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 1980 1 ). 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 228 Ra. 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 228 Ra 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 " 8 RA and 228 Th and bomb-produced 14 C. 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. Bowman 1 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 7 3 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/d 2 . 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: A L = 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 2 5 O i STOMACH TISSUE WEIGHT/ BODY WEIGHT EXPONENTIAL CURVE FIT r 2 = 091, =0 107, b *0092 x STOMACH CONTENT WEIGHT/ BODY WEIGHT ,2 ., EXPONENTIAL CURVE FIT r^ = O 94, a = 006, b = 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 = ae bx ). 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 2 V- V) \ 1 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 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 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 1 0.4 0.1 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 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 0.435 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 1 + + Crustacea 8.7 100 75 2 27.9 39.9 13.0 2 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 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 — — 7 16 19 9 0.1 — Euphausiacea 7.5 4 66 8 3.3 13 5 5 2 0.6 1 Mysidacea 0.9 — — 0.3 3.8 05 + + Cumacea — — — 0.1 + + Copepoda — — — — — — Other Crustacea — 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 6.6 20 8 — Phycis chesten — — — — — — — — Ammodytes amencanus 81.9 — — — 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 3 45 83 1 20 54 47 Mean stom. cont. wt. (g) 099 152 670 0.571 0.673 1.597 8.185 17 826 Mean fish FL (cm) f^0 12.0 18.5 23.4 28 32.9 37.7 46 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 1 + 0.1 Decapoda 49 26 6.5 Crangonidae 2.4 1 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 7 + Pisces 87 5 78.4 76.4 Scomberesox saurus — — 6.1 Clupeidae 32 1.3 5 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 F 005 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 — 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 5 2.0 44 59 03 11.1 258 Pandalidae 1.0 2.4 — 7 26 — 11.7 13.3 Pasiphaeidae 0.6 — — — — — — — Other Decapoda 1 0.4 1.7 + 04 0.1 0.1 75 Euphausiacea — 0.2 4.4 — 14.4 3 + — Mysidacea — + — — — — 5.4 — Cumacea — + + + 0.1 — + — Copepoda — — + + — + — — Other Crustacea + 03 3 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 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 1 0.8 02 0.5 Ampeliscidae — 1 6 + 1.5 0.1 0.2 + + Oedicerotidae — — — — + — + + Hyperndae — 1 — i 9 + 0.5 0.1 5 Other Amphipoda + + + 6 + 1 0.1 + Decapoda 1 8 8 4 1 13 7 6 9 9.7 5.5 1.2 Crangonidae 02 0.9 + 4 5 2.0 04 4 7 2 Pandalidae 0.9 7 3 — 7.0 1.8 9 1 8 1.0 Pasiphaeidae — — — — — — — — Other Decapoda 0.7 2 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 6 902 107 952 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 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 + 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) — 0.108 180 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 0.976 603 0.515 Number in sample 17 67 51 33 47 49 63 58 Small fish Adjusted weight (g) 0256 0.036 200 0.074 0.414 184 0205 0.149 0.269 111 Number in sample 73 15 4 49 35 62 39 58 Georges Bank Large fish Adjusted weight (g) 0400 743 0.916 0.576 1 239 0506 0.735 0.734 0823 640 Number in sample 43 58 50 53 32 57 27 51 Small fish Adjusted weight (g) — 140 0.321 0.325 0566 0.106 0473 117 0453 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): 826 Developing 1.004 Ripe 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 — - OD 1 - 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. Larson 1 and Edward E. DeMartini 2 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. Carr 3 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. 3 M. 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 (Dean 4 ) 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). 4 T. 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 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 Light 5 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 rockfish 2 — — 400 Embiotocidae Sebastes serranoides, olive rockfish 2 4 175 — Brachyistius frenatus, kelp perch — — 25 Sebastes spp,, juvenile rockfish 2 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. 2 Members 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- W h (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 m 3 ) 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 m 2 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 D m = 2 W h D h , where D m was the estimate of stratified mean density in the 3 m, 7.6 m, and 12 m strata; W,„ the weighting factor; and D h , the mean density on that day in stratum h (Snedecor and Cochran 1980). The mean (D wc ) and variance (S 2 U J of these daily estimates were then computed. The mean {D b ) and variance (S 2 6 ) 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 XW h D n /!