IVIBL U.S. Department of Commerce Volume 88 Number 1, 1990 arine Biological Laboratory LIBRARY I -^UL 2 mo t i?iB! A«B ^IV. .lAk K^H .jm^ .It U.S. Department of Commerce Robert Mosbacher Secretary National Oceanic and Atmospheric Administration John A. Knauss Under Secretary for Oceans and Atmosphere IMational IVIarine Fisheries Service William W. Fox, Jr. Assistant Administrator for Fisheries Scientific Editor Dr. Andrew E. Dizon La Jolla Laboratory, Southwest Fisheries Center National Marine Fisheries Service, NOAA P.O. Box 271, La Jolla, California 92038 The Fishery Bulletin (ISSN 0090-0656) is published quarterly by the Scientific Publications Office, National Marine Fisheries Service, NOAA, 7600 Sand Point Way NE, BIN C15700, Seattle, WA 98115-0070. Second class postage is paid in Seattle, Wash., and additional offices. POSTMASTER send address changes for subscriptions to Fishery Bulletin, Super- intendent of Documents, U.S. Govern- ment Printing Office, Washington, DC 20402. Although the contents have not been copyrighted and may be reprinted entire- ly, reference to source is appreciated. The Secretary of Commerce has deter- mined that the publication of this period- ical is necessary in the transaction of the public business required by law of this Department. Use of funds for printing of this periodical has been approved by the Director of the Office of Management and Budget. For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. Subscrip- tion price per year: $16.00 domestic and $20.00 foreign. Cost per single issue: $9.00 domestic and $11.25 foreign. Editorial Committee Dr. Jay Barlow National Marine Fisheries Service Dr. William H. Bayliff Inter-Amencan Tropical Tuna Commission Dr. George W. Boehlert National Marine Fisheries Service Dr. Bruce B. Collette National Marine Fisheries Service Dr. Robert C. Francis University of Washington Dr. James R. Kitchell University of Wisconsin Dr. William J. Richards National Marine Fisheries Service Dr. Tim D. Smith National Marine Fisheries Service Managing Editor Nancy Peacock National Marine Fisheries Service, Scientific Publications 7600 Sand Point Way NE. BIN CI 5700, Seattle, Washington 98! 15-0070 The Fishery Bulletin carries original research reports and technical notes on investiga- tions in fishery science, engineering, and economics. The Bulletin of the United States Fish Commission was begun in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. Begin- ning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin. Beginning with volume 70, number 1 , January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. It is also available free in limited numbers to libraries, research institutions, State and Federal agencies, and in exchange for other scientific publications. U.S. Department of Commerce Seattle, Washington Volume 88 Number 1, 1990 Fishery Biological Laburai LIBRARY ■'•" 2 1990 Contents Woods Hole, Mass. 1 Able, Kenneth W., R. Edmond Matheson, Wallace W. Morse, Michael P. Fahay, and Gary Shepherd Patterns of summer flounder Paralichthys dentatus early life history in the Mid-Atlantic Bight and New Jersey estuaries 1 3 Barry, James P., and Mia J. Tegner Inferring demographic processes from size-frequency distributions: Simple models indicate specific patterns of growth and" mortality 21 Benfield, Mark C, Jaap R. Bosschieter, and Anthony T. Forbes Growth and emigration of Penseus indicus H Milne-Edwards (Crustacea.Decapoda.-Penaeidae) in the St Lucia Estuary, southern Africa 29 Carls, Mark G., and Stanley D. Rice Abnormal development and growth reductions of pollock Theragra chalcogramma embryos exposed to water-soluble fractions of oil 39 DeBrosse, Gregory A., Adam J. Baldinger, and Patsy A. McLaughlin A comparative study of the megalopal stages of Cancer oregonensis Dana and C, productus Randall (Decapoda: Branchyura: Cancridae) from the northeastern Pacific 51 Fisher, Joseph P., and William G. Pearcy Distribution and residence times of juvenile fall and spring chinook salmon in Coos Bay, Oregon 59 Graves, John E., Michelle J. Curtis, Paul A. Oeth, and Robin S. Waples Biochemical genetics of southern California basses of the genus Paralabrax Specific identification of fresh and ethanol-preserved individual eggs and early larvae Fishery Bulletin 88(1), 1990 ^ 67 Hayse, John W. Feeding habits, age, growth, and reproduction of Atlantic spadefish Chaetodipterus fsber (Pisces: Ephippidae) in South Carolina 85 Herrick, Samuel F., Jr., and Dale Squires On measuring fishing fleet productivity: Development and demonstration of an analytical framework 95 Ho, Ju-shey Phylcrgeny and biogeography of hakes [Merluccius Teleostei): A cladistic analysis 105 Holt, Rennie S., and Stephanie N. Sexton Monitoring trends in dolphin abundance in the eastern tropical Pacific using research vessels oyer a long sampling period: Analyses of 1986 data, the first year 113 Jackson, George David Age and growth of the tropical nearshore loliginid squid Sepioteuthis lessoniana determined from statolith growth-ring analysis 119 Jefferson, Thomas A. Sexual dimorphism and development of external features in Dall's porpoise Phocoenoides dalli 133 Kendall, Arthur W. Jr., and Susan J. Picquelle Egg and larval distributions of walleye pollock Thersgra chalcogramma in Shelikof Strait, Gulf of Alaska ^155 Maillet, Gary L., and David M. Checkley, Jr. Effects of starvation on the frequency of formation and width of growth increments in sagittae of laboratory- reared Atlantic menhaden Brevoortia tyrannus larvae 167 Podesta, Guillermo P. Migratory pattern of Argentine hake Merluccius hubbsi and oceanic processes in the southwestern Atlantic Ocean 179 Steimie, Frank W. Population dynamics, growth, and production estimates for the sand dollar Echinarachnius parma 191 Wespestad, Vidar G., and Eriend Moksness Observations on growth and survival during the early life history of Pacific herring Clupea pallasii from Bristol Bay, Alaska, in a marine mesocosm Motes 201 Gauldie, Robert W., and Zophie Czochanska Hyperostosic bones from the New Zealand snapper Chrysophrys auratus (Sparidae) 207 Ratty, Frank J., R. Michael Laurs, and Raymond M. Kelly Gonad morphology, histology, and spermatogenesis in South Pacific albacore tuna Thunnus alalunga (Scombridae) 217 Uchiyama, James H., and Jeffrey D. Sampaga Age estimation and composition of pelagic armorhead Pseudopentaceros wheelen from the Hancock Seamounts Abstract.- Tlie simimer flounder Pnralichthii^ dentatu!< spawned throughout the Mid-Atlantic Bight and Georges Bank during 1977-85. Spawaiing peaked in fall hut extended from September through January. Planktonic larvae (2-13 mm) were most abundant in the Mid-Atlantic Bight September-May. At approx- imately 11-14 mm, some larvae entered New Jersey estuaries; but their occurrence, especially during winter and early spring, was spora- dic. Young-of-the-year were more frequently collected after May. Dur- ing the first summer inshore they grew rapidly and reached 160-320 mm TL. Young-of-the-year emigrated from the estuaries in fall and were most abundant on the shallow por- tions of the adjacent continental shelf. Some evidence suggests that young-of-the-year in the northern Mid-Atlantic Bight reach a larger size than those from the southern portion. An hypothesis to explain the observed rarity of small juveniles in northern estuaries in some years is that some juveniles utilize the conti- nental shelf as a nursery and enter estuaries at a larger size. This hypoth- esis requires testing. Patterns of Summer Flounder Paralichthys dentatus Early Life History in the Mid-Atlantic Bight and Mew Jersey Estuaries Kenneth W. Able Institute of Marine and Coastal Sciences and Biological Sciences Rutgers University, New Brunswick, New Jersey 08903 Present address Marine Field Station, Rutgers University, Tuckerton, New Jersey 08087 R. Edmond Matheson Institute of Marine and Coastal Sciences and Biological Sciences Rutgers University. New Brunswick, New Jersey 08903 Present address Florida Marine Research Institute St Petersburg, Florida 33701 Wallace W. Morse Michael P. Fahay Sandy Hook Laboratory, Northeast Fisheries Center National Marine Fisheries Service, NOAA, Highlands, New Jersey 07732 Gary Shepherd Woods Hole Laboratory, Northeast Fisheries Center National Marine Fisheries Service, NOAA, Woods Hole, Massachusetts 02543 Manuscript accepted 18 September 1989. Fishery Bulletin, U.S. 88:1-12. The summer flounder ParaUchtlujs dentatus inhabits estuarine and continental shelf waters from Nova Scotia (Leim and Scott 1966) to Florida (Gutherz 1967). Over much of this range it supports important commercial and recreational fisheries (Grosslein and Azarovitz 1982). The economic importance of P. dentatus has prompted research on various aspects of its biology (see Scarlett 1982, Grosslein and Azarovitz 1982, Rogers and Van Den Avyle 1983), but the spatial and temporal patterns of spawning and the nursery areas are not well known. Our understand- ing of the timing and distribution of spawning in the Mid-Atlantic Bight is based on examination of adult gonads (Morse 1981) and larval surveys (Smith 1973, Smith et al. 1975, Morse et al. 1987). Spawning occurs at temperatures of 12-19°C (Smith 1973), eggs and larvae are pelagic, hatcliing in the laboratory occurs approximately 48-96 hours after fertilization at 15-21°C (Johns and Howell 1980, Johns et al. 1981), and the pelagic larvae begin trans- formation at 9-12 mm SL (Smith and Fahay 1970, Smigielski 1975). Major nursery areas are assumed to occur in estuaries from Virginia and south (Poole 1966, Powell and Schwartz 1977, Grosslein and Aza- rovitz 1982), although juveniles also occur in estuaries in the northern Mid-Atlantic Bight (Poole 1961, Pearcy and Richards 1962, Pacheco and Grant 1973). Juveniles, pre- sumably, leave estuaries in fall, migrate offshore to overwinter, and return to the estuaries in spring with the adults (Hamer and Lux 1962, Murawski 1970), Although some individuals have been found north and east of their estuarine and offshore tagging locations, Fishery Bulletin 88(1 1990 Figure 1 Study area with suharea boundaries and important localities mentioned in the text. others returned to the original tagging locations in estuaries. Connicting evidence concerning the distribution of different stocks or subpopulations hampers interpreta- tion of the patterns of reproduction and identification of nursery areas. Several authors have suggested that Mid-Atlantic Bight and North Carolina populations may be separate (Ginsburg 1952, Smith and Daiber 1977, Wilk et al. 1980). An alternate interpretation is that inshore populations from Virginia to North Carolina may form a separate population from those to the north and offshore (Delaney 1986). Herein we attempt to clarify the patterns of repro- duction and early life history of P. dentatus in the Mid-Atlantic Bight. Our interpretations are based on extensive collections of eggs, larvae, and first- year juveniles from continental shelf and estuarine waters. Materials and methods Data for eggs, larvae and first year juveniles of P. den- tatus collected over the Mid-Atlantic Bight continen- tal shelf and in New Jersey estuaries (Fig. 1) derive from sources given in Table 1 and from Marine Resources Monitoring, Assessment and Prediction (MARMAP) surveys (Sherman 1980, 1986) by the National Marine Fisheries Service (NMFS). These surveys were conducted at monthly to bimonthly in- tervals over the continental shelf from Cape Hatteras. North Carolina, to Cape Sable, Nova Scotia. Sampling methodology can be found in Sibunka and Silverman (1984) and Morse et al. (1987). Collections made with 61 -cm bongo frames fitted with ().505-mm mesh nets were corrected for the depth of tow and volume of water filtered and expressed as number per unit vol- ume of sea surface. For graphic display of the egg and larval distribution and abundance, the catches per Able et al : Paralichthys dentatus early life history Table 1 Summary and sources oi Paralichthys dentat-us data and specimens of young-of-the-year from New Jersey, reexamined for this study. Study Collection Collection Number of Length range area method period individuals (mm TL) Source Southern New Jersey Miscellaneous trawls. 1930S-1972 538 10-320 Allen et al. 1978 estuaries and seines de Sylva et al. 1961 Delaware Bay Milstein and Thomas 1976 Acad. Nat. Sci. Phila. collections New Jersey inlets 1-m plankton net (O.ri mm mesh), trawls l%2-72 380 5-21 Festa 1974 Maurice River and 1-m plankton net (0.5 197.5-77 6 15-19 Himchak 1982 vicinity of Atlantic City mm mesh) Barnegat Bay Power plant screens, trawls 1976-80 1091 10-330 Vouglitois 1983 Tatham et al. 1977 Great Bay Recreational fishery creel survey 1967-76 1905 230-300 Festa 1979 New Jersey offshore Otter trawl 1985 173 16-300 Halgren and Scarlett 1985 Manasquan River Trawl 1984-86 128 35-320 Scarlett 1989 10 m- within each 25-km- block in the survey area were averaged for all years combined. Additional P. dentatus collections were made with a 0.5-m net (1.8 mm mesh) during 1987 from Little Sheepshead Creek which is immediately adjacent to Little Egg Inlet, New Jersey (Fig. 1). Lengths of lan'ae (Table 4) were record- ed as millimeters (mm) notochord length (NL) or stan- dard length (SL). Lengths for transforming individuals and larger young-of-the-year (YOY) (Figs. 4,5) are pre- sented as total length (TL). For purposes of com- parison, SL = 0.650 + 0.778 TL. Our assignment of P. dentatus specimens to the YOY group from inshore and offshore data sets was based on size frequencies and ages derived from scales. Age- ing criteria were based, in part, on the protocol devel- oped by Smith et al. (1981). Data for YOY collected over the continental shelf were taken during fall liottom-trawl surveys conducted by the Northeast Fisheries Center, NOAA. Sampling methodology for these surveys is described by Azarovitz (1981). \Kr^ .■■'4? #/ lOm^ /rectangle 40 38 0 • 01-25 . 26- 50 • 51-10 36 • > 1 0 74 7 2 7.0 68 66_ Figure 2 Distribution and abundance (cumulative mean no./ 10 nr of sea surface) oi Paralichthijs dentatus eggs from MARMAP collections during 1979-81. 1984. and 1985. Results Offshore egg distribution Composite collections of Paralichthys dentatus eggs during 1979-85 indicate that spawning occurs from Georges Bank to Cape Hatteras from nearshore to the edge of the continental shelf (Fig. 2). Eggs were most abundant in samples from subareas II-V (Fig. 1, Table 2). Eggs were collected as early as September (except in subarea I) and as late as December and January in subareas L IL and IV, although the number of stations sampled was small (subarea III) or zero (subareas IV and V) in December (Table 2). In all subareas the highest frequency of occurrence and greatest abundance occurred in October and November (Table 2). Offshore larvae distribution The spatial distribution of larvae was similar to that of the eggs except that larvae were much less numer- ous in subarea I (Fig. 3, Table 3). In general, larvae were most abundant in subareas II, III, and V (Table 3). Fishery Bulletin 88(1), 1990 Table 2 Abundance of Paraliclithi s dentatuf. eKgs( mean no./lOO ni" of sea surface ) from MARMAP survey.- during 1979-81 1984, and 1985. 1 by subarea and month, occ = number of stations wher e eggs occurred; nst = number of stations- sampled dash indicates no eggs collected; ns indicates no samples. Subarea Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun I Georges Bank — — — 45.2 42.8 2.0 0.9 — — — - — (occ) 0 n 0 42 17 1 1 0 0 0 0 0 (nst) 149 176 79 345 143 97 109 80 191 252 121 75 II Southern New England - - 1.3 171.8 97.1 19.8 — — — — — - (occ) 0 0 1 46 27 3 0 0 0 0 0 0 (nst) 152 116 35 143 63 58 74 13 183 109 49 57 III New Jersey — — 2.8 4.4 133.7 — — — — — — — (occ) 0 0 2 25 15 0 0 0 0 0 0 (1 (nst) 102 64 39 83 54 5 38 57 118 76 40 61 IV Delmarva Peninsula — - 14.3 86.3 83.2 ns 1.2 — - — — — (occ) 0 0 12 10 21 1 0 0 0 0 (1 (nst) 97 (;2 79 34 64 42 84 110 71 43 67 V Virginia Capes to Ca pe Hatteras - - 1.6 48.2 121.0 ns — — — — — — (occ) 0 0 o 5 17 0 0 0 0 0 0 (nst) 64 71 5S 25 48 34 66 83 137 162 74 Table 3 Abundance of Panillrhtl J/.S llflll iitus larvae (mean no /loo nr of sea surface) from MARMAP surveys during 1977-85 l)y subarea and month, occ = number of Stat ons where larvae occurred; nst = number of stati ms samji led; dash indii ates no arvae collected; ns indicates no samples. Subarea Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun I Georges Bank — — - 7.0 15.6 5.9 1.8 1.8 — — — — (occ) 0 0 0 20 12 7 2 2 1) 0 0 0 (n.st) 162 184 31 371 164 137 106 118 200 241 225 109 11 Southern New England - - 5.9 75.5 177.7 112.6 4.7 11.4 0.2 — — — (occ) 0 0 3 73 26 54 3 6 1 0 0 0 (nst) 169 170 29 186 49 113 49 56 207 184 154 124 III New Jersey - - 0.3 13.4 141.7 47.9 — 5.9 0.6 - — - (occ) (1 (1 1 26 23 10 0 5 •} 0 0 0 (n.st) 141 122 44 116 64 30 79 119 160 144 135 128 IV Delmarva Peninsul a - - - 53.8 52.9 27.3 — 2.3 0.1 0.7 0.2 (occ) 0 0 0 13 23 (> 0 7 1 2 1 0 (nst) 91 113 83 46 65 17 79 130 154 93 129 103 V Virginia Capes to C ape Hatteras - - - 3.7 176.8 ns — 1.3 3.6 17.7 3.4 - (occ) 0 0 0 4 30 0 3 8 14 8 0 (nst) 65 80 72 34 48 79 113 103 56 97 73 Able et al , Parahchthys dentatus early life history 44 76 74 72 70] 68 ME / LARVAE NH ( 1-3 mm >.»<: 42 «-4 ria J- a ' '• ■ 7 4 7 2 7. 0 6.8 66 44 76 74 72 10 j 68 ME/ 4-6 mm ^»i'' MA NY CT ^\tiji^ .\y y'42- 40 38 36 74 72 70 68 66 74 7 2 7.0 6.8 66_ 44 76 74 72 70i68 #/ 10m^ /rectangle 38 0 • 01-25 • .26- 50 • .5 1-10 • > 1 0 36 'A, ■-'^ _' 7 4 7.2 7.0 6.8 66_ Figure 3 Distribution and abundance (cumulative mean no. /10m- of sea surface) of Punflichthnx lU-xtatii^ larvae by length during 1977-84. Small larvae (<6 mm) were clumped in three regions (Fig. 3): Subarea II to northern subarea III, northern subarea IV, and subarea V. The northern and southern- most of these groupings are also apparent for larvae of 7-10 mm, but there is no well defined pattern for larger larvae. Larvae were collected September-May. depending on location, but in most subareas the peak abundance occurred in November (Table 3). Larvae were captured as early as September in subareas II and III. The latest captures were during May in subareas IV and V. Lar- vae occurred over 5-7 months within the study area (Table 3). The shortest duration of occurrence was in subarea I (October-February), and the longest was in subareas IV and V (October-May). For most subareas, larval catches were low or zero in January followed by a slight increase in February. The most southerly subarea (V) showed a smaller peak in April. Individual years (see Morse et al. 1987) revealed similar patterns in timing and location of eggs and larvae as found in the composite data (Table 3). As expected from the timing of peak egg abundance, the smallest larvae (<6 mm) were most abundant October-December (Table 4). Small larvae (4-8 mm) occurred in April, although we collected no eggs February-April which could have accounted for these larvae. The largest larvae (>11 mm) were abundant November-May with peaks in November-December and March-May. Inshore occurrence of larvae and juveniles Transforming F. (h')it(itNs have been collected from most of the major inlets adjacent to subarea III along the New Jersey coast. Reexamination of data collected in New Jersey estuaries (Table 1) and our own collect- ing efforts indicate transfoi'ming larvae have occurred Fishery Bulletin 88(1), 1990 Abundance length. (mean no./lOO m- of sea surface) of Paral irhthy. Table 4 ■ dentatKS larv le from MARMAP survey s (luring 1977-85, by month and Length (mm SL) Sep Oct Nov Dec Jan Feb Mar Apr May 2 73.6 53.0 7.2 3 9.5 1048.0 815.6 141.1 4 563.1 853.0 291.7 2.2 6.5 5 238.1 ,509.7 308.6 4.5 12.1 12.5 6.5 6 7 4.6 106.8 75.7 461.5 318.2 213.9 194.6 8.9 4.5 13.7 29.0 4.6 28.9 8 4.6 46.8 117.8 166.0 10.0 94.1 8.9 4.5 8.8 9 1.9 11.8 80.6 4.5 23.0 5.0 12.6 10 22.1 56.5 5.6 30.6 15.7 4.0 11 1.9 3.5 12.9 4.4 3.8 14.0 16.9 12.1 12 3.5 9.1 3.4 5.4 13 5.3 3.1 1.7 11.8 2.0 >13 14.7 1.5 in Sandy Hook Bay, Shark River Inlet, Manasquan Inlet, Barnegat Bay, Little Egg Inlet, Absecon Inlet, Corson's Inlet, and the Maurice River (Fig. 1). These larvae were in the range 8-21 mm (n = 378), but most were 11-14 mm. In Manasquan Inlet larval abundance was quite high, with counts per half-hour set reaching 31 larvae on 24 October 1967, 44 on 28 November 1963, 34 on 2 December 1966, and 64 on 5 December 1963 (Festa 1974). Movement of transforming individuals through inlets in New Jersey occurs primarily October- December (Fig. 4). However, a few larvae have been collected as late as February (15-17 mm Th, n = 2) in Little Sheeps- head Creek inside Little Egg Inlet (this study); in March and May (15.0-19.0 mm TL, n = 4) in the Maurice River off Delaware Bay; in March in Manas- quan Inlet {n = 1) and Corson Inlet (n = 1); and Feb- ruary, March, and April in Absecon Inlet (10.6-13.8 mm TL, n = 7) (Table 1). The extensive collections in the Manasquan and Shark River inlets (Festa 1974) produced relatively large numbers that are depicted as the first peak of 10-20 mm individuals in October in Figure 4. Other collections suggest that some of these individuals may attain a size of 30-50 mm by Decem- ber, but these occurrences have been sporadic. During 1975-80, P. dentatus of 30-50 mm only occurred in winter 1975. The available data are sparse for YOY January-April (Fig. 4). This same cohort was repre- sented by a few individuals in May (30-50 mm and perhaps up to 90 mm), more individuals in June (30-140 mm) and accounts for the dominant peak July-October when they reach a size of 160-320 mm. As a result, in October there were two well-defined length- frequency modes: a mode around 10 mm that repre- sents transforming individuals, and a larger mode (160-320 mm) that represents individuals 1 year older. In a subsample (n = 111) of P. dentatus collected dur- ing the period September- November (larger mode in Figure 4), 97% of the scales lacked an annulus; thus, by convention (Smith et al. 1981), they are considered to be YOY, although they are 1 year older than the transforming .specimens (smaller mode in Figure 4). Fall distribution of YOY Young-of-the-y ear (160-320 mm) apparently move out of estuaries in fall with adults. By November, YOY are less abundant inshore (P"'ig. 4) but are well represented in NMFS continental shelf trawl surveys (Fig. 5). This same size class appeared as a definite peak ('^ 200-270 mm) in September 1985 trawl collections off the coast of New Jersey (Halgren and Scarlett 1985). YOY in- dividuals are distributed inshore from Long Island to south of Cape Hatteras at this time of the year (Fig. 6) with larger individuals apparently more abundant north of Delaware Bay (Fig. 7). Discussion Timing and location of spawning and development The geographical patterns of reproduction observed during 1979-85, as measured by the distribution and abundance of eggs (Fig. 2), were similar to those reported for 1965-66 (Smith 1973). An exception is the report (possibly a result of misidentifications) of spawn- ing in July in Narragansett Bay (Herman 1963). Our sampling in 1980-86 extended farther north and east Able et at : Parslichthys dentatus early life history Figure 4 Cumulative iiKJiithly leiij:jth frequencies of juvenile Pnra- Urhthya ileiitatm from northern New Jersey (Barnegat Bay anrl north) estuaries over several years. See Table 1 for liata sources. 60 NMFS FALL SURVEY. 1975-1985 50 n=466 ■■ 40- 30 - 20- _ m 5 10- 0- 60 - L NEW JERSEY ESTUARIES. SEPT-NOV 50- n=108 40- 30 20- 10- 120 140 160 180 200 220 24U 2b0 2B0 300 320 340 1 TOTAL LENGTH (MM) Figure 5 Length-frequency distribution of juvenile (0 age) Para- tichlhys dentatus from the Mid-Atlantic Bight continen- tal shelf (NMFS September-November bottom-trawl surveys) and New Jersey estuaries. Fishery Bulletin 88(1), 1990 Figure 6 Geographical distribution of young-of-the-year (Oage) Paralichthys dentatus (NMFS Septem- ber-November bottom-trawl surveys), 1982-86, expressed as number per half-hour tow. than Smith's (1973), thus we can confirm that exten- sive spawning during this period occurred in subarea II, and extended to subarea I (Fig. 2, Table 2). In fact, subarea II (Fig. 1) provided the largest collections of eggs (Table 2) and larvae (Table 3). Although Smith (1973) reported the center of abundance to be off New Jersey and New York in 1965-1966, we found exten- sive reproduction during 1980-1986 occurred from New York to Massachusetts. Spawning was most pro- nounced in fall, with the earliest in the northern and the latest in the southern subareas (Table 2). The spawning period observed in this study is consistent with that of other studies (Hildebrand and Schroeder 1928, Smith 1973, Smith et al. 1975, Morse 1981) in the Mid-Atlantic Bight. 160 210 260 TOTAL LENGTH (MM) Figure 7 Length-frequency distribution of y(jung-of-the-year PdniUchthys dentatus in NMFS fall bottom-trawl surveys north and south of Delaware Bay. The above locations and timing of spawning, deduced from egg collections, were corroborated by collections of larvae. Small larvae were most alnmdant in early fall in subarea II but occurred into early spring in the southern subareas (Table 3). Larvae of all sizes were found over the continental shelf throughout the study area (Fig. 3). Thus, there appeared to be no pronounced movements by larvae out of the major spawning areas as might l)e predicted from the general southerly or southwesterly flow reported for much of the Mid-Atlantic Bight (Bumpus and Lauzier 1965, Norcross and Stanley 1967). The only apparent exception is subarea I where egg collec- tions (Fig. 2, Table 2) indicated spawning occurred, yet few larvae were collected (Fig. 3). Stronger currents in the prevailing flow off the southern and western por- tions of Georges Bank into the northern Mid-Atlantic Bight (Colton and Temple 1961, Butman and Beards- ley 1987) might explain this exception. Estuarine recruitment and growth The movement of larval, transforming, and possibly juvenile P. dentatus into estuaries occurs over an ex- tended time period. The collection of P. dentatus in New Jersey inlets October- April is consistent with the prolonged spawning period in the adjacent waters of subareas II and III (Table 3). Also, the peak in spawn- ing in these areas in October and November may be reflected in the relatively abundant catches of small in- dividuals (<20 mm TL) in October-December (Fig. 4). Most authors have assumed that transforming P. dentatus move into estuarine nursery areas (see Rogers and Van Den Avyle 1983 for review). From Able et al • Paralichthys denatus early life history the available evidence, we conclude that at least a portion of the P. dentatus do this in New Jersey estuaries. It is clear that at sizes larger than approx- imately 11 mm SL, they become scarce in continental shelf plankton collections (Fig. 3) and that they can be collected in New Jersey inlets, usually at sizes of 11-14 mm TL. In addition, the larger juveniles have been collected, at least from power plant screens, with some regularity (Fig. 4). In some years larger YOY (240-300 mm) are clearly represented in recreational catches in New Jersey estuaries (Festa 1979). Despite this pat- tern, it is not always easy to collect small juvenile P. dentatus in New Jersey estuaries. As we pointed out earlier, the collections of small juveniles (30-50 mm, Fig. 4) occurred in a single year (1975) but not in subse- quent years (1976-80). In addition, small juveniles are rarely encountered in the winter (January- April), following recruitment to the estuary (Fig. 4). Limited data might lead to the assumption that YOY P. den- tatus could not survive winter temperatures in New Jersey inshore waters. Johns et al. (1981) indicate that larval P. dentatus fail to develop past yolksac absorp- tion when reared at temperatures below 5°C. Estu- arine temperatures in New Jersey regularly drop to several degrees below this level in winter (Rutgers Univ., Mar. Field Stn., Tuckerton, NJ 08067, unpub- lished data). Thus, if transforming individuals suffer the same mortality as the yolksac larvae, individuals that have moved into the estuaries may not survive winter. Also, several studies indicate that P. dentatus in Chesapeake Bay may succumb to infections of the hemoflagellate Trypanoplasma hullocki at low tem- peratures (Burreson and Zwerner 1982, 1984; Sypek and Burreson 1983). Effective immune response to the parasite was not noted in natural infections below 10°C (Sypek and Burreson 1983). This parasite occurs in New Jersey waters (G. Burreson, Va. Inst. Mar. Sci., Gloucester Pt., VA 23062, pers. commun. July 1987), but its effect on New Jersey populations of P. dentatus is unknown. Some authors have assumed that because of the perceived paucity of YOY in New Jersey and other estuaries in the northeastern United States, the impor- tant nurseries for P. dentatus occur in Virginia and North Carolina (see Rogers and Van Den Avyle 1983 for review). We find the evidence to support this con- clusion ambiguous and in some cases contradictory. First, the evidence for abundant juvenile P. dentatus is based primarily on studies in South Atlantic Bight estuaries that no doubt support P. dentatus nurseries (Tagatz and Dudley 1961, Miller and Jorgenson 1969, Burns 1974, Powell 1974, Cain and Dean 1976, Bozeman and Dean 1980). However, some of these data sources (Weinstein 1979, Weinstein et al. 1980) also include southern flounder P. lethostigma and gulf flounder P. albigutta. but the patterns observed are considered characteristic of P. dentatus (see Rogers and Van Den Avyle 1983). Second, there is evidence that estuaries in the northern Mid-Atlantic Bight do provide nurseries for juvenile P. dentatus (Grosslein and Azarovitz 1982). This has been substantiated for Connecticut (Pearcy and Richards 1962), Long Island (Poole 1961), New Jersey (Table 1 and this study), Delaware Bay (this study), and coastal Delaware (Pacheco and Grant 1973). That nurseries occur in these areas would certainly be consistent with the ex- tensive spawning and larval development that occur in the northern Mid-Atlantic Bight (Smith 1973 and this study), especially considering the lack of evidence for larval transport or advection from this area. Third, the habitats utilized by small juvenile flounder may be difficult to sample. There is evidence that juvenile P. dentatus use eelgi-ass beds in North Carolina (Adams 1976), Chesapeake Bay (Orth and Heck 1980, Wein- stein and Brooks 1983) and New Jersey (senior author's pers. observ.). Sampling flatfishes in these structural- ly complex habitats with conventional gears (trawls, seines) is inefficient and difficult. To our knowledge, no one has presented data to con- firm or deny the possibility that P. dentatus uses the continental shelf as a nursery area. Certainly spawn- ing in the Mid-Atlantic Bight and on Georges Bank occurs great distances from estuaries (Fig. 2). Perhaps many of the larvae undergo transformation, descend to the bottom over the continental shelf, and then move into estuarine areas at a variety of sizes, beginning with early transforming individuals. This could help explain the relative scarcity of small juveniles and the large numbers of YOY (150-320 mm) that appear during summer (Fig. 4). These YOY occur in the recreational fishery in New Jersey in some years (Festa 1979). A continental shelf nursery area is more tenable in the northern Mid-Atlantic Bight, given the broad continen- tal shelf (~ 150 km) relative to the narrow shelf (~ 50 km) off North Carolina. We know of no adequate continental-shelf sampling progi-am, using appropriate collecting gear, to support or refute this possibility, but suggest it should be vigorously tested. The growth of YOY P. dentatus in New Jersey estuaries appears to be very fast, with individuals reaching 160-320 mm 1 year after spawning (October, Fig. 4). An almost identical growth rate has been found in Long Island estuaries (Poole 1961), where length fre- quencies reported for July, August, and September are similar to those from New Jersey (Fig. 4). Additional- ly, the modal sizes of males (25.1 cm) and females (27.1 cm) of YOY reported from Long Island (Poole 1961) and those for both sexes from New Jersey (Fig. 4) are similar to the modes of YOY captured in fall over the continental shelf (Fig. 5). Assuming little or no 10 Fishery Bulletin 88(1), 1990 growth during winter, these lengths would approx- imate those of individuals that are forming the first annulus during the winter. Similarly, a laboratory study of P. deiitatus collected from North Carolina found a mean size of 232.8 mm TL at the end of 1 year under constant conditions, although most growth occurred June-November (Klein-MacPhee 1979). Earlier attempts to review the age and growth of P. dentatus along the east coast of the United States (Smith et al. 1981) were not completely successful because of variation in size at first annulus formation. Our data support the estimates of size at first annulus formation proposed by Poole (1961), an interpretation that was not accepted by those at the age and growth workshop (Smith et al. 1981). The latter favored a slower growth estimate based on data from North Carolina estuaries (Powell 1974, 1982). An alternate interpretation of the available P. den- tatus growth data, and one that would resolve the above discrepancy, is that YOY P. dentatus from New Jersey grow at a faster rate than do those from North Carolina. This is consistent with the workshop inter- pretation that overall fish growth rate tended to increase from south to north (Smith et al. 1981) and with the presence of larger YOY individuals in more northern fall collections (Fig. 7). The possibility of geographical differences in gr-owth patterns is also con- sistent with the view that northern Mid-Atlantic Bight populations of P. dentatus are distinct from those south of Cape Hatteras (Wilk et al. 1980, Delaney 1986). Acknowledgments Preparation of this manuscript was supported in part by grants from the New Jersey Fisheries Development Commission, New Jersey Fisheries and Aquaculture Technology Exchange Center, NOAA Sea Grant, New Jersey Marine Science Consortium, and the Center for Coastal and Environmental Studies, Rutgers Univer- sity. P. Berrien provided information on egg distribu- tion, G. Burreson discussed parasite/temperature effects on mortality, and S. Szedlmayer, J. Musick, and J. Defosse pi-ovided comments on an earlier draft. We are grateful to all of the above. Citations Adams, S.M. 1976 The ecology of eelgras.s Z»:iti-ni tiKiritin (L.) fish com- munitics. I. Structural analysis. J. Exp. Mar. Biol. Ecol. 22:269-291. Allen. D.M., .I.P. Clymer, and S.S. Herman 1978 Fishes of the Hereford Inlet estuary, .southern New- Jersey. Lehigh Univ. and Wetlands Inst.. Bethlehem, PA 1801.5, l.'?8 p. Azarovitz, T.R. 1981 A brief historical review of the Woods Hole laboratory trawl survey time series. In Doubleday, W.G., and D. Rivard (eds.). Bottom trawl surveys, p. 62-67. Can. Spec. Pulil. Fish. Aquat. Sei. .58. Bozeman, E.L., and J.M. Dean 1980 The abundance of estuarine larval and juvenile fish in a South Carolina intertidal creek. Estuaries 2:89-97. Bumpus, D.F., and L.M. Lauzier 1965 Surface circulation on the continental shelf of eastern North America between Newfoundland and Florida. Am. (leogr. Soc. Ser. Atlas. Mar. Environ. Folio 7. NY. Burns, R.W. 1974 Seasonal abundance and diversity of larval fishes in a high-marsh tidal creek. M.S. thesis, Univ. South Carolina. Columbia. 54 p. Burreson, E.M., and D.E. Zwerner 1982 The role of host biology, vector biology, and temperature in the distribution of Tiypanoplasma hidlocki infections in the lower Chesapeake Bay. J. Parasitol. 68:306-313. 1984 .luvenile summer flounder, Paralichthyf; dentatus, mor- talities in the western Atlantic Ocean caused by the hemo- flagellate Trypanoplasmn buUorki, evidence from field and experimental studies. Helgol. Meeresunters. 37:343-352. Butman, B., and R.C. Beardsley 1987 Physical oceanography, pp. 88-98. In Backus. R.H. (ed.), Georges Bank. p. 88-98. MIT Press, Cambridge. Cain, R.E., and J.M. Dean 1976 Annual occurrence, abundance and diversity of fish in a South Carolina intertidal creek. Mar. Biol. (Berl.) 36:369-379. Colton. J.B., Jr., and R.F. Temple 1961 The enigma of Georges Bank spawning. Limnol. Oceanogr. 6:280-291. Delaney, G.R. 1986 Morphometric and meristic stock identification of sum- mer flounder {Paralichthys dentatus). M.S. thesis. Coll. William and Mary. Williamsburg, VA, 47 p. de Sylva, D.P., F.A. Kalber, Jr.. and C.N. Shuster, Jr. 1961 Fishes and biological conditions in the shore zone of the Delaware River estuary, with notes on other species collected in deeper water. Info. Ser. Publ. 5, Univ. Delaware Mar. Lab., Lewes, DE 19958, 164 p, Festa, P.J. 1974 A study of the distribution of young and larval summer flounder in New Jersey estuarine waters. Misc. Rep. 1 1 . New Jersey Dep. Environ. Prot., Trenton, NJ 08625-0402, 22 p. 1979 Creel census of the summer flounder {Paralirhthi/s den- latim) sportfishery in Great Bay, New Jersey. Tech Rep. 19M, New Jersey Dep. Environ. Prot., Game and Wildl.. Trenton, NJ 08625-0400, 62 p. Ginsburg, I. 1952 Flounders of the genus Pantlirhthy^ and related genera in American waters. U.S. P'ish Wildl. Serv., Fish. Bull. .52:267-351, Grosslein, M.D., and T.R. Azarovitz 1982 Fish distribution. MESA New York Bight Atlas Mono- graph 15. New York Sea Grant Inst., Albany. 182 p. Gutherz, E.J. 1967 Field guide to the flatfishes of the family Bothidae in the western North Atlantic. U.S. Fish Wildl. Serv. Circ. 263, 47 p. Halgren, B., and P.G. Scarlett 1985 Briefing document for mesh selectivity for summer flounder. Unpubl. rep.. New Jersey Dep. Environ. Prot., Div. Fish Game Wildl., Mar. Fish. Admin., Trenton, NJ 08625-0400, 16 p. Able et al : Parahchthys dentatus early life history 1 1 Hamer, P.E., and F.E. Lux 1962 Marking experiments of fluli- tatits) embryos and larvae reared at two temperatures. Mar. Ecol. Prog. Ser. 2:1-8. Johns, D.M., W.H. Howell, and G. Klein-MacPhee 1981 Yolk utilization and growth to yolk-sac absorption in sum- mer flounder {Paralichthys dentatus) larvae at constant and cyclic temperatures. Mar. Biol. (Berl.) 63:301-308, Klein-MacPhee. G. 1979 Growth, activity and metabolism studies of summer flounder, Paralichthys dentatus (L.), under laboratory condi- tions. Ph.D. diss., Univ. Rhode Island, Kingston, 99 p. Leim, A.H., and W.B. Scott 1966 Fishes of the Atlantic Coast of Canada. Fish. Res. Board Can. Bull. 15.5, 731 p. Miller, G.L., and S.C. Jorgenson 1969 Seasonal abundance and length frequency distributions of some marine fishes in coastal Georgia. U.S. Fish Wildl. Serv. Data Rep. 35, 102 p. Milstein, C.B., and D.L. Thomas 1976 Ecological studies in the bays and other waterways near Little Egg Inlet and in the ocean in the vicinity of the proposed site for the Atlantic Generating Station. New Jersey. Pro- gress Report for period Jan. -Dec. 1975. Ichthyological Assoc, Inc., Absecon, NJ 08201, 572 p, Morse, W.W. 1981 Reproduction of the summer flounder. Paralichthys den- tatus (L.). J. Fi.sh. Biol. 19:189-203. Morse, W.W., M.P. Fahay, and W.G. Smith 1987 MARMAP surveys of the continental shelf from Cape Hatteras, North Carolina to Cape Sable, Nova Scotia (1977-1984). Atlas No. 2. Annual distribution patternsof fish larvae. NOAA Tech. Memo. NMFS-F/NEC-47, Northeast Fish. Cent., Natl. Mar. Fish. Serv., NOAA, Woods Hole, MA 02543, 215 p. Murawski, W.S. 1970 Results of tagging experiments of summer flounder. Paralichthys dentatus. conducted in New Jersey waters from 1960-1967. Misc. Rep. 5M, New Jersey Dep. Environ. Prot.. Div. Fish Game Wildl., Trenton, NJ 08625-0400, 73 p. Norcross. J.J. and E.M. Stanley 1967 Circulationof shelf waters of the Chesapeake Bight. II. Inferred surface and bottom drift, June 1963 through October 1964. U.S. Dep. Commer., Environ. Sci. Serv. Adm., Prof. Pap. 3, p. 11-42, Wash., DC. Orth, R.J., and K.L. Heck, Jr. 1980 Structural components of eelgrass {Zostera marina) meadows in the lower Chesapeake Bay— fishes. Estuaries 3:278-286. Pacheco, A.L., and G.C. Grant 1973 Immature fishes associated with larval Atlantic menhaden at Indian River Inlet, Delaware, 1958-61. In Pacheco. A.L. (ed.). Proceedings of a workshop on egg, larval, and juvenile stagesof fish in Atlantic Coast estuaries, p. 78-117. Mid. Atl. Coast. Fish. Cent. Tech. Publ. 1. Sandy Hook Lab., Natl. Mar. Fish. Serv., NOAA, Highlands, NJ 07732. Pearcy, W.G., and S.W. Richards 1962 Distribution and ecology of fishes of the Mystic River estuary. Connecticut. Ecology 43:245-259. Poole, J.C. 1961 Age and growth of the fluke in Great South Bay and their significance to the sport fishery. N.Y. Fish Game J. 8:1-18. 1966 A review of research concerning summer flounder and needs for further study. N.Y. Fish Game J. 13:226-231. Powell, A.B. 1974 Biology of the summer flounder, Paralichthys dentatus, in Pamlico Sound and adjacent waters, with comments on P. lethostigma and P. albigutta. M.S. thesis, Univ. North Carolina, Chapel Hill, 145 p. 1982 Annulus formation on otoliths and growth of young sum- mer flounder from Pamlico Sound, North Carolina. Trans. Am. Fish. Soc. 111:688-693. Powell. A.B.. and F.J. Schwartz 1977 Distribution of paralichthid flounders (Bothidae: Para- lichthys) in North Carolina estuaries. Chesapeake .Sci. 18:334-.3.39. Rogers, S.G., and M.J. Van Den Avyle 1983 Species profiles: Life histories and environmental re- quirements of coastal fishes and invertebrates (South Atlan- tic)-summer flounder. U.S. Fish Wildl. Serv. FWS/OBS-82/ 11.15. U.S. Army Corps of Engineers. TR EL-82-4, 14 p. Scarlett, P.G. 1982 Annotated bibliography and subject index on the sum- mer flounder, Paralichthys dentatus. NOAA Tech. Rep. NMFS SSRF-755, Natl. Oceanic Atmos. Adm., Natl. Mar. Fish. Serv., 12 p. 1989 Results of finfish sampling in the Manasquan River, 1984-1986. New Jersey Dep. Environ. Prot., Div, Fish Game Wildl.. Draft Rep., proj. F-15-R-30, Trenton, NJ 08625-0400, 97 p. Sherman. K. 1980 MARMAP, a fisheries ecosystem study in the Northwest Atlantic: Fluctuations in ichthyoplankton-zooplankton com- ponents and their potential for impact on the system. In Diemer, F.P., F.J. Vernberg, and D.R. Mirkes (eds.), Advanced concepts in ocean mea.surements for marine biology, p. 9-37. Belle W. Baruch In.st. Mar. Biol. Coastal Res., Univ. South Carolina Press. 1986 Measurement strategies for monitoring and forecasting variability in large marine ecosystems. In Sherman, K., and L.M. Alexander (eds.). Variability and management of large marine ecosystems, p. 203-231. AAAS Selected Symp. 99, Am. Assoc. Adv. Sci., Wash., DC 20005. Sibunka. J.D., and M.J. Silverman 1984 MAR.MAP surveys of the continental shelf from Cape Hatteras, North Carolina to Cape Sable, Nova Scotia (1977- 1983). Atlas No. 1. Summary of Operations. NOAA Tech, Memo, NMFS-F/NEC-33, Northeast Fish. Cent., Natl. Mar, Fish. Serv.. NOAA. Woods Hole, MA 02543, 306 p. Smigielski, A.S. 1975 Hormone induced spawnings of the summer flounder and rearing of the larvae in the laboratory. Prog. Fish-Cult. 37:3-8. Fishery Bulletin 1990 Smith, R.W., and F.C. Daiber 1977 Biology of the summer flounder, Paralichthy>< denial us. in Delaware Bay. Fish. Bulh, U.S. 7.=5:823-830. Smith. R.W., L.M. Dery. P.G. Scarlett, and A. Jearld, Jr. 1981 Proceedings of the summer noun\) usually result in strongly unimodal size-fre- quency distributions. Stable bimodal distributions require shifts from mor- tality-dominated to growth-dominated conditions via age-related changes in Z. K, or both. Non-equilibrium con- ditions or events such as pulses in recruitment or mortality can also modify size-frequency distributions, but these effects are usually transi- ent. These results indicate that infer- ences concerning the demographic dynamics of a population may be derived sinijily by obsemng the shape of its size-frequency distribution. Inferring Demographic Processes from Size-Frequency Distributions: Simple Models Indicate Specific Patterns of Growth and Mortality James P. Barry Mia J. Tegner A-OOI Scnpps Institution of Oceanography La Jolla, California 92093 Manuscript accepted 24 August 1989. Fishery Bulletin, U.S. 88:13-19. Population ecologists and wildlife managers are often interested in identifying spatial and temporal pat- terns of growth and mortality in order to understand the dynamics of populations. Because of logistic or other constraints in their abilities to directly quantify mortality and growth, however, ecologists often rely on the analysis of size distribu- tions to infer these parameters (Cassie 1954; Beverton and Holt 1956; Ricker 1958, 1975; Ebert 1973; Van Sickle 1977b; Pauly and Morgan 1987). A population's size-frecjuency distribution (hereafter referred to as size distribution or size structure) results from its recent history of re- cruitment and mortality, integrated with the growth rates of individuals. Temporal or spatial changes in the size structure of a population must therefore reflect changes in one or more of these parameters. For exam- ple, red sea urchins Strongylocentro- tus Jranciscfuius are long-lived and have size distributions that are rela- tively stable through time, but vary in space due to geographic changes in recruitment and mortality related to the distribution of predators and dispersal of larvae (Tegner and Barry Unpubl.). In some locations urchins have persistently bimodal size struc- tures, while in others, the population is unimodal or amodal. Techniques for size-frequency analysis usually combine models of the growth rates of individuals with mortality rates to describe observed size-frequency distributions. Because growth and mortality both affect the shape of the distribution, knowledge of one parameter may allow deduc- tion of the other from the shape of the size-frequency distribution. Since growth, mortality, and recruitment can vary considerably, as well as in- dependently of one another, analyses of size distributions typically utilize several simplifying assumptions. In this paper, we show that al- though simplified models commonly employed for growth and mortality result in a limited range of size struc- tures, they can nonetheless be in- dicative of the demographic patterns of populations. Thus, examination of size-frequency distributions may allow reasonable inferences concern- ing the dynamics of a population. In particular, bimodal size distributions, that are "impossible" under typical model assumptions but are typical of red urchins in the Southern Califor- nia Bight, must arise from particular patterns of age or size-specific changes in demographic parameters. In order to generate these size structures, assumptions concerning constant mortality and growth coefficients are not tenable. The most basic assumption of most models is that the population is stable and has a stationary age structure. Thus, recruitment is taken to be invar- iant from year to year, or continuous, and mortality rates are presumed 13 Fishery Bulletin 88(1). 1990 constant over time. In addition to these assumptions, the population is usually required to conform to strict- ly deterministic equations that describe growth and abundance. Often, mortality rate is also assumed to be independent of age (Green 1970, Ebert 1973), resulting in a type II survivorship curve (Deevey 1947). Growth rates of individuals are frequently assumed to fit a Brody-Bertalanffy function (von Bertalanffy 1957, Ricker 1975) in which size increases at a rate propor- tional to the distance from the maximum size. For models in which one or another of these assumptions is relaxed, see Van Sickle (1977a, b), DeAngelis and Mattice (1979), DeAngelis and Coutant (1982), and several papers in Pauly and Morgan (1987). The applicability of these simplifying assumptions varies considerably among species and populations. For some species (e.g., certain long-lived fishes of arctic lakes), age structures are stable and size distributions are stationary, with fairly constant recruitment from year to year (Johnson 1976). For many species, how- ever, the assumption of a stable age structure is un- realistic, due to considerable interannual variability in recruitment to the zero age class (Johnson et al. 1986, Pearse and Mines 1987, Raymond and Scheibling 1987, Barry 1989). For red urchins, although recruitment varies between years, the shape of its size distribution is characteristically constant from year to year (Tegner and Barry Unpubl.). The assumption of an age-invari- ant mortality rate is also probably unwarranted for many, if not, most species. A more common pattern is high juvenile mortality, followed by high adult sur- vival {ty[)e III survivorship). Models incorporating these simplifying, but perhaps unrealistic, assumptions (dis- cussed above) can, nevertheless, be of value in iden- tifying patterns of individual growth as well as in estimating recruitment and mortality for the popula- tion. In many cases these properties may otherwise be unobtainable. In this paper we are concerned with population dynamics that result in a bimodal distribution of sizes. Bimodal size distributions have been reported for several species, and are of considerable interest (John- son 1976; Tegner and Dayton 1977, 1981; Shelton et al. 1979; Timmons et al. 1980; Stein and Pearcy 1982; Wilson 1983; Pollard 1985; Page 1986; Tegner and Barry Unpubl.), especially for populations that are apparently stable. Intuitively, bimodality may develop and persist under equilibrium conditions by a combina- tion of rapid growth of individuals to adult size and high survival rates. Thus, a mode of juveniles may be distinct from an adult mode comprised of several age classes that overlap in size, with relatively few intermediate- sized, rapidly .growing individuals. DeAngelis and Mat- tice (1979) and Power (1978) suggested that bimodality may arise from this sort of "pileup" of individuals at larger size classes due to a decrease in growth rate at adult size, even with constant mortality. Mortality decreases the number of older individuals, but a large number are left clustered near the upper size limit. Here we show that bimodal size stnicture must develop from particular patterns of age-specific growth, mor- tality, or both, that are not possible with commonly employed models. Specifically, bimodality can develop only from an increase in survivorship with age or an increase in the growth coefficient with age, or both. Even though simple models are limited in their range of size distributions, we can use these models to iden- tify deviations from them that are necessary or suffi- cient to produce particular size distributions, such as bimodal or strongly unimodal distributions. Derivation of a simple size-frequency model The change in abundance of a cohort can be repre- sented as, dN dt = -ZN (1) where A'' is the number of individuals alive in the cohort at time /, and Z is the instantaneous mortality rate for the population. Assuming that Z is constant over time and independent of age and size, this equation can be integrated to obtain a simple decreasing exponential function for the number of individuals versus time. A^, A„c-2('-'"' (2) where f,, is the time of recruitment or birth and A,, is the abundance of the population at time ^|. The equa- tion can be simplified slightly l)y defining at)undance in terms of age rather than time; age (t) equals t-t-Q, or time since recruitment. Hence, equation (2) becomes A, = /V„ f -Zi (3) If we now assume that the population is stable, with constant recruitment, this function describes both the time series of abundance for a single cohort and the stable age structure of the population. Brody-Bertalanffy growth is characterized by expo- nentially decaying growth in size, with no lag during early life. The general form of this deterministic equa- tion is S(l_6e-A-„-u) (4) where S, is the size of an individual at age t (i.e., at time / after /„, the time of iiirth or recruitment), S is Barry and Tegner: A predictive model for size-frequency distribution 15 VON BERTALANFFY GROWTH AGE (t) Figure 1 Brody-Bertalanffy growth for various values of the Brody growth coefficient K. Values of K are listed adjacent to each curve. the maximum size, 6 is a scaling factor to account for a size at recruitment larger than 0 (for a recruitment size of 0, 6 = 1), and K is the Brody growth coefficient (Ricker 1975) which constrains the shape of the func- tion. Higher values of A' result in a more rapid approach to asymptotic size (Fig. 1). If 6 is considered to be unity and age (t) rather than time is used, the equation simplifies to S^a-hei^^). (5) Assuming these functions adequately describe the mortality and growth schedules of the population, we can combine equations (3) and (5) to derive an expres- sion for the size-frequency distribution of the popula- tion. By definition, the number of individuals alive in an age interval tj to to in equation (3) is equal to the number in the corresponding size interval Sj to S._. , as determined by equation (5). Ns dr = N, dS, and rearranging, "-"■i- (6) Because we assume these relationships to be strictly deterministic, we can solve equation (5) for t K In 1 (7) Next, we form the first derivative of equation (7) with respect to t, yielding d^ dS, 1 KS^ 1 (8) Combining equations (6), (3), (7), and (8) yields an ex- pression for number as a function of size. Ns = Noe z s. KS^ 1- S. which simplifies to: N. = -I: (i-O (9) This size distribution function (9) describes popula- tion abundance as a function of size, rather than age, for any conditions of constant mortality (Z) and growth coefficient (A'). By evaluating its first derivative at dNIdS = 0, we can identify conditions necessary for the existence of a zero slope (modes or troughs) in the size distribution. This derivative is dN dS Nu AS„ 1 1 S..P-1 S. ■i.^(-f g-) (10) The conditions where dN/ds = 0 are: TV,, = 0 : trivial S, = S^ : trivial S = oo : trivial Z = K : growth is balanced by mortality. The only non-trivial condition where the size-fre- c}uency function has a slope of zero is when Z = K. In this case all size classes are equally abundant, since the solution to equation (9) indicates that when Z = K. Ng is constant and independent of 5^. Therefore, there are no conditions of growth and mortality that are capable of producing a bimodal distribution when using these simplifying assumptions. Thus, the hypothesis that bimodality arises from rapid growth to adult size Fishery Bulletin 88(1). 1990 CONSTANT MORTALITY A CC O SIZE Figure 2 Hypothetical size-frequency distributions for populations exhibiting constant mortahty and von Bertalanffy growth. (A) Mortality [Z] is constant for all sizes. The growth coefficient (K) is also assumed to be a constant. Due to these assumptions, the shape of size- frequency distributions are limited to those general forms presented in b-d. according to the relative magnitudes of Z and K. (B) Z is less than K over all S. (C) Z equals K. (D) Z is greater than A'. and high (constant) survivorship (Van Sicl\) conditions produces a strongly unimodal pattern, with the position of the mode determined by the size where Z = K. Obviously, variation in the ratio of ZZ/C can arise from size or age-specific changes in the value of Z, A', or t)oth. Although Bertalanffy-type growth curves assume K to be independent of age, the growth intervals of species are frequently shown to exhibit variations in K with age, with K commonly decreasing slightly with age (Ricker 1975). For example, decreases in K with size or age are evident from the increasing slope of Walford plots for sea urchins Strongylocentrotus purpuratus presented by Russell (1987). Estimates of A from Rus- sell's Figure 3 decrease from approximately 0.5 to 0.05 from small to large urchins at most locations. Assum- Barry and Tegner: A predictive model for size^frequency distribution ing the reported instantaneous mortality rates (Z-O.l) to be constant, the age-specific Z/K ratio indicates that the size-specific population dynamics switch from growth-dominated {Z/K<1) to mortality-dominated (ZIK>1), which might account for the observed strong- ly unimodal size-frequency distributions. Other species of urchins (Tegner and Dayton 1977, 1981; Himmelman 1986; Tegner and Barry Unpubl.) as well as many species of long-lived arctic fishes (DeAngelis and Mat- tice 1979) exhibit bimodal size-frequency distributions and must, under steady-state conditions, undergo a shift from mortality-dominated to growth-dominated population dynamics. In that the growth rates of individuals usually decrease with age (Ricker 1975, but see Campbell 1979, Himmelman 1986), the most likely cause of persistently bimodal size distributions under steady-state conditions is an even greater decrease in the mortality rate of large individuals. A reduction in mortality, assuming a constant or slightly decreasing growth coefficient, could allow ZIK to shift from >1 to <1: conditions necessary for bimodality. Evidence for lower mortality with larger size or age is common. Many species exhibit type III survivorship curves of decreasing mortality with age (Deevey 1947, Odum 1971, Wilson and Bossert 1971). For example, because lobsters preferentially consumed small-sized red urchins Strongylocentrotus franciscanus (Tegner and Levin 1983) and sheephead Semicossyphus pulcher repeatedly select ?,vasL\\ev S. franciscanus when offered a choice of sizes (Tegner and Dayton 1981), predation mortality apparently decreases with size ( = age) in urchins. In addition, geographic differences in the size structure of red urchins are related to the distribution of predators, with bimodal size-frequency distributions found where these predators are most abundant (Tegner and Barry Unpubl.). Intraspecific competition for resources and high adult survivorship appear to limit the growth or survival rates of juveniles or both for arctic fishes (Johnson 1976), green sea urchins (Himmelman 1986), as well as forest trees (Harper 1977), often leading to a bimodal distribution of sizes; however, bimodality in these populations may arise from stochastic, age-specific changes in growth (DeAn- gelis and Coutant 1982). Unstable or non-equilibrium conditions There are conspicuous alternative causes of bimodal size-frequency structures for populations with unstable or non-stationary age compositions. In particular, species that have seasonal recruitment and live for only two years (e.g., blue crabs; Hines et al. 1987) have per- sistent, but recruitment-controlled bimodality. For longer-lived species, relaxation of the assumptions of a stable age structure and stationary size distribution allow for transient, but perhaps persistent, variations in age and size structures due to temporal variation in recruitment and mortality. Interannual variation in recruitment, to a population normally dominated by an adult mode, can skew the age structure and occasional- ly produce a mode of juveniles, thereby resulting in bimodality. This feature will, however, deteriorate as the juveniles grow and merge into the adult mode. Similarly, a mortality-dominated population can become bimodal when a large pulse of juveniles grows to adult size, before being eventually depleted by mor- tality. In both cases bimodality is a transient feature of the size structure. How long it will persist is deter- mined by the rapidity with which individuals grow to asymptotic size as well as the range of variation in recruitment. Recruitment pulses leading to unstable size and age distributions have been reported for several species (Hjort 1914, 1926; Ebert 1983; Cowen 1985; Johnson et al. 1986; Paine 1986; Pearse and Hines 1987; Raymond and Scheibling 1987). Variation in growth and mortality within an age cohort, due to stochastic processes and genetic vari- ability, can disrupt the deterministic character of growth and survivorship processes leading to a highly variable age and size structure, even within a single cohort. Intraspecific competition for resources can in- duce bimodality within a single cohort if growth to a particular size confers a great competitive advantage, leading to even more rapid gi'owth (DeAngelis and Cou- tant 1982). For example, if recruitment by juveniles into the adult size classes is regulated by stochastic pro- cesses that provide limiting resources to a few juveniles upon the removal of adult individuals, the size struc- ture of a cohort may become bimodal. This is appar- ently ty]3ical of arctic fishes (Johnson 1976), large- mouth bass (Shelton et al. 1979; Timmons et al. 1980), gi'een sea urchins (Himmelman 1986), stalked barnacles (Page 1986), and many species of forest trees (Harper 1977). Value of simple size-frequency distribution models As shown in this analysis, even very simple models of size-frequency distributions, with perhaps unrealistic assumptions, can provide valuable information con- cerning the growth and mortality schedules of popula- tions. Although the range of potential size structures is constrained by model assumptions such that bimodal size distributions, or unimodal distributions with the mode centered away from Sd or S^ , are not possible, we can still utilize these models to identify likely 18 Fishery Bulletin 1990 patterns of age-related mortality and growth; "impossi- ble" size-frequency distributions are indicative of spe- cific violations of the assumptions. Even though model assumptions impose severe restrictions upon the shape of size distributions, the shape of an observed distribu- tion, coupled with marginal information concerning recruitment and growth for the species, can be used to infer age-related changes in growth, mortality, or both, leading to better directed research efforts on the population. This is particularly important for species such as red urchins which show strong geographical variation in their population dynamics (Tegner and Barry llnpubl.); fishery managers often must rely on easily collected size-frequency information to infer demogT'aphic parameters, rather than costly jiopulation studies at several locations. Acknowledgments We thank P. Dayton, J. Enright, C. Lennert, A. McCall, T. Ragen, W. Stockton, G. Sugihara, W. Wakefield, W. Wright, and an anonymous reviewer for advice and editorial help. This research was sponsored in part by the National Science Foundation, National Sea Grant College Program, Department of Commerce, under grant NOAA 04-8-M()M89, project number R/F-36, through the California State Resources Agency. The U.S. Government is authorized to reproduce and dis- tribute for governmental purposes. Citations Barry, J. P. 1989 Response (if a marine annelid to winter storms: an analog to t"ire-adaptation in plants? Mar. Ecol. Prog. Ser. .54:99-107. Beverton, J.H., and S.J. Holt 19.56 A review of methods for estimating mortality rates in exploited fish populations, with special reference to sources of bias in catch sampling. Rapp. P.-V. Reun. Cons. Int. Hxpior. Mer 140:67-83. Campbell. R.N. 1979 Ferox trout, Salnio truttn L., and charr, Stilvelinus (ilinnu,s (L.), in Scottish lochs. J. Fish. Biol. 14:1-29. Cassie, R.M. 1954 Some uses of probability paper in the analysis of size fre- quency distributions. Aust. .1. Mar. Freshwater Res. 5:513-522. Cowen, R.K. 1985 Large scale patterns of recruitment by the lahri o z UJ Z) o 111 COHORT 1 60 40 08. 06.82 n=267 COHORT 2 COHORT 1 8° ^ 01.18.83 V \ n=117 I1.I1III1IIIIII1I.I,. 5 10 15 20 25 30 35 09.18.82 n=84 ...l..,l,.hllll.i,i. 5 10 IS 20 25 30 35 80 1 11.01.82 n=353 60 40 20 1 1 n J III. 80 60 40 20 5 10 15 20 25 30 35 11.30.82 n=197 .lilli.ll 5 10 15 20 25 30 35 60 40 20 0 ■■ \ \ .\ \ .ll.....l.lllMI<....„... 5 10 15 20 25 30 35 \ \ 80 \ \ 03.01.83 , n=84 60 40 20 1,^ 0 ■lllll,.U..I..^... 5 10 15 20 25 30 35 1 04.12.83 n=113 5 10 15 20 25 30 35 COHORT 3 80 60 40 20 0 \ 07.06.83 n = 339 I ill 5 10 15 20 25 30 35 COHORT 3 COHORT 2 80 \ I 80 60 40 20 0 80 60 40 20 0 80 60 40 20 O 08.24.83 n = 233 ll.lll.l.l.ll.,... 5 10 15 20 25 30 35 mill.... 5 10 IS 20 25 30 35 \ 1 \ \ \ 1 , , 12 04 83 I 1 n=258 \ I \ 1 \ 1 \ ..mini llllllll.. 5 10 15 20 25 30 35 \ \ \ 1 01.19.84 ' n=248 .lllllllllllllli 5 10 15 20 25 30 35 CARAPACE LENGTH (mm) Figure 3 Length-frequency histograms for three cohorts ofPenaeus indtcus in the St. Lucia system between 6 August 1982 and 19 January 1984. Dashed Mnes indicate positions of fitted cohort means. three cohorts overwintered in the estuary and emi- grated during December-March (summer-autumn). PenaeuH ivdicu^i began to leave the system upon attaining approximately 18 mm CL with most emigra- tion occurring between 20 and 25 mm CL (Fig. 6). Declines in the proportions of 20-25 mm CL shrimp in both the beam trawl (Fig. 3) and bait fishery samples (Fig. 7) supported the December-May (summer- autumn) emigration hypothesis. Emigrating size classes declined sharply in the baitfishery samples over April-May 1983 and January- April 1985 (Fig. 7). Emigration did not appear to be related to salinity fluctuations. The wet and dry seasons are not well defined in St. Lucia, which receives 62% of its annual precipitation (1971-81 data for the estuary mouth, Natal I'arks Board, unpubl.) during summer and autumn when most emigration appears to occur. Dur- ing our study period, salinities in the lower Narrows showed little fluctuation during the emigration of the first cohort, but did decline prior to and during the emigration of the second (Fig. 4). A more convincing relationship exists between emigration and declining water temperatures. Water temperatures decreased during the emigration of cohort 1 (Figs. 4 and 7) and may have begun to cool by the onset of emigration in the second cohort. The Benfield et al Growth and emigration of Pemeus indicus 25 40 1 i 35 O ^ 30 • .,' , , J * . . > t 25 ^ 20 V) 15-1 ^ 30 m ^ 25 < 20 m 1 15 t^ loJ { I ■'•Jl ' 1 1 25 1 ? i. 20 ^ X 1- O Z 15 lU -I r-^l lU O 10- < a. < O 0 / M|J|J|A|S|0|N|D J|F|M|A|M|J|J|A|S|0|N|D J 1982 1983 84 Figure 4 Mean carapace length oi Pcnai'us indicus cohorts in the St. Lucia system over time; vertical bars indicate 1 standard deviation. I'pper scales indicate corresponding mean temperatures and salinitie.s aver- aged across sampling stations 1-7; vertical bars indicate maxima and minima. _ 0.12 ■? 0.10 r, =0.746 TH RATE (mm c.l o o o bob .b a> a> ■ ■ O 0.02 O ■ ■ 1 8 20 22 24 26 MEAN TEMPERATURE (° C) 28 ^ BEAMTRAWL C3 BAITFISHERY 15n n=2923 n=9522 ElO- n\ l\ FREQUENCY 3 U\ J \j 5 10 15 20 25 30 SIZE CLASS (mm carapace length) Figure 6 Cumulative length-freiiuency histogram of Penaews indicut: for beam trawl (August 1982-November 1984) and bait fishery (September 1982-August 198.5) in the St. Lucia system. Plot truncated at 30 nmi CL because larger size classes contributed < 0.2.5% of total catch. FREQUENCY (%) o 8 § § § III ll 1 ll. S N J M M J S N J M M J S N J M 1982 1983 1984 1985 Figure 5 Relationship between growth rates oi Penaeus indicus (7.25-17.66 mm CL) and corresponding mean water temperatures in the St. Lucia system determined for consecutive sampling intervals. Figure 7 Monthly frequency oi F-'uku-uk imliciis >20 mm CL in the baitllshery catch in the St. Lucia system. Blank intervals indicate no data available for that month. emigration of the third cohort would likely have coin- cided with declining water temperatures if it had not been forced by the cyclone. Discussion Estimation of growth rates and emigi-ation size depend on a representative sample of the population at each sampling interval. Trawl samplers typically have a low catch efficiency (Zimmerman et al. 1984) and a selec- tive bias towards certain size classes (Staples 1980). Our beam trawl captured P. indicus over the range 26 Fishery Bulletin 88(1), 1990 4-33 mm CL. The percentage contribution by size class to the total beam trawl catch showed relatively little change between 8 and 18 mm CL (Fig. 6) suggesting that selection bias was not a problem within this range. On this basis, we are confident in using a restricted range of the catch (7-17.66 mm CL) for the growth estimates. However, our growth-rate estimates for sizes above this range may be influenced by emigra- tion and avoidance, and should be used with caution. Similarly, the growth rates determined for small shrimp between the time of recruitment into the estu- ary and vulnerability to our trawl should be regarded as rough approximations because of the reduced effi- ciency of our gear and absence of a precise immigra- tion time. Growth The mean growth-rate estimates obtained in our study are lower than those reported for P. indicus in other areas. This may be a consequence of the lower tem- peratures which prevail at the edge of this species' distribution. Parrack (1979) suggested that differences in the growth rates of P. aztccus populations from different latitudes could be explained by temperature differences in the bottom waters of each region. Juve- nile P. i7idicvfi grew at 0.102 mm CL per day in Singa- pore ponds (Hall 1962) while Le Reste and Marcille (1976) recorded more rapid growth of 0.125 mm CL per day in Madagascar. A review of the growth rates recorded for this species (Champion 1983) suggested an average growth of 30 mm total length per month which equates to 0.176 mm CL per day when the con- version factor of Prahhakara Rao (1967) is applied and a 30-day month assumed. Application of the von Bertalanffy growth function may be justified for some penaeid shrimp (Garcia and Le Reste 1981); however, its use is questionable in areas where growth is reduced at times by low temper- atures. The reduced growth of overwintering cohorts accounts for the relatively long residency period of the shrimp in St. Lucia. Between August and October 1983, the water temperatures in the Narrows were 18-23°C. The lowest growth rates from our study oc- curred over 18-22° C, which may provide an estimate of the low-temperature growth threshold for P. indicuf^ in St. Lucia. Comparable data for P. indicus were unavailable; however, our estimate of the low-temper- ature gi'owth threshold is similar to estimates reporteci for other penaeid species. Latapie et al. (1972) found negligible growth in P. aetiferus below 20°C, and Phares (1980) used 17°C as the zero growth tem- perature in a model for the same species. Below 22°C, the growth of juvenile P. vannamei was reduced in Mexico (Edwards 1977). In contrast, Zein-Eldin and Aldrich (1965) reported slow growth of postlarval P. aztecuH at 18°C and a lower growth threshold of 11°C. In several instances, the interval growth rates were inversely related to temperature. This may be a con- sequence of size-specific changes in growth rates over- riding the effects of temperature. Hall (1962) reported linear growth up to 25 mm CL for P. indicus; however, there is no evidence that more complex growth models were fitted. In juvenile P. vannamei (Menz and Blake 1980) and juvenile P. sefifenis (Phares 1980) growth was negatively correlated with size. While growth rates probably decline with increasing size in P. indicus, the observed fluctuations may also be due to changes in the thermal history of each cohort not manifested in the mean interval water temperature. Emigration The life cycle of P. indicus in St. Lucia appears to follow the generalized pattern outlined for penaeids by Garcia (1985). There appear to be two cohorts each year, the first entering during early summer and emigrating from the system over autumn and early winter; the second recruits during autumn, also over- winters in the estuary, and emigi'ates during the follow- ing spring and summer. A similar pattern of recruit- ment and emigration was observed for P. indicus in Singapore by Hall (1962) and in Madagascar by Le Reste (1978). This pattern has also been noted for P. /*/f'r(/«(>riSi's in Australia (Dredge 1985, Rothlisberg et al. 1985). In an earlier study of the St. Lucia pen- aeids, Joubert and Davies (1966) reported emigration through summer and autumn but did not indicate that this protracted emigration actually represented the overlapping departures of two cohorts. Penaeus indicus appears to emigrate from St. Lucia and other systems over similar size ranges. In Singa- pore, Hall (1962) attributed a decline in the frequency of 20-25 mm CL shrimp to emigration, and in Mada- gascar P. indicus began to emigrate at approximately 20 mm CL. Prabhakara Rao (1967) suggested that P. indicus left Chilka Lake, India, between 18.3 and 22.9 mm CL. Jayakody and Costa (1988) repotted that P. indicus may emigrate from Sri Lankan estuaries in response to osmotic stress imposed by monsoon rains. The onset of emigration by P. mer-guiensis has been linked to elevated rainfall and declines in salinity (Rothlisberg et al. 1985, Staples and Vance 1986). Joubert and Davies (1966) suggested a link between migration and salinity in St. Lucia but had no experimental data to support this connection. Emigration appeared to be related to late-summer and autumn water cooling; however, the apparent correlation does not inifjly Benfield et al Growth and emigration of Penseus indicus 21 causality. Other factors associated with declining water temperatures such as food availability or predation (Matylewich and Mundy 1985) may also be contributing agents. Acknowledgments The authors gratefully acknowledge the financial assis- tance of the South African National Committee for Oceanographic Research, the cooperation of the Natal Parks Game and Fish Preservation Board, D.S. Smith for construction and maintenance of sampling gear, Dr. P.D.M. MacDonald for providing access to his cohort analysis program, the Texas A&M University Com- puter Graphics Lab staff, and the comments of two anonymous reviewers. Citations Champion, H.F.B. 1983 The reproductive biology of Penaeus indicus with notes on its autecology. M.S. thesis. Rand Afrikaans Univ., Johan- nesburg. 160 p. Chen, Y.L. 1983 Emigration of penaeid shrimp from the once-through coohng lal.,5 2,6-dimethylnapthalene tl.ll 1.4-dimethylnapthalene* O.i) 1.2-dimethylnapthalene' 0.0 Total mononuclear aromatics* 97.7 Total dinuclear aromatics* 2.3 0.72 1.62 0.17 0.4.5 0.28 (1.15 0.15 0.06 0.02 0.0(1 0.00 0.00 51.0 30.3 0.2 4.5 2.8 0.1 10.5 0.0 0.3 0.0 0.1 0.2 88.9 11.1 66 89 22 27 18 .03 .87 00 12 00 .04 04 in the WSF, concentrations are reported as total mono- plus diaromatic hydrocarbons in ppm. The comi)osition of the WSF was similar in 1982 and 1983; the principal difference was a higher percentage of naphthalene in 1983 (10.5%) than in 1982 (1.2%) (Table 2). Developing eggs were examined daily; mortality and number hatched were recorded. A few eggs (2-10/ concentration • treatment) were subsampled daily for observation and measurement. Development was staged according to Yusa (1954). Blastopore diameters were measured with an occular micrometer near the time of closure in the 1-21 day treatment. The presence or absence of morphological abnormalities was recorded. If present, the size of vesicles was coded as very small, small, medium, or large. Dead eggs were also examined for abnormalities and development stage. Immediately after hatch, larvae were examined for abnormalities with a microscope; total body length and yolk lengths were measured with an occular microm- eter, except measurements were not made in the 1-21 day treatment. Yolk lengths were measured along the major body axis, and were considered proportional to yolk volume. Mortality and the number abnormal were counted for each replicate. Carls and Rice: Oil-exposed Theragra chalcogramma embryos 31 Table 3 Distribution of individual aromatic hydrocarbons in water-soluble fractions of Cook Inlet crude oil during an 18-day static exposure (1-21 day treatment) of walleye pollock eggs. Concentrations are means of three replicates measured by gas chromatography. Hydrocarbon 0 days 4 days ppm % 1.424 51.0 0.845 30.3 ().007 0.2 0.125 4.5 0.077 2.8 0.004 0.1 0.292 10.5 0.000 0.0 O.OOS 0.3 0.000 0.0 0.004 0.1 O.O06 0.2 2.482 88.9 0.310 11.1 2.792 100.0 100.0 ppm % 0,808 50.6 0.432 27.0 0.000 0.0 0.060 3.7 0.041 2.6 0.000 0.0 0.243 15.2 0.000 0.0 0.007 0.4 0.000 0.0 0.003 0.2 0.004 0.3 1.341 83.9 0.257 16.1 1.598 100.0 18 days ppm % 0.057 26.4 0.000 0.0 0.000 0.0 0.000 0.0 0.000 0.0 0.000 0.0 0.153 70.8 0.000 0.0 0.000 0.0 0.000 0.0 0.002 0.9 0.004 1.9 0.057 26.4 0.159 73.6 0.216 100.0 7.7 benzene toluene ethylbenzene m- and p-xylene o-xylene mesitylene naphthalene 2-niethylnapthalene 1 -methylnapthalene 2. 6-dimethylnapthalene 1 , 4-dimethylnapthalene 1, 2-dimethylnapthalene Total mononuclear aromatics Total dinuclear aromatics Total mono- and dinuclear aromatics Percent total hydrocarbons remaining Median lethal concentrations (LCr,,,) and median concentrations causing abnormalities (ABr,,,) were cal- culated with logit analysis (Berkson 1957) or Spearman- Karber analysis (Hamilton et al. 1977), corrected by control response (Abbott 1925). Results Test conditions Monoaromatic hydrocarbons were initially predomi- nant in the static WSF tests (89%), but they declined significantly more rapidly than diaromatic hydrocar- bons (P« 0.001) (Table 3). After 18 days monoaroma- tics comprised only 26% of the remaining hydrocarbons (Table 3). The rate of total aromatic hydrocarbon loss from solution was not linear; rates were rapid initial- ly, but slowed over time. Approximately one-half of the hydrocarbons were lost in the first 10 days. Concen- trations reported in this paper are based on initial values. Lethal effects Egg survival and hatching success were slightly reduc- ed (up to 17%) by exposure to WSF (Table 4). Reduc- tion in survival and hatching success was significant in the 1-21 day treatment, but not in the 0-21 day and 7-21 day treatments. Mortality after hatch was strong- ly dependent on concentration (Fig. 1) and began to differ significantly from controls 29 days after fertiliza- tion, or about 10 days after hatch. The LC51) was 2.2 + 0.8 ppm 10 days after hatch and stabihzed at 1.8 + 0.6 ppm 16 days after hatch. Growth Exposure to WSF before blastopore closure slowed early embryonic development. Blastopores in control embryos closed earlier than in embryos exposed to WSf'(> 1.65 ppm) in the 0-21 day treatment (Table 5). Embryos in the 1-21 day treatment tended to have larger diameters at high WSF concentrations, but diameters were highly variable (Table 6). Pore diam- eters at 2.7 ppm were significantly greater than con- trol diameters, but were significantly smaller at the lowest concentration (0.4 ppm). After blastopore clos- ure, embryonic development in the upper concentra- tions could not be distinguished from control develop- ment, and the time of hatch was not influenced by exposure to WSF (Fig. 2). Yolk and body lengths of larvae that hatched from eggs in the 0-21 day and 7-21 day treatments were significantly reduced by exposure to WSF (P<0.03 [untransformed ANOVA]). (Lengths were not measured in the 1-21 day treatment.) Yolk lengths were reduced a maximum of 9%, total body lengths were reduced a maximum of 23% (Fig. 3), and the yolk- to-length ratio increased a maximum of 20%. Mean yolk lengths at concentrations >2.8 ppm were significant- ly (P<0.05) smaller than controls. Reductions in lar- val body length in both treatments overlapped closely 32 Fishery Bulletin 88(1), 1990 Table 4 Walleye pollock egg survival and hate liing success for treatment groups exposed to water-soluble fi actions of Cook Inlet crude oil. ANOVA indicates signil icance of survival and hatching by analysis of variance; are sine transformations were used w th proportional data. NS P>005, '*P <0.01, •••p< 0.005, t significant differences from control (Dunnette test, 9 5% confidence). and ± 95% con- fidence interval. 0- -21 day treatment 1-21 day treatment 7-21 day treatment ppm n % survival % hatch ppm n % survival % hatch ppm n % survival % hatch 0.00 3 95.3 + 5.0 88.0 + 5.3 0.00 3 98.1 + 5.2 97.1 + 6.9 0.00 3 100.0 + 0.0 100.0 + 0.0 0.19 3 93.9 + 8.6 86.9 ± 15.8 0.41 3 97.9 + 6.4 96.8 + 11.1 0.26 3 99.2 + 3.5 99.2 + 3.5 0.57 3 95.3 ± 8.3 87.4 + 2.3 0.82 3 95.5 + 6.9 95.5 + 6.9 0.68 3 98.6 + 3.7 98.6 + 3.7 1.65 3 95.5 + 5.0 84.9 + 7.8 1.53 3 92.1 + 2.5 89.7 + 2.8 1.83 3 98.7 + 3.0 98.7 + 3.0 2.30 3 95.0 + 8.3 87.6 + 11.7 2.17 3 85.7 ± 11. 3T 85.7 ± 11. 2t 2.80 3 95.3 + 1.3 95.3 ± 1.3 3.59 3 91.0 ± 4.1 84.3 ± 11.4 2.75 3 81.9 ± 17.0T 80.4 ± 18.4T 3.14 3 95.2 ± 16.1 95.2 ± 16.1 ANOVA NS NS * ** ** NS NS o c 0) u 0) Q. 100 90 80 70 60 50 40 30 20 - 10 Peak hatch 14^ 12 Figure 1 Percent mortality of walleye pollock larvae after hatch when embryos were exposed to water- .soluble fractions of Cook Inlet crude oil (1-21 day treatment). Vertical bars are .standard error. Table 5 Blastopore closure in walleye pollock embryos 1 6 days after fertilization in the 0 -21 day ex- posure to water-soluble fractions of Cook In- let crude oil. ppm n n closed % closed 0.00 6 6 100.0 0.19 8 8 100.0 0.57 8 8 100.0 1.65 8 7 87.5 2.30 8 0 0.0 3.. 59 8 0 0.0 Table 6 Blastopore diameters of walleye pollock embryos 5 days after fertiliza- tion in the 1 -21 day exposure to water-soluble fractions of Cook In let crude | oil. ANOVA indicated signi fieance of blastopore diameters by analysis of variance "*P< 0.005, t significant differences from control (Dun- nette test, 95% confidence) and +95% confidence interval. ppm n Diameter (mm) 0.00 30 0.13 + 0.03 0.41 30 0.09 ± 0.02t 0.82 30 0.15 ± 0.02 1..53 30 0.15 ± 0.03 2.17 30 0.16 ± 0.02 2.75 30 0.19 + 0.02t ANOVA » • • Carls and Rice Oil-exposed Theragra chalcogramma embryos 33 12 Control 60 - 2.3 ppm r. > ro T3 o ro 50 - 40 - 3.6 ppm / ' u / ' \ ^ c o Q. 30 - 20 - 10 - 1 / 1 / ' \ ^ / ' \ * / ' \ ^ / ' \ * ' V * ' \ * 0 - / I \-^., 14 16 18 20 Days after fertilization 22 Figure 2 Timing of hatcli (% per day) of walleye pollock when eggs were exposed to water-soluhle frac- tions of Cook Inlet crude oil in the 0-21 day treatment. 1.7 5.0- 1%^*^ ■ ' _] 1..., ■ 1 A— A 0-21 day . i E 4.6- £ " +-^ ^^^■, ■--■ 7-21 day ■ g 4 2 : <0 > ■o -^ O 3 8 ^ ^""^""""■-■^.M - CO ^^^^ 3 4 - 1 2 3 Concentration (ppm) Figure 3 Length of yolks (measured along the major body axis) and total larval body lengths of walleye pollock plotted against concentrations of water-soluble fraction of Cook Inlet oil to which eggs were exposed. Vertical bars indicate 95% confidence interval. (Fig. 3). Mean body lengths at concentrations > 2.3 ppm were significantly (P<0.05) smaller than controls. Abnormalities Embryos exposed to WSF developed abnormalities. The earliest detected abnormality was the formation of membranous vesicles about the time of blastopore closure (days 6-9). In the 0-21 day treatment, embryos developed 1-5 vesicles along the ventral surface, usual- ly posterior near the blastopore (Fig. 4). These vesicles were roughly spherical, apparently formed by a single membrane. Interior fluid was indistinguishable from surrounding fluid by observation of living specimens with a microscope. Frequency of occurrence and quan- tity of vesicles were correlated with concentration (r = 0.82 and r=0.67, respectively), and vesicle size tended to increase with concentration (Table 7). The median concentration causing vesicle abnormalities was 2.4 + 0.4 ppm. Vesicle formation was not observed in the 1-21 day or 7-21 day treatments; embryos in the 7-21 day treatment were beyond blastopore closure when exposure began. Yolksac abnormalities were observed in exposed eggs shortly before hatch and other abnormalities occurred at and after hatch (Table 8). These abnormalities included body curvatures; deformations of yolk, eye, brain, jaw, intestine, and pericardial sac; absence of lower jaw; fin erosion; yolksac bloating; and light pigmentation. Yolksac bloating caused inverted floating. Correlation of abnormalities with concen- tration was strong (r = 0.9 [logit transformation]). Cor- relation between abnormalities at hatch and vesicle 34 Fishery Bulletin 88(1), 1990 ^^=^^^^^^^ r ¥w^^ I t 0^ A^ 1 M ■ "1 V /^^¥ ■ 7 V ::S . •■::/ OPyrf^ "^^V • V^ 1 1 s|^^^^ 0.2 mm Figure 4 Walleye pollock embryos exposed to water-soluble fraction (3.6 ppm in the embryo shown) of Cook Inlet crude oil developed abnormal vesicles (arrow) along the ventral surface, usually posterior near the blastopore. abnormalities was also high (r = 0.99). Median concen- trations causing abnormalities at hatch ranged from 1.6 ±0.2 to 2.1 + 0.6 ppm (Table 8). Deformed larvae did not recover from their abnor- malities in clean water. In many cases the abnormalities became more pronounced as developing structures failed to form properly: 21 days after hatch a few abnormal larvae were still alive, but most (88%) had died (control mortality was 10%). Discussion Early embryonic development, as indicated by differ- ences in blastopore diameters, was delayed by exposure to WSF, but the timing of hatch was not affected. Developmental stages (Yusa 1954) of controls and treated embryos were similar at all times except dur- ing blastopore closure. In other studies involving WSF tests of crude oils (Venezuelan, Iranian, Libyan), devel- opment of cod Gadus morhua embryos slowed, depend- ing on concentration (Kiihnhold 1974). Genetic differences or differences in WSF prepara- tions may account for the greater sensitivity observed in 1983 than in 1982. The WSF contained a higher pro- portion of diaromatic compounds in 1983 than in 1982, and diaromatics are generally more toxic than mono- aromaties (Rice et al. 1977). However, these sensitivity differences were only in magnitude of response; basic response patterns were the same. Increased metabolic demand may explain the ob- served reductions in yolk sizes of newly hatched larvae exposed to WSF during embryonic development. Yolk size was negatively correlated with concentration in both treatments, and yolk sizes tended to be smaller in embryos exposed for the longest period of time (0-21 day treatment), suggesting increased consumption of yolk and elevated metabolism. Linden (1978) also con- cluded that shorter lengths of larval Baltic herring Clupea harenguf! membras at hatching probably re- sulted from increased energy demands during exposure to the WSF of several oils. Other researchers have measured increases in metabolic rate directlv: for Table 7 Occurrence of vesicle abnormalit ies in f eveloping walleye } Hillock embryos near time of l)lastoport; closure in the 0- -21 day exposure to water-soluble fraction of Cook Inlet crude oil. ANOVA indicates significance of abnormalities by analysis of variance; arc .sine transfor- mations were u.sed with proportional data (% occurrence) or a square root transformation to control heterogeneity in the no./embryo | test; *F<0.05 ••P<0.01, * *»P<0.005, t significant diffei ences from control (Dunnette test, 95% confidence). Size codes: vs = very small, s = small, m = med um. and 1 = large Large vesicles were 8.2% of the egg diameter (~ ).ll mm). ppm % occurrence « /embryo Relative size n X SE Range X SE 0.00 14 0 (0) 0-0 0.0 (0.0) 0.19 17 0 (0) 0-0 0.0 (0.0) — 0..57 16 0 (0) 0-0 11.11 (0.0) — 1.6.5 16 18 (6) 0-1 0.2 (O.U vs-m 2.;?0 16 75 I25)t 0-2 1.1 (0.2)T .s-1 3.59 14 78 (ll)T o-r> 1.7 (0.4)T m-l ANOVA * * ** * Regression /• = 0.8 2 * r = 0.67 ** Carls and Rice: Oil-exposed Theragra chalcogramma embryos 35 Table 8 Percentage of walleye pollock larvae abnormal at the time of hatch following treatment in water-soluble fractions of Cook Inlet crude oil. ANOVA indicates significance of abnormalities by analysis of variance; arc sine transformations were used with proportional data. ***P<0.005, t significant differences from control (Dunnette test. 95% confidence), and +95% confidence. 7-21 day treatment 0- 21 da y treatment ppm n % abnormal 0.00 3 2.4 + 3.7 0.19 3 5.3 ± 9.5 0.57 3 5.4 + 6.0 1.65 3 10.6 + 13.6 2.30 3 86.4 ± 5.5t 3.59 3 99.5 ± 2.2t ANOVA ... 1-21 day treatment ppm n % abnormal 0.00 0.41 0.82 1.53 2.16 2.75 3 3 3 3 3 3 1.0 ± 4.2 1.0 ± 2.3 2.3 ± 7.0 51.2 ± 2.7t 73.6 + 15.4t 90.3 ± 11.5T ppm % abnormal 0.00 0.26 0.68 1.83 2.80 3.14 6.4 8.6 6.6 16.4 91.8 90.9 ± 5.1 ± 13.4 ± 23.2 ± 10.7 ± 12.4 1 ± 9.2t example, metabolic rates were elevated when larval Pacific herring Clupea harengus pallasi and northern anchovy Engraulis mordax were exposed to benzene (Struhsaker et al. 1974). Decreases in larval size and yolk reserves caused by sublethal exposure of developing fish eggs to WSF potentially reduces survival potential. Predator avoid- ance capability tends to increase exponentially with larval length; however, decreases in avoidance behavior may be important only when the difference in size between predator and prey is small (Hunter 1981). Lar- vae with smaller yolk reserves generally have less time to begin feeding before onset of irreversible starvation (Blaxter and Hempel 1963). The abnormalities observed in this study are the same, or similar, to those observed in other studies with a variety of species (Rosenthal and Alderdice 1976; Linden 1976; Kiihnhold 1974, 1977) and support the observation by Rosenthal and Alderdice (1976) that fish embryos tend to respond developmentally in similar ways to external stress. Some of the developmental ab- normalities (vesicle formation, body curvatures, and bloated yolks) caused by exposure of pollock eggs to WSF in this experiment were also caused by elevated temperatures in a pollock egg study by K. Krieger and L. Sonenberg (Auke Bay Lab., Alaska Fish. Sci. Cent., Natl. Mar. Fish. Serv., NOAA, P.O. Box 210155, Auke Bay, AK 99821, pers. commun. May 1988). Cold temperatures may cause similar problems in herring larvae: abnormal optic vesicles, enlarged pericardial areas, and jaw abnormalities (Ojaveer 1981). Lighter pigmentation in exposed cod Gadus morhua embryos has been observed in studies involving hydroxylated aromatic hydrocarbons (Falk-Petersen et al. 1985). Ion imbalance due to changes in membrane osmo- regulation (Ernst and Neff 1977, Linden 1978) provides a good explanation for yolksac and pericardial defor- mities, because yolksac volumes are related to water content (Rosenthal and Alderdice 1976). Aromatic hydrocarbons alter the surface properties of cell mem- branes, possibly modifying their permeability (Roubal and Collier 1975, Aronovich et al. 1975). Structural changes in cell and mitochondrial membranes of larval mummichogs Fundulus heteroclitus exposed to the WSF of Prudhoe Bay crude oil occurred even in larvae which did not show visible abnormalities (Cameron and Smith 1980). Previous cell damage, rather than the continued presence of sequestered hydrocarbons, is most likely responsible for the development of abnormalities after hatch, because newly hatched larvae were transferred to clean water and most of the hydrocarbon depurated rapidly (59-83% in the first 8 hours; Carls and Rice 1988). In another study, cod larvae also depurated naphthalene quickly, but heavier organic compounds (phenanthrene, benzo(a)pyrene, and 2, 4, 5,2', 4', 5'- hexachlorobiphenyl) were depurated progressively more slowly as molecular weights increased (Solbak- ken et al. 1984). In our study, however, compounds with molecular weights greater than the methylnaph- thalenes were present only at extremely low concen- trations, if at all. An indication that the damage hap- pened early in development, but was expressed later as malformations, was the occurrence of abnormal membranous vesicles 6-9 days after fertilization. At hatch, other malformations were observed which could have been the result of the same or similar mechanisms responsible for vesicle formation, and correlation between early and later abnormalities was very high (r=0.99). The significant mortality of larvae exposed to WSF during egg development was probably caused by biochemical changes, structural malformations, or disruption of tissue and organ development rather than a continued presence of hydrocarbons or insufficient yolk reserves. Larval mortality did not exceed 2% 36 Fishery Bulletin 88(1), 1990 until 3 days after the majority (89%) of the hatch was complete, ample time for most sequestered hydrocar- bons to be depurated (Carls and Rice 1988). Some abnormal larvae (12% at 2.7 ppm) survived on their yolksac energy reserves to the end of the experiment (21 days after hatch), suggesting that depletion of en- dogenous energy was not responsible for mortalities occurring soon after hatch. Oil accidently spilled in the marine environment rare- ly reaches concentrations (0.4-2.3 ppm) necessary to cause the effects observed in this study. However, the probability of oil spills is greater in nearshore waters, and this oil tends to concentrate in surface layers where walleye pollock eggs occur. For example, in the Bering Sea pollock eggs are spawned at depth and rise to the pelagic zone (Incze et al. 1984). Pollock eggs were most abundant in the upper 5-10 m, but abundance of older eggs (stage IV-VI) peaked about 20 m (Nishiyama et al. 1986, Serobaba 1974). Large, single-event, nearshore oil spills have released sufficient quantities of hydrocarbons to affect plank- tonic fish eggs. For example, after the Amoco Cadiz spill, water entering the Aber Wrac'h estuary con- tained more than 1 ppm hydrocarbons and 0.5 ppm throughout the estuary (Calder and Boehm 1981). After the Ar-go Merchant grounded on Nantucket Shoals, about one-half the chorions of cod Gadus rnorhua and pollock Pollachius virens eggs were contaminated with oil droplets or tar, and approximately 20-46% of the eggs were dead or moribund, compared with a 4% control mortality (laboratory-spawned cod) (Longwell 1977). In conclusion, exposure of pelagic marine walleye pollock eggs to WSF during development can cause em- bryos to develop abnormalities, and reduces size. Although they generally survive and hatch, embryos exposed to WSF produce abnormal larvae that have poor survival potential. Acknowledgments We thank Jeff Short for sample analysis by gas chroma- tography. Citations Abbott, W.S. 1925 A method of computing the effectiveness of an insecticide. .]. Econ. Entomol. 18:26.5-267. .'Vronovich, T.M., S.I. Doroshev. L.V. Spectorova, and V.M. Makhotin 1975 Egg incubation and larval rearing of navaga (Eleginuii naragn Pali.), polar cod {Eurt'ngadus saida Lepechin) and arc- tic flounder (Liopxi'tta fihicUilis Pall.) in the laboratory. A(|ua- culture 6:233-242. Berkson. J. 1957 Tables for the maximum likelihood estimate of the logistic function. Biometrics 13:28-34. Blaxter, J. U.S., and G. Hempel 1963 The inlluence of egg size on herring larvae (Clupea harengus L.). Cons. Perm. Int. E.xplor. Mer 28:211-240. Calder, J.A., and P.D. Boehm 1981 Thechemistry of Amoco Cadiz oil in the Aber Wrac'h. In Amoco Cadiz consequences d'une pollution accidentelle par les hydrocarbures, fates and effects of the oil spill, p. 149-1.58. Proceedings of the Internation Symposium (.'entre Ooeano- logique de Bret.agne, Brest, France. 1979. Public Par Le Center National Pour L'exploitation des Oceans. Cameron, J. A., and R.L. Smith 1980 llltrastructural effects of crude oil on early life stages of Pacific herring. Trans. Am. Fish. Soc. 109:224-228. Carls, M.G. 1987 Effects of dietary and water-borne oil exposure on lar- val Pacific herring (Cbipea harengus pallasi). Mar. Environ. Res. 22:2.53-270. Carls. M.G.. and S.D. Rice 1988 Sensitivities of walleye pollock. Therngra chnlcogi'mnma , eggs and larvae to hydrocarbons. Mar. Environ. Res. 26: 285-297. Ernst, v., and J.M. Neff 1977 The effects of the water-soluble fractions of no. 2 fuel oil on the early development of the estuarine fish, Funduius grnndis Baird and Girard. Environ. Pollut. 14:25-35. Falk-Petersen. I., E. Kjorsvik, S. Lonning, A.M. Nalev, and L.K. Sydnes 1985 Toxic effects of hydroxylated aromatic hydrocarbons on marine embryos. Sarsia 70:11-16. Hamilton, M.A.. R.C. Russo, and R.V. Thurston 1977 Trimmed S[)earman-Karber method for estimating me- dian lethal concentrations in toxicity bioassays. Environ. Sci. & Techol. 11:714-719. Hunter, J.R. 1981 Feeding ecology and predation of marine tlsh larvae. In Lasker, R. (ed.). Marine fish larvae. i>. 33-77. Univ. Wa.sh. Press, Seattle. Incze, L.S., M.E. Clarke, J.J. (Joering, T. Nishiyama, and .\.J. Paul 1984 Eggs and larvae of walleye pollock and relationships to the planktonic environment. In Ito, D.H. (ed.), Proceedings of the workshop on walleye pollock and its ecosystem in the eastern Bering Sea. p. 109-160. NOAA Tech. Memo. NMFS F/NWC-62, Alaska Fish. Sci. Cent.. Natl. Mar. Fish. .Serv.. NOAA, Seattle, WA 9811.5-0070. Kuhnhold, W.W. 1970 The influence of crude oils on fish fry. In Ruiro. M. (director), Marine pollution and sea life, p. 315-318, FAO Fishing News Ltd.. London. 1974 Investigations on the toxicity of seawater-extracts of three crude oils on eggs of cod {Gaduti morhua L.). Ber. Dtsch. Wiss. Komm. Meersforsch. 23:165-180. 1977 The effect of mineral oils on the development of eggs and larvae of marine species. A review and comparison of ex- perimental data in regard to possible damage at sea. Rapp, P,-v Reun, Cons. Int. Explor. Mer 171:175-183. Linden. 0. 1976 The influence of crude oil and mixtures of crude oil/disper- sants on the ontogenic development of the Baltic herring, Chipen harcngus membras L. Ambio 5(3):136-140. 1978 Biological effects of oil on early development of the Baltic herring Clupea harcngus mcmbrns. Mar. Biol. (Berl.) 45: 273-283, Carls and Rice: Oil-exposed Theragra chalcogramma embryos 37 Linden, O.. R. Laughlin, Jr.. J.R. Sharp, and J.M. Neff 1980 The combined effect of salinity, temperature and oil on the growth pattern of embryos of the killifish. Fundulus liitcnirlitiin Walbaum. Mar. Environ. Res. 3:129-144. Longwell. A.C. 1977 A genetic look at fish eggs and oil. Oceanus 20(4):46-58. Moles, A., S.D. Rice, and S. Andrews 1985 Continuous-flow devices for exposing marine organisms to the water-soluble fraction of crude oil and its components. Can. Tech. Rep. Fish. Aquat. Sci. 1368:53-61. Moore, S.F., and R.L. Dwyer 1974 Fjffects nf oil on marine organisms: A critical assessment of published data. Water Res. 8:819-827. National Marine Fisheries Service 1982 Fisheries of the United States. 1982. Current Fishery Statistics No. 8300. Natl. Mar. Fish. Serv.. NOAA. Wash., DC, 117 p. 1983 Fisheries of the United States, 1983. Current Fishery Statistics No. 8320. Natl. Mar. Fish. Serv., NOAA, Wash., DC, 121 p. Nishiyama, T., K. Ilirano, and T. Haryu 1 986 The early life history and feeding habits of larval walleye pollock Theragra chrdcoyramma (Pallas) in the southeast Bering Sea. Int. North Pac. Fish. Comm. Bull. 4.5:177-227. Ojaveer, E. 1981 Influence of temperature, salinity, and reproductive mix- ing of Baltic herring groups on its embryonal development. Cons. Int. Explor. Mer 178:409-41,5. Rice, S,D, 1985 Effects of oil on fish. In Engelhardt, F.R. (ed.). Petro- leum effects in the .f^rctic environment, p. 1.57-182. Elsevier. London. Rice, S.D., J.W. Short, and J,F. Karinen 1977 Comparative oil toxicity and comparative animal sensi- tivity. 7« Wolfe, D.A. (ed.). Fate and effects of petroleum hydrocarbons in marine organisms and ecosystems, p. 78-94. Pergamon Press, NY. Rosenthal, H.. and D.F. Alderdiee 1976 Sublethal effects of en\'ironmental stressors, natural and pollutional, on marine fish eggs and larvae. J. Fish. Res. Board Can. 33:2047-2065. Roubal, W.T., and T.K. Collier 1975 Spin-labeling techniques for studying mode of action of petroleum hydrocarbons on marine organisms. Fish. Bull.. U.S. 73:299-305. Serobaba, I.I. 1974 Spawning ecology of the walFeye pollock {Theragra chalcogramma) in the Bering Sea. .J. Ichthyol. 14(4):544-5.52. Smith, G.B. 1981 The biology of walleye pollock. In Hood. D.W.. and ,J.A. Calder (eds.). The eastern Bering Sea shelf: Oceanography and resources, vol. 1., p. 527-551. Univ. Wash. Press. Seattle. Solbakken, J.E,, S. Tilseth, and K.H. Palmork 1984 Uptake and elimination of aromatic hydrocarbons and a chlorinated biphenyl in eggs and lai-vae of cod Ga mm (B-G), and I).! mm (H-.I). 44 Fishery Bulletin 88(1), 1990 DeBrosse et al Megalopal stages of Cancer oregonensis and C productus 45 Figure 4 Megalopa uf (_'iuiccr productus Randall for Puget Sound Basin. (A) whole animal (dorsal view; ( B) cheliped; (C) 2nd pereopod; (D) coxae and bases (ventral view); (E) 3rd pereopod; (F) 4th pereopod; (G) 5th pereopod; (H) 2nd pleopod; (I) 5th pleopod; (J) telson and uropods. Scales 1.0 mm (A), 0.5 mm (B-G), and 0.1 mm (H-.J). Figure 3 Megalopa of Ciirtcer productus Randall from Puget Sound Basin. (A) antennule; (B) antenna; (C) mandible; (D) maxillule; (E) maxilla; (F) 1st maxilliped; (G) 2nd maxilliped; (H) 3rd maxilliped. Scale 0.1 mm. 46 Fishery Bulletin 88(1), 1990 Table 1 Differences between Puget Sound Basin and California populations of Cancer productus. Puget Sound Basin California (Present study) (Trask 1979) Setal formulae Antenna peduncle plus 3,2,4,0,0.4,0,4,0,3.4 to 5.4,4,0.0.3,2,3,1.3,5 tlafjellum 6, 3, 5. 0,0, .5. 0.5, 0.4, 5 Maxilla scaphognathite 62-72 62-64 First maxilliped basal endite 28-38 38-40 coxal endite 15-19 9 Third maxilliped epipodite 20-29 13 exopod 5 or 6 + 5-9 -/ + 6 endopod segment 1 25-30 26' segment 2 11-14 10* segment 3 12-17 19* segment 4 12-1(; 12* segment 5 9-12 23* Pleopodal exopod 2nd 20-22 21 ■M;\ 21 or 22 19 4th 20-23 19 r.th 18-22 19 Armature Cheliped Ischial process — 2nd pereopod Coxal, ischial processes —Not reported. * Derived from figure. Redescription: Cancer productus Randall TL 5.6-6.4 mm. CL 3.1-3.6 mm. Carapace (Fig. 4A) Considerably longer than broad, naked, lateral spines reduced to small, sometimes in- distinct knobs, pair of gastric prominences present. Rostrum well developed and with ventromedial tuber- cle, dorsal spine projecting posteriorly over 1st and 2nd abdominal somites. Eyes well developed, corneae slightly dilated. Antennule (Fig. 3A) Biramous; peduncle 3-seg- mented, proximal segment broadly expanded and with 6-14 scattered setae, penultumate segment with 3-5 distal plumose setae, ultimate segment with 1 seta. Exopod 4-segmented, basal segment naked, 2nd seg- ment with 7-12 aesthetascs, 3rd segment with 5-9 aesthetascs and 2 or 3 marginal setae, 4th segment with 3-5 aesthetascs, 1 submarginal and 1 terminal seta. Endopod indistinctly 2-segmented, proximal seg- ment naked, distal segment with 2 marginal and 3 or 4 terminal plumose setae. Antenna (Fig. 3B) Peduncle 3-segmented: tlagellum with 8 articles; setal formula varying from 4,2,4.0.0, 4,0,4,0,3,4 to 6,3,4,0,0,5,0,5,0,4,5. Mandible (Fig. 3C) Molar and incisor processes not distinguishable. Mandibular palp 2-.segmented. prox- imal segment naked, distal segment with 10-12 mar- ginal setae. Maxillule (Fig. 3D) Endopod indistinctly 2-seg- mented, proximal segment with 1 or 2 marginal setae, distal segment with 0-2 terminal setae. Coxal endite with 10-19 setae. Basal endite with 22-24 terminal setae/spines and 4 or 5 marginal setae basally. Maxilla (Fig. 3E) Endopod expanded basally and with 4 marginal setae in proximal half, 1 short terminal seta. Proximal lobe of coxal endite with 2-4 terminal and 1 or 2 subterminal plumose setae, distal lobe with 2-5 terminal and 2 or 3 subterminal plumose setae. Prox- imal lobe of basal endite with 9 or 10 terminal and 1-3 subterminal plumose setae, distal loiie with 8 or 9 ter- minal and 1 subterminal plumo.se setae. Scaphognathite DeBrosse et al : Megalopal stages of Cancer oregonensis and C productus 47 with 62-72 marginal setae and 3-5 scattered surface setae (not shown in illustration). First maxilliped (Fig. 3F) Epipod with 12-20 mar- ginal and/or submarginal setae. Exopod 2-segmented, pro.ximal segment with 3 or 4 marginal plumose setae distally, distal segment with 4-6 terminal plumose setae. Endopod with 1-3 marginal setae basally and 4-7 distally, 1 terminal seta; coxal endite with 15-19 plumose setae; basal endite with 28-38 plumose setae. Second maxilliped (Fig. 3G) Epipod with 6-10 mar- ginal setae; protopod with 1-7 scattered setae. Exopod 2-segmented, proximal segment with 1 or 2 marginal, short spine-like setae, distal segment with 4 or 5 ter- minal plumose setae. Endopod 4-segmented, merus with 2-6 setae, carpus with 2 or 3 setae, propodus with 5-9 setae, dactyl with 3-5 submarginal setae and 4-6 terminal serrate setae. Third maxilliped (Fig. 3H) Epipod with 20-29 mar- ginal setae. Protopod with 21-32 scattered plumose setae. Exopod 2-segmented, proximal segment with 5 or 6 marginal setae, distal segment with 5-9 terminal plumose setae. Endopod 5-segmented, ischium with 25-30 setae, merus with 11-14 setae, carpus with 12-17 setae, propodus with 12-16 setae, dactyl with 9-12 setae and 0-2 distinctly toothed bristles. Pereopods (Figs. 4B, C, E-G) Segments of all pereo- pods with scattered short setae. Cheliped with ischium armed with acute spine on ventrodistal margin, cutting edge of fixed finger and dactyl each with 2-4 teeth. Second pereopod with coxa and ischium each armed with acute spine on distoventral margin (4D), ischial spine smaller than coxal spine, dactyl with 5-7 spines on ventral margin. Third pereopod with ischium fre- quently armed with minute process on ventrodistal margin, dactyl with 6 or 7 spines on ventral margin. Fourth and 5th pereopods with coxae and ischia un- armed, dactyls with 5-7 and 2-4 spines on ventral margin respectively; dactyl of fifth also with 3 terminal setae. Abdomen and pleopods (Figs. 3 A, 4H,I) Abdomen six-segmented. Second pleopod with 3-5 and 3rd through 5th each with 3 or 4 hooks on appendix inter- nae. Exopods with 20-22, 21 or 22, 20-23, 18-22 plumose setae respectively. Uropods 2-segmented, peduncle naked or with 1 basal marginal plumose seta. Exopod with 12-13 plumose setae, endopod absent. Telson (Fig. 4J) Dorsal surface with 3 or 4 pairs of short setae in midline distally, terminal margin usual- ly slightly rounded, without marginal setae. Discussion Of concern to field biologists is the fact that larval studies primarily report characters found in specimens reared under laboratory conditions. Even though en- vironmental conditions have been varied to ascertain their influence on larval development, there often re- mains a question as to the similarities of characters reported for these laboratory-reared organisms and those that would be found in naturally occurring popu- lations of the same species. Although the megalopal characters of Cancer oregonensis and C. productus reported in this study were based upon specimens col- lected from naturally occurring populations, counts for some characters of C. oregonensis have been compared with counts derived from laboratory-reared animals. No appreciable differences were found between the two populations; however, in carapace length, the labor- atory-reared individuals fell in the lower half of the length-frequency curve determined for the natural population. As previously indicated, a number of differences be- tween the megalopae of C. productus examined in this study and those described by Trask (1970) have been observed. Most obvious are the differences in setal for- mulae for the coxal endite of the first maxilliped and the epipod of the third maxilliped. Differences in anten- nal setation were also observed; however, Trask did not give a range for the number of setae occurring on each article, thus overlap between C. oregonensis and C. productus remains a possibility. Trask described the endopod of the 3rd maxilliped as four-segmented; however, five segments are clearly illustrated (1970, Fig. 7i). Quintana and Saelzer (1986) suggested that if the apparently contradictory descriptions of the megalopal stage of Co nceranthonyi Rathbun presented by Trask (1974) and Anderson (1978) were based on correct identifications and observations, geographical differences probably accounted for the great differ- ences in the setation reported. A similar situation may account for the differences in C. productus megalopae. Although Quintana and Saelzer (1986) recommend the use of the antennal setation, as well as the seta- tion of the epipods of the maxillipeds, for distinguishing between Cancer megalopae, these characters cannot be used to separate Puget Sound populations with any degree of reliability. The antennal setation found in C. oregonensis overlaps that reported for C. gracilis; only the setae of the antepenultimate article differ between C. oregonensis and C. productus. The setation reported for C. magister that we have determined from the figure presented by Poole (1966, Fig. 6C) differs from counts made for northern specimens. With the exceptions of larger numbers of setae on the epipods 48 Fishery Bulletin 88(1), 1990 of the first and second maxillipeds of C magister, setal numbers for the maxillipedal epipods overlap among all four local species. In his description of the larval development of Cancer magister, Poole stated that the megalopa possessed five abdominal segments and a telson, and this number was repeated by both Ingle (1981) and Iwata and Konishi (1981) in their reviews of megalopal characters of Cancer species. As all other described larvae possess a six-segmented abdomen, this single character would be expected to provide a very simple and easy means for recognizing megalopae of C. magister. Poole's (1966) description of the megalopae was based on two specimens collected from Drakes Bay, California; how- ever, Poole indicated that he had compared these indi- viduals with laboratory-reared specimens. He noted no significant differences. His figure (1966: fig 6A) indi- cates no suture between the sixth abdominal somite and the telson, and he refers to the uropods as the "pleo- pods of the telson." Unless a major, and evolutionar- ily significant, variation occurs in the postlarvae of this species over its range, Poole's description is incorrect. We have examined a substantial number of C. magister megalopae from the Puget Sound Basin, and in all cases the sixth abdominal somite is clearly separated from the telson by a well marked suture; the uropods arise, as in other megalopae, from the distal margin of the sixth somite. Megalopae of Cancer have been broadly grouped by Orensanz and Gallucci (1988) into the size categories small, medium, and large. In the species of local in- terest, C. gracilis is grouped among those species with small megalopae: C. oregonensis and C. productus in the medium-sized group and C. magister as the single representative in the large category. However, Oren- sanz and Gallucci have reported a bimodal recruitment of C. magister, with the late-summer megalopae being of considerably smaller size. Thus the use of size in dis- tinguishing C. ynagister from other local species may be less reliable during certain periods of the year than previously assumed. The megalopae of C. oregonensis and C. productus are morphologically very close and no definitive means of distinguishing between the two species at this stage has been available. During the course of this study an apparently constant and easily recognizable character has been found that will distin- guish C. productus from C. oregonensis, i.e., the pres- ence in the former species of an acute process on the ventral surface of the ischium of the cheliped (Fig. 4D) that is absent in the latter species (Fig. 2D). In fact, the absence of a spine or process on the ischium of the cheliped distinguishes C. oregonensis megalopae from all three of the other local species. Poole (1966) de- scribed a spine only on the "basi-ischipoidite of the first walking leg" of C. magister; however, in our northern populations we have found that a strong, acute spine is present on the ventrodistal margin on the ischium of the cheliped and on the ventrodistal margin of the coxa of the 2nd and 3rd pereopods. Ally (1975) reports a spine ("hook") on the ventral surface of the ischium of the cheliped in C. gracilis. We have not been able to examine local megalopae of this species; therefore, in preparing the following key to the local species, we have relied on the completeness and accuracy of Ally's description. Key to the megalopae of northern populations of Ca ncer 1 No spine or process on ventrodistal surface/ margin of ischium of cheliped . . . . C oregonensis Acute spine or process on ventrodistal surface/margin of ischium of cheliped 2 2 Megalopae of small size (<3.0 mm); 2nd and 3rd pereopods lacking coxa) spine or process on ventrodistal surface C. gracilis Megalopae of medium to large size (<4.0 mm); 2nd pereopod with coxal spine or process on ventrodistal surface 3 3 Megalopae with acute spine on ventrodistal surface of coxa of 3rd pereopod C magister Megalopae usually without process, or rarely with very small process on ventrodistal sur- face of coxa of 3rd pereopod C. productus Acknowledgments The senior autlior expresses his deep appreciation to Dr. G. Jamieson and to the staff at the Pacific Bio- logical Station, Nanaimo, B.C., for providing the oppor- tunity to participate in the station's sampling progi-am. This research was supported in part by Washington Sea Grant Program #R/F-73-pd and by a National Science Undergraduate Research Grant to Dr. S. Sulkin, Western Washington University. Particular thanks are due Dr. R.B. Manning, National Museum of Natural History, and R. Van Syoc, California Academy of Sciences, for making specimens of Cancer tirannerl and C. oregonensis available to us. We also wish to gratefully acknowledge the assistance and sup- port provided by the staff of the Shannon Point Marine Center. The helpful suggestions of two anonymous reviews substantially improved the manuscrijit. DeBrosse et al Megalopal stages of Cancer oregonensis and C productus 49 Citations Ally, J.R.R. 1975 A description of the laboratory-reared larvae of Cancer granlifi Dana, 1852 (Decapoda, Brachyura). Crustaceana 23(3):231-24(;. Anderson, W.R. 1978 A description of laboratory-reared larvae of the yellow crab, Cancer anthonyi Rathbun (Decapoda, Brachyura), and comparisons with larvae of Cancer magister Dana and Cancer proflurtus. Randall. Crustaceana 34(l):55-68. Botsford, L.W. 1986 Population dynamics of the Dungeness crab {Cancer magister). In Jamieson, G.S., and N. Bourne (eds.). North Pacific workshop on stock assessment and management of in- vertebrates. Can. Spec, Publ. Fish. Aquat. Sci. 92:140-153. Garth, J.S., and D.P, Abbott 1980 25. Brachyura: The true crabs. In Morris, R.IL, U.l'. Abbott, and E.G. Haderlie (eds,), Intertidal invertebrates of California, p. 594-630. Stanford Univ. Press, Palo Alto. Hart, J.F,L. 1982 Crabs and their relatives of British Columbia. British Columbia Provincial Museum Handbook 40, Victoria, 267 p. Ingle, R.W. 1981 The larval and post-larval development of the edible cral), Cancer pagurus Linnaeus (Decapoda: Brachyura). Bull. Mus. (Nat. Hist.) Zool. 40(5):211-236. Iwata, F., and K. Konishi 1981 Larval development of Cancer amphwetus Rathbun, in comparison with those of seven other species of Cancer (Deca- poda, Brachjan-a). Publ. Seto Mar. Biol. Lab. 26:369-391. Kozloff. E.N. 1974 Keys to the marine invertebrates of Puget Sound, the San Juan Archipelago, and adjacent regions. Univ. Wash. Press, Seattle, 226 p. 1987 Marine invertebrates of the Pacific northwest. Univ. Wash. Press, Seattle, 511 p. Lough, G.R. 1976 Lan-al dwamics of the Dungeness crab. Cancer magister. off the central Oregon coast, 1970-1971. Fish. Bull., U.S. 74:353-376. Mason, J.C., and A.C. Phillips 1986 An improved otter surface sampler. Fish. Bull., U.S. 84:480-484. Nations. J.D. 1975 The genus Cancer (Crustacea: Brachyura): Systematics, biogeography and fossil record. Nat. -Hist. Mus. Los Ang. Cty. Sci. Bull. 23, 104 p. Orensanz. J.M., and V.F. Gallucci 1988 Comparative study of postlarval life-history schedules in four sympatric species of Career (Decapoda: Brachyura: Can- cridae). J. Crustacean Biol. 8(2): 187-220. Poole, R.L. 1966 A description of larlioratory-reared zoeae of Cancer magister Dana and megalopae taken under natural conditions (Decapoda Brachyura). Crustaceana n(l):S3-97. (juintana, R., and H. Saelzer 1986 The complete larval development of the edible crab. Cancer setosus Molina and observations on the prezoeal and first zoeal stages of C. coronatus Molina (Decapoda: Brachyura, Caneridae). J. Fac, Sci. Hokkaido Univ., Ser. 6 Zool. 24(4): 267-303. Rathbun, M.J. 1904 Decapod crustaceans of the northwest coast of North America. Harrinian Alaska Expedition 10, 210 p. (reprinted 1910). 1930 The cancroid cral)s of America of the families Euryalidae, Portunidae, Atelecyclidae, Caneridae and .Xanthidae. U.S. Natl. Mus. Bull. 152 , 609 p. Reilly, P.N. 1983 Dynamics of Dungeness crab. Cancer fiiagislcr. larvae off central and northern California. In Wild. P.W., and R.N. Tasto (eds.), Life history, environment, and mariculture studies of the Dungeness crab. Cancer magister, with emphasis on the central California fishery resource. Calif. Dep. Fish Game, Fish. Bull. 172, p. 57-84. Schmitt, W.L. 1921 The marine decapod Crustacea "f California. Univ. Calif. Publ. Zool. 23, 470 p. Shirley, S.M., T.C. Shirley, and S.D. Rice 1987 Latitudinal variation in the Dungeness crab. Cancer magister: Zoeal morphology explained by incubation temper- ature. Mar. Biol. 95:371-376. Trask. T. 1970 A description of laboratory-reared larvae of Cancer pro- ilnclus Randall (Decapoda, Brachyura) and a comparison to lar- vae oi Cancer magister Dana. Crustaceana 18(1):33-146. 1974 Laboratory-reared larvae of Cancer anthonyi (Decapoda: Brachyura) with a brief description of the internal anatomy of the megalopa. Mar. Biol. 27:63-74. Wencker, D.L., L.S. Incze, and D.A. Armstrong 1983 Distinguishing between Chinnocetes hairdi and C. opilio zoeae collected in the southeast Bering Sea. In International Symposium on the genus Chionoecetes, p. 219-230. Alaska Sea Grant Rep. 82-10, Alaska Sea Grant Coll. Prog., Univ. Alaska, Fairbanks. Abstract.- Peak beach-seine i-atclies o( small (60-120 mm FL) juvenile fall chinook salmon in Coos Bay occurred about 30-45 days after peak seine catches farther upstream in the Coos and Millacoma Rivers. The average time between release and capture in the bay of marked hatchery fall chinook salmon was about 30 days, but ranged up to 83 days. Thus, many small hatchery and wild fall chinook salmon remained in Coos Bay for about 1 month befoi'e entering the ocean. Most captures of large (123-156 mm mean FL) tagged spring chinook salmon released di- rectly into the bay occurred within 10 days following release, indicating a shorter period of residence in the bay for spring chinook salmon than for fall chinook salmon. Catches of juvenile spring chinook salmon were very patchy. Potential for competi- tion between juvenile fall and spring chinook salmon in Coos Bay may be reduced l.)ecause of diffei'ences in the timing and locations of maximum abundance of these two groups. Fin- clipped fall chinook salmon grew at least 0.2-0.5 mm per day. Distribution and Residence Times of Juvenile Fall and Spnng Chinook Salmon in Coos Bay, Oregon Joseph P. Fisher William G. Pearcy College of Oceanography, Oregon State University Corvallis, Oregon 97331 Manuscript accepted 31 July 1989. Fishery Bulletin, U.S. 88:51-58. Estuaries are important rearing habi- tats for subyearling chinook salmon (Healey 1982, Reimers 1973, Myers 1980, Nicholas and Hankin 1988, Levy and Northcote 1982). Residence in estuaries of subyearling chinook salmon can be as long as 3 months or more (Reimers 1973, Myers 1980). Growth rates of subyearling chinook salmon are high in some estuaries (Healey 1980, Levings et al. 1986, Argue et al. 1986), but low in others, perhaps because of competition for food (Reimers 1973, Neilson et al. 1985). Because of the long period of residence and active feeding of juve- nile chinook salmon in estuaries, the release of large numbers of hatchery chinook salmon into a system could impact survival and growth of wild fish. The interactions among groups of chinook salmon depend on their overlap in the estuary in time and space. Many large subyearling spring chi- nook salmon smolts were released in 1987 into Coos Bay, C^regon, by Anad- romous. Inc., a private salmon ranch- ing facility. These fish ranged from about 120 to 160 mm mean fork length (FL) and were generally larger than wild subyearling chinook salmon found in Oregon estuaries (Reimers 1973, Myers 1980, Nicholas and Hankin 1988). The large spring chinook salm- on may compete with smaller fall chi- nook salmon for estuarine resources. However, estuarine dependency and emigration rates of spring and fall chinook salmon in Coos Bay are not known. To assess the potential for competition between juvenile spring chinook salmon released by Anad- romous. Inc. and fall chinook salmon, we studied their temporal and spatial overlap in beach-seine samples. Methods We collected juvenile chinook salmon in Coos Bay (lat. 43°21'N, long. 124°20'W) with a 60 x 2.5 m beach seine between 25 April and 10 Octo- ber 1987. The seine had 19- and 13-mm mesh (stretch-measure) in the wings and bunt, respectively. The net was set with the current using a 6.1-m dory powered by a 50 hp out- board motor. Juvenile salmon were counted, measured, and checked for fin-clips immediately after capture. When catches were large, juvenile salmon were kept alive in a floating net pen during processing. Salinity was estimated with an American Op- tical (Model TS) refractometer to the nearest "Am and temperature mea- sured to the nearest 0.1 °C. Five stations were sampled regu- larly (Fig. 1). Stations 1, 2, 3, 4, and 5 were 2.0, 3.3, 6.5, 9.8, and 14.8 km, respectively, from the mouth of the bay. The Anadromous, Inc. release facility was located on North Spit between sites 3 and 4. On most dates each station was sampled twice. Reference to trade names does not imply en- dorsement by the National Marine Fisheries Service, NOAA. 51 52 Fishery Bulletin 88(1). 1990 I ... V I 125- 124" Figure 1 Coos Bay, Oregon, showing five stations routinely sampled and two sites occasion- ally sampled (lA and 3 A) for juvenile Chi- nook salmon with a beach seine in 1987. The Anadromous, Inc. release site on North Spit is indicated by an arrow. The Milia- coma River (not shown) is a tributary of the Coos River. Also shown are other estuaries mentioned in the Discussion. Generally, we sampled from the lower to the upper bay in the morning and in the opposite direction in the afternoon. Occasionally sets were made at two othei- sites (lA and 3A in Fig. 1). The substrate at all stations was sand except at sta- tion 5 where it was gravel and mud. During low tide the seine usually sampled parts of eel grass beds at sta- tions 2, 3, and 4. About 415,000 subyearling fall chinook salmon were released in 1987 by the Salmon and Trout Enhance- ment Program (STEP) into tidewater tributaries of the Coos River between 27 and 35 km above station 5 (Fig. 1). These subyearling fall chinook salmon were released between 30 April and 28 June (half before and half after 23 May). Average fork length (FL) of fish at release ranged from approximately 48 to 94 mm (converted from mean weights, T. Rumreich and R. Bender, Oreg. Dep. Fish Wiidl., P.O. Box 5430, Charleston, OR 97420, pers. commun., March 1988). Of STEP fail Chinook salmon released in 1987, 74% were supposed to be fin-clipped. However, the actual percentage of fish with recognizable marks was not known because marking efficiency was not evaluated. Wild fall chinook salmon caught in the lower Millacoma and Coos Rivers were about the same size as STEP fish. Over 5 million subyearling spring chinook salmon were released into Coos Bay in eight groups between 19 June and 1 October 1987 from Anadromous, Inc.'s holding and release facility located on North Spit (Fig. 1). These spring chinook salmon were consider- ably larger than the STEP or wild fall chinook salmon and averaged 123-156 mm FL at release. Between 0 and 5.9% (average 3.6'7o) of the fish in each group were coded-wire tagged (CWT) and had clipped adipose fins. We sampled the bay 1-2 days before and after each Anadromous, Inc. release and at about weekly intervals between releases. Fisher and Pearcy Distribution and residence of juvenile chinook salmon 53 1— Ld a: UJ a. u < o 2 1 2 1 2 1 5 4 3 2 1 A May 3 1 1tS_ 1 1 1 2 1 6 5 4 3 2 1 2 1 2 1 10 8 6 4 2 1 1 1 6 (m .rlliTn T July 30 t JklYn ^ "^9 ' June 7 „ .JVTr.- ^^9 ' M 2 1 June 1 9 3^_ n ^_ _ ^ Aug 5 -r ^1 - [si June 20 r 4s s -| Aug 1 3 AA -1 3 2 1 iis June 21 s iH n rii rfl Aug 20 J -rfi Aug 30 4 3 2 1 3 2 1 Jd June 29 "If n S A 4 aIs] niJ iTl Sept 3 aaa|^ ^AAi i Sept 4 '1 1 A AA -h jfss July 7 s Vnmnu J-, -v^ July 1 5 r s r. Sept 10 ^ July 17 2 1 _r s ^ Sept 22 Oct 3 1 _r. : July 29 _ Oct 10 6 0 100 140 180 FORK LENGTH 0 100 140 180 Figure 2 Len^h-lVequency distributions of juvenile chi- nook salmon caught in Coos Bay, Oregon, in beach seines in 1987. Numbers of marked fall chinook salmon released by the Salmon and Trout Enhancement Program (S) and spring chinook .salmon released from Anadromous. Inc. (A) caught in each length category are indicated. An "A" or "S" without a number represents a single fish. Heavy horizontal lines indicate lengths ( ± 2 SD) of production groups of spring chinook salmon released by Anadromous, Inc. Arrows indicate the lengths used to separate fall and spring chinook. Results Immigration of fall chinool< salmon Juvenile wild and STEP fail chinooi< salmon were caught in the bay in large numbers starting in late May (Fig. 2). Only two juvenile chinook salmon were cap- tured during our first two sampling trips on 25-26 April (8 sets) and 9 May (6 sets). Catches of juvenile wild and STEP fish in the lower reaches of the Milla- coma and Coos Rivers entering the bay peaked on 15 May and 26 May, respectively, and declined to low levels after mid- June (R. Bender, Oreg. Dep. Fish Wildl., P.O. Box 5430, Charleston, OR 97420, pers. commun., March 1988). Size-frequency distributions Distinct modes in the length-frequency distribution of juvenile chinook salmon provided a basis for separating small STEP and wild fall chinook salmon from large Anadromous, Inc. chinook salmon until the end of July (Fig. 2). Recoveries of fin-clipped STEP fall chinook salmon (S in Figure 2) confirmed that the mode of small 54 Fishery Bulletin 88(1), 1990 fish was mainly fall chinook salmon from 31 May through the end of July. Almost all fall chinook salmon captured before the first release of spring chinook salmon on 19 June were less than 105 mm FL. Al- though we caught some large spring chinook salmon after the first release on 19 June, a large, distinct mode of these fish was not apparent until 10 days later on 29 June. Recoveries of Anadromous, Inc. spring chinook salmon with CWTs (A in Figure 2) indicated that this second mode of large fish was mostly spring chinook salmon. This mode of large-sized fish was also obvious a week later on 7 July, but by 16 and 17 July catches of large fish had decreased to low levels. Catches of large chinook salmon did not increase im- mediately following the 17 and 29 July releases. We classified fish as either fall or spring chinook salmon to estimate their relative abundances in the bay. On 19, 20, and 21 June fish that were < 105 mm FL were considered fall chinook salmon and the larger fish were considered spring chinook salmon. (Arrows in Figure 2 indicate the division between these groups). On 29 June and 7 July, we used the valley between the two distinct modes to separate fall and spring chinook salmon. To account for gi-owth of fish, we distinguished fall and spring chinook salmon at slightly larger lengths on 16, 17 July and 29, 30 July (120 and 125 mm FL, respectively). In August, length ranges of fall and spring chinook salmon overlapped and the two groups could not be separated by length. However, catch per set of fall chinook salmon on 29 and 30 July was lower than it had been in June and earlier in July, suggesting that abundance of fall chinook salmon probably peaked in June and July. Catches of fin-clipped fall chinook salmon (S in Figure 2) in August were also low com- pared with earlier periods. Therefore, we assumed that subsequent increases in catch per set of juvenile chi- nook salmon in August (Fig. 2) were due almost ex- clusively to releases of large spring chinook salmon from Anadromous, Inc. In the following discussion we have treated all fish caught starting 3 August as spring chinook salmon, realizing that this probably overesti- mates the abundance of spring chinook salmon in the bay, especially in early August when a few marked fall chinook salmon were caught. The proportions of adipose clipped or CWT fish in our catches of spring chinook salmon were usually similar to the proportions of CWT fish in the immedi- ately preceding releases of spring chinook salmon from Anadromous, Inc. (r tests, p>0.05) (Table 1). This supports the conclusion that most fish we classified as spring chinook salmon originated from the Anadro- mous, Inc. facility. An exception was during the period 4-30 August when the proportion of CWT fish in our catch of spring chinook salmon was significantly lower than the proportion released by Anadromous, Inc. on Table I Percentage of coded-wire tagged (('WT) (ir adi|inse-cli|i|ied fish in the catch of spring chinook salmon from Coos Bay, Oregon, vs. the percentage of CWT fish in the preceding release from Anadromous, Inc. Catch Anadromous Inc. % CWT in release Period n % CWT" x' 6/19-7/7 232 2.6 2.3 0,08 7/17-8/3 92 1.1 2.5 0.75 8/4-8/30 751 0.7 1.9 6.14" 9/3-9/22 775 2.3 2.9 0.93 10/3-10/10 50 4.0 5.9 0.32 ' Includes 16 Anadromous, Inc. CWT fish, 1 1 fish with unread- al)le CWTs probably from Anadromous Inc., and 5 adiposc- eli|jped fish without tags. * Observeil frequency of tags is significantly difl'ircnt frotn expected fre()uency, ;i< 0.025. 4 August (x- = 6.14, p<0.025). This result may be explained by lack of complete mixing of marked and unmarked Anadromous, Inc. fish and their patchy distribution. Over 77% of the catch in August occurred in just five sets, three of which were on the same day. Large numbers of fall chinook salmon in the bay in August mistakenly classified as spring chinook salmon also could have produced the low proportion of CWT fish. However, there is little direct evidence that fall chinook salmon were abundant in the bay during Aug- ust since very few fin-clipped fish were recovered, although regeneration of fins may have made recogni- tion of marks difficult (R. Bender, Oreg. Dep. Fish Wildl., P.O. Box 5430, Charleston, OR 97420, pers. commun., March 1988). Catch distribution Before mid-June, roughly e(iual numbers of fall chinook salmon were caught at each of the five standard sta- tions (Fig. 3). After mid-June, fall chinook salmon were concentrated near the mouth of the bay at station 1, although they were also caught at the other stations. At station 1, catch per set of fall chinook salmon peaked in June and July and dropped to low levels at the end of July (Fig. 3). Catches of Anadromous, Inc. spring chinook salmon were extremely patchy, with large catches at one or two stations and low catches at the others. After the first release on 19 June, almost all spring chinook salm- on were found upbay at station 5, but later releases were caught at the lower bay stations 1, lA (not shown), and 2 (Fig. 3). Catch per set of spring chinook Fisher and Pearcy Distribution and residence of juvenile Chinook salmon 55 IS 12 ANADROMOUS. INC. SPRING CHINOOK A FALL CHINOOK »'l- 5PRING CHINOOK B STA. 5 STA. 4 ,^•4 ^. 111 100 if) HIIt 11- UJ Q. 60 I 40 ( ) 1— 29"/(». after mid-June), indicating little influence of freshwater at our sampling sites. (See also Biu't and McAlister 1959.) Residence in the bay Unmarked STEP or wild fall chinook salmon resided in the shallow nearshore areas of the bay for about 1-2 months. Catch per set of fall chinook salmon was highest between 20 June and 17 July (Fig. 3), about 1.0-1.5 months later than the peak catches in the river systems just above the bay (R. Bender, Oreg. Dep. Fish Wildl., P.O. Box 5430, Charleston, OR 97420, pers. commun., March 1988). Mean duration of residence in the bay of 63 fin- clipped STEP fall chinook salmon was about 1 month (Table 2). The average number of days (weighted by catch per set of the marked fish) between release of the median fish in a mark gi'oup and recovery in the bay of fish from that same mark group was 29 days (n =63, range -6 to 83 days). Eight fin-clipped fish were caught more than 50 days after release. The number of days between release of the last fish in a mark-group and the recovery in the bay of fish from that same mark-group, a minimal estimate of time since release, averaged 24 days {)i =63, range - 13 to 81 days). Average recovery date of right-pelvic and anal fin-clipped fish was 27 and 32 days, respectively, after release of the median fish (Table 2). Peak catch per set of spring chinook salmon usually occurred within 1-10 days after Anadromous, Inc. releases, and catches declined ra|)idly afterwards, sug- gesting that spring chinook sahiKni had a much shorter period of residence in shallow waters of the bay than fall chinook salmon (Fig. 3). Catches returned to low levels within about 25 days following the releases on 19 Jime and 29 July-4 August and within 7-10 days following the releases on 31 August and 3 Sej)tember. The very rapid decline in catch per set after the 3 1 Aug- ust and 3 September releases, together with the low catches of fish from the last two releases on ^9 Sep- tember and 1 October, indicate that movement into deepwater channels or out of the bay for these late summer releases may have been more rapid than for earlier releases. All but one of the 27 CWT Anadromous, Inc. spring chinook salmon caught in beach seines were captured 10 or fewer days after release* (range 1-18 days); *Tags from eleven of these fish were unreadable; however, these unreadable tags were prohalily from the 31 Aug. ;ind 1 Oct. releases (Mary McGowan, Anadromous. Inc., P.O. Ko.\ 1(107, North Bend, OR 974.'')9. pers. commun., ,Ian. 1989). Fisher and Pearcy Distribution and residence of juvenile chinook salmon 57 another indication that residence of spring chinook salmon in the bay is relatively short. Moreover, CWT spring chinook salmon from each release were recov- ered only during the period before the next release. Growth rates Growth rates of anal and right-pelvic clipped fall chi- nook salmon, estimated from the slopes of the linear regressions of fork length on days since release of the median fish, were 0.54 mm/day {n = 19, r- =0.74) and 0.29 mm/day (h =33, r- = 0.33), respectively. Because emigration from the bay may be positively related to fish size, these observed growth rates probably under- estimate the true average rates of growth of fish in the mark groups. Discussion Our data indicate that small fall chinook salmon reside in Coos Bay for a longer period than do larger spring chinook salmon. Duration of residence in estuaries may be related to the size or stock of fish. Dawley et al. (1986) found that rates of downstream movement of subyearling chinook salmon from lower Columbia River stocks were positively related to fish length. Movement of small subyearling fish through the Columbia River estuary decreased by 30% relative to movement rates farther upstream, while larger yearling fish moved through the estuary at the same rate as through the river. Neither subyearling nor yearling fish, however, reared for long periods in the Columbia River estuary. Because our collections were made by beach seine in shallow nearshore areas of Coos Bay, we can say little about the utilization of deeper channel areas by juvenile fall and spring chinook salmon. We do not know whe- ther peak abundances of fall or spring chinook salmon in channels coincide with peak abundances in shallow areas. Neither do we know what fraction of fall or spring chinook salmon at any given time are in shallow or channel areas. Temporal and size-related differences in utilization of nearshore and channel areas by juvenile chinook salmon have been found in other estuaries and also may occur in Coos Bay. (See Figure 1 for locations of other estuaries discussed.) In Yaquina Bay, Myers (1980) found that although catches in shallow areas peaked in late July and early August, catches continued to increase in channels into October. In both Yaquina Bay and the Columbia River estuary, small fish and large fish preferentially utilized nearshore and chan- nel areas, respectively. The mean lengths of wild juve- nile chinook salmon in beach seine (nearshore) and lam- para net (channel) catches in Yaquina Bay during June were 88 and 106 mm FL, respectively (Myers 1980). In the upper Columbia River estuary, many more year- ling chinook salmon (large fish) were caught in chan- nels than in nearshore areas and, conversely, more subyearling chinook salmon (smaller fish) were found in nearshore areas than in channels (Dawley et al. 1986). In addition, subyearling fall chinook salmon caught in nearshore areas were 10-20 mm shorter than those caught in channel areas, and catch rates of sub- yearling chinook salmon in nearshore areas were in- versely related to length (Dawley et al. 1986). If similar size-related distributional patterns occur in Coos Bay, then the large spring chinook salmon released from Anadromous, Inc. may utilize channel areas much more extensively than shallow nearshore areas. Delays of up to 10 days occurred between releases of spring chinook salmon from the Anadromous, Inc. facility and peak catches of juvenile spring chinook salmon in nearby (<8 km distant) shallow areas (Figs. 2 and 3). This suggests that spring chinook salmon may stay in channels for several days following release and then disperse into shallow waters. Some of the spring chinook salmon occurring at stations 1 and 2 near the mouth of the bay, especially those found several days after a release, may even have reentered the shallows from the ocean. Two CWT juvenile chinook salmon released earlier in 1987 in Yaquina Bay were found later in Coos Bay at station 1. This demonstrates that juvenile chinook salmon, after they have entered the ocean, sometimes reenter estuaries. The apparent rates of growth in length for two groups of fin-clipped STEP fall chinook salmon caught in Coos Bay (0.29 and 0.54 mm/day) were similar to rates reported for a group of marked subyearling fish caught in the upper Columbia River estuary (Dawley et al. 1986; 0.60 mm/day) and reported by Levings et al. (1986) for fry caught in the Campbell River estuary (0.46-0.70 mm/day) but lower than reported by Healey (1980) for fry in Nanaimo Estuary (1.32 mm/day) or by Argue et al. (1986) for smolts in Cowichan Bay (0.97 mm/day). All these calculations of growth rates were based on changes in length of marked fish with time and should have a similar bias (i.e., possible faster emigi-ation of fast-growing or large fish that results in underestimates of actual mean growth rates attained in the bay). Reimers (1973) reported little change in size of marked and unmarked fish in the Sixes estuary June through August, after which growth rates in- creased dramatically. He attributed the slow growth of juvenile fish during the June- August period to their high densities in the bay. Myers (1980) found substan- tial increases in size of wild chinook salmon with time and between the upper and lower bay, indicating sub- stantial growth of wild chinook salmon. Thus growth of juvenile chinook salmon appears to vary among estuaries, perhaps depending on their density and food supply. 58 Fishery Bulletin 88(1). 1990 All interesting feature of the catch distribution of fail and spring chinook salmon in Coos Bay (Fig. 3) is that when the catch of fall chinook salmon was highest at the downbay station 1 (20 June-29 July), the catch of spring chinook salmon at this station was low. Con- versely, when the catch of spring chinook salmon was high at the upbay station 5 (29 June-7 July), the catch of fall chinook salmon at this station was relatively low. Later in the summer when most spring chinook salmon were released, fewer fall chinook salmon were caught in the bay. Apparently peak abundances of juvenile fall and spring chinook salmon differed in time and place. The potential for competition between these two stocks was probably greatest in June and July after the large release of spring chinook salmon and when large num- bers of fall chinook salmon were present in shallow nearshore areas of the bay. However, since the two groups were found at different sites, direct competi- tion may have been limited. Acknowledgments We thank Karl Brookins for his skillful boat handling and beach seining, Mr. and Mrs. Jack Brookins for their kind hospitality, and Alton Chung. Matt Wilson, Ann Raich, and Karen Young for their help during sampling operations. Ron Gowan and Dan VanSlyke of Anadro- mous, Inc. and Reese Bender and Tom Rumreich of Oregon Department of Fish and Wildlife (ODFW) pro- vided information on sampling and hatchery releases of fish in Coos Bay. Jay Nicholas of ODFW and two anonymous reviewers provided helpful comments on the manuscript. This research was funded by the Northwest and Alaska Fisheries Center (Contracts NA-85-ABH-00025 and NA-87-ABH-00014), National Marine Fisheries Service, NOAA. Citations Argue, A.W., B. Hillaby, and CD. Shepard 1986 Distribution, timing, change in size, and stdniach ocmtents of juvenile chinool< and coho salmon caught in Cowichan estuary and bay, 1973, 197.5, 1976. Can. Tech. Rep. Fish. Aquat. Sci. 1431. KiS |i. Burt, W.V., and W.B. McAlister 1959 Recent studies in the hydrograpliy of Oregon estuaries. Oreg. Fish Comm. Res. Briefs 7(l):14-27. Davvley, E.M., R.I). Ledgerwood. T.H. Blahm, C.W. Sims, J.T. Durkin, R.A. Kirn, A.E. Rankis, (i.K. Monan, and K..I. Ossiander 1986 IVIigrational characteristics, liiological ol)servations, and relative survival of juvenile salmonids entering the Columbia River estuary, 19B6-1983. Res. rep., U.S. Dep. Energy, Bon- neville Power Adm., Div. Fish Wildl., P.O. Box 3G21, Portland, OR 97208, 256 p. Healey, M.C. 1980 Utilization of the Nanaimo River estuary by juvenile chinook salmon, Onairhynchu-'i trihaurytHcha. Fish. Bull., U.S. 77:653-668. 1982 Juvenile pacific salmon in estuaries: The life system, fn Kennedy, V.S. (ed.), Estuarine comparisons, p. 315-341. Acad. Press. NY. Levings, CD., CD. McAllister, and B.D. Chang 1986 Differential use of the Campbell River estuary, British Columbia, liy wild and hatchery-reared juvenile chinook salmon (Oncorhynrlms tnkawytxchd). Can. .1. Fish. A(|uat. Sci. 43: 1.386-1397. Levy, D.A., and T.G. Northcote 1982 Juvenile salmon residency in a marsh area of the Fraser River Estuary. Can. J. Fish. Aquat. Sci. .39:270-276. Myers, K.W.W. 1980 An investigation of the utilization of four study areas in Yaquina Bay, Oregon, l)y hatchery and wild juvenile salmonids. M.S. thesis. Oregon State Univ., Corvallis, 234 p. Neilson, J.D., G.H. Geen, and D. Bottom 1985 Estuarine growth of juvenile chinook salmon (Oncorhi/n- chus tahawytacha) as inferred from otolith microstructure. Can. J. Fish. Aquat. Sci. 42:899-908. Nicholas, J.W., and D.G. Hankin 1988 Chinook salmon populations in Oregon coa.stal river basins: description of life histories and assessment of recent trends in run strengths. Info. Rep. 88-1, Oreg. Dep. Fish Wildl., Fish. Div., Portland 97201, 3.59 p. Reimers, P.E. 1973 The length of residence of juvenile fall chinook .salmon in Sixes River, Oregon. Oreg. Fish Comm. Res. Rep. 4(2), Oreg. Dep. Fish. Wildl.. Portland 97201, 43 p. Abstract. - Three species of basses iif the genus Paralahrax are found in southern California coastal waters. Adults can be identified on the basis of moqjhological characters, but the eggs and early larval stages of the three species are extremely similar in appearance. This paper re- ports an investigation of biochemical genetic characters for specific iden- tification. An electrophoretic analy- sis of 43 presumptive gene loci dem- onstrated several genetic differences between any two of the three spe- cies, but no single locus was able to unambiguously distinguish the three species. In contrast, an analysis of Parnlahrax mitochondrial DNA (mtDNA) demonstrated that 8 of 13 informative restriction endonuleases produced species-specific fragment patterns. A technique is described for the relatively rapid enrichment and analysis of mtDNA from both fresh and ethanol-preserved indi- vidual eggs and early larvae which allows specific identification on a cost-effective basis. Biochemical Genetics of Southern California Basses of the Genus Paralabrax: Specific Identification of Fresh and Ethanol-preserved Individual Eggs and Early Larvae John E. Graves Department of Biology, University of San Diego, San Diego, California 921 10 Present address: Virginia Institute of Marine Science, College of William and Mary Gloucester Point, Virginia 23062 Michelle J. Curtis Paul A. Oeth Department of Biology, University of San Diego San Diego. California 921 10 Robin S. Waples Department of Biology, University of San Diego, San Diego, California 921 10 Present address Nortfiwest Fishieries Center National Marine Fisfieries Service, NOAA 2725 Montlake Blvd. East, Seattle. Washington 981 12 Manuscript accepted 21 September 1989. Fishery Bulletin, U.S. 88:.59-66. Tlie serranid genus Paralabrax is endemic to the eastern Pacific Ocean and comprises seven species. Three species, the kelp bass P. clathratus, barred sand bass P. nebulifer, and spotted sand bass P. maculatofascia- tus, are common in southern Cahfor- nia coastal waters and all are impor- tant components of the California sport fishery (Oliphant 1979). Adults of these three species can be separ- ated on the basis of morphological differences (Miller and Lea 1972); however, it is not possible at this time to determine the specific identity of field-caught Pa /•/a6ra.r eggs, larvae, or early juveniles using morpholo- gical characters (Butler et al. 1982; R. Lavenberg, Los Ang. Cty. Mus. Nat. Hist., Los Angeles, CA 90007, pers. commun., June 1988). Ecolo- gical studies of the early life history of the three species, as well as stock assessment studies based on the early-life-history stages, have been hampered by the inability to specif- ically identify Paralabrax eggs and larvae. A variety of biochemical characters have been used to identify the eggs and larval forms of closely related fishes in the absence of discrimin- ating morphological characters. Elec- trophoresis of water soluble proteins (allozyme analysis) has been used to differentiate species of marine fishes which have morphologically similar eggs, larvae and early juvenile stages (Morgan 1975, Smith and Crossland 1977, Sidell et al. 1978, Smith et al. 1980, Mork et al. 1983, Graves et al. 1989). However, due to the small sizes of eggs and larvae, and the pos- sibility of low enzyme activity during early-life-history stages, some studies have reported difficulty in resolving the electrophoretic bands of eggs and very small larvae (Smith and Cross- land 1977, Sidell et al. 1978, Smith et al. 1980, Graves et al. 1989). Restriction endonuclease analysis of mitochondrial DNA (mtDNA) has 59 60 Fishery Bulletin 88(1). 1990 been used to demonstrate intra- and interspecific genetic relationships among and within various marine and freshwater fishes (Ferris and Berg 1987). This technique provides an additional tool for the investiga- tion of egg and larval identities. Although original pro- tocols were costly, time-consuming, and generally not applicable to the small amount of tissue available with early-life-history forms, a modification of existing methodologies has provided a relatively simple method for distinguisliing the eggs and larvae of closely related (congeneric) species. In this paper, we report the results of an investigation of water-soluble protein and mtDNA restriction fragment differentiation among the three southern California species ofParalabrax to find a reliable genetic marker to discriminate the early-life- history stages of the three species. A cost-effective means of analyzing mtDNA restriction fragments of both fresh and ethanol-preserved individual eggs is presented. Materials and methods Adult specimens of the three Paralabrax species were collected by hook and line, pole spear, and beach seine off San Diego and La Jolla, California. After collection, fish were transported to the laboratory alive or on ice. Samples of liver, eye, and muscle tissue were removed and frozen at -25°C until electrophoretic analysis was undertaken. Gonads and livers were removed and used fresh or stored at -70°C for mtDNA analysis. Eggs and larvae of the three species were obtained from a captive brood stock maintained at the Southern California Edison research facility in Redondo Beach, California. Eggs were collected within 12 hours of spawning and either transported to the laboratory in San Diego in seawater or preserved in 95% ethanol. Protein electrophoresis Four specimens each of Paralabrax niacidatofasciatus and P. nebulifer were collected for the electrophoretic investigation. The sample of P. clathrntu:^ (45) was larger because individuals were collected as part of a more extensive investigation of gene flow within south- ern California coastal fishes (Waples and Rosenblatt 1987). Samples of muscle, liver, and eye tissue were individually macerated in an approximately equal vol- ume of 0.1 M potassium phosphate buffer, pH 7.0, before centrifugation for 10 minutes at 16,000 g, 5°C. Supernatants were loaded on horizontal starch gels using procedures similar to those described by Selander et al. (1971). Staining recipes for the 24 enzyme and protein systems studied (Table 1) were modified from Shaw and Prasad (1970) and Harris and Hopkinson (1976). A more detailed description of the electrophor- etic procedures, including all staining recipes, is pre- sented in Waples (1986). Proteins encoded by multiple genes (isozymes) were numbered according to decreasing anodal mobility. For each gene locus the mobility of the most common allele in Paralabrax clathratus was arbitrarily designated 100, and alternate alleles were numbered in accordance with their mobility relative to this standard. Genetic similarity and genetic distance were computed from the allele frequency data using Nei's (1978) method, which corrects for small sample size. Mitochondrial DMA analysis Mitochondrial DNA for restriction analysis was jiuri- fied from fresh or frozen gonad or liver samples from six individuals of each species following the protocols for equilibrium density gradient ultracentrifugation of Lansman et al. (1981), with minor modifications. Typi- cal yields were 1 microgram of mtDNA per gram of fresh tissue. A mini-prep procedure was developed for isolating mtDNA from small tissue samples based on the tech- niques described by Chapman and Powers (1984). Tissue samples as small as an individual egg (0.6 mg) were homogenized in 0.15 mL 10 mmol/L TRIS, 10 mmol/L EDTA, pH 7.4 (TE buffer) in a 0.2-mL ground glass homogenizer. The homogenate was transfered to a 1.5 mL microfuge tube and an additional 0.5 ml of TE buffer added. The tube was spun at 800 g, 4°C, for 3 minutes to remove nuclei and cellular debris. The supernatant was transferred to a second microfuge tube and centrifuged at 12,000 g, 4°C, for 20 minutes to pellet mitochondria. The mitochondrial pellet was resuspended in 0.4 mL TE buffer and lysed with 0.04 mL 10% SDS. The mitochondrial lysate was extracted with an equal volume of a 25:24:1 phenol/chloroform/ isoamyl alcohol solution and then an equal volume of 24:1 chloroform/isoamyl alcohol. The DNA from the aqueous layer, after the addition of ammonium acetate, was precipitated in two volumes of 95% ethanol at -70°C for at least 2 hours. The following restriction endonucleases employed in this study were purchased from Bethesda Research Laboratories (BRL) and used according to the sup- plier's instructions: Aral. ArnW, BamWl, B(fl\\, BslEll, Clal, Ecom, HuicU, HitidlU, Hinjl, HpaU, Kpnl, Pstl, PvhU. Sail. Smal, SkU. Xbal. andXhoL Endlabelling was performed accoi-ding to the protocols of Brown (1980), with the exception that an ethanol precipitation was not necessary for digestions with restriction endonucleases that recognized six liase pairs. Electrophoresis of restriction fragments was per- formed on both large and mini-submarine horizontal Graves et al Biochemical genetics of Paralabrax 61 Table 1 Enzymes and proteins surveyed and tissues and buffers used in the electrophoretic analysis c f three species of Paralabrax. Protein (EC number) Locus Buffer* Tissue'* Aconitate hydratase (4.2.1.3) Aeon 1 1 Adenosine deaminase (3.5.4.4) Ada 1 m Adenylate kinase (2,7.4.3) Ak 1 1 Alcohol dehydrogenase (1.1.1.1) Adh-1 2 1 Adh-2 2 1 Aspartate aminotransferase (2.6.1.1) Aat-1.2 1,2 1 Creatine kinase (2.7.3.2) Ck-A 2 m Ck-B 2 e Esterase (3.1.1.-); o-naphthyl acetate Est-1,2,3 3 m Fumarte hydratase (4.2.1.2) Fum 1 1 Glucose-6 phosphate dehydrogenase (1.1.1.49) G6pdh 1 1 Glucosephosphate isomerase (.5.3.1.9) Gpi-A 2,3 m,l Gpi-B 2,3 m Glutamate dehydrogenase (1.4.1.2) Gdh 1 1 Glyceraldehyde-phosphate dehydrogenase (1.2.1.12) Gapdh-1 1 1 Gapdh-2 1 m Glycerol-3-phosphate dehydrogenase (1.1.1.8) G3pdh-A 1 m G3pdh-B 1 1 Iditol dehydrogenase (1.1.1.14) Iddh 1 1 Isocitrate dehydrogenase (NADP) (1.1.1.42) Icdh-s 1 m Icdh-m 1 1 Lactate dehydrogenase (1.1.1.271 Ldh-A 1 m,e Ldh-B.C 1 e Malate dehydrogenase (1.1.1.37) Mdh-A.B,m 1 ni Mannosephosphate isomerase (5.3.1.8) Mpi 1 1 Phosphoglucomutase (5.4.2.2) Pgm 2 m Phosphogluconate dehydrogenase (1.1.1.44) Pgdh 1 m Peptidase (3.4.11.-); leucyl-tyrosine Pep-1 3 ni Peptidase (3.4.11.-); leucylglycyl-glycine Pep-2 3 m Peptidase (3.4.11.-); non-specific Pep-3 3 m Superoxide dismutase (1.15.1.1) Sod 2 1 Xanthine dehydrogenase (1.1.1.204) Xdh 1 1 General proteins Pro-1,2,3.4,5 3 Tris-citric acid, pH 8.0 (Selandei m et al. 1971); "Buffers: 1 = Tris-citric acid, pH 6.9 (Whitt 1970); 2 = 3 = Lithium hydroxide (Selander et al. 1971). **ni = mu.scle, 1 = liver, e = eye. agarose gels (0.8-1.2%) and vertical polyacrylamide gels (3.5-5.0%) at 3 volts/cm. The Southern transfer and hybridization protocols of Maniatis et al. (1982) were followed. Hybridization probe mtDNA (purified Pamlahrnx mtDNA) was nick translated with biotinylated dATP using the proce- dures of the BRL Nick Translation System. Visualiza- tion of the fragments followed the procedures included with the BRL BluGene Gene Detection Kit. The mean mtDNA nucleotide sequence divergence between the three Paralahrnx species was calculated Reference to trade names does not imply endorsement by the Na- tional Marine Fisheries Service, NOAA. from the numl^er of shared fragments using the algo- rithm of Nei and Tajima (1983). Results In the electrophoretic analysis, 43 presumptive gene loci could be resolved in all three species (Table 1). The genetic interpretation of observed banding patterns was guided by expectations based on known patterns of tissue specificity of expression and subunit composi- tion of enzymes in fishes and other vertebrates. Discus- sion of banding patterns for each enzyme system can be found in Waples (1986). Twenty-two loci were mono- morphic (fixed for the same allele) in all individuals 62 Fishery Bulletin 88(1), 1990 Table 2 Allele frequencies at variable loci in three Ptira.liibrax species. P. P. P. P. P. P. Locus/Allele clathratus nebulifer maculatofasciatus Locus/Allele clathratus nebulifer maculatofasciatus Aat-1 130 0.167 Ic 2N 90 8 8 Est-3 108 - - 1.0 Pep-1 lOS 0.012 _ 100 0.988 1.0 — 100 0.988 1.0 1.0 94 0.012 — — 2N 82 s 8 2N 82 6 6 Pep-2 110 0.125 Gapdli-2 100 0.989 — 1.0 100 1.0 0.875 1.0 - 12O0 O.OU 1.0 — 2N 42 8 8 2N 90 8 6 Pep-3 110 0.011 G3pdh-2 100 1.0 0.125 1.0 100 0.989 1.0 1.0 (G3pdh-A 60 2N 90 0.875 8 8 Pr(.-2 2N 110 90 8 8 1.0 Gpi-2 150 — 0.25 — 100 1.0 1.0 (Gpi-B) 100 1.0 0.75 1.0 2N 90 8 8 2N 90 8 8 Pro-4 100 1.0 1.0 G6p(:l)i 101) 1.0 — 1.0 80 1.0 90 — 1.0 — 2N 78 6 6 2N 50 4 4 Sod 105 _ 1.0 1.0 Icdlvl 100 1.0 — — 100 1.0 _ (Icdh-s) 95 2N 10 1.0 6 1.0 6 2N 80 8 8 sampled: Aat-2, Ak, Aeon. Adh-1. Ck-B, Est-1, Est-2, Fum, Gpi-A, Gdh, Gapdh-1, G.3pdh-B, Ldh-A, Ldh-B, Ldh-C, Mdh-A, Mdh-B, Mdh-m, Pro-1, Pro-3. Pro-5, and Xdh. The remaining 21 loci were polymorphic either within or between species (Table 2). Genetic similarity values for the three pairwise com- parisons (Table 3) were between 0.70 and 0.85, which is near the high end of the range of genetic similarity values found between congeneric fish species (Shaklee et al. 1982, Thorpe 1983). Nevertheless, each species pair can be distinguished on the basis of apparent fixed differences at multiple gene loci. The samples of Paralahrax clathratus and P. nebulifer shared no alleles at seven presumptive gene loci (Adh-2, Ck-A, G6pdh, Icdh-S, Pgdh, Pro-4, Sod); P. clathratus and P. maculatofasciatus were distinct at six loci (Ck-A, Table 3 Values of genetic dist mce (al ove diag inal) and nitONA nucleotide sequence differential on (below diagonal) between three Parahibrax species. Standard errors of the estimates | are in parentheses. P.c. P.n. P.m. Pa rnlahrax rhilfi riifi/^ - 0..304 (0.090) 0.265 (0.085) P(i riihihrnx nebulifer 0.145 (0.026) — 0.165 (0.063) Ptiriiltibrdx maeuUiUtfdnriatux 0.142 0.069 — (0.027) (0.019) Graves et al ■ Biochemical genetics of P^^lsbrsx 63 Est-3, Icdh-S, Pgdh, Pro-2, Sod); and P. iiehitlifer and P. maculatofasciatus were distinguished at six loci (Adh-2, Est-3, Gapdh-2, G6pdh, Pro-2, Pro-4). A number of loci were also identified that are capable of separating one species from each of the other two. For example, P. clatfrratus is fixed for alleles not found in P. nebulifer or P. maculatofasciatus at four gene loci (Ck-A, Icdh-S, Pgdh, Sod). Thus all three species can be distinguished using only two loci: one of the four for which P. clathratus has unique alleles, and one of the six that separates P. nebulifer and P. ■maculato- fasciatus. A similar procedure can be used with the two loci (Adh-2, G(ipdh) that distinguish P. nebuUfer from the other two species or the two loci (Est-3, Pro-2) for which P. maculatofasciatus has unique alleles. No single locus was identified that by itself completely distinguishes all three Paralabrax species. The closest to a completely diagnostic locus is Ada; at this locus, a different allele predominates in each of the species at a frequency of 0.80 or more. Considerable mtDNA sequence differentiation was demonstrated between the three Paralabrax species (Table 4). Six of the restriction endonucleases (Bglll. Clal, Kpnl, Sail, Smal, and A'6al) failed to cleave the circular mtDNA molecule more than once in each of three species and were therefore uninformative. Of the 13 restriction endonucleases which produced two oi' more fragments in the Paralabrax species, eight {Am I, Avail, BstEll. Hindi, Hindlll, Hinjl. Hpall, and Xhol) were able to distinguish all three species, while three enzymes (Bam HI, EcoRl, and Sstl) were able to distinguish one of the species from the other two, and two enzymes (Pstl and P(mll) produced similai- mtDNA fragments in all three species. Thus 8 of the 13 informative restriction endonucleases were diag- nostic for all three species. The mean nucleotide sequence divergences between the three Paralabrax species, in pairwise comparisons, are presented in Table 3. Because digestions with Hinfl, Hpall, and Avail (restriction endonucleases which recognize four or five nucleotide base pairs) pro- duced so many fragments, it was not possible to assure homology of similarly migrating fragments, and the results from these enzymes were not used in the cal- culation of interspecific divergences. Little intraspecific mtDNA nucleotide sequence vari- ation was detected in this study. Of the 18 fish analyzed with the 13 informative restriction endonucleases, only two variants were encountered, each in a single in- dividual (Table 4). The species-specific fragment patterns produced from the digestion of Paralabrax mtDNA with the restriction endonuclease Hivdlll (Fig. 1) were con- sidered the most practical to use for specific identifica- tion because there were diagnostic differences in the Table 4 Restriction fragment sizes (in lase pairs) in three species | oi Paralabrax Not all fragments were resolved in digestions | with /Ira II. Restriction Paralabrax endonuclease species Fragment sizes Ara\ flathratus 7.4, 4.1, 2.1, 1.9 ma rii la tofascia t us 11.9, 2.6, 2.3 >i rt III lifer 9.2, 3.6, 2.7, 1.3 AirnU rintliral lis 4.9, 2.7, 1.8, 1.2, 1.1. 0.7, 0.7, 0.5 iiiitriilittiiJasciatHS 2.6, 2.2, 1.7, 1.5, 1.5. 1.5, 1.3, 1.1, 1.0 nebulifer 5.7, 1.9, 1.8. 1.5. 1.5. 1.3, 1.1, 1.0 BamUl clathratus 16.8 maculatofasciatus 10.0, 6.8 nebulifer 10.0, 6.8 BstEll clathratus 6.5, 5.3, 5.0 maculatofa scia tus 9.4, 7.4 nebulifer 14.3, 2.5 (5) or 10.1, 4.2, 2.5 (1) EcoRl clathratus 8.8, 7.9 ma culatofa sciat us 8.8, 7.9 nebulifer 7.9, 4.6, 2.4, 1.9 Hnirll clathrnttis 4.2, 2.8, 2.7, 2.4, 1.3, 1.2, 1.0, 0.7, 0.3 initciiliitufasciatus 4.6, 3.6, 3.0, 2.7, 1.5, 0.7, 0.6 nebulifer 4.0, 3.5, 3.0, 2.4, 1.6, 1.5, 0.7 Hnidin clathratus 6.2, 4.0, 2.2, 1.8, 0.7, 0.6 maculatofasciatus 6.2, 4.7, 2.2, 2.0, 1,8 nebulifer 6.2, 6.2, 2.2, 1.8, 0.4 Pstl clathratus 12.6, 4.2 maculatofiscintus 12.6, 4.2 (5) or 9.4, 4.2, 3.0(1) nebulifer 12.6. 4.2 Pm II clathratus 11.6, 5.2 maculatofa sciat us 11.6, 5.2 nebulifer 11.6, 5.2 Ss( I clathratus 14.0, 2.8 maculatofasciatus 16.8 nebulifer 16.8 A7).)I clathnaus 12.1, 2.9, 1.8 macutatafisciatus 16.8 nebulifer 9.8, 4.8, 2.2 larger (most easily visualized) bands. Fresh or frozen tissue samples as small as 0.01 g could be identified from a Hindlll digestion of the mtDNA enriched in a miniprep procedure and run on a 1.0% agarose minigel stained with ethidium bromide. It was not possible to identify tissue samples smaller than 0.01 g on ethidium bromide-stained minigels 64 Fishery Bulletin 88(1), 1990 Figure 1 Species-specific fragment patterns of nitDNA isolated from three species of P(iralahrn,r and digested with the restriction endonucJeaseZ/iHrflll. Note that the upper (larger) bands distinguish the three species. From left to right: P. ctathratu.-i (1). P. niaculatofasriatus (2). P. nebulifer (2), and the size standard (BRL 1 Kb ladder). The mtDNA restriction fragments were endlabeled with radioactive nucleotide triphosphates, separated on a 1.0% agarose gel and visualized by autoardiography. Figure 2 A Southern blot of //i»i/ III digests of genomic ONA isolated from individual Paralabrax eggs and probed with biotinylated Paralahrax mtDNA. The eggs can easily be identified on the basis of the frag- ment patterns by comparison with those in Figure 1. One prepara- tion (lane .5) did not contain sufficient DNA for identification. From left to right: P. maculatofasciatus (2), P. nebidifer (2), and P. cUi- because there was not sufficient mtDNA fcir visual- ization of the restriction fragments. However, samples as small as 0.6 mg(one Pdntlahrax egg) could be iden- tified after Southern transfer and hybridization with a biotinylated probe (Fig. 2). Although greater sensi- tivity could have been obtained with a radioactively labeled probe, the results obtained with the biotinylated probe were sufficient to identify an individual fresh or ethanol preserved egg, and avoided the cost and prob- lems of probe half-life and waste disposal associated with radioactively labeled probes. Graves et al ' Biochemical genetics of Paralabrax 65 Discussion The results of the electrophoretic and mtDN A analyses are similar in several respects. First, the absolute level of genetic differentiation between the Paralabrax species is relatively small and is consistent with values typically found between other closely related organisms (Avise and Aquadro 1982, Ferris and Berg 1987). Sec- ond, the comparison of P. maculatofasciatus and P. nebulifer yields the smallest genetic distance of the three pairwise comparisons. A major focus of the electrophoretic and mtDNA analyses was to find a biochemical character which would unambiguously differentiate the early-life- history stages of the three Paralabrax species found in southern California coastal waters. The electro- phoretic investigation revealed several apparently fixed allelic differences between any two of the three species and demonstrated that a screening of a minimum of two electrophoretic loci will result in a positive specific identification. This methodology can be used to iden- tify late larvae, juveniles, or adults in which there is sufficient tissue (or enzyme activity) to survey two or more loci. However, it was not always possible to score the diagnostic loci in Paralabrax eggs and larvae smaller than 10 mm total length. Although it may have been possible to improve the resolution of the electro- phoretic techniques, we chose to rely upon mtDNA restriction fragment differences to identify Paralabrax eggs and early larvae. The restriction endonuclease analysis of Paralabrax mtDNA, like the electrophoretic survey, demonstrated considerable genetic differentiation among the three species. Several restriction endonucleases were each capable of distinguishing the three species of basses. Furthermore, the mtDNA analysis not only worked with relatively large adult tissue samples, but with quantities as small as an individual egg. Consequent- ly, the mtDNA analysis is the method of choice for the identification of Paralabrax early-life-history stages. Several ecological and assessment studies involving Paralabrax eggs and early larvae have not been possi- ble because of an inability to specifically identify the early-life-history forms. The mtDNA isolation and analysis techniques presented in this paper will facili- tate such studies. Furthermore, because the methods work with ethanol-preserved as well as fresh speci- mens, there is no need to separate Paralabrax eggs and larvae from a fresh plankton tow while at sea. Acl), where t is age in years, and l, = total length at age. Mean asymptotic total length was calculated to be 490 mm. The oldest fish examined was age 8 and ,504 mm TL. Histological examina- tion indicated that fish matured at age 1 and spawned fi-om May through August. Feeding Habits, Age, Growtli, and Reproduction of Atlantic Spadefisli Chaetodipterus faber (Pisces: Epiiippidae) in Soutfi Carolina* John W. Hayse Grice Marine Biological Laboratory 215 Fort Johnson Rd . Charleston, South Carolina Present address: Department of Zoology, Miami University, Oxford, Ohio 45056 The Atlantic spadefish Chaetodipterus faher (Broussonet) is the only mem- ber of the family Ephippidae native to the western Atlantic Ocean. This species inhabits coastal waters from Chesapeake Bay to southeastern Brazil, including the Gulf of Mexico (Johnson 1978), and has also been in- troduced into the waters surrounding Bermuda (Burgess 1978). It is a com- mon fish in South Carolina, par- ticularly from early spring through late fall, and all life stages have been collected in South Carolina. Since 1978 the Recreational Fish- eries Section of the South Carolina Wildlife and Maiine Resources Depart- ment (SCWMRD) has promoted C. faber as a sportfish, primarily due to the attraction of large numbers of adult Atlantic spadefish to offshore artificial reefs that have been created and/or maintained by SCWMRD. Atlantic spadefish have traditionally been a relatively minor recreational species in South Carolina and were only occasionally caught by hook-and- line fishermen using shrimp or squid as bait. Observations of feeding be- havior by SCWMRD personnel in the early 1980s suggested that Atlantic spadefish might eat Stomolophus meleagris, the cannonball jellyfish. As a result, a technique using can- Manuscript accepted 11 August 1989. Fishery Bulletin, U.S. 88:67-83. •Contribution No. 86 of the Grice Marine Bio- logical Laboratory, and Contribution No. 274 of the South Carolina Marine Resources Re- search Institute. nonball jellyfish as bait was devel- oped and has proven extremely ef- fective in attracting and capturing C. faber (Moore et al. 1984). This method has since been the subject of some recreational fishing publica- tions (Ogle 1985, 1987). Little, how- ever, is known about the life history, feeding habits, age and growth, or reproductive biology of C. faher, even though such information is pre- liminary to proper assessment and management. The present study was imdertaken with three major objectives. The first was examination of stomach con- tents from specimens of C. faher in order to describe the diet. In addi- tion to a qualitative and quantitative diet analysis, I wished to determine if the mode of collecting fish (hook- and-line vs. net and spearfishing) might bias the results of diet analy- sis. Since Atlantic spadefish appeared to inhabit different areas depending upon body size and age, I also wished to deteiTnine if ontogenetic and habi- tat differences existed in the diets of the fish collected. The second por- tion of the study dealt with age and growth of Atlantic spadefish in South Carolina by finding a suitable ageing method and estimating growth rates. The final segment of the research explored the reproductive biology of C. faber, specifically determination of spawning period, sex ratios, and the age at sexual maturity of Atlan- tic spadefish off South Carolina. 