoperiod was an important factor affecting the timing of sea- ward migration. Lunar phase has been suggested to affect the onset of migration, based on the correlation between peaks in plasma thyroxine levels and lunar phase (Grau et al. 1981). Assuming maximal migration occurred on 22 May in both 1977 and 1978, this date correspond- ed to the time of a new moon in 1977 and that of a full moon in 1978. These brief data do not support the hypothesis that the migration is influenced by the lunar phase. Temperature may have had a dual influence on migration. Temperature has been suggested as a releasing factor for salmon migration (Hoar 1958; Baggerman 1960), but we were unable to show a statistical relationship between daily migration and average daily temperature (Fig. 4). Temperature also serves to increase growth rates in salmonids in the presence of abundant food supplies. Wagner (1974) suggested that a critical size was required in steelhead if migration were to take place. The importance of size on migration of spring chinook salmon can be seen by comparing the extent of migration of the slow- and fast-reared fish in 1978 (Fig 3). The slow-reared fish may have failed to mi- 162 EWING ET AL.: EFFECTS OF SIZE AND RELEASE TIME ON SALMON grate because they did not reach a critical size and/or growth rate by the appropriate photoperiod. Migra- tion from Pelton ladder seemed to occur as fish reached a particular size, since during a 3-wk period of migration, there was no difference in average fork length of the fish recaptured (Table 5). From estimated growth rates (Table 3), fish at the end of the migration period might be expected to be nearly 2 cm larger than those at the beginning. This influence of size on migration could be best demonstrated in fish recaptured at the Dalles Dam after a migration distance of 213 km. Apparent growth rates were much higher than that of fish reared at Round Butte Hatchery, suggesting that a selection for larger fish occurs during the long migration distance. A major concern in utilizing a closed system for studying seaward migration is the importance of aggressive behavior by resident fish toward newly introduced fish. Chapman (1962) found that aggres- sive behavior of resident fish may be partly respon- sible for emigration of fish introduced into the system. Aggressive behavior may have caused the rapid movement immediately following release for the March and April release groups in both 1977 and 1978. Further movement of these fish was not ob- served until May. Alternatively, migration in these fish immediately after release may have been due to disorientation of the fish upon release and a passive drifting downstream with the current. Fish released earliest into Pelton ladder migrated first in both 1977 and 1978. The importance of determining appropriate times for hatchery releases of spring chinook salmon in order to obtain maximum seaward migration is de- monstrated by the short time during which maximum migration occurred (Tables 1, 2). In both 1977 and 1978 peak migration occurred within a period of a few weeks. Releases made on either side of this time period exhibited decreased migratory activity. The use of model systems, such as the Pelton ladder, to determine when peak migration occurs can benefit hatchery programs by suggesting sizes and times for release of salmonids which maximize seaward mi- gration. ACKNOWLEDGMENTS We thank the members of the Deschutes River Salmon Study, Round Butte Hatchery personnel, and biologists of Portland General Electric Company for their help and cooperation throughout this study. We acknowledge the special assistance of Ray Hill, Jerome Diamond, Zeke Madden, Garet Soules, and Richard Aho. This study was supported by a grant from Portland General Electric Company to the Oregon Department of Fish and Wildlife. LITERATURE CITED Baggerman. B. 1960. Factors in the diadromous migrations of fish. Symp. Zool. Soc. Lond. 1:33-58. BUCKMAN, M., AND R. D. EWING. 1982. Relationship between size and time of entry into the sea and gill (Na+K)-ATPase activity for juvenile spring chinook salmon. Trans. Am. Fish. Soc. 111:681-687. Chapman, D. W. 1962. Aggressive behavior in juvenile coho salmon as a cause of emigration. J. Fish. Res. Board Can. 19:1047-1080. 1966. Food and space as regulators of salmonid populations in streams. Am. Nat. 100:345-357. Diamond, J., and H. J. Pribble. 1978. A review of factors affecting seaward migration and survival of juvenile salmon in the Columbia River and ocean. Oreg. Dep. Fish Wildl. Inf. Rep. Ser., Fish. 78-7. Ewing, R. D.. C. A. Fustish, S. L. Johnson, and H. J. Pribble. 1980. Seaward migration of juvenile chinook salmon without elevated gill (Na+K)-ATPase activities. Trans. Am. Fish. Soc. 109:349-356. Grau, E. G., W. W. Dickhoff, R. S. Nishioka, H. A. Bern, and L. C. Folmar. 1981. Lunar phasing of the thyroxine surge preparatory to seaward migration of salmonid fish. Science (Wash., D.C.) 211:607-609. Hart, C. E., G. Concannon, C. A. Fustish, and R. D. Ewing. 198 1 . Seaward migration and gill (Na+K)-ATPase activity of spring chinook salmon in an artificial stream. Trans. Am. Fish. Soc. 110:44-50. Hoar, W. S. 1958. The analysis of behaviour of fish. In P. A. Larkin (editor), The investigation of fish-power problems. Univ. Br. Columbia, Inst. Fish., p. 99-111. Jefferts, K. B„ P. K. Bergman, and H. F. Fiscus. 1963. A coded wire identification system for macro- organisms. Nature (Lond.) 198:460-462. Krcma, R. F., and R. F. Raleigh. 1970. Migration of juvenile salmon and trout into Brownlee Reservoir, 1962-65. U.S. Fish Wildl. Serv., Fish. Bull. 68:203-217. Lorz, H. W., and B. P. McPherson. 1976. Effects of copper or zinc in fresh water on the adapta- tion to sea water and ATPase activity, and the effects of copper on migratory disposition of coho salmon (Oncorhynchus kisutch). J. Fish. Res. Board Can. 33:2023-2030. Mains, E. M., and J. M. Smith. 1964. The distribution, size, time and current preferences of seaward migrant chinook salmon in the Columbia and Snake Rivers. Wash. Dep. Fish., Res. Pap. 2:5-43. Miller, R. B. 1952. Survival of hatchery-reared cutthroat trout in an Alberta stream. Trans. Am. Fish. Soc. 81:35-42. 1955. The role of competition in the mortality of hatchery- trout. J. Fish. Res. Board Can. 15:27-45. Phinney, D. E., D. M. Miller, and M. L. Dahlberg. 1967. Mass-marking young salmonids with fluorescent pigments. Trans. Am. Fish. Soc. 96:157-162. Raymond, H. L. 1979. Effects of dams and impoundments on migrations of 163 FISHERY BULLETIN: VOL. 82. NO. 1 juvenile chinook salmon and steelhead from the Snake River, 1966 to 1975. Trans. Am. Fish. Soc. 108:505- 529. Reimers, P. E. 197:!. The length of residence of juvenile fall chinook salmon in Sixes River, Oregon. Res. Rep. Fish Comm. Oreg. 4(2):3-43. SCHLUCHTER, M. D., AM) J. A. LlCHATOWICH. 1977. Juvenile life histories of Rogue River spring chinook salmon (Oncorhynchus tshawytscha Walbaum) as deter- mined by scale analysis. Oreg. Dep. Fish Wildl., Inf. Rep. Ser., Fish. 77-5. Sholes, W. h.. and R. J. Hallock. 1979. A evaluation of rearing fall-run chinook salmon, Oncorhynchus tshawytscha, to yearlings at Feather River Hatchery, with a comparison of returns from hatchery and downstream releases. Calif. Fish Game 65:239-255. Solomon, D. J. 1978. Some observations on salmon smolt migration in an chalkstream. J. Fish Biol. 12:571-574. Wagner, H. H. 1974. Photoperiod and temperature regulation of smoltingin steelhead trout {Salmn gairdneri). Can. J. Zool. 42:219- 234. Zaugg, W. S., and H. H. Wagner. 1973. Gill ATPase activity related to parr-smolt transforma- tion and migration in steelhead trout (Salmo gairdneri): Influence of photoperiod and temperature. Comp. Biochem. Phvsiol. 45B:955-965. 164 INTERACTIVE EFFECTS OF AGE AND ENVIRONMENTAL MODIFIERS ON THE PRODUCTION OF DAILY GROWTH INCREMENTS IN OTOLITHS OF PLAINFIN MIDSHIPMAN, PORICHTHYS NOTATUS Steven E. Campana' ABSTRACT Plainfin midshipman, Porichthys notatus, were reared in the laboratory under three environmental regimes to determine the influence of certain variables upon otolith growth increment formation. Both larval and juvenile midshipman were used to test diel cycles and constant conditions of light and temperature. Daily growth increments were formed upon hatch unless a diel photoperiod was absent. However, under constant light, an endogenous circadian rhythm became evident aftera 2-3 week acclimation period, resulting in daily increment production. With increasing age, the influence of light as a zeitgeber decreased, while daily increments became more prominent in all environments. Temperature fluctuation affected increment appearance, but did not entrain increment deposition. Daily growth increments in the otoliths of fishes have been observed in a large number of species (Pannella 1971; Brothers et al. 1976; Taubert and Coble 1977; Wilson and Larkin 1980). These concentrically formed increments may be counted or measured to provide a chronological record of past fish growth. Information on hatching date/age (Ralston 1976; Struhsaker and Uchiyama 1976), daily growth rates (Methot 1981), and timing of life history transitions (Pannella 1980; Brothers and McFarland 1981) has been derived from the examination of otolith micro- structure. Such data are difficult to obtain from larval and juvenile fishes by other means. Daily increments are produced through a diel periodicity in the deposition of calcium carbonate on the otolith (Mugiya et al. 1981). However, there is some controversy as to the zeitgeber behind the daily cycle of deposition, if indeed one exists. In a series of experiments upon larval Lepomis, Taubert and Coble (1977) determined that a 24-h light-dark cycle was necessary to entrain an endogenous rhythm of increment production. Reversal of the light-dark cycle reversed the daily sequence of increment for- mation in larval Tilapia (Tanaka et al. 1981). However, 36-h "days" and constant light conditions had no effect on daily increment production in juvenile starry flounders, Platichthys stellatus (Cam- pana and Neilson 1982). Similarly, constant light or institute of Animal Resource Ecology, University of British < Columbia, Vancouver, British Columbia, Canada V6T 1W5; pre- sent address: Marine Fish Division, Bedford Institute of Ocean- ography, P.O. Box 1006, Dartmouth, Nova Scotia, Canada B2Y 4A2. Manuscript accepted .July 198.'!. FISHERY BULLETIN: VOL. 82. NO. 1. 1984. dark conditions did not inhibit the formation of daily increments in young chinook salmon, Oncorhynchus tshawytscha (Neilson and Geen 1982). The con- tradictory results of the above studies suggest that photoperiod effects on increment production may vary with age or species of fish. Other environmental variables may influence the daily rhythm of otolith deposition. Diel temperature fluctuation has been implicated as a factor in daily increment production of temperate stream fishes (Brothers 1981), although this suggestion has not been supported by other studies (Campana and Neilson 1982; Neilson and Geen 1982). Feeding fre- quency may also influence otolith increment produc- tion; fish given multiple daily feedings have been reported to produce nondaily increments (Pannella 1980; Neilson and Geen 1982), although recent studies suggest that feeding effects are limited (Tanaka etal. 1981 ; Marshall and Parker 1982; Cam- pana 1983). Confidence in the reliability of otolith microstruc- ture examination requires knowledge of those factors that may influence otolith increment production. Conflicting results in the literature suggest that age influences the response of daily increment produc- tion to environmental variables such as photoperiod and temperature. This study was undertaken to test that hypothesis. Plainfin midshipman, Porichthys notatus, were reared from the egg stage under various light and temperature regimes; constant conditions and diel cycles of each variable were tested. The effect of the regimes on otolith microstructure was noted for both newly hatched and juvenile fish. 165 FISHERY BULLETIN: VOL. 82, NO. 1 Juveniles were then subdivided and transferred to different regimes, allowing an examination of the interactive influence of greater age and novel environment on increment production. MATERIALS AND METHODS Fertilized Porichthys eggs were collected inter- tidally from White Rock, British Columbia, on 9 and 22 June 1982. Yolk-sac larvae remain attached to the rock upon which the eggs were originally deposited (Arora 1948), necessitating the collection of both rocks and egg masses. Upon return to the laboratory, eight separate egg masses (50-250 ova each) were isolated in individual saltwater aquaria and main- tained under a diel photoperiod and a temperature of 13°C. Small amounts of methylene blue, strep- tomycin sulphate, and penicillin G were used to con- trol bacterial and fungal infection. Embryo development varied both among and within egg masses, but the difference appeared to be <2-3 d. On 1 July, egg masses were exposed to an experimental environment. Environmental regimes were selected to provide a diel periodicity of either photoperiod or temperature. A third regime main- tained constant conditions of both variables. In this manner, the influence of both factors on increment formation could be determined for newly hatched lar- vae. Daily increment production in the constant environment would suggest the presence of an endogenous circadian rhythm. Regimes were as follows: 14L:10D at a constant temperature of 19°C (14L:10D/CT) 24L with 14 h at 21CC and 10 h at 19 C (24L/ 14TV10T,) 24L at a constant temperature of 19°C (24L/ CT) Duplicate aquaria, each containing an egg mass (or 2 small masses, if at similar developmental stages), were kept in light-proof, temperature-controlled cubicles under each of the above environments. All lighting was fluorescent (30 jiiEs/m2/s). Temperature fluctuations were timer-controlled and conducted parallel to the light cycle. New temperatures were reached VA h after initiation. Mean temperatures approximated those of the egg collection site; diel temperature fluctuations were present at the site, but were not recorded. Aquarium water was changed at 7-10 d intervals. Hatching date varied among and within egg masses, beginning between 7 and 1 1 July. Release from the rock (before completion of yolk-sac resorption) was more variable, and occurred between 23 July and 9 August. Live adult Artemia were first provided as food on 30 July and were consumed by both released and attached larvae. Thereafter, Artemia were maintained in all aquaria at all times, with the exception of two 3-d periods when food was not available. Food abundance did not differ among the aquaria. Observations of feeding fish indicated that the accessibility of Artemia did not limit growth. By 10 August, all fish were about 32-d old (posthatch) and had become juveniles (i.e., had assumed the appearance of an adult). To test the effect of an altered photoperiod or temperature cycle on juveniles, one tank from each of the environmental regimes was subdivided (Fig. 1). About 25 fish were transferred from one aquarium ("cohort") to each of the remaining environments, while leaving 25 fish in the original environment as a control. Sagittae were removed from up to 25 of the excess fish to determine the effect of the original environment on newly hatched larvae. In order to remove any intercohort variability of hatching dates, only one of the two avail- able cohorts from each environment was subdivided and sampled. However, low numbers of 1 4L: 1 0D/CT fish necessitated the transfer of an entire cohort. For processing, the sagittae were brushed free of tissue and glued sulcus-side up with instant glue on a standard microscope slide. Sagittae were ground and polished with metallurgical lapping film (grit size 30 AUG 10 JULY 1- AUG 10 TRANSFER 14L 10D/CT 24L/ 14T, 10T 24L/CT AUG 10- SEPT 10 24L/14T, 10T2 141 10D/CT 24L/CT Fk;i RE 1. — Summary of experimental environmental regimes of plainfin midshipman through time. Fish transferred to new environ- ments on 1(1 August came from the same egg mass as that sampled on 10 August. 166 CAMP ANA: DAILY GROWTH INCREMENTS IN OTOLITHS OF PLAINFIN MIDSHIPMAN jum to 0.3 jixm) until the growth increments in the region of maximal growth were most visible. I defined a growth increment as a bipartite structure, consist- ing of a narrow opaque band and an adjacent broad translucent region. Growth increments between the otolith periphery and the hatch check were counted at least twice through a compound microscope at a magnification of 400X. Duplicate counts of an otolith never differed by more than 10fA. The use of a hand counter eliminated the possibility of a count converg- ing on an expected value. There was little doubt con- cerning the nature of the hatch check; its radius matched that of radii of otoliths removed at hatch. Growth increments in 14L:10D/CT fish sampled 10 September were counted as above. However, a second series of counts was made from the hatch check to the prominent 10 August check; the second data set served as a substitute for the actual sampling of 14L:10D/CT fish on 10 August. Increment counts were made from both the left- and right-hand side sagittae. Since the two sides did not differ systematically under any of the environments (paired /-test, P > 0.05), the means were used in all data analyses. Increment widths were measured from photo- graphs with a micrometer. Expected increment widths were calculated from radial measurements (central nucleus to rostral tip) of otoliths from all environments and a variety of sampling dates (N = 10 per date). Values for mean increase in radial otolith growth per day were then compared to ob- served values. Since individual otoliths often displayed erratic but discernable width trends through time, a measure of the similarity of the widths of two adjacent daily increments was calculated: IR, w,- w;._, (W, + W^/2 where IR, is the index of increment width regularity for day/, and W, is the increment width for day (.Such an index gives low values when adjacent increments are similar in width, despite any trends in the data. Index values were calculated for a range of ages in otoliths from a given environment. RESULTS Porichthys larvae and juveniles survived and grew under all laboratory environments. Survival ex- ceeded 95 '/< after hatch. Fish sampled about 1 mo after hatch (10 August) did not differ significantly in standard length (ANOVA, P > 0.05). By the end of the study, only those fish maintained in the 24L/ 1 4T, : 1 0T2 environment were significantly smaller in length (Scheffe's test P < 0.01); the difference was apparently due to unintentional overcrowding from the date of transfer. Hatching was initiated simultaneously in two of the three initial environments, but started 4 d later in the 24L/CT aquaria. The delay did not appear to be due to the artificial environment, since embryo develop- ment among the 24L/CT egg masses lagged behind that of the others at the time of collection. In the aquarium, about 959? of the viable ova hatched within 4 d of hatch initiation. Intratank hatch-date variance would be expected to affect the variance of increment counts. However, the 17-d range of larval release dates (from the rock) was not reflected in the otolith microstructure. Unground sagittae derived from both pre- and posthatch fish were extremely lobulated in structure. The origin of the numerous lobes was 5-10 "peripheral" nuclei, from which the majority of the growth increments emanated. A central nucleus also had growth increments associated with it, although these were incorporated into the peripheral incre- ments within 10-20 d/increments. A prominent hatch check occurred within 5-10 major increments of the central nucleus. The most prominent check of the older otoliths was that associated with the sub- division/transfer date of 10 August. Many growth increments were visible in the polished otoliths sampled after hatch. When plotted as a function of time, total increment counts were significantly greater than those expected of daily pro- duction (P < 0.05) (Fig. 2). Diel light and tempera- ture cycles both produced an increment: age slope of about 3.0, suggesting that numerous subdaily increments were being counted with any daily increments present. Increment clarity, prominence, and width varied substantially within an otolith. However, most increments could be assigned to one of two "levels" — visually prominent/relatively wide and visually faint/relatively narrow. To determine if the first level consisted primarily of daily increments, the expected width of a daily increment was calculated. 23 July 30 July 9 Aug. 10 Sept. Mean otolith radius (jam): 270 430 620 875 Daily increments on the order of 12-23 and 5-8 fim wide would be expected in the first and second month posthatch, respectively. These expected increment widths were similar to those observed in the first "level" of growth increments. 167 FISHERY BULLETIN: VOL. 82, NO. 1 200-1 • =14L10D/CT a = 24L/14Ti:10T2 40 AGE (DAYS) 50 70 FIGURE 2.— Total otolith increment count as a function of age for plainfin midshipman from two cyclic experimental environments. A straight line has been fitted to the data, although the relationship is probably curvilinear. N = 5 for each data point. Criteria for distinguishing daily from subdaily increments have been reported previously (Taubert and Coble 1977; Campana and Neilson 1982; Marshall and Parker 1982). Nevertheless, no objec- tive criteria have yet been defined which can be applied to all otoliths. In this study, I have used visual prominence and increment width as guides for dif- ferentiating daily and subdaily increments. In- crements assigned as daily were 1) of similar visual prominence (contrast) to adjacent daily increments (±30%), 2) of similar increment width to adjacent daily increments (±50%), 3) not merged with adja- cent daily increments in the nearest radial groove of the sagitta. Some increments met only some of the criteria and were subjectively assigned as daily or subdaily. The observed widths of daily increments, as classified above, were similar to those expected on the basis of otolith growth calculations (see previous paragraph). Diel Light Cycle Otoliths offish reared under a diel photoperiod and constant temperature ( 1 4L: 1 OD/CT) produced clear daily growth increments from the time of hatch. Regression of major increment number against elapsed time produced a slope not significantly different from 1.0 (P> 0.05); a slope of 1.0 would indicate that one increment was formed every day. Increment width varied with location on the otolith and fish age (Fig. 3). Subdaily increments were com- mon at all ages, numbering up to 5 between adjacent daily increments. They were most abundant in the first month after hatch. The distinction between daily and subdaily increments was generally clear; however, increments produced 5-20 d after hatch were the most irregular on the otolith, and were sometimes difficult to interpret. Subdaily incre- ments tended to be prominent in this region, so that distinction was a matter of degree (Fig. 4A). 5 I H Q i 5 cr U z z < 16 12 4- 24L/14T, 10T2 24L/CT 14L 10D/CT i 10 i i 20 30 AGE (DAYS) — r~ 40 - 1 50 Figure 3. -Daily increment width as a function of age for otolith samples of plainfin midshipman from each of the three experimental environments. At a given age, mean widths do not differ significant- ly among environments, with the exception of values at age 40 d (P < 0.05). Fish transferred to a constant environment (24 L/ CT) as juveniles produced posttransfer increments that were very different from those produced prior to transfer. Posttransfer increments were visually faint and, in some cases, virtually invisible (Fig. 5A). Subdaily increments were also present. Transfer to a constant environment was not associated with a recognizable lag period during which increments gradually shifted their appearance. Increments pro- duced within 1-2 d of transfer were virtually nonexis- tent. Nevertheless, posttransfer increments were daily in nature, as indicated by increment counts similar to those expected of daily increment produc- tion (Table 1). Daily increments gradually became more prominent after about 15 d posttransfer, their visual contrast improving until the end of the experiment. 168 CAMP ANA: DAILY GROWTH INCREMENTS IN OTOLITHS OF PLArNFIN MIDSHIPMAN TABLE 1 .—Growth increment counts in otoliths of plainfin midshipman, Porichthys notatus, in rela- tion to elapsed time for various experimental environments. Fish were transferred to new environ- ments (or kept in the original environment as a control) on 10 August, 10 Aug. samples Environ- 10 Sept. samples Fnviron- Days after No. major Days aft€ ir No. major ment 1 hatch increments SE ment 2 hatch increments SE 14L10D/CT 34 '34.3 057 14L10D/CT 65 66.7 0.80 14L10D/CT — — — 24I7CT 65 65.1 1.21 24L/14Tp:10T2 34 41.1 1.29 241VCT 65 71 2 0.70 2417CT 30 49.1 1.33 24L7CT 61 76.9 1.04 2417CT — — — 14L10D/CT 61 72 7 1.10 24UCT — — — 24L'14T,:10T2 61 69.3 0.92 'This value was derived from 14L: 10D/CT otoliths sampled 10 September; counts were made from the hatch check to the prominent subdivision/transfer check Diel Temperature Cycle Fish hatched under a 24L/14T1:10T2 regime deposited growth increments that differed in many respects from those produced under a cyclic photoperiod (14L:10D/CT). Increments produced up to 8 d medial and distal of the peripheral nuclei were characterized by a high incidence of prominent subdaily increments (Fig. 4B), more so than was the case under a cyclic photoperiod. Daily/subdaily similarities are reflected in the data of 10 August (Table 1), where the observed major increment count was significantly different from that expected of daily increments (P < 0.05). The high increment count indicates that some subdaily increments were promi- nent enough to be classified as daily. Increments produced in the 15-20 d before transfer were generally distinct and regular in appearance. Increment width and the incidence of subdaily increments were similar to those observed in the cor- responding region of the cyclic photoperiod otoliths (Fig. 3). However, the appearance of the major increments was unusual in that the opaque portion of each increment was relatively broad and sharply delineated (Fig. 6). Fish maintained in the 24L/14T,:10T2 environ- ment after 10 August were overcrowded and did not grow well. As a result, posttransfer otolith growth was limited, increments were very narrow, and reliable counts could not be made. However, increment counts of representative otoliths suggested that daily increments were deposited after the transfer date. Juvenile fish transferred from the fluctuating tem- perature regime to a constant environment (24L/CT) produced posttransfer increments similar to those of fish transferred from 14L:10D/CT to 24L/CT (Fig. 5B). The difference between August and September increment counts corresponds to that expected of daily increment deposition (P > 0.05) (Table 1). The first five posttransfer increments were faint and vir- tually nonexistent; subsequent increments became more distinct and regular with time. Opaque regions within each increment never became as broad and discrete as was observed prior to transfer. Constant Environment Otoliths of fish hatched under constant conditions (24L/CT) initially resembled those of the other two environments (with respect to the first 5-8 increments). The subsequent region resembled that of 24L/14T,: 10T2 fish in that subdaily increments were prominent (Fig. 4C). Although the difference was not significant (Scheffe's test, P = 0.07), incre- ment widths tended to be more irregular than those of 1 4L: 1 OD/CT fish of similar age (Fig. 7) . The confu- sion of daily and subdaily increments in the early lar- val region resulted in a high variance and a mean increment count that was significantly higher than would be expected of daily increments (P < 0.05) (Table 1). After age 10-25 d, daily increments de- creased in width (Fig. 3) and became more regular in width (Fig. 7) and appearance, although subdaily increments were still present. Increments with broad, discrete opaque portions were not observed in the 24L/CT fish, as they were in the fluctuating tem- perature regime. For an unknown reason, otolith growth (but not fish growth) under a 24L/CT regime significantly exceeded that observed under 14L:10D/CT(P<0.05). Fish remaining in a constant environment after the 10 August transfer date continued to produce dis- tinct increments, although daily and subdaily increments were occasionally difficult to differen- tiate. Increment width was significantly more irregular than in the posttransfer region of 1 4L: 1 0D/ CT fish (<-test, P < 0.05) (Fig. 7). Major increments in the posttransfer region were daily; the regression of increment number against elapsed time resulted in a slope not significantly different from unity (P > 0.05). Posttransfer increments of fish hatched and reared under constant conditions were prominent, although 169 FISHERY BULLETIN: VOL. 82, NO. 1 1 lilt I III I I 111 I I FlGl KK 4. — Growth increments on the polished sagittae of larval plainfin midshipman. Subdaily increments are visible between some of the indicated daily increments. Daily increments became more clear with age, but were most prominent/consistent in width in (A). Bar = 'AO fim. 170 CAMPANA: DAILY GROWTH INCREMENTS IN OTOLITHS OF PLAINFEM MIDSHIPMAN PN = peripheral nucleus. (A) Hatched under a diel light cycle; (B) hatched under a diel temperature cycle; (C) hatched under a constant environment. irregular in width (Fig. 5C). In contrast, increments of fish transferred to the constant environment as juveniles were visually faint, becoming more promi- nent after 2-3 wk. Juveniles transferred from a cons- tant environment to a cyclic regime deposited similar-appearing increments before and after transfer. However, posttransfer increments tended to be more regular in width than in constant environ- ment fish; the change generally became apparent 2-4 d after transfer. Visual contrast of daily increments may have increased in the fluctuating temperature regime, but the change was not consistent among all otoliths. No such change was evident among the post- transfer increments of fish shifted from 24L/CT to 14L:10D/CT, although the incidence of subdaily increments appeared to decrease. Fish transferred from the constant environment to either of the cyclic regimes produced daily increments after transfer; high increment counts (Table 1) were derived from the irregular, pretransfer region of the otolith. DISCUSSION Daily growth increments were deposited on the otoliths of plainfin midshipman under a variety of environmental conditions. My results indicated that light, temperature, age, and an endogenous circadian rhythm may all influence the production and/or appearance of daily and subdaily increments. However, some of the variables tested interacted to a large degree, and their influence on increment pro- duction was subject to alteration through time. A cyclic light regime influenced increment produc- tion in larval fish more than any other variable tested. Under a natural photoperiod, daily increments were produced from the time of hatch. In contrast, con- stant light conditions disrupted the production of posthatch increments, resulting in a high incidence of prominent nondaily increments (> 1 increment/24 h) and irregular increment widths. My observations are consistent with those of Taubert and Coble (1977), who observed numerous, nondaily increments in lar- val Tilapia hatched under constant light conditions. Those authors concluded that light acted as a zeitgeber for an endogenous rhythm and that without a cyclic photoperiod, daily increment production was not possible. My results only partially support their conclusion. Photoperiod entrained daily increment production in newly hatched midshipman. However, in the absence of cyclic light or temperature stimuli, an endogenous circadian rhythm of increment deposition became apparent after an acclimation 171 FISHERY BULLETIN: VOL. 82, NO. 1 \ A 1 * i | , T ' ' ' ■■ B Kiel kk 5. — Growth increments in sagittae of plainfin midshipman produced before and after transfer to a constant environment. Fish hatched under 24L/CT produced clearer daily increments than those transferred from a different 172 CAMPANA: DAILY GROWTH INCREMENTS IN OTOLITHS OF PLAINKIN MIDSHIPMAN '* environment. Daily increments are indicated, as is direction of sagittal growth (arrow). T = transfer check. Bar= 30ju.ni. (A) 14L:10D/CTto24L/CT; (B) 24L/14T,:10T, to 24L/CT; (C) 24L/CT to 24L/CT. period of 2-4 wk. Therefore, photoperiod acted as a zeitgeber for an endogenous rhythm during the early larval stages, but became unnecessary with increas- ing age. The nondaily increments produced after hatch in this study (and that of Taubert and Coble 1977), probably comprised both daily and subdaily increments. The combination resulted in the deposi- tion of more than 1 increment/24 h. If a constant photoperiod was present at hatch, an endogenous rhythm of increment deposition became apparent after an acclimation period. Acclimation also occurred when older fish were transferred from a natural light cycle to constant light conditions. However, the pattern of increment production during acclimation differed at the two ages (Table 2). The larval fish acclimation period may be analogous to that of newborn rats transferred from a diel photoperiod to constant conditions. An arhythmic activity pattern continues for almost 2 wk in rats before an endogenous circadian rhythm becomes apparent (Davis 1981). The length of the acclimation period could not be determined with accuracy. A shift in increment appearance after transfer from a constant to a cyclic environment generally occurred in 2-5 d. The reverse transfer resulted in almost nonexistent increments Table 2. — Age effects on growth increment production in otoliths of plainfin midshipman, Porichthys notatus, reared under three artifi- cial environments. Larvae Juveniles Light important as zeitgeber Daily & subdaily increments similar during acclimation to 24L Long acclimation to 24L Immature circadian rhythm Light not important as zeitgeber. Faint daily increments, but subdaily increments dissimilar during acclima- tion to 24L Short acclimation to 24L Mature circadian rhythm for a period of 5 d, but the visual contrast of the growth patterns improved over the subsequent 10- 15 d. Therefore, the critical stage of the adaptation process appears to have been completed in 2-5 d. This result is consistent with that of Tanaka et al. ( 1981), who observed a 6-d transitory period of incre- ment formation when a 24-h light-dark cycle was sud- denly reversed. Age-related changes in endogenous circadian rhythms have not been examined in fishes. Mam- malian studies indicate that endogenous rhythms often appear after birth; once present, cycle amplitude tends to increase with time until the rhythm is "mature" (Davis 1981). Porichthys larvae hatched under constant light appear to fit this pat- tern. Daily and subdaily increments were not easily 173 ^.HERY BULLETIN: VOL. 82, NO. 1 X Fic.i'RE 6. — Daily growth increments produced on the sagittae of plainfin midshipman after 15-25 d of rearing under a diel temperature cycle. The increments were visually prominent and sharply delineated relative to those produced under other environmental regimes. Bar = 20 /i in. O 5i O 4- cc < _l D a 0 3 DC I t- n 0 2 3 L^ o « LU O I c 2 o 24L/14T, 10T; 24L CT 14L 10D CT 10 20 i 30 40 50 AGE (DAYS FlGi RE 7. Index of daily increment width regularity as a function of age for otolith samples of plainfin midshipman from each of the three experimental environments. Bars represent ±1 SE. differentiated at first, suggesting that the circadian deposition rhythm was not yet mature. Maturation apparently occurred by days 10-20. Early larval increments were only indistinct temporarily in the 14L:10D/CT fish, suggesting that the cyclic photoperiod entrained the maturing rhythm fairly quickly. In addition, very young animals may be more responsive to a diel light cycle, due to age-related characters of the rhythm cycle (Sacher and Duffy 1978). For instance, the metabolic rate of newly hatched rats is very sensitive to changes in light level, while older rats are less affected. In this study, larval fish exposed to a constant environment took longer to produce daily increments than did juvenile fish, sug- gesting an analogy with the rat study. Similar age- related results were reported by Gibson et al. (1978) in an ontogenetic study of flatfish activity cycles. A constant photoperiod eliminated a diel activity cycle in larval plaice (Pleuronectes platessa), but had no such effect on juveniles of the same species. Increasing age of midshipman was correlated with decreasing increment width and fewer subdaily 174 CAMPANA: DAILY GROWTH INCREMENTS IN OTOLITHS OF PLAINFIN MIDSHIPMAN increments in all environments. However, foremost among the age-associated effects (Table 2) was the prominence of daily increments in juveniles relative to larvae. Distinction between daily and subdaily increments was seldom difficult in juveniles (outside of the acclimation period) unlike the situation in lar- val otoliths. If this age-related difference in daily increment formation is universal, daily increment counts in larvae may be unreliable relative to slightly older fish. This suggestion has serious implications for the application of growth increments in aging lar- val fish. Similarly, the absence of definitive criteria for differentiating daily and subdaily increments could cause problems in aging field-collected fish. Subdaily increments can be numerous and confusing in some species (Campana, unpubl. data). The demonstration of and age-related rhythm and the existence of an acclimation period may have resolved some of the conflicting results in the litera- ture concerning the zeitgeber effect of light. In a pre- vious study, a constant light regime did not influence the production of daily increments in juvenile starry flounders (Campana and Neilson 1982). The floun- ders were about 8 mo-old, suggesting that the necessary acclimation period would be short. In addi- tion. the fish were exposed to the experimental environment for 2 wk prior to tetracycline injection (marking the start of the experiment); it is probable that acclimation occurred during this period, resulting in clear daily increment production by the time the experiment began. An analogous explanation may explain the results of another study, where chinook salmon eggs, reared in darkness, produced daily increments after hatch (Neilson and Geen 1982). The embryos were held in total darkness for 50 d before hatch, suggesting that their endogenous circadian rhythm had time to acclimate before hatch. A fluctuating temperature regime did not entrain increment production under constant light con- ditions. Fish reared in this environment produced more increments than would be expected of daily production, similar to those of 24L/CT fish. The variance of larval increment counts was similar to that produced under a constant environment, both of which were significantly larger than the 1 4L: 1 OD/CT variance (Bartlett's test,P< 0.01). Once acclimation occurred, daily increments were produced through an apparently endogenous periodicity, and not through temperature entrainment of an internal clock. However, the formation of a broad, optically dense, sharply delineated opaque zone in postac- climation daily increments indicates that tempera- ture fluctuation did affect increment production. The opaque portion of a daily increment consists of calcium carbonate and a proteinaceous matrix, with the latter component predominating (Brothers 1981;Mugiyaetal. 1981). Falling temperatures, such as would occur at night, may have increased the pro- portion of protein deposited in the opaque region, resulting in an increment that had increased visual contrast. Accentuation of contrast renders in- crements visually prominent, and could easily be interpreted as an entraining mechanism. Diel tem- perature fluctuations noticeably accentuated incre- ment contrast in young chinook salmon otoliths ( J. D. Neilson2). A correlation of increasing protein deposi- tion with decreasing temperature suggests that the broad opaque zone formed during the low tempera- ture, 1 0-h, experimental "night", overlaid the opaque zone formed under circadian control through a 3-h period (Mugiya et al. 1981). If temperature does exert a "masking" effect (Enright 1981), a low temperature-induced opaque zone would appear independently of any endogenous circadian rhythm of deposition. Therefore, multiple daily oscillations in temperature could conceivably produce a distinct increment after each cycle, in addition to the daily increment formed under endogenous control. In some situations, the masking effect of temperature fluctuations may be substantial, obscuring most of the increments formed through the action of an endogenous rhythm of deposition (E. B. Brothers3). This hypothesis is consistent with studies that demonstrated that temperature cycles do not entrain daily increment production (Campana and Neilson 1982; Neilson and Geen 1982), but can influence increment formation (Brothers 1981). My results suggest that a diel light cycle entrains an endogenous circadian rhythm of increment deposi- tion. Increasing age mitigated the zeitgeber effect of photoperiod, while temperature fluctuation influ- enced increment appearance, rather than perio- dicity. In other studies, the incidence of subdaily increments was correlated with feeding periodicity (Neilson and Geen 1982; Campana 1983). The fact that so many variables may affect increment deposi- tion suggests that the environment does not influence the rhythm of otolith deposition directly, but acts through some penultimate process. Metabolic rate is susceptible to environmental influence, as well as being subject to an endogenous circadian rhythm (Matty 1978) that changes with age (Davis 1981). However, metabolic rate is in turn -J. D. Neilson, Marine Fish Division. Biological Station, St. Andrews, New Brunswick, Canada EOG 2X0, pers. comraun. Jan- uary 1983. 'K. H. Brothers, Division of Biological Sciences, Cornell Univer- sity, Ithaca, XV 1 1850, pers. comraun. May 198 3 175 FISHERY BULLETIN: VOL. 82. NO. 1 regulated by endocrine levels, and it may be the environmental modulation ofendocrine rhythms that ultimately controls increment periodicity on the otolith (Menaker and Binkley 1981). Endocrine se- cretion often follows a circadian pattern (Simpson 1978) and, in mammals at least, is closely linked to the circadian pacemaker itself (Menaker and Binkley 1981). Hormones regulate many aspects of meta- bolism and growth, including skeletal calcification (Simpson 1978). Therefore, it seems reasonable to postulate that those factors that entrain and/or mod- erate the circadian rhythm of endocrine secretion will have a subsequent effect on increment deposition in the otolith. ACKNOWLEDGMENTS Jim McNutt of Ayerst Laboratories kindly donated the antibiotics used in this study. I appreciate the assistance and technical innovation of Doug Begle in the field and laboratory. John D. Neilson and Nor- man J. Wilimovsky provided many helpful comments on an earlier draft of the manuscript. This study was supported by a grant from the Natural Sciences and Engineering Research Council of Canada to Norman J. Wilimovsky. LITERATURE CITED Arora, H. L 1948. Observations on the habits and early life history of the batrachoid fish Porichthys notatus Girard. Copeia 1948:89-93. Brothers, E. B. 1981. What can otolith microstructure tell us about daily and subdaily events in the early life history of fish? Rapp. P.- V. Reun. Cons. Int. Explor. Mer 178:393-394. Brothers, E. B., C. P. Mathews, and R. Lasker. 1976. Daily growth increments in otoliths from larval and adult fishes. Fish. Bull., U.S. 74:1-8. Brothers, E. B., and W. N. McFarland. 1981. Correlations between otolith microstructure, growth, and life history transitions in newly recruited French grunts \Haemulon flavoUneatum (Desmarest), Haemulidae]. Rapp. P-V Reun. Cons. Int. Explor. Mer 178:369-374. Campana, S. E. 1983. Feeding periodicity and the production of daily growth increments in otoliths of steelhead trout (Salmo gairdneri) and starry flounder (Platichthys stellatus). Can. J. Zool. 61:1591-1597. Campana, S. E., and J. D. Neilson. 1982. Daily growth increments in otoliths of starry flounder {Platichthys stellatus) and the influence of some environ- mental variables in their production. Can. J. Fish. Aquat. Sci. 39:937-942. Davis, F. C. 1981. Ontogeny of circadian rhythms. In J. Aschoff (editor). Handbook of behavioral neurobiology, Vol. 4, p. 257- 274. Plenum Press, N.Y. ENRIGHT, J. T. 1981. Methodology. In J. Aschoff (editorl. Handbook of be- havioral neurobiology. Vol. 4, p. 11-19. Plenum Press, N.Y. Gibson, R. N., J. H. S. Blaxter, and S. J. de Groot. 1978. Developmental changes in the activity rhythms of the plaice (Pleuronectes platessa L.). In J. E. Thorpe (editor). Rhythmic activity of fishes, p. 169-186. Acad. Press, N.Y. Marshall, S. L., and S. S. Parker. 1982. Pattern identification in the microstructure of sockeye salmon {Oncorhynchus nerka) otoliths. Can. J. Fish. Aquat. Sci. 39:542-547. Matty, A. J. 1978. Pineal and some pituitary hormone rhythms in fish. In J. E. Thorpe (editor), Rhythmic activity of fishes, p. 21- 30. Acad. Press, N.Y. Menaker, M., and S. Binkley. 1981. Neural and endocrine control of circadian rhythms in the vertebrates. In J. Aschoff (editor). Handbook of behavioral neurobiology, Vol. 4, p. 243-256. Plenum Press, N.Y. Methot, R. D„ Jr. 1981. Spatial covariation of daily growth rates of larval northern anchovy, Engraulis mordax, and northern lampfish, Stenobrachius leucopsarus. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 178:424-431. Mugiya, Y., N. Watabe, J. Yamada, J. M. Dean, D. G. Dunkelberger, and M. Shimuzu. 1981. Diurnal rhythm in otolith formation in the gold fish, Carassius auratus. Comp. Biochem. Physiol. 68A:659- 662. Neilson, J. D., and G. H. Geen. 1982. Otoliths of chinook salmon (Oncorhynchus tsha- wytscha): daily growth increments and factors influencing their production. Can. J. Fish. Aquat. Sci. 39:1340-1347. PANNELLA, G. 1971. Fish otoliths: daily growth layers and periodical pat- terns. Science (Wash., D.C.) 173:1124-1127. 1980. Growth patterns in fish sagittae. In D. C. Rhoads and R. A. Lutz (editors). Skeletal growth of aquatic organisms: Biological records of environmental change, p. 519- 560. Plenum Press, N.Y. Ralston, S. 1976. Age determination of a tropical reef butterflyfish utiliz- ing daily growth rings of otoliths. Fish. Bull., U.S. 74:990- 994. Sacher, G. A., and P. H. Duffy. 1978. Age changes in rhythms of energy metabolism, activity and body temperature in Mus and Peromyscus. In H. V. Samis, Jr., and S. Capobianco (editors). Aging and biologi- cal rhythms, p. 105-124. Plenum Press, N.Y. Simpson, T. H. 1978. An interpretation of some endocrine rhythms in fish. //; J. E. Thorpe (editor). Rhythmic activity of fishes, p. 55-68. Acad. Press, N.Y. Struhsaker, P., and J. H. Uchiyama. 1976. Age and growth of the nehu, Stolephorus purpureus (Pisces: Engraulidae), from the Hawaiian islands as indicated by daily growth increments of sagittae. Fish. Bull., U.S. 74:9-17. Tanaka, K., Y. Mugiya, and J. Yamada. 198 1. Effects of photoperiod and feeding on daily growth pat- terns in otoliths of juvenile Tilapia nilotica. Fish. Bull., U.S. 79:459-466. Taubert, B. D., and D. W. Coble. 1977. Daily rings in otoliths of three species of Lepomis and 176 CAMPANA: DAILY GROWTH INCREMENTS IN OTOLITHS OF PLAINFIN MIDSHIPMAN Tilapia mossambica. J. Fish. Res. Board Can. 34:332- 340. Wilson, K. H., and P. A. Larkin. 1980. Daily growth increments in the otoliths of juvenile sock- eye salmon [Oncorhynchus nerka). Can. J. Fish. Aquat. Sci. 37:1495-1498. 177 ASPECTS OF THE LIFE HISTORY AND FISHERY OF THE WHITE CROAKER, GENYONEMUS LINEATUS (SCI AENIDAE), OFF CALIFORNIA Milton S. Love,1 Gerald E. McGowen,2 William Westphal ' Robert J. Lavenberg,2 and Linda Martin' ABSTRACT White croaker. Genyonemus lineatus (Ayres), was a dominant species off southern California in nearshore, sandy substratum waters, and comprised 29.7% of all fish taxa taken in otter trawl hauls. Juveniles occurred in waters < 27 m and the mean length of all individuals increased with depth. The maximum depth of capture was 183 m. White croaker live to 1 2 years, exhibiting rapid growth which is essentially constant throughout the species' life. Females grew at a slightly faster rate than males. Von Bertalanffy age-length parameters for females wereL„ = 60.7,fc = 0.04,(0=-7.6,andformalesLoc = 59.2,& = 0.03,ro=-8.7.Afterlyear,morethan50% of the individuals are mature, but others delay maturity for 4 years. Larger females had longer spawning seasons than did smaller individuals. Although spawning occurred throughout the year, principal spawning occurred between November and April, with a February-March peak. White croaker are batch spawners; females spawned 18-24 times a season. Batch fecundities ranged from 800 to 37,200 eggs. White croaker reproduction off Monterey differed significantly from that off southern California. Large-scale spawning occurred from at least July through February, and continued throughout the year. Colder water off Monterey may have allowed for extended spawning activity. White croaker larvae were a significant constituent of the southern California ichthyoplankton fauna, second in abundance to northern anchovy, Engraulis mordax, in waters <36 m deep. Data from ichthyoplankton surveys indicated two spawning centers, one located from Redondo Beach to Laguna Beach and a smaller one centered about Ventura. Highest larval densities were found near the substratum in 15-22 m of water. White croaker is an important part of the skiff sportfishery and the basis of a growing commercial gill net fishery. Size frequencies of white croaker taken in both fisheries indicated that few juveniles were captured. Fishes of the family Sciaenidae (drums) are a major constituent of the fauna of the eastern temperate Pacific coast off California (Skogsberg 1939; Frey 1971). Eight species have been recorded off Califor- nia, primarily in inshore waters. With the exception of the shortfin corvma,Cynoscionparvipinnis, and black croaker, Cheilotrema saturnum, all six of the other species known from off California (white seabass, Atractoscion nobilis; white croaker, Genyonemus lineatus; California corbina, Mentieirrhus undulatus; spotfin croaker, Roncador stearnsii; queenfish, Seriphus politus; yellowfin croaker, Umbrina ron- cador) are of sport or commercial importance. The white croaker is an abundant species that associates with soft (primarily sand) substrata in the coastal zone. White croaker are small (reaching 'Vantuna Research Group, Department of Biology, Occidental College, Los Angeles, CA 90041. 'Natural History Museum of Los Angeles County, 900 Exposition Blvd., Los Angeles, CA 90007. 'Moss Landing Marine Laboratory, P.O. Box 223, Moss Landing, CA 95039. lengths of 41.4 cm total length, Miller and Lea 1972) and active fishes that range from the surf zone to depths of 183 m between Vancouver Island, British Columbia, Canada, south to Magdalena Bay, Baja California, Mexico. Within this geographic range, they are most abundant between San Francisco Bay and northern Baja California. White croaker are omnivores, feeding on a variety of benthic and epibenthic forms (crustaceans, clams, polychaetes, and small fishes, particularly the northern anchovy, Engraulis mordax (Phillips et al. 1972; Morejohn et al. 1978; Ware 1979)). White croaker are the mainstay of pier and small boat sportfish catches in both southern (Pinkas et al. 1968; Wine and Hoban 1976) and central California (Miller and Gotshall 1965). In addition, commercial catches have increased in recent years to 200,000 kg/ yr.4 Despite this, G. lineatus is a much maligned species, as it is small and adept at bait-stealing. More- Manuscript accepted June 1983. FISHERY BULLETIN: VOL. 82, NO. 1, 1984. 4M. Oliphant, California Department of Fish and Game, Long Beach, CA 90802, pers commun. July 1981. 179 FISHERY BULLETIN: VOL. 82, NO. 1 over, there is a firmly held belief that white croaker are unusually wormy. In fact, the frequency of occurrence of nematodes (larval Anisakis and Phocanema) in white croaker muscle is lower than that for at least some other important sport and com- mercial species such as California halibut, Paralichthys californicus, and chilipepper rockfish, Sebastes goodei (Dailey et al. 1981). Because white croaker are abundant around sewage outfalls and tolerant of degraded environments, much of the recent research on this species has been pollution-centered. Several published works deal with pesticide levels (Castle and Woods 1972; MacGregor 1972; Stout and Beezhold 1981) and pollution-implicated diseases and abnormalities (Russell and Kotin 1957; Mearns 1974, 1979; Mearns and Sherwood 1977; Sherwood 1978). Five small-scale studies have been conducted on its life history (Issacson 1964, 1967; Goldberg 1976; More- john et al. 1978; Ware 1979) This contribution represents a summation of unpublished white croaker data obtained from three sources: a life history and fishery study by Love, ichthyoplankton work by McGowen and Lavenberg, and a trawling survey by Westphal. METHODS Collection of Juveniles and Adults Samples were collected monthly (3-6 per month) from October 1978 to February 1981 with a 7.6 m or 4.9 m headrope otter trawl in 15-65 m of water be- tween Palos Verdes and Huntington Beach, Calif. Reduced numbers of white croaker also were collect- ed monthly from April 1979 to September 1981 in Monterey Bay with a 4.9 m otter trawl in 10-60 m of water or were purchased from local fishermen. All of these specimens were frozen for later dissection. After thawing, all fish were measured (total length, fork length, standard length), weighed, sexed, and the gonads were weighed. Collection of Depth Preference Data for Adults and Juveniles Information on white croaker depth preference was based on data from a trawling program aboard the RV Vantuna . Trawling was conducted at a speed of 2- 3 kn for 20 min with a 7.6 m (occasionally 4.9 m) otter trawl having a net of 0.6 cm stretch mesh. From Sep- tember 1972 through December 1980, 18 stations (Fig. 1) were sporadically sampled at 10 depths, although most of the trawling effort was performed at depths between 59 and 91 m. After shipboard sort- ing, fishes were measured (board standard length) and discarded. All lengths were converted to total length (TL) using conversion factors based on measurements of 100 white croaker (Table 1). TABLE 1. — Conversion factors between standard (SL), fork (FL), and total (TL) lengths (cm), based on measurements of 100 white croaker from southern California. SL = 0.442 + 0 79 TL = 0.379+ 0 82 FL FL = 0.088 + 0 96TL 0.849+ 1 14 SL TL=0892 + 1.19 SL = 0.023+ 1.04 FL Techniques for Aging Juveniles and Adults Sagitta were removed from each side of the head, and the otoliths were cleaned, air dried, and stored in vials. Because whole croaker otoliths are difficult to age, they were sectioned on a Buehler Isomer5 low speed saw, Otoliths were placed on wood blocks and completely embedded in clear epoxy (Ciba 825 hard- ener and Ciba 6010 resin). Each block with its otolith was emplaced on the saw and a dorsal-ventral 0.05 cm wafer was cut through the otolith, using two diamond-edge blades separated by a stainless steel shim. Wafers were stored in water for a few days to soften the epoxy (which was removed), then the wafers were placed in a black-bottomed water glass filled with water and read under a dissecting micro- scope at a magnification of 10X. All otoliths were read twice, about 4 mo apart, by Love. When readings did not agree, the otoliths were read again. The value of two coincident readings was accepted as the best estimate of age. Fifteen percent of all otoliths were unreadable due to a lack of recogniz- able annuli. Procedures for Determining the Timing of Maturation and Reproduction We estimated length at first maturity by classifying gonads as immature or mature based on the tech- niques of Bagenal and Braum (1971). Smallermature fish and fish just entering their first mature season become reproductive later in the spawning season. Hence we estimated length at first maturity during the peak spawning period of January, February, and March. To ascertain spawning season duration and its relation to body size, we sampled at least 150 females/mo in 1 cm size intervals throughout the 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 180 LOVE ET AL.: LIFE HISTORY AND FISHERY OF WHITE CROAKER Fliti'RE 1.— Location of white croaker sampling sites. ^C^>> 100km year. A gonadosomatic index (gonad weight)/(total body weight) X 100 was computed from frozen specimens to quantify changes in gonad size with season. Ovaries for use in fecundity studies were fixed in modified Gilson's fluid (Bagenal and Braum 1971) for 4-8 mo. We measured fixed egg diameters from 11 individuals, all of which contained some hy- drated eggs. Batch fecundity was estimated by the gravimetric method of Bagenal and Braum (1971). The time between spawning events per female was computed by estimating the percent of females with hydrated eggs on any given night during the spawning season. We computed condition factor 1UU (W u W), L3 where W = body weight in grams, GW — gonad weight in grams, and L = total length in centime- ters— of mature southern California and Monterey croaker. Condition factor was computed using body weight with gonad weight subtracted to mini- mize the effects of seasonal changes in gonad size. Larval Sampling Ichthyoplankton data presented here were collect- ed monthly between August 1979 and July 1980 along 20 sites within the Southern California Bight aboard the RV Seawatch (Table 2). Stations were established at 8 and 22 m along each transect (with exception of Palos Verdes and Laguna Beach where 15 m was substituted for 8 m). Additional stations at 15 and 36 m depths were maintained at three sites (Ormond Beach, Redondo Beach, San Onofre). Oblique bongo tows from the bottom to the surface were made at all stations. A 70 cm diameter bongo net sampler (McGowan and Brown 1966), equipped with wheels to prevent damage when the sampler encountered the bottom, was lowered to the bottom with canvas 181 FISHERY BULLETIN: VOL. 82, NO. 1 TABLE 2.— Southern California ichthyoplankton collection sites, August 1979-July 1980. Location abbreviations used in Figures 13-15 are in parentheses. Collection sites Lat. N Long. W Coho Bay (80) 34 26' 120c26' Refugio to El Capitan. 8 m (DR) 34°27' 120' 02'- 120°05' North of Refugio. 22 m 34°27' 1 20r 06' Santa Barbara to Goleta Pt. (8. 1 5) 34°25' 119 44'- 119°51' Pt. Gorda to Rincon Pt |RN) 3422'- 34 23' 119°28' Ventura (83) 34° 1 6' 119°17' Ormond Beach (OB) 34°07' 1 1 9 1 0' Arroyo Sequit (85) 34°03' 118°57' Mahbu Beach (MU| 34°02' 118°41' Playa del Rey (87) 33°57' 118°27' Redondo Beach (RB) Redondo Breakwater, 8, 1 5, and 22 m 33°51' 118°24' Hermosa Pier. 36 m 33°52' 118°25' Palos Verdes (PV) 33°43' 118°25' Huntington Harbor (88) 33°4T 11 8° 04' Balboa (BA) 33° 36' 1 1 7°54' Aliso Creek (Laguna Beach) (90) 33°31' 1 1 7°46' San Onofre (SO) 33°21' 1 1 7°33' Santa Margarita River (91 ) 33° 15' 1 1 7°28' Agua Hedionda (Carlsbad) (CD) 33°08' 117 23' San Dieguito River (Del Mar) (93) 32° 58' 117 16' Mission Beach (MB) 32°48' 117"16' San Diego (95) 32°38' 1 1 7°09' doors over the mouth openings. The canvas doors were removed by a cable messenger, allowing the nets to fish. Immediately thereafter the sampler was retrieved at a constant rate of about 10 m/min (0.17 m/s); a wire angle of 51 ± 5° was maintained. The ship's speed (0.95 ± 0.03 m/s) plus the retrieval rate brought the net speed to about 1.12 m/s. In addition, stratified (surface, midwater, bottom) tows were made at each of the four stations on tran- sects at Ormond Beach, Redondo Beach, and San Onofre. Horizontal midwater tows were made with the previously described bongo sampler towed at a rate of 1.06 ± 0.06 m/s. For these tows the sampler was lowered to a depth about half-way between the surface and the bottom, opened via cable messenger, fished, closed via cable messenger, and retrieved. Surface samples were taken with a manta sampler (Brown 1979) towed at a rate of 1 .07 ± 0.06 m/s. This net had a rectangular opening (88 X 16 cm). Bottom collections were taken using an auriga net6 with a 200 X 50 cm mouth. The auriga net fished a zone 2 mwide by 0.5 m deep, about 0.25 m above the substratum, and was fished at a rate of 1.07 ± 0.46 m/s. All nets were equipped with 335 ju mesh. A General Oceanics flowmeter was mounted in the mouth of each net. The field program is described in greater detail by Laven- berg and McGowen.7 Additional data from a 4-yr study off Redondo Beach were derived from monthly surface tows made from January 1974 to February 1977, using meter nets with 335 /i mesh. A TSK flowmeter was mount- ed in the mouth of each net. This field program is de- scribed in greater detail by McGowen.8 Fishery Although white croaker are usually the most impor- tant species in the private vessel sportfishery, no size-frequency data were available. For this reason, 4,941 croaker taken by anglers aboard skiffs and other small private vessels were measured during the period June 1980 to July 1981, between Oxnard and Dana Point. From September 1980 through August 1981, 1,748 white croaker were taken off southern California by commercial gill net vessels and were measured. RESULTS Depth Preference Our trawling study indicated that white croaker pre- ferred nearshore habitats and their abundance declined in deeper waters. Ranking first of all species taken, white croaker was the dominant species at the shallowest (18-27 m) stations (Table 3), and com- posed 29.7% by number of the total catch and appeared in 68% of the trawls. At the 59-73 m stations, white croaker catches had declined to 3.3% of total catch, frequency of occurrence 20.7%, and at the 91-109 m station, the species made up 1.2% of total catch, frequency of occurrence 14.0%. At stations between 165 and 183 m, white croaker com- prised 0.6% of the total catch, with a frequency of occurrence of 1.7%. On the basis that no individuals were captured at greater depths, we accept 183 m as their maximum depth. Though white croaker was supplanted as the domi- nant species at deeper stations, it remained an important community component to depths of 109 m. Two other species, the California tonguefish, Symphurus atricauda, and the Pacific sanddab, Citharichthys sordidus, were among the 10 most abundant species throughout these depths. Pacific ''Mitchell, C. T. Auriga: A wheeled epibenthic plankton sampler for rocky bottoms. Unpubl. rep., 12 p. Marine Biological Con- sultants Inc., 947 Newhall Street, Costa Mesa, CA 92627. 'Lavenberg, R. •!.. and (1. E. McGowen. Coastal ichthyoplankton of the Southern California Right: temporal and spatial distribution (Augusl 1979-July 1980). Manuscr. in prep. Los Angeles County Museum of Natural History, 900 Exposition Blvd., Los Angeles, CA 9 7. 8McGowen, G. E. 1978. Effects of thermal effluent from Southern California Edison's Redondo Beach steam generating plant on the warm temperate fish fauna of King Harbor Marina. SCE Research and Development Series: 78-RD-47, 65 P. 182 LOVE ET AL.: LIFE HISTORY AND FISHERY OF WHITE CROAKER Table 3. — The 1 0 most abundant fish species taken by otter trawls in three depth intervals off Southern California, 1972-80. Total no % total % frequency taken no. occurence Depth interval. 18-27 m Number of collections. 109 Total no. of fish, 14.313 Total Species. 80 Genyonemus /meatus 4.252 297 679 Cithanchthys stigmaeus 2.221 15 5 63.3 Symphurus atncauda 2.031 14.2 60.6 Senphus politus 1.341 94 44.0 Phanerodon furcatus 595 4.2 59.6 Engraulis mordax 591 4.1 22.0 Pleuromchthys verticalis 476 33 62.4 Hyperprosopon argenteurn 395 2.8 33.0 Cithanchthys sordidus 301 2.1 12.8 Synodus lucioceps 206 1.4 38.5 Depth interval. 59-73 m Number of collections. 82 Total no. of fish, 1 3,337 Total species. 62 Cithanchthys sordidus 3.196 24.0 72.0 Microstomus pacificus 2,769 20.8 65.9 Sebastes dalh 1.565 11.3 65.9 Sebastes saxicola 867 6.5 29.3 Ponchthys notatus 786 5.9 59.8 Sebastes jordani 694 52 17.1 Symphurus atncauda 512 3.8 51.2 Scorpaena guttata 506 38 63.4 Genyonemus /meatus 436 3.3 207 Icelmus quadnseriatus 297 2.2 25.6 Depth interval. 91-109 m Number of collections, 1 72 Total no. of fish, 35,488 Total species, 77 Microstomus pacificus 12.386 34.9 762 Cithanchthys sordidus 9.655 27.2 73.8 Sebastes saxicola 4,262 12.0 65.1 Ponchthys notatus 1,688 4.8 63.4 Glyptocephalus zachirus 1,249 3.5 30.2 Scorpaena guttata 875 2.5 44.2 Sebastes jordani 802 2.3 21.5 Genyonemus tmeatus 441 1.2 14.0 Symphurus atncauda 377 1.1 24.4 Zaniotepis frenata 299 0.8 250 where L, k sanddab dominated in waters between 59 and 109 m, declining in numbers both inshore and offshore. California tonguefish exhibited an abundance pat- tern like white croaker, with numbers peaking in inshore waters and declining with greater depth. Most juvenile white croaker (50% mature by 15 cm) were limited to the inshore (18-27 m) stations (Fig. 2). Larger individuals inhabited greater depths. In fact, the mean size of white croaker was successively larger as depth increased (ANOVA, F = 284.2, P < 0.001). Age and Growth Lengths at ages were estimated by direct observa- tion of otolith annuli and through the von Bertalanffy growth curve model Lt = Lx [1 - exp -k {t-t0)\ = length at time t = theoretical maximum length = constant expressing the rate of ap- proach to Lx = theoretical age at which L, = 0 to the direct observation age-length was fitted data. We transformed male and female growth equations to linear form (Allen 1976) and compared these by analysis of variance. Females were found to grow significantly faster than males (F = 16.8, P < 0.05), hence we separated growth data by sex (Table 4). TABLE 4.— Parameters of the von Bertalanffy equation for white croaker off southern Calfornia. Sex Lm SE * SE to SE Female Male 60.72 59.17 0.23 0.29 0037 0033 0.02 0.03 -7.54 -866 1.1 1.3 The oldest male and female white croaker we examined were 12 yr old (Fig. 3). Females grew slightly faster than males and reached a greater size. Females from age 1 (at which over 50% of the fish were mature) outgrew males. White croaker grew at a fairly constant rate throughout their lives, exemplified in their very low/? values. No asymptote was reached within the observed 12-yr life span. Thus, the maximum predicted lengths were longer than both published (41.4 cm TL, Miller and Lea 1972) and unpublished (44.2 cm9) records, although the r values for the von Bertalanffy equations were high (0.84 for both sexes). Length - Weight Relationships A total of 58 1 males and 665 females from southern California and a total of 94 males and 161 females from Monterey Bay were weighed and measured. The relationships between total length and weight fit the relationship W = aLh, where W = weight in grams, L = total length in centimeters, and a and b are constants, with values determined using log10 transformation and fitting the values to a straight line by least squares (Figs. 4, 5). In southern California, males tended to be heavier at a given length than females (analysis of variance, F — 10.18, P < 0.01), whereas off Monterey no significant difference was found (analysis of variance, F = 0.67, P > 0.4). To test whether this difference was an artifact caused by seasonal and gender-related factors, we subtracted 'R. N. Lea, California Department of Fish and Game, 2201 Garden Road, Monterey, CA 93940, pers. commun. May 1982. 183 50 - 50 - UJ < I- I £2 Li. LL O en CO D z 50 300 200 100 SAMPLES=1 N=60 X=26.1 DEPTH 165-183 m SAMPLES=19 N=308 X=23.9 DEPTH 91-109 m SAMPLES=13 N=286 X=17.3 DEPTH 59-73 m SAMPLES=69 N= 3,764 X=16.2 DEPTH 18-27 m 6-6.9 8-8.9 FISHERY BULLETIN: VOL. 82, NO. 1 10- 12- 14- 16- 18- 20 22- 24 26 28- 30- 10.9 12.9 14.9 16.9 18.9 20.9 22.9 24.9 26.9 28.9 3 30.0 32- 32.9 TOTAL LENGTH INTERVALS (cm) FIGURE 2.— Length intervals of white croaker taken by otter trawl off southern California. gonad weight from body weight, generated the length-weight relationships for each sex and tested these between sexes. Again, differences between sexes existed in southern California (ANOVA, F = 1 1.13, P < 0.01), but not in Monterey Bay (ANOVA, F= 1.33, P> 0.05). Condition Factor Both male and female southern California white croaker displayed differences in condition between peak spawning and resting seasons (Table 5). In both sexes, fish were more robust during the resting season, perhaps because energy normally utilized for somatic maintenance and growth was shifted to egg and sperm production and spawning behavior. Over all seasons, whereas southern California females were more robust than males (Table 5), no such sex- TaBLE 5. — Condition factor (K) of white croaker from southern California 1978-81 and Monterey Bay. Calif., 1979-81. N K SD F P Southern California Males Jan -Mar 264 0.34 0.53 1 17.4 <0.001 May-Aug. 91 0.98 0.32 Females Jan-Mar 280 0.46 0.56 24,4 <0.001 May-Aug 76 0.80 0.49 Sexes Combined Jan-Mar 544 0.40 0.55 118.3 <0.001 May-Aug. 167 0.90 0 41 All Seasons Males 535 0.71 056 4 5 <0.05 Females 617 078 0.54 Monterey — All seasons Males 80 1 03 0.09 1 29 >0.2 Females 142 1 02 0.10 Southern California and Monterey Males Monterey 80 1 03 0.09 26.54 <0.001 S Calif 535 0.71 0.56 Females Monterey 142 1.02 0.10 27.83 <0.001 S. Calif. 617 0.78 0.54 184 LOVE ET AL.: LIFE HISTORY AND FISHERY OF WHITE CROAKER 30 25 5 U a H O 20 15 - 10 FEMALES MALES 10 11 12 AGE (YEARS) FIGURE 3. — Von Bertalanffy growth curves of female and male white croaker. Also included are mean lengths at ages (females-circles, males-triangles) computed from direct observation of otolith annuli. Based on 332 females and 250 males taken off southern California, 1977-81. ual dimorphism was observed off Monterey. Both males and females off Monterey were more robust than their southern California counterparts (Table 5). Maturation and Reproduction Although a few white croaker matured before 1 yr (12.9-13.4 cm TL), over 50% of the males were mature by 14 cm TL and over 50% of the females by about 15 cm TL, which equals an age of 1 yr (Fig. 6). All fish were mature by 19 cm TL (3-4 yr). Larger females (greater than about 1 7 cm TL and 1 - 2+ yr) spawned earlier in the year and continued to spawn later than smaller and younger individuals (Table 6). The smallest spawning females may spawn for 3-4 mo whereas larger individuals may spawn for as long as 7 mo. Off Long Beach, white croaker spawned primarily from November through April, with January through March the peak months, based on the occurrence of hydrated eggs within ovaries. A few individuals (> 18 cm TL) spawned from May through October. Ovaries increased in size in the fall and peaked in January, when they averaged 4.67c of body weight (maximum 11.8%, minimum 0.8%). Thereafter, ovarian size declined in summer to a minimum of about 1.0% of body weight (maximum 1.3%, minimum 0.07%) and remained constant through August (Fig. 7). Similarly, testes were small during summer months 185 FISHERY BULLETIN: VOL. 82, NO. 1 350 H 300 250 200 o I- I (J UJ 150 100 FEMALES W = .0109 L30239 R = .9836 • • 2 A •2 4/ 3 • 335»**2 22 •2234/222. 3.33*534 • ... 5>24*«3 ♦ •••3J794.82.2. • 4>5276«'2« 2' 2059424 •3 5WS25525 2. 4 6<2962«2« 345; >Jf>272 2 • 29 70< ) 352. 2 .5^5533* 50 22.&ZB82.2 • 24 5/TC • • ■ 4*^2»« 222 _L _L _L 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 TOTAL LENGTH (cm) FIGURE 4. — Length-weight relationship based on 665 female white croaker sampled off southern California, 1978-81. TABLE 6. —The percent per month of female white croaker from southern California (1978-81) that were sexually mature. Mean total length (cm) Percen sexually mature Sept. Oct. Nov. Dec Jan. Feb. Mar. Apr. May June July Aug 13.0 0 0 0 2 16 15 6 0 0 0 0 0 140 0 0 0 1 1 26 26 8 0 0 0 0 0 15 0 0 0 0 21 73 72 15 0 0 0 0 0 16 0 0 0 0 18 88 88 27 0 0 0 0 0 17.0 0 0 2 20 91 90 35 tr1 1 0 0 0 18.0 0 0 6 21 96 94 61 tr tr 0 0 0 19.0 0 0 7 21 100 1 ( 10 83 48 tr 0 0 0 20.0 0 tr 7 23 100 100 82 52 tr 0 tr tr 21 0 0 0 5 31 100 100 94 51 2 tr 0 0 22.0 tr 0 6 32 99 99 93 58 1 0 0 0 23.0 0 0 7 48 100 100 95 60 tr 0 0 tr 24.0 tr 0 6 47 100 100 93 58 2 0 0 0 25.0 0 0 6 47 100 100 99 57 2 0 0 0 26.0 0 tr 6 46 100 100 98 59 1 0 tr 0 'Trace <1%. 186 LOVE ET AL.: LIFE HISTORY AND FISHERY OF WHITE CROAKER 300 - 250 200 O \- I o LLI 150 100 50 MALES W = .0111l3-0114 R = .9750 2 / 2» •2 »2 > 2 y/ r • * • 4 5^ • • . «2>«2.. 2 23. ..2 36» 3.22 3, . ».*<.32. 232/222. .. 4%*.24 ••

*32 «5623« 252* S43» J_ _1_ _L _L _L _L J_ X 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 TOTAL LENGTH (cm) FIGURE 5.— Length-weight relationship based on 581 male white croaker sampled off southern California, 1978-81. (Fig. 8), averaging 0.39c of body weight (maximum 0.99c, minimum 0.057c), and increased in the fall to a January peak averaging 2.69c of body weight (max- imum 7.7%, minimum 0.4%). In contrast, white croaker off Monterey Bay spawned over a longer period and may have winter and summer spawning peaks. Ovarian weights were highest in January and September (averaging about 6.5 and 7.0% of body weight, respectively) and lowest in May (1.3% of body weight). Ovaries never shrank to the minimum sizes typical of individuals in the southern California population during summer months. Testes grew to a much larger maximum size (4.6% vs. 2.6%) off Monterey. Northern white croaker spawned nearly every month, though spring spawning might have been limited. In limited sam- pling during the following year,10 the second (January) peak was not evident, and thus may not be an annual event. Batch fecundities ranged from an estimated 800 eggs in a 15.5 cm female to about 37,200 in a 26 cm female (Fig. 9). During the spawning period about 19% of all mature female white croaker sampled con- tained hydrated eggs, implying that a female spawned about once every 5 d. Females of ages 1 and 2 (13-18 cm) have a spawning season of 3 mo (Table 6) and spawn about 18 times per season, whereas older fish ( 1 9 cm and larger) spawn over a period of 4 mo, about 24 times/season. 10T. Keating, Moss Landing Marine Laboratory, P.O. Box 223, Moss Landing, CA 95039, pers. commun. January 1982. 187 14 15 16 17 18 TOTAL LENGTH (cm) FIGURE 6.— Length-maturity relationship in 995 female and 941 male white croaker collected off southern California, 1978-81. FISHERY BULLETIN: VOL. 82, NO. 1 Larvae Data from our ichthyoplankton surveys showed that white croaker spawning occurs every month of the year (Fig. 10). However, a distinct seasonal spawning period can be deduced from findings that few larvae were collected from June through November, whereas high densities were encountered from January through April with a strong peak in March. Results of our study in King Harbor, Redondo Beach (Fig. 11), confirm the peak densities of white croaker larvae in January, February, and March. White croaker larvae constitute an important com- ponent of the neritie ichthyoplankton fauna of the Southern California Bight, ranking second in overall abundance behind the northern anchovy, Engraulis mordax. On a per transect basis (Fig. 12), white croaker larvae ranked first in abundance at all tran- sects between Palos Verdes1 ' and Laguna Beach and nGenyonemus and Engraulis were virtually tied for first place at Redondo Beach with 40.1% and 40.39;, respectively. MONTEREY SOUTHERN CALIFORNIA 1 I i 1 1 1 _L JAN FEB MAR APR MAY JUNE JULY AUG SEPT OCT NOV DEC Figure 7. — Seasonal changes in the gonosomatic index (GSI-gonad weight as a percentage of total body weight) of female white croaker (based on 720 southern California and 223 Monterey individuals). Vertical lines indicate 95',' confidence intervals of the mean. 188 LOVE ET AL.: LIFE HISTORY AND FISHERY OF WHITE CROAKER 5 - MONTEREY < 5 4 - 3 - 2 - SOUTHERN CALIFORNIA JAN FEB MAR APR MAY JUNE JULY AUG SEPT OCT NOV DEC FIGURE 8. — Seasonal changes in the gonosomatic index of male white croaker (based on 631 southern California and 1 14 Monterey individuals). Vertical lines indicate 95'* confidence intervals of the mean. 40 «" 36 - b o 2 32 z D 24 1 20 u 2 16 1? a 4 F = .000093 L R = 0.80 6.08 14 15 16 17 19 20 21 £?. TOTAL LENGTH 25 26 Figure 9.— Batch fecundity— total length relationship for 44 white croaker collected off southern California during February and March 1979-81. second at the remaining transects, except Mission Beach where it ranked third behind Engraulis and an unidentified goby. In the King Harbor study, white croaker larvae ranked either fourth or fifth depend- ing on the year and the stations sampled. Larval density data (number of individuals per unit volume of water) indicate two spawning centers be- tween Point Conception and the U.S. -Mexican bor- der (Fig. 13): The larger one extends north and south of the Palos Verdes Peninsula, from Redondo Beach 7000 6000 o o o 5000 Q- UJ < > DC < O a. 4000 3000 3 Z 2000 1000 — AUG SEP OCT NOV DEC JAN FEB MAR APR MAYJUN JUL 1979 1980 FIGURE 10.— Mean density of white croaker larvae collected in the oblique bongo tows per month between August 1979 and July 1980. to Laguna Beach, whereas the smaller one is further north around Ventura. That area from San Onofre south to the international border was striking for its low densities of white croaker larvae. Along this sec- 189 FISHERY BULLETIN: VOL. 82, NO. 1 M E o O 550 - 500 - 450 - 400- 350- HI a. 300 HI < > < 250 - O jjj 200- D 150 - 100 - 50 - 1974 MIJIJIAIS 1977 FIGURE 1 1 . — Mean densities of white croaker larvae in the vicinity of King Harbor, Redondo Beach, Calif., between January 1974 and February 1978. tion of the coast, white croaker larvae accounted for 11.7% of the larval fishes collected in oblique tows versus 43.6% from Laguna Beach to Redondo Beach and 17.9% from Playa del Rey to Point Conception (Fig. 14). Our data indicate that highest densities of white croaker larvae occur near the bottom (Fig. 15). In the coastal zone, between the 15 and 36 m isobaths, rela- tive densities indicate little variation through the water column, being 1.5-3.5% with surface waters, 55.0-58.0% in the bottom waters, and 40.0-42.5% in middepth waters (Fig. 16). Relative densities in the surface waters at the shallow 8 m stations dramatically increased to 17.5% with a correspond- ing decrease in both bottom and middepth waters. White croaker larval densities peaked at stations located at 1 5 and 22m depths (Fig. 15). The densities declined sharply at the deeper (36 m) and shallower stations (8 m). The only exception to this trend was in surface water where densities steadily decreased in an offshore direction. Only 1 5 of our 20 transects had stations at 8 and 22 m isobaths. Data in Figure 15 suggest that an abun- dance estimate based on the 8 and 22 m stations may approximate one based on the 15 and 36 m stations. If so, an estimate based on either of those station pairs should approximate one based on all four. We examined this at the three transects (OB, RB, SO), where data for all four stations were available. We tested the data from each transect for each of the 12 mo of the sampling program using the sign test (Dixon and Massey 1957). The estimated number of white croaker larvae per 1 ,000 m3 based on the 8 and 22 m stations was compared with the estimate based 190 LOVE ET AL.: LIFE HISTORY AND FISHERY OF WHITE CROAKER on the 8, 15, 22, and 36 m stations; no statistically significant difference was found (N = 26; P > 0.05). The similarities of the overall estimates based on these two station groupings are shown in Figure 13. On the basis of our 8 and 22 m stations we have extrapolated density estimates to 36 m. Estimates were made for the truncated Palos Verdes and Laguna Beach transects as well, which are likely to be upwardly biased as they are based on two high den- sity stations (15 and 22 m). Data in Figure 13 show that these two transects are not high density ones; in fact, Palos Verdes is low for that section of the coast. The Laguna Beach transect is lower than the next two transects to the north. We included the Laguna Beach transect in the portion of the Southern Califor- nia Bight where white croaker larvae are in high abun- dance, on the basis that the density would still be higher than the portion of the coast from San Onofre to San Diego, based on just the 8 and 22 m stations, even if the 8 m station contributed no larvae. We estimated, from oblique bongo tows taken at the 8 and 22 m stations (15 and 22 m stations at Palos Verdes and Laguna Beach), the average density of white croaker larvae between August 1979 and July 1980 to have been 740/1,000 m-\ 2,203/1,000 m\ and 4 1 1/1,000 m3 for the regions between Point Con- ception and Playa del Rey, Redondo Beach and Laguna Beach, and San Onofre and the international border, respectively. On the basis that there is no significant difference between estimates based on the 8 and 22 m stations and one based on the 8, 15, 22, and 36 m stations, we use the 8 and 22 m density estimates to project the average number of white croaker larvae to the 36 m isobath. It has been estimated (Lavenberg and McGowen footnote 7) that about 3 1 km3 of water are located in a band along the coast between Point Conception and the U.S. -Mexican international border and extend- ing seaward to the 36 m isobath. Of this, 15.6 km3 (50.6%) is located in the region between Point Con- ception and Playa del Rey, 7.9 km3 (25.9%) between Redondo Beach and Laguna Beach, and 7.2 km3 ■j-i \- u LU w 2 < 80 DR 81.5 RN 83 OB 85 MU 87 RB PV 88 BA 90 SO 91 CD 93 MB 95 SANTA BARBARA 3 2 RANKINGS SAN DIEGO FIGURE 12. — Rank abundance of white croaker larvae collected in oblique bongo tows taken along 20 transects in the Southern Califor- nia Bight between August 1979 and July 1980. See Table 2 for sta- tion abbreviation definitions. DR 81.5 RN 88 BA 90 SO 91 93 MB 95 Based on 2 Stations Based on 4 Stations r T SAN DfEGO 500 1000 1500 2000 2500 3000 3500 NUMBER OF LARVAE PER 1000 m3 FIGURE 13. — Mean densities of white croaker larvae along 20 tran- sects in the Southern California Bight between August 1979 and July 1980. See Table 2 for station abbreviation definitions. 191 FISHERY BULLETIN: VOL. 82, NO. 1 BO DR 81 5 KN B3 OB h- u I/) z < MU 87 RB PV 88 BA 90 SO 91 93 MB 95 10 — I- 20 — 1 1 — 30 40 PERCENT SAM DtEGO < > < 3 3300 - 3000 2700 2400 2100 1800 1500 1200 900 600 300 - • = MANTA □ = MID-DEPTH ♦ = BENTHIC A= OBLIQUE "I 1 I- 08 15 22 BOTTOM DEPTH (m) 36 60 FIGURE 15.— Mean density of white croaker larvae collected with each of four different tow types along four isobaths — 8, 15, 22, and 36 m— in the Southern California Bight between August 1979 and July 1980. FIGURE 14. — The percentage contributed by white croaker to the total number of larvae collected along each of 20 transects in the Southern California Bight between August 1979 and July 1980. See Table 2 for station abbreviation definitions. (23.57c) between San Onofre and the international border. Based on these values plus the density estimates, we project the average number of white croaker larvae in each of the three areas during this period to have been 1.15 X 1010, 1.75 X 1010, and 2.97 X 109, respectively. Thus, about 55% of the white croaker spawned in the area between Redondo Beach and Laguna Beach, 36% between Playa del Rey and Point Conception, and about 9% between San Onofre and the border. Fishery Most of the white croaker retained by sportfisher- men were adults (Fig. 17), being 21-25 cm and 5-7 yr Figure 16. — Mean percentage of white croaker larvae collected near the surface, near the bottom and in between along each of four isobaths — 8, 15, 22, and 36 m — in the Southern California Bight be- tween August 1979 and July 1980. 6IJ 50 - O 4(1 30 M 10 | NEUSTON | | MID-DEPTH I BENTHIC 4 08 15 22 BOTTOM DEPTH (m) q 36 192 LOVE ET AL.: LIFE HISTORY AND FISHERY OF WHITE CROAKER 900 800 700 600 O 500 CC 111 co E Z 400 300 200 - 100 100% !=□_ 10 12 14 16 18 20 22 24 TOTAL LENGTH (cm) M 28 30 32 FIGURE 17. — Lengths of white croaker retained by skiff sportfishermen off southern California, 1980-81, with length at IOO'a maturity noted. old. Small fish were only occasionally hooked, and rarely retained. Within the Southern California Bight, about 10 vessels fished white croaker full time. Two areas, Long Beach south to Dana Point and Oxnard to Santa Barbara, were fished most heavily, which cor- responded to the sites of peak white croaker larvae concentrations reported here. This is a gill net fishery, and an informal agreement among fishermen sets the net mesh at 7.0 cm (2.75 in) stretch. Nets are 1.3 km (0.8 mi) long and are set on the bottom in depths of 5.5-37 m (3-20 fathoms). Mean catches of white croaker are 270-400 kg (600-900 lb) per set with maximum catches of 680-770 kg (1,500- 1,700 lb). Largest catches occurred in January and February, during spawning season, when white croaker aggregated in large numbers. The prices for 1982 to fishermen were 13-18«/kg (30-40*/lb). Most fish taken during our study were 26-29 cm long (Fig. 18) and 8-10 yr old. We found no immature fish. DISCUSSION Depth Preference Though most species of Sciaenidae prefer inshore waters, white croaker are distributed over a wider depth range than other northeastern Pacific species. Queenfish was the fourth most abundant species taken in our survey at the shallowest station (Table 3); its abundance declined rapidly with depth. Though it was present in deeper water, it contributed <0.1% of the fishes taken at 59-73 m. The white seabass is common within the 30 m contour (though they are taken as deep as 90 m during winter months). Umbrina roncador, Roncador stearnsi, and Men- ticirrhus undulatus prefer sandy beaches and bays to 193 FISHERY BULLETIN: VOL. 82, NO. 1 500 400 < I 300 m 3 Z 200 100 FIGURE 18.— Lengths of white croaker retained by com- mercial gill net fishermen off southern California, 1980- 81. 100% 18 20 22 24 26 28 TOTAL LENGTH (cml 30 32 34 depths of perhaps 9 m (Skogsberg 1939), whereas Cheilotrema saturnum are common over reefs to perhaps 15 m (occasionally to 45 m) (Limbaugh 1961). Most eastern Pacific drums are limited to the warmer waters south of Point Conception (Miller and Lea 1972) or, like the queenfish and white seabass, are rare north of the Point. Conversely, white croaker are abundant north to San Francisco. Temperature pre- ference experiments12 indicate that juvenile white croaker have wide metabolic thermal optima (11°- 17 C, based on routine oxygen consumption rates) that may account for their wide depth and latitudinal ranges. Though white croaker are most abundant over sandy, featureless substrata, they are occasionally found in large numbers in kelp beds. This is particularly the case in beds anchored on sand, such as those off San Onofre and Santa Barbara. Similarly, though they spend most of their time near the bottom, we have noted schools in midwater, 20-40 m or more above the substrata. We have also seen white croaker at the surface, chasing anchovy schools. Maturation and Reproduction We computed the length-maturity relationship using standard length to compare our results with 1!Hose, J. E., and W.H.Hunt. 1981. Physiological responses of juvenile marine fish to temperature. Occidental College Annual Report submitted to Southern California Edison, 17 p. those of Issacson (1967). We found 50% of the males mature by 12.0 cm SL and 50% of females by 13.0 cm SL, both at 1 yr. This was in sharp contrast to Issac- son's statement that "The white croaker matures between 147 and 164 mm standard length at an age of 3 to 4 years." Why such a disparity should exist is unclear. White croaker is the only southern California drum that spawns in the winter. Winter spawning is unusual even among tropically derived temperate species off California. All species in the families Blenniidae, Carangidae, Labridae, Pomacentridae, Scombridae, and Sphyraenidae are either summer spawners or spring and summer spawners with a summer spawning peak. An exception are the rock- fishes (Scorpaenidae), the vast majority of which spawn in winter and/or spring. The more or less continuous (or perhaps dual- peaked) spawning season seen in white croaker in Monterey Bay is an interesting phenomenon. Most California marine fishes have restricted spawning seasons. If spawning does continue for extended periods (as in the bocaccio, Sebastes paucispinis), there is usually only one peak spawning period. An exception is the northern anchovy, Engraulis mor- dax, that may spawn year-round and which exhibits a major peak in late winter-early spring and a minor one in early fall. Fishes of the northeastern Pacific tend to have a longer spawning season in the southern part of their range, as favorable conditions are usually more re- 194 LOVE ET AL.: LIFE HISTORY AND FISHERY OF WHITE CROAKER stricted in northern waters (Westrheim 1975). However, on examination, the water temperatures in Monterey Bay more closely approximate optimal white croaker spawning conditions than those off southern California. The peak spawning periods, based on gonosomatic indices and ichthyoplankton surveys, in southern California occur between January and March, when mean surface tem- peratures decrease to 13°-14°C (U.S. Department of Commerce 1956). Off Monterey, the mean tem- peratures of the warmest months are 13°-14°C (June-October), whereas the other months are 1°- 3°C cooler. Thus white croaker encounter tem- peratures conducive to spawning for more months off Monterey than off southern California. White croaker reproductive behavior is in some re- spects the opposite of the cooccuring queenfish. White croaker spawn almost entirely during late winter and early spring (peak February-March), but our ichthyoplankton survey gives a March-April peak, whereas queenfish are spring and summer spawners (peak April-May, DeMartini and Fountain 1981). Most egg hydration in white croaker takes place dur- ing the night, with spawning occurring from just before dawn to midmorning. Queenfish spawn be- tween late afternoon and evening. We have not ascer- tained the extent that habitat partitioning has played in this separation. Off Monterey, where queenfish are rare, white croaker spawn virtually year round. As discussed before, this is perhaps a reflection of a more favorable temperature regime. It would be instructive to know if in the absence of queenfish, egg hydration and spawning time are simlar to those off southern California. Larvae Data from both gonosomatic indices and ichthyoplankton surveys show white croaker spawn year-round in southern California waters. However, peak spawning clearly is in the winter and spring. Our data, combined with Watson's (1982), indicate that peak densities of white croaker larvae were in either January, February, or March from 1974 through 1980. This is out of phase with other southern California sciaenids, all of which spawn primarily in the spring and summer (Lavenberg and McGowen footnote 7). White croaker larvae are an important component of the southern California neritic ichthyoplankton fauna. Along the three sections of the Southern California Bight, defined and studied during this investigation, white croaker larvae contributed 11.7, 43.6, and 17.9% of the total larvae from south to north. Highest densities were found at stations located in 15-22 m depths (Fig. 15). The decreasing densities, as one moves shoreward of the 15 m isobath, apparently continues into the enclosed bays and estuaries of southern California. McGowen (1981) did not collect any white croaker larvae in south San Diego Bay during a 13-mo study. Larval white croaker ranked sixth, contributing 0.6% of the larvae collected in Newport Bay during an 18-mo study by White (1977). The percentage reported by White may have a bias toward lower values because the period of peak spawning was sampled only once during the 18 mo. However, even a doubling of White's percentages does not make white croaker larvae dominant members of the Newport Bay ich- thyoplankton assemblage. Leithiser (1981) reported white croaker to contribute 1.9% of the total catch of larval fishes in Anaheim Bay during a 12-mo study. King Harbor is typical of the estuarine-enclosed bay habitat rather than that of the open coast and is dominated by blennies, clinids, gobies, and engraulids (McGowen footnote 8). White croaker lar- vae ranked either fourth or fifth in the King Harbor study, depending on the year and the stations sampled. Densities of white croaker larvae also decreased between the 22 and 36 m isobaths (Fig. 15). This indication that white croaker larvae are not common in offshore waters is supported by CalCOFI data. The highest any sciaenid ranked in these collections between 1955 and 1958 was 18th, contributing 0.30% of the total larval catch (Ahlstrom 1965). This pattern of white croaker larvae being dis- tributed in a narrow band along the coast, between the 15 and 22 m isobaths, is similar to the pattern reported by Watson (1982) and Barnett et al.13 off San Onofre. They designated white croaker larvae as having an inner nearshore epibenthic pattern. Bar- nett et al. (footnote 13) indicated highest densities on the bottom, shoreward of the 22 m isobath, and the second highest densities in the water column be- tween the 12 and 22 m isobaths and on the bottom between the 22 and 45 m isobaths. The major dis- crepancy between their data and ours is the higher epibenthic densities that they report shoreward of the 12 m isobath and seaward of the 22 m isobath. This discrepancy may be partially explained by dif- "Barnett, A. M., A. E. Jahn, P. E. Sertic, and W. Wat- son. 1980. Long term spatial patterns of ichthyoplankton off San Onofre and their relationship to the position of the SONGS cooling system. A study submitted to the Marine Review Committee of the California Coastal Commission, July 22, 1980, Unpubl. rep., 32 p. Marine Ecological Consultants of Southern California, 533 Stevens Ave., Suite D-57, Solana Beach, CA 92075. 195 FISHERY BULLETIN: VOL. 82, NO. 1 ferences in sampling strategy. They sampled within blocks defined by depth contours whereas we sampled at specific isobaths. Thus, part of their block D (be- tween the 22 and 45 m isobaths) is located at a depth where we found high densities (22 m) and part of it where we found low densities (36 m). All of their block B (between 9 and 12 m) is located at depths where we did not sample. Their block A (between 6 and 9 m) is located in a zone where our data suggest lower densities. Our trawling data also support this narrow band as important for the young stages of white croaker. Almost all of the juvenile white croaker taken during our study were collected at stations located between the 18 and 27 m isobaths (Fig. 2). In summary, these data suggest that adult white croaker migrate shoreward (larger adults were taken at deeper depths; Fig. 2) and spawn in a narrow band along the coast. This band has its shoreward boun- dary located between the 8 and 12 m isobaths, and its seaward boundary located between the 22 and 36 m isobaths. Furthermore, the pelagic stages remain pri- marily within this band. At the end of the pelagic phase young white croaker move into 3-6 m and take up residence near the bottom. As these juvenile fish mature, they migrate to deeper waters (Fig. 2). Based on this hypothesis, we believe that a realistic evaluation of the spawning activities of the white croaker can be based on data collected from the shore to the 36 m isobath. We have done this and found that about 9% of the spawning by white croaker occurred along the coast from San Onofre to the international border, about 55% from Laguna Beach to Redondo Beach, and around 36% from Playa del Rey to Point Conception. If this represents the typi- cal annual pattern, the portion of the Southern California Bight from Laguna Beach to at least Point Conception is important for white croaker, especially the region around the Palos Verdes Peninsula from Redondo Beach to Laguna Beach. However, that portion of the bight from San Onofre to the border is relatively insignificant. The only remaining coastal zone in the U.S. portion of the Southern California Bight is around the Channel Islands. We have not investigated the coastal zones of these islands and cannot appraise their significance to the spawning activities of white croaker in the Southern California Bight. Fishery Historically, the commercial white croaker fishery has been minor, rarely exceeding 1 million lb/yr (Frey 1971). Most fish were caught and landed in the Long Beach-San Pedro region and Monterey Bay. Southern California accounted for about two-thirds of the catch and Monterey one-third, although during World War II, Monterey produced over one-half the total catch. Until recently, white croaker were taken commercially by otter trawl, round haul net, mul- tifilament gill net, and hook and line. However, in the past few years, significant changes have occurred in the fishery. Gill nets, particularly monofilament nets, have almost entirely supplanted other methods. The ubiquity of white croaker along the southern California mainland makes this species accessible to small boat sportfishermen. The ease with which it may be taken, using minimum skill or equipment, ensures that this species will be caught in consider- able numbers. We commonly found two fishermen with at least 50 or more white croaker after a half day's effort. Though traditionally scorned by many, we found that the species is popular with a number of ethnic groups. The Monterey fishery has been revived in the past 2-3 yr by newly arrived Vietnamese fishermen.14 White croaker are fished throughout Monterey Bay, over the entire year, in 12-24 m (40-80 ft), occasionally to 37 m (120 ft) with 1.6-2.4 km (1-1.5 mi) long monofilament gill nets [6.3 cm (2.5 in) stretch mesh]. Nets are tended daily, and 450-900 kg (1,000-2,000 lb) catches are common with maximum catches to 1,800 kg (4,000 lb). Depending on catch size and fish condition, payment to fishermen ranges from 6 to 22C/kg (15 to 50C/lb). These white croaker are sold principally within central California (par- ticularly the San Francisco area), although a small amount is shipped to southern California. Demand is increasing, particularly among various Asian com- munities.15 SUMMARY In this study, white croaker was the most abundant species in nearshore (18-27 m) otter trawl collections in southern California. This species dwelled prin- cipally in shallow water and juveniles were restricted to the shallower (<27 m) parts of the species depth range. Living to 12 yr, white croaker grew at a nearly l4D.J. Miller, California Department of Fish and Game, 2201 Gar- den Road, Monterey, CA 93940, and T. Keating, Moss Landing Marine Laboratory, P.O. Box 233, Moss Landing, CA 95039, pers. commun. August 1981. "Though most white croaker are retailed fresh, there is reason to believe that a potential market exists for them as surimi (fish cakes). A fish cake plant existed in Ventura during 1979, processing 3,000- 4,000 lb (1,360-1,800 kg) of white croaker per day. All cakes were sold to the Asian community in Los Angeles. Demand for the pro- duct was very strong and the plant closed for reasons unrelated to profitability. 196 LOVE ET AL.: LIFE HISTORY AND FISHERY OF WHITE CROAKER constant rate throughout the species' life. A majority of both males and females matured at about 1 yr and all were mature by 4 yr. We noted a difference in spawning season between southern and central California. Off southern California, significant spawning occurred between November and April, while central California individuals spawned all year, with large-scale activity occurring from July through February. Our ichthyoplankton survey indicated that two spawning centers occurred off southern California — one located from Redondo Beach to Long Beach and the other centered about Ventura. White croaker larvae, which were second in abun- dance to northern anchovy in nearshore waters, were found in greatest abundance near the substratum in 15-22 m of water. The abundance of white croaker and its ease of capture make it a major sportfish in the skiff fishery and a growing component of the com- mercial gill net fishery. Our study indicates that the vast majority of fishes taken in both fisheries were adults. ACKNOWLEDGMENTS We thank J. Stephens for his continual support of our work. J. Palmer, T. Sciarrotta, and J. Stock of the Southern California Edison Company and G. Brewer of the University of Southern California assisted in project design and logistical support. L. McCluskey helped estimate batch fecundities, and L. Natanson and E. Taylor conducted the small vessel creel census. T. Keating supplied numerous Monterey specimens, and J. Balesteri supplied data on the southern California commercial operation. Majority of the larval identifications were made by D. Carlson, D. Chandler, D. Eto, R. Feeney, S. Good- man, N. Singleton, D. Winkler, and R. Woodsum of the University of Southern California and the Natural History Museum of Los Angeles County. E. Gray and L. Games of the Southern California Edison Com- pany and the Natural History Museum of Los Angeles County, respectively, assisted with data reduction and computer programming. We also thank M. Butler (illustrations) and R. Meier (photography) of the Los Angeles County Natural History Museum. Lastly, we thank the many people who assisted in the sorting and collecting of samples, especially the crews of RV Vantuna and RV Seawatch. LITERATURE CITED AHLSTROM, E. H. 1965. Kinds and abundances of fishes in the California Current region based on egg and larval surveys. Calif. Coop Oceanic Fish. Invest. Rep. 10:31-52. Allen, R. L. 1976. Method for comparing fish growth curves. N.Z.J. Mar. Freshw. Res. 10:687-692. BAGENAL, T. B.. AND E. Braum. 1971. Eggs and early life history. In W. E. Ricker (editor), Methods for assessment offish production in fresh waters, p. 166-198. (Int. Biol. Programme) Handb. 3., Blackwell Sci. Publ., Oxf. and Edinb. Brown, D. M 1979. The manta net: quantitative neuston sampler. Inst. Mar. Res. Mar. Tech. Handb. IMR TR 64, 15 p. Castle, W. T., and L. A. Woods, Jr. 1972. DDT residues in white croakers. Calif. Fish Game 58:198-203. Dailey, M. D., L. A. Jensen, and B. W. Hill. 1981. Larval anisakine roundworms of marine fishes from southern and central California, with comments on public health significance. Calif. Fish Game 67:240-245. DeMartini, E. E., and R. K. Fountain. 1981. Ovarian cycling frequency and batch fecundity in the queenfish, Seriphus politus: attributes representative of serial spawning fishes. Fish. Bull., U.S. 79:547-560. Dixon, W. J., and F. J. Massey, Jr. 1957. Introduction to statistical analysis. 2d ed. McGraw- Hill Book Co., Inc., N.Y., 488 p. FREY, H. W. 1971. California's living marine resources and their utilization. Calif. Dep. Fish Game, 148 p. Goldberg, S. R. 1976. Seasonal spawning cycles of the sciaenid fishes Genyonemus lineatus and Seriphus politus. Fish. Bull., U.S. 74:983-984. ISSACSON, P. A. 1964. Length-weight relationship of the white croaker. Trans. Am. Fish. Soc. 93:302-303. 1967. Notes on the biology of the white croaker, Genyonemus lineatus (Ayres). Trans. Kentucky Acad. Sci. 28:73-76. Leithiser, R. M. 1981. Distribution and seasonal abundance of larval fishes in a pristine southern California salt marsh. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 178:174-175. LlMBAUGH, C. 1961. Life-history and ecologic notes on the black croaker. Calif. Fish Game 47:163-174. MacGregor, J. S. 1972. Pesticide research at the fishery-oceanography cen- ter. Calif. Coop. Oceanic Fish. Invest. 16:103-106. McGowan, J. A., and D. M. Brown. 1966. A new opening-closing paired zooplankton net. SIO Ref. 66-23, 56 p. McGowen, G. E. 1981. Composition, distribution, and seasonality of ichthyoplankton populations near an electricity generat- ing station in south San Diego Bay, California. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 178:112-114. MEARNS, A. J. 1974. Southern California's inshore demersal fishes: diver- sity, distribution, and disease as reponses to environmen- tal quality. CalCOFI (Calif. Coop. Oceanic Fish. Invest.) Rep. 17:140-148. 1979. Responses of coastal fishes and invertebrates to waste- water discharges. Prog. Water Technol. 4:19-32. Mearns, A. J., and M. J. Sherwood. 1977. Distribution of neoplasms and other diseases in marine 197 FISHERY BULLETIN: VOL. 82, NO. 1 fishes relative to the discharge of waste water. Ann. N.Y. Acad. Sci. 298:210-224. Miller, D. J., and D. Gotshall. 1965. Ocean sportfish catch and effort from Oregon to Point Arguello, California. Calif. Dep. Fish Game, Fish Bull. 130, 135 p. Miller, D. J., and R. N. Lea. 1972. Guide to the coastal marine fishes of California. Calif. Dep. Fish Game, Fish Bull. 157, 235 p. Morejohn, G. V., J. T. Harvey, and L. T. Krasnow. 1978. The importance oiLoligo opalescens in the food web of marine vertebrates in Monterey Bay, California. In C. W. Recksiek and H. W. Frey (editors). Biological, ocean- ographic, and acoustic aspects of the market squid, Loligo opalescens Berry, p. 67-98. Calif. Dep. Fish Game, Fish Bull. 196. Phillips, L. C, C. Terry, and J. S. Stephens. 1972. Status of the white croaker (Genyonemus lineatus) in the San Pedro Bay region. South Calif. Coastal Water Proj. Tech. Rep. 109, 49 p. Pinkas, L., M. S. Oliphant, and C. W. Haugen. 1968. Southern California marine sportfishing survey: private boats, 1964; shoreline, 1965-66. Calif. Dep. Fish Game, Fish Bull. 143, 42 p. Russell, F., and P. Kotin. 1957. Squamous papilloma in the white croaker. Nat. Can. Inst. J. 18:857-861. Sherwood, M. J. 1978. The fin erosion syndrome. South. Calif. Coastal Water Res. Proj., Annu. Rep., p 203-221. Skogsberg, T. 1939. The fishes of the family Sciaenidae (croakers) of California. Calif. Dep. Fish Game, Fish Bull. 54, 62 p. Stout, V. F., and F. L. Beezhold. 1981. Chlorinated hydrocarbon levels in fishes and shell- fishes of the northeastern Pacific Ocean, including the Hawaiian Islands. Mar. Fish. Rev. 43(1):1-12. U.S. Department of Commerce. 1956. Surface and water temperatures at tide stations, Pacific coast, North and South America and Pacific Ocean islands. U.S. Coast Geod. Surv. Spec. Publ. 280, 74 p. Ware, R. R. 1979. The food habits of the white croaker Genyonemus lineatus and an infaunal analysis near areas of waste dis- charge in outer Los Angeles Harbor. M.A. Thesis, California State Univ., Long Beach, 113 p. Watson, W. 1982. Development of eggs and larvae of the white croaker, Genyonemus lineatus Ayres (Pisces: Sciaenidae), off the southern California coast. Fish. Bull., U.S. 80:403-417. Westerheim, S. J. 1975. Reproduction, maturation, and identification of larvae of some Sebastes (Scorpaenidae) species in the northeast Pacific Ocean. J. Fish. Res. Board Can. 32:2399-2411. White, W. S. 1977. Taxonomic composition, abundance, distribution and seasonality of fish eggs and larvae in Newport Bay, California. M.A. Thesis, California State Univ., Fuller- ton, 107 p. Wine, V., and T. Hoban. 1976. Southern California independent sportfishing survey annual report, July 1, 1975-June 30, 1976. Calif. Dep. Fish Game, 109 p. 198 FEEDING HABITS OF BLACKSMITH, CHROMIS PUNCTIPINNIS, ASSOCIATED WITH A THERMAL OUTFALL Pamela A. Morris1 ABSTRACT The availability and use of food by blacksmith, Chromispunctipinnis, were examined at a thermal outfall and a control site in King Harbor, California. Stomach analysis showed that blacksmith from the outfall area con- sumed a significantly greater amount of food, consist ing of larger prey items, than control fish. Movements of water created by the outflow may provide dietary benefits by reducing zooplankton predator avoidance and by entraining and entrapping organisms not normally planktonic. This dietary enrichment may result in attraction of blacksmith to the King Harbor outfall. An increased demand for energy resulting in growth of coastal power plant activity has created concern for the effects of heated effluents upon the fish com- munity (Miller 1977; Stephens 1978,2 19803; Stephens and Palmer 19794). Few studies have examined the factors attracting fish to outfall areas. White et al. (1977) found less diversity and lower abundance of fish at an outfall station, while Kelso (1976) and Minns et al. (1978) reported a clustering offish in the vicinity of thermal outfalls. Underwater observations suggest that fish are attracted to ther- mal outfalls to feed. Kelso (1976) found that fish in proximity to a thermal discharge exhibited a complex swimming behavior that could represent feeding activity. Moreover, this behavior continued when unheated effluent was discharged. The blacksmith, Chromis punctipinnis (family Pomacentridae), an abundant planktivorous tem- perate reef inhabitant, has been regularly observed feeding at the thermal outfall of a steam electrical generating station in King Harbor, Redondo Beach, Calif. Recent studies on the effects of thermal effluents upon blacksmith have concentrated on behavioral 'VANTUNA Research Group, Department of Biology, Occidental College, Los Angeles, CA 90041. 'Stephens, J. S., Jr. 1978. Effects of thermal effluent from Southern California Edison's Redondo Beach steam generating plant on the warm temperate fish fauna of King Harbor Marina. Fish and laboratory study reports for Phase III. VANTUNA Research Group, Department of Biology, Occidental College, Los Angeles, CA 90041. 'Stephens, J. S., Jr. 1980. Effects of thermal effluent from Southern California Edison's Redondo Beach steam generating plant on the warm temperate fish fauna of King Harbor Marina. Fish and laboratory study reports for 1977-1978. VANTUNA Research Group, Department of Biology-, Occidental College, Los Angeles, CA 90041. 4Stephens, J. S., Jr., and J. B. Palmer. 1979. Can coastal power stations be designed to offset impacts by habitat enrichment? Gen. Tech. Rep. RM-65, p. 446-450. Paper presented at Mitigation Symposium, U.S. Department of Agriculture, Fort Collins, Colo. responses to intermittent chlorination (Hose and Stoffel 1980; Hose et al. in press). The objective of this study was to examine the feeding habits of black- smith and determine whether the discharge was attracting them through dietary enrichment. MATERIALS AND METHODS This study was conducted at King Harbor, Redondo Beach, Calif., at the southern end of Santa Monica Bay, just north of the Palos Verdes Peninsula (Fig. 1, lat. 33°51'N, long. 188°24'W) (Terry and Stephens 1976; Stephens and Zerba 1981). Situated just offshore is the head of the Redondo Submarine Canyon, a source of cold upwelling water for the har- bor. In contrast, thermal effluent from Units 7 and 8 of Southern California Edison's Redondo Beach steam electrical generating plant is discharged just inside the harbor mouth. The thermal outfall study site consists of a vertical conduit, 4 m in diameter, out of which the effluent is pumped. The circular outlet is level with the sub- strate at a depth of 7 m. Effluent is discharged at a rate of 1.78 X 106 1/min during peak operation. A control site was chosen about 500 m from the dis- charge. This area, referred to as the Point, is located at the tip of the breakwater that partially encloses the harbor. This site has been surveyed by Stephens and Zerba (1981) who note that blacksmith are an abun- dant resident species. A form of presence/absence monitoring was used as an indicator offish abundance at the discharge. Mean estimates (0-25, 26-50, 51-75, 76-100, or >100) were made by two scuba divers swimming a circular tran- sect around the discharge. The position of fish was recorded: in the plume (the column of water directly over the discharge), in the outer plume (the area of Manuscript accepted June 1983. FISHERY BULLETIN: VOL. 82. NO. 1. 1984. 199 FISHERY BULLETIN: VOL. 82. NO. 1 Santa Monica Bay Redondo Beach Los Angel es 100 500 I 1 1 1 1 1 I meters FIGURE 1.— Study area at King Harbor, Redondo Beach, Calif. water immediately surrounding the plume), or at the base (the substrate surrounding the discharge). The abundance of fishes at the Point has been documented since 1974 (Stephens and Zerba 1981), and work continued at this area during the same time period the discharge was examined. Two divers equipped with slates and depth gauges swam in one direction along the rock breakwater at a fixed depth for 5 min, counting all fish seen 1.5 m above and below them and within sight to either side. Transects were run at depths of 1.5, 4.5, 7.5, and 10.5 m, with replicates at each depth. In order to determine the nature of the feeding habits of blacksmith at the discharge versus those feeding at the Point, utilization of food items based on stomach analysis was examined for each area. General availability of food was estimated by sam- pling plankton at both sites. Stomach analysis closely followed methods em- ployed by Ellison et al. (1979). Fish were collected from each study site by scuba divers using pole spears. During fish collection, a temperature profile was taken using a temperature probe coupled to a telethermometer (Yellow Springs Instruments Co., Model 43 ID5). After capture the fish were placed on ice. The body wall was cut open and the stomach injected with a 207c Formalin solution. The fish were then preserved in a 1 0% Formalin solution for at least 48 h, rinsed in running water for 2 h, and placed in 70% isopropyl alcohol. Within 2 wk from date of capture, fish stomachs were removed and placed in vials of 70% isopropyl alcohol. At this time the standard length, wet weight, and sex of each fish were noted. Each stomach was then blotted dry (with special care taken to remove the internal fluid) and weighed, food items dissected out, and the empty stomach weighed again. Stomach fullness was estimated using a scale from 0 (empty) to 5 (full). Individual prey items were separated into the lowest identifiable taxa and counted, and the percent of the total volume estimated. In most cases, only whole organisms or whole organism indicators were counted. In prey items which were not eaten whole (i.e., algae and ectoprocts), only the percent volume was estimated. In 1979-80, 73 fish were collected at the discharge area from 13 sampling days during a 15-mo period. Four sampling days were in the afternoon (1430- 1830 h) and 10 were in the morning (0830-1100 h). A total of 35 blacksmith were collected from the Point area before noon (1000-1130 h). During the study period, 28 plankton samples from the discharge plume and 13 plankton samples from the Point were collected. The mean rank order abun- dance of prey items from each site was determined for comparison with blacksmith stomach contents. Observations comparing different prey items from two locations were tested using contingency table analysis, the G-test (Crow 1982), and Kendall's coef- ficient of rank correlation. When only one variable (fish weight, stomach fullness etc.) was tested be- tween two locations, a two-sample f-test was used, assuming separate variances. Values of the Index of Relative Importance (IRI) were calculated for con- sumed prey from the sum of the percent number and the percent volume, multiplied by the frequency of occurrence (Foe) (Pinkas et al. 1971). Dietary overlap between blaeksmth from the Point and discharge was examined using the formula of Schoener (1970): a 1-0.5 ( Z \PxrPy, 'Reference to trade names does not imply endoresement by the National Marine Fisheries Service, NOAA. where n is the number of food categories, x, is the average percentage of estimated volume that food category ;' contributed to species at location x, andy, is the average percentage of estimated volume that food category i contributed to species at location y. 200 MORRIS: FEEDING HABITS OF BLACKSMITH An estimate of mean prey size was obtained by dividing the total number of prey consumed into the stomach weight for each fish collected. RESULTS Thirty species of fish were identified from the area surrounding the discharge. Blacksmith were the most abundant and frequently occurring fish (mean estimate of abundance per transect >100 in- dividuals, rank of the mean number per transect = 1 , and frequency of occurrence per transect = 92.3). Large schools arrived in the morning and began feed- ing in the plume and outer plume. When feeding in the outer plume, blacksmith would orient themselves toward the plume, surrounding it, and feed on the organisms that settled out of the rising effluent. When in the plume, blacksmith were in constant motion, being tossed about by the irregular flow, but it was apparent from mouth action that these fish were also feeding on suspended food items. The mean abundance per transect of blacksmith at the Point for the quarterly sampling days in 1979 and 1980 was 148.4. They ranked first in mean number per transect, with a mean frequency of occurrence of 86.2, and used the breakwater as their primary noc- turnal sheltering site. There were no significant differences in either fish length or fish weight, but there were significant dif- ferences in stomach weight and stomach fullness be- tween the two collection sites (Table 1). Fish collected from the discharge had a greater amount of food in their stomachs (an increase of 138%). Stomach fullness was not influenced by collection time. The stomach weight and stomach fullness were TABLE 1. — Comparison of blacksmith, Chromis punctipinnis, collect- ed from the discharge (thermal outfall) and the Point (Control Site), King Harbor, Calif. Discharge Point n = 73 n = 35 Fish weight (g| Mean = 175.6 g Mean = 168.3 g SD= 38.4 SD = 44 7 r = 0819 P = 0 416 Fish length (SL mm) Mean = 1 72.8 mm Mean = 1 72.2 mm SD= 12 9 SD= 14,3 t = 0 569 P = 0.571 Stomach weight (g) Mean = 1 10 g Mean = 0.30 g SD = 0.53 SD = 0.25 t= 9 726 'P<0.001 Stomach fullness (0-5) Mean = 3.89 Mean = 1.63 SD= 1.06 SD= 1 .09 f = 10.175 P<0.001 'Note: The statistical package (SPSS) used was unable to compute P values lower than 0.001 . Values below this number are represented as P < 0.001 . not significantly different between morning and afternoon collections (£-test:£ = 1. 359, P= 0.181 and t = 1.471, P = 0.147, respectively). Consequently, the data collected from the discharge samples were combined. The mean prey abundance, percent number, per- cent volume, frequency of occurrence, and the calculated IRI value of the 30 most abundant prey items from each location are given in Table 2. A con- tingency table analysis of the mean abundance indicates thatthere was a significant difference in the stomach contents between the two locations (G = 570.6, P <0.001, df = 17). The 10 most abundant prey from each site (eliminating the smaller values) are significantly different (G = 56 1 . 1 , P < 0.00 1 , df = 12). A comparison of the 10 highest IRI values from each site are not significantly correlated (Kendall's tau, t = 0.1868, P = 0.324, n = 14). A pictorial rep- resentation of the IRI values is given in Figures 2 and 3. A comparison of the mean prey weight from each sampling site revealed that blacksmith from the dis charge ate larger prey than blacksmith from the Point (discharge mean prey weight = 3.22 mg, SD = 4.01, Point = 0.82 mg, SD = 0.81, t = 4.439, P <0.001). Temperatures from the discharge plume and base were compared with surface and bottom tem- peratures at the Point. The mean plume temperature (26.3°C, SD = 3.3, n = 15) was significantly greater (t-test: t = 5.69, P <0.001) than the mean surface temperature from the Point (20.8°C, SD = 2.5, n = 30). Similarly, the mean base temperature (18.2°C, SD= 2.4, n = 17) was significantly greater (t = 4.12, P < 0.001) than the mean bottom temperature from the Point (15.2° C, SD = 2.4, n = 30). The rank of the 10 most abundantly consumed prey items was compared with the rank of the 10 most abundant plankton items for both the discharge and Point. There was no significant correlation for either study site (discharge t = 0.01 10, P = 0.956, n = 14; Point t= 0.2051, P= 0.329, n = 13). Between-site comparisons of the mean abundance of six abundantly consumed prey items from both stomach contents and plankton samples (Table 3) show that two prey items, gammarids and Polyophthalmus pictus, had a significantly higher usage and availability at the discharge than the Point, and that Calanus sp. and mysids had a higher usage at the discharge but were not significantly more avail- able. There was no significant difference in the usage or availability of Oikopleura sp. between the Point and discharge (although blacksmith from the Point tended to eat a greater amount). The diets of blacksmith at the discharge and Point 201 FISHERY BULLETIN: VOL. 82. NO. 1 Table 2. — The 30 most abundant food items consumed by blacksmith, Chromis punctipinnis , at the dis- charge (thermal outfall) and the Point (control site), King Harbor, Calif. Foc= frequency of occurrence; IRI = index of relative importance. Point Discharge ■ % % X % % no no. vol. Foe. IRI no. no vol. Foe. IRI Oikopleura 43049 77.5 41.3 77.1 9,159 5 29063 33.7 108 87 7 3.902.7 Ac am a 5903 10.6 3.2 71.4 985.3 46.38 5.4 2.1 67,1 503.3 Calanoids. misc. 26.1 1 *1 3.2 71.4 564 1 14.86 1.7 1.4 75 3 2334 Polychaeta. misc. 8.03 14 1 6 57 1 171.2 2.53 0.3 1.2 50 7 76 .1 Corycaeus 6.51 1.2 1.4 45 7 118 8 3.68 0.4 0,5 42.5 38.3 Calanus 529 0 9 5.7 62 9 415.1 298 36 34.6 110 76.7 3.497.5 Chaetognath 4 51 0.8 3 4 57.1 239 8 4,14 0.5 0.5 34.2 27.4 Labidocera 2 60 0 5 :> 2 48.6 131 2 2.12 0.2 0.6 52 1 41.7 Brachyuran zoea 1 89 03 0 9 22.9 27 5 8 14 09 1.2 46.6 97 9 Gammandae 1.83 0 3 1 1 42.9 60 1 1 11.33 12 9 25.3 91.8 3,506.8 Pagurid zoea 1.63 0 3 0 h 25.7 23.1 3 27 0.4 0.5 43.8 394 Cladocera 1.60 0.3 0 4 28.6 200 0.49 0.1 0,1 20,5 5.7 Hhincalanus 1 .17 0 2 06 34.3 27.4 2.27 0 3 0,6 41.1 30.9 Euphausids 097 0 2 04 25.7 154 082 0.1 0 I 26.0 5 2 Tortanus 077 0.1 1.2 286 37 2 1 52 0.2 0,4 32.9 17 2 Cypns larvae 0.54 0 1 02 22 9 69 1.10 0 1 0.2 37,0 11.1 Fish eggs 0.49 0.1 0.2 28.6 8 6 0.53 0 1 0.1 24,7 4 9 Cirnpide exoskel. 0.46 (I I 0.1 143 2.9 0.42 1.) 1 0.6 27.4 10 0 Polyophthalmus pictus 0.34 0.1 0 1 2.9 0 3 25.70 3.0 6 8 28.8 282.2 Gastropoda 0.34 0.1 0 2 22.9 6 9 0.37 0 1 0 1 205 4 1 Fish larvae 0.31 0.1 0.3 17 1 6.8 3 29 0.4 1 4 35.6 64.1 Mysids 0.31 0 1 02 200 6.0 3601 4.2 7 7 80.8 961.5 Ophelndae 0.14 0.1 0.1 86 1.7 0.90 0 1 03 30.1 12.0 Decapoda. misc. 0.06 0.1 0 2 5.7 1.1 0.55 0 1 0 9 30.1 30 1 Caprellidae 003 0 1 0 1 2 9 0 3 1.90 0 2 1.0 466 55 9 Porcellanid zoea 003 0.1 0 1 2 9 0 3 0.89 0 1 0.2 32.9 9 9 Pelecypoda n 0 0 0 0 0.60 0 1 0.4 24,7 12 4 Anemone 0 0 0 0 0 3.29 0.4 0,5 15.1 13.6 Ecto-Entoprocta — — 0.1 2 9 0.3 — — 1 5 8.6 12 9 Unidentified, misc. — — 11 2 35.3 395 4 — — 69 42.9 296.0 TABLE 3. — Usage and availability of selected prey items from the Point (control site) and discharge (thermal outfall), King Harbor, Calif. In stomachs1 In plar ikton2 Prey items Discharge Point Discharge Point Polyophthalmus pictus Mean 25 70 0.34 30.59 0 SD 67.49 2.03 86 52 0 t= 3.207 P = 0.002 3p<0.001 Acartia Mean 46.38 5903 181,987.13 167,487.59 SD 112.02 130.95 323.297.81 133,525.13 t = 0.492 P= 0.625 r = 0.031 P> 0.840 Calanus Mean 29936 5 29 364.41 721.00 SD 753.29 11.96 717.27 1,260.84 t = 3.323 P= 0.001 t = -0.959 ^ = 0.353 Mysidacea Mean 3601 031 943.28 306 38 SD 75.38 0.72 3.562.00 568.29 t = 4.046 P < 0.001 r = 0981 P= 0.333 Grammandae Mean 111.33 1.83 6.291.81 472.92 SD 174.51 4.52 10.784.44 845.73 f = 5.357 P < 0 001 t = 3.029 P = 0.005 Oikopleura Mean 290.63 43049 6.82981 4,582,08 SD 471.00 557.59 19.821.55 9,906 22 t = -1.281 P= 0 205 r = 0 505 P= 0616 'Mean number of prey consumed per fish. 2Mean number per 100 m3 of water sampled. 3Note: The statistical package (SPSS) used was unable to compute P values low- er than 0.001 . Values below this number are represented as P <0.001 , did not overlap (a = 0.522, with a value >0.60 con- sidered significant, Zaret and Rand 1971). DISCUSSION Blacksmith were a numerically dominant species at both study sites. The daytime abundance of black- smith was similar at the discharge and the Point. Blacksmith may travel to the discharge from the breakwater and other nearby jetties during the day, since they do not seek shelter around the discharge at night. Such diel migrations of blacksmith between the Units 7 and 8 intake of Southern California Edison's Redondo Beach Station and the nocturnal rocky shelters at the Point have been previously observed.6 The feeding habits of blacksmith were significantly different between the Point and discharge (Figures 2 and 3 best illustrate this difference). At the Point, Oikopleura and calanoid copepods (primarily Acar- tia) were the most heavily utilized organisms. At the discharge, blacksmith consumed larger organisms, gammarids, calanoid copepods of the genus Calanus, 6M. Helvey, VANTUNA Research Group, Occidental College, Los Angeles, CA 90041, pers. commun. 1980. 202 MORRIS: FEEDING HABITS OF BLACKSMITH 50- 01 n 40 E c 30 >> n 20 10 0 0) 10 E > >. 30- a ^ 40 CALANOIDS (1 Calanus) OIKOPLEURA ? i i GAMMARIDS POLYCHAETE Other CRUSTACEANS MISC 75 43 88 92 88 92 FREQUENCY OCCURRENCE FIGURE 2.— Graphic representation of the Index of Relative Importance of prey items consumed hy blacksmith, Chromis punctipinnis, at the discharge (thermal outfall) in King Harbor, Calif. -D E 3 C >< E 3 o > >> n 80 70 60 50 40 30 OIKOPLEURA 20 GAMRDS P0LYCHAETE CRUSTACEAN 10 0 10 " CALANOIDS i O 1 i (1°Acartia) 43 57 ' 71 MISC 35 20 83 30 40 77 50 P 40 - FREQUENCY OCCURRENCE FIGURE 3. — Graphic representation of the Index of Relative Importance of prey items consumed by blacksmith, Chromis punctipinnis, at the Point (control site) in King Harbor, Calif. large polychaetes, other crustaceans, as well as Oikopleura. At both sites blacksmith were selective in their planktonic feeding, consuming the largest prey items available. Brooks (1968) stated that there is selection for larger zooplankters, with smaller ones eaten as the larger ones become scarce. At the Point, Oikopleura was the largest prey item found in abun- dance, while at the discharge other larger food items 203 FISHERY BULLETIN: VOL. 82, NO. 1 were common along with Oikopleura (gammarids, Polyophthalmus pictus, and mysids). The amount of dietary overlap between the two locations was not considered significant. Although more abundant at the Point, a significant- ly greater amount of Calanus sp. was eaten by black- smith at the discharge than at the Point. A possible explanation for the high usage of Calanus at the dis- charge could be the increased susceptibility of zooplankton to predation as a result of turbulent out- flow. Entrained Calanus are more accessible to planktivorous fishes, since the mortality rate of copepods passing through a power plant may reach 70% (Carpenter et al. 1974). Dead or damaged copepods would appear as viable prey upon dis- charge from the plant and could be easily consumed. Increased mortality from turbulence has also been shown for other zooplankters (Gregg and Ber- gersen 1980). There is evidence that alterations in plankton dis- tributions at outfall areas are the result of upward vertical displacement of deep-water organisms. Evans (1981) noted that deeper living zooplankton are carried vertically upward to the turbulent waters over the discharge jets. Although analysis of plankton sampled did not prove the existence of such currents, in a previous study at King Harbor dye injections were carried to the plume from bottom water 20 m away from the discharge.7 Large gammarids, polychaetes, and juvenile anemones, all of which were common in stomachs of blacksmith from the discharge, are not normal con- stituents of King Harbor plankton. The force of the swirling effluent is strong enough to detach and entrap these organisms from their normal habitat inside and around the discharge pipe. Once entrapped in the plume, these large invertebrates are accessible to the planktivorous blacksmith. Zooplankton avoid predation through escape movements upon detection of suction currents created by predatory fish (Dreeneretal. 1978; Kettle and O'Brien 1978). Once entrained in the effluent plume, the ability of zooplankton to detect these currents becomes impaired (Evans 1981). As a result, fish frequenting the plume have the potential for feeding on a high concentration of zooplankton with limited predator avoidance. The greater stomach weight and stomach fullness of blacksmith feeding at the discharge support this theory. Results from other studies examining the feeding 7Kinnetic Laboratories, Inc. 1981. Hydrodynamic characteris- tics of offshore intake structures. Field verification studies. Kin- netic Labs., Inc., P.O. Box 1040, 1 Potrero St., Santa Cruz, CA 95061. habits of blacksmith appear to be similar to those found at the Point. The food items consumed by blacksmth at Santa Catalina Island are (listed in de- creasing abundance) Oikopleura, calanoid and cyclopoid copepods, fish eggs, cladocerans, and other crustaceans (Hobson and Chess 1976). At Naples Reef, off Santa Barbara, Calif., Bray (1981) found the diet of blacksmith to consist of larvaceans (Oikopleura), copepods, cladocerans, chaetognaths, decapods, and polychaetes. In the two above- mentioned studies and from the Point, blacksmith consumed at least twice as many Oikopleura as any of the other food items, while at the discharge, Calanus was the most abundantly consumed prey and gam- marids comprised the greatest volume of prey eaten (Table 2). When Calanus, gammarids, mysids, and the polychaete Polyophthalmus pictus are removed from the analysis of the 10 most abundant prey con- sumed, no significant difference was observed be- tween the two locations (G — 9.4, n.s. atP= 0.05, df = 7). It has long been recognized that blacksmith forage on plankton in areas where currents are present (Limbaugh 1955, 1964; Feder et al. 1974; Ebeling and Bray 1976; Hobson and Chess 1976; Bray 1981). The tropical species of damselfish (family Pomacen- tridae) also prefer feeding in areas where currents are strong (Hobson and Chess 1978). Blacksmith have been shown to prefer incoming currents (Limbaugh 1955, 1964; Ebeling and Bray 1976; Bray 1981), and Limbaugh believed they materially affected the amount of plankton entering the kelp beds. In Bray's (1981) study, stomach fullness was greater in fish at the incurrent end of the reef than in fish at the excurrent end. Areas of strong currents are rich in zooplankters (Hobson and Chess 1978) as is the discharge which receives both entrained and entrapped organisms. Although the discharge releases warm water, the current created by the outflow is the major attract- ant. Blacksmith, a species which prefers warm water (mean preferred temperature = 14°-15°C), are found in 26°-32°C discharge plume water, above their upper temperature avoidance limit of 23°-25°C (Shrode et al. 1982). In the presence of food, black- smith will disregard their normal avoidance limits for chlorine, intermittently present in most power plant effluents (Hose and Stoffel 1980). It can be concluded that the outflowing effluent and its related phenomena attract blacksmith to the dis- charge. This theory is further supported by documentation of similar attraction and rheotropic behavior by blacksmith at an offshore water intake structure (Helvey and Dorn 1981). 204 MORRIS: FEEDING HABITS OF BLACKSMITH ACKNOWLEDGMENTS For their help in the collection of field data, I wish to thank K. Zerba, D. Terry, K. Shriner, C. Rand, T Wong, B. Johnson, J. Hough, and L. McCluskey. I also wish to thank M. Love, J. E. Hose, G. Martin, and R. N. Bray for their help in reviewing this manuscript. My utmost appreciation goes to J. Stephens, Jr., for his support and guidance throughout this project and to S. Warschaw for her generous help in preparing the final copy. I wish to acknowledge the support of J. Palmer and Southern California Edison Research and Develop- ment Project No. C0650901 to J. Stephens, Jr. LITERATURE CITED Bray, R. N. 1981. Influence of water currents and zooplankton densities on daily foraging movements of blacksmith, Chromis punctipinnis, a planktivorous reef fish. Fish. Bull., U.S. 78:829-841. Brooks, J. L. 1968. The effects of prey size selection by lake plank- tivores. Syst. Zool. 17:272-291. Campbell, R. C. 1974. Statistics for Biologists. 2d. ed. Cambridge Univ. Press, 385 p. Carpenter, E. J., B. B. Peck, and S. J. Anderson. 1974. Survival of copepods passing through a nuclear power station on northeastern Long Island Sound, USA. Mar. Biol. (Berl.) 24:49-55. Crow, M. E. 1982. Some statistical techniques for analyzing the stomach contents of fish. //; G. M. Cailliet and C. A. Simenstad (editors), Fish food habits studies, p. 8-15. Proceedings of the Third Pacific Workshop. Washington Sea Grant Publ., Univ. Washington. Dreener, R. W., J. R. Strkkler, and W. J. O'Brien. 1978. Capture probability: the role of zooplankter escape in the selective feeding of planktivorous fish. J. Fish. Res. Board Can. 35:1370-1373. Ebeling, A. W., and R. N. Bray. 1976. Day versus night activity of reef fishes in a kelp forest off Santa Barbara, California. Fish. Bull., U.S. 74:703- 717. Ellison, J. P., C. Terry, and J. S. Stephens, Jr. 1979. Food resource utilization among five species of embiotocids at King Harbor, California, with preliminary estimates of caloric intake. Mar. Biol. (Berl.) 52:161- 169. Evans, M. S. 1981. Distribution of zooplankton populations within and adjacent to a thermal plume. Can. J. Fish. Aquat. Sci. 38:441-448. Feder, H., C. H. Turner, and C. Limbaugh. 1974. Observations on fishes associated with kelp beds in Southern California. Calif. Dep. Fish Game, Fish Bull. 160, 144 p. Gregg, R. E., and E. P. Bergersen. 1980. Mysis relicta: Effects of turbidity and turbulence on short-term survival. Trans. Am. Fish. Soc. 109:207-212. Hklvey, M., AND P. Dorn. 1981. The fish population associated with an offshore water intake structure. Bull. South. Calif. Acad. Sci. 80:23-31. Hohson, E. S., and J. R. Chess. 1976. Trophic interactions among fishes and zooplankters near shore at Santa Catalina Island, California. Fish. Bull., U.S. 74:567-598. 1978. Trophic relationships among fishes and plankton in the lagoon at Enewetak Atoll, Marshall Islands. Fish. Bull., U.S. 76:133-153. Hose, J. E., AND R. J. STOFFEL. 1980. Avoidance response of juvenile Chromis punctipinnis to chlorinated seawater. Bull. Environ. Contam. Toxicol. 25:929-935. Hose, J. E., R. J. Stoffel, and K. E. Zerba. In press. Behavioral responses of selected marine fishes to chlorinated seawater. Mar. Environ. Res. Kelso, J. R. M. 1976. Movement of yellow perch (Perca flavescens) and white sucker (Catostomus commersoni) in a nearshore great lakes habitat subject to a thermal discharge. J. Fish. Res. Board Can. 33:42-53. Kettel, D., and W. J. O'Brien. 1978. Vulnerability of arctic zooplankton species to preda- tion by small lake trout (Saluelinus namaycush). J. Fish. Res. Board Can. 35:1495-1500. Limbaugh, C. 1955. Fish life in the kelp beds and the effects of kelp harvest- ing. Univ. Calif. Inst. Mar. Resour. Ref. 55-9, 158 p. 1964. Notes on the life history of two Californian pomacen- trids: Garibaldis, Hypsypops rubicunda (Girard), and blacksmiths, Chromis punctipinnis (Cooper). Pac. Sci. 18:41-50. Miller, S. 1977. The impact of thermal effluents on fish. Environ. Biol. Fish. 1:219-222. Minns, C. K., J. R. M. Kelso, and W. Hyatt. 1978. Spatial distribution of nearshore fish in the vicinity of two thermal generating stations, Nanticoke and Douglas Point, on the Great Lakes. J. Fish. Res. Board Can. 35:885-892. Pinkas, L., M. S. Oliphant, and I. L. K. Iverson. 1971. Food habits of Albacore, Bluefin Tuna, and Bonito in California waters. Calif. Dep. Fish Game, Fish Bull. 152, 105 p. Schoener, T. W. 1970. Nonsynchronous spatial overlap of lizards in patchy habitats. Ecology 51:408-418. Shrode, J. B., K. E. Zerba, and J. S. Stephens, Jr. 1982. Ecological significance of temperature tolerance and preference of some inshore California fishes. Trans. Am. Fish. Soc. 111:45-51. Stephens, J. S., Jr., and K. E. Zerba. 1981. Factors affecting fish diversity on a temperate reef. Environ. Biol. Fish. 6:111-121. Terry, C. B., and J. S. Stephens, Jr. 1976. A study of the orientation of selected embiotocid fishes to depth and shifting seasonal vertical temperature gradients. Bull. South. Calif. Acad. Sci. 75:170-183. White, J. W., W. S. Woolcott, and W. L. Kirk. 1977. A study of the fish community in the vicinity of a ther- mal discharge in the James River, Virginia. Chesapeake Sci. 18:161-171. Zaret, T. M., and A. S. Rand. 1971. Competition in tropical stream fishes: support for the competitive exclusion principle. Ecology 52:336-342. 205 CALIBRATION OF DENTAL LAYERS IN SEVEN CAPTIVE HAWAIIAN SPINNER DOLPHINS, STENELLA LONGIROSTRIS, BASED ON TETRACYCLINE LABELING Albert C. Myrick, Jr.,1 Edward W. Shallenberger,2 Ingrid Rang,2 and David B. Mac-Kay' ABSTRACT To calibrate dentinal and cemental growth layer groups (GLGs) with real time, a study was conducted on the teeth from seven captive Hawaiian spinner dolphins that had been treated clinically with tetracycline (TCL) at numerous times over multiple years at Sea Life Park, Hawaii. To monitor layer accumulation as it occurred for 1 year, we gave single injections to three animals every 3 months and pulled a tooth from each every 6 months. By comparing dental-layer patterns between TCL labels that had been introduced at 6-month and 1-year intervals, annual patterns were distinguished. In the dentine, a thin, light layer (the first being the neonatal line) was formed about every 6 months. Each annual GLG contained 13 lunar monthly layers (LMLs). Using LML or light-layer counts, age, month, and year of birth were estimated for each of the seven specimens. All seven deposited nearly the same dentinal GLG thickness in the same year of life. Estimates of birth months indicated that five of the animals were born in late summer or early autumn and two were born in spring. Com- parisons of dentinal labels with clinical records for a captive-born animal showed that TCL given to its mother was imparted via milk to the nursing calf. Time calibration of cemental GLGs showed that usually one cemental GLG was deposited annually, but in some cases a GLG was formed every second year or twice a vear. The technique of "reading" layers or growth layer groups (GLGs, terminology of Perrin and Myrick 1980) in teeth, developed to determine ages for pin- nipeds in the early 1950's by Scheffer (1950) and Laws ( 1952), is now used routinely in dolphin studies (see reviews by Klevezal' and Kleinenberg 1967; Jonsgard 1969; Scheffer and Myrick 1980). Early work on dolphins (e.g., Nishiwaki and Yagi 1953; Sergeant 1959), showing a correlation between apparent age and number of GLGs led to the working assumption that GLG-deposition cycles are con- stant, each GLG usually, but not always, interpreted as representing 1 yr. Critical analysis of this assump- tion has been impaired by a lack of suitable material. Three approaches have been used in efforts to calibrate dental GLGs with time and to determine their deposition rate: 1) In vivo labeling of tooth layers, 2) multiple extractions of teeth over time, and 3) examination of teeth from animals of known age. Nishiwaki and Yagi (1953) labeled the layered den- tine in four wild-caught striped dolphins, Stenella coeruleoalbo, by intramuscular injection of lead ace- tate paste. None of the four survived long enough for the labels to provide useful data. 'Southwest Fisheries Center, National Marine Fisheries Service, NOAA, La Jolla, CA 92038. 2Sea Life Park, Makapuu Point, Waimanalo, HI 96795. 'Kaneohe Veterinary Clinic, Kaneohe, HI 96744. Nielsen (1972) treated a young wild-caught harbor porpoise, Phocoena phocoena, with tetracycline (TCL) three times over a 370-d period. Three fluorescent labels were found in thin sections of its teeth examined in ultraviolet (UV) light "... but the uniform [unlayered] dentine made it impossible to determine the number of growth-layers formed per year" (Nielsen 1972:72). Best (1976) administered oral doses of TCL hy- drochloride, "Mysteclin-V", on each day over an 8-d period to each of three wild-caught dusky dolphins, Lagenorhynchus obscurus. Labels were detected in teeth of two of the three specimens after their deaths. In one specimen, dentine accumulated for 703 d be- tween treatment and death averaged 200 jum/yr and 0.56 jum/d. In the other (older) specimen, the average deposition rate in dentine between the treatment label and the pulp-cavity wall was 77 jiim/yr and 0.21 ftm/d. Best concluded that the thickness of GLGs decreases significantly with age in dusky dolphins. Gurevich et al. (1980) successfully introduced a single TCL label into the teeth of three of four wild- caught adult common dolphins, Delphinus delphis. The three labeled animals died 328, 354, and 441 d, respectively, after the date of treatment. By estimat- ing the dentinal pattern laid down in about 1 yr, the investigators characterized an annual GLG. They estimated the ages of the animals by assuming that the GLGs in the unlabeled regions of the teeth rep- Manuscript accepted August 1983. FISHERY BULLETIN: VOL. 82, No. 1, 1984. 207 FISHERY BULLETIN: VOL. 82, NO. 1 resented the same amount of time as the single GLG interpreted from the labeled region of each tooth. A study by Hui (1978) included two tooth extrac- tions made 2.5 yr apart from a captive male bot- tlenose dolphin, Tursiops truncatus (No. 10,"Kona"). Comparisons of longitudinal thin sections of the two teeth led Hui to conclude that "... almost three den- tin layers [GLGs] had been deposited during the intervening period. . . " (p. 1 1). Other than indicating GLG boundaries in figures of the two thin sections (his fig. 3), Hui did not describe the GLGs or their components. Three published studies (Sergeant 1959; Sergeant et al. 1973; Hui 1978) have attempted to demonstrate time content in GLGs using teeth of known-age, i.e., captive-born dolphins. All three had access to only a small number of specimens, all of Tursiops truncatus. Apparently, the investigators knew the ages of the specimens before defining and counting dentinal GLGs in the teeth, and no assurance was provided that the GLGs counted cor- responded to annual periods between birth and death. Hui's study demonstrated that GLGs may be defined in such a manner as to verify the age that is already known for a specimen (Myrick 1980a). The incorrect age data (3.3 yr) provided to Hui for one of two "known-age" specimens studied by him (Hui 1978) led to his subsequent division of its dentinal layering pattern into three GLGs and a small fraction (Hui4). The original clinical records for the specimen (Hui's No. 29, LACM 54698) show, however, that the dolphin was born on 28 August 1965 and died on 8 August 1969, at nearly 4 yr of age. Used independently, teeth of known-age animals, single-labeled teeth, or teeth extracted on two dates do not provide reliable means by which to determine tissue accumulation rates fully or to define GLGs with precision. Each method yields only two dates bracketing a segment of layered tissue into which the known elapsed time is divided. Myrick (1980b) de- scribed approaches that combine the use of two or more labels and two or more tooth extractions over an extended period to monitor rates and calibrate GLGs. The present paper is an account of such a study which used TCL-labeled teeth from seven cap- tive Hawaiian spinner dolphins, Stenella longi- rostris. MATERIALS AND METHODS The study consisted of two phases. The first was a retrospective examination of TCL labels in the dolphins' dental tissues produced incidentally by clinical treatments administered during their cap- tivity at Sea Life Park, Hawaii. Teeth were used from four frozen carcasses (Nos. WFP 606, 669, 670, and 67 15), including one specimen of known age, and three live animals (Nos. ACM 103, 104, and 106) from which teeth were extracted in early 1980. The second phase was a 1-yr monitoring of tissue- accumulation rates in teeth of three live animals. Each animal was given intramuscular injections of TCL at about 3 -mo intervals and underwent three tooth extractions during the monitored period. To restrain the dolphins during injections and extractions, an elevated rigid litter was placed near the edge of the dolphin holding tank in which the water level had been lowered to a depth of 0.5 m. The sloped tank bottom inclined the litter at an angle of 20° relative to the water surface. Each dolphin in turn was guided on its belly onto the litter until the front half of its body was above the water surface. In this position the dolphin could be held firmly with little apparent discomfort to the animal. The procedure used to extract teeth was adapted for the spinners from the method described by Ridgway et al. (1975) for bottlenose dolphins. The dolphin's mouth was held open by moistened rolled toweling placed around the upper and lower jaws. Carbocaine6 (5-10 cc) was injected into the right or left interalveolar nerve immediately behind the anterior border of the mandibular foramen. After allowing about 10 min for the anesthetic to take effect, a tooth was removed from the middle of the corresponding mandibular tooth row using an elevator and an extractor. The vacated alveolus was packed with cotton soaked with a ferric solution to control bleeding and promote healing. Liquamycin 100, a form of TCL, was injected into the dorsal musculature between the dorsal fin and the blow hole. To reduce the possibility of local inflammation of the tissue — a problem known to result from concentrations of TCL — each dose (25 mg/kg body weight) was distributed along the dor- sum at three separate sites. Untreated (cut or ground) thin sections and decalcified and haematoxylin-stained (D/S) thin sec- tions are the two most widely used preparations for dolphin teeth in age determination studies (see Perrin and Myrick 1980: 21 ff.). D/S sections pro- duce simpler, more uniform GLG patterns, but de- 4Clifford Hui, Naval Ocean Systems Center, San Diego, Calif., pers. commun. 