> where A st was the stratified mean estimate of integrated abundance over 100 m 2 of bottom, W h was the weighting factor, and D h was the mean density in either the above-bottom strata {D wc ) or in the bottom stratum (D h ). The term in the summation is the estimate of stratified mean density (per 1,000 m 3 ) over all strata, and the ratio (1,500/1,000) converts this value to abundance over 100 m 2 of bottom. The standard error of A st was calculated as 4 " v v 1,000 ' h h '' h where S 2 h was the variance of daily density estimates in either the above-bottom (S 2 ,,,.) or bottom (S 2 b ) 42 LARSON and DeMARTINI: DISTRIBUTION OF FISHES IN KELP FOREST strata; W h , the weighting factor; and n h , 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 m 3 ) 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 m 2 . 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 . < 1 284 + 1 893(0.582 * <8> * -*e e- • INTO SUN : AWAY FROM SUN 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, P 2 , and P 3 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 , ? 7 l P 3 (m) square Logistic All 284 1.89 0.582 3.52 0369 Y Away 0.259 2.63 560 3.86 250 Y= 1/(P, + P 3 ) Into 317 1.20 0618 3.15 0355 Gompertz All 1.27 -3 19 0647 ' 3.56 0370 Y=e |P . + P . p J Away 1 37 -3.88 648 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, + P 2 /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 m 3 ) 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 m 3 ) Biomass densi ty (kg/1.000 m 3 ) 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 1 372 372 barred sand bass 0.02 0.02 0.13 0.04 3.30 0.70 0024 0.024 0.173 046 4.434 0930 kelp perch 1 39 0.26 0.23 13 0.02 0.02 0.035 0.007 0006 0.003 neg. black perch 12 0.07 2.25 065 0046 0.028 0.717 209 white seaperch 1.91 1.21 3.16 1.20 2.33 0.86 3.07 59 0.319 0210 0491 0.209 287 0.105 0.376 0108 pile perch 002 0.02 0.08 0.05 0.66 0.11 009 0.009 0039 0.025 0263 0.079 rubberlip seaperch 0.04 0.03 1.08 35 028 0.017 0634 0.265 rainbow seaperch 2.02 0.92 0.167 068 senorita 2695 6.53 2445 5.78 4.66 2 22 14.16 5.95 0950 0.223 1.103 0225 0.241 110 0566 0.237 California sheephead 013 0.06 4.87 1 16 0058 0.040 1.561 0338 rock wrasse 1.20 1.24 0.237 0022 opaleye 0.03 0.03 0033 0.033 halfmoon 0.27 0.20 0.08 0.05 0.02 02 006 006 0068 0050 020 012 0006 0.006 0.015 0.015 blacksmith 002 0.02 neg. garibaldi giant kelpfish 0.35 0.08 009 0.04 0.08 0.04 0.18 0.12 018 0.007 0004 0003 0.015 008 0.014 0.012 cabezon 0.06 0.06 0089 0089 Calif, scorpionfish rockfish spp. 0.04 0.02 0.06 0.06 0.003 0.003 0.024 0.024 black croaker salema silversides 4.21 1.54 0092 0.029 jack mackerel 0.09 0.90 8 77 8.74 0.50 0.36 0010 0.010 1 008 1.005 0057 0.041 Pacific barracuda leopard shark 0.06 0.06 119 0.119 thornback bat ray 006 0.06 0.397 0.397 Pacific electric ray 001 0.01 003 002 0.02 0.02 0136 0136 320 196 0.154 0.154 45 FISHERY BULLETIN: VOL. 82, NO. 1 TABLE 7.— Mean numerical and biomass densities (per 1,000 m 3 ) 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 m 3 ) Biomass density (kg/1 .000 m 3 ) 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 12.87 395 726 162 1.101 0.440 1621 0.672 2 363 0675 barred sand bass kelp perch 0.83 0.20 0.19 0.09 0.12 0.11 0.02 0.08 3 14 029 0.021 005 0005 0002 0.178 0.029 0003 0.002 3446 0.577 black perch white seaperch pile perch 3.50 2.38 0.04 0.04 415 1 52 18 4.83 0.23 0.11 94 0.13 4 77 0.63 3.64 148 1.74 0.18 0.582 407 013 0.013 0681 0269 040 0.017 693 180 0.105 0.068 1.401 089 399 0.137 0.682 0.056 njbberlip seaperch rainbow seaperch sehonta California sheephead rock wrasse 19 46 2.82 0.02 02 2104 3 57 0.60 0.23 0.02 0.02 568 1.52 0.06 1.82 0.42 0.03 064 0.24 249 0.71 13.31 7.77 13.66 1.29 1.86 0.49 0569 0.078 0.017 0.017 1.039 0.158 0.181 0.119 0005 0.005 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 1.09 0.44 2.92 1.83 0.35 0.19 0.12 0.12 0.237 1 10 0730 0.457 0087 0047 0.030 0.030 blacksmith 0.03 0.04 neg. garibaldi giant kelpfish cabezon 0.28 0.06 10 0.02 0.12 0.07 012 0.07 007 0.06 0024 0007 0008 0004 0.014 0.009 0.012 0.010 0099 0099 Calif, scorpionfish rockfish spp. black croaker <) 006 006 11.85 11.85 0.033 0033 2.667 2.667 salema silversides jack mackerel Pacific barracuda 5 99 3.96 2096 9.05 0.13 0.13 19.34 17 69 0.61 0.61 3.32 3.32 889 5.93 0.120 0.079 2.410 1.040 0.019 0.019 2.224 2.035 0.092 0092 0.381 0.381 0.667 0.444 leopard shark thornback bat ray Pacific electric ray 0.12 0.12 0.794 0.794 TABLE 8. — Mean numerical densities (per 1,000 m 3 ) 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 m 3 ) 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 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 62 0.98 0.12 033 1.24 2.88 0.49 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 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 6 E. 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 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 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/lOOOm 3 ) ','« 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/1000m 3 ) 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 m 2 at SOK-U and SOK-D, respec- tively, with corresponding biomass values of 3.9 and 6.5 kg/100 m 2 (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 m 2 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) m 2 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 80 0.25 14 0.67 15 1.83 041 12 004 barred sand bass 0.55 0.1 1 0.52 0.04 116 0.37 0.74 14 0.58 0.09 1 69 0.55 kelp perch 085 0.20 0.57 05 0.02 0.01 0.01 0.01 black perch 038 0.10 0.78 10 54 032 0.13 003 023 0.02 0.19 1 1 white seaperch 3 76 1.43 6.05 1 54 046 31 0.55 0.24 93 030 007 0.06 pile perch 14 0.03 0.37 007 005 002 0.06 0.02 0.15 003 002 001 rubberlip seaperch 18 0.05 0.10 0.04 03 002 0.11 0.04 0.07 02 0.02 0.01 