67 68 Fishery Bulletin 1990 Materials and methods Data collection Specimens were collected July 1985 through May 1987, excluding the periods from December 1985-April 1986 and December 1986-April 1987 when Atlantic spade- fish were absent from collection sites (Fig. 1). Most of the specimens less than 200 mm total length were pro- vided by other research programs which obtained them as bycatch in trawl tows. Collections were from three different habitat strata (Fig. 1). Stratum 1 included estuarine habitats. The fish in this stratum were primarily young-of-the-year and were collected by a variety of gear, including dipnets, seines, stopnets. and trawl nets. Stratum 2 consisted of shallow nearshore habitats less than 20 m in depth with bottoms consisting primarily of mud or sand and shell, with rare patches of live bottom (sponges and soft corals, low rock relief). The fish in stratum 2 were all collected by trawling. Stratum 3 contained fish col- lected primarily by spearfishing while SCUBA diving on artificial reefs located off South Carolina's coast and on the Murrell's Inlet jetties. Some specimens from the artificial reefs were collected by hook-and-line using Stomolophus meleagris as bait. The artificial reefs and the jetties were similar habitats in that all were man- made structures providing relatively high relief and all were covered with a variety of fouling organisms such as algae, sponges, corals, hydroids, sea anemones, bryo- zoans, and ascidians. The fish from stratum 3 were all adults. Fish (n = 832), subsampled from 84 collections, were sexed, weighed to the nearest gram, and total length (TL) and standard length (SL) measured to the nearest millimeter. Subsamples, consisting of 177 stomachs and 233 gonads, were excised from a representative por- tion of the collections, preserved in buffered 10% seawater formalin, and later transferred to 50% iso- propanol. Otoliths (sagittae) were removed from 643 specimens and stored in 95%i ethanol. Stomach contents Stomachs of Chaetodipteriis faber are well-defined structures, and contents of the digestive tracts between the pharynx and the pyloric sphincter were examined using compound and stereoscopic microscopes. Stom- achs were scored for fullness and food items were sorted, identified to the lowest practical taxa, and counted. Colonial organisms, such as hydroids and bryozoans, and multiple fragments of individual species were counted as single individuals unless numbers could be estimated from the fragments by counting structures such as eyes or legs. After removing excess water by blotting on absorbent paper, the volume of -vjLillle Rii • STRATUM 1 ■ STRATUM 2 • STRATUM 3 ^\^^^ Charleston Harbor l^j-^* SI- Helena Sound ""^^■TZ::^^::^* ''<"' Royal Sound ' -y'A ■'' Cal.boque ^' • _ Sound Figure 1 Positions of collection sites in South Carolina for Oiiuiodipteriif. fnber. Symbols indicate general collection areas, and some indicate more than one collection in a particular area, ("ircles (•) represent estuarine sites (stratum 1). .s<)uares (■) represent shallow nearshore sites with sandy bottoms (.stratum 2). and stars (•) represent ar- tificial reef habitats (stratum 3). Hayse: Feeding, age, growth, and reproduction of Chaetodipterus faber 69 each taxon was determined by displacement, except in the case of small items. Volumes of small items were estimated using a 1 x 1 mm grid. Percent frequency of occurrence (%F), percent of total number (%A^), and percent of total volume {%V) of stomach contents were calculated by higher taxonomic categories for the en- tire data set and for the data separated by length in- tervals and collection strata, and these values were used to calculate an index of relative importance {IRI) (Pinkas et al. 1971): IRI = (%N + %V) X %F. Shortly after laboratory analysis of stomach contents commenced, it became apparent that most of the prey items observed were either colonial or fragments of soft-bodied organisms that could only be counted as one individual per stomach each time they were encoun- tered. The primary countable organisms were amphi- pods and copepods, which were present in large num- bers, but accounted for only a small portion of the volume. In the opinion of some researchers (Lagler 1956, Crow 1982), percent number can lead to a some- what distorted view if small organisms are present in large numbers even though they are of minor impor- tance to the diet of a fish. Consequently, a modified index (MI) of relative importance which did not incor- porate %N into the formula was also calculated for each taxonomic category: MI = %F X %V. Stomach contents from 27 fish taken by hook-and-line were analyzed separately in the above manner and com- pared with the data from fish collected by nets and spearfishing. Age and growth Sagittae were chosen as structures for age determina- tion since they were easily accessible, easily stored, and exhibited well-defined growth zones. Other structures were deemed unsuitable as ageing structures for vari- ous reasons. Scales were small, frequently regenerated, and the marks on them were considerably more dif- ficult to interpret than those on the sagittae. Vertebrae were difficult to obtain and time-consuming to prepare for analysis. No differences were observed between the left and right sagittae of individual fish and, unless damaged or unavailable, the left sagitta was used in age analysis. Cleaned whole sagittae were examined using both reflected and transmitted light, as necessary, in order to establish the location of presumed annuli. Whole otoliths were used since 100% agreement was obtained between counts of opaque zones from a subsample {n =18) of otolith sections and counts from the cor- responding whole otoliths. Sagittae were placed, concave-side-up, in a dish containing cedar oil and view- ed with reflected light against a dark background at 25 X using a stereoscope equipped with an ocular micro- meter (1 ocular unit = 0.04 mm). The distance (in ocular units) from the kernel to each opaque mark and from the outermost mark to the edge of the sagitta (marginal increment) was determined from measurements made along the ventral radius of each sagitta (Fig. 2). These measurements were performed on two occasions sep- arated in time by at least 2 weeks, and the two sets of measurements were compared to see if otolith mea- surements were repeatable. Both least-squares linear regressions (Sokal and Rohlf 1981) and geometric mean (GM) functional regressions (Ricker 1973, Sokal and Rohlf 1981) were used to describe the relationship of TL to ventral otolith radius (OR), SL to TL, and lengths to weight. Mean back-calculated TL at age was derived by the Fraser- Lee method using the intercept of the OR-TL relation- ship (Poole 1961, Carlander 1982). Mean back-calcu- lated TLs were weighted by the reciprocal of the standard error of the mean squared, and the von Ber- talanffy growth equation (Bertalanffy 1938) was then fitted to mean back-calculated TL at age by using the SAS NLIN procedure while employing Marquardt's algorithm and the SAS NLIN weight statement (Helwig and Council 1979). Reproduction Reproductive tissues were processed in an Auto- Technicon 2A Tissue Processor, vacuum infiltrated, blocked in paraffin, and sectioned (6-10 i^m thick) on a rotary microtome. Sections were stained with Har- ris hematoxylin and counter-stained with eosin-Y. Mounted sections were then examined with a com- pound microscope, and sex and maturity stages were assigned according to the criteria of Waltz et al. (1979). The ratio of males to females and maturity stages by month were examined to determine sex ratios and spawning period of C. faber off South Carolina. Results Analysis of stomach contents Eighty percent of the stomachs obtained by hook-and- line contained food, and at least seven prey species Reference to trade name.s does not imply endorsement by the National Marine Fisheries Service, NOAA. 70 Fishery Bulletin 88(1), 1990 1 mm ^■ \S> 1 mm f^ >MARG Figure 2 Measurements taken from each sagitta of Chiietiidiiilfnisfiibi'r used in age analysis. (1) Photograph of the left sagitta from a young-of-the-year (115 mm TL) specimen. (2) Photograph of a whole left sagitta from an age-6 (396 mm TL) specimen. Orientation of the sagittae within the neurocranium is indicated: A anterior; P posterior; D dorsal; V ventral. A1-A6 represent successive measurements to each annulus; OR is the ventral sagitta radial measurement; MARG is the marginal increment; K is the kernel of the sagitta. were present (Table 1). Stomolophufi mdeagriii was, by far, the dominant organism according to all numeric indicators (%F, %N, %V, IRI, and MI). Hydroids, epi- faunal amphipods, and anthozoans (sea anemones) were observed in considerably lower volumes, numbers, and frequencies than S. meleagris. Ninety percent of the stomachs from Atlantic spade- fish obtained by other collection methods (nets and spearing) contained tood, and at least 75 prey species were represented (see Appendix). In sharp contrast to the hook-and-line samples, jellyfish occurred in only one stomach collected by nets or spearing even though S. meleagris and other jellyfish were common in the artificial reef areas during many of the spearfishing ex- peditions and were abundant in trawl catches. Further- more, the species of jellyfish eaten was identified, by Hayse Feeding, age, growth, and reproduction of Chsetodipterus faber 71 Table 1 Percent frequency of occurrence (%F). per cent of total number (%N), percent of total volu me{%V), index of relative importance (IRI), and modified index of impor tance (A//) for food items of Chaetodipterus faber obtained by hook-and-l ne in South Carolina using Stomolophus meleagris as bait. Prey item %F %N %V IRI MI Cnidaria Hydrozoa Endendriiini sp. 18.18 9.52 0.69 185.62 12.54 Olu'lin sp. 4.55 2.38 0.01 10.88 0.05 Hydrozoa undetermined 4.55 2.38 0.04 11.03 0.18 Total Hydrozoa 27.27 14.29 0.74 409.67 20.17 Anthozoa Actiniaria 4.55 2.38 0.55 13.33 2.50 Sc_yi:ihozoa Stonioloj)h us mrlFngris 95.45 50.00 98.66 14.189.60 9417.10 Crustacea Amphipoda CiijircHa penantis 13.64 19 05 0.03 260.25 0.41 Erichthon ius 6»y/.>;i7iVh.>.7s 9.09 9.52 0.01 86.63 0.09 Jagsa faJcata 4.55 4.76 <0.01 21.67 0.01 Total Amphipoda 27.27 33.33 0.04 910.12 1.09 Number of stomachs examined: 27 Examined stomachs with food: O'? 60n 40- 20- 20- ° 40- 60- 80 A C B 79% 17% 29% ^il^°/°F TAXON 46% IRI Ml A Hydrozoa 2464 2105 B Anthozoa 813 533 C Polychaeta 436 421 D Amphipoda 3647 163 E Porifera 88 85 F Algae 40 30 examining tentacle arrangement and nematocyst mor- phology, as Chiropsalmus quadrumanus, a cubomedu- sa known as the sea wasp. In stomachs obtained by nets and spearing, hydroids were dominant by volume (27%) and were the most frequently observed organisms (seen in about 80% of the stomachs examined) (Fig. 3). Am- phipods were in 45% of the stomachs and accounted for about 76% of the individual organisms, but had a low volume (4%). Nearly all of the amphipods were epi- faunal, species primarily caprellids and tubicolous gam- marids, that are frequently associated with hydroids. Anthozoans, consisting nearly equally of anemones (Acti- niaria) and sea pansies Ren ilia reniformis were pres- ent in 29% of the stomachs, accounted for 10% of the individuals, and were the third-ranking taxon in terms of percent volume. Polychaetes were seen in 17% of the stomachs and accounted for 24% of total volume, but comprised less than 1% of the number of individuals. This category consisted primarily of terebellid feeding tentacles, which were probably cropped by C. faber Figure 3 Percent frequency (%F), percent number (%A'). percent volume (%V). index of relative importance (IRI). and modified index of im- portance {MI) for higher taxonomic groups of food in the diet of Chaetddiptenisjhher collected in South Carolina waters by nets and spearing. 72 Fishery Bulletin 88(1), 1990 > Z > ^ > ^ 80 70 40 20 ri 20 40 50 Stratum 1 (n=49) TAXON C D 22% 10% 88% %F A B C D E F G H 29% : : 1 rr 31% 60-| B 40- A Stratum 2 A (rr=55) 20- C D E G 2% 4h,^,J% 64% 45% 20- 30* , ^ %F 50 60 30-1 20 1 t 64% Stratum 3 (n=51) B C 88% 10% 16% 20 80 100 J H 4% 2 7% 1 8% 2% %F 63'! JHL Hydrozoa Anthozoa Algae Amphlpoda Copepoda Polychaeta Porlfera Scyphozoa 8251 222 164 643 32 _ML 6875 122 77 0 0 0 66 14 0 0 0 JHi ML A 2142 1539 B 7328 3610 C 1 1 D 750 15 E 8 1 F 426 339 IRI Ml A B C D E F G H 2454 2224 69 57 6002 1 802 332 26 48 49 297 1 786 326 26 Figure 4 Percent frequency (%F). percent number (%N), percent volume (%V), index of relative importance {IRI), and modified index of im- portance (MI) for higher taxonomic groups of food in the diet of ChaetoHiiiti'rusJhhrr col- lected in South Carolina waters, by strata. Stratum 1 is composed of inshore estuarine habitats; stratum 2 encompasses shallow (<20 m), nearshore habitats with sandy bot- toms; stratum 3 includes offshore artificial reef areas and the Murrells Inlet jetties. while these polychaetes were feeding. Since heads or mouthparts of the polychaetes were usually absent, there were often no countable body parts, making it impossible to determine the number of individuals eaten. I had similar problems in enumerating hydroids, sponges, algae, bryozoans, and occasionally sea ane- mones and sea pansies. The IRI ranked amphipods as the dominant prey, followed by hydroids, anthozoans, |)olychaetes, and sponges. The MI ranked hydroids as the most important taxon, followed by anthozoans, polychaetes, amphipods, and sponges. Hydroids were the dominant food item for fish from stratum 1, as shown by both the IRI and the MI (Fig. 4). Less important were anthozoans, algae, amphipods Hayse: Feeding, age. growth, and reproduction of Chaetodipterus faber 73 > ^ > Z > z 60 40 20 20 40-" 50 40 20 20- 40v 80 90 B 1-100 mm SL (n=80) 78% TAXQN IHL _ML 9% 4% F ■o— '15% %F A Hydrozoa 2874 1872 B Anthozoa 2818 1861 C Polychaeta 68 53 D Amphipoda 840 39 E Eggs 25 23 F Algae 45 12 G Porlfera 0 0 H Scyphozoa 0 0 39% 31% B 78% 101-200 mm SL (n=40) S D 18% 45% F 10% ■%F A B C D F IRI 2404 2330 230 4484 45 Ml 2234 1890 222 455 42 u 48% 4U- A c 20- 37% 86% 20: > 200 mm SL (n=35) D F -B 1=1. H JHL ML 1 V 26% 3% ■%F 2602 1239 A C D 5384 F G H 30 556 42 2246 1193 113 23 535 42 80^ 90^ 60% Figure 5 Percent frequency (%F), percent number (%A'). percent volume (%V), index of rela- tive importance (IRI), and modified index of importance (MI) for higher taxonomic groups of food in the diet of Chaetodipterus faber collected in South Carolina waters, by standard length (SL) intervals. and copepods. In stratum 2, the relative-importance in- dices indicated anthozoans were tlie major component of the diet, followed by hydroids. Anthozoans in stomachs of fish from stratum 2 consisted mainly of Renilla 7'eniforwis. Jellyfish (Scyphozoa) were never seen in stomachs of fish from stratum 1 or stratum 2. The IRI indicated that amphipods were the most impor- tant prey in stratum 3, while the MI ranked hydroids as the dominant food item. Amphipods composed 91% of the individuals and were found in 63% of the stom- achs containing food, but accounted for only 5% of total volume of food in stomachs from stratum 3. Hydroids, on the other hand, were observed in 88% of the stomachs from stratum 3 which contained food and 74 Fishery Bulletin 88(1), 1990 100 150 200 250 300 350 400 450 500 TOTAL LENGTH (mm) Figure 6 I'lTcent frequency of* 'liiiihidi/ilcrux Jdhrr witliin Kl-mni TL size classes. composed 25% of total volume in stratum 3. Percent number was low for hydroids in stratum 3, but they could only be counted as one individual per stomach each time they were encountered, whereas there were usually many amphipods per stomach. The IRI ranked hydroids as the second-most important food source, followed by polychaetes (feeding tentacles) and sponges. TheMZ ranked polychaetes second, followed by sponges and amphipods. Although Scyphozoa ranked as one of the top seven food items in stratum 3, it occurred in only one stomach, which was distended with 16 mL (after blotting) of jellyfish. Most stomachs from stratum 3 that were considered full contained 3-6 mL of food. The MI indicated that hydroids were the dominant food for all size classes oiC.faber. being found in over 75% of the stomachs from each size class and accoun- ting for approximately 25% of the volume in each of the groups (Fig. 5). The MI also showed that antho- zoans were nearly as important as hydroids for fish up to 200 mm SL, but were not important to fish >2()0 mm SL. The second-ranking prey for fish >200 mm SL was polychaete-feeding tentacles according to the MI. The IRI denoted amphipods as a major source of food for fish of all sizes (dominant, by far, for fish > 100 mm SL), but this was certainly due to the numerical bias of the IRI. The MI indicated amphipods were a minor dietary component except for fish 100-200 mm SL, where they were the third-most important prey. Un- identified egg masses, probably gastropod eggs, were found in 4% of the stomachs from fish up to 100 mm, but were a relatively minor component of the diet from fish this size. Sponges were observed in 26% of the stomachs from fish >200 mm SL and accounted for 21% of total volume of prey for these fish, making it the third-ranked food item for fisli in this grouji accord- ing to the MI. Age and growth All but 6 of the 643 pairs of otoliths obtained were used in age analysis, since 6 pairs were deemed unreadal)le due to deformities in otolith structure. Discernment of check marks on whole otoliths oiCfaber was relatively easy in either reflected or transmitted light, and there was 93% agreement between the two series of counts and measurements. The differences between the re- maining 7% (attributable to measurement and record- ing errors) were satisfactorily resolved, allowing all 637 otolith pairs to be utilized in ageing analysis. The pat- tern of alternating opaque and translucent l)ands which were observed on the sagittae were believed to be permanent records of the physiological growth of the Atlantic spadefish from which they had been obtained. Length-frequency distributions are often used to separate fish of different lengths into age groups and are also utilized to validate ageing techniques, particu- larly for younger age groups (Lagler 1956, MacDonald 1987). Good agreement was seen between the TLs of fish determined to be age 0 and age 1 by counting opaque zones on otoliths and the range of TLs for the first and second groups in a length-frequency distribu- tion (Fig. 6). Consequently, the innermost opaque zone on each sagitta was believed to be the first annulus. Since marginal increments on the otoliths should approach zero during the time of annulus formation, the monthly means were calculated to determine if one opaque band was laid down on the sagittae during each year. Generally, the mean marginal increment was low- est in May and ijicreased steadily, by month, through Hayse Feeding, age. growth, and reproduction of Chaetodipterus faber 75 16-, 17 23 15- ^ ,4. ^ 13- S ., 53 UJ 12- CC ,1. 39 68 II f 1 —si 1 ■ *— ' CO ,3 1 z -I-io- 1 ■ ^§9- J t Z to Dcge. 1 1 \ \ \\\ 1 2 5. If 4- |l Z Jl < 3- -|- ^ 2 2' 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oc t Nov Dec MONTH Figure 7 Mean marginal increments, in ocular units, on the otoliths of Chnetndipterus faber by month of cap- ture in South Carolina. Heavy horizontal bars repre- sent mean values, heavy vertical bars represent 95% confidence intervals, and vertical lines represent ranges of measurements. Numbers over each series of values are sample sizes. Table 2 Sample sizes M. means {x). and stan( ard deviations (SD) for observed lengths mm), and weight (g) by age for Chaetodiptenis faber. Age Total length Standard length n (x) SD Weight w (5) SD w (X) SD 0 421 84 24 424 67 20 424 30 21 1 121 163 21 121 133 17 121 185 74 O 36 265 29 36 216 25 36 670 202 3 21 309 43 21 250 36 21 1063 469 4 10 396 61 ](i 332 53 10 2107 810 .5 17 440 27 17 367 24 17 3025 .586 6 7 439 34 7 369 30 7 3087 867 7 0 — — 0 — — 0 — — 8 1 472 — 1 405 — 1 — — November (Fig. 7). Although a marginal increment of zero was never observed, there was a strong indica- tion that the opaque zones on the sagittae formed once a year between December and April, and, therefore, justified their use as annular marks. However, these were months when Atlantic spadefish were absent from collection sites and, consequently, no specimens were obtained. The oldest fish was age 8, although no age-7 fish were collected. Observed mean lengths (TL and SL) and weight increased with age (Table 2), and ail least- squares regression relationships were very highly significant (P<0.001) (Table 3). Analysis of the age composition of Atlantic spadefish by stratum showed that all fish taken in stratum 1 were young-of-the-year (age 0) and the fish from stratum 2 were predominantly age 1, although age-0 fish were seen in stratum 2 in increasing numbers from late summer to late fall. All fish taken on the artificial reefs (stratum 3) were age 2 or older. Mean back-calculated TLs at age were lower than mean observed TLs for all age groups Mean back- calculated lengths at preceding annuli, weighted mean back-calculated lengths, and growth increments for all fish age 1 and older are in Table 4. Growth rates dif- fered little between sexes. The von Bertalanffy growth equation, derived from back-calculated mean total lengths at age was: I, = 490(1 - e-o.34(/ + o.i8))_ where t is age in years and /, is total length at age (Table 5). There was good agreement between the mean observed, mean back-calculated, and theoretical 76 Fishery Bulletin 88(1), 1990 Table 3 Least-squares linear and geometric mean (GM) regression equations of weight (WT) on total length (TL) and standard length (SL), length-length, and TL-ventral otolith radius (OR) for Chaetodi ptcrun fiibcr from South Carolina waters. Weight unit in grams, SL and TL units in millimeters, and OR expressed as ocular units (1 ocular unit = 0.04 mm). All least-squares linear regressions were significant atP0.05) deviations from lailQ ratios for C. faber of all ages combined or for fish grouped by ages 0-6. Histological examination of reproductive organs revealed that 64% of age-0 males were mature and all males age 1 and older were mature. The smallest spent male was age 0 and 94 mm TL. Ovaries of all age-0 females were immature, while all females age 1 and older were mature. The smallest sj)ent female was 120 mm TL. Analysis of maturity stages by month indicated spawning activity for C. faber lasted from May through September, and the greatest percentage (97%) of devel- oping and ripe gonads was observed in May (Fig. 9). P'roni June to October the percentage of developing and ripe gonads decreased steadily to 0%. All reproduc- tive organs from fish captured in October and Novem- ber were resting, and no developing gonads were collected after August. Collapsed follicles in various Hayse: Feeding, age, growth, and reproduction of Chaetodipterus faber 11 500' 1 .--^ ^ 400 E / .'" E / '''' ^^ / ■' X / / H // C3 300. yk 2 LU / .' • • OBSERVED _l f •' • • BACK-CALCULATED / / ■---■ THEORETICAL _l / / < o =^°°' //' 1- / ; z ^ / < m ^ 2 100 12 3 4 5 6 7 8 AGE (years) 9 Figure 8 Mean observed, mean back-caicuhited. and theoretical (von Berta- lanffy) total lengths (mm) at age for both sexes of Chaetodipter'us faber from South Carolina waters. states of atresia were observed in some histological sections of ovaries which also contained developing oocytes, indicating that some serial spawning took place during the period May- August. Discussion Analysis of stomach contents Dietary analysis indicated that Chaetodipterus faber in South Carolina eat mostly hydroids, anthozoans such as sea anemones and sea pansies, and polychaete tenta- cles. Pearce and Stillway (1976) reported the presence of an unusual fatty acid in the liver of C. faber. Since coelenterates appear to be the source of this fatty acid in certain fishes and marine turtles, the authors sug- gested that the elevated levels in Atlantic spadefish could be a reflection of a high dietary intake of coelen- terates. Gallaway et al. (1981) believed that the hy- droids and other fouling organisms they observed in the stomachs of Atlantic spadefish were being taken from the water column after being sloughed off the understructure of oil platforms. My underwater obser- vations while SCUBA diving provided no evidence that Atlantic spadefish are primarily planktivores as sug- gested by Gallaway et al. (1981) and I never observed anything to suggest that the ingested hydroids were 17 22 1U0- 35 ■ Developing of Ripe o LU 90- n Spent or Resting 20 z 80- 5 8 < 70 • X ^^M LU BO- ^M '' 16 16 tn H ^ R a bU- 1 H-i. ^r- < ■ H^ ■ 2 41)- 1^ ■ ■ o H-i B ■ o 30- B ■ ■ li. ■ ■ ■ o ■M- ■ ■ ■ 3 3? 10- 2 1 1 1 0 0 0 MAY JUN JUL AUG SEPT OCT MONTH OF CAPTURE NOV Figure 9 Maturity stages of age-1 and older (_'h1200 individuals in one stomach) found in three stomachs from a single collection at the Murrell's Inlet jetties; a large volume of hydroids was also present in these stomachs. Overall, major changes in food habits within the size range examined were not apparent; even the smallest fish examined (19 mm SL) had eaten primarily hydroids. Differences that existed probably reflected changes in habitat with increasing size. No information is currently availal)le about food habits of larval C. faber. Age and growth Some information is available about early development of C. faber (Ryder 1887, Hildebrand and Cable 1938, Johnson 1978); however this study presents the only known information of this species' later growth with age. Although the location of the first annular mark on the otoliths has not been thoroughly validated, the size of fish with only one opaque zone on the sagittae corresponded well with the lengths of fish in the sec- ond group of a size-frequency distribution (Fig. 6). This also agreed with Hildebrand and Cable (1938), who be- lieved Atlantic spadefish in North Carolina attained a length of 55-100 mm by the end of the first summer and a length of 135 mm (type of measurement unspe- cified; probably TL) during their second summer. Con- sequently, I am confident that the first opaque zone is the first annulus. Since young-of-the-year Atlantic spadefish are easily obtained from South Carolina estuaries during the summer and they can survive relatively well in captivity, more thorough validation of the first annulus, as well as additional yearly marks, may be obtained through daily growth studies and marking of otoliths with a chemical such as tetracycline (Beamish and McFarlane 1987). Marginal increment analysis supported the hypothe- sis that the opaque zones on sagittae are annuli since they appeared to be formed once a year (over the winter months), although fish were available at collec- tion sites only between May and December. In addi- tion, the data showed (1) a strong relationship l)etween otolith radius and length of the fish; (2) a decrease in growth rate in length with age (except where obscured by small sample sizes); and (3) close agreement between back-calculated lengths and observed lengths at age. The mean asymptotic TL (L^ = 490 mm TL) was reasonable for South Carolina's Atlantic spadefish population, but small for tropical regions where Atlan- tic spadefish with lengths up to 900 mm (presumably TL) have been reported (Breder 1948, Johnson 1978). Hayse Feeding, age, growth, and reproduction of Chaetoclipterus faber 79 Reproduction Few studies have dealt with the reproductive biology of C. faber. Hildebrand and Cable (1938) observed fully mature ripening females at around 135 mm, a size thought to have been attained during the second year (age 1); similarly, I found that all females age 1 and older were mature. Spawning of C. faber off South Carolina occurred May to October, and some females spawned more than once (serially) during this period. Serial spawning may allow fishes to produce a greater number of eggs than would be possible if they spawned only once during the year (DeMartini and Fountain 1981). The largest percentage of the fish were prepared to spawn in May, and histological evidence indicated that spawning continued periodically until August. Her- rema et al. (1985) found Atlantic spadefish in spawn- ing condition March through June along the east coast of Florida; however, few Atlantic spadefish were col- lected or examined. Chapman (1978) observed spawn- ing aggregations of C. faber during late July off the coast of Georgia. No fecundity estimates for C. faber exist, but researchers investigating this would probably wish to use batch fecundity methods which utilize ova diameters to establish groups of eggs that will develop and be spawned at different times (DeMartini and Fountain 1981, Hunter et al. 1985). The complete life history of the Atlantic spadefish is not well known and warrants further investigation. This study indicated that Atlantic spadefish in South Carolina waters spawn offshore late spring through early fall, with the juveniles subsequently moving into and inhabiting estuarine areas. In the fall, young-of- the-year Atlantic spadefish move into shallow offshore areas, which are also inhabited by age 1 individuals. Historical trawl data from SCWMRD indicated that C. /a6er juveniles were rarely taken in trawls off the coast of South Carolina during winter months, although they were taken in increasing numbers when proceeding southward toward Cape Canaveral, Florida (C.A. Wen- ner. Mar. Resour. Res. Inst., SCWMRD, Charleston, SC 29412, pers. commun., March 1987). In addition, analysis of trawl data from 1973 to 1980, by depth, showed that during the summer young (TLs corre- sponded to age 0 and age 1) Atlantic spadefish off the coasts of Georgia and Florida were located in depths of 20 m or less, while during the winter the majority of the catches were from depths of 28-56 m (G.R. Sedberry, Mar. Resour. Res. Inst., SCWMRD, Charles- ton, SC 29412, pers. commun., March 1987). This sug- gests that the age 0-age 1 group may move southward and into deeper water as the nearshore water cools. Age-2 and older Atlantic spadefish, which are common on South Carolina's artificial reefs and in high-relief live-bottom areas during the summer, are apparently rare in these areas during the winter. Practically nothing is known of the whereabouts of South Carolina's Atlantic spadefish during the winter, although a commercial trawler reported capturing several thousand C. faber in a single trawl during the winter 50 km off South Carolina's coast (Ogle 1987). Perhaps the fish from artificial reef areas move into deeper water during the winter, returning to these structures the following spring as water temperatures increase along the coast. Presently, there is an effort by SCWMRD person- nel to promote tagging of Atlantic spadefish by recrea- tional fishermen (D.L. Hammond, SCWMRD, Charles- ton, SC 29412, pers. commun.. May 1987). Many more fish will probably have to be tagged before adequate returns are seen, but tagging could lead to an increased understanding of C. faber movements, allow population estimates to be made, and, if used in conjunction with tetracycline marking of otoliths, could lead to thorough validation of the ageing technique presented here. Acknowledgments Financial support for this project was provided by the Sport Fishery Research Foundation and the Slocum- Lunz Foundation. Additional data and many specimens were provided by the MARMAP contract between SCWMRD and NMFS and the Inshore Recreational Fish research progi-am at SCWMRD. C.K. Biernbaum, D. Knott, C. O'Rourke, and P. Wendt helped with iden- tification of stomach contents. I especially thank W.A. Roumillat for help in analyzing histological sections of reproductive tissues. Citations Barnes, R.D. 1980 Invertebrate zoology, 4th ed. Saunders College/Holt, Rinehart, and Winston, Philadelphia. 1089 p. Beamish, R.J., and G.A. McFarlane 1987 Current trends in age determination methodology. In Summerfelt. R.C., and G.E. Hall (eds.), Age and growth offish, p. 15-42. Iowa State Univ. Press. Ames. Breder, CM., Jr. 1948 Field book of marine fishes of the Atlantic coast from Lalirador to Texas. G.P. Putnam's Sons, NY, 332 p. Burgess, W.E. 1978 Ephippidae. In Fisclier, W. (ed.), FAO species identifica- tion sheets for fishery purposes. Western central Atlantic, vol. 4. FAO. Rome. Calder. D.R. 1976 The zonation of hydroids along salinity gradients in South Carolina estuaries. In Mackie. G.O. (ed.), Coelenterate ecologj' and behavior, p. 16.5-174. Plenum Press, NY. 80 Fishery Bulletin 88(1), 1990 Carlander, K.D. 1982 Standard intercepts for calculating lengths from scale measurements for some centrarchid and percid fishes. Trans. Am. Fish. Soc. 111:332-336. Chapman, R.W. 1978 Observations of spawning behavior in Atlantic spadefish. Chaetodipterus faher. Copeia 1978:336. Crow, M.E. 1982 Some statistical techniques for analyzing the stomach con- tents offish. In Caillet, G.M., and CM. Simenstad (eds.). Gut- shop '81: Fish food habits studies; Proceedings of the third Pacific workshop, p. 8-15. Washington Sea Grant Prog., Univ. Wash., Seattle. 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. Gallavvay, B.J., L.R. Martin, R.L. Howard, G.S. Roland, and G.D. Dennis 1981 Effects on artificial reef and demersal fish and macro- crustacean communities. In Middleitch, B.S. (ed.). Environ- mental effects of offshore oil production, p. 237-299. Plenum Publ. Corp., NY. Helwig, J.T., and K.A. Council (editors) 1979 SAS user's guide. 1979 ed. SAS Inst., Inc.. Gary, NC, 294 p. Herrema, D.J., B.D. Peery, and N. Williams-Walls 198.5 Spawning periods of ccimmon inshore fishes on the Florida east coast. Northeast Gulf Sci. 7:153-155. Hildebrand, S.F., and L.E. Cable 1938 Further notes on the development and life history of some teleosts at Beaufort. N.C. Bull. U.S. Bur. Fish. 24:505-642. Hildebrand, S.F., and W.C. Schroeder 1928 Fishes of Chesapeake Bay. Bull. U.S. Bur. Fish. 48: 1-.366. Hunter, J.R., N.C.H. Lo, and R.J.H. Leong 1985 Batch fecundity in multiple spawning fishes. In Lasker, R. (ed.), An egg production method for estimating spawning biomass of pelagic fish: Application to the northern anchovy, Engraulis mordax. p. 67-77. NOAA Tech. Rep. NMFS 36, Natl. Oceanic Atmos. Adm., Natl. Mar. Fish. Serv. Johnson, G.D. 1978 Development of fishes of the mid-Atlantic Bight, vol. IV. Carangidae through Ephippidae. U.S. Fish Wildl. Serv. Biol. Serv. Prog.. 314 p. Lagler, K.F. 1956 Freshwater fishery biology. Wm. C. Brown Co., Dubu(|ue. lA, 421 p. MacDonald, P.D.M. 1987 Analysis of length-frequency distributions. In Summer- felt, R.C., and G.E. Hall (eds.). Age and growth of fish, p. 371-384. Iowa State I'niv. Press. Ames. Moore, C.J., M. Bell, and D. Hammond 1984 Atlantic spadefish. Saltwater Conversation 1984:19-22. S.C. Mar. Resour. Cent.. Charleston. SC 29412. Ogle, T. 1985 Fishing in spades. Saltwater Sportsman 46(4):76-79. 1987 Deal me spades. Southern Outdoors' Saltwater Maga- zine I(5):72-73. Pearce, RE., and L.W. Stillway 1976 rn;H.s-6-hexadecanoic acid in the spadefish Chnctodip- terus faher. Lipids 11:248-249. Pinka.s, L., M.S. Oliphant, and I.L.K. Iver.son 1971 Food habits of albacore. bluefin tuna, and horiito in Cali- fornia waters. Calif. Dep. Fish Game Fish Bull. 152. 105 p. Poole, J.C. 1961 Age and growth of the fluke in Great South Bay and their significance in the sport fishery. N.Y. Fish Game J. 8:1-11. Randall, J.E., and W.D. Hartman 1968 Sponge-feeding fishes of the West Indies. Mar. Biol. 1:216-225. Ricker, W.E. 1973 Linear regression in fishery research. J. Fish. Res. Board Can. 11:559-623, Ryder, J. A. 1887 On the development of osseous fishes, including marine and freshwater forms. U.S. Fish. Comm. Rep. 1885:488-604. Shealy, M.H., Jr., B.B. Boothe, Jr., and CM. Bearden 1975 A survey of the benthic macrofauna of Fripp Inlet and Hunting Island, South Carolina, prior to beach nourishment. Tech. Rep. 7, S.C. Mar. Resour. Cent., Charleston. SC 29412, 30 p. .Sokal, R.R., and F.J. Rohlf 1981 Bii.nietry, 2n(l ed. W.H. Freeman and Co., NY. 859 p. Sutherland, J. P. 1977 Effect of Sch izuporclla (Ectoprocta) removal on the foul- ing community at Beaufort, North Carolina, USA. In Coull, B.C. (ed.). Ecology of marine benthos, p. 155-17G. Univ. South Carolina Press, Columbia. Van Dolah, R.F., D.R. Calder, and D.M. Knott 1983 Assessment of benthic macrofiuma in an ocean disposal area near Charleston. South Carolina. Tech. Rep. 56. S.C. Mar. Resour. Cent., Charleston, SC 29412, 97 p. von Bertalanffy, L. 1938 A (juantitative theory of organic growth. II. Inquiries on growth laws. Hum. Biol. 10:181-213. Waltz, W., W.A. Roumillat, and P.K. Ashe 1979 Distribution, age structure, and sex composition of the black sea bass, Centwprifttis striald. sampled along the .southeastern coast of the United States. Tech. Rep. 43, S.C. Mar. Resour. Cent., Charleston, SC 29412, IS p. Wendt, P.H., D.M. Knott, and R.F. Van Dolah 1989 Community structure of the sessile biota on five artificial reefs of different ages. Bull. Mar. Sci. 44:1106-1122. Hayse: Feeding, age. growth, and reproduction of Chaetodipterus faber 81 Appendix Percent frequency of occurrence (%F), percent of total number %A'), percent of total vol jme(%V'), index of relative importance (IRI), and the modified index of importance (MI) for food items of | Chaetodipterus faber collected in South Carolina waters using nets or pole spears. Prey item %F %N %V IRI MI Algae Cladophora laetiverens 0.65 0.03 0.01 0.03 0.01 GraHlaria folifera 3.23 0.14 1.25 4.49 4.04 Hypnea musciformis 3.87 0.17 0.82 3.83 3.17 Ulva sp. 0.65 0.03 0.03 0.04 0.02 Algae undetermined 8.39 0.37 0.12 4.08 1.01 Total Algae 12.90 0.73 2.33 39.56 30.06 Porifera Leucoselenia sp. 0.65 0.03 0.02 0.03 0.01 Scypha barbadieyisis 0.65 0.03 0.50 0.34 0.33 Scypha sp. 0.65 0.03 0.42 0.29 0.27 Porifera undetermined 5.16 0.23 11.74 61.76 60.58 Total Porifera 6.45 0.31 13.25 87.50 85.46 Cnidaria Hydrozoa AijUwjjhenia rigida 0.65 0.03 0.07 0.06 0.05 Aglaophenia sp. 1.29 0.06 0.05 0.14 0.06 Campanularia sp. 0.65 0.03 0.01 0.03 0.01 Clytia sp. 1.29 0.06 0.04 0.12 0.05 Dyiinmena conucviin 0.65 0.03 0.22 0.16 0.14 Dynamenii sp. 0.65 0.03 <0.01 0.02 <0.01 Eudend ri u m ra meum 1.29 0.06 0.67 0.94 0.86 Eudendrium ramofntm. 3.87 0.17 1.98 8.31 7.66 Eudendrium sp. 21.94 0.96 12.55 296.39 275.35 Halocordyle disticha 3.23 0.14 0.38 1.67 1.23 Lnfoea fruticosa 1.29 0.06 0.53 0.76 0.68 Lytocarpus sp. 1.29 0.06 0.02 0.10 0.03 Monostaechas quadridens 0.65 0.03 0.08 0.07 0.05 Ohelia dichotoma 1.94 0.08 0.18 0.51 0.35 Ohelia genieulata 1.94 0.08 0.41 0.95 0.80 Ohelia sp. 3.87 0.17 0.72 3.46 2.79 Plumularia sp. 0.65 0.03 0.56 0.38 0.36 Sertularia distans 5.81 0.25 0.04 1.70 0.23 Sertularia sp. 3.23 0.14 0.07 0.68 0.23 Sertulariidae 2.58 0.11 1.69 4.66 4.36 Tubulariidae 0.65 0.03 0.22 0.16 0.14 Hydrozoa undetermined 43.87 1.92 4.91 299.29 215.40 Total Hydrozoa 79.35 4.51 26.53 2463.62 2105.16 Anthozoa Actiniaria 14.84 3.41 7.73 165.34 114.71 Octocorallia 0.65 0.03 0..50 0.34 0.33 Renilla ren iformis 16.77 6.20 9.33 260.56 156.46 Total Anthozoa 29.03 9.64 18.35 812.54 532.70 Scyphozoa Chiropsalmus quad nnnanMx 0.65 0.03 9.37 6.06 6.09 Ctenophora 4.52 0.20 0.36 2.52 1.63 Annelida Polychaeta Sabellaria sp. 1.29 0.11 <0.01 0.15 <0.01 Sabellariidae 0.65 0.08 0.01 0.06 0.01 Terebellidae 12.90 0.56 22.72 300.39 293.09 Polychaeta undetermined 3.23 0.14 0.39 1.70 1.26 Total Polychaeta 17.42 0.90 24.14 436.17 420.52 82 Fishery Bulletin 88(1), 1990 Appendix (continued) Prey item %F %N %V IRI MI Mollusca Gastropoda Anachis sp. 0.65 0.06 0.01 0.04 0.01 MitreUa lunata 0.65 0.03 0.01 0.02 D.Ol Nudihranchia 0.65 0.11 0.01 0.08 0.01 Tiirhonilla sp. 0.65 0.03 <0.01 0.02 <0.01 Gastropoda undetermined 2.58 0.11 0.03 0.37 0.08 Total Gastropoda 5.16 0.34 0.06 2.03 0.31 Pelecypoda A nadara transversa 0.65 0.03 <0.01 0.02 <0.01 Musculus lateralis 4.. 52 0.31 0.02 1.47 0.09 Pelecypoda undetermined 5.16 11.42 0.14 2.89 0.72 Total Pelecypoda 9.68 0.76 0.16 8.91 1.55 Arthropoda Pycnogonida Anoplodactylus insignis 0.65 0.03 0.02 0.03 0.01 Crustacea Ostraeoda 1.94 0.11 <0.01 0.22 <0.01 Copepoda Calanoida 1.29 fl.OS <0.01 0.11 <0.01 Cyclopoida 1.29 0.37 <0.01 0.48 <0.01 Harpacticoida 9.68 3.35 0.03 32.76 0.29 Copepoda undetermined 0.65 0.17 <0.01 0.11 <0.01 Total Copepoda 12.26 3.98 0.04 49.17 0.49 Cirripedia Balanus sp. 5.16 0.37 0.07 2.24 o.3(; Cirripedia undetermined 1.94 0.14 <0.01 0.28 <0.01 Total Cirripedia 7.10 0.51 0.07 4.13 0..',0 Mysidacea 0.65 0.06 <0.01 0.04 <0.ol Tanaidacea 0.65 0.03 <0.01 0.02 <0.01 Isopoda Paracerceis caudala 0.65 0.17 0.02 0.12 0.01 Amphipoda Ampithoe sp. 5.16 1.30 0.06 7.00 0.31 Batea catherinensis 7.10 1.01 0.03 7.41 0.21 Caprella equUibra 11.61 3.07 0.15 37.44 1.74 Caprella penantis 17.42 42.46 2.25 778.74 39.20 Caprellidae undetermined .3.23 0..54 0.01 1.76 0.03 Cera p us t u h u I a r is 1.94 8.09 0.18 16.01 0.35 Dulichielta appendiculata 1.94 0.28 0.04 0.62 0.08 Erichthonius brasiliensis 18.71 6.06 0.22 117.50 4.12 Gammaridea undetermined 11.61 0.90 0.03 10.83 0.35 Gammaropsis sp. 2..58 0.23 0.01 0.60 0.03 Hyperiidae undetermined 0.65 0.03 <0.01 0.02 <0.()1 Jassa falcata 7.74 8.91 0.31 71.39 2.40 Lembos sp. 1.29 0.14 <0.01 0.19 <0.01 Luconacia incerta 2.58 0.54 0.02 1.44 0.05 Phtisica marina 4.52 1.16 (1.05 5.43 0.23 Stenothoe georgiana 5.16 1.10 0.03 5.83 0.15 Stevothoe minula 3.23 0.23 0.01 0.74 0.03 Stenothoe sp. 0.65 0.03 <0.01 0.02 <0.01 Total Amphipoda 45.81 76.06 3.55 3646.63 162.63 Decapoda Natantia undetermined 1.29 0.11 0.05 0.21 0,06 Xanthidae undetermined 0.65 0.03 0.01 0.02 0,01 Brachyura undetermined 2.58 0.17 0.02 0.48 0.05 Decapoda undetermined 1.29 0.06 0.01 0.09 0.01 Total Decapoda 4.52 0.37 0.09 2.05 0.41 Hayse Feeding, age. growth, and reproduction of Chaetodipterus faber 83 Appendix (continued) Prey item %F %N %V IRI MI Bryozoa Anguinella palmata 1.94 0.08 0.02 0.19 0.04 Bugula nfritina 7.74 0.34 0.15 3.75 1.16 Biigula sp. 1.29 0.06 0.01 0.09 0.01 Schizoporella errata 0.65 0.03 <0.01 0.02 0.01 Sundanella sibogae 1.29 0.06 0.07 0.16 0.09 Bryozoa undetermined 7,10 0.31 0.07 2.68 0.50 Total Bryozoa 17.42 0.87 0.33 20.92 5.75 Echinodermata Ophiuroidea undetermined 1.94 0.08 0.04 0.24 0.08 Chordata Urochordata Ascidiacea D idem n inn candid urn 0.65 0.03 <0.01 0.02 <0.01 D istapiia bermudiensis 1.29 0.06 0.07 0.16 0.09 Eudistoma carolinense 1.94 0.08 0.30 0.75 0.58 Ascidiacea undetermined 0.65 0.03 0.04 0.04 0.03 Total Ascidiacea 3.23 0.20 0.43 2.03 1.39 Pisces (scales) 0.65 0.03 <0.01 0.02 <0.01 Eggs undetermined 1.94 0.08 0.84 1.72 1.63 Number of stomaclis examined: 177 Examined stomachs with food: 155 Abstract. -This study develops and demonstrates a framework for measuring changes in the total-fac- tor productivity of fishing fleets op- erating in common-property, open- access fisheries. This approach is distinguished from previous efforts to measure productivity growth in fisheries by our explicit treatment of the fishery resource as a constraint on production and by the incorpora- tion of recent advances in produc- tivity measurement that take into account variations in the degree of capacity utilization in an industry. The approach is developed in suffi- cient detail for the non-economist fishery analyst to follow and imple- ment. The empirical analysis of total- factor productivity growth in the U.S. tropical tuna fleet reveals that this approach eliminates a significant amount of bias in fleet productivity measures which is otherwise intro- duced when using traditional methods of productivity analysis. On Measuring Fishing Fleet Productivity: Development and Demonstration of an Analytical Framework Samuel F. Herrick, Jr. Dale Squires Southwest Fisheries Center, National Marine Fisheries Service, NOAA PO. Box 271, La Jolla, California 92038 Maiiuscri])t accepted 31 August 1989. Fishery Bulletin, U.S. 88:85-94. Measuring changes in productivity has long been an important compo- nent in evaluating an industry's eco- nomic performance. Such measures in fishing industries can signal the need for, as well as indicate the suc- cess of, policy or management ac- tions. Measurement of productivity growth or technical progress in ma- rine fishing industries has received attention by Bell and Kinoshita (1973), Davis et al. (1987), Duncan (n.d.), Kirkley (1984), and Norton et al. (1985). Two considerations not addressed by previous researchers are impor- tant for evaluating productivity gi-owth in fishing industries. First, tradition- al measures of productivity implicit- ly assume that a fishing industry's capital stock is being utilized, in an economic sense, at its long-run ecjui- librium or capacity output level. Thus, traditional measures fail to adjust for variations in the degree of utilization of the industry's productive capacity. Second, the effect of changes in abundance of the fish stock on pro- ductivity growth in fishing industries has not been specifically accounted for in the traditional analysis. As a result, changes in resource abun- dance are not disentangled from changes in productivity. In this study we develop a non- parametric framework utilizing the method of growth accounting and economic index numbers to analyze the productivity growth of fishing fleets operating in common-property, open-access fisheries. The framework is then demonstrated by analyzing productivity growth in the U.S. trop- ical tuna purse seine fleet over the years 1981-1985.^ We also demon- strate a method for deriving implicit aggregate output and input price in- dices for the purse seine fleet. In developing the productivity as- sessment framework, we introduce further refinements to the standard procedure described by Denny et al. (1981) and Cowing et al. (1981)- which has seen continual improve- ment since the pioneering work of Solow (1957)— by adjusting for vari- ations in capacity utilization and changes in resource abundance. The exposition includes the technical detail necessary to provide the non- economist fishery analyst with a com- prehensible and useful means of tracking and analyzing productivity gi'owth and performance in fisheries. 'The method is non-parametric because param- eters of the production technology or produc- tion function are not econometrically esti- mated. The method of growth accounting and economic inde.x numliers is discussed in a later section of the te.xt. Econometric estimation of productivity growth does not impose the conditions of a constant-return.s-to-scale pro- duction technology and Hicks-neutral tech- nical change. However, this comes at the ex- pen.se of more demanding data requirements: either a longer time-series of aggregate data or more vessel-level observations in any given year. In addition, some fairly sophisticated econometrics and economics are required to estimate and interpret the results. 85 86 Fishery Bulletin 88(1). 1990 In the following section we introduce the methodol- ogy used to analyze total factor productivity in fishing fleets. The potential bias from failing to account for variation in economic capacity utilization is investigated in the third section. We discuss the data, empirical issues, and the construction of index numbers for the U.S. tropical tuna purse-seine fleet in the fourth sec- tion, and then we develop the implicit price indices and report and interpret results of the empirical analyses in the fifth section. Total-factor productivity The standard procedure for estimating total-factor pro- ductivity is derived from the economic theory of pro- duction. In common-property natural resource indus- tries, the production function expresses a stock-flow relationship between the resource stock and the flow of resource extraction or output from the production activity in any given time-period. An important con- sideration when measuring productivity growth in fishing industries is defining the role of the common- property resource stock in the production technology. Rather than treating the common-property resource as a conventional input (Scott 1954, Clark 1976, Dasgupta and Heal 1979), it is more appropriately specified as a constraint to the production technology. The fish stock is a biological constraint on the produc- tion technology because its abundance affects the pro- duction environment within which fishing firms oper- ate, but it is beyond the control of any individual firm. That is, the use of conventional inputs such as capital, labor, and energy is conditional upon expected resource-abundance levels. Changes in resource abun- dance shift the production technology. McFadden (1978) develops this approach by treating environmen- tal parameters (such as resource abundance) in a man- ner similar to disembodied technical change; i.e., technological progress due to more efficient use of ex- isting inputs. Finally, resource abundance is a techno- logical constraint because the total catch cannot exceed the abundance available, and an increase (decrease) in resource abundance allows an increase (decrease) in catch for any given level of input usage and state of technology.- ^Gordon (19.54) similarly treated biological abundance as a techno- logical constraint. Gordon (p. I'M')) states, "For each given level of population, a larger fishing effort will result in larger landings. Each population contour is, then, a production function for a given popula- tion level." Moreover, as Gordon noted, this approach does not preclude the impact that increases in catch typically have in reduc- ing the resource stock. If resource abundance was in.stead another factor of production, like labor or capital, then changes in resource abundance would imply movements along the existing production function rather than shifts up or down the production function. When the fish stock is treated as a technological con- straint, the production function, F, relates the max- imum flow of output in time t (such as tons of fish extracted), Y{t), to the flow of iV-i-1 inputs, Xi{t), X2{t ), . . . ,X^ + 1 , the state of technology represented by A{t ),^ and abundance of the fish stock indexed by Bit): Y{t) = F[Xi{f). X.it),. . .,X^.,,(t)', A(t), B(t)]. (1) Growth-accounting framework Total-factor productivity measures are derived from equation (1) using the growth-accounting framework and economic index numbers.'' This approach accounts for the growth in output flow over time by partition- ing this output growth among the growth in inputs, technical progress, and changes in resource stock abun- dance. Total-factor productivity is then measured as the residual in the growth of output flow after account- ing for all of the measurable sources of growth. Under the growth-accounting framework, a constant- re turns-to-scale production function is as.sumed, so that a proportional increase in inputs yields a proportional increase in output flow for any given level of resource stock abundance and state of technology. Movement in time t is assumed to lead to improvements in the state of technology, so that 6FI6A (t )>0 (Solow 1957). Following conventional practice, we further assume a particular form of technological change, Hick's-neutral disembodied technical change.^ Moreover, 6F/6B{t)>0, so that increases (decreases) in resource stock-size allow an increase (decrease) in the flow rate of extrac- tion for any given input bundle and state of technical progress. We assume a Schaefer (1957)-tyf:)e produc- tion technology, and further assume that changes in resource stock-size are Hick's neutral, so that the ^The state of technology refers to the current level of technology or kind of productiVi;,-, where S^ is capital's cost share. 0.5 2 (S„ ■^ S,,_,)ln(A'„/A', ln(B,/B, ,). 'The alternative approach is that of fixed-base indices in which TFP in any time / is directly compared with TFP in the initial or base period. See Squires (1988) for an extensive discussion within fisheries. 88 Fishery Bulletin 1990 Failure to capture this additional source of variation creates a potential bias in the true productivity resid- ual, A /A, which varies with the rate of capacity utilization. Methods of capacity utilization adjustment When the firm is in temporary or short-run equilibrium rather than long-run equilibrium, the productivity residual formula can be adjusted for variations in CU in two different ways. The first relaxes the assump- tion that service flows are proportional to stocks by ad- justing the stock of a quasi-fixed input to reflect its flow of services^ (Berndt and Fuss 1986). Thus, rather than specifying capital as a stock (e.g., the number of vessels in the fleet), capital is measured by its flow of services or total time of utilization (e.g., as fleet vessel-days fished). Such a flow adjustment corresponds to an economic notion of CU, because we assume that a pro- ducer's decision to increase or decrease running and fishing time is the outcome of an economic optimiza- tion process. The second approach to adjusting the productivity residual for CU variations uses engineering notions of the proportion of available productive capacity that is actually being utilized. Let capacity output Y* repre- sent the maximum possible output level corresponding to "normal" input usage, existing technology, and the stocks of quasi-fixed inputs. A measure of capacity utilization is then obtained by the identity: CU= YIY*, where Y is the observed output level. Data and empirical issues Tornqvist output indices for the U.S. ti'opical tuna purse-seine fleet were constructed using annual deliv- eries of skipjack and yellowfin tuna by the fleet to U.S. canneries and the corresponding dollar values of these deliveries. Weighted exvessel implicit prices for skip- jack and yellowfin tuna were calculated by dividing the total dollar value of cannery receipts for each species by the total volume of cannery receipts. Revenue and cannery receipts data were obtained from the South- west Region, National Marine Fisheries Service (NMFS). Dollar values were deflated by the GNP implicit price index. To construct the input indices, four major categories of factors used in owning and operating a tropical tuna •An alternative approach to accounting for variations in economic capacity utilization adjusts the cost or income shares rather than the flows of capital services. Berndt and Fuss (1986) and Hulten (1986) develop this approach, and Squires and Herrick (1988) pro- vide a fisheries application. This approach is well suited when ac- curate and detailed data on running and fishing time are unavailable. purse seiner wei'e identified: labor, capital, fuel, and other intermediate inputs (transshipment services, repairs, gear, insurance, helicopter services, travel, and other). Constant-dollar unit prices for these inputs were estimated based on purse-seine expenditure data reported by the U.S. International Trade Commission (ITC 1986). The labor index incorporates the flow of lal)or ser- vices derived by multiplying estimated total days ab- sent (at sea or absent from port) per vessel per year by 19 crew members, which is the assumed average crew size in each year of the period. The unit price of labor, cost per crew-day-absent, was estimated by dividing the sum of the ITC's reported annual crew- related expenditures per vessel by a measure of annual crew-days-absent per vessel. Three different capital indices were constructed for the total-factor productivity analysis of the U.S. tropical tuna fleet. The first two capital indices both assumed that firms were in long-run equilibrium and that the flow of capital services was proportional to the capital stock. The first capital index specified capital as the annual number of vessels in the fleet. The sec- ond capital index captured the effect of different vessel sizes (where size is a measure of the vessel's hold or carrying capacity) upon catch rates by measuring capital as the annual carryitig capacity of the fleet. The third capital index not only captured the effects of different-sized vessels, but also accounted for actual changes in the flow of services from this size-differen- tiated capital stock. In this third case, the flow of capital services was measured in annual ton-days- absent, an aggregation of each vessel's cai-rying capa- city multiplied by the number of days it spent at sea during the year. Measures of annual ton-days-absent were derived from purse-seine-tleet activity data com- piled by NMFS. The cost share for capital used in construction of all the capital indices was its market rental price, the sum of the annual interest expense and reported annual depreciation per vessel from the ITC sample. The fuel index was constnicted by dividing the annu- al fuel expenditure per vessel (from the ITC sample) by average fuel prices provided by the American Tuna Boat Association. Fuel consumption per vessel was then multiplied by the number of vessels in the fleet resulting in the aggregate annual fuel consumption. The index of other intermediate inputs was derived by deflating the fleet's nominal expenditure on this category of inputs by the producer price index for in- dustrial commodities. This approach represents the collective use of these inputs in real terms. The nominal expenditure for this category of inputs divided by the corresponding deflated ex])enditure was used as a proxy for the unit price of other intermediate inputs. Herrick and Squires Measuring fishing fleet productivity 89 Table 1 Total-factor productivity growth in tiie U.S. tropical tuna fleet. 1981-85. No biomass adjustment Engineering-CL'-adjusted carrying capacity' Biomass adjusted Ton-days-absent; biomass adjustment^ Capital input alternatives No. of Period vessels' Fleet carrying capacity^ Ton-days- absent* 1981-82 -0.1037 1982-83 0.3739 1983-84 0.0298 1984-85 0.0619 Mean 0.0905 -0.1241 -0.0863 -0.1159 0.1006 0.3643 0.3378 0.1753 0.2120 0.0278 0.0767 0.1056 -0.4330 0.0499 0.0452 0.0932 0.0192 0.0795 0.0933 0.0645 -0.0253 capital, capital represented by the number of vessels; no biomass adjustment. capital, capital represented by fleet carrying capacity: no biomass adjustment. i by the actual flow of capital services, ton-days-absent: no biomass adjustment. arrying capacity corrected for variations in capacity utilization iCU); no biomass adjustment. 3n-days-absent; biomass adjusted. qvist bilateral indices using the logarithmic form of Equation (3) given in Footnote 6. 'Long-run equilibrium in 'Long-run equilibrium in 'Capital input represente ^Capital represented by c ''Capital represented by t Note: Calculated as Torn Because the U.S. tropical tuna purse-seine fleet oper- ates almost exclusively in the Pacific Ocean, resource abundance measures relied on the extensive biological database that has been compiled by the Inter-American Tropical Tuna Commission (lATTC, La Jolla, CA 92038). From the lATTC database, we obtained esti- mates of annual yellowfin tuna biomass and estimates of catchability coefficients for yellowfin in the Commis- sion's eastern tropical Pacific yellowfin regulatory area (CYRA). The CYRA was the only region of the Pacific for which such data were available. Moreover, there were no such data for skipjack tuna from any area of the Pacific. Because of these circumstances, it was not possible to explicitly account for fluctuations in biomass over the full range of the fishery and across all species harvested. Therefore, only the annual yellowfin bio- mass in the CYRA weighted by the catchability coeffi- cient was used to adjust fleet productivity for changes in resource abundance. Empirical results Table 1 reports changes in annual total-factor produc- tivity growth for the U.S. tropical tuna fleet over the period 1981-85 using Tornqvist bilateral-chain indices. Each of the total-factor productivity growth rates pre- sented in Table 1 is distinguished by its specification of the capital input and adjustments for changes in re- source abundance. Treatment of outputs and the other inputs is the same in all cases. To anticipate our results, we find that adjusting the traditional productivity mea- sures for variations in economic capacity utilization and changes in resource abundance pares away sources of output growth that are not due to technical progress, giving a more accurate measure of productivity. Productivity measures under full equilibrium Columns (1) and (2) of Table 1 provide productivity growth measures assuming long-run equilibrium— eco- nomic capacity is fully utilized— without accounting for changes in resource abundance. The total factor pro- ductivity growth rates in column (1) use the first capital index: capital is represented by the number of vessels included in the fleet. The growth rates reported in col- umn (2) are based on the second capital index: capital is represented by the carrying capacity of the fleet. Because vessels leaving the fleet during the period tended to be older and in the smaller size-categories, capital expressed in number of vessels decreased at a greater rate than capital measured in aggregate car- rying capacity (columns [2] and [3] of Table 2). Hence, total-factor productivity growth based on the second capital index is lower than that based on the first capital index for those years in which the fleet declined.' 'It might be argued that technical progress is embodied in the capital stock, so that different ages of vessels, embodying different advances in technical progress, should be ex[3licitly considered. Tliis is referred to as vintage effects, and is certainly the case for some types of technical progress such as purse seining versus pole-and-line harvesting. However, in recent years, much of the technical pro- gress has been in the form of vessel electronics. While this type of technological change in a narrow sense, represents embodied tech- nical change, so that vintage effects could theoretically be impor- tant, the volume of investment is negligible in comparison with the vessel's value, and much of the technical change is fundamentally related to the managerial function, information, and learning-by- doing: Hick's-neutral technical change. 90 Fishery Bulletin 88(1). 1990 Table 2 Annual growth in output, capital, CYRA* yellowfin tuna biomass, and capital stock capacity utilization rates, 1981-85. CYRA No. of Fleet Ton-days- yellowfin tuna Rate of capacity Period Output' vessels'- carrying capacity" absent'' biomass' utilization'' 1981-82 -0.0090 0.0947 0.1151 0.0773 -0.1869 -0.0327 1982-83 0.1869 -0.1870 -0.1774 -0.1508 0.1258 0.3898 1983-84 -0.0165 -0.0464 -0.0444 -0.0932 0.5096 - 0.0244 1984-85 -0.0800 -0.1420 -0.1299 -0.1252 0.3167 -0.0125 Mean 0.0204 -0.0702 -0.0592 -0.0730 rea, lATTC. 0.1913 0.0801 'Eastern tropical Pacific yellowfin regulatory a 'Annual change in deliveries of U.S. -caught yellowfin and skipjack tuna to U.S. canneries "Annual change in number of vessels comprising U.S. tropical tuna purse-seine fleet. •'Annual change in U.S. tropical tuna purse-seine fleet's hold capacity. 'Annual change in the fleetwide flow of capital services for the U.S. tropical tuna purse-s eine fleet. 'Change in lATTC estimates of the annual CYRA yellowfin biomass. '''Annual change in ratio of actual U.S. tropical tuna purse-seine fleet output to fleet engi neering capacity. Note: In natural log form (cf. footnote 6). Productivity measures adjusted for capacity utilization The total-factor productivity growth rates presented in cohimn (3) of Table 1 incorporate the third capital index: the annual flow of capital services from the size- differentiated capital stock, ton-days-absent. Since the number of ton-days-absent directly reflects the degree to which economic capacity is utilized, resulting mea- sures of total-factor productivity growth are not sub- ject to a capacity-utilization bias as are the growth rates shown in columns (1) and (2) of Table 1. The effects of adjusting for CU are revealed in Table 2 by comparing the rates of change in the fleet's capital stock and the fleet's flow of capital services: fleet carrying capacity and ton-days-absent reported in columns (3) and (4), respectively. Between 1982 and 1983, fleet carrying capacity decreased 18%, while ton- days-absent decreased 15%. This means that the re- duced fleet was fishing more intensively, and that a measure of productivity growth based on the stock of capital or fleet carrying capacity, without correcting for the degree of capacity utilization, will be biased up- wards. Similarly, comparing growth in fleet carrying capacity and ton-days-absent between 1983 and 1984 reveals a greater decline in the latter relative to the former. Therefore, a smaller fleet capital stock is util- ized (fished) proportionately less, and productivity growth based simply on the capital stock (carrying capacity) will be understated. Column (4) of Table 1 reports growth in total factor productivity for the U.S. tropical tuna fleet, where changes in fleet carrying capacity— changes in the stock of capital— are corrected for variations in the rate of capacity-utilization. In this case, the capacity utiliza- tion adjustment is based on vessel design or engineer- ing characteristics which act to establish an upper limit on fleet output in a physical sense. To estimate fleet engineering capacity, we assumed that each vessel was capable of making three fishing trips annually, filling its hold on each trip. Thus, maximum fleet output in each year is three times the fleet's carrying capacity. The ratio of actual fleet output (the total quantity of tuna delivered to canneries) to maximum potential fleet output estimates the degree of capacity utilization. Changes in the rate of capacity utilization using the engineering adjustment are shown in column (6) of Table 2. These capacity-utilization rates were then used to derive the engineering-adjusted total-factor produc- tivity growth rates presented in Table 1, column (4). Comparing the total-factor productivity growth rates in columns (2) and (4) of Table 1 discloses the extent of the bias introduced by failing to account for varia- tions in the degree of capacity utilization. The capacity-utilization adjustment is made to ap- proximate the actual flow of services from the quasi- fixed factor, the capital stock. Therefore, one might expect the total-factor productivity growth rates using ton-days-absent (Table 1 , column [3]) to closely corre- spond to those based on correcting for cajjacity utiliza- tion using the engineering approach (Table 1, column [4]). The fact that they do not points out that, in an economic sense, capacity-utilization adjustments ex- plicitly recognize that quasi-fixed factors are not always utilized at the long-run equilibrium or full economic- capacity-output level, the level of output which mini- mizes the per-unit-cost of production. Under these cir- cumstances, engineering-capacity output, as we have Hernck and Squires Measuring fishing fleet productivity 91 Table 3 Implicit aggregate output and input price and total-factor- productivity chain indices. Total- Implicit Implicit factor aggregate aggregate Year productivity output price input price 1981 1.0000 1.0000 1.0000 1982 1.1058 0.9504 1.0221 1983 1.3669 0.8833 0.9016 1984 o:8865 0.7961 1.1941 1985 0.9037 0.7049 1.1099 Note: Implicit price indices formed by Fisher's weak factor- reversal relationship. See text for details. Total-factor pro- ductivity, calculated as Tornqvist bilaterial chain indices, ad- justed for variations in biomass and capacity utilization (using ton-days-absent). A / \ \ Total Factor / Productivity \ Aggregate \ Input Price \ Figure 1 Total-factor productivity and implicit price indices. defined it, should be greater than full economic-capa- city output. Thus, the engineering capacity-utilization rates should be biased downward, and the correspond- ing productivity growth rates will inherit this bias. The preferred capacity-utilization correction, consistent with economic theory, should adjust the capital stock by its time in use to provide a flow measiu-e of capital services. Productivity measures adjusted for biological abundance The total-factor productivity growth rates presented in column (5) of Table 1 are based on the actual flow of capital services and the growth of the yellowfin biomass in the CYRA (Table 2, column [5]). Compar- ing columns (3) and (5) of Table 1 indicates that the in- crease in the CYRA yellowfin biomass during most of the period acts to partially offset gains in total-factor productivity otherwise attributable to technical pro- gress. The resource and capacity utilization adjusted total-factor productivity growth rates from Table 1 , col- umn (5) are used to compute the index of total factor productivity for the U.S. tropical tuna fleet shown in Table 3 and in Figure l.^" Implicit output and input price indices Fleet economic performance depends upon the real prices of outputs and inputs in addition to total-factor "These period-to-period changes in total-factor productivity, i.e., pro- ductivity growth, were computed following footnote 6. Next, the exponent is taken of each value in column (5) of Table 1, giving TFP,i^. The Tornqvist bilateral-chain index is then formed following Equation (4), to give the first column of Table 3. Note that the value 1.000 for 1981 is the exponent of zero, where zero refers to the zero change for the first period. productivity. Corresponding to the aggregate output and input quantity indices are implicit aggregate out- put and input prices. These are calculated by a rela- tionship due to Fisher (1922), which states that the pro- duct of the price index times the quantity index equals the expenditure ratio between the two time-periods. The implicit price index for an output (or input) can be interpreted as the ratio of the price level in period / -i- 1 to the price level t. Fisher's relationship for the price (PI) and quantity (QI) indices for aggregate output i' can be written as: P/(P,,P,,i,F,,F,,,)Q/(P,,P,,,,y,, }',,,) = I,p„,,F,„,/I,p„y„. (6) Given either a price index or quantity index, the other function can be defined implicitly by Equation (6). Implicit Tornqvist bilateral-aggregate output-price and input-price chain indices for the U.S. tropical tuna purse-seine fleet, along with the resource-abundance adjusted total-factor productivity index, are presented in Table 3 and in Figure 1. Increases in total-factor pro- ductivity or output prices, or both, improve the fleet's economic performance, while increases in input prices worsen the fleet's economic performance. Taken together, the changes in total-factor produc- tivity, aggregate input-price, and aggregate output- price indices shown in Figure 1 and Table 3 indicate that the 1981-85 period was highly unstable with re- gard to the fleet's economic performance. Herrick and Koplin (1986, 1987) point out that this was a time of massive restructuring in the U.S. tuna industry, dur- ing which the U.S. fleet began a significant shift of its operations from the eastern to the western Pacific 92 Fishery Bulletin 88(1), 1990 Ocean, ami the U.S. canned tuna market was inundated by imports. These events undoubtedly contributed to the unstable pattern of productivity growth, and would have introduced additional, unmeasurable disturbance into the system which we are unable to disentangle from the pn.xluctivity ivsidual in the gi'owth-accounting framework." Concluding remarks We have shown how the nonparametric growtli-accoujit- ing framework can be modified to measure the growth in total-factor productivity or technical [irogress of fishing fleets. This framework can provide useful in- formation for tracking and analyzing economic growth and its causal factors in a fishing industry, particular- ly where short tinie-sei'ies of aggregate data are all that is available. Our approach can be readily implemented in fishing industries in both developed and developing countries, even by non-economist fishery analysts. Unique to our approach is the treatment of the fishery resource, not as a conventional input, but as a techno- logical constraint on production. Our empirical analysis of total-factor productivity growth in the U.S. tropical tuna fleet demonstrates that disentangling the produc- tivity residual from changes in I'esoiu'ce alumtlance pro- vides markedly different results. We consider capital the most important component of aggregate input because it is represented by the basic unit of production, the fishing vessel. The capital stock— the fishing fleet— determines capacity output in both an economic and engineering sense. Theoi'etical- ly, it is the flow of services from the capital stock that should serve as the capital input when measuring total- factor productivity. In practice, however, one may not have measures of the flow of capital services, in which case proper specification of the capital input, and ac- counting for temporary equililirium effects such as variations in the degree of capacity utilization, becomes extremely important. "Because our biomass adjustment is based only on changes in yellowfin tuna resource abundance in the CYRA. the variation in total-factor productivity should become more pronounced as the fleet moved from the eastern to the western Pacific Ocean during the 1981-8.5 period. Furthermore, there were likely to have been some initial technical inefficiencies as the fleet began fishing in the relatively unfamiliar western Pacific. New investment and industry restructurings can have very real detrimental effects upon the time-path of productivity growth. Therefore, we would expect to see unstable productivity growth as firms adapt over time to changing industrial conditions. Under such circumstances, the assumptions underlying our model— constant-returns-to-scale, disembodied Hick's-neutral technical change conditions, and technical efficiency— may not fully apply. Nonetheless, these are limitations of virtually any application using the growth-accounting framework. Acknowledgments Senior authorship is not assigned. The comments of Andy Dizon, Roberto Enriquez, Susan Hanna, Dan Huppert, Bruce Rettig, and two anonymous referees have substantially improved the paper and are grateful- ly acknowledged. We wish to thank Jeffrey Lee and Patrick Tomlinson for technical assistance in the prep- aration of this manuscript. The authors remain respon- sible for any errors. The views expressed are not neces- sarily those of the National Marine Fisheries Service. Citations Bell, F., and K. Kinoshita 1973 Productivity gains in U.S. fisheries. Fish. Bull., U.S. 71:911-919. Berndt, E., and M.A. Fuss 1986 Productivity measurement with adjustments for varia- tions in capacity utilization and other forms of temporary eciuililirium. .1. Econom. 1^3:7-29. Clark. C. 1976 Mathematical bioecoiiomics. Wiley, NY.. Cowing, T., J. Small, and R. Stevenson 1981 Comparative measures of total factor productivity in the regulated sector: The electric utility industry. In Cowing, T., and R. Stevenson (eds.). Productivity measurement in regulated industries, p. 162-177. Acad. Press, NY. Dasgupta, P.. and G. Heal 1979 Economic theory and exhaustible resources. Cambridge Univ. Press, Cambridge. Davis, L.E., R.E. Gallman, and T.D. Hutchins 1987 Technology, prmluctivity and profits: British-American whaling competition in the North Atlantic. 181(;-1842, Oxford Econ. Papers ,39:738-7.S9. Denny, M., M. Fuss, and L. Waverman 1981 The measurement and interpretation of total factor pro- ductivity in regulated industries, with an application to Cana- dian telecommunications. In Cowing, T.. and R. Stevenson (eds.). Productivity measurement in regulated industries, p. 179-212. Acad, Press. NY, Duncan, I. n.d. New Zealand's inshore fishery: A summary of economic cimclusions and management options. N.Z. Fishing Industry Board, Wellington, mimeo. Fisher, I. 1922 The making of index numbers. Houghton Miftlin. Bo.ston. Gordon, U.S. 19.54 The economic theoi-y of a common property resource: The fishery. .1. Political Econ. 62:124-142. Herrick, S.. Jr.. and S. Koplin 1986 U.S. tuna trade summary, 1984. Mar, Fish, Rev, 48(3): 28-37. 1987 U.S. tuna trade summary, 1985. Mar. Fish. Rev. 49(3): 73-84, Hulten. C.R. 1974 Divisia index numbers. Econom. 41:1017-1025. 1986 Productivity change, capacity utilization, and the sources of efficiency growth, .1, Econom, 33:31-50. International Trade Commission (ITC) 1986 Competitive conditions in the U.S. Tuna industry: Report to the President on investigation no. 332224 under section 332 oftheTariff Actof 1930. U.S. Int. Trade Comm,. Wash., DC 20436, 320 p. Hernck and Squires Measuring fishing fleet productivity 93 Kirkley. J. 1984 Productivity in fisheries. Discussion pap., Woods Hole Lab.. Natl. Mar. Fish. Serv., NOAA, Woods Hole, MA 02543, 20 p. McFadden, D. 1978 Cost, revenue, and profit functions, /h Fuss, M., and D. McFadden (eds.). Production economics: A dual approach to theory and applications, vol. 1, p. 3-109. North-Holland Press, Amsterdam. Nelson, R. 1981 Research on productivity growth and productivity dif- ferences: Dead ends and new departures. J. Econ. Lit. 19:1029-1064. Norton, V.. M. Miller, and E. Kinney 1985 Indexing the economic health of the U.S. fishing in- dustry's harvesting sector. NOAA Tech. Memo. NMFS- F/NEC-40, Northeast Fish. Cent., Natl. Mar. Fish. Serv., Woods Hole, MA 02543, 42 p. Schaefer, M.B. 1957 A study of the dynamics of the fishery for yellowfin tuna in the eastern tropical Pacific Ocean. Bull. 6, Inter- Am. Trop. Tuna Comm., La Jolla, CA 92038, p. 247-285. Scott, A. 1954 The fishery: The oLijectives of sole owiiership. J. Political Econ. 63:116-124. Solow. R. 1957 Technical change and the aggregate production func- tion. Rev. Econ. Stat. 39:312-320. Squires, D. 1988 Index numbers and iiroductivity measurement in multi- species fisheries: An application to the Pacific coast trawl fleet. NOAA Tech. Rei:.. NMFS 67, Natl. Oceanic Atmos. Adm., Natl. Mar. Fish. Serv., 34 p. Squires D., and S. Herrick, Jr. 1988 Productivity measurement in common property resource industries: An application to the Pacific coast trawl fishery. Mimeo rep.. Southwest Fish. Cent., Natl. Mar. Fish. Serv., NOAA, La Jolla, CA 920:38, 34 p. Tornqvist, M.A. 1936 The Bank of f^inland's consumption price index. Monthly Bull. Bank of Finland 10:27-34. Appendix The growth-accounting framework is developed by dif- ferentiating Equation (1) with respect to time t: dY(t) dt N+l = I 6F_ 6F 6B(t) dX;{t) dt dt 6F dA(t) dA(t) dt (A.l) where (/i'(^ )ldt represents the growth rate of output (i.e., extraction rate) due to technical progress. Putting the left side of Equation (A.l) into percentage terms (and suppressing the notation for each time period / ) gives: dYl dt Y 1 F Af-fl y 6F^dX, 6FdA 6FdB ,T"i 6X, dt ^ 6A dt '^ 6B dt (A.2) J To obtain proportionate growth rates for all vari- ables, substitute [dYldt \ [1/7] = YIY, [dX,ldt ] [1/A',] = A',/A',. [dAldt] [IIA] = AIA, and [dBldt] [IIB]^B/B into Equation (A.2) to give: Y N+l = I ' 6F X- ^' , 6F A 6X^ F X, _ 6A F 6FB 6B F B b' (A.3) In equation (A.3), by convention (Denny et al. 1981, Solow 1957) the term [6FI6A ] [AIF] is the proportional shift in the production function with time and is set equal to unity. Thus a 1% increase in the index of tech- nical progress increases the flow rate of output by 1%. This shifting of the production function through time is called technical change or the time rate of growth of technical progress. The term [dFI6B] [BIF] is set to unity because it is a technology-shift parameter for a Schaefer (1957)-type production technology in which catch rates are proportional to resource abundance for any given vector of inputs. i- Define E, = [6FI6X,] [X,IF]. This is the output elasticity of input A', , representing the proportional change in output flow for a given change in A, within some level of resource stock abundance and state of technological progress. Substituting the output elas- ticity expression into (A.3) gives the long-run rate of output growth as: Y Y X ^1 + + - , ,=1 A, A B (A.4) ^This corresponds to the production function siiecified by Schaefer (1957), }' = FlE.B) = qEB. where q denotes the catchahility coef- ficient and E represents effort or an aggregate input index so that E = 3(A', Xy^ I ), where g is a linearly homogeneous aggi-egator function and the A' -i- 1 A', are the inputs. Thus, di'/dB = 1. We are grateful to Jim Kirkley who pointed this out to us. Moreover, the Schaefer production function implicitly assumes constant-returns- to-scale in inputs, because dVldE = 1, so that a proportionate in- crease in £ generates an equal proportionate increase in Y. In addi- tion, changes in the resource stock affect the production function in a Hick's-neutral manner, similar to technical progress. 94 Fishery Bulletin 88(1). 1990 or after rearranging: A Y TTx X, B where the dots over the variables represent time derivatives. Equation (A.5) is the fundamental equation of growth accounting in its continuous time form. Thus full, long- run equilibrium total-factor productivity growth, A I A, identified with technical progress, is a residual after the sources of output growth have been allocated among intertemporal changes in inputs and resource abundance for a constant-returns-to-scale production technology. Equation (A.5) allocates the growth rate of Y{t ) among A (t ), X{t ), and B{t ) as required, but because the E, are not observable, two additional steps are re- quired for empirical analysis. Assuming that all inputs are paid the value of their marginal product, then 6FI6XXt ) = P,{t )IP(t )■ where P(t ) and P,(t ) are the full equilibrium prices of output and inputs, respec- tively. This implies that: S,(t) = E,(t) = PXt)X,(t)l P{t)Y(t), where S,(t) is the income or cost share of input Z,. Under constant returns to scale, total costs equal total revenue; i.e., S,P,(/ )X,(t ) = P(t )Y{t ), and 1,5,(0 = 1. The final step is to substitute E,{t ) = S,{t) into equa- tion (A.5), which gives an equation in which all variables are measurable except A/A, which is calculated as a residual: A Y ^4-\ X, B = - 1 S, ' - -, (A.6) A Y f~i X, B where the notation for time-period t is again suppressed and 'Z,S,(X,IX,) represents aggregate input growth. Productivity growth equals the rate of change of out- put flow minus a share-weighted index of rates of change of inputs minus the rate of change of the re- source stock. The index of productivity growth in Equa- tion (A.6) is also called a Divisia index (Hulten 1974). Abstract.- The phylogeny and historical biogeography of hakes {Merhicciu^) are reexamined using a cladistic analysis of Inada's (1981) osteological data. One of the 188 most parsimonious trees (cladograms), which also has the lowest (best) F value, is congruent with the scheme of evolution proposed by Ho (1974) for the hake-specific copepod para- sites. Offshore hake M. albidus is the sister group of all other extant hakes. Silver hake M. bilmearis is the sister group of the three eastern Pacific hakes: Chilean hake M. gaiji, Pana- manian hake M. angustimanus, and Pacific hake M. produetus. The five Eastern Atlantic species, European hake M. nierluccius, Senegalese hake M. senegalensis, Benguelan hake M. polli, shallow-water cape hake M. capensis, and deep-water cape hake M. paradoxus, are monophyletic and constitute the sister group that forms a trichotomy with Argentine hake M. huhbsi and New Zealand hake M. australiH. The biogeography inferred from the most parsimonious phylogenetic hypothesis of hakes dif- fers from the views given by both Inada (1981) and Kabata and Ho (1981). The ancestral hake in the eastern seaboard of North America diverged into two stocks. Subse- quently, one of them vicariated into a northern and a southern popula- tion. The southern population moved southward first and then separated into three stocks, with one of them moving eastward and crossing the Atlantic in the region of low lati- tudes. One of the two remaining stocks moved further southward and entered the Pacific through the Drake Passage. After further diver- gence, a descendant stock of the northern population off the coast of North America also moved south- ward; it did not enter the south Atlantic, but, instead, moved into the Pacific over the submerged Panama- nian Isthmus. Some geologic events are discussed for their possible ef- fects in the formation of the present pattern of hake distribution. Phylogeny and Biogeography of Hakes [Merluccius; Teleostei): A Cladistic Analysis Ju-shey Ho Department of Biology, California State University Long Beacfi, California 90840-3702 Manuscript accepted 28 August 1989. Fishery Bulletin. U.S. 88:95-104. In 1981, two views on the historical biogeography of hake were proposed: one by T. Inada and the other jointly by Z. Kabata and J.S. Ho. Although inferred from different sets of bio- logical information, ichthyological vs. parasitological, the two views are re- markably ahke in most major points. Both views (Fig. 1) suggest that the hakes originated in the eastern North Atlantic and dispersed southward via two routes: one along the west coast of Europe and the other along the east coast of North America. These authors also arrived at the same con- clusion that hakes entered the Pacific over the submerged Panamanian Isthmus. The major discrepancy be- tween the two views lies in their ac- counts of the origin and dispersal of Argentine hake Merluccius hubbsi. Using the coevolutionary relation- ships between hakes and their cope- pod parasites, Kabata and Ho (1981) suggested that the Argentine hake was derived from the western North Atlantic stock because it shares the subspecies of copepod parasite Neo- brarhidla insidiosa lagenifor-mes with the silver hake M. bilinearis occur- ring off the east coast of North Ameri- ca. However, in contrast, Inada (1981) proposed that the Argentine hake was derived from an eastern South Pacific stock that rounded Cape Horn to reach Argentina. The Pacific ori- gin of Argentine hake was first pro- posed by Szidat(1955, 1961), but Ho (1974) questioned its validity based on his studies on the copepod para- sites of European hake M. merluc- cius, silver hake Af. bilinearis, Chile- an hake M. goyi, and shallow-water cape hake M. capensis. Contrary to the conclusion of Kabata and Ho (1981), Ho (1974) speculated that hakes originated in the western North Atlantic. Recently, Fernandez (1985) studied the parasites of a population of New Zealand hake occurring at Guafo Island, Chile (lat. 43°36'S. long. 74°43"W), and corroborated the view of Kabata and Ho on hake biogeog- raphy. Since her study methods are the same as that of Kabata and Ho (1981), the above-mentioned disagree- ment between ichthyologist and parasitologist on the origin and dis- persal of Argentine hake remains unresolved. A third method of inves- tigation is needed. Cladistic analysis is a systematic method that attempts to discover genealogical (phylogenetic) relation- ships of taxa (Hennig 1966, Wiley 1981). Since a detailed phylogenetic hypothesis for a group of organisms can and shotild serve as a basis for in- ferring the biogeographic history (Hennig 1966, Brundin 1966, Nelson and Platnick 1981, Humphries and Parenti 1986), I have used this ap- proach to reexamine the phylogenetic relationships and biogeography of hakes. This paper reports my results. Character analysis In his revision of Merluccius Rafines- que, Inada (1981) recognized 12 spe- cies of hake (Table 1) and provided detailed redescription of each spe- cies, including measurements of 28 95 96 Fishery Bulletin 88(1), 1990 Figure 1 Distriliution ami movenit'iits of hakes as proposed by Inada (1981) (—) and Kabata and Ho (1981) ( «). body parts and counts for nine meristic characters. Moreover, he added osteological information for 14 groups of bones, and discussed the distribution, biol- ogy, and fishery of ail the species recognized. It is the most complete revision ever published on the genus Merluccius. The morphological differences and diagnostic char- acters of the 12 species of hake were succinctly listed by Inada (1981, tables 37 and 38). Because quantitative variables are difficult to analyze cladistically (Pimentel and Riggins 1987), data in table 37 of Inada (1981) were not included in the present analysis. Thus, only the osteological information summarized in his table 38 was used. Steindachneria is a problematical gadiform genus that has been treated as a merluciid or as a monotypic family with uncertain affinities. Recently, Fahay (1989) presented evidence from ontogenetic and osteological grounds to support the removal of the genus from the Merlucciidae. Nevertheless, in his report on the sys- tematics of Merlucciidae, Inada (1989) concluded that the family contains four genera in two subfamilies, Steindachneriinae (Steindachneria) and Merlucciinae (Lyconus, Macruronus, Merluccius). This treatment is in line with Nolf and Steurbaut's (1989) result based on analysis of otolith features. Therefore, I included Steindachneria in my analysis of outgroup genera to determine the sister group of Merluccius. The osteological information given in Inada's (1989) Table 1 was used in my search for outgroup genera. Lotinae was selected as the outgroup to polarize the character states of the four merlucciid genera, because both the works of Dunn (1989) and Nolf and Steurbaut (1989) indicate that it is the nearest sister group of Merlucciidae. Since six of the thirteen enumerated characters in table 1 of Inada (1989: 199) show no dif- Table 1 Extant species of hakes (Mciiiicciiitt) recognized by Inada (1981), and their distribution. Species Common name Distribution merluccius European hake Europe, western North Africa •iericqnicnsis Senegalese hake Western north Africa potli Benguelan hake Mauritania to Angola capc7isiK shallow-water hake Angola to South Africa purailnxHs deep-water cape liake Nambia, South Africa hilincaris silver hake Atlantic coast of North America iilhi<1ii!i offshore hake Western Atlantic, Gulf of Mexico, Caribbean lirodiirtus Pacific hake Pacific coast of North America (ingii.tlimfuiiis Panamanian Baja California to hake Columbia 9"y> Chilean hake Peru, northern Chile hiibhai Argentine hake Argentina austraiis New Zealand Southern Argentina, hake southern Chile. New Zealand ferences in their states among the five genera under consideration, they were excluded from the present analysis. The BRANCH AND BOUND algorithm from the phylogenetic computer package PAUP (version 2.4.1, David L. Swofford, 111. State Nat. Hist. Surv., Urbana 61820), which guarantees the finding of all the most parsimonious trees, was used in this analysis. Two trees were obtained, both with a consistency index of 0.667. The one with the lower F value (= 0.174) is selected and reproduced in Figure 2. Unexpectedly, it shows that Merluccius is a sister group of the rest of the merlucciid genera. This disagrees with the phylog- eny proposed by Inada (1989) and Okamura (1989). According to the outgroup procedure as proposed by Maddison et al. (1984), all three genera— Steindach- neira, Macrouronus, and Lyconus— were included in the analysis of the "outgroup node" for polarization of character states of the hake. Dr. Tadashi Inada (Tohoku Reg. Fish. Res. Lab., Same, Hachinohe 031, Japan, pers. commun., Dec. 1988) has kindly provided osteological information for these three genera. Appen- dix 1 gives the coding of osteological characters (Inada 1981) for the hake species, and Appendix 2 shows the matrix of the character states in the twelve extant species of hakes. A new species of hake, Merluccius hernandezi, has been reported from the Gulf of Califor- nia (Mathews 1985) since revision of Inada. However, it is not included in the present analysis due to the lack of osteological information. Ho: Phytogeny and biogeography of Merluccius 97 Figure 2 Cladogram showing hypothesized relationships of the merlucciid genera. Character codes: 1 = V-shaped crest on skull present; 1' = V-shaped crest on skull absent; 2 = foramen for trigemino-fascialis nerve present; 3 = shape of anterior parasphenoid vertical; 4 = up- per window on suspensorium closed; 5 = two dorsal fins; 5' = one dorsal fin; 6 = one pseudospine in dorsal fin; 7 = two pseudospines in dorsal fin; 8 = caudal fin tapering. Phylogeny An unusually large number of trees (cladograms) were obtained, 188 in total, all with a consistency index of 0.522. The F value of these most parsimonious trees ranges from 0.277 to 0.482. However, there is only one tree with the lowest (best) F value, reproduced in Fig- ure 3. Since the development of host specificity in parasites is the result of coevolution between the host and its parasite, an hypothesis of the host phylogeny should also apply to the phylogeny of its specific parasite. In other words, host-specific parasites can be useful to corroborate host phylogeny. Brooks et al. (1986) suggested that the lower the F-ratio ( = F value), the greater is the degree of his- torical constraint on the data. In this regard, the tree with the lowest F value (Fig. 3) should exhibit the highest degree of congruence with the parasite data set. To check this congruency, a parasite summary cladogram needs to be constructed. Figure 4 is such a cladogram, generated by replacing the hake species on the cladogram with each of their respective sub- species of Neobrachiella insidiosa, a host-specific copepod parasite of hakes. Since Chondracanthus palpifer WUson was recently found by Villalba and Fer- nandez (1985) to parasitize not only Merluccius but also another merlucciid (Macruronus magellanicus Lonn- berg) in Chilean waters, it can no longer be treated as aMer/McciMS-specific parasite as previously proposed. Accordingly, only the three subspecies oi Neobrachiella insidiosa are considered. Figure 3 Cladogram showing hypothesized relationships among species of Merluccius. For character codes 1-23, see Appendix 1. A i\ .:. i\ A A A A A A A A ^ Pacilici A l3geniIormis A 'fsidt Figure 4 Parasite summary cladogram showing relationships among sub- species of Neobrachiella insidiosa inferred from hypothesized rela- tionships among species of hake {Merluccius). Arrows indicate the changes (evolution) of parasites. As in Figure 4, the result agrees, indicating that the hake-specific Neobrachiella insidiosa had changed (evolved) from A'^. insidiosa lageniformes (western Atlantic form) to A'', insidiosa insidiosa (eastern Atlan- tic form) on one hand, and from the western Atlantic form (original form) to A'^. indisiosa pacifica (Pacific form) on the other. This scheme of evolution for Neobrachiella insidiosa contradicts the one proposed by Kabata and Ho (1981) but, nevertheless, agrees with the hake-parasite evolution proposed by Ho (1974). The phylogenetic hypothesis predicts a Pacific form of N. insidiosa, but it is yet to be found in M angustimanus. The discovery of A'', insidiosa pacifica on the Panama- nian hake would strongly support the cladogram shown in Figure 3. 98 Fishery Bulletin 88|l). 1990 Figure 5 Examples of some most parsimonious cladograms of hakes {Merlur- ciiis). All have 23 steps and a consistency index of 0.522. Abbrevia- tion of species names: ALBl = albidus; ANGU = angustimanus; AUST = aiistralis; BILI = bilinearis; CAPE = capensis; HUBB = hubbsi; MERL = merlucciu.r, PARA = paradoxus; POLL = polli; PROD = pradurtu.t; SENE = Keveiidlnisis. With currently available information, the tree in Fig- ure 3 is considered to be the better hypothesis than the other existing h>T30theses of hake phylogeny. Admit- tedly, it has a high ratio of homoplasy (17/23 or 73.9%). Such frequent homoplasy is somewhat undesirable in a cladistic analysis. Nevertheless, using different sets of characters (number of gill-rakers, vertebrae, and fin rays), Inada (1981) identified five groups in his key to the species of Merluccins, with four species {au^tralis, bilinearis, hubbsi, polli) in two groups, two species (gayi, senegalensis) in three groups, and one species (capensis) in four groups. This repeated sharing of many characters (high frequency of homoplasy) seems to be the norm in Merluccius. Close examination of each of the 188 most parsi- monious trees revealed that European hake M. merluc- cius, Senegalese hake M. senegalensis, Benguelan hake M. polli, shallow-water cape hake M. capensis, deep- water cape hake M. paradoxus, Argentine hake M. hubbsi, and New Zealand hake M. aiistralis are always included in a monophyletic group. However, relation- ships among the three eastern Pacific hakes— Chilean hake M. gayi, Panamanian hake M. angtuitimanus, and Pacific hake A/, prnductus—and the two western North Atlantic species— silver hake Af. bilinearis and offshore hake M. albidus are not as consistent. Some represen- tative trees are reproduced in Figure 5 for comparison. Offshore hake (ALBI, Fig. 5) appears to occupy the most plesiomorphic node in almost all cases. This is an Figure 6 Strict consensus tree of hakes showing information common to the 188 equally parsimonious trees. For abbreviation of species names, see Figure 5. indication that offshore hake is the extant species closest to the ancestral form. The strict consensus tree in Figure 6 was obtained using the CONTREE algorithm contained in the PAUP program. It summarizes the point of agreement in all 188 equally parsimonious trees. Remarkably, the mono- phyletic relationships of the seven species of hakes occurring off the coasts of Europe, Africa, and eastern South America are also supported by this consensus tree. Biogeography According to the principle of vicariance biogeography (Nelson and Platnick 1981. Humphries and Parenti 1986), the pattern of spatial distribution attained by the hakes can be deduced from the phylogenetic hypothesis. The area summary cladogram in Figure 7 illustrates how the present pattern of hake distribution was attained. According to the progression rule of Hen- nig (1966), it implies that the ancestral hake, residing in the western North Atlantic, diverged into two line- ages. One gave rise to the modern offshore hake (ALBI, Fig. 7), and the other formed the species A, which is the ancestor of all the remaining extant hake species. Species A vicariated into two populations: the northern population (C, Fig. 7) whose descendants later occupied the Pacific Ocean, and the southern popula- tion (B, Fig. 7), whose descendants gave rise to all the hakes in the western South Atlantic and entire eastern Atlantic. I interpret the cladogram to mean that (1) hake originated in the western North Atlantic, (2) migrated southward and then eastward in the Atlantic, and (3) entered the Pacific in two ways, over the submerged Ho; Phytogeny and biogeography of Merluccius 99 Figure 7 Area summary cladogram of hakes (Merliicrius) with ancestral species (A-H). For abbreviation of species names see Figure 5. Panamanian Isthmus and through the Drake Passage. Assertion (1) agrees with the proposal of Ho (1974) con- cluded from a work on host-specific copepod parasites of hake, but it contradicts Inada (1981) and Kabata and Ho (1981) who suggested otherwise. Assertion (3) is roughly the same as those suggested by Inada (1981) and Kabata and Ho (1981), differing only in the details. However, the 2nd assertion is a new discovery; this track has never been proposed. Inada (1981) speculated that hakes had migrated out of the western Atlantic only once, when they dispersed to the southern hemisphere during the Pliocene over the submerged Panamanian Isthmus. He believed that the hakes could not have migrated to the South Atlan- tic along the coast of Brazil because the low salinity in the estuarine area of the Amazon River would have acted as a natural barrier to their -dispersal. However, the Amazon River estuary was not an effective deter- rent to hake dispersal in the early Tertiary. In work on the coevolution between freshwater stingrays and helminth parasites, Brooks et al. (1981) concluded that the Amazon Basin became a separate eastward-flowing entity in the Miocene. Moreover, Damuth and Kumar (1975) reported that the Brazilian plate developed a geosyncline downfold when the Andes began to fold up in the Pliocene, and, as the result of this lowering of the eastern side of the plate, the water from the land- locked sea (of the Pre-Amazonas) started to flow eastwards to the Atlantic. Therefore, during and prior to the Miocene the salinity in this area could not have been as low as it was in the Pliocene and the ancestral hakes could have migrated to the South Atlantic along the coast of Brazil. If true, migration pathways of hake would be different from those pro- posed by Inada. Both the works of Inada (1981) and Kabata and Ho (1981) indicated that European hake M. merluccius occur in the region of the North Atlantic inhabited by the original stock (Fig. 1). However, the phylogenetic hypothesis (Fig. 3) implies that European hake is the most derived species among the eastern Atlantic hakes, and, consequently, could not be the one occupying the original habitat (Hennig 1966, Wiley 1981). Before discussing further the biogeography of hake inferred from the phylogenetic hypothesis given in Fig- ure 7, some points about the nature of the movement and range expansion of hake need to be explained. Hakes of the genus Merluccius are coastal inhabi- tants occurring in the waters above the continental shelf and slope. Most species exhibit a seasonal on- shore-offshore bathymetric migration correlated with water temperature. Adults and juveniles move inshore during the spring. When winter cooling occurs on the shelf, they migrate to warmer waters on the continen- tal edge and slope (Grinols and Tillman 1970). Addi- tionally, the silver hake and Pacific hake apparently undertake a latitudinal migration (Bailey et al. 1982, Leim and Scott 1966, Nelson 1969). During the fall they move southward and then return to more northerly waters during the late spring. Clearly, paleotempera- ture changes in the ocean could be one of the factors in altering the limits of the distribution of hake in the past: they expanded onto the continental shelf or slope of lower latitudes when there was a cooling trend, and retreated to the waters of higher latitudes when the cooling trend subsided. The copepod parasite Neohrachiella iyjsidivsd Ingeni- formes has been recorded from widely separated species of hake, e.g., the silver hake off the coast of Florida and the New Zealand hake in the western South Pacific (Kabata and Ho 1981). Since this parasite is found mostly on the adult hake, rarely on juveniles and never on the larvae, the dispersal of this parasite from North America to New Zealand via the South Ameri- can coast must have been effected by the adults. Besides, this parasite belongs to the family Lernae- opodidae which is noted for a brief stage of free-living during development. It is logical to assume that adult hakes, not larvae, are responsible for extending their range. The offshore hake M. albidus, primarily an inhabi- tant off the U.S. east coast and in the Gulf of Mexico and Caribbean (Inada 1981), does not share its clade with any extant hake (Fig. 7). The distributional ranges of its ancestor must have waxed and waned like other ancestral hakes with the changes of ocean temperature and sea level. However, it is very likely that this stock of hake stayed mostly in the Gulf of Mexico and 100 Fishery Bulletin 88(1)^ 1990 Caribbean for most of the time since its evolution in the Eocene. Geologic events that cause fragmentation of the contiguous, ancestral distribution are considered the major means of distributional pattern formation (Nelson and Platnick 1981). Although not all vicariant events are identifiable at present, those known geologic events that could have produced the present pattern of hake distribution are explored below. Range expansion and vicariance of species B The closeness between fossil members of the genera Palaeogadus and Merluccius has led Fedotov and Ban- nikov (1988) to speculate that hake originated from a species related to Palaeogadns intergerinus in the Mid- dle Miocene. However, since fossil records indicate the minimum age, the discovery of fossil Merluccius in- feni:^ from the Tethys area of the Soviet Union in the Middle Oligocene deposit may not necessarily signify that Merluccius originated in the Middle Oligocene. The genus could have made its appearance much earlier. It is unclear when the hakes living on the western North Atlantic seaboard (off North America) diverged into two stocks, but species B must have expanded its range southward and moved into the continental shelves of low latitudes in the late Eocene (about 40 million years ago [MA]), an age characterized by a major decrease in annual temperature (Frakes 1979, Wolfe and Poore 1982). According to Savin et al. (1975), equatorial temperatures may have been as low as 20°C throughout the Oligocene (35-25 MA); conse- quently, species B most likely migrated into Brazilian waters during this long cooling period. In his discussions on Caribbean biogeogi'aphy, Rosen (1975, 1978, 1985) concluded that there was an isth- mian land bridge between North and South America in the late Cretaceous or early Cenozoic. A similar in- tercontinental link was also suggested by Savage (1966, 1982), Axelrod (1975), and White (1986) in their bio- geographic studies on the herpetofauna, plants, and silverside fishes, respectively. Alternatively, recent geologic reports suggested that this isthmian link was most developed sometime in the Eocene (Coney 1982, Pindell and Dewey 1982). In that case during their southward range expansion, hakes of the species B would have been prevented from entering the Pacific. The area cladogram (Fig. 7) obtained from the adopted phylogenetic hypothesis (Fig. 3) indicates that species D had expanded across the Atlantic Ocean to become the common ancestor of hakes in the eastern Atlantic. The occurrence of such range expansion would invoke a geologic event with a shallow-water (<200 m) connection between South America and Africa, because, from the parasitological point of view. it is the demersal adult hakes (and not the planktonic larvae) that are responsible for this range expansion. Since the fossil hake Me^iuccius inferus is known from the middle Oligocene deposit in Europe (Svetovidov 1940), this eastward crossing of the Atlantic must have occurred in the early Tertiary. However, no such ex- tension of continental shelf in the Tertiary has been proposed by geologists, except Vail et al. (1977) who recognized five major lowstands in the Tertiary (Mid- Paleocene, Early-Mid Eocene, middle Late Oligocene, Late Miocene, and Late Pliocene-Early Pleistocene). According to them, during these lowstands the sea level fell below the edge of the continental shelf in most regions; thus, the eastward expansion of species D across the Atlantic Ocean in the low latitudes was possi- ble. After reaching the eastern Atlantic, it spread north and south along the coast of Africa. In the Early Miocene between 22 and 20 MA, the pro- longed cooling trend initiated in the late Eocene was reversed (Haq 1982). Temperatures in low latitudes rose to modern values and might have exceeded 30°C in some areas. This change in global climate, even if it did not affect the adult hakes because of their pref- erence for deep water (Grinols and Tillman 1970), could have effectively prevented hatching of their eggs (nor- mally at 11-14°C) and the normal growth of their epipelagic larvae (found usually at 40-60 m); thus, the hakes were prevented from occupying waters of low latitudes. This unusual warming of the tropical waters can be viewed as a vicariance event that split species D into a northern segment which later developed into another ancestral form (species E, Fig. 7), and a st)uth- ern segment representing the common ancestor (spe- cies F, Fig. 7) of the shallow-water and deep-water cape hake. Species E must have invaded again the waters of low latitudes during another cooling, which started in early Middle Miocene (about 15 MA) when a major enlargement of the East Antarctic ice-sheet developed (Savin 1977, Woodruff et al. 1981). This would account for the presence of Benguelan hake M. polli, one of the decendants of species E, in the tropical waters of the eastern Atlantic. However, further divergence of spe- cies E is unclear given the present knowledge of geo- logic history. The ancestor of New Zealand hake M. australis is a sister group of species D. It must have occupied the continental shelf off Argentina when the latter crossed the Atlantic. The extinct hake Merluccius fimbriatus, known from the Miocene deposit in Victoria, Australia (Stinton 1958), may be a descendant of this stock since the Drake Passage had remained open since the Oligocene, between 29 and 28 MA (Haq 1984). Accord- ing to Inada (1981), there are two distinct populations of New Zealand hake: the New Zealand population living in New Zealand waters, and the Patagonian Ho. Phytogeny and biogeography of Merluccius 101 Figure 8 Dispersal routes of hakes infeiTed from the area summary cladogram in Figure 7. population living on both sides of the southern part of South America south of 40° S. Interestingly, both populations, in spite of being distantly separated by the Pacific, carry the same species of copepod parasite— Neobrachiella insidiosa lageniformes— on their gills (Kabata and Ho 1981), a clear indication that the New Zealand population must have originated from the hakes in South American waters. Range expansion and vicariance of species C According to the proposed history of the relative motion of the plates in the Gulf of Mexico/Caribbean region, Pindell and Dewey (1982) concluded that the large-scale eastward migration of the Caribbean plate started in the Oligocene (about 36 MA) and eventually placed the Greater Antilles in their present positions. It is assumed that species C, as in species B, expanded southward prior to the initiation of this Caribbean plate movement. However, being a northern population, it did not reach as far south into the South Atlantic as Species B. Instead, it expanded into the Pacific when the Panamanian seaway was opened in the Oligocene. As with the hake in the eastern Atlantic, species C was prevented from inhabiting the waters of low lati- tudes when the climate warmed in the Early Miocene. By this restriction, species C was divided into an Atlan- tic population, which eventually gave rise to silver hake M. bilinearis and a Pacific population (species G, Fig. 7) which became the ancestor of the three eastern Pacific species. The fossil remains of the Pacific hake M. pro- dudus are common in the Pliocene deposit of Califor- nia (Fitch 1969, Fitch and Reimer 1967, Zinsmeister 1970), indicating that the divergence of species G into species H and Pacific hake (M. prodiietus or its imme- diate ancestor) took place either in Miocene or Pliocene. Species H was very likely confined to the North Pacific off the coast of Mexico until the Pliocene when the Panamanian isthmus was reestablished at about 3 MA (Haq 1984). Panamanian hake M. angustimanus is found from Baja California to Colombia, but not in the Caribbean; therefore, species H must have moved southward after closing of the Panamanian seaways. The vicariance event responsible for the separation of species H into Panamanian hake and Chilean hake is unclear, but the influence of the Ice Ages during the late Pliocene and Pleistocene may have played a role. Conclusion The phylogenetic hypothesis of hake (Fig. 3) presented here is the most parsimonious scheme derived from the cladistic analysis of the osteological data of Inada (1981). It is congiiient with the scheme of evolution pro- posed by Ho (1974) for the hake-specific copepod parasites. This is a testable model; it can be corrobor- ated by including more anatomical data (e.g., muscu- lature), ontogeny, karyology, DNA sequences, and allozymes. Parasitological information from proto- zoans, helminths, and crustaceans can also provide fur- ther corroboration, particularly when hake-specific parasites are found among them. The vicariance model proposed for the hake biogeog- raphy is based on the adopted most-parsimonious tree that shows congruence with the available parasito- logical information. The inferred pattern of biogeog- i-aphy, particularly the track across the Atlantic in the low latitudes (Fig. 8), is yet to be reported for demer- sal fishes. However, it should be mentioned that Van der Spoel and Heyman (1983) considered that the planktonic faunas of the Panama Passage and South American inland sea as ancestral to all eastern Atlan- tic distant neritic taxa, and proposed a similar eastward dispersal. The ranges of ancestral hake waxed and waned with the fall and rise of paleoceanographic temperatures in the Tertiary, dispersed during the cooling periods and fragmented by the development of warming trends. Development of major lowstands in sea level is also viewed as an effective vicariant event. This model can be tested with additional details of the plate-tectonic models and paleoceanographic-climatic history. Works on other marine life with similar distributions would test the validity of the three inferred general tracks. Both the phylogenetic hypothesis and biogeographic model differ from current views, but they are viewed as the better explanation of available data. They are presented here as a working model subject to modifica- tion as more exact infoi-mation becomes available. 102 Fishery Bulletin 1990 Acknowledgments I wish to thank Tadashi Inada of Tohoku Regional Fisheries Research Laboratory, Japan, for providing the indispensable osteological data oi Steindachneria, Macrouronus, Lyconus, and Gadomus; and Masahiro Dojiri of Hyperion Treatment Plant, Playa del Rey, California, and Gregory B. Deets of Institute of Para- sitology, California State University, Long Beach, for their suggestions on the first draft of this paper. I am particularly indebted to Daniel M. Cohen of the Natural History Museum of Los Angeles County for providing information and literature on the gadiform fishes. Both Daniel Cohen and Tadashi Inada have also read and commented on the first draft of this paper. Five anon- ymous reviewers gave their helpful comments on the manuscript. However, the author bears the sole respon- sibility for the interpretation presented herein. The preparation of this work was supported by a grant from California State University, Long Beach Foundation. Citations A.xelrod, D.I. 1975 Evcilution anri biogeography of Madrean-Tethyan scleni|)hyll vegetation. Ann. Mo. Bot. Garri. 62:280-334. Bailey. K.M., R.C. Francis, and P.R. Stevens 1982 The life history and fishery of Pacific whiting, Merliir- niis productus. Calif. Coop. Oceanic Fish. Invest. Rep. 22:81-98. Brooks, D.R., T.B. Thorson, and M.A. Mayes 1981 Fresh-water stingrays (Potamotrygonidae) and their helminth parasites: Testing hypothesis of evolution and coevolu- tion. In Funk, V.A.. and D.R. Brooks (eds.), Advances in cladistics, p. 147-175. New York Bot. Card., NY. Brooks, D.R., R.T. O'Grady, and E.O. Wiley 1986 A measure of the information content of phylogenetic trees, and its use as an optimality criterion. Syst. Zool. .3."):ri71-.581. Brundin, L. 1966 Transantarctic relationships and their significance, as evidenced by chironomid midges. Kungl. Svenska Vetenskap. Handl. 11, 472 p. Coney, P.J. 1982 Plate tectonic constraints on the biogeography of Mid- dle America and the ('aribbean region. Ann. Mo. Bot. Card. (;) zoogeography. Proc, 3rd Int. ('ongr. Parasitol. 3:1633-16.34. Humphries, C.J., and L.R. Parenti 1986 Cladistic biogeography. Clarendon Press. Oxford. Inada. T. 1981 Studies on the merluccid fishes. Bull. Far Seas Fish. Res. Lab. (Shimizu) 18, 172 p. 1989 Current status of the systematics of Merlucciidae. In Cohen, D.M. (ed.). Papers on the systematics of Gadiform fishes, p. 197-208. Nat. Hist. Mus. Los Ang. Cty., Sci. Ser. 32. Kabata, Z., and J.S. Ho 1981 The origin and dispersal of hake (Genus Merlucciua: Pisces: Teleostei) as indicated by its copepod parasites. Oceanogr. Mar. Biol. Annu. Rev. 19:381-404. Leim, A.H., and W.B. Scott 1966 Fishes of the Atlantic coast of Canada. Bull. Fish. Res. Board Can. 155:1-485. Maddison. W.P., M.J. Donoghue. and D.R. Maddison 1984 Outgroup analysis and parsimony. Syst. Zool. 33: 83-103. Mathews. C.P. 1985 Meristic studies of the Gulf of California species oiMerhw- cms, with a description of a new species. J. Nat. Hist. 19; 697-718. Nelson. G., and N. Platnick 1981 Systematics and biogeography, cladistics and vicariance. Columbia Univ. Press, NY. Nelson. M.O. 1969 Availability of Pacific hake {Merlun'ius frmdml-ux) related to the harvesting process. FAO Fish. Rep. 62(3) E/34: 619-644. Ho: Phytogeny and biogeography of Merluccius 103 Nolf. D., and E. Steurbaut 1989 Evidence from otoliths for establishing relationships within Gadoformes. In Cohen. D.M. (ed.). Papers of the systematics of Gadiform fishes, p. 89-112. Nat. Hist. Mus. Los Ang. Cty., Sci. Ser. 32. Okamura, O. 1989 Relationships of the suborder Macrouroidei and related groups, with comments on Merlucciidae and Steindachneria. In Cohen, D.M. (ed.). Papers of the systematics of Gadiform fishes. Nat. Hist. Mus. Los Ang. Cty., Sci. Ser. 32. Pimentel, R.A., and R. Riggins 1987 The nature of cladistic data. Cladistics 3(3):201-209. Pindell, J., and J.F. Dewey 1982 Permo-Triassic reconstruction of western Pangea and the evolution of the Gulf of Mexico/Caribbean region. Tectonics 1:179-211. Rosen, D.E. 1975 A viacariance model of Caribbean biogeography. Syst. Zool. 24:431-464. 1978 Vicariant patterns and historical explanation in biogeog- raphy. Syst. Zool. 27:159-188. 1985 Geological hierarchies and biogeographic congruence in the Caribbean. Ann. Mo. Bot. Card. 72:636-659. Savage, J.M. 1966 The origin and history of the Central American herpeto- fauna. Copeia 1966:714-766. 1982 The enigma of the Central American herpetofauna: Dispersals or vacariance? Ann. Mo. Bot. Gard. 69:464-547. Savin, S.M. 1977 The history of the Earth's surface temperature during the last 100 million years. Annu. Rev. Earth Planet. Sci. 5: 319-355. Savin, S.M., R.G. Douglas, and F.G. Stehli 1975 Tertiary marine paleotemperature. Bull. Geol. Soc. Am. 86:1499-1511. Stinton. F.C. 1958 Fish otoliths from the Tertiary strata of Victoria, Aus- tralia. Proc. R. Soc. Vic. N.S. 70:81-93. Svetovidov, A.N. 1940 The problem of distribution of the Gadidae and of the other families of Gadiformes. Bull. Soc. Nat. Moscow, Sec. Biol. 9Ul):,50-60. Szidat, L. 1955. La faima parasitos de "Merlurrius hiihhsi" como caracter auxiliar para la solucion de los problemas sistematicos y zoogeograficos del genero "Merluccius" L. Comun. Inst. Nac. Invest. Cienc. Nat., B. Aires 3, 54 p. [in Spanish]. 1961 Versuch einer zoogeographie des SudAtlantik mit Hilfe von Leitparasiten der Meeresfische. Parasitol. Schriften. 13, 98 p. [in German]. Vail, P.R.. R.M. Mitchum, and S. Thompson 1977 Seismic stratigraphy and global changes of sea level, pt. 4: Global cycles of relative changes of sea level. Mem. Am. Assoc. Petrol. Geol. 26:83-97. Van der Spoel, S., and R.P. Heyman 1983 A comijarative atlas of zooplankton. Biological patterns in the oceans. Springer- Verlag. NY. Villalba, C, and J. Fernandez 1985 Contribucion al conocimiento de la familia Chondracanth- idae en Chile (Copepoda: Poecilostomatoida). Gayana Zool. 49(l-2):31-58. White, B.N. 1986 The Isthmian link, antitropicality and American biogeog- raphy: Distributional history of the Atherinopsinae (Pisces: Atherinidae). Syst. Zool. 35(2):176-194. Wiley, E.O. 1981 Phylogenetics: The theory and practice of phylogenetic systematics. J. Wiley and Sons, NY. Wolfe, J. A., and R.Z. Poore 1982 Tertiary marine and nonmarine climatic trends. /;/ Climate in earth history, p. 154-158. Natl. Acad. Sci. Press, Wash., DC. Woodruff, F.. S.M. Savin, and R.G. Douglas 1981 Miocene stable isoto|ie record: A detailed deep Pacific Ocean study and its paleoclimatic implications. Science (Wash., DC) 212:665-668. Zinsmeister, W.J. 1970 A late Pliocene macrofossil fauna of Newport Beach, Orange County, California. Bull. South. Calif. Acad. Sci. 69:121-125. Coding of osteological 1981). Appendix characters fo 1 ■ the hake species (Imada Character States of character (code) Gill raker length 0: short 1: medium {10)(19)(23) 2:long(ll)(20) Size of teeth on jaws 0: small (22) 1: large (4) Hyomandibular intermuscular process length 0: none 1: short (1)(21) 2: long (5) Postcleithrum head 0: large (7) 1: medium (2)(14) 2: small (9)(12)(16) Urohyal 0: thin 1: thick (17) Cleithrum groove 0: none 1: deep (3) (15) 2: shallow (13)(18) Infraorbital foramen width 0: wide 1: medium (6) 2: narrow (8) 104 Fishery Bulletin 88(1), 1990 Appendix Matrix of character states in the Iwe 2 Ive e xtant species of hakes Species Character 1 2 3 4 5 6 7 8 9 10 11 12 Gill raker length 0 Size of teeth on jaws 1 Hyomandibular intermuscular process length 2 Postcleithrum head (.1 Urohyal 0 Cleithrum groove 1 Infraorbital foramen width 2 1 0 2 0 0 1 0 nus 0 0 o 0 0 1 2 10 11 12 1 1 1 () 0 1 1 .'/".'/! it list SI 2 1 >> 0 II 1 1 Its 2 ] 1 2 0 1 0 0 1) 1 o 0 1 0 2 (1 1 ■? (1 '> 0 2 0 1 1 (1 1 (1 2 (1 1 1 () ') 0 0 1 2 1 1 1 0 1 2 2 (1 1 1 1 merluccius 4 niitni^is 7 nlhidus 2 se)wgalensis 5 parudoxii.-; 8 produdus 3 polli 6 hilinearin '.) tnigustim.a Abstract.- During 1986, the Na- tional Marine Fisheries Service be- gan conducting long-term research vessel surveys to determine trends in population size of dolphin stocks taken incidentally by tuna piu-se sein- ers in the eastern tropical Pacific. Line transect methodology was used by observers aboard two vessels for 120 days each. We assumed the vari- ability associated with the abundance estimates would be relatively con- stant during the sampling period and investigated (1) annual changes in population size of the northern off- shore spotted stock that could be detected within a 5-year (six survey) sampling period, and (2) the number of years required to detect a 10% an- nual decline with a and /? error levels of 10%. The abundance estimate of the northern offshore spotted dol- phin stock using the first year's data was 929,000 animals. After 5 years, a minimum 17.6% annual decline could be detected during which 62% of the stock would decrease. A 10% annual decline can be detected in a minimum of 8 years. Data from sub- sequent surveys will be investigated that may improve our ability to de- tect smaller annual declines. Monitoring Trends in Dolphin Abundance \n the Eastern Tropical Pacific Using Research Vessels Over a Long Sampling Period: Analyses of 1986 Data, The First Year Rennie S. Holt Stephanie N. Sexton Southwest Fisheries Center National Marine Fisheries Service, NOAA La Jolla, California 92038 Manusci-i|it accepted 21 Septeml)er 1989. Fishery Bulletin, U.S. 88:105-111. The National Marine Fisiieries Ser- vice (NMFS) is responsible for assess- ing the status of dolphins taken inci- dentally by tuna purse seiners in the eastern tropical Pacific (ETP) (Rich- ey 1976). The status of spotted dol- phins Stenella attenuata is of special concern because it is the major spe- cies taken by the fishery (Smith 1979). Of the three stocks of spotted dol- phins, the northern offshore stock has been fished more frequently than any other stock. The four stocks of spinner dolphins S. longirostris, and the four stocks of common dolphin Delphinufi delphis, are also taken. The four stocks of striped dolphin 5. coeruleoalha and the Eraser's dolphin Lagciiodelplns hosei are occasionally caught (Holt and Powers 1982). These five species are herein grouped and termed target species. NMFS conducted assessments of target populations in 1976 (SWFC 1976) and again in 1979 (Smith 1979) using absolute stock abundance. The validity of the absolute estimates de- pended on several assumptions (i.e.. all schools located directly on the trackline were detected, schools did not respond to the ship before being detected, etc.). Unfortunately, not all of these underlying assumptions were met (Holt and Cologne 1987, Holt 1987). Therefore, in 1984 Congress amended the Marine Mammal Pro- tection Act and mandated that an alternative approach for assessing stocks be used that was less sensitive to biases. By using consecutive popu- lation estimates and establishing the first estimate as a base to which all subsequent estimates would be rela- tive, we can detect trends in stock sizes over a long sampling period. These relative estimates can provide an assessment of stock condition if the biases in the abundance estimates are consistent over the sampling period. In 1986, NMFS initiated a research program to monitor dolphin popula- tions in the ETP which would utilize two research vessels for at least 5 years during which six sui-veys would be conducted. The research design (Holt et al. 1987) indicated that a 10% annual rate of decrease in northern offshore spotted dolphins could be detected (a total 41%) decrease over six surveys). Herein, we present the population estimates for the first year's survey data. We also discuss effects of several factors on these base estimates, and, assuming data for subsequent surveys will have the same level of precision (coefficient of variation levels) as the first year, ex- amine changes in population sizes that can be detected in 5 years or, conversely, the number of years re- cjuired to detect various levels of change. 105 106 Fishery Bulletin 1990 LONGITUDE Figure 1 Tracklines traversed by the NOAA RV Darid Starr Jordiin (solid) and McA rthur (dash) during the 198G dol- phin survey, eastern tropical Pacific. Tracklines generated using noontime positions. Materials and methods Study area and survey coverage The study area was described by Au et al. (1979) (Fig. 1). We partitioned the area into four strata: inshore, middle, and west located north of 1°S, and a south stratum. These strata were selected based upon pre- liminary examinations of historical distributions of dolphin stocks and oceanographic features. The NOAA research vessels David Starr Jordan and Mc Arthur traversed predetermined tracklines in the ETP during 29 July-5 December and 6 December, respectively (Fig. 1). Each vessel spent approximate- ly 120 days at sea. Detailed operations, survey pro- cedures and preliminary data summaries for each vessel are presented elsewhere (Holt and Sexton 1987, Holt and Jackson 1987). On each ship, two observers used 25 x l)inoculars located on each side of the ship to search from directly ahead to abeam of their respective sides of the ship. A third observer served as data recorder and searched directly ahead of the ship when not recording data. Two teams of three observers each alternately occupied the three duty stations. Each team was on duty for 2-hour shifts. During each shift, members spent approximately equal time occupying each duty station. Whenever possible, schools were approached and observers re- corded independent "best" estimates of school size. If an observer could not obtain a best estimate, a "mini- mum" estimate was made. Independent estimates were averaged to obtain mean minimum and best estimates. Abundance estimation Estimates of population abundance of the target spe- cies (N,j) are computed as (Holt and Powers 1982): N,j = 1 \D,SaP,kP,kA,IA,,][A,j, ^P\,A',,,] A-=l where Dj; = estimate of density of all dolphin schools, both identified and unidentified to species, in area A', S,i; = estimate of mean size of target schools in area A', Pi I; = estimate of proportion of dolphin schools which are target schools in area A', P,i; = estimate of proportion of individuals of species i in target schools in area A-, P' II = estimate of proportion of individuals of stock j of species i in target schools in overlap region containing two stocks of species i (overlap region discussed in text), Al- = total area inhabited by the target species in area A, A, I; = area inhabited by species i in area A, Ajjt = area inhabited by species ?', stock,/, in area A-, and A'lji; = area inhabited by species i, stocky, in over- lap region of area k. The variance of N,j was calculated using boot- strapped methods. For each stratum, the number of legs (segment of time during which all sighting Holt and Sexton Dolphin abundance in eastern tropical Pacific 107 conditions were consistent) of searching effort was tabulated, and then effort legs equal to that number were randomly selected with replacement. This effort and the associated sightings were used to calculate school density, school size, species proportions, and finally estimates of Nij. This process was repeated 100 times. The bootstrapped variance oi N,i for each stock was calculated using these 100 estimates. Formulae used to estimate school density are from Burnham et al.'(1980). Holt (1985, 1987), and Hayes and Buckland (1983). The Fourier series (Grain et al. 1979) and hazard rate (Hayes and Buckland 1983, Buckland 1985) models both provided adequate fits to the perpendicular (sighting) distance data; however, the hazard rate model was used because, unlike the Fourier series model, it does not require subjective selection of the number of terms in the model and, therefore, could be used in the bootstrapped procedures. Of schools containing both target and non-target species, only the proportion of individuals of the target species was used in the school size estimates. Estimates of the proportion of all dolphin schools that were target schools (Pi I;) were calculated using formulae from Holt and Powers (1982). Formulae to estimate the pro- portions (Pa) of the number of individuals for each species of all target individuals are given by Barlow and Holt (1986). All species of dolphins encountered in the study area were included in the density analyses. Estimates were calculated using only schools containing 15 or more animals. Smaller schools were not used because we believe small schools both on and off the ships' track- lines may be difficult to detect, especially during rough weather and may have been missed at a variable rate depending on prevailing weather conditions (Holt and Powers 1982). Schools detected at increasing distances from the trackline tend to include disproportionately more large schools because there is a direct correlation between the size of a school and the probability of it being detected (Drummer 1985). This biases school size esti- mates upward and species proportions toward species which tend to occur in large schools. We attempted to adjust for this bias by weighting school size and species- proportion estimates by the inverse of the logarithm of school size (Holt and Powers 1982). Schools for which there were no "best" estimates of size were not used in the school-size or species-proportions calculations. Because a 3.7 km (2.0 nm) truncation point provided the best fit of the hazard model to the data, only schools detected within 3.7 km perpendicular distance of the trackline were used to estimate school density. Schools detected greater than 3.7 km have little affect on the density estimates, and the perpendicular distance distributions of schools at greater distances were "spiked". These spikes are a result of the ob.servers' tendency to round-off estimates of radial distances and sighting angles of schools detected at large distances from the vessel in multiples of 0.5 nm or 5°, respectively. Some stocks of the same species have overlapping ranges (A',i^.). These overlapping stocks include (1) coastal and northern spotted, (2) eastern and whitebelly spinner, and (3) Baja Neritic and northern common dolphins (Perrin et al. 1984). The relative number of dolphins of each overlapping stock (P,^) was calcu- lated for data pooled over the range and, if applicable, over strata. The relative proportions of coastal and northern spotted, and of eastern and whitebelly spin- ner stocks, within their respective areas of overlap, were calculated as the average of their relative abun- dances (percent occurrence). Few data were available to determine relative proportions of the overlapping Baja Neritic and northern common dolphins. Therefore, population estimates for Baja Neritic were combined with northern common dolphins. The area inhabited by each target species (/I,/,) and each stock (/I,,/,) used to calculate the population abundance estimates were those defined liy Au et al. (1979) and Perrin et al. (1984). The size of each stratum (Aj;) and the size of the area occupied by each stock in each stratum were calculated liy counting the numlier of 1° quadrilateral squares in the stratum at each degree of latitude; partial squares were approximated. Next, the number of 1° squares was multiplied l:)y the area in a 1° square for that latitude as described by Holt and Powers (1982). Detecting trends in abundance The variability associated with the population abun- dance estimates of northern offshore spotted dolphins during this first survey was examined to determine changes which may be detected from subsequent surveys using methods presented by Holt et al. (1987) and Gerrodette (1987). For Type I "(a) and Type II ((i) error levels of 0.10, we computed the number of years required to detect a minimum annual decrease of 0.10 and the minimum annual decrease in northern offshore spotted dolphins which could be detected at the end of the planned 5-year (6-survey) period. In addition, we calculated the total population decrease that would occur over the 5-year period given that annual level of decrease. Results During the entire survey, observers aboard both vessels searched 30,339 km and detected 1150 marine mam- mal schools. Dolphins were present in 749 of these 108 Fishery Bulletin 88(1). 1990 Table 1 Summary ol 198(i dolphin survey in eastern ti opical Pacific. Da ta were truncatec at 3.7 km perpendicular distance. Schools with less | than 15 animals were omitted from analyses School sizes and species proportions weighted by inverse of logaritl im of school size. Effort collected during sea states 0-5 were included in analyses Data summed for both vessels. Ti tal estimates calculated using effort summed over all four strata. Inshore Middle West South Total Survey area (1000 km-') 5,693 3,798 5,298 4,3.59 19,148 Percent of total survey area 30 20 28 22 100 Trackline searched (km) 11,889 7.846 3,877 4,056 27,669 Percent of searching effort 43 28 14 15 100 Dolphin schools Density (Z), ) (sehools/lOOO km-) 3.62 2.56 1.89 2.32 2.89 Number detected' 165 77 28 36 306 Mean target-species school size (S,J 89.41 83.97 104.55 179.04 99.13 Number target schools- 126 71 25 27 249 Proportion target schools (P, , | 0.790 0.906 0.897 0.744 0.825 Proportion of target animals by species Spotted 0.239 0.378 0.347 0.170 0.266 Spinner 0.293 0.242 0.351 0.054 0.235 Common 0.296 0,108 0.008 0.661 0.305 Striped 0.163 0.272 0.126 0.056 0.160 Fraser's 0.009 0.000 0.168 0.059 0.035 Proportion of animals in overlap area Coastal spotted 0.395 Offshore spotted 0.605 Eastern spinner 0.703 Whitebelly spinner 0.297 Number of schools in overlap area Spotted 29 Spinner 96 'Includes unidentified dolphin schools. -Includes schools identified as target species but for which a 1 est estimate of sc hool size was not made. schools. While searching in the study area (Fig. 1), observers on both vessels searched 27,(i(39 km and detected 306 dolphin schools containing 15 or more animals located within 3.7 km perpendicular distance of the trackline during Beaufort sea states of 5 and less (Table 1). The amount of effort and schools detected varied among strata; 43% of the total trackline searched and 54% of all dolphin schools detected were in the inshore area. Abundance estimation The estimate of/(0) for data in the total area (pooled) was 0,522. Density estimates (Di^) in the four strata, calculated using the pooled /(O), ranged from 1.89 to 3.62 schools/1000 km^ (Table 1). Estimates of mean school size (S,;.) of target species ranged from 83.97 to 179.04 animals (Table 1). The proportion of identi- fied dolphin schools that included target species iP,k) ranged from 0.744 to 0.906 among strata (Table 1). The proportions of individuals of all target schools that were spotted dolphins (P,^.) in the four strata ranged from 0.170 to 0.378 (Table 1). The proportions of the other target species among strata also varied greatly. For example, the proportion of common dolphins ranged from 0.008 to 0.661 (Table 1). Only 29 spotted dolphin schools were detected in the overlap region of the coastal and offshore spotted stocks. The proportion (P',j) of these that were offshore spotted dolphins was 0.605 (Table 1). Of 96 spinner dolphin schools detected in the area of overlap of eastern and whitebelly spinner stocks, the proportion of these that were eastern spinner dolphin individuals was 0.703. The areas inhabited by each stock (/l,jj.) of each tar- get species (A,^.) in each stratum (/l^.), and the over- lapping areas (A',,;,.) inhabited by (1) coastal and off- shore spotted dolphins and (2) eastern and whitebelly spinner dolphins, are shown in Table 2. Relatively, common dolphins were the most abundant species (Table 3). The estimate of 929,000 northern off- shore spotted dolphins represented 78% of the estimate of all stocks of spotted dolphins. The coefficient of variation (CV) of the abundance of the northern off- shore spotted dolphon stock, CV(iV|j), was 0.255. The Holt and Sexton Dolphin abundance in eastern tropical Pacific 109 Table 2 Area inhabited (knr) by each dolphin species/stock in inshore, middle. west, and south strata, and in total survey area eastern tropical Pacific. Dolphin species/stock Inshore Middle West South Total area Spotted Coastal, non-overlapping 113,660 8,557 122,217 Northern offshore, non-overlapping 4,073,493 3,788,767 4,016,325 11.878,585 Coastal and northern offshore, overlapping 909,279 909,279 Southern offshore, non-overlapping 3,175,138 3,175,138 Coastal and southern offshore, overlapping 68,455 68,455 Total 5,096.432 3.788,767 4.016,325 3,252,150 16,153,674 Spinner Costa Rican 248,366 248,366 Northern whitebelly. non-overlapping 1,806,239 1,806,239 Eastern and northern whitebelly, overlapping 4.762,844 3,401,601 2,889,195 11,053,640 Southern whitebelly. non-overlapping 9,825 346.337 47,898 2,.543,829 2,947,889 Eastern and southern whitebelly, overlapping 190,322 751.964 942,286 Total 5,211,357 3,747,938 4.743.332 3,295,793 16,998,420 Common Northern tropical 1,477,237 477,581 1,954,818 Western central tropical 17.886 1,273,546 2.207.221 3.498,6.53 Eastern central tropical 3,502,375 420,.581 3,922,956 Southern tropical 829,881 847,176 1,239,588 2,916.645 Total 5,827,379 3,018.884 2.207,221 1.239.588 12,293,072 Striped Northern tropical 1.11)4.177 681,546 1.785,723 Western central tropical 2,870,874 2,870,874 Eastern central tropical 3,549,019 1,066,116 4,615,135 Southern central tropical 510,371 1,658,074 514.537 3,792,278 6.475.260 Total 5,163,567 3,405,736 3.385,411 3,792.278 15.746,992 Eraser's* 5,211,357 3.747,938 4.743,332 3.295.793 16.998.420 All species/stocks 5,848,469 3,797.734 5.298,266 4.203,366 19.147,835 'Used total spinner dolphin area. Table 3 Estimates of base population sizes (A^, ) (10' animals) by stock for target dolphin species in total survey area. eastern tropi ■al Pacific. Estimates weighted by size of each stratum. Dolphin species/stock N. SE{N,^) cv(iv,; Dolphin species/stock N.J SE(Ar,^) CV(iV,^) Spotted Common Coastal 36.0 8.6 0.2.39 Northern tropical 124.9 39.7 0.318 Northern offshore 929.0 236.5 0.255 West central tropical 42.5 22.2 0.,522 Southern offshore 218.5 120.3 0.551 East central tropical 277.3 85.6 0.309 Total 1183.5 365.4 0.309 Southern tropical 943.2 388.8 0.412 Spinner Costa Rican Eastern Northern whitebelly .Southern whitebelly Total 20.9 579.6 333.1 83.1 1016.7 6.1 192.2 127.5 32.9 358.7 0.292 0.332 0.383 0.396 0.3.53 Total Striped Northern tropical West central tropical East central tropical Southern tropical Total 1387.9 92.4 99.9 230.6 212.3 635.2 536.3 20.2 40.7 48.6 62.4 171.9 0.386 0.219 0.407 0.211 0.291 0.271 Eraser's 247.9 202.5 0.817 Total 4471.2 1634.8 0.366 Fishery Bulletin 88(1). 1990 abundance estimate for eastern spinner dolphins was 579,600 animals with a CV of 0.332. Trends in abundance Assuming the CV(A^,j) for subsequent surveys will be constant during the sampling period, and with a and P errors equal to 0.10, a 10% annual decrease in abun- dance of offshore spotted dolphins can be detected in a minimum of 8 years. After 5 years, and assuming the CV{N,j) of 0.255 will remain constant, a minimum an- nual decline of 17.6% may be detected during which a 62% decrease of the offshore spotted stock would have occurred. Discussion A biased estimate may be acceptable to detect trends in population changes if the bias is constant among an- nual surveys and if it results in a more precise estimate. For example, the estimate of school density that was calculated by fitting a detection function to data pooled over all strata is biased upwards because searching ef- fort was not allocated to strata uniformly but propor- tionately to historical estimates of density. The inshore stratum, which historically had the largest density (Holt et al 1987), received a greater proportion of the search- ing effort (43% of total effort) compared with its rela- tive size (30% of total area). However, the pooled esti- mate is more precise than the stratified estimate, and the bias should be consistent during subsequent years. Our estimate of target school size (99.13 animals. Table 1) was half the estimate from aerial data collected in 1979 (199.8 animals per school) (Holt 1985). Although school size may have declined between 1979 and 1986, our estimate is similar to a previous school size estimate that used data collected 1979-83 aboard research vessels (119.9 animals per school) (Holt 1985). In addi- tion, we used only schools detected within 3.7 km perpendicular distance of the trackline, while the pre- vious aerial and ship studies used schools detected within 11.1 km perpendicular distance (approximate distance to horizon from ship). Our estimate using the 11.1 km perpendicular distance was 1 1 1 .97 animals per school. The previous estimates from airplanes and ships may have been biased upward because large schools are more likely detected at greater distances than are small ones. The inverse log weighting factor used to adjust biases in the school-size and species proportion estimates caused by detecting disproportionately more large schools may have over- or undercompensated by an unknown degree. Recent work by Drummer (1985) investigating size biases may be utilized during com- parisons of this data and subsequent years' data. Estimates of the relative proportions {P'ljk) of coastal and offshore spotted dolphins and of eastern and whitebelly spinner stocks in their respective areas of overlap were based on data pooled over all strata and included all schools occurring within 11.1 km (6 nm) perpendicular distance from the trackline; however, they were based on small sample sizes (29 spotted and 96 spinner schools). These estimates may change with collection of additional survey data. However, the proportion of eastern spinner dolphins to whitebelly spinner dolphins in their overlap area was 0.703 dur- ing 1986 (Table 1) and was 0.714 when all research vessel data collected from 1976-86 were combined (207 schools). Our population abundance estimates are intended to serve as the baseline estimates for relative comparisons using data collected during subsequent surveys. When our estimates of northern offshore spotted dolphins are compared with previous estimates, ours are much smaller than those made for data collected through 1979 (2,775,000 animals) (Holt and Powers 1982) and for data collected through 1984 (2,533,300 animals) (Holt 1985). Both the latter absolute and our current base estimates may be biased because they share com- mon data collection constraints— failure to detect all trackline schools— which bias the density estimates downward. The older estimates may also contain other biases which result in relatively higher values. For example, those estimates used a combination of data collected aboard airplanes, research vessels and tuna vessels, and survey coverage was pooled over several years, seasons, and areas. However, some variables used to calculate our estimates used small data sets that may have resulted in less precise results. Because our population estimates are intended to serve as the baseline estimate for relative comparisons using data collected during subsequent surveys, con- sistent bias during the sampling period will not jeopar- dize the results. Therefore, several options will be reviewed in analyzing subsequent years' data which may reduce variability associated with the population estimates. Sample sizes to calculate school sizes and species proportions may be increased by utilizing all schools detected at perpendicular distances from the trackline out to the horizon. We will investigate spotted dolphin abundance estimates using only schools of spotted dolphins. A computerized binocular system (Holt and Sexton 1987) may yield more precise esti- mates of radial distance and sighting angles to dolphin schools, and we have incorporated use of a ship-based helicopter to obtain aerial photographs of dolphin schools to calibrate observer estimates of school size. Hopefully, some or all of these factors may reduce the CViNii) levels to around ]2%, as anticipated l)y Holt et al. (1987) in initial survey design. Holt and Sexton Dolphin abundance in eastern tropical Pacific I I Acknowledgments We acknowledge the biologists who collected the data: S. Benson, C. Bisbee, K. Brownell, S. Buckland, A. Dizon, W. Irwin, R. LeDuc, M. Lynn, M. Newcomer, R. Pitman, L. Robertson, D. Skordal, S. Sinclair, P. Stangl, and M. Webber. We are grateful to A. Jack- son, W. Parks, and S. Reilly for directing the at-sea operations, and .W. Parks and P. Stangl for their shoreside support. We thank J. Barlow, S. Buckland, K. Burnham, D. DeMaster, M. Hall, L. O'Brian, and S. Reilly for reviewing this manuscript. Citations Au. D., W.L. Ferryman, and W. Perrin 1979 Dolphin distribution and the relationship to environmental features in the eastern tropical Pacific. Adni. Rep. L,J-79-4.3, Southwest Fish. Cent., Natl. Mar. Fish. Sei-v., NOAA. La.Jolla. CA 92038, .59 p. Barlow. J., and R.S. Holt 1986 Proportions of species of dolphins in the eastern tropical Pacific NOAA-TM-NMFS-SWFC-.56, Southwest Fish. Cent., Natl. Mar. Fish. Serv., NOAA, La Jolla, CA 920.38, 44 p. Buckland. S.T. 1985 Perpendicular distance models for line transect sampling. Biometrics 41:177-195. Burnham, K., D. Anderson, and J. Laake 1980 Estimation of density from line transect sampling of l)iological populations. Wildl. Monogr. 72, 2(12 |). Grain. B.. D. Burnham. D. Anderson, and J. Laake 1979 A Fourier series estimator of population density for line transect sampling. LItah State LIniv. Press, Logan, 25 p. Drummer, T.D. 1985 Size-bias in line transect sampling. Ph.D. Diss., Univ. Wyoming, Laramie. 143 p. Gerrodette, T. 1987 A power analysis for detecting trends. Ecology 68: 1364-1372. Hayes, R.J., and S.T. Buckland 1983 Radial distance models for the line transect method. Biometrics 39:29-42. Holt. R.S. 1985. Estimates of abundance of dolphin stocks taken inciden- tally in the eastern tropical Pacific yellowfin tuna fishery. Adm. Rep. LJ-85-20, Southwest Fish. Cent., Natl. Mar. Fish. Serv., NOAA, La Jolla, CA 92038, 32 p. 1987 Estimating density of dolphin schools in the eastern tropical Pacific ocean by line transect methods. Fish. Bull., U.S. 85:419-434. Holt, R.S., and J.B. Cologne 1987 Factors affecting line transect estimates <:if dolphin school density. .]. Wildl. Manage. 51(4):836-873. Holt. R.S.. and A. Jackson 1987 Report of a marine mammal siuT/ey of the eastern tropical Pacific aboard the Research Vessel McA rthu r July 29-Decem- ber 6, 1986. NOAA-TM-NMFS-SWFC-77, Southwest Fish. Cent., Natl. Mar. Fish. Serv., NOAA, La Jolla, CA 92038, 161 p. Holt, R.S., and J.E. Powers 1982 Abundance estimation of dolphin stocks in the eastern tropical Pacific yellowfin tuna fishery determined from aerial and ship surveys to 1979. NOAA-TM-NMFS-SWFC-23, Southwest Fish. Cent., Natl. Mar. Fish. Serv., NOAA, La Jolla, CA 92038, 95 p. Holt, R.S., and S.N. Sexton 1987 Rejxirt of a marine mammal survey of the eastern tropical Pacific aboard the Research Vessel David Slarr Jnnlini July 29-December 5, 1986. NOAA-TM-NMFS-SWFC-76, South- west Fish. Cent., Natl. Mar. Fish. Serv.. NOAA, La Jolla, CA 92038, 171 |i. Holt, R.S., T. Gerrodette, and J.B. Cologne 1987 Research vessel survey design for monitoring dolplnn abundance in the eastern tropical Pacific. Fish. Bull., U.S. 85:4.35-446. Perrin. W.F.. M.D. Scott. G.J. Walker, and V.L. Cass 1984 Review of geographical stocks of tropical dolphins (Stenrtia spp. and DAphituis del])his) in the eastern Pacific. Adm. Rep. LJ-87-02, Southwest Fish. Cent., Natl. Mar. Fish. Serv., NOAA, La Jolla, CA 92038, 68 p. Richey. C.R. 1976 Memorandum of opinion. CA NO. 74-1465 and CA No. 7.5-0227 U.S. District Court, District of Columbia, May 11, 1976. Smith. T.D. 1979 Report of the status of the porpoise stock workshop (August 27-31, 1979, La .Jolla, CA). Adm. Rep. LJ-79-41. Southwest Fish. Cent., Natl. Mar. Fish. Serv., NOAA, La Jolla, CA 92038, 120 p. SWFC 1976 Report of the workshop on stock assessment (jf porpoises involved in the eastern tropical Pacific yellowfin tuna fishery. Adm. Rep. LJ-76-29, Southwest Fish. Cent., Natl. Mar. Fisii. Serv., NOAA, La Jolla, CA 92038, 60 p. Abstract.— Age and growth pa- rameters of the tropical loliginid squid Sepioteuthis lessoniana in eastern Australian waters were determined from statolith growth-ring analysis. Juvenile specimens were captured, maintained alive, and their statoliths were chemically marked in fn'tii with either tetracycline or calcein. These chemicals produced a fluorescent mark within the statolith microstruc- ture when viewed under UV light. Statoliths were mounted in ther- moplastic cement and subsequently ground and polished. This process allowed rings to be visualized with- out any further preparation. It was thus ])ossible to validate that distinct statolith rings were formed daily and that less-distinct thinner rings were, in fact, subdaily rings. The results of the age analysis of field-captured individuals revealed that the population of S. lefn^oniiniii in the study area grows at a very fast rate. Maturity in both sexes was achieved in less than 100 days. All specimens aged were less than 6 months old. The size of large in- dividuals was within the range of S. lessoniana captured in other areas, with size ranges being 75-213 mm and 75-184 mm for males and fe- males, respectively. Growth rates determined for S. lessoniana based on statolith ageing are considerably different from pre- vious estimates based on length- frequency data. Age and Growth of the Tropical Mearshore Loliginid Squid Sepioteuthis lessoniana Determined from Statolith Growth-Ring Analysis George David Jackson Department of Marine Biology, James Cook University of North Queensland Townsville. Queensland 481 I. Australia Manuscript accepted 28 .liily 1989. Fishery Bulletin, U.S. 88:113-118. Estimation of growth rates along with estimates of age-at-maturity and life span are the key elements in understanding the population dynam- ics of marine organisms. In many instances, such data are difficult to obtain and approximations often have broad confidence intervals. The provision of accurate age estimates is one method by which reliable data on demographic processes can be ob- tained. For example, research into fish demography has developed ra- pidly (for relevant reviews see Pan- nella 1980, Campana and Neilson 1985, Jones 1986) following the dis- covery and validation of daily micro- structure increments within the oto- lith (Pannella 1971). Pelagic cephalopods have many fish-like features, yet differ consider- ably in fundamental ways with re- spect to metabolism and growth (O'Dor and Webber 1986). Because of their rapid growth and the lack of population statistics in comparison with fish studies, cephalopods are potentially one of the more interest- ing targets of demographic analysis. Ageing techniques of pelagic cephalo- pods, employing statolith microsti-uc- ture analysis, are currently devel- oping along lines similar to those applied to fish biology. Cephalopod statoliths and fish otoliths are physio- logically analogous structures (Rad- tke 1983). Although statolith gi-owth rings were recognized over 20 years ago (Clarke 1966), it has not been un- til relatively recently that some de- tailed microstructure analysis has taken place (see Jackson [1989] for relevant literature). Ongoing research into cephalopod demography is revealing that many species are short-lived and exhibit rapid growth and early maturity (Saville 1987). These features make cephalopods particularly amenable to ageing using daily growth rings. Re- cent work on growth and ageing on the small tropical sepioid Idiosepius pygmaeus has revealed that very small tropical species can have rela- tively fast growth and maturity rates (Jackson 1989) compared with pre- vious estimates. The majority of research on squid demography has taken place in tem- perate waters. Life spans for large temperate squids are generally be- lieved to be annual (Voss 1983, Natsukari et al. 1988, Amaratunga 1987). It was thus of interest to ex- amine age, growth and maturity parameters of a large tropical cepha- lopod to see if these parameters paralleled similar work on their tem- perate counterparts. The tropical loliginid squid Sepio- teuthis lessoniana was selected for age and growth analysis. This species is one of the larger tropical nearshore squids, reaching lengths of over 30 cm and occurs throughout much of the Indo-Pacific (Roper et al. 1984). 113 14 Fishery Bulletin 88(1), 1990 The environmental distribution of S. lessoniana in North Queensland waters includes offshore reef en- vii-onments as well as nearshore/estuarine habitats. Its distribution thus overlaps that of /. pygniaeus. Materials and methods The two primaiy components of this study involved (1) the collection of juvenile Sepioteuthis lessoniana specimens, which were maintained alive and stained with tetracycline or calcein to determine statolith ring periodicity, and (2) the collection of field specimens, which were fixed shortly after capture and subsequent- ly used for age and gonad analysis. Thus all age esti- mates were taken from field specimens that were not exposed to artificial maintenance. Juvenile S. lessoniana were often seen along break- waters in the Townsville region and occasionally in a local estuary. Specimens commonly sheltered under flotsam or floating pieces of ropes, singly or in schools of up to 20. Juveniles were dipnetted and transported back to the lab for subsequent tetracycline staining and maintenance. Collections were made between May 1988 and February 1989. Larger individuals for ageing were captured using squid jigs at night off the Picnic Bay jetty (lat. 1 9° 1 1 'S. long. 146°50'E) on Magnetic Island (a continental island ~7 km off Townsville) and the Australian In- stitute of Marine Science jetty south of Townsville (lat. 19°17'S, long. 147°03'E). Night jigging took place between November 1988 and February 1989. Squids were aged (i.e., statolith rings counted) and examination of reproductive sti-uctures was carried out to ascertain the level of maturity. Maturity was deter- mined by the presence of spermatophores in the sper- matophore sac in males and the presence of mature oocytes in females. Maintenance and tetracycline staining Captured 5. lessoniana were transported back to the lab and were maintained in 1500-L and 2500-L round tanks located outside so that squids would maintain normal diel periodicity. Specimens were kept alive as long as possible to provide maximum statolith growth before being sacrificed for examination, although in many instances the maintenance was terminated by death of the squid or sometimes by individuals jump- ing out of the tank. Squids which survived the first few days of captivity were maintained for between 9 and 39 days. Seawater from a closed recirculating system continually flowed through the tanks. Feeding was (/(/ libitum. Common food organisms maintained in the tanks were the crustacean Acetes sibogae ansfralis and Ambassid, Mugilid, and Clupeid fish species. On several occasions squids were induced to eat previously frozen prawns. Up to seven squids were maintained concur- rently in one tank. Their typical behavior was to hover motionless in a school, aligned in the same direction. The methods for tetracycline staining were the same as used for Idiosepius pygmaeus (Jackson 1989). Speci- mens were normally stained on the day of capture, and if they survived long enough they were exposed to a second staining at least 9 days later, with an interval of 9-20 days between stainings. In one experimental treatment, five squid were initially exposed for 1.5 hours to 100 mg of dissolved calcein per liter of sea- water, and then stained a second time 11 days later with tetracycline (250 mg/L, 2 hours). The calcein stain- ing produced a very faint green fluorescence compared with the very strong more yellow-green tetracycline hydrochloride fluorescence. Exposure to calcein at a higher concentration or for a longer time period is likely to produce more prominent fluorescence under UV irradiation. Preparation and observation of statolith microstructure Sepioteufliis lesso)iiana specimens were preserved in 10% borax-buffered seawater formalin to fix tissues and gonads. All length measurements refer to dorsal mantle length (DML). Statoliths were removed the next day within 12 hours to avoid damage from prolonged exposure to formalin. Statoliths were exposed to a 1% bleach solution for several minutes to remove excess tissue, rinsed in distilled water, dehydrated with 100% ethanol and mounted in the thermoplastic cement Crystal Bond. Crystal Bond was found to be an ex- cellent mountant, as it is completely translucent, does not fluoresce under UV irradiation, and statoliths can be easily manipulated or turned-over because it ijuick- ly melts at a low temperature and hardens relatively rapidly after removal from heat. Statoliths were then ground by hand on wet 1200- grade carborundum paper. The scratches from the grinding were removed by polishing the specimens either by hand on wet suede with 0.05 jl* alumina powder, or by using a modified gem polishing machine equipped with a microscope slide-holding arm, in which the statolith was lowered onto a fi-inch rotating disc to which was attached a wet Leco Lecloth impregnated with alumina powder. Statoliths rings were usually best visualized by grind- ing and polishing the statolith on the anterior (concave) surface only; however, in some statoliths that were particularly opaque or thick, the ring structure was Referenoo to tradi' n;imes (iocs not imply tTKiorst'iiiciit liy the National IMarine Fisheries Service, NOAA Jackson Age and growth of Sepioteuthis lessoniana Figure 1 (A) IIV micrograph of statolith from Sepioteuthis lessoniana from eastern Australian waters (age 68 days, DML 67 mm) stained with tetracycline twice at a 20-day interval and then allowed to grow for a further 18 days. Scale bar = 100 fjm. (B) Light micrograph of same specimen in Fig. lA, ground and polished on both anterior and posterior surfaces. Scale bar =100 ^m. (C) Light micrograph of lateral region of dor- sal dome (indicated by arrow in Fig. IB). Arrows indicate checks that correspond to tetracycline stainings. Scale bar = 50 (jm. 116 Fishery Bulletin 88(1), 1990 Table 1 Statolith ring validation information for Scpioteidhis lessouitnui. Mantle length Date experiment Number of Number of (mm) Stain 1 Stain 2 terminated days increments Comments 53 11 May 88 20 May 88 9 9 50* 18 May 88 — 27 May 88 9 9 Not taken 14 Oct. 88 (calcein) 25 Oct. 88 8 Nov. 88 11 11 Between stains 52 14 Oct. 88 (calcein) 25 Oct 88 9 Nov. 88 11 11 Between stains 73 14 Oct. 88 (calcein) 25 Oct. 99 20 Nov. 88 2fi 26 F^nnri 2n(i stain to edge 67 20 Oct. 88 9 Nov. 88 28 Nov. 88 20 20 50 8 Feb. 89 of hectocotylus 20 Feb. 89 observed on this 20 Feb. 89 specimen. 12 11 Died during 2nd staining 'Development clarified by grinding and polisliing on both surfaces. Growth rings were counted on the polished surface with no further preparation. Rings were accentuated by observing the specimen under polarized light and were enumerated with a hand counter by following the ring sequence with a pencil via a camera lucida. Speci- men age in days (i.e., statolith ring number) was estab- lished by taking the mean of at least three counts that deviated less than 10%. Specimen age was rounded to the nearest whole number. The statolith size and shape of a newly hatched S. lessoniana were determined from a specimen which was hatched from an egg trawled-up in Cleveland Bay off Townsville. This specimen further provided infor- mation on DML at hatching (5.3 mm). Growth rate of aged specimens was thus taken as the increase in man- tle length in mm/day minus 5.3 (i.e., DML-5.3/age). Results The notable characteristics oi Sepioteuthis lessoniana statoliths are a rounded dorsal dome and a relatively long, thin rostrum (Fig. IB). Unlike the sepioid Idio- sepius pygmaeus in which rings were most obvious in the lateral region of the statolith (Jackson 1989) or Loligo (Photololigo) edulis (Natsukari et al. 1988) which has a clearly countable ring sequence in the rostrum, S. lessoniana statolith rings were most discernable in the dorsal dome. Rings were very difficult to discern in the rostrum. The ring sequence within the statolith commenced from a prominent check, although some faint ring structure was visible inside this check. This check ring corresponded closely in size to the outer margin of the statolith of the newly hatched S. lessoniana (which also had some ring structure at hatching), indicating that this ring represents a hatching check. This feature has also been documented in the loliginid .s(iuids/l//o/f')(///i.s suhulata (Lipinski 1986) and Loliga (Phatoliiligo) edulis (Natsukari et al. 1988). Tetracycline staining The growth ring sequence could be clearly visualized and counted in the vicinity of the tetracycline or cal- cein marks in seven specimens maintained in captiv- ity. The counts of these specimens corresponded to a daily periodicity in ring formation (Table 1). Rings were generally more easily counted between two stain marks (P^ig. IC) than from the stain inark to the edge, as the edge rings are the most difficult to discern. A statolith check was often induced within the statolith, which cor- responded to the date of capture and staining (Figs 1A,C). When using high magnification or very sharp focus, numerous subdaily rings could be discerned which often made counting of daily rings difficult. This was espe- cially true in areas of the statolith where rings were quite thick (wide). Using a lower magnification or changing the plane of focus helped to delineate the true daily rings that were superimposed over the numerous subdaily rings. A similar phenomena has been shown to exist in the otoliths of the freshwater fish Coregonus spp. (Eckmann and Rey 1987). Age and growth A total of 23 individual .squids were aged (Table 2) to produce a growth curve (Fig. 2) and to ascertain maturity. (Jrowth inS. lessoniana is rapid, with a large size reached in less than 6 months. Growth rates deter- mined for larger specimens were considerably greater Jackson: Age and growth of Sepioteuthis lessoniana 17 Table 2 Variability in replicate statolith ring counts for field-captured Sepioteuthis lessoniana specimens. SD = standard deviation. Specimen no. Sex Replicate Mean SD 1 J 37, 36, 35 36.0 1.00 2 J 10, 10, 10 10.0 0 3 J 34, 37, 37 36.0 1.73 4 J 37, 37, .39 37.7 1.15 5 J 34, 34, 35 34.3 0.58 6 J 38, 40, 41 39.7 1.53 7 F 191, 190, 182 187.7 4.93 8 M 89. 85, 85 86.3 2.31 9 F 66, 65, 75 68.7 5.51 10 M 67, 67, 67 67.0 0 11 M 62, 55, 62 59.7 4.04 12 M 98, 95, 101 98.0 3.00 13 M 85, 96, 89 90.0 5.57 14 M 85, 77, 81 81.0 4.00 15 M 87, 89, 90 88.7 1.53 16 M 117, 119, 119 117.7 1.15 17 F 53, 57, .54 54.7 2.08 18 F 68, 64, 68 66.7 2.31 19 F 90, 90, 87 89.0 1.73 20 F 121. 122, 119 120.6 1.53 21 M 151, 159, 150 153.3 4.93 22 M 132, 142, 132 135.3 5.78 23 M 55, .57, 41 57.7 3.05 250 200 r 150 - 100 50 ■ ■ ■ ■ 4- ■1- ■+ + ■ 4 + /. ■ + MALES FEMALES JUVENILES • 50 100 150 AGE (days) 200 250 Figure 2 Age-length relationship for field-captured male, female, and juvenile Sepioteuth i'.s lesson ia na. than in the paralarvae or juveniles (Table 3). Maturity was found to take place as young as 67 days and 69 days in males and females, respectively. Discussion Ageing research with Sepioteuthis lessoniana provides further evidence that statolith growth rings in tropical loliginids are a valuable tool that can be applied to tropical squid demogi-aphy. Daily statolith ring periodi- city has also been demonstrated to occur in Illex ille- cebrosus (Hurley et al. 1985, Dawe et al. 1985), Allo- teuthis subidata (Lipinski 1986), Loligo opalescens (Yang et al. 1986), and Idiosepius pygmaeus (Jackson 1989). Furthermore, statolith ring sequences have been observed and counted in other temperate-water squid species (Kristensen 1980, Rosenberg et al. 1981, Nat- sukari et al. 1988). The presence of subdaily rings within the statolith microstructure of S. lessoniana illustrates the need for validation when attempting to enumerate growth rings. Validation is the only way to conclusively delineate whether less-prominent rings are in fact subdaily rings. Statolith growth-ring analysis suggests that S. lesso- niana in Australia has a short life cycle and rapid growth. Estimates of age and growth of this species Age, length, and growth Sepioteuth is lesson ia na . standard deviation. Table 3 -rate parameters for field-captured DML = dorsal mantle length; SD = Age range N (days) DML range (mm) Growth rate Range (mm/day) Mean SD Paralarva Juveniles Males Females 1 10 5 34-40 9 58-153 9 60-188 6.1 29-35 75-213 75-184 0.08 0.59-0.78 1.11-1.65 0.95-1.65 0.71 0.073 1.42 0.197 1.34 0.239 from size-frequency analysis at other localities suggest a very different pattern of growth. Earlier work on growth analysis in India (Rao 1954, Silas et al. 1982) has estimated that S. lessoniana takes up to three years to reach maximum size. This has also been further sup- ported by ELEFAN 1 computer program analysis of Rao's (1954) length-frequency data (Longhurst and Pauly 1987). Data from these different tropical populations of S. lessoniana suggest different patterns in the life cycle and growth of this species. The question is raised as to whether this is a locality difference or a reflection of different methodologies in analysis of growth rate. 118 -ishery Bulletin 1990 This study emphasizes the need to critically compare different methods of estimating growth in tropical squids. It is essential to review size-frequency analysis in the light of statolith age findings at a variety of dif- ferent geographic localities in the range of a species. Growth comparisons incorporating both size frequen- cy and statolith ageing methods of other tropical Aus- tralian squid species are currently being investigated. Artificial culture conditions provide a third method to estimate squid growth. Based on laboratory culture and field observations, Segawa (1987) has reached con- clusions similar to the present study of a short life span and rapid growth for S. lessoniana in temperate waters around Japan. Segawa further provides growth data from a captive female S. lessoniana spawning at 113 days and 143 mm in tropical waters in the Philippines. This value fits well in the size-age correlation for female squid in tropical Australia. Forsythe and Van Heukelem (1987) have suggested that development and refinement of ageing techniques are the most important tool needed in the study of cephalopod growth in natural populations. While stato- lith rings are a valuable tool, it is important to use a variety of methods to assess growth on the same population to distinguish between locality and method- specific differences. Acknowledgments I would like to thank Prof. J.H. Choat for assistance during the research and critical examination of the manuscript, C.H. Jackson for assistance with collection of field specimens, and the Australian Institute of Marine Science for providing access to study sites on Cape Ferguson. This research was supported by grants through James Cook University of North Queensland. Citations Amaratunga, T. 1987 Population biology, hi Boyle, P.T, (ed.). Cephalopod life cycles, vol. II: Comparative reviews, p. 239-2.52, Acad. Press. London. Campana, S.E., and J.D. Neilson 1985 Microstructure offish otoliths. Can. J, Fish. Aquat. Sci. 42:1014-1032. Clarke, M.R. 1966 A review of the systematics and ecolog-y of oceanic squids. Adv. Mar. Biol. 4:91-300, Dawe, E.C., R.K. O'Dor. P.H. Odense, and G.V. Hurley 1985 Validation and application of an ageing technique for short-finned sciuid (Illex illecebrosus). J. Northwest Atl. Fish Sci. ti:107-n. Mar. Biol. Ecol. 116:177-190. ODor, R.K., and D.M. Webber 1986 The constraints on cephalopods: why squid aren't fish. Can. J. Zool. 64:1591-1605. Pannella, G. 1971 Fish otoliths: daily growth layers and periodical pat- terns. Science (Wash.. DC) 173:1124-1127. 1980 Growth patterns in fish sagittae. In Rhoads, D.C.. and R.A. Lutz (eds.). Skeletal growth of aquatic organisms. Bio- logical records of environmental change, p, 519-560. Plenum Press, NY. Radtke, R.L. 1983 Chemical and sti"uctural characteristics of statoliths from the short-finned squid Illex illecebrosus. Mar. Biol. 76:47-54. Rao, K.V. 1954 Biology and fishery of the Palk-Bay s((uid, St'pinleulhis lUTtiptniiix Gimld. Indian J. Fish. 1:37-66. Roper, C.F.E., M.J. Sweeney, and C.E. Nauen 1984 Cephalopods of the world. An annotated and illustrated catalogue of species of interest to fisheries. FAO Fish. Synop. 12.5(3), 277 p. Rosenberg, A. A., K.F. Wiborg, and LM. Bech 1981 Growth of T'orfarorf<'.s,sa^)7(a^!