1981. 'Skeletons are in the synoptic collection at Southwest Fisheries Center, NMFS, La Jolla, Calif. ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 208 MYRICK ET AL.: DENTAL LAYERS IN HAWAIIAN SPINNER DOLPHINS calcification removes TCL labels (Nielsen 1972). We prepared the spinner teeth using both methods. Untreated, mid-longitudinal thin sections, 100 [xm thick, were prepared by hand grinding and polishing teeth using 240 and 600 grit AI2O3 on a glass plate. Other teeth were decalcified in RDO7 for 6-8 h, rinsed, and cut with a microtome in longitudinal plane to produce 30 /xm thick sections that were stained in Mayer's haematoxylin for 15-30 min. Untreated and D/S preparations were mounted on slides in Permount or glycerin gel and covered with coverslips. To determine the pattern components of GLGs, the D/S and untreated thin sections were examined in plain transmitted light 39X and 150X with a Zeiss photomicroscope. TCL labels were viewed at the same magnifications with UV reflected light using a Zeiss fluorescent vertical illuminator with a filter- reflector No. 44-75-05 combination attached to the same instrument. Retrospective Calibration of Dentinal GLGs Dates and durations of treatment, date of birth (for one specimen) or capture, and dates of death (for four carcasses) were taken from clinical records main- tained for each dolphin during its captive life at Sea Life Park (for summaries see Myrick et al. in press). Data for each specimen were transcribed onto a cali- bration chart as the chronological series of event blocks, the relative width of a given block corresponding to the length of a given period of treatment. In each thin section showing distinct fluorescent labels under UV light, label thicknesses and interlabel distances were measured. Label- measurement data for each dolphin were entered on its chart as a series of blocks below the event blocks, with spacing and thickness scaled to the correspond- ing measurements. The treatment and label blocks were compared for spacing and thickness to identify the date each label was introduced. Connecting lines were drawn from the beginning and the end of each matched pair of blocks (Fig. 1C). A UV photograph of each thin section was used to identify and letter key labels that enclosed 6- or 12- mo segments of dentine. Labels and structural landmarks in the UV photograph were traced with a china marker on an overlay of transparent plastic. Using the landmarks, the tracing was lined up on the corresponding plain-light photograph onto which the 1 A commercial rapid decalcifying agent available through Dupage Kinetic Laboratories, Inc., Plainfield, 111. labels were reproduced to delineate layering pat- terns within the time segments. Each marked photograph was then inspected for repeating layer components to define GLGs and their subunits in the untreated thin section. GLGs defined in the labeled dentine of each thin section were used as a basis for identifying similar GLGs in the unlabeled regions of the dentine and permitted a complete series of GLG- thickness measurements and an estimate of dentinal age in years to be made for each animal. Dentinal GLGs in dolphin teeth are most easily dis- cerned in the region of the "shoulder", i.e., along a transect from near the base of the neonatal line (the first layer of the postnatal dentine), downward and inward at about a 30°-40° angle to the margin of the pulp cavity (for examples see Perrin and Myrick 1980: fig. 2; Hui 1978: figs. 1, 2, 3). For consistency, measurements of GLG and label thickness, taken perpendicular to the long axis of the teeth of the Hawaiian spinner dolphins, were made along tran- sects at a similar position and angle (Figs. 1A, B). However, a GLG or label may vary in thickness in localized regions of the dentine and may not be the same on both sides of a tooth because of tooth asym- metry. For these reasons, measurements were made on the most symmetrical side of a tooth and in regions where GLGs and labels were clearest and least vari- able in thickness; departing slightly from a uniform angle of transect. GLGs in the dentine of the corres- ponding D/S thin sections were defined and counted with the aid of GLG-thickness measurements obtained from the untreated section. Retrospective Calibration of Cemental GLGs Because fewer labels were observed in the cemen- tum than in the dentine of the same untreated thin section, it was assumed that those visible represent- ed condensed forms of only the brightest, thickest, or closely spaced groups of dentinal labels. This has been verified in bottlenose dolphins (Myrick 1980b) and recently in the present sample of Hawaiian spin- ners by observations that bright dentinal labels at the tooth base are continuous with cemental labels. Hence, cemental labels were lettered to correspond to the brightest dentinal labels, and the cemental layers between labels were calibrated using the time segments represented between the dentinal labels. The annual GLG pattern was defined as precisely as possible using the calibrated segments of the tissue, and the cemental GLG definition was tested by com- paring the dentinal GLG count with the cemental GLG count in untreated thin sections. In D/S thin 209 FISHERY BULLETIN: VOL. 82, NO. 1 Enamel Enamel Prenatal dentine Neonatal line Approximate region and angle of GLG and label measurements Cementum ABCP Pulp cavity Tetracycline margin labels , 654 3 2 1 i i Dentinal GLGs VIEW IN UV LIGHT VIEW IN PLAIN LIGHT B Periods of treatment Dentinal labels ig78 197g 1980 .i i i i i i i i i i I Death Pulp cavity FIGURE 1 .—Line drawing of hypothetical dolphin tooth in thin section showing appearance of TCL labels, A, B, C, D, under ultraviolet light (1A, left-hand side) and dentinal growth layer group (GLG) layering patterns under plain transmitted light (IB, right-hand side), and standard positions in tooth where label and GLG thickness are measured. 1 C illustrates method of identifying labels in tooth section with TCL treatment dates by comparing relative thickness and spacing of labels with treatment periods. 210 MYRICK ET AL.: DENTAL LAYERS IN HAWAIIAN SPINNER DOLPHINS sections, cemental GLGs were defined indirectly by comparing them with the pattern and number of cemental GLGs determined in untreated sections. Direct Monitoring Calculation of depositional rates and calibration and definition of GLGs in dentine and cementum were achieved by comparing tooth specimens con- taining successively introduced labels and/or additional tissue accumulated over the 1-yr period of monitoring. To make determinations, for cases in which labels were not distinct or not successfully pro- duced, the additional tissue was measured from structural landmarks or labels in the extracted series of thin sections. RESULTS Dentinal labels. — The untreated thin sections for all seven specimens contained multiple labels. Most attempts to match labels with treatments were suc- cessful (Figs. 2-6). However, in four specimens more labels occurred than could be accounted for from clinical records. In the only captive-born specimen, WFP 670, numerous TCL labels were observed (Fig. 7A, B), but only three were found to have been caused by intentional therapeutic treatments (Fig. 7D, labels C, F, and G). Labels A and B apparently were a result of TCL impaired to the then-calf through the milk of its mother, who was treated with the drug for two periods while the calf nursed. The other labels appear to have resulted from frequent ingestion of stolen TCL-dosed smelt intended for other dolphins being treated at various times while sharing a com- mon tank with this animal. No treatment was recorded for label A found in the dentine of dolphin carcass WFP 669 (Fig. 4A, C) and live dolphin ACM 104 (Fig. 6A, C). Judging from the relative positions of the "A" labels to the other labels for which matches were found with recorded treat- ments, "A" labels were introduced into both specimens at or about their respective dates of cap- ture. It is a fairly common practice in commercial aquaria to give medication (often tetracycline) to newly captured dolphins recovering from stress of capture and adjusting to the captive environment8. Labels B and G in the dentine of dolphin carcass WFP 671 could not be identified from clinical records (Fig. 5A, C), although the numerous other "William A. Walker, Los Angeles County Museum of Natural His- tory, Los Angeles, Calif., pers. commun. 1982. labels match well in relative thickness and spacing with the treatment dates for this specimen. In teeth of live dolphin ACM 103 the labels were indistinct. The presence of TCL, introduced clinically during three periods of treatment over 2 yr and experimentally at 3-mo intervals in 1980, was indicated only by several areas of hazy fluorescence in the dentine near the pulp cavity. Dentinal GLG pattern. — The use of plastic overlays of key labels enclosing 6-mo or 1 -yr segments of den- tine on plain-light photographs of the dentine for each specimen permitted repeated calibrations of the annual dentinal layering pattern for six of the seven specimens (the seventh specimen, ACM 103, had no discrete labels). In untreated thin sections, a dentinal GLG contained four major components deposited in the following sequence: 1) A thin, light (GLG-boundary) layer, 2) a thicker dark layer, 3) another thin, light (mid-GLG) layer, and 4) a second thick, dark layer (Figs. 3A, 4B, 5B, 6B). In addition to the four components, many of the earliest deposited GLGs had an infrastructure com- posed of finer alternating dark and light layers. Counts made at 150X under low transmitted light showed that each of these annual GLGs contained 13 pairs of fine layers (Figs. 3A, 4B, 6D, 7C). Where layers were sufficiently distinct to be counted be- tween labels (e.g., between label B and M, Fig. 4A, B), counts indicated that each pair |"LML," (lunar monthly layer) Myrick 1980b] represented about 1 lunar month. The full complement of LMLs was vis- ible throughout the dentine in the captive-born specimen, WFP 670, i.e., 13 LMLs in each of the first three complete GLGs and 9 in the incomplete fourth GLG (Fig. 7C). In specimen ACM 103, 13 LMLs were observed in the first 12 of the 14.5 GLGs present (Fig. 8). But in other specimens, LMLs were clear enough to be counted only in the first five or six GLGs. In D/S thin sections, the annual GLG pattern con- sisted of two lightly stained and two darkly stained layers. The thin, light, GLG-boundary layers and mid-GLG layers in untreated thin sections corres- ponded to the lightly stained layers in D/S thin- sections (Fig. 9A, B). LMLs were indistinct in almost all GLGs in D/S preparations. Age-specific GLG thickness.— Table 1, showing dentinal GLG thickness measurements made from the most symmetrical side of the tooth of each of the seven dolphins, indicates that for each animal a GLG of a specific thickness was produced that appears to be related to the year of life in which the GLG was formed, i.e., an age-specific GLG thickness. 211 FISHERY BULLETIN: VOL. 82, NO. 1 t> o ft 5 s cu ■•-> ** CM — Oi "3 "-1 k £ -a cu O TO O .> t- ^_> CU _TO I E H o o i_, _c -r CO cu — Cu -a Eh o U P £ 5 212 MYRICK ET AL.: DENTAL LAYERS IN HAWAIIAN SPINNER DOLPHINS 100 pm I I > A B * * Treatment Labels 1976 1977 i i i I i i n li i_i i i 1979 1980 I "in ii Hi lllllllllll T"c"dTT g\ IIIIIIIIIMI b l d t (■ b-\pC of second too(h FIGURE 3. — Labeled tooth taken from dolphin carcass WFP 606. A. Untreated thin section in plain light showing about eight annual GLGs in dentine (separated by arrows). GLGs divided approximately in half by thin, light mid- GLG layers (heavy dark marks). GLGs 6, 7, and 8 were interpreted from positions of tetracycline labels (lettered). Finer dark layers represent lunar monthly layers ( 1 50X). B. Dentine labels in UV light ( 1 50X) . C. Chart showing dates labels were introduced. 213 FISHERY BULLETIN: VOL. 82, NO. 1 O) - IN _ O) _ ■- -J c o CQ (D .S o is a. « * a > >,-2 to O -§■ >> ft ^ £ o a a ° a; co o> a> -C i3 ft — <*- . CD O .2 c O js c > CO ffl CO 0> — 05 tr to PL, CO a §1E b5 ■n ■> § en O co o CO ^ CO &c .s >> o K -a T3 5 >= p, T3 cc 0> Cut a> T3 u o .o CO o o "3 3 c c CO a '% o -C to < CO ft CO ft , ° CO >> en -O O T3 J .2 o 5 C o> "^ XI CO — CU tO £ "co 214 MYRICK ET AL.: DENTAL LAYERS IN HAWAIIAN SPINNER DOLPHINS B H h 200pm 1974 I I i I l l l l Treatments Labels 1975 1976 1977 I I I I i I I n I ii i I 1979 1980 l 1 1 I I I ii l 1 1 i I I i i A^B C No record E F ^G H No record FIGURE 5.— Labeled tooth of dolphin carcass WFP 671. A. Untreated thin section in UV light showing TCL labels (39X). B. Thin section in plain light showing almost eight complete GLGs as interpreted from labels. Light GLG boun- dary layers appear to have been deposited in or about March (39X). C. Chart showing match between labels and treat- ments. TABLE 1. — Mean age-specific thicknesses (fim) of completed dentinal growth layer groups (GLGs) in teeth of seven Hawaiian spinner dolphins, Stenella longirostris. Values are averages of at least three measurements per specimen, taken perpendicular to the long axis of the tooth in a stairstep fashion downward and inward from the base of the neonatal line to the pulp-cavity wall. Specimen GLG number no 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 WFP 670 240 240 180 — — — — — — — — — — — — WFP6691 240 240 172 5 150 1275 1275 92.5 85 65 57.5 — — — — — WFP 671 240 240 165 140 120 95 80 ______ __ ACM 104 240 230 170 140 130 90 90 70 70 60 55 — — — — ACM 103 240 240 180 160 110 80 80 70 60 60 60 65 55 40 40 ACM 106 240 240 1 50 1 30 1 1 0 90 90 60 60 60 55 — — — — WFP 606 240 240 1 80 1 50 1 20 90 90 — — — — — — — — /V 7776666444311 11 * — 2386 1710 1450 1196 95 4 87 0 713 613 58 1 55 0 — — — — SD 3.8 110 10 5 8 4 16 5 5 6 10 3 2 5 2.4 5.0 — — — — SE 1.4 4 2 4 3 34 6 7 2.3 5.2 1.3 12 2.9 — — — — 'Mean values ol measurements in untreated and D/S (decalcified and haematoxylin-stained) sections 215 FISHERY BULLETIN: VOL. 82, NO. 1 O 1976 1977 INlMll.lllllllll Treatments Labels 1978 1979 1980 PC of 3rd Tooth 100 pm I'll FIGURE 6.— Tooth of live dolphin ACM 104 extracted 2 February 1981. A. Untreated thin section in UV light showing location of TCL labels (150X). B. Same section as in 6A in plain light showing position of key labels bracketing last 4 yr of deposition. Light GLG boundary layers appear to have been deposited in or about August. C. Chart showing match of labels and treatments. D. Thin section showing 1 1 complete annual GLGs (separated by dark marks) as interpreted from labels (39X). 216 MYRICK ET AL.: DENTAL LAYERS IN HAWAIIAN SPINNER DOLPHINS 200|jm N 100pm D Treatments 1975 1976 I i I i I I i I II i I ll I I I I 1977 1978 labels I Ml I M I I I I I LUUJ I I I I I I I m^J I I III 1979 l i I I I I I Treatments to otners in same tank \ ABC v Neonatal line D E FG\. PC FIGURE 7.— Teeth from dolphin carcass WFP 670 (captive-born animal). A. Untreated thin section showing TCL labels in den- tine. Labels A-B apparently represent TCL imparted to this animal through its mother's milk (UV, 39X). N =neonatal line; PC = pulp cavity margin. B. Portion as shown in 1A showing numerous labels from TCL-dosed smelt stolen from other dolphins occupying the same tank. Labels F and G represent direct treatments administered shortly before death (UV, 150X). C. Thin- sectioned tooth showing three entire and one partial GLGs (indicated by heavy dark marks) in the postnatal dentine as inter- preted from TCL labels. LMLs are indicated by fine dark markers (plain transmitted light, 150>^. D. Chart showing dates of direct and presumed incidental introduction of TCL and corresponding labels identified in the dentine by relative label position and thickness. 217 FISHERY BULLETIN: VOL. 82, NO. 1 200jum B / /*>..J /K'&~ ^ >• ft 50/im Figure 8.— Untreated tooth of live dolphin ACM 103 extracted 25 January 1980. A. About 14'/2GLGs indicated (heavy dark marks) (39X). B. GLGs 8-14% showing thin, light boundary layers with dark margins. Thirteen LMLs indicated in each GLG 8-12 are particularly well developed (150X). Comparisons of age-specific GLG thickness among the specimens suggest that the animals deposited a GLG of similar thickness in the same year of life. In the first and second year, 240 /xm thick GLGs were deposited. In the third, fourth, fifth, sixth, and seventh years, thickness of GLGs averaged 171, 145, 218 119, 95, and 87 ju.m respectively. From the 8th to the 1 1th year, GLGs were between 71 and 55 jum thick. The data in Table 1 represent averages of at least three measurements per GLG per specimen. Cemental labels.— Relatively few TCL labels were MYRICK ET AL.: DENTAL LAYERS IN HAWAIIAN SPINNER DOLPHINS B FIGURE 9.— Comparison of GLG patterns in teeth from dolphin carcass WFP 606 prepared by two methods: A. Untreated thin section (39X). B. De- calcified and stained thin section (39X). Pi 76 r found in the cementum compared with those in the dentine of the specimens. In the captive-born specimen, WFP 670, with about 25 dentinal labels, the cementum contained only three labels. In specimen WFP 669, only four cemental labels were observed (Fig. 10A) compared with 30 dentinal labels (Fig. 4A). The cementum in the other specimens had either zero or 1 label, despite the numerous dentinal labels observed for each. Cemental GLG pattern. — In untreated thin sec- tions, a cemental GLG consisted of a dark layer and a light layer (Fig. 10B). In D/S sections it was com- posed of a dark-stained layer, corresponding to the dark layer in untreated sections, and a lightly stained layer (Fig. 1 1). In both types of preparations, the dark layers contained larger concentrations of ceraen- tocytes than did the light layers. Calibration of cemental GLGs. — Calibrations of cemental GLGs with those in the dentine were car- ried out using the assumption that cementum is a less sensitive recording structure than dentine (Klevezal' 1980) and that labels occurring in the cementum cor- responded only to the brightest and thickest labels or label groups in the dentine. Thus, for example, the four labels detected in the cementum of specimen WFP 669 (Fig. 10A) were flagged with the same let- ters used to identify multiple label concentrations in the dentine (Fig. 4A). In some cases, such as in WFP 669, plastic overlays were used to determine that a cemental GLG rep- 219 FISHERY BULLETIN: VOL. 82, NO. 1 B FIGURE 10.— Tooth cementum of dolphin carcass WFP 669. A. TCL labels interpreted as corresponding to lettered dentinal labels (150X). B. Positions of TCL labels (arrows) in layered cement. About 10 GLGs are indicated (150X). resented the same amount of time as a dentinal GLG, i.e., 1 yr (e.g., Fig. 10B). In other cases, where labels were absent or where only one label occurred, cali- bration of cemental GLGs with dentinal GLGs was made indirectly by comparing GLG counts from both tissues. This method usually demonstrated a one-to- one relationship of GLGs in dentine and cementum, but in a few regions of the cementum of the captive- born specimen, WFP 670, there were twice as many GLGs as in the dentine (Fig. 11), indicating that a GLG may have been deposited twice a year in the cementum. In expanded regions of the cementum in another specimen (ACM 104; see Table 2), the cemental count was equal to the dentinal count; but in thinner regions the cemental count was only half that of the dentine, suggesting a cemental GLG being deposited every 2 yr in some cases. Direct monitoring. — The results of examinations of thin sections of the series of three teeth from each of three live animals, taken at the beginning, at mid- point, and at the end of a 1-yr monitored period are 220 MYRICK ET AL.: DENTAL LAYERS IN HAWAIIAN SPINNER DOLPHINS FIGURE 11.— Cementum of dolphin carcass WFP 670 tooth in de- calcified and stained thin section. The number of dark layers is eight, about double the age in years of this captive-born specimen (150X, plain transmitted light). presented in Table 2. Although distinct labels were not always successfully introduced, dentinal and cemental GLGs continued to be accumulated at a uniform rate of one per year. A comparison of accumulated dentine and labels in the first two extracted teeth of specimen ACM 106 (Fig. 12) showed two experimental treatments and one (unscheduled) clinical treatment accounted for in the second tooth (Fig. 2A-C). Specimen ACM 103, in which premonitor labels were indistinct, showed no experimental labels but clearly showed continued accumulation of dentine, the third extracted tooth having added about one GLG over the 1-yr period. No experimentally introduced labels were observed in the (less sensitive) cementum in any teeth of the three animals, but cemental deposition of about one complete GLG occurred in each animal for the period. Seasons of birth. — By determining the dates of key dentinal labels introduced at or near the thin, light component layers of GLGs and by noting the approx- imate time of formulation of component layers in the teeth extracted during the monitor period, it was found that GLG-boundary and mid-GLG layers were formed at about 6 mo intervals. In five specimens, GLG-boundary layers were deposited in or about August and the mid-GLG layers were deposited in or about March. In the two other specimens the timing was reversed, i.e., GLG-boundary layers formed in March, mid-GLG layers in or near August, Proceed- ing on the assumption that the timing of layer forma- tion (determined from the labeled or monitored Table 2.— Results of examinations of teeth extracted from three live Hawaiian spinner dolphins, Stenella longirostris , over a 1-yr period monitoring accumulation of layers and labels. GLGs = growth layer groups. Specimen no. and tooth Date of tooth Date label Dentine Cementum extraction introduced Additional labels No. GLGs Additional labels No. GLGs ACM 103 First 25 Jan. 1980 25 Jan 1980 30 Apr 1980 — 14 5 " 14.5 Second 30 Julv 1980 30 July 1980 30 Nov 1980 indistinct 150 None 150 Third 2 Feb. 1981 — indistinct 15 5 None 15.5 ACM 106 First 19 Mar. 1980 19 Mar. 1980 11-28 Apr 1980' 5 June 1980 103 10 Second 30 July 1980 30 July 1980 30 Nov. 1 980 3 10.7 None 10+ Third 2 Feb 1981 — indistinct 112 None 11 ACM 104 First 25 Jan 1980 25 Jan 1980 30 Apr 1980 — <2> (2) Second 30 July 1980 30 July 1980 30 Nov. 1980 2 10.7 None 35/10 Third 2 Feb. 1981 — 4 11.3 None 36/11 Unscheduled medical treatment 1 8 d in duration. Not examined because of poor preparation of section. 3Cementum showed a number of GLGs equal to that of the dentine as well as half that of the dentine. 221 FISHERY BULLETIN: VOL. 82. NO. 1 222 MYRICK ET AL.: DENTAL LAYERS IN HAWAIIAN SPINNER DOLPHINS regions of the dentine) was uniform throughout all of the dentine for a given specimen, GLG-boundary and mid-GLG layers were counted in reverse order of deposition up to the first boundary layer, the neona- tal line, to estimate month and year of birth. Table 3 summarizes the month- and year-of-birth estimates made from boundary-layer counts in six specimens and birth dates taken from park records of two captive-born specimens, WFP 670 and the calf of ACM 104. Six were born in late summer/early autumn and two in March. TABLE 3.— Estimated birthdates of eight cap- tive Hawaiian spinner dolphins, Stenella long- imstris.. Specimen Month and year no. of birth ACM 103 August 1964 ACM 106 August 1969 WFP 669 August 1969 ACM 104 September 1969 WFP 606 March 1972 WFP 671 March 1973 WFP 670' 8 September 1975 Calf of ACM 1042 21 July 1977 'Born in captivity. 2Born in captivity, survived 3 d DISCUSSION Age-Specific GLG Thickness Dentinal GLG thickness appears to be age-specific for the Hawaiian spinner dolphin teeth examined. There was little variability from tooth to tooth or from animal to animal in the sequence of GLG thick- ness through the 11th GLG, despite deposition of a specific GLG in some specimens while still in the wild and in other specimens during their captive lives. This suggests that, to some extent at least, the amount of dentine deposited by animals at a given age may be predetermined and that animals of a given stock, species, or higher common phylogenetic affinity may follow the same or similar pattern of age- specific GLG deposition unaffected by environ- ment. Used in conjunction with the GLG component-layer pattern, the regularity in thickness of age-specific GLGs may be useful as an aid in locating GLG boun- daries and counting GLGs in teeth of wild Hawaiian spinner dolphins and dolphins of related species in which GLG thickness and component-layer patterns are found to be similar. When measurements are taken at standard positions in the teeth of such dolphins, one may make fairly rapid age estimates without having to examine each GLG in detail (see Myrick et al. 1983). Lunar Monthly Layers (LMLs) Laws ( 1 962) was the first to suggest that the system of fine layers within dentinal GLGs of pinniped teeth corresponded to lunar monthly cycles. Putative LMLs have been reported in dentine of dugongs (Kasuya and Nishiwaki 1978; Marsh 1980), in den- tine of beaked whales ("short cycles," Kasuya 1977; "accessory layers," Perrin and Myrick 1980:3, 5), in fossil dolphin teeth (Myrick 1979), and in the man- dibular bone (Myrick 1980b) and dentine of modern dolphins (Myrick 1980b; Hohn 1980a, b). Hui (1978) reported finding no relationship between the fine layers that he counted in a tooth from a known-age bottlenose dolphin and its age in lunar months; but with no prior knowledge of its age, Myrick (1980b) made dentinal LML counts in the same specimen that closely agreed with its known age. The present study has furnished verification that LMLs are deposited with lunar-monthly regularity in the animals studied. In the 3.7-yr-old captive-born spinner dolphin (WFP 670), 13 LMLs were counted in each of the three complete annual dentinal GLGs and 9 were counted in the partial fourth GLG. Where LMLs were visible between TCL labels in the den- tine in this and other specimens, they were found to correspond consistently in number to the time in months represented between labeling dates. Where LMLs could be seen clearly, no departure from the 13 LML/GLG pattern was detected in the teeth used in the present study. Variability has been reported in studies of other marine mammals. Marsh (1980:197) found only "about 12 [LMLs] per GLG" in the dentine of the deciduous incisor of a dugong. Ten to 15 LMLs/GLG were observed in dugong tusks by Kasuya and Nishiwaki (1978). Kasuya (1977) found between 11 and 13.4 LMLs ("short cycles")/ GLG in teeth of Baird's beaked whales, Berardius bairdii. Hohn ( 1 980b) counted 10-13 LMLs/dentinal GLG in Atlantic bottlenose dolphin teeth. Pre- sumably, LML variability will be found to occur also in Hawaiian spinner dolphins when larger samples are examined. Relationship of Cemental GLGs to Dentinal GLGs None of the teeth of the studied specimens had reached the stage of pulp-cavity occlusion or dentinal irregularity that necessitated age estimation solely from cemental GLG counts (Kasuya 1976; Myrick et al. 1983). Although the pulp cavities were small in some specimens and some later-administered TCL 223 FISHERY BULLETIN: VOL. 82, NO. 1 failed to produce distinct labels, none showed evidence of cessation of dentine deposition. All cemental GLG counts corresponded in number to dentinal annual GLG counts except in the case of specimen WFP 670, where some regions of the cementum showed double the number, and in specimen ACM 104, where in some places the cemental count was half that of the dentine. The find- ing that in some cases cemental GLGs may form at half or twice the rate of dentinal GLG deposition points up the problem of using cemental GLGs to estimate ages without reference to the dentine (Myricket al. 1983). Evidence for an Internal Clock In the dentine of the animals studied, a thin GLG boundary layer, beginning with the neonatal line, was formed in the month of birth and on anniversaries of the month of birth. Mid-GLG layers were formed about 6 mo after formation of boundary layers. Where LMLs could be calibrated, one was found to form about every (lunar) month with high uniformity in relative spacing. Such a cycle of deposition is indicative of an internal clock, or clocks. The pattern commences at birth and apparently is reset with solar and/or lunar regularity without perceptible altera- tion by fluctuation in the dolphins' natural or captive environment or in calendric season of birth. That it may not be a totally free-running system, i.e., not without external cues, is suggested by the precisely synchorized deposition of the fine and coarse pat- terns of the dentine repeated over many years. Age at Sexual Maturity Perrin et al. (1977) indicated that sexual maturity may be reached in females of Stenella longirostris at an average 5.5 yr (range of 5-9 yr) and the average period of gestation may be about 1 1 mo. From the study of dentinal GLGs and TCL labels of specimen ACM 104, it was possible to determine that this animal was about 8-yr-old when she gave birth to her calf. Assuming an 11 -mo gestation, we estimate that she would have been 7-yr-old when she conceived. It is not known whether the pregnancy resulted from fertilization at her first or subsequent ovulations. ACM 104 remains alive. This precludes examination of her ovaries for ovulation scars. Reproductive Seasonality Based on the birth records of specimen WFP 670 and the calf of ACM 104 and the deductions made from dentinal layers, six animals were born in late summer/early fall, and two were born in March. Since all animals in the study represented the same popula- tion off Kona, Hawaii, the early-spring and early-fall birth patterns might indicate a corresponding two- cycle pattern of reproductive peaks for the wild pop- ulation generally. Such a seasonal pattern has been suggested by Norris and Dohl (1980, fig. 16), but Wells (in press), who has studied the population in considerable detail, concluded that the breeding season occurs from spring to fall, with most births in the fall. Our sample was too small to verify Wells' findings. Tetracycline Exposure to the Calf Through the Milk The first two labels found in the dentine of specimen WFP 670, the captive-born animal, were interpreted as having been introduced through milk received by the calf while the mother was being treated with TCL. This recommends a possible prac- tical application in indirectly treating newborns in ill health. Excessive handling of such animals fre- quently results in a worsening of their condition, making the treatment more dangerous than the malady. (Nursing calves not on solid food cannot be treated with TCL-dosed fish and must be force-fed or injected with drugs.) Separating the young calf from its mother may produce additional com- plications.9 If treatments for the calf could be administered through the milk by treating the mother with TCL-dosed food, it seems likely that most of the problem could be minimized. The question invites further study. ACKNOWLEDGMENTS We thank D. G. Chapman, R. L. Brownell, Jr., D. B. Siniff, A. Wild, W. F. Perrin, F. Hester, A. Dizon, and J. Barlow for their critical reviews of the manuscript. K. Raymond and R. Allen prepared the figures. M. DeWitt typed the manuscript. LITERATURE CITED Best, P. B 1976. Tetracycline marking and the rate of growth layer for- mation in the teeth of a dolphin (Lagenorhynchus obscurus). S. Afr. J. Sci. 72:216-218. Gurevich, V. S., B. S. Stewart, and L. H. Cornell. 1980. The use of tetracycline in age determination of com- 9L. H. Cornell, Sea World, Inc., San Diego, Calif., pers. commun. 1980. 224 MYRICK ET AL.: DENTAL LAYERS IN HAWAIIAN SPINNER DOLPHINS mon dolphins, Delphinus delphis . In W.F.PerrinandA. C. Myrick, Jr. (editors). Age determination of toothed whales and sirenians, p. 165-169. Rep. Int. Whaling Comm., Spec. Issue 3. HOHN, A. 1 980a. Age determination and age related factors in the teeth of western North Atlantic bottlenose dolphins. Sci. Rep. Whales Res. Inst. Tokyo 32:39-66. 1980b. Analysis of growth layers in teeth of Tursiops trun- catus, using light microscopy, microradiography, and SEM. In W. F. Perrin and A. C. Myrick, Jr. (editors), Age determination of toothed whales and sirenians, p. 155- 160. Rep. Int. Whaling Comm. Spec. Issue 3. HUI, C. A. 1978. Reliability of using dentin layers for age determination in Tursiops truncatus. U.S. Marine Mammal Comm. Rep. No. MMC-77/09 (Nat. Tech. Inf. Serv. PB288444), 25 P.o JONSGARD, A 1969. Age determination of marine mammals. In H. T. Anderson (editor), The biology of marine mammals, p. 1- 30. Acad. Press, N.Y. Kasuya, T. 1976. Reconsideration of life history parameters of the spot- ted and striped dolphins based on cemental layers. Sci. Rep. Whales Res. Inst. 28:73-106. 1977. Age determination and growth of Baird's beaked whale with a comment on the fetal growth rate. Sci. Rep. Whales Res. Inst. Tokyo 29:1-20. Kasuya, T., and M. Nishiwaki. 1978. On the age characteristics and anatomy of the tusk of Dugong dugon. Sci. Rep. Whales Res. Inst. Tokyo 30:301- 311. Klevezal, G. A. 1980. Layers in the hard tissues of mammals as a record of growth rhythms of individuals. In W. F. Perrin and A. C. Myrick, Jr. (editors) , Age determination of toothed whales and sirenians, p. 89-94. Rep. Int. Whaling Comm. Spec. Issue 3. Klevezal', G. A., and S. E. Kleinenberg. 1967. Opredelenie vozrasta mlekopitayushchikh po sloityn strukturam zubov i kosti (Age determination of mammals by layered structure in teeth and bone). Izdatel'stvo Nauka, Moscow, 144 p. Laws, R. M. 1952. A new method of age determination in mammals with special reference to the elephant seal (Mirounga leonina Linn.). Falkland Is. Dep. Surv. Sci. Rep. 2, 11 p. 1962. Age determination of pinnipeds with special reference to growth layers in the teeth. Saugetier. Mitt. 27(3):129- 146. Marsh, H. 1980. Age determination of the dugong {Dugong dugon (Muller)) in northern Australia. In W. F. Perrin and A. C. Myrick, Jr. (editors), Age determination of toothed whales and sirenians, p. 181-201. Rep. Int. Whaling Comm. Spec. Issue 3. Myrick, A. C, Jr. 1979. Variation, taphonomy, and adaptation of the Rhabdos- teidae (=Eurhinodelphidae) (Odontoceti, mammalia) from the Calvert formation of Maryland and Vir- ginia. Ph.D. Thesis, Univ. California Los Ang., 41 1 p. 1 980a. Some approaches to calibration of age in odontocetes using layered hard tissues. In W. F. Perrin and A. C. Myrick, Jr. (editors) , Age determination of toothed whales and sirenians, p. 95-97. Rep. Int. Whaling Comm. Spec. Issue 3. 1980b. Examination of layered tissues of odontocetes for age determination using polarized light microscopy. In W. F. Perrin and A. C. Myrick, Jr. (editors), Age determination of toothed whales and sirenians, p. 105-112. Rep. Int. Whaling Comm. Spec. Issue 3. Myrick, A. C, Jr., A. A. Hohn, P. A. Sloan, M. Kimura, and D. D. Stanley. 1983. Estimating age of spotted and spinner dolphins (Stenella attenuate and Stenella longirostris) from teeth. NOAA-TM-NMFS-SWFC-30, 17 p. Myrick, A. C, Jr., E. W. Shallenberger, and I. Kang. In press. Records used in the calibration of dental layers in seven captive Hawaiian spinner dolphins, Stenella longirostris. NOAA-TM-NMFS-SWFC. Nielsen, H. G. 1972. Age determination of the harbour porpoise Phocoena phocoena (L.) (Cetacea). Vidensk. Medd. Dan. Naturhist. Foren. 135:61-84. Nishiwaki, M., andT. Yagi. 1953. On the age and the growth of teeth in a dolphin, (Pro- delphinus caeruleo-albus). (I). Sci. Rep. Whales Res. Inst. Tokyo 8:133-146. NORRIS, K. S., AND T. P. DOHL. 1980. Behavior of the Hawaiian spinner dolphin, Stenella lon- girostris. Fish. Bull., U.S. 77:821-849. Perrin, W. F., D. B. Holts, and R. B. Miller. 1977. Growth and reproduction of the eastern spinner dolphin, a geographical form of Stenella longirostris in the eastern tropical Pacific. Fish. Bull, U.S. 75:725-750. Perrin, W. F., and A. C. Myrick, Jr. (editors). 1980. Age determination of toothed whales and sireni- ans. Rep. Int. Whaling Comm. Spec. Issue 3, 229 p. Ridgway, S. H., R. F. Green, and J. C. Sweeny. 1975. Mandibular anesthesia and tooth extraction in the bot- tlenose dolphin. J. Wildl. Dis. 11:415-418. SCHEFFER, V. B. 1950. Growth layers on the teeth of Pinnipedia as an indica- tion of age. Science (Wash., D.C.) 112:309-311. SCHEFFER, V. B., AND A. C. MYRICK, JR. 1980. A review of studies to 1 970 of growth layers in the teeth of marine mammals. In W. F. Perrin and A. C. Myrick, Jr. (editors), Age determination of toothed whales and sirenians, p. 51-63. Rep. Int. Whaling Comm. Spec. Issue 3. Sergeant, D. E. 1959. Age determination in odontocete whales from dentinal growth layers. Nor. Hvalfangst-tidende 48:273-288. Sergeant, D. E., D. K. Caldwell, and M. C. Caldwell. 1973. Age, growth, and maturity of bottlenosed dolphin (Tur- siops truncatus) from northeast Florida. J. Fish. Res. Board Can. 30:1009-1011. Wells, R. S. In press. Reproductive seasonality and social behavior of Hawaiian spinner dolphins, Stenella longirostris. In W. F. Perrin and D. DeMaster (editors), Cetacean reproduc- tion. Rep. Int. Whaling Comm. Spec. Issue 6. 225 REPRODUCTION OF THE BANDED DRUM, LARIMUS FASCIATUS, IN NORTH CAROLINA1 Steve W. Ross ABSTRACT The reproductive biology of Larimus fasciatus was examined in coastal North Carolina from September 1975 through September 1976. Spawning occurred in nearshore waters from April through Sept ember with a peak in August. Maturity in fameles was reached by the first year between 120 and 130 mm SL. Generally the larger, older fish matured earlier and also continued spawning later in the season than the younger ones. Fecundity ranged from 12,750 to 320,819 ova with first spawners preducted to have between 31 ,088 and 65,038 eggs. Fecundity was best predicted by ovary weights during August. Sex ratios generally favored more females. As fish grew the sex ratio changed from predominately males to predominately females. The banded drum, Larimus fasciatus Holbrook, occurs from Massachusetts to southeastern Florida and along the northern Gulf of Mexico from the Florida west coast to Mexico. Unlike other drums it appears to be largely restricted to nearshore coastal waters at all sizes and is rarely collected in estuaries or from the outer continental shelf (Gunter 1938; Dahlberg 1972; Chao 1978: Powles 1980). Larimus fasciatus is a small sciaenid reported by Holbrook (1860) to reach 305 mm TL (total length), but it seldom grows larger than 220 mm (Chao 1978). Its small size, low abundance, and lack of status as a food or game fish afford this species little commercial or recreational value, although it was reported as a com- ponent of the North Carolina (Wolff 1972) and Gulf of Mexico (Gutherz et al. 1975) industrial fisheries. Published data on life history aspects of L. fasciatus are largely lacking. Hildebrand and Cable (1934) reported limited information on spawning, growth, and juvenile descriptions of North Carolina specimens, and Powles (1980) presented data on lar- val description, spawning seasons, and areas in the South Atlantic Bight. Feeding habits were briefly examined by Welsh and Breder (1923) and Chao and Musick (1977). Standard and Chittenden (in press) have studied banded drum life history off of Texas. This study describes the following aspects of L. fas- ciatus life history in North Carolina: 1) spawning seasonality, 2) age and size at maturity, 3) fecundity, and 4) sex ratios. METHODS Most banded drum were collected in the ocean near the mouth of the Cape Fear River, N.C., about 4-6 km off Oak Island in depths of 4-14 m (Fig. 1). Bottom topography was uniform with sediments of fine sand and mud. Hydrographic conditions were heavily influenced by discharge from the Cape Fear River (Ross 1978). This area was sampled weekly from September 1975 through September 1976, except only monthly samples were made during January, June, July, and August. Each sample consisted of repetitive (4-12) 30-min trawls with a 12.4 m semiballoon otter trawl of 3.85 cm stretched mesh during daylight hours. Additional specimens were collected from Septem- ber 1975 through September 1976 during twice month- ly, daylight sampling between Beaufort Inlet and Cape Lookout, N.C. (Fig. 1), except that there was no sampling in December 1975 and only monthly sampling in January and February 1976. Repetitive trawls were made in this area in a depth range of 9-12 m over a flat, sand bottom using the aforementioned gear and tow times. Specimens were also collected near Cape Hatteras (9-17 m depth) in November and December 1975 and April 1976 by the North Car- olina Division of Marine Fisheries (Fig. 1). Larimus fasciatus were preserved in the field in 1 0% Formalin3 and later stored in 40% isopropanol. Total length (TL) and standard length (SL) were measured to the nearest mm. Body weights (BW) were deter- mined to the nearest 0.1 g, and gonads >0.01 g were 'Adapted from part of a thesis submitted to the Zoology Depart- ment, University of North Carolina, in partial fulfillment of the requirements for the MA degree. 2North Carolina Division of Marine Fisheries, P.O. Box 769. Morehead City, NC 28557. 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Manuscript accepted August 1983. FISHERY BULLETIN: VOL. 82. NO. 1, 1984. 227 FISHERY BULLETIN: VOL. 82, NO. 1 •ianUiSrt *y CaPe inlet .r Lookout ©-Larimus fasciatus collection sites 34 Cape Fear FIGURE 1. — Collection sites for Larimus fasciatus in North Carolina. blotted dry and weighed (gonad weight (GW)) to the nearest 0.01 g. Gonad indices (GI) were calculated as follows: GI = GW (both) X 100/(BW - GW) and were used to determine spawning seasons and maturity. Fecundity was determined for both maturing gonads by relating the number of eggs in a subsample to the whole gonad. Each subsample (weighed to the nearest 0.001 g), removed from the middle and both ends of each alcohol-preserved gonad, represented roughly 5% of the total gonad weight. All eggs (excluding those <0.01 mm in diameter and atretic eggs) in the subsample were counted, and the modal ovum diameter was measured to the nearest 0.05 mm. Total fecundity used in the analysis equaled the number of eggs in both gonads combined. RESULTS Spawning Larimus fasciatus spawned from April through Sep- tember with peak activity in August as indicated by female gonad indices (n = 126, Fig. 2). Male gonad indices (n = 53) somewhat mirrored the female pat- tern, but the spawning cycle was not clearly illus- trated because the testes composed a small per- centage of the body weight at any maturity stage in all months (Fig. 2). Some running ripe males were ob- served in the field from June through August. Since the mean gonad index was still high in September (Fig. 2), spawning may have continued after Septem- ber, although I have no collections to substantiate this. The large size range of juveniles and the collection of young-of-the-year <40 mm SL in all months ex- 228 ROSS: REPRODUCTION OF BANDED DRUM UJ Q 4.5 4. OH 3.5 3.0 §2.5 z O 2.0 C3 -Z- 1.5 < UJ 5 1.0 0.5 0 7 - P" N F M A MONTH M FIGURE 2.— Monthly mean gonad index of male and female banded drum from October 1975 to September 1976 in North Carolina, including sample size and ±1 standard error of the mean. cept December 1975 and January, April, May, and June 1976 (Fig. 3) support an extended late spring through early fall spawning season. Major young-of- the-year (1976 year class) recruitment, evidently from Spring spawning, first appeared in July 1976 and continued through September 1976. Young-of- the-year from the 1975 year class were evident from September 1975 through November 1975 and ap- peared again in February 1976 (Fig. 3). This young- of-the-year recruitment over a long period with a lack of bimodal length frequencies indicated sustained spawning effort. Other collections in and near the lower Cape Fear River of Larimus fasciatus <40 mm SL in January, February, April, June, July, Septem- ber, November, and December also indicated ex- tended spawning (K. A. MacPherson4). The majority of the reproductively active adults were collected near the Cape Lookout area (Fig. 1), especially during August and September where bot- tom water temperature averaged 27° (August) to 20°C (September). A high percentage (48.9-100%) of the total number of females collected in the Cape Lookout area exhibited maturing or ripe gonads while corresponding percentages from Cape Fear 4K. A. MacPherson, biologist, Carolina Power and Light Company, Brunswick Biological Laboratory, P.O. Box 10429, Southport, N.C. 28461, pers. commun. 1977. FIGURE 3. — Length frequencies of Larimus fasciatus collected in North Carolina from September 1975 through September 1976. SEP 1975 n=118 25 55 85 115 145 175 STANDARD LENGTH(mm) 229 FISHERY BULLETIN: VOL. 82, NO. 1 were low (0-8.1%) (Table 1). Although sampling effort in the Cape Fear area was half of that near Cape Lookout from June through August, more female banded drum were collected near Cape Fear; however, the percent of females with large gonads was much greater in the Cape Lookout area (Table 1 ) . Cape Fear area sampling effort doubled over that near Cape Lookout in September and yielded many more female banded drum, but only 0.7'r were re- productively active compared with 48.9% in the Cape Lookout area (Table 1). Irregular sampling from the Cape Hatteras area (Fig. 1) yielded matur- ing or ripe L. fasciatus only during April when 82.4% of the females collected had gonad indices between 1.7 and 6.1 (Table 1). Bottom water temperature in this area was 17°C. Ovum diameter is often an indication of sexual maturity (Higham and Nicholson 1964), and the relationship between egg size (OD) and gonad index (GI) for banded drum {n = 90) was OD - 0.34 + 0.1 1 (In GI), r = 0.77 (Fig. 4). This relationship is an objective, quantita- tive way to determine degree of maturity (Yuen 1955; Schaeferand Orange 1956) and was used to differen- tiate maturing from immature female banded drum. The point on the graph (Fig. 4) where gonad index began to increase more rapidly than egg size was used as the boundary between immature and maturing gonads and occurred around a gonad index of 1.0 and an ovum diameter of 0.35 mm. Mean ova diameters peaked from July through September at 0.48 mm (Table 2), which also coincided with the highest gonad indices. Maturity Female banded drum reached sexual maturity be- tween 120 and 130 mm SL (n = 112). All fish <120 mm SL were immature (GI <1.0) and 97% of those >130 mm were mature, with 607t between 120 and 130 mm reaching maturity (Table 3). During the spawning season, females between 120 and 130 mm indicated increased gonad activity. Females smaller than 120 mm displayed no seasonal gonad activity, while only three fish >130 mm were not maturing during the spawning season (Fig. 5). Only the larger adults >150 mm matured and spawned early (April), and generally a higher proportion of the older Table 1.— Percent of female Larimus fasciatus with gonad indices >1.0 and sample size collection area during the spawning months of 1976. TV) from each Area April May June July Aug. Sept. Total Cape Fear Cape Lookout Cape Hatteras Total 0 (274) 0 (1) 82.4 (17) 4.8 (292) 0(219) 0(219) 0 (9) 53.8(26) 40.0(35) 2.8 (36) 100 (12) 27 .1 (48) 8.1 (11 1) 75.0(28) 21.6(139) 0.7 (153) 48.9(45) 11 6(198) 1 4(802) 61 6(112) 82.4 (17) 10.1 (931) 0.6 0 5 £ E o o i ••• • ••« OD = 0.34 + 0.11ILnGII r = 0.77 n=90 3 4 GONAD INDEX FIGURE 4. — Relationship between famale gonad index and ova diameters of North Carolina Larimus fasciatus. 230 ROSS: REPRODUCTION OF BANDED DRUM TABLE 2. — Mean monthly ova diameters of Larimus fasciatus from March through September 1976. M •■in ova diameter Month (mm) Sa mple size March 0.01 1 April 0.41 7 June 0.46 16 July 0.48 13 August 0.48 32 September 0.48 21 females continued spawning later (September) (Fig. 5). Most of the smallest reproductively active females (between 120 and 130mmSL) matured from June to August (Fig. 5). Using age-length relationships of Ross (1978), Larimus fasciatus reached maturity shortly after turning 1-yr-old. They continued spawning through- out life until age 3, which was the maximum age encountered. Fecundity TABLE 3. — Number and percentage of mature and immature female banded drum by 10 mm size cate- gories off North Carolina, April-September 1976. Maturity was judged by gonad index (GI) value. Standard Immature Mature Percent length (mm) GI < 1 .0 GI > 1.0 mature <90 1 0 00 90-99 0 0 0.0 1 00- 1 09 4 0 0.0 110-119 4 0 0.0 120-129 6 9 60.0 130-139 1 12 92.3 140-149 1 22 95.7 150-159 0 17 100.0 160-169 1 26 963 170-179 0 7 100.0 180-189 0 1 100.0 Total 18 94 130 mm SL and predicted fecundity in this size range is 31,088-65,038 ova. Body weight (BW) minus the gonad weight (GW) was regressed onto fecundity yielding the equation: Number of ova increased with increasing fish size, ranging from 12,750 ova in a 118 mm SL female to 320,819 in a 179 mm female. The relationship be- tween fecundity (F) and SL for 86 females was linear and expressed by the equation: F = -376,312 + 3,395 (SL), r = 0.76 (Fig. 6). Length at first spawning is between 120 and F = -52,741 + 1,887 (BW), r = 0.76, n = 85. Gonad weight varies seasonally and is closely related to fecundity; therefore, eliminating it from body weight reduced the possibility of autocorrelation. Even without the gonad weight, body weight varies seasonally and to some extent daily as a function of diet; therefore, body weight is not the best predictor of fecundity. The fecundity to ovary weight (0 W) 7- 6- 5- X LU Q 4- Q I3 O o 1- 67634^ *** 88 8 4 9 9 8 8 7 6 4 8 8 74 6 9 99 8 8 8 '788 8 8 9 6 8 4 9 9 8 6 «8> 9 6 9 89 999 998 9 4 « 6 ^ ' 6 97 7 6 6 6 6 6 6 7 8 4 4 4 — r~ 90 100 110 120 130 140 150 160 170 180 STANDARD LENGTH(mm) FIGURE 5. — Relationship between famale gonad index and standard length by month for banded drum during January (l)-September (9) 1976 (n = 124). 231 FISHERY BULLETIN: VOL. 82, NO. 1 35- F = -376,312 + 3395ISL] r = 0.76 • 30- n : 86 * • / 25- ^^ • • / • "<* • / O • / --20- / X • / ' — • y > • • * • t 15- • / • •• • •• • • • • Q • • /• • 2 • Am* • • •* • z> • / • Q10- • • /• •• • • ••• • • /•• • • LLI U_ * 4 • • 5 - t '• / • • * • • • • o- 1 1 1 T— 100 125 150 175 200 STANDARD LENGTH(mm) FIGURE 6. — Relationship of fecundity to standard length for banded drum collected in North Carolina from April through September 1976. relationship was expressed by F = 15,490 + 28,024 (OW), r = 0.94, n = 85 and had a much higher correlation coefficient than either the length or body weight regressions. To minimize monthly variation (Morse 1980) the most accurate prediction of fecundity was derived from ovary weights only from the peak spawning month, August, expressed by F = 18,532 + 28,181 (OW), r = 0.97, n = 31 (Fig. 7). Sex Ratios Sex was determined for 2,729 banded drum and the overall ratio of males to females varied significantly from 1:1 in favor of females (Table 4). This non- homogeneity of total sex ratios could not be account- ed for by any consistent pattern of seasonal ratio differences. The two largest size groups exhibited sex ratios significantly in favor of females. The disparity between sexes in the size range 100-139 mm SL was accounted for during winter, spring, and summer, while that in the fish > 140 mm SL was accounted for during fall and winter (Table 4). Contingency table analysis indicated strong dependency between sex and size group (x2 = 17.84, df — 3,P < 0.001), even though differences in the smallest two size groups were nonsignificant (Table 4). As fish grew, the population shifted from more males to more females. There were more total females than males in all seasons except summer; however, the differences were only significant in the fall. The fall divergence from a 1:1 ratio was explained by differences in the 60-99 mm and >140 mm SL size groups (Table 4). 33- 30- 27- o 24- 21- X 18- > Q 15- 12 D O 9- LU Ll_ 6 3- 0- F = 18,532+ 28,181I0WI r =0.97 n = 31 -i 1 1 1 1 I r- 2 3 4 5 6 7 8 OVARY WEIGHT (g) 10 11 FIGURE 7. — Relationship of fecundity to overy weight during August 1976 for North Car- olina banded drum. 232 ROSS: REPRODUCTION OF BANDED DRUM Table 4. — Larimus fasciatus male/female sex ratios by season and size group from North Carolina (September 1975-September 1976) with chi-square values from testing a 1:1 ratio. Season Size group Fall Winter Spring Summer (mm SL) (Sept. -Nov.) (Dec. -Feb.) (Mar. -May) (June-Aug.) Total df X1 <59 151/144 22/27 104/100 65/34 342/305 3 5.23 60-99 77/103 15/12 492/479 65/64 649/658 3 2.14 100-139 71/67 19/34 87/123 48/74 225/298 3 8.04' >140 21/64 14/23 19/22 49/50 103/149 3 8.47' Total 320/268 70/96 702/724 227/222 1.319/1.410 df 3 3 3 3 X1 928* 3.64 7.28 763 10.38* 'P < 0 05 DISCUSSION The prolonged April-September spawning season of L. fasciatus in this study is supported by the few published references to its reproduction. From analysis of larval occurrence in North Carolina, Hildebrand and Cable (1934) proposed a May through October spawning season. Powles (1980) reported a May to October spawning in the South Atlantic Bight also based on larval collections. Gun- ter (1938) suggested April spawning for banded drum in Louisiana. Standard and Chittenden (in press) found two spawning peaks forL. fasciatus off Texas, a minor one in the spring (April-June) and the major one in the fall (September-November). They did not find significant evidence of spawning in July or August. My data suggested a prolonged spawning effort in North Carolina beginning as early as April, peaking in August, and possibly continuing after September. This major departure from Standard and Chitten- den's (in press) biomodel spawning was supported by 1) a steady increase in gonad indices with a single August peak, 2) a single peak mode of ova diameters of 0.48 mm from July through September, 3) con- tinuous recruitment of young-of-the-year through the summer and fall months, and 4) the collection of larvae in all months except March (Powles 1980; K. A. MacPherson footnote 4). Although it is fairly cer- tain that spawning begins in April, at least for larger fish, I did not determine if spawning continued into October because samples of adults were lacking. Although the September gonad index declined, young-of-the-year recruitment in North Carolina in February and larval collections in November, December, January, and February (K. A. MacPher- son footnote 4) indicated that spawning may last at least through October. Protracted spawning is also characteristic of many other Sciaenidae (Welsh and Breder 1923; Thomas 1971; Merriner 1976; Warlen 1980). Maturation at an early age is typical in sciaenids (Schaefer 1965; Meriner 1976; Shlossman and Chit- tenden 1981) and in short-lived fishes in general which tend toward r strategy life histories (Adams 1980). Since L. fasciatus is a short-lived sciaenid, rarely completing a fourth year, the small size (120 mm SL) at first maturity, attained shortly after reach- ing 1 yr of age, is not surprising (Ross 1978). Larimus fasciatus off of Texas apparently live only 2 yr and consequently mature earlier (80 mm TL) than North Carolina individuals (Standard and Chittenden in press) . In addition to short life and early maturation, r strategists' traits are rapid growth, high fecundity (even at early ages), small maximum size, high mor- tality, and low maximum age (Adams 1980), all of which are related to emphasizing reproductive pro- ductivity. Banded drum have all of these characteris- tics as indicated in this study and by Ross (1978) and Standard and Chittenden (in press). As banded drum became older their growth rate slows (Ross 1978; Standard and Chittenden in press), as is typical of most fishes, and they can devote relatively more energy toward reproductive activity than at earlier ages. Only the largest fe- males (>150 mm) appeared to spawn as early as April and continue spawning into September. Although the phenomenon of older fish having a lon- ger spawning season has not been reported in United States east or gulf coasts sciaenids, it does occur in other fishes (Quast 1968; Grimes and Huntsman 1980). Larimus fasciatus spawns as far north as Cape Hat- teras. Although larvae have been collected off Chesapeake Bay (Berrien et al. 1978), there are no records of reproductively active adults north of Cape Hatteras and this species is rare north of Chesapeake Bay (Hildebrand and Schroeder 1928; Johnson 1978); therefore, Cape Hatteras is probably the northern limit of banded drum reproduction. Larimus fasciatus in spawning condition were most often collected in the nearshore waters between 233 FISHERY BULLETIN: VOL. 82, NO. 1 Beaufort Inlet and Cape Lookout, larval dis- tributions have not clarified the preferred spawning depth range, since larvae have been collected over a wide range of the continental shelf (Berrien et al. 1978; Powles 1980); there is, however, some ten- dency toward increased abundance over the inner shelf (Powles 1980). Miller et al. (in press) suggested that onshore transport by currents into estuarine nurseries of offshore spawned larvae is most favor- able during the winter off North Carolina south of Cape Hatteras. Several winter spawners with estuarine dependent young spawn along the outer continental shelf {Leiostomus xanthurus, Dawson 1958; Mugil cephalus, Anderson 1958; Breuoortia tyrannus, Nelson et al. 1977; Micropogonias undulatus, Warlen 1980); thus, the young could take advantage of the inshore directed currents. A cor- ollary to this theory indicates that summer spawners should reproduce near shore or in the estuary if lar- vae are to be retained in the more productive shallow waters because net current movement is offshore (Miller et al. in press). In addition to L. fasciatus, other fishes also spawn in nearshore or estuarine waters south of Cape Hatteras during the summer (Cynoscion regalis, Merriner 1976; C. nebulosus, Mahood 1975; Stellifer lanceolatus and Bairdiella chrysoura, Powles 1980). ACKNOWLEDGMENTS I am especially grateful to Sheryan P. Epperly for her time and statistical advice. I also thank D. R. Colby, G. W. Link, K. A. MacPherson, and F. C. Rohde for advice and field assistance. J. W. Gillikin, B. F. Holland, S. G. Keefe, J. B. Sullivan, and J. Vaughn made special efforts to collect specimens for this study. C. S. Manooch III and C. B. Grimes con- tributed much through review of a preliminary manuscript. I wish to acknowledge my thesis commit- tee, F. J. Schwartz, E. A. McMahan, and A. F. Chestnut, for their support. Major support for this project was through a grant from Carolina Power and Light Company. LITERATURE CITED Adams, P. R. 1980. Life history patterns in marine fishes and their conse- quences for fisheries management. Fish. Bull., U.S. 78:1- 12. Anderson, W. W. 1958. Larval development, growth, and spawning of striped mullet (Mu?>il cephalus) along the South Atlantic coast of the United States. U.S. Fish Wildl. Serv., Fish. Bull. 58:501-519. Berrien, P. L., M. P. Fahay, A. W. Kendall, Jr., and W. G. Smith. 1978. Ichthyoplankton from the RV Dolphin survey of con- tinental shelf waters between Martha's Vineyard, Massachusetts and Cape Lookout, North Carolina, 1965- 66. U.S. Dep. Commer., NOAA, Sandy Hook Lab. Tech. Serv. Rep. 15, 152 p. CHAO, L. N. 1978. Family: Sciaenidae.Lon>m/.s fasciatus, Holbrook, 1860. In W. Fischer (editor), FAO species identification sheets for fishery purposes, western central Atlantic (fish- ing area 31), Vol. 4, unpaged. FAO, Rome. CHAO, L. N., AND J. A. MUSICK. 1977. Life history, feeding habits, and functional morphology of juvenile sciaenid fishes in the York River estuary, Virginia. Fish. Bull., U.S. 75:657-702. Dahlberg, M. D. 1972. An ecological study of Georgia coastal fishes. Fish. Bull., U.S. 70:323-353. Dawson, C. E. 1958. Study of the biology and life history of the spot, Leio- stomus xanthrus Lacepede, with special reference to Smith Carolina. Contrib. Bears Bluff Lab, 28, 48 p. Grimes, C. B., and G. R. Huntsman. 1980. Reproductive biology of the vermilion snapper, Rhnm- boplites aurorubens, from North Carolina and South Car- olina. Fish. Bull., U.S. 78:137-146. GUNTER, G. 1938. Seasonal variations in abundance of certain estuarine and marine fishes in Louisiana, with particular reference to life histories. Ecol. Monogr. 8:313-346. Gutherz, E. J., G. M. Russell, A. F. Serra, and B. A. Rohr. 1975. Synopsis of the northern Gulf of Mexico industrial and foodfish industries. Mar. Fish. Rev. 37(7): 1-1 1. Higham, J. R., and W. R. Nicholson. 1964. Sexual maturation and spawning of Atlantic men- haden. U.S. Fish Wildl. Serv., Fish. Bull. 63:255-271. Hildebrand, S. F„ and L. E. Cable. 1934. Reproduction and development of whitings or kingfish, drums, spot, croaker, and weakfishes or seatrouts, family Sciaenidae, of the Atlantic coast of the United States. Bull. [U.S.| Bur. Fish. 48:41-117. Hildebrand, S. F., and W. C. Schroeder. 1928. Fishes of Chesapeake Bay. Bull. U.S. Bur. Fish. 43(1), 366 p. Holbrook,.]. E. 1860. Ichthyology of South Carolina. 2d ed. Russell and Jones, Charleston, S.C., 205 p. Johnson, G. D. 1978. Development of fishes of the mid-Atlantic Bight: an atlas of egg, larval and juvenile states, Vol IV. Carangidae through Ephippidae. U.S. Fish Wildl Serv. Biol. Serv. Program FWS/OBS-78/12, 314 p. Mahood, R. K. 1975. Spotted seatrout in coastal waters of Georgia. Proc. 29th Annu. Conf. Southeast. Assoc. Game Fish Comm., p. 195-207. Merriner, J. V. 1976. Aspects of the reproductive biology of the weakfish, Cynoscion regalis (Sciaenidae), in North Carolina. Fish. Bull., U.S. 75:18-26. Miller, J. M., J. P. Reed, and L. T. Pietrafesa. In press. Patterns, mechanisms, and approaches to the study of migrations of estuarine dependent fish larvae and juveniles. In J. D. McCleave (editor), Mechanisms of migration in fishes. NATO Adv. Res. Inst, 13-17 Dec. 234 ROSS: REPRODUCTION OF BANDKI) DRl M 1982. Acquafredda, Italy. Morse. W. W. 1980. Maturity, spawning, and fecundity of Atlantic croaker, Micropogonias undulatus, occurring north of Cape Hat- teras. North Carolina. Fish. Hull., U.S. 78:190-1 Of,. Nelson, W. R.. M. C. Ingham, and W. E. Schaaf. 1977. Larval transport and year-class strength of Atlantic menhaden. Brevoortia tyrannus. Fish. Hull., U.S. 75:23- 41. POWLES, H. 1980. Descriptions of larval silver perch, Bairdiella chrysoura, banded drum, Larimus fasciatus, and star drum. Stellifer lanceolatus (Sciaenidae). Fish. Hull., U.S. 78:119-136. Ql AST, J. C. 1968. Observations on the food and biology of the kelp bass, Paralobrax clathratus with notes on its sportfishery at San Diego. California. In W. J. North and C. I.. Hubbs (com- pilers and editors), Utilization of kelp-bed resources in southern California, p. 81-108. Calif. Fish Game, Fish Bull. 139. Ross. S. W. 1978. The life history of the banded drum, Larimus fasciatus, in North Carolina waters. M.A. Thesis, Univ. North Car- olina, Chapel Hill, 143 p. Schaefer. m. b.. and C. .J. Oranok. 1956. Studies of the sexual development and spawning of yellowfin tuna (Neothunnus macropterus) and skipjack {Katsuwanus pelamis) in three areas of the eastern Pacific Ocean, by examination of gonads. [In Engl, and Span.] Inter-Am. Trop. tuna Comm., Bull. 1:281-349. Schaefer, R. H. 1965. Age and growth of the northern kingfish in New York waters. N.Y. Fish. Game J. 12:191-216. Sholssman, P. A., and M. E. Chittenden, Jr. 1981. Reproduction, movements, and population dynamics of the sand seatrout, Cynoscion arenarius. Fish. Bull.. U.S. 79:649-669 Standard, G. W., and M. E. Chittenden, Jr. In press. Reproduction, movements, and population dy- namics of the banded drum,Larimus fasciatus, in the Gulf of Mexico. Fish. Bull., IS Thomas, D. L. 1971. The early life history and ecology of six species of drum (Sciaenidae) in the lower Delaware River, a brackish tidal estuary. Ichthyol. Assoc. Bull. 3, Part III, 247 p. WARLEN, S. M. 1980. Age and growth of larvae and spawning time of Atlantic croaker in North Carolina. Proc. Annu. Conf. Southeast. Assoc. Fish Wildl. Agencies 34:204-214. Welsh, W. W., andC. M. Breder. L923. Contributions to life histories of Sciaenidae of the east- ern United States coast. Hull. U.S. Bur. Fish. 39:141- 201. Wolff, M. 1972. A study of North Carolina scrap fishery. N.C. Div. Comm. Sport Fish., Dep. Nat. Econ. Resour., Spec. Sci. Rep. 20, 29 p. Yuen, H. S. H. 1955. Maturity and fecundity of bigeye tuna in the Pacific. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 150, 30 p. 235 NOTES MARKING GROWTH INCREMENTS IN OTOLITHS OF LARVAL AND JUVENILE FISH BY IMMERSION IN TETRACYCLINE TO EXAMINE THE RATE OF INCREMENT FORMATION Age determination of fishes by counting daily growth increments in their otoliths is becoming a widely used technique in growth and population studies. Daily formation of otolith increments was first reported by Pannella (1971) for three species of temperate fish. Since then a number of workers, using three basic techniques for confirming the periodicity of incre- ment formation, have reported the presence of daily increments in larval or adult otoliths of at least 15 species of marine and freshwater fishes. Laboratory rearing from eggs to larvae of known age was used to confirm daily increments by brothers et al. (1976), Taubert and Coble (1977), Barkman (1978), Tanaka etal. (1981), and Laroche et al. (1982). The change in the mean number of increments over time in fish cap- tured in the wild and held in captivity was used to validate daily increments by Struhsaker and Uchiyama (1976), Wilson and Larkin (1980), and Uchiyama and Struhsaker (1981). The third method makes use of chemical agents to mark the growing margin of calcified structures in order to examine their rate of growth (Harris 1960). Tetracycline is one of the best chemical markers because it is relatively nontoxic and produces a fluorescent mark which is easily viewed in ultraviolet light (Harris 1 960; Weber and Ridgway 1962). It has been administered to fish by feeding (Choate 1964; Weber and Ridgway 1967; Trojnar 1973; Odense and Logan 1974) and by injec- tion (Kobayashi et al. 1964 and others below). Tet- racycline has been used in two studies to determine the rate of increment formation in otoliths. Wild and Foreman (1980) injected the drug into large juveniles and adult skipjack tuna, Kotsuwonus pelamis, and yellowfin tuna, Thunnus albacares, in a mark- recapture program in the tropical eastern Pacific. They found that otoliths of yellowfin tuna of 40-110 cm FL showed daily average increment formation, but that skipjack tuna of 42-64 cm FL showed <1 increment/d. Campana and Neilson (1982) injected tetracycline into juvenile starry flounders, Platich- thys stellatus, and found that daily increments were subsequently produced in both field and laboratory conditions. These authors briefly mentioned obtain- ing similar marking results by immersion, but did not detail their procedure. This paper presents a technique for marking otolith increments by immersing larval and juvenile fish in a solution of tetracycline in seawater, and reports the rate of increment formation under laboratory con- ditions for two species from the Great Barrier Reef, Australia: Hypoatherina tropicalis (Altherinidae) and Spratelloides dellicatulus (Dussumeriidae). Materials and Methods The experiments were conducted between July 1980 and February 1982 at One Tree Island Field Station and Lizard Island Research Station, during a field study of the population dynamics of these species. Achromycin (a brand of tetracycline HC11) was used in all experiments. The concentration that would mark the otoliths but not kill the fish was determined by testing three concentrations (400 mg/1, 250 mg/1, and 40 mg/1) using//, tropicalis from 12.8 to 23.0 mm SL. The otoliths of survivors were compared with untreated specimens to assess the effectiveness of the mark. The appropriate concentration, 250 mg/1, was then used in a series of similar experiments to determine the rate of increment formation (Table 1). The experiment number (I-IV) designates a group offish collected at the same time. In each experiment, fish were killed at two different times, designated as A or B, to compare the number of increments in fish held for different time periods. In experiment IV, the treatment times also differed, but in all other experiments the treatment time was the same for both groups A and B. Both species are small (adults <7 cm SL), mid- water, reef-associated, schooling fishes which do not undergo a marked metamorphosis between larval and juvenile stages (pers. observ.). Both attain their full complement of fin elements and begin to form scales and adult pigmentation at a standard length of 17-19 mm. Following the convention of Ahlstrom (1968), I consider this to be the size at which larvae become juveniles. Hypoatherina tropicalis used in the rate-determination experiments ranged from 12.8 to 27.2 mm SL, with 10 of 21 fish classed as larvae (<17.0 mm SL). Spratelloides delicatulus ranged from 15.5 to 22.9 mm SL, with 2 of 29 being larvae (Table 1). 'Manufactured by Lederle Labs, a division of Cyanamid Australia Pty. Ltd. References to trade names do not imply endorsement by the National Marine Fisheries Service, NOAA. FISHERY Bl'LLKTIN: VOL. 82, NO. 1, 1984. 237 The fish were collected at night with a light and a dip net, and placed in 25 1 aquaria without aeration or running seawater as soon as possible after collection. The aquaria were located outdoors under an awning, and therefore were exposed to the ambient diel light cycle, but not to direct sunlight. The fish were allowed to acclimatize for 12-24 h before treatment. Usually there was mortality during this period, but the proportion was not determined. All dead fish were removed prior to treatment. The fish were exposed to 250 mg tetracycline/1 seawater for 12 h from sunset to sunrise, except in experiment IVB when the immersion period was from sunrise to sunset (Table 1). After an immersion period, the aquarium was flushed with 90% water changes until no visible color remained. The tetracycline-seawater solution is yellow until ex- posed to sunlight for more than ~3 h, when it turns pink, due to oxidative photolysis. Following the treat- ment, fish were maintained in clean seawater for 2-6 d by feeding either fresh wild plankton > 125 mm diameter once a day (experiment I) or Artemia salina nauplii 3-4 times/d (all subsequent experiments). Artemia nauplii were more convenient for frequent feedings than fresh wild plankton. Ninety percent of the water in each aquarium was changed each morn- ing by siphoning, to minimize handling the fish. Tank water temperatures were measured over the diel cycle during February 1982 (summer) at One Tree Island. The temperature ranged from 25°C at 0630 h to 30°C at 1800 h. Replacement water, added at 0700 h from the surface of the lagoon, measured 27°C. Larvae were killed at the end of each experiment by placing them into 707c ethanol. Fish were subsequent- ly measured to the nearest 0. 1 mm SL. Their otoliths (both sagittae and lapilli) were removed and mounted whole on glass slides without coverslips, using Protexx. The following terms are used in this report for the concentric rings seen in otoliths. A growth zone is a wide ring which appears light or hyaline under transmitted light. A discontinuous zone is the nar- rower ring between two growth zones, often called the opaque zone because it appears dark under trans- mitted light. A growth increment, or simply an incre- ment, is a growth zone plus a discontinuous zone. Otoliths were examined at 250-l,000X magnifica- tion with a combination of incident ultraviolet light to reveal the fluorescent tetracycline-marked rings, and polarized transmitted light to count the rings. The fluorescence microscope used ultraviolet light from a 50W mercury lamp. Excitation wavelength was limited by a band pass filter (450-490 nm) and a long pass suppression filter (515 nm). In most cases, one sagitta from each fish was read, although occasionally the lapillus was used if its rings were clearer. The area to be counted was selected by scanning the margin of each otolith to find the place where the greatest number of distinct rings could be seen between the innermost fluorescent increment and the edge. A datum was considered valid only if identical counts were obtained in at least two out of three blind readings. No other otoliths were con- sidered in the analysis. Of 21 H. tropicalis otoliths Table 1. —Summary of tetracyeline-marking experiments to determine the rate of increment formation in H. tropicalis and S. delicatulus. Experiment No. of fish with various Standard length Predicted no deviations from the (mm) Treatment Date and of discontin- predicted number N Mean (range) period time of killing uous zones . Hypoathenna tropicalis IA IB MA IIB IMA Total 2(1) 4 6(21) 6 3 21|21) Spratelloides delicatulus IIIA 6 1MB IVA IVB Total 5(22) 9 9(21) 29(23) 14.0 (13.6-14.4) 13.7 (12.8-14.7) 20.5 (16.2-27.2) 18.9(16.8-20.7) 16.1 (15.4-17.2) 17.5 (15.5-19 .1) 17 9(17.6-18.2) 19 9(18.8-22.8) 20 5 (17.9-22 9) 2130, 8 July to 0830. 9 July 1980 2130, 8 July to 0830. 9 July 1980 1830. 31 Oct. to 0630. 1 Nov. 1980 1830. 31 Oct. to 0630, 1 Nov 1980 2000, 6 Nov to 0700. 7 Nov 1981 2000. 6 Nov to 0700. 7 Nov. 1981 2000, 6 Nov. to 0700. 7 Nov. 1981 1800. 31 Jan to 0630. 1 Feb. 1982 0600 to 1 800 31 Jun 1982 0830, 12 July 1730, 14 July 0730, 6 Nov. 0600, 7 Nov. 0545, 12 Nov. 0545, 12 Nov. 1800, 9 Nov. 1 800. 6 Feb. 0715, 6 Feb 2+1 5 4+1 5+1 4+1 4+1 2 5 4+1 1 2 5 6 3 0 17 5 3 2 6 7 12 'Otoliths of two treated fish were destroyed by poor preservation. 2Number of fish discarded because of inconsistency between otolith readings. 238 examined, 1 (4.8%) was discarded. Of 29 S. delicatulus, 3 (10.3%) were discarded (Table 1). Results and Discussion Marking Technique In the experiment to determine an effective tetracycline-marking concentration, all fish (n = 17) in 400 mg/1 died during the 12-h immersion period. Of 10 fish treated with 250 mg/1, 1 died during treatment, and 1 died during the subsequent holding period. Of 10 fish treated witlv50 mg/1, 1 died dur- ing treatment. Otoliths of untreated specimens showed faint fluorescence around the edge and occasionally along cracks and surface irregularities (Fig. 1A); this is a naturally occurring autofluorescence (Campana and Neilson 1982). Otoliths of fish in 50 mg/1 were indis- tinguishable from those of untreated specimens. Otoliths of fish in 250 mg/1 showed a strong fluores- cent band medial to the edge, in addition to the weak fluorescence at the edge (Fig. IB, C). This strong band consisted of two growth zones and one discon- tinuous zone (Fig. 2). It is not known how long it takes for tetracycline to be incorporated into the growing otoliths when administered by immersion. Campana and Neilson (1982) reported that after injection, 50% of fish showed fluorescent otoliths after 10 h, and 100% after 24 h. If one assumes similar or slightly longer incorporation times in the present study, then the inner fluorescent growth zone was probably formed the day after the immersion period. The subsequent discontinuous zone and growth zone were formed while there was residual tetracycline in the water or fish. Another possible explanation is that the appearance of fluorescence in two growth zones is an artifact of viewing whole otoliths. The results of this experiment indicate that immer- sion in a concentration of 250 mg Achromycin/1 of seawater for 12 h is adequate to mark one or more growth increments in//, tropicalis and S. delicatulus larvae and juveniles. The overall mortality rate in experiments I, II, and III (total n = 37), was 5.4% during treatment and 2.7% during the holding phase. To determine whether fluorescent marking would occur if the tetracycline immersion period was during daylight hours, an experiment was conducted using S. delicatulus from 17.9 to 22.9 mm SL (experiment IV). The fish were collected and divided between two tanks. One tank received tetracycline from 1800 h to 0630 h, the other from 0600 to 1800 h. Mortality due to treatment was not monitored. After 6 d, the fish FIGURE 1. — Flourescence photomicrographs of sagittae of larval Hypaatherina tropicalis. A. Untreated otoliths, showing autofluores- cence around the edge (10.1 mm SL). B. Tetracycline-marked oto- lith, showing fluorescent band medial to the edge (16.2 mm SL). C. Marked otolith under higher magnificaton (17.6 mm SL). were killed and examined. The fluorescent bands medial to the edges were similar in width and inten- sity to those in previous experiments, and showed no difference between the two treatments. This indicates that tetracycline is incorporated into grow- ing otoliths and produces fluorescent increments equally well during the day and night, regardless of whether the solution is yellow or has oxidized to pink. 239 FIGURE 2.— Tetracycline-marked otolith from H. tropicalis (17.6 mm SL), photographed with a combination of fluorescent and trans- mitted polarized light. Arrows indicate the fluorescent band pro- duced by the marking technique. This individual is from experiment IIB, and shows six discontinuous zones between the innermost fluorescent growth zone and the edge. The edge appears to be a growth zone. In summary, tetracycline can be administered by three techniques: feeding, injection, and immersion. Feeding has apparently not been used in otolith studies. The immersion technique presented here has advantages over injection in some situations. It can be used on fish which are too small or fragile for inj ection. The fluorescent mark obtained is relatively narrow, covering only two increments, compared with the wider mark resulting from injection (Kobayashi et al. 1964; Campana and Neilson 1982). Therefore, it is distinguishable from edge autofluores- cence after a shorter period of time, and allows finer resolution of increment formation, which may be use- ful in some experimental situations. Also, immersion requires minimum equipment, facilities, and han- dling of fish. Rate of Increment Formation In interpreting the results of my experiments, the number of discontinuous zones between the inner- most fluorescent growth zone and the edge was com- pared with the number predicted if one dis- continuous zone formed every day from ca. 0700 to 1000 h. Tanaka et al. (1981) found that growth zones in juvenile Tilapia nilotica held under various photoperiods started forming a few hours after lights-on, continued through the dark period, and stopped or slowed down about the time of the follow- ing lights-on. The discontinuous zone was formed in the few hours after lights-on. Mugiya et al. (1981) demonstrated that the deposition of calcium in goldfish, Carassius auratus, slowed down or stopped at sunrise and resumed in 3 h. Since otoliths are made of a matrix of organic fibers, which are calcified in the growth zones and not calcified in the discontinuous zones (Panella 1980; Watabe et al. 1982), the find- ings of Mugiya et al. (1981) support Tanaka et al. (1981). Whether this rhythm of increment formation is found in most fish remains to be investigated. The results for all experiments are presented in Table 1. For fish that were killed between 0545 and 0830 h, the predicted number includes an additional discontinuous zone that should have been forming at the time of death, although this ring was probably not always sufficiently formed to be counted. In these cases, an otolith was considered to show daily incre- ment formation even if the number of discontinuous zones was one less than predicted. One growth increment was formed each day in 85% of H. tropicalis (n — 20); the rest had one more than the predicted number of increments. In S. delicatulus, 46% (n = 26) showed daily formation of growth increments; 27% showed one less, and 27% showed one more, than expected if increments form daily. Thus, the variability in rate of increment for- mation was greater in S. delicatulus than in H. tropicalis, but the average rate for S. delicatulus was still 1 increment/d. This apparent difference in the rate of increment formation between species may be partially due to a difference between larvae and juveniles. Almost all (93%) of the S. delicatulus treated were juveniles, but only about half (52%) of the H. tropicalis were juveniles. However, no conclusion can be drawn from these data because the experiments were not designed to examine this factor, and the numbers are too small to compare larvae with juveniles. It is possible that tetracycline may affect the rate of increment formation. Some workers have reported that tetracycline inhibits mineralization in scales and bone (Harris 1960; Kobayashi et al. 1964), although others note neither growth promotion, retardation, nor structural weakness in bone as a result of tet- racycline administration (Weber and Ridgway 1967). The possibility that the tetracycline treatment inter- feres with growth of otoliths or fish was not con- sidered in this study, but should be examined before further use is made of this technique. In conclusion, the rate of increment formation has been examined in only a small number of species under a limited range of conditions. Recent evidence suggests that increment formation may be affected in some species by temperature, food availability and feeding frequency, photoperiod, and developmental stage (Taubert and Coble 1976; Brothers 1978; Pan- nella 1980; Wild and Foreman 1980; Geffen 1982; 240 Lough et al. 1982; Neilson and Geen 1982). It is therefore desirable to examine the rate of increment formation under various conditions before using otoliths for age determination (Brothers 1979). The technique presented here is a tool for studying incre- ment formation in otoliths of young fish under laboratory and possibly field conditions. It can be used for reef and nearshore benthic species which can be captured while larvae or juveniles and kept in containers or enclosures. Acknowledgments I want to thank Jeffrey M. Leis, Keith A. McGuin- ness, Richard Methot, Jr., Peter F. Sale, and two anonymous reviewers for their helpful comments on early drafts of the manuscript. I am grateful to J. M. Leis for the initial suggestion which led to this work, and to Peter Clarke for assistance with fluo- rescence microscopy. Literature Cited Ahlstrom, E. H. 1968. Reviews and Comments: Development of fishes of the Chesapeake Bay region, an atlas of egg, larval, and juvenile stages, Part 1. Copeia 1968:648-651. Barkman, R. C. 1978. The use of otolith growth rings to age young Atlantic silversides, Menidia menidia. Trans. Am. Fish. Soc. 107:790-792. Brothers, E. B. 1978. Exogenous factors and the formation of daily and sub- daily growth increments in fish otoliths. Am. Zool. 18:631. 1979. Age and growth studies on tropical fishes. In S. B. Saila and P. M. Roedel (editors), Stock assessment for tropical small-scale fisheries, p. 1 19-136. Proceedings of international workshop held Sept. 19-21, 1979, at Univer- sity of Rhode Island. (Copies available from Agency for International Development, AID-DIHF/ARDA, 7222 - 47th St., Chevey Chase, MD 20815.) Brothers, E. B., C. P. Mathews, and R. Lasker. 1976. Daily growth increments in otoliths from larval and adult fishes. Fish. Bull., U.S. 74:1-8. Campana, S. E., and J. D. Neilson. 1982. Daily growth increments in otoliths of starry flounder (Platichthys stellatus) and the influence of some environ- mental variables in their production. Can. J. Fish. Aquat. Sci. 39:937-942. Choate, J. 1964. Use of tetracycline drugs to mark advanced fry and fingerling brook trout (Salvelinus fontinalis). Trans. Am. Fish. Soc. 93:309-311. Geffen, A. J. 1982. Otolith ring deposition in relation to growth rate in her- ring (Clupea harengus) and turbot (Scophthalmus max- imus) larvae. Mar. Biol. (Berl.) 71:317-326. Harris, W. H. 1960. A microscopic method of determining rates of bone growth. Nature (Lond.) 188:1038-1039. KOBAYASHI, S., R. YUKI, T. FURUI, AND T. KOSUGIYAMA. 1964. Calcification in fish and shell-fish - I. Tetracycline labelling patterns on scale, centrum and otolith in young goldfish. Bull. Jpn. Soc. Sci. Fish. 30:6-13. Laroche, J. L., S. L. Richardson, and A. A. Rosenberg. 1982. Age and growth of a pleuronectid, Parophrys vetulus, during the pelagic larval period in Oregon coastal waters. Fish. Bull., U.S. 80:93-104. Lough, R. G., M. Pennington, G. R. Bolz, and A. A. Rosenberg. 1982. Age and growth of larval Atlantic herring, Clupea harengus L., in the Gulf of Maine-Georges Bank region based on otolith growth increments. Fish. Bull., U.S. 80:187-199. Mugiya, Y., N. Watabe, J. Yamada, J. M. Dean, D. G. Dunkelberger, and M. Shimizu. 1981. Diurnal rhythm in otolith formation in the goldfish, CaraxsiiLs auratus. Comp. Biochem. Physiol. 68A:659-662. Neilson, J. D., and G. H. Geen. 1982 Otoliths of chinook salmon {Oncorhynchus tshawytscha): daily growth increments and factors influencing their production. Can. J. Fish. Aquat. Sci. 39:1340-1347. Odense, P. H., and V. H. Logan. 1974. Marking Atlantic salmon (Salmo salar) with oxytet- racycline. J. Fish. Res. Board Can. 31:348-350. Pannella, G. 1971. Fish otoliths: daily growth layers and periodical pat- terns. Science (Wash., D.C.) 173:1124-1127. 1980. Growth patterns in fish sagittae. In D. C. Rhoads and R. A. Lutz (editors), Skeletal growth of aquatic organisms. Biological records of environmental change, p. 519- 560. Plenum Press, N.Y. Stri hsaker, P., and J. H. Uchiyama. 1976. Age and growth of the nehu, Stolephorus purpureus (Pisces: Engraulidae), from the Hawaiian Islands as indicated by daily growth increments of sagittae. Fish Bull., U.S. 74:9-17. Tanaka, K., Y. Mugiya, and J. Yamada. 1981. E ff ects of photoperiod and feeding on daily growth pat- terns in otoliths of juvenile tilapia nilotiea. Fish. Bull., U.S. 79:459-466. Taubert, B. D., and D. W. Coble. 1977. Daily rings in otoliths of three species of Lepomis and Tilapia mossambica. J. Fish. Res. Board Can. 34:332-340. Trojnar, J. R. 1973. Marking rainbow trout fry with tetracycline. Prog. Fish-Cult. 35:52-54. Uchiyama, J. H, and P. Struhsaker. 1981. Age and growth of skipjack tuna, Katsuwonus pelamis, and yellowfin tuna, Thunnus albacares, as indicated by daily growth increments of sagittae. Fish. Bull., U.S. 79:151-162. Watabe, N., K. Tanaka, J. Yamada, and J. Dean. 1982. Scanning electron microscope observations of the organic matrix in the otolith of the teleost fish Fundulus heteroclitus (Linnaeus) and tilapia nilotiea (Linnaeus). J. Exp. Mar. Biol. Ecol. 58:127-134. Weber, D., and G. J. Ridgway. 1962. The deposition of tetracycline drugs in bones and scales offish and its possible use for marking. Prog. Fish- Cult. 24:150-155. 1967. Marking Pacific salmon with tetracycline anti- biotics. J. Fish. Res. Board Can. 24:849-865. Wild, A., and T. J. Foreman. 1980. The relationship between otolith increments and time for yellowfin and skipjack tuna marked with tet- 241 racycline. [In Engl, and Span.] Inter.-Am. Trop. Tuna Comm. Bull. 17:509-560. Wilson, K. H., and P. A. Larkin. 1980. Daily growth rings in the otoliths of juvenile sockeye salmon (Onchorhynchus nerka). Can. J. Fish. Aquat. Sci. 37:1495-1498. P. D. SCHMITT School of Biological Sciences, University of Sydney Sydney, N.S.W. 2006, Australia Present address: Southwest Fisheries Center La Jolla Laboratory National Marine Fisheries Service, NOAA P.O. Box 277, La Jolla, CA 92038 TAG-RECAPTURE VALIDATION OF MOLT AND EGG EXTRUSION PREDICTIONS BASED UPON PLEOPOD EXAMINATION IN THE AMERICAN LOBSTER, HOMARUS AMERICANUS Techniques for molt prediction based upon epider- mal and setal development in pleopods (Aiken 1973) and for egg extrusion prediction based upon pleopod cement gland development (Waddy and Aiken 1980; Aiken and Waddy 1982) provide opportunities for more comprehensive studies of growth and re- productive potential in natural American lobster, Homarus americanus, populations than have pre- viously been possible. These laboratory-developed techniques have only recently been applied to field samples from a number of areas of Atlantic Canada (Robinson 1979; Campbell and Robinson 1983; Ennis 1984). Although the methodologies are fairly straightforward and may be applied in field studies quite readily, in practice the investigator will some- times be faced with specimens for which predictions can only be made with some degree of uncertainty. A study of Newfoundland lobsters using these tech- niques has included the tagging of animals from which pleopods were obtained. This paper presents results from observations on recaptured lobsters which validate the predictions that were made at the time of tagging that molting or egg extrusion would or would not occur during the current molting/spawning period. Materials and Methods Pleopods were obtained from American lobsters (ranging from 33 mm to 130 mm CL (carapace length)) caught in traps and by scuba divers near Arnold's Cove, Placentia Bay, Newfoundland, be- tween 24 June and 17 July 1981. These were examined for molt and cement gland stages accord- ing to the methodologies of Aiken (1973), Waddy and Aiken (1980), and Aiken and Waddy (1982). It is clear from Aiken (1973) that one can predict with considerable confidence that lobsters with pleopod stages 3.0 and higher just prior to or early in the molting season will molt that year. It is also clear, however, that for animals with pleopod stages 1.0-2.5 one cannot predict with confidence that molting will or will not occur. Molt prediction for these stages is not reliable because of development plateaus that occur during D0 (i.e., molt stages 1.0-2.5). However, most such plateaus occur at stages 1.5-2.0, and a lobster will rarely remain at stage 2.5 for more than 2 wk. Once an animal has passed beyond stage 2.5, there will be no further plateaus, and proecdysis will proceed at a rate that is regulated by temperature (Aiken 1973). Aiken (1980) also stated that at stage 2.5, the epidermis in the general integument begins to show signs of activity, indicating imminent transi- tion from indecisive D„ into the irreversible premolt development of D,. Considering that animals with stage 2.5 pleopods should molt in 48-52 d at 10°C (Aiken 1973) plus the fact that at Arnold's Cove the July-August temperatures on the lobster grounds average in excess of 10°C (mean daily temperatures from 24 June to 31 August averaged 12. 1°C in 1981), it appeared more likely that lobsters with stage 2.5 pleopods during the 24 June-17 July sampling at Arnold's Cove would molt. As a working hypothesis, it was decided to predict that lobsters with pleopod stages 2.5 and higher would molt during the 1981 molting season at Arnold's Cove and that those with pleopod stages 0-2.0 would not molt. Cement glands were initially staged according to the classification scheme of Waddy and Aiken (1980). These stages were subsequently converted to their more recent scheme (Aiken and Waddy 1982). It is clear from these papers that for lobsters with stage 0 or stage 1 cement glands just prior to or early in the spawning season one can confidently predict that egg extrusion will not occur that year, whereas for those with stage 2 or higher cement glands one can confidently predict that egg extrusion will occur. During the sampling at Arnold's Cove, 356 of the lobsters from which pleopods were removed for molt and cement gland staging were tagged with "sphyrion" tags, which are designed to remain attached through ecdysis (Scarratt and Elson 1965), and released within a few minutes of being taken from the water very close to where they were captured. Observations on 171 of these lobsters recaptured subsequent to the molting/spawning period (mainly during the 1982 fishing season, 20 April-30 June) 242 FISHERY lU'LLETIN: VOL. 82, NO. 1, 1984 provide a basis for validating the molt or egg extru- sion predictions. TABLE 1. — Summary of molt predictions and subsequent valida- tions for American lobsters sampled and tagged at Arnold's Cove, Newfoundland, 24 June-17 July, 1981. Results Molt Predictions Four of the 11 males (36.4%) and 11 of the 27 females (40.7%) with pleopod stages 0-2.0 molted instead of not molting as was predicted (Table 1). Even some with pleopod stage 0 molted. Of the 16 females which did not molt, 1 4 extruded eggs, and the 2 females which did not extrude eggs had stage 1 cement glands, indicating that egg extrusion would not occur. Six out of 21 males (28.6%) with pleopod stages 2.5 and 3.0 did not molt, whereas all with pleopod stages >3.5 and all females with pleopod stages >2.5 did molt (Table 1). Overall, 78.4% of the predictions which could be validated were correct. There was greater success with predicting that molt- ing would occur (89.8% correct predictions) than with predicting it would not (60.5% correct predic- tions). There was no pleopod stage at and below which none molted; however, at stage 3.5 and higher all molted. Validations of molt prediction are available for males ranging in size from 73 to 104 mm CL. Except for one animal at 99 mm, it was only for animals smaller than 8 1 mm that any of the predictions were incorrect. The size range for which validations are available for females is limited (75-82 mm CL). Egg Extrusion Predictions All of 17 females with either stage 0 or stage 1 cement glands did not extrude eggs, and all of 7 with stage 3 cement glands did extrude eggs as predicted. However, 2 out of 9 with stage 2 cement glands, which were predicted would lay eggs, did not do so (Table 2). Overall, 93.9% of the predictions which could be validated were correct. The 2 females which failed to extrude eggs as predicted, molted, despite having molt stage 0 pleopods. Discussion There have long been problems associated with growth rate and functional maturity determinations in American lobsters. Reliable data on annual pro- portions molting (or molt frequency) and proportions laying eggs in relation to size are difficult to obtain. Both these parameters are essential in assessing the impact of size limit and/or exploitation rate changes Number of molt predictions/va lidat ions Ma les rem3les 1 Pleopod Cor- Cor- Cor- Cor- stage Yes rect No rect Yes rect No rect 0 14 9 1.0 1 1 2 2 1.5 8 5 8 4 2.0 2 1 3 1 2.5 7 4 2 2 3.0 14 11 3.5 13 13 1 1 4.0 11 1 1 3 3 4.5 1 1 5.0 2 2 5.5 5 5 'This table does not include 69 females which were ovigerous with old eggs at the time of sampling/tagging, all of which subsequently molted Table 2. — Summary of egg extrusion predic- tions and subsequent validations for female American lobsters sampled and tagged at Arnold's Cove, Newfoundland, 24 June- 1 7 July, 1981. Sixty-nine (69) females which were ovigerous with old eggs at the time of sampling/ tagging, all of which subsequently molted, are not included in the table. Number of egg extrusion Cement g land predictions/va lidat ions stage Yes Correct No Correct 0 8 8 1 9 9 2 9 7 3 7 7 in a lobster fishery on yield per recruit and reproduc- tive potential. Such assessments are important to proper lobster fishery management. The techniques used here to predict molting and egg extrusion provide new approaches to the study of lobster growth and maturity that have only recently been used in studies of lobster populations. Results of this validation study, however, clearly indicate that caution has to be used in their application. In the case of molt prediction it appears that the time of sampling in relation to the molting period is critical. The ideal situation would be a very short annual molting period with sampling just prior to the start of molting when all animals going to molt would have well-developed (stage 3 or higher) pleopods. American lobsters reach the northern limit of their range in Newfoundland waters, and it is probably here that their annual molting period is the shortest. In the Arnold's Cove area, molting starts early in July and is virtually completed by early September. In the present study, 5 out of 1 4 lobsters (all females, Table 1), sampled and tagged between 24 June and 17 July 1981 and had stage 0 pleopods (for which it was pre- 243 dieted that molting would not occur that year), had molted when recaptured prior to the molting period the following year. For these animals premolt development must have occurred very rapidly during the 1 98 1 molting period. This indicates that periodic sampling throughout the molting period along with a validation study are required in order to use these molt prediction techniques as a basis for estimating annual proportions molting in a lobster population. The overall success rate with predicting egg extru- sion was much greater than with molt prediction (94% cf. 78%). The small number of incorrect predic- tions may have resulted from loss of eggs rather than failure of the animals to extrude. One of 6 ovigerous females with newly laid eggs that were tagged during the 24 June-1 7 July sampling period had molted and was nonovigerous when recaptured. While egg extru- sion prediction based upon the cement gland staging technique provides a reliable basis for estimating annual proportions laying eggs in a lobster popula- tion, it is clear that such estimates should be ad- justed, using the kind of information that can be obtained from a validation study before being used in an assessment of reproductive potential in a population. Ennis, G. P. 1984. Comparison of physiological and functional size- maturity relationships in two Newfoundland populations of lobsters Homarus americanus. Fish. Bull, U.S. 82: 244-249. Robinson, D. G. 1979. Consideration of the lobster (Homarus americanus) recruitment overfishing hypothesis; with special reference to the Canso Causeway. In F. D. McCracken (editor), Canso marine environment workshop, Part 3 of 4 Parts, Fishery impacts, p. 77-99. Fish. Mar. Serv. Tech. Rep. 834. SCARRATT, D. J., AND P. F. ELSON. 1965. Preliminary trials of a tag for salmon and lobsters. J. Fish. Res. Board Can. 22:421-423. Waddy, S. L., andD. E. Aiken. 1980. Determining size at maturity and predicting egg extru- sion from cement gland development in Homarus americanus. CAFSAC (Can. Atl. Fish. Sci. Advis. Comm.) Res. Doc. 80/43, 9 p. G. P. ENNIS Department of Fisheries and Oceans Fisheries Research Branch P.O. Box 5667 St. John's. Newfoundland, Canada A 1C 5X1 Acknowledgments I am grateful to S. L. Waddy for her courtesy and cooperation in teaching P. W. Collins the pleopod and cement gland staging techniques. I am indebted to Collins who, in addition to examining all the pleopods in this study, participated in the field work involved in obtaining samples, tagging, and recovery of tagged animals and provided the data summaries. Assistance with field work was provided by G. Dawe and D. G. Badcock to whom I am also very grateful. Literature Cited AlKEN, D. E. 1973. Proecdysis, setal development, and molt prediction in the American lobster (Homarus americanus). J. Fish. Res. Board Can. 30:1337-1344. 1980. Molting and growth. In J. S. Cobb and B. F. Phillips (editors), The biology and management of lobsters. Vol. I, Physiology and behavior, p. 91-163. Acad. Press, N.Y. Aiken, D. E., and S. L. Waddy. 1982. Cement gland development, ovary maturation, and reproductive cycles in the American lobster Homarus americanus. J. Crust. Biol. 2:315-327. Campbell, A., and D. G. Robinson. 1983. Reproductive potential of three American lobster (Homarus americanus) stocks in the Canadian Mari- times. Can. J. Fish. Aquat. Sci. 40:1958-1967. COMPARISON OF PHYSIOLOGICAL AND FUNCTIONAL SIZE-MATURITY RELATIONSHIPS IN TWO NEWFOUNDLAND POPULATIONS OF LOBSTERS HOMARUS AMERICANUS Lobster (genus Homarus) fisheries are characterized by excessive exploitation rates and small, minimum legal sizes in relation to sizes at maturity (Anonymous 1977, 1979). Under such conditions, widespread recruitment overfishing is a distinct possibility and in eastern Canada appears to be the cause of stock collapses in certain areas (Robinson 1979). Stock-recruitment relationships as such are poorly known for the genus Homarus; however, since current levels of landings are well below historical levels in most fisheries, it is reasonable to assume that, within the limits of habitat carrying capacity, increased egg production will result in increased re- cruitment. It is clear that increasing the minimum legal size and/or reducing exploitation rates will result in increased egg production within a lobster stock; however, detailed knowledge of size-fecundity and size-maturity relationships is required to pro- perly assess the impact of changes in fishery regulatory measures on annual egg production within a given stock. 244 FISHERY Bl'LLKTIN: VOL. 82. NO. 1. 1984. Size-maturity relationships, based mainly on obser- vations of ovary color and ova size in nonovigerous females for five Newfoundland lobster populations, indicate 100% maturity (physiological) for sizes at which tagging results show that substantially < 100% of the nonovigerous females lay eggs in a given spawning season (Ennis 1980). Resorption of the mature ovary near the expected time of oviposition is a common phenomenon in//, americanus (Aiken and Waddy 1980a) and presumably is the main reason for failure on the part of physiologically mature females to express their maturity by extruding eggs. Clearly, it is an "expressed" or functional size-maturity relationship that is required to assess the impact of size limit and/or exploitation rate changes in a fishery on annual egg production. Using the pleopod cement gland staging technique described by Aiken and Waddy (1982) as a basis for predicting egg extrusion, such a relationship was derived for two Newfound- land populations. These are compared with physio- logical size-maturity relationships for the same populations. Materials and Methods 52° -I — 52° 50° 48° $2° FIGURE 1. — Map of Newfoundland showing location of Arnold's Cove and Comfort Cove. Pleopods were obtained from 172 nonovigerous female lobsters caught between 24 June and 17 July 1981 and 77 caught between 14 and 18 June 1982 near Arnold's Cove, Placentia Bay, and 246 caught between 1 and 7 July 1982 at Comfort Cove, Notre Dame Bay, Newfoundland, (Fig. 1) using traps and by scuba diving. Sizes ranged from 40 to 111 mm CL (carapace length) at Arnold's Cove and from 58 to 113 mm at Comfort Cove. Pleopods were examined for molt stage according to the method of Aiken (1973) and for cement gland development according to the method of Aiken and Waddy (1982) to deter- mine whether molting or egg extrusion would occur during the current molting/spawning period. In this study it was predicted that females with cement glands in stages 0 and 1 would not extrude eggs dur- ing the current spawning period whereas those with stage 2 or higher cement glands would (see Aiken and Waddy 1982 for descriptions of cement gland stages). A validation study (Ennis 1983) has demonstrated that egg extrusion prediction based on cement gland staging is quite reliable. Of the predic- tions that could be validated, 947c were correct. The only incorrect predictions were for females with stage 2 cement glands of which 2 out of 9 (22%) failed to extrude eggs. Accordingly, in the data analyzed here the number of animals with stage 2 cement glands in each size group was reduced by 22% to obtain a more accurate estimate of the number that would actually extrude eggs. Where 22% of the number was < 0.5, nothing was subtracted. The two Arnold's Cove samples were combined. For each area the numbers examined and numbers functionally mature (i.e., going to extrude eggs during the current season) were grouped by 1 mm CL and subjected to probit analysis. Although good statisti- cal fits were obtained (P values >0.9), the fitted curves did not approximate the data very well at the upper and lower ends. Proportions from the same data were analyzed using the logistic equation Y = 1 +eb+cX (1) An SAS1 program, which performs this analysis by means of a nonlinear regression procedure using the Marquardt method, was used. Curves were obtained with substantially improved visual fits to the data. Previously published size-maturity relationships for Arnold's Cove and Comfort Cove lobsters (Ennis 1980) were based mainly on detailed examination of the gonads of nonovigerous females, but ovigerous females in the samples were included as mature animals. For this paper the ovigerous specimens were excluded from these samples and the data *SAS User's Guide: Statistics, 1982 ed. SAS Institute Inc., Cary, N.C., 584 p. 245 reanalyzed using the above equation. The size maturity relationships thus derived are a more accurate reflection of the proportions of non- ovigerous females whose gonads are developing for extrusion during the upcoming spawning season (i.e., physiologically mature). Results The smallest female lobsters with cement glands in stage 2 (or higher), indicating that egg extrusion would occur during the current spawning period, were 73 mm CL at Arnold's Cove and 7 1 mm at Com- fort Cove (Tables 1, 2). All smaller animals had stage 0 or 1 cement glands, indicating that egg extrusion would not occur. The largest female lobsters with cement glands in stage 0 or 1 were 96 mm CL at Arnold's Cove and 88 mm at Comfort Cove. All larger animals had stage 2 (or higher) cement glands. Functional and physiological size-maturity re- lationships were derived for each area and plotted together (Figs. 2, 3). Sizes at 50% functionally mature female lobsters from the relationships were 81 mm CL at Arnold's Cove and 80 mm at Comfort Cove. These compare with sizes at 50% physiologically mature female lobsters of 74 mm and 76 mm for Arnold's Cove and Comfort Cove, respectively. Observations taken from the data indicate that at Arnold's Cove the shift in physiological maturity from none to all occurred over a 9 mm CL size range (71-80 mm) compared with a 25 mm size range (72- 97 mm) for functional maturity. The equivalent size ranges for Comfort Cove lobsters were 22 mm CL (64-86 mm) for physiological maturity and 23 mm (70-93 mm) for functional maturity. Examination of the fitted curves shows considerable disparity be- tween proportions of physiologically mature and functionally mature lobsters at given sizes over much of the size range in each area. In order to quantify this disparity, points on the curves were treated as num- bers (out of 100) rather than percentages and the dif- ference determined between the two curves at any given size. The greatest disparities were for 73 mm CL lobsters at Arnold's Cove (Fig. 2) and for 70 mm lobsters at Comfort Cove (Fig. 3) where this com- parison of the curves indicates that 60% and 41%, respectively, of the physiologically mature animals fail to extrude eggs. This percentage decreases with increasing size in each area. To derive an estimate of this percentage for the population as a whole, the Table 1 . - Cement gland stages for female lobsters caught at Arnold's Cove, Newfoundland, 24 June - 17 July 1981 and 14-18 June 1982. Carapace length Cement gland si age (mm) 0 1 2 3 4 Total 40-69 31 31 70 2 1 3 72 I 1 73 3 1 1 1 6 74 2 :• 75 2 2 /)_, 2 3 2 7 77 5 3 3 4 15 78 3 4 3 3 2 15 79 3 6 7 6 9 31 80 3 3 B 4 6 24 81 2 6 9 4 1 22 82 1 1 1 1 4 83 4 2 5 1 1 13 84 1 1 1 3 85 2 2 3 4 1 12 86 2 1 3 1 7 87 1 1 1 3 6 88 1 9 1 1 1 89 1 1 3 1 6 90 2 1 3 91 1 3 4 92 I 1 93 1 1 2 94 1 1 95 2 1 3 96 1 1 2 4 97 1 1 2 98 2 2 4 100 1 1 102 2 2 107 1 1 109-111 1 1 2 TABLE 2. — Cement gland stages for female lobsters caught at Comfort Cove, Newfoundland, 1-7 July 1982. Carapace length Cement gland stage (mm) 0 1 2 3 4 Total 58-69 7 1 8 70 2 1 3 71 1 1 1 3 72 3 1 4 73 2 1 3 74 1 2 1 4 75 5 5 76 1 2 3 77 2 1 3 78 1 2 1 4 79 1 1 7 9 80 1 7 8 81 2 7 9 82 3 2 8 13 83 1 4 16 4 25 84 4 15 19 85 1 2 15 1 19 86 4 4 4 12 87 1 2 8 6 17 88 1 13 1 15 89 8 1 9 90 5 5 10 91 4 1 5 92 3 2 5 93 2 4 6 94 1 2 1 4 95 1 1 2 96 4 4 97 1 1 98 2 2 100 1 2 3 101-1 13 7 2 9 246 100 i I I 1 1 I T 1 1 1 X^_X-*-*-X — X-» — * — X I 1 — x x-x 90 / x / - 80 / / - 70 PHY 1 ■ SI0L0GICAL MATURITY 10404 \ / / x / / x IX X IX I - 14-0943 l + e -0I896X £ 60 >— 3 N = 167 x / / " X " -50 X XX - O x / x "" 40 - 30 X X / \ FUNCTIONAL MATURITY ' N 0-9694 - 15 9542 -0-1983 X l + e 20 N = 230 - 10 - 0 — 1 =r»Hr«- "fx XX 1 XXX X 1 1 X XX 1 i i i i i i i 50 55 60 65 70 75 80 85 90 95 100 105 110 115 CARAPACE LENGTH (mm) FlGl'RK 2. — Physiological and functional size-maturity relationships for female lobsters at Ar- nold's Cove, Newfoundland. Functional maturity data only are provided. i 1 r X DM X » » — «-« ■ ■ X « « 100 90 80 70 - 60 50 40 - 30 - 20 10 PHYSIOLOGICAL MATURITY 10254 16-0316 -0-21 10 X l + e N = 250 FUNCTIONAL MATURITY . 0-9801 ' 15-9997 -0-2019 X l + e N=246 XXXXXXXX X I I 1 50 55 65 70 75 80 85 90 95 100 105 110 CARAPACE LENGTH (mm) FIGURE 3. — Physiological and functional size-maturity relationships for female lobsters at Com- fort Cove, Newfoundland. Functional maturity data only are provided. 247 above procedure was followed for those sizes be- tween the largest with 100% functionally immature and the smallest with 100% functionally mature (from the data) and the numbers added. The result- ing estimates were 25% at Arnold's Cove and 20% at Comfort Cove. Discussion This study has demonstrated that, failure on the part of physiologically mature female lobsters to "express" their maturity by extruding eggs is quite common in the wild. Resorption of the mature ovary near the expected time of extrusion appears to be the main reason. Resorption occurs when the molting and reproductive cycles conflict (Aiken and Waddy 1976, 1980a, b). These cycles are normally syn- chronized by temperature and photoperiod regimes so that conflict between them is minimized. However, final ovary maturation is disrupted, if it coincides with middle to late premolt, and the ovary is resorbed prior to the impending molt. Not only would this ensure the conservation of energy, but it might also serve to resynchronize the molt and re- productive cycles (Aiken and Waddy 1980b). Nonfertilization may also be a cause of resorption. In Jasus lalandii, for example, oviposition will not occur in unfertilized females (Heydorn 1969). While oviposition will occur in H. americanus even if the female has not successfully mated (Aiken and Waddy 1980a), it is not clear if this is the rule or the excep- tion. Physiologically mature H. americanus females which are unfertilized (i.e., empty seminal recep- tacles) occur in the wild (Krouse 1973; Ennis 1980). In sampling from January to June 1973 at St. Chads, Bonavista Bay, on the northeast coast of Newfound- land, Ennis (1980) found 6 (11.5%) of 52 physiologically mature females to be unfertilized. At Arnold's Cove in August and September 1981, 98 of 100 females >79 mm CL were fertilized as deter- mined by the presence of spermatophores in seminal receptacles. While nonfertilization may be a con- tributing factor in some areas, it does not appear to be a major cause of ovary resorption in wild H. americanus. A validation study (Ennis 1983) has demonstrated that the cement gland staging technique enables a reliable prediction of whether a female lobster will extrude eggs during the upcoming spawning season. However, caution has to be exercised in applying a functional size-maturity relationship based on these predictions because there is substantial loss of eggs subsequent to spawning. For example, 2 of 15 females with well-developed (stages 3 and 4) cement glands, indicating extrusion to be imminent, and 1 of 6 females with newly laid eggs (all tagged during the 24 June to 17 July 1981 sampling period at Arnold's Cove) had molted and were nonovigerous when recaptured prior to the 1982 molting/spawning period. There is also substantial loss of eggs other than through molting. Some of this loss may be the result of eggs not being fertilized. Unfertilized eggs do not attach securely and may be lost soon after oviposi- tion, but in some cases a fair number will remain attached for several months (Aiken and Waddy 1980a, b). However, it is common for fertizlied eggs to be lost as well (Aiken and Waddy 1980a , b). Nor- mal attrition of properly attached (fertilized) eggs over the 9-12 mo incubation period has been estimated at around 36% (Perkins 1971); however, some females lose up to 100% of their eggs. The six ovigerous females referred to above (i.e., tagged dur- ing 24 June to 17 July 1981 at Arnold's Cove) had apparently normal clutches of eggs when tagged, but, of the five that had eggs when recaptured, four had normal clutches and one had < 200 eggs remaining. A normal clutch for this particular animal, which was 79 mm CL, would have been about 10,000 eggs (Ennis 1981). Similar observations were made on animals tagged between 1 and 14 August 1981 at Arnold's Cove. Of six females with newly laid, normal-sized clutches of eggs, one had just a few hundred eggs remaining when recaptured. Another female, which had well-developed (stage 4) cement glands, had no eggs but had pleopods covered with cement when recaptured, indicating that eggs had been extruded and subsequently lost (Templeman 1940). These observations demonstrate that there is sub- stantial loss of eggs subsequent to extrusion over and above that attributed to normal attrition. This loss of eggs should be taken into account in any assessment of the impact of changes in fishery regulatory measures on reproductive potential (i.e., annual egg production) in a population. Acknowledgments I am grateful to P.W. Collins who was responsible for collecting the samples and examining the pleopods for molt stage and cement gland develop- ment and to G. Dawe and D. G. Badcock who assisted in the collection of the samples. Literature Cited Aiken, D. E. 1973. Proecdysis, setal development, and molt prediction in 248 the American lobster (Homarus americanus). J. Fish. Res. Board Can. 30:1337-1344. AIKEN, D. E., AND S. L. WADDY. 1976. Controlling growth and reproduction in the American lobster. In J. W. Avault, Jr. (editor), Proceedings of the 7th Annual Meeting World Mariculture Society, p. 415- 430. Louisiana St. Univ. Press, Baton Rouge. 1980a. Reproductive biology. In J. S.Cobb and B. F.Phillips (editors), The biology and management of lobsters. Vol. I, Physiology and behavior, p. 215-276. Acad. Press, N.Y. 1980b. Maturity and reproduction in the American lobster. In V. C. Anthony and J. F. Caddy (editors), Pro- ceedings of the Canada-U.S. Workshop on Status of Assessment Science for N.W. Atlantic Lobster {Homarus americanus) Stocks, St. Andrews, N.B., Oct. 24-26, 1978, p. 59-71. Can. Tech. Rep. Fish. Aquat. Sci. 932, St. Andrews, Can. 1982. Cement gland development, ovary maturation, and reproductive cycles in the American lobster Homarus americanus. J. Crust. Biol. 2:315-327. Anonymous. 1977. Report of the working group onHomarus stocks. ICES CM. 1977/K:11, 19 p. 1979. Report of the Homarus working group. ICES CM. 1979/K:8, 49 p. ENNIS, G. P. 1980. Size-maturity relationships and related observations in Newfoundland populations of the lobster (Homarus americanus). Can. J. Fish. Aquat. Sci. 37:945-956. 1981. Fecundity of the American lobster, Homarus americanus, in Newfoundland waters. Fish. Bull., U.S. 79:796-800. 1983. Tag-recapture validation of molt and egg extrusion predictions based upon pleopod examination in the American lobster, Homarus americanus. Fish. Bull., U.S. Heydorn, A. E. F. 1969. The rock lobster of the South African west coast Jasus lalandii (H. Milne-Edwards). 2. Population studies, behaviour, reproduction, moulting, growth and mi- gration. S. Afr. Div. Sea Fish. Invest. Rep. 7:1-52. KROUSE, J. S. 1973. Maturity, sex ratio, and size composition of the natural population of American lobsters, Homarus americanus, along the Maine coast. Fish. Bull., U.S. 71:165-173. Perkins, H. C. 1971. Egg loss during incubation from offshore northern lobsters (Decapoda: Homaridae). Fish. Bull., U.S. 69:451- 453. Robinson, D. G. 1979. Consideration of the lobster (Homarus americanus) recruitment overfishing hypothesis; with special reference to the Canso Causeway. In F. D. McCracken (editor), Canso Marine Environment Workshop. Part 3 of 4 parts. Fishery impacts, p. 77-99 Fish. Mar. Serv. Tech. Rep. 843. TEMPLEMAN, W. 1940. The washing of berried lobsters and the enforcement of berried lobster laws. Newfoundland Dep. Nat. Resour. Res. Bull. (Fish.) 10,21 p. G. P. ENNIS Fisheries Reasearch Branch Department of Fisheries and Oceans P.O. Box 5667 St. John's, Newfoundland, Canada AlC 5X1 CONVERSIONS BETWEEN TOTAL, FORK, AND STANDARD LENGTHS IN 35 SPECIES OF SEBASTES FROM CALIFORNIA In recent years, the rockfishes (Scorpaenidae: Sebas- tes) of the northeastern Pacific Ocean have been investigated extensively. With many institutions studying diverse aspects of their biology and fisheries, a lack of standardized methods has ham- pered attempts to synthesize the data. A particular problem has been the reporting of different length measurements. To provide the means to convert one of these length measurements to another, we report here the linear regression statistics necessary for conversions in 35 species of Sebastes. Specimens were collected from fishery catches be- tween Cape Blanco, Oreg., and San Diego, Calif., dur- ing 1977-82. The sample included five fish for each centimeter of body length throughout the size range of each species. Measurements were taken on a meter board in millimeters on frozen, then thawed, carcasses. Standard length was measured from the anterior tip of the upper jaw to the posterior end of the vertebral column (Hubbs and Lagler 1970:25); fork length was measured from the anterior tip of the longest jaw to the median point of the caudal fin; and the total length was measured from the most anterior tip of the longest j aw to the most posterior part of the tail when the caudal rays are squeezed together (Holt 1959:71). Linear regressions were run on all com- binations of the measurements of length. Outliers (±3.0 standard deviations) from the line were noted by the computer program, then checked for data entry error and corrected when possible. If a data entry error was not found, an outlier was assumed to result from measurement error and the observation was deleted. Statistics reported for each species arey-intercept (a), slope (/?), standard error of estimate (Svx), cor- relation coefficient (r), range in length, and the sam- ple size used in the regression (n) (Tables 1-3). Estimates of a imply impossible values for the dependent variable when the independent variable is zero. The impossible results could be caused by ran- dom error in estimation of a or nonlinearity for values less than those observed. The high values of r and examination of scattergrams indicate that the length relationships are linear over the observed range of values. The standard precaution of limiting the application of these regressions to the ranges of observed values is advised. To calculate the total length (TL) of S. alutus, given a standard length (SL) of 250 mm, the regression values from Table 1, total length on standard length, are used so that FISHERY BULLETIN: VOL. 82. NO. 1. 1984. 249 Table 1. — Results of linear regressions of standard length versus total length for Sebastes. Measurements are in millimeters. Species of Sebastes n r Star len dard gth Total length standard len on gth Standard length on total length Mm Max a P Vx a P Vx alutus 49 0995 232 361 1454 1 249 3.746 2.056 0.792 2.984 aunculatus 1 16 1 000 72 426 -1423 1 240 3.787 1.369 0806 3054 aurora 43 0.991 164 324 0.098 1 220 4 709 4.398 0.806 3.827 babcocki 74 0 999 185 532 6.478 1 196 4 833 -4.614 0.834 4035 cam at us 105 0999 75 292 3.676 1 201 2.206 -2.866 0832 1 836 caunnus 113 0997 1 1 1 443 3 873 1 209 5 769 -0.653 0820 4.568 chlorostictus 107 0999 107 382 5316 1 202 3 636 -3.931 0830 3.023 chrysomelas 60 0998 77 316 1 007 1 211 3 161 -0.123 0.822 2 605 constellatus 105 0999 148 365 4497 1 175 3 119 -3.204 0849 2.651 cramen 102 0 999 102 394 -0.304 1 266 4.153 0.737 0.788 3 278 diploproa 80 0999 87 308 1 286 1 242 2 718 -0.740 0 804 2 188 elongatus 108 0998 107 317 15.238 1 165 3 543 -12.144 0855 3036 entomelas 105 0998 194 435 9.496 1 211 5.679 -6.296 0822 4.679 fl avid us 193 0.997 191 453 0.468 1 247 5.700 1 379 0.798 4.558 goodei 99 1 000 101 449 4.199 1 224 2 870 -3.196 0816 2.344 hopkinsi 71 0993 99 251 3059 1 200 4 788 -0.195 0.822 3 964 lordani 145 0998 77 260 4 610 1 216 2 903 -3.128 0819 2 382 levis 31 1 000 190 717 -4.500 1 248 4 907 3.813 0.801 3.932 mahger 42 0996 174 397 1463 1 220 5 639 1.120 0.813 4.604 melanops 138 0999 74 495 7 724 1 221 5.193 -5 596 0817 4.247 me/anostomus 8/ 0.994 207 421 -0954 1 244 6897 4 780 0.794 5 508 mimatus 109 0.994 237 550 9629 1 229 9.765 -3.095 0804 7 900 mystmus 163 0998 102 387 2.930 1 238 5694 -1.192 0.804 4,588 nebulosus 69 0995 213 366 4.294 1 196 3.962 -0.731 0828 3296 ovahs 83 0.997 181 375 0550 1 225 4.374 1.329 0.81 1 3 558 paucispims 163 0999 103 649 -5.035 1 262 7.550 4882 0.790 5.974 p/nn/ger 136 0 997 196 565 11476 1 239 8002 -7 447 0803 6.443 rosaceus 83 0.996 132 263 3.917 1 199 2 867 -1 794 0 828 2 383 rosenblatti 104 0 999 132 428 9.567 1 182 3.653 -7.374 0844 3086 ruberrtmus 118 0996 203 565 5.856 1 202 9465 -1 71 7 0826 7.843 rufus 26 0999 152 447 12.946 1 1 77 5 963 -10.316 0.848 5.061 saxtcola 68 0999 109 288 3.226 1 242 2 456 -2.252 0.804 1.976 semicmctus 31 0.979 101 147 8179 1 170 3.617 -1.752 0.820 3027 serranoides 129 0995 190 441 8.292 1 209 7 277 -3542 0.819 5988 witsom 48 0 999 71 126 0572 1 234 1 071 -0.231 0808 0 868 Table 2. — Results of linear regressions of standard length versus fork length for Sebastes. Measurements are in millimeters. Species of Sebastes n r Star lei dard gth Fork 1 stand ength on ard length Standard length on fork length Mm Max a P Vx a P Vx alutus 48 0996 232 361 -0281 1 195 3024 2.492 0.831 2 521 aunculatus 114 0 999 72 426 -0.369 1 228 4.126 0575 0.813 3 358 aurora 44 0 993 164 324 -3 046 1 201 4 237 6237 0.821 3 502 babcocki 76 0999 185 532 9034 1 153 5 190 -6 860 0 865 4.496 carnatus 104 0999 75 292 4.601 1 194 2 425 -3 613 0 836 2.030 caunnus 1 17 0 996 11 1 448 5896 1 187 6 764 -2 272 0 836 5 674 chlorostictus 107 0999 107 382 5 289 1 171 3.719 -3 987 0 852 3 173 chrysomelas 58 0997 77 226 1.137 1 209 3 209 -0009 0 822 2.647 constellatus 107 0999 148 365 3 883 1 .152 2 964 -2.774 0.866 2 571 cramen 103 0999 102 394 1 390 1 205 4.282 -0.565 0.828 3 550 diploproa 82 0999 87 308 2 092 1 181 2.627 -1 460 0845 2 223 elongatus 116 0998 107 317 14 186 1 116 3469 -11 724 0 892 3 102 entomelas 106 0997 194 435 1 6 964 1 124 5602 -13 326 0885 4 970 tl avid us 198 0998 191 453 -0 918 1 213 5367 2 363 0.820 4.412 goodei 99 1.000 101 449 1 515 1 159 3 085 -0988 0862 2 660 hopkinsi 72 0 994 99 251 3.011 1 153 4 465 -0.372 0856 3 847 jordani 154 0998 77 260 5.645 1 124 2.519 -4 418 0887 2 238 levis 34 0 999 190 717 0 033 1 1 77 8 446 0688 0 848 7.169 maliger 41 0997 1 74 397 11835 1 173 4867 -8202 0848 4.138 melanops 135 0999 74 495 7 149 1 197 5.042 -5 247 0834 4 209 me/anostomus 86 0.994 207 421 -0.828 1 201 6.853 4.912 0822 5670 mimatus 106 0.994 237 550 16 442 1 168 9.200 -9445 0.847 7836 mystmus 164 0998 102 387 0 352 1 192 4.975 0 644 0836 4.166 nebulosus 71 0 993 213 366 6934 1 181 4623 -1.852 0835 3 888 ovahs 83 0996 181 375 -3 554 1 187 4 677 5 130 0 836 3925 paucispims 162 0999 103 649 -4 082 1 209 6819 4 183 0 826 5636 pmniger 138 0998 196 565 12880 1 164 7,440 -9 326 0 855 6378 rosaceus 83 0997 132 263 1399 1 187 2.730 0.225 0 837 2 293 rosenblatti 104 0999 132 428 9 938 1 147 3.347 -8.023 0870 2.915 ruberrtmus 1 18 0 996 203 565 6 665 1 181 9028 -2 664 0841 7.620 250 TABLE 2.— Continued Stand ard Fork length on Standard length Species of Se hastes leng rh standard ten gth on fork length n r Mm Max f> fi Vx a P Vx rufus 26 0999 152 447 14 246 1.112 4.416 -12 392 0 898 3 969 saxicola 77 0.999 109 288 3 234 1 200 2.511 -2 315 0831 2.090 semicmclus 31 0.978 101 147 6486 1.128 3 562 -0.343 0849 3.091 serranoides 126 0995 190 441 4 422 1.184 6.779 -0672 0837 5.700 wilsoni 53 0999 71 126 0671 1 203 0884 -0.372 0.830 0.734 Table 3. — Results of linear regressions of fork length versus total length for Sebastes. Measurements are in millimeters. Fork Total ength on Fork I ength on Species of Sebastes length fork length total length n r Mm Max a P Vx a P Vx alutus 48 0999 278 430 -0.003 1 050 1 483 1.321 0.949 1.272 aunculatus 113 1 000 90 529 -0.586 1 007 1 637 0634 0.993 1.626 aurora 43 0998 198 388 2 293 1 019 2 349 -0.917 0.977 2 300 babcocki 72 1.000 222 635 -1.146 1 032 2.392 1.336 0968 2.316 carnatus 101 1.000 92 351 -0.759 1 005 0.510 0.768 0.995 0.507 caurinus 107 0999 135 538 0629 1 010 3.022 0005 0988 2 988 chlorostictus 106 1 000 127 449 -0 723 1 028 1.905 0858 0.972 1.852 chrysomelas ' constellatus 104 1.000 174 422 -0 134 1 023 1.504 0.301 0.977 1.470 era men 99 1000 124 480 -1.700 1 051 2 002 1 756 0.951 1.904 diploproa 80 1 000 106 364 -0558 1 049 1 704 0669 0.953 1 625 elongatus 102 1.000 129 360 -0 552 1 047 1 449 0.701 0954 1.383 entomelas 100 0999 231 496 -6 845 1 072 3 251 6954 0.931 3.029 it avid us 191 1.000 226 551 2.358 1 025 2 439 -1 906 0.974 2.377 goodei 96 1 000 122 527 2 468 1 057 2 647 -2.096 0.945 2.503 hopkinsi 70 0999 115 292 0002 1 041 1 917 0.428 0959 1.840 jordani 140 0999 89 296 -1.872 1 086 1 885 2.036 0.920 1.735 levis 34 1 000 228 855 -3.335 1 055 4452 3369 0.947 4.219 matiger 40 0999 215 480 -8.696 1 034 2 782 9075 0965 2 687 melanops 132 1.000 91 599 1595 1 017 2.099 -1.421 0983 2 063 melanostornus 82 0.999 247 519 -0.635 1 036 2.181 1.065 0 964 2.103 mmtatus 103 0999 293 654 -7 857 1 054 4638 8 665 0946 4394 mystmus 158 1 000 122 463 2 495 1 039 2.329 -2 164 0962 2.241 nebulosus 71 1 000 256 498 0.854 1 001 1.423 -0.487 0998 1.420 ova lis 78 0999 225 438 3.914 1 033 1.996 -3.311 0967 1.931 paucispinis 157 1.000 123 781 -0.870 1 045 2.273 0.930 0.956 2.174 pinniger 132 1 000 235 586 -4.107 1 070 2.822 4.108 0934 2 638 rosaceus 79 0999 158 316 1.409 1 015 1.173 -1.085 0984 1.155 rosenblatti 103 1.000 155 497 -0453 1 030 2.026 0692 0970 1 966 rubernmus 118 1.000 243 680 -0 758 1 018 3.640 1.296 0.981 3.573 rufus 24 1.000 182 517 -2.197 1 057 1 659 2.135 0946 1.569 saxicola 69 0999 136 347 -0.669 1 038 2009 0.921 0963 1.935 semicmctus 29 0998 119 174 -0 422 1 050 1 178 1.010 0 949 1.120 serranoides 125 0999 222 518 1419 1 029 2.623 -0862 0.971 2 548 wilsoni 45 1.000 86 151 -1.141 1 035 0.560 1.182 0 966 0.541 1No regression was run because total length and fork length are equal. TL = a + p (SL) TL= 1.454 + (1.249) (250) TL= 313.7 mm. HlJBBS, C. L.. AND K. F. Lagler. 1970. Fishes of the Great Lakes region. Univ. Michigan Press, Ann Arbor, 213 p. Literature Cited Holt, S. J. 1959. Report of the international training center on the methodology and technqiues of research on mackerel (Rastrelliger). FAO/ETAP Rep. 1095, 129 p. Tina Echeyerria William H. Lenarz Southwest Fisheries Center Tiburon Laboratory- National Marine Fisheries Service, NOAA 3150 Paradise Drive Tiburon, CA 94920 251 NOTICES NOAA Technical Reports NMFS published during first 6 months of 1983 Circular 448. Synopsis of biological data on the grunts Haemulon aurolineatum and H. plumieri (Pisces: Haemulidae). By George H. Darcy. February 1983, iv + 37 p., 33 figs., 26 tables. 449. Synopsis of biological data on the pigfish, Orthopristis chrysoptera (Pisces: Haemulidae). By George H. Darcy. March 1983, iv + 23 p., 22 figs., 15 tables. 450. The utility of developmental osteology in taxonomic and systematic studies of teleost larvae: A review. By JeanR.Dunn. June 1983, iii+ 19p., 7 figs., 5 tables. Special Scientific Report — Fisheries 761. Sea level variations at Monterey, California. By Dale Emil Bretschneider and Douglas R. McLain. January 1983, iii + 50 p., 16 figs., 3 tables, App. A, B. 762. Abundance of pelagic resources off California, 1963-78, as measured by an airborne fish monitoring program. By James L. Squire, Jr. February 1983, v + 75 p., 65 figs., 4 tables. 763. Climatology of surface heat fluxes over the California Current region. By Craig S. Nelson and David M. Husby. February 1983, iii + 155 p., 21 figs., 1 table, App. I, II, III. 764. Demersal fishes and invertebrates trawled in the northeastern Chukchi and western Beaufort Seas, 1976-77. By Kathryn J. Frost and Lloyd F. Lowry. February 1983, iii + 22 p., 4 figs., 6 tables, App. A. 765. Distribution and abundance of larvae of king crab, Paralithodes camtschatica, and pandalid shrimp in the Kachemak Bay area, Alaska, 1972 and 1976. By Evan Haynes. April 1983, iii + 64 p., 29 figs., 1 table, App. 766. An atlas of the distribution and abundance of dominant benthic inverte- brates in the New York Bight apex with reviews of their life histories. By Janice V. Caracciolo and Frank W. Steimle, Jr. March 1983, v + 58 p., 69 figs., 5 tables. 767. A commercial sampling program for sandworms, Nereis virens Sars, and bloodworms, Glycera dibranchiata Ehlers, harvested along the Maine coast. By Edwin P. Creaser, Jr., David A. Clifford, Michael J. Hogan, and David B. Sampson. April 1983, iv + 56 p., 16 figs., 30 tables, App. A. 768. Distribution and abundance of east coast bivalve mollusks based on speci- mens in the National Marine Fisheries Service Woods Hole collection. By Roger B.Theroux and Roland L.Wigley. June 1983, xvi+ 172 p., 121 figs., 327 tables. Some NOAA publications are available by purchase from the Superintendent of Documents, U.S. Govern- ment Printing Office, Washington, DC 20402. Individual copies of NOAA Technical Reports (in limited numbers) are available free to Federal and State government agencies and may be obtained by writing to Publication Services Branch (E/Al 13), National Environmental Satellite, Data, and Information Service, NOAA, 3300 Whitehaven Street, N.W., Washington, DC 20235. Ctmtcnts — conlmtivti LOVE, MILTON S., GERALD E. McGOWEN, WILLIAM WESTPHAL, ROBERT J. LAVENBERG, and LINDA MARTIN. Aspects of the life history and fishery of the white croaker, Genyonemus lineatus (Sciaenidae), off California 179 MORRIS, PAMELA A. Feeding habits of blacksmith, Chromis punctipinnis , associated with a thermal outfall 199 MYRICK, ALBERT C., JR., EDWARD W. SHALLENBERGER, INGRID KANG, and DAVID B. MacKAY. Calibration of dental layers in seven captive Hawaiian spinner dolphins, Stenella longirostris, based on tetracycline labeling 207 ROSS, STEVE W. Reproduction of the banded drum, Larimus fasciatus, in North Carolina 227 Notes SCHMITT, P. D. Marking growth increments in otoliths of larval and juvenile fish by immersion in tetracycline to examine the rate of increment formation 237 ENNIS, G. P. Tag-recapture validation of molt and egg extrusion predictions based upon pleopod examination in the American lobster, Homarus americanus 242 ENNIS, G. P. Comparison of physiological and functional size-maturity relationships in two Newfoundland populations of lobsters Homarus americanus 244 ECHEVERRIA, TINA, and WILLIAM H. LENARZ. Conversions between total, fork, and standard lengths in 35 species of Sebastes from California 249 # GPO 693-007 ^°^o. Sr4T£S 0< * Fishery Bulletin "\ r Vol. 82, No. 2 April 1984 ROPES, JOHN W., STEVEN A. MURAWSKI, and FREDRIC M. SERCHUK. Size, age, sexual maturity, and sex ratio in ocean quahogs, Arctica islandica Linne, off Long Island, New York 253 BRODEUR, RICHARD D., and WILLIAM G. PEARCY. Food habits and dietary overlap of some shelf rockfishes (genus Sebastes) from the northeastern Pacific Ocean 269 COLVOCORESSES, J. A., and J. A. MUSICK. Species associations and community composition of Middle Atlantic Bight continental shelf demersal fishes 295 HAYNES, EVAN B. Early zoeal stages of Placetron wosnessenskii and Rhinolith- odes wosnessenskii (Decapoda, Anomura, Lithodidae) and review of lithodid larvae of the northern North Pacific Ocean 315 ZIMMERMAN, ROGER J., THOMAS J. MINELLO, and GILBERT ZAMORA, Jr. Selection of vegetated habitat by brown shrimp, Penaeus aztecus, in a Galveston Bay salt marsh 325 STANDARD, GARY W, and MARK E. CHITTENDEN, Jr. Reproduction, move- ments, and population dynamics of the banded drum, Larirnus fasciatus, in the Gulf of Mexico 337 TETTEY, ERNEST, CHRISTOPHER PARDY, WADE GRIFFIN, and A. NELSON SWARTZ. Implications of investing under different economic conditions on the profitability of Gulf of Mexico shrimp vessels operating out of Texas 365 BUCK, JOHN D. Quantitative and qualitative bacteriology of elasmobranch fish from the Gulf of Mexico 375 BOTTON, MARK L., and HAROLD H. HASKIN. Distribution and feeding of the horseshoe crab, Limulus polyphemus, on the continental shelf off New Jersey .... 383 PEARCY, W, T NISHIYAMA, T. FUJII, and K. MASUDA. Diel variations in the feeding habits of Pacific salmon caught in gill nets during a 24-hour period in the Gulf of Alaska 391 RUGGERONE, GREGORY T, and DONALD E. ROGERS. Arctic char predation on sockeye salmon smolts at Little Togiak River, Alaska 401 Notes MAULE, ALEC G., and HOWARD F. HORTON. Feeding ecology of walleye, Stizostedion vitreum vitreum, in the mid-Columbia River, with emphasis on the interactions between walleye and juvenile anadromous fishes 411 (Continued on back coven Seattle, Washington U.S. DEPARTMENT OF COMMERCE Malcolm Baldnge, Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION John V. Byrne, Administrator NATIONAL MARINE FISHERIES SERVICE William G. Gordon. Assistant Administrator Fishery Bulletin The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. The Bulletin of the United Stales Fish Commission was begun in 188), it became the Bulletin of the Bureau of Fisheries in 19dl and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1 103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in ,. single issue of the bulletin instead of being issued individually. Beginning with volume 70. number 1. January 1972. the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents. U.S. Government Printing Office. Washington DC 20402. It is also available free in limited numbers to libraries, research institutions. State and Federal agencies, and in exchari nl ific publications SCIENTIFIC EDITORS. Fishery Bulletin Dr. Carl J. Sindermann Northeast Fisheries Center Sandy Hook Laboratory National Marine Fisheries Service, NOAA Highlands, NJ 07732 Dr. William J. Richards Southeast Fisheries Center Miami Laboratory National Marine Fisheries Service, NOAA Miami, FL 33149-1099 Editorial Committee Dr. Bruce B. Collette National Marine Fisheries Service Dr. Edward D. Houde Chesapeake Biological Laboratory Dr. Merton C. Ingham National Marine Fisheries Service Dr. Reuben Lasker National Marine Fisheries Service Dr. Donald C. Malins National Marine Fisheries Service Dr. Jerome J. Pella National Marine Fisheries Service Dr. Jay C. Quast National Marine Fisheries Service Dr. Sally L. Richardson Gulf Coast Research Laboratory Mary S. Fukuyama, Managing Editor ■\ Bulletin - published quarterly by the Scientific Publications Office, National Marine iinl Way NE, BIN C15700, Seattle, WA 98115. Second class postage is paid at Seattle. Wash., and additional offices. POSTMASTER send addre - for subscriptions to Superintendent of Docui S Government Printing Office. Washington, DC 20 10 Although the contents have nm been copyrighted and may be reprinted entirely, reference to source is appreciated. The - mmerce has determined that the publication of this periodical is necessary in the transaction of the public iiss required by law of this Department. I is for printing of this periodical has been approved by the Director o( the Office of Management and Budget through 1 April 19 Fishery Bulletin CONTENTS Vol. 82, No. 2 April 1984 ROPES, JOHN W., STEVEN A. MURAWSKI, and FREDRIC M. SERCHUK. Size, age, sexual maturity, and sex ratio in ocean quahogs, Arctica islandica Linne, off Long Island, New York 253 BRODEUR, RICHARD D., and WILLIAM G. PEARCY. Food habits and dietary overlap of some shelf rockfishes (genus Sebastes) from the northeastern Pacific Ocean 269 COLVOCORESSES, J. A., and J. A. MUSICK. Species associations and community composition of Middle Atlantic Bight continental shelf demersal fishes 295 HAYNES, EVAN B. Early zoeal stages of Placetron wosnessenskii and Rhmolith- odes wosnessenskii (Decapoda, Anomura, Lithodidae) and review of lithodid larvae of the northern North Pacific Ocean 315 ZIMMERMAN, ROGER J., THOMAS J. MINELLO, and GILBERT ZAMORA, Jr. Selection of vegetated habitat by brown shrimp, Penaeus aztecus, in a Galveston Bay salt marsh 325 STANDARD, GARY W., and MARK E. CHITTENDEN, Jr. Reproduction, move- ments, and population dynamics of the banded drum, Larimus fasciatus, in the Gulf of Mexico 337 TETTEY, ERNEST, CHRISTOPHER PARDY, WADE GRIFFIN, and A. NELSON SWARTZ. Implications of investing under different economic conditions on the profitability of Gulf of Mexico shrimp vessels operating out of Texas 365 BUCK, JOHN D. Quantitative and qualitative bacteriology of elasmobranch fish from the Gulf of Mexico 375 BOTTON, MARK L., and HAROLD H. HASKIN. Distribution and feeding of the horseshoe crab, Limulus polyphemus , on the continental shelf off New Jersey .... 383 PEARCY, W, T NISHIYAMA^ T FUJII, and K. MASUDA. Diel variations in the feeding habits of Pacific salmon caught in gill nets during a 24-hour period in the Gulf of Alaska 391 RUGGERONE, GREGORY T, and DONALD E. ROGERS. Arctic char predation on sockeye salmon smolts at Little Togiak River, Alaska 401 Notes MAULE, ALEC G., and HOWARD F. HORTON. Feeding ecology of walleye, Stizostedion vitreum uitreum, in the mid-Columbia River, with emphasis on the interactions between walleye and juvenile anadromous fishes 411 (Continued on next page) Seattle, Washington 1984 Uborate; APR 1 7 1985 For sale by the Superintendent of Documents, U.S. Government Printing Office, Washing- ton DC 20402 — Subscription price per year: $21.00 domestic and $26.25 foreign. Cost per single issue: $6.50 domestic and $8.15 foreign. Contents — continued ROCKETTE, MARK D., GARY W. £¥A*ffiARD, and MARK E. CHITTENDEN, Jr. Bathymetric distribution, spawning periodicity, sex ratios, and size compositions of the mantis shrimp, Squilla empusa, in the northwestern Gulf of Mexico 418 CROWE, BARBARA J. Distribution, length-weight relationship, and length- frequency data of southern kingfish, Menticirrhus americanus , in Mississippi .... 427 RADTKE, RICHARD. Scanning electron microscope evidence for yearly growth zones in giant bluefin tuna, Thunnus thynnus, otoliths from daily increments . . . 434 PAYNE, P. MICHAEL, and DAVID C. SCHNEIDER. Yearly changes in abun- dance of harbor seals, Phoca vitulina, at a winter haul-out site in Massachusetts . 440 GOLDBERG, STEPHEN R., VICTOR HUGO ALARCON, and JUERGEN ALHEIT Postovulatory follicle histology of the Pacific sardine, Sardinops sagax from Peru 443 SHIMEK, RONALD L., DAVID FYFE, LEAH RAMSEY, ANNE BERGEY, JOEL ELLIOTT, and STEWART GUY. A note on spawning of the Pacific market squid, Loligo opalescens (Berry, 1911), in the Barkley Sound region, Vancouver Island, Canada 445 EPPERLY SHERYAN P., and WALTER R. NELSON. Arithmetic versus expo- nential calculation of mean biomass 446 Notice NOAA Technical Reports NMFS published during last 6 months of 1983. The National Marine Fisheries Service (NMFS) does not approve, recommend or endorse any proprietary product or proprietary material mentioned in this pub- lication. No reference shall be made to NMFS, or to this publication furnished by NMFS, in any advertising or sales promotion which would indicate or imply that NMFS approves, recommends or endorses any proprietary product or proprietary material mentioned herein, or which has as its purpose an intent to cause directly or indirectly the advertised product to be used or purchased because of this NMFS publication. SIZE, AGE, SEXUAL MATURITY, AND SEX RATIO IN OCEAN QUAHOGS, ARCTICA ISLANDICA LINNE, OFF LONG ISLAND, NEW YORK John W. Ropes, Steven A. Murawski, and Fredric M. Serchuk1 ABSTRACT Ocean quahogs, A rctica islandica. were collected off Long Island. New York, in 1978 for a determination of sexuality and gonadal condition. A microscopic examination of histologically prepared tissues of 133 clams. 19-60 mm in shell length, revealed that 36 were in an undifferentiated condition and could not be sexed. Sexual differentiation was evident in 97 clams; of the latter, 69 were in two types of intermediate development: those with sparse (20) and moderate (49) tubule development. Only 28 clams were fully mature. Age and growth were assessed from acetate peels of shell cross sections. Determinations of sex of these, and of specimens 57-103 mm in shell length collected from the same area in 1980. indicated that the smallest and youngest ocean quahogs were predominantly male, but the largest and oldest were predominantly female. Ocean quahogs, Arctica islandica, like most other bivalves, lack external characteristics for a de- termination of sex, maturation, and gonadal con- dition. Sex determination has been made for other bivalves, such as the surf clam, Spisula solidis- sima (Ropes 1979a), from microscopic examina- tions of gametogenesis in histological prepara- tions of gonadal tissues. Similar examinations were lacking for ocean quahogs. The resource has become an important fishery within the past half-decade (Ropes 1979b; Serchuk and Murawski 19802). In most bivalves that have been studied, sexual maturity occurs at a young age and small size, but species differences have been observed (Altman and Dittmer 1972). Thompson et al. (1980a, b) found that the ocean quahog is a slow growing, long-lived species which exhibits considerable variability in maturation with respect to size and age. The latter conclusion was based on examina- tions of 39 specimens, 87% of which were 40 mm or longer in shell length. The samples were collected in April-May, 3-4 mo before the spawning period reported for this species by Loosanoff (1953). It seemed reasonable to assume that mature, older quahogs in the sample would produce large num- ■Northeast Fisheries Center Woods Hole Laboratory. National Marine Fisheries Service, NOAA, Woods Hole, MA 02543. 2Serchuk, F M.. and S. A. Murawski. 1980. Evaluation and status of ocean quahogs, Arctica islandica (Linnaeus) popula- tions off the Middle Atlantic coast of the United States. Woods Hole Lab. Ref. Doc. 80-32. 4 p. Northeast Fisheries Center Woods Hole Laboratorv, National Marine Fisheries Service, NOAA, Woods Hole. MA 02543. Manuscript accepted September 1983. FISHERY BULLETIN: VOL. 82. NO. 2. 1984. bers of sex cells, but it was not possible to deter- mine whether most of the undifferentiated gonads in the sample would do likewise. Their contribu- tion to the reproductive potential of the species was an enigma, and our knowledge of maturation was incomplete. In late July and early August 1978, the National Marine Fisheries Service marked large numbers of ocean quahogs at a location near a site sampled in the study of sexual maturity reported by Thompson et al. (1980b). This was an opportunity to collect specimens for a reexamination of gonadal condition at about the time of maximum ripeness, as Loosanoff (1953) had reported finding many ocean quahogs in the partial spawning con- dition in mid-August. The time of collection, then, seemed favorable for obtaining sexually mature quahogs with fully developed, ripe gonads that could be clearly separated from immature quahogs with undifferentiated sex cells in the gonads. METHODS A commercial clam dredge vessel, MV Diane Maria, was chartered for the marking project dur- ing 25 July-5 August 1978. The hydraulic clam dredge had a 100-in (2.54 m)-wide knife and was modified by lining the inner surfaces with 1/2-in (12.7 mm) square-mesh hardware cloth to retain small clams. Sample tows were of 4-5 min duration and usually resulted in a dredge filled with clams, shells, and bottom substrata. 