rainbow seaperch 030 14 037 1 1 001 12 0.03 0.01 04 001 001 001 senorita 28 86 4.63 2388 2.06 2 16 77 1.17 0.21 0.95 05 0.06 004 Calif, sheephead 78 0.18 2.89 0.30 0.61 0.20 0.26 005 1.12 17 18 006 rock wrasse 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 0.02 001 0.01 0.02 0.02 0.01 001 halfmoon 0.20 13 2 04 1 03 11 005 0.05 003 051 026 0.03 0.01 blacksmith 001 0.01 001 0.04 neg neq garibaldi 0.02 0.01 neq giant kelpfish 028 007 021 0.04 0.02 001 0.02 001 cabezon 0.02 001 0.01 0.01 0.01 0.01 0.02 0.02 Calif, scorpionfish 001 001 0.02 001 neq 0.01 001 rockfish spp 0.02 0.01 0.04 0.03 0.01 001 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 (F 2 12 = 9.42, P < 0.01), but an a priori comparison of SOK-U and SOK-D versus the cobble area was not significant (F x 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 (F 2>11 = 0.25, P > 0.75), even though the point estimate of 2.4 kg/100 m 2 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 m 2 at the "kelpless" cobble area, 23.11 ± 1.47 plants/100 m 2 at SOK-U, and 30.18 ± 1.69 plants/100 m 2 at 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 m 3 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 m 2 (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- tini 7 Unpubl. data). In particular, the densities of resi- 7 E. 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/m 2 ; 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 SHRIMPS 1 James A. Coyer 2 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], 93 f ; |B|) and biomass (90 [C|, 89 [M|, 86 f i [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 19 42 002 002 040 10 10 10 16 15 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/m 2 ) 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 ±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 (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 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) 1 ±<0.1 Asteroids (juv.) 0.2 ±0.1 1.1 ±1.3 •(2.7.2.0) 6±08 Jaeropsid isopods 0.1 ±0.1 0.2 ±0.3 "(2.3.0.3) 1 4 + 09 Cumaceans — 02±0 2 3 ±0.3 Brachyurans (zoea) — — 0.1 ± <0.1 Ophiuroids (juv.) — <0.1 ± <0 1 <0 1 ± <0 1 Tanaids — <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 ^ I W^: . :.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 n 1 : 1 : 1 : 1 .". ::9s; ^\\\\\\\^\\\\\\^^ -E ::£S:: 4^k\\\\\\\\\\\\\\\^^ ^2 o O Z3 Q. < o O Q. < O I s - 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 — £ I s - ■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. DeLong 1 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 octopus 1 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 japonicus 2 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 stilbius 2 California smooth- tongue 2 1.0 Senphus pohtus queenfish 2 10 Zalembtus rosaceus pink surf perch 0.5 Anoplopoma fimbria 2 sablefish 05 Poncbthys notatus plainfin midshipman 0.5 Ictchthys lockingtoni 2 medusafish 5 Stenobrachius leucopsarus 2 northern lampfish 05 Octopus bimacu/atus 2 two-spotted octopus 0.5 1 Pelagic life stage 2 Not 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 R 2 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 84 977 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 ' 10 20 30 40 50 60 70 80 90 100 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 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 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.Kato 4 ). 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- son 6 ). 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. 4 S. 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". fi E. 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. Griswold 2 and Thomas W. McKenney 3 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. 2 Northeast 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 m 1 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 Formalin 4 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. 4 Reference 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 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 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 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 c i 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 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 88 8 54.9 507 25.4 7 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 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 36.0 f /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 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. Hettler 2 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% glycerin 4 . "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) 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- 36 r 32 28 e I | 20 LLI _l D 16 DC < D I 12 co 8 ./: 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 70.2 — 8 2 5 — 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 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 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 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 1.9 54 5.0-5.9 15 81 — — 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 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 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 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 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 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 14 2 15 7 106 23.2 4 4 7.4 19 0-199 9 76 9 57 3 48 16 1 16 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 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 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"—' / ■y p M* # CO a UJ Q > Q O CO 30 20 10 40 r 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 Watson 1 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. 