(.s- (Lamarck) (Cephalopoda, Ommastrephidae) from the northeast Atlantic, based on counts of statolith growth rings. Sarsia 66:53-57. Saville, A. 1987 Comparisons between cephalopods and fish of those aspects of the biology related to stock management. In Boyle, P.R. (ed.). Cephalopod life cycles, vol. II: ('omparative reviews, p. 277-290. Acad. Press. London. Segawa, S. 1987 Life history of the oval squid Sepiolcnlhis lesfioniinui in Kominato and adjacent waters central Honshu. Japan. .1. Tokyo Univ. Fish. 74(2):67-105. Silas. E.G.. K.S. Rao. R. Sarvcsan. K.P. Nair, and M.M. Meiyappan 1982 The exploited squid and cuttlefish resources in India: A review. Tech. Ext. Ser. Mar. Fish. Inf. Serv., Cochin 34, 17 p. Voss, G.L. 1983 A review of cephalopod fisheries biology. Mt'm. Nat. Mus. Vic. 44:229-241. Yang, W.T.. R.F. Hixon. P.E. Turk. M.E. Krejci. W.H. Unlet, and K.T. Hanhin 1986 Growth liehavior. and sexual m.-ituration of the market S(|uid, LdIii/ii iipaUvci-nx, cultured through the life cycle. Fish. Bull., U.S. 84:771-798. Abstract.- Examination of 95 Dail's porpoise specimens incidental- ly caught in gill nets, and 4 collected as beach strandings, indicate signifi- cant sexual dimorphism and develop- mental variation in several aspects of external morphology and colora- tion. The dorsal fins of males become extremely canted in adulthood, and mature males can be distinguished by this feature alone. Size of the post- anal hump of connective tissue and caudal peduncle depth also become exaggerated in adult males. The uro- genital color pattern is highly vari- able, and frosting variations on the dorsal fin and flukes can be used to discern the general age of the indi- vidual. The strong sexual dimor- phism and small testes of Ball's por- poise indicate a polygynous mating system. It is suggested that Dail's porpoise secondary sexual character- istics are used in male-male competi- tion or female choice. Sexual Dimorphism and Development of External Features in Dail's Porpoise Phocoenoides dalli* Thomas A. Jefferson Moss Landing Marine Laboratories, P O Box 450. Moss Landing, California 95039 Present address: Marine Mammal Research Program Texas ASM University at Galveston PO Box 1675. Galveston, Texas 77553-1675 Behavioral studies of many mammals have been greatly facilitated by the ability of researchers to distinguish different age and sex classes in the field (see Schaller 1963 for mountain gorillas Gorilla gorilla heringei; Geist 1968 for mountain sheep Ovis dalli and 0. canadensis; Schaller 1972 for lions Panthera leo; Smith 1988 for mountain goats Oreamnos ameri- canus). Knowledge of the behavior and social systems of cetaceans is not as advanced as that of terrestrial mammals, largely due to difficulties in studying these animals at sea and in identifying individuals and age/sex classes (but see Bigg 1982, Bigg et al. 1987 for killer whales Orcinus orca). Dail's porpoise Phocoenoides dalli (True, 1885) presents particular prob- lems for behavioral ecologists. It is generally an open-ocean species, seen most commonly several kilometers from shore; lives in small groups that are relatively hard to detect; and is extremely fast-swimming and unpre- dictable in its movements, often out- swimming or evading research ves- sels. In addition to these problems, detailed studies of sexual and age- related variation in external mor- phology of Ball's porpoise have not been conducted, and thus identifica- tion of different age/sex classes has been limited. Although few (juantitative studies have been done, comments in the Manuscript accepted 22 August l'J8y. Fishery Bulletin, U.S. 88:119-132. 'Contribution no. 4 of the Marine Mammal Re- search Program, Texas A&M University at (Jalveston. literature and my past observations led me to believe that there may be reliable external indicators of age and sex in this species. In particular, five features seemed promising: (1) dorsal fin shape (pers. observ.), (2) caudal peduncle shape (Mizue and Yoshida 1965, Houck 1976), (3) post- anal hump size (Scheffer 1949), (4) size of the thoracic epaxial muscle mass (Newby 1982), and (5) colora- tion patterns, especially the frosting variations of the dorsal fin and flukes (Mizue and Yoshida 1965, Morejohn et al. 1973, Morejohn 1979, Kasuya 1982). These characteristics were chosen specifically for their potential use in identifying age/sex classes in the field. Another objective of this study was to quantify sexual, developmen- tal, and individual variation in exter- nal morphology for use in population studies. Finally, it was hoped that this information, if combined with data on testis weight, would shed light on the type of social system possessed by Ball's porpoise. Materials and methods While working as a scientific ob- server in the U.S. -Japan Cooperative Research Program on Ball's Por- poise, between 12 June and 14 July 1986, 1 examined 95 Ball's porpoises aboard the mothership Nojima Mam of the Japanese high seas salmon driftnet fishery (see description of this fishery in Jones 1984). All of the animals died after becoming inciden- I 19 120 Fishery Bulletin 88(1), 1990 tally entangled in gill nets set for salmon south of the Near Island group at the western end of the Aleutian Islands, Alaska. The capture locations of these specimens are plotted by 1-degree block in Figure 1. All specimens were of the dalli-type color morph (see Houck 1976) and were presumably from the north- western North Pacific population that calves south of the western Aleutians (Kasuya and Ogi 1987). In addi- tion to these specimens, four calves collected as beach strandings along the California coast in 1987 and 1988 were used in the analyses. Although these four speci- mens were from a separate population (Kasuya 1982, Kasuya and Ogi 1987), they were included because the sample size of calves was small, and geographic varia- tion at the calf stage was assumed to be insignificant. Little work has been conducted on geographic varia- tion in subadult animals. However, some differences in fin shapes between calves of different stocks of spin- ner dolphins Stenella longirostris have been identified, although these are minor compared with differences in adults (W.F. Perrin, Southwest Fish. Cent., Natl. Mar. Fish. Serv., NOAA, La Jolla, CA 92038, pers. commun., Aug. 1989). Each porpoise was weighed and sexed, a specimen number was assigned, and black-and-white photo- graphs were taken. A series of seven measurements were obtained to the nearest 0.5 cm (Fig. 2, Table 1), and the animal was dissected. The reproductive status of females was assessed as pregnant, lactating, resting, or immature, based on gross examination. The distribu- tion of total lengths of the specimens is shown in Fig- ure 3. Each specimen was placed into one of seven age/sex classes (Table 2). Most of the classes were based on the total length of the porpoise, generally following the growth curve presented by Newby (1982). Because there was only one neonate male, and only one feature (Canting Index) was measured on it, it was pooled with the females. The assumption was made that there are no sexual differences at the neonate stage. The cutoff between immature and mature males was based on Jones et al. (1987), who gave 180 cm as the average length at sexual maturity. Attainment of sexual matur- ity in males is considered to be more strongly corre- lated with length than with age, and all males in this study over 180 cm had enlarged testes, indicative of maturity (Kasuya and Shiraga 1985). Female sexual maturity is not well correlated with length, so average length at maturity was not used to separate immature and mature females. If a female was pregnant, lac- tating, or had apparent corpora upon gross examina- tion of the ovaries, it was classified as mature. If it fulfilled none of these criteria, but was ^151 cm, it was placed into the immature class. Because over 95% of all mature females in the study area are pregnant or Attu 1 Agattu 1 ■c 3 7 3 11 25 8 6 3 14 15 530N 520 51° 500 171" 172« 173° 1740 49- 175"E Figure 1 (-'apture locations of Dall's porpoise specimens from the northwestern North Pacific. lactating during the fishing season, these are very good indicators of sexual maturity for this population (Jones et al. 1983). In order to test quantitatively for variation in dorsal fin shape, a value similar to that used by Perrin (1975) for spinner dolphins was computed as the ratio alb (where a is measurement 7 and h is measurement fi in Table 1). This quantity is termed the Canting Index, and for Ball's porpoise will generally fall between Jefferson: Sexual dimorpfiism and development in Phocoenoides dalli 121 Figure 2 Measurements taken of Ball's porpoise specimens (see also Table 1). a females n=66 t:ii Total Length (cm) c 01 10 b males n=33 Total Length (cm) Figure 3 Frequency histograms of total lengths of Dall's porpoise specimens. Table 1 Measurements of Ball's porpoise specimens taken in this study, illustrated in Figure 2. Numbers in parentheses cor- respond to numbers of standard measurements in Norris (1961). No. (Fig. 2) Measurement 1 Total length (1) 2 Thoractic girth, midway between blowhole and dorsal fin 3 Bepth of caudal peduncle, posterior to postanal hump 4 Bepth of postanal hump 5 Height of dorsal fin (32) 6 Base of dorsal fin (33) 7 Length from posterior insertion of dorsal fin t(j a perpendicular from tip Table 2 Summary of age/sex classes into which Ball's porpoise speci- mens were separated. Age/sex class Symbol Criteria n Calf N < 130 cm 5 .Juvenile female JF 131-150 cm 2 Juvenile male JM 131-150 cm 6 Immature female IF ^151 cm, not pregnant 16 or lactating, no corpora Immature male IM 151-180 cm 12 Mature female MF ^151 cm, pregnant, 44 lactating, or corpora present Mature male MM 3^181 cm 14 Total 99 zero and one. The higher the Canting Index, the larger the degree of foreward canting. Sexual differences in morphology between immature males and females and between mature males and females were tested with a f-test, and age variation was tested separately for males and females with a one- way analysis of variance (ANOVA). Variation in dorsal fin coloration was tested with a chi-square test. Be- cause proportions are known not to be normally dis- tributed, an arc sine transformation was performed on all proportions for statistical analyses (the resulting data approached normality). Results External morphology Dorsal fin shape The dorsal fins of Dall's porpoises cant foreward to varying degrees (Fig. 4). The Cant- ing Index of the dorsal fm shows a tendency to increase 122 Fishery Bulletin 88(1), 1990 Figure 4 Dorsal fin shapes of Dall's porpoise specimens: (a| calf (TAJ 171). (b) im- mature male (TAJ IfiS), (c) mature fe- male (TAJ 156), (d) mature male (TAJ 161). Figure 5 External features of Dair,s porpoise specimens plotted by age/sex class. Points are mean, boxes are + ISD, ver- tical bars are range, and numbers in- dicate sample size. ANOVAs test for developmental variation separately within each sex: significance '** //< 0.001. " p<0.01. • /)<0.05, ns = not significant. 0,8- a 13 T 0 20 - c 14 0,7- X 06- T5 C — 05- O) c 'c °''- (0 " 03- - 1 1 5 r n 4 Y 4/ +** t □. 0) O 0) u c 3 T3 a 0 18 - 0 16 - 0 14 - 0.12 - 4 }- 6 16 .^ X *** 4 ** -r'T^i \\ ns r - < 02- J. - I J - N JF JM IF IM MF MM Age/sex Class N JF JM IF IM MF MM Age/sex Class Post-anal Hump Depth (/TL) J o o o ! 2 S S b .1 1 J *** b o o 2 o 1- 08- 0,7- 06- d 4 -r J 1 6 1 H 4 r - 4 , 1 u I v. /J 4 / ^ - r 1 ns N JF JM IF IM MF MM Age/sex Class N JF JM IF IM MF MM Age/sex Class Jefferson: Sexual dimorphism and development in Phocoenoicles dalli 123 0,8- a ; 0.7- Y=0.021+0.22X 1 X o ■o £ 06- r2=0.79 I ■ ^^^ ■ ^ c 'I 05- O 0 4 - ■^-"""'^ " " ■ 1 ^ ■ • ^^^^'^ ■ ■ ■ ■ ' ■ ■ " 130 140 150 160 170 180 190 200 210 Total Length (cm) Depth (/TL) o p o o b I Y=0.006+0.06X i r^'=0.82 I ■ B Q. E 3 I ■ ■ \^^^ • ■ ■ Post-anal 130 140 150 160 170 180 190 200 210 Total Length (cm) Figure 6 Development of external features in male Dall's porpoise specimens. Broken line represents average length at sexual maturity. with age of the animal in males (Figs. 5a, 6a). In fe- males, however, the dorsal fin apparently does not change shape once it reaches the immature stage. The difference between the mean Canting Indices for im- mature males and females was significant {t =3.079, df = 25, /><0.01), and the difference between mature males and females was highly significant (/ = 11.539, df = 55, /j<0.001). In fact, there was little or no overlap between adult males and any other age/sex class, and individuals with a Canting Index of 0.55 or greater can be assumed to be adult males. At a casual glance, it appears that the dorsal fins of adult males have a wider base than those of females. However, the mean base/height ratio for mature females was 2.15 (SD =0.18, range = 1.83-2.57, n =44), and for mature males it was 2.16 (SD=0.15, range = 1.94-2.41, «=13), a non-significant difference (< =0.182, df = 55, p>0.05). The difference in appear- ence results from the more anterior position of the tip of the fin f)f arlult males, not a wider liase. Postanal hump depth A small to moderate hump composed of connective tissue, just posterior to the anus, could be seen in nearly every Dall's porpoise, in- cluding newborn animals. This feature was even evi- dent in most nearterm fetuses. In some females of all female age classes, the hump was <0.5 cm, and thus not measurable (Fig. 5b). Im- mature and mature males all had significant postanal humps, but the feature was exaggerated only in adult males, in which it was >1.2% of the total length. Dif- ferences between immature males and females {t = 4.086, df = 26, 7J<0.001), and between mature males and females (/ =8.599, df = 56, jD<0.001) were both highly significant. Postanal hump depth of males in- creases with increasing length (Fig. 6b). and is corre- lated with the degree of foreward canting of the dor- sal fin (Fig. 7). Caudal peduncle depth Depth of the caudal pedun- cle (as a proportion of total length) showed a great deal 124 Fishery Bulletin 88(1), 1990 r =0.88 p <0.001 0 000 0 005 0 010 0 015 0 0 Post-anal Hump Depth (/TL) Figure 7 Canting Index vs. postanal hump liejith for male Dali's porpoise specimens. Dashed box comprises juvenile and immature males. of overlap for all age/sex classes, except mature males (Fig. 5c). There was no significant difference in this feature between sexes in immatures (t = 1.683, df = 26, p>0.05), but the difference between mature males and females was highly significant (f = 14.935, df = 56, p<0.001). The caudal peduncle only becomes greatly deepened in adult males, and again this appears to be an absolute difference. Peduncle depth represents >15% of the total length in mature males, and <15% in all other age/sex classes. Size of thoracic epaxial muscle mass There was a noticeable "hump" on the back of most adult males ex- amined in this study (Fig. 8f). Dissections of this hump revealed that it resulted from an increase in the size of the thoracic epaxial muscle mass. It did not seem to be asociated with an increase in the thickness of the blubber layer. Measurement of the thoracic girth in Dali's porpoise was not a good indication of the size of this muscle mass. There was no difference in the size of the thoracic girth (relative to total length) either between immature males and females {t =0.245, df = 26, p>0.05) or be- tween mature males and females (^=0.130, df = 56, /;>(}. 05) (Fig. 5d). After an increase in thoracic girth from the neonate to the juvenile stage, it leveled off in both sexes. Flul0.55). There is almost always a large postanal hump (>1.2% of the total length), and the peduncle is deepened (>15% of the total length). Often, an enlargement of the thoracic epaxial muscle mass is apparent as a hump between the blowhole and the dor- sal fin. Many (possibly older) adult males appear to have flukes with a convex rear border, and they are usually very robust (the head may look disproportionately small). The assumption is made that results from animals of this northwestern North Pacific population are ap- plicable to other Dall's porpoise populations as well. Growth characteristics of reproductively isolated, small cetacean populations are known to sometimes differ Jefferson Sexual dimorphism and development in Phocoenoides dalli 129 widely (Perrin 1984, Kasuya and Shiraga 1985)1 how- ever the same general trends seem to be apparent in photographs of animals from other populations. Only future studies of these features in other P. dalli popula- tions will reveal if this assumption is justified. Sexual dimorphism, testis weight, and mating system The moderately extreme sexual dimorphism apparent in this species deserves some discussion. Not only are males significantly longer and heavier than females (Kasuya 1978, Morejohn 1979, Newby 1982), but sev- eral morphological features are exaggerated in adult males. The mating system of Ball's porpoise is un- known, but sexual dimorphism among mammals, in which the male is larger, is usually related to intra- sexuai competition for females in a polygynous mating system. In fact, Ralls (1977) determined that extreme sexual dimorphism is a very good predictor of extreme polygyny. Newby (1982) implied a polygynous system for P. dalli, and Landino (1985) suggested a unimale system (either polygyny or monogamy). Landino's prediction is based on the finding that the relative testes weights of various primate species are related to their breeding systems (Short 1979, Harcourt et al. 1981, Harvey and Harcourt 1984). Males with relatively large testes tend to have multimale systems (mostly promiscuity), while those with smaller testes have unimale systems. This is thought to be related to the need to deliver large loads of sperm in promiscuous species, in which the potential for sperm competition exists (see Trivers 1985 for a general description of sperm competition theory). Kenagj' and Trombulak (1986) raised the question of whether mating system/testis size predictions apply to cetaceans, but Brownell and Ralls (1986) found sperm- competition theory to be quite useful in explaining mating systems in mysticete cetaceans. Kenagy and Trombulak (1986) compared relative testis weights for 133 species of mammals, and found the harbor porpoise Phocoena. phocoena, a closely related species, to have the highest value. This evidence points to extreme multimale (probably promiscuous) breeding for this species, which is consistent with mor- phological evidence. Harbor porpoises show reverse sexual dimorphism, with females larger than males (Gaskin et al. 1984), and males do not appear to possess any secondary sexual characteristics. Unfortunately, Kenagy and Trombulak (1986) did not include data on testis weight in Dall's porpoise. In order to compare Dall's porpoise with their information on other species, data on nine adult males presented by Subramanian et al. (1986, 1987) were used in a manner consistent with that of Kenagy and Trombulak. Only sexually mature males taken during the season of male sexual activity were used (August-September; Jones et al. 1988). When plotted on Figure 1 of Kenagy and Trombulak, the Ball's porpoise data point fell close to the regression line for all 133 mammalian species, and well below that of the other five species of small odon- tocete plotted. The testes of P. dalli comprise only about 0.22% of their body weight, as opposed to 4.00% in P. phocoena. Small testes relative to body weight suggests a single-male system with low copulatory frequency (either monogamy or polygyny) for Ball's porpoise. Monogamy is rare in mammals (<3% of all mammal species are monogamous), and it is generally associated with sexual monomorphism (Kleiman 1977). Thus it is unlikely to be the case for P. dalli. This leaves polygy- ny, and this conclusion is further supported by the sex- ual dimorphism in size (males larger) and the striking secondary sexual characteristics of males in this spe- cies. Although some adult males are present, Kasuya and Jones (1984) and Kasuya and Ogi (1987) reported that the majority of mature males are segregated from mating areas during the mating season in the western Pacific. This finding is also more consistent with polygyny than with monogamy. Several species of odontocete cetaceans show sexual dimoi-phism in these same charcteristics. For instance, extremely canted dorsal fins are seen in adult male spinner dolphins, especially those of the eastern form of the eastern tropical Pacific (Perrin 1972, 1975). Killer whales and spectacled porpoises Australopho- caena dioptrica have a significant degree of sexual dimorphism in both size and shape of the dorsal fin (Fraser 1968, Brownell 1975, Heimlich-Boran 1986, Leatherwood et al. 1982). Adult male long-finned pilot whales Globicephala melas have more rounded fins, with thicker leading edges than females and young (Sergeant 1962). Postanal humps appear to be common in many spe- cies of odontocetes, although they have been properly described and correlated with age and sex in only a few (published photographs show them in several genera, including Peponocephala, Lagenorhynckus, and Lage- nodelphis). Rough-toothed dolphin S^eno bredanensis. spotted dolphin Stenella attenuata. spinner dolphin, and some common dolphin Delphinus delphis adult males exhibit prominent humps (Norris 1967; Perrin 1972, 1975; Evans 1975; Leatherwood et al. 1982). Again, this feature is most exaggerated in eastern spinners. Deepened caudal peduncles (not including the post- anal hump) are apparent in photographs of adult male spinner dolphins (Perrin 1972), and Norris et al. (1985) were sometimes able to use this feature to distinguish adult male Hawaiian spinners in the field. To my 130 Fishery Bulletin 1990 knowledge, thoracic humps of the extent seen in adult male Ball's porpoises have not been reported in other cetaceans, but they may exist in a more subtle form in some species. Sexual dimoiphism in fluke shape sim- ilar to that in P. dalli has been reported by Nishiwaki et al. (1963) for sperm whales Physeter macrocephalus. Overall, the dimorphism described here for Ball's porpoise most resembles that described for eastern and whitebelly spinner dolphins by Perrin (1972, 1975). Per- rin (1972) proposed that the strikingly canted dorsal fins and enlarged postanal humps of adult male eastern Pacific spinners function as species-recognition signals within mixed schools of spinner dolphins and spotted dolphins. Norris et al. (1985) interpreted these second- ary sexual characteristics as signals that make adult males easily recognizable, and possibly allow them to mimic gray reef sharks Carcharhinus amblyrhynchos. Subgroups of adult male spinners seem to play an or- dering role in dolphin schools, and often achieve this by means of overt aggression. Thus their striking ap- pearance would assist in this aggressive role. Mating appears to be promiscuous in Hawaiian spin- ner dolphins (Norris et al. 1985), but understanding the functions of sexual dimorphism seems to be compli- cated by other social pressures than just mating system considerations. Sexual dimorphism may have originally evolved in association with a polygynous mating sys- tem. However, if Hawaiian spinners tended more toward promiscuity than other populations, the sexual- ly dimorphic features might become reduced, but could still function in social ordering (as proposed by Norris et al. 1985). Perhaps eastern spinners have remained strongly polygynous and thus dimorphic, and Hawaiian spinners have moved more towards promiscuity, with whitebellys intermediate. In this scenario, Hawaiian, whitebelly, and eastern spinner stocks would show a grade of increasing degrees of polygyny and sexual dimorphism, and decreasing testis size. In general, spinner dolphins have relatively large testes (Perrin et al. 1977, Perrin and Henderson 1984). However, eastern spinners have smaller testes than whitebelly spinners, and this is con- sistent with sperm competition predictions. There is little information on Hawaiian spinners (the form with the least pronounced dimorphism), but if this hypoth- esis is correct, one would expect them to have the least polygynous system and the largest testes of these three S. longirostrif: races. I propose that despite the similarity in appearance, the sexual dimorphism in Ball's porpoise evolved and is used largely in polygynous mating. Ball's porpoises do not school with other species, and thus have no need for the type of species-recognition signals suggested by Perrin (1972). Also. Ball's live in small groups, and the need for aggressive ordering of schools is probably nonexistent or very weak. I have suggested above that P. dalli is polygynous, and I believe these sexually dimorphic features are mostly related to male-male competition for females in this species. Newby (1982) first suggested this after discovering that adult male Ball's taken closer to the "rutting season" were heavier, and had a larger thor- acic girth, than those taken earlier. The relationship between androgen levels, social aggression, and neck girth has been well established in some polygynous mammals (Lincoln 1971, Bouissou 1983). It is easy to imagine how an increase in size of the thoracic muscu- lature could benefit fighting males, but how the other features could relate to this requires some speculation. If canting of the dorsal fin increases its rigidity, this could aid males that use their dorsal fins in aggressive encounters. A large postanal hump may assist some- how in mating with the female. Of course, these features may not be used in fighting at all, but may instead function as visual signals, allowing males to gauge each other or in being attractive to females, and thus operating in terms of female choice. It is not possible to evaluate the validity of these speculations at present, because almost nothing is known of the social behavior of Ball's porpoise. Long- term detailed observation of wild porpoises, especial- ly including" identification of individuals, is needed to understand how these features operate in wild Ball's porpoise groups. Acknowledgments This study was completed in partial fullfillment of the requirements for an M.Sc. degree at Moss Landing Marine Laboratories. I would like to thank the staff of the National Marine Mammal Laboratory (NMFS, NOAA), for allowing me to collect this "extra" data, for providing equipment, and for suggesting improve- ments. Atsushi Endoh, Masahiko Katsumata, Wes Armstrong, Scott Brainerd, Jim Thomason, and the crew of the Nojima Maru helped in data collection. The California specimens were obtained through the assistance of Nancy Black, Sal Cerchio, Barbara Curry, the California Marine Mammal Center, and Marine- world Africa/USA. Jill Schoenherr. Lynn McMasters, and Greg Silher helped during write-up. Reviews of earlier drafts were provided by Bernd Wiirsig, Mike Foster, Bob Brownell, Toshio Kasuya. and an anon- ymous reviewer. Jefferson, Sexual dimorphism and development in Phocoenoides dalli 131 Citations Best. P.B. 1988 The external appearance of Heaviside's dolphin, Cephalo- rhynchiis h^ainsidii (Gray, 1828). Rep. Int. Whaling Comm. (Spec. Iss. 9):279-299. Bigg. M. 1982 An assessment of killer whale (Orcinus orca) stocks off Vancouver Island. British Columbia. Rep. Int. Whaling Comm. 32:655-666. Bigg. M.A.. G.M. Ellis. J.K.B. Ford, and K.C. Balcomb 1987 Killer whales: A study of their identification, genealogy and natural history in British Columbia and Washington State. Phantom Press and Publishers, Nanaimo, B.C., 79 p. Bouissou, M.F. 1983 Androgens, aggressive behaviour and social relationships in higher mammals. Horm. Res. (Basel) 18:43-61. Brownell. R.L., Jr. 1975 Pkocoena dioptrica. Mamm. Spec. 66:1-3. Brownell. R.L.. Jr., and K.R. Ralls 1986 Potential for sperm competition in baleen whales. Rep. Int. Whaling Comm. (Spec. Iss. 8):97-112. Evans, W.E. 1975 Distribution, differentiation of populations, and other aspects of the natural history of Delphitnif: delphis Linnaeus in the northeastern Pacific. Ph.D. diss.. L'niv. Calif., Los Angeles. 145 p. Fraser. F.C. 1968 Notes on a specimen oiPhocnena dioptrini from South Georgia. Br. Antarct. Surv. Bull. 16:51-56. Gaskin. D.E.. G.J.D. Smith. A. P. Watson. W.Y. Yasui. and D.B. Yurick 1984 Reproduction in the porpoises (Phocoenidae): Implications for management. Rep. Int. Whaling Comm. (Spec. Iss. 6): 135-148. Geist, V. 1968 On the interrelation of external appearance, social behaviour and social structure of mountain sheep. Z. Tier- psychol. 25:199-215. Goodall. R.N, P.. A.R. Galeazzi, S. Leatherwood. K.W. Miller. I.S, Cameron. R.K. Kastelein. and .A. P. Sobral 1988a Studies of Commerson's dolphins, Cephulnrhyiwhus com- mersonii, off Tierra del Fuego, 1976-1984, with a review of information on the species in the South Atlantic. Rep. Int. Whaling Comm. (Spec. Iss. 9):3-70. Goodall. R,N,P.. K,S, Norris. A.R. Galeazzi. J, A, Oporto, and I.S. Cameron 1988b On the Chilean dolphin, Cephnlorhynchiis eutnipid (Gray, 1846). Rep. Int. Whaling Conun. (Spec. Iss. 9): 197-257. Hall, J.D. 1981 Aspects of the natural history of cetaceans of Prince William Sound, Alaska. Ph.D. diss., Univ. Calif., Santa Cruz, 148 p. Harcourt, A.R., P.H. Harvey. S.G. Larson, and R.V. Short 1981 Testis weight, body weight and breeding system in pri- mates. Nature (Lond.) 293:55-57. Harvey. P.H.. and A.H. Harcourt 1984 Sperm competition, testis size, and breeding systems in primates. In Smith, R.L. (ed.). Sperm competition and evolu- tion of animal mating systems, p. .589-600. Acad. Press, NY. Heimlich-Boran. J.R. 1986 Photogrammetric analysis of growth in Puget Sound Orcinua orca. hi Kirkevold, B.C., and J.S. Lockard (eds.), Behavioral biology of killer whales, p. 97- 111. Alan R. Liss. Inc., NY. Houck. W.J. 1976 The taxonomic status of the species of the porpoise genus Phocoenoides. Unpubl. pap. ACMRR/MM/SC/114, presented to the Food Agric. Org. (FAO). Humboldt State Univ., Areata. CA. 13 p. Jones, L.L. 1984 Incidental take of the Dall's porpoise and the harbor por- poise by Japanese salmon driftnet fisheries in the western North Pacific. Rep. Int. Whaling Comm. 34:531-538. Jones, L.L., D.W. Rice, and A. A. Wolman 1983 Biological studies of Dall's porpoise taken incidentally by the Japanese salmon mothership fishery. Unpubl. Rep. SC/35/Sm9, Int. Whaling Comm. Sci. Comm.. The Red House, Station Rd., Histon, Cambridge, CB4 4NP, U.K., 21 p. Jones. L.L., G.C. Bouchet, and B.J. Turnock 1987 Comprehensive re[jort on the incidental take, biology and status of Dall's porpoise. Unpubl. Doc. 3156, Int. North Pac. Fish. Comm. Natl. Mar. Mammal Lab., Natl. Mar. Fish. .Serv., NOAA, Seattle, WA 98115-0070, 78 p. Jones, L.L.. M.E. Gosho. D.W. Rice. A. A. Wolman. M.M. Muto. and W.A. Walker 1988 Report on biohjgy of male Dall's porpoise in the western North Pacific and Bering Sea. Unpubl. Doc. 3289, Int. North Pac. Fish. Comm. Natl. Mar. Mammal Lab., Natl. Mar. Fish. Serv., NOAA. Seattle, WA 98115-0070, 25 p. Kasuva. T. 1978 The life history of Dall's porpoise with special reference to the stock off the Pacific coast of Japan. Sci. Rep. Whales Res. Inst. Tokyo 30:1-63. 1982 Preliminary report of the biology, catch and populations of Phocoenoides in the western North Pacific. In Mammals of the seas, vol. 4: Small cetaceans and sirenians, p. 3-19. FAO Fish. Ser. 5. Kasuva. T., and L.L. Jones 1984 Behavior and segregation of the Dall's porpoise in the northwestern North Pacific. Sci. Rep. Whales Res. Inst. Tokyo 35:107-128. Kasuya. T.. and H. Ogi 1987 Distribution of mother-calf Dall's porpoise pairs as an in- dication of calving grounds and stock identity. Sci. Rep. Whales Res. Inst. Tokyo 38:125-140. Kasuya. T.. and S. Shiraga 1985 Growth of Dall's porpoise in the western North Pacific and suggested geographical growth differentiation. Sci. Rep. Whales Re.s. Inst. Tokyo 36:139-152. Kenagy. G.J.. and S.C. Trombulak 1986 Size and function of mammalian testes in relation to body size. J. Mammal. 67:1-22. Kleiman, D.G. 1977 Monogamy in mammals. Q. Rev. Biol. 52:39-69. Landino. S.W. 1985 Cetacean breeding systems revealed by interspecific com- parison [Abstr.] Six-th Bienn. Conf. Biol. .Mar. Mammals, \'an- couver, B.C., Canada. Soc. Mar. Mammal. Leatherwood. S.. and W.A. Walker 1979 The northern right whale dolphin, Lissodelphis horealis Peale in the eastern North Pacific. In Winn, H.E., and B.L. 011a (eds.). Beha\ior of marine animals, vol. 3: Cetaceans, p. 85-141. Plenum Press. NY. Leatherwood, S., R.R. Reeves, W.F. Perrin, and W.E. Evans 1982 Whales, dolphins and porpoises of the eastern North Pacific and adjacent Arctic waters. NOAA Tech. Rep. .NMFS Circ. 444, Natl. Oceanic Atmos. Adm., Natl. Mar. Fish. Serv., 245 p. 132 Fishery Bulletin 88(1). 1990 Lincoln, G.A. 1971 The seasonal rejjroduclive changes in the red deer stag (CcTW-s ehiphus). J. Zool. (Lond.) 163:105-123. Mizue. K., and K. Yoshida 1965 On the porpoises caught by the salmon fishing gill-net in Bering Sea and the North Pacific Ocean. Bull. Fac. Fish., Nagasaki Univ. 19:1-36. Morejohn, G.\. 1979 The natural liistory of Dall's porpoise in the North Pacific Ocean. In Winn, H.E., and B.L. 011a (eds.). Behavior of marine animals, vol. 3: Cetaceans, p. 4.5-83. Plenum Press, NY. Morejohn, G.V., V. Loeb, and D.M. Baltz 1973 Coloration and sexual dimorphism in the Dall porpoise. J. Mammal. 54:977-982. Newby. T.C, 1982 Life history of Dall porpoise {Phococnoides dalli, True 1885) incidentally taken by the Japanese high seas salmon mothership fishery in the northwestern North Pacific and western Bering Sea, 1978 to 1980. Ph.D. diss., Univ. Wash., Seattle, 155 p. Nishiwaki, M., S. Ohsumi, and Y, Maeda 1963 Change of form in the sperm whale accompanied with growth. Sci. Rep. Whales Res. Inst. Tokyo 17:1-14. Norris. K.S, 1961 Standardized methods for measuring and recording data on the smaller cetaceans. J. Mammal. 42:471-476. 1967 Aggressive behavior in Cetacea. /nClemente, CD., and D.B. Lindsley (eds.), Aggression and defense, p. 225-241. Univ. Calif. Press, Berkeley. Norris, K.S., and J.H, Prescott 1961 Observations on Pacific cetaceans of Californian and Mex- ican waters. Univ. Calif. Publ. Zool. 63:291-402. Norris, K.S., B. Wursig, R,S, Wells, M, Wiirsig. S.M. Brownlee, C. Johnson, and J. Solow 1985 The behavior of the Hawaiian spinner dolphin. Stenelta longirostris. Unpubl. Admin. Rep. LJ-85-06C, Southwest Fish. Cent.. Natl. Mar. Fish. Serv., NOAA, La Jolla, CA 92038, 213 p. Perrin, W.F. 1972 Color patterns of spinner porpoises {Stenella cf. S. longi- rostris) of the eastern Pacific and Hawaii, with comments on delphinid pigmentation. Fish. Bull., U.S. 70:983-1003. 1975 Variation of spotted and spinner porpoise (genus Stenella) in the eastern tropical Pacific and Hawaii. Bull. .Scripps Inst. Oceanogr. Univ. Calif. 21, 206 p. 1984 Patterns of geographical variation in small cetaceans. Acta Zool. Fenn. 172:137-140. Perrin, W,F.. and J.R. Henderson 1984 Growth and reproductive rates in two populations of spin- ner dolphins, Stenella longirostris, with different histories of e.xploitation. Rep. Int. Whaling Comm. (Spec. Iss. 6):417-430. Perrin. W.F., D.B, Holts, and R.B. Miller 1977 Growth and reproduction of the eastern spinner dolphin, a geographical form of Stejielln longirostris in the eastern tropical Pacific. Fish. Bull., U.S. 75:725-750. Ralls. K. 1977 Sexual dimorphism in mammals: Avian models and un- answered questions. .'\m. Nat. 111:917-938. Robineau, D. 1984 Morphoigie externe et pigmentation du dauphin de Commerson, Cepkalorhynchus commersonii (Lacepede. 1804), en particulier celui des iles Kerguelen. Can J. Zool. 62: 2465-2475. Schaller, G.B, 1963 The mountain gorilla: Ecology ami behavior. Univ. Chicago Press, 429 p. 1972 The Serengeti lion: A stuity of pri'dalnr-prcy relaticms. Univ. Chicago Press, 480 |). Scheffer, V.B. 1949 The Dall porpoise, Phocoenoidrs dalli. in Alaska. J. Mammal. 30:116-121. Sergeant, D.E. 1962 ( )n the external characters of the bhuklish or pilot whales (genus Globirephala). J. Mammal. 43:394-413. Short, R.V. 1979 Sexual selection and its component parts, somatic and genital selection, as illustrated by man and the great apes. Adv. Study Behav. 9:131-158. Slooten, E,, and S.M. Dawson 1988 Studies on Hector's dolphin, Cephalorlninchiis hcrlori: A progress report. Rep. Int. Whaling Comni. (Spec. Iss. 9):325-3,38. Smith, B.L. 1988 Criteria for determining age and sex of American moun- tain goats in the field. J. Mammal. 69:395-402. Subramanian, A., S. Tanabe, Y. Fujise, and R. Tatsukawa 1986 Organochlorine residues in Dall's and True's porpoises collected from northwestern Pacific and adjacent waters. Mem. Natl. Inst. Polar Res.. Ser. E. Biol. Med. Sci., Spec. Iss. 44:167-173. .Subramanian, A. S. Tanabe. and K. Tatsukawa 1987 Age and size trends and male-female differences of PCBs and DDE in dalli-type Dall's porpoises, Phoroeytnidcs dalli of northwestern North Pacific. Proc, Natl. Inst. Polar Res Symp. Polar Biol. 1:205-216. Trivers, R.L. 1985 Social evolution. Benjamin/Cummings Publ. Co., Menlo Park, CA, 462 p. Abstract.- We determined the patterns of disti'ibutit)n for eggs and larvae of walleye pollock by analyz- ing 1,929 ichthyoplankton samples collected on 32 cniises in the western Gulf of Alaska between 1972 and 1986. The combined effects of addi- tions of recently spawned eggs, mor- tality, dispersion, and advection deter- mine the location and concentrations of eggs and larvae. The vast major- ity of the eggs were found in spring samples, primarily in April. Larvae were found mainly in late April and May. Eggs occurred mostly in a small area near Cape Kekurnoi in Shelikof Strait, and larvae were centered progressively to the south- west of this area as they grew and were moved by advection. By the end of May, they were usually found near the Semidi Islands. However, comparisions of larval distributions in late May showed significant inter- annual variations in the area of con- centration and larval size. Egg and Larval Distributions of Walleye Pollock Theragra chalcogramma \n Shelikof Strait Gulf of Alaska Arthur W. Kendall, Jr. Susan J. Picquelle Alaska Fisheries Science Center, National Marine Fisheries Service, NOAA 7600 Sand Point Way N E , Seattle. Washington 981 15-0070 Manuscript accepted 18 September 1989. Fishery Bulletin, U.S. 88:133-1,54. Walleye pollock Theragra chalcogram- ma is the dominant gadid in the sub- arctic Pacific Ocean and in the Ber- ing Sea. It is a moderate-sized (to about 70 cm standard length) pelagic or semidemersal fish that frequently occurs in large aggi'egations. It feeds primarily on copepods and euphau- sids, although fish, including its own young, enter its diet (see Lynde 1984). Intensive multinational midwater trawl fisheries are conducted for walleye pollock in the Gulf of Alaska, the Bering Sea, and the northwest Pacific Ocean. In recent yeai's, cat Kekurno. ^— . ?5^T N" W1 SuT.v,kl. J ,- _ _, I ' I ' I ' 1= 60 OON I- 59 00 58 00 - 57 00 56 00 55 00 54 00 65 OOW ' 50 00 I 5 5 00 150 0 0 1 4 5 00 Figure 1 Gulf of Alaska showing the six strata used to analyze distribution of walleye pollock eggs and larvae. Gridded area was used in detailed analyses of the Shelikof Strait region. The 32 cruises were conducted for a variety of pur- poses, so not all station patterns are ideal for the pres- ent analysis. Some results from these cruises have been reported elsewhere (Bates and Clark 1983, Dunn et al. 1984, Kendall and Dunn 1985, Bates 1987, Kendall et al. 1987). Although several ships were used, stan- dard MARMAP oblique tows with 60-cm bongo nets (Smith and Richardson 1977) were. conducted at all sta- tions, with two exceptions: smaller bongo nets (20 cm) were used in 1984 aboard the NOAA ship Chapman and in April 1985 aboard the Miller Freeman. Mesh size was generally 505 j^m, although 333-/jm mesh was used aboard the NOAA Ship Miller Freeman in 1985 and 1986. On most cruises, sampling was conducted to a maximum depth of 200 m, as water depth permitted, but on the Miller Freeman in 1985 and 1986 sampling was conducted to within about 10 m of the bottom. Data collected from flowmeters in the mouth of the nets and bathykymograph records (for most tows) allowed us to determine the maximum depth and volume of water filtered for each tow. More details of the conduct of these cruises are found in Dunn and Rugen (1989). Samples were preserved at sea in a 5% buffered for- malin/seawater solution and returned to shore for pro- cessing. All fish eggs and larvae were then sorted from the samples, identified, counted, and the larvae were measured (mm standard length, SL): total sample when larvae numbered fewer than 50, or a random subsam- ple of at least 50 when more than 50 larvae were in a sample. In each sample, developmental stages were determined for 100 randomly selected walleye pollock eggs according to the scheme of Matarese (Alaska Fish. Sci. Cent., Natl. Mar. Fish. Serv., NOAA, 7600 Sand Point Way NE, Seattle, WA 98115-0070, unpubl. manuscr.) (see Kendall and Kim 1989). Egg develop- ment data were then grouped into the six age groups used by Kendall and Kim (1989). Numbers of eggs and larvae in the tows were standardized to numbers per 10-m- sea surface by calculating the volume of water filtered during the tow and dividing this number by the maximum depth of the tow. Since most tows were taken to a maximum depth of 200 m, and substantial numbers of eggs occurred below this depth, we adjusted catches of eggs, based on data in Kendall and Kim (1989), to account for eggs below 136 Fishery Biilletin 88(1). 1990 T" I 80 30 1 TT I I 90 10 — r I 100 —r I no 301 I 120 I I 0 -r I 140 30 1 TTi I 150 10 T- Julian Day "T I 70 ; 89 H 79 I 160 0 TIME INTERVAL Figure 2 Cruise periods and time intervals used to analyze the distribution of walleye pollock eggs and larvae in Shelikof Strait. Numljers of tows are shown in blocks indicating cruise times. the sampling depth. The percent of eggs that are ex- pected to occur in the depth interval {A,B) is given by: P.4,j = (0.06883) • [exp(0.02913-B) - exp(0.02913-A)] where P4 g = percent of eggs in depth interval (A,B); A = upper limit of depth interval, >40 m; and B = lower Hmit of depth interval, <250 m. We assumed that no eggs occur at depths shallower than 40 m or deeper than 250 m. The adjustment was made by multiplying the catch per 10 m- by the ratio of the theoretical percentage of eggs that occur be- tween 40 m and the bottom depth or 250 m (whichever is shallower) to the theoretical percentage of eggs that occur between 40 m and the maximum tow depth: Correction factor 40, bottom depth or 250 "40. tow depth In order to compare walleye pollock egg and larval catches at time of maximum abundance in various parts of the western ( lulf of Alaska, we divided the gulf into six strata and examined egg and larval catches for each stratum (P"'ig. 1). For years in which more than one stratum was sampled (and at least three stations were sampled in a stratum), we used catches of eggs during April and catches of larvae from 16 May to 8 June for that year. P^or these cases, we calculated an area- weighted mean abundance (Sette-Ahlstrom method [see Appendix 1]) and percent contribution for each stratum. For detailed analysis of the Shelikof Strait area, only those stations from Shelikof Strait to the Shumagin Islands were considered (see Fig. 1). Since the major spawning of walleye pollock in the western Gulf of Alaska occurs in Shelikof Straii in early spring (as we will show later), only cruises conducted in March through early June were considered in the detailed analysis of Shelikof Strait. In order to establish the Kendall and Picquelle Egg and larval distributions of Thersgrs chslcogramrm 137 general pattern of occurrence of eggs and larvae as the season progressed, this time period was divided into five intervals (12-28 March, 29 March-13 April, 14-29 April, 30 April- 15 May, and 16 May-2 June), and samples collected during each interval were grouped for analysis regardless of cruise or year (Fig. 2). In- tervals of a 16-18 day duration were chosen because the egg incubation period is about this duration at am- bient temperatures, and lai^vae would grow about 3 mm in this interval (at an observed growth rate of about 0.2 mm per day [Kendall et al. 1987]). Thus eggs or lar- vae within a 3-mm length increment from each time interval can be considered a cohort when comparing their distribution with that of adjacent time intervals. Similarly, since sampling patterns and positions were different for each cruise, the Shelikof Strait area was divided into a number of sectors, and the samples that were collected in each time interval and each geo- graphic sector were grouped so they could be analyzed regardless of cruise or year. The grid of sectors was laid out with axes parallel and perpendicular to the axis of Shelikof Strait (Fig. 1). Within Shelikof Strait prop- er, where heaviest concentrations of stations and abun- dances of eggs and larvae occurred, the sectors were 10 X 10 miles (18.5 x 18.5 km); southwest of this area the sectors were 20 x 20 miles (37 x 37 km). The geographic distributions of walleye pollock eggs and larvae were summarized by mapping their cen- troids of distribution (see Appendix 2). Centroids were computed for total walleye pollock eggs and larvae by time interval regardless of year, and by egg age-group and 3-mm larval length-interval (3.0-5.9, 6.0-8.9, 9.0-11.9, 12.0-14.9, 15.0-17.9 mm SL). The centroid can be considered the center of mass of the distribu- tion and is computed as the weighted mean location of the sectors, where the weights are the estimated total number of eggs or larvae for each sector (Koslow et al. 1985): Centroid = (A', f ) I A^, • X, where A' = , I AT, 2 A^, • Y. Y I AT,. N, = estimated total number of pollock in sector (■ A, ■ J.N.., A, = area of sector i (in units of 10 m-) A'^,, = number of pollock per 10 m- for sample j in sector (', w, = number of samples in sector i, X, = center position of sector i in the east- west axis X,i = position of sample J in sector ; in the east-west axis, Y, = center position of sector / in the north- south axis I N,j ■ Y,j J and Y, , = position of sample j in sector ; in the north-south axis. The distribution was further analyzed by rotating the A' and Y axes so that one axis was in the direction of the greatest geographic range of walleye pollock eggs or larvae in the study area. This axis is called the prin- cipal axis. The second axis is orthogonal to the prin- cipal axis. An ellipse was drawn around the centroid that is one standard deviation away from the centroid along the rotated_axes. The ellipse is defined as having the center at (A', }') and the major axis parallel to the principal axis. The major and minor axes are each two standard deviations long. The ellipse is the two-dimensional analogue of a mean and standard error bars; it shows the center and the orientation of the animal's distribu- tion in space and the amount of dispersion about the center. If the distibution of animals in space follows a bivariate normal distribution, then the ellipse is a 40% confidence contour. Total abundance in the Shelikof Strait area was estimated for eggs and larvae for each time interval. Total abundances were also estimated for the six egg age-groups and the five larval length-increments for each time interval. The sectors were treated as strata to correct for the higher density of samples in areas of high egg and larval abundances. Not all sectors were sampled in each time interval, so the estimate of total abundance was adjusted to the total area. This was ac- complished by grouping the sectors into two regions, (one included all the 10 x 10 mile sectors, and the other included all the 20 x 20 mile sectors), estimating a stratified mean density for each region, then multiply- 138 Fishery Bulletin 88(1). 1990 Table 3 Abundance of walleye pollock eggs and larvae in Shelikof Strait bv time interval. Totals are expanded to total area of all sectors: (5.556 X lO'W). Eggs Larvae Time No. sectors with samples interval Dates Total (10'2) % SD (10'-) Total (W) % SD (10'-) 1 12-28 Mar. 0.710 1 0.0314 <0.001 0 <0.001 48 2 29 Mar.-13 Apr. 89.6 67 10.0 0.236 1 0.0385 68 3 14-29 Apr. 38.9 29 1.69 21.1 66 2.01 64 4 30 Apr.- 15 May 2.91 2 0.198 5.32 17 1.24 67 5 16 Mar.-2 June 0.681 1 0.0427 5.09 16 0.515 74 ing the density by the total area of the region (Jessen 1978). N =I.A,, ■ Ni, h where A'' = estimate of total abundance, Aji = area of region h (in units of 10 ni-) A'';, = mean density in region h (number/ 10 m-) ^Ah, ■ N,,, A;, , = area of sector i in region /; (in units of 10 m-) Nil, = mean density in sector / in region h (number/10 m-) Nfiij = number/10 m- in sample ,/ in sector / in region h. and n.),, = number of samples in sector / in region /;. The estimate of variance was adjusted to correct for sectors with one or no sample because a minimum of two samples are required to estimate variance for each sector (see Appendix 3). Var A^ = Z .4,,- • Var A^,, where I A,,,:- Var TV, /,,- Var A^,, A,2 1 + 'o 1+ refers to sectors with at least two samples, i„ refers to sectors with one or zero samples, and VariV,,,^ Results I.(N„,^^-N,,J^ ("A - 1) • n. General features of egg and larval distribution Season of occurrence Of the 696 tows taken during the surveys from 1972 to 1979, 91 contained walleye pollock eggs and 76 contained larvae (Table 1). Be- tween 28 March and 24 May, 205 tows were taken (29% of the total) and these included 82 and 75% of the tows containing walleye pollock eggs and larvae, respective- ly. A few of the tows taken between October and March also contained eggs, and the 88 tows in June and July accounted for an additional 20% of those containing larvae. From these results it appears that nearly all walleye pollock spawning and the egg and larval period in the Gulf of Alaska occurs in April and May (Table 2). Limited numbers of eggs are spawned earlier than this, and some larvae are available to plankton sam- pling after this. Among the five time intervals between 12 March and 2 June established to investigate occurrences of eggs and larvae in the Shelikof Strait area, the second time interval (29 March to 13 April) accounted for 67% of the eggs, and the third time interval (14 to 29 April) accounted for 66% of the larvae (Table 3). The third time interval accounted for 29% of the eggs, and the fourth and fifth intervals accounted for 17 and \6%i of the larvae, respectively. These results indicate that spawning in Shelikof Strait occurs mainly in early April, and that the larvae resulting from this spawn- ing are present in the area in decreasing numbers into June. Areas of occurrence Comparing the catches of wall- eye pollock eggs from April and larvae from 16 May to 8 June from the six strata, H9% of the eggs and 45% Kendall and Picquelle: Egg and larval distributions of Thersgra chatcogramma 139 Table 4 Abundance estimates (x 10'") of walleye pollock eggs and larvae in six strata of the western G number of stations sampled, and the bottom number is the number of eggs or larvae. alf of Alaska. The top number is the Year Amatuli Kodiak Shelikof Chirikof Semidis- Shumagin Sanak Eggs (April) 1981 [32.7]- 15 1.16 104 1670 13 102 16 1.60 [57.9] 1982 [47.2] 12 1.24 14 907 11 99.0 24 21.6 20 48.2 1984 73 29.0 3 1.61 88 2240 23 0.510 28 0.744 [19.0] 1985 23 28.8 18 2.74 160 1220 20 177 5 0.434 [54.1] 1986 45 11,3 15 2.47 82 635 17 13.5 22 13.1 30 88.3 Mean no. eggs Percent 29.8 1.98 1.84 0.122 1330 88.6 78.4 5.20 7.50 0.498 53.5 3.55 Larvae (16 iMa.v-8 June) 1979 11 0.0409 13 0.00 8 .38.4 9 0.0432 10 0.196 7 0.520 1981 [24.2] 11 (1. 0181 93 99.7 13 27.9 15 42.6 [10.4] 1982 [4.80] 4 2.37 15 1.14 13 1.13 23 3.39 6 0.997 1985 40 55.7 3 0.807 49 5.06 23 0.892 30 0.181 29 2.00 Mean no. larvae Percent 21.2 0.798 36.1 7.50 11.6 26.3 0.990 44.7 9.30 14.4 )r abundance estimates in brackets: these values were predicted using a technique from analy ta in a randomized block experiment (Cochran and Cox 1957). where the strata were considered observed abundance estimates were log transformed. 3.47 4.30 sis of variance as treatments. *No data were available ft for estimating missing da the years as blocks, and of the larvae occurred in the Shelikof stratum (Table 4). The ne.xt most important stratum for eggs was Chirikof, where 5.2% occurred. The Amatuli stratum contained 2(3% of the larval catch, and the Semidis- Shumagin stratum contained 14%. Larval catches in the Amatuli stratum were based mainly on sampling done in 1985, a year in which larval occurrences in Shelikof Strait were dissimilar to all other years sam- pled (see later). Distribution of eggs in Slielil, = day of year when sample i was taken. 146 Fishery Bulletin 88(1), 1990 cruise permits a comparison among years, and provides an estimate of the increase in length of fish in the population with time (Fig. 9). The mean length of larvae in the population at a particular time results from a combination of time since spawning and individual growth, as well as the effects of mortality. Older lar- vae will have been subjected to mortality for longer periods than younger larvae, so the change in length of larvae in the population should be less than individual growth rate. In 1978, 1979, and 1983, only one cruise sampled larvae. However, in 1981 and 1985, five and four cruises, respectively, sampled larvae. Altogether, 20 cruises were available to assess larval-population length increase. Before the third time interval, little increase in the length of larvae in the population or variation among years in size of larvae was evident. The mean lengths were 3-6 mm. Larvae at the end of April were smallest in 1981 and largest in 1985. After the middle of May, the larval population length seemed to increase at a fairly uniform and linear rate through the end of May. In 1981, 1982, and 1985, larvae in late May were be- tween 7 and 9 mm. In 1983, however, larvae were over 10 mm in late May. Excluding the larval lengths from 1983, the length data from the third through fifth time intervals fit the straight line of length (mm SL) = -9.60 + 0.122 ■ day of year. 1 1 1 1 1 1 1 1 1 4/14 4/30 Date Veat 78 A 79 + 83 -«^ 84 -A -B- 32 -E- 86 Figure 9 Lengths of walleye pollock larvae l)y .sanipling date and year. with r- = 89.7. Thus the population increased in length at a rate of about 0.12 mm/day. Based on analysis of larval otolith daily growth increments, in- dividual growth in 1983 was 0.2 mm/day (Kendall et al. 1987). Distribution of cohorts of eggs and larvae in Shelikof Strait The five time intervals of 16-18 days were about equal to the incubation period of the eggs, and the larvae grew about 3 mm (~0. 1 7 mm/day) during each time in- terval, after hatching at 3-4 mm. Thus eggs from one time interval were assumed to be the 3-6 mm larvae observed in the following time interval, the 6-9 mm larvae in the time interval after that, and so on. Under this assumption three cohorts (1-3) were established, based on eggs from the first, second, and third time intervals, regardless of year, in order to examine the geographic displacement of the eggs and larvae dur- ing development (Table 6). Changes in abundance with time in the cohorts is influenced by differences among years in sampling and overall population abundances, which prevent estimation of mortality rates. Eggs in the first cohort, those from the first time in- terval, were slightly northeast of larvae in this cohort, which were represented by 3-6 mm larvae from the second time interval (Fig. 10). The centroid of larvae that were 3-6 mm long in the second time interval was in the southern end of Shelikof Strait, midway between the southern end of Kodiak Island and the Alaska Peninsula. In the third time interval the first cohort (6-9 mm larvae) was centered further to the southwest, just north of a line between Sutwik and Chirikof Islands. In tlie fourth time interval the 9-12 mm lar- vae were centered near the position of the centroid of the 3-6 mm larvae in the second time interval. The ellipse for these larvae was quite elongate (Table 6), indicating considerable dispersion on the along-strait axis. This pattern reflects differences in the distribu- tion of this size larvae among years, particularly in the abundance of these larvae in the northern part of Shelikof Strait in 1985. The centroid of the 1 2- 1 5 mm larvae from the fifth time interval is positioned in the vSemidi Islands, and the ellipse is broadened in the across-strait dimension compared with the previous ones. Thus, excluding the fourth time interval, which was influenced by the aberrant pattern of 1985, the first cohort of eggs and larvae appears to have been advected to the southwest in the center of the strait from the southern end of Kodiak Island to the Semidi Islands between early April and late May. Kendall and Picquelle Egg and larval distributions of Theragra chalcogramma 147 Table 6 Abundance ( x 10'') and dimensions of e lipses (of three cohorts) of walleye pollock eggs and larvae in Shelikof Strait by time interval and 3-mm length increment for larvae Cohorts are indicated by the diagonal numbered boxes. Areas in km'-; axes in km. Time interval Dates Cohort Eggs Larvae Length increments (mm SL) 3-6 6-9 9-12 12-15 15-18 1 12-28 Mar. Abundance Area Major axis Minor axis \ 1 . ().71{)\ \ 1340 \ \ ^''•^ \ \ 21 <0.001 0.000 0.000 0.000 0.000 2 29 Mar.-13 Apr. Abundance Area Major axis Minor axis \,. 89. 6\ \ 2080 \ \ ■'■^ \ \ 28 ^ 0.23 1\ 3400 \ 132 \ 33 0.1)04 0.000 0.000 0.000 3 14-29 Apr. Abundance Area Major axis Minor axis \ \ ^^'^ \ 19. 7\ 2440 \ 104 \ 30 0.204\ 1.570 \ \ 95 ^ \ "' 0.001 0.000 0.000 4 30 Apr.- 15 May Abundance Area Major axis Minor axis 2.9 1\ 4.79\ 2030 \ lOS \ ^^ 0.512\ 4920 \ \ 100 \ \ 39 0.012\ 6310 \ \ -•^" \ \ 35 0.000 0.000 5 Ifi May-2 June Abundance Area Major axis Minor axis 0.681 0.27.5\ 3.40\ 4780 \ \ 117 \ \ 52 1.27\ 9220 \ 183 \ \ 111 ^ 0.137\ 8910 \ ItiO \ 0.004 \ The second cohort, represented by eggs from the sec- ond time interval and 3-6 mm larvae from the third time interval, resulted from the period of heaviest spawning activity. Most of the eggs and larvae collected were in this cohort. The centroid of the eggs of this cohort, in the second time interval, was in the middle of Shelikof Strait, off Cape Kekurnoi (Fig. 10). Dur- ing the third time interval, the centroid of larvae of the second cohort (3-6 mm long) was located east of Wide Bay, southwest of the centroid of eggs of this cohort. In the fourth time interval, 6-9 mm larvae were further to the southwest, off the south end of Kodiak Island. By the fifth time interval, 9-12 mm larvae were near a line between Sutwik and Chirikof Islands. The ellipses for this cohort of eggs and larvae expanded between each successive time interval (Table 6). The centroid of eggs of the third cohort was found in the center of Shelikof Strait, off Wide Bay (Fig. 10). The centroid of this cohort as 3-6 mm larvae (during the fourth time interval) was positioned at the southern end of Shelikof Strait, closer to Kodiak Island than to the Alaska Peninsula which was slightly south of where the eggs were. In the fifth time interval the centroid of this cohort as 6-9 mm larvae was between Sutwik Island and the Trinity Islands. 148 Fishery Bulletin 88(1), 1990 , n 1 1 . . I . ' ' .1 ■ ' . I . . I . ..I . . I . .-I- I I I I I I I I' I ■ I ■ ■ I ■ ' I ' ■ I Figure 10 ('entrriiiJs of tlirt-t' ccjhnrts of walleye piillock egg and larval distributions in Shelikof Strait. Labels beside dots indicate stage (egg) i>r length increment, with time interval in parentheses. Discussion As we compiled data for analysis, we made several assumptions which may have affected our results. Most of the tows used to describe egg distributions did not exceed a maximum depth of 20(1 m and, therefore, a correction factor was applied to account for eggs in deeper parts of the water column. Our key assumption for this correction was that the depth distribution of eggs found in Shelikof Strait in 1985 and 1986 was ap- plicable throughout the study area and for all years ex- amined. The validity of this assumption needs to be verified by determining geographic and interannual variability in depth distribution and bouyancy of eggs. Specific gravity and depth distribution of the eggs change with development (Kendall and Kim 1989), but these changes were not accounted for in our analysis. In many of the analyses, there was an implicit as- sumption that samples were drawn randomly from within a sector or stratum. Some of what appears to be random variability in the data may be features of distribution on a scale too fine for our analyses. In com- l)ining data from various years with slightly different sampling times and station patterns, small-scale inter- annual differences may have been blended. Effects of different sampling intensities among the years may have interacted with changes in overall abundance of eggs and larvae, so that patterns in low adundance years may have been overwhelmed by patterns from high abundance years. However, aside from the un- usual spatial distribution of larvae in 1985, and the large size of larvae in late May lif8.3, timing and geo- graphic areas of occurrence of eggs and larvae seemed (juite consistent among years. Knowledge of the distribution of eggs and larvae of walleye pollock in the Gulf of Alaska has increased dramatically in recent years. The larvae could not be separated from those of Pacific cod Gadus macroceipha- lus prior to studies of Matarese et al. (1981). Kendall and Dunn (1985) summarized knowledge of ichthyo- Kendall and Picquelle Egg and larval distributions of Ther^gra chalcogramma 149 plankton, including walleye pollock in the Gulf of Alas- ka based on sampling prior to 1980. Very little sam- pling had been done before 1977, when a series of cruises was conducted mainly southeast of Kodiak Island. Eggs were found mainly nearshore in spring, and to a lesser extent in fall. Larvae were found in spring and summer throughout the shelf and slope area sampled. The spawning concentration in the Shelikof Strait was discovered inadvertantly in 1980. The mean number of eggs and larvae per 10 m- in cruises re- ported by Kendall and Dunn (1985) did not exceed 40 and 228 respectively, whereas we found mean numbers per cruise as high as 177,000 for eggs and 13,700 lar- vae in Shelikof Strait (Table 2). Based on the patterns of distribution elucidated here, several topics for future work are indicated. By com- paring distributions of cohorts of eggs and larvae, we inferred advection rates and directions. The reason- ableness of these drift patterns needs to be investigated based on the physical oceanography of the area. Although these results show that most spawning oc- curs in Shelikof Strait, evidence of some spawning south of Chirikof Island was observed. Surveys of adults producing the eggs in this area need to be con- ducted so that the importance of the area can be evaluated. The source of the eggs and larvae found in the Amatuli area and the northern part of Shelikof Strait in 1985 also needs to be established. The cruises available for this study were made only through early June. By this time the larvae had drifted a considerable distance southwest of the spawning area and grown to about 8-11 mm. The distribution of lar- vae later in spring and summer is virtually unknown. To understand recruitment in this population, the lar- vae need to be followed through summer to find the location of their nursery area. Such information is a necessary prerequisite to studies of factors contribut- ing to survival throughout the planktonic stages of walleye pollock. During the years of study here, the adult spawning biomass of walleye pollock in Shelikof Strait declined from 3.77- 10*^ mt in 1981 to 0.62- 10'^ mt in 1986 (Nelson and Nunnallee 1987). The planktonic egg data from these years can be analyzed to see how closely they reflect this decrease in spawning biomass. Based on the analysis here, a model of seasonal and geographic egg production could be developed that would allow extrapolation of expected egg catches when sampling did not cover the entire area of their occurrences. Given the interannually consistent patterns of egg distribution in Shelikof Strait described here, there is a reasonable opportunity to establish egg and larval mortality rates based on changes in population abun- dance during April and May. The timing of changes in mortality could be determined, and differences in mor- tality rates among years could be investigated. This can lead to studies of the environmental and biological causes of mortality and how these vary during the season and interannually. Knowledge of these varia- tions in mortality rates is critical to an understanding of recruitment patterns in this population. Summary Based on analysis of walleye pollock egg and larval distributions in 1,929 MARMAP bongo tows taken on 32 cruises in the northern Gulf of Alaska from 1972 to 1986, we made the following conclusions: 1 Most eggs occur in April, and larvae occur in late April through May. 2 Most eggs and larvae result from spawning in Shelikof Strait near Cape Kekurnoi. 3 Little difference in geographic distribution of eggs was seen through the season or among the years sampled. 4 Larvae occurred progressively to the southwest of the area of egg occurrence, as the larvae developed through the season. 5 Larval population averaged from 4.2 to 4.9 mm in mid April, and reached 7.5 to 9.1 mm by the end of May, except in 1983 when the population averag- ed 10.5 mm at the end of May. 6 Considerable variation in distribution of larvae was seen by late May among the years sampled, in- dicating differences in amount of advection. Acknowledgments Sample collection and processing, as well as the data entry and editing that went into producing this publica- tion, required contributions of persons too numerous to mention; however, without them we could not have considered undertaking this project. Suffice it to say that the environment that walleye pollock find suitable for spawning is often extremely inhospitable to those on ships collecting their eggs and larvae, and we ap- preciate the efforts of all involved. Most samples were sorted and fish eggs and larvae identified, counted, and staged (eggs) or measured (larvae) at the Polish Plank- ton Sorting Center under the direction of Dr. Leonard Ejsymont, whose diligence in overcoming the many problems associated with such demanding work, as well as the challenges of a cooperative program between two countries, is gratefully acknowledged. Many peo- ple at NWAFC assisted in processing egg and larval samples, but we particularly want to thank Ann Matarese, Beverly Vinter, and Deborah Blood. Jay Clark and Richard Bates at NWAFC made the data 150 Fishery Bulletin 88(1), 1990 required ready for analysis and Bates performed some of the analyses. Dr. Suam Kim, University of Wash- ington and NWAFC, advised us on depth adjustments of egg catches and on the use of centroids for our analysis. Dr. Paul Smith, Southwest Fisheries Center, and Dr. Don Gunderson, University of Washington, reviewed and made valuable comments on an earlier draft. Citations Alton, M.S. 1987 Walleye pollock fishery in the Gulf of Alaska, 1985-86. In Major, R.L. (ed.), Condition of groundfish resources of the Gulf of Alaska region as assessed in 1986, p. 1-13. NCAA Tech. Memo. NMFS-F/NWC-119, Northwest and Alaska Fish. Cent.. Natl. Mar. Fish. Serv., Seattle. WA 9811,5-0070. Bates, R.D. 1987 Ichthyoplankton of the Gulf of Alaska near Kodiak Island. April-May 1984. NWAFC Proc. Rep. 87-11, Northwest and Alaska Fish. Cent.. Natl. Mar. Fish. Serv., NOAA. Seattle. WA 98115-0070, ,53 p. Bates, R.D.. and J. Clark 1983 Ichthyoplankton off Kodiak Island and the Alaskan Penin- sula during spring 1981. NWAFC Proc. Rep. 83-09. North- west and Alaska Fish. Cent., Natl. Mar. Fish. Serv.. NOAA, Seattle, WA 98115-0070, 105 p. Cochran. W.G., and G.M. Cox 1957 E.xperimental desigiis. 2nd ed. John Wiley & Sons, NY, 611 p. Dunn, J.R., and A.C. Matarese 1987 A review of the early life history of northeast Pacific gadoid fishes. Fish. Res. 5:163-184. Dunn, J.R., and W, Rugen 1989 A catalog of Northwest and Alaska Fisheries Center ichthyoplankton cruises 1965-1988. NWAFC Proc. Rep, 86-08, Northwest and Alaska Fish, Cent.. Natl. Mar, Fish, Serv,. NOAA, Seattle, WA 98115-0070, 78 p, 85 figs. Dunn, J.R., A,W, Kendall. Jr.. and R.D. Bates 1984 Distribution and aliundance patterns of eggs and larvae of walleye pollock {Theragra chdlcogrtniitna) in the western Gulf of Alaska. NWAFC Proc, Rep. 84-10, Northwest and Alaska Fish. Cent.. Natl, Mar, Fish, Serv., NOAA, Seattle, WA 98115- 0070, 66 p. Jessen, R,J, 1978 Statistical survey techniques. .John Wiley & Sons. NY. 520 p. Kendall, A.W. Jr., and J.R. Dunn 1985 Ichthyoplankton of the continental shelf near Kodiak Island, Alaska. NOAA Tech. Rep. NMFS 20. Natl. Oceanic Atmos. Admin.. Natl. Mar, Fish. Serv., Seattle, WA 9811.5- 0070, 89 p. Kendall, A.W. Jr., and S. Kim 1989 Buoyancy of walleye pollock {Theragra chalcogramma) eggs in relation to water properties and movement in Shelikof Strait, Gulf of Alaska. Jn Beamish. R.J., and G. A. McFarlane (eds.). Effects of ocean variability on recruitment and an evalua- tion of parameters used in stock assessment models, p. 169-180. Can. Spec. Publ. Fish. Aquat. Sci. 108, Kendall, A.W. Jr., M.E. Clarke, M,M, Yoklavioh, and G.W. Boehlert 1987 Distribution, feeding, and growth of larval walleye pollock, Theragra chalcogramma, from Shelikof Strait, Gulf of Alaska. Fish. Bull,, U,S, 85:499-521, Koslow, J, A., S. Branlt. J. Dugas, and F. Page 1985 Anatomy of an apparent year-class failure: the early life history of the 1983 Browns Bank haddock Mdanogrammas neglefiniix. Trans. Am. Fish, Soc. 114:478-489. Lynde. CM. 1984 Juvenile and adult walleye pollock of the eastern Bering Sea: literature review and results of ecosystem workshop. hi Ito, D.H. (ed.). Proceedings of the workshop on walleye pollock and its ecosystem in the eastern Bering Sea, p. 43-108. NOAA Tech. Memo. NMFS-F/NWC-62. Northwest Alaska Fish. Cent,. Natl. Mar. Fish. Serv., Seattle, WA 98115-0070, Matarese, A.C, S,L, Richardson, and J.R. Dunn 1981 Larval development of Pacific tomcod. Minngadux prnx- imua. in the northeast Pacific Ocean with comparative notes on larvae of walleye pollock, Theragra chalcogramma and Pacific cod, Gndiis mairocephatu.;h(-rv Gullenn 88(I|, 1990 And the variance of the total abundance is estimated by Var N = A- ■ War N. Appendix 2 The centroid is simply the weighted bivariate mean of the sampled locations of the eggs or larvae. Its estimated value depends on where the samples were taken and the size of the catch. If the samples are not evenly or ran- domly spaced over the study area, then {X, Y) will be biased towards those areas with the highest sample density. This bias can be avoided_by_either grouping the individual samples into strata (sectors) as done here, or by modify- ing the computation of (X, Y) to account for unequal sample density as was done to estimate the mean and variance for the Sette-Ahlstrom method (see Appendix I). The rotation of the ,Y and Y axes is specified by the slope of the principle axis, 6, (Sokal and Rohlf l'.)81): Cov{X,Y) Oi = . A, - Var(.Y) J.N, ■ (X,-X} ■ {Y,-Y) where Cov(X,Y) = — '^ NA -1 Ai = the first latent root of the variance-covariance matrix of A" and Y, = 0.5 ■ [Var(A) + Var(r) + v^(Var(A) + Var(F))- - 4 ■ (Var(A) ■ Var(y) - Cov(;(',y)2)], I.N, ■ (A,-X)2 Var (A') = ' ^ , and 2! A' ^N, ■ (F, -}•)■- Yard") = I.N,\ - 1 I j The ellipse about the centroid is specified by the standard deviations along the rotated axes. The standard deviation in the direction of the major axis is equal to A] as defined above. The standard deviation in the direc- tion of the minor axis is equal to X-,, which is the second latent root of the variance-covariance matrix of A and Y: I, = Var(A) + Var(}') - A,. Sokal and Rohlf (1981, box 15.5) give a formula for an ellipse of this shape but of a different size: Var(F) ■ (A -A)- - 2 Cov(A,y) • (Y-Y) ■ (A -A) + Var(A) ■ {Y-Y)- = C. By setting C to the appropriate value, this formula can define the ellipse as specified above and shown in Appen- dix Figure 1. The value for C is derived from the formula for the point (A, 7) given in Sokal and Rohlf (1981, box 15.5): A = A -h 1 and Y = Y + 6, ■ (A -A). A, -(1 + 6,2) Kendall and Picquelle Egg and larval distributions of Theragra chslcogrsmnw 153 Appendix Figure 1 Ellipse with the center at (A'. Y ) and the major and minor axes each two standard deviations long. The criterion that the leng:th of the major axis is 2 ■ A] implies that A, = (A- -A')- + (Y-Y)- based on the Pythagorean theorem. By substituting in the formulas for A' and Y given above, this formula can be rewritten as C 6r • C C solving for C: C A, = A.,. A2-(l+V) A2-(l + 6i2) A2 Thus the equation for the ellipse is Var(F) • {X-X)- Cov(A,i') • (}'-!') ■ (A -A') + Var(A) • (Y -Y)- = A, Ao, This ellipse is estimated by substituting the estimated values for the variances, covariance, means, and latent roots. No distributional assumptions ^ire needed to estimate the centroid, the slope of the principal axis, or the shape of the ellipse. The centroid is the standard measure of central tendency and is valid as long as the stations repre- sent a valid sample of the survey area. The slope of the principal axis is defined by the line that minimizes the sum of the squared perpendicular distances between the data and the line. The size and shape of the ellipse are set by the variance of the data about the centroid. The ellipse is simply a two-dimensional version of a mean with standard error bars. In the univariate case, the mean and standard error provide information about the location of the data's center of gravity and the amount of dispersion about that center, no matter what probability function the data come from (Shuster 1982). However, if the data are from a normal distribution, a confidence interval may be computed directly from the mean and standard error. If the data are from a bimodal or skewed distribution, the mean and standard error are still valid, but viewing the mean with standard error bars can give the false impression that the data are distributed symmetrically about the mean, when in fact the confidence interval would be asymmetrical if the data are asymmetrical. Similarly, caution should be employed in interpreting the centroid and ellipse. No matter what probability function describes the spatial distribution of the eggs and larvae, the centroid is the measure of central tendency and the ellipse shows the orientation of the data and the amount of dispersion along the two axes. However, the ellipse can only be considered a contour or an "equal frequency ellipse" about the centroid if the data are from 154 Fishery Bulletin 88(1), 1990 a bivariate normal distribution. In tliis case the proportion of the population expected to be enclosed by the ellipse may be estimated as follows. Sokal and Rohlf (1981) state that if Ai ■ A2 • (n-l) ■ 2 C = - - ■ i'„|2,„-2| (n - 2) then the ellipse is expected to enclose 100 • (1 - a)% of the observations. The a that corresponds to the ellipse used in this study can be determined by finding the value for a such that A, -A,- (n-l) • 2 A, • A^ = C = ^ ^;; ^a[2.n-2\< which can be rewritten as o[2.»i-2] (w-2) in - 1) • 2 If /( is very lai-ge, then (n -2)/(n - 1) is close to 1 and ^ o|2.a>| which corresponds to a = 0.6065. Thus the ellipse is expected to enclose approximately 40% of the observations. Appendix 3 The usual stratified estimate of variance is Xa,,;" • YavN, hi YarN,, = A,, (Jessen 1978) which may be rewritten as 1a,,,- ^ • VariV;,,. + Ia,,,,,^ • VariV, h L VkrNi, = ^ where ?'+ refers to sectors with at least 2 samples, and ('n refers to sectors with 1 or 0 samples. There are no data to estimate Var N),, directly, so it is estimated with the average variance from sectors where there are at least two samples. ^A,,J ■ YavNu, Var N|^^^ = — — r^^ , for all i„ Let This estimate of Var N/,,^^ is substituted back into the formula estimating Var N/,, giving I A,,/ • VariV,.^ TAi,,J ■ VariV;,,^ + 1a h,„ VarN,. = l/u. which simplifies to 1A;,, 2 . VariV,,. Var N,, A,2 4,2 1 + Ai,. Abstract. — We investigated in- crement formation in sagittae of At- lantic menhaden Brevoortia. tyrannus in laboratory experiments, and found that the age of individual larvae can be estimated within + 3 days over the first month of life using counts of growth increments. Sagittae were first obsei"ved dmnng embryonic devel- opment. The first prominent growth increment was formed at first feed- ing, and the frequency of increment formation of fed and starved larvae ranged from 0.86 to 0.98 increments per day thereafter. Starvation did not appear to systematically alter the periodicity of increment formation from one increment per day, although it consistently modified the width of growth increments among different age groups of lai^vae. Microstructural growth pattems in sagittae responded rapidly (days) to changes in feeding: larvae starved for 1-3 days formed narrow, poorly-defined increments compared with fed larvae that formed wide, well-defined increments. Stan- dard length and estimated diy weight of larvae were related to sagittal radius by asymptotic and logistic func- tions, respectively. Sagittal radius of larvae was related to days after first feeding by a logistic function. Our re- sults for Atlantic menhaden confirm the potential of otoliths in provid- ing information about age, stressful events, and growth history of in- dividual fish larvae. Effects of Starvation on the Frequency of Formation and WidtPi of Growth Increments \n Sagittae of Laboratory-Reared Atlantic Menhaden Brevoortia tyrannus Larvae Gary L. Maillet Department of Marine, Earth, and Atmospheric Sciences North Carolina State University, Box 8208, Raleigh, North Carolina 27695-8208 Present address Department of Biology, McGill University 1205 Ave Docteur Penfield. Montreal. Quebec, Canada H3A IBl David A/I. Checkley, Jr. Department of Marine, Earth, and Atmospheric Sciences North Carolina State University. Box 8208, Raleigh, North Carolina 27695-8208 Manust'ript accepted 20 September 1989, Fi.shery Bulletin, U.S, 88:15,5-165, Rates of growth and survival of young fish are hypothesized to affect the abundance of the incoming year- class (Lasker 1985, Rothschild 1986). Both biotic (e,g., prey resources) and abiotic (e.g., water temperature) fac- tors have a direct effect on the growth and survival of freshwater and marine fish larvae. Microstruc- tural growth patterns in otoliths of teleost fish may provide a record of environmental and physiological con- dition throughout the larval and juve- nile stages and hence important in- formation about processes regulating recruitment in fish (Pannella 1980, Houde 1987, Rice et al. 1987). The examination of microstruc- tural growth patterns in otoliths for making inferences about the ecology of young fishes has become a popular technique since Pannella (1971) pos- tulated that annuli (yearly growth zones) consisted of growth incre- ments formed on a daily basis. Subse- quently, microstructtiral growth pat- terns have been used to estimate age and growth histories of fish (Methot and Kramer 1979, Penney and Evans 1985), infer the temperattire chronol- ogy of larval and juvenile life stages (Radtke 1984, Gauldie et al. 1986), de- tect life history transitions (Brothers and McFarland 1981, Campana 1984a), and investigate patterns of recruit- ment and mortality (Crecco et al. 1983, Essig and Cole 1986) and stock identification (Mulligan et al. 1987). Age validation studies of lar-val and juvenile fishes have shown that micro- structural characteristics are species- specific and may be influenced by nutrition and/or environmental vari- ables (Campana and Neilson 1985, Rice et al. 1985, Jones 1986). Fish lar- vae subjected to periods of stress (e.g., starvation) or cyclic environ- mental variables (e.g., diel fluctua- tions in water temperature) may have their increment deposition disrupted, resulting in apparent nondaily forma- tion (Taubert and Coble 1977, Jones 1984, Neilson and Geen 1985). These results suggest that validation stud- ies are necessary for a species before analysis of otolith microstructure can be used to age individuals in nature. Substantial errors may be incorpor- ated into the analysis if daily incre- ment formation is assumed but non- daily deposition occurs (Campana and Neilson 1985). 155 156 Fisheiy Bulletin 88(1), 1990 Few studies have assessed the effect of short-term starvation on the accuracy of age estimates and growth histories of the early life stages of fish derived from the analysis of otolith microstructure. This paper ex- amines the reliability of sagittal microstructure of lar- val Atlantic menhaden to estimate age, detect stressful events, and infer the growth chronology of individual larvae. Two age groups of laboratory-reared larvae were subjected to short periods of starvation and optimal feeding conditions to determine the relation- ship between age, growth history, and microstructural growth patterns in sagittae. The Atlantic menhaden is a commercially and eco- logically important species along the Atlantic east coast (Reintjes 1969, Reish et al. 1985). Recruitment of this species has undergone marked fluctuations since moni- toring of the fishery began in the 1940's (Ahrenholz et al. 1987). Investigations of the early life stages of Atlantic menhaden indicate that physical and biological factors operating during larval drift may influence the growth and survival and hence recruitment (Nelson et al. 1977, Checkley et al. 1988). Information derived from these laboratory experiments will serve as a basis for interpreting the microstructure of sagittal otoliths of Atlantic menhaden larvae collected in nature (Maillet 1988). Methods Spawning and rearing conditions Fertilized eggs (henceforth called the "stock popula- tion") were obtained from an induced spawn of a cap- tive stock (Hettler 1981, 1983) and placed into three circular tanks containing 60-L of filtered (20 pim) seawater. Eggs and larvae were incubated at 19°C (18.9 ± 0.1°C; X ± SE) in lightly aerated, static water with overhead fluorescent lighting on a 12 L: 12 D photoperiod. Salinity ranged from 29 to 33 g/kg. Furan II (7 mg/L, Aquarium Pharmaceuticals) was added to retard the growth of bacteria and fungi. During early development (first feeding to 12 days postfertilization), larvae were offered cultured algae Nanochhris spp., rotifers 5rr(c/(/o*i'M.s plicatilis, and wild microzooplank- ton (70-250 ^m size range) ad libitum. During later development, larvae were offered only wild microzoo- plankton. Dead larvae and settled plankton were re- moved every 1-3 days. Water level was maintained by removal of seawater each time food was added. Reference to trade names does not imply endorsement by tl National Marine Fisheries Service. NOAA. Experimental procedures Eggs were sampled daily during development and pre- served in 95% ethanol to investigate otolith formation in embryos. A total of 15 eggs were used in the anal- yses. To estimate the time to first increment forma- tion and to test for subsequent daily increment forma- tion, 10 to 15 larvae were sampled from the stock population at 3, 4, 5, 6, 7, 8, 11, 14, 25, 34, and 35 days postfertilization. The effect of starvation on the period- icity of increment formation was investigated by ex- posing two different age groups of larvae (13-20 days, and 28-36 days postfertilization) to short periods wit h no food. For each group, larvae of various sizes were randomly selected from the stock population and trans- ferred to eight 10-L experimental tanks filled with 20-Mm filtered seawater. Initial densities of larvae in experimental tanks were 5 larvae/L (13-20 days) and 4 larvae/L (28-36 days). Environmental conditions, in- cluding illumination, water temperature, salinity, and the use of Furan II, were identical to those of the stock tanks. Four experimental groups, consisting of (a) con- tinuous feeding (controls), (b) 1-day starved, (c) 2-day starved, and (d) 3-day starved treatments, were ran- domly assigned to duplicate experimental tanks. Con- trol larvae were fed wild microzooplankton ad libitum immediately after transfer; treatment larvae received food at the end of the respective starvation interval and were allowed to continue feeding for several days after this period. All larvae sampled from the stock and experimental containers were first anesthetized with tricaine me- thanesulfonate (Cresent Research Chemical), measured for standard length (tip of upper jaw to end of noto- chord) to the nearest 0.1 mm, and then preserved in 95 % ethanol. Larvae were stored individually in 10-mL vials. The preservative was changed once after 48 hours to maintain sagittae in optimal condition. Otolith preparation and analysis Sagittae were examined within two months after pres- ervation. The right sagitta was teased from the inner ear with minuten needles, cleaned of excess tissue, and mounted medial-side-up in Flo-Texx (Lerner Labora- tories). Specimens were examined with transmitted light under a compound microscope fitted with a lOOx objective and a video camera and monitor, thereby in- creasing the total magnification (monitor image/actual size) to 3600 X. An electronic caliper was used to measure growth increments on the video monitor. This system allowed electronic enhancement of otolith images (e.g., contrast between the incremental and discontinuous zones). Increment counts were made in triplicate on masked (i.e., of unknown origin) samples, to minimize bias, and Maillet and Checkley Effects of starvation on Brevoortia tyrannus 157 the mean calculated. If triplicate counts deviated by ± 3 increments or more, the specimen was excluded from fui'ther analysis. Five percent of the samples were thereby excluded. Another 16% of the specimens were excluded from the analysis because of our inability to resolve increments in some specimens and/or improper orientation for viewing specimens. Linear regressions of mean increment count on days after first feeding were computed for each experimental group. Stu- dent's t was used to determine whether increment for- mation is initiated at first feeding (i.e., regression in- tercept = 0) and growth increments are formed daily (i.e., slope = 1.0). Statistical power to detect a devia- tion of 0.1 from a slope of 1.0 at /) = 0.05 level (two- sided test) was estimated for each linear regression (Rice 1987, Steel and Torrie 1980). Analysis of variance was used to test the homogeneity among all slopes. The width of growth increments formed before, dur- ing, and after starvation was measured for 5-10 lar- vae from each experimental tank. Increment width was measured from the jjerimetei- moving inward along the maximal radius, the inner end of the radius being defined as the center of the nucleus. This line was consistently the best for enumerating increments. To eliminate bias in measurement of increment width the focal plane was adjusted, if necessary, for the measure- ment of each increment. Increment width was aver- aged for each larva over each of the three treatment intervals (i.e., fed, starved, refed), except for the 1-day starved treatment during starvation. Analysis of co- variance (ANOCOVA) was used to compare the mean width of growth increments between larvae from the control and ti'eatment tanks during and after starva- tion (Steel and Torrie 1980). Since the duration of the starvation and recovery intervals varied with length of starvation, the ANOCOVA tested for differences in mean increment width between larvae from control and treatment tanks within intervals. The interval duration was identical between comparisons of mean increment width of larvae from the control and treatment tanks. Comparison of increment width between fed and starved larvae within intervals was necessary because of age-related trends in increment width. The covariate included in the ANOCOVA was mean increment width prior to the starvation and recovery intervals. The ANOCOVA took into account any differences in the width of growth increments of larvae allocated to the experimental tanks and provided a more accurate com- parison between individuals from control and treat- ment groups. The relationships between sagittal radius, standard length, estimated dry weight, and days after first feeding were also investigated. We pooled all of the larvae in the treatment groups (i.e., 1-, 2-, and 3-day starved larvae) for analysis, except that starved larvae were excluded when estimating the dry weight— otolith size relation since these measurements were not made directly on individual larvae. Sagittal radius was mea- sured from the center of the nucleus to the perimeter along the maximal radius. Dry weight of individual lar- vae was estimated from a linear regression of log- transformed dry weight {DW, pig) on log-transformed standard length (SL, mm) : In DW = -3.041 -i- 3.799 In SL , n = 195, r- = 0.95. This equation was derived for laboratory-reared Atlantic menhaden larvae rang- ing in size from 5 to 25 mm (Checkley et al., ms in prep.). Linear and nonlinear regressions of standard length and estimated dry weight on sagittal radius, and sagittal radius on days after first feeding were com- pared to determine the best predictive models. Average growth rates {AVG, mm/day) of larvae for both age groups were estimated from first feeding (ff) to experi- ment termination hy.AVG = (SL - 4.7)/(days after ff), where 4.7 is the average standard length (mm) at first feeding (Powell and Phonlor 1986). Results Increment description and larva growth rate Sagittae were first obser-ved dui'ing embryonic develop- ment and consisted of the dark and apparently pro- teinaceous primordium (Fig. 1). Examination of sagit- tae with the light microscope revealed that no growth increments were formed during this period. At hatch- ing, the sagitta resembled a flattened spheroid or hemi- sphere with a mean radius of 4.8 ± 0.3 fxm, n = 15. The first increment surrounding the primordium was typically characterized by a wide discontinuous zone and coincided with hatching. In some cases, 3-4 nar- row, poorly defined increments were observed outside the first increment. Limited resolution of the light microscope did not always allow these increments to be resolved, counted, and/or measured. The first prom- inent increment in sagittae formed 3-4 days after hatching and coincided with the initiation of exogenous feeding. This prominent growth increment was used as the starting point for subsequent counts along the maximal radius. Mean size at hatching for this laboratory population of Atlantic menhaden was 3.59 + 0.22 mm SL, n = 100 (range 3.08-4.00 mm). The average growth rate of menhaden larvae from first feeding to experiment termination was 0.368 ±0.006 mm/day (13-20 days) and 0.360 + 0.004 mm/day (28-36 days). Size range for the two age groups examined was 6.4-13.0 mm and 10.3-20.0 mm SL. This rate of growth slightly ex- ceeded the estimate of 0.32 mm/day based on data of Powell and Phonlor (1986) for Atlantic menhaden larvae reared at 20°C. Larvae did not metamorphose 158 Fist-iery Bulletin 88(1). 1990 Figure 1 Light micrograph of sagittal otolith of a 22 day-old (postfertilization) laboratory- reared Atlantic menhaden larva showing 18 growth increments. The primordium (p) is delineated hy an innermost protein- rich layer. The nucleus (n) is delineated by the first continuous growth increment surrounding the primordium. The first growth increment formed at hatching (h) is characterized by a thick protein-rich layer, and the first prominent growth in- crement is formed at first feeding (fl"). Scale bar represents 10 nm. in this experiment, but estimates in the literature in- dicate that wild Atlantic menhaden transform to the juvenile life stage at 25-30 mm (Lewis et al. 1972, Nelson et al. 1977). Initiation and frequency of increment formation The time to first increment formation and frequency of increment deposition were estimated from linear regressions of mean increment count on days after first feeding. The regression intercepts of larvae from the stock population and control and treatment groups were not significantly different from zero, except for the 1 -day starved treatment, indicating that the first prominent increment is formed at or near the time of first feeding (Table 1). Frequency of increment formation varied from 0.86 to 0.98 increments/day among experimental groups (Table 1, Fig. 2). There did not appear to be any sys- tematic effect of starvation on frequency of increment formation. The rate of increment formation for both the control and 1-day starved treatments were signif- icantly different from one increment/day, but the stock population, 2-, and 3-day starved treatments formed increments at a rate not sigTiificantly different from one increment per day. Computations of statistical power indicated low variability among the increment count— age regressions and provided additional con- fidence that growth increments are formed daily. The test for homogeneity of slopes indicated a significant difference between larvae from the control and pooled treatments {p<0.05, n = 108). The difference between the estimated slopes for individuals from the contrt)l and 1-day starved treatment tanks accounted for this difference. The estimated pooled slope for the control and 2- and 3-day starved treatments was not signifi- cantly different in the test for homogeneity among slopes (;)>0.fi3, n =68). Response of increment width to changes in larva feeding Since there were no significant differences in increment width between duplicate containers within different levels of treatments (/*> 0.05, ANOVA), measurements Maillet and Checkley: Effects of starvation on Brevoortia tyrannus 159 Table 1 Least-squares regressions of mean increment count on days after first feeding (ff) of laboratory-reared Atlan- | tic menhaden: Increment count = 0-1-6 (days after ff). Asterisks denote significant deviations from the | hypothesized values (H ,: a = 0, 6 = 1.0, Student' s t: *p<0.05, **?;<0.01 •**5D<0.001 . Experimental Intercept Slope 95% group n a±SE 6+SE ^2 CL' Power^ Stock 79 -0.01±0.11 0.98±0.01 0.99 ±1.8 0.99 Control * 0-day starved 27 0.73 + 0.63 0.92 + 0.03 0.97 ±2.4 0.90 Treatments * ** *** 1-day starved 40 2.15±0.39 0.86 ±0.02 0.98 ±2.2 0.99 2-day starved 20 -0.19-1-0.58 0.96 + 0.03 0.98 ±2.4 0.91 3-day starved 22 -0.47±0.74 0.94±0.04 0.97 ±2.6 0.82 (CL) for estimating increment count of an individual larva from time of first feeding. '95% confidence limits -Estimate of statistical power to detect a deviation of 0.1 from a slope of 1.0 at the p = 0.05 level. ZJ o o c CO Coniro 30 20- 10' 0 — I — I — 1 I I I I r 2-day starved 1 — I — I I I I I I r n — I — I — I — I — I — I I r . . .-0 10 20 30 40 3-day starved '/ ^.-V.' Days after first feeding T — I — I — I — I — I — I — I — r 0 10 20 30 40 Days after first feeding Figure 2 Least-squares regression (— ) of incre- ment count on days after first feeding and 95% confidence limits ( — ) for estimating mean increment count of an individual Atlantic menhaden larva. Some symbols represent more than one observation. Statistics are given in Table 1. of increment width of larvae from duplicate control and treatment tanks were pooled. Prior to starvation, no significant differences (/j>0.05) were observed in mean increment width between larvae from control and treatments except for the control and 3-day starved treatment in the older age class. The mean width of growth increments differed significantly (/)<0.05) between fish from control and treatments during the respective periods of starvation and recovery (Fig. 3). The width of growth increments was significantly larger in fed compared with starved larvae. In general, the magnitude and significance of these differences 160 Fishery Bulletin 88(1), 1990 13-20d 28 - 36 d ^a^ 5.0- (a) 28^ \°i 40. 2,2^ + ■ .. 1 ? 3 0 \ 16. \ ' ■ 1 2.0 i ■i ^ 10- 10. E 1 ' ' ' ' 1 „ 2.8 (b) 50 (b) ^ 4 0 ? 2 2 g c 1.6- E f } I 30- 20 * i 2 10- f 1 1.0 c ' ' ' ' 1 — (C) 50 (c) 2.8- 40 2.2- " 30 ^ " 1.6- 1 { * .0 + ^ 1 0 } ^ 1.0 B 1 efore During 1 1 After Before During After Treatment Treatment Figure 3 Mean increment width ol' Atlanlic ineiihailen larvae from control (fed, ■) and treatment (1-3 day starved, O) tanks before, during, and after the respective starvation inter- val, (a) 1-day starved, (h) 2-day starved, and (c) 3-day starved larvae. Vertical bars represent standard errors. increased with longer starvation, and both age classes of larvae responded similiarly to these stressful events. Starved larvae in both age classes displayed narrow [1.4±0.1 fim, n =60, (0.7-2.2 ^^m, range), 13-20 days; 1.8 ±0.3 ^m, n =63, (1.0-3.2 j^m), 28-36 days] poorly defined increments while control larvae exhibited wider [2.1±0.2Mm, n = 30, (1.2-2.8 fim), 13-20 days; 2.7 ±0.6 /im, n =57, (1.3-4.6 j:.im), 28-36 days] well-defined in- crements during the starvation interval. After starva- tion, mean increment width of starved larvae in both age classes increased during the 3-6 day recovery in- terval (Fig. 3). The results of the ANOCOVA, which adjusted for mean increment width prior to the initiation of star- vation, indicated that mean increment width differed significantly (])<0.05) between individuals from control and treatment tanks during starvation in both age classes (Table 2). This result suggests that differences in mean increment width arose during the starvation interval and were not due to differences prior to this interval. However, during the recovery interval, no significant differences (;)>0.05) were observed between larvae from control and treatment tanks when adjusted by the covariate. This result suggests that differences observed between individuals from control and treat- ments during the recovery interval were the result of differences formed during the starvation interval, and indicates that increment width of larval Atlantic men- haden responds rapidly to short-term variations in feeding. Relationship between sagittal size and body size and sagittal size and age Visual inspection of residuals for regressions of stan- dard length and estimated dry weight on sagittal radius, and sagittal radius on days after first feeding, indicated that nonlinear models were superior to linear models in all cases. Asymptotic regressions of standard length on sagittal radius were fit separately for larvae from control and pooled treatments (Fig. 4a). The regressions were highly significant (p<0.0001) and residuals were distributed at random over the entire size range examined, indicating that the models fit the data well. The regression coefficients for fed larvae were slightly larger than for starved larvae. A logistic function was used to regress estimated dry weight on sagittal radius (Fig. 4b). Data were log-transformed to stabilize the variance and reduce the inlluence of larger values on the regression. Starved larvae were excluded from this relationship since dry weights were estimated from standard length. The regression was highly Maillet and Checkley Effects of starvation on Brevoortia tyrannus 161 Table 2 Comparison of least-squares means (adjusted for the covariate in the ANOCOVA) of increment width during 1 and after starvation between control (fed) and treatment (starved) Atlantic menhaden larvae aged (a) 13-20 1 days, and (b) 28 -36 days. N refers to the number of larvae examined, while n corresponds to the total number of observations of increment width from which the mean was calculated. Asterisks denote significant devia- tions between 1 ^rvae from control and treatment tanks; * p<0.05, ** p<0.01, *•* p<0.001. Experimental Mean increment width (/jm) During treatment After treatment group N i±SE(n) X ± SE (n) (a) Control 10 , 1.89 + 0.13(10) 2.08±0.12 (69) 1-day starved 10 1.48±0.13(10) 2.01 + 0.12(70) Control 10 ,, 2.05 + 0.11 (20) 2.00 + 0.10 (59) 2-day starved 10 1.49±0.11 (20) 1.89±0.10 (60) Control 10 ... 2.12 + 0.12(30) 1.94±0.11 (49) 3-day starved 10 1.33±0.12(30) 1.81±0.11 (50) (b) Control 19 . 2.63 ±0.09 (19) 2.97±0.12 (109) 1-day starved 19 2.28 + 0.12(13) 2.73±0.14 (84) Control 19 ... 2.68 + 0.10(38) 2.73±0.14 (90) 2-day starved 18 2.03±0.14 (12) 2.97±0.19(42) Control 19 .. 2.53 ±0.10 (57) 2.94±0.10(71) 3-day starved 111 2.03 ±0.14 (30) 3.32±0.16(3S) (a) 24- fed „ SL = 23,208 - 21 762 e' ° 0^0 (SR) ^_^--- E 20- n = 120, r=0 86 ^^^^...-^-zr^r^ - - ^ E^ ^ i& rfflltwrr ^ D) Ur^jfjagBrp' ~ C li J(|tirt PI _QJ 12- EL^hK starved CO 8- ^P^ SL = 21,668 - 19.480 0'°°^^'^'^' C j,BB n = 107. r^= 0,96 ro ,j^ff W 4- w 0- 1 1 1 1 1 1 1 1 1 1 1 . 1 3 20 40 60 80 100 120 9 Maximal sagittal radius (|im) * ' ted ^ In DW = 9.570 / [1 + 26 0918"'' ^^^ *'" ^"' ] ^^ 8 n = 120, r'= 0 93 *^1^^^^ D) jf^^^ i^t" "*" - ' -kJ^ -w- J^ ^ CD 6 CD ^^ 5 5 4^ >% +^ Q 4 x^* c + Wt 3 < 40- starved SR = 72 552/(1 + 10 6156-" 109 (days tf), 2 1 4_ 1 n = 52, r = 0.91 II;^ \/ " re E 2 ^^J^*--^^ 0- ^-rr 1 1 1 1 1 1 1 1 0 5 10 15 20 25 30 35 \ 40 Days after first feecjing Figure 5 Nonlinear regression of sagittal radius (SR) on days after first feeding (days ./J) of Atlantic menhaden lar- vae from control (solid line, -f ) and treatment (1-3 day starved, broken line, D) tanks. Model coefficients and other descriptive statistics are also given. Some sym- bols represent more than one observation. significant (/)<0.0001) and resitluals were distributed at random over the entire size range examined, in- dicating a good fit of tlie logistic model to the log- transformed data. The relationship between sagittal radius and days after first feeding was also fit with a logistic function (Fig. 5). The regressions were highly significant (<0.001) and residuals were distributed at random, but the variance increased with age of the lar- vae. Predictions of sagittal radius fi-om the regressions indicated that increment width increased from 0.6 \xm. at first feeding to 3.8 ^m at 30 days in larvae from the control tanks, while increment width increased from 0.7 tol.6 \ixvi in larvae from the treatment tanks. Discussion The results of our laboratory experiments indicate that microstructural growth patterns in sagittae of Atlan- tic menhaden can be used to accurately estimate age from first feeding, detect short-term variations in growth rate caused by starvation, and estimate the growth chronologj' of individual larvae. Age of in- dividual larvae from first feeding can be estimated within +3 days during the first month of life (based on inverse regression: Days after first feeding = o -i- h (increment count); Draper and Smith 1966, Rice 1987). Atlantic menhaden larvae appear to initiate in- crement formation at hatching, although growth incre- ments formed prior to first feeding are narrow (< 1 ycca), poorly defined, and could not be consistently resolved with light microscopy. Poorly-defined increments ob- served during yolk feeding in the Atlantic menhaden have been observed during similiar periods in other species (Lough et al. 1982, McGurk 1984, Bolz and Lough 1983). The formation of these increments may be related to an immature circadian rhythm (Campana 1984b) and/or the presence of increments too narrow to detect using light microscopy (Campana et al. 1987). Formation of the first prominent growth increment in sagittae at first feeding may be related to the shift to exogenous feeding and related circadian activity patterns. Transitions within the egg and larval stages of Atlan- tic menhaden were characterized by particular micro- structural features. Hatching was characterized by a wide discontinuous zone, and the transition to exog- enous feeding coincided with a prominent growth in- crement. Similiar patterns have been observed in other species (Brothers and McFarland 1981. Campana 1983, Lagardere and Chaumillon 1988). Particular micro- structural features formed during the early life stages may be the result of changes in physiological metab- olism, stress, and/or growth cycles that are generally associated with these transitions. Studies that have examined the rate of increment for- mation in fish larvae reared under conditions pro- moting rapid growth have shown that growth incre- ments form daily in most cases ( Jones 1986). Atlantic menhaden larvae fed ad libitum formed increments at a rate of unity, consistent with the hypothesis that in- crements are formed daily in sagittae of well-nourished fish larvae. Estimates of statistical power obtained in our study indicate low variability about the relation- ship between increment count and days after first feeding and provide additional support that growth Maillet and Checkley Effects of stawation on Brevoortia tyrannus 163 increments are formed daily in sagittae of Atlantic menhaden larvae. A closely related species, the Gulf menhaden 5. patronus, initiated increment deposition at first feeding and formed an average of one growth increment per day in sagittae (Warlen 1988). The fre- quency of increment formation in juvenile Atlantic menhaden held in enclosures also indicated that sagit- tal growth increments are formed daily (Simoneaux and Warlen 1987). Starvation of larval Atlantic menhaden for 1-3 days did not result in the cessation of sagittal growth nor systematically alter the periodicity of increment for- mation from one increment per day. Larvae starved for 2- and 3-days formed increments at rates similiar to larvae which were fed continuously. Similiar results have been reported for other species reared under various feeding regimes, including starvation (Marshall and Parker 1982, Campana 1983, Eckmann and Rey 1987). However, other studies investigating formation of otoliths have found that periods of starvation may affect the rate of increment formation (Townsend and Graham 1981, Geffen 1982, McGurk 1984, 1987). Growth increments formed during starvation and periods of slow growth may be difficult to resolve due to their small size , generally < 1 ^^m, which is near the limit of resolution of most light microscopes. This may have resulted in underestimation of the number of in- crements in two of the experimental groups of Atlan- tic menhaden larvae in our study and contributed to the apparent nondaily deposition rate. Theoretical daily increment width of larval Atlantic herring during early growth (first 2 weeks) was predicted to be below the limit of resolution of light microscopy (Campana et al. 1987). Examination of otoliths of fish exposed to sub- optimal conditions indicated that age was underesti- mated using light microscopy compared with high reso- lution viewing by SEM (Jones and Brothers 1987, Bailey and Stehr 1988). Microstructural growth patterns observed in sagittae of larval Atlantic menhaden support the hypothesis that variations in nutrition and environmental factors affecting fish growth are manifest as variations in in- crement width (Methot and Kramer 1979, Neilson and Geen 1982, 1985, Radtke 1987). The response of incre- ment width to short intervals of starvation indicate that sagittae of Atlantic menhaden larvae may provide a reliable record of short-term (e.g., days) changes in feeding. The response time to starvation of Atlantic menhaden larvae examined in our study was rapid. In- crement width of larvae starved for only 1 day declined significantly compared with fed larvae. The response time of increment width to cessation of starvation varied with the age of larvae. Mean increment width of larvae in the young age class increased to that of the controls within 3-4 days after starvation. Mean increment width of larvae in the older age class in- creased less rapidly after starvation than did that of the younger age class and did not reach that of con- trols during the recovery interval (4-6 days). Our results indicate that microstructural growth in- crements in sagittae of Atlantic menhaden can be used to infer the dynamics of larval growth. In particular, examination of sagittal microstructure can be used to estimate the age of individual larvae from first feeding, detect stressful periods, and reconstruct the growth rate chronology of Atlantic menhaden larvae. We are currently investigating short-term (e.g., days) varia- tions in growth rate of sea-caught Atlantic menhaden larvae, inferred from analysis of sagittal microstruc- ture, in relation to meteorological and oceanographic variables (Checkley et al. 1988, Maillet 1988). Measure- ments of individual growth increments, although tedi- ous to obtain, may allow inference about important events during the early life stages of fish. Variations in the width of growth increments from otoliths of fish larvae collected in nature could be analyzed for corre- lations with the timing of particular developmental and environmental events to determine their relative importance. Acknowledgments We would like to thank the Beaufort Laboratory of the Southeast Fisheries Center, National Marine Fisheries Service, for use of their larval fish rearing facilities. Assistance in laboratory experiments with Atlantic menhaden was given by W.F. Hettler and A.B. Powell who provided fertilized eggs of Atlantic menhaden and helpful advice on rearing larvae in the laboratory, and C. Lewis who supplied algae and rotifers for feeding young larvae. K.M. Mason provided expertise in the laboratory and constructive criticism of the manuscript. Drs. S. Warlen, S.E. Campana , and J.D. Neilson provided helpful advice in otolith preparation techniques. We thank Dr. D.L. Kamykowski for use of the camera and video equipment for viewing otoliths. Dr. D.A. Dickey provided assistance with statistics. Drs. L.B. Crowder and D.L. Kamykowski made helpful and constructive comments on the manuscript. This research was sup- ported by a grant from the National Science Founda- tion to D.M. Checkley, Jr. (OCE-85-16799). 164 Fishery Bulletin 88(1)^ 1990 Citations Ahrenholz, D.W., W.R. Nelson, and S.P. Epperly 1987 Population and fishery characteristics of Atlantic men- haden. Brevoiirtia tyrannus. Fish. Bull.. U.S. 85:569-600. Bailey, K.M.. and C.L. Stehr 1988 The effects of feeding periodicity and ration on the rate of increment formation in otoliths of larval walleye pollock Theragra ckalcogramma (Pallas). J. Exp. Mar. Biol. Ecol. 122:147-161. Bolz, G.R., and R.G. Lough 1983 Larval cod {Gadus morhua) and haddock (Melamigram- mus aeglefimis) growth on the Georges Bank, spring 1981. Fish. Bull., U.S. 81:827-836. Brothers, E.B., and W.N. McFarland 1981 Correlation between otolith microstructure, growth and life transitions in newly recruited French grunts {Haemulon Jlavolineatum). 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 the otoliths of steelhead trout (Salrrw gairdwri) and starry flounder (Platichthya stellatus). Can. J. Zool. 62: 1507-1512. 1984a Microstructural growth patterns in the otoliths of larval and juvenile starry flounder. Plutichthy^ stellatus. Can. J. Zool. 62:1507-1512. 1984b Interactive effects of age and environmental modifiers on the production of daily growth increments in the otoliths of plainfin midshipman. Pnrirhthys notatm. Fish. Bull., U.S. 82:165-177. Campana, S.E., and J.D. Neilson 1985 Microsti-ucture of fish otoliths. Can. J. Fish. Aquat. Sci. 42:1014-1032. Campana, S.E., J. A. Gagne, and J. Munro 1987 Otolith microstructure of larval herring (Clupea haren- gii.^:}: Image or reality? Can. .J. Fish. Aquat. Sci. 44:1922-1929. Checkley, D.M., Jr., S. Raman, G.L. Maillet, and K.M. Mason 1988 Winter storm effects on the spawning and larval drift of a pelagic fish. Nature (Lond.) 335:346-348. Crecco, V., T. SAvoy, and L. Gunn 1983 Daily mortality rates of larval and juvenile American shad (Alosa sapidissima) in the Connecticut River with changes in year-class strength. Can. J. Fish. Aquat. Sci. 40:1719-1728. Draper. N.R., and H. Smith 1966 .ViJiilied regression analysis. John Wiley and Sons, NY. 709 p. Eckmann, R., and P. Rev 1987 Daily growth increments on the otoliths of larval and juvenile Coregonus spp., and their modification by environmen- tal factors. Hydrobiologia 148:137-143. Essig, R.J., and C.F. Cole 1986 Methods of estimating larval fish mortality from daily growth increments in otoliths. Trans. Am. Fish. Soc. 115: 34-40. Gauldie, R.W., D.A. Fournier, D.E. Dunlop, and G. Coote 1986 Atomic emission and proton microprobe studies of the ion content of otoliths of chinook salmon aimed at recovering the temperature life history of individuals. Comp. Biochem. Physiol. 84A:607-615. Geffen, A.J. 1982 Otolith ring deposition in relation to growth rate in larval herring (Clupeti hnrerigvs) and turbot (Sniphthdlmu.'i inaxunus) larvae. Mar. Biol. 71:317-.326. Hettler, W.F. 1981 Spawning and rearing of Atlantic menhaden. Prog. Fish-Cult. 43:80-84. 1983 Transporting adult and larval (Uilf menhaden and tech- niques for spawning in the laboratory. Prog. Fish-Cult. 45: 4.5-48. Houde, E.D. 1987 Fish early life dynamics and recruitment variability. Am. Fish. Soc. Symp. 2, p. 17-29. Jones, CM. 1984 The otolith increment technique: ajiplication in larval i1sh. Ph.D. thesis, Univ. Rhode Island, Kingston, 123 p. 1986 Determining age of larval fish with the otolith increment technique. Fish. Bull., U.S. 84:91-103. Jones, CM., and E.B. Brothers 1987 Validation of the otolith increment ageing technique for stri|je(l bass, Morone saj^ctilis, larvae reared under suboptimal feeding conditions. Fi.sh. Bull., U.S. 85:171-178. Lagardere, F., and G. Chaumillun 1988 F^ffects of delayed initial feeding on the growth of sole. Solea rulgdris. larvae and on the incremental pattern of their otoliths. C.R. Acad. .Sci. -Ser. Ill Sci. Vie, t. 306, p. 601-607. Lasker, R. 1985 What limits clupeoid production? (.'an. J. F'ish. Aiiuat. Sci. 42(Suppl. 1):31-3,S. Lewis, R.M., P.H. Wilkens, and H.R. Gordy 1972 A description of young Atlantic menhaden. Brevanrtia tynriinus, in the White Oak River Estuary, North Carolina. Fish. Bull., U.S. 70:115-119, Lough, R.G., M. Pennington, G.R. Bolz, and A. Rosenberg 1982 Age and growth of larval Atlantic herring, Clupea hareugus L., in the Gulf of Maine-Georges Bank region based on otolith growth increments. Fish. Bull., U.S. 80:187-200. Maillet, G.L. 1988 Otolith increment growth patterns of Atlantic menhaden Bremortm tyraniius larvae and their relation to meteorological an95% during most of the period studied. The proportion of shortfin squid (Illex argen.tmus) in the catches increased during May-July as a result of seasonal overlap in the distributions of both species. Even then, the Argentine fleet did not generally target on squid, which could be considered as bycatch (Juanico 1982). The use of fishery data to investigate migration relies on the fishermen's ability to locate the densest aggrega- tions of hake. Several lines of evidence support this assumption. The hake fishery has been taking place for several years and the displacements of fish are well known. Monthly histograms of standardized hake CPUE show a moderately positive skewness (data not shown) that may be partly associated with a widespread knowledge of fish distribution (Quinn 1985). A com- parison of alternative formulations for a monthly CPUE index, i.e., an average-of-ratios vs. a ratio-of- averages, confirms that effort was concentrated on areas of higher-than-average hake abundance (Otero 1986). This conclusion is also supported by spatial statistics of fishing effort and CPUE (Rothschild and Yong 1970). To visualize better the patterns of hake movement, monthly latitudinal and longitudinal weighted averages of fishing effort and hake CPUE were computed ac- cording to the procedure described by Rothschild and Yong (1970). In some months, fishing effort was deployed in distinctly separate regions throughout the shelf. This may have reflected true operational dif- ferences in the fishing fleet as a consequence of separ- ate centers of fish abundance. To take into account the discontinuity, I grouped all contiguous 1-degree areas, calculated the total fishing effort for each cluster of contiguous areas, and considered only major clusters (those in which effort was >5% of the effort in the most heavily exploited cluster). Most months showed only one valid cluster of areas and, at most, two clusters fulfilled the previous criteria. I computed the spatial statistics of effort for each valid cluster. For the CPUE statistics, the same clusters delimited above were used. (T) Boenos Atres Slope Norlti ^} Buenos Aires Slope Soulh (?) Rio de 10 Ploto Moutli (4) Pologonian Slope Norlh (5) Pologonian Slope Soulh (b) Buenos Aires Coost (7) Intermediole Shelf (e) Southern Spawning Grounds 65* 55*W Figure 1 Geographic regions in southwL'stern Atlantic Ocean used for inves- tigating interannual differences in hake migration. Stippled regions are discussed in the text. but data were included only from areas where five or more trips had been reported, to reduce variability in CPUE values due to low levels of sampling. As with most hakes (Jones 1974), tagging Argentine hake has proven so far to be impossible and, conse- quently, there are no estimates of migratory speed derived from tag returns of individual fish. As an ap- proximation, I estimated hake migration speed by using the spatial statistics of fishing effort calculated above. I calculated the distances between the centers of ef- fort in consecutive months, and then estimated daily migratory speeds by dividing the distances by 30 days (regardless of the specific months involved). If the migratory pattern of hake is associated with environmental factors, year-to-year differences in their distribution may reflect fluctuations in oceanic condi- tions. I examined differences in the distributional pat- tern of hake between 1978 and 1979, and tried to relate those differences to environmental conditions. A direct comparison of the commercial fishery effort and CPUE data is impossible because (a) there was a change in the format of the fishing tickets (the forms turned in by fishermen upon return to port) between 1978 and 1979, and Oi) fishing effort had to be stan- dardized separately for each year because of software limitations (see Podesta 1987). A new fishing ticket was introduced in 1979 which, due to its design, increased uncertainty on the reported location of effort. To reduce the possibility of error, I grouped the data, originally reported for 1-degree squares, into eight larger geographical regions (Fig. 1). Because fishing effort had been standardized separ- ately for 1978 and 1979, the cumulative time density approach introduced by Mundy (1982) was used, which Podest.3: Migratory pattern of Merluccius hubbsi 169 - 9- 45 S Figure 2 Schematic description of the annual migratory pattern of Argentine hake in southwestern Atlantic Ocean. Approximate locations of some oceanic features mentioned in the text are also shown. Figure 3 [A) Temperature section (°C) along the northern Argentine shelf break, July 1978. (B) Log of Argentine hake CPUE at fishing sta- tions occupied at the same latitude as the hydrographic stations, but slightly shoreward. Date from the RV Wnlthcr Henvig. does not require between-year consistency in the mea- sure of effort. The technique is based on the fluctua- tions in fish abundance as a function of time in a fixed geographic reference frame and originally employed catch or CPUE data (Mundy 1982). However, the regions considered showed no fishing effort during some months and CPUE was undefined; the zero ef- fort value, in contrast, was meaningful. Therefore, I used effort values (hours trawled per month) to con- struct the cumulative time density series in the man- ner described by Mundy (1982). Results Migratory pattern of hai4°C, except in very shallow regions. This may have been due to the development of a shallow thermocline (~10 m) in inshore waters following a few days of calm weather. The location of the front is in- dicated instead by the transition from a shallow to a fleeper (~35 m) thermocline (data not shown) which occurred in the vicinity of water column temperature differences of 7-8°C, as depicted in Figure 4. Highest hake CPUEs were observed near this transition; in con- trast, hake CPUEs inshore of the front were very low. The last portion of the migration, which closes the circuit, is the northward movement. This stage begins in February and lasts through May-June, when hake reach the wintering grounds. Fishing-effort maps sug- gest that the fish move north and across the shelf up to 40° S during February, and then continue north- wards along the shelf break. Spatial statistics The spatial statistics are intended to represent "centers of mass" of the distributions of fishing effort or hake relative abundance during each month. Due to the un- certainty in effort location reports in 1979, the centers of fishing effort and hake CPUE are shown only for 1978 (Fig. 5). The average positions of CPUE and effort showed cyclical north-south and east-west movements during a year. The monthly centers of CPUE and effort were quite similar, implying coincidence in the spatial dis- tributions of fishing effort and apparent hake abun- dance. This suggests that the fleet was successful in locating dense aggregations of hake and lends con- fidence to the use of fishing-effort data to follow the commercial densities of migrating hake. The multiple centers of effort in December, January, and February may be related to the existence of sepa- rate spawning groups of hake. On the other hand, the two centroids may merely reflect operational tactics of the fleet; although effort has been standardized to take into account differences in catching power, smaller vessels may lack the endurance or hold capacity re- quired for a trip to the southern spawning grounds. Alternatively, fishermen may perceive that the catch rates in the northern and southern grounds are not so different as to make the longer trip worthwhile. Speed of hake migration Distances between monthly centers of effort and esti- mated speeds of migration are shown in Table 1 . There was a bimodality in estimated displacement speeds throughout the year; the two modes corresponded. Podest^: Migratory pattern of Merluccius hubbsi 71 8 ' 6 2 S' 12 66* 64- 62- 60- 58° 56° 54'W 34* 36' 38° 40 42S 12 10 11 2 12 1 66' 64* 62* 60- 58" 56- 5?V\r se- as- 42 S Figure 5 Monthly weighted latitude and longitude means of (a) fishing effort and Oj) Argentine hake CPUE in southwestern Atlantic Ocean, com- puted from 1978 commercial fishery statistics. Numbers indicate months. Table I Distance between consecutive monthly centers of fishing effort | in southwestern Atlantic Ocean du •ing 1978 and Argentine hake swimming speeds required to cover those distances in 30 days. When two centers existed n a month, they were in- dicated as northerr (N) and southei n (S). Segments marked | with an asterisk are considered to be inconsistent with the migratory pattern of hake. Speed required Distance Segment (km) m/sec km/day Jan (N)- Feb (N) 128 0.05 4.3 (*) Jan(N)-Feb(S) 297 0.11 9.9 (*) Jan(S)-Feb(N) 636 0.25 21.2 Jan(S)-Feb(S) 244 0.09 8.1 Feb(N)-Mar 144 0.06 4.8 Feb(S)-Mar 276 0.11 9.2 Mar-Apr 131 0.05 4.4 Apr-May 99 0.04 3.3 May-June 122 0.05 4.1 June-July 48 0.02 1.6 July-Aug 37 0.01 1.2 Aug-Sep 351 0.13 11.7 Sep-Oct 285 0.11 9.5 Oct-Nov 49 0.02 1.6 Nov-Dec(N) 188 0.07 6.3 Nov-Dec(S) 264 0.10 8.8 latter value was only slightly higher than that com- puted for the active northward migration, 9.2 km/day (February-March). If a significant difference in south- ward vs. northward migrating speed exists, it probably cannot be detected using coarse fishery statistics. The migration speeds estimated above must be con- sidered only as a first approximation, but they general- ly agree with values reported for a similar species. Francis (1983) estimated swimming speeds for Pacific hake Merluccius productus of different ages based on timing of sequential appearance at different latitudes. His estimated speeds for the age classes 4-7, those predominant in Argentine hake catches, range from 9 to 12 km/day. Ermakov (1974; cited in Bailey et al. 1982) concluded from direct observation of a school that the northward migration of Pacific hake is at speeds of 5-11 km/day. respectively, to speeds of 6-11 km/day and 1-4 km/day (Table 1). Higher speeds, obviously, are associated with periods of active migration. The estimated speeds of the northward and south- ward (i.e., feeding and spawning) migrations were sim- ilar. Estimated speeds during the most active periods of southward displacement were 11.7 km/day (August- September) and 9.5 km/day (September-October). The Interannual variability in the migratory pattern of hake Cumulative time density of fishing effort is shown in Figin-e 6 for three regions: Buenos Aires slope north (region 1), Patagonian slope north (region 4), and south- ern spawning grounds (region 8). The values shown in the figures for each month represent the proportion of yearly regional effort accumulated up to the end of 172 Fishery Bulletin 1990 that month. Monthly values are joined by line segments in the figures: segments with steep slopes are associ- ated with a relatively large proportion of total regional effort. Conversely, if little or no effort took place within a month, the corresponding segment will be nearly horizontal. The interpretation of the cumulative time density curves is related to the pattern of hake migration. For example, effort started to accumulate in region 1 (Fig. 6a) around April or May, when fish first arrived in this area on the northward leg of their migration. The S- shaped 1978 curve indicated that there was a single peak in the deployment of effort during winter months (May-August). In contrast, in 1979 most of the effort was deployed between March and June, with a second- ary peak in September-October. In region 4 (Fig. 6b), effort accumulated rapidly dur- ing February, March, and April of 1978, as the hake were migrating northwards along the shelf break. Afterwards, effort decreased greatly. In 1979, the ini- tial effort was much like the previous year's, but it tapered off until June, July, and, particularly, August (winter). Effort in winter months was not observed in 1978. In region 8 (Fig. 6c), the effort patterns were re- markably similar for both years. The curves were flat between March and September, as both the fish and the fleet were absent from the region. During October, November, and December the spawners arrived in the area and fishing effort accumulated rapidly. Interannual differences between cumulative time density patterns were statistically analyzed using the Kolmogorov-Smirnov (K-S) two-sample test. Because of the large N values (the annual total number of hours trawled in a region), all the K-S tests were significant (see Rugolo 1984), including those for regions not shown. As a consequence, the discussion will focus on regions 1 and 4, where the patterns appeared to differ most between the two years. From the patterns in regions 1 and 4, it appears that the fishing fleet (and, thus, hake) was located further south in 1979 than in 1978. The spatial statistics of fishing effort also reflect a more southerly deployment of effort in 1979. The latitudinal means of fishing effort for July and August 1978 were, respectively, 36.4°S and 36.6° S; the corresponding values for 1979 were 38.2°S and 40.2°S. That is, the "center of mass" of fishing effort in the winter of 1979 was located 2-4 degrees of latitude further south along the shelf break than in 1978. The changes in hake distribution may reflect interannual differences in environmental conditions. The northernmost extension of the hake feeding migration, reached during June- August, is associated with the Brazil-Malvinas confluence. The confluence is C 80 o o Q. 60 O > TO 3 E D o Buenos Aires Slope North C 80- O 8. 60 O 0) > — 20 E D O 0 c o g_60i o qI > 3 5 o 2 i A b6789 Month Patagonian Slope North 1979 23456789 Month Southern / Spawning Grounds ^1978_^ // 1979 C 234 56769 Month Figure 6 ( 'uinulative time densitifs for (.1 ) Bueno.s Aires sliipe iKjrth (region 1), (/?) Patagonian slope north (region 4) and ((') southern spawning grounds (region 8). Location of regions is shown in f^igure 1. an effective barrier for adult hake of commercial size: aggregations of hake found north of it are mainly formed by juveniles (Rojo and Silvosa 1969). Olson etal. (1988) showed considerable fluctuations in the Podest^ Migratory pattern of Merluccius hubbsi 173 latitude of the Brazil-Malvinas confluence along the shelf break. Shifts in the position of the confluence plausibly could have been responsible for the differ- ences in hake distribution between 1978 and 1979. Hydrographic stations occupied by the RV Walther Herwig during July 1978 placed the Brazil-Malvinas confluence along the shelf break at about 36° S. In August 1978, the confluence must have been located north of 37 °S, as temperatures characteristic of Brazil Current waters -were not observed during a cruise car- ried out by the RV Sh inkn i Ma.ru that reached approx- imately that latitude (data not shown). No hydrographic data were available to locate the confluence in July- August 1979. Instead, I used weekly GOSSTCOMP (global ocean sea surface temperature computation) charts for the period 3 July- 28 August 1979. The charts were produced from satellite data by the U.S. National Oceanic and Atmospheric Adminis- tration (Brower et al. 1976). The 10°C and 12°C surface isotherms were considered to represent the Brazil- Malvinas confluence at this time of the year. The ap- proximate latitudes at which those isotherms occurred along the shelf break are displayed in Figin-e 7. In early July 1979, the confluence was located at about 37°S, only slightly further south than in July 1978. During the last 2 weeks of July 1979, however, it moved to the south, reaching about 42 °S by the end of the month. During August 1979, the confluence was generally located around 40°S. The deployment of fishing effort during July- August 1979 coincided roughly with the confluence locations deduced from the GOSSTCOMP charts. This seems to support the hypoth- esis linking shifts in the position of the Brazil-Malvinas confluence and hake distribution in 1978 and 1979. Discussion Feeding migration The need to find an adequate food supply for main- tenance, growth, and reproduction may have contrib- uted to shape the migi'atory pattern of Argentine hake. Adult hake feed actively after spawning (Hart 1946). The postspawning offshore migi-ation of adults towards the shelf break, then, may correspond to a search for abundant food sources. The western edge of the Malvinas Current, from about 40° S to its confluence with the Brazil Current, plays an important role in the northward migration of hake, as the fish are closely associated with it for 5-6 months of the year. The boundary between shelf and Malvinas Current waters can be characterized as a shelfbreak front. Shelfl^reak fronts elsewhere have been associated with high biological productivity and Ashing intensity (Fournier et al. 1979). ? 35°S- < -A O ^ 40- . < LlI O ^ K / /^ NE 3 10 17 24 31 7 14 21 28 SW -A' 1 JULY 1 AUGUST Figure 7 Approximate latitude of the Brazil and Malvinas currents (.'(jniluence along the shelf break, July- August 1979, derived from GOSTCOMP charts. Each vertical line corresponds to a weekly sea-surface tem- perature chart. Shaded area represents projection of the 10- and 12°C isotherms along the shelf break, assumed to represent the position of the confluence. The high productivity of the Malvinas front has been described from ship data (Hubold 1980a,b and refer- ences therein). More recently, remote sensing tech- niques have confirmed this feature: images collected by the Coastal Zone Color Scanner (CZCS) were userl to monitor nearsurface concentrations of phytoplank- ton pigments and, indirectly, phytoplankton liiomass. High phytoplankton concentrations occurred along the shelfbreak through spring and summer; in contrast, phytoplankton biomass across most of the shelf de- creased rapidly folk)wing the late September-early October spring bloom. The shelfbreak, therefore, pos- sibly represents one of the few places where forage is abundant during the summer, following the hake's spawning. After a second bloom ends on the shelf in the fall (late March-early April), phytoplankton biomass along the shelfbreak remains high into late April. This may in turn support elevated zooplankton concentrations in subsequent winter months (Hubold 1980b), although at that time phytoplankton biomass is low. The large phytoplankton biomass along the shelf break is probably the result of enhanced supply of nutrient-rich MaKnnas Current waters into the euphotic zone, which could happen through a variety of pro- cesses. Small-scale eddies along the edge of the Mal- vinas Current may upwell nutrient-rich waters. This has been shown to occur along the edge of the Gulf Stream, resulting in an increase in phytoplankton and zooplankton biomasses (Paffenhoffer et al. 1984). Dick- son et al. (1980) suggested that the interaction between coastally trapped waves propagating along the slope and bottom topography at the shelfbreak could enhance upwelling. The generation of internal tides at the shelfbreak, coupled with episodic wind stress, may 174 Fishery Bulletin 88(1), 1990 increase vertical mixing, again injecting nutrients into the upper layers (Maze et al. 1986). Additionally, the tilt of the isopycnals or the interleaving of water masses at the front would enhance vertical stability, retaining phytoplankton cells in the euphotic zone (Fournier et al. 1979). The increased phytoplankton bioniass along the shelf break supports large concentrations of zooplankton (Hubold 1980a); species on which i^elagnc fish feed (e.g., calanoid copepods) are very abundant (Carreto et al. 1981a, b). Species from higher trophic levels also aggre- gate at the front. The distribution pattern of anchoita Eugriiulis an- choita. an extremely important prey for hake, is close- ly associated with the shelflireak front (Brandhorst et al. 1971). While hake and anchoita coincidentally migrate offshore and then northward along the slope in autumn, the former actively feed on the latter (Ciechomski and Sanchez 1983). During the northward migration, other components may appear in the hake's diet, such as myctophids associated with the Malvinas Current (e.g., Gynuiosco- pelus spp., Lanipnnijctuii spp., and Myctophum spp.) and shortfin squid Illex argerifiin^s (Cordo 1981). Dense concentrations of shortfin squid occur along the shelf- break in late fall (Otero et al. 1981). Squid is increas- ingly found in the stomachs of medium and large hakes between April and July (Angelescu and Cousseau 1969, Cordo 1981). Given their high food assimilation efficien- cy (Caddy 1983), squid probably play an important eco- logical role in rapidly transferring energy from lower trophic levels to hake. Wlien hake reach the northernmost end of their feed- ing migration in winter, they liave access to a large abundance of prey from a variety of closely located en- vironments. In addition to the species found along the shelllireak, they may feed on anchoita wintering in mid- shelf waters or on subtropical myctophids from Brazil Current waters. Further south or inshore, in contrast, pelagic fish become rare and the overall abundance of prey decreases rapidly (Angelescu and Cousseau 1969). Spawning migration The present migratory pattern of hake may have evolved to maximize chances of reproductive, as well as feeding, success. Starvation is a major source of lar- val mortality for most fish species. The choice of spawn- ing sites with adequate concentrations and size ranges of larval food may have a great influence on survival. Hake larvae feed almost exclusively on zooplankton, particularly small calanoid copepods (Ciechomski and Weiss 1974). Regional shifts in hake spawning activity may be related to the seasonal pattern of productivity along the shelf. The phytoplankton spring bloom begins in the north and progresses southward (Carreto et al. 1981a). The spawning activity between 36 °S and 38 °S in early spring, then, may be tied to increased zoo- plankton availability resulting from the bloom. How- ever, satellite imagery shows that postbloom ])hyto- plankton biomass here remains low throughciut late spring and summer and apparently does not support an extended spawning season. Later in the season, the bulk of the spawning activity takes place between 42° S and 44°S, where production possibly may continue dur- ing the summer because of the presence of tidal fronts. The tidal fronts appear to be associated with en- hanced phytoplankton biomass: Carreto et al. (1981a) observed integrated chlorophyll concentrations >200 mg/m- at the beginning of October. The increased biomass may be due to enhanced nutrient supply by horizontal mixing from the mixed side or by vertical mixing through the pycnocline. Carreto et al. (1986) found high surface concentrations of nitrates (>6 pmi) on the mixed side. The existence of the tidal fronts has only recently been noticed and, consequently, there is relatively little information on their biological productivity. Never- theless, high concentrations of zooplankton (express- ed in dry weight) have been reported along the fronts (Angelescu and Prenski 1987). There is additional in- direct evidence for the relatively high productivity of these features: hake larvae in this area are found to be in an excellent nutritional state, better than elsewhere on the shelf (R. Sanchez, INIDEP, C.C. 175, Playa Grande, Mar del Plata, Argentina, pers. com- mun.. Feb. 1989). Additionally, adult hake find abun- dant anchoita and squid in the vicinity of the fronts (Angelescu and Prenski 1987). The tidal fronts occur throughout the summer, for as long as the offshore side of the front remains strati- fied. Satellite imagery shows that this area, together with the shelfbreak, are the only places where near- surface phytoplankton biomass remains high through- out the summer. The integrated spring-summer pro- duction of the tidal fronts area, therefore, may possibly be higher than in most parts of the shelf, although this requires confirmation. Another factor relevant to the selection of a spawn- ing site is the subsequent transport of eggs and larvae to (or, alternatively, their retention within) areas suit- able for further development. Argentine hake show a wide temporal and spatial spawning range: hake eggs are found on the shelf practically year-round. Never- theless, the gonads do not mature more than once dur- ing the main spawning period of spring-summer (Chris- tiansen and Cousseau 1971), suggesting that a number of groups of spawners release eggs throughout a large area and an extended time interval. Argentine hake Podest^: Migratory pattern of Merluccius hubbsi 175 exhibit a ubiquitous spawning strategy throughout the shelf and, therefore, do not appear to be dependent on larval drift and retention conditions. A possible excep- tion is the shelfbreak area: despite its importance in supporting abundant forage supplies, this area is not a suitable spawning location. Eggs and larvae spawned along the shelfbreak would most likely be carried offshore. The circulation pattern of the Argentine shelf, as it contributes to larval drift or retention, has not been described in adequate detail. Circulation gyres or topo- graphical features that may contribute to the separa- tion of hake spawning groups on the Argentine shelf are not apparent. If different spawning groups exist, the specificity of their place and time of spawning or, equivalently, the degree of mixing among them, re- mains unclear. This work identified apparent differences in the migratory pattern of hake between 1978 and 1979, which were tied to the location of the confluence of the Brazil and Malvinas currents along the shelllireak. Sosa et al. (1976) mention 60-100 mile changes in Argen- tine hake distribution associated with wind-induced short-term fluctuations in the position of the conflu- ence. Repeated monitoring of the confluence location through satellite imagery (cf. Olson et al. 1988) would make it possible to further investigate this association. In summary, the cyclic migration of commercial den- sities of Argentine hake was roughly described using data from commercial fishing operations. The migra- tory pattern has evolved to maximize the chances of successful spawning, while ensuring an adequate food supply. During their migi-atory cycle, hake take advan- tage of productive ocean features. A number of issues remain unclear, such as the ex- istence of different spawning groups and the degree of mixing among them. There is little information on the recruitment variability in Argentine hake, so one cannot speculate on the relationship between migi'atory pattern and recruitment success. The circulation on the Argentine shelf and its influence on larval drift or retention requires much study. Finally, the continued availability of prey for hake larvae throughout the sum- mer in the tidal-fronts environment (which would lead to the bulk of spawning taking place there) needs to be confirmed. Acknowledgments Comments by F. Williams, N. Ehrhardt, D. Olson, J. Powers, and three anonymous reviewers are ap- preciated. Support was provided by the Office of Naval Research and by a National Research Council Research Associateship at NASA's Goddard Space Flight Center. Data from the Argentine fishing fleet were supplied by Subsecretan'a de Pesca, Buenos Aires, Argentina. The Shinkai Maru and Walther Herwig data were provided by Institute Nacional de Investiga- cion y Desarrollo Pesquero, Mar del Plata, Argentina. Travel to compile the data was funded by the Tinker Center of the University of Miami. Citations Angelescu, V., F.S. Gneri, and A. Nani 19.58 La merluza (k'l Mar Argentine (biologia y taxonomi'a). Publicacion Servieio Hidrografi'a Naval H1004. Buenos Aires, 224 p. (in Spanish]. Angelescu, V., and M.B. Cousseau 1969 Alimentacii)!! de la merluza en la region del talud con- tinental argentine, epoca invernal (Merlucciidae, Merluccius merluccius huhhsi). Bol. Inst. Biol. Mar. Mar del Plata 19, 84 p. [in Spanish]. Angelescu, V., and L.B. Prenski 1987 Ecologia tnifica de la merluza comun del Mar Argentine (Merlucciidae, Merluccius merluccius hubbsi). Parte 2. Dina- mica de la alinientacion analizada sobre la base de las condi- ciones ambientales. la estructura y las evaluaciones de los efec- tivos en su area de distribucion. Ser. Contrib. INIDEP .561, Inst. Invest. Desarrollo Pesq., Mar del Plata, Argentina, 205 p. [in Spanish]. Bailey, K.M., R.C. Francis, and P.R. Stevens 1982 The life history and fishery of Pacific whiting, Merhwci/us producfus. Calif. Coop. Oceanic Fish. Invest. Rep. 23:81-98. Bellisio. N.B., R. Perrotta, J. Aenlle, A. Fortuny, and G. Padilla 1978 Merluza. Secretaria de Estado de Intereses Maritimos. Subsecretan'a de Pesca, Buenos Aires, 95 p. [in Spanish]. Brandhorst, W., J.P. Castello, R. Perez Habiaga, and B.H. Roa 1971 Evaluacion de los recursos de anchoita (Engraulis an- choita) frente a la Argentina y Uruguay. IV. Abundancia relativa entre las latitudes 34°40' y 42°10'S en relacion a las condiciones ambientales en mayo-junio de 1970. Proyecto Desarrollo Pesq., Ser. Inf. Tecnicos 34. 39 p. [in Spanish]. Brovver, R.L., H.S. Gohrband. W.G. Pichel, T.L. Signore, and C.C. Walton 1976 Satellite derived sea-surface temperatures from NOAA spacecraft. NOAA Tech. Memo. NESS 78. Natl. Environ. Satellite Serv.. Natl. Oceanic Atmos. Adm.. Wash., DC, 74 p. Caddy, J.F 1983 The cephalopods: Factors relevant to their population dynamics and to the assessment and management of stocks. hi Caddy, J.F. (ed.). Advances in assessment of world cephalo- pod resources, p. 41(;-452. FAO Fish. Tech. Pa[). 231, 452 p. Carreto, J. I.. R. Negri, and H. Benavides 1981a F^itoplancton. jjigmentos y nutrientes. Resultados cam- panas III y VI del B/I "Shinkai Maru". 1978. In Angelescu, V. (ed.), Campanas de investigacion pesquera realizadas en el Mar Argentine por los B/1 "Shinkai Maru" y "Walther Herwig" y el B/P "Marburg", anos 1978 y 1979. Resultados de la parte argentina, p. 181-120. Ser. Contrib. INIDEP 383, Inst. Invest. Desarrollo Pesq., Mar del Plata, Argentina, 339 p. [in Spanish. Engl. summ.]. Carreto, J. I., F. Ramirez, and C. Date 1981b. Zooplancton y produccion secundaria. Parte II. Distribucion y variaciAn estacional de la biomasa zooplanc- tdnica. In Angelescu, V. (ed.), Camparias de investigaci6n pes- quera realizadas en el Mar Argentine por los B/I "Shinkai Maru" y "Walther Herwig" y el B/P "Marburg", afios 1978 y 1979. Resultados de la parte argentina, p. 213-232. Ser. Contrib. INIDEP 383, Inst. Invest. Desarrollo Pesq., Mar del Plata. Argentina, 339 p. [in Spanish, Engl. summ.]. 176 Fishery Bulletin 88(1). 1990 Carreto, J. I.. H.R. Benavides, R.M. Negri, and P. Glorioso 1986 Toxic red-tide in the Argentine Sea. Plivtoplankton distribution and survival of the toxic dinoflagellate Guniiaidax ezcavata in a frontal area. J. Plankton Res. 8:15-28. Christiansen, H.E., and M.B. Cousseau 1971 La reproduccion de la merhiza en el Mar Argentine (Mer- lucciidae, Merluccius merluccius hubbsi). 2. La reproduccion de la merluza y su relaci6n con otros aspectos biologicos de la especie. Bol. Inst. Biol. Mar. Mar del Plata 20:43-73 [in Spanish]. Chung, S.-C, and S. Tanaka 1985 On the seasonal migration of Argentine hake, Moiucciux hubbsi Marini. Bull. Korean Fish. Soc. 18:571-580 [only English abstract consulted]. Ciechomski, J.D. de, and G. Weiss 1974 Distribucion de huevos y larvas de merluza. Merluccius merluccius hubbsi, en las aguas de la plataforma de la Argen- tina y Uruguay en relacion con la anchofta. EngmuUs anchoita, y las condiciones ambientales. Physis (B. Aires) (Sec. A) 33:185-198 [in Spani.sh]. Ciechomski, J.D. de, M.D. Ehrlich, C.A. Lasta, and R.P. Sanchez 1979a Campaiias realizadas por el buque de investigaci6n "Walther Herwig" en el Mar Argentine, desde mayo hasta noviembre de 1978. Organizacion y reseha de datos basicos obtenidos. Ser. Contrib. INIDEP 374, Inst. Invest. Desarrollo Pesq., Mar del Plata. Argentina, 311 p. ]in Spanish]. Ciechom.ski, J.D. de, R.P. Sanchez, M.D. Ehrlich, and C.A. Lasta 1979h Distriliucion de huevos y larvas de merluza {MfHuccius hubbst) en el Mar Argentine en diferentes epocas del aho y evaluacion de sus efectivos adultos desovantes. Ser. Contrib. INIDEP 380, Inst. Invest. Desarrollo Pesq., Mar del Plata. Argentina, 12 p. [in Spanish]. Ciechomski, J.D. de, and R.P. Sanchez 1983 Relationship between ichtyoplankton abundance and asso- ciated zooplankten biomass in the shelf waters off Argentina. Biol. Oceanogr. 3:77-101. Cordo, H.D. 1981 Resultados sobre la alimentacibn ile la merluza del Mar Argentine {Merhwcius hubbsi). Analisis biologico y estadistico de los datos obtenidos de las campahas de los B/I "Shinkai Maru" y "Walther Herwig" (1978-1979). In Angelescu, V. (ed.), Campanas de investigacion pesquera realizadas en el Mar Argentine por los B/I "Shinkai Maru" y "Walther Herwig" y el B/P "Marburg", anos 1978 y 1979. Resultados de la parte argentina. Ser. Contrib. INIDEP 383, Inst. Invest. Desar- rollo Pesq.. Mar del Plata, Argentina, 339 p. [in Spanish, Engl, summ.]. Cousseau, M.B., J.E. Hansen, and D.L. Gru 1979 Campanas realizadas por el buque de investigacion "Shinkai Maru" en el Mar Argentine, de.sde abril de 1978 hasta abril de 1979. Oganizacion y resena de datos basicos obtenidos. .Ser. Contrib. INIDEP 373, Inst. Inves. Desarrollo Pesq., Mar del Plata, Argentina, 625 p. [in Spanish]. Dickson, R.R., P.A. Gurbutt, and V.N. Pillai 1980 Satellite evidence nf enlianceii upwelling along the Euro- pean continental slope. .1. I'hys. Oceanogr. 10:813-819. Ermakov, Y.K. 197 1 The biology and fishery of Pacific hake, Merluccius pro- ductus. Ph.D. Diss., Tikhoekean. Nauchno-Issled. Inst. Tybn. Khoz. Okeanogr. (TINRO), Vladivo.stok, USSR. [Original' not seen, cited in Bailey et al., 1982]. FAO 1975 I'royecto de Desarrollo Pe.squero Argentina. Resultados y recomendaciones del proyecto. FI:DP/ARG/().5/510, 94 p. Fournier, R.O., M. van Del, J.S. Wilson, and N.B. Hargreaves 1979 Influence of the shelf-break front off Nova Scotia on phytoplankton standing stock in winter. J. Fish. Res. Board Can. 36:1228-1237. Francis, R.C. 1983 Population and trophic dynamics of Pacific hake {Merliu:- cius produclus). Can. J. Fish. Aquat. Sci. 40:1925-1943. Glorioso, P. 1987 Temperature distribution related to shelf-sea fronts on the Patagonian Shelf. Continental Shelf Res. 7:27-34. Hart, T.J. 1946 Report on trawling surveys on the Patagonian continental shelf. Discovery Rep. 23:227-408. Hubold, G. 1980a Hydrography and plankton off southern Brazil and Rio (le la Plata, August-November 1977. Atlantica 4:1-21. 1980b Second report on hydrography and plankton off south- ern Brazil and Rio de la Plata; autumn cruise: April-June 1978. Atlantica 4:23-42. Jones, B.W. 1974 World resources of hakes of the genus Merluccius. In Harden Jones, F.R. (ed,), Sea fisheries research, p. 139-166, John Wiley and Sons, NY. Juanico, M. 1982 South American squid fisheries: Some new aspects. In Proceedings of the International Squid Symposium, August 9-12, 1981, Boston, MA, p. 245-264. Prepared by N. Engl. Fish. Dev. Found. L'nipuh, NY. Maze, R., Y. Camus, and J.-Y. Le Tareau 1986 Formation de gradients thermiques a la surface de I'ocean, au-dessus d'un talus, par interaction entre les ondes internes et le melange du au vent. J. Cons. Cons. Int. Ex- plor. Mer 42:221-240 [in French]. Mundy, F.R. 1982 Computation of migratory timing statistics for adult Chinook salmon in the Yukon river, Alaska, and their relevance to fi.sheries management. N. Am. J. Fish. Manage. 2:359-370. Olson, D.B., G.P. Podesta, R.H. Evans, and O.B. Brown 1988 Tem|inral variations in the .separation of the Brazil and Malvinas currents. Deep-Sea Res. 35:1971-1990. Otero, H.O. 1986 Determinacion del cicio migratorio de la merluza comiin (Merluccius hubbsi) mediante el analisis de indices de densidad poblacional y concentraci6n del esfuerzo de pesca. Publica- cienes de la Comision Tecnica Mixta del Frente Marftimo 1. Montevideo, Uruguay, p. 75-92 [in Spanish]. Otero, H.O., and T. Kawai 1981 The stock assessment on commnn hake (Mi-rliiccius hubb- si) in the south-west Atlantic. Bull. Tokai Reg. Fish. Lab. 104:35-53. Otero, H.O., S.L Bezzi, R.G. Perrotta, J. A. Perez Comas, M.A. Simonazzi, and M.A. Renzi 1981 Los recursos pesqueros demersales del Mar Argentine. Parte III. Distribuci6n, estructura de la poblaci6n, biomasa y rendimiento potencial de la polaca, el bacalae austral, la merluza de cola y del calamar. In Angelescu, V. (ed.), Campahas de investigacion pesquera realizadas en el Mar Argentine por les B/I "Shinkai Maru" y "Walther Herwig" y el B/P "Marburg", anos 1978 y 1979. Resultados de la parte argentina, p. 28-41. Ser. Contrib. INIDEP .383, Inst. Invest. Desarrollo Pesq., Mar del Plata, Argentina, 339 p. [in Spanish, Engl. summ.(. Paffenhoffer, G.-A.. B.T. Wester, and W.D. Nicholas 1984 Zooplankten abundance in relation to state and type of intrusions onto the southeastern United States shelf during summer. .1. Mar. Res. 42:995-1017. Podest^: Migratory pattern of Merluccius hubbsi 177 Podesta. G.P. 1987 The relative fishing power of the Argentinian offshore fleet fishing for hake (Merburius hubbsi) in 1978 and 1979. J. Cons. Cons. Int. Explor. Mer 43:268-271. Quinn II, T.J. 1985 Catch-per-unit-effort: A statistical model for Pacific halibut (Hippoglosus stenolepis). Can. J. Fish. Aquat. Sei. 42:1423-1429. Rojo, A.L., and J.M. Silvosa 1969 Stock in vernal de la merluza (Merlwccins merluccivs huhh- si) del talud del-sector bonaerense. Campana "Merluza 69-1", 30 julio-9 agosto 1969. Proyecto Desarrollo Pesq., Ser. Inf. Tecnicos 20, 42 p. (in Spanish]. Rothschild, B.J., and M.Y.Y. Yong 1970 Apparent abundance, distribution and migrations of albacore, Thiinnus alahinga, on the North Pacific longline grounds. U.S. Fish Wildl. Serv. Spec. Sci. Rep. Fish. 623, 37 p. Rugolo, L.J. 1984 Methods for the comparison of timing behavior applied to the pink salmon fisheries of Prince William Sound, Alaska. Tech. Rep. 84-7, Old Dominion Univ., School of Sciences and Health Professions, Dep. Oceanogr., Norfolk, VA, 22,5 p. Sosa, M., R. Valdes, and H. Ramis 1976 Recursos pesqueros— pesquerfas en el Atlantico sudocci- dental. Rev. Invest., Cent. Invest. Pesq. (Cuba) 2:201-266 [in Spanish]. Abstract.— Knowledge of popu- lation dynamics and production of dominant benthic organisms on the northwestern Atlantic continental shelf is important for developing or interpreting models addressing eco- logical or fishery management ques- tions. This paper estimates the re- cruitment frequency and success, life span, growth rate, and production of a major component of the sand- bottom benthic community in this re- gion: the common sand dollar Echl- naracluiius parma (Echinodermata: Echinoidea). These estimates are based on size-frequency data col- lected at least annually between 1977 and 1985 from three sites in the Mid- dle Atlantic Bight and one site on Georges Bank. Larval recruitment of E. parvia to the sites was mostly an- nual, but persistent cohorts were ir- regular. Average life span was about 8 years with some rare, larger, and presumably older animals present; the ma.ximum size found was ,54 mm. A time series of modal size progres- sions suggests mean growth in body width ranged from 3.5 to 6 mm/yr over a period of at least 5 years, with growth curves for cohorts appearing to be sigmoidal. Annual production: biomass ratios for definable cohorts varied with age, ranging from -0.04 for senescent older cohorts to 8.10 for juveniles. Population Dynamics, GrowtPi, and Production Estimates for the Sand Dollar Echinarachnius parma Frank W. Steimie Sandy Hook Laboratory. Northeast Fisheries Center National Marine Fisheries Service, NOAA Highlands. New Jersey 07732 Manuscript accepted 18 .July 1989. Fishery Bulletin, U.S. 88:179-189. Information on growth and produc- tion rates of major marine taxa can support the development of better ecological and fishery production theories or models (Winberg 1971, Greze 1978). Echinoderms are one of the major marine benthic taxa; how- ever, information on their gi'owth and production rates is scarce. For exam- ple, production estimates are cur- rently available for little more than a dozen species worldwide (Richards and Riley 1967, Zaika 1972, Miller and Mann 1973, Buchanan and War- wick 1974, Greze 1978, Warwick et al. 1978, Warwick and George 1980). On the Atlantic continental shelf north of Cape Hatteras, North Caro- lina, the sand doWar Echinamchniufi parma is an abundant echinoderm of fine to medium-sand habitats (Coe 1972, Wigley and Theroux 1981). Be- cause of its relatively large size (>50 mm) and population densities which can reach hundreds of adult individu- als per m- (Caracciolo and Steimie 1983), it may be a "keystone" species as defined by Paine (1969), i.e., a ma- jor factor in structuring the benthic community by dominating habitat and detrital food use. Stanley and James (1971) conclude this species to be the second most important factor, after major storms, in reworking surface sediments. This sediment reworking can disturb smaller epifaunal benthic species and larval recruitment, which can include shellfish species of com- mercial interest. For example, Rich- ardson et al, (1983) studied an E. par- ma community off Rhode Island, and studies of similar Pacific and Atlantic sand dollar species Dendraster excen- t7-icus and Mellita quinquiesperfora- ta by Smith (1981), Creed and Coull (1984), and J. A. Reidenauer (Dep. Oceanogr., Fla. State Univ., Tallahas- see, FL 32306, pers. comm., July 1985) found that aggregations of sand dol- lars substantially alter benthic macro- fauna community structure, especi- ally for tube dwellers and meiofauna. Besides its potential role in control- ling benthic community structure, E. parma also occurs frequently in the diets of some commercially/recrea- tionally valuable fish, including had- dock Melanogrammus aeglefinus, summer flounder Pa ralichthys denta- tus, American plaice Hippoglossoides platessoides, scup Stenotomns chry- sops. tautog Tautoga onitis, and yel- lowtail flounder Limandaferruginea (Bigelow and Schroeder 1953, Coe 1972, Maurer and Bowman 1975, Collie 1987). However, its food value (Kcal/g) is half that of other common benthic prey, e,g., polychaetes or crustaceans (Steimie and Terranova 1985). Despite wide use of £". parma as prey, extensive populations of this echinoderm, primarily of individuals too large to be eaten by most preda- tors, may be an energy sink (Mercer 1982), For example, E. parma can comprise 40-50% of the total benthic macrofaunal biomass on Georges Bank (Steimie 1987). 179 180 Fishery Bulletin 88(1), 1990 CONN NEW YORK , ^l- MASS ~^ S\ ISLAND ,»/;/•• ty^^ k y® ■ «^" .vA^" .^^ DEL , 70 Ot^ t>«»«^ 1. „, ,■ joo' 0 50 100 KILOMETERS 68" Figure 1 Station locatitms of the four study sites in the Middle Atlantic Bight-Georges Bank area. Other sites examined which did not produce reliable samples oi Erhinarachnius parma are also included ( • ). Several studies have examined aspects of £". parma's life history (for review, see Caracciolo and Steimle 1983). Only two studies have examined its growth (Cocanour 1969, Graef 1977) and none its production. Since E. parnia can be a major contributor to the overall benthic community biomass, estimates of its production could greatly support an estimate of total community production. Warwick (1980) has shown that a major proportion of total community production is often attributed to one or a few species. The purpose of this study was to examine E. parmn size frequencies from a series of at least annual collec- tions to provide information on the species' population structure, dynamics, and growth, and to estimate its production at stations in the Middle Atlantic Bight and on Georges Bank. The term "production" is used in this paper to indicate the amount of biomass added to a unit of bottom area per year. Material and methods Echinarachnius parma were obtained from archived benthic collections made at three stations (11, 15A, 17) in the Middle Atlantic Bight and one station (23) on southwestern Georges Bank (Fig. 1). Water depths at these stations were 20 m (Stn. 17), 30 m (Stn. 15 A), 50 m (Stn. 11), and 70 m (Stn. 23). The archived collections consisted of 3-5 replicate 0.1-m^ Smith-Mclntyre grab collections at each station sampled on a quarterly to annual basis, 1978 to 1985, as part of the Northeast Monitoring Program (NEMP). Collection periods varied slightly between stations (see Figures 2-5 for specifics). Grabs were handled to minimize loss of sample by surficial material washout; grabs suspected of being excessively disturbed were resampled. The samples were washed through l.O-mm mesh sieves prior to 1980, and through 0.5-mm mesh sieves thereafter. While E. parma larvae are reported to undergo metamorphosis and settle to the seabed from the plankton at a width of about 0.4 mm (Coca- nour 1969), the l.O-mm sieve collected most postlar- vae because few were collected on the 0.5-mm sieve when both sieve sizes were used together in 1980 and 1981. The sieved benthic samples were fixed in lO'Fo buffered formalin, transferred to 70% ethanol within a few days, and later sorted. Size-frequency distributions were based on measur- ing the diameter of each individual, parallel to the anal pore margin of the test, using either an ocular micro- meter or vernier caliper. To estimate the ash-free dry weight (AFDW) of different cohort size classes, addi- tional E. parma specimens were collected at the above sites and frozen, then sorted into 5-mm size groups covering the 5-50 mm size range, with each group con- taining 3-10 individuals, as available. Each size group was dried for 12 hours at 60°C and ashed for 4 hours at 500°C to produce a mean AFDW value per mean width (L in mm) and the regression; AFDW = 0.02L- '•'' (/•- =0.991). Overlapping cohorts were initially sepa- rated by probit graphical analysis (Cassie 1954), and these estimates were refined by the NORMSEP pro- gram (Tomlinson 1971) to separate overlapping dis- tributions. Growth rates were estimated by modal progression analysis of the time series of size-frequen- cy histograms (Figs. 2-5) for collections at each sta- tion. Production estimates are based on Crisp's (1971 ) growth-increment survivorship-curve method for po])U- lations with distinguishable recruitment tuid age classes (cohorts). Recruitment events in the Middle Atlantic Bight were assumed to l)e the result of fall-winter spawning, based on when recruits are first detected in Steimie Population, growth, and production o^ Echinarachnius parma 181 lOOi 50 40 20 APR 1978 {1 N = 33 1.o|n n DEC 1980 N-725 A^- SEP 1978 '° N:71 4 r' Uj^jw^ -TV, „ „ „ JUL 1981 N=338 40 > 20 z iBri&i ~ ~- APR 1979 ^°^ N=23 2| FEB 1982 N=250 JUL 1982 N = 60 3 40i o UJ 20 ir li. H 40, z ^ on O 20 a. 40 20 40 20 n 0 JUL 1979 N = 4 '0 „Jl/h^ _ SEP 1979 ''°1 N--19 20 _n JT_.-T_, _ _ _ „ „ _ DEC 1982 N = 69 iy\„, . . - DEC 1979 '^° N=82 20 rftJL J 15 JUL 1983 N=15 JUL 1980 ,„ N=85 10 _ fl Eid^T^Pfifc n AUG 1984 N = 26 10 20 30 40 SIZE(mm) 50 60 10 20 30 40 SIZE(mm) 50 60 Figure 2 Ech inarachnius parma size-frequen- cy histograms per collection at station 11. Middle Atlantic Bight. 40 20 60- 20^ 40 20 40 20 40 20 40 20 40 20 40 20 .^^S^ Lw A^ -Jh^ 20 30 SIZE(mm) JUL 1977 40 N=137 20 N = 228 APR 1978 *° N=467 20 SEP 1978 40 20 APR 1979 40] N=212 20 JUL 1979 '*0 N--115 20 SEP 1979 40 N=293 20- DEC 1979 ""O N=259 20 JUL 1980 N: 147 DEC 1980 N = 186 AUG 1981 N = 48 . r-^lTyg rvn— -^ FEB 1982 N=125 JUL 1982 N:363 JUL 1983 N:89 NOV 1983 N=122 n AUG 1984 N:5 ^v.^^-^^^^. JUL 1985 N:213 20 30 SIZE(mm) Figure 3 EchiNiirm-hniu.'i parma size-frequency histograms per collection at station 17. Middle Atlantic Bight. the samples (winter-spring) and thus identified by the pair of years covering the probable peak spawning period. The recruitment season was estimated, in some cases, from back-calculations based on growth curves. Ages assigned to each cohort are based on inspecting the length-frequency histograms and noting when recruitment, i.e., >2 mm size mode, was apparent. These can be considered relative ages, although it is highly probable they represent absolute age ( + 3 months) if it is assumed that a delayed metamorphosis of the larvae is not normally a major problem; there is no evidence to suggest otherwise. The reliance on size-frequency analysis for growth and production estimates assumes (1) frequency modes represent cohorts, a reasonable assumption when re- cruitment period can be determined and modal shifts can be traced over a period of time, and (2) that the size distribution within each cohort was normal or representative of the true population and unbiased. A departure from normality may not be critical to the ac- curacy of mean size estimates (McNew and Summerfelt 1978). The sample sizes were often below the minimum (400) suggested by Cohen (1966) to separate accurate- ly two cohorts in a single size-frequency analysis, with a larger sample needed for separating more than two 182 Fishery Bulletin 1990 100 50 i 5 80 S 80 or u- 40 g 80 ^y^ r^^. 1- 80 40 \L 20 30 40 SIZE(mm) APR 1978 '°° N=141 50- JUL 1979 SEP 1978 N=29 40 i APR 1979 80 N-97 40 80 "n = 47 40 SEP 1979 80 N=20 40 DEC 1979 80 N=2211 40 DEC 1980 N=121 AUG 1981 N-226 ft- FEB 1982 N=929 AUG 1982 N^416 J^ -, --,JT-„ JUL 1980 N = 70 10- rx iH JUL 1983 N:59 Aug 1984 N=36 JUL 1985 N=33 20 30 40 SIZE(mm) JUL Figure 4 Ech inii rac/i » iiis pn rinii size-frcinKTioy histograms piT ci)llectioii al station 15A, Middle Atlantic Bight. groups. The time series of population size-frequencies available in this study could estimate poorly defined cohorts and modal peaks by visual examination of the overall modal patterns in the time series. Results The size-frequency distributions, grouped into 2-mm size classes for all collections at each station, show a high degree of variability in population structure (Figs. 2-5). The distributions for Station 1 1 , off Delaware Bay (Fig. 2), show a population with a variety of size classes evident in all collections; cohort definition is often weak, largely because of small sample sizes for most collections. Size-frequency distributions at Station 17. off Central New Jersey (Fig. 3), in contrast to Station 11, show a very limited population, restricted entirely to two annual recruitments in 1976 and 1978. Both of the two cohorts at Station 17 appear to persist until at least 1985 (the last collection of the study), although they were inseparable after 1981. There is no evidence of any additional recruits, even though the population reached a size, >27 mm, when spawning should occur (Ruddell 1977). The situation at Station 15A, off Fire Island, is similar to that of Station 11, with several sizes present in the population subsequent to initial collec- tions (Fig. 4). There appear to be two recent strong cohorts, that of 1978 and 1982. Some non-persistent recruitment was evident in other years, however. The E. parma population structure at station 23, on north- east Georges Bank, was similar to that of Station 17, with no evidence of a population prior to 1978 and with the 1978 and 1979 cohorts the strongest, although non-persistent recruitment was noted for other years sampled. APR 1978 90p N = 117 80 5 FEB 1982 SEP 1979 80-5 N=19 1,0 10 20 SIZECmm) Figure 5 Er!u»in-i:chiii}is [iiirmii sizA'-frequency histograms per collection at station 23. southwestern (ieorges Bank. The estimated growth curves (based on change in cohort's modal size progressions over time) of persis- tent and prominent cohorts at Stations 15A, 17 and 23, were sigmoidal (Figs. 6-8). The highly variable size- frequency distributions at Station 1 1 (Fig. 2) did not produce reliable modal trends to use for growth esti- mates. The mean cohort growth rates, approximated from modal progression analysis for definable cohorts ranged from 4.0 to 6.5 mm per year (Table 1). Table 2 summarizes the estimated annual production (P), mean biomass (B), and P:B ratio for each cohort, and the total annual population production and P:B ratio for each station, showing much variability. A com- parison of individual-cohort annual P:B ratios with Steimie Population, growth, and production of Echinsrachnius parms 183 Figure 6 Growth curves from the shift of the mean widths of prom- inent Echuiarachnius parma cohorts, over time, at sta- tion 17. Middle Atlantic Bight. Figure 7 Growth curves from the shift of the mean widths and ash-free dry-weight values of tlie prominent 1977-78 Echbmrachniiis parma cohort, over time, at station 15A, Middle Atlantic Bight. 30n ^^ .^ E E -5 20- X 1- Q K""' '" .A z 10- < LU "^ ^1 5 -'"ii -' ^ 1 ."5"' 1978 1979 1980 1981 1982 Figure 8 Growth curves from the shift of the mean widths of the prominent 1978 Echinarachmus parvia cohort, over time, at station 23, south- western Georges Bank. Table 1 Mean estimates of Erhinarachnius parma growth rate, bas- ed on cohort-size modal shifts at three stations in the Middle Atlantic Bight and southwestern Georges Bank. Station Approx. mean growth rate Cohort (mm/yr) 17 1.5A 2.3 1976-77 6.0 1977-78 4.5 1977-78 5.5 1978 4.n estimated cohort age (Fig. 9), based on the modal pro- gression analysis or estimated growth rates, for all sta- tions combined produced a skewed normal curve. Discussion The population dynamics of E. parma, with variable annual recruitment and cohort survival patterns (Figs. 2-5), are in general agreement with population dyna- mics reported for other echinoderms (Ebert and Dex- ter 1975, Lane and Lawrence 1980, Ebert 1983, Beu- kema 1985). Recruitment is evident as peaks in the 0-2 mm size-frequency columns of Figures 2-5. Station 17 (Fig. 3) was the exception, however, to the almost annual or frequent recruitment pattern; detectable recruitment was evident there in only 2 years. This recruitment pattern may be associated with anoxic con- ditions that defaunated this area during the summer of 1976 (Steimie and Radosh 1979). The definition of the first definable cohort at this station in July 1977 included some animals larger (to 12 mm) than the 6-mm estimated maximum annual growth rate, although the size-frequency mode was about 6 mm for the initial July 1977 collection (Fig. 3). It must be assumed any preanoxia recruitment in 1976 did not survive the event, as few living E. parma were found in the area 184 Fishery Bulletin 1990 Table 2 1 Estimates of annual cohort pro duction and P:B ratios for Erliiiinnichiniis parma and total annual popul ition production and F:B at 1 all stations in the Middle Atlantic Bight and southwestern Georges Bank. Production (P) values are mg ash-free dry we ght (AFDW)/ m ■ • year, and biomass (B) va^ ues are mg AFDW. Conversions to energy equivalents are based on a mean conversion of 24 KJ/g | AFDW (5.74 Kcal/g AFDW) (f rom Steimle and Terranova 1985). Cohort Cohort Cohort Cohort Cohort Station/ 1972-73? 1976-77 1978-79 1979-80 1981-82 Total P/m- annual period P B P:B P B P:B P B P:B P B P:B P B P:B mg AFDW Kcal P:B Station 11 Nov-Oct 1977-78 184 1217 0.15 21 71 0.30 205 1.18 0.16 1978-79 599 2843 0.21 270 208 1.30 180 94 1.91 1049 6.02 0.31 1979-80 781 717 1.09 1004 810 1.24 952 541 1.75 30 10 3.00 2767 15.88 1.34 1980-81 211 2288 0.09 102 429 0.24 611 848 0.72 170 185 0.92 9 345 0.03 1103 6.33 0.27 1981-82 369 697 0..53 974 722 1.35 371 344 1.08 14 9 1.56 1728 9.92 0.98 1982-83 179 518 0.35 2 1 2.00 X = 181 1172 1.04 6.72 0.35 0.57 Cohort Cohort Station 17 1976-77 1977-78 l'.t7fi 77 X02 667 1.19 802 4.60 1.19 1977-7>; 1 ri.'