253 *1 FISHERY BULLETIN: VOL 82. NO. 2 The sample site was 48 km SSE of Shinnecock Inlet, Long Island, N.Y., at lat. 40°21'N, long. 72°24'W, and 53 m deep. This location contained high densities and a wide size range of ocean quahogs and had a low probability of being dis- turbed by the fishery: criteria important for suc- cess in the marking experiment (Murawski et al. 1982). The wide size range of ocean quahogs found at and near the site included more small individu- als for a study of maturity than elsewhere in the Middle Atlantic Bight. Small quahogs ( ^ 65 mm shell length) were sorted from the catch during the marking opera- tion, and the soft bodies were immediately re- moved from the shells for preservation in Bouin's fixative; shells were saved and coded for reference to corresponding tissues. Slides of the gonadal tis- sues were prepared for microscopic examination using standard histological techniques. The clam bodies were cut dorsoventrally through the mid- section, and the anterior and posterior pieces of each clam were embedded to produce two sections for examination. The 6 /xm thick sections were stained in Harris' hematoxylin and eosin. Recog- nition of gametogenic stages was based on previ- ous studies of bivalve reproduction by Loosanoff ( 1953 ); Ropes and Stickney ( 1965 ); Ropes ( 1968a, b; 1971; 1979a ); Thompson et al. ( 1980b ); Jones ( 1981 1; and Mann (1982). The shells were processed for observation of internal age/growth lines in acetate peels by methods similar to those reported in Thompson et al. 1 1980a, b) and reported more fully by Ropes (1982)3. A radial section was made from the umbo to ventral margin of left valves, since these contain a single prominent tooth that Thompson et al. ( 1980a, b) found had growth lines corresponding in number to those in the valve. Proper orientation of the valve for sectioning to retain the umbonal por- tion and broadest tooth surface in the anterior portion of the valve was a critical procedural step. The sections were made on a low-speed saw and by a 10.2 cm diameter by 0.03 cm thick diamond wa- tering blade. The cut edges were hand polished on wetable carborundum paper (240, 400, and 600 grits) to remove saw marks, polished to a high luster on a vibrating lap machine charged with aluminum oxide, then etched in a \c/< HC1 solution for one min. Peels were produced by flooding the ;)Ropes, J. W. 1982. Procedures for preparing acetate peels of embedded valves of Arctica islandica for ageing. Woods Hole Lab. Ref. Doc. 82-18, 8 p. Northeast Fisheries (inter Woods Hole Laboratory, National Marine Fisheries Service. NOAA, Woods Hole, MA 02543 etched surfaces with acetone and applying 0.127 mm thick acetate film. After a 15-min drying period, the film was peeled off and sandwiched between glass slides. Peel images were enlarged on a microprojector to 40 x . Age/growth lines were counted and the exit location of each at the exter- nal edge was marked on the peel for a comparison with the external bands by placing the anterior valve portion on the peel image. This procedure clearly demonstrated correspondence between the number and location of internal lines and external bands. It also delimited sequential increments be- tween external bands for measurement to the nearest 0.1 mm with calipers. Periodic age/growth phenomena in the shells of ocean quahogs have been called "bands" for incre- ments of darker periostracum deposits on the ex- ternal shell surfaces and "lines" for those accreted in the shells. The latter have been identified as prismatic microstructures that demark bound- aries of growth increments (Ropes et al. in press); the external pigmented bands varied in intensity and width (from 0 to -2 mm). A slight concentric depression often outlined the shell shape in the bands and corresponded to the location of internal lines. This and the method of marking the acetate peel aided in measuring increments of growth. After completing the study of the gonadal tis- sues of small ocean quahogs, it was evident that the sex ratio of larger clams from the same area should be examined. Therefore, squashes of thawed gonadal tissues from 199 marked clams 57-103 mm shell length recaptured in August 1980 were examined microscopically at the laboratory for determination of sex. RESULTS Observations of Age The shells and acetate peels of 137 clams were examined. Bands on the external shell surfaces were not equally distinct for all clams in the sam- ple. The bands were widely separated for small clams, but crowded together at the ventral margin for large clams. A few shells had poorly defined bands, but lines in the peels aided in locating them. Age annuli formed during the earliest on- togeny of ocean quahogs are difficult to detect on the valve surface and must be carefully exposed in the sectioned shell. A quahog 20.0 mm in shell length had three barely detectable bands on the surface of its valves; the two most recent annuli in peels of the valve and hinge tooth were most obvi- 254 ROPES ET AL : SIZE. AGE. AND SEX OF OCEAN QUAHOGS OFF LONG ISLAND. NY ous and the first was confounded by a secondary incomplete line that had formed slightly later (Fig. la, b). The formation of secondary lines is not typical at a young age. The formation of a complete line is, however, important in detecting annul i. Three clams had shell abnormalities related to an injury An ocean quahog with six bands had a slight depression at the anterior end of the left valve that was not detected as unusual growth lines or increments in the peel of either the valve or tooth; the right valve showed no evidence of an injury (Fig. 2a, b). Another quahog had a deep indentation, and part of the ventral margin was missing in the left valve before band six had been formed, but the right valve showed a slight inden- tation and darkening as evidence of an injury ( Fig. 3a, b). The peel of the left valve showed age lines before and immediately after the site of the injury (Fig. 3c). The sixth annuli in the hinge tooth was very prominent (Fig. 3d). The valve of a quahog with seven bands had definite surface indenta- tions associated with annuli, and the hinge tooth showed regularly spaced growth increments (Fig. 4a, b, c). An injury was not clearly evident. The annuli in peels of all these clams were easily re- lated to bands on the valve surface for mea- surements of growth. For 9 clams (47.5-60.4 mm long), all annuli in the peels were counted, but only some bands were measured because those near the ventral margin were too crowded and poorly defined. The shells of 3 clams (39.7-64.0 mm long) pro- duced a confused pattern of lines in the ventral third of the peels and extensive ridging and poorly defined bands on the external valve surfaces (Fig. 5a, b, c). It was not evident that these clams had been injured, but they were omitted in analyses, since growth appeared to be aberrant. In all, 134 clams, 18.7-60.4 mm long and averaging 38.9 mm (S.D. ± 8.65), were used. 5mm i i i i 5 mm . FIGURE 1. — im I ' i i i FIGURE 2. — (a) Right valve of a 6-yr-old ocean quahog, Arctica islan- dica. 31.1 mm shell length, ibi Photomicrograph of the acetate peel image of the hinge tooth showing six annuli. 255 FISHERY BULLETIN: VOL. 82, NO. 2 •■^ 5mm i i i i ,500 um, 1 i i r i I FIGURE 3. — iai Right valve of a 6-yr-old ocean quahog, Arctica islandica, 33.1 mm shell length, lb) Sectioned anterior portion of the left valve showing injury, ic) Three serial photomicrographs of the acetate peel image. Arrows point to annuli formed before and after the injury. (d> Photomicrograph of the acetate peel image of the hinge tooth showing six annuli. Size measurements at age of the clams are shown in Figure 6. The mean shell length, one standard deviation from the mean, and range are given for clams 3-8 yr old. The bands on the shells and lines in the peels indicated rapid growth through age 8. From age 3 and a mean size of 23.4 mm, the clams increased about 5 mm in shell length each year to age 8 and a mean size of 46.1 mm. Thereafter, growth seemed to decrease in rate. The bands were well separated to age 13. The bands at the ventral margin of 14-yr-old and older clams were too indistinct for accurate mea- surements, but the growth lines in peels were clearly separated and easily counted. The oldest 14- to 18-yr-old specimens may have been the smallest and slowest growing individuals in their year classes, but mean lengths were not progres- sively smaller than means for clams 9-13 yr old. Thus, a significant bias was not clearly indicated in the selection of older specimens. 256 ROPES ET AL.: SIZE, AGE, AND SEX OF OCEAN QUAHOGS OFF LONG ISLAND, NY. £. E. 10. fclD c . > 1^ O *t\ — KjHkV ^ 299ft -c y^t. .« 0 \# ° y* -*-* tC &> V t>£ &Jy " c . A ■ ~ ^. ' ^ ^ * ^ "^ 0) CO > £ d> "~" -C "3 4-3 cu c a. 0/ ■-; 4-3 CO *3 3 C cu c CO CO CU o 4J X -l-> bo c~ a o *2 J3 c a. CO — bd y. o in S CJ o ;- s CO o +3> -C o ^ X 5 0- — CD ,Q -- CO > = So - c iri — o cc CO .J -C J^ — CO -^-, £ ;i Q o ./; o a E w c *^ — o i- ~ ^ a. SI -, c CU CO CO 3 cr ~ c c CO -3- cu . o CJ o ~ o Sh 3 >, C t^- C cfl CO c*-. c o cu cu _> > cu CO "cO > -*-> _c M w OS D o 257 FISHERY BULLETIN: VOL. 82, NO. 2 £_ in 4 J3 — H ^ -— 4-3 « S . o ■$ & 1 * *■ o> 0> ""T m a. a £ '3 o 01 - -E — ' ■w -a <*-< c ° 5 — ai 13 "° 0) >-. o-t; 1/ o ■y o . 4) V- o o CO OJ 0) c -B g 4_> N "3 =° -B s b- 2 ca -B i» to M M O *J - Cu .5 ■* r- O o ■- o w -B C Oh -5 _ bo X! b _ CS E e a,- 5s "cO .3 Sh o : "? c OJ in -5 > ^ 0; '« — C *J ~ +J *o CO ^ O < z ■ CO bo > 0 quah ioned B t3 cO a; 1; X 8 jj _C +-> f <*- be 0 Js be — CS 01 B x. ■- en — * _ 4> £ 0 % 0 *^ -f g S CO S^ 0) -4-> > <~ lo ° > CO *> -s -B S1 .5° 2 OS gi — *■« CO u 1 £ '. 0 lO -w 0 W J= « a P O 258 ROPES ET AL.: SIZE. AGE, AND SEX OF OCEAN Ql AH( >GS OFF LONG ISLAND, N.Y. 5 6.' 5.: 4C 30 20- 10- :■■' 2 1 10 1 3 29 27 W 26 14 ; 9 N = 134 X =3 89 SO = 8.65 MEAN RANGE 8 10 12 AGE IN YEARS 14 16 FIGURE 6. — Observed shell lengths at age of ocean quahogs, Arctica islandica, off Long Island, N.Y., late July-early August 1978. Observations of Gonadal Condition Gametogenesis Gametogenesis in pelecypod molluscs exhibits similar basic characteristics. Each reproductive cycle begins with the production of the smallest, earliest cells at the basement membrane of folli- cles or alveoli. These infiltrate the lumina during maturation. Spermiogenesis through meiotic di- visions is completed within male gonadal alveoli; oogenesis undergoes mitotic division of the oogonia and growth of the primary oocytes within the female gonadal alveoli. Oocytes may reach metaphase of the first meiotic division in the ducts of spawning females, but are blocked from com- pleting maturation until after spawning and sperm penetration (Raven 1958). Most pelecypods expel the ripe cells into the surrounding environ- mental water where fertilization and larval devel- opment occur. A few pelecypods, and most notably female oysters of the genus Ostrea, are exceptions, since the eggs are held in the inhalent cavity dur- ing fertilization and initial developmental stages (Yonge 1960). A reproductive cycle corresponds to the initiation and completion of gametogenic stages and spawning. Single annual cycles have been described for many pelecypods, including the ocean quahog, although biannual and continuous cycles have been discribed for others (Sastry 1979). In some species, such as the ocean quahog, succes- sive reproductive cycles begin at or soon after spawning; in others, activation of a cycle is de- layed and the gonads are considered to be in a quiescent or resting stage (Sastry 1979). The latter condition frustrates determination of sex, since secondary sexual characteristics are generally lacking in most pelecypods. Spermiogenesis Spermatogonia about 5.5 /xm in diameter are the initial germinal cells produced by male Arc- tica islandica during a mitotic phase of sper- miogenesis. Successive meiotic stages follow and include primary and secondary spermatocytes ( ~ 3.7 and 4.0 fj.m in diameter, respectively ), sper- matids (-2.2 /u.m), and flagellated spermatozoa. The respective cells proliferate into the lumina of alveoli. Sperm have conical heads —4.8 ^.m long. Oogenesis Oogonia are the initial germinal cells produced by female Arctica islandica during oogenesis. These are embedded in the basement membrane and are comprised of cytoplasm and a conspicu- ous nucleus or germinal vesicle with a basophilic nucleolus surrounded by a network of loose chro- matin. The distinction between oogonia, sper- matogonia, and other cells in the basement mem- brane is not obvious. Primary oocytes begin protruding from the basement membrane into the lumina of alveoli and retain an attachment with it during the growth stage. The large spherical, ve- sicular nucleus of primary oocytes is surrounded by a coarse cytoplasm containing granules of the golgi apparatus and acidophilic granules of pro- teid yolk (Raven 1958; Kennedy and Battle 1964). The nucleolus differentiates into an amphinu- cleolus with maturation. Mature oocytes appear free in the lumina of alveoli and are often of ir- regular shape and have a distinct vitelline mem- brane. Measurements of the diameter of 50 clearly spherical oocytes that were sectioned through the nucleus and amphinucleolus ranged from 49.4 to 65.0 /ttm and averaged 56.6 fxm. Thirty-six gonadal tissues were in an undif- 259 FISHERY BULLETIN: VOL. 82, NO 2 ferentiated condition (Table 1, Fig. 7a, b). Gonadal tubules were of small diamater, few in number, and surrounded by an abundant loose vesicular connective tissue. Gonia embedded in the germi- nal epithelium lacked definite cellular structures for sex determinations. The lumina of tubules were empty. Sex determinations were possible for 97 quahogs, but in most (69) the gonads appeared to be in an intermediate stage and not fully devel- oped. These latter tissues were separated into two categories: Those with either sparse or moderate tubule development. Differentiated gonads with sparse tubule devel- opment were characterized by a limited number of gametogenic cells, as well as a limited number of tubules. The 16 male tissues examined were pro- ducing a few sperm; the 4 female tissues examined were producing a few oocytes. Abundant loose ve- sicular connective tissue occurred between the widely spaced gonadal tubules. In males, sper- matogenic cells at the germinal epithelium were about one layer thick, but were absent in portions of the epithelium (Fig. 8a, b). Some sperm were in close contact with the spermatogenic cells and a few were scattered in the lumina of tubules. In females, the few small oocytes occurred at the germinal epithelium, none were in the tubule lumina, and all were in an early developmental stage (Fig. 8c, d). For differentiated gonads with moderate tubule development, 39 males examined were producing sperm, while 10 females examined were producing oocytes. The gonadal tubules were more numerous than in gonads of sparse tubule condition, and some exhibited an expanded alveolar condition. Loose vesicular connective tissue clearly sepa- rated the tubules. In males, several layers of sper- matocytes proliferated from the germinal epithelium with some sperm forming a fringe ex- tending toward the empty lumina; however, por- tions of the germinal epithelium in some tubules TABLE 1. — Gonadal condition relative to age. sex. and size of three categories of ocean quahogs. An tua islandica — sexually immature, intermediate, and mature — off Long Island. N.Y.. late July-early August 197s. M = male; F = female. No clams (%] i Immature Intermediate Tubule development (undiffer- entiated) Spai se Moderate MatL ire Total M F M F M F no. Age (yO 2 K08) 1 3 4(3 0) 2(1.5) 2(1.5) 8 4 7(5.3) 5(3.7) 2(1.5) 14 5 11(8.2) 4(3.0) 1(0 8) 9(6.7) 1(0.8) 26 6 9(6.7) 3(22) (1 5) 10(7.5) 2(1.5) 1(0 8) 27 7 3(2 2) 2(1 5) 1(0 8) 12(9 0) 9(6.7) 2(1.5) 29 8 1(08) 3(22) 1(0.8) 5(3.7) 10 9 1(0.8) 1 10 1(0 8) 2(1.5) 3 11 1(08) 1(0 8) 2 12 1(0 8) 2(1 5) 3 13 1(0.8) 1 14 1(0 8) 1(0.8) 2 15 16 4(3.0) 4 17 18 2(15) 2 Age range 2-8 3-7 5-7 3-10 7-8 5-18 6-16 2-18 Mean 5.03 4 63 6 00 6.08 7 .10 9 79 13.22 6 71 Shell length (mm) 20 2 0 0 0 0 0 0 2 20-29 8 4 0 3 0 0 0 15 30-39 16 9 3 18 2 2 0 50 40-49 10 3 1 18 8 12 1 53 50-59 0 0 0 0 0 5 6 11 -59 0 0 0 0 0 0 2 2 Length range 19-46 21-44 36-42 20-48 39-45 36-58 41-60 19-60 Mean 34.4 33.8 384 37.2 41.8 47.1 55.0 39.0 Total no. 36(27.1) 16(12 0) 4(3.0) 39(29.3) 10(7.5) 19(14.3) 9(6.8) '133 'The tissues of a 21.1 mm. 3-yr-old clam were too poorly prepared for examination 260 ROPES ET AL.: SIZE, AGE, AND SEX OF OCEAN QUAHOGS OFF LONG ISLAND, NY. i* c? IOO>im • t **& • lie if • * ■ * o .■?--«. aft ■ j» . ,, - S . **. lOjJrn- FIGURE 7. — (a) Undifferentiated gonadal tissue section from a 5-yr-old ocean quahog, Arctica islandica, 37.2 mm shell length. (b) Enlargement of a gonadal tubule from the same clam. I \ s X >:« lift ■: *•*'*' *"" ►• » ctf\*. > if.* • < » J II • / i 9tf » •*. L_l lOO^m >»«i :v? f lOjjm FIGURE 8. — (a) Differentiated gonadal tissue section in the sparse condition from a 3-yr-old male ocean quahog, Arctica islandica, 21.0 mm shell length, (b) Enlargement of spermiogenesis in a portion of a gonadal tubule, (c) Differentiated gonadal tissue section in the sparse condition from a 5-yr-old, 37.5 mm shell length, ocean quahog. (d) Enlargement of oogenesis in a gonadal tubule. 261 FISHERY BULLETIN: VOL. 82, NO. 2 again lacked obvious spermatogenic cells (Fig. 9a, b). Oocytes in females were at the same stage of development as seen for females with sparse gonadal tubules, but more were growing from the germinal epithelium and some portions of the germinal epithelium lacked obvious oogenic cells (Fig. 9c, d). The sexually mature condition was found in 19 males and 9 females. In these quahogs the tubules were greatly expanded and filled the gonadal area; little connective tissue occurred between adjacent tubules. Developmental stages similar to those de- scribed for other bivalves by Ropes and Stickney ( 1965) were recognized. Two males and one female were in an early gonadal condition. Sper- miogenesis and oogenesis had cellular charac- teristics as in gonads of moderate tubule develop- ment, but the tubules were more numerous and crowded together. Six males were in a late gonadal condition. Primary and secondary spermatocytes and spermatids were proliferating from the ger- minal epithelium, filling about half of the tubules and sperm crowded into the lumina. No females were found in the late gonadal condition, but 11 males and 2 females were in an advanced late stage. In males, spermatocytes and spermatids proliferated from the germinal epithelium and sperm predominated in the lumina of the tubules (Fig. 10a, b). In females, oocytes crowded into the lumina of tubules and a few seemed to be attached to the germinal epithelium. No ripe males and only six ripe females with numerous ripe oocytes crowding into the tubules were found (Fig. 10c, d). The potential for developing large numbers of germinal cells was most evident and indicative of full sexual maturity in all of these quahogs. a ,< , v . % :S ■ m V \ *>. V I 0 /"** ■AS'4 3< .•SO*- 'ft** SI *? '■ V t. * lOO^m* d &-?§£# U u FIGURE 9. — (a) Differentiated gonadal tissue section in the moderate condition from a 7-yr-old ocean quahog, Arctico islandica, 42.9 mm shell length, (bi Enlargement of spermiogenesis in a portion of a gonadal tubule, ic) Differentiated gonadal tissue section in the moderate condition from an 8-yr-old female ocean quahog, 43.3 mm shell length, (d ) Enlargement of oogenesis in a portion of a gonadal tubule. 262 ROPES ET AL.: SIZE. AGE. AND SEX OF OCEAN QUAHOGS OFF LONG ISLAND, NY. I " m ■ ft ft 7 * i I > FIGURE 10. — i a i Differentiated gonadal tissue section in the mature condition from an 18-yr-old male ocean quahog, Arctica islandica , 57.8 mm shell length, (b) Enlargement of spermiogenesis in a portion of a gonadal tubule, (c) Differentiated gonadal tissue section in the mature condition from a 16-yr-old female quahog, 59.8 mm shell length, (d) Enlargement of ripe oocytes in a tubule. Gonadal Condition vs. Size and Age In an analysis of gonadal condition relative to age and size, quahogs in the undifferentiated, immature condition ranged from 2 to 8 yr old, averaged 5.0 yr old, and were from 19 to 46 mm long and averaged 34.4 mm (Table 1). This condi- tion was found in 27% of the gonads in the sample. For the three types of differentiated gonads, quahogs with sparse tubule development com- prised 15% of the sample. Males ranged from 3 to 7 yr old, averaged 4.6 yr old, and were from 21 to 44 mm long and averaged 33.8 mm; females ranged from 5 to 7 yr old, averaged 6.0 yr, and were from 36 to 42 mm long and averaged 38.4 mm. This category contained the smallest and youngest female in the sample: 38 mm long and 5 yr old. Quahogs with moderate tubule development comprised 37% of the sample. Males ranged from 3 to 10 yr old, averaged 6.1 yr, and were from 20 to 48 mm long and averaged 37.2 mm; females ranged from 7 to 8 yr old, averaged 7.1 yr, and were from 39 to 45 mm long and averaged 41.8 mm. This cate- gory contained the smallest and youngest male in the sample, which was 20 mm long and 3 yr old (Fig. la, b). Sexually mature quahogs comprised 21% of the sample. Males ranged from 5 to 18 yr old, averaged 9.8 yr, and were from 36 to 58 mm long and aver- aged 47.1 mm; females ranged from 6 to 16 yr old, averaged 13.2 yr, and were from 41 to 60 mm long and averaged 55.0 mm. The smallest mature quahog found was a male 36 mm long and 6 yr old, although a 5-yr-old, 41 mm long male was also mature; the smallest and youngest mature female found was 41 mm long and 6 yr old. None of the gonads contained germinal cells 263 FISHERY BULLETIN: VOL. 82, NO. 2 suggestive of ambisexuality. This is consistent with the conclusion of Loosanoff (1953) that the sexes are separate. The sex ratio, however, was particularly imbalanced in favor of males. In the 69 quahogs considered less than fully mature, 55 were males and 14 were females, while in the 28 sexually mature specimens, 19 were males and 9 were females; the observed ratios were 4:1 and 2:1, respectively. The data were subjected to goodness of fit tests under the hypothesis of a 1:1 ratio between the sexes; results indicated highly signif- icant (P<0.01) and significant (P<0.05) dif- ferences, respectively. Microscopic examinations of gonadal tissue squashes of the 199 clams collected in 1980 re- vealed an overall sex ratio of 96 males and 103 females. These results were not significantly dif- ferent from parity (1 male:1.07 female), but by separating the data into 10 mm size groups, a significant difference (P 0.05) in favor of males was indicated in the size group 80-89.9 mm, and a highly signficant difference (P < 0.01) in favor of females was indicated in the 100-110 mm size group (Table 2). Figure 11 shows the combined observations of clam size and sex obtained from the 1978 and 1980 samples. In these samples, males tended to decrease in occurrence relative to females with increasing shell size. TABLE 2. — Occurrence of male and female ocean quahogs, Arctica islan- dica, within 10 mm size groups off Long Island, N.Y., August 1980. too Size group (mm) Number Males Females 50-59 4 0 60-69 44 32 70-79 12 21 80-89 16 5 90-99 19 33 100-109 1 12 Total 96 103 DISCUSSION The time of sampling, sample size, and capture of small quahogs provided a basis for detection of the differentiated and sexually mature stage at younger ages and smaller sizes as compared with the study of Thompson et al. (1980b). In the pres- ent study, 5- and 6-yr-old quahogs 41 and 36 mm long, respectively, were considered sexually ma- ture; the youngest mature quahog reported by Thompson et al. (1980b) was a 42 mm male 11 yr old. The intermediate gonadal condition was 80 - -J 60 1 k 0. .' N= 170 d*o* 126 $9 1 :074 J_ _L _L L < 20-29 40-49 60-69 SIZE CROUPS 80-89 100-109 FIGURE 11. — Sex of ocean quahogs. A rctica islandica, relative to shell length (mm) in collections off Long Island, N.Y., 1978 and 19S0 found to occur at lower ages and smaller sizes than by Thompson et al. (1980b), and slightly smaller sizes were found for sexually mature quahogs. Variability in attainment of sexual maturity at age/size was observed in both studies. The onset of sexual maturity at young ages has been reported for several bivalves. The bay scal- lop, A rgopecten irradians, attains maturity at 1 yr; the hard clam, Mercenaria mercenaria, soft clam, Mya arenaria, and blue mussel, Mytilus edulis, matures at 1-2 yr ( Altman and Dittmer 1972). Surf clams, Spisula solidissima, from an inshore habitat showed precocious sexuality in a few post- larvae or juveniles; they spawned at 1 yr, but reached full maturity at 2 yr (Ropes 1979a). Sea scallops, Placopecten magellanieus, spawned at about 1.5-2 yr after forming the first growth ring (Naidu 1970). In apposition to more mature gonadal conditions, some scallops in his collec- tions were considered undifferentiated and dif- ferentiated male and female immature specimens. Lucas (1966) observed precocious sexuality in a scallop (Chlamys varia) and two clams (Glycymeris glycymeris and Venus striatula) from waters off France. The development of the repro- ductive potential during the early life history of these several bivalves seems consistent with esti- mates of their life span, which are as short as 2 yr for the bay scallop and as long as 30 yr for the surf clam (Belding 1906; Ropes 1979a). In contrast, the present study revealed that ocean quahogs attain maturity at 5-10 yr of age, and Thompson et al. 264 ROPES ET AL.: SIZE. AGE, AND SEX OF OCEAN QUAHOGS OFF LONG ISLAND. NY. (1980a) reported a longevity of about 150 yr. They found that growth was vigorous at old age and that there were no obvious indications of reproductive senility. A small abyssal nuculoid bivalve, Tin- daria call isti for mis, studied by Turekian et al. (1975) seems most exceptional with regard to age and size at sexual maturity They found a longev- ity of about 100 yr for a large specimen (8.4 mm shell length) by radiometric techniques and counts of shell growth bands, but gonadal devel- opment was not recognized until the clams were about 4 mm long and 50-60 yr old. The attainment of sexual maturity about midway in the life span of Tindaria sets it apart from other species that re- produce at a younger age. Nevertheless, all have the potential to reproduce for many years. Repro- duction during a long life span of a species may be an evolutionary strategy in response to uncertain larval and juvenile survival (Krebs 1972). Repro- duction during a particularly long life span is most obvious for Arctica islandica. For the 69 gonads containing sexually differen- tiated germinal cells and sparse-to-moderate tubule development, some morphologically ripe sperm were present. In contrast, oogenesis never exceeded an early developmental state. Jones (1981), Loosanoff (1953), von Oertzen (1972), and Mann (1982) reported that mature ocean quahogs spawn each year. Thus, the sperm may be spawned, but the fate of the oocytes remains an enigma. In American oysters, Crassostrea vir- ginica, germinal cells remaining in the gonads after spawning are reabsorbed (Galtsoff 1964), but viable, nearly ripe, or ripe germinal cells may be retained by hard clams throughout the fall, winter, and into the following spring (Loosanoff and Davis 1951). Thus, bivalves appear to differ greatly in this respect. No conclusion can be drawn relative to retention of germinal cells after spawn- ing for ocean quahogs which were intermediate between the immature and mature condition in the absence of collected data. Gonadal development in 28 mature clams suggested that many (46% ) were approaching ripeness or were ripe (21% ). Later development probably resulted in a spawning which was begun in late August-September. This seems reasonable based on observations by Mann (1982) of the re- productive cycle of Arctica islandica from sample locations in Block Island Sound. At the beginning of his study in September 1978, most (69% ) were in the partially spent or spent condition and spawn- ing was indicated until mid-November. An exact correspondence of the time and duration of spawn- ing may be a hazardous assumption, since the two sample sites are about 110 km apart and some of the samples taken by Mann (1982) were at shallow depths (36 m). A disparity in the initiation of gametogenesis was observed between the sexes. Male ocean quahogs began producing germinal cells at a smaller size and younger age than females. This suggests that females require a longer period of development and growth. The later development of female sexuality is a probable explanation for the highly significant difference obtained in tests of the sex ratio of quahogs in the intermediate gonadal condition. The significant difference ob- served for fully mature quahogs may be due to the small number in the sample (Dixon and Massey 1957), but Jones (1981) observed a similar dispar- ity (P = 0.008) for quahogs > 75 mm from offshore New Jersey. In his collections 184 were males and 136 were females, a ratio of 1:0.74. Mann (1982) examined ocean quahogs that were mostly 80-100 mm long and found 185 males and 169 females, a ratio of 1:0.91. These observations suggest that spatial variation may occur in the sex ratio of ocean quahog populations, but that males are more numerous than females. Pelseneer (1926) investigated the sex ratio of several mollusc species, including bivalves. He found more females among the older individuals of some populations and the converse among younger individuals. Coe ( 1936) recognized the existence of such disparities in molluscs and proposed the fol- lowing hypotheses as possible explanations: 1) That males have a shorter longevity than females, because of a differential mortality rate or less re- sistance to unfavorable environmental conditions; 2) that the development of alternative sexual con- ditions is environmentally determined; and 3) that sex change may occur. Loosanoff (1953), von Oertzen ( 1972 ) , Thompson et al .( 1980b ) , and Jones (1981) all considered the species to be strictly dioecious, as did Mann (1982), although he found two hermaphrodites. These are anomalous, "acci- dental functional hermaphrodites" by the ter- minology of Coe (1943). Although Sastry (1979) hypothesized that a failure in the genetic sex- differentiating mechanism may produce some hermaphrodites, he found no evidence of a phenotypic or genetic basis for sex determination in pelecypods. It is unlikely that ocean quahogs are protandric. This condition in a typically hermaphroditic species is characterized by the development of male organs or maturation of their products before 265 FISHERY BULLETIN: VOL. 82, NO. 2 the appearance of corresponding female products. In Ostrea lurida, for example, spermatogonia are proliferated first throughout the follicles, but be- fore the sperm mature oogonia have developed into numerous oocytes in the same follicles and the gonad has a definite intersexual character (Coe 1932). More than 90c/c of the young oysters exhibit the bisexual condition and no strictly male or female specimens occur. Old oysters in the female phase retain sperm balls and spermatogonia, and those in the male phase retain large and small oogonia. The two anomalous ocean quahogs found by Mann (1982) were examples of bilateral her- maphroditism, i.e., the germinal cells for each sex were in separate follicles. None of the inves- tigators of the reproductive cycle in ocean quahogs suggested finding ambisexual conditions (Loosanoff 1953; von Oertzen 1972; Jones 1981; Mann 1982 ). Thus, the characteristic germinal cell development for protandry is lacking in ocean quahogs. Sex reversal in some molluscs has been linked to castration from parasites invading the gonads, but evidence of causality was considered inconclusive by Noble and Noble (1961) and Malek and Cheng ( 1974 ). Except for the occurrence of the commensal nemertean, Malacobdella grossa, in ocean quahogs (Gibson 1967; Jones 1979), parasites in the species have not been reported (Ropes and Lang 1975 )4. The causality of hermaphroditism in ocean quahogs, then, remains uncertain and evi- dence is unavailable that sex may be environmen- tally determined. The hypothesis that female ocean quahogs may live longer than males has some support from determinations of the sex of specimens recovered from the marking site in August 1980. Based on predicted ages of ocean quahogs at the marking site reported by Murawski et al. ( 1982 ), the largest and oldest notched ocean quahogs were predomi- nantly female. Since this may be atypical for the extensive population of ocean quahogs inhabiting the Middle Atlantic Bight, samples from other lo- cations are being examined to determine possible spatial variations. ACKNOWLEDGMENTS We gratefully acknowledge the technical assis- 4Ropes, J. W, and H. S. Lang. 1975. An annotated bibliog- raphy of the ocean quahog, Arctica islandica (Lin- naeus). Xeroxed manuscr., 67 p. Northeast Fisheries Center Woods Hole Laboratory, National Marine Fisheries Service, NOAA, Woods Hole, MA 02543. tance of Dorothy W. Howard and Cecelia S. Smith of the Northeast Fisheries Center Oxford Labora- tory, National Marine Fisheries Service, NOAA, Oxford, Md., for histological preparations of gonadal tissues; and Taina Honkalehto, Frances Lefcort, and Miranda Olshansky, student trainees from Smith College, Northampton, Mass., for as- sistance in preparing acetate peels of the shells of ocean quahogs. LITERATURE CITED ALTMAN, p. l. and d. s. Dittmer. 1972. Biological data book. 2d ed. Fed. Am. Soc. Exp. Biol., p. 154-155. BELDING, D. L. 1906. Preliminary report upon the shellfisheries of Mas- sachusetts. Comm. Fish. Game, Mass., 15 p. COE, W. R. 1932. Development of the gonads and the sequence of the sexual phases in the California oyster (Ostrea lurida i. Bull. Scripps. Inst. Oceanogr., Tech. Ser. 3:119-144. 1936. Sex ratios and sex changes in mollusks. Mem. Mus. Hist. Nat. Belg. 3:69-76. 1943. Sexual differentiation in mollusks. I. Pelecypods. Q. Rev. Biol. 18:154-164. DIXON, W. J.. AND F J. MASSEY, JR. 1957. Introduction to statistical analysis. 2d ed. McGraw-Hill. N.Y.. 488 p. GALTSOFF, P. S. 1964. The American Oyster, Crassostrea virginica Gme- lin. U.S. Fish Wildl. Serv, Fish. Bull. 64, 480 p. GIBSON, R. 1967. Occurrence of the entocommensal rhynchocoelan, Malacobdella grossa , in the oval piddock, Zirfaea cnspata , on the Yorkshire coast. J. Mar. Biol. Assoc. U.K. 47:301- 317. JONES, D. S. 1979. The nemertean, Malacobdella grossa, in the ocean quahog, A rcti lea islandica (Bivalvia). Nautilus 93:29-30. 1981. Reproductive cycles of the Atlantic surf clam Spisula solidissima , and the ocean quahog Arctica islandica off New Jersey. J. Shellfish Res. 1:23-32. KENNEDY, A. V, AND H. I. BATTLE. 1964. Cyclic changes in the gonad of the American oyster, Crassostrea virginica (Gmelin). Can. J. Zool. 42:305-321. KREBS, C. J. 1972. Ecology; the experimental analysis of distribution and abundance. Harper and Row, Publ., N.Y., 694 p. LOOSANOFF, V L. 1953. Reproductive cycle in Cyprina islandica. Biol. Bull. (Woods Hole) 104:146-155. LOOSANOFF, V. L., AND H. C. DAVIS. 1951. Delaying spawning of lamellibranchs by low tem- perature. Sears Found., J. Mar. Res. 10:197-202. LUCAS, A. 1966. Manifestation precoce de la sexualite chez quelques mollasques bivalves. Estratlo Lav. Soc. Malacol. Ital. 3:153-158. Malek, E. a., and C. Cheng. 1974. Medical and economic malacology. Acad. Press, NY, 398 p. 266 ROPES ET AL : SIZE, AGE, AND SEX OF OCEAN QUAHOGS OFF LONG ISLAND, NY. Mann, R. 1982. The seasonal cycle of gonadal development in Arctica islandica from the Southern New England shelf. Fish. Bull.. U.S. 80:315-326. MURAWSKI, S. A., J. W. ROPES, AND F. M. SERCHUK. 1982. Growth of the ocean quahog. A rctica islandica, in the Middle Atlantic Bight. Fish. Bull., U.S. 80:21-34. NAIDU, K. S. 1970. Reproduction and breeding cycle of the giant scallop Placopecten magellanicus (Gmelini, in Port au Port Bay, Newfoundland. Can. J. Zool. 48:1003-1012. NOBLE. E. R.. AND G. A. NOBLE. 1961. Parasitology, the Biology of Animal Parasites. Lea and Febiger, Phila., 767 p. OERTZEN, J. A. VON. 1972. Cycles and rates of reproduction of six Baltic Sea bivalves of different zoogeographical origin. Mar. Biol. (Berl.) 14:143-149. PELSENEER, R 1926. La proportion relative des sexes chez les animau et particulierement chez les mollusques. Mem. Acad. R. Belg., Classe Sci. 8:1-258. RAVEN, C. R 1958. Morphogenesis: The analysis of molluscan devel- opment. Pergamon Press, N.Y., 311 p. ROPES, J. W. 1968a. Reproductive cycle of the surf clam, Spisula solidis- sima, in offshore New Jersey. Biol. Bull. (Woods Hole) 135:349-365. 1968b. Hermaphroditism in the surf clam, Spisula solidis- sima. Proc. Natl. Shellfish. Assoc. 58:63-65. 1971. Maryland's hard clam studied at Oxford labora- tory. Chesapeake Bay Affairs, Commer. Fish. News 4(6):2-3. 1979a. Shell length at sexual maturity of surf clams, Spisuhi solidissima, from an inshore habitat. Proc. Natl. Shellfish. Assoc. 69:85-91. 1979b. Biology and distribution of surf clams (Spisula so- lidissima) and ocean quahogs (Artica islandica) off the Northeast Coast of the United States. Proc. Northeast Clam Ind.: Management for the future, p. 47-66. Univ. Mass. and Mass. Inst. Tech., Sea Grant Prog. SP-112. ROPES. J. W, AND A. P. STICKNEY. 1965. Reproductive cycle of Mya arenaria in New En- gland. Biol. Bull. (Woods Hole) 128:315-327. ROPES, J. W, D. S. JONES, S. A. MURAWSKI, F M. SERCHUK, AND A. JEARLD, JR. 1984. Documentation of annual growth lines in ocean quahogs, Arctica islandica Linne. Fish. Bull., U.S. 82:1-19. SASTRY, A. N. 1979. Pelecypoda (excluding Ostreidae). In A. C. Giese and J. S. Pearse (editors), Reproduction of marine inver- tebrates, Vol. V, p. 113-292. Acad. Press, N.Y. THOMPSON, I., D. S. JONES, AND D. DREIBELBIS. 1980a. Annual internal growth banding and life history of the ocean quahog Arctica islandica (Mollusca: Bival- via). Mar. Biol. (Berl.) 57:25-34. THOMPSON, I., D. S. JONES, AND J. W ROPES. 1980b. Advanced age of sexual maturity in the ocean quahog Arctica islandica (Mollusca: Bivalvia). Mar. Biol. (Berl.) 57:35-39. TUREKIAN, K. K., J. K. COCHRAN, D. R KHARKAR, R. M. CER- RATO, J. R. VAISNYS, H. L. SANDERS, J. F GRASSLE, AND J. A. ALLEN. 1975. Slow growth rate of a deep-sea clam determined by 228Ra chronology. Proc. Natl. Acad. Sci. 72:2829-2832. YONGE, C. M. 1960. Oysters. Collins, Lond, 209 p. 267 FOOD HABITS AND DIETARY OVERLAP OF SOME SHELF ROCKFISHES (GENUS SEBASTES) FROM THE NORTHEASTERN PACIFIC OCEAN Richard D. Brodeur and William G. Pearcy1 ABSTRACT Euphausiids were the major food of five co-occurring species of rockfishes I Sebastes spp. ) along the west coast of North America from Vancouver Island to northern California. Copepods, decapods, cephalo- pods, amphipods, fishes, and other pelagic prey were also consumed but were less important to the overall diet. Two species, S. flavidus and S. diploproa, were relatively euryphagous, utilizing a high number of prey taxa. The other species, S. pinniger, S. alutus, and S. crameri, had a more restricted diet comprised mostly of euphausiids. The numerical composition of prey in the diet of all species was similar due to the preponderance of the two dominant euphausiid species. Diet overlaps based on weight composition were high for S. pinniger, S. diploproa, and S. alutus but were moderate for most comparisons involving S. flavidus and S. crameri. The diets of S. flavidus and S. pinniger were examined in more detail to explain some of the vari- ability associated with their food habits. Both species exhibited peak feeding periods at the same time during the day. They consumed about the same mean size of prey, although S. flavidus consumed a wider size range of prey. Size of prey and dietary composition did not vary much with size offish. There were significant seasonal, geographical, and diel differences in food composition for both species, which may be a function of varying food availability. Factors that allow coexistence of a large number of morphologically similar species have been the focus of numerous studies and continued debate in the ecological literature. Competition and re- source partitioning have been reviewed in general by Schoener (1974), and for fishes by Helfman (1978). Potential competition for resources is thought to be most common in three aspects of the ecological niche in fish communities: habitat, food, and time of activity (Tyler 1972; Bray and Ebeling 1975; Ross 1977; Werner 1979; Larson 1980; McPhersonl981). Rockfishes (Sebastes spp.) of the family Scor- paenidae are, a priori, interesting subjects for examining the various modes of resource parti- tioning. This genus is extremely speciose, with about 100 species reported from the North Pacific Ocean. At least 69 of these species are known to occur in the eastern North Pacific (Chen 1975). In addition to the large number of species, rockfishes also exhibit a high degree of overlap in their geographical distributions, with as many as 50 species occurring in a narrow latitudinal band (lat. 34°-38°N) off central California (Chen 1971). Several of these congeners are morphologically 'School of Oceanography, Oregon State University, Marine Science Center, Newport, OR 97365. similar and occupy similar habitats, so the poten- tial for resource overlap and competition is high (Larson 1980). Many of these species are abundant enough and of sufficient size to contribute substantially to commercial trawl landings in the northeastern Pacific (Alverson et al. 1964; Alton 1972; Gabriel and Tyler 1980; Gunderson and Sample 1980). De- spite their abundance in the northeastern Pacific, relatively few quantitative studies exist on rock- fish feeding habits. Most of the studies to date have dealt with shallow-water, neritic species often taken in recreational fisheries or accessible to in situ observations and sampling by scuba divers (Gotshall et al. 1965; Larson 1972; Hobson and Chess 1976; Love and Ebeling 1978). Descrip- tions of the diet of offshore species of Sebastes generally either lack taxonomic or quantitative detail (Phillips 1964) or encompass limited geo- graphical area or collection times (Pereyra et al. 1969; Lorz et al. 1983). Skalkin (1964) and Somer- ton et al. (1978)2 described food habits of rock- fishes from the Bering Sea and Gulf of Alaska, far Manuscript accepted October 1983. FISHERY BULLETIN: VOL. 82, NO. 2, 1984. 2Somerton, D., F. Funk, K. Mesmer, L. J. Bledsoe, and K. Thornburgh. 1978. A comparative study of the diets of Pacific ocean perch [Sebastes alutus) and walleye pollock [Theragra chalcogramma ) in the Gulf of Alaska. NORFISH Tech. Rep. NPB8, Wash. Sea Grant, 25 p. 269 FISHERY BULLETIN: VOL. 82, NO. 2 north of our study area which extends from off northern California to off Vancouver Island, Brit- ish Columbia. This study represents the first attempt to exam- ine broad geographical and seasonal patterns in food utilization and overlap by several commer- cially important species of rockfish on the outer continental shelf. The species considered include the yellowtail rockfish, Sebastes flavidus; canary rockfish, S. pinniger; Pacific ocean perch, S. alutus; splitnose rockfish, S. diploproa; and the darkblotched rockfish, S. crameri, all important members of the demersal shelf rockfish complex (Gabriel and Tyler 1980). In addition, variability in the diet of two of these species, S. flavidus and S. pinniger, was examined for the purpose of determining the effects of factors such as sea- son, geographic area, time of capture, and pred- ator size. MATERIALS AND METHODS Sampling Methods The food utilization patterns of the five rockfish species were determined by examining stomach contents. Fishes were obtained by two different survey methods (hereafter referred to as the summer and seasonal surveys). As the collection methods differ, they will be discussed separately The laboratory methods are similar and will be presented together. Summer Survey Methods Collections for the summer survey were made during the National Marine Fisheries Service (NMFS) 1980 West Coast Survey which took place from 12 July to 28 September 1980. The purpose of this survey was to assess the distribution and abundance of commercially important rockfishes. The area encompassed by the survey included much of the continental shelf and inner slope (ranging in depth from 55 to 366 m) between Monterey, Calif., (lat. 36°48'N) and the northern end of Vancouver Island, British Columbia (lat. 50° 00' N). Two commercial stern trawlers, the FV Mary Lou and the FV Pat San Marie, were utilized for the survey. A Nor'Eastern3 high- opening bottom trawl with an estimated 13.4 m horizontal and an 8.8 m vertical mouth opening was used on both vessels. The main body was constructed of 127 mm stretched mesh with 89 mm mesh in the cod end. The cod end also contained a 32 mm mesh liner. Half-hour tows were made at random depth-stratified stations chosen by a method described in Gunderson and Sample (1980). The majority of the stomach samples used in this study were collected in August and Septem- ber from north of lat. 43° N (Table 1, Fig. 1). Complete station data are given in Brodeur (1983). Stomachs were removed at sea from a random subsample of the catch of the five target species (Table 1). Sebastes pinniger and S. flavidus were the primary target species, and stomachs of these species were collected first and the other species sampled as time allowed. Altogether, 480 stom- achs were collected during the survey, all from adult fish ( > 200 mm FL). Fork length (measured to the nearest millimeter) and sex were recorded for all fish sampled, and stomachs were then removed, individually wrapped and labeled, and preserved in a 10% Formalin-seawater mixture. The intestinal tracts of many of the fish were examined at sea but few contained any recogniz- able food and none were retained. Total elapsed time between bringing the fish on board and preserving the stomachs was < 1 h. The oral cavities of all fish were examined for signs of stomach eversion and regurgitation; any fish showing such signs were discarded. Individual fish weights were not recorded at sea but were later calculated using the length-weight relation- ships of Phillips (1964). TABLE 1. — Number of rockfish stomachs analyzed from the 1980 National Marine Fisheries Service summer survey. The approximate latitudinal ranges covered by each leg were I, lat. 37°-42°N; II, lat. 43°-46°N; III, lat. 46°-50°N. Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Leg Sampling dates Species Number 1 12-20 July S pinniger S flavidus 9 8 17 II 4-29 Aug. S pinniger S. flavidus 85 127 S. alutus 54 S. diploproa 52 S. crameri 30 348 III 4-28 Sept S pinniger S flavidus 36 50 S. alutus 19 S diploproa 10 115 Total number analyzed 480 270 BRODEUR and PEARCY: FOOD HABITS AND DIETARY OVERLAP OF SOME SEBASTES 48 FIGURE 1. — Location of sampling sta- tions from which stomach collections were taken. + sign denotes collections made during the National Marine Fish- eries Service's summer survey and the stippled area (inset) shows the sampling area on Heceta Bank of the Oregon De- partment of Fish and Wildlife's seasonal collections. All depth contours are in meters. 4t> 42" CALIFORNIA t l( EUREKAo^, A I 'I 271 FISHERY BULLETIN: VOL. 82, NO. 2 Seasonal Survey Methods Stomachs for the seasonal study were collected during rockfish surveys conducted by the Oregon Department of Fish and Wildlife (ODFW) on Heceta Bank off the central coast of Oregon. These surveys obtained hydroacoustic and environmen- tal data along with the trawl catches. A total of 317 stomach samples was collected during seven sur- veys conducted in 1980-81 (Table 2). All surveys used trawling gear similar to that used in the summer surveys. Locations of the tows were chosen on the basis of high concentrations offish found during acoustic surveys over the outside edge of Heceta Bank between lat. 44°20'N and 44°00'N between the 128 m and 238 m bathymetric contours (inset, Fig. 1). The duration of tows was variable but averaged < 1 h. No tows were attempted at night because of the lack of acoustical targets near the bottom at this time. Stomachs were collected as described earlier. TABLE 2. — Number of rockfish stomachs analyzed from the seasonal Oregon Department of Fish and Wildlife collections on Heceta Bank. All dates are in 1980 unless otherwise noted. Vessel Cruise Sampling dates Species Number Ronnie C 1 23-24 April S, pinniger 42 Bay Islander 1 17-18 June S pinniger 24 Queen Victoria 1 15-16 July S pinniger S flavidus 47 16 Ronnie C II 26-28 Sept. S. pinniger S. flavidus 60 23 New Life 1 27 Oct. S. pinniger S. flavidus 21 2 Ronnie C III 17-18 Dec S pinniger S flavidus 33 25 New Life II 25 Jan. 1981 S- pinniger S flavidus 11 13 Total number analyzed 317 Analysis of Stomach Contents The stomachs were opened and their contents transferred to 50% isopropyl alcohol in the labora- tory. Contents were examined using a variable power dissecting microscope. Individual stomach fullness was estimated according to a subjective rating ranging from 0 (empty) to 5 (stomach fully distended with food). The condition of the contents was assigned a value from 0 (well-digested, barely identifiable to phylum) to 4 (fresh). Prey were identified to the lowest possible taxon and enumerated. In stomachs containing many small prey, such as euphausiids, any large or rare prey items were removed first. The remaining contents were then subdivided by means of a 272 Folsom plankton splitter (McEwen et al. 1954), and the contents of one subsample were used to estimate the stomach contents of small prey. The digested state of the contents of many stom- achs made precise counts of some prey difficult. Some paired parts of prey animals (e.g., eyes of euphausiids, otoliths of teleosts) were more resis- tant to digestion and total counts of these parts were halved to yield minimum counts of prey in- gested. Total lengths or greatest dimensions of intact prey found in the stomach were measured to the nearest 0.1 mm for the total sample (or a sub- sample of at least 15 individuals) using a stage ruler or ocular micrometer. All prey were blotted dry with absorbent paper and wet weights of each taxon were recorded to the nearest milligram. Analysis of Food Habits The minimum number of stomach samples needed to adequately describe the diet of a species was determined for all five rockfish species, using a cumulative prey species curve. A subset of stomachs of a particular species was randomly chosen and the cumulative number of unique prey taxa were then plotted versus the number of stom- achs which produced these taxa. The point on the abscissa where the curve begins to level off is considered the minimum number of stomachs nec- essary to describe the diet of that species. An example of the cumulative prey curves for the first 28 stomachs of each of the species in this study is shown in Figure 2. Although the curves assume different shapes, all approach an asymptote at sample sizes less than those analyzed. The contributions of the different prey items to the total diet of the rockfishes were expressed as percent frequency of occurrence, percent numeri- cal composition, and percent gravimetric composi- tion. Breadth and overlap were calculated for the five rockfishes from the summer surveys and for S. pinniger and S. flavidus from the seasonal surveys, using the pooled p;'s (relative proportion of the total number or biomass of resource i used by each species) for the major taxa. These include all taxa identified to at least generic level that exceeded Q.\c/< of the total weight or number of all identified foods. Resource breadth was computed for each species using the following formula: B = 1 -I Pi where B equals R (the total number of prey taxa BRODEUR and PEARCY: FOOD HABITS AND DIETARY OVERLAP OF SOME SEBASTES 25- 20- < x o < 10- 5- 0- s pinniger s flavidus s olutus s crameri s diploproa .._./. / 1 i i i T T T T 8 12 16 20 NUMBER OF POOLED STOMACHS 24 28 FIGURE 2— Cumulative prey curves for the first 28 stomachs of each of the 5 rock- fish species. in a food spectrum) when all items are in equal proportion in the diet (Levins 1968). These values were normalized as Bn = BIR, which ranges from 0 (most uneven distribution) to 1 (totally even distribution among the prey present). This index assumes equal availabilities of the different prey to all predators. Several indices of dietary overlap have been proposed and tested with known distributions of prey organisms (see Cailliet and Barry 19794; Linton et al. 1981; Wallace 1981). The coefficient of overlap described by Colwell and Futuyma (1974; identical to Schoener's (1970) index but not ex- pressed as a percentage) was chosen as it was found to be realistic for a wide range of true over- laps (Linton et al. 1981). This coefficient is as follows: dh = 1.0 -0.5(2, Pij - phj I where ptJ and phj are the proportions of prey j found in the diets of species i and h respectively. This coefficient has a minimum of 0 (no overlap 4Cailliet, G. M., and J. P. Barry. 1979. Comparison of food array overlap measures useful in fish feeding habits analysis. In S. J. Lipovsky and C. A. Simenstad (editors). Fish food habits studies, p. 67-79. Proc. 2d Pac. Northwest Tech. Workshop, Wash. Sea Grant. of prey) and a maximum of 1 (all items in equal proportions). Analysis of Diet Variations The sample sizes of S. pinniger and S. flauidus were sufficient to permit detailed analyses of their food habits, including seasonal, latitudinal, diel, and predator-size variations. The 368 specimens of S. pinniger and 264 of S. flavidus were grouped into 10 mm length catego- ries (Fig. 3). The distribution of S. pinniger lengths from the two surveys was similar and no significant differences in the means were found (Student's t-test; P > 0.05). Specimens of S. flavidus collected during the seasonal survey were significantly larger (P < 0.001) than those of the summer survey. Sebastes pinniger averaged about 40 mm larger than S. flavidus for both surveys combined. Corrections were made for this difference where appropriate in the analyses. To simplify the analysis of dietary variation in S. pinniger and S. flavidus, eight major types of prey were selected for comparison, based on their gravimetric importance or frequency of oc- currence. Numerical abundances were not used because of the great disparity in prey sizes en- countered and the problem of making counts on 273 FISHERY BULLETIN: VOL. 82, NO. 2 350 4O0 450 500 550 FORK LENGTH (mm) ■ i i r 600 650 x o 3 300 350 400 450 500 550 600 FORK LENGTH (mm) FIGURE 3. — Size distributions of Sebastes pinniger and S. Ilavidus from summer (National Marine Fisheries Service) and seasonal (Oregon Department of Fish and Wildlife) surveys. incomplete animals. These prey categories include the two most important euphausiid species and other major taxonomic groups (Table 3). Other planktonic prey (e.g., copepods, chaetognaths, pteropods) were occasionally present in the diet of one or both species, but their contributions to the overall diets were minor. Cephalopods did not TABLE 3. — The major prey categories used in the analysis of diet variations and their respective size ranges found in the stomachs of S. pinniger and S. flavidus. Prey Category size range (mm) Inclusive taxa or life stages Euphausia pacifica Thysanoessa spinifera Total euphausuds 8-26 8-30 8-30 juvenile and adult stages juvenile and adult stages above two and other species, Decapods 3-87 unidentified euphausuds adult shrimp, crab zoea and Amphipods Cephalopods' Fishes Gelatinous zooplankton 3-30 18-150 + 16-150 + 10-22 megalopae, shrimp mysis mostly hypernd but some gammand squid and octopods larvae, juvenile and adult stages ctenophores, thaliaceans, medusae, and siphonophores ' Found in S flavidus stomachs only. occur in the diet of S. pinniger; thus only seven prey categories were used for this species. We analyzed four factors that may affect the diet of these two species: season, geographic area, time of day, and size of fish. Each factor was subdivided into four classes to elucidate the gen- eral trends within each factor. Stomach content data for all cruises were grouped into four sea- sons, based on major periods in the hydrographic regime on the continental shelf off Oregon (Huyer et al. 1975; Huyer 1977): spring (March-May), summer (June-August), fall (September- Novem- ber), and winter (December-February). The collec- tion stations for all cruises were divided into one of four latitudinally defined shelf regions: North- ern California-Southern Oregon (lat. 41° 00' to 43°50'N), Heceta Bank-Central Oregon (lat. 43° 50' to 45°00'N), Columbia Region (lat. 45°00' to 47°00'N), and Northern Washington-Vancou- ver (lat. 47° 00' to50°00'N). For the analysis of diel variation of feeding, the local mean sampling time was adjusted to account for latitudinal, longitudinal, and seasonal differ- ences in daylight. Each collection time was stan- dardized to an equinox day with 12 h between sunrise and sunset, based on solar table values. These adjusted collection times were assigned to one of four time periods: morning (0800-1200 h), early afternoon (1200-1600 h), late afternoon (1600-1800 h), and night (1800-0700 h). Only a small number of S. pinniger and S. flavidus were collected at night despite extensive nighttime trawling effort on several occasions during the summer survey. Since the length distributions of the two species were roughly normal (Fig. 3), dividing the length range into four equal size groups would result in disproportionately large sample sizes in the middle size ranges. On the other hand, setting the sample sizes of the four groups equal would result in narrow size ranges around the mode. As neither of these options seemed desirable, compromise groupings were chosen. For S. pinniger, we used the following size classes: <45 cm, 45-<50 cm, 50- < 55 cm, and s 55 cm. Similar size classes were selected for S. flavidus but were offset 5 cm to reflect the smaller mean size of this species. To test whether significant within-factor varia- tion occurred in the diet of each species, contin- gency tables were constructed comparing the occurrence of food or a particular prey category versus the absence of food or that prey category. A variance test for homogeneity of binominally distributed data (Snedecor and Cochran 1967) was 274 BRODEUR and PEARCY: FOOD HABITS AND DIETARY OVERLAP OF SOME SEBASTES used for testing differences among the classes within each factor. Any comparisons which ex- ceeded the tabulated 0.05 \2 percentage caused a rejection of the null hypothesis of similar diets. RESULTS General Food Habits The results of the stomach content analysis are presented for both surveys and all five species in Tables 4 through 8. Each species will be discussed in detail in this section. Sebastes flavidus preyed on a diverse assem- blage of planktonic and micronektonic prey (Table 4). Dominating the diet in terms of frequency of occurrence (F.O.), percent by number, and, to a lesser extent, percent by weight were euphausiids, principally Euphausia pacifica and Thysanoessa spimfera. Many species of hyperiid amphipods were represented in the diet, but these were not numerous and did not comprise a major portion of the food on a weight basis. Decapods and cephalo- pods were moderately important in stomachs examined from both surveys. Copepods and larval decapods occurred only in the stomachs from the summer survey, while gelatinous zooplankton were found only in the seasonal study, and were common during late fall and winter. Fish were an important component on a weight basis; they were mainly mesopelagic species and juvenile stages of predominantly benthic species, although many adult Pacific herring, Clupea harengus pallasi, and some smelts were also found. The mean number of taxa and mean number of myctophids per stomach were higher in fish from the seasonal than those from the summer survey. Sebastes pinniger had a much more limited diet both in number of prey species and major prey categories consumed than S. flavidus (Table 5). Euphausiids were again the dominant prey con- sumed with proportional abundances and weights exceeding 90% of the total in both surveys. Many stomachs were distended with adult euphausiids (>1,000 individuals). Hyperiid and gammarid amphipods were common but did not appear to be important components of the diet. Mesopelagic fishes, including myctophids and stomiatoids, con- tributed to the biomass consumed during the fall and winter months of the seasonal survey. There was a low number of taxa represented in each stomach, especially in the summer survey. Because of the advanced stage of digestion of most of the stomach contents (mean digestion score = 1.05), many taxa were not identified to species in the stomachs of S. alutus, although many major prey categories were represented (Table 6). Euphausiids were the principal prey by weight and number. Of the remaining prey species, amphipods were relatively common and numerous. The oceanic shrimp, Sergestes similis, appeared in a significant number of stomachs and may constitute an important prey item. Remains of fishes were found in only a few stomachs, a noteworthy difference compared with the other four species examined. Sebastes diploproa utilized a spectrum of prey items as wide as that of S. flavidus, but the smaller mean size of this species is reflected in generally smaller prey taken (Table 7). Euphau- siids were less important, and amphipods, cope- pods, and decapods were more important on a numerical and percentage occurrence basis than for the other species. Sergestes similis contributed heavily in all respects and was found in almost half the stomachs examined. The small hyperiid amphipod, Vibilia propinqua, was common and numerous but contributed little to the bulk of the diet. The mean number of prey found per stomach was second only to the seasonal number of S. flavidus. The diet of S. crameri was characterized by very few prey taxa, perhaps because only 30 stomachs were examined (Table 8). Of these, one-third of the stomachs were empty and only about one-third of the total biomass found in these stomachs was identifiable, resulting in very low mean fullness and digestion scores (1.03 and 1.05, respectively). This identifiable fraction was composed of equal numbers of euphausiids, amphipods, and cope- pods. Euphausiids contributed a greater share to the total biomass, however, and completely domi- nated the identifiable contents. Few prey taxa were found, overall, in the stomachs of S. crameri. Diet Breadth and Overlap In order to quantify the relative food resource used by the various species, niche breadth mea- sures were calculated for all species. The principal prey types (proportional biomasses exceeding 1.07c of the total biomass), and niche breadth values (overall and normalized) are given in Table 9 for all species analyzed from the summer sur- veys and for S. pinniger and S. flavidus collected during the seasonal surveys. Sebastes flavidus utilized the greatest number of prey types (R), had the widest niche breadth 275 FISHERY BULLETIN: VOL. 82, NO. 2 TABLE 4. — Summary of yellowtail rockfish, Sebastes flavidus, stomach contents from the Oregon Department of Fish and Wildlife's seasonal and the National Marine Fisheries Service's summer samplings. F.O. = frequency of occurrence. Seasonal Summer FO (%) Number Weight (g) FO (%) Number Weight (g) Prey organism Mean % Mean % Mean % Mean % Euphausiacea Euphausia pacifica (juv.) Euphausia pacifica (adults) Thysanoessa spinifera (juv ) T. spinifera (adults) T longipes Thysanopoda acutifrons Euphausild unidentified Amphipoda Phronima sedentana Paraphronima gracilis Parathemisto pacifica Hyperia medusarum Hyperoche medusarum Streetsia challenger/ Vibilia propmqua Primno macropa Hyperndea unidentified Rhacotropis sp. Decapoda Sergestes similis Pandalus jordani Munida quadrispina (|uv) Pinnothendae megalopae Cancer sp megalopae Decapod mysis larvae Copepoda Calanus pacificus C marshallae Neocalanus sp. Euchirella sp Copepod unidentified Cephalopoda Abraliopsis felis Gonatus sp Loligo opalescens Japatella heathi Octopus sp. (juv.) Cephalopod unidentified Miscellaneous invertebrates Sagitta elegans bmacina helicma Alciopid polychaete Siphonophora Ctenophora Cnidana Osteichthyes Clupea harengus pallasi Thaleichthys pacificus Spirinchus starksi Stenobrachius leucopsarus Diaphus theta Tarletonbeania crenularis Symbolophorus californiensis Protomyctophum crocken Myctophidae unidentified Argyropelecus aculeatus Chauliodus macouni Nectoliparis pelagicus Lipandidae unidentified Stichaeidae unidentified (juv) Sebastes sp (juv) Glyptocephalus zachirus Lyopsetta exilis (juv) Psettichthys melanostictus (juv) Unidentified fish larvae Fish remains Unidentified animal remains 367 24.5 6.4 0.34 2.3 — — — 608 120 1 52.2 2 57 28.3 40 5 37 4 513 190 26.4 — — — — — 6.0 11.5 2.3 0 43 0 9 683 406 19.8 1.32 167 23 2 88 68 0 80 6.4 — — 0.5 1.0 — 0.01 — 1.3 1.0 — 0.12 — — — — — — 49.4 56.3 19 9 108 9.9 16 7 61.1 34 7 2 60 14 9 7.6 12 — 0.11 0.2 3.2 1.8 0.2 0.06 — 1.3 10 — 001 — 1.1 10 — 0.01 — 1.3 10 — 0.01 — 2.7 1.0 0.1 001 — 2.5 1.0 — 0.01 — 2.7 1.2 0.1 0 01 — 2.5 40 — 0.02 — 4.9 1.7 03 0 89 1.5 38 1.3 — 0.03 — 05 1.0 — 0.04 — 1.3 2.0 — 0.16 — 0.5 1.0 — 0.01 — 3.8 1.0 — 0.02 — 0.5 2.0 — 0 02 — 1.3 2.0 — 0.02 — — — — — 1.3 1.0 — 0.01 — — — — — — 7.6 2.5 0.1 0.71 1.0 2.7 2.8 0.2 0.75 0.7 1.3 12.0 0.1 370 0.9 1.1 1 0 — 5.19 1.9 3.8 93 0.2 0.22 02 27 5.6 0.5 0 .12 0.1 — — — — — 0.5 1.0 — 0.10 — 4.3 1.9 03 0 02 — 1.6 2.7 0.1 0.04 — — — — — 05 1 0 001 1.6 5.7 0.3 0.01 — 2.7 44 0.4 0.01 — — — — — — 0.5 10 — 0.02 — 22 4.5 0.3 001 — 1.3 1 0 — 0.93 02 1.3 30 — 068 0.2 1.1 10 — 057 — 38 1.0 — 21.24 149 22 1.5 0 1 2.26 1.7 1 6 1.0 — 068 0.3 6.3 26 0.1 071 08 6.5 6.5 0.3 1.33 2.1 11.4 1.2 0.1 1.82 3.8 22 2.2 0.2 234 1.7 2.5 2.5 — 0.16 — 05 10 — 0 03 — 1.3 1.0 — 0.01 — 5.4 1.5 0.3 0.04 — 1.3 10 — 0.27 — — — — — — 2.5 5.5 0.1 0.54 0.3 — — — — — 1.3 8 0 — 1.56 0.3 — — — — — 1.3 10 — 0 27 — — — — — — — — 3.8 1 6 0 1 14.17 184 1.3 10 — 1.22 0.3 — — — — — 2.5 2.0 — 1.65 0.7 0 5 10 — 0.28 — 1.3 1.0 — 0.84 — 0.5 1.0 — 0.99 0.2 2.5 1.0 — 1.97 0.9 0.5 2 0 — 9.96 18 5.1 1.7 0.1 4.94 4.6 — — — — — 1.3 10 — 0.07 — — — — — — 1.3 10 — 0.14 — — — — — — 11.4 1.3 0.1 1.37 2.9 0.5 1.0 — 0.71 0.1 1.3 10 — 2.49 0.6 — — — — 1.3 10 — 3 81 0.8 — — — — — — — 1.6 1.3 — 0.22 01 1.3 10 — 0.17 — — — — — — 1.3 20 — 0.44 0.1 1.1 10 — 026 01 2.5 10 — 0.37 0.2 1.1 2.0 — 1.07 0 4 1.3 10 — 0.88 0.2 — — — — — — — 1.1 1.5 — 0.29 01 — — 0.5 1.0 — 0.06 — — — — — — 1.1 1.0 — 0.14 — 152 — — 1.31 3.7 81 — 4.19 11.6 30.4 — — 066 3.7 38 4 — 0 64 84 276 BRODEUR and PEARCY: FOOD HABITS AND DIETARY OVERLAP OF SOME SEBASTES TABLE 4.— Continued. Predator characteristics Number of stomachs examined 79 Number of empty stomachs 4 Mean weight per stomach 5.192g ± 7 004 (SD) Mean total length 478 35 mm r 29 06 (SD) Mean fullness score: 2.87 Mean digestion score 290 Mean no. prey taxa per fish 281 185 38 2.905 g ± 6.032 (SD) 444 86 mm • 51 43 (SD) 1.92 1.95 1.55 TABLE 5. — Summary of canary rockfish, Sebastes pinniger, stomach contents from the Oregon Department of Fish and Wildlife's seasonal and the National Marine Fisheries Service's summer samplings. FO. = frequency of occurrence. Seasonal Summer Number Weight (q) Number Weight (g) FO (%) FO Prey organism Mean O o Mean °0 (%) Mean % Mean % Euphausiacea Euphausia pacitica (|uv ) 21.0 88 1 12.1 065 48 38 125 0.4 0.09 0 1 E pacitica (adults) 546 141 3 503 3.21 62.0 408 124 3 51.6 5 74 435 Thysanoessa spinilera 22.3 240 3.5 0.59 4.6 14 6 64 7 9.6 740 20.1 Thysanopoda sp 04 34 — 003 — — — — — — Euphausnd unidentified 374 139 6 34 0 1 54 20 4 26 1 141 5 37 7 579 28.1 Mysidacea Inusitatomysis sp — — — — — 08 2.0 — 0.02 — Amphipoda Parathemisto pacifica 08 2 0 — 0 01 — 46 1.2 — 0 01 — Hyperoche medusarum — — — — — 08 20 — 001 — Phronima sedentana 0.4 1 0 — 0.03 Streetsia challenger* 04 20 — 0 03 Hypemdea unidentified — — — — — 1.5 1.0 — 001 — Rhacotropis sp 04 4 0 — 005 — 1 5 60 0 1 007 — Atylus tndens 04 1 0 — 001 Anonyx sp 08 1 5 — 0 18 Lysianassidae unidentified — — — — — 0.8 1 0 — 0.02 — Decapoda Sergestes similis 2.9 1.7 — 0.09 0.1 1.5 140 0.2 1.89 0.6 Pandalus /ordani 04 1 0 — 1 05 0 1 1.5 1 0 — 1.55 0.4 Crangon sp — — — — — 08 1 0 — 0.03 — Munida quadnspina (juv.) 2.5 50 0.1 006 Chaetognatha Sagitia elegans 04 6.0 — 0.07 Osteichthyes Stenobrachius leucopsarus 0.8 1.5 — 0 78 0.2 08 10 — 0.59 0.1 Tarletonbeania crenulans 0.4 1.0 — 1 70 0.3 Myctophidae unidentified 1.3 1.3 — 1.47 0.6 Tactostoma macropus 0.4 1.0 — 1.73 0.2 Argyropelecus aculeatus 0.4 1.0 — 021 Ammodytes hexapterus 3.1 4 5 0.1 0 76 0.4 Sebastes /ordani 08 1.0 — 1904 56 Fish remains 84 — — 0.39 1.2 108 — — 300 6.0 Unidentified animal remains 122 — — 003 0.1 42.3 — — 0.09 0.7 Predator characteristics Number of stomachs examined 238 130 Number of empty stomachs: 39 18 Mean weight per stomach: 2 828 g ± 4.440 (SD) 5.385 g± 11 .297 (SD Mean total length. 191.45 mm ± 51.07 (SD) 504 07 mm ± 50.34 (SD) Mean fullness score. 202 1.68 Mean digestion score 1.89 1.55 Mean no. prey taxa per fish: 1.27 1.00 277 FISHERY BULLETIN: VOL. 82, NO. 2 TABLE 6. — Summary of Pacific ocean perch, Sebastes alutus, stomach contents from the National Marine Fisheries Service's summer sampling. F.O. = frequency of occurrence. TABLE 7. — Summary of splitnose rockfish, Sebastes diploproa, stomach contents from the National Marine Fisheries Service's summer sampling. F.O. = frequency of occurrence. Number of prey Weight of prey (g) Prey organism (%) Mean % Mean % Euphausiacea Euphausia pacifica 52.1 20.5 62.4 1 .12 63.1 Thysanoessa spinifera 19.2 7.1 80 0.47 9.8 Euphausnd unidentified 20.6 16.9 20.3 0.55 12.2 Amphipoda Phronima sedentaria 2.7 2.0 0.3 0.11 03 Paraphronima gracilis 1.4 1 0 — 0.02 — Parathemisto pacifica 6.8 3.2 1.3 0 03 0.2 Vibilia propinqua 1.4 11.0 09 0.12 0.2 Pnmno macropa 2.7 10 02 0.02 — Hyperndea unidentified 6.8 24 1.0 0.01 — Cyphocans challenger/ 2.7 1.0 0.2 0.03 0.1 Copepoda Neocalanus plumchrus 4.1 1.3 0.3 001 — Euchaeta sp. 2.7 3.0 0.4 0.01 — Decapoda Sergesles similis 20.6 3 1 3.7 0.34 7.5 Pasiphaea pacifica 1.4 1 0 — 0.03 — Decapod mysis larvae 1 4 1.0 — 0.01 — Crustacea remains 2.7 — — 0.19 0.6 Cephalopoda Loligo opalescens 1.4 1.0 — 0.53 0.8 Cephalopod unidentified 6.8 14 0.5 0.22 16 Osteichthyes remains 5.5 — — 0.04 0.1 Predator characteristics Number of stomachs examined: 73 Number of empty stomachs: 26 Mean weight per stomach 0923 g ± 1 954 (SD) Mean total length 365 36 mm ± 60.01 (SD) Mean fullness score 1.49 Mean digestion score: 1.05 Mean no prey taxa per fish: 1.68 (B), and had the most even distribution among prey types (Bn ) of all rockfish examined from the summer survey. Sebastes diploproa preyed on fewer taxa than S. flavidus but had moderately high overall and normalized food breadth values. Sebastes pinniger, S. crameri, and S. alutus uti- lized a similar number of distinct prey items and had similar breadth and evenness values with S. alutus having a more equitable distribution of prey than the other two. The seasonal results for the S. flavidus and S. pinniger were more divergent and represent the extreme values found among the species. Seven- teen principal prey types were important in the seasonal diet of S. flavidus, contributing toward a high B value. However, the dominance of a few species yielded a low evenness value for this species. Sebastes pinniger preyed on few taxa in fairly unequal proportions yielding fairly low niche breadth and evenness values. These low evenness values could be caused by the prepon- derance of euphausiids found in the guts of both species during the summer months. The individual overlap coefficients and the mean overlap for each species are presented for Number of prey Weight of prey (g) Prey organism (%) Mean % Mean % Euphausiacea Euphausia pacifica 46.8 26.5 41.2 1.53 42.1 Thysanoessa spinifera 145 29 1.4 0.16 1.4 Euphausnd remains 290 349 33.7 1 79 306 Amphipoda Parathemisto pacifica 1 6 1.0 — 001 — Hyperoche medusarum 32 1.0 — 0.01 — Paraphronima gracilis 3.2 1.0 — 002 — Streetsia challenger/ 1.6 1.0 — 0 02 — Vibilia propinqua 32.3 10.3 111 0.10 19 Pnmno macropa 3.2 10 — 0.01 — Hyperndea unidentified 97 1.5 0.4 0.02 0 1 Cyphocans challenger/ 4.8 1.7 03 0.02 — Lysianassidae unidentified 1.6 1.0 — 003 — Gammandea unidentified 1 6 1.0 — 003 — Isopoda unidentified 1.6 1.0 — 0.02 — Copepoda Neocalanus cristatus 6.5 3.2 0.7 0.02 0.1 Euchaeta elongata 48 33 0.5 0.03 0 1 Euchirella sp 3.2 1.5 0.2 001 — Candacia bipinnata 4.8 3.6 0.6 001 — Metndia sp. 32 1.0 0.1 0 01 — Decapoda Sergestes similis 46.8 44 68 0 60 165 Pasaphaea pacifica 1.6 1.0 — 0 59 0.6 Benthogenema burkenroadi 1.6 1.0 — 0 12 0.1 Munida quadnspina 1.6 60 0.3 0.12 0.1 Cancer sp. megalopae 97 1.3 04 002 0 1 Decapod mysis larvae 1.6 10 — 001 — Mollusca Pteropoda unidentified 1.6 10 — 0.03 — Gonatus sp. 1.6 1.0 — 0 07 — Octopus sp. (juv.) 1.6 10 — 0.17 0.2 Osteichthyes Stenobrachius leucopsarus 1.6 1.0 — 036 0.3 Myctophidae unidentified 6.5 1.0 0.2 0.13 0.5 Tactostoma macropus 1.6 1.0 — 2.28 22 Lipandidae unidentified 1.6 1.0 — 0.15 0.1 Fish remains 9.7 — — 0.38 0.1 Unidentified animal remains 17.7 — — 0.03 03 Predator characteristics Number of stomachs examined: 62 Number of empty stomachs: 15 Mean weight per stomach: 1 698 g ± 3 449 (SD) Mean total length: 264 82 mm ± 41 .82 (SD) Mean fullness score: 250 Mean digestion score: 1.25 Mean no prey taxa per fish: 2.48 both the weight and numerical abundance of prey in Table 10 for the summer surveys. As overlap indices are affected by the level of taxonomic specificity at which the prey have been identified, no unbiased means for testing the significance of these values are available. We adopted the con- vention that overlap values from 0.00 to 0.29 are considered low, 0.30 to 0.60 considered medium, and those above 0.60 show highly similar diets (Langton 1982). The coefficients for numerical composition show high values for all possible combinations except those involving S. crameri. Very similar propor- tions of the major euphausiid prey groups resulted in an extremely high overlap value (0.93) between 278 BRODEUR and PEARCY: FOOD HABITS AND DIETARY OVERLAP OF SOME SEBASTES TABLE 8. — Summary of darkblotched rockfish, Sebastes cra- meri, stomach contents from the National Marine Fisheries Service's summer sampling. F.O. = frequency of occurrence. Numbe rof Weighl of FO (%) prey 1 prey ( Mean 9) Prey organism Mean % % Euphausiacea Euphausia pacifica 133 80 372 042 26.2 Thysanoessa spimfera 3.3 1 0 1.2 0.06 09 Euphausnd remains 33 1 0 12 0.01 02 Amphipoda Parathemisto pacifica 16.7 44 256 0 01 1.0 Cyphocans challengeri 6.7 1.0 2.3 0.01 0.3 Lysianassidae unidentified 3.3 10 1 2 004 08 Copepoda Neocalanus cristatus 3.3 1 0 1.2 0 01 02 Euchaeta elongata 100 30 105 0 01 0.5 Copepod unidentified 16 7 30 174 001 0.8 Decapoda Sergestes similis 33 1 0 1 2 007 11 Osteichthyes Ammodytes hexaplerus 3.3 1 0 1 2 028 4.3 Unidentified animal remains 53 3 — — 025 625 Predator characteristics Number of stomachs examined 30 Number of empty stomachs 10 Mean weight per stomach 0 246 g i 0.389 (SD) Mean total length 330.36 mm ± 77.17 (SD) Mean fullness score 1 03 Mean digestion score 1 05 Mean no prey taxa per fish 1.26 pinniger, S. diploproa, and S. alutus are all rela- tively high (0.58, 0.56, and 0.61, respectively). Overlaps between S. pinniger and S. flavidus for the seasonal cruises are similar to the results of the summer surveys (dh = 0.80 by number; 0.46 by weight). A possible explanation for the lower values may be changes in availability of both predator and prey (i.e., no S. flavidus stomachs were collected during spring and early summer when the euphausiid populations are generally the highest). The variability associated with the different cruises was examined by calculating the overlaps between these two species for the four seasonal cruises that contained at least 10 speci- mens of each species. The July cruise had the highest overlap of all on a weight basis (dh = 0.88) and the September cruise had the lowest (dh = 0.32), while the December and January cruises had intermediate overlaps (dh = 0.52 and 0.46), suggesting seasonal variations in prey availability for these species. For comparative purposes, the dietary composi- TABLE 9. — Principal prey types making up >1.09t of the diet and food breadths of the five species of Sebastes. R is the total number of distinct prey items identified to at least genus level and that make up 0.1' < ip, 0.0011 of the identified fraction of the total weight. These prey were used to calculate the overall diet breadth (B i and the evenness of distribution of the prey items in the diet iBn I. The seasonal values for S. flavidus and S. pinniger are given in parentheses. Sample size Principal prey types (Pi's > 0.01) Pooled species values Species R B Bn S. flavidus 185 (79) Euphausia pacifica , Thysanoessa spinifera . hypernd amphipods. Sergestes similis. Loligo opalescens. myctophids. Clupea harengus pallasi 12 (17) 3.64 (377) 0.303 (0.222) S diploproa 62 E. pacifica. T spinifera, S similis. Vibilia propinqua 8 228 0.285 S pinniger 130 (238) E pacifica. T spinifera. Sebastes jordani 8 (6) 1.86 (133) 0.232 (0.222) S cramen 30 E pacifica. calanoid copepods, hypernd amphipods. Ammodytes hexapterus 8 1.80 0.225 S. alutus 73 E pacifica . T spinifera . S similis 7 1.73 0.247 S. pinniger and S. flavidus, although the diets are not similar for other prey items. Overlap on the basis of weight, which may be a better measure of the energy obtained from the various food items, indicates high overlap be- tween S. pinniger and S. diploproa and between S. alutus and S. pinniger, S. diploproa, and S. crameri. The rest of the values were <0.60, in- cluding S. pinniger with S. flavidus (dh = 0.48). The diet of S. flavidus overlaps the least with the other species (dh = 0.42) mainly due to its more piscivorous habits. The mean overlap values of S. tion of the five most important prey categories for each of the rockfish species is presented by percent number and percent weight in Figures 4 and 5. Both figures show the importance of euphausiids in all five species. The stomachs of S. crameri contained a more equitable distribution of num- bers of the major prey groups than the other species of rockfishes, with higher proportions of amphipods and copepods. Some of this difference may be ascribed to the smaller sample size. On a weight basis, S. flavidus was unique in that fishes and cephalopods were of greater importance in the 279 FISHERY BULLETIN: VOL. 82, NO. 2 TABLE 10. — Overlap matrix for the five species of Sebastes. Only those prey that have proportional abundances exceeding 0 V , were used in the analysis. Values above the rules are for proportional weight overlap and values below are for propor- tional abundance. The mean overlap for each species by weight and number is given in parentheses directly above and below the rules. S. pinniger S. flavidus S. diploproa S. crameri S alutus (0.58) S pinniger (0.71) 0.48 (042) 0.72 0.47 066 W S flavidus 0.93 (0.70) 0.44 (0.56) 0.30 046 E I S diploproa 0.76 0.78 (0.65) 0.48 (0.48) 0.63 G H S cramen 040 0.40 0.42 (0.41) 0.69 T (0.61) S alutus 074 070 N 0.63 UMBER 0.42 (0.62) diet of this species than any of the other rockfish. Decapods were of moderate importance to S. diploproa and, to a lesser extent, S. alutus. Fishes were an important food source by weight for all rockfishes but S. alutus. Seasonal Variation Differences in the diet of S. pinniger and S. flavidus are summarized in Table 11 for the four seasons. The spring cruise shows an extreme dom- inance of one prey item, Euphausia pacifica , in the diet of S. pinniger. This prey species was found in about three-quarters of the stomachs and made up almost all the prey biomass. Decapod shrimp and fishes were rarely found in the diet at this time. Euphausiids also dominated the diet in the sum- 100-,-= 90- 80- 70- m i 60. 50- « 40-| r- z Ld g 30. Ld Q- 20-1 10- S . pinniger S . f lavidus S . diploproa S crameri S . alutus EUPHAUSIIDS DECAPODS AMPHIPODS COPEPODS PRINCIPAL PREY ITEMS FISHES FIGURE 4. — The proportions of the five major prey taxa found in the five rockfish species based on numerical composition. 280 BRODEUR and PEARCY: FOOD HABITS AND DIETARY OVERLAP OF SOME SEBASTES TABLE 11. — Variation in major prey taxa composition with season for Sebastes pinruger and S. flauidus. F.O. = frequency of occur- rence; % W = percent gravimetric composition; + = a prey category was present but made up O.V'r of the total weight. Species and No. of fish (% empty) Euphausia Thysanoessa Total Gelatinous pacifica FO % W spimfera F.O. % W euphausiids F.O. % W Decapods F.O. % W Amphipods FO %W Cephalopods F.O. % W Fishes zoopla F.O. ikton season FO. % W % W Sebastes pinniger Spring' 42 73.8 99.6 — — 738 996 2.4 + — — — — 24 04 — — (Mar -May) (14.3) Summer 165 61.2 75.1 79 0.5 68.5 94.7 36 11 2.4 + — — 67 42 3.6 + (June-Aug.) (10.9) Fall 117 35.9 138 32.5 378 55.6 84.2 9.4 0.2 120 0.1 — — 179 15.1 222 0.3 (Sept-Nov) (205) Winter 44 500 25.9 477 120 81 8 94.2 — — 6.8 0.4 — — 13.6 5.5 4.5 + (Dec. -Feb) (22.7) Sebastes flavidus Spring2 0 (Mar -May) Summer 151 523 372 23.2 0.5 58.3 66.7 9.9 08 132 0.1 152 65 139 25.6 3.3 04 (June-Aug.) (16.6) Fall 75 34.7 6.1 427 29.6 54.7 42.2 93 13.8 26 7 0 8 6.7 18 28.0 40.5 10.7 0.8 (Sept -Nov) (22.7) Winter 38 81.6 195 92.1 15.3 94 7 46.7 0.5 0.7 10.5 + 289 306 52.6 154 21 .1 6.6 (Dec -Feb.) (0.0) 'All collections taken during one cruise All other seasons represent the means of at least two cruises spaced a minimum of 1 mo apart (Tables 1 and 2 give the exact dates and samples collected on each cruise). 2No stomachs of S flavidus were collected during this season. 1 00 90- 80- 70 -I x uj 60 50- 40- o 30 cc Id 20 -I 10- lE 5ta 1 jn ^0 El r I EUPHAUSIIDS DECAPODS AMPHIPODS CEPHALOPODS FISHES FIGURE 5. — The proportions of the five major prey taxa found in the five rockfish species based on gravimetric composition. 281 FISHERY BULLETIN: VOL. 82, NO. 2 mer but to a lesser degree. Thysanoessa spinifera appeared in the stomachs at this time, but E. pacifica continued to be the most important euphausiid consumed. Shrimp and fishes were slightly more important but together made up only a minor portion of the diet. A low percentage of empty stomachs occurred in the summer. The diet of S. pinniger in the fall showed substantial shifts in prey composition. Although the frequencies of occurrence were about equal for the two species of euphausiids, T. spinifera great- ly exceeded E. pacifica by weight. Decapods were common but were represented mainly by small shrimp (Sergestes similis) and juvenile pelagic crabs (Munida quadrispina), which contributed little on a weight basis. Amphipods and gelatinous zooplankton occurred frequently but were not important by weight. Fishes were important by occurrence and weight and consisted mostly of mesopelagic species and several adult Sebastes jordani which made a large contribution to the biomass consumed. Almost one-quarter of the fish collected in the winter had empty stomachs and contained much digested material. Euphausia pacifica and 71 spinifera occurred in about the same number of stomachs, but E. pacifica contributed over twice as much of the total weight as T. spinifera. Subadult E. pacifica were very numerous at this time. The fishes consumed were mostly meso- pelagic species. Sebastes flavidus showed similar trends in food resource utilization among the three seasons from which collections were made (Table 11). Euphausiids, consisting mostly of E. pacifica, made up two-thirds of the diet by weight in the summer. Fishes were common and contributed heavily to the total biomass. Cephalopods were next in importance by either occurrence or weight. The diet in the fall showed the same shift in euphausiid species as was apparent for 8. pinni- ger, with T. spinifera the dominant species. Fishes were almost as important by weight as euphau- siids, but their weight total was mostly composed of adult clupeids. Cephalopods were least impor- tant in the fall months. Euphausiids represented about half the diet during the winter, but the remainder was shared mostly by cephalopods and fishes. Both species of euphausiids were commonly found, but E. pacifica (mostly subadults) were slightly more important in the overall diet. Cephalopods (mostly adult Loligo opalescens and juvenile copepods) did show a substantial increase in weight and occurrence during these months. Fishes were found in over half the stomachs but were mainly juveniles of relatively small myctophids. Gelatinous zoo- plankton were most common, and decapods were least common, during this season. In contrast to S. pinniger, all stomachs of this species contained some food and many stomachs were full during this season. Geographic Variation Several trends were evident when comparing the diet of S. pinniger between regions (Table 12). The two southernmost regions had similar diets dominated by E. pacifica with T. spinifera repre- senting only a minor portion of the diet. Meso- TABLE 12. — Variation in major prey taxa composition with geographic area for Sebastes pinniger and S. flavidus. F.O. = frequency of occurrence; % W = percent gravimetric composition; + = a prey category was present but made up < 0.1% of the total weight. No of fish (% Eupha usia Thysanoessa Total Gelatinous Area pacif: < ^-r-6-cJ- -6-0- 100 CO X u < o (- co 60 >- 80 40 20 0 0 0400 0800 1200 1600 2000 2400 SR SS TIME (hours) FIGURE 6. — Feeding intensity indices for Sebastes pinniger at adjusted times of the day. See text for explanation of indices. FIGURE 7. — Feeding intensity indices for Sebastes flavidus at adjusted times of the day. See text for explanation of indices. 284 BRODEUR and PEARCY: FOOD HABITS AND DIETARY OVERLAP OF SOME SEBASTES ratios at the individual times were subjected to an analysis of covariance which compared the weight ratios adjusted for fish size (Jenkins and Green 1977). Both S. pinniger (F(12,354> : 5.68, P < 0.001) and S. flavidus (F(U,262) = 6.51, P < 0.001) showed significant differences in the mean weight ratios over the times tested, implying that feeding varied during the diel period. No significant dif- ferences (P > 0.05) in stomach fullness were associated with size or sex of the predator for either species. Predator-Size Variation The proportion of empty S. pinniger stomachs as well as the percent frequency of occurrence and percent weight of prey taxa were remarkably invariant among the four predator size classes (Table 14). Only the largest size class (2:55+ cm) shows any substantial variation with a larger pro- portion by weight of fishes and a commensurate decrease in weight of euphausiids consumed. Much of this fish weight was contributed by a few individual fish of large relative size (mostly adult S. jordam); the frequency of occurrence of fishes is only slightly higher for this largest size class. Few obvious size-related trends were apparent for S. flavidus. The two smallest size classes consumed the largest proportion of euphausiids. Euphausia pacifica were less important for large fish. Decapods and cephalopods showed similar trends except that the frequencies of occurrence were highest for cephalopods but lowest for deca- pods in the largest size class. Fishes were consis- tent in their weight and occurrence proportions except that one size class (40-<45 cm) had much lower proportions than the others. Few trends were apparent for either amphipods or gelatinous zooplankton although both groups were commonly found. To determine if different sizes of rockfish se- lected different sizes of prey, all fish that con- tained measurable prey were grouped into 10 mm length intervals and the means and ranges of their prey were plotted against fish size (Fig. 8). Al- though some exceptions exist, the majority of the prey of S. pinniger are found within a narrow range of prey sizes, a range (15-27 mm) largely determined by adult euphausiids, the dominant prey category (Fig. 8). Fishes of the largest two size classes consumed larger prey on average, and their prey had the largest variation in size due to high numbers of both small and large prey con- sumed by these fish. No significant relationship was found between length of fish and either overall size of prey or size of euphausiid prey. Sebastes flavidus showed a much greater range in the sizes of prey consumed with the variation and range in prey length increasing with size of predator (Fig. 8). The mean size of prey eaten did not appreciably increase until the very largest size classes. Although the maximum prey size increases with fish size, the minimum size varies little throughout the length ranges examined. Again for this species, no relationship was found between fish length and overall or euphausiid prey lengths. The size distribution of prey is shown for both species in Figure 9. The prey-size spectrum of S. pinniger was distributed fairly normally with the TABLE 14. — Variation in major prey taxa composition with size of predator for Sebastes pinniger and S. flavidus. F.O. = frequency of occurrence; '/ W = percent gravimetric composition; + = a prey category was present but made up < O.l'X of the total weight. Size range (cm) No of fish (% empty) Euphausia pacifica FO. % W Thysanoessa spinifera FO. % W Total euphausiids F.O. % W Decapods F.O. % W Amphipods F.O. % W Cephalopods F.O. % W Fishes Gelatinous zooplankton F.O. % W F.O. % W Sebastes pinniger 45 64 484 43.2 21 9 9.7 687 91.4 4.7 3.1 6.2 0.3 14.1 48 4.7 0.4 45- 50 (172) 102 51.8 46.1 176 186 676 92.5 4.9 0.9 4.9 0.1 _ _ 13.7 6.5 5.9 0.1 50- < 55 (17.6) 146 47.3 65.6 21.2 12.0 61.1 94.9 5.5 0.2 4 1 + 11.0 4.7 8.9 0.1 '55 (14.4) 56 589 49.3 125 7.6 67.9 83.4 5.4 02 7.1 + _ 16.1 163 143 0.2 Sebastes flavidus (14.3) - 40 35 886 44.7 57.1 11.7 943 61.3 11.4 0 1 14.3 0.1 14.3 2.8 31.4 34.6 29 0.1 40- 45 (0.0) 61 459 50.2 29 5 12.4 525 83.9 13.1 1.2 246 0.2 13.1 8.4 9.8 58 49 04 45-<50 (22.9) 126 47.6 22.9 38.1 13.7 57.1 46.8 11.1 3.7 12.0 0.1 127 17.4 29.3 29.3 79 24 ^50 (21.4) 42 (2.4) 47.6 25.8 30.9 83 54.7 51.9 7.1 2.6 238 0.3 21.4 136 45.2 303 214 0.1 285 FISHERY BULLETIN: VOL. 82, NO. 2 150-7 125- E E 100 H UJ M OT 75H >- Q- 504 25- 0- 400 S. pinniger U?t )ii> f^ + i|t llll 450 500 —I — i — i — •- 550 600 FIGURE 8. — Mean (horizontal lines) ±95% confidence limits (boxes) and ranges (vertical lines) of prey sizes found for each 10 mm interval of Sebastes pinniger and S. flavidus. 1507 125- 100- <" 75 >- UJ cr Q- 50 25- S flavidus it^nHt iiim «fi .mi 0 } i i i I I « i I I i ' I i ■ I I I ' I I I I I I I i I t I I 300 350 400 450 500 550 600 PREDATOR LENGTH (mm) mode coinciding with the mean (x = 10.38 mm), although disjunct groups of small and large prey were found (Fig. 9). The prey-size spectrum of S. flavidus was slightly skewed toward the larger sized prey with the mean size (x = 18.44 mm) less than the mode. A smaller peak also appeared around 25 mm. No significant differences were found in the mean prey sizes utilized by the two species (Student's f-test, P > 0.05). Analysis of Variation The results of the chi-square analyses for S. pinniger showed that none of the factors analyzed had a significant effect on the occurrence of food in the stomachs (Table 15). At least one of the factors was related to the occurrences of all seven prey categories examined. Seasonal effects were the most significant (all P ^ 0.01) and were due to the higher occurrences of hyperiid amphipods, fishes, and gelatinous zooplankton in fall and winter. Area and time of capture showed both highly significant (P s 0.001) and insignificant effects depending on the prey category, but most compar- isons were significant at the 0.05 level. In none of the prey categories examined did the size of the predator have a significant effect on the relative proportions consumed. For S. flavidus, season of capture and size of predator affected the proportion of empty stom- achs found (Table 15). Again season had the most significant influence on prey occurrence and was significant in all eight prey categories. Highly significant differences were found in area of cap- ture and size of predator especially in the euphau- siid and fish categories. Differences in occurrence of prey with time of capture deviated from ex- pected the least of all the factors analyzed. DISCUSSION The five species of rockfishes examined rely heavily, if not exclusively, on pelagic macrozoo- plankton and micronekton. Although some ben- thic species appear in the prey lists (e.g., Lyopsetta exilis, Munida quadrispina, Psettichthys melan- 286 BRODEUR and PEARCY: FOOD HABITS AND DIETARY OVERLAP OF SOME SEBASTES TABLE 15. — Results of chi-square analyses testing for differences in the occurrence of food and specific prey categories within the various factors. All significances are with three degrees of freedom except where noted. Factor analyzed Occurrence of food Euphausia pacifica Thysanoessa spinifera Total euphausiids Decapods Amphipods Cephalopods Fishes • = P - 0.05. " = P • 0.01. •" = P ■ 0 001 1 Significance with two degrees of freedom. Gelatinous zooplankton Sebasfes pinniger Season 659 23.09"- 72.48— 11 26" 113 28-" 1676*" 1587— 39.96— Area' 0.19 8.61- 5.22 11 42— 0.46 22.44— — 7.95* 35.66— Time 432 28 18*" 53.77— 11.86" 1.37 13.05— — 1.43 1921 — Size 0.67 3.12 324 626 0.04 098 1.26 4.35 Sebastes flavidus Season' 979" 1332— 51 33— 10.02" 30.27— 11.15*" 6.65- 21 43— 11.67— Area 5.76 20.83— 15.81 — 1423— 1.21 2.76 11 62" 22.41 — 6.43 Time 1 12 3.21 542 238 10.43* 12.30" 8.25* 9.33* 5.60 Size 1750"- 13 78— 6.35 14.17— 202 11 69" 0.94 12.92"- 886" 25 20 ■• < o (J rr UJ a. o t- u. O o 15-' S pinniger n 25 20 ■■ 15 •• 10 • 5 •• S flavidus r^ o \ > ■ ■ i^ ■ ■' ' ■ l m ■ ■ ■ | ■ ■ i i i i i i i i ■ i 1 1 1 | i i i i 0 5 10 15 20 25 30 35 PREY LENGTH INTERVAL (mm) FIGURE 9. — Prey size spectra in percent for Sebastes pinniger and S. flavidus. ostictus ), they were represented by postlarval or juvenile forms commonly found in the plank- ton. Several comparatively large nektonic fishes and cephalopods (e.g., Clupea harengus pallasi, Sebastes jordani, Loligo opalescens) were eaten, but their occurrences were relatively rare. Con- versely, the virtual absence of many common benthic and epibenthic organisms of appropriate size such as mysids, cumaceans, and gammaridean amphipods further implies that these fish do not normally feed on benthic animals. These findings concur with the limited number of previous studies dealing with food habits of off- shore rockfish. Phillips (1964) reported on the diet of all the species included here except S. alutus. Although little taxonomic detail and no quanti- tative data on prey consumption were given, euphausiids were listed as important forage items for all four species. Fishes were also important prey for several species, especially S. flavidus. Skalkin (1964), in a study of S. alutus in the Bering Sea, found mostly euphausiids and cope- pods in the stomachs, but also stated that a few nektobenthic species and "fragments" of benthic echinoderms were present. The food habits of S. flavidus have been de- scribed in several studies off Oregon and Wash- ington. Pereyra et al. (1969) found unusually high abundances and volumes of the mesopelagic fish, Stenobrachius leucopsarus, in S. flavidus stom- achs collected near Astoria Canyon and hypothe- sized that local hydrographic conditions may have aggregated these prey at high densities. Gunder- son et al. (1980)5 reported that S. flavidus off the coast of Washington fed mostly on fishes, includ- ing some pleuronectid fishes possibly eaten near the bottom along with benthic polychaetes. Lorz et al. (1983) found euphausiids dominating the diet of S. flavidus off Washington and Queen Char- lotte Sound, with fishes of greater importance in the latter region. Another deepwater species, S. marinus, found in the North Atlantic Ocean, also fed chiefly on pelagic prey (Lambert 1960). Euphausiids, hyperiid amphipods, and copepods were the most abundant prey, but mesopelagic fishes were also found in large numbers. Among the species considered here, two diver- gent feeding patterns are apparent, assuming that 5Gunderson, D. R., G. L. Thomas, P. Cullenberg, D. M. Eggers, and R. Thome. 1980. Rockfish investigations off the Wash- ington coast. Ann. Rep., prep, for NMFS, Univ. Wash., 68 p. 287 FISHERY BULLETIN: VOL. 82, NO. 2 the same prey items are equally available to all species. These can be seen most clearly in the divergence of the cumulative curves of the number of prey species (Fig. 2). Three species (S. pinniger, S. alutus, and S. crameri) tend to be steno- phagous, with very few prey items represented in large volumes of prey organisms. Euphausiids appear to be the most sought after or available prey, and other prey taxa occur in low numbers. These three species show similar low food breadth values. Sebastes flavidus and S. diploproa , on the other hand, have steadily rising prey curves that con- tinue to rise and approach an asymptote beyond the limits of the figure. These curves are charac- teristic of euryphagous predators which show high overall prey diversity as well as high within- stomach diversity. This high prey diversity can be seen in the greater food breadth values attained by these two species. Although euphausiids pre- dominate in these stomachs, high abundances of other prey, which may be preferred but have lower abundances and availabilities than euphausiids, also occur. The diet overlap measurements calculated here may be useful in comparing how similar the food habits of two species are but may be of limited use when interpreted in an ecological sense. The interaction of factors that affect or determine the diet of a particular species is complex and may include such factors as temporal and spatial dis- tribution of prey, behavioral adaptations of pred- ator and prey, prey detection capabilities, and feeding morphologies of predators (Hyatt 1979). Caution should be exercised when inferences are made about possible species interactions based on diet overlap measurements alone. Two species may have broadly overlapping diets in terms of prey composition but segregate with respect to prey sizes selected, time of feeding, or habitat utilization (Schoener 1974; Ross 1977; Werner 1979; Macpherson 1981). Sebastes pinniger and S. flavidus are two of the most abundant rockfish species within the geo- graphical confines of this study. They inhabit similar depth ranges, latitudinal ranges, and show broadly overlapping areas of peak abun- dances according to trawl survey data (Alverson et al. 1964; Richardson and Laroche 1979; Gunderson and Sample 1980). Adams (1980) found that these two species had the highest positive association in trawl catches using presence-absence data of the seven abundant species he examined. Little is known, however, about their small-scale hori- zontal and vertical distribution. Although they may occupy similar bottom habitat, S. flavidus may be more pelagic (Alton 1972). Seasonal, geographical, and diel variations in the abundance and availability of the important prey of S. pinniger and S. flavidus could be a major cause of the variations in the diet of these species. These variations may be the result of intrinsic prey population fluctuations with sea- son, behavioral adaptations such as diel and ontogenetic vertical migration, or may stem from the prevailing oceanographic conditions either concentrating, dispersing, or transporting prey so that all prey are not equally available in the limited time and space frame of the individual predator. Current patterns alone are known to vary with season, depth, and geographic area (Huyer et al. 1975; Ingraham and Love 1978) and may affect the availability and concentration of prey. Quantitative estimates of the seasonal and areal distributions of the total prey spectrum consumed by these rockfishes are limited. Day (1971) sampled macrozooplankton and micronek- ton from the northern part of the range of this study (lat. 46° 45 '-50° 02' N) using a 0.9 m Isaacs- Kidd midwater trawl in the upper 150 m of the water column during the spring and fall. He found a peak in the biomass of catches at the outer edge of the continental shelf. Euphausiids dominated the catch at most stations, and E. pacifica and T. spinifera together accounted for 90% of the total abundance of all organisms collected, which is similar to the abundances found in the stomachs of several species examined here. Although the pro- portional abundance of E. pacifica varied greatly relative to T. spinifera, E. pacifica dominated the catches and was most concentrated during the spring when it comprised the largest proportion of the stomach weights in our study. Mesopelagic fishes were commonly collected in Day's sampling, but mostly at the offshore stations. Pearcy (1972) reviewed the species composition, vertical and horizontal distribution, and varia- tions in abundance of the macrozooplanktonic and nektonic fauna derived from 8 yr of sampling off Oregon. Annual and seasonal changes in the abundance and distribution of many species could be correlated with changes in oceanographic con- ditions. Following the cessation of upwelling in fall, surface waters flow predominantly inshore and northward, transporting shrimps and myc- tophids onto the shelf. We found that shrimp and myctophids became more important in the diets of 288 BRODEUR and PEARCY: FOOD HABITS AND DIETARY OVERLAP OF SOME SEBASTES S. pinniger and S. flavidus at this time. An inshore-offshore peak in the biomass of midwater collections occurred on the edge of the continental slope off the central Oregon coast (lat. 44° 39' N), a zone where oceanic macrozooplankton and micro- nekton may be concentrated by advection (Pearcy 1976). This is the region where pelagic-feeding rockfishes are often concentrated (Gabriel and Tyler 1980). The majority of the prey species found in the stomachs of the rockfish species examined are pelagic species that undertake extensive diel vertical migrations and are important compo- nents of the biological sound scattering layer in the Northeast Pacific (Pearcy and Laurs 1966; Brinton 1967; Pearcy and Mesecar 1971; Pearcy 1972; Alton and Blackburn 1972). In this study, both of the euphausiid species of interest, E. pacifica and T. spinifera, have been found to have substantially different daytime and nighttime vertical distributions. According to Alton and Blackburn (1972), catch rates of T. spinifera off the coast of Washington were the highest near the bottom during the early evening hours (1800- 2000 h) and at the surface a few hours later (2100- 2300 h). The diurnal downward migration of these orga- nisms over the continental shelf may result in a substantial biomass in close proximity to near- bottom predators, such as rockfishes, which feed on pelagic prey during the day. Deeper migration to daytime depths typical of their more open ocean conspecifics is restricted by the shelf, especially in shoaler areas such as Heceta Bank. Isaacs and Schwartzlose (1965) found dense populations of predators, including many rockfishes, on shallow banks off California; these predators presumably take advantage of net inshore transport by cur- rents of oceanic organisms over the bank. Pereyra et al. (1969) reported high incidences of predation on mesopelagic organisms by aggregations of S. flavidus residing on the shelf edge near a deep canyon. Vertically migrating mesopelagic orga- nisms may also constitute an important food source for many species of slope fishes (Sedberry and Musick 1978). Diel vertical distribution patterns of offshore rockfishes are not well documented. Based on acoustic observations, Westrheim (1970) con- cluded that schools of Pacific ocean perch move off bottom at night. Pereyra et al. (1969) and Love (1981) caught rockfishes that were apparently feeding well off the bottom at night. Lorz et al. (1983) concluded that S. flavidus off Washington fed on euphausiids during night or early morning hours, when these euphausiids would be expected to be in surface waters. Similar migrations were seen on Heceta Bank during this study. Figure 10 shows an acoustic 33 kHz transect taken across Heceta Bank during the late morning (about 1023- 1050 h PST). Many large "spikes" offish aggrega- tions were apparent extending over 100 m above bottom. Some of these were probably caused by rockfish ascending in the water column to feed. Figure 11 is a 33 kHz echogram on Heceta Bank made around 1800 h PST The "haystacks" shown are characteristic of tight aggregations of S. pinniger just above bottom (Barss6) and may represent feeding aggregations. Also visible in this echogram is more diffuse scattering in the water column (20 m off bottom) probably caused by euphausiids. The tow made concurrently with this trace did yield a large catch of rockfish (97% S. pinniger), most of which had stomachs full of fresh euphausiids. This stratification of large sound scatterers below diffuse midwater scatter- ing prey was often observed during the acoustic surveys. Atlantic cod appear to interact with pelagic prey in a similar fashion (Brunei 1965; Pearcy et al. 1979; Falk-Peterson and Hopkins 1981). The two primary species examined in detail in this study appear to forage mainly during the midday and evening dusk periods, although sam- pling was limited during nighttime. The similar diel patterns of feeding intensity suggest that temporal partitioning of feeding time is not occur- ring between S. pinniger and S. flavidus. The differing utilization patterns of euphausiids and fishes seen for the two species (Table 13) may be related to the vertical positioning of the two species in the water column. Sebastes flavidus may feed high in the water column, prey upon adult herring and pelagic juvenile fishes during the daytime, and intercept euphausiids during crepuscular periods, whereas S. pinniger may stay nearer the bottom where they may feed al- most exclusively on increased daytime aggrega- tions of euphausiids. The occurrence of a high percentage of empty stomachs and generally low feeding intensity indices in S. alutus, which were caught mainly in late afternoon in our study, suggests that this species is more nocturnal in its feeding patterns, assuming that this species has similar regurgita- 6W. Barss, fishery biologist, Oregon Department of Fish and Wildlife, Newport, OR 97365, pers. commun. December 1980. 289 FISHERY BULLETIN: VOL. 82, NO. 2 Q. LU Q _200 FIGURE 10. — Echo sounder (Krupp-Atlas Elektronik Model 611, 33 kHz) transect across Heceta Bank showing characteristic rough topography and schools offish in the water column. The larger signal at the extreme right (arrow) is believed to be a large concentration of forage fish, possibly myctophids. The bottom section of the figure shows an expanded version of the layer just above the seabed. tion and digestion rates as the other species studied. Skalkin (1964) found that the feeding intensity of immature S. alutus in the Bering Sea was highest around midday with a smaller peak shortly after dusk as found for S. pinniger and S. flavidus in this study He also hypothesized that larger fish may feed higher in the water column at night on euphausiids, but his nighttime sampling was also limited. Similar pelagic feeding by S. alutus at night may be occurring in our study area, but these fish would not be available to bottom trawls at this time. The rockfishes considered here are just a few species in an extensive guild (sensu Root 1967) 290 BRODEUR and PEARCY: FOOD HABITS AND DIETARY OVERLAP OF SOME SEBASTES 0__ / 2S~ Sfx^"^ *f-*-3Ut~&C 20. 'Cusf S X t- Q. UJ Q 100_ - 40 _ 80 _120 _160 O m ■o FIGURE 11. — Smoother bottom profile made during a tow showing the "haystacks" of rockfish in close association with the bottom and possibly preying upon the food organisms (arrow) directly above them. of organisms which feed in varying degrees on euphausiids. Other pelagic predators in this study- area known to feed intensively on euphausiids include Pacific hake (Alton and Nelson 1970), myctophids (Tyler and Pearcy 1975), juvenile salmon (Peterson et al. 1982), and squid (Karpov and Cailliet 1978). Standing stocks and production rates of euphausiids in northern latitudes may be of such magnitude that many predators often subsist on them in coexistence rather than com- pete for other more limited resources. More re- search is needed on the biology and distribution of these abundant prey species and their importance to fishery resources. In complex, multispecies fisheries such as those utilizing rockfishes, it may be possible to treat several species with similar 291 FISHERY BULLETIN: VOL 82, NO. 2 life histories and which prey on similar resources as a biological unit for management purposes. ACKNOWLEDGMENTS We are indebted to the many individuals who assisted with the collection of the stomach sam- ples and in particular to Mary Yoklavich and Ray Taylor of Oregon State University, Bill Barss and Steve Johnson of the Oregon Department of Fish and Wildlife (ODFW), and numerous scientists at the Northwest and Alaska Fisheries Center (NWAFC) Seattle Laboratory. We thank Bob Demory of ODFW and Tom Dark of NWAFC for allowing us to participate on their research cruises and George Boehlert, Carl Bond, and an anonymous reviewer for their helpful criticisms on the manuscript. This study was funded by Contract No. 81-ABC-00192 from the NWAFC, National Marine Fisheries Service. LITERATURE CITED ADAMS, P. B. 1980. Morphology and distribution patterns of several important species of rockfish (Genus Sebastes). Mar. Fish. Rev. 42(3-4):80-82. ALTON, M. S. 1972. Characteristics of the demersal fish fauna inhabit- ing the outer continental shelf and slope off the northern Oregon Coast. In A. T. Pruter and D. L. Alverson (edi- tors), The Columbia River Estuary and adjacent ocean waters, p. 583-634. Univ. Wash. Press, Seattle. ALTON, M., AND M. O. NELSON. 1970. Food of the Pacific hake, Merluccius productus, in Washington and northern Oregon coastal waters. In Pacific hake, p. 35-42. U.S. Fish Wildl. Serv, Circ. 332. Alton, M. S., and C. J. Blackburn. 1972. Diel changes in the vertical distribution of the euphausiids, Thysanoessa spimfera Holmes and Euphau- sia pacifica Hansen, in coastal waters of Washington. Calif. Fish Game 58:179-190. ALVERSON, D. L., A. T. PRUTER, AND L. L. RONHOLT 1964. A study of the demersal fishes and fisheries of the northeastern Pacific Ocean. H. R. MacMillan Lectures in Fisheries, 190 p. Institute of Fisheries, University of British Columbia, Vancouver. Bray, R. N., and A. W. Ebeling. 1975. Food, activity, and habitat of three "picker-type" microcarnivorous fishes in the kelp forests off Santa Barbara, California. Fish. Bull., U.S. 73:815-829. BRINTON, E. 1967. Vertical migration and avoidance capability of euphausiids in the California Current. Limnol. Ocean- ogr. 12:451-483. BRODEUR, R. D. 1983. Food habits, dietary overlap and gastric evacuation rates of rockfishes (Genus Sebastes). M.S. Thesis, Oregon State Univ., Corvallis, 102 p. BRUNEL, P. 1965. Food as a factor or indicator of vertical migrations of cod in the western Gulf of St. Lawrence. Int. Comm. Northw. Atl. Fish., Spec. Publ. 6:439-448. CHEN, L.-C. 1971. Systematics, variation, distribution, and biology of rockfishes of the subgenus Sebastomus (Pisces, Scorpaen- idae, Sebastes ). Bull. Scripps Inst. Oceanogr. 18, 115 p. 1975. The rockfishes, genus Sebastes (Scorpaenidae), of the Gulf of California, including three new species, with a discussion of their origin. Proc. Calif. Acad. Sci. 40: 109-141. COLWELL, R. K., AND D. J. FUTUYMA. 1971. On the measurement of niche breadth and overlap. Ecology 52:567-576. DAY, D. S. 1971. Macrozooplankton and small nekton in the coastal waters off Vancouver Island (Canada) and Washington, spring and fall of 1963. U.S. Dep. Commer., Natl. Mar. Fish. Serv., Spec. Sci. Rep. Fish. 619, 94 p. FALK-PETERSON, S., AND C. C. E. HOPKINS. 1981. Zooplankton sound scattering layers in North Nor- wegian fjords: Interactions between fish and krill shoals in a winter situation in Ullsfjorden and Oksfjorden. Kieler Meeresforsch. 5:191-201. Gabriel, W L., and A. V Tyler. 1980. Preliminary analysis of Pacific Coast demersal fish assemblages. Mar. Fish. Rev. 42(3-4):83-88. GOTSHALL, D. W, J. G. SMITH, AND A. HOLBERT 1965. Food of the blue rockfish Sebastodes mystinus. Calif. Fish Game 51:147-162. GUNDERSON, D. R., AND T. M. SAMPLE. 1980. Distribution and abundance of rockfish off Washing- ton, Oregon and California during 1977. Mar. Fish. Rev. 42(3-4):2-16. HELFMAN, G. S. 1978. Patterns of community structure in fishes: summary and overview. Environ. Biol. Fish 3:129-148. HOBSON, E. S., AND J. R. CHESS. 1976. Trophic interactions among fishes and zooplankters near shore at Santa Catalina Island, California. Fish. Bull., U.S. 74:567-598. HUYER, A. 1977. Seasonal variation in temperature, salinity, and density over the continental shelf off Oregon. Limnol. Oceanogr. 22:442-453. HUYER, A., R. D. PILLSBURY, AND R. L. SMITH. 1975. Seasonal variation of the alongshore velocity field over the continental shelf off Oregon. Limnol. Oceanogr. 20:90-95. Hyatt, K. D. 1979. Feeding strategy. Chap. 2. In W. S. Hoar, D. J. Randall, and J. R. Brett (editors), Fish physiology. Vol. VIII. Bioenergetics and growth, p. 71-119. Academic Press, N.Y. INGRAHAM, W. J., JR., AND C. M. LOVE. 1978. Oceanographic conditions off California to Vancou- ver Island in the summer of 1977. Mar. Fish. Rev. 40(21:24-28. ISAACS, J. D., AND R. A. SCHWARTZLOSE. 1965. Migrant sound scatterers: interaction with the sea floor. Science (Wash., D.C.) 150:1810-1813. JENKINS, B. W, AND J. M. GREEN. 1977. A critique of field methodology for determining fish feeding periodicity. Environ. Biol. Fish 1:209-214. 292 BRODEUR and PEARCY: FOOD HABITS AND DIETARY OVERLAP OF SOME SEBASTES KARPOV, K. A., AND G. M. CAILUET. 1978. Feeding dynamics of Loligo opalescens. Calif. Dep. Fish Game, Fish Bull. 169, p. 45-65. Lambert, D. G. I960. The food of the redfish Sebastes marinus (L.) in the Newfoundland area. J. Fish. Res. Board Can. 17: 235-243. LANC.TON, R. W. 1982. Diet overlap between Atlantic cod, Gadus morhua, silver hake. Merluccius bilinearis, and fifteen other Northwest Atlantic finfish. Fish. Bull., U.S. 80:745-759. LARSON, R. J. 1!<72 The food habits of four kelp-bed rockfishes (Scor- paenidae, Sebastes I off Santa Barbara, California. M.S. Thesis, Univ. California, Santa Barbara, 57 p. 1980. Competition, habitat selection, and the bathymetric segregation of two rockfish [Sebastes) species. Ecol. Monogr. 50:221-239. LEVINS, R. 1968. Evolution in changing environments. Princeton Univ. Press, Princeton, 120 p. LINTON, L. R., R. W. DAVIES, AND F. J. WRONA. 1981. Resource utilization indices: an assessment. J. Anim. Ecol. 50:283-292. LORZ. H. V. W. G. PEARCY, AND M. FRAIDENBURG. 1983. Notes on the feeding habits of the yellowtail rock- fish, Sebastes flavidus, off Washington and in Queen Charlotte Sound. Calif. Fish Game 69:33-38. LOVE, M. S. 1981. Evidence of movements of some deepwater rock- fishes (Scorpaemdae: Genus Sebastes i off Southern Cali- fornia. Calif. Fish Game 67:246-249. LOVE, M. S., AND A. W. EBELING. 1975. Food and habitat of three switch-feeding fishes in the kelp forests off Santa Barbara, California. Fish. Bull., U.S. 76:257-271. MACPHERSON, e. 1981. Resource partitioning in a Mediterranean demersal fish community. Mar. Ecol. Prog. Ser. 4:183-193. MCEWEN. G. F. M. W. JOHNSON, AND T. R. FOLSOM. 1954. A statistical analysis of the performance of the Folsom plankton sample splitter, based on test observa- tions. Arch. Meteorol. Geophys. Bioklimatol., Ser. A. 7:502-527. PEARCY, W. G. 1972. Distribution and ecology of oceanic animals off Oregon. In A. T. Pruter and D. L. Alverson (editors), The Columbia River estuary and adjacent ocean waters, p. 351-377. Univ. Wash. Press, Seattle. 1976. Seasonal and inshore-offshore variations in the standing stocks of micronekton and macrozooplankton off Oregon. Fish. Bull.. U.S. 74:70-80. PEARCY, W G., C. C. E. HOPKINS, S. GRONVIK, AND R. A. EVANS. 1979. Feeding habits of cod, capelin, and herring in Bals- fjorden. northern Norway, July-August 1978: the impor- tance of euphausiids. Sarsia 64:269-277. PEARCY, W G., AND R. M. LAURS. 1966. Vertical migration and distribution of mesopelagic fishes off Oregon. Deep-Sea Res. 13:153-165. PEARCY, W. G., AND R. S. MESECAR. 1971. Scattering layers and vertical distribution of oceanic animals off Oregon. In G. B. Farquhar (editor), Proceed- ings of an international symposium on biological sound scattering in the ocean, p. 381-394. Maury Center for Ocean Science, Wash., D.C. PEREYRA, W. T, W G. PEARCY, AND F E. CARVEY, JR. 1969. Sebastodes flavidus, a shelf rockfish feeding on mesopelagic fauna, with consideration of the ecological implications. J. Fish. Res. Board Can. 26:2211-2215. PETERSON, W T„ R. D. BRODEUR, AND W G. PEARCY. 1982. Food nabits of juvenile salmon in the Oregon coastal zone, June 1979. Fish. Bull., U.S. 80:841-851. PHILLIPS, J. B. 1964. Life history studies on ten species of rockfish (genus Sebastodes >. Calif. Dep. Fish Game, Fish Bull. 126, 70 p. RICHARDSON, S. L., AND W. A. LAROCHE. 1979. Development and occurrence of larvae and juveniles of the rockfishes Sebastes crameri, Sebastes pinniger, and Sebastes helvomaculatus (Family Scorpaenidae) off Oregon. Fish. Bull., U.S. 77:1-46. ROOT, R. B. 1967. The niche exploitation pattern of the blue-gray gnat- catcher. Ecol. Monogr. 37:317-350. ROSS, S. T. 1977. Patterns of resource partitioning in searobins (Pis- ces: Triglidae). Copeia 1977:561-571. SCHOENER, T. W 1970. Nonsynchronous spatial overlap of lizards in patchy habitats. Ecology 51:408-418. 1974. Resource partitioning in ecological communities. Science (Wash., D.C.) 185:27-39. SEDBERRY, G. R.. AND J. A. MUSICK. 1978. Feeding strategies of some demersal fishes of the continental slope and rise off the Mid-Atlantic Coast of the USA. Mar. Biol. (Berl.) 44:357-375. SKALKIN, V. A. 1964. Diet of rockfish m the Bering Sea. In R A. Moiseev (editor), Soviet fisheries investigations in the Northeast Pacific, Part II, p. 159-174. (Transl. Isr. Program Sci. Transl.; available U.S. Dep. Commer., Clearinghouse Fed. Sci. Tech. Inf., Springfield, VA 22151.) SNEDECOR, G. W. AND W. G. COCHRAN. 1967. Statistical methods. 6th ed. Iowa State Univ. Press, Ames, 593 p. TYLER, A. V. 1972. Food resource division among northern, marine, demersal fishes. J. Fish. Res. Board Can. 29:997-1003. TYLER, H. R., JR., AND W G. PEARCY. 1975. The feeding habits of three species of lanternfishes (family Myctophidae) off Oregon, USA. Mar. Biol. (Berl. ) 32:7-11. Wallace, R. K., Jr. 1981. An assessment of diet-overlap indexes. Trans. Am. Fish. Soc. 110:72-76. WERNER, E. E. 1979. Niche partitioning by food size in fish communities. In H. Clepper (editor). Predator-prey systems in fisheries management, p. 311-322. Sport Fish. Inst., Wash., D.C. WESTRHEIM, S. J. 1970. Survey of rockfishes, especially Pacific ocean perch, in the northeast Pacific Ocean, 1963-66. J. Fish. Res. Board Can. 27:1781-1809. 293 SPECIES ASSOCIATIONS AND COMMUNITY COMPOSITION OF MIDDLE ATLANTIC BIGHT CONTINENTAL SHELF DEMERSAL FISHES1 J. A. Colvocoresses and J. A. Musick2 ABSTRACT Cluster analyses of seasonal NMFS Groundfish Survey bottom trawl catches on the Middle Atlantic Bight continental shelf revealed consistent species associations and faunal zones over a 9-year period. Boundaries between faunal zones tended to follow isotherms and isobaths. During the late winter-early spring, the following faunal zones were found: Northern inner shelf, northern mid-shelf, southern inner- and mid-shelf, and outer shelf-shelf break. Five species groups were identified: A small cryophilic group restricted to the first zone, a cold-water boreal group found in the first two zones, a ubiquitous boreal/resident group containing the major dominants, a warm-temperate group confined to the warmer southern and outer shelf waters, and a group of slope residents confined to the deepest zone. During the fall, five faunal zones were found: Southern inner/mid-shelf, northern inner shelf, northern mid-shelf, outer shelf, and shelf break. Five species associations were largely analogous to those in the spring, with the following exceptions: The cryophilic group was absent, the ubiquitous group contained mixed boreal and warm-temperate elements, and a second outer shelf group was recognized. The most notable change in the distribution of groups from the spring was a general northward shift and a sharply defined inshore movement of the temperate group. Communities of fishes on the continental shelf have rarely been studied beyond the compilation of species lists for given areas. This is enigmatic when one considers the large amount of survey data that has been collected from much of the world's continental shelf waters in connection with fishery exploration and monitoring. While trawl survey data have traditionally been col- lected with the primary aim of assessing commer- cially harvestable stocks, they also provide an excellent base for evaluating the interspecific re- lationships among trawlable organisms. The few studies which have previously ad- dressed community structure of open continental shelf fishes have found clearly definable species associations with distributions related to en- vironmental parameters. Demersal fish species assemblages found using objective mathematical measures have been described for the continental shelves in the Gulf of Guinea ( Fager and Long- hurst 1968), northwest Pacific coast of the United States (Day and Pearcy 1968), and Campeche Bank off Mexico (Sauskan and Ryzhov 1977). Since 1967 the National Marine Fisheries Ser- vice (formerly Bureau of Commercial Fisheries) Contribution No. 1151 from the Virginia Institute of Marine Science of the College of William and Mary. 2Virginia Institute of Marine Science and School of Marine Science, College of William and Mary, Gloucester Point, VA 23062. Manuscript accepted September 1983. FISHERY BULLETIN: VOL. 82, NO. 2, 1984. has conducted a semiannual bottom trawl survey of the continental shelf waters from Nova Scotia to Cape Hatteras (Grosslein 1969). This program has produced a data base which offers a unique oppor- tunity for the analysis of the composition and var- iability of the fish communities in this region. In the present study, that portion of these data collected in the Middle Atlantic Bight (Cape Cod to Cape Hatteras) during the cruises from fall 1967 through spring 1976 were analyzed with the aim of defining the composition of fish communi- ties present within this area and how they vary geographically, thermally, and seasonally. METHODS Sampling Groundfish Survey cruises were conducted by the National Marine Fisheries Service during the fall and spring from fall 1967 through spring 1976, aboard either the RV Albatross IV or RV Delaware II. The survey area extended from the 15-fathom (27 m) contour offshore to 200 fathoms (365 m). A stratified random sampling design was utilized, based on depth and geographical zones (Fig. 1). Catch data from strata 1-12 and 61-76 (Middle At- lantic Bight) were analyzed in the present study. Sampling intensity in each stratum was allocated according to the geographic area of each stratum 295 30 FIGURE 1. — Northwest Atlantic area sampled by NMFS Groundfish Survey. In the present study, data collected from the Middle Atlantic Bight area (strata 1-12, 61-76) between fall 1967 and spring 1976 were examined. (2-16 stations per stratum). At each station a tow of l/2-h duration at a speed of 3.5 kn was made along the bottom. A standard #36 Yankee trawl was utilized except during the spring cruises from 1973 to 1976, when a modified high-opening #41 Yankee trawl was used. The fishes captured were identified, counted, and weighed by species. A bathythermograph cast was made at each station. Further details of sampling design and sample processing may be found in Clark and Brown (1977) and Grosslein (1969). Analyses Clustering Catch data were initially analyzed separately for each of 18 cruises, using numerical classifica- tion (clustering). Assemblages of fishes were de- fined by computing a similarity coefficient, Srv k )> among species from the species-station matrix and subsequently classifying species into clusters or groups (Sneath and Sokal 1973). Stations were clustered in the same manner from the inverted matrix, and species and station (site) groups were compared by nodal analysis (Lambert and Wil- liams 1962). Matrix values entered were counts of individuals, as biomass measurements are overly influenced by the presence of relatively rare but large, motile fishes (which are poorly sampled by trawls) in the collections. The similarity coefficient used was the Can- berra metric (Lance and Williams 1967), which is particularly effective when the organisms under study are contagiously distributed (Clifford and Stephenson 1975) as most fishes are. Also, to further reduce the effects of contagion, the numer- ical abundance data were transformed [log10 (x + D] before analysis (Taylor 1953). Species were eliminated from cluster analysis if they occurred at <5<7c of the stations occupied during a sampling period. Although this is a more severe data reduc- tion than is commonly employed, examination of the raw matrix and trial runs at various cutoff levels indicated that species occurring below this level showed highly inconsistent distributions. The clustering strategy used was flexible fusion with beta set at the conventional value of -0.25 (Boesch 1977). Calculations were performed on an IBM 370-1153 at the Virginia Institute of Marine Science using the Fortran IV program COMPAH 3Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 296 COLVOCORESSES and MUSICK: CONTINENTAL SHELF DEMERSAL FISHES (Combinatorial Polythetic Agglomerative Hierar- chical Program) developed at the institution. Out- put was in the form of similarity matrices and computer generated dendrograms. The choice as to which branches in the dendro- grams were to be identified as biologically signifi- cant groups was based on the following procedure. Each branch of the dendrogram, composed solely of fusions involving only one entity as at least one-half of each fusion, was considered to consti- tute a minimal grouping. The distribution of each minimal grouping was then map-plotted, with logarithmic keyed symbols being used for plots of abundances of minimal species groupings. The plot of each grouping was then compared with that of the grouping with which it next fused; if no significant differences in distribution were evi- dent, the fusion was considered to be intragroup. This procedure was repeated until all minimal groupings had been fused into groups showing evident distributional differences. In cases where there was any doubt as to whether two groups should be fused, nodal analysis diagrams were generated and compared for the two cases and the decision producing the "crisper" result (Clifford and Stephenson 1975) utilized. While this method obviously involves some subjectivity in the recog- nition of groups, it has been pointed out by several authors that all methods of interpreting numeri- cal classifications require a certain degree of sub- jectivity and that fixed stopping rules are espe- cially inappropriate with fusion strategies which introduce a group size-dependent element into in- tergroup relative affinities (Boesch 1977; Pielou 1977; Clifford and Stephenson 1975). Two methods of nodal analysis were performed. The patterns of "constancy" and "fidelity" of species groups to site groups were expressed as relative densities of cells of a two-way table (Stephenson et al. 1972). Constancy is the propor- tion of the number of occurrences of each species group in the site group to the total number of occurrences possible (Boesch 1977). The index has a value of 1 when all members of a species group occur in all collections in a site group and a value of 0 when a species group does not occur in a given site group. Fidelity is a measure of the degree to which species groups are limited to site groups. The fidelity index used in this study was the constancy of a species group within a site group divided by the average constancy over all site groups. This index is unity when the con- stancy of a species group in a site group is equiv- alent to its overall constancy, >1 when its con- stancy in the site groups is greater than that overall, and between 0 and 1 when its constancy is less than its overall constancy. A chi-square test was applied to the fidelity value of each cell to determine whether it varied significantly (a = 0.05) from 1. Fidelity values significantly >1 indicate a positive association of species in a group with a site group, while values significant- ly <1 suggest a "negative" association. In the present analyses, a highly positive (or strong) association was inferred if the number of occur- rences of a given species group within a site group was twice that necessary to produce a fidel- ity value significantly >1, and a highly negative association was assumed when the number of occurrences was less than half that necessary to produce a fidelity value significantly <1. All nodal diagrams were drawn with the width of the rows and columns proportional to the number of entities in the respective site and species groups. Species Dominance Numerically dominant species have been used by ecologists for many years to characterize com- munities (Thorson 1957), and changes in domi- nant species often reflect faunal changes. In the present study, we have compared patterns of species dominance among site groups. A species was included in dominance comparisons if it oc- curred among the five most abundant species in at least 20% of all the stations from a site group. Faunal Affinities The faunal affinities of fishes captured were de- termined by examining published records of their usual ranges of occurrence (Bigelow and Schroeder 1953; Leim and Scott 1966; Struhsaker 1969; Musick 1972). Most warm-temperate species had resident populations south of Cape Hatteras in the "Carolinian" faunal province (Hazel 1970) and had their normal northern range limit some- where within the Middle Atlantic Bight south of Cape Cod. Boreal species had permanent popula- tions north of Cape Cod, and most had their south- ern range limit somewhere within the Middle At- lantic Bight north of Cape Hatteras. A few boreal species transcend Hatteras through bathymetric submergence. Certain components of the fauna tended to be residents on the inner shelf (Scophthalmus aquosus) or outer shelf (Par- alichthys oblongus). Many species were resident on the shelf edge and upper slope (Musick 1976). 297 Pooling of Within-Season Cruises Size of the data matrix was too large for simul- taneous clustering of either of the two multiple- year seasonal data sets. However, the results of the cluster, nodal, and dominance analyses of the individual cruises revealed a high degree of within-season repetition in the composition and distribution of species groups and in the faunal, geographic, and hydrographic attributes of site groups. Major repetitive species groups were rec- ognized for each season, and site groups for each year were referred to generalized seasonal site groups. The validity of these groups was examined by subjecting the pooled seasonal data sets to nodal and dominance analyses based on these groupings and comparing these results to those for the individual cruises. FISHERY BULLETIN: VOL. 82, NO. 2 RESULTS AND DISCUSSION Thermal Regime The geographic patterns of bottom-water tem- peratures were variable among years within both of the sampling seasons, although these differ- ences were minor compared with the seasonal vari- ation within a given year. Variability among years within a season can be attributed to two sources: Climatic differences among years and sampling artifacts (i.e., differences in the dates and duration of the sampling periods, and stochastic differences arising from the location of stations and the tem- poral sequence in which they were done). The spring cruises were conducted in March and April, the period during which water tempera- tures in the Middle Atlantic Bight are at a mini- "T ^TfOO - FIGURE 2.— Bottom isotherms for spring 1969 (A) and 1976 298 COLVOCORESSES and MUSICK: CONTINENTAL SHELF DEMERSAL FISHES mum (Walford and Wicklund 1968), and therefore it is more appropriate to consider these cruises as having sampled the late winter distribution of fishes (Musick and Mercer 1977). There was a definite trend toward warmer temperatures during the study period for this season which cannot be completely attributed to sampling artifacts (Davis 1979). Bottom isotherms extrapolated from the collection data are shown for two cruises representative of the warmer (1976) and cooler (1969) extremes (Fig. 2). During the 1969 cruise, inshore and mid-shelf temperatures were <4°C north of Delaware Bay and between 4° and 6°C between Delaware Bay and Cape Hatteras, with an increasing gradient present along the entire outer shelf. In 1976 temperatures of <6°C were encountered only at northern inshore stations. South of Chesapeake Bay there was a southwardly increasing thermal gradient perpendicular to the shoreline, and the outwardly increasing gradient was distributed across a greater portion of the shelf. Bottom temperatures for the other spring cruises exhibited patterns intermediate be- tween these two (Davis 1979). Fall sampling cruises were conducted primarily in October. Because of water column turnover, this is the time of maximum temperature for mid-shelf bottom waters in this region (Bigelow 1933); how- ever, coastal waters undergo rapid cooling during the fall (Parr 1933), initiating migrations for many fishes that spend the summer inshore. Bot- tom isotherms for a typical warm (1973) and cool (1971) fall sampling period are shown in Figure 3. In 1971 a strong thermal gradient was encoun- tered along the mid-shelf from New York to Cape Hatteras. A pocket of cooler water (6°-9°C) was present northward and offshore of this gradient, where turnover was in progress or just beginning. •" Vc 7**00 9 4«M^ CA»t "A'TfUAS —LLU (B) extrapolated from NMFS Groundfish Survey cruises. 299 FISHERY BULLETIN: VOL. 82, NO. 2 „. 42° 00 FIGURE 3.— Bottom isotherms for fall 1973 (A) and 1971 During 1973 bottom-water temperatures were less stratified and 2°-4°C warmer throughout most of the study area. Inshore temperatures exceeded 16° C along the entire Bight, with temperatures above 18°C occurring only south of Chesapeake Bay The coolest temperatures were found again on the mid-shelf off New Jersey and Long Island, but the "pocket" was much less clearly defined and was composed of waters between 10° and 12°C, indicating that turnover had already occurred. The other fall cruises showed thermal regimes intermediate to those of 1971 and 1973 (Davis 1979). Site Groups Spring Cruises Station groups based on cluster analysis were 300 determined for the nine spring cruises (Col- vocoresses and Musick 1979). Most groups were geographically contiguous and tended to be ther- mally and bathymetrically restricted. Site groups were not precisely comparable from one year to the next, but could, however, be categorized on the basis of faunal similarity, geographic location, bathymetry, and temperature. During all nine cruises there was a group of site clusters of similar depth and temperature regimes which were contained between the shore and ap- proximately the 8°C isotherm. The geographic ex- tent of these groups varied from year to year, but generally covered the inner- and mid-shelf out to about 70 m from Cape Cod south to between Dela- ware Bay and Cape Hatteras, depending upon the southward extent of waters cooler than 8°C. These site groups were assigned to site group I ( Fig. 4 ) for the pooled analyses. Adjacent to this group were COLVOCORESSES and MUSICK: CONTINENTAL SHELF DEMERSAL FISHES "T ,T6' 00 •OQ, (B) extrapolated from NMFS Groundfish Survey cruises. two other categories of groups: Northern outer shelf groups extending from the cold-water group to the shelf break (150 m) (pooled group II), and southern groups (pooled group III) which occupied the remaining shelf both outward and south of the 8°C isotherm. The boundary between these two categories was generally off the New Jersey coast, at which point there was often considerable over- lap. The remaining outermost groups were located along the shelf break at depths of 150-350 m (pooled group IV). In general, areas of geographic overlap between site groups can be related to variations in the thermal regime. For example, there is consider- able overlap between groups I and III on the inner- and mid-shelf south of Delaware Bay. This area showed the greatest temperature variation among years, with group I station clusters predominating in the area in colder years and group III station clusters in warmer years. Hydrographic parame- ters and basic catch data for each stratum are summarized in Table 1. The hydrographic parameters (depth, temperature) within a site group are much better represented by the mean and standard deviation than by the range of val- ues encountered. At a small percentage of stations only a few species were taken, and in cases where these species occurred within all or several strata, some misclassifications occurred. Because the in- cidence of these obvious misclassifications was low, they have been ignored rather than introduc- ing an arbitrary system of reclassification. Virtu- ally all extremely variant values of depth and temperature and strong deviation in geographic location within a site group were attributable to stations where only a few ubiquitous species were taken. Figure 5 illustrates temperature-depth en- velopes for each site group. In order to reduce dis- 301 FISHERY BULLETIN: VOL. 82, NO. 2 A ' 74°00' 42°00 76°00' 36°00 36°00' 71°00^ FIGURE 4.— Pooled site groups based on cluster analysis for spring NMFS Groundfish Survey cruises, 1968-76. tortions introduced by misclassified stations, points which exceeded 2 standard deviations from either mean were not included. As may be seen by a comparison of Figures 4 and 5, groups I and IV are geographically, bathymetrically, and ther- mally discrete from one another with groups II and III occupying the intermediate area and somewhat overlapping the first two groups in terms of bathymetry and thermal regime. Groups II and III are largely separable on the basis of latitude (as well as faunal composition). Fall Cruises Station groups recognized from cluster analysis of the fall cruises (Colvocoresses and Musick 1979) were not as geographically contiguous or as ther- mally restricted as during the spring cruises, but could still be readily grouped into categories based on faunal attributes. During seven of the nine cruises there was a distinct southern inshore site group between shore and about 60 m extending from Cape Hatteras northward to the region off 302 COLVOCORESSES and MUSICK: CONTINENTAL SHELF DEMERSAL FISHES TABLE 1. — Hydrographic and average catch parameters by site group for Spring NMFS Groundfish Survey cruises, Middle Atlantic Bight, 1968-76. The 1968-72 cruises used a #36 Yankee trawl, the 1973-76 cruises a #41 Yankee trawl. Numbers in parentheses are retransformed values. I II IV Site group 1968-72 1973-76 1968-72 1973-76 1968-72 1973-76 1968-72 1973-76 No. of stations Abundance log (x + 1) Biomass (kg) log(x * 1) No. of species Depth (m) Temperature (CC) x SD x SD x SD range x SD range x SD 237 2.19(154) 040 1.70(50) 0.40 10.1 28 18-101 50.0 17.3 2-9 46 1.5 188 2.47(296) 045 1.97(95) 036 12.4 3.0 18-90 46.5 15.4 3-11 6.0 1.5 92 2.40(252) 0.54 1.79(62) 0.54 97 29 24-329 117.9 46.8 4-14 100 2.1 90 2.62(416) 0.40 2.14(138) 0.43 12.1 2.5 29-152 78.9 21 9 7-16 9.6 2.0 138 2.17(149) 068 1.58(38) 0.59 8.1 29 18-349 84.1 542 5-13 8.9 22 53 2.35(224) 0.67 1 92(84) 0.51 88 3.3 27-152 75.2 332 5-14 10.2 2.1 110 2.16(144) 054 1.56(37) 0.57 9.1 3.3 66-379 222 1 78.3 5-16 10 1 2.0 112 2.36(230) 0.51 1.65(45) 0.46 13.1 4.3 53-341 194.0 72.1 5-15 11.4 16 Delaware Bay. This group was generally con- tained behind a strong thermal gradient and ex- 20 - U rr is z> < or UJ a. 'o UJ 1 -i i i SPRING - 1 — . — IV / / / / \ X \ s \ \ / 1 /— \ y 4 ^ y J 1 -- — — f .- — ^ / — ^ ^_ s 1 7 \ / i 1 1 1 100 200 300 400 DEPTH (m) FIGURE 5. — Temperature-depth envelopes for pooled spring site groups. Middle Atlantic Bight area, 1968-76. To avoid distortions introduced by misclassified stations, points falling over two standard deviations from either mean were excluded. hibited the warmest bottom temperatures in the study area. These groups were assigned to site group I for the pooled analyses (Fig. 6, Table 2). Extending northward from these groups along the inner shelf was a second, colder site group which tended to be constricted toward shore between northern Long Island and Cape Cod (group II). During 1973 and 1974, when thermal stratifica- tion was weaker and inshore water temperatures in the north were higher, there was no distinct break between northern and southern inshore sta- tion groups (groups I and II), but instead there were two station groups with members in both northern and southern inshore and mid-shelf waters. One group from each of these years was assignable to each of the two major pooled groups based on faunal similarity; but such assignment, of course, led to the geographical overlap between groups I and II seen off Long Island and Chesapeake Bay in Figure 6. One or two site groups each year occurred on the northern mid-shelf primarily between 35 and 90 m, in the region of the coolest shelf waters (group TABLE 2. — Hydrographic and average catch parameters by site group for Fall NMFS Groundfish Survey cruises, Middle Atlantic Bight, 1967-75. All cruises used a #36 Yankee trawl. Numbers in parentheses are retransformed values. Site group 1 II III IV V No. of stations 114 176 209 382 114 Abundance log(x + 1) X SD 2 19(130) 0.73 2.30(200) 0.59 2.20(249) 0.57 2.05(111) 0.67 1 92(84) 0.54 Biomass (kg) loglx + 1,1 X SD 1 .55(36) 0.58 1.78(61) 0.56 1.73(54) 0.59 1.09(11) 056 0.86(6) 0.39 No of species X SD 82 3.7 10.8 3.6 10.8 4.1 6.8 2.9 9.3 3.7 Depth (m) range X SD 18-123 33.8 12.7 20-80 42.6 12.4 31-192 61 5 17.1 16-397 110.6 60.2 71-433 2496 77.4 Temperature (°C) range X SD 8-23 16.7 3.5 6-25 13.4 3.5 5-22 10.7 2.6 6-21 11.7 2.2 6-18 10.4 1.9 303 FISHERY BULLETIN: VOL. 82, NO. 2 *74°00 ^42°00' 76°00 CAPE HAT TERAS 3 6° 0 0 \. 