2 Marine 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 1979 3 ; Bartlet et al. 198 1 4 ). 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. 4 Barnett,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 m 3 were adequate to attain asymptotes of numbers of taxa per tow. A sampled volume of 400m 3 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 m 3 . 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 study 5 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 m 2 ) was used. A General Oceanics 6 (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 6 Reference 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. Sowby 8 ) 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 m 1 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 m 3 (bongo), 400 m 3 (Auriga), and 200 m 3 (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 m 3 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 Sertic 9 ); 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 - 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 m 3 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. Smith 10 ). 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 m 3 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- ,0 P. 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 10 f /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% MgS0 4 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 00m 1 . To describe the cross-shelf abundance patterns of ichthyoplankton, a procedure was adopted involving Hotelling's T 2 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 T 1 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 m 3 (0.75/100 m 3 ). 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 m 2 ) 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 5 10 15 25 20 FISHERY BULLETIN: VOL. 82, NO. 1 Hypsoblennius spp. 22 September 1978 Mean Number/ 100 m 15 10 5 I i i l T~ V 1 5 ~r 10 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 10 20 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 m 2 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 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 90 103 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 08 66 53 3 Paralabrax spp. 0.1 08 34 3 97.8 84 .1 Parophrys vetulus 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, N b = N L, where N is number under 1 00 m 2 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. 100m 3 ) 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 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 73 3 15 11 52 33 1.11 2.93 004 26 0.64 flexion and postflexion stage larvae 7.46 3088 125.51 53 1 56 3 97 10 062 1.90 02 0.08 0.14 Senphus politus preflexion stage larvae 58 1.37 2.90 15 0.49 1.12 004 16 0.33 flexion and postflexion stage larvae 4 31 2064 95 57 0.50 1 90 5.86 0.10 0.22 105 FISHERY BULLETIN: VOL. 82, NO. 1 DISTANCE FROM SHORE (km) 12 3 4 5 6 7 Engrauhs mordax eggs 10 20 30 -|40 50 60 70 J 80 DISTANCE FROM SHORE (km) 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 10 20 £ 30" - 40 fE ^°£ 60 o 70 J 80 Genyonemus hneatus\ preflexion stage larvae 10 20 30 40 50 60 - 70 J 80 Genyonemus hneatus% flexion and postflexion stage larvae 10 20 £ 30- - 40 f -50 a- 60 £ H70 80 Senphus pohtus preflexion stage larvae o 10 20 30 40 - 50 - 60 - 70 J 80 Senphus politus flexion and postflexion stage larvae 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 m 5 (0.75/100 m 3 ). 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 1979 11 ; 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, + CT M + CD I'* + TD m + DP [mlllJk] + CTD (l/k) + e, lkm i /km M c , oT OP trnt,,lk) crb m 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 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 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 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 4 180 0030 097 4 36 1.2 1980B 13.1 -0.004 0.019 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 T H 2h (3 z 22 HI 1 ia Q CC 14 < a 10 z < b \- rn in 1980B 1980C 1981B 40 60 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 > (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 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 08 099 5 60 1.3 1982A 1.45 12 0.08 0.83 3 38 1.1 1982B 4.90 10 11 053 4 39 1.2 116 McGURK: PACIFIC HERRING OTOLITH RINGS not significantly different from (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 (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 (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 i l( ) 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. Methven 4 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 0-16.0 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. Hull 5 ); 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 rh i a n= 83 20 1 r> = 2 . d n = 128 10 — .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 20 40 60 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 P p , n,n^ , , J j , ^m^rn* p a , p A , ' A E*4 | | | I 200 r f-i*V- r A i— I 6 ^ I I I 100 50 Melanogrammus aegleflnus -rt, , f , JU q ^ rrfWYf^ r' AM i | ■*■ i i I'l' V^V t i [ M. i M '*' »' ■* * i A T*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 - Bay of Fundy - n -. 