-iii 1092 1.42 310 101 3.07 1866 10.71 1.56 1978-79 i'M)7 1420 2.09 1479 1.523 0.97 4446 25.52 1.51 1979-80 i4or, 4707 0.72 4675 3564 1.31 8080 46.38 0.98 1980-81 l(i32 6072 0.27 5305 5534 0.96 6937 39.82 0.60 1981-82 2801 7054 0.40 2796 27870 0.10 5597 32.13 0.16 X = 5068 29.09 0.94 Cohort Cohort Cohort Cohort Cohort Station 15 1972-74? 1977-78 1978-79? 1979-80? 1980-81 A 1977-78 -13 1584 - -0.01 49 13 3.77 36 0.21 0.02 1978-79 -65 1936 - -0.03 550 556 0.99 485 2.78 0.19 1979-80 -58 4554 - -0.01 2369 965 2.45 86 44 1.95 2397 13.76 0.43 1980-81 1887 2404 0.78 1887 10.83 0.78 1981-82 -65 2403 - -0.03 1927 5092 0.38 218 27 8.10 239 101 2.37 2319 13.31 0.30 1982-83 53 1362 0.04 371 14960 0.02 202 25 8.10 A' = 626 1292 3.59 7.41 0.04 0.29 Cohort Cohort Cohort Cohort Cohort Station 23 1976-77 1977-78 1978-79 1979-80 1980-81 May-April 1978-79 27 8 3.39 27 0. 1 5 3.39 1979-80 70 107 0.65 83 84 0.99 153 0.88 0.80 1980-81 246 353 0.711 310 377 0.82 30 9 3.33 7 06 1.17 593 3.40 0.80 1981-82 466 579 0.8U 1019 1327 0.77 40 41 0.98 23 3 7.67 2 2 1.00 1 550 8.90 0.79 1982-83 236 1838 0.13 1178 3131 0.38 .V = 1414 747 8.12 4.29 0.28 1.21 during the fall of 1976. These few were found at sites, nearby, that were on the tops of sand ridges near or in the thermocline, offering less anoxia-stressed con- ditions (Steimle and Radosh 1979). This suggests there may have been accelerated growth of some recruits spawned in nearby areas unaffected by anoxia, possi- bly a response to a rich detrital food su|)ply remain- ing from the related phytoplankton bloom (Mahoney and Steimle 1979) and an absence of competitors oi' predators. Steimie Population, growth, and production of Echinsrschnius parma 185 8.1 8.1 4.5h N = 1 1 '■ 35- m a 2.5- ^ _^ z < tu S 1.5- y 1 N: 8 N N = 8 N--7 1 -2 N = 12 , ■ 0.5- 0.0- N = 1 1 "^-^-^ N = 2 N=1 1 2 3 4 5 6 7 APPROXIMATE AGE OF COHORT(Years) 8 9 Figure 9 Relationship between i;ir(.iduction (P) to biomass (B) ratios (mean and ranges) and estimated age for all definable Echhiarachnius parma cohorts from all stations, Middle Atlantic Bight and Georges Bank. The period of peak recruitment varied slightly be- tween the Middle Atlantic Bight stations and that of Georges Bank. Recruitment was heaviest or most con- sistent in the winter to early spring (December-April) at Stations 11 and 15A (Figs. 2,4) which is consistent with most reports of fall E. pnrmn spawning (Fewkes 1886. Cocanour and Allen 1967, Ruddell 1977). Recruit- ment of the two cohorts found at Station 17 (Fig. 3) iwurred prior to when sampling began, but the 1978 ^■ot\in was probably recruited in the winter of 1977-78. (.''r: (.ux^rges Bank (Fig. 5), however, recruitment ap- peared in all collections except those of September, and most did not persist. Collie's (1987) observation of yellowtail iloiinder selectively feeding on < 12 mm size E. pcDDia on Georges Bank is relevant to the poor sur- vival of most new cohorts there. The life span for persistent cohorts in the Middle Atlantic Bight, estimated from recruitment in 1976 to the end of the study, was at ieas,t 7-8 years, e.g., see Figure 3. The presence of a few larger individuals, to 54 mm, suggests a maximum life span of perhaps 15-1- years, given the probable very slow growth rates of the larger/older segment of the population. However, max- imum widths were generally below 40 mm which sug- gests that most of the population did not live much more than 8 years. At the Georges Bank station, recruitment prior to the 1978 cohort was not evident, thus reasonable estimates of maximum size, life span, or other population variables were not possible because collections there covered less than 5 years. The estimated life span of E. parma in the study area is most likely less than the estimated 21 years reported by Brykov (1975) for a Pacific population of this dis- junctly trans-subboreal species, based on assumed an- nual growth-ring counts to estimate age. However, individuals of E. parma have been measured that generally exceed 80 mm (maximum reported is 92 mm) in the northern Gulf of Maine (Lohavanijaya 1964, Cocanour 1969), and if one uses a mean growth rate of about 3 mm per year (range 1.5-4 mm/year) as deter- mined from tagged, middle-sized (33-55 mm) indivi- duals measured after 2 years in the Gulf of Maine (Cocanour 1969), then 20 -i- year life spans could occur in boreal waters. If annual ring counts to size estimates of Cocanour (1969) alone were used, it appears a lesser age would be reasonable because 60-70 mm animals had only 6-7 growth rings. The reduced growth rates of a cohort as it reaches senescence may not produce distinguishable growth rings and juvenile rings may also be weak or obscure, so ring counts could under- estimate age. Also, Cocanour (1969) has suggested that mature E. parma may alternate years of growth and gametogenesis, at least in the Gulf of Maine; there is no evidence of this alternate-year growth pattern in the growth curves (Figs. 6-8). The maximum width measured in the Middle Atlantic Bight was 54 mm. This apparent difference in maximum size, and possibly life spans, between populations in the temperate Middle Atlantic Bight and the boreal Gulf of Maine suggests a latitude or temperature-related size cline for this species, as predicted for echinoids in general by Ebert (1975). Other factors, e.g., food supply, genetic vari- ability, differences in methodologies, or habitat or predation pressure, could also be involved. Although the collection periods caused some prob- lems in accurately defining the specific time of peak recruitment, the quarterly-to-semiannual sampling over at least the initial 5-year period of the study was probably sufficient to estimate growth and produc- tion. Parsons et al. (1977), for example, suggest the 186 Fishery Bulletin 88(1), 1990 sampling period for growth and production estimates should be no greater than 10% of the generation time (from egg to sexual maturity) of a species. Ruddell (1977) reports E. panna reaches sexual maturity at a width of about 27 mm in the Middle Atlantic Bight. This width suggests an age of approximately 5 years from the growth curve at Stations 17 and 15 A (Figs. 7 and 8); thus, semiannual sampling appears to meet this sug- gested minimum. The slightly sigmoidal growth curves, for most prom- inent cohorts (Figs. 6-8), differ from the more parabolic growth curves reported for other sand dollar species (Ebert and Dexter 1975, Lane and Lawrence 1980) and sea urchins (Ebert 1982). However, Sime and Cranmer (1985) and Nichols et al. (1985) report sigmoidal growth curves in North Sea echinoid species. Some of the ir- regularities in growth curves in the present study may be due to a sigmoidal. intra-annual, seasonal growth cycle that would have highest rates in the early sum- mer and lowest in the early winter (Cocanour 1969). These curves, based on defined cohort means, may be somewhat imprecise because of the small sample sizes; however, the modes are basically congruent with the means and support the general shapes of the curves. The apparent growth curves for Stations 17 (1976/77 cohort. Fig. 6) and 15A (1977/78 cohort. Fig. 7) are almost congruent although the Station 17 (1977/78) cohort has a lower slope, similar to that of the Station 15A (1981/82) cohort or that of the Station 23 (1977/78) cohort curve. The cause of difference in these two sets of slopes is unknown, at present. The mean annual growth rates (Table 1), estimated by modal progression analysis, ranged from 4.0 to 6.5 mm per year and were greater than the average 1.5- 4.0 mm per year rates reported by Cocanour (1969), based on tagged individuals, for an intertidal popula- tion along the northern Maine coast. However, the ratios of number of apparent growth rings to diameter were generally higher, ranging from about 7.5 to 15 mm per ring. The rates estimated in the present study are less, however, than the approximate 7 mm per year from Woods Hole, Massachusetts (Durham 1955) or the 7.5 mm per year off northern New Jersey (Graef 1977), both determined by growth ring counts for 45-48 mm specimens. The diameter-to-ring ratio ranged from 4.3 to 11.3 mm per ring for eight smaller specimens in Graef's study and appeared to peak at about 34 mm with three rings evident. The size range examined by these two studies is near the predicted size asymptote, and senescence may be involved with zero or negative growth which would underestimate the age and over- estimate overall growth rates. Negative growth (shrinkage), evident as slope declines on the right side of curves found for some older cohorts, agrees with similar findings by Cocanour (1969) for this species and for other sand dollars (Lane and Lawrence 1980). Senescence is thought by these authors to be the cause of this negative growth in older populations, although Gordon (1929) and Cocanour (19(i9) rei)ort some shrink- age in all age classes during winter months. There is better agreement between the growth-ring count estimates and the size-frequency estimates of growth rate if it can be assumed that the relatively slow juvenile growth, suggested by the sigmoid growth curves and reported by Gordon (1929) and Highsmith and Emlet (1986) for a Pacific population, may leave an obscure ring for the first or perhaps even second year of survival, thus lowering the size-age estimates based on ring counts. On Georges Bank, the most per- sistent or dominant cohort was the 1978 recruitment. The growth patterns (Fig. 8) suggest a mean growth rate of about 4 mm per year for the first 4 years of this cohort, although there was some irregularity in July 1979. This rate is slightly lower than most rates for similiar cohorts at other stations. The overall, estimated annual cohort and population production varied from station-to-station and from year-to-year (Table 2), reflecting the variable dynamics of recruitment, growth, and mortality of each popula- tion. It is interesting to note that population produc- tion peaked in the November 1979-October 1980 pro- duction year at all three Middle Atlantic Bight stations and later, during the May 1981 -April 1982 production year for Georges Bank. The E. parma population at Station 17 was the most productive for the period ex- amined, and its total production peaked in 1979-80 at 8080 mg AFDW/m^ per year This production level (46 Kcal/m- per year), when converted to energy equiva- lence (24 KJ or 5.7 Kcal/g AFDW; Steimie and Ter- ranova 1985), is greater than the total benthic com- munity |)roduction reported in many North Atlantic continental shelf areas (table 5, Steimie 1985); even the mean value for this station, 29 Kcal/m- per year is comparable to some total community values. The high production at Station 17 is probably related to the 1976 anoxia episode, mentioned previously. The P:B ratios (Table 2) varied from negative values for older cohorts to a maximum of about 8.1 for some juvenile cohorts at Station ISA. A comparison of mean cohort P:B ratios against approximate cohort age (Fig. 9) for all stations comliined shows a skewed normal curve, with mean ratios per age group ranging from a maximum of about 2.4 for 1-2 year-old cohorts to less than 0.5 for cohorts older than 5 years. The declining part of this curve appears to be common in other marine organisms, e.g., Warwick (1980). The low P:B ratios for 0-1 year-old cohorts could suggest a relative- ly slow growth of young-of-the-year juveniles (Figs. 6-8). The apparently low initial ratios could also be in- fluenced by accurately establishing time of recruitment. Steimie' Population, growth, and production of Echinarachnius parnna 187 Table 3 Summary of production (P) to biomass (B) ratios for some echinoderms. Class Species P:B (range) Reference Holothuroidea Cuntmaria eUmgata 0.26 Zaika (1972)* Echinoidea Stningyloce)itnitii>! dri)rbachiensis 0.80 Miller and Mann (1973) Ech inoca rd i u m co rda t u m (-0.002-3.7) Warwicl< et al. (1978) Brissopsis lyrifera 0.30 Buchanan and Warwick (1974) Moira atrops 0.82-1.0 Moore and Lopez (1966) Echinarachnius pa mm (0.02-3.39) Present study Stelleroidea Astropecten irregularis 0.005 Warwick et al. (1978) Asterias forhesi 2.64 Richards and Riley (1967) Asterias i-uhens (3.6-7.3) Zaika (1972)' Ophiura alhida 0.84 Arntz (1971) Ophiura tejiiurata 0.68 Warwick et al. (1978) 0.48 Warwick and George (1980) Am.phiura filiformis 1.97 O'Connor et al. (1986) Amphiopius eoniortode 2.26 Moore (1972) Amphipholus squamata 1.8 George and Warwick (1985) Ophiothrix frag His 1.8 George and Warwick (1985) Oph ionepthys Urn icola 2.33 Moore (1972) Amphoidia arctica (0.6-0.8) Zaika (1972)*) 'Ratios calculated from published data. Mean P:B ratios for E. parma, although generally declining with the age of the cohort (Fig. 9), are com- parable to the range of P:B ratios determined for other echinoderms (Table 3). Because of the variable relation- ship of P:B ratios to the (1) specific population age structure (Fig. 9), (2) segment of the life span, or (3) other factors, the significance of this comparison is uncertain. This problem has attracted the attention of other authors who have attempted to improve the utility of interspecific comparisons of P:B ratios by scaling them to either temperature (Parsons et al. 1977), body mass at maturity (Banse and Mosher 1980), or life span (Winberg 1971, Zaika 1972). The results could be affected by a number of sources of error or variability that commonly affect similar studies and are usually difficult to assess, e.g., sam- pling error. We did not have samples near the end of the presumed winter-spring peak of larval settlement to obtain accurate estimates of the time and density of larval recruitment. This could affect the growth curves or production estimates. The production of any missed recruits during the few weeks or months before they were sampled would probably be minor because of their small size and high mortality rate. It could influence the cohort P:B-to-age curve (Fig. 9) and par- tially explain the initial low ratio and rise. Later, semi- annual sampling also made it difficult to determine if E. parma spawns more than once per year. There are suggestions of substantial recruitment outside of the approximate peak periods (or delayed recruitment) in some of the histograms, especially for the Georges Bank station (Fig. 5). Another source of error may be systematic, e.g., an- nual mean production or P:B values do not adequately consider a cyclic annual growth and production pat- terns. These appear to have maximum values in the warmer months, based on the study of Cocanour (1969), when specific rates would be much higher than the mean annual values. The final source of error is statistical, especially con- sidering the small sampling sizes available to estimate many of the variables, e.g., the size mode of each cohort. The relatively long time series of data exam- ined, however, allowed a better definition of cohort size modes and growth than is normally possible by at- tempts to intei-pret a limited number of collections over a short time and reduced the necessity for alternate aging estimates. Fishery Bulletin 88( I 1990 Acknowledgments The author wishes to thank B. Baker and R. Terranova for their assistance in the laboratory, and M. Montone, A. Gruber, D. Schibik, D. Ralph, and M. Cox for their patient assistance in producing this paper. C. MacKen- zie, P. Fallon, A. Calabrese, R. Reid, A. Bejda, and anonymous reviewers provided several useful com- ments and suggestions which are greatly appreciated. Citations Arntz W.E. 1971 Biomasse und production des makrobenthos in den tieferen Teilen der Kieler Bucht im Jahr 1968. Kiel. Meeres- forsch. 27:36-61 [cited in Greze 1978). Banse. K., and S. Mosher 1980 Adult liody mass and annual produotion/liioniass relation- ships of field populations. Ecol. Monogr. 50:335-379. Beukema, J.J. 1985 Growth and dynamics in populations of Echhiorardium cordatum living in the the North Sea off the Dutch north coast. Neth. J. Sea Res. 19:129-134. Bigelow, H.B., and W.C. Schrocder 1953 Fishes of the Gulf of Maine. U.S. Fish Wildl. Sen-., Fish. Bull. 53, 577 p. Brykov, V.A. 1975 Individual age and life span of certain species of sea ur- chins in the Sea of Japan. Sov. J. Mar. Biol. 1:111-116 [Engl, transl. from Biol. Morya 1:39-44). Buchanan. J.B., and R.M. Warwick 1974 An estimate of benthic macrofaunal production in the off- shore mud of the Northumberland Coast. J. Mar. Biol. Assoc. UK 54:197-222. Caracciolo. J.V.. and F.W. Steimle. Jr. 1983 An atlas of the distribution and abundance of dominant benthic invertebrates in the New York Bight apex with reviews of their life histories. NOAA Tech. Rep. NMFS SSRF-766. Natl. Oceanic Atmos. Adm., Natl. Mar. Fish. Serv., 58 p. Cassie, R.M. 1954 Some uses of probability paper in the analysis of size frequency distributions. Aust. J. Mar. Freshwater Res. 5: 513-522." Cocanour. B.A. 1969 Growth and reproduction of the sand dollar, Echinn- rachniini finrma (Echinodermata:Echinoidea). Ph.D. thesis, Univ. Maine, Orono, 96 p. Cocanour, B.A., and K. Allen 1967 The breeding cycles of a sand dollar and a sea urchin. Comp. Biochem. Physiol. 20:327-331. Coe. W. 1972 Starfishes, serpent stars, sea urchins and sea cucumbers of the Northeast (Echinoderms of C(mnecticut). Dover Press, NY, 152 p. Cohen, A.C. 1966 Discussion of "Estimation of [larameters for a mixture of normal distributions" by Victor Hasselblad. Technometrics 83:445-446. Collie. J.S. 1987 F\)od selection liy yellowtail flounder (LuiKuidn ftrru- ginni) on Georges Bank. ('an. J. Fish. Aquat. Sci. 44:3.57-367. Creed, E.L.. and B.C. Coull 1984 Sand dollar, Melltta quinquiesperforahi (Leske), and sea pansy, Rerulln reniformia (Cuvier) effects on meiofaunal abun- dance. J. Exp. Mar. Biol. Ecol. 84:225-234. Crisp, D.J. 1971 Energy flow measurements. In Holme, N.A., and A.D. Mclntyre (eds.). Methods for the study of marine benthos, p. 197-279. Blackwell Sci. Publ., Oxford. Durham, J.W. 1955 Clas.sification of Clypeasteroid Echinoids. Univ. Calif. Publ. Geol. Sci. 31:73-198. Ebert, T.A. 1975 Growth and mortality of post-larval echinoids. Am. Zool. 15:755-775. 1982 Longevity, life history, and relative body wall size in sea urchins. Ecol. Monogr. 52:353-394. 1983 Recruitment in echinoderms. In Jangoux. M., and J.M. Lawrence (eds.), Echinoderm studies 1, p. 16Vt-2(l3. A. A. Balkema, Rotterdam. Ebert, T.A., and D.M. Dexter 1975 A natural history study of Enmpr grandis and AMIita grantii. two sand dollars in the Northern Gulf nf (':ilif(irnia. Mexico. Mar. Biol. .32:397-407. F'ewke.s, J.W. 1886 Preliminary observations on the development of (^/i/i/o- pholis and Echinarachnius. Bull. Mus. Comp. Zool. 12: 105-152. George. C.L.. and R.M. Warwick 1985 Annual macrofaunal production in a hard-bottom reef community. J. Mar. Biol. Assoc. UK 65:713-735. Gordon, I. 1929 Skeletal development in Arhacui, Eclmianirhniux. and LeptuMn-MS. Philos. Trans. 217:289-334. Graef, J.G. 1977 A preliminary investigation of the growth rate and natiu-al history oiEehitKimchniua pnnmi (Lamarck) in the New York Bight apex area. Unpubl. manuscr., Sandy Hook Lab., Northeast Fish. Cent., Natl. Mar. Fish. Serv., NOAA. High- lands, NJ 07732, 25 p. Greze, V.N. 1978 Production in animal po|iulations. In Kinne, 0. (ed.). Marine ecology, vol. IV. Dynanucs, p. o c CT U- O.t Hatching Week 3 Termination O.J 0.2 A J / \ 0,1 i \ Ar 1a . \\ ex 6 e 10 12 U 16 B 20 ?2 2(. lb It 10 i; n (5 65 n Length (mm ) Figure 8 Length-frequency distribution of Pacific herring from Bristol Bay. Alaska, in the 2000-m'' basin at Fl0(levigen Biological Station, Nor- way, at hatching, week-3, and study termination on day-63. from an essentially unimodal population at hatching (Fig. 8). The cause of distinct size modes in this experi- ment is unknown, but it may be related to success at first feeding or be genetically based, because size dif- ferences developed early. Blaxter and Hunter (1982) report that strong size hierarchy has been observed in some aquarium studies of herring, which they speculate may involve crowding or food competition; however, in a previous study of Atlantic herring larvae in a 4400 nr'' basin a size hier- archy did not develop (0iestad and Moksness 1981). McGurk (1984) found that length and weight of larvae of the same hatch differed at 30 days posthatch as a result of withholding food from 0 to 14 days. This sug- gests that modes may be a reflection of feeding suc- cess. Moksness et al. (1989) observed that in ocean catfish differences in feeding success at first feeding was expressed as differential growth that persisted throughout the larval period. Although most of the literature suggests that differ- ential growth is related to early feeding success genetic factors may also influence larval growth and survival. Christopher et al. (1988), in a study of capelin Multotus viUosus larvae from known females crossed with a single male, found that the amount of yolk varied among the females and that growth and survival was correlated with yolk quantity. Whatever the cause, one would tend to conclude that the modes were distinct "cohorts" from different hatchings if observed in the wild. The significance of these results is that a method such as otolith daily ring counts is required to validate larval age and the use of length-frequency analysis to determine the occur- rence of cohorts may lead to erroneous assumptions. If differential growth among a brood is a common Wespestad and Moksness: Growth and survival of Clupea pallasii 199 occui-rence in nature, it may complicate analyses and results of larval cohort studies such as those of Lambert (1984). Acknowledgments We would like to thank Dr. D. Hay and two anonymous reviewers for their constructive comments and sugges- tions, Jeff Skrade and Dr. Phil Mundy of the Alaska Department of Fish and Game for assistance in obtain- ing and transporting Bristol Bay herring eggs, and Dr. Per Hognestad, Director of the Fl0devigen Bio- logical Station, and his staff for logistical and technical assistance. This research was partially funded by the Norwegian Fisheries Research Council. Citations Alderdice, D.F., and A.S. Hourston 1985 Factors influencing development and survival of Pacific herring (Clupea harengus pallasi) eggs and larvae to begin- ning of exogenous feeding. Can. J. Fish. Aquat. Sci. 42 (Suppl. l):.56-68. Arai, M.N., and D.E. Hay 1982 Predation by medusa on Pacific herring (Clupea harengus pallasi) larvae. Can. J. Fish. Aquat. Sci. .39:1.537-1540. Bla.xter, J.H.S., and J.R. Hunter 1982 The biology of the clupeoid fishes. Adv. Mar. Biol. 20:1-123. Checkley, D.M. 1982 The aging of juvenile Pacific herring by otolith analysis. Final rep.. NOAA contract 82-ABA-lOOl. Northwest and Alaska Fish. Cent., Natl. Mar. Fish. Serv., NOAA, 7600 Sand Point Way NE, Seattle, WA 98115-0070, 9 p. 1983 Ecology of Pacific herring, Clupea harengus pallasi. larvae in the Herring Sea. Final data rep., Rapid Response Project, Alaska Sea Grant Coll. Proj. RR/80-04, Univ. Alaska, Fairbanks, AK 99775, 24 p. Christopher. R.. W.C. Leggett, and J. A. Brown 1988 Maternal influences on, and correlations among, egg size, size at hatching, age at hatching, and age at death in unfed capelin (Mallotus mtlosus). Int. Counc. Explor. Sea Early Life History Symposium Contrib. 40. " Fried, S.M., and V.G. Wespestad 1985 Productivity of Pacific herring (Clupea harengus pallasi) in the eastern Bering Sea under various patterns of exploita- tion. Can. J. Fish. Aquat. Sci. 42 (Suppl. 1):181-191. Grant, W.S., and F.M. Utter 1984 Biochemical popiJation genetics of Pacific herring (Clupea pallast). Can. .J. Fish. Aquat. Sci. 41:856-865. Hay, D.E. 1981 Effects of capture and fixation on gut contents and body size of Pacific herring larvae. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 178:39.5-400. 1982 Fixation shrinkage of herring larvae: Effects of salinity, formalin concentration and other factors. Can. J. Fish. Aquat. Sci. .39:1138-1143. Hourston, A.S., H. Rosenthal, and S. Kerr 1981 Capacity of juvenile Pacific herring (Clupea harengus pallasi) to feed on larvae of their own species. Can. Tech. Rep. Fish. Aquat. Sci. 1044, 9 p. Ingraham, J. 1981 Shelf environment. In Hood, D.. and J. Calder (eds.). The eastern Bering Sea shelf: Oceanography and resources, vol. 1, p. 455-470. U.S. Gov. Print. Off., Wash., DC. Ivlev, V.S. 1961 Experimental ecology of the feeding of fishes. Yale Univ. Press, New Haven, 302 p. Keegen, T.P., B.S. Miller, and D.R. Gunderson 1986 Nearshore distribution and growth of Pacific herring larvae near Birch Point, Washington. In Haegle, C.W. (ed.). Proceedings, Fifth Pacific coast herring workshop. Oct. 29-30. 1985. Can. Manuscr. Rep. Fish. Aquat. Sci. 1871:76-82. Ki0rboe, T., P. Munk, and J.G. St0Urup 1985 First feedmg by larval herring (Clupea harengus L.). Dana 5:95-107. Lambert, T.C. 1984 Larval cohort successi(jn in herring (Clupea harengus) and capelin (Mallotus villosus). Can. ,1. Fish. Aquat. Sci. 41: 1.552-1,564. McGurk, M.D. 1984 Effects of delayed feeding and temperature on the age of irreversible starvation and on the rates of growth and mor- tality of Pacific herring larvae. Mar. Biol. 84:13-26. Messieh, S.N., D.S. Moore, and P. Rubec 1987 Estimation of age and growth of larval Atlantic herring as inferred from examination of daily growth increments of otoliths. In Summerfelt, R.C., and G.E. Hall (eds.). The age and growth of fish, p. 433-442. Iowa State Univ. Press, Ames. Moksness, E. 1982 Food uptake, growth and survival of capelin larvae (Mallotus villosus Muller) in an outdoor constructed basin. Fiskeridir. Dir. Skr. Ser. Havunders. 17:267-285. Moksness, E., and V. 0iestad 1987 Interaction of Norwegian spring-spawning herring lar- vae (Clupea harengus) and Barents Sea capelin larvae (Mallotus villostis) in a mesocosm study. .1. Cons. Int. Explor. Mer 44:42.5-436. Moksness, E., and V. Wespestad 1989 Ageing and back-calculating growth rates of Pacific her- ring, Clupea harengus jjallasi, larvae by reading daily otolith increments. Fish. Bull.. U.S. 87:.509-513. Moksness, E.. J. Gj0seater, I.S. Fjallstein, and A. Reinert 1989 Start feeding and on-growing of ocean catfish (Anar- ehicus lupus) in the laboratory. Aquaculture 77(23):221-228. Morita. S. 1985 History of herring fishery and review of artificial prop- agation techniques for herring in .lapan. Can. J. Fish. Atjuat. Sci. 42 (Suppl. ]):181-191. 0iestad, V. 1983 Growth and survival of herring larvae and fry (Clupea harengus L.) exposed to different feeding regimes in ex- perimental ecosystems: Outdoor basin and plastic bags. Ph.D. thesis, Univ. Bergen, Bergen, Norway, 299 p. Oiestad. V., and E. Moksness 1981 Study of growth and survival of herring larvae (Clupea harengus L.) using plastic bags and concrete basin enclosures. Rapp. P.-V. Reun. Cons. Int. E.xplor. Mer 178:144-149. Rosenthal, H. 1969 Verdauungsgeschwindigkeit, Nahrungswahl und Narh- rungsbedarf bei den larven des herings, Clupea harengus Ber. Dtsch. Komm. Meeresforschung 20:60-69. Stevenson, J.C. 1962 Distribution and survival of herring larvae (Clupea palkisi Valenciennes) in British Columbia waters. J. Fish. Res. Board Can. 19:735-810. 200 Fishery Bulletin 88(1), 1990 Talbot, G.B.. and S.I. Johnson 1972 Rearing Pacific herring in the laboratory. Prog. Fish- Cult. 34:2-7 Ware D.M. 1975 Relationship between egg size, growth and natural mor- tality of larval fish. J. Fish. Res. Board Can. 32:2503-2512. Zweifel, J.R., and R. Lasker 1976 ['rehatch and posthatch growth of fishes— a general model. Fish. Bull., U.S. 74:609-(321. Hyperostosic Bones from the Mew Zealand Snapper Chrysophrys auratus (Sparidae) Robert W. Gauldie Fisheries Research Centre, P O Box 297, Wellington, New Zealand Present address: Hawaii Institute of Geophysics University of Hawaii at Manoa, Honolulu, Hawaii 96822 Zophie Czochanska Organic Chemistry Division Department of Scientific and Industrial Research Private Bag. Gracefield, Wellington, New Zealand General hvperostoses ai'e well known in fossil fish bone literature (Tiffany et al. 1980). Hyperostosic bones have been described as occurring as nodules on the ventral pterygio- phores as well as other bones of Re- cent fishes (Olsen 1971, Fierstine 1968, Konnerth 1966). Consistent occurrence of hyperostosic bones in many species has been taken to in- dicate that they are not pathological (Fierstine 1968). Various roles have been suggested for hyperostosic bones ranging from aids in fin erection (Fierstine 1968) to hydrostatic correction (Breder 1952). The hydrostatic suggestion arose from the observation that hyperostosic bones are "filled with fat" (Breder 1952), and oils in fish bones have been proposed as an aid to neutral buoyancy (Lee et al. 1975). Hyperostoses in sparids of the genus CJirysophrys have also been used as characters to separate Pagrus major from Chrysophrys auratus (Yasuda and Mizuguchi 1969). Hyperostosic bones occur in the snapper Chrysophrys auratus and are usually highly vascularized. We were interested in (1) the possible relationship between size of hyper- ostosic bones and fish size aiKl (2) the significance of the fatty-acid composition of hyperostoses to the biology of the fish. Materials and methods Snapper were measured as fork length, filleted along one side to reveal the hyperostosic bones, and photographed on a measuring board. The h,\T3erostosic bones and a sam- ple of the vertebrae from each fish were dissected out and frozen at - 15°C until the fat content was analyzed. The relative sizes of hy- perostosic bones were measured from the photographs. Two hyper- ostosic bones usually occur in the snapper, commonly on the seventh (preural) haemal spine and less com- monly on the sixth haemal spine from the tail in specimens of the size range (31-48 cm) used in this study. Both were measured. The eighth and ninth vertebrae from the tail were removed and analyzed as non- hyperostosic comparisons. One bone was cut through the hy- perostosis, mounted in epoxy resin, and ground down to about 50-^4m thickness and photographed with a WILD M400 stereomicroscope. Whole fish were x-rayed using a Faxitron Model 804 Radiographic Inspection Unit. Ten of the speci- mens examined were aged fi'om the otoliths using an annual check-ring method (Paul 1976). Reference to trade names does not imply en dorsement by the National Marine Fisheries Service, NOAA. Flesh and connective tissues were scraped off the bones. Prior to chop- ping, bones bearing the hyperosto- ses were cut diagnoally. Bones were weighed, chopped into small pieces, and extracted three times with chlo- roform-methanol (2:1 v/v) in an ul- trasonicator bath (Sanophon Ultra- sonic Ind.) containing iced-water for 10 minutes and finally extracted once with chloroform. The extracts were combined, filtered, and evapo- rated at 40° C in a rotary evapora- tor and weighed. Extracted lipids were stored under nitrogen at - 15°C pending analysis. Fat-free bone residues were dried to constant weight. The lipid ex- tract is expressed as a percentage on a dry matter basis (Table 1). Lipids were separated into classes by column chromatography and preparative thin-layer chromatog- raphy. All solvents used were redis- tilled prior to use. Portions of total lipids dissolved in chloroform were initially separated into fractions on 30 g and 12 g columns (depending upon sample sizes) packed with sili- ca acid (Koch-Light Lab. 5024h, 100-200 mesh) previously activated at 110°C. The elution procedure in- volved chloroform for neutral lipids and methanol for polar lipids. By combination of silicic-acid column chromatography and preparative thin-layer chromatography, neutral lipids were resolved into their con- stituent categories (Tables 1, 2). Preparation of methyl esters and gas-liquid chromatography (GLC) was carried out in the following way. Aliquots of total lipid extracts, which contained fatty acids com- bined as glycerides and phospholi- pids, and transesterified with BF;i- methanol (Van Wijngaarden 1967). Fatty-acid methy esters were ana- lyzed by GLC using a Pye Unicam GCV and a Philips PU 4500 capil- lary chromatograph fitted with flame ionization detectors. Both Manuscript accepted 9 August 1989. Fishery Bulletin, U.S. 88:201-206. 201 202 Fishery Bulletin 88(1). 1990 Table 1 Hyperostoses (H) and vertebrae (C), extraction and column chromatography in Cht 'ysoph ••yt: aurata. Average Fish 1 Fish 2 Fish 3 Fish 4 Fish 5 Fish 6 Fish 7 Fish 8 values H C H C H C H C H C H C H C H C H C Total lipid extract Weight (ing) 17.4 4.2 58.9 34.7 24.4 12.0 31.4 12.3 16.3 8.6 37.3 12.2 7.3 5.5 51.4 49.4 % (dry 3.11 4.57 16.91 17.73 11.50 10.71 2.04 5.10 6.96 8.14 22.70 16.90 8.00 14.29 29.30 22.93 12.57 12.55 matter basis) Column chromatography (%) Triglycerides 58.0 57.1 58.4 .58. 1 58.2 57.2 58.0 57.7 58.3 58.1 58.4 48.2 58.9 58.2 58.6 58.5 58.35 57.89 Sterols and 29.9 30.9 30.9 30.6 30.6 30.0 30.6 30.9 30.7 30.2 30.0 30.3 30.1 30.9 30.0 29.9 30.35 30.46 alcohols Total 87.9 88.0 88.9 88.6 88.6 87.5 88.6 88.6 89.0 88.3 88.4 88.5 89.0 89.1 88.6 88.4 88.63 88.38 neutral lipids Total polar 9.8 9.5 10.7 10.1 10.2 10.(1 10.2 10.6 10.4 10.5 9.6 10.6 9.6 10.6 10.1 9.9 10.08 10.23 liyiids Table 2 Fatty acid composition of the lipi Is of hyperostoses (H) and vertebrae (C)( e.xpressec as percentage of GLC peak areas) in Ch •ysofili ryu aurntus. Major components as fatty acid nr ethyl esters Fish 1 Fish 2 Fish 3 Fish 4 Fish 5 Snapper fillef* H C H C H C H C H C Saturated 14:0 4.1 4.4 3.9 4.2 5.8 6.3 5.1 4.7 4.3 3.8 4.4 15:0 0.8 0.7 1.0 0.8 0.7 0.7 0.8 1.0 0.8 0.9 1.0 16:0 .35.9 34.2 31.2 28.0 27.8 26.8 28.5 27.9 35.1 30.8 36.1 17:0 0.9 0.9 0.9 0.9 1.0 1.0 0.9 1.0 0.8 0.8 0.8 18:0 10.9 11.2 11.4 11.0 11.2 12.4 12.0 11.6 11.0 12.6 11.3 Other 0.3 0.7 0.4 1.9 4.4 0.8 1.9 2.8 0.3 1.1 0.9 Total saturated 52.9 52.1 48.8 46.8 50.9 48.0 49.2 49.0 .52.3 50.0 54.5 Unsaturated 16:1 a;;* 7.9 8.3 8.5 10.4 7.9 6.6 8.4 9.2 8.5 7.9 6.7 16:3 w. 0.9 1.0 1.2 1.4 0.9 2.5 1.5 1.2 0.9 1.0 0.8 18:1 w., 22.1 21.8 24.9 25.0 21.9 20.9 23.2 21.8 23.0 24.8 22.2 18:2 coc 1.0 1.3 1.9 1.0 1.2 2.1 1.6 1.4 2.0 1.7 0.8 18:3 w; 2.9 2.6 2.1 1.9 2.9 2.6 2.4 2.2 2.9 2.5 0.6 20:1 oj. 1.5 1.6 1.0 1.0 2.0 2.2 1.6 1.6 1.0 1,6 1.2 20:4 aj„ 1.7 1.7 1.3 0.5 1.0 1.5 1.1 1.4 1.2 1.1 2.1 20:4 w., 1.2 1.9 1.2 1.5 1.3 1.7 1.4 1.6 1.7 1.5 0.2 20:5 w.. 0.9 1.1 1.9 1.5 2.2 1.9 1.9 2.0 2.1 1.8 0.3 22:5 ojs 1.0 0.8 1.1 1.1 0.8 1.2 1.1 1.0 0.9 1.0 1.3 22:5 a;^ 0.8 0.7 0.7 1.1 1.1 1.3 1.0 1.2 1.0 1.2 1.8 22:6 oj., 1.2 1.4 0.3 1.4 0.8 1.1 0.9 0.7 0.4 (1.7 1.5 Other 2.1 1.7 3.1 3.6 3.3 4.3 2.8 3.7 0.2 1.4 4.2 Total unsaturated 45.2 45.9 49.2 51.4 47.3 49.9 48.9 49.0 45.8 48.2 43.7 Total branched 1.9 bond on of 2.0 2.0 1.8 nearest to terminal methyl grc snapper fillet reported by Hugh 1.8 up. es et al 2.1 (1980). 1.9 2.0 1.9 1.8 1.8 *ai, position of doublt **Fatty acid compositi NOTES Gauldie and Czochanska Hyperostosic bones from Chrysophrys auratus 203 Figure 1 Radiograph of Chrys^ophrys auratus hyperostosic bones showing the enlarged haemal spines. polar and non-polar columns were employed, the former containing 10% EGSS-X (210 x 0.254 cm i.d.) and the latter a BP-1 capillary column (25 x 0.33 mm i.d.). Eight pairs of hyperostoses and vertebrae were ana- lyzed (Table 1) for fat content and general fat composi- tion. A subset of pairs of hyperostoses and vertebrae was used in the identifications of fatty-acid methyl esters based upon comparison of retention times with those of authentic reference methyl esters run under identical conditions. Where particular reference stan- dards were not available, equivalent chain-length values reported by Hofstetter et al. (1965) and by Jamieson (1969) were accepted as criteria of identity. Identification was corroborated by hydrogenation of an aliquot of the esterified samples. Saturated and branched chain products were then analyzed by GLC on a polar column. The results of fatty-acid analyses were expressed as the percentage area occupied by each component methyl-ester peak relative to the total peak area. In the fatty-acid composition analyses re- ported here, the percentages were based on peak areas obtained with the 10% EGSS-X column (Table 2). Fatty-acid composition of snapper fillets (Hughes et al. 1980) is included for comparison in Table 2. Results Eight sets of snapper hyperostosic bones and controls were examined. An example of a pair of snapper hyper- ostosic bones is shown in Figure 1. In life, the hyper- ostosis has a reddish color and appears to be well vas- cularized. Commonly, hyperostoses occur both on the sixth and seventh haemal spine from the tail, as has also been described for Chrysophrys unicolor (Yasuda and Mizuguchi 1969). Although both hyperostoses may be enlarged, it is more common to find that the one 204 Fishery Bulletin 88(1), 1990 Figure 2 A section of a hyperostosis bone in ('hn/ftiiphri/s niiratiis shows a vacuolated structure (arrows) encased with bone. furthest from the tail shows less enlargement, as in Figure 1. Section of the snapper hyperostosic bone shows the vacuolated appearance (Fig. 2) characteristic of most fish bones. The relative size of the most caudal hyperostosis was measured as the ratio of the width of the hyperostosis to the width of the spinal vertebra to which the haemal spine, on which the hyperostosis occurs, was attached. A plot of the ratio of hyperostosis width/vertebra width to fork length is shown in Figure 3. The correlation between relative size of the hyperostosis and fork length was low, ?■ = 0.58. The correlation between the ratio of hyperostosis width/vertebra width to annual check-ring age was low, r = 0.53. Eight pairs of hyperostoses/vertebrae were examined for fat content and fat composition (Table 1 ). The mean fat content of the hyperostoses (% dry matter) was 12.57 + 19.34, and the mean fat content of vertebrae (% dry matter) was 12.53 + 13.04. Although there was a certain amount of variation in fat content between hyperostoses and vertebrae, it was unlikely to be sig- nificant in any biological sense because fish lipid com- position and content have been shown to be affected by diet to a greater extent than is shown in our data (Worthington and Lovell 1973). The fatty-acid composition was almost identical (±5%) for all samples and very similar to the table presented in Love (1980: 414) for marine fish fatty-acid composition. A typical fatty composition for five pairs of hyperostoses and vertebrae is shown in Table 2. There was no significant difference in fatty-acid com- position between the hyperostoses and vertebrae. Discussion Hypoerostosic bones of the snapper Ckrysophrys aurata increased in size with increase in size of the fish, NOTES Gauldie and Czochanska' Hyperostosic bones from Chrysophrys suratus 205 100^ ■ ^ ■ 5 5 2 80- X3 ID u > • £ 60- ■ • B • 5 ■ w • ■ CO o 40- . ■ ■ w o • 0 a > -c 20- o to DC 0- 1 ' 1 < 1 1 1 . 1 1 I 1 1 1 1 1 1 ' 30 32 34 36 38 40 42 44 Fork length (cm) 46 48 Figure 3 The ratio of hyperostosis/vertebra width in Chrysojihrys auratus is plotted against forlv length. but the correlation between hyperostosic bone size and fish size was low. The relative size of snapper hyper- ostosic bones did not appear to be related with any more significance to age than to size. The fat content of the snapper hyperostosic bones was not significantly different from that of the verte- brae. Nor was the composition of fat in the snapper hyperostosic bones significantly different from either the vertebrae of snapper or fats found in marine fishes generally. The difference in fat content between snap- per hyperostosic bones and vertebrae is low (Worth- ington and Loveil 1973) compared with the potential range of differences in fat content known in fishes. Snapper hyperostoses are vacuolated, but no more so than other bones in fish. The contribution of fat in hyperostosis to buoyancy is difficult to assess. The amount of lipid as a percent dry weight varies between individuals, but in some individuals is of the same order as the lipid content of the bones of fish species in which bone lipid may have a role in neutral buoyancy (Lee et al. 1975). However, the low correlation with both size and age suggests that if hyperostoses are related to buoyancy correction, then that relationship is tem- pered by factors other than size. The use of hyperostoses to discriminate species both in Recent genera, such as Chrysophrys (Yasuda and Mizuguchi 1969) and fossil species (Fierstine 1968) points to another possible e.xplanation. Hyperostoses occur with great regularity in certain bones in those species which display hyperostoses. We interpret such regularity as indicating that hyperostoses are under genetic control and are therefore good species char- acters, as are other genetically determined characters. But the expression of hyperostoses results from an in- determinate kind of ontogenetic process in which there is only a loose correlation between size of the hyper- ostoses and size or age. Thus the hyperostoses involve the local proliferation of cells (Breder 1952) but in a way that could be seen as physiologically between a normal bony growth and a bony tumor, almost a kind of controlled tumor which contains both fat cells and bony trabeculae. Snapper hyperostoses are readily ob- tainable and may provide a useful model to study the ultrastructural interactions between proliferating bone and fat cells. Acknowledgments The snapper used in this study came from a tagging study conducted by Arthur Hoare of the Fisheries Management Division, MAF, New Zealand. Annual check-ring ages were provided by Larry Paul of the Fisheries Research Division. Expert technical help was provided by Kevin Mulligan. Citations Breder. CM. 1952 The problem of directives to cellular proliferation as illus- trated liy ontogenetic processes in certain fishes. Growth 16: 189-198. Fierstine, H.L. 1968 Swollen dorsal fin elements in living and fossil Caranx (Teleostei: Carangidae). Contrib. Sci. Los Ang. 137:1-10. Hofstetter, H.H., N. Sen, and R.T. Holman 1965 Equivalent chain lengths of fatty acids. J. Am. Oil Chem. Soc. 42:537-.587. Hughes, J.T., Z. Czochanska. L. Pickston, and E.L. Hove 1980 The fatty acid content of some New Zealand fish. N.Z. J. Sci. 23:4.3-51. Jamieson, G.R. 1969 Ec^uivalent chain lengths of fatty acids. In Gunstone, F.I), (ed.), Topics in lipid chemistry. Vol. I, p. 214-232. Logos Press Ltd., London. Konnerth, A. 1966 Tilly bones. Oceanus 12:6-9. Lee. R.F., C.F. Phleger, and M.H. Horn 1975 Composition of oil in fish bones: possible function in neutral buoyancy. Comp. Biochem. Physiol. 50:13-16. Love, R.M. 1980 The chemical biology of fishes. Acad. Press. NY, 579 p. Olsen, S.J. 1971 Swollen bones in the AUantic cutlass fish Trirhmnis lep- turns Linnaeus. Copeia 1971(1):174-175. 206 Fishery Bulletin 88(1-), 1990 Paul. L. 1976 A study on age. growth and population structure of the snapper Chnjsophrys auratus (Firster), in the Hauraki Gulf, New Zealand. N.Z. Fish. Res. Bull. 13, 67 p. Tiffany, W.J.. R.E. Pelham, and F.W. Howell 1980 Hyperostosis in Florida fossil fishes. Fla. Sci. 43:4.5-49. Van Wijngaarden, D. 1967 Transesterification of fatty acids. Anal. Chem. 39: 838-8.57. Worthington, R.E., and R.T. Lovell 1973 Fatty acids of channelcatfish (Ictalurus pmictatus): Variance components related to diet, replications within diets, and variability among fish. J. Fish. Res. Board Can. 30: 1604-1608. Yasuda, F., and K. Mizuguchi 1969 Specific characters of three sparid fishes referred to the genus Chrysophrys in the Indo-Pacific. Jpn. J. Ichthyol. 16: 24-30. Gonad Morphology, Histology, and Spermatogenesis In South Pacific Albacore Tuna Thunnus alalunga (Scombridae) Frank J. Ratty Department of Biology, San Diego State University San Diego, California 92182 R. Michael Laurs Coastal Fishieries Resource Division, Soutfiwest Fisfieries Center National Marine Fisheries Service, NOAA, PO Box 271 La Jolla, California 92038 Raymond M. Kelly School of Medicine, M-OI I. University of California, San Diego La Jolla, California 92093-001 I The reproductive biology of tuna, and tlie albacore Thunnus alalunga in particular, is not well understood with respect to functional morphol- ogy of the gonads in relation to sex- ual maturity. In the course of an in- vestigation of the genetic variability of albacore, differences in morphol- ogy were observed between the right and left gonad of male and fe- male fish and the size of the fat body associated with the gonads. Histo- logical examinations were made to determine the relationship between these morphological differences and the reproductive state of the go- nads. In this paper we report find- ings made on gonad morphology, histology, and spermatogenesis and relate them to the reproductive biol- ogy of this species. Materials and methods Albacore specimens were collected onboard the U.S. NOAA RV Town- send Cromwell and the New Zea- land RV Ka.haroa during coopera- tive studies conducted in 1986 and 1987 to investigate the biology/ecol- ogy and fishery exploration of alba- core tuna in the South Pacific Ocean (Laurs 198G, Laurs et al. 1987). The fish were caught using standard al- bacore troll-fishing methods (Dot- son 1980). In addition, specimens were collected during commercial fishing operations conducted in 1987 by Japanese longline vessels. Locations of specimen collection are summarized in Table 1. Soon after the fish were caught, fork lengths were measured to the nearest millimeter and the gonads and associated fat body were dis- sected from each fish. The right and left gonads were preserved sepa- rately in 10% neutral buffered for- malin. Histological preparations were initiated by washing the tissue in running water for 12 hours. An American Optical T/P tissue pro- cessor was used to dehydrate and embed the tissue in paraplast. The tissue was sectioned at 5 microns and stained with Harris' hematox- ylin followed by eosin counter stain (Humason 1979, Gabe 1976). Each testis subsample histological preparation was classified accord- Reference to trade names does not imply en- dorsement by the National Marine Fisheries Service, NOAA. ing to (a) estimates of the relative abundances of spermatocytes, sper- matids, and sperm; (b) sperm-duct development; and (c) sexual matur- ity. In addition, the cross-sectional area of the medial section of each testis was determined. Electron mi- crographs were prepared of sperm, which had been washed from the sperm duct of selected testes with physiological saline and fixed in 3% gluteraldehyde in 0.1 M phosphate buffer. Results We observed asymmetrical differ- ences in the size of both ovaries and testes and in the size of the fat body associated with gonads from both sexes offish. Histological examina- tion showed that all stages of sex- ual maturity were represented in the specimens of testes, allowing us to describe spermatogenesis. No sexually mature ovaries were found, limiting the results for females to asymmetrical size differences. Number and size of fish examined Gonads from 197 albacore were examined histologically during the study. Sex ratio was nearly 1:1, with 106 males and 91 females. The fish ranged in fork length (FL) from 50 to 101 cm. Over 90% of females and over 80% of males were less than 90 cm FL, the size at which sexual maturity is believed to first occur in female albacore (Ueyanagi 1955, Otsu and Uchida 1959, Otsu and Hansen 1961). Length-frequen- cy distributions of the fish examined are given in Figure 1. Gonad general morphology As in most teleosts, the gonads in male and female albacore are paired, elongate organs, located in the dor- sal portion of the body cavity. They Manuscript accepted 18 September 1989. Fishery Bulletin, U.S. 88:207-216. 207 208 Fishery Bulletin 88(1). 1990 Table I Summary of information on Tliiinniif: nlaliinija specimens used in study. No. species Fork length (cm) Location in South Pacific Ship Date Range Median RV Townsend Cromwell 1986 33 50.6-84.4 71.5 Central 17-21 Feb. lat. 38°-40°S long. 146°-152°W RV TiiivnacntI C roniiri'll 1987 25 Jan-l Feb. 33 50.6-97.5 63.8 Central lat. 35°-39'=S long. 151°-157°W RV Kiilinnid 1986 15 Feb.-28 Mar. 13 50.0-76.0 70.0 Western lat. 33''-44°S long. 166°E-178°W F\' \Vakaf:hi(i 1987 21 June 9 63.0-98.0 88.0 Western lat. 37°S long. 179°E FV Z III ho 1987 20-2:^ .June 18 85.2-101.1 94.5 Western lat. 37°-38°S long. 179°E Q FEMALE/SUBSURFACE 0 MALE/SUBSURFACE O FEMALE/SURFACE □ MALE/SURFACE \R n 0-54 55-59 60-64 6569 70-74 75-79 80-84 85-89 90-94 95-99 -10 0 FORK LENGTH {cm) Figure 1 Length-frequency distributions for male and female South Pacific albacore tuna examined in this study. Surface troll-caught fish and subsurface longline-caught fish are differentiated. are suspended by the mesentery which contains a fat body closely associated with the gonad between it and the dorsal body wall (Godsil and Byers 1944). The fat body will be discussed in a later section. Testes are thin and ribbon-like in immature fish, and develop into somewhat flattened, whitish-yellow organs which are relatively solid as the fish advance in maturity. Ova- ries in immature fish closely resemble immature testes in appearance. They become progressively enlarged in length and girth and tend to be somewhat pinkish in color as the fish progresses in maturity. We initially made casual observations, when collect- ing gonad tissue samples for genetic-variability studies of North Pacific albacore, that the right gonads were most often larger than those on the left side. The pres- ent study of nearly 200 South Pacific albacore showed that while right and left gonads of both sexes were about the same length, the cross-sectional area and volume of the right ovaries and testes were regularly larger than those on the left side. Figure 2 shows a pair of ovaries and associated fat bodies, portion of the digestive tract, and ventral fin dissected from a 94-cm FL North Pacific albacore. The disparity in size be- tween the right and left ovaries and associated fatty bodies is typical of what we have observed in this study and in casual observations. Size differences between right and left testes Measurements of the cross-sectional area (mm-) of the medial sections of the right and left testes provide quantitative estimates of size differences between the two. Data summarized in Table 2 show that the right testis was larger than the left in 72% of male fish ex- amined, the two testes were equal in size in 2%, and the left was larger than the right in 23%. The results given in Figure 3 show that the dissimi- larity in size increases with increasing fork length. The right testis is about 28-82% larger than the left in fish less than 80 cm FL, but increases to about 6(5-75% larger in fish over 80 cm. Data on displacement vol- umes measured on a sample of the right and left testes showed results similar to those made on the measure- ments of cross section. Examination of relationships between fork length and cross-sectional area of the NOTES Ratty et al : Reproductive biology of Thunnus alalunga 209 Right Fat Body. Right Ovary Portion of Digestive Track Left Fat Body Left Ovary •Ventral Fin Figure 2 Pair of ovaries and associated fat bodies, portion of digestive system, and ventral fin from a 94-cm FL albacore tuna caught in the North Pacific. 210 Fishery Bulletin 88(1). 1990 Relative size of right (re = 106). Table 2 and left testes Thuiiniis alalunga Relative size of testes Percent occurrence Right larger than left Right equal to left Right smaller than left No data 72 2 23 3 < LU < 80 70 - 6 LU o z LU cr UJ HI O < \- z LU o cc LU Q. - i - i - 1 - « <70 71-80 81-90 FORK LENGTH (cm) >90 Figure 3 Percentage difference in cross-sectional area of the medial sections of right and left testes of South Pacific albacore, by fork-length. right and left testes showed that the sizes of both gonads increase exponentially with increasing length of the fish, and that the area of the right testis does not increase at a faster rate than the left. In summary, the right testis is predominately larger than the left; however, they both appear to increase in size at similar rates as the fish increases in length. Histology of testes The albacore testis has a lohular-type cellular ar- rangement, characteristic of most teleosts (Billard et al. 1982). We classified the germ cells into the follow- ing phases of development: spermatogonia, primary spermatocytes, secondary spermatocytes, spermatids, and spermatozoa. A. B^ PS, .V* c. sm ■^1^ «»'«../ Figure 4 Spermatogenic cells oli.servcd in testes of South Pacific albacore. (A) Spermatogonia (sg) situated on lobule wall ( x 400); (B) primary spermatocyte (ps) cells and secondary spermatocyte (ss) cells (x GOO); (C) spermatids (sm) and spermatozoa (s) (x400). NOTES Ratty et al : Reproductive biology of Thunnus alalunga Figure 5 Electron micrograph of South Pacific albacore sperma- tozoa (scale indicates 1 ^im). The primary spermatogonia are the largest cells with diameters of 12-16 jum. They have a prominent nucleo- lus, are basophilic, and are found in cysts with vary- ing numbers of individual cells. The cysts are usually found singly near the periphery of the testis; a cyst situated on the lobule wall is shown in Figure 4A. Primary spermatocytes are oval or spherical with diameters of 8-12 yim. They have no visible nuclear membrane and the chromatin material occupies most of the cell. Figure 4B shows examples of primary sper- matocytes. Secondary spermatocytes are very small, spherical cells with diameters of 4-7 ^m. Unlike primary spermatocytes, the chromatin is found in a clumped condition, similar to the spermatids. The cells occur in groups and are strongly basophilic. Figure 4B shows secondary spermatocytes. Spermatids are strongly basophilic spherical cells with diameters of 2-4 fim. As they mature, they become smaller and the chromatin becomes uniformly con- densed. Figure 4C shows that after detaching from the lobule wall, the spermatids remain in dense clusters. The spermatozoa have a rounded nucleus and are morphologically subdivided into head, neck piece, short midpiece, and tail; as in other teleosts, there is no acrosome. They measure 1-2 ^m in diameter, excluding the tail. An electron micrograph of an individual sperm is shown in Figure 5. Table 3 Differences in meiotic activity between right and left testes, Thunnus alalunga (n = 106). Meiotic activity Percent occurrence Left < right Left = right Left > right No data 10 26 61 3 Meiotic activity Relative abundances of sperm, spermatids, and sper- matocytes were taken as a measure of meiotic activ- ity. As previously discussed, the albacore testes were generally asymmetrical in size with the right testis more often larger than the left. However, we found that meiotic activity and sperm-duct development showed the opposite relationship. The left testis was meiotically more active than the right in 61% of males examined, the left and right testes had about equal ac- tivity in 26%, and the right was more active than the left in 10% (see Table 3). Examination of data for each pair of testes showed a 60% occurrence of greater meiotic activity in the left testis relative to larger size of the right testis. Other combinations of meiotic activity and relative size of testes showed a 15% or less occurrence (see Fig. 6). Relative abundance of sperm, spermatids, and spermatocytes Absolute counts of the number of spermatogenic cells would be exceedingly difficult to make. Instead we estimated relative abundances of spermatocytes, sper- matids, and sperm present using the categories none, few, sonw, and many. The results show that for all sizes of fish, the relative abundance of sperm was higher in the left testis than the right testis. For example, there were nearly 2 times more many and 2V2 times fewer none categories of sperm observed for the left than the right testis. In both the left and right testes of fish in all sizes, relative abundance of sperm was highest in the caudal portion and decreased towards the rostral end (Figure 7). The findings are similar for spermatids and sperma- tocytes, but the differences between testes are not as pronounced as those for sperm. For all sizes of fish, relative abundances of both spermatids and spermato- cytes were higher in the left than the right testis. For the left testis, there were about 20% and 30% more many and 40% and 50% fewer none, categories ob- served for spermatids and spermatocytes, respectively. 212 Fishery Bulletin 88(1 1990 ISl LU o z UJ (T CE O o o 70 60 50 40 30 20 10 Gonad size R * L AND Meiotic activity R^L RELATIVE SIZE AND MEIOTIC ACTIVITY Figure 6 Percentage occurrence of relative level of meiotic activity and relative size of right (R) and left (L) gonads of South Pacific albacore. On the abscissa, upper expression refers to relative gonad size, and lower expression refers to relative level of meiotic activity. 28 26 24 22 20 18 16 14 12 10 a 6 4 RIGHT LEFT ROSTRAL RIGHT LEFT MEDIAL TESTIS SECTION RIGHT LEFT CAUDAL Figure 7 Relative abundance of sperm in rostral, medial, and caudal portions of the right and left testes of South Pacific albacore. A. ■ J *> .• -i*'. it >t t '>^' I Figure 8 (A) Empty sperm (iuct of an immature South Pacific albacore; (B) sperm duct of a mature albacore filled with sperm. xlOO. than for the right testis. The relative abundance of each group of germ cells was highest in the caudal portion and decreased towards the rostral portion in both the left and right testes in all sizes of fish. Sperm duct development During testicular morjihogenesis, teleost sperm ducts are formed by somatic cells derived from the somatic wall (Nagahama 1983). The main duct is present in im- mature albacore testes (Fig. 8A) in which no sperm are found. In mature albacore, we observed that the sperm duct is located near the center of the testis (Fig. 8B) and that a branching system of tubules radiates toward and end blindly at the testicular periphery. These obser- NOTES Ratty et al: Reproducwe biology of Thunnus .^lalunga 213 70 71-80 81-90 >90 FORK LENGTH (cm) Figure 9 Comparison of sexual maturity levels of male South Pacific albacore by fork length. Table 4 Percentage sexual maturity of fish by fork length Stage of Fork length (cm) sexual maturity <70 71-80 81-90 >90 No sperm present 51 18 40 37 and immature Intermediate 42 4(3 20 42 Mature and fully 7 36 40 21 mature vations differ from those reported for many other tele- osts, where the main duct is located along the dorsal surface of the testis and may not be present in imma- ture stages of the testis (Grier et al. 1980). As the production of sperm increases, morphological changes occur in the main sperm duct. In the immature testis the main sperm duct is thin-walled and highly convoluted (Fig. 8A). In the mature testis the main duct, when filled with sperm, has much thicker wails and is less convoluted and more rounded (Fig. 8B). The production of sperm is also associated with increased vascularization of the testis, first in the caudal portion followed by the medial and rostral portions. Ratios of observations characterizing immature and mature stages of sperm-duct development (thin:thick. convoluted;round, large:small) showed a higher propor- tion of mature sperm ducts in the left testis than the right. The results also reveal that sperm-duct develop- ment is greatest in the caudal, intermediate in the medial, and least in the rostral portions of both the right and left testes. As expected, there was an increas- ing trend in sperm-duct development with increasing fork length. Sexual maturity Fish were classified into five levels of sexual maturity based on relative abundance of sperm observed in the testes: (1) No sperm present in any portion of either testis; (2) immature, few sperm observed in one or more portions of either or both testes; (3) intermediate, few or some sperm in more than one portion of both testes, but not the sperm duct; (4) mature, many sperm ob- served in most portions of both testes and the sperm duct; and (5) fully mature, many sperm observed in all portions of both testes and the sperm duct. Information on sexual maturity in relation to fork- length size categories is given in Figure 9 and Table 4. Slightly more than 50% of fish 50-70 cm FL were immature, including 28% (derived from Figure 9) with no sperm present in any portion of either testis. Over 40% were in the intermediate level and 7% were judged sexually mature, but none was fully mature. Fish in size groups larger than 70 cm FL had a lower proportion of immature individuals than did fish less than 70 cm, ranging from 18% to 40% (Table 4), including only 3-10% with no sperm present in any portion of either testis (derived from Figure 9). The proportion of fish in the intermediate level of maturity ranged between 20% and 46% for fish in size groups larger than 70 cm FL. The percentages of fish in the mature and fully matui'e categories were nearly the same for size groups 71-80 cm and 81-90 cm FL, 36% and 40%, respective- ly, but only 21% for the size group over 90 cm. The 71-80 cm FL size group had the highest proportion of fully sexually mature individuals, 23%, while only about 5% of fish larger than 90 cm were fully sexually mature (derived from Figure 9). The lower proportion of mature and fully mature in- dividuals in the size group larger than 90 cm FL ap- pears to be related to geograpliic region and/or type of fishing gear. Figure 10 compares maturity levels be- tween fish caught in the midocean region of the cen- tral South Pacific by surface trolling and those caught in coastal waters off New Zealand by subsurface longline. It is evident in Figiu'e 10 that a higher pro- portion of immature and a lower proportion of mature and fully mature individuals were caught in coastal waters than in the midocean region, even though the longline-caught fish were much larger, averaging 92.3 214 Fishery Bulletin 88(1). 1990 cm FL, than troll-caught fish, averaging 69.2 cm. Figure 1 shows that virtually all of the fish in the size category greater than 90 cm were caught in coastal waters by subsurface longline. Fat body A lobulated mass of fat-like tissue was usually observed in the mesentery attached to testes and ovaries. In some cases it was extremely large and well developed, but quite rudimentary in others. It was generally white- to-creamy in color and always formed in segmental lobes. The extent of the fat body was usually coinci- dent with that of the gonad, but sometimes it extended a variable distance anteriorly beyond the gonad. The mass of the fat body was generally correlated with size of the gonad. We observed that the right testis or ovary of the albacore, which is generally larger than the left, usually had a larger fat body (see Figure 2). In addi- tion, we observed that the fat body was proportionately larger in immature fish that were meiotically inactive than in fish with gonads actively producing sperm. Discussion 50 30 PERCENT OF FISH 10 0 10 30 50 I I I I COASTAL, LONGLINE n = 27 MEAN F.L. - 92.3 1 1 1 1 MID-OCEAN. TROLLING n =65 MEAN F.L. : 69.2 y/y M A I U H t / Z^- FULLY MATURE r Figure 10 Comparison of sexual maturity levels of male albacore caught in coastal South Pacific waters by subsurface longline (left) with tho.se caught in midocean South Pacific region by surface trolling (right). According to literature pertaining to the general struc- ture and function of teleost gonads, the development of gonads is usually similar on both sides of the fish. However, asymmetry in gonad size has been observed sometimes in some species, usually at or near the time of breeding (Turner 1919, Robertson 1957, Sanwal and Khana 1972, Dalela et al. 1977). Asymmetry in gonad size appears to be common for albacore, at least in prespawning stages. We observed asymmetry in 95% of fish examined, with the right gonad larger than the left in 72% of the cases. The disparity in size increased with increasing fork length. The right testis averaged about 30% larger than the left in fish less than 80 cm and averaged more than 70% larger in fish greater than 80 cm. We do not have data collected during active spawning to ascertain if this relationship exists at that stage. Otsu and Uchida (1959) reported that the right ovary was usually slightly larger than the left in albacore specimens from the central North Pacific, but they did not quantify the difference. Ueyanagi (1955) noted that the right and left ovaries from a mature female alba- core caught in the Indian Ocean differed in weight by 100 grams. However, he did not indicate weights or relative size of the ovaries. Partlo (1955) conducted histological studies on 44 albacore from the eastern North Pacific, but did not publish data on gonad size or mention asymmetry. The gonads of other tunas are more or less symmetrical in size and usually have no conspicuous fat body (Kurt Schaefer, Inter-Am. Trop. Tuna Comm., P.O. Box 271, La Jolla, CA 92038, pers. comm., 1 May 1989). The finding that meiotic activity is higher in the smaller of a pair of testes is puzzling, and we have no ready explanation for it. It may be that this relation- ship holds true only for prespawning individuals and possibly during other periods when albacore are not spawning. We are unable to determine whether this condition continues through the spawning cycle be- cause our samples were collected over an insufficient period and not in the primary geographic region where South Pacific albacore are believed to spawn (Otsu and Hansen 1961). A remote possibility is that the smaller testis is kept in an advanced reproductive state for "opportunistic" spawning and that the larger testis matures only when the fish is in the more prevalent location of spawning. If this is so, the species could presumably extend its reproductive potential with relatively low expenditures of energy. It is also possi- ble that sperm are not shed by precocious fish, but are reabsorbed. Progressive reabsorption of sperm by Sertoli cells has been observed in rainbow trout (van den Hurk et al. 1978). Our results show that some males less than 70 cm may be sexually mature and that the proportion of sex- ually mature individuals increases with increasing fork length, up to 90 cm. There is a decrease in the percent- NOTES Ratty et al. Reproductive biology of Thunnus alalunga 215 age of sexually mature males in the size category above 90 cm FL which may be related to the method of cap- ture and/or the type of habitat where collected. All fish larger than 90 cm, except one, were caught during sub- surface longline fishing operations within about 75 miles of the coast of the New Zealand north island, whereas nearly all other fish were caught by surface trolling in midocean waters. Oceanographic observa- tions made concurrently with albacore troll-fishing op- erations indicate that albacore in the midocean region of the South Pacific are associated with Subtropical Convergence Waters (Laurs 1986, Laurs et al. 1987). Female albacore have been reported to attain sex- ual maturity and first spawn at about 90 cm FL (Ueya- nagi 1955, Otsu and Uchida 1959); however, size at maturity for males has not been well established. Otsu and Hansen (1961) found that some male albacore are probably mature when they attain a length of about 90 cm FL. Based on the general appearance of testes and oozing of milt, Ueyanagi (1957) postulated the smallest mature male to be 97 cm FL in a sample of albacore collected in the western North Pacific. Brock (1943) reported that "Milt could be squeezed from the testes of some of the more mature males, but no females were found in which eggs could be discerned by the unaided eye. ..." Brock (1943) examined fish caught off the Oregon coast that ranged in fork length between 53 and 90 cm, but did not specify the size of fish that contained milt. Partlo (1955), in the only published histological study that includes albacore testes as well as ovaries, con- cluded that both males and females in his age groups V and VI were in a condition approaching spawning. He calculated that the mean fork length for males and females in age-group V is approximately 70 cm, and for age-group VI is about 85 cm for males and 79 cm for females. He also reported observing small numbers of fully mature sperm restricted to tubules of the posterior portion of testes in age-group IV, which he reported as having a mean fork length of 62 cm. Partlo (1955) examined a total of 44 testes and ovaries, but unfortunately did not specify the sex or number of in- dividuals in each age group. Godsil and Byers (1944) reported that the fat body was present in albacore collected throughout the North Pacific, except in three specimens from the area around the Hawaiian Islands, with its greatest development always on the right side of the fish. However, they made no mention of asymmetry in gonad size. There appears to be a functional relationship between meiotic activity and the amount of fat in the mesentery. We presume the fat body provides an energy source for spermatogenesis, although we have no information from the present study on the timing of sperm release from the albacore testis. Godsil and Byers (1944) also speculated that the fat body acts as an energy reserve for female albacore that is expended during the spawn- ing season. They based their supposition on the absence of the fat body in three recently spawned albacore from the Hawaiian Islands. Schaefer (1987) noted that spawning in the black skipjack, Eiifhymrus UtieatHs, is probably regulated by energy available in fat storage. When the fat stores fall below a minimum level, ovarian atresia probably occurs over a short period of time. Conclusions We provide quantitative evidence that a substantial portion of male albacore caught in midocean areas of the South Pacific Subtropical Convergence Zone about lat. 35°-40°S, may be sex-ually mature when they attain a fork length of 71-80 cm. Also, the proportion of ma- ture fish increases with increasing length. A lower pro- portion of males caught by subsurface longline fishing in coastal waters off New Zealand showed development in sexual maturity, although not as accelerated as males caught by surface trolling in the midocean region. Fe- male fish 55-95 cm FL caught in both regions showed few signs of sexual maturation. South Pacific albacore are believed to spawn generally in the region of the Southern Tropical Convergence waters about lat. 20°-10°S (Knox 1970). This region is about 1000-1500 miles north of the Subtropical Convergence Zone where the albacore investigated in this study were caught. The adaptive significance of sexual maturity in male albacore at times, locations, and ages when females ai-e not in spawning condition is not clear. Sampling of both males and females over the entire spawning cycle and in locations where albacore are believed to spawn will be required to understand the reproductive biology of the South Pacific albacore population. Acknowledgments The authors wish to express their appreciation to Dar- lene Pickett and Christine Hopkins for their assistance in making histological preparations, to E. Roger Mar- chand for preparing photomicrographs, to Robert S. Garrett for producing scanning electron micrographs, and to Lana Nimmo for supplying technical help in image analyzer operations. In addition, we extend spe- cial thanks to Mark Hess for skillful computer applica- tions used in processing and analyzing results and to Kurt M. Schaefer for constructive comments on a draft of the manuscript. We also acknowledge Talbot Mur- ray for handling the logistics of collecting gonad sam- ples on the RV Kaharoa and Japanese longline fishing vessels, and the scientific party, crew and officers of the RV Townsend Cromwell. 216 Fishery Bulletin 1990 Citations Billard, R.. A. Fostier, C. Weil, and B. Breton 1982 Endocrine control of spermatogenesis in teleost fish. Can. ,). Fish. Aquat. Sci. 39:6S-7!). Brock, V.E. 1943 Contribution to the biology of the albacore {Germ ala- lunga) of the Oregon coast and other parts of the North Pacific. Stanford Ichthydl, Bull. (:2)(;: 199-248. Dalela, R.C., M. Rana. and S.R. Verma 1977 Sea.son histological changes in the gonads of two teleost fishes, Nolopterus nutoptcriis (Pallus) and Colisa fasciatus. Gegenbaurs Morphol. Jahrb. 132:128-140. Dotson. R.C. 1980 Fishing methods and equipment of the U.S. west coast albacore fleet. NOAA Tech. Memo TM-NMFS-SWFC-8, Southwest Fish. Cent., Natl. Mar. Fish. Serv., NOAA, La Jolla, Ca 92038, 126 p. Gabe, M. 1976 Hi.*>logical techniiiues [Transl. R.E. Blaclieith|. Springer- Verlag. NY. HOG p. Godsil, H.C.. and R.D, Byers 1944 A systematic study of the Pacific tunas. Calif. Dep. P^ish Game Fish. Bull. 50, 131 p. Grier, H.J.. J.R. Linton. J.F, Leatherland, and V,L. de Vlaming 1980 Structural evidence for two different testicular types in teleost fishes. Am. ,T. Anat. 1.59:331-34.5. Humason, G.L, 1979 Animal tissue techniques. W.H. Freeman, San Fran- cisco, 661 p. Kno-x, G,A, 1970 Biological oceanography of the South Pacific. /// Wooster, W,S, (ed.). Scientific exploration of the South Pacific, p. 155-182. Natl. Acad. .Sci., Wash., DC. Laurs, R.M. 198G U.S. albacore trolling e.xploi-ation conducted in the South Pacific during February-March, 1986. NOAA Tech. Memo. TM-NMFS-SWFC-66, Southwest Fish. Cent., Natl. Mar. Fish. Serv., NOAA, La .Jolla, CA 92038, 30 p. Laurs, R.M., K. Bliss, J. WeatheralL and B. Nishimoto 1987 South Pacific albacore fishery exploration conducted by U.S. jig boats during early 1987. Admin. Rep. LJ-87-22, Southwest Fish. Cent., Natl. Mar. Fish. Serv., NOAA, La .Jolla, CA 92038, 31 p. + append. 21 p. Nagahama. Y. 1983 The functional morphology of teleost gonads. In Hoar. W.S.. et al. (eds.). Fish physiology. Vol. IX A, p. 22.3-274. Acad. Press. NY. Otsu. T.. and R.J. Hansen 1961 Sexual maturity and spawning of the albacore in the Cen- tral South Pacific Ocean. Manuscr. to Pacific Tuna Biology Conf.. Honolulu, Aug. 1961. Honolulu Lab., Natl. Mar. Fish. Serv., NOAA, Honolulu, HI 96822-2396, 26 p. Otsu, T., and R. Uchida 1959 Sexual maturity and spawning of albacore in the P.acific Ocean. U.S. Fish Wildl. Serv. Fi.sh. Bull. ]48:287-.305. Partlo. J.M. 1955 Histological studies on albacore (Thunnux alalungii) gonads from the eastern Pacific. J. Fish. Res. Board Can. 12(l):61-67. Robertson, D.H. 1957 Accelerated development of testes after unilateral gona- dectomy, with observations of the normal testes of rainbow trout. U.S. Fish Wildl. Serv. Fish. Bull. 127:9-30. SanwaL R., and S.S. Khanna 1972 Seasonal changes in the testes of a freshwater fish Chnmia gachua. Acta Anat. 83:139-148. Schaefer, K.M. 1987 Reproductive biology of black skipjack, Eidhynnus linea- tus, an eastern Pacific tuna. Inter-Am. Trop. Tuna Comm. Bull. 19(2): 164-260. Turner, C,L. 1919 The seasonal cycle of spermary of the perch. .J. Mor- phol. 32:681-711. Ueyanagi. S. 1955 On the ripe ovary of the albacore, Crnmi ijtrmii (Lace- pede). taken from the Indian Ocean. Bull. .Ipn. Soc. Sci. Fish. 20(12):1050-1053. 1957 Spawning of the albacore in the Western Pacific. Reji. Nankai Reg. Fish. Res. Lab. 6:1I3-J24. van den Hurk, R., J. A. Vermeij, J. Stegenga, J. Peule, and P.G.W.J. van Oordt 1978 Cyclic changes in the testis and vas deferens of the rain- bow trout (Salnw gairdneri) with special reference to sites of steroidogenesis. Ann. Biol. Anim. Biochim. Biophys. 18: 899-904. Age Estimation and Composition of Peiagic Armorliead Pseudopentaceros wheeleri from the Hancoci< Seamounts James H. Uchiyama Jeffrey D. Sampaga Honolulu Laboratory, Southwest Fisheries Center National Marine Fisheries Service, NOAA 2570 Dole Street, Honolulu. Hawaii 96822-2396 The pelagic armorhead Pseudopen- taceros wheeleri is widely distrib- uted in the North Pacific, but rare- ly collected until fishermen from the Soviet Union discovered large con- centrations on the central North Pacific seamounts in the late 1960s (Sakiura 1972). The Soviet Union started a trawl fishery at the south- ern Emperor-northern Hawaiian Ridge (SE-NHR) seamounts (29°N, 179°E to 35°N, 171°E) in 1967, and the Japanese entered the fishery in 1969 (Sasaki 1986). Since then,' in- tensive fishing occurred at the Northwest and Southeast Hancock Seamounts, the southeast limit of the pelagic armorhead seamounts. During the 1970s, the pelagic ar- morhead catch rate by Japanese fishing vessels at the Hancock Sea- mounts declined dramatically (Uchi- da and Tagami 1984). Commercial fishing at the Hancock Seamounts ceased in 1984, and finally, a 6-year fishing moratorium went into effect in August 1986 (NMFS 1986). Pelagic armorhead have an un- usual life history. At SE-NHR sea- mounts, adults spawn during De- cember-March, as indicated by ova- rian studies (Sasaki 1974, BiJim et al. 1978) and the occurrence of lar- vae and young juveniles in Febru- ary and March (Borets 1980). As juveniles, they apparently are pela- gic and occur to the north and northeast of the seamounts as far as the Alaskan gyre before return- ing to the seamounts (Fujii 1986, Boehlert and Sasaki 1988). At the SE-NHR seamounts, most of the pelagic armorhead caught in the fishery are 26-33 cm fork length (PL), although 15-40 cm PL speci- mens occur (Humphreys and Taga- mi 1986). Larger specimens (>40 cm PL), but none 26-33 cm PL, have been caught on the northern Hawaiian Ridge, from Kure Atoll to French Frigate Shoals (Tagami and Humphreys, in prep.). Pelagic armorhead at the sea- mounts exhibit three morphological types: fat, intermediate, and lean (Humphreys et al. 1984, 1989; Hum- phreys and Tagami 1986). During their pelagic phase and when they recruit to seamounts, they are "fat" and their gonads are in an early stage of development. When gonads begin developing and spawning oc- curs, they transform from an "inter- mediate" to a "lean" stage. It has not been determined whether the lean fish recover to spawn again. In this study, we determined which hard part is the most suitable for estimating armorhead age and determined the temporal meaning of perceived daily growth incre- ments and apparent annual check marks on the sagitta, the preferred hard part. Chikuni (1970) and Vasil'kov and Borets (1975) used annuli on scales to estimate age, but their estimates did not agree. Fur- thermore, neither study used vali- dation procedures, nor did they document the area of collection. The criteria we developed were used to estimate the age composition of pelagic armorhead captured on Southeast Hancock Seamount. Methods and materials Specimen collection Pelagic armorhead were obtained at the Hancock Seamounts in 1976-85 from the Japanese trawler Kifa- kami Maru, and by bottom trawl and handline from the NOAA ship Townsend Cromwell. Random sam- ples were obtained by pooling speci- mens from bucket scoops (range 2-100 kg/haul) on the Kitakami Maru, during 1 August-15 October 1981, and from a single haul on the Townsend Cromwell on 24 July 1984. Fish were measured for PL to the nearest millimeter and sexed by gonadal examination. Fish were classified, on the basis of color and body depth, to one of three body types: fat, intermediate, or lean (Humphreys et al. 1984). We also in- cluded a few very large specimens caught by handline on the northern Hawaiian Ridge seamounts between Kure Atoll and French Frigate Shoals and postlarvae caught in a surface tow with a Tucker trawl at Southeast Hancock Seamount, 23- 24 February 1985. All postlarva! specimens were preserved and stored in 70% ethanol. Hard-part selection and check-mark counts To determine which hard part was most suitable for estimating age, 13 structures (Table 1) were dissected from 5 specimens collected at the Hancock Seamounts. The stractui'es were prepared for examination fol- lowing modified procedures of Six and Horton (1977). The hy]oural Manuscript accepted 25 Julv 1989. Fishery Bulletin, U.S. 88:217-222. 217 218 Fishery Bulletin 88|l). 1990 Table 1 Comparison of check marks on various hard parts of five Pseudopentaceros | wheeleri from the Hancock Seamounts. Hard part No. check marks by fish identification number 343 350 338 385 383 Sagitta 3 3 — 3 2 Centrum (vertebra) 3 3 3 3 2 Dorsal spine 3 3 3 3 2 Dorsal ray 0 0 0 0 0 Anal spine 3 0 3 3 9 Anal ray 0 0 0 0 Pelvic spine 0 0 3 3 Pelvic ray 0 0 0 (1 Pectoral spine 0 0 0 0 Pectoral ray 0 0 0 0 Hypural 3 3 3 3 2 Maxillary 0 0 3 3 2 Scales n — — — 0 Figure 1 Sagitta oi Pseudopentdcerun iv)ifiUri with annul! marked and numbered. plate, maxillary, and vertebrae were cleaned of tissue and dried. Spines and rays of the dorsal, pelvic, pec- toral, and anal fins were cleaned of tissue, embedded in resin, cross-sectioned in a continuous series from the base, and examined under a compound microscope. Both surfaces of a section were examined. Tissues were also examined under a dissecting: microscope with reflected and transmitted light. Saj^ttae were sanded lightly on the surfaces with no. 400 ffc'\i carborundum sandpaper, then, with a dissect- ing microscope, were examined whole against a black background while submerged in water. Check marks appeared as pairs of bright (opaque) and dark (translu- cent) concentric bands under reflected light (Fig. 1 ). Scales were taken from nine different areas of the body. Three sites were along the lateral line: below the NOTES Uchiyama and Sampaga: Age estimation and composition of Pseudopentsceros wheelen 219 first dorsal spine, midway along the lateral line, and below the posterior end of the dorsal fin. Three sites were midway between the lateral line sites and the dor- sal fin, and three were midway between the lateral line sites and the ventral border. Scales were taken from 20 fish of three different body types and from 13 fish, caught in February 1985, with three check marks on their sagittae. After being cleaned, scales were mounted in Euparal under a cover glass on a glass slide or flattened between two glass slides after being air dried. Selection of the hard part for ageing was based on the consistency of perceived annual check mark counts and ease of preparation (Table 1). Only the sagittae, vertebral centrum, hypural plate, and dorsal and anal fin spines showed check marks. Counts of check marks from each part were identical (Table 1), indicating these structures could be used for estimating age if their check marks could be verified as annuli. The number of check marks on vertebral centrum, hypural plate, and ttiaxillary was reduced by half upon drying. Soft rays of dorsal, pelvic, pectoral, and anal fins showed no marks that could be interpreted as check marks. Scales were without check marks, and their circuli ap- peared evenly spaced. Of all the structures compared, the sagitta was considered the best structure for estimating the age of pelagic armorhead. Counts of daily growth increments Sagittae removed at sea were stored in fresh water in vials and refrigerated; those not extracted at sea were frozen within the head and later removed in the labor- atory. Sagittae were sanded lightly on both sides with no. 400 carborundum sandpaper until translucent and etched in a very dilute solution of HCl for 3-5 minutes. The acid was diluted until only 1-2 bubbles per second were formed in reaction to the calcium carbonate of the sagitta. This process was repeated until the incre- ments became visible from the core to the tip of the postrostrum. The sagittae of postlarval specimens were teased from the skull, cleaned, and mounted in Euparal without further processing. To support the coverslip, each sagitta was mounted on a microscope slide in Euparal by using short segments of monofilament line slightly greater in diameter than the thickness of the otolith. Sagittae were examined under a compound microscope at 300 x magnification. Increment counts often took a circuitous route be- tween the core and the tip of the postrostrum; the mean of 10 counts was used as the age estimate. Increment counts were routine for sagittae with two check marks Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. but were more difficult for those with three check marks. Increment width in translucent zones decreased with increasing number of check marks (8.3 ytva dur- ing year 1, 1.7 yim during year 2, and 1.5 \xm during year 3), and incremental pattern was more variable after year 2. Age estimation Check marks couhl ndt be validated as annuli by con- ventional means (Beamish and McFarlane 1983). Mean increment counts were used to estimate the hatching time, assuming the first increment formed at or soon after hatching. This technique, though low in accuracy, may suffice when direct validation is unattainable (Gjt^saeter et al. 1984). To verify check marks as an- nuli, mean counts of growth increments for fish caught in February were then compared with counts of check marks, which were multiplied by 365. Age composition of catches of pelagic armorhead at Southeast Hancock Seamount was based on check- mark counts (i.e., perceived annuli). Mean FLs between age-sex groups were compared using the Mann- Wliitney paired test, since unequal variances occurred in one sample. To increase sample size of age-sex groups from the Kitakami Maru catch, randomly sampled specimens from Northwest and Southeast Hancock Seamounts were combined. Results and discussion Increment counts from demersal and oceanic adults and juvenile pelagic armorhead suggested that incre- ments were deposited daily (Tables 2 and 3). A total of 88% of the estimated hatching dates of adults oc- curred within the spawning period, and the other three hatching dates were reasonably close to the spawning period. Check marks appeared to be valid annuli based on number of daily growth increments for fish caught in February (Table 2). Sagittae with two check marks had approximately 730 increments (range 728-773, A^ = 5), and those with three check marks had about 1095 in- crements (range 1037-1128, A/^ = 4). Check marks were also verified as annuli by a second but less precise method (Pannella 1971). Partial counts of 365 apparent daily growth increments appeared to coincide closely with the first pair of opaque and translucent zones of sagittae of age-1 and -2 fish. Partial counts of about 730 apparent daily growth increments corresponded with the outer margin of the second translucent zone on sagittae of age-2 fish. Fish caught in summer appeared to have completed check marks on their sagittae. Sagittae from fish caught in July with two check marks had increment 220 Fishery Bulletin 88(1). 1990 Table 2 Age estimates of pelagic arnwrhead Faeudopei tnceros wheeleri by counting apparent daily growth increments (GI) on sagittae. Fork length Mean GI Estimated No. Date captured (mm) Sex count SD hatching date check marks Body type 29 Oct. 1976 277» M 623 25.63 8 Feb. 1975 2 — 18 June 1983 340 F 902 13.04 29 Dec. 1980 3 Fat 12 July 1984 300 F 576 24.03 13 Dec. 1982 2 Lean 12 July 1984 322 F 621 21.67 30 Oct. 1982 2 Intermediate 13 July 1984 305 M 549 43.83 10 Jan. 1983 2 Intermediate 13 July 1984 308 F 937 58.00 18 Dec. 1981 3 Lean 24 July 1984 309 M 907 94.29 28 Jan. 1982 3 Fat 24 July 1984 310 M 544 37.28 26 Jan. 1983 2 Fat 24 July 1984 322 M 923 92.46 13 Jan. 1982 3 Fat 24 July 1984 329 F 605 37.94 26 Nov. 1982 2 Fat 25 July 1984 307 F 586 37.16 16 Dec. 1983 2 Intermediate 25 July 1984 319 F 914 39.99 22 Jan. 1982 3 Intermediate 29 July 1984 303 M 853 51.39 28 Mar. 1982 3 Intermediate 16 Feb. 1985 282 M 728 46.10 18 Feb. 1983 2 Lean 16 Feb. 1985 290 M 1037 26.80 15 Apr. 1982 3 Lean 16 Feb. 1985 293 M 762 50.96 15 Jan. 1983 2 Fat 16 Feb. 1985 304 F 758 40.53 19 Jan. 1983 2 Lean 16 Feb. 1985 311 F 761 45.73 16 Jan. 1983 2 Lean 16 Feb. 1985 317 M 773 27.29 4 Jan. 1983 2 Lean 16 Feb. 1985 319 F 1128 79.66 14 Jan. 1982 3 Lean 16 Feb. 1985 320 F 1101 84.65 10 Feb. 1982 3 Lean 23 Feb. 1985 355 F 1065 51.32 25 Mar 1982 3 Fat 25 July 1984 322" M 951 69.04 16 Dec. 1981 3 Fat 25 July 1984 305'^ _*Mi»M Female age 1 / \ Male age 2 /^ A \ O40 UJ / i\ \ Female age 2 / i - \ / / \\ 3 o r^ 1 M UJ IE 30 u. 20- // ^' 'A 10- 0 ^ .«««»»^* . ..-^-^ — "-..^ '*"■ ■'^* jM;«^«e»^i«»WjSl:^--"l ■— ■!-— 1 1 -f -\'-\- 1 23 24 25 26 27 28 29 30 31 32 33 34 FORK LENGTH (cml Figure 2 Length-frequency distribution of Pseiidopei^taceros wheeleri sampled at the Hancock Seamounts, 1 August-15 October 1981. LEGEND 45 All 40 Male age 1 /\ 35 ..»..„. Female age i / \ Male age 2 / \ > 30- U z Female age 2 / \ UJ 3 25 a UJ cc "- 20 h\ \ 15- / A\ 10- 5- ^ ^ 23 24 JS 26 27 28 29 30 31 32 33 34 FORK LENGTH (cml Figure 3 Length-frequency distribution oi Pseudopentaceros wheeleri sampled at the Southeast Hancock Seamount, 24 July 1984. We could not determine when the opaque zone formed because specimens were not obtainable throughout the year and the edge always appeared translucent. Pelagic armorhead from the summer trawl fishery at the Southeast Hancock Seamount consisted of fish approximately ages 1.5 and 2.5, and yearlings were dominant in both 1981 (96%) and 1984 (66%) (Table 4). Yearling females appeared to be larger than males in both 1981 and 1984 samples (Figs. 2 and 3; Table 4). Differences between male and female mean lengths were significant in 1981 (P<0.001) and in 1984 (P = 0.025), but difference in mean lengths for year-2 females and males in 1984 was not significant (P = 0.107). The latter could have been due to small sample size; hence, larger samples should be examined before requiring separate growth curves for males and females. Our samples were not perfectly suited for group comparison because sex and age were unknown when the random samples were taken. Thus, unequal group size and unequal variances occurred. Due to our small sample size, slight skewness and kurtosis were encountered in comparing central tendencies of some age-sex groups. However, only the unequal variance between groups was serious enough to have affected the analyses. Pelagic armorhead may attain sexual matiuity before they reach the seamount. We base this on the age- length relationships of the three oceanic specimens (Table 2). After pelagic armorhead settle on the sea- mount, it appears that little growth occurs. Length- frequency distribution reported by other investigators (Sasaki 1986, Wetherall and Yong 1986) tends to con- firm this observation. The lack of older (>age 3) indi- viduals on the seamounts suggests that pelagic armor- head may spawn only one or two seasons and then die. Indeed, the presence of emaciated individuals suggests 222 Fishery Bulletin 88(1 1990 some postspawning mortality (Humphreys et al. 1984, Humphreys and Tagami 1986). Conversely, the exis- tence of larger and older specimens in the north- western Hawaiian Islands would suggest that longer life is possible and different life histories occur in dif- ferent environments. Of the sagittae from large pelagic armorhead examined for comparison purposes, fish 43-48 cm FL {N = 3) had four check marks, fish 48-54 cm FL (A^ = 2) had five check marks, and the largest fish, 54.7 cm FL, had eight-plus check marks (Fig. 1). Although the relationship of these large pelagic ar- morhead to those on the seamounts is unknown, it is plausible that during the pelagic stage juveniles were displaced to the east, entered the subtropical gyre, and grew larger while spending 1-3 years longer in the open ocean before settling in the northern Hawaiian Ridge seamounts, southeast of the SE-NHR sea- mounts (Boehlert and Sasaki 1988). Acknowledgments We thank G. Kobayashi of the Oshoru Maru for col- lecting the oceanic pelagic armorhead specimens. B. Mundy verified the postlarval specimens. We also thank G.W. Boehlert, G.M. Cailliet, R.D. Methot, D.A. Somerton, and especially the anonymous reviewers for critically reviewing this manuscript. Citations Beamish, R.J., and G.A. McFarlane 1983 The forgotten requirement for age validation in fisheries biology. Trans. Am. Fish. Soc. 112:735-743. Bilim, L.A., L.A. Borets, and L.K. Platoshina 1978 Characteristic of ovogenesis and spawning of the boar- fish in the region of the Hawaiian Islands. Izv. Tikhookean. Nauchno-Issled. Inst. Rybn. Khloz. Okeanogr. (TINRO) 102: .51-.57 [In Russ., Engl, abstr.] (Engl, transl. by W.G. Van Campen, 1986, Transl. 106. 9 p.; avail. Honolulu Lab., Natl. Mar, Fish. Serv., NOAA, Honolulu, HI 96822-2396.) Boehlert, G.W., and T. Sasaki 1 988 Pelagic biogeography of the armorhead, Pseudopentaceros trhfelcri., and recruitment to isolated seamounts in the North Pacific Ocean. Fish, Bull., U.S. 86:4.53-46.5, Borets, L.A. 1980 The distribution and structure of the range of the boar- fish PcnUircrox richardsoni. ,]. Ichthyol. 20(3):141-143. Chikuni, S. 1970 The "phantom fish," "Kusakari tsubodai"— An outline. Enyo (Far Seas) Fish. Res. Lab. News 3:1-4, Feb, 1970. (Engl, transl. by J.H. Shohara, Natl. Mar. Fish. Serv,, NOAA, Ter- minal Island. CA 90731.) Fujii. E. 1986 Zoogeographical features of fishes in the vicinity of sea- moimts. In Uchida, R.N„ S. Hayasi. and G.W, Boehlert (eds.). Environment and resources of seamounts in the North Pacific, p. 67-69. NOAA Tech. Rep, NMFS 43, Natl. Oceanic Atmos. A^rSc» ^*< U.S. Department of Commerce Robert Mosbacher Secretary National Oceanic and Atmospheric Administration John A. Knauss Under Secretary for Oceans and Atmosphere National Marine Fisheries Service William W. Fox, Jr. Assistant Administrator for Fisheries ,v>^'*^°'c-o. IFusIa}(SI7^ n ©Gam Scientific Editor Dr. Linda L. Jones National Marine Mammal Laboratory National Marine Fisheries Service, NOAA 7600 Sand Point Way NE Seattle, Washington 98! 15-0070 Editorial Committee Dr. Andrew E. Dizon National Marine Fisheries Service Dr. Charles W. Fowler National Marine Fisheries Service Dr. Richard D. Methot National Marine Fisheries Service Dr. Theodore W. Pietsch University of Washington Dr. Tim D. Smith National Marine Fisheries Service Dr. Mia J. Tegner Scripps Institution of Oceanography The Fishery Bulletin (ISSN 0090-0656) is published quarterly by the Scientific Publications Office, National Marine Fisheries Service, NOAA, 7600 Sand Point Way NE, BIN CI 5700, Seattle, WA 98115-0070. Second class postage is paid in Seattle, Wash., and additional offices. POSTMASTER send address changes for subscriptions to Fishery Bulletin, Super- intendent of Documents, U.S. Govern- ment Printing Office, Washington, DC 20402. Although the contents have not been copyrighted and may be reprinted entire- ly, reference to source is appreciated. The Secretary of Commerce has deter- mined that the publication of this period- ical is necessary in the transaction of the public business required by law of this Department. Use of funds for printing of this periodical has been approved by the Director of the Office of Management and Budget. For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. Subscrip- tion price per year; $16.00 domestic and $20.00 foreign. Cost per single issue: $9.00 domestic and $11.25 foreign. Managing Editor Nancy Peacock National Marine Fisheries Service Scientific Publications Office 7600 Sand Point Way NE. BIN CI 5700 Seattle, Washington 981 15-0070 The Fishery Bulletin carries original research reports and technical notes on investiga- tions in fishery science, engineering, and economics. The Bulletin of the United States Fish Commission was begun in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. Begin- ning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. It is also available free in limited numbers to libraries, research institutions, State and Federal agencies, and in exchange for other scientific publications. U.S. Department of Commerce Seattle, Washington Volume 88 IMumber 2, 1990 Fishery Bulletin Contents 223 Matthews, Kathleen R. A compa;ative study of habitat use by young-of-the-year, subadult, and adult rockfishes on four habitat types in central Puget Sound 241 Nizinski, Martha S., Bruce B. Collette, and Betsy B. Washington Separation of two species of sand lances, Ammodytes amencanus and A. dubius. in the western North Atlantic 257 Kope, Robert G., and Louis W. Botsford Determination of factors affecting recruitment of Chinook salmon Oncorhynchus tshawytscha in central California 271 Selpt, Irene E., Phillip J. Clapham, Charles A. Mayo, and Mary P. Hawvermale Population characteristics of individually identified fin whales Balaenoptera physalus in Massachusetts Bay 279 Shields, Jeffrey D., and Armand M. Kuris Carcinonemertes wickhami n, sp. (Nemertea), a symbiotic egg predator from the spiny lobster Panulirus interruptus in southern California, with remarks on symbiont-host adaptations 289 Cockcroft, Victor Gavin, and Graham James Berry Ross Age, growth, and reproduction of bottlenose dolphins Tursiops truncatus from the east coast of southern Africa 303 Kim, Suam, and Bohyun Bang Oceanic dispersion of larval fish and its implication for mortality estimates: Case study of walleye pollock larvae in Shelikof Strait, Alaska 313 Hampton, John, and Geoffrey P. Kirkwood Tag shedding by southern bluefin tuna Thunnus maccoyii Fishery Bulletin 88(2), 1990 323 Shelton, Peter A. and Larry Hutchings Ocean stability and anchovy spawning in the southern Benguela Current region 339 Silber, Gregory K. Occurrence and distribution of the vaquita Phocoens sinus in the northern Gulf of California 347 Waring, Gorton T., Patricia Gerrior, P. Michael Payne, Betsy L. Parry, and Jolin R. iViicolas Incidental take of marine mammals in foreign fishery activities off the northeast United States 361 Sturm, Maxwell G. de L. Age, growth, and reproduction of the king mackerel Scomberomorus cavalla (Cuvier) in Trinidad waters 371 Smith, Susan E., and Norman J. Abramson Leopard shark Tnakis semifssciata distribution, mortality rate, yield, and stock replenishment estimates based on a tagging study in San Francisco Bay Motes 383 Iversen, Edwin S., Scott P. Bannerot, and Darryl E. Jory Evidence of survival value related to burying behavior in queen conch Strombus gigas 389 Eggleston, David B., and Eleanor A. Bochenek Stomach contents and parasite infestation of school bluefin tuna Thunnus thynnus collected from the Middle Atlantic Bight, Virginia 397 Holland, Kim, Richard Brill, and Randolph K.C. Chang Horizontal and vertical movements of Pacific blue marlin captured and released using sportfishing gear 403 Ojeda, F. Patricio, and John H. Dearborn Diversity, abundance, and spatial distribution of fishes and crustaceans in the rocky subtidal zone of the Gulf of Maine 411 Stergiou, Konstantinos I. A seasonal autoregressive model of the anchovy Engraulis encrasicolus fishery in the eastern Mediterran an 415 Barshaw, Diana E., and Kenneth W. Able Tethering as a technique for assessing predation rates in different habitats: An evaluation usingjuvenile lobsters Homarus amencanus 418 List of recent NOAA Technical Reports Abstract.— Seasonal habitat use of young-of-the-year, subadult, and adult rockfishes (Sebastes caurinus, S. maliger, andS. auriculatug) were compared for four habitat types: high- relief rocky reefs, low-relief rocky reefs, high-relief artificial reefs, and sand/eelgrass. Diving surveys con- ducted December 1986 through Octo- ber 1988 on two representative sites of each habitat ty]3e revealed signifi- cant seasonal changes in rockfish densities and habitat use. Young-of- the-year (YOY) recruitment varied between the two survey years: YOY were observed on all haliitat types in the summer and fall of 1987, where- as they were observed at only one site (artificial reef) in a similar time period of 1988. High-relief rocky reefs had the most consistent den- sities of the three rockfish species, mostly fish > 200 mm TL. Adult and YOY copper, quillback, and brown rockfishes were observed on the low- relief rocky reefs primarily in the summer months coincident with summer algal growth; when the kelp died back in the fall, most rockfishes left these reefs. The highest densities of rockfishes, primarily 80-200 mm quillback rockfish (up "to 420/90-nr'' transect) and large copper rockfish (up to 56.3/transect), were observed on the artificial reefs. Here, also, density fluctuations were dramatic; copper rockfish densities peaked in fall and winter and declined (to 0/transect) during the summer, and quillback rockfish densities also seasonally fluctuated. Sand/eelgrass areas were the least-utilized habitat type; only during July and August wei , young-of-the-year and low den- sities of adult copper and brown rockfishes observed on one sand/eel- grass site. Although all four habitats were used, natural reefs may repre- sent source habitats that are used by and maintain rockfishes on less pro- ductive sink (artificial reef) habitats. Thus, the recent use of artificial reefs as mitigation for the loss of natural reefs could have negative impacts on rockfish populations. A Comparative Study of Habitat Use by Young-of-tFie-year, Subadult, and Adult Rockfishes on Four Habitat Types in Central Puget Sound* Kathleen R. Matthews Washington Department of Fisheries, 7600 Sand Point Way NE Seattle. Washington 981 15-0070 School of Fisheries WH-10, University of \A a^ytgn^JetogtCoWashington 98195 Present address: Pacific Biological Station, I )epartment ^(gif^pes and Oceans Nanaimo, British Columbia, Canada V9li 5K6 SEP 12 mo Rockfishes are a successful group of marine fishes represented by about 100 species (Barsukov 1981), which occupy a variety of habitats ranging from the intertidal to the edge of the continental shelf (Lea 1983). This speciose and interesting group dis- plays extremes in habitat use and movement patterns: Some species live deep (500 m) over sandy bottoms, some make long-distance movements (to 550 km), whereas others are pe- rennially sedentary on shallow (<10 m) rocky reefs (Larson 1980, Love 1980, Culver 1987). One group of rockfishes is solitary and demersal and inhabits shallow reefs (Ebeling et al. 1980a); in Puget Sound this group of rockfishes is represented by cop- per Sebastes caurinus, quillback S. maliger, and brown S. auriculatus. Copper, quillback, and brown rock- fishes are common shallow-water benthic species found along the Pacif- ic coast of North America (Hart 1973). They are morphologically sim- ilar to one another but differ in color- ation. The three species have overlap- ping geographic ranges; copper and brown rockfishes are found from Baja California to Alaska, whereas quillback are found from central Cali- fornia to Alaska (Hart 1973). All three species occur in Puget Sound, Manuscript accepted 1 December 1989. Fishery Bulletin, U.S. 88:223-239. 'Contribution No, 804 of the School of Fish- eries, University of Washington. Washington, and comprise approx- imlflfSp^i^J-tePitheipecBeationkl bot- romllah thatch in central Puget Sound (Palsson 1988). These species inhabit a variety of habitat types although highest densities are reported on ar- tificial reefs, natural rocky reefs, and rock piles in water less than 30 m (Moulton 1977, Buckley and Hueckel 1985). Usually these species are found directly on the bottom closely associated with rock, artificial sub- strate, or vegetation (Patten 1973, Moulton 1977). Although copper and quillback rockfishes are common throughout Puget Sound, the San Juan Islands, and Strait of Juan de Fuca, brown rockfish are more limited in distribution and are found only within central and south Puget Sound (Moulton 1977). Rockfishes are viviparous (Boehlert and Yoklavich 1984), undergoing in- ternal fertilization and subsequently releasing pelagic larvae. Little is known regarding the length of time that young rockfishes are pelagic before they recruit to adult or tran- sitional habitat. Parturition reported- ly occurs April through June for the three species in Puget Soimd (DeLacy et al. 1964, Washington et al. 1979, Dygert 1986). An influx of postlarval or yoimg-of-the-year (YOY) rockfishes into isolated areas in Puget Sound has been observed (Patten 1973, Gowan 1983); assessment of habitat 223 224 Fishery Bulletin 88(2), 1990 SE 1 • vRedondo Figure I Map of study sites within central Puget Sound. HRl Orchard Rocl;tes pnucispinus, S. pinniger, and S. minlatns), that recruit to kelp beds in central California. Because rockfish produce pelagically dispersed larvae and juve- niles, local adult density is probably a poor predictor of local recruitment and future adult density. Rather, the limited availability of habitat with which young-of-the- year and juvenile associate may be the critical determi- nant of local recruitment. Determining the habitat re- quirements of young rockfishes can be crucial to our understanding of local rockfish dynamics. If algae, sea- grasses, or other habitats provide YOY rockfishes essential shelter from predation and increased access to food, then recruitment could be suppressed by elimi- nating such shelter or enlianced by adding more shelter. Habitat affinities and requirements are an unknown for YOY and juvenile rockfishes in Puget Sound. Copper, quillback, and brown rockfishes inhabit a variety of habitats, yet the relative importance of each habitat and seasonal uses are unknown. Additionally, differences in habitat use by different life-history stages are poorly understood. For example, do young rockfishes settle on reefs and remain there for life, or do the different age groups (young-of-the-year, sub- adult, and adult) later move to separate or different habitats? Understanding how and when the different habitats are exploited provides needed information on ecological requirements for habitat management; rec- ognition and protection of important habitats are necessary for effective fisheries management. Similar- ly, further descriptions of the different habitats are required to adequately assess the relative importance of habitat features to the different life-history stages of rockfishes. This study was designed to provide a quantitative comparison of habitat use for copper, quill- back, and brown rockfishes on four habitat types com- mon in Puget Sound. Materials and methods Study sites For each of four habitat types— high-relief rocky reef, low-relief rocky reef, high-relief artificial reef, and sand/eelgrass— two representative replicate study sites were selected. The two high-relief natural rocky reefs were characterized by steep vertical relief to 5 m off the bottom, surface canopies of the annual bull kelp Nereocystis leutkeana May through November, and understories of the perennial kelps Agarum fimbria- fun) and Pferygophora ralifnniicn. The high-relief rocky reefs range in depth from 12-20 m (gauged from mean lower low water). The Orchard Rocks high-relief rocky reef (HRl) is located on the southeastern side of Bainbridge Island (Fig. 1) approximately 600 m off- shore in the middle of Rich Passage, which is swept by high currents up to 8.1 km/hour (4.5 knots) (U.S. Dep. Commer. 1987, current tables). The entire Or- chard Rocks reef covers approximately 5 ha. The portion of reef used as the study site was about 12-18 m in depth and consisted of large boulders and rocky ledges that rise up to 5 m off the bottom. The second high-relief reef (HR2), Blakely Point, is a series of rock outcroppings separated by sand and shell gravel. The outcroppings are oriented perpendicularly offshore from the north entrance of Port Blakely Harbor on the eastern side of Bainbridge Island. Transects extended across the reef from 12-18 m in depth; each transect covered a similar depth range. Similar to HRl , the HR2 reef consists of steep walls and crevices with vertical relief of 5 m, but has minimal current. Matthews: Habitat use by rockfishes in central Puget Sound 225 1 2 3 4 finolysis with 0 Ho=No ditterenc hobiTo Mo= No a.ttcer.c sues Habitat Type High relief rocky reef 1 1 1 Si'e 1 -Orchofd RocKS Ti T2 T3 Sue 2 -Point Blakely Tl T2 T3 Low rel e( rocky reef 1 1 1 S"e 1 -Boinbndge Islona -1 T2 '3 Site 2-Bloke Island Tl T2 t; H.qh relie arlif 1 cial reef 1 1 Sife 1 -Blake islond It T2 T3 Si'e ?-Boeinq Creek Tl J 2 T3 Sand/eelgross 1 1 S.te 1 - Reaondc T 1 T ^' T 3 Site 2 -Poft Modison Tl T2 TJ les'ed ANOv/A to test e between meari Oersr'ie types e Detweer meon densiiie vithifi a habito' type of OOliI s o( odii , subadui' and vOi rockftsties on di"ereni t.suDodutt ond fOl rochdshes on lepiico'e Figure 2 Nested experimental design showing habitat types in central Puget Sound, sites within habitat types, and the three transects per site. The two low-relief rocky reefs were characterized by flat, featureless cobble and rock bottom with a few isolated areas of vertical relief (1-2 m). Transects ex- tended from 8-10 m in depth, but not deeper than 10 m where the reef drops off into sand/cobble substrate. The low-relief reefs were more uniform in depth when compared with the high-relief reefs. Both low-relief reefs have dense canopies of Nereocystis leutkeana May through November and understories of perennial kelps Agarum fimhriatum and Pterygophora californica. Low-relief habitats underwent dramatic changes in the fall and winter when they were virtually devoid of all fish and algal structure. Bainbridge Island low-relief rocky reef (LRl) is 500 m inshore from HRl, and the two reefs are separated by cobble and sand. Blake Island low-relief reef (LR2) is located 5 km south of Bainbridge Island. Both low-relifef reefs are swept by high currents up to 8.1 km/hour (4.5 knots). The two high-relief artificial reefs were character- ized by vertical relief of up to 5 m and no surface canopies of bull kelp, but with isolated patches of under- story kelps. Transects extended from 15-20 m in depth. The Blake Island artificial reef (ARl) was constructed in 1980 (Lauile 1982) and is located 5 km south of Bain- bridge Island on the southwestern side of Blake Island approximately 500 m offshore of LRl. The reef con- sists of concrete rubble, slabs, rectangular boxes, and tires forming vertical relief up to 4 m. The Boeing- Creek artificial reef (AR2) was constructed in late 1982 and is located 8 km north of Shilshole marina approx- imately 600 m offshore. The reef consists of concrete rubble with vertical relief of up to 6 m. The two reefs differ, as Blake Island is swept by strong currents up to 8.1 km/hour (4.5 knots) while Boeing Creek ex- periences little current. The sand/eelgrass areas were characterized by flat, shallow, unconsolidated substrate with dense growth of eelgrass Zostera marina May through November. Eelgrass typically grows in low-current, sheltered water and is restricted to depths less than 6 m (Phillips 1984), so the transects were placed at 5 m. The Redon- do (SEl) sand/eelgrass area is located approximately 500 m north of Saltwater State Park, 24 km south of Seattle. The Port Madison eelgrass area (SE2) is located 500 m north of the Suquamish boat ramp, off the northern side of Bainbridge Island. Survey methods From December 1986 through February 1988, I con- ducted monthly SCUBA surveys and estimated rockfish densities along three (30-m long, 3 m wide, and 1 m high) permanent transect lines at each of the eight reefs (Fig. 2). After the February 1988 survey, surveys were conducted in April, June, July, and August of 1988. In October 1988, surveys quantifying only YOY rockfishes were conducted on four reefs where YOY had been observed in the summer and fall of 1987: HR2, LRl, AR2, and SE2. All eight sites were sam- pled each month, and all three transects were sampled on the same day. I swam along the transect line and recorded individuals of each species as YOY (<80 mm TL). subadult (80-200 mm TL), and adult (>200 mm TL). During the summer and early fall when kelp cover was the highest, I also swam through the canopy to search for YOY rockfish. The ability to designate these size categories was verified by periodically capturing and measuring fish. Because YOY copper, quillback, and brown rockfishes could not be identified to species underwater, they were combined into one group. A total of 456 transects was completed on all reefs: 24 transects conducted monthly for 15 consecutive months December 1986-February 1988, and 24 transects con- ducted monthly during April, June, July, and August 1988. Fifteen additional YOY surveys were completed in October 1988. Other information collected along the transect lines included estimates of vegetative cover and water tem- perature. Kelp and eelgrass cover was qualitatively assessed by estimating, at the end of each transect, the percent cover of bullkelp, understory kelps, and the height of eelgrass present. The percent cover ranged from 0 (no vegetation present) to 100% (transect line completely covered by vegetation). Temperatures were recorded along the transect lines on each survey with a submersible thermometer. 226 Fishery Bulletin 88(2). 1990 Survey analyses From the three transects at each site, a mean number ( + 95% CI) of fish per transect was computed for each species and size category. To determine whether den- sities of rockfishes differed among the four habitat types and within the two replicate sites within each habitat type, I used a nested analysis of variance (ANOVA) (Zar 1984) (Fig. 2) that tested two null hypotheses: (1) There is no difference in rockfish densities between habitats (F„o.5( us (habitat types), 4 (.sites)), and (2) there is no difference in rockfish densities be- tween the two replicate sites within each habitat type (F().05(i)4(sites).i6(error))- The degTces of freedom for error was calculated as (4 (types) • 2 (sites) ■ 3 (transects) - 4 ■ 2) = 16 (Zar 1984, p. 147). To reduce the number of analyses, data were analyzed as a separate ANOVA for each of six seasonal periods using the middle month: Winter 1 = December 1986-February 1987; spring = March 1987-May 1987; summer 1 = June 1987-August 1987; fall = September 1987-November 1987; winter 2 = December 1987-February 1988; and summer 2 = June 1988- August 1988. Although density estimates were similar within quarters (Figs. 3-6), the sedentary behavior of rockfish could mean that the same indivi- duals were counted in February that had previously been counted in January, which would constitute a repeated sample. Thus to eliminate the possibility of non-independence of data, as I have no evidence that rockfish redistribute themselves randomly from month to month, I used only the middle month of each quarter. Because many months and sites contained zero den- sities and, in some cases, the variances were propor- tional to the means, the data were transformed using a square root transformation (A' ' = square root (A' + 0.5)). If a significant difference in densities was de- tected among the four habitat types, the types were then compared using Tukey's multiple range test (Zar 1984) to determine which types were different. To determine whether densities were different be- tween years at each site, quarterly estimates were com- pared using a Student's t-test (Zar 1984). For these tests, densities from the summer of 1987 were com- pared with the summer of 1988 for each site testing the null hypothesis: There is no difference between den- sities of each site between summer 1987 and summer 1988. Similarly, densities from the winter of 1987 were compared with the winter of 1988. I conducted in- dividual tests for copper, quillback, brown, and YOY rockfishes for summer and winter comparisons. Results Habitat type and site comparisons All size categories of copper, quillback, and brown rockfishes and YOY rockfish were seen at the high- relief natural habitat type (Fig. 3). Consistent monthly densities of large (>200 mm) copper (mean 3.7-13.0 fish/90-m'^ transect) and quillback (mean 3.0-7.0 fish/ transect) rockfishes were observed throughout the year on the high-relief rocky reefs. Brown rockfish (>200 mm) (mean 1.3-4.0 fish/transect) were seen at HRl (Or- chard Rocks) but infrequently at HR2 (Blakely Point). Low densities (mean<1.3 fish/transect) of subadult (80-200 mm) copper, quillback, and brown rockfishes were observed on both reefs until the late summer and fall 1987. YOY (<80 mm) rockfishes were observed (mean 8.0-57.0 fish/transect) from August-November 1987 but only on HR2. YOY were not observed in the summer or fall of 1988. After the intlux of YOY dur- ing summer and fall 1987 on HR2, an increase in 80-200 mm rockfishes was subsequently observed (mean 6.0-15.0 fish/transect). Rockfishes primarily utilized low-relief rocky reefs during the summer, coincident with peak Nereocyt^fis leufkeana and understory cover, and were infre(iuent- ly observed during other seasons (Fig. 4). Large cop- per (up to a mean 6.0 fish/transect) and brown rock- fishes (up to a mean 7.7 fish/transect) were observed mainly during the summer, whereas large quillback rockfish were not seen at LRl and only infrequently at LR2. Low densities (mean 0-3.3 fish/transect) of small copper and quillback rockfishes were seen on LR2, while small brown rockfish were never seen on either reef. YOY rockfishes were observed (mean 16.7 fish/transect) during August 1987 on LRl and (mean 1.3 fish/transect) on LR2 during October 1987, but not during the summer or October of 1988. On the artificial reefs, densities of large copper rock- fish fluctuated throughout the year. During both 1987 and 1988, high densities (mean 12.0-56.3 fish/transect) were seen September through May; however, low den- sities (mean 0.3-2.3 fish/transect) of copper rockfish were observed June through August (Fig. 5). Variable densities (mean 0-12.3 fish/transect) of large quillback rockfish were observed on ARl while they were lower in number (mean 0-2.3 fish/transect) on AR2 (Fig. 5). Large brown rockfish were infrequently (mean 0-0.3 fish/transect) seen on either artificial reef. Low den- sities (mean 0-5.0 fish/transect) of subadult (80-200 mm) copper rockfishes were observed on both artificial reefs throughout the year. Extremely high densities (mean 8.0-420.0 fish/transect) of subadult quillback rockfish were observed on both ARl and AR2, al- though densities were much higher on AR2 (up to 420 Matthews Habitat use by rockfishes in central Puget Sound 227 High relief rocky reefs Copper rockfish % 20 D86 F87 A87 J87 A87 087 D87 FB8 A88 J88 ASS Survey month CD HR 1 ^M HR 2 Brown rockfish :: 5- 0-1 BO-200 mm >> 3 * 3 ^ o h II 1 1 I 1 10 D86 F87 A87 J87 A87 087 D87 F88 A88 JSa A88 >200 mn D86 F87 A87 J87 A87 087 D87 FSB A88 J88 ASS Survey month tZD HR 1 H HR 2 High relief rocky reefs Ouillback rockfish 80-200 mm ±1. Ala. D86 F87 A87 J87 A87 0S7 DS7 F88 A88 J88 ASS >200 mm 5- 150t 100 50 D86 F87 A87 J87 AS7 087 DS7 F88 Survey month i i YOY rockfish <80 mm ASS JS HR ASS I HR 2 11 rB7 A87 J87 A87 087 D87 F88 A88 J88 A88 088 Survey month CZJ HR 1 ■■ HR 2 Figure 3 Mean monthly densities ( + 95% CI) of young-of-the-year (< 80 mm), subadult (80-200 mm), and adult (> 200 mm) rockfishes on the two high-relief reefs in central Puget Sound, December 1986-August 1988; YOY surveys December 1986-October 1988. 228 Fishery Bulletin 88(2), 1990 30 T 20 10 Low relief rocky reefs Copper rockfish 80-200 mm ^1^ 15 D86 FB? A87 J87 A87 087 D87 F8B A88 J88 A88 >200 mm iJii IOt 5-- A87 J87 A87 087 087 F88 A88 J88 A88 Survey month CZI LR 1 IB LR2 Brown rockfish )200 mm A87 je7 A87 087 087 F88 A88 J88 A88 Survey month c^ LR 1 ^B LR 2 5t Low relief rocky reefs Ouillbock rockfish 80-200 mm D86 F87 A87 J87 A87 087 087 F88 A88 J88 A88 5t >200 mm D86 F87 A87 J87 A87 087 087 F88 A88 J88 A88 Survey month CZ) LR 1 ^ LR 2 HU YOY rockfish (80 mm 20 3 e 2. n- .1 AS? 087 D87 FSS Survey month ABB J6e ABB 088 C^ LR 1 ^ LR 2 Figure 4 Mean monthly densities ( + 95% CI) of young-of-the-year (<80 mm), subadult (80-200 mm), and adult (>200 mm) rockfishes on the two low-relief reefs in central Puget Sound, December 1986-August 1988; YOY surveys December 1986-October 1988. No subadult brown rockfish were observed. fish/transect). Small brown rockfish were infrequent- ly (mean 0-0.3 fish/transect) seen on either artificial reef. YOY rockfishes were observed (mean 1.7-63.0 fish/transect) only on AR2 during the spring of 1987, and during the following fall, winter, spring, and sum- mer (mean 4.3-80.0 fish/transect). Low densities (mean 2.0 fish/transect) were observed in the October 1988 survey. YOY were never observed on ARl. Sand/eelgrass areas had the lowest densities of all habitats sampled; large copper rockfish (mean 5.0 in July 1987, mean 0.7 in August 1988), brown rockfish (mean 1.0, August 1987) and YOY (mean 111.0. July Matthews Habitat use by rockfishes in central Puget Sound 229 20 10-- Artificial reefs Copper rockflsh 80-200 mm J III k I II \\i ^i il hi k z z 100 086 F87 A87 J87 A87 087 D87 F88 ABB J88 A88 )200 mm 75-- 50 25-- -■ «i 1 ±^ A87 J87 A87 087 087 F88 ABB JBB AB8 Survey month CU AR 1 ^ AR 2 Brown rockfish 80-200 mm No survey No survey D86 F87 A87 J87 A87 087 D87 F88 A88 J88 A88 )200 m m No survey No survey 086 F87 AB7 J87 A87 087 087 F88 ABB J88 A88 Survey month CD AR 1 ^ AR 2 Artificial reefs Ouillback rockfish 600 i400-- £200 jLsl 80-200 mm 086 F87 A87 J87 A87 087 087 FB8 ABB J88 ABB )200 mm 20- 10- 3 3 z z n ll , 1 uii ..i I i i i xi 1 I D86 F87 A87 J87 A87 087 087 F88 ABB JBB ABB Survey month CD AR 1 ^ AR 2 JUU - YOY rockfish (80 mm a 1 e 150- n E Z V I o z i D z c c 3 .... 1 i i \ i D86 r87 AB7 J87 Aa? 087 087 F88 A88 J88 A88 088 Survey month CZl AR 1 ^m AR 2 Figure 5 Mean monthly densities ( + 95% CI) of young-of-the-year (<80 mm), subadult (80-200 mm), and adult (> 200 mm) rockfishes on the two artificial reefs in central Puget Sound, December 1986-August 1988; YOY surveys December 1986-October 1988. 230 Fishery Bulletin 88(2), 1990 1987) were the only rockfishes seen and only during surveys conducted in late July and August (Fig. 6). These observations of rockfishes coincided with the peak growth of eelgrass up to 1.5 m high; in the fall and winter the height of eelgrass beds was reduced to <0.5 m. No quillback rockfish were ever observed on sand/eelgrass. Nested analysis of variance Habitat type comparison When density differences were detected, the highest densities were observed on the high-relief rocky and artificial reefs (i-'<0.05 nested ANOVA, Table 1). Highest densities of large copper rockfish were observed on high-relief rocky reefs dur- ing both the summer of 1987 and 1988, whereas ar- tificial reefs had the highest densities in the fall and winter. Similar low densities of 80-200 mm copper rockfish were observed on all four habitat types all seasons except fall 1987. High-relief rocky reefs had the highest densities of large quillback rockfish all seasons; sand/eelgrass and low-relief rocky reefs were similar as large quillback rockfish were rarely observed. Artificial reefs had the highest densities of 80-200 mm quillback rockfish all seasons except winter and sum- mer 1988. No density differences were observed for both size groups of brown rockfish on any of the habitat types. Similarly, YOY rockfishes were observed on all four habitat types, and no differences were detected in densities of YOY rockfish among the four habitats. Replicate site comparison For many (23/42 groups) size groups of copper, quillback, brown, and YOY rock- fishes, there were significant differences in densities between the two replicate sites for most seasons (P<0.05 nested ANOVA, Table 1). The major excep- tion was that similar densities of large copper rockfish were observed on replicate sites all seasons except spring 1987. Year-to-year comparison For most size categories of copper, quillback, and brown rockfishes, there were no significant difference between their densities for the two periods tested: Summer 1987-summer 1988, and winter 1987-winter 1988 (Table 2). For the summer comparisons, the prin- cipal differences in densities were for YOY rockfish on all sites where they were observed: HR2, LRl, LR2, AR2, and SE2. YOY were observed at all five reefs in the summer of 1987; however, no YOY were ob- served in the summer of 1988, except at AR2. Other differences between summer 1987-1988 densities were 80-200 mm quillback rockfish on HR2; densities of Sand/eelgrass Copper rockfish )200 mm O a D86 F87 A87 J87 Ae7 087 D87 F8B A88 J8B Survey month CZ) SE 1 ^B Brown rockfish )200 mm ABB SE 2 _ 2 D86 Fa7 AS? J87 AB7 087 DB? FB8 ABB J8B ASS Survey month CU SE 1 ■■ SE 2 YOY rockfish (80 mm 086 ra? AB7 J87 A87 087 087 F88 A86 J88 A8B 088 Survey month ^B SE 2 Figure 6 Mean monthly densities (-1-95% Cl)of y(iung-of-the-year(<80 mm), suhadult (80-20(1 mm), and adult (>200 mm) rockfishes on the two sand/eelgrass areas in central Puget Sound, December 1986-AugT.ist 1988; YOY surveys December 1986-()clober 1988. No quillback rockfish (80-200 mm or >200 mm), subadult copper rockfish, or subadult brown rockfish were observed. small quillback rockfish were higher in the summer of 1988 after the YOY recruitment of 1987. Similar to the summer 1987-summer 1988 compari- sons, there were few significant density differences for winter 1987-winter 1988 (Table 2). The principal dif- ferences were that large copper rockfish and small Matthews Habitat use by rockfishes in central Puget Sound 231 Table 1 Results of the nested ANOVAs (P<0.05) testing for < iifferences between •ockfish densities among the four habitat types: High-reHef rocky reef (HR), low-relief rocky reef (LR), high-relief artificial reef (AR), md sand/eelgrass (SE). If there was a significant difference between type densities (Test 1), the results of Tukey s multiple range tes are Hsted from highest to lowest densities. Test 2 lists the significant differences (*) in densities between sites. Copper Quiliback Brown Young-of- 80-200 >200 mm 80-200 mm >200 mm 80-200 mm >200 mm <80 mm Test 1 1987 Winter AR>HR>SE=LR HR>AR>LR=SE Spring AR>HR>SE=LR HR>AR>LR=SE Summer HR>LR>AR = SE AR>HR>SE=LR HR>AR>LR=SE Fail AR>HR>LR = SE AR>HR>LR = SE AR>HR>SE=LR HR>AR>LR=SE 1988 Winter AR>HR>LR = SE HR>AR>LR=SE Summer HR>LR>AR = SE HR>AR>LR=SE Test 2 1987 Winter • Spring * * * * « Summer * * * * Fall * * * * 1988 Winter * * * * Summer * * * * * Table 2 Summary of year-to-year comparisons testing (Student's (-test) for differences in rockfish densities between summer (June. July. Aug.) 1987-summer 1988 and winter (Dec, Jan., Feb.) 1987-winter 1988. Listed are habitat types where a significant difference was detected. HR = high-relief rocky reef, AR = high-relief artificial reef, LR = low-relief rocky reef, SE = sand/eelgrass. Summer 1987-Summer 1988 Winter 1987- Winter 1988 Copper 80-200 mm >200 mm HR2 ARl HR2 HR2, ARl, AR2 Quiliback 80-200 mm >200 mm HR2, ARl, AR2 No differences ARl, AR2 HRl Brown 80-200 mm >200 mm No differences LRl HR2 No differences Young-of-the-year <80 mm HR2, LRl, LR2, AR2. SE2 AR2 quiliback rockfish densities were higher in the winter of 1988 on both artificial reefs. In addition, both small and large copper rockfish increased on HR2, Species comparison Small copper rockfish were infrequently seen on any habitat (Figs. 3-6), Small copper rockfish were first observed on the high-relief rocky reefs after the influx of YOY in the fall and winter. They were also sporad- ically observed on the artificial reefs and Blake Island low-relief reef. Large copper rockfish were seen on all habitat types and were observed all seasons. Consis- tent densities of large copper rockfish were observed 232 Fishery Bulletin 88(2). 1990 HIGH RELIEF ROCKY REEFS 50- • plOO 40- // ^^ -80 30- /^r""''^ / -60 •--' ■""^ / ''^\ / id 20- y cl / -"0 S CO > ■z. o < 10- i' ^Vi -20 ^ (r y o 2 y / \ \ / -60 ci 'Z. / / \ \ / UJ < LU / 1 \ \ / -40 5 5 / / \ \ / UJ u 0-' 1 \ 1^-^ 2 " — ■ ■ ^ ■ •-0 < WTR ( =.P SU 1 FALL WTR 2 SU 2 LJ 5 Figure 7 Mean percent of algal or eelgrass cover (O) compared with mean number ( • . combination of 80-200 mm and >200 mm) of copper, quillback. and brown rockfishes on the four habitat types in central Puget Sound for winter 1 (wtrl), spring (sp), summer 1 (sul), fall (fall), winter 2 (wtr2), and summer 2 (su2). Seasons are described in Methods section. Density increase in high-relief rockfish (*) follow- ed the influx of young-of-the-year rockfish. Percent eelgrass cover was relative to maximum height (1.5 m) of eelgrass which occurred during the summer. Small quillback rockfish were primarily observed on the artificial reefs all seasons and were infrequently observed on all other habitats (Figs. 3-5). Large quillback rockfish were most common on the high-relief rocky reefs and artificial reefs. Large quillback rockfish were rarely observed on the low-relief reefs. Neither small nor large quillback rockfish were observed on the sand/eelgrass. Small brown rockfish were observed only on the high- relief reefs and on one artificial reef (Figs. 3, 5). Large brown rockfish were seen all seasons on the high-relief reefs although highest numbers were observed at HRI (Orchard Rocks) (Fig. 3). They were also seen during the summer on low-relief reefs and sand/eelgrass (Figs. 4,6). YOY rockfish were observed on all four habitat types (Figs. 3-6). On the high-relief low-relief reefs, and the sand/eelgrass, they were first seen in the summer and fall of 1987. YOY were observed year-round on Boe- ing Creek artificial reef. YOY were not observed in the summer or October of 1988 except at AR2. Macrophyte cover varied seasonally; the densest cover was observed June through October on all reefs (Fig. 7). The high-relief reefs had dense canopies of Nereocystis leufkeuna and subcanopies o( Pterygophora califomica, and Agaru m that during the summer cov- ered 50-100% of the bottom along the transect lines. The two low-relief reefs had 100'7o bullkelp and under- story coverage June through October. The annual bullkelp was gone November through May and only the stipes of the understory perennial kelps remained; the blades of the understory kelps had eroded or were otherwise lost. There was no Nercocystis on the two artificial reefs; however, understory kelps covered ap- pro.ximately 0-20% of the transects, with peak growth in the summer. Sand/eelgrass transects were always covered with the perennial eelgrass throughout the year, although the height varied seasonally. The height peaked during the summer at apjjroximately 1 .5 m, and during the winter many of the blades were gone and the eelgrass was prostrate with little (0-0.5 m) vertical structure. Monthly mean temperatures were highest August through October (up to 13°C) and lowest (down to 7.1°C) in January and Februai-y (Fig. 8). There was little qualitative difference in temperatures between the different reefs, although de|)th ranges were 5-20 m along transect lines. on the high-relief reefs all seasons. In contrast, their densities fluctuated seasonally on other habitats; they were only observed during the summer and fall on the low-relief reefs during all seasons, although in lower densities during the summer on the artificial reefs, and only during the summer on the sand/eelgrass. Discussion Seasonal habitat use Habitats that underwent seasonal vegetation and sub- sequent structural changes had the most dramatic Matthews Habitat use by rockfishes in central Puget Sound 233 1 s 1 D — T I I i t J o 'Ml I £r^ J i in ± cr> to- r i i J i +1 >• >- ' >^ 01 i r : ; i i 2 : 1 - 3 D i I 1 t o O O z z z QJ Cl E 5- QJ C o a- :5 0- D86 r87 A87 ja7 A87 087 087 Date F38 ABB J88 A88 088 Figure 8 Mean monthly bottom temperatures (°C) ±95% CI measured along transect lines on the eight reefs in central Puget Sound, December 1986-October 1988. density differences (Fig. 7). On the low-relief rocky reefs and sand/eelgrass, the highest densities, mainly >200 mm copper, brown, and YOY rockfishes, were observed during the summer coincident with the dens- est growth of bull kelp, understory kelps, and eelgrass. When the vegetation died back, lowest densities of rockfishes were observed, presumably due to move- ment away from the reefs. Movement away from the Blake Island low-relief reef in the fall onto the Blake Island artificial reef was confirmed in a tag-recapture study (Matthews In press). Richards (1987) also noted a decrease in copper rockfish densities off Vancouver Island during the winter, but argued that it was not fish movement that caused the decrease in densities but rather that during the winter fish were simply more difficult to see as they presumably hid in crevices. On the low-relief reefs and sand/eelgrass areas I surveyed, this explanation was not the case. Once the vegetation died back there were few places for fish to hide un- detected by divers; there are no crevices or holes in the flat-bottom rock reef and sand. Fishing pressure was not responsible for the seasonal density fluctua- tions. Although I never directly measured the amount of fishing activity, I observed the highest number of sportfishing boats (both divers and anglers) at the high- relief reefs, whereas the lowest fishing effort occurred at the low-relief and artificial reefs. Larson and DeMar- tini (1984) compared two low-relief areas in southern California with and without giant kelp Macrocystis pyrifera. They found a higher biomass of fishes on the reef with kelp and concluded that the presence of kelp on low-relief reefs enhances fish biomass by providing prey and structure. The seasonal change in kelp on high-relief reefs had no effect on fish densities; presumably there is adequate structure (rocks and crevices) and prey on these reefs regardless of kelp cover. Other research in California has demonstrated that kelp on high-relief reefs had less effect on fish abundance when compared with their dramatic effect on low-relief reefs; when kelp cover declines on low- relief reefs, the shelter is lost but refuge in high-relief rock is permanent regardless of kelp cover (Quast 1968, Stephens et al. 1984, Ebeling and Laur 1988). Large (>200 mm) copper rockfish densities on the artificial reefs declined dramatically during the summer coinci- dent with higher densities of large copper rockfish on the low-relief rocky reefs. Copper rockfish may leave the artificial reefs in the summer because of the lack of vegetation (bullkelp and understory kelps) and its associated prey. The summer is an important feeding time for rockfishes when fat reserves are stored to be used as energy sources during the winter (Guillemot et al. 1985). Although the artificial reefs had isolated patches of perennial kelps (Agarum and Pter-ygophora) during the summer, this habitat type had the sparsest vegetation growth of all the habitats surveyed. Because low-relief rocky reefs, artificial reefs, and sand/eelgrass were not suitable year-round habitats, rockfish move to utilize alternate habitats during the winter (Mat- thews In press). Habitat comparison Apparently, high-relief rocky reefs were suitable habi- tat for copper, quillback, brown, and YOY rockfishes: This was the only habitat type where all size categories of the three species as well as YOY were observed, although few were observed in the 80-200 mm range. Actually, one should not expect to see high numbers of 80-200 mm rockfish. Copper and quillback rock- fishes are relatively slow-growing and may live up to 55 years (Richards and Cass 1987). Rockfishes release thousands of larvae, and the highest predation and mor- tality is on the youngest fish; presumably only a few survive to adults. Rockfish >200 mm represent several age groups— those rockfish about 5 years and older— whereas the 80-200 mm group represents only a few age classes of 1-5 years (Sandra Oxford, Wash. Dep. Fish., Seattle, WA 98115, pers. commun., summer 1988). Thus, one would expect to see more fish >200 mm than 80-200 mm on the high-relief natural reefs where the size groups co-occur. High-relief rocky reefs also had the most consistent densities of large rockfish and were the most structurally complex of all habitats surveyed. Although the vertical relief was similar to that on artificial reefs, natural rocky reefs had more cracks, crevices, and holes for fish to hide. Moreover, the algal diversity and cover was considerably greater; thus high-relief rocky reefs provided the most struc- turally diverse and persistent habitat. 234 Fishery Bulletin 88(2), 1990 Low-relief rocky reefs and sand/eelgrass were tem- porary habitats primarily utilized during the summer, coincident with summer vegetation growth. Presum- ably rockfishes utilize these habitats when structure and prey availability are highest and subsequently move to other more suitable habitats (Matthews In press). Although rockfish densities were low, these habitats cover considerably more area than artificial reefs or high-relief rocky reefs in Puget Sound, and thus are important to rockfish. The highest densities of rockfishes were observed on the artificial reefs, primarily 80-200 mm quillback rock- fish. Artificial reefs had high densities of 80-200 mm quillback rockfish throughout the year, although some fluctuations occurred, and high densities of large cop- per rockfish in the fall, winter, and spring. No other reefs surveyed had such high densities of 80-200 mm rockfishes of any species, and it is not known whether this results from higher survival of small (juillback rockfish on artificial reefs. Furthermore, it is puzzling that more large quillback rockfish were not found on the artificial reefs, considering the abundance of 80- 200 mm quillback rockfish. Several hypotheses could explain the low numbers of large quillback rockfish on the artificial reefs: (1) Artificial reefs are not suitable habitat for large quillback rockfish, i.e., they leave when they grow large; (2) the fish have not yet grown to the larger size; (3) some factor is preventing the fish from growing; or (4) there is a high mortality of fish once they reach the 200-mm size. I ruled out the pos- sibility that large copper rockfish competitively ex- cluded large quillbacks, as the two species coexist on high-relief natural reefs, although competition may be reduced at high-relief habitats if resources are not limiting. The shortage of large quillback rockfish on the artificial reefs is not depth-related since large quillbacks have been observed at similar depths in other studies (Moulton 1977, Richards 1987) and at HRl and HR2. Again, the lack of large quillback rockfish was not due to fishing pressure: I observed low levels of fishing at both artificial reefs. Furthermore, if fishing pressure was responsible for the lack of large fish, there would be few large copper rockfish. In any case, this apparent refuge for 80-200 mm quillback rockfish should be in- vestigated to determine if rockfish leave once they reach a certain size and eventually contribute to recrea- tional fisheries. Although it has been well established that artificial reefs attract high densities of fish, in this case 80-200 mm quillback rockfish, there is no infor- mation that verifies whether there is adequate food or if growth or mortality is similar to that observed on natural reefs (Ambrose and Swarbrick 1989). Year-to-year comparison In a 4-year study of kelp-bed fishes in southern Califor- nia, Ebeling et al. (1980) found little annual variation in numbers of bottom assemblages, including rock- fishes. Although my study demonstrated significant seasonal changes within a habitat (artificial reefs, low relief, sand/eelgrass), when analyzed between seasons there were few differences between years, with the ex- ception of YOY. Ebeling et al. (1980) suggested that the low annual variation in kelp-bed fish assemblages was characteristic of stable communities in predictable environments. Although more recent work (Ebeling and Laur 1988) demonstrated the dramatic effect storms can have on fish populations, Puget Sound is a relatively protected environment not subjected to open ocean waves and surge. Thus year-to-year varia- tion in Puget Sound rockfish populations may be small. Species comparison Copper rockfish occurred on all habitats and can be con- sidered a habitat generalist. Large copper rockfish were observed on all four habitats, although they left the artificial reefs in the summer and moved into shallower low-relief and sand/eelgrass areas. Buckley and Hueckel (1985) also noted a seasonal decrease of copper rockfish densities during the summer on Gedney Island artificial reef in central Puget Sound, and highest densities were noted during the fall and winter. They speculated rockfish movement was in response to prey occurrence (surfperch). Moulton (1977) specu- lated that copper rockfish move to deeper water dur- ing the fall in northern Puget Sound due to seasonal depth preference or prey availability. Turbulence pre- sumably does not contribute to movement; in Puget Sound most water motion on the bottom is from cur- rents, not surge or turbulence, and water motion is similar in the summer and winter. In the summer, cop- per rockfish utilized several habitats: high-relief rocky reefs, low-relief reefs, and sand/eelgrass. Their appear- ance on the shallower reefs during the summer and disappearance from the deeper artificial reefs is not simply a shallower depth preference in the summer; copper rockfish were observed during the summer at the high-relief reefs even at comparable artificial reef depths. Quillback rockfish were more restricted in tlieir dis- tribution when compared with copper rockfish. Similar- ly, quillback rockfish were not as widely distributed as copper rockfish at 12 study sites in the Strait of Georgia (Richards 1987). The small (juillback rockfish were seen in very high densities on the artificial reefs, but infre- quently on other habitats. Large quillback rockfish were primarily seen on high-relief rocky and artificial Matthews Habitat use by rockfishes in central Puget Sound 235 reefs; Richards (1987) noted qiiOlback rockfish densities were correlated with rehef, and highest densities were found on complex habitats. The lack of large quillback rockfish on low-relief and sand/eelgrass areas is pre- sumably due to their preference for high relief and com- plex habitats. Brown rockfish displayed the most restricted and perplexing distribution, primarily being found on Or- chard Rocks high-relief reef and the inshore low-relief reefs during the summer, and rarely observed on arti- ficial reefs or sand/eelgrass. Brown rockfish have been observed on artificial reefs in south Puget Sound (Greg Hueckel, Wash. Dep. Fish., Olympia, WA 98504, pers. commun., summer 1988) and it is unclear why they did not inhabit the artificial reefs in central Puget Sound. Additionally, brown rockfish are relatively uncommon on rocky reefs in northern Puget Sound and the San Juan Islands (Moulton 1977). In California, brown rock- fish are primarily found on sandy, low-relief areas (Mat- thews 1985); their different habitat use in Puget Sound could be due to local hybridization with congeners (L. Seeb, Southern 111. Univ., Carbondale, IL 62901, pers. commun., summer 1987). YOY were distributed differentially among all habitat types. The sand/eelgrass was a temporary YOY habi- tat, as YOY were observed in July and never again seen. Either the YOY left the area or died. Similarly, YOY were observed on the low-relief reefs but subse- quently emigi'ated or suffered high mortality; their con- tribution to recruitment on other reefs is unknown. On the other hand, YOY were first observed on the tran- sects on HR2 in August, although they were previous- ly seen in July off the reef on adjacent (within 25 m) sand/Agarum. The numbers of YOY increased over the next few months, peaked in November, and initial set- tlement was followed by an increase in the 80-200 mm copper, quillback, and brown rockfishes, presumably, the result of recruits staying and growing into the larger size category. A similar pattern of YOY influx was followed by an increase in the 80-200 mm group on AR2, although two periods of YOY settlement were observed, spring and fall. Parturition of these rockfishes reportedly occurs April through June, as most female rockfishes captured April and May near Bainbridge Island had embryos and ovaries that were in the transitional stage during the summer (DeLacy et al. 1964, Washington et al. 1979, Dygert 1986). On SCUBA surveys, I saw pregnant rockfish late April through late June. Therefore, the YOY that were observed during the summer and fall of 1987 on HR2 presumably were released between April and July, spent some unknown amount of time in pelagic regions or on some other habitat, and then settled to a demersal existence in July and August. Ex- aminations of the otolith microstructure of eight 45-60 mm TL rockfish (M. Yoklavich, Natl. Mar. Fish. Serv., Seattle, WA 98115, pers. commun., fall 1987) that I collected 29 August 1987 on HR2 revealed approx- imately 120-160 growth rings. Similar growth rings have been shown to be deposited daily in black rockfish (Yoklavich and Boehlert 1987). If the rings examined on my rockfish were also daily, it would confirm that parturition occurred around April. Additionally, these fish appeared to grow quite quickly. When first ob- served in July and August they were approximately 45-60 mm TL and by November were 90-100 mm TL (fish wei'e captured to verify these measurements). This scenario describes the parturition on the natural reefs which was quite different from the YOY settlement on Boeing Creek artificial reef. YOY were assumed to be primarily quillback rockfish, as YOY settlement was followed by an increase of 80-200 mm quillback rockfish. They were observed in the spring and the fall on Boeing Creek artificial reef. It is unclear when the spring recruits were released or if there are possibly two reproductive periods for quillback rockfish in Washington as noted for some California rockfishes (Wyllie Echeverria 1987). Brown rockfish sampled from one location in north-central California had two distinct seasons of larval extrusion, December and June, and Wyllie Echeverria (1987) concluded that rockfishes have a flexible reproduction system that enables individuals to adaptively respond to environ- mental factors. On the other hand, in my study young quillback rockfish could arrive from another source- northern Puget Sound, Strait of Juan de Fuca, or even the outer coast of Washington— causing the second pulse. On Vancouver Island, Haldorson and Richards (1987) noted an infliLX of YOY (< 50 mm) copper rockfish dur- ing August and September; the YOY utilized four habitat types: (1) Nereocysfis leutkeana, (2) Agarum slopes, (3) eelgrass, and (4) sand. They observed the highest densities of young copper rockfish first in the bullkelp canopies. Subsequently, the yoimg rockfish left the kelp canopy and were found on the floor of the kelp forest in September and October, coincident with fall storms and the annual decomposition of the bullkelp. After the initial association with the kelp canopy, the YOY shifted their distribution to a demersal habitat with the perennial macvo\>hyte?, Agarum and eelgrass. Similarly, Carr (1983) first observed YOY copper rock- fish in the upper canopy of giant kelp Macrocystis pyrifera in central California kelp beds. The YOY cop- per rockfish subsequently moved toward the bottom over the following weeks. Carr (1983) and Haldorson and Richards (1987) always observed YOY in close pro- ximity to drift or attached kelp or eelgrass and sug- gested that young rockfishes strongly associated with plant cover to avoid predation and find food resources. 236 Fishery Bulletin 88(2). 1990 Interestingly, I never observed YOY in bullkelp cano- py. The young fish were only seen on the bottom closely associated with the perennial understory macrophytes Agarum, Pterygophora, and eelgrass, or closely asso- ciated with rocks at AR2, which has little plant cover. The YOY could have spent a short period of time in the bullkelp canopy before I first saw them in my surveys or they may never use the bullkelp canopy in my study sites. Young rockfish may not be able to main- tain position in the water column in areas with current. It is not surprising that YOY rockfishes were seen in 1987, yet not in 1988. The numbers of first-year rockfish recruits reportedly vary greatly year-to-year along the coast of central and northern California (Hob- son et al. 1986, Gaines and Roughgarden 1987). Studies in California have documented wide fluctuations in YOY recroiitment sometimes influenced by warm-water years and changes in distribution after kelp removal (Bodkin 1988). Temperature did not appear to be a fac- tor in the absence of YOY in 1988; bottom tempera- tures were similar between 1987 and 1988 (Fig. 8). In addition, few YOY were observed off eastern Vancou- ver Island during the same time period (summer and fall 1988) (L. Richards, Pac. Biol. Sttn., Nanaimo, B.C., Canada V9R 5K6, pers. commun., spring 1989). Ap- parently, copper, quillback, and brown rockfish recniit- ment is episodic; this has important implications to our understanding of the population dynamics and manage- ment of Puget Sound rockfishes. The most numerous aggregations of YOY were ob- served on low-current reefs: HR2, AR2, and SE2. YOY were never observed on the highest current reefs, HRl and ARl. Possibly YOY settle out on high-current reefs but lack the ability to remain on the reef and are swept away by strong current. In contrast, larvae may accum- ulate at sites of low current velocity. Leaman (1976) collected significantly higher numbers of Sebastes larvae in sheltered waters than in open, more exposed channels off the west coast of Vancouver Island. Similarly, Haldorson and Richards' (1987) observations of YOY rockfishes on eastern Vancouver Island were in low-current areas. Experimental design Because underwater transects are an effective means of describing large diurnally active fish populations (Brock 1954, Brock 1982), my transects were probably effective for estimating abundances of adult rockfishes, although I probably underestimated YOY rockfishes and the very dense aggregations of 80-200 mm quillback rockfish on the artificial reefs. For many size categoi-ies of the three species during most seasons, there were density differences between the two repli- cate sites (reefs) within each habitat type. This variabil- ity suggests there may not be a typical rocky reef or artificial reef. In most cases, though, this difference was due to variability in densities and not to different species present on the reef. For example, both ARl and AR2 had very high densities of 80-200 mm quillback rockfish, perhaps a characteristic of artificial reefs in Puget Sound. High densities of 80-200 mm rockfishes were not observed on any other reef type, but AR2 had significantly larger densities, perhaps due to location or age of the reef. Habitat quality Which is the highest quality habitat? One view of habi- tat quality would predict that densities are directly pro- portional to quality; rockfish densities are highest where habitat quality is highest. This reasoning forms the basis of many current habitat assessment models in resource management (Van Home 1983). In the absence of long-term information on growth, survival, stability, and other features of how species respond to different habitats, habitat quality is evaluated in short- term studies describing densities (Van Home 1983). Those habitats with the highest densities will be those most important to the maintenance of that species and should ultimately be protected as critical habitat. It is important to note that densities can be misleading in designating habitat quality (Van Home 1983). While my habitat surveys documented the highest rockfish densities on artificial reefs, the densities underwent major fluctuations. According to classic ecological theory, as density increases habitat quality declines (Svardson 1949, Fretwell 1972). If resources are limited, then at high rockfish densities, food, hiding, and resting places would be in short supply, thereby reducing growth and survival. In my study, changes in availability of resources presumably were the causes for rockfish density fluctuations. Therefore, a true mea- sure of habitat quality should include factors other than densities, such as availability of essential resources, stability over time, survival, reproductive output, growth, and the animal's preference for that habitat. Nevertheless, copper, quillback, and brown rockfish utilized all four habitat types surveyed, and each habitat is important. But some of these habitats vary seasonally vary their suitability, and so rockfishes move to exploit alternate habitats when suitability is low (Matthews In press). Thus, for rockfishes, I would argue that high densities should not be the primary in- dicator of habitat quality, particularly in reference to the high ratio of small/large quillback rockfish on ar- tificial reefs. Documenting densities of rockfishes on- ly during a short time frame (few months) would not be representative of that habitat's importance. In addi- tion, low-relief rocky reefs, although only important as Matthews Habitat use by rockfishes in central Puget Sound 237 rockfish habitat in the summer, Hkely represent essen- tial summer feeding area for rockfishes. Although den- sities are low, this habitat type is widespread in Puget Sound. In order to protect low-relief habitats, they would obviously have to be preserved year-round, i.e.. not altered or destroyed. Thus, in terms of seasonal stability in densities, presence of all size categories, and seasonal resource availability, I suggest the following ranks for the different habitats, listed from highest to lowest: High-relief rocky reefs, low-relief reefs (impor- tant as summer feeding area and thus contribute to year-round growth), artificial reefs (pending further research and assessment), and sand/eelgrass. Regardless of which habitat has the highest quality, the habitat types examined in this study were utilized differently by the various size groups and species. Recently, however, Hueckel et al. (1989) have sug- gested that artificial reefs can be used as mitigation for the loss of natural rocky-reef habitat. Mitigating the loss of natural habitats with artificial reefs uses the rationale that artificial reefs provide rocky-reef type substrate that replicates natural rocky reefs. Hueckel and Buckley (1989) reported that artificial reefs in Puget Sound replicate processes on natural reefs and provide as evidence the similarity in the number of fish and prey species present on artificial and natural reefs. My research indicated that rockfish species composi- tion was similar (with the exception of brown rockfish) between natural and artificial reefs. However, these species utilize artificial reefs quite differently than natural reefs: Large copper rockfish leave artificial reefs during the summer, large quillback rockfish are found in small numbers, and artificial reefs are domi- nated by extremely high densities of small quillback rockfish, unlike any natural reef I surveyed. In fact, the artificial reefs I studied seem to represent an anomalous habitat unlike any natural habitat. Furthermore, the difference in resource availability and habitat use may result in different birth and death rates on different habitats. Thus, the different habitats could be viewed as sources, where reproductive out- put exceeds deaths, or sinks, where a deficit exists (Pulliam 1988). On the superficial observation of high densities of quillback rockfishes on artificial reefs, it could be determined that the destruction of a nearby rocky reef would have little impact on rockfishes. Pulliam (1988) points out that if a habitat (e.g., artificial reef) being preserved was a sink and the one being destroyed was a source (e.g., natural rocky reef), destruction of a relatively small source habitat could then have disastrous results. For example, if a low- relief natural reef was destroyed for a development project, the loss of the productive feeding area used by artificial-reef rockfishes during the summer could result in reduced growth and less reproductive output. Again, whether or not artificial reefs provide adequate food for such high densities of rockfishes is unknown. The largest rockfish inhabiting artificial reefs were those that made use of natural habitats; their move- ment was confirmed in a tagging study (Matthews 1988). Thus copper rockfish may be maintained by a source habitat when they make use of kelp beds dur- ing the summer. In addition, artificial reefs are con- siderably smaller than natural reefs, so they would not compensate for the total loss of abundance (Ambrose and Swarbrick 1989). Thus it is premature and specu- lative to suggest that artificial reefs should be used as mitigation. Summary Habitat surveys comparing monthly densities of cop- per, quillback, and brown rockfishes on high-relief rocky reefs, low-relief rocky reefs, high-relief artificial reefs, and sand/eelgi'ass areas demonstrated strong dif- ferences in how rockfishes utilize these habitats. High- relief rocky reefs had the most consistent densities of the three species of rockfishes, mostly fish >200 mm. Low-relief rocky reefs were primarily inhabited in the summer months coincident with the summer growth of Nereocystis leutkeana. YOY rockfishes were also observed on low-relief reefs; however, most fish left these reefs in the fall. The highest densities of rock- fishes, primarily 80-200 mm quillback rockfish (up to 420/90-m'^ transect), were observed on artificial reefs and high densities of large copper rockfish were also observed. On artificial reefs, density fluctuations were dramatic; copper rockfish densities peaked in the fall and winter and declined (to 0/transect) during the sum- mer and quillback rockfish densities also seasonally fluctuated. Brown rockfish were rarely seen on the artificial reefs. Sand/eelgrass areas were the least util- ized habitat type; only during July and August were YOY, adult copper rockfish, and brown rockfish ob- served on one sand/eelgrass habitat. Although all four habitats were used, natural reefs may represent source habitats that are used by and maintain rockfishes on less productive sink (artificial reef) habitats. Thus the recent use of artificial reefs as mitigation for the loss of natural reefs could have negative impacts on rockfish populations. Acknowledgments I am grateful to the many people who helped me com- plete this study, which was part of my dissertation research at the University of Washington School of Fisheries. I was fortunate to have the expert diving assistance of Robert Reavis throughout the entire 238 Fishery Bulletin 88(2), 1990 course of my field work; other divers included Vince Macurdy, Wayne Palsson, and Amy Unthank. Dr. Loveday Conquest and Kevin Lohman assisted me with statistical consultation. Funds for this project were pro- vided by the Federal Aid to Sport Fish Restoration Act (Wallop-Breaux), Project F-81-R; Segments 1-3, Moni- toring and Assessment of Puget Sound Recreational Bottomfish Stocks, through the efforts of Greg Barg- mann, Cyreis Schmitt, and Wayne Palsson at Wash- ington Department of Fisheries. Support during manuscript preparation was provided by a Canadian Government Postdoctoral Fellowship at the Pacific Biological Station. This manuscript was greatly im- proved by constructive reviews from Greg Bargmann, Dee Boersma, Mark Carr, Don Gunderson, Bruce Miller, Wayne Palsson, Tom Quinn, Cyreis Schmitt, and two anonymous reviewers. Citations Ambrose R.F., and S.L. Swarbrick 1989 Comparisons of fish assemblages on artificial and natural reefs off the coast of southern California. Bull. Mar. Sci. 44:718-73.3. Barsukov, V.V. 1981 A brief review of the subfamily Sebastinae. .J. Ichthyol. 21:1-26. Bodkin, J.L. 1988 Effects of kelp forest remc>val on associated fish assem- blages in central California. J. Exp. Mar. Biol. Eool. 117: 227-2.38. Boehlert. G.W., and M.M. Yoklavich 1984 Reproduction, embryonic energetics, and the maternal- fetal relationship in the viviparous genus Seiostes (Pisces: Scor- paenidae), Biol. Bull. (Woods Hole) 167:354-370. 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. Buckley. R.M., and G.J. Hueckel 1985 Biological processes and ecological development on an artificial reef in Puget Sound. Washington. Bull. Mar. Sci. 37:.S0-69. Carr. M.H. 1983 Spatial and temporal patterns of recruitment of young- of-the-year rockfishes (Genus Sebastes) into a central Califor- nia kelp forest. Master's thesis, San Francisco State Univ., San Francisco, CA 94132, 104 p. 1989 Effects of macroalgal assemblages on the recruitment of temperate zone reef fishes. J. Exp. Mar. Biol. Ecol. 126: ,59-76. Culver. B. N. 1987 Results from tagging black rockfish {SehaKtes mi'lanops) off the Washington and northern Oregon coast. In Proc. Int. Rockfish Symp., p. 231-240. Univ. Alaska Sea Grant Rep. 87-2. Fairbanks, Anchorage, Alaska, 99701. Delacy, A.C., C.R. Hitz. and R.L. Dryfoos 1964 Maturation, gestation, and birth of rockfish {SrhaatiuUv) for Washington and adjacent waters. Wash. Uep. Fish., Fish. Res, Pap. 2:51-57. Dygert. P.H. 1986 Management implications of variability in reproduction and growth of commercial marine fishes. Ph.D. diss., Univ. Wash., Seattle, 98195, 275 p. Ebeling, A.W., and D.R. Laur 1985 The infiuence of plant cover on surfperch abundance at an offshore temperate reef. Environ. Biol. Fish. 12:169-179. 1988 Fish populations in kelp forests without sea otters: Ef- fects of severe storm damage and destructive sea urchin graz- ing. In Van Blaricom, G.R., and .J. A. Estes (eds.). The com- munity ecology of sea otters, p. 169-191. Ecol. Stud. 65, Springer- Verlag. 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 Power, D.M. (ed.), Multidisciplinary symposium on the California islands, p. 404-431. Santa Barbara Mus. Nat. Hist.. Santa Barbara. CA. 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. Fretwell, S.D. 1972 Populations in a seasonal environment. Princeton l^niv. Press, Princeton, N.J, 217 p. Gaines, S.D., and J. Roughgarden 1987 Fish in offshore kelp forests affect recruitment to inter- tidal barnacle populations. Science (Wash. DC) 235:479-481. Go wan, R.E. 1983 Population dynamics and exploitation rates of rockfish {Sebastes spp.) in central Puget Sound, Washington. Ph.D. diss., Univ. Wash., Seattle, 98195, 90 p. Guillemot, P.J., R.J. Larson, and W.H. Lenarz 1985 Seasonal cycles of fat and gonad volume in five species of northern California rockfish (Scorjiaenidae: Sdui.'ites). Fish. Bull., U.S. 83:299-312. Haldorson. L.. and L.J. Richards 1987 Habitat use by young-of-the-year copper rockfish {Si'bastcs caurinus) in British Columbia. In Proc. Int. Rockfish Symp., Anchorage, Alaska, p. 1 29- 1 4 1 . Alaska Sea Grant Rep. 87-2, Fairbanks 99701. Hart, J.L. 1973 Pacific fishes of Canada. Fish. Res. Board Can. Bull. 180. 740 p. Hobson, E.S. 1972 Activity of Hawaiian reef fishes during the evening and morning transitions between daylight and darkness. Fish. Bull., U.S. 70:71.5-740. Hobson. E.S.. P. Adams. J. Chess. D. Howard, and W. Samiere 1986 Temporal and spatial variations in the numbers of fir.st- year juvenile rockfishes off northern Cahfomia. Admin. Rep. T-86-02, Southwest Fish. Cent., Natl. Mar. Fish. Serv., NOAA, La Jolla, CA 92038, 16 j.. HueckeL G.J.. and R.M. Buckley 1989 Predicting fish species on artificial reefs using indicator biota from natural reefs. Bull. Mar. Sci. 44:873-880. Hueckel, G.J., R.M. Buckley, and B.L. Benson 1989 Mitigating rocky luibit^it loss using artificial reefs. Bull. Mar. Sci. 44:913-922. Larson. R.J. 1980 Competition, habitat selection, and the bathymetrie segregation of two rockfish {Sebnxifs) s[.)ecies. Ecol. Monogr. 50:221-2,39. Matthews Habitat use by rockfishes in central Puget Sound 239 Larson, R.J., and E.E. DeMartini 1984 Abundance and vertical distribution of fishes in a cobble- bottom kelp forest off San Onofre, California. Fish. Bull, U.S. 82:37-55. Laufle, J.C. 1982 Biological development and materials comparison on a Puget Sound artificial reef. Wash. Dep. Fish. Tech. Rep. 72. 183 p. Lea, R.N. 1983 Current approaches to and problems with rockfish sys- tematics [Abstract]. In Henry, F. (ed.), Proc. 1983 Ground- fish Workshop, Monterey, California. Calif. Dep. Fish Game, Menlo Park, CA 94025 [unpaginated], Leaman, B.M. 1976 The association between black rockfish (Sebastes melaruyps Girard) and beds of the giant kelp {Macrocystis integrifolia Bory) in Barkley Sound, British Columbia. Master's thesis, Univ. British Columbia. Vancouver, 108 p. Love, M.S. 1980 Isolation of olive rockfish, Sebastes serrayioides. popula- tions off southern California. Fish. Bull., U.S. 77:975-983. Matthews. K.R. 1985 Species similarity anti movement of fishes on natui'al and artificial reefs in Monterey Bay. California. Bull. Mar. Sci. 37:252-270. 1988 Habitat use and movement patterns of copper, quillhack, and brown rockfishes in Puget Sound. Washington. Ph.D. diss., Univ. Wash., Seattle 98195. 121 p. In press An experimental study of habitat preferences and movement patterns of copper, quillback, and brown rockfishes. Environ. Biol. Fish. Moulton, L.L. 1977 An ecological analysis of fishes inhabiting the rocky near- shore regions of northern Puget Sound, Washington. Ph.D. diss., Univ. Wash., Seattle 98195, 181 p. Palsson. W.A, 1988 Bottomfish catch and effort statistics from boat-liased recreational fisheries in Puget Sound. 1970-1985. Wash. Dep. Fish. Prog. Rep. 261. 104 p. Patten, B.G. 1973 Biological information on cojiper rockfish in Puget Soimd, Washington. Trans. Am. Fish. Soc. 102:412-416. Phillips, R.C. 1984 The ecology of eelgrass meadows in the Pacific northwest: A community profile. U.S. Fish. Wildl. Serv. FWS/OBS-84/24, 85 p. Pulliam, H.R. 1988 Sources, sinks, and population regulation. Am. Nat. 132:6.52-661. Quasi, J.C. 1968 Effects of kelp harvesting on the fishes of the kelp beds. In North, W.J., and C.L. Hubbs (eds.). Utilization of kelp bed resources in southern California, p. 143-149. Calif. Dep. Fish Game Fish. Bull. 139. Richards, L.J. 1987 Copper rockfish (Sebastes caurhtus) and quillback rockfish (Sebastes mnliger) habitat in the Strait of Georgia, British Columbia, Can. .1. Zool. 65:3188-3191. Richards, L.J., and A.J. Cass 1987 The British Columbia inshore rockfish fishery: Stock assessment and fleet dynamics of an unrestricted fishery. In Proc. Int. Rockfish Symp., Anchorage, Alaska, p. 299-308. Univ. Alaska Sea Grant Rep. 87-2, Fairbanks 99701. Stephens, J.S., P. A. Morris, K.E. Zerba, and M.S. Love 1984 Factors affecting fish diversity on a temperate reef: The fish assemblage of Palos Verdes Point, 1974-1981. Environ. Biol. Fish. 11:259-275. Svardson, G. 1949 Competition and habitat selection in birds. Oikos 1: 157-174. U.S. Department of Commerce 1987 Tidal Current Tables 1987. Pacific Coast of North America and Asia. Natl. Ocean Serv., Natl. Oceanic Atmos. Adm., Rockville. MD 20852, 279 p. Van Home, B. 1983 Density as a misleading indicator of habitat quality. J. Wildl. Manage. 