7 4° 00 \ FIGURE 6. — Pooled site groups based on cluster analysis for fall NMFS Groundfish Survey cruises, 1967-75. III). The remaining site groups could be classified as outer shelf-shelf break (group IV) or upper slope (group V). The outer shelf-shelf break group dis- played a wide depth range and a temperature range very similar to groups II and III, but oc- curred consistently offshore of those two groups (Fig. 6). The upper slope group had the most re- stricted temperature range and was bathymetri- cally discrete from the inner- and mid-shelf groups. The temperature-depth envelopes for the first four site groups (Fig. 7) show a large amount of overlap in the shallower portion of the study area, but much of this overlap is an artifact of combining data across years and over a wide area (i.e., thermal ranges and boundaries between groups varied between years, and bathymetric boundaries varied with latitude). Species Associations Between 6 and 11 species clusters were recog- nized for each cruise (Colvocoresses and Musick 1979). As with the station clusters, although there was some variation in group composition and dis- 304 COLVOCORESSES and MUSICK: CONTINENTAL SHELF DEMERSAL FISHES FALL I — . - I I ill iv V 100 200 300 400 DEPTH (m) FIGURE 7. — Temperature-depth envelopes for pooled fall site groups, Middle Atlantic Bight area, 1967-75. To avoid distortions introduced by misclassified stations, points falling over two standard deviations from either mean were excluded. tribution from year to year, the groupings were largely consistent over the 9-yr period of this study. Figure 8 shows the number of times the 30 most commonly occurring and dominant demersal species occurred within the same species group during the spring and fall cruises. The species are arranged so as to be closest to those species they occurred with most often in the clusters, i.e., so that the densest cells fall along the diagonal bor- der of the diagram. Four strongly recurring species groups are evi- dent from this diagram. Myoxocephalus oc- todecemspinosus, Scophthalmus aquosus, Raja erinacea, and Limanda ferruginea frequently ap- peared in the same group during both seasons. In the spring they were often joined by Macrozoarces americanus, a species generally absent from the FIGURE 8. — Cooccurrences within the same species cluster group for major species, spring and fall NMFS Groundfish Survey cruises. Middle Atlantic Bight area. 1967-76. NUMBER OF COOCCURENCES 305 FISHERY BULLETIN: VOL. 82, NO. 2 clusters in the fall, while Squalus acanthias and Pseudopleuronectes americanus were common co-group members during the fall cruises. In the spring the latter species regularly occurred in a separate group which included Gadus morhua and Hemitripterus americanus. Except for Scophthal- mus aquosus, an inshore resident, all of these species are of boreal faunal affinity and are re- stricted to cold water (Bigelow and Schroeder 1953; Leim and Scott 1966). Prionotus carolinus, Stenotomus chrysops, Paralichthys dentatus, and Centropristis striata, all warm-temperate species, were regularly clas- sified in the same group during both seasons. Dur- ing the fall this group was often joined by Mustelus canis, another warm-temperate species which was only rarely taken during the spring cruises. Two other warm-temperate species, Peprilus triacan- thus and Urophycis regia, regularly cooccurred with this group in the spring. Merluccius bilinearis and Urophycis chuss were the two most consistently cooccurring species, ap- pearing in the same group in all but one cruise. These two species formed the core of a third species group which was ubiquitous in the spring and widespread across the deeper portion of the study area in the fall. Both of these boreal species have broader temperature tolerances than the cold- water groups noted above (Musick 1974; Bigelow and Schroeder 1953). Abundances of these two species were greater on the outer shelf and shelf break, and they often clustered with Paralichthys oblongus ( = Hippoglossina oblonga), an outer shelf resident, and, in the fall, with Citharichthys arctifrons, a slope resident which also occurs on the outer shelf (Richardson and Joseph 1973) and Lepophidium cervinum, another outer shelf resi- dent. The warm-temperate species Peprilus triacanthus and Urophycis regia were also com- mon group members in the fall, while Lophius americanus regularly occurred in this group in the spring. The fourth clearly defined recurring species group was an upper slope group composed of Helicolenus dactylopterus, Chloropthalmus agas- sizi, and Merluccius albidus, which appeared con- sistently during both seasons. Urophycis tenuis commonly cooccurred with members of this group during the spring, while in the fall this species was more widely distributed across the outer shelf and tended to appear in small groups with Lophius americanus and Glyptocephalus cynoglossus . The major recurring species groups described above are listed for each season in Table 3. The groups are ordered in the same manner as the generalized station groups, that is, from shal- lowest to deepest (distribution) while still main- taining nearest neighbor intergroup relationships as determined in the clusters. Figures 9 and 10 show the distributional relationships between the major site and species groups as determined by nodal analyses. As noted above, these relation- ships are more sharply defined during the spring cruises than in the fall, but in both cases the nodal analyses show clear differences in the faunal com- position of site groups and the distribution of species groups. The interrelationships seen here are also highly representative of those noted dur- ing analyses of the individual cruises. Dominance The dominant species for each of the pooled site groups are given in Tables 4 and 5. During the spring Limanda ferruginea was the major domi- nant at the cold-water, inshore site group (I), Squalus acanthias and Merluccius bilinearis were among the major dominants at all site groups, and Peprilus triacanthus was a major dominant at all but the cold-water site group. Stenotomus chrysops was a major dominant along the southern outer shelf (group III). In the fall, the southern inshore site group (I) was strongly dominated by three warm-temperate species: Prionotus carolinus, Stenotomus chrysops, and Peprilus triacanthus. These three species persisted as major dominants at the northern inshore site group, but were joined there in roughly equal dominance by three boreal species: Limanda fer- ruginea, Squalus acanthias, and Merluccius bilinearis. Peprilus triacanthus and the latter group were major dominants on the northern mid-shelf (group III). Peprilus triacanthus and Merluccius bilinearis were also major dominants at the outer shelf stations (group IV), where they were joined by Urophycis regia. The shelf-break stations (group V) were dominated by Merluccius bilinearis, Citharichthys arctifrons , and Helicolenus dactylopterus. There were few major changes in species domi- nance throughout the study, and Tables 4 and 5 are representative of those for the individual cruises. Merluccius bilinearis, Peprilus triacan- thus, and Squalus acanthias were consistently the three most dominant species during both major seasons. Although Merluccius bilinearis ac- counted for only around 107c of the individuals taken, it was the most consistently dominant 306 COLVOCORESSES and MUSICK: CONTINENTAL SHELF DEMERSAL FISHES TABLE 3. — Major recurrent species groups, NMFS Groundfish Survey, Mid-Atlantic Bight area, 1967-76. Faunal affinity is designated after each species name: Boreal, Bo; warm temperate, WT; inner shelf resi- dent, IS; outer shelf resident, OS; slope resident, SI. Spring cruises Fall cruises Gadus morhua Bo Hemitripterus amencanus Bo Pseudopleuronectes americanus Bo B Limanda ferrugmea Bo Macrozoarces americanus Bo Myoxocephalus octodecemspinosus Bo Ra/a erinacea Bo Scopthalmus aquosus IS Lophius americanus Bo Merluccius bilineans Bo Paralichthys oblongus OS Squalus acanthias Bo Urophycis chuss Bo Centropristes striata WT Paralichthys dentatus WT Pepnlus tnacanthus WT Prionotus carolmus WT Stenotomus chrysops WT Urophycis regia WT Chloropthalmus agassizi SI Helicolenus dactylopterus SI Merluccius albidus SI Urophycis tenuis Bo-SI Centropristes striata WT Mustelus canis WT Paralichthys dentatus WT Prionotus carolinus WT Stenotomus chrysops WT B Umanda ferrugmea Bo Myoxocephalus octodecemspinosus Bo Pseudopleuronectes americanus Bo flaya erinacea Bo Scophthalmus aquosus IS Squalus acanthias Bo Cithanchthys arctifrons OS Lepophidium cervmum OS Merluccius bilinearis Bo Paralichthys oblongus OS Peprilus triacanthus WT Urophycis chuss Bo Urophycis regia WT Glyptocephalus cynoglossus Bo-SI Lophius americanus Bo Urophycis tenuis Bo-SI Chloropthalmus agassizi SI Helicolenus dactylopterus SI Merluccius albidus SI species, reflecting a very uniform distribution. Limanda ferruginea was the only major species to undergo a notable change in dominance, showing a pronounced decline only during the last 2 yr of the study. Parrack4 has carefully linked the de- cline of this valuable commercial species to overfishing. Squalus acanthias and Peprilus triacanthus, two of the most dominant species, showed strong seasonal differences in the groups with which they clustered. Squalus, a boreal cold-water species, was widespread in the spring and occurred in the ubiquitous group, but during the fall cruises this species was restricted to the cooler waters on the northern shelf and generally clustered with the Limanda -dominated cold-water group. Peprilus triacanthus generally appeared in the same group as the other warm-temperate species in the spring when it was distributed along the outer shelf, but in the fall this species was widespread across the 4Parrack, M. L. 1973. Current status of the yellowtail flounder fishery in ICNAF Subarea 5. Int. Comm. North w. Atl. Fish., Res. Doc. 73 104, Ser. No. 3067, 3 p. shelf and tended to be more concentrated in the cooler portions of the study area and usually clus- tered with the semi-ubiquitous Merluccius bilinearis-Urophycis chuss group. Peprilus triacanthus is considerably more tolerant of cooler temperatures than the other warm-temperate species encountered in this study (Horn 1970). Urophycis regia, another warm-temperate species which inhabits cooler waters (Struhsaker 1969), clustered similarly to Peprilus triacanthus, occur- ring with the warm-temperate group in the spring and with the semi-ubiquitous group in the fall; however, it appeared to have slightly narrower temperature tolerances, as it was more restricted to the southern portion of the outer shelf in spring and tended to be more concentrated in deeper, warmer waters in the fall. Absolute abundances, both of total catches and of individual species, varied to a much greater extent than did the relative abundances between species throughout the study. Because abundance trends for the fall cruises have been well docu- mented by Clark and Brown (1977) and the change 307 FISHERY BULLETIN: VOL. 82. NO. 2 SITE GROUPS a. O O a vl I H ur bz: ^■$^v**5s5&&&,'s*»sIkHk!^H ^m//////////y fl 1111 CONSTANCY ■ Very 70% High PH i 50% High ^ > 30% Moderote Low 10% □ < Very 10% Low B O a o UJ o Llj a. FIDELITY ■ -> 2 2 Highly Positive Positive [J 0 9, < Neutral n <- ■ Negat ve n < 045 Highly Negotive 1 1 SPRING CRUISES FIGURE 9.— Nodal constancy 1A1 and fidelity 1B1 diagrams showing the inter- relation between pooled site and species groups. NMFS Groundfish Survey spring cruises. 1968-76. TABLE 4. — Dominant species by site group for Spring NMFS Groundfish Survey cruises. Mid- Atlantic Bight area, 1968-76. A species was considered dominant if it occui red among the five most abundant species at at least 20'/ of all stations in the site group. Figures given are percentage of stations within each site group at which a species occurred ('< i and the average percentage that the species contributed towards total abundance of nonpelagic fishes (x7< i within the site group. Faunal affinities and species groups are as given in Table 3. Site group Species Faunal Species affinity group Bo A Bo A Bo B Bo B Bo B Bo B IS B OS C Bo C Bo C Bo C Bo C WT D WT D WT D WT D WT D WT D SI E SI E SL E Bo-SI E IV Gadus morhua Pseudopleuronectes americanus Limanda Ferruginea Macrozoarces americanus Myoxocephalus octodecemspinosus Rata ennacea Scophthalmus aquosus Hippoglossina oblonga Lophius americanus Merluccius bilmeans Squalus acanthias Urophycis chuss Centropristes striata Paralichthys dentatus Pepnlus triacanthus Pnonotus carolmus Stenotomus chrysops Urophycis regius Chloropthalmus agassizi Helicolenus dactylopterus Merluccius albidus Urophycis tenuis 44 1 4 38 2.2 88 28 5 22 1 4 68 5.2 56 5.0 77 11 9 52 1 5 70 4.7 29 1 1 84 76 44 24 63 52 58 0.7 53 1.4 79 20 5 97 22 4 66 130 90 272 73 1 1 1 87 30.0 82 24 6 58 122 54 3.9 84 9.3 25 40 47 1 9 42 2.2 74 9.5 75 148 57 12.3 56 194 50 7.1 51 9.8 24 2.1 50 156 48 7.2 35 31 59 38 38 2.0 1.6 6.7 3.1 1.6 308 COLVOCORESSES and MUSICK: CONTINENTAL SHELF DEMERSAL FISHES A SITE GROUPS co B Q- O K v> C UJ o UJ w D E wmmr CONSTANCY | 2 70 % Very High bi S > 50% High f^j > 30% Moderate D > l0% Low ] < 10% Very Low B FIGURE 10.— Nodal constancy (A) and fidelity (B) diagrams showing the inter- relation between pooled site and species groups. NMFS Groundfish Survey fall cruises, 1967-75. SITE GROUPS nr m CO 0- o o cc o CO UJ o UJ 0. CO FIDELITY ) 2 2 Highly Positive > I I Positive > 09, < I I Neutral < 09 Negative ] <_ 045 Highly Negative FALL CRUISES TABLE 5. — Dominant species by site group for Fall NMFS Groundfish Survey cruises, Mid-Atlantic Bight area, 1967-75. A species was considered dominant if it occurred among the five most abundant species at at least 20'/r of all stations in the site group. Figures given are percentage of stations within each site group at which a species occurred (% ) and the average percentage that the species contributed towards total abundance of nonpelagic fishes (x9c ) within the site group. Faunal affinities and species groups are as given in Table 3. Site group I II III IV V Faunal Species Species affinity group % x% % x% % x% % x% % x% Centropristes striata WT A 50 4.7 Mustelus canis WT A 38 5.7 Paralichthys dentatus WT A 61 6.7 51 13.5 Pnonotus carohnus WT A 75 33.3 61 10.1 Stenotomus chrysops WT A 53 17.5 51 13.5 30 2.4 Limanda ferruginea Bo B 65 14.1 79 18.0 Myoxocephalus octodecemspmosus Bo B 52 27 Pseudopleuronectes amencanus Bo B 69 3.3 47 1.7 Raja erinacea Bo B 64 3.8 58 2.7 Scophthalmus aquosus IS B 52 29 62 3.4 25 1.0 Squalus acanthias Bo B 69 16.8 77 11.3 23 5.1 Cithanchthys arctifrons OS C 35 27 49 6.4 68 11.8 Hippoglossina oblonga OS C 59 22 70 3.3 53 3.6 34 1.4 Lepophidium cervmum OS C 32 2.9 27 2.9 Merluccius bilmearis Bo C 67 10.2 92 23.0 80 20.8 58 11.3 Peprilus triacanthus WT C 59 15.8 65 11.1 72 196 65 26.5 28 5 1 Urophycis chuss Bo C 35 2.7 74 82 37 42 30 2.1 Urophycis regius WT C 40 6.5 57 14.8 40 7.9 Lophius americanus Bo D 31 22 57 3.0 Chloropthalmus agassizi SI E 44 5.9 Helicolenus dactylopterus SI E 84 21.2 Merluccius albidus SI E 65 7.7 309 FISHERY BULLETIN: VOL. 82, NO. 2 in nets makes a similar analysis of the spring cruises tenuous at best, we will not consider the topic further other than to note that average abundance and biomass were higher in the north- ern and inshore portion of the study area during both seasons (Tables 1, 2). CONCLUSIONS Despite large variation in the abundances of individual species, cluster analyses of 9 yr of sur- vey data have shown clear and consistent patterns of community composition and distribution among demersal fishes of the Middle Atlantic continental shelf. Allowing for thermal variation and misclas- sification of small catches, persistent site and species clusters have indicated the presence of four relatively constant and well-defined areas of faunal homogeneity in the spring and five more general areas in the fall, and five strongly recur- ring species associations during both seasons. The spring site groups can be described approx- imately as northern inner- and mid-shelf (I), ex- tending from shore out to about 60-80 m from Cape Cod to south of Delaware Bay; northern mid-shelf (ID, occupying from around 60-80 m out to about 150 m from Cape Cod to Hudson Canyon; southern outer shelf (III), 60-150 m, from Delaware Bay to Cape Hatteras; and outer shelf-shelf break (IV), >150 m. The southern inner and mid-shelf is a thermally related transition zone between groups I and III. The outer shelf between Delaware Bay and Hudson Canyon was also a transition zone (between groups II and III), but this discontinuity does not appear to be related directly to tempera- ture, but rather to the extent to which the north- ward migration of the warm-temperate species group has progressed by the time of the survey The five spring species groups contained one group specific to this season and four which con- tained common elements and properties with analogous fall groups. The first group (A) can be characterized as highly cryophilic, being virtually restricted to site group I and containing two mem- bers [Gadus morhua and Hemitripterus americanus) which were relatively absent from the study area during the fall. None of these species were major dominants, even within group I. The second group (B) is also composed of primar- ily boreal, cold-water species, but in this case is not completely restricted to site group I (although primarily distributed there) and contains the major dominant for that site, Limanda ferruginea . The third group (C) may be described as ubiqui- tous throughout the study area with moderate or better constancy to all site groups (Fig. 9). All members of this group are boreal or resident, and the major dominants, Merluccius bilinearis and Squalus acanthias , are the nuclear members. The fourth group (D) is composed entirely of warm- temperate members and is restricted to the warmer southern and outer shelf waters (site groups II-IV). Peprilus triacanthus and Stenotomus chrysops are the major dominants from this group. The last group (E) is composed strictly of weakly dominant slope species mostly confined to the shelf break site group (IV). The spring warming trend noted during the study period appeared to have no major effect on the composition and distribution of fish com- munities in the area other than the latitudinal division between the inshore site groupings. The results of the present study are very much in ac- cordance with the conclusions of Taylor et al. (1957) and Colton (1972) who found that while the ranges and distributions of certain species did shift with a changing thermal regime, there were no obvious overall changes in faunal composition. This is understandable when one considers that the average change encountered (about 2°C) is relatively small compared with the temperature tolerances of the species involved and the seasonal and geographic temperature variation encoun- tered. The five fall site groups can best be described as southern inner- and mid-shelf (I), extending out to about 60 m from Cape Hatteras to Delaware Bay and containing the area of warmest temperatures; northern inner shelf (II), extending northward from group I along a similar depth regime and containing cooler waters; northern mid-shelf (III), extending from group II out to about 90 m and occupying the area of the cold pool; outer shelf (IV), occupying the area between groups I and III and about 150 m; and shelf break (V), >150 m. While, again, with these groups there is some overlap (particularly with groups I and II as dis- cussed above), their definition is fairly good con- sidering the rapidly changing environmental con- ditions and migratory activity offish during this period. The fall species associations, as noted above, have much in common with those noted in the spring. The small cryophilic group is absent, but the terms applied to the other four spring groups may be applied here as well. An exclusively boreal-resident group (B) persists on the northern inner- and mid-shelf, including four members of 310 COLVOCORESSES and MUSICK: CONTINENTAL SHELF DEMERSAL FISHES the spring cold-water group B, one member of the cryophilic group, and Squalus acanthias, a ubiquitous dominant in the spring found only in the northern portion of the study area in the fall. The ubiquitous spring group (C) persists with Mer- luccius bilinearis the major dominant, and two other common members from the spring group, but the fall group is no longer exclusively boreal- resident in faunal affinity and the group is dis- tributed primarily in more northerly and deeper waters. Two warm-temperate species, Peprilus triacanthus and Urophycis regia, join this group as major dominants, while the other warm- temperate species, dominated by Prionotus carolinus and Stenotomus chrysops, continue to occur in the same group (A) but show a dramatic change in distribution, occurring on the inner shelf rather than the outer as in the spring. The shelf break group (E) shows the same composition and distribution as in the spring, while the fifth group (D), which did not occur in the spring, is composed of nondominant eurybathyic species which occur sporadically across all but the south- ern inner site group. It is obvious that although the two sampling periods included the two extremes of average water temperatures in the study area, the fall (warm extreme) is a much more dynamic period than the spring (cool extreme) for the fish com- munities in the region. This appears to be related in large part to the much less stable thermal re- gime encountered in the fall, particularly in shal- lower portions of the study area. Thermal gra- dients developed during the warmer months on the inner shelf are much stronger than those en- countered on the mid-shelf during the spring, and because cooling waters mix or turn over while warming waters stratify, the fall gradients break down much more quickly than those in the spring. As a result, a fish community in this region may be subjected to rapidly changing environmental temperatures by a number of factors. A relatively small shift of water masses in the vicinity of a strong thermal gradient, migration across a gra- dient, or rapid cooling and mixing along the gra- dient all subject these communities to abrupt changes of temperature (Parr 1933), and it is not surprising that the site groupings based on faunal similarities found in the inner portion of the study area during the fall exhibited wide temperature ranges (Fig. 7). Parr (1933) pointed out that the temperature-related distributions of organisms in the vicinity of a strong thermal gradient may be more influenced by the magnitude of short-term temperature changes than by the actual tempera- tures encountered. This concept may well have application to the formation of the three inner- most site groups identified during the fall; for al- though the groups strongly overlap with respect to the temperature ranges encountered, there is a considerable difference in the strength of the thermal gradients and presumably the short-term temperature variations encountered within each, with group I being primarily sited in the region of the sharpest gradients and group III being located in the most thermally stable area. The distributional patterns noted in this paper lead to the conclusion that continental shelf de- mersal fish communities in the Middle Atlantic Bight are largely structured by temperature on the inner- and mid-shelf and by depth on the outer shelf and shelf break. This is not at all unexpected considering the sedimentary and topographical uniformity of the inner- and mid-shelf (Emery and Uchupi 1972) and the large annual variation in bottom-water temperature in the inshore region, with the converse holding true along the outer shelf and shelf break. Scott (1982) found the dis- tributions of a number of groundfish species on the Scotian Shelf to be related to bottom sediment type. Although substrate preference indices were not generated during the present study, compari- sons of species group distribution with bottom sed- iment type maps do not indicate any strong species group-sedimentary relationships. This contrast may be the result of two major differences between the continental shelves in the Middle Atlantic Bight and off Nova Scotia; there is a much more variable sedimentary environment and a consid- erably smaller annual range of bottom-water temperatures on the Scotian Shelf. Tyler (1971) examined latitudinal variation in the regular and seasonal components of several nearshore Atlantic marine fish communities, and concluded that the proportion of seasonal and occasional components to regular components var- ied directly with annual variation in water tem- perature. The results of the present study are cer- tainly in accord with this conclusion, in that the most highly variable area in terms of annual wa- ter temperature variation (the southern inner- and mid-shelf) was also the most variable area in terms of community composition, but it is also evident that Tyler's statement cannot be taken axiomatically The outer shelf, although very homothermic, was also subject to considerable seasonal variation in community structure be- cause of the changing relationship between the 311 FISHERY BULLETIN VOL. 82, NO. 2 stable thermal regime on the outer shelf and the highly varying regime in adjacent inshore waters. During the spring, when inshore water tempera- tures were depressed well below those on the outer shelf, the outer shelf served as a refuge for the warm-temperate species association which occurs largely inshore when water temperatures there become elevated above those on the outer shelf. It is also interesting to note that while for the most part the communities observed here are structured by species associations that behave as a group in response to environmental variation, two of the most successful species (Peprilus triacan- thus and Squalus acanthias) are those which show the least permanent group affinities. As noted above, the success of P. triacanthus may be due in part to the species' very wide thermal tolerance, but S. acanthias was one of the more thermally restricted species encountered in the study, being restricted to waters less than 14°C. ACKNOWLEDGMENTS We would like to thank Marvin Grosslein and the staff of the Groundfish Survey Unit, Northeast Fisheries Center (NEFC) Woods Hole Laboratory, for their helpful assistance in making this data base and supporting information available to us, and also to express our appreciation to the hun- dreds of individuals who have participated in the survey over the years. Eric Foell and William Bly- stone provided a large measure of assistance in data analyses and computer programming, re- spectively. Marvin Grosslein and the staff at NEFC, particularly William Overholtz, provided a helpful review of the manuscript, for which we are also indebted to Jim Price and Eric Anderson of the Virginia Institute of Marine Science. Data analyses for this study were supported by contract No. AA550-CT6-62 of the Bureau of Land Man- agement. LITERATURE CITED BIGELOW, H. B. 1933. Studies of the waters on the continental shelf, Cape Cod to Chesapeake Bay. I. The cycle of temperature. Pap. Phys. Oceanogr. Meterol. 2(41:1-135. BIGELOW, H. B., AND W. C. SCHROEDER. 1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 53:1-577. BOESCH, D. F. 1977. Application of numerical classification in ecological investigations of water pollution. E.RA., Ecol. Res. Ser. EPA-600/3-77-033, 115 p. 312 Clark, S. H., and b. E. Brown. 1977. Changes in biomass of finfishes and squids from the Gulf of Maine to Cape Hatteras, 1963-74, as determined from research vessel survey data. Fish. Bull., U.S. 75:1- 21. CLIFFORD, H. T., AND W. STEPHENSON. 1975. An introduction to numerical classification. Acad. Press, N.Y., 229 p. COLTON, J. B., JR. 1972. Temperature trends and the distribution of ground- fish in continental shelf waters, Nova Scotia to Long Is- land. Fish. Bull., U.S. 70:637-657. COLVOCORESSES, J. A., AND J. A. MUSICK. 1979. Historical community structure analysis of finfishes. Section II: NMFS Groundfish Survey. Va. Inst. Mar. Sci. Spec. Rep. Appl. Mar. Sci. Ocean Eng. 198, 211 p. DAVIS, C. W. 1979. Bottom-water temperature trends in the Middle At- lantic Bight during spring and autumn, 1964-76. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-739, 13 p. Day, d. S., and W. G. Pearcy. 1968. Species associations of benthic fishes on the conti- nental shelf and slope off Oregon. J. Fish. Res. Board Can. 25:2665-2675. Emery, k. o., and e. uchupi. 1972. Western North Atlantic Ocean: topography, rocks, structure, water, life, and sediments. Am. Assoc. Petr. Geol, Tulsa, Okla.,532p. FAGER, E. W, AND A. R. LONGHURST. 1968. Recurrent group analysis of species assemblages of demersal fish in the Gulf of Guinea. J. Fish. Res. Board Can. 25:1405-1421. GROSSLEIN, M. D. 1969. Groundfish survey program of BCF Woods Hole. Commer. Fish. Rev. 31(8-9):22-35. HAZEL, J. E. 1970. Atlantic Continental Shelf and slope of the United States — Ostracod zoogeography in the southern Nova Scotian and northern Virginian faunal provinces. U.S. Geol. Surv. Prof. Pap. 529-E, 21 p. HORN, M. H. 1970. Systematics and biology of the stromateid fishes of the genus Peprilus. Bull. Mus. Comp. Zool. 140:165-261. Lambert, J. M., and W. T. Williams. 1962. Multivariate methods in plant ecology. IV. Nodal analysis. J. Ecol. 50:775-802. Lance, G. N., and W. T. Williams. 1967. Mixed-data classificatory programs. I. Agglomera- tive systems. Aust. Computer J. 1:15-20. leim, a. h., and w. b. Scott. 1966. Fishes of the Atlantic coast of Canada. Fish. Res. Board Can. Bull. 155, 485 p. MUSICK, J. A. 1972. Fishes of Chesapeake Bay and the adjacent coastal plain. In M. L. Wass (compiler), A check list of the biota of lower Chesapeake Bay, p. 175-212. Va. Inst. Mar. Sci., Spec. Sci. Rep. 65. 1974. Seasonal distribution of sibling hakes, Urophycis chuss and U. tenuis (Pisces, Gadidae) in New En- gland. Fish. Bull., U.S. 72:481-495. 1976. Community structure of fishes on the continental slope and rise off the middle Atlantic Coast of the United States. Abstr. Joint Oceanogr. Assembly, Edinburgh, Scotland, 146 p. MUSICK, J. A., AND L. R MERCER. 1977. Seasonal distribution of black sea bass, Centropristis COLVOCORESSES and MUSICK: CONTINENTAL SHELF DEMERSAL FISHES striata, in the mid- Atlantic Bight with comments on the ecology and fisheries of the species. Trans. Am. Fish. Soc. 106:12-25. Parr, A. E. 1933. A geographic-ecological analysis of the seasonal changes in temperature conditions in shallow water along the Atlantic coast of the United States. Bull. Bingham Oceanogr. Collect.. Yale Univ. 4(31:1-90. PIELOU, E. C. 1977. Mathematical ecology. 2d ed. John Wiley & Sons, Inc., N.Y., 385 p. Richardson, s. L., and e. b. Joseph. 1973. Larvae and young of western north Atlantic bothid flatfishes Etropus microstomas and Citharichthys arcti- frons in the Chesapeake Bight. Fish. Bull., U.S. 71:735- 767. SAUSKAN, V. I., AND V. M. RYZHOV. 1977. Investigation of communities of demersal fish of Campeche Bank. Oceanology 17:223-227. SCOTT, J. S. 1982. Selection of bottom type by groundfishes of the Sco- tian Shelf. Can. J. Fish. Aquat. Sci. 39:943-947. SNEATH, P. H. A., AND R. R. SOKAL. 1973. Numerical taxonomy. W. H. Freeman and Co., San Franc, 573 p. STEPHENSON, W, W. T. WILLIAMS, AND S. D. COOK. 1972. Computer analyses of Petersen's original data on bottom communities. Ecol. Monogr. 42:387-415. STRUHSAKER, P 1969. Demersal fish resources: Composition, distribution, and commercial potential of the continental shelf stocks off Southeastern United States. U.S. Fish Wild!. Serv., Fish. Ind. Res. 4:261-300. Taylor, C. C. 1953. Nature of variability in trawl catches. U.S. Fish Wildl. Serv., Fish. Bull. 54:145-166. Taylor, C. C, H. B. Bigelow, and H. G. Graham. 1957. Climatic trends and the distribution of marine ani- mals in New England. U.S. Fish Wildl. Serv.. Fish Bull. 57:293-345. THORSON, G. 1957. Bottom communities tsublittoral or shallow shelf). In J. W Hedgepeth (editor), Treatise on marine ecology and paleoecology, Vol. I, 461-534. Geol. Soc. Am., Mem. 67. Tyler, a. V 1971. Periodic and resident components in communities of Atlantic fishes. J. Fish. Res. Board Can. 28:935-946. WALFORD, L. A., AND R. I. WICKLUND. 1968. Monthly sea temperature structure from the Florida Keys to Cape Cod. Am. Geogr. Soc. Ser. Atlas Mar. Envi- ron., Folio 15, 2 p. 313 EARLY ZOEAL STAGES OF PLACETRON WOSNESSENSKII AND RHINOLITHODES WOSNESSENSKII (DECAPODA, ANOMURA, LITHODIDAE) AND REVIEW OF LITHODID LARVAE OF THE NORTHERN NORTH PACIFIC OCEAN Evan B. Haynes1 ABSTRACT Stage I zoeae of Placetron wosnessenskii , and Stage I and Stage II zoeae of Rhinolithodes wosnes- sensku , which were reared in the laboratory, can be distinguished from other described zoeae of Lithodidae: P. wosnessensku have long, blunt spines on posterior margins of abdominal somites 2-5 and sinuate curvature of long, blunt, posterolateral spines on abdominal somite 5; R. wosnessensku zoeae have a spine in the middorsal, posterior portion of the carapace. Zoeae of Lithodidae can be distinguished from zoeae of Pagunnae by body shape, size of the eyes, spines on the carapace, devel- opment of uropods, and presence or absence of the anal spine. Stages of lithodid zoeae can be distin- guished by eye attachment, number of natatory setae on maxillipeds, and development of pleopods, uropods. and telson. Keys, based on spination of the carapace, rostrum, abdomen, and telson, distin- guish between zoeae and glaucothoe of each described species of Lithodidae from the northern North Pacific Ocean. Crabs of the family Lithodidae constitute a major component of the reptant decapod fauna of the northern North Pacific Ocean. Of about 25 species of Lithodidae in the northern North Pacific Ocean, larvae have been described, at least in part, for eight species: Dermaturus mandtii Brandt, Cryptolithodes typicus Brandt, Hapalogaster grebnitzkii Schalfeew, H. mertensii Brandt, Lith- odes aequispina Benedict, Paralithodes breuipes (Milne Edwards and Lucas), P. camtschatica (Tilesius), and P. platypus Brandt. Most descrip- tions are scattered in foreign scientific journals, however, and published reviews of the larvae are limited in species and scope. This report describes and illustrates Stage I zoeae of Placetron wosnes- sensku Schalfeew and Stages I and II zoeae of Rhinolithodes wosnessenskii Brandt reared in the laboratory from ovigerous females. I characterize the morphological differences between zoeae of the Lithodidae and subfamily Pagurinae (family Paguridae), compare the morphology of lithodid larvae of the northern North Pacific Ocean, and provide keys for identifying the described larvae to species and stage. 'Northwest and Alaska Fisheries Center Auke Bay Labora- tory, National Marine Fisheries Service, NOAA, P.O. Box 210155. Auke Bay, AK 99821. Manuscript accepted October 1983. FISHERY BULLETIN: VOL. 82. NO. 2, 1984. METHODS AND RESULTS In March 1982, ovigerous females of Placetron wosnessenskii and Rhinolithodes wosnessenskii were collected near Auke Bay, Alaska, in traps and by divers using scuba. The females were transported to the laboratory and kept in filtered seawater (about 6°C) until the zoeae hatched about 1 wk later. After hatching, about 50 zoeae of each species were transferred to each of four 4 1 glass jars containing about 2,500 ml of seawater at 6.1° C. Seawater in the jars was changed about every other day. Zoeae were fed live plankton strained through a 0.333 mm mesh. About 10 ml of live plankton was added to each jar every other day. The live plankton consisted mostly of phyto- plankton and barnacle nauplii. A more detailed description of the rearing system and type and duration of illumination is given in Haynes and Ignell (1983). Zoeae of Placetron wosnessenskii hatched at night, and samples of zoeae were taken the follow- ing morning. No prezoeae of P. wosnessenskii were seen. Rhinolithodes wosnessenskii zoeae hatched at night and during the day, and those examined about 10 min after hatching had rem- nants of prezoeal exuviae attached to the cephalo- thorax and telson. The remnant exuviae are not described in this paper. 315 3^1 FISHERY BULLETIN: VOL. 82, NO. 2 Although food was seen in the guts of some Placetron wosnessenskii zoeae, none molted to Stage II. Zoeae of Rhinolithodes wosnessenskii fed actively and molted to Stage II about 20 d after hatching. Failure to change the seawater on schedule prevented rearing the zoeae of R. wos- nessenskii beyond Stage II. DESCRIPTION OF ZOEAE Terminology, methods of measuring zoeae and their appendages, techniques of illustration, and nomenclature of appendages follow Haynes ( 1979). Carapace length refers to the straight-line dis- tance from posterior margin of eye orbit to mid- dorsal posterior margin of carapace, excluding the middorsal spine. Spines on the telson are num- bered from the outermost to innermost (medial) pair. Setation formulae are the number of setae per segment from the distal segment to the proxi- mal segment. For clarity in the illustrations, setules on plumose setae are usually omitted, but 0. 5 mm FIGURE l.— Stage I zoea of Placetron wosnessenskii: A, whole animal, right side; B, carapace, dorsal; C, antennule, ventral; D, antenna, ventral; E, mandibles (left and right), posterior; F, maxillule, ventral; G, maxilla, dorsal; H, first maxilliped, lateral; 316 HAYNES: EARLY ZOEAL STAGES OF LITHODID CRABS spinulose setae are shown. Five zoeae were used to verify segmentation and setation; 10 zoeae were used for measurements. Only those morphological characteristics useful for readily identifying each stage are given. Placet rou wosnessenskii — Stage I Zoeae Mean carapace length, 2.12 mm (range 2.08- 2.21 mm, 10 specimens); mean total length, 6.22 mm (range 5.90-6.70 mm, 10 specimens) (Fig. 1A, B). Live zoeae orange throughout except for color- less appendages and posterolateral spines on carapace and abdomen. Carapace with medially curved, long (>l/4 carapace length), posterolat- eral spines; markedly pronounced lateral ridges and dorsoventral ridge; and two middorsal angu- lar prominences: one near center of carapace and other at posterior edge. No supraorbital spines. Eyes sessile. 1 .0 mm I, second maxilliped, lateral; J, third maxilliped, lateral; K, pereopods 1-5, lateral; L, abdomen and telson, dorsal. 317 FISHERY BULLETIN: VOL. 82, NO. 2 Antennule (Fig. 1C). — First antenna (antennule) with unsegmented tubular basal portion (pedun- cle) and distal conical projection. Peduncle with ventral plumose seta. Conical projection with sev- en aesthetascs and two simple setae terminally. Antenna (Fig. ID). — Second antenna (antenna) with inner flagellum (endopodite) and outer an- tennal scale (exopodite). Flagellum unsegmented, shorter than scale, and tipped with two simple setae. Antennal scale without distal joint, has fringe of 10 heavily plumose setae along terminal inner margin and prominent spine on distal outer margin. Ventral surface of protopodite with spin- ulose spine at base of flagellum and naked spine at base of antennal scale. Mandible (Fig. IE). — Incisor process of right mandible a tooth; left mandible with biserrate in- cisor process. Anterior margins of each mandible with premolar denticles. Mandibles without sub- terminal processes or movable premolar denticle (lacinia mobilis). Maxillule (Fig. IF). — First maxilla (maxillule) with coxopodite, basipodite, and endopodite. Cox- opodite (proximal lobe) unsegmented with four large spinulose spines and three simple spines terminally. Basipodite (median lobe) with three spines terminally (each spine with several spin- ules) and two simple spines subterminally. Three- segmented endopodite originates from lateral margin of basipodite. Endopodite with three setae terminally, a long distal seta on second segment, and a short distal seta on first segment. Fine hairs on inner and outer margins of exopodite, outer margin of endopodite, both lobes of basipodite, and distal lobe of coxopodite. Maxilla (Fig. 1G). — Second maxilla (maxilla) with platelike exopodite (scaphognathite). Exo- podite with three long plumose setae terminally and a subterminal plumose seta on outer margin; no proximal expansion of exopodite. Endopodite unsegmented, setation formula 3, 1, 3. Basipodite and coxopodite bilobed. Basipodite with four setae on distal lobe and five setae on proximal lobe. Coxopodite with four setae on distal lobe and eight (sometimes seven) setae on proximal lobe. Fine hairs on inner and outer margins of scaphogna- thite, outer margins of endopodite, and distal mar- gins of basipodite and coxopodite. First maxilliped (Fig. 1H). — Exopodite partially 318 segmented with four natatory setae. Endopodite slightly longer than exopodite and distinctly five segmented; setation formula 5, 3, 1, 2, 3. Protopo- dite unsegmented with 10 setae. Second maxilliped (Fig. II). — Similar to first maxilliped except endopodite slightly shorter than exopodite. Endopodite four segmented, seta- tion formula 5, 2, 2, 2. Protopodite with three lateral setae. Third maxilliped (Fig. 1J). — Exopodite and en- dopodite undeveloped. Exopodite partially seg- mented, with three undeveloped setae terminally. Endopodite with undeveloped seta terminally. Pereopods (Fig. IK). — Poorly developed, without exopodites. First pereopod bilobed. Fifth pereo- pod arises medially between first and second pereopods. Abdomen and telson (Fig. 1A,L). — Abdomen with five somites and telson (somite 6 fused with telson). Somites 2-5 have six bluntly tipped spines on posterior margin and two minute dorsal setae. Outer pair of posterior spines on somite 5 are long (about 1.2 times somite width), blunt, and some- what sinuate. Telson with medial invagination posteriorly and 6 + 6 spines. Third pair of telsonic spines longest (about 3/4 maximum telson width). All spines jointed with telson. Minute seta be- tween spinal pairs 1 and 2 originates from ventral surface; seta often without setules; spinules on spinal pairs 2-6. No uropods or anal spine. Rhinolitbodes wosnessenskii Stage I Zoeae Mean carapace length, 1.29 mm (range 1.21-1.34 mm, 10 specimens); mean total length, 4.45 mm (range 4.02-4.62 mm, 10 specimens) (Fig. 2A, B). Cephalothorax and base of maxillipeds orange; remainder of maxillipeds, most of rostrum, and all of abdomen colorless. Carapace with middorsal angular prominence and spine at middorsal poste- rior margin; medially curving, long (>l/4 cara- pace length), posterolateral spines; markedly pro- nounced lateral ridge. No supraorbital or anal spine. Eyes sessile. Antennule (Fig. 2C). — Distal conical projection unsegmented from peduncle. Peduncle with ven- tral plumose seta. Conical projection with seven HAYNES: EARLY ZOEAL STAGES OF LITHODID CRABS 0. 5 mm 0. 5 mm 0.5 mm 0. 5 mm 1 .0 mm FIGURE 2. — Stage I zoea of Rhinolithodes wosnessenskii: A, whole animal, right side; B, carapace, dorsal; C, antennule, ventral; D, antenna, ventral; E, mandihles (left and right*, posterior; F. maxillule. ventral; G, maxilla, dorsal; H, abdomen and telson, dorsal. 319 FISHERY BULLETIN: VOL. 82, NO. 2 aesthetascs and two simple setae (one terminal and one lateral). Antenna (Fig. 2D). — Antenna with inner flagel- lum and outer antennal scale; flagellum without setae and shorter than scale. Antennal scale un- jointed distally, fringed with six heavily plumose setae along terminal and inner margins, and prominent spine distally on outer margin. Ventral surface of protopodite with spinulose spine at base of flagellum and smaller naked spine at base of scale. Mandible (Fig. 2E). — Incisor processes of left and right mandibles a single tooth. Anterior margins of each mandible with premolar denticles. No sub- terminal processes or movable premolar denticles. Maxillule (Fig. 2F). — Similar to Stage I Placetron wosnessenskii except spines of basipodite less spinulose and proximal segment of endopodite with two simple setae terminally instead of one. Maxilla (Fig. 2G). — Scaphognathite with seven long plumose setae on outer margin, no proximal expansion, setation formulae of endopodite, basi- podite, and coxopodite same as in Stage I Place- tron wosnessenskii. Fine hairs on margins of basi- podite and coxopodite. Maxillipeds 1-3 and pereopods 1-5. — Nearly iden- tical in shape and number of setae to those of Stage I Placetron wosnessenskii. Pleopods. — Absent. Abdomen and telson (Fig. 2A, H). — Short blunt spines on abdominal somites 2-5; length of outer pair on somite 5 about 0.8 times maximum width of somite. Telson with medial invagination and 7 + 7 posterior spines. Third pair of telsonic spines longest, about 3/4 maximum telson width; minute seta between spinal pairs 1 and 2 sometimes with- out setules; spinules on spinal pairs 2-6. No uro- pods or anal spine. Stage II Zoeae Mean carapace length, 1.30 mm (range 1.21-1.34 mm, 10 specimens); mean total length, 4.81 mm (range 4.02-5.03 mm, 10 specimens). No supra- orbital spine. Eyes stalked. Characters not men- tioned are nearly identical to characters of Stage I. Antennule. — Distal conical projection segmented from peduncle. Antenna. — Tip of flagellum may have small spine. Antennal scale with seven plumose setae along terminal and inner margins. Mandible. — Right mandible with five teeth be- tween incisor and molar processes. Maxillule. — Basipodite with five terminal spines (four slightly spinulose). Maxilla. — Scaphognathite with 11 plumose setae on outer margin; no proximal expansion. Setation formulae of endopodite 4, 1, 3. First and second maxillipeds. — Exopodite with seven natatory setae terminally. Third maxilliped (Fig. 3). — Exopodite with six plumose setae terminally. Endopodite with two plumose setae terminally. Pleopods. — May be present as minute buds. Abdomen and telson. — Identical to Stage I except joint between somite 6 and telson indicated by small indentation in lateral margins. 0.5 mm FIGURE 3.— Stage II zoea of Rhinolithodes wosnessenskii: third maxilliped, lateral. 320 HAYNES: EARLY ZOEAL STAGES OF LITHODID CRABS DISTINCTION BETWEEN ZOEAE OF LITHODIDAE AND PAGURINAE Zoeae of the family Lithodidae have long been considered similar morphologically to those of the subfamily Pagurinae (family Paguridae) and dif- fer only in reduction or disappearance of the uropods (Gurney 1942; MacDonald et al. 1957). Recent descriptions of zoeae of Cryptolithod.es typicus, Lithodes aequispina, L. antarctica, and Paralomis granulosa (Hart 1965; Haynes 1982; Campodonico 1971; Campodonico and Guzman 1981) have extended the range of zoeal characters of the Lithodidae and show that zoeae of the Lithodidae and Pagurinae can be distinguished by size of the eyes and morphology of the carapace and abdominal appendages (Table 1). In general, zoeae of the Lithodidae (except Cryptolithodes typicus ) are characterized by stoutness, small eyes, posterolateral spines in middle or lower half of carapace, uniramous uropods, and no anal spine. Zoeae of the Pagurinae are characterized by slenderness, large eyes, posterolateral spines in the middle or upper half of the carapace, biramous uropods, and an anal spine. The glau- cothoe of the Lithodidae and Pagurinae are read- ily distinguished from each other by their simi- larity to the adults (Haynes 1982). TABLE 1. — Characters useful for distinguishing between zoeae of Lithodidae and zoeae of Pagurinae from the northern North Pacific Ocean. Zoeae of Cryptolithodes typicus (Lithodidae i are an exception and are not characterized in this table. Lithodidae Pagurinae 1 . General appearance, stout. 2 Longitudinal diameter of eye less than width of abdomen. 3. Posterolateral spines of carapace in middle or lower half of posterior margin. 4 Lateral margins of carapace nearly parallel 5. No anal spine in any stage 6. More than eight pairs of telsonic spines in some species (excluding minute seta). 7. Uropods (when present) uniramous and terminal margin blunt with (usually three or four) short setae. General appearance, slender Longitudinal diameter of eye greater than width of abdomen. Posterolateral spines of carapace in middle or upper half of posterior margin. Lateral margins of carapace converge posteriorly. Anal spine present until Stage III in some species Never more than eight pairs of telsonic spines (excluding seta). Uropods biramous. exopodite styhform terminally with usually more than three or four long setae along medial margin. MORPHOLOGY OF LITHODID LARVAE Lithodidae of the northern North Pacific Ocean have four zoeal stages and a glaucothoe. Stage I zoeae are characterized by sessile eyes, four nata- tory setae on maxillipeds 1 and 2; maxilliped 3 is undeveloped and without natatory setae; pleopods and uropods are absent; and the telson and abdom- inal somite 6 are fused. Beginning in Stage II, the eyes are movable, and maxillipeds have at least six natatory setae. In Stage III, undeveloped pleopods and uropods are present, and the telson and abdominal somite 6 are articulated. In Stage IV, the pleopods are biramous, and the uropods are two segmented and usually have three or four apical setae. Table 2 and the keys are provided for distinguishing described zoeae and glaucothoe of Lithodidae of the northern North Pacific Ocean. Glaucothoe of H. grebnitzkii, P. wosnessenskii, and R. wosnessenskii have not been described. TABLE 2. — Characters useful for distinguishing between Stages I-IV of lithodid zoeae of the northern North Pacific Ocean. Paralithodes brevipes may have only three zoeal stages (Kurata 1956), thus, may not always conform to the descriptions in this table. Stage Characteristic 1 11 III IV Eyes sessile movable movable movable Natatory setae First maxilliped 4 -6 •6 ■6 Second maxilliped 4 •6 •6 •6 Third maxilliped 0 ?6 ■6 •6 Pleopods absent absent absent or present as buds present Uropods absent absent present; present; two unsegmented segmented Telson and sixth abdominal somite fused fused articulated articulated Described Lithodid Zoeae of the Northern North Pacific Ocean la. Carapace without posterolateral spines; uropods absent in all stages; posteri- or margin of telson without medial invagination Cryptolithodes typicus lb. Carapace with posterolateral spines; uropods present in later stages (usually Stages III and IV); posterior margin of telson with medial invagination 2 2a. Posterolateral spines of carapace short (< 1/4 carapace length) 3 2b. Posterolateral spines of carapace long ( > 1/4 carapace length) 5 3a. Posterolateral spines and denticles on abdominal somites 3 and 4 about same length; posterior margins of carapace concave Hapalogaster grebnitzkii 3b. Posterolateral spines obviously longer than denticles on abdominal somites 321 FISHERY BULLETIN: VOL. 82, NO. 2 3 and 4; posterior margin of carapace convex 4 4a. Carapace length 1.2-1.4 mm; antennal flagellum and scale (including distal spine) about same length; antennal scale < 5 times as long as wide Dermaturus mandtii 4b. Carapace length 1.4-1.7 mm; antennal flagellum longer than antennal scale (including distal spine); antennal scale about 9 times as long as wide Paralithodes brevipes 5a. Carapace with middorsal posterior spine Rhinolithodes wosnessenskii 5b. Carapace without middorsal posterior spine 6 6a. Antennal scale with s 6 markedly short ( 1/2 scale width), heavily plumose setae 7 7a. Posterolateral spines of carapace project somewhat laterally; telson has ^11 pairs of spines (excluding minute hair); longest (third) pair of telsonic spines fused to telson Lithodes aequispina 7b. Posterolateral spines of carapace do not project laterally; telson has ^ 8 pairs of spines (excluding minute hair); longest (third) pair of telsonic spines jointed with telson 8 8a. Spines on posterior margins of abdominal somites 2-5 markedly long and tips blunt; posterolateral spine on abdominal somite 5 blunt and sinuate Placetron wosnessenskii 8b. Spines on posterior margins of abdom- inal somites 2-5 typically short and tips pointed; posterolateral spine on abdom- inal somite 5 pointed and not sinuate .... 9 9a. Telsonic spines 8 + 8 (excluding minute hair) Paralithodes platypus 9b. Telsonic spines 7 + 7 (excluding minute hair) Paralithodes camtschatica Described Lithodid Glaucothoe of the Northern North Pacific Ocean. Glaucothoe of Hapalogaster grebnitzkii, Placetron wosnessenskii, and Rhinolithodes wosnessenskii have not been described. la. Dorsal surface of carapace without spines 2 322 lb. Dorsal surface of carapace with spines 4 2a. Carapace triangular Cryptolithodes typicus 2b. Carapace rectangular 3 3a. Lateral margin of carapace with teeth in branchial region but not in hepatic region Hapalogaster mertensii 3b. Lateral margin of carapace with teeth in branchial and hepatic regions Dermaturus mandtii 4a. Tips of anterolateral spines of rostral complex spinulose; most, if not all, spines on dorsal surface of carapace bifid Lithodes aequispina 4b. Tips of anterolateral spines of rostral complex styliform or bifid; most, if not all, spines on dorsal surface of carapace styliform 5 5a. Carapace with 15 pairs of spines on dorsal surface Paralithodes platypus 5b. Carapace with < 15 pairs of spines on dorsal surface 6 6a. Carapace with 14 pairs of spines on dorsal surface . . Paralithodes camtschatica 6b. Carapace with 13 pairs of spines on dorsal surface Paralithodes brevipes Paralithodes brevipes may have three stages; thus, Table 2 may not always be appropriate for distinguishing the stages of this species. Kurata (1956) reared and described the larvae of P. brevipes from ovigerous females collected in Jap- anese waters. In Kurata's description, P. brevipes has three zoeal stages instead of the four that characterize the genus, and Stage III zoeae cor- respond morphologically to Stage IV zoeae of the genus. Makarov (1967), however, found four zoeal stages of P. brevipes, including a Stage III zoea, in plankton of the west Kamchatkan coast that correspond morphologically to Stage III zoeae of the genus. Kurata's zoeae may have skipped Stage III of the genus because growing conditions in the laboratory were especially favorable (Mak- arov 1967). Only Stage I zoeae of Placetron wosnessenskii, and Stages I and II zoeae of Rhinolithodes wos- nessenskii have been described (this report). Because these zoeal stages are morphologically typical of lithodid species with four zoeal stages, P. wosnessenskii and R. wosnessenskii like- ly have the four zoeal stages characterized in Table 2. HAYNES: EARLY ZOEAL STAGES OF LITHODID CRABS Brief descriptions and comparisons of previous- ly described lithodid zoeae follow. Cryptolithod.es typicus. — Based on the descrip- tion by Hart (1965), Cryptolithodes typicus zoeae are markedly different morphologically from other described lithodid zoeae. In C. typicus zoeae, the carapace lacks posterolateral spines in all stages, the proximal expansion of the maxilla is present in Stage II, and the telson does not have a medial posterior invagination. In all other lithodid zoeae, the carapace has posterolateral spines in all stages, the proximal expansion of the maxilla is absent until Stage IV, and the telson has a medial posterior invagination. The large eyes of C. typicus, however, are typical of zoeae of the Pagurinae, and the absence of posterolateral spines on the carapace is similar to zoeae of some species of the Diogenidae. The shape of the telson, the fused abdominal somite 6 and telson in Stages III and IV, and the absence of uropods in C. typicus are characters similar to those of some porcellanid zoeae. Hapalogaster grehnitzkii , Dermaturus mandtii, and P. brevipes. — Makarov (1967) briefly de- scribed larvae collected off west Kamchatka that he provisionally identified as Hapalogaster greh- nitzkii, based on distribution of adults. Zoeae of H. grehnitzkii are morphologically similar to zoeae of Dermaturus mandtii and Paralithodes brevipes but can be distinguished by length of the posterolateral spines on abdominal somites 3-5. In zoeae of H. grehnitzkii , posterolateral spines on somites 3 and 4 are short (slightly longer than the denticles that fringe the posterior margin), and the posterolateral spines on somite 5 are shorter than the width of somite 5. In zoeae of D. mandtii and P. brevipes, posterolateral spines on somites 3 and 4 are long (at least twice the length of the denticles), and posterolateral spines on somite 5 are longer than the width of somite 5 (Kurata 1956). Based on Kurata's (1956) brief descriptions, zoeae of P. brevipes can be distinguished from zoeae of D. mandtii by size of the carapace and morphology of the antenna. Paralithodes brevipes zoeae are slightly larger (carapace length, 1.4-1.7 mm) than D. mandtii zoeae (carapace length, 1.2- 1.4 mm). The antennal fiagellum of P. brevipes zoeae is noticeably longer than the antennal scale (including distal spine), and the antennal scale is about nine times as long as wide. The antennal fiagellum and antennal scale of D. mandtii zoeae are about the same length, and the scale is not more than five times as long as wide. Hapalogaster mertensii. — Larvae of Hapalogast- er mertensii were collected from ovigerous fe- males at Fidalgo Island, Wash., and then reared and described by Miller and Coffin (1961). Unfor- tunately, their description is brief and lacks de- tail and, therefore, has limited value. Apparently, the only characters useful for distinguishing H. mertensii zoeae from zoeae of other lithodid spe- cies are size and number of setae on the antennal scale. In zoeal Stages I-III of H. mertensii, the antennal scale has six setae and, in Stage IV, four setae. In all stages of H. mertensii, the setae are markedly short ( ~^, ** «i*- '. i I * — . *T- ^. * . <>i >\ .*#" U> *+' J** ft***: til > l K \& , tf , •»- .r # **'*/.K £ «J V> '4 M 1 '•■ 1 > i* 5* : 3 ;i 327 FISHERY BULLETIN: VOL. 82, NO. 2 period of maximum seasonal immigration for P. aztecus as described by Baxter and Renfro (1967). Accordingly, seven sets of samples were taken be- tween 29 March and 23 July 1982. Each set was obtained over a period of 3 d, and sets were taken biweekly (29 March through 28 May) and monthly (28 May through 23 July). Ordinarily, a set con- tained 12 sample pairs that were subdivided to sample the inner, middle, and outer marsh zones equally, i.e., during each of three sampling days four vegetated-nonvegetated sample pairs were taken from a single zone. Sample sites within zones were chosen randomly each month from subunits in a grid superimposed on a map of the area. The map and aerial photographs were used to estimate percent coverage of vegetated and nonvegetated habitats within different zones. A t-test of paired observations (Steel and Torrie 1960) provided the primary means for evaluating differences in shrimp density between habitats. Other analyses were performed using Pearson product-moment correlations and ANOVAs across sample sets, and Kendall's nonparametric concor- dance tests (Tate and Clelland 1957) within sam- ple sets. Analyses across sets incorporated an ele- ment of temporal variability that was specifically eliminated in analyses within sets. Data were log transformed for ANOVAs to assure homogeneity of variances. Procedures A drop sampler (Fig. 3) was designed to operate in the marsh from the bow of a skiff. The device was an open-ended fiber glass cylinder, reinforced on one end with galvanized metal, that enclosed 2.8 m2 of marsh bottom. The sampler was deployed endwise and pushed at least 15 cm into the sub- strate to insure a good seal against leakage. After marsh grass was removed, water was pumped from the sampler and the enclosed bottom was swept with dip nets to capture the entrapped or- ganisms. The water and the contents of the dip nets were placed into a 1 mm square mesh plankton net with a removable cod end bag. When all sample contents were washed, the cod end bag was detached, labelled, and stored in a container with Formalin4 and Rose Bengal stain. Two identical sampling cylinders were used to obtain sample pairs. Typically, the first sampler was hoisted above the bow of the skiff and quietly maneuvered into position over either vegetated or barren substrate. The device was released and allowed to free fall to the bottom. After disconnect- ing the first sampler, the second sampler was hoisted and the operation repeated in the opposing habitat. The sequence of habitats was reversed from pair to pair so that one would not continually precede the other. Sample pairs were always within two sample diameters of each other (3.6 m) and care was taken to not disturb the site until the second sampler was deployed. Within all samples, the water temperature, •"Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. FIGURE 3— A hand-operated drop sampler used to estimate Penaeus az- tecus densities in a Galveston West Bay salt marsh. 1.83 M DIA FIBER GLASS CYLINDER METAL SKIRT BOOM RELEASE LINE 328 CROSS SECTION OF SAMPLER ZIMMERMAN ET AL.: SELECTION OF HABITAT BY PENAEUS AZTECUS oxygen (YSI oxygen meter, Model 51 B) and maximum and minimum depth were recorded. Water samples (500 ml) were also procured in order to measure turbidity (HF Instruments, Model DRT-15). In vegetated samples, emergent plant material was cut and removed to measure plant biomass and to facilitate capturing the macrofauna. Tide level was recorded from a per- manent station at the beginning and end of each sampling operation. All field work was done dur- ing daylight within about 2 h before and after high tide. In the laboratory, shrimp were identified, sorted, and measured to the nearest millimeter from rostrum tip to end of telson. Shrimp numbers for each millimeter size interval were recorded for each sample. Associated macrofauna from each sample, including fish, crabs, and other shrimp, were identified, measured, and counted. Gut con- tents of the fish were examined for penaeid shrimp as well as other identifiable material. Plant biomass from each sample was dried in sunlight until weight change was negligible. Sediments and epiphytes were allowed to fall away as the material dried. The resulting dry weight was taken using a Mettler K-7 toploading balance and reported as grams above-ground dry plant biomass. Stem density was calculated by weighing a subsample (about 20*7 of the total) and counting the number of culms. Sampler Effectiveness Since the experimental design assumed no sam- pling bias, the method was tested for recovery efficiency both in vegetated and nonvegetated habitats. Fifty shrimp, in the size range of 23 to 91 mm, were marked by clipping a uropod and placed into deployed samplers. After a 30-min adjust- ment period, the usual sampling procedure was followed and recovery was recorded. Since our density data were compared with other surveys, it was useful to test the effective- ness of the drop sampler in relation to other col- lecting devices. These included aim beam trawl, a 5.5 m bag seine, and a 3.7 m otter trawl. During the initial test, eight replicate vegetated- nonvegetated sample pairs were taken using the 1 m beam trawl (3.0 m2) and the drop sampler (2.8 m2). Later, 10 nonvegetated sample replicates were obtained for each of the following: the drop sampler, a 5.5 m bag seine (110 m2), and a 3.7 m otter trawl (75 m2). The data were reported as mean and standard deviation of shrimp density (per m2) for each sampler. The efficiency for each device was calculated relative to the drop sampler. RESULTS A total of 3,277 penaeid shrimp (97% P. aztecus) were collected in 81 paired samples taken between 29 March and 23 July 1982. Shrimp densities in the marsh were significantly higher in S. alter- niflora habitat than adjacent nonvegetated habitat (P < 0.001, t-test, 81 paired observations). The magnitude and integrity of the relationship between shrimp density and habitat type held consistently throughout all sampling dates (Table 1, Fig. 4) and zones within the marsh, except for the outer zone during March and April (Table 2). Comparison of marsh zones (Table 2) revealed highest P. aztecus densities and greater selection for vegetated habitat in the innermost marsh di- minishing toward the outer zone. Shrimp densi- ties in nonvegetated habitat were highest in the outer zone and diminished significantly toward the inner zone (ANOVA, P < 0.001). TABLE 1. — Percent of Penaeus aztecus in vegetated iSpartina alterniflora) and non- vegetated habitats of a Galveston West Bay salt marsh, 29 March through 23 July 1982. Habitat Sampling number Vegetated Nonvegetated period (n) (%n) (%n) 3/29-4/1 355 94.4 5.6 4/13-15 519 81.7 18.3 426-28 802 88.3 11.7 5/11-14 309 90.3 9.7 5 26-28 388 91.8 82 6,22-24 237 97.0 3.0 7/21-23 559 90.2 9.8 S 20 - 5 | 10- o z "r- I H MAR 29 APR 13 APR 26 MAY 11 MAY 26 JUN 22 JUL 21 SHRIMP 35S 519 802 309 388 237 559 SAMPLE 9 12 12 12 12 12 12 FIGURE 4. — Mean densities of Penaeus aztecus compared be- tween vegetated Spartina alterniflora habitat and adjacent non- vegetated habitat. 329 FISHERY BULLETIN: VOL. 82, NO. 2 TABLE 2. — Mean number of Penaeus aztecus per m2 by zone in vegetated and nonvegetated salt marsh habitats from Galveston West Bay 29 March through 23 July 1982. Marsh zone and habitat Sampling Outer Middle Inner Overall period Veg/Non' Veg/Non Veg/Non Veg/Non 3/29-4/1 2 [2.7/1 .9"! [88/5. 5j 12.3/20 16.7/1.1 12.6/2.8 4 26-28 12.3/6 8 28.5/1.3 22 4/0.4 21.1/2.8 5/11-14 7.2 12 9.6/1.3 8.0/0.3 8.3/0.9 5/26-28 12.0/1.5 10.6/0.9 9.2/0.4 10.6/1.0 6/22-24 3.8/0.2 9.7/0.3 7.0/0.2 6.8/0.2 7/21-23 10.9/1.8 13.8/1.9 20.3/1.3 15.0/1.6 Overall 7.5/23 11.0/1.0 16.6/0.6 11.7/1.4 'Veg = Spartina alterniflora habitat; Non = Nonvegetated habitat. difference within brackets not significant between vegetated and nonvegetated pairs; for all others, the difference was highly significant (P < 0.001 . r-test, paired observations). Penaeus aztecus densities for each 20 mm size interval were more abundant in Spartina habitat than adjacent nonvegetated bottom (Fig. 5). Vege- tated habitat contained 89 to 96% of all shrimp in size classes under 50 mm and 75 to 78% of larger size classes (Table 3). Those under 30 mm in length comprised 77,% of all shrimp and those under 60 mm made up 98% of the total (Table 3). Size class distributions differed between habitats (Kol- mogorov-Smirnov test, P = 0.02; Fig. 5), but the very small sample size from nonvegetated habi- tat decreased the strength of this observation. The highest P. aztecus densities in vegetation and the lowest on nonvegetated bottom were characteristic of the innermost zone (Table 1). The degree of vegetated-nonvegetated differences suggested an apparent selection for vegetated habitat and greater selection in the inner zone compared with the outer zone. The increase in vegetated to nonvegetated shrimp densities coin- cided with an increase in S. alterniflora coverage between the outer and inner marsh (Fig. 6). Areal coverage of vegetation, determined from aerial photographs (Fig. 2), differed by a factor of 3 be- tween the outer and inner marsh, and selection, as measured by the ratio of shrimp density in vege- tated habitat to density in nonvegetated habitat, differed by a factor of 9 from outer to inner zones (Fig. 6). In addition, the ratio differed between the middle and inner zone, but shrimp densities within vegetation between those zones (Table 2) did not change significantly (ANOVA, Duncan's multiple range test, 0.05 level). Due to the inter- tidal nature of vegetated habitat, shrimp were forced into subtidal areas at low tide and redis- TABLE 3. — Percent abundance among size classes for Penaeus aztecus in a Galveston West Bay salt marsh. 29 March through 23 July 1982. n = number of shrimp per size interval; N = total number of shrimp collected. Size Shrimp abundance class Overall Spartina (%n) Nonvegetated (mm) n % N Cum. % (%n) <20 1,117 47.7 47.7 89.4 10.6 21-30 683 29.2 76.9 95.6 4 4 31-40 234 10.0 86.9 94.9 5.1 41-50 184 7.8 94.7 886 11.4 51-60 86 3.7 98.4 77.9 22.1 61-70 25 1.1 99.5 76.0 24.0 71-80 8 0.3 99.8 75.0 25.0 81-90 4 0.2 100 75.0 25.0 Total (A/) = 2,341 MAR 29. APR I APR 13-15 APR 2628 MAY 11.14 MAY 26 28 R- 2 lOh JUN 22 24 JUL 21-23 , Cb_ P NONVEGETATED HABITAT n n~^ «=i-R- 1 n-i . n^ SPARTINA VEGETATED HABITAT H- TT„ *=T- fckfi 305°7090 ,0305°7090 ">305°7090 '°305° 9° '°3050709° '° SIZE CLASSES PENAEUS AZTECUS 1 50 90 10 1 50 90 30 70 30 70 FIGURE 5. — Densities of Penaeus aztecus by size class in adjacent vegetated and nonvege- tated habitats from Galveston West Bay during 1982. Size class distributions differed between habitats (Kolmogorov-Smirnov test, P = 0.02). 330 ZIMMERMAN ET AL.: SELECTION OF HABITAT BY PENAEUS AZTECUS 0. 2 en I (/) > \- z LJ o o u I- < I- Ld o LJ > z o z \ Q Ul H < LJ > 28r- 24 20 16 12 8 - 4 - INNER MARSH TABLE 4. — Within habitat densities of Penaeus aztecus from a salt marsh in Galveston West Bay, 29 March through 23 July 1982. n = number of samples. MIDDLE MARSH 10 20 30 PERCENT SPARTINA COVERAGE FIGURE 6. — Selection by Penaeus aztecus for vegetated habitat compared against percent coverage of Spartma alterniflora. tnbuted anew on each subsequent flood tide. Differential predation by fish did not account for shrimp differences between habitats. Of four species preying on shrimp, 328 were in vegetation versus 48 on nonvegetated bottom. Among these, 18 from vegetated ( 5% ) and 3 from nonvegetated (69c ) contained shrimp in gut contents. The pred- ators, in order of vegetated/nonvegetated abun- dance, were Lagodon rhomboides (pinfish 246/36), Fundulus grandis (gulf killifish 45/0), Cynoscion nebulosus (spotted seatrout 22/2), and Paralichthys lethostigma (southern flounder 15/ 10). Only southern flounder contained shrimp in gut contents (3 of 10) from nonvegetated habitat. In vegetated habitat, 8 of 15 southern flounder, 10 of 22 spotted seatrout, 1 of 45 gulf killifish, and 3 of 246 pinfish contained shrimp. Mean density of P. aztecus in vegetation was 11.7/m2 overall with a range of 0.7 to 43.2/m2 (Ta- ble 4). Densities were highest in the innermost marsh (x = 16.6/m2; range = 1.8 to 43.2/m2) and lowest in the outer marsh (x = 7.5/m2; range = 0.7 to 28.2/m2). The overall variance was less than the overall mean. Among marsh zones, shrimp patch- iness in vegetation decreased slightly from the outer to inner marsh (Table 4). Density of P. aztecus in nonvegetated habitat was 1.4/m2 with a range of 0 to 18.2/m2 (Table 4). Densities on nonvegetated bottom were highest in the outer marsh (x = 2.3/m2; range = 0 to 18.2/m2) and lowest in the inner marsh (x = 0.6/m2; range = Marsh habitat Individuals, m2 and zone n X Median 1 SD Coeff. var. (% ) Range With vegetation Outer 27 7.5 6.4 68 90 0.7-28.2 Middle 26 110 11.4 8.9 81 0.4-39.6 Inner 28 16.6 13.8 12.5 75 1 8-43.2 Overall 81 11.7 10.5 9.4 80 07-43.2 Without vegetation Outer 27 2.3 1.4 36 157 0-18.2 Middle 26 1.0 0.7 1.2 120 0- 4.6 Inner 28 0.6 1.0 1.5 56 0- 2.1 Overall 81 1.4 1.1 1.9 136 0-18.2 0 to 2.1/m2). Overall distribution on nonvegetated bottom, as reflected by the variance to mean ratio (coefficient of variation, Table 4), was patchier (more clumped) than on vegetated bottom. Shrimp distributions also were patchier in nonvegetated outer and middle zones, than in the nonvegetated inner zone. Stem density and above-ground biomass of S. alterniflora were positively correlated (Table 5). The overall range of values was 41 to 784 g m2 for biomass and 33 to 629 stems/m2 with respective means of 298 g/m2 (1 SD = 175, n = 81) and 234 stems/m2 ( 1 SD — 72, n =81). Between zones, plant biomass from the outer to inner zone increased from 258 to 348 g/m2. The weight per stem in- creased (larger diameters) from outer to inner marsh. Although the trend suggested a negative relationship between shrimp density and vegeta- tional density and biomass, correlation was not significant over the range examined. Abiotic Relationships Water depth between vegetated and nonvege- tated sample pairs was significantly different (P < 0.01, t-test of 81 paired observations). The mean water depth was 22.1 cm (1 SD = 10.0, n = 81) in TABLE 5. — Density and biomass of Spartma alterniflora from a salt marsh in Galveston West Bay, 29 March through 23 July 1982. n = number of samples. Biomass and density n X 1 SD Coeff. var. (% ) Range Biomass (g/m2) Outer zone 27 258 164 64 41-634 Middle zone 26 289 187 65 41-784 Inner zone 28 348 174 50 69-731 Overall 81 298 175 59 41-784 Density (stems m:) Outer zone 28 234 88 38 37-576 Middle zone 26 231 65 28 33-629 Inner zone 28 236 64 27 47-496 Overall 81 234 72 31 33-629 331 FISHERY BULLETIN: VOL. 82, NO. 2 vegetated samples compared with 41.8 cm (1 SD = 11.8, n = 81) in nonvegetated samples. Changes in tide level were not large (about 30 cm) but were important relative to sample depths. Since sam- pling was executed at high tide, tide station mea- surements were comparable between sampling periods and useful for establishing variability in high-water level. Mean high water during the summer was 12 cm lower than in the spring reflect- ing seasonally variable tidal inundation (Hicks et al. 1983) and greater accessibility to vegetation (Provost 1976) in the spring. A weak negative relationship between shrimp density and temperature within a range of 17.0° to 34.0°C was apparent (r = - 0.34 in vegetation, P < 0.01, n = 57). Since temperature and oxygen levels were inversely related, the trend, attributed to temperature, also extended to an observed rela- tionship between oxygen concentration and shrimp density. However, oxygen levels were al- ways near saturation (vegetated* = 8.2 ppm, 1 SD = 1.4, n = 81; nonvegetated x = 8.1 ppm, 1 SD = 1.4, n = 81) and unlikely to have influenced shrimp distribution. Shrimp densities did not correlate well with salinities (range of 19 to 35 ppt), tur- bidities (range of 3.0 to 55 nephelometer turbidity units), or water depths (overall range of 5.5 to 76 cm). In addition, temperature, salinity, oxygen, and turbidity did not differ between habitats la- test of 81 paired observations for each). Sampler Performance Test results suggested that shrimp recovery from the drop sampler was more variable and somewhat less effective in vegetation (x = 91% recovery, 1 SD = 6.6%, n = 4) than in habitat without vegetation (jc = 97.5% recovery, 1 SD = 2.5%, n = 4). However, a t-test between means by habitat revealed no significant difference (P > 0.1) and justified combining means (94%, 1 SD = 5.8%, n = 8). Mean shrimp densities on nonvegetated bottom, comparing our 1.8 m diameter drop sampler, a 5.5 m wide bag seine, and a 3.7 m wide otter trawl, were 0.285/m2, 0.104/m2, and 0.054/m2, respec- tively. Assuming 97.5% recovery and no avoidance with the drop sampler, conservative estimates of efficiency were 33% for the bag seine and 17% for the otter trawl. Clearly, the data from the drop sampler were more accurate (Table 6). DISCUSSION Habitat Selection Significant differences in habitat-related shrimp densities from a Galveston salt marsh (Ta- ble 2, Fig. 4) demonstrate that P. aztecus may select for S. alterniflora habitat. In support, laboratory data of Giles and Zamora (1973) suggest that P. aztecus and P. setiferus each prefer S. alterniflora as opposed to barren substrate. In addition, marsh grass transplanted on a dredge spoil in Galveston Bay increased shrimp numbers (Trent et al. 1969) and elimination of marsh habitat to create waterfront housing diminished shrimp abundance (Mock 1966; Gilmore and Trent 1974; Trent et al. 1976). In other instances, P. az- tecus has been associated with vegetation includ- ing Ruppia and Vallisneria in Mobile Bay (Loesch 1965), seagrasses in the Laguna Madre (Stokes 1974), and Juncus, Spartina, and seagrasses in Mississippi Sound (Christmas et al. 1976). The latter reported movement of postlarvae into marsh vegetation during tidal inundation. The determinants of selection may have less to do with S. alterniflora per se than with other characteristics of vegetated habitat. For example, in our case, shrimp numbers were not related to the density or biomass of marsh grass (Table 5) but simply to its presence or absence. Also, attraction to vegetation differed between outer and inner marsh (Table 2). Other studies have shown that TABLE 6. — Comparative gear efficiencies for sampling Penaeus aztecus in a Galves- ton West Bay salt marsh. Area sampled and number of replicates for each device are as follows: Drop sampler 2.8 m2 (n = 22); beam trawl 3.0 m2 (n = 12); bag seine 109 m2 (/! = 10); otter trawl 72 m2 (n = 10). x Efficiency Drop Beam Bag Otter Habitat type sampler trawl seine trawl Spartina vegetation 94% 23% not not (Shrimp count, x/m2 ±SD) (8.9±3.7) (2.2±2.2) operable operable Nonvegetated 98% 82% 33% 17% (Shrimp count, x/m2 ±SD) (0 30-0.3) (0.25 ±0.46) (0.10±0.06) (0.05 ±0.04) 332 ZIMMERMAN ET AL.: SELECTION OF HABITAT BY PENAEUS AZTECl 'S the presence of estuarine macrophytes can be as- sociated with an increase in epifaunal abundance (Heck and Wetstone 1977; Heck and Orth 1980) as well as providing protective cover (Vince et al. 1976; Nelson 1979; Coen et al. 1981; Heck and Thoman 1981). For shrimp selecting vegetated marsh, this may translate into a greater variety and abundance of food and some degree of protec- tion from predation. Zonal and Areal Relationships Penaeus aztecus demonstrated a greater degree of attraction to vegetated habitat in the inner than the outer marsh. Accordingly, shrimp densities were higher among vegetation and lower on non- vegetated bottom in the innermost zone compared with the outer zone. This relationship is adequately reflected by comparing ratios of vege- tated with nonvegetated shrimp density. Using the ratios, the change in selection from the outer, mid- dle, to inner zone was 3.3:1, 11.0:1, and 27.7:1, re- spectively. The percent area covered by S. alter- ni flora (Fig. 2) also increased (by a factor of three) from outer to inner marsh, but as vegetational coverage increased arithmetically selection by P. aztecus increased geometrically (Fig. 6). This im- plies that salt marshes with more vegetational coverage have disproportionately greater attrac- tive value to P. aztecus than do those with less coverage. On a larger scale, Turner ( 1977 ) revealed a positive correlation between extensiveness of es- tuarine vegetation and offshore shrimp yield. However, the relationship may not be simple; it is likely to depend upon characteristics such as the configuration, accessibility, and quality of vegeta- tional patches within a marsh. For instance, an edge effect has been identified which associates large numbers of shrimp with the nonvegetated zone adjacent to vegetation (Mock 1966; Christmas et al. 1976). Since our Spartina habitat was inter- tidal, and often not inundated during low tides, the nonvegetated subtidal habitat provided a ref- uge against stranding. We have assumed that it did and that shrimp redistributed accordingly each tidal cycle. It is evident that an increase in the amount of ecotone edge (between habitats) would facilitate movement for the shrimp popula- tion. It is also evident that the amount of edge is proportionally related to the degree of retic- ulation in the marsh (Fig. 2). Thus, reticulation may be an important mechanism for increasing the accessibility of intertidal vegetation to P. aztecus. Shrimp Densities Density estimates for penaeid shrimp in S. al- terniflora vegetation have not been reported pre- viously. We found a density range for P. aztecus in Spartina habitat of 0.7 to 43.2/m2 with an overall mean, from March through July, of 11.7/m2 (1 SD = 9.4, n = 81). Comparable densities from adjacent nonvegetated habitat ranged between 0 and 18.2/m2. All densities were taken when P. aztecus numerically dominated the shrimp population. By August, when P. setiferus first began to dominate, the combined mean for both species in vegetation increased to 50.8/m2 (1 SD = 31.6, n = 12) and a single sample attained a density of 118.6 shrimp/ m2. These data may indicate a potential for higher P. aztecus densities earlier in the season and suggest that P. aztecus were not restricted by lack of space. To our knowledge, we have provided the first accurate estimates of shrimp density in marsh vegetation, and our densities are among the few available for any estuarine system. Due to method limitations, most researchers have only reported relative abundances of restricted sizes, usually over nonvegetated bottom. The single exception was data by Allen and Hudson (1970), using a suction sampler in seagrasses in Florida Bay. From 43 trials, they reported a mean of 6.2/m2 ± 3.4 SD for P. duorarum. Estimates off! aztecus densities from nonvege- tated bottom in three other Galveston Bay salt marshes were available from the Texas Parks and Wildlife Department (TPWD) from 1976 through 1981 (Benefield 1982, footnote 5.). The data were taken using a marsh net (Renfro 1963) which was relatively effective for capturing shrimp on non- vegetated bottom (Table 6 compares a beam trawl, similar to the marsh net, with other sampling devices). Mean TPWD densities for P. aztecus dur- ing the latter half of March were 10.4/m2 for 1976, 5.2/m2 for 1977, 0.3/m2 for 1978, 1.3/m2 for 1979, 8.7/m2 for 1980, and 5.1/m2 for 1981 with an overall mean of 5.2/m2. In our study, on nonvegetated bot- tom, the March mean for P. aztecus was 0.9/m2 and overall (March through July) the mean was 1.4/m2. It is evident that our nonvegetated densities for P. aztecus were within the range, but low compared with the mean calculated from TPWD data. These densities of P. aztecus may not be strictly 5R. L. Benefield, Bay Shrimp Project Leader, Texas Parks and Wildlife Department, Coastal Fisheries Branch, RO. Box 8, Seabrook, TX 77586, pers. commun. September 1982. 333 FISHERY BULLETIN: VOL. 82, NO. 2 comparable, since sampling was executed during unknown variable tidal stages and the degree of flooding in intertidal vegetation appears to greatly influence shrimp densities on nearby non- vegetated subtidal bottom. Perhaps the only meaningful density estimates are those taken dur- ing low tide in nonvegetated habitat or those taken in vegetated habitat at flood tide. In any case, tide stage must be uniform for data to be comparable. Sampling Integrity The sampling approach in our investigation provided more realistic density estimates than traditional methods for sampling shrimp in es- tuaries (Table 6). We agree with Loesch et al. (1976) in concluding that techniques such as the area-swept method using an otter trawl are among the poorest for quantifying P. aztecus. Past recog- nition of this problem stimulated development of the push net (Allen and Inglis 1958), small beam trawl (Renfro 1963; Loesch 1965), and marsh net (Pullen et al. 1968). These samplers improved ac- curacy on nonvegetated bottom, but were ineffec- tive when vegetation was present and did not solve avoidance problems. Further improvement came for sampling in seagrasses, but not salt marshes, with the invention of a sled-mounted suction sam- pler (Allen and Hudson 1970) and modification of a drop net technique (Hoese and Jones 1963; Gil- more et al. 1976). Our methodology has been de- signed to minimize escape, improve recovery from the area sampled (including burrowed shrimp), and to operate in salt marsh habitats. The drop- sampler method proved to be nearly as effective among vegetation as on nonvegetated bottom. CONCLUSION We contend that differences in P. aztecus den- sities between vegetated and nonvegetated marsh bottom were due to habitat selection. In support, we refer to Loesch (1965), Trent et al. (1969), and Stokes (1974) who have associated brown shrimp distributions with estuarine vegetation, and a laboratory experiment by Giles and Zamora ( 1973 ) demonstrating P. aztecus prefer S. alterniflora in- stead of barren substrate. Finally, our fish gut examinations indicate that immediate effects of predation did not account for the density differ- ential. Since S. alterniflora is characteristically inter- tidal, and not continuously available to shrimp, the adjacent subtidal zone provided an important alternate habitat during low tide. We propose that the amount of edge between habitats facilitated shrimp movement, and the reticulated nature of the salt marsh was an important feature for in- creasing the amount of edge. In addition, intertid- al vegetation was more accessible and its potential for utilization greater during spring and fall high tides. This interaction may in part account for seasonal peaks in P. aztecus populations. In our investigation, recruitment began abruptly with equinox tides. The shrimp population during the spring and early summer was dominated entirely by P. aztecus. Our shrimp densities from vegetated habitat were higher than any previously reported includ- ing those from seagrass and mangrove systems. The high densities in vegetation were possibly governed by the amount of total marsh, ratio of vegetated to nonvegetated habitat, and size of re- cruitment. The densities on nonvegetated marsh bottom were probably controlled by the relative accessibility of nearby vegetated habitat. In any case, the observed density differential strongly implies that marsh vegetation provides a vital function for juvenile brown shrimp. ACKNOWLEDGMENTS Edward F. Klima and the staff of the Southeast Fisheries Center Galveston Laboratory, National Marine Fisheries Service, are especially acknowl- edged for their support of this investigation. In particular, C. Albrecht, S. Dent, D. Gleason, K. Griffith, E. Martinez, and J. Wellborn are recog- nized for field and laboratory assistance. Equip- ment and important logistical support were kindly provided by K. Baxter and Z. Zein-Eldin. The manuscript was reviewed by E. Klima, S. Ray, G. Thayer, and Z. Zein-Eldin, and final prepara- tion was assisted by D. Patlan, J. Doherty, and B. Richardson. LITERATURE CITED ALDRICH, D. V, C. E. WOOD, AND K. N. BAXTER. 1968. An ecological interpretation of low temperature re- sponses in Penaeus aztecus and P. setiferus postlar- vae. Bull. Mar. Sci. 18:61-71. ALLEN, D. M., AND J. H. HUDSON. 1970. A sled-mounted suction sampler for benthic or- ganisms. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 614, 5 p. ALLEN, D. M., AND A. INGLIS. 1958. A pushnet for quantitative sampling of shrimp in shallow estuaries. Limnol. Oceanogr. 3:239-241. 334 ZIMMERMAN ET AL.: SELECTION OK HABITAT BY PENAEUS AZTECUS ANDERSON, W. W, J. E. KING, AND M. J. LINDNER. 1949. Early stages in the life history of the common marine shrimp, Penaeus setiferus ( Linnaeus >. Biol . Bull. ( Woods Hole) 96:168-172. ANDREWARTHA, H. G. 1971. Introduction to the study of animal populations. Univ. Chicago Press, Chicago, 283 p. Barrett, B. B.. and M. C. Gillespie. 1973. Primary factors which influence commercial shrimp production in coastal Louisiana. La. Wild Life Fish Comm.Tech. Bull. 9, 28 p. Baxter, K. N., and W. C. Renfro. 1967. Seasonal occurrence and size distribution of postlar- val brown and white shrimp near Galveston, Texas, with notes on species identification. U.S. Fish Wildl. Serv., Fish. Bull. 66:149-158. BENEFIELD. R. L. 1982. Studies of shrimp populations in selected coastal bays of Texas. Part I. Investigations of brown shrimp i Penaeus aztecus) populations in Texas Bays, 1979- L980. Tex. Parks Wild]. Coast. Fish. Branch, Manage. Data Ser. 41, p. 1-20. BROWDER. J. A., AND D. MOORE. 1981. A new approach to determining the quantitative re- lationship between fishery production and flow of fresh- water to estuaries. U.S. Fish Wildl. Serv. Rep. 81/04, Vol. 1, p. 40.3-431. CHRISTMAS. J. Y , W LANGLEY, AND T VAN DEVENDER. 1976 Investigations of commercially important penaeid shrimp in Mississippi. Gulf Coast Res. Lab.. Ocean Springs, Miss., 66 p. COEN, L. D.. K. L. HECK. JR.. AND L. G. ABELE. 1981. Experiments of competition and predation among shrimps of seagrass meadows. Ecology 62:1484-1493. DE LA CRUZ, A A 1965. A study of particulate organic detritus in a Georgia salt marsh-estuarine ecosystem. Ph.D. Thesis. Univ. ( leorgia, Athens. 141 p. FALLER. K. H. 1979. Shoreline as a controlling factor in commercial shrimp production. NASA Tech. Memo. 72-732, 33 p. Giles. J. H„ and G. zamora 1973. Cover as a factor in habitat selection by juvenile brown {Penaeus aztecus) and white iP. setiferus) shrimp. Trans Am. Fish. Soc. 102:144-145. gilmore, r. g., g. r. Kulczychi. p a. Hastings, and w C. magley. 1976. Studies of fishes of the Indian River Lagoon and vicinity. //; D K. Young (editor), Indian River coastal zone study, third annual report, p. 133-147. Harbor Branch Consortium, Fort Pierce, Fla. GILMORE, G., AND L. TRENT. 1974. Abundance of benthic macroinvertebrates in natural and altered estuarme areas. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-677. 13 p. GUNTER, G. 1961. Habitat of juvenile shrimp (Family Penaei- dae). Ecology 42:598-600. HAINES, E. B. 1977. The origins of detritus in Georgia salt marsh es- tuaries. Oikos 29:254-260. HECK, K. L., JR., AND R. J. ORTH. 1980. Seagrass habitats: The roles of habitat complexity, competition and predation in structuring associated fish and motile macroinvertebrate assemblages. //; V. S. Kennedy (editor), Estuarine perspectives, p. 449- 464. Acad. Press, Inc., N.Y HECK, K. L., JR., AND T. A. THOMAN. 1981. Experiments on predator-prey interactions in vege- tated aquatic habitats. J. Exp. Mar. Biol. Ecol. 53:125- 134. HECK, K. L., JR., AND G. S. WETSTONE. 1977. Habitat complexity and invertebrate species rich- ness and abundance in tropical seagrass meadows. J Biogeogr. 4:135-142. HICKS, S. D., H. A. DEBAUGH, JR., AND L. E. HICKMAN, JR. 1983. Sea level variations for the United States 1855- 1980. NOS NOAA, Tides and Water Levels Branch, Rockville, Md., 170 p. HILDEBRAND, H. H., AND G. GUNTER. 1952. Correlation of rainfall with the Texas catch of white shrimp Penaeus setiferus i Linnaeus i. Trans. Am. Fish. Soc. 82:151-155. HOESE, H. D., AND R. S. JONES. 1963. Seasonality of larger animals in a Texas turtle grass community. Publ. Inst. Mar. Sci. Univ. Tex. 9:37-46. JONES. R. R, JR. 1973. Utilization of Louisiana estuarine sediments as a source of nutrition for the brown shrimp, Penaeus az- tecus. Ph.D. Thesis, Louisiana State Univ., Baton Rouge. 131 p. KUTKUHN, J. H. 1966. Dynamics of a penaeid shrimp population and man- agement implications. U.S. Fish Wildl. Serv.. Fish. Bull. 65:313-338. LOESCH, H. 1965. Distribution and growth of penaeid shrimp in Mobile Bay, Alabama. Publ. Inst. Mar. Sci., Univ. Tex. 10:41- 58. LOESCH, H., J BISHOP. A. CROWE, R. KUCKYR. AND P. WAGNER. 1976. Technique for estimating trawl efficiency in catching brown shrimp {Penaeus aztecus*. Atlantic croaker {Mi- cropogon undulatus) and spot {Leiostomus xanth- urus). Gulf Res. Rep. 5(2):29-33. MOCK. C.R. 1966. Natural and altered estuarine habitats of penaeid shrimp. Proc. Gulf Canbb. Fish. Inst.. 19th Annu. Sess., p. 86-98. NELSON, W. G. 1979. Experimental studies of selective predation on am- phipods: consequences for amphipod distribution and abundance. J. Exp. Mar. Biol. Ecol. 38:225-245. PROVOST, M. W. 1976. Tidal datum planes circumscribing salt marshes. Bull. Mar. Sci. 26:558-563. PULLEN, E. J., C. R. MOCK, AND R. D. RlNGO. 1968. A net for sampling the intertidal zone of an es- tuary. Limnol. Oceanogr. 13:200-202. PULLEN, E. J.. AND W L. TRENT. 1969. White shrimp emigration in relation to size, sex. temperature and salinity. FAO Fish. Rep. 57. p. 1001- 1014. RENFRO. W C. 1963. Small beam net for sampling postlarval shrimp. U.S. Fish Wildl. Serv. Circ. 161. p. 86-87 ST AMANT, L. S., J. G. BROOM, AND T. B. FORD. 1966. Studies of the brown shrimp Penaeus aztecus. in Barataria Bay, Louisiana, 1962-1965. Proc. Gulf Caribb. Fish. Inst., 18th Annu. Sess., p. 1-17. STEEL, R. G. D . AND J. H. TORRIE. 1960. Principles and procedures of statistics with special 335 FISHERY BULLETIN: VOL. 82, NO. 2 reference to the biological sciences. McGraw-Hill, N.Y., 481 p. STOKES, G. M. 1974. The distribution and abundance of penaeid shrimp in the lower Laguna Madre of Texas, with a description of the live bait shrimp fishery. Tex. Parks Wildl. Dep., Ser. 15, 32 p. Tate, M. w, and R. C. Clelland. 1957. Nonparametric and shortcut statistics in the social, biological, and medical sciences. Interstate Printers and Publishers, Inc., Danville, 111., 171 p. TEAL, J. M. 1962. Energy flow in the salt marsh ecosystem of Geor- gia. Ecology 43:614-624. THAYER, G. W, H. H. STUART, W. J. KENWORTHY, J. F. USTACH, AND A. B. HALL. 1978. Habitat values of salt marshes, mangroves and sea- grasses for aquatic organisms. In P. E. Greeson, J. R. Clark, and J. E. Clark (editors), Wetland functions and values: The state of our understanding, p. 235-247. Proc. Natl. Symp. Wetlands, Am. Water Res. Assoc, Min- neapolis, Minn. TRENT, L., E. J. PULLEN, AND R. PROCTOR. 1976. Abundance of macrocrustaceans in a natural marsh and a marsh altered by dredging, bulkheading, and fill- ing. Fish. Bull.. U.S. 74:195-200. TRENT, W L., E. J. PULLEN, C. R. MOCK, AND D. MOORE. 1969. Ecology of western Gulf estuaries. In Report of the Bureau of Commercial Fisheries Biological Laboratory, Galveston, Texas, fiscal year 1968, p. 18-24. U.S. Fish Wildl. Serv.,Circ. 325. TURNER, R. E. 1977. Intertidal vegetation and commercial yields of penaeid shrimp. Trans. Am. Fish. Soc. 106:411-416. VINCE, S., I. VALIELA, N. BACKUS, AND J. M. TEAL. 1976. Predation by the salt marsh killifish Fundulus heteroclitus (L.) in relation to prey size and habitat struc- ture: consequences for prey distribution and abun- dance. J. Exp. Mar. Biol. Ecol. 23:255-266. WEINSTEIN, M. P. 1979. Shallow marsh habitats as primary nurseries for fishes and shellfish. Cape Fear River, North Caroli- na. Fish. Bull., U.S. 77:339-357. ZEIN-ELDIN, Z. R, AND G W GRIFFITH. 1966. The effect of temperature upon the growth of laboratory-held postlarval Penaeus aztecus. Biol. Bull. (Woods Hole) 131:186-196. 336 REPRODUCTION, MOVEMENTS, AND POPULATION DYNAMICS OF THE BANDED DRUM, LAR1MUS FASCIATUS, IN THE GULF OF MEXICO1 2 Gary W. Standard3 and Mark E. Chittenden, Jr.4 ABSTRACT Collections of banded drum, Larimus fasciatus, were made from 5 to 100 m in the Gulf of Mexico along a cross-shelf transect off Texas during the period October 1977- August 1981. Larimus fasciatus mature at 80-130 mm as they approach age I. Spawning occurs during two discrete periods, a major fall period (September- November) and a very minor spring period (April-June), coincident with downcoast along- shore currents (toward Mexico) and onshore surface Ekman transport. Fish first spawning at 12-14 months old produce the predominant fall-spawned groups. Fall-spawned fish spawning at 19-21 months old produce the minor spring-spawned groups, so that temporal reproductive isolation does not exist between spring and fall cohorts. Larimus fasciatus in the northwestern Gulf range from <5 to 55 m but are most abundant at 5-16 m. Adults occupy the 13-24 m bathymetric range, while the young recruit in waters of <5-16 m when 2-4 months old. Larger, older, spawning or postspawning individuals may undergo more or less permanent emigration from the northwestern Gulf to the north central area as they approach age I. Apparent mean sizes of fall-spawned fish were 130-150 mm at age I and 155-180 mm at or approaching age II. Von Bertalanffy parameters for fall-spawned fish were 201 and 176 mm for Lx and 1.15 and 1.34 for K (annual), respectively. Maximum size is about 180 mm in the northwestern Gulf, but more typically only 160-165 mm. Typical maximum life span Ul ) in the northwestern Gulf is only 1-2 years but may be 2-3 years if the the stock ranges in both the northwestern and north central Gulf. Apparent mean time-specific and cohort-specific total annual mortality rates are 92-100% in the northwestern Gulf but true values probably are 80-90% for a stock that ranges in both the northwest- ern and north central Gulf. Fecundity, weight, girth, and length relationships are presented. The banded drum, Larimus fasciatus, is a common demersal fish that ranges along the Altantic coast of the United States from Chesapeake Bay to southern Florida and in the Gulf of Mexico (Gulf) from the west coast of Florida to Campeche Bay (Hildebrand and Schroeder 1928; Hildebrand 1954; Briggs 1958). It primarily occurs in near- shore marine waters (Hildebrand and Cable 1934; Powles 1980) and only occasionally enters the lower reaches of estuaries (Gunter 1938; Swingle 1971; Dahlberg 1972). In the northern Gulf this species is most abundant off Louisiana (Gunter 1945; Behre 1950; Hildebrand 1954). The life history of L. fasciatus is poorly known despite its common occurrence. No detailed study describes its life history in the Gulf, although Ross 'Based on a thesis submitted by the senior author in partial fulfillment for the M.S. degree, Texas A&M University. ^Technical Article TA 18596 from the Texas Agricultural Ex- perimental Station. College Station, Tex. :iDepartment of Wildlife and Fisheries Sciences. Texas A&M University, College Station. Tex.; present address: Texas Parks and Wildlife Department, Route 2, Box 537, Brownsville, TX 78520. 4Department of Wildlife and Fisheries Sciences, Texas A&M University. College Station, TX 77843; present address; Virginia Institute of Marine Science, Gloucester Point, VA 23062. Manuscript accepted October 1983. FISHERY BULLETIN: VOL. 82, NO. 2, 1984. ( 1978 ) did so for North Carolina. Life history notes appear in Hildebrand and Cable ( 1934), Miller and Jorgenson (1969), Christmas and Waller (1973), Chao and Musick (1977), Johnson (1978), and Powles (1980). This paper describes maturation, spawning periodicity, bathymetric distribution, recruit- ment, movements, age determination and growth using length frequencies, maximum size, life span, mortality, sex ratios, fecundity, and length- weight, length-girth, and standard-total length relationships of L. fasciatus in the northwestern Gulf. METHODS Larimus fasciatus were collected in 71 monthly or twice monthly cruises from October 1977 through August 1981 along a transect in the Gulf off Freeport, Tex., (Fig. 1) aboard a chartered shrimp trawler using twin 10.4 m (34-ft) trawls with a 4.4 cm stretched mesh cod end and a tickler chain. Initial stations usually were located at depths of 9, 13, 16, 18, 22, 27, 36, and 47 m. Sam- pling was expanded to include stations at 5 and 24 337 FISHERY BULLETIN: VOL. 82, NO. 2 94 !0 r28 00 ?8 10 94° 30 FIGURE 1. — Location of sampling area off Freeport, Tex. Station depths and bathymetric contours are indicated in meters. Starred area in insert indicates where collections were made in the north central Gulf. 338 STANDARD and CHITTENDEN: REPRODUCTION. MOVEMENT. AND POPULATION OF BANDED DRUM m after November 1978 and at 55, 64, 73, 82, 86, and 100 m after May 1979. Collections were made during the day through September 1978; there- after, a day and a night cruise usually were made each month. Two 10-min tows (bottom time) were made at each depth except that 1 tow was made at most depths prior to October 1978, usually 8 tows were made at 16 m, and usually 24 tows were made at 22 m. All L. fasciatus were culled from the catch, mea- sured to the nearest millimeter total length, fixed in 10'<^ Formalin5, and later preserved in 70' , ethanol. For the period October 1979- April 1981, if available, 300 fish each month were selected for intensive processing using stratified random sam- pling in which a stratum included an individual spawned group (Standard 1983: app. 1). The follow- ing data were taken on the first 200 fish selected: total length (TL), standard length (SL), girth at origin of dorsal fin (G), total weight (TW>, gonad weight (GW), sex, and ovary maturity stage. Only sex and ovary maturity stage were recorded for the remaining 100 fish. Maturity stages (Table 1) were assigned to immature and female fish using a slight modification of Kesteven's system (Bagenal and Braum 1971). Gonadosomatic indices (GSI) were calculated for individual females as GSI = 100 GW TW. Supplemental collections were made in the north central Gulf from 24 October to 5 November 1982 aboard the FRS Oregon 11 (NMFS) using standard 12.2 m (40-ft) 4-seam semiballoon shrimp trawls at depths of 9-91 m between long. 88°00' and 89°00'W and at depths of 347-549 m between long. 87°50' and 88°00'W (Rohr et al.6). Total length was measured on all L. fasciatus cap- r'Reference to trade names does not imply endorsement by the National Marine Fisheries Service. NOAA. TABLE 1. — Description of gonad maturity stages assigned to Larimus fasciatus. Stage and name Description 1 Immature 2 Maturing Virgin 3 Early Developing 4 Late Developing 5 Gravid 6 Ripe 7 Spawning Spent 8 Resting Gonads barely or not visible, sexes indistin- guishable to the naked eye. Ovaries small, thin, confined to postenodorsal wall of body cavity. Ovaries solid, opaque, occupy 30% of body cavity Individual eggs not visible to naked eye. Ovaries occupy -30% of body cavity. Individ- ual eggs opaque, distinguishable to naked eye by close examination. Ovaries occupy -50°o of body cavity. Individ- ual eggs distinct, 50% translucent. Ovaries completely fill body cavity, -50% of the eggs translucent. Ovaries flaccid, remaining eggs translucent. Ovaries firm, occupy <30% of body cavity, translucent eggs may persist. Fish large enough to have spawned. tured to compare with size compositions from the northwestern Gulf. Age in years was determined by length- frequency analysis, e.g., the Petersen Method (Lagler 1956). Spawned groups (intrayear class cohorts) were specified by the season and year when they hatched, e.g., fall 1980. Descriptions of spawning periodicity (beginnings and ends) using length frequencies assume the following size and age combinations predicted from quadratic re- gression of total length on age, years pooled, noted below: 15 mm TL at 1 mo, 30 mm at 2 mo, and 45 mm at 3 mo. The same combinations were predicted from regressions for individual fall- spawned groups. Duration of the spawning period was approxi- mated for fall-spawned groups following Geoghe- gan and Chittenden (1982) as Time-specific mean size range early in life Mean growth/day early in life Calculations were based on April-June data, the first months when fall groups appeared fully re- cruited. Time-specific size range was estimated for each fall group as the mean of the 99c/c confidence intervals for observations in April- June (Table 2). Growth per day was estimated as the mean of the growth per day values between successive collec- tions in the April-June period (Table 2). This pro- cedure assumes large fish hatch before small ones and that all grow at the same rate ( Geoghegan and Chittenden 1982). The latter assumption appears valid because 999? confidence intervals for obser- vations (Table 3) were fairly constant within cruises in the April-June period when sample sizes were large. Hatching dates used to set time scales to calcu- late growth of fall-spawned fish were determined by a one-step iteration process. A hatching date of 1 October was assigned to start the process be- cause D fish 20-40 mm TL, which we assumed were 1-3 mo old, first appear in November-December, and 2) slopes for regressions of ovary weight on total length (Fig. 2) and mean GSI (Fig. 3) were greatest in September-October. Quadratic regres- sions of total length on age in days were then used as a simple model to estimate initial .x- intercepts for each fall-spawned group. Final hatching dates 6Rohr. B. A.. A. J.. Kemmerer, and W. H. Fox. Jr. 1983. FRS Oregon II Cruise 130. 10-12-11 24 82. Cruise Rep.. 22 p. Southeast Fisheries Center Pascagoula Facility, National Marine Fisheries Service, NOAA. P.O. Drawer 1207, Pascagoula, MS 39567-0112. 339 FISHERY BULLETIN: VOL. 82, NO. 2 TABLE 2. — Calculations (see Methods ) to estimate duration of fall spawning periods ofLarimus fasciatus during 1977, 1978, 1979, and 1980. Collection dates, sample sizes, means, and 999S confidence limits for observations were obtained from Table 3. 99% Growth Mean confidence increment Group and total limits of between first Time Growth/Day collection date n length observations and last date (days) (mm) 1977 fall-spawned group 14 Apr. 1978 5 69.0 33.9-104 0 8 May 1978 37 87.5 63.9-111 1 Means 58.65 (interval) Spawning period = 58.65/0.77 = 76 d 1978 fall-spawned group 5 Apr. 1979 197 67.6 39.3-95.9 20 Apr 1979 328 78.7 50.6-106.8 14 May 1979 490 94.8 68 0-121 6 6 June 1979 65 115.3 853-145.3 21 June 1979 151 122.4 104.0-1408 Means 52.64 (interval) Spawning period = 52.64/0.69 = 76 d 1979 fall-spawned group 1 Apr 1980 746 72.6 49.3-95.9 16 Apr 1980 437 79.6 53.1-106.0 5 May 1980 1.202 79.7 55.2-104.2 19 May 1980 239 888 59.8-117.8 2 June 1980 614 103 8 81 3-126.3 19 June 1980 254 115.9 90.2-141 6 Means 50.48 (interval) Spawning period = 50.48/0.58 = 1980 fall-spawned group 7 Apr 1981 110 600 33.1-869 20 Apr 1981 69 74.2 457-1027 4 May 1981 97 80.4 57.0-103 8 19 May 1981 165 89.7 68 8-110.6 2 June 1981 128 102.5 77.9-127.1 15 June 1981 174 111.8 90.4-133.2 Means 48.57 (interval) 18.5 24 0.77 18.5 24 077 11 1 15 0.74 16.1 24 067 20.5 23 0.89 7.1 15 0 47 13.7 8.7 19 16 069 7.0 15 0.47 0.1 19 0.01 9.1 14 0 65 15.0 14 1 07 12.1 17 0.71 0.58 142 13 1.09 6.2 14 0.44 9.3 15 0 62 12.8 14 0.91 9.3 13 0 72 104 14 0 76 .76 = 64 d TABLE 3. — Growth data (mm TL) by spawned group for Larimus fasciatus from the Gulf off Freeport, Tex. Night and day cruises are indicated by night (N> and day (D). Observed size ranges delineate spawned group boundaries used in growth and mortality calculations and indicated in Figure 6. Collection dates with asterisks indicate collections pooled to esti- mate mean size at age. 95% 99% Unadjusted Observed Mean confidence confidence growth Group and size range length limits of limits of increment collection date n (mm) (mm) s2 the mean observations (mm) 1976 fall-spawned group 1 Oct. 1977 D- 9 3 Dec 1977 D 2 1977 fall-spawned group 21 Mar. 1978 D 10 14 Apr 1978 D 5 8 May 1978 D 37 15 July 1978 D 286 15 Sept. 1978 D" 8 1 Dec. 1978 N 14 13 Dec 1978 D 1 6 June 1979 N 1 21 June 1979 D 1 1978 fall-spawned group 1 Dec. 1978 N 98 13 Dec 1978 D 51 24 Feb. 1979 D 19 5 Apr 1979 N 197 20 Apr 1979 D 328 14 May 1979 N 490 6 June 1979 N 65 21 June 1979 D 151 5 July 1979 N 3 19 July 1979 D 83 22 Aug 1979 D 83 22 Sept 1979 D" 23 2 Oct 1979 N' 59 16 Oct. 1979 D' 198 3 Nov. 1979 N 9 15 Nov 1979 D 7 120-143 136-141 47-97 62-78 70-104 113-141 145-152 141-162 155 156 174 34-99 34-98 40-97 45-98 51-115 68-134 78-130 95-140 115-131 111-138 111-155 132-164 133-156 127-164 138-158 144-155 133.6 1385 623 69.0 87.5 127.5 148 0 1492 155.0 156.0 174.0 64.3 738 663 67.6 78.7 948 115.3 122.4 121.0 123.5 139.1 144 8 1456 144.1 146.1 1489 465 12.5 261.3 58.0 75.0 26.5 86 33.3 220.7 266.6 343 8 120.3 118.7 108.0 128.1 51.3 76.0 42.7 54.5 388 298 34.6 43.4 13.8 1284-1388 106.7-170 3 507-73.9 59 5-78.5 84.6-90.4 126.9-128.1 145.5-150.5 145.9-152.5 61.3- 69.2- 57.4- 66.1- 77.5- 93.9 112.5 121.3- 99.3- 122 1 137.5- 142.1- 1442- 143.3- 141.0- 145.5- 67.3 78.4 75.2 69.1 79.9 95.7 118.1 1235 142.7 124.9 140.7 147.5 147.0 144.9 151.2 152.3 110.7-1565 -86 6-3636 98-1148 33.9-104 0 63.9-111.1 114.2-140.8 137.7-158.3 131 8-166.6 25.3 30.1 12.9 39.3 50.6- 68 0- 85.3- 104.0- 34.5- 106.3- 119 6- 127.2- 131.1- 128 9- 124 0- 135.1- 103.3 117.5 1197 95.9 1068 121.6 145.3 140.8 207.5 140.7 158.6 162.4 160.1 159 3 1682 162.7 + 49 + 6.7 + 18.5 +40.0 + 20 5 + 1.2 + 5.8 + 10 + 18.0 + 9.5 - 7.5 + 1.3 + 11.1 + 16.1 + 20.5 + 7.1 - 1.4 + 2.5 + 15.6 + 5.7 + 0.8 - 1.5 + 2.0 + 2.8 + 2.8 340 STANDARD and CHITTENDEN: REPRODUCTION, MOVEMENT, AND POPULATION OF BANDED DRUM TABLE 3.— Continued Observed Mean 95% confidence 99% confidence Unad|usted growth Group and size range length limits of limits of increment collection date n (mm) (mm) S2 the mean observations (mm) 1 Dec 1979 N 3 148-155 151 7 123 143.0-1604 1169-186.5 +-11.3 - 60 -11 0 14 Dec 1979 D 1 163 163.0 — — — 3 Jan. 1980 4 Feb 1980 N N 1 1 157 146 157.0 1460 — — — 5 Mar 1980 N 1 150 150.0 — — — - 4 0 + 1.8 - 1.3 -21.0 -130 - 1.3 - 2.0 19 Mar 1980 D 45 135-163 151.8 298 150.2-153.4 137.1-166.5 16 Apr 1980 D 8 132-166 1505 91 4 142.5-158 5 117.0-1840 19 May 1980 2 June 1980 D N 2 22 161-182 146-174 171.5 1585 220 5 374 38 1-304 9 155 8-161 2 -773.8-1.116.8 141 2-175 8 19 June 1980 D 26 147-180 157.2 50.8 154.3-160.1 137.3-177 1 7 July 1980 N 45 150-172 1592 26 1 157.7-160.7 1454-1730 5 Aug 1980 N" 11 156-175 1633 31.6 159.5-167 1 145 5-181.1 + 4.1 7 Sept. 1980 N' 1 177 177.0 — — — -137 1979 fall-spawned group 3 Nov 1979 N 4 21-41 32.5 703 19 2-45 8 16 5-81 5 - 8 0 +14.8 - 4 5 - 5.8 - 25 - 4.0 - 2.0 + 0.9 -109 - 0.7 - 70 - 0 1 - 9 1 -15.0 -12 1 - 46 - 38 - 66 +26.1 -11.0 - 57 - 3.0 - 77 +11.9 - 1.0 - 04 - 4.5 - 40 + 14.0 -170 - 5 5 - 4.5 + 14.8 - 28 - 60 0.0 - 0.5 15 Nov 1979 D 2 36-45 40.5 40.5 16 7-97.7 -364.6-445 6 1 Dec. 1979 N 24 33-87 55.3 2696 484-62.2 9.2-101.4 14 Dec. 1979 D 74 29-88 50 8 1464 48 0-53 6 18 8-82 8 3 Jan 1980 N 39 34-110 566 313.2 50 9-62.3 8 7-104.5 16 Jan 1980 D 246 31-124 54 1 108 8 52.8-554 27.2-81 0 4 Feb 1980 N 581 33-109 58 .1 99.2 57.3-58.9 324-838 15 Feb. 1980 D 1723 30-97 60.1 77.3 59.7-60.5 37.5-82.7 5 Mar 1980 19 Mar 1980 N D 507 217 39-95 47-121 61 0 71 9 805 103.5 60 2-61 8 70.5-733 37.9-84 1 457-98.1 1 Apr 1980 N 746 45-115 72 6 81 7 72.0-73 2 49.3-95 9 16 Apr 1980 D 437 44-108 79.6 105 1 78 6 80.6 53 1-106 0 5 May 1980 N 1202 54-114 797 90.5 79.2-802 55.2-1042 19 May 1980 2 June 1980 D N 239 614 59-116 72-139 88 8 1038 127 1 76 5 87.4-90.2 103.1-104 5 59.8-1178 81 3-126 3 19 June 1980 D 254 86-139 115.9 99 6 114 7-117 1 90 2-141.6 7 July 1980 N 750 90-149 120.5 842 119.8-121 2 96.9-144 1 21 July 1980 D 142 104-153 124 3 846 1228-1258 100 6-148 0 5 Aug 1980 26 Aug 1980 N D 794 1 102-155 157 130 9 157.0 73.0 130.3-131 5 108 .9-152.9 7 Sept 1980 N' 116 123-158 1460 42 1 144 8-147 2 129 0-163 0 22 Sept 1980 D- 12 143-162 151 7 392 1477-155.6 132 3-171 1 6 Oct 1980 N" 82 137-168 148 7 36.7 147 4-150 0 132 7-164.7 20 Oct 1980 D" 5 118-156 141 0 251 0 121 3-160.7 48.5-233.5 3 Nov. 1980 N 27 140-173 152 9 67 1 149.7-156 .1 130 1-175.7 18 Nov 1980 D 18 143-157 151.9 14.6 150 0-1538 1408-163.0 1 Dec 1980 N 4 149-153 151.5 3.7 148 4-154.6 140 3-1627 15 Dec 1980 D 1 147 147 0 — — — 7 Jan 1981 2 Feb. 1981 16 Feb. 1981 N N D 7 1 2 127-167 157 130-150 1430 157.0 140.0 211 7 200.0 129 5-156 5 12 9-267.1 89 1-196.9 -760 2-1.040.2 7 Apr 1981 N 2 145-146 145 5 05 139.1-151 9 100 5-190.5 20 Apr 1981 D 3 133-148 141.0 57 0 122.2-1598 66.1-215.9 4 May 1981 N 20 148-167 155.8 23 5 153 5-158.1 141 9-169.7 2 June 1981 N 5 143-161 153.0 43 5 144 8-161 2 122.6-1834 15 June 1981 D 1 1 146-174 159.0 532 154 1-1639 135 .9-182 1 1 July 1981 N 9 143-171 159.0 850 151 9-166 1 128 1-189 9 20 July 1981 D 199 146-176 158.5 40.5 157.6-159.4 142 1-1749 1980 spring-spawned group 3 Aug 1980 N 7 58-80 654 640 58 0-72 8 35 7-95.1 1980 fall-spawned g oup 18 Nov 1980 1 Dec 1980 D N 1 62 68 22-108 68.0 526 267 0 48 4-56 8 15 0-90.2 -154 -20.1 +19.6 - 22 - 7.5 -153 - 64 - 05 - 3.8 -142 - 62 - 9.3 + 128 + 9.3 + 8.4 -130 + 8.3 + 1.8 15 Dec 1980 D 2 29-36 32.5 24 5 -12.0-770 -282 6-347 6 7 Jan 1981 N 191 31-91 52.1 231 2 49 9-54 3 12.9-91 3 21 Jan. 1981 D 25 33-82 54.3 2140 48.3-60.3 134-952 2 Feb. 1981 N 255 21-118 46.8 93.7 45.6-48.0 21 .9-71.7 16 Feb. 1981 D 49 33-108 62.1 330 2 56.9-67.3 13.4-110.8 2 Mar 1981 N 32 42-86 55.7 156.5 51.2-60.2 21 4-90.0 16 Mar 1981 D 11 43-87 56.2 1298 48 5-63 9 20 1-92.3 7 Apr 1981 20 Apr. 1981 N D 110 69 39-87 56-107 600 74.2 105.7 115.7 58 1-61.9 71 6-76.8 33 1-86.9 45.7-102.7 4 May 1981 N 97 64-107 80.4 797 78.6-82.2 57.0-103.8 19 May 1981 D 165 70-115 89.7 66.0 88.5-909 68.8-110.6 2 June 1981 N 128 78-129 102.5 91 1 100 8-104.2 77 9-127 1 15 June 1981 D 174 92-133 111.8 688 110.6-113.0 90.4-133.2 1 July 1981 20 July 1981 N D 48 152 106-137 107-145 120.2 133.2 398 41 8 118.4-1220 132.2-134.2 103.3-137 1 116 .5-149.9 3 Aug 1981 N 132 120-166 141.5 62.0 140.2-142.8 121 2-161.8 16 Aug 1981 D 428 105-170 143.3 49.3 142.6-144 0 125.2-161 4 1981 spring-spawned group 15 June 1981 D 2 42-47 44.5 12.5 12.7-763 -180.6-269 6 + 9.5 -145 20 July 1981 D 9 41-72 540 110.8 45.9-62.1 187-893 3 Aug 1981 N 2 58-79 685 220.5 -64 9-201 9 -876.8-1.013.8 341 4 0 30 H I C3 LLI < > O Oct 79 (88, 0.08) Dec 79 (8, 0.85) Jan 80 (14. 0.51 Feb 80 (40, 0.52 Mar 80 (54, 0.91 Apr 80 (92. 0.77) May 80 (99. 0.40 Jun 80 (108, 0.47 Jul 80 (91, 0.62 Aug 80 (103, 0.36) Sep 80 (74. 0.29) Dec 80 (12, 0.78) Jan 81 (38, 0.84 Feb 81 (22, 0.87 Apr 81 (43, 0.61) FISHERY BULLETIN: VOL. 82, NO. 2 Sep 80 Jun 80 Mar 80 Dec 80 Jan & Feb 81 Jan 80 120 150 TOTAL LENGTH (mm) 180 Rgure 2. — Monthly regressions of ovary weight on total length for Larimus fasciatus. Length of each line shows the observed size range. Sample sizes and r2 values are tabulated for each period. Regressions presented were significant at a = 0.05. were calculated by using initial x- intercepts to readjust the initial x- variable (time) scale, so that each final growth curve passed through the origin (Table 4). Final calculated hatching dates are mean values because regressions predict aver- ages. Rate parameters — e.g., regression coeffi- cients, growth/30 d, and von Bertalanffy K values — fitted to observed size data are the same within rounding error regardless of the hatching dates used, because curves are fitted to the same time dimension between the initial and last collec- tion of a cohort. Total annual mortality rates (1 — S) were calcu- lated on a time-specific and cohort-specific basis from the expression S = Nt/N0 where S = rate of survival, and Nt and N0 are the numbers offish at 342 STANDARD and CHITTENDEN: REPRODUCTION, MOVEMENT, AND POPULATION OF BANDED DRUM 18.0 X Ul Q O o z < LU 5 16.0 14.0 120 O \- < 5 10.0 o o Q < 8.0 6.0 4.0 2.0 - 0.0 95% Co onfidence Limits j I About the Mean l Mean Range -tiU $ ¥ T * OCT NOV DEC JAN FEB MAR APR MAY M 1979 1980 I ai JUL AUG SEP OCT NOV DEC JAN FEB MAR APR 1981 FIGURE 3. — Monthly mean gonadosomatic indices, ranges, and 95' , confidence limits about means for female Larimus fasciatus, spawned groups pooled, October 1979-April 1981. TABLE 4. — Summary of iterative process used to calculate final hatching dates and set time scales for growth calculations. Equations describe regressions of observed mean total length (TL) on age in days Initial age values and growth equations were scaled to a 1 October hatching date. Final fitted regressions are in Figure 7. Initial growth Initial Final growth Final calculated Spawned group equation x-intercept equation hatching date Fall 1978 y = 28.59 - 0.40825X - 0.00031X2 -67 y = -0 13- 0 44924x - 000031X2 27 July 1978 Fall 1978, initial two y = -8.28 * 0.58365x 14 y = -0.21 + 0.56983x 15 Oct. 1978 collections deleted - 0.00049x2 - 0.00049X2 Fall 1979 y = 7.41 + 0.51937x - 0.00045x2 -14 y = 0.13 - 0.53162X - 0 00045x2 17 Sept. 1979 Fall 1980 y = 12.94 + 0 36704X -35 y = 0 09 - 0 36704x 27 Aug. 1980 Fall 1980, initial two collections deleted y = -7.47 + 0.45303x 16 y = -0.22 + 0.45303X 17 Oct. 1980 343 FISHERY BULLETIN: VOL. 82, NO. 2 age each month. Analyses excluded several months in which estimates were not reliable be- cause of incomplete recruitment (2 mo, time- specific; 6 mo, cohort-specific), seeming immigra- tion ( 1 mo, time-specific), or some stations were not occupied in one cruise (1 mo, cohort-specific). Pooled estimates of S were calculated using Heincke's procedure (Ricker 1975) and were con- verted to 1 - S and Z using relationships in Gul- land (1969:59). Observed estimates were compared against theoretical values calculated from the ex- pression Z = 4.6/number of years in life span (Royce 1972:238). Typical maximum life span was approximated by the Beverton-Holt yield model parameter tL (Gulland 1969), and typical maximum size was approximated as a correspond- ing length (l/J following Alverson and Carney's (1975) definition that only 0.5-1% of the catch ex- ceeds age tL. Values of li were calculated from the cumulative length frequency for all fish captured in the period October 1977- August 1981. We calcu- lated specific values of ti from lL by solving for time in von Bertalanffy (Gulland 1969:40) and quadratic regression equations. Total mortality rates and growth data presented are termed ap- parent because they may be affected by emigration as noted; if so, they overestimate mortality but underestimate sizes at age. Ovaries were prepared to estimate fecundity (FEC) using procedures similar to Bagenal (1957) and Simpson (1959). Entire ovaries of 60 Early Developing, Late Developing, Gravid, or Ripe fish were removed, split, everted, placed in Gilson's solution for 1-3 mo, and agitated using a magnetic stirrer to enhance separation of ova from connec- tive tissue. Connective tissue was removed and supernatant siphoned off until only ova remained. Eggs were then placed in a beaker, filled to 200 ml with water, and magnetically stirred to be uni- formly dispersed. Fecundity was determined for each fish by taking a 2 ml sample from each of three fixed levels in the beaker (at 25, 100, 175 ml) to enumerate eggs. Samples were pooled to calcu- late a mean/2 ml for each fish because, although significant, differences in mean egg count per level over all fish were small (627, 592, and 598, respec- tively; n = 60 for each level). Mean counts/2 ml were expanded to determine fecundity as the number of eggs in the total water volume. Regression relationships were calculated fol- lowing standard procedures (Helwig and Council 1979; Snedecor and Cochran 1980). Von Ber- talanffy growth was calculated using Fabens' (1965) program. All length measurements pre- sented herein are total length unless stated otherwise, and all length frequencies are moving averages of three. Conversions between standard length and total length used regressions presented herein. We use the verb "recruit" and the noun "re- cruitment" herein to describe areas in which young L. fasciatus descend to the bottom from their pelagic (Johnson 1978) early stages. This usage conforms to Beverton and Holt's (1957) meaning of recruitment, because these bottom areas are exploited, and to Ricker 's (1975) mean- ing, because fish also then enter the exploited phase of life. RESULTS Maturation and Spawning Periodicity Larimus fasciatus from the northwestern Gulf mature at 80-130 mm as they approach age I. Gonad development was distinct at 80-150 mm when most females entered the Early Developing stage (Fig. 4). All fish in Late Developing and later stages were >130 mm. These data are supported by regressions of ovary weight on length (Fig. 2) in which extrapolated x- intercepts were 75-125 mm during the April-October period when spawning occurs. Age compositions and sizes presented later indicate L. fasciatus mature to first spawn at 12-14 mo. Little somatic growth seemingly occurs after L. fasciatus enter late stages of gonad development. Mean lengths of fish were 146 mm in the Late Developing stage, 147 mm when Gravid, 149 mm when Ripe and Spawning/Spent, and 150 mm when Resting (Fig. 4). Maximum and minimum sizes also remained constant through these stages. Larimus fasciatus spawn within a broad period from April through November. Fish in Ripe or Gravid stages occurred throughout this period (Fig. 5), slopes and elevations of regressions of ovary weight on length generally were highest (Fig. 2), and GSI maximums usually were high (Fig. 3). Gonad analyses are supported by 1) recur- rent collections of fish 20-40 mm from November through February each year in the period 1978-81 which probably were 1-3 mo old and indicate spawning from September through November (Fig. 6), and 2) collections of distinct groups offish 40-80 mm in the period mid-June through August in 1980 and 1981 which probably were 3-5 mo old and indicate spawning from April through June. Little or no spawning of L. fasciatus occurs from 344 STANDARD and CHITTENDEN: REPRODUCTION. MOVEMENT, AND POPULATION OF BANDED DRUM 60 45 30 15 0 20 15 10 5 0 10 5 0 15 10 5 0 10 5h 0 10 5 0 5 0 5 0 Immature n=929 x=50 1 1 1 1 1 Maturing Virgin n = 487 x = 87 Early Developing n =175 x = 120 ,^\ i ' i Late Developing n 119 x = 146 ~I r— Gravid n = 69 x = 147 ■> r A. Ripe n = 57 x = 149 -i 1 1 r— Spawning/Spent n = 31 x=149 A. — i — > — i — r Resting n=47 x-150 ^-V /V\ FIGURE Larimus maturit\ 1 — — ' 1 40 80 120 160 200 TOTAL LENGTH (mm) 4. — Length frequencies of immature and female fasciatus by gonad maturity stage. See Table 1 for stage criteria. December through March. Almost all fish were in the Immature, Maturing Virgin, Early Develop- ing, or Resting stages in that period (Fig. 5). Few were in the Late Developing stage then and none were Gravid or Ripe. In addition, mean and maximum GSI values were lowest during the December-March period (Fig. 3) as were the slopes and elevations of regressions of ovary weight on length (Fig. 2). Gonad analyses are supported by the absence offish 20-80 mm from late February to mid-June each year (Fig. 6) with the exception of fall-spawned fish whose growth in late winter and spring is clearly followed. Although L. fasciatus spawn within a broad time period, spawning primarily occurs during what we interpret as two discrete periods, a major fall period (September-November) and a very minor spring period (April-June). Mean and/or maximum GSI values were highest in the periods May-June and September- November, and these peaks were separated by troughs in the periods July-August and December- April (Fig. 3). Al- though few fish were Ripe in August, nearly all were in Late Developing, Gravid, Ripe, and Spawning/Spent stages from September through November (Fig. 5); few were in Immature, Matur- ing Virgin, or Early Developing stages then. Fall- spawned fish greatly predominated each year and formed length-frequency modes easily followed through the spring and summer after their re- cruitment in fall and winter (Fig. 6). A minor spring spawn is indicated by distinct, but not abundant, groups offish 40-80 mm in the periods August 1980 and June- August 1981 ( Fig. 6 ) and by the occurrence of a few (14) Gravid and Ripe stage fish from April through June and Late Developing fish in March (Fig. 5). No recently recruited spring-spawned fish were evident after August (Fig. 6), but they may be represented by the few intermediate-sized fish from an unclear spawned group in January and March 1980 and February 1981. Little or no spawning of L. fasciatus occurs in July and August. No fish 20-40 mm (1-3 mo old) were captured from July through October ( Fig. 6). Only two fish were Gravid or Ripe during July and August and few were in the Late Developing stage then (Fig. 5). Calculated hatching dates agree with the major fall-spawning period — September- November — indicated by gonad and length-frequency analyses. Depending on data points included, cal- culated hatching dates were 27 July or 15 October 1978, 17 September 1979, and 27 August or 17 October 1980 (Table 4). Hatching dates of 15 Oc- tober and 17 October seem most realistic for 1978 and 1980. The earlier dates for those years are based on regressions fitted to all collections. The earliest two collections in those years, however, probably contribute upwardly biased size data 345 FISHERY BULLETIN: VOL. 82, NO 2 IMi 111(1 50- Ocl 79 n 100 ' oh^mi 12 3 4 5 6 7 Nov 79 n 9 150 -, 100- 50 > U z 3 12 3 4 5 6 o w rx 150 100 I 50-A 0- Dec 79 n 93 Mk- 150 100 50 150 Feb 80 n 181 50- I50 IMC 50 12 3 4 5 6 7 8 Mar 80 n 150 Jun 80 n 162 r - V 150 100- 50- to^.fti Ocl 80 n 124 150 100 -T-GP- *-A-f ™i a Feb 81 n 180 ?3 150-1 / A irm 50 12 3 4 5 6 7 JillHd n 132 1 2 3 4 5 6 7 8 150 150 Nov 80 N 26 , i.i-r^n , , , u o 100- 50 100- 50- 150 I ( ](j 50- i i t T1 0 12 3 4 5 6 7 8 1 2 3 4 5 6 7 Apr 80 n 166 150 100 50- AT^TT" , — i — r- o LJ , I,, 12 3 4 5 6 7 Aug 80 n 153 0 150-, H 50 1 1 r 12 3 4 5 6 Dec 80 n 60 W^ 0 -(4-1 — i — i — i — --*- 1 2 3 4 5 6 7 8 Mar 81 n 38 JU- 12 3 4 5 6 7 8 150 i Kill 50-, 12 3 4 5 6 7 150 100- 50 Jan 80 n 193 150 100 50- i — i — i — i — i — r- 0 1 2 3 4 5 6 7 8 i 50 May 80 n 141 /. I 00 50- 253 "T— i 1 r 12 3 4 5 6 7 8 0 12 3 4 5 6 Sep 80 n 74 1 2 3 4 5 6 7 ^ 0 Apr81 n 128 H "T^ — i — r 2 3 4 5 6 7 150 100 50 1 , ^-i, iyrw> f\ r*- 1-4 Dec 79 N F78.1 n 27 5 0 5 0 15 10 5h 0 40 30 - 20 - 10 - •F79.0- . ■ 14-19 Dec 79 D ■/Vy F78-l ° 75 I F79.0— I "> 3-6 Jan 80 N F78.1 n 40 «! ■ 1 F79.0 16-20 Jan 80 D n 246 -F79.0 4-11 Feb 80 N n 582 TOTAL LENGTH (mm) FIGURE 6. — Monthly length frequencies of Larimus fasciatus off Freeport, Tex. Day and night cruises are indicated by D and N. Bars in each panel depict size range of indicated spawned group; further detail is in Table 3. The letter and first two digits within a bar indicate spawned groups; the last digit is age in years, e.g., F77, 1 347 30 - 20 - 10 0- U z LU o LU a. 15r 10 - 5 15-20 Feb 80 D n 1723 F79.0- 5-8Mar80N n 508 F78.1 19-23 Mar 80 D n 262 F78.1 I 1 -A ^ r^ — | 1 I F79.0 \ 1-5 Apr 80 N n = 746 F79.0 ( 16-20 Apr 80 D n = 445 F790 i 5-10May80N n=1202 40 19-22 May 80 D F79.0 i n = 241 F78.1 i- \ -r — -* — i 80 120 160 200 40 30 20 ioh 0 I5r 10 5 FISHERY BULLETIN: VOL. 82, NO. 2 - F7.9.0 1 2-6Jun80N n 636 F79.0- 19-24 Jun 80 D n-280 r 21-24 Jul 80 D n 142 5 0 F79.0 5r 0 10 5 0 5 0 26-29 Aug 80 D n 1 9 7-11 Sep 80 N n=117 F79.H F78.2 — ^ 22-25 Sep 80 D n = 12 F79.1 I < 10 r 6-9 Oct 80 N n = 82 5h 0 T T F79.1 l^ \ iv 20-31 Oct SOD n = 5 F79.1 I 1 tJ=Y- 40 80 120 160 200 TOTAL LENGTH (mm) FIGURE 6. — Continued — represents the fall 1977-spawned group when age I. Age designation for each cohort changed in August to simplify reference between Figures 6 and 10 D; true ages in August only approach those indicated. 348 STANDARD and CHITTENDEN: REPRODUCTION. MOVEMENT, AND POPULATION OF BANDED DRUM 5 0 5 0 5" 0L 5r 0 3-5 Nov 80 N n 27 -i 1 < 1 ' 18-21 Nov 80 D n 19 F80.0 -"-! ' F79.1 i 1 T F79.1 I 1 F80.0 1-4 Dec 80 N n 66 ,~a^W\^ J^ ^ T F80.0 A A F79.1 15-19 Dec 80 D F79.1 n 3 — i T T 7-13 Jan 81 N n 198 F-F79.H ^w"w> **■ ^ , 21-24 Jan 81 D n 25 2-8 Feb 81 N n 256 7 T F79.1 ^1 - 5" o1- 16-19 Feb 81 D ^80.0-^ 9 , (F79ln51 ry\^r^' - 1 -. - , — . — i 2-6 Mar 81 N n 32 5 " o- F80.0- t-4~- 16-19 Mar 81 D n 11 10r 5 0 •F80.0- f /W^n. 7-10 Apr 81 N n 112 F79.1 40 i — F80.0- 20-23 Apr 81 D F79.1 n72 WV\ 10 5 0 4-8 May 81 N F80.0-I n 117 F79.1 A^ A 15 - 10 - 5 - o' — r 10 - 5- o1 — ■ — r 15-19 Jun 81 D 10r n 187 19-27 May 81 D F80.0— i n 165 F80.0- — ■ 1 ' 1 2-9 Jun 81 N n 133 F79.1 ^r F80.0 5h 0 10 5 0 S81.0 1-8 Jul 81 N n 57 F80.0 20 20-24 Jul 81 D n 360 F80.0 F79.1 15 1 u 10 1 A 5 S81.0 .A \ 0 1 ■ 1~^ i^. U-U 10 0 3-9 Aug 81 N n 134 S81.0 i < ■ F80.1 -<* — r- 16-20 Aug 81 D 1-F8O, I 1 30 n=428 f\ 10 i > i .-./ V-. 0 80 120 160 200 TOTAL LENGTH (mm) 40 80 120 160 200 FIGURE 6.— Continued. 349 FISHERY BULLETIN: VOL. 82, NO. 2 that reflect gear selection for large fish or incom- plete recruitment of smaller fish; mean sizes in those collections were as large as or larger than means in subsequent collections and seem to be outliers (Fig. 7A, C). Coefficients of determination were higher when the earliest two collections were excluded (Fig. 7A, C). The period of successful spawning spans 60-90 d within the major September through November interval. Based on mean 99% confidence limits for observations and growth per day in each April- A. Fall78 All Collections: y=- 13- 44924x- 00031 x- x}-- 90 Two Collections Deleted: y = - 21+ 56983x- 00049x- r'= 91 z LU < o r- 200 l 80 160 140 120 1 00 80 60 40 20 0 mtial Two Collections Deleted 95% Confidence Limits I t About The Observed Mean j t Mean Range 60 120 180 240 300 360 42U 480 540 600 660 720 AGE (DAYS) June period, calculated durations of fall spawning periods were 76 d in 1977, 76 in 1978, 87 in 1979, and 64 in 1980 (Table 2). These estimates fall within and agree with the broad 90-d duration of the fall-spawning period indicated by gonad maturity and weight data. The predominant fall-spawned groups are pro- duced by fish that first spawn when 12-14 mo old and the minor spring-spawned groups by fish 19-21 mo old. Fall-spawned fish apparently do not spawn when 5-7 mo old in their first spring, because GSI values for fall 1979 and fall 1980 fish remained low during their initial springs (Fig. 8B, C). Peaks in GSI values from September through November in 1979 and 1980 were formed by fall 1978 and fall 1979 fish, first spawning at 12-14 mo of age (Fig. 8 A, B). Peaks in GSI values from April through June in 1980 and 1981 were formed by fall 1978 and fall 1979 fish that spawned when 19-21 mo old (Fig. 8A, B). We were not able to determine age when spring-spawned groups spawn, because these fish were clearly identifiable only until 3-5 mo old (see section on Age Determination and Growth). FIGURE 7. — Mean observed and predicted sizes at age (days) for the 1978 (A), 1979 (B I, and 1980 (C ) fall-spawned groups of Larimus fasciatus. Mean sizes at age were regressed on age scaled to calculated hatching dates of 27 July and 15 October 1978, 17 September 1979, and 27 August and 17 October 1980. Observed mean lengths and their confidence limits are from Table 3. Regressions were significant at a = 0.001. ONDJ FMAMJ J ASONDJ FMAMJ J ASO 1 — i — ' — i — i — i — i — i — i — i— i — i — i — i — i — i — i i—j I ' ' ' i__i 1978 I 1979 I 1980 COLLECTION DATE B. Fall79 All Collections: y= .13+ 53162x-00045x- rJ=.94 0 60 120 180 240 300 360 420 480 540 600 660 720 AGE (DAYS) SONDJ FMAMJJASONDJFMAMJJAS ' — I — i — i — J — i — i — i — i — i — i i i i i i 1979 I 1980 I 1981 COLLECTION DATE (3 Z LU < o 180 160 - 140 120 100 80 60 - 40 - 20 0 C. Fall80 All Collections: y 09* 36704x r'~- 82 Two Collections Deleted: y = - 22 + 45303x r'= 93 »/ /■•' * d i." -Initial Two Collections Deleted ■All Collections Included 95% Confidence Limits 1 1 About The Observed Mean | t Mean Range J_ _i_ _i_ -L- _l_ 60 120 180 240 300 360 420 480 540 600 660 720 AGE (DAYS) ONDJ FMAMJ J ASONDJF MAMJ J ASO i — I — i — i — i I i I I i i 'I'' i_i i 1980 I 1981 I 1982 COLLECTION DATE 350 STANDARD and CHITTENDEN: REPRODUCTION, MOVEMENT, AND POPULATION OF BANDED DRUM Bathymetric Distribution, Recruitment, and Movements The bathymetric distribution of L. fasciatus in the northwestern Gulf off Freeport extends from <5 to 55 m. This species was most abundant at 5 m, the shallowest depth occupied (Fig. 9). Abun- dance declined sharply between 5 and 18 m and remained low from 18 to 36 m. Only one specimen was collected deeper than 36 m. Young-of-the-year L. fasciatus in the north- western Gulf recruit in waters <5-16 m when 2-4 mo old. Fall-spawned young 30-100 mm recruited almost exclusively in 5-16 m from November through April (Fig. 10A, B). Only four young-of- the-year specimens at 18-22 m were collected deeper than 16 m. Similarly, spring-spawned young 60-80 mm recruited only at 5-16 m in the period August-October (Fig. 10D). Greatest recruitment of fall-spawned L. fas- ciatus occurred in the shallowest depths sampled. Recent fall-spawned recruits were most abundant by far at 5 m in November- April (Fig. 10A, B). Abundance then sharply declined with depth and was very low deeper than 16 m in that period. 110 100 90 80 70 O 60 50 40 30 20 - 10 s • Pooled D 1977 A 1978 o 1979 T 1980 ♦ 1981 o Absent i afi 8t e 5 20 18 16 14 12 10 X UJ z 6 o < s O 2 V) O O 0 < Z 12 o C3 z 10 < ,-13-14 mo I— I 86 A Fall78 95% Confid Abo fence Limits ( J ut the Mean | 1 Mean 19-21 mo I 1 26 Range 22 15 10 B Fall79 -i — i — i — i — i — i — i — i — i — i — i — i — i I | 12-14 mo 64 74 5-7 mo T 1 r 40 ,, go 76 14 I 3,2 B6 7 82 I 91 ! i t i i i i i 26 19 mo I o 2r C Fall80 — i 1 r~ 5-7 mo |- 21 8 36 XX J* isA ONDJFMAMJJASONDJFMA 1979 1980 1981 FIGURE 8. — Monthly mean gonadosomatic indices, sample sizes. ranges, and 95'V confidence limits about means for the 1978 i Ao 1979 (B), and 1980 (C) fall-spawned groups of Larim us fasciatus. Age in months is indicated for each spring and fall spawning period. 10 20 30 40 50 60 DEPTH (m) 70 80 90 100 FIGURE 9. — Catch/effort (mean number of individuals per 10 min tow) by depth for Larimus fasciatus off Freeport, Tex., each year and pooled, October 1977- August 1981. 351 FISHERY BULLETIN: VOL. 82. NO. 2 Fall-spawned L. fasciatus gradually disperse toward deeper water in late spring or summer. The distribution of young-of-the-year in May- July was similar to that in November- April (Fig. IOC). Fish approaching age I showed a clear offshore shift in distribution by August-October when abundance greatly declined at 5-13 m, became highest at 16-22 m, and was as high at 24-27 m as at 5-13 m (Fig. 10D). Fish approaching age I became distrib- uted to 36 m, a depth they previously did not oc- cupy, in August-October. Larger L. fasciatus lead the offshore dispersal as they approach age I. Size compositions of the young-of-the-year were uniform with depth in November-April (Fig. 10A, B). They became skewed toward the right in May-July and show a gradient of increasing size with depth in August- October which suggests larger, presumably older, fish move offshore first. Adult fall-spawned L. fasciatus in the north- western Gulf occupy the 13-24 m bathymetric range from November through April while A. NOVEMBER-JANUARY 15r 10 5 0 o Z 0 LU D O 15 LU CC 10 - LL 5 0 5r 0 •0 C/f =5 6 vv 9M n 171 I C/f<0 1 C/f 1 6 0 1 /VVV ' I ■4-^ ~i ■ — 13M n 53 ■I— I C/f = 0.2 C/f = 1 8 t 16M n 282 A~t ^ C/! = 0 3 -*A 18M r n 3 C/f<0 1 0 C'f<0 1 1 — I — " ' 1 1 — °~> 1 1 — 1 r-O— o z LU D O LU rr ""OSpnng ^A^AA^JK,^ -i r 24M n 4 "I ' C f 0 1 hlH "T1 5r 27M n 6 0 40 — i — 80 C f<0 1 0 H~l i — -r < """> C f 0 2 5M n 28 C/f-0 2 Spring ^0—1 — c-a — Q- C/f=1 0 9M n 16 C f<0 1 Spring 0 "I r C f = 06 -I I — - C f<0 1 5r 13M Spring n = 68 0 ** C/f= 2 6 0 T C f = 5 9 16M 30 1_ n 668 20- 10 0 C/f<0.1 Spring 0 * 15r 18M 10- n 154 5- 0 40 30 20- 10- 0 22M n 929 ,_ 24 Ml n = 47 — | r C f--2 1 A^ r- 27M 0 n 51 -i 1 1 r C f 2 0 36M n 10 C f-04 120 160 200 0 40 80 — i S*^ — q , 120 160 200 TOTAL LENGTH (mm) FIGURE 10. — Continued — period October 1977-August 1981 because length frequencies were similar each year. Designated ages are for fall-spawned fish except where noted. Age design n for each cohort changed in August. 353 FISHERY BULLETIN: VOL. 82, NO. 2 negligible even then. Although two fall-spawned groups, represented by young-of-the-year and age I fish, often were captured, only one predominated in any month except in July 1981 when fish ap- proaching age II were abundant but disappeared thereafter. Once fully recruited, each fall-spawned group predominated to age I and then nearly dis- appeared as the next fall group began to recruit. Larimus fasciatus is not abundant after 12 mo of age in the northwestern Gulf and reaches only 21-23 mo there. Fall-spawned fish were abundant after 12 mo of age (Fig. 6) only in July 1981 when they approached age II. Fall-spawned groups dis- appeared at 15-23 mo of age (Table 5, Fig. 6). The few spring-spawned fish captured were identifi- able only until 3-5 mo of age (Table 5, Fig. 6). Fish of intermediate size between clearly defined fall- spawned groups in January and March 1980 and February 1981 could have been spring-spawned, but their identity is not clear. Slightly larger L. fasciatus occur in the north central Gulf than in the northwestern area. (Harding 1949), moreover, do not and cannot re- solve this situation because of the original prob- lem: the underlying length frequency is not abso- lutely clear. However, the 48 mm range of sizes (139-187 mm) for fish in the north central Gulf is only slightly larger than a 35 mm range (130-165 mm) that tightly brackets most fish in the north- western Gulf where only one fall group predomi- nated (Fig. 11). Moreover, sizes in the north central Gulf in the period October-November were only slightly larger than and greatly overlap those for northwestern Gulf fish which were just age I. These facts suggest only one or at most two spawned groups predominated in the north cen- tral Gulf, fish just age I and age II. This interpreta- tion is supported by our findings noted later that 1) the largest fish we captured in the northwestern Gulf (182 mm) was only 20 mo old, 2) von Ber- talanffy predictions indicated mean sizes of 164 and 181 mm at age II in the northwestern Gulf depending upon variation between fall-spawned groups, and 3) the observed size range was 143-176 Table 5.- -Periods of time, sizes, and age when spawned groups of Larimus fasciatus were last captured. Spawned Period Size Age group last captured (mm TL) (mo) Comments Fall 1976 Early December 1977 136-141 15-16 Very few ever captured Fall 1977 Late June 1979 174 21-22 Few ever captured Fall 1978 Early September 1980 177 22-23 Few captured after October 1979. Fall 1979 Mid- July 1981 146-176 22-23 Few captured after October 1980, except for late July 1981 Spring 1980 Early August 1980 58-80 3-4 Collected only in August 1980 Fall 1980 Mid-August 1981 105-170 10-11 Still dominant in last col- lection Spring 1981 Early August 1981 58-79 3-4 Very few ever collected Maximum and mean sizes were greater in the north central Gulf (max. = 187 mm, x = 160 mm) than in the northwestern area (max. = 173, x = 146) during the period October and November, ig- noring the seven recently hatched recruits cap- tured in the latter area (Fig. 11). Only one or two spawned groups of L. fasciatus apparently predominate in the north central Gulf, probably fish that became age I and age II in the fall. We are not able to confidently identify modal groups to assign ages and, particularly, delineate sizes where age groups overlap in that area (Fig. 11), because we made only one cruise there, not the time-intensive series that permits confident age designations for the northwestern Gulf. Analyses such as linear transformation of cumulative per- centage frequencies using probability paper 30 20 > O 10 z 111 § ° 111 Northwestern Gulf n = 414 x = 146 mm s2 = 144.28 Recently Hatched I 1 North Central Gulf n 64 x = 160 mm s2 = 141.61 i (?) ii i 1 1 1 ^W^Vy. 40 80 120 160 TOTAL LENGTH (mm) 200 FIGURE 11. — Length frequencies and age designations for all Larimus fasciatus captured in the period October-November in the northwestern (1977-80) and north central (1982) Gulf of Mexico. Means, n, and s2 ignore seven recent recruits (< 70 mm) in the northwestern Gulf. 354 STANDARD and CHITTENDEN: REPRODUCTION. MOVEMENT, AND POPULATION OF BANDED DRUM mm in July for the abundant fall 1979-spawned group as it approached age II in the northwestern Gulf. We have assumed fish in the north central Gulf were produced in the major fall-spawning period, not the minor spring one, and that differ- ences in size compositions do not reflect only possi- ble intra-Gulf differences and greater size at age in the north central Gulf. Comprehensive collec- tions in that area are needed to resolve the latter assumption. Apparent growth of L. fasciatus in the north- western Gulf was similar between fall-spawned groups, mean sizes being 130-150 mm at age I and 155-180 mm at or approaching age II. Observed mean sizes and ranges at age I, based on pooled data from September and October (Table 3), were 134 mm (120-143) for fall 1976 fish, 148 mm (145- 152) for fall 1977 fish, 145 mm (127-164) for fall 1978 fish, and 147 mm (118-168) for fall 1979 fish. These sizes at age I agree with quadratic regres- sion predictions of 142 and 134 mm (Fig. 7) and von Bertalanffy model predictions of 137 and 131 mm for fall 1978 and fall 1979 fish, respectively. Ob- served mean size and ranges at or approaching age II (Table 3) was 159 mm (150-172) in July 1980 for fall 1978 fish, 159 mm ( 143-176 1 in July 1981 for fall 1979 fish, and 164 mm (156-177) in August- September 1980 for fall 1978 fish. These values are only slightly larger than a quadratic regression prediction of 155 mm at age II for fall 1978 fish (Fig. 7A), and are the same as or a little smaller than von Bertalanffy model predictions of 164 and 181 mm at age II for fall 1979 and fall 1978 fish, respectively. Fitted von Bertalanffy equations based on hatching dates of 15 October for fall 1978 fish and 17 September for fall 1979 fish were Fall 1978: I, = 201 (1 - e Fall 1979: lt = 176 (1 - e -0.003162(/-1.574> -0.003670U + 4.696) young, or a pattern of incomplete recruitment in which smaller, younger fish gradually recruit to join early recruits that are slightly larger and older. Growth increments for age 0 fish were greatest in March-June, peaking in early June (26.7 and 32.1 mm/30 d). Growth increments de- creased as maturation occurred from July through September, became small after spawning in October-January, and increased slightly in June-August as the fish approached age II. Apparent sizes of L. fasciatus at or approaching age I in the northwestern Gulf reached a plateau in August-September and for many months later. Mean sizes of fish at or approaching age I re- mained constant then, and 99% confidence limits for observations generally remained uniform at 130-160 mm (Table 3, Fig. 6). This pattern suggests an exodus of larger individuals and/or cessation of growth coincident with gonad maturation (see sec- tion on Maturation and Spawning Periodicity). Maximum Size, Life Span, and Mortality The maximum size L. fasciatus reach in the northwestern Gulf is about 180 mm, but more typ- ically individuals reach only 160-165 mm. The largest of the 13,676 fish we collected there was 182 mm; 99% were <161 mm and 99.5% were <164 mm (Fig. 13), these sizes being estimates of lL. Typical maximum life span of L. fasciatus ap- pears to be only 1-2 yr in the northwestern Gulf. A value of ti, = 1-2 yr is reasonable for that area because 1) fish average 155-180 mm at or approach- ing age II and 130-150 mm at age I with the upper 99% confidence limits for observations at age I generally being 160-165 mm (Table 3 ), 2 ) li values of 161 and 164 mm predict ti values of 1.3-2.0 yr (Table 6), 3) the largest specimen was 20 mo old when collected in May 1980, and 4) L. fasciatus disappeared off Texas at 15-23 mo of age (Table 5), where /, = length in millimeters, and t = time in days. Annual K values were 1.15 and 1.34, respec- tively. Respective annual ^0 values (0.00431 and -0.01287) were small which may reflect our forc- ing the curve through the origin. Apparent growth of fall-spawned L. fasciatus follows an S-shaped intrayear pattern and is greatest in the spring and summer. Observed mean sizes at age showed a clearly S-shaped pat- tern (Fig. 7). Adjusted growth increments were small early in life (Fig. 12) and may reflect cool- water temperatures then, gear selection for larger TABLE 6. — Values of fj, (yr) for Larimus fasciatus cal- culated from li i mm TL) using quadratic and von Ber- talanffy equations scaled to hatching dates of 15 October for fall 1978 and 17 September for fall 1979 fish. The apex of the parabola for fall 1979 fish was 157 mm so that ti. values could not be calculated for that cohort. lL 'L calculated from: Spawned group Quadratic regression von Bertalanffy equation Fall 1978 Fall 1978 Fall 1979 Fall 1979 161 164 161 164 1.33 1 44 1.40 1.47 1 83 1.99 355 FISHERY BULLETIN: VOL. 82, NO. 2 35 . 30 (/> >. •a 8 25 E £ 20 h- Z UJ UJ 15 CC O z i o o cc o ° 1978 Fall Spawned Group a 1979 Fall Spawned Group o o r> A A_ 1 Nov 'Dec' Jan1 Feb 'Mar' Apr'May'jun1 Jul 'Aug1 Sep' Oct ' Nov1 Dec1 Jan ' Feb1 Mar! Apr 'May1 Jun'jul' 'Aug' Sep1 AGEO AGE 1 GROWTH PERIOD Oct FIGURE 12. — Monthly growth increments for 1978 and 1979 fall-spawned Larimus fasciatus. Unadjusted growth increments (Table 3) were converted to growth/30 d, omitting collections of five or less fish. Negative growth is rounded to zero. Values denoted by darkened symbols may reflect incomplete recruitment, gear selection for larger young, or cool-water temperatures early in life. 225 100 n = 13.676 / 200 AJ fi / o 175 /'Mm / c / I / 75 2 O 150 f \ / C r- Z \i y . > UJ 125 ^r A H 2 1°° 1 A )\ Ml < / \ K r HJLa rn cc / w( iV./ V^n ■o U. 75 I / V 1 m 20 50 / / \ ^ m z 25 jJ V H 50 1 00 1 50 TOTAL LENGTH (mm) 200 FIGURE 13. — Length frequencies and cumulative percentage of all Larimus fasciatus collected off Freeport, Tex., October 1977- August 1981. and few ever approached age II except in July 1981 as previously noted. The latter instance suggests larger, older specimens of L. fasciatus may occur elsewhere; if so, our estimate of tL may be too low for a stock that also ranges outside the northwest- ern Gulf. Larimus fasciatus from fall-spawned groups have an apparent total annual mortality rate of 90-100% in the northwestern Gulf, mean time and cohort-specific values being 92-100%. Time- specific values of 1 — S were 100% in 17 of the 38 mo in which fish were collected (Fig. 6) because only one fall-spawned group was present and N, was zero in the ratio N,/N0- Time-specific mortal- ity estimates for the 18 remaining months gave ratios whose percentage values ranged from 89.93 to 99.96%; 14 exceeded 94%. Pooled estimates using Heincke's procedure were 97.18-100% de- pending on the spawned groups compared (Table 7). By pooling the Heincke numerators and de- nominators from each comparison an average time-specific 1 - S was 98.24%, or 96.86% if data from July 1981 are included. Cohort-specific val- ues of 1 - S were 100% in 12 of the 24 mo for fall 1978 fish and in 2 of the 12 mo for fall 1979 fish because Nt was zero. Cohort-specific estimates for the remaining 6 mo in the fall 1978 cohort ranged from 97.83 to 99.81% except in June and July 1980 when estimates were 64.84 and 55.05% . Estimates for the remaining 7 mo in fall 1979 cohort ranged 356 STANDARD and CHITTENDEN: REPRODUCTION. MOVEMENT. AND POPULATION OF BANDED DRUM TABLE 7. — Pooled time-specific and cohort-specific mortality estimates for Larimus fasciatus using Heincke's procedure (Ricker 1975). Symbols represent: ./Vo> youngest spawned group in Heincke's estimate; S, annual survival rate; 1 - S, annual mortality rate; and 2, instantaneous mortality rate. N0 S 1 - S Z Time-specific Fall 1977 0 1 0000 X Fall 1978 0.0083 0 9917 4.79 Fall 1979 00180 0 9820 4.02 Fall 1980 00282 09718 3.57 Heincke's Pooled 00176 0 9824 404 Cohort-specific Fall 1978 00810 0 9190 251 Fall 1979 00317 0 9683 345 from 94.44 to 99.84** except in July 1981 when it was 74.43rr. The low estimates in June and July reflect the unusual instance previously noted of immigration by fish approaching age II. Pooled cohort-specific estimates of 1 - S using Heincke's procedure were 91.90°r for fall 1978 fish and 96.839? for fall 1979 fish (Table 7). Sex Ratio and Fecundity Male and female L. fasciatus appear equally abundant. The observed sex ratio of 1.00 males to 1.02 females among 2,502 mature or maturing fish examined in the period October 1979-April 1981 did not differ significantly from 1:1 ( \2 = 0.19; df = 1. a = 0.05). Mean fecundity of L. fasciatus in Gravid and Ripe stages was 70,453 eggs. Observed fecundity ranged from 32,333 to 143,800 eggs/female. Un- transformed and log-log transformed linear re- gressions of fecundity on total length and total weight and related statistics are presented in Table 8; the former regression is depicted in Fig- ure 14. The untransformed regression is a better fit (100r2 being 29.7 vs. 27.1 for length and 35.1 vs. 28.4 for weight), but the transformed regression permits extrapolation over a broader size range. Fecundity statistics were based only on Gravid and Ripe fish, because residual plots for untrans- formed data indicated a relationship between fecundity and maturity stage (Fig. 15): maximum fecundity occurred in the Gravid and Ripe stages. Weight, Girth, and Length Relationships Total weight-total length, girth-total length, and standard length-total length regressions are presented in Table 8 with related statistics. Total length-total weight regressions for males and females were not significantly different in slope (F = 0.35, df = 1, 1936, a = 0.05) or in elevation (F = 1.62, df - 1, 1936, a == 0.05) so one pooled equation is presented for them. Total length-total weight regressions for males and females pooled and for immatures, males, and females pooled were significantly different in slope (F = 44.87, df = 1, 4808, a = 0.05), but one equation that pools all sizes may be useful and is presented. Calcu- lated slopes significantly exceeded (3 = 3.0 at a = 0.05 for both length-weight relationships (males and females pooled, t = 53.06; immatures, males, and females pooled, t = 60.41). DISCUSSION Spawning Periodicity We found that the broad April to November period within which L. fasciatus spawns generally agrees with many studies, including Hildebrand and Cable (1934), Miller (1965), Christmas and Waller (1973), and Ross (1978). However, our in- terpretation is new that little or no spawning oc- TABLE 8. — Fecundity, total weight-total length, girth-total length, and standard length-tota Larimus fasciatus with supporting statistics. All regressions were significant at « = 0.01; v is regression. Measures are grams and millimeters. See Methods for symbols. length regressions for from Ricker 's 11973) GM Equation TL range 100r2 Residual MS Corrected total SSx Corrected total SSy GM FEC = -295,307 + 2.498 36 TL log10FEC = -5.5049 - 4 7689 log10TL FEC = -32.999 * 2.223.57 TW log10FEC = 2.6564 - 1.3013 log,0TW log10TW = -5.5981 + 3.3481 log10 TL (males + females) log,0TW = -5.4761 + 3.2883 log, 0 TL (males + females f immatures) G = -1.21 + 0.77 TL TL = 2.35 + 1 29 G SL = -5.63 + 0 83 TL TL = 6.96 + 1.20 SL 40 136-163 297 4.966 « 105 1.280.00 2,686 x 107 146.4 70,453 4,581.2 40 136-163 27.1 00178 0.0111 0 9280 2.17 4.82 9.1573 40 136-163 35 .1 4.588 x 105 1.906.00 2,686 x 107 46.5 70.453 3.753.7 40 136-163 28.4 00175 0 1554 09280 1 67 482 2.4435 1.938 52-179 99.3 0.0015 3442 38867 2.03 1.20 3.3605 2,874 22-179 99.4 00025 108 91 1.18474 1.92 0.84 32983 2,871 22-179 99.2 775 2.628.914 (G) 92 09 69 35 0.77 2,871 22-179 99.2 1309 4.439.633 (TL) 69.35 92 09 1 30 2.875 22-179 998 2 23 3,066.599 (SL) 92 08 7076 0.83 2.875 22-179 99.8 3.24 4.446.964 (TL) 70.76 92.08 1.20 357 FISHERY BULLETIN: VOL. 82, NO. 2 160 140 120 (A O) en a> — 100 o M b o 2 80 a z D O LU 60 40 20 0 160 r 140 120 100 80 60 40 20 j Ol i I i_ J i i i_ j i 135 140 145 150 155 160 165 34 37 40 43 46 49 52 55 58 61 64 67 TOTAL LENGTH (mm) WEIGHT (g) FIGURE 14. — Regressions of fecundity on total length and total weight with 95% confidence limits for yx for Larimus fasciatus. 