6 Passamaquoddy Bay ,n -i 1 M " l " . 1 1" 1 1 1 Aug. 13, 1980 10 r n= 3 a n = 4 -p — i r"-i 1 — t P-°- 10 Oct. 9, 1980 n = 2 □ n = 4 -i r r 20 Dae. 13, 1980 10 ■1 10=35 30 1 r l i i i Jan. 15, 1981 1 1 20 - JL ,„•:!? 10 - 60 #1 i ' i ' I i i ' I SO ~ 40 _ April 29, 1981 30 - n ; 3 , on: 163 20 - 10 - ^ <"i 1 i r - i i 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 100 r 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-jlV v P < Pi i a M *i , i . ^ n i^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 Ma i , " 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 — 10 200 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 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. ELLERTSEN 1 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 Counter 2 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 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/m 2 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 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 21 JU H 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 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 30 J 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/m 2 ) 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/m 2 (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. q 50 >150 20- 40 r~ft~^~ Particle density \C- 150 nos. per liter ' 100 40 fe\ 40 30 l^^ x 20 7 rO 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 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/M 2 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/M 2 SURFACE 24 APRIL 1981 1000M LARVAE/M 2 SURFACE 120 100 80 60 H 40 20 LARVAE /M 2 SURFACE HO- 120 - 100 8 60H 40 20 H 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/m 2 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 5 10 0-«. 10- """^-° °c 20- 30- ./" 19-20 HOURS 5 10 15 22-23 HOURS 01-02 HOURS 0^-05 HOURS 09-10 HOURS / 10 15 \ O / 10 10 a \ O J \ \ JOLarvae/rrr 10 Q- 20- LU Q 3 50 100 "I 2 50 100 v ? 4 1 \ \ • 3 a 50 100 J L. \ j a \ 50 ip o q_ 1 2 \\ / /I 50 ip o g_ 1 2 50 100 4 I / V »/oFI Naupl/larv K FIGURE 1 1 .—Distribution of first feeding cod larvae (per m 3 ) (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 20 00 ,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 £ £ 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 15 40 H g-IOH Q 20 30- 40 J 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 17 14 H Myrland 26 Apr. 81 21°° 2 n. miles 17 50 H • 0- Q. 2 ? >20 ^~""— 10 Q 10- ] J / 20 -^ 20- V A 10 <10 30- 10 z 10 ^ > 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 22 47 Skiva 29-30Apr. 81 2 n. miles 40 J 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 Heath 4 incubators in 10 U C 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 10 U C 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, R 2 = 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 14 N 16 X LENGTH 7 3 8 5 91 97 IOI 112 12 134 14 9 168 18 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 5 5/1-5/15 1 5 1.0 3 S 'i 370 5/16-5/31 34.5 420 27 19.0 5 5 6/1-6/15 8.0 16 180 37.0 355 78 7.0 6/16-6/30 0.0 1.0 0.5 6.0 3.0 40 35 7/1-7/15 5 0.0 0.0 1)0 5 1 20 40.0 7/16-7/31 5 1.0 0.0 0.0 o 5 18.5 8/1-8/15 00 00 o o 0.0 05 260 8/16-8/31 00 0.0 5 1 1.0 5 7 5 9/1-9/15 0.0 0.0 0.0 1.0 o o 1.0 3 5 b 2.0 29 9/16-9/30 o o 0.0 0.0 00 1 3 10/1-10/15 0.0 0.0 i) 5 o 0.0 0.0 00 10/16-10/31 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 00 0.0 11/16-11/30 0.0 0.0 oo 00 05 0.5 2 5 50 12/1-12 15 0.0 0.0 00 00 0.0 0.0 00 1.5 1.5 0.5 12/16-12 31 00 on I 1 (III 0.0 0.0 0.0 00 00 00 1/1-1/15 0.0 00 0.0 0.0 2 1.5 1 Total percentage migration 505 61 62 73.0 81.5 85.0 490 60.0 37 36 13.0 6 5 Percent residuals 0.0 2 00 2.0 14.5 31.5 43 5 55 Total percentage recovered 50.5 61 62 750 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 1 3/1-3/15 3/16-3/31 3 4/1-4/15 4/16-4/30 0.1 1 37 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 33 8 6/1 6-6/30 1.7 1.0 3.1 77 7/1-7/15 5 00 7/16-7/31 1 3 1.0 0.0 2.0 55.0 8/1-8/15 Oil 0.0 0.0 0.0 o 5 8/16-8/31 7 0.0 5 1.0 5 54 5 9/1-9/15 2.2 ..I \, 2 5 1 2 5 50 Total percen m gration 92 5 84.8 85 3 81 6 81 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 85.3 f /(: of the fish released from 15 March through 15 June were recovered. Only 58.5 and 59.5 f 7f 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 r l6 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 5/26 097 -*~»— • -^ti -10 3 Hatchery 5/15 6/7 097 8 10 12 14 16 18 20 22 24 26 28 30 13 5 7 JUN 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) 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 21 C C 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/m 2 /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, 10T 2 141 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 10 f A. 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 0T 2 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:10T 2 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, 10T 2 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/14T p :10T 2 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,:10T 2 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/14T 1 :10T 2 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,:10T 2 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,: 10T 2 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 3 DC I t- n 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. Neilson 2 ). 