47:893-901. Washington, P.M., R.G. Gowan. and D.H. Ito 1979 A biological report on eight species of rockfish (Set}astes) from Puget Sound, WA. Proc. Rep., Northwest Alaska Fish. Cent., Natl. Mar. Fish. Serv., NO.AA, 7600 Sand Point Way NE, Seattle, WA 9811.5-0070, .50 p. Wyllie Echeverria, T.W. 1987 Thirty-four sj.iecies of California rockfishes: Maturity and seasonality of reproduction. Fish. Bull., U.S. 85:229-250. Yoklavich, M.M., and G.W. Boehlert 1987 Daily growth increments in otoliths of juvenile black rockfish, Sebastes rrwlarwps: An evaluation of autoradiography as a new method of validation. Fish. Bull.. U.S. 85:826-832. Zar, J.H. 1984 Biostatistical analysis. Prentice-Hall. Inc., Englewood Cliffs, NJ, 718 p. Abstract. — Two species of sand lances are recognized in the western North Atlantic, the inshore Ammo- dytes americanus DeKay 1842 and the offshore A. duhiua Reinhardt 1838. The best separation of the two species is achieved by using the num- ber of plicae (oblique folds of skin on the lateral body surface) singly or in combination with the number of ver- tebrae. Ammodyte^ ameyicanus has fewer meristic structures than A. dubius: number of lateral plicae 106- 126, X 117.4 vs. 124-147, 132.1; total vertebrae 62-70, x 66.4 vs. 68-76, 70.8; dorsal fin rays 52-61, X 57.4 vs. 56-67, 61.8; anal fin rays 26-33, J 29.4 vs. 28-35, 31.1; pec- toral fin rays 11-15, x 13.2 vs. 12- 16, 14.0; gill rakers on first arch 21- 28, X 24.3 vs. 23-31, 26.6. Meristic differences between species were summarized with principle compo- nent analysis. In addition to con- siderable variation within samples, there is geographic variation in num- bers of vertebrae, plicae, and dorsal and anal fin rays, particularly in the offshore A. dubius. Specimens from the Scotian Shelf north have higher counts than do specimens from more southern populations. Based on spec- imens e.xamined, A. americanus oc- curs from southern Delaware north to Labrador in shallow coastal waters as well as in protected bays and es- tuaries. Ammodytes dubius is found in deeper, open waters from North Carolina to Greenland. Separation of Two Species of Sand Lances, Ammodytes americanus and A. dubius, in thie Western iMortli Atiantic Martha S. Nizinski Bruce B. Collette Betsy B. Washington National Systematics Laboratory, National Marine Fisheries Service. NOAA National Museum of Natural History, Washington, DC 20560 Manuscript accepted 26 January 1990. Fishery Bulletin. U.S. 88:24 1-255. Members of the genus Ammodytes, or sand lances, are small elongate fishes abundant over shallow, sandy areas of the continental shelves of northern oceans. They are important prey items for several commercial fishes (e.g., American plaice, cod, haddock, silver hake, yellowtail flounder, and Atlantic salmon (Reay 1970, Meyer et al. 1979, Winters 1983) as well as fin and humpback whales (Overholtz and Nicolas 1979) and various kinds of seabirds (Reay 1970, Powers and Backus 1987). Additionally, in the North Sea and off Japan, sand lances are the basis for an important fish-meal industry (Macer 1966). Western North Atlantic popula- tions of sand lances have increased dramatically in recent years (Sher- man et al. 1981, Winters 1983). This population explosion was correlated with a decline in stocks of herring Clupea harengus and mackerel Scom- ber scombrus along the eastern coast of the United States (Population Dynamics Branch, Conservation and Utilization Div., Northeast Fish. Cent., Natl. Mar. Fish. Ser\'., NOAA, Woods Hole, MA 02543) (Fig. 1). The opportunistic sand lances seemed to have replaced these stocks. Concur- rently, piscivorous fishes increased their consumption of sand lances. Peak abundance of sand lances in this region was readied in 1981, and numbers have since decreased (Nelson and Ross 1987). Again, the shift in sand lance abundance was correlated with mackerel numbers; mackerel populations have been steadily increasing since 1983 (Fig. 1). In view of the ecological impor- tance and population dynamics of sand lances, it is important that the taxonomic status of these fishes be resolved. Although the taxonomy of the majority of fish species in the west- ern North Atlantic Ocean is known reasonably well, Ammodytes is a major exception despite the results of several detailed studies (e.g., Richards et al. 1963, Scott 1972, Winters and Dalley 1988). Recogni- tion of two species of Ammodytes in the western North Atlantic dates back to at least Jordan and Ever- mann (1896). Since then, the major- ity of researchers have fundamental- ly accepted this finding (Bruun 1941, Backus 1957, Richards et al. 1963, Leim and Scott 1966, Winters 1970, Scott 1972, Winters and Dalley 1988, Scott and Scott 1988). However, final conchisions on species names, syn- onomies, and meristic and geograph- ic ranges vary between studies. Ammodytes americanus DeKay 1842 and .4. dubius Reinhardt 1838 are currently accepted as the appropriate names for these sand lances (Leim and Scott 1966, Reay 1970, Richards 1982, Winters and Dalley 1988), and until a worldwide systematic revision 241 242 Fishery Bulletin 88(2), 1990 of the family is completed these names should be used. Taxonomic confusion results mainly from the mor- phological similarity and large variability in characters traditionally used to separate and identify the different species ofAmmodytes. Generally, distinguishing char- acters have been limited to meristic ones, especially numbers of vertebrae and dorsal and anal fin rays. Meristic overlap and variability are further complicated by the trend for more northerly populations of both species to have higher counts. Meristic characters show clinal variation, increasing both with latitude and distance offshore (Richards et al. 1963, Scott 1972, Richards 1982). Body depth and maximum total length have also been used to separate species (Richards et al. 1963, Winters and Dalley 1988), although this char- acter combination has been shown to vary greatly with stage of maturity and age. In fact, intraspecific varia- tion is gi'eater than interspecific variation in some cases (Scott 1972). Additionally, the phase of the reproduc- tive cycle and the amount of food present in the digestive tract also affect body depth in these fish. With numerous hypotheses and species names cir- culating in the literature (see Richards et al. 1963 and Winters and Dalley 1988 for additional reviews), a need to alleviate some of the confusion associated with this genus is obvious. The western North Atlantic sand lances were studied in detail to determine appropriate species definitions, delimit geographic distributions, and describe meristic variation. Historical background Systematic problems involving Annnodyfes are preva- lent at all levels of taxonomic complexity. Researchers still are not sure of the phylogenetic relationships be- tween genera of sand lances and the systematic place- ment of the Ammodytidae among perciforms. Pietsch and Zabetian (1990), however, believe ammodytids to be trachinoids with the family Ammodytidae the sister group of the Trachinidae plus Uranoscopidae. At the alpha taxonomy level, 23 nominal species of the genus Arnmodytes have been described. However, only the following six species have been consistently recognized in the literature: A. americanus DeKay 1842 and A. dubius Reinhardt 1838 in the western North Atlantic, A. mariyiiix Raitt 1934 and /I. tohianxs Linnaeus 1758 in the eastern North Atlantic, and A. hexapterus Pallas 1811 and A. personatus Girard 1857 in the North Pacific (Reay 1970). However, Reay's (1970) synopsis of valid species names and delimitations of geographic ranges has not always been accepted; synonymies and geographical range adjustments are abundant. For instance, several workers including ^soo 1 1 / ■ ^ Sand Lance /\ ,— . MacVutol ,• \ / '\ / \ -- V::/--7-<.J-.: fo lyaa ibhj Figure 1 Abundance (10" metric tons) of mackerel Scomber scombrus, herring Clufn'ii hart'ngu^, and sand lance Arnmodytes dubiun populations. 1963-87, in the western North Atlantic (Northeast Fi.sh. Cent., Natl. Mar. Fish. Serv., NOAA, Woods Hole, MA). Andriashev (1954), Walters (1955), McAllister (1960), and Richards et al. (1963) have proposed that A. hex- apterus is circumpolar and synonymous with A. ameri- canus and/or /I. marinus. Regarding the sand lances that occur in the extreme northern oceans (particularly Greenland), some investigators suggest that the low- count inshore form should be regarded as A. ameri- canus (Reay 1970). Winters and Dalley (1988) proposed that these northern populations were A. wnrinus. Thus, these two nominal species may be conspecific, but researchers disagree on which name to use for the species in the western North Atlantic. Ammodyfes duhius also occurs off the coast of Greenland but generally is considered to be a distinct species (e.g., Backus 1957, Reay 1970, Winters and Dalley 1988). Additionally, Jensen (1941, 1944) suggested that the broad overlap in characters used to separate Arn- modytes species is so great that there may be only one extremely variable, polymorphic Atlantic species of sand lance. Furthermore, clinal variation adds to the taxonomic confusion since some investigators recognize this varia- tion as worthy of species designation while others pro- pose the use of subspecies or only recognize isolated populations. Richards et al. (1963) suggested that clinal variation may be due to effects of environmental fac- tors such as temperature, a phenomenon that has been documented in many other marine fishes. These trends have been reported in all species oi Arnmodytes, again leading to the suggestion of a single heterogeneous species with meristic and morphological differences attributable to environmental variables related to distribution. Nizinski et al,: Separation of Ammodytes amencanus and A dubius in western North Atlantic 243 Methods Approximately 1500 specimens from a range of loca- tions (North Carolina to Greenland) along the western North Atlantic coast were examined in this study (Fig. 2; Appendix 1). Total number of gill rakers were counted on the first arch on the right side; pectoral fin rays were counted on the left. Dorsal and anal ray counts were obtained from radiographs. The anterior 1-3 pterygiophores of the dorsal fin may have no asso- ciated fin ray; therefore, dorsal fin counts began at the first visible ray. Several cleared and stained specimens were examined to verify counts obtained from radiographs. Precaudal, caudal, and total vertebral counts were also taken from radiographs. The first caudal vertebra was defined as the centrum bearing the first elongate hemal spine. Total vertebral counts exclude the hypural plate following the practice of most recent workers on the genus. Sand lances possess distinctive rows of oblique folds of skin or plicae which occur on the lateral body sur- faces and are lined on the undersides by cycloid scales. The rows of plicae characteristically extend from an area above the pectoral fin base to the caudal pedun- cle. Plicae length and the angle of direction at which plicae run are highly variable near the head and tail (Fig. 3). Total plicae counts proved difficult to make and were not repeatable. Therefore, we modified plicae counts to begin with the first plica posterior to the first 70< ~~Vt-,'° '■ "Sv ^0 V 40 0) Baffi 1 Island \ ^\, Greenland 0a«(8 Strait ® > \ ^-— ^ a \® \ < so J Hudaon Bay 1 \<' ) \ \ \ Labrador ^-^-v, \ - ®® y \ \ ''^ \ \ ^^ <> Newfound- K \ ® K 50 V \ \ \ \ Ouetivc ) MalneJ>» ( / I ® \ ®® ^- \ rjl^ ^^^-- .. - \ C ■^ s J9® S® 1 \ \ \ \ \ 30 y \ --yf ' \:: A americanus V dubius \ \ Geo rgia yao Figure 2 Distribution o{ Ammodytes aintru-anua and /I. dubius in western North Atlantic based on specimens examined for this study. .__^ , ' . , ;/ >i fi J >U i, .. ' . — , , „ . 1' I.' ..■ .■ — L — : Figure 3 Drawing of Ammodytes dubius shows how phcae were counted (between arrows). The irregular plicae present anteriorly and posteriorly (see insets) were not counted. 244 Fishery Bulletin 88(2), 1990 pored lateral line scale and continuing posteriorly to the plica associated with the last pored scale (Fig. 3). Because plicae are much easier to count than scales, the folds themselves were counted between the two end points. If, however, the regular serial arrangement of plicae was interrupted (i.e., size and direction of slope of the plicae become irregular), the pored lateral line scales were counted in order to maintain consistency in counts between specimens. Plicae counts made in this manner are much more consistent. Meristic data were divided into four geographic regions in order to describe and analyze geographic variation: Labrador, Quebec-Nova Scotia, Maine- Massachusetts, and New York-North Carolina. Since three species of Ammodytes may be present in Green- land waters, Greenland specimens were analyzed sep- arately. Conclusions made with regard to the other geographic regions were applied to Greenland collec- tions in an effort to determine which species are pre- sent in this region. To determine if two species could be distinguished objectively, principle component analysis (PCA) was conducted on a covariance matrix for meristic char- acters (plicae, vertebrae, dorsal, anal, and pectoral fin rays, and gill rakers) for 332 individuals. These speci- mens, all with a complete complement of meristic values, were pooled from locations throughout the geographic range, exclusive of Greenland. An attempt was made to give equal representation to each area; however, there were not many offshore collections from Labrador available. Since the data set contained both individuals with low meristics and those with high counts, the assumption was made that both forms were present within the sample. An individual's score on the first component was determined by entering its values for meristic characters into an equation. Component scores were plotted and two groups are clearly present (Fig. 4). The point of least overlap was chosen as the boundary between the two species. Data points were then coded using earlier tentative identifications (based on observed meristic trends) to determine the effec- tiveness of the methodology. The equation and the value of the score at the boundary between the two species were cross-validated by testing an additional 54 specimens. Results Number of lateral plicae was the most useful single character in separating the two species, even though plicae numbers varied considerably between individuals and slightly from side to side in individual specimens (Table 1). Of 723 specimens included in the analysis, 98.8% could be separated into species at a line of Figure 4 Histogram of principal component analysis (PCA) scores for si.x meristic characters oi Ammodytes. Ranges: A. am.ericanus 1199- 1378, A. duhius 1378-1572. Component scores were calculated using the following equation: (8.09 x number of plicae) + (2.44 x number of dorsal rays) + (2.31 x total number of vertebrae) + (1.23 x total number of gill rakers) + (1.03 x number of anal rays) + (0.45 x number of pectoral rays). separation between 124 and 125 plicae (Fig. 5). Ammo- dytes americanus had fewer plicae, ranging from 106 to 126 (x 117.4) with virtually no geographic varia- tion; A. dubius had more plicae (124-147, x 132.1) and exhibited geographic variation. Plicae means for A. dubius ranged from 131.2 in the New York-North Carolina region to approximately 132-136 from Massa- chusetts northward. Number of vertebrae, a standard character examined by previous researchers interested in Amnwdytes, also proved to be a relatively good character in separating species. Although some overlap in vertebral numbers exists, particularly south of Nova Scotia, 92.3% separa- tion was achieved at a line of separation between 68 and 69 vertebrae (Table 2). Ammodytes americanus had fewer vertebrae, with mean values ranging from 65.2 in specimens from New York to North Carolina and increasing to a mean value of 67.2 in specimens from Quebec to Labrador. Ammodytes dubius characteris- tically had more vertebrae than A. americanus, with means ranging between 70-71 in the southern portions of its range (Maine to North Carolina) to 73.8 in the Quebec-Nova Scotia region and 71.9 off Labrador. There is a distinct increase in vertebral number {x 72.9) in specimens of A. dubius taken off the Scotian Shelf and northward. Dorsal ray counts were also effective in separating specimens. Using this character alone, 92.3% of the specimens could be separated at a line of separation between 59 and 60 dorsal rays. Am.modytes americanus had relatively consistent dorsal ray counts throughout the entire geographic range (Table 3; 52-61, x 57.4); however, in comparison, number of dorsal rays for A. dubius was higher for specimens collected north of Nizinski et al Separation of Ammodytes amencanus and A dubius in western North Atlantic 245 Table 1 Frequency distribution of plicae counts for Ammodytes am^rieanus and A. dubius n four geographic regions. A. americanus A. dubius Quebec- Maine- New York- Quebec- Maine- New York- Labrador Nova Scotia Massachusetts N. Carolina Total Labrador Nova Scotia Massachusetts N. Carolina Total 106 1 1 107 1 1 108 1 1 2 109 1 — 1 2 110 1 1 1 2 5 111 2 5 — 1 8 112 5 1 7 3 16 113 2 2 1 8 13 114 5 4 8 16 33 115 7 4 7 7 25 116 10 4 3 11 28 117 10 7 8 10 35 118 18 9 5 12 44 119 13 2 6 6 27 120 8 2 6 5 21 121 8 9 6 5 28 122 8 3 6 2 19 123 1 2 4 7 14 124 1 2 8 5 16 1 2 2 5 125 1 2 3 3 1 — 8 12 126 1 1 2 — 2 10 14 127 4 — 2 25 31 128 1 — 3 14 18 129 1 3 4 24 32 130 2 2 8 33 45 131 1 2 6 20 29 132 2 3 9 18 32 133 — 1 1 22 24 134 1 6 8 23 38 135 1 2 2 19 24 136 — 3 3 9 15 137 1 8 3 6 18 138 — 4 4 5 13 139 4 4 1 3 12 140 1 2 2 1 6 141 — 2 — _ 2 142 _ — — — — 143 2 3 _ 1 6 144 1 1 1 3 145 1 — 1 146 _ — 147 1 1 X 117.6 117.4 117.7 116.1 117.4 131.8 136.0 132.6 131.2 132.1 N 103 57 79 103 342 27 48 62 244 381 C3 A , tnoficano s EZZ3 A dubius Lr 1 n Ln ri \ k ^ i k \K. 140 145 Figure 5 Histogram of counts of lateral plicae ior Ammodytes americanus (106-126, x 117.4) and A. dubius (124-147. X 132:1). 246 Fishery Bulletin 88(2), 1990 Table 2 Frequency distribution of vertebral count s for Ammodytes americaniis and A. dubiu. in four geographic regions. A. americanus A. dubius Quebec- Maine- New York- Quebec- Maine- New York- Labrador Nova Scotia Massachusetts N. Carolina Total Labrador Nova Scotia Massachusetts N. Carolina Total 62 2 2 63 1 8 9 64 6 1 11 18 36 fif) 4 3 9 40 56 66 17 8 14 16 55 67 24 21 16 7 68 68 29 16 18 3 66 2 16 18 69 15 4 7 4 30 3 12 39 54 70 1 1 2 1 5 5 23 87 115 71 3 1 22 64 90 72 2 3 3 23 31 73 2 15 1 4 22 74 6 14 20 7S 4 8 12 76 3 8 X 67.2 67.2 66.6 65.2 66.4 71.9 73.8 70.3 70.2 70.8 N 97 M 77 99 :«7 27 44 61 238 365 Table 3 Frequency distribution of dorsal ray counts for Ammodytes americanus and A. dubius in four geogra phic regions. A. americanus A. dubius Quebec- Maine- Ne w York- Quebec- Maine- New York- Labrador Nova Scotia Massachusetts N. Carolina Total Labrador Nova Scutia Massachusetts N. Carolina Total 52 1 1 53 1 2 1 4 54 1 2 3 2 8 55 1 4 5 7 17 56 12 5 9 7 33 1 1 57 15 16 8 8 47 _ _ 58 21 10 8 5 44 1 1 2 4 59 19 10 8 2 39 4 1 11 16 60 12 3 2 17 2 3 35 40 61 1 1 3 5 1 12 77 90 62 3 2 19 80 104 63 1 7 6 23 37 64 2 14 6 24 65 2 11 13 66 3 6 9 67 1 2 3 X 57.9 57.2 56.9 56.9 57.4 62.4 64.4 61.5 61.3 61.8 N s:-! ■17 48 37 215 20 42 42 235 341 Maine. Overall, dorsal ray counts for A. dithiu:^ varied from 56 to 67 rays. Specimens examined had means of 62.4 and 64.4 for Labrador and Quebec-Nova Scotia, respectively, compared with 61.5 in Maine-Massachu- setts and 61 .3 in the New York-North Carolina region. Anal rays, the least successful character for separat- ing specimens (Table 4), provided only 75.7% separa- tion of individuals (at a line of separation between 29 and 30 anal rays). Means for A. americanus ranged from 29 in southern regions to 29.7 off the coast of Quebec northward to Labrador, with total number of rays varying between 26 and 33 (overall x 29.4). Geo- graphic variation was again evident in anal ray counts for A. dubius. Counts averaged between 30 and 31 rays Nizinski et al.: Separation of Ammodytes amencanus and A dubius in western North Atlantic 247 Table 4 Frequency distribution of anal ray counts for Amtnodytes nmericaniis and ,4. dubius in four geograp lie regions. A. americanus A. dubius Quebec- Maine- New York- Quebec- Maine- New York- Labrador Nova Scotia Massachusetts N. Carolina Total Labrador Nova Scotia Massachusetts N. Carolina Total 26 2 2 27 4 2 6 6 18 28 10 4 15 3 32 4 4 29 23 13 8 15 59 1 27 28 30 27 18 8 9 62 2 17 76 95 31 13 9 7 4 33 6 2 15 99 122 32 6 9 3 11 6 6 11 32 55 33 1 1 4 17 5 26 34 3 14 17 35 1 7 8 X 29.7 29.7 29.0 29.1 29.4 32.1 33.4 30.8 30.6 31.1 N 84 48 49 37 218 22 46 44 243 .355 Table 5 Frequency distribution of pectoral ray counts for Ammodytes americanus and A. dubius in four geographic regions. A. americanus A. dubius Quebec- Maine- New York- Quebec- Maine- New York- Labrador Nova Scotia Massachusetts N. Carolina Total Labrador Nova Scotia Massachusetts N. Carolina Total 11 1 1 12 12 5 14 10 41 1 4 5 13 62 40 49 53 204 17 2 10 35 64 14 20 11 14 35 80 6 25 39 155 255 15 1 1 1 5 8 3 21 14 47 85 16 2 2 X 13.1 13.1 13.0 13.3 13.2 13.4 14.4 14.1 14.0 14.0 N 9.''. .^7 79 103 334 27 48 63 243 381 for specimens collected south of the Scotian Shelf (New York-North Carolina, x 30.6; Maine-Massachusetts, X 30.8) while specimens collected further north had mean values of 33.4 (Quebec-Nova Scotia) and 32.1 (Labrador). The number of pectoral rays varied little geograph- ically (Table 5) and was one of the least successful char- acters for separating the species. Modally, A. ameri- canus had 13 rays (11-15, x 13.2) while A. dubius had one more ray, 14 (12-16, x 14.0). Only 78.0% of the specimens could be separated solely on this char- acter (at a line of separation between 13 and 14 pec- toral rays). Gill rakers did not show pronounced geographic trends as was evident for vertebral counts (Table 6). For A. americanus, counts varied between 21 and 28, with means ranging from 23.7 (Labrador) to 24.9 (Mas- sachusetts-Maine). Mean values for .4. duhiu^s clustered around 26 or 27 gill rakers (x 26.6, range 23-31 over- all). At a line of separation between 25 and 26 gill rakers. 81.4% of the specimens could be separated into species using only this character. Ammodytes dubius showed geographic variation in all the meristic characters examined. Meristic values increased northward with the exception of Labrador. Offshore Labrador collections were sparse; therefore, sample size is only half as large as other regions. Ex- amination of additional material is necessary to deter- mine if meristic values for A. dubius collected off Labrador are consistently lower. The meristic characters used in this study separate the western North Atlantic sand lances into low- and high-count species. Each character, however, exhibited some overlap (ranging from approximately 3'^i for plicae counts to 49% for anal rays) and did not separate 100% of the specimens. Combinations of characters then became important in verifying identifications, par- ticularly for individuals with intermediate counts. 248 Fishery Bulletin 88(2). 1990 Table 6 Frequency distributio n of gill raker coun ts for Aminiidytes americanus and A. duhius in four geographic regions. A. americanus A. dubius Quebec- Maine- New York- Quebec- Maine- New York- Labrador Nova Scotia Massachusetts N Carolina Total Labrador Nova Scotia Massachusetts N. Carolina Total 21 1 1 99 9 5 1 15 23 30 18 7 13 68 4 1 5 24 28 22 19 29 98 5 1 4 7 17 25 18 7 23 22 70 4 3 8 37 52 26 9 5 14 15 36 4 9 18 70 101 27 5 5 10 4 10 7 66 87 28 1 1 2 9 13 6 50 71 29 7 — 9 16 .'?(! 1 1 3 5 31 2 2 X 23.7 23.8 24.9 24.7 24.3 25.2 27.4 26.1 26.6 26.6 N 88 57 7n 85 ,300 23 46 45 242 356 , ...... . ..„ 7^ - mai.idu^l i-lU (na.vrOu.iK • * I : : . . . 74 C' „„„„,..„.,. • • =• : •• > . : : : . . ••i" .;:■ . 1... . : * ■ V •••• CSB^ ••!■: S 68 •• ''-v:^r»-' ' ™a..a E .0 = 8 x.j.ij- ;§=._ 66 sss|.:«jr-'M#i s • = 8-1 r>8o||«.». . 105 no lis 1^0 125 110 lis 140 145 ,50 Numbei ol Plicae Figure 6 Scattergram of vertebral count plotted against number of lateral plicae for Atnmodytes americanus and A. dubius. Plicae and vertebral counts, when plotted against one another, were the most useful combination of char- acters for the entire geographic range producing vir- tually 100% separation (Fig. 6). The results of a PC A run on meristic data confirmed the usefulness of these characters. Since factor 1 ac- counted for 91% of the variation and all the meristic characters loaded heavily on this factor, factor 1 was the only factor used. Plicae had the highest component loading score (8.09) followed by dorsal rays (2.44), vertebrae (2.31), gill rakers (1.23), anal rays (1.03), and pectoral rays (0.45). The histogram of component scores (Fig. 4) produced a bimodal distribution with little overlap above or below a score of 1378 for the first set of individuals evalu- ated. When the scores were compared with the earlier tentative identifications (based only on meristics), only 10 (5 of each species) were either borderline or placed into the wrong species grouping. These 10 individuals were then reexamined and either the proposed iden- tification was verified (3 cases) or errors in the counts of certain meristic characters were found (7 cases). At most, 2% of the specimens were misclassified using the component scores. Of 54 additional specimens tested against the ec|uation, none were misclassified. This methodology, therefore, was consistent and reliable. Based on these results, diagnoses for the two species follow: Ammodytes americanus Diagnosis Ammodytes ajnericnmis tends to have lower counts than A. dubius: lateral plicae 106-126 (X 117.4); vertebrae 62-70, usually 64-69 {x 66.4); dorsal fin rays 52-61 {x 57.4), anal fin rays 26-33 (x 29.4), and pectoral fin rays 11-15 (x 13.2), with the most common dorsal, anal, and pectoral ray counts being 55-59, 27-31, and 13 rays, respectively; gill rakers on the first arch 21-28 (x 24.3). Six meristic characters in combination provide a good separation of this species from A. duhius. Number of plicae, however, is the single best character for distin- guishing >!. (nrieriranus from A. dubius in the western North Atlantic. Vertebrae plotted against plicae separ- ate individuals of the two species with little overlap (Fig. 6). Distribution Based on our samples, A. ampriranus ranges coastally from southern Delaware north through Nizinski et al : Separation of Ammodytes amencanus and A dubius in western North Atlantic 249 Labrador (Fig. 2). Previous studies reported its occur- rence in Chesapeake Bay (Richards 1982, Norcross et al. 1961); however, no specimens were examined from Chesapeake Bay during this study. This species occurs in shallow coastal waters and in protected bays and estuaries. Frequently, specimens were collected with seines or dipnets on sandy beaches in less than 2 m of water. Identification of individuals based only on locality (i.e., inshore vs. offshore) can be made with discretion, but some collections (approx- imately 20%), mainly from inshore stations, contain both species. Ammodytes dubius Diagnosis Ammodytes dubius shows geographic vari- ation in meristic features (Tables 1-6), and counts for this species are higher than those of A. americanus: lateral plicae 124-147 (i 132.1); vertebrae 68-76 (x 70.8) with 69-74 being the most common; dorsal fin rays 56-67 (x 61.8), anal fin rays 28-35 {x 31.1), and pectoral fin rays 12-16 {x 14.0). The majority of indi- viduals examined have >60 dorsal, >29 anal, and 14 pectoral rays. More gill rakers (23-31, i 26.6) are present on the first arch than in A. americanus. As described previously, plicae count is the best single distinguishing character; however, vertebrae plotted against plicae separate the species with little overlap (Fig. 6). Distribution Ammodytes dubius ranges from North Carolina to Greenland (Fig. 2) and is found in deeper, more offshore waters than A. americanus. Ammodytes dubius is occasionally found inshore but is generally taken in deeper, open waters. This species has a broad bathymetric distribution in coastal waters and ranges in depth from 7 to 80 m. Both species were taken together in approximately 20% of the samples exam- ined, with the majority of these samples being collected at inshore stations or around islands just offshore. Winters and Dalley (1988) reported co-occurrence in- shore oiA. americanus andyl. dubius, particularly in Nevirfoundland waters. Although none of the collections that we examined from Quebec-Nova Scotia contained both species, mixed collections were found in all other geographic areas. Several mixed collections (6 of 11 collections) contained only one or two individuals of A. dubius, with the remainder of the lot being A. americanus. Two of the 11 collections were all A. dubius except for one specimen of ^. americanus. Greenland Ammodytes Since doubts exist with regard to the number of species and the appropriate names of species occurring in Greenland, these speci- mens were analyzed separately using the PCA equa- tion and diagnoses used in analyzing and describing the species in other regions. Only those specimens (A^ = 51) with complete meristic data were included in the analysis. The majority of specimens fit our definition of A. dubius. Vertebral counts ranged from 66 to 75 (x 70.8); plicae counts from 124 to 156 (x 133.4); PCA scores from 1388 to 1636. Two specimens, however, fit our definition oi A. americanus; one specimen had 70 vertebrae, 123 plicae, and a PCA score of 1374, and the other had 68 vertebrae, 124 plicae, and a score of 1372. Discussion Results of this study demonstrate, in accordance with the majority of previous research (Richards et al. 1963, Leim and Scott 1966, Winters 1970, Scott 1972, Rich- ards 1982), that two species of sand lances occur in the western North Atlantic: an inshore species, A. ameri- canus, with low meristic features, and an offshore species, A. dubius, characterized by high meristics. No unequivocal method has been demonstrated previously to consistently identify individuals of A. americamis and A. dubius. Counts purportedly delineating the two species varied between studies, and considerable variation in meristic features for either one or both of these species has been reported. Addi- tionally, earlier studies relied principally on vertebral counts (with supporting data from dorsal and anal ray counts) but this approach was inadequate to accurate- ly identify all individuals. Plicae count, the most useful character in our study, was not used in the majority of previous studies. Furthermore, the lack of published detailed locality data in the majority of previous studies, consistent inshore/offshore designations, and a full understanding of migration patterns (Reay 1970) have added to the confusion in taxonomic status and ability to accurately identify species. With no clear-cut definition of the western North Atlantic species, identification problems have hindered previous investigators. A case in point is the reported discrepancy in geographic distribution of A. dubius and resultant interpretations of species. Since the earliest studies by Reinhardt (1838) and DeKay (1842), species designations have not been consistent. Reinhardt (1838) described A. dubius from Greenland. Jordan and Evermann (1896) later reported that this species oc- casionally reached as far south as Cape Cod. In the revision of the Ammodytidae by Duncker and Mohr (1939), A. dubius was reported from throughout the North Atlantic. However, the majority of recent re- searchers disagree and have limited the range of A. dubius to the western North Atlantic. In 1963, Richards et al. reported A. dubius as extending 250 Fishery Bulletin 88(2), 1990 further south to Virginia (37°N). Later Richards (1982), for no stated reason, reported a more Hmited distribu- tion for A. dubius, with no individuals found further south than the Scotian Shelf. Winters and Dalley (1988) also reported A. dubius as ranging no further south than Georges Bank. Restricting the southern limit of the range for A. dubius to Georges Bank creates several problems, in particular the explanation of a southern offshore, high- meristic form. Richards et al. (1963), Richards (1982), and Winters and Dalley (1988) recognized two popula- tions of A. americanus in southern waters: one an in- shore low-meristic form, and the other an offshore, intermediate- to high-meristic form. This conclusion seems to stem from Perlmutter's (1940) recognition of two offshore populations, one occurring north and one south of Cape Cod, which were designated as subspe- cies of A. fobitinus. In those studies (Richards et al. 1963, Richards 1982, Winters and Dalley 1988), A. americanus was probably chosen for the species name for southern offshore specimens because meristic values (vertebrae, dorsal and anal fin rays) were more similar to the counts they found for /I. americanus than for northern A. dubius. However, all of these studies reported and accepted the existence of a north-south and/or an inshore-offshore cline for at least one (but not always the same) species. Furthermore, the off- shore A. americanus occupied the habitat commonly inhabited by A. dubius. Also, the morphometric description given by Winters and Dalley (1988) for their southern offshore A. americanus was the same as that for their .4. dubius. Meristic differences between in- shore and oHshore Am modyfes in southern waters are consistent with a hypothesis of two distinct sympa- tric— but not necessarily syntopic— species, A. ameri- canus and A. dubius, occurring in these waters. Addi- tionally, meristic variation between northern and southern offshore forms supports the hypothesis of geographic intraspecific variation for A. duhi)is. The data are not consistent with a hypothesis recognizing only A. americanus in the southern region. Misconceptions regarding species designations of western North Atlantic Ammodytes are further com- pounded by the tendency of some authors (Richards et al. 1963, Winters and Dalley 1988) to create three groups— a low, intermediate, and high meristic group- within their data sets to explain the variation, instead of recognizing geographic variation in the offshore species. Recognizing three groups results in consider- able overlap between groups, thus adding to the diffi- culty of identifying species. Modal analysis (only for vertebrae; Winters and Dalley 1988) in some cases seems to identify modes that are not immediately obvious in the accompanying figures or data. The ac- curacy of identifying groups within collections, let alone individuals, is questionable using this method. The number of components and first estimates of the modes and standard deviation must be specified first. Addi- tionally, as Winters and Dalley (1988) point out, the reported high standard error for certain modes in- dicates that these means are estimated poorly. To eliminate confusion or guesswork surrounding identification of individuals of Ammodytes, it was necessary to devise an objective method of identifica- tion. PCA achieves this goal since this method iden- tifies patterns of variation between individuals without regard to the groups represented. PCA scores were utilized as a tool to summarize the data and to verify the usefulness of these meristic characters. A plot of the component scores of 332 individuals clearly showed two groups (Fig. 4). Further testing showed that these groups represent the two species. Thus, the two species can be separated using meristic characters instead of morphometric ratios of questionable validity (i.e., length-weight; see Scott 1972) with little overlap. Furthermore, by designating a component score boundary between the two groups, the two species are clearly defined and meristic variation can be delimited for each of the two species. This is particularly impor- tant for individuals with intermediate counts, especially those occurring in collections containing both species. These individuals were the hardest to identify, but defined limits for plicae, vertebral, and dorsal ray counts made identifications easier. To support our data analysis, a comparison with Perlmutter's 1940 data was made (Table 7; in the pres- ent study. Northern = Labrador and Quebec-Nova Scotia, Southern = Maine-Massachusetts and New York-North Carolina). The Perlmutter (1940) data were utilized because the data collected and geographic locations were similar. Designations of north-south as well as inshore-offshore locations were given also. Means for meristic data were surprisingly similar between the two studies. Perlmutter (1940) recognized north-south and in- shore-offshore clines in his data, yet misinterpreted these results. He designated several subspecies within a single worldwide polytypic form instead of recogniz- ing geographic variation within two parapatric species. Evidence for recognition of two species is that both species have been collected in sympatry in coastal waters, particularly in the Newfoundland area. It has been postulated that /I. dubius spawns inshore along the Newfoundland coast (Dalley and Winters 1987). The continued occurrence of high meristic individuals inshore suggests that these species are reproductive- ly isolated (Winters and Dalley 1988). Our data provide additional support for recognition of two parapatric species since there tend to be ecological differences between the species, in particular, habitat preference. Nizinski et al Separation of Ammodytes amencanus and A dubius in western North Atlantic 251 Table 7 Comparison of mean meristic values for characters used by | Perlmutter (1940) and in the present study. Perlmutter (1940) Nizinski et al.' X X Vertebrae^ A. americanus Northern 67.3 67.0 Southern 65.5 65,2 A. dubius Northern 73,5 72.0 Southern 70.0 70.2 Dorsal rays A. americanus Northern 57.0 57.3 Southern 55.3 56,9 A. dubi-us Northern 63.3 62,8 Southern 60.3 61.3 Anal rays A. americanus Northern 28.8 29.5 Southern 27.6 29.1 A. dubi-us Northern 32.3 32.1 Southern 29.6 30.6 Pectoral rays A. americanus Northern 13.2 13.1 Southern 13.7 13.3 A. dubius Northern 14.4 14.0 Southern 14.2 14.0 or and Quebec-Nova Scotia 'Northern = Labrac Southern = Maine- -Massachusetts and New York-North Carolina "Hypural plate excluded. The majority of individuals have been collected either inshore (^4. americanus) or offshore (A. dubius) with little syntopic occurrence. We conclude, therefore, that two distinct species, A. americanus and A. dubius, occur in the western North Atlantic Ocean. A similar case has been reported for the species of Ammodytes {A. personatus and A. hexapterus) found off Japan. Populations of morphologically and meris- tically similar individuals are difficult to distinguish from one another. Electrophoretic data, however, suggest that two sympatric, genetically distinct groups do occur among these Ammodytes species (Okamoto 1989). Perhaps a similar analysis is needed for the western North Atlantic ammodytids to confirm species designations. As previously mentioned, this study is a preliminary step toward understanding a more complex problem. Many other questions, including the taxonomic status of Greenland j4mmod2/te.s, remain unanswered. Speci- mens examined in our study (A'^ = 51) had high counts, similar to those obtained for A. dubius. But counts for these individuals were highly variable and did not agree completely with those of ^. dubius from the western North Atlantic. In particular, some specimens had low vertebral counts and extremely high plicae counts. Others with low plicae counts, albeit still in the range for A. dubius, tended to have high vertebral counts. Based on meristic combinations and computed PCA scores, the Greenland sand lance is most similar to A. dubius. This finding adds to the existing conflict over which species are found in Greenland. Published accounts indicate two species of sand lances occur in this region: high-meristic A. dubius, and a low-meris- tic, inshore species that has been called A. marinus by Winters and Dalley (1988) as were all western Atlan- tic A. am£ricanus. However, there is some question to the identity of the low-meristic sand lance occurring in Greenland. Counts for these specimens match those of A. americanus as well as the European A. marinus. The low-meristic Greenland species has reportedly been dipnetted and seined on shallow beaches and in pro- tected fjords. Ammodytes marinus, on the other hand, is described as the commonly occurring offshore, deeper-water species in the eastern North Atlantic (Richards et al. 1963, Wheeler 1969, Reay 1970), al- though this species has also been reported from inshore stations (Raitt 1934, Kirillov 1936). A limited number of A. marinus from the British Isles were examined; however, the data from these fish posed more questions than solved existing problems. Meristic features of these specimens spanned the ranges recorded for A. americanusl dubius, and mor- phological differences between these three species were not distinctive. The problem is further compli- cated since A. dubius and A. marinus are believed to be on opposite shores of the Atlantic (Reay 1970) yet both have been reported from Greenland. The tax- onomic status of Ammodytes occurring in Greenland waters cannot be resolved until meristic and geograph- ic ranges of European A. marinus are determined. Taxonomic confusion is not restricted to Atlantic species; the taxonomy of North ¥ac\i\c Ammodytes is problematical as well. Researchers agree that A. hex- apterus and A. personatus occur in Japanese waters (Kitaguchi 1979, Hashimoto 1984), with A. hexapterus reported as the more northern species and A. persona- tus as the more southerly species. But these species also are similar morphologically and have overlapping counts (vertebrae and dorsal and anal fin rays). Addi- tionally, high variability with two existing modes in the meristic data suggest the presence of two subpopula- tions in the southern A. personatus group (Hashimoto 252 Fishery Bulletin 88(2) 1990 and Kawasaki 1981, Hashimoto 1984). Isozyme differ- ences have indicated three separate genetic stocks (one A. hexapterus and two /I. personatus; Hashimoto 1984), and recent electrophoretic analysis confirms the ex- istence of a northern and southern population of ^4. per- sonatus as well as a possible new species or subspecies (Okamota et al. 1988, Okamota 1989). Morphological differences, however, are not significant between pop- ulations and between species, and it remains to be decided if designation of subspecies within this species complex is appropriate (Hashimoto 1984). Problems also occur in defining limits between North Pacific and western North Atlantic species o{ Ammo- dytes. Some investigators (Lindberg 1937, Andriashev 1954, Walters 1955, Richards et al. 1963) have sug- gested that A. hexapterus is circumpolar and extends from the Pacific into the Arctic and North Atlantic oceans. These workers have argued that/l. hexapterus is synonymous with A. americanus and/or >!. marinus. A limited number of A. hexapterus (N = 5) from Alaska and A. marinus (N = 15) from the British Isles were examined and no distinct morphological or meristic characters were found to clearly separate these species from those occurring in the western North Atlantic. Bothyl. hexapterus and A. mminus, however, are usual- ly characterized as occurring in deeper, offshore waters. Obviously, the entire genus is in need of revision. Acknowledgments For loan or donation of material, we thank Kenneth W. Able (Rutgers University), Eugenia B. Bohlke (ANSP), Bruce Burns (NMFS Narragansett Labora- tory), Barry Chernoff (FMNH), Michael P. Fahay (NMFS Sandy Hook Laboratory), William L. Fink (UMMZ), Karsten E. Hartel (MCZ), Don E. McAllister (NMC), John A. Musick (VIMS), Gareth Nelson (AMNH), J0rgen G. Nielsen (ZMUC), and Lou Van Guelpen (ARC). The NMFS Woods Hole Laboratory assisted in the collection of fresh material. A large file of radiographs of Ammodytes, made to search for vertebral anomalies, was loaned to us by the NMFS Oxford Laboratory. James D. Felley provided statis- tical advice. Frank P. Almeida, NMFS Woods Hole Laboratory, supplied the data for Figure 1. Figure 3 was drawn by Keiko Hiratsuka Moore. Douglas W. Nelson (UMMZ) provided a xerox copy of Perlmutter's unpublished dissertation. Drafts of the manuscript were reviewed by Kenneth W. Able, Laurence Buckley, Michael P. Fahay, Irv Kornfield, Thomas A. Munroe, Theodore W. Pietsch, and Sarah W. Richards. G.H. Winters critically reviewed several drafts of the manuscript and many of his comments were useful in revising the paper, although we do not agree on the interpretations of some aspects of geographic variation in Ammodytes. Citations Andriashev, A. P. 19.54 Fishes of the northern seas of the USSR. Zool. Inst. Akad. Nauk Moscow Opred. Faune SSSR 53, 566 p. [transl., Israel Prog. Sci. Transl., IPST Cat. No. 836, 617 p.|. Backus, R.H. 1957 The fishes of Labrador. Bull. Am. Mus. Nat. Hist. 1 13(4): 273-338. Bruun, A. 1941 Observations on North Atlantic fishes. 2, The Ammo- dytes lancea group. Vidensk. Medd. Dan. Naturhist. Foren. 104:329-340. Dalley, E.L., and G.H. Winters 1987 Early life history of sand lance {Ammodytes) with evi- dence for spawning oi A. duhius in Fortune Bay, Newfound- land. Fish. Bull., U.S. 85:631-641. DeKay, J.E. 1842 Zoology of New- York, or the New-York fauna, Pt. IV. Fishes. W. and A. White and .J. Visscher, Albany. 415 p. Duncker, G., and E. Mohr 1939 Revision der Ammodytidae. Mitt. Zool. Mus. Berl. 24(1): 8-31. Hashimoto. H. 1984 Population structure of the sandeel around Japan. Bull. .Ipn. See. Sci. Fish. 50(8): 1357- 1365. Hashimoto, H., and T. Kawasaki 1981 Population studies of the sand ee\, Ammodytes personatus (Girard), in Sendai Bay and its neighborhood. Tohoku J. Agric. Res. 31(4):173-197. Jensen, A.S. 194 1 On subspecies and races of the lesser sand eel {A mmodytes lancea s. lat.). A contribution to the discussion of the species problem in fishes. K. Dan. Vidensk. Selsk.. Biol. Medd. 16(9), 33 p. 1944 On specific constancy anil segregation into races in sea- fishes. K. Dan. Vidensk. Selsk., Biol. Medd. 19(8), 19 p. Jordan, D.S., and B.W. Evermann 1896 The fishes of North and Middle America. Bull. U.S. Nat. Mus. 47, pt. 1, 9.54 p. Kirillov. F.N. 1936 The sand-eel of Novaya Zemlya {Ammodytes marimis Raitt). Dokl. Akad. Nauk SSSR 4(13), 5:233-235. Kitaguchi, T. 1979 A taxonomic study of sand lances (genus 4 7«morf!/(f.s) in the waters of northern coast of Hokkaido, Japan. Sci. Rep. Hokkaido Fish. Exp. Stn. 21:17-30. Leim, A.H., and W.B. Scott 1966 Fishesof the Atlantic coast of Canada. Fish. Res. Board Can. Bull. 155, 485 p. Lindberg, G.U. 1937 On the classification and distribution of sandlances genus Ammodytes (Pisces). Bull. Far East. Branch Acad. Sci. USSR 27:85-93 [in Russian. Engl, transl.]. Macer, C.T. 1966 Sand eels (Ammodytidae) in the south-western North Sea; their biology and fishery. Fish. Invest. U.K. Ser. II, 24(6), 55 p. McAllister, D.E. 1960 List of marine fishes of Canada. Bull. Natl. Mus. Can. 168, 76 p. Nizinski et al Separation of Ammodytes amencanus and A dubius in western North Atlantic 253 Meyer, T.L., R.A. Cooper, and R.W. Langton 1979 Relative abundance, behavior, and food habits of the American sand lance, Ammodytes americanus, from the Gulf of Maine. Fish. Bull., U.S. 77:243-253. Nelson, G.A., and M.R. Ross 1987 The population dynamics of sand lance {Ammodytes dubius) in the Northwest Atlantic. I. Abundance and distribu- tion. Unpubl. rep., Natl. Mar. Fish. Serv., NOAA, Woods Hole. MA 02543, 66 p. Norcross, J.J., W.H. Massmann, and E.B. Joseph 1961 Investigations of inner continental shelf waters off lower Chesapeake Bay. Part II. Sand lance larvae, Ammodytes lunericanus. Chesapeake Sci. 2(l-2):49-59. Okamoto, H. 1989 A genetic comparison of sympatric populations of sand lance {genus Ammodytes) from the region east of Cape Soya, Japan. Can. J. Fish. Aquat. Sci. 46:1945-1951. Okamoto, H., F. Yamazaki, and S. Mishima 1988 Genetic divergence among sand-lance Ammodytes per- sonatus populations in Japan. Bull. Jpn. Soc. Sci. Fish. 54(8): 1297-1304 [in Jpn., Engl, sumni.]. Overholtz. W.J., and J.R. Nicolas 1979 Apparent feeding by the fin whale, Balaenoptera physa- lus, and humpback whale, Megaptera novaengliae, on the American sand \a.nce, Ammodytes americanus, in the northwest Atlantic. Fish. Bull., U.S. 77:285-287. Perlmutter, A. 1940 Variation of American North Atlantic marine fishes correlated with the environment. 7. Sand eels of the genus Ammodytes. D.Sci. diss., Univ. Michigan, p. 16-20. Pietsch, T.W., and C.P. Zabetian 1990 Osteology and relationships of the sand lances (Teleostei: Ammodytidae). Copeia 1990:78-100. Powers, K.D., and E.H. Backus 1987 Energy transfer to seabirds. In Backus, R.H. (ed.), Georges Bank, p. 372-374. MIT Press. Cambridge. Raitt, D.S. 1934 A preliminary account of the sandeels of Scottish waters. J. Cons. Perm. Int. Explor. Mer 9(3):365-372. Reay, P.J. 1970 Synopsis of biological data on North Atlantic sand eels of the genus Ammodytes {A. tobianus, A. dubius. A. ameri- canus. and A. marinus). FAO Fish. Synop. 82, 47 p. Reinhardt, J. 1838 Icthyologiske bidrag til den Gronlandske fauna. K.Dan. Vidensk. Selsk., Nat. og Math. Afhand. 7:131-132. Richards, S.W. 1982 Aspects of the biology of Ammodytes americanus from the St. Lawrence River to Chesapeake Bay, 1972-75, including a comparison of the Long Island Sound postlarvae with Am- modytes dubtus. J. Northwest Atl. Fish. Sci. 3:93-104. Richards, S.W., A. Perlmutter, and D.C. McAneny 1963 A taxonomic study of the genus Am.modytes from the east coast of North America (Teleostei: Ammodytes). Copeia 1963:358-377. Scott, J.S. 1972 Morf)hoiogical and meri.stic variation in northwest Atlan- tic sand lances {Amntodytes). J. Fish. Res. Board Can. 29:1673-1678. Scott, W.B., and M.G. Scott 1988 Atlantic fishes of Canada. Can. Bull. Fish. Aquat. Sci. 219, 731 p. 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. Walters, V. 1955 Fishes of the western Arctic America and eastern Arc- tic Siberia. Taxonomy and zoogeogi'aphy. Bull. Am. Mus. Nat. Hist. 106:255-368. " Wheeler, A. 1969 The fishes of the British Isles and northwest Europe. MacMillan, London. 613 p. Winters, G.H. 1970 Meristics and morphometries of sand lance in the New- foundland area. J. Fish. Res. Board Can. 27:2104-2108. 1983 Analysis of the biological demographic parameters of northern sand lance, Ammodytes dubius, from the New- foundland Grand Bank. Can. J. Fish. Aquat. Sci. 40:409-419. Winters, G.H., and E.L. Dalley 1988 Meristic composition of sand lance {Ammodytes spp.) in Newfoundland waters with a review of species designations in the Northwest Atlantic. Can. J. Fish. Aquat. Sci. 45: 516-529. Appendix 1: Material examined Abbreviations of institutions AMNH American Museum of Natural History, New York ANSP Academy of Natural Sciences, Philadelphia ARC Atlantic Reference Centre, Huntsman Marine Laboratory, St. Andrews, New Brunswick FMNH Field Museum of Natural History, Chicago MCZ Museum of Comparative Zoology, Harvard University, Cambridge NMC National Museum of Natural Science, Ottawa, Canada UMMZ University of Michigan Museum of Zoology, Ann Arbor, MI USNM National Museum of Natural History, Smithsonian Institution, Washington, DC VIMS Virginia Institute of Marine Science, Gloucester Point, VA ZMUC Kobenhavns Universitet Zoologisk Museum, Copenhagen, Denmark Material is listed by species, geographically from north to south. Ammodytes americanus Labrador 177 specimens (49-157 mm SL) from nine collections. FMNH 31106-31116,31118-31132(25, 49-126) Anatalak Bay, Nain, Labrador; July-Aug. 1928. MCZ 12471 (8, 98-143) Labrador. MCZ 12480 (52, 80-145) Labrador. MCZ 12481 (17, 105-157) Labrador; Packard; 1865. MCZ 12482 (30, 85-155) Labrador. MCZ 51802 (4, 69-76) Hamilton Inlet, Collingham's Cove; Blue Dolphin; 7 Feb. 1952. USNM 165262 (1, 96) Indian Cove, Assyis Run, St. Lewis Sound, 52°15N 55°04'W; Blue Dolphin; 12 July 1949. USNM 165300 (35, 96-147) Labrador, Hare Harbor, 53°43N 56°46'W: Blue 254 Fishery Bulletin 88(2), 1990 Dolphin; 2 July 1950. USNM 165371 (5, 92-104) Labrador, mouth of Tessiujarsuk near Nain; Blue Dolphin: 8 Aug. 1951. Newfoundland, IMova Scotia, and Quebec 90 specimens (52-147 mm SL) from 11 collections. MCZ 49725 (1, 147) Newfoundland, Nolan's Beach, Sweet Bay, BonavisUi Bay; R.H. Backus; 7 June 1948. NMC 73-380 (3, 121-140) Newfoundland; Bonne Bay at Norris Point Wharf; B.E. Bowen; 9 July 1973. NMC 66-174 (1, 80) North New- foundland, Grigiiet Harho'r at Griguet, 5r32.5'N 55°28.5'W; D.E. McAllister and W.H. Vliet; 0-1 m; 18 June 1966. NMC 59-266 (38, 96-140) Quebec, Perce on tip of Gaspe Peninsula, Gaspe Co. 48°30'N 64°15W; D.E. and N.A. McAlhster; 0-1 m; 10 June 1959. NMC 81-0888 (11, 76-101) Quebec, Fleuve St. Laurent, 49''35'N 67°25'W; J.D. Dutil and B. Legare; 24 June 1981. USNM 88849 (1, 131) Quebec, north side of Matamer River; Amory-Bowman Expedition; 1927. ARC 8600807 (12, 97-124) Tabusintac estuary (giilley); M.J. Dadswell and G, Melvin; 0-1 m; 20 July 1977. ARC 8600809 (1, 90) New Brunswick; New River Beach; M.J. Dadswell et. al.; int.ertidal; 16 Sept. 1977. ARC 8600803 (1,114) New Brunswick, St. Andrews; Oct 1949. UMMZ 193387 (20, 52-63) New Brunswick, Miramichi, inside West End Bay du Vin Island; R.A. McKenzie; 22 July 1942, ARC 8600802 (1, 125) Prince Edward Island; Aug. 1953. Maine and Massachusetts 285 specimens (60-168 mm SL) from 14 collections. FMNH 17181-17185 (5, 141-168) Boothbay Harbor, ME; July 1931. MCZ 44877 (2, 95-102) MA, Barnstable Harbor; J. Morin. MCZ 57161 (14, 60-121) Chatham, upper Cockle Cove off Mill Creek Rd.; K.E. Hartel. MCZ 12464 (15, 86-123) Yarmouth; L. Agassiz. UMMZ 140537 (15, 106-133) MA, beach near Scituate between second cliff and third cliff; C,L. and L.C. Hubbs; 0-3 ft; 17 June 1928. USNM 302255 (126-I-, 7,5-105) Nahant, East Point, Marine Science Institute; B.B. Collette, BBC 1759; 0-1 m; 7 Aug. 1981, USNM 3022,56 (1 1, 76-108) 41° 18'N 70°28'W; Gloria Michelle Cr. 8592, Sta. 57; 12 Sept. 1985. USNM 83720 (15, 91-128) MA, Truro; W.C. Kendall; 14 Sept. 1892. USNM 36925 (20, 101-119) MA, Bass Rocks, New Gloucester; A.H. Clark; Oct. 1879. USNM 132092 (20, 73-115) ME, Wood Lsland; Grampus; surface; 13 Oct. 1915. USNM 73499 (3, 97-105) MA. "The Cut" and Pavillion Beach, Gloucester; Grampus; 29 July 1895. USNM 302254 (10, 79-112) Newberry Port, MA, Merrimack R. USNM 302253 (22, 73-122) Martha's Vineyard Soimd, MA. Fahay collection (7, 80-130) Nauset Marsh, Cape Cod; Fahay and Able. IMew Yorl< to Virginia 199 specimens (83-137 mm SL) from three collections. AMNH 36590 (72, 89-120) NY, Suffolk Co., Sunken Meadows' State Park, east of Jetty Seine; G.J. Nelson et al.; 19 May 1977. AMNH 37692 (30, 83-118) Suffolk Co., Gardener's Island, tidal outlet, Bostwick's Pond; D.E. Rosen et al.; 22 Apr. 1977. ANSP 165786 (97, 83-137) approx. 2.5 naut. mi. SE of Little Egg Inlet, along axis of lump-5252; C.B. Milstein. CBM 72-145. Ammodytes dubius Labrador 14 .specimens (121-203 mm SL) from two collections. USNM 165263 (11, 121-187) Pack's Harbor, 53°54N 56°59W; Btiw Dolphin; 24 July 1949. USNM 165264 (3. 143-203) Hare Harbor; Blue Dolphin; 2 July 1950. IMewfoundland, IMova Scotia, and Quebec 72 specimens (107-244 mm SL) fn.m five colkrtii.ns. NMC 6 1-763 (16. 178-209) off New- foundland, Grand Banks, 45°06'30"N 49''0r00"W; A.T. Canwron; 14 Oct. 1964. NMC 64-764 (11, 169-244) off Newfoundland, Grand Banks, 54°02'50"N 49°0415"W; .4.7'. Cameron; 5 Oct. 1964. VIMS 1224(30. 192-230) 44°050.5'N60°0430"W; .4.7. Cameron Cr. 176, Sta. 16. UMMZ 201715 (14. 107-195) Nova Scotia, caught ..n Mid- dle Bank of Atlantic Ocean, 44°35'N 60°25'W; S.T. Venosta; Nov. 1938. ARC 8600808 (1, 155) Grand Bank, SE shoal water; G. Somer- ville; 8 July 1954, Maine and Massachusetts 264 specimens (77-253 mm SL) from 15 collections, MCZ 4(1699 (1, 177) N, end Stellwagen Bank; F. Buinelt. MCZ 62955 (3, 77-102) Martha's Vineyard, just off south coast between Chilmark and Edgartown; Gloria Michelle Cr. 8592, Sta. 53-56. VIMS 2293 (2, 249-2.53) 42°36'N 66° 17' W; Albatross /yCr. 6911, Sta. 214. USNM 67636 (4, 83-98) MA, North Truro, Cape Cod; Grampus; 9 June 1896. USNM 163734 (1, 256) Georges Bank. 41°2rN 67°33W; R.L. Wigley; 1950. USNM 302242 (6, 80-128) 40°59'N 69°(26-28)'W; Delaware 11 Cr. 8207, Sta. 91; 44 m; 13 Oct. 1982. USNM 302245 (3, 112-122) 40''49'N 69°04'W - 40°48'N 69°05'W; Delaware II Cr. 8207, Sta. 92; 69 m; 14 Oct. 1982. USNM 302247 (15, 98-123) 41°21'N 67°2G'yN; Delaware II Cr. 8207, Sta. 123; 38 m; 21 Oct. 1982. USNM 302258 (55, 87-125) 41°28'N 67°46'W; Delaware II Cr. 8207, Sta. 124; 40 m; 21 Oct. 1982. USNM 302250 (26, 86-142) 4r33'N 67°58W; Delaware II Cr. 8207. Sta. 125; 30 m; 21 Oct. 1982. USNM 302240 (17, 96-140) 4r38'N 68°27'W; Delaware II Cr. 8207, Sta. 128; 60 m; 22 Oct. 1982. USNM 302249 (10, 87-128) 41°35'N 66°58'W; Delaware II Cr. 8207, Sta. 143; 58 m; 23 Oct. 1982. USNM 302248 (50, 121- 175) 41°5rN 70°29'W; Gloria Michelle Cr. 8592, Sta. 3; 13 m; 6 May 1986. USNM 302246 (45, 59-184) 41°35'N 69°56'W; Gloria Michelle Cr. 8592, Sta. 77; 7 m; 18 May 1986. USNM 302251 (46, 100-161) 42°05'N 70°08W; Gloria Michelle Cr, 8592, Sta. 91; 10 m; 18 Sept. 1985. USNM .302257 (17, 84-108) 41°18'N 70''28W; Glirria Michelle Cr. 8592, Sta. 57; 12 Sept. 1985. IMew York to IMorth Carolina 415 specimens (54-249 mm SL) from 24 collections, ANSP 36662-73 (13, 146-207) NJ, Carson's Inlet, Cape May Co.; R.J. Phillips; 1 Nov. 1908. VIMS 2294 (3, 175-193) 41''2,5'N 69°00'W;^W«i/ro.ss/yCr. 7002. VIMS 2295(19, 194-249) :«°03'N 73°41'W; Albatross IV Cr. 6908, Sta. 161. VIMS 2390 (1. 217) 28°28N 74°37W; Sea Breeze. Sta. T350. VIMS 2807 (3, 98-109) 37°51'N 75°08'W; McEachran. VIMS 2808 (9, 140-178) 38°07'N, 74°44W; Albatross IV Cr. 6920, Sta. 2. VIMS 7377 (1, 186) 37°0rN 74°57'W, Virginia Shelf; Captain Wool. VIMS 7751 (4, 120-190) 40°47'N 69°40'W; ^/ia(ross A' Cr. 8005, Sta. 36, ANSP 165785 (153, 92-138) approx. 2.5 naut, mi. SE of Little Egg Inlet, along a.xis of lump-.52.52; C.B. Milstein. CBM 72-145. UMMZ 212624 (50, 100-185) NY, Plum Beach, Long Island; A. Perlmutter; 20 Dec. 1937. USNM .302243 (1, 159) 40°00'N 73°47'W - 40°0rN 73°44'W; Delaware II Cr. 8207, Sta. 44; 29 m; 8 Oct, 1982. USNM 302244 (3, 119-143) 41°1 UN 71°24'W-41°13'N71°24'W;DW«il/6(!(ross/V' Cr. 8809, Sta. 60; 5.7 fm; 19 Sept. 1988. USNM 302232 (15, 144-171) 39°24'N 73°29'W - 39°25'N 73°26'W; Albatross IV Cr. 8809, Sta. 95; 38 m; 23 Sept. 1988. USNM 302233 (25, 117-146) 38°47N 74°08'W - 38°49'N 74°09'W; .Albatross IV Cr. 8809. Sta. 84; 44 m; 22 Sept. 1988. USNM 302234 (1, 132) 40°40'N 72°58'W - 40°4rN 72°56W; Albatross IV Cr, 8809, Sta. 130; 15 m; 26 Sept. 1988, USNM 3022,35 (20, 144-171) 39°23'N 73°43'W - 39°24'N 73°42'W; Albatross IV Cr. 8809. Sta. 94; 32 m; 23 Sept. 1988. USNM 302236 (2, 121-122) 37°(04-06)'N 75° 18' W; Albatross IV Cr. 8809, Sta. 56; 27 m; 18 Sept. 1988. USNM 302237(1, 125)39°19'N 74°(21-23)'W; Albatross IV Cr. 8809. Sta, 88; 10 m; 22 Sept. 1988. USNM 302238 (24, 100-138) .36''46N 75°24'W - 36°47N 75°22'W; Albatross IV Cr. 8809, Sta. 54; 23 m; 18 Sept, 1988, USNM .302239 (18, 104-189) Nizinski et a\ Separation of Ammodytes americsnus and A dubius in western North Atlantic 255 38°47N 73°15'W - 38°48'N IZ'^VX'^; Albatross IV Cr. 8809, Sta. 3; 80 m; 13 Sept. 1988. Ammodytes sp. Greenland .5.5 specimens (48-192 mm). USNM 87373 (5, 48-55) P.H. Sorenson. MCZ 63338 (4, 63-77) 64°47.3'N 30°37.4-W Endevor Cr. 133, Sta. 5. MCZ 63339 (2, 55-59) 65°23'N 29°22.8'W Endevor Cr. 133. Sta. 12. ZMUC various lots (44, 57-192): 61150 store Hellefiskebanke, 24 Hum v.s.v. f. Rifkol, smaa sten; 10 July 1912: Beskyttemi. P611.52. P61153. P611.54. P61158; Gtxihavn: Porsild. P61159; Bunden af. Kapisigdlitfjord 0f lak seelven Aale- haandvaad; Dann. Sta. 2325. P61166; Greenland 66°44N .53°W; Dana. Sta. 2349. P61312; Pr0vens Havn, NV Gr0nland; Finn Salo- monsen. P61322. P61324. 73; Gr0nland. 76. 78. 79. 84; 23/2, 49; N. 101. 89; Gr0nland. 90d. 67°53'N 54°02W. 121. 160; Jakobshavn; P. Muller. 161; Jakobshavn; P. Muller. 162; Claus- havn; Ad. Jensen. 163; Disko Biigten. 165; Marrak i Sydostbugten, Opskyllet i Beg. af., Aug. 1906; Ad. Jensen. 169; Tkamiut i Sydost- bugten; Lohmann. 170; Godhavn; Porsild. 176; Sukkertoppen; Bistnip. 178; Sukkertoppen; Bistrup. 179; Sukkertoppen; Bistrup. 180; Sukkertoppen; Bistrup. Uncat (15 spec); Kigdiuf Iluat, Gr0n- land, 63°53N 51°22W. Ammodytes marinus USNM 302259 (2, 115-126) Brunswick Wliarf; Thames and Wheeler. USNM 302260 (8, 88-1481 Littlebrook and W. Thurrock; Thames. USNM 108812 (5, 160-171) British Isles; Firth of Forth, 56°12N 2°43'W. A. hexapterus USNM 2U7553 (1, 159) AK, Cook Inlet, N. Barren Islands, S of Kenai Peninsula, 59°09N 1.52°17W; Yaquina; 60 fm; 30 July 1963. USNM 235313 (2, 91-143) AK, Aleutian Islands, NE Adak Island, southside of Clam Lagoon; J. Rosewater and P.R. Greenhall; 8 June 1979. USNM 266655 (2, 123-129) AK; Miller Freeman. Sta. B44; 14 Aug. 1982. Abstract. — We computed corre- lations between various population estimates for Central California chinook salmon Oncorhynchus tshawytscha and both freshwater and maiine environmental variables using methods that account for intraseries correlation in a more accurate and conservative way than those used previously. These indicated a nega- tive influence of ENSO (El Nino- Southern Oscillation) conditions in the year during which most of these fish are caught or leave the ocean to spawn. Although freshwater envi- ronmental influences have been pre- viously proposed on the basis of correlation analysis, and have been demonstrated using direct survival estimates based on marked fish, they were not detectable using correlation techniques that accurately account for intraseries correlation. There was also weak evidence for an influ- ence of conditions associated with a negative upwelling index at the time chinook salmon enter the ocean. However, because these conditions are associated with high river flows in addition to oceanographic effects, these correlations may merely result from the influence of freshwater flows. To further describe oceano- graphic influences we computed the principal components of upwelling index, sea level height, and sea sur- face temperature. The first principle component, which reflected the effects of ENSO conditions in the equatorial Pacific during the previ- ous winter, was significantly corre- lated with chinook salmon abundance in their final year, and marginally correlated with abundance during the first ocean summer. This work demonstrates new techniques for reducing spurious correlations and' the practical difficulties involved in sorting out the multivariate influ- ences on populations subject to re- mote forcing through oceanographic and meteorological conditions. Determination of Factors Affecting Recruitment of Clninool< Salmon Oncorhynchus tshawytscha in Central California Robert G. Kope Department of Wildlife and Fisheries Biology University of California, Davis, California 95616 Present address: Tiburon Laboratory, Southwest Fisheries Center National Marine Fisheries Service, NOAA, 3150 Paradise Drive Tiburon, California 94920 Louis W. Botsford Department of Wildlife and Fisheries Biology University of California, Davis, California 95616 Manuscript accepted 23 January 1990. Fishery Bulletin, U.S. 88:257-269. Understanding the influences of envi- ronment on survival and abundance of chinook salmon Oyicorhynchus tshaurytscha is necessary for manage- ment of both fisheries and surface water in central California. The salmon fisheries are managed by allocating predicted preseason abun- dance between catch and spawning escapement, and flow rates in central California rivers are currently man- aged to allow adequate flows for spawning and rearing of chinook salmon. Better understanding of en- vironmental influences on salmon populations could improve the ability to forecast preseason abundance and more clearly define the impacts of water management on salmon abundance. Chinook salmon spawn in fresh- water, but spend most of their lives in the ocean. In California's Central Valley, there are four distinct runs of chinook salmon named for the timing of their spawning migrations: fall, late-fall, winter, and spring. Typically over 90% of the spawning salmon are from the fall run, and over 80% spawn in the Sacramento River and its tributaries (USFWS 1987). Fall run chinook salmon migrate up- stream July through November, and spawn October through January. Eggs incubate in the gravel October through March, and young fish rear in the streams and migrate down- stream December through June (USFWS 1987). Most fish spawn at age 3, with some spawning at ages 2 and 4, and a few fish at age 5 (Dett- man et al. 1986, Reisenbichler 1986, Kope 1987). The precocious 2-year- old spawners are primarily males and are called "jacks." While in the ocean. Central Valley chinook salmon are harvested by commercial and recreational troll fisheries. Catches in both fisheries are dominated by 3-year-old fish (Denega 1973, Kope 1987). In addition to natural production, several hatcheries and spawning facilities produce fish to enhance the fisheries and mitigate for losses of spawning and rearing habitat result- ing from water development projects (Fig. 1). The contribution of hatchery production to spawning runs has in- creased in recent years, and has been estimated to be approximately 15% of the total run in the upper Sacra- mento River (Reisenbichler 1986), 78% of the run in the Feather River, and 87% of the total in the American River (Dettman and Kelley 1986). Numerous studies have examined relationships between environmental 257 258 Fishery Bulletin 88(2). 1990 Figure 1 Chinook salmon spawning streams and fish facihties of California's Central Valley. CNH = Coleman National Hatchery; RBDD = Red Bluff Diversion Dam; FRH = Feather River Hatchery; NH = Nim- bus Hatchery; MRFI = Mokelumne River Fish Installation; MRFF = Merced River Fish Facility; SWP = State Water Project pump- ing plant: CVP = Central Valley Project pumping plant. variables and recruitment or abundance of Pacific sal- monids through correlation analysis. Many of these studies have focused on chinook salmon (e.g., Van Hyn- ing 1973, Barton 1980), and a few have dealt with the stock of chinook salmon that spawns in California's Central Valley. Most of these studies have detected en- vironmental influences on recruitment; but in the case of Central Valley chinook salmon, reasonably long time- series of recruitment data have been lacking. Some analysts have therefore used estimates of spawning escapement as a proxy for past recruitment (Dettman et al. 1986, Dettman and Kelley 1986, Reisenbichler 1986, USFWS 1987). In general, these studies have reported positive correlations with flow variables and negative correlations with temperature during the period when the spawners were migrating downstream as smolts. Because of a negative correlation between flow variables and temperatures (Dettman et al. 1986), it is impossible to distinguish between effects of temperature and effects of flow using existing data. Stevens and Miller (1983) used two time-series of smolt estimates based on smolt catches in the delta (Fig. 1), and showed a positive correlation with flows in the previous winter. The most convincing studies used estimates of mortality based on the difference in sur- vival between tagged fish released above the delta and tagged fish released below the delta to show a positive influence of flow rates in the delta and negative in- fluences of temperature and diversions (Kjelson and Brandes 1989, Kjelson et al. 1981, USFWS 1987, Dett- man et al. 1986). In contrast to those of freshwater, potential marine influences on Central Valley chinook stocks have not been explored. Oceanographic conditions off the Cali- fornia coast vary dramatically providing potentially important interannual differences in predators, prey, transport, and ocean temperature during the marine phase of chinook salmon life. The normal pattern of cir- culation in the northeastern Pacific involves north- westerly winds of varying strength along the coast of California in the spring and summer months which rein- force the geostrophic flow of the California Current and transport surface waters offshore (Mysak 1986, Hickey 1989). This causes varying degrees of upwelling of deeper water and longshore advection of subarctic water along the coast of California and Oregon, which are associated with colder temperatures and depressed sea level heights. The amount of upwelling is estimated by an upwelling index based on winds computed from measured atmospheric pressure gradients (Bakun 1975). Occasional El Nino-Southern Oscillation (ENSO) events, characterized by a warming of equatorial Pacific waters, have a dramatic effect on California coastal waters. These effects may result from either coastal trapped waves which propagate poleward from the Equator, or an atmospheric teleconnection between the Equator and midlatitudes (Emery and Hamilton 1985, Mysak 1986). The former would cause higher sea surface temperatures and sea level heights, but would not necessarily affect Bakun's upwelling index (Enfield and Allen 1980, Emery and Hamilton 1985, Mysak 1986). The latter involves an increase in strength of the Aleutian low-pressure system and a weakening of the north Pacific high-pressure system. Changes in the atmospheric pressure gradients, associated with ENSO events, weaken the northerly winds which results in less upwelling, warmer sea temperatures, higher sea level, and a lower upwelling index (Norton 1987, Mysak 1986). ENSO events and anomalous strengthening of the Aleutian low pressure system may occur at the same time, but each also occurs alone (Emery and Hamilton 1985, Mysak 1986). Marine and freshwater influences on California Cen- tral Valley chinook salmon could be confounded. Large- scale oceanographic and meteorological conditions in Kope and Botsford: Recruitment of Oncorhynchus tshawytscha in central California 259 the Pacific have been related to precipitation in Califor- nia, which could in turn influence flow rates. Markham and McLain (1977) found a positive correlation between California rainfall and sea surface temperature off the coast of Washington. However, rainfall in California is only weakly related statistically to equatorial condi- tions associated with ENSO events (Douglas and Englehart 1981, Cayan and Peterson 1989, Redmond 1988). Environmental influences on fish populations are often explored by computing correlations between en- vironmental and population time-series; however, these are computed and the significance of resulting correla- tions are evaluated in a variety of ways. In attempting to detect environmental influences by computing corre- lations, the probability of obtaining a significant result is often artificially inflated by two problems: (1) intra- series correlation and (2) multiple tests (cf. Walters and Collie 1988). The latter arises from the fact that many environmental series can be tested for correlation with the population series; hence the probability of falsely identifying a significant relationship is greater than the probability level of each test. Methods exist for deal- ing with both problems, but we focus on the former here [see Hollowed et al. (1987) and Drinkwater and Myers (1987), for examples of dealing with the latter]. The former, intraseries correlation, is a lack of statis- tical independence of observations, which implies that the true number of degrees of freedom that should be used in determining the significance of a computed cor- relation is less than the number of points in the series. There are two steps that can be taken to deal with this problem: (1) removing as much of the intraseries cor- relation as possible, and (2) accounting for the remain- ing correlation in the computation of confidence limits. Removal of intraseries correlation can be accom- plished by arbitrarily "pre-whitening" the series (i.e., filtering the series so that correlation between adja- cent points is removed) (Box and Jenkins 1976, Fogarty 1988), but is more directly justified when there is a physical basis for doing so and the new series repre- sents a meaningful quantity. In the case of an abun- dance or catch series, one can compute a recruitment time-series from the original series using deconvolu- tion (Kope and Botsford 1988). Deconvolution is a pro- cedure based on the observation that catch or abun- dance is a sum over age-classes which is essentially a weighted sum over past recruitment (i.e., a moving average), in which the weighting factors are the (assumed constant) proliabilities of survival to each age. If the abundance or catch series is established using a size limit, rather than an age limit, the numbers at each age that are larger than the size limit must also be accounted for in the weighting factors. Deconvolu- tion inverts this summing process to give recruitment in terms of catches or abundances. Deconvolution pro- vides two advantages over using the catch or abun- dance series without deconvolution: (1) it removes the intraseries correlation due to summing ages in abun- dance or catch data, and hence provides more conser- vative estimates of the significance of correlations between recruitment and the environment; and (2) it yields an estimate of recruitment from the abundance or catch data, hence it can also increase the magnitude of estimated correlation coefficients. However, decon- volution cannot yield an exact value of recruitment when noise is present in the abundance or catch data. The error present in the recruitment estimates depends on the error present in the catch or abundance series and on the stability of the deconvolution (Kope and Botsford 1988). Thus, there is a trade-off between the two advantages of removing the effects of multiple age- classes and the disadvantage of amplifying errors that depends on the stability of the deconvolution and the magnitude of measurement errors in abundance. Con- sequently, deconvolution may not always be useful. Intraseries correlation is accounted for in the com- putation of confidence limits by estimating the variance of the computed correlation coefficient in a way that accounts for the intraseries correlation. This often in- volves using one of the many approximations to an ex- pression originally due to Bartlett (1946) (Box and Jenkins 1976, Chelton 1984, Botsford 1986). This ap- proach is functionally equivalent to adjusting the num- ber of degrees of freedom using a variation of the expression originally due to Bayley and Hammersley (1946) (Sutcliffe et al. 1976, Drinkwater and Myers 1987). We use another method for dealing with this problem that leads to an actual probability of detection that is closer to the specified probability than other methods (Kope and Botsford 1988, Botsford and Wain- wright unpubl.). In spite of the fact that greater intraseries correla- tion artificially inflates the probability of detecting a significant relationship, some researchers have added intraseries correlation to the time-series before com- puting correlations. For example, in the freshwater Chinook salmon studies cited above, Dettman et al. (1986) and Reisenbichler (1986) used 2-year moving averages of both spawning escapement and environ- mental variables in an effort to reduce the effects of age structure in the spawning escapement data. Hol- lowed et al. (1987) used five-point moving averages in their analysis of recruitment patterns in the northeast Pacific. Because the effects of intraseries correlation were not accounted for in either of these studies, the use of moving averages would decrease the probability of detecting real correlations and increase the probabil- ity of spurious correlation (Kope and Botsford 1988). Others have used moving averages before computing 260 Fishery Bulletin 88(2). 1990 correlations, but then have attempted to account for increased intraseries correlation [cf. Drinkwater and Myers (1987), who use a modified form of the Bayley and Hammersley (1946) approach]. The goals of the study reported here are to explore potential environmental influences on chinook salmon populations in California's Central Valley, and to dem- onstrate the effective use of some recently developed methods of correlation analysis that are more realistic and conservative than those used previously. We iden- tify potential oceanographic environmental influences on these populations and show that the freshwater in- fluences, which were established by direct survival estimates (USFWS 1987, Kjelson and Brandes 1989), are not detectable from correlation analysis of the available environmental and population data. Data and methods Environmental data were obtained from four sources. Average monthly streamllow data were obtained from published USGS records for gauging stations at Verona on the upper Sacramento River, Nicholas on the Feather River, and Fair Oaks on the American River. Flow and diversion data for the Sacramento-San Joa- quin delta were obtained from the California Depart- ment of Water Resources DAYFLOW hydrological model. Data on sea surface temperature at the Farallon Islands, southern oscillation index, and tidal height at San Francisco (37.48 N 122.22 W) were obtained from D. Cayan (Scripps Inst. Oceanography, La JoUa, CA 92037), and calculated coastal upwelling index at lat. 39°N (Bakun 1975) was obtained from A. Bakun (Pacific Fish. Environ. Group, Natl. Mar. Fish. Serv., NOAA, Monterey, CA 93940). Population data were taken from the recreational fishery, the commercial fishery, and spawner esti- mates. We used spawner data from the upper Sacra- mento River, the Feather River, the Yuba River, and the American River. In the available data, adults had been separated from jacks on the basis of length. Fish less than 60.7 cm fork length were taken to be jacks and larger fish were counted as adults. For the period 1970-86, spawning stock estimates were obtained from Pacific Fishery Management Council (PFMC) reports. For the years 1962-69, total fall-run spawners were taken from Fr\' and Petrovich (1970), and adult spawn- ers were estimated by multiplying the total estimate of fall-run spawners in each stream by the fraction of carcasses classified as adults in the spawning stock surveys. In addition to abundance indices from indivi- dual streams, total adult spawners for the entire Cen- tral Valley were estimated by multiplying the fraction of adults reported in all spawning stock surveys each year by the number of spawners in all nms for all rivers combined. Commercial and sport catch south of Point Arena were obtained from PFMC reports for the years 1971-86, and from L.B. Boydstun (Calif. Dep. Fish Game, Region II, Rancho Cordova, CA 95670, unpubl. data) for the years 1962-70. Commercial effort, measured in thousands of landings, and sport effort, measured in thousands of angler days, were obtained or calculated from the same sources. Fishing effort was used to calculate catch-per-unit-effort (CPUE) in an at- tempt to remove some of the effects of variable fishing effort from the catch data. Recruitment of year-classes at the beginning of the commercial fishing season in their second year of life was estimated by deconvolution of the abundance in- dices using the procedure described in Kope and Bots- ford (1988). This involved calculating the contribution of recruitment in each year to the abundance indices in subsequent years using an age-structured catch and escapement model with population parameters esti- mated by separable virtual population analysis of marked hatchery fish (Kope 1987). In addition to the recruitment estimates obtained by deconvolution of the individual abundance series, we estimated recruitment by deconvolving the combined abundance series. This combined deconvolution was obtained by adding to- gether commercial catch, sport catch, and total spawn- ing escapement for each year, and deconvolving the combined abundance series (Kope and Botsford 1988). Examination of the different recruitment estimates revealed that the deconvolved spawner series contained a great deal of high-frequency noise which resulted from the marginal stability of the deconvolutions of spawner indices. Because of differences in the relative contribution of each age-class, the deconvolution of total spawners was inherently much less stable than the deconvolution of adult spawners (Kope and Bots- ford 1988). Because recruitment estimates from spawn- ers (adults and jacks) and adult spawners were highly correlated with one another (hence probably represent the same signal, and the deconvolved adult spawners contain less introduced error), only deconvolved adult spawners were used as recruitment estimates derived from spawning escapement. No attempt was made to separate natural production from hatchery fish in the spawning escapement. Be- cause of the large contribution of hatchery fish to the spawning runs (cf., Reisenbichler 1986, Dettman and Kelley 1986) and the straying of hatchery fish (Hallock and Reisenbichler 1979, Sholes and Hallock 1979), ac- curate estimation of the natural component of the runs was not possible. Before computing correlations with environmental data, we tested for an influence of density on the pop- Kope and Botsford: Recruitment of Oncorhynchus tshawytscha in central California 261 600, • 65 i2i 500- 400- 83 75 •66 „ 82^i-^ "—---.JL* " 2 « 300- 200- y^2 "•70^76 .68 ■~~-~-- ^ ^ 100- 0 / • 69 c 50 100 150 Adult Spawners (thousands ol lish) 200 Figure 2 Chinook salmon stock-recruitment curves for the main stem of the Sacramento River. Upper cui-ve was fitted to the data for brood years 1962-66 and the lower curve was fitted to 1967-83. degrees of freedom is less than the number of years in the series. Linear trends in the time-series reflect variability on time scales that: (a) can easily be removed by detrending, and (b) are not detectable with existing data. We have removed linear trends from our time- series because they may obscure a relationship on faster time scales. Intraseries correlation due to variability on inter- mediate time scales was accounted for in evaluating the significance of computed correlation coefficients. The variance of computed crosscorrelation coefficients can be approximated by Var [r,„(0] ri h1 P,,{i) P,nM) (1) ulation series. Stock-recruitment relationships for the Sacramento, Feather, and American Rivers were ex- amined by plotting recruitment against spawning stock, and fitting Ricker stock-recruitment curves to the data by the standard linear-regression method. This approach can result in substantial biases in the esti- mates of stock-recruitment parameters that result from the correlation between recruitment and subsequent spawning stock (Walters 1985, Kope 1988). However, because we use stock-recruitment curves only as a visual aid in interpreting the stock-recruitment data, potential bias was not evaluated. To examine possible relationships between recruit- ment and the environment, correlations were computed between time-series of recruitment and quarterly aver- ages of the environmental variables at all lags that had potential biological ineaning. These quarterly averages are referred to as: winter, January-March; spring, April-June; summer, July- September; fall, October- December. Correlations with individual streatn flows were computed at lags that corresponded to the winter, and to spring while the fish were resident in the streams as fry or migrating downstream. Correlations with delta flows and diversions were computed for the spring while the fish were migrating to sea, and cor- relations with oceanographic variables were computed from the spring in the year that fish migrated to sea through the summer 2 years later when most fish mature and leave the ocean to spawn. Correlations in- volving abundance indices (i.e., catch and spawning escapement series) were computed at lags that as- sumed the indices consisted primarily of 3-year-old fish. Variability on all time scales can contribute to the variance of time-series and to the magnitude of calcu- lated correlation coefficients. However, variability on greater than annual time scales involves intraseries correlation, which implies the effective number of where n is the number of data pairs, P,rj{i) and Pyy{i) are the autocorrelation functions of the two time-series variables at lag i, and r\y{t ) is the computed cross- correlation coefficient between the two time-series at lag t (Botsford and Wainwright unpubl., Kope and Botsford 1988). When no intraseries correlation is pres- ent, this expression simplifies to Var[r,,(0] = 1 n (2) Because the real values of the autocorrelation functions of time-series variables are unknown, and computed autocorrelation coefficients must be used instead, ex- pression (1) can sometimes produce variance estimates smaller than expression (2). However, expression (2) places a lower bound on the possible variance of com- puted crosscorrelation coefficients because it approx- imates the variance in the best possible case, when no intraseries correlation is present. We estimated the significance of computed crosscorrelation coefficients by assuming a normal distribution with variance given by the greater of expressions (1) and (2). This strategy has performed the best in giving appropriate rejection rates in Monte Carlo simulations for independent ran- dom series with varying degrees of intraseries correla- tion (Botsford and Wainwright, unpubl. data). Results Exploration of possible density-dependence revealed no clear relationships that could be removed prior to examining the influence of environmental variables. The stock-recruitment relationship for the upper Sacra- mento River appears to show a decrease in the equi- librium stock size that coincides with the closure of Red Bluff Diversion Dam in 1966 (Fig. 2). This supports the 262 Fishery Bulletin 88(2). 1990 Table 1 Correlations of chinook salmon popu ation estimates with freshwater flow variables in the Sacramen- to Valley, California. All correlations significant at the 0.1 level using the uncorrelated series test | are shown. Asterisks refer to levels of significance using the test that accounts for intraseries cor- | relation [equation (1)]. Index Variable Age of influence Correlation Feather River adult spawners Delta diversions 1st spring 0.4.56" Feather River recruits (deconvolved adult spawners) Delta diversions 1st spring 0.424** Total adult spawners Delta inflow 1st spring 0.402* Delta outflow 1st spring 0.393* *p<0.10, •*p<0.05 Table 2 Correlations with combined California Central Vail ev fall run chinook salmon spawning stock indices. | UPW = upwelling index, SLH = sea level height, SST = sea surface temperature. All correlations significant at the 0.1 level using the uncorrelated series test are shown. Asterisks refer to levels of 1 significance using the test that accounts for intraseries correlation [equation (1)1- Index Variable Age of influence Correlation Total adults UPW 1st spring 1st summer -0.542*** -0.444* SST 3rd spring 3rd summer -0.427** -0..593*** Recruits (deconvolved adults) SLH 3rd summer -0.425** SST 3rd summer -0.541'** *p<0.10, **;)<0.05, ♦**/)^* ^^Ai^\ Figure 2 An example of a year-to-year match of an individual fin whale in Massachusetts Bay. The photos show catalogue number 0143 in 1983 (top two photos) and in 1987 (bottom two photos). Seipt et al : Identification of Balaenoptera physalus in Massachusetts Bay 275 Table 1 The number of individual fin whales (n) identified in Massachusetts Bay for each year of the study period, and the percentage of those individuals that were reidentified within the same year, and in each subsequent year. For example, of the 22 individuals identified in 1981, 50.0% were seen more than once that year, 59.1% were reidentified in 1982, and 50% in 1983. The mean of all same-year resightings is 49.2%. The mean of all consecutive year values (e.g., 1980-81, 1981-82) is 44.5%. 1980 1981 1982 1983 1984 1985 1986 1987 1980 n = 21 33.3% 38.9% 38.9% 50.0% 44.4% 33.3% 66.7% 61.1% 1981 re = 22 50.0% 59.1% 50.0% 40.9% 45.5% 40.9% 45. .5% 1982 re = 38 42.1% 42.1% 39.5% 42.1% 47.4% 47.4% 1983 « = 33 48.5% 42.4% 36.4% 54.5% 60.6% 1984 re = 38 50.0% 36.8% 44.7% 50.0% 1985 re = 31 .58.1% .54.8% 51.6% 1986 n = 67 52.2% 37.3% 1987 re = 64 59.0% I I III 'l,' "is \ S I 15 \ 15 I 15 I 15 15 I 15 MAR APH MAY JUN JUL AUG SEP OCT Figure 3 Observed temporal occurrence of individual fin whales in Massachu- setts Bay during 1987. Individual whales are ordered (top to bot- tom) by date of first sighting; each row of marks represents the dates on which one individual was seen during the year. A black dot to the left of the vertical a.xis signifies that the individual concerned had been observed in at least one previous year of the study period. Table 2 Resighting intervals, grouped by 10-day periods. of individual fin whales photographed in Massa- chusetts Bay for the three years 1985, 1986 and 1987 combined. The zero-interval category encom- passes whales that were observed | only once during a year. Interval No. of (days) intervals 0 81 1-10 146 11-20 24 21-30 14 31-40 9 41-50 6 51-60 4 61-70 2 71-80 2 81-90 2 91-100 2 101-110 0 111-120 0 121-1.30 0 131-140 0 141-1.50 0 151-160 1 161-170 1 171-180 1 276 Fishery Bulletin 88(2), 1990 Table 3 Sighting histories oi individual fin whales observed in Massachusetts Bay in more than one year of the s tudy. 0 = observed. Note: 0180 and 0394 were also observed in 1979. No. 1980 1981 1982 1983 1984 1985 1986 1987 No. 1980 1981 1982 1983 1984 1985 1986 1987 0017 0 0 0 0 0 0 0 0227 0 0 0 0027 0 0 0 0 0 0 0 0228 0 0 0 0045 0 0 0 0230 0 0 0061 0 0 0 0 0 0232 0 0 0062 0 0 0 0 0233 0 0 0 0 0064 0 0 0 0 0241 0 0 0 0068 0 0 0261 0 0 0 0 0069 0 0 () 0266 0 0 0 0081 0 0 0 0 0 0 0 U 0267 0 0 0 0 0 0086 0 0 0 0 0275 0 0 0 0 0089 0 0 0 0 0 0 0 0281 0 0 0 0 0090 0 0 0286 0 0 0093 0 0 0 0 0287 0 0 0096 0 0 0299 0 0 0110 0 0 0303 0 0 0 0113 0 0 0318 0 0 0114 0 0 0 0319 0 0 () 0 0126 0 0 0 0 0322 0 0 () 0 0131 0 0 0334 0 0 0 0137 0 0 0 0357 0 0 0141 0 0 0367 0 0 0 0 0143 0 0 0 0 0 0.391 0 0 0 0146 0 0 () 0 0 0 0394 0 0 0 0 0174 0 0 0 0 0420 0 0 0179 0 0 0 0 0434 0 0 0180 0 0 () () 0 0 0 0443 0 0 () 0183 0 0 () (_) 0 0452 0 0 0189 0 0 0 0 0 0 0512 0 0 0213 0 0 0 0 0 0520 0 0 () 0215 0 0 0 0 0 0521 0 (1 0218 0 0 0542 0 0 0222 0 0 0 0 0 0 0575 0 0 0224 0 0 0582 0 0 0225 0 0 0 0584 0 0 0226 0 0 0 0592 0 0 backs. Furthermore, while a humpback can often be recognized from a single poor-quality photograph, iden- tification of most individual fin whales requires a series of high-quality photographs which must be taken from approximately the same position relative to the whale. Despite this, the frequency with which individual fin whales were resighted is high. Roughly half of the fin whales seen in one year were seen again the same season, and about half were also resighted the follow- ing year. Resighting intervals provide some evidence for a degree of residency on the part of some whales, and of bimodal occupancy by others. However, the many gaps in the sighting histories of individuals are difficult to interpret. Do the gaps represent whales that were resident in the area for extended periods but were not photographed, or do they indicate movement to other areas between sightings? At this point, we can say only that, given the obvious bias in effort, it seems likely that rates of within-season occurrence and an- nual return of fin whales in Massachusetts Bay are con- siderably under-represented in our data. When only 1 out of every 83 fin whales sighted is identified (as occurred in 1981), it is unlikely that even an individual which remained in the area for many weeks would be recaptured more than a few times. This is supported by the fact that the subset of individuals towards which there was clear observer bias— those bearing large, prominent scars— were among the most frequently observed animals during this study. For example, whale number 0081, named "Braid" (an animal with very large propeller scars on its left side), was observed in all 8 years, with a mean occupancy of 116 days, and mean occurrence of 6 days. Seipt et al Identification of Balaenoptera physalus in Massacfiusetts Bay 277 It is clear, however, that many of the gaps in our sighting histories of individual whales represent real absences from the area. Resightings of individuals out- side our study area confirm that the summer ranges of these whales are often, if not regularly, extensive. In a preliminary analysis of photographs from the North Atlantic Fin Whale Catalogue, Agler et al. (1990) report a number of instances of movement of indivi- duals between Massachusetts Bay and the waters of Maine and the Bay of Fundy. It is possible that some individuals undertake extensive movements outside the Gulf of Maine, but the current lack of observer effort beyond this region precludes investigation of this idea. From the regional perspective of this study, it is tempting to compare the results reported here with the much more complete information available on Gulf of Maine humpback whales and to conclude that the two populations are broadly similar in their patterns of occurrence and distribution. The annual return rate of humpback whales to Massachusetts Bay is extremely high, with as many as 85% of individuals observed in one year returning the next (Mayo 1983, Mayo et al. 1988). While no individual humpback remains in Mas- sachusetts Bay for an entire season, many appear to spend prolonged periods in the area between making wider forays elsewhere in the Gulf of Maine (Clapham and Mayo 1987, Mayo et al. 1988). Other humpbacks are observed less often, and presumably frequent other habitats for much of the year, a fact which reflects con- siderable variation among individuals. It is also clear that overall individual fidelity to the Gulf of Maine is maternally directed (Clapham and Mayo 1987). It is not unreasonable to expect that fin whales might exhibit similar patterns of occurrence. Given that the distribution of fin and humpback whales in high lati- tudes must be predominantly related to the distribu- tion of their prey (Payne et al. 1986), it is to be expected that the occurrence of both species should be character- ized by individuals returning repeatedly, both during a season and from year to year, to consistently produc- tive habitats such as Massachusetts Bay. From the standpoint of energetics, it would make little sense for an individual to abandon an area of high productivity and search elsewhere if the resources found in the former habitat were adequate for its needs, although it is possible that, as suggested by Watkins et al. (1984), social imperatives play an important role in the move- ment of individuals. Overall, the data presented here suggest that there are more similarities than differences in the high- latitude population characteristics of humpback and fin whales, although data from other sources suggest that significant differences between the two species do exist. In addition to the above-mentioned studies which documented extensive movements on the parts of individuals (Brown 1962, Mitchell 1974, Watkins et al. 1984), other observers have noted evidence for spatial segregation by length in certain areas (Mackintosh 1942, Mitchell 1974, Rorvik etal. 1976, Sergeant 1977), a phenomenon that has not been demonstrated for humpbacks. Our own data are regional in nature and do not permit us to address these broader questions of population structure at the oceanic level. However, with increased photographic effort in other areas, studies based upon the identification of individual fin whales should provide clearer insights into the popula- tion biology of this species in the North Atlantic. Acknowledgments The authors are grateful to the staff of the Center for Coastal Studies who, by ignoring humpback and right whales as often as possible, helped to provide the photographic data upon which this paper is based. We also thank Bill Rossiter for additional photographs and data, and for his consistent support of fin whale studies over the years. We are grateful to the captains and crew of the Dolphin Fleet for all their assistance in the field; we are particularly indebted to Captain Aaron Avellar, who initially suggested that variations in the blaze and chevron patterns of fin whales might be useful in the recognition of individuals. The manuscript benefitted from a thoughtful review by Scott Baker. This study was funded in part by the National Marine Fisheries Service, Northeast Region, under contract number 50-EANF-6-00059. Additional support from the International Wildlife Coalition and the Seth Sprague Educational and Charitable Foundation is gratefully acknowledged. Citations Agler, B.A., J. A. Beard, R.S. Bowman, H.D. Corbett, S.E. Frohock, M.P. HawTermale, S.K. Katona, S.S. Sadove, and I.E. Seipt 1990 Finback whale. Balaenoptera physalus, photographic identification: Methodology and preliminary results from the western North Atlantic. In Hammond, P.S.. et al. (eds.). In- dividual recognition and the estimation of cetacean population parameters. Rep. Int. Whaling Comm., Spec. Issue 12 (in press). Brown, S.G. 1962 The movements of fin and blue whales within the Antarc- tic zone. Discovery Rep. .33:1-54. CeTAP 1982 A characterization of marine mammals and turtles in the mid- and North Atlantic areas of the U.S. outer continental shelf. Final Rep. Cetacean and Turtle Assessment Program, Univ. Rhode Island, Bur. Land Manage, contract AA551-CT8- 48. U.S. Dep. Interior, Wash., DC, 450 p. 278 Fishery Bulletin 88(2). 1990 Clapham, P.J. (editor) 1987 North Atlantic fin whale research. Proceedings of a meeting to review methodology and to coordinate information exchange, October 1986. North Atl. Mar. Mamm. Assoc, Pi-ovincetown. MA. 33 p. Clapham, P. J., and C.A. Mayo 1987 Reproduction and recruitment of individually identified humpbaclf whales, Megaptern noviwantjliae, (ihsei"ved in Massa- chusetts Bay, 1979-19S.S. Can. .1. Zool. (i.S:28.53-28(53. Hammond. P.S., S.A. Mizroch, and G.P. Donovan (editors) 1990 Individual recognition of cetaceans: Use of photo-identi- fication and other techniques to estimate population param- eters. Rep. Int. Whaling Comm., Spec. Issue 12 (in press). Hawvermale, M.P. 1987 Identification of individual fin whales, its use and tech- nique. Unpubl. Master's thesis. Southern Connecticut State Univ., New Haven, 49 p. Mackintosh. N.A. 19-12 The southern st(}cks of whalebone whales. Discovery Rep. 22:197-300. Mayo, C.A. 1982 Observations of cetaceans: Cape Cod Bay and Stellwagen Bank, Massachusetts, 197.5-1979. Natl. Tech. Inf. Serv. Rep. MMC-80/07, Springfield. VA. 68 p. 1983 Patterns of distribution and occurrence of humpback whales in the southern Gulf of Maine (abstract). In Pro- ceedings, Fifth biennial conference on the biology of marine mammals. Boston. MA. p. 64. Society for Marine Mam- malogy, Lawrence, KS 66044. Mayo, C.A., M.P. Hawvermale, C.A. Carlson, D.K. Mattila, and P.J. Clapham 1985 Identification of individual fin whales: The technique and its uses (abstract). In Proceedings, Sixth biennial conference on the biology of marine mammals, Vancouver. BC. Society fur Marine Mammalogy, Lawrence, KS 66044 [unpaginated]. Mayo, C.A., D.K. Mattila, S. Pittman, and L. Baraff 1988 Abundance, distributidn ami habitat use of large whales in Massachusetts Bay and the Great South Channel. Final rep. to NOAA Fisheries under contract .50-EANF-6-000.59. Avail. Northeast Fish. Cent., Natl. Mar. Fish. Serv.. NOAA, Woods Hole. MA 02543, 151 p. Mitchell, E.D. 1974 Present status of northwest Atlantic fin and other whale stocks. In Schevill, W.E. (ed.). The whale [irobleni. A Status Report, p. 108-161, Harvard I'niv. Press. Payne, P.M., J.R. Nicolas, L. O'Brien, and K.D. Powers 1986 The distriliution of the humpback whale, Meijaptera novaeangliae, on Georges Bank and in the Gulf of Maine in rela- tion to densities of the sand eel.Ammvdyte.f nmeriaviiii:. Fish. Bull., U.S. 84:271-277. Rorvik, C.J., J. Jonsson. O. Mathisen, and A. Jonsgard 1976 Fin whales, Balaenoptvni phjisaliis (L.), off the west coast of Iceland. Distribution, segregation by length and exploita- tion. Rit Fiskideildar 5:1-30. Sergeant, D.E. 1977 Stocks of fin whales, Bahifnoplcra physalus, L. in the North Atlantic Ocean. Rep. Int. Whaling Comm. 27:460-473. Watkins, W.A., K.E. Moore, D. Wartzok, and J.H. Johnson 1981 Radio tracking of finback {Balncnuptera physaliiti} and humpback (Megaptera novaeanyline) whales in Prince William Sound. Alaska. Deep-Sea Res. 28:577-588. Watkins, W.A., K.E. Moore, J. Sigurjonsson, D. Wartzok, and G.N. di Sciara 1984 Fin whale (Balni'nuplirii ///i?/.sy//h.s') tracked by radio in the Irminger Sea. Rit P'iskideildar 8:1-14. Abstract. — A symbiotic nemer- teaii worm found on spiny lobsters is described and compared with other members of the genus Carcinone- mertes. The new species oi Carcino- nemertes has a relatively large basis, stylet, anterior and posterior probos- cis chambers, and a voluminous sem- inal vesicle. In addition, lateral intes- tinal diverticula project anterior to the level of the middle proboscis chamber. These characters distin- guish this species from others in the genus. The life-history patterns of sLx spe- cies of Carcinonemertes appear re- lated to the developmental timing of host embryogenesis. Portunid crabs with a short duration of embryogen- esis are infested by the species C. carcinophila and C. mit^iukurii. The worms settle only on mature female hosts: after eclosion the worms mi- grate to the branchial lamellae of the host where they lie dormant until the host oviposits a new clutch. Cancrid and grapsid crabs with an interme- diate duration of embryogenesis are infested by the species C. epialfi and C. errans. The worms settle on both sexes of crabs and are sexually trans- mitted to female hosts. At eclosion, the worms die or regress and mi- grate to the limb axillae. Lithodid crabs and panulirid lobsters with a long duration of embryogenesis are infested by C. regicides and C. wick- hanii. The nemerteans settle on ovig- erous hosts and die or leave the host after eclosion. Carcinonemertes wickhami n. sp, (Nemertea), a Symbiotic Egg Predator from the Spiny Lobster Panulirus interruptus in Soutliern California, witli Remarl32 78x47 C. epialti Coe, 1902 F&M 4.0-6.0 30.0 13.5 61-68 C. epialti * (Shields, pers. observ.) F M 4.3 2.2 31.2 14.5 75 63x41 C. errans Wickham. 1978 F&M 4.0-6.0 35.2 11.0 >46 100 X 50 C. mitsukurii Takakura, 1910 F M 30.0 10.0 27.0 8.0 48" 86x28** C. regicides Shields et al.. 1989 F M 2.1 1.6 40.0 17.0 76 82x62 C. wickhami n. sp. F M 30.0 10.0 40.0 20.0 98 79 125 X 42 149x47 *ex. Cancer antennarius "Shields, pers. observ. bases, and stylet apparatus. These species prey on large host eggs. In addition, the morphology of the stylet of C. wickhami is similar to that of C. regicides (i.e., in the form of a broad, flat dagger). The above adapta- tions may allow penetration of the thick outer coat of these large eggs (Wickham and Kuris 1988). The presence of ovarian pores may be of taxonomic value to the genus. Distinct ovarian pores were not observed on female C. wickhami. Similarly, Strieker (1986) did not find distinct ovarian pores on C. epialti. Takakura (1910) and Shields et al. (1989) reported distinct ovarian pores from C. mitsukurii and C. regicides, respectively. The ovarian pores of C. regi- cides are typically found prior to ovulation, but are not apparent prior to oviposition (Shields et al. 1989); C. mitsukurii appears to follow a similar pattern (Shields, pers. observ.). The larvae of C. wickhami resemble the typical hoplo- nemertean larval form (Gibson 1972, p. 150), and are similar to the larvae of C. epialti in gross morphology (Strieker and Reed 1981). Larvae most likely undergo direct development into the juvenile stage upon settling onto their host (Strieker and Reed 1981). The larva of C. wickhami is similar to that of C. carcinophila, C. epialti, and C. regicides; the larvae do, however, differ in size. Carcinoyiemertes ivickharni has the smallest larva yet described for the genus. Mature C. wickhami were found only in the broods of ovigerous lobsters. Further, these nemerteans have been collected only from lobsters with eggs in relatively advanced stages of development (development of eye placodes initiated— 8 of 19 lobsters examined). No nemerteans were recovered from 6 lobsters with rela- tively early broods nor from 3 lobsters held for dissec- tion after eclosion. Carcinonemertes wickhami does not appear to migrate to the branchial chamber and into the branchiae after host eclosion, nor does it move to the limb apodemes and axillae (see below). Three distinct life-history patterns have emerged for sb{ of the eight species of Carcinonemertes. These pat- terns appear related to the developmental timing of host embryogenesis (oviposition to eclosion): embryo- genesis in various reptantian decapods can be of short (e.g., 13-16 days), moderate (e.g., 40-120 days), or long duration (e.g., 120-300-1- days). In addition, life-history patterns may be useful taxonomic characters as few differences in morphology aid in distinguishing be- tween species. 1 Portunid crabs have a short duration of embryo- genesis (Churchill 1919, Sandoz and Rogers 1944). Careinonemertid larvae settle primarily on ovigerous female crabs where they quickly metamorphose and mature. After eclosion, adult worms migrate to the 286 Fishery Bulletin 88(2). 1990 branchial chamber, encyst between the branchial lamel- lae, and lie dormant until the female crab oviposits (Humes 1942). Adult worms then migrate back to the clutch (Humes 1942, Hopkins 1947). Two species of Carcinonemertes follow this pattern: C. carcinophila and C. mitsukurii. 2 Embryogenesis of cancrid and grapsid hosts is of moderate duration (Kuris 1971, Roe 1979, Wickham 1980, Shields et al. In press). Carcinonemertid larvae settle on male and female crabs, metamorphose into juveniles, and migrate to the limb apodemes, axillae, and abdomens of their hosts. Juveniles transfer from male to female crabs during host copulation (Wickham et al. 1985). The juveniles then lie dormant, absorbing amino acids and other nutrients (Roe et al. 1982), un- til the host oviposits her clutch. Juvenile worms then migrate into the egg clutch and mature after eating crab eggs. Two nemerteans follow this pattern: C. errans and C. epialtl. 3 Spiny lobsters and king crabs brood their eggs for long periods (Marukawa 1933; Shields, pers. observ.). Nemertean larvae settle, metamorphose, and mature only on ovigerous female crabs. Adult worms die or leave the host at eclosion. Circumstantial evidence sug- gests that planktonic larvae may, in at least one species (C. regicides), reinfect hosts immediately uymn hatch- ing (Kuris et al. In prep.). Two species follow this pat- tern: C. regicides and C. wickhami. Acknowledgments We thank UCSB divers/collectors Shane Anderson and Jim McCullough for their able collection of ovigerous lobsters. Jason Daniel Shields and Emma Ramoy pro- vided moral support. Drs. Ian Whittington and Tom Cribb improved the manuscript. This work is a result of research sponsored in part by NOAA, National Sea Grant College Program, Department of Commerce, under grant number NA80AA-D-()0120, project num- ber R/F-75, through the California Sea Grant College Program, and in part by the California State Resources Agency. The U.S. Government is authorized to repro- duce and distribute for governmental purposes. Citations Campbell, A., K. Gibson, and L.H. Evan 1989 A new species of Carcinonemertes (Nemertea: Car- cinonemertidae) ectohabitant on Panidiru.'i n/gnus (Crustacea: Palinuridae) from Western Australia. Zool. J. Linn. Soc. 95: Churchill, E.P. 1919 Life hi.story of the blue crab. Bull. t'.S. Bur. Fish. ,'«5: 96-123 (Doc. #X70). Coe. W.R. 1902 Nemertean i)arasites of crabs. Am. Nat. 36:431-450. Gibson, R. 1972 Nftnertcitna. Hutchinson & Co. Ltd., London. 224 p. Hopkins, S. 1947 The nemertean Carcinonemerte)^ as an indicator of the spawning history of the host, C'dllinerten fiapidus. J. Para- sitol. 33:146-150. Humes, A.G. 1942 The moi-phology , taxonomy, and bionomics of the nemer- tean genus Carcinonemertes. III. Biol. Monogr. 18:5-105. Kolliker, A. von 1845 IJber drei neue Gattungen von VVurmern. Lineola, Chloraima, Polycystia, neue Wurgattungen, und neue Arten von Nemertes. Verh. Schweiz. Naturforseh. Ges. Chur 29 (1844):86-9S fin C,erman|. Kuris, A.M. 1971 Population interactions between a shore crab and two symbionts. Ph.D. thesis. Lhiiv. Calif., Berkeley. 451 p. 1978 Life cycle, distribution and abundance of CarciiMnemnien cpiatti, a nemertean egg predator of the shore crab, Hemi12 >12 Table 3 Length, mass, and brain mass of six strandec or captive-born neonate bottlenose dolphins captured off southern Africa. Length (cm) Mass (kg) Brain mass (g) 100 12.4 562 105 17.7 680 106 14.8 738 106 595 108 13.0 635 109 13.4 625 Mean 106 14.3 639 shows a clear discontinuity between 10 and 13 years of age (Fig. 4), indicating that a two-stage growth curve may better fit these data. The mean length and mass of the first five year- classes (calculated from calves displaying whole, com- pleted GLGs only) showed that growth rate during the first year far exceeded that in any other. Mass in- creases in the first year by some 255% of mean birth mass, but increases over the subsequent 4 years slow to 49%, 13.5%, 10.6% and 3.8%, respectively. In con- trast, length increase is much less rapid over this same period and is approximately 57%, 15.2%, 3.7%, 4%, and 5.5%, respectively. In both males and females the relationship between mass and length, up to asymptotic values, is well defined by power ctirves (males, log mass = -5.1 -f 3.06 X log length, r=0.97; females, log mass = -4.7-1- 2.9 x log length, r=0.95). Both male and female bottlenose dolphins appear to reach physical maturity between 12 and 15 years of age (Table 2). The youngest physically mature animal was 12 years old while the oldest nonphysically mature animal was 14 years old (Table 2). Female reproduction Only six (2.8% of the total catch and 12% of mature females) of the captured females were pregnant, and the growth of the six fetuses— length (cm) against mass (kg)— was well defined by a linear regression (Fig. 5). Although no estimate of gestation period was possible 120 n 100 ^ 80 ^ i 60 40 20 I ■ ^ ''"^ ■ 0 2 4 6 8 Mass (Kg) 10 12 14 Figure 5 Fetal growth in length and mass of bottlenose dolphins captured off southern Africa, and a regression of the form (length = 23.914 -t- 0.006 X mass) drawn through these data. from these data, an estimate was obtained from the relationship between neonatal and adult brain mass (Sacher and Staffeldt 1974, in Perrin et al. 1977). A neonatal brain mass of 639 g was estimated from the mean brain mass of six stranded or captive-born neonates in which the umbilicus was unhealed (Table 3). An asymptotic brain mass of 1460 g was calculated from the mean brain volume of physically mature females, although the variation in maximum size of females would obviously affect this. Application of the Sacher and Staffeldt equation gives a gestation period of some 372 days. An alternative method of estimating gestation period based on the relationship between birth length and gestation period (Perrin et al. 1977) yielded an estimate of 12.3 months or 374 days, assum- ing a birth length of 103 cm. Birth date, to the nearest month, was back-calculated for 25 captured calves less than 1-year-old by measur- ing the width of deposited dentine as a proportion of the mean width of the first GLG (Fig. 6). The mean width of the first GLG layer was 279 ^m (A'^ = 29, range 229-331 \im) with a 95% CI of 10 \m\, indicating a max- imum error of about 13 days. The subtraction of age from date of capture (Fig. 6) suggests that most births occurred in summer. In all females the majority of ovarian scars (80%) oc- curred in the left ovary. The maximum number of scars in any ovary was 1 1 , with no indication that ovary mass decreased with the number of scars (Fig. 7). Age-, mass-, and length-related ovulation rates appear ex- tremely varied in bottlenose dolphins (Fig. 8). In only one female was there one scar, so calculation of mean age at first ovulation was impossible. This 13-year-old 294 Fishery Bulletin 88(2). 1990 Figure 6 Birth month of 25 bottlenose dolphin calves captured off southern Africa back-calculated from the width of the first dentinal growth- layer-group. 18 1 w ro E 16 14 12 - ■ ■ . ■ • I • . . , ■ 1 ■o n E o 10 - 8 - 6 - 4 2 - 0 - • • ■ ' ' C ) 2 4 6 8 10 12 14 16 18 Ovarian scars Figure 7 Relationship between combined ovary mass and the total number of ovarian scars for female bottlenose dolphins captured off south- ern Africa. female had a corpus albicans measuring 10x8x8 mm (index volume 640 mm'') and was lactating, indicating she probably had a suckling calf. Extrapolation from a regression of calf length on an index of largest cor- pora volume, for mother-calf pairs (Fig. 9), suggests that the calf would have been some 184 cm in length or 18 months old (Figs. 3,4), implying that her first o 80 100 120 140 Mass (kg) 160 180 20 18 16 14 12 10 8- 6- 4 2 0 140 160 180 200 220 Length (cm) 240 260 Figure 8 Increase in the number of ovarian scars with mass, age, and length of female bottlenose dolphins captured off southern Africa. A linear regression (no. ovarian scars = 0..58 - 0.29 x age) fitted to this rela- tionship in postpubertal females is shown. ovulation and conception occurred at approximately 10. .5 years of age. Cockcroft and Ross: Tursiops truncatus off southern Africa 295 220 n E o c 200 180 160 - 140 • 120 D D D """""-^ a 1^ 100 80 ■ 60 40 20 - 0 - ( D 5 10 15 20 25 30 Corpora volume index ,mm^ X 100) Figure 9 Relationship between the length of a bottlenose dolphin calf and an index of the volume of the most recent or largest corpus in the mother. The fit of a linear regression (calf length in cm = 209.61 - 0.032 X corpora volume index) to these data is shown. 2500- ■ ■ ^^ "e 2000 3 ■ 0) E 3 o ' 1500 ■o c ■ a o> •■ ■ >< E 1000 ■ E ■ ■ ro ■ E o J SOO- TS = 20 30 40 50 60 70 Mass of Calf (kg) 80 90 Figure 10 Relationship between the mass of a bottlenose dolphin calf and an index of the volume of the mother's mammary glands. Two other females, 12 and 13 years old, which had each undergone two ovulations, were iactating and the index volumes of their largest corj)ora albicans were 900 mm'' and 2100 mm-\ respectively. Extrapolation from Figure 9 suggests that the calves of these females were approximately 181 cm or 18 months old, and 143 cm or 6 months old, respectively. These data imply that at conception of the calves, these two females were about 10.5 and 9.5 years old, respectively. The above data, and the presence of 10-year-old females that had not ovulated, imply that first ovulation occurs between 9.5 and 11 years of age. One 17-year-old female had undergone two ovulations, was Iactating, and had a 170-cm calf of approximately 1-year-old. This suggests that this female was 15-years-old at the time of con- ception of this calf, although she may have undergone a previous pregnancy. A regression fitted to the number of ovarian scars on age of sexually mature females has a slope of 0.29 (r = 0.6, A'^ = 32), implying that mature females, in general, ovulate at least every third year (Fig. 8). A similar regression of log of age on log of number of cor- pora albicantia. was linear (log age = -0.61 + 1.03 • log # scars, r = 0.62), indicating no decrease in ovula- tion rate with age. Length of lactation calculated from the catch statis- tics of animals on the Natal coast (Cockcroft and Ross In press) where 27% and 5.2% of females were Iac- tating and pregnant, respectively, gave an estimate of 5.2 years (proportion lactating/proportion pregnant; Perrin and Reilly 1984). As this is obviously exagger- ated by the overabundance of Iactating females and dearth of pregnant females, alternative means of estimating lactation period are required. The relation- ship between calf mass and an index of mammary gland volume (length x height x depth) in 13 mother and calf pairs suggests that the mammary glands increase in size during lactation, until the calf's mass is between 60 and 70 kg, after which the mammary volume decreases (Fig. 10). Extrapolation from the growth curves suggests that calves of this mass are about 18-months-old. An examination of the stomach contents of captured calves and juveniles showed that solids first appear in stomachs at about 6 months of age, although milk remains were still evident in calves of up to 3 years of age. Of the 20 known mother and calf pairs, ten calves were 1 year old or less, five were 1-2 years old and a further five were greater than 2 years old; the mother of one of the latter was pregnant with a fetus of only 38.5 g. Although only solids were found in the stomach of her 69-kg calf, she was still Iactating, the only one of six pregnant females simultaneously Iactating. These facts imply a mother and calf association of up to 3 years before a subsequent pregnancy. Female resting period calculated from gestation period (1 year), proportion of females resting (5.3%) and the proportion of females pregnant (Perrin and Reilly 1984) yielded an estimate of about 1 year. 296 Fishery Bulletin 88(2). 1990 25-. 2500 n 1 • • ■ ■ 20 2000 ■■ •• ". • E 3 s 1 ■6 4J 15 10J S '500 i 2 1000 ■ 0 * V ■ o °o 0° o o ° o° ° c ° ■ 0 ■■ o o o o o o £3 3 1 5- Combined « o Oo 0 0 10 20 30 40 50 Age Figure 1 1 Increase of mean seminifenjus tubule diameter (■) and combined testis mass (O) with age in male bottlenose dolphins captured off southern Africa. Table 4 Mean age. mass, and length of male bottlenose dolphins cap- tured off southern Africa, with testes showing no mature tubule development (stage 1), some ("^half) mature tubule development (stage 2), and 75-100% of tubules mature (stage 3/4). Testis development stage 1 3/4 Age Mass (kg) Length (cm) 4.68 90.8 201 99 14.3 1.51 228 20.6 174 241 2.5-1 ■ 2 ■ ■ ■ 1.5- ■ ■ ■ ■ ■ ■ 8 O X 1- ■■-.I . ■ ■ ■ ■ ■ tn E I/) Q5- : ■ ■ ■ ■ 160 180 200 220 240 260 .E s 2.5n 2- 1.5- 1- Length (cm) ■ ■ ■ ■ ..1 ■ ■..'■■■■ ■ Q5 ■ ■ ■ ■ 40 60 80 100 120 140 160 180 200 220 Mass (kg) Figure 12 Increase in combined testis mass with mass and length in male bottlenose dolphins captured off southern Africa. Male reproduction No consistent differences were found between mean seminiferous tubule diameter of samples taken from the outer, middle, and inner sections of testes. In gen- eral, left and right testes were of similar mass (r = 0.95, A'' = 46). Combined testis mass remained low (< 100 g) up to approximately 10 years of age, approximately 140 kg body mass, and a length of 225 cm, but there- after increased rapidly (Figs. 11. 12). Tubule diameter growth and state of testes development (Mitchell and Kozicki 1984) showed a similar pattern of growth with age, mass, and length (Table 4, Fig. 11), although development of tubules appeared to occur earlier than did the increase in testis mass. The maximum mature testis (stage 4) mass of any male was 1160 g and the largest apparently immature testis (stage 1) was 1.30 g. The smallest testis at stages 2, 3, and 4 were 140 g, 25 1 25t 0 o o o ■; 2 20 jj ■= ' □ = e B ^ 15i "e § 15- „ a i ° c» H Combined testis 1 j"" : • ; =' 05- 5- 0- ' 2 I 4 5 6 7 8 9 10 11 12 Months Figure 13 Influence of season (month) on combined testis mass (A) and mean testis seminiferous tubule diameter (D) in male bottlenose dolphins caiJtured off southern Africa. Cockcroft and Ross. Tursiops truncatus off southern Africa 297 320 g and 310 g, respectively. Mean age, mass, and length of males according to testis development stage are given in Table 4. Puberty in male Tursiops from the Indian Ocean may begin as early as 9 years of age but primarily between 10 and 12 years (Fig. 11). How- ever, sexual maturity (50% each of stages 2/3 and 4) occurs only at about 14.5 years of age, a length of 240 cm and a mass of 165 kg. There was no evidence of a seasonal pattern in either combined testis mass and tubule size (Fig. 13) or the presence of sperm in the epididymis. Discussion The similarity of dentinal GLG counts to "Dolfie's" actual age indicates an annual deposition of dentinal GLGs for at least the first 6 years. Ross (1984) superim- posed the growth rate of this animal on a length-versus- age (dentinal GLGs) relationship in bottlenose dolphins from Natal and the Eastern Cape, and concluded that "the best fit of the points to the curve is reached when the accumulation rate of dentine layers is equal to one per year." Similarly, when the growth rate of another captive Indian Ocean animal (Cockcroft and Ross 1990b) is fitted to the curve, the relationship is best explained by an annual deposition of dentinal GLGs, also proposed for Tursiops in other areas (Sergeant et al. 1973, Hui 1980). In contrast to apparently continuously deposited dentinal GLGs, cemental GLGs appear to be rapidly deposited and only accumulate as whole layers in the teeth of Tursiops. Nevertheless, the strong correlation between dentinal and cemental age estimates up to occlusion of the pulp cavity, and the similarity of den- tinal and cemental GLG counts in teeth of the captive animal, suggest that cemental GLGs also accumulate annually and are reliable estimators of age in bottlenose dolphins from the Indian Ocean, at least up to an age of 12 years. Despite the lack of direct evidence, it seems likely that cemental GLGs are deposited annually even after 12 years. Male and female Indian Ocean bottlenose dolphins may attain an age greater than 40 years (cf. Ross 1977) with little difference in the apparent maximum ages of the sexes. In excess of 20% of the Natal net catch were older than 20 years, indicating a long-lived spe- cies. Few data are available on the longevity of Tur- siops elsewhere, and most existing studies have used dentinal age estimates. Sergeant et al. (1973) estimated the longevity of Tursiops from northeast Florida to be about 25 years, with no apparent differences in the life expectancy of males and females. Hohn (1980), in a study of Tursiops from the southeast coast of the United States, found animals with up to 27 dentinal GLGs, males and females reaching similar ages. In con- trast, the maximum age of spotted dolphins, estimated from cemental GLGs, is in excess of 45 years (Kasuya 1976). Thus, the use of cemental GLG age estimates in future studies of bottlenose dolphins may yield greater estimates of maximum age. The estimated mass and length at birth of Indian Ocean bottlenose dolphins in this study are comparable with those calculated by Ross (1984). Subsequent to birth, growth is rapid, particularly in terms of mass, but decreases gradually with age. The proportional length increase is similar to that previously recorded for bottlenose dolphins from Natal and the eastern Cape (Ross 1977, 1984; Cockcroft and Ross 1990b) and also those from other areas such as northeast Florida (Sergeant et al. 1973), the western North Atlantic (Hohn 1980) and captive animals from the Pacific coast of Japan (Kasuya et al. 1986). An enormous increase in mass during the first (suckling) year is well known in seals and balaenopterid whales but has not often been recorded for delphinids. Presumably, a large ini- tial mass increase reflects the rapid development of the calf and its need to reach thermoregulatory equilibrium as well as some social and motor independence from its mother (Cockcroft and Ross 1990b) before the female's involvement with the next pregnancy and calf. The asymptotic length and mass values obtained in this study and those of Ross (1977, 1984) are less than those for bottlenose dolphins from the western North Atlantic (Hohn 1980), Florida (Sergeant et al. 1973), and the Japanese Pacific (Kasuya et al. 1986). It is unclear why bottlenose dolphins from different areas have varying asymptotic sizes. Some authors have suggested that this may warrant the separation of the various populations to the specific level (Ross 1977), although Ross and Cockcroft (1990) have sug- gested that there is little morphometric evidence to suggest this and that such differences may only have resulted from environmental conditions, particularly temperature. The asymptotic lengths of male and female Indian Ocean bottlenose dolphins are only slightly different (243 cm and 238 cm, respectively). In the western North Atlantic, Hohn (1980) found no difference in the maximum lengths of males and females. These results support the findings of Sergeant et al. (1973) that the total lengths of male and female delphinids, in general, do not appear to be different, although they found that the asymptotic length of male Tursiops from Florida was 20 cm greater than that of females. In contrast, fully grown male Indian Ocean bottle- nose dolphins are considerably heavier (9%) and more robust than females (176 kg and 160 kg, respectively). Lactating and nonlactating females had a similar mean mass, indicating that this mass difference cannot be 298 Fishery Bulletin 88(2). 1990 attributed to stress and blubber mass loss through lac- tation (Cockcroft and Ross 1990a, b). Some of this mass difference between the sexes may be a direct conse- quence of the male mass growth spurt between 10 and 12 years of age. It is unclear why this growth spurt is not reflected in the length of males; it may be that robusticity and not length is important in male and female interaction. Despite this, the male growth spurt at the onset of puberty may be similar to the two-stage growth that Perrin et al. (1976, 1977) described for male and female spotted and spinner dolphins in the eastern tropical Pacific where growth showed a pubertal secondary growth spurt. In male bottlenose dolphins, this spurt occurs 4-5 years later than in spotted or spinner dol- phins but also at the onset of puberty, suggesting that such growth spurts may be directly related to the at- tainment of sexual maturity. It is possible that this growth spurt may also be evident in female bottlenose dolphins but is not discernible owing to the small sam- ple size. Females mature sexually some 2 or 3 years prior to the attainment of physical maturity in contrast to males, where sexual maturity is attained just before physical maturity. In both sexes, however, physical maturity occurs almost in concert with the occlusion of the tooth pulp cavity, supporting previous sugges- tions that animals with occluded pulp cavities are sex- ually and physically mature (Ross 1977, 1984). Females attain sexual maturity at least 2 years earlier and at a lesser length and mass than do males, although the reduced number of first-time ovulators and high occur- rence of females with multiple ovulations will have biased this upwards. It is not unusual for female del- phinids to attain sexual maturity sometime before and at a smaller size than males. Female Tur slops in the western North Atlantic also appear to follow this pattern (Sergeant et al. 1973). Female spotted and spinner dolphins reach sexual maturity about 3 or 3-4 years, respectively, before males and both are smaller than their male counterparts (Perrin et al. 1976, 1977). It has been proposed that this disparity ensures more sexually mature females than males in the population (Bryden 1972). Intensive behavioral field work is needed before this suggestion can be confirmed. Although no direct estimate of gestation period was available from the fetal growth data, both derived estimates were in excellent agreement, about 373 days. A 373-day gestation period is slightly longer than previous estimates for Indian Ocean Tursiops, which range from minimums of 342 and 341 days (Saayman and Tayler 1977) to maximums of 364 and 368 days (Ross 1984) in captive bottlenose dolphins. Similar estimates of the gestation period in captive Tursiops from other areas have been given by Tavolga and Essapian (1957), McBride and Kritzler (1951), and Kasuya (1985). Although the sample was too small for an assessment of the age at first ovulation, the onset of ovulation in females is apparently rapid. Thereafter, there was con- siderable variability in the ovulation rates. Some 12- or 13-year-old females had particularly high ovulation rates, possibly a result of several initial infertile ovula- tions (Harrison et al. 1972). Others of the same age had low corpora counts, were all lactating, and one was pregnant, indicating that fertilization occurred on the first or second ovulation. Indirect evidence, which shows that lactational transfer of organochlorines in female bottlenose dolphins occurred after one or two ovulations (Cockcroft et al. 1989), supports the view that the majority of females conceive after one or two ovulations. The variation in ovulation rates of older females may be due, in part, to the same factors which apply to younger females and to additional reasons such as calf mortality or aborted pregnancies. Overall, the calculated annual ovulation rate for female bottlenose dolphins was 0.28, a substantially lower rate than that observed for Tursiops from north- east Florida (Sergeant et al. 1973). There was little sign of reproductive senescence in females from Natal, as ovulation rate did not appear to decline with age, and there was no reduction in ovary mass with an increas- ing number of ovarian scars. Also, the oldest captured female was lactating, and a number of older females were captured with calves and had enlarged corpora in their ovaries. Results from organochlorine residue studies in these females also indicate that older females do not become senescent (Cockcroft et al. 1989). These data imply that older females do not act as "wet nurses," which is contrary to suggestions for several other species of odontocetes that manifest age-related declines in fecundity. Senescent females invest more in quality calf-rearing and a longer lactational commit- ment than in quantity calf-bearing, as their reproduc- tive potential fails (Kasuya and Marsh 1984, Marsh and Kasuya 1986). However, sample numbers in this study were small and certain female-calf pairs showed an extended relationship, though there was no indication that this was restricted to older females. Taken in combination, these facts indicate that some Indian Ocean bottlenose dolphin females are probably reproductively active until an advanced age. A similar conclusion was reached by Kasuya (in Marsh and Kasuya 1986) who found that although the annual pregnancy rate and the number of resting females in a sample of Tursiops from the Pacific declined with age, pregnant and lactating females were present in all age groups, presenting no conclusive evidence of senescence. Cockcroft and Ross: Tursiops truncatus off southern Africa 299 Although births apparently occur throughout the year, there is a peak in summer, between November and February, when over 60% of births occur. How- ever, as birth dates were back-calculated, this may reflect the greater catch of dolphins in these months and the bias of the net catch for larger calves (Cock- croft and Ross In press), although previous work in this region noted that births occurred predominantly in late spring and summer (Ross 1977). No seasonal cycle of either testis mass, tubule diameter, or occurrence of sperm in epididymis was evident in mature males. These data also imply no distinct mating or breeding season in Indian Ocean Tursiops. In Florida waters, the main mating and calving season is apparently February to May (Essapian 1963), or spring to early fall (Irvine et al. 1981), which is similar to that found in the present study. In contrast, bottlenose dolphins off Argentina show a distinct summer calving and mating season (Wursig 1978). These geographical vari- ations, however slight, indicate the adaptability of coastal Tursiops to local conditions. Although it has been suggested that suckling as a nutritional source probably only lasts 1 year (Kasuya and Marsh 1984, Cockcroft and Ross 1990b), there is evidence that suckling may last at least 18 months in Tiirsiops and that non-nutritional suckling may con- tinue for as long as 3 years for some mother and calf pairs. An estimate of the duration of lactation is dif- ficult, where suckling extends over long periods and may serve a non-nutritional purpose such as enhanc- ing the cow-calf bond (Brodie 1969). Of the calves from lactating female-calf pairs, 25% were over 1 year old and a further 25% were over 2 years old, and some of the latter had both milk and solids in their stomachs. The mammary glands from these lactating females, only one of which was pregnant, increased in size with calf size, until calves were at least 18 months old. In combination, these data indicate that lactation in In- dian Ocean bottlenose dolphins lasts more than 1 year and in some instances may extend to more than 2 years. This is slightly longer than previous estimates of lac- tation length and age at weaning based on studies of captive and captured free-ranging bottlenose dolphins (McBride and Kritzler 1951, Gurevich 1977, Saayman and Tayler 1977, Kasuya 1985, Cockcroft and" Ross 1990b) and suggests a prolonged mother-calf associa- tion that may extend in free-ranging bottlenose dol- phins for at least 15 months (Irvine et al. 1981). Such extended mother-calf interaction may indicate a stable school structure, such as that postulated for short- finned pilot whales off the Pacific coast of Japan, which may be indicative of late maturing, long-lived animals (Kasuya and Marsh 1984). This may equally apply to Tursiops where a lengthy mother-calf bond may be im- portant in the calf's development and be a reflection of the smaller school size and inshore habitat, mastery of which may require greater maternal care and a longer learning period (Cockcroft and Ross 1990b). During a study of captive bottlenose dolphin mother and calf association, Cockcroft and Ross (1990b) have shown that the calf's suckling rate decreased with age, although its energy requirements probably grow with its level of independence and activity. As there was no evidence of energy changes in delphinid milk during lactation (Arvy 1974) to compensate for this, the authors proposed that the quantity of milk ingested may increase as the calf's stomach volume increased (Cockcroft and Ross 1990b). This explanation is sup- ported by the present findings that a female's mam- mary glands increase in size, probably increasing the volume of milk produced, during lactation. In view of the extended lactation period of female bottlenose dolphins and the early and probably increas- ing intake of solid food by the calf, it is unlikely that females require a substantial interval between the end of lactation and the next pregnancy. A 1-year resting period, estimated from the catch statistics data, is almost certainly an overestimate due to catch bias. Kasuya (1985) estimated a 3-month resting period for Tursiops in the western north Pacific, and it is probable that Indian Ocean Tursiops are similar. Considering that gestation lasts about 1 year and that lactation probably lasts 18 months to 2 years, a calving interval of around 3 years can be estimated for Indian Ocean bottlenose dolphins. This estimate is in good agreement with the projected ovulation rate of one every 3 years, but assumes that all calves survive and ignores the effects of differential calf mortality (Perrin and Reilly 1984) that would lower the mean calving interval considerably. Nothing is known of the age and sex structure of the Natal bottlenose dolphin population. The only available information is from the catch of these animals in the Natal shark nets, the sex, size, and age structure bias of which have been discussed (Cockcroft and Ross In press). Given these biases, attempts to calculate repro- ductive parameters from these data are flawed but pro- vide the only means of calculating the reproductive potential of this population. The relevant proportions in the net catch of females- mature, lactating, pregnant, and resting— are 56%, 43%, 27%, 5.2%, and 5.8%, respectively (Cockcroft and Ross In press). Annual pregnancy rates (APR; Perrin and Reilly 1984) calculated from these catch data and lactation period (either 1 or 2 years) range between 5.2% and 27%. Changes in either the proportion of females lactating or the length of lactation greatly in- fluence this calculation, but even the highest estimate is low in comparison with values calculated for Tursiops in other areas; 63% in the Black Sea (Danilevsky and 300 Fishery Bulletin 88(2). 1990 Tyutyunnikov 1968, in Perrin and Reilly 1984) and 43.6% and 40.4% for the western north Pacific (Kasuya and Izumisawa 1981, in Perrin and Reilly 1984; Kasuya 1985). Gross annual reproduction rate (GARR) (Perrin and Reilly 1984) calculated from catch statistics and the range of APR values yields estimates between 0.043 and 0.065. Although biases in the catch will influence these GARR estimates, they are useful for comparative purposes. The former GARR estimate is greater than that calculated for an unexploited stock of Tursiops from eastern Australian waters, although this was based on an unreliable technique of estimating calf numbers from aerial surveys (Lear and Bryden 1980, in Perrin and Reilly 1984). The latter GARR figure, although probably an overestimate, is some 40% and 500% lower than those estimated for exploited popula- tions of Tursiops off Iki Island, Japan (Kasuya 1985) and in the Black Sea (Danilevsky and Tyutyunnikov 1968, in Perrin and Reilly 1984), respectively. The probable biases in the calculated APR and GARR estimates suggest that an assessment of the theoretical maximum natural rate of increase (ROI) of the Natal bottlenose dolphin population would be more practical. Assuming a calving interval of 2-3 years, age at first breeding of 10 years, and an annual survival rate of less than 0.97, an ROI of 4-6% can be calculated (Reilly and Barlow 1986). The ROI makes allowances for adult and calf mortality not accounted for by a GARR esti- mate, which, therefore, infers that even the greater GARR figure may be an underestimate. Given an an- nual increase of as much as 6% of the estimated 900 population, the mean annual mortality of bottlenose dolphins in shark nets— 32 dolphins per year including about 4 reproductive females— in conjunction with whatever other sources of man-induced mortality, such as the probable death of first-born neonates through pollutant toxicity (Cockcroft et al. 1989), implies that mortalities may be close to or exceed the likely replace- ment rate of this population. However, this conclusion should be viewed with some caution, as it is based on an estimated population of only some 900 dolphins, although biases of aerial counts suggest that numbers may be greater (Cockcroft et al. In press). Additional- ly, other factors may also influence understanding of the reproductive capacity of this population. If bottle- nose dolphins on the Natal coast are geographically separated for long periods (Cockcroft et al. 1989) with little mixing even of adjacent groups (Cockcroft et al. In press), then reproductive parameters for females in different areas may vary and have a profound effect on calculated replacement potentials. The incidental mortality and probable depletion of long-lived dolphins, that invest many years in the care and socialization of their young and are resident in areas with which they are familiar, is of concern. The future management of the Natal bottlenose dolphin population requires accurate population figures and an unbiased estimate of age and sex structure. Regular aerial, boat, and shore-based surveys along the Natal coast are needed to define the former. The latter is best obtained through a combination of intensive field obser- vational work on free-ranging dolphins and a continued monitoring of captured animals. Acknowledgments We gratefully acknowledge Shantal Koch and Sabine Klages for many hours spent in preparing histological slides, counting dentine layers, and sectioning ovaries. We appreciate the help of Dr. T. Kasuya for reading a selection of teeth so that we could calibrate our own techniques. Our thanks to the Director and staff of the Natal Sharks Board for their cooperation in collecting animals from the nets. Citations Arvy, L. 1974 Mammary glands, milk and lactation in cetaceans. In Pilleri, G. (ed.), Investigations on Cetacea, vol. V, p. 1.57-202. Inst. Brain Anat., Univ. Berne, Berne, Switzerland. Beddington, J.R., R.J.H. Beverton, and D.M. Lavigne (editors) 1985 Marine mammals and fisheries. George Allen & I'lnvin, London, 354 p. Brodie, P.F. 1969 Duration of lactation in Cetacea: An indicator of required learning? Am. Wildl. Nat. 821:312-314. Bryden, M.M. 1972 Growth and development of marine mammals. /« Har- rison, R.J. (ed.), Functional anatomy of marine mammals, vol. 1, p. 1-79. Acad. Press, London and New York. Cockcroft, V.G., and G.J.B. Ross 1990a Food and feeding of the Indian C)cean bottlenose dolphin off southern Natal, South Africa. In Leatherwood. S., and R. Reeves (eds.). The bottlenose dolphin, p. 295-308. Acad. Press. NY. 1990b Observations on the early development of a captive bottlenose dolphin calf. In Leatherwood, S., and R. Reeves (eds.), The bottlenose dolphin, p. 461-478. Acad. Press, NY. In press Bottlenose dolphins in the Natal shark nets, 1980 to 1987: Catch rates and associated contributing factors. In Leatherwood, S., and G.P. Donovan (eds.). Cetaceans and ceta- cean research in the Indian Ocean sanctuary. Rep. Int. Whal- ing Comm. Cockcroft, V.G., A.C. De Kock, D.A. Lord, and G.J.B. Ross 1989 Organochlorines in bottlenose dolphins from the east coast of southern Africa. S. Afr. J. Mar. Sci. 8:207-218. Cockcroft. V.G., G.J.B. Ross, and V.M. Peddemors In press Bottlenose dolphin distribution in Natal's coastal waters. S. Afr. J. Mar. Sci. 9. Cockcroft and Ross: Tursiops truncatus off soutfiern Africa 301 Danilevsky, N.N., and V.P. Tyutyunnikov 1968 Nekotorye dannye sovremennom sostoyanii zapasa del'finov Chemogo Morya. (Present state of Black Sea dolphin described.) Rybn. Khoz. Nov. 1968:25-27. [Translations on USSR Fishing Industry and Marine Resources. No. 24, Joint Research Service, 1969, 9 p.] (In Perrin and Reilly 1984.) Essapian, F.S. 1963 Observations on abnormalities of parturition in captive bottlenosed dolphins, Tursiops truncatus, and concurrent behaviour of other porpoises. J. Mammal. 44:405-414. Gurevich, V.S. 1977 Post-natal behavior of an Atlantic bottlenosed dolphin calf {Tursiops truncatus, Montagu) born at Sea World. In Ridgeway, S.H., and K. Benirschke(eds.), Breeding dolphins: Present status, suggestions for the future, p. 168-184. Rep. MMC-76/07, U.S. Mar. Mammal Comm., Wash. DC. Harrison, R.J., R.L. Brownell, Jr., and R.C. Boicc 1972 Reproduction and gonadal appearances in some odon- tocetes. In Harrison, R.J. (ed.). Functional anatomy of marine mammals, vol. 1, p. 361-429. Acad. Press, NY. Hohn, A.A. 1980 Age determination and age related factors in the teeth of western North Atlantic bottlenosed dolphins. Sci. Rep. Whales Res. Inst. Tokyo 32:39-66. Hui, C.A. 1980 Variability of dentin deposits in Tursiops truncatus. Can. J. Fish. Aquat. Sci. 37:712-716. Irvine, A.B., M.D. Scott, R.S. Wells, and J.H. Kaufmann 1981 Movements and activities of the Atlantic bottlenose dolphin, Tursiops truncatus, near Sarasota, Florida. Fish. Bull., U.S., 79:671-688. Kasuya, T. 1976 Reconsideration of life history parameters of the spotted and striped dolphins based on cemental layers. Sci. Rep. Whales Res. Inst. Tokyo 28:73-106. 1985 Fishery-dolphin conflict in the Iki Island area of Japan. In Beddington, J.R., R.J.H. Beverton, and D.M. Lavigne, (eds.). Marine mammals and fisheries, p. 253-272. George Allen & Unwin, London. Kasuya, T., and Y. Izumisawa 1981 The fishery-dolphin conflict in the Iki Island, Japan area. Rep. to U.S. Mar. Mammal Comm., Contract NM1533791-7 (NTIS). 31 p. (/« Perrin and Reilly 1984.) Kasuya, T., and H. Marsh 1984 Life history and reproductive biology of the short-fmned pilot whale, Glohicephala macrorhynchiis, off the Pacific coast of Japan. Rep. Int. Whaling Comm., Spec, issue 6, p. 259-310. Kasuya, T., T. Tobayama, T. Saiga, and T. Kataoka 1986 Perinatal growth of delphinoids: Information from aqua- rium reared bottlenose dolphins and finless porpoise. Sci. Rep. Whales Res. Inst. Tokyo, 37:85-97. Lear, R.J., and M.M. Bryden 1980 A study of the bottlenose dolphin, Tursiops tru7icat>is, in Eastern Australian waters. Occas. Pap. 4, Aust. Nat. Parks Wildl. Serv., 25 p. (In Perrin and Reilly 1984.) Marsh, H., and T. Kasuya 1986 Evidence for reproductive senescence in female ceta- ceans. In Donovan, G.P. (ed.), Behaviour of whales in rela- tion to management, p. 57-74. Rep. Int. Whaling Comm., Spec, issue 8, p. 57-74. McBride, A.F., and H. Kritzler 1951 Observations on pregnancy, parturition, and post-natal behavior in the bottlenose dolphin. J. Mamma!. 32(3):251-266. Mitchell, E. 1975 Porpoise, dolphin and small whale fisheries of the world: Status and problems. Monogr. 3, Int. Union Conserv. Nat. Nat. Res., Unwin Bros., Surrey, England, 129 p. Mitchell, E., and V.M. Kozicki 1984 Reproductive condition of male sperm whales, Physeter macrocephalus, taken off Nova Scotia. Rep. Int. Whaling Comm., Spec, issue 6, p. 243-252. Norris, K.S. (editor) 1961. Standardized methods for measuring and recording data on the smaller cetaceans. J. Mammal. 42:471-476. Northridge, S., and G. Pilleri 1986 A review of human impact on small cetaceans. In Pilleri, G. (ed.). Investigations on Cetacea, vol. XVIII, p. 221-261. Inst. Brain Anat., Univ. Berne, Berne, Switzerland. Perrin, W.F., and A.C. Myrick (editors) 1980 Age determination of toothed whales and sirenians. Rep. Int. Whaling Comm., Spec. Issue 3, 229 p. Perrin, W.F., and S.B. Reilly 1984 Reproductive parameters of dolphins and small whales of the family Delphinidae. Rep. Int. Whaling Comm., Spec, issue 6, p. 97-133. Perrin, W.F., J.M. Coe, and J.R. Zwiefel 1976 Growth and reproduction of the spotted porpoise, StetwUa attenuata, in the offshore eastern tropical Pacific. Fish. Bull., U.S. 74:229-269. 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. Reilly, S.B.. and J. Barlow 1986 Rates of increase in dolphin population size. Fish. Bull., U.S. 84:527-533. Ross, G.J.B. 1977 The taxonomy of bottlenosed dolphins Tursiops species in South African waters, with notes on their biology. Ann. Cape Prov. Mus. Nat. Hist. 11(9): 135-194. 1984 The smaller cetaceans of the south east coast of southern Africa. Ann. Cape Prov. Mus. Nat. Hist. ll(14):259-327. Ross, G.J.B.. and V.G. Cockcroft 1990 Comments on Australian bottlenose dolphins and the tax- onomic status of Tursiops aduncus (Eherenberg. 1832). In Leatherwood. S., and R. Reeves (eds.). The bottlenose dolphin, p. 101-128. Acad. Press, NY. Ross, G.J.B., V.G. Cockcroft, D.A. Melton, and D.S. Butterworth 1989 Population estimates for bottlenosed dolphins Tursiops truncatus in Natal and Transkei waters. S. Afr. J. Mar. Sci. 8:119-130. Saaynuui, G.S., and C.K. Tayler 1977 Observations on the sexual behavior of Indian Ocean bottlenosed dolphins (Tursiops aduncus). In Ridgeway, S.H., and K. Benirschke (eds.). Breeding dolphins: Present status, suggestions for the future, p. 1 13-129. Rep. MMC-76/07, U.S. Mar. Mammal Comm., Wash. DC. Sacher, G.A., and E.F. Staffeldt 1974 Relation of gestation time to brain weight for placental mammals: Implications for the theory of vertebrate growth. Am. Nat. 108(963):593-615. (In Perrin et al. 1977.) Sergeant, D.E., D.K. Caldwell, and M.C. Caldwell 1973 Age, growth, and maturity of bottlenosed (Tursicrps trun- catus) from northeast Florida. J. Fish. Res. Board Can. 30:1009-1011. 302 Fishery Bulletin 88(2). 1990 Tavolga, M.C., and F.S. Essapian 1957 The behavior of the bottle-nosed dolphin {Tursiops trun- catus): Mating, pregnancy, parturition and mother-infant behavior. Zoologica (NY) 42(1):11-31. Wursig, B. 1978 Occurrence and group organization of Atlantic bottlenose porpoises {Tiirswps truncatus) in an Argentine bay. Biol. Bull. (Woods Hole) 154:348-359. Abstract.- To estimate mortality and abundance of walleye pollock Theragra chalcogranuna larvae in Shelikof Strait, Alaska, during spring 1981, a diffusion-advection model, combined with growth and death of fish larvae, was applied. Physical parameters (diffusion coefficients and advection rates) were derived from the distributional variances and centroids of fish larvae collected in ichthyoplankton surveys. The diffu- sion coefficient and the advection rate in the along-strait direction were 65.2 km-/day and 2.7-4.2 km/day, respectively, which compared favor- ably with values obtained from moored current-meter data. Simula- tion revealed that the expected dis- tribution of larvae was similar to that observed from ichthyoplankton sam- plings, and that around 20% of the larvae drifted out of the survey area in Shelikof Strait within the 1 -month sampling period. The larval fraction dispersed out of the survey area was used to revise larval mortality and abundance estimates. Revised mor- tality (0.070/day) was close to that (0.063/day) determined from examin- ing larval patches. The simulation in this paper resulted in an increase by a factor of 1.5 in the estimated total larval abundance compared with ear- lier estimates and field observations. Oceanic Dispersion of Larval Fish and Its Implication for Mortality Estimates: Case Study of Walleye Pollock Larvae in Shelikof Strait, Alaska Suam Kim Alaska Fisheries Science Center. National Marine Fisheries Service, NOAA 7600 Sand Point Way NE, Seattle, Washington 981 15-0070 Present address: Korea Ocean Research and Development Institute Polar Research Laboratory. An San, P O Box 29, Seoul 425-600, KOREA Bohyun Bang Division of Physical Oceanography and Environmental Engineering Virginia Institute of Marine Science, College of William and Mary Gloucester Point, Virginia 23062 Manuscript accepted 24 January 1990. Fishery Bulletin, U.S. 88:303-311. Most marine fish have a period of planktonic existence during egg and larval stages. Since early in this cen- tury it has been believed that survival during early life determines year- class strength and recruitment vari- ability to fisheries (Hjort 1914). Fur- thermore, recruitment processes are quite complex because the biological and environmental factors which act on eggs and lar-vae are closely related (Wooster et al. 1983). Hence, the relationship between organisms and their environment is critical for un- derstanding recruitment variability. The observed patterns of egg and lar- val distributions can be considered to be the result of a combination of fun- damental processes, including spawn- ing time and location, advection, dif- fusion, growth and mortality. These parameters could be identified using biological data. The early-life stages of walleye pol- lock Theragra chalcogramma whose biomass is the largest of a single spe- cies in world fisheries (Sharp 1987), have been the objects of considerable research in recent years. In Shelikof Strait (Fig. 1), about 90% of the eggs were produced between 25 March and 15 April 1981, and they have approximately a 2-week embryonic period at 5°c"(Kim 1989). Spawning produces a patch of planktonic eggs and larvae that can be followed as they develop and are advected in pre- vailing currents toward the south- west. In Shelikof Strait walleye pol- lock eggs exist at depths below 150 m due to their high specific gravity. Their transport rate from the spawn- ing area is very small because of weak circulation in deep water (Ken- dall and Kim 1989). Eggs of late developmental stage would move upward fast due to the decreased specific gravity of old eggs, and the eggs hatch mid-depth in the water column (Kim 1987). Also the specific gravity of newly hatched larvae is continuously decreasing, so that most larvae are found within the upper 60 m of the surface (Kendall et al. 1987). Kim and Kendall (1989) de- scribed the distribution and transport pattern of larvae in Shelikof Strait during spring. Young larvae occupy a relatively small area and form a dense patch, whOe older lai'vae spread over a broader area in Shelikof Strait, showing the importance of diffusion on the larval patch. These larval patches have been identified for at 303 304 Fishery Bulletin 88(2). 1990 157" W - 57° 30' )80" 160" W 140" 120" ^"■•^F': 60" N 50" -^ii"«A SHEUKOF \ STRAIT t- 58°30N' - 56° 30 155° 153° Figure 1 Typical locations of sampling stations (•) and the grid pat- tern used for the computer simulation. The flux of larvae through the solid lines is zero. The dashed lines are open boundaries of the model that permit the loss of larvae. least a month after hatching (Incze et al. 1989). Assum- ing no drift or dispersion of larvae away from the sam- pling area, the instantaneous daily mortality was esti- mated as 0.086 (Kim and Gunderson 1989). The effects of advection and diffusion on the larval mass, however, can cause errors in estimation of mortality and abun- dance. Hence, the distribution and abundance of wall- eye pollock larvae in Shelikof Strait should be recon- sidered in the light of oceanic diffusion-advection theory (McGurk 1989). The main objectives of this paper are: 1 Estimation of diffusion coefficients and advection rates of walleye pollock larvae based on ichthyoplank- ton distribution in Shelikof Strait; 2 Description of expected larval distribution and abundance with time from a computer simulation using a diffusion-advection model; and 3 Reestimation of larval mortality and expected lar- val abundance reported in Kim and Gunderson (1989) and comparison of these results with field observations. Background of theories and model In general, the simplest approach to the diffusion prob- lem for particles in a fluid medium follows from the assumption that the rate of diffusion is directly pro- portional to the local concentration gradient (Okubo 1980). Walleye pollock larvae in Shelikof Strait are advected and diffused in the upper 60 m during early larval stages (Kendall et al. 1987), and the swimming ability of larvae less than 10 mm is assumed to be in- consequential. Therefore, a horizontal two-dimensional diffusion-advection model is applicable to changes in the distribution and abundance of young walleye pol- lock larvae in Shelikof Strait. Current speeds (ii and v) and diffusion coefficients (K^ and A',J are set con- stant, with the assumption of a homogeneous turbu- lence field and negligible horizontal divergence. Newly hatched larvae (i.e., the source material in the diffusion- advection model) are assumed to be produced daily at a fixed location near the center of Shelikof Strait (Kim and Kendall 1989). Once larvae that dispersed from the source point arrived at the Alaska Peninsula coast, they were required to remain there, since the shallow coastal region is considered a nursery area of young fish (Walters et al. 1985). Since larvae grow as they drift, the simulated distrilnition and abundance of larvae can he divided into several size groups. The partial differen- tial equation, with initial condition Ci{x,y, Ti ) = 0 and boundary conditions as described above, is Kim and Bang: Simulation of walleye pollock distribution in Shelikof Strait 305 dCi a2c, a^c, ac, dt 3x2 ^ Qy2 gj. - V - rCi dy (1) where C, = larval concentration of t-th cohort (number/m^), X and y = along- and cross-strait coordinates, K^ and Ky = along- and cross-strait turbulence diffusion coefficients, u and V = along- and cross-strait velocity components, r = instantaneous daily mortality of larvae, and Ti = starting time of i-th cohort larval production. The solution using Laplace transform is C,{x,y,t) = ^= f ^^i^ent-T) (^-K iKAt-T) (y-B,f AKJt-T) ■ dT (2) where C,{x,y,t larval concentration of i-th cohort at time t at point {x,y), P,{T) = i-th cohort's larval production rate at time T {T^ + ln(l-/) AT (3) where Z*(L) = revised instantaneous daily larval mortality, AT = time difference in days (24 days) between surveys, ,-~. / 306 Fishery Bulletin 88(2), 1990 di and dp = observed larval densities during late April and late May surveys, respectively, and / = out-fraction. Note that the first term of the right side means the daily mortality rate for the closed population, which was used in Kim and Gunderson (1989), and that the second term is the correction factor due to dispersion. Once the new mortality rate was determined, it was used for the second simulation to obtain a revised estimate of larval abundance in late May 1981. Parameter estimation Some parameters for the model were available from other studies. Kim and Gunderson (1989) found two dominant larval cohorts during April and May surveys, and assumed that 4-5 mm larvae in late April had grown to 8-9 mm, and regarded this group as cohort 1. Also, the 5-6 mm size class in late April and the 9-10 mm size class in late May were treated as cohort 2. By comparing the larval abundances of these two cohorts, they estimated an instantaneous daily larval mortality of 0.086 and a daily growth rate of 0.17 mm. Daily larval production was derived from daily egg production and time-specific egg mortality during development (Kim and Gunderson 1989). A rapid in- crease in daily larval production occurred in mid-April, and most larvae were produced during late April (Table 1). Toward the end of the spawning season, larval pro- duction decreased but tended to be prolonged due to decreased egg mortality late in the spawning season. Advective velocities and turbulent diffusivities were derived by examining changes in the distributional cen- troids and variances of several larval size groups. We treated the distribution of a certain size of larvae at a specific time as a single dispersing system in Shelikof Table 1 Estimates of daily larval product on of walleye jollock after 14-i:iay iiic ubation period in Shelikt f Strait during spring 1981, using estimated daily egg produ ■tion and time -specific egg mortality in Kim and Gundersoi (1989). Dates 95 and 144 1 denote 5 April and 24 May, respectively. Daily larval Daily larval Date of production Date of production year (xlO" year (xlO") 9.5 23 120 1103 96 48 121 1027 97 77 122 938 98 111 123 835 99 151 124 717 100 197 125 584 101 250 126 435 102 311 127 270 103 380 128 37 104 458 129 38 105 546 130 37 106 1016 131 38 107 1557 132 38 108 2174 133 39 109 2873 134 39 110 3087 135 38 111 3311 136 38 112 3545 137 38 113 3790 138 37 114 3294 139 37 115 2711 140 36 116 2034 141 35 117 1257 142 34 118 1217 143 33 119 1166 144 32 Strait. The change in the centroids in along- and cross- strait coordinates for three size groups (4-5, 5-6, and 6-7 mm) from each survey was used to estimate advec- tion, assumming that larvae were hatched in the same area and that these larvae were not flushed out of Comparison of walleye pollock larv; Shelikof Strait during spring 1981. d mortal Table 2 ity rates, advection rates. and diffusioi coefficients in Instantaneous daily mortality Advection rate (km/day) Diffusion coefficients (km^/day) K, a; u V April May April May This paper 0.070 Reed et al. (1989) 0.063 4.2 4.3 2.7 1.3 0.4 65.2 3.6 43.2 Kim and Bang Simulation of walleye pollock distribution in Shelikof Strait 307 Shelikof Strait. Estimates of u and v, from regression analysis (see Table 1 in Kim and Kendall (1988) for data) were 4.2 and 1.3 km/day in April, and 2.7 and 0.4 km/day in May, respectively (Table 2). Because the variance of the horizontal distribution is a suitable measure of the spread of the substance (Bowden 1983, Okubo 1971), the change in variance (i.e., Sf and S^ for along- and cross-strait directions) with time pro- vides a reasonable measure of the diffusion coefficient; K, K„ IrfSL Q2 e2 2 dt 2At 1 dS\ ^ q2 Q2 2 dt 2At (4) (5) Table 3 Spatial variances of larval distributions of two maj or cohorts of walleye pollock larvae in Shelikof Strait during late April and late May 1981. The larval sizes of cohorts 1 and 2 in late April are 4-5 mm and 5 -6 mm, and they grow to 8- -9 mm and 9-10 mm in late May, respectively. Variance (km^) late April late May Cohort 1 Si 1292 5114 SI 563 609 Cohort 2 SI 3066 5500 s; 302 606 where S?,o. •Sjn- Sjjo, and Sjn are the spatial vari- ances at time ^0 and time il in along- and cross-strait directions, and M = tl - tO. The variances of two dominant larval cohorts were calculated as in Kim (1987). Variances of larval distribution were much in- creased in a month, and along-strait components were dominant compared with those in the cross-strait direc- tion (Table 3). The estimated Kj. for these two cohorts were 79.6 and 50.7 km-/day (average 65.2). For Ky, the values were 1.0 and 6.3 km-/day (average 3.6) (Table 2). Simulation results Spatial distribution and abundance of larvae The model distribution of larvae in late May after 50 days of simulation was similar to that observed. Two major larval cohorts, 8-9 mm and 9-10 mm size groups, were selected for comparing the model and observed distributions (Fig. 2). In general, the simulated cen- troids of distribution were close to the densest patches of the larvae, and the area of larval distribution and contour levels of the larval concentrations were not very different from those observed. The simulation demonstrated the elliptical pattern of distribution, which was elongated in the along-strait direction, and the southwesterly movement of centroids from the main hatching area. The first (8-9 mm) and second (9-10 mm) cohorts drifted 92 km and 114 km, respec- tively, from the source point after hatching. The maximum size of larvae was 12 mm, and the abundance of larvae and the out-fraction in each size group were computed. Comparing the simulated values with the observed ones, we found that the results were very close (Table 4). Also, as expected, the effect of out-fraction on larval distribution was more important for the larger size group than the smaller group. This was caused by the larger larval size having a higher out-fraction value. Negligible amounts of small larvae were advected from the simulation box, but over 50% of the large larvae were removed. Among total larval abundance, about 20% were flushed out of the simula- tion area. This concept of out-fraction due to diffusion and advection might change the larval abundance and mortality previously reported by Kim and Gunderson (1989). Even though their derived estimates agreed well with observed ones, their results should be recon- sidered because the areas involved for abundance esti- mates were not the same. Our study revealed that their sampling area, which was similar to our simulation box, was only part of the area of larval occurrence in Sheli- kof Strait. Therefore the total larval abundance should be higher than they observed, and the simulated abun- dance within the simulation box should be close to that observed. In Table 4, the reason for the smaller larval abundance in the simulation box (3.37 x 10'-) than observed (4.15 x 10'^) might be due to their overesti- mate of the mortality rate. Re-estimation of larval mortality and abundance The first approximation of larval mortality has been recalculated using the out-fraction in Table 4. By ap- plying the out-fractions of the two major cohorts to Equation (3), we computed revised instantaneous daily mortalities of 0.081 and 0.059 from the first and sec- ond cohorts, respectively (average 0.070). Assuming no significant change in diffusion coefficients, the revised mortality rate was used for the second computer simu- lation of the diffusion-ad vection model. Figure 3 re- vealed that the second simulation resulted in a 50% increase in the total larval abundance (6.21 x 10'-) in Shelikof Strait, compared with the first simulation (4.23 X 10'- from Table 4). The size-specific abundance 308 Fishery Bulletin 88(2). 1990 Alaska Peninsula Alaska Peninsula Figure 2 Contours of observed ( ) and estimated ( — ) distributions of (a) 8-9 mm and (b) 9-10 mm walleye pollock larvae during late May 1981. Source points of larvae are indicated by ( + ) in each map. of larvae within the simulation box was compared with the observed values, because the area used in the simulation could be regarded as the survey area. In general, the trends in size abundance curves, as well as the absolute abundances, were very similar to one another. The abundance of two major cohorts, both observed and simulated, consisted of about 50% of the total abundance. Discussion The application of a diffusion-advection model to ex- amine the dispersal of larvae helps not only to explain spatiotemporal distribution of abundance but also to revise estimates of population parameters such as lar- val mortality. Difficulties in determining parameter values, however, often arise in this kind of study. In describing plankton distribution, the physical proper- ties (diffusion coefficients and advection rates) often Kim and Bang Simulation of walleye pollock distribution in Shelikof Strait 309 Table 4 Simulated abundances of walleye pollock larvae in Shelikof Strait during late May 1981 resulted from the first computer simulation, using a daily instantaneous mortality rate of 0.086 and observed values derived from Kim (1987). Simulated larval abundance (xlO ') Observed larval Larval size (mm) abundance (xlO'=) Total Simulation box Out-fraction <4 0.033 0.033 0.0000 0.004 4-5 0.157 0.157 0.0006 0.057 5-6 0.112 0.111 0.0128 0.111 6-7 0.627 0.593 0.0538 0.496 7-8 1.011 0.881 0.1281 1.221 8-9 1.509 1.120 0.2576 1.350 9-10 0.701 0.436 0.3786 0.627 10-11 0.073 0.035 0.5200 0.204 11-12 0.008 0.003 0.6367 0.058 Total 4.230 3.369 0.2036 4.149 24 23 22 2 1 20 1 8 17 „ 1.6 % 1 5 X 1 4 g 13 C I 12 I 1 1 !^ 10 TO I 09 08 07 06 05 04 03 02 0 Total abundance Observed: 4 15 X 10 im Simulated: 4.85 X lo'^ (sampling area) \ 1 Simulated : 6 21 X 10'' (total area) 6 7 7 8 8-9 Larval size (millimeters) 11-12 Figure 3 Simulated larval abundance from the second computer simulation using a revised mortality rate of 0.070/day, and observed larval abundance rn Table 3. Notice the difference in the simulated abundances from the sampling area and the total area. are calculated from oceanic current data, wind speed, temperature distribution, or results of dye experiments (Talbot 1974, Talbot 1977, Power and McCleave 1983, Sundby 1983), even though they do not represent the actual diffusion and advection of dispersing organisms. The collection of diffusion coefficients and advection rates from several sources is very important to under- stand the characteristics of the oceanic situation. Aside from measuring the dispersion using inert tracers, the use of plankton sampling data to derive these param- eters has been limited because of the complexity of bio- logical systems in the sea. The mobility and mortality of larvae may bias in estimating such parameters. The em- phasis of this paper is how biological sampling data can be used for estimating physical properties, when the swimming ability of the larvae is not significant. 310 Fishery Bulletin 88(2), 1990 Reed et al. (1989) estimated parameters based on cur- rent measurement and larval patch in Shelikof Strait. By examining the abundance in small areas around the densest larval patches found in April and May surveys, they estimated an instantaneous mortality rate of 0.063/day and an advection rate of 4.3 km/day (Table 2), which are in excellent agreement with our esti- mates. Also, the estimation of eddy diffusivity in the along-strait direction of 43.2 km^/day by Reed et al. (1989) did not differ greatly from our mean value of 65.2 km^ (Table 2), even though the former was de- rived from moored current meter data in the surface layer (56 m) and the latter from larval distribution. These estimates are realistic only in a mean sense, because they vary in both time and space. The diffusion process tends to destroy larval aggre- gation until larvae reach a certain size, so that patch- iness will decrease with time. The Lloyd Patchiness Index (LPI) has been frequently used for describing ag- gregations of organisms (Lloyd 1967), and Kim (1987) discussed changes in LPI of walleye pollock larvae as they grew. The smaller size groups of larvae usually had a higher LPI, but it decreased until a size of about 10 mm because of dispersal of the larval patch. For sizes greater than 10 mm, although their contribution to the total larval abundance was very small, the LPI increased with length, perhaps because of reaggrega- tion of the larvae as their swimming ability increased. If a larval retention mechanism worked for large lar- vae (10-12 mm) in Shelikof Strait due to increased swimming ability, that might explain the higher larval abundance from field samplings than that from the simulation shown in Figure 3. Similar examples of lar- val aggregation (i.e., initially patchy, dispersing until a larval size of around 10 mm, and then patchy again) were reported for northern anchovy and jack mackerel off California (Hewitt 1982). The expected total larval abundance from the first simulation was almost identical to that in Kim and Gunderson (1989) because the same parameter values were used. By adding the concept of diffusion and advection to their model, elaborations on mortality and expected abundance of walleye pollock larvae were made. Based upon good agreement between observed and simulated results, this paper has emphasized that dispersion (or emigration) of organisms is important in the field of population dynamics. Acknowledgments We are indebted to many people for helping us com- plete this manuscript. We would like to thank P. Sta- beno, R. Reed, and J. Schumacher at the NOAA Pacific Marine Environmental Laboratory and A. Okubo at State University of New York at Stony Brook for their comments and advice. Also we are greatful to all FOCI (Fisheries Oceanography Coordinated Investigation) members of AFSC (Alaska Fisheries Science Center) who have prepared and provided us the necessary data. Special thanks is given to A. Kendall (AFSC) who funded this project and gave us valuable comments. Citations Bowden, K.F. 1983 Physical oceanography of coastal waters. John Wiley, NY, 302 p. Gradshteyn, I.S., and I.M. Ryzhik 1980 Table of integrals, series, and products. Acad. Press, NY, 1160 p, Hewitt, R.P. 1982 Spatial pattern and survival of anchovy larvae: Implica- tion of adult reproductive strategy. Ph.D, diss.. Univ. Calif,, San Diego, 187 p. Hjort, J. 1914 Fluctuations in the great fisheries of northern Europe viewed in the light of biological research. Rapp. P.-V, Cons. Inst, Explor, Mer 20:1-13, Incze. S.I., A.W. Kendall, Jr., J.D. Schumacher, and R.K. Reed 1989 Interactions of a mesoscale patch of larval fish {Tlwnignt chdlriiiframma) with the Alaska Coastal Current. Continen- tal Shelf Res, 9:269-284, Kendall. A.W., Jr.. M.E. Clarke, M.M. Yoklavich, and G.W. Boehlert 1987 Distribution, feeding, and growth of larval walleye pol- lock, Theragra chalcogramma, from Shelikof Strait, Gulf of Alaska. Fish. Bull., U,S, 8.5:499-.521. Kendall, A.W., Jr., and S. Kim 1989 Buoyancy of walleye pollock (Tlwragra chalcogramma) eggs in relation to water properties and movement in Shelikof Strait, GuLf of Alaska, In Beamish, R..1,, and G,A, MacFarlane (eds,). Effects of ocean variability on recruitment and an evalua- tion of parameters used in stock assessment model, p. 169- 180. Can. Spec. Publ, Fish, Aquat, Sci. 108, Kim. S. 1987 Spawning behavior and the early life history of walleye pollock. Theragra chalcogramma, in Shelikof Strait. Gulf of Alaska, in relation to oceanographic factors, Ph,D, diss,. Univ, Wash,, Seattle, 221 p, 1989 Early life history of walleye pollock. Theragra chalco- gramma, in the Gulf of Alaska, In Proceedings of the Inter- national Symposium on the Biology and Management of walleye pollock. Anchorage, Alaska, Nov, 14-16, 1988, p. 117- 139, Alaska Sea Grant Coll, Prog., Univ, Alaska, Fairbanks. Kim, S., and D.R. Gunderson 1989 Cohort dynamics of walleye pollock (Theragra chalco- gramma) in Shelikof Strait, Gulf of Alaska, during the egg and larval period. Trans, Am, Fish, Soc, 118:264-273, Kim. S.. and A.W. Kendall. Jr. 1988. Distribution and transport of walleye pollock. Theragra chalcogramma, in relation to environmental factors. Early Life History Symposium 1988 Paper 90, Int, Counc, Explor, Sea, Bergen. 18 p. 1989 Distribution and transport mechanism of walleye pollock. Theragra chalcogramma, in relation to environmental factors, Rapp, P,-V, Reun, Cons, Int, Explor, Mer 191:127-136. Kim and Bang Simulation of walleye pollock distribution in Shelikof Strait Lloyd, M. 1967 Mean crowding. J. Anim. Ecol. 36:1-30. McGurk. M.D. 1989 Advection, diffusion and mortality of Pacific herring larvae, Clupea harengus pallasi, in Bamfield Inlet, British Columbia. Mar. Ecol. Prog. Ser. .51:1-18. Okubo, A. 1971 Oceanic diffusion diagrams. Deep-Sea Res. 18:789-802. 