10 0 >- 10 o z w 0 O 10 LU (X Li. 10 A TL Early Developing - .L.J.B Gravid Ripe RESIDUALS (1000s) 10 0 Late Developing 10 0 10 0 10 B. TW Early Developing Late Developing Gravid Ripe ■ J. .1. RESIDUALS (1000s) FIGURE 15. — Residual plots of maturity stages of Larimus fasciatus for relationship between fecundity and total length (A) and total weight (B) regressions. curs in July and August and that spawning primarily occurs in two discrete periods, a major fall peak in September-November and a minor spring peak in April-June. The existence of dis- tinct spring and fall spawning periods is supported by 1) Hoese's (1965) collection of larval fish only in June (11.5-25 mm SL modal length) and October (14-39 mm SL modal length) off Port Aransas, Tex., 2) Ross's (1978) collection off North Carolina of presumably fall-spawned fish in February (79 mm modal TL = 60 mm SL) and what must be spring-spawned fish in July (75 mm modal TL = 57 mm SL), 3) Powles's (1980) collections of larval fish in the periods April-May and August- September but not in June-July between Cape Canaveral and Cape Fear, and 4) collections of 358 STANDARD and CHITTENDEN: REPRODUCTION. MOVEMENT. AND POPULATION OF BANDED DRUM larval fish in the periods April-June and August- October between Cape Lookout and Chesapeake Bay (Berrien pers. commun. footnote 7 in Powles 1980). This bimodal pattern, moreover, is similar to findings of distinct spring- and fall-spawning periods in other Gulf fishes such as Cynoscion arenarius (Shlossman and Chittenden 1981), C. nothus (DeVries and Chittenden 1982), and Pep- rilus burti (Murphy 1981). Murphy and Chitten- den7 integrated biological findings with hydro- graphic data of Kelly and Randall* and suggested hydrographic reasons for this pattern in P. burti and C. nothus, which probably apply also to L. fasciatus and many other species: spawning is timed to coincide with the periodicity of downcoast alongshore currents (toward Mexico) and onshore Ekman transport at the surface. These phenomena probably transport pelagic eggs and larvae inshore and "downstream" to nurseries in the northwestern Gulf from spawning grounds lo- cated "upstream" in or toward the north central Gulf. Current transport mechanisms reverse in the summer (a variable period but about mid- June-early August) and would carry pelagic eggs and larvae offshore, which presumably is un- favorable to survival in many species, or toward the north central Gulf. If spawning is substantial, and successful, during summer, our length- frequency data and those for the other species cited indicate summer-spawned individuals do not subsequently appear in the northwestern Gulf. Presumably, their existence would be reflected as unimodal, or at least not clearly bimodal, length frequencies when adequate data become available for the north central Gulf. Our finding that the few spring-spawned fish disappeared after August at 3-5 mo of age agrees with Hildebrand and Cable (1934) who collected what must have been spring-spawned fish in July (x = 34 mm, range 3-70 mm ?L) and August (x = 54 mm, range 3-77 mm ?L) off North Carolina and noted they were absent later. Because the mag- nitude of spring spawning appears so small, for practical stock assessment purposes, our data could just as well be interpreted as one long period 7Murphy. M. D.. and M. E. Chittenden. Jr. Unpubl. manu- scr. Reproduction, movements, and population dynamics of the gulf butterfish, Peprilus burti. 66 p. Marine Research Laboratory, Florida Department of Natural Resources, 100 Eighth Avenue, S.E.. St. Petersburg. FL 33701. HKelley. F. J., Jr., and R. E. Randall. 1980. Physical oceanography. In R. W. Hann, Jr. and R. E. Randall (editors >. Evaluation of brine disposal from the Bryan Mound .te of the Strategic Petroleum Reserve Program, p. (l-l)-(l-93). National Technical Information Service, Springfield, VA 22150 (DOE/ P010114-1). with little or no spawning from April to August. It seems more meaningful, however, to regard the spawning of L. fasciatus as occurring during two discrete periods because Murphy and Chittenden (footnote 7) suggested a hydrographic basis for that pattern. Our findings are new that fall-spawned groups spawn in both spring and fall periods and that spring spawning is the product offish about 20 mo old. Shlossman and Chittenden (1981) noted that temporal isolation of spawned groups in C. arenarius implied reproductive isolation and might indicate separate populations. The tempor- ally separate spawned groups in L. fasciatus, how- ever, are not reproductively isolated and appar- ently do not form separate populations because the same spawned group spawns in both periods. This simplifies management, because separate data may not be necessary for both spawning periods, especially considering that one is very small. Shlossman and Chittenden (1981) and DeVries and Chittenden (1982) noted that the existence of two spawned groups in C. arenarius and C. nothus might buffer population stability as a multiple year class structure does in longer lived species. In L. fasciatus, however, the contribution that spring-spawned groups make to total population size is probably too small to buffer fluctuations at any reasonably "normal" stock size. Bathymetric Distribution Larimus fasciatus primarily is restricted to the inner continental shelf. Our finding, that they range from <5 to 55 m but are most common from 5 to 16 m, agrees with Hildebrand (1954), Miller (1965), Burns (1970), Milstein and Thomas (1976). and Wenner et al. (1979a, b). Franks et al. (1972) captured most specimens in 37-55 m off Missis- sippi which supports Chittenden and McEachran's (1976) suggestion that the white shrimp commun- ity, of which L. fasciatus is a member, penetrates into deeper water in the north central Gulf than it does in the northwestern area. However, Springer and Bullis (1956) collected fish at 106 m off both Mississippi and Texas. Age Determination and Growth Little literature exists on age determination and growth in L. fasciatus. We determined age by length-frequency analysis because our data came from a long-term set of cruises close enough to- gether in time that modes were easily followed. 359 FISHERY BULLETIN: VOL. 82. NO. 2 Moreover, we observed so little overlap of lengths from different spawned groups that few individu- als could have been incorrectly aged and basic conclusions on apparent growth and mortality would be little affected by such error. As Geoghe- gan and Chittenden (1982) found for Stenotomus caprinus, length-frequency analysis can be a superior method to age L. fasciatus because 1) little spawning occurs in other than one major discrete period each year, 2) length frequencies within spawned groups are reasonably described by a normal distribution, 3) growth of large and small fish within a spawned group appears uni- form since the variance was generally constant between cruises, and 4) life span is short so age determination need be applied only to a few ages, the ideal situation for using length frequencies (Lagler 1956; Tesch 1971). Ross (1978) successfully used scales and otoliths to determine age of North Carolina fish, but we were not able to do so in limited trials and did not pursue these methods further because it seemed unnecessary. Larimus fasciatus reach slightly larger sizes at age off North Carolina than apparent sizes we found in the northwestern Gulf. Von Bertalanffy predictions of size at age off North Carolina (Ross 1978) were 153 mm (121.3 mm SL) at age 1, 188 mm ( 151.1 mm SL) at age II, and 209 mm (168.0 mm SL) at age III compared with quadratic and von Ber- talanffy predictions of 130-150 mm at age I and 155-180 mm at or approaching age II for the northwestern Gulf. Our finding that growth is greatest in the spring and summer agrees with Ross (1978). Maximum Size, Life Span, and Mortality The largest L. fasciatus we found in the north- western Gulf (182 mm) is smaller than most maximum sizes reported from the north central Gulf (Louisiana: 208 mm by Hildebrand 1954, 195 mm ?L by Dunham 1972, and Mississippi: 202 mm by Franks et al. 1972, 189 mm by Christmas and Waller 1973, 187 mm in our data), and much smaller than those reported from the Atlantic coast of the United States (New York to Florida: 220 mm by Wilk and Silverman 1976, Chesapeake Bay: 215 mm ?L by Hildebrand and Schroeder 1928, off North Carolina: 205 mm ?L by Hilde- brand and Schroeder 1928, 206 mm ?L by Hilde- brand and Cable 1934, and 225 mm = 182 mm SL by Ross 1978). The largest record is a 271 mm specimen collected off Mississippi (Franks 1970). The larger size off the Atlantic coast of the United 360 States may reflect greater longevity there, espe- cially from about Cape Lookout or Cape Hatteras north where zoogeographic change in population dynamics may occur (White and Chittenden 1977). Ross (1978) collected age III fish off North Carolina, but the oldest fish we collected only ap- proached age II. The appearance of larger L. fasciatus in the north central Gulf than in the northwestern area follows a pattern apparent in a variety of species (Murphy and Chittenden footnote 7) including C. nothus, P. burti, S. caprinus, Brevoortia patronus, and Micropogonias undulatus. These authors suggested this could reflect 1) small but funda- mental percentage composition and population dynamics differences between these areas, 2) greater biomass at all ages in the north central Gulf, not necessarily population dynamics differ- ences, so that greater numbers of large fish might be captured there even if percentage compositions did not vary, and/or 3) probable permanent emi- gration from the northwestern to the north central Gulf by larger, older, spawning or postspawning fish as they approach age I. They suggested the last explanation applied to C. nothus, P. burti, and probably other fishes, and that it would be man- ifested as between area population dynamics dif- ferences. The following findings also suggest that L. fasciatus too more or less permanently emi- grates from the northwestern to the north central Gulf as spawning and age I approaches 1) the plateaus in length formed in August and seeming cessation of somatic growth in later stages of gonad development, and 2) the appearance in the northwestern Gulf in July 1981 of an abundant fall-spawned group approaching age II. This older spawned group, and parallel spawned groups in other years, was absent or rare in all other months even though our data were based on 71 cruises and 3,390 tows over 4 yr. Because larger, older L. fasciatus probably emigrate to the north central Gulf, the typical maximum life span of 1-2 yr we observed for the northwestern area may be a little low for a stock that ranges over both areas. With the exception of the very large specimen that Franks (1970) found, the largest individuals reported from the north central Gulf are only 189-208 mm as noted. This is not much larger than our largest specimen (182 mm) from the northwestern area, which was 20 mo old. Moreover, these maximums are similar to von Bertalanffy predictions of mean sizes at age II (165-185 mm) or at age III (175-195 mm) that we found, and sizes of 188 and 204 mm Ross (1978) STANDARD and CHITTENDEN: REPRODUCTION. MOVEMENT, AND POPULATION OK BANDED DRUM found at ages II and III, respectively. Therefore, a tL value of 2-3 yr may be realistic for a stock that ranges over the north central and northwestern Gulf. We assume in suggesting this, that differ- ences in size compositions do not reflect only possi- ble intra-Gulf growth differences and greater size at age in the north central Gulf. The mean apparent time-specific and cohort- specific total annual mortality rates we observed for the northwestern Gulf (92-100% ) agree with theoretical estimates (Royce 1972:238) of 90-100% if maximum life span typically is only 1-2 yr as we found for that area. Because larger, older L. fas- ciatus probably emigrate to the north central Gulf, our observed mortality estimates are probably too high for a stock that ranges over both areas. Theoretical values of 80-90% based on a 2-3 yr typical maximum life span may be more realistic, a magnitude which agrees with the lowest values tenable for other sympatric species such as C. arenarius (Shlossman and Chittenden 1981), C. nothus (DeVries and Chittenden 1982), S. cap- rinus (Geoghegan and Chittenden 1982), and P. burti iMurphy 1981). Even values of 80-90% are higher than the three lowest mortality rates we found for the northwestern Gulf (55-74%) and rates of 57 and 81% that Ross (1978) reported off North Carolina; the latter range of values is theoretically appropriate as an average over life spans of 3-5 yr, although present data suggest 4-5 yr is too large a value of tL for the Gulf. General Population dynamics of L. fasciatus are similar to those reported from the northwestern Gulf for M. undulatus (White and Chittenden 1977), C. arenarius (Shlossman and Chittenden 1981), C. nothus (DeVries and Chittenden 1982), S. cap- rinus (Geoghegan and Chittenden 1982), P. burti (Murphy 1981; Murphy and Chittenden footnote 7), and in Centropristis philadelphica ignoring its hermaphroditism (Ross and Chittenden9). Our findings support the suggestions that 1) ground- fishes of the white and brown shrimp communities in the Gulf have evolved a common pattern of population dynamics characterized by small size, early age at maturity, short life spans, high mor- 9Ross, J. L., and M. E. Chittenden, Jr. Unpubl. manu- scr. Reproduction, movements, and population dynamics of the rock sea bass, Centropristis philadelphica, in the northwestern Gulf of Mexico. 50 p. North Carolina Division of Marine Fisheries, North Carolina Marine Resources Center, Manteo, NC 27954. tality rates, and rapid turnover of biomass (Chit- tenden and McEachran 1976; Chittenden 1977), and 2) more or less permanent spawning or post- spawning emigration may occur from the north- western Gulf to the north central area as fish ap- proach age I (Murphy and Chittenden footnote 7). Because typical maximum life spans may be closer to 2-3 yr than 1-2 yr, these fishes may be a little more sensitive to growth overfishing than Chit- tenden's (1977) simulations, based on a 2 yr life span, suggest for M. undulatus. ACKNOWLEDGMENTS We would like to thank R. Baker, M. Burton, T Crawford, D. DeVries, V Fay, P. Geoghegan, R. Grobe, S. Harding, M. Murphy, J. Pavela, M. Rockett, J. Ross, P. Shlossman, B. Slingerland, H. Yette, and Captains H. Forrester, M. Forrester, R. Forrester, R Smirch, and A. Smircic for assistance in field collections and data recording. R. Darnell, J. McEachran, and K. Strawn reviewed the man- uscript. B. Rohr made it possible to use data from the NMFS groundfish survey 120. R. Case, M. Cuenco, and J. Cummings wrote and assisted with computer programs. Financial support was pro- vided by the Texas Agricultural Experiment Sta- tion; by the Strategic Petroleum Reserve Program, Department of Energy; and by the Texas A&M University Sea Grant College Program, supported by the NOAA Office of Sea Grant, Department of Commerce. LITERATURE CITED ALVERSON, D. L.. AND M. J. CARNEY. 1975. A graphic review of the growth and decay of popula- tion cohorts. J. Cons. Int. Explor. Mer 36:133-143. BAGENAL, T. B. 1957. The breeding and fecundity of the long rough dab Hippoglossoides platessoides (Fabr. > and the associated cycle in condition. J. Mar. Biol. Assoc. U.K. 36:339- 375. BAGENAL. T. B., AND E. BRAUM. 1971. Eggs and early life history. 2d ed. In W. E. Ricker (editor), Methods for assessment of fish production in fresh waters, p. 166-198. IBP (Int. Biol. Programme) Handb. 3. BEHRE. E. H. 1950. Annotated list of the fauna of the Grand Isle Region, 1928-46. Occas. Pap. Mar. Lab.. La. State Univ. 6:1-66. BEVERTON. R. J. H.. AND S. J. HOLT. 1957. On the dynamics of exploited fish populations. Fish. Invest. Minist. Agric, Fish. Food (G.B.i. Ser. II. 19:1-533. BR1GGS. J. C. 1958. A list of Florida fishes and their distribution. Bull. Fla. St. Mus. 2:223-318. 361 FISHERY BULLETIN: VOL. 82, NO. 2 BURNS, C. 1970. Fishes rarely caught in shrimp trawls. Gulf Res. Rep. 3:110-130. CHAO, L. N.. AND J. A. MUSICK. 1977. Life history, feeding habits, and functional morphol- ogy of juvenile sciaenid fishes in the York River estuary, Virginia. Fish. Bull., U.S. 75:657-702. CHITTENDEN, M. E., JR. 1977. Simulations of the effects of fishing on the Atlantic croaker. Micropogon undulatus. Proc. Gulf Caribb. Fish. Inst. 29th Annu. Sess., p. 68-86. CHITTENDEN, M. E., JR., AND J. D. MCEACHRAN. 1976. Composition, ecology, and dynamics of demersal fish communities on the northwestern Gulf of Mexico conti- nental shelf, with a similar synopsis for the entire Gulf. Sea Grant Publ. No. TAMU-SG-76-208, 104 p. Christmas, J. Y., and R. S. Waller. 1973. Estuarine vertebrates, Mississippi. In J. Y. Christmas (editor). Cooperative Gulf of Mexico estuarine inventory and study, Mississippi, p. 320-434. Gulf Coast Res. Lab., Ocean Springs, Miss. DAHLBERG, M. D. 1972. An ecological study of Georgia coastal fishes. Fish. Bull., U.S. 70:323-353. DEVRIES, D. A., AND M. E. CHITTENDEN, JR. 1982. Spawning, age determination, longevity, and mor- tality of the silver seatrout, Cynoscion nothus, in the Gulf of Mexico. Fish. Bull., U.S. 80:487-500. DUNHAM, F. 1972. A study of commercially important estuarine- dependent industrial fishes. La. Wildl. Fish. Comm., Tech. Bull. 4:1-63. FABENS, A.J. 1965. Properties and fitting of the von Bertalanffy growth curve. Growth 29:265-289. Franks, J. S. 1970. An investigation of the fish population within the inland waters of Horn Island, Mississippi, a barrier island in the northern Gulf of Mexico. Gulf Res. Rep. 3:3-104. Franks, J. S., J. Y. Christmas, W. L. Siler, R. Combs, R. WALLER, AND C. BURNS. 1972. A study of nektonic and benthic faunas of the shal- low Gulf of Mexico off the state of Mississippi as related to some physical, chemical and geological factors. Gulf Res. Rep. 4:1-148. Geoghegan, p., and M. E. Chittenden, Jr. 1982 . Reproduction , movements, and population dynamics of the longspine porgy Stenotomus caprinus. Fish. Bull., U.S. 80:523-540. GULLAND, J. A. 1969. Manual of methods for fish stock assessment. Part 1: Fish population analysis. FAO Man. Fish. Sci. 4:1-154. GUNTER, G. 1938. The relative numbers of species of marine fish on the Louisiana coast. Am. Nat. 72:77-83. 1945. Studies on marine fishes of Texas. Publ. Inst. Mar. Sci. Univ. Tex. 1:1-190. Harding, J. P 1949. The use of probability paper for the graphical analy- sis of polymodal frequency distributions. J. Mar. Biol. Assoc. U.K. 28:141-153. HELWIG, J. T, AND K. A. COUNCIL. 1979. SAS user's guide, 1979 ed. SAS Institute Inc., Raleigh, N.C., 494 p. HILDEBRAND, H. H. 1954. A study of the fauna of the brown shrimp (Penaeus aztecus Ives) grounds in the western Gulf of Mexico. Publ. Inst. Mar. Sci. Univ. Tex. 3:233-366. HILDEBRAND, S. F, AND L. E. CABLE. 1934. Reproduction and development of whitings or king- fishes, drums, spot, croaker, and weakfishes or seatrouts, family Sciaenidae, of the Atlantic coast of the United States. Bull. U.S. Bur. Fish. 48:41-117. HILDEBRAND, S. F, AND W C. SCHROEDER. 1928. Fishes of Chesapeake Bay. Bull. U.S. Bur. Fish. 43:1-366. HOESE, H. D. 1965. Spawning of marine fishes in the Port Aransas, Texas area as determined by the distribution of young and larvae. Ph.D. Thesis, Univ. Texas, Austin, 150 p. JOHNSON, G. D. 1978. Development of fishes of the mid-Atlantic Bight: an atlas of egg, larval and juvenile stages, Vol. IV Carangidae through Ephippidae. U.S. Fish Wildl. Serv. Biol. Serv. Program FWS/OBS-78/12, 314 p. LAGLER, K. F. 1956. Freshwater fishery biology. 2d ed. Wm. C. Brown Co., Dubuque, Iowa, 421 p. MILLER, G. L., AND S. C. JORGENSON. 1969. Seasonal abundance and length frequency distribu- tion of some marine fishes in coastal Georgia. U.S. Fish Wildl. Serv., Data Rep. 35, 102 p. MILLER, J. M. 1965. A trawl survey of the shallow Gulf fishes near Port Aransas, Texas. Publ. Inst. Mar. Sci. Univ. Tex. 10:80- 107. MILSTEIN, C. B., AND D. L. THOMAS. 1976. Fishes new or uncommon to the New Jersey coast. Chesapeake Sci. 17:198-204. MURPHY, M. D. 1981. Aspects of the life history of the gulf butterfish, Pep- rilus burti. M.S. Thesis, Texas A&M Univ., College Sta- tion, 77 p. POWLES, H. 1980. Descriptions of larval silver perch, Bairdiella chrysoura, banded drum, Larimus fasciatus, and star drum, Stellifer lanceolatus (Sciaenidae). Fish. Bull., U.S. 78:119-136. RICKER, W E. 1973. Linear regressions in fishery research. J. Fish. Res. Board Can. 30:409-434. 1975. Computation and interpretation of biological statis- tics offish populations. Fish. Res. Board Can., Bull. 191, 382 p. ROSS, S. W 1978. The life history of the banded drum, Larimus fas- ciatus, in North Carolina waters. M.A. Thesis, Univ. North Carolina, Chapel Hill, 143 p. ROYCE, W F. 1972. Introduction to the fishery sciences. Acad. Press, N.Y., 351 p. SHLOSSMAN, P A., AND M. E. CHITTENDEN, JR. 1981. Reproduction, movements, and population dynamics of the sand seatrout, Cynoscion arenarius. Fish. Bull., U.S. 79:649-669. SIMPSON, A. C. 1959. Method used for separating and counting the eggs in fecundity studies on the plaice (Pleuronectes platessa ) and herring (Clupea harengus). FAO Indo-Pac. Fish. Counc, Occas. Pap. 59/12, 3 p. Snedecor, g. W, and W G. Cochran. 1980. Statistical methods. 7th ed. Iowa State Univ. 362 STANDARD and CHI ITKNDEN: REPRODUCTION, MOVEMENT, AND POPULATION OF BANDED DRUM Press, Ames, Iowa, 507 p. SPRINGER, S , AND H. R. BULLIS, JR. 1956. Collections by the Oregon in the Gull of Mexico. U.S. Fish Wildl. Serv, Spec. Sci. Rep. Fish. 196, 134 p. STANDARD, G. W. 1983. Life history and population dynamics of the banded drum, Larimus fasciatus, in the northwestern Gulf of Mexico. M.S. Thesis, Texas A&M Univ., College Station, 89 p. SWINGLE, H. A. 1971. Biology of Alabama estuarine areas - cooperative Gulf of Mexico estuarine inventory. Ala. Mar. Resour. Bull. 5, 123 p. TESCH, F W. 1971. Age and growth. 2d ed. //; W. E. Ricker (editor). Methods for assessment offish production in fresh waters, p. 98-130. Blackwell Sci. Publ., Oxf. WENNER, C. A., C. A. BARANS. B. W STENDER. AND F. H. BERRY. 1979a. Results of MARMAP otter trawl investigations in the South Atlantic Bight. I. Fall, 1973. S.C. Mar. Res. Cen.,Tech. Rep. 33, 79 p. 1979b. Results of MARMAP otter trawl investigations in the South Atlantic Bight. II. Spring, 1974. S.C. Mar. Res. Cen.,Tech. Rep. 40, 78 p. WHITE, M. L., AND M. E. CHITTENDEN, JR. 1977. Age determination, reproduction, and population dynamics of the Atlantic croaker, Micropogonias un- dulatus. Fish. Bull., U.S. 75:109-123. WILK, S. J., AND M. J. SILVERMAN. 1976. Fish and hydrographic collections made by the re- search vessels Dolphin and Delaware II during 1968-1972 from New York to Florida. U.S. Dep. Coramer., NOAA Tech. Rep. NMFS SSRF 697, 159 p. 363 IMPLICATIONS OF INVESTING UNDER DIFFERENT ECONOMIC CONDITIONS ON THE PROFITABILITY OF GULF OF MEXICO SHRIMP VESSELS OPERATING OUT OF TEXAS1 Ernest Tettey, Christopher Pardy, Wade Griffin, and A. Nelson Swartz2 ABSTRACT Due to the inflationary trend in recent years coupled with fluctuating shrimp prices, the shrimp business has become a highly uncertain undertaking. The financial performance of a sample of the Gulf of Mexico shrimping fleet, operating out of the Texas coast, was examined over a 10-year period (1971-80). The results indicate that investments made in the early part of the 1970s performed better than those made in the latter part. Periods of low inflationary levels appeared to be more favorable to investments in the shrimp fishery than periods of high inflationary levels. In terms of economic profits, steel vessels generally did better than wooden ones. Medium-sized vessels (18.6-20.0 m in overall length) were the most efficient vessels to operate in the Gulf of Mexico. The Gulf of Mexico supplies a major share of the shrimp landed by commercial shrimp producers in the United States. From 1977 to 1981, Gulf shrimp landings accounted for 62% of the U.S. total. In 1981, 161 million kg of commercial shrimp valued at $463.4 million are landed in the United States. The Gulf of Mexico accounted for 76% of these landings and 87% of the value. Although the Gulf shrimp fishery is the most valuable in the United States, individual harvesters within the industry are not without their financial problems. Of late, the high variability in shrimp landings and prices has created short-run uncertainty among shrimp producers (Caillouet and Patella 1978; Warren and Griffin 1980). Coupled with this, operating costs have been significantly increasing over the years, to the extent that, it has become quite difficult for fishermen to stay in business (Griffin et al. 1978). There have been several costs and returns and/or investment analyses conducted on fishing vessels in recent years (Gates and D'Eugenio 1975; Noetzel 1977; Jones et al. 1979; Roberts and Saas 1979; Prochaska and Cato 1981); however, none have been concerned with the effect of inflation on investing in a fishing vessel. This paper uses the period 1971-80 to draw conclusions about the effect of low, medium, and high inflationary periods on return to investment. The study further examines 'Technical Article No. 18678 of the Texas Agricultural Exper- iment Station, Texas A&M University, College Station, Tex. 2Department of Agricultural Economics, Texas A&M Univer- sity, College Station, TX 77843. Manuscript accepted October 1983. FISHERY BULLETIN: VOL. 82, NO. 2, 1984. the implication of unstable shrimp prices and ris- ing costs of operations on the profitability of the shrimp industry in the Gulf of Mexico. Finally, the performance of wooden and steel hulled vessels in various size classes is compared. METHODS Data Description The data used in this study are an accumulation of 5 yr of data collection, which have been reported in previous publications (U.S. Department of Commerce 1971-1980, 1971-1981; Griffin et al. 1974, 1976; Griffin and Nichols 1976; Warren and Griffin 1978). Although data were collected for other Gulf states, only data for vessels operating out of Texas are used in this study, since it is the only state for which data were available for all 5 yr that data were collected. In the original studies, data were collected by personal interview in ports from Galveston to Port Isabel for 1971, 1973, 1974, 1975, and 1977 and estimated for the remaining years. Additional in- formation was obtained from officials of various lending institutions which engage in shrimp ves- sel financing, from boat builders, and from the National Marine Fisheries Service. Cost The variable cost items for which data were gathered included ice, fuel, nets, supplies, repairs 365 FISHERY BULLETIN: VOL. 82, NO. 2 and maintenance, crew shares, payroll taxes, and packing charges. Fuel consumption by vessel class, for which primary data were unavailable, was estimated according to the following relation- ship: ECn = WC; (TLITLj) * (*<,•*«,•) where Y, X„ = N„ = IN estimated number of gallons ( 1 gal = 3.79 1) used per year by vessel class i. actual number of gallons (1 gal = 3.79 1) used in year; by vessel class i. number of vessels in vessel class i in yearj. The nets, supplies, and repairs and mainte- nance variables were similarly adjusted like Y- to account for unequal numbers of vessels in the sampling years. These variables were deflated into real terms, using the industrial price index, and the weighted average was determined. From the weighted average of the 5 yr for which data were available, adjusted nominal values were deter- mined for the other years. Crew shares were esti- mated at 33% of value of catch. Packing charge was set at $0.10 per pound (0.454 kg) of shrimp landings. Fixed costs include insurance, depreciation, interest, and opportunity cost. Reported data were used to determine fixed charges for over- head items, while charges relating directly to investment — depreciation, insurance, and interest — were calculated in nominal dollars for new vessels. Overhead values for years that data were not available were calculated using the in- dustrial price index in the same manner as with variable cost. Insurance charges were set at 47c of new vessel cost. The standard straight-line formula was used to determine depreciation. In this study the ter- minal value of the vessel was calculated in two ways: 1) at 100% of original cost, and 2) at salvage value at the end of 1980. The market rates, which prevailed over the years, were employed in deriv- ing the cost and returns budget. where WC , RCy EC, 7 NU RC, 1 j N, and TL} = TL = RCu = ACu = N;; = ACyKTLi/TL). estimated catch by vessel class i for year j where j = 1971, 1972, ...,1980. Texas landing's for year j, = 1971,1972, ...,1980. " average Texas landings for the 1971-80 period. real catch of vessel class i in year j for the 5 yr vessel data were available. actual catch of vessel class i for year j for the 5 yr vessel data were available. number of vessels of class i for year j for the 5 yr vessel data were available. Exvessel3 prices per pound of shrimp were ad- justed according to the formula given below using the average value from the National Marine Fisheries Service data for Texas and that from the survey to generate exvessel prices for those 5 yr that data were not collected. AP, {A, - B) + TP/ where AP, TP, AI B = adjusted exvessel price per pound (0.454 kg) of shrimp for vessel i in yearj. = Texas prices as reported by the National Marine Fisheries Service for year j, where j = 1971, ...,1980. = average exvessel price per pound (0.454 kg) of shrimp for vessel i over the 5 yr data were available. = average exvessel price per pound (0.454 kg) of shrimp Revenue Catch relationships were estimated for years for which no data were available by utilizing the fol- lowing formulation: 3It is recognized that exvessel price of shrimp is greatly influ- enced by seasonal fluctuation in local supply as well as the size composition of catch (Caillouet and Patella 1981). However, this study implicitly accounts for such trends by employing primary data collected for 1971, 1973, 1974, 1975, and 1977. Therefore, the price represents a weighted average between the various shrimp sizes. 366 TETTEY ET AL.: PROFITABILITY OF GULF OF MEXICO SHRIMP VESSELS reported by the National Marine Fisheries Service. Analyses A computer program referred to as a budget generator was devised to organize and assimilate the data for various analyses. The program al- lowed data reports to be produced according to the desired vessel classifications, interest rate, per- cent financed, number of years financed, number of loan payments per year, depreciation method, crew share agreement, rate of packing charges, payroll tax rate, discount rate, and planning hori- zon. The program reported results in the form of annual costs and returns budgets and projected cash flow budgets. The following analysis first examines a detailed annual income and cash flow statement for a ves- sel purchased new in 1971. This detailed annual income and cash flow is then compared for the TABLE 1. — Number and types of Gulf of Mexico shrimp vessels surveyed operating out of Texas ports. Vessel type1 1971 1973 1974 1975 1977 Wooden, 17.1-18.5 m 1 1 3 3 1 Wooden, 18.6-20.0 m 1 9 8 4 1 Wooden, 20.1-21.5 m 1 26 24 26 4 Steel, 18.6-20.0 m 0 14 19 21 17 Steel, 20.1-21.5 m 3 13 41 41 18 Steel 21 .6-23.0 m 2 4 10 5 2 1 Coast Guard registered length same vessel operated under identical conditions, but purchased in 1977 and 1979. Next, six different types of vessels (three wood and three steel; Table 1) are compared by examining their net returns during the three different periods of investment. Finally, investment performance is analyzed through net present value (NPV). In all the above analyses three investment periods (1971, 1977, 1979) are considered. A given vessel is assumed to be operated under identical conditions regardless of the investment period. Since the actual sale price of the vessel at the end of the investment period is determined by the economic environment at that time, a comparison is made between the effects of selling the vessel at a salvage value of 35 and 100% of the original price. RESULTS Detailed Annual Budgets and Cash Flow Tables 2 and 3 represent detailed annual income and cash flow budgets for a newly financed 1971 steel vessel, 20.1-21.5 m in overall length. Over the 10-yr study period, annual revenue doubled al- though it decreased by 18.9% in 1974 and by 7.4% in 1980. The decrease could be attributed, in part, to the decrease in exvessel price for shrimp from $4.23/kg to $3.62 in 1974 and from $8.38 to $7.06 in 1980. Another contributing factor was poor landings recorded in those periods (Fig. 1). TABLE 2. — Annual income statement for a steel vessel, 20.1-21.5 m. long, operating out of Texas ports, 1971 to 1980. 1971 1972 1973 1974 1975 1976 1977 1978 1979 19801 19802 Revenues 78.100 95,798 99,903 80,978 99,498 132.799 142.272 155.733 164.638 152,479 152,479 Variable costs Ice 1,502 2,062 2,052 1,656 1.903 2,169 2.859 4,371 4.002 4.105 4.105 Fuel 6.515 6,515 9,523 20,549 21,551 22.052 23.055 25,060 40,095 47,613 47.613 Nets, supplies, groceries 8,348 8,186 9.220 11,263 12,559 13,357 14,287 15,334 17.319 20,028 20,028 Repair and maintenance 5.563 5.455 6,144 7,505 8.369 8.901 9,521 10,219 11,541 13,347 13,347 Crew shares 25.773 31.613 32.968 26,723 32,834 43,824 46,950 51,392 54.331 50,318 50.318 Packing 2,580 2.899 2,440 2,679 3.003 80,219 3,178 93,481 3,891 3,586 109.962 2,873 3.143 3,143 Total 50,281 56,730 62,347 70.375 100.563 130,161 138,554 138.554 Fixed costs Depreciation 9.084 9.084 9,084 9,084 9,084 9,084 9,084 9.084 9,084 9,084 9,084 Insurance 5,191 5.191 5,191 5,191 5,191 5,191 5,191 5,191 5.191 5,191 5,191 Interest (vessel loan)3 6,910 6,412 5.879 5,308 4,697 4.042 3.340 2,589 1,785 923 923 Overhead 3.423 3,357 3,781 4,619 5,150 5,477 23,794 5.859 6.288 7,102 8,213 8,213 Total 24,608 24,044 23,935 24,202 24,122 23.474 23,152 23,162 23,411 23,411 Total operating costs 74,889 80.774 86,282 94,576 104,341 117,275 124,037 133,114 153,323 161,965 161,965 Net revenue 3.211 15.024 13.621 -13,599 -4.843 15.524 18,235 22,619 11,315 -9.487 -9,487 Net return after tax" 266 9,405 8,158 -17,106 -8,672 - 8,916 11,140 14,377 5,100 -15,383 45,273 Current equity 33.611 39.101 42,531 28,724 19,718 32,096 41,324 43,328 39.530 43,282 73,609 Required return to equity 3,946 4,391 4,636 3,326 2,686 4,250 4,992 5,273 5,190 5,376 9,142 Economic profit -3,680 5,014 3,522 -20,433 -11.357 4,667 6,148 9,104 -90 -20,758 36,130 'Vessel sold for salvage value. 2Vessel sold for original purchase price. 3Vessel was purchased for $129,767; 75% financed at 7.1% "The difference between net revenue and net return after tax interest. es includes owner s salary, social security tax for owner, and income tax. 367 FISHERY BULLETIN: VOL. 82, NO. 2 TABLE 3. — Annual cash flow statement for a steel vessel, 20.1-21.5 m. long, operating out of Texas ports, 1971 to 1980. 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980' 22.052 23.055 25.060 71,429 77.507 84,901 20.205 18,145 19,722 40.095 90.065 18.510 47,613 90.942 19.300 70.896 7.509 6,412 84.817 76,783 83.692 8.042 5,879 90,704 19802 29.011 29,011 152,479 152,479 38,930 129.767 98.137 113,222 103.496 105,381 132,799 147.463 170,568 191.602 220.420 311,257 47.613 90.942 49.481 Revenues Beginning cash balance 0 2.339 13,319 22,518 5,883 0 5.192 14.835 26,964 Receipt from shrimp 78,100 95.798 99,903 80,978 99,498 132,799 142,272 155,733 164,638 Capital receipts 000000000 Total cash inflow 78,100 Operating expenses Fuel ' 6.515 6,515 9,523 20,549 21,551 Other variable expenses 43,766 50,216 52,824 49,826 58,669 Fixed cash expenses 11.559 14,165 14.436 13,317 14,169 Total 61.840 Long- and short-term debt Long-term debt (principle) 7.01 1 Long-term debt (interest) 6,910 Total cash outflow 75.761 Cash situation Net cash balance 2,339 10.980 9.199 -16,635 -8.812 8,121 9,643 12.129 2,047 Cash available 2,339 13.320 22,518 5,883 -2,929 5,192 14,835 26.964 29.011 Ending cash balance 2,339 13.320 22,518 5,883 0 5,192 14,835 26,964 29.011 Net present value 0. 1 1 74 94,389 113,686 118.707 129,683 148.670 157,855 188,036 9,224 9,879 10.581 1 1 ,332 12.136 12,998 12.998 4,697 4,042 3.340 2.589 1,785 923 923 8,613 5,308 97,613 108.310 127.607 132,628 143,604 162,591 171,776 201,957 19.633 80.288 48.644 109.300 48,644 109.300 22,572 42,560 'Vessel sold for salvage value 2Vessel sold for original price 18CM K 140- G S r N 120 D D • L 103 L fi R 5 80 I N T H 60 0 U S R N 40- D S I97B 1979 1980 TERRS LEGEND: OPTIONS LANDINGS -VALUE FIGURE 1. — Total Texas landings and value of landings for a typical steel vessel 20.1 to 21.5 m long by year. There was no significant variation in total fixed costs during the 10 yr for two reasons: Deprecia- tion and insurance charges were set at fixed levels, but while overhead charges increased gradually over the years, interest payments on vessel loans decreased by about the same margin. The highest total fixed cost incurred was $24,608 in 1971 and the lowest was $23,152 in 1978 (Table 2). Total variable costs, on the other hand, in- creased by 175.5% over the 10-yr period. The high- est increase of 18.4% occurred in 1979, and it was in direct response to an increase in fuel charges from $25,060 in 1978 to $40,095 in 1979 (Table 3). Expenditure on fuel showed a tremendous in- crease in 1974 by 115.7%; both of these increases were due to the sharp rise in fuel prices (Fig. 2). In response to an increase in variable costs and relatively constant fixed costs, total operating 6- 0 5 0 L L R R 4 S 3- 0 l 1970 1971 1972 1973 1974 1975 1976 19"' 1978 1979 1380 TERRS LEGEND: 3PT13NS FUEL -- SHRIMP FIGURE 2. — Price of fuel and shrimp for a steel vessel 20.1 to 21.5 m long by year operating out of Texas ports. 368 TETTEY ET AL.: PROFITABILITY OF GULF OF MEXICO SHRIMP VESSELS costs more than doubled over the period 1971-80, from $74,889 to $161,965 (Table 2). The highest net revenue before taxes recorded was $22,619 in 1978. Losses were recorded in 1974, 1975, and 1980, and this could be due to the effects of inflation and poor harvest recorded in those years. Terminal value has a critical effect on profitabil- ity. Decreased landings led to a substantial economic loss of $20,758 in 1980 (Table 2), when vessels were salvaged at 359£ of their original val- ues. Increasing the terminal value to 100% of cost resulted in a positive economic profit of $36,130, the highest obtained (Table 2). It was further ob- served that the net present value for the invest- ment project increased by 72.5% when vessels were salvaged at their original costs rather than at the 35% level of their original values. Investment in Different Time Periods Tables 2 and 3 present the results of investing in a shrimp vessel in 1971. Considerable investment has been made in new shrimp vessels since 1977. Table 4 shows the annual increase and cash flow statements for purchasing the same steel vessel in 1977. Since the level of operations is held constant, revenue and variable costs are the same; fixed costs, however, changed dramatically. In 1971, the value of the 20.1-21.5 m steel vessel was $129,767. In 1977, that same vessel cost $222,084, a 71.1% increase in price. The loan pay- ment in 1971 was a little over $ll,600/yr assuming a 10 yr note, but increased to $24,074/yr. If the vessel was purchased in 1971, profits were made in 1977 and 1978; for the vessel purchased in 1977, losses were incurred over the entire investment period. Losses were particularly substantial if the vessel was sold for salvage value at the end of 1980. The cash flow statement shows that cash avail- able was very low for 1977-79. There was about $35,000 difference in cash available in 1980, de- pending on whether the vessel had been sold for salvage value or for its original purchase price. For the vessel purchased in 1971, the net cash increase in 1977 was $9,643; and only $2,223 for the vessel purchased in 1977. In 1979 under the 1971 scenario, cash available would have increased $2,047; here it declined $6,252. Despite poor economic conditions, NPV was positive for both vessel sale prices, implying vessels would have had a greater return on investment than bonds pur- chased in 1977. Table 5 shows the results of purchasing a vessel in 1979. The price of the vessel went up by about 26.8% since 1977, and substantial losses were in- curred. The NPV is negative if the vessel is sold for its salvage value and positive if it is sold for 100% of its original purchase price. Short-term borrow- ing occurred when the vessel was purchased in 1979. In fact, the vessel owner had to borrow more than he was paying in principal on his original purchase note. Economic Performance by Size and Construction Figures 3 through 8 show variations in net rev- enue for the various vessel types used in the analy- sis. These variations follow a general pattern. For vessels purchased in 1971, net revenue peaked in 1972 and 1978 and dropped to minimum levels in 1974 and 1980. A major reason for this trend is that TABLE 4. — Summarized annual income and cash flow state- ments for steel vessels, 20.1-21.5 m long, operating out of Texas ports, 1977 to 1980. 1977 1978 1979 1980' 19802 Income statement Value of landings 142,272 155,733 164.638 152,479 152,479 Total variable cost 100,562 109,961 130.160 138,555 138,555 Total fixed cost3 43.497 43.011 42.838 42,883 42,883 Net revenue -1,787 2,761 -8,359 -28,959 -28,959 Economic profit4 -8,628 -5,006 -15.587 -43.663 -158 Cash flow statement Total cash inflow 142,272 157,956 172,165 313,654 375,838 Total cash outflow 140.049 150.429 170,890 180.396 187.041 Net cash balance 2.223 5,304 -6,252 131,983 187,522 Cash available 2,223 7,527 1.275 133,258 188,797 Net present value - 12,759 47,954 1 Salvage value set at 35% of original cost. 2Vessel sold for original purchase price, 3Vessel was purchased for $222,084; 75°o financed at 7 93% interest. 4Economic profit is the net revenue adjusted for any changes in the value of operating inventories and capital items. TABLE 5. — Summarized annual income and cash flow statements for steel vessels, 20.1-21.5 m long, operating out of Texas ports, 1979 to 1980. 1979 1980' 19802 Income statement Value of landings 164.638 152,479 152,479 Total variable cost 130.160 138.555 138,555 Total fixed cost3 61,449 63,668 63.668 Net revenue -26,971 -49,745 -49,745 Economic profit4 -34,528 -68,559 -45.932 Cash flow statement Total cash inflow 164,638 383,999 421.689 Total cash outflow 184,132 215.631 218,132 Net cash balance -19,493 187,861 223,049 Cash available -19,493 168,368 203.556 Net present value -9,355 18.140 'Vessel sold for salvage value. 2Vessel sold for original purchase price. 3Vessel was purchased for $269.21 0; 75% financed at 1 2.25% interest. "Economic profit is the net revenue adjusted for any changes in the value of operating inventories and capital items. 369 FISHERY BULLETIN: VOL. 82, NO. 2 -21 I N V I S T H E N T -2U- P E R -27 I a o -33- -3B-1 T ■ ■ 1971 1972 1973 197M Investment Period: 1975 1976 TIME ITERRS1 1971 TO 1980 ♦ 1977 TO 1980 » 1979 TO 1980 ZERO LINE fiDDED FOR REFERENCE FIGURE 3. — Net revenue for a wooden vessel 17.1 to 18.5 m long operating out of Texas ports by investment period. while tremendous increases in the value of land- ings were recorded in 1972 and 1978, 1974 and 1980 were the only periods in which the value of land- ings actually decreased. It can further be shown that fuel charges increased dramatically in 1974 and 1980 ( Fig. 2 ), making those years particularly bad ones for Texas shrimp producers in terms of net revenue. In general, steel vessels performed better economically than wooden vessels irrespective of the investment period. This may be attributed to the durability of steel vessels and their ability to operate under more adverse weather conditions than wooden vessels. The performance of 18.6-20.0 m steel vessels was particularly outstanding (Fig. 6) while the same size wooden types performed very poorly, recording losses throughout the study period (Fig. 4). As explained earlier, steel vessels are more durable and can spend more days offshore fishing than wooden vessels. Besides, steel vessels generally call for less maintenance and repair costs and attract a better quality crew than wooden ones. With the exception of the 17.1- 18.5 and 18.6-20.0 m wooden vessels, which re- corded losses over most of the period, all the other vessels performed satisfactorily. Variation in Net Present Value Evaluating the investments based on the net present value criterion, the 18.6-20.0 m steel ves- sels would be ranked as the best investments in the Gulf shrimp fishery (Table 6). Compared with the other vessel types, they consistently showed the highest net present values under all invest- ment conditions examined. At the other end of the continuum lie the 18.6-20.0 m wooden vessels, which showed the poorest net present values under each investment condition, actually show- ing negative net present values and implying that investing in 18.6-20.0 m wooden vessels is not a feasible endeavor (Table 6). With the exception of 18.6-20.0 m wooden ves- sels, 1971 investments showed the highest net present values, followed by those made in 1977. Investments made in 1979 were the least feasible; this may be attributed to unusually high capital 1971 1972 1973 197U 1975 1976 Investment Period: TIME (TEfiRSi ■ ■ ■ i 1979 1980 1971 TO 1980 » 1977 TO 1980 » 1979 TO 1980 ZERO LINE fiDDED FOR REFERENCE FIGURE 4. — Net revenue for a wooden vessel 18.6 to 20.0 m long operating out of Texas ports by investment period. 370 TETTEY ET AL.: PROFITABILITY OF GULF OF MEXICO SHRIMP VESSELS and high vessel costs, resulting in higher annual principal and interest payments. SUMMARY AND CONCLUSIONS Due to the inflationary trend in recent years, investments made in the early part of the last decade performed better than those made in the latter part. Steel vessels generally showed higher economic profits than wooden ones, and medium- sized vessels (18.6-20.0 m in overall length) were the most efficient vessels to operate in the Gulf of Mexico. Shrimp production in the Gulf of Mexico has 15- 10- 5- N E R E V -5- E N U E -10- \ / +'' .-+ $ \ 0 -15- 0 0 V \ B -20- T \ \ 1 N -25- V e : 5 T -30- M E N T -35- 1 \ \ \ \ I \ \ . \ \ \ p ; E R -110- I 1 0 D -«5- \ * 1 1 V V \ \ -50- \ \ I -55- 1971 1972 1973 197M Investment Period: 1975 1976 TIME (TERRS) 1971 TO 1980 ♦ 1977 TO 1980 ' 1979 TO 1980 ZERO LINE RDDEO FOR REFERENCE FIGURE 5. — Net revenue for a wooden vessel 20.1 to 21.5 m long operating out of Texas ports by investment period. -50- 1971 1972 1973 1974 1975 Investment Period: time iyer^si 1971 TO 1SE • 1977 TO 1980 m 1979 TO 1980 ZERO LINE RODEO FOR REFEF I FIGURE 6. — Net revenue for a steel vessel 18.6 to 20.0 m long operating out of Texas ports by investment period. significant seasonal variations — periods of low shrimp landings and periods of abundant catch. Although the general trend in real prices was upward, high variability has created short-run uncertainty for shrimp producers. In general, after-tax net revenue was lowest in 1974; 1978 was the most favorable year of operation. Vessel terminal value plays a major role in de- termining overall returns to investment. The NPV for the investment increased when the terminal value rose from 35 to 100% of the original vessel cost (Table 6). Based on the net present value criterion, the 18.6-20.0 m steel vessels once again proved to be the most feasible investment. Inves- TABLE 6. — Net present value for each investment period, salvage value, and vessel type for vessels operating out of Texas ports Investment period: 1971 to 1980 1977 to 1980 1979 to 1980 Salvage value: 35% 100% 35% 100% 35% 100% Vessel type Wooden. 17 1-18.5 m -14,303 4,113 -16,166 2,019 -17,766 4,122 Wooden, 18.6-20.0 m -70.622 -37,348 -49,898 -30,731 -32,431 -17,612 Wooden, 20 1-21 5 m 4,993 23,089 -18.368 6,834 -27,829 -9,667 Steel, 18.6-20.0 m 89,154 106,585 51,483 84,543 16,250 44,462 Steel. 20 1-21.5 m 22,572 42,560 12,759 47,954 -9,355 18,140 Steel, 21 6-23.0 m 13,385 35,352 -18,161 19,655 -27,528 1,212 371 FISHERY BULLETIN: VOL. 82, NO. 2 30 ; 20 ■ 10 : ,-\ Tj *' \ / *''' \ ', \ \ \ « 1 30 - \ * \ \ '. \ \ \ \ \ \ ' \ \ * : ■ , 1975 1976 1977 197 1 Investment Period: tike [TEBRSi « 1971 TO 1960 * 1977 TO 1950 - 1979 10 1380 2EF;0 LIME BODED FOR REFE FIGURE 7. — Net revenue for a steel vessel 20.1 to 21.5 m long operating out of Texas ports by investment period. tors should, however, bear in mind that based on this sample data similar-sized vessels built of wood are not feasible ventures. It can be inferred from this study that in high inflationary periods, shrimp producers should avoid newly financed vessels. The resultant in- creases in costs of equity and debt capital are such that economic profit is eliminated. Investing in used vessels may be a viable alternative. Results from this study further indicate that the economic performance of steel vessels is far superior to that of wooden vessels. Steel vessels can withstand adverse weather conditions much better than wooden ones and as a result, can spend longer days in offshore fishing. Steel vessels showed higher landings per trip than similar-sized wooden vessels. Therefore, the extra expense to purchase steel vessels may prove a worthy invest- ment. ACKNOWLEDGMENTS This work is the result of an accumulation of research sponsored in part by the Texas A&M 1171 19 1973 1 974 1975 1976 1977 1978 1979 19S0 Investment Period: time iTERRSi 1971 TO 1980 ♦ 1977 TO 1980 <■ 1979 TO 1980 ZEF;0 LINE BODED FOR REFERENCE FIGURE 8. — Net revenue for a steel vessel 21.6 to 23.0 m long operating out of Texas ports by investment period. University Sea Grant College Program and the National Marine Fisheries Service, NOAA, with the Texas Agricultural Experiment Station. LITERATURE CITED CAILLOUET, C. W.. AND F J. PATELLA. 1978. Relationship between size composition and ex-vessel value of reported shrimp catches from two Gulf Coast States with different harvesting strategies. Mar. Fish. Rev. 40(2):14-18. GATES, J. M., AND J. M. D'EUGENIO. 1975. Costs and returns of fisherman in the Massachusetts inshore lobster fishery. Univ. R.I., Dep. Resour. Econ., Mar. Reprint 60. Griffin, w. L., R. D. lacewell, and w. a. Hayenga. 1974. Estimated costs, returns, and financial analysis: Gulf of Mexico shrimp vessels. Mar. Fish. Rev. 36( 12):l-4. Griffin, W. L., and J. P. Nichols. 1976. An analysis of increasing costs to Gulf of Mexico shrimp vessel owners: 1971-1975. Mar. Fish. Rev. 38(3):8-12. Griffin, W L., n. J. Wardlaw, and J. R Nichols. 1976. Economic and financial analysis of increasing costs in the Gulf shrimp fleet. Fish. Bull., U.S. 74:301-309. Griffin, w. l., J. P. Nichols, R. G. Anderson, J. E. BUCKNER, AND C. M. ADAMS. 1978. Costs and returns data: Texas shrimp trawlers. Gulf of Mexico, 1974-1975. Texas A&M Univ., Sea Grant Coll. Program Rep. TAMU-T-78-003. 372 TETTEY ET AL.: PROFITABILITY OF GULF OF MEXICO SHRIMP VESSELS Hopkin, J., P Barry, and C. b. Baker. 1973. Financial management in agriculture. Interstate Printers and Publishers, Inc., Danville, 111., 459 p. JONES, T. M., J. W. HUBBARD, AND K. J. ROBERTS. 1979. Productivity and profitability of South Carolina shrimp vessels, 1971-75. Mar. Fish. Rev. 41(4):8-14. NOETZEL, B. G. 1977. Revenues, costs, and returns from vessel operation in major U.S. fisheries. Div. Fish. Manage. Open, Nat. Mar. Fish. Serv.. NOAA, Wash., D.C., 23 p. PROCHASKA, F. J., AND J. C. CATO. 1981. Economic conditions in the Gulf of Mexico shrimp industry: 1960-1981. Univ. Fla., Food Resour. Econ. Dep., Staff Pap. 180. ROBERTS, K. J., AND M. E. SAAS. 1979. Financial aspects of Louisiana shrimp vessels, 1978. Louisiana State Univ., Sea Grant Coll. Pro- gram LSU-TL-79-007, 9 p. ROBINSON, R. I., AND D. WRIGHTMAN. 1974. Financial markets: the accumulation and allocation of wealth. McGraw-Hill Book Co., New York, N.Y., 439 P- U.S. DEPARTMENT OF COMMERCE. 1971-1981. Current fisheries statistics. Natl. Oceanic Atmos. Admin., Wash., D.C., var. p. WARREN, J. P., AND W L. GRIFFIN. 1978. Costs and returns trends for Gulf of Mexico shrimp vessels. Texas A&M Univ., Dep. Agric. Econ., Staff Pap. DIR 78-1, SP-4, 23 p. 1980. Costs and returns trends in the Gulf of Mexico shrimp industry, 1971-1978. Mar. Fish. Rev. 42(2):l-7. 373 QUANTITATIVE AND QUALITATIVE BACTERIOLOGY OF ELASMOBRANCH FISH FROM THE GULF OF MEXICO1 John D. Buck2 ABSTRACT Twelve species of elasmobranch fish (8 sharks, 2 rays, 1 skate, and 1 guitarfishi taken from the Gulf of Mexico off Sarasota. Florida, were studied. Numbers of bacteria on skin were recorded, as were types of bacteria occurring on skin, gills, teeth, and in intestinal contents. Comparative observations were made on eight species of osteichthyan fish and seawater. Counts/cm2 of elasmobranch skin varied greatly both among genera and within a given species. In general, skin displayed relatively high counts which could be of significance in subsequent flesh spoilage. One brief study of spoilage of nurse shark meat at 5°C and room temperature (24°-26°C) showed that, after 7 days, species of Pseudomonas, Vibrio, and Micrococcus were dominant at the lower temperature while Micrococcus and Proteus vulgaris were recovered at 24°-26°C. Various types of bacteria found in or on the several areas of elasmobranch fish examined were compared with the little information available in the literature. Overall, Gram negative bacteria, particularly the genera Pseudomonas and Vibrio, were most common although several species of Gram positive bacteria were found also. Planococcus isolates from skin may represent important organisms because they have been implicated in shrimp spoilage. Three genera of hemolytic bacteria (Proteus. Staphylococcus, Streptococcus* were recovered from teeth of several elasmobranchs and may present a hazard to bite victims. Also, a variety of enteric bacteria potentially pathogenic to humans was found in intestinal contents; therefore, caution is suggested in handling shark material. Considerable information is available on the nor- mal and spoilage microflora of marine fish (e.g., Shewan 1961, 1971; Horsley 1977). However, the bacteriology of the elasmobranchs (sharks, skates, rays) is less understood despite a widespread pre- sent commercial fishery in local areas (Riedel 1961; McCormick et al. 1963) and its future poten- tial (Juhl 1973; U.S. Department of Commerce 1982). Venkataramen and Sreenivasan (1955) studied the bacterial flora of skin of one shark caught off India; Johnson et al. (1968) charac- terized the intestinal microflora of five species of sharks obtained in the Indian Ocean; and Yap (1979) reported on skin isolates of two sharks freshly caught off Australia. Liston (1957) studied the bacteria associated with slime and gills of fresh North Sea skate. Spoilage bacteria in shark flesh were noted by Wood (1950) in Australia and Velankar and Kamasastri (1955) in India. Al- though the number of shark attacks on humans worldwide is statistically small (Baldridge 1974; Coppleson 1975), there are no substantive data on the potential bacteriological hazard of shark bites other than brief notations of hemolytic bacteria 'Contribution No. 162 from The University of Connecticut Marine Research Laboratory, Noank, CT 06340. 2The University of Connecticut, Marine Sciences Institute, Marine ResearchLaboratory, P.O. Box 278, Noank, CT 06340. Manuscript accepted November 1983. FISHERY BULLETIN: VOL. 82. NO. 2, 1984. being recovered from the teeth of sharks (Davies 1960; Davies and Campbell 1962). Consequently, this study was initiated to characterize the numbers and types of bacteria associated with a wide variety of elasmobranch fish common to the Gulf of Mexico. Comparative data were recorded for water and osteichthyan fish caught in the same area. These results will have relevance to the potential spoilage of elasmo- branch meat and the pathobiology of shark bites. METHODS Sampling Sites All fish were obtained from the Gulf of Mexico within several kilometers off Sarasota, Fla., or in the contiguous waters of Sarasota Bay. Small elasmobranchs were caught by use of a long, monofilament gill net set from the surface to a depth of about 1 m. Larger sharks were caught using baited longlines farther offshore. The one sand tiger shark, Odontaspis taurus, studied was obtained from the Mystic Marinelife Aquarium (Mystic, Conn.) and had been dead and refriger- ated for 4 h. This shark, caught off the coast of New Jersey 3 d previously, was maintained in chlorinated brine water at the aquarium for 2 d 375 FISHERY BULLETIN: VOL. 82, NO. 2 prior to death. Only teeth and intestinal contents were sampled. For comparative bacteriological studies, several osteichthyan fish were caught by rod and reel or retrieved from the gill net men- tioned above. Some elasmobranchs were occasion- ally maintained in large concrete or fiberglass tanks containing seawater piped from Sarasota Bay. All fish examined from tanks were either alive or had been dead for less than 1 h. Fish caught in the Gulf of Mexico were either iced if dead or kept in a wet hold until examined, which was routinely less than 1 h. In one case (see below) sharks were dead for 3 h before sampling. Water samples were collected from either Sarasota Bay or the tanks containing fish. Quantitative Analysis Swabbing was compared initially with two other methods involving the use of membrane fil- ters in a quantitative sampling of elasmobranch skin. In the swabbing technique, a sterile alumi- num foil template containing a 3.1 x 3.1 cm square opening (9.6 cm2) was placed on the skin, on the side of each fish just posterior to the gills. A sterile polyester-tip swab (Falcon No. 20693) was used over the exposed area in all directions, and the tip was broken off in a screw-capped tube containing 10 ml of sterile seawater. Decimal dilutions of this were prepared in 9 ml of sterile seawater and 0.1 ml volumes spread (Buck and Cleverdon 1960) on Bacto-Marine Agar (Difco Laboratories, Detroit, Mich.). One procedure with membrane filters in- volved placing sterile 0.45 /xm gridded membranes (Millipore No. HAWG047SO) on shark skin and pressing them down by rolling a sterile glass rod across the membrane. The membrane was then placed grid uppermost on the surface of an agar plate. The second membrane filter technique was similar to the first except that the membrane, after exposure to the skin, was placed in a sterile plastic screw-cap centrifuge tube with sterile sea- water and agitated on a vortex mixer for 30 s. Decimal dilutions and platings were then made as indicated above. All plates were incubated at room temperature (24°-26°C) for 3-5 d. Counts/cm2 of skin were calculated. Qualitative Analysis In addition to skin, other body areas including 3Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. teeth, gills, and intestinal contents were sampled by use of a swab. The upper third of plates of eosin-methylene blue, tryptic-soy, brain heart in- fusion, and marine agar (all Difco) was swabbed, and the remaining two portions of the plate were streaked sequentially with a sterile wire loop to isolate colonies. Plates were incubated at room temperature (24°-26°C) and 37°C for 1-5 d and colonies were selected based on differences in morphology. Qualitative changes in the bacterial flora on shark flesh were assessed as a function of time and temperature. Pieces of nurse shark, Ginglymos- toma cirratum , flesh (about 2 cm2) were cut asepti- cally from one area of one side of the fish after the skin had been removed. These pieces were placed in sterile petri dishes and incubated at room tem- perature (24°-26°C) and 5°C. Initially and after 3, 4, 5, and 7 d incubation, the surface of the flesh was sampled using a sterile loop, which was used to directly inoculate agar plates to obtain well- isolated colonies. Tank and bay samples were collected at a depth of about 30 cm in sterile bottles. One ml volumes were diluted in 9 ml of sterile seawater and addi- tional decimal dilutions prepared in a similar manner. Spread plates (see above) were made on marine agar. Representatives of various colonial types were selected and identified after incubation for 3-5 d at room temperature. All isolates were maintained on slants of either marine agar or tryptic-soy agar. Gram reactions were recorded by both conventional staining and the KOH technique (Buck 1982). Gram negative enteric bacteria were identified using either the Enterotube II (Roche Diagnostics, Nutley, N.J.) or API 20E ( Analytab Products, Plainview, N.Y.) sys- tems. Hemolysis was detected on tryptic-soy agar containing 5% horse blood. Other bacteria were characterized using the methods of Shewan et al. (1960) and Oliver (1982). RESULTS AND DISCUSSION Quantitative Analysis Counts by the swab technique averaged 115% higher than those obtained by membrane filters applied directly to agar plates (two experiments) and 910% higher than counts by agitating the membrane in seawater followed by dilution and plating (three experiments). All subsequent counts of skin bacteria on both elasmobranch and osteichthyan fish were made using the swab 376 BUCK: BACTERIOLOGY OF ELASMOBRANCH FISH technique (Yap 1979), although this technique has inherent weaknesses (Horsley 1977), especially in examining shark skin, because it is abrasive and essentially three-dimensional. Nonetheless, counts obtained by swabbing a known area were always higher than those achieved by pressing a membrane filter on the skin, probably because the membrane did not recover bacteria associated with the lower portion of the denticles. The fibrous texture of the swab, while prone to some shredding unless care was used, may have allowed penetra- tion into the skin. Perhaps an agar-coated slide or "paddle" which could be pressed onto shark skin would be more effective, although such a proce- dure might be unwieldy in the field. Table 1 shows the number of bacteria recovered from the skin of various elasmobranch and os- teichthyan fish. Elasmobranch skin showed a very wide range of counts both among genera and within a given species. Data in Table 1 indicated that there was no obvious correlation between TABLE 1 . — Number of bacteria on fish skin ( Bacto-marine agar i Taxon Source No. /cm2 Elasmobranchs Florida smoothhound. Mustelus nomsi Gill netted '120 '840 Nurse shark. Ginglymostoma cirratum Gill netted '8.200 Atlantic sharpnose shark, Rhizopnonodon terraenovae Gill netted '2.300 220.000 '1,000 1 1 ,000 Bonnethead or shovelhead shark. Sphyrna tiburo Gill netted 400,000 42,000 1,100 410,000 Brown or sandbar shark. Carcharhmus plumbeus Longline caught 2,300 Blacktip shark. Carcharhmus limbalus Gill netted 530.000 Blacknose shark. Carcharhmus acronotus Gill netted 330,000 Tiger shark. Galeocerdo cuvieri Longline caught 240 Atlantic guitarfish. Rhmobatos lentigmosus Gill netted '260 '460 Clearnose skate. Ra/a eglantena Gill netted '100,000 80.000 Southern stingray. Dasyatis amencana Gill netted 16,700 Cownose ray, Rhmoptera bonasus Gill netted 50.000 Osteichthyes Gulf menhaden. Brevoortia patronus Gill netted 23.000 Southern flounder, Parahchthys lethostigma Rod caught 26.000 Spanish mackerel. Scomberomorus maculatus Gill netted 15.000 Ladyf ish . Elops saurus Gill netted 19.000 Searobm. Pnonotus tribulus Gill netted 100 Black drum. Pogonias cromis Gill netted 15.000 Permit. Trachinotus falcatus Gill netted 7,100 Atlantic spadefish, Chaetodipterus faber Gill netted 14.000 'Tank held bacterial counts on tank-held and freshly caught elasmobranchs, although bacterial counts on skin of three tank-held Atlantic sharpnose sharks were about two orders of magnitude lower than counts for one specimen of the same species caught in a gill net. Three of the highest counts on shark skin noted in Table 1 (bonnethead, 410,0007cm2; blacktip, 530,000/cm2; blacknose, 330,000/cm2) were from fish which had been dead for 3 h. This suggests that bacteria rapidly colonize skin of dead sharks. Yap (1979) reported varying counts on skin of freshly caught shark which depended on the area of the fish sampled. His counts (310- 1,900/cm2) were, in general, lower than those re- ported herein, and were estimated from broth di- lutions and not plates. The shark skin sampled here displayed relatively high bacterial counts which are of considerable significance if the sharks are to be used for food. Large numbers of bacteria could be deposited onto flesh which sub- sequently may undergo more rapid spoilage if not adequately washed and/or refrigerated. With the exception of the bighead searobin, Prionotus tribulus, the counts on osteichthyan fish were quite similar and within the range reported in other studies (Horsley 1977). Qualitative Analysis Table 2 shows the number of isolates and genera of bacteria recovered from elasmobranch and os- teichthyan fish and from waters where the fish were taken or held. The Gram negative bacteria, especially Pseudomonas, Vibrio, and Cytophaga, accounted for 89% of the 111 isolates from skin. In other studies, pigmented Gram positive isolates of Micrococcus, Bacillus, and Corynebacterium were the most common (30 strains) on the skin of one shark (Carcharhinus sp.) caught off India; only 5 cultures of Gram negative Achromobacter, Flavobacterium, and Vibrio were recovered (Ven- kataraman and Sreenivasan 1955). Pseudomonas (40% ), Micrococcaceae (309c ), and Moraxella (15% ) were dominant on skin of a freshly caught shark off Australia (Yap 1979). The data here for skin (Table 2) show a similar percentage for Pseudomonas but considerably fewer isolates of Moraxella and Gram positive cocci. Table 2 shows that for Gulf of Mexico sharks, fewer genera were recovered from intestines than from other areas, but that the Gram negative gen- era were predominant and species of Photobac- terium, Pseudomonas, and Vibrio accounted for 57% of the isolates. Gram positive bacteria (one 377 FISHERY BULLETIN: VOL. 82, NO. 2 TABLE 2.— Number of isolates and percentage of bacterial genera recovered from elasmobranch and osteichthyan fish and water of the Gulf of Mexico. Elasmobranchs All elasmobranch Osteichth^* Skir Gills Teeth Intestines samples No. % skin Water Genus No. % No. % No. % No. % No % No. % Gram negative Aeromonas 1 1 2 3 3 1 2 4 Acinetobacter 4 4 2 10 6 8 1 7 13 6 3 7 Alcaligenes 2 2 5 3 1 2 5 4 8 Cytophaga 11 10 5 1 1 1 7 14 6 1 2 Flavobacterium 1 1 4 5 1 7 6 3 1 2 Flexibacter 1 1 1 1 Moraxella 4 4 1 1 5 2 2 5 3 6 Photobactenum 4 4 4 18 5 7 4 29 17 8 6 12 Pseudomonas 35 32 5 10 13 2 14 48 22 8 18 15 30 Vibrio 30 27 10 46 17 23 2 14 59 27 16 36 15 30 Xanthomonas 6 5 5 2 3 9 4 4 9 2 4 Gram positive Arthrobacter 4 4 3 4 1 7 8 4 1 2 Bacillus 4 5 1 7 5 2 4 9 1 2 coryneforms 1 1 5 1 1 3 1 2 5 1 2 Micrococcus- Staphylococcus 4 4 5 8 11 13 6 2 5 1 2 Planococcus 3 3 2 3 5 2 Streptococcus 9 12 1 7 10 5 Total 111 22 75 14 222 45 51 isolate each of Arthrobacter, Bacillus, and Strep- tococcus) represented 21% of the total. In a study of intestinal material from five species of sharks caught in the Indian Ocean, 10 isolates of Bacillus were found, and 1 each of Corynebacterium, Al- caligenes, Vibrio, Spirillum, and Xanthomonas; one animal showed no bacteria (Johnson et al. 1968). No data are available in the literature on bacte- rial types recovered from shark gills, although the gills and skin of North Sea skates have been studied (Liston 1957). Gram negative bacteria were dominant with Pseudomonas most common on both skin and gills. Qualitative observations of skin agreed with the present data (Table 2), but skate gills showed a much higher percentage of Pseudomonas (60%) compared with this study (5%). The other Gram negative bacteria from skate gills were also found in this study (Table 2). Hemolytic bacteria were isolated from the teeth of sharks in the present study. Streptococcus spp. were recovered from teeth of shovelhead, Sphyrna tiburo, and sand tiger, Odontaspis taurus, sharks; Staphylococcus spp. were found on the teeth of a cownose ray, Rhinoptera bonasus; and Providencia rettgeri was recovered from teeth of two shovelhead sharks. All of these bacteria were from sharks taken in the Gulf of Mexico except for the sand tiger shark which was caught off New Jersey and had been in captivity for only 3 d. In addition, several hemolytic species of Vibrio have been iso- lated recently from the teeth of a white shark, Carcharodon carcharias, caught off Block Island, R.I. (Buck et al. unpubl. data4). Hemolytic bacteria were found in the mouths of sharks from South African waters, and it was suggested that bacterial infections of bites could have been a contributing factor in the deaths of victims (Davies 1960). The hemolytic bacterium recovered from teeth of Carcharhinus zambezensis (leucas?) was described as a "Paracolon bacillus" (Davies and Campbell 1962). The present observations not only confirm the occurrence of hemolytic organisms on teeth of sharks in nature but also extend these types to include bacteria not reported previously from sharks and the number of species of sharks which harbor them. They suggest that shark bites could possibly introduce potentially pathogenic bacteria into the tissues of victims. A variety of enteric bacteria was found as- sociated with the intestinal contents and occa- sionally the teeth of elasmobranchs; none were recovered from the gills or skin. These data are presented in Table 3. Three cultures only, all Shigella species, were isolated from bony fish. One was found in pinfish, Lagodon rhomboides, intes- tine, and two strains were isolated from a black drum, Pogonias cromis — one on the gills and the other from intestinal contents. 4Buck, J. D., S. Spotte, and J. J. Gadbaw, Jr. Manuscr. in prep. Bacteriology of the teeth from "Jaws": Medical implica- tions for shark-bite victims. 378 BUCK: BACTERIOLOGY OF ELASMOBRANCH FISH Enterobacteria are found frequently on os- teichthyan fish, but there are no reports in the literature on their occurrence in (on) elasmo- branchs. If waters contain domestic wastes, then the fish will almost certainly be contaminated also (Shewan 1971; Horsley 1977). Coliform counts in Sarasota Bay are generally low, although counts of 1,800/100 ml have been recorded in one bayou re- ceiving treated sewage effluent (Buck, unpubl. data5 ). Areas north (Tamplin et al. 1982) and south (Peterson and Yokel 1983) of Sarasota Bay have shown the presence of potentially pathogenic en- teric bacteria. Consequently, the elasmobranch fish studied here may well have been in contact with sources of enterobacteria. The enteric bac- teria encountered on the teeth and in the intes- tines of several elasmobranchs probably reflected feeding habits and originated on smaller prey which had passed through waters receiving human and/or animal excretions. Enteric bacteria do not multiply in passage through rainbow trout but temperature may be an important factor (Lesel and Peringer 1981; Lesel and LeGac 1983). The internal temperature of some sharks (Lam- nidae) (Carey et al. 1981; Smigh and Rhodes 1983) is significantly warmer than the surrounding water. In subtropical areas, increased water tem- 5J. D. Buck. University of Connecticut Marine Research Laboratory. Noank, Conn., unpubl. data, 1982. TABLE 3. — Enterobactenaceae isolated from elasmobranch fish. Taxon Bacteria Nurse shark. Ginglymostoma cirratum^ Intestine Shovelhead shark. Sphyrna tiburo Intestine Teeth Sandbar shark. Carcharhmus plumbeus Intestine Teeth Blacktip shark. Carcharhmus limbalus Intestine Teeth Sand tiger shark. Odontaspis taurus] Intestine Cownose ray. Rhinoptera bonasus Intestine Teeth Clearnose skate. Ra/a eglantera'' Intestine Proteus vulgaris Escherichia coli Enterobacter agglomerans Escherichia coli Shigella sp. Citrobacter Ireundii Providencia rettgen Providencia sp. Shigella sp Proteus vulgaris Providencia rettgeri Escherichia coli Providencia alcalifaciens Shigella sp Escherichia coli Proteus vulgaris Citrobacter freundu Morganella morgann Proteus vulgaris Shigella sp. Serratia liquetaciens Escherichia coli 'Tank held. perature and that of the interior tissues of elasmo- branchs might provide an environment that encourages bacterial multiplication, including potential pathogens. While none of the entero- bacteria, except perhaps Shigella species, recov- ered from intestines and teeth of elasmobranchs represent primary pathogens, members of the other genera are commonly found as secondary or opportunistic pathogens in humans. Thus, caution should be exercised when handling dead shark material, particularly internal organs such as the digestive tract. The genera Vibrio and Pseudomonas were pre- dominant bacteria in combined data for all elas- mobranch samples (Table 2). When isolates for tank-held and open-water fish were compared, these two genera were the most common in each group. The occurrence of other microbes did not vary more than 69c for any genus of bacteria be- tween tank-held and freshly caught elasmo- branchs, except for Photobacterium species which represented 119c of the isolates from the former and 39c of the latter. The bacterial flora of osteichthyan fish and sea- water consisted largely of Gram negative bacte- ria (829c and 949c, respectively), with Vibrio and Pseudomonas predominating. No substantial dif- ferences in generic composition were noted be- tween Sarasota Bay water and fish holding tanks. Fewer numbers of several other Gram negative forms were found; these results agree with those of others (e.g., Shewan 1961). Small populations of Gram positive bacteria (Arthrobacter, Bacillus, cocci) were noted and probably represented ter- restrial influence because the fish were taken from nearshore waters. This assumption may require reevaluation because there may be a widespread distribution of Gram positive bacteria in seawater (Gunnetal. 1982). The microflora of spoiling shark muscle (no species indicated) from Australia have been studied, and the genus Corynebacterium was the dominant organism; Pseudomonas species and Gram positive cocci were also found in large num- bers (Wood 1950). Few coryneforms were isolated in the present study, although Pseudomonas and Gram positive organisms were commonly recov- ered. In the brief study here of nurse shark flesh, the dominant bacteria found initially were species of Vibrio and Pseudomonas. After 7 d of incuba- tion at 5°C, the flora were composed principally of Pseudomonas, Vibrio, and Micrococcus. When flesh was held at room temperature (24°-26°C), Gram positive cocci and Proteus vulgaris were 379 FISHERY BULLETIN: VOL. 82, NO. 2 predominant after 7 d. The latter is capable of hydrolyzing urea, and several species of Micrococ- cus are urease-positive (Buchanan and Gibbons 1974); hence, both of these groups are potential contributors to shark tissue spoilage. This en- richment of Gram positive types in elasmobranch spoilage was noted by Wood (1950). Bacteria were found in 12 samples of shark mus- cle (Scoliodon sp.) allowed to spoil at 27°-30°C (Velankar and Kamasastri 1955). No coryneforms and only one Micrococcus isolate were found; all others were unidentified Gram negative nonpig- mented rods. The spoilage of iced abdominal wall muscle of Australian school shark, Galeorhinus australis, was studied by Yap (1979). Pseudomonas and Moraxella (459c and 20%, respectively) were the dominant bacteria recovered after 10 d although the Gram positive cocci represented 15c/c of the total. The data presented here for the flesh spoilage experiment, albeit limited, confirm the observa- tions of Wood (1950) and Yap (1979), but none of these parallel the findings of Velenkar and Kamasastri (1955) which also concerned sharks from subtropical waters. Perhaps the local marine microflora or experimental conditions influenced their observations. Although the number of isolations was rela- tively small, the genus Planococcus was found as- sociated with elasmobranch skin and teeth in this study. All the cultures recovered were yellow- pigmented and were probably Planococcus citreus, the only accepted species (Buchanan and Gibbons 1974). This proteolytic bacterium has been impli- cated in shrimp spoilage (Alvarez 1982) and may be a significant spoilage organism of elasmo- branch flesh. ment on flesh did not correlate well in all respects with results of other studies which in some in- stances were limited to one or a few fish or differ- ent species than those considered here. Also, little information was provided in the literature on cul- tural conditions and other variables which could affect development of various bacteria reported. The data here substantiate the occurrence of cer- tain potential spoilage bacteria on skin and in- clude the genus Planococcus which has been im- plicated in shrimp spoilage. The present study also confirms and extends other observations on the occurrence of hemolytic bacteria on shark teeth. In addition, potentially pathogenic enterobacteria were recovered from teeth and intestinal contents of several elasmobranch species. It is hoped that future studies will include larger numbers of addi- tional shark species for a clearer assessment of the role of bacteria in both spoilage and public health aspects of a valuable and underutilized marine resource. ACKNOWLEDGMENTS The majority of this research was conducted at the Mote Marine Laboratory, Sarasota, Fla., while the author was on sabbatical leave. I am especially indebted to William H. Taft, President, for provid- ing space and facilities. Appreciation is extended also to Carl Luer and Perry Gilbert of the Mote Laboratory for valuable advice and background information. Jack Schneider furnished the shark from the Mystic Marinelife Aquarium. I thank Denise Baird and Shannon Kelly for their time and patience in assisting with laboratory identifi- cation of bacterial isolates. LITERATURE CITED CONCLUSIONS The observations reported here have shown that elasmobranch fish contain a large and diverse bacterial flora. Because there is little information on the microbiology of sharks, skates, and rays, assessing the relative significance of the data is difficult. In many cases, counts of bacteria on the skin were an order of magnitude higher than those noted on osteichthyan fish caught in the same waters. In other samples, counts were two orders of magnitude lower. Considerable variation was seen in individual species of elasmobranchs. Types of bacteria recovered from different areas of fresh fish and during one controlled spoilage experi- ALVAREZ, R. J. 1982. Role of Planococcus citreus in the spoilage of Pen- naeus shrimp. Zentralbl. Bakteriol. Parasitenkd. In- fektionskr. Hyg. Abt. I Orig. C3:503-512. BALDRIDGE, H. D. 1974. Shark attack. Berkley Publ. Corp., N.Y., 263 p. BUCHANAN, E. G., AND N. E. GIBBONS (editors). 1974. Bergey's manual of determinative bacteriology. 8th ed. The Williams and Wilkins Co., Baltimore, 1268 p. BUCK, J. D. 1982. Nonstaining (KOH) method for determination of Gram reactions of marine bacteria. Appl. Environ. Mi- crobiol. 44:992-993. BUCK. J. D., AND R. C. CLEVERDON. 1960. The spread plate as a method for the enumeration of marine bacteria. Limnol. Oceanogr. 5:78-80. Carey, F. G., J. M. Teal, and J. W. kanwisher. 1981 . The visceral temperatures of mackerel sharks ( Lam- 380 BUCK: BACTERIOLOGY OF ELASMOBRANCH FISH nidae). Physiol. Zool. 54:334-344. COPPLESON.V. M. 1975. Patterns of shark attack for the world. In P. W. Gil- bert (editor), Sharks and survival, p. 389-421. D.C. Heath and Co., Boston, Mass. DAVIES, D. H. 1960. The Oceanographic Research Institute. So. Afr. Assoc. Mar. Biol. Res. Bull. 1, p. 11-15. [Durbin, South Africa.] DAVIES, D. H., AND G. D. CAMPBELL. 1962. The aetiology, clinical pathology and treatment of shark attack. J. R. Nav. Med. Serv. 48:110-136. Gunn, B. A., F. L. Singleton, E. r. peele, and r. r. Colwell. 1982. A note on the isolation and enumeration of gram positive cocci from marine and estuarine waters. J. Appl. Bacteriol. 53:127-129. HORSLEY, R. W. 1977. A review of the bacterial flora of teleosts and elas- mobranchs, including methods for its analysis. J. Fish Biol. 10:529-553. JOHNSON, R. M., R. M. SCHWENT, AND W. PRESS. 1968. The characteristics and distribution of marine bac- teria isolated from the Indian Ocean. Limnol. Oceanogr. 13:656-669. JlHL. R. 1973. Fishery resources of the Caribbean and their poten- tial //; C. O. Chichester and H. D. Graham (editorsi. Microbial safety of fishery products, p. 25-40. Acad. Press, NY. LESEL. R., AND P LE GAC. 1983. Transit of enterobactena originating from homeo- therms in fish living at low temperature. Aqua- culture 31:109-115. LESEL. R . AND R PERINGER. 1981. Influence of temperature on the bacterial microflora in SalniD gairdneri Richardson. Arch. Hydrobiol. 93:109-120. LISTON.J. 1957. The occurrence and distribution of bacterial types on flatfish. J. Gen. Microbiol. 16:205-216. MCCORMICK, H. W, T. ALLEN, AND W. E. YOUNG. 1963. Shadows in the sea. Chilton Book Co., Philadelphia. 415 p. Oliver, J. D. 1982. Taxonomic scheme for the identification of marine bacteria. Deep-Sea Res. 29:795-798. PETERSON, M. E., AND B. J. YOKEL. 1983. Recovery of pathogens from Naples Bay, Flori- da. Abstr. Annu. Meet., Am. Soc. Microbiol., p. 278. RlEDEL, D. 1961. World fisheries. In G. Borgstrom (editor), Fish as food, Vol. 1, p. 41-75. Acad. Press, N.Y. SHEWAN, J. M. 1961. The microbiology of sea-water fish. In G. Borgstrom (editor), Fish as food, Vol. 1, p. 487-560. Acad. Press, N.Y. 1971. The microbiology of fish and fishery products — a progress report. J. Appl. Bacteriol. 34:299-315. SHEWAN, J. M., G. HOBBS, AND W. HODGKISS. 1960. A determinative scheme for the identification of cer- tain genera of Gram-negative bacteria, with special refer- ence to the Pseudomonadaceae. J. Appl. Bacteriol. 23:379-390. SMITH, R. L., AND D. RHODES. 1983. Body temperature of the salmon shark, Lamna dit- ropis. J. Mar. Biol. Assoc. U.K. 63:243-244. TAMPLIN, M., G. E. RODRICK, N. J. BLAKE, AND T. CUBA. 1982. Isolation and characterization of Vibrio vulnificus from two Florida estuaries. Appl. Environ. Microbiol. 44:1466-1470. U.S. DEPARTMENT OF COMMERCE. 1982. Status of the fishery resources off the northeastern United States for 1981. NOAA Tech. Memo. NMFS-F NEC-12, Woods Hole, Mass., 114 p. VELANKAR, N. K., AND P. V KAMASASTRI. 1955. Shark spoilage bacteria. Curr. Sci. (Indiai 24:272- 273. VENKATARAMAN, R., AND A. SREENIVASAN. 1955. Bacterial flora of fresh shark. Curr. Sci. ( India i 24:380-381. WOOD. E. J. F 1950. The bacteriology of shark spoilage. Aust. J. Mar. Freshw. Res. 1:129-138. YAP, A. S. J. 1979. Microbiological considerations in shark han- dling. Food Technol. Aust. 31:297-300. 381 DISTRIBUTION AND FEEDING OF THE HORSESHOE CRAB, LIMULUS POLYPHEMUS, ON THE CONTINENTAL SHELF OFF NEW JERSEY1 Mark L. Botton2 and Harold H. Haskin3 ABSTRACT The horseshoe crab, Limulus polyphemus , population was assessed during hydraulic dredge surveys of the surf clam resource in the inshore 5.5 km (3 nautical miles) of the continental shelf off New Jersey from 1976 through 1979. Frequency of occurrence and abundance was higher off the southern half of the state, which may be a function of its proximity to Delaware Bay, a principal spawning site. Horseshoe crabs consumed various benthic organisms, primarily bivalves, arthropods, and polychaetes. Surf clams, Spisula solidissima, were important in the diet of Limulus; individual valves ranged in length from < 1 mm to about 35 mm. In the laboratory, horseshoe crab predation was observed on surf clams as long as 46 mm. This report describes the distribution of the horse- shoe crab, Limulus polyphemus, on the inshore continental shelf off New Jersey, and the diets of a sample of these animals. Previous studies of the horseshoe crab on the continental shelf are limited to distributional records (Wolff 1977; Shuster 1979) or tagging studies conducted close to es- tuarine spawning areas (Baptist et al. 1957; Rud- loe 1980), although crabs have been found at depths as great as 200 m according to National Marine Fisheries Service surveys (J. W. Ropes4). Since the early 1960's, an intensive surf clam, Spisula solidissima, fishery has developed along the New Jersey coast (Ropes 1982). The junior au- thor (Haskin) and his colleagues have inven- toried the surf clam resource in the New Jersey waters, to 5.5 km (3 nmi) offshore yearly since 1972. All macroinvertebrates, including L. polyphemus, captured in hydraulic dredge hauls from 1976 through 1979 were counted. Since a percentage of the horseshoe crab population mi- grates from the continental shelf to estuaries and back again (Shuster 1982), we analyzed both tem- poral and spatial variability. Separating these ef- fects was difficult because the sampling program was designed primarily to inventory a sessile clam resource, rather than a migratory one. However, the data, based on over 1,100 stations, still repre- sent the most systematic survey of L. polyphemus distribution on the inshore continental shelf, and since exploitation of these crabs for biomedical research and bait is increasing (Pearson and Weary 1980), our study provides baseline informa- tion should future population assessment studies be warranted. Information on the feeding biology of horseshoe crabs is limited (Lockwood 1870; Fowler 1908; Shuster 1950; Smith and Chin 1951; Smith 1953: Smith et al. 1955; Botton 1981). In this study, stomach contents from 36 horseshoe crabs from the continental shelf were examined to supple- ment a more intensive study of the food habits of animals from Delaware Bay (Botton 1982); in Au- gust 1980, predation by crabs on surf clams about 4 cm long was examined in the laboratory. MATERIALS AND METHODS 'No. D-32503-1-83 of the New Jersey Agricultural Experiment Station, Rutgers University, Piscataway, N.J. 2Department of Biological Sciences and Oyster Research Laboratory, New Jersey Agricultural Experiment Station, Rut- gers University, Piscataway, N.J.; present address: Excel Divi- sion. Fordham University, The College at Lincoln Center, New- York, NY 10023. department of Biological Sciences and Oyster Research Laboratory, New Jersey Agricultural Experiment Station, Rut- gers University, P.O. Box 1059, Piscataway, NJ 08854. 4J. W. Ropes, Northeast Fisheries Center Woods Hole Labora- tory, National Marine Fisheries Service, NOAA, Woods Hole, MA 02543, pers. commun. February 1983. Manuscript accepted October 1983. FISHERY BULLETIN: VOL. 82, NO. 2, 1984. Population Survey Stations were sampled with a hydraulic dredge (Meyer et al. 1981), adjusted to retain surf clams >88 mm. This gear retained both adult and sub- adult horseshoe crabs. Catch data, as number of animals per tow, were normalized for dredge width and tow time. The standard tow (ST) is defined as a 5-min haul using a 152 cm knife (width of dredge). 383 FISHERY BULLETIN: VOL. 82, NO. 2 This standard tow covered an area of about 418 m2. The New Jersey coastline from Cape May to Shark River Inlet was subdivided into 19 areas (Table 1). Stations were located by 3-point sextant fixes and/or loran C, and grouped in intervals of 0-1.8 km (0-1 nmi) (0-0.9 and 0.9-1.8 km north of Beach Haven Inlet), 1.8-3.7 km (1-2 nmi) and 3.7- 5.5 km (2-3 nmi) based on distance from land. Inlets were used as latitudinal break points (Fig. 1 1. Because of the reduced sampling effort from 1.8 km offshore, these areas were larger than areas inshore of 1.8 km. For statistical analysis, all tows on the same day in a given area were considered replicates. The Analysis of Variance (ANOVA) for the number of crabs per tow had three sources of variation: Area, time nested within area, and replicate tow nested within time within area. Because each year's de- sign was unbalanced, a pseudo-F procedure ( Hicks 1973) tested the significance of the area effect. Data were log-transformed to stabilize the var- iances. When areas were sampled more than once in a given year, we tested differences between sample dates using a completely randomized one-way classification ANOVA. If the F-test was significant, a Student-Neumann-Keuls proce- dure for unequal group sizes tested for differences between the means for each sample date (Zar 1974). TABLE 1 . — Description of areas of the New Jersey coast surveyed from 1976 to 1979. Distance offshore (km) Area Southern boundary Northern boundary 0-1 8 1.8-3.7 3.7-5.5 1 Cape May Inlet 2 Hereford Inlet 3 Stone Harbor 4 Townsends Inlet 5 Corson Inlet 6 Great Egg Harbor Inlet Absecon Inlet 8 Beach Haven Inlet1 9 Beach Haven Inlet2 10 Barnegat Inlet 1 1 Barnegat Inlet2 12 Cape May Inlet 13 Townsends Inlet 14 Absecon Inlet 15 Beach Haven Inlet 16 Cape May Inlet 17 Townsends Inlet 18 Absecon Inlet 19 Beach Haven Inlet Hereford Inlet Stone Harbor Townsends Inlet Corson Inlet Great Egg Harbor Inlet Absecon Inlet Beach Haven Inlet Barnegat Inlet Barnegat Inlet Shark River Inlet Shark River Inlet Townsends Inlet Absecon Inlet Beach Haven Inlet Shark River Inlet Townsends Inlet Absecon Inlet Beach Haven Inlet Shark River Inlet '0-0.9 km 20 9-1 8 km. Stomach Contents Thirty-six adult L. polyphemus were collected between 10 July and 25 August 1978 for analysis of stomach contents. The results are grouped for three locations: Stone Harbor (1 station, 5 indi- viduals), Atlantic City (12 stations, 24 individu- als), and Point Pleasant (3 stations, 7 individuals) (Fig. 1). Complete digestive tracts were removed from crabs aboard ship or shortly after returning to the laboratory, fixed in 10% Formalin5 seawater, and later transferred into 70% ethanol until examina- tion. Food, much of which was entangled with mucus, was sorted under a 10 x stereoscope. The number of bivalves was determined by counting the number of umbones and dividing by 2. Shells were measured by ocular micrometer or vernier caliper. RESULTS Population Surveys 1976 Survey Sampling commenced in mid-July and was most extensive in late August and early September; no areas north of Beach Haven Inlet were sampled. Horseshoe crabs were present in over 909c of all hauls in the first 1.8 km between Hereford Inlet and Townsends Inlet, and from 1.8 to 3.7 km be- tween Cape May and Townsends Inlet (Table 2). More than 10 animals/ST were dredged from 1.8 to 3.7 km offshore between Cape May and Absecon 5Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. TABLE 2. — 1976 Limulus polyphemus survey results. Area means are expressed as the number of crabs per standard tow, as defined in the text. CV = coefficient of variation. Data were log transformed prior to Analysis of Variance. Area locations are shown in Table 1. Distance offshore Area N sta- tions % with crabs Mean CV Maximum 0-1.8 km 1 27 85.2 7.7 2 12 66.7 41 3 21 905 60 4 7 714 11.2 5 9 88 9 4 8 6 8 62.5 2.4 7 10 70.0 7.2 1.8-3 7 km 12 31 90.3 14.6 13 28 78.6 125 14 10 800 3.0 3 7-5.5 km 16 18 83.3 6 6 17 16 62.5 25.4 18 7 42.9 11 Analysis of Variance: Source df SS MS Total 202 25994 Area 12 40.08 3.34 Time (area) 38 80.77 2.13 Station (time (area)) 152 139 09 0 92 081 1 27 065 1 14 1 05 1 01 1 01 069 236 068 0.97 2.72 1.30 1.12 2.32 25.7 107 15 0 35.3 15.0 5.4 20.0 47.1 145.7 6.0 268 2774 3.2 ns 005 384 BOTTON and HASKIN: DISTRIBUTION AND FEEDING OF HORSESHOE CRAB NEW JERSEY Shark River Inlet Barnegat Inlet Beach Haven nlet Absecon In let . ,%** H k ATLANTIC Great Egg Harbor OCEAN Corson Inlet Townsends Inlet Hereford Inlet Cape May Nautical mi 0 5 10 10 20 Ki lometers FIGURE 1. — Map of the New Jersey coast. Filled circles show the locations of stations from which Limulus polyphemus stomach contents were obtained. Inlets, and in several areas from 0 to 1.8 km and 3.7 to 5.5 km between these two inlets. Two hundred seventy-seven crabs were found in a single dredge tow, 5.4 km off Townsends Inlet. In 1976, both time within area and station within time within area were greater sources of variability than area itself (Table 2). The inshore 1.8 km from Cape May to Hereford Inlets was sam- pled on 10 and 30 July and again on 4-5 September. The mean number of crabs collected per standard tow on each date was 8.3, 9.3, and 2.1, respectively (F = 2.86, 0.10 < P <0.05). There were no signifi- cant differences between sampling dates within any other individual areas. 1977 Survey Most sampling in southern New Jersey was done in June and July, but the areas north of Beach Haven Inlet were sampled in August. Thus, tem- poral variability was confounded with geographic variability. Areas north of Beach Haven Inlet were 385 FISHERY BULLETIN: VOL. 82, NO. 2 relatively depauperate; the area effect was sig- nificant at P < 0.07 (ANOVA); the time within area effect was significant at P < 0.05 (Table 3). Limulus polyphemus occurred in 70% to 100% of all tows between Cape May and Absecon Inlet (Table 3). Crabs were most numerous from 0 to 1.8 km offshore between Cape May and Hereford In- lets. When sampled on 15 June (23 stations), there was an average of 15.52 crabs/ST, but in late August-early September (9 stations), only 2.17 crabs/ST (F = 5.006, P < 0.05). Abundance from 1.8 to 3.7 km between Cape May and Townsends Inlets declined over the same time, from 13.08 to 3.3 crabs/ST (F = 4.805, P < 0.05). From 1.8 to 3.7 km offshore, between Townsends Inlet and Abse- con Inlet, crabs declined between early July (x = 25.33, n = 6), late July-early August {x = 3.95, n = 14), and late August (x = 7.6, n = 5), but the differences were marginally significant [F = 3.03, 0.10

4 mm. Other important bivalves were Tellina sp. and Siliqua costata. Twelve polychaete taxa were identified, of which Nereis sp. was the most frequently occurring, while the most abundant were unidentified Spionidae. Fif- teen digestive tracts contained one or more speci- mens of brachyuran crabs, which in several cases were identified as young rock crabs, Cancer ir- roratus. Stomachs of the seven horseshoe crabs from the Point Pleasant series contained little food. Only four bivalves (one S. solidissima and three M. edulis), a gastropod (Nassarius trivittatus), and a brachyuran were identified. Polychaete setae were TABLE 5. — 1979 Limulus polyphemus survey results. Area means are expressed as the number of crabs per standard tow, as defined in the text. CV = coefficient of variation. Data were log transformed prior to Analysis of Variance. Area locations are shown in Table 1. Distance N sta- % with offshore Area tions crabs Mean CV Maximum 0-1.8 km 1 30 96.7 20.2 1.20 107.8 2 11 100.0 96 0.70 18.3 3 21 100.0 155 0.63 33.3 4 20 100.0 20.6 0.96 92 5 5 9 100.0 9.4 0.67 19.2 6 19 68.4 38 1.13 14.2 7 20 55.0 1.4 1.15 5.1 8 20 55.0 1.7 1.38 7.5 9 11 63.6 2.0 1.49 10.3 10 22 45.5 3.0 1.50 13.3 11 4 25.0 0.2 2.00 0.7 1.8-3.7 km 12 23 95.7 79 1.25 41.7 13 20 70.0 29 1.92 25.0 14 19 73.7 2.4 0.94 8.3 15 8 500 1.5 1.48 5.1 3.7-5.5 km 16 15 100.0 7.8 1.05 32.5 17 26 73.1 2.6 1.55 20.0 18 16 43.8 1.3 1.52 6.2 19 3 00 00 — 0.0 Analysis of Variance: Source df ss MS F p Total 312 401 42 Area 18 206.37 11.46 4.74 0.01 Time (area) 47 6230 1.33 2.46 0.01 Station (time (area)) 247 132.76 0.54 TABLE 6. — Ranking of food items by total abundance and frequency of occurrence, from 24 Limulus polyphemus collected in the Atlantic City series, summer 1978. Item Foraminifera Unidentified bivalve Spisula Tellina Brachyura Siliqua Spionidae Nematoda Cancer Fecal pellets Plant material Gemma Glycera Polychaete setae Ensis Polynoidae Mytilus Nereis Cirnpedia Spiophanes Ampharetidae Anomia Capitellidae Isopoda Mulmia Nemertea Ostracoda Turbellana Unidentified gastropod Unidentified oligochaete (Tie- 17 items) Number of specimens Rank Number of occurrences Rank 136 1 9 5 65 2 13 2 48 3 14 1 42 4 10 4 16 5 11 3 16 5 6 6 15 6 3 9 10 7 4 8 9 8 4 8 9 8 9 5 9 8 9 5 8 9 5 7 7 10 3 9 6 11 6 6 6 11 5 7 6 11 3 9 4 12 4 8 4 12 4 8 3 13 3 9 3 13 2 10 2 14 2 1 2 14 2 10 2 14 2 10 2 14 2 10 2 14 2 10 2 14 2 10 2 14 2 10 2 14 2 10 2 14 2 10 2 14 1 11 1 15 1 11 387 FISHERY BULLETIN: VOL. 82, NO. 2 noted in two samples and unidentified shells in three. The most numerous item was Foraminifera (n = 21), and no other item was found more than three times. In a laboratory experiment, a 20.3 cm (prosomal width) male horseshoe crab ate one 40.6 mm surf clam; the same crab consumed two clams, 43.8 and 42.4 mm, several days later (see Botton 1982 for procedural details). A 27.9 cm female ate two clams, 46.0 and 36.2 mm. Clams of this size are manipulated by the walking legs so that the ven- tral shell margin is held against the gnathobases. The chitinous gnathobases chip the ventral mar- gin, eventually resulting in the fracture of one of the valves. Cracking of the valves continues until the crab is able to remove the meat from the shell using the pincer-tipped walking legs or the chelicerae. Ingestion of the shell of 4 cm S. solidis- sima is apparently incidental. DISCUSSION A latitudinal gradient in horseshoe crab abun- dance along the New Jersey coast during the spring and summer months was recognized as a decrease in abundance with distance north from Delaware Bay, and an onshore-offshore gradient was apparent in northern New Jersey. The transi- tion between areas of high and low density takes place between Great Egg Harbor Inlet (Ocean City ) and Absecon Inlet (Atlantic City). Horseshoe crabs were more abundant inshore in the late spring and early summer than in the late summer and fall. Why are adult L. polyphemus concentrated in southern New Jersey, at least during the spring and summer? Since Delaware Bay, in southern New Jersey, contains the largest spawning popula- tion of horseshoe crabs in North America (Shuster 1982), we believe that the distribution on the New Jersey continental shelf may be related to the mi- gration of deep-water crabs to those beaches for reproduction. However, horseshoe crabs spawn elsewhere in New Jersey and are widely distrib- uted on the middle Atlantic continental shelf (Shuster 1979); based on electrophoretic evidence (Selander et al. 1970), there is gene flow between widely separated populations. Hydraulic surf clam dredges are efficient samplers of large benthic infauna (Meyer et al. 1981), but an evaluation of this dredge as a means of capturing L. polyphemus is lacking. Given its sluggish habits, it is unlikely that gear avoidance by horseshoe crabs significantly affects our re- sults; indeed, much more active lady crabs, Ovalipes ocellatus, are caught in large numbers ( Meyer et al. 1981; Haskin, unpubl. data). However, in the absence of direct observations, it is perhaps best to consider our results as relative, rather than absolute abundances of horseshoe crabs off New Jersey. Because the temporal sequence of sam- pling varied yearly and because the effect of time on abundance was statistically significant, we do not encourage speculation on year-to-year vari- ability based on these data. The horseshoe crab is a dietary generalist; based on the limited number of animals dissected, molluscs, arthropods, and polychaetes are the major food items. Although Foraminifera were numerous, they are probably ingested inadver- tently while digging out infauna. Opportunistic foraging was shown from the Stone Harbor group, which fed almost exclusively on M. edulis. Smith (1953) noted that crabs could locate discrete patches of soft-shell clam, Mya arenaria, but the behavioral basis for patch selection is un- known. Horseshoe crab predation may be an important source of juvenile surf clam mortality. In aquaria, crabs ingested only the meats of 4 cm S. solidis- sima\ this implies that this species may be more important as food than is apparent from visual stomach content analysis, which relies heavily on shell remains. Young S. solidissima may have been underestimated because many small (0.5-2.0 mm) shells were categorized only as "unidentified bivalves." Further studies of the food habits of horseshoe crabs, and of the abundance and diets of other predators, are necessary to evaluate the importance of predation in the survivorship of juvenile surf clams in New Jersey. ACKNOWLEDGMENTS We thank the many captains, mates, and Rut- gers University assistants for their help in the field, Michael Friedman for his aid in the statisti- cal analysis, and Carl Shuster and three anony- mous referees for helpful comments on the manu- script. This research was supported by grants from the National Marine Fisheries Service (contract 03-4-043-356) and the Surf Clam Inventory Fund of the New Jersey Department of Environmental Protection to the junior author, and from the James and Anna Leathern Fund. This is publica- tion #D-32503-l-83 of the New Jersey Agricul- tural Experiment Station. 388 BOTTON and HASKIN: DISTRIBUTION AND FEEDING OF HORSESHOE CRAB LITERATURE CITED BAPTIST, J. R, O. R. SMITH, AND J. W. ROPES. 1957. Migrations of the horseshoe crab, Limulus poly- phemus. in Plum Island Sound, Massachusetts. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 220, 15 p. BOTTON. M L. 1981. Food habits of breeding horseshoe crabs in Delaware Bay. Bull. N.J. Acad. Sci. 26:68. 1982. Predation by adult horseshoe crabs, Limulus polyphemus (L.), and its effect on benthic intertidal com- munity structure of breeding beaches in Delaware Bay, New Jersey. Ph.D. Thesis, Rutgers University, New Brunswick, N.J., 466 p. FOWLER, H. 1908. The king crab fisheries in Delaware Bay, and further note on New Jersey fishes, amphibians and reptiles. Annu. Rep. N.J. State Mus., 1907, Part 3, p. 349-432. HICKS, C. R. 1973. Fundamental concepts in the design of experi- ments. 2ded. Holt, Rinehart and Winston, N.Y., 349 p. LOCKWOOD, S. 1870. The horse foot crab. Am. Nat. 4:257-274. MEYER. T. L., R. A. COOPER, AND K. J. PECCI. 1981. The performance and environmental effects of a hy- draulic clam dredge. Mar. Fish. Rev. 43(9):14-22. PEARSON, F. C, AND M. WEARY. 1980. The Limulus amebocyte lysate test for endotox- in. BioScience 30:461-464. ROPES, J. W 1982. The Atlantic coast surf clam fishery, 1965-1974. Mar. Fish. Rev. 44(81:1-14. RUDLOE, A. 1980. The breeding behavior and patterns of movement of horseshoe crabs, Limulus polyphemus, in the vicinity of breeding beaches in Apalachee Bay, Florida. Estuaries 3:177-183. Selander, R., S. Yang, R. lewontin, and W. Johnson. 1970. Genetic variation in the horseshoe crab (Limulus polyphemus), a phylogenetic "relic." Evolution 24:402- 414. SHUSTER, C. N., JR. 1950. Observations on the natural history of the American horseshoe crab, Limulus polyphemus. Third report on investigations of methods of improving the shellfish re- sources of Massachusetts. Woods Hole Oceanogr. Inst., Contr. 564, p. 18-23. 1979. Distribution of the American horseshoe "crab," Limulus polyphemus (L.>. In E. Cohen (editor). Biomedi- cal applications of the horseshoe crab (Limulidae), p. 3-26. Liss, N.Y. 1982. A pictorial review of the natural history and ecology of the horseshoe crab Limulus polyphemus. with reference to other Limulidae. In J. and C. Bonaventura (editors), Physiology and biology of horseshoe crabs: Studies on normal and environmentally stressed animals, p. 1-51. Liss, N.Y. Smith, O. R. 1953. Notes on the ability of the horseshoe crab, Limulus polyphemus, to locate soft-shell clams, Mya arenaria. Ecology 34:636-637. Smith, O. R., and E. Chin. 1951. The effects of predation on soft clams. Mya arenaria Conv. Address Natl. Shellfish. Assoc. 1951, p. 37-44. Smith, O. R., J. R Baptist, and E. Chin. 1955. Experimental farming of the soft-shell clam, Mya arenaria, in Massachusetts, 1949-1953. Commer. Fish. Rev. 17(6):1-16. WOLFF, T. 1977. The horseshoe crab {Limulus polyphemus) in north European waters. Vidensk. Medd. Dan. Naturhist. Fo- ren. 140:39-52. ZAR, J. H. 1974. Biostatistical analysis. Prentice-Hall, Englewood Cliffs, N.J.. 620 p. 389 DIEL VARIATIONS IN THE FEEDING HABITS OF PACIFIC SALMON CAUGHT IN GILL NETS DURING A 24-HOUR PERIOD IN THE GULF OF ALASKA W. Pearcy,1 T. Nishiyama,2 T. Fujii,3 and K. Masuda3 ABSTRACT Changes in prey composition and stomach fullness indicate diel variations in feeding behavior of sockeye, pink, and coho salmon caught in surface gill nets set for 2 hours each over a 24-hour period at a station in the Gulf of Alaska. All of these species of salmon switched from feeding primarily on squids, fishes, and amphipods during the day to euphausiids at night. Apparently dense concentrations of euphausiids can be exploited by salmon in surface waters at very low light intensities, even during an overcast night. Day-night changes were less obvious in the food of chum salmon, which fed largely on salps. Total catches of salmon and catches in the near-surface portion of the gill nets were highest between sunset and sunrise, suggesting that diel vertical movements contribute to the higher night than day catches of surface gill nets. Although many studies have been published on the feeding habits of salmonids in oceanic waters of the North Pacific Ocean (Andrievskaya 1957; Allen and Aron 1958; LeBrasseur 1966; Ito 1964; Manzer 1968; Takeuchi 1972 ), most studies of daily feeding patterns have been conducted on juvenile salmon in fresh water or in coastal waters. These have generally shown that juvenile pink, sockeye, and chum salmon are diurnal or crepuscular (dawn and dusk) feeders (see Godin 1981 for re- view). The few studies conducted on diel feeding variations of adult or maturing Pacific salmon in oceanic waters of the northwestern Pacific Ocean have not revealed a consistent pattern (Machidori 1968; Shimazaki and Mishima 1969; Ueno et al. 1969). To further elucidate the diel feeding patterns of these fishes, we collected and examined stomach contents of four species of Pacific salmon caught in the Gulf of Alaska during one 24-h period. METHODS Two gill nets, each 800 m long and 6 m deep, with 300 m of 115 mm, 250 m of 121 mm, and 250 m of 130 mm (stretch) mesh, were alternately fished 'School of Oceanography, Oregon State University, Corvallis, OR 97331. institute of Marine Science, University of Alaska, Fairbanks, AK 99701. 3Faculty of Fisheries, Hokkaido University, Hakodate, Hok- kaido, Japan. for about 2-h periods over a 24-h period in the Gulf of Alaska from the Oshoro Maru, training ship of the Faculty of Fisheries, Hokkaido University, Hakodate, Hokkaido. The first net was set at 1200 h local time ( GMT - 9 h ) on 13 July; the last set was hauled at 1206 h on 14 July 1981 (Table 1). The time that the gill nets were fishing varied in the first 11 sets from 140 to 152 min (from start of set to start of TABLE 1. — Summary of gill net sets and catches for salmon. 13-14 July 1981. Start of set and Number of salmon Set Sock- Steel- no haul (h) eye Chum Pink Coho head Total 1 1200 1422 8 4 1 0 0 13 2 1400 1629 2 0 1 5 0 8 3 1600 1821 2 8 4 7 0 21 4 1800 2020 7 1 5 2 0 15 5 1957 2227 15 1 5 3 1 25 6 2158 0025 9 5 11 11 1 37 7 2359 0224 11 7 8 7 0 33 8 0159 0430 17 8 7 8 0 40 9 0358 0627 6 2 11 5 2 26 10 0600 0832 11 3 2 1 2 19 11 0758 1026 7 1 6 1 0 15 12 0957 1206 12 4 7 1 0 24 Total 107 44 68 51 6 276 Manuscript accepted November 1983. FISHERY BULLETIN: VOL. 82, NO. 2, 1984. 391 hauling) and 129 min in the last set. Five to eight minutes were required to set the nets, 12-20 min to retrieve them. The study area was between lat. 54° 51. 5' and 54°57.9'N, long. 144°55.1' and 14511.3' W. On con- secutive sets the gill nets were set 1.5-30 km apart to reduce the possibility of one net influencing the catch of another. Gill nets were set along a ship course of 040°, except for the first two nets which were set along 230°. In general, nets drifted 0.4-6.5 km northeastward during the sets. The vertical location in the gill net (upper, mid- dle, and lower 2 m) and species of each captured salmon were noted as the gill net was hauled aboard. Fish were removed from the gill nets, measured (fork length), and weighed with a beam balance. Stomachs were removed, weighed to the nearest gram with a beam balance, placed in a tray, and cut open with scissors. The fullness of cardiac and pyloric portions of the stomach was estimated visually as a) empty, b) trace amounts (few individual organisms with cumulative weights of a gram or less), c) ::•.. COHO S ^'•- CHUM - 1-5 6-8 9-12 200 TIME (hrs) -1 Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. FIGURE 1. — Average percent of the total catch of the four species of salmon caught in the upper 2 m of the gill net during afternoon (sets 1-5), nighttime (sets 6-8), and morning hours (sets 9-12) (upper panel), and the total number of all species of salmon caught in the upper 2 m of the gill nets per set during the 24-h period (lower panel). 392 PEARCY ET AL: DIEL FEEDING HABITS OF PACIFIC SALMON I5r 10 _J Q > 0 15 10- 5- SOCKEYE -CD [H Li. o cr UJ m PINK XH WW COHO "i n n m i i cl I0r CHUM u _cl n_ o 35 40 45 50 55 60 65 70 FORK LENGTH (cm) FIGURE 2. — Length-frequency histograms for the four species of salmon caught in the 24-h study. ages are represented for sockeye and chum salm- on. All pink and coho salmon were probably be- ginning their second year of ocean life. Compari- sons of length-frequencies between day (sets 1-4 and 10-12) and twilight and night (sets 5-9) were not significantly different for sockeye, pink, and chum salmon, but were significant for coho salmon (Kolmogorov-Smirnov test, P < 0.05). Coho salm- on were 2.4 cm larger in the twilight-night sets. Feeding Habits Stomach fullness of the four species of salmon, calculated as a percentage of body weight, were usually variable, ranging from 09c (empty) to a maximum of 49c for sockeye, 3.09c for chum, 3.39c for coho, and 2.39c for pink salmon (Fig. 3). Some individuals of all species had empty stomachs dur- ing most sets, regardless of time of day. Although ranked differences of the stomach weight:fish weight ratio were not significantly different be- tween day (sets 1-4 and 10-12) and night-twilight (sets 5-9) for each of the four species of salmon (Mann- Whitney U-test, P > 0.05), the highest per- centages of stomach weight to body weight for sockeye (>39c) and coho and pink salmon (>2%) were obtained from nighttime sets (Fig. 3). Moreover, our visual estimates of stomachs also indicated that full, distended stomachs of sockeye, coho, and pink salmon occurred only at night. There were no suggestions of diel periodicity of stomach fullness for chum salmon, however. The frequency of occurrence and percent com- position of the most common prey taxa (euphausiids, amphipods, squids, fishes) in the cardiac portions of salmon stomachs containing more than trace amounts of food are summarized in Table 2. All species of salmon consumed all of the four major categories of food. The most fre- quently occurring major taxa was euphausiids in sockeye and coho salmon, amphipods in pink salmon, and "other taxa" (mainly salps, but often unidentified material and sometimes pteropods and polychaetes) in chum salmon stomachs. Am- phipods were the second most frequent taxa in SET NUMBER 4- 3- 2- O O I 0- 3- .— I I UJ Q o o 12 3 4 5 6 7 8 9 10 II 12 l 1 i i 1 1 1 1 1 1 1 1 SOCKEYE " . . • •~ - . • - i : i . 1 J _:_ -.i - 1 — i — l i s — ' - - . CHUM - . ■ • : •J •i- • : - _ _ COHO v •• i • i. • - PINK , • . - • • - i- . .. i ■ 1800 " 0000 '|~0600 SS SR TIME(hrs) 1200 FIGURE 3. Stomach fullness, expressed as a ratio of food weight to fish weight, for the four species of salmon caught in each of the 12 gill net sets during the 24-h period. 393 FISHERY BULLETIN: VOL. 82, NO. 2 TABLE 2. — Frequency of occurrence and mean percent composition of major prey taxa in the gastric portions of salmon stomachs containing more than trace amounts of food. No. of % occurrence % volume Euphau- Amphi- Euphau- Amphi- stomachs siids pods Squids Fishes Other suds pods Squids Fishes Other Sockeye 92 73 40 17 18 17 53 18 9 14 5 Pink 35 49 63 17 17 11 37 35 9 14 4 Chum 27 41 37 4 7 78 29 10 2 4 56 Coho 30 63 33 43 3 0 54 3 39 4 0 sockeye, euphausiids in pink and chum salmon, and squids in coho salmon stomachs. The same taxa that ranked first and second on a frequency of occurrence basis usually ranked first and second on the basis of mean percent volume. Euphausiids (mainly Euphausia pacifica and Thysanoessa longipes) were most important for sockeye, pink, and coho salmon; "other taxa" were most important for chum salmon. Amphipods (mainly Parathemisto pacifica and Hyperia medusarum ) ranked second in sockeye and pink salmon. Gonatid squids ranked second in coho stomachs and euphausiids ranked second in chum salmon stomachs. Thus sockeye fed primarily on euphausiids and secondarily on amphipods and myctophid fishes. Pink salmon fed mostly on euphausiids and amphipods. Coho fed mainly on euphausiids and squids, and chum on salps and euphausiids (see Table 2). Squids comprised only 2% of the volume of the stomach contents of chum salmon, and fishes comprised only 4% of the vol- ume for chum and coho salmon. Copepods were not important (<1% of volume) for any species of salmon captured during the study. Dietary overlap, based on the sum of minimum percentage volumes (percent similarity index, PSI, Sanders 1960) of the four main prey taxa, was 78% between sockeye and pink, 69% between sockeye and coho, and 53% between pink and coho. Because chum salmon had the most unique diet of the four species consuming mainly salps and gelatinous zooplankton, they had overlap values of only 45% with sockeye and pink and 38% with coho. Although all species of salmon fed on a variety of taxa, individual fish usually contained only a few prey taxa. Only two major prey taxa were found in 85% , 89% , 93% , and 89% of the cardiac portions of sockeye, pink, coho, and chum salmon stomachs, respectively, containing more than trace amounts of food. Most sockeye and pink salmon had only one taxon of food in their stomachs. When salmon had only one food type in their stomachs, it was euphausiids in 65%, 52%, 85%, and 28%- of the individual sockeye, pink, coho, and chum salmon, respectively. Euphausiids were obviously the most important prey for sockeye, pink, and coho salmon during this study. They were often the exclusive prey. Sometimes the contents of the cardiac portion of sockeye and pink salmon stomachs were clearly divided with one type of prey in the anterior and one in posterior portion of the stomach. Generally this "stratification" involved euphausiids and am- phipods, or euphausiids and squid. Usually, how- ever, the cardiac and pyloric portions of the stomach had similar percentage compositions of major taxa (excluding empty stomachs and stomachs with trace amounts). Cardiac and pyloric contents were similar for 70% of the sock- eye, 72% of the pink, and 60% of the coho and chum salmon. When sockeye and coho had the same prey composition in cardiac and pyloric stomachs, both portions usually contained only euphausiids. When pink salmon had the same prey composi- tion, amphipods or euphausiids were found. The relative composition of major prey taxa in the stomachs of each species caught in the 12 gill net sets is illustrated in Figure 4 and is discussed below. Open circles in Figure 4 indicate when fresh prey were common, except for amphipods which usually showed little evidence of being digested. Sockeye Salmon Prey composition of sockeye salmon had a dis- tinctive diel pattern. Sockeye caught at night (2158-0430 h) contained a high percentage of euphausiids compared with the afternoon and morning sets (Fig. 4). In these night sets, euphausiids averaged over 80% of the volume of the stomach contents, and about 90% of the sock- eye contained only euphausiids. Fish caught dur- ing and after sunset (1957-0224 h) also contained large numbers of freshly ingested euphausiids. Some fish in set 5 (1957-2227 h) had a clear divi- sion between euphausiids in the fore portion and amphipods in the posterior portion of the cardiac 394 PEARCY ET AL: DIEL FEEDING HABITS OF PACIFIC SALMON 100 50 O SET NUMBER I 2 3 4 5 6 7 8 9 10 II 12 n 1 1 1 1 p 1 1 1 1 1 <~ SOCKEYE -fe^y" , "v;w^ 6 0 I 6 15 9 9 15 5 10 4 10 (NO. OF FISH) 100- 50- '-■■^■■' V CHUM 3 0 4 1 14 2 5 1 3 0 4 100 50 ° SO COHO 2 4 1800 | 0000 SS SR TlME(hrs) 0600 1200 sunset (2158-0430 h), when they comprised over 65% of the food and were often in fresh condition. All three fish with over trace amounts of food in the set that fished from 2359 to 0224 h contained 100% euphausiids. Squids and amphipods are most important in the afternoon (1600-2020 h), and fresh squids were found in stomachs of pink salmon caught from 1800 to 2020 h, just before sunset. Fishes and amphipods were the most important prey during the morning daylight period. Chum Salmon A diel trend for this species, which fed on a variety of prey taxa, was less obvious than for other species of salmon (Fig. 4). Salps composed over 75% of the stomach contents during the after- noon (1200-2020 h). Euphausiids were the most common prey taxa from sunset to the last set at midday, with the exception of a single chum salm- on caught at 2158-0025 h whose stomach con- tained many salps and a salmon caught at 0358- 0627 h whose stomach contained 95% fish. Most euphausiids in the stomachs of fish caught about the time of sunset (1957-2227 h) appeared to be recently ingested. Squids, which were only a minor part of the stomach contents, are not indi- cated in Figure 4. FIGURE 4. — Diel variations in the percent composition of major prey taxa in the stomachs of the four species of salmon contain- ing more than trace amounts of food. E = euphausiids, A = amphipods, F = fishes, SQ = squids, S = salps. Open symbols show when fresh prey was common. The number under each figure indicate the number of fish with more than trace amounts of food in their stomachs. SS = sunset; SR = sunrise. stomach, indicating a switch from amphipods to euphausiids during dusk. Euphausiids comprised <30% of the food in sets before sunset and after sunrise, and no fresh euphausiids were noted dur- ing these daytime periods. Amphipods and fishes formed the highest percentage of the food during daytime. Squids were also eaten by sockeye salm- on and were most important during late afternoon and sunset (1800-2227 h) and during sunrise (0350-0627 h). Fresh squids were noted in stomachs of fish caught in sets that fished from 1957 to 0025 h. Pink Salmon As with sockeye salmon, euphausiids attained peak importance as prey for pink salmon after Coho Salmon Coho salmon fed mainly on euphausiids during the night and on squids during the day. Euphausiids were not observed in stomachs of coho salmon during the afternoon but increased in importance from 0 to 100% of the stomach volumes between 1800 and 0240 h (Fig. 4). Most of the euphausiids during this period were in fresh con- dition. Euphausiids also comprised most (>60% > of the stomach contents during the morning hours (0159-0832 h) but were never fresh. Squids were the most important prey of coho salmon caught during the afternoon-daylight period and in the last set in late morning. Amphipods and fishes were of minor importance. DISCUSSION The larger catches of salmon in surface gill nets during twilight-night periods than in daytime periods have three possible explanations: Avoid- ance of nets during the daytime when visual acuity of salmon is highest, increased swimming 395 FISHERY BULLETIN: VOL. 82. NO. 2 activity in surface water at night compared with daytime, and diel vertical ascent of salmon into near-surface waters at night. The higher catches in the upper 2 m of the gill net at night than day lend support to the last possibility, but not to the exclusion of the other possibilities. Most other authors favor vertical migration as an explanation for diel peaks in gill net catches (Taguchi 1963; Manzer 1964; Mishimaet al. 1966). Birman (1964) noted visual avoidance of "sweep nets" by day, but concluded that salmon migrate into upper waters primarily as a response to verti- cal movements of their zooplanktonic prey which they feed on during periods of low light intensity, chiefly before dawn. Swimming activity could also influence catch- ability, but neither Ichihara et al. (1975) nor Ichi- hara and Nakamura (1982) found large differences in day-night swimming speeds of chum salmon tagged with ultrasonic transmitters. The most interesting finding of our study is the distinct diel change in composition of major prey. Stomach contents of sockeye, pink, and coho salm- on were comprised largely of euphausiids after sunset and during the night (Fig. 4). The largest number of full stomachs, usually containing only fresh euphausiids, were also found during the nighttime. These three species of salmon preyed intensively on euphausiids at night, often to the exclusion of other types of prey. This change to feeding on euphausiids was first observed in the salmon caught during the time that the 24-kHz sonic scattering layer ascended into surface waters (Fig. 5). We assume that euphausiids were an important component of this scattering layer (see Suzuki and Ito 1967). A 1.8 m Isaacs-Kidd midwater trawl collection (three mesh sizes: 70, 11, and 4 mm stretch) in the upper 10 m at night at the 24-h gill net station caught mainly salps and medusae, but euphausiids were abundant (19g/l,000 m3). Euphausiids were also abundant in a 1.3 m ring net (1.0 mm mesh) towed at the surface after sun- set at this station. The most common euphausiids caught were Euphausia pacifica and Thysanoessa longipes, the same species common in salmon stomachs. Euphausia pacifica were found to undertake diel vertical migrations at the Cana- dian Weather Station located at lat. 50°N, long. 145°W (Marlowe and Miller 1975). Frost and McCrone (1974) also found evidence for diel verti- cal migration of E. pacifica at this location but not for T. longipes. The intense predation on euphau- siids at night is therefore thought to be related to their dense concentration and increased vulner- ability in surface waters after dark. Most of the studies of the diel periodicity or chronology of feeding in salmon have been juveniles in fresh or coastal waters (Godin 1981). In general, these indicate diurnal or crepuscular feeding patterns for juveniles of pink salmon (Ali 1959; LeBrasseur and Barner 1964; Bailey et al. 1975; Parker and Vanstone 1966; Parker 1969; Godin 1981), chum salmon (Bailey et al. 1975; M. C. Healey as cited in Godin 1981), sockeye (Narver 1970; McDonald 1973; Doble and Eggers 1978), and coho salmon (Mundie 1971). Bailey et al. (1975) concluded that pink and chum salmon fry did not feed during cloudy moonless nights. Nighttime feeding by sockeye apparently occurs during moonlight but not on cloudy or moonless nights in Babine Lake (Narver 1970). Experiments con- ducted by Brett and Groot (1963) and Ali (1959) indicated that juvenile pink salmon changed their mode of capturing prey below 10° mc (meter can- dle), an intensity where the change from photo- topic to scotopic vision apparently occurs, and their feeding activity decreased between inten- sities of 10° to 10 4 mc and most ceased between 10~3 and 10 5 mc. Experiments by Bailey et al. (1975) showed almost no feeding by pink salmon fry at light Intensities below 101 mc. In our study, salmon fed intensively on euphausiids at night under an obscured, overcast sky. From the general data given by Brown (1952) and Blaxter (1970) we estimated that the light intensities on this night were between 10 3 and about 10 "5 mc. But, despite these low light inten- sities, with attendant reduction in contrast of prey and sighting range to prey (Eggers 1977; Anthony 1981), salmon were capable of actively feeding on small, euphausiid-sized prey. At night, larger prey such as squids and fishes are probably encoun- tered less frequently than euphausiids and evade capture more easily because of reduced sighting and tracking ranges of salmon. Euphausiids may not be as capable of active predator evasion and, when abundant in near-surface aggregations at night, are encountered frequently and actively selected. Bioluminescense produced by euphausiids may facilitate detection and capture by salmon. Thus, escape responses and sighting ranges at different light intensities may influence the size and type of prey selected at different times of a diel period. Machidori (1968) reported that the indices of stomach:body weight of sockeye and chum salmon caught in gill nets that fished different depths 396 PEARCY ET AI.: DIEL FEEDING HABITS OK PACIFIC SALMON CO CD o v a CD >> M _g cd -^ CO u c o Cfl CO o c cu u O) CO cu be C '% o 00 *3 CO 1— ( c c o CN CN CN O E c CO I- tUD O CM OS D o 397 FISHERY BULLETIN: VOL. 82, NO. 2 between 0 and 50 m in the northwestern Pacific were usually highest in near-surface depths by day and below 10 m at night. Euphausiids were an important food taxa only in salmon caught below 10 m during the day. Since average stomach full- ness indices were higher during the day than the night, Machidori concluded that light was neces- sary for salmon to feed. Takagi (1971) reported that surface longlines and gill nets caught salmon during morning and evening, but during the night salmon were caught in gill nets but not by longlines. These observations indicate reduced feeding activity of salmon at night. Shimazaki and Mishima (1969) concluded from diel trends in the feeding of maturing pink and chum salmon at two locations in the Sea of Okhotsk that feeding activity was high in the evening before and after sunset and low in day- time. They found peak stomach fullness values after sunset. In three of four instances these peak values were the first values after sunset, and may have been the result of crepuscular feeding. In one instance involving pink salmon, however, stomach fullness increased from 1917-2040 h to a peak at 2119-2245 h, indicating active feeding at night. Amphipods, squids, and fishes were the dominant food on a wet weight basis. Additionally, Ueno et al. (1969) found that pink and chum salmon had full stomachs during the late afternoon as well as after dark in waters off Kamchatka. Suzuki (1970) compared the volume of food in stomachs of chum salmon caught in gill nets off the Kamchatka Peninsula during night (2100-2330 h) and morning daylight hours (0330- 0610 h) and concluded that no major differences existed. He found that myctophid fishes always comprised a larger percentage of the stomach con- tents during the morning and pteropods usually comprised a larger percentage at night. Thus the above studies plus our own clearly document that salmon are capable of feeding dur- ing both day and night periods in oceanic waters. Their feeding behavior is flexible and variable, permitting opportunistic exploitation of a profit- able food resource regardless of when it is en- countered. ACKNOWLEDGMENTS We heartily thank all the crew and cadets of the Oshoro Maru who worked throughout the 24-h period of this study and made the research possi- ble. The research was supported in part by grants from the Sea Grant College Program (No. NA81AA-D-00086, Project No. R/OPF-17) and the National Science Foundation (INT80-00665/R- XMB0102). LITERATURE CITED ALI, M. A. 1959. The ocular structure, retinomotor and photo- behavioral responses of juvenile Pacific salmon. Can. J. Zool. 37:965-996. ALLEN, G. H., AND W. ARON. 1958. Food of salmonid fishes of the western North Pacific Ocean. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 237, lip. ANDRIEVSKAYA, l. d. 1957. Pitanie tikhookeanskikh lososei v severo-zapadnoi chasti tikhovo okeana (The food of Pacific salmon in the northwestern Pacific Ocean). [InRuss.] From: Materi- aly po biologii morskovo perioda zhizni dalnevostochnykh lososei, p. 64-75. Publ. by: Vses. Nauchno-Issled. Inst. Morsk. Rybn. Khoz. Okeanogr. (VNIRO), Mos- cow. [Translation: Fish. Res. Board Can. No. 821. | ANTHONY, P. D. 1981. Visual contrast thresholds in the cod Gad us morhua L. J. Fish Biol. 19:87-103. BAILEY, J. E., B. L. WING, AND C. R. MATTSON. 1975. Zooplankton abundance and feeding habits of fry of pink salmon, Oncorhynchus gorbuscha, and chum salm- on, Oncorhynchus keta, in Traitors Cove, Alaska, with speculations on the carrying capacity of the area. Fish. Bull., U.S. 73:846-861. BIRMAN, LB. 1964. Vertical migrations and vertical distribution of salmon in the sea. Dokl. Biol. Sci. Proc. Acad. Sci. USSR 156:346-349. BLAXTER, J. H. S. 1970. Light. Animals. Fishes. In O. Kinne (editor), Marine ecology, Vol. 1, No. 1, p. 213-320. Wiley- Interscience, Lond. BRETT, J. R., AND C. GROOT. 1963. Some aspects of olfactory and visual responses in Pacific salmon. J. Fish. Res. Board Can. 20:287-303. BROWN, D. W. E. 1952. Natural illumination charts. Dep. Navy, Bur. Ships Rep. 371-1, 11 p., 43 plates. DOBLE, B. D., AND D. M. EGGERS. 1978. Diel feeding chronology, rate of gastric evacuation, daily ration, and prey selectivity in Lake Washington juvenile sockeye salmon (Oncorhynchus nerka). Trans. Am. Fish. Soc. 107:36-45. EGGERS, D. M. 1977. The nature of prey selection by planktivorous fish. Ecology 58:46-59. FROST, B. W, AND L. E. MCCRONE. 1974. Vertical distribution of zooplankton and myctophid fish at Canadian Weather Station P, with a description of a new multiple net trawl. IEEE International Confer- ence on Engineering in the Ocean Environment, p. 159- 165. GODIN, J-G. J. 1981. Daily patterns of feeding behavior, daily rations, and diets of juvenile pink salmon (Oncorhynchus gorbuscha) in two marine bays of British Columbia. Can. J. Fish. Aquat. Sci. 38:10-15. 398 PEARCY ET AL: DIEL FEEDING HABITS OF PACIFIC SALMON ICHIHARA, T, AND A. NAKAMURA. 1982. Vertical movement of mature chum salmon con- tributing to improvement of set net structure on the Hok- kaido coast. Proc. North Pac. Aquaculture Symp., p. 39- 50. Univ. Alaska Sea Grant Rep. 82-2. ICHIHARA, T., T. YONEMORI, AND H. ASAI. 1975. Swimming behavior of a chum salmon, Oncorhyn- chits kt'ta, on the southern migration off Etorofu Island, the southern Kurile Islands. Bull. Far Seas Fish. Res. Lab. 13, p. 63-77. ITO, J. 1964. Food and feeding habit of Pacific salmon (Genus Oncorhynchus) in their oceanic life. Bull. Hokkaido Reg. Fish. Lab. 29. p. 85-97. (Fish. Res. Bd. Canada Transl. Ser. 1309, 1969. 1 LEBRASSEUR. R. J. 1966. Stomach contents of salmon and steelhead trout in the northeastern Pacific Ocean. J. Fish. Res. Board Can. 23:85-100. LEBRASSEUR, R. J., AND W. BARNER. 1964. Midwater trawl salmon catches in northern Hecate Strait November. 1963. Fish. Res. Board Can. Manuscr. Rep. Ser. 176, 1-7. MACHIDORI, S. 1968. Vertical distribution of salmon i Genus Oncorhyn- chus' in the Northwestern Pacific. III. Hokkaido Reg. Fish. Res. Lab. Bull. 34. p. 1-11. MANZER, J. I. 1964. Preliminary observations on the vertical distribu- tion of Pacific salmon (Genus Oncorhynchus) in the Gulf of Alaska. J. Fish. Res. Board Can. 21:891-903. 1968. Food of Pacific salmon and steelhead trout in the Northeast Pacific Ocean. J. Fish. Res. Board Can. 25:1085-1089. MARLOWE. C. J.. AND C. B. MILLER. 1975. Patterns of vertical distribution and migration of zooplankton at Ocean Station "P". Limnol. Oceanogr. 20:824-844. McDonald. J. 1973. Diel vertical movements and feeding habits of undervearling sockeye salmon (Oncorhynchus nerka I, at Babine Lake. B.C. Fish. Res. Board Can. Tech. Rep. 378, 55 p. MISHIMA, S.. S. SAITO. AND K. SHIMAZAKI. 1966. Study on the daily vertical movement of salmon - I. On the tendency of netting by the gill-net ll). Bull. Jpn. Soc. Sci. Fish. 32:922-930. MUNDIE. H. 1971. The diel drift of Chironomidae in an artificial stream and its relation to the diet of coho salmon fry, Oncorhyn- chus kisutch iWalbaum). Can. Entomol. 103:289-297. NARVER, D. W 1970. Diel vertical movements of feeding of underyearling sockeye salmon and the limnetic zooplankton in Babine Lake, British Columbia. J. Fish. Res. Board Can. 27:281-316. Parker, r. r. 1969. Foods and feeding of juvenile pink salmon in Central British Columbia waters. I. 1966 diel series. Fish. Res. Board Can. Manuscr. Rep. 1017. Parker, r. R., and W. E. Vanstone. 1966. Changes in chemical composition of Central British Columbia pink salmon during early sea life. J. Fish. Res. Board Can. 23:1353-1384. Sanders, H. L. I960. Benthic studies in Buzzards Bay III. The structure of the soft-bottom community. Limnol. Oceanogr. 5:138- 153. SHIMAZAKI, K., AND S. MISHIMA. 1969. On the diurnal change of the feeding activity of salmon in the Okhotsk Sea. Bull. Fac. Fish. Hokkaido Univ. 20:82-93. SUZUKI. T. 1970. Swimming behaviour of chum salmon in the south- eastern area off Kamchatka Peninsula. Bull. Jpn. Soc. Sci. Fish. 36:19-25. SUZUKI. T. AND J. ITO. 1967. On the DSL in the northwestern area of the North Pacific Ocean - I. Relationship between vertical migra- tion of DSL. submarine illumination and plankton bio- mass. Bull. Jpn. Soc. Sci. Fish. 33:325-337. TAGUCHI, K. 1963. Some factors having effects on the behavior of sal- mon in the time of gill-netting. Bull. Jpn. Soc. Sci. Fish. 29:434-440. TAKAGI. K. 1971. Information on the catchable time period for Pacific salmon obtained through simultaneous fishing by longlines and gillnets. Bull. Far Seas Fish. Res. Lab. 5, p. 177-194. TAKEUCHI. T. 1972. Food animals collected from the stomachs of three salmonid fishes {Oncorhynchus) and their distribution in the natural environments in the northern North Pacific. Hokkaido Reg. Fish. Res. Lab. Bull. 38:1-119. L'ENO. M., S. KOSAKA, AND H. USHIYAMA. 1969. Food and feeding behavior of Pacific salmon - II. Sequential change of stomach contents. Bull. Jpn. Soc. Sci. Fish. 35:1060-1066. 399 ARCTIC CHAR PREDATION ON SOCKEYE SALMON SMOLTS AT LITTLE TOGIAK RIVER, ALASKA1 Gregory T. Ruggerone and Donald E. Rogers2 ABSTRACT Observations of Arctic char feeding on migrating sockeye salmon smolts at Little Togiak River, Alaska, indicate a Type II functional response where the number of smolts consumed increased with smolt abundance. The number of smolts migrating was usually low <■ 20,000 smolts/24 hours) and the corresponding consumption of smolts averaged 0.8 smolts/char per 24 hours. When large smolt migra- tions occurred ( -80,000 smolts/24 hours), char consumption of smolts generally increased to 5.6 smolts/char per 24 hours. In addition to smolt abundance, smaller smolts and longer char were correlated with an increase in the number of smolts consumed. Estimates of percent smolt mortality, based on two hypothetical char numerical responses to varying smolt abundances, indicate that smolts were migrating at densities most susceptible to predation. A comparison of length of smolts consumed by char with those in the migration shows that char consumed larger than average smolts when their stomachs were not full and smaller than average smolts when char approached stomach fullness. This may be explained by the migration of larger smolts during the feeding period of char and the possibility of char feeding less effectively when approaching fullness. Although major hatchery releases often exceed 100,000 smolts per day, these data suggest that hatchery-released smolts may be less susceptible to predation in small rivers when released during the night in large numbers (> 20,000 smolts/24 hours). The relationship between predation on juvenile salmon and relevant biological and environmen- tal factors is important to the understanding of salmon population dynamics. Development of these relationships may be useful for establish- ment of "optimal" escapement levels and for maximum production from salmon enhancement projects. A few investigations have related preda- tion rates to juvenile salmon abundance and have reported up to 85% juvenile mortality (Neave 1953; Hunter 1959; Parker 1968; Peterman and Gatto 1978). Other investigations have examined the ef- fect of biological or environmental variables such as juvenile salmon size ( Parker 1971 ), predator size (Ricker 1941; Hunter 1959; Rogers et al. 1972), infection by parasites (Burke 1978), river velocity and turbidity (Ginetz and Larkin 1976), thermal stress (Sylvester 1972; Coutant 1973), or several variables independent of juvenile salmon density (Fresh et al. 1980). No investigation has analyzed predation while concurrently assessing the par- tial effect of prey density along with the partial effect of biological and environmental factors. This investigation represents a 5-yr study of predation by Arctic char, Saluelinus alpinus, on 'Contribution No. 646, School of Fisheries, University of Washington, Seattle, Wash. 2Fisheries Research Institute, WH-10, University of Washing- ton, Seattle, WA 98195. emigrating sockeye salmon, Oncorhynchus nerka, smolts at Little Togiak River, Alaska (Fig. 1). Predator-prey interaction appears to be especially refined in this river. With the onset of the smolt emigration each spring, char migrate to the inter- connecting rivers in the lake system where mi- grating smolts are most vulnerable (McBride 1979). After the smolt migration ends, char return to their spawning streams in the fall. The objec- tives of this investigation, which were tested dur- ing this brief period of predator-prey interaction, were 1) to empirically model the daily functional response of char (i.e., the relationship between smolt abundance and number of smolts consumed/char per 24-h period (Fig. 2)) while con- currently measuring the effect of biological and environmental variables; 2) to estimate percent smolt mortality in relation to smolt abundance; and 3) to test for disproportional consumption of large or small smolts by char that differ in stomach fullness and fork length. Numerous biological and environmental vari- ables are likely to influence char predation on salmon smolts. The variables concurrently tested in the functional response model were 1) the num- ber of migrating smolts during the 24-h period prior to sampling the char; 2) the number of mi- grating smolts during the 24-48 h period prior to sampling; 3) smolt weight; 4) char length; 5) Manuscript accepted November 1983. FISHERY BULLETIN: VOL. 82, NO. 2, 1984. 401 FIGURE l.— Little Togiak River in relation to the Wood River lake system and the State of Alaska. number of smolts migrating during the daylight; 6) percent of migration during the daylight; 7) presence of adult sockeye in the river; 8) days after ice-out; 9) river temperature; 10) river depth; 11) light intensity at dusk; and 12) incident solar radiation. METHODS Description of Study Site Little Togiak River is a small nonturbid river located in the Wood River lakes system, Alaska (Fig. 1). River length is about 200 m as it flows from the smaller Little Togiak Lake to the larger Lake Nerka. River width is about 20 m and average depth ranges from about 2 m during spring high 402 water to 0.25); however, there were significant differences in diets (numbers and volumes) between summer (July-September) 1980 and summer 1981 (P < 0.05) and between the spring and summer of each year (P < 0.01). We examined seasonal changes in IRI to detect which prey could account for significant differ- ences in walleye diet (Table 4). From 1980 to 1981 there is a reduction in the importance of sal- monids, cottids, and shad, and an increase in importance of catostomids, cyprinids, and inver- tebrates. Seasonal changes are not consistent; however, there are reductions in importance of cottids, increases in importance of cyprinids, and no changes in importance of invertebrates from spring to summer each year. Our CPUE data (Table 5) reflect annual and seasonal changes in the abundance of juvenile shad, juvenile TABLE 3. — Percent by volume, percent by number, and percent frequency of occurrence of foods found in the stomachs of walleye collected in the John Day pool of the Columbia River, April-September 1981. Sample size equals 236 walleye, with 39. (Fr empty stomachs. (Raw data are in parentheses.) Prey taxon % volume (ml)1 % number' % occurrence Salmonidae (juvenile) 3.6 (62) 4.4(14) 7.0(10) Oncorhynchus Ishawytscha 2.8 (48) 3.2 (10) 14 (2) Unidentifiable Salmonidae 0.8 (14) 1.2 (4) 5.6 (8) Castostomidae 32.5 (563) 11.4(36) 18.1 (26) Catostomus columbianus 11.6 (201) 2.5 (8) 4.2 (6) C macrocheilus 1.2 (21) 0.6 (2) 1.4 (2) Unidentifiable Catostomidae 19.7 (321) 8.3 (26) 12.5 (18) Cyprinidae 34.1 (590) 13.0(41) 25.7 (37) Acrocheilus alutaceus 28.3 (490) 5.7 (18) 11.1 (16) Mylocheilus caurinus 1.8 (32) 2.5 (8) 5.6 (8) Ptychocheilus oregonensis 1.7 (30) 1.6 (5) 2.8 (4) Cyprinus carpio 0.3 (6) 0.3 (D 0.7 (1) Carassius auratus 0.5 (8) 0.6 (2) 1.4 (2) Unidentifiable Cyprinidae 1.5 (24) 2.3 (7) 4.1 (6) Miscellaneous fishes 29.3 (508) 58.2(184) 77.8(112) Cottus asper 22.5 (390) 25.6 (81) 36.8 (53) Alosa sapidissma (juvenile) 0.1 (1) 0.3 (1) 0.7 (1) Ictalundae 0.2 (4) 0.3 (D 0.7 (1) Unidentifiable 6.5 (113) 32.0(101) 39.6 (57) Invertebrates 0.5 (8.23) 13.0(41) 11.1 (16) Ephemendae 0.3 (5.98) 12.0 (38) 10.4 (15) Chironomidae 0.01 (0.20) 0.3 (D 0.7 (1) Gammandae <0.01 (0.05) 0.3 (1) 0.7 (1) Astacidae 0.1 (2.00) 0.3 (1) 0.7 (1) 1 Volumes and numbers of individual prey taxa were significantly different from those of 1 980 (P < 0.005). 