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. Brothers 3 ). 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,( =-7.6,andformalesL oc = 59.2,& = 0.03,r o =-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. 4 M. 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 + 79 TL = 0.379+ 82 FL FL = 0.088 + 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 Isomer 5 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) , L 3 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' 120 c 26' Refugio to El Capitan. 8 m (DR) 34°27' 120' 02'- 120°05' North of Refugio. 22 m 34°27' 1 20 r 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 net 6 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. 8 McGowen, 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 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 L t = L x [1 - exp -k {t-t )\ = length at time t = theoretical maximum length = constant expressing the rate of ap- proach to L x = theoretical age at which L, = 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 L m 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 cm 9 ) 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 = aL h , where W = weight in grams, L = total length in centimeters, and a and b are constants, with values determined using log 10 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 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 L 30239 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 2 16 15 6 140 1 1 26 26 8 15 21 73 72 15 16 18 88 88 27 17.0 2 20 91 90 35 tr 1 1 18.0 6 21 96 94 61 tr tr 19.0 7 21 100 1 ( 10 83 48 tr 20.0 tr 7 23 100 100 82 52 tr tr tr 21 5 31 100 100 94 51 2 tr 22.0 tr 6 32 99 99 93 58 1 23.0 7 48 100 100 95 60 tr tr 24.0 tr 6 47 100 100 93 58 2 25.0 6 47 100 100 99 57 2 26.0 tr 6 46 100 100 98 59 1 tr '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. 10 T. 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 Verdes 1 ' and Laguna Beach and n Genyonemus 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 m 3 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 m 3 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 km 3 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 km 3 (50.6%) is located in the region between Point Con- ception and Playa del Rey, 7.9 km 3 (25.9%) between Redondo Beach and Laguna Beach, and 7.2 km 3 ■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 m 3 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 10 10 , 1.75 X 10 10 , and 2.97 X 10 9 , 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 experiments 12 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 l4 D.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. Morris 1 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 1980 3 ; Stephens and Palmer 1979 4 ). 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. 4 Stephens, 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 10 6 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 ID 5 ). 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 (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 \Px r Py, '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 = 416 Fish length (SL mm) Mean = 1 72.8 mm Mean = 1 72.2 mm SD= 12 9 SD= 14,3 t = 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 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 5 :> 2 48.6 131 2 2.12 0.2 0.6 52 1 41.7 Brachyuran zoea 1 89 03 9 22.9 27 5 8 14 09 1.2 46.6 97 9 Gammandae 1.83 3 1 1 42.9 60 1 1 11.33 12 9 25.3 91.8 3,506.8 Pagurid zoea 1.63 3 h 25.7 23.1 3 27 0.4 0.5 43.8 394 Cladocera 1.60 0.3 4 28.6 200 0.49 0.1 0,1 20,5 5.7 Hhincalanus 1 .17 2 06 34.3 27.4 2.27 3 0,6 41.1 30.9 Euphausids 097 2 04 25.7 154 082 0.1 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 1 02 22 9 69 1.10 1 0.2 37,0 11.1 Fish eggs 0.49 0.1 0.2 28.6 8 6 0.53 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 Polyophthalmus pictus 0.34 0.1 1 2.9 3 25.70 3.0 6 8 28.8 282.2 Gastropoda 0.34 0.1 2 22.9 6 9 0.37 1 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 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 1 03 30.1 12.0 Decapoda. misc. 0.06 0.1 2 5.7 1.1 0.55 1 9 30.1 30 1 Caprellidae 003 1 1 2 9 3 1.90 2 1.0 466 55 9 Porcellanid zoea 003 0.1 1 2 9 3 0.89 1 0.2 32.9 9 9 Pelecypoda n 0.60 1 0.4 24,7 12 4 Anemone 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 stomachs 1 In plar ikton 2 Prey items Discharge Point Discharge Point Polyophthalmus pictus Mean 25 70 0.34 30.59 SD 67.49 2.03 86 52 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 < 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= 205 r = 505 P= 0616 'Mean number of prey consumed per fish. 2 Mean number per 100 m 3 of water sampled. 3 Note: 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, 6 M. 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) 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 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 7 Kinnetic 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. 2 Sea 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 (Hui 4 ). 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 1 5 ), 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. Carbocaine 6 (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- 4 Clifford 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 RDO 7 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 ABC P 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 environment 8 . 