1980 Diffusion and ecological problems: Mathematical models. Springer- Verlag, Berlin, Heidelberg, NY, 254 p. Power, J.H. 1986 A model of the drift of northern anchovy, Engra idis mor- (i«.r, larvae in the California current. Fish. Bull., U.S. 84: ,585-603. Power, J.H., and J.D. McCleave 1983 Simulation of the north Atlantic ocean drift oi Anguilla leptorephali. Fish Bull., U.S. 81:483-500. Reed, R., L.S. Incze, and J.D. Schumacher 1989 Estimation of the effects of flow on the modes of disper- sion of larval pollock, Theragra chatcogramma, in Shelikof Strait, Alaska. In Beamish, R.J., and G.A. McFarlane (eds.), Effects of ocean variability on recruitment and an evaluation of parameters used in stock assessment models, p. 239-246. Can. Spec. Publ. Fish. Aquat. Sci. 108. Scheid, F. 1968 Schaum's outline of theory and problems of numerical analysis. McGraw-Hill, 422 p. Sharp, G.D. 1987 Climate and fisheries: Cause and effect or managing the long and short of it all. In Payne, A.I.L., et al. (eds.), The Benguela and comparable ecosystem. S. Afr. ,J. Mar. Sci. 5:811-8.38. Sundby, S. 1983 A one-dimensional model for the vertical distribution of pelagic eggs in the mixed layer. Deep-Sea Res. 30(6A): 645-661. Talbot. J.W. 1974 Diffussion studies in fisheries biology. In Harden Jones, F.R. (ed.). Sea fisheries research, p. 31-54. Wiley & Sons, NY. 1977 The dispersal of plaice eggs and larvae in the southern Bight of the North sea. J. Cons. Int. E.\-plor. Mer 37:221-248. Walters, G.E.. G.B. Smith, P.A. Raymore, Jr.. and W. Hirschberger 1985 Studies of the distribution and abundance of juvenile groundfish in the northwestern Gulf of Alaska, 1980-82. Part II, Biological characteristics in the extended region. NOAA Tech. Memo. NMFS F/NWC-77, Northwest Alaska Fish. Cent., Natl. Mar. Fish. Serv., NOAA, Seattle, WA 98115-0070, 95 p. Wooster, W.S.. K. Banse. and D.R. Gunderson 1983 On the development of strategies for the study of ocean fish interactions. In Wooster. W.S. (ed.), From year to year, p. 199-206. Wash. Sea Grant Publ, Univ. Wash., Seattle. Abstract.- Models are formu- lated for estimating tag-shedding rates from recapture records for eight dou- ble-tagging experiments with south- ern bluefin tuna (Thunniis maccoyii) in three Australian fishing areas. These models incorporate either a constant or time-varying rate of tag shedding and allow for the possibil- ity of immediate tag shedding. Likeli- hood ratio tests are used to select the most parsimonious model for each data set. The probability of a tag be- ing shed after 4 years at liberty (the length of the most recent experi- ments) varied from around 0.2 for the 1983 and 1984 experiments to 0.5-0.7 for the experiments carried out in the 1960s and 1970s. Although a single, best-fitting model was se- lected in all but one experiment and despite the large total numbers of recoveries, precise estimates of long- term shedding rates could not be ob- tained because there were relatively few long-term data. This has signifi- cant implications for analyses of tag- return data that give heavy weight- ing to long-term recaptures. Tag Shedding by Southern Bluefin Tuna Thunnus maccoyii John Hampton Geoffrey P. Kirkwood CSIRO Division of Fisheries, Marine Laboratories GPO Box 1538, Hobart, Tasmania 7001, Australia Manuscript accepted 11 December 1989. Fishery Bulletin. U.S. 88:313-321. Southern bluefin tuna Thunnus mac- coyii were tagged throughout the 1960s in three areas of major Austra- Han commercial fishing activity: off the south coast of Western Australia (WA), in the Great Australian Bight off South Australia (SA), and off the south coast of New South Wales (NSW). During the 1970s, tagging took place mainly off WA. The aim of these experiments was to provide information on movements, growth, and mortality of the southern bluefin. In 1983 and 1984, new experiments were initiated off WA and SA in order to answer more specific ques- tions relating to fishery interactions, yield-per-recruit, and schooling be- havior, as well as to update informa- tion obtained from the previous ex- periments (Majkowski and Murphy 1983). Apart from the period prior to 1963, almost all fish were double tagged. Quantitative analyses of tagging data that require an estimate of the total number of recaptures of tagged fish (e.g., fishery interactions, yield- per-recruit, and mortality) need to take account of tag loss due to shed- ding. As reviewed by Wetherall (1982), this problem can be approached in two ways. If estimation of mortality rates is the main objective of the analysis, one can attempt to model the entire process of removal of tags from the population directly by incor- porating into the model parameters that account for all souj-ces of removal, i.e., natural mortality, fishing mor- tality, permanent emigration, non- reporting of recaptured tags, and tag shedding. Alternatively, tag-shed- ding rates may be estimated indepen- dently from a double-tagging experi- ment and the recovery data adjusted accordingly before proceeding with further analyses (other losses, e.g., non-reporting, may also require in- dependent estimation). A disadvan- tage of the former approach is that, even if simple functions are chosen to describe mortality, emigration, and tag shedding, the model will inevitably be parameter-laden, and therefore difficult to estimate. In any case, such models will not necessar- ily lend themselves to analyses of interactions and yields-per-recruit, where the mortality and emigration processes are embedded in the recap- ture statistics and do not necessarily require explicit resolution (Majkow- ski et al. 1984). The aim of this paper is to estimate tag-shedding rates for a number of southern bluefin tuna double-tagging experiments, with a view to using these estimates as the basis for cor- recting for tag shedding in other analyses. Hynd (1969) made a prelim- inary estimate of the tag-shedding rate, assumed constant, based on recoveries up to 1968 only. Subse- quently, Kirkwood (1981) developed generalized tag-shedding models and analyzed recoveries of southern blue- fin tuna that were double tagged dur- ing the period 1962-76 but excluded all tagging that was contracted to fishermen. In both cases, these anal- yses used tag-recovery data that were grouped by time of recovery. A reevaluation of southern bluefin tuna tag shedding is now appropriate because, (1) for a number of present 313 314 Fishery Bulletin 88(2). 1990 Table 1 Double-taggi ng experiments of southern bluefin tuna in the New So iith Wales (NSW), South Australia (SA). and Western Australia | (WA) areas. Number' recovered Experiment Fishing Tagging Number number Area Years Carried out by method method released with 2 tags with 1 tag 1 NSW 1963-70 CSIRO Pole Board' 2770 253 153 2 NSW 1963-70 Fishermen under contract to CSIRO Troll Board 9513 3199 862 3 SA 1964-69 CSIRO Pole Board 7328 929 300 4 SA 1977 CSIRO Pole Board 908 164 54 5 WA 1963-67 CSIRO Pole Board 12826 264 264 6 WA 1970-78 WA Dep. Fisheries Pole Board 5692 236 120 7 WA 1983 CSIRO Pole Cradle^ 6907 1853 472 8 SA 1984 CSIRO veries for which an accurate recapture date was not Pole available. Cradle 3211 1117 176 'Excludes re CO -Fish placed or a measuring board for tagging. 'Fish placed in a specially designed cradle for tagging. applications, it is necessary that the data be analyzed as a series of separate release sets rather than as a pooled data set, and corresponding shedding-rate esti- mates are thus required; (2) there are now reliable recovery data for 10,416 double-tagged southern blue- fin available for analysis (compared with the 1511 re- coveries considered by Kirkwood 1981); and (3) models are now available that utilize exact periods at liberty rather than data aggregated by time period. Provided accurate recapture times are available (as is the case here), these new models are preferable because of their greater ability to deal with low recovery numbers towards the end of an experiment. Tagging data and tagging methods The tagging data used in this analysis consist of all records received by 31 March 1987 of southern bluefin that were originally double-tagged and that had accu- rate recapture dates. Recovery dates were considered accurate if at least the month of recapture was known with certainty. In total, there were 10,416 recoveries that met this criterion. Of these, there were 671 recov- eries for which the month of recapture was known but the exact day was uncertain; however in almost all of these cases, the uncertainty was estimated to be no more than +5 days. The data were grouped into eight double-tagging experiments: (1) NSW releases 1963-70 (CSIRO*); 'Tagged by staff of the Commonwealth Scientific and Industrial Research Organization (CSIRO). (2) NSW releases 1963-70 (fishermen contracted by CSIRO); (3) SA releases 1964-69 (CSIRO); (4) SA releases 1977 (CSIRO); (5) WA releases 1963-67 (CSIRO); (6) WA releases 1970-78 (WA Dep. Fish.); (7) WA 1983 (CSIRO); and (8) SA 1984 (CSIRO). This classification of the data was made to ensure that, within experiments, the geographical area, fish size, tagging and fishing methods, and tagging personnel were as similar as possible. Except for experiment 2, the primary method used to catch fish for tagging was commercial bait and pole; in experiment 2, trolling was used. In all experiments, fork lengths of the fish selected for tagging were mea- sured before tagging. For the first six experiments, this was done by placing the fish on a measuring board. In experiments 7 and 8, the fish were placed in a specially designed vinyl cradle supported by a metal frame and their fork lengths measured using gr-aduations marked on the cradle. While the fish were restrained on the measuring board or in the cradle, two numbered tuna tags of a standard type (Williams 1982) were inserted forward into the musculature at an angle of about 45°, 1-5 cm below either side of the posterior insertion of the second dorsal fin. Ideally, the tag barb anchored behind the second dorsal fin ray extensions or the neural spines, and in experiments 7 and 8 greater effort was made to achieve this than in earlier experiments. A summary of the fishing method, tagging method, and the numbers of tuna released and recovered is given for each experiment in Table 1. For ease of pre- sentation, the numbers of recoveries in each experi- ment are grouped by period at liberty in Table 2; as noted above, exact times at liberty for each recovery were used in the subsequent analyses. Hampton and Kirkwood: Tag shedding by Thunnus maccoyii 315 Table 2 Numbers of double-tagged southern bluefin tuna, for which accurate recapture dates are known, that were reported to have retained | both tags (D) or only one tag (S; on recapture, listed by period at liberty. Experiment number 1 2 3 4 5 6 7 8 Period at liberty (yr) D S D S D S D S D S D S D S D S 0-1 175 61 2638 467 831 158 128 23 166 84 205 73 1428 314 644 64 1-2 70 61 517 329 51 76 30 15 72 108 23 31 306 102 381 75 2-3 6 21 37 52 15 19 3 11 18 27 2 7 lO: 40 82 35 3-4 1 1 4 5 11 9 2 4 — 8 3 4 18 16 10 2 4-5 1 1 1 1 10 11 — — 3 15 1 9 5-6 — — — 1 • ) 10 — — 3 9 _ 2 6-7 — 2 1 2 2 7 — — 1 4 2 1 7-8 — 2 — 1 9 1 — _ _ _ _ — 8-9 — 1 — 1 2 6 1 1 — 4 — _ 9-10 — 1 — 1 2 _ _ — — 1 — — 10-11 — — 1 1 — 1 — _ — — 11-12 — — — — 1 1 — 1 _ — 12-13 — — — — — — — — — — 13-14 — — — — — — — 2 — — 14-15 — 2 — — — — — _ _ _ 15-16 — — — — _ _ — 1 — 16-17 — — — 1 — 1 — _ — — 17-18 — — — — — — 1 — Tag-shedding models Various models have been proposed to describe the pro- cess of tag loss (Beverton and Holt 1957 and reviews by Ricker 1975; Wetherall 1982). Tag losses can be classified as type I losses, which effectively reduce the number of tags released (immediate tag shedding, im- mediate tagging mortality, and non-reporting), and type II losses which occur steadily over time (natural mortality, fishing mortality, permanent emigration, and long-term tag shedding). Considering the shedding process only, the simplest model is that proposed by Beverton and Holt (1957) for type II shedding, where the instantaneous rate of shedding (L) is constant over time. In this model, if for a fish originally single-tagged, Q{t) is the probability of the tag being retained at time t after release, then Qit) = e-^' (1) If immediate tag shedding occurs, with a probability 1-P of this occurring, then Q{t) = p e-^' (2) In some cases, it may be inappropriate to assume a constant shedding rate. Some tags may deteriorate over time, causing their shedding rate to increase (Baglin et al. 1980). Alternatively, some tags may be come more securely fixed over time, e.g., through the gradual laying down of tissue around the tag, causing the shedding rate to decrease (Kirkwood 1981). More generally, Kirkwood (1981) has noted that if individual tags have a different propensity for shedding, then the apparent average shedding rate will decrease with time at liberty, as the tags with the higher shedding rate are lost first. Kirkwood modeled this process by allow- ing the shedding rate to be a gamma-distributed ran- dom variable, rather than a constant; however, his model did not allow for an increasing shedding rate. An alternative approach was adopted by Wetherall (1982), who assumed that the probability of a tag be- ing shed was time-dependent. He proposed a flexible function that allowed either an increasing or decreas- ing shedding rate. In this paper, we adopt a time-dependent shedding- rate model that allows the probability of an individual tag being shed to decrease over time in an identical fashion to the average shedding rate of the Kirkwood (1981) model. In this case, we now assume that the rate of shedding at time t after release follows the functional form Lit ) = - bk b + U 316 Fishery Bulletin 88(2), 1990 It then follows that Q{t) Note that as 6-*°°, - f L{u)du b + kt (3) >e-*'andL^A. Thatis, b + Xt the variable-rate model reverts to the constant-rate model. If a proportion 1-p of tags are shed immediate- ly, then Q(0 = p b + U (4) Parameter estimation Because only tag returns with accurate recapture dates are considered, an extension of the maximum likelihood estimation procedure used by Kirkwood and Walker (1984) can be used. Suppose fish are double tagged and all tags not immediately shed have identical shedding probabilities that are independent of their companion tags' status (already shed or still retained). Then, if Pzii ). Pi (t ). ^nd poit ) are the probabilities of a fish re- taining two, one, and zero tags, respectively, at time t(Q ■o SI CO 06- y 0) 0.6- . f I >. S 04- f\ ~ 0,4- CO 1 n n 1 (0 o 1 J3 qI 0 2 - i o 02- iJ r yCX^ Models r 0 0 -1 — ' — ' — ' — ' — 1 — ' — ' — ' — ■ — 1 — • — • — '■ ■ ' 1 — ■ — ■ — ■ — • — 1 OOT''''r -^.,.. 1 0 5 10 15 20 0 5 10 15 20 Period at liberty (yr) Period at liberty (yr) (f) Experiment 6 (h) Experiment 8 1.0- LO- 08- y ^ Model 2 c GS- T3 X T3 ■D / ■D 06- / r 0) SL 06- w f ifi o / o >. > ' >> 04- / -.^- 04- S f d nj 1 to JD r .Q O / O i 02- ( qI 02- ^ ' Model 2 0 0- 1 .... 1 ... . T -T 1 1 1 0.0- . . r f 1 ■ . 1 ■ 1 . . . . 1 ■ -^ ' ' 1 0 5 10 15 20 0 5 10 15 20 Period at liberty (yr) Period at liberty (yr) Figure 1 (continued) Table 4 Maximum likelihood estimates of tag-shedding parameters for southern bluefin tuna double-tagging | experiments using grouped data. Parameters Experiment Selected no. model A±SE /3+SE P±SE 1 2 0.26±0.05 OO 0.98 + 0.05 2 1 0.18 + 0.01 oo 1.0 3 3 0.27 + 0.07 0.55-^0.37 1.0 4 1 0.20 + 0.04 c» 0.96 + 0.01 5 3 0.56±0.14 0.66 ±0.32 1.0 6 3 0.43 + 0.09 0.53-1-0.27 1.0 7 2 0.053±0.010 oo 0.93 + 0.01 8 2 0.055 + 0.011 oo 0.98±0.01 320 Fishery Bulletin 88(2|. 1990 result in a higher rate of shedding. Tags used in the latest experiments do not appear to have suffered from such defects. In experiments 7 and 8, the use of a spe- cially designed tagging cradle may have resulted in more effective and less traumatic tagging than in earlier experiments, where fish were tagged on a mea- suring board. Also in experiments 7 and 8, a deliberate effort was made to anchor the tags behind the second dorsal fin ray extensions. Tags successfully attached in this way should have a very small probabDity of shed- ding. Unfortunately, little quantitative information is available on the efficacy of the technique of tag attach- ment adopted for experiments 1-6, but it is known that greater care was taken in tag attachment in the later experiments. If the proportion of ideally attached tags in experiments 1-6 differed markedly from those for experiments 7 and 8, then commensurate differences in shedding-rate estimates would result. While the lower tag shedding rates in experiments 7 and 8 are obvious on first inspection, the situation is less clear for the earlier experiments. In general terms, the precision of estimates of shedding rates, par- ticularly long-term rates, will increase with the size of the database. On these grounds, pooling of the data for experiments 1-6 would appear to be an attractive op- tion. However, it would be appropriate to do so only if the assumptions underlying the analysis of the pooled data remained valid. Critical among these assumptions is that all tags have identical and independent shed- ding probabilities. The primary reason for classifying the tag releases in the 1960s and 1970s as six separate experiments was a suspicion that these probabilities may well have been heterogeneous. As mentioned earlier, with this classification we were attempting to minimize within-experiment differences in geograph- ical area, fish size, capture and tagging methods, and tagging personnel. While in principle further subdivi- sion is possible, for example by tagging vessel in order to take account of different tagging teams, we felt this may leave too few data in each subset to obtain reliable estimates of shedding rates. Also, no assertion is made that the classification chosen is an optimal one; indeed it is by no means obvious that suitable criteria for op- timality could be defined. A statistical way of examin- ing whether or not it is appropriate to pool data for experiments 1-6 is to carry out a likelihood ratio test of the hypotheses that the tag shedding parameters for each experiment are equal. This hypothesis proved to be resoundingly rejected (P<0.001). It appears that pooling is not a viable option for these experiments. One of the subsequent analyses that we had envis- aged for recovery data adjusted for tag shedding in- volved application of the method of Hearn et al. (1987) to obtain estimates of natural and fishing mortality rates. This method, which is essentially a cohort anal- ysis of the tagging data, gives particular weight to long- term recaptures and therefore requires accurate esti- mates of long-term shedding rates. Despite the large numbers of recoveries in the experiments described in this paper, precise estimation of long-term shedding rates has not been possible. This is best exemplified in the results of experiment 1, where no single best fit- ting model was obtained, and for which the point esti- mates of long-term shedding rates differ markedly. Even where a single best-fitting model was available, considerable uncertainty still remained about the long- term shedding rates. It therefore seems essential to take account of this uncertainty in any subsequent anal- yses using methods such as that of Hearn et al. (1987). This is the subject of further research. Acknowledgments Dr. G. Eckert, Dr. V. Mawson, and Ms. S. Wayte reviewed and provided helpful comments on an earlier draft of this manuscript. We also thank Dr. W.S. Hearn for useful comments and discussion, particularly in rela- tion to the assumption of independent and identical shedding rates. Citations Baglin, K.E.. Jr.. M.I. Farber, W.H. Lenarz, and J.M. Mason, Jr. 1980 Siieddiiig rates of [ilastic and metal dart tags from Atlan- tic bluefin tuna. Thunniu^ thywius. Fish. Bull., U.S. 78: 179-185. Bard. Y. 1974 Nonlinear parameter estimation. Academic Press, NY, 341 p. Bayliff, W.H., and L.M. Mobrand 1972 Estimates of the rates of shedding of dart tags from yellowfin tuna. Inter-Am. Trop. Tuna Comm. Bull. 1.5:465- .503 (Engl, and Span.]. Beverton, R.H.J.. and S.J. Holt 1957 (_)n the dynamics of exploited fish populations. Fish. In- vest. Ser. II, Mar. Fish. G.B. Minist. Agric. Fish. Food 19, 533 p. Hearn, W.S.. R.L. Sandland, and J. Hampton 1987 Robust estimation of the natural mortality rate in a com- pleted tagging experiment with variable fishing intensity. J. Cons. Int. Explor. Mer 43:1(17-117. Hynd, J.S. 1969 New evidence on southern bluefin stocks and migra- tions. Aust. Fish. 28(5):26-30. Hynd, J.S., B.B. Jones, G.L. Kesteven, and N. Sproston 1967 Tunas. /« CSIRO Division of Fisheries and Oceanog- raphy Annual Report 1966-67, p. 20-24. CSIRO Div. Fish. Oceanogr., Cronulla, Australia. Kendall, M.G., and A. Stuart 1961 The advanced theory of statistics. Vol. 2: Inference and relationship. Charles Griffin, London, 676 p. Hampton and Kirkwood: Tag shedding by Thunnus msccoyii 321 Kirkwood, G.P. 1981 Generalized models of the estimation of rates of tag shed- ding by southern bluefin tuna (Thunnus maccoyii). J. Cons. Int. Explor. Mer 39:256-260. Kirkwood, G.P., and M.H. Walker 1984 A new method for estimating tag shedding rates, with application to data for Australian salmon, Arripis trutta esper Whitely. Aust. J. Mar. Freshwater Res. 35:601-606. Laurs, R.M., W.H. Lenarz, and R.N. Nishimoto 1976 Estimates of rates of tag shedding by North Pacific albacore, Thunnu.-i alalunga. Fish. Bull., U.S. 74:675-678. Lenarz. W.H., F.J. Mather III, J.S. Beckett, A.C. Jones, and J.M. Mason Jr. 1973 Estimation of rates of tag shedding by northwest Atlantic bluefin tuna. Fish. Bull., U.S. 71:1103-1105. Lewis, A.D. 1981 Population genetics, ecology and systematics of Indo- Australian scombrid fishes, with particular reference to skip- jack tuna (Katsuwonus pelamis). Ph.D. thesis, Australian Natl. Univ., Canberra, Australia. Majkowski, J., and G.L Murphy 1983 CSIRO southern bluefin tuna tagging: scientific achieve- ments and future objectives. Aust. Fish. 42(ll):26-27. Majkowski, J., W.S. Hearn, and R. L. Sandland 1984 An experimental determination of the effect of increas- ing the minimum age (or size) of fish at capture upon the yield per recruit. Can. J. Fish. Aquat. Sci. 41:736-743. Ricker, W. E. 1975 Computation and interpretation of biological statistics of fish populations. Bull. Fish. Res. Board Can. 191, 382 p. Wetherall. J. A. 1982 Analysis of double-tagging experiments. Fish. Bull., U.S. 80:687-701. Williams, K. 1982 Tagging method. In Majkowski, J. (ed.), CSIRO data base for southern bluefin tuna [Thunnus maccoyii (Castlenau)), p. 7-8. Aust. CSIRO Mar. Lab. Rep. 142. Abstract.— Seasonal advection of warm western-boundary current water into the southern Benguela eastern-boundary current region from late spring to early autumn coincides with increased coastal upwelling ac- tivity. The combined effect of these two processes is to stabilize the sys- tem through intensification of tem- perature fronts and thermoclines, pro- viding conditions favoring anchovy spawning and larval survival. These conditions appear to be robust to short periods of intense storm mixing and downwelling during summer, and weaken and disappear only with the onset of winter when there is dimin- ished influence of the warm western- boundary current water, downweUing conditions, and increased frequency of storms. Ocean Stability and Anchovy Spawning in the Southern Benguela Current Region Peter A. Shelton Science Branch, Department of Fisheries and Oceans PO Box 5667, St John's. Newfoundland AlC 5X1, Canada Larry Mulchings Sea Fisheries Research Institute Private Bag X2, Rogge Bay 8012, Cape Town, South Africa Manuscript accepted 24 November 1989. Fishery Bulletin. U.S. 88:323-338. The extension of the Hjort (1914) hypothesis by Lasker and his cowork- ers (Lasker 1975, 1978, 1981a, b) to include the role of a stable ocean in determining fish recruitment has had a profound influence on research in- to the early life history of marine fish. Lasker's extended hypothesis suggests that ocean stability influ- ences food aggregations for larvae and larval drift and, hence, recruit- ment strength (Lasker 1981b). In sup- port of Lasker's hypothesis, strong wind events have been found to dis- rupt anchovy larval food aggrega- tions, and, by inference, to effect lar- val mortality off California (Lasker 1975) and Peru (Walsh et al. 1980). During such events, larvae can also be entrained within increased off- shore Ekman flow and placed outside suitable feeding areas (Walsh et al. 1980, Lasker 1981b). A significant linear relationship between anchovy larval mortality rate and the frequen- cy of calm, low wind-speed periods during the spawning season off Cali- fornia (Peterman and Bradford 1987) provides further evidence of the im- portance of event-scale disruption of ocean stability. Parrish et al. (1983) compared the seasonality and geography of sardine and anchovy reproduction in the Cali- fornia, Peru, Canary, and Benguela Current systems with corresponding features of the environment and sug- gested that patterns of correspon- dence may indicate the most crucial processes affecting reproductive suc- cess. They found that spawning was adapted to avoid times and places char- acterized by intense ttirbulent mixing and strong offshore transport. In this paper we use survey data to provide a finer-scaled description of ocean stability and anchovy spawn- ing than that of Parrish et al. (1983), and suggest that in the southern Ben- guela Current system anchovy spawn- ing is adapted to seasonal patterns of ocean stability that are resistant to event-scale processes. Local condi- tions are briefly compared with those in the Southern Californian Bight where Lasker's studies took place, and conclusions are drawn with re- spect to the influence of ocean sta- bility on survival of the planktonic stages and recruitment in the south- ern Benguela Current region. Methods Diu-ing the Cape Egg and Larval Pro- gramme (CELP), plankton and the environment were sampled at approx- imately monthly intervals between August 1977 and August 1978 from a grid of stations in the southern Benguela Current region. The pur- pose of the program was to deter- mine general spawning habitat and time for the major fish species of com- mercial importance (Shelton 1986). The grid comprised 120 stations posi- 323 324 Fishery Bulletin 88(2|. 1990 Figure 1 Cape Egg and Larval Programme (CELP) survey grid, southern Benguela Current region, sampled monthly August 1977-August 1978, and general topography of the survey area (depth contours in meters). tioned at 10-nautical-mile intervals (18.5 km) on 20 transects (Fig. 1). Vertical profiles of temperature and water samples for measurement of salinity, chlorophyll-a concentra- tion, and microplankton density were collected from the water column. Chlorophyll analysis, described in Shannon et al. (1984), was by means the standard photometric method of SCOR/UNESCO Working Group 17 (1966). Microplankton particle concentration was determined by prefiltering 2 liters of the water sample through a 100-/im filter and collecting those particles retained by a 37-;.(m filter. Particles between 37 and 100 ^m correspond to the size fraction consid- ered to be important for first-feeding anchovy larvae (Arthur 1976, Hunter 1977). Particles were counted under a light microscope at 20 x magnification and ex- pressed as number per liter. Although all particles were enumerated, only the maximum density at each station of total "esculent" (nutritious to fish larvae [Sharp 1980]) particles are presented, such as copepod eggs, nauplii, copepodite stages, and dinoflagellates. Wind strength and direction were estimated by the light- house keeper at Cape Point throughout the period. In order to examine surface drift patterns, 20 plastic drift cards (Dimcan 1965) were released monthly at each sta- tion sampled. Fish eggs and larvae were sampled by means of a double oblique haul of a Bongo sampler down to a max- imum depth of 100 m. The sampler had a mouth open- ing of 57 cm and was fitted with 300-^m and 500-/im mesh nets. Counts of anchovy eggs and larvae from the 300-f4m unit were standardized to volume or num- bers under 10 m^ using the formula given in Smith and Richardson (1977). In November 1979 and Novem- ber 1984 the vertical distribution of anchovy eggs and temperature profiles was measured at two localities. Eggs were sampled with Miller nets (Miller 1961) as described in Shelton and Hutchings (1982). Anchovy egg and early-stage larval abundance data from a more extensive grid of stations sampled in November 1983 using a CalVET net (Smith et al. 1985) are also presented. Shelton and Hutchings: Ocean stability and anchovy spawning in southern Benguela Current region 325 Results and discussion Seasonality in temperature and spawning Both the mean and the coefficient of variation (CV) of sea surface temperature measured over the CELP giid peaked in summer (Fig. 2). This is caused by the sea- sonal advection of warm water orginating from the western-boundary Agulhas Current into the southern Benguela Current region (Shelton et al. 1985) and the simultaneous increase in coastal upwelling (Andrews and Hutchings 1980). Strong cross-shelf temperature gradients, or fronts, are set up between advected and upwelled water, reflected by the increase in the sur- face temperature CV. In winter the influence of west- ern-boundary current water is diminished and upwell- ing is less frequent and less intense (Shelton et al. 1985, Andrews and Hutchings 1980). Consequently, cross- shelf temperature gradients weaken and disappear and mean SST values are lower and have a lower CV, re- flecting cool, more isothermal conditions. Horizontal structure The changes in temperature structure described above, as well as associated changes in the patterns of abun- dance of microplankton and chlorophyll-a, can be seen in contour diagrams of values for summer and winter (Fig. 3). The strong temperature front that develops along the west coast in summer is clearly visible. Al- though this front is most persistent between Cape Point and Cape Columbine, and was considered by Bang (1973) to "pivot" at Cape Point, it may extend as far south as Cape Agulhas when upwelling occurs at capes east of Cape Point. At Cape Columbine isotherms tend to diverge offshore, and further north the front is generally weaker and more variable. The density of potential food particles for early-stage larval feeding, as indicated by microplankton and chloro- phyll-a concentrations, are highest inshore of the tem- perature front in summer, indicating that the front constrains the offshore Ekman drift of productive upwelled water. In winter when the front disappears, microplankton and chlorophyll-a concentrations are generally lower and more dispersed. Arrows indicating major directions of surface drift over the survey area in summer (January 1978) and winter (August 1977) based on drift-card recoveries are plotted in Figure 4. Of the 24,000 cards released in each month there was a 5% recovery in January 1978 and a 7% recovery in August 1977. There was no clear seasonality in the overall recovery rate. In summer the southern Agulhas Bank area was characterized by on- shore surface flow with a westerly component. Some cards released over the western Agulhas Bank moved XJ -20 18 u ^ 16 to < LiJ 14 40 SST 0^ CV Z ^ny < _ q; < > _ VO u. o t— 1 z ^^ LU » 10 ^ II u^ ■^-n UJ C) r. U ASONDJ FMAMJ MONTH J A Figure 2 Monthly means and coefficents of variation (CV) of sea-surface temperature (SST) values over the CELP survey grid, southern Benguela Current region, for the period August 1977-August 1978. around Cape Point and up the west coast, and the gen- eral pattern of flow inshore of the strong temperature front on the west coast was northwards. North of Cape Columbine there was some evidence of onshore flow. There were very few southern African recoveries off- shore of the front, although several cards were recov- ered along the eastern coast of South America, indi- cating general offshore flow into the South Atlantic gyre at stations offshore of the front. In winter, flow in the south was again predominant- ly onshore, although this time with a marked easterly component east of Cape Agulhas and a westerly com- ponent to the west. Several cards released on the western Agulhas Bank were found substantial dis- tances up the west coast, although one card released in the extreme north of the grid travelled against this flow and was recovered on the east coast. Again, few cards were recovered on the coast of southern Africa from releases on the west coast, although there were 10 recoveries from the coast of South America and one recovery from southern California. In comparison with summer, several of these recoveries came from cards released at inshore stations, indicating widespread off- shore flow along the west coast in the absence of the front. Vertical structure Summer and winter vertical structure of temperature, microplankton density, and chlorophyll-a concentration are shown for three representative sections over the survey area in Figures 5-7. A section over the Agulhas Bank (Fig. 5) shows the existence of a strong thermo- cline in summer between warm water of Agulhas 326 Fishery Bulletin 88(2). 1990 SUMMER WINTER mbert's Bay SEA SURFACE TEMPERATURE (°C) I I I I I I I I I I I I I I I I I I I I I \7° 18° 19° 20° 21° E 17° 18° 19° 20° 21° E MICROPLANKTON (particles-f') ' I I '■ I I I I I I I I I I I i ■ ■iTTiN J I 1 I I I I I I I I I I I I I I I I I I I I I I I I 17° 18° 19° 20° 21° E 17° 18° 19° 20° 21° E CHLOROPHYLL o CONCENTRATION (mgm 18° 19° 20° 21° 22° E Figure 3 Typical summer (January 1978) and winter (August 1977) patterns of sea-surface temperature, microplankton particle concentrations, and chloropliyll-a concentrations over the CELP survey grid, southern Benguela Cur- rent region. Shelton and Hutchings: Ocean stability and anchovy spawning in southern Benguela Current region 327 3? 33' 34' 35' SUMMER SURFACE FLOW 32 - 33' 34' 35' Figure 4 Summer and winter patterns of surface flow in tlie southern Benguela Current region, gauged from drift-card recoveries from releases of 20 cards at each station in January 1978 and August 1977. Arrows may represent more than one recovery and indicate general direction of flow. 328 Fishery Bulletin 88(2), 1990 50 100 50 X I— a. LU Q 100 50 100 SUMMER WINTER DISTANCE OFFSHORE (n. miles) 20 10 50 40 30 1 1 1 20 — 1 — 10 ISOTHERMAL 15 °C Temperature ( C) Microplankton (particles litre } Chlorophyll a (mg-m-') Figure 5 Summer (January 1978) and winter (August 1977) sections of temperature, microplankton, and chlorophyll-a values along line 7G over ttie Agulhas Bank, southern Benguela Current region. Current origin forming an upper mixed layer and cold water that moves onto the Bank along the bottom in summer. The warm layer had very low plankton den- sities, but enhanced microplankton density and phyto- plankton-a concentration were associated with the thermocline. With the onset of winter, the influence of the warm Agulhas Current water over the Agulhas Bank be- comes less and heat loss to the atmosphere increases, causing a cooling of the surface layer. Surface cooling coinbined with the movement off the shelf of the cold bottom layer makes the water column less stable and susceptible to mixing by winter storms (Shannon et al. Shelton and Hutchings: Ocean stability and anchovy spawning in southern Benguela Current region 329 50 100 50 X 100 50 — r- 40 SUMMER WINTER DISTANCE OFFSHORE (n. miles) 30 20 10 50 40 30 20 r 50 100 Figure 6 Summer (January 1978) and winter (August 1977) sections of temperature, microplankton. and chlorophyll-o values along line 44 just north of Cape Town, South Africa. 1984, Largier and Swart 1987). Microplankton parti- cle concentrations tend to be low in the well-mixed water column but chlorophyll-a concentrations can be quite high, either if nutrients accumulated in bottom sediments are mixed into the euphotic zone by strong storm action (Shannon et al. 1984), or if periods of calm weather persist long enough for the water column to stabilize, albeit weakly, and phytoplankton is thereby allowed to remain in the euphotic zone long enough to grow. A section just north of Cape Town (Fig. 6) shows the strong temperature front in summer between cold upwelled water inshore and warm water offshore. En- hanced levels of microplankton occurred in the vicinity 330 Fishery Bulletin 88(2), 1990 SUMMER WINTER DISTANCE OFFSHORE (n. miles) 10 50 40 30 20 n Figure 7 Summer (January 1978) and winter (August 1977) sections of temperature, microplankton, anil elilorophyll-u values along line 16 in the vicinity of Lambert's Bav. South Africa. of the front as well as close inshore, and high levels of chlorophyll-a were found inshore near the surface, declining in concentration and following the isotherms into deeper water in the vicinity of the front. Offshore of the front, both chlorophyll-*/ and microplankton levels were negligible. A similar pattern was noted in a transect of the front off Cape Columbine in December 1984, with copepod nauplii and copepodites concen- trated in the front and chlorophyll-*; values increasing towards the coast, inshore of the front (Armstrong et al. 1987). Although the front may persist as a subsur- face feature in winter, even during prevalent downwell- ing conditions, it is much weaker as a consequence of cooler water offshore. Microplankton and chlorophyll-a Shelton and Hutchings Ocean stability and anchovy spawning in southern Benguela Current region 331 01 I 02 I 03 I 04 I 05 I 06 I 07 I 08 I 09 I 10 , 11 . '2 , AUGUST 1977 = 10 msec" ^ vfe^ I 13 I 14 I 15 I 16 I 17 I 18 I 19 I 20 I 21 I 22 I 23 I 24 I l^m.. 23 , 24 I 25 I 26 I 27 I 28 29 I 30 31 ^ 01 I 02 I 03 I DECEMBER 1977 JANUARY 1978 ^i^^MMi^^hglkM teSi^^M,^^ I 04 I 05 1 06 I 07 I 08 I 09 I 10 I 11 I 12 1 1 = 10 m sec 13 I 14 I 15 I JANUARY 1978 Ffgure 8 Wind stick diagram for the period 23 December 1977-15 January 1978 and 1-24 August 1977, showing tyical summer and winter patterns of wind stress measured at the lighthouse at Cape Point, South Africa. values are generally much lower and more dispersed over the west coast between Cape Point and Cape Col- umbine in the absence of the surface front in winter. A section on the west coast off Lambert's Bay (Fig. 7) shows a moderately strong thermocline with coastal upwelling in simimer. Moderate-to-high levels of micro- plankton and chlorophyll-a extending to the limit of the grid in the top 50 m are t>'])ical of the area at this time. In winter the thermal structure is similar, but wath less upwelling activity and cooler water offshore. Although microplankton levels were low in the section shown, widespread moderate levels of chlorophyll-o are quite 332 Fishery Bulletin 88(2). 1990 Figure 9 Variability in the vertical temperature struc- ture at station 64-06 on the Agulhas Bank, southern Benguela Current region. S 32' 34' Position of the surface thermal front, October 1977 to ,0 I. May 1978 16° 18- 20 22° E Figure 10 Variability in the position of the surface temperature front in the southern Benguela Current region, spring 1977 through autumn 1978. (Numbers refer to months of the year, e.g., 10 = October 1977, 2 = February 1978.) typical for the recruitment area in winter, as shown in Figure 3. Robustness to event-scale processes Wind stick diagrams of wind strength and direction estimated by the lighthouse keeper at Cape Point (Fig. 8) show that winter is characterized by frequent periods of strong northerly wind interspersed with southerly winds and calm, whereas in summer strong upwelling- favorable southeasterlies dominate, alternating with brief periods of calm and occasional wind reversals. A SONDJ FMAMJ MONTH J A Figure 1 1 Monthly mean abundance of anchovy eggs and larvae over the CELP survey grid, southern Benguela region, for the period August 1977-August 1978. Despite event-scale forcing of this nature, contoured monthly temperature profiles through the water col- umn over the Agulhas Bank show strong stratification persisting over several months of the spring-to-autumn period, weakening and disappearing with the onset of winter (Fig. 9). Similarly, although erratic in appear- ance and position to the north and south, the tempera- ture front in the transport area between Cape Point and Cape Columbine persists as a stable feature for up to 8 months of the year (Fig. 10), despite event-scale variability in the wind field. During lulls in upwelling the front may weaken at the surface and move inshore, at which time the temperature discontinuity may be most apparent in the form of a thermocline, but it per- sists as a subsurface feature further offshore (Shelton Shelton and Hutchings: Ocean stability and anchovy spawning in southern Benguela Current region 333 Figure 12 Anchovy egg and larval abundance patterns over the CELP survey grid, southern Benguela Current region, in January 1978, a month of peak spawning. 1986). With the next pulse of upwelhng, the isotherms flip to a more vertical orientation and the front moves offshore again, to an average position over the inner shelf break (200-m contour). Anchovy spawning, transport, and recruitment in relation to ocean stability Anchovy eggs are generally absent or in very low abun- dance in the plankton in autumn and winter when the region is cool and isothermal (March- August, Fig. 11). With the onset of spring (September-November) both egg and larval abundance increases rapidly, reaching a peak in late spring or early summer when both the mean temperature and the CV are high, and then decreasing to low numbers over the autumn. Anchovy spawning is typically most intense south and east of Cape Point over the broad Agulhas Bank (Fig. 12). The vertical distribution of anchovy eggs in the spawning area on the Agulhas Bank (Fig. 13a) is restricted almost entirely to the warm upper mixed layer, above the strong thermocline that forms in sum- mer at about 50 m. On the west coast a transect bisect- ing the front in the vicinity of the Cape Peninsula shows that anchovy eggs extending up the west coast are con- centrated within the frontal zone (Fig. 13b). Anchovy larvae (Fig. 12) are generally more widespread over the Agulhas Bank than the eggs as a result of disper- sal, and in particular extend further up the west coast, often as far as Cape Columbine, in the vicinity of the front. Recent samples from a more extensive grid of stations confirm these patterns (Fig. 14). Based on these patterns of anchovy eggs and larval abundance and the catch pattern of recruits in the com- mercial fishery (Crawford et al. 1980 and Crawford 1980), the habitat of the anchovy in the southern Benguela region can be divided into spawning, trans- port, and recruitment areas (Fig. 15). The role of ocean stability in each of these three areas can be evaluated. Anchovy spawning is highly seasonal, peaking in summer when the southern Benguela region is most structured by temperature fronts and thermoclines under the simultaneous influence of advected warm western-boundary current water and coastal upwell- ing. The spawning area is characterized by a strongly stratified, stable water column throughout summer. Stratification is resistant to event-scale mixing pro- cesses and only weakens with the reduced influence of eastern-boundary current water on the Bank and the retraction of the cold bottom layer with the onset of winter. Winter storms are then able to mix the entire water column creating isothermal conditions which per- sist until the onset of summer. The warm, uniform sur- face layer over the Agulhas Bank in summer provides conditions conducive to rapid anchovy egg develop- ment; below 14 °C anchovy egg development in the 334 Fishery Bulletin 88(2). 1990 (a) 200 100 (n 100m-3) 100 200 200 100 100 200 10 20 I '° J 40- O- LU ^ 50- 60- 70- 10 16 20 10 TEMPERATURE (°C) DISTANCE (n miles) 20 80 60 Egg density (n 100 m 7 n 5-50 □ 50-200 ^M >200 20 Figure 13 (a) Vertical distribution of eggs and temperature at depth for two stations in the anchovy spawning area, southern Benguela Current region, November 1984. (b) Vertical sections of temperature and anchovy egg abundance measured along a transect bisecting the temperature front off the Cape Peninsula, South Africa, November 1979. Figure 14 Pattern of abundance of anchovy eggs and larvae in November 198.'-! from an extensive survey grid covering nearly the entire anchovy spawning area in the southern Benguela Current region. Shelton and Hutchings Ocean stability and anchovy spawning in southern Benguela Current region 335 336 Fishery Bulletin 88(2|. 1990 Figure 15 A generalized representation of anchovy spawning, transport, and recruitment areas in the southern Benguela Current region. laboratory is very slow and hatching larvae are de- formed (King et al. 1978). Below the warm surface layer a strong thermocline at about 50 m is associated with subsurface maxima of microplankton and chloro- phyll-fl, facilitating the survival of early-stage larvae that require high concentrations of food particles within a suitable size range (Lasker et. al. 1970). Spawning conditions in the southern Benguela ap- pear to be in marked contrast with conditions in the spawning ground of the Northern anchovy EngrauUs mordax in the Southern California Bight where Lasker's studies were carried out. Surface tempera- tures in the Southern California Bight given in Lasker (1981b) are some 3-5°C lower than over the Agulhas Bank during the respective spawning seasons. Vertical temperature structure during the spawning season off southern California, as shown in Lasker (1978), is quite variable, with generally weak stratification susceptible to mixing. The chlorophyll-a layer disrupted by storm mixing in April 1984 was at about 15 m (Lasker 1975), in contrast to the deeper maxima on the Agulhas Bank. Although maximum chlorophyll-a values before storm mixing in the layer off California in April 1974 are similar to those recorded in the Agulhas Bank layer, microplankton particle concentrations over the Agulhas Bank do not reach the high values recorded by Lasker (1981b) because dinoflagellates are not as common as in the Southern California Bight, possibly due to per- sistent strong winds (Parrish et al. 1983), which thoroughly mix the relatively deep, upper mixed layer. From the the spawning area in the south, the drift of eggs and early-stage larvae is primarily northwest- wards, into the west coast transport area. Bang (1973) and Bang and Andrews (1974) have documented the presence of a strong northward-flowing jet current associated with the front which is a prominent feature of this area in summer. The role of this jet in transport- ing anchovy eggs and early-stage larvae from the Agulhas Bank spawning ground to the west coast nursery area has been examined and described in Shelton and Hutchings (1982). Drift-card recoveries for summer support their interpretation. In the transport area, enhanced levels of chlorophyll-a and microplank- ton associated with the inshore and frontal zone may be important in early-larval survival. Despite predomi- nantly southeast winds in summer, facilitating offshore Ekman flow and upwelling at the coast, the offshore loss of eggs and larvae along the west coast may be small because of the constraining influence of the front. However, there is frequently an offshore divergence of the front off Cape Columbine, following the orien- tation of bottom contours, and this may be associated with the "leaking" of water containing larvae from the productive coastal region (Shelton et al. 1985, Shelton 1986). Onshore flow north of Cape Columbine facilitates penetration of the productive recruitment area by larvae. Positioned downstream of the major sites of upwelling at Cape Columbine and Cape Point, the fron- tal structure dominating the transport area in summer gives way to more gently sloping isotherms north of Cape Columbine where the shelf widens. Because of lower average wind speeds (Hutchings et al. 1988), upwelling is milder than in the transport area and a vertically stable water column is a persistent feature. Enhanced levels of microplankton and chlorophyll-a characterize the inner shelf in this area throughout the year (Hutchings 1981), providing suitable conditions for the survival and growth of anchovy recruits. The patterns of ocean stability described above, based largely on survey data, appear to have played a major role in shaping the spawning behavior observed by anchovy in the southern Benguela region. Upwell- ing generates the high levels of plankton production that support the spawning biomass of up to 2 million tons of anchovy estimated for the area (Armstrong et al. 1988), but the associated processes of turbulence and offshore Ekman flow present special problems to the plankton stages that have to be solved through evolutionary adaptation. In the southern Benguela region this appears to have been brought about by the selection of times and areas of spawning which make best use of seasonal and spatial patterns of ocean stability to provide suitable food concentrations and favorable transport. Survey data for the southern Benguela therefore tends to confirm the conclusions from the larger-scale Shelton and Hutchings, Ocean stability and anchovy spawning in southern Benguela Current region 337 comparative study of eastern-boundary currents by Parrish et al. (1983) that the spawning habits of an- chovy (and sardine) attempt to minimize both wind- induced turbulent mixing and offshore-directed trans- port. Bakun (1985) argues that the results of Parrish et al. (1983) strengthen the relevance of Lasker's hypothesis to modeling recruitment and exploratory data analysis. However, if anchovy populations in the different eastern-boundary current regions have essen- tially "solved" the problems of disruptive turbulence and unfavorable transport, "mistakes" should be rare and difficult to find in the field. The fact that event- scale disruptive turbulence has been observed in the spawning area off southern California and that the in- cidence of calm periods can be statistically related to daily larval mortality rate (Lasker 1975, 1981b; Peter- man and Bradford 1987) suggests that such "mistakes" do occur despite adaptations. Even so, they may be undetectable in recruitment measurements, either because of large and often unquantified measurement error, or because the "mistakes" are masked by a well- developed suite of "bet-hedging" traits (Shelton 1987). Estimates of anchovy recruitment (with error bars) for the southern Benguela population based on direct surveys (Butterworth 1989, Fig. 16) show less variabil- ity than anticipated in the literature (e.g., Beddington and Cooke 1983), although the time series is still too short to be conclusive. If the southern Benguela anchovy population is adapted to local persistent patterns of ocean stability and is well-buffered against event-scale variations through bet-hedging, a prerequisite for detectably poor recruitment at moderate spawning stock sizes may be season-long weakening or disruption of the stability patterns in either the spawning, transport, or recruit- ment areas. Such conditions would probably arise if the influence of Agulhas Current water over the Agulhas Bank and on the west coast were reduced, particular- ly over the summer period. Acknowledgments We acknowledge the input of several of our colleagues at the Sea Fisheries Research Institute, Cape Town, to the Cape Egg and Larval Programme. In particular. Garth Newman and Vere Shannon provided much of the impetus for the initiation of CELP, and Alan Robertson, Hennie Crous, Frans Kriel, Deon Horst- man, Donald Alexander, Robert Cooper, Gustav Weh- meyer, Jan Van der Westhuizen, Chris Jeenes, John Giddey, and Christine Illert gave excellent technical support during its execution. Colleagues at the NMFS Southwest Fisheries Center, La Jolla, particularly the group formally headed by Reuben Lasker, are thanked 600- z LU S 400- 200- 81 82 83 84 85 YEAR 86 87 88 Figure 16 Estimates (median and 95% confidence interval) of annual anchovy recruitment for the southern Benguela population, deduced from direct survey results to July 1988 (Butterworth 1989). for their advice and encouragement with respect to the studies of anchovy early-life history in the Benguela Current system. We dedicate this paper to the memory of Reuben. Citations Andrews, W.R.H., and L. Hutchings 1980 Upwelling in the southern Benguela Current. Prog. Oceanogr. 9:1-81. Armstrong, D.A., B.A. Mitchell-Innes, F. Verheye-Dua, H. Waldron, and L. Hutchings 1987 Physical and biological features across an upwelling front in the southern Benguela. In Payne. A.I.L, J. A. Gulland, and K.H. Brink (eds.), The Benguela and comparable ecosystems. S. Afr. J. Mar. Sci. .5:171-190. Armstrong, M., P. Shelton, I. Hampton, G. Jolly, and Y. Melo 1988 Egg production estimates of anchovy biomass in the southern Benguela system. Calif. Coop. Oceanic Fish. Invest. Rep. 29:137-157. Arthur. D.K. 1976 Food and feeding of larvae of three fishes occurring in the California Current, Sardinops sagax, Engraulis mordax and Tmchurus fiijinmetricus. Fish. Bull., U.S. 74:517-530. Bakun, A. 1985 Comparitive studies and the recruitment problem: Searching for generalizations. Calif. Coop. Oceanic Fish. Invest. Rep. 24:30-40. Bang, N.D. 1973 Characteristics of an intense ocean frontal system in the upwelling regime west of Cape Town. Tellus 25:256-265. Bang, N.D., and W.R.H. Andrews 1974 Direct current measurements of a shelf-edge frontal jet in the southern Benguela system. J. Mar. Res. 32:405-417. 338 Fishery Bulletin 88(2), 1990 Beddingrton, J.R., and J.G. Cooke 1983 The potential yield of fish stocks. FAO Fish. Tech. Pap. 242, 47 p. Butterworth. D.S. 1989 The Benguela Ecology Programme: Successful and appropriate? S. Afr. J. Sci. 85:633-643. Crawford, R.J.M. 1980 Seasonal patterns in South Africa's Western Cape purse seine fishery. J. Fish. Biol. 16:649-664. Crawford, R.J.M. , P.A. Shelton, and L. Hutchings 1980 Implications of availability, distribution and movements of pilchard {Sardinops ocellata) and anchovy (Engraulis capen- sis) for assessment and management of the South African purse-seine fishery. Rapp. P.-V. Reun. Cons. Int. E.xplor. Mer 177:355-373. Duncan, C.P. 1965 Disadvantages of the Olson Drift Card and description of a newly designed card. .J. Mar. Res. 23:350-354. Hjort, 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. 1977 Behaviour and survival of northern anchovy Engraulis mordax larvae. Calif. Coop. Oceanic Fish. Invest. Rep. 19: 138-146. Hutchings, L. 1981 The formation of plankton patches in the southern Benguela Current. In Richards, F.A. (ed.). Coastal and estu- arine sciences 1. Coastal upwelling, p. 496-506. Am. Geophys. Union, Wash. DC. Hutchings, L., S. Kuster, and J. Taunton-Clark 1988 Wind patterns in the south-west Cape. In Macdonald, I.A.W., and R.J.M. Crawford (eds.), Long-term data series relating to southern Africa's renewable natural re.sources, p. 28-32. S. Afr. Natl. Sci. Prog. Rep. 157, CSIRO, Pretoria, S, Afr. King, D,P.F,, A. A. Robertson, and P.A. Shelton 1978 Laboratory oliservations on the early development of the anchovy Engraulis capensis from the Cape Peninsula. Fish. Bull. S. Afr. 10:37-45. Largier. J.L., and V.P. Swart 1987 East-west variation in thermocline breakdown on the Agulhas Bank. In Payne, A.I.L., J. A. Gulland, and K.H. Brink (eds.), The Benguela and comparable ecosystems. S. Afr. J. Mar. Sci. 5:263-272. Lasker, R. 1975 Field criteria for survival of anchovy larvae: the relation between inshore chlorophyll maximum layers and successful first feeding. Fish. Bull., U.S. 73:453-462. 1978 Ocean variability and its biological effects - Regional review - Northeast Pacific. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 173:168-181. 1981a Factors contributing to variable recruitment of the northern anchovy (Engraulis mordax) in the California Cur- rent: Contrasting years, 1975 through 1978. Rapp. P.-V. Reun. Cons. Int. E.xplor. Mer 178:375-388. 1981b The role of a stable ocean in larval fish survival and subsequent recruitment. In Lasker, R. (ed.). Marine fish lar- vae, p. 80-87. Wash. Sea Grant Prog., tiniv. Wash., Seattle. Lasker, R., H.M. Feder, G.H. Theilacker, and R.C. May 1970 Feeding, growth and survival oi Engraulis mordax lar- vae reared in the laboratory. Mar. Biol. (Berl.) 5:345-353. Miller, I). 1961 A modification of the small Hardy plankton sampler for simultaneous high-speed plankton hauls. Bull. Mar. Ecol. 45:165-172. Parrish, R.H., A. Bakun, D.M. Husby, and C.S. Nelson 1983 Comparative climatology of selected environmental pro- cesses in relation to eastern boundary current pelagic fish reproduction. In Sharp, G. D., and J. Csirke (eds.). Proceed- ings of the expert consultation to examine changes in abun- dance and species composition of neritic fish resources, San Jose, Costa Rica, 18-29 April 1983. FAO Fish. Rep. 291: 731-777. Peterman, R.M., and M.J. Bradford 1987 Wind speed and mortality rate of marine fish, the north- ern anchvoy (Engraulis mordax). Science (Wash.. DC) 235: 354-3.56. SCOR/UNESCO Working Group 17 1966 Determination of photosynthetic pigments. In Deter- mination of photosynthetic pigments in sea-water. UNESCO Monographs on Oceanographic Methodology 1:9-18, Paris. Shannon, L.V., L. Hutchings, G.W. Bailey, and P.A. Shelton 1984 Spatial and temporal distribution of chlorophyll in southern African waters as deduced from ship and satellite measurements and their implications for pelagic fisheries. S. Afr. J. Mar. Sci. 2:109-130. Sharp. G.D. 1980 Oilonization: Modes of opportunism in the ocean. In. Sharp, G.D. (ed.). Workshop on the effects of environmental variation on the survival of larval pelagic fishes, Lima, Peru, April-May 1980, p. 15-59. IOC Workshop Rep. Ser. 28, IOC/UNESCO, Paris. Shelton, P.A. 1986 Fish spawning strategies in the variable southern Benguela Current region. Ph.D. thesis, Univ. Cape Town, Capetown, S. Afr., 327 p. 1987 Life-history traits displayed by neritic fish in the Benguela Current ecosystem. In Payne, A.I.L., J. A. Gulland, and K.H. Brink (eds.). The Benguela and comparable ecosystems. S. Afr. J. Mar. Sci. 5:235-242. Shelton, P. A., and L. Hutchings 1982 Transport of anchovy Etigraiilis rapensis tJilchrist eggs and early larvae by a frontal jet current. J. Cons. Int. Ex- plor. Mer 40:185-198. Shelton, P. A., A.J. Boyd, and M.J. Armstrong 1985 The influence of large-scale environmental processes on neritic fish populations in the Benguela Current system. Calif. Co(ip. Oceanic Fish. Invest. Rep. 26:72-92. Smith, P.E., and S.L. Richardson 1977 Standard techniques for pelagic fish egg and larva surveys. FAO Fish. Tech. Pap. 175, p. 100. Smith, P.E., W. Flerx, and R.P. Hewitt 1985 The CalCOFI Vertical Egg Tow (CALVET) net. In Lasker, R. (ed.). An egg production method for estimating spawning biomass of pelagic fish: Application to the northern anchovy, Engraulis mordax, p. 99. NOAA Tech. Rep. NMP'S 36, Natl. Oceanic ."Xtmos. .Ailm.. Natl. Mar. Fish. Serv. Walsh, J.J., I.E. Whitledge, W.E. Esaias, R.L. Smith, S.A. Huntsman, H. Santander, and B.R. de Mendiola 1980 The spawning liabitat of the Peruvian anchovy, Engraulis rmgens. Deep-Sea Res. 27A:l-27. Abstract.- A total 3236 km of boat and aircraft surveys was con- ducted in the northern Gulf of Cali- fornia in search of the vaquita Pho- coena sinus in 1986-1988. Vaquita were seen on 51 occasions, represent- ing an estimated 96 individuals. Forty-three porpoises (19 sightings) occurred during 1715 km of vessel transects, a rate of 2.51 individuals/ 100 km of surveys. The number of sightings relative to the extent of the survey emphasized the rarity of the porpoise and is a cause for concern regarding the vulnerability of the population. All sightings of vaquita occurred north of lat. 30°45'N, and all but two sightings (96.1%) occurred less than 40 km from San Felipe, Baja California Norte. Porpoises were observed in small groups (<3 individuals) in all but five sightings. The mean group size was 1.9 ± 1.24 SD individuals, and the mode was 1. All sightings occurred in water depths of 13.5-37.0 m (j 26.1 ±6.18 SD), with water clarity at these localities ranging from 0.9 to 12 m. Calves represented 9.37% of all individuals sighted. Unconfirmed sightings sug- gest that the porpoise is present in the southern Gulf of California. Fish- ermen interviewed in San Felipe, Baja California Norte, were familiar with vaquita, and some have caught the porpoise in gillnets. Occurrence and Distribution of thie Vaquita Phocoena sinus in the IMortliern Gulf of California Gregory K. Silber Institute of Marine Sciences. University of California Santa Cruz. California 95064 Manuscript accepted 23 January 1990. Fishery Bulletin, U.S. 88:339-346. The vaquita Phocoena sinus (Norris and McFarland 1958) is among the rarest and least understood of the Cetacea. In the 30 years following its initial description (Norris and McFar- land 1958) very little new informa- tion about the porpoise had been ob- tained. Brownell (1986) reported that only 45 confirmed records of vaquita existed. These consisted of skulls and skeletons (Noble and Eraser 1971, Orr 1969) and about 20 reported sight- ings obtained between 1958 and 1986 (Norris and McFarland 1958, Norris and Prescott 1961, Villa-R. 1976, Wells et al. 1981). However, Brown- ell (1986) argued that only four of the sightings were valid. Increased study of the porpoise in recent years has yielded new infor- mation through descriptions of recov- ered specimens (Magatagan et al. 1984, Brownell et al. 1987), reported sightings (Robles et al. 1986, Vidal et al. 1987, Silber 1988), and behavioral accounts (Silber et al. 1988). Norris and McFarland (1958) described the range of P. sinus as occurring in the upper Gulf of California and probably extending south along the Mexican coast, but Brownell (1986) main- tained that the range was limited to the upper Gulf of California. The vaquita population has been impacted by gillnetting activities and other forms of habitat degi'adation in recent decades (Brownell 1983, Bar- low 1986). The species continues to experience mortality at an unknown rate in gillnets (Brownell 1983, Silber 1988) set primarily for totoaba Toto- aba rfiacdonaldi and various species of shark. The objective for the initial year of the present study (1986) was to locate P. sinus, and subsequently (1987, 1988) to gather information on the ecology and distribution of the por- poise. Visual surveys were conducted from boat and aircraft in an attempt to determine the vaquita's distribu- tion, range, and habitat utilization. Reported here are the results of three seasons of study on the vaquita conducted in the northern Gulf of California in the springs of 1986-88. I have also included a discussion of unconfirmed sightings in the south- ern Gulf that may reflect the histor- ical range of P. sinus. Materials and methods Vessel surveys and sighting data Boat surveys for Phocoena sinus were conducted in the northern Gulf of California using an 8-m Boston Whaler. A total of 1715 km of survey transects was conducted on 71 days during the three years (Figs. 1, 2a). The study was conducted in the spring months because (1) operational costs and logistic (e.g., boat time) restric- tions limited field work duration, (2) conversations wath fishermen in the upper Gulf in 1986 suggested that va- quita were most abundant (perhaps only present) in March-May, and (3) the number of sightings that we ob- tained (two porpoises in 485 km sur- Reference to trade names does not imply en- dorsement by the National Marine Fisheries Service, NOAA. 339 340 Fishery Bulletin 88(2). 1990 Figure I Number of days per week spent conducting transects and observations of Phocoena sinus in 1986. 1987, and 1988. veyed) in February 1986 appeared to support informa- tion gathered from fishermen. For these reasons, the decision was made to concentrate survey efforts in March-May, when the likelihood of encountering the porpoise was believed greatest. During vessel transects, two to four observers posi- tioned 3.5 m above the water surface searched with unaided eyes and binoculars (7 and 10 x) the area in front, and about 200-300 m on either side, of the ves- sel's track. Observers also regularly scanned to greater distances. During all surveys the vessel traveled con- sistently 10-11 km/hour. Transects conducted in 1986 were restricted primarily to nearshore areas. In subse- quent field work inshore surveys were augmented by transects farther from shore. For each porpoise sighting, the number of individuals seen was estimated and sighting locations were deter- mined by triangulation from landmarks. Water depths were obtained with a JRC color video depth sounder or from a nautical chart. Water clarity (using a Secchi disc) and temperature were measured at most sighting locations. Sea states were obtained by visual estimate according to the Beaufort scale. No surveys were con- ducted when sea state exceeded Beaufort 3, and the majority were conducted in sea states of 1. Aerial surveys Limitations imposed by time, weather, and financial constraints prevented repeated vessel surveys south of lat. 30°30'. However, aircraft surveys were con- ducted as far south as lat. 29°34'N (Fig. 2b). A total of 1521 km was flown on three days, 3-5 May 1988, concentrating on the upper Gulf. Fishermen interviews Fishermen in San Felipe, Baja California Norte, and in the vicinity of La Paz, Baja California Sur, were in- terviewed in an attempt to obtain qualitative informa- tion with regard to the natural history, range, and distribution of the vaquita. Questions were asked re- garding type of fishing, target species, years of em- ployment, and location of fishing. The men were asked if they recognized vaquita after viewing photographs of living and dead P. sinus and other odontocetes common to the area. The interviews were intended as general informational surveys of local knowledge to direct future study of vaquita and preliminary assessments of fishing effort and vaquita entanglement rates. Information derived from interviews with fish- ermen has been presented, although questioning methodology and sample sizes were insufficiently rigorous to draw definitive conclusions with regard to vaquita distribution. Results Vaquita sightings Vaquita were seen on 51 occasions, representing an estimated 96 individuals during boat and aircraft surveys (Table 1). A total of 43 porpoises was seen while conducting vessel transects, a sighting rate of 2.51 individuals/100 km surveyed. The remainder of the sightings occurred while tracking porpoises, collecting other types of data, or otherwise not engaged in visual transects. The paucity of sightings relative to the extent of the survey emphasized the rarity of the porpoise. All sightings of vaquita occurred north of lat. 30°45'N, and all but two sightings (96.1%) occurred less than 40 km from San Felipe, Baja California Norte (Fig. 2a). Most sightings (94.2%) occurred in sea states 0 or 1 (Fig. 3). In 90.2% (n = 46) of all sightings, por- poises were observed in small groups (< 3 individuals). Mean group size was 1.9±1.24 SD individuals per group (Table 1) and the group size mode was one (Fig. 4). For 86.3% (n = 44) of all porpoise sighting locations, water depths ranged from 21 to 35 m (x 25.7 + 6.36 SD; overall range 13.5-37.0) (Fig. 5), and most sight- ings occurred 11 to 25 km from the nearest shore Silber: Phocoena sinus in the northern Gulf of California 341 BAJA CALIFORNIA Figure 2 Phocoena sinus sightings and vessel transects conducted in 1986, 1987, and 1988 (top) and aircraft surveys on 3-5 May 1988 (bottom). (x 18.3 ±6.10 SD; range 2.4-32.2) (Fig. 6, Table 1). Water clarity at sighting locations ranged from 0.9 to 12 m. During four vaquita sightings the boat's depth sounder indicated concentrated layers at 15 {n = 2), 23, and 25 m, possibly representing schooling bait fish or squid upon which the porpoises may have been feed- ing. On numerous occasions, vaquita surfaced in or alongside surface slicks or long bands of flat water sur- rounded by rippled water. Calves represented 9.37% of all individuals. One very young calf was observed on 9 April 1987. It was esti- mated at <2 days old because of its size, a dorsal fin 342 Fishery Bulletin 88(2), 1990 Table 1 Summary of all Phocoena sinus sightings in he northern Gulf of California, 1986-88. Group sizes are represented by mean, one standard deviation, and range. Distance to shore Depth Water temperature Water clarity Year and dates Sightings (km) (m) CO (m) 1986 N 16 16 16 12 12 2 Feb.-27 Mar. x 1.7 14.3 21.0 21.5 2.4 SD 0.87 4.50 5.00 1.83 0.81 Range 1-4 2.4-20.0 13.5-25.0 17.0-23.0 0.9-4.0 Est. total ind. 27 1987 A' 22 22 22 12 13 8 Apr. -7 May ? 2.1 19.3 27.5 22.9 8.2 SD 1.57 5.71 4.73 2.09 2.46 Range 1-7 15.3-32.2 20.0-37.0 21.0-26.0 5.0-12.0 Est. total ind. 46 1988 N 13 13 13 8 7 21 Mar. -5 May x 1.8 20.7 29.5 19.2 4.1 SD 1.01 6.58 5.87 0.46 1.07 Range 1-4 7.0-29.0 22.0-40.0 18.5-20.0 2.5-6.0 Est. total ind. 23 All years N 51 51 51 32 32 X 1.9 18.3 25.7 21.5 5.0 SD 1.24 6.10 6.36 2.23 3.10 Range 1-7 2.4-32.2 13.5-40.0 17.0-26.0 0.9-12.0 Est. total ind. 96 Figure 3 Phocoena sinv.'i sighting rate (individuals/100 km) relative to sea state. =■ 1(1 Group size (number of iiuiivitluals) N = 51 Figure 4 Phocoena sinus group size. that was not yet fully erect, and the ungainly manner in which it ]ifte(i its hea(^ from the water with each surfacing. Aerial surveys Eleven Phocoena sinus (five sightings) were seen dur- ing aircraft surveys. All aerial sightings took place between San Felipe and Rocas Consag, and none oc- curred farther south along the Baja peninsula coast (Fig. 2b), although much less effort was dedicated to this area in relation to the extreme upper Gulf. The number of aircraft surveys was relatively small and non-random. Nonetheless, they served to support P. sinus distribution patterns observed during vessel sur- veys, failed to yield sightings south of the area sur- veyed by vessel, and illustrated the utility of aircraft as a platform of observation in future study and cen- sus of the species. Sympatric vertebrates Other marine vertebrates were seen near P. siyms; Silber: Phocoena sinus in the northern Gulf of California 343 15 - E ^^ 10- = "ri S -o ^ 1 5- ul 1 1 ■ <16 16-20 21-25 26-30 31-35 36-40 >40 Water Depth (ni) Figure 5 Phocoena sinus sighting rate (individuals/ 100 km) relative to water depth. 20 ■ E ^ S as =;' Ul > .= "S c« .g 15 - 10 - 5 - 0 - ill.. <10 11-15 16-20 21-25 26-30 31-35 36-40 Dislunce from ntarest shore (lini) >40 Figure 6 Phocoena sinus sighting rate (individuals/100 km) relative to distance from nearest shore. affiliations that probably represent utilization of a simi- lar habitat by two or more species rather than active association. Twice, Bryde's vfha\es Balaenoptera edeni were seen < 1 km from P. sinus, and three times com- mon dolphins Delphinus delphis were observed <1.5 km from the porpoise. Numerous times black storm- petrels Oceanodroma melania and Bonaparte's gulls Larus Philadelphia dipped into the wake of surfacing vaquita. Manta rays Mania birostris were seen once near porpoises. In addition to vaquita, five other marine mammal species were common within the study area, including bottlenose dolphins Tursiops truncatus, common dol- phins, fin whales Balaenoptera physalus, Bryde's whales, and California sea lions Zalophus californi- anus. There appeared to be a general segregation by location and water depth among the three most com- mon odontocet species. The distribution of P. simis overlapped that oi Delphinus and B. edeni, but not that of Tursiops and B. physalus (Silber et al. In prep.). Bottlenose dolphins were consistently seen in more shallow water than were vaquita, and common dolphins were generally seen in greater, but comparable, water depths as vaquita. Data on sympatric marine mammals will be presented in greater detail elsewhere (Silber et al. In prep.). Unconfirmed sightings and fishermen interviews In February and April 1983 cetaceans that may have been P. sinus were seen near Cerralvo Island (24°10'N; 109°55'W) by a scientist who has extensive experience with marine mammals and Gulf of California fauna (G. Notarbartolo di Sciara, Piazza Duca d'Aosta 4, 20124 Milano, Italy, pers. commun., May 1984). In both sight- ings, the animals were 10-50 m from the observer. They were described as being small, possibly possess- ing a blunt rostral profile, and surfacing "unobtrusive- ly" in small groups. They were "quite distinct from Tursiops, Delphinus, or Lagenorhynchus" (G. Notar- bartolo di Sciara, pers. commun.. May 1984). If true, these unconfirmed sightings occurred in an area >850 km south of our southernmost sightings, and extend the present known range considerably. Based on these observations, the decision was made to interview fishermen in the southern Gulf about the presence of vaquita. In 1987, 17 fishermen from La Paz were interviewed about their fishing practices and their knowledge of the vaquita (Table 2). In most cases the fishermen had spent considerable time on the water (fishing 4-6 days/week. 3-12 months/year, for 5-51 years; x = 20.6 yrs). Fishermen in the La Paz area generally had no knowledge of vaquita; however, one man was familiar with the porpoise and indicated that he had seen it several times near San Jose Island (25°00'N; 110°40'W), 80 km north of La Paz. Unlike others interviewed, this man was one of two who had been fishing for over 50 years and he had worked in an area (the coast near San Jose Island) different from the other fishermen. In addition, his knowledge of the natural history of marine mammals was regarded as being accurate by the interviewer (D. Aurioles, Cen- tro de Investigaciones, Biologicas B.C.S., APO Postal 128, La Paz, Baja California Sur, Mexico, pers. commun., June 1987). One fisherman described netting near San Jose Island a small cetacean that had rounded ("acorn-like") teeth, a feature characteristic of the genus Phocoena. All fishermen interviewed in San Felipe (n = 7) expressed a knowledge of vaquita, and two said that they had entangled the porpoise while fishing for totoaba. Although taking totoaba is illegal, three men indicated that they continue to use gillnets for totoaba. 344 Fishery Bulletin 88(2). 1990 Summary of interviews with fisherme and knowledge of Phocoena sinus. n in La Paz Table 2 , Baja California Sur, and San Felipe , Baja California Norte, regarding fishing practices San Felipe La Paz Number fishermen interviewed 7 17 Distance from home port (km) X SD Range 37.4 39.48 1-100 37.8 22.32 5-112 Number of years engaged in fishing X SD Range 17.8 12.50 6-35 23.4 15.51 5-51 Type of fishing Gillnet, longline, hook-and-line Longline, gillnet, trawl, diving Target species Shrimp {n =6) Totoaba {n = 3) Shark (n = 2) Marlin {n = 1) Dorado (n = 1) Yellowtail jack (» = = 1) Spanish mackerel (n = 14) Shark {n = 9) Pacific manta (h = 1) Snapper (n = 1) Striped mullet (n = 1) Fishing effort 12 mo/yr (n = 4) 3 mo/yr (« = 3) 12 mo/yr (n = 11) 10 mo/yr (n = 1) 3-6 mo/yr {n = 3) Recognition of photos of P. sinus Yes (n = 7) Yes(n= 1); No (n = 16) Estimated total porpoises killed 3-4 0 Common names used Cochito, cochinito, and vaquita = P. simis Vaquita = Kogia or Phocoena spp. Duende = Lagenorhynchus Cochinito = Delphinus Cochito = Globicephata The vernacular name for P. sinus tended to vary between communities and individuals. Fishermen in San Felipe referred to P. sinus as "cochito," "cochi- nito," and "vaquita." In contrast, for La Paz fishermen "vaquita" refers to Kogia simus or Phocoena spp., whereas "cochinito" refers to Delphinus delphis, "cochito" to Glohicephala macrorhynchus, and "duende," also meaning P. sinus in some communities (Magatagan et al. 1984), refers to Lagenorhynchus obliquidens (D. Aurioles, pers. commun., June 1987). Discussion Although surveys were conducted throughout the up- per Gulf, sightings of Phocoeyia sinus were limited almost entirely to the western side. One or a combina- tion of the following features may account for vaquita distribution patterns in the upper Gulf. Relative to the eastern shore, the western boimdary exhibits lower sur- face temperatures (Robinson 1973) resulting partly from northward-bearing upwelled water from the Midriff Island region. With respect to the eastern coast- line of the upper Gulf, the western perimeter is char- acterized by stronger north-south tidal currents year- round (Hendrickson 1973) which may contribute to increased mixing on the west coast. Vaquita often surfaced near surface slicks, which are caused by tidally induced internal waves (Ewing 1950, Shanks 1983) or areas of convergence of different water masses, a common feature of the northern Gulf of California (Hendrickson 1973, Lepley et al. 1975). Surfacing near slicks has been noted in other cetaceans, including Feresa attenuata, Steyio breda.nensis, and Pseudorca crassidens (K.S. Norris, Cent. Mar. Stud., Univ. Calif., Santa Cruz, CA 95064, pers. commun., Aug. 1987). Concentrations of small marine organisms are associated with internal waves (Shanks 1983, 1988; Kingsford and Choat 1986). Vaquita and other ceta- ceans may be drawn to aggregated prey in these features. It is believed by some researchers that the entire P. sinus population is limited to the upper Gulf, which represents the smallest range of any marine cetacean (Brownell 1986, Barlow 1986); however, the actual range remains unsubstantiated. Although presently Silber Phocoena sinus in the northern Gulf of California 345 untested, I suggest that vaquita occur in the upper Gulf year-round, and although the majority of the popula- tion may have receded to the upper Gulf following a relatively recent decline in population size, some in- dividuals are probably scattered throughout the Gulf as separate subpopulations or a sparsely dispersed single population. Reported sightings of P. sinus out- side the northern Gulf (Scammon 1874, Norris and McFarland 1958, Norris and Prescott 1961, Villa-R. 1976; presented here) suggest the possibility of a greater historical range. However, extensive field work on cetaceans near the Midriff Islands, Guaymas, and Kino Bay areas has not yielded sightings of vaquita (L. Ballance, Dep. Biol., Univ. Calif., Los Angeles, CA 90024, pers. commun.. May 1989; L.T. Findley, Escu- ela de Ciencias Man'timas y Alimentarias, Institute Tecnologico y de Estudios Superiores de Monterrey, Unidad Guaymas, A. P. 484, Guaymas, Sonora, 85400 Mexico, pers. commun., Nov. 1985; B. Tershy, Dep. Biol., Cornell Univ., Ithaca, NY 14853, pers. commun., Sept. 1987). Most phocoenids are cold-temperate water animals (Gaskin 1982), with ranges restricted to <20°C. Northern Gulf of California surface water temperatures in the summer and fall can exceed 28 °C (Hendrickson 1973, Robinson 1973). If the porpoise is limited to this region, it deviates remarkably from other phocoenids in its ability to tolerate seasonal water temperature fluctuations. Management recommendations Exposure of the depleted P. sinus population to cur- rent levels of human impact may result in the demise of the species. Habitat degradation and pesticide con- tamination of the northern Gulf may contribute to the decline of P. sinus. However, vaquita mortality in gillnets (Brownell 1983, Robles et al. 1986, Silber 1988) may represent the most direct threat to the population, and increased efforts to reduce porpoise entanglement is imperative. Nets set for totoaba probably account for the highest percentage of incidental porpoise deaths. The totoaba is an endangered species and can- not be legally caught, sold, or exported. Nonetheless, illegal fishing continues due to the inability to enforce restrictions. Additional study is needed to quantify incidental mortality rates and to explore possible modifications of fishing amounts, timing, and tech- nique. Vaquita sightings occurred predominantly in discrete locations and water depths. If widespread fishing bans are not possible, emphasis should be placed on excluding fishing from these areas. However, because maximum catch of totaba also occurs in spring (Flanagan and Hendrickson 1976), there will be difficul- ty achieving compliance by fishermen. Rigorous sanc- tions on importation of totoaba bound for U.S. markets are required. Additional boat and aircraft census surveys should be conducted to determine the vaquita population size, and to enhance current knowledge of the ecology of the porpoise. Large areas of the northern Gulf have been recommended for consideration as a Marine Sanctuary (B. Villa-R., Inst. Biol., Univ. Nal. Auton. Mexico, A.P. 70-153, Mexico, D.F. 04510, pers. commun., Feb. 1986), and these plans should be en- dorsed. Mexican and U.S. scientists should engage in joint efforts to monitor changes in the vaquita popula- tion size. Action by Mexican and U.S. governments, fishing cooperatives, cetacean biologists, and resource management personnel is necessary to the preserve the species. Acknowledgments This study would not have been possible without im- portant contributions by the following field assistants: T. Silber, M. Newcomer. G. Ellis, H. Perez-Cortes, G. Barros, D. Breese, T. Jefferson, L. Torrez-M., A. Velazquez-R., and A. Robles. These people endured the rigors of field work and high expectations of a demanding captain. 1 extend thanks to K. Norris, B. Villa-R., and R. Wells who provided encouragement and advice. I am grateful to D. Aurioles and H. Perez- Cortes for interviewing fishermen in La Paz and San Felipe, respectively, and for observations provided by G. Notarbartolo di Sciara. The project received finan- cial support from the Nature Conservancy, the Center for Marine Conservation, and the American Cetacean Society (Los Angeles Chapter), and logistical support from the Center for the Study of Deserts and Oceans, Project Lighthawk, and West Coast Whale Research Foundation. The study was conducted under Scientific Research Permits 301856, 300422, and 400036 issued by the Secretaria de Pesca, Mexico. The paper was improved by comments from K. Norris, B. Wursig, R. Wells, M. Newcomer, and T. Jefferson. This paper is dedicated to my wife Trish. Citations Barlow, J. 1986 Factors affecting the recovery of Phocoena sinus, the vaquita or Gulf of CaUfornia harbor porpoise. Adm. Rep. LJ-86-37, Southwest Fish. Cent., Natl. Mar. Fish. Serv., NOAA. P.O. Box 271, La .Jolla, CA 92038, 19 p. Brownell, R.L., Jr. 1983 Phocoena sinus. Mamm. Species 198:1-3. 1986 Distribution of the vaquita. Phocoena sinus, in Mexican waters. Mar. Mammal Sci. 2:299-305. 346 Fishery Bulletin 88(2), 1990 Brownell, R.L.. Jr., L.T. Findley, 0. Vidal, A. Robles. and S. Manzanilla N. 1987 External morphology and pigmentation of the vaquita, Phocoena sinu^ (Cetacea: Mammalia). Mar. Mammal Sci. 3:22-30. Ewing, G. 1950 Slicks, surface films and internal waves. J. Mar. Res. 9:161-187. Flanagan, C.A., and J.R. Hendrickson 1976 Observations on the commercial fishery and reproduc- tive biology of the totoaba. Cynoscion mafdonaldi, in the north- ern Gulf of California. Fish. Bull., U.S. 74:531-544. Gaskin, D.E. 1982 The ecology of whales and dolphins. Heinemann Educ. Books. Exeter, NH. 4.59 p. Hendrickson, J.R. 1973 Study of the marine environment of the northern Gulf of California. Contract rep. NAS5-21777, Goddard Space Flight Cent., Greenbelt, MD 20771. Avail. NTIS Publ. N74-16008, 95 p. Kingsford, M.J., and J.H. Choat 1986 Influence of sui-face slicks on the distribution and onshore movements of small fish. Mar. Biol. (Berl.) 91:161-171. Lepley, L.K., S.P. Vonder Haar, J.R. Hendrickson, and G. Calderon-RiveroU 1975 Circulation in the northern Gulf of California from or- bital photographs and ship investigations. Cienc. Mar. 2:86-93. Magatagan, M.D., E.H. Boyer, and B. Villa-Ramirez 1984 Revision del estado que guarda Phocuena sinus Norris and McFarland y descripci6n de tres nuevos ejemplares. An. Inst. Biol. Univ. Nal. Auton. Mex., Ser. Zool. 55:271-294 [in Spanish, Engl, abstract]. Noble, B.A., and F.C. Fraser 1971 Description of a skeleton and supplementary notes on the skull of a rare porpoise Phocoena sums Norris & McFarland, 1958. J. Nat. Hist. 5:447-464. Norris, K.S., and W.N. McFarland 1958 A new harbor porpoise of tlie Genus Phocoena from the Gulf of California. J. Mammal. 39:22-39. Norris, K.S.. and J.H. Prescott 1961 Ob.se r\-at ions on Pacific cetaceans of Califomian and Mex- ican waters. Univ. Calif. Publ. Zool. 63:291-370. Orr, R.T. 1969 An additional record of Phocoena simis. J. Mammal. .50:384. Robinson, M.K. 1973 Atlas of monthly mean sea surface and subsurface temperatures in the Gulf of California, Mexico. San Diego Soc. Nat. Hist. Mem. 5. p. 4-97. Robles, A., L.T. Findley. O. Vidal, R.L. Brownell, Jr., and S. Manzanilla 1986 Registros recientes y apariencia externa de la marsopa del Golfo de California, o vaquita. Phocoena sinus, Norris and McFarland, 1958. XI Reuni6n Intemacional Sobre Mamiferos Marinos, Guaymas, Sonera. Avail. Instituto Tecnologico y de Estudios Superiores de Monterrey, Unidad Guaymas, A.P. 484, Guaymas, Sonera, 85400 Mexico [Spanish and English abstract]. Scammon, CM. 1874 The marine mammals of the northwestern coast of North America. J.H. Carmany. San Francisco, 319 p. Shanks. A.L. 1983 Surface slicks associated with tidally forced internal waves may transport pelagic larvae of benthic invertebrates and fishes shoreward. Mar. Ecol. Prog. Ser. 13:311-315. 1988 Further support for the hypothesis that internal waves can cause shoreward transport of larval invertebrates and fish. Fish. Bull., U.S. 86:703-714. Silber. G.K. 1988 Recent sightings of the Gulf of California harbor porpoise, Phocoena sinus. J. Mammal. 69:430-433. Silber, G.K., M.W. Newcomer, and G.J. Barros 1988 Observations (m the behavior and ventilation cycles of the vaquita, Phocoena sinus. Mar. Mammal Sci. 4:62-67. Silber. G.K., M.W. Newcomer. H. Perez-Cortes M., P.O. Silber, and G. Ellis In prep. Distribution of marine mammals of the northern Gulf of California. Inst. Mar. Sci., Univ. Calif., Santa Cruz, CA 95064. Vidal. O.. A. Aguayo, L. Findley, A. Robles, L. Bourillon, I. Vomend, P. Turk, K. Garete, L. Maronas, and J. Rosas 1987 Avistamientos de mamiferos marinos durante el crucero "Guaymas I" en la region superior del Golfo de California, Primavera de 1984. Memoria X Reunion Intemacional Sobre Mamiferos Marinos, 24-27 de Marzo de 1985, La Paz, Baja California Sur. Avail. Instituto Tecnol6gico y de Estudios Superiores de Monterrey, Unidad Guaymas, A.P. 484, Guaymas, Sonora, 85400 Mexico, 196 p. [in Spanish, Engl. abstr.[. Villa-R., B. 1976 Report on the status of Phocoena sinus, Norris and McFarland 1958, in the Gulf of California. An. Inst. Biol. Univ. Nal. Aut6n. Mi^x., .Ser. Zool. 47:203-208, Wells, R.S., B.G. Wiirsig, and K.S. Norris 1981 A survey of the marine mammals of tlie Upper Gulf of California, Mexico, with an assessment of the status of Pho- coena sinus. Final rep. MMC-79/07 to U.S. Mar. Mamm. Comm., Wash. DC. Avail. NTIS PB81-168791, 51 p. AbStTelCt. — Fishery observers aboard foreign commercial fishing vessels collected information on the incidental catch of marine mammals in the Exclusive Economic Zone (EEZ) off the northeastern United States since March 1977. Observer cover- age on foreign vessels was 25-35% dm-ing 1977-82, and incre^ised to 58%, 86%. 95%, 98%, 100%, and 100%, respectively, in 1983-88. Dm-ing 1981- 88, observers have covered most joint- venture fishing operations. During 1977-88, observers reported 538 ma- rine mammals captured incidental to direct and joint-venture fishing activ- ities. Eight cetacean species and three unidentified baleen whales were cauglit, principally in the fish- eries for Atlantic mackerel Scomber scombms, and squid Illex illecebrosus and Loligo pealei. Pilot whales Globi- cephala spp. (297/538) and common dolphins Delphinus delphis (203/538) comprised 93% of the catch. Chi- square tests indicate that significant differences in diel rates of capture occurred between the two species. The number of Globicephnla spp. captured at night (2000-0400 h) in the Atlantic mackerel fishery was significantly less (j- = 8.28, P<0.03) than the number caught during day (0800-1(500 h) or dawn/dusk (1600- 2000 h, 0400-0800 h). The number of D. delphis captured during day- light in the Loligo squid fishery was significantly less (x- = 44.48, P < 0.001) than the number caught at night or dawn/dusk. A minke whale B. acutorostmta (released alive) and individuals of two endangered spe- cies, a humpback whale Megaptera novaeangliiic (released alive) and a right whale Eubalaena glacialis, were also captured incidental to fish- ing activity. Dui'ing December 1986-. February 1988, observers collected whole, dead, non-endangered mam- mals for detailed shoreside e.xamina- tion. Trawl contents at the time of capture and subsequent analysis of mammal stomach contents suggest that L. pealei is a major component of the mid-shelf and shelf-edge diet of common dolphins and pilot whales. Further, pilot whales, considered prin- cipally as teuthophagous, were ob- served to selectively feed on mack- erel while (in the Continental Shelf. Manuscript accepted 16 January 1990. Fishery Bulletin, U.S. 88:347-360. Incidental Take of Marine Mammals in Foreign Fishery Activities Off the Mortheast United States, 1977-88 Gordon T. Waring Patricia Gerrior Northeast Fisheries Science Center, National Marine Fisheries Service, NOAA Woods Hole, Massachusetts 02543 P. Michael Payne Manomet Bird Observatory, Manomet, Massachusetts 02345 Betsy L. Parry Department of Land i'Jesources, Institute of Environmental Studies University of Wisconsin, Madison, Wisconsin 53715 John R. Nicolas Northeast Fisheries Science Center, National Marine Fisheries Service, NOAA Woods Hole, Massachusetts 02543 Marine mammal/fishery interactions in United States waters have received widespread attention in recent years (Bonner 1982, Fowler 1982, Lowry 1982, Loughlin et al. 1983, Loughlin and Nelson 1986). These interactions, generally involving commercial fish- eries (Mate 1980), are of two types: (1) direct or operational, and (2) in- direct or ecological (Lowry 1982). In the shelf waters off the north- eastern United States, marine mam- mal/commercial fishery interactions have been described only for fisheries occurring in nearshore waters (Gil- bert and Wynne 1985). These interac- tions occur in the fixed-gear fisheries for American lobster Homwrus ameri- ciutus, the surface-gillnet fishery for Atlantic herring Clupea harengus and Atlantic mackerel Scomber scotn- brus, and the groundfish-gillnet fish- ery for assorted finfish, principally Gadidae and Pleuronectidae. Two principal marine mammals taken in- cidentally in these fisheries are the harbor porpoise Phocoena phocoena and the harbor seal Phoca vitulirta. The gray seal Halichoerus grypus is infrequently captiu'ed. These three spe- cies are known to feed on fish caught in nets and to tecome entangled, there- by damaging fishing gear. Marine mammal/fishery interactions in the deeper, offshore waters off the northeastern United States have not been previously documented. Under the provisions of the United States Marine Mammal Protection Act (MMPA) of 1972 a General Permit system was established by the Nation- al Marine Fisheries Service (NMFS) allowing incidental taking of marine manmials in commercial fishing oj)er- ations. All domestic and foreign fish- ing vessels were required to have a valid permit on board that established an allowable limit on the number of non-endangered marine mammals that could be taken within a specified fish- ery (i.e., mackerel, squid). The Mag- nuson Fishery Conservation and Man- agement Act of 1976 as amended in 1983, (MFCMA, Public Law 94-265) mandated the placement of Fisheries Compliance Inspectors, or observers. 347 348 Fishery Bulletin 88(2). 1990 beginning in 1984, on 100% of all foreign fishing ves- sels operating within the 200-mile outershelf and slope waters off the United States. (This area is fur- ther referred to as the Exclusive Economic Zone, or EEZ). Since that time in 1984, reporting of marine mammal takes incidental to foreign fishing activities has been more effectively monitored by species and fishery. In this paper we summarize the observed incidental take of marine mammals from all foreign fishing ac- tivities that operated within the EEZ off the north- eastern United States from 1977 to 1988. The inciden- tal take of common dolphins Delphinus delphis and of pilot whales Globicephala spp.*, the two most frequent- ly caught species, are examined relative to the type of fishery in which they were taken, the nationality in- volved in the fishery operation, and the time of the in- cidental take (day versus night). Morphological mea- surements, sex, and stomach contents for 33 of the common dolphins taken during 1986 and 1988 are also reported in this study. These measurements and records are the first such information taken from any marine mammal specimens recently killed in the waters off the northeastern United States. Summary of foreign fisheries Fishing operations Large-scale foreign fishing activity began off the U.S. east coast in the early 1960s with the arrival of the distant-water fleet (DWF) from Europe and Japan. Ini- tially, the DWF harvested groundfish (Atlantic cod Gadus morhiia, haddock Melanogrammus aeglefinus, hakes, flounders, etc.) and pelagics (Atlantic herring and Atlantic mackerel) (Brown and Halliday 1983). Declining fish stocks, coupled with management mea- sures implemented in the late 1960s, under the auspices of the International Commission for the Northwest Atlantic Fisheries, resulted in the DWF shifting em- phasis to "under-utilized" species, such as long-finned squid Loligo pealei, short-finned squid Illex illecebrosus, and Atlantic butterfish Peprilns triacanthus. The MFCMA, which became effective March 1977, significantly altered DWF activities within the EEZ. Fishing effort was immediately reduced, and the num- *The distribution of the Atlantic pilot whale Globicephala melaena, the northern species, overlaps with the short-finned pilot whale G. macrorhyncha, mainly a southern species, between about 3.5°30N to 38°00'N (Leatherwood et al. 1976). Although, G. melaena is the most common and the most lil8°C (Lange and Sissenwine 1980). During spring and summer Loligo moves into nearshore waters to spawn (Summers 1967). Commercial concentrations are found primarily off southern New England and the mid-Atlantic regions from about Georges Bank to Baltimore Canyon (Lange 1980, Lange and Johnson 1981). DWF fishing prior to the MFCMA occurred at depths of 100-500 m along the shelf edge from Novem- ber through March (Lange 1980), but since 1977 fishing operations have been restricted in area and season. Since 1987, the allocation iov Loligo has been reduced strictly to bycatch in the Atlantic mackerel fishery; therefore, DWF Loligo fishing has been suspended. ///ex or short-finned squid The short-finned squid (hereafter referred to as Illex) is a more northern species than Loligo, ranging from Florida to Newfoundland (Squires 1957). In late autumn (October-December), Illex moves offshore toward the shelf edge and beyond (Lange and Sissen- wine 1980). Spawning is believed to occur offshore at great depths, primarily from December to March (Lange and Sissenwine 1980). During warmer months Illex move nearshore to feed (Lange 1980). Commer- cial concentrations occur from the mid-Atlantic, near Baltimore Canyon, to Newfoundland (Lange and Johnson 1981). The DWF Illex fishery took place from June through October, while this species is on the shelf (Lange 1980). In 1987, Illex allocations for DWF fisheries were suspended. At present, there is no direct foreign fishing effort for this species. Atlantic mackerel Atlantic mackerel, a major target species of the DWF, overwinter along the edge of the continental shelf between Cape Hatteras and Sable Island, Nova Scotia (Anderson and Paciorkowski 1980). The overwintering population consists of two components (southern and northern), which separate in spring. The southern com- ponent moves nearshore in early spring along the 350 Fishery Bulletin 88(2), 1990 Virginia Capes and proceeds northeastward to spawn off the New Jersey and Long Island coasts from mid- April to early May. By midsummer, the southern com- ponent reaches the Gulf of Maine where it remains throughout summer (Anderson and Paciorkowski 1980). Migration out of the Gulf of Maine begins in autumn as the southern component returns to deeper, offshore waters. The northern component migrates fur- ther north in spring along the Nova Scotian shelf to the Gulf of St. Lawrence where spawning occurs dur- ing June- July (Anderson and Paciorkowski 1980). The northern component begins leaving the Gulf of St. Lawrence in September and returns to the mid- Atlantic region to overwinter. The seasonal DWF mackerel fishery occurs in the mid-Atlantic region during the winter when both com- ponents of the mackerel population concentrate primar- ily along the shelf edge (Anderson and Paciorkowski 1980), although commercial quantities are sometimes encountered in waters as shallow as 30 m. ceans killed per day (kill-rate) in the 1984-88 Nether- lands and the GDR mackerel fishery. Data for Poland were not included, because that country did not begin commercial fishing until 1988. Because a Mann-Whit- ney test examines only differences in paired data sets, a Friedman rank sums test (Hollander and Wolfe 1973) was used to examine the kill-rates between Italy, Spain, and Japan in the 1984-86 Loligo squid fishery. Randomly selected samples of 334 trawl logs from the 1985 Atlantic mackerel fishery (from Netherlands) and 1018 trawl logs from the 1985 Loligo squid fishery (from Italy) were used to determine the expected level of fishing effort (number of tows) for each of three time periods-Day (0800-1600), Dawn/Dusk (0400-0800 and 1600-2000)," and Night (2000-0400)-within each fish- ery. The Chi-Square Statistic was then used to deter- mine whether there were significant differences in the numbers of common dolphins and pilot whales inciden- tally taken in each fishery (relative to fishing effort) due to time of day of the take. Methods and materials Data collection on incidental take Observers monitor compliance with applicable fishing regulations, collect biological data and samples, and serve in a liaison role, when needed, between foreign and domestic captains. In the northeast region, ob- server training is the responsibility of the NMFS Observer Program. Between 1977 and 1985, observers routinely recorded the number of marine mammals taken incidentally in foreign fishing activities. In 1986, a new sampling protocol was implemented to collect additional information on marine mammals, sea turtles, and marine debris. Obsei-vers are now required to com- plete sighting forms, document the circumstances of capture and obtain biological data on incidentally cap- tured marine mammals* and sea turtles. Marine mam- mal biological sampling includes five straight-line body measurements (e.g., snout to fluke notch, flipper length, flipper width, fluke width, and dorsal fin height), girth at pectorals, sex, and total weight. Additionally, when feasible, incidentally caught marine mammals are frozen whole, brought ashore, and later examined by researchers at the Smithsonian Institution. Data analysis on incidental take A paired Mann-Whitney test (Zar 1974) was used to test for differences between the total number of ceta- *In this paper, an incidental take i.s defined as any live marine mam- mal or any carcass caught during foreign fishing activity. Data analysis on stomach contents, sex and total length A total of 95 common dolphins and 169 pilot whales taken between 1986 and 1988 were measured (total standard length) and sexed at sea. Stomach contents from 33 of the common dolphins were examined by Smithsonian Institution personnel, and the information was provided for this paper. Prior to 1986, stomach contents and sex data were not collected. Results Observers reported 538 marine mammals incidentally caught during foreign fishing activities in the EEZ off the northeast United States between March 1977 and December 1988 (Table 1). Pilot whales were the most frequently caught marine mammal with 297 caught, representing 55% of the total marine mammal take between 1977 and 1988 (Table 1). Common dolphins were the next most-frequently-taken cetacean with 203 caught, composing 38% of the total incidental takes. Approximately 5% of the total catch consisted of three other members of the Delphinidae: Atlantic white-sided dolphin Lagenorhychus acutus, bottlenose dolphin Tur- siops trvncatus, and Risso's dolphin Graynpuft griseus (Table 1). Six large whales were reported caught or entangled during fishing operations. One subadult or juvenile right whale Eubalaena glacialis (based on observer identification and estimated length of 6 m) and one unidentified baleen whale were taken in the Loligo fishery. A humpback Megaptera novaeangliae, minke Waring et al : Incidental take of marine mammals off northeast United States 351 Table 1 Summary of incidental take of marine mammals as reported by U.S. fisheries observers on foreign vessels in the shelf and slope waters | of the northeastern United States. Values in parantheses represent animals released alive, (jv = joint venture) Year/ Common Pilot White-sided 1 3ottlenose Risso's Unid. Right Humpback Minke Unid. Fishery Country dolphin whale dolphin dolphin dolphin dolphin whale whale whale whale Total 1977-83 lUex Italy Japan Mexico Spain USA Gv) 1 1 1 3 6 2 3 1 1 3 6 Loiigo Italy Japan Mexico Spain 2 1 1 2 3 5 1 1 1 1(1) 4 5 1 8(1) Tuna Japan 2 1 1(1) 1 5(1) Hake USSR 2 1 1 4 Mackerel Poland 4 13 1 18 Totals 8 39 3 4 1 1(1) 3(1) 59(2) 1984 Illex USA (jv) 1 1 Loiigo Italy Spain 3 3 7 2 8 7 Tuna Japan 1 1 Mackerel GDR Netherlands Poland 1 11 1 1 11 1 Totals 3 23 2 1 1 30 1985 Loiigo Italy Japan Spain 56 1 4 1 4 1 1 58 1 9 Mackerel GDR Netherlands Poland USA (jv) 5 8 32 1 1 3 3 11 40 1 1 Totals 66 47 3 3 1 1 121 1986 Loiigo Italy Spain USA (jv) 54 1 1 2 1 56 1 2 Tuna Japan 1(1) 1(1) Mackerel GDR Poland USA Gv) 4 3 14 12(1) 1(1) ^ 4 16(1) 4(1) 18 Totals 76 20(2) 1 1(1) 98(3) U)87* lUex Italy 1 1 Mackerel GDR Netherlands USA av) 11 7 1 12(1) 11 3 23(1) 18 4 Totals 19 26(1) 1 46(1) 1988 Mackerel GDR Netherlands 20 27 50 4 1 47 55 Poland 11 58(3) 4 1 1 75(3) USA Gv) 7 7 Totals 31 142(3) 8 2 1 184(3) 1977-88 Tc tals 203 297(6) "ishing operations 13 ceased in 1986 9 4 6 1 1(1} 1(1) 3(1) 538(10) •Directed foreign Loiigo 352 Fishery Bulletin 88(2|. 1990 Figure 2 Distribution of incidental takes of common dolphin in foreign fishing activities off the northeastern United States, 1977-88. Balaenoptera acutorostrata, and an unidentified baleen whale were also reported taken in the tuna longline fishery (Table 1). Another unidentified baleen whale was taken in the hake fishery. Location of incidental takes of common dolphins and pilot whales A total of 68% (n = 136) of the common dolphins cap- tured in foreign fishing activities were caught along the shelf edge, represented by the 100-m isobath, be- tween 37°30'N and 40°00'N latitudes (Fig. 2), and 69% (n = 138) of the common dolphins were taken between December and February. Most (n = 152) of the pilot whales captured during foreign fishing activities also occurred along the shelf edge (Fig. 3), and 83% {n = 246) were captured between March and July. Of the total number of pilot whales taken incidentally to fishing, 8% (n = 24) were captured between December and February, as compared with 69% for common dolphins. Incidental takes by fishery and country During 1977-83, 54% (32/59) of the incidental bycatch of marine mammals in foreign fishing activities was taken in the Loligo and Illex fisheries (Table 1). This high percentage is consistent with the distribution of DWF fishing effort during this period. Since 100% observer coverage of most fisheries did not begin until 1984, the number of marine mammals taken during this period is probably under-represented. Lotlgo and Illex fisheries A total of 61% (n = 123) of the total number of common dolphins taken during 1977-88 {n = 203) were caught in the Loligo fishery, and 94% (w = 1 16/123) of these takes occurred during 1985-86. Again, takes prior to 1984 were probably under-represented. Italian vessels caught 93% of the common dolphins taken in the Loligo fishery. The total number of pilot whales caught in the Loligo fishery (n = 28) was 9.4% of the total number of pilot whales taken incidentally to all foreign fishing activ- ities. Of the pilot whales taken in the Loligo fishery between 1984-86, 61% (»= 11/18) were taken by Spain. No common dolphins were taken in the Illex fishery; however, 13 pilot whales were taken between 1977 and 1984. Of the pilot whales taken, 46% (n = 6) were caught by U.S. joint-venture vessels. Atlantic mackerel fishery During 1984-88, 76 com- mon dolphins were taken in the fishery for Atlantic mackerel (Table 1). Of the common dolphins taken in this fishery, 46% (« = 35/76) were caught by GDR vessels, 19% (« = 15) by U.S. joint-venture vessels, 18% (n = 14) by Polish vessels, and 17% {)i = 12) by Dutch vessels. Waring et al.: Incidental take of marine mammals off northeast United States 353 Figure 3 Distribution of incidental takes of pilot whales in foreign fishing activities off the northeastern United States, 1977-88. Of the DWF fisheries mortality of pilot whales since 1984, 93% occurred in the Atlantic mackerel fishery (Table 1). During 1984-87, the Netherlands and the GDR caught 54 and 33 animals, respectively, which combined represent 90% (n = 97) of the pilot whales taken in the mackerel fishery during that 4-year period. In 1988, Poland began commercial fishing for Atlan- tic mackerel and caught 41% (n = 58) of the pilot whales taken that year. Although the Netherlands did not fish in the EEZ during 1986, they still caught 44% (104/239) of the pilot whales taken since 1984 in the Atlantic mackerel fishery. Summary of kill rates by fishery and country Loligo fishery The total number of marine mammals (all species) killed per days fished (k/d rate) by Italy in the directed Loligo fishery was more than six times that of either Japan or Spain between 1984 and 1986 (period of nearly 100% observer coverage) (Table 2). During the years 1985 and 1986, this rate was approximately 10 times greater than the same rate for Spain, in spite of comparable number of days fished (Table 2). The k/d rates for 1984-86 between the three countries, how- ever, are not significantly different using the Fried- man test (P<0.19). Atlantic macl'* ^l"' ^a" n,»- •»,»- n,W LENGTH CLASSES (CM) tfi ,°-aP Figure 5 Length distribution of incidentally captured pilot whales measured by observers aboard foreign fishing vessels off the northeastern United States, 1986-88. Size at attain- ment of sexual maturity is from Perrin and Reilly (1984). Incidental take of common dolphins and pilot whales relative to time of day and fishery Significantly more common dolphins were captured during the Loligo fishery at night than during other time periods throughout the day. Possibly the diurnal movement upward at night would concentrate Loligo nearer the surface, thereby also concentrating common dolphins feeding on this prey. This would result in an increased likelihood of common dolphins capture due to a narrowing of the fishing area within the water column. If common dolphins follow squids downward during daylight hours, then this would account for the lack of surface-feeding observations during the daytime. Con- versely, if common dolphins feed principally at night, then common dolphins might become spatially separ- ated from squid during daylight hours, resulting in a decreased likelihood of incidental capture. Although the reasons that common dolphins are caught at night in the DWF Loligo fishery are not readily apparent, it does seem that the day/night differences in capture are also related to a behavioral phenomenon of the dol- phins, and not simply fishing practices. Conversely, the take of pilot whales was significant- ly greater during daylight hours. Pilot whales were observed on numerous occasions in active pursuit and opportunistically feeding in and around the mouth of the net during haulback operations. Similar observa- tions have been reported for the Pacific bottlenose dolphin and other species in Leatherwood (1975). The large size and widespread opening of pelagic mackerel trawls might serve to corral larger delphinid species such as pilot whales. Current data do not allow con- elusive determination of whether the high incidence of pilot whale mortality during daylight hours is behavi- oral or the result of fishing practices by the Nether- lands, GDR, and Poland. Food habits Common dolphins have been reported to feed on a wide variety of epipelagic and mesopelagic schooling fin- fishes and squids (Collett et al. 1981, Fiscus 1982, Fiscus and Niggol 1965, Fitch and Brownell 1968, Jones 1981, Nishiwaki 1972, Norris and Prescott 1961, Major 1986). Prey items collected from stomachs of common dolphins captured in the DWF indicate that Loligo and Atlantic mackerel are important prey items for dolphins during midwinter in the shelf waters of the mid-Atlantic. Information on the dietary habits of pilot whales is limited, but they are considered teuthophagous, feeding principally on squid with fish as an alternative. Atlan- tic cod and Greenland turbot Reinhardtius hippoglos- soides. which were taken off Newfoundland by over- wintering pilot whales when squid were not available, are the only finfish prey items reported from the north- west Atlantic (Sergeant 1962, Mercer 1975). We have not examined any stomachs of pilot whales taken in the DWF squid fisheries. However, it seems likely, based on qualitative examination of the relative abundance and co-occurrence of pilot whales and Loligo and known preferences for squid from the literature. 358 Fishery Bulletin 88(2), 1990 that Loligo is a principal prey item of pilot whales in the shelf waters of the mid-Atlantic during late winter and early spring. It also seems apparent that Atlantic mackerel could be considered an important prey of pilot whales in the mid-Atlantic. This conclusion is based on the observations of pilot whales feeding around and in the opening of the mackerel trawls, the occurrence of mackerel in the stomachs of two pilot whales taken in the Atlantic mackerel fishery, and the high incidence of mortality in the Atlantic mackerel fishery. It is possi- ble, however, that feeding on Atlantic mackerel is an opportimistic phenomenon related to fisheries only, and that pilot whales do not otherwise prey significantly on mackerel. Sex and maturity of common dolphins and pilot whales taken in foreign fishing operations It is likely that most of the common dolphins killed in the Loligo and Atlantic mackerel fisheries were sex- ually immature. Estimates of total length at sexual maturity for male common dolphins are extremely variable between populations (Hui 1979, Perrin and Reilly 1984, Collett and Saint-Girons 1984). Collett and Saint-Girons (1984) found that, in the northeast Atlan- tic, sexual maturity in male common dolphins is reached at 200 cm, with animals <190 cm prepubescent. Sev- eral species of dolphins, including common dolphins, are known to travel in herds segregated by age and sex. In the northwest Atlantic, segregation by age and sex has been found in bottlenose dolphins (Irvine et al. 1981, Shane et al. 1986) and the harbor porpoise (Gaskin 1982). Irvine et al. (1981) found that subadult males formed bachelor groups, and sexually mature adult males rarely mixed with subadult males. Also, male and subadult female dolphins may follow fishing nets more often than females with calves. The dolphins reported in Table 6 were not captured simultaneously (i.e., having come from one group or pod) but rather were captured on several separate occasions. Although female common dolphins with calves have been ob- served in the areas of foreign fishing activity, none have been captured during trawling operations. The size/sex composition of pilot whales killed in the Atlantic mackerel fishery indicates that most of the animals were sexually mature females. The size and sex ratios of pilot whales killed in the Atlantic mackerel fishery are similar to those ratios observed from pilot whale mass strandings in New England waters (Greg Early, New England Aquarium, Boston, MA, pers. commun.. May 1988). Pilot whales are gregarious and social groupings are believed to be comprised of several sexually mature females and a few mature males (Mar- tin et al. 1987). Incidental take of common dolphins and pilot whales relative to available abundance estimates for each species Estimates of the total number of common dolphins and pilot whales are available both for the shelf-edge alone and the combined all-shelf and shelf-edge waters of the northeastern United States (an area which is contained within the EEZ) by season (CETAP 1982). These esti- mates are based on standardized aerial surveys con- ducted during November 1978-January 1982, using line-transect methodology. Althf)ugh these estimates are not for the same time period as the majority of the incidental take described in this paper, and these esti- mates have very high degrees of uncertainty (e.g., stan- dard deviations of the estimates equal to the estimates themselves), they are the only population estimates available which can provide an indication of the relative magnitude of these incidental takes. The shelf-edge abundance estimate for common dolphins during winter (most incidental takes for this species occurred during Deceml)er and February) is 15,703 (CV 0.78) (CETAP 1982, table 17, p. 261).'The average take per year for 1977-88 was 17 and 39 in 1984-88. These takes represent 0.11% and 0.25% of the winter abundance estimate. The maximum annual rate of take based on this winter abundance estimate was 0.48 (« = 76) in 1986 and was nearly twice the 1984-88 average. Similar CETAP shelf-edge estimates for pilot whales during spring and summer are 6823 (C V 0.52) (spring) and 5251 (CV 079) (summer) (CETAP 1982, table 15, p. 233). The average take per year for 1977-88 was 25 and 46 in 1984-88. These takes represent 0.37% and 0.67% (spring) and 0.48% and 0.88% (summer) of the average seasonal abundance estimates for this species. The maximum annual rate of take based on the spring and summer abundance estimates, respectively, were 2.01% and 2.70% (n = 142) in 1988 and were three times the 1984-88 spring and summer rates of take. The shelf-wide abundance estimates for common dolphins 31,124 (CV 0.59) (winter), and pilot whales 1 1 ,4 17 (C V 0.37) (spring) and 9808 (C V 0.66) (summer) are nearly twice the shelf-edge estimates, which re- duces the above estimated rates of incidental take by nearly 50% if animals from the shelf-edge and shelf are considered as a single population. To place pilot whale takes within the EEZ in perspec- tive, present pilot whale incidental kill levels are com- parable with recent mass strandings in New England waters (Greg Early, New England Aquarium, Boston, MA, pers. commun., May 1988). They do not approach the mortalities reported by Mercer (1975) for the historical Newfoundland drive fisheries (1948-71: Total 54,248; average 2260/yr). The impact of trophic Waring et al : Incidental take of marine mammals off northieast United States 359 changes in the offshore cetacean communities, and the unknown DWF incidental mortality levels on compo- nents of the dolphin stocks when the DWF is fishing outside the EEZ, may be more of a factor in structur- ing present dolphin population trends than the levels of incidental takes that occur within the EEZ. Also, we do not know the historical rates of take dur- ing the period 1960-76, when over 100 fishing vessels operated inside the EEZ per year. Since these fleets were not restricted by time or space, it is not unrea- sonable to assume that the number of dolphins taken pre-MFCMA were equal to or greater than the present. However, extrapolation of current levels of take to historical DWF fishing effort off the U.S. east coast (beyond speculation) would not be appropriate for several reasons. These include differences in the spatial/seasonal components of the fisheries over time and changes in the principal targeted species that have occurred during the past several decades. Likewise, the marine mammal/finfish/squid associations might have, likewise, changed during this time period. Also, tech- nological changes in fishing gear and vessels have im- proved greatly during the past two decades, which might have increased the rate of take in recent years. Changes in the nationalities, and their associated fishing techniques, represented in the DWF have also occurred and could effect the rate of take of mammal species. Acknowledgments We thank Doug Beach, Steve Murawski, Tom Pola- check, Tim Smith, and two anonymous referees for reviewing this manuscript. We acknowledge and thank the foreign fishery obsei'vers for collecting the inciden- tal take data, and their dedication to obtaining bio- logical samples. Citations Anderson, E.D., and A.J. Paciorkowski 1980 A review of the northwest Atlantic mackerel fishery. Rapp. P.-V Reun. Cons. Int. Explor. Mer 177:175-211. Bonner, W.N. 1982 Seals and man, a study of interactions. Univ. Wash. Press, Seattle, 170 p. Brown, B.E., and R.G. Halliday 1983 Fisheries resources of the Northwest Atlantic — some responses to extreme fishing perturbations. In Por. E. (ed.), Proceedings of the expert consultation to examine changes in abundance and species composition of neritic fish resources. San Jose. Costa Rica, 18-29 April 1983. FAO Fish. Rep. 291(2):96-109. CETAP 1982 A characterization of marine mammals and turtles in the mid- and North Atlantic areas of the U.S. outer continental shelf. Final Rep. AA551-CT8-48 of Cetacean and Turtle Assessment Prog. (CETAP) to Bur. Land Manage., U.S. Dep. Int.. Wash. DC, 538 p. Collett, A., and H. Saint-Girons 1984 Preliminary study of the male reproductive cycle in com- mon dolphins. Delphinus delphis, in the eastern North Atlan- tic. In Perrin, W.G. (ed.), Reproduction in whales, dolphins, and porpoises, p. 355-360. Rep. Int. Whaling Comm., Spec. Iss. 6. Collett, A., M.H. DuBuit, and R. Duguy 1981 Regime alimentaire de Delphinus delphis dans le nord- est Atlantique. Int. Counc. Explor. Sea CM. 1981/N:5, 4 p. [in French, Engl, abstr.]. Fiscus, C.H. 1982 Predation by marine mammals on squids of the eastern North Pacific Ocean and the Bering Sea. Mar. Fish. Rev. 44(2):1-10. Fiscus. C.H., and K. Niggol 1965 Observations of cetaceans off California, Oregon and Washington. U.S. Fish. Wildl. Serv. Spec. Sci. Rep. Fish. 490. 27 p. Fitch. J.E.. and R.L. Brownell 1968 Fish otoliths in cetacean stomachs and their importance in determining feeding habits. J. Fish. Res. Board Can. 25:2561-2574. Fowler. C.W. 1982 Interactions of northern fur seals and commercial fisheries. Trans. 47th N. Am. Wildl. Nat. Resour. Conf., p. 278-292. Gaskin, D.E. 1982 The ecology of whales and dolphins. Heinemann Educ. Books, London, 459 p. Gilbert, J.R., and K.M. Wynne 1985 Harbor seal populations and fisheries interactions with marine mammals in New England. 1984. Interim, rep. NOAA NA-84-EAC-00070, Northeast Fish. Cent., Natl. Mar. Fish. Serv., NOAA, Woods Hole, MA 02.543, 15 p. Hain, J.H., R.K. Edel, H.E. Hays, S.K. Katona. and J.D. Roanowicz 1981 General distribution of cetaceans in the continental shelf waters of the northeastern LInited States. In A characteriza- tion of marine mammals and turtles in the mid- and North Atlantic areas of the U.S. outer continental shelf. Annu. rep. 1979, Univ. Rhode Island [Avail, as PB83-215855, Natl. Tech. Inf. Serv., Springfield, VA 22161]. Hollander, M., and D.A. Wolfe 1973 Nonparametric statistical methods. John Wiley, NY, 503 p. Hui, C.A. 1979 Correlates of maturity in the common dolphin, Delphinus delphis. Fish. Bull., U.S. 77:295-300. Irvine. A.B., M.D. Scott, R.S. Wells, and J.H. Kaufmann 1981 Movement and activities of the Atlantic bottlenose dolphin, Tursiops truncatus, near Sarasota, Florida. Fish. Bull., U.S. 79:671-688. Jones, R.E. 1981 Food habits of smaller marine mammals from northern California. Proc. Calif. Acad. Sci. 42:409-433. Lange, A.M.T. 1980 The biology and population dynamics of the squids, LoHgo pealei (Le'&eur) and Illex illecebrosus (LeSeur), from the north- west Atlantic, M.S. thesis, Univ. Wash., Seattle, 178 p. 360 Fishery Bulletin 88(2). 1990 Lange, A.M.T., and K.L. Johnson 1981 Dorsal mantle leng^th — Total weight relationships of squids Loligo pealei and Ilkx lUecebrostis from the Atlantic coast of the United States. NOAA Tech. Rep. NMFS SSRF-745, 17 p. Lange, A.M.T., and M.P. Sissenwine 1980 Biological considerations relevant to the management of squid Loligo pealei and Ilkx illecebrosits of the Northwest Atlantic. Mar. Fish. Rev. 42(7-8):23-28. Leatherwood, S. 1975 Some observations of feeding behavior of bottle-nosed dolphins (Tursiops truncatus) in the northern Gulf of Mexico and {Tursiops cf T. gill.i) off southern California, Baja Califor- nia, and Nayarit, Mexico. Mar. Fish. Rev. 37(9):10-16. Leatherwood. S., K.K. Caldwell, and H.E. Winn 1976 Whales, dolphins, and porpoises of the western North Atlantic. A guide to their identification. NOAA Tech. Rep. NMFS Circular 396, Natl. Oceanic Atmos. Adm., Natl. Mar. Fish. Serv.. 176 p. Loughlin, T.R., and R. Nelson, Jr. 1986 Incidental mortality of northern sea lions in the Shelikof Strait. Alaska. Mar. Mammal Sci. 1:14-33. Loughlin, T.R.. R.L. Consiglieri, R.L. DeLong, and A.T. Actor 1983 Incidental catch of marine mammals by foreign fishing vessels, 1978-1981. Mar. Fish. Rev. 45 (7-9):44-49. Lowry, L.F. 1982 Documentation and assessment of marine mammals- fishery interactions in the Bering Sea. Trans. 47th N. Am. Wildl. Nat. Resour. Conf, p. 300-311. Major. P.F. 1986 Notes on a predator-prey interaction between common dolphins {Delphinus delphis) and short-finned squid (Illex illecebrosus) in Lydonia submarine Canyon, western North Atlantic Ocean. J. Mammal. 67:769-770. Martin, A.R., P. Reynolds, and M.G. Richardson 1897 Aspects of the biology of pilot whales {Globicephala mAaeyia) in recent mass strandings on the British coast. J. Zool. (Lond.) 211:11-23. Mate. B.R. 1980 Workshop on marine mammal-fisheries interactions in the northeastern Pacific. U.S. Mar. Mamm. Comm., Wash. DC, 48 p. [Avail, as PB80-159536, Natl. Tech. Inf. Ser., Springfield, VA 22161]. Mercer. M.C. 1975 Modified Leslie-DeLury population models of the long- finned pilot whale {Globicephala melaena) and annual produc- tion of the short-finned squid {Ilkx illecebrosus) based upon their interaction at Newfoundland. J. Fish. Res. Board Can. 32:1145-11.54. Nishiwaki, M. 1972 General biology. In Ridgway, the sea: Biology and medicine, p. Springfield, IL. Norris, K.S., and J.H. Prescott 1961 Obser\'ations on Pacific cetaceans of California and Mex- ican waters. Univ. Calif Publ. Zool. 63:291-402. Payne. P.M.. L.A. Selzer, and A.R. Knowlton 1984 Distribution and density of cetaceans, marine turtles and seabirds in the shelfwaters of the northeastern United States, June 1980-December 1983, based on shipboard observations. Final Rep. NA-81-FA-C00023, Northeast Fish. Cent., Natl. Mar. Fish. Serv., NOAA, Woods Hole, MA 02543, 246 p. S.H., (ed.). Mammals of .3-214. C.H. Thomas, Perrin, W.F.. and S.B. Reilly 1984 Reproductive parameters of dolphins and small whales of the family Delphinidae. hi Perrin, W.G. (ed.). Reproduc- tion in whales, dolphins, and porpoises, p. 97-133. Rep. Int. Whaling Comm. Spec. Iss. 6. Sergeant, D.E. 1958 Dolphins in Newfoundland waters. Can. Field-Nat. 72:156-159. 1962 The biology of the pilot or pothead whale Globicephala melaena (Traill) in Newfoundland Waters. Bull. Fish. Res. Bo.ard Can. 132, 84 p. Sergeant, D.E.. and H.D. Fisher 1957 The smaller cetacea of eastern Canadian waters. J. Fish. Res. Board Can. 14:83-115. Shane. S.H., R.S. Wells, and B. Wursig 1986 Ecology, behavior and social organization of the bottle- nose dolphin: A review. Mar. Mammal Sci. 2(l):34-63. Squires. H.J. 1957 Squid, Ilkx ilkcebrosiis (LeSueur) in the Newfoundland fishing area. J. Fish. Res. Board Can. 14:693-728. Summers, W.C. 1967 Winter distribution of Loligo pealei determined by ex- ploratory trawling. Biol. Bull. (Woods Hole) 133:489-501. Zar. J.H. 1974 Biostatistical analysis. Prentice-Hall, Englewood Cliffs, NJ, 620 p. Abstract. — A study was made of age, growth, and reproduction of the king mackerel Scomberomoru^ cavalla in Trinidad waters captured by hook-and-line and drift gillnets. Ages, estimated from otolith rings, ranged from 0 to VII in males and from 0 to X in females. Mean back- calculated lengths in 99 males ranged from 43.7 cm FL at the end of year I to 85.9 cm FL at the end of year VI, and in 233 females from 41.8 cm FL at the end of year I to 105.6 cm FL at the end of year VIII. The von Bertalanffy growth equations were for males, L„ = 112.3 [1 - exp(-0.18 (t + 1.80))] and for females L„ = 140.1 [1 - exp(-0.15(( + 1.52))]. Based on gonad examination of 97 males and 224 females, spawning takes place throughout the year around the island, with peak spawning from October through March, a period associated with low salinity. First spawning takes place at ages I-II for both sexes. Females predominated in all size groups, with the proportion of males increasing during the peak spawning Age, Growth, and Reproduction of the King Maclcerel Scomberomorus cavalla (Cuvier) in Trinidad waters Maxwell G. de L. Sturm Premila Salter Institute of Marine Affairs, P O Box 3160 Carenage Post Office, Trinidad. West Indies Manuscript accepted 1 1 December 1989. Fishery Bulletin, U.S. 88:361-370. The king mackerel Scomberomorus cavalla (Cuvier) (Scombridae) occurs in coastal waters of the western At- lantic ranging from the Gulf of Maine to Rio de Janiero (Briggs 1958). It has also been reported throughout most of the West Indies (Erdman 1949), and its distribution is said to extend eastwards to Africa (Jordan et al. 1930). However, more recent reviews (Collette and Russo 1979) make no mention of S. cavalla in the Eastern Atlantic, although it has been reported from St. Paul's Rocks, a group of small islands in the mid- Atlantic (Lubbock and Edwards 1981). Scombermyvonis cavalla has commer- cial and recreational value through- out its range. Griffiths (1971) consid- ered Scomberomorus spp., especially S. cavalla and S. maculatus ( = S. brasiliensis Collette et al. 1978) to have greater commercial potential than the timas in Venezuela. The gen- eral biology of S. cavalla has been in- vestigated in North American waters (Beaumariage 1973) and off the state of Ceara, northeastern Brazil, includ- ing age and growth studies (Nomura and Rodrigues 1967, Ximenes et al. 1978). Manooch et al. (1978) provided a useful annotated bibliography on this species, and many aspects of its biology and fishery have been sum- marized by Collette and Russo (1984). More recent work includes studies on reproduction (Finucane et al. 1986) and growth (Manooch et al. 1987) in North American waters. In Trinidad, Whiteleather and Brown (1945) and more recently Sturm et al. (1984) reported that S. cavalla. ap- proached S. maculatus ( = S. brasili- ensis) as the most abundant continen- tal pelagic species in commercial land- ings. The Scomberomorus fishery in Trinidad is artisanal and seasonal from March through October, and S. cavalla is caught mainly by hook-and- line (trolling and live-bait fishing) and drift gillnets (Sturm et al. 1984). This paper provides information on age, growth, and reproduction of S. caval- la (the kingfish or thazard) in Trini- dad waters. Materials and methods The climate of Trinidad is tropical, with a wet season from June to De- cember. The surrounding waters are mainly estuarine. Surface tempera- tures vary little, ranging from 27 to 29°C (van Andel and Postma 1954, Edwards 1983). Further details of the study area may be found in Sturm (1978). Fish were bought from artisanal fishermen in 1981-82 at six beaches: Maracas and Las Cuevas in the north, Mayaro and Guayaguayare in the southeast, and Cedros and Icacos in the southwest of the island. Artisan- al catches consisted mainly of both Scomberomorus species, which on be- ing landed were separated into bask- ets. There were usually not more than about a dozen boats fishing on any 361 362 Fishery Bulletin 88(2), 1990 one day from a beach, and catch per boat of S. cavalla was usually small (<20 fish). Not more than three catches were sampled up to a maximum of about 25 fish per sampling day, to enable accuracy of sample treatment in the field. Entire catches were sampled, and when this was not possible, large fish (>90 cm FL) and/or small fish (<50 cm FL), when present, were selected to direct attention to the ends of the distribu- tion, and the rest of the sample was taken at random. A total of 363 fish was obtained, 190 from hook-and- line, 165 from gillnets (average stretched mesh size 113 mm), and eight from beach seines. In addition, 264 fish were measured at the Port of Spain Fish Market. Fish were measured to the nearest 0.5 cm fork length (FL) and weighed to the nearest 20 g. Gonads were weighed to the nearest 10 g. Otoliths (sagittae) were used to age 5. cavalla. When viewed in a black dish containing water in reflected light under a binocular microscope (x 10), the otoliths revealed opaque and translucent rings. The non-mar- ginal opaque rings were counted as annuli following Beaumariage (1973), Ximenes et al. (1978), Johnson et al. (1983), and Manooch et al. (1987), and distances were measured along the longitudinal axis from the focus to the distal edge of each opaque ring and to the edge of the otolith. One person read the otoliths. The relationship between fork length and otolith radius was determined by regression analysis using both linear and quadratic models. When the relation- ship was established, fork length at age was back- calculated using methods of Bagenal and Tesch (1978). The von Bertalanffy growth equation (1938)— L, = L^ [l-exY)(-K{t + to))] where L, = fork length at age t, L^ = asymptotic fork length, K = growth co- efficient, and tfi = time when length is theoretically zero— was fitted to weighted mean back-calculated lengths using an MS-DOS/BASIC version of a compu- ter program by Sparre (1987). Macroscopic stages of gonad development for both sexes were identified as described for S. maculatus (=S. brasiliensis) by Sturm (1978). They were imma- ture-inactive, immature-developing, ripe, and ripe- running. The gonadosomatic index (GSI) was computed by dividing gonad weight by total fish weight and multiplying by 100. Results Size composition of material examined Length frequencies oi Scomberomorus cavalla samples from gillnets and hook-and-line were approximately similar. Hook-and-line apparently selected slightly 40 O GILL NET /V- 165 • HOOK ANO-LINE A/- 190 35 10 / •''t \ 25 V'\ 20 \ \ \ \ IS i 1 / ^ 10 / / 1 1 \ \^ s \i V ^'^''^^^ / 1 . J. .^. . . .X /C:^^. 30 40 50 60 70 80 90 100 110 Fork Length Figure I Length frequencies of gill net and hook-and-line samples of Scom- beromontft cavalla from Trinidad waters. Lengths are grouped into 5-cm intervals. N = number of fish sampled. larger fish— 45-114.5 cm FL, with a modal length be- tween 66 and 70 cm FL— than gillnets, which selected fish between 38.5 and 105 cm FL, with a modal length of between 61 and 65 cm FL (Fig. 1). The largest fish sampled was a 127-cm FL female caught in a beach seine. The size range of fish measured at the Port of Spain Fish Market was 37-135 cm FL with a modal length of 61-65 cm FL. Age and growth Only 2 of 341 otoliths were considered illegible. How- ever, not all of the remaining 339 otoliths had clear rings, though 85% were read with confidence. The main difficulty was determining the extent of the first opaque ring. The remaining 15% (52) were reread, with 79% (41) agreement between the two readings. The re- maining 11 otoliths were then given new readings. For tracing the frequency of opaque-edged otoliths through- out the year, the data were grouped in bimonthly intervals due to the scarcity of data in non-seasonal months (Fig. 2). Otoliths with opaque edges were found throughout the year. The minimum percentage (20%) occurred in August-September after which there was a large increase to maximum percentages (60%, 58%, and 57%) from October through March, followed by 36% and 35% in April-May and June-July, respective- ly. These results suggest that opaque rings are formed Sturm and Salter: Scomberomorus cavalts m Trinidad waters 363 so 50 l~~ ■ \ 10 30 / ' \ 20 . J ', H- 51 55 38 , 51 62 76 as ON DJ FM AM JJ Months Figure 2 Bimonthly percentages of otoliths of Sromheromorus cavalla with opaque edges. A'^ = number of fish sampled. annually, mainly from October through March. Time of ring formation is similar to that recorded for Brazil- ian S. cavalla (Ximenes et al. 1978). The relationship between fork length (FL) and oto- lith radius (OR) was represented by the following linear equations: Males: FL = - 1.73 -f 1.490R (r = 0.86) Females: FL = - 11.64 -i- 1.760R (r = 0.90) The intercepts were significantly different at the P<0.001 level {F = 14.76, df = 1, 336); therefore, the data were treated separately. Second-degree quadratic models were fitted to the separate data, but analysis of variance did not show significant curvature for males (F = 3.03, df = 1, 98, P>0.05) nor females (F= 1.65, df = 1, 235, P>0.1), therefore the linear equations were used to back-calculate length from age for males and females separately. Observed and back-calculated lengths for 99 male and 233 female S. cavalla are shown in Table 1 . Five age-0 fish (38.5-57.5 cm FL) were also recorded. The oldest male was age VII and the oldest female age X, but these were not included in the back-calculations. Length variation within age groups was large, as was the case for southeastern United States populations (Beaumariage 1973, Johnson et al. 1983). For exam- ple, age-II females ranged from 43.0 to 85.5 cm FL. From age II onwards, females grow faster than males. Figure 3 compares back-calculated growth of S. cavalla from Trinidad with that from southeastern United Table 1 Mean observed and back-calculated lengths cm) for Scomberon orus caval a from Trinidad waters. Mean back-calculated length at age Mean observed Age group N length Range 1 II III IV V VI VII VIII Males I 23 .57.3 44.0-72,5 44,9 II 34 64.5 53.5-75.0 42.5 57,8 III 25 71.4 62,5-78.0 44,6 58,4 66,9 IV 9 72.2 65,.5-81.0 41,5 54,5 61,8 68.3 V 6 81.8 77,5-86.5 44,2 58,1 67,5 73.6 78.8 VI 2 87.5 80.5-94.5 46,4 58,5 68.9 74.6 81.0 85.9 Weighted mean 43.7 57,6 66.0 70.9 79.3 85.9 N 99 76 42 17 8 2 SD 5,0 5,3 5.2 5.5 4.3 9.8 Annual increment 13,9 8.4 4.9 8,4 6.6 Females 1 46 58.2 42,0-70.5 43,7 II 68 67.4 43.0-83,5 42,3 59.7 III 48 71.8 62,0-93,5 39,8 57.1 66.7 IV 33 80.2 65.0-101.0 41.1 57.6 67.4 75.1 V 19 87.2 77.5-103,0 42,2 58,0 68.7 76.7 83,8 VI 7 95.8 92.0-101.5 43,2 60,1 70.5 78.6 85,7 91.5 VII 7 102.9 92,5-109,0 41,6 62,2 73.4 81,5 87,5 92.7 99.2 VIII 5 109.7 101,0-127,0 39,9 59,8 70.0 79.6 88,4 95.8 101,2 105,6 Weighted mean 41,8 58,6 68.0 76.8 85,1 93.1 100,1 105.6 N 233 187 119 71 38 19 12 5 SD 5.7 6,2 7,0 6.8 6.2 5.4 1,9 9.9 Annual increment 16,8 9.4 8.8 8.3 8.7 7,0 5.5 364 Fishery Bulletin 88|2). 1990 VII VIII IX Age (years ) XII XIII XIV Figure 3 Growth (mean back-calculated fork lengths) of Scomberomorus canal- la (a) males and (b) females. 1 = Northeastern Brazil (Ximenes et al. 1978); 2 = Northeastern Brazil (Nomura and Rodrigues 1967), ages III and above observed lengths; 3 = Trinidad, this study; 4 = U.S. Gulf of Mexico (Manooch et al. 1987); 5 = Southeastern U.S., excl. La. (Johnson et al. 1983); 6 = Florida (Beaumariage 1973), fork length transformed from standard length by his formula FL = 1.096SL - 17.143; 7 = Southeastern U.S. incl. La. (Johnson et al. 1983). States and northeastern Brazil. For both sexes, there is wide separation of lengths at age I, with those of Trinidad and North American fish being larger than those of Brazilian fish. The largest male lengths are from the Florida sample (Beaumariage 1973) up to age IV, and the smallest are for fish sampled in Brazil by Nomura and Rodrigues (1967) up to age VI. For fe- males, the largest lengths are again for Beaumariage's sample except for an anomalous population of large females sampled in Louisiana (Johnson et al. 1983), and the smallest lengths are similarly for Nomura and Rodrigues' Brazilian sample. Back-calculated lengths at age of Trinidad fish lie between the larger lengths of North American fish and the smaller lengths of Brazilian fish, up to age III for males (82.8% of the sam- ple) and age V for females (91.9% of the sample). The growth curve of Trinidad males closely resembles those from the southeastern United States presented by Johnson et al. (1983) and Manooch et al. (1987) up to age V, whereas the growth of Trinidad females is more simOar to that of BrazOian females sampled by Ximenes et al. (1978). Growth rates of males from the different areas are similar up to age II, with Brazilian males showing less incremental decrease in the later ages. The same trend is shown for females, with marked in- cremental decrease being shown by southeastern United States (excluding Louisiana) females sampled by Johnson et al. (1983). The von Bertalanffy growth parameters of Trinidad S. cavalla are presented in Table 2 along with those from southeastern United States and Brazil. Asymp- totic lengths are larger for females than males. The asymptotic lengths of Trinidad and Brazil males are close, and much larger than those from North America, except for the recent study from the Gulf of Mexico (Manooch et al. 1987). For females, the pattern is some- what similar, with the Trinidad and Gulf of Mexico values being the largest, except for the additional sam- ple of large females from Louisiana (Johnson et al. 1983). For both sexes the K values are fairly close, the lowest coming from Trinidad, Brazil, and the Gulf of Mexico, with again the exception of the Louisiana females. Size and age at first maturity There were 14 ripe males (11 ripe -i- three ripe-running) and three ripe females in the age-I group, indicating first spawning for both sexes to take place in this age group. The shortest ripe male was 54.5 cm, and the largest immature developing male was 80 cm. The shortest ripe female was 58.5 cm, and the largest im- mature developing female was 114.5 cm. These sizes at maturity indicate that first spawning for both sexes may also take place at age II (Table 1). Gonad analysis and GSI Gonad stages and corresponding GSI values are shown in Table 3. No ripe-running females nor spent fish were observed in this study. Figures 4 and 5 show gonad analysis of the 97 males and 224 females that were capable of spawning as indicated by size at first matur- ity. Of these, 70 males and 164 females were analyzed for GSI, and the results are also shown in Figures 4 and 5. Male gonad analysis was more difficult because milt was present in the early stages of gonad develop- ment; this explained the trend to higher percentages of ripe males. Ripe fish were present during all months Sturm and Salter: Scomberomorus cavalla in Trinidad waters 365 Table 2 von Bertalanffy growth parameters for Scomberomorus cavalla from Trinidad, southeastern United States, and Brazil. L K ',1 Area and source (FL cm) (yr) Males Trinidad (thiis study) 112.3 0.180 -1.79 Nortiieastern Brazil (Nomura and Rodrigues 1967) 116.0 0.180 -0.22 Nortiieastern Brazil (Ximenes et al. 1978) 113.3 0.229 -1.50* Florida (Beaumariage 1973) 90.3" 0.350 -2.50 Southeastern U.S. (Johnson et al. 1983) 96.5 0.280 -1.17 U.S. Gulf of Mexico (Manooch et al. 1987) 111.3 0.208 -1.48 Females Trinidad (this study) 140.1 0.150 -1.52 Northeastern Brazil (Nomura and Rodrigues 1967) 137.0 0.150 -0.13 Northeastern Brazil (Ximenes et al. 1978) 131.7 0.164 -2.00* P'lorida (Beaumariage 1973) 124.3" 0.210 -2.40 Southestern U.S., excl. La. (Johnson et al. 1983) 106.7 0.290 -0.97 Southeastern U.S., La. (Johnson et al. 1983) 152.9 0.140 -2.08 U.S. Gulf of Mexico (Manooch et al. 1987) ^ 141.7 0.136 -1.98 'Assumed negative. "Calculated from Beaumariage's formula FL = 1.096SL- - 17.143. Table 3 Gonad stages and gonadosomatic indices of Scomberomorus cavalla from Trinidad waters. Gonad stage Males Females N Mean GSI Range GSI N Mean GSI Range GSI Immature inactive Immature developing Ripe Ripe running 2 21 18 29 0.77 0.38 0.92 1.11 0.32-1.37 0.14-0.78 0.30-1.92 0.33-3.73 10 113 41 0.58 0.63 1.75 0.41-1.16 0.16-1.37 0.41-5.68 except December (when only nine fish were sampled), and ripe females were absent in April. Maximum per- centages of females occurred from September to Octo- ber and January to March. GSIs for females peaked in September, November, January, and February. Males showed similar patterns, though data were more limited. The above observations indicated that S. cavalla spawns throughout the year with more intense spawning from September through March. Sex ratio Females were dominant in all size groups, both in gill- net and hook-and-line samples, and in the peak spawn- ing and non-peak spawning seasons (Table 4). The degree of dominance varied with size class, type of gear, and season. Dominance increased with size class, and all but a single specimen over 90 cm FL was female. Females predominated to a greater extent in the hook-and-line samples compared with gillnet sam- ples. In gillnet samples, the proportion of females decreased from 74% during the non-peak spawning season to 62% during the peak spawning season, which corresponded with an overall decrease in females from 77% to 63% in the respective seasons. In the south- eastern United States, the sex ratio was also found to favor females (Fischer 1980, Trent et al. 1983). Discussion Evidence for time of annulus formation in Trinidad Scomberomorus cavalla is weak. Tracing of otolith edges throughout the year suffered from a paucity of 366 Fishery Bulletin 88(2). 1990 20 - (0) 15 to o - [ A A c o g 10 / A A 05 - ^ V ' \ 0 *- ,',',•',«,',',',', ^' ,',",'» , 100 -(b)^^ 1 1 90 80 70 ^ 60 - 5? 50 - 40 - 30 20 10 - - 0 *- , , 5 , 15 , « , " , 5 , •■ 31 , 4 1 13 , 10 MJJ aSONOJFMA Months 20 - (o) T f \ _ 15 tn - A ; I y- \ o 6 10 - Y V V 05 - ^^i~~54.5 cm) by (a) mean monthly GSI and (b) monthly percentages of ripe (ripe and ripe-running) gonads. N = number of fish sampled. Figure 5 Seasonal maturation cycle of female Scmnheromortis cavalla (>58.5 cm) by (a) mean monthly GSI and (b) nicmthly percentages of ripe gonads. N = number of fish sampled. data when the annulus is apparently formed as fish were scarce from November to February. Maximum (bimonthly) percentages of opaque otolith margins compare well with those (monthly) of North American studies (Beaumariage 1973, Johnson et al. 1983, Manooch et al. 1987). However, in the present study the minimum (bimonthly) percentage (20%) was much higher than in North America where in all the studies there were one or more months wath no edged otoliths. In Trinidad the period September to March, when maximum percentages of opaque otolith margins are found, corresponds with the period of greatest spawn- ing intensity. In Brazil, the opaque ring is laid down from November to March corresponding to a period of intense spawning (Ximenes et al. 1978). Otoliths of North American fish generally had the largest percent- ages of opaque margins from April through Jiuie (Beau- mariage 1973, Johnson et al. 1983, Manooch et al. 1987) at the start of a well-defined spawning season (Beau- mariage 1973, Finucane et al. 1986). The process of spawning may therefore be related to annulus forma- tion in S. cavalla. If first spawning takes place at age II, as might occur on occasion for Trinidad S. cavalla, an annulus in immature (age-I) fish may be explained by some internal physiological rhythm, together with the environmental changes that trigger spawning, causing annulus formation. This explanation was put forward for annulus formation in immature fish for hake Merluccius meriuccius (Hickling 1935) and red snapper Lutjanus campechanus (Nelson and Manooch 1982). As spawning is year-round in the tropical waters around Trinidad, the finding of relatively large num- bers of opaque margined otoliths in the months out- side the period of their maximum occurrence during peak spawning can be expected. These, however, ob- scure the results of tracing the frequency of otolith edges throughout the year compared with the more clear-cut results from the more temperate North American waters. The differences in length at age and theoretical growth parameters between Trinidad, Brazilian, and North American fish may result from different environ- Sturm and Salter: Scomberomorus cavalla in Trinidad waters 367 Table 4 Sex ratio of Scomberomorus cavalla from Trinidad waters Length range Gillnets Hook-and-line Totals Male Female Male Female Male Female (cm) (N) (N) (%) (N) (N) (%) (N) (N) (%) Peak spawning (September-March) 30.0-49.5 1 0 0 0 1 100 50.0-69.5 24 36 60 10 24 71 70.0-89.5 18 12 40 9 29 76 90.0-109.5 0 1 — 1 11 92 110.0-130.5 0 0 — 0 0 — Total 43 49 53 20 65 76 63 114 60 Non- jeak spawning (April-August) 30.0-49.5 1 6 86 0 0 — 50.0-69.5 14 39 74 14 26 65 70.0-89.5 4 7 64 8 42 84 90.0-109.5 0 2 100 0 13 100 110.0-130.5 0 0 — 0 2 100 Total 19 54 74 22 83 79 41 137 77 Totals 62 103 62 42 148 78 ments, feeding habits, exploitation rates, methods of capture, sample sizes, etc. The fishery in North Ameri- ca is better developed than in Brazil (Collette and Russo 1984) and Trinidad, which enabled larger samples for age and growth studies to be collected from the former region. There were considerable differences between sample sizes of older fish in the various studies, which could introduce bias in the comparison of von Berta- lanffy parameters. Best representation of growth of older fish came from Manooch et al. (1987) and the anomalous group of Louisiana females (Johnson et al. 1983). Changes in populations between studies brought about by exploitation and other factors could also in- fluence the results. Different methodologies could also contribute to these differences. Nomura and Rodrigues (1967) counted translucent rings to age Brazilian fish, thus omitting part of a year. This probably resulted in their obtaining the slowest growth of all the studies. Back- calculated lengths were fitted to the von Bertalanffy equation by Ximenes et al. (1978), Johnson et al. (1983), and in this report. Beaumariage (1973) fitted observed standard lengths and advised caution in extrapolation of the theoretical growth curve for size at older ages, because it was derived from Walford plots that ex- cluded older fish; this resulted in conservative esti- mates. Nomura and Rodrigues (1967) used both back- calculated and observed lengths, which probably resulted in large values of L^. Other differences in methodology, e.g., the weighting of mean back-calcu- lated and observed lengths and more exact fitting of the von Bertalanffy curve using computer programs in the later studies, and the non-use of an intercept value for back-calculation in some of the studies, could lead to minor differences in the results. Comparisons of results between the three areas evidently cannot be properly done due to differing sample sizes and methodologies. Estimates of maturity of 58.5 cm FL or age I-II for Trinidad female S. cavalla generally agree well with those of other studies. In Brazil, first maturity was found to take place at 58.6 cm FL (Alves and Tome 1967) and 63.5 cm FL or age III (Gesteria and Mes- quita 1976). Another study showed that females first matured between 43.5 and 56.5 cm FL or age III, with 50% maturity at 77 cm FL or age V-VI (Ivo 1972). Size of maturity of Trinidad females compares well with that in Brazil. However, greater age at maturity of Brazilian fish could be due to overestimated age at length based on the use of translucent rings for age- ing by Nomura and Rodrigues (1967). Finucane et al. (1986) reported size at maturity at 44.9 cm FL in the southeastern United States. In Florida waters, Beau- mariage (1973) round ripe eggs in age-I females (61.4 cm SL = 65.6 cm FL) similar to Trinidad results. How- ever, he concluded that these eggs were aborted or reabsorbed, and spawning did not really take place until age IV (88.0 cm FL). As with females, ripe males 368 Fishery Bulletin 88(2). 1990 first appeared at age I (63.4 cm FL) as they did in Trinidad, but Beaumariage believed that males spawned initially at age III (77.0 cm FL) due to the greater development of spermatogenesis in the testes of older fish. However, in Trinidad initial spawning takes place at age I, as shown by the presence of ripe- running males in this age group. The spawning pattern of S. cavalla in Trinidad is ap- parently similar to that off the coast of Ceara, in north- eastern Brazil. Gesteria and Mesquita (1976) observed year-round spawning with maximum intensity from October through March, as in Trinidad. Spawning throughout the year was also recorded by Ivo (1972), but his period of maximum activity was from January through June. Another study (Menezes 1969) indicated that they spawn during the period October through March. Further north, the spawning season is reversed, where in the northeastern Caribbean it lasts from April through November (Erdman 1976) and in the south- eastern United States, from April through October (Beaumariage 1973, Finucane et al. 1986). Spawning migrations in North America are determined by tem- perature (Moe 1972, Beaumariage 1973). In Trinidad, however, spawning and migration may be influenced by salinity changes because there is little variation in the water temperature. Peak spawning starts after the rains have set in, and may be triggered by a drop in salinity; minimum salinities have been recorded in August and September (van Andel and Postma 1954, Edwards 1983). The presence of ripe-running males in the samples indicates spawning in local waters, but the spawning grounds remain to be discovered. The scarc- ity of the fish during November through February remains unexplained. Spawning on the outer Continen- tal Shelf, 50-60 km offshore, as occurs in the north- western Gulf of Mexico (McEachran et al. 1980) would place part of the population beyond the reach of the artisanal fishing fleet which is limited to some 40 km offshore. Another possible factor could be migration along the Venezuelan coast to the northwest and/or the southeast, although part of the population may be resi- dent throughout the year. Decreased vulnerability to hook-and-line due to decreased feeding activity during spawning, and/or gillnets due to spawning in waters deeper than that in which the gillnet is effective, are other possible reasons to be considered. Beach landing data indicate that a southerly migra- tion takes place in Trinidad during the months when the fish is seasonally abundant (Sturm et al. 1984). In the present study, gonad analysis did not suggest any migratot7 trends since ripe fish were taken around the coast throughout the year. Moreover, observations on stomach contents showed no clear feeding patterns as evidence of migi-ation (unpubl. data). The spawning and abundance patterns of S. cavalla approximately cor- respond to those of S. maculatus ( = S. brasiliensis), which spawns throughout the year with more intense spawning from October through April and is season- ally abundant from May through September (Sturm 1978). S. brasiliensis also moves in a southerly direc- tion during peak abundance, part of likely clockwise movements around the island (Sturm 1978, Sturm et al. 1984). In Florida, Williams and Sutherland (1979) and Sutherland and Fable (1980) have shown that S. cavalla undertakes long-range migrations compared with the shorter range migrations of S. m.aculatus, which is closely related to S. brasiliensis (Collette et al. 1978). In Trinidad, a similar situation would explain the dif- ficulty in recognizing migratory trends of S. cavalla, compared with the more local movements described for S. brasiliensis (Sturm 1978). Female dominance in samples may be the result of more females than males being hatched or mortality being higher in males than females. Higher mortality probably is associated with slower growth in males. Alternatively, female dominance may be a function of gear selection in gillnets or of behavioral differences between the sexes. Increased voracity in females was recorded in Brazil (Menezes 1969) and may explain why the sex ratio favored females in hook-and-line samples. Also, males may inhabit greater depths than females and are less vulnerable to gillnets. If this was the case, a meeting of the sexes for spawning by an upward migration of males and/or a downward migration of females could explain the increase in the male:female ratio observed for gillnets during peak spawning. Acknowledgments Thanks are due to Michele Julien-Flus for assistance in field work, to Sherry Manickchand-Heileman for useful comments and discussion, to Bruce Lauckner of the Caribbean Agricultural Research and Development Institute for advice on statistical matters, to Rodney Ramkissoon for assistance in figure preparation, and to Avril Siung-Chang and Roland Bailey (Kings Col- lege, London University) for reading the manuscript. Citations Alves, M.I.M, and G.S. Tome 1967 Alguns aspectos do desenvolvimento maturativo das gonadas da cavala, Sromheromoru!; ravnlln (Cuvier, 1829). Arq. Estac. Biol. Mar. Univ. Fed. Ceara 7(1): 1-9 [in Portugese. Engl. summ.]. Bagenal, T.B.. and F.W. Tesch 1978 Age and growth. In Bagenal, T.B. (ed.). Methods for assessment of fish production in fresh waters, 3rd ed., p. 101-1.36. Blackwell Sci. Publ., Oxford. Sturm and Salter: Scomberomorus cavalla in Trinidad waters 369 Beaumariage, D.S. 1973 Age, growth and reproduction of king macl 1 10 cm total length (TL) collected by Rusher (1987) in Elkhorn Slough and Monterey Bay and San Francisco Bay showed signs of either pregnancy, recent birth, or embryo abortion. Therefore, we use the assumption of an annual reproductive cycle in later sections dealing with stock replacement by reproduction. The leopard shark is primarily a benthic feeder (Russo 1975, Talent *Drs. M. Moser and J. Sakanari (Long Marine Lab.. Univ. Calif., Santa Cruz, pers. commun., Sept. 1984) report that in a sample of nine pregnant females taken on 10 September 1984, presumably in midterm, the mean embryo length w;is 11.26 cm(1.51 SD), (« = 51 embryos). Smith and Abramson: Leopard shark tag-recovery data from San Francisco Bay 373 1976). Prior to the work described here, nothing was known of its movements or the degree of exchange with other leopard shark populations along the Califor- nia coast. In 1979, a tagging study was initiated in San Fran- cisco Bay to obtain information on age validation, growth, and movements of this species. Tag recaptui'es were monitored over a 9-year period. This report gives results of movements that were deduced from the tem- poral and geographic distribution of tag recoveries. In addition, although beyond the planned design of this study, we decided to utilize the tag recovery data together with published information to estimate the ef- fect of management by size limit on stock replenish- ment and on yield per recruit. The lack of suitable statistical information on catches, as mentioned previ- ously, prevented us from performing analyses that involve weighting tag recoveries by catch or effort. Results of the age validation segment of the study have been published elsewhere (Smith 1984), and results on age and growth are also being published separately (Kusher et al. In prep.). Methods All sharks were tagged off Hunters Point in south San Francisco Bay in 1979. Collections were made with a 183-m longline rigged with an average of 150 baited hooks fished on the bottom at depths of 15-20 m. Prior to release, total and precaudal lengths were recorded to the nearest centimeter, and each fish was given an intraperitoneal injection of oxytetracycline hydrochlor- ide to mark vertebral centra for age verification pur- poses (Smith 1984). A record was made of the sex and general physical condition of each fish; seriously injured animals were not tagged. Those with minor hook in- juries or with partially everted stomachs were classified as "injured"; the rest were classified as "healthy." A plastic rototag of the type recommended by Kato and Carvallo (1967) was applied to the first dorsal fin and the fish released at the capture point. The fin tags were imprinted with a legend informing the recoverer that a reward (amount unspecified) was offered for return of the tag and the fish or a section of its vertebral col- umn for age verification purposes. The legend also provided an address and phone number to contact to arrange delivery. Mortality estimation Fishing mortality rates were estimated from the tag- ging data using the concept described by Murphy (1965), Gulland (1965), and Tomlinson (1970), which is now commonly referred to as virtual population anal- ysis (VPA), though it differs from the original VPA procedure of Fry (1949). The computer program COHORT, written by John Geibel and Phil Law (Calif. Dep. Fish Game, 411 Burgess Drive, Menlo Park, CA 94025) was used to calculate the estimates. The natural mortality estimate was based on Hoenig's (1983) regression equation log (Z) = 1.46-1.01 log (f,„„j:) where Z is the instantaneous annual total mortality coefficient and t,,,,,,. is maximum age attained by the species. If the maximum age was determined from a period when there was virtually no fishing directed at the species, then one could assume the estimated Z approximates the instantaneous annual natural mor- tality coefficient, M. The basic procedure involved assuming values of M over each 1-year time interval, taking a trial value of F„ , the instantaneous annual fishing mortality coeffi- cient, for the ultimate interval, and executing the backward VPA computation on the tag recoveries to obtain an estimate of A^o > the number of tagged fish at the beginning of the first interval. Trial values were then iterated until the series converged on N(t . Before conducting the VPA, it was necessary to con- sider two additional factors which would cause adjust- ments to the actual observations used in the analysis: (1) the likely rate of tag loss and (2) the level of tag recovery nonreporting. Since the tagging experiment was not designed for this type of analysis, there were no built-in procedures to estimate these factors. We therefore used what we judged to be the best available information from outside sources. Yield per recruit Yield per recruit was calculated by piecewise integra- tion of the yield curve. The yield in weight at each age was taken to be the product of the annual rate of exploitation, the midpoint between an individual's weights at the beginning and end of the age interval, and the population size at the beginning of the interval. Y = exp[-M(^. -I- 1)] [1 - exp(-Z)] (F/Z) X X {exp[-Z(t-t,)]}w, (1) t^i. where Y is yield per recruit in weight (kg), t is age, <,. is age at first capture, and w, is the midpoint be- tween the weights at t and t + 1. Weight at age was computed by using pi-edicted values from the von Bertalanffy length equation from Kusher (1987) and the weight-length formula in Smith (1984). Note that equation (1) assumes cbnstant M and Z except that natural mortality is doubled during 374 Fishery Bulletin 88(2). 1990 PACIFIC OCEAN HALF MOON BAY Vi; Figure 1 Map of San Francisco Bay showing location of release area (arrow) and major Bay Area divisions and names referred to in the text. year 1. This gives some weight to a higher mortality these young fish must suffer relative to the adult sharks. The phenomenon of the young being preyed on by larger sharks has been cited by Springer (1960, 1967) and by Holden (1974). Stock replacement Also calculated for various levels of age at first cap- ture (knife-edge selection) and fishing mortality were the percentages of stock which would be replaced due to reproduction. Assuming an even division among the sexes, this was done by summing over ages, from age at first maturity to maximum age, the products of the number of female survivors per recruit and the number of pups produced by females at each age above the age of maturity and multiplying this sum by 100. For given t, and f" /? = X 50exp[-M(^ + l)-F(/-^.)d]P, (2) (>(„ where t,n is female age at maturity, R is stock replace- ment per recruit in percent, P, is estimated number of progeny produced annually by an age t female, and 1 t>tr' 0 t 1-year-old juveniles. It seems reasonable that these very young conch do not migrate over long distances to the cays where we find the large juveniles, but are car- ried during their planktonic life by currents to these locations. These findings suggest that conch may be buried almost continuously until they reach a shell length of ~5 cm. The purpose of this study is to use tag-recapture data to examine the relationship between survival of juve- nile conch and their burying activity. Methods and materials The data used in this study were col- lected in the Berry Islands (Bahamas) about 190 km east of Miami, Florida (Fig. 1). A wide range of conch sizes (3.5-22.0 cm) was tagged during these exTDeriments at all seasons and at several different cays to account for the effects of different habitats, conch sizes, and seasonal variation. The sampling area, consisting of shallow sand flats with abundant turtle grass Thalassia testudinum, is described in more detail by Iver- sen et al. (1987). Individual conch were tagged with thin, numbered plastic tags affixed to the spire vdth underwater epoxy. Shell length was taken using a mea- suring board (Iversen et al. 1987). Shallow intertidal waters were sam- pled by wading and the deeper off- shore water by snorkeling. Counts of tagged conch were made on 23 sampling trips lasting 4-5 days made approximately every 5 weeks be- tween May 1980 and February 1983. A conch was considered to be buried when it was not found on one or more sampling trips, but was found alive on subsequent visits (Table 1). We believe this to be a reasonable assumption, because the aggrega- tions of conch we studied remained in close proximity to the cays and when not buried were easOy located. Conchs as large as 10-15 cm in length which we placed in 25-m2 pens would bury and were not found on some visits despite intensive search- ing. On subsequent visits a month or more later, they would reappear on the surface of the substrate. For example, of 27 tagged conch in one pen, seven were buried on at least one visit between April and Decem- ber 1980. Tagged individuals released at various cays that were not subse- quently found were considered to have fallen prey to predators; large juveniles (^16 cm) were assumed to have migrated offshore to deep water (Iversen et al. 1987). Results Our tagging data show a significant positive correlation (Fig. 2) between survival and percent buried for conch released in the subtidal region of about-2 m depth south of Vigilant Cay (shell length range, SLR, 3.5- 16.5 cm; r = 0.80; P<0.01), and for conch from intertidal zones at Cat Cay (SLR 4.5-18.0 cm; r = 0.56; P<0.05) and Little Cockroach Cay (SLR 8.0-22.0 cm; r = 0.81; P<0.05). Discussion Predation is probably the most im- portant cause of natural mortality on stocks of > 1-year-old juvenile conch. Based on our searches for parasites in juveniles and adult conch, and our field observations, we do not believe that parasites or diseases play an important role in conch survival. There are reports of isolated mass mortalities of conch Manuscript accepted 28 December 1989. Fishery Bulletin, U.S. 88:383-387. 383 384 Fishery Bulletin 88(2), 1990 28' N 27° 26° 25' 240. 23° /Miami ;Bimini BERRY ISLANDS Little Cockroach • Coy Cockrooch Cay V Coy 'S-WhaieCay Channel FRAZERS hog// . '^"'VSU " /P^Coy ^°l^^ ^ CAY /Chonnei ^ Scale Islands [-^ / Nassau. N ANDROS S V. 80°W 79° 78° 77° 76° 75° 74° Figure 1 Location of Strombus gigas study site in the Berry Island (from Iversen et al. 1987). Table 1 Example of database used in this study of queen conch burying behavior. Tagging location: Little Cockroach Cay, Berry Islands. Bahamas. Tagging date: 27 May 1980. Field observations Tag no. 1980 1981 05/27 06/23 08/21 09/05 10/03 11/01 12/16 01/23 03/03 0617 0618 0619 0620 0621 0622 0623 0624 0625 0626 0627 0628 14.9 15.3 16.0 (B) (B) 17.6 15.8 (B) (B) (B) (B) (B) 16.5 16.8 17.6 15.2 15.7 16.6 (B) 17.8 17.1 (B) 18.1 18.6 19.0 15.1 15.7 16.6 17.2 (B) 17.7 16.2 (B) 17.4 17.7 (B) 18.8 16.7 17.3 14.7 15.6 16.2 17.0 17.1 14.4 15.1 15.9 16.4 16.8 15.2 16.0 16.5 17.4 (B) 18.1 14.1 (B| 15.3 (B) 16.1 16.0 (B) 18.6 A tagged conch is considered to be buried (B) if not found on one or more sub.sequent trips, but is eventually found. It is assumed to have been killed or to have migrated offshore if not subsequently found. No tags from this group (0617-0628) were found in the vicinity of Little Cockroach Cay during the period February 1981 to the end of the study, February 1983, despite 11 field searches made every 4-6 weeks during this period. Conch of about 18-20 cm shell length generally migrate into deep offshore waters from this cay (Iversen et al. 1987). NOTES Iversen et al : Survival related to burying behavior in Strombus gigas 385 00 -| CAT CAY r = 0.56 N = 14 80- . .. 60- • •_/ 40- / ^ :■■ 20- ' n- lOOi VIGILANT CAY r = 0.80 N = 9 80 H w u 02 W 60 40 20 100-1 LITTLE COCKROACH CAY r = 0.81 N = 7 80- 60- 40- .^^ 20- n- .