413 TABLE 4. — Index of Relative Importance (IRI) (Pinkas et al. 1971) of foods found in spring (April through June) and summer (July through September) 1980 Prey taxon Combined' Spring Summer Salmonidae 182 (1.7) 293 (2.5) 148 (1.2) Oncorhynchus Ishawytscha 37 (0.9) 99 (1.7) 104 (2.4) Unidentifiable Salmonidae 42 (1.0) 51 (0.9) 4 (0.1) Catostomidae 545 (5.0) 548 (4.6) 749 (6.1) Catostomus columbianus 40 (0.9) 12 (0.2) 121 (2.7) C macrocheilus 36 (0.8) 12 (0.2) 66 (1.5) Unidentifiable Catostomidae 113 (2.6) 226 (3.9) 37 (0.8) Cyprinidae 250 (2.3) 98 (0.8) 491 (4.0) Acrocheilus alutaceus 37 (0.8) 2( ■0.1) 162 (3.7) Mylocheilus caurinus 34 (0.8) 62 (1.1) 16 (0.4) Ptychocheilus oregonensis 1 (<0.1) 4 (0.1) Other Cyprinidae 3 (0.1) 6 (0.1) Miscellaneous fishes 9,879(9 10,859(9 10,795(88.3) Cottus asper 2,090 (48.6) 3,621 62.8) 1,335 (30.7) Alosa sapidissma 91 (2.1) 480 (11.1) Other (unidentifiable fish; Lampetra spp.. Ictalundae) 1,784 (41.3) 1,677 (J 2,012 (46.4) Invertebrates 34 (0.3) 74 (0.6) 40 (0.3) 1 1RI s are not additive across columns TABLE 5. — Catch-per-unit-effort (CPUE) for various juvenile (Juv. i and adult fishes caught in the John Day pool of the Colum- bia River, April-September 1980-81. Semes Dates Effort Sets Juv Chinook CPUE Juv shad Juv peamouth Apr-Jun 1980 Apr-Jun 1981 Jul -Sept 1980 Jul -Sept 1981 Gill nets 45 37 35 39 16 65 10 70 2.65 1.36 0 0 9276 42.87 0 0 688 5.77 Largescale Bridgelip Dates Hours Chiselmouth sucker sucker Apr-Jun 1980 122 0.23 0 86 0.40 Apr-Jun 1981 212 0.32 090 0 67 Jul -Sept 1980 330 0.37 0.54 0 74 Jul. -Sep 1981 154 031 0.84 0.61 peamouth, and juvenile chinook salmon. However, Sims et al. (1982) reported no significant sea- sonal differences in the estimated numbers of juvenile salmonids emigrating past the John Day Dam (53,000 and 44,000 daily from 21 April to 30 June and from 1 July to 28 September 1981, respec- tively), 90 km downstream of our study area. Simi- lar estimates are not available for 1980, but the smolt emigration past the John Day Dam was estimated at 8.3 million (Sims et al. 1981) and 7.7 million (Sims et al. 1982) in 1980 and 1981, respec- tively. Unfortunately cottids, the most important food, were rare in our CPUE data for 1980, and in 1981, the electroshock CPUE was <0.1, a level too low to detect changes. Seasonal shifts in walleye diets are often the result of high spring-time availability of aquatic insects and/or increased availability of prey fish in the summer (Eschmeyer 1950; Parsons 1971). In this study, invertebrates represented 4-13% of the numbers of prey items (Tables 2, 3), however they contributed little to the total caloric intake of the walleye because of their almost negligible volume and poor assimilation by walleye (Kelso 1972). Moreover, invertebrates did not exhibit significant seasonal variation in walleye dietary importance (Table 4). Diel Periodicity The mean index of fullness, measured as the volume of stomach contents (ml) divided by wall- eye body weight (kg), for all walleye sampled is plotted against time of capture (2-h intervals) in Figure 3. The shape of this curve suggests the '_ 24.0 fZD-Salmonids QShad 12.0 - 1000 1400 Sunrise 1800 2200 Sunset Time (hours) FIGURE 3. — Index of fullness and numbers of juvenile salmonids and shad consumed per walleye captured during 2-h intervals. Data collected from the John Day pool of the Columbia River, April to September 1980 and 1981. (Sample size in parentheses.) 414 stomachs of walleye collected in the John Day pool of the Columbia River in the 1980 and 1981. Numbers in parentheses are percent IRI. 1981 Prey taxon Combined' Spring Summer 56 (0.6) 53 (0.5) 58 (0.6) 8 (0.2) 4 (0.1) 11 (0.3) 26 (0.5) 58 (1.5) 795 (8.8) 88 (0.9) 816 (8.3) 59 (1.4) 65 (1.1) 69 (1.7) 3 (0.1) 1 (<0.1) 8 (0.2) 350 (8.4) 418 (7.3) 293 (7.3) 1.211 (13.4) 649 (6.3) 2,776 (28.4) 377 (9.0) 462 (8.1) 298 (7.5) 24 (0.6) 1 (<0.1) 162 (4.1) 9 (0.2) 2(<0.1) 41 (1.0) 34 (0.8) 1 (<0.1) 234 (5.9) 6.808 (75.5) 9.480 (9 6,028(61.6) 1,770 (42.2) ?,376 (41.7) 1.314 (32.9) <1 (<0.1) 2 (<0.1) 1.544 (36.8) 2.336 (41.0) 1.511 (37.9) 150 (1.7) 113 (1.1) 106 (1.1) Salmonidae Oncorhynchus tshawytscha Unidentifiable Salmonidae Catostomidae Catostomus columbianus C macrocheilus Unidentifiable Catostomidae Cypnnidae Acrocheilus alutaceus Mylocheilus caurmus Ptychocheilus oregonensis Other Cypnnidae Miscellaneous fishes Cottus asper Alosa sapidissma Other (unidentifiable fish; Lampetra spp . Ictalundae) Invertebrates same bimodal feeding periodicity as reported for other walleye populations during times of high prey densities (Swenson 1977). We found no an- nual or seasonal variation in this periodicity. Numbers of juvenile salmonids and shad con- sumed per walleye at various times of the day peak from late night to midmorning, drop to a low level at midday, and remain low through the evening peak in walleye feeding. Size of Prey Consumed Parsons (1971) showed a positive relationship between walleye length and length of prey con- sumed in Lake Erie. Walleye in the mid-Columbia River exhibit the same relationship, and size of prey is correlated to different prey taxa. The change in the percent of the IRI of various prey groups, as a function of walleye fork length, is 100 KS2S2828ogo8ogogogogogogogogogogo^^^ Walleye Fork Length FIGURE 4.— Change in percent of total Index of Relative Importance (IRI) (Pinkas et al. 1971) of prey components as a function of walleye fork length (100 mm increments). Walleye collected in the John Day pool of the Columbia River, April through September 1980 and 1981. (Sample size in parentheses.) 415 charted in Figure 4. Small walleye (200-400 mm FL) primarily consume salmonids, cottids, and shad, while midrange walleye (400-600 mm FL) rely more heavily on cyprinids, cottids, and catos- tomids. For large walleye ( >600 mm FL), suckers are the most important prey and the importance of cyprinids and cottids is reduced. Figure 5 contains the length frequencies of walleye prey collected in 1981 and shows peaks which correspond to the size of walleye most likely to consume that prey, i.e., cottids, juvenile shad, and juvenile salmonids are small (25-125 mm FL); cyprinids, excluding juvenile peamouth, are midrange in length (125- 300 mm FL); and catostomids are present in a large range of sizes (150-450 mm FL) with peaks 300 mm FL. 40 30 c m u Si Q. 20 - 10 - |\ . />\/\A/' A i • / \SJ3 \° /R o Peamouth (n 362) • Chiselmouth (n = 553) a Largescale Sucker (n =1211) □ Bridgelip Sucker (n = 775) ,D ££.A < **&*?. jmbk V iP-n LtA^ c (J 0.05). Sex ratios gener- ally were equal from 30 to 80 mm, but females increasingly predominated at larger sizes (Fig. 6). Sex ratios did not differ significantly from a 1:1 ratio until S. empusa exceeded 80 mm (Table 3). Maximum Size and lntrayear Variations in Size Typical maximum size reached by S. empusa in the northwestern Gulf is 110-115 mm. The largest of 9,400 specimens we captured was only 132 mm, TABLE 3. — Observed sex ratios and chi-square statistics for Squilla empusa divided into 10 mm length groups, July 1979-June 1980. Asterisks indicate significant x2 at a = 0.05. Length range (mm) No. individuals Male Female X2 21-30 3 1 1.00 31-40 28 29 0.02 41-50 117 138 1.73 51-60 381 382 0.00 61-70 835 837 0.00 71-80 1.353 1,357 0.01 81-90 1.069 1.339 30.27* 91-100 387 699 89.64* 101-110 94 241 32.25* 111-120 19 69 28.41* 121-130 4 16 7.20* 131-140 — 2 2.00 Total 4.290 5,110 — 99% were <110 mm, and 99.5% were <114 mm (Fig. 7). Size compositions showed little change throughout the sampling period. Except for the 422 200 CO CO CNJ V 100-^ JUL 79 n = 443 J 200n 100- o OS 12 3 4 AUG 79 n = 1 r— — i — i — r- 12 3 4 £ 200 100 lO CO CNJ i I SEP 79 n = 269 f ». " 12 3 4 !UU- CD OCT 79 co / / n = 814 00- n '4 ~™ 2 3 4 200 100- N0V79 n = 370 1 0 T r i 200 12 3 4 DEC 79 n = 122 100-2 1 12 3 4 uu- l-~ JAN 80 co n = 508 00- 0 i 1 i ? J 12 3 4 200 100 Ol FEB 80 I n = 631 I g \ 1 2UU- m [ MAR 80 CO co n = 623 100- 1 ,111 200 1 12 3 4 APR 80 n = 283 12 3 4 200 CO MAY 80 100- | n = 592 1 0 £4 y, 12 3 4 200 100- 0 CO JUN80 n = 505 vx 7- 4 -A 12 3 4 12 3 4 Maturity Stage 200 1 100- JUL80 n = 173 v\ mx V, 12 3 4 200 1 100 P2 AUG 80 n = 245 fc v n vj/T/y, 200 100 2 3 4 SEP 80 n = 58 V 4 r " i"^-* 12 3 4 -.' NO. 2, 1984. 427 92°00 31°003 30° 29° S8°00' 87°00' 30° ?9° 28' 00' 92°00' 87°00' I- K.I KE 2.— Area of the Gulf of Mexico sampled for southern kingfish during the fall offshore cruise (25 October 1979 through 14 November 1979' and during the spring offshore cruise (10-18 April 1980). Grids represent station sites established by the National Marine Fisheries Service. temperatures of the waters in these regions are influenced greatly by input from rivers and fresh- water runoff. The estuaries drain into the Missis- sippi Sound which is bordered on the south by barrier islands. A complete description of the Mis- sissippi Sound is given by Eleuterius (1978). Areas sampled varied geomorphologically with location. Bottom types ranged from mud and silt in the rivers and upper estuaries to sandy tidal zones on the north side of the barrier islands. Overall depths varied from 0.3 m at seine stations to 91.4 m at offshore trawl stations. Sample areas in the estuaries ranged from 0.3 m to 12.2 m with rela- tively small tidal variations (0.1-0.6 m). with the Fisheries Monitoring and Assessment Program at the Gulf Coast Research Laboratory, Ocean Springs, Miss. (Christmas 1978). The type of gear used was dictated by the bottom topography and geographic location. At five sta- tions along the barrier islands and in the es- tuaries, a 15.2 m bag seine with 6.4 mm bar mesh was used to collect juveniles. Larval and postlar- val fish were collected by towing a Renfro beam plankton net (BPL) with 50 holes/cm2 and a 1.8 m diameter mouth, in a 45.7 m semicircle at five stations. During the 10-min tows, young-of-the- year and adult southern kingfish were sampled with a 4.9 m standard otter trawl with 19.9 mm Materials and Methods Monthly collections of southern kingfish were made from October 1979 through September 1980 (Fig. 1, Table 1), with the exception of stations 93 through 98 which were sampled monthly from February 1980 through September 1980. Twenty- eight of the stations were sampled in conjunction TABLE 1. — Bottom salinities at monthly stations where south- ern kingfish were collected. Maximum Minimum Average Number of salinity salinity salinity observations (%.) (%.) (%.) Rivers 10 34.5 9.0 26.1 Estuaries 27 27.5 2.0 17.1 Mississippi Sound 37 33.0 14.5 24.7 Barrier Islands 29 29.0 5.0 22.2 Offshore 35 33.5 6.0 26.7 428 mesh, 6.4 mm tail mesh, and 0.9 m doors at 18 stations in the Mississippi Sound and with a 12.2 m standard otter trawl with 19.1 mm mesh, 6.4 mm tail mesh, and 1.3 m doors at 6 stations in the passes and outside the barrier islands. One sample was taken at each station per month, except at Fort Point (station 30) where two samples were taken: one with the 15.2 m seine and one with the BPL. In the fall of 1979 and the spring of 1980, collec- tions were made during National Marine Fisheries Service groundfish cruises (U.S. De- partment of Commerce 1979, 1980). Samples were taken by the RV Oregon II in depths of 9.1-91.4 m (5-50 fathom lines) from Mobile Bay, Ala., to Ship Shoal, La., extending to the east and west beyond Mississippi coastal waters (Fig. 2). Three succes- sive 10-min tows were made at each station with a 12.2 m standard otter trawl with 19.1 mm mesh, a 1.22 m vertical opening, and 2.44 m doors, and samples were randomly taken. Specimens were frozen on board the ship, except ripe females which were preserved in 10% Formalin2 for future use in fecundity estimates. In the laboratory, each fish was weighed to the nearest 0.1 g and standard length measured to the nearest 1 mm. If the sample exceeded 50, a sub- sample of 50 was weighed and measured individu- ally. The remainder of the catch exceeding 50 was counted and gross weight taken. Salinity was determined with a refractometer (accuracy of IX) at stations sampled monthly, and on the groundfish cruises by titration at the Na- tional Marine Fisheries Service Laboratory (U.S. Department of Commerce 1979, 1980). Tempera- ture was taken with a YSI Model 54 oxygen meter at monthly stations and with a centigrade ther- mometer on the groundfish cruises. Where a trawl was used, bottom water samples were collected with a Niskin or a Kemmerer bottle depending on the location and vessel used. The length-weight relationship was calculated for southern kingfish by following the procedure of Rounsefell and Everhart (1953). Results Geographic Range The distribution of southern kingfish in Missis- sippi extends from as far north as Bayou Bernard 2Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. (station 36) to offshore beyond the barrier islands (Fig. 1). Fish were captured at all monthly sample sites at some time during the year except at seven stations where water depth ranged from 0.3 to 19.5 m. During the fall groundfish cruise, southern kingfish were captured from 4.8 km south of the barrier islands to below the Mississippi River Delta and from long. 88°10'-91°25'W (Fig. 2). Dur- ing the spring cruise the geographic range was about the same; however, populations were denser near the Chandeleur Islands and at the southern tip of the Mississippi River Delta. Catches of shrimp, a primary food of southern kingfish, were high in these areas, indicating the fish could have concentrated because of an abundant food sup- ply- Size of the southern kingfish captured varied with the geographic location. Specimens 50-150 mm SL frequented estuaries and inshore waters, while adults exceeding 150 mm SL were found on the Gulf of Mexico side of the barrier islands and offshore in waters as deep as 69.5 m. Distribution and size of the fish captured were also related to gear selectivity — trawls were used to capture adults in the passes and offshore, whereas smaller mesh gear types were more effective for obtaining juveniles and young-of-the-year in estuaries and inshore waters. Trawls accounted for 42r/c of the total catch at monthly sample sites. Seasonal Distribution A total of 1,554 southern kingfish was captured from October 1979 through September 1980. The numbers of fish gradually increased from De- cember with a peak in February; from March through September there was no apparent numer- ical pattern (Fig. 3). On the other hand, the monthly catch by weight peaked in April with a lesser peak in January. Weight gradually declined during the remainder of the year. Although overall catches for January and Feb- ruary were high, stations in the passes between the barrier islands and offshore accounted for the majority of southern kingfish in these months. Few southern kingfish were found in the estuaries during winter, and no kingfish were landed at any of the offshore stations during May, June, and July. During the fall offshore groundfish cruise, 121 southern kingfish (100-280 mm SL) were captured. The spring cruise was more productive with 307 fish (70-270 mm SL) taken. No previous data on 429 MONTHS (OCTOBER 1979 THROUGH SEPTEMBER 19801 FIGURE 3. — Summary by number and by weight of the catch data for southern kingfish at stations sampled monthly from October 1979 through September 1980. the distribution of this species in these offshore areas of Mississippi are available for comparison. Salinity Range Southern kingfish were captured in waters ranging in salinity from 2.0 to 36.61., with the majority found above 20.0 (Table 1). The mean salinity for the Mississippi coastal waters where southern kingfish occurred at regularly sampled stations was 23.4/.. During the fall groundfish cruise, bottom salinities where southern kingfish were taken ranged from 31.4 west of the Missis- sippi River Delta to 36.6 south of Horn Island, with an average bottom salinity of 34.51 . No bot- tom salinities were available during the spring offshore cruise because of equipment problems. Temperature Range Young-of-the-year and adult southern kingfish were captured in waters with bottom tempera- tures ranging from 8.0°C in December to 37.3°C in August. The largest catches were taken in waters ranging from 20.0° to 30.0°C. Larval and postlar- val fish (0.5-20.0 mm SL) were found from May through November in the shallow inshore waters and northern tidal zones of the barrier islands at temperatures ranging from 12.0° to 37.3°C. The bottom water temperatures where southern kingfish were taken for the fall groundfish cruise ranged from 22.4° to 25.5°C (average of 23.2°C) and for the spring cruise ranged from 22.4° to 25.5°C (average of 22.3°C). In general, bottom temperatures on offshore cruises decreased with increasing depths. Length- Weight Relationship The length-weight relationship was calculated for the 1982 southern kingfish (ranging from 0.5 to 291.0 mm SL) by the following regression equa- tion: Log W = -4.48683 + 2.92908 Log L where W = weight in grams and L = standard length in millimeters (Fig. 4). The coefficient of determination R2 was 0.9779. Length-Frequency Data At stations sampled monthly, southern kingfish juveniles (<50 mm SL) were most common from May through October; fish in the 100-150 mm SL range were fairly constant all year with a peak in March; and fish in the 150-250 mm SL range were rare during all months except April (Fig. 5). August was the only month where fish 10 mm SL or less were captured. The majority of the fish cap- tured during the two offshore cruises were > 100 mm SL with an average of 174 mm SL (Fig. 6). The mean sizes of southern kingfish taken at the monthly stations were generally much smaller than those from the offshore groundfish cruises because young-of-the-year utilize the estuaries and inshore waters as nursery grounds. Gear selectivity must also be taken into account be- cause the larger trawls used offshore are ineffi- cient for capturing juveniles. Discussion Young-of-the-year frequented estuaries and in- shore waters, while adult southern kingfish were found to be more abundant offshore in deeper waters. This was also reported by Pearson (1941) for the Chesapeake Bay area and by Geagan ( 1962 ) for the coastal waters of Louisiana. Irwin (1970) found the most common habitat for juveniles <50 mm SL to be open surf on sandy beaches, whereas in this study they occurred most often in sandy tidal zones of the barrier islands and estuaries. High offshore catches and very low inshore catches from December through April support reports by Gunter (1938, 1945) and Christmas and Waller (1973) that adult southern kingfish migrate offshore during winter months and are summer residents of the estuaries. Similar reports were made by Mcllwain (1978) for recreational 430 2.731 -, 2340 - 1.948 - X \- cd z LU _l CD o 1.557- 1.165- 0.774 1.843 1.970 2.096 2.223 LOG WEIGHT 2.349 2476 FIGURE 4. — Length-weight regression for all southern kingfish caught from October 1979 through September 1980. catches from Biloxi Bay, Miss., and by Miller (1965) while trawling off Port Aransas, Tex. The salinity range of 2.0-36.61 is slightly broader than the range of 5.0-35.5 "L reported by Christmas and Waller (1973). Loman (1978) re- ported that the highest catches of southern kingfish in Mississippi were between 15.0 and 30.0 and that mean length increased as salinity increased. The latter was also true in this study since larger fish were captured offshore where salinities were higher, and only postlarval and young-of-the-year fish were found in salinities below 15.0 . I concur with most authors that this species is eurythermal (Gunter 1945; Franks 1970; Christ- mas and Waller 1973; Loman 1978). While Loman (1978) reported southern kingfish in Mississippi coastal waters with temperatures as low as 7.0°C, the highest previously recorded temperature was 31.0°C reported by Springer and Woodburn (1960) 431 40- 30- 20 10 40 30- 20 ■ 10 I I M | I W MAR 1980 n = 126 "n~m q i n 1 1 n i ii m i in FEB 1980 n = 423 :N=FFFn=Ffh-TT -f SEP 1980 n=47 Tl IT-rKI n n 1 1 M Hip TTT AUG 1980 n=5B ~l 1 1 1 1 n i m~i l~l i ■ 1 1 1 1 1 1 Mill. 40- 30- 20- 10- 40- 30- 20 10 JAN 1980 n= 146 TTT ^fl fl piTTTn i i | 1 1 i i_~r4 JUL 1980 n= 33 fl I | I I I I I I M I I | I I | | I | | I I DEC 1979 n= 195 1 1 rn HI tffltbn 40- 30 20- 10- 40- 30- 20- 10- m 1 1 1 1 1 c NOV 1979 n=103 ■FFFF 4W I I I I I I | I I I I- flfl OCT 1979 n = 109 nrp , n-| mi 1 1 1 n 1 1 1 in 50 100 150 200 250 JUN 1980 n= 33 TiirrrT fl ^fTtrrrjT-rn- MAY 1980 n = 79 Iflfl* n ' i ' 1 1 1 1 1 1 1 1 1 1 APR 1980 n = 191 m infT M m T1 1 1~| i i M 50 100 150 200 250 STANDARD LENGTH (mm) FIGURE 5.— Standard length frequencies of southern kingfish captured at regular monthly stations from October 1979 through September 1980. Total number of fish per month is represented by "n". 432 30 _ : 10 . y o 20^ 10-1 I I I I | I I I I r^ Fall 1979 n =121 "h. HI i i i i i < 50 f=H=- Spring 19B0 n = 307 :P^=t -n- 100 150 200 STANDARD LENGTH (mm] 300 FIGURE 6. — Standard length frequencies of southern kingfish caught during the fall and spring offshore cruises in the Gulf of Mexico. Total number of fish per cruise is represented by "n". in the Tampa Bay area. Loman (1978) reported a much narrower temperature range of 24.0-30.0°C for larval and postlarval fish (4.0-20.0 mm SL). The length-frequency distribution recorded dur- ing 1979-80 is comparable to the published reports for 1973-76 (Loman 1978). Growth of southern kingfish is fairly consistent among individuals with the most rapid growth during the first year, as also reported by Hildebrand and Cable (1934) and Bearden ( 1963 ) for southern kingfish on the Atlan- tic coast. Young-of-the-year averaged about 100 mm SL by November. Bearden (1963) also reported southern kingfish in South Carolina to reach about 100 mm SL by November of the first year, whereas Hildebrand and Cable (1934) reported a slightly higher average of 135 mm TL by November of the first year in Beaufort. N.C. Acknowledgments A debt of special thanks is due to Ronald Fritzsche of Humboldt State University, Luther Knight of the University of Mississippi, and Thomas Mcllwain of the Gulf Coast Research Laboratory for their advice and help throughout this study. I thank all the personnel of the Gulf Coast Research Laboratory for their help, cooper- ation, and the use of their facilities. Thanks go also to the members of the Southeast Fisheries Center of the National Marine Fisheries Service for their assistance in obtaining offshore samples. This study was supported by a grant from the Missis- sippi Department of Wildlife Conservation, Bureau of Marine Resources. Literature Cited BEARDEN, C. M. 1963. A contribution to the biology of the king whitings, genus Menticirrhus , of South Carolina. Contrib. Bears Bluff Lab. 38, 27 p. Christmas, J. Y. 1978. Fisheries assessment and monitoring - Mississip- pi. PL. 88-309, Proj. 2-215-R Completion Rep., p. 120- 167. Gulf Coast Res. Lab., Ocean Springs, Miss. CHRISTMAS, J. Y., AND R. S. WALLER. 1973. Estuarine vertebrates, Mississippi. In J. Y. Christmas (editor!, Cooperative Gulf of Mexico estuarine inventory and study, Mississippi, p. 320- 434. Gulf Coast Res. Lab., Ocean Springs, Miss. ELEUTERIUS, C. K. 1978. Geographical definition of the Mississippi Sound. Gulf Res. Rep. 6:179-181. FRANKS, J. S. 1970. An investigation of the fish population within the inland waters of Horn Island, Mississippi, a barrier island in the northern Gulf of Mexico. Gulf Res. Rep. 3:3-104. GEAGAN, D. W. 1962. An investigation of the sport fishes in the coastal waters of Louisiana. Bienn. Rep., La. Wildl. Fish. Comm. 9:125-132. GUNTER, G. 1938. Seasonal variations in abundance of certain es- tuarine and marine fishes in Louisiana, with particular reference to life histories. Ecol. Monogr. 8:313-346. 1945. Studies on marine fishes of Texas. Publ. Inst. Mar. Sci. Univ. Tex. 1:1-190. HILDEBRAND, S. F, AND L. E. CABLE. 1934. Reproduction and development of whitings or king- fishes, drums, spot, croaker, and weakfishes or sea trouts family Sciaenidae, of the Atlantic coast of the United States. Bull. U.S. Bur. Fish. 48:41-117. HILDEBRAND, S. F, AND W. C. SCHROEDER. 1928. Fishes of Chesapeake Bay. Bull. U.S. Bur. Fish. 43:1-366. IRWIN, R. 1970. Geographical variation, systematics, and general biology of shore fishes of the genus Menticirrhus, family Sciaenidae. Ph.D. Thesis, Tulane Univ., New^ Orleans, 295 p. JOHNSON, G. D. 1978. Development of fishes of the Mid- Atlantic Bight. Vol. IV. Carangidae through Ephippidae. U.S. Fish Wildl. Serv. FWS OBS-78/R. 314 p. U.S. Gov. Printing Office. Wash.. D.C. LOMAN, M. 1978. Other finfish. In J. Y Christmas (editor), Fisheries assessment and monitoring - Mississippi, p. 120-167. PL. 88-309, Proj. 2-215-R Completion Rep. Gulf Coast Res. Lab., Ocean Springs, Miss. MClLWAIN, T. D. 1978. An analysis of salt water angling in Biloxi Bay 1972-1974. Ph.D. Thesis, Univ. Southern Mississippi, Hattiesburg, 156 p. MILLER, J. M. 1965. A trawl survey of the shallow Gulf fishes near Port Aransas, Texas. Publ. Inst. Mar. Sci. Univ. Tex. 10:80- 107. 433 PEARSON, J. C. 1941. The young of some marine fishes taken in lower Chesapeake Bay, Virginia, with special reference to the gray sea trout Cynoscion regalis (Bloch). U.S. Fish Wildl. Serv., Fish. Bull. 50:70-102. RICHARDS, C. E., AND M. CASTAGNA. 1970. Marine fishes of Virginia's eastern shore (inlet and marsh, seaside water). Chesapeake Sci. 11:235-248. ROUNSEFELL, G. A., AND W. H. EVERHART. 1953. Fisheries science: its methods and applications. John Wiley and Sons, Inc.. N.Y., 444 p. SPRINGFR, V. G„ AND K. D. WOODBURN. 1960. An ecological study of the fishes of the Tampa Bay Area. Fla. Board Conserv. Mar. Lab. Prof. Pap. Ser. 1. STEVENS, J. R. 1962. Analysis of populations of sport and commercial finfish and of factors which effect these populations in the central hays of Texas. Tex. Game Fish. Proj. Rep. (1961- 1962 '.Job 1:1-61. I S DEPARTMENT OF COMMERCE. 1977-78. Mississippi landings. U.S. Dep. Commer., NOAA, Natl. Mar Fish. Serv., Curr. Fish. Stat. 7519, 7 198, 7561, 7580, 7600. 7620. 7639, 7659. 1979. FRS Oregon II cruise 101 report. NOAA, Natl. Mar. Fish. Serv.. Southeast Fish. Cent., Pascagoula, Miss., p 1-.") 1981 1 FRS ( )regon II cruise 106 report. NOAA, Natl. Mar. Fish Serv., Southeast Fish. Cent.. Pascagoula, Miss., p. 1-4 WELSH, W M . AND ( ' M BREDER. JR. 1924. ( lontributions to the life history of Sciaenidae of the Eastern United States coast. Bull. U.S. Bur. Fish 39:141-201 Barbara J. Crowe 2072 West Second Street, Apt. 11-E Beach. MS .{9560 growth of tunas (probably bluefin) was by Greek fishermen nearly 2,000 yr ago as documented in Aristotle's "Historia Anumalium" (Bell 1964). In recent times, the aging of tunas has become much more important and has been critiqued by Hayashi (1958), Bell (1964), and Shomura (1966). These reviews point to the problems and difficul- ties in aging tuna. These problems and difficulties appear to be more evident in aging bluefin tuna. Bluefin tuna are usually aged by counting growth increments on their hard parts. Vertebrae have provided acceptable ages (Rodriguez-Roda 1964; Butler 1971; Nichy and Berry 1976; Berry et al. 1976), but the aging of large or "giant" (>250 kg) bluefin tuna is suspect because the outer in- crements appear very close together. Otoliths have also been used to study age and growth of bluefin tuna (Butler et al. 1977). Berry et al. (1976) com- pared otolith age estimates with vertebra esti- mates and discovered a discrepancy. They found corresponding marks on both vertebrae and otoliths for the first 10 yr, but not thereafter, when otoliths had more incremental zones. They hypothesized that more than one incremental zone was deposited yearly in otoliths after the first 10 yr. Daily increments in yellowfin tuna, Thunnus albacares, and skipjack tuna, Katsuwonus pelamis, otoliths were studied by Wild and Fore- man (1980) and Uchiyama and Struhsaker (1981). Taubert and Tranquilli (1982) used daily incre- ments to verify annuli in the otoliths of large mouth bass, Micropterus salmoides salmoides, and it is proposed that an analogous investigation would provide corroborative evidence for the an- nual nature of outer major increments in giant bluefin tuna otoliths. SCANNING ELECTRON MICROSCOPE EVIDENCE FOR YEARLY GROWTH ZONES IN GIANT BLUEFIN TUNA, THUNNUS THYNNUS, OTOLITHS FROM DAILY INCREMENTS Atlantic bluefin tuna, Thunnus thynnus, are found throughout the Atlantic Ocean, the Mediterranean Sea, and the Gulf of Mexico (Gibbs and Collette 1967 ). Bluefin tuna are both commer- cially and recreationally important. Thus, it is important that the population dynamics of this species be understood in order that international policies can be developed. Age determination and subsequent growth es- timation are critical for tuna management. How- ever, confusion and controversy surround age es- timation in tunas. The earliest record of age and 404 FISHERY BULLETIN: VOL. 82, NO. 2, 1984. Methods and Materials Sagittal otoliths were collected in November 1978, from giant bluefin tuna which were reared in the sea ranching program of St. Margaret's Bay, Nova Scotia, Canada. Fish were weighed and mea- sured (TL) and the otoliths were collected as de- scribed by Caddy et al. (1976). All otoliths were washed in water and stored dry. Whole otoliths from four fish were placed in epoxy resin and sectioned on a Buehler Isomet1 saw. Sections 200 /xm thick were acquired from the region judged to contain the core. A diagrammatic view of a cross section of a bluefin tuna otolith is 1 Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. shown in Figure 1. Ten to 15 sections were sawed from each otolith. The number of sections viewed was dependent upon the clarity of the increments. Each otolith section was fastened to an alumi- num scanning electron microscope (SEM) stub with 5-min epoxy. The otolith section was highly polished with 0.3 /u.m alumina paste and etched with 67c EDTA (ethylenediaminetetraacetatic acid, adjusted to pH 8 with NaOH) for 1 to 20 min. The otolith sections were washed in water, dried, coated with gold, and viewed on a SEM at various magnifications. Observations and counts were made while the otolith section was in the SEM. It was discovered that different areas of the ros- tral lobe of the otoliths were made clear by differ- ent etching times. Sequential etching made it pos- sible to view microincrements in the outer 10 major increments. Individual sections were etched for different periods of time, with 15- to 20-min etching times showing the inner increments more clearly. The 10 outermost major increments were clearly visible in all sections and could be followed from section to section regardless of etching time. Each major increment was chosen to be from the center of one ridge to the center of the successive ridge; sequential etching revealed the micro- increments between the ridges. It was not possible to count the microincrements from the edge of the otolith inward past the 10th major increment on any individual section. Consequently, sequential cross sections from each otolith were etched for different periods of time, in steps of 1 min, in order to follow the progression of the microincrements. In the present study, a microincrement was de- fined as an unbroken incremental zone with dis- continuous zones as boundaries (Radtke and Dean 1982) and was considered to be a daily increment. Results and Discussion SEM techniques made it possible to view micro- increments in bluefin tuna otoliths from four indi- vidual fish. The most visually distinct increments were found on the rostral lobe of the otolith cross section (Fig. 1). Thus this area was used predomi- nantly for SEM observations. The major incre- ments of the otolith can readily be seen in Figure 2. Higher magnification (10,000 x) revealed that the major increments were constructed of smaller increments which in turn were composed of micro- increments (Figs. 3, 4). Differential etching caused the problem that not all the increments could be viewed at the same time. This was overcome through the use of suc- FIGURE 1. — Cross section of a bluefin tuna otolith showing the area (arrow) studied for microincrements. This area is on the rostral lobe of the otolith. 435 ■' 200jjm / FIGURE 2. — Bluefin tuna otolith etched with EDTA which shows distinctive major increments. A short etching time gave good resolution to the outermost increments. FIGURE 3. — Protein ridges of microincrements from a bluefin tuna. Strands of protein can be seen to interconnect the ridges. 436 Wx FIGURE 4. — Microincrements detected on the slope of a major protein ridge from a bluefin tuna. Differences in widths cause the yearly increments. cessive cross sections which were etched for dif- ferent time periods. This sequential etching made it possible to follow microincrements within the major increments. A difference in etching can be seen in Figures 2 and 5. Although major incre- ments were clear in most etching times, the micro- increments were not. Through the utilization of these techniques it was possible to obtain microincremental numbers for major increments (Table 1). The microincrement counts in each major in- crement varied from 273 to 385 with the lowest count being found on the edge of the otolith. The summations of the microincrement counts for each fish were remarkably close and not significantly different (P > 0.05). Also, means of micro- increments for each fish were not significantly dif- ferent (P > 0.05) from the expected of 365 per year. These data increase the credibility of the micro- increments being daily and present a plausible verification of the major increments as being annual. Each microincrement is composed of a protein matrix with calcium carbonate crystals, in the aragonite crystal configuration, deposited within the matrix. Etching with EDTA dissolves the TABLE 1. — Numbers of microincrements found in the major increments on the outer edge of the rostral lobe of the sagittae of four bluefin tuna, Thunnus thynnus. Fish 1 2 3 4 Weight (kg) Fork length (cm) Sex 496 275 M 381 216 F 405 251 F 470 268 F Estimated age1 25 19 19 24 Ma|or increment Counts Counts Counts Counts 1 278 273 300 289 2 368 375 337 321 3 355 310 366 374 4 339 370 339 342 5 385 344 376 323 6 366 376 370 372 7 313 355 347 349 8 369 356 358 373 9 341 369 315 348 10 328 348 365 329 Total 3.442 3,476 3,473 3.420 Mean ± SD 344 ±32 348 ±33 347 ±25 342 ±27 'From counts of major increments by light microscopy. aragonite crystals leaving areas with a higher protein content to form discernible increments (Figs. 3, 4). Extended etching (times varied de- pending on the area of the otolith) can cause the protein ridges to collapse and prevent counting of the microincrements. Thus, etching times were critical to the acquisition of viewable increments. 437 - v - ■» 4* > FIGURE 5. — Neighboring area of a bluefin tuna otolith shown in Figure 2 which demonstrates the uneven effects of etching. The width of each microincrement varied in accor- dance with its position within a major increment. Microincrement width was probably a function of the time of the year when deposited. The widest microincrements were displayed between the ridges. Furthermore, the microincrements formed at the edge of the sagittae were wide and deposited during a time when the fish were fed large amounts of mackerel as part of the sea ranching operations. Observations on microincrement width suggest that wide microincrements were deposited during summer feeding and growth, while finer microincrements were deposited dur- ing the winter. It was these differences in width that accounted for the formation of yearly increments. Most fish species investigated for daily age es- timates have been found to possess daily incre- ments in their otoliths (Pannella 1971; Brothers et al. 1976; Struhsaker and Uchiyama 1976; Taubert and Coble 1977; Methot and Kramer 1979; Steffen- sen 1980; Wild and Foreman 1980; Townsend and Graham 1981; Uchiyama and Struhsaker 1981; Radtke and Dean 1982). Thus, it is conceivable that the microincrements displayed in bluefin tuna otoliths are also daily. In tunas, Wild and Foreman (1980) studied daily increments in yellowfin and skipjack tuna, and Uchiyama and Struhsaker (1981) also investigated daily incre- ments in yellowfin and skipjack tuna. Yellowfin tuna are found to deposit daily increments in both studies, whereas Wild and Foreman (1980) sug- gested that skipjack tunas have 25% fewer in- crements than would be expected if the increments occurred daily, while Uchiyama and Struhsaker ( 1981) advocated that daily increments did occur in skipjack tuna. In light of the present data, Wild and Foreman (1980) may have not detected incre- ments formed during winter or colder periods. For giant bluefin tuna it is suggested that the micro- increments are formed daily. If bluefin tuna did not deposit microincrements on a daily schedule, it would be expected that fewer daily increments would be detected in each major increment. Since this is not the case, it corroborates the idea that daily increments are formed in bluefin tuna otoliths and groups of daily increments form an- nual increments. Otoliths may be the most useful hard structure for aging fish. Vertebrae and other hard structures are much more susceptible to resorption during times of physiological stress, while otoliths are 438 capable of permanently storing important ecologi- cal information since they are not susceptible to resorption (Mugiya and Watabe 1977). Otoliths have been shown to be the more accurate method of age determination in several fish species (Six and Horton 1977; Kimura et al. 1979). Otoliths are probably the most accurate means of age resolu- tion in bluefin tuna. In conclusion, the observation that micro- increments in the sagittae of giant bluefin tuna about 365 in number for each outer major incre- ment verifies the annual nature of these struc- tures and strongly suggests that the micro- increments are daily. Although it is not feasible to view large numbers of tuna otoliths by SEM techniques, the application of such techniques can provide answers to important biological questions. Acknowledgments Thanks are due to P. Hurley, Biological Station, St. Andrews, New Brunswick, for providing the samples, to J. Bell for critical review and discus- sion, and to R. Kawamoto for artistic work. This work, a part of the "Program Development, Management, and Administration" project (PM/M-1), is the result of research sponsored by the University of Hawaii Sea Grant College Pro- gram under Institutional Grant NA79AA-D- 00085 from NOAA, Office of Sea Grant, Depart- ment of Commerce, and NSF grant OCE-8208441. Literature Cited BELL. R. R. 1964. A history of tuna age determinations. Proceedings of the Symposium on Scombroid Fishes held at Mandapam Camp, Jan. 12-15, 1962. Mar. Biol. Assoc. India. Part 2:693-706. BERRY, F. H.. D. W. LEE, AND A. R. BERTINOLINO. 1976. Age estimation in Atlantic bluefin tuna - An objec- tive examination and intuitive analysis of rhythmic markings on vertebrae and in otoliths. ICCAT Doc. SCRS 76 67, 13 p. BROTHERS. E. B., C. P MATHEWS, AND R. LASKER. 1976. Daily growth increments in otoliths from larval and adult fishes. Fish. Bull., U.S. 74:1-8. BUTLER, M. J. A. 1971. Biological investigations on aspects of the life his- tory of the bluefin tuna, 1970-1971. Newfoundland and Labrador Tour. Dev. Off., 169 p. BUTLER, M. J. A.. J. F. CADDY, C. A. DICKSON, J. J. HUNT, AND C. D. BURNETT. 1977. Apparent age and growth, based on otolith analysis of giant bluefin tuna (Thunnus thynnus) in the 1975-76 Canadian catch. ICCAT Collect. Vol. Sci. Pap. 6 (SCRS- 19761:318-330. CADDY, J. F, C. A. DICKSON, AND M. J. A. BUTLER. 1976. Age and growth of giant bluefin tuna (Thunnus thynnus) taken in Canadian waters in 1975. Fish. Res. Board Can., Manuscr. Rep. Ser. 1395, 17 p. GIBBS, R. H., JR., AND B. B. COLLETTE. 1967. Comparative anatomy and systematics of the tunas, genus Thunnus. Fish. Bull., U.S. 66:65-130. HAYASHI, S. 1958. A review on age determination of the Pacific tunas. Proc. Indo-Pac. Fish. Counc, 7th Sess., Sec. 2-3:53-64. KIMURA, D. K., R. R. MANAPAT, AND S. L. OXFORD. 1979. Method, validity, and variability in the age determi- nation of yellowtail rockfish (Sebastes flavidus I, using oto- liths. J. Fish. Res. Board Can. 36:377-383. METHOT, R. D., JR., AND D. KRAMER. 1979. Growth of northern anchovy, Engraulis mordax, larvae in the sea. Fish. Bull., U.S. 77:413-423. MUGIYA, Y., AND N. WATABE. 1977. Studies on fish scale formation and resorption — II. Effect of estradiol on calcium homeostasis and skeletal tissue resorption in the goldfish, Carassius auratus, and the killifish, Fundulus heteroclitus. Comp. Biochem. Physiol. 57(2A): 197-202. NICHY, F, AND F H. BERRY. 1976. Age determination in Atlantic bluefin tuna. IC- CAT Collect. Vol. Sci. Pap. 5 (SCRS-1975):302-306. PANNELLA, G. 1971. Fish otoliths: daily growth layers and periodical pat- terns. Science (Wash., D.C.) 173:1124-1127. RADTKE, R. L., AND J. M. DEAN. 1982. Increment formation in the otoliths of embryos, lar- vae, and juveniles of the mummichog, Fundulus hetero- clitus. Fish. Bull, U.S. 80:201-215. RODRIQUEZ-RODA, J. 1964. Biologia del Atun, Thunnus thynnus I L. i de la costa sudatlantica de Espaha. Invest. Pesq. 25:33-146. SHOMURA, R. S. 1966. Age and growth studies of four species of tunas in the Pacific Ocean. In T. A. Manar (editor). Proceedings of the Governor's Conference on Central Pacific Fishery Re- sources, Honolulu-Hilo. February 28-March 12, 1966, p. 203-219. State of Hawaii, Honolulu. SIX, L. D., AND H. F HORTON. 1977. Analysis of age determination methods for yellow- tail rockfish, canary rockfish, and black rockfish off Ore- gon. Fish. Bull., U.S. 75:405-414. STEFFENSEN, E. 1980. Daily growth increments observed in otoliths from juvenile East Baltic Cod. Dana Rep. Carlsberg Found. No. 88, p. 29-37. STRUHSAKER, P., AND J. H. UCHIYAMA. 1976. Age and growth of the nehu, Stolephorus purpureus (Pisces: Engraulidae), from the Hawaiian Islands as indi- cated by daily growth increments of sagittae. Fish. Bull.. U.S. 74:9-17. TAUBERT, B. D., AND D. W. COBLE. 1977. Daily rings in otoliths of three species of Lepomis and Tilapia mossambica. J. Fish. Res. Board Can. 34:332-340. TAUBERT, B. D., AND J. A. TRANQUILLI. 1982. Verification of the formation of annuli in otoliths of largemouth bass. Trans. Am. Fish. Soc. 111:531-534. TOWNSEND, D. W., AND J. J. GRAHAM. 1981. Growth and age structure of larval Atlantic herring, Clupea harengus harengus, in the Sheepscot River Es- 439 tuary, Maine, as determined by daily growth increments in otoliths. Fish. Bull., U.S. 79:123-130. UCHIYAMA, J. H., AND P. STRUHSAKER. 1981. Age and growth of skipjack tuna, Katsuwonus pelamis, and yellowfin tuna, Thunnus albacares, as indi- cated by daily growth increments of sagittae. Fish. Bull., U.S. 79:151-162. WILD, A., AND T. J. FOREMAN. 1980. The relationship between otolith increments and time for yellowfin and skipjack tuna marked with tet- racycline. [In. Engl, and Span.] Inter-Am. Trop. Tuna Comm. Bull. 17:509-560. RICHARD RADTKE Hawaii Institute of Marine Biology P.O. Box 1346 Kaneohe, HI 96744 YEARLY CHANGES IN ABUNDANCE OF HARBOR SEALS, PHOCA V1TULINA, AT A WINTER HAUL-OUT SITE IN MASSACHUSETTS Information on the abundance of the harbor seal, Phoca vitulina concolor, population in New England consists of outdated estimates in the lit- erature (King 1964; Maxwell 1967; Hewer 1974; Bonner 1976). A more recent series of unpublished reports (Richardson1; Knapp and Winn2; Kraus3; Gilbert and Stein4) suggests a harbor seal popula- tion which is increasing in numbers from its present breeding range north of Massachusetts southward into southern New England. A primary research need identified by Prescott et al.5 was confirmation of this suspected increase in the har- bor seal population throughout New England. 'Richardson, D. T. 1973. Distribution and abundance of harbor and gray seals in Acadia National Park. Final report to National Park Service and Maine Department of Sea and Shore Fisheries, State of Maine Contract No. MM4AC009, 59 p. 2Knapp, C. L., and H. E. Winn. 1978. Harbor seals, New Hampshire to Long Island. Unpubl. rep., University of Rhode Island, Graduate School of Oceanography, Kingston, RI 02881, 36 p. 3Kraus, S. 1980. The population of harbor seals (Phoca vit- ulina) in southern New England. Unpubl. rep. of harbor seal workshop, 5 March 1980, Boston, Mass. New England Aquarium, Boston, MA 02109, 9 p. ••Gilbert, J. R., and J. L. Stein. 1981. Harbor seal popula- tions and marine mammal fisheries interactions. University of Maine, Department of Forestry and Wildlife Resources, Orono, Maine. Annual Report to NEFC/NMFS/NOAA, Contract No. NA-80-FA-C-00029, 55 p. ^5Prescott, J. H.,S. D. Kraus, and J. R. Gilbert. 1980. East Coast/Gulf Coast cetacean and pinniped workshop. Final Re- port for Marine Mammal Commission, contract 79/02. Avail- able National Technical Information Service, Springfield, VA 22151 as PB80-160104, 142 p. This study summarizes available data on annual fluctuations in seal numbers since 1972 at one site in southeastern Massachusetts. The study was conducted at Stage Point, Man- omet, Mass. (lat. 41°55'N, long. 70°32'W). Harbor seals occur seasonally at Stage Point from late October through May (Schneider and Payne 1983). A rapid decrease in numbers occurs at this site in May (Schneider and Payne 1983 ), prior to the pup- ping season which occurs mid-May to mid-June in Maine (Richardson footnote 1; Wilson6). A few seals are reported throughout the summer but most move northward out of the study area by June. The study site consists of a shoreline with a sandy cliff to 25 m. Sand, rock, and cobble extend from the base of the cliff into the water. Seals haul out exclusively on the larger rocks in the im- mediate subtidal zone from about 1-2 h before to 1-2 h after low tide (Schneider and Payne 1983). A similar haul-out pattern has been described at other rock-ledge sites in New England (Richardson footnote 1; Wilson footnote 6). Be- cause of the synchronized haul out observed at Stage Point, the number of seals seen on the rocks is considered representative of the number of seals in the immediate vicinity (Schneider and Payne 1983) and, therefore, a useful index for monitoring changes in the abundance of harbor seals at this location. Methods Counts at Stage Point were made by direct ob- servation within 2 h of low tide from the cliffs above the haul-out site. Schneider and Payne (1983) found that during 1979-80 the average number of seals observed at Stage Point peaked in January; therefore, the average number of seals ( ±SE) seen per daily count in January of each year was used in analyses among years. We trans- formed the January averages into logarithmic values, and the coefficient of correlation (r) from the linear regression was used to describe the rela- tionship between the average number of seals seen per daily count in January 1972 and 1983. In addition, air temperature, wave intensity, and human disturbance influence the total number of seals seen per daily count at Stage Point 6Wilson, S. C. 1978. Social organization and behavior of harbor seals Phoca vitulina concolor in Maine. Final Report to Marine Mammal Commission, Contract No. GPO PB 280- 3188. Available National Technical Information Service, Springfield, VA 22151 as PB 280 188, 103 p. 440 FISHERY BULLETIN: VOL. 82, NO. 2, 1984. ( Schneider and Payne 1983 ) . Prior to the winter of 1979-80, a record of environmental conditions at the time of the count was not maintained. Since it is not known to what extent weather or human disturbance near the haul-out site had on zero or near-zero counts previous to 1979-80, all daily counts in January with less than five seals were considered unreliable and excluded from the analyses. There were no available data for January 1973 or January 1977. Results and Discussion The average number of seals observed per daily count in January (Table 1) ranged from 9.3 seals (1974) to 88.25 seals (1980) with considerable var- iability among years. However, the observed number of seals was not randomly distributed among years; the January averages increased sig- nificantly (P <0.05, r = 0.63, df = 9) between 1972 and 1983 (Fig. 1). The average annual rate of increase since 1972 at Stage Point (based on expected values from the semilogarithmic regression, Table 1) was 11. 9% /yr. The expected average number of seals per daily count in January at Stage Point (Table 1) doubled between 1973 and 1980. The observed increase in the average number of seals at Stage Point has followed the termination in 1962 of a Massachusetts bounty on harbor seals and passage in 1972 of the Marine Mammal Pro- tection Act. Rapid expansion of seal populations after the passage of protective legislation has been observed in the past (Hewer 1974; Bonner 1975; Everitt and Beach 1982) and has likely facilitated the increase since 1972 of the number of seals seen at Stage Point. TABLE 1. — January averages of seals observed per daily count, 1972-83, at Stage Point, Man- omet, Mass. n.d. = no data. No. xno. (±SE) Expected x no. daily seals/daily seals/daily Year counts count count1 1972 2 12.5 (6.52) 12.86 1973 n.d. 14.39 1974 3 9.3 (1.85) 16.11 1975 15 18.8 (2.32) 18.02 1976 2 34.0 (4.00) 20.17 1977 n.d. 22 58 1978 9 20.0 (3.07) 25.28 1979 9 35.56(7.00) 28.29 1980 28 88.25(6.06) 31.66 1981 18 21.67(6.92) 35.43 1982 18 21.88(2.94) 39.66 1983 19 4800(5.63) 44.39 An increase in seal populations (after protec- tion) due to unrestricted dispersion of juvenile seals has also been noted elsewhere (Bonner and Witthames 1974; Reijnders 1983). Bonner and Witthames (1974) suggested that the population of common seals, P. v. vitulina, located at the Wash in England, acted as a reservoir from which other reduced populations were replenished. Existence of a seal population in the Dutch Wadden Sea depends on unrestricted dispersal of juvenile seals from adjacent rookeries (Reijnders 1983). Since no rookeries occur south of Maine, it is apparent that the population increase seen at Stage Point (and throughout southern New England) has occurred through the southward dispersion of seals from Maine rookeries, after protection was established in Massachusetts. _l < tr UJ CD 13 y=H4898e r = 0626 df = 9 p<005 (OII263)x 'From the linear regression: r= 0.628. P< 0.05 (Fig. 1). 1 1.4898 e<011263>*, — I 1 1 1 1 1 i 1 1 1 1 r 72 73 74 75 76 77 78 79 80 81 82 83 YEAR FIGURE 1 . — Semilogarithmic plot of the average number of seals observed per daily count in January at Stage Point, Manomet, Mass., 1972-83. 441 Several investigators have reported an increase in seal numbers elsewhere in New England over the past decade. Gilbert and Stein (footnote 4) reported a total of 10,483 seals counted in June 1981 between Isles of Shoals on the Maine-New Hampshire border and the Canadian border. This nearly doubled the 1973 census of 5,786 seals re- ported for the same area by Richardson ( footnote 1). Our data at Stage Point confirm this increase in southern New England. The possibility does exist that the increase ob- served at Stage Point is merely the result of more thorough survey coverage in recent years; how- ever, coastal bird observations were made regu- larly at Stage Point before 1973-74 by staff at the Manomet Bird Observatory. Any large number of seals would have been noticed during such counts. The present harbor seal distribution, abun- dance, and breeding status in Massachusetts have changed considerably from the past. Allen (1869) reported "hundreds" of seals during the summer in Boston Harbor. As late as the 1930s and 1940s, harbor seals were permanent residents on Cape Cod (Prescott 1981) and pupping occurred throughout Massachusetts. Katona et al. (1983) suggested that the retention of the bounty until 1962 led to the extirpation of breeding activity in Massachusetts. The continued protection of an in- creasing harbor seal population throughout New England may result in expansion of the present breeding range southward into areas formerly used for pupping. Acknowledgments We wish to thank many people who helped in the counts: Brian A. Harrington (1972-74), Pam Cot- ton and Frank Gardner (1975-76), Kevin D. Pow- ers (1977-79), and Ann M. Frothingham (1978-79). James R. Gilbert, Kate M. Wynne, Kevin D. Pow- ers, and anonymous reviewers commented on pre- vious drafts of this manuscript. This study was funded by National Marine Fisheries Service/ Northeast Fisheries Center Contract Nos. NA- 80-FA-00005 and NA-82-FA-00007, and private funding. Bonner, W. n. 1975. Population increase of grey seals at the Fame Is- lands. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 169:366- 370. 1976. The stocks of grey seals (Halichoerus grypus) and common seals (Phoca vitulina) in Great Britain. Nat. Environ. Res. Counc. Publ. Ser. C, No. 16, 16 p. BONNER, W. N., AND S. R. WlTTHAMES. 1974. Dispersal of common seals ( Phoca vitulina), tagged in the Wash, East Anglia. J. Zool., Lond. 174:528-531. EVERITT, R. D., AND R. J. BEACH. 1982. Marine mammal-fisheries interactions in Oregon and Washington: An Overview. In K. Sabol (editor). Transactions of the American Wildlife and Natural Re- source Conference, p. 265-277. Wildlife Management In- stitute, Wash., D.C. HEWER, H. R. 1974. British Seals. William Collins, Press, Lond., 256 p. katona, s. K., v. Rough, and d. t Richardson. 1983. A field guide to whales, porpoises and seals of the Gulf of Maine and eastern Canada, Cape Cod to New- foundland. 3d ed. Scribner's Sons, N.Y., 255 p. KING, J. E. 1964. Seals of the world. Trustees of the British Museum, Lond., 153 p. Maxwell, G. 1967. Seals of the world. Constable and Co. Press, Lond., 151 p. PRESCOTT, R. 1982. Harbor seals, mysterious lords of the winter beach. Cape Cod Life 3(41:24-29. REIJNDERS, R J. H. 1983. The effect of seal hunting in Germany on the further existence of a harbour seal population in the Dutch Wad- den Sea. Z. Saeugetierkd. 48:50-54. SCHNEIDER, D. S.. AND R M. PAYNE. 1983. Factors affecting haul-out of harbor seals at a site in southeastern Massachusetts. J. Mammal. 64:518-520. Manomet Bird Observatory Manomet, MA 02345 Ecology and Evolutionary Biology University of California Irvine, C A 92717 P. Michael Payne David C. Schneider Literature Cited Allen, J. a. 1869. Catalog of the animals of Massachusetts with a criti- cal revision of the species. Bull. Mus. Comp. Zool. No. 1, p. 143-252. 442 POSTOVULATORY FOLLICLE HISTOLOGY OF THE PACIFIC SARDINE, SARDINOPS SAGAX, FROM PERU' The use of the postovulatory follicle as a means of estimating incidence of spawning in multiple spawning fishes was originally developed by Hunter and Goldberg (1980) for the northern an- chovy, Engraulis mordax, from southern Califor- nia. This technique has proven to be quite useful for biomass assessment using the "Egg Production Method" (Parker 1980). As a result of this work, the postovulatory folli- cle has assumed new importance. In the work of Hunter and Goldberg (1980), E. mordax were spawned artificially in the laboratory (Leong 1971). Fish were sacrificed at different time inter- vals, and histological conditions of the postovula- tory follicles were noted. As an alternative to this method, in the current report we have aged post- ovulatory follicles of the Pacific sardine, Sar- dinops sagax, from Peru by establishing the time of spawning (egg collections) and by making periodic collections of S. sagax. Methods Samples of Sardi?2ops sagax were collected dur- ing September-October 1982 near Chimbote, Peru (lat 09°05', long. 78°35' ). Ovaries were preserved immediately on collection in 10** neutral, buffered Formalin2. Later, samples from a total of 270 ovaries were dehydrated in ethyl alcohol and em- bedded in Paraplast. Histological sections were cut at 6/x. Slides were stained with Heidenhain's iron hematoxylin or Harris' hematoxylin followed by eosin counterstain. Sardine egg samples from Peru indicated 0100 h to be the midpoint of the daily spawning interval ( Smith3 ). Therefore, by knowing the hour of collec- tion and assuming that spawning occurred around 0100 h, we calculated the approximate age of post- ovulatory follicles. Results and Discussion The sardine is a multiple spawning fish (Clark 1934), and during the spawning season we typi- ■Publication No. 11 of PROCOPA. 2Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 3P.E. Smith, Southwest Fisheries Center La Jolla Laboratory, National Marine Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA 92038, pers. commun. March 1983. cally observe a mature yolk-filled mode of eggs representing the next spawning session and a vi- tellogenic mode for a subsequent spawning. Postovulatory follicle. Day 0 (0-6 h after spawning) The new S. sagax postovulatory follicles (Fig. 1A, B) were striking in their strong resemblance to the age 0-day postovulatory follicles of E. mordax (elapsed time from spawning <24 h) (Hunter and Goldberg 1980). The newly formed follicles of S. sagax contained many involutions or corrugations and were composed of columnar epithelium rest- ing on a connective tissue theca. Nuclei had a basal location. The lumina occasionally contained eosinophilic granules of unknown origin (Hunter and Goldberg 1980) similar to those reported in the newly formed postovulatory follicles of E. mordax. Postovulatory follicle, Day 1 (7-30 h after spawning) These structures showed the beginning ( Fig. 1C ) of a breakdown in organization in comparison to day-0 postovulatory follicles. This included a size decrease to about one-half and marked degenera- tion of the columnar epithelial cell lining. Many epithelial cells had irregular shapes, vacuoles, and pycnotic nuclei. The convoluted structure was not as distinct as in day-0 postovulatory follicles. The linear arrangement of columnar epithelial cells was still evident. This is important, and con- stitutes the chief character that should be used for distinguishing day-1 from day-2 postovulatory fol- licles in S. sagax. This linear arrangement was absent in day-2 S. sagax postovulatory follicles. Postovulatory follicle, Day 2 (31-53 h after spawning) Degeneration of the S. sagax postovulatory fol- licle was clearly more advanced (Fig. ID) at this stage. Distinguishing them from old atretic folli- cles is now a critical problem. Lumina were typi- cally occluded and contained irregularly shaped cells with pycnotic nuclei, representing the final stages in the degeneration of the columnar epithe- lial cells that were previously so evident (Figs. 1A, B) in day-0 S. sagax postovulatory follicles. Vac- uoles may be present. While the greatly convo- luted structure that characterized earlier post- ovulatory follicles is no longer pronounced, there FISHERY BULLETIN: VOL. 82, NO. 2 443 C *miM FIGURE 1. — Photomicrographs of Sardinops sagax postovulatory follicles. (A) Day 0, showing highly convoluted morphology. Note cluster of eosinophilic granules I arrow) in lumen 1 250 x ). (P>i Day 0, showing columnar epithelial cell lining (arrow) (400x ). (C)Day 1. columnar epithelial cell lining (arrow) undergoing degeneration. Underlying layer is connective tissue theca (400 x ). (D) Day 2, lumen contains scattered degenerated columnar epithelial cells ( arrow i (400 x ). should be some suggestion of it. We therefore rec- ommend careful observation of the convoluted structure of day-0 and day-1 structures before at- tempting to identify day-2 structures. A useful criterion for distinguishing day-2 S. sagax postovulatory follicles from advanced atre- tic follicles would be the presence of yellow granules (irrespective of staining) that are found in advanced atretic structures (delta atresia) (Lambert 1970). These were occasionally noted in S. sagax. The presence of these yellow granules which appear in nucleated clusters conclusively indicates atretic structures. We did not use the artificial spawning technique (Leong 1971) for aging postovulatory follicles in S. sagax. However, we feel that estimating their age from periodic collections of fish, after the spawn- ing time is established from collections of egg samples (as done herein), will prove to be a useful alternative method. This is particularly true in situations where facilities are lacking for laboratory-induced spawning. Laboratory- induced spawning studies using S. sagax will be useful to provide estimates of the accuracy of our classification scheme. While there are numerous accounts of the oc- currence of postovulatory follicles in marine fishes, there are few reports describing their longevity and subsequent degeneration. They have been described previously as being short- lived structures by Yamamoto and Yoshioka (1964) and Hunter and Goldberg (1980). More studies are needed of a wide variety of fishes before our knowledge of their histology and function is com- pleted. Of utmost value will be investigations on how to distinguish conclusively between old post- ovulatory follicles and old atretic follicles. Acknowledgments We thank J. R. Hunter and B. J. Macewicz for their constructive comments. 444 Literature Cited CLARK, F. N. 1934. Maturity of the California sardine {Sardina caerulea (, determined by ova measurements. Calif. Dep. Fish (^ame, Fish Bull. 42, 49 p. HUNTER, J. R., AND S. R. GOLDBERG. 1980. Spawning incidence and batch fecundity in northern anchovy, Engraulis mordax. Fish. Bull., U.S. 77:641- 652. LAMBERT, J. G. D. 1970. The ovary of the guppy, Poecilia reticulata . The atre- tic follicle, a corpus atreticum or a corpus luteum praeovulationis. Z. Zellforsch. 107:54-67. LEONG, R. 1971. Induced spawning of the northern anchovy, En- graulis mordax Girard. Fish. Bull., U.S. 69:357-360. PARKER. K. 1980. A direct method for estimating northern anchovy, Engraulis mordax, spawning biomass. Fish. Bull., U.S. 78:541-544. YAMAMOTO, K., AND H. YOSHIOKA. 1964. Rhythm of development in the oocyte of the medaka, Oryzias latipes. Bull. Fac. Fish.. Hokkaido Univ. 15:5-19. STEPHEN R. GOLDBERG Department of Biology Whit tier College Whittier, CA 90608 Institute del Mar del Peru Apartado 22, Callao, Peru Programa Cooperatwo Peruano-Aleman de Investigation Pesquera tPROCOPA) Instituto del Mar del Peru Apartado 22. Callao, Peru VICTOR HUGO ALARCON JUERGEN ALHEIT A NOTE ON SPAWNING OF THE PACIFIC MARKET SQUID, LOLIGO OPALESCENS (BERRY, 1911), IN THE BARKLEY SOUND REGION, VANCOUVER ISLAND, CANADA In California, Loligo opalescens (Berry, 1911), has large spawning schools and spawn masses (McGowan 1954; Fields 1965; Hobson 1965; Cous- teau and Diole 1973; Hochberg and Fields 1980). Spawns and spawning effort of this squid in the Pacific Northwest are poorly known and, to our knowledge, large spawns or spawning events have not been quantitatively described. Loligo opalescens spawns regularly in Barkley Sound near Bamfield, British Columbia, (lat. 48°50.0'N, long. 125°07.5'W) in spring. We examined and measured portions of a spawn using scuba during early June 1982. The largest single capsule mass aggregation in our 200 x 50 m sur- vey area was measured. Adjacent areas of smaller solitary egg capsule masses were surveyed using transects to determine overall spawn dimensions and percent cover of individual capsule masses. Dimensions of 23 typical masses were determined. Four representative masses were collected; the number of capsules in each was counted; and from each, 10 capsules were randomly selected and the number of eggs in each capsule was determined. These eggs were examined microscopically to determine the developmental stage, which was compared with the embryological stages illus- trated in Fields (1965) to estimate the time of deposition. The spawn, including areas of continuous and solitary egg capsule masses, was larger than the area surveyed, as the spawn extended below our deepest possible survey depth. Within our survey area, the largest capsule mass aggregation cov- ered about 69.3 m2 and averaged 0.28 ± 0.09 m (n = 4) in thickness. The mean density of the indi- vidual masses was 1.3 ± 0.1/m2, and the mean area covered by 23 masses was 0.28 ± 0.14 m2/mass, with a range of 0.13-0.66 m2. The mean number of egg capsules per solitary mass was 1,937 ± 912 (n = 4), with 149 ± 35 eggs/capsule (n = 40). Thus, the total number of eggs per isolated mass was 288,000 ± 125,000. For the large areas of isolated masses, the potential number of larvae produced per 100 m2 ranged from 19 to 58 x 106, with a mean of 37 x 106. The number of potential larvae from the single large aggregation of 69.3 m2 ranged from 27 to 204 x 106, with a mean of 72 x 106. Based on embryological stages observed, dep- osition probably occurred during the night of 31 May-1 June 1982. Small squid schools were ob- served spawning near the survey area on that date. None of the embryos were old enough to be deposited before 31 May, and all were of the same embryological stage. Female squids from Californian populations de- posited about 21 capsules, each containing about 200 eggs, in one night (Fields 1965); fecundity data from our region are not available. Hochberg and Fields (1980) stated that L. opalescens females produce 180-300 eggs/capsule. Our data indicate a lower mean value of about 150 eggs/capsule. If each female deposited 20 capsules, the large mea- sured aggregation would be the result of about 24,000 females. FISHERY BULLETIN: VOL. 82, NO. 2. 1984. 445 In conclusion, northern Loligo opalescens popu- lations form large spawning schools and deposit massive egg capsule masses similar to those ob- served in the Californian populations. Acknowledgments We thank the directors of the Bamfield Marine Station and the Friday Harbor Laboratories for the use of the respective facilities. Two anonymous reviewers provided helpful suggestions. Literature Cited COUSTEAU, J.-Y, AND R DIOLE. 1973. Octopus and squid: the soft intelligence. Double- day, Garden City, N.Y., 304 p. FIELDS, W. G. 1965. The structure, development, food relations, repro- duction, and life history of the squid Loligo opalescens Berry. Calif. Dep. Fish Game, Fish. Bull. 131:1-108. HOBSON, E. S. 1965. Spawning in the Pacific Coast squid, Loligo opales- cens. Underwater Nat. 3(3):20-21. HOCHBERG, F G, AND W. G FIELDS. 1980. Cephalopoda: The squids and octopuses. In R. H. Morns. D. R Abbott, and E. C. Haderlie (editorsi, Interti- dal invertebrates of California, p. 432-444. Stanford Univ. Press. Stanford. MCGOWAN, J. A. 1954. Observations on the sexual behavior and spawning of the squid, Loligo opalescens, at La Jolla, Califor- nia. Calif. Dep. Fish Game, Fish. Bull. 40:47-54. RONALD L. SHIMEK DAVID FYFE LEAH RAMSEY ANNE BERGEY JOEL ELLIOTT STEWART GUY Bamfield Marine Station Bamfield. British Columbia Canada VOR 1B0 or exponentially as ARITHMETIC VERSUS EXPONENTIAL CALCULATION OF MEAN BIOMASS Mean biomass (B) within a time interval (t) is used in the Ricker method of estimating yield per recruit and can be calculated either arithmeti- cally as (i) 446 B< Bt + Bt+1 (ii) Bt Bt(eGt~Zt- 1) Gt-Zt (Ricker 1975). The choice of calculation method may influence the yield estimates and con- sequently the determination of optimal levels of exploitation. Ricker (1975) and Paulik and Bayliff (1967) al- luded to the importance of the difference in mag- nitude between instantaneous growth and total mortality rates (Gt - Zt). They indicated that if the difference was small, arithmetic and exponen- tial calculations approached one another. Ricker suggested using small intervals if the rates are rapidly changing. In this paper we 1) examine the difference in the two estimates of mean biomass as a function of the instantaneous rates of growth and mortality, and 2) reexamine the consequences of the choice of mean biomass estimates on esti- mates of equilibrium yield per recruit using data previously employed by Ricker (1975) and Paulik and Bayliff (1967), showing that under many con- ditions, exponential estimates of mean biomass are preferable. The difference between arithmetic and expo- nential estimates of mean biomass increases rapidly as Gt - Zt increases in a positive direction, but diverges less rapidly when Gt - Zt increases in a negative direction. When Bt is arbitrarily taken in unity, the relationship is satisfactorily rep- resented by a polynomial regression (Fig. 1). With many fisheries it is only possible to esti- mate instantaneous fishing mortality (Ft) on an annual basis. Thus, a large interval must be used. The larger the interval, the more likely it is that Gt - Zt is of a magnitude that would cause sig- nificant differences in estimates of Bt calculated arithmetically and exponentially. Also, in heavily exploited fisheries there may be a large difference between growth and mortality rates within an interval especially at older ages. We employed Ricker's (1975:242-243, table 10.3) example of bluegills from Muskellunge Lake to illustrate the difference between the two methods of computing mean biomass. This set was chosen because Ricker's data have been used previously as a historical data set and are readily available through his text. Mean biomass was computed arithmetically in the text example and also by Paulik and Bayliff (1967), who used the same data to introduce their computer program. We used the data in two separate runs to compute yield per FISHERY BULLETIN: VOL. 82, NO. 2, 1984. < LU Q -0 42 -0 36 -0 30 -0 24 - 0 18 -0.12 -0.06 0 00 -1.6 -1.2 -0.8 -0.4 0.0 G -Z FIGURE 1. — Difference between arithmetic and exponential calculations of mean biomass when dt = 1.0 and Bt = 1.0, DELTA =BtiB, , exp - B(,anth) = Bt* ,0.0061 + 0.0037 - 0.1095 UGt- Zt>2 1 - 0.0491 ( \ X\ " \ s \\ ^ IV-'' \\ ^ \\ \\ x \\ \\ N «n o m o mo mo 0, f • t' will not peak at Bg = 0 (for the situation described in Equation (5)). Instead, it will peak at some intermediate level of B and will then decline, Figure 2E. Thus, even for the hypothetical equilibrium fishery, f ■ t' would have to "voluntarily" decline from a maxi- mum level as Be approaches zero. DISCUSSION The assumptions inherent in Equation (5) limit its direct application as a qualitative model of existing fisheries. However, the general behavior described in Figure 2B, D has been noted in sev- eral recent papers (Fox 1974; Pope and Garrod 1975; Schaaf 1975b; MacCall 1976; Ultang 1976; Garrod 1977; Peterman 1980; Peterman and Steer 1981; Bannerot and Austin 1983). Two important examples occur with the Pacific sardine and At- lantic menhaden fisheries. In their analysis of the available catch and ef- fort data on the California based fishery on Pacific sardine, Fox (1974) and MacCall (1976) were forced to relax the usual restriction of a constant catch- ability coefficient which is independent of popula- tion size. Rather, they applied a density-dependent catchability coefficient of the form aN? (6) where a and /3 are constants, assuming a constant catch per unit effort. The general patterns pre- dicted by these analyses are similar to those in Figure 2B, D, with MacCall noting an inverse re- lationship between the apparent q and population abundance, and Fox noting a collapse of the fishery in plots of catch versus vessel-months, Fig- ure 3. Both of these patterns may be the result of an inability to describe mathematically how the time available for searching increased as the population of sardines declined. It is generally assumed that Atlantic menhaden have been overfished since the early sixties. Sup- port for this conclusion was derived from the surplus-production work of Schaaf and Huntsman (1972), later updated by Schaaf (1975a). In both studies the available effort index (vessel-weeks) was modified in an attempt to correct for changes in fishing efficiency with time. Under the assump- tion of a constant q and lacking detailed informa- tion on vessel characteristics and catch, the au- thors adjusted effort by "multiplying the effort observed in each year by the relative change in q", using either 1965 (Schaaf and Huntsman) or 1971 (Schaaf 1975a) as a base year. The resulting pat- tern (Fig. 4A) strongly suggests that the fishery was operating on the descending arm of the catch-effort curve. In another paper, Schaaf (1975b) observed an inverse relationship between his estimates of catchability coefficient and the population density, generating a pattern, like MacCall's (1976), simi- lar to Figure 2D and at least partially explained by the lack of information on density-dependent searching time. However, Schaaf's apparent density-dependent estimate of q violates his early assumption of a constant q for use in standardiz- ing the available effort data. The point is not triv- ial. Without this adjustment, the available effort data suggest that the Atlantic menhaden fishery is operating on the ascending arm of the catch- effort curve, Figure 4B. 450 CONDREY: DENSITY-DEPENDENT SEARCHING TIME 25 - 20 40 60 80 EQUILIBRIUM POPULATION DENSITY 100 20 - 0 U L L 15 I U M 10 - 5- B y\ *■-: }-<. /? / / !/ }/ ! ' / / / / /I /; /j / / / A ' / I / / Y / / / I : / / i - / / ,'.V' I.// 2 4 "5 8~ EQUILIBRIUM (ft') 10 12 0 20 -U.D 1 0- *c l\ »N v>\ V> \ s ° 8" v. \ E A R C H • 0.6- g \\\ * \ \ ^ T \\ \ I E 0.4* \ x- *» — — 1 X X ~-., ^^ ***'--. ~~~-^ .2 0.2- ""■-««. '""**—■ ~_ 3 ~~ — 4 ""' — T." 0.0- L , , . r ■ t 20 40 60 80 EQUILIBRIUM POPULATION DENSITY 100 1 1 8 - \\\ ' z w 0 16 y u. u, 8 0 14 O 5 0 12i < z o lu 0 08 K 2 < O 06 0 04 0 02 |\ W\ \ \\\ \ \\ \ \ \\ \ \\ \ \ \ \ \ \ V3 4 \ X 20 40 60 80 EQUILIBRIUM POPULATION DENSITY 100 E 12- / \ /' \ 10- />^"^X. \ r 8' a ' / \ \ //,---" '"---.. \\ = -i 5 o " 4, // \ \\ "\ \\\ \ \\ ^ \ v 2- 0 V\\\ FIGURE 2. — Effect of variation in the ratio of handling time W. ) to total fishing time (t ' ) on the relationship between (A) equilib- rium yield, Y , and the equilibrium biomass of the stock, Bg, (B) Y and the product of t' and equilibrium fishing gear; (C) searching time andB ; (D) catchability incorrectly estimat- ed as ql(\ + q ■ t' ■ B ) andB ; and (E) equilibrium/"' ■ t' and B . The model used is Equation (5) where Bx = 100, k = 1, and q = .2. 0 20 40 60 80 100 EQUILIBRIUM POPULATION DENSITY 451 FISHERY BULLETIN: VOL. 82, NO. 3 800 O 600 o c 400 IB o 2 200 4) ' S A S \ s \ * s V A r— — •/ 1 i / ' **^\ I jr } \ i v i ', A i ^i /l / 1 .' \ i / 1 I \ / • / \ / ' # r - ~-\ / . \ ' ■*' \ / • 1932-33 \ > \ V '"\ / ' •- •! 1 / •' "~--i - / 1954-55 / / + * — -"""'^ f 1 ** ** 1 1 1 1 i i i 0 2 4 6 8 10 12 14 16 18 Fishing Effort (thousand boat months) FIGURE 3. — Catch versus a nominal effort index for the California Pacific sardine fishery (after Fox 1974). IMPLICATIONS In this simple treatment, I have shown how some available estimates of fishing effort are in- adequate to describe a fishery where searching time is stock-density dependent. More realistic models can be based upon an examination of de- tailed log-book data for inferences as to the rela- tionships between searching time, stock abun- dance, prey and fleet distribution, and cooperation among the fleet. Hassel (1978) offers a comprehen- sive review of current approaches. Even with better estimates of fishing effort a catastrophic collapse may be unavoidable because of parameters that cannot be measured or con- trolled (e.g., Clark and Manguel 1979). These es- timates are important, however, in order to use surplus-production models for estimating the causes of such catastrophic collapses as that of the Pacific sardine fishery. Such estimates may also be necessary to suggest whether existing fisheries such as the Atlantic menhaden are actually being overfished. ACKNOWLEDGMENTS I am indebted to the captain and crew of the Sea Wolf, Wallace Menhaden Company, for allowing me to participate in a fishing expedition off Cam- eron, La.; the Menhaden Advisory Committee of the Gulf States Marine Fisheries Commission; and T. B. Ford, W. W. Fox, Jr., J. Geaghan, and an anonymous reviewer for their interest and advice. This report resulted from research supported by the Louisiana Sea Grant Program, Gulf States Marine Fisheries Commission, and Louisiana Board of Regents. 452 800 A 7 00- ,'\ (0 z 5-V \ \ °. 600- •*--,< ■f . \ y X — - - — . ^ 50 0- .*- \ \ Ui s O 400- X^ \ \ o X. \ o /N ^»v r- >»3x Z 300- / / I ^ 200 X N < % \ u 100i 0 I 1 » r 1 1 1 1 r 400 500 600 700 800 900 1000 1100 1200 1300 EFFORT IN "1971" VESSEL-WEEKS 800- B £ 700- 0 h (j 600- E r- W 500- 8.1 ( O O 400- 1 1 1 z h\ 1 I 300^ \\ O h 5 200h A V / v 100- 0- 0 400 800 1200 1600 1000 2400 2800 EFFORT IN VESSEL-WEEKS FIGURE 4. — Comparing the difference between fitting catch data for Atlantic menhaden when the effort is measured as (A) Schaaf's (1975a) 1971-vessel weeks and (B) vessel weeks. Curves were fit after the technique of Marchesseault et al. (1976). LITERATURE CITED BANNEROT, S. E, AND C. B. AUSTIN. 1983 . Using frequency distributions of catch per unit effort to measure fish-stock abundance. Trans. Am. Fish. Soc. 112:608-617. BEDDINGTON, J. R. 1979. On some problems in estimating population abun- dance from catch data. Rep. Int. Whal. Comm. 29:149- 154. CLARK, C. W, AND M. MANGEL. 1979. Aggregation and fishery dynamics: a theoretical study of schooling and the purse seine tuna fisheries. Fish. Bull., U.S. 77:317-337. FOWLER, C. W. 1980. A rationale for modifying effort by catch, using the CONDREY: DENSITY-DEPENDENT SEARCHING TIME sperm whale of the North Pacific as an example. Rep. Int. Whal. Comm. (Spec. Issue 2), p. 99-102. FOX, WW., JR. 1974. An overview of production modeling. U.S. Dep. Commer., Natl. Mar. Fish. Serv. Southwest Fish. Cent. Adm. Rep. LJ-74-10. GARROD, D. J. 1977. The North Atlantic cod. In J. A. Gulland (editor), Fish population dynamics, p. 216-242. John Wiley & Sons, N.Y. Graham, M. 1935. Modern theory of exploiting a fishery, and applica- tion to North Sea trawling. J. Cons. Cons. Int. Explor. Mer 10:264-274. Gulland, J. A. 1956. On the fishing effort in English demersal fisheries. Fish. Invest. Minist. Agric. Fish. Food (G.B.) Ser. II, 20(5), 41 p. 1964. Catch per unit effort as a measure of abun- dance. Rapp. P-V. Reun. Cons. Perm. Int. Explor. Mer 155:8-14. HASSEL, M. P 1978. The dynamics of arthropod predator-prey systems. Princeton Univ. Press, Princeton, N.J., 237 p. HOLLING, C. S. 1959a. The components of predation as revealed by a study of small mammal predation of the European pine saw- fly. Can. Entomol. 91:293-320. 1959b. Some characteristics of simple types of predation and parasitism. Can. Entomol. 91:385-398. MARCHESSEAULT, G. D., S. B. SAILA, AND W J. PALM. 1976. Delayed recruitment models and their application to the American lobster (Homarus americanus) fishery. J. Fish. Res. Board Can. 33:1779-1787. MACCALL, A. D. 1976. Density dependence of catchability coefficient in the California Pacific sardine, Sardinops sagaz caerulea, purse seine fishery. Calif. Coop. Oceanic Fish. Invest. Rep. 18, p. 136-148. PALOHEIMO, J. E., AND L. M. DICKIE. 1964. Abundance and fishing success. Rapp. P-V. Reun. Cons. Int. Explor. Mer 155:152-163. PETERMAN, R. M. 1980. Dynamics of native Indian food fisheries on salmon in British Columbia. Can. J. Fish. Aquatic Sci. 37:561- 566. PETERMAN, R. M., AND G. J. STEER. 1981. Relationship between sport-fishing catchability co- efficients and salmon abundance. Trans. Am. Fish. Soc. 110:585-593. POPE, J. G., AND D. J. GARROD. 1975. Sources of error in catch and effort quota regulations with particular reference to variations in the catchability coefficient. Int. Comm. Northwest Atl. Fish. Res. Bull. 11, p. 17-30. RICKER, W E. 1975. Computation and interpretation of biological statis- tics of fish populations. Fish Res. Board Can. Bull. 191, 382 p. ROTHSCHILD, B. J., AND A. SUDA. 1977. Population dynamics of tuna. In J. A. Gulland (editor), Fish population dynamics, p. 309-334. John Wiley & Sons, N.Y. SCHAAF, W E. 1975a. Status of the Gulf and Atlantic menhaden fisheries and implications for resource management. Mar. Fish. Rev. 37(9):l-9. 1975b. Fish population models: Potential and actual links to ecological models. In C. S. Russell (editor), Ecological modeling in a resource management framework: the pro- ceedings of a symposium, p. 211-239. John Hopkins Uni- versity Press, Baltimore, Md. SCHAAF, W E., AND G. R. HUNTSMAN. 1972. Effects of fishing on Atlantic menhaden stock: 1955- 1969. Trans. Am. Fish. Soc. 101:290-297. ULTANG, 0. 1976. Catch per unit effort in the Norwegian purse seine fishery for Atlanto-Scandian (Norwegian Spring Spawn- ing) herring. FAO Fish. Tech. Pap. 155, p. 91-101. 453 COMMUNITY AND TROPHIC ORGANIZATION OF NEKTON UTILIZING SHALLOW MARSH HABITATS, YORK RIVER, VIRGINIA1 Stephen M. Smith,2 James G. Hoff,2 Steven P. O'Neil,3 and Michael P. Weinstein3 ABSTRACT Nekton were collected by trawl and Wegener ring at each of two stations within tidal creeks and at adjacent shoal stations in the polyhaline and oligomesohaline zones of the York River estuary, Virginia. Species richness was significantly higher (P < 0.05) in Goalders Creek (oligomesohaline) compared with Blevins Creek (polyhaline) and may reflect the general absence of stenohaline marine "southern" taxa which seasonally occupy the tidal creeks of warm-temperate estuaries. In general, diversity was low in both creeks with dominance mainly shared by two species, Leiostomus xanthurus and Anchoa mitchelli. Trinectes maculatus also was abundant at the Goalders Creek shoal station, about 200 m outside of the creek mouth. A detailed analysis of the distribution of L. xanthurus indicated that after recruitment ceased, this species was largely resident in the creeks for several months, only emerging in the fall (October). Furthermore, emigration from Blevins Creek occurred earlier than at the upstream locality. Of the "transient" marine species encountered, L. xanthurus seemed to be the most tidal creek dependent. However, this may be due partly to the collection methodology employed. Diet composition of six dominant species comprising >98% of the total number of individuals collected indicated that all were essentially trophic opportunists feeding on a wide variety of food items. Ontogenetic shifts in diet were also observed for the five most abundant species. Lack of dietary specialization and the consequently large degree of diet overlap in all species may reflect the nonlimit- ing nature of food abundance in the primary nurseries, however a seasonal change in relative fullness values may indicate periodic food scarcity. The structural and functional role of shallow es- tuarine habitats has received increasing attention in the past few years. Although widely recognized as primary nurseries, two of these habitats, marshes and seagrass meadows, have only re- cently come under scrutiny in the lower Chesa- peake Bay Much of the impetus for these efforts was derived from priorities established by the Chesapeake Bay Program (Environmental Pro- tection Agency 1979). As a result, areas covered with submerged aquatic vegetation (SAV) were investigated between 1977 and 1981. The role of SAV as primary nurseries, especially for blue crabs, Callinectes sapidus, was confirmed (Orth4). It was suggested (and, in some cases, demon- strated experimentally) that a principal function of vegetated habitats was that of predation re- 'Contribution No. 1160, Virginia Institute of Marine Science. 2Virginia Institute of Marine Science and School of Marine Science, College of William and Mary, Gloucester Point, VA 23062. 3Virginia Institute of Marine Science and School of Marine Science, College of William and Mary, Gloucester Point, Va.; present address: Lawler, Matusky and Skelly Engineers, 1 Blue Hill Plaza, Pearl River, NY 10965. 4R. J. Orth, Marine Scientist, Virginia Institute of Marine Science, Gloucester Point, VA 23062, pers. commun. May 1982. Manuscript accepted January 1984. FISHERY BULLETIN: VOL" 82, NO. 3, 1984. fugium for the early life stages of many species (Heck and Thoman 1981; Lascara 1981; see also Nelson 1979 and Coen et al. 1981). Although sea- grass meadows were contrasted with immediately adjacent unvegetated areas, their value compared with tidal salt marshes was not established. To place the utilization of SAV and tidal creeks by the immature life stages of dominant species into better perspective, Weinstein and Brooks (1983) undertook a direct comparison of these areas along a contiguous marsh-seagrass ecosys- tem on the eastern shore of Virginia. Two results of their study were the observations that the domi- nant finnsh in the area — spot, Leiostomus xanthurus — was nearly four times more abun- dant in the tidal creek throughout the study period, and that larger juvenile and adult blue crabs made nearly equal use of both habitats. A further outcome of this study was the obvious need for additional inventories of shallow waters of the lower Chesapeake Bay with regard to the relative value of different habitats to resident species. For this reason we have extended our program to include a survey of habitat utilization by nekton occupying oligohaline and polyhaline tidal creeks 455 vol FISHERY BULLETIN: VOL. 82. NO. 3 of the York River estuary and have included a comparison of the tidal creeks with adjacent shoal areas (about 200 m outside of the tidal creek mouths). Along with data on community composi- tion and structure, we have begun baseline studies of diet composition of dominant species. These as- pects of community ecology of York River tidal marshes are reported herein. STUDY AREA AND METHODS The York River is one of six major tributaries which enter the Chesapeake Bay along its western shoreline. The narrow estuary covers about 208 km2 and extends 46 km from Tue Marsh Light to West Point, Va. (Fig. 1). The upper portion of the York is characterized by broad, shallow flats and tidal creeks dominated by Spartina spp. along the shoreline. Upstream the river channel averages 8-9 m in depth, but broadens downstream and reaches a maximum depth of 18-23 m. The Guinea Marshes (Fig. 1) is a major Spartina alterniflora- dominated marsh system located near the estuary mouth. Much of the adjacent shallows in this re- gion is carpeted with dense stands of eelgrass, Zostera marina, and widgeon grass, Ruppia maritima. Salinities are usually in the polyhaline range. Two tidal creeks were selected for study. Goal- ders Creek (Fig. Llocation A) was located in the oligohaline-mesohaline zone just below the city of West Point. Blevins Creek (Fig. l:location B), a part of the Guinea Marsh system, was situated in the polyhaline zone where salinities always ex- ceeded 16%o. In each creek, one station was estab- lished as far upstream as possible, one near the creek mouth, and one about 200 m offshore. All collections were made monthly on consecutive days (March-October 1983), with sampling ini- tiated as close to daytime high tides as possible. Miller and Dunn (1980) collected a greater propor- tion of fish with stomachs containing food at this time. Creek bottoms were of the mud-silt type and ranged from 1 to 1.5 m deep. The primary collecting device utilized in this study was a 4.9 m otter trawl consisting of wings and body of 19 mm mesh and a liner of 6.3 mm mesh. This gear was towed for 2-min intervals at a speed of about 1.0 m/s at each station. In an earlier study at Guinea Marshes, Orth and Heck (1980) demonstrated that six hauls of the trawl were necessary in seagrass habitats to attain asymp- totic returns on community information ( as judged by several diversity indices). Because of the ex- pectedly lower diversity in the tidal creeks (Wein- stein and Brooks 1983), it was determined that four consecutive hauls at each tidal creek station would be sufficient to attain the same level of community information. Ancillary collections were taken in the tidal creek with a modified Wegener ring (Wegener et al. 1973). The gear was used in depths <1 m, in the vicinity of the trawling stations. The side walls of the ring consisted of 1.5 mm woven netting, with the original design of the gear being changed to include a 305 mm "skirt" and chain attached to the FIGURE 1. — Location of study areas in York River estuary, Va. Three perma- nent sampling stations were estab- lished— near the headwaters, at the mouth, and 200 m offshore of an oligo- mesohaline (A = Goalders Creek) and polyhaline (B = Blevins Creek) tidal creeks. 456 SMITH ET AL.: ORGANIZATION OF NEKTON leadline hoop which helped fill contours along the bottom. Samples were obtained by tossing the ring from shore and then applying rotenone at a 30 ppb concentration within the confines of the net (Weinstein and Brooks 1983). Stricken fishes were then captured with dip net or swept off the bottom. Fishes were sorted into 20 mm (or less) size classes and up to 20 randomly selected individuals from each class used for trophic analyses. Stomach fullness was recorded as a relative fullness index (RFI) value (Hyslop 1980). Stomach contents were subsequently analyzed using a modified Carr and Adams (1972) sieve fractionation technique. Total dry weights for each sieve fraction were then ob- tained and proportioned among the prey taxa identified from a five drop subsample taken before drying. On the assumption that particles of equal size have approximately the same weight, this method agglomerates food particles of roughly the same size. The Carr and Adams technique provided rapid, accurate identification of food items for a large number of stomachs and has been used with suc- cess by several investigators (Sheridan 1979; Sheridan and Livingston 1979; Stoner 1980; Livingston 1982). A useful modification employed in this study was the application of a low pressure stream of compressed air delivered through a Pas- teur pipette which greatly aided in removing food particles adhering to the finer screens of the sieves. Numerical classification analysis used here is similar to the procedures employed by Weinstein (1979) and Weinstein and Brooks (1983). Briefly, marsh creek communities and trophic ecology of dominant species were compared by classification methods using "normal" and "inverse" classifica- tion (Clifford and Stephenson 1975). The former method groups sites (or predators) by their species (or prey taxon) attributes; while inverse classifica- tion (used only for community analysis purposes here) groups species according to their site of oc- currence (i.e., the sites become the attributes of the species). Similarity between sites (or pred- ators) was calculated as the complement of the Canberra metric index: N [l/n][!]\xij - x2j\Kxij + x2j) (1) where n = number of attributes, and x\j and x2j are the values of the jth attribute for any pair of entities. The merits of the Canberra metric index have been discussed by Clifford and Stephenson (1975). Separate matrices were constructed for each comparison from untransformed, pooled monthly data and clustered by the unweighted pair, group-average strategy (Clifford and Stephenson 1975). Species occurring at only one station (sin- gletons) were eliminated prior to the community analyses. Combined trawl and Wegener ring data were used separately in these procedures. Den- drograms for site and species dissimilarity (com- munity analyses) were constructed and cross- tabulated in a two-way coincidence table. All nekton were preserved in 10% buffered For- malin5. Standard length (SL, carapace width for blue crabs) was recorded for all taxa. Up to 30 individuals/species were measured from each col- lection, subsampling for lengths was employed when sorted collections contained more than 30 individuals of a given species. Prior to each collec- tion, temperature and salinity were recorded with an immersion thermometer and a temperature- compensated refractometer. RESULTS Abundance and Seasonality Only two species — spot and the bay anchovy, Anchoa mitchelli — comprised >90% of the total number of individuals captured at Blevins Creek and adjacent shoals. Using this same criterion, upstream densities were more equitably distrib- uted with four species in the creek and six on the shoal sharing dominance (Table 1). Blue crabs; white perch, Morone americana; and the hog- choker, Trinectes maculatus, were in this group in Goalders Creek, while in late summer and fall the Atlantic croaker, Micropogonias undulatus, and the weakfish, Cynoscion regalis, were also abun- dant at the shoal station (Table 1). Species richness (S) was also greater in all months in the Goalders Creek system compared with the polyhaline Blevins Creek (Fig. 2) and was significantly greater for the entire study period (Wilcoxon sign-ranks test; P < 0.05). No apparent trend, however, was evident in the number of indi- viduals captured at each locality (Fig. 2), except that peak catches of two dominant species, spot and bay anchovy, were greater in Blevins Creek and resulted in the large disparity in catches in 5 Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 457 Table i. FISHERY BULLETIN: VOL. 82, NO. 3 -Pooled species abundance and percent composition for all trawl collections, York River estuary, Va. N = number of individuals. Species N % Species N % Species N % Goalders Creek upstream Goalders Creek downstream Goalders Creek shoal Leiostomus xanthurus 709 '70.69 Leiostomus xanthurus 1,646 '60.09 Trinectes maculatus 1,439 '50.31 Callinectes sapidus 83 '8.27 Trinectes maculatus 385 '14.05 Leiostomus xanthurus 412 '14.40 Anchoa mitchilli 69 '6.88 Anchoa mitchilli 373 '13 62 Anchoa mitchilli 310 '10.84 Morone americana 48 14.79 Callinectes sapidus 144 '5.26 Callinectes sapidus 308 '10.77 Alosa aestivalis 35 3.49 Ictalurus catus 75 2.74 Micropogonias undulatus 91 '3.18 Morone saxatilis 27 2.69 Morone americana 57 208 Cynoscion regalis 77 '2.69 Ictalurus catus 25 2.49 Brevoortia tyrannus 17 0.62 Ictalurus catus 59 2.06 Brevoortia tyrannus 3 0.30 Cynoscion regalis 16 0.58 Morone americana 57 1.99 Pomatomus saltatrix 1 0.10 Morone saxatilis 8 029 Anchoa hepsetus 33 1.15 Peprilus alepidotus 1 0.10 Paralichthys dentatus 3 0.11 Opsanus tau 28 0.98 Anguilla rostrata 1 0.10 Micropogonias undulatus 3 0.11 Ophidion margmata 13 0.45 Trinectes maculatus 1 0.10 Alosa aestivalis 3 0.11 Paralichthys dentatus 12 0.42 Total 1,003 100.00 Opsanus tau 3 0.11 Anguilla rostrata 8 0.28 Pomatomus saltatrix 2 0.07 Morone saxatilis 5 0.18 Peprilus alepidotus 1 0.04 Menticirrhus saxatilis 3 0.11 Alosa sapidissima 1 0.04 Pomatomus saltatrix 2 0.07 Anguilla rostrata 1 0.04 Peprilus alepidotus 1 0.04 Syngnathus fuscus 1 0.04 Brevoortia tyrannus 1 0.04 Total 2,739 100 00 Gobiosoma bosci Total 1 2.860 0.04 100.00 Blevins Creek upstream Blevins Creek downstream Blevins Creek shoal Leiostomus xanthurus 1,301 '46.51 Leiostomus xanthurus 1,501 '73.69 Leiostomus xanthurus 1,935 '80.03 Anchoa mitchilli 1,248 '44.62 Anchoa mitchilli 414 '20.32 Anchoa mitchilli 385 '15.92 Callinectes sapidus 162 5.79 Callinectes sapidus 103 5.05 Callinectes sapidus 49 2.03 Trinectes maculatus 35 1.25 Paralichthys dentatus 6 0.30 Trinectes maculatus 10 0.42 Brevoortia tyrannus 23 0.82 Menidia menidia 5 0.24 Bairdiella chrysoura 10 0.42 Menidia menidia 17 0.61 Gobiosoma bosci 2 0 10 Paralichthys dentatus 6 0.25 Fundulus heteroclitus 5 0.18 Trinectes maculatus 2 0.10 Opsanus tau 4 0.17 Gobiosoma bosci 3 0.11 Cobiesox strumosus 2 0.10 Micropogonias undulatus 3 0.12 Cynoscion regalis 1 0.04 Peprilus tnacantus 1 0.05 Stenotomus chrysops 2 0.08 Morone americana 1 0.04 Microgobius thalassinus 1 0.05 Anchoa hepsetus 2 0.08 Microgobius thalasslnus 1 0.03 Total 2,037 100.00 Syngnathus fuscus 2 0.08 Total 2,797 100.00 Gobiosoma bosci Menidia menidia Synodus foetena Centropnstis striata Rachycentron canadum Brevoortia tyrannus Urophyas regia Total 2 2 2 1 1 1 1 0.08 008 0.08 0.04 0.04 0.04 0 04 2,418 100.00 ' Species comprising -90°o of the total number of individuals. May and July. Except for the spike seen in Figure 2, resultant from a large influx of bay anchovy into Blevins Creek in July, combined catches of all other taxa were at a minimum for the summer months (June-August) coincident with peak summer temperatures. Seasonal abundance of the more common species was typically associated with recruitment of young-of-year individuals into the tidal creeks and adjacent shoals. Young spot dominated in both creeks but were subsequently replaced by post- larval and juvenile bay anchovy (July), and there- after at Goalders Creek by hogchoker (August), weakfish (August-September), and Atlantic croaker (October). In addition, white catfish, Ic- talurus catus, and white perch were frequently captured in the Goalders Creek vicinity during early spring when salinities were at their lowest recorded levels. Because of the overall seasonal abundance, it was possible to examine spatial and temporal distributions of spot in greater detail (Fig. 3, Table 2). As expected, spot were more abundant outside of the creeks very early in re- cruitment; but by June had established a greater degree of residency within the creeks compared with the adjacent shoals. This pattern of large creek to shoal abundance ratios (Fig. 3) continued (with a single exception) until October when spot emerged from Blevins Creek. Note, however, that at the termination of the sampling program, this emigration had not taken place upstream. Similar TABLE 2. — Relative abundance of Leiostomus xanthurus at tidal creek stations, York River estuary. Location Mar. Apr. May June July Aug Sept. Oct Goalders Creek Upstream 0 1 401 85 51 58 68 45 Downstream 0 52 624 318 236 172 93 151 Blevins Creek Upstream 0 19 493 326 292 60 105 6 Downstream 0 18 586 337 102 150 221 87 458 SMITH ET AL.: ORGANIZATION OF NEKTON patterns of earlier downstream emigration were observed during 1976-78 in the Cape Fear estuary, N.C. (Weinstein and Walters 1981; Weinstein pers. obs.), the cause of which remains unexplained. It is evident from Table 2 that spot may have been more 2000r 1800- MAR APR MAY JUN JUL AUG SEP OCT FIGURE 2. — Total numbers of individuals and species captured in monthly tidal creek collections. Temperatures and salinities are mean values recorded at each creek in each month. restricted in their upsteam movement in Goalders Creek where salinities averaged about 2"L lower at the upsteam station than at the creek mouth. The ratio of upstream to downstream station catch was twice as high at Blevins Creek, supporting this pattern. Other species which seem to prefer a specific portion of the tidal creek to shoal habitat gradient included hogchoker, weakfish, and At- lantic croaker which had creek-to-shoal ratios (over all months) of 0.13, 0.10, and 0.02, respec- tively. Moreover, these species were far more abundant at upstream sites (Table 1). Community Composition A two-way coincidence table, using a similarity value of 0.200 to define clusters (Clifford and Stephenson 1975), was constructed in order to summarize species and site relationships for pooled monthly collections at each station (Table 3). Included in this analysis are samples collected with the Wegener ring, a gear which was expected to be more successful in collecting both cryptic species (e.g., Gobiidae) and shore-zone taxa (e.g., cyprinodonts [Cyprinodontidae] and silverside [Atherinidae]). It should be pointed out, however, that any comparisons between the Wegener ring and trawl samples are qualitative since no at- tempt was made to compare gear selectivity, effi- ciency, and area encompassed by a unit effort for each sampling device (Weinstein and Brooks 1983). Species group IV (Table 3) was generally the most ubiquitously distributed assemblage over the range of environmental factors (particularly FIGURE 3. — Relative densities of Leiostomus xanth- urus at tidal creek (values shown are monthly means for both creek stations) and shoal sampling localities. May values for Blevins Creek are drawn to half-scale. Values appearing above histograms are ratios of creek to shoal densities. 20.5 0. 52* 500- GOALDERS CK [10311 BLEVINS CK ■ CREEK a SHOAL ■ CREEK a SHOAL 400- 0.13 1.4 300- 15401 200- 100- o.: 28 i.; » 2.9 1 v V 0. 16 6 0 1.C f 4 5 1 1 1 III J 1 ll APR MAY JL IN JUL AUG SEP OCT APR MAY JUN JUL AUG SEP OC :f1 459 FISHERY BULLETIN: VOL. 82, NO. 3 TABLE 3. — Two-way coincidence table comparing stations (Groups A-D) and species (groups I-VI) associations at York River estuary sites. Clustering by unweighted-pair group-average; similarity index = Camberra metric, all data log10(x + 1) transformed, single station occurrences dropped. G = Goalders Creek, B = Blevins Creek, U = up- stream, D = downstream, S = shoal stations, W = Wegener ring samples. Species A B C D Group GU GD GS BU BD BS GUW BDW BUW GDW 1 Paralichthys dentatus 3 12 6 6 1 Micropogonias undulatus 3 91 3 Opsanus tau 3 28 4 Anchoa hepsetus 33 2 Cynoscion regalis 16 77 1 Syngnathus fuscus 1 2 1 II Pomatomus saltatrix Peprilus alepidotus Anguilla rostrata Morone saxatilis Ictalurus catus 1 1 1 27 24 2 1 1 8 75 2 1 8 5 59 1 Brevoortia tyrannus 3 17 1 23 1 7 III Alosa aestivalis 35 3 1 1 5 Fundulus heteroclitus 5 1 3 23 1 Menidia beryllina 3 6 3 IV Morone americana 48 57 57 1 2 1 10 Leiostomus xanthurus 709 1,646 412 1,301 1.501 1,935 9 14 14 3 Anchoa mitchilli 69 373 310 1,248 414 385 3 4 32 Tnnectes maculatus 1 385 1,439 35 2 10 1 Callinectes sapidus 83 144 308 162 103 49 Gobiosoma bosci 1 3 2 2 2 13 4 106 Menidia menidia 17 5 2 24 V Fundulus majalis 2 1 VI Microgobius thalassinus 1 1 salinity) examined. Nonetheless, within this group were several species which displayed area- specific distributions, in either relative numbers or presence/absence in a given creek system. Examples of the former include the hogchoker and bay anchovy and of the latter, white perch, which was far more prevalent upstream. The Goalders Creek nekton community was also dominated by members of groups I and II, whose members were rare or absent at downstream localities. Remain- ing species were generally not captured in suffi- cient numbers to depict their role in defining community structure in each area. Trophic Analysis Six species (Fig. 4) were sufficiently abundant in time and space to allow a comparative trophic analysis to be undertaken. Collectively, they ex- ceeded 989r of the total number of individuals cap- tured during this study. Prey taxa were defined on the basis of 39 categories (Table 4). All but two — miscellaneous (MISC) and unidentified (UID) — were mutually exclusive. These two did not exceed 179c of the total diet composition of any one species and were generally much lower than this amount. The dietary relationships of these six species are summarized across all sampling strata by the dendrograms appearing in Figure 4. With the ex- ception of summer flounder, Paralicthys dentatus, sufficient numbers of individuals were captured to also allow partitioning by size classes. Such on- togenetic summaries are shown in Figure 5. Although more than 2,600 specimens were examined for diet composition, sample sizes were not sufficient in the first year of the study to examine details of seasonal nor spatial food utili- zation in all species except spot (O'Neil 1983). In addition, several species were only abundant in a restricted area (Table 1) or attained peak abun- dance in a relatively narrow time frame, preclud- ing dietary comparison of dominant species (Fig. 6). Diet overlap was greatest between white perch and hogchoker (Fig. 4). Dietary items shared by these species included clam siphons, Leoptochirus plumulosus, and other gammarid amphipods. Both predators were generally habitat-specific, with young-of-year white perch more abundant in the creeks and hogchoker prevalent on the river shoals (Table 1). Largely because of fish (TEL) included in the diet of larger individuals, white catfish displayed somewhat less overlap in its diet compared with the former species, but because of its partial piscivorous habits shared this similar- ity with the major fish predator captured in our trawl samples — the summer flounder. Both small summer flounder and white perch (<61 mm) also consumed substantial quantities of the mysid shrimp Neomysis americana. 460 SMITH ET AL.: ORGANIZATION OF NEKTON PAR DEN 81 160 rPLA NEM NER CLS Na PAL TEL MOR AME 21 160 PLA NER i-MISC CLS Na -AMP Lp GAM UID -COR 28 28 146 163 Me MISC TRI MAC 2 1 I60 NER CLS TEL UID DET MISC -CAP 179 27 8 ICTCAT 8I 280 r-NER MAC 126 158 LEIXAN I6 I25 PLA -CHL FOR NEM -Eh NER MAL OLI -SPI CLS OST HA 1 HA 2 LP MISC J-CAL La s CYA l-Me DET 1722 1753 ANC MIT 2I I00 t— PI -BIV CRZ wr_ HA1 }-HA2 CAL Na LP UID :-GAM 242 330 i-DET MISC FIGURE 4. — Trophic comparisons among dominant predators ( those comprising > 98% of the total number of individu- als captured), pooled across all sampling strata. PAE DEN = Paralichthys dentatus; MOR AME = Morone americana; TRI MAC = Trinectes maculatus; ICT CAT = Ictalurus catus; LEI XAN = Lewstomus xanthurus; ANC MIT = Anchoa mitchilli. Ratios appearing below histograms represent stomachs with food as a proportion of total stomachs. Values above histograms are size ranges (standard lengths). Diet composition was compared using Can- berra metric and unweighted-pair group-average clustering strategy, data untransformed. Prey designations are defined in Table 4. TABLE 4. — Prey categories used for trophic comparisons. All but unidentified (UID) and miscellaneous (MISC) are mutually exclusive feeding categories. AMP Amphipoda La Leucon amencanus BIV Bivalves Lp Leptocheirus plumulosus BRA Branchipoda MAC Macoma sp CAL Calanoids MAL Maldanidae CAP Caprellidae Me Monoculodes edwardsi CLS Clam siphons MISC Miscellaneous CHI Chironomidae Na Neomysis americana CHL Chlorophyta NEM Nematoda COR Corophiidae NER Nereidae Crs Crangon septemspmosa OLI Oligochaeta CRZ Crab zoea OST Ostracods Cs Callinectes sapidus PAL Palaemonidae CYA Cyathura PI Polydora ligni DET Detritus PLA Plant matter Eh Eteone heieropoda POL Polychaeta Et Edotea tribola SPI Spionidae FOR Foraminifera TEL Teleosteii GAM Gammaridae UID Unidentified remains HA1 Harpacticoid 1 XAN Xanthidae HA2 Harpacticold 2 Among the six species examined, spot and bay anchovy displayed the least dietary overlap, with the former exhibiting the greatest dietary diver- sity (consisting mainly of benthic prey items), and the latter including a greater percentage of plank- tonic food items in its diet. However, the large variety of prey items consumed by all species indi- cates that each is a trophic opportunist (Darnell 1958, 1961; Carr and Adams 1973; Sheridan and Livingston 1979; Livingston 1982). Ontogenetic shifts in diet were evident for each of the species examined (Fig. 5). In spot, dietary importance of calanoid copepods declined in fish >20 mm SL, while harpacticoid copepods in- creased in importance in fish 21-80 mm SL. Con- currently, the percentage of various polychaetes and gammarid amphipods slowly increased in their diet. Nematodes also became less important with increasing size. Interestingly, spot stomachs at all sizes contained clam siphons and maldanid tails, indicating that specific parts of larger prey were important dietary items. Ostracods and crab zoea were abundant food items for small (<21 mm SL) bay anchovy; larger 461 16 9x UJX 2< it < (7) uji- w2 ZQC o< (7) 3 QC 3C/) -13 -NER ~CLS Na GAM msc -COR -UID 30 101-1 BESD PLA ns Me MISC ~5T 20 FOR L-NER SPI t-OST HE2>HA1 La -CAL ^CYA DET POL Eh CLS BRA UID CLS Na GAM NER -AMP -COR 14 81 ^.NER ^Et -Lp 55 101 22 120 + NER Lp GAM Lp CAP GAM GAM XAN TEL XAN UID UID DET MISC 1-UID -MISC 56 FIGURE 5. — Ontogenetic comparisons in diet among the five most dominant species captured in this study. Sample size (stomachs with food) appear below histograms, size increment (standard length) above. Prey designations are defined in Table 4. 462 SMITH ET AL.: ORGANIZATION OF NEKTON 10 < Q > a a. UJ m 2 FIGURE 6. — Seasonality of selected taxa at tidal creek and shoal stations, York River estuary, Va. Numbers on x axis are sample sizes too small to plot. Paralichthys dentatus ■ GOALDERS CK G BLEVINS CK ■ 0 0 0 ■ 0 0 0 50 25 0 84 Ictalurus catus A _*o |o I 107 Morone americana 1 0 ■O 1500 1000 500 « a & MAR APR MAY JUN JUL AUG SEP OCT MONTH individuals incorporated more Neomysis ameri- cana, calanoid copepods, and gammaridean am- phipods into their diet. In hogchokers, however, gammarids predominated in smaller individuals, <61 mm SL, but with increasing size became somewhat less important and were replaced by nereid polychaetes and clam siphons. Hogchokers had the highest proportion of unidentified re- mains (UID) of any predator examined due to the high level of maceration characteristic of this species. Neomysis americana was clearly the dominant prey item of small (<60 mm SL) white perch, but became less important in the diets of larger indi- viduals which fed increasingly on gammarid am- phipods and clam siphons. White catfish <120 mm SL also fed conspicuously on amphipods but, uniquely among the predators examined, also fed upon xanthid crabs, and at larger sizes incorpo- rated a substantial proportion of fishes into their diet. Relative fullness indices displayed varying trends on a species-specific basis (Table 5). Values declined in the later part of the study for white perch and white catfish, whereas no apparent trends were observed for other species. TABLE 5. — Monthly relative fullness index (RFI) for six domi- nant fishes in the York River estuary, Va. Values are means for all individuals examined. Species Mar. Apr. May June July Aug. Sept. Oct Anchoa mitchilli 0.66 2.73 1.60 2.16 2.72 2.43 1.03 Ictalurus catus 2.53 2.20 1.84 1.50 1.19 1 74 1.00 Leiostomus xanthurus 2.45 2.45 2.62 2.59 2.77 2.51 1.98 Morone americana 265 1.59 1.14 1.00 1.00 0.50 Paralichthys dentatus 2.60 3.33 2.50 3.50 4.00 Trinectes maculatus 1.82 1.91 1.45 1.47 1.60 0.88 1.30 DISCUSSION Species Composition and Abundance Along with other recent studies of shallow- water nekton (Orth and Heck 1980; Weinstein and Brooks 1983; Heck and Thoman 1984), the present effort provides additional information on the use of inshore habitats of the lower Chesapeake Bay A striking characteristic of the marsh nekton com- munity in the lower Chesapeake Bay is the gener- ally low diversity of the constituent fauna and the high level of dominance of only a few species. On both the eastern shore of Virginia (Delmarva Peninsula) (Weinstein and Brooks 1983) and in the 463 FISHERY BULLETIN: VOL. 82, NO. 3 York River estuary, no more than two species com- prised >90% of the total number of individuals captured at polyhaline tidal creek stations. On the average, spot comprised 71.8% of this total. Only in the oligomesohaline Goulders Creek was species dominance shared by more than two species (Table 1), but once again, spot predomi- nated with 65.4% of the total. By comparison, Hackney et al. (1976) reported a mean of seven species totaling >90.0% at four trawl stations in their study of a mesopolyhaline tidal creek in Georgia, while in several studies in South and North Carolina (Cain and Dean 1976; Bozeman and Dean 1980; Weinstein et al. 1980) three to nine species (x = 7) comprised this total. Species rich- ness was significantly greater in Goalders Creek than in the polyhaline Blevins Creek system. This is somewhat surprising since previous studies have often shown that diversity decreases for both fishes and invertebrates in the upstream direction (Dahlberg 1972; Boesch 1977; Gainey and Green- berg 1977; Weinstein et al. 1980). The absence or scarcity of stenohaline marine species derived largely from the seasonally abundant southern Carolinian ichthyofauna may partially explain this difference in the York River estuary In North Carolina, for example, these taxa increased species richness in polyhaline tidal creeks, espe- cially near the estuary mouth (Weinstein 1979; Weinstein et al. 1980). Also present in estuaries below the Chesapeake Bay are species with warm-temperate affinities which share domi- nance with spot and bay anchovy, including Mugil cephalus, M. curema, Lagodon rhomboides, Paralichthys lethostigma, Bairdiella chrysura, and the brown shrimp, Penaeus aztecus. These species are much less common in the Chesapeake Bay Another noteworthy finding is that species re- placement does not occur from regional and north- erly taxa. For example, Ophidion marginata, Stenotomus chrysops, Urophysis regia, and Cen- tropristis striata were only rarely encountered in our studies (Weinstein and Brooks 1983). In the present investigation, these species were re- stricted to shoal stations outside of the tidal creek mouths (Table 1). Thus, there appears to be an underutilization of shallow nursery habitats by transient marine fishes in the Chesapeake Bay compared with the lower latitude estuaries (for more detailed discussion see Weinstein and Brooks 1983 and Heck and Thoman 1984). This difference is perhaps due partly to the unique loca- tion of the Chesapeake Bay in the transition zone between faunal provinces (Briggs 1974) with neither taxonomic group able to adapt fully to conditions (primarily temperature regimes and their variance?) associated with this transition zone. The recent geological and evolutionary his- tory of northern estuaries, including the Chesa- peake Bay (Shubel and Hirschberg 1978), may also play a role in determining the degree of estuarine dependency of local faunas. A unique aspect of this study was the opportu- nity to compare utilization of the tidal creeks with adjacent shoal areas. Previously, these compari- sons had to be made among collections with differ- ent gears (and their associated selectivity and effi- ciency) or in different years or by different inves- tigators (Chao and Musick 1977; Markle 1976). The results for spot are of interest because of the general dominance of this species in many es- tuaries along the Atlantic and Gulf coasts. Recently, Weinstein and Walters (1981), Wein- stein and Brooks (1983), and Weinstein (1983) de- scribed the importance of marshes, specifically tidal creeks, to this species and the relationship between productivity and energy export via sev- eral fish vectors — from the marshes ultimately to the marine environment (Weinstein 1981). Spot were recruited into upstream marshes of the York River estuary earlier than to downriver sites and tended to remain there longer (Fig. 3). Once re- cruited into the marshes (by June) spot reside here until fall, when they emigrated into deeper water, and finally (for most individuals) return to the marine environment. As expected, however, there is an upstream limit to utilization in oligohaline tidal creeks where we found densities of spot de- creased (Table 2) as salinities became more vari- able (approaching 0%o) and where temperature re- gimes became more unstable (Hackney et al. 1976). Whether the lower abundance of spot out- side of the creeks is due to differential mortality and/or habitat selection remains unknown. Other taxa, e.g., young-of-year Atlantic croaker, weak- fish, and hogchoker, apparently prefer the shoals and generally deeper water. There is little ques- tion that they are more abundant outside of the marshes (Chao and Musick 1977; Orth and Heck 1980; Weinstein and Brooks 1983; Middleton, unpubl. data). If, as many would argue, predation is a major regulator of local abundance and ulti- mately community structure, what protection would the homogenous, relatively unstructured shoals and flats afford these species? Considering the apparent physical and behavioral similarities, as well as recruitment dynamics between spot and Atlantic croaker, there does not seem to be any 464 SMITH ET AL.: ORGANIZATION OF NEKTON significant adaptive feature of the latter that would provide better survivorship in open waters (at least with regard to predation). Just how this species and others minimize the effects of preda- tion in open waters is an important research ques- tion for the future. Trophic Comparisons The six species examined in detail are clearly trophic opportunists and overlap in many food categories. In addition, each goes through distinct ontogenetic stages in feeding which include sig- nificant shifts in the portions of the water column searched. The prey taxa have been categorized by Darnell (1961), Qasim (1972), and Chao and Musick (1977) according to their vertical occur- rence in the water column from open waters to the bottom: fishes, macrozooplankton (e.g., Neomysis americana), microzooplankton (e.g., calanoid copepods), epibenthos (e.g., harpacticord copepods), infauna, and organic matter. At sizes <21 mm SL all five of the species examined appar- ently spent considerable periods foraging in the water column. Between 21 and 40 mm SL several species continued to feed on "pelagic" prey, al- though by this size the transition to benthic feed- ing was nearly complete in spot and hogchoker. Whether resource partitioning or dietary specialization (Chao and Musick 1977) occurs in these taxa as a means of reducing interspecific interactions is a matter of speculation. Without question, there are differences in feeding localities of the fishes examined — e.g., white perch and white catfish are generally restricted to oligo- haline habitats, while Atlantic croaker and hog- choker are more abundant on the shoals. Also noted are differences in seasonal abundance (Fig. 6), size related feeding distributions reflecting on- togenetic shifts (Fig. 5), morphological differences among predators (Chao and Musick 1977), etc. But whether or not any of these traits reflect past or present competitive pressures remains unknown. Food that is generally limiting for several of these species and others is currently an area of con- troversy. Currin et al. (in press) have suggested that predation, not resources, limits production rates of spot and Atlantic croaker in shallow marsh embayments in Albermarle Sound, N.C. In contrast, Weisberg and Lotrich (1980) found that growth rates of the mummichog, Fundulus heteroclitus, could be altered by manipulating fish density. Increased growth rates were also demon- strated with food enrichment experiments in sub- tidal areas. Similar findings were reported by Miklas and Reed ( in press ) for F. heteroclitus popu- lations in a tidal tributary of the Rhode River, Del. Our own findings of a seasonal decline in relative fullness index values in several species, along with a parallel decline in benthic biomass (T Fre- dette6), tend to support the possibility of periodic food scarcity. Trophic opportunism has often been cited in studies of estuarine fishes (Darnell 1958, 1961; Livingston 1982). Several investigators have pointed to the importance of omnivorous and on- togenetic progressions in feeding stages (Sheridan 1979; Stoner 1980; Livingston 1982) as obscuring distinct trophic relationships in nektonic food webs. Along with these difficulties are problems associated with the "snapshot" view often gained of the system. Numerically abundant species are likely to play the major role in conversion and production of organic materials in estuaries (and are, therefore, mainly responsible for the con- struct of food webs and energy flow therein), yet the identification of these species often comes from the sampling program itself. Thus, although the dominant species in this study, spot, is undoubtedly important in this re- gard, the selective nature of our sampling effort does not allow us to place this importance in proper perspective. It is probable that dominance, expressed in numbers and/or biomass of those species captured in this study, is shared and some- times surpassed by other local species not sampled quantitatively by this program. These include young-of-year bluefish, Pomatomus saltatrix; var- ious cyprinodonts, especially F. heteroclitus; an- chovies; and silversides. On an estuary-wide basis, we also do not completely "track" species distribu- tions in time and space (Purvis 1976) so that our already distorted view of local habitats cannot easily be extrapolated to system-wide consider- ations. Such difficulties occur in most studies and must be recognized and eventually accounted for in considerations of fish community ecology in estuaries. SUMMARY Tidal creeks of the York River estuary were characterized by distinct nekton communities dis- playing low diversity and dominated by relatively 6T. F. Fredette, Marine Scientist, Department of the Army Corps of Engineers, P.O. Box 631, Vicksburg, MS 39189, pers. commun. September 1982. 465 FISHERY BULLETIN: VOL. 82, NO. 3 few taxa. Of the many transient marine species ( Weinstein 1979; Weinstein et al. 1980) that utilize marsh creeks along the Atlantic coast, only the spot, Leiostomus xanthurus, seemed to actively select this habitat. Within creeks there also was an apparent upstream limit in abundance of this species at low salinities. Although not captured quantitatively in this study, Atlantic menhaden, Brevortia tyrannus, were often observed in large numbers in the creeks, especially upstream. Other species, particularly Atlantic croaker, weakfish, and hogchoker were captured in greater numbers in low salinity shoal waters (<5 m) adjacent to marshes. Compared with the polyhaline marshes and shoals, stations sampled upstream in oligo- mesohaline waters were more diverse and had a larger variety of taxa apparently utilizing this area as a primary nursery habitat. In addition to the species mentioned above, white perch; striped bass, Morone saxatilis; and white catfish were sea- sonally present as young-of-year in the area. Dietary composition of the six species examined in detail reflected that of trophic opportunism, with maximum dietary diversity displayed by spot. Ontogenetic progressions in diet also were observed in all species. Two species, summer flounder and white catfish, were piscivorous at larger sizes, feeding mainly on Anchoa spp. The apparent absence of specialization in any of these predators may reflect the general adequacy of food supplies in the primary nurseries. ACKNOWLEDGMENTS We thank K. Anderson, H. Brooks, and S. Webb for their aid in the field and laboratory and T Fredette and R. Diaz for giving freely of their time in helping us identify difficult food items. Any errors in their identification, however, are solely our responsibility. All drafts of the manuscript were typed by G. Dunaway. We also wish to express our gratitude to R. Anderson for providing several computer programs for our use. This study was funded by EPA Grant #R808707 to Michael Weinstein. LITERATURE CITED BOESCH, D. F. 1977. A new look at the zonation of benthos along the estuarine gradient. In B. C. Coull (editor), Ecology of marine benthos, p. 245-266. Univ. South Carolina Press, Columbia. BOZEMAN, E. L., JR., AND J. M. DEAN. 1980. The abundance of estuarine larval and juvenile fish in a South Carolina intertidal creek. Estuaries 3:89-97. BRIGGS, J. C. 1974. Marine zoogeography. McGraw-Hill Co., N.Y., 475 p. CAIN, R. L., AND J. M. DEAN. 1976. Annual occurrence, abundance and diversity offish in a South Carolina intertidal creek. Mar. Biol. (Berl.) 36:369-379. CARR, W. E. S„ AND C. A. ADAMS. 1972. Food habits of juvenile marine fishes: Evidence of the cleaning habit in the leatherjacket, Oligoplites saurus, and the spottail pinfish, Diplodus holbrooki. Fish. Bull., U.S. 70:1111-1120. 1973. Food habits of juvenile marine fishes occupying seagrass beds in the estuarine zone near Crystal River, Florida. Trans. Am. Fish. Soc. 102:511-540. CHAO, L. N., AND J. A. MUSICK. 1977. Life history, feeding habits, and functional morphol- ogy of juvenile sciaenid fishes in the York River estuary, Virginia. Fish. Bull., U.S. 75:657-702. CLIFFORD, H. T., AND W. STEPHENSON. 1975. An introduction to numerical classification. Acad. Press, N.Y, 229 p. COEN, L. D., K. L. HECK, JR., AND L. G. ABELE. 1981. Experiments on competition and predation among shrimps of seagrass meadows. Ecology 62:1484-1493. CURRIN, B. M., J. P. REED, AND J. M. MILLER. In press. Growth, production, consumption and mortality of juvenile fishes; A comparison of tidal and non-tidal nursery areas. Estuaries. DAHLBERG, M. D. 1972. An ecological study of Georgia coastal fishes. Fish. Bull., U.S. 70:323-353. Darnell, R. M. 1958. Food habits of fishes and larger invertebrates of Lake Pontchartrain, Louisiana, an estuarine communi- ty. Publ. Inst. Mar. Sci., Univ. Tex. 5:353-416. 1961. Trophic spectrum of an estuarine community, based on studies of Lake Pontchartrain, Louisiana. Ecology 42:553-568. ENVIRONMENTAL PROTECTION AGENCY. 1979. Second Annual Workshop Chesapeake Bay Pro- gram. U.S. Environ. Prot. Agency, Hampton, Va., 579 p. GAINEY, L. F, JR., AND M. J. GREENBERG. 1977. Physiological basis of the species abundance- salinity relationship in molluscs: A speculation. Mar. Biol. (Berl.) 40:41-49. HACKNEY, C. T, W. E. BURBANCK, AND O. P HACKNEY. 1976. Biological and physical dynamics of a Georgia tidal creek. Chesapeake Sci. 17:271-280. HECK, K. L., JR., AND T. A. THOMAN. 1981. Experiments on predator-prey interactions in vegetated aquatic habitats. J. Exp. Mar. Biol. Ecol. 53:125-134. 1984. The nursery role of seagrass meadows in the upper and lower reaches of the Chesapeake Bay. Estuaries 7:70-92. HYSLOP, E. J. 1980. Stomach contents analysis — a review of methods and their application. J. Fish Biol. 17:411-429. LASCARA, V. J. 1981. Fish predator-prey interactions in areas of eelgrass (Zostera marina). M.S. Thesis, Coll. William and Mary, Williamsburg, Va., 81 p. 466 SMITH ET AL.: ORGANIZATION OF NEKTON LIVINGSTON, R. J. 1982. Trophic organization of fishes in a coastal seagrass system. Mar. Ecol. Prog. Ser. 7:1-12. MARKLE, D. F. 1976. The seasonality of availability and movements of fishes in the channel of the York River, Virginia. Chesapeake Sci. 17:50-55. MIKLAS, J., AND C. REED. In press. Comparisons of production dynamics in tidal marsh creek and open water shoreline populations of Fundulus heteroclitus. Trans. Am. Fish. Soc. Miller, J. M., and M. L. Dunn. 1980. Feeding strategies and patterns of movements in juvenile estuarine fishes. In V S. Kennedy (editor), Es- tuarine perspectives, p. 437-448. Acad. Press. N.Y. NELSON, W. G. 1979. Experimental studies of selective predation on am- phipods: consequences for amphipod distribution and abundance. J. Exp. Mar. Biol. Ecol. 38:225-245. O'NEIL, S. P. 1983. The distribution and trophic ecology of young-of- year spot (Leiostomus xanthurus, Lacepede) in polyhaline versus meso-oligohaline tidal creeks and adjacent shoals of the York River, Virginia. M.S. Thesis, Virginia Com- monwealth University, Richmond, 54 p. ORTH, R. J., AND K. L. HECK, JR. 1980. Structural components of eelgrass (Zostera marina ) meadows in the lower Chesapeake Bay — Fishes. Estuaries 3:278-288. PURVIS, C. 1976. Nursery area survey of northern Pamlico Sound and tributaries. Div. Mar. Fish. Rep. (prepared for U.S. Dep. Commer, NOAA, Natl. Mar. Fish Serv.), 62 p. QASIM. s. z. 1972. The dynamics of food and feeding habits of some marine fishes. Indian J. Fish. 19:11-28. SCHUBEL, J. R.. AND D. J. HlRSCHBERG. 1978. Estuarine graveyards, climatic change, and the im- portance of the estuarine environment. In M. L. Wiley (editor), Estuarine interactions, p. 285-303. Acad. Press, N.Y SHERIDAN, P. F 1979. Trophic resource utilization by three species of sci- aenid fishes in a northwest Florida estuary. Northeast Gulf Sci. 3:1-14. SHERIDAN, P. F, AND R. J. LIVINGSTON. 1979. Cyclic trophic relationships of fishes in an unpol- luted, river-dominated estuary in north Florida. In R. J. Livingston (editor), Ecological processes in coastal and marine systems, p. 143-161. Plenum Press, NY. STONER, A. W. 1980. Feeding ecology of Lagodon rhomboides (Pisces: Sparidae): Variation and functional responses. Fish. Bull., U.S. 78:337-352. WEGENER, W, D. HOLCOMB, AND V. WILLIAMS. 1973. Sampling shallow water fish populations using the Wegener ring. Proc. Annu. Conf. Southeast. Game Fish Comm. 27:663-673. WEINSTEIN, M. R 1979. Shallow marsh habitats as primary nurseries for fishes and shellfish, Cape Fear River, North Caroli- na. Fish. Bull., U.S. 77:339-357. 1981. Plankton productivity and the distribution of fishes on the southeastern U.S. continental shelf. Science (Wash.. D.C. 1214:351-352. 1983. Population dynamics of an estuarine-dependent fish, the spot {Leiostomus xanthurus), along a tidal creek — seagrass meadow coenocline. Can. J. Fish. Aquat. Sci. 40:1633-1638. WEINSTEIN, M. P, AND H. A. BROOKS. 1983. Comparative ecology of nekton residing in a tidal creek and adjacent seagrass meadow: community compo- sition and structure. Mar. Ecol. Prog. Ser. 12:15-27. WEINSTEIN, M. P, AND M. F WALTERS. 1981. Growth, survival and production in young-of-year populations of Leiostomus xanthurus Lacepede, residing in tidal creeks. Estuaries 4:185-197. WEINSTEIN, M. R, S. L. WEISS, AND M. F WALTERS. 1980. Multiple determinants of community structure in shallow marsh habitats, Cape Fear River estuary, North Carolina, USA. Mar. Biol. (Berl.) 58:227-243. WEISBERG, S. B., AND V. A. LOTRICH. 1980. Food limitation of the mummichog. Fundulus heteroclitus, in a Delaware salt marsh. [Abstr] Am. Zool. 20:880. 467 DISTRIBUTION, ABUNDANCE, AND GROWTH OF JUVENILE DUNGENESS CRABS, CANCER MAGISTER, IN GRAYS HARBOR ESTUARY, WASHINGTON1 Bradley G. Stevens2 and David A. Armstrong3 ABSTRACT Dungeness crabs, Cancer magister, were collected biweekly or monthly from May 1980 to July 1981 in Grays Harbor, Washington. Age of each crab was estimated from width-frequency analyses, and the population density and growth rate were monitored for each age class over the 14-month period. In April 1980 and 1981, crabs entered the estuary either as megalops larvae that metamorphosed to first instar postlarvae or directly as first instars. Intertidal mudflats with beds of eelgrass (Zostera spp.) were important habitats for the first few postlarval stages. Some crabs may have emigrated from the estuary during their second year of life, whereas others dispersed throughout the estuary and appeared to emigrate at sexual maturity (about 2 years). No gravid females were ever found in the bay. Population size was estimated to range from 3.3 million crabs (winter) to 39.0 million crabs (summer); 74% of the summer population were early instars. Growth of early instars was rapid and resulted in a 282-fold increase in dry weight from May to September, but little growth occurred during the remain- der of the year. Based on summer population abundance, it is estimated that this estuary could account for a substantial portion of recruitment to the offshore commercial fisheries. The biology of the Dungeness crab, Cancer magis- ter, has been studied by numerous investigators for several decades (Weymouth and MacKay 1936; MacKay 1942) because of its importance to com- mercial fisheries and its position as a benthic predator in estuaries and offshore communities (Gotshall 1977; California Department of Fish and Game 1981). Previous studies of C. magister biol- ogy have been conducted largely along the open coast (MacKay 1942; Cleaver 1949; Butler 1960, 1961; Gotshall 1978b, c). The few studies of crab populations in estuaries or shallow- water habitats (Butler 1956; Tegelberg and Arthur 1977; Gotshall 1978a; California Department of Fish and Game 1981) have indicated that such areas may be ex- tremely important nursery grounds, but the size and dynamics of estuarine populations have not been statistically determined and, furthermore, the contribution of estuarine habitats to offshore stocks has not been adequately assessed. Orcutt et al. (1978) estimated that 50-80% of crabs caught by the fishery in the Gulf of the Farallones spend some of their life cycle in the San Francisco-San 'Contribution No. 629, School of Fisheries, University of Washington, Seattle, Wash. 2School of Fisheries, WH-10, University of Washington, Seattle, Wash.; present address: National Marine Fisheries Service, P.O. Box 1638, Kodiak, AK 99615. 3School of Fisheries, WH-10, University of Washington, Seattle, WA 98195. Manuscript accepted January 1984. FISHERY BULLETIN: VOL. 82, NO. 3, 1984. Pablo Bay complex. Benefits derived from es- tuarine early life history may include enhanced growth rates, more abundant food, and refuge for postlarval and juvenile crabs from larger, older age classes that act as competitors and predators (Botsford and Wickham 1978). Quantitative studies of C. magister in major estuaries are timely and imperative. The demise of the San Francisco fishery prompted a 5-yr investi- gation of C. magister biology in that region (California Department of Fish and Game 1981), and hypotheses for the decline include alterations of estuarine habitat and water quality. In addition, channel dredging practices in west coast estuaries kill hundreds of thousands of crabs annually (Ste- vens 1981). Armstrong et al. (1982) estimated that a proposed channel modification project in Grays Harbor, Wash., could entrain and kill 2.5 million crabs over a 2-yr period. Knowledge of estuarine crab population dynamics and ecology of juveniles is required to gauge the relative importance of such habitat to the species, and to mitigate im- pacts of estuarine development on juvenile stages. MATERIALS AND METHODS Study site Grays Harbor is a shallow drowned river basin 469 t/#3 FISHERY BULLETIN: VOL. 82, NO. 3 estuary with an extensive littoral zone and a wide variety of substrate types (Fig. 1). Sampling sta- tions were established primarily along the exist- ing navigation channel in accord with a concur- rent study to assess the impact of dredging on Dungeness crabs (Armstrong et al. 1982). Habitats represented include deeper sandy channels (sta- tions 1-3), shallow sand (station 4), sand-mud (sta- tions 6 and 7), mud (stations 8-12), and those adja- cent to eelgrass (Zostera marina and Z. noltii) beds (station 5) (Table 1; Fig. 1). Fifteen sublittoral strata were established for the purpose of estimat- ing population abundance, and sampling stations were located approximately at the center of each (Fig. 1). However, strata 14 and 15 contained no regularly sampled stations because these areas were outside the primary focus of our contract. The boundaries of each stratum were defined by the midpoint between sampling stations, or the bot- tom contours at -5.5 m or 0.0 m (for detailed 50° N 4 0° N 3(T N — FIGURE 1. — A. West coast of North America. Arrow indicates site of Grays Harbor, Wash.; B. Map of Grays Harbor, showing sites of Cancer magister collection (1-13). Stations 14 and 15 represent unsampled strata. Lines separating strata were defined arbitrarily for use in determining crab population size. TABLE 1. — Location and description of sampling sites for Cancer magister in Grays Harbor, Wash. Station Area Gear Latitude Longitude Depth no. (ha) type' N W (m) Bottom type Comments 1 837 T 46°54'30" 125° 9' 0" 15-18 Sand, cobble Not sampled in 1981 due to rough water conditions. 2 496 T 46°55'15" 124° 7 '20" 13-15 Hard sand Dredged sediment disposal site. 3 1.507 T 46°55'15" 124° 4'20" 10-15 Sand Outer harbor channel bottom. 4 1,120 T 46°56'40" 124° V10" 3-5 Sand, mud Shallow outer harbor habitat. 5 658 T 46°51'55" 124° 4'15" 5-8 Sand, mud Adjacent to extensive eelgrass beds. 6 680 T 46°55'40" 123°59' 5" 6-8 Sand, mud, leafy debris Near mudflats without eelgrass. 7 656 T 46°57'30" 123°59' 0" 11-14 Mud, sand Adjacent to eelgrass beds 8 418 T 46°58'10" 123°55'20" 11-15 Mud Inner harbor channel bottom. 9 221 T 46°57'40" 123°50'50" 12 Mud Numerous snags; adjacent to shipping terminals, inner harbor. 10 96 RN 46°56'20" 123°54'15" 3-4 Mud, snags 11 86 RN 46°57'12" 123°51'15" 3-4 Mud, snags 12 — RN 46°57'37" 123°46'18" 3-4 Mud, cobble Along shore of deep (20 m) river channel. 13 T 46°54'45" 124° 5'15" 0-3 Sand Intertidal sand flat; no eelgrass. 14 2,653 Mud, sand, eelgrass Not sampled. Used for population estimate 15 414 Sand Not sampled. Used for population estimate 'T = otter trawl; RN = ring net. 470 STEVENS AND ARMSTRONG: CANCER MAGISTER IN GRAYS HARBOR descriptions see Stevens 1982). The area of each stratum was determined by planimetry at the level of mean lower low water (NOAA Chart No. 18502 Grays Harbor, 1979 edition). Sampling Design Crabs were sampled at stations 1-9 and 13 with a 4.9 m, 4-seam, semiballoon otter trawl net, having 38 mm stretch nylon mesh throughout and a 6 mm cod end liner. Working width of the net was about 3.0 m. Distance towed was measured between buoys placed at the beginning and end points of each trawl, by compass triangulation to station- ary objects whose positions were predetermined and located on 7.5-min topographic maps (U.S. Geological Survey). Distances were then con- verted to area swept and catches expressed as crabs/ha. At stations 10-12, underwater snags pre- vented trawl operation so crabs were collected by setting collapsible ring nets (76 cm diameter) cov- ered with 12 mm mesh. A "set" consisted of 4 baited nets set 50 m apart and fished for 20 min. Catches were expressed as crabs/net. Trawls and ring net sets were made within 1-2 h of slack low tide in daylight. Occasional plankton tows were made with a 0.5 m diameter conical net of 500 /xm mesh, in the spring of 1980 and 1981, to determine if crabs entered the bay as larvae. Stations 3, 6, 8, and 9 were sampled biweekly from May through October 1980 and at intervals of 4-5 wk thereafter through July 1981. Other sta- tions were generally sampled monthly except when weather or boat problems precluded opera- tions. Most stations were sampled on 13-19 occa- sions during the 14-mo field study (May 1980 to late June 1981), with the exception of stations 1 and 2 (6 and 10 samples, respectively). Station 13 was sampled quarterly on a diel basis, and com- plete results from that diel study are reported elsewhere (Stevens and Armstrong 1984). No sam- ples were taken at stations 14 and 15 which were used only to calculate crab populations based on data from adjacent areas (see below). All crabs were measured to the nearest milli- meter across the carapace between the notches just anterior to the 10th anterolateral spines ("carapace width" or cw), sexed, and released. Sub- samples were used for width frequencies only in May and June of 1980 and 1981, when early instars were collected in large quantities. Surface and bottom-water samples were collected during each trawl with a modified Van Dorn bottle; tempera- ture was measured to 0.1°C, and salinity deter- mined with a refractometer at room temperature. Growth Analysis Cumulative width frequencies of all crabs caught during a given week were plotted on proba- bility paper, and width limits were subsequently defined as the curve inflection points (arbitrarily nonoverlapping) to delimit the size range of each year class through time, according to the method of Cassie (1954). These were compared with fre- quency graphs for verification. Values were inter- polated during weeks in which too few crabs were caught for accurate analysis. Each crab was then assigned to an age-group on the basis of the width limits for each sampling week. Age was defined as the number of years since metamorphosis. Mean widths were calculated for each age group (0+, 1+, 2+, and 3 + ) and plotted by sampling week. Eighty-seven males (12-132 mm cw) and 74 female crabs (15-115 mm cw) were frozen and returned to the University of Washington where they were opened at the epimeral line and dried to constant weight at 60°C (48-72 h). Only hardshell intermolt crabs were used. Log10 dry weight (g) was plotted against log10 carapace width (mm) and regression equations determined for each sex. Mean weights for each age group of crabs were calculated at monthly intervals from mean widths using the regression equation (the 1977 year class was omit- ted because the regression equation did not repre- sent these larger animals). Weight-specific growth rates {k) per month were calculated by use of the equation W, = W0 ekt. The monthly percent weight increase was calcu- lated as ek - 1. Crab Density Analysis Because counts of benthic invertebrates usually show a contagious distribution (Elliott 1977), all density data were transformed prior to analysis of variance or regression by Xt = Log10 (density + 1), where Xt is the transformed variable. Density was plotted against bottom-water salin- ity, temperature, and estimated Chehalis River flow by a stepwise multivariate procedure (SPSS REGRESSION) for all trawl and ring net samples. 471 FISHERY BULLETIN: VOL. 82, NO. 3 The effects of season and location on crab density were examined by analysis of variance (SPSS ANOVA procedure). The sampling year was di- vided into two seasons: spring-summer (March- August) and fall-winter (September-February). The navigation channel was divided into two areas: the outer estuary (stations 2, 3, and 4: sta- tion 1 deleted due to lack of winter data points), and the inner estuary (stations 7, 8, and 9). A two-way ANOVA was performed with these two seasons and two station groups as the independent variables, and crabs/ha as the dependent variable. Population Estimation Two basic assumptions were made concerning the trawl data: 1) Sampling efficiency of the net was not 100% and varied for each age class of crabs. Efficiency was estimated to be 0.33 for the 0+ age class during summer, and 0.25 in winter, based on comparisons between net catches and visual counts of young instars on mudflats at low tide (see Discussion). Sampling efficiency was es- timated to be 0.50 for all other age groups in accor- dance with Gotshall (1978a). 2) Sampling effi- ciency was assumed to remain constant and not to vary as a function of changes in crab behavior (e.g., burial or diel activity variations). Data on crab densities were used from a 12-mo period, June 1980 to May 1981, which was divided into three "seasons": summer (June-August 1980), fall-winter (September 1980-February 1981), and spring (March-May 1981). Population estimates were made for three age groups (0+, 1 +, and 2+, the latter including all 3+ animals which were identifiable only in summer 1980) in each of the three defined seasons. A stratified random technique was used, using the following variables (see Cochran 1953): h <*ih nh Xih s2(xh) = stratum of harbor = catch of crabs in tow /', stratum h = area (ha) covered by tow /, stratum h = number of tows in stratum h = individual estimates of crabs ha-1, from tow i, stratum h = mean catch (crabs ha-1) in stratum h for a given season = area of harbor in stratum h (ha) = variance of mean x^ in stratum h = total number of crabs in harbor. that would have led to complications in the deter- mination of confidence intervals, but only minor changes in the resultant mean densities of crabs. Mean crab density in each stratum was calculated for each age group and season by Xh I t=i (clh/aih) nh The total number of crabs in each stratum was calculated as T^ = A^ (xh), and the total for the harbor by the sum of all stratum totals, 15 T = !Th . h=l For strata 1-9, the variance of each stratum total was calculated by V(Th) = A%s2(xh) n Data used for population estimates were not trans- formed as done for ANOVA comparisons because and the variance of the total was calculated by summing the individual variances 9 V(T) = 2 V(Th) ■ h=l Confidence intervals for T were approximated at the 95% level by T± *(df, 0.05) ^(T) • Crab abundances in strata which were not sam- pled by trawl (10, 11, 14, and 15) were calculated using mean density values from nearby strata of similar ecological characteristics. Data were used from strata 6 (for 10 and 11), 5 (for 14), and 3 (for 15). Totals by age group and season for those strata were added to totals for strata 1-9, to obtain totals for the entire estuary. The estuary totals were divided by the estimated trawl efficiency factors to obtain final corrected estimates of crab abundance by age group and season. Confidence intervals for these final estimates could not be computed. RESULTS Temperature-Salinity Profile Grays Harbor has a strong horizontal salinity gradient (Fig. 2). Temperature and salinity were 472 STEVENS AND ARMSTRONG: CANCER MAGISTER IN GRAYS HARBOR 4000 3500 300 0 ■ 2500 - 2000 » • J o o o ■ 1000 - ) ■ o T • 0 800 u • z 1 1 o o o o -2 600 - 0 1 1 1 1 o 0 Z 400 - 1 1 o c 0 ■ * 200 - 1 1 i II t o r30 25 20. E 15 O o m 10 o « s 1 2 3 4 5 6 7 8 9 10 11 12 Sire FIGURE 2. — Temperature-salinity profile for crab sampling sites 1-12, Grays Harbor, Wash. Filled circles indicate mean density +1 SE of crabs for the entire study, as determined by trawl; open circles indicate mean bottom salinity at low tide. Crab densities are not plotted for stations 10-12 where ring nets were used. more stable in the outer estuary, but less so as distance increased eastward from the harbor mouth. At station 3, bottom temperatures ranged from 7CC (winter) to 14°C (summer), while at sta- tion 9 they ranged from 5° to 18°C. Vertical stratifi- cation was greater toward the head of the estuary and less so in the outer estuary as a result of turbulent mixing. Greatest vertical salinity dif- ference measured during the study was 17%» at station 9. Grays Harbor receives 70-100 in of rain- fall annually, and stratification was greatest dur- ing November-March, the period of peak rainfall. Flow rates of the Chehalis River, which contrib- utes 80% of the freshwater inflow, varied from 22.3 mHr1 in August 1980 to 2,322 mHr1 in Febru- ary 1981 (data provided by U.S. Geological Sur- vey). Spatial Distribution of Crab Population Complete records (width, sex, age) were ob- tained for 14,556 crabs. Coefficients of variation averaged 0.53 for the trawls and 0.46 for the ring nets, implying that both techniques had a similar degree of precision. Mean density of crabs during the 14-mo sam- pling period was greatest at station 1 (2,190 crabs ha-1), and catches declined with increasing dis- tance from the estuary mouth and decreasing bot- tom salinity (Figs. 2, 3A). Notable exceptions to this pattern were low densities at stations 6 and 2 (120 and 290 crabs ha"1, respectively; Fig. 3). Sta- tion 2 was concurrently being used as a dredged sediment disposal site by the U.S. Army Corps of Engineers. Crabs caught by ring net were more abundant at station 11, near the eastern (upstream) end of the estuary than at station 10, and averaged 22.9 and 12.7 crabs net- \ respectively, from June to October (Fig. 3F). No crabs were caught at station 12 except in August and September 1980. Temporal Distribution of Crab Population Megalops larvae were found as early as 1 April 1980 at station 6, and in densities up to 810/1,000 m3 at station 5 on 22 April 1980. Crab densities at all stations were greatest from May to August 1980 (Fig. 3A-E) and declined from September 1980 through January 1981. Lowest densities occurred in October and November 1980, none being >200 crabs ha-1 except at station 1. Although monthly variation was great at each station, this general decline in crab density dur- ing fall-winter occurred throughout the estu- ary. Crab abundance at the three ring-net stations (10, 11, and 12) increased dramatically from June through October 1980, then dropped in November 1980 to a low of <1.0 crabs net- 1 at all three sta- tions (Fig. 3F). No crabs were caught at station 12 except during August and September 1980, when the salinity reached 9 and TL, respectively. Salin- ity at station 12 was 1.0'/, or less during all other sampling periods. The F- tests showed that mean crab density in the outer estuary (stations 2, 3, and 4) was sig- nificantly greater (P = 0.011) than in the inner estuary (stations 7, 8, and 9; Table 2). Crab density at all six stations (2, 3, 4, 7, 8, and 9) was signifi- cantly greater (P = 0.001) in spring-summer than in fall-winter 1980-81. Regression analysis of log10 -transformed density data on bottom salinity, temperature, and Chehalis River flow rate showed no significant dependence of trawl catches on these variables, but salinity alone was responsible for about 40% of the variance in crab abundance at the ring net stations (10, 11, and 12; r2 = 0.398, P = 0.001). 