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-\p C 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 — — — — — — — — — — — — WFP669 1 240 240 172 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 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 87 713 613 58 1 55 — — — — 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 L UU J 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 3 5/10 Third 2 Feb. 1981 — 4 11.3 None 3 6/11 Unscheduled medical treatment 1 8 d in duration. Not examined because of poor preparation of section. 3 Cementum 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 104 2 21 July 1977 'Born in captivity. 2 Born 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- 9 L. 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 CAROLINA 1 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% Formalin 3 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. 2 North 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 •i a n U iS rt *y Ca P e 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 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. MacPherson 4 ). 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 4 K. 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 (274) (1) 82.4 (17) 4.8 (292) 0(219) 0(219) (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 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 00 90-99 0.0 1 00- 1 09 4 0.0 110-119 4 0.0 120-129 6 9 60.0 130-139 1 12 92.3 140-149 1 22 95.7 150-159 17 100.0 160-169 1 26 963 170-179 7 100.0 180-189 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 I 3 O o 1- 6 7 63 4 ^ *** 88 8 4 9 9 8 8 7 6 4 8 8 7 4 6 9 9 9 8 8 8 '788 8 8 9 6 8 4 9 9 8 6 «8> 9 6 9 89 9 99 9 9 8 9 4 « 6 ^ ' 6 9 7 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> • / • Q 10- • • /• •• • • ••• • • /•• • • 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 (x 2 = 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 X 1 <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 X 1 928* 3.64 7.28 763 10.38* 'P < 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 HC1 1 ) 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( 2 1) 6 3 21| 2 1) Spratelloides delicatulus IIIA 6 1MB IVA IVB Total 5( 2 2) 9 9( 2 1) 29( 2 3) 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 17 5 3 2 6 7 12 'Otoliths of two treated fish were destroyed by poor preservation. 2 Number 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 D (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 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 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 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 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 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 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 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 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 +e b+cX (1) An SAS 1 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 or 1 cement glands, indicating that egg extrusion would not occur. The largest female lobsters with cement glands in stage 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) 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) 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 - — 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 D M 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 (S vx ), 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 74 999 185 532 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 102 999 102 394 -0.304 1 266 4.153 0.737 0.788 3 278 diploproa 80 0999 87 308 1 286 1 242 2 718 -0.740 804 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 136 997 196 565 11476 1 239 8002 -7 447 0803 6.443 rosaceus 83 0.996 132 263 3.917 1 199 2 867 -1 794 828 2 383 rosenblatti 104 999 132 428 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 48 999 71 126 0572 1 234 1 071 -0.231 0808 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 114 999 72 426 -0.369 1 228 4.126 0575 0.813 3 358 aurora 44 993 164 324 -3 046 1 201 4 237 6237 0.821 3 502 babcocki 76 0999 185 532 9034 1 153 5 190 -6 860 865 4.496 carnatus 104 0999 75 292 4.601 1 194 2 425 -3 613 836 2.030 caunnus 1 17 996 11 1 448 5896 1 187 6 764 -2 272 836 5 674 chlorostictus 107 0999 107 382 5 289 1 171 3.719 -3 987 852 3 173 chrysomelas 58 0997 77 226 1.137 1 209 3 209 -0009 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 -11 724 892 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 72 994 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 34 999 190 717 033 1 1 77 8 446 0688 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 102 387 352 1 192 4.975 644 0836 4.166 nebulosus 71 993 213 366 6934 1 181 4623 -1.852 0835 3 888 ovahs 83 0996 181 375 -3 554 1 187 4 677 5 130 836 3925 paucispims 162 0999 103 649 -4 082 1 209 6819 4 183 826 5636 pmniger 138 0998 196 565 12880 1 164 7,440 -9 326 855 6378 rosaceus 83 0997 132 263 1399 1 187 2.730 0.225 837 2 293 rosenblatti 104 0999 132 428 9 938 1 147 3.347 -8.023 0870 2.915 ruberrtmus 1 18 996 203 565 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 -12 392 898 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 1.065 964 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 1.010 949 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 1.182 966 0.541 1 No 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. Sr 4T£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. Serchuk 1 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 1980 2 ). 