^--"^^^ ' "20 40 60 80 PERCENT SURVIVAL 100 Figure 2 Relationship between Strongii.s gigas burying activity and survival. Vigilant Cay: oi^shore, 171 conch tagged 3 October 1980; SLR 6.7-9.3 cm. Cat Cay: 455 conch tagged 27 May 1980; SLR 8.9-19.1 cm. Little Cockroach Cay: 84 conch tagged 8 April 1981; SLR 9.2-15.5 cm. N = number of months the experiment lasted. during the summer on tide flats when temperatures reached extremely high levels (Weil and Laughlin 1984). In the few trials made to date, survival of hatchery- reared small conch (2.0-7.0 cm in length) released in different locations in the Bahamas and the Caribbean has been extremely low. For example, in St. Croix, Coulston et al. (1987) reported that unprotected juvenile conch did not survive. And in a small-scale preliminary trial, Iversen et al. (1986) reported com- plete mortality of 1.3-3.7 cm long, hatchery-reared conch placed under protective screens laid on the substrate and held in place by rocks. To fully interpret our results and put them in con- text with what is known about mortality and predation in juvenile queen conch, it is necessary to consider the factors affecting both burying behavior and mortality/ predation. Concerning the former, we first note that very small conch are almost never found unburied. Although emphasis of our research was on animals larger than 5 cm in length, we searched for small conch during all field trips. Of the 491 small conch (1-9 cm) we found, all but a few (~3%; A''= 15) were buried, some as deep as 20 cm, while others were barely covered by broken shells and rubble in shallow depres- sions on a large tidal flat. With only a few exceptions, those found on the substrate were larger than ~5 cm. Iversen et al. (1987) showed significant differences in burying activity within a tidal cycle at various sites in the Berry Islands. Significantly more conch buried on high tides than on low tides, which they suggested may be a response to possible increased predator ac- tivity during high tides when large swimming predators can reach the upper intertidal zone. This apparent tidal- height behavioral response should not be triggered in the deeper (subtidal) water offshore of Vigilant Cay where we released tagged conch. Our results showing a relationship between survival and burying behavior suggest that conch over a wide range of sizes are less vulnerable to predation when buried. This may explain why the highly vulnerable, thin-shelled young-of-year bury for extended periods. In addition to the effects of conch size and tide stage, burying behavior may be affected by time of year (a proxy which encompasses a number of environmental circumstances, such as water temperature and wind and sea conditions). According to Hesse (1979), sea- sonal variation in burying behavior is also a factor, with more active burying during the winter when the waters are cooler and the winds stronger than during the re- mainder of the year. However, no clear-cut seasonal trend in burying activity is evident from our data. Rather, there are generally wide variations in the num- ber of animals buried, which we believe are possibly related to different environmental conditions in the dif- ferent habitats of the cays where we collected our data. Our size-frequency data suggest that burying activity is not related to the size of the animals, after they at- tain a shell length of ~5-6 cm. With respect to mortality, Appeldoorn (1985) showed that juvenile queen conch mortality rates were highest during summer and lowest during winter in Puerto 386 Fishery Bulletin 88|2), 1990 Rico. Jory and Iversen (1983) found in their studies in the Berry Islands that the highest levels of predation occurred during the summer and the lowest during the winter. They also suggested that reduced mortality in the winter may be due to reduced predator activity and a decrease in conch activity. Small conch do not grow in length during the winter period of burying (Appeldoorn 1985, Iversen et al. 1986). Therefore, when they emerge in the spring to begin active foraging on the substrate they would have gained no obvious "size" survival advantage as a result of their "hibernation." Johnson et al. (1964) demon- strated that the shell thickness of queen conch in- creases with activity, and because buried conch would not be expected to be very active during the winter, shell thickness would probably not increase substan- tially over this time. Conch survival estimates show trends common to many mollusks: low survival for small animals, increas- ing as the size of the mollusk increases (Jory et al. 1984; Appeldoorn 1984, 1988; Iversen etal. 1986). Recoveries from 259 tagged wild conchs (SLR 5.0-21.0 cm) at Little Whale Cay, Bahamas, over a 6-month period showed a significant positive correlation between length at tagging and percent recovered (r = 0.95; jD<0.05). This correlation suggests that within this size range mortality decreased with increased conch size. Using floating cages to provide protection from pred- ators and increase survival, we demonstrated that as high as 96% survival was achieved for conch in the size range 5.0-8.0 cm over a 1-year period, while survival of tagged non-caged small conch of similar size after 6 months was only between 0 and 10%. Similar evi- dence on differential survival by conch size was gathered in Puerto Rico by Appeldoorn (1984) who, by using short-term (~8 weeks) mark-recapture data for both queen conch and milk conch (S. costatus), deter- mined that larger individuals suffered less mortality than smaller ones. Jory and Iversen (1988) provided direct evidence of a lack of increase in shell thickness in small juveniles by examining the relationship be- tween shell breaking strength and shell size of queen conch. They found little increase in breaking strength up to a size of ~5.5 cm, suggesting that queen conch shells are relatively thin below this length. Above ~5.5 cm, pressure to crush shells increased rapidly with in- creasing shell length. As queen conch increase in size, their spines increase in length and become more robust. Spines may reduce predation on mollusks in various ways (Palmer 1979). They increase the overall size of the shell, thereby limiting predation to larger predators. Conch spines also distribute crushing pressure over a greater area of the shell, thus increasing the pressure required to crush the shell (Jory and Iversen 1988). Spines also serve as additional area for attachment of epibionts which may help conceal mollusks from predators (Feifarek 1987). The amount of attached organisms on the shells of conch increases with the size of the conch, suggesting a decrease in long-term burying activity with increase in conch size (Iversen et al. 1986). All of the smallest conch we collected had very clean shells with no growth of attached organisms, in contrast to the rapid and heavy growth of algae we observed on the shells of similar-sized conch in hatchery tanks at the University of Miami laboratory and in floating cages at Little Whale Cay, Bahamas. These results further support the hypothesis that young conch spend most of their early life (up to ~1 year of age) below the surface of the substrate, a strategy that may provide haven from some predators (Hesse 1979, Appeldoorn 1985, Iversen et al. 1986). The relationships between shell size and strength and between survival and burying have practical applica- tion in suggesting place and time of year to release hatchery-reared conchs, in order to both minimize predation and increase the efficiency of planting pro- grams using hatchery-reared juveniles of queen conch and several other mollusk species. Placement of young hatchery-reared conch in relatively safe, natural areas optimal for burying and with certain physical char- acteristics such as strong currents, fine-to-coarse sand, and abundant intact and broken mollusc shells, where red and green algae and sponges occur (as described in Iversen et al. 1986) would obviate the need for over- wintering in an expensive hatchery facility. Further, as suggested by Jory and Iversen (1983), fall planting of small hatchery-reared conch would provide time for them to acclimate to their environment before the spring or early summer "emergence." This is par- ticularly relevant in view of evidence available for other mollusk species, where hatchery-reared juveniles are more vulnerable to predators when released in nature than wild juveniles (Tegner and Butler 1985). Develop- ment of an inexpensive enclosure for holding large numbers of young-of-the year under the surface of the substrate could further increase survival of hatchery- reared conch juveniles. Acknowledgments This research was funded jointly through the Univer- sity of Miami by the Wallace Groves Aquaculture Foun- dation and the Kirby Foundation. This support is deep- ly appreciated as is the field assistance provided by many individuals. Employees on Little Whale Cay gave generously of their time and knowledge of the area. NOTES Iversen et al.: Survival related to burying behavior in Strombus gigas 387 Citations Appeldoorn, R.S. 1984 The effect of size on mortality of small juvenile conchs (Strombxts gigas Linne. and S. costatus Gmelin). J. Shellfish Res. 4(l):37-43. 1985 Growth, mortality and dispersion of juvenile laboratory- reared conchs, Strombus gigas and S. costatus, released at an offshore site. Bull. Mar. Sci. 37(3):785-793. 1988 Ontogenetic changes in natural mortality rate of queen conch, Strombus gigas, (Mollusca: Mesogastropoda). Bull. Mar. Sci. 42(2);159-165. Brownell, W.N., and J.M. Stevely 1981 The biology, fisheries and management of the queen conch, Stromb-us gigas. Mar. Fish. Rev. 43(7):1-12. Coulston, M.L., R.W. Berey, A.C. Dempsey, and P. Odum 1987 Assessment of the queen conch (Strmnbus gigas) popula- tion and predation studies of hatchery reared juveniles in Salt River Canyon, St Croix, U.S. Virgin Islands. Proc. Gulf Caribb. Fish. Inst. 38:294-306. Davis, M., and C. Hesse 1983 Third world level conch mariculture in the Turks and Caicos Islands. Proc. Gulf Caribb. Fish. Inst. 35:73-82. Feifarek, B.P. 1987 Spines and epibionts as antipredator defenses in the thorny oyster Spo-ndylus americanus Hermann. J. Exp. Mar. Biol. Ecol. 105:39-56. Hesse, K.O. 1979 Movement and migration of the queen conch Strombus gigas in the Turks and Caicos Islands. Bull. Mar. Sci. 29: 303-311. Iversen, E.S., D.E. Jory, and S.P. Bannerol 1986 Predation on queen conchs, Strortibus gigas in the Baha- mas. Bull. Mar. Sci. 39(l):61-75. Iversen, E.S., E.S. Rutherford, S.P. Bannerot, and D.E. Jory 1987 Biological data on Berry Islands (Bahamas) queen conchs, Strombus gigas, with mariculture and fisheries management implications. Fish. Bull., U.S. 85:299-310. Johnson, R.F., J.J. Carroll, and L.J. Greenfield 1964 Some sources of carbonate in molluscan shell formation. Limnol. Oceanogr. 9(3):377-381. Jory, D.E., and E.S. Iversen 1983 Queen conch predators: Not a roadblock to mariculture. Proc. Gulf Caribb. Fish. Inst. 35:108-111. 1988 Shell strength of queen conch (Strombus gigas L.): Aqua- culture implications. Aquacult. Fish. Manage. 19:45-51. Jory, D.E., M.R. Carriker, and E.S. Iversen 1984 Preventing predation in molluscan mariculture. J. Worid Aquacult. Soc. 15:421-432. Palmer. A.R. 1979 Fish predation and the evolution of gastropod shell sculpture: Experimental and geographic evidence. Evolution 33:697-713. Randall, J.E. 1964 Contribution to the biology of the queen conch, Strom- bus gigas. Bull. Mar. Sci. Gulf Caribb. 14:246-295. Robertson, R. 1959 Observations on the spawn and veligers of conchs (Strom- bus) in the Bahamas. Proc. Malacol. Soc. Lond. 33(4):166-171. Tegner, M.J., and R.A. Butler 1985 The survival and mortality of seeded and native red abalones (Haliotis rufescens) on the Palos Verdes Peninsula. Calif. Fish Game 71(3):150-163. Weil, E., and R. Laughlin G. 1984 Biology, population dynamics, and reproduction of the queen conch Strombus gigas Linne in the Archipielago de los Roques National Park. J. Shellfish Res. 4(l):45-62. Stomach Contents and Parasite Infestation of School Bluefin Tuna Thunnus thynnus Collected from the Middle Atlantic Bight, Virginia* David B. Eggleston Virginia Institute of Marine Science, School of Marine Science The College of William and Mary, Gloucester Point. Virginia 23062 Eleanor A. Bochenek Virginia Institute of Marine Science, School of Marine Science The College of William and Mary, Gloucester Point, Virginia 23062 Present address: Louis Berger and Associates, Inc 100 Halsted Street, East Orange, New Jersey 07019 In the western Atlantic Ocean, north- ern bluefin tuna Thunnus thynnus are distributed from Labrador and Newfoundland to the Gulf of Mex- ico, Caribbean Sea, and off Vene- zuela and Brazil. The northern blue- fin tuna is epipelagic and usually oceanic, but seasonally strays near the coast (Collette and Nauen 1983). During June through October, these tuna are common off the eastern United States and Canada (Squire 1962) and support both commercial and recreational fisheries. From the end of May to August, many pods of small school bluefin tuna (< 100 kg) migrate past Virginia on their way to more northern feed- ing grounds. These tuna are caught 30 to 60 km off the Virginia coast in the vicinity of numerous shoals or "hills," by recreational anglers trolling dead bait or lures on or near the surface (Figley 1984). In 1986, 886 boats participated with some degree of regularity in the recreational fishery for tuna and billfish out of Virginia ports (Boch- enek and Lucy In press). Further north, there is a recreational fish- ery for giant (>200 kg) and me- dium (100-200 kg) bluefin tuna, •Contribution no. 1580 of the Virginia Insti- tute of Marine Science. and, to a lesser extent, school blue- fin tuna (Figley 1984). Bluefin tuna are opportunistic pred- ators that prey upon fishes, moUusks, crustaceans, and salps (Crane 1936; Bigelow and Schroeder 1953; Knmi- holz 1959; Dragovich 1969, 1970; Mason 1976; Matthews et al. 1977; Holliday 1978). Pacific bluefin tuna Thunnus thynnus orienialis caught off California and Baja California preferred the same prey as the At- lantic Ocean subspecies Thunnus thynnus thynnus (Pinkas 1971). The spawning stock of the west- em Atlantic bluefin tuna has declined sharply since 1970, and both recruit- ment and juvenile stock size are still substantially lower than in 1970 (ICCAT 1987). Thus, information about life-history characteristics, such as trophic habits, is essential for developing sound management plans for this important commercial and recreational fish. Mason (1976) and Holliday (1978) studied the feed- ing behavior of school bluefin tuna captured along the eastern coast of the United States; however, only 68 bluefin tuna stomachs were collec- tively examined from fish caught off or near the Virginia coast (lat 36- 38°N and long. 75°W). Therefore, knowledge of the feeding habits of school bluefin tuna off the Virginia coast is relatively sparse. The pres- ent paper describes the findings of stomach content analysis for juve- nile bluefin tuna collected during the summer of 1986 by recreational fishermen along the mid-Atlantic coast off Virginia. Methods and materials During June and July 1986, stom- ach samples of 97 bluefin tuna were obtained from recreational fisher- men as they landed their catch at Rudee Inlet, Virginia Beach, and at Wachapreague on the eastern shore of Virginia. Curved fork length (mm) and the area of capture (Fig. 1) were recorded for each fish. Fish that could not be identified with a specific area of capture were elim- inated from the sample. Weights (kg) were recorded for tuna official- ly weighed on certified marina scales. Stomachs were removed and placed in 10% buffered formalin. Stomachs were opened and designated in the laboratory as either containing food or empty. Stomachs containing only parasites were classified as empty. Stomach contents were rinsed in water and stored in 10% ethanol im- til identification. Prey items were sorted into ma- jor food groups (fishes, crustaceans, mollusks, and unidentifiable remains), enumerated, and identified to the lowest possible taxon with the aid of a binocular dissecting scope. Vol- umes were determined by water dis- placement using a graduated cylin- der and measured to the nearest 0.5 mL. Fishes too far digested for cer- tain identification were placed in an unidentified teleost category and used in estimating total prey vol- ume. The majority of unidentified teleost material resembled remnants of sand lance {Ammodytes spp.) more than any other recognizable species. Manuscript accepted 18 December 1989. Fishery Bulletin, U.S. 88:389-395. 389 390 Fishery Bulletin 88(2). 1990 Figure 1 Tuna and billfish grounds off Virginia. To help evaluate the relationship of the various food items found in stomachs we employed an index of relative importance (IRI) (Pinkas 1971): IRI = {N + V) F, where N = numerical percentage, V = volumetric per- centage, and F = frequency of occurrence percentage. Percent IRI consisted of the IRI value of each prey category (unidentified fish and cephalopods excluded) divided by the sum of the IRI values (unidentified fish and cephalopods excluded). To determine if the quan- tity of the key trophic group differed by area of capture (Fig. 1), displacement volume was compared against three areas sampled with a one-way ANOVA model (with displacement volume as the dependent variable and area of capture as the independent factor). The key trophic group was composed of pooled volumetric con- tributions of both identified and unidentified teleost re- mains. The three areas were: (1) the "Hot Dog," (2) '"Fish hook" and "S.E. Lumps," and (3) "21 Mile Hill." The "Fish hook" and "S.E. Lumps" areas were pooled because boat captains generally fished both areas dur- ing the same trip. The remaining areas, "26 Mile Hill" and "V-Buoy," were both eliminated from the hypoth- esis test because of low sample sizes (N = 2 for both). Significant differences were contrasted by a Student- Neuman-Kuels (SNK) multiple range test set at an experiment-wise error rate (EWER) of 0.05 (Zar 1984). Tests for normality and equality of variance (Zar 1984) NOTES Eggleston and Bochenek. Stomach content analysis of juvenile Thunnus thynnus off Virginia 391 Table 1 List of prey species or class groups occurring in stomachs of juvenile bluefin tuna Thunnus thynnus from the Mid-Atlantic Bight, | Virginia, 1986. % Frequency of occurrence No. of individual based on stomachs Index of prey items from Percent Volume Percent containing food relative importance 72 stomachs by number (mL) volume (N = 72) ilRI) %IRI Teleosts Ammodytes spp. (sand lance) 403 84.1 1028.0 30.75 48.6 5583.2 0.968 Peprilus triacanthus (butterfish) 11 2.3 84.0 2.51 2.8 13.5 0.002 Hippocampus erectus 9 1.9 2.5 0.07 1.4 2.7 0.0004 (lined seashore) Alute!-us scriptus 1 0.2 1.5 0.04 1.4 0.4 0.00006 (scrawled filefish) Unidentified teleosts NA NA 1653.5 49.46 75.0 NA NA (primarily sand lance) Cephalopods LoUiguncula hrevis 48 10.0 432.0 12.92 6.9 158.3 0.027 (Atlantic brief squid) Loligo pealeii (longfin squid) 2 0.4 124.0 3.71 2.8 11.6 0.002 Unidentified cephalopods NA NA 5.0 0.15 1.4 NA NA Miscellaneous Salpidae 12.3 0.37 8.3 NA NA Idotea sp. (Isopod) 5 1.0 0.5 0.01 1.4 1.5 0.0002 Totals 479 3343.3 5771.0 1.0 Total stomachs analyzed 97 No. (%) containing food materials 72(74.2) (identified and unidentified) No. (%) empty 25(25.8) indicated that the logarithmically transformed volumes were appropriate for ANOVA. Results Food analysis Of the 97 juvenile bluefin tuna stomachs examined, 72 (74%) contained food. These tuna averaged 21.3 kg (n = 7, SD 7.7, range 15-39 kg) with a mean fork length of 90 cm (ft = 85, SD 13, range 70-132 cm). Stomach contents consisted of two primary food groups: teleosts and cephalopods. Teleosts contributed over five times the percent volume to the diet (82.8%) compared with cephalopods (16.8%) (Table 1). Teleosts occurred in 91% of those stomachs containing food items and accounted for 86% of the total identified prey items (Table 1). Major subgroups of identifiable teleosts by percent frequency of occurrence (based on number of stomachs containing food), IRI, and percent IRI, listed in decreasing order were sand lance, butterfish Pepri- lus triacanthus, lined seahorse Hippocampus erectus, and scrawled iWeixsh Aluterus scriptus (Table 1). Sand lance was the predominant teleost occurring in stom- achs, especially considering that the unidentified tele- ost category (probably primarily sand lance) contrib- uted the greatest volume of all prey species found (Table 1). Cephalopods occurred in 14.4% of those stomachs containing food (Table 1). This group was represented by two species, the Atlantic brief squid LoUiguncula hrevis and the longfm squid Loligo peakii. Unidentified cephalopod remains accounted for the lowest percent volume (0.2%) of prey items in stomachs containing food, whereas Atlantic brief squid contributed the highest (12.9%) (Table 1). A third, miscellaneous prey category included salps and one immature species of isopod. A cigarette wrap- per and piece of Sa rgassum weed were each present in two of the stomachs. The combined volumetric contributions of teleost re- mains to the gut were significantly affected by area of capture (ANOVA; F = 8.93, df 2, 82, P<0.0003). Stomach contents of tuna landed from "21 Mile Hill" had significantly higher volumes of teleost remains than did stomachs taken from either the "Hot Dog" 392 Fishery Bulletin 88(2). 1990 80 E w 70 d 60 1 50. § 40 i 30 a 1 20 lij -1 r 1 ^ '43-' ^ 20 ' • 1 ; 20 : J. i 1 HOT DOG FISH HOOK-SE LUMPS 21 MILE HILL AREA LANDED Figure 2 Mean combined displacement volume of identified and unidentified teleost remains from stomachs of bluefin tuna collected from three different areas off Virginia. Numbers within each bar indicate number of specimens sampled; vertical lines indicate + 1 SE. or "Fish hook and S.E. Lumps" areas (SNK: EWER 0.05) (Fig. 2). Digenetic trematodes Hirudinella ventricosa were found in 8 (11.1%) of the stomachs and averaged 10 mm in length and 2-3 mm in width. The worms were never attached to the lining of the stomach and were typically found at the posterior end. The number of worms per stomach ranged from 1 to 2 with a mean of 1.14 H. ventricosa per individual stomach. The possi- ble effects of area landed on the number of trematodes occurring in the stomachs were not evaluated because of the relatively low rate of parasitism. Discussion Diet This study indicates that school bluefin tuna, captured off the Virginia coast, feed predominantly on the sand lance. Mason (1976) was the first to report sand lance as a prey item of school bluefin tuna caught off the U.S. East Coast. He also reported sand lance to be the most important prey of school bluefin tuna caught off Virginia, but for fish taken north of Virginia, mackerel (Scomber spp.) replaced sand lance as the dominant prey. Holliday (1978) also found the sand lance to be the predominant food item for bluefin tuna captured by trolling along the U.S. East Coast. The IRI of sand lance in this study (IRI = 5583) is very similar to that reported by Holliday (1978) (/ie/ = 4896). Sand lance form dense schools over New England and mid- Atlantic Continental Shelf areas. They occur throughout the water column during daylight hours, and are available to tuna predation at various depths (Meyer et al. 1979, Auster and Stewart 1986). Tuna predation on sand lance may affect the populations of this important forage species off Virginia. Reproduc- ing populations of sand lance, as indicated by egg and larvae counts, exist on the Virginia shelf (Norcross et al. 1961); hence, the Virginia coast is an important habitat to the species. The sand lance serves as an im- portant link between secondary producers and higher trophic-level fish and mammals in marine food chains (Bigelow and Schroeder 1953); thus, extensive preda- tion by tuna could affect marine mammal populations. A cause-and-effect relationship may exist between low mackerel and herring stocks (resulting from heavy fishing mortality) and the observed population explo- sion of sand lance larvae in the mid- to late 1970s (Sherman et al. 1981); thus, tuna predation on sand lance could be beneficial to the return of mackerel and herring stock abundance. The Atlantic brief squid was the second most impor- tant item consumed by school bluefin tuna examined in our study. Mason (1976) found two squid in the 20 fish he examined from Virginia waters. Holliday (1978) also reported similar species of cephalopods in stomach contents of the bluefin tuna taken off the U.S. East Coast. Krumholz (1959), working near the Bahamas, reported salps as the second most important food item. In the western North Atlantic, Dragovich (1970) noted molluscs (mainly cephalopods) as second in trophic importance. Similarly, Matthews et al. (1977) also reported cephalopods, pteropods, and heteropods as being the most frequent invertebrate forage group after fishes. For California bluefin tuna, the second most important food item was the California market squid Loligo opalescens or the pelagic swimming crab Pleuroncodes planipes, depending upon the area of cap- ture (Pinkas 1971). The butterfish, lined seahorse, and scrawled filefish were found in very few stomachs, being rare con- tributors to the diet of bluefin tuna in this study. These prey species demonstrate considerable diversity in their foraging locations, including nearsurface, mesopelagic, and demersal habitats. It is possible that the butter- fish, lined seahorse, and scrawled filefish are associated with drifting Sargassum weed; thus, tuna may feed in part around drifting Sargassvm. communities. Other miscellaneous items found in the stomachs were salps, the isopod Idotea spp., a cigarette wrapper, and Sargassum weed. Holliday (1978) also reported the occurrence of Idotea spp. in stomachs of bluefin tuna caught trolling near Sargassum communities. He hypothesized that the isopod and Sargassum- weed were accidently ingested by the tuna while pursuing other prey. NOTES Eggleston and Bochenek: Stomach content analysis of juvenile Thunnus thynnus off Virginia 393 Parasites The digenetic trematode Hirudimlla ventricosa occur- red in 11% of the stomachs examined in this study. Mason (1976) also found an annulated hemiurid trema- tode in 2% of the bluefin tuna stomachs he examined from the western Atlantic Ocean. The trematodes in Mason's (1976) study were found in both empty stom- achs and stomachs which contained food. Crane (1936) reported Distoma-Vike worms in 25% of giant bluefin tuna stomachs he examined off Maine. Hirudinella ventricosa (= marina) occurred in 9% of school bluefin tuna and 48% of giant bluefin tuna stomachs collected from North Carolina to Massachusetts (Holliday 1978). Giant trematodes of the genus HinidineUa frequent- ly parasitize scombroid fishes (Nigrelli and Stunkard 1947, Nakamura and Yuen 1961, Watertor 1973, Manooch and Hogarth 1983). Adult parasites typical- ly attach to the stomach lining and remain near the site of attachment throughout this life-stage (Manooch and Hogarth 1983). These digenetic endoparasites have complicated life cycles involving an alternation of generations and hosts; however, the life cycle oi Hiru- dinella spp. is still unknown (Manooch and Hogarth 1983). Attempts to evaluate the incidence of parasitism by size and sex of the host and by geographical area of collection have demonstrated mixed results (Nakamura and Yuen 1961, Manooch and Hogarth 1983). Naka- mura and Yuen (1961) examined the occurrence of the parasite H. ventricosa ( = marina) in the stomachs of skipjack tuna Euthynnus pelamis collected from Hawaii and off the Marquesas. They concluded that significant differences in the occurrence of trematodes collected from these two areas were attributable to time (year of collection) rather than area. Manooch and Hogarth (1983) reported distinct differences in the in- cidence of parasitism by H. ventricosa between wahoo Acanthocybium solanderi from the coast of Florida- South Florida and wahoo from the rest of the south- eastern Atlantic. They suggested that this difference may reflect two subpopulations of wahoo along the southeastern U.S. coast: a northern population char- acterized by high incidence of trematodes, and a south- ern population with a much lower incidence. Watertor (1973) examined 258 bluefin tuna captured off the East Coast of the United States Oat. 35-40°N; long. 65-75°W) and off the northeast coast of South America (lat. 0-18°N; long. 50-82°W). Of these, 51 were infected with H. ventricosa ( = marina), nearly twice the infection rate noted in this study. There are two possible explanations for this difference. First, the parasites described by Watertor (1973) were pooled from both the eastern U.S. and northeastern South American samples. Inclusion of a South American group, with possibly a higher prevalence of parasitism, similar in nature to that described for wahoo by Manooch and Hogarth (1983), might have biased the values reported by Watertor (1973). Secondly, Water- tor (1973) did not report the overall size ranges of blue- fin tuna used in his study. Inclusion of giant bluefin tima, with a higher prevalence of parasitism (see Crane 1936) may also contribute to apparent differences in levels of infection. Area effects Environmental factors such as temperature and ocean- ographic frontal zones have been shown to markedly influence the distribution, abundance and catchability of tunas (Murphy 1959, Uda 1973, Laurs and Lynn 1977, Rockford 1981, Sund et al. 1981, Laurs et al. 1984). Murphy (1959) suggested that the aggregation of albacore Thunnus alalunga in clear water on the oceanic side of fronts in nearshore areas may reflect an inability to efficiently capture large, mobile prey in turbid coastal waters. This same mechanism may help to explain the higher combined displacement volumes of sand lance and unidentified teleost remains in the stomachs of tuna taken from the "21 Mile Hill" com- pared with the "Hot Dog" or "Fishook and S.E. Lumps" areas (Fig. 2). Turbidity associated with ef- fluent from Chesapeake Bay might have reduced the ability of bluefin tuna to detect mobile forage such as the sand lance and other teleosts. The effluent from the Chesapeake Bay appears in shelf waters as a lens of freshened water (with high concentrations of bay water constituents) extending offshore and towards the south as a part of the general shelf circulation (Ruzecki 1981). The three areas in question are directly offshore of the Chesapeake Bay mouth (Fig. 1). Differences in the diet of bluefin tuna have also been attributed to depth of capture, availabOity and type of food in a given area, time of day or year, spawning, atmospheric con- ditions, physiological conditions of predator fish, size of prey, and size of the bluefin tuna (Dragovich 1970). We conclude that the sand lance is the most impor- tant forage of school bluefin tuna off the Virginia coast and suggest that this prey species, as well as teleosts in general, may become more vulnerable to tuna preda- tion in areas least affected by the turbid waters of the Chesapeake Bay plume. In addition, the occurrence of the digenetic trematode Hirudinella ventricosa in a small but significant number of bluefin tuna off Virginia suggests that variation in infestation rates of this para- site might provide a mechanism to help distinguish among possible subpopulations of bluefin tuna occur- ring in the Western Atlantic. 394 Fishery Bulletin 88|2). 1990 Acknowledgments We thank J. Lucy, N. Chartier, and B. Sweeney for assistance in field collections, and B. Cornett and M. Peterson, participants in the Governor's School for Marine Science of the Virginia Department of Educa- tion for help in collecting and processing tuna stomachs and in enumeration of food contents. Special thanks to J. Lucy for his helpful advice throughout the dura- tion of this study. Helpful reviews of the manuscript were provided by J. Colvocoresses, J. Lucy, J. Musick, and two anonymous referees. This work was supported by funds from the Virginia Institute of Marine Science Advisory Services Program and National Marine Fisheries Service Port Sampling Program for Large Pelagics, Northeast Fisheries Center. First authorship was determined by a coin toss. Citations Auster, P.J.. and L.L. Stewart 1986 Species profiles; Sand Lance: life histories and environ- menta! requirements of coastal fishes and invertebrates North Atlantic. U.S. Fish Wildl. Serv. Biol. Rep. 82(11.66), 11 p. Bigelow, H.B., and W.C. Schroeder 1953 Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 74 (vol. 53), 577 p. Bochenek, E., and J. Lucy In press A comparison of two sampling methods for analyz- ing Virginia's recreational marlin and tuna fishery. In Stroud, R. (ed.). Proceedings, Second International Billfish Symposium, Kailua-Kona, HI. Part II, Marine Recreational Fisheries 13. National Coalition for Marine Conservation Inc., Savan- nah, GA. CoUette, B.B., and C.E. Nauen 1983 FAO species catalogue. Vol. 2. Scombrids of the world, an annotated and illustrated catalogue of tunas, mackerels, bonitos, and related species known to date. FAO Fish. Synop. 125, vol. 2, 1.37 p. Crane, J. 1936 Notes on the biology and ecology of giant tuna, Thun- nus thynnus Linnaeus, observed at Portland, Maine. Zoo- logica (NY) 212:207-212. Dragovich, A. 1969 Review of studies of tuna food in the Atlantic Ocean. U.S. Fish. Wildl. Serv. Spec. Sci. Rep. Fish 593, 21 p. 1970 The food of bluefin tuna (Thiinnus thynmia) in the West- em North Atlantic Ocean. Trans. Am. Fish. Soc. 99:726-731. Figley. W. 1984 Recreational fishery for large offshore pelagic fishes of the Mid- Atlantic Coast. NJ Div Fish Game Wildl. Tech. Ser. 84-1, Trenton, NJ, 66 p. Holliday, M. 1978 Food of Atlantic bluefin tuna, Thunnm thynnm (L.), from the coastal waters of North Carolina to Massachusetts. M.S. thesis, C.W. Post College, Long Island Univ., Long I., NY, 27 p. ICCAT 1987 Newsletter 17(3):2 (November), Int. Comm. Conserv. Atl. Tunas, Madrid, Spain. Kumholz, L.A. 1959 Stomach contents and organ weights of some bluefin tuna, Thunnus thynnus (Linnaeus), near Bimini, Bahamas. Zoologica (NY) 44:127-131, Laurs, R.M., and R.J. Lynn 1977 Seasonal migration of North Pacific albacore, Thunnus alalunga, into North American coastal waters: Distribution, relative abundance, and association with transition zone waters. Fish. Bull., U.S. 75:795-822. Laurs, R.M., P.C. Fiedler, and D.R. Montgomery 1984 Albacore tuna catch distributions relative to environmen- tal features observed from satellites. Deep-Sea Res. 31(9): 1085-1099. Manooch, C.S., and W.T. Hogarth 1983 Stomach contents and giant trematodes from wahoo, Acanthocybimn solanderi, collected along the south Atlantic and Gulf Coasts of the United States. Bull. Mar. Sci. 33(2): 227-238. Mason, J.M. 1976 Food of small, northwestern Atlantic bluefin tuna, Thun- nus thynnus (L.) as ascertained through stomach content analysis. M.S. thesis, Univ. Rhode Island, Kingston, 27 p. Matthews, F.D., D.M. Dankaer, L.W. Knapp, and B.B. Collette 1977 Food of Western North Atlantic tunas {Thunnus) and lancetfishes (Alepisaurvs). NOAA Tech. Rep. NMFS SSRF- 7006, Natl. Oceanic Atmos. Adm., Natl. Mar. Fish. Serv., 19 p. Meyer, T.L., R.A. Cooper, and R.W. Langton 1979 Relative abundance, behavior, and food habits of the American sand lance, Ammodytes americanus, from the Gulf of Maine. Fish, Bull., U.S. 77:243-253. Murphy, G.I. 1959 Effect of water clarity on albacore catches. Limnol. Oceanogr. 4:86-93. Nakamura, E.L., and H.S.H. Yuen 1961 Incidence of the giant trematode. HirudineUa marina Garcin, in skipjack tuna, Euthynnus pelamis (Linneaus), from Marquesan and Hawaiian waters. Trans. Am. Fish. Soc. 90: 419-423. Nigrelli, R.F., and H.W. Stunkard 1947 Studies on the genus Hu-udinella, giant trematodes of scombriform fishes. Zoologica (NY) 31:185-196. Norcross, J. J., W.H. Freeman, and E.B. Joseph 1961 Investigations of inner continental shelf waters off lower Chesapeake Bay. Part II. Sand lance larvae, Aminodytfs ameri- canus. Chesapeake Sci. 2:49-59. Pinkas, L. 1971 Bluefin tuna habits. In Pinkas, L., M.S. Oliphant. and I.K. Iverson (eds.). Food habits of albacore, bluefin tuna, and bonito in California waters, p. 47-63. Calif. Dep. Fish Game, Fish. Bull. 152. Rockford D.J. 1981 Anomalously warm sea surface temperatures in the Western Tasman Sea. Their causes and effects upon southern bluefin tuna catx?h 1966-1977. Rep. 114, Div, Fish, Oceanogr., CSIRO, Oonulla, Australia, 21 p. Ruzecki, E.P. 1981 Temporal and spatial variations of the Chesapeake Bay Plume. In Campbell, J.W., and J. P. Thomas (eds.), Chesa- peake Bay Plume study, Superflux 1980, p. 111-130. NASA Conf. Publ. 2188, Wash. DC. 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. NOTES Eggleston and Bochenek Stomach content analysis of juvenile Thunnus thynnus off Virginia 395 Squire, J.L. Jr. 1962 Distribution of tunas in oceanic waters of the north- western Atlantic. Fish. Bull., U.S. 62:323-341. Sund, P.N., M. Blackburn, and F. Williams 1981 Tunas and their environment in the Pacific Ocean: A review. Oceanogr. Mar. Biol. Annu. Rev. 19:443-512. Uda, M. 1973 Pulsative fluctuation of oceanic fronts in association with the tuna fishing grounds and fisheries. J. Fac. Mar. Sci. Technol. Tokai Univ. 7:245-265. Waterier, J.L. 1973 Incidence o{ Hirudinella marina Garcin, 1730 (Trema- toda: Hirudinellidae) in tunas from the Atlantic Ocean. J. Parasitol. 59(l);207-208. Zar. J.H. 1984 Biostatistical analysis. Prentice-Hall. Englewood Cliffs. NJ, 718 p. Horizontal and Vertical Movements of Pacific Blue Marlln Captured and Released Using Sportfishing Gear* Kim Holland Hawaii Institute of Marine Biology, P O Box 1346 Kaneohe, Hawaii 96744 Richard Brill Randolph K.C. Chang Honolulu Laboratory, Southwest Fisheries Center National Marine Fisheries Service, NOAA, 2570 Dole Street Honolulu, Hawaii 96822-2396 Despite the commercial and recre- ational importance of Pacific blue marlin Makaira nigricans, little is known about their biology or be- havior. This is due mainly to their large size and pelagic habitat and the difficulty in maintaining them in captivity or observing their behav- ior in the wild. However, two tech- niques are available to elucidate their movements: capturing and re- leasing marlin fitted with identifica- tion tags, and tracking of fish car- rying ultrasonic transmitters. The recovery of tagged marlin has enhanced our understanding of the long-term geogi'aphical range of individual fish and their minimum rates of travel (Squire 1974, Squire and Nielsen 1983, Bayliff and Hol- land 1986), and growing interest in the tag and release of marlin by sportfishermen should produce an increasingly precise picture of mar- lin movements. However, this tech- nique cannot answer questions re- garding the vertical movements of marlin, and questions remain con- cerning both the survival of tagged fish following the trauma of capture and the nature of their behavior im- mediately upon release. •Sea Grant Publication UNIHI-SEAGRANT JC-90-13. Fine-scaJe observations of the hori- zontal and vertical movements of pelagic fish can be obtained for periods of up to a few days by track- ing fish equipped with depth-sensi- tive ultrasonic transmitters (Hol- land et al. 1985, 1990; Bayliff and Holland 1986). For billfishes, this technique has been used to track swordfish Xiphius gladius (Carey and Robison 1981) and striped mar- lin Tetrapturus audax (Holts and Bedford In press). Similarly, Yuen et al. (1974) tracked Pacific blue marlin but used temperature-sensi- tive transmitters to monitor ambi- ent water temperature from which depth was later calculated using bathythermograph data. However, this study heightened concerns about tagging mortality rates because three of the five tagged fish died soon after release. Here we report on the movements of six Pacific blue marlin tracked in the waters around the Hawaiian Islands. Of particular interest were survivorship and behavior of the fish immediately upon release, their patterns of vertical movement, and their overall patterns of horizontal movement. To discern any common patterns of movement associated with one particular area of ocean, three fish were caught and tracked from one well-defined location on the Kona coast of Hawaii. For com- parison, two other tracks were in- itiated several miles away along the same coast, and one marlin was tracked off the Waianae coast of Oahu. Methods The ultrasonic tracking techniques employed were identical to those used previously to track yellowfin tuna (HoDand et al. 1985, 1990; Bay- liff and Holland 1986). The transmit- ters used in the present study had a nominal life span of 3 days and a maximum working pressure of 500 psi (Vemco, Halifax County, Nova Scotia). The signal, encoding depth information by variable pulse inter- val, was recorded on audiotapes for later onshore plotting of vertical movements of the fish. Horizontal location was determined every 15 minutes (or more frequently when necessary) by using a combination of Loran-C, radar, visual, and bathy- metric fixes. Water temperature was measured by expendable bathy- thermographs deployed approx- imately every 3 hours. Aggregate depth and tempera- ture distributions were calculated (with 10-m and 1°C bins, respective- ly) as percentages of the total track time spent at any particular depth or temperature. For temperature, the combined distributions of all fish were calculated as the percentage of time spent in the various tem- perature strata relative to the up- per mixed layer. Each transmitter was attached with a 10-cm length of 130-lb test monofilament to a stainless steel "arrowhead" (~2.75 x 1.75 x 0.1 cm; Fig. 1) modified from the type Reference to trade names does not imply en- dorsement by the National Marine Fisheries Service, NOAA. Manuscript accepted 5 March 1990. Fishery Bulletin, U.S. 88:397-402. 397 398 Fishery Bulletin 88(2). 1990 10 CM TRANSMITTER I^-- Figure 1 Tag attachment system for Pacific blue marlin. A stainless steel ar- rowhead, with downcurved tines and central notch to accommodate the prong of the applicator pole, is attached to the transmitter body with a 10-cm length of 130-lb test monofilament crimped in two places. used by Yuen et al. (1974) and Carey and Robison (1981). The arrowhead fitted into the tip of a standard tagging pole to which the transmitter was loosely bound with rubber bands (Yuen et al. 1974, Holland and Bayliff 1986). The fish was harpooned in the trunk musculature, thereby allowing the transmitter to lie against the surface of the fish alongside the forward half of the dorsal fin. Once the transmitter was at- tached, the fish was released by cutting the leader, leav- ing the hook in place. Fish were caught by surface trolling with artificial lures and standard sportfishing rods and reels. Two of the fish were caught on the tracking research vessel Kaahele'ale, and four were caught by teams par- ticipating in the 1988 and 1989 Hawaii International Billfish Tournaments (HIBT). In these latter cases, the transmitter and applicator pole were passed to the anglers' vessels during the fighting of the fish. Crew members then tagged the fish while the tracking vessel stood by ~100 m away. Once the fish had been re- leased, tracking began. To replicate tracks from one well-defined area, three tracks were initiated on the fishing grounds off Keahole Point, Hawaii. This is an area approximately 5 nautical miles (nmi) long and 2 nmi wide that is renowned for high catch rates of marlin during certain periods of the summer. For comparison, two other tracks were ini- tiated from locations ~12 nmi south of the Keahole grounds, and one fish was tracked off Barbers Point, Oahu. Results Six Pacific blue marlin were tracked. The three Kea- hole fish were all caught within a 1-nmi radius over a span of 19 days in 1988; one track was 24 hours, and two were of 42 hours duration each (Fig. 2). In 1989, Figure 2 Horizontal movements of five Pacific blue marlin tagged off the Kona coast of Hawaii: three off Keahole Point and two off Keauhou. Lines perpendicular to the tracks represent hourly positions. over a period of 4 days, two marlin were tracked from starting points ^^4 nmi apart off Keauhou and were tracked for 26 and 29 hours (Fig. 2). The Barbers Point fish was tracked for 7 hours. Thus, all fish survived for at least 7 hours after release, and there is no reason to believe that any of the fish subsequently died as a result of the capture and tagging procedure. Synopsis of tracks Fish 8710 (weighing ~150 kg) was caught on a single- hooked artificial lure trolled behind the tracking vessel Kaahele'ale at a location 7 nmi west of Barbers Point, at 0904, 18 October 1987. After a fight time of 30 minutes, the fish was brought to the side of the boat. The fish was cleanly hooked through the bill and was immobile. The transmitter was embedded dorsolater- ally, about a third of the way back along the dorsal fin. After release, the fish glided downwards and away from the boat and swam away on a southwesterly (off- shore) course, which it maintained for the duration of the track. After 7 hours the fish was lost while at the surface. Data on the rate of horizontal movement for NOTES Holland et al.: Capture and release behavior of Pacific blue marlin off Hawaii 399 Table 1 Rates of horizontal movement of captured and released Pacific | blue marlin off Hawaii. Average speed (kn + SD) Dura- tion Track First ID (hr) Day Night Total hour 8710 7.0 1.21 + 0.27 NA 1.21 + 0.27 1.50 8803 24.0 1.48 + 0.31 1.37 + 0.32 1.43 + 0.30 1.20 8804 42.0 2.04±0.43 1.78 + 0.37 1.90 + 0.42 3.25 8807 42.0 2.25 + 0.67 2.09 + 0.67 2.18 + 0.65 3.25 8903 26.0 0.95 + 0.54 0.65 + 0.36 0.83 ±0.49 1.80 8904 29.0 1.62±0.37 1.68 + 0.39 1.65 + 0.37 3.00 this fish and the other five marlin are summarized in Table 1. In terms of vertical behavior, this fish dove when released and spent the next 4 hours in the top 3 or 4 degrees of the thermocline at depths varying between 50 and 70 m. The fish then moved closer to the sur- face, spending 92% of the last 3 hours of the track within the surface mixed layer, moving between the surface and 35 m. Fish 8803 (60 kg) was caught and tagged 2 nmi off Keahole Point at 1120, 8 August 1988, by a team par- ticipating in the HIBT. The fish was fought for ~20 minutes and described as tired but in good condition when released by the anglers. This marlin moved steadily west for ~ 12 hours before curving north dur- ing the night (Fig. 2). The fish was moving northwest when the track was terminated after 24 hours because of deteriorating sea conditions. The vertical movements of this fish (Fig. 3A) had a consistent "floor" where 50 100 150 -■ 50 J 100 •i 150 50. I 100 I 150- ■^...jimM^MM^L^M^ r --W -n 26 C 12 13 14 15 16 17 18 19 20 21 22 23 0 2 3 4 5 6 7 irnrr 9 10 11 12 13 15 17 19 21 23 I J ^•^•^•^ ^r rr26 c m^ 9 11 13 15 17 19 21 23 1 3 5 7 9 11 13 15 17 19 21 23 1 3 iT^:^. w -26 C i^^^^ ^■^^^ ^ii""P 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 0 1 2 3 4 5 6 7 n v^- /^[J", VWiuMil ^^^ "To i'i 12 13 14 15 16 17 18 19 20 21 22 23 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 REAL TIME Figure 3 Vertical movements of marlin tracked off Keahole Point and Keauhou, Hawaii. A = fish 8803; B = 8804; C = 8807; D = 8903; E = 8904. Solid horizontal line represents the sharpest inflection in water temperature at the bottom of the mixed layer, whereas dashed horizontal lines represent 1°C isotherms, starting at 1°C below surface temperature. Horizontal bars represent nighttime. 400 Fishery Bulletin 88(2), 1990 the surface mixed layer met the top of the thermocline at depths varying between 75 and 100 m. During the daytime, the fish moved in and out of the top of the thermocline, with occasional excursions towards the surface; during the night, movements were more com- pletely confined within the mixed layer with more time spent very close to the surface. Fish 8804 (also estimated to weigh 60 kg) was caught and tagged by an HIBT team at 1310, 11 August 1988, 3 nmi off Keahole Point. Fight time and release condi- tion were similar to fish 8803. After an initial, brief movement towards shore, this fish also proceeded directly west for about 8 hours before curving south- ward after sunset on the first night (Fig. 2). This direc- tion was maintained all day during the second day until after sunset when it turned more shoreward. However, a southerly course was resumed after sunrise of the third day. The track was terminated after 42 hours. The daytime vertical movements of this fish were focused around the interface between the mixed layer and the thermocline. From that depth (around 90 m) intermittent excursions were made to the surface. Nighttime distribution was shifted markedly towards the surface (Fig. 3B). Fish 8807 was caught by the Kaahele'ale at 0900 on 27 August 1988, 3 nmi off Keahole Point. The fish, which was fought for 35 minutes, was caught using an artificial lure with a single hook that lodged in the base of the marlin's bill. The fish, weighing in excess of 160 kg, was completely immobile and floating belly-up at the side of the boat when the transmitter was attached near the midline in line with anterior insertion of the anal fin. When the leader was cut, the fish sank slowly upside down under the boat but, after ~30 seconds, righted itself and began to swim slowly downwards. During the next hour, the fish traveled 3.25 nmi and, as with the previous two Keahole fish, headed steadily west, in this case for 7 hours. After sunset, the fish slowed and assumed a southeasterly course which it retained for the remainder of the track. After sunset on the second night, the fish returned close to the island and briefly visited fish aggregating device (FAD) TT off Milolii (Fig. 2). The fish was lost in rapidly deteriorating sea conditions off South Point, Hawaii, after 42 hours of tracking. At this time, the fish appeared to be con- tinuing in a southerly direction. Immediately upon release, this fish descended into the uppermost 2 or 3 degrees of the thermocline, where it remained for 6 hours as it swam steadily offshore. Upward excursions towards the surface started 3 hours before sunset, and movements on the second day were mostly in the thermocline-mixed layer interface with frequent upward movements, some of which reached the surface. As with the other Keahole marlin, night- time distribution was shifted towards the surface, with considerable amounts of time spent at the surface (Fig. 3C). Fish 8903, weighing ~75 kg, was caught and released 3 nmi off Keauhou Bay by a team in the 1989 HIBT. The fight time was ~20 minutes and the fish was live- ly when the transmitter was attached. Immediately upon release, the fish headed offshore for 5.5 hours before turning north (Fig. 2). This initial deviation in direction occurred ~11.5 nmi offshore and coincided with the marlin's first movement to the surface from the top of the thermocline, where it had been since the initial release (Fig. 3D). Compared with other fish tracked in this study, this fish spent large amounts of time swimming slowly very close to the surface, often with the tips of its dorsal and caudal fins protruding above the surface. Although this "lazy" surface behavior is not uncommonly ob- served by fishermen, this was the only fish to demon- strate this behavior in the current study. The track was terminated after 24 hours, with the fish continuing to swim slowly at the surface in a northwesterly direction. Fish 8904, also weighing ~75 kg, was caught and released by an HIBT team at a location ~2 nmi off- shore of Kealakekua Bay, after a 36-minute fight. This fish also ran west (for ~6 hours) to a point 11.5 nmi offshore, at which time contact was lost for ~2 hours. When relocated, the fish had adopted a northwesterly course which it maintained for the remainder of the 29-hour track (Fig. 2). Vertical movements were similar to those of other fish, being largely constrained by the top of the thermocline and the surface, and averaging closer to the surface at night than during the day (Fig. 3E). Temperature and depth distribution Pooling the data from all six fish indicates that 82% of the daytime and 97% of nighttime distribution oc- curred in the mixed layer and top 2 degrees of the ther- mocline. In terms of depth, approximately 36% of the daytime and 60% of nighttime was spent between the surface and 30 m, the rest of the time being spent at greater depths (Fig. 4). Discussion Although the sample size reported here is quite small, these data represent a significant increase in the number of Pacific blue marlin tracked. Also, several aspects of the behavior of these marlin show remark- able consistency. For instance, the behavior of all the fish tracked in this study was strongly influenced by the interface NOTES Holland et al : Capture and release behavior of Pacific blue marlin off Hawaii 401 70-1 60- A 50- 40- -30. 1- 20- _i < 10- O n "XrCC , , I 1 I S^, 1- 0^ 1 1 in o ^°- 1- 60- Z LlJ 50- U ^ 30- rK B 20- 10 0 r^, ;i!i!C , ^^ ,.,,._ , o oooooooo iM n-^inuDr--a3 en 1 1 1 1 1 1 1 1 !^ CNirO'^LDUDr-- oo o o o o O — CM fO o o 1 o 1 1 1 1 1 1 cn o — CM rO 1- DEPTH Nl "EF ?V/ \L (m) Figure 4 Aggregate depth distribution of six Pacific blue marlin off Hawaii during da,v1ime (A) and nighttime (B). between the surface mixed layer and the top few degrees of the thermodine, even though the depth of this interface varied between tracks. Unlike yellowfin tuna Thurinus albMares, which are also associated with this feature (Carey and Olson 1982, Yonemori 1982, Holland et al. In press), the marlin only rarely dove below the top of the thermodine. Thus, a major por- tion of the distribution of these marlin was bracketed by the surface and the bottom of the mixed layer. This behavior— not making deep dives, and spending most time in the mixed layer and upper thermodine— is similar to that of striped marlin tracked off California (Holts and Bedford In press), but differs greatly from that of swordfish which make frequent deep dives into cold water (Carey and Robison 1981). Compared with the striped marlin tracked off California, most of the blue marlin in Hawaii spent a relatively small amount of time very close to the surface. This may be due to the shallower depth of the thermodine off California (15-25 m) compared with Hawaii (35-90 m). With the exception of fish 8903, which spent an atypical amount of daytime very close to the surface, the marlin tracked in the present study moved closer to the surface at night. This differs from the striped marlin data, but is consistent with the behavior reported for skipjack tuna Katsuwonus pelamis (Yuen 1970), swordfish (Carey and Robison 1981), yellowfin tuna (Yonemori 1982, Holland et al. In press), and bigeye tuna T. obesus (Holland et al. In press). Upon release, all six marlin dove into the upper layers of the thermodine and remained there for several hours. Onset of consistent upward excursions from the thermodine appeared to represent the end of the post- capture recovery period, usually between 4 to 6 hours. The recovery period was initiated by comparatively high-speed swimming, with four of the six fish travel- ling farther in the first hour following release than at any other time during the tracks. Because marlin are obligate ram-ventilators, these fish may be swimming fast enough to repay the anaerobic metabolic debt in- curred during the fight, but not so fast as to acquire new debt. The fast speeds of the first hour are even more remarkable considering that at least two of the fish appeared to be completely exhausted when tagged and released. Holts and Bedford (In press) also report deeper than normal depths and heightened activity levels immediately upon the release of striped marlin. With the exception of slow-moving fish 8903, the range of swimming speeds of individual fish in our study (1.2-2.18 kn) are similar to the blue marlin (1.2- 1.9 kn) of Yuen et al. (1974). This is somewhat faster than the striped marlin speeds (0.75-1.45 kn) observed by Holts and Bedford (In press). In the present study, most of the fish swam slightly slower at night than dur- ing the day. One of the most remarkable features of the current study is the consistent direction of movement displayed by the marlin in the first several hours of each track. Even though the three Keahole fish could have moved anywhere within a 240 "-sector without running into land, all three swam along parallel westerly courses that took them directly offshore. Also, initial deviation from these parallel tracks occurred in the same general area ~12 nmi offshore. Similarly, both fish caught off Keauhou swam offshore for '^-■6 hours along almost identical paths, and both made their first major direc- tional changes between 11 and 12 nmi offshore. This point in these two tracks was separated by only ~0.5 nmi, even though the release sites were over 4 nmi apart. The previously longest track of a Pacific blue marlin (22.5 hours; Yuen et al. 1974) was also initiated off Keauhou, and also commenced with a 7 nmi-move- ment offshore, as did the Barbers Point marlin tracked in the current study. These results suggest that direct movement offshore may be a common response to the trauma associated 402 Fishery Bulletin 88(2). 1990 with capture and release. The first major horizontal direction changes often coincide with changes in ver- tical behavior and may represent the end of the recovery period. Previous blue marlin tracks, and those acquired in the present study, share the characteristic of essential- ly straight or slowly curving azimuths. This contrasts with the movements of nearshore tuna (Yuen 1970, Holland et al. In press) and swordfish (Carey and Robi- son 1981) which frequently display cyclical diel move- ments that bring the fish back to their starting points. If Pacific blue marlin have such a pattern in Hawaiian waters, the cycle time for this behavior is longer than any of the tracks so far obtained. At present, the im- pression is that these animals swim along straight or slightly curving courses as they move through the area. It is difficult to explain the difference in tagging mor- tality between this study in which no fish died and the previous one in which three of the five blue marlin died (Yuen et al. 1974). Holts and Bedford (In press) also report that none of their 1 1 striped marlin died as a result of tagging trauma. Their results, and the results reported here, support the practice of releasing marlin caught on artificial lures by sportfishing techniques, even if the fish are apparently exhausted. Acknowledgments This work was supported by the University of Hawaii Sea Grant Program (Ultrasonic telemetry of Horizon- tal and Vertical Movements of Pelagic Fish Associated with FADs, project MR/R-25) under Institutional Grant No. NA85AA-D-SG082 from the NOAA office of Sea Grant, Department of Commerce; the Honolulu Labor- atory, National Marine Fisheries Service, NOAA; and the Hawaii Institute of Marine Biology, University of Hawaii. We also gratefully acknowledge the assistance of Lts.(jg.) Mark Ablondi and Scott Sullivan, NOAA; the Pacific Gamefish Research Foundation; the Hawaii International Billfish Foundation; and the anglers and captains of the 1988 and 1989 Hawaii International Billfish Tournaments. Citations Bayliff, W.H., and K.N. Holland 1986 Materials and methods for tagging tuna and billfishes, recovering the tags, and handling the recapture data. FAG Fish. Tech. Pap. 279. 36 p. Carey, F.G., and R.J. Olson 1982 Sonic tracking experiments with tunas. ICCAT Collec- tive Volume of Scientific Papers XVIII (2):458-466. Int. Comm. Conserv. Atl. Tunas, Madrid. Carey, F.G., and B.H. Robison 1981 Daily patterns in the activities of the swordfish Xiphias gladius, observed by acoustic telemetry. Fish. Bull., U.S. 79:227-292. Holland, K.N., R.W. Brill, S. Ferguson, R.K.C. Chang, and R. Yost 1985 A small vessel technique for tracking pelagic fish. Mar. Fish. Rev. 47(4):26-,32. Holland, K.N., R.W. Brill, and R.K.C. Chang In press Horizontal and vertical movements of yellowfin and liigeye tuna associated with fish aggregating devices. Fish. Bull, U.S. 88(3). Holts, D., and D.W. Bedford In press Activity patterns of striped marlin in the Southern California Bight. In Stroud, R. (ed.). Planning the future of billfishes: Research and management in the 90's and beyond. Part II: Contributed papers. National Coalition for Marine Conservation, Atlanta, GA. Squire, J.L., Jr. 1974 Migration patterns of Istiophoridae in the Pacific Ocean as determined by cooperative tagging programs. In Shomura, R.S.. and F. Wiiliams(eds.), Proc, Int. Billfish Symp., Kailua- Kona, Hawaii, 9-12 Aug. 1972. Pt. 2, Review and contributed papers, p. 226-237. NOAA Tech. Rep. NMFS SSRF-67.''), Natl. Oceanic Atmos. Adm., Natl. Mar. Fish. Serv. Squire, J.L., Jr., and D.V. Nielsen 1983 Results of a tagging program to determine migration rates and patterns for black marlin, Makaini indira, in the southwest Pacific Ocean. NOAA Tech. Rep. NMFS SSRF- 772, Natl. Oceanic Atmos. Adm., Natl. Mar. Fish. Serv., 19 p. Yonemori, T. 1982 Study of tuna behavior, particularly their swimming depths, by use of ultrasonic tags. Far Seas Fish. Res. Lab. (Shimizu) Newsletter 44:1-5. [Engl. Transl. No. 70, T. Otsu, 1982, 7 p. Avail. Honolulu Lab., Southwest Fish. Cent., Natl. Mar. Fish. Serv., NOAA, Honolulu, HI 96822.] Yuen, H.S.H. 1970 Behavior of skipjack tima, Katsuumms pdami^, as deter- mined by tracking with ultrasonic devices. J. Fish. Res, Board Can. 27:2071-2079. Yuen, H.S.H., A.E. Dizon, and J.H. Uchiyama 1974 Notes on the tracking of the Pacific blue marlin Makaini nigricans. In Shomura, R.S., and F. Williams (eds.), Proc, Int Billfish Symp., Kailua-Kona, Hawaii, 19-12 Aug. 1972. Part 2. Review and contributed papers, p. 265-268. NOAA Tech. Rep. NMFS SSRF-675, Natl. Oceanic Atmos. Adm., Natl. Mar. Fish. Serv. Diversity, Abundance, and Spatial Distribution of Fishes and Crustaceans in the Rocl0.05). In contrast to mackerel and spiny dogfish, cunners were very rarely observed to be active at night, indicating they are primarily diur- nal predators in these environments, as previously shown by Chao (1973), 011a et al. (1975), and Dew (1976). Temporal patterns Surface-water temperature of the littoral zone followed a seasonal cycle typical of cold waters of the Gulf of Maine (Fig. 1 A), as shown by Tyler (1971) and MacDon- ald et al. (1984). Maximum values of about 18°C were observed during summer (July-September), and minima of about 1-2°C occurred during winter (Feb- ruary-March). Both the abundance and diversity of fish showed a clear seasonal pattern closely paralleling the seasonal variation of water temperature (Fig. IB). Fish abundance peaked during summer, reaching values of 30-34 fish/hour, and was lowest (near zero) during winter. Species diversity followed a similar seasonal pattern (Fig. IB). These results suggest a seasonal migration (inshore-offshore) of most components of this fish assemblage, which appears to be correlated with temperature changes. 406 Fishery Bulletin 88|2). 1990 A. LOBSTERS 10- I 8 6 /\ \ 1- u 4 f ^ \k 4-^ N LU 2 ^K CO \^ z 0 ' 1 1 1 1 1 — 1 r ■ "t * ■ ■ < CL 1- B. CRABS 16 cc x\ 10- 5- ^•"' / \ ^(^ {^^^ t-i LU m i ^ \ ■* , , 5 ^ C. FISH z 100 80- / — K T 60- / / t\, 40- J \ Y > 20 V \ A-"''^ ^ ♦< , , , 0 J J A 5 O N D J F M A M J J A S O N 1985 1986 Figure 2 Seasonal variations in the abundance of lobsters (A), crabs (B), and benthic fish (C) along bathynietric underwater tran- sects at Pemaquid Point, Maine. Each vertical bar indicates \'l SE. The abundance of major components of the catch- lobsters, crabs, and benthic fishes— also varied sea- sonally (one-way ANOVA, P<0.01 for all species; Fig. 2). Maximum abundances of lobsters occurred in sum- mer (July-September) at densities of 5-8 individuals per transect (~1 adult lobster per 26 m-) (Fig. 2 A). No lobsters were observed in the winter transects (Febru- ary 1986). The seasonal distribution of lobsters close- ly follows the seasonal variation in water temperature (compare Figures lA and 2A). Highest densities of crabs, however, occurred in late fall (November) 1985, with ~14 individuals per transect (~1 adult crab per 12 m"), and during the summer of 1986, though their densities were markedly lower than those observed in 1985 (~5 crabs per transect or ~1 crab per 33 m^) (Fig. 2B). Few crabs were observed in winter; of those observed, most remained inactive and semiburied in the sand. Benthic fish such as cunners, rock gunnels, sculpins, shannys, flounders, and rays were markedly most abundant during fall months, reaching densities of ~75 individuals per transect in 1985 and ~60 in 1986 (Fig. 2C). Their abundance decreased significantly in winter to 2-4 fish per transect, and progressively increased from mid-spring (May- June) until fall as temperature increased. Exhaustive underwater surveys for large predators were carried out in winter (February 1986) with the aid of underwater lights. No cunners or lob- sters were observed along these transects despite the special care taken to examine crevices, caves, and spaces underneath boulders and rocks. In addition to a few crabs, three species of fish were commonly ob- served along these winter transects: rock gunnels, sea ravens, and winter flounders. Rock gunnels were the most abundant fish species, and they were usually ob- served up to 20 m depth, often associated with clumps oi Modiolus modiolus. During the rest of the year, this fish species was rarely found at depths greater than 10 m. This mobile predator assemblage can be divided into distinctive seasonal components. The first is comprised of "summer-fall residents," species consistently pres- ent from late spring until late fall, including pollock, cunners, longhorn and shorthorn sculpins, grubbys. NOTES Ojeda and Dearborn Populations of fishes and crustaceans in the Gulf of Maine 407 < Q O 0 z A. L O B S T E n s c FISH Be R A B S 2 4 6 8 10 12 14 16 18 20 D E P T H (m) D. GUNNERS 2 4 6 8 10 12 14 16 It 20 D E P T H (m) 30 20 O '0 a. Ill OD 0 2 z 1 4 t I 10I2M 1619 20 D E P T H (m) Figure 3 Bathymetric distribution in the abundances of lobsters (A), crabs (B), benthic fish (C), cunners (D), and num- l_ier of shelters (E) in the underwater transects con- ducted at Pemaquid Point, Maine. Each vertical bar indicates ±2 SE. Asterisks in D indicate juvenile indi- viduals. cods, and lobsters and crabs. Lobsters and crabs, how- ever, occurred earlier than these fishes. Crab species are included in this group because their abundance was extremely low during winter, and the few individuals that remained were not active. Summer-fall resident species moved into shallow water or became active (crabs) when the water temperature increased to 8-12°C, generally reaching maximum abundance in midsummer or early fall. The second component was the "regular residents," species captured or observed active at any season, such as sea ravens, rock gunnels, and winter flounders. These species, however, were generally much less abundant than the summer-fall residents (Table 1). The third temporal component of this fauna was the "summer periodicals" (sensu Tyler 1971), and refers to those species that occurred period- ically in the samples during summer. The most abun- dant and conspicuous species in this group were mack- erel, spiny dogfish, bluefish, and, less abundantly, black seabass and yellowtail flounder (Table 1). Finally, there was a group of species that occurred at infrec}uent in- tervals and in low numbers during this study. Fish species belonging to this group have been called "occa- sionals" (Tyler 1971) and include shad, little skate, lumpfish, scup, smelt, Atlantic herring, sUver hake, and tomcod. Spatial patterns The abundances of most groups of mobile predators in- creased with depth, as did the number of shelters (Fig. 3). The number of adult lobsters significantly increased with depth in the first 8 m, then remained relatively constant thereafter to 19 m depth (one-way ANOVA, P<0.01; a posteriori SNK test; Fig. 3A). There was no relationship between depth and lobster size, since both large and small individuals were observed at all depths along the transects. Crab species (mostly Cancer irroratus and C. boralis) showed a somewhat differ- ent bathymetric pattern (Fig. 3B). Their abundance markedly increased between 2 and 5 m depth, remained relatively constant at 5-11 m, and decreased at depths greater than 14 m (one-way ANOVA, P<0.01, a pos- teriori SNK test). The collective abundance of benthic fishes (e.g., cunners, rock gunnels, radiated shannys, sculpins, and flounders) progressively increased along the bathymetric gradient, reaching highest density at the deeper zone (18-20 m depth; one-way ANOVA, P<0.01; a posteriori SNK test) (Fig. 3C). A large pro- portion of these fish, however, were cunners which comprised about 80% of the total number of individuals observed in the transects. Cunners exhibited a distribu- tion pattern similar to the one described for total fish 408 Fishery Bulletin 88(2). 1990 (one-way ANOVA, F<0.01; a posteriori SNK test; Fig. 3D). In contrast to adult cunners, juvenile cunners (<5 cm) comprised more than 90% of the total cunners sighted in the first three depth stations (i.e., 2-8 m depth) along the transects. The number of shelters (a measure of substrate heterogeneity) significantly increased with depth (one- way ANOVA, P<0.01; Fig. 3E). This increment is due to the increasing number of large rounded rocks ob- served at the deeper (10-20 m) transects which create large numbers of interstices (holes) suitable for pred- ators' occurrence. Discussion A diverse assemblage of mobile predators inhabits the shallow rocky subtidal zone at Pemaquid Point, Maine. These subtidal habitats are utUized by these species for multiple purposes such as nursery grounds, feeding grounds, shelter, and reproductive activities (011a et al. 1975, MacDonald et al. 1984, Keats et al. 1987, Ojeda 1987). The occurrence and abundance patterns of these species are, however, markedly seasonal and closely follow the temperature regime typical of northwest Atlantic waters. This suggests that temperature is one of the major abiotic factors affecting the distribution of these predator species on this coast, a finding in agreement with other studies (Tyler 1971; Hacunda 1981; Ennis 1984b; MacDonald et al. 1984; Keats et al. 1986, 1987). Four temporally distinct components were recog- nized in this predator assemblage: summer-fall resi- dents, regular residents, summer periodicals, and occasionals. Similarly, Tyler (1971) distinguished four temporal groups in the demersal fish assemblage oc- curring in Passamaquoddy Bay, New Brunswick, which have also been recognized in other demersal ichthyo- faunas of the Gulf of Maine (Hacunda 1981, MacDon- ald et al. 1984). Three of the four categories of Tyler (1971) were recognized in this study: "regulars," "sum- mer periodicals," and "occasionals." No representa- tives of Tyler's (1971) fourth category ("winter peri- odicals"; i.e., species occurring only during winter) were found in this study. This result, however, should be taken cautiously because no samples were taken in midwinter. The most conspicuous and abundant species of mobile predators in this study were summer-fall residents (e.g., cunners, pollock, lobsters, and sculpins), which occurred for most of the warm period in shallow rocky habitats. Summer-fall residents undergo small- scale migrations, moving into deeper, warmer water (>4°C) during winter and returning to shallow subtidal areas in spring. These seasonal movements occur in response to changes in water temperature and prob- ably to physical disturbances such as strong water surges and storms along some e.xposed coasts. This seasonal migratory behavior represents a behavioral stratety to avoid freezing that involves no physiological adjustment. An alternative strategy involves the elab- oration of macromolecules which possess unique anti- freeze properties (e.g., glycopeptides in sea raven and winterflounder: Duman and De Vries 1974, Slaughter et al. 1981). There are conflicting results in the literature concern- ing winter migration and activity of cunners and lob- sters. Several authors have shown that cunners and lobsters do not migrate into deeper waters during winter (Green and Farwell 1971, Cooper et al. 1975, OUa et al. 1975, Dew 1976). However, investigations have shown that cunners (Chao 1973) and lobsters (En- nis 1984b) move to deeper waters in winter as sug- gested in this study. Recently, Ennis (1984b) showed that small-scale movements of lobsters in Newfound- land were related to increased turbulance due to storms. This is a likely explanation for the absence of lobsters and cunners during winter at Pemaquid Point in this study. As mentioned above, Pemaquid Point is an exposed site (Ojeda 1987). During winter, this area is severely affected by periodic storms and heavy water motion that usually generate strong turbulence over a wide depth range. For organisms that remain in dor- mant states in shallow waters during winter (such as cunners), water movements may severely restrict their distribution. Most studies documenting the presence of cunners in inshore habitats during winter have been conducted in bays or in other protected areas away from heavy water motion (e.g.. Green and Farwell 1971, Olla et al. 1975, Dew 1976). In contrast, docu- mentation of offshore movements of cunners during winter comes from studies conducted on exposed coasts, such as that of Chao (1973). A similar situation seems to occur in lobsters (Ennis 1984b) and probably with crab species which strongly suggests that water movement from turbulence and heavy surge is an im- portant factor, in addition to water temperature, af- fecting the temporal and spatial distribution of large organisms in shallow subtidal environments of the Gulf of Maine. Shallow rocky subtidal habitats along the northern New England coast harbor a diverse community of marine organisms providing abundant food resources to seasonal fish residents, occasional migratory fish species, and large crustacean predators (Ojeda 1987). This is so despite the general paucity of kelp beds in this coast, which are an important determinant of the abundance and diversity of nearshore fish and large decapod crustaceans on other temperate coasts (Quast 1968, Moreno and Jara 1984). In addition to acting as feeding grounds, shallow rocky environments also NOTES Ojeda and Dearborn: Populations of fishes and crustaceans in the Gulf of Maine 409 provide large mobile predators with microhabitats for shelter against predation, and bases for reproductive activities and nursery grounds. Of particular relevance in these environments is the spatial heterogeneity of the bottom. Large rocks and boulders which typically occur at depths greater than 10 m at the study site are important microhabitats for territorial predatory spe- cies such as cunners and lobsters (Pottle and Green 1979, Ennis 1984a). The increased number of lobsters and fish (mostly cunners) observed along the bathy- metric gradient is probably related to the availability of such microhabitats (see Figure 3). As shown else- where (Ojeda and Dearborn 1989), the substrate hetero- geneity at Pemaquid Point progressively increases wnth depth as bottom irregularities such as cracks, crevices, holes, and rocks become more common. In addition to microhabitat availability, the observed bathymetric distribution of large predators may represent an avoid- ance response by these species to strong water turbu- lance and wave surge that commonly affect the shal- lower end of this subtidal zone (especially at low tides). The occurrence in this study of numerous juvenile pollock and cod provides a good example of the impor- tance of shallow, rocky, subtidal zones as nursery grounds as shown by MacDonald et al. (1984) and Keats et al. (1987). Juvenile pollock were the most abundant species of the nearshore fish assemblage (Table 1). These findings, in addition to feeding data presented elsewhere (Ojeda 1987), suggest that pelagic fish species may play important roles in rocky nearshore benthic communities by affecting the distribution and abundance of their benthic prey. Moreover, the one- way offshore migration exhibited by these species indicates that they may be important linkages in the transfer of energy from nearshore to offshore ecosystems. Acknowledgments This paper represents a portion of the Ph.D. disserta- tion submitted by F.P.O. to the Department of Zool- ogy, University of Maine, Orono. We are grateful to Drs. Hugh DeWitt, Bob Elner, Bill Glanz, Fabian Jaksic, Irv Kornfield, and Les Watling, and to Dave Tapley for critically reading this paper. We appreciate the diving-assistance and logistic support given by I. Babb, M. Dunn, P. Garwood, C. Gregory, J. Guy, S. Hacker, D. Knowles, M. Lesser, C. Moody, G. Pod- niesisnski, B. Richardson, K. Scully, D. Tapley, R. Vadas, and R. Wahle. This research was funded by grants from the Migratory Fish Research Institute (MFRI), the Graduate Student Board (GSB), and the Department of Zoology (all of them to F.P.O.) of the University of Maine, Orono. The work of F.P.O. at the University of Maine was funded by an ODE PLAN Chilean Scholarship, by a University Graduate Re- search Fellowship, and by the Center for Marine Studies of the University of Maine. Citations Bigelow, H.B., and W.C. Schroeder 1953 Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 74(vol. .53), .577 p. Breen, P. A., and K.H. Mann 1976 Changing lobster abundance and the destruction of kelp beds by sea urchins. Mar. Biol. 34:137-142. Campbell. A., and A.B. Stasko 1986 Movements of lobsters (Homams americamis) tagged in the Bay of Fundy, Canada. Mar. Biol. 92:393-404. Chao, L.N. 1973 Digestive system and feeding habits of the cunner, Tautogolabrus adspersus a stomachless fish. Fish. Bull., U.S. 71:56.5-586. Cooper, R.A., R.A. Clifford, and CD. Newell 1975 Seasonal abundance of the American lobster, Hoinarus americamw, in the Boothbay region of Maine. Trans. Am. Fish. Soc. 104:669-674. Dew, C.B. 1976 A contribution to the life history of the cunner, Tauto- golabnts adspersus, in Fishers Island Sound, Connecticut. Chesapeake Sci. 17:101-113. Duman, J.R., and A.L. DeVries 1974 Freezing resistance in winter flounder, Pseudopleuro- nectes americanus. Nature (Lond.) 247:237-238. Ennis, G.P. 1984a Territorial behavior of the American lobster Homarus americanus. Trans. Am. Fish. Soc. 113:330-335. 1984b Small-scale seasonal movement of the American lobster Homarus americayius. Trans. Am. Fish. Soc. 113:336-338. Green. J.M.. and M. Farwell 1971 Winter habits of the cunner, Tautogolabrus adspersus (Walbaum 1792), in Newfoundland. Can. J. Zool. 49:1497- 1499. Hacunda, J.S. 1981 Trophic relationships among demersal fishes in a coastal area of the Gulf of Maine. Fish. Bull., U.S. 79:775-788. Keats, D.W.. D.H. Steele, and G.R. South 1986 Atlantic wolfish {Anarhichas lupus L.; Pisces: Anarhichi- dae) predation on green sea urchins {Strongylocentrotus droe- bachiensis (O.F. Mull); Echinodermata: Echinoidea) in eastern Newfoundland. Can. J. Zool. 64:1920-1925. 1987 The role of fleshy macroalgae in the ecology of juvenile cod (Gadus morhua L.) in inshore waters off eastern Newi'ound- land. Can. J. Zool. 65:49-53. MacDonald, J.S.. M.J. Dadswell, R.G. Appy. G.D. Melvin, and D.A. Methven 1984 Fishes, fish assemblages, and their seasonal movements in the lower Bay of Fundy and Passamaquoddy Bay, Canada. Fish. Bull., U.S. 82:121-139. Mann, K.H., and P.A. Breen 1972 The relation between lobster abundance, sea urchins, and kelp beds. J. Fish. Res. Board Can. 29:603-609. Miller, R.J. 1985 Seaweeds, sea urchins, and lobsters; A reappraisal. Can. J. Fish. Aquat. Sci. 42:2061-2072. 410 Fishery Bulletin 88|2). 1990 Moore, I. A., and J.W. Moore 1974 Food of shorthorn sculpin Myoxocephalus scorpius, in the Cumberland Sound area of Baffin Island. J. Fish. Res. Board Can. 31:.355-3.S9. Moreno, C.A., and H.F. Jara 1984 Ecological studies on fish fauna associated with Macrocystis pyrifera belts in south of Fueguian Islands, Chile. Mar. Ecol. Prog. Ser. 15:99-107. Ojeda, F.P. 1987 Rocky subtidal community structure in the Gulf of Maine: The role of mobile predators. Ph.D. thesis, Univ. Maine, Orono, 192 p. Ojeda. F.P., and J.H. Dearborn 1989 Community structure of macroinvertebrates inhabiting the rocky subtidal zone in the Gulf of Maine: Seasonal and bathymetric distributions. Mar. Ecol. Prog. Ser. 57:147-161. Olla, B.L., A.J. Bejda, and A.D. Martin 1975 Activity, movements, and feeding behavior of the cun- ner, Tautogolabrus adspersus and comparison of food habits with young tautog, Tautoga onitis, off Long Island, New York. Fish. Bull., U.S. 73:895-900. Paine, R.T. 1969 The Pisaster-Teguln interaction: Prey patches, predator food preference and intertidal community structure. Ecology 50:9.50-961. Pottle, R.A., and J.M. Green 1979 Territorial behaviour of the north temperate labrid, Tautogolabrus adspersus. Can. J. Zool. 57:2337-2347. Quast, J.C. 1968 Fish fauna of the rocky inshore zone. Calif. Dep. Fish Game, Fish. Bull 139:35-36. Slaughter, D., G.L. Fletcher. V.S. Ananthanarayanan, and C.L. Hew 1981 Antifreeze protein from the sea raven, Hemitripterus am,:riranu». ,]. Biol. Chem. 256:2022-2026. Sokal, R.R., and F.J. Rohlf 1969 Biometry. W.H. Freeman, San Francisco, 776 p. Tyler, A.V. 1971 Periodic and resident components in communities of Atlantic fishes. J. Fish. Res. Board Can. 28:935-946. A Seasonal Autoregresslve Model of the Anchovy Engraulis encrasicolus Fishery In the Eastern Mediterranean Konstantinos I. Stergiou National Centre for Marine Research Agios Kosmas, Hellenikon, Athens 16604. Greece Accurate forecasting is a difficult issue mainly because forecasters are confronted by all sorts of uncertain- ty (in resource production, budgets, political decisions, etc.) (Hilbom 1987). Although not widely employed, Auto- Regressive Integrated Moving Aver- age (ARIMA) stochastic models (Box and Jenkins 1976) have been successful in describing and fore- casting the fishery dynamics of a wide variety of species (lobster: Boudreault et al. 1977, Saila et al. 1979; tuna: Mendelssohn 1981; pil- chard: Stergiou 1989a). Although multivariate, deterministic models (e.g., regression models: Ryan 1986, Koslow et al. 1987) are more com- mon in fishery forecasting, these deterministic models often suffer Figure I Monthly catches of anchovy in Greek waters, January 1964-December 1986. I 2ee6 rv^V^ 2 4 S MONTH <1=JAN, . le 12 Figure 2 Seasonal subseries plot of monthly catches of anchovy in Greek waters, 1964-86. Horizontal lines represent the average catch of each month. Vertical lines are plotted from the average catch to the actual catch of each year of the 1964-86 period. from (1) artificial correlations intro- duced to the data, (2) residual auto- correlation, (3) high residual vari- ance, and (4) colinearity between the independent variables that may bias the fit (Stergiou 1989b, In press). In this note, a seasonal autoregres- slve model is presented that pro- duces forecasts of the monthly purse- seine catches of anchovy Engraulis encrasicolus in Greek waters. The anchovy is one of the most impor- tant pelagic fish, in terms of bio- mass, in the Mediteranean Sea. Mean annual anchovy catch in Greek waters during 1982-85 amounted to 12 820 t, representing 17% of the total marine catch (Stergiou 1989b). Of the anchovy catch in Greek waters, 96% is attributed to the purse-seine fishery, accounting for 26% of the total purse-seine catch (Stergiou 1986a, b). Monthly catches of an- chovy for January 1964-December 1986 (Fig. 1) (National Statistical Service of Greece 1968-88) show a marked seasonal pattern (Figs. 1, 2) and an increasing trend in the vari- ability of monthly catches for the years following 1980. The latter may indicate that anchovy has suf- fered recruitment overfishing in re- cent years (since the late 1970s, purse-seine fishing in Greek waters is anchovy-oriented rather than pil- chard-oriented due to the higher price of anchovy [Stergiou 1986b, 1989a, 1990]). The ARIMA processes (Box and Jenkins 1976, Makridakis et al. 1983) apply to stationary series (time series with no systematic change in mean and variance and free of peri- odic variations). First- or second- order differencing (nonseasonal and/or seasonal) handles problems of nonstationarity in the mean, and logarithmic (or power) transforma- tion of the raw data handles nonsta- tionary variance. The general form of the ARIMA models, Manuscript accepted 28 December 1989. Fishery Bulletin, U.S. 88:411-414. 411 412 Fishery Bulletin 88(2). 1990 ARIMA (j),d,q){P,D,Q)S, can be described by the following equation: {l-niBP){l-N^BP) a - B^iXl - B^>)X, = (l - UiBi)(l - U^BQ)et where X, = the value at time t, Bi' is a backward shift operator that is used as Bt'Xj = X,_f,, rii, N^, Ui, and Ui = arithmetic coefficients, e, = error term at time t, p = order of autoregressive term (AR term), d = degree of differencing involved to achieve stationar- ity (I term), q = order of moving average term (MA term), and 5 = seasonality (number of periods per season); P. D, Q = seasonal terms (corresponding to p,q,d respectively). Based on the examination of autocorrelation and par- tial autocorrelation functions (not shown here) of the original logarithmically transformed series, the follow- ing model was fitted to the logarithms of the raw data, ARIMA (1,1,0)(1, 1,0)12 (Table 1), or, after substituting the autoregressive coefficients and expanding the backward shift operator (Xf = logarithm of raw data), Xt = 0.792X,_i-0.208X,_2 + 0.684X(_i2 -0.862;^,_13-(-0.138X,_i4-h0.336X,_24 -0.266X,_25 -0.07X,.2fi + e,. The unit of time (t) is one month. The model was estimated using the approximate maximum-likelihood algorithm of McLeod and Sales (1983). Parameters were estimated using backcasting with length of 13. The examination of the cumulative periodogram of the residuals (not shown here) (Box and Jenkins 1976) indicated that residuals approximated random noise. Actual catches for January 1985-December 1986, not used in the development of the model, and forecasts for those years are plotted in Figure 3. The coefficient of determination (Table 1) was found to be r- = 0.94 for January 1985-December 1986 (for the untrans- formed series). Mean absolute percentage error (MAPE) (Table 1) for January 1985-December 1986 was 20.4%. Except for March and October 1985 and February, October and November 1986, when the absolute percentage error (APE) was higher than 26% (Table 1), monthly catches were predicted reasonably accurately within an APE range of 2-26% (MAPE of the remaining 19 forecasts was 11%) (Table 1). Table 1 Parameter estimates of the anchovy fishery model. MAPE = mean absolute percentage error, APE = absolute percent- age error, r ^ = coefficient of determination. APE = Absolute (actual catch at time t - forecast at time t ) 100 Actual catch at time t MAPE = mean of APE, and r- = 1 - (variance of residuals)/(variance of variable) Parameter Estimate SE Significance n, -0.208 -0.336 0.067 0.059 p<0.01 p<0.001 r -■ for 1985-86 Residual mean MAPE 1985 MAPE 1986 MAPE 1985-86 APE > 26% March 1985 October 1985 February 1986 October 1986 November 1986 MAPE excluding APE > 26 0.94 0.007 13.3 27.4 20.4 31.5 33.5 47.7 65.9 95.6 11 32 _ ^ 3 ^ \\ c A T 28 _ 26 _ \\ C 24 w H 22 . 2 ^ u N 1 1 8 16 . 1 4 ^ /, w 0 1 2 II \ ' 0 I 1 0 08 _ tl T 0 6 ' // 04 . 0 2 J--,-^ n 1 2 3 4 S 6 7 8 9 10 11 12 13 14 15 16 17 1 B 19 20 21 22 /! ; 4 MONTH 1=J«N 1985 24 = DEC1986 „ ACTUAL , FORECAST Figure 3 Comparison between actual monthly catches (D) of anchovy in Greek waters during January 1985-December 1986 and the forecasts ( + ) estimated from the model. The error in predicting the February-March and October-November catches may be attributed to the highly variable timing of the initiation of the inshore (prespawning migration, which occurs sometime in March) and the offshore migrations (in fall) both of which are expected to be affected greatly by ocean- NOTES Stergiou Forecasting the anchovy fishery in the eastern Mediterranean 413 ographic conditions. Catches exhibit a marked seasonal pattern (Figs. 1, 2) which is hkely related to these seasonal offshore and inshore migrations. Purse-seine fishing activity in Greek waters does not operate in the open sea but is mainly restricted to coastal areas where schools of anchovy migrate on a seasonal basis. The an- chovy starts its inshore migration in early spring, but the peak occurs in coastal waters in May- August and schools disperse again during late summer-fall (Fig. 2; Tsimenidis and Caragitsou 1984). In Greek waters trawling and coastal seining are pro- hibited from the 1st of June to the 30th of September. As a result, the landings of demersal species (e.g., Merluccius merluccius, Micromesistius poutassou, Lophius sp., MuUus sp., Pagellus sp.) are low. This com- bined with the increased demand for fish in summer, drives up fish prices in Greece. In this context, accurate forecasts of the catches of anchovy (and of pelagic fish in general) during June-September are essential for market and industrial planning. APE in June-Septem- ber was <11.2% in 1985 (mean 7.7%) and < 10.6% in 1986 (mean 7.8%). Strict monitoring and accurate forecasts are essen- tial for pelagic fisheries that are heavily dependent on a single year-class, inasmuch as they are prone to col- lapse when conditions for recruitment in a particular year are not favorable and fishing is intense (as for anchovy in Greek waters). In this context, accurate forecasts of the annual anchovy catch together with information related to optimal management of an- chovy estimated by deterministic fishery-management models, can be used by resource managers for the preseasonal adjustment of anchovy fishing mortality. The model produced reasonable forecasts of the annual 1985 and 1986 catches. Total observed annual catches in 1985 and 1986 were 17 544 t and 18 339 t, respec- tively, while the model predicted 17 369 t and 20 210 t (APE 1% and 10%, respectively). It must be pointed out, however, that although the model will most likely produce accurate forecasts if the anchovy stock is under equillibrium, it may fail to produce reliable forecasts for years characterized by weak year-classes of anchovy. (In fact there may be a lag of some months between the occurrence of a turning point and its recognition by the model; for a general discussion on turning points and out-of-sample forecasting, see Schlegel 1985.) The model presented here has also an interesting biological interpretation. The seasonal-difference term of the model indicates the seasonal migratory nature of anchovy in Greek waters. Moreover, climate-plank- ton-anchovy interactions in the Eastern Mediterranean have been found to involve time lags of 2 to 3 years (Pucher-Petkovic et al. 1971). Yet, a 2-3 year cycle has been identified in the variability of different biotic (zooplankton, phytoplankton, fish) and abiotic (air and sea temperature, salinity and air pressure) components of the Eastern Mediterranean/Black Sea ecosystem (Polli 1955; Regner and Gacic 1974; S. Regner 1982, 1985; D. Regner 1985; Dement'Eva 1987; Petrova- Karadjova and Apostolov 1988). The autoregressive terms (X,_24, -Y,_2.5, Xt_2(,) of the model seem to be consistent with these biological/oceanographical observations. Citations Boudreault, F.R., J.N. DuPont, and C. Sylvain 1977 Modeles lineaires de prediction des debarquement de homard aux Iles-de-la-Madeleine (Golfe du Saint-Laurent). J. Fish. Res. Board Can. 34:379-383 [in French. Engl, abstr.]. Box, G.E.P.. and G.M. Jenkins 1976 Time series analysis, forecasting and control. Holden- Day, San Francisco, 575 p. Dement'Eva, T.F. 1987 A method for correlation of environmental factors and year-class strength of fishes. J. Icthyol. 27:55-59. Hilborn, R. 1987 Living with uncertainty in resource management. N. Am. J. Fish. Manage. 7:1-5. Koslow, A.J., K.R. Thompson, and W. Silvert 1987 Recruitment to Northwest Atlantic cod (Gadus morhua) and haddock {Melanogramrrms aeglefiniis) stocks: Influence of stock size and climate. Can. J. Fish. Aquat. Sci. 44:26-39. Makridakis, S., S. Wheelwright, and V. McGee 1983 Forecasting: Methods and applications. John Wiley. NY. 926 p. McLeod, A.I., and P.R.H. Sales 1983 An algorithm for approximate likelihood calculation of ARMA and seasonal ARM A models. Algorithm AS 191. Appl. Stat. 1983:211-223. Mendelssohn, R. 1981 Using Box-Jenkins models to forecast fishery dynamics: Identification, estimation and checking. Fish. Bull., U.S. 78: 887-896. National Statistical Service of Greece 1968-88 Results of the sea fishery survey by motor vessels. Section G, 20 issues (for the years 1964-86), Athens, Greece. Petrova-Karadjova, V.J., and E.M. Apostolov 1988 Influence of solar activity upon the diatoms of Black Sea plankton. Rapp. Comni. Int. Mer Medit. 31, p. 224. Polli. S. 1955 I cicli cHmatici di 5.6 e 8.0 anni e la loro realta' fisca. Riv. Meteor. Aeronautica 2. 12 p. Pucher-Petkovic, T., M. Zore-Armanda, and I. Kacic 1971 Primary and secondary production of the Middle Adriatic in relation to climatic factors. Thalassia Jugosl. 7:301-311. Regner, D. 1985 Seasonal and multiannual dynamics of copepods in the middle Adriatic Sea. Acta Adriat. 26:11-99. Regner, S. 1982 Investigations of qualitative and quantitative composi- tion of the larval fish stages in the plankton at the high sea of the middle Adriatic. Stud. Mar. Fauna 11/12:45-60. 1985 Ecology of planktonic stages of anchovy, Engraulis en- crasicolus (Linnaus, 1758), in the central Adriatic. Acta Adriat. 26:5-113. 414 Fishery Bulletin 88(2). 1990 Regner, S., and J. Gacic 1974 The fluctuation of sardine catch along the eastern Adriatic coast and solar activity. Acta Adriat. 15, 15 p. Ryan, P.M. 1986 Prediction of angler success in an Atlantic salmon, Salmon salar, fishery two fishing seasons in advance. Can. J. Fish. Aquat. Sci. 43:2531-2534. Saila. S.B., M. Wigbout, and R.J. Lermit 1979 Comparison of some time series models for the analysis of fisheries data. J. Cons. Cons. Int. Explor. Mer 39:44-52. Schlegel, G. 1985 Vector autoregressive forecasts of recession and recov- ery: Is less more? Econ. Rev. IIQ/1985:2-12. Stergiou, K.I. 1986a On the anchovy and pilchard fishery in Greek waters, 1964-1982. Rapp. Comm. Int. Mer Medit. 30, p. 241. 1986b On the assessment of the pelagic fishery resources in Greek waters. Rapp. Comm. Int. Mer Medit. 30, p. 241. 1989a Modeling and forecasting the fishery of pilchard, Sar- dina pilchardits in Greek waters using ARIMA time series models. J. Cons. Cons. Int. Explor. Mer 46:16-23. 1989b Multivariate analysis and trends in Greek fishery. Fishbyte 7(2):4-7. ICLARM, Manila. 1990. On the Greek fishery production. Greek Fishing News 103:31-38 [in Greek]. In press On different statistical methods for the analysis of fishery time series. Proc. 4th Conf., Hellenic Soc. Ichthyol., .June 1988 [in Greek, Engl, abstr.j. Tsimenidis, N., and H. Caragitsou 1984 The state of the sardine and anchovy resources in Greek Seas. Proc. 1st Hellenic Symp. Ocean. Fish., p. 578-589. Natl. Cent. Mar. Res., Athens [in Greek]. Tethering as a Technique for Assessing Predation l?ates in Different Habitats: An Evaluation using Juvenile Lobsters Homarus amehcanus Diana E. Barshaw Rutgers University Marine Field Station Great Bay Boulevard. Tuckerton. New Jersey 08087 Present address: Center for Maritime Studies University of Haifa. Mount Carmel. Haifa 31999, Israel Kenneth W. Able Rutgers University Marine Field Station Great Bay Boulevard. Tuckerton. New Jersey 08087 Tethering has been used successful- ly to assess predation rates of a variety of predator-prey systems in several different habitats. The ma- jority of these experiments have used tethered crabs as prey (Heck and Thoman 1981, Wilson 1985, Wilson et al. 1987, Heck and WOson 1988, Wilson et. al. 1990, Barshaw and Able In press). Fish have also been tethered in different habitats; however, in these experiments the tethered fish could not act natural- ly, and therefore the technique only showed the differences in predator encounter rate in different habitats (Shulman 1985, Mclvor and Odum 1988). Other organisms are present- ly being used in tethering experi- ments including molluscs (R.N. Lip- cius and L.S. Marshall, Jr., Coll. William and Mary, Va. Inst. Mar. Sci., Gloucester Ft., VA 23062, un- publ. data) and spiny lobsters (Herrnkind and Butler 1986). We determined if tethering was an appropriate technique to assess predation on species that burrow (i.e., juvenile lobsters Homarus americanus). Lobsters were chosen for this study, in part, because their behavior has been well studied and, therefore, a basehne of "normal" behaviors is available (Botero and Atema 1982, Barshaw and Bryant- Rich 1988). Lobsters use different methods of constructing burrows in different habitats; therefore we tested three habitats known to be important for early juvenile lobsters: mud, cobble, and Spartina peat (Able et al. 1988, Barshaw and Lavalli 1988). Methods and materials Six "ant farm" aquaria (10 cm wide, 30 cm long, 45 cm deep) were 2/3 filled with either cohesive mud, cob- ble of a natural size distribution, or Spartina peat substrates (two repli- cates per substrate type) and pro- vided with running, unfiltered sea- water. Early juvenile lobsters (8-14 mm carapace length) were individu- ally tethered to monofilament line using "super glue" to attach it to their carapace. Individual tethered lobsters were placed into half the tanks while similar-sized untethered control lobsters were placed into the remaining tanks. A discrete reading of each lob- ster's behavior was recorded every Reference to trade names does not imply en- dorsement by the National Marine Fisheries Service, NOAA. minute for the first 20 minutes, every 5 minutes for the next 35 minutes, and then every hour for 6 hours. A final assessment of each lobster's burrow was made after 24 hours. Therefore, each lobster was observed 33 times over 24 hours in each test. Seven tests were run using all the substrates, with two extra tests run only with mud; thus observations were made on a total of 14 lobsters in cobble, 14 in peat, and 26 in mud. The behaviors ob- served are described in Table 1. The behaviors of the tethered and untethered lobsters were compared by calculating the percent of the 33 observations in which the lobsters were engaged in each of the behav- iors for each test. Since this experi- ment was designed only to compare tethered and untethered lobsters, no comparisons were made between different behaviors or between dif- ferent substrates (comparisons of that nature were studied in Bar- shaw and Bryant-Rich 1988). The percent of observations was trans- formed using the arc-sign trans- formation, and analyzed with a Stu- dent's t test. The numbers of the tethered and untethered lobsters that had burrows at the end of the experiment were analyzed for each substrate using 2x2 contingency tables and chi-square tests. Results The tethering of juvenile lobsters resulted in substrate-specific dif- ferences in behavior and the ability to construct burrows. There were no significant differences between the behavior of tethered and unteth- ered lobsters in the peat or cobble substrates (Fig. 1); in both peat and cobble, all the tethered (7/7) and all the untethered (7/7) lobsters had burrows that they constructed and maintained throughout the experi- ment. Manuscript accepted 22 January 1990. Fishery Bulletin, U.S. 88:415-4:7. 415 416 Fishery Bulletin 88(2). 1990 Table 1 Description of behaviors of tethered and untethered juvenile Homarus aviericanus observed in this experiment. Behavior Description Rest (RST) No movement for at least 30 seconds. Groom- ing was not considered movement and was not recorded separately from RST. Pleopod fan (PPF) Movement of the pleopods which caused sedi- ment to be moved; i.e., was being used to build or repair the burrow. In this study PPF which only moved water was not recorded. Bulldoze (BLD) Pushing sediment forward with the claws and first walking legs spread apart. Dig Loosening sediment by pushing claws into it. Walk (WLK) Walking on the sediment; does not include "walking" in the burrow. Swim Swimming in the water column. In the mud substrate, however, the tethered lobsters walked significantly more and bulldozed significantly less than the untethered lobsters {t test, p<0.05. Fig. 1). (There was a trend in the tethered lobsters to bulldoze less in all of the substrates). Also, in the mud substrate most tethered lobsters were unable to make or maintain a burrow. Only 7.7% (2/13) of the lobsters had constructed a complete burrow, deep enough to hide the lobster at the end of the 24-hour period. This result was significantly different from the untethered group in which 69.2% (9/13) had constructed burrows by the end of the experiment (chi-square test, /j<0.05). Discussion These observations demonstrate a substrate-specific difference in the effect of tethering on the burrowing behavior of juvenile lobsters. While tethering did not substantially effect the behavior or burrow construc- tion of lobsters in peat or cobble, lobsters tethered in mud spent more time walking, less time bulldozing, and were unable to build and maintain a burrow as well as the untethered lobsters. The tethered lobsters in mud would therefore be more vulnerable to predation than the untethered controls, since being without a burrow has been demonstrated to increase mortality due to predation (Barshaw and Lavalli 1988). Heck and Thoman (1981), Heck and Wilson (1988), and Barshaw and Able (In press) conducted predator-prey experi- ments using tethered crabs as the prey. Neopanopi sayi, Panopeus herbsti, Libinia diibia, Ovalipefs ocella- tus, and Callinectes sapidus were tethered using similar techniques as were used in our study. These investi- z o l- < > cr LU (/) m O u. O LU O cr 111 Q. (f) < > CO UJ en m O u. O UJ o tr 111 a. (/) < > LU (/) m O UJ o DC UJ CL lOOn 90 80 70 60 50 40 30 20 loi 0 100- 90- 80 70 60 50 40 30 20- 10- 100 90- 80- 70 60- 50- 40- 30- 20- 10 0 MUD ■ NOT TETHERED 0 TETHERED WLK BLD COBBLE PPF RST ■ NOT TETHERED E3 TETHERED PPF RST ■ NOT TETHERED E3 TETHERED WLK BLD PPF RST Figure 1 Average percent of the 33 observations in which lobsters were ob- served doing each behavior in each substrate. WLK = walk. BLD = bulldoze, PPF = pleopod fan, RST = rest. NOTES Barshaw and Able: Effect of tethering on predatlon rates of Homarus amencanus 417 gators did not observe any differences in the beiiavior of crabs tethered in different substrates. The differ- ences in tethering effects between these investigations and ours are probably due to the fact that Homarus americanus construct burrows in the substrate while the crabs used in the above tethering experiments simply bury themselves directly into the substrate. The results of this study suggest that tethering to assess predation in different habitats should be evalu- ated for each new species under consideration because species-specific behavior patterns could create habitat- specific tethering artifacts. In particular, caution should be used when interpreting survival rates of lobsters in mud relative to other substrates. Similar caution should be considered when tethering is used to assess predation in other burrowing forms. Acknowledgments We thank Donald Bryant-Rich, Dan Roelke, and Steve Weiss for help in collecting substrates and making observations. Mike Syslo and Kevin Johnson from the Martha's Vineyard State Lobster Hatchery and Kari Lavalli provided us with juvenile lobsters. Support for this study was provided by a minigrant from the New Jersey Marine Sciences Consortium. Citations Able, W.K., K.L. Heck, Jr., M.P. Fahay, and C.T. Roman 1988 Use of salt-marsh peat reefs by small juvenile lobsters on Cape Cod, Massachusetts. Estuaries 11:83-86. Barshaw, D.E., and K.W. Able In press Deep burial as a refuge for lady crabs Ovalipes ocellatus: Comparisons with blue crabs Callinectes sapidus. Mar. Ecol. Prog. Ser. Barshaw, D.E., and D.R. Bryant-Rich 1988 A long-term study on the behavior and survival of early juvenile American lobster, Homarus americanus, in three naturalistic substrates: Eelgrass, mud, and rocks. Fish. Bull., U.S. 86:789-796. Barshaw, D.E., and K.L. Lavalli 1988 Predation upon postlarval lobsters Homarus americanus by cunners Tautogokibrus adspersus and mud crabs Neopamtpi sayi on three different substrates: Eelgrass, mud and rocks. Mar. Ecol. Prog. Ser. 48:119-123. Botero, L., and J. Atema 1982 Behavior and substrate selection during larval settling in the lobster, Homarus americanus. J. Crustacean Biol. 2:59-69. Heck, K.L., Jr., and T.A. Thoman 1981 Experiments on predator-prey interactions in vegetated aquatic habitats. J. Exp. Mar. Biol. Ecol. 53:125-134. Heck, K.L., Jr., and K. A. Wilson 1988 Predation rates on decapod crustaceans in latitudinally separated seagrass communities: A study of spatial and tem- poral variation using tethering techniques. J. Exp. Mar. Biol. Ecol. 107:87-100. Herrnkind, W.F., and M.J. Butler IV 1986 Factors regulating postlarval settlement and juvenile microhabitat use by spiny lobsters Panulirus argus. Mar. Ecol. Prog. Ser. 34:23-30. Mclvor, C.C, and W.E. Odum 1988 Food, predation risk and micro-habitat selection in a marsh fish assemblage. Ecology 69:1341-1351. Shulman, M.J. 1985 Recruitment of coral reef fishes: Effects of distribution of predators and shelter. Ecology 66:1056-1066. Wilson, K.A. 1985 Physical and biological interactions that influence habitat use of mangrove crabs. Ph.D. diss., Univ. Pennsylvania, Philadelphia. 187 p. Wilson, K.A., K.L. Heck, Jr., and K.W. Able 1987 Juvenile blue crab, Callinectes sapidus, survival: An evaluation of eelgrass, Zostera marina, as refuge. Fish. Bull., U.S. 85:53-58. Wilson, K.A., K.W. Able, and K. L. Heck, Jr. 1 990 Predation rates on juvenile blue crabs in estuarine nursery habitats: Evidence for the importance of macroalgae (Ulva lac- tuca). Mar. Ecol. Prog. Ser. 58(3):243-251. The National Marine Fisheries Service (NMFS) does not approve, recommend or en dorse any propnetary product or proprietary material mentioned in this publication 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 ap proves, recommends or endorses any proprietary product or proprietary material mentioned herein, or which has as its purpose an intent to cause directly or indirect ly the advertised product to be us«d or purchased because of this NMFS publication U.S. Department of Commerce Seattle, Washington IMOAA Technical Reports IMMFS Series Recent publications in the NOAA Technical Report Series Some NOAA publications are avail- able by purchase from the Superin- tendent of Documents, U.S. Govern- ment Printing Office, Washington, DC 20402. 77 78 79 80 81 82 83 84 Keppner, Edwin J., and Armen C. Tarjan Illustrated key to the genera of free-living marine nematodes of the order Enoplida July 1989, 26 p., 118 figs. Pearson, Donald E. Survey of fishes and water properties of south San Francisco Bay, California, 1973-82, August 1989, 21 p , 8 figs. Wenner, Charles A., and George R. Sedberry Species composition, distriPution, and relative abundance of fishes in the coastal habitat off the southeastern United States. July 1 989, 49 p., 35 figs, Matarese, Ann C, Arthur W. Kendall, Jr., Deborah A/I. Blood, and Beverly M. Vinter Laboratory guide to early life history stages of northeast Pacific fishes. October 1989, 652 p., 272 figs. Estrella, Bruce T., and Daniel J. McKiernan Catch-per-unit-effort and biological parameters from the Massachusetts coastal lobster [Homarus amencanus] resource. Description and trends. September 1989, 21 p., II figs. Shaffer, Rosalie Vaught, and Eugene L. Nakamura Synopsis of biological data on the cobia Rachycentron canadum (Pisces: Rachycentridae) December 1989, 21 p., 8 figs. Fiscus, Clifford H., Dale W. Rice, and Allen A. Wolman Cephalopods from the stomachs of sperm whales taken off California. December 1989, 12 p., 2 figs. Ahrenholz, Dean W., James F. Guthrie, and Charles W. Krouse Results of abundance surveys of juvenile Atlantic and gulf menhaden, Brevoortia tyrannus and B patronus. December 1 989, 1 4 p., 16 figs. 0 5 4U Fishery Bulletin Guidelines for Contributors Form of manuscript The document should be in the following sequence: Title Page, Abstract (not re- quired for Note), Tbxt, Acknowledgments, Citations, Text footnotes. Appendices, Tables, Figure legends, and Figures. Title page should include authors' full names and mailing addresses and the senior author's telephone and FAX numbers. Abstract Not to exceed one double- spaced typed page. Should include a sentence or two explaining to the gen- eral reader why the research was under- taken and why the results should be viewed as important. Abstract should convey the main point of the paper and outline the results or conclusions. No footnotes or references. Text A brief introduction should por- tray the broad significance of the paper. The entire text should be intelligible to readers from different disciplines. All technical terms should be defined, as well as all abbreviations, acronyms, and symbols in text, equations, or formulae. Abbreviate units of measure only when used with numerals or in tables and figures to conserve space. Measure- ments should be expressed in metric units, with other equivalent units given in parentheses. Follow the U.S. Govern- ment Printing Office Style Manual, 1984 ed., and the CBE Style Manual, 5th ed. Fishery and invertebrate no- menclature should follow the American Fisheries Society Special Publication 12 (for fishes), 16 (for mollusks), and ^^(for decapod crustaceans). Text footnotes should be numbered in Arabic numerals and typed on a sep- arate sheet from the text. Footnotes are not used for reference material or per- sonal communications, but rather to explain or define terms in the text and for contribution numbers on the title page. Informal sources Personal communi- cations, unpublished data, and untitled manuscripts in preparation are noted parenthetically in the text (full name. affiliation, brief address including zip code, and month and year when appro- priate). Acknowledgments Gather all ac- knowledgments into a brief statement at the end of the text. Give credit only for exceptional contributions and not to those whose contributions are part of their normal duties. Citations All titled sources should be listed in the Citations section, including unpublished and processed material. In text, cite as Smith and Jones (1977) or (Smith and Jones 1977); if more than one citation, list chronologically (Smith 1936, Jones 1975, Doe In press). All sources cited in the text should be listed alphabetically by the senior authors' surnames under the heading CITA- TIONS. Abbreviations of periodicals and serials should conform to Serial Sources for the BIOSIS Data Base'". Indicate whether som-ces are in a langu- age other than English. For informal literature, include address of author or publisher. Authors are responsible for the accuracy of all citations. Tables should supplement, not dupli- cate, the text. Each table should be numbered and cited consecutively, with headings short but amply descriptive so that the reader need not refer to the text. For values less than one, zeros should precede all decimal points. In- dicate units of measure in column head- ings;'do not deviate from the unit of measure within a column. Table foot- notes should be noted consecutively in Roman letters across the page from left to right and then down. Since all tables are typeset, they need not be submitted camera-ready. Figures Photographs and line draw- ings should be of professional quality- clear and concise— and reducible to 42 picas for full-page width or to 20 picas for a single-column width, and to a max- imum 55 picas high. All graphic ele- ments in illustrations must be propor- tioned to insure legibility when reduced to fit the page format. Line weight and lettering should be sharp and even. Let- tering should be upper and lower case. and vertical lettering should be avoided whenever possible (except for vertical, y, axis). Zeros should precede all deci- mal points for values less than one. Re- productions of line artwork are accepted in the form of high-quality photographic prints from negatives or photomechan- ical transfer (PMT). Halftones should be sharply focused with good contrast. Micron rules should be inserted on elec- tron micrographs, even when magnifi- cation is included in the figure legend. There should be clear distinction be- tween identifying letters (press-on or overlay) and background of photograph. Label each figure in pencil on the back. Send only xerox copies of figures to the Scientific Editor; originals or photo- graphic prints will be requested later when the manuscript is accepted for publication. Copyright Government publications are in the public domain, i.e., they are not protected by copyright. Submission of manuscript Oisl> E E < z cc 3 ? .1 u o 3 14 12 10L /l At hatch r- i 3 k ^"y 2 1 a 0 ^ 10 10 -Hi 10 .4 .3 .2 .1 0 15 12 9 6 3 0 At first feeding 10 10 10 10 Figure 1 Standard length, yolk volume, and biochemical composition of Pseudopleuronectes nmericanus larvae at hatching and first feeding. O Tsp = 2°C; ▲ Tsp = 7°C. 424 Fishery Bulletin 88(3). 1990 Table 5 Analysis of variance for factors affecting size and chemical content o{ Pseudvpleuronectes americanus larvae at hatch and first feeding. Tsp = adult acclimation and spawning temperature; Tine = embryo and larval incubation temperature; T x T = temperature inter- action; Block = female tag number. *P<0.05, **P^0.01, ***P<0.001. Where the temperature interaction is significant. F values for Tsp and Tine are not given. Tsp Tine TxT Block Total Larval length (mm) df F value 1 At Hatch 2 2 25.17*** 4 13.66*** 139 Yolk volume (mnr') df F value 1 0.00 2 12.26*** 2 0.05 4 2.40* 139 RNA (Mg/mg) df F value 1 4.68* 2 15. 94*** 2 0.18 4 10.16*** 49 DNA (Hg/mg) df F value 1 2 2 24.15*** 4 8.41*** 49 Protein (^ig/mg) df F value 1 0.00 2 1.11 1 0.01 4 15. 59*** 36 Larval length (mm) df F value At First Feeding 1 2 2 5.07** 2 8.08*** 169 RNA (^ig/mg) df F value 1 0.27 2 1.38 2 3.13 4 1.90 39 Protein (jjg/mg) df F value 1 2 9 6.32** 4 11. 93*** 34 are independent class variables (SAS Institute 1985). A is the acclimation and spawning temperature (Tsp); B is the embryo and larval incubation temperature (Tine); C is the block effect or individual female tag number (Block). This maternal factor is nested within the acclimation temperature and represents the vari- ability among individual spawning females. This fac- tor was added to the model to block out fish-to-fish variability. It was significant for all dependent vari- ables tested (Table 5). A *B is the interaction of acclima- tion and incubation temperatures. Significant interactions between spawning and in- cubation temperature were observed at hatch for length and DNA content, and at first feeding for length, RNA, and protein content. In these cases, where the effects of spawning temperature and incuba- tion temperature were nonadditive, simple main effects were tested at each spawning and incubation tempera- ture using Tukey's studentized range test. Where no temperature interaction was observed (yolk volume, RNA, and protein content at hatch, and DNA content at first feeding) main-effects means were calculated and differences among main-effects means were tested for significance using Tukey's studentized range test. At hatching, standard length showed a significant temperature interaction (Table 5). At a spawning temperature of 2°C, standard length decreased with increasing incubation temperature (Table 3). At a spawning temperature of 7°C, standard length was higher at an incubation temperature of 7°C than at 4°C. Yolk volume showed no significant temperature interaction, was unaffected by spawning temperature, and was highest at the lowest incubation temperature of 4°C. RNA content at hatch showed no significant temperature interaction and decreased with both in- creasing spawning and incubation temperature. DNA content at hatch showed a significant temperature in- teraction. At a spawning temperature of 2°C, DNA content was lowest at the intermediate incubation temperature (7°C). At a spawning temperature of 7°C, DNA content was highest at this intermediate incuba- tion temperature. Protein content at hatch showed no temperature interaction and was unaffected by either spawning or incubation temperature. At first feeding, standard length showed a signifi- cant temperature interaction. At both spawning tem- peratures (2° and 7°C) standard length decreased with increasing incubation temperature (Table 4). RNA con- tent at first feeding showed a significant temperature interaction. At a spawning temperature of 2°C, RNA content was lowest at the highest incubation tempera- ture (10°C). At a spawning temperature of 7°C, RNA Buckley et al Effects of water temperature on Pseudopleuronectes amencanus 425 content was lowest at the intermediate incubation tem- perature (7°C). DNA content at first feeding showed no significant temperature interaction and was unaf- fected by either spawning or incubation temperature. Protein content at first feeding showed a significant temperature interaction. At both spawning tempera- tures, protein content decreased with increasing in- cubation temperature. Discussion Winter flounder spawn and embryos survive over a wide range of temperatures. In the laboratory, Wil- liams (1975) observed survival of embryos between - 1.8° and 15°C. Rogers (1976) reported survival be- tween 3° and 14°C. While 2°C is close to the temper- ature of maximum egg production, winter flounder larvae require water temperatures above 2°C for sur- vival to metamorphosis (Laurence 1975, Buckley et al. 1982). Spawned through the late winter and early spring, winter flounder embryos and larvae generally experience gradually increasing water temperatures. In shallow estuaries and bays, however, embryos and larvae may be subjected to large fluctuations in salin- ity and water temperature over relatively short periods of time, or to prolonged periods of abnormal warming or cooling. The effects of such changes in water tem- perature on survival and growth of the early-life stages of winter flounder are largely unknown. Inverse relations between water temperature (dur- ing late winter and early spring) and indices of recruit- ment have been reported for winter flounder (Jeffries and Johnson 1974, Jeffries and Terceiro 1985, North- east Utilities 1988). The shape of larval abundance curves, a parameter affected by water temperature, has also been related to recruitment; cold years with broad shallow abundance curves produce good year- classes (Northeast Utilities 1988). The size and chemical composition of larvae at initia- tion of feeding provide useful criteria for evaluation of these temperature effects, and may provide insight into causal mechanisms. Feeding initiation is the end point of the period of reliance on endogenous energy reserves that commences at ovulation. Increased lar- val size confers increased potential for survival, since larger larvae are better able to avoid size-dependent predation, capture food, and survive periods of star- vation (Blaxter and Hempel 1966, Rosenberg and Haugen 1982, Bailey and Battey 1984, Knutsen and Tilseth 1985). DNA content is an index of cell number and RNA content an index of the capacity for protein synthesis and hence growth (Bulow 1987). Buckley et al. (In prep.) demonstrated direct relations between size and RNA content of yolksac winter flounder larvae and survival for the first month of life in the laboratory. Size of larvae at first feeding is dependent upon egg size at spawning and the efficiency of production of lar- val tissue (yolk-conversion efficiency). Based on observations of the timing of yolk absorb- tion and first feeding, and the size and composition of winter larvae produced by a single female spawned at 2°C, Buckley (1982) speculated that eggs spawned at low temperatures (2°) may be poorly suited for growth and survival at high temperatures (10°C) within the range of tolerance of winter flounder. The present fac- torial study of six females spawned at low (2°C) and high (7°C) temperatures confirmed and extended these earlier findings. However, unlike the earlier study, first feeding occurred within one day of completion of yolk absorbtion at all combinations of adult acclimation (prespawning) and incubation temperature (Table 2). The present study, with the spawn from six winter flounders, demonstrated the importance of variability at the level of individual females in determining size and composition of winter flounder larvae (Tables 3 and 4). Adult acclimation (prespawning) temperature was important alone or in combination with embryo incuba- tion temperature in determining length and RNA and DNA content at hatch, and in determining length and RNA and protein content at first feeding (Table 5). Em- bryo incubation temperature was important alone or in combination with adult acclimation temperature in determining larval size, yolksac volume, RNA and DNA content at hatch, and in determining length and RNA and protein content at first feeding. At first feeding, size and chemical content of eggs spawned by adults acclimated to 2°C were maximized at the lowest in- cubation temperature (4°C). Eggs spawned at 7°C pro- duced the longest first-feeding larvae at 4°C, while DNA and protein content were highest at the inter- mediate incubation temperature (7°C) and RNA con- tent was highest at the warmest incubation tempera- ture (10°C). The largest larvae, whether measured by length or chemical content, were produced at the low- est combination of acclimation and rearing tempera- tures. RNA content, which is critical to protein syn- thesis and growth, was highest at fn-st feeding in larvae incubated at 4°C for eggs produced at 2°C, and at 10°C for eggs produced at 7°C. RNA content was lowest in first-feeding larvae produced from eggs spawned at 2°C and incubated at 10°C. This group represented the largest difference between spawning and incubation temperature. Hempel and Blaxter (1967) reported differences in fecundity and egg size between Atlantic herring stocks spawning at different temperatures. Stocks spawning at colder temperatures produced fewer but larger eggs. Tanasichuk and Ware (1987) working with Pacific her- ring found that both size-specific fecundity and egg 426 Fishery Bulletin 88(3), 1990 size, but not size-specific ovary weight, were related to seawater temperature 60-90 days before spawning. Again, fecundity increased with increasing tempera- ture while egg size decreased. Results from this study demonstrate that water temperatures during the latter stages of gamete maturation (48-51 days prior to spawning) and during embiyo and larval development affect the size and com- position of winter flounder larvae produced. Further, the interaction between acclimation temperature of adults and incubation temperature of embryos and lar- vae also appears to have a strong effect on size and composition of larvae. While the number of oocytes entering vitellogenesis is probably determined earlier in the reproductive cycle (Brown 1957, Dunn 1970, Tyler and Dunn 1976, Burton and Idler 1984), several important functions occur during the latter stages of gamete development, including further deposition of yolk and final meiotic division. Our data on size and chemical composition of winter flounder at hatching and first feeding suggest a more complex relation between water temperature and lar- val size than observed between water temperature and egg size. These data may help explain some of the variability observed in the relation between yolk-con- version efficiency or maximum larval size and incuba- tion temperature (Sweet and Kinne 1964, Alderdice and Forrester 1968, Laurence and Rogers 1976, Linden et al. 1980, Johns et al. 1981, Laurence and Howell 1981, Buckley 1982). Our data suggest that lar- val size and composition at hatch and first feeding are dependent not only upon incubation temperature but also upon water temperature during the final stages of gamete maturation and upon the interaction of water temperature during these two time-periods. This im- plies that the contents of the egg are modified in some way in response to water temperature prior to spawn- ing. Most likely, this temperature response goes beyond simply producing a larger or smaller egg with com- ponents in the same proportion. The ratio of major organic components including protein, lipids, carbo- hydrates, and nucleic acids may be altered in response to temperature. More subtle, but possibly more signifi- cant, changes in the composition of the developing oocyte in response to water temperature may include alterations in the content, composition, activity, or stability of enzymes, hormones, maternal messenger RNA, and stable RNA (tRNA and rRNA). Any conclu- sions about the efficiency of yolk utilization at different incubation temperatures should take into account the thermal history of the spawning adults. Winter flounder reproductive strategy has most like- ly evolved to exploit the dramatic increase in water temperature during the spnng in the shallow estuaries along the northwest margin of the Atlantic Ocean. Between spawning and metamorphosis, a period of about 2 months in winter flounder, water temperatures warm an average 10°C (from ~2° to 12°C) affecting not only the abundance and composition of predators and prey but also the rates of metabolic processes (Laurence 1975, 1977). This maximizes size and condi- tion of first-feeding larvae at low temperatures (Table 4), allows relatively long resistance to starvation at in- termediate temperatures, and facilitates rapid larval and postlarval growth at higher temperatures (Buckley 1982). Mortality of late embryos and larvae at cold temperatures (<2°C) (Laurence 1975, Buckley 1982) indicates that good survival of winter flounder is de- pendent upon the expected spring warming. While gametogenesis and embryo and larval development show a wide range of temperature tolerance in winter flounder, these processes appear to have been opti- mized for cold winter temperatures followed by gradual spring warming. Our data suggest that cold winters followed by gradual spring warming favor good sur- vival and recruitment of winter flounder by facilitating production of the largest larvae at first feeding (high standard length and DNA content) in the best condi- tion (high RNA and protein content). These data may explain in part the observed correlation between cold years and strong year-classes (Jeffries and Johnson 1977, Jeffries and Terceiro 1985, Northeast Utilities 1988). Citations Alderdice, D.F., and C.R. Forrester 1968 Some effect.s of salinity ami teni|ieraturt' on early develop- ment and .survival of the English sole (Pa roph rys pentuhis). J. Fish. Res. Board Can. 25:49.5-521. Bagenal, T.L. 1971 The interrelation of the size of fish eggs, the date of spawning and the production cycle. J. Fish. Biol. 3:207-219. Bailey, K.M.. and R.S. Batty 1984 Lal_H)ratory study of predation by .4 urclin aurita on lar- vae of cod, flounder, plaice and herring: development and vulnerability to capture. Mar. Biol. (Berl.) 8.3:287-291. Blaxter, J.H.S., and G. Hempel 1966 Utilization of yolk by herring larvae. J. Mar. Biol. Assoc, U.K. 46:219-234. Brett, J.R. 1970 Temperature— fishes, hi Kinne, 0. (ed.), Marine ecology, vol. 1, Environmental factors, p. 515-560. Wiley, London. Brown, M.E. 1957 Experimental studies on gi'owth. /n Brown. M.E. (ed.). The physiology of fishes, chap. 9, p. 361-400. Acad. Press, NY. Buckley, L.J. 1979 Relations between RNA-DNA ratio, prey density, and growth rate in Atlantic cod (Gadus morhua) larvae. J. Fish. Res. Board Can. 36:1497-1502. 1982 Effects of temperature on growth and biochemical com- position of larval winter flounder Pxeudupleurimectes nmeri- ainus. Mar. Ecol. Prog. Ser. 8:181-186. Buckley et al Effects of water temperature ori Pseudopleuronectes amencanus 427 Buckley. L.J., T.A. Halavik, A.S. Smigielski, G.C. Laurence, and E.M. Calderone In prep. Variability in size, biochemical composition, and sur- vival of winter flounder (Pseudopleuronectes americanus) lar- vae reared in the laboratory. Narragansett Lab., Northeast Fish. Cent., Natl. Mar. Fish. Serv.. NOAA, Narragansett, RI 02882-1199. Bulow. F.J. 1987 RNA-DNA ratios as indicators of growth in fish. In Summerfelt. R.C., and G.H. Hall (eds.). Age and growth offish, p. 4,5-64. Iowa State Univ. Press, Ames. Burton, M.P.. and D.R. Idler 1984 The reproductive cycle in winter flounder, Pseudopleuro- nectes americanus (Walbaum). Can. J. Zool. 62:2563-2567. Gushing. D.H. 1967 The grouping of herring populations. J. Mar. Biol. Assoc. U.K. 47:193-208. Dunn, R.S. 1970 Further evidence for a three-year oocyte maturation time in the winter flounder {Pseudopleuronectes americanus). J. Fish. Res. Board Can. 27:957-960. Folkvord, A., and J.R. Hunter 1986 Size-specilic vulnerability of northern anchovy, Engraulis mordax, larvae to predation by fishes. Fish. Bull., L'.S. 84: 859-869. Heming, T.A., and R.K. Buddington 1988 Yolk absorbtion in embryonic and larval fishes. In Hoar, W.S., and D.J. Randall (eds') Fish physiology. Vol. 11, The physiology of developing fish. Part A, Eggs and larvae, p. 408-446. Acad. Press, NY. Hempel, G., and J.H.S. Blaxter 1967 Egg weight in Atlantic herring (Clupea ha rengus L.). J. Cons. Perm. Int. E.x'plor. Mer 31:170-195. Jeffries, H.P.. and W.C. Johnson 1974 Seasonal distributions of bottom fishes in the Nar- ragansett Bay area: Seven-year variations in the abundances of winter flounder (Pseudopleuronectes a7ne7Hcanus). ,1. Fish. Res. Board Can. 31:1057-1066. Jeffries. H.P.. and M. Terceiro 1985 Cycle of changing abundances in the fishes of the Nar- ragansett Bay area. Mar. Ecol. Prog. Ser. 25:239-244. Johns, D.M., W.H. Howell, and G. Klein-MaePhee 1981 Yolk utilization and growth to the yolk-sac absorption in the summer flounder (Paralichthys doitatks) larvae at con- stant and cyclic temperatures. Mar. Biol. (Berl.) 63:301-308. Knutsen, G.M., and S. Tilseth 1985 Growth, development, and feeding success of Atlantic cod larvae Gadus morhua related to egg size. Trans. Am. Fish. Soc. 114:507-511. Lasker, R. 1962 Efficiency and rate of yolk utilization by developing em- bryos and larvae of the Pacific sardine, Sardinops caerula (Girard). J. Fish. Res. Board Can. 19:867-875. Laurence, G.C. 1973 Influence of temperature on energy utilization of em- bryonic and prolarval tautog. Tautoga onitis. J. Fish. Res. Board Can. 30:435-442. 1975 Laboratory growth and metabolism of winter flounder. Pseudopleuronectes americanus, from hatching through meta- morphosis at three temperatures. Mar. Biol. (Berl.) 32: 223-229. 1977 A bioenergetic model for the analysis of feeding and sur- vival potential of winter flounder. Pseudopleuronectes ameri- canus, larvae during the period from hatching to metamor- phosis. Fish. Bull.. U.S. 75:529-546. Laurence. G.C. and W.H. Howell 1981 Embryology and influence of temperature and salinity on early development and survival of yellowtail flounder Limanda fen-ugmea. Mar. Ecol. Prog. Ser. 6:11-18. Laurence, G.C, and CA. Rogers 1976 Effects of temperature and salinity on comparative em- bryo development and mortality of cod (Gadus morhua L.) and haddock (Melanogrammus aeglefinus [L.]). J. Cons. Cons. Int. Explor. Mer 63:220-228. Linden. O.. R. Laughlin Jr.. J.R. Sharp, and J.M. Neff 1980 The combined effects of salinity, temperature and oil on the growth pattern of embryos of the killifish, Fundulus heteroclitus Walbaum. Mar. Environ. Res. 3:126-144. May, R.C 1974 Effects of temperature on yolk utilization in Bairdiella icistia (Jordan & Gilbert) (Pisces: Sciaenidae). J. Exp. Mar. Biol. Ecol. 16:213-225. Miller, T.J., L.B. Crowder, J.A. Rice, and E.A. Marschall 1988 Larval size and recruitment mechanisms in fishes: Toward a conceptual framework. Can. J. Fish. Aquat. Sci. 45:1657-1670. Northeast Utilities 1987 Monitoring the marine environment of Long Island Sound at Millstone Nuclear Power Station. Waterford, CT. Summary of studies prior to Unit 3 operation. Northeast LItilities, En- vironmental Lab., Waterford, CT. 151 p. 1988 Monitoring the marine environment of Long Island Sound at the Millstone Nuclear Power Station. Waterford. Connec- ticut. Three-unit operational studies 1986-1987. Northeast Utilities, Environmental Lab., Waterford, CT. 355 p. Rogers, CA. 1976 Effects of temperature and salinity on the survival of winter flounder embryos. Fish. Bull.. U.S. 74:52-58. Rosenberg. A. A., and A.S. Haugen 1982 Individual growth and size-selective mortality of larval turbot (Sco)}hthahnus nuuimus) reared in enclosures. Mar. Biol. (Berl.) 72:7.3-77. SAS Institute Inc. 1985 SAS/STAT guide for personal computers, Version 6 Edi- tion. SAS Institute Inc., Gary, NC, 378 p. Sissenwine, M.P. 1984 Why do fish populations vary? /m May. R. (ed.). Work- shop on e,xploitation of marine communities, p. 59-94. Springer- Verlag, Berlin. Smigielski, A.S. 1975 Hormonal-induced ovulation of winter flounder. Pseudopleui'07U'ctes americanus. Fish. Bull., U.S. 73:431-438. Smigielski, A.S., and CR. Arnold 1972 Separating and incubating winter flounder eggs. Prog. Fish-Cult. 34:113. Southward. A. J., and N. Demir 1974 Seasonal changes in dimensions and viability of the devel- oping eggs of the Cornish pilchard (Sardinia pilchardus Walbaum) off Plymouth. In Blaxter. J.H.S. (ed.). The early life history of fish. p. 53-68. Springer- Verlag, NY. Sweet, J.G., and O. Kinne 1964 The effects of various temperature-salinity combinations on the body form of newly hatched Cyprinidon macularis (Teleostei). Helgol. Wiss. Meeresunters. 11:49-69. Tanasichuk, R.W., and D.M. Ware 1987 Influence of interannual variations in water temperature on fecundity and egg size in Pacific herring (Clupen harengus pallasi). Can. J. Fish. Aquat. Sci. 44:1485-1495. 428 Fishery Bulletin 88(3), 1990 Tyler, A.V., and R.S. Dunn 1976 Ration, growth, and measures of somatic and organ con- dition in relation to meal frequency in winter flounder, Pseudo- plerironectes americanus, with hypotheses regarding popula- tion homeostasis. J. Fish. Res. Board Can. 33:63-75. Ware, D.M. 1975 Relation between egg size, growth, and natural mortali- ty of larval fish. J. Fish. Res. Board Can. 32:2503-2.512. Williams, G.C. 1975 Viable embryogenesis of the winter flounder Pseudo- pleuronect.es nmericanu.50%. Greig et al. (1977) reported a cadmium con- centration of 1.4 mg/kg, a copper concentration of 175.0 mg/kg, a mer- cury concentration of 0.7 mg/kg, and a lead concentration of 110.0 mg/kg in the sediment. NOAA (1988) re- ported a total PCB concentration of 255.5 /.(g/kg and a total PAH concen- 429 430 Fishery Bulletin 88(3). 1990 Connecticut New Haven. MKforc -4r30 -4roo' Figure 1 Long Island Sound stations sampled for Scophthalniu,'^ nqmnnts: Station 1, off Shoreham, NY, 5.5 m; Station 2, off Milford, CT. 15.2 m; Station 3, Hempstead Harbor, NY. 10. T m. tration of 5366 i^glkg in the sediments of Hempstead Harbor. The windowpane flounder Scophthalrmis aquosus was chosen as a test animal because it is one of the few species available in the Long Island Sound area throughout the year. The possibility of some movement between stations is a consideration in this type of study. However, Moore (1947) reported on the basis of tag- ging studies that, although individuals have been shown to move over large distances, no seasonal migration of windowpane flounder is indicated and that this species is relatively stationary. This study utilized blood chemistry as an indication of stress in the fish. Several investigators have sug- gested that the study of blood may ultimately be as useful in assessing the health of fish as it is in diagnos- ing human health (Blaxhall and Daisley 1973, Hickey 1976). Fish blood chemistry has not received the same critical study as has human (Wedemeyer and Yasutake 1977). A given species of fish in its normal habitat is likely to be subjected to a wide range of natural condi- tions, any of which may be reflected in its blood chem- istry. Johansson-Sjobeck et al. (1975) noted that hematocrit increased and hemoglobin decreased in eels subjected to starvation, both returning to control levels within 164 days. DeWilde and Houston (1967) reported an increase in hematocrit and hemoglobin with an in- crease in acclimation temperature in rainbow trout; the degree of response depended on season. Effects of heavy-metal exposure on fish blood have been demon- strated (Christensen et al. 1972; Calabrese et al. 1975; Dawson 1979, 1982). Within the same species, fresh- water-adapted fish often exhibit blood chemistry which is different from that of fish adapted to seawater (Cour- tois 1976). Snieszko (1960) emphasized the need to con- sider two sets of standards in using hemotological methods in hatcheries: one general standard for the species, and a second standard that determines a nor- mal value for the parameter at a particular hatchery. Although the objective of the present study was to obtain information on pollutant-related stress, it was necessary to consider seasonal changes as well in order to distinguish between the effects of pollutants and those that represent a normal response to changing natural conditions. Although seasonal changes in fish blood have received little attention in the literature, such changes have been documented in the winter flounder (Bridges et al. 1976), the striped bass (Loch- miller et al. 1989) and the rainbow trout (DeWilde and Houston 1967). In addition, the values for hematologi- cal measurements vary from species to species, but few species have received sufficient study to allow a real- istic estimate of their normal ranges under a reasonable variety of conditions. The field study described here was supplemented by a series of three laboratory exposures of the same species to heavy metals. Although laboratory exposures are subject to criticism on the grounds that they do not correspond closely to natural conditions, they do make it possible to attribute changes to a particular pollu- tant, which is not often possible in the field. Materials and methods Field study Windowpane flounder for the field study were collected by otter trawl using 30-minute tows at the three desig- Dawson Blood chemistry of Scophthalmus aquosus in Long Island Sound 431 nated stations and held in running seawater aboard tiie boat until blood samples were taken, generally within 1 hour after the fish were collected. On occasion, be- cause of inclement weather, it was necessary to return to the laboratory and sample the fish at dockside; in that case, the interval between capture and sampling was up to 4 hours. The stations were sampled monthly unless inclement weather or boat repairs precluded sampling. General- ly, blood samples were taken from 20 fish at each sta- tion; if fewer fish were obtained, sample numbers were necessarily smaller. Of 102 collections used, 67 con- sisted of 20 fish each, 33 of 10-19 fish, and one each of 8 and 9 fish. Respective lengths of fish sampled at Station 1, 2, and 3 were 26.7 ± 0.13, 25.0 ± 0.16, and 25.6 ±0.15 cm. Hematocrit (Hct), hemoglobin (Hb), plasma osmolality, sodium, potassium, and calcium were measured on each blood sample. Prior to each fish collection, a bottom-water sample was collected for measurement of temperature, salinity, and dissolved oxygen. Four preliminary collections were made at each station in 1979; intensive sampling continued from May 1980 through April 1983. Exposure studies Windowpane flounder were exposed to mercury, cop- per, or cadmium in the laboratory. Fish used in the exposure studies were collected by otter trawl using 15- or 30-minute tows in the vicinity of Milford, Con- necticut, and transported to the laboratory in running seawater. They were held at the laboratory for at least 2 weeks prior to exposure. In general, exposures were conducted during the colder months when the fish do not normally feed much. However, they were given small amounts of minced surf clam Spisula solidissima weekly. The fish measured 24.8 + 0.21 cm in length and 158.4 ± 4.2 g in weight. Three 60-day exposures were performed: the first used HgClo at a nominal mercury concentration of 5 or 10 Mg/L. the second used CuClo-2H20 at a nominal copper concentration of 10 or 20 ^g/L, and the third used CdCl2-2V2H20 at a nominal cadmium concentra- tion of 5 or 10 Mg/L. Measured concentrations were 5.1 + 0.60 ppb and 12.7 ± 2.20 ppb for low and high cad- mium concentrations, and 19.2 ±1.4 for the high cop- per concentration. Backgi'ound metal levels were below 1 uglL for mercury, below 5 juglL for copper, and below 1 (lig/L for cadmium. The fish were exposed in 285-L fiberglass tanks filled to 225 L with sand-filtered Mil- ford Harbor seawater by a proportional dilution ap- paratus (Mount and Brungs 1967). The diluter con- trolled the intermittent delivery of toxicant-containing or control seawater at a flowrate of 1.5 L every 2.5 minutes throughout the exposure period. This provided a flow of 864 L per day and an estimated 90% replace- ment time of 15 hours (Sprague 1969). The seawater was at ambient salinity during the exposures. The temperature was slightly above ambient because the water was held in the heated building during filtration and delivery. Duration of the mercury exposure was 22 December 1979-20 Febraary 1980. The salinity range diunng that time was 25.0-26.9"Ain. Temperature in the exposure tanks was 8°C at the beginning of the exposure, drop- ping to 3-4 °C throughout most of the month of Febru- ary when the exposure was completed. During the mer- cury exposure, each tank held 5 fish for a total of 20 fish per concentration and 20 controls. The copper exposure ran from 7 January through 8 March 1983. Salinity during the exposure period was 24.4-26.4"/i«i. Temperature was 9°C at the beginning of exposure, dropped to a low point of 6°C in mid- February, and gradually rose to 9°C at the end. Three tanks at each concentration held 4 fish per tank. Duration of the cadmium exposure was 27 January- 28 March 1984. Salinity range during that time was 22.6-25.9'7o„. Temperature was 6.5°C at the begin- ning of the exposure, rising to 8.5 °C at the end. The lowest temperature recorded was 5.0 °C on 3 February. Twenty fish were exposed per concentration, 5 fish per tank. Blood chemistry The parameters measured were hematocrit (Hct), hemoglobin (Hb), mean corpuscular hemoglobin concen- tration (MCHC), plasma osmolality, sodium, potassium, and calcium. Erythrocyte counts (RBC) and calcula- tions of mean corpuscular hemoglobin (MCH) and mean corpuscular volume (MCV) were performed only on animals used in exposures. Blood was collected from each animal by cardiac puncture using a 3-mL plastic syringe and a 20- or 22-gauge needle. The sample was transferred gently into an 8-mL glass vial containing 150 units of dried ammonium heparinate as an anticoagulant. Immediately following collection of the last blood sample, a portion of each blood sample was centri- fuged at 12 000 g and the plasma frozen for later determination of osmolality, sodium, potassium, and calcium. The remaining whole-blood sample was used for the determination of hemoglobin, hematocrit, and erythrocyte counts. Hemoglobin was determined by the cyanmethemoglobin method using Hycel chemicals. Microhematocrits were determined following cen- trifugation for 5 minutes at 13 500 g. Erythrocyte counts were made in a hemacytometer; blood samples were diluted 1:200 with Yokoyoma's solution (Katz 1950). Plasma osmolality was determined with an 432 Fishery Bulletin 88(3). 1990 Table 1 Effects of 60-day exposure to mercury on blood of windowpane flounder Smphthabnus aquosus in Long Island Sound. Values are mean ± standard error, with number of samples in parentheses. Number of samples is often fewer than that of fish exposed because of clotting or pooling of samples. Control Mercury concentration Test 5fig/L 10 pig/L Hematocrit (%) 22.00 ± L3 (17) 25.00 ±1.4 (15) 24.00 ±1.6 (18) Hemoglobin (g/100 mL) 4.10 ±0.2 (17) 4.30 ±0.3 (15) 3.40 ±0.2 (18) Erythrocytes (10'' cells/mm^) 2.14 ±0.17 (9) 2.13 ±0.14 (9) 2.41 ±0.29 (10) MCV (fi^/cell) 96.60 ± 5.7 (9) 109.80 ± 4.0 (9) 103.90 ±9.1 (10) MCH (pg/cell) 20.20 ±1.6 (9) 21.10 ±1.2 (9) 18.90 ±1.2 (10) MCHC (g/100 mL) 19.10 ±1.0 (17) 18.40 ±0.6 (15) 18.10 ±0.5 (18) Na (mEq/L) 179.00 ±3.9 (11) 181.00 ±3.5 (14) ♦199.00 + 3.3 (13) K (mEq/L) 8.21 ±0.85 (11) 6.93 ±0.66 (14) 8.52 ±0.81 (13) Ca (mEq/L) 4.95 ±0.23 (11) •3.91 ±0.18 (14) •3.60 + 0.19(13) Osmolality (mOsm) 338.00 ± 7.7 (9) 328.00 ±4.0 (14) 331.00 ±2.9 (12) MCV = Mean corpuscular volume MCH = Mean corpuscular hemoglobin MCHC = Mean corpuscular hemoglobin concentration •Significantly different from controls (P<0.05). Advanced 3L osmometer; the effect of the added hepa- rin on osmolahty was negligible. Plasma sodium, potas- sium, and calcium concentrations were measured with a Coleman 51 flame photometer. At times it was neces- sary to pool the small plasma samples, resulting in slightly fewer samples for plasma ions and osmolality. Mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), and mean corpuscular hemoglobin concentration (MCHC) were determined using the following equations (Piatt 1969): MCV in M'Vcell = 10 Hct/RBC MCH in pg/cell = 10 Hb/RBC MCHC in g/00 mL packed red cells 100 Hb/Hct Statistical differences were determined by one-way analysis of variance and the Scheffe test (Hicks 1973). Results Exposure studies Results of the 60-day exposure of windowpane flounder to mercury are summarized in Table 1. Red blood cells in the species appear to be unaffected by mercury at the test concentrations used in this study. Control flounder had a mean hematocrit of 22 + 1.3%, a mean hemoglobin of 4.1 ± 0.2 g/100 mL, and a red cell count Reference to trade names does not imply endorsement by the Na- tional Marine Fisheries Service. NOAA. of 2.4 + 0.17 X 10'' cells/mnr^ Following a 60-day ex- posure of the fish to 5 or 10 ^g/L mercury, none of these variables differed signficantly from control values, nor did the indices calculated from the three variables (namely, MCH, MCV, and MCHC). The mean plasma sodium level in flounders exposed to 10 uglL mercury was 199 + 3.3 mEq/L, significantly higher than the con- trol value of 179 ± 3.9 mEq/L. Plasma calcium de- creased from the control value of 4.95 mEq/L to 3.91 ± 0.18 and 3.60 + 0.19 m/Eq/L in animals exposed to 5 and 10 fxgIL mercury, respectively. There were no significant differences between controls and mercury- exposed flounder in plasma potassium or osmolality (Table 1). No significant difference was demonstrated between controls and animals exposed to either concentra- tion of cadmium or copper in any of the parameters measured. Field study The data from each season were divided by sex and compared using the Scheffe test; no significant sex- related differences were noted. The data were grouped into four size-classes (<20.0 cm, 20.1-25 cm, 25.1-30.0 cm, >30 cm) and compared for each season; no signifi- cant size-related differences were noted during any season. Because of the lack of differences, data from both sexes and all size-classes are shown together. Overall, windowpane flounder taken from Station 3 had significantly higher hematocrits and hemoglobin concentrations than did those taken from Station 1. Dawson Blood chemistry of Scophthalmus aquosus in Long Island Sound 433 Table 2 Hematocrits of windowpane flounder Scophthalmus aquosus in error, with number of fish in parentheses. Long I sland Sound. Values are percent ± standard Station Winter Spring Summer Fall 1 "■''20.61 0.9 (78) "23.210.3(172) 2 ''22.1 ± 0.4 (105) 24.0 ± 0.3 (182) 3 ''23.2 10.6 (93) "■"26.6 10.4(178) given tion. •'24.910.4 (140) 24.6 1 0.5 (132) •'25.4 ±0.4 (136) season. 23.8 ±0.4 (139) 23.9 1 0.5 (137) "25.9 1 0.4 (139) "Significantly different (P<0.05) from all other stations in a ''Significantly different from all other seasons at a given sta ' ''Significantly different pairs of seasons at a given station. Table 3 Hemoglobin concentrations of windo-wpane flounder Scophthalmus aquosus in Long Island Sound Units are g/100 1 mL 1 standard error, -with number of fish sampled in parentheses. No significant seasonal variations were demon- strated. Values from Station 1 are signif cantly lower than those from Stations 2 and 3; however, no contrasts between stations were significant when the data were separated by season. Station Winter Spring Summer Fall 1 3.5 1 0.13 (89) 3.6 ±0.11 (173) 4.2 ±0.11 (164) 3.6 ±0.12(133) O 4.0 ±0.13 (113) 4.0 1 0.11 (188) 4.0 ±0.12 (145) 3.8 lO.ll (149) 3 3.8 ±0.14 (84) 3.9 ±0.10 (169) 4.1 lO.lO (153) 4.2 1 0.14 (1.50) Table 4 Seasonal variations in blood chemistry of windowpane flounder Scophthalmus aquosus in Long Island Sound. Values are mean ± standard error. Because of the lack of significant differences between stations in these variables, data from all stations are combined. Variable Winter Spring Summer Fall MCHC (g/100 mL) Sodium (mEq/L) Potassium (mEq/L) Calcium (mEq/L) 17.00 1 0.04 (256) "186.00 1 0.4 (285) "4.10 ±0.18 (285) 4.22 ±0.12 (285) liar hemoglobin concentration. pairs. pairs. 15.60 1 0.03 (508) 170.001 0.3 (515) ''5.16 ±0.14 (515) "4.43 ±0.10(515) 15.30 ± 0.04 (408) 14.40 i 0.04 (419) "157.0010.3 (409) "157.0010.3 (405) "■"5.48 ± 0.15 (409) ''4.25 i 0.16 (405) 3.79 ±0.10 (409) "3.4910.11(405) MCHC = Mean corpusc " ''Significantly different '■''Significantly different When the data from each station were separated by season, the difference between Stations 1 and 3 in hematocrit was significant during the winter, spring, and fall (Table 2). When the data on hemoglobin levels were separated by season, the difference between sta- tions for any particular season was not significant (Table 3). Variations in hemoglobin from season to season were not significant (Table 3). At each station, the hematocrits of flounders collected during the winter were significantly lower than those of fish collected at any other season (Table 2). Seasonal means in MCHC ranged from 14.4 to 17.0. There were no significant differences either between stations or between seasons (Table 4). Plasma sodium levels varied by season but not by sta- tion. Because of the lack of difference between stations, the data from the three stations were treated together. The mean sodium value for winter-collected animals at all stations was 186 + 0.4 mEq/L; this is significantly higher than the summer and fall values, both of which were 157 ± 0.3 mEq/L (Table 4). Plasma potassium levels varied significantly by season but not by station. The highest levels were found in the summer. The mean value for summer-collected animals of 5.48 + 0.15 mEq/L was significantly higher than either the winter value of 4.10 + 0.18 mEq/L or the fall value of 4.25 ± 0.16 mEq/L (Table 4). 434 Fishery Bulletin 88(3), 1990 Table 5 Pkisma osmolalities of windowpane f\o\mder Scophthalmus iiqiitufiix in Long error, with number of fish in parentheses. Within each station, groups underlined. Island Sound. Units are mOsm ± standard which are not significantly different are Station Winter Spring Summer Fall 1 2 3 *392 ± 5 (89) *387 + 2 (167) 390 ± 2 (129) *401 ±2(124) *389±3(118) 377 ± 2 (109) 377 ± 2 (170) 393 ±3 (113) 380 ± 2 (87) 374 ± 2 (176) •383 ± 2 (134) •377 ± 2 (132) different (f - X O 10- o lU > o ^ CO 0- CO 10 STATION 1 — I — I — I — I — STATION 2 — I 1 — I — I r — r STATION 3 J F —I — M — 1 — M -1 1 — I — r— A S O N MONTH Figure 3 Yearly cycle of dissolved o.xygen concentrations at the three stations in Long Island Sound. Each point is the mean of all samples taken from a given month and station throughout the study. bass exhibited a significant increase in sodium and a decrease in calcium following mercury exposure, and winter flounder also had a drop in plasma calcium con- centration (Dawson 1979, 1982). The lesser effects of mercury on the window])ane flounder suggest that this fish is generally less susceptible than striped bass or winter flounder or that the distribution of the metal within the animal differs from species to species. Pen- treath (1976) reported that both gill and kidney ac- cumulated high concentrations of mercury during the exposure of plaice Pleuronectes platessa L. to this metal. Accumulation in the gill and kidney would be likely to affect plasma chemistry; kidney accumulation may affect renal hematopoiesis as well. This study produced no significant changes in win- dowpane flounder blood following copper exposure. 436 Fishery Bulletin 88(3). 1990 Other investigators have reported hematological ef- fects of copper exposure in various fish species. Christensen et. al. (1972) reported significant increases in hematocrit and hemoglobin in the brown bullhead Ictalurus nebulosus following copper exposure. The ex- posure levels ranged from 27 to 107 t^glL, compared with our 10 and 20 i^glL, and the changes were noted in the fish exposed to 40-107 ^g/L copper. It is possi- ble that the lack of copper effects in the present study simply reflects a low exposure level. The results of the cadmium exposure were similar to those obtained in an earlier exposure of the winter flounder Pseudopleuronectes americanus, to cadmium in this laboratory (Calabrese et al. 1975). In the earlier study, winter flounder were exposed for 60 days to 5 or 10 piglL Cd, as were the windowpane flounder in the present study. No hematological changes were ob- served in winter flounder, although Na, K, and Ca were not measured. Larsson (1975) reported decreased hematocrits and hemoglobin concentrations following a 9- week exposure of the flounder Pleuronectes Jlesus L. to concentrations of cadmium as low as 5 /jg/L. The same study demonstrated decreased calcium and potassium, but these were observed only at higher Cd concentrations. Both the laboratory exposures and the field-sampling portion of this study suggest that the windowpane flounder is a hardy animal. The higher hematocrit and hemoglobin at the most polluted station compared with the least polluted station suggest an increase in hematopoiesis. The plasma osmolality, normally well below the osmolality of the surrounding water, was lower in fish collected from the most polluted station than in those collected from the cleanest station, in- dicating no loss of osmoregulatory ability. The limited effects of cadmium, copper and mercury exposures in- dicate that the species is not particularly vulnerable to these pollutants. These facts taken together suggest that the windowpane flounder at the most polluted sta- tion were subject to a stress, perhaps pollutant-induced, sufficient to challenge their metabolism, but a stress to which they were capable of adjusting. Acknowledgments The author thanks S. Schurman, D. Kapareiko, and D. Tucker for technical assistance. Citations Blaxhall, P.C., and K.W. Daisley 197.3 Routine heniatolo^cal methods for use with fish blood. J. Fish Biol. ,5:771-781. Bowman. M.J. 1976 Tidal locks across the East River: An engineering solu- tion to the rehabilitation of western Long Island Sound. In Wiley, M. (ed.). Estuarine processes, Vol. I. Uses, stresses, and adaptation to the estuary, p. 28-43. Acad. Press, NY. Bridges, D.W., J.J. Cech Jr., and D.N. Pedro 1976 Seasonal hematological changes in winter flounder, Pseudopleuronectes americanus. Trans. Am. Fish. Soc. 105: 596-600. Calabrese, A., P.P. Thurberg, M.A. Dawson, and D.R. Wenzloff 1975 Sublethal physiological stress induced by cadmium and mercury in the winter floun(ier, Pseudopleuronectes ameri- canm. In Koeman, J.H., and J.J.T.W.A. Strik (eds.). Sublethal effects of toxic chemicals on aquatic animals, p. 15-21. American Elsevier. NY. Christensen, G.M., J.M. McKim, W.A. Brungs, and E.P. Hunt 1972 Changes in the blood of the brown bullhead Ictalurus nebulosus (Lesueur) following short and long term exposure to copper (II). Toxicol. Appl. Pharmacol. 23:417-427. Chytalo, K.N. 1979 PCBs in dredged materials .ind benthic organisms in Long Island Sound. M.S. thesis. State Univ. New York at Stony Brook, 109 p. Courtois, L.A. 1976 Hematology of juvenile striped bass. Morone saxatilis (Walbaum), acclimated to different environmental con- ditons. Comp. Biochem. Physiol. 54A:221-223. Dawson, M.A. 1979 Hematological effects of long-term mercury ex-posure and subsequent periods of recovery of the winter flounder, Pseudfi- pleuronwtes americanus. In Vernberg. W.B., F.P. Tliurberg, A. Calabrese, and F.J. Vernberg (eds.). Marine pollution: Func- tional responses, p. 171-182. Acad. Press, NY. 1982 Effects of long-term mercury exposure on hematology of striped bass, Morone saj-atili.^. Fish. Bull.. U.S. 80:389-392. DeWilde, M.A., and A.H. Houston 1967 Hematological aspects of the thermoacclimatory process in the rainbow trout, Sahno gairdneri. J. Fish. Res. Board C,in. 24:2267-2281. Greig, R.A.. R.N. Reid. and D.R. Wenzloff 1977 Trace metal concentrations in sediments from Long Island Sound. Mar. Pollut. Bull. 8:183-188. Hickey. C.R. Jr. 1976 Fish hematology, its uses and significance. N.Y. Fi.sh Game J. 23:170-175. Hicks, C.R. 1973 Fundamental concepts in the design of experiments. Holt. Rinehart, and Winston, NY. Hunt, CD. 1979 The role of phytoplankton and particulate organic car- lion in trace metal deposition in Long Island Sound. Ph.D. thesis, Univ. Connecticut, Storrs, 265 p. Interstate Sanitation Commission llnpubl. Long Island Sound estuarine study historical data analysis for toxic contaminants in the water column. Unpub- lished report for the U.S. Environ. Prot. Agency, Office of Water. OMEP. Wash. DC, Agreement no. CX812842. In- sterstate Sanitation Comm.. 311 W. 43d. NY 10036. 68 p. Johansson-Sjobeck, M-L., G. Dave, A. Larsson, K. Lewander, and J. Lidman 1975 Metabolic and hematological effects of starvation in the European eel. Anffuilla ani^uiUa L. — II. Hematology. Comp. Biochem. Physiol. 52A:431-4.34. Dawson: Blood chemistry of Scophthalmus aquosus in Long Island Sound 437 Katz, M. 1950 The number of erythrocytes in the blood of silver salmon. Trans. Am. Fish. Soc. 80:184-193. Larsson, A. 1975 Some biochemical effects of cadmium on fish. In Koeman. J.H., and J.J.T.W.A. Strik (eds.). Sublethal effects of toxic chemicals on aquatic animals, p. 3-13. American Elsevier, NY. Lochmiller, R.L., J.D. Weichman, and A.V. Zale 1989 Hematological assessment of temperature and oxygen stress in a reservoir population of striped bass (Morone sax- atilis). Comp. Biochem. Physiol. 93A:535-541. Moore. E. 1947 Studies on the marine resources of southern New Eng- land. VI. The sand flounder, Lophopsetta aquosa (Mitchell); a general study of the species with special emphasis on age determination by means of scales and otoliths. Bull. Bingham Oceanogr. Collect. Yale Univ. 11(3): 1-79. Mount, D.I., and W.A. Brungs 1967 A simplified dosing apparatus for fish toxicology studies. Water Res. 1:21-29. NOAA 1988 A summary of selected data on chemical contaminants in sediments collected during 1984, 1985, 1986, and 1987. NOAA Tech. Memo. NOS OMA 44. Natl. Oceanic Atmos. Adm., Natl. Ocean Serv., Rockville, MD, 15 p. + 4 appendices. Pentreath, R.J. 1976 The accumulation of inorganic mercury from seawater by the plaice, Pleuratiectes platessa L. J. Exp. Mar. Biol. Ecol. 24:103-119. Piatt, W.R. 1969 Color atlas and textbook of hematology. J.B. Lippincott, Philadelphia, 445 p.. Reid. R.N., A.B. Frame, and A.F. Draxler 1979 Environmental baselines in Long Island Sound, 1972-73. NOAA Tech. Rep. SSRF-738, Natl. Oceanic Atmos. Adm., Natl. Mar. Fish. Serv., 31 p. Snieszko, S.F. 1960 Microhematocrit as a tool in fishery research and manage- ment. U.S. Fish Wildl. Serv. Spec. Sci. Rep. Fish. 341, 15 p. Sprague, J.B. 1969 Measurement of pollutant toxicity of fish. I. Bioassay method for acute toxicity. Water Res. 3:793-821. Wedemcyer, G.A., and W.T. Yasutake 1977 Clinical methods for the assessment of the effects of en- vironmental stress on fish health. U.S. Fish Wildl. Serv. Tech. Pap. 89, 18 p. Abstract-— Trawl collections were made for Atlantic threadfin Polydac- tylus octonemus from 5 to 100 m in the Gulf of Mexico along a cross-shelf transect off Texas during October 1977-August 1981. Threadfin gener- ally mature at 165-210 mm TL as they approach 7-9 months of age. Spawning primarily occurs in one period. mid-December-mid-March, and spans 45-120 days overall; 90% of successful spawning may occur in only 59% of that period. Threadfin in the northwestern Gulf range from < 5 to 27 m depths in the demersal stage but are most abundant at <5 to 16 m. Young-of-the-year recruit in waters <5-16 m when 2-4 months old. Fish begin to disperse to deeper waters in early summer and form a positive size gradient from the estu- aries seaward. Threadfin in the de- mersal phase are not abundant in the northwestern Gulf after 9-11 months of age and reach only 15 months there. Observed mean and predicted sizes were 135-165 mm TL at 6 months, 165-215 mm at 9 months, and 180-205 mm at 12 months. Fitted von Bertalanffy parameters were 2.17-2.92 (A', annual). 195-230 {LJ. and -0.03-0.08 years {(„). Maximurn size in the demersal phase is 230 mm TL in the northwestern Gulf, but more typically only 200-205 mm. Typical maximum life span (/; ) is about 1 year but may exceed that if individuals survive in a pelagic stage after spawning. Apparent mean time and cohort-specific total annual mor- tality rates are 97-100% in the north- western Gulf. Population dynamics parameters presented are termed apparent because of the unknown effects of recruitment, movements, random variation, gear selectivity, etc. Spawning grounds seemingly lie along the Outer Continental Shelf, slope, or further offshore, and cur- rents of the cyclonic shelf gyre off Texas and western Louisiana trans- port the young to estuarine and in- shore nurseries. Reproduction, Movements, and Apparent Population Dynamics of the Atlantic Threadfin Polydactylus octonemus in the Gulf of Mexico* Michael W. Dentzau Department of Wildlife and Fisheries Sciences. Texas ASM University, College Station, Texas 77843 Present address Florida Department of Environmental Regulation 2269 Bay Street, Fort Myers, Florida 33901 Mark E. Chittenden Jr. Department of Wildlife and Fistieries Sciences, Texas A&M University, College Station, Texas 77843 Present address' College of William and Mary, Virginia Institute of Marine Science Gloucester Point, Virginia 23062 Manuscript accepted 26 March 1990. Fishery Bulletin, U.S. 88:439-462. The Atlantic threadfin Polydactylus octonemus occurs along continental shelves from Massachusetts to Brazil, and in the Gulf of Mexico (Breder 1948, Briggs 1958, Dahlberg 1975, Fischer 1978). Although uncommon on the Atlantic coast (Hildebrand and Schroeder 1928, Anderson and Geh- ringer 1965, Anderson 1968, WOk and Silverman 1976), it is abundant in Gulf of Mexico (Gull) coastal, sui-f, and estu- arine waters (Hildebrand 1954, McFar- land 1963, Chittenden and McEachran 1976). However, its annual abundance seems to flucttiate greatly (Reid 1956, Gallaway and Strawn 1974, Ogren and Brusher 1977). Despite its abundance, the life his- tory of P. octonemus has not been studied in detail, and little supporting data have been published. General notes occur in many faunal studies in- cltiding Gunter (1938a, 1945), Reid (1955), Miller (1965), and Juneau (1975). These notes indicate P. octo- nemus is a small, inshore species that occtirs from spring through fail in the northern Gulf and spawns from fall through spring. •Contribution no. 1586 of The College of Wil- liam and Mary, Virginia Institute of Marine Science. Our paper describes maturation, spawning periodicity, bathymetric distribution, recruitment, movements, age determination and growth using length frequencies, maximum size, life span, mortality, and relationships of length-weight, length-girth, stan- dard-total length and fork-total length of P. octonemus in the northwestern Gulf. It also discusses spawning areas and larval dispersal in relation to Gtilf hydrography, as well as how recruit- ment, movements, and randomiza- tion affect parameter estimation and interpretation. Materials and methods Collections for Polydactylus octone- mus were made from nearly 3000 trawl tows in 71 monthly or twice- monthly cruises, October 1977-Aug- ust 1981, along a cross-shelf transect in the Gulf off Freeport, Texas (Fig. 1) aboard a chartered shrimp trawler. Gear used were twin 10.4-m (34-ft) trawls with a 4.4-cm stretched-mesh cod end and a tickler chain. Initial stations usually were located at depths of 9, 13, 16, 18, 22, 27, 36, and 47 m. Sampling was expanded to in- clude stations at 5 and 24 m after November 1978 and at 55, 64, 73, 82, 439 440 Fishery Bulletin 88(3), 1990 Figure 1 Location of sampling area off Freeport, Texas. Sta- tion depths and bathymetric contours are indicated in meters. 86, and 100 m after May 1979. Collections were made during the day through September 1978; thereafter, a day and a night cruise were usually made each month. Usually, two tows of 10 minutes bottom time were made at each depth, except that 8 were made at 16 m, 24 were made at 22 m, and only 1 tow was made at most depths prior to October 1978. We emphasize our spatial sampling design was a single, cross-shelf transect from a sampling frame that encompassed much of the northern Gulf. All P. octonemus were culled from the catch, mea- sured to the nearest millimeter total length, fixed in 10% formalin, and then preserved in 70% ethanol. All individuals collected January-December 1979 were processed to determine the following: total length (TL), fork length (FL), standard length (SL), girth at origin of the dorsal fin (G), total weight (TW), gonad weight (GW), and sex and gonad maturity stage. If available, 200 fish were randomly selected for similar process- ing in all other months, except standard length and girth were not recorded. Gonad maturity (Table 1) was determined by a modification of the Kesteven classifica- tion (Lagler 1978). Fish in the Early Developing and later stages were considered mature to calculate maturity curves (Bowering 1983), which indicate the length when half were mature (Mr,,,). Gonadosomatic indices (GSI) were calculated for individual fish as: GSI = 100 GW/TW. Age in years was determined by length-frequency analysis, e.g., the modal group progression analysis modification (Jearld 1983) of the Petersen Method (Lagler 1956). Knowing from this that essentially all Dentzau and Chittenden Population dynamics of Polydactylus octonemus in the Gulf of Mexico 441 Table I Descriptions of gonad maturity stages assigned to Polydac- | tyliLS octonemus. Stage and name Description 1 Immature Sex undetermined; gonads small or not visible. 2 Maturing Virgin Sex distinguishable; gonads thin, occupy < 10% of body cavity; eggs not visible. 3 Early Developing Sex distinguishable to naked eye; gonads not flat, occupy 10-25% of body cavity; eggs opaque if present, not visible to naked eye. 4 Late Developing Gonads occupy 2.5-.50% of body cavity; opaque eggs present, visible to naked eye. 5 Gravid Gonads occupy >50% of body cavity; up to .50% of eggs translucent. 6 Ripe Gonads occupy >50% of body cavity; >50% of eggs translucent. 7 Spawning/Spent Ovaries flaccid, completely or partially empty. fish were age 0, we forced through the origin a regres- sion to approximate age in days at observed sizes; this procedure is detailed under the methods used to calculate hatching dates. Cohorts (year classes) were specified on length-frequency figures by the season and year when they were hatched, e.g., winter 1981. Arith- metic means were used to describe central tendencies of length for each cruise, because length frequencies within each cohort were generally normally distributed. Descriptions of spawning periodicity (beginnings and ends) using length frequencies assumed the following size and age combinations predicted from regressions of total length on age (Fig. 2): 25-40 mm TL at 1 Figure 2 Mean observed and predicted sizes at age (days) for Polydactylus octonemus cohorts with a summary of the iterative process used to calculate final hatching dates (H.D.) and set time-scales for final growth equations (F.E.). Initial growth equations (I.E.) were scaled to a 1 February hatching date. All regressions are significant at o = 0.005. The first two and three collections of the 1980 and 1981 cohorts, respectively, were deleted from calculations; see text on "Maturation and Spawning Periodicity" for explanation. E 150 E A Winler 78 SS^a Conlidence Interval \ About Observed Mean I Range l.E.:v= 19.44t0 89526«-0 00114 k' x-Int = 21 F.E. :v=0 14»0 94298X-0 00TI4«' r' =0.96 H.0.= 11 JAN 78 SO 120 180 240 300 360 420 460 AGE (DAYS) ~F M A M J J A S 6 N D J F M A M 1976 1979 COLLECTION DATE B Winter 79 I E v = -21 62»1 41077X -0 00239«' «-lnt = 16 F.E. :v=0 34+1 33423«-0 00239«' r'=0.97 H.D = 16 FEB 79 ) 60 120 180 240 300 360 420 480 AGE (DAYS) MAMJ J ASONDJ FMAMJ 1979 1980 COLLECTION DATE C Winter 80 LE.:y=4.8V0. 785241 X -lnt.= 6 F.E.:v=0 10+0.78524X t'=0 95 H O = 26 JAN 80 0 60 120 180 240 300 360 420 480 AGE (DAYS) F M A M J J A S O N 5 J F M A M ' 1980 1981 COLLECTION DATE D Winter 81 I 150 / .^ /' I.E.:v=-5.22t0.926S3x x-lnl,= 6 F E v=0 34*0 92653X r' =0 94 HO =7 FEB 81 60 120 180 240 300 360 420 480 AGE (DAYS) ~fi A M J J A S O i5 5 J F M A M T 1981 1982 COLLECTION DATE 442 Fishery Bulletin 88(3). 1990 month, 45-70 mm at 2 months, 70-100 mm at 3 months, and 95-125 mm at 4 months, depending upon cohort. These sizes were reasonable because growth averaged 28-32 mm/30 days at 3-6 months of age in May-July (Dentzau 1985). Apparent duration of the overall spawning period was approximated following Geoghegan and Chit- tenden (1982), as Time-specific mean size range early in life Mean growth/day early in life Calculations were based on fish collected early in life, i.e.. May- June; too few were collected before then (1-4 fish/cruise in March and April) to estimate growth. For the numerator, time-specific size range was estimated for each cohort as the mean of the 99% confidence in- tervals for observations in that period. For the denom- inator, individual growth increments were estimated as the difference between (1) observed mean total lengths on successive dates and (2) total lengths on suc- cessive dates predicted from regressions. Individual estimates of growth/day were then made as the in- dividual growth increment divided by the time between collection dates. Mean growth/day early in life was then estimated as the average of individual growth/day values between successive collections during May- June. Dentzau (1985, table 3) details these calculations. Calculations were also made using 90% confidence in- tervals to compare how much successful spawning oc- curred in a period shorter than the overall period. Hatching dates used to approximate time scales to calculate growth were determined by a one-step itera- tion process following Standard and Chittenden (1984). An initial hatching date of 1 February was used to start the approximation, because fish 50-90 mm, assumed to be 2-3 months old, first appeared in March-April. Quadratic regressions of total length on age in days after 1 February were then used as models to estimate an initial .r-intercept for each cohort; linear regression was used if the quadratic term was not significant at a = 0.05. Final hatching dates were calculated by using the ir-intercept for each cohort to readjust the initial .r-variable (time) scale, so each final growth curve passed through the origin (Fig. 2). Recruitment patterns and movements in the Gulf were determined by analyzing length frequencies and catch-per-unit-effort against depth at specified "sea- sons" (months) of the year; length-frequency data from Galveston Bay, Texas (Gallaway and Strawn 1974), ex- tended our analyses to estuaries. We use the words "recruit" and "recruitment" in two ways: (1) to de- scribe movement to areas by young P. octonemus descending to the bottom from their pelagic early stages, and (2) horizontal movements from estuaries 100 - r 100 /- 90 ■ l/ 80- 1 - 75 70- j 1 o c s >- o 60 - c n > z H 3 o lij Ml - 50 < m T3 cc 10 1 / \ m u. Ji / O / m .10 - / Z H .?0 - / \ -25 10 - ^ V ',11 100 I'lO JOO 250 1 0 TOTAL LENGTH (mm) Figure 3 Length frequencies and cumulative percentage of all Polydnctyhis octom-Tnus collected off Freeport, Texas, October 1977- August 1981. to the Gulf, or within the Gulf, by fishes already in the exploited phase. The former conforms to Beverton and Holt's (1957) meaning of recruitment (t^), because these areas are generally exploited, and to Beverton and Holt's (1957) and Ricker's (1975) meaning (t,), be- cause fish also then enter the exploited phase of life. Some gear selection for older, larger fish may occur as part of these processes. Typical maximum life span was approximated by the Beverton-Holt model parameter //^ (Gulland 1969), and typical maximum size was approximated as a cor- responding length (//J following the definition that only 0.5-1.0% of the catch exceeds age t^ (Alverson and Carney 1975, DeVries and Chittenden 1982). Values of l^ were calculated from the cumulative fre- quency for all fish captured (Fig. 3). Specific values of ti were calculated from li by solving for time in von Bertalanffy (Gulland 1969) and regression growth equations (Fig. 2). Time and cohort-specific total annual mortality rates (1 - S) were calculated using S = N/INo, where S = rate of survival, and A'', and N,t are the numbers of fish at age each month or per tow. Observed estimates were compared against theoretical values calculated from the expression Z = 4.6/number of years in life span (Royce 1972:238). Total mortality rates, typical max- imum life spans, maximum sizes, sizes at age, spawn- ing period durations, and von Bertalanffy parameters Dentzau and Chittenden: Population dynamics of Polydactylus octonemus in the Gulf of Mexico 443 > o z lU o UJ ir 30 -1 20 ■ 10- 15 n 10 - 5 10 - 5 - 5- 5 -| 5- Immalure n = 1260 X = 127 Maturing Virgin n= 378 Early Developing n - 118 A Late Developmg n - 12 X = 197 Gravid n = 3 X = 196 Spawning/Spent n 1 X - 189 60 120 180 24C TOTAL LENGTH (mm) 300 100- 90- 80^ ^ ^0- D < 60- 1- 50- z LU O 40- CL LU °- 30- 20- 10- 0- .' 