473 FISHERY BULLETIN: VOL. 82, NO. 3 — i — i — i — i — i — i — i — i — i — i — i — i — i — i — ^, MJJASON Dlj F MAMJJ O „ 1980 1981 O 20 1980 1981 M J J A S 0 '■QOi>f iQ, , i N D I J F MAMJ 1 — I — I — I — I" 1980 1981 MJJASON DU FMAMJJ 1980 1981 n r M J J A S r i-«t i w — i — i i i i ON D ' J F MAMJJ 1980 1981 .o SITE 11 O' : F C 9 l\. u 0) a D CO • SITE 10 L. •'SITE 12 t V o U .1 . a «; r> u n [ %;*S-i iVi O -C E 3 o r 1 m m MAY 1980 N =373 fiN=74.44 90 _m L^ft 16 MRY 1980 490 MERN=69-52 T>u-n-L MERN=69.52 r Tl ru. m MEF)N=48. Ti fil 11 HTrnmrmTnTrmmrn ntfe-Bi 4 JUNE 1980 = 290 MERN=48.46 44 iThhr-rTI 1 1 Ml iTTTfTTTTTTfTTl 16 JUNE 1980 N =295 in HERN=60-23 fflmTflrftfrifn3 ^.TTMTjfflffl I [fllHTTmrfU^ rhsn°-g: 30 JULY 1980 N =146 MEflN=61.89 28-41 Hi] 14 RUG 1980 N =613 rrrrhm: MERN=70-12 5.D 12 SEPT 1980 rfTTTMTIIMlTi^HTTi =219 MERN=61 .53 S-0.=21.88 'rrff mil "Mm taM 1 JULV 1980 N =456 „J1EflN=56-85 mm-' 1 rfli-T ftHf n , 15 JULY 1980 "T-Tr-i N =875 " -n rfl-rh-n MEflN=59.46 1 unnnrtTh^-rft,^ 21.39 29 RUG 1980 N =176 n _r-Tl rh-ITU MERN=71 .53 ^mM'Tki^WinrM;27-19 nriTn rriHilTllTTm 13 OCT 1980 N =45 HERN=68-95 S.D.=16.67 27 OCT 1980 N =8 HEAN=40.25 S.D.=30.I8 JU n n M >rfTh 12 NOV 1980 N =142 MERN=77.71 S.D.=16.21 r^ -H- n^ril rtfHTTTTTMTh 15 0EC 1980 N =133 HERN=75.1 S-D. =18-94 rfH>in n rf 9 FE8 1981 -n-n N =361 Tt-| NEAN-88.81 lrnnftfTrrriTi rTfTfTTThfTTTTfTTTT 11 MflR 1981 JTh N =404 -n MERN=75.79 n nmfT T It 1 21 APRIL 1981 l__ N =612 >L_ MERN-74.16 fcrfrfrffl^'m DB 12 MAY 1981 N 1 Qfl MN=68-19 2 JULY 1981 M N = 351 Wmm, MEflN=58-37 S-D-=25-40 nil n~n I 1 1 1 1 1 1 1 1 1 I 1 0 15 30 45 60 75 90 105 120 135 150 165 — i ■ i i i 1 i i i i i 15 30 45 60 75 90 105 120 135 150 165 Carapace Width in 3 mm Intervals FIGURE 4. — Carapace width frequencies of all crabs caught by trawl at stations 1-9 in Grays Harbor, Wash., 1980-81. Numbers expressed as log10 (catch + 1). Box for each sampling period shows date, number of crabs measured, mean carapace width overall (mm), and standard deviation. 475 TABLE 3. — Upper limit of carapace width range (mm) of Cancer magister in Grays Harbor, Wash., for each age/sex group. Selection method was cumulative probability (P) or interpolation (I). The upper limit for age 2 + crabs (lower limit of 3 + ) was not distinguishable in fall-winter due to low numbers caught, nd = not distinguishable. Males Females Date Method Age groups 0+ 1+ 2 + Age groups Method 0+ 1+ 2 + 5/4/80 5/16/80 6/4/80 6/16/80 6/21/80 7/1/80 7/15/80 7/30/80 8/14/80 8/29/80 9/12/80 9/26/80 10/13/80 10/27/80 11/12/80 12/15/80 1/17/81 2/9/81 3/11/81 4/4/81 4/21/81 5/21/81 7/1/81 P I P I P I P I P I I P P I P P P I P P I P P 25 26 27 29 30 32 34 37 40 43 45 46 46 46 46 46 44 44 45 47 15 26 29 60 65 70 70 70 75 85 88 92 92 94 96 105 106 107 101 121 121 121 120 55 70 75 115 120 124 132 136 136 136 134 132 nd nd nd nd nd nd nd nd nd nd nd nd 127 nd P I P I P I P I P I I P P I P P P P P P I P P 30 28 25 27 28 31 37 36 36 41 45 50 50 52 54 52 55 54 61 56 15 29 29 60 69 77 79 80 87 90 91 92 93 94 95 96 98 100 104 125 126 126 133 63 75 86 120 124 127 126 126 127 130 nd nd nd nd nd nd nd nd nd nd nd nd nd 120 120 nd FISHERY BULLETIN: VOL. 82, NO. 3 abundant at all stations except 3, averaging 268 crabs ha"1 and 54.7% of all crabs over the entire sampling period. Greatest densities occurred at station 1, but these crabs were also abundant at stations 3, 4, 5, and 7 (Fig. 5), i.e., the outer es- tuary. This group was least abundant at stations 6, 8, and 9, but comprised the largest proportion (73-78%) at the ring net stations (10-12). The average density of the 2+ age group was 121 crabs ha" * (stations 1-9) equal to 21.3% of all crabs caught. Greatest densities occurred at stations 1 and 3 (Fig. 5). This group was the most abundant at station 3, the only area where the 1 + age group did not predominate. The 3+ age group was difficult to separate from the 2+ group because the former were caught in low numbers. Of all samples taken during the study, they represented 3% with an average den- sity of 17 crabs ha" l. This group occurred primarily at stations 1-3, with greatest densities at station 1 (Fig. 5). SITE 1 1 7£n 15.4 0 12 3 4 CRABS HECTARE"1 x 100 n 12 p 2 4 6 8 CRABS PER RING NET 10 FIGURE 5.— Mean number of crabs per hectare (sites 1-9) and crabs per ring net (sites 10-12) in Grays Harbor, Wash., over the entire study period, by age group. Age Class 0+ 1 + 9.5 I n i i i r-r-r 2+ ESS^J ^kss^^^S 3 + A 4 2 0 4 2 0 4 2 0 f 2 0 r10 8 6 4 2 0 O O o a> a O 123456789 Trawl Site FIGURE 6. — Actual density of crabs caught by trawl at sites 1-9 in Grays Harbor, Wash., June 1980, by age class. Sites 2, 5, and 7 sampled on 4 June; sites 1, 4, and 9 sampled on 16 June; sites 3, 6, and 8 sampled on both dates and averaged. Note greatest abun- dance of 0+ age group at site 5. 476 STEVENS AND ARMSTRONG: CANCER MAGISTER IN GRAYS HARBOR Growth:Width Crabs in the 0+ age group (1980 year class) increased in width by a factor of 4, from 10 mm in May to 40 mm in October 1980, but growth slowed from then to the following April 1981 when they were about 50 mm wide (Fig. 7). The same pattern of rapid growth during spring and summer was evident among the 1+ and 2 + age groups (1979 and 1978 year classes, respectively), although measurable increases in carapace width were re- corded into winter 1981 for these older crabs. Crabs of the 1979 year class grew from 45 to 73 mm between May and October 1980 (factor of 1.6), while the 1978 year-class crabs grew from 84 to 118 mm during the same period, an increase of 1.4. Females had slightly greater mean widths than males up to about 125 mm cw, but the differences were minor (Table 3). Growth: Weight Regression equations for log10 dry weight (g) on log10 carapace width (mm) were derived sepa- rately for male and female crabs, but were not significantly different. Therefore, a pooled regres- sion equation was calculated for both sexes combined: LL < a. < HO 120 100 80 60 40- 20 1979 1980 1981 J J A S 1980 N D J F M A M 1981 J J FIGURE 7. — Mean width of four age groups of crabs (0+, 1 + , 2+, 3+ ) in Grays Harbor, Wash., 1980-81. Mean width ( + 1 SE) shown for each sampling period was determined by graphical analysis of width-frequency data or interpolation (see Table 3). or Log10 Weight (g) = -4.064 + 2.832 (Log10 Width, mm) Weight (g) = (8.63 x 10" 5) Width2 832 (r2 = 0.985, P = 0.0001; Fig. 8). Differences in width/weight and width/age rela- tionships between male and female crabs would probably increase at sexual maturity, which oc- curs about 2 yr after metamorphosis, and at widths of 93-122 mm for males and 100-105 mm for females (Butler 1960, 1961; Poole 1967). Growth data presented herein are probably valid only for male crabs <132 mm and female crabs <115 mm width. Changes in mean weight with time (Fig. 9) probably represent a continuous curve, but there appeared to be an inflection point in late August that separated spring-summer and fall-winter growth stanzas. Therefore, k values were calcu- lated for the periods May-August and September- April . Monthly weight-specific growth rates were greater in spring-summer than in fall-winter for all age groups but decreased with size (Fig. 9). Specific growth rates were greatest for 0+ age group crabs in their first summer during which the average monthly weight increase was 206% (Table 4). Growth decreased to an average 15.8% per month during the winter. Growth rates increased again for age 1+ crabs in their second summer (31% per month), but were lower than experienced in their first year. This pattern was found for all age groups. Crabs in the 2 + age group (probably at sexual maturity) increased in weight 25% per month in the summer of 1980, but only 6.5% per month during the following winter (Table 4). First TABLE 4. — Weight-specific growth rates (k) and percent weight increase of three year classes of Cancer magister in Grays Harbor, Wash. Weight calculated from mean carapace widths of each year class by regression equation (see Figure 8). Growth per month calculated for spring-summer (May-August) and fall-winter (September- April) growth stanzas (see also Figure 9). Dry weight (g) Mean growth per month Spring-Summer % weight k increase Fall-Winter Year class 4 May 1980 29 Aug. 1980 22 Apr. 1981 % weight k increase 1980 1979 1978 0.02 4.02 24 30 1.75 11.75 59.70 5.56 21.46 98.71 1.118 206 0268 31 0.225 25 0147 15.8 0.075 7.8 0.063 6.5 477 FISHERY BULLETIN: VOL. 82, NO. 3 CO CL t — i LU >- CD o » — I o 2.0 1.5 ■■ 1.0 ■■ 0.5 ■■ 0.0 ■■ 0.5 1.0 1.0 1.2 1.4 1.6 1.8 2.0 L0G10 (CRRRPRCE WIDTH IN MM) 2.2 FIGURE 8. — Regression of loglo dry weight (g) on log,0 carapace width (mm) for 87 male (12-132 mm) and 74 female (15-115 mm) Cancer magister from Grays Harbor, Wash. Observations ( + ), regression line (solid), and 95% confidence interval about the regression line (dashed) are shown. instar crabs (7 mm cw) of 0.02 g dry weight in- creased in dry biomass 282 times by the time they reached the 6th or 7th instar (about 50 mm cw), weighing 5.7 g the following April (Fig. 9). Some may have reached 70 mm by that time, weighing 14.7 g, an increase of over 700-fold. Second-year crabs increased in dry biomass 5.3 times, from 4.0 to 21.5 g. Third-year crabs increased from 24 to 99 g, a dry biomass increase of 4.1 times. DISCUSSION Recruitment and Distribution in the Estuary Megalops larvae probably metamorphosed to the first postlarval stage in Grays Harbor, since trawl collections included second instars on 4 May 1980 and first instars in April and May 1981. Cast exuviae of these stages were abundant on beaches of the outer estuary in early May 1982. Larval densities in the estuary were at the low end of the range of densities found by Lough (1976) off the Oregon coast in 1970-71 (100-8,000/1,000 m3). In contrast, no megalops larvae were found in San Francisco-San Pablo Bays during 4 yr of surveys by the California Department of Fish and Game, which concluded that crabs entered that estuarine system only after metamorphosis (Orcutt et al. 1975, 1976). Once inside Grays Harbor, C. magister showed an ontogenetic change in habitat selection, i.e., centers of abundance changed with age. Eelgrass beds may be the preferred habitat of the first post- larval stages, because catches of 0+ crabs were most abundant near those areas (Figs. 5, 6). Butler (1956) also found that the most abundant concen- trations of early instars along the northern shore of Graham Island, Canada, were associated with the presence oiZostera marina in sheltered inlets. However, this age class was widely distributed from stations 1 to 9. Crabs in the 1+ age group (size range 50-90 mm cw) were the most abundant. Although their dis- 478 STEVENS AND ARMSTRONG: CANCER MAGISTER IN GRAYS HARBOR M J SOND'JFMA 1980 1981 M J FIGURE 9.— Growth rates of Cancer magister from Grays Har- bor, Wash. Points represent log10 mean dry weight calculated from mean width by regression equation. Straight lines (fitted visually) show slope of growth curve (k) for summer and winter growth stanzas of 1980 (0+), 1979 (1 + ), and 1978 (2 + ) year classes. Growth stanzas arbitrarily separated at 28 August 1980. tribution greatly overlapped that of the 0+ age class, age 1+ crabs showed proportionally less use of low-salinity stations such as 6, 8, and 9. Al- though this group was also abundant near eel- grass beds, they were restricted to the subtidal channels and showed only intermittent use of mudflat areas during high tides, whereas many crabs of the 0+ group remained in the littoral zone at low tide. The 2+ age group, consisting of sexually mature crabs (Poole 1967), was abundant only at the outer estuary stations (1, 3, 4, 6, and 7). Many crabs probably migrate out of the harbor before reaching age 3 + . This hypothesis is supported by the scar- city of age 3+ crabs east of station 3 and the total absence of gravid females from trawls taken in the estuary, although many trawls were made during the spawning season (October-March). Appar- ently, most mature females leave the harbor to spawn. Stressful salinity and temperature, high larval predation in the estuary, or inadequate lar- val food supplies might have created selection pressures for spawning females to seek water offshore with the proper environmental conditions for higher egg and larval survival. Crab Population Estimates A more important question than that of local crab densities at several stations is that of total population abundance throughout the estuary, de- termined for different seasons and age classes. Such a calculation is of interest in order to gauge 1) the potential use of the estuary by the species and the 0+ age class in particular; 2) the theoretical contribution made by the estuarine population to the commercial fishery; and 3) the potential impacts of estuarine development (e.g., dredging and landfill) on resident populations. The former two points are addressed in this discussion. Trawl Efficiency Estimates In order to extrapolate crab density values at each station to abundance within the correspond- ing stratum, some measure of trawl efficiency was needed. Examinations of mudflat areas (stations 5 and 14) in May of 1981 and 1982 showed very high densities of 0+ age crabs, ranging from 1 to 5 crabs m-2. This estimate was conservatively reduced to 1 crab m- 2, and we assumed that only 50% of the available estuarine bottom (Fig. IB) was utilized by early instars (excluding the inner estuary and perimeter which had lower salinities and had pro- duced few or no crabs of this age group). This "corrected" density of 0.5 crabs m~2 (5,000 crabs ha-1) was about 30 times the mean summer den- sity of age 0+ crabs at station 5 (162 crabs ha- x) as estimated by trawl; thus trawl efficiency in that season was about 0.033 (more recent studies have shown early instar densities to equal or exceed 10 rrr2 in 1983; D. Armstrong, unpubl. data). In winter and spring, 0+ age group crabs were large enough to be sampled more effectively, but proba- bly not as effectively as larger crabs; thus a factor of 0.25 was used in both of those seasons. An effi- ciency factor of 0.5 was applied to all other age groups, in accordance with Gotshall (1978a). Abundance Calculation For the nine strata of Grays Harbor sampled by 479 FISHERY BULLETIN: VOL. 82, NO. 3 trawl, the total number of trawl-catchable crabs present in 1980-81, ±95% confidence intervals, were summer, 4.3 ± 1.7 million crabs; winter, 1.3 ± 0.7 million crabs; and spring, 2.6 ± 1.2 million crabs (Table 5). However, stratum 1 (837 ha) was excluded from the spring estimate due to lack of data, but was reincluded later, using density esti- mates from adjacent stratum 3. Totals for the other nonsampled strata (10, 11, 14, and 15) were added to totals for the trawl-sampled strata, and the sums for each age group were divided by the trawl efficiency estimates described above. Final calculated numbers for the total crab population were summer, 39.0 million crabs; fall-winter, 3.3 million crabs; spring, 7.8 million crabs (Table 5). The 1980 year class, which was extremely abun- dant in the summer of 1980, virtually disappeared during the following winter, and reappeared in spring of 1981. Some hypotheses for this decline and recovery include winter hibernation, migra- tion to nonsampled areas (e.g., stratum 14), and temporary egress from the estuary. Natural mor- tality probably contributed substantially to the decline as well. Due to the speculative nature of these estimates and the underlying assumptions concerning trawl efficiency and proportion of habitat utilized, it was not possible to compute confidence limits on these final estimates. The estimates of total population abundance in the estuary suggest a tremendous increase in summer with the influx of 0+ crabs as megalopae and first instars, and an increase in 1 + animals as well. This estimate of 39 million crabs is the highest estuarine crab population abun- dance yet reported. The only other reported esti- mate, that of 9.3 million crabs in the San Francisco-San Pablo estuary complex during 1975 (Orcutt 1978), is based on a much less systematic survey than ours and does not correct for poor gear efficiency in regards to the small size of early in- stars. Furthermore, this latter estuarine system represents an area (500 km2) five times that of Grays Harbor. The accuracy of our estimates of population abundance can be qualitatively assessed by com- parison of trawl density data with that of other studies (Gotshall 1978a; Orcutt et al. 1975, 1976; Orcutt 1977, 1978; Table 6). Generally, there is great seasonal variation, but densities estimated in Grays Harbor are in accord with values for Humboldt Bay and San Francisco Bay. Extrapola- tions to total abundance indicate that large popu- lations of juvenile crabs may use other coastal estuaries as well. Even relatively small estuaries in Oregon, such as Tillamook, Netarts, Yaquina, and Coos Bay, could support large populations of 0+ crabs, considering their small biomass (0.2 g dry weight). The principal benefits of these es- tuaries are probably refuge from larger can- nibalistic conspecifics (Botsford and Wickham 1978; Stevens et al. 1982), more abundant food, and possibly accelerated growth as a result of food supplies and warmer temperatures than offshore waters. TABLE 5. — Estimation of Cancer magister population in Grays Harbor, Wash., for 1980-81. All values are numbers of crabs except efficiency factors and percentages. C.I. = confidence interval. Strata sampled by trawl (sites 1-9) Strata Sum Variance not of Effi- Total % Season/ n of n C.I.2 sampled3 crabs ciency crabs of Age Class' (x103) (x109) df (x103) (nxio3) (x103) factor (x103) total Summer 0 + 485 9 36 188 470 955 0033 28,942 74.2 1 + 2,979 555 36 1,511 982 3,961 0.5 7,922 20.3 2 + 851 19 36 279 228 1,079 0.5 2,160 5.5 Total 4,315 687 36 1,681 1.680 39,024 100.0 Winter 0 + 182 14 28 244 6 188 0.25 753 23.0 1 + 1,070 82 28 588 87 1,157 05 2,311 70.7 2 + 97 10 28 65 7 104 0.5 207 6.3 Total 1,349 123 28 720 100 3,271 100.0 Spring 0 + 146 8 13 189 87 233 0.25 931 11.9 1 + 1,176 181 13 918 307 1,483 0.5 2,965 38.0 2 + 1,246 342 13 1,262 707 1,953 05 3.904 50.1 Total 2,568 290 13 1,163 1,101 7,800 100.0 'See Table 3 and Figure 7 for size of these age classes throughout the year 1980-81 2Values for t(0.05)- summer = 2.029, winter = 2.048, spring = 2.160. 3See text for explanation of estimates based on data of adjacent trawl stations 480 STEVENS AND ARMSTRONG: CANCER MAGISTER IN GRAYS HARBOR TABLE 6. — Comparison of Cancer magister densities in Grays Harbor, Wash, (this report); Humboldt Bay, Calif. (Gotshall 1978a); and San Francisco-San Pablo Bay, Calif. (Orcutt et al. 1975, 1976; Orcutt 1977). Data are not corrected for gear efficiency. Season Year Transect No. Bay Method Area (m2) crabs/ha San Francisco- Summer 1975-77 Trawl '1,500 90-340 San Pablo, Calif September 1977-78 Trawl 1.500 13-170 Humboldt Bay. Calif. January 1967 Trawl 2,400 4,910 August 1967 Trawl 2,400 300 April 1968 Trawl 2,400 140 August 1968 Trawl 2,400 1,280 October 1968 Trawl 2.400 930 (Mean of trawl samples, 1967-68 = 890) August 1967 Scuba 140 520 April 1968 Scuba 140 0 August 1968 Scuba 140 4.480 October 1968 Scuba 140 280 (Mean of scuba samples. 1967-68 = 1,080) Pacific Ocean, near October 1968 Trawl 26,667 0-9,400 Humboldt Bay, Calif. November 1968 Trawl 6,667 0-36,000 (x = 800) Grays Harbor, Wash. June 1980 Trawl variable 200-1,000 Outer Harbor December 1980 Trawl 1,400 310 May 1981 Trawl 2,000 1.320 'Distance estimated as 50 m/min. 2Area estimated as distance (given) x % (headrope length). Growth Dry weight increased 282 times between first instar (0.2 g) and sixth instar (5.7 g) during the first year. Other authors have not presented growth data as changes in weight, but rather as increases in carapace width. Crabs in Grays Har- bor grew from about 7 mm to 50+ mm cw during 1980-81, which is similar to values reported by Cleaver (1949) and Butler (1961). However, Poole (1967) concluded that crabs in Bodega Bay, Calif., reached 75 mm (range 50-100 mm) by 1 yr after metamorphosis. This would represent fairly rapid growth, but close to the upper limits of crab growth rates in Grays Harbor. In contrast to Grays Harbor, Tasto (1983) stated that juvenile crabs spend only 1 yr in San Pablo Bay, and reach 100 mm by the end of that time (twice the growth rate of ocean crabs and Grays Harbor crabs). He concluded that the estuarine population was a single year class and was almost completely replaced by a new year class each spring, a situation very different from Grays Har- bor where at least three year classes are present constantly The San Francisco data may have been misinterpreted, perhaps caused by use of collect- ing gear (mostly ring nets) that selected larger crabs and resulted in a frequency mode near 100 mm cw that may have actually represented older 1+ age group crabs. Unfortunately, growth data are not available for 0+ age Dungeness crabs that metamorphose di- rectly offshore for comparison with estuarine populations. Presumably, colder bottom-water temperatures offshore (8°-10°C) would cause metabolic, growth, and general energetic depres- sion of these animals relative to rates in warmer (14°-18°C) estuaries. Studies of offshore juvenile populations are much needed in this regard. Importance of Grays Harbor to Commercial Fisheries. The potential contribution of Grays Harbor to the commercial landings of Cancer magister was calculated by assigning various mortality rates to the 1980 year class for a period of 3.5 yr, i.e., until recruitment to the fishery. Jow (1965) estimated annual natural mortality of adult male crabs to be 15% per year (M = 0.165, exponential). Mortality rates for juveniles are unknown, so we have as- sumed a range of 0.5-0.8. From an initial popula- tion (N0) of 28.9 million juvenile crabs in summer of 1980, the number surviving 3.5 yr (N) was cal- culated from the equation N = N0e-Zt where z represents the annual mortality rate and t is the time interval. Values of 2 used were 0.8 for the first half year it = 0.5), 0.5 for the second half year, and 0.2 (as above) for the remaining 2.5 yr necessary to reach legal size, assuming crabs enter the fishable population at that age, as sug- gested by Cleaver (1949). At these mortality rates, about 9.2 million adult crabs might remain by 481 FISHERY BULLETIN: VOL. 82, NO. 3 December 1983, of which about half, 4.6 million, would be males subject to the commercial fishery. If an equivalent number of crabs were available from the 1980 recruitment to Willapa Bay, a large bay equaling or exceeding Grays Harbor in area and located about 20 km south, then about 9.2 million legal male crabs of estuarine origin might be avail- able to the commercial fishery in 1984-85 from larvae and early instars that utilized these two Washington estuaries in 1980-81. Washington coastal crab landings for the period 1971-80 have averaged 3,500 t/yr (PMFC 1981), or about 3.85 million crabs (at 0.9 kg/crab). Thus, these two bays could theoretically serve as nursery grounds for more than enough crabs necessary to maintain a viable commercial fishery in Washing- ton. However, landings over the past 40 yr have fluctuated from 1,000 to 8,000 t, with a 9-12 yr period, so it is impossible to predict how the esti- mated contribution of the 1980 year class will compare to 1984 commercial landings. ACKNOWLEDGMENTS This study represents a portion of a Ph.D. dis- sertation submitted to the University of Washing- ton School of Fisheries by B. G. Stevens, and was supported by Contract Nos. DACW-67-80-C-0086 and DACW-67-81-M-1096 from the U.S. Army Corps of Engineers (USACE) and Washington De- partment of Fisheries (WDF). Support for B. Ste- vens was also received from the University of Washington Graduate School in the form of a Sarah Denny Graduate Research Fellowship. We are greatly indebted to J. C. Hoeman for his assis- tance with field sampling, and to J. Armstrong and R. Thorn, USACE, for their advice and support throughout the study. We extend appreciation to S. Barry and R. Westley, WDF, for their roles in project administration. We are also grateful to J. Little and R. Cusimano for their field and labora- tory assistance, to the many paid and volunteer students who assisted us, especially G. McKeen, W Latan, and J. Garcia, and to four anonymous reviewers for their helpful comments. We also thank V. Munzlinger (University of Wash.) for typ- ing the first draft, and S. Cooke, Y. Wilson, and D. Miller (Old Dominion University, Norfolk, Va.) for assistance with graphics. LITERATURE CITED ARMSTRONG, D. A., B. G. STEVENS, AND J. C. HOEMAN. 1982. Distribution and abundance of Dungeness crab and Crangon shrimp and dredging-related mortality of inver- tebrates and fish in Grays Harbor, Washington. Final report to Washington State Department of Fisheries and U.S. Army Corps Engineers, Contract No. DACW67-80- C-0086, by the University of Washington School of Fisheries, 349 p. BOTSFORD, L. W, AND D. E. WlCKHAM. 1978. Behavior of age-specific, density-dependent models and the northern California Dungeness crab (Cancer magister) fishery. J. Fish. Res. Board Can. 35:833- 843. BUTLER, T. H. 1956. The distribution and abundance of early post-larval stages of the British Columbia commercial crab. Fish. Res. Board Can., Prog. Rep. 107, p. 22-23. 1960. Maturity and breeding of the Pacific edible crab, Cancer magister Dana. J. Fish. Res. Board Can. 17:641- 646. 1961. Growth and age determination of the Pacific edible crab Cancer magister Dana. J. Fish. Res. Board Can. 18:873-891. California Department of Fish and Game. 1981. A summary of the Dungeness crab research pro- gram, 1974-1980. Mar. Res. Admin. Rep. 81-3, 13 p. CASSIE, R. M. 1954. Some uses of probability paper in the analysis of size frequency distributions. Aust. J. Mar. Freshw. Res. 5:513-522. CLEAVER, F. C. 1949. Preliminary results of the coastal crab (Cancer magister) investigation. Wash. Dep. Fish., Biol. Rep. 49A, p. 47-82. COCHRAN, W. G. 1953. Sampling techniques. John Wiley and Sons, N.Y., 330 p. ELLIOTT, J. M. 1977. Some methods for the statistical analysis of samples of benthic invertebrates. Sci. Publ. 25, Freshwater Biological Assoc, Ferry House, Ambleside [Eng.], 160 p. GOTSHALL, D. W. 1977. Stomach contents of northern California Dungeness crabs, Cancer magister. Calif. Fish Game 63:43-51. 1978a. Relative abundance studies of Dungeness crabs, Cancer magister, in Northern California. Calif. Fish Game 64:24-37. 1978b. Catch-per-unit-of-effort studies of Northern California Dungeness crabs, Cancer magister. Calif. Fish Game 64:189-199. 1978c. Northern California Dungeness crab, Cancer magister, movements as shown by tagging. Calif. Fish Game 64:234-254. JOW, T. 1965. California-Oregon cooperative crab tagging study. In Pacific Marine Fisheries Commission 16th and 17th Annual Reports for the years 1963-64, p. 51-52. LOUGH, R. G. 1976. Larval dynamics of the Dungeness crab, Cancer magister, off the central Oregon coast, 1970-71. Fish. Bull., U.S. 74:353-375. MACKAY, D. C. G. 1942. The Pacific edible crab, Cancer magister. Fish. Res. Board Can. Bull. 62, 32 p. Orcutt, h. g., r. n. tasto, p w. Wild, c. Haugen, and P COLLIER. 1975. Dungeness crab research program. Calif. Dep. Fish Game, Mar. Res. Admin. Rep. 75-12, 31 p. 482 STEVENS AND ARMSTRONG: CANCER MAG1STER IN GRAYS HARBOR ORCUTT, H. G., R. TASTO, P W. WILD, C. HAUGEN, AND E. EBERT. 1976. Dungeness crab research program. Calif. Dep. Fish Game, Mar. Res. Admin. Rep. 76-15, 42 p. ORCUTT, H. G. (compiler) 1977. Dungeness crab research program. Calif. Dep. Fish Game, Mar. Res. Admin. Rep. 77-21, 55 p. 1978. Dungeness crab research program. Calif. Dep. Fish Game, Mar. Res. Admin. Rep. 78-16, 24 p. PMFC (PACIFIC MARINE FISHERIES COMMISSION). 1981. Dungeness crab fishery, 1979-80. In 33rd Annual report of the Pacific Marine Fisheries Commission for the year 1980, p. 22-23. POOLE, R. L. 1967. Preliminary results of the age and growth study of the market crab (Cancer magister) in California: The age and growth of Cancer magister in Bodega Bay. In Pro- ceedings of the Symposium on Crustacea, Part II, p. 553- 567. Mar. Biol. Assoc. India, Symp. Ser. STEVENS, B. G. 1981. Dredging-related mortality of Dungeness crabs as- sociated with four dredges operating in Grays Harbor, Washington. Report to the U.S. Army Corps Engineers, Seattle District, Contract No. DACW-79-C-0045, by the Washington State Department of Fisheries, 148 p. 1982. Distribution, abundance, and food habits of the Dungeness crab, Cancer magister, in Grays Harbor, Wash- ington. Ph.D. Thesis, Univ. Washington, Seattle, 213 p. [Available Univ. Microfilms, Int., Ann Arbor, Mich.] STEVENS, B. G., AND D. A. ARMSTRONG. 1984. Diel activity of an estuarine population of Dunge- ness crabs, Cancer magister, in relation to feeding and environmental factors. J. Crust. Biol. 4:390-403. STEVENS, B. G., D. A. ARMSTRONG, AND R. CUSIMANO. 1982. Feeding habits of the Dungeness crab Cancer magis- ter as determined by the Index of Relative Impor- tance. Mar. Biol. (Berl.) 72:135-145. TASTO, R. N. 1983. Juvenile Dungeness crab, Cancer magister, studies in the San Francisco Bay area. In P W. Wild and R. N. Tasto (editors), Life history, environment, and maricul- ture studies of the Dungeness crab, Cancer magister, with emphasis on the central California fishery resource, p. 135-154. Calif. Dep. Fish Game, Fish Bull. 172. TEGELBERG, H., AND R. ARTHUR. 1977. Distribution of Dungeness crabs (Cancer magister) in Grays Harbor, and some effects of channel maintenance dredging. Appendix N. In Maintenance dredging and the environment of Grays Harbor, Washington. Final Report to U.S. Army Corps Engineers, Seattle District, 94 p. WEYMOUTH, F. W, AND D. C. G. MACKAY. 1936. Analysis of the relative growth of the Pacific edible crab, Cancer magister. Proc. Zool. Soc. Lond. 1936 (Pt. l):257-280. 483 AGE, GROWTH, AND MORTALITY OF GRAY TRIGGERFISH, BALISTES CAPRISCUS, FROM THE NORTHEASTERN GULF OF MEXICO Allyn G. Johnson and Carl H. Saloman1 ABSTRACT Age, growth, and mortality of gray triggerfish, Batistes capriscus, from the northeastern Gulf of Mexico were estimated from sections of the first dorsal spine of 1,746 fish. The oldest female was estimated to be 12 years old and the oldest male was 13 years old. The von Bertalanffy growth equa- tions, using weighted means, were as follows: males, lt = 491.9 (1 - e -0.382U- 0.227) > ancj females, lt = 437.5 (1 - e-0.383(<- 0.150)) where / = fork length in millimeters and t = age in years. The mean annual mortality rate as determined by four methods of analyses (based on number of fish at age) ranged from 0.32 to 0.53. The weight-length relationships of gray triggerfish were males, W = 6.71505 x 10" 6 L3 -187, and females, W = 1.3939 x 10" 5 L3065, where W = weight in grams and L = fork length in millimeters. Exploitation offish from the northeastern Gulf of Mexico by recreational and commercial fisher- men has created a demand for underutilized fish resources. One of the abundant fish resources that is being subjected to exploitation is the gray triggerfish, Balistes capriscus. A dramatic in- crease in demand for this species can be seen in the commercial landings on the west coast of Florida: 7.8 t in 1967 and 26.7 t in 1977 (Anon- ymous 1967, 1977). This species is known to occur in the western and eastern Atlantic. In the western Atlantic, its range is from Nova Scotia to Argentina, includ- ing the Gulf of Mexico (Briggs 1958; Moore 1967). In the Gulf of Mexico, the gray triggerfish is a primary reef fish inhabiting the area between 12 and 42 m in depth (Smith 1976), except for its first year of life when it is planktonic and associ- ated with Sargassum (Dooley 1972). The harvest of the gray triggerfish in the northeastern Gulf of Mexico and its utilization of reef habitats has created a need to know more about the biology of this species, especially age, growth, and mortality. Age and growth of gray triggerfish, using the first dorsal spine, has been reported only for the southwestern coast of Africa (Anonymous 1980; Caveriviere et al. 1981). This paper reports the results of our investigation on age, growth, and mortality, using the first dorsal spine of gray triggerfish from the northeastern Gulf of Mexico. METHODS AND MATERIALS The hook and line fishery for gray triggerfish off Panama City, Fla., was sampled from May 1979 to March 1982. During this period, 2,808 fish were sampled and from each the fork length in millimeters and total weight in grams measured and recorded. The sexes of the fish were also recorded when determinable by gross examina- tion of the gonads. First dorsal spines were avail- able from 1,746 of the 2,808 fish in the collection. Total length (TL), standard length (SL), and fork length (FL) were measured in millimeters from 100 fish to develop length conversion formulas. The first dorsal spines were processed for examination as follows: 1) removing the first 5 mm of spine shaft above the condyle with a Dremel2 tool; 2) placing the shaft section on a mounting tag using Lakeside No. 70c thermo- plastic cement and sectioning the shaft using the method described by Berry et al. (1977); 3) re- moving three 0.18 mm thick serial sections from the cement with 50% isopropanol; and 4) mount- ing the clean sections in 209c Piccolyte cement (20% Piccolyte, 809c xylenes) on glass slides. Spine cross sections were examined and mea- sured using a closed-circuit television using a 50 'Southeast Fisheries Center Panama City Laboratory, Na- tional Marine Fisheries Service, NOAA, 3500 Delwood Beach Road, Panama City, FL 32407-7499. Manuscript accepted February 1984. FISHERY BULLETIN: VOL. 82. NO. 3, 1984. 2Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 485 FISHERY BULLETIN: VOL. 82, NO. 3 mm 3.5 macro lens which projected an image of the section on to a monitor screen at 20 x magni- fication. Illumination was by transmitted light. Translucent (light) bands on the cross sections were counted and the distances (in millimeters) from the center of the spine to the distal edge of each band was measured. The spine radius (R) was defined as the maximum distance (in milli- meters) from the center of the section (appears as a small hole) to the posterior distal edge (Fig. 1). Additionally, before the spines were sectioned, the anterior-posterior thickness (T) of 200 dorsal spines was measured to the nearest 0.01 mm at the sectioning site. The type of growth (opaque = dark, translucent = light) of the margin of each section was noted. The sections were read three times. The relationships of R, T, and body weight to FL and the relationships between TL, SL, and FL were determined by least squares methods fol- lowing the suggestions of Ricker (1975). A com- puter program by Abramson (1971) was used to fit weighted back-calculated mean length at age to von Bertalanffy growth curves. The growth equa- tion (von Bertalanffy 1938, 1957) and values were as follows: lt= I (1 - e -K(t-t0) ) where It = length at age t, I = asymptotic length, K = growth coefficient, ^o = time when length would theoretically be zero. Estimates of annual mortality (a), annual sur- vival (s), and instantaneous mortality (i) were developed for the total collection (2,808 fish) (Ricker 1975). Length-frequency data were con- verted to age-frequency distribution (Ny = num- ber of fish caught in age class y ) by applying age- FIGURE 1. — Sections of gray triggerfish first dorsal spines from fish collected off Panama City, Fla. (A) Spine section from a 1-yr-old male (263 mm FL) collected 4 September 1980 with spine radius R labeled. (B) Spine section from a 2-yr-old female (336 mm FL) collected 11 September 1980. (C) A 3-yr- old female (315 mm FL) collected 8 August 1980. (D) A 4-yr- old male (350 mm FL) collected 13 August 1980. (E) A 5-yr-old female (331 mm FL) collected 24 September 1980. (F) A 6-yr- old male (477 mm FL) (seventh mark forming on margin) collected 25 June 1980. length keys. Ages III through IX of the resultant catch curves were analyzed by the methods of Heincke (1913), Jackson (1939), Robson and Chap- man (1961) and by finding the slope (m) of a regression line fitted to ln(Ny) and y and substi- tuting the equation a = 1 — em. RESULTS A positive relationship was found between the growth of the first dorsal spine and FL. The relation between FL and R was as follows: FL = 4.58 i?0951 with a correlation coefficient (r) of 0.84. The relation between FL and T was as follows: FL = 24.87 T1422 with r = 0.89. The variation in the two relationships probably re- sulted from the slight tapering of the spine in the area from which the sections were taken and the effect of sectioning. The Fh-R relationship was used for back-calculation of sizes at previous ages. The spine sections possessed distinct dark- light banding patterns (Fig. 1) and the agreement between readings as to the numbers of bands was 98% [Beamish and Fournier's (1981) index of average error was 0.0072]. The translucent (light) band formation occurred during spring and summer (April to October with a peak during June-July), and the mean marginal opaque in- crement was least during this period of time (Table 1). We thus considered the translucent bands on the first dorsal spines to be annular TABLE 1. — Percent frequency of dorsal spines with translucent (light) margins and mean marginal measurements of opaque (dark) margins in millimeters for gray triggerfish from northeast Gulf of Mexico, 1979-82. Month 1 2 3 4 5 6 7 8 9 10 11 12 Percent of fish with translucent margins 0.00 0.00 0.00 5227 41.86 56.70 53.59 23.79 2.40 0.90 0.00 0.00 Mean opaque marginal increment for fish 2 light bands — — 1.40 1.24 1 09 0.74 0.82 0.64 1.00 1 70 — — 3 light bands — 0.79 0.49 0.59 0.60 0.58 0.56 0.36 0.50 — — — 4 light bands — 0.56 0.47 0.43 0.52 0.48 0.56 038 0.60 — — — 5 light bands — 0.44 0.24 0.26 0.06 0.24 0.37 0.037 — — — — Total number of fish 6 19 83 88 215 194 209 248 250 111 97 13 486 JOHNSON and SALOMAN: AGE, GROWTH. AND MORTALITY OF GRAY TRIGGERFISH . «N ■ . 487 FISHERY BULLETIN: VOL. 82, NO. 3 deposits and suitable for age determination. Lengths varied within age classes and length ranges overlapped between age classes (Tables 2, 3, 4). For example, males with three annuli (translucent bands) ranged from 258 to 537 mm FL and those with four annuli ranged from 250 to 549 mm FL. There was, however, a general trend of increasing modal length with increase in age. The gray triggerfish is a moderately long-lived species. The oldest male was estimated to be 13 yr old (544 mm FL) and the oldest female was 12 yr old (561 mm FL). The back-calculated and empirical sizes at age are presented in Tables 5, 6, and 7. The average mean back-calculated length at age for males was slightly (5-50 mm FL) larger than that for fe- males at age 1-9, after which the females were larger. Only three fish were collected that were older than 9 yr; thus the reversal of the trend is probably an artifact caused by few samples. The von Bertalanffy growth parameters varied slightly between males, females, and all fish. The von Bertalanffy equations were: males It = 491.9(1 - e 0.382(^-0.227) ), females lt = 437.5(1 - e-o.383«-o.i50))) all fish lt = 466.0(1 -eo.382U-o.i89>K TABLE 2. — Length composition, in percent, of male gray triggerfish by age groups from northeast Gulf of Mexico, 1979-82. Length Age in years Total group (FLmm) 0 1 2 3 4 5 6 7 8 9 of fish 150-199 50.00 50 00 2 200-249 50 00 50 00 18 250-299 14.46 44.58 36.14 3.62 1.20 135 300-349 0.«6 26 32 42.10 2566 2.63 1.97 0.66 307 350-399 926 35.80 32.72 18.52 3.70 339 400-449 1.48 1852 29 62 25.19 15 56 5.93 3.70 233 450-499 11.02 1224 26 53 29.59 16.34 9.18 408 1.02 155 500-549 3.30 27.47 30.77 27.47 8.79 1.10 121 550 and larger 66 67 8 33 1667 8.33 24 Total 1 23 184 371 339 229 117 47 19 4 1,334 TABLE 3. — Length composition, in percent, of female gray triggerfish by age groups from northeast Gulf of Mexico, 1979-82. Length Age in years Total group (FLmm) 1 2 3 4 5 6 7 8 9 10 11 12 of fish 200-249 41.67 33.33 25.00 19 250-299 5.04 4286 33.61 15.97 0.84 1.68 207 300-349 0.47 1408 4461 26.29 7.98 4.69 1.41 047 453 350-399 4 64 2384 3576 23.18 10.60 1.98 304 400-449 15 38 27 69 21.54 18.46 9.23 3.08 4.62 122 450-499 2.63 7.89 21.06 28.95 2632 789 526 74 500-549 4.55 22 73 22.73 18.18 909 9.09 13.63 40 550 and larger 20 00 40.00 20.00 20 00 9 Total 20 175 374 321 168 107 35 10 17 0 0 1 1,228 TABLE 4. — Length composition, in percent, of gray triggerfish (all fish) by age groups from northeast Gulf of Mexico, 1979-82. Length Ac ie in years Total group (FLmm) 0 1 2 3 4 5 6 7 8 9 10 11 12 13 of fish 150-199 50.00 50 00 2 200-249 47.83 39.13 13.04 32 250-299 9.13 4247 35.16 11.42 0.91 091 365 300-349 0.47 1827 41 69 26 93 7.72 3.75 0.94 0.23 848 350-399 7.12 31 05 34 47 19 09 6.84 1.43 691 400-449 090 16 59 28 70 25.11 16.59 7.17 3.14 1.38 0.45 418 450-499 1.36 10.20 26 53 28 57 19.73 884 2.73 204 244 500-549 4.23 27 37 28 20 24 79 8.36 258 4.37 0.10 172 550 and larger 5.00 50 00 15.00 10.00 5.00 10.00 5.00 36 Total 1 53 373 805 737 445 249 92 29 22 1 0 1 1 2,808 488 JOHNSON and SALOMAN: AGE, GROWTH, AND MORTALITY OF GRAY TRIGGERFISH The weight-length relationships for gray trig- gerfish computed for the equation W = ah , where W is weight in grams and L is FL in millimeters, were as follows: males W = 6.7105 x 10 6 L3187, n = 169, females W = 1.393 x 10" 5 L n = 167, 3.065 r = r = 0.97, 0.93, TABLE 5. — Back-calculated fork lengths (mm) at age ■ for male gray triggerfish from the northeastern Gulf of Mexico, 1979-82. Age group Mean length ±1SD at capture N Average back-calculated FL at age 1 2 3 4 5 6 7 8 9 10 11 12 13 1 250.0 ±29.0 18 137.7 II 313.4±41.8 99 1239 2480 III 357.4 ±55.8 192 119.2 243 0 319.6 IV 407.8 ±66.6 186 124.1 2442 322 8 376.3 V 450.6 ±56.2 134 1282 245 .1 324.7 381 8 4270 VI 461 9 ±56.0 72 1322 245.3 3097 3643 4098 443.2 VII 4745±51.5 28 126.3 231.8 301.2 350.7 390 7 430.5 4587 VIII 462.5 ±27.4 10 134.0 236 5 3074 3443 372.8 399.5 4297 4486 IX 511 5±78.5 2 1658 293.1 323.0 3620 3898 424.7 4687 4899 504.0 XIII 544.0 ± — 1 91.0 308 5 395 0 452 7 461.5 479.0 492.1 4965 500.8 509.5 513.8 526.8 535.4 Weighted mean 124.9 244 3 3196 373.7 415.5 436.2 4529 458.7 502.9 509.5 513.8 526.8 535.4 ±1SD ±41.0 ±49.4 ±55.5 ±60.9 ±60.1 ±56.8 ±486 ±32.3 ±54.5 — — — — N 742 724 625 433 247 113 41 13 3 1 1 1 1 Annual increment 1249 119.4 75.3 54.1 41 8 20.7 16.7 5.8 442 6.6 4.3 13.0 86 TABLE 6. — Back-calculated fork lengths (mm) at age for female gray triggerfish from the northeastern Gulf of Mexico, 1979-82. Age group Mean length ±1SD at capture N Average back-calculated FL at age 1 2 3 4 5 6 7 8 9 10 11 12 1 259.0 ±34.0 12 1556 II 300.4 ±35.2 93 120.1 241.4 III 330.8 ±46.2 187 117.0 227.8 298.3 IV 360.1 ±57.5 161 117.9 2263 2924 335.6 V 398 5 ±64 0 85 128 2 2265 291.7 340.7 378.5 VI 402.5 ±63.6 55 113.5 2184 2760 320.1 357.5 387.1 VII 419.8 ±73.6 16 1249 2053 2609 307.2 345.9 3790 405.1 VIII 448.8 ±79.6 5 109.0 193.3 262 3 305.7 342 4 3768 415.1 437.4 IX 457.8 ±71 2 10 101.8 208.1 2648 306 9 340.2 373.7 402 5 431.0 447.9 XII 561. 0± — 1 121.2 258 2 390.4 463.2 481.2 494.6 516.8 5257 534 6 547.8 552.2 556.6 Weighted mean 119.4 227.3 291.1 332 0 366 1 384.7 4093 438.9 455.8 547.8 552.2 5566 ±1SD ±37.8 ±43.1 ±45 1 ±53.3 ±61.1 ±62.4 ±70.0 ±71.4 ±73.6 — — — N 625 613 520 333 172 87 32 16 11 1 1 1 Annua increment 1194 107.9 638 409 34.1 18.6 24.6 29.6 169 92.0 44 44 TABLE 7. — Back-calculated fork lengths (mm) at age for all gray triggerfish collected from the northeastern Gulf of Mexico, 1979-82. Age group Mean length ±1SD at capture Average back-calculated FL at age N 1 2 3 4 5 6 7 8 9 10 11 12 13 1 259.3 ±29.7 34 140.9 II 308 6 ±42.3 210 121.7 247.4 III 343 9 ±52.1 424 118.0 2354 308 9 IV 384 0 ±66 1 398 122.0 2353 307.3 356.1 V 425 8 ±70 7 243 128.1 236.6 309.1 361.8 403.9 VI 434.7 ±67 1 141 123.9 232.6 2950 3460 386.6 417.7 VII 541.0 ±64.6 49 123.6 221.0 286 7 334 8 373.2 4090 435.7 VIII 464 8 ±54.2 16 134.2 227.1 296.3 335.3 372.2 401.3 432.6 452.2 IX 478.1 ±72.3 14 117.4 233 5 282 6 324.1 3544 3888 420.1 4484 468 6 X 439 0 ± — 1 108.3 2040 251 .5 2748 286 3 309.2 331.9 354.9 384.1 4208 XII 561 0± — 1 212.2 2582 3904 463 2 481.2 4946 5168 525.7 5346 5478 5522 556.6 XIII 544.0 ± — 1 91.9 3084 3950 4527 461.5 4790 492.1 496.5 5008 509.5 513.8 5267 535.4 Weighted mean 112 8 236 5 3058 354.1 3929 412.9 432.8 451.2 460 4 492 7 533 0 541.7 535.4 ±1SD ±40.0 ±47.8 ±52.6 ±61.0 ±65.9 ±66.0 ±64.0 ±61.8 ±72.0 ±65.1 ±27.2 ±21.1 — N 1,523 1,498 1,288 864 466 223 82 33 17 3 2 2 1 Annual increment 1128 123.7 69.3 48.3 38.8 20.0 19.9 18.4 9.2 32.3 40.3 8.7 -6.3 489 FISHERY BULLETIN: VOL. 82, NO. 3 all fish W = 2.146 x 10 5 L2"2, r = 0.96, n = 175. Conversions between different length measures were linear and expressed as follows: FL vs. TL: FL = 29.704 + 0.774 TL, r = 0.97, n = 100, FL vs. SL: FL = 22.823 + 1.171 SL, r = 0.99, n = 100, TL vs. SL: TL = 9.666 + 1.446 SL, r = 0.96, n = 100. Estimates of mortality (a, s, and i) varied slightly between estimation methods (Table 8). Full recruitment to the fishery was considered to be at 3 yr for both sexes. Estimates of a were between 0.32 and 0.53 with i between 0.39 and 0.75 (Table 8). TABLE 8. — Estimated annual mortality (a), annual survival (s), and instantaneous mortality (i) by esti- mation technique for gray triggerfish from the north- eastern Gulf of Mexico, 1979-82. Estimation technique Robson and Hemcke Jackson Chapman Regression (1913) (1939) (1961) analysis Males a 0.33 0.32 044 0.53 s 0.67 0.67 056 047 / 0 40 0.39 0.57 0.75 Females a 036 0.32 045 0.47 s 0.64 068 0.55 0.53 / 0.45 0.38 0.59 0.64 All fish a 0.34 0.33 0 44 0.49 s 0.66 0.64 0.56 0.51 / 041 0.40 0.58 0.67 DISCUSSION The variation in length at age and overlays of length ranges between ages found in gray trig- gerfish is not unusual in fish from southeastern U.S. waters. Many species such as king mackerel, Scomberomorus cavalla; Spanish mackerel, S. maculatus; red grouper, Epinephelus morio; sail- fish, Istiophorus platypterus ; and black sea bass, Centropristis striata, have large variations in size within age groups (Beaumariage 1973; Powell 1975; Moe 1969; Jolley 1977; Waltz et al. 1979). Our gray triggerfish growth rates are similar to growth information from the Gulf of Mexico, but not information from Africa. Beaumariage (1969) reported growth rates for three tagged fish (250, 270, and 332 mm TL) from the northeastern Gulf of Mexico. His fish grew at a rate of 187.2, 153.6, and 51.6 mm/yr. If one considers Beaumar- iage's fish to be 2, 2, and 3 yr old, respectively, then his growth increments are similar to ours (Table 7). Gray triggerfish age and growth have been reported from southwestern Africa (Ivory Coast-Ghana-Togo area) by Anonymous (1980). We took the information in Anonymous' figure 11 and converted it to mean length at capture per age which gave the following approximate values: age I, 148 mm; age II, 203 mm; values are about 100 mm less than ours for each age (Table 7). Caveriviere et al. (1981) provided comprehen- sive information on the age and growth of gray triggerfish off Senegal and the Ivory Coast. Two hypotheses with regard to band formation were suggested: A) one band per year, and B) two bands per year. The sizes (FL) at age (in years) for Senegal fish by hypotheses were age I, 153 mm for hypothesis A, 90 mm for B; age II, 231 mm for A, 170 mm for B; age III, 285 mm for A, 238 mm for B; age IV, 322 mm for A, 290 mm for B; age V, 348 mm for A, 324 mm for B. Sizes at age of our Gulf of Mexico fishes (Table 7) were larger after the first year than predicted by both of the above hypotheses for Senegal fish. The sizes at age for Ivory Coast fish were smaller than both the Senegal and the Gulf of Mexico fish using the hypothesis of one band formed per year. These differences may be the result of different envi- ronments, biology, methods of capture, or aging. Anonymous (1980) suggested that the African fish have a seasonal offshore migration to avoid the cold coastal water (during the third quarter of the year) which is the result of upwelling. Gulf of Mexico fish are not known to have migratory habits, and thus might not be subject to the energy expense such movements incur. More in- formation on the life histories and environments of these groups of gray triggerfish is needed to explain the observed variations. The K values (growth coefficient) of gray trig- gerfish varied between 0.382 and 0.383. These values were similar to, but higher (about 0.1) than, those reported for other demersal marine fish from the southeastern United States (see Manooch 1982 and Pauly 1978 for a listing of values). The K values estimated for gray trigger- fish may be high because of the low asymptotic lengths that were found. Additional investigation of the larger and older fish is needed to evaluate the growth coefficients of this species. The esti- 490 JOHNSON and SALOMAN: AGE, GROWTH, AND MORTALITY OF GRAY TRIGGERFISH mates of mortality (Table 8) were similar to those of demersal marine fish such as the white grunt, Haemulon plumieri, where a = 0.37-0.51 and the red porgy, Pagrus pagrus, where a = 0.32-0.55 (Manooch 1976; Manooch and Huntsman 1977) that inhabit similar habitats. The mortality rates for gray triggerfish probably reflected the exploi- tation level on this species in the northeast Gulf of Mexico. Nelson and Manooch (1982) reported similar values (/ = 0.39-0.50) for red snapper, Lutjanus campechanus, from the Carolinas and Florida coasts where the fishing pressure is light to medium and much higher values (i — 0.78- 0.94) from the fishery off Louisiana where the commercial fishing pressure is high. The effect of fishing on gray triggerfish populations were therefore assumed to be similar to the effects of fishing on these other reef fish resources. ACKNOWLEDGMENTS We thank H. F. Horton, Oregon State Univer- sity, Corvallis, Oreg.; E. D. Prince, National Marine Fisheries Service, Southeast Fisheries Center, Miami, Fla.; and R. S. Nelson, National Marine Fisheries Service, Southeast Fisheries Center, Beaufort, N.C., for their valuable criti- cisms of the manuscript. LITERATURE CITED ABRAMSON, N. J. 1971. Computer programs for fish stock assessment. FAO Fish. Tech. Pap. 101, 146 p. ANONYMOUS. 1967. Florida landings. U.S. Dep. Inter. Bur. Commer. Fish., Annu. Summ. Curr. Fish. Stat. 4660, 18 p. 1977. Florida landings. U.S. Dep. Commer., Natl. Mar. Fish. Serv., Annu. Summ. Curr. Fish. Stat. 7517, 14 p. 1980. Report of the ad hoc working group on Sardinella off the coast of Ivory Coast-Ghana-Togo. FAO, CECAF/ ECAF Series 8012 (EN), 72 p. BEAMISH, R. J., AND D. A. FOURNIER. 1981. A method for comparing the precision of a set of age determinations. Can. J. Fish. Aquat. Sci. 38:982-983. BEAUMARIAGE, D. S. 1969. Returns from the 1965 Schlitz tagging program including a cumulative analysis of previous results. Fla. Dep. Nat. Resour., Mar. Res. Lab., Tech. Ser. 59, 38 p. 1973. Age, growth, and reproduction of king mackerel, Scomberomorus cavalla, in Florida. Fla. Mar. Res. Publ., No. 1, 45 p. BERRY, F. H., D. W. LEE, AND A. R. BERTOLINO. 1977. Age estimates in Atlantic bluefin tuna — an objec- tive examination and an intuitive analysis of rhythmic markings on vertebrae and in otoliths. Int. Comm. Conserv. Atl. Tunas, Collect. Vol. Sci. Pap. 6 8 KENT ISLAND • = WINTER t - BREEDING • = SUMMER i i i i 7 8 9 10 I I 12 13 14 15 16 17 18 19 20 TIME OF DAY (PST) FIGURE 2. — Graph of time of day and the mean number of seals hauled out per time period for winter, breeding, and summer seasons at Kent Island. Disturbance Camera and field observations of Kl recorded 539 actual and zero-seal disturbances. Of those with identifiable cause, 33.1% were nonpower boats, 10.0% people on foot, 7.8% powerboats, 3.4% clam diggers or bait harvesters, 2.8% dogs, and TABLE 1. — The relationship between time of day and the number of seals hauled out per season on Kent Island (Kl) and Pickleweed Island (PWI); x is the mean number of seals per hour, SD is the standard deviation, and n is the sample size. Time Season 0700 0800 0900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 Winter Kl x 6.8 8.3 84 85 84 9.0 9.5 9.8 9.8 8.7 5.6 SD 9.5 11.6 11.9 11.6 11.7 11.4 12.0 12.5 12.0 11.5 8.9 n 84 83 84 84 87 88 87 88 87 87 12 PWI x 0 0.8 3.5 69 63 5.5 6.2 5.4 2.2 22 SD 0 2.1 4.9 6.7 6.2 6.6 6.8 6.0 39 3.6 n 9 8 10 10 11 11 10 10 10 10 Breeding Kl x 69 10.6 13.3 134 15.1 15.1 16.1 16.4 16.9 159 14.3 11.6 SD 10.2 13.8 16.5 16.0 16.8 16.3 16.8 16.1 16.5 16.2 15.3 13.8 n 105 108 112 114 118 119 118 119 116 116 116 94 PWI x 12 1.9 26 3.3 4.0 4.2 4.7 49 49 3.8 3.3 0.1 SD 3.1 4.8 58 66 7.3 80 83 84 85 7.3 6.5 0.3 n 87 89 86 89 91 89 92 91 91 90 83 37 Summer Kl x 12.7 15.2 16.5 173 16.8 19.1 18.3 17.8 16.4 16 1 14.8 13.1 120 106 SD 14.4 15.9 17.4 17.7 16.5 18.0 18.2 18.2 17.8 16.8 153 14.9 15.1 15.3 n 77 83 92 98 104 102 101 105 105 103 86 51 27 PWI x 0.4 1.2 2.1 3.5 4.7 54 5.0 4.5 36 24 1.3 0.8 0.6 0.7 SD 10 4.0 5.5 9.2 10.8 12.0 12.8 11.1 10.6 7.4 3.9 3.0 2.5 2.2 n 91 90 91 94 101 101 99 99 98 98 98 92 70 14 496 ALLEN ET AL.: EFFECT OF DISTURBANCE ON HARBOR SEALS 10 tr z> o I Or UJ Q. < tr LU 4 CO Z Z 3 < UJ 2 2 PICKLEWEEO ISLAND • = WINTER ' = BREEDING ■ = SUMMER 10 II 12 13 14 15 16 17 18 19 20 TIME OF DAY (PST) FIGURE 3. — Graph of time of day and the mean number of seals hauled out per time period for winter, breeding, and summer seasons at Pickleweed Island. 0.7% helicopters. The camera did not record the cause in 40% of the disturbances. During De- cember, January, and February, commercial bait harvesters accounted for 25.7% of the distur- bances. Disturbances from aircraft were detected only in field observations. The seals were dis- turbed at least once on 71% of 356 d when the KI camera was functioning, and during these days, 72.7% of 539 disturbances caused the seals to dis- perse (Table 3). On 211 d during which the PWI camera was functional, seals reacted to 57% of 236 disturbances (Table 3). The frequency of ac- tual disturbances per day averaged highest during the summer for both KI and PWI. Of the actual disturbances of known cause, most occurred within 100 m of the KI site and resulted more from nonpower boats than from any other source (Table 4). The deletion of terms [13] and [14] from the log-linear model in backward selection, however, indicated that only the distance of a dis- turbance significantly affected seal behavior (Ta- ble 5). Seals did not react differentially to any disturbance type and did not react more to distur- bances during weekend/holidays than during weekdays. The relative magnitude of the weights associated with the distance/seal response in- teraction term [12] denoted that seals responded to TABLE 2. — Correlation coefficients between tide level and the number of seals hauled out at hourly intervals per season at Kent Island (Kl)and at Pickleweed Island (PWI), from camera data; n = number of censuses. Insufficient data were available for PWI during the winter. A positive correlation indicates a positive relationship between seal number and low tide, and a negative correlation indicates a negative relationship between seal number and high tide. Winter Breec ing Summer KI KI PWI KI PWI Time (n = 27) (n = 26) (n = 16) (n = 24) (n =21) 0800 054" 0.35 023 0.49" 0.36 0900 075" 034 0.17 0.69" 0.20 1000 0.67" 042" -0.17 0 55" 0.62" 1100 0.71" 0.31 -0.30 0.57" 0 52" 1200 0.64" 0.14 006 068" 048* 1300 0.65" -0.12 0.34 0.54" 0.45- 1400 0.58" -0.15 0.42 0.62" 0.37 1500 0.55" 0.001 022 0.67" 0.23 1600 0.20 0.11 067" 0.11 • Significant at P = 0.05 "Significant at P = 0.01. TABLE 3. — The frequency of disturbances on Kent Island (KI) and Pickleweed Island (PWI) by season; n is the number of days the camera was functional, A is the number of actual and zero- seal disturbances combined, A 1 is the number of actual distur- bances type I, A In is the mean number of all disturbances per day, and A \ln is the mean number of the actual disturbances per day. KI PWI Season n A A1 Aln AMn n A A\ Aln AMn Breeding Summer Winter Totals 150 216 158 1.4 114 215 167 1.9 92 108 67 12 356 539 392 1.5 1.1 1.5 0.7 1.1 96 101 54 1.1 102 121 75 1.2 13 14 6 1.1 211 236 135 1.1 0.6 07 0.5 0.6 TABLE 4. — Data from the Kent Island camera used in log-linear model analysis (Bishop et al. 1975). The numbers in the table are the number of disturbances where seals were present for each category. Y = response (at least one seal left site); N = no response (no seals left site). Weekday Weekend/holiday Distance: 100 m 101- 200 m 201- 300+ m :100 m 101- 200 m 201- 300+ m Y N Y N Y N Y N Y N Y N People/dog Nonpower boat Power boat 12 35 7 0 2 1 1 8 3 0 5 3 0 1 1 0 7 0 12 0 32 1 8 0 2 5 0 0 1 1 0 1 1 0 6 0 TABLE 5. — Backward selection of log-linear model using data in Table 4; model variables are 1 (response), 2 (distance), 3 (day of week), and 4 (disturbance type). An asterisk indicates a term that is significantly different from zero. Term Model G df P deleted [12] [13] [14] [234] 11.15 8 >0.10 [12] [13] [234] 154 13 >0.25 [14] [12] [14] [234] 12.18 9 >0.25 [13] [13] [14] [234] 57.52 14 <0.01 [12] * [12] [234] 16.91 14 >0.25 [14] [14] [234] 58 98 11 <0.01 [12] • [1] [234] 70.38 16 <0.01 [12] • Final model = [12] [234] 497 disturbances at =£100 m more than at distances 101-200 m and 201-300 m, and were least reactive to disturbances at 201-300 m (Table 6). TABLE 6. — Weights associated with the distance/seal response interaction term [12] of the log-linear model that fits the data in Table 5. Relative mag- nitude of weight indicates importance of the variable. Sign of the weight in- dicates direction of effect ( + is more, — is less). Seal response (m) Yes No 0-100 101-200 201-300 '1.116 -0201 09ie> -1.116 0.201 20.915 'Seals were most reactive to disturbance 2Seals were least reactive to disturbance. After actual disturbances (type I), the number of seals that eventually rehauled was always lower than the original number. On KI, the average time it took seals to rehaul regardless of season, was 28 ± 20.8 min (range 5-100, n = 187). In 96 instances, no seals rehauled, primarily due to tidal height. During rising tides, they rehauled only 16.2% of the time (n = 37 disturbances), at low slack tide, 55.6% of the time (n = 124), and on falling tides 61.5% of the time (n = 26; x2 = 13.82, P < 0.001). Disturbances were of short duration and seals rehauled after the disturbance source had left the area, except for disturbance from commercial bait harvesters, who remained in the vicinity for entire low tide cycles. Bait harvesters likely prevented seals from hauling out at all (zero-seal distur- bance). During December, January, and February, we recorded the presence of the harvesters on 13 d. After being disturbed, seals did not return to the haul out site on eight of those days. They were disturbed briefly by the harvesters and then re- hauled on 3 d, and there was no change in seal numbers on 2 d. PWI apparently was not an important alterna- tive site when KI was disturbed. A weak correla- tion between seal numbers at PWI after they were disturbed from KI existed during the winter (r = - 0.42, n = 7) and summer (r = - 0.40, n = 152), but not during the breeding season (r = - 0.14, n = 123). In 45 of these instances, however, (winter 3, breeding 14, and summer 28), disturbances oc- curred simultaneously at KI and PWI, thereby precluding seal movement to PWI. During field observations, the movement from KI to PWI after disturbance was actually observed on 11 oc- casions. FISHERY BULLETIN: VOL. 82, NO. 3 DISCUSSION The population of harbor seals at Bolinas La- goon is much higher than previously recognized, and in contrast to seasonal peaks during the breeding season at other seal haul out sites (Fancher 1979; Johnson and Jeffries 19779; Loughlin 1978; and Allen and Huber 198310), the peak at Bolinas Lagoon occurred during summer after the pupping season. The peak at Bolinas may be caused in part by an influx of seals, possibly from San Francisco Bay, only 24 km south, or from Double Point, 10 km north, where numbers decline after the pupping season (Risebrough et al. 197811; Allen and Huber 1983 footnote 9). The summer increase also coincides with a marked increase in fish abundance in Bolinas Lagoon and Bolinas Bay; fish abundance and species diversity are greater in the lagoon from May to September than from November to February (J. Gustafson12). Scheffer and Sperry (1931), Spalding (1964), and Pitcher (1977 footnote 5), suggested that harbor seals are opportunistic, preying primarily upon small schooling fish. In a study by Brown and Mate (1983), peak abundance of seals in Netarts Bay, Oreg., also occurred in the fall and coincided with the seasonal abundance of chum salmon. Move- ment to Bolinas Lagoon at a time of high food availability may be a consequence of the seal's opportunistic feeding strategy. Time of day and tide were important factors that influenced daily haul out patterns of seals. The peak in numbers during early afternoon is consis- tent with studies on the Farallon Islands (Ainley et al. 1977 13) and in San Francisco Bay (Fancher 1979). Though seals were seen hauled out on KI at night on 10 occasions, the sharp drop in numbers during late afternoon suggests that diurnal haul- ing out is preferred. The diurnal pattern may also 9 Johnson, M. L., and S. J. Jeffries. 1977. Population evalu- ation of the harbor seal (Phoca vitulina richardi ) in the waters of the State of Washington. U.S. Dep. Commer., N.T.I.S. PB-270 376, 27 p. 10Allen, S. G., and H. R. Huber. 1983. Pinniped assessment in the Point Reyes/Farallon Islands National Marine Sanctuary, 1982-83. Annual Report to U.S. Department of Commerce, Sanctuary Programs Office, 64 p. "Risebrough, R. W, D. Alcorn, S. G. Allen, V. C. Alderlini, L. Booren, R. L. DeLong, L. E. Fancher, R. E. Jones, S. M. McGinnis, and T. T. Schmidt. 1978. Population biology of harbor seals in San Francisco Bay, California. U.S. Dep. Commer., N.T.I.S. PB81-107963, 67 p. 12J. Gustafson, Environmental Consultant, Resources and Ecology Projects, Mill Valley, CA 94941, pers. commun. August 1979. 13Ainley, D. G„ H. R. Huber, R. P Henderson, T J. Lewis, and S. H. Morrell. 1977. Studies of marine mammals at the Faral- lon Islands, California, 1975-76. U.S. Dep. Commer., N.T.I.S. PB-266 249, 32 p. 498 ALLEN ET AL.: EFFECT OF DISTURBANCE ON HARBOR SEALS be related to seal feeding habits as discussed by Antonelis and Fiscus (1980) and Spalding (1964) who noted that seals fed primarily in the late af- ternoon. The weak inverse correlation between tide level and seal numbers during the breeding season is likely related to the tendency of females with pups to haul out at irregular times to nurse. These patterns were interrupted by disturbance from boats, pedestrians, dogs, and aircraft. People in nonpower boats were the greatest source for disturbance possibly because they are more mobile than people in power boats or on foot. Dis- tance of disturbance, however, rather than type or season was the significant element at KI since at distances >100 m seals tended not to leave the hauling out site. The response of seals at distances >100 m may have been precipitated by the nature or unpredictability of the disturbance source. For example, a boat advancing directly toward the seals or lingering nearby caused flight more often than a boat moving by. The source of current disturbances is a small but stable resident and tourist human population; however, a variety of changes in the seal's behavior may be expected if disturbance levels increase. Both Paulbitsky (1975) and Woodhouse14 docu- mented a change from diurnal to nocturnal haul- ing out patterns in seals at Strawberry Spit, Tibu- ron, and at Atascadero State Beach, Morro Bay, Calif., which was believed to be a response to an increase in the local human populations. The re- sponse of seals to the prolonged activities of com- mercial bait harvesters on Bolinas Lagoon is indic- ative of the potential disruption of seal haul out patterns. Excessive disturbance may also lead to in- creased pup mortality. According to Kenyon (1972), 7 of 18 Hawaiian monk seals, Monachus schaunislandi, died before weaning on heavily disturbed pupping grounds on Midway Atoll, Hawaii. In contrast, for harbor seals at a relatively undisturbed pupping ground in British Columbia, Bigg ( 1969) estimated that pup mortality was only 12<7c . We do not know to what extent disturbance is affecting pup mortality rates at Bolinas Lagoon. In 1979, 3 of 12 pups were found dead; at least 1 of those 3 was killed by a dog. Site abandonment is a third possible response to increased disturbance. Newby (1971) attributed harbor seal abandonment of a site in Puget Sound 14C. Woodhouse, Santa Barbara Museum of Natural History, 2559 Puesta del Sol, Santa Barbara, CA 93105, pers. commun. May 1977. in part to increased disturbance from recreational boating. Kenyon (1972) postulated for the monk seal that site abandonment results in overall population losses because other traditional haul out sites probably cannot absorb the emigration. The same could apply to harbor seal populations in Marin County, if other sites are currently filled to capacity. ACKNOWLEDGMENTS This study was funded by the Marine Mammal Commission under Contract No. MMAC012, the Marin County Department of Parks and Recre- ation, and the Point Reyes Bird Observatory. Special thanks are extended to L. and D. Roush and D. and K. Beacock, who allowed us to place cameras in their homes. We are also grateful to J. Bausor, C. Hale, M. Rousch, B. Sorrie, L. Stenzel, and S. Tallen for help in field observations. K. and D. Reichard and R. Rains provided advice on camera maintenance and operation. R. E. Jones of Museum of Vertebrate Zoology, Berkeley, analyzed pup carcasses and offered valuable advice and as- sistance. LITERATURE CITED ANTONELIS, G. A., JR., AND C. H. FISCUS. 1980. The pinnipeds of the California current. CalCOFI Rep. XXL68-78. BIGG, M. A. 1969. The harbour seal in British Columbia. Fish. Res. Board Can. Bull. 172, 33 p. BISHOP, Y. M. M., S. E. FlENBERG, AND P. W. HOLLAND. 1975. Discrete multivariate analysis: theory and practice. MIT Press, Cambridge, Mass., 557 p. BRADLEY, J. V. 1968. Distribution-free statistical tests. Prentice-Hall, Englewood Cliffs, N.J., 388 p. BROWN, R. F, AND B. R. MATE. 1983. Abundance, movements, and feeding habits of har- bor seals, Phoca vitulina, at Netarts and Tillamook Bays, Oregon. Fish. Bull., U.S. 81:291-301. Carlisle, J. G., Jr., and J. A. Alpin. 1966. Sea lion census for 1965 including counts of other California pinnipeds. Calif. Fish Game 52:119-120. 1971. Sea lion census for 1970, including counts of other California pinnipeds. Calif. Fish Game 57:124-126. FANCHER, L. E. 1979. The distribution, population dynamics, and behav- ior of the harbor seal, Phoca vitulina richardu, in south San Francisco Bay, California. MS Thesis, California State University, Hayward, 109 p. FlENBERG, S. E. 1981. The analysis of cross-classified categorical data. 2d ed. MIT Press, Cambridge, Mass., 198 p. KENYON, K. W. 1972. Man versus the monk seal. J. Mammal. 53:687-696. 499 FISHERY BULLETIN: VOL. 82, NO. 3 LOUGHLIN, T. R. 1978. Harbor seals in and adjacent to Humboldt Bay, California. Calif. Fish Game 64:127-132. NEWBY, T. C. 1971. Distribution, population dynamics and ecology of the harbor seal of the southern Puget Sound, Seattle. MS Thesis, University of Puget Sound, Tacoma, Wash., 75 p. PAULBITSKY, P. A. 1975. The seals of Strawberry Spit. Pac. Discovery 28(4): 12-15. SCHEFFER, T. H., AND C. C. SPERRY. 1931. Food habits of the Pacific harbor seal, Phoca rich- ardii. J. Mammal. 12:214-226. SCHEFFER, V. B., AND J. W. SLIPP. 1944. The harbor seal in Washington State. Am. Midi. Nat. 32:373-416. SNEDECOR, G. W, AND W G. COCHRAN. 1967. Statistical methods. 6th ed. Iowa State Univer- sity Press, Ames, 593 p. Spalding, D. J. 1964. Comparative feeding habits of the fur seal, sea lion and harbour seal on the British Columbia coast. Fish. Res. Board Can. Bull. 146, 52 p. VENABLES, U. M., AND L. S. V. VENABLES. 1955. Observations on a breeding colony of the seal Phoca vitulina in Shetland. Proc. Zool. Soc. Lond. 125:521-532. 500 REPRODUCTION OF WEAKFISH, CYNOSCION REGALIS, IN THE NEW YORK BIGHT AND EVIDENCE FOR GEOGRAPHICALLY SPECIFIC LIFE HISTORY CHARACTERISTICS Gary R. Shepherd1 and Churchill B. Grimes2 ABSTRACT Reproduction characteristics for weakfish, Cynoscion regalis, in the New York Bight were examined. Spawning in 1980-81 occurred from May to early July with spawning time dependent on parental size. Maturity for both sexes occurred by age 1 but at a greater size in females. Annual fecundity estimates were compared with literature values for North Carolina weakfish and were found to be considerably lower at size, yet cumulative fecundities were nearly equivalent. The latitudinal variations in fecun- dity may be a behaviorally and environmentally induced phenomena, and influence the long-term population stability of weakfish. Weakfish, Cynoscion regalis, are a member of the family Sciaenidae and are a common inshore species occurring between Cape Cod, Mass., and southern Florida. The species undergoes a spring migration from offshore waters of Virginia and the Carolinas to appropriate estuarine spawning areas, then a return migration in late fall to over- wintering grounds (Nesbit 1954). The center of greatest abundance occurs within the Middle At- lantic Bight in quantities sufficient to support a recreational and commercial fishery. In 1979 commercial fishermen landed 13,000 metric tons (t) of weakfish and nearly 5,000 t were caught by recreational anglers (Wilk 1981). Abundance has, however, undergone some dramatic fluctuations over the last several decades. Commercial land- ings averaged 8,800 t from 1940 to 1949, then dropped to 2,915 t by 1950, and remained at these low levels until the mid-1970's (Wilk 1981). The exact cause of these variations remains a mystery, although speculations include overfishing, DDT- induced mortality, and environmentally induced recruitment failure (Massman 1963; Joseph 1972; Merriner 1976). To adequately assess the mecha- nisms controlling recruitment success or failure, we must first have a thorough understanding of the reproductive biology of weakfish. Merriner (1976) has examined reproduction of weakfish in North Carolina, and Daiber (1957) mentioned spawning behavior of weakfish in Delaware Bay, but the reproductive biology of 'Rutgers University, New Brunswick, N.J.; present address: Northeast Fisheries Center Woods Hole Laboratory, National Marine Fisheries Service, NOAA, Woods Hole, MA 02543. 2Rutgers University, New Brunswick, NJ 08903. weakfish in their northern range has never been fully investigated. Furthermore, there is reason to believe that reproduction may vary throughout the geographic range. Leggett and Carscadden (1978) have shown latitudinal variations in repro- duction and growth of American shad, Alosa sapidissima, and White and Chittenden (1977) have likewise shown geographic differences in another sciaenid, the Atlantic croaker, Micro- pogonias undulatus. Growth differences have al- ready been established for weakfish between the New York Bight and North Carolina (Perlmutter et al. 1956; Shepherd and Grimes 1983), so there is reason to suspect possible reproductive differ- ences. The purpose of the study was to investigate weakfish reproduction in the Middle Atlantic re- gion, to determine if any geographic variations exist, and to consider possible reasons for geo- graphically specific characteristics. METHODS AND MATERIALS Sample Collection Samples {n = 1,208) were collected during the National Marine Fisheries Service (NMFS) groundfish survey from 1980 to 1983 at stratified random stations north of Chesapeake Bay (Fig. 1). Fish were collected with a #41 Yankee trawl in spring and a #30 Yankee trawl in summer and fall at depths between 5 and 200 m (Grosslein 1969). NMFS samples were supplemented by 461 fish collected between May 1980 and June 1981 from commercial pound nets in Gardiners Bay, N.Y, (n = 61) and Sandy Hook Bay, N.J., (n = 115) and Manuscript accepted February 1984. FISHERY BULLETIN: VOL. 82, NO. 3, 1984. 501 FISHERY BULLETIN: VOL. 82, NO. 3 from pair and otter trawl fisheries in Delaware Bay (n = 285). At each NMFS station, weakfish catches were FIGURE 1. — Sampling locations for weakfish, Cynoscion regalis, in the New York Bight and North Carolina site of Merriner (1976). stratified into 10 cm size intervals and 10-15 fish sampled per interval. Total length to the nearest centimeter, sex, and maturity stage were recorded. Weakfish from Gardiners Bay were randomly sampled, total length measured to the nearest mil- limeter, sex, and maturity stage recorded, and gonads removed and weighed to the nearest 0.1 g. Sandy Hook and Delaware Bay samples were col- lected by random selection of 50-lb (22.5 kg) boxes in each market category available from the catch, and total length to the nearest millimeter and weight (whole and gutted) to the nearest gram recorded. Gonads and livers were also removed and weighed to the nearest 0.1 g. Gonads were preserved in modified Gilson's solution (Bagenal and Braum 1971) for several weeks, then removed, washed with distilled water and stored in 95% isopropyl alcohol. Maturity and Fecundity Methods The maturity stage of each sample was catego- rized as immature, developing, ripe (spawning), spent, or recovering as modified from Nikolsky (1963). The maturity stages were further sub- divided as mature or immature for calculation of length at 50% maturity using probit analysis (Finney 1971). Seasonality of reproduction was determined from changes in the gonad condition. A gonosoma- tic index (GSI) was calculated to show changes in gonad weight relative to somatic weight (gutted body weight). The index was computed as GSI = (gonad weight/somatic weight) x 100. To examine physiological changes associated with spawning, liver condition was assessed using a hepatosomatic index (HSI) computed by sub- stituting liver weight for gonad weight in the above relationship (Htun-Han 1978). The differ- ences in monthly mean HSI between sexes were analyzed statistically using a Wilcoxon test (Patz- ner and Adam 1981). The number of spawnings in the season were investigated by analyzing the seasonal frequency distribution of oocyte diameters (Hickling and Rutenberg 1936). In each of 15 samples collected between 5 May and 22 July 1980-81 in Delaware and Gardiners Bays, three subsamples were taken per ovary and about 500 oocyte diameters were randomly measured using an ocular micrometer. Fecundity estimates were determined from 28 fish macroscopically classified as developing, 502 SHEPHERD and GRIMES: REPRODUCTION OF WEAKFISH which were captured during May 1981 in Dela- ware and Gardiners Bays. An oocyte diameter of 0.20 mm was determined from diameter frequency distributions and the degree of yolk accumulation as the size between oogonia and developing ova, and was used as the lower size limit of ova in the fecundity estimates. Each sample was diluted with distilled water, stirred, then several aliquots removed from the solution to provide a density of about 10,000 ova in a 6 x 6 cm gridded petri dish. About 800-1,000 ova were counted from six ran- domly selected squares, then adjusted for a total subsample count. Two subsamples were counted per sample and three if the ova sample was from a large fish (>60 cm). The sample and subsample were oven dried at 40°C for a minimum of 24 h then weighed to the nearest 0.001 g on a Mettler3 balance. Total fecundity was calculated as Total fecundity = [number in subsample x (sample wt/subsample wt)] + number in subsample. Predictive equations of fecundity from length and weight were fit to a geometric mean (GM) func- tional regression (Ricker 1975) following log-log transformation. RESULTS Seasonality The changes in maturity stages during the year indicate spawning takes place from May to mid- July for weakfish in Delaware Bay and north to Long Island (Fig. 2). In May, all 122 mature fish examined of both sexes were in the developing or ripe stage of gonad development. The 50 fish that were inspected in June included 100% of the males and 70% of the females in a developing or ripe stage. Several of the ovaries examined in a June sample from Gardiners Bay had the appearance of being partially spent. The ovary was flaccid, slightly hemorrhaged and the lumen filled with fluid, but a few transparent ova were still visible. Spawning weakfish in Delaware Bay were cap- tured as late as 12 July, when 11% of the females and 76% of the males were classified as ripe. In the same month, 84% of the females were in spent condition. By August, all of the fish examined at all locations were in postspawning condition. MAY JUNE JULY AUG SEPT OCT NOV DEC FEMALE 30 17 2U U MAY JUNE JULY AUG SEPT OCT NOV DEC RECOVERING [ IMMATURE ^DEVELOPING | | RIPE SPENT 3Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. FIGURE 2. — Changes in visual maturity stage of weakfish, Cynoscion regalis, in 1980-81 for the New York Bight. TV indi- cated for each month. Weakfish from Sandy Hook Bay were not available until June and all the samples examined from this area were in postspawning condition, either spent or recovering. This seasonality of spawning at each of the sam- ple locations was suggested by seasonal changes in the GSI. Mean GSI values (percent gonad weight relative to somatic weight) ± 95% C.I. (Confidence Interval) of all females rose from 5.75 ± 1.29 in May to a peak value of 6.04 ± 1.43 in June, then declined to 1.76 ± 0.40 by July (Fig. 3). Male GSI were at a yearly high of 4.88 ± 1.45 in May but declined to only 2.51 ± 0.85 by July. The mean GSI values for females were consistently higher in all months except July. Gonad size reached the lowest levels for both sexes in August with mean GSI values of 0.71 ± 0.07 and 0.19 ± 0.02 for females and males, respectively. The GSI values remained low until the last samples were collected in November. Specific spawning time, as reflected in GSI, was dependent on the size of the individual fish (Fig. 4). 503 FISHERY BULLETIN: VOL. 82, NO. 3 aug Month FIGURE 3. — Mean monthly gonosomatic indices ±95% C.I. of weakfish, Cynoscion regalis, for 1980 in the New York Bight. N values at each point. Mean index values and lengths ±95% C.I. for de- veloping fish in Delaware and Gardiners Bays dur- ing May was 2.48 ± 0.67 at a mean length of 551 ± 47.0 mm for females and 1.13 ± 0.85 at a length of 391 ± 98.9 mm for males. Spawning fish captured during May had GSI of 8.65 ± 1.64 at mean length of 702 ± 25.2 mm for females and 6.49 ± 1.58 at mean length of 619 ± 33.1 mm for males. In June, spawning fish had mean GSI values of 6.72 ± 1.83 with mean length of 544 ±95.6 mm for females and 4.54 ± 1.04 with mean length of 570 ± 56.3 mm for males. These index variations confirmed observations of earlier spawning by large fish. Seasonal variation in mean HSI of Delaware and Sandy Hook Bay samples indicates that changes in liver weight occurred at the time of spawning and prior to fall migration (Fig. 5). Mean HSI values ± 95% C.I. for females decreased from a high in May of 2.80 ± 0.35 to a low in September of 0.87 ± 0.30. The index values for males followed the same pattern declining from 1.89 ± 0.27 in May to 0.96 ± 0.18 in September, but the maximum values occurred in November with a mean HSI of 2.57 ± 0.29. The indices were tested for differences between sexes and the differences were found to be significant in May and June (P < 0.001) with the values for females being greater (Table 1). No significant differences were found in a comparison of the July through November samples. Frequency distributions of oocyte diameters from 15 fish, ranging from 55 to 81 cm and collected TABLE l.— Wilcoxon (Mann-Whitney) test comparing hepatosomatic indices between sexes of weakfish, Cynoscion regalis, from the New York Bight for June through November 1980 and May 1981. S = Wilcoxon test statistic, Z = critical value. Probability s N Z level 1980 June 6220 50 5.081 0.0001 — July 194.0 34 -1.767 0.0865 August 106.0 28 0.265 0.7928 September 274.0 37 -0904 0.3658 October 252.0 35 0.017 0.9867 November 2835 40 -1.215 0.2318 1981 May 332 0 48 -3.782 0.004— ■ highly significant differences. between 6 May and 22 July show one seasonally progressing mode of developing ova (Fig. 6). The position of the mode varied according to the devel- opment stage of the individual ovary. A sample from 5 May had a mode between 0.02 and 0.45 mm and was skewed toward a prominent peak at 0.12 mm. In the 13 May sample, a second mode ap- peared around 0.35 mm and this mode increased to a size of 0.63 mm in the 20 June sample. The 22 July sample contained only oogonia. Maximum ova diameter observed was 0.95 mm from a sample on 26 May. That ovum was filled with fluid be- tween the yolk and chorion, indicative of an ova immediately prior to release (Bagenal 1967). This corresponds to the size of weakfish eggs, 0.870- 0.975 mm, which have been identified in the water column (Harmic 1958). Maturity, Sex Ratio, and Fecundity Length ±95% C.I. at which 50% of the total sampled population reached maturity was similar for both sexes. Females attained 50% maturity at 25.6 ± 1.2 cm, while males were slightly lower at 25.1 ± 1.1 cm (Table 2). The corresponding age at maturity for both sexes was 1 yr. The smallest mature male and female was 20 cm. The maximum size of immature weakfish was 40 cm for females and 33 cm for males. The overall population sex ratio approached equality (Table 3). The sex ratio of the population, divided into 5 cm length intervals, was 48:52 females to males and was not significantly differ- ent from 50:50 as determined from a chi-square analysis (x2 = 1.81, x2o.o5 = 3.84, n = 1,669). Sex ratio at size data did reveal significant differences from 50:50 for several length intervals. At 40 cm, the ratio was 52:48 female to male but increased to 504 SHEPHERD and GRIMES: REPRODUCTION OF WEAKFISH Male X "O c 14 12 10 8 6 4 2 14 12 10 8 6 4 2 2 14 a § 12 in o § 10 o 8 6 4 2 14 12 10 8 6 4 2 May x » x r X X X X* June *x July August -% tr- 10 20 30 40 50 Length (cm) 60 70 16 14 May Female x ** « 12 10 X X X X X 8 X * 6 X " • • XX X 4 2 14 June 12 X 10 8 X X X X 6 X Index to * X •• *" X X * • o!4 E Six o §10 July 8 6 4 2 X X X • • • • • • • • • 14 August 12 10 8 6 4 2 • • «■■• • 10 20 30 40 50 Length (cm) 60 70 FIGURE 4. — Monthly gonosomatic indices of individual weakfish, Cynoscion regalis, by sex for 1980. Spawning fish indicated by x. 505 FISHERY BULLETIN: VOL. 82, NO. 3 30 ti & 20 o 16 E I 10 x TABLE 2. — Length at maturity for weakfish, Cynoscion regalis, from 1980 to 1983. may jun jlil aug Month sep oct FIGURE 5. — Mean monthly hepatosomatic indices ±95% C.I. of weakfish, Cynoscion regalis, for 1980 with N indicated at each point. May 5 ":833 Developing .25 .35 .45 .55 .65 Ova Diameter (mm) .75 .85 FIGURE 6. — Monthly frequency of oocyte diameters for weakfish, Cynoscion regalis, in the New York Bight for 1980-81. 60.5% (x2 = 5.25, xVos = 3.84, n = 119) and 63.8% (X2 = 7.19, x2o.oi = 6.64, n = 94) males at 45 and 55 cm, respectively. The sex ratio of combined length intervals above 55 cm had a ratio of 58.5:41.5 Female Male (cm) N % mature N % mature 18 15 0.0 16 0.0 19 7 0.0 11 0.0 20 8 12.5 11 9.1 21 11 18.2 5 40.0 22 8 37.5 9 33.3 23 13 538 16 31.3 24 11 45.5 12 50.0 25 13 46.2 12 75.0 26 14 71.4 15 86.7 27 17 88.2 14 71.4 28 10 80.0 20 65.0 29 17 82.4 22 63.6 30 19 100.0 17 82.4 31 18 88.9 14 85.7 32 16 938 20 85.0 33 10 90.0 17 94.1 34 14 85.7 15 100.0 35 7 85.7 3 100.0 36 10 80.0 8 1000 37 9 77.8 12 100.0 38 15 100.0 39 14 85.7 40 13 92.3 41 4 100.0 42 10 100.0 43 10 100.0 Total 313 269 Size at 50% mature 25.6 25.1 95% Confidence Interval 24.4-268 24.0-26.2 TABLE 3. — Sex ratios (female:male) of weakfish, Cynoscion regalis , in the New York Bight. Length1 Ratio 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 Total 27:40 2 52 72:73 0.01 149:174 1.93 148:157 0.27 72:73 0.01 71:65 0.26 42:72 5.25 35:40 0.33 34:60 7 19 36:35 0.01 35:27 2.00 31:28 0 .15 28:17 2.69 18:1 15.21 4:0 4.0' 807:862 1.81 'Midpoint of length interval (12.5 to 17.5, etc.; * = significant difference P < 0.05. ** = significant difference P < 0 01. *** = significant difference P < 0.001. which was significantly dominated by females (x2 = 7.45, x2o.oi = 6.64, n = 260). Regression models indicated that length and weight were equally predictive of fecundity. The relationships between total length or gutted weight, and fecundity were best described by a power curve. The GM functional regression (Ricker 1975) of the log^ transformed data ± stan- dard error of the regression were 506 SHEPHERD and GRIMES: REPRODUCTION OF WEAKFISH 2250 1750 o 1250 -o 750 3 250 350 450 550 Length (mm) 650 750 FIGURE 7. — Fecundity as a function of length for weakfish, Cynoscion regalis, in New York Bight for 1981. In fecundity = -16.322 + 4.659 In total length (mm) ±0.368 r2 = 0.835 n = 28 Figure 7 In fecundity = 1.975 + 1.542 In gutted weight (g) ± 0.364 r2 = 0.839 n = 28. Fecundities were also estimated for six weakfish collected in June from Gardiners Bay which ap- peared to be spent, although they had ova remain- ing in the gonad. These ova were 0.55 to 0.65 mm which were similar in size to those present in a ripe ovary. All six samples were significantly dif- ferent (P < 0.001) than expected values based on a Students t~test (Table 4). These June samples were 60-75% lower than the predicted fecundities at length. DISCUSSION Weakfish spawning in northern estuaries is a seasonal event which occurs following the spring inshore migration. (Welsh and Breder 1923; Hil- debrand and Schroeder 1928). Our study found the spawning period for weakfish in Delaware Bay and Gardiners Bay, as determined from maturity stages, to be from May to mid-July Ichthyoplank- ton surveys have found weakfish eggs and larvae present in New York Bight from May to July (Col- ton et al. 1979). The spawning period was further substantiated by changes in GSI. The mean GSI values reached a maximum in May for males and in June for females, then both declined to mini- TABLE 4. — Difference between expected fecundity (based on In fee = -16.322 + 4.659 In TL) and ob- served fecundity for 6 weakfish, Cynoscion regalis, collected 27 June 1981, in Gardiners Bay, N.Y. Total length (mm) Fecundity estimate % Expected Observed Difference 590 613 615 704 737 644 650,171 777,718 789,675 1,487,262 1,843,210 979,893 196,891 93,579 105,919 438,909 650,600 313,054 69.7 88.0 86.6 70.5 64.7 68.1 x = 74.6 mum values by August. The gonad weights were maintained at this low level until the fish disap- peared from the coast in late November. The spawning period of weakfish in North Carolina is somewhat longer, extending from March to Sep- tember with the peak period from April to June (Merriner 1976). The duration of the ripe maturity stage in Delaware and Gardiners Bays was greater for males, with ripe males captured as late as 22 July. The protracted spawning season of males was also evident in the GSI values. July was the only month in which male indices were greater than those of females, indicating that the gonad weight per unit body weight was larger in males. In other months, female GSI values were as much as 2.8% greater than males. Sex differences were also evident in physiologi- cal changes associated with spawning. HSI values were near maximum in May, during spawning season, and as expected higher in females. Devel- 507 FISHERY BULLETIN: VOL. 82, NO. 3 opment of ovaries demands higher energetic costs than testicular development (Nikolsky 1963); therefore, the necessary energy reserves in the liver would be proportionately larger in females (Timashova 1982). After spawning, the sexual dimorphism disappears and HSI values for both sexes increase similarly prior to the fall migration. The peak spawning period in the New York Bight estuaries varied with different size weakfish. Generally, among migratory fishes, the larger individuals will return to an estuary prior to their smaller counterparts (Briggs 1955; Nikolsky 1963). Weakfish exhibit a similar behav- ior, as the largest individuals or "tiderunners" enter the bays in early May and spawn by mid- May, whereas the smaller weakfish arrive later and reach peak spawning during June. The GSI values for fish > 60 cm generally decline from May to June, while smaller fish increase to maximum values in June. This differential spawning based on parental size results in two spawning peaks and the subsequent appearance of bimodal length- frequency data for juvenile weakfish (Daiber 1954; Thomas 1971; Shepherd and Grimes 1983). Sex Ratio, Maturity, and Fecundity The overall sex ratio of the population is not different from 1:1, but one sex was dominant at certain length intervals. We believe the deviations from a 1:1 ratio at various lengths were attribut- able to differential growth between sexes ( Wenner 1972). Female growth begins to exceed male growth at about 45-55 cm (Shepherd and Grimes 1983), at which point the sex ratio becomes domi- nated by males. Females grew beyond that size interval faster and occupied the majority of the 60-85 cm length intervals. Although males have growth potential equal to females, the numbers attaining maximum size were greater for females. In northern waters, the size at which weakfish attained 50% population maturity was similar for both sexes. Females matured by 25.6 cm at age 1 while males attained maturity at 25.1 cm, also at age 1. Apparently, greater differences exist be- tween northern weakfish (Delaware Bay and north) and southern weakfish (North Carolina). Although ages were similar, southern females spawned at 23 cm and males by 18 cm (Merriner 1973). Estimates of fecundity for New York Bight weakfish differ from estimates for southern weakfish. Weakfish in North Carolina did not reach sizes much beyond 45 cm but had fecun- dities, relative to length, several times higher than northern weakfish (Merriner 1976) (Table 5). For example, at 50 cm TL female weakfish from New York Bight produced 306,159 ova, while the fecundity of southern fish of the same size was 2,051,080 ova. In spite of these large differences, lifetime fecundity would be approximately equal. Southern weakfish can potentially reproduce until age 5 and produce about 9,913,080 ova (based on the equation, fecundity = 0.152 TL? 6418, from Mer- riner 1976), whereas northern weakfish reproduc- ing for 10 yr have nearly equivalent total ova pro- duction of 10,008,167. Batch spawning, involving two distinct groups of ova, was found for weakfish in North Carolina (Merriner 1976). In samples examined in 1980-81 from Delaware and Gardiners Bays, multiple spawns were not evident. The ova diameter fre- quencies of developing and ripe ovaries contained two modes, one consisting of reserve oocytes and another of developing ova. The developing ova had a wide size range ( — 0.3 mm) and may have been released during consecutive spawning events, but did not constitute separate batches within an ov- ary. Furthermore, all ova produced annually by weakfish may not be released during spawning. A study of Delaware Bay weakfish in 1954 proposed batch spawning, but spent ovaries were not examined to determine if all ova were released (Daiber 1954). Ovaries classified as spent which we examined still contained 25-40% of the ova expected for a fish of that size. These results suggest fertility may be 60-75% of estimated po- tential fecundity. Foucher and Beamish (1980) re- ported similar conclusions from studies of Pacific hake, Merluccius productus. We did not examine TABLE 5. — Comparison of fecundity for weakfish, Cynoscion regalis, between Cape Hatteras (Merriner 1976) and New York Bight. Fecundity comparisons Cape Hatteras New York Bight Age SL Fecundity TL Fecundity 0 159 149.429 1 225 391 ,688 203 4,593 2 286 762,258 323 39,978 3 357 1,410,550 479 250.685 4 421 2,229,220 578 601,544 5 562 4,969,940 638 953,027 6 677 1,256,487 7 701 1,477,902 8 728 1,762,445 9 758 2,127,278 10 763 2,193,448 Cumulative fecundity 9,913.085 10,667,387 508 SHEPHERD and GRIMES: REPRODUCTION OF WEAKFISH enough spent weakfish ovaries to determine if this phenomenon was consistent from year to year. The variable reproductive and age and growth characteristics for weakfish in different geo- graphic areas suggest specific physiological re- sponses to different environmental conditions. North Carolina weakfish had smaller sizes at maturity, smaller length at corresponding age after age 1, reduced longevity and maximum size, and higher relative fecundity than New York Bight weakfish. However, the lifetime reproduc- tive potential was nearly equal for both groups. These life history characteristics for weakfish are similar to clinal variations between Labrador and Florida described for American shad (Leggett and Carscadden 1978), differences in reproductive characteristics between Atlantic herring, Clupea harengus harengus, in the Norway and the Baltic Sea (Schopka 1971), and north to south variations in North American populations of Pacific herring, Clupea harengus pallasi (Paulson and Smith 1977). The different reproductive strategies in weakfish may result from varying environmental demands. When weakfish spawning occurs in northern estuaries, water temperatures are un- predictable and subject to sudden drops which are potentially lethal to eggs and larvae (Harmic 1958). Table 6 shows minimum estuarine temper- TABLE 6. — Surface water temperatures (°C) for April-July in Plum Island, N.Y.; Cape May, N.J.; Gloucester Point, Va; and Beaufort, N.C.1 April May June July Plum Island. N.Y X 5.6 97 14.6 18.8 Max 9.0 18.0 200 24.0 Min 20 4.0 9.0 150 •*max 7.7 13.4 17.7 21.4 xmin 33 64 11.4 16.3 Cape May. N.J. X 10.2 14.5 20.0 22.6 Max 17.0 210 25.0 28.0 Mm 60 9.0 12.0 18.0 *max 14.6 18.9 238 25.6 xmin 7.0 108 157 199 Gloucester Point Va. X 13.1 19.1 23.8 26.5 Max 21 0 260 30.0 32.0 Mm 7.0 13.0 17.0 21 0 'max 17.1 224 27.0 29.0 xmin 96 156 20.4 24.1 Beaufort. N C X 172 21 4 25 2 274 Max 21.0 260 320 30.0 Mm 14.0 140 18.0 24.0 *max 20.3 24.2 283 29.2 xmin 14.3 17.5 21.9 25.8 'National Ocean Survey 1972 Surface water temperature and density; Atlantic coast, North and South America. 4th ed NOS Publ .31-1, p. 1-109. atures during the spawning season in northern waters may drop below the temperature limits of 12°-16°C necessary for successful hatching (Har- mic 1958). The higher probability of prereproduc- tive mortality of progeny in northern estuaries results in a "bet-hedging" strategy in which fewer eggs are produced each year, but the possible number of annual spawnings are increased (Stearns 1976; Giesel 1976), thus maximizing po- tential contributions to the gene pool throughout a fish's lifespan. In contrast, southern weakfish spawn in a more predictable estuarine environ- ment (Table 4) and, consequently, there is less chance of environmentally induced prereproduc- tive mortality. However, southern fish are faced with greater postreproductive mortality (longev- ity observed by Merriner (1973) was 5 yr). Greatest reproductive success in this situation requires maximizing annual gamete production in the few years possible. In addition, weakfish migrating to northern estuaries (Nesbit 1954) may utilize energy reserves otherwise available for gonad growth, whereas southern fish having less dis- tance to travel may reallocate energy towards re- production. Consequences of the area specific reproductive characteristics may be a reduced population sta- bility for weakfish in the northern end of the range. Apparently, northern fish have a strategy to cope with potentially higher egg and larval mortality by spreading reproduction over more years and reducing annual fecundity, i.e., a "bet- hedging" strategy (Stearns 1976). Therefore, to fulfill their reproductive potential they must avoid premature adult mortality. If adult mortality (natural and fishing) in weakfish becomes exces- sive, the larger, most fecund individuals will be lost or reduced, thus shifting the burden of spawn- ing to the smaller, less fecund fish. When the value of b in the fecundity equation F = aTLb is >3, as in New York Bight weakfish, then truncation of the size/age structure in a spawning population will also result in a reduction of population fecundity (Hempel 1979). Concurrent high adult mortality and high prereproductive mortality could contrib- ute to a decline in population abundance. The large fluctuations which have occurred in weakfish populations over the last several decades (Wilk 1979) may be due in part to these cir- cumstances (i.e., high adult and prereproductive mortality). Although the correlation between population fecundities and recruitment is not usually strong for marine fishes (Cushing 1977), a decrease in population fecundity may eventually 509 FISHERY BULLETIN: VOL. 82, NO. 3 reduce any buffer that weakfish have of with- standing natural fluctuations in egg, larval, and juvenile survival. Therefore, if management prac- tices are to effectively regulate the weakfish re- sources, geographic variations in reproductive po- tential should be considered. ACKNOWLEDGMENTS The authors would like to thank Ken Able and J. Richard Trout for their helpful guidance. Dan Dzenkowski and crew of the Dan Dee, Greenport, Long Island, were more than generous in allowing us to sample their catches, and personnel of NMFS, Woods Hole, Mass., especially Ambrose Jearld, kindly provided samples from the ground- fish surveys. We would also like to thank Steve Turner, Chuck Idelberger, and Susan Toogood for their assistance and Hannah Goodale for her as- sistance in preparing the manuscript. Funding for this project was provided by NMFS (contract No. NA-79-FAC-00041) and the New Jersey Agricul- tural Experiment Station (NJAES publication No. D-12409-14-83). LITERATURE CITED BAGENAL, t. b. 1967. A short review of fish fecundity. In S. D. Gerking (editor), The biological basis of freshwater fish production, p. 89-111. John- Wiley & Sons, N.Y. BAGENAL, T. B., AND E. BRAUM. 1971. Eggs and early life history. In W. E. Ricker (editor), Methods for assessment offish production in fresh waters, p. 159-181. Blackwell Sci. Publ., Oxford, Engl. BRIGGS, J. C. 1955. Behavior pattern in migratory fishes. Science (Wash., D.C.) 122:240. Colton, J. B., Jr., W. G. Smith, A. W. Kendall, Jr., R L. BERRIEN, AND M. R FAHAY. 1979. Principal spawning areas and times of marine fishes, Cape Sable to Cape Hatteras. Fish. Bull., U.S. 76:911- 915. CUSHING, D. H. 1974. The natural regulation offish productions. In F. R. Harden-Jones (editor), Sea fisheries research, p. 399- 412. John- Wiley and Sons, Lond. DAIBER, F C. 1954. Fisheries research program. Mar. Lab. Dep. Biol. Sci., Univ. Del. Biennial Rep. 1958 and 1954, Publ. 2:50-64. 1957. Sea trout research. Estuarine Bull. 2(5):l-6. FINNEY, D. J. 1971. Probit analysis. Cambridge Univ. Press, Cam- bridge, Engl., 333 p. FOUCHER, R. P, AND R. J. BEAMISH. 1980. Production of nonviable oocytes by Pacific hake (Merluccius productus). Can. J. Fish. Aquat. Sci. 37:41- 48. GIESEL, J. T. 1976. Reproductive strategies as adaptations to life in temporarily heterogeneous environments. Ann. Rev. Ecol. Syst. 7:57-79. GROSSLEIN, M. D. 1969. Groundfish survey programs of BCF Woods Hole. Commer. Fish. Rev. 31(8-9):22-30. HARMIC, J. L. 1958. Some aspects of the development and the ecology of the pelagic phase of the grey squeteague, Cynoscion re- galis (Bloch and Schneider), in the Delaware es- tuary. Ph.D. Thesis, Univ. Delaware, Newark, 168 p. HEMPEL, G. 1979. Early life history of marine fish: The egg stage. Univ. Washington Press, Seattle, 70 p. HICKLING, C. F, AND E. RUTENBERG. 1936. The ovary as an indicator of the spawning period in fishes. J. Mar. Biol. Assoc. U.K. XXI:311-316. HILDEBRAND, S. F, AND W. C. SCHROEDER. 1928. Fishes of Chesapeake Bay. U.S. Bur. Fish. Bull. 43(l):l-366. HTUN-HAN, M. 1978. The reproductive biology of the dab, Limanda limanda (L.) in the North Sea: gonosomatic index, hepatosomatic index and condition factor. J. Fish Biol. 13:369-378. JOSEPH, E. B. 1972. The status of the sciaenid stocks of the Middle Atlan- tic coast. Chesapeake Sci. 13:87-100. LEGGETT, W. C, AND J. E. CARSCADDEN. 1978. Latitudinal variation in reproductive characteristics of American shad (Alosa sapidissima): evidence for popu- lation specific life history strategies in fish. J. Fish. Res. Board Can. 35:1469-1478. MASSMAN, W. H. 1963. Age and size composition of weakfish, Cynoscion re- galis, from pound nets in Chesapeake Bay, Virginia 1954- 1958. Chesapeake Sci. 13:87-100. MERRINER, J. V 1973. Assessment of the weakfish resource, a suggested management plan, and aspects of life history in North Carolina. Ph.D. Thesis, North Carolina State Univ., Raleigh, 12 p. 1976. Aspects of the reproductive biology of the weakfish, Cynoscion regalis (Sciaenidae), in North Carolina. Fish. Bull., U.S. 74:18-26. NESBIT, R. A. 1954. Weakfish migration in relation to its conserva- tion. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 115, 81 p. NIKOLSKY, G. V. 1963. The ecology of fishes. Academic Press, N.Y, 352 p. PATZNER, R. A., AND H. ADAM. 1981. Changes in weight of the liver and the relationship to reproduction in the hagfish, Myxine glutinosa (Cyclo- stomata). J. Mar. Biol. Assoc. U.K. 61:461-464. PAULSON, A. C, AND R. L. SMITH. 1977. Latitudinal variation of Pacific herring fecundi- ty. Trans. Am. Fish. Soc. 106:244-247. Perlmutter, a., W. S. Miller, and J. C. Poole. 1956. The weakfish (Cynoscion regalis) in New York waters. N.Y. Fish Game J. 3:1-43. RICKER, W. E. 1975. Computation and interpretation of biological statis- tics offish populations. Bull. Fish. Res. Board Can. 191, p. 1-382. 510 SHEPHERD and GRIMES: REPRODUCTION OF WEAKFISH SCHOPKA, S. A. 1971. Vegeleichende Untersuchungen sur Fortpflanz- ungsrate bei Herings- und Kabeljaupopulationen (Clupea harengus L. und Gadus morhua L.). Ber. Dtsch. Wiss. Komm. Meeresforsch. 22:31-79. |Also in Hempel, G. 1979. Early life history of marine fish: the egg stage. Univ. Washington Press, Seattle, 70 p.| SHEPHERD, G., AND C. B. GRIMES. 1983. Geographic and historic variations in growth of weakfish, Cynoscion regalis, in the Middle Atlantic Bight. Fish. Bull., U.S. 81:803-813. STEARNS, S. C. 1976. Life-history tactics: A review of the ideas. Q. Rev. Biol. 51:3-47. Thomas, d. l. 1971. The early life history and ecology of six species of drum (Sciaenidae) in the lower Delaware River, a brackish tidal estuary. Ichthyol. Assoc. Bull. 3, 247 p. TIMASHOVA, L. V. 1982. Seasonal changes in the structure of the liver of the plaice, Pleuronectes platessa. J. Ichthyol. 21:145-151. WELSH, W W, AND C. M. BREDER, JR. 1923. Contribution to the life histories of Sciaenidae of the eastern United States coast. Bull. U.S. Bur. Fish. 39:141-201. WENNER, A. M. 1972. Sex ratio as a function of size in marine Crustacea. Am. Nat. 106:321-350. WHITE, M. L., AND M. E. CHITTENDEN, JR. 1977. Age determination, reproduction, and population dynamics of the Atlantic croaker, Micropogonias un- dulatus. Fish. Bull., U.S. 75:109-123. WILK, S. J. 1979. Biological and fisheries data of weakfish, Cynoscion regalis (Block and Schneider). Northeast Fish. Cent., Sandy Hook Lab., NOAA Tech. Ser. Rep. 21. 1981. A review of the fisheries for Atlantic croaker, spot, and weakfish, 1940-1979. In H. Clepper (editor), Marine Recreational Fisheries Symposium VI, p. 15-27. Sportfishing Inst., Wash., D.C. 511 FIELD AND LABORATORY OBSERVATIONS ON DIURNAL SWIM BLADDER INFLATION-DEFLATION IN LARVAE OF GULF MENHADEN, BREVOORTIA PATRONUS D. E. Hoss1 AND G. Phonlor2 ABSTRACT Diurnal swim bladder inflation-deflation in gulf menhaden larvae was studied at sea and in the laboratory. At sea, the larvae filled their swim bladders at night and deflated them during the day. Laboratory experiments in which the larvae were either prevented or allowed to reach the air-water interface demonstrated that the larvae fill their swim bladder each night by swallowing air. These results agree with the findings of other investigators and suggest that diurnal swim bladder inflation may be a common characteristic in the late stage larvae of clupeoid species. The swim bladder in fishes has been assigned var- ious functions, the most widespread being the reg- ulation of buoyancy. Recent work on clupeoid species, however, has shown that the function of the swim bladder changes with ontogeny In adults of at least two clupeoids — Atlantic herring, Clupea harengus, and Atlantic menhaden, Bre- voortia tyrannus — the swim bladder is thought to serve as a reserve of gas for the adjustment of hydrostatic pressure in the gas-filled bulla, allow- ing the bulla membrane to maintain acoustic sensitivity independently of depth. The swim bladder's role as a buoyancy regulating organ is secondary (Blaxter and Hunter 1982). During the late larval stages of some clupeoids, however, buoyancy provided by an inflated swim bladder may have an important function. Hunter and San- chez (1976), working with the pelagic larvae of the northern anchovy, Engraulis mordax, proposed that an observed diurnal inflation and deflation of the swim bladder by larvae is an energy-sparing mechanism. In this case, one function of the in- flated swim bladder is to provide buoyancy that allows the larvae to "rest" during the night when they are unable to see to feed. This diurnal infla- tion and deflation of the swim bladder has also been reported for other larval clupeoids by Uotani (1973). The objective of this study was to determine if a diurnal swim bladder inflation-deflation rhythm exists in larval gulf menhaden, Brevoortia patro- 1 Southeast Fisheries Center Beaufort Laboratory, National Marine Fisheries Service, NOAA, Beaufort, NC 28516-9722. 2Fundacao Universidade do Rio Grande, 96.200 Rio Grande, RS, Brazil. nus, under natural conditions and, if so, to evalu- ate the mechanism of inflation in the laboratory. The swim bladder in gulf menhaden is similar to that described for Atlantic menhaden by Hoss and Blaxter (1981). In the Atlantic species the anlage of the swim bladder is present at 10 mm standard length (SL), and the pro-otic bullae first appear at 12.5 mm and may contain gas soon after. The swim bladder first contains bubbles of gas in 13 mm SL larvae, and the lateral line first appears in larvae of about 17 mm. In the fully developed system, narrow ducts connect the swim bladder to the gas-filled bullae which are close to the labyrinth of the inner ear. The bullae-swim bladder system is in turn connected to the extensive lateral line on the head of adult fish through a membrane in the skull. As in other clupeoid species (Blaxter and Hunter 1982), menhaden apparently swallow air to initially fill both the bullae and the swim blad- der. As there is no evidence for gas secretion in menhaden, it is also assumed that they replace lost gas by regularly swallowing air into the alimentary canal and then by transferring it to the swim bladder through the pneumatic duct. The swim bladder is deflated by diffusion and by reversing gas movement in the above pathway. Unlike some clupeoids, menhaden have no direct connection between the swim bladder and the anal opening (Tracy 1920). METHODS At Sea Gulf menhaden larvae were obtained in the northern Gulf of Mexico off Southwest Pass, La., Manuscript accepted January 1984. FISHERY BULLETIN: VOL. 82, NO. 3, 1984. 513 FISHERY BULLETIN: VOL. 82, NO. 3 at the mouth of the Mississippi River in Decem- ber 1981 (RV Oregon II cruise 123) and December 1982 (RV Oregon II cruise 131). On cruise 123, larvae were collected by oblique or surface tows of either a multiple-opening closing net and envi- ronmental sensing system (MOCNESS) ( Wiebe et al. 1976), a neuston net, or an opening-closing paired BFN-1 net3. The sampling scheme was modified on cruise 131 in that only the MOCNESS system was used and duplicate sam- ples were taken at fixed depths (1, 8, and 20 m) during a 24-h period of time. On board the ship, menhaden larvae were removed from the sam- ples, measured to the nearest 0.1 mm SL, and examined for the presence of gas in the swim bladder. Gas in the swim bladder was easily observed before pigmentation of larvae (Fig. 1). In inflated bladders, a light-refractive bubble is ob- vious while a deflated bladder appears under the microscope as a flattened sac (Doroshev et al. 1981). Maximum width and length of the swim bladder (with or without gas) was measured to the nearest 0.02 mm, and volume was calculated by the equation for a prolate spheroid, V = 4/3 tt ab2, where a is half the maximum bladder length and b is half the maximum bladder width (Hunter and Sanchez 1976). Approximate changes in volume of the swim bladder due to increased pressure at increased depth of capture were cal- P V culated from Boyle's law — - = — - where P is P2 v, pressure and V is volume, and temperature is assumed to be constant. After being measured, larvae were preserved in 5% Formalin4. In the Laboratory Experiments to determine if larvae filled the swim bladder by gulping in air at the water sur- face utilized larvae hatched from eggs in the laboratory (Hettler 1983). Larvae were reared on the rotifer Brachionus plicatilis, also cultured in the laboratory. As larvae grew older, their diet was supplemented with newly hatched Artemia naup- lii. Before being used, larvae were held in 80 1 tanks at a water temperature of 20° C, salinity of ca. 25%o, and a 12 h light-12 h dark photoperiod without a dawn or dusk transition. Three hours before the start of the experiment, 15-20 larvae were transferred from the rearing tanks to each of eight 10 1 tanks, and 10 larvae were measured and observed for the presence of gas in the swim bladder. A 500 /xm screen was then placed below the water surface (Fig. 2) in four of the experimental tanks to prevent access of the larvae to the air-water interface. In the other four tanks, larvae had access to the air-water interface. During the experiment, the 12-h-light photoperiod was continued. Larvae sampled at ca. 1800, 2100, 0630, 0900, and 1230 h were measured and ob- served for gas in the swim bladder. 3Tarez and Co., 8460 S.W. Street, Miami, FL 33143. "•Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Bulla Gas Bubble ln Gut Swimbladder FIGURE l.— Larval gulf menhaden, Brevoortia patronus. Inflated swim bladder, gas-filled bullae, and gas bubble in foregut are indicated by arrows. 514 HOSS and PHONLOR: SWIM BLADDER OF GULF MENHADEN LARVAE , 10 L CIRCULAR TANKS \ / /—AIR-WATER INTERFACE— y \ '//A '// h^ri WATER BATH CONTROL 1 777 EXPERIMENTAL /7 0.05) from daylight samples. Gulf menhaden larvae showed a diurnal pattern of depth distribution that seemed to contrast the pattern of swim bladder inflation (Table 2). Night sampling (1700-0600) indicated that larvae were present from the surface down to at least 20 m, the deepest samples taken. Maximum water depth at r^ ss rDay— r or LU Q Q < _l i< WI LU t O 5 LU o a. LU O- 1600 2000 M 0400 0800 N 1600 2000 M 0400 Midnight Noon TIME OF DAY FIGURE 3. — Diurnal change in percent of gulf menhaden, Brevoortia patronus, larvae with gas in their swim bladders (S.S. = sunset, S.R. = sunrise). Numbers offish examined are in parentheses. Data from two cruises (Dec. 1981, open circles; Dec. 1982, closed circles) are combined and presented in chronological order. TABLE 1. — Swim bladder volume of gulf menhaden, Brevoortia patronus, larvae measured immediately after capture. Samples are from oblique and surface net tows. Length Night samples Day samples class Sw m bladder vol. (mm3) Swim bladder vol. (mm) N (mean ± 2 SE)' N (mean ± 2 SE)' f df P <14.9 7 0.44 ±0.19 1 0.03 — — — 15-16.9 13 0.89 ± 0.21 9 0.05 ± 0.02 7.97 20 -0.001 17-18.9 18 1.71 ± 0.52 12 0.17 ± 0.10 5.81 28 <0.001 >19.0 11 1.78 ± 0.75 8 0.32 ± 0.19 3.77 17 <0.01 'SE = standard error of the mean. 515 FISHERY BULLETIN: VOL. 82, NO. 3 TABLE 2. — Swim bladder volume of gulf menhaden, Brevoortia patronus, larvae captured in MOCNESS tows and measured immediately after capture. Volumes were corrected for expansion of the swim bladder due to the change in pressure (0.1 atm/m of water). Night samples Corrected Day samples Corrected Water Sw m bladder vol. (mm3) for % Swim bladder vol. for depth (mean ± 2 SE)1 pressure with (mean ± 2 SE)1 pressure % (m) N At surface change gas N At surface change gas 1 55 0.422 ±0.148 NA 97 54 0 033 ± 0.013 NA 0 8 75 0 593 ±0.136 0.329 ± 0 075 99 3 (2) - — 0 20 21 0.336 ±0.123 0.126 ± 0.044 97 0 — — — — 1SE = standard error of the mean 2Swimbladder volume of the three fish captured was not measured. this station varied between 23 and 27 m. During the day nearly all the larvae were taken at the surface and only three larvae were captured as deep as 8 m. Without exception, fish examined from daylight samples did not have gas in the swim bladder, while almost all of the fish from the night samples contained some gas. In some cases, the volume of the swim bladder was such that it constricted the gut (Hunter and Sanchez 1976) or burst through the body wall. The volume of gas in the larval swim bladder would, of course, be reduced due to increased pres- sure as the larvae moved deeper in the water. As the volume of the swim bladder decreased, its ca- pacity as a buoyancy organ would decrease, caus- ing the larvae to expend more energy in swimming or to sink more rapidly Since our measurements were all made at the surface, we corrected the volumes of the swim bladders in larvae collected at 8 and 20 m to reflect the increased pressure at these depths (Table 2). Since swim bladder volume is related to size of the fish, a £-test was used to compare the mean standard length of the larvae from each depth. This test showed no significant difference in lengths of fish captured at the three depths U-test, P > 0.05). In the Laboratory Swim bladder volume was much greater in tanks where the larvae had direct access to air (Fig. 4). Swim bladder volume of the larvae with- out access to air remained essentially the same throughout the experiment. It appears from this experiment that gulf menhaden larvae, like a number of other clupeoid species, fill their swim bladders by swallowing air at the surface. DISCUSSION Our findings for swim bladder inflation in larval gulf menhaden generally agree with the findings of Hunter and Sanchez (1976) and Uotani (1973) for other clupeoid species. Our field studies showed conclusively that gulf menhaden inflate their swim bladders at night and deflate them during the day. Hunter and Sanchez (1976) suggested that nighttime swim bladder inflation in larvae of the northern anchovy is an energy-sparing mechanism that allows larvae to reduce swim- ming activity during nonfeeding periods while maintaining their depth in the water column. These authors further suggest that a reduction in swimming activity may reduce predation, since some predators of larval fish (e.g., chaetognaths) use the water movement caused by swimming activity to detect their prey. In the laboratory we found that larvae were unable to fill their swim bladders when they were prevented from reaching the air-water interface. This too agrees with the previous hypothesis on 1600 2000 2400 0400 0800 1200 1600 TIME OF DAY FIGURE 4.— Swim bladder volume (X ± 2 SE) of gulf menhaden, Brevoortia patronus, held in the laboratory with access (solid line) and without access (dashed line) to the air- water interface. 516 HOSS and PHONLOR: SWIM BLADDER OF GULF MENHADEN LARVAE how larvae of physostomatous clupeoid fishes initially inflate the swim bladder (Blaxter and Denton 1976). Under field conditions, gulf menhaden larvae began to fill and empty their swim bladders during the approximately 45 min of twilight preceding sunset and sunrise. The numbers of fish with gas in their bladders increased and decreased gradu- ally (i.e., it is not an all or none phenomenon) prior to darkness (or daylight). This observation suggests to us that larvae are responding as indi- viduals to gradually changing light levels. This response is probably better developed in larger larvae. The relation of diurnal vertical migration to swim bladder inflation that we found is different from the generally accepted position that larvae are near the sea surface with well-inflated swim bladders at night and are deeper in the water with deflated swim bladders during the day. We cap- tured menhaden larvae at three discrete depths at night (down to 20 m) and over 95% of the larvae captured at 8 and 20 m had gas in their swim bladders. On four previous cruises, collections in the same location also showed that menhaden lar- vae were distributed throughout the water column at night but concentrated at the surface during the day (unpubl. data5). In conclusion, the swim bladder of the larval stages of gulf menhaden acts as a buoyancy reg- ulator that allows the fish to maintain a position in the water column at night without movement. By day the swim bladder is deflated, and the larvae must actively swim to maintain their position near the water surface. At some point during de- velopment, the swim bladder's primary function switches to that of a pressure-adjusting mechanism for the otic bullae. 5Sogard, Susan M., Donald E. Hoss, and John J. Govoni. In prep. Density and depth distribution of larval fishes at selected sites in the northern Gulf of Mexico. Beaufort Lab., Natl. Mar. Fish. Serv. 28516-9722. Southeast Fish. Cent. NOAA, Beaufort, NC ACKNOWLEDGMENTS Robin Berry and Susan Sogard assisted in the data analyses and laboratory work. John Govoni, William Hettler, and Susan Sogard provided crit- ical review of the manuscript, and W Hettler provided laboratory-reared gulf menhaden. MOCNESS sampling was conducted by Shailer Cummings (Atlantic Oceanographic and Mete- orological Laboratories, NOAA) with able assis- tance from the crew of the RV Oregon II. This research was supported in part by a fellowship to the Junior author from the Brazilian National Science Council. LITERATURE CITED BLAXTER, J. H. S., AND E. J. DENTON. 1976. Function of the swimbladder-inner ear-lateral line system of herring in the young stages. J. Mar. Biol. As- soc. U.K. 56:487-502. BLAXTER, J. H. S., AND J. R. HUNTER. 1982. The biology of clupeoid fishes. Adv. Mar. Biol. 20:1-223. DOROSHEV, S. I., J. W. CORNACCHIA, AND K. HOGAN. 1981. Initial swim bladder inflation in the larvae of physoclistous fishes and its importance for larval cul- ture. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 178:495- 500. HETTLER, W. F. 1983. Transporting adult and larval gulf menhaden and techniques for spawning in the laboratory. Prog. Fish- Cult. 45:45-48. HOSS. D. E., AND J. H. S. BLAXTER. 1981. Development and function of the swimbladder-inner ear-lateral line system in the Atlantic menhaden, Bre- voortia tyrannus (Latrobe). J. Fish. Biol. 20:131-142. HUNTER, J. R., AND C. SANCHEZ. 1976. Diel changes in swim bladder inflation of the larvae of the northern anchovy, Engraulis mordax. Fish. Bull., U.S. 74:847-855. Tracy, H. C. 1920. The membranous labyrinth and its relation to the precoelomic diverticulum of the swimbladder in clupeoids. J. Comp. Neurol. 31:219-257. UOTANI, I. 1973. Diurnal changes of gas bladder and behavior of post- larval anchovy and other related species. Bull. Jpn. Soc. Sci. Fish. 39:867-876. WIEBE, P. H., K. H. BURT, S. H. BOYD, AND A. W. MORTON. 1976. A multiple opening/closing net and environmental sensing system for sampling zooplankton. J. Mar. Res. 34:313-326. 517 NOTES COMPARISON OF AMERICAN EEL GROWTH RATES FROM TAG RETURNS AND LENGTH-AGE ANALYSES Estimates of growth rates of American eel, Anguilla rostrata, have been largely indirect, based on projections from length-age regressions or comparisons of mean lengths at particular ages (Smith and Saunders 1955; Boetius and Boetius 1967; Ogden 1970; Bieder 1971; Gray and Andrews 1971; Hurley 1972; Harrell and Loyacano 1980; Kolenosky and Hendry 1982). Although valid for many fish species, these two approaches may be questionable in eel studies because of high vari- ability in lengths at given ages and because of considerable overlap in lengths among ages (e.g., Bertin 1956; Fahay 1978; Facey and LaBar 1981; Moriarty 1983). Testing the accuracy of a length- age regression as an estimator of growth rate requires a simultaneous mark-recapture study. Our objective was to mark and recapture eels in a Georgia estuary and to compare growth data from recaptures with growth estimates derived from length-age regressions and mean-length-at-age calculations for eels from the same population captured at the same time. We also sought infor- mation on seasonal growth patterns and differ- ences in growth rates among size classes. Materials and Methods All American eels were captured in tidal Friday- cap Creek (lat. 31°21'N, long. 81°24'W) which enters the South Altamaha River, Ga., about 11 km from the river mouth (see Helfman et al. 1983). Salinities and water temperatures ranged from 0 to 22/., and 5.5° to 31°C, respectively. Baited eel traps were set at or before sunset and pulled shortly after sunrise the next day. Animals were anesthetized in an ice slurry or in tricaine meth- anesulfonate, measured (total length), weighed, tagged with 25 mm long Floy1 FD-68B anchor tags, and released where captured. We tagged 659 animals on eight occasions between October 1980 and December 1982. Growth data from eels at large < 20 d were not used because of possible confusion with measurement error, which aver- aged ± 1 mm (range = 0-5 mm, N = 35 measure- ments of seven eels). Age determinations are based on sagittal oto- lith analyses from 305 eels captured concurrently with tagged animals. Most otoliths had distinct opaque and transparent zones, with few apparent supernumerary zones. Seasonal analysis of otolith margins indicated that presumed annuli were deposited on an annual basis and were a reason- able chronicle of age (Helfman et al. in press). Fish used in the mark-recapture study of growth were not collected for histological examination of gonads, and we therefore could not determine if sex-related differences in growth occurred (Tesch 1977). Results We recaptured 101 individuals, for an overall recapture rate of 15% . Time at large ranged from 8 to 493 d. Recapture frequencies were 84 fish recaptured once, 14 recaptured twice, 2 recaptured three times, and 1 recaptured four times. Growth rates of recaptured eels were variable but fell into two apparent seasonal categories (Table 1): 1) Slow growth from November through February (0.0-0.08 mm/d,x = 0.026, SD = 0.024, N = 13 recaptures) and 2) fast growth during spring, summer, and fall (0.01-0.63 mm/d, x = 0.221, SD = 0.152, N = 78 recaptures); fast period growth was significantly greater U-test, P < 0.001). Combining averages, and assuming a slow period of 4 mo, yield an average annual growth TABLE 1. — Growth rates of recaptured American eels as a function of season and year. Values in the body of the table are numbers of animals with particular growth rates. Intervals for fastest and slowest rates are subdivided by 0.05 mm/d; other intervals are 0.10 mm/d. Slow growth period (Nov.-Feb.)1 Fast growth period (Mar -Nov.) Growth (mrr i/d) 1980-81 1981 1982 0.00-0.05 12 2 4 006-0.10 1 5 12 0.11-0.20 10 9 0.21-0.30 7 9 0.31-0.40 3 4 0.41-0.50 7 0.51-0.60 1 4 0.61-0.65 1 x growth (mm/d) 0026 0.182 0.246 SD 0.024 0.107 0.172 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 'An additional 26 eels at large from late November 1982 to early May 1983, i.e., encompassing primarily the slow growth period, grew an average of 0.054 mm/d (SD = 0.034 mm/d). FISHERY BULLETIN: VOL. 82, NO. 3, 1984. 519 rate of 57 mm for eels 270-500 mm long over the 2-yr studied (95% C.I. [confidence interval] = ± 8.4 mm, growth periods treated as independent random variables, Bliss 1967). Growth as percent increase in length for an average eel 347 mm long was 16%. Accuracy of the extrapolated annual estimate can be tested against independent, long-term growth data from animals whose recapture inter- vals included both growth periods. Seven animals had recapture intervals of 6-16 mo (Table 2); average growth was 62 mm/yr (95% C.I. = ± 20.1 mm) or a 17% increase in length. TABLE 2. — Annual growth rates of eels in Fridaycap Creek, Ga., based on long-term recaptures. Data are from eels that were at large for more than a 180-d interval that included both fast and slow growth periods. Date Days Length (mm) Growth Percent Eel First Second at rate in- no. capture capture large Initial Final (mm/yr) crease '1 26-X-80 27-VIII-81 313 353 429 83 23 1 26-11-81 24-II-82 363 362 460 98 27 2 26-X-80 30-IV-81 186 328 351 45 14 3 7-111-81 24-II-82 353 378 450 74 20 4 26-X-80 27-VIII-81 306 487 511 24 5 5 19-11-81 24-II-82 369 307 352 45 15 6 7-111-81 24-II-82 354 433 504 73 17 7 7-111-81 13-VII-82 492 393 492 56 14 X 381 62 17 SD 58 24 7 'Eel No. 1 was captured four times; growth between first and third and be- tween second and fourth captures were analyzed separately. When the data are grouped into 50 mm size classes, animals in the 350-400 mm class grew faster than smaller animals (Fig. 1); 95% confi- dence intervals for other size classes overlapped, although some overlap may result from small samples of larger animals. Similar trends in relative growth (percent increase in length) were apparent (Fig. 1): values overlapped in the smaller size classes, and the largest size class grew slower than the fast-growing 350-400 mm group. Growth rates during fast growth periods (Table 1) suggest that animals grew faster in 1982 than in 1981 ( t-test, one tail, df = 76, P < 0.05). Maximum growth rates also differed: the 5 fastest growth rates, as well as 13 of the 15 fastest rates, occurred in 1982 (Table 1). Information on weight gain is less complete but shows a similar seasonal trend. Average weight increase between recaptures was 0.223 g/d (SD = 0.222, N = 47) for the fast growth period; limited data suggest lesser gains for the slow growth period (0.017-0.144 g/d, A7 = 2). When seasonal data are summed and a 4-mo slow growth period is assumed, annual weight increase was 63 g. Long-term weight change data from two animals at large 299 and 371 d indicate an average weight increase of 76 g/yr (range = 67 to 86 g/yr). Mean lengths at different ages were Age class (yr): II III IV V VI VII x length (mm) 242 310 361 403 442 460 Range (mm) 197- 214- 233- 256- 297- 386- 278 446 548 570 559 500 N 6 51 134 78 32 5 The mean values project an average annual in- crease of 44 mm (range = 39-68 mm). The related length-age regression for all eels aged at this locale during the study period was length = 183.3 + 43.5 x age (N = 305, r = 0.492, P < 0.01) which also projects an average annual increase of 44 mm (95% C.I. = ± 8.7 mm) for an average eel 370 mm long. .50 .40 E E I O (Z .30 .20 • 10 .00 T I I 19 34 i I I • ' 5 I I I H f.13 I Li2 ♦ r.io K08 T I I I I I \ I I I I1 I 275 325 375 425 475 LU < LU o UJ o q: 06 LU [-.04 LENGTH CLASS (mm) FIGURE 1. — Growth rates of recaptured American eels as a function of size at initial capture, Fridaycap Creek, Ga., October 1980-November 1982. Growth is expressed as the actual daily rate of increase (solid vertical lines, means ±95% confidence intervals) and the percent increase as a function of initial length (dashed vertical lines, means ± 95% confidence intervals). Data are from fast growth periods, 1981 and 1982 combined. Numbers beside each mean are the total animals comprising each 50 mm size class. Midpoints of length classes are shown on x-axis. 520 Discussion Comparisons of Growth Measures Different procedures yielded different estimates of growth rate. The two independent, direct mea- sures based on recaptures — seasonal summation and long-term recaptures — produced similar val- ues (57 mm/yr and 62 mm/yr, respectively). The indirect measures — length-age regression and mean-length-at-age analysis — both projected an- nual growth rates of 44 mm/yr. All estimates are complicated by extreme variability in growth, with overlap in lengths among four to six year classes common (Smith and Saunders 1955; Ogden 1970; Gray and Andrews 1971; Hurley 1972). Growth rate estimates based on recapture data were apparently higher than those derived from length-at-age analyses, but confidence intervals overlapped among all estimates. However, we feel that the direct measures are more accurate. First, the sample size for the length-age analyses was more than three times larger than for the seasonal summation analysis, but the confidence intervals were very similar (17.4 mm and 17.8 mm, respec- tively), suggesting less variability in the recap- ture data. Second, growth rates derived from recaptured animals are based on actual growth between captures; variability in calculated growth rates should therefore reflect real variability in growth among animals. In length-age analyses, age classes are commonly resolved at no finer than an annual level. Consequently, growth subse- quent to day 1 of each year increases the variance around the estimate rather than increasing the accuracy of the estimate. Finally, the accuracy of age determinations from otoliths in some eel populations is questionable (Moriarty and Stein- metz 1979; Deelder 1981; Casselman 1982), placing length-age analyses in doubt unless annulus for- mation can be verified. Limited growth data from other mark-recapture studies of American eels are available. Hurley ( 1972 ) tagged 1,418 American eels in Lake Ontario, Canada, and reported recapture intervals for 13 large individuals (730-874 mm), which increased an average of 34 mm/yr. At two Louisiana fresh- water locales, Gunning and Shoop (1962) tagged 43 American eels; only four recaptures provided usable data, indicating an average growth of 140 mm/yr (growth range = 46-325 mm, initial lengths = 255-915 mm). R. L. Haedrich2 tagged 148 American eels in a Massachusetts estuary. Four individuals (initial lengths = 500-700 mm) had an average annual growth rate of 6% (range = 4.1-8.4%). An inverse latitudinal trend in growth is suggested (see also Harrell and Loyacano 1980), but direct comparison is complicated by different initial lengths, small sample sizes, and high variability in growth. Length-related differences in growth have also been found for other populations. A shift from allometric to symmetric growth occurred at 800 mm for American eels in Lake Ontario (Hurley and Christie 1982). Those authors, as well as Smith and Saunders (1955), related such a growth change to physiological preparation for matura- tion and migration. Gray and Andrews (1971) found that American eels in New Brunswick, Canada, estuaries grew slowly after age XI. Helf- man et al. (in press) suggested that maturation of Fridaycap Creek eels occurred at around age IV (mean length = 387 mm). An apparent decrease in growth rates of Fridaycap Creek animals longer than 400 mm (Fig. 1) supports their interpretation. Causes of Seasonal Differences Seasonal and annual differences in growth rate can be linked to fishing success as affected by climate. Eel fishing in Georgia estuaries is typi- cally poor at water temperatures below 10° C and above 24° C. In 1980-81, estuarine water tem- perature fell below 10° C during December 1980, but average 1981-82 monthly temperatures were higher and did not reach the 10° minimum until January 1982. In addition, rainfall in 1981 was 45 cm below average, and mean water temperatures were 2°C higher during June through September than in 1982 (R. Arnsdorff3 and T. E. Targett4). The winter slow growth period may therefore result from colder water temperatures and re- duced feeding. Faster growth in 1981 than in 1982 may have resulted from elevated temperatures during much of the fast growth period of 1981. High water temperatures — leading to reduced feeding, interrupted growth, and poor fishing — occurs in European eels, Anguilla anguilla (Deelder 1981). Interrupted summer growth may R. L. Haedrich. Department of Biology, Memorial University of Newfoundland, St. John's, Newfoundland, Canada A1B 3X9, pers. commun. April 1983. 3R. Arnsdorff, Georgia Department of Natural Resources, Environmental Protection Division, 270 Washington St. S.W, Atlanta, GA 30334, pers. commun. October 1982. 4T. E. Targett, Skidaway Institute of Oceanography, P.O. Box 13687, Savannah, GA 31406, pers. commun. October 1982. 521 have occurred in our study population, but be- cause we lack growth data from midsummer only, we cannot test for it. Acknowledgments We thank K. Benson, J. Biggers, E. Brown, R Christian, J. Crim, and S. Pierce for their con- tributions during field and laboratory work, and D. Facey and S. Hilliard for editorial comments. The University of Georgia Marine Extension Ser- vice and the staff of Two-Way Fish Camp, Darien, Ga., have been most helpful throughout our inves- tigations. This work is a result of research spon- sored by the National Oceanic and Atmospheric Administration's Office of Sea Grant, Department of Commerce, under Grant NA80AA-00091. Literature Cited BERTIN, L. 1956. Eels. A biological study. Cleaver-Hume Press, Lond.,192p. BIEDER, R. C. 1971. Age and growth in the American eel, Anguilla rostrata (LeSueur), in Rhode Island. MS Thesis, Univer- sity of Rhode Island, Kingston, R.I., 39 p. BLISS, C. I. 1967. Statistics in biology. Vol. I. McGraw-Hill, N.Y., 558 p. BOETIUS, I., AND J. BOETIUS. 1967. Eels, Anguilla rostrata, LeSueur, in Bermuda. Vi- densk. Meddr. Dansk. Naturh. Foren. 130:63-84. CASSELMAN, J. M. 1982. Chemical analyses of the optically different zones in eel otoliths. In K. H. Loftus (editor), Proceedings of the 1980 North American Eel Conference, p. 74-82. Ont. Fish. Tech. Rep. Ser. No. 4. DEELDER, C. L. 1981. On the age and growth of cultured eels, Anguilla anguilla (Linnaeus, 1758). Aquaculture 26:13-22. FACEY, D. E., AND G. W. LABAR. 1981. Biology of American eels in Lake Champlain, Ver- mont. Trans. Am. Fish. Soc. 110:396-402. FAHAY, M. P. 1978. Biological and fisheries data on American eel, Anguilla rostrata (LeSueur). Natl. Mar. Fish. Serv., Sandy Hook Lab., Highlands, N.J., Tech. Ser. Rep. 17, p. 1-82. GRAY, R. W., AND C. W. ANDREWS. 1971. Age and growth of the American eel (Anguilla rostrata (LeSueur)) in Newfoundland waters. Can. J. Zool. 49:121-128. GUNNING, G. E., AND C. R. SHOOP. 1962. Restricted movements of the American eel, Anguilla rostrata (LeSueur), in freshwater streams, with comments on growth rate. Tulane Stud. Zool. 9:265-272. harrell, r. m., and h. a. loyacano, Jr. 1980. Age, growth and sex ratio of the American eel in the Cooper River, South Carolina. Proc. Annu. Conf. S.E. Assoc. Fish Wildl. Agencies 34:349-359. HELFMAN, G. S., E. L. BOZEMAN, AND E. B. BROTHERS. In press. Size, age and sex of American eels in a Georgia river. Trans. Am. Fish. Soc. HELFMAN, G. S., D. L. STONEBURGER, E. L. BOZEMAN, P A. CHRISTIAN, AND R. WHALEN. 1983. Ultrasonic telemetry of American eel movements in a tidal creek. Trans. Am. Fish. Soc. 112:105-110. HURLEY, D. A. 1972. The American eel (Anguilla rostrata) in eastern Lake Ontario. J. Fish. Res. Board Can. 29:535-543. HURLEY, D. A., AND W. J. CHRISTIE. 1982. A re-examination of statistics pertaining to growth, yield and escapement in the American eel (Anguilla rostrata ) stocks of Lake Ontario. In K. H. Loftus (editor), Proceedings of the 1980 North American Eel Conference, p. 83-85. Ont. Fish. Tech. Rep. Ser. No. 4. KOLENOSKY, D. R, AND M. J. HENDRY. 1982. The Canadian Lake Ontario fishery for American eel (Anguilla rostrata). In K. H. Loftus (editor), Pro- ceedings of the 1980 North American Eel Conference, p. 8-16. Ont. Fish. Tech. Rep. Ser. No. 4. MORIARTY, C. 1983. Age determination and growth rate of eels, Anguilla anguilla (L). J. Fish Biol. 23:257-264. MORIARTY, C, AND B. STEINMETZ. 1979. On age determination of eel. Rapp. P.-V Reun. Cons. Int. Explor. Mer 174:70-74. OGDEN, J. C. 1970. Relative abundance, food habits, and age of the American eel, Anguilla rostrata (LeSueur), in certain New Jersey streams. Trans. Am. Fish. Soc. 99:54-59. SMITH, M. W, AND J. W. SAUNDERS. 1955. The American eel in certain fresh waters of the Maritime Provinces of Canada. J. Fish. Res. Board Can. 12:238-269. TESCH, F.-W. 1977. The eel. J. Greenwood, translator. Chapman and Hall, Ltd., Lond./J. Wiley & Sons, N.Y., 434 p. GENE S. HELFMAN Earl L. Bozeman Zoology Department and Institute of Ecology University of Georgia Athens, GA 30602 Edward B. Brothers Section of Ecology and Systematics Cornell University Ithaca, NY 14853 522 DESCRIPTION OF EARLY STAGE ZOEAE OF SPIRONTOCARIS MURDOCHI (DECAPODA, HIPPOLYTIDAE) REARED IN THE LABORATORY Larvae of Spirontocaris murdochi Rathbun have not been described in the literature. During stud- ies on rearing larvae in the laboratory for descrip- tive purposes, I succeeded in rearing zoeae of S. murdochi through Stage III. The first three zoeal stages of S. murdochi are described, illustrated, and compared with descriptions of morphological- ly similar hippolytid zoeae. Methods and Rearing Results I obtained an ovigerous female Spirontocaris murdochi carrying late-stage embryos while sam- pling pandalid shrimp in Auke Bay, Alaska, for toxicity studies. The female was caught 2 April 1979 at a depth of 18 m at lat. 58° 21.6' N, long. 134° 39.3' W. Stage I zoeae released from the female were reared in 250 ml jars containing about 200 ml of filtered seawater. The jars were checked daily for exuviae, and a few zoeae were preserved every other day. The zoeae were offered live plankton strained through a 0.333 mm mesh, but there was no evidence that the zoeae ate the plankton. (For a more complete description of the methods, see Haynes 1982.) Most of the zoeae molted to Stage II, but only two zoeae molted to Stage III. Techniques of measurement and illustration are those of Haynes (1976, 1979). At least five zoeae of Stages I and II were used to verify seg- mentation and setation. Description The terms used in the descriptions and nomen- clature of appendages are from Haynes (1976, 1979). Only those morphological characteristics useful for readily identifying each stage are giv- en. Setation formulae are the number of setae per segment from the distal to the proximal segment. The telsonic setae are numbered as pairs begin- ning with the inner (medial) pair. For clarity, setules on setae are usually omitted, but spinu- lose setae are shown. The following characteristics apply to zoeal Stages I, II, and III. The rostrum is slender, spini- form, without teeth, about one-fourth the length of the carapace, and projects horizontally. The ventral and posterior margins of the carapace are smooth except for pterygostomian spines. Man- dibles are without palps; there is no proximal setose seta on the maxillule; and the maxillipeds are without epipodites. Abdominal somites 4 and 5 have posterolateral spines (the fifth pair is slightly longer than the fourth pair in Stage I, but both pairs are nearly the same length in Stages II and III). An anal spine is present. Stage I Zoea Mean total length of Stage I zoea (Fig. 1A), 3.4 mm (range 3.2-3.6 mm; six specimens). Eyes sessile. Carapace with two minute rounded prom- inences: One at posterior edge, other at base of rostrum. ANTENNULE (Fig. IB). — Protopodite of first antenna, or antennule, simple, unsegmented, tubular, with heavily plumose seta terminally. Conical projection tipped with four aesthetascs: Three long, one of intermediate length. ANTENNA (Fig. 1C). — Second antenna, or an- tenna, with inner flagellum (endopodite) and outer antennal scale (exopodite). Flagellum un- segmented, slightly shorter than scale, styliform, tipped by plumose seta and shorter spine. Anten- nal scale distally divided into five joints (distal joint incomplete) and fringed with 10 heavily plu- mose setae along terminal and inner margins. Tip of antennal scale curved toward outer margin. Protopodite with spine only at base of flagellum. MANDIBLES (Fig. ID).— Incisor process of left mandible has four teeth in contrast to triserrate incisor process of right mandible. Both left and right mandibles with movable premolar denticle (lacinia mobilis). Left mandible with subterminal tooth. MAXILLULE (Fig. IE). — First maxilla, or max- illule, with coxopodite, basipodite, and endo- podite. Coxopodite (proximal lobe) with seven spines: Five spinulose, two simple. Basipodite (median lobe) with 10 short, smooth spines ter- minally. Two-segmented endopodite originates from lateral margin of basipodite: Proximal seg- ment with two spinulose spines, distal segment with three spinulose spines. MAXILLA (Fig. IF). — Second maxilla, or max- illa, has platelike exopodite (scaphognathite) with five plumose setae along outer margin and FISHERY BULLETIN: VOL. 82, NO 3, 1984. 523 FIGURE 1. — Stage I zoea of Spirontocaris murdochi: A, whole animal, right side; B, antennule, dorsal; C, antenna, ventral; D, man- dibles (left and right), posterior; E, maxillule, ventral; F, maxilla, dorsal. 524 FIGURE l. — Continued — G, first maxilliped, lateral; H, second maxilliped, lateral; I, third maxilliped, lateral; J, telson, ventral. 525 hairs along medial margin. Unsegmented endo- podite with nine setae (four setae slightly spinu- loses Coxopodite and basipodite bilobed. Coxopo- dite with eight setae on each lobe. Basipodite with 16 setae: 4 on distal lobe, 12 on proximal lobe. FIRST MAXILLIPED (Fig. 1G). — Most setose of natatory appendages. Bilobed protopodite with 5 setae on proximal lobe, 20 setae on distal lobe (three of setae on distal lobe spinulose). Endopo- dite four-segmented; setation formula 4, 2, 1, 2. Exopodite (a long, slender ramus jointed at base) has four natatory setae. SECOND MAXILLIPED (Fig. 1H).— Protopodite bisegmented: Distal segment with seven setae, proximal segment without setae. Endopodite four-segmented; setation formula 6, 2, 1, 3. Exo- podite with five natatory setae. THIRD MAXILLIPED (Fig. II).— Unsegmented protopodite with three setae. Five-segmented en- dopodite about two-thirds length of exopodite; setation formula 5, 2, 1, 1, 2. Exopodite with five natatory setae. PEREOPODS.— Poorly developed, anteriorly di- rected under body. PLEOPODS. —Absent. ABDOMEN AND TELSON (Fig. 1A, J).— Abdo- men with pair of posterolateral spines on somites 4 and 5, pair on somite 4 somewhat shorter than pair on somite 5. Telson emarginate posteriorly, fused with abdominal somite 6. Telson with 7 + 7 densely plumose setae, minute spinules at base of each seta except outermost pair, larger spinules along terminal margin of telson between bases of four inner pairs of setae. Uropods visible and enclosed. Stage II Zoea Mean total length of Stage II zoea, 3.7 mm (range 3.5-4.0 mm; three specimens). Eyes stalked. Carapace identical to Stage I carapace. ANTENNULE (Fig. 2A).— Two-segmented, with large outer flagellum and smaller inner flagel- lum on terminal margin. Flagella not segmented; inner flagellum conical, with long spine termi- nally; outer flagellum with four aesthetascs ter- minally. Proximal segment with large spine pro- jecting slightly downward from ventral surface. Both proximal and distal segments have two plumose setae each. L 0.25 mm FIGURE 2. — Stage II zoea of Spirontocaris murdochi: A, antennule, ventral; B, antenna, ventral; C, telson, ventral. 526 ANTENNA (Fig. 2B). — Flagellum styliform, about same length as antennal scale, tipped by short spine. Antennal scale about 3.5 times as long as wide, fringed with 15 plumose setae along terminal and medial margins. Antennal scale with four joints distally (proximal joint incom- plete), lateral projection on distal portion. Tip of antennal scale not curved laterally as in Stage I. Protopodite with spine at base of flagellum. MANDIBLES. — Same as in Stage I except have slightly developed molar lip. MAXILLULE, MAXILLA, AND MAXILLIPEDS. — Similar to Stage I except scaphognathite of maxilla has six setae; exopodites of maxillipeds 1-3 have 5, 8, and 10 natatory setae, respectively. PEREOPODS.