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. 2 Serchuk, 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 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. .« \# ° 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,- 5 s "cO .3 Sh o : "? c OJ in -5 > ^ 0; '« — C *J ~ +J *o CO ^ O < z CO bo > quah ioned B t3 cO a; 1; X 8 jj _C +-> f <*- be Js be — CS 01 B x. ■- en — * _ 4> £ % *^ -f g S CO S^ 0) -4-> > <~ lo ° > CO *> -s -B S 1 .5° 2 OS gi — *■« CO u 1 £ '. lO -w 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 2 20-29 8 4 3 15 30-39 16 9 3 18 2 2 50 40-49 10 3 1 18 8 12 1 53 50-59 5 6 11 -59 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 /"** ■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 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 90 c /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- 4 Ropes, 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 228 Ra 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. Pearcy 1 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. 2 Somerton, 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'Eastern 3 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 (empty) to 5 (stomach fully distended with food). The condition of the contents was assigned a value from (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 B n = BIR, which ranges from (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 1979 4 ; 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 p tJ and phj are the proportions of prey j found in the diets of species i and h respectively. This coefficient has a minimum of (no overlap 4 Cailliet, 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 43 9 683 406 19.8 1.32 167 23 2 88 68 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 01 — 2.5 40 — 0.02 — 4.9 1.7 03 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 — 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 — 5.19 1.9 3.8 93 0.2 0.22 02 27 5.6 0.5 .12 0.1 — — — — — 0.5 1.0 — 0.10 — 4.3 1.9 03 02 — 1.6 2.7 0.1 0.04 — — — — — 05 1 001 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 — 0.93 02 1.3 30 — 068 0.2 1.1 10 — 057 — 38 1.0 — 21.24 149 22 1.5 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 — 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 — 1.56 0.3 — — — — — 1.3 10 — 27 — — — — — — — — 3.8 1 6 1 14.17 184 1.3 10 — 1.22 0.3 — — — — — 2.5 2.0 — 1.65 0.7 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 — 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 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 — 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 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 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 — 01 — 46 1.2 — 01 — Hyperoche medusarum — — — — — 08 20 — 001 — Phronima sedentana 0.4 1 — 0.03 Streetsia challenger* 04 20 — 03 Hypemdea unidentified — — — — — 1.5 1.0 — 001 — Rhacotropis sp 04 4 — 005 — 1 5 60 1 007 — Atylus tndens 04 1 — 001 Anonyx sp 08 1 5 — 18 Lysianassidae unidentified — — — — — 0.8 1 — 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 — 1 05 1 1.5 1 — 1.55 0.4 Crangon sp — — — — — 08 1 — 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 — 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 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 — 0.02 — Parathemisto pacifica 6.8 3.2 1.3 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 — 0.03 — 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 (B n ) 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 — 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 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 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 01 — Decapoda Sergestes similis 46.8 44 68 60 165 Pasaphaea pacifica 1.6 1.0 — 59 0.6 Benthogenema burkenroadi 1.6 1.0 — 12 0.1 Munida quadnspina 1.6 60 0.3 0.12 0.1 Cancer sp. megalopae 97 1.3 04 002 1 Decapod mysis larvae 1.6 10 — 001 — Mollusca Pteropoda unidentified 1.6 10 — 0.03 — Gonatus sp. 1.6 1.0 — 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 1.2 0.06 09 Euphausnd remains 33 1 12 0.01 02 Amphipoda Parathemisto pacifica 16.7 44 256 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 1.2 01 02 Euchaeta elongata 100 30 105 01 0.5 Copepod unidentified 16 7 30 174 001 0.8 Decapoda Sergestes similis 33 1 1 2 007 11 Osteichthyes Ammodytes hexaplerus 3.3 1 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 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 iB n 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 B n 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 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 Spring 2 (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 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). 2 No 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 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 , 3 54> : 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 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 } 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 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 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 5 Gunderson, 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 (Barss 6 ) 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- 6 W. 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. 38