— Slightly larger than in Stage I, extend somewhat vertically, have naked exopo- dites on pereopods 1 and 2. PLEOPODS. — Absent. ABDOMEN AND TELSON. — Posterolateral spines on abdominal somites 4 and 5 nearly same length. Telson (Fig. 2C) still fused with ab- dominal somite 6, has 8 + 8 densely plumose setae. Enclosed uropods somewhat longer than in Stage I. Stage III Zoea Mean total length of Stage III zoea, 4.1 mm (range 3.9-4.3 mm; two specimens). Carapace identical to Stage II carapace, except has supra- orbital spine. ANTENNULE. — Similar to Stage II antennule except outer flagellum has subterminal seta, proximal segment with four setae around distal joint and three plumose setae laterally, distal segment with four large plumose setae. ANTENNA, MANDIBLES, MAXILLULE, AND MAXILLA. — Similar to Stage II but with follow- ing differences. Antennal scale, without joints terminally, has 20 plumose setae along terminal and medial margins; subterminal spine extends just beyond tip of scale. Mandibles with a few additional teeth between incisor and molar pro- cesses. Maxillule with 9 spinulose spines on coxo- podite and 12 short smooth spines on basipodite. Scaphognathite of maxilla has 9-12 plumose setae. MAXILLIPEDS. — Exopodites of maxillipeds 1-3 have 5, 10, and 10 natatory setae, respectively. PEREOPODS (Fig. 3 A, B). — Exopodites only on pereopods 1 and 2; chelae present but undevel- oped; endopodites of pereopods 1 and 2 with terminal seta. PLEOPODS. — Present as buds. ABDOMEN AND TELSON.— Telson (Fig. 3C), 0. 5 mm B FIGURE 3.— Stage III zoea of Spirontocaris murdochi: A, pereopod 1, lateral; B, pereopod 2, lateral; C, telson, dorsal. 527 jointed with abdominal somite 6, has 8 + 8 densely plumose setae. Most of spinules at setal bases and along terminal margin absent. Uro- pods free. Endopodite of uropod about one-half length of exopodite, has four setae along terminal margin. Exopodite usually with 13 distal mar- ginal setae and an outer subterminal spine. Comparison of Zoeal Stages with Descriptions by Other Authors Of the described Spirontocaris zoeae, those of S. murdochi are most similar to zoeae of S. spinus (Sowerby), S. spinus intermedia Makarov, and S. phippsii (Kr0yer): all have relatively long, spiniform rostrums; exopodites on pereopods 1 and 2; and posterolateral spines on abdominal somites 4 and 5. However, larvae of S. spinus (described by Pike and Williamson 1961) and S. spinus intermedia (described by Ivanov 1971) are distinguishable from larvae of 99 i i i i > f i i i 75 80 85 Premolt Coropoce Length IOO FIGURE 1. — Percentage of non-ovigerous female lobsters that molt and lay eggs in the same season in relation to size at Arnold's Cove, Newfoundland. 529 size (in relation to size at maturity) and exploita- tion rates in the fishery. In a fishery with a small minimum legal size and high exploitation rates, most of the ovigerous females in the population would be small animals laying for the first time. The incidence of new-shelled ovigerous females in autumn sampling at Arnold's Cove has ranged from 0 to 38.5% of the total ovigerous specimens examined (Table 1). This year-to-year variability, which has also been observed elsewhere in New- foundland (Ennis 1980), could be accounted for by variation in relative abundance of prerecruit ani- mals caused by annual fluctuation in recruitment and exploitation rate. TABLE 1. — Percentage of ovigerous lobsters with new shells in autumn sampling at Arnold's Cove, Newfoundland, 1975-82. tions of lobsters, Homarus americanus. Fish. Bull. U.S. 82:244-249. G. P. ENNIS Carapace length (mm) No. % ovigerous Range of ovigerous with Range of new-shelled Year examined new shell ovigerous ovigerous' 1975 75 10.7 72-103 73-83 19752 16 12.5 65-92 65-71 1976 31 6.5 73-92 83-90 19762 26 19.2 68-91 68-77 1977 78 385 71-101 76-88 1978 12 16.7 71-95 82-83 1979 31 25.8 72-99 72-90 1980 18 0.0 73-99 — 1981 31 6.5 71-101 71-81 1982 27 3.7 75-94 75 'These are postmolt carapace lengths 2Diver-caught samples obtained during the same period as the trap-caught samples. Acknowledgments I am indebted to R W. Collins and G. Dawe who, with assistance from other technical staff and casual employees, carried out the field work and data processing associated with this study. Literature Cited AIKEN, D. E., AND S. L. WADDY. 1976. Controlling growth and reproduction in the Amer- ican lobster. Proc. 7th Annu. Meet. World Mariculture Soc, p. 415-430. 1980. Reproductive biology. In J. S. Cobb and B. F. Phillips (editors), The biology and management of lob- sters, Vol. I, Physiology and behavior, p. 215-276. Acad. Press, N.Y. ENNIS, G. P. 1978. Growth curves for Newfoundland lobsters from data on molt increment and proportion molting. Research Document 78/29, Canadian Atlantic Fisheries Scientific Advisory Committee, Halifax, Canada, 11 p. 1980. Size-maturity relationships and related observa- tions in Newfoundland populations of the lobster (Homa- rus americanus) . Can. J. Fish. Aquat. Sci. 37:945-956. 1984. Comparison of physiological and functional size- maturity relationships in two Newfoundland popula- Fisheries Research Branch Department of Fisheries and Oceans P.O. Box 5667 St. John's, Newfoundland, Canada A1C 5X1 PARASITES OF OLIVE ROCKFISH, SEBASTES SERRANOIDES, (SCORPAENIDAE) OFF CENTRAL CALIFORNIA The olive rockfish, Sebastes serranoides, inhabits reefs from Del Norte County, Calif., to San Benito Island, Baja California, Mexico. Olive rockfish are large (to 64 cm TL), active predators, usually found in the water column, but occasionally hover- ing over or resting upon rocky substrates. Juveniles are primarily midwater feeders, preying upon zooplankton and small fishes, though some demersal feeding (e.g., isopods, caprellid and gammarid amphipods, etc.) has been noted (Hob- son and Chess 1976; Love and Ebeling 1978; Love and Westphal 1981). Adults feed almost entirely on nektonic forms of squid and fish and on substrate- dwelling octopus (Love and Westphal 1981). Little is known about the parasite fauna of olive rockfish, as previous reports are either descrip- tions of newly discovered species (Cressey 1969; Moser and Love 1975; Love and Moser 1976; Moser et al. 1976) or surveys of particular parasites throughout a fish community (Turner et al. 1969; Hobson 1971; Dailey et al. 1981). As part of a life- history study, we investigated the parasite popula- tion of central California olive rockfish. Methods Specimens were collected monthly from April 1975 to February 1976 at a group of shallow- water pinnacles, about 11 km west of Avila Beach, San Luis Obispo Co., Calif., (Fig. 1). These pinnacles, at depths of 20-30 m, are situated 100-300 m offshore from Diablo Cove and North Cove and rise to within 5-10 m of the surface. Six hundred olive rockfish, ranging from 8.6 to 49.2 cm TL, were collected by hook and line or spear, placed in plastic bags, and frozen for later dissection. After thawing, each specimen was 530 FISHERY BULLETIN: VOL. 82, NO. 3, 1984. n .. f* % //'A /* pV .' r? •' fXi. til''' ' #'*• m'< £•••" /?5? v Jf% \a% ' m \i'.V NORTH % COVE \& \3»f X «. J>j ^YiSAN FRANCISCO o£> fl ^^^^^^ V3».-' \{;?-- A v*\ vifc". <&" . w % w DIABLO Sv X?'~: ■ ' X COVE £% \*-'.>\ . /**'•'■ c\ Jf/-^1 9 V\ jkr AVI L A '*>{•. V *' C^Sv. V^* . CONCEPTlOlii%SANTA • ^^^^ARBARA lnSv**' 500M ^sStMbj; • \-'»!". •&«• ^"%-V • v? ;'/ m. ^ >£>. J&$' • V*' j/jftzstt* * 1 /?$»X*'''' ■^ \ s *•* 200 KM ^ \ FIGURE 1. — Location of sampling sites ( marked with x ) for olive rockfish off Diablo Cove, Calif. measured (total length) to the nearest millimeter and examined for parasites on external surfaces, gills, gill cavities, mouth, mesentery, heart, gallbladder, stomach, intestine, and muscle. Copepods and monogenetic and digenetic trem- atodes were fixed in alcohol-formaldehyde-acetic acid (AFA). The trematodes were stained with Harris' hematoxylin, cleared with xylene, and mounted. Nematodes were cleared in lactophenol. Protozoans were studied unpreserved after thawing. Most parasites were identified to the lowest pos- sible taxon. However, the microsporida, copepods of the genera Caligus and Lepeophtheirus , and larval cestodes, nematodes, and acanthocephalans were not identified to species. Copepods of the gen- era Caligus and Lepeophtheirus were only identi- fied to genus, as an earthquake destroyed most of the specimens before they were identified to species. To facilitate our analysis, we grouped together the relatively uncommon gallbladder myxozoans ( = myxosporidans) (Ceratomyxa sebasta, Lep- totheca informis, L. longipes, L. macrospora, Myxidium incurvatum, Zschokkella ilishae) and the hemiurid trematodes (Lecithaster gibbosus, Parahemiurus merus, Lecithochirium exodicum, and Tubulouesicula lindbergi). We analyzed the incidence of infection of all parasites over the entire range of host lengths throughout the year. Differences in prevalence be- tween size classes and between monthly samples were examined using the Kruskal- Wallis test. Results For our analyses we divided the specimens into 11 size classes. Table 1 shows the number of speci- mens taken per month per size class. 531 TABLE 1. — The number of olive rockfish taken per month per size class, April 1975-February 1976. Size class (cm TL) Month 5-10 11-15 16-20 21-25 26-30 31-35 36-40 41-45 46 + Total April 6 6 3 6 8 10 6 7 3 55 May 4 7 4 4 5 8 8 8 4 52 June 5 5 5 8 6 9 5 7 6 56 July 6 5 7 8 6 8 9 8 7 64 Aug. 4 4 7 6 10 8 7 6 5 57 Sept 5 8 6 9 5 9 9 9 3 63 Oct. 4 7 5 4 8 7 10 5 4 54 Nov. 5 6 5 6 7 8 8 7 3 55 Dec. 4 6 6 5 7 11 8 8 5 60 Jan. 2 3 5 6 4 6 6 7 5 44 Feb. 2 1 6 4 5 8 6 4 4 40 Total 47 58 59 66 71 92 82 76 49 600 Thirty-six parasite species were recovered from olive rockfish (Table 2). Five species were found in <1% of the individuals examined. These inciden- tal parasites were an unidentified microsporidan and the copepods Neobrachiella robusta, Chon- dracanthus pinguis, Naobranchia occidentalis , and Sarcotaces arcticus. Found in <10% of the hosts were Davisia reginae, Kudoa clupeidae, Lep- totheca longipes, and L. macrospora (Myxozoa); Trochopus marginata (Monogenea); Aporocotyle macfarlani, Lecithaster gibbosus, Lecithochirium exodicum, Parahemiurus merus, and Tubulovesi- cula lindbergi (Digenea); Anisakis sp. and Phoca- nema sp. (Nematoda); Caligus sp. and Clavella parva (Copepoda); and Rhabinorhynchidae gen. sp. (Acanthocephala). Larval cestodes were the most commonly en- countered parasites, infecting 98% of all individu- als >20 cm in length. Larval Contracaecum sp. were found in 62% and cysticanths of immature Corynosoma sp. in 18% of fishes >20 cm. Of parasites which use Sebastes serranoides as a final host, Microcotyle sebastis had the highest prevalence, occurring on more than 90% of hosts >20 cm. It was most prevalent on the filaments of the first gill arch (Table 3), declining in number through successive arches (x2 = 108.1, P < 0.001). No significant differences in infection intensities were noted between left and right arches. Other commonly encountered ectoparasites were Holobomolochus spinulus, Neobenedenia girellae, and Lepeophtheirus sp. Adult metazoan endoparasites were not abundant, though three, Deretrema cholaeum, Opechona sebastodis, and Hysterothylacium aduncum, were often found in larger fish. Three myxozoans, Henneguya sebasta, Lep- totheca in for mis, and Zschokkella ilishae, were found in more than 10% of hosts. Henneguya sebasta was found in 93% of hosts >35 cm. Nine percent of H. sebasta infections were sufficiently severe to virtually occlude the bulbous arteriosus. Although no histological sections were made, no evidence of gross pathogenic effects were noted, as these heavily infected individuals were of an age and weight indistinguishable statistically (analy- sis of variance) from lightly or non-infected indi- viduals. TABLE 2. — Parasites recovered from olive rockfish, Sebastes serranoides, off Diablo Cove, Calif. *denotes first host records. Parasite Location Protozoa (Myxozoa) Ceralomyxa sebasta Davisia reginae Henneguya sebasta 'Kudoa clupeidae Leptotheca informis Leptotheca longipes Leptotheca macrospora Myxidium incurvatum Zschokkella ilishae 'Protozoa (Microsporida) Monogenea 'Microcotyle sebastis 'Neobenedenia girellae 'Trochopus marginata Digenea 'Aporocotyle macfarlani 'Deretrema cholaeum 'Lecithaster gibbosus 'Lecithochirium exodicum 'Opechona sebastodis 'Parahemiurus merus 'Podocotyle sp. 'Tubulovesicula lindbergi Cestoda "Tetraphyllidea (immature) Nematoda Anisakis sp. (immature) 'Contracaecum sp. (immature) 'Hysterothylacium ( =Thynnascans) aduncum Phocanema sp. (immature) Copepoda 'Neobrachiella robusta Caligus sp. 'Chondracanthus pinguis 'Clavella parva Holobomolochus spinulus 'Lepeophtheirus sp. 'Naobranchia occidentalis 'Sarcotaces arcticus Acanthocephala 'Corynosoma sp (immature) "Echinorhynchus gadi "Rhabdinorhynchidae gen. sp. Gallbladder Urinary bladder Bulbus arteriosus, gallbladder (rarely) Muscle Gallbladder Gallbladder Gallbladder Gallbladder Gallbladder Urinary bladder Gills Skin, mouth Gills Afferent branchial arteries Gallbladder Stomach Stomach Intestine Stomach Stomach Stomach Viscera Viscera Viscera Stomach, intestine Muscle Gills Skin, gills Gills Dorsal and anal fin rays Gills, inner surface of gill opercula Skin, gills Gills Body cavity near anus Viscera Intestine Intestine TABLE 3. — Position and number of Microcotyle sebastis on 32 olive rock- fish, Sebastes serranoides, off Diablo Cove, Calif. Arch position number Left Right Total 1 114 124 238 2 60 63 123 3 29 37 66 532 We found numerous cases of multiple species myxozoan infections in the gallbladder, particu- larly in individuals >35 cm. Twenty-two percent of all infections were comprised of two species, 5.1^ of three, and 1.29£- of four. The occurrence of myxozoans in Deretrema cholaeum-infected gallbladders occurred less frequently than ex- pected (y2 = 123.3, P < 0.0001). The species of parasites infecting olive rockfish by host length and age is shown in Table 4. Sebastes serranoides harbors a maximum number of parasite species between 31 and 40 cm or 4 and 10 yr of age [compared with 3-6 yr in S. alutus and S. caurinus (Sekerak 1975)]. Of the five species of parasites found in the smallest size class, four exhibited direct life cycles, whereas in fish of 20 cm (1-2 yr old) 6 of 11 species had indirect life cycles. In the largest class (41-50 cm), slightly less than half (15 of 34) of the species had indirect life cycles. By 20 cm, representatives of all the parasite groups, with the exception of microsporida, were found in olive rockfish. The prevalence rates of six parasite species and one parasite group increased signifi- cantly with increasing host length (Fig. 2). Seven species or species groups showed significant an- nual changes in prevalence (Fig. 3). Discussion Of the seven parasite species showing increas- ing prevalence with increasing host length, six (Lepeophtheirus sp., Neobenedenia girellae, Micro- cotyle sebastis, Holobomolochus spinulus, Henne- guya sebasta, and the gallbladder myxozoans) had direct life cycles and one (Hysterothylacium adun- cum) had an indirect cycle. A change in diet to- TABLE 4. — Parasite species infecting five size classes of olive rockfish off Diablo Cove, Calif. See Table 1 for number of specimens per size class. Host length cm TL (age in years) 5-10(0) 11-20(1-2) 21-30(1-4) 31-40(4-10) 41-50(7-20) Acanthocephala Acanthocephala Acanthocephala Acanthocephala Corynosoma sp Corynosoma sp Corynosoma sp Corynosoma sp. Rhabdinorhynchidae gen. sp Rhabdinorhynchidae gen. sp. Rhabdinorhynchidae gen sp Cestoda Cestoda Cestoda Cestoda Cestoda Tetraphyllidea Tetraphyllidea Tetraphyllidea Tetraphyllidea Tetraphyllidea Copepoda Copepoda Copepoda Copepoda Copepoda N. robusta H spinulus Caligus sp Caligus sp. Caligus sp. C. parva Lepeophtheirus sp. H. spinulus C. pinguis C. pinguis N occidentalis Lepeophtheirus sp. H. spinulus H. spinulus S. arcticus Lepeophtheirus sp. N. occidentalis S. arcticus Lepeophtheirus sp. N. occidentalis S. arcticus Digenea Digenea Digenea Digenea D. cholaeum D cholaeum A macfarlani A macfarlani O. sebastodis L gibbosus D- cholaeum D. cholaeum L exodicum L. gibbosus L gibbosus O sebastodis L exodicum L exodicum P. merus O. sebastodis O sebastodis Podocotyle sp. P. merus P. merus T. lindbergi Podocotyle sp. T lindbergi Podocotyle sp. T lindbergi Monogenea Monogenea Monogenea Monogenea Monogenea M sebastis M sebastis M sebastis M. sebastis M. sebastis /V girellae N. girellae N. girellae N. girellae T. margmata T. margmata T. marginata Nematoda Nematoda Nematoda Nematoda Contracaecum sp Anisakis sp Anisakis sp. Anisakis sp. H. aduncum Contracaecum sp. Contracaecum sp Contracaecum sp. H. aduncum H. aduncum Phocanema sp H aduncum Phocanema sp Protozoa (Myxozoa) Protozoa (Myxozoa) Protozoa (Myxozoa) Protozoa (Myxozoa) H sebasta C sebasta C. sebasta C. sebasta H. sebasta D reginae D reginae K clupeidae H sebasta H. sebasta L. informis K. clupeidae K. clupeidae L longipes L informis L. informis L. macrospora L longipes L macrospora M. incurvatum L. macrospora L sebasta Z ilishae L. sebasta M. incurvatum Z ilishae M. incurvatum Z. ilishae Protozoa (Microsporida) Protozoa (Microsporida) Protozoa (Microsporida) Total number of specimens: 5 11 29 35 34 533 O100 E 90 U 80 u. 70 Z 60 H 50 5 40 U 30 cc UJ Q. 20 10 OlOOr £ 90 0 80 ff 70 1 60 (- 50 z 40 O 30 £ 20 Q. 10 gioo u in 90 80 70 £ 60 \- 50 g 40 y 30 2 20 °> 10 H. aduncum 510 1620 26 30 36 40 TOTAL LENGTH (CM) M.sebastis 46 510 1620 2630 3640 TOTAL LENGTH CM) 46 Lepeophtheirus sp. 510 1620 2630 3640 TOTAL LENGTH (CM) INFECTED CD -vl CD (£>0 OOOOO h- 50 5 40 O 30 85 20 E 10 -4T q100 UJ 90 O 80 £ 70 Z 60 t 50 5 *oi- (3 30 £ 20 Q. 10 O100 £ 90 O 80 £ 70 Z 60 K 50 Z 40 ui ^v' O 30 £ 20 Q- 10 q100 UJ 90 O 80 £ 70 Z 60 H 50 40 30 20 10 UJ UJ o. N. girellae ./' 510 1620 TOTAL 2630 3640 LENGTH (CM) Myxozoa sp. 46 / / 5-10 1620 2630 3640 TOTAL LENGTH (CM) H. sebasta 46 / / ./ 510 H. spinulus 1620 2630 3640 TOTAL LENGTH (CM) 46 1620 2630 36 40 TOTAL LENGTH (CM) 46 FIGURE 2. — The relationships between host length and percent prevalence of infection by seven parasite species from olive rockfish taken off Diablo Cove, Calif. All relationships show significant difference at P =£ 0.05. See Table 3 for numbers offish examined in each length interval; see Table 1 for the number of specimens per size class. ward fish and away from zooplankton (Love and Westphal 1981) probably accounts for the increase in Hysterothylacium aduncum infections, as fish are thought to be intermediate hosts for this species (Margolis 1970). The prevalence of Clauella parva was the opposite — it was found only in hosts <10 cm in length. Clauella parva attaches to dorsal, anal, and caudal fin rays. Perhaps structural barriers (such as ray diameter or surface characteristics) or increased water flow over the fins in larger fish prevent infection. Simi- lar infection patterns were noted in Sebastes alutus and S. caurinus by Sekerak (1975). Among species with seasonal patterns of infec- tion, winter maximum infections were exhibited by Lepeophtheirus sp., Holobomolochus spinulus, the gallbladder myxozoans, and Deretrema cholaeum. The first three forms listed have direct life cycles. Olive rockfish are winter and early spring spawners (December-March) with internal fertilization occurring from November to Feb- ruary. It is possible that these parasites time their 534 O100r P 90 ft 80 u! 70 ? 60 t- 50 5 40 O 30 £ 20 Q- 10 OlOOr £ 90 O 80 £ 70 Z 60 H 50 Z 40 O 30 gj 20 Q- 10 O.100 uj 90 O 80 £ 70 Z 60 t- 50 g 40 O 30 £ 20F o. 10 H. aduncum M J MJJASONDJF MONTH Lepeophtheirus sp. \ /"' ■ MJJASONDJF MONTH q100 UJ 90 O 80 £ 70 Z 60 . 50 Z 40 S 30 a. 20 q. TO J A S O N D J MONTH F Q 0. sebastodis 1- U UJ u. z "\ 1- z UJ o UJ a. 100r 90 80 70 60 50 40 30 20 10 Q100 l= 90 ft 80 £ 70 - 60 \- 50 £ 40 O30 S 20 °- 10 D. choleum \y y ^__ i___^ AMJJASONDJF MONTH Myxozoa sp. X M J J A S 0 N MONTH Hemiundae D J / \ MJJASONDJ MONTH lOO O 90 £80 H. spinulus ft 70 u. 60 \ ./ * 2 50 V ^ ^ / !r 40 5 30 fc> 20 ft 10 0- M J J A S O MONTH FIGURE 3. — The relationships between month of capture and percent prevalence of infection by seven parasite species from olive rockfish taken off Diablo Cove, Calif. All relationships show significant differences at P =s 0.05. See Table 1 for the number of specimens per month. movements and reproduction to coincide with that period when their hosts may be at closest proxim- ity with each other. This phenomenon was ob- served between the Monogenea, Dactylogyrus vas- tator and Mazocraes alosae, and their respective hosts, Cyprinus carpio and Alosa sapidissima (Kennedy 1975). Parasites with maximum prevalence during other periods all had indirect life cycles. The hemiurid trematode infections peaked in autumn, Opechona sebastodis in summer, and Hys- terothylacium aduncum was most abundant in April and August. Seasonality among hemiurids has been reported by Shotter (1973) in whiting, Odontogadus merlangus, of the Irish Sea and in staghorn sculpin, Leptocottus armatus, from Ore- gon by Burreson and Olson (1974). In both cases infections were greatest in late summer or early fall. The infection patterns we observed may reflect differences in the oceanographic conditions off central California. Water conditions in this region 535 may be divided into two periods (Bakun 1973), "upwelling" (March-August) and "oceanic" (September-February). Upwelling periods are characterized by an increase in the flow of nutrient-rich bottom water to the surface and in- creased plankton abundance. Olive rockfish food habits change with these seasons (Love and Westphal 1981). Zooplankton (particularly pelagic tunicates and euphausids), squids, and juvenile rockfish are eaten in greater quantity during the upwelling season, due to increased availability. The secondary intermediate hosts for hemiurids and Opechona sebastodis are planktonic (Shotter 1973; Yamaguti 1971), and the prevalence increase may be due to greater seasonal predation on the planktonic intermediate host. Similarly, as fish are possible intermediate hosts of Hystero- thylacium aduncum, its infection pattern may re- flect the rockfish's heavy predation on juvenile rockfish during the upwelling period. All of the parasite species infecting olive rock- fish infect at least some other rockfish species. The genus Sebastes has exhibited an explosive radia- tion in the northeast Pacific (Kabata 1970). This rapid speciation has occurred relatively recently, probably during and after the Miocene1. Despite extreme morphological and behavioral differences between species, there is little species specificity among rockfish parasites, perhaps because of the rapidity of the host speciation events. Eighty-nine adult metazoan parasites have been reported from rockfishes between Alaska and California (Love and Moser 1983). Of these, 30 species have been found to infect only Sebastes spp. and 10 of the 89 species were unique to one host species. Some of this specificity may be a function of host behavior or habitat preference rather than physiological differences (Kennedy 1975) between rockfish hosts. Holmes (1971) found that a rock- fish's proximity to rocky reefs influenced the prev- alence of the digenetic trematodes Psettarium sebastodorum and Aporocotyle macfarlani. Aporocotyle macfarlani was found in species as- sociated with inshore reefs, P. sebastodorum in those hosts living away from rocks or in deeper waters. Olive rockfish, limited to relatively shal- low reefs, occasionally harbored A. macfarlani but was not infected with P. sebastodorum . Some parasites, particularly ectoparasites, are widespread among rockfishes. For example, Mi- crocotyle sebastis has been reported from 22 1 W. Eschmeyer, California Academy of Science, Golden State Park, San Francisco, CA, pers. commun. 1982. species, Naobranchia occidentalis from 13, Neo- branchiella robusta from 21, and Chondr acanthus pinguis from 19. These and other parasites exhibit an extensive latitudinal range, considerably longer than some of their hosts. Further surveys of those species only lightly studied (such as Sebastes eos, S. melanostomus, S. mystinus, S. rastrelliger, and S. semicinctus) will show that some of those parasites infect nearly all rockfish species. Acknowledgments We thank Mike Moser and Larry Jensen for their assistance in parasite identification. Dave Behrens, Richard Bray, Craig Fusaro, Robert Henderson, Lew Halderson, Ralph Larson, Dave Laur, Gerry Robinson, and Jerry Wellington as- sisted in collecting specimens. We are also indebt- ed to Waheedah Muhammad for typing the final manuscript. Literature Cited Bakun, a. 1973. Coastal upwelling indices, west coast of North America, 1946-71. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-671, 103 p. BURRESON, E. M., AND R. E. OLSON. 1974. Seasonal variations in the populations of two hemiurid trematodes from the Pacific staghorn scul- pin, Leptocottus armatus Girard, in an Oregon estuary. J. Parasitol. 60:764-767. CRESSEY, R. F. 1969. Five new parasitic copepods from California inshore fish. Proc. Biol. Soc. Wash. 82:409-428. DAILEY, M. D., L. A. JENSEN, AND B. W. HILL. 1981. Larval anisakine roundworms of marine fishes from southern and central California, with comments on public health significance. Calif. Fish Game 67:240-245. HOBSON, E. S. 1971. Cleaning symbiosis among California inshore fishes. Fish. Bull., U.S. 69:491-523. HOBSON, E. S., AND J. R. CHESS. 1976. Trophic interactions among fishes and zooplankters near shore at Santa Catalina Island, California. Fish. Bull, U.S. 74:567-598. HOLMES, J. C. 1971. Habitat segregation in sanguinicolid blood flukes (Digenea) of scorpaenid rockfishes (Perciformes) on the Pacific Coast of North America. J. Fish. Res. Board Can. 28:903-909. Kabata, Z. 1970. Some Lernaeopodidae (Copepoda) from fishes of British Columbia. J. Fish. Res. Board Can. 27:865-885. KENNEDY, C. R. 1975. Ecological animal parasitology. Wiley, N.Y., 163 p. LOVE, M. S., AND A. W EBELING. 1978. Food and habitat of three switch-feeding fishes in the kelp forests off Santa Barbara, California. Fish. Bull., U.S. 76:257-271. 536 LOVE, M. S., AND M. MOSER. 1976. Davisia reginae sp. n. (Protozoa: Myxosporidai from four California marine fishes. J. Parasitol. 62:982-983. 1983. A checklist of parasites of California, Oregon, and Washington marine and estuarine fishes. U.S. Dep. Commer, NOAA Tech. Rep. NMFS SSRF-777, 576 p. LOVE, M. S., AND W. V. WESTPHAL. 1981. Growth, reproduction, and food habits of olive rock- fish, Sebastes serranoides, off central California. Fish. Bull., U.S. 79:533-545. MARGOLIS, L. 1970. Nematode diseases of marine fishes. In S. F. Sniezko (editor*, A symposium on diseases of fishes and shellfishes, p. 190-208. Am. Fish. Soc. Spec. Publ. 5. MOSER, M., AND M. S. LOVE. 1975. Henneguya sebasta sp. n. (Protozoa, Myxosporidai from California rockfish, Sebastes spp. J. Parasitol. 61:481-483. MOSER, M., M. S. LOVE, AND L. A. JENSEN. 1976. Myxosporida (Protozoa) in California rockfish, Sebastes spp. J. Parasitol. 62:690-692. SEKERAK, A. D. 1975. Parasites as indicators of populations and species of rockfishes {Sebastes: Scorpaenidae) of the Northeastern Pacific Ocean. Ph.D. Thesis, Univ. Calgary, Calgary, Alberta, 251 p. SHOTTER, R. A. 1973. Changes in the parasite fauna of whiting Odonto- gadus merlangus L. with age and sex of host, season, and from different areas in the vicinity of the Isle of Man. J. Fish Biol. 5:559-573. Turner, C. h., e. e. ebert, and r. r. given. 1969. Man-made reef ecology. Calif. Dep. Fish Game, Fish Bull. 146, 221 p. YAMAGUTI, S. M. 1971. Synopsis of digenetic trematodes of vertebrates. Keigaku Publ. Tokyo, 1074 p. Milton S. Love kimberly shriner Pamela morris VANTUNA Research Group Department of Biology Occidental College Los Angeles, C A 90041 SENSITIVITY OF THE POPULATION GROWTH RATE TO CHANGES IN SINGLE LIFE HISTORY PARAMETERS: ITS APPLICATION TO MYA ARENARIA (MOLLUSCA:PELECYPODA) The question of sensitivity analyses in demo- graphic studies was first addressed by Lewontin (1965), and since that time, Hamilton (1966), Demetrius (1969), Emlen (1970), Goodman (1971), Keyfitz (1971), and Mertz (1971) have made con- tributions in the area. More recently, Caswell (1978) has given general formulae for the sensi- FISHERY BULLETIN: VOL. 82, NO. 3, 1984. tivity of the population growth rate (A) derived from a Leslie model, to changes in single life history parameters written as formulae involving eigenvectors of the Leslie matrix. The application of such analyses to the study of the population dynamics of commercially important species can provide useful information to those interested in resource management. The work presented here describes the sensitiv- ity of the population growth rate to changes in the settlement rate (Brousseau et al. 1982) and in the age-specific fecundity and survivorship rates of the soft-shell clam, Mya arenaria, using a mod- ified Leslie matrix model and an extension of the sensitivity formulae derived by Caswell (1978). Predictions concerning the effect that changes in these life history parameters will have on K and the implications of these results to the manage- ment of this species are discussed. Results Leslie Model The population of females is divided into n age classes. The Leslie matrix, M, has the following form: M = ax a-2 a3 . . . On-i an rsb\ 0 0 . .. 0 0 0 b2 0 . .. 0 0 0 0 0 bn 0 (1) Here, ax is the mean number of female eggs pro- duced annually by a female in class i (age i - 1 to i); assuming a one-to-one sex ratio, a; is one- half the total egg production. The parameter bt is the probability of a clam in class i surviving to class 2, 3, ..., n - 1. The survivorship from age class 1 to age class 2 is divided into 2 factors, rs and bi. The factor rs is the settlement rate or the probability that an egg will survive the plank- tonic larval stage and develop into a clam with a 2 mm shell length (0-2 mo of age); 6i is the prob- ability that a clam with a 2 mm shell length will survive the remainder of the year (about 10 mo). If x is a column vector with n components such that Xi is the number of females in age class i immediately following spawning, then Mx rep- resents the population 1 yr from now. For benthic marine invertebrates possessing planktotrophic larval stages, the events surround- ing metamorphosis and settlement are extremely 537 important for the development and maintenance of species populations. Settlement rates are no- toriously variable in nature, however, making it impossible to determine a fixed rs and hence a fixed A . For this reason, we have studied sensitiv- ities over a range of A's (0.25-3.0). Values of a, and bi used for My a arenaria are empirically derived (Brousseau 1978a, b). Sensitivity Formulae The population growth rate, A, is the eigen- value of M with maximum modulus. In general, A is unique and is a positive real number. This follows from the Perron-Frobenius Theorem, which may be referenced, for example, in Deme- trius (1969). The sensitivity of A to a life history parameter is defined to be the derivative of A with respect to that parameter. Following Caswell (1978), let u and v be col- umn vectors satisfying Mv = kv u'M = ku' (u, v) = 1 (2) (3) (4) where u' denotes the transpose of u, and (•,) denotes inner product. Statement (2) indicates that v is a right eigenvector, while Statement (3) indicates that u is a left eigenvector, each asso- ciated with A. Statement (4) is used as a normal- ization device. While Statements (2) through (4) do not define u and v uniquely, they are sufficient to make the sensitivity formulae below well de- fined. Explicit calculations for the components of vectors u and v start with «i ut= 1 ajbj-1...bik-(J-i+1\i>l (5) and y, =1 vi = k I6j-1...61rs, i > 1 (6) and then normalize using Statement (4) above. With these definitions, Caswell (1978) shows dkldmij = uivj, ij = l...n, (7) where mij is the parameter in the ij position of 538 the Leslie matrix M , ut is the itb. component of vector ui , and vj is the jth component of vector v. Of course, the components of M of interest to us are those in the first row (the fecundity param- eters) and those in the main subdiagonal (the survivorship parameters). Further, since position m2i equals rsbx in our notation, the sensitivity formulae become dkldai = UiUi , i = l,2,...,n (8) dkldbt = Ut + ivt , i — 2,3,..., n - 1 (9) dk/dbi = rsdk/dm2i = rsu2Vi (10) dkldrs = bidk/dm2i = biU2V\ (11) In particular, notice that A is not equally sensitive to rs and bx unless the two values are equal. For the present study, we hold at and bt fixed and allow A to vary. The settlement rate, rs, then becomes a function of A, specifically, rs = (A - ax )/(A~ 1 a2bi + k~2 a3b2bi + ... +kn+1 dnbn-i ... bi), (12) and is used in the Leslie matrix, M. We then compute u and v satisfying Statements (2)-(4) for the given A , and the sensitivity values Statements (8)-(ll). Relationships among the sensitivity formulae above have been derived by Demetrius (1969) and Caswell (1978). Of particular interest are dkldai > dk/daj , i 1 (13) dkldai < dk/daj , i bj d A Idbj , i < j (14) dk/dbi A - ai d A Ida i b i (15) Statement (13) can actually be made stronger, as proven by Demetrius (1969, Statement (8) ); State- ment (14), in the case i = 1, and Statement (15) follow from Demetrius (1969, Statement (11) ) and Caswell (1978, Statement (22)); and Statements (5), (10), and (12) above. Calculation of Sensitivity Values Settlement Rate.— Using the data in Table 1, the sensitivity of the population growth rate of TABLE 1. — Life history statistics used in the derivation of the Leslie matrix for Mya arenaria (data from Brous- seau 1978a, b). Age (yr) Age class Shell length (mm) Fecundity1 (a/) Probability of survival {bi) 0-1 1 2.0-299 0.0 0.177 1-2 2 30.0-44.9 3.744 0 0.912 2-3 3 45.0-59.9 17.170.0 0.904 3-4 4 600-64.9 31.1590 0952 4-5 5 65.0-699 39,957.0 0 949 5-6 6 70 0-74.9 50,341 0 0 969 6-7 7 75.0-79.9 62.450.0 0 984 7 + 8 + 800-84.9 76,465.0 0.911 'Fecundity = number of female eggs produced per individual assuming a 1:1 sex ratio. Mya arenaria to changes in the settlement rate (rs) can be calculated. Results are summarized in Table 2 for a range of values of A. . As expected, rs increases as the population growth rate increases, while the sensitivity of A to changes in rs de- creases as A increases. The population growth rate, A. , is far more sensitive to changes in rs than it is to changes in any other single life history parameter. A further discussion of this point is given below. TABLE 2. — Sensitivity of various population growth rates (A.) to changes in the settlement rate (rs). The intrinsic growth rate = log A . Population growth rate Intrinsic Settlement rate Sensitivity of (a) growth rate (rs) A to rs 025 -1.386 6.535 ■ 10~'2 3.291 x 109 0.5 -0.693 1.700 x 10"e 2.697 • 106 075 -0.288 1.139 x 10 ~6 6 822 ■ 10" 1.0 0.0 '1.462 ■ 10 ~5 8.562 > 103 1.25 0.223 7.320 • 10 "5 2.646 ■ 103 1.5 0405 2 141 • 10"4 1.307 ■ 103 1.75 0.560 4.618 *■ 10~4 8.158 x 102 2.0 0.693 8.310 x 10"4 5.765 ■ 102 3.0 1 099 3.702 ■ 10"3 2.454 x 102 Vs = Equilibrium settlement rate, rseq Fecundity and Other Survivorship Rates. — The sensitivity of A. to changes in fecundity are illus- trated in Figure 1. Under equilibrium conditions (A = 1.0), sensitivity to fecundity changes over the reproductive life span of the individual are slight. If the population is actually growing (A > 1.0), the magnitude of the sensitivity to changes in the fecundity decreases with increasing age, while the reverse is true if the population is actually declining (A < 1.0). This follows from Statement (13). For declining populations this is probably due to the combined effects of an increasing reproductive value with increasing age and a shift in the age structure to older individuals as the population declines. The sensitivity of A to changes in survivorship parameters other than rs is illustrated in Figure 2, where it is evident that A is more sensitive to changes in bi, the survivorship of an individual from 2 mo to 1 yr of age, than to other values of bi for i > 1. As above, these curves illustrate a general result. Since bi is >6i for i > 1 in the Mya arenaria model, dkldbi is >dkldbi using State- ment (14). By comparing Figures 1 and 2, it seems evident that the population growth rate is more sensitive to changes in survivorship than to changes in fecundity. This result may be made precise if the population is actually growing (A > 1), since using Statements (13) and (15) it follows that dk/dbi is >dk/da{ for all values of i. Finally, by examining Statements (10) and (11), it is clear that A is more sensitive to rs than to b\ for the Mya arenaria model as long as rs is <6i. Hence, A is more sensitive to rs than to all other survivorship parameters, and, at least for growing populations, more sensitive to changes in rs than to any other fecundity parameter as well. Discussion Fisher (1958) in his fundamental theorem of FECUNDITY -4- X=3.0 -5- -6- -7- _ \ = 1.0 O ® -8- o 8 ■J _ X=0.75 SENSITIVITY o o •r X=0.5 -11- -12 / X0.25 3 4 5 6 7 8 9 AGE-CLASS 10 11 12 FIGURE l.— Sensitivity of a range of As (0.25-3.0) to changes in the fecundity (a,) of Mya arenaria in each age class. The first age class is not included since Mya arenaria are not mature until after the first year of age. 539 natural selection states that natural selection will favor genotypes which increase the popula- tion growth rate, A . Since the A for a population is based on the life history parameters of age- specific fecundity and survivorship, the greater the sensitivity of A to changes in a particular life history value, the greater the potential for effect- ing evolutionary change through that parameter. Existing evidence indicates that population growth rate is more sensitive to changes in sur- vival rates than to changes in reproductive out- put. Cole (1954) reached this conclusion when he suggested that in species with repeated repro- duction and relatively large litter size, there is little selection pressure favoring increased fe- cundity. Similarly, Caswell (1978) using Harts- horn's (1975) data for Pentaclethra macroloba, a tropical rain forest tree, illustrates by the use of models, that, for this species, population growth rate is more sensitive to changes in growth and survival than to changes in fecundity. The data reported here for Mya arenaria follow the same pattern; A is relatively insensitive to changes in egg production. The most interesting results of the sensitivity analyses are produced by changes in the survivorship parameters. SURVIVORSHIP Based on our analyses, two important general- izations can be made regarding the sensitivity of A. First, whenever rs < bi < bi, the population growth rate will be more sensitive to changes in rs than to changes in any of the other survivor- ship parameters. Second, in growing populations, i.e., where A > 1, A is always most sensitive to changes in the settlement rate. In terms of Dee- vey's (1947) categorization of generalized survi- vorship curves, the relationship rs < 6i < bi is probably operative in most Type III curves, which are characterized by extremely heavy mortality early in life. Consequently, these generalizations are of interest since the types of life history features exhibited by Mya arenaria are likely to be common to other species of marine organisms, many of which are also commercially important. On a more practical level, the ability to identify those life history stages to which the population growth rate is most sensitive may serve as a use- ful tool in directing the efforts of those interested in shellfish management. For instance, the mod- els described here indicate that larval settlement is the most critical stage in Mya arenaria's life history. Developing a better understanding of the factors surrounding metamorphosis and settle- ment and implementing a method for inducing spatfall would probably be the single most effec- tive way to increase clam yields. Another area for consideration centers around the survivorship of the first year class. Since the population growth rate of Mya arenaria is also very sensitive to changes in the b\ parameter, a second way to increase clam productivity is to improve the survivorship of clams 2 mo to 1 yr of age (ca. 2-25 mm shell length). This age class corresponds to that postlarval stage in Mya arenaria which is the most vulnerable to both biotic (predation) and abiotic (wash-out, temper- ature and salinity fluctuations) factors in the environment. The practice of transplanting juvenile Mya arenaria from one flat to another has been used by managers since the turn of the century (Bel- ding 1930) in efforts to 1) replenish depleted clam beds or 2) reduce densities in "overcrowded" beds. Currently, there is a renewed interest in this procedure1 even though in the past, these efforts have met with varying degrees of success (Bel- ding 1930; Turner 1951; Smith et al. 1955). The FIGURE 2.— Sensitivity of a range of X's (0.25-2.0) to changes in the survivorship (ft,- ) of Mya arenaria in each age class. 'D. E. Wallace, Director, Department of Marine Resources, State of Maine, Boothbay Harbor, ME 04538, pers. commun. April 1983. 540 reasons may be related to the vulnerability of juveniles as discussed above. To insure success with transplanting tech- niques, it is essential to reduce mortality among transplanted clams either by protecting them from significant sources of mortality in the field or by retaining them in protective "nurseries" until they pass this critical phase. Past attempts to protect juveniles in the field by building fences to exclude predators have proven costly, difficult to carry out, and unreliable (Smith et al. 1955). More promising are recent advances in aqua- culture techniques for commercially important bivalves. By employing "nurseries" for the young and field "grow-out" procedures for adults, sources of juvenile mortality can be reduced while still utilizing natural sources of food during the greater part of the individual's growth period. Literature Cited BELDING. D. L. 1930. The soft-shelled clam fishery of Massachusetts. Mass. Dep. Conserv., Div. Fish Game, Mar. Fish. Serv. 1, 65 p. BROUSSEAU, D. J. 1978a. Spawning cycle, fecundity, and recruitment in a population of soft-shell clam, Mya arenaria, from Cape Ann, Massachusetts. Fish. Bull., U.S. 76:155-166. 1978b. Population dynamics of the soft-shell clam Mya arenaria. Mar. Biol. (Berl.) 50:63-71. BROUSSEAU, D. J., J. A. BAGLIVO, AND G. E. LANG, JR. 1982. Estimation of equilibrium settlement rates for ben- thic marine invertebrates: its application to Mya are- naria (Mollusca: Pelecypoda). Fish. Bull., U.S. 80: 642-644. Caswell. H. 1978. A general formula for the sensitivity of population growth rate to changes in life history parameters. Theor. Pop. Biol. 14:215-230. COLE, L. C. 1954. The population consequences of life history phe- nomena. Q. Rev. Biol. 29:103-137. DEEVEY, E. S., JR. 1947. Life history tables for natural population of ani- mals. Q. Rev. Biol. 22:283-314. DEMETRIUS, L. 1969. The sensitivity of population growth rate to pertur- bations in the life cycle components. Math. Biosci. 4: 129-136. EMLEN, J. M. 1970. Age specificity and ecological theory. Ecology 51: 588-601. FISHER, R. A. 1958. The genetical theory of natural selection. 2d ed. Dover, N.Y.,291p. GOODMAN, L. A. 1971. On the sensitivity of the intrinsic growth rate to changes in the age-specific birth and death rates. Theor. Pop. Biol. 2:339-354. Hamilton, W. D. 1966. The moulding of senescence by natural selection. J. Theor. Biol. 12:12-45. HARTSHORN, G. S. 1975. A matrix model of tree population dynamics. In F. B. Golley and E. Medina (editors), Tropical ecological systems: trends in terrestrial and aquatic research, p. 41-51. Springer-Verlag, N.Y. KEYFITZ, N. 1971. Linkages of intrinsic to age-specific rates. J. Am. Stat. Assoc. 66:275-281. LEWONTIN, R. C. 1965. Selection for colonizing ability. In H. G. Baker and G. L. Stebbins (editors), The genetics of colonizing species, p. 77-94. Acad. Press, N.Y. MERTZ, D. B. 1971. Life history phenomena in increasing and decreas- ing populations. In G. P. Patil, E. C. Pielou, and W. E. Waters (editors), Statistical ecology, Vol. 2, p. 361-399. Penn. State Univ. Press, University Park, Pa. SMITH, O. R., J. P. BAPTIST, AND E. CHIN. 1955. Experimental farming of the soft-shell clam, Mya arenaria, in Massachusetts, 1949-1953. Commer. Fish. Rev. 17(61:1-16. TURNER, H. J. 1951. Fourth report on investigations on the shellfisheries of Massachusetts. Mass. Div. Mar. Fish. (Woods Hole Oceanogr. Inst. Coll. Reprints, Contrib. No. 615.) DIANE J. BROUSSEAU Department of Biology Fairfield University Fairfield, CT 06430 Department of Mathematics Fairfield University Fairfield, CT 06430 JENNY A. BAGLIVO THE OCCURRENCE OF PISCINE ERYTHROCYTIC NECROSIS (PEN) IN THE SEA LAMPREY PETROMYZON MARINUS, FROM SEVERAL MAINE LOCALITIES The sea lamprey, Petromyzon marinus, is an anadromous fish found in the North Atlantic Ocean from Iceland and northern Europe to northwestern Africa, and from the Grand Banks and the Gulf of St. Lawrence to Florida (Hubbs and Lagler 1949). The sea lamprey has adopted an entirely freshwater life cycle in the Great Lakes where it has seriously depleted fish popu- lations (Everhart 1976). The lamprey feeds on other fishes by hanging on with its sucking mouth. Once attached, it FISHERY BULLETIN: VOL. 82, NO. 3, 1984. 541 rasps its victim with its tongue to obtain nourish- ment in the form of blood and other body fluids (Everhart 1976). Secretions from a pair of rela- tively large salivary glands below the tongue retard coagulation of host blood and also dissolve tissue (Lagler et al. 1977). Bigelow and Schroeder (1953) reported that in saltwater, lampreys have been found preying on mackerel, shad, cod, had- dock, pollock, salmon, basking sharks, various anadromous herrings, swordfish, hake, sturgeons, and eels. Piscine erythrocytic necrosis (PEN), a condi- tion characterized by cytoplasmic inclusions and nuclear abnormalities in erythrocytes, has been shown to be of viral etiology in the Atlantic cod, Gadus morhua, and Atlantic herring, Clupea harengus harengus, from the Atlantic coast and the chum salmon, Oncorhynchus keta, pink salm- on, O. gorbuscha, and Pacific herring, Clupea harengus pallasi, from the Pacific coast of North America (Walker 1971; Appy et al. 1976; Walker and Sherburne 1977; Philippon et al. 1977; Reno et al. 1978; Evelyn and Traxler 1978; MacMillan and Mulcahy 1979). In addition, PEN has been reported in 15 other marine teleost species from the Atlantic coast of North America, but confir- mation as to viral etiology has not been made (Laird and Bullock 1969; Walker and Sherburne 1977; Sherburne 1977; Sherburne and Bean 1979). PEN has also been evident in the Atlantic mack- erel, Scomber scombrus, (Sherburne, unpubl. data). This report documents the first finding of PEN in a host from the most primitive group of fishes, the Agnatha. Materials and Methods A total of 142 lampreys, Petromyzon marinus, was obtained for blood analysis from 5 Maine localities (Table 1). Live lampreys were measured for total length and sexed. Slides were prepared by severing the caudal peduncle and taking the blood into a heparinized capillary tube, from which a small drop of blood was placed on a microscope slide and the smear made. Air-dried smears were Giemsa-stained and thoroughly examined for PEN using light microscopy at 1000 x magnification. Results Of the total lampreys sampled in this study, 50.7% (72/142) had red cell lesions characteristic of PEN (Table 1). By light microscopy, PEN lesions of lamprey red cells often showed the nuclear chromatin condensed into round blebs, and there was evidence of nuclear vacuolization (Fig. 1). Red acidophilic cytoplasmic inclusions were occasionally seen in an infected cell (Fig. 2). Individual infections were light, with only one or two infected cells evident in most smears. Among the 72 infected lampreys, the severest infection involved 2% of the red cells and it occurred in a 69 cm male from the Coopers Mills Fishway on 26 May 1980. From a total of 139 lampreys sexed, 47.9% of the males and 53.0% of the females had PEN. The smallest infected lamprey was 62.4 cm (24.6 in) long; the largest was 81.5 cm (32.1 in). TABLE 1. — The occurrence of piscine erythrocytic necrosis (PEN) in the sea lamprey, Petromyzon marinus, from several Maine localities. Location Date PEN Mean length, SC range (cm) of sa Sample source Incidence in sample Percent incidence i, and mple Nequasset Lake Fishway Woolwich 16 June 1977 0/2 0.0 67.0 ± 1.4(66.0 - 68.0) Kennebunk River Kennebunk 11 May 1978 1/1 100 63.1 (63.1) Sheepscot Pond Fishway Palermo 9 June 1978 1/3 33.3 65.4 ± 2.7 (62.5 - 67.8) Sheepscot Pond Fishway Palermo 15 June 1978 0/1 0.0 65 (65) Sheepscot River Coopers Mills Fishway Coopers Mills 30 May 1979 5,28 17.8 72.4 ± 4.1 (62.4 - 82) Sheepscot River Coopers Mills Fishway Coopers Mills 26 May 1 980 34/46 73 9 72.4 ± 4.6 (66 - 88.5) Sheepscot River Head Tide Alna 6 June 1 983 7/18 38.8 71.4 ±4.3(64.8 - 78.8) Sheepscot River Coopers Mills Fishway Coopers Mills 13 June 1983 1/4 25.0 73.5 + 2.3 (70.0 - 75.0) Sheepscot River Head Tide Alna 14 June 1983 5/8 62.5 70.6 + 3.4 (66.7 - 76.0) Sheepscot River Head Tide Alna 15 June 1983 18/31 58.1 70.5 ± 4.0 (63.5 ■ - 78.0) 542 FIGURE 1. — Sea lamprey erythrocytes with PEN lesions. In- fected cells show characteristic chromatin condensation with evidence of nuclear vacuolization. From a female lamprey 80 cm in total length from the Coopers Mills Fishway, Coopers Mills. Me., on 26 May 1980. Sea lamprey erythrocytes are rounded in shape in contrast to most other fish species which have elliptical shaped red cells. i * * FIGURE 2. — Red acidophilic inclusions are occasionally seen in PEN infected sea lamprey erythrocytes. From a female lamprey 67.8 cm in total length from the Sheepscot Pond Fishway, Palermo. Me., on 9 June 1978. Discussion microscopy. Consequently, viral etiology of the condition still remains to be confirmed. Fish obtained from lakes where alewives spawn have shown typical PEN lesions (Sherburne, un- publ. data), but whether alewives contribute to this is unknown. MacMillan and Mulcahy (1979) reported transferring VEN to chum salmon, Oncorhynchus keta, and brook trout, Salvelinus fontinalis, by waterborne virus. Perhaps infected anadromous species can transmit PEN to fresh- water species via body fluids such as urine and reproductive products as well as by direct contact. Lampreys could conceivably transmit PEN to a variety of marine and freshwater species because of their feeding habits, their diversity of prey, and their ability to become adapted to an entirely freshwater environment. The high prevalence of infection and the low intensity of infection sug- gest that lampreys might readily spread the in- fection without suffering a high mortality rate from PEN. I prefer to use the term PEN in species where the cellular pathology has not yet been confirmed as associated with a virus, and viral erythrocytic necrosis (VEN) after confirmation. By light mi- croscopy, PEN lesions of lamprey red cells resem- ble those of VEN-infected Atlantic cod. As with alewives, Alosa pseudoharengus, (Sherburne 1977) and smelt, Osmerus mordax, (Sherburne and Bean 1979), lampreys have a relatively high percentage of individuals affected with PEN, but individual infections are very light. The blood of the sea lamprey must be examined by electron microscopy to determine if the PEN seen is an ICDV infection. Unfortunately, the individual infections observed in this study were so light as to preclude their detection by electron Literature Cited APPY, R. G., M. D. B. BURT, AND T. J. MORRIS. 1976. Viral nature of piscine erythrocytic necrosis (PEN) in the blood of Atlantic cod iGadus morhua). J. Fish. Res. Board Can. 33:1380-1385. BIGELOW, H. B., AND W. C. SCHROEDER. 1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv, Fish. Bull. 53:1-577. EVELYN, T. R T., AND G. S. TRAXLER. 1978. Viral erythrocytic necrosis: natural occurrence in Pacific salmon and experimental transmission. J. Fish. Res. Board Can. 35:903-907. EVERHART, W. H. 1976. Fishes of Maine. 4th ed. Maine Dep. Inland Fish. Wildl., Augusta, Me., 96 p. HUBBS, C. L., AND K. F. LAGLER. 1949. Fishes of the Great Lakes region. Cranbrook Inst. Sci., Bull. 26, 186 p. 543 lagler, k. e, J. E. Bardach, r. r. miller, and d. r. m. passino. 1977. Icthyology. 2d. ed. John Wiley & Sons, Inc., N.Y., 506 p. Laird, M., and w. l. Bullock. 1969. Marine fish hematozoa from New Brunswick and New England. J. Fish. Res. Board Can. 26:1075-1102. MACMILLAN, J. R., AND D. MULCAHY. 1979. Artificial transmission to and susceptibility of Puget Sound fish to viral erythrocytic necrosis (VEN). J. Fish. Res. Board Can. 36:1097-1101. PHILIPPON, M., B. L. NICHOLSON, AND S. W. SHERBURNE. 1977. Piscine erythrocytic necrosis (PEN) in the Atlantic herring (Clupea harengus harengus ): evidence for a viral infection. Fish Health News 6:6-10. reno, r, m. philippon-fried, b. l. nicholson, and s. w. Sherburne. 1978. Ultrastructural studies of piscine erythrocytic necrosis (PEN) in Atlantic herring {Clupea harengus harengus ). J. Fish. Res. Board Can. 35:148-154. Sherburne, S. W. 1977. Occurrence of piscine erythrocytic necrosis (PEN) in the blood of the anadromous alewife, Alosa pseudo- harengus, from Maine coastal streams. J. Fish. Res. Board Can. 34:281-286. Sherburne, S. W, and L. L. bean. 1979. Incidence and distribution of piscine erythrocytic necrosis and the microsporidian, Glugea hertwigi, in rainbow smelt, Osmerus mordax, from Massachusetts to the Canadian maritimes. Fish. Bull., U.S. 77:503-509. Walker, R., and S. W. Sherburne. 1977. Piscine erythrocytic necrosis virus in Atlantic cod, Gadus morhua, and other fish: ultra-structure and dis- tribution. J. Fish. Res. Board Can. 34:1188-1195. Walker, r. 1971. PEN, a viral lesion of fish erythrocytes. Am. Zool. 11:707. Stuart W Sherburne Maine Department of Marine Resources Fisheries Research Laboratory West Boothbay Harbor, ME 04575 544 ERRATA Fishery Bulletin: Vol. 81, No. 3 Livingston, R A., "Food habits of Pacific whiting, Merluccius productus, off the west coast of North America, 1967 and 1980," p. 629-636. Page 635, Equation (2) should read: R = 0.0416e° -1057: Fishery Bulletin: Vol. 81, No. 4 Grant, W. S., R. Bakkala, F. M. Utter, D. J. Teel, and T. Kobayashi, "Biochemical genetic population structure of yellowfin sole, Limanda aspera, of the North Pacific Ocean and Bering Sea," p. 667-677. Page 672, Figure 2, the top portion should read: ADA -107 * -100 -95 *- 90 K. 0 0) ? s £ *> >o 5 85 05 Ol o> 0 Phenotype origin 05 MDH Position of homodimer 1 'i mm 125 1 -120 _I05 ~I00 „ ME MDH-3 % B .. -200 J heterodirner between loci -100 ] MDH-2 MDH-I 100/100 100/125 100/120 100/105 100/100 100/200 origin Phenotype MDH-3 MDH-2 .<*""' PGD •O IT) Position of homodimer -115 -105 -100 - 95 - 85 ^ O O Q> Phenotype origin 05 Sherman, K., J. R. Green, J. R. Goulet, and L. Ejsymont, "Coherence in zooplankton of a large northwest Alantic ecosystem," p. 855-862. Pages 858 and 889 should read: FISHERY BULLETIN: VOL. 81, NO. 4 150 100 50 O 150 - > 100 < 5 50 150 100 50 Gulf of Maine Georges Bank S. New England Mid-Atlantic Bight JFMAMJ JASOND MONTH 100 50- LU 100 O z < 50 z 5 O 0 Q UJ o CO UJ "' 1 1 ' * :):::::::::i-: Georges Bank •i '•''' t'-"' 'V' • i nV-i'ir i '•'• T ■ ■ i'n nl LW ESp LSp ESu LSu EA LA EW 100 50 S. New England o 100 50 Mid-Atlantic Bight LW ESp LSp ESu LSu EA LA EW SEASON FIGURE 2. — Patterns of zooplankton in four northeastern U.S. continental shelf subareas — Gulf of Maine, Georges Bank, Southern New England, and the Mid-Atlantic Bight, (a) Seasonal patterns in mean zooplankton standing stock (cc/100 m3) for the 5-yr MARMAP time series. Solid line represents the mean, short dashed line is one standard duration, and long dashed line is the range. SHERMAN ET AL.: COHERENCE IN ZOOPLANKTON fjjSfe Pseudocalanus minutus — - — Calanus finmarchicus Centropages typicus Centropages hamatus Penilia avirostris Z other Metridia lucens I, I, Sagitta elegans Balanidae Temora longicornis Acartia sp Calanus sp. Evadne nordmanni Appendicularia Doliolidae Brachyura Echinodermata Tha/iacea io3- Calanus /Inmarchlus Pseudocalanus minutus Centropages typicus E o CO cc 111 m io IO io' IO io- 10' o5h Gulf of Maine i i — i i i i i i i i i '//I ^v% •I \\ 1 ^K \ I I I I I I I I I I I Georges Bank i i i i i i i i i i < 10° Ul 2 io2 - / S. New England ii' i io- io Mid-Atlantic Bight i i i i i i i i i i i i i i i i i i i i i i i i i i i i i J_L i i i i i i I i i I I i i i l i i I ' ' Ill J FMAMJ JASOND JFMAMJ JASOND JFMAMJ JASOND MONTH FIGURE 2. — Continued — (b> Seasonal patterns of dominance of zooplankters by subarea shown as a percentage of the samples with a dominant taxon in the 5-yr MARMAP time series. LW = late winter, ESp = early spring, ESu = early summer, EA = early autumn, LA = late autumn, and EW = early winter, (c) Seasonal pulses in abundance of the three dominant copepod species Calanus fin- marchicus, Pseudocalanus minutus, and Centropages typicus (No./lOO m3) in each of the subareas for the 5-yr time series. Solid line represents the mean, short dashed line is one standard duration, and long dashed line is the range. U S Po«i«l S«rvic« STATEMENT OF OWNERSHIP, MANAGEMENT AND CIRCULATION Required by J9 USC JngS) 1A. TITLE OF PUBLICATION Fishery Bulletin IB PUBLICATION NO 3 6 6 2 DATE OF FILING 3 7 0 I 1 October 1984 3 FREQUENCY OF ISSUE Quarterly 3A NO OF ISSUES PUBLISHED 3B ANNUAL SUBSCRIPTION ANNUALLY PRICE $21.00 COMPLETE MAILING ADDRESS OF KNOWN OFFICE OF PUBLICATION /Street. City. County. State and UP+4 Code) (.Vol printers) Scientific Publications Office, NMFS 7600 Sand Point Way N.E., 8IN C15700 (Bldg. 1), Seattle, WA 98115 5 COMPLETE MAILING ADDRESS OF THE HEADQUARTERS OF GENERAL BUSINESS OFFICES OF THE PUBLISHER (Hot printer. National Marine Fisheries Service, NOAA, DOC 2001 Wisconsin Ave., NW, Washington, DC 20235 6 FULL NAMES AND COMPLETE MAILING ADDRESS OF PUBLISHER. EDITOR. AND MANAGING EDITOR /This Hem MUST HOT be blank I PUBLISHER (Same and Complete Mailing Addressj Jack McCormick, Scientific Publications Office, NMFS 7600 Sand Point Way N.E., Seattle, WA 98115 EDITOR (Name and Complete Mailing Address) Mary Fukuyama, Scientific Publications Office, NMFS 7600 Sand Point Way N.E., Seattle, WA 98115 MANAGING EDITOR (Name and Complete Mailing Addressj Mary Fukuyama, Scientific Publications Office, NMFS 7600 Sand Point Way N.E., Seattle, WA 98115 7. OWNER lift, wned by a corporation, ill name a no address must be stated and also immediately thereunder the names and addresses of stockholders owning or holding I percent or more of total amount of slock If not owned by a corporation, the names and addresses of the individual owners must be given If owned by a partnership or other unincorporated firm, its name and address, as well as that of each individual must be given If the publica- tion u published by a nonprofit organisation, us name and address must be stated I (Item must be completed. I FULL NAME COMPLETE MAILING ADDRESS U.S. Department of Commerce 14th St., NW, Washington, DC 20230 KNOWN BONDHOLDERS. MORTGAGEES. AND Oi HER SECURITY HOLDERS OWNING OR HOLDING 1 PERCENT OR MORE OF TOTAL AMOUNT OF BONOS, MORTGAGES OH OTHER SECURITIES 111 there are none, so stale) FULL NAME COMPLETE MAILING ADDRESS None 9 FOR COMPLETION BY NONPROFIT ORGANIZATIONS AUTHORIZED TO MAIL AT SPECIAL RATES (Section 423.12 DMM onlvl The purpose, function, and nonprofit status of thu organization and the exempt status for Federal income tax purposes iCheck one) H HAS NOT CHANGEO DURING PRECEDING 12 MONTHS □ HAS CHANGED DURING PRECEDING 12 MONTHS (If changed publisher must submit explanation of change with this statement > EXTENT AND NATURE OF CIRCULATION (See instrvc nons on reverse side) AVERAGE NO COPIES EACH ISSUE DURING PRECEDING 12 MONTHS ACTUAL NO COPIES OF SINGLE ISSUE PUBLISHED NEAREST TO FILING DATE A TOTAL NO COPIES (Net Press Run 2216 2241 b paio and/or requested circulation (handled by U.S 1 Sales through dealers and carriers, street vendors and counter sales Govt. Printing 2 Mail Subscription (Paid and or requested! Office, Wash. , DC 20402) C TOTAL PAID AND/OR REQUESTED CIRCULATIONS pri nted /Sum of 10B1 and /OS.' f q r S a 1 e S ) 800 800 D FREE DISTRIBUTION BY MAIL. CARRIER OR OTHER MEANS SAMPLES. COMPLIMENTARY, AND OTHER FREE COPIES ( requested 1416 1441 E TOTAL DISTRIBUTION /Sum of C and D) copiet ) 2216 22A1 F COPIES NOT DISTRIBUTED 1 Office use. left over, unaccounted, spoiled after printing 2. Return from News Agents G TOTAL (Sum of E. Fl and 2-shoutd equal net press run shown in A) ??41 I cartify that the statements made by me above are correct and complete ANAGER. OR OWNER PS Form 3526, Julv 1984 INFORMATION FOR CONTRIBUTORS TO THE FISHERY BULLETIN Manuscripts submitted to the Fishery Bulletin will reach print faster if they conform These are not absolute requirements, of course, but desiderata. to the following instructions. CONTENT OF MANUSCRIPT The title page should give only the title of the paper, the author's name, his affiliation, and mailing address, including ZIP code. The abstract should not exceed one double-spaced page. In the text, Fishery Bulletin style, for the most part, follows that of the U.S. Government Printing Office Style Manual. Fish names follow the style of the Amer- ican Fisheries Society Special Publication No. 12, A List of Common and Scientific Names of Fishes from the United States and Canada, Fourth Edition, 1980. Text footnotes should be typed separately from the text. Figures and tables, with their legends and headings, should be self-explanatory, not requiring reference to the text. Their placement should be indicated in the right-hand margin of the manuscript. Preferably figures should be reduced by photography to 5% inches (for single-column figures, allowing for 507c reduction in printing), or to 12 inches (four double- column figures). The maximum height, for either width, is 14 inches. Photographs should be printed on high quality or glossy paper. Do not send original drawings to the Scientific Editor; if they, rather than the photographic reductions, are needed by the printer, the Scientific Publications Office will request them. Each table should start on a separate page. Consis- tency in headings and format is desirable. Vertical rules should be avoided, as they make the tables more expensive to print. Footnotes in tables should be num- bered sequentially in arabic numerals. To avoid confu- sion with powers, they should be placed to the left of numerals. Acknowledgments, if included, are placed at the end of the text. Literature is cited in the text as: Lynn and Reid (1968) or (Lynn and Reid 1968). All papers referred to in the text should be listed alphabetically by the senior author's surname under the heading "Literature Cited." Only the author's surname and initials are required in the literature cited. The accuracy of the literature cited is the responsibility of the author. Abbreviations of names of periodicals and serials should conform to Serial Sources for the BIOSIS Data Base. (Chemical Abstracts also uses this system, which was developed by the American Standards Association. I Common abbreviations and symbols, such as mm, m, g, ml, mg, °C (for Celsius), ind so forth, should be used. Abbreviate units of measure only when used with numerals. Periods are only rarely used with abbreviations. We prefer that measurements be given in met- ric units; other equivalent units may be given in parentheses. FORM OF THE MANUSCRIPT The original of the manuscript should be typed, double-spaced, on white bond paper. Please triple space above headings. We would rather receive good dupli- cated copies of manuscripts than carbon copies. The sequence of the material should be: TITLE PAGE ABSTRACT TEXT LITERATURE CITED TEXT FOOTNOTES APPENDIX TABLES (Each table should be numbered with an arabic numeral and heading provided) LIST OF FIGURES (Entire figure legends) FIGURES (Each figure should be numbered with an arabic numeral; legends are desired) ADDITIONAL INFORMATION Send the ribbon copy and two duplicated or carbon copies of the manuscript to: Dr. William J. Richards, Scientific Editor Fishery Bulletin Southeast Fisheries Center Miami Laboratory National Marine Fisheries Service, NOAA 75 Virginia Beach Drive Miami, FL 33149-1099 Fifty separates will be supplied to an author free of charge and 50 supplied to his organization. No covers will be supplied. *GPO 593035 1 985 S w \ Sf+ns o« *" Fishery Bulletin "N r Vol. 82, No. 4 October 1984 COLLETTE, BRUCE B., and JOSEPH L. RUSSO. Morphology, systematics, and biology of the Spanish mackerels (Scomberomorus , Scombridae) 545 SHAKLEE, JAMES B., and PAUL B. SAMOLLOW. Genetic variation and pop- ulation structure in a spiny lobster, Panulirus marginatus , in the Hawaiian Archipelago 693 SHAKLEE, JAMES B., and PAUL B. SAMOLLOW. Genetic variation and pop- ulation structure in a deepwater snapper, Pristipomoides filamentosus, in the Hawaiian Archipelago 703 HENDRICKX, M. E. Distribution and abundance of Sicyonia penieillata Locking- ton, 1879 in the Gulf of California, with some notes on its biology 715 Index 721 1985 Notices NOAA Technical Reports NMFS published during the first six months of 1984. Seattle, Washington U.S. DEPARTMENT OF COMMERCE Malcolm Baldrige, Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION John V. Byrne, Administrator NATIONAL MARINE FISHERIES SERVICE William G. Gordon, Assistant Administrator Fishery Bulletin The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. The Bulletin of the United States Fish Commission was begun in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin instead of being issued individually. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. It is also available free in limited numbers to libraries, research institutions, State and Federal agencies, and in exchange for other scientific publications. SCIENTIFIC EDITOR, Fishery Bulletin Dr. William J. Richards Southeast Fisheries Center Miami Laboratory National Marine Fisheries Service, NOAA Miami, FL 33149-1099 Editorial Committee Dr. Bruce B. Collette National Marine Fisheries Service Dr. Edward D. Houde Chesapeake Biological Laboratory Dr. Merton C. Ingham National Marine Fisheries Service Dr. Reuben Lasker National Marine Fisheries Service Dr. Donald C. Malins National Marine Fisheries Service Dr. Jerome J. Pella National Marine Fisheries Service Dr. Jay C. Quast National Marine Fisheries Service Dr. Carl J. Sindermann National Marine Fisheries Service Mary S. Fukuyama, Managing Editor The Fishery Bulletin (ISSN 0090-0656) is published quarterly by the Scientific Publications Office, National Marine Fisheries Service, NOAA, 7600 Sand Point Way NE, BIN C15700, Seattle, WA 98115. Second class postage is paid at Seattle, Wash., and additional offices. POSTMASTER send address changes for subscriptions to Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. Although the contents have not been copyrighted and may be reprinted entirely, reference to source is appreciated. The Secretary of Commerce has determined that the publication of this periodical is necessary in the transaction of the public business required by law of this Department. Use of funds for printing of this periodical has been approved by the Director of the Office of Management and Budget through 1 April 1985. Fishery Bulletin CONTENTS Vol. 82, No. 4 October 1984 COLLETTE, BRUCE B., and JOSEPH L. RUSSO. Morphology, systematics, and biology of the Spanish mackerels (Scomberomorus, Scombridae) 545 SHAKLEE, JAMES B., and PAUL B. SAMOLLOW. Genetic variation and pop- ulation structure in a spiny lobster, Panulirus marginatus, in the Hawaiian Archipelago 693 SHAKLEE, JAMES B., and PAUL B. SAMOLLOW. Genetic variation and pop- ulation structure in a deepwater snapper, Pristipomoides filamentosus , in the Hawaiian Archipelago 703 HENDRICKX, M. E. Distribution and abundance of Sicyonia penicillata Locking- ton, 1879 in the Gulf of California, with some notes on its biology 715 Index 721 Notices NOAA Technical Reports NMFS published during the first six months of 1984. •ne Bfoii Laboratory OCT 30 1985 L WOQds HolflJUaee^J Seattle, Washington 1984 For sale by the Superintendent of Documents. U.S. Government Printing Office, Washing- ton DC 20402 — Subscription price per year: $21.00 domestic and $26.25 foreign. Cost per single issue: $6.50 domestic and $8.15 foreign. The National Marine Fisheries Service (NMFS) does not approve, recommend or endorse any proprietary product or proprietary material mentioned in this pub- lication. No reference shall be made to NMFS, or to this publication furnished by NMFS, in any advertising or sales promotion which would indicate or imply that NMFS approves, recommends or endorses any proprietary product or proprietary material mentioned herein, or which has as its purpose an intent to cause directly or indirectly the advertised product to be used or purchased because of this NMFS publication. MORPHOLOGY, SYSTEMATICS, AND BIOLOGY OF THE SPANISH MACKERELS (SCOMBEROMORUS , SCOMBRIDAE) Bruce B. Collette1 and Joseph L. Russo2 ABSTRACT The Spanish mackerels and seerfishes of the genus Scorn beromorus constitute the most speciose group of the 44 genera in six families that comprise the suborder Scombroidei. As in higher scom- brids, Scomberomorus , Acanthocybium , and Grammatorcynus have a well-developed median keel on the caudal peduncle, but there is no bony support as is present in the Sardini and Thunnini. Acanthocybium and Scomberomorus share 17 osteological characters and are considered sister- groups. The relationships of Grammatorcynus are not clear but it is clearly more primitive than Scomberomorus; therefore, we have used it as the outgroup for a cladistic analysis of Scomberomorus. Scomberomorus differ from all other scombrids in having a spatulate anterior extension of the vomer. There are 18 species in the genus, nearly 40% of the 49 species of scombrids: Eastern At- lantic— tritor (Cuvier); western Atlantic — brasiliensis Collette, Russo and Zavala-Camin, cavalla (Cuvier), maculatus (Mitchill), and regalis (Bloch); eastern Pacific — concolor Lockington and sierra Jordan and Starks; and Indo-West Pacific — commerson (Lacepede), guttatus (Bloch and Schneider), koreanus (Kishinouye), lineolatus (Cuvier), munroi Collette and Russo, multiradiatus Munro, niphonius (Cuvier), plurilineatus Fourmanoir, queenslandicus (Macleay), semifasciatus (Macleay), and sinensis (Lacepede). A cladistic analysis of 58 characters shows six monophyletic species-groups in Scomberomorus. The sinensis group is monotypic and is defined by the presence of an abrupt downward curve in the lateral line under the first dorsal fin and by its retention of a swim bladder. The commerson species-group contains commerson, niphonius, queenslandicus, and cavalla and is defined by the presence of an intercalar spine of at least moderate length. Scomberomorus cavalla and S. commerson share two additional specializations, the pterosphenoid bones are close together and the lateral line curves abruptly downward under the second dorsal finlets. The munroi species-group is monotypic and is defined by the loss of the anterior process on the outer surface of the head of the maxilla. The semifasciatus species-group contains semifasciatus, plurilineatus, and lineolatus, and is defined by the presence of a greatly expanded posterior end of the maxilla. Scom- beromorus lineolatus and S. semifasciatus share an additional specialization, a wide parasphenoid, but this character state appears independently in several other lines. The guttatus species-group contains guttatus, multiradiatus, and koreanus and is defined by a high supraoccipital crest. Auxil- iary branches extend off the anterior part of the lateral line in S. guttatus and S. koreanus. The regalis species-group contains regalis, tritor, maculatus, concolor, sierra, and brasiliensis and is defined by the presence of nasal denticles. All but the most primitive species in this group (S. tritor) have an artery arising from the fourth left epibranchial artery. The four most advanced species (all except tritor and maculatus ) have developed a long posterior process on the pelvic girdle. The three most advanced species (sierra, brasiliensis, and regalis) have a coeliaco-mesenteric shunt connect- ing the fourth right epibranchial artery with the coeliaco-mesenteric artery. The purposes of this paper are to define the 18 species of Scomberomorus, to clarify their rela- tionships, and to assess the systematic position of Scomberomorus within the Scombridae. The methods used are similar to those of Collette and Chao (1975) in a revision of the bonitos and of Gibbs and Collette (1967) in a revision of Thun- 1 National Marine Fisheries Service, Systematics Laboratory, National Museum of Natural History, Washington, DC 20560. National Marine Fisheries Service, Systematics Laboratory, National Museum of Natural History, Washington, DC 20560, and Department of Biological Sciences, The George Washington University, Washington, DC 20006; present address: Office of Information Research Resource Management, Smithsonian Institution, Washington, DC 20560. nus, and rely on previous work by Kishinouye (1923), Munro (1943), Mago Leccia (1958), and Devaraj (1977). The Spanish mackerels have been placed by Collette and Chao (1975) and Collette and Russo (1979) in a tribe (the Scomberomorini) along with Acanthocybium and Grammatorcynus, interme- diate between the more primitive mackerels (Scombrini) and the more advanced bonitos (Sar- dini). Acanthocybium is clearly the specialized sister group of Scomberomorus, but the phylo- genetic position of Grammatorcynus has been unclear. Until recently, the number of valid species of Manuscript accepted November 1983. FISHERY BULLETIN: VOL. 82, NO. 4, 1984. 545 Scomberomorus has been in doubt. In his revision of Australian species, Munro (1943) recognized 15 species in the world (excluding Cybiosarda ele- gans, a bonito, and Lepidocybium flavobrunneum , a gempylid). Fraser-Brunner (1950) recognized only nine species, placing five valid species in synonymy. In the course of this revision, we have discovered two previously undescribed species, S. brasiliensis (Collette et al. 1978), which was con- fused with S. maculatus, and S. munroi (Collette and Russo 1980), which was confused with S. niphonius. Emphasis was placed on obtaining fresh or frozen specimens for dissection from several pop- ulations of each species. Standard counts and measurements were taken, color pattern was recorded, and a search made for parasitic cope- pods. Results of the copepod study have been reported by Cressey and Cressey (1980), and analysis of these data from a host-parasite point of view has been completed (Cressey et al. 1983; Collette and Russo 1985). The viscera were ex- amined and illustrated in situ following remov- al of an oval portion of the ventral body wall. The viscera then were removed and drawings were made of the liver and other selected organs. The kidneys and anterior parts of the arterial system then were drawn. Counts of ribs and intermuscular bones were made and the speci- men was then skeletonized, facilitated by immer- sion in hot water. The base measurement for morphometric com- parisons of fresh, frozen, and preserved specimens was millimeters fork length (mm FL). This paper is divided into three major parts. The first part contains descriptions and illustra- tions of morphometry, meristic characters, soft anatomy, and osteology of the species of Scom- beromorus. Comparisons with Acanthocybium solandri and Grammatorcynus bilineatus are in- cluded. All references to Grammatorcynus in this paper refer to G. bilineatus. The validity of the second species, G. bicarinatus, was only estab- lished recently (Collette 1983). The second part comprises separate species accounts including synonymy, types of nominal species, diagnosis (based on characters from the first section), de- scription, size, color pattern, summaries of pub- lished information on biology and interest to fisheries, geographic distribution, and material examined. The most important references to each species are marked with asterisks in the syn- onymies. The third part is an analysis of the relationships of Acanthocybium and the spe- FISHERY BULLETIN: VOL. 82, NO. 4 cies of Scomberomorus based on a cladistic analysis of characters described in the first part, using Grammatorcynus as the plesiomorphic out-group. MATERIAL The material examined is listed by general lo- cality under four or five headings in the accounts for each of the 18 species of Scomberomorus. Comparative material of Acanthocybium and Grammatorcynus is listed at the end of this section. The numbers under these headings are not additive but are included to give some degree of confidence in the morphological data presented in the body of the paper. "Total specimens" is the total number of individuals examined whether preserved, dissected, or skeletonized. "Dissected" are fresh or frozen specimens for which data on the viscera and usually other characters were recorded. Specimens were subsequently made in- to skeletons. "Measured and counted" includes specimens that were subsequently dissected as well as the preserved museum specimens used for detailed morphometric and meristic examination. "Counts only" are additional museum specimens used only for meristic examination. "Skeletons" refer to all the skeletal material examined, both specimens that were dissected and additional skeletal material already in museums. Asterisks indicate type-specimens of nominal species. Material was examined from the following institutions: AMNH American Museum of Natural His- tory, New York AMS Australian Museum, Sydney ANSP Academy of Natural Sciences, Philadelphia BMNH British Museum (Natural History), London CAS California Academy of Sciences, San Francisco CSIRO CSIRO Marine Biological Laborato- ry, Cronulla, N.S.W., Australia DASF Department of Agriculture, Stock, and Fisheries, Port Moresby, Pap- ua New Guinea FMNH Field Museum of Natural History, Chicago HUMZ Laboratory of Marine Zoology, Hokkaido University, Hakodate, Hokkaido 546 COLLETTE and RUSSO: SPANISH MACKERELS GCRL Gulf Coast Research Laboratory and Museum, Ocean Springs, Miss. LACM Los Angeles County Museum of Natural History, Los Angeles MCZ Museum of Comparative Zoology, Harvard MNHN Museum National d'Histoire Natu- relle, Paris MPIP Museu de Pesca do Instituto de Pes- ca, Santos MSUF Museo de La Specola, Universita di Firenze, Florence MZUSP Museu de Zoologia da Universidade de Sao Paulo, Sao Paulo NHMV Naturhistorisches Museum, Vienna NMC National Museum of Natural Sci- ences, Ottawa QM Queensland Museum, Brisbane RMNH Rijksmuseum van Natuurlijke His- toric, Leiden ROM Royal Ontario Museum, Toronto RUSI J. L. B. Smith Institute of Ichthyolo- gy, Rhodes University, Grahams- town, South Africa SAM South African Museum, Capetown SIO Scripps Institution of Oceanography, La Jolla, Calif. TABL Miami Laboratory (formerly Tropi- cal Atlantic Biological Laborato- ry), NMFS, Miami, Fla. [Most specimens now at UF] UDONECI Universidad de Oriente, Nueva Es- parta, Centro de Investigaciones, Venezuela UF Florida State Museum, University of Florida, Gainesville UMMZ University of Michigan Museum of Zoology, Ann Arbor USNM United States National Museum, Washington, D.C. WAM Western Australia Museum, Perth ZMA Zoological Museum, Amsterdam ZMH Zoologisches Institut und Zoolog- isches Museum, Hamburg ZMK Zoological Museum, Copenhagen ZSI Zoological Survey of India, Calcutta Acanthocybium solandri. — Total 47 (536-1,500 mm FL). meas.: 26 (536-1,500): W Atlantic (8); St. He- lena (1); S. Africa (3); Indian Ocean (4); Caroline Is. (6); Tuamotu Is. (1); E Pa- cific (3). heads: 8 (202-380): Bahama Is. (1); St. Helena (2); Australia (1); Marshall Is. (1); E Pacific (2). counts: 36. diss.: 11 (943-1,420): W Atlantic (7); Indian O. (3); Revillagigedos (1). Grammatorcynus bilineatus. — Total 52 (23.5-575 mm FL). meas.: 34 (226-575): Red Sea (13, *Thynnus bilineatus); Indian Ocean ? (1); Anda- man Sea (3); Celebes (1); New Guinea (3); Australia (8); Philippine Is. (5, *Nesogrammus piersoni); Solomon Is. (1); Caroline Is. (3); Marshall Is. (8); Fiji (2). counts: 44. diss.: 10 (382-453): Indian Ocean ? (1); Timor Sea (2); Bismarck Arch. (1); Marshall Is. (2); Queensland, Australia (4). Grammatorcynus bicarinatus. — Total 9 (306-825 mm FL). meas.: 9 (306-825): Western Australia (5); Queensland (4). counts: 9. diss.: 2 (521 and 563): Queensland. KEY TO GRAMMATORCYNUS, ACANTHOCYBIUM, AND SCOMBEROMORUS la. Two lateral lines, the lower joining the upper behind the pectoral fin base and at the caudal fin base; interpelvic pro- cess single; teeth in jaws slender, coni- cal, not compressed; vertebrae 31 .... Grammatorcynus 2 lb. One lateral line; interpelvic process double; teeth in jaws strong, com- pressed, almost triangular or knife- like; vertebrae 39-64 3 2a. Gill rakers 14-15; small eye, 3-4% FL; frequently with small dark spots on lower sides of body G. bicarinatus (Quoy and Gaimard) 2b. Gill rakers 19-24; large eye, 7-9% FL; seldom with dark spots on sides of body G. bilineatus (Riippell) 3a. Snout as long as rest of head; no gill rakers; 23-27 spines in first dorsal fin; posterior end of maxilla concealed un- 547 FISHERY BULLETIN: VOL. 82, NO. 4 der preorbital bone; vertebrae 62-64 . . Acanthocybium solandri (Cuvier) 3b. Snout much shorter than rest of head; gill rakers 1-27; 12-22 spines in first dorsal fin; posterior end of maxilla exposed; vertebrae 41-56 Scomberomorus 4 4a. Lateral line abruptly curving down be- low first or second dorsal fin; verte- brae 41-46 5 4b. Lateral line straight or descending gra- dually posteriorly; vertebrae 44-56 .... 7 5a. Lateral line abruptly curving down be- low first dorsal fin; total gill rakers on first arch 12-15; caudal vertebrae 21-22 S. sinensis (Lacepede) 5b. Lateral line abruptly curving down be- low second dorsal fin; total gill rakers on first arch 2-13; caudal vertebrae 23-27 6 6a. Total gill rakers on first arch 7-13, usual- ly 9 or more; spines in first dorsal fin 12-18, usually 15 or fewer; precaudal vertebrae 16-17 S. cavalla (Cuvier) 6b. Total gill rakers on first arch 3-8, usual- ly 6 or fewer; spines in first dorsal fin 15-18, usually 16 or more; precaudal vertebrae 19-20 S. commerson (Lacepede) 7a. Total gill rakers on first arch 21-27; no bars on body . . . . S. concolor Lockington 7b. Total gill rakers on first arch 1-18; spots, bars, or other markings usually pres- ent on sides of body 8 8a. Anal fin rays 25-29; second dorsal fin rays 21-25, usually 23 or more; gill rakers on first arch 1-4; total verte- brae 54-56; no pattern on body S. multiradiatus Munro 8b. Anal fin rays 15-24; second dorsal fin rays 15-24; total gill rakers on first arch 3-18; total vertebrae 44-53; sides of body usually with spots or other markings 9 9a. Dorsal fin spines 19-22, usually 19 or more 10 9b. Dorsal fin spines 13-19, usually 18 or fewer 11 10a. First dorsal fin black only on first 5-7 interspinous membranes, white pos- teriorly; intestine straight, with no folds; total vertebrae 48-50 S. niphonius (Cuvier) 10b. First dorsal fin black to, or almost to, posterior end; intestine with 2 loops and 3 limbs; total vertebrae 50-52 .... S. munroi Collette and Russo 11a. Lateral line with many small auxiliary branches anteriorly 12 lib. Lateral line without auxiliary branches or with only a few anteriorly 13 12a. Dorsal fin spines 15-18, usually 16 or more; intestine with 2 loops and 3 limbs; total vertebrae 47-52, usually 48 or more; head longer, 20.2-21.5% FL; body depth less, 22.8-25.2% FL . . . S. guttatus (Bloch and Schneider) 12b. Dorsal fin spines 14-17, usually 15 or fewer; intestine with 4 loops and 5 limbs; total vertebrae 46-47, usually 46; head shorter, 19.7-20.4% FL; body depth greater, 24.4-26.7% FL S. koreanus (Kishinouye) 13a. Sides of body with spots and at least one stripe, the stripes may be short, wavy or interrupted 14 13b. Sides of body without any stripes, spots usually present 16 14a. One long stripe on sides with spots or interrupted lines above and below the stripe; total vertebrae 47-48, usually 48; total gill rakers on first arch 12-18, usually 15 or more . . . . S. regalis (Bloch) 14b. Sides with several short stripes; total vertebrae 44-47, usually 46; total gill rakers on first arch 9-15, usually 14 or fewer 15 15a. Sides with a series of short straight stripes and few if any spots; total gill rakers on first arch usually 11 or fewer; second dorsal fin rays 15-19, usually 18 or fewer; distance from 2D origin to caudal base 46.2-54.5% FL, 3c 50.0% S. lineolatus (Cuvier) 15b. Sides with a series of short wavy mark- ings plus many small spots; total gill rakers on first arch usually 12 or more; 548 COLLETTE and RUSSO: SPANISH MACKERELS second dorsal fin rays 19-21, usually 20 or more; distance from 2D origin to caudal base 51.8-57.5% FL, x 54.8% . . S. plurilineatus Fourmanoir 16a. Sides with bars or large spots, larger than the diameter of the eye 17 16b. Sides with small round spots, about the diameter of the eye, orange colored in life 19 17a. Sides with large spots or blotches; total gill rakers on first arch 3-9, usually 7 or fewer S. queenslandicus Munro 17b. Sides plain or with bars; total gill rakers of first arch 6-15, usually 9 or more . . 18 18a. First dorsal fin spines 13-15; second dor- sal fin rays 19-22, usually 20 or more; total gill rakers on first arch 6-13, usually 11 or fewer; total vertebrae 44-46, usually 45; base first dorsal fin 17.0-23.6% FL S. semifasciatus (Macleay) 18b. First dorsal fin spines 15-18, usually 16 or more; second dorsal fin rays 16-19, usually 17; total gill rakers on first arch 12-15; total vertebrae 46-47, usu- ally 46; base first dorsal fin 23.8- 30.4% FL S. tritor (Cuvier) 19a. Total vertebrae 51-53; second dorsal fin rays 17-20, usually 18 or more S. maculatus (Mitchill) 19b. Total vertebrae 46-49; second dorsal fin rays 15-19, usually 18 or fewer 20 20a. Pectoral fin rays 21-24, usually 22 or more; pelvic fin short, 2.9-5.9% FL, x 4.5% . . . . S. brasiliensis Collette, Russo, and Zavalla-Camin 20b. Pectoral fin rays 20-24, usually 21 or fewer; pelvic fin longer, 3.2-6.4% FL, x 5.3% S. sierra Jordan and Starks COMPARATIVE MORPHOLOGY The morphological characters useful for distin- guishing the species of Scomberomorus and for evaluating their phylogenetic relationships are divided into six categories: lateral line, nasal denticles, morphometry, meristics, soft anatomy, and osteology. Lateral Line In most species of Scomberomorus, the lateral line runs posteriorly above the pectoral fin and then gradually descends to the middle of the body at about the level of the second dorsal fin. Gram- matorcynus differs from Scomberomorus, Acan- thocybium , and all other members of the family by having a second lateral line that joins the upper lateral line at a right angle behind the pectoral fin base and then courses ventrally and posteriorly along the ventral surface of the body to join the dorsal lateral line on the caudal peduncle. In Acanthocybium and three species of Scomberomorus, the lateral line moves abruptly downward under the first or second dorsal fin. The abrupt downward curve is under the first dorsal fin in Acanthocybium and S. sinensis (see Figure 68); it is under the second dorsal in S. cavalla and S. commerson (see Figures 50 and 52). Scomberomorus guttatus and S. koreanus dif- fer from other members of the genus in having many fine branches from the anterior part of the lateral line, both dorsally and ventrally (see Figures 54 and 56). Acanthocybium and S. ni- phonius (see Figure 62) may have branches from the lateral line but they are not as numerous or distinct. Nasal Denticles Nasal denticles (Fig. la, b) are small general- ized teeth found within the olfactory chamber on the medial surface surrounding the posterior nares and on the skin covering the anterior surface of the lateral ethmoid. Nasal denticles are similar to the small villiform teeth present within the mouth cavity and adjoining regions of stomadeal origin and on the skin covering the cleithrum (Fig. lc, d) and on the isthmus where they are contacted by the opercular membrane. These teeth typically fit into sockets in pads of fine spongelike bone. They point posteriorly and are aligned with presumed flow of water from the anterior naris through the olfactory chamber and out the posterior naris. Nasal denticles were found only in the six species of the Scomberom- orus regalis species-group {brasiliensis, concolor, maculatus, regalis, sierra, and tritor). Nasal denticles are not present in Acanthocybium or Grammatorcynus. We do not know their function and are not aware of such structures in other fishes. 549 FISHERY BULLETIN: VOL. 82, NO. 4 co m CO CO O o o o S g -o S o CO c B a u ■5 x 2 S ->-> o -a c CO X CO Z CO _ D cd a XI ca •— tut O O o, c o M c '8 c cd u CO w as s o 550 COLLETTE and RUSSO: SPANISH MACKERELS Morphometric Characters In addition to fork length, 26 measurements routinely were made on all specimens destined to be dissected, to insure that these data would be available if needed. Preserved material also was measured until an adequate sample was obtained. Measurements follow the methods of Marr and Schaefer (1949) as modified by Gibbs and Collette (1967) and Collette and Chao (1975). Morphometric characters can be used to separate species and populations within species. Tables showing the 26 characters as thousandths of fork length and 8 characters as thousandths of head length are presented in the systematic section of the paper (see Tables 13-30). Most of the charac- ters are best used at the species level; therefore, only a summary table of the means of proportions (Table 1) is presented in this section. Where there was sufficient material from two or more poten- tially different populations, analysis of covari- ance (ANCOVA) was carried out on the regres- sions of body parts on fork length. Results are reported, under a section entitled Geographic Variation, in 11 of the 18 species accounts. Tests of significance were made by Newman-Keuls Multiple Range Test. Meristic Characters Countable structures are of special value sys- tematically because they are relatively easy to record unambiguously and are easy to summarize in tabular fashion. Meristic characters that have proved valuable systematically in Scomberomo- rus include numbers of fin rays (first dorsal spines, second dorsal rays, dorsal finlets, anal rays, anal finlets, and pectoral rays), gill rakers, teeth on the upper and lower jaws, vertebrae (precaudal, caudal, and total), and lamellae in the olfactory rosettes. Olfactory lamellae are dis- cussed as the next to last section under soft TABLE 1. — Morphometric comparison of the species of Scomberomorus. Means as thousandths of fork length or head length. Species arranged alphabetically by the first three letters of their names. Ranges for the species given in Tables 13-30. Character bra cav com con gut kor lin mac mul mun nip plu que reg sem sie sin tri Min. spp. Max. spp. Fork length Snout -A 538 539 542 524 517 493 507 536 505 546 563 502 525 548 506 537 584 533 493 kor 584 sin Snout-2D 511 506 510 507 481 467 501 503 477 528 536 473 501 521 472 510 559 513 467 kor 559 sin Snout- 1 D 242 258 243 236 239 242 252 241 249 222 248 221 234 255 245 241 291 246 221 mun, plu 233 plu 291 sin Snout- P2 253 258 257 242 251 248 245 257 243 249 263 233 251 265 250 252 290 266 290 sin Snout-Pi 219 232 237 209 209 210 212 217 213 201 225 193 229 234 219 221 258 222 1 93 plu 258 sin P,-P2 108 106 96 100 106 114 93 110 102 105 105 103 99 109 105 104 113 111 93 lin 114 kor, sin 255 sin Head length 213 223 229 202 205 208 206 212 208 198 216 193 220 223 213 212 255 217 193 plu Max. body depth 198 191 187 187 209 237 181 197 229 190 172 206 188 197 211 190 218 206 172 nip 237 kor Max body width 82 89 94 89 93 100 97 91 95 100 84 97 101 91 94 84 102 90 82 bra 104 mun Pi length 123 129 122 125 109 133 139 129 131 109 111 123 120 126 147 123 158 134 109 gut, mun 40 mul 158 sin P2 length 45 65 56 50 59 60 55 52 40 54 68 51 55 56 50 53 83 60 83 sin P2 insertion-vent 273 271 273 261 251 227 241 263 247 281 285 243 254 267 237 267 273 250 237 sem 285 nip P2 tip-vent 225 212 217 212 191 164 185 211 207 225 218 186 198 210 187 222 189 190 164 kor 225 bra, mun 307 mun Base 1 D 263 245 261 254 235 218 231 256 216 307 282 240 263 257 210 260 260 262 210 sem Height 2D 117 109 103 111 131 166 124 125 167 112 98 148 114 114 159 123 145 126 98 nip 167 kor, mul Base 2D 118 106 104 127 141 160 114 128 178 115 113 128 113 114 138 120 121 122 104 com 178 mul Height anal 114 106 100 107 127 160 117 118 164 108 97 135 112 112 156 117 145 125 97 nip 164 mul Base anal 113 108 100 134 133 154 122 123 216 105 107 125 108 110 145 119 122 120 100 com 216 mul Snout (fleshy) 82 87 89 72 72 70 81 80 77 77 81 67 86 87 81 79 97 81 67 plu 97 sin Snout (bony) 72 79 81 63 64 62 74 70 67 70 75 59 80 79 72 70 91 72 59 plu 91 sin Maxilla length 123 132 131 113 108 111 113 119 125 104 120 96 125 124 119 121 147 123 96 plu 1 47 sin Postorbital 95 98 104 96 96 101 91 96 86 90 102 94 102 98 95 98 117 96 86 mul 117 sin Orbit (fleshy) 37 38 35 32 37 34 32 34 34 25 34 34 31 41 35 33 35 38 25 mun 41 reg Orbit (bony) 54 51 49 46 53 50 48 51 52 39 47 45 49 56 51 49 52 53 37 mun 56 reg Interorbital width 57 60 62 49 59 60 57 56 58 56 57 56 63 58 57 55 63 59 49 con 63 que, sin 2D-caudal 490 477 481 484 527 550 500 487 494 468 465 548 496 480 517 475 445 476 445 sin 550 kor Head length Snout (fleshy) 386 392 390 353 351 339 395 376 372 386 376 348 391 390 378 371 382 376 339 kor 395 lin Snout (bony) 343 357 355 313 310 301 359 335 321 351 346 306 363 351 339 331 355 333 301 kor 363 que Maxilla length 581 591 571 555 526 532 547 562 603 521 553 496 568 556 555 570 578 568 496 plu 603 mul Postorbital 446 438 455 476 464 489 442 454 415 456 473 485 463 439 447 461 460 443 415 mul 489 kor Orbit (fleshy) 175 168 147 159 174 157 156 160 165 134 150 179 142 178 162 158 138 173 129 mun 178 reg Orbit (bony) 249 229 211 226 252 238 231 242 252 199 215 232 223 247 238 235 202 245 191 mun 252 gut, mui Interorbital width 270 268 270 241 284 292 276 266 280 282 264 290 286 262 267 253 249 272 241 con 292 kor 551 FISHERY BULLETIN: VOL. 82, NO. 4 anatomy. The other meristic characters are dis- cussed in the relevant osteological sections of the paper. Soft Anatomy The relative position, shape, and size of the various internal organs provide valuable diag- nostic characters. For purposes of discussion, the characters in the soft anatomy are divided into five sections: viscera, vascular system, urogenital system, olfactory organ, and pharyngeal muscles. VISCERA Emphasis was placed on the appearance of the viscera in ventral view, after removal of an oval segment of the belly wall (Figs. 2, 3). Previous papers on the viscera include Kishinouye (1923, 5 Japanese species of Scomberomorus , and Acan- thocybium and Grammatorcynus), Munro (1943, 4 Australian species), Silas (1963, Gramma- torcynus), Mota Alves and Tome (1967a, S. caval- la), Mota Alves (1969, S. brasiliensis), Tongyai (1971a, S. guttatus and S. commerson), and Col- lette and Russo (1979, preliminary review of the genus). The anterior end of the liver abuts the trans- verse septum anteriorly in the body cavity. The liver has three lobes. The left and right lobes are longer than the middle lobe in all three genera (Fig. 4). The right lobe is longest in Scomberomo- rus and Grammatorcynus. The left and right lobes are about equal in length in Acanthocyb- ium. Two efferent (venous) vessels lead directly from the anterior surface of the liver into the sinus venosus in all species. The short esophagus leads into the stomach. The stomach is sometimes visible in ventral view but this is dependant on the amount of food present, rather than showing differences between species. The pyloric portion of the intestine arises from the anterior end of the liver INTESTINE FIGURE 2. — Viscera in ventral view. a. Scomberomo- rus maculatus, Georgia, 290 mm FL. b. Acanthocybium solandri, Campeche Banks, Mexico, 1,280 mm FL. c. Grammatorcynus bilineatus, Marshall Is., 424 mm FL. CAECAL MASS STOMACH GALL BLADDER .'.'.I GONAD x&axd URINARY BLADDER 552 COLLETTE and RUSSO: SPANISH MACKERELS stomach. At this point the main branches of the pyloric caeca join the intestine. The caeca branch and form a dense dendritic conglomeration, the caeca] mass. Cells in the pyloric caeca are histo- logically similar to those in the intestine and produce enzymes such as lipase, maltase, trypsin, and pepsin (Mota Alves and Tome 1970). The intestine continues posteriorly and its course appears to be species-specific. The intestine may be a simple straight tube from stomach to anus, have two descending and one ascending arm, or have four bends with three descending and two ascending arms. The spleen is prominent in ven- tral view in most species but is hidden in others. The gall bladder, an elongate tubular sac which is usually green, arises from the right lobe of the liver and usually lies along the first descending arm of the intestine on the right side. A swim bladder is present in Grammatorcynus , Acantho- cybium, and S. sinensis (Fig. 5) but is absent in the other 17 species of Scomberomorus. The Spanish mackerels can be divided into three groups based on the number of folds in the intestine. Grammatorcynus , Acanthocybium (Fig. 2b, c), and S. niphonius (Fig. 3k) have a straight gut not folded back on itself. Scomber- omorus koreanus (Fig. 3f) has four folds and five distinct arms. The other species all have two folds and three long arms (Fig. 3). Collette and Russo (1980) used this character to differentiate S. munroi from the North Pacific S. niphonius. The spleen is large and centrally located in ventral view in four species: guttatus, koreanus, munroi, and plurilineatus. The spleen is smaller and distinctly on the left side in ventral view in seven species: brasiliensis , commerson, lineola- tus, maculatus, multiradiatus , queenslandicus , and sinensis. It is not visible in ventral view in Grammatorcynus, Acanthocybium, and seven species of Scomberomorus: cavalla, concolor, ni- phonius, regalis, semifasciatus , sierra, and tri- tor. VASCULAR SYSTEM The only published work on the vascular sys- tem of the Spanish mackerels is on Japanese species by Kishinouye (1923). No specialized sub- cutaneous vascular system and no cutaneous ar- teries or veins are present as they are in the higher tunas, Thunnini, Auxis to Thunnus (Col- lette 1979). Therefore, this description will be confined to the anterior portion of the dorsal aorta and the postcardinal vein. The efferent branchial (epibranchial) arteries and coeliaco-mesenteric artery form a unit at the anterior end of the dorsal aorta (Figs. 6, 7). Two anterior epibranchials on each side unite to form a common trunk, and these trunks join as the "Y" of the aorta beneath the posterior part of the skull or the first or second vertebra. The posterior two epibranchials of each side unite immediately before they join the aorta, usually ventral to the second or third vertebra. As the aorta proceeds posteriorly, it gives rise to the large coeliaco- mesenteric artery on the right side ventral to the second to fourth vertebrae. The coeliaco- mesenteric artery has two or three main branches which lead to the liver and other viscera. The postcardinal vein runs along the ventral surface of the kidney (Fig. 8) from the vicinity of the first complete haemal arch anteriorly in the median line to the pectoral region. There it curves to the right and discharges into the right Cuvierian duct. Posteriorly, the postcardinal re- ceives a pair of small veins at the level of each vertebra. The postcardinal is composed of two main branches that join anterior to the Y of the ureter. The main branch leaves the haemal arch dorsally and the small branch runs under the surface of the kidney from the urogenital area. Five species of Scomberomorus (brasiliensis, concolor, maculatus, regalis, and sierra) have unique specializations of the right and/or left fourth epibranchial arteries (Fig. 7c-g). Each of these species has an artery arising from the fourth left epibranchial artery. Other species of the genus (e.g., S. guttatus and S. tritor, Fig. 7a, b) lack these specializations. In S. concolor and S. brasiliensis this branch is small and goes into the muscular tissue surrounding the left dorsal portion of the esophagus (Fig. 7d, 0. In S. macu- latus and S. sierra, this branch is large and becomes the dorsal left gastric artery (Fig. 7c, e). In S. regalis this branch goes into the left lobe of the liver (Fig. 7g, hepatic branch). Scomberomo- rus maculatus and S. sierra have lost the connec- tion between the dorsal left gastric artery and the coeliaco-mesenteric artery. It is replaced by a connection to the fourth left epibranchial artery. In S. regalis, the left dorsal gastric artery seems to have been reduced. Scomberomorus brasiliensis, S. sierra, and S. regalis share a specialization of the right fourth epibranchial artery. In these species an artery connects the fourth right epibranchial artery with a branch of the coeliaco-mesenteric artery (coeliaco-mesenteric shunt, Fig. 7e-g). 553 FISHERY BULLETIN: VOL. 82, NO. 4 a * <-- ""^ 1 * -*** ~^W 1 * ^#J •ii b I • --•I * ^ * * , • FIGURE 3. — Viscera in ventral view of representative specimens of the 18 species of Scomberomo- rus. a. S. brasiliensis, Belem Fish Market, Brazil, 556 mm FL. b. S. cavalla,off Miami, Fla., 797 mm FL. c. S. commerson, Gulf of Papua, 580 mm FL. d. S. concolor. Gulf of California, 495 mm FL. e. S. guttatus. Gulf of Mannar, 405 mm FL. f. S. koreanus, locality unknown, 812 mm FL. g. S. lineolatus, Cochin, India, 786 mm FL. h. S. maculatus, St. Andrews Bay, Fla., 323 554 COLLETTE and RUSSO: SPANISH MACKERELS LIVER CAECAL MASS V^yy\ INTESTINE I SPLEEN STOMACH I GALL BLADDER GONAD SWIM BLADDER mm FL. i. S. multiradiatus , Gulf of Papua, 272 mm FL. j. S. munroi, Gulf of Papua, 512 mm FL, USNM 219374. k. S. niphonius, Korea, 235 mm FL. 1. S. plurilineatus , Durban, S. Africa, 490 mm FL. m. S. queenslandicus , Exmouth Gulf, Western Australia, 466 mm FL. n.S. regalis , Bahamas, 456 mm FL. o. S. semifasciatus , Gulf of Papua, 715 mm FL. p. S. sierra, Baja Cali- fornia, 516 mm FL. q. S. sinensis, China, 711 mm FL. r. S. tritor, Gulf of Guinea, 415 mm FL. 555 FISHERY BULLETIN: VOL. 82, NO. 4 urinary bladder blood vessels ureter lateral muscle peritoneum swim bladder urogenital opening FIGURE 5. — Swim bladder and urinary bladder in ventral view of Scomberomorus sinensis (body wall and viscera removed), off Zhoushan Is., China, 714 mm FL, USNM 220856. FIGURE 4. — Livers in ventral view. a. Scomberomorus maculatus , Florida, 712 mm FL. b. Acanthocybium solandri, Florida, 1,403 mm FL. c. Gramma- torcynus bilineatus, Marshall Is., 444 mm FL. 556 COLLETTE and RUSSO: SPANISH MACKERELS ANTERIOR EPIBRANCHIAL POSTERIOR EPIBRANCHIAL \ COELIACO- MESENTERIC a b c FIGURE 6. — Anterior arterial system in ventral view. Numbers indicate vertebral centra, stippled areas where pharyngeal muscles originate, a. Scomberomorus multiradiatus , off the Fly River, Gulf of Papua, 272 mm FL. b. Acanthocybium solandn, Revillagigedos Is., 1,068 mm FL. c. Grammatorcynus bilineatus, Timor Sea, 453 mm FL. UROGENITAL SYSTEM The only reference to the anatomy of the uro- genital system in Scomberomorus (other than fishery biology studies of the gonads) is Kishi- nouye (1923) on Japanese species and Acantho- cybium. The paired gonads lie along the dorsolat- eral body wall and are visible in ventral view in mature adults. The kidney lies dorsal to the layer of fibrous connective tissue which forms the dor- sal wall of the peritoneum. Anteriorly, the kidney divides into a pair of narrow projections which extend along the sides of the parasphenoid and usually reach the posterior end of the "midridge" of the prootic. The anterior ends of the kidney surround the origins of the pharyngeal muscles on the vertebral column and usually separate along the middle of the vertebral column. In the vicinity of the esophagus, the kidney expands laterally and forms two projections which may extend anteriorly to the upper end of the gill slits. Posteriorly, near the posterior fifth of the body cavity, the kidney narrows to an elongate trian- gle (Fig. 8). The branches of the "ureter" (meso- nephric ducts) join to form a common trunk just before entering the urinary bladder. The ureters enter the urinary bladder either at its anterior end or on its dorsal surface. The urinary bladder (Figs. 9, 10) is either ovoid or elongate, depending on degree of inflation, and is located in the mesenteries between the gonads in all species except S. sinensis. Scomberomorus sinensis has a specialization of the urinary bladder unique to scombrids and, so far as we know, vertebrates in general. In this species the urinary bladder has become hypertrophied and occupies the space inside the swim bladder (Fig. 5). Acanthocybium (Fig. 2b) has an elongate urinary bladder that extends anteriorly one-third to two-thirds the length of the visceral cavity. OLFACTORY ORGAN Kishinouye (1923) provided a generalized ac- count of the olfactory organ of several scombrids. More detailed studies have been made on Scom- ber scombrus (Burne 1909), Sarda sarda (Treti- akov 1939), Allothunnus fallai (Nakamura and Mori 1966), Katsuwonus pelamis (Gooding 1963), Thunnus (Iwai and Nakamura 1964a; Gibbs and Collette 1967), and the bonitos, Sardini (Collette and Chao 1975). As in other scombrids, the olfac- tory cavity in Scomberomorus has a small ante- rior naris and a slitlike posterior naris. No infor- mation on the supplementary sacs, or accessory olfactory cavity (Iwai and Nakamura 1964a), was obtained from the present study comparable with that of Tretiakov (1939), who described three supplementary sacs (middle, maxillary, and ros- tral sacs) in Sarda sarda. The central axis of the 557 FISHERY BULLETIN: VOL. 82, NO. 4 anterior epi branchials posterior epibranchials coeli aco- mesenteric artery dorsal left gastric artery coeli aco-mesenteric — shunt dorsal left gastric artery hepatic branch coeliaco-mesen teric shunt 9 FIGURE 7. — Anterior arterial system in dorsal view of seven species of Scomberomorus . a. S. guttatus, Pakistan, 545 mm FL. b. S. tritor, Gulf of Guinea, 494 mm FL. c. S. maculatus, Chesapeake Bay, 312 mm FL. d. S. concolor, Gulf of California, 455 mm FL. e. S. sierra, Ecuador, 512 mm FL. f. S. brasiliensis, Belem market, Brazil, 588 mm FL, USNM 217557, paratype. g. S. regalis , Bahama Is., 490 mm FL. olfactory rosette is located beneath the anterior naris. Leaflike lamellae radiate from the central axis and occupy the anterior dorsal third of the olfactory cavity. Gooding (1963) studied the mor- phology and histology of the olfactory organ of Katsuwonus pelamis and found olfactory cells on the olfactory epithelium of the lamellae. Iwai and Nakamura (1964a) found that the number of lamellae per rosette varies among specimens of species of Thunnus but that there were differ- ences among species in the shape of the nasal laminae. Most species of bonitos have 21-39 la- mellae in each nasal rosette but Gymnosarda unicolor is distinct in the group in having 48-56 (Collette and Chao 1975:532). The number of olfactory lamellae was counted on both sides in Scomberomorus and a wide range of variation was observed, 24-76 (Table 2). In bonitos, the number of lamellae increases from small specimens to adults but does not appear to change after a certain size is reached, as Collette and Chao (1975:532) showed for Gymnosarda 558 COLLETTE and RUSSO: SPANISH MACKERELS left postca rdina! vein k i d n TABLE 2. — Number of lamella in nasal rosettes of species of Scomberomorus. ey FIGURE 8. — Kidney and postcardinal vein in ventral view of Scomberomorus queenslandicus, Palm I., Queensland, 641 mm FL. Overall Species Side Min. Max X N X Rank brasiliensis L 24 40 33.67 12 33.88 6 R 25 42 34.08 13 cavalla L 30 56 42.92 24 43.13 12 R 31 55 43.35 23 commerson L 42 58 48.92 12 49.32 13 R 43 60 49.80 10 concolor L 26 35 30.92 13 30.92 1 R 26 34 30.92 13 guttatus L 30 76 53.41 27 53.43 16 R 31 73 53.46 26 koreanus L 47 56 50.67 3 54.75 18 R 48 73 57.20 5 lineolatus L 30 35 32.50 4 32.18 3 R 30 34 32.00 7 maculatus L 25 38 33.43 14 33.44 5 R 30 37 33.45 11 multiradiatus L 32 40 36.75 4 36 00 10 R 25 44 34.50 2 munroi L 54 54 54.00 3 53.84 17 R 54 57 53.67 3 niphonius L 25 42 33.67 15 34.41 7 R 26 42 35.21 14 plurilineatus L 45 53 49.50 4 50 50 14 R 44 56 51.50 4 queenslandicus L 43 59 49.75 4 50.67 15 R 43 61 51.40 5 regalis L 28 41 34.00 9 35.11 9 R 30 43 3622 9 semifasciatus L 31 37 34.00 3 34.78 8 R 31 38 35.17 6 sierra L 30 36 32.64 14 3207 2 R 28 34 31.50 14 sinensis L 38 38 38.00 1 42.50 11 R 41 47 44.00 3 tritor L 27 48 33.40 10 32.57 4 R 24 37 31.83 11 unicolor and Orcynopsis unicolor. We have not examined many small Scomberomorus nasal ro- settes but did find 23 lamellae in an 80 mm FL S. guttatus, a species for which the minimum count of lamellae for specimens larger than 100 mm was 30. Three species of Scomberomorus (koreanus, munroi, and guttatus) had high counts, overall means 53.4-54.8. The highest counts per side were for S. koreanus (73) and S. guttatus (76). Ten species had low counts, overall means 31.0- 36.0. These 10 included all 6 species of the regalis group as well as lineolatus, multiradiatus, ni- phonius, and semifasciatus. PHARYNGEAL MUSCLES The paired pharyngeal (retractor dorsalis) muscles originate on the ventral surface of one or two vertebrae between the third and the sixth abdominal vertebrae and insert on the upper pharyngeal bones (Fig. 2). We did not find any differences between species as Collette and Chao (1975) did for the bonitos. 559 FISHERY BULLETIN: VOL. 82, NO. 4 ovary FIGURE 9. — Urogenital system in ventral view of Scomberomorus (body wall and viscera removed). Composite illustration. peritoneum urinary bladder intestine anal pore genital pore urogenital papilla peritoneum gonad urinary b I a d d e arge intestine ureter urinary bladde anus FIGURE 10. — Urogenital system in ventral view of Scomberomorus queenslandicus , Palm I., Queensland, 641 mm FL. a. With intestine opened, b. Urinary bladder and ureters. Osteology Osteological characters proved to be useful in determining relationships among the 18 species of Scomberomorus and between this genus and its presumed closest relatives, Acanthocybium and Grammatorcynus. The osteological portion of the paper is divided into five sections: skull, axial skeleton, dorsal and anal fins, pectoral girdle, and pelvic girdle. Osteological terminology gen- erally follows Gibbs and Collette (1967) and Col- lette and Chao (1975). Organization within sec- tions is similar to that of Collette and Chao (1975) and the two earlier papers of most importance to the osteology of Scomberomorus: Mago Leccia (1958) on three western Atlantic species (caualla, maculatus, and regalis) and Devaraj (1977) on four Indian species (commerson, guttatus, ko- reanus, and lineolatus) and Acanthocybium. SKULL Description of the skull is presented in two sections: neurocranium (Figs. 11-19) and bran- chiocranium. 560 COLLETTE and RUSSO: SPANISH MACKERELS Neurocranium Following a general description of the neuro- cranium, the four major regions are discussed: ethmoid, orbital, otic, and basicraniai. GENERAL CHARACTERISTICS.— In dorsal view, the neurocranium of Scomberomorus is more or less trapezoidal in shape. It is elongate and flat, particularly at the anterior region and is deepest at the hind end of the orbit. The dorsal surface is marked by a median ridge and three grooves on each side: dilator, temporal, and su- pratemporal (Allis 1903:49). These grooves are SPHENOTIC PTEROTIC FRONTAL NASAL ETHMOID INTERCALAR EXOCCIPITAL VOMER LATERAL ETHMOID FIRST VERTEBRA EPIOTIC SUPRAOCCIPITAL a FIGURE ll. — Skulls in dorsal view. a. Scomberomorus commerson, Coffs Harbour, New South Wales, 1,155 mm FL. b. Scom- beromorus munroi, Cairns, Queensland, 800 mm FL, USNM 219372, paratype. 561 separated from each other by ridges of bone. Thus, there are six grooves and five ridges in all. The median ridge is carried forward on the fron- tals to the ethmoid and is prolonged posteriorly in a large supraoccipital crest. This crest extends down over the exoccipital suture more broadly than in any other genus of the Scombridae. The internal ridge or temporal ridge almost reaches anteriorly to the posterior portion of the FISHERY BULLETIN: VOL. 82, NO. 4 nasal, and it is not interrupted above the eyes by any transverse ridge. Posteriorly, the ridge ends at the epiotic where the medial process of the posttemporal attaches. The external or pterotic ridge extends forward to the midlevel of the orbit and develops anterior- ly a small auxiliary ridge that extends laterally and posteriorly toward the temporal ridge. The dilator groove is shorter than the other two a FIGURE 12.— Skulls in dorsal view. 562 a. Scomberomorus koreanus, Singapore, 480 mm FL. California, 495 mm FL. b. Scomberomorus concolor, Gulf of COLLETTE and RUSSO: SPANISH MACKERELS and can be detected easily in lateral view. The temporal groove is the middle one and is deeper than either of the other two. The remaining groove, the supratemporal, is the largest of the three and opens posteriorly between the supra- occipital crest and the middle portion of the epiotic. The interorbital and otic regions are not as broad as in the more advanced genera of the Sardini (Collette and Chao 1975) and Thunnini (Gibbs and Collette 1967). The median and tem- poral crests are higher in Scomberomorus than in other scombrids. The bonitos, particularly Or- cynopsis unicolor (Collette and Chao 1975:fig. 21), a FIGURE 13. — Skulls in dorsal view. a. Acanthocybium solandri, Caribbean Sea, 1,240 mm FL. b. Grammatorcynus bilineatus, Scott Reef, Timor Sea, 453 mm FL. 563 EPIOTIC FISHERY BULLETIN: VOL. 82, NO. 4 SUPRAOCCIPITAL CREST PARIETAL FRONTAL ETHMOID VOMER LATERAL ETHMOID PTEROSPHENOID PARASPH PTEROTIC INTERCALAR FIRST VERTEBRA EXOCCIPITAL PROOTIC a FIGURE 14. — Skulls in lateral view. a. Scomberomorus commerson, Coffs Harbour, New South Wales, 1,155 mm FL. b. Scomberomorus munroi. Cairns, Queensland, 800 mm FL, USNM 219372, paratype. have the next highest crests. ETHMOID REGION. —This region is composed of the ethmoid, lateral ethmoid, and vomer. The nasal bone lies lateral to the ethmoid and lateral ethmoid and, therefore, is included here. Ethmoid. — The ethmoid (dermethmoid) is a forked median bone overlapped by the frontals above and bounded by the vomer and lateral ethmoid ventrally The concave anterior surface articulates with the ascending process of the premaxilla. At its anterolateral aspect, the eth- moid bone supports the nasals. In Scomberomorus, only the most anterior part of the ethmoid bone is exposed in dorsal view, while the rest of it is overlapped by the frontals. In Acanthocybium, only the lateral aspects of the bone are overlapped by the frontals and a V-shaped dorsal median portion is exposed. The ethmoid bone is longer in A. solandri than in Scomberomorus. Lateral ethmoid. — The lateral ethmoids (par- 564 COLLETTE and RUSSO: SPANISH MACKERELS ethmoids) are massive paired bones which form the anterior margin of the orbit and the posterior and mesial walls of the nasal cavity. The lateral portion of each bone extends downward from the middle region of the frontals. The ventral surface of this wall mesially bears an articulating surface for the palatine and laterally another articulat- ing surface for the first infraorbital (lachrymal). The inner walls of the lateral ethmoids come closest to each other at the ventral median line of the skull and contact the anterior edge of the parasphenoid. The median half of each lateral ethmoid extends downward about three-fourths as far as the lateral portion and has a large round foramen for the olfactory nerve which is promi- nently seen on the anterior surface. On the dorsal surface, they abut the nasals anteriorly, the fron- tals posteriorly, and articulate with the ethmoid mesially. On the anterior surface, ventral to the foramen, each lateral ethmoid bears a process a FIGURE 15. — Skulls in lateral view. a. Scomberomorus koreanus, Singapore, 480 mm FL. b. Scomberomorus concolor, Gulf of California, 495 mm FL. 565 FISHERY BULLETIN: VOL. 82, NO. 4 that extends anteriorly and mesially to contact the dorsolateral surface of the spear-shaped pos- terior portion of the vomer. No appreciable differ- ence was noted in the lateral ethmoids of the different species. Vomer. — The vomer is the most anteroventral- ly located bone of the cranium. The spatula- shaped anterior process bears a large oval patch of fine teeth on its ventral surface. The vomerine tooth patch extends posteriorly as a narrow ridge in some specimens of some species, e.g., S. con- color (Fig. 15b). The vomer articulates with the ethmoid dorsally and lateral ethmoid dorsolater- al^. The pointed posterior process is firmly anky- losed dorsally with the parasphenoid. On each side of the vomer, dorsolaterally and behind the spatulate anterior process, is a prominent articu- lar surface for a loose articulation with the head of the maxilla. Posterior to this articular surface, facing ventrolateral^, is a prominent sulcus for a similar movable articulation with the ventral branch of the anterolateral fork of the palatine. The spatulate anterior process of the vomer is very long and extends beyond the anterior mar- gins of the nasal and ethmoid bone in Scomber- a FIGURE 16. — Skulls in lateral view. a. Acanthocybium solandri, Caribbean Sea, 1,240 mm FL. bilineatus, Scott Reef, Timor Sea, 453 mm FL. b. Grammatorcynus 566 COLLETTE and RUSSO: SPANISH MACKERELS omorus. No other scombrid has such a spatulate anterior extension of the vomer. In fact, the vomer is either not visible in dorsal view or protrudes anteriorly slightly beyond the ethmoid in other scombrid genera. Nasal. — The nasal bones (Fig. 20) are flat, roughly triangular bones with thickened lateral edges. The mesial edges are irregular and almost serrate in some species to form a firm immovable articulation with the lateral edge of the frontals. PTEROSPHENOID LATERAL ETHMOID ETHMOID VOMER FRONTAL PROOTIC > PTEROTIC TERCALAR PARASPHENOID SPHENOTIC a BASIOCCIPITAL FIRST VERTEBRA EXOCCIPITAL FIGURE 17. — Skulls in ventral view. a. Scomberomorus commerson, Coffs Harbour, New South Wales, 1,155 mm FL. b. Scom- beromorus munroi, Cairns, Queensland, 800 mm FL, USNM 219372, paratype. 567 FISHERY BULLETIN: VOL. 82, NO. 4 a FIGURE 18. — Skulls in ventral view. a. Scomberomorus koreanus, Singapore, 480 mm FL. b. Scomberomorus concolor, Gulf of Cali- fornia, 495 mm FL. The anterior margins fit neatly beside the ante- rior branches of the forked ethmoid bone as can be seen in the dorsal views of the skulls (Figs. 11, 12). They are nonprojecting in that their anterior margin is at the level of the ethmoid bone except in Grammatorcynus where they project well be- yond the anterior end of the neurocranium (Fig. 13b). Length divided by width ranges from 2.0 to 4.2 in the three genera. The widest nasal bones are in S. koreanus (2.0-2.1) and S. sinensis (2.0- 2.3, Fig. 20b). The most elongate nasals are in Acanthocybium (3.1-4.2, Fig. 20c), Grammator- cynus (2.8-3.4, Fig. 20d), S. cavalla (2.8-3.1, Fig. 20a), and S. regalis (2.8-3.0). The other 14 species of Scomberomorus are intermediate (2.0-2.9). The anterior end of the nasal bone is rounded and heavy in Scomberomorus and Acanthocybium (Fig. 20a-c). The anterior end has a short, slightly angled arm in Grammatorcynus (Fig. 20d). ORBITAL REGION.— The orbit is surrounded by the posterior wall of the lateral ethmoid, the ventral side of the frontal, the pterosphenoid, sphenotic, prootic, suborbital, and lachrymal 568 COLLETTE and RUSSO: SPANISH MACKERELS FIGURE 19. — Skulls in ventral view. a. Acanthocybium solandri, Caribbean Sea, 1,240 mm FL. bilineatus, Scott Reef, Timor Sea, 453 mm FL. b. Grammatorcynus bones. The left and right orbits are partially separated by the basisphenoid. The sclerotic bones enclose the eyeballs. Frontal. — The frontals are paired bones that form the largest portion of the dorsal surface of the neurocranium. Anteriorly they are pointed, and posteriorly they become expanded. Anterior- ly, the frontals overlap the dorsal surface of the ethmoid bone, the inner edge of the nasals, and the dorsal surface of the lateral ethmoid. The midlateral aspect is thickened to form the orbital roof. Posteriorly, they are bounded by the supra- occipital and parietals. Posterolaterally, they overlap the pterotics and just anterior to the pterotics, cover the sphenotics. Ventrally, each frontal bears a sheet of bone, the orbital lamella, which is bounded by the sphenotic posteriorly, lateral ethmoid anteriorly, and pterosphenoid mesially. On the base of the orbital lamella may 569 FISHERY BULLETIN: VOL. 82, NO. 4 be seen a number of small foramina for the branches of the supraorbital nerve trunk. The laterosensory canals of the frontals are evident on the pterotic crests as a series of pores. In Acanthocybium, the frontals are separated from each other by the dorsomedian pineal fenes- tra lying just in front of the supraoccipital at the level of the pterosphenoids and another anterior fontanel just posterior to the ethmoid bone (Fig. 13a). A smaller, more oval pineal opening is present between the posterior ends of the frontals in Grammatorcynus (Fig. 13b). When viewed through the pineal fenestra, a part of the dorsal surface of the parasphenoid is visible through the opening of the brain chamber between the ptero- sphenoids. There is a deep depression on the frontals mesially, just anterior to the pineal fe- nestra. This depression becomes shallower ante- riorly, becoming confluent with the dorsal surface of the frontals. In Scomberomorus , the frontals join mesially along the median line on the neuro- cranium where they form the anterior half of the median ridge whose posterior half is composed of the supraoccipital crest. In all but three species of Scomberomorus, the left and right frontals are attached very closely to each other such that there is no gap between them. However, in S. commerson and S. cavalla, there is a long narrow slit between the left and right frontals, but it is not a fenestra in the true sense, as the lower parts of the bones are very closely approximated. A third condition is found in S. sinensis. Here the anterior part of the median ridge is almost absent and there is a wide gap between the left and right frontals. The interorbital commissures of the lat- eralis system are developed a little anterior to the middle of each frontal in the form of two pores at the margin of the median ridge which lead into oblique tubes downwards and posteriorly. An- other pair of commissures of the lateralis sys- tem is developed along the anterolateral margin of the frontals. These sensory canals are not developed in Acanthocybium and Grammatorcy- nus. Pterosphenoid. — The pterosphenoids (alisphe- noids) form the posterodorsal region of the orbit. They abut the basisphenoid and prootics posteri- FlGURE 20. — Left nasal bones in lateral view. a. Scomberomorus cavalla, Miami, 797 mm FL, 2x . b. Scomberomorus sinensis, Tokyo, 1,850 mm FL, lx. c. Acanthocybium solandri, Revillagigedos Is., 1,068 mm FL, 1.5 x. d. Grammatorcynus bilineatus, Queensland, 521 mm FL, 3 x . 570 COLLETTE and RUSSO: SPANISH MACKERELS orly and the frontals and sphenotics laterally. There is a space between the left and right pterosphenoids opening into the brain chamber just anterior to the basisphenoid. In most species of Scomberomorus, there is an anterior medially directed lobe on each pterosphenoid. These lobes meet along the median line or at least come very close to each other in adults of three species: commerson (over 1,000 mm FL), caualla (over 550 mm FL), and lineolatus (over 750 mm FL). Small- er specimens of these 3 species and all sizes of the other 15 species have a wide gap or fenestra between the left and right lobes. The gap is about equal to the width of the parasphenoid or slightly larger in three species: brasiliensis, koreanus (Fig. 12a), and concolor (Fig. 12b). The gap is largest in S. multiradiatus , so large that there is virtually no medially directed lobe. This causes the window into the brain chamber to be almost rectangular in this species. Sclerotic. — The sclerotic bones consist of two thickened semicircular segments connected by cartilage on the inner lateral surface and by corneal membranes on the outside. The inner rim of the sclerotic bones appears elliptical externally as in the bonitos (Collette and Chao 1975) and Thunnus (e.g., T. atlanticus, de Sylva 1955:fig. 7). The sclerotic bones of Grammatorcynus are relatively larger, thinner, and close to circular. In Acanthocybium, the sclerotic bones are ellipti- cal as in Scomberomorus, but they are heavier and extend further medially. The only species of Scomberomorus that appeared to differ from the other species is S. sinensis. The sclerotics are especially thick in this species and there is a thick bony lump in the middle of the posterior surface of one of the two sclerotics. Other species of Scomberomorus have a thickening of the bone in the same region but it does not form a distinct protrusion as it does in S. sinensis. Basisphenoid. — The basisphenoid is a small, median, Y-shaped bone that connects the para- sphenoid, prootics, and pterosphenoids. The com- pressed median vertical base bears an anterior median process but lacks a posterior process as is present in other scombrids such as Thunnus (Gibbs and Collette 1967) and most bonitos (Col- lette and Chao 1975). In most species of Scomber- omorus there is at least a trace of a lateral ridge that extends laterally and posteriorly on each side of the anterior process. There is great varia- tion in the length of the anterior process and in the relative degree of development of the lateral ridges. Both features are best developed in S. commerson where the length of the anterior pro- cess is greater than the height of the vertical axis of the bone. Infraorbitals. — The infraorbital (suborbital) series of Scomberomorus consists of from 9 to 13 elements which enclose the infraorbital branch of the lateral sensory canal system (Fig. 21a). Only 9 elements were observed in S. munroi, S. sierra, and S. sinensis, but 13 elements were observed in S. brasiliensis. The canal enters the infraorbital series at what is usually considered the last element (dermosphenotic) and continues around the orbit to terminate on the first infraorbital (lachrymal). The first infraorbital (lachrymal or IOl) is the first and largest element in the infraorbital se- ries. Anteriorly, several canal tubes open on the laminar, platelike surface of the bone. Posterior- ly, the canal tube continues directly to the second infraorbital. The first infraorbital is an elongate bone (length/height = 2.8-3.5) that covers part of the maxilla and is attached to the lateral ethmoid dorsally by a mesially directed articular process. The anterior portion is forked with a thin ante- rior process. This process is a point of attachment for a ligament connected to the nasal. The projec- tion is present in all species of Scomberomorus except S. lineolatus and S. tritor. The portion posterior to the articular process is elongate, pointed, and longer than the anterior portion. The general shape of the first infraorbital in Scomberomorus is similar to that in the bonitos (Collette and Chao 1975:fig. 28), particularly Cybiosarda elegans, except that the anterior pro- cess is smaller and more dorsally directed than in Cybiosarda. Acanthocybium differs from Scom- beromorus in having the posterior portion of the first infraorbital short and broad, shorter than the anterior portion (Fig. 21b). Grammatorcynus has a feebly forked anterior end (Fig. 21c), lack- ing a distinct anterior process such as is present in Scomberomorus and Acanthocybium. As Devaraj (1977) noted, the dorsal margin of the anterior part of the first infraorbital is straight, or nearly so, in S. cavalla and S. com- merson but clearly concave in the other species. Mago Leccia (1958:pl. 4, fig. 7) indicated that S. cavalla lacked the characteristic anterior projec- tion, but we have found it to be present in our material. In other respects, there seems to be as much variation between individuals of a species 571 FISHERY BULLETIN: VOL. 82, NO. 4 d e r m o s p h e n o t i c cheek scales FIGURE 21. — Left infraorbital bones in lateral view. a. Scomberomorus maculatus, Cape Hatteras, N.C., 534 mm FL. b. Acantho- cybium solandri, Revillagigedos Is., 1,068 mm FL. c. Grammatorcynus bilineatus, Timor Sea, 453 mm FL. as between species in the shape of the first infraorbital. The second infraorbital (102) sits firmly on the dorsal edge of the anterior portion of the first infraorbital. It is a flat, somewhat compressed bone. The third infraorbital (103) is an elongate, tubular bone. It has no platelike extensions, but has a large mesial shelflike extension (subocular shelf of Smith and Bailey 1962). Although not reported by those authors, we have found this shelf to occur in all species of Scomberomorus as well as in all other genera of scombrids. The shape of this shelf varies among specimens of the same species as well as between right and left sides of a single specimen. The fourth through the penultimate elements (postorbitals) usually are simple tubelike bones which may have pores accommodating canal tubes to the skin and cheek scales. The fourth 572 COLLETTE and RUSSO: SPANISH MACKERELS through about the seventh elements may be ex- panded laterally as laminar plates which cover the anterior end of the cheek scales. There may be 10-16 rows of specialized cheek scales posterior to the infraorbitals. These scales originate mesial to the infraorbital canal tubes and extend poste- riorly as flat, sometimes pointed, platelike ele- ments. These platelike scales may themselves be covered with more typical cycloid scales and exhibit the same morphology as the corselet scales of higher scombrids. The cheek scales of Scomberomorus may represent the primitive condition of the corselet. OTIC REGION.— This region encloses the otic chamber inside the skull, and is formed by the parietal, epiotic, supraoccipital, prootic, pterotic, sphenotic, and intercalar (opisthotic) bones. Parietals. — The parietals articulate with the frontals anteriorly, the supraoccipital mesially and the pterotics laterally, sphenotics ventrally, and epiotics posteriorly. The inner lateral crest that originates at the middle of the frontal bones continues through the parietals to terminate at the epiotics. This crest is typical of scombrids and is particularly well developed in Scomberomorus. These crests originate on the parietals, instead of the frontals, in Acanthocybium and Grammator- cynus and are not as high as in Scomberomorus. The parietals of all the species of Scomberomorus are similar. There is a gap or fenestra on the dorsal surface of the skull where the parietal, epiotic, and pter- otic bones come together. It varies in shape from roughly triangular to rectangular in most spe- cies. There is wide variation from specimen to specimen that tends to obscure potential inter- specific differences. The gap is very small in some specimens of eight species: commerson , concolor (Fig. 12b), koreanus (Fig. 12a), maculatus , mun- roi (Fig. lib), plurilineatus , queenslandicus , and sierra. It is usually larger in the other species and in most specimens of S. commerson (Fig. 11a). Epiotics. — The epiotics are massive, irregular, and bounded by the parietals anteriorly, the supraoccipital mesially, the exoccipitals posteri- orly, and the pterotics laterally. The inner lateral crests terminate at the posterior end of the epi- otics. The medial process of the posttemporal bone attaches here on a rough process. There are slight differences between the species of Scom- beromorus in the attitude of the attachment area and its roughness. In many species, the lateral crest continues posteriorly almost perpendicular to the skull. In some species such as S. commer- son (Fig. 11a) and S. queenslandicus, the area of attachment is flatter. This area is flat and rough in Acanthocybium (Fig. 13a) and forms a separate process in Grammatorcynus (Fig. 13b). Supraoccipital. — The supraoccipital forms the dorsomedial portion of the posterior end of the neurocranium and bears a well-developed crest which continues anteriorly on the frontals and is pronounced posteriorly as a strong supraoccipital crest. The supraoccipital can be divided into two parts: a thin, elongate triangular crest and a roughly hexagonal base. The crest extends down over the exoccipitals along the median line where the dorsal walls of the exoccipitals suture with each other, but it is not interposed between the exoccipitals. The hexagonal base is bounded an- teriorly by the frontals and laterally by the parietals and epiotics. The crest extends posteri- orly over the first vertebral centrum usually to a level past the posterior margin of the centrum (Figs. 14-16). The height of the crest varies among species of Scomberomorus and is highest in three species, S. guttatus, S. koreanus (Fig. 15a), and S. multiradiatus. Dividing the height of the su- praoccipital crest (ventral margin of supraoccipital to edge of crest) by skull length (tip of vomer to posteroventral margin of basioccipital) gives a ratio of 0.46-0.57 for these three species, com- pared with 0.34-0.45 in the other 15 species. Low ratios are found in S. cavalla and S. commerson (0.35-0.40) and in all six species of the regalis group (0.34-0.42). Prootics. — In ventral view, the prootics con- nect with all bones on the ventral side of the skull which compose the posterior part of the neuro- cranium (Figs. 17-19). Each prootic is bordered ventrally by the parasphenoid; posteriorly by the basioccipital, exoccipital, and intercalar; lateral- ly by the pterotic and sphenotic; and anteriorly by the pterosphenoid and basisphenoid. The pro- otic bones are irregular in shape and meet each other along the ventromedian line of the brain case to form the anterior portion of the poste- rior myodome. On the ventral surface, extending from the lateral wing of the parasphenoid to the sphenotic, the prootic forms a thick bridge which strengthens the trigemino-facialis chamber (Allis 1903). A prootic foramen is present anterolateral - 573 FISHERY BULLETIN: VOL. 82, NO. 4 ly between the tip of the parasphenoid wing and the sphenotic. There is no trace of the prootic pit characteristic of the Thunnini and Allothunnus (Gibbs and Collette 1967; Collette and Chao 1975). Specimens differ in the number and arrangement of foramina leading into the brain cavity from inside the anterior opening of the trigemino- facialis chamber, but these do not seem to be useful interspecific differences. Pterotics. — The pterotics form the lateral pos- terior corners of the neurocranium. Posteriorly, each pterotic is produced into a truncate process or pointed spine. The pterotics articulate with the epiotics and parietals medially and with the exoccipitals and intercalars posteriorly. A ridge, the pterotic ridge, originates on the dorsal sur- face of the posterior third of the frontal and continues posteriorly, diverging to the posterior corner of the pterotic, just anterior to the pterotic spine. In ventral view, the pterotics articulate with the sphenotics anteriorly and the prootics and intercalars medially. Two contiguous fossae, one at the posterior half of the pterotic bone and one at its joint with the sphenotic, seat the dor- sal and anterior condyles of the hyomandibula. Three closely situated lateral sensory canal pores open on each pterotic at the posteriormost region of the pterotic crest. The largest pore is the most posterior and opens dorsally; lateral to this is the next largest opening laterally on the outside of the pterotic crest; the smallest is the most ante- rior of the three, lying along the crest and usually more elongate in shape. The lengths and widths of the pterotic spines vary among the species. In eight species (brasili- ensis, guttatus, koreanus (Fig. 18a), multiradi- atus, plurilineatus , regalis, semifasciatus , and tritor), there is essentially no pterotic spine, merely a rounded posterior area of the skull. In six species (concolor (Fig. 18b), lineolatus, macu- latus, munroi (Fig. 17b), niphonius, and sierra), there is a blunt posteriorly projecting spine. Scomberomorus sinensis is similar to this group, but the posterior projection is broader and less like a spine. The pterotic spines are longest in three species (cavalla, commerson (Fig. 17a), and queenslandicus), all of which also have promi- nent posterior projections of the intercalars. Grammatorcynus (Fig. 19b) is similar to the lat- ter group, but the spine is thinner and sharper. Acanthocybium (Fig. 19a) has a longer and thin- ner pterotic spine than do Grammatorcynus and the species of Scomberomorus. Sphenotics. — The sphenotics form the most posterior dorsolateral part of the roof of the orbit. They continue the outer lateral shelf from the frontals and articulate with the pterosphenoid medially and the prootic and pterotic posteriorly. A segment of the articular fossa for the head of the hyomandibula is afforded by the lateral wall of the sphenotic on the ventral surface. The sphenotic is pierced by a foramen for the ramus oticus nerve (Allis 1903). When viewed dorsally, the sphenotics spread out on both sides more prominently in Scomberomorus than in Acan- thocybium, as noted by Devaraj (1977). Devaraj stated that the "midlateral projection" was large in koreanus, guttatus, maculatus, and regalis; small in lineolatus, cavalla, and commerson; and absent in Acanthocybium, but we are not clear as to what he was referring. Intercalars. — The intercalars (opisthotics) are flat bones that form part of the posterior border of the neurocranium interposed between the pter- otics and exoccipitals. The anterior portion on the dorsal surface is concealed by the overlapping pterotic, thus exposing the bone on the dorsal surface less than on the ventral side. Each inter- calar bears a protuberance on the dorsal surface to receive the lateral arm of the posttemporal. This protuberance is followed by a posterior projection of the intercalars in some species of Scomberomorus but not in Acanthocybium or Grammatorcynus. Species of Scomberomorus may be roughly di- vided into three groups based on the size of the posterior projection from the intercalar as Devaraj (1977) noted for Indian species. Eight species lack any posterior projection or have only an insignificant projection: guttatus, koreanus (Fig. 18a), lineolatus, multiradiatus , munroi (Fig. 17b), plurilineatus, semifasciatus, and si- nensis. In each of these species, except S. multi- radiatus, the pterotic spine protrudes further posteriorly than does the intercalar region. In S. multiradiatus, the posterior corners of the skull are rounded and there is no pterotic spine so the intercalars project further posteriorly. Eight spe- cies have a distinct posterior projection from the intercalar: brasiliensis , cavalla, concolor (Fig. 18b), maculatus, niphonius, regalis, sierra, and tritor. The posterior projection is smaller in some specimens of S. niphonius, placing it somewhat between groups 1 and 2. The posterior projection is a little longer in S. cavalla, between groups 2 and 3. Two species, commerson (Fig. 17a) 574 COLLETTE and RUSSO: SPANISH MACKERELS and queenslandicus, have a prominent truncate process. BASICRANIAL REGION.— This region con- sists of the parasphenoid, basioccipital, and ex oc- cipital bones, and forms the posteroventral base of the skull. Parasphenoid. — The parasphenoid is a long cross-shaped bone (Figs. 17-19) which articulates with the vomer anteriorly and forms the ventral axis of the skull. The lateral wing of the para- sphenoid extends dorsolaterally along the ventral ridge of the prootic bones on either side, and has a pointed end which forms part of the antero ventral wall of the posterior myodome. Posteriorly, the parasphenoid bifurcates into two lateral flanges which attach dorsally to the corresponding pos- teroventral flanges of the basioccipital bone and surround the posterior opening of the posterior myodome. A ventrally projecting median keel is present in the area anterior to the origin of the lateral flanges. In ventral view, the general characteristic of the parasphenoid is a gradual narrowing of the bone from anterior to posterior. The broadest portion of the parasphenoid is lo- cated usually at or before the tip of the V-shaped joint with the vomer. Broad parasphenoids are also present in Acanthocybium and the bonitos, Sardini (Collette and Chao 1975). In lateral view (Figs. 14-16), the parasphenoid forms the ventral border to the orbits and connects with the lateral ethmoids, basisphenoid, prootics, and basioccipi- tal bones dorsally. The shaft of the parasphenoid is distinctly wider in seven species: S. commerson (Fig. 17a), lineolatus, munroi (Fig. 17b), niphonius, queens- landicus, semifasciatus, and sinensis. Devaraj (1977) included S. cavalla along with S. lineola- tus and S. commerson as having a broad para- sphenoid, based on Mago Leccia (1958). We find S. cavalla to have a broader parasphenoid than some members of the regalis species group but not as broad as in the group of seven species listed above. Basioccipital. — The basioccipital is the most posteroventrally located bone of the skull. It is shaped like an inverted U with lateral flanges on either side of the skull and forms the roof and lateral walls of the posterior myodome. Anterior- ly, the basioccipital is attached to the prootic bones and dorsally with the exoccipital bones. Its lateral flanges expand ventrally to meet the flat posterior flanges of the parasphenoid. Posterior- ly, the lateral flanges fuse to form a circular margin in a slightly backward oblique position and attach to the margin of the first vertebral centrum. There are a variable number of small pores in a shallow depression on the lateral surfaces of the basioccipital. This depression is deepest in S. sinensis but does not approach the basioccipital depression characteristic of the bonitos, Sardini (Collette and Chao 1975). Exoccipital. — The exoccipitals connect the skull to the first vertebra dorsally. The exoccipi- tal articulates with the epiotic and supraoccipital bones anterodorsally, the intercalar laterally, and with the other exoccipital posterodorsally In ventral view, the exoccipital articulates with the prootic anteriorly, basioccipital medioventrally, and intercalar laterally. In posterior view, the foramen magnum is framed by the exoccipitals. Laterally, there are two foramina. The small anterior glossopharyngeal foramen (Allis 1903) lies close to the posterior border of the prootic. The large posterior vagal foramen lies just under the overhanging shelf formed by the posterior margin of the exoccipital. Dorsally, a small fora- men which opens into the brain cavity is present at the medioposterior corner of the exoccipital. Branchiocranium The branchiocranium is divided into five sec- tions: mandibular arch, palatine arch, hyoid arch, opercular apparatus, and branchial apparatus. MANDIBULAR ARCH. —The mandibular arch is composed of the upper jaw (premaxilla, maxil- la, and supramaxilla) and the lower jaw (dentary, angular, and retroarticular). Teeth are borne on the premaxilla and dentary, and the number of teeth on these bones is a useful taxonomic char- acter (see Dentition section). Dentition. — Large, triangular, laterally com- pressed teeth are present in the upper and lower jaws of Scomber omor us. Acanthocybium has sim- ilar teeth that are a little blunter and more tightly packed. Grammatorcynus has long thin teeth that are slightly compressed laterally. Bo- nitos have conical teeth that are larger than the conical teeth of the higher tunas (Thunnini). Tooth replacement in Scomberomorus cavalla was studied by Morgan and King (1983). The number of jaw teeth in Scomberomorus varies widely with a range of 5-39 in the upper jaw, 4-37 575 FISHERY BULLETIN: VOL. 82, NO. 4 in the lower jaw (Tables 3, 4). Two species of Scomberomorus stand out from the rest, S. multi- radiatus with the fewest teeth (5-10, x 8.0 on the upper jaw; 5-11, x 7.8 on the lower jaw) and S. concolor with the most teeth (13-37, x 22.2 on the upper jaw; 12-34, x 19.7 on the lower jaw). The 18 species can be ranked from lowest to highest as follows (mean for upper jaw followed by mean for lower jaw): 1) multiradiatus (8.0, 7.8); 2) queenslandicus (13.3, 10.6); 3) semifasciatus (12.8, 11.2); 4) cavalla (14.0, 10.9); 5) koreanus (13.7, 11.2); 6) commerson (14.1, 11.3); 7) sinensis (13.4, 12.2); 8) brasiliensis (14.0, 11.9); 9) lineo- latus (15.1, 12.9); 10) guttatus (16.9, 14.4); 11) sierra (17.3, 14.1); 12) maculatus (16.8, 14.6); 13) munroi (17.5, 15.0); 14) plurilineatus (17.9, 15.4); 15) tritor (18.6, 15.4); 16) regalis (19.3, 15.8); 17) niphonius (19.6, 15.9); and 18) con- color (22.2, 19.7). The species with the fewest teeth,