1 a ;. IS ■"• 1 1 1 1 1 1 II 1 100 120 140 160 180 200 220 TOTAL LENGTH (mm) Figure 4 Length frequencies of immature ani:l female Polydactylus octotiemiis by gonad stage. Figure 5 Percentage of mature female Polydactylus octonemus as a function of size. dard, fork, and total lengths used regressions pre- sented herein. presented were termed apparent, because they may have been affected by recruitment of larger young from estuaries to the Gulf, by emigration and/or gear avoidance of fish about age I, or by a change from demersal to pelagic behavior; if so, they overestimated mortality, K, and spawning period durations and underestimated life spans, maximum sizes, average sizes at age, and L^. Von Bertalanffy parameters were calculated using Fabens' (1965) program and the same data used in TL on age regressions (Fig. 2). These points described a curvilinear regression and evidenced an asymptote, so they met the minimum requirements for a von Berta- lanffy fit (Knight 1968, Gallucci and Quinn 1979). Un- less stated, all length frequencies were moving aver- ages of three and all lengths were total length. The symbol "?L" was used for instances where the type of length was not reported. Conversions between stan- Results Maturation and spawning periodicity Polydactylus octonemus begin to mature at 165-210 mm in late summer-early fall. Sex usually could be determined by eye at 165 mm as many females entered the Early Developing stage (Fig. 4). Fish entered later stages at 180-210 mm. M50 was 185-195 mm, about 7-9 months old (Fig. 5), in agreement with gonad-stage length frequencies. Little somatic growth seemingly occurs after P. oc- tonemus enter later stages of gonad development. Mean sizes were 188 mm in the Early Developing stage (Fig. 4), 197 mm when Late Developing, and 196 mm when Gravid. The only fish collected in the Spawn- ing/Spent stage was 189 mm. Minimum and maximum sizes remained constant after the Early Developing stage. 444 Fishery Bulletin 88(3), 1990 >- I5n 10 5- lOn 5 ► W77.0 . 1 Oct 77 D n=221 -1 — ^ r W77 0 5 Nov 77 D n=7 1 3 Dec 77 D n=8 W770 I- ] . 1 W78 0 14 Apr 78 D n=4 W78 0 ■-^" — r 8 May 78 D 11=8 W78 0 14 Jun 78 D 11=1 1 ' — ^ -W78 0- 15 Jul 78 D n=60 -1 1 r" — ' — n~ ^W78 0- 15 Sep 78 D n=267 13 Oct 78 N n=97 -W78 0- 1 Dec 78 N n=72 ~i ^^ . W78 0^ -1 f^-'H — "^ 19 Dec 78 D n=10 -W78 0- "^1 "-""1 W79 0 5 Apr 79 N n=4 W78 1 20 Apt 79 D 11=1 —\ — 60 120 — I 1 1 240 300 10 5 • W79.0 15 May 79 N n=96 ^ W79,0 10 Jun 79 N n=53 5l -W790^ 24 Jun 79 D n=57 ^-W79 0^ ■ 9 Jul 79 N n=59 1 ' 1 ' 1 1 ' 1 W79 0 22 Jul 79 D n=5 "1 r 15 10- 5 22 Aug 79 D n=337 -W790^ 15 10- 5 24 Sep 79 D n=234 W79 0- -1 ^^ r 4 Oct 79 N 11=15 lOl 5 19 Oct 79 D n=160 - W79 0 - -W79 0' A— 5^ 6 Nov 79 N n=10 W79J) "I' '""I r W80 0 21 Mai 80 U 11=1 — I — 60 — I — 120 — I — 180 1 ' 1 240 300 TOTAL LENGTH (mm| Figure 6 Monthly length frequencies oi Polydacti/lus dctitnemus off FreepMrt. Texas, in day (D) and night (N) cruises. Bars in each panel depict cohort size ranges. The letter and first two digits above or within a bar indicate cohorts; the digit to the right of the decimal, age in years. Dentzau and Chittenden: Population dynamics of Polydactylus octonemus in the Gulf of Mexico 445 W80 0 21 Mar 80 D n=1 n 1 r ^W80 0^ 5 May 80 N n=58 T 1 r 5n ^W80-0 — 19 May 80 D n=45 W80,( — W80 0- 19 Jun80D n=22 — ^ -W80 0 I*- 11 Jul SON n=209 -W800- .,.,A.^^.„ 24 Jul 80 D n=121 10-| 5- T ' r — W80 0- Y^-' 5-1 11 Sep SON n=35 -W80 0^ 5i 25 Sep 80 0 n=36 — I 1 1 r 6 Oct 80 N n=4 ^W80 0- W80 0 6n 21 Oct 80 D n=1 W80 0 6 Nov 80 Nl n=1 W810 W81 0 n ' 1 20 Api81 D n=1 60 — I — 180 240 300 ^W810^ 4 May 81 N n=31 T 1 r 2Jun80N lOn n=6 W81 0 ■ W81,0 . 19 May 81 D n=54 -I—' 7 Jiin81 N n=133 1-^^^ I . W81 0 • 16 Jun81 D n=127 ^WBIO- T ' 1 1 Jul 81 N n=165 AQ, - 13 Aug 80 N lO- n=182 W81 0- 20 Jul 81 0 n=342 307 20 10 ^W81 0- 3 Aug 81 D n=566 20 1 15- 10 5- 20 Aug 81 D n=395 W81,0 . Apr 81 N 10 n=2 1 ' 1 18 May 82 D n=219 120 180 240 300 TOTAL LENGTH (mm) Figure 6 (continued) Polydactylus octonemus primarily spawn in one dis- crete period from mid-December through mid-March. The well-defined modes in length frequencies from throughout the study (Fig. 6) suggest spawning oc- curred in one discrete period. Major spawning must have occurred during mid-December-mid-March, because small, young fish recruited mainly during mid- March-mid-May each year, e.g., fish 75-105 mm and 446 Fishery Bulletin 88(3), 1990 APR 78 n=4 o-¥i 1 1 M . MAY 78 11=6 0CI78 n=45 DEC 78 11=38 APR 79 n=5 MAY 79 n=94 IUN79 n=100 IUL79 SEP 79 n=71 MAR 80 n=2 MAY 80 11=100 OCT 8 11=22 APR 81 n=2 MAY 81 n=79 11=175 SEP80 n=30 10181 n=l59 MATURITY STAGE A0G81 n=t37 Figure 7 Monthly gonad maturity stages (defined in Table 1) of immature and female Pi)lydaclylux uctoui'- mus. 3-4 months of age recruited in abundance in mid- April-early May 1978, fish 50-70 mm and 2 months old in early April 1979, fish 65-115 mm and 2-4 months old in mid-May 1979, fish 75 mm and 3 months old in mid-March 1980, fish 55-115 mm and 2-4 months old in May 1980, and fish 60-115 mm and 2-4 months old in early April-early May 1981. Few fish 50-115 mm were collected from late August through February, in- dicating little spawning during late April-late October or November. Calculated hatching dates were 11 January 1978, 16 February 1979, 26 January 1980, and 7 February 1981 for the 1978-1981 cohorts, respectively (Fig. 2). Other dates could have been calculated depending on data points included; however, the basic spawning period and coefficients of determination would be similar (Dentzau 1985). We deleted the earliest two collections in 1980 and earliest three in 1981 from calculations because mean sizes in these collections were as large or larger than those in subsequent ones and may have reflected incomplete recruitment or gear selection for larger fish, causing an upward size bias. Gonad maturity data support the major mid-Decem- ber-mid-March spawning period indicated by length frequencies. Except for one gravid age-I fish collected in April 1979, all fish collected in March-June were im- mature or maturing virgin (Fig. 7); none were in the Early Developing or later stages. Early and/or Late Developing stage fish were most common in October- December. Except for the one individual caught in April 1979, gravid fish were collected only in December as they approached age I. No ripe fish were collected, and few were gravid or late developing. Mean and/or maximum GSI values, although usually low, were high- Dentzau and Chittenden Population dynamics of Polydactylus octonemus in the Gulf of Mexico 447 ,.. 95\Conlit]ence Limits AbouHheMean Range SEP OCT DEC APR JUN JUL AUG SEP OCT NOV JUN JUL AUG SEP OCT MAY JUN JUL AUG Figure 8 Monthly mean gonadosomatic indices, ranges, and 95% confidence limits about means for female Polydactylus octonemus, September 1978-AugTast 1981 . est November-December (Fig. 8), except in August 1980 and 1981, when the few late-developing fish great- ly skewed the means, and in mid-April 1979, when one age-I fish was gravid. The latter fish suggests some spawning in April and possibly May. The apparent overall spawning period spans 45-120 days. Calculated overall spawning periods were 144 and 80 days in 1978, 53 and 80 days in 1979, 46 and 103 days in 1980, and 64 and 121 days in 1981 based on respective observed and predicted growth/day values. The broad 144-day interval for 1978 was based on only 9 fish and may be unreliable. Ignoring that value, spawning encompassed a 45-120 day interval which brackets the 90 day duration length frequencies indicated. A large fraction of the successful spawning may oc- cur in a period much shorter than the calculated overall spawning period duration. Using mean 90% instead of 99% confidence limits for observations, spawning periods were 78 and 43 days in 1978, 33 and 50 days in 1979, 27 and 60 days in 1980, and 40 and 76 days in 1981, based on respective observed and predicted growth/day values. On average, the duration estimated using 90% confidence limits was 59% of that using 99% limits. This suggests most spawning occurs in a rela- tively small part of the overall spawning period. Cohorts are apparently produced by fish that first spawn when 10-14 months old, assuming the hatching dates and sizes at early age noted previously. Spawn- ing at 10-14 months of age is supported by (1) the disappearance from collections of all but one fish sometime during late October-late December when 9-11 months old (Fig. 6), (2) the occurrence of large mean and maximum GSI values in October-December (Fig. 8), and (3) the collection of early developing and more sexually mature fish during September-Decem- ber (Fig. 7). ' Polydactylus octonemus exhibit a sex ratio of 1.00 male to 1.78 females. This ratio was observed in 800 fish and differed significantly from 1:1 (x" = 64.98, a = 0.05, df= 1). Bathymetric distribution, recruitment, and movements Polydactylus octonemus in the northwestern Gulf oc- cur from <5 m depths to at least 27 m off Freeport. Only one specimen was collected deeper than 27 m, a 75-mm fish at 86 m in April, which may represent a pelagic young captured as the net was being retrieved. Greatest abundance generally occurred at 5 m, the shallowest depth occupied (Fig. 9), except in 1978 when this depth was occupied only in December, a time when 448 Fishery Bulletin 88(3). 1990 o o o §§ i • Pooled a 1977 a 1978 O 1979 V 1980 O 1981 « AbserH Figure 9 Catch/effort (mean numher/ 10-min tow) by depth for Piilydartylut; octonemux off Freeport, Texas, each year and pooled, October 1977- August 1981. Tows in Janu- ary and February were excluded in calculating ef- fort because no fish were captured then. fish had already begun to move offshore as noted be- low. Abundance was generally high at 9-16 m, although it often declined with increasing depth in this range. Abundance declined sharply between 18 and 22 m. Young-of-the-year P. octonemus in the northwestern Gulf recruit to estuaries and coastal waters <5-16 m deep. Many young recruited to Galveston Bay in April and May when 50-110 mm (Fig. 10). In the Gulf, re- cruitment was greatest at the shallowest depths oc- cupied. Fish 50-120 mm and 2-4 months old mainly recruited to 5-9 m in March-May, although some recruited as deep as 16 m (Fig. 11 A, Table 2). Fish were most abundant at 5 m. Their abundance declined with increasing depth and was very low deeper than 13-16 m. Only one recruit was taken deeper than 16 m, the 75-mm fish at 86 m in April that may have been in a pelagic stage. Polydactylus octonemus gradually disperse to deeper water in early summer and thereafter. Few age-0 fish were captured deeper than 9-13 m in April and May, but they were found at 18-27 m in June and were abun- dant at 16 m in July (Table 2). This gradual offshore dispersal is supported by the size gradient that existed seaward from Galveston Bay in June-December (Table 3), i.e., the smallest mean sizes were usually in Gal- veston Bay and the shallowest Gulf depths. Larger P. octonemus lead the offshore dispersal. Minimum sizes were similar at each depth in the Gulf and in Galveston Bay during May (Fig. 10, 11 A). In con- trast, size compositions showed gradients of increas- ing size with depth in June, July, August, September, and October, suggesting larger, presumably older, fish move offshore first (Fig. 11A,B,C; Table 3). Peak abundance of P. octonemus shifts towards deeper waters as they disperse offshore. Abundance was greatest by far at 5-9 m in May and June, then included 13 and 16 m in July as fish dispersed offshore (Table 2). Abundance peaked at 16 m in August and September, although fish were abundant from 5 or 9 to 18 m, and at 16-24 m in October (Table 2). Polydactylus octonemus abandon estuaries and in- shore portions of the Gulf (<5-9 m) in late fall and occupy only the offshore portions (13-24 m) before disappearing in November and December. Two fish at 9 m in November were the only individuals captured shallower than 13 m in the Gulf in November and December (Table 2), although peak abundance was at 5-9 m in May and June. Gallaway and Strawn's (1974) data show a similar abandonment of estuaries by late fall (Fig. 10). Polydactylus octonemus were most abundant in July-September and least abundant November-April Dentzau and Chittenden: Population dynamics of Polydactylus octonemus in the Gulf of Mexico 449 im Api 68 10 Oct 68 r 11 = 481 11 = 28 160 20- r to_ 8(1 J I . Nov 68 1 11=2 JBO 320 May 68 11= 1187 ^ Dec 68 ^ Jan 69 1 "-'} n 160 |: ■5_ 5 Api 69 1 ^ " = 5 May 69 660- n Jun68 " 11= 137 >- o z LU 3 o UJ cc LL 440- 220 i. n= 1455 25- 100- I ■ kn. Jun69 n = 386 80 40 Jul 68 n = 250 \ 50- fH; h ■ 40- Jul 69 n= 107 50 1 Aug 68 n= 136 ■ 20- n 25 50 25 '-.''■■ ;;.;:3_ 4 i..-i\ . 20- Aug 69 Sep 68 n= 167 f- 5- Sep 69 n=3 — 1 — 1 — , — . — p-T — p^ P [; :^ 5, Oct 69 n = 2 1 40 80 130 160 0 40 30 1?0 160 TOTAL LENGTH |mni| TOTAL LENGTH |mm| ) 20 40 60 80 100 ) 20 40 60 80 100 SI ^ANDARD LENGTH |mm| STANDARD LENGTH (mm) Figure 10 Monthly length frequencies oiPolydnctylii,s octonemus in Galveston Bay. Texas. 1968 and 1969. Adapted from seine and trawl data of Gallaway and Strawn (1974); length frequencies are not moving averages of three. (Fig. 12). Usually no P. octonemus were captured December-February. Age determination and growth Only one cohort of P. octonemus usually occurs at any one time in the northwestern Gulf. The only exceptions to this we found were in April 1979 when two cohorts appeared, albeit on separate cruises, and possibly in November of 1977 and 1980 (Fig. 6), when the second cohort was represented by only one individual. Polydactylus octonemus is not abundant in the demer- sal phase in the northwestern Gulf after 8-11 months of age and apparently reaches only 15 months. Cohorts were present only until their first November or Decem- ber after which they were not captured again (Fig. 6), except for the 194-mm fish in April 1979 that was the only fish older than 12 months. Apparent growth of P. octonemus varied between cohorts, but observed mean sizes, and von Bertalanffy and quadratic regression predictions, averaged 135-165 mm at 6 months, 165-215 mm at 9 months, and 180-205 mm extrapolated to 12 months (Table 4). Regardless of the growth model, for a given cohort, predicted sizes agreed within 6 mm or less at 6 and 9 months and within 11 mm at 12 months. Fitted von Bertalanffy equations were: (1978):^, = l95.2[l-e-«oo™9"-3'''»0)]; annual K = 2.92; annual ^, = 0.0833 (1979):/, = 230.4[l-e-"005955(f-ii.69)]. annual K = 2.17; annual /,, = -0.0320 where /< = TL in millimeters at time t in days. Maximum size, life span, and mortality Polydactylus octonemus in the demersal phase ap- parently reach a maximum size of about 230 mm in the northwestern Gulf but more typically only 200-205 mm. The largest of 4324 specimens collected was 229 mm; 99% were <202 mm and 99.5% were <206 mm (Fig. 3). The latter two sizes estimate an apparent li. The apparent typical maximum life span in the demersal phase of P. octonemus in the northwestern Gulf is only 1 year. A value of /^ = 1 seems reasonable because (1) /^ values of 202 and 206 mm can be substituted into hatching-date regressions and von Ber- talanffy equations to predict values of 0.69-1.00 year; (2) observed and predicted mean sizes at 9 months are 165-215 mm and predicted sizes at 12 months are 180-205 mm; (3) the largest specimen was about 11 months old when collected in December 1978; and 450 Fishery Bulletin 88(3). 1990 A. MARCH-JUNE Mar-Apr 5M n=7 C/t=035 'V'"" ' " I — I 1 1 9M n=4 C/t=01R -r 1 1 r 5n 16 M n=1 C/f<0 1 1 ' r- 86M n=1 C/f<0 1 -1 ^"i T" 5- May V 9M n=64 C/f=5 8 —I 1 r- , — 0- 13M n=15 C/t=l 4 -1 1 r 5i ^0- 16M n=27 C/t=0 5 o 10 UJ =3 5- Jun 5M n=167 C/t=14 8 1 — ^ 1 ' "i • r . 0 • 9M n=83 C/t=8 6 ,.. - jaA/-^ 13M n=29 C/f=2 5 1 5i ^..-yV^'V^s. 16M n=65 C/f=14 18M n=9 C/(=0 8 5l "r "■''- I ^^^^,^„- 22-24M n=45 C/f=04 5n 27M n=1 C/f=0 1 — r- 60 — I ' — — I ' 1 1 1 120 180 240 300 5 Jul B. JULY-AUGUST -0- 5i 15-1 5 M n = 186 C/f=18 7 lo- -1 5- -1 5 5] Aug 40 1 30- 20- 10- 1 10 5 Hi''- " " 1 r- 5M n=178 C/f=17 2 0- 9M n=210 C/f=107 ' 0— 13M n=94 C/M8 0 16M n=460 C/f=9 1 1 1 1 — ^ n . -0 ^ 18M n=6 C/t=0 5 0' 22-24M n=13 C/f=0 1 T ' T -. .Y^h. A ,n , j_ 5M n=15 C/f=1 8 'f-'"''T'ir' — f^ gM n=73 C/f=9 9 V ■^. ^(^ " 13M n=112 C/f=11 1 -1 ' r 18M n=126 C/t=12 6 ' 16M n=970 C/f=278 ^^* , . 1 -1 1 1 1 ' I 1 ^^ r 10-1 5 — 0 22 -241^1 n=174 C/t=1 6 -1 — — 1 r- 5l TOTAL LENGTH (mm| • 0 • 27M n=10 C/f-12 -T T ' 1 i?n 180 :'4n .wo Figure 1 1 Length frequencies by depth for Polydactylus octonernus off Freeport, Texas, in March-June (A). July-August (B), September-October (C), and November-Decem- ber (D). Data in each panel were pooled over the period October 1977-AugTJSt 1981. Dentzau and Chittenden Population dynamics of Polydactylus octonemus in the Gulf of Mexico 451 C. SEPTEMBER-OCTOBER Sep 5M n=17 C/f=3 8 T ' 1 1 ''i ' — I 1 < 1 5i 9M n=22 C/t=3 9 n^ A-A^" -1 ' 1 13M n=16 C/f=4 0 -1 ' 1 '1"^ 'f ' r 15i 10 5- 16M n=322 C/f=12 2 T ^ 1 1 1 18M n=25 C/f=4 1 -1 V '' IT' r- >- 5' 22-24M n=170 C/t=2 7 1 1 ' 1 J 1 K 1 ^^v — 1 — • — 1 Oct 5M n=2 C/f=l 1 """-T 1 1 1 1 9M 1^ 11=7 C/t=0 9 5i ■n ' 1 13M n=6 C/t=13 -T T '"'N" 15 10 5- 16M n=259 C/f=5 0 -n "-I r- 18M 0"^ — ' n=38C/f=5 0 ,-,.--■■ AM, „^ T 1 1 15 ID- S' 22-24M n=186 C/f=2 3 -1 r 1 1 ^ 60 120 180 240 300 D. NOVEMBER-DECEMBER 51 No Summer > 0 — • 9M n=2 C/t=0 2 ISM • 0' n=9 C/f=04 -\ ' r 5i ^0 22-24M n=7 C/t=0 1 Dec 13M n=1 C/f=0 1 T ' r 16M n=23 C/t=0 5 '-A" "^f -1 r'-'-'V v^^-^ 18M n=1 C/f=0 1 22-24M n=65 C/f=0 5 0- l.'il 130 :-10 300 TOTAL LENGTH Imml Figure 1 1 (continued) (4) P. octonemus almost completely disappeared from trawl catches off Freeport at 9-11 months of age. The collection of one specimen seemingly 15 months old in April 1979 (Fig. 6) suggests a few age-I fish survive and remain in the white shrimp community, an inshore fauna described by Hildebrand (1954) and Chittenden and McEachran (1976). However, this would not great- ly affect our estimate of ti. 452 Fishery Bulletin 88(3). 1990 Table 2 Monthly catch/effort (mean number/10- nin tow) of Polydac- | tylus octonemus by depth off Freeport. Texas. Period Depth (m) 5 9 13 16 18 22-24 27 Mar.-Apr. 0.4 0.2 0.1 0.1 0 0 0 Mav 18.7 5.8 1.4 0.5 0 0 0 June 14.8 8.6 2.5 1.4 0.8 0.4 0.1 July 17.2 10.7 18.0 9.1 0.5 0.1 0 Aug. 1.8 9.9 11.1 27.8 12.6 1.6 1.2 Sep. 3.8 3.9 4.0 12.2 4.1 2.7 0 Oct. 1.1 0.9 1.3 5.0 5.0 2.3 0 Nov. 0 0.2 0 0.4 0 0.1 0 Dec. 0 0 0.1 0.5 0.1 0.5 0 Polydactylus octonemus in the demersal phase have an apparent total annual mortality rate off Texas that approaches 100%, mean time- and cohort-specific values being 97-100%. Only one winter cohort was pre- sent in 31 of 32 months off Freeport (Fig. 6), so time- specific estimates were 100% in each of these instances. Cohort-specific values were 100% in 6 of 7 months for 1978, 8 of 8 months for 1979, and 5 of 5 months for 1980 fish because A^, was zero. A time-specific esti- mate for April 1979 and a cohort-specific estimate for the 1978 cohort in April of 1978 and 1979 were 75% and 95% (Z=1.39 and 3.00), respectively, but the former value may be an underestimate because the younger cohort was incompletely recruited in April. Table 3 Mean total length (mm) f jr Pdlydn iyhis octonemus in Galveston Bay, 1968 and 1969 (adapted from Gallaway and Strawn 1974). and by depth in the Gulf off Freeport, Te.xas. Data at each depth in the Gulf are for age 0 fish pooled over the period October 1977- August 198] Month Galveston Bay 1968 1969 Gulf of Mexico (depth in m) 5 9 13 16 18 22-24 27 86 Mar. 74.0 Apr. 77.2 65.7 70.7 80.5 75.0 May 78.3 81.6 84.9 86.7 76.9 80.7 June 86.6 92.7 109.4 127.1 132.1 123.2 139.2 147.7 164.0 July 93.1 117.5 133.6 148.6 142.8 156.4 163.5 157.0 Aug. 102.3 129.5 140.6 157.1 159.9 164.5 173.4 165.5 164.9 Sep. 109.7 132.0 168.2 169.4 168.5 164.7 168.8 175.7 Oct. 114.1 140.3 163.0 160.1 167.2 165.0 173.6 185.5 Nov. 104.8 194.7 189.7 Dec. 123.9 164.0 185.9 186.0 181.8 Jan. 1.58.0 o 12 10 8 6 4 2 H 0 • 1979 D 1980 ^ 1981 M n 1 r M J J MONTH A O Figure 12 Catch/effort (mean number/10-min tow) by month o{ Polydncfylux urdmemus off Freeport, Texas, each year. Dentzau and Chittenden Population dynamics of Polydactylus octonemus in the Gulf of Mexico 453 Table 4 Observed mean sizes (mm TL), size ranges, and predicted sizes by hatching-date regressions (Fig. 2) and von Bertalanffy models for 1977-80 Polydactylus octonemus cohorts at 6, 9, and 12 months of age. Observed values are pooled data from July-August at 6 months and October-November at 9 months. Dentzau (1985, table 4) gives more detail. Year Observed sizes 6 mo 9 mo Mean Size range Mean Size range 1977 1978 1979 1980 1981 1978 1979 1980 153 154 148 160 125-176 107-194 86-197 103-223 178 178 183 207 144-200 160-214 157-212 194-214 Predicted sizes 6 mo 9 mo 12 mo Regression von Bertalanffy Regression von Bertalanffy Regression von Bertalanffy 133 163 141 136 157 172 186 212 166 187 192 181 205 Table 5 Total weight-total length, girth-total length. and length- ength regressions for P. Aydactylus octonemus with supporting statistics. All regressions were significant at 0, = 0.05. Th« > symbohi is from Kicker's 1 (1973) GM regression. Measures are grams and millimeters. See Methods for symbols. Equation n Range lOOr^ V Log,,, TW = -5.72 + 3.27 Log,, TL 2398 51-226 99.0 3.29 (males + females + immatures) Log,,, TW = -6.29 + 3.53 Log,, TL 800 107-226 94.2 3.64 (males + females) Log,,, TW = -6.50 + 3.63 Log,, TL 286 126-207 92.8 3.76 (males) Log,,, TW = -6.25 + 3.51 Log,, TL 514 107-226 94.8 3.60 (females) TL = 15.10 + 1.70 G 847 51-226 94.8 1.74 G = -4.22 + 0.56 TL 847 51-226 94.8 0.57 TL = -1.04 -K 1.42 SL 847 51-226 99.1 1.43 SL = 1.74 + 0.70 TL 847 51-226 99.1 0.70 TL = -2.77 + 1.26 FL 2398 51-226 99.5 1.27 FL = 2.80 + 0.79 TL 2398 51-226 99.5 0.79 Weight, girth, and length relationships Total weight-total length regressions for males and females (Table 5) were not significantly different in slope (F = 2.91, df = 1, 798), but they differed in eleva- tion (F= 7.73, df = 1, 799). Therefore, equations are presented for each sex and for males and females pooled. Total weight-total length regressions for all fish and for males and females pooled were significantly different in slope (F = 99.10, df = 1, 2056); however, the one equation that pools all fish may be useful at this stage of management. Discussion Spawning periodicity The primary winter-spring spawning period we found for Polydactylus octonemus in the northwestern Gulf, 454 Fishery Bulletin 88(3). 1990 Table 6 Available information on larval, young (see text footnotes 1 and 4), and juvenile specimer s of Pvlydnrtylus octonemus collected in western Gulf of Mexico surface waters by plankton net dip net, meter net, and neuston net. Reference numbers correspond to locations in Figure 13. Location collected Reference Collection Depth Length number (lat.°N) (long.°W) date (m) (mm) Reference Texas and Loui siana 1 Gulf off Barataria Bay, Louisiana Apr. 33 <22 50 TL Gunter 1938b 2 26°40' 92°00' 8 May 54 1602 — Springer and Bullis 1956 3 26°05' 95°25' 27 May 54 1890 — Springer and Bullis 1956 4 26°15' 95°00' 7 Mar. 65 S2300 44-54 TL Pequegnat (text footnote 3) 5 27°52' 93°48' Mar. 72 137 57-71 TL Bright and Cashman 1974 6 27°33' 96°06' Jan. -Feb. 76 134 35-36 TL Pequegnat et al. 1977 7 26°57' 96°32' Jan.-Feb. 76 106 38-47 TL Pequegnat et al. 1977 8 26°10' 96°24' 27 Feb. 76 91 21-52 TL Pequegnat et al. 1977 9 27n7' 96°23' 25 Mar. 76 131 40 TL Pequegnat et al. 1977 10 27°30' 96°44' 8 Apr. 76 49 56 TL Pequegnat et al. 1977 11 27°17' 96°23' 10 Apr. 76 131 43-80 TL Pequegnat et al. 1977 12 26°58- 97''ir 29 May-8 June 76 25 — Pequegnat et al. 1977 13 27''17' 96°23' 20 Apr. 77 131 65-72 TL Wormuth et al. 1979 14 27°34' 96°07' 15-21 May 77 134 — Wormuth et al. 1979 15 26° 10' 96°24' 15-21 May 77 91 — Wormuth et al. 1979 16 Gulf off Freeport 6 May 79 54 56-66 TL Rockett' 17 28°14' 94°57' 23 Apr. 59 47 — Bullis and Thompson 1965 18 26°00' 96°00' 22 Apr. 82 200 44 SL Lieby' 19 26°30' 94°30' 24 Apr. 82 200 .39-51 SL Lieby 20 27°00' 92°00' 25 Apr. 82 200 40-48 SL Lieby 21 27°30' 92°30' 10 May 82 200 41-44 SL Lieby 22 27°30' 93°30' 10 May 82 200 45 SL Lieby 23 28°00' 94°00' 9 May 82 Mexico 58 38-45 SL Lieby 24 2o°or 92=27' 14 May 54 1467 — Springer and Bullis 1956 25 19°13' 95°34' 17 May 54 1260 — Springer and Bullis 1956 26 20°34' 95°37' 20 May 54 2160 — Springer and Bullis 1956 27 21°42' 93°35' 22 May 54 2736 — Springer and Bullis 1956 28 24°00' 96°50' 25 May 54 1035 — Springer and Bullis 1956 29 24°54' 96°05' 27 May 54 1530 — Springer and Bullis 19.56 30 25°08' 94°58' 9 Mar. 65 — 48 TL Pequegnat (text footnote 3) 31 23°14' 96°08' 23 Apr. 75 — 49-60 TL Kennicutt (text footnote 3) 32 23°00' 96°52' 24 Apr. 75 — 53-66 TL Kennicutt (text footnote 3) 33 20°40' 92°55' 25 Nov. 56 2088 - Bullis and Thompson 1965 34 20°40' 96°05' 2 Apr. 56 1782 — Bullis and Thompson 1965 35 19°35' 95°28' State Univ., Boone, 11 Apr. 56 2.300 - NC 28608. pers. commun.. March 1981. Bullis and Thompson 1965 'M. Rockett. Appalachian -M. Lieby, B ur. Mar. Res . St. Petersburg. FL 33701. pers. commun.. Feb. 1986. mid-December-mid-March and continuing at a lower level through April or May, generally agrees with the limited literature. Gunter (1938b, 1945) suggested fall- winter or late winter-early spring spawning, because 28-100 mm young appeared from April through June. Winter-spring spawning is also supported by: (1) col- lection of neustonic "larvae"' 21-80 mm^ in water 25-134 m deep over the south Texas continental shelf primarily from mid-January through early April, but also into early June (Pequegnat et al. 1977); (2) collec- tion at the surface of fish (a) 44-54 mm in water 2300 m deep in early March off south Texas (26°15'N, 95°00'W)3, (b) 57-71 mm (41-51 mm SL) in water 137 'Larvae as used by Pequegnat et al. (1977) and Wormuth et al. (1979) may encompass the juvenile stage defined by Moyle and Cech (1982). -We obtained sizes from specimens deposited at the Texas Coop- erative Wildlife Collection (TCWC), Texas A&M University, Col- lege Station, TX. 'Fish were collected by W. Pequegnat and M. Kennicutt (Dep. Oceanogr., Texas A&M Univ.). and deposited in TCWC. Dentzau and Chittenden: Population dynamics of Polydactylus octonemus in the Gulf of Mexico 455 m deep in March over the West Flower Garden Bank off Texas (27°52'N, 93°48'W; Bright and Cashman 1974), (c) 50 mm in water < 22 m in April off Barataria Bay, LA (Gunter 1938b), and (d) 56-66 mm in water 54 m deep in early May off Freeport (M. Rockett, Ap- palachian State Univ., Boone, NC 28608, pers. com- mun., March 1981); (3) collection at the surface in early May of "young"^ off south Texas (26°05'N, 95°25'W) and off the Louisiana continental shelf (26°40'N, 92°00'W) in water 1890 and 1602 m deep, respective- ly (Table 6; Springer and Bullis 1956); and (4) collec- tion of neustonic "larvae" (footnote 1) 65-72 mm in water 91-134 m deep in mid-April-mid-May off south Texas (Wormuth et al. 1979). Our calculated spawning-period duration of 45-120 days assumes large fish hatch before small ones and all grow at the same rate (Geoghegan and Chittenden 1982). The latter assumption appears reasonable, be- cause 99% confidence intervals for observations (in Dentzau 1985, table 3) were fairly constant between cruises within each May-June period. Our calculated duration may be an overestimate because continued recruitment of small fish to the Gulf would tend to depress estimates of mean growth/day, the denomin- ator in our spawning-duration equation. The numer- ator, mean size range, is less affected; larger recruits gradually disperse offshore to deeper waters, but we collected throughout and beyond the bathymetric range of the species. Comparisons of spawning-period durations from 90 and 99% confidence intervals are valid, however; such calculations use the same values except t. Our calculated hatching dates indicate mid-January- mid-February spawning, these being mean spawning dates because regression predicts averages. They are in error to the extent that regression curves do not pass through the origin and that they are asymptotic to the X- or ^/-axis in very early life. The latter is an unknown; the former problem does not seem serious, because Shirota (1970, in Hunter 1981) found length at first feeding in many species is four times the egg diameter. Egg diameter is not known in P. octonemus (Martin and Drewry 1978, deSylva 1984), but it averages 0.76 mm in the closely related Eleutheronema tetradadylum (deSylva 1984). This suggests P. octonemus is 3-3.5 mm at hatching, so only a few days error occurs in x, at most. Spawning areas Any spawTiing of P. octonemus in the northern or north- western Gulf presumably occurs in the water column. ■•Springer and Bullis (1956) defined young as postlarval. juvenile, and young specimens. Our data and the literature (Chittenden and McEach- ran 1976; Wohlschlag et al. 1977, 1979) indicate they basically disappear as demersal fish from the white and brown shrimp communities of the northern Gulf in November-December. They may assume a pelagic behavior then, which would explain their apparent absence in January-February, because Hastings et al. (1976) observed this species in December in habitat typical of pelagic fish— open water areas around an off- shore platform in water 18 m deep in the northeastern Gulf off Panama City, Florida. Catch records for young in the literature and the disappearance of P. octonemus from the white and brown shrimp communities in their primary winter spawning period, if the latter does not simply reflect gear selectivity following assumption of a pelagic habit, suggest this species spawns along the outer continen- tal shelf, the continental slope, or further offshore. Although pelagic "larvae" (footnote 1) and juveniles have been collected at the surface in or near waters of the white shrimp community off Texas and Louisiana in April-early June (collections 1 and 12, Fig. 13) or the brown shrimp community in late February-May (collections 8, 10, 15, 16, and 23, Fig. 13), more fre- quent collections in U.S. waters have been made along the outer continental shelf at 106-137 m in January- May (collections 5, 6, 7, 9, 11, 13, and 14, Fig. 13), and even further offshore at > 200-2300 m in March-May (collections 2, 3, 4, 18, 19, 20, 21, and 22, Fig. 13). Several collections of "young" (footnote 4) and juve- niles have been made in Mexican waters 1035-2736 m deep between Laguna Madre Tamaulipas and Cam- peche Bank in March-May (collections 24-32, Fig. 13). Similarly, young of other Polydactylus spp. also occur off continental shelves although adults are common in inshore waters; e.g., Klawe and Alverson (1964) found young P. opercularis and P. approximans <48 mm 250 nm offshore in the eastern tropical Pacific Ocean. In contrast to the abundance of P. octonemus larvae and juveniles in the western and northwestern Gulf, they have not been identified in extensive collections off the west coast of Florida from Pensacola on the north, south to the Florida Keys (Vick 1964; Houde et al. 1979; M. Lieby, Bur. Mar. Res., St. Petersburg, FL 33701, pers. commun., Feb. 1986). This suggests little or no spawning in that area and in the eastern side of the Loop Current, which contributes water to the western Florida shelf (Austin and Jones 1974). Polydactylus octonemus eggs, larvae, and juveniles, pelagic like other polynemids (Breder and Rosen 1966, Kagawade 1970), presumably use current transport to reach their estuarine and shallow Gulf nurseries. Stan- dard and Chittenden (1984), based on Murphy and Chit- tenden (unpubl.), suggested spawning in Larimusfas- ciatus and other Gulf fishes is timed to coincide with 456 Fishery Bulletin 88(3). 1990 30°- 25°- 20°- Figure 13 Summary of available collections of lar- val, young (see text foonotes 1 and 4. and juvenile Polydactylus octonemus from surface waters in the Gulf of Mex- ico. Reference numbers correspond to those in Table 7 where collection infor- mation is given. the occurrence of downcoast alongshore coastal cur- rents (towards Mexico) and onshore surface Ekman transport caused by downcoast wind stress along much of the Texas and western Louisiana coast from Aug- ust-September through April-May (Kelly et al. 1981). When alongshore currents are downcoast, an eastward or northeastward counterflow along the shelf break forms a cyclonic gyre on the continental shelf off Texas and western Louisiana (Kelly et al. 1984). The along- shore components of this gyre probably transport, in merry-go-round fashion, eggs, larvae, and juvenile P. octonemus from the outer shelf and slope to their shal- low nurseries. Portions of this current not diverted westward at the Mississippi Delta (Kelly et al. 1983) may carry eggs and larvae east towards the west coast of Florida where older stages of P. octonemus have been collected in the surf and estuaries (Powell et al. 1972, Ogren and Brusher 1977) even though, as noted, larvae have not been found. Shaw et al. (1985) sug- gested a similar model for Gulf menhaden Brevoortia pat r onus. Onshore surface Ekman transport components of the gyre seemingly could transport young P. octonemus directly inshore in the northwestern Gulf; however, this does not seem to be the case. Unlike P. octonemus, fishes of the brown shrimp community which spawn offshore and have pelagic eggs and larvae do not ap- pear to be transported into the white shrimp commu- nity; they are basically absent there (Chittenden and McEachran 1976). Depending on how long the young remain pelagic, current regimes are such that the spawning areas that produce young P. octonemus of the northwestern Gulf can lie almost anywhere in the western or central Gulf. Recent studies (Parker et al. 1979) cited in Rezak et al. (1983) found almost all surface drifters released west of a line from the Mississippi delta to the middle of the Yucatan Straits washed ashore along Texas; those released to the east were almost all found outside the Gulf along the Atlantic coast of Florida. This probably reflects a well-documented picture of Gulf circulation (Nowlin and McLellan 1967, Nowlin 1972, Behringer et al. 1977, Merrell and Morrison 1981, Merrell and Vasquez 1983): The major driving force for nearsur- face circulation in the deep eastern Gulf is the Loop Current which enters the GuJf via the Yucatan Straits, loops toward Alabama, and turns back to exit the Gulf via the Florida Straits. The major driving force in the western Gulf is a large, permanent anticyclonic gyre, centered at 23.5°N, and maintained by consistent pinching off and westward drift of Loop Current rings (Merrell and Morrison 1981) or wind stress curl across the western Gulf (Sturges and Blaha 1976, Merrell and Vasquez 1983). South of this gyre is a cyclonic gyre in the Bay of Campeche from which water joins a pre- dominantly northerly flow along the Mexican shelf (Nowlin 1972, Merrell and Morrison 1981, Rezak et al. 1983). Pelagic early stages of P. octoneynus have been collected along much of the western Gulf margin (Fig. 13). The northerly flow there could transport pelagic Dentzau and Chittenden Population dynamics of Polydactylus octonemus in the Gulf of Mexico 457 young towards Texas. The coastal area off south Texas and northern Mexico has converging currents much of the year, water which at least in part heads easterly or northerly along the shelf edge after leaving the coast near 26°N. Whatever their source, pelagic young trans- ported to and entrained in the gyre of the Texas- west- ern Louisiana shelf could be transported to inshore and estuarine nurseries there. Bathymetric distribution, recruitment, and movements Our finding that P. octonemus is most abundant at depths < 22 m in the northwestern Gulf agrees with Hildebrand (1954) and Chittenden and McEachran (1976), who only captured it in the demersal phase and considered it a member of the white shrimp commu- nity. Maximum depths recorded for the demersal form of this species in the Gulf are 66 m off Louisiana (Springer and Bullis 1956) and 65 m off Texas (Wohl- schlag et al. 1979). Our finding that P. octonemus appears late March- April and basically disappears late October-late De- cember agrees with the literature for Texas and Loui- siana including Gunter (1938b, 1945), Perret et al. (1971), and Gallaway and Strawn (1974). Some reports indicate a few fish occur in winter in the white and brown shrimp communities throughout the northern Gulf: (1) in January in Galveston Bay, Texas (Gallaway and Strawn 1974, Sheridan 1983), (2) at 18-22 m in January off Texas (Chittenden and McEachran 1976), (3) in shallow waters in January and February at Sabine Lake and Holly Beach, Louisiana (Perry and Carter 1979), and (4) at 40 m in early February off Alabama (Bullis and Thompson 1965). The similar periodicity of reported occurrence suggests similar movements and spawning periodicity throughout the northwestern and north-central Gulf. The spring recruitment we found in P. octonemus agrees with data in several studies from the northern Gulf including Gunter (1945) and Miller (1965) in Texas, Gunter (1938b) and Perret et al. (1971) in Louisiana, and Ogren and Brusher (1977) in northwestern Florida. Recruitment appears concurrent in estuaries and the shallow Gulf, because we collected fish 50-105 mm in the Gulf off Texas at < 9 m in April, the same period when Gallaway and Strawn (1974) and Landry (1977) collected fish in Galveston Bay 56-97 mm (40-69 mm SL) and 74-101 mm (53-72 mm SL), respectively. Moreover, demersal phase fish first appeared the same month in both Barataria Bay and the shallow Gulf off Louisiana (Gunter 1938b). Our finding that P. octonemus disperse offshore beginning in early summer agrees with the literature. Gunter (1945), Gallaway and Strawn (1974), and Per- ret et al. (1971) found abundance declined in estuaries in June- August or September, at about the same time (July-September) we found greatest abundance in the Gulf and size gradients from estuaries to offshore. We found the largest, oldest fish in the deeper areas of the white shrimp community or in the transition commu- nity (Chittenden et al. 1982) as they dispersed offshore. This pattern has also been suggested for other species of the Gulf white shrimp community such as Micropo- gonias undulatiis, Peprilus burii, Cynoscion arenariu^, and Larimus fasciatus (White and Chittenden 1977, Murphy 1981, Shlossman and Chittenden 1981, Stan- dard and Chittenden 1984). Age determination and growtln Age determination, size at age, and von Bertalanffy parameters have not been previously described for P. octonemus. Our calculations for growth are based on sizes at known time scaled to calculated hatching dates to give age. Regression coefficients and K values are the same regardless of hatching date, however, because curves are fitted to the same time dimension between the first and last collections of a cohort. We determined age in months and years by length- frequency analysis because we had almost 4 years of data from cruises so close in time that modes were easi- ly followed, and only one cohort usually occurred at a given time. As in Stenotmnus caprinu^ (Geoghegan and Chittenden 1982) and Larmius fasciatus (Standard and Chittenden 1984), modal-group progression analysis can be a superior method of age determination in P. octonemus because (1) spawning primarily occurs in one discrete period, (2) growth of large and small fish in a cohort appears uniform because variances were generally constant between cruises, (3) length frequen- cies are reasonably normally distributed within cohorts, and (4) age need only be determined for 1 or 2 years, ideal conditions for using length frequencies (Lagler 1956, Bagenal and Tesch 1978). Although it would be desirable to support our aging findings by the more time-consuming analysis of daily otolith increments (Jones 1986), the growth and other parameters we present provide useful upper or lower boundaries for true values (see comments in Methods). We reiterate that our sizes at age and von Bertalanffy parameters are apparent ones; they are affected to an unknown degree by a combination of recruitment, and, especially as the fish approach age I, emigration, gear avoidance, and change from demersal to pelagic behav- ior. Because of similar problems, Knudsen and Herke (1978) suggested length frequencies should not be used to estimate growth in Micropogonius undulatus. We do not agree, however, because the fundamental prob- lem is not the length-frequency method. Rather, it is 458 Fishery Bulletin 88(3|. 1990 how best to draw a random sample in time and space to characterize the population. The time dimension has three major aspects of variation (diel, annual, and monthly or seasonal), as does the spatial (along- shelf, across-shelf and estuary, and vertical). No pub- lished Gulf study to our knowledge addressed all six aspects, and few addressed any to the extent we did. Our design directly addressed all three time aspects and the across-shelf spatial dimension. We addressed along-shelf and estuarine variation indirectly by referencing the considerable literature. We were not able to address vertical variation, but apparently no published work has done that. The problem of getting representative data for P. ocioneTOws— indeed, for many other Gulf species— may not be fully solved till all six aspects of sampling are concurrently addressed in one study. Maximum size, life span, and mortality The largest P. octonemus we collected in the north- western Gulf (229 mm) is very close to that reported from off Virginia (233 and 235 mm ?L, Hildebrand and Schroeder 1928), the only Atlantic coast location with published sizes. Breder (1948) noted this species reaches one foot in length but gave no details to sup- port the figure; it is only 2-3 inches larger than our maximum sizes and might reflect a rough rounding-off by him. The largest P. octonemus we found is slightly larger than maximum sizes in other Gulf studies [off Texas: 213 mm (Gunter 1945), 210 mm (Compton and Bradley 1963), 213 mm (171 mm FL) (Chittenden and McEachran 1976); Off Mississippi: 197 mm (Christmas and Waller 1973); Off Florida: 209 mm ?L (Powell et al. 1972)]. However, these reported maximum lengths are similar to the apparent typical maximum size of 202-206 mm we found. Smaller maximum sizes in other Gulf reports such as Gunter (1938b), Ferret et al. (1971), Gallaway and Strawn (1974), and Perry and Carter (1979) may reflect gear selectivity t)r the fact they collected largely in estuaries, surf, and shallow Gulf areas. The studies cited, however, took place over much of the northern Gulf so the demersal stage of P. octonemus appears to show little, if any, east-west spatial variation in maximum sizes. Our finding the largest, oldest P. ncto)temus in the deeper areas of the white shrimp community may reflect a broad life-history pattern for species there, many of which show a positive size gradient towards offshore. This has been reported for strongly estuarine dependent species such as Micropogonius undulatus and Cynoscion arenarius (White and Chittenden 1977, Shlossman and Chittenden 1981), and more marine forms such as Peprilus burti and Larimus fasciatus (Murphy 1981. Standard and Chittenden 1984). The apparent typical maximum life span of 1 year we found for P. octonemus in the northwestern Gulf agrees with Chittenden and McEachran (1976) and appears reasonable, at least for the demersal stage, because the maximum and typical maximum sizes we observed are similar to, or larger than, all other values reported for the Gulf. The value oi ti = \ year might be a little low because P. octonemus disappears at 9-11 months of age to spawn in winter, and because we col- lected one fish 15 months of age in April. That one fish is the only instance in 71 cruises in which a second year class co-occurred with young-of-the-year, so our data clearly indicate that at least the demersal phase of life basically ends at 1 year. Though the literature indicates no east-west spatial variation in maximum size, and age by inference, P. octonemus might survive their first spawning and assume a pelagic behavior. Bullis (1961) observed a dense ball of small Polydactylus sp., which could have been young or adult P. octoyiemus, P. vir- ginicus, or P. oligodon (Fischer 1978), at the surface along the 500-fm curve in August in the western Caribbean. Midwater and surface collections are needed to resolve the duration of a pelagic phase in adult P. oc- tonemus; that would clarify the validity of the several life-history parameter values we have termed "ap- parent." In addition, collections with trawls larger than we and other workers have used would clarify if and how gear-avoidance affects parameter values. Despite these possible problems, however, it seems clear that few adult P. octonemus resume a demersal existence in the white or brown shrimp communities after age I. We collected only one fish older than 12 months, and Chittenden and McEachran (1976) caught no fish older than 1 2 months even though they caught 33 specimens on the white shrimp grounds in January. The apparent mean time- and cohort-specific total an- nual mortality rates we observed for P. octonemus (97-100%) agree with theoretical estimates of 90-100% (Royce 1972:238) if maximum life span is typically only 1-2 years. Though these are also apparent values, they are not that far from the range of values for another well-studied, exploited pelagic species, Brevoortia patronus. In that species, total annual mortality rate is 83-95% depending on age (Nelson and Ahrenholz 1986). Moreover, the population dynamics of P. octo- nemus in the northwestern Gulf appear generally similar to those reported for many other species there including Micropogonius undulatus (White and Chit- tenden 1977), Cynoscion arenarius (Shlossman and Chittenden 1981), Cynoscion nothuJi (DeVries and Chit- tenden 1982), Peprilus burti (Murphy 1981), Steno- toniiis caprinus (Geoghegan and Chittenden 1982), Larimus fasciatus (Standard and Chittenden 1984), and Leiostomus xnyithurus (Hata 1985). This similar- Dentzau and Chittenden Population dynamics of Folydactylus octonemus in the Gulf of Mexico 459 ity over a variety of species supports the suggestions (Chittenden and McEachran 1976, Chittenden 1977) that fishes of the white and brown shrimp communities, at least in the northwestern Gulf, have evolved a com- mon pattern of population dynamics which stresses small size, early maturity, short life spans, high mor- tality rates, and rapid turnover of biomass. Acknowledgments This manuscript is based on a thesis submitted by the senior author in partial fulfillment of the requirements for the M.S. degree, Texas A&M University. Prepara- tion of this manuscript was completed while the senior author was at the Florida Department of Environmen- tal Regulation and the junior author at College of William and Mary, Virginia Institute of Marine Science. We would like to thank R. Baker, M. Burton, T. Craw- ford, D. DeVries, V. Fay, P. Geoghegan, R. Grobe, D. Hata, M. Murphy, J. Pavela, M. Rockett, J. Ross, P. Shlossman, B. Slingerland, G. Standard, H. Yette and Captains H. Forrester, M. Forrester, R. Forrester, P. Smirch, and A. Smircic for assistance in field col- lections and data recording. H. Austin, R. Darnell, J. McEachran, J. Olney and D. Stilwell reviewed the manuscript and/or made helpful suggestions. E. Houde and M. Lieby provided information on egg and larval collections in the eastern Gulf. R. Case, M. Cuenco, and J. Cummings wrote and assisted with computer programs. Financial support was provided by the Texas Agricultural Experiment Station; by the Strategic Petroleum Reserve Program, Department of Energy; and by the Texas A&M University Sea Grant College Program, supported by the NO A A Office of Sea Grant, Department of Commerce. Citations Alverson, D.L.. and M.J. Carney 1975 A graphic review of the growth and decay of population cohorts. .J. Cons. Int. Explor. Mer .36:1.33-143. Anderson, W.W. 1968 P^ishes taken during shrimp trawling along the south Atlantic coast of the United States. 1931-3.5. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 570. 60 p. Anderson, W.W., and J.W. Gehringer 1965 Biological-statistical census of the species entering fish- eries in the Cape Canaveral area. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 514, 79 p. Austin, H.M., and J.I. Jones 1974 Seasonal variation of physical oceanographic parameters on the Florida Middle Ground and their relation to zooplankton biomass on the West Florida Shelf. Fla. Sci. 37:16-32. Bagenal, T.B., and F.W. Tesch 1978 Age and growth. 3rd ed. In Bagenal, T.B. (ed.). Methods for assessment of fish production in fresh waters, p. 101-136. Blackwell Sci. Puhl., Oxford, England. Behringer, D.W., R.L. Molinari, and J.F. Festa 1977 The variability of anticyclonic current patterns in the Gulf of Mexico. J. Geophys. Res. 34:5469-5476. Beverton, R.J.H.. and S.J. Holt 1957 On the dynamics of exploited fish populations. Fish. In- vest. Minist. Agric, Fish. Food (G.B.), Ser. II, 19. 533 p. Bowering, W.R. 1983 Age, growth, and sexual maturity of Greenland halibut, Reinhardtiiis hippoglossoides (Walhaum), in the Canadian Northwest Atlantic. Fish. Bull., U.S. 81:599-611. Breder, CM. Jr. 1948 Field book of marine fishes of the Atlantic Coast from Labrador to Texas. G.P. Putnam's Sons, NY, 332 p. Breder. CM. Jr., and D.E. Rosen 1966 Modes of reproduction in fishes. Nat. Hist. Press, Garden City, NY. 941 p. Briggs, J.C 1958 A list of Florida fishes and their distribution. Bull. Fla. State Mus. Biol. Sci. 2:223-318. Bright, T.J., and C Cashman 1974 Fishes. In Bright, T.J.. and L.H. Pequegnat (eds.). Biota of the West Flower Garden Bank, p. .339-409. Gulf Publ. Co., Houston. Bullis. H.R. Jr. 1961 Observations on the feeding behavior of white-tip sharks on schooling fishes. Ecology 42:194-195. Bullis, H.R. Jr., and J.R. Thompson 1965 Collections by the exploratory fishing vessels Oregon, Sili-Vr Bay. Combat, and Pelican made during 1956 to 1960 in the southwestern North Atlantic. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 510, 1.30 p. Chittenden. M.E. Jr. 1977 Simulations of the effects of fishing on the Atlantic croaker, Micropogon undutntus. Proc. Gulf Caribb. Fish. Inst. 29:68-86. Chittenden. M.E. Jr., and J.D. McEachran 1976 Composition, ecology, and dynamics of demersal fish com- munities on the northwestern Gulf of Mexico continental shelf, with a similar .synopsis for the entire Gulf. Sea Grant Publ. TAMU-SG-76-208, Sea Grant Coll. Prog., Texas A&M Univ.. College Station. 104 p. Chittenden, M.E. Jr., J.F. Pavela, and J.L. Ross 1982 Nekton. In. Hann, R.W. Jr., and R.E. Randall (eds.). Evaluation of brine disposal from the Bryan Mount Site of the Strategic Petroleum Reserve: Final report of 18 month post- disposal studies, p. (4-1H4-155). Contract DOE/P01O114-5, Texas A&M Univ. College Station (avail. NTIS, Springfield, VA]. Christmas, J.Y., and R.S. Waller 1973 Estuarine vertelirates, Mississippi, /n Christmas, J. Y. (ed.). Cooperative Gulf of Mexico estuarine inventory and study, Mississippi, p. 320-434. Gulf Coast Res. Lab.. Ocean Springs, MS. Compton. H.. and E. Bradley 1963 Analysisof populations of sports and commercial fin-fish and of factors which affect these populations in the coastal bays of Texas. Coastal Fish. Proj. Reps., Proj. MF-R-4, Job 14, Tex. Parks Wildl. Dep., Austin. 14 p. Dahlberg, M.D. 1975 Guide to coastal fishes of Georgia and nearby states. Univ. Georgia Press, Athens, 186 p. 460 Fishery Bulletin 88(3), 1990 Dentzau, M.W. 1985 Aspects of the life history of the Atlantic threadfin, Polydactylus octonemus. MS thesis, Texas A&M Univ., Col- lege Station, 69 p. DeSylva, D.P. 1984 Polynemoidei: Development and relationships. In Moser, H.G., et al. (eds.). Ontogeny and systematics of fishes, p. 540-541. Spec. Publ. 1, Am. Soc. Ichthyol. Herpetol., Allen Press, Lawrence, KS, 760 p. DeVries, D.A., and M.E. Chittenden Jr. 1982 Spawning, age determination, longevity, and mortality of the silver seatrout, Cynoscion nothus, in the Gulf of Mex- ico. Fish. Bull., U.S. 80:487-500. Fabens, A.J. 1965 Properties and fitting of the von Bertalanffy growth curve. Growth 29:265-289. Fischer, W. (editor) 1978 FAG species identification sheets for fishery purposes. Western Central Atlantic (fishing area 31), vols. 1-7. FAO, Rome, pag. var. Gallaway, B.J., and K. Strawn 1974 Seasonal abundance and distribution of marine fishes at a hot-water discharge in Galveston Bay, Texas. Contrib. Mar. Sci. Univ. Tex. 18:71-137. Gallucci, V.F., and T.J. Quinn II 1979 Reparameterizing, fitting, and testing a simple growth model. Trans. Am. Fish. Soc. 108:14-25. Geoghegan, P., and M.E. Chittenden Jr. 1982 Reproducti(in, movements, and population dynamics of the longspine porgy, Stenotomuis capriniL'i. Fish. Bull., U.S. 80:.523-.540. Gulland, J.,\. 1969 Manual of methods for fish stock assessment. Part I. Fish population analysis. FAO Manuals in Fish. Sci. 4, 154 p. Gunter, G. 1938a The relative numbers of species of marine fish on the Louisiana coast. Am. Nat. 72:77-83. 1938b Seasonal variations in abundance of certain estuarine and marine fishes in Louisiana, with particular reference to life histories. Ecol. Monogr. 8:313-346. 1945 Studies on marine fishes of Texas. Publ. Inst. Mar. Sci. Univ. Tex. 1, 190 p. Hastings, R.W., L.H. Ogren. and M.T. Mabry 1976 Observations on the fish fauna associated with offshore platforms in the northeastern Gulf of Mexico. Fish. Bull, U.S. 74:387-402. Hata. D. 1985 Aspects of the life history and population dynamics of the spot, Leiostomus xanthurus. in the northwestern Gulf of Mexico. MS thesis, Texas A&M Univ., College Station, 87 p. Hildebrand, H.H. 1954 A study of the fauna of the brown shrimp (Penneus aztecus Ives) grounds in the western Gulf of Mexico. Publ. Inst. Mar. Sci. Univ. Tex. 3:229-366. Hildebrand. S.F., and W.C. Schroeder 1928 Fishes of Chesapeake Bay. Bull. U.S. Bur. Fish. 43, 366 p. Houde, E.D., J.C. Leak, C.E. Dowd. S.A. Berkeley, and W'.J. Richards 1979 Ichthyoplankton abundance and diversity in the eastern Gulf of Mexico. A report to the Bureau of Land Management, Contract AA5.50-CT7-28, by Rosenstiel School Mar. Atmos. Sci., LIniv. Miami, 546 p. Hunter, J.R, 1981 Feeding ecology. In Lasker, R. (ed). Marine fish lar\-ae, p. 34-.S9. Univ. Wash. Sea Grant Prog., Seattle, 131 p. Jearld, A. Jr. 1983 Age determination. In Nielsen, L.A., and D.L. Johnson (eds.). Fisheries techniques, p. 301-324. Am. Fish. Soc, Bethesda, 468 p. Jones, C. 1986 Determining age of larval fish with the otolith increment technique. Fish. Bull., U.S. 84:91-103. Juneau, C.L. Jr. 1975 An inventory and study of the Vermilion Bay-Atcha- falaya Bay complex. Phase II, Biology. Tech. Bull. 13, La. Wildl. Fish. Comm., Baton Rouge. 153 p. Kagawade, P.V. 1970 The polynemid fishes of India. Bull. Cent. Mar. Fish. Res. Inst. 18, 69 p. Kelly, F.J. Jr., J.E. Schmitz, R.E. Randall, and J.D. Cochrane 1981 Physical oceanography. In Hann, R.W. Jr., and R.E. Randall (eds.). Evaluation of brine disposal from the Bryan Mound site of the Strategic Petroleum Reserve Program. Final report of twelve month post-disposal studies, p. 1-105. Con- tract DOE/P010114-4, Texas A&M Univ., College Station [avail. NTIS, Springfield. VA]. 1983 Physical oceanography. In Hann, R.W. Jr., and R.E. Randall (eds.). Evaluation of brine disposal from the Bryan Mound site of the Strategic Petroleum Reserve Program: An- nual report for September 1981 through August 1982, p. 1-134. Contract DOE/P010114-6, Texas A&M Univ., College Station [avail. NTIS, Springfield, VA]. 1984 Physical oceanography. In Hann, R.W. Jr., C.P. Giam- mona, and R.E. Randall (eds.). Offshore oceanographic and en- vironmental monitoring services for the Strategic Petroleum Reserve: Annual report for the Bryan Mount Site from Septemlier 1982 through August 1983, p. 1-122. Contract DOE/P010850-2, Texas A&M Univ., College Station [avail. NTIS, Springfield, VA], Klawe, W.L., and F.G. Alverson 1964 Occurrence of two species of young threadfin, Polydac- tylus vpercularis and P. approxi>nan.<:, in the offshore waters of the eastern tropical Pacific Ocean. Pac. Sci. 18:166-173. Knight, W. 1968 Asymptotic growth: an example of nonsense disguised as mathematics. J. Fish. Res. Board Can. 25:1303-1307. Knudsen, E.E., and W.H. Herke 1978 Growth rate of marked juvenile Atlantic croakers, Micropugon undulatus, and length of stay in a coastal marsh nursery in southwest Louisiana. Trans. Am. Fish. Soc. 107: 12-20. Lagler, K.F. 1956 Freshwater fishery biology. 2d ed. Wm. C. Brown Co., Dubuque, I A, 421 p. 1978 Capture, sampling and examination of fishes. 3rd ed. In Bagenal, T.B. (ed.). Methods for assessment of fish produc- tion in fresh waters, p. 7-47. Blackwell Sci. Publ., Oxford, England. Landry, A.M. Jr. 1977 Life history and susceptibility of fishes in Galveston Bay, Texas to power-plant cooling-water operations. Ph.D. thesis, Texas A&M Univ.. College Station, 546 p. Martin, F.D., and G.E. Drewry 1978 Development of fishes of the mid-Atlantic Bight, Vol. VI. U.S. Fish Wildl. Serv., Biol. Serv. Prog., 416 p. McFarland, W.N. 1963 Seasonal change in the number and the biomass of fishes from the surf at Mustang Island, Texas. Publ. Inst. Mar. Sci. Univ. Tex. 9:91-105. Dentzau and Chittenden: Population dynamics of Polydactylus octonemus in the Gulf of Mexico 461 Merrell, W.J. Jr., and J.M. Morrison 1981 On the circulation of the western Gulf of Mexico with observations from April 1978. J. Geophys. Res. 86:4181-4185. Merrell. W.J. Jr., and A.M. Vasquez 1983 Observations of changing mesoscale circulation patterns in the western Gulf of Mexico. J. Geophys. Res. 88:7721-7723. Miller, J.M. 1965 A trawl survey of the shallow Gulf fishes near Port Aran- sas, Texas. Publ. Inst. Mar. Sci. Univ. Tex. 10:80-107. Moyle, P.B., and J.J. Cech Jr. 1982 Fishes: An introduction to ichthyology. Prentice-Hall, Englewood Cliffs, NJ, 593 p. Murphy, M.D. 1981 Aspects of the life history of the gulf butterfish, Peprilus hia-ti. M.S. thesis, Texas A&M Univ., College Station, 77 p. Murphy. M.D.. and M.E. Chittenden Jr. Unpubl. Reproduction, movements, and population dynamics of the gulf butterfish, Peprilus burti. Mar. Res. Lab., Fla. Dep. Nat. Resour.. St. Petersburg, 66 p. Nelson, W.R., and D.W. Ahrenholz 1986 Population and fishery characteristics of gulf menhaden, Bremortin patnmus. Fish. Bull.. U.S. 84:311-326. Nowlin. W.D. Jr. 1972 Winter circulation patterns and property distributions. In Capurro, L.R.A., and J.L. Reid (eds.), Texas A&M Univer- sity oceanographic studies, vol. 2, p. 3-51. Gulf Publ. Co.. Houston. Nowlin, W.D. Jr.. and H.J. McLellan 1967 A characterization of the Gulf of Mexico waters in winter. J. Mar. Res. 25:29-59. Ogren, L.H., and H.A. Brusher 1977 The distribution and abundance of fishes caught with a trawl in the St. Andrew Bay system, Florida. Northeast Gulf Sci. 1:83-105. Parker, R.P., J.M. Morrison, and W.D. Nowlin Jr. 1979 Surface drifter data from the Caribbean Sea and Gulf of Me.xico. 1975-1978. Office Nav. Res. #N00014-75-C-0537. Tech. Rep. 79-8-T. Dep. Oceanogr., Tex. A&M Univ.. College Station. 157 p. Pequegrnat, L.H., J.H. Wormuth, J.D. McEachran. S.P. Berkowitz. A.D. Hart, and M.A. Daher 1977 Neuston project. In Groover, R.D. (ed.). Environmen- tal studies, south Texas outer continental shelf, biology and chemistry. The 1976 final report to the Bureau of Land Man- agement, p. (7-1H7-83). (E-lHE-75). Contract AA550- CT6-17, Texas A&M Univ.. College Station. Perret. W.S.. W.R. Latapie, J.F. Pollard. W.R. Mock, G.B. Adkins, W.J. Gaidry. and C.J. White 1971 Fishes and invertebrates collected in trawl and seine samples in Louisiana estuaries. In Perret, W.S., B.B. Bar- rett, W.R. Latapie. J.F. Pollard, W.R. Mock, G.B. Adkins, W.J. Gaidry, and C.J. White (eds.). Cooperative Gulf of Mexico estuarine inventory and study. Louisiana. Phase 1, Area description, and phase 4, Biology, p. 39-105. La. Wild). Fish. Comm., Baton Rouge. Perry. W.G.. and B.J. Carter 1979 Seasonal occurrence of fishes collected from beach sein- ing, southwest Louisiana. Proc. La. Acad. Sci. 42:24-38. Powell, D., L.M. Dwinell, and S.E. Dwinell 1972 An annotated listing of the fish reference collection at the Florida Department of Natural Resources Marine Research Laboratory. SSR 36, Mar. Res. Lab., Fla. Dep. Nat. Resour., St. Petersburg, 179 p. Reid, G.K. 1955 A summer study of the biology and ecology of East Bay, Texas. Part H. The fish fauna of East Bay, the Gulf beach, and summary. Tex. J. Sci. 7:430-453. 1956 Observations on the eulittoral ichthyofauna of the Texas Gulf coast. Southwest. Nat. 1:157-165. Rezak, R.. T.J. Bright, and D.W. McGrail 1983 Circulation, hydrography, and distribution of suspend- ed sediment on the continental shelves in the Gulf of Mex- ico. In Norman, R. (ed.), Reefs and banks of the northwestern Gulf of Mexico: Their geological, biological and physical dynamics. Final report, p. 31-94. Tech. Rep. 83-1-T, Dep. Oceanogr., Texas A&M Univ., College Station. Ricker, W.E. 1973 Linear regressions in fishery research. J. Fish. Res. Board Can. 30:409-434. 1975 Computation and interpretation of biological statistics of fish populations. Fish. Res. Board Can. Bull. 191, 382 p. Royce. W.F. 1972 Introduction to the fishery sciences. Acad. Press. NY, .351 p. Shaw, R.F., W.J. Wiseman Jr., R.E. Turner, L.J. Rouse Jr., and R.E. Condrey 1985 Transport of larval Gulf menhaden Bj-evoortia patronus in continental shelf waters of western Louisiana: A hypothesis. Trans. Am. Fish. Soc. 114:452-460. Sheridan, P.P. 1983 Abundance and distribution of fishes in the Galveston Bay system. 1963-1964. Contrib. Mar. Sci. Univ. Tex. 26:143-163. Shirota. A. 1970 Studies on the mouth size of fish larvae. Bull. Jpn. Soc. Sci. Fish. 36:3.53-368. [Transl. by Fish. Res. Board Can. Transl. Serv., 1978.) Shlossman. P. A., and M.E. Chittenden Jr. 1981 Reproduction, movements, and population dynamics of the sand seatrout, Ci/noscion arenaritis. Fish. Bull., U.S. 79:649-669. Springer. S.. and H.R. Bullis Jr. 1956 Collections by the Oregon in the Gulf of Mexico. U.S. Fish Wildl. Serv.. Spec. Sci. Rep. Fish. 196, 134 p. Standard. G.W.. and M.E. Chittenden Jr. 1984 Reproduction, movements, and population dynamics of the banded drum, Lanmus fasciatus, in the Gulf of Mexico. Fish. Bull.. U.S. 82:.337-363. Sturges. W.. and J. P. Blaha 1976 A western boundary current in the Gulf of Mexico. Science (Wash. DC) 192:367-369. Vick. N.C. 1964 The marine ichthyofauna of St. Andrews Bay, Florida, and nearshore habitats of the northeastern Gulf of Mexico. Proj. 286-D, Ref 64-19T, Dep. Oceanogr. Meteorol.. Texas A&M Univ., College Station. 77 p. White. M.L.. and M.E. Chittenden Jr. 1977 Age determination, reproduction, and population dynam- ics of the Atlantic croaker, Micropugonias undulatus. Fish. Bull., U.S. 75:109-123. Wilk. S.J.. and M.J. Silverman 1976 Fish and hydrographic collections made by the research vessels Dolphin and Delaware II during 1968-1972 from New York to Florida. NOAA Tech. Rep. NMFS SSRF 697, Natl. Oceanic Atmos. Adm., Natl. Mar. Fish. Serv., 159 p. 462 Fishery Bulletin 88(3), 1990 Wohlschlag, D.E., J.F. Cole, M.E. Dobbs. E.F. Vetter, and R.H. Moore 1977 Demersal fishes. In Groover, R.D. (ed.), Environmen- tal studies south Texas outer continental shelf, biology and chemistry. The 1976 final report to the Bureau of Land Man- agement, p. (10-1)-(10-162), (H-lHH-131). Contract AA550- CT6-1T, Mar. Sci. Inst., Univ. Tex., Port Aransas. Wohlschlag. D.E., R.M. Yoshiyama, M. Dobbs, E. Findley, and E.F. Vetter 1979 Demersal fishes, /k Flint, R.W., and C.W. Griffin (eds.). Environmental studies, south Texas outer continental shelf, biology and chemistry. The 1977 final report to the Bureau of Land Management, p. (18-1H18-99), (Q-lHQ-108). Contract AA.5.50-CT7-11, Mar. Sci. Inst., Univ. Tex., Port Aransas. Wormuth, J.H., L.H. Pequegnat, J.D. McEachran, A.D. Hart, J. Cummings. M.A. Daher. and C. Odumodu 1979 Neuston project. /«. Flint, R.W., and C.W. Griffin (eds.), Environmental studies, south Texas outer continental shelf, biology and chemistry. The 1977 final report to the Bureau of Land Management, p. tl5-lH15-66), (N-lHN-168). Contract AA550-CT7-11, Mar. Sci. Inst., Univ. Tex.. Port Aransas. Abstract.- The feeding habits of late-larval and early-juvenile wall- eye pollock Theragra chalcogramma from a coastal nursery habitat in the western Gulf of Alaska were exam- ined in relation to fish size (10.0-29.9 mm SL) and site of collection. Pseu- docalanus sp. was the dominant prey. Copepod nauplii and Pseudocalanus copepodids decreased in importance in the diet as fish size increased, con- comitant with an increase in impor- tance oi a.du\t Pseudocalanus sp. and Calanus spp. Other prey showed no clear relation to fish size. A compari- son of diet between collection sta- tions revealed considerable variabil- ity. Along two transects perpendicu- lar to the Alaska Peninsula, between- station diet overlap was observed to be high for one transect, and low for the other. Along a transect parallel to the Alaska Peninsula, some adja- cent stations yielded high diet over- lap, and other yielded low overlap, while the most distant stations pro- duced high diet overlap. Patchiness of food resources likely contributed subtantially to these patterns. One possible consequence of the ingestion of different prey in different locales is that differential growth rates could result. Feeding Ecology of Late-Larval and Early-Juvenile Walleye Pollock Theragra chalcogramma from the Gulf of Alaska in 1987* Jill J. Grover College of Oceanography, Oregon State University Hatfield Marine Science Center, Newport, Oregon 97365 The recent discovery of large spawn- ing aggregations of walleye pollock Theragra chalcogramma in Shelikof Strait from late March to mid-April suggests that this is the principal spawning center for the species in the Gulf of Alaska (Dunn et al. 1984, Kendall et al. 1987). The concentrated mass of eggs produced in Shelikof Strait in early spring and the ensu- ing dense patch of larvae provide un- common access to the early life his- tory of the species. The spawning aggregations produce distinct cohort patches (20 x 50 km or larger) of eggs at depths greater than 150 m (Incze et al. 1989). After about 14 days, the eggs hatch (Dunn and Matarese 1987, Kim 1989) and larvae ascend to the upper 50 m (Incze et al. 1989). The larval patches can be followed for up to 30 days as they drift with prevailing currents to the southwest (Kendall et al. 1987, Hinckley et al. 1989, Incze et al. 1989). The cohesive- ness of the patches likely results from the restricted spawning area within Shelikof Strait (Hinckley et al. 1989) as well as Alaska Coastal Current cir- culation patterns (Incze et al. 1989). Ichthyoplankton has been surveyed regulai'ly in and around Shelikof Sti'ait since 1979, and much is known re- garding the early life history of wall- eye pollock in this area (e.g., Kendall et all 1987, Incze et al. 1989, Kim 1989, Yoklavich and Bailey 1989). However, considerably less is known Manuscript accepted 7 May 1990. Fishery Bulletin, U.S. 88:463-470. •Contribution FOCI-0105 to Fisheries-Ocean- ography Coordinated Investigations, NOAA. about life history of late-larval and juvenile stages within nursery habi- tats in the western Gulf of Alaska. A survey in June and July 1987 col- lected late-larval and early-juvenile pollock from a coastal nursery area over the continental shelf along the Alaska Peninsula (Hinckley et al. 1989). While the center of distribu- tion was between the Shumagin and Semidi Islands, a large number of late larvae and juveniles were collected over a broad area running nearly the length of the Alaska Peninsula. The broad distribution enabled us to ex- amine whether diet varied along the length of the Alaska Peninsula and along a nearshore-offshore transect. Although the feeding habits of the late larvae have been previously re- ported from several locations, in all cases results have been pooled across collection stations (Kamba 1977, Clarke 1978, Cooney et al. 1980. Lee 1985, Kendall et al. 1987). In the present study, feeding habits of late larvae were analyzed by size and these data are compared to previous reports. Further, diet was analyzed by collection station, and between- station differences were examined using a measure of diet overlap. Materials and methods Pollock larvae were collected in the Gulf of Alaska during June 1987 by the RV Miller Freeman. Collections were made using a 5-m- Methot frame trawl (Methot 1986) (2 x 3 mm oval mesh) that was fished obliquely 463 464 Fishery Bulletin 88(3), 1990 59° N 54° N I64°W I52°W Figure 1 Location of collection stations in the Gulf of Alaska. Stations represent one transect parallel to the Alaska penin- sula, and two perpendicular offshore transects, with stations 38 and 48 located near the shelf break. from 15 m off-bottom (to a maximum of 300 m depth) to the surface (Hinckley et al. 1989). Samples from 10 stations, representing 2 nearshore-offshire transects (stations 38, 40, and 42 and stations 44, 45, and 48) and one transect parallel to the Alaska Peninsula (stations 21, 30, 32, 42, 44, and 58; see Figure 1), that were collected from 20 to 25 June, were the focus of this study. The total number of fish examined was 490. Samples were preserved in 5% formalin at sea, and transferred to 70% ethanol prior to their examination. Specimen shrinkage most likely occurred as a result of fixation. After the standard length (SL) of each larva was measured, the digestive tract was removed. Contents of the entire digestive tract were evaluated. Gut con- tents were teased out and prey items were identified: invertebrate eggs, copepod nauplii, barnacle nauplii, euphausiid larvae, pteropods, other non-copepod prey, and copepods mduding Pseudocalanus sp. adults (CVI), Pseudocalanus copepodids (CI-CV), Acartia longire- mis, Acartia sp., Calanus marshallae, Calanus sp., Centropages abdominalis, Eucalanus bungii, Metridia sp., Oithcma sp., Oncaea sp., unidentified calanoid cope- podids < 1.5 mm, and unidentified calanoid copepodids >1.5 mm. Copepod eggs comprised over 99% of the "invertebrate eggs" category and Oithona similis com- prised over 99% of the "Oithona sp." category. While the taxonomy of the genus Pseudocalanus has recent- ly been clarified (Frost 1989), no attempt at specific identifications of Pseudocalanus was made in the pres- ent study. For all copepods except Pseudocalanus sp., copepodid (CI-CV) through adult stages were consid- ered together. Copepodids were inconsequential in all species except Calanus sp. Diet was analyzed in terms of numerical percentage composition (%A''), volumetric percentage composition (%VOL), and percent frequency of occurrence (WoFO). Prey volumes were estimated using prey dimensions (Grover and 011a 1987). Over 8000 prey items were measured for prey volume calculations. The three anal- yses {%N, %VOL, and %F0) were combined to yield a more comprehensive assessment of prey importance, the index of relative importance (IRI = (%N + % VOL ) X %i^0) (Pinkas et al. 1971). Data were pooled across stations for analysis by size class. The four size classes were defined as 10.0-14.9, 15.0-19.9, 20.0-24.9, and 25.0-29.9 mm SL. For station-by-station analyses, data were pooled across size classes. Diet overlap was calculated using an index of pro- portional similarity (Wallace 1981, Linton et al. 1981, Kohn and Riggs 1982) defined as / " PS = 1 0.5 I 1^^ where PS is the index of proportional similarity or over- lap, Pj., is the proportion of prey category i in the diet at location x, Py, is the proportion of prey category i in the diet at location y. and n is the number of prey categories. Mean percent volume was used as a mea- sure of diet for these calculations (Wallace 1981). Al- though this index has been traditionally used to exam- ine diet overlap between species at one location, I have used it to examine diet overlap between locations for Graver: Feeding ecology of Theragra chalcogramma in the Gulf of Alaska 465 Table I Mean lengths of late-larval and early-juvenile walleye pollock 1 collected at 10 stations in the Gulf of Alaska in 1987 (with standard deviations), and date and time of collection. Station Standard length Collection Time no. (mm) Date fli) 21 16.1 (2.11) 20 June 2156 30 18.2 (2.27) 22 June 0144 32 17.4 (1.74) 22 June 0704 38 19.3 (3.27) 22 .June 2205 40 18.6 (2.06) 23 June 0114 42 18.4 (2.44) 23 June 0558 44 17.1 (2.47) 24 June 1756 45 19.9 (2.40) 24 June 1956 48 20.2 (2.55) 25 June 0117 58 19.0 (3.36) 25 June 2314 a single species. This index was chosen because it does not require resource-availabihty data (Wallace 1981), it is independent of sample size (Kohn and Riggs 1982), and it estimates overlap accurately for real overlaps between 7 and 85%, whereas three other indices per- formed poorly (Linton et al. 1981). PS values greater than 60% are generally considered to demonstrate significant overlap (Mathur 1977, Wallace and Ramsey 1983). In order to minimize size-related bias in the measurement of diet overlap, proportional similarity indices were calculated for only the 15.0-19.9 mm SL size class. Results Standard length tended to increase with date of col- lection (Table 1, correlation coefficient r = 0.678, P <0.05). The dominant size class for all stations was 15.0-19.9 mm SL, accounting for 62% of the larvae, with a minimum of 21 larvae of this size collected at each station. The mean lengths of larvae from five sta- tions located along a transect parallel to the Alaska Peninsula were significantly different (P< 0.001) (ANOVA, Zar 1974), while of the two transects perpen- dicular to the Alaska Peninsula, larval lengths were significantly different (P<0.001) along only one of these transects (stations 44, 45, and 48). Table 2 Diet of walleye pollock from the Gulf of Alaska in 1987, by size class, expressed as %N, %VOL , %F0, and %IRI. 10.0-14.9 mm SL 15.0-19.9 mm SL 20.0-24.9 mm SL 25.0-29.9 mm SL (Ar= 50) (iV = 302) (JV = 127) (N- = 7) %N %VOL %F0 %IRI %Ar %VOL %F0 %IRI %N %VOL %F0 %IRI %N %VOL %F0 %IRI Invertebrate eggs 15.8 0.8 66.0 7.3 25.8 0.9 62.6 12.4 27.1 0.6 67.7 13.3 15.9 0.3 42.9 4.7 Copepod nauplii 25.3 2.8 86.0 16.1 9.3 0.7 63.2 4.7 1.5 0.1 40.9 0.5 0.5 <0.1 14.3 <0.1 Barnacle nauplii 0.1 <0.1 4.3 <0.1 0.1 <0.1 4.7 <0.1 Euphausiids 1.1 6.4 24.0 1.2 0.8 3.0 20.5 0.6 0.7 1.9 20.5 0.4 0.7 1.3 28.6 0.4 Pteropods 0.6 0.3 8.0 <0.1 1.6 0.5 16.9 0.3 0.4 0.1 6.3 <0.1 Other 0.2 0.1 5.3 <0.1 <0.1 <0.1 3.1 <0.1 Pseudocalanus sp. 8.8 30.7 82.0 21.6 14.6 35.5 79.5 29.5 26.6 43.1 87.4 43.3 37.2 40.7 100.0 52.9 Pseudocalanus copepodids 6.7 y.i 76.0 8.0 7.5 7.2 66.6 7.3 7.8 5.0 69.3 6.3 8.6 3.7 57.1 4.8 Acartia longiremis 6.4 13.8 64.0 8.6 9.2 14.0 66.2 11.4 6.9 7.0 69.3 6.8 2.1 1.4 42.9 1.0 Acartia sp. 1.0 1.5 22.0 0.4 0.9 1.0 18.2 0.2 0.6 0.4 11.0 0.1 Calanus marshallae <0.1 0.9 2.0 <0.1 0.3 3.5 10.3 0.3 0.9 7.2 27.6 1.6 3.2 18.3 57.1 8.4 Calanus sp. 0.1 2.5 5.0 0.1 0.4 4.6 11.8 0.4 0.7 5.7 14.3 0.6 Cenfropages ahihiminalis 0.2 1.3 6.0 <0.1 0.2 0.9 8.9 0.1 0.3 0.8 14.2 0.1 Eucfdanus hungii 0.2 6.5 5.0 0.2 0,6 11.3 15.0 1.2 0.9 12.3 14.3 1.3 Metridia sp. <0.] 0.9 2.0 <0.1 <0.1 0.4 1.3 <0.1 Unident. copepodids < 1.5 mm 14.4 20.3 96.0 22.1 15.5 15.3 95.7 21.8 17.6 11.6 95.3 19.8 19.4 8.6 85.7 16,3 Unident. copepodids >1.5 mm <0.1 0.6 2.0 <0.1 0.4 2.9 12.3 0.3 0.8 4.3 24.4 0.9 1.6 6.1 42.9 2.2 Oithona sp. 18.5 10.2 76.0 14.5 12.9 4.9 80.8 10.7 7.6 1.9 76.4 5.2 9.2 1.6 100.0 7.4 Oncaea sp. 1.0 0.4 18.0 0.1 0.3 0.1 7.3 <0.1 0.1 <0.1 2.4 <0.1 466 Fishery Bulletin 88(3), 1990 o a: LJ 100- 150- 200- 250 149 199 24 9 299 10 0- 15 0- 20 0- 25 0- 14 9 19 9 24 9 29 9 SIZE CU\SS (mm) (ZZ) (ZZ) rTTTTI ALL OTHER PRE^|• OITHONA SP UNIDENT COPEPODIDS <1 5mm ALL CALANUS SPP ACARTIA LONGIREMIS PSEUD0CAL6NUS COPEPODIDS PSEUDOCALANUS SP COPEPOD NAUPLII INVERTEBRATE EGGS 10 0- 15 0- 20-0- 25 0- 14 9 19.9 24 9 29 9 SIZE CLASS (mm) Figure 2 The diet of larval pollock in the Gulf of Alaska in June 1987, t)y size class, in terms of %N, %l'OL, and %IRL Diet as a function of size Small unidentified calanoid copepodids(i.e.. <1.5 mm) were the primary prey {IRl) for larvae 10.0-14.9 mm SL, followed by P^eudocalanus sp. (Table 2, Fig. 2). Copepod nauplii, which ranked as a major food item for only this size class, was third in importance, fol- lowed by Oitiwna sp. and Acartia longiremis. For larvae 15.0-19.9 mm SL, the same two prey items dominated the diet, but Paeudocalanus sp. was the most important item and small (unidentified) cala- noid copepodids were second (Table 2, Fig. 2). Inverte- brate eggs, A. hngiremis. and Oitfuma sp. ranked third through fifth. For fish 20.0-24.9 mm SL, Pseudocalanus sp. con- tributed twofold more to the diet than any other item (Table 2, Fig. 2). Small calanoid copepodids, inverte- brate eggs, and yl. longiremiH were second, third, and fourth most important prey, followed hy Pneudocalanus copepodids. Pseudocalanus sp. continued to he the dominant prey for the largest size class that was examined, 25.0-29.9 mm SL, comprising over 52% of the diet, with small calanoid copej>odids ranking second in importance (Table 2, Fig. 2). These were followed by Calunuji mar- shallae. a species whose only major contribution to the diet was in this size class, Oifhoua sp., and Psrudo- calanus copepodids. Considering adult and copepodid forms together, Pseudocalanus sp. was the principal prey for ail sizes of larvae and early juveniles (Table 2, Fig. 2). The con- tribution of Pseudocalanus sp. adults and all Calnnus spp. increased directly in relation to fish size. On the Grover Feeding ecology of Theragrs chalcogramma in the Gulf of Alaska 467 Table 3 Diet of larval walleye pollock in the Gulf of Alaska in 1987, by station, expressed as %IRI. Station number 21 30 32 38 40 42 44 45 48 58 Invertebrate eggs 6.5 0.3 16.7 12.4 1.4 28.3 13.3 26.1 6.9 0.2 Copepod nauplii 9.9 5.5 9.1 0.6 0.6 1.2 1.3 0.6 0.7 10.2 Barnacle nauplii <0.1 0.1 <0.1 0.1 <0.1 Euphausiids 5.2 0.4 0.7 1.2 0.5 0.2 0.7 <0.1 <0.1 <0.1 Pteropods 10.7 0.9 <0.1 <0.1 <0.1 Other <0.1 0.4 <0.1 <0.1 <0.1 <0.1 Pseudocalanus sp. 23.3 4.5 7.4 32.8 5.2 31.4 39.2 50.0 49.1 41.2 Pseudocalanus copepodids 11.3 0.1 2.8 13.2 0.3 5.3 11.1 7.1 4.6 6.0 Acartia longiremis 1.7 22.2 7.9 5.6 53.7 4.5 5.8 1.0 6.1 11.2 Acartia sp. 0.1 0.2 5.1 <0.1 3.6 <0.1 <0.1 Calanus marshaUae 0.5 7.0 0.2 <0.1 0.7 0.1 0.4 0.5 Calanus sp. <0.1 0.1 <0.1 0.2 6.6 Centropages abdominalis <0.1 0.1 0.3 <0.1 0.1 0.2 <0.1 0.4 <0.1 Eucalaniis bungii 3.1 13.5 Metridia sp. <0.l <0.1 <0.1 <0.1 Unident. copepodids <1.5 mm 16.1 37.8 15.7 16.7 17.2 22.7 23.3 13.2 21.7 20.1 Unident. copepodids >1.5 mm <0.1 21.1 0.8 0.7 <0.1 <0.1 <0.1 0.4 0.1 Chthoiia sp. 25.7 7.2 22.7 6.5 5.6 2.4 4.4 1.6 3.1 10.4 Oiicaea sp. 0.1 0.3 <0.1 0.1 <0.1 <0.1 <0.1 other hand, the contribution of Pseudocalanus cope- podids and copepod naupHi decreased with increasing fish size. Other major prey categories (i.e., small cala- noid copepodids, invertebrate eggs, A. longiremis, and Oithona sp.) showed no clear relation to fish size. Diet 3S related to site of collection Pseudocalanus sp. was the most important prey at 6 of the 10 stations, accounting for over 30% (IRI) of the diet at each of these locations (Table 3, Fig. 3). Small calanoid copepodids were of secondary importance, ranging from 13.2 to 23.3%, while invertebrate eggs, copepod nauplii, Pseudocalanus copepodids, and A. longiremis and Oithona sp. each contributed more than 10% to the diet at one or more of these stations. At two of the remaining four stations Oithona sp. was the primary prey (22.7-25.7%). Small calanoid copepodids were most important at another staion (37.8%i), and A. longiremis was dominant at the re- maining one, comprising more than 53% of the diet. Additionally, invertebrate (mostly copepod) eggs, cope- pod nauplii, pteropods, Pseudocalanus sp., Pseudocala- nus copepodids, A. longiremis, Eucalanus bungii, and small and large unidentified calanoid copepodids (i.e., >1.5 mm), were well represented in the diet at one or more of the stations that were not dominated by Pseu- docalanus sp. Diet overlap between stations Values of the index of proportional similarity (PS) varied widely, franging from 0.252 to 0.833 (Table 4). Of 11 pairs of adjacent stations, 45% showed no sig- nificant diet overlap, while of 34 pairs of non-adjacent stations 50% showed no significant overlap. Two sta- tions, 30 and 40, were quite dissimilar from all other stations, and station 32 showed significant overlap only in relation to stations 21 and 58. Wlien examining diet overlap for the six stations that comprise the transect parallel to the Alaska Peninsula, a high degree of variation was seen in PS values (Table 5). At the southwestern end of the transect the lowest PS value was obtained for stations in closest proximity (21 and 30), while at the northeastern end of the tran- sect the greatest PS value was obtained for stations in closest proximity (58 and 44). At the same time, diet at the two most distanct stations (21 and 58) showed great similarity. Along one of the two transects perpendicular to the Alaska Peninsula (stations 44, 45, and 48), PS values indicated considerable overlap (Table 6). In contrast, a different pattern was obtained along the other per- pendicular transect (stations 38, 40, and 42) where PS values were wide-ranging. For each of the two tran- sects the greatest diet overlap was between the near- shore and offshore exti'emes of the transect (i.e., sta- tions 44 and 48, and stations 38 and 42). 468 Fishery Bulletin 88(3). 1990 30 32 38 «0 42 4< COLLECTION STAT 45 48 ION 21 30 32 38 40 42 44 45 48 58 COLLECTION STATION I 1 ALL OTHER PRPt mTm OITHONA SP EZ] UNIDENT COPEPODIDS <1 5mm ^m ALL CALANUS SPP ITTn ACARTIA LONCIREMIS ESa PSEUDOCAUNUS COPEPODIDS rm PSEUDOCAusNus sp imm COPEPOD NAUPLII ^S INVERTEBRATE ECCS Figure 3 The diet of larval pollock in the Gulf of Alaska in June 1987, by collection station in terms of %A', %VOL. and %1R1. Table 4 Index of proportional similarity (PS ) for walleye pollock 15.0-19.9 mm SL, at all stations, calculated using mean volume data. PS values for adjacent stations are underlined. Sta. no. 21 30 32 38 40 42 44 45 48 58 21 30 0.384 32 0.605 0.504 38 0.652 0.475 0.507 40 0.303 0.577 0.433 0.519 42 0.695 0.458 0.591 0.682 0.328 44 0.770 0.474 0.524 0.742 0.353 0.833 45 0.6.58 0.287 0.399 0.628 0.252 0.689 0^5^ * 48 0.676 0.415 0.490 0.676 0.368 0.760 0.805 0.728 58 0.782 0.547 0.626 0.810 0.441 0.7.54 0.831 0.645 0.706 Table 5 Inde.x of |)ro portional similar ty (PS) for walleye | polloc k 15.0- 19.9 mm SL, ak ng tran sect parallel to the Alaska Peninsula, cal :ulated using mean volume data. PS values for adjacent stations are | underlined. Sta. no. 21 30 32 42 44 58 21 3(1 0.384 * 32 0.605 0.504 * 42 0.695 0.4.58 0..591 • 44 0.770 0.474 0.524 0.833 • 58 0.782 0..547 0.626 0.754 0.831 • Grover Feeding ecology of Theragra chalcogramma in the Gulf of Alaska 469 Table 6 Index of proportional similarity (PS) for walleye pollock. 15.0-19.9 mm SL. along two transects perpendicular to the Alaska Peninsula, calcu- lated using mean volume data. PS values for adjacent stations are underlined. Discussion The portion of the present study dealing with the rela- tionship between diet and fish size complements the study by Kendall et al. (1987) conducted in the Gulf of Alaska during May 1983. As the samples in 1983 were collected earlier in the year using different gear (a 505-^m mesh bongo net) than in 1987, the size range of larvae was substantially different in the two studies. However, a comparison of the diet of larvae 10.0-19.0 mm can be made between the two years. From this it was clear that copepod eggs and nauplii were of some consequence in the diet in both years. A between-year comparison of the relative importance of Pseudocala- nus sp. adults and copepodids cannot be made because the 1983 dietary analysis considered adult and imma- ture stages together as copepodids (CI-CVI). In both years Pseudocalanus sp. was the dominant copepod, followed by Acartia sp. and Oithona sp. As Pseudo- calanus sp. carries its eggs (Corkett and McLaren 1978), the majority of copepod eggs in the diet were probably incidentally ingested attached to adult females. However, berried females were not observed in pollock stomachs. While the delicate structure of Pseudocalanus egg sacs may preclude larval fish from ingesting females with eggs intact (A.J. Paul, Univ Alaska, Seward, pers. commun., Nov. 1988), it is also possible that some of the eggs that were ingested could have been from other genera such as Acartia or Cala- nus that broadcast their eggs into the sea (C.B. Miller, Oregon State Univ., Corvallis, pers. commun., June 1989). The source of copepod eggs cannot be resolved without observing female copepods with eggs in the diet or identifying eggs to species. Neither condition was met in the present study. The dietary importance of copepod nauplii decreased and Pseudocalanus minuius increased with increasing larval size for larvae 9-20 mm SL in Uchiura Bay, Hok- kaido (Kamba 1977), as was the case in the present study. However, while /I. longiremis, Oithona sp., and various copepodids were important in the Gulf of Alas- ka, they were not in Uchiura Bay. Since copepod eggs were excluded from the dietary analyses by Kamba, it is not possible to compare their relative importance in Uchiura Bay with the Gulf of Alaska. In the Bering Sea, Clarke (1978) reported that copepod nauplii and Oithona similis were the primary prey for pollock 9-18 mm SL, and Cooney et al. (1980) reported that cope- podids (CI-CVI) and copepod nauplii were dominant. Oithona similis was the dominant copepod species, followed by Pseudocalanus sp. and A. longiremis. In both Bering Sea accounts, copepod eggs accounted for approximately 5% of the diet of this size class, less than in the Gulf of Alaska. While the diet of 20.0-29.9 mm SL fish in the Gulf of Alaska appeared to be more diverse than in Uchiura Bay (Kamba 1977), in both areas Pseudocalanus sp. was the dominant prey, and Calanus sp. was well repre- sented in the diet. In the Bering Sea Pseudocalanus sp. copepodids were the dominant prey and copepod eggs were also very important in the diet of early juveniles, although a broader size class, 20.1-60.0 mm SL, was utilized (Cooney et al. 1980). In a more recent study, also in the Bering Sea, cyclopoid copepods proved to be the primary prey for fish 20.0-29.9 mm TL, with Pseudocalanus sp. adults second in impor- tance (Lee 1985). In the Gulf of Alaska, Oithona sp. (the dominant cyclopoid) was of highly localized importance. The manner in which diet of larval and juvenile pollock varied between different locales and between years in the same locale illustrates how diet may vary in response to available food resources, which vary in response to oceanographic conditions (e.g., Grover and 011a 1987). As plankton data for the present study were unavailable, the extent to which diet resulted from prey selectivity cannot be determined. However, assuming that diet was at least reflective of available food re- sources, the wide range in PS values would indicate that food resources were not homogeneous. Diet varied widely from station to station, with nearly as many non- adjacent as adjacent station pairs showing significant dietary overlap. This type of variation would suggest areas of planktonic patchiness, perhaps reflecting underlying hydrographic heterogeneities. Low PS values may indicate small prey patches, while high values that were seen across several stations may be indicative of large prey patches. Regardless of whether prey were energetically equi- valent at each station, the incidence of feeding was 470 Fishery Bulletin 88(3), 1990 100%. It is unclear whether differential growth rates may have resulted from the ingestion of different prey since growth rates were not measured in the present study. However, it remains a possibility since Yokla- vich and Bailey (1989) reported differential growth rates for late larval and early juveniles at three loca- tions in the Gulf of Alaska in 1987. These fish, which ranged from 12.8 to 40.8 mm SL, were collected in the same survey of late-stage larval pollock (Hinckley et al. 1989) as the present study, from 18 June to 15 July 1987. While specific collection stations differed be- tween the two studies, Yoklavich and Bailey found that growth rates increased from southwest to northeast along the Alaskan Peninsula. Acknowledgments I would like to thank Bori 011a for valuable discussions, encouragement, and reviewing early drafts of this manuscript. I would also like to thank Art Kendall for his continued interest in this project, C.B. Miller for his help in identifying copepodids (CI V-C V) of Calan iis sp. and Pseudocalanus sp., and Sarah Hinckley for col- lecting specimens. Thanks also to M.E. Clarke and two anonymous reviewers for their comments. This work was supported by the Alaska Fisheries Science Center, National Marine Fisheries Service, NOAA Contracts NA-88-ABH-00005 and NA-89-ABH- 00039. Citations Clarke, M.E. 1978 Some aspects of the feeding biology of larval walleye pollock, Theragra chalcogramma (Pallas), in the southeastern Bering Sea. M.S. thesis, Univ. Alaska. Fairbanks, 44 p. Cooney. R.T.. M.E. Clarke, and P. Walline 1980 Food dependencies for larval, post-larval, and juvenile walleye pollock, Theragra chalcogramma (Pallas), in the south- eastern Bering Sea. hi PROBES: Processes and resources of the Bering Sea shelf. Progress Report 1980, Vol. 2, p. 169- 189. Inst. Mar. Sci., Univ. Alaska, Fairbanks. Corkett, C.J., and I. A. McLaren 1978 The hiologj- oi Pseudocalanus. Adv. Mar. Biol. 15:1-231. Dunn, J.R., and A.C. Matarese 1987 A review of the early life history of northeast Pacific gadoid fishes. Fish. Res. 5:163-184. Dunn, J.R., A.W. Kendall Jr., and R.D. Bates 1984 Distribution and abundance patterns of eggs and larvae of walleye pollock (Theragra elm Irogram ma) in the western (iulf of Alaska. NWAFC Proc. Rep. 84-10, Northwest .Alaska Fish. Cent., Natl. Mar. Fish. Serv., NOAA, Seattle, WA 98115- 0070. Frost, B.W. 1989 A taxonomy of the marine calanoid copepod genus Pseudocalanux. Can. J. Zool. 67:.525-551. Grover. J.J., and B.L. 011a 1987 Effects of an El Nifio event on the food habits of larval sablefish, Anoplopoma fimbria, off Oregon and Washington. Fish. Bull., U.S. 85:71-79. Hinckley. S., K. Bailey, J. Schumacher, S. Picquelle, and P. Stabeno 1989 Preliminary results of a survey for late-stage larval wall- eye pollock and observations of larval drift in the western Gulf of Alaska, 1987. In Proc. Int. Symp. Biol. Manage. Walleye Pollock, Nov. 1988, Anchorage, Alaska, p. 297-306. Alaska Sea Grant Prog., Univ. Alaska, Fairbanks. Incze, L.S.. A.W. Kendall Jr., J.D. Schumacher, and R.K. Reed 1989 Interactions of a niesoscale patch of larval fish (Tlieragra chalcogramma) with the Alaska Coastal Current. Continen- tal Shelf Res. 9:269-284. Kamba, M. 1977 Feeding habits and vertical distribution of walleye pollock, Theragra chalcogramma (Pallas), in early life stage in Uchiura Bay. Hokkaido. Res. Inst. N. Pac. Fish., Hokkaido Univ., Spec. Vol.. p. 175-197. Kendall, A.W. Jr., M.E. Clarke, M.M. Yoklavich, and G.W. Boehlert 1987 Distribution, feeding, and growth of larval walleye pollock, Theragra chalcogramma. from Shelikof Strait, Gulf of Alaska. Fish. Bull., U.S. 85:499-521. Kim. S. 1989 Early life history of walleye pollock. Theragra chalco- gramma. in the Gulf of Alaska. In Proc. Int. Symp. Biol. Manage. Walleye Pollock, Nov. 1988, Anchorage, Alaska, p. 117-139. Alaska Sea Grant Prog., Univ. Alaska, Fairbanks. Kohn. A.J., and A.C. Riggs 1982 Sample size dependence in measures of proportional similarity. Mar. Ecol. Prog. Ser. 9:147-151. Lee, S.S. 1985 A comparison of the food habits of juvenile Pacific cod and walleye pollock in the southeast Bering Sea. M.S. thesis, Univ. Alaska, Fairbanks, l.'-iO p. Linton, L.R., R.W. Davies, and F.J. Wrona 1981 Resource utilization indices: An assessment. J. Anim. Ecol. 50:283-292. Mathur. D. 1977 Food habits and competitive relationships of the band- fin shiner in Halawakee Creek, Alabama. Am. Midi. Nat. 97: 89-100. Methot, R.D. 1986 Frame trawl for sampling pelagic juvenile fish. Calif. Coop. Oceanic Fish. Invest. Rep. 27:267-278. Pinkas, L.. M.S. Oliphant. and I.L.K. Iverson 1971 Food habits of albacore, bluefin tuna, and bonito in Califoniia waters. Calif. Dep. Fish Game Fish Bull. 152, 105 p. Wallace. R.K. Jr. 1981 Anassessment of diet-overlap indexes. Trans. Am. Fish. Soc. 110:72-76. Wallace, R.K. Jr.. and J.S. Ramsey 1983 Reliability in measuring diet overlap. Can. ,1. Fi.sh. Aquat. Sci. 40:347-.351. Yoklavich. M.M., and K. Bailey 1989 Growth of larval and juvenile walleye pollock from Shelikof Strait, Gulf of Alaska, as determined from daily in- crements in otoliths. In Proc. Int. Symp. Biol. Manage. Wall- eye Pollock, Nov. 1988, Anchorage, Alaska, p. 241-251. Alaska Sea Grant Prog., Univ. Alaska, Fairbanks. Zar. J.H. 1974 Biostatistical analyses. Prentice-Hall, Englewood Cliffs, NJ, 620 p. Abstract.— Egg size has been shown to relate to survival and growth in the early life stages of fish. Com- bined field and laboratory studies ex- amined variation in the egg size of walleye pollock Tlwragi-a chalcogram- ma. a commercially important North Pacific gadoid; and a preliminary as- sessment of the effect of egg size on larval size was made. Differences in egg size were found over parts of the geographical range, among years, and over the spawning season in Shelikof Strait. Egg size did not ap- pear to correlate with female length, age, or condition. Individual female's egg size decreased significantly over the course of the spawning cycle. Preliminary studies in the laboratot^ showed that egg size is correlated with larval size and may therefore affect mortality in the early larval stages. Variation of Egg Size of Waileye Pollocl< Theragra chalcogramma \N\th a Preliminary Examination of the Effect of Egg Size on Larval Size Sarah Hinckley Alaska Fisheries Science Center, National Marine Fisheries Service, NOAA 7600 Sand Point Way NE, Seattle, Washington 981 15 Variability in egg size may be an im- portant factor in the survival of early life stages of fish. It has been proposed by numerous authors (Blaxter and Hempel 1963, Bagenal 1969, Ware 1975, Hunter 1981, Knutsen and Til- seth 1985) that egg size (dry weight or diameter) influences larval survi- val through its effects on larval size, growth rates, and activity. Larger eggs provide more energy for growth and development (Hempel and Blax- ter 1967, Hempel 1979) and general- ly produce larger larvae which are able to avoid predators more effec- tively (Miller et al. 1988), survive longer without feeding (Hunter 1981), search a greater volume of water for prey (Blaxter 1986, Webb and Weihs 1986), and eat prey of a greater vari- ety of sizes (Hunter 1981). Differences in egg size are thought to reflect adaptations by the spawn- ing stock to varying conditions met by early larvae. Bagenal (1971) pro- posed that there is a relationship among egg size, time of spawning, and availability of food. Gushing (1967) stated that egg size in Atlantic her- ring Clupea harengus harengus is re- lated to the type of production cycle and its variability. Hunter (1976) indicated that egg size may be important in studies of starvation and predation as cause of larval fish mortality, because of its ef- fects on time to the onset of irrever- sible starvation, on feeding success. Manuscript accepted 26 April 1990. Fishery Bulletin, U.S. 88:471-483. •Contribution FOCl-0078 to Fisheries-Ocean- ography Coordinated Investigations, NOAA. and on the ability of larvae to avoid predators. Larval size and its relation to egg size has been proposed as a unifying factor integrating the dynam- ics of larval survival and the mechan- isms of recruitment. Miller et al. (1988), in their review of this subject, state that "... even small size differ- ences of larvae at hatching can have significant ecological implications, and that even within a species (e.g., herring), size at hatching might be an adaptive response to local geographic conditions." Walleye pollock Theragra chalco- gramma is a semidemersal species ranging from Japan around the North Pacific rim to central California. This species supports important commer- cial fisheries in the Bering Sea and the Gulf of Alaska. Walleye pollock spawn pelagic eggs which have an in- cubation time of approximately 15 days in 5°C water. Egg diameters have been reported to range from 1.0 to 1.9 mm (Table 1). Larvae are 3-4 mm standard length (SL) at hatching (Dunn and Matarese 1987). Egg size in walleye pollock varies across its geographic range (Table 1). It also varies within populations (Gorbunova 1954, Yusa 1954, Nishiyama and Ha- ryu 1981, Sakurai 1982), and over the course of the spawning season (Sero- baba 1968). The causes and signifi- cance of egg size variation in walleye pollock are not well known. This field and laboratory study as- sesses sources of variation in walleye pollock egg size and presents results of a preliminary study of the relation- 471 472 Fishery Bulletin 88(3). 1990 Table 1 Geographic variation in walleye pollock egg size. Average Egg size (mm) Temperature* latitude (°C) Location (°N) Mean Range (Mean) Reference Eastern Pacific Puget Sound 47 1.2 1.0-1.3 8.5 This study Shelikof Strait 57 1.37 1.3-1.4 5.0 This study Bering Sea 57 1.73 1.3-1.9 3.25 Nishiyama and Haryu 1981 Bering Sea 59 1.57 1.5-1.7 0.95 Serobaba 1968 Bering Sea 59 1.59 1.4-1.8 2.75 Serobaba 1974 Western Pacific Japan 38 1.45 1.2-1.7 8.0 Yusa 1954 Peter the Great Bay 43 1.46 1.3-1.7 2.0 Gorbunova 1954 S. Kurile Islands 45 1..53 1.3-1.8 1.43 Gorbunova 1954 N. Kurile Islands 48 1.61 1.4-1.8 1.17 Gorbunova 1954 Saklialin Island 50 1.58 1.5-1.7 -0.08 Gorbunova 1954 W. Kamchatka 56 or depth of highest 1.57 1.3-1.8 egg abundance (averaged using 1.5 minimum and maximum Gorbunova 1954 of reported range). 'Temperature at bottom ship between egg size and larval size. Size variation in walleye pollock eggs due to annual, seasonal, and geo- graphical factors was examined from eggs caught in ichthyoplankton surveys in the Gulf of Alaska from 1981 to 1986. Variation due to female age, length, and condition was studied using eggs taken from spawn- ing females caught at sea and from females held in the laboratory. Effects of egg size on larval size at hatching and at yolksac absorption were investigated in rear- ing experiments, using eggs from captive females and eggs taken in ichthyoplankton surveys. Methods Field studies Walleye pollock eggs were collected in the Gulf of Alaska between 1981 and 1986 with 60-cm bongo nets (0.505-mm mesh), deployed with standard MARMAP procedures (Posgay and Marak 1980). Eggs selected for measurement were chosen from archived samples (Table 2) taken in different years (1981-86), months (March-May), and areas (Shelikof Strait, the Semidi Islands to the Shumagin Islands, and west of the Shumagin Islands; Fig. 1). Variation in mean egg size among individual sampling stations was examined by measuring 25-50 eggs from up to 10 sta- tions per year/month/area combination. Eggs at similar stages of development (early blastula to early germ ring) were measured with an ocular micrometer on a dissecting microscope at 40 x magnification. All eggs were preserved in 3-5% buffered formalin. Because mean egg sizes were significantly different {F = 4.75, p<0.0005) among sampling stations, nested analysis of variance models with unequal cell sizes were used to test year, month, and area effects, with sta- tions as the nested factor. Variation in egg size among years (within area and month), among months (within year and area), and among areas (holding year and month constant) had to be tested separately, due to the incompleteness of the available dataset, including miss- ing months and areas in some years (Table 2). The General Linear Models procedure of SAS-PC (SAS Inst. 1985) was used in this analysis. The F- statistic was derived using the Satterthwaite approx- imation (Gaylor and Hopper 1969 as cited in Sokal and Rohlf 1981), except in one case where the criteria for using this test were not satisfied; in this case a simple approximation test (Sokal and Rohlf 1981) was used. Effects of maternal age, size, and condition on egg diameter and dry weight were examined using eggs from females caught by midwater trawl in Shelikof Strait in late March, 1986 (Table 2). Hydrated (fully ripe) eggs were collected from spawning females, and preserved in 5% buffered formalin. For each female, fork length (FL), gonad- and liver-free body weight, and liver weight were determined and otoliths removed for ageing. Body weight and liver weight were used as indices of condition. Sixty eggs per female were measured. Because hydrated eggs taken from the ovary are sometimes slightly nonspherical, egg diameters were measured on the axis that fell along the transect to make sure mea- surements were random. Five to ten eggs per fish were subsampled for dry weight; eggs were measured. Hinckley Egg size variation of Theragra chalcogramma 473 Table 2 Stratification of walleye pollock egg samples measured for differences in egg size by year, area, and month, and to ex- amine the effect of female size, age, and condition on egg size. Samples were collected on ichthyoplankton cruises in the western Gulf of Alaska. Semidi Is. to West of Shelikof Strait Shumagin Is. Shumagin Is. Year March April May April April 1981 1982 1984 1985 1986 M Y,M M M Y,A,M M M Y,M M Y Samples used to study the effect of year (Y), area (A), month (M), and female (F) age, size, and condition. washed with distilled water, placed individually in pre- weighed aluminum boats, dried for 24 hours at 60°C, and weighed on a Cahn 29 electrobalance to the nearest 0.001 mg. An analysis of variance procedure was used to exam- ine egg size within and among fish. A multiple regres- sion model was used to examine the effect of female length, age, and condition on egg diameter and egg weight. The factors in the regression were fork length, age, body weight (minus ovary weight and liver weight), and liver weight. The model used was Y = Bo + 5iA'i + B0X2 + 5,3X3 + B4.Y4 where Y = egg diameter (mm) or egg dry weight (mg), Xi = fork length (cm), Xo = age (yr), Xg = body weight (minus ovary and liver weight, g), X4 = liver weight (g). Interaction terms were not included in the model until after a test of the overall significance. Reference to trade names does not imply endorsement by the Na- tional Marine Fisheries Service, NOAA. -60 DON 166 OOW 164 00 IS"! 00 Figure 1 Western Gulf of Alaska, including Shelikof Strait, a major spawning area for walleye pollock. Shaded areas indicate regions from which samples were taken. 474 Fishery Bulletin 88 13). 1990 Laboratory experiments Laboratory investigations were conducted using cap- tive walleye pollock in March and April 1987. Adult prespawning walleye pollock were caught in January and February by handline in the Tacoma Narrows region of Puget Sound. Fish were transported to the laboratory and held in net pens under ambient light and temperature conditions and fed chopped herring ad libidum until they were nearly in spawning condition (i.e., eggs and milt were expelled with slight pressure to the abdomen). These fish were then transferred to individual 1700- or 2500-L tanks. One female and one or two males were placed in each tank and held throughout their spawning cycle. Water temperatures in the tanks were ambient and varied over the spawn- ing season from 9.5 to 11.5°C (x 10.08°C). Ambient water temperature increased significantly (i?^ = 20.9%) over the experimental period. Salinity was constant at 28 ppt. A 20-cm plankton net with a hard codend was suspended in each tank to catch a sample of eggs from each spawning event. Plankton nets were checked each morning for the presence of eggs. If eggs were pres- ent in the codend, they were saved, and the water in the tanks replaced. In 1988, pollock caught in Puget Sound were held in- dividually in tanks to further examine the effect of fe- male length on egg size. Eggs from the first batch from each of seven females were collected and measured in the same manner as described for the 1987 studies. Egg diameters were measured on 15-25 unpreserved eggs, and egg dry weights were determined for 5-10 eggs from each spawning event. The remainder of the eggs were incubated separately O^y female and spawn- ing event) in 4-L jars filled with filtered seawater. The jars were placed in a water bath with running seawater at ambient water temperatures. Light levels were am- bient. Water in these jars was changed several times during incubation (which lasted 6-9 days) and dead eggs were removed daily. Six to fifteen larvae were collected on the day of hatch, anaesthetized with MS222, and their standard lengths measured. The rest of the larvae were held in the 4-L jars until yolksac absorption (defined as the point when individual larvae had used up 90-100% of their yolksacs). Standard lengths were measured (on anaesthetized larvae) and dry weights determined for 5-10 larvae at this stage. A sample of live eggs collected from the ichthyo- plankton in Shelikof Strait was transported to Seattle in April of 1987. Diameters and dry weights were determined on a subsample of these eggs and the rest were incubated. Larvae were measured at hatching and at yolksac absorption in the same manner as described above. Table 3 Differences in mean walleye pollock egg diameter among years in April in the Shelikof Strait region of the Gulf of Alaska. Egg diameter Year* Mean SE SNK N grouping* 1981 1986 1985 1984 1982 1.360 1.317 1.299 1.298 1.296 0.003 320 0.004 250 0.004 250 0.003 250 0.004 249 A B C c c 'Data not available from 1983. **Student-Newman-Keuls multiple range test(SAS Institute 1985). A, B, C indicate significantly different groups. Eggs collected from the Gulf of Alaska ichthyoplank- ton surveys for examination of yearly, seasonal, and regional differences in egg size were preserved in 3-5% formalin. No correction factor was used to correct for changes in egg size due to preservation, as all eggs used in these analyses were preserved in the same manner. The field samples taken from spawning adults and used to examine the relationship between egg size and fish size, age, and condition were also preserved in 3-5% formalin. These, however, were collected before spawn- ing, and the diameters may not be directly comparable to sizes of spawned eggs collected from the ichthyo- plankton, due to possible changes in egg size at fertiliza- tion and activation (Fleming and Ng 1987, Kj0rsvik and Lonning 1983). Eggs from the laboratory were measured fresh and are therefore not directly com- parable with eggs from the field studies. The com- parison between sizes of Gulf of Alaska eggs and Puget Sound eggs was made only on the fresh samples col- lected in 1987 from both areas. Results Geographical, seasonal, and regional differences in egg size Examination of 1319 walleye pollock eggs from April plankton samples (the month when most spawning occurs; Kim 1987) collected in Shelikof Strait demon- strated a significant difference in mean egg diameters among years (F = 19.862, p<0.0005). Mean egg diam- eter was largest (Table 3) in 1981 and smallest in 1982; there was a 4.7% difference in size between these years. Mean egg size was not statistically different in 1982, 1984, and 1985 (SNK test. Table 3). Mean egg size was intermediate in 1986. Hinckley: Egg size variation of Theragra chalcogramma 475 Table 4 Differences in mean walleye pollock egg diameter over the 1 spawning season in Sheiikof Strait for 1981, 1982, and 1984. Egg diameter SNK Year Month Mean SE N grouping 1981 March 1.405 0.007 103 A April 1.360 0.003 320 B May 1.297 0.00.5 204 C 1982 March « April 1.296 0.004 249 A May 1.314 0.006 159 B 1984 March 1.383 0.002 792 A April 1.299 0.004 2.50 B May from th ?se months. ♦Data not available A significant difference in mean egg diameters was also observed over the course of tiie spawning season (March-May) in Sheiikof Strait (Table 4) in 1981 (F = 23.389, p<6.0005), 1982 (F= 16.3387, 0.01<0.0005*). In 1981 and 1984 egg sizes decreased over the course of the spawning season by about 7% per month (Table 4, 0.02 0.25, respectively). A stepwise regression eliminated all variables as nonsignificant (;.)>0.05). No interactions or nonlinear relationships among variables were found. Fourteen females (29.5- 47.0 cm PL) were held in the laboratory to investigate the female length-egg size relationship. As in previous regression analyses, no relationship was seen between female length and egg diameter for these fish (Fig. 2, r = 0.006). Relationship between egg size and female characteristics The mean diameter of hydrated eggs taken from Sheii- kof Strait females in 1986 was 1.41 mm (range 1.29- 1.57). The mean dry weight of these eggs was 0.119 mg (range 0.095-0.148). Variation in egg diameter and egg dry weight among fish was more significant (/j *Satterthwaite approximation not used in this test as the criteria were not met. A simple approximation test (Sokal and Rohlf 1981) was used in this case. Spawning characteristics and the relationship of egg size to larval size The number of egg batches spawned per female in the laboratory in 1987 ranged from 2 to 21 over 3-26 days (Table 7). Excluding the two females which died pre- maturely, the average number of batches was 14.4 (range 9-21), and the average duration of spawning was 21.4 days (range 18-26). The average interval be- tween batches for all fish was 2.1 days (range 1-5). The relationship between egg diameter and egg dry weight was linear (7.><0.0005, R'~ = 0.630). 476 Fishery Bulletin 88(3). 1990 1.35 r- . 1 1.30 0) 0) I E 3 1.25 ■o - i ' 5 m c S 1.20 ■ 1 I J } 1 , , , 1 , , , 1 I 1 25 30 35 40 45 Female (ork length (cm) 50 Figure 2 Relationship between fetiiale fork length an90% spent* 2 2 5 5.0 <10% spent** 3 13 18 1.4 1-2 spent* 4 15 23 1.5 1-3 spent* 5 2 3 3.0 <10% spent** 6 14 20 1.4 1-2 spent* 7 21 26 1.2 1-2 spent* Mean 14. 4**' 21.4*** of cessation of spawning. 2.1 'Sacrificed within 1 week ••Died after 2 batches. ** 'Spent fish only. 1.30 1.25 ? ^ 1.20 E .2 1.15 1,10 1.05 1.00 Y-1, 247-0. 008X R^ - 0 72 J ,_ 10 15 Spawning event (batch) 20 25 Figure 3 Change in mean egg diameter over the batch spawm- ing cycle of female walleye pollock. Solid line indicates fitted regression line. Each symbol represents a separate female held in the laboratory over the spawn- ing cycle. Hinckley Egg size variation of Theragra chalcogramma 477 5.0 _ 4 8 _ • Alaska eggs • Pugat Sound eggs * E 4.6 - E Y . 0463 ♦ 2.613X ^ 4,4 ~ r2. 0.385 "J 4.2 " 4.0 - _^ JZ • • ^ g 3.8 • - • -.r^^^ T5 3-6 - • • > _---* ' ** -•^-^^^'^ 3.2 "■^ • 1.00 1.05 1.10 1.15 1.20 1.25 1.30 Egg diameter (mm) 1.35 1.40 Figure 4A Relationship between mean diameter of a batch of walleye pollock eggs spawned in the laboratory, and the mean stan- dard length of larvae hatched from that batch. 5.0 _ 4.8 _ ♦ Alaska eggs E 46 - • Puget Sound eggs * E ~ 44 w ™ 4.2 - Y - 2.503 ♦ 14.167X b2. 0.351 ^^ ^ " 4 0 f 3.8 H 3.6 I 3,4 3.2 •, • • „-- • 1 1 1 1 1 1 1 1 0.05 0.06 0.07 0.08 Egg d 0.09 0.10 y weight (mg) 0.11 0.12 0.13 Figure 4B Relationship between mean dry weight of a batch of walleye pollock eggs spawned in the laboratory, and the mean standard length of larvae hatched from that batch. Mean egg diameter per batch of eggs spawned de- clined over the spawning cycle of an individual for all females (Fig. 3). The average decline in egg diameter over the batch spawning cycle (for females that com- pleted spawning) was 0.14 mm, or 11,5% of the initial egg diameter. Thirty-three batches of eggs were incubated and hatched. Larval standard length at hatch was positively correlated with egg diameter (Fig. 4A) and egg dry weight (Fig, 4B). The regressions of larval length at hatch on egg diameter and dry weight were both sig- nificant (0.001 5 39 - • 1 1 1 • 1 1 0.05 0.06 0.07 0.08 Egg d 0.09 0.10 0. y weight (mg) 11 0.12 0.13 Figure 5B Relationship between mean dry weight of a liateh of walleye pollock eggs spawned in the laboratory, and the mean standard length of larvae at yolksac absorption. val between batches of eggs spawned in Atlantic cod Gadus m,orhua (Kjesbu 1988). Egg size variation within and among females The laboratory studies showed that egg diameter in walleye pollock declined significantly over the spawn- ing period of an individual. This was also shown by Sakurai (1982) for walleye pollock (over a limited por- tion of the spawning cycle) and for other gadoids by Hislop et al. 1978 and Moksness and Vestergard 1982 (Melanogrammus aeglefinus), by Grauman 1965 and Solemdal 1970 (Gadus morhua), and Hislop 1975 (Mer- langius meriangus). The amount of decrease seen in walleye pollock, about 12%, was comparable with that seen in Gadus morhua by Kjesbu (1988) (7-15%) and with that in Merlangius meriangus by Hislop (1975) (9-14%). This decline may represent an adaptation by the parent stock to conditions faced by the larvae, as discussed later in this section. Egg size in walleye pollock does not appear to be cor- related with female length. The measurement of eggs from the first batch spawned (a method also used by Kjesbu (1988) for Gadus morhua) removes variation in egg size caused by the decline in egg size over the course of the spawning cycle, a problem which may have confused the results of other studies of the rela- tionship between egg size and female size. Egg size in walleye pollock did not appear to correlate with female age or condition. Wet weights of body, gonad, and liver are not particularly sensitive indicators of condition in fish, however, and use of these measures may have obscured a correlation between these indices and egg size. There are many contradictory studies on the egg size- female size relationship. Sakurai (1982, Theragra chal- cogramyna). Marsh (1984, Etheostomxi specfahilis), Zilstra (1973, Clupea harengus). Solemdal (1970, Gadus morhua) and Ossthuizen and Daan (1974, G. morhua), and others did not find egg size-female size or age Hinckley Egg size variation of Theragra chalcogramma 479 E 0.17 —■ c o • Alaska eggs • a 0.15 o - • Puget Sound eggs .o y - -0249 ♦ 0 284X ^^ I 0.13 - r2. 0.462 • 5 O 0.1 1 _ ^^■^ %^^^ S 0.09 01 0) t >■ 0.07 "3 IfT^ *' i 0.05 -1 1.05 1.10 1.15 1.20 1.25 Egg diameter (mm) 1.30 1.35 1.40 Figure 6A Relationship between mean diameter of a batch of walleye pollock eggs spawned in the laboratory, and the mean dry weight of larvae at yolksac absorption. 'to E 0.1 1 c 0 ♦ Alaska eggs Q. 0.10 - • Puget Sound eggs * o n Y • -0.055 * 1.550X ^ o 0.09 - r2. 0.852 ^^ CO o 0.08 - • • ffl ^^ ^ 0.07 >. 0.06 -/^ T3 • ^^ "5 -i 0.0 1 ^^ 1 1 1 1 1 1 1 60 0.065 0.070 0.075 0.080 0.085 0.090 0.095 0.100 Egg dry weight (mg) Figure 6B Relationship between mean dry weight of a batch of walleye pollock eggs spawned in the laboratory, and the mean dry weight of larvae at yolksac absorption. relationships in their studies. On the other hand, there are numerous studies in which egg size-female size rela- tionships have been seen, such as Blaxter and Hempel (1963, Clupea harengus), Hislop et al. (1978, Melano- grammus aeglefinus), Grauman (1964, Gadus morhua), Kjesbu (1988, G. morhua), and many on salmonids. Some studies on the same species contradict one another with regards to this relationship. For those species showing a seasonal decline in egg size, an egg size-female size relationship may have been obscured in studies where this was not taken into account. Within- and among-population variation in egg size The population of walleye pollock spawning in Shelikof Strait shows significant seasonal and annual differ- ences in egg size. The seasonal decline in egg size prob- ably covaries with the decline over the spawning cycle. Annual differences in egg size during the peak spawn- ing month (April) in Shelikof Strait appear to be sig- nificant. These yearly differences are apparently not attributable to changes in the size or age composition of the spawning stock, as no relationship between these factors and egg size is apparent. It is possible that differences in the date of peak spawning in different years could change the mean egg size in the first week of April (when eggs measured for this study were collected). There is evidence that the mean date of peak spawning was later in 1986 and 1987 than in some earlier years (Yoklavich and Bailey 1990). The apparent difference in mean April egg size seen in this study could therefore be a reflection of the variability in time between peak spawning and egg col- lection and the decline in egg size over the spawning season, rather than a true interannual difference. 480 Fishery Bulletin 88(3)^ 1990 There appears to be a positive correlation between egg size and latitude in walleye pollock (Table 1). This correlation was also noted by Gorbunova (1954). In this study, this correlation was seen both among widely separated areas such as Puget Sound and the Gulf of Alaska, and within the Gulf of Alaska itself. Latitudinal clines in egg size have been seen by Demir (1963), Chiechomski (1973), and others, and have been dis- cussed by Rass (1941), Marshall (1953), and Thresher (1988). As noted by Thresher, "Latitudinal variation in life histories of marine organisms typically are at- tributed to selection acting on local populations along a latitudinal environmental gradient. . ." Several studies have noted an inverse relationship between egg size and temperature (Southward and Demir 1974, Ware 1977, Marsh 1984, Houghton et al. 1985, Imai and Tanaka 1987). Imai and Tanaka (1987), working with Japanese anchovy Engraulis japonicn, found yearly changes in egg size which correlated with temperature, and were able, experimentally, to alter egg size during the spawning cycle of anchovy held in the laboratory by changing water temperature. Ware (1977) noted that peak spawning of Scotnher scomhrus occurs later if water temperatures warm more slowly than usual in the spring. Temperature may be the latitudinal environmental gradient along which selection for egg size acts on local populations. The relationship between latitude and egg size appears to be correlated to temperature in the dif- ferent spawning regions over the range of walleye pollock (Table 1). In the laboratory experiments on spawning done for this study, ambient water tempera- ture increased over the spawning period, while egg size declined significantly for each female over this period. Mean monthly egg size and mean monthly water temperatui'e at the depth where eggs are spawned (150 m to the bottom) within Shelikof Strait, however, show no apparent inverse correlation. There was also no in- verse correlation between mean egg size in April and mean April water temperature at the depth of spawn- ing over the years examined (J.D. Schumacher, Pacific Mar. Environ. Lab., Seattle, WA 98115-0070, unpubl. data). It may be that temperature acts on the parent stock at some earlier date, for example while the stock is migrating to the spawning grounds. Data are insuf- ficient to investigate this question. Consequences of egg size variation for larval survival A correlation between egg size and larval size has also been noted in other species {Clupea harengus, Blaxter and Hempel 1963; Salvelinus alpinus, Wallace and Aasjord 1984; Etheostoma spectabile, Marsh 1986; Gadus morhua, Solemdal 1970, Knutsen and Tilseth 1985). Only a few studies have reported no correlation (Zonova 1973, Reagan and Conley 1977, Lagomarsino et al. 1988). Egg size has been correlated with other larval fac- tors. Growth rate, for example, has been seen to in- crease with increased egg size (Blaxter and Hempel 1963, Bagenal 1969, Wallace and Aasjord 1984, Moodie et al. 1989). This may be related to increased feeding success due to larger mouth sizes or increased search area (from increased reactive distance or swimming speed). Mouth size has been positively correlated with egg size (Blaxter and Hempel 1963, Shirota 1970, Knutsen and Tilseth 1985). Larvae with larger mouths are able to take larger or more varied sizes of prey items. Length of time from hatch to starvation (due to increased endogenous yolk reserves; Blaxter and Hempel 1963, Bagenal 1969, Theilacker 1981, Marsh 1986, Wallace and Aasjord 1984), and survival (Blax- ter and Hempel 1963, Bagenal 1969, Pitman 1979, Small 1979, Moodie et al. 1989) have also been positive- ly correlated with egg size. If larval survival is affected by larval size at hatch (Miller et al. 1988) and other fac- tors, then variation in egg size in walleye pollock may cause differences in larval mortality rates. It has been proposed that changes in egg size may be an adaptation to the timing of the production cycle, to ensure that larvae are produced that are able to take advantage of the available food supply (Gushing 1967; Hempel and Blaxter 1967; Bagenal 1971; Jones and Hall 1974; Ware 1975, 1977). The reproduction of zoo- plankton (such as copepods, e.g., Pseudocalanus and Oithona spp., whose naupliar stages constitute the most common food of first-feeding walleye pollock lar- vae in the Bering Sea and the Gulf of Alaska; Clarke 1978. 1984; Nishiyamaand Hirano 1983; Kendall etal. 1987) is temperature and food related (Checkley 1980a,b; Durbin et al. 1983; Runge 1985; Corkett and McLaren 1978; Landry 1976). This implies that tem- perature would provide information on the timing of zooplankton reproduction, and therefore on the size of the available larval food supply (Ware 1977). Informa- tion on the changes in species and size composition of larval food sources in the western Gulf of Alaska is not presently available. The observed trends in egg size would be explained in terms of larval food supplies if the abundance of prey items increased and the size of food particles decreased and became more uniform as the season progressed. The observed trends in egg size may also be a result of an adaptation to relieve predation pressure on eggs and larvae. Seasonal changes in egg size could result in changes in the predator-prey size ratio for eggs and larvae, affecting predation rates (Bailey and Houde 1 989). This may not be important if egg predators are mainly planktivorous fishes. With zooplankton pred- Hinckley: Egg size variation of Theragrs chalcogramma 481 ators, however, small changes (10-20%) in egg diam- eter may affect their ability to handle or consume eggs. Also, if the abundance of egg predators increases, or they co-occur more frequently with eggs, then smaller but more numerous eggs (assuming a tradeoff between egg size and fecundity; Svardson 1949, Bagenal 1978) may be an adaptation to offset higher rates of preda- tion later in the season. Miller et al. (1988) have shown that predation rates are negatively correlated with larval size, possibly due to differences in predator mouth size, encounter rates of larvae and their predators, or the ability of larvae to escape predators. Trends in egg size and the result- ing differences in larval size at hatch may, therefore, also be an adaptation to relieve predation on larvae, if the size, species composition, and co-occurrence of predators and larvae changes seasonally. Acknowledgments I wish to thank Annette Brown for her assistance in the laboratory, Susan Picquelle for her advice on statistical analysis and on the manuscript, and Drs. Kevin Bailey and Arthur Kendall for their advice on the manuscript and for general support. Citations Bagenal, T.B. 1969 Relationship between egg size and fry survival in brown trout Salmo trutta L. J. Fish. Biol. 1:349-353. 1971 The interrelation of the size of fish eggs, the date of spawning and the production cycle. J. Fish. Biol. 3:207-219. 1978 Aspects of fish fecundity. In Gerking, S.D. (ed.), The ecology of freshwater fish production, p. 75-101. John Wiley, NY. Bailey, K.M., and E.D. Houde 1989 Predation on eggs and larvae of marine fishes and the recruitment problem. Adv. Mar. Biol. 25, 83 p. Blaxter, J.H.S. 1986 Development of sense organs and behavior of teleost lar- vae with special reference to feeding and predator avoidance. Trans. Am. Fish. Soc. 115:98-114. Blaxter, J.H.S., and G. Hempel 1963 The influence of egg size on herring larvae (Clupea harenffiis L.}. J. Cons. Perm. Int. Explor. Mer 28:211-240. Checkley. D.M. Jr. 1980a The egg production of a marine planktonic copepod in relation to its food supply: Laboratory studies. Limnol. Oceanogr. 25:430-446. 1980b Food limitation of egg production by a marine planktonic copepod in the sea off southern California. Limnol. Oceanogr. 25:991-998. Chiechomski, J.D. de 1973 The size of the eggs of the Argentine anchovy, Engraulis anchoita (Hubbs and Marini) in relation to the season of the year and to the area of spawning. J. Fish. Biol. 5:393-398. Clarke, M.E. 1978 Some aspects of the feeding ecology of larval walleye pollock, Theragra chalcogramma in the southeastern Bering Sea. M.S. thesis, Univ. Alaska, Fairbanks, 44 p. 1984 Feeding behavior of larval walleye pollock, Theragra chalcogramma (Pallas) and food availability to larval pollock in the southeastern Bering Sea. Ph.D. thesis, Univ. Calif., San Diego. 208 p. Corkett, C.J., and I. A. McLaren 1978 The biology of Pseudocatanits. Adv. Mar. Biol. 15, 248 p. Gushing. D.H. 1967 The grouping of herring populations. J. Mar. Biol. Assoc. U.K. 47:193-208. Demir, N. 1963 Synopsis of biological data on anchovy, Engraulis en- crasichoh4s (Linnaeus) 1758. (Mediterranean and adjacent seas). FAO Fish. Synop. 26, Rev.l, 48 p. Dunn, J.R., and A.C. Matarese 1987 A review of the early life history of northeast Pacific gadoid fishes. Fish. Res. 5(2-3): 163-184. Durbin, E.G., A.G. Durbin, T.J. Smayda, and P.G. Verity 1983 Food limitation of production by adult Acartia tonf^a in Narragansett Bay, Rhode Island. Limnol. Oceanogr. 28: 1199-1213. Fleming, I.A., and S. Ng 1987 Evaluation of techniques for fixing preserving, and mea- suring salmon eggs. Can. J. Fish. Aquat. Sci. 4:1957-1962. Gaylor, D.W., and F.N. Hopper 1969 Estimating the degrees of freedom for linear combina- tions of mean squares by Satterthwaite's formula. Tech- nometrics 11:691-705. Gorbunova, N.N. 1954 The reproduction and development of walleye pollock, Theragra chakogramma (Pallas). Akad. Nauk SSSR, Tr. Inst. Okean. 11:132-195. [Transl. by S. Pearson, Alaska Fish. Sci. Cent., Natl. Mar. Fish. Serv.. NCAA, Seattle, WA 98115-0070.] Grauman, G.B. 1964 The importance of the size of the eggs of the Baltic cod for survival of foetuses. ICES CM. (Gadoid Fish Committee) Pap. 85, 5 p. 1965 Changes in the egg size of cod {Gadus morhua calla rius L.) within the spawning period. ICES CM. (Baltic-Belt Seas Committee) Pap. 47, 4 p. Hempel, G. 1979 Early life history of marine fish. The egg stage. Univ. Wash. Press, Seattle, 70 p. Hempel, G., and J.H.S. Blaxter 1967 Egg weight in Atlantic herring (Ciupcateren5r!(,sL.). J. Cons. Cons. Int. Explor. Mer 31:170-195. Hinckley, S. 1987 The reproductive biology of walleye pollock, Theragra chalcogramma, in the Bering Sea, with reference to spawn- ing stock structure. Fish Bull., U.S. 85:481-498. Hislop. J.R.G. 1975 The breeding and growth of whiting, Merlangivs mer- langus in captivity. J. Cons. Cons. Int. Explor. Mer 36: 119-127. Hislop. J.R.G.. A.P. Robb. and J. A. Gauld 1978 Observations on effects of feeding level on growth and reproduction in haddock, Melanogrammus aeglefmus (L.) in captivity. J. Fish. Biol. 13:85-98. Houghton, R.G.. J.M. Last, and P.J. Bromley 1985 Fecundity and egg size of sole {Solea solea (L.)) spawn- ing in captivity. J. Cons. Cons. Int. Explor. Mer 42:162-165. 482 Fishery Bulletin 88 13). 1990 Hunter, J.R. 1976 (editor) Report of a colloquium on larval fish mortality studies and their relation to fishery research, January 1975. NOAA Tech. Rep. NMFS CIRC-395, Natl. Oceanic Atmos. Adm., Natl. Mar. Fish. Serv., 5 p. 1981 Feeding ecology and predation of marine fish larvae. In Lasker, R. (ed.), Marine fish larvae, p. 33-77. Wash. Sea Grant Prog., Univ. Wash. Press, Seattle. Imai, C, and S. Tanaka 1987 Effect of sea water temperature of egg size of Japanese anchovy. Nippon Suisan Gakkaishi .53:2169-2178 [in Engl.]. Jones, R., and W.B. Hall 1974 Some observations on the population dynamics of the lar- val stage in the common gadoids. In Blaxter, J.H.S. (ed.). The early life history of fish, p. 87-102. Springer- Verlag, Berlin. Kendall, A.W. Jr., M,E. Clarke, M.M. Yoklavich, and G.W. Boehlert 1987 Distribution, feeding, and growth of larval walleye pol- lock, Theragrn chalcogramma, from Shelikof Strait, Gulf of Alaska. Fish. Bull., U.S. 85:499-521. Kim, S. 1987 Spawning behavior and early life history of walleye pollock, Tkeragra chalcogramma, in Shelikof Strait, Gulf of Alaska, in relation to oceanographic features. Ph.D. Diss., Univ. Wash., Seattle, 221 p. Kjesbu. O.S. 1988 Aspects of the reproduction in cod {Gadns morhiKi L.): oogenesis, fecundity, spawning in captivity and stage of spawn- ing. Dr. Sci. thesis. Dep. Fish. Biol., Univ. Bergen, Norway, 147 p. Kj0rsvik E., and S. Lonning 1983 Effects of egg quality on normal fertilization and early development of the cod. Gadus morhua L. J. Fish. Biol. 23:1-12. Knutsen, G.M.. and S. Tilseth 1985 Growth, development, and feeding success of Atlantic cod larvae Gadus murhau related to egg size. Trans. Am. Fish. Soc. 114:.507-511. Lagomarsino, I.V., R.C. Francis, and G,W. Barlow 1988 The lack of correlation lietween size of egg and size of hatchling in the Midas cichlid, Cwhkisvma citruwltum. Copeia 1988:1086-1089. Landry, M.R, 1976 Population dynamics of the planktonic marine copepod Acartia clauM Giesbrecht. in a small temperate lagoon. Ph.D. Diss., Univ. Wash., Seattle, 200 p. Marsh. E. 1984 Egg size variation in central Texas populations of Etheo- stoma speetabile (Pisces: Percidae). Copeia 1984:291-301. 1986 Effects of egg size on offspring fitness and maternal fecundity in the orangethroat darter, Etheosloma spectahile (Pisces: Percidae). Copeia 1986:18-30. Marshall. N.B. 1953 Egg size in Arctic. Antarctic and deep-sea fishes. Evolu- tion 7:328-341. Miller, T.J., L.B. Crowder, J. A. Rice, and E.A. Marschall 1988 Larval size and recruitment mechanisms in fishes: Toward a conceptual framework. Can. J. Fish. Aquat. Sci. 45:1657-1670. Moodie, G.E.E., N.L. Loadman. M.D. Wiegand. and J. A. Mathias 1989 Influence of egg characteristics on survival, growth and feeding in larval walleye {Stizo.'^tediion rilreiini). Can. J. Fish. Aquat. Sci. 46:516-521. Moksness, E., and J.R. VestergSrd 1982 .Spawning of haddock {Mi'hinognniimu.'i atylffiiui.t) in cap- tivity. Flcjdevigen Rapp. 2, 9 p. Nishiyama, T., and T. Haryu 1981 Distribution of walleye pollock eggs in the uppermost layer of the southeastern Bering Sea. In Hood, D.W.. and J. A. Calder (eds.). The eastern Bering Sea shelf: Oceanography and resources, vol. 2, p. 993-1012. Univ. Wash. Press, Seattle. Nishiyama, T., and K. Hirano 1983 Estimation of zooplankton weight in the gut of larval walleye pollock (Theragra chidcognimina). Bull. Plankton Soc. Jpn. 30:1.59-170. Ossthuizen, E., and N. Daan 1974 Egg fecundity and maturity of North Sea cod, Gadus niorhiiii. Neth. J. Sea Res. 8(4):378-397. Pitman, R,W. 1979 Effects of female age and egg size on growth and mor- tality in rainbow trout. Prog. Fish-Cult. 41:202-204. Posgay. J. A., and R.R. Marak 1980 The MARMAP bongo zooplankton samplers. J. North- west Atl. Fish. Sci. 1:91-99. Rass, T.S. 1941 Analagous or parallel variations in structure and develop- ment of fishes in the northern and arctic seas. Jubilee Publ. Mosc. Soc. Natural. 1805-1940. 60 p. Reagan, R,E., and CM. Conley 1977 Effect of egg diameter on growth of channel catfish. Prog. Fish-Cult. 39:133-134. Rover, T.C. 1986 Temperature fluctuation in the Northeast Pacific from 1954 to 1985 in response to El Nino/Southern Oscillations and longer period forcing, p. 203-208. Int. North Pac. Fish. Comm. Bull. 47. Runge, J. A. 1985 Relationship of egg production of Calanus pacificus to seasonal changes in phytoplankton availability in Puget Sound, Washington. Limnol. Oceanogr. 30:382-396. Sakurai, Y. (formerly known as T.H. Yoon) 1982 Reproductive ecologj' of walleye pollock Theragra chnko- graintna (Pallas). Ph.D. Diss., Hokkaido Univ., Hakodate, Hokkaido, Japan. [Transl. by T. Nishida, N.D. Davis, and T. Nishiyama for Natl. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., Seattle, WA 98115-0070.] SAS Institute, Inc. 1985 SAS/STAT guide for personal computers. Version 6 Edi- tion. SAS Institute Inc., Gary, NC, 378 p. Serobaba. I.I. 1968 .Spawning of the Alaska pollock Tht-ragra chalcogram- ma (Pallas) in the northeastern Bering .Sea. .1. Ichthyol. 8:789-798. 1974 Spawning ecology of the walleye pollock {Tkeragra chal- cogramma) in the Bering Sea. J. Ichthyol. 14:544-552. Shirota, A. 1970 Studies on the mouth size of fish larvae. Bull. Jpn. Soc. Sci. Fish. 36:353-368. Small, T. 1979 Trout eggs — look for size and service. F'roc. 1 1th Two Lakes Fish Synip., p. 127-132. Jans.sen Services, London. Sokal, R.R., and F,J. Rohlf 1981 Biometry, 2(1 ed. W.H. Freeman, NY, 859 p. ■Soiemdal, P. 1970 Intraspecific variations in size, buoyancy and growth of eggs and early larvae of Arcto-Norwegian cod, Gadua morhau L., due to parental and environmental effects. ICES CM. 1970/F: 28. Demersal Fish (Northern) Committee, 12 p. Hinckley Egg size variation of Theragrs chalcogramma 483 Southward, A.J., and N. Demir 1974 Seasonal changes in dimensions and viability of the developing eggs of the Cornish pilchard (Sardinia pilchardiis Walbaum) off Plymouth. In Blaxter, J.H.S. (ed.), The early life history of fish, p. 53-68. Springer-Verlag, Berlin. Svardson, G. 1949 Natural selection and egg number in fish. Rep. Inst. Freshwater Res. Drottningholm 29:115-122. Theilacker, G.H. 1981 Effects of feeding history and egg size on the morphology of jack mackerel, Trachurus sym.met>Hcu3, larvae. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 178:432-440. Thresher, R.E. 1988 Latitudinal variation in egg sizes of tropical and sub- tropical North Atlantic shore fishes. Environ. Biol. Fishes 21:17-25. Wallace, J.C., and D. Aasjord 1984 An investigation of the consequences of egg size for the culture of Arctic charr, Salvelinus alpinu.-< (L.). J. Fish. Biol. 24:427-435. Ware, D.M. 1975 Relation between egg size, growth, and natural mortal- ity of larval fish. J. Fish. Res. Board Can. 32:2503-2512. 1977 Spawning time and egg size of Atlantic mackerel, Scomber scombrus, in relation to the plankton. J. Fish. Res. Board Can. 34:2308-2315. Webb, P.W., and D. Weihs 1986 Functional locomotor morphology of early life history stages of fishes. Trans. Am. Fish. Soc. 115:11.5-127. Yoklavich, M.M., and K.M. Bailey 1990 Hatching period, growth and survival of young walleye pollock Theragra chalcogramma as determined from otolith analysis. Mar. Ecol. Prog. Ser. 64:12-23. Yusa, T. 1954 On the normal development of the fish, Theragra chalco- gramma (Pallas), Alaska pollock. Bull. Hokkaido Reg. Fish. Res. Lah. 1(1:1-15 [in Jpn.]. Zilstra, J.J. 1973 Egg weight and fecundity in the North Sea herring {Clupea harengus). Neth. J. Sea Res. 6(l-2):173-204. Zonova, A.S. 1973 The connection between egg size and some of the characters of female carp (Cypriniis carpio L.). J. Ichthyol. 13:679-689. Abstract. — Determining absolute survival rates for larval fishes is ex- tremely difficult. However, many eco- logical questions concern relative sur- vival of two groups. For example, we might ask: (1) Do older larvae have higher survival than younger larvae? and (2) Do faster growers have high- er survival than slower growers? We present a simple model and several estimation schemes for the ratio of survival rates based on monitoring relative abundance of the two groups over time. When the logarithm of the ratio of abundances is regressed on time, the resulting estimate of slope is an estimate of the difference in in- stantaneous mortality rates. An esti- mate of the ratio of survival rates is obtained by exponentiating the slope. The model is shown to be a logistic model and can be fitted by maximum likelihood methods. Estimating Relative Survival Rate for T\A/o Groups of Larval Fishes from Field Data: Do Older Larvae Survive Better Than Young? John M. Hoenig Pierre Pepin Science Branch, Department of Fisheries and Oceans PO Box 5667. St John's. Newfoundland, AlC 5X1, Canada William D. Lawing Departments of Statistics and Industrial Engineering University of Rhode Island, Kingston, Rhode Island 02881 Manuscript accepted Ki April 1990. Fishery Bulletin. U.S. 88:485-491. Determining absolute survival rates for larval fishes is extremely difficult, even when cohorts can be accurate- ly identified by means of daily growth rings in the otoliths. For example, survival between times 1 and 2 might be estimated by catch/tow at time 2 -^ catch/tow at time 1. Due to ran- dom sampling error alone, this esti- mate may be nonsensical (>1), and the chances of obtaining nonsensical estimates increases with increasing patchiness in the distribution of lar- vae over space. In some cases, it may not be neces- sary to estimate absolute survival rates. Estimates of relative survival rate may be sufficient and easier to obtain. Many ecological questions concern the relative survival rates of two or more groups. For example, at a given point in time, the older lar- vae present should be larger than the yotmger larvae and, hence, may have a higher survival rate (Peterson and Wroblewski 1984, McGurk 1986). On the other hand, it has been suggested that larvae born later (i.e., younger larvae) may have a higher survival than larvae born earlier (Victor 1983, Methot 1983, Crecco and Savoy 1985, Rice et al. 1987). Also, it has been suggested that faster-growing larvae survive better than slower-growing ones (Rosenberg and Haugen 1982). In this paper, we consider how rela- tive survival rates can be estimated for two groups occurring at the same time and place from field data con- sisting of the composition of the catch at two or more times. The intuitive basis for the method is this: Changes in the relative abundance of two groups over time reflect differences in mortality rates (assuming no emi- gration or immigration occurs). The methods we present allow for catch- ability to differ for the two groups and to vary over time. However, the relative catchability (ratio of the catch- ability coefficients) of the groups can- not change over the time period con- sidered. Thus, if factors such as wind, currents, boat speed, or net clogging vary among sampling periods, then the same proportional change in the catchability coefficient must occur for both groups. This is less restric- tive than assuming constant catch- ability over time (as when estimating absolute survival rate by the decline in catch-per-unit-sampling-effort over time). Development of the model Suppose that at time t, a sample of larvae is obtained and that examina- tion of the otoliths reveals that all lar- vae are of the same (approximate) 485 486 Fishery Bulletin 88(3), 1990 age. It is also determined that larvae can be classified as either fast or slow growing on the basis of the widths of the growth increments. Thus, we can determine the proportion of the larvae that are fast growing. If we sample the same population at time t + l,we can again separate the larvae into fast- and slow-growing groups on the basis of the width of the otolith at time t (not at time t + 1). Alternatively, suppose that a sample of larvae at time t has two cohorts, where a cohort is defined to be all larvae hatching in a given week. We can follow the relative abundance of the two cohorts in the catch over time by taking repeated samples from the population. We will let the subscripts E and L refer to the two cohorts (e.g., E for early- and L for late-spawned). Let the number of larvae in the population's two cohorts at time t be A^^^, and Ni, . Suppose that the size of each cohort declines exponentially over time such that, for the ;th cohort (; 6 {E,L}), iV„ = A,„exp(-Z,0 where Z, is the instantaneous mortality rate (time^ ■) and Nn) is the initial abundance of the (' th gi'oup. Also suppose that the expected catch of animals from group i at time t (C,(), for a standard unit of effort, is pro- portional to the abundance, i.e., C,( = q,t N„ where 7,, is the time- and group-specific catchability coefficient. Then the ratio of expected catches, R, , is Rt = Cei to A^£„ exp(-Z£0 (1) We assume that the ratio of the catchability coeffi- cients (qLflqEt) is constant over the course of the study. Since the ratio of initial abundances (A^o/Afu) is also a constant, equation (1) can be rewritten R, = q exp(-Z^O exp(-Z£0 qexp{(ZE - Zi)t] (2) where g is a nuisance parameter that subsumes the catchability coefficients and initial abundances. Tak- ing logarithms of (2) results in a linear relationship with respect to time: \ogAR,) = loge(9) + (Z,: - Z,)t. (3) Thus, regressing the logarithm of the observed ratio of abundances (Rt) against time results in a linear relationship with slope equal to an estimate of the dif- ference in the instantaneous mortality rates. (The proper weighting to use in a weighted regression is discussed below.) Note that it is sometimes necessary to add a small constant to the numerator and denom- inator to avoid dividing by 0 or taking the logarithm of 0. Exponentiating the slope estimated by (3) provides an estimate of the ratio of the finite survival rates: SJSe = e'^'"/'.- where Si = e~^i; S^ = e'^t:. By the Taylor's series (delta) method, the asymptotic variance of the estimated survival ratio can be approx- imated by (Seber 1982): V{SJSe) = e<- •^■'"'"■' V(slope). Diagnostics Under the assumptions given above (constant ratio of catchabilities over time and constant difference in in- stantaneous mortality rates), a plot of the logarithm of the ratio of catches versus time would be expected to be linear (assuming the sample sizes are reasonably large). A departure from linearity suggests violation of one or both of the assumptions. This provides a diagnostic procedure to check on the assumptions. If catchabilities vary in a nonsystematic fashion, the fit of the regressions would be low and would not be in- fluenced by increasing sample size. Two-sample estimator A special case is when only two samples have been ob- tained. Then, exponentiating the slope of the line described by equation (3) reduces to the change-in-ratio estimator of relative survival described by Paulik and Robson (1969). Thus, Se R2 _ Cl2_^^E1 Rl Cix Ce2 (4) where the ~ symbol indicates estimated quantities. If the proportion of early spawners in samples 1 and 2 are denoted by P| and P2, respectively, then an esti- mate of the variance can be found by the Taylor's series method to be (Seber 1982, p. 382) V(SJSe) = lP,a-P,)]-' {(l-P,)^Po-r(P,) + (1-P,)^P,^ V{P.,)}. Hoenig et at Estimating survival rate over time for larval fishes 487 Here, V(Pi) and ViP-,) are the variances of the pro- portions, found by the usual formula for variance of a binomial, i.e.. ViP, where A'', is the number of larvae in sample ;', and i can take on the values of 1 and 2. Weighted regression Equation (3) can be seen to be a logistic model by noting that \oge(Rf) is the logistic transformation. That is, letting P, be the proportion in the sample at time t rj which belongs to group "L" (i.e., ), the logit Cl + Ce transformation of the proportion is l0git(P,) = l0ge|j^-^| = loge (/?,). This allows us to refer to known results for logistic models to determine the optimum weighting scheme. Not all observations on catch composition are of equal value in estimating the relative survival rate. This is because the variance of the ratio of catches for any par- ticular sampling date will be a function of the sample size and the proportions in the population. One method for specifying weights to be used in the regression, which explicitly accounts for this, is: weight, = Cei''^ + Cu'l When these weights are used in the regression, the resulting estimates are known as minimum logit chi- square estimates. Once the logistic regression model has been fitted, new weights can be computed using the predicted catch composition, rather than the ob- served composition. The regression can then be recom- puted, the weights updated, and the regression recom- puted, until adequate convergence is achieved. This procedure results in maximum likelihood estimates (McCullagh and Nelder 1983). Most standard statistical packages will perform logistic regression so that the user need not specify explicitly the iterative weighting scheme. The above weighting scheme is appropriate when lar- vae are sampled randomly, i.e., each larva is sampled independently of all other larvae. This situation is ap- proximated when the expected catch per tow is small (< 1) and tows are random over space. When the ex- pected catches are large, the sampling procedure re- sults in cluster samples. Consequently, the theoretical binomial variance is too small. However, the binomial variance becomes increasingly small as the sample size increases and as the proportions approach the extremes (0 or 1), and this is what would be expected for cluster sampling. Thus, weights computed from the binomial variance should be reasonably appropriate. Estimation when there are many small samples It sometimes occurs that catch rates are extremely low and many small samples are obtained. For example, sampling may be conducted daily with low intensity. In this case, the logistic regression procedure will not work well due to the occurrence of many zero catches. Estimation under the logistic model can still be accom- plished by constructing the likelihood function and solv- ing directly for the difference in mortality rates that maximizes the likelihood function. The probability that an individual, randomly selected from the catch at time t, is from group E (early-spawn- ing group) is equal to Pt(E) C. Et Cei + Cit QEt Neo e-'^fi' to Neo e-^f.' + qit Nio e" 1 + qe i^t (5) where q is the "apparent" initial ratio of abundances (actual initial ratio of abundances if the ratio of catch- ability coefficients is equal to 1) and AZis the difference in instantaneous mortality rates {Ze-Zi). The probability that an animal is from group L is the complement of (5): P,{L) = 1 \ + qe tiZt qe t£t 1 -I- qe^^ The likelihood function can then be constructed as the product of the probabilities for each animal in each sample: A = n (=1 1 \ + qe tiZi, n qe bZt. qe kZt, q«L e n (1 + qe'^'') (6) 488 Fishery Bulletin 88(3). 1990 tr 1980 cohort 2 vs 1 cohort 3 vs 2 cohort 4 vs 3 0- 0 r 234567" '3 45678 '3 45678 cohort 5 vs 4 cohort 6 vs 5 cohort 7 vs 6 . 2r CD 456789 '5 6789 10 "6 789 10 11 CJ) cohort 8 vs 7 cohort 9 vs 8 cohort 10 vs 9 O If . Ir ' "''7 8 9 10 11 12 "8 § 10 11 12 13 ~ '9 10 11 12 13 14 cohort 1 1 vs 10 cohort 12 vs 11 cohort 1 3 vs 1 2 1r lr ir 0 9o 11 12 13 14 15 ^1 12 13 14 15 16 '^2 13 14 15 16 17 sampling period Figure 1 Plots of the logarithm of the ratio of abundance (cohort ( + 1 -H cohort i ) versus sampling period (measured as 5-day intervals) in 1980. Slopes of the linear regression lines estimate Zf - Z^. (Note that the scale of the ordi- nate varies among plots.) In all 12 comparisons of cohorts, the slope is positive (later cohort has the better survival). where %- and ni are the total number of animals ob- served in groups E and L, respectively, t, is the time when the tth individual from group 1 was caught, and similarly for tj. The major statistical packages have routines for maximizing this function. Generalizations The logistic model is easily extended to enable one to incorporate the effects of covariates. For example, sup- pose one has information on the time elapsed since the sampling program began (/ ) and also on the number of days (t *) a power plant, which can be an additional source of mortality, has been operating for each of the sampling dates. Then instead of having t as the only explanatory variable, we can have a linear combination of explanatory variables in the exponent AZt, + 6] t* where 6i is the differential mortality rate (per time) of the two groups attributable to power plant operation. Example We consider data on American shad Alosa sapidissima in the Connecticut River, Connecticut, USA, kindly supplied by Victor Crecco and Thomas Savoy (Dep. Environ. Prot., Mar. Fish Office, Waterford, CT). A description of the study area and sampling methods, and a careful analysis of the data, are given in Crecco et al. (1983). Our purpose in considering these data is to illustrate the use of our method and to explore the kinds of questions that can be asked by study- Hoenig et at Estimating survival rate over time for larval fishes 489 cr. 1981 cohort 2 vs 1 cohort 3 vs 2 cohort 4 vs 3 2 3 4 5 6 7 '345678 ^3 X 5 6 7 8 cohort 5 vs 4 cohort 6 vs 5 cohort 7 vs 6 0) '456789 56 "6 7 8 9 to ^ 7 8 9 10 CTi cohort 8 vs 7 cohort 9 vs 8 cohort 10 vs 9 O 2 -1.7 —1 ' ^ 8 9 10 11 12 '8 9 10 11 12 13 '9 10 11 12 13 14 cohort 11 vs 10 cohort 12 vs 11 1o 11 12 13 14 15 9l 12 13 14 15 16 sampling period Figure 2 Plots of the logarithm of the ratio of abundance (cohort I + 1 -i- cohort i ) versus sampling period (measured as 5-day intervals) in 1981. Slopes of the linear regression lines estimate Z^ - Z[^. (Note that the scale of the ordi- nate varies among plots.) In 8 of the 11 comparisons of cohorts, the slope is positive (later cohort has the better survival). ing relative survival. For instance, Crecco and Savoy suggest that cohorts spawned later in the season have a higher relative survival based on comparison of lar- val and juvenile numbers. Our method provides a tool with which we can test this hypothesis for the larval stage while overcoming problems which may be associ- ated with sampling variability or limited sample sizes. Ichthyoplankton were sampled throughout the spring of 1980 and 1981. Approximately 30-40 larvae were aged for each 5-day sampling period. We compared the mortality of cohort i + 1 with that of cohort i for all possible i where cohort i is defined to be animals born in the ; th 5-day period of the season. This resulted in 12 comparisons for 1980 and 11 comparisons for 1981 (Figs. 1, 2). Because information on exact sizes of the samples was not available to us, we computed un- weighted regressions. This provides imbiased estimates of the regression parameters, but the estimates are not of minimum variance and the standard errors are not accurate (Weisberg 1980). The coefficients of determination were poor (r- range 4-98%, mean 55%, for comparisons with three or more observations) suggesting we cannot place much confidence in the magnitude of the estimates of dif- ferential mortality (Z^ - Zi). This is undoubtably due to the very small sample sizes. However, it is worth noting that in 12 of the 12 comparisons using the 1980 data, the differential mortality was positive (see Figure 3), i.e., the survival rate of cohort i + 1 (the later spawned cohort) was higher than the survival rate of cohort I (the earlier spawned cohort). Also, in 8 out of 11 comparisons using the 1981 data, the later-spawned cohort had the higher survival rate. If there were no differences in survival rates of the two cohorts in each 490 Fishery Bulletin 88 (3|, 1990 1980 2 3 4 5 6 7 8 910111213 cohort number i+1 1981 0) Q. O 23456789 1011 1213 cohort number i+1 Figure 3 Plot of estimated differential mortality (i.e.. Z for cohort i + 1 minus Z for cohort ( ) versus cohort number, i + 1 . Differential mortalities are estimated as the slopes of the regressions in Figure 1 (top) and Figure 2 (bottom). Note that in all but three cases, the later-spawned cohort is estimated to have lower mortality (higher survival) than the earlier-spawned cohort. pair, then one would expect half the comparisons would have a positive estimate of differential mortality and half would have negative estimates. Discussion We have presented three methods for estimating rela- tive survival rates of larval fishes. The first method is based on a logistic regression model. It provides a graphic way to check the assumptions of constant relative survival and constant ratio of catchabilities. A special case of this method is the two-sample estimator of Paulik and Robson (1969) (our eq. 4). Explicit spec- ification of the likelihood function (eq. 6) is necessary when many small samples are obtained, e.g., from a daily sampling program where the catch per tow is very small. An intuitive method for estimating relative survival rate would be to estimate the absolute survival rate of each group by the decline in catch-per-unit-effort be- tween two sampling times, and then to take the ratio of the two survival estimates. Thus, alternative estimate of S^ IS^ = Cl2 Ce2 Cfii C'j L2 Cpo c. .(7) £2 '-'LI The assumption necessary for the estimation of each survival rate is that the catchability has not changed over time. If one were to obtain an estimate of survival that is unfeasible (>1.0) one would be tempted to dis- card the data without computing relative survival. However, the expression to the extreme right in (7) is exactly equivalent to the two-sample estimator in (4). Thus, one can validly estimate relative survival rates even when estimates of absolute survival are obtained which are nonsensical. This is because the relative survival estimators do not require the catchabilities to remain constant over time, only the relative catch- abilities. On the basis of existing information, it is not possi- ble to state quantitatively in what proportion of fish populations or in which situations late-spawning larvae will survive better than early-spawning larvae. In our example, sample sizes were quite small (~30-40 age determinations per 5-day sampling period) so that esti- mates of relative survival rate were imprecise. It is in- teresting to note, however, that in 12 of 12 comparisons in the 1980 data and in 8 of 11 comparisons in the 1981 data the later-spawned cohort (week / -i- 1) had a higher survival rate than the earlier-spawned larvae (cohort from week 0 (Fig. 3). However, our results indicate that late-hatching (smaller) larvae survive a given calendar period better than early-hatching (larger) lar- vae which is opposite to the general findings that lar- val fish mortality rates decrease with increasing size and/or age (Peterson and Wroblewski 1984, McGurk 1986). This suggests that whatever the cause of mor- tality (e.g., predation, transport), the vulnerability of larval shad in the Connecticut River to this factor in- creases with age and/or size during the period studied. Hoenig et a\ Estimating survival rate over time for larval fishes 491 Our method was developed for the situation in which cohorts can be sampled simultaneously. Thus, if in a particular week gear efficiency was better than nor- mal, any changes in catchabilities would tend to affect both cohorts so that the ratio of catchabilities might stay relatively constant. It may be tempting to use our method to compare cohorts over the same part of the ontogeny (i.e., cohorts of the same age occurring at dif- ferent times of the season). However, in this situation there is no reason to believe that variations in catch- ability over time of the first cohort will be tracked by variations in catchability of the later cohort. Hence, there is no advantage in using our method over tradi- tional methods of estimating absolute survival based on declines in catch-per-unit-effort. Finally, we note that methods for estimating relative survival have wide applicability beyond the study of lar- val fishes. A classical problem is Lee's phenomenon in which back-calculated sizes at the first annulus do not agree with the observed sizes of the young fish. Jones (1958) proposed as explanation that faster-growing fish may have a different mortality rate than slower growers. This can be studied by observing the propor- tion at each age that have a small back-calculated size at the first annulus. Hoenig and Lavdng (1983) describe how the logistic model might explain the occurrence of progressively skewed sex ratios with age. Differen- tial survival has been studied by a variety of mark- recapture methods in the context of fitness and natural selection (see Manly 1985) and as a means of estimating impacts of power plants (see Burnham et al. 1987). Another application is to the evaluation of stocking suc- cess as related to genetic strain of fish or hatchery treatment. That is, the ratio of abundances of two strains can be monitored over time to study whether one strain survives better than the other. Acknowledgments We would like to thank Russell Millar, Geoffrey Evans, Dennis Heisey, William W. Fox Jr., and an anonymous reviewer for helpful comments. Victor Crecco and Thomas Savoy generously provided data on shad lar- val abundance for our example. Citations Burnham, K.P., D.R. Anderson, G.C. White, C. Brownie, and K.H. Pollock 1987 Design and analysis methods for fish survival experiments based on release-recapture. Am. Fish. Soc. Monogr. 5. Crecco, V., and T. Savoy 1985 Effects of biotic and abiotic factors on growth and relative survival of young American shad, Alosa sapidissima, in the Connecticut River. Can. J. Fish. Aquat. Sci. 42:1640-1648. Crecco, V., T. Savoy, and L. Gunn 1983 Daily mortality rates of larval and juvenile American shad (Alosa sapidissivia) in the Connecticut River with changes in year-class strength. Can. J. Fish. Aquat. Sci. 40:1719-1728. Hoenig, J.M., and W.D. Lawing 1983 Using incomplete tagging data to estimate the total mortality rate when the sex ratio is skewed. ICES CM. 1983/D:24. Jones, R. 1958 Lee's phenomenon of "apparent change in growth rate", with particular reference to haddock and plaice. Int. Comm. Northwest Atl. Fish.. Spec. Publ. 1:229-242. Manly, B.F.J. 1985 The statistics of natural selection. Chapman and Hall. NY, 484 p. McCullagh, P., and J. A. Nelder 1983 Generalized linear models. Chapman and Hall, NY, 261 p. McGurk. M.D. 1986 Natural mortality of marine pelagic fish eggs and larvae: Role of spatial patchiness. Mar. Ecol. Prog. Ser. 34:227-242. Methot, R. 1983 Seasonal variation in survival of larval northern anchovy, Engraulis mordcuc, estimated from age distribution of juveniles. Fish. Bull, U.S. 81:741-7,50. Paulik. G.J., and D.S. Robson 1969 Statistical calculations for change-in-ratio estimators of population parameters. J. Wildl. Manage. 33:1-27. Peterson, I., and J.S. Wroblewski 1984 Mortality rate of fishes in the pelagic ecosystem. Can. J. Fish. Aquat. Sci. 41:1117-1120. Rice, J. A., L.B. Crowder, and M.E. Holey 1987 Exploration of mechanisms regulating larval survival in Lake Michigan bloater: a recruitment analysis based on characteristics of individual larvae. Trans. Am. Fish. Soc. 116:703-718. Rosenberg, A. A., and A.S. Haugen 1982 Individual growth and size-selective mortality of larval turbot reared in enclosures. Mar. Biol. (Beri.) 72:73-77. Seber. G.A.F. 1982 The estimation of animal abundance and related param- eters. Macmillan, NY, 654 p. Victor, B.C. 1983 Recruitment and population dynamics of coral reef fish. Science (Wash. DC) 219:419-420. Weisberg, S. 1980 Applied linear regression. John Wiley, NY. 283 p. Abstract.- The horizontal and vertical movements of yellowfin tuna Thunnus albacares and bigeye tuna T. obes2is captured near fish-aggre- gating devices (FADs) were deter- mined using pressure-sensitive ultra- sonic transmitters. The movements of these FAD-associated fish were compared with the tracks of yellow- fin tuna not associated with FADs. Tracks from 1 1 yellowfin and 4 big- eye tuna were obtained; these in- cluded 23 complete 24-hour periods of observation. Whether associated with FADs or a 40-fathom (75-m) reef dropoff, most yellowfin and big- eye tunas exhibited similar diurnal patterns. The fish tended to remain tightly associated with FADs or the reef dropoff during the day, move away at night, and return the next morning. The maximum range of these nighttime excursions averaged approximately 5 nmi. These tunas apparently treated the FADs as out- liers of the coastal topography. This may not be the same behavior that results in the association of these species with drifting objects such as logs. Tuna can learn FAD positions and navigate precisely between FADs that are at least 10 nmi apart. When not associated with FADs or the 40-fathom dropoff, yellowfin tuna oriented to the bottom of the mixed layer (50-90 m) in daytime, whereas the bigeye occupied depths between 190 and 250 m. The daytime distribution of bigeye tuna seemed to be influenced by the depth of the 15°C isotherm. Both species swam closer to the surface at night. Swim- ming strategies possibly associated with energy and thermoconservation were observed. Horizontal and Vertical Movements of Yellowfin and Bigeye Tuna Associated with Fish Aggregating Devices* Kim N. Holland Hawaii Institute of Marine Biology, University of Hawaii PO Box 1346, Coconut Island, Kaneohe, Hawaii 96744 Richard W. Brill Randolph K.C. Chang Honolulu Laboratory, Southwest Fisfneries Science Center National Marine Fisheries Service, NOAA 2570 Dole Street. Honolulu, HI 96822-2396 Tropical and subtropical pelagic fishes aggregate around natural, drifting debris such as logs and mats of algae. Man-made, anchored floating objects, known as fish aggregating devices (FADs), have also proven effective in attracting and holding commercially important pelagic species (Shomura and Matsumoto 1982, Matsumoto et al. 1981, Brock 1985). Species com- monly found around FADs in Hawaii are yellowfin tuna Tliunnus aWamres, bigeye tuna T. obesus, skipjack tuna Katsuwcmus pelamis, dolphin or mahi- mahi Coryphaena hippurus, and wa- hoo or ono Acanthocybium solandri. The first commercial FADs were de- ployed in the calm waters of the Phili- ppines in the early 1970s to attract yellowfin tuna (Kihara 1981), and in 1977 experimental FADs designed for use in high-energy, deep-water environments were anchored around the Hawaiian islands (Matsumoto et al. 1981). Due to the success of these bouys in aggregating fish, FADs have come to play an important role in the commercial, subsistence, and recrea- tional fisheries of all the tropical and sub-tropical oceans of the world. Manuscript accepted .5 February 1990. Fishery Bulletin, U.S. 88:493-507. *Sea Grant Publication UNIHI-SEAGRANT- JC-90-01, and Hawaii Institute of Marine Biology Contribution No. 774. Despite the widespread use of FADs, left unanswered are important ques- tions concerning their range of influ- ence, optimal placement, and impact on surrounding fish populations. The behavioral patterns that result in fish being associated with floating objects of any type are poorly understood and not widely agreed upon (Gooding and Magnuson 1967, Hunter and Mit- chell 1967, Fedoryako 1982). How- ever, the underlying mechanisms of aggregation have generally been as- sumed to be the same for both free- floating phenomena such as logs, and anchored man-made objects such as FADs (Shomura and Matsumoto 1982). To better understand the be- havior of tunas associated with FADs, the movements of yellowfin and big- eye ttina caught within 500 m of these devices were tracked using pressure- sensitive ultrasonic transmitters. The movements of these fish were com- pared with those of tunas that were not associated with FADs, but with the reef perimeter surrounding the islands. In Hawaii and elsewhere in the Pacific, yellowfin tuna can be found in the vicinity of the 40-fathom (75-m) isobath where the island topog- raphy abruptly descends into very deep water. In fact, the dropoff of the outer reef edge is often so steep that 493 494 Fishery Bulletin 88(3), 1990 the horizontal distance between the 40-fathom and 100-fathom (187-m) isobaths is measured in tens of meters. To observe the behavior of non FAD-associated tuna, three yeliowfin tuna were tracked that were caught on or near the 40-fathom isobath around Oahu, Hawaii. The influence of FADs on the vertical movements of these species was investigated by comparing the depth distributions of the two species in on-FAD (within 500 m of a FAD) and off-FAD (beyond 500 m of a FAD) situations. The depth distributions were analyzed with respect to ambient ocean temperatures, as measured by expendable bathythermographs. Methods The horizontal and vertical movements of individual fish were monitored by pressure-sensitive, 50 KHz ultrasonic transmitters (Vemco, Halifax, Nova Scotia, Canada). Fish caught on the tracking vessel by troll- ing and pole-and-line techniques were fitted with trans- mitters by passing two nylon "tie-wraps" through the dorsal pterygiophores and trunk musculature adjacent to the second dorsal fin (Fig. 1). The fish were out of the water for approximately 1 minute during this pro- cedure. The fish were then released and followed using a directional hydrophone mounted on the tracking boat. Whenever possible, a distance of approximately 200 m was maintained between the fish and the boat. Geo- graphic position was determined every 15 minutes by a combination of Loran-C, radar, bathymetric, and visual fixes. These techniques evolved from pioneer- ing work in tuna tracking by Yuen (1970), Laurs et al. (1977), Carey and Olson (1982), and Carey (1983). A summary of previous ultrasonic tracks of tuna has been compiled by Hunter et al. (1986). A detailed account of the methods used in the current study has been published previously (Holland et al. 1985, Bayliff and Holland 1986). The transmitters were equipped with pressure sen- sors, which modulated the rate of pulse transmission in response to changes in water pressure (depth). Thus, vertical movements of the fish were determined by measuring the time between signal pulses. These pulses were recorded on audiocassettes for analysis ashore. Expendable bathythermographs deployed approx- imately every 3 hours provided ocean temperature pro- files, which were superimposed on the vertical move- ment plots. Pooled time-at-depth and time-at- temperature histograms were constructed from all Reference to trade names does not imply endorsement by the National marine Fisheries Service, NOAA. Figure 1 Transmitters were attached to yeliowfin and bigeye tuna with two nylon straps inserted through the dorsal musculature and pterygio- phores associated with the second dorsal fin (from Holland et al. 1985). tracks combined, using 10-m and 1°C bins averaged every 10 minutes. The data were analyzed for daytime and nighttime depth distributions. Differences in the temperature and depth distributions for day-versus- night and on-FAD versus off-FAD were analysed by factorial analysis of variance (SAS General Linear Models Procedure, SAS 1985). Portions of tracks oc- curring on-FAD or near the reef dropoff were not in- cluded in these calculations. Thus, the depth histograms and overall averages are compiled from movements that were not constrained by bottom topography or in- fluenced by the presence of floating objects. The tracking techniques used in this study are not capable of detecting small-scale changes in the horizon- tal swimming direction of the fish. Consequently, sus- tained swimming speeds have been calculated only from sections of tracks in which fish demonstrated pro- longed straight-line movements. Swimming speeds were calculated as distance traveled per unit time and in terms of body lengths (fork length, FL) per second. Swimming speeds were calculated for each complete hour of "staight" running; these hourly rates were averaged to yield a sustained swimming speed for the relevant section of each track. Vertical movements were not included in these calculations. Daytime and nighttime time-at-temperature histo- grams were generated for both yeliowfin and bigeye tuna. Because of differences in absolute ocean-surface temperatures between different tracks, and because of Holland et al Movements of Thunnus albscares and T obesus near fish-aggregating devices 495 previous indications that ocean thermoclines may be important orientation cues for tuna (Carey and Olson 1982), the time-at-temperature data were analyzed relative to the surface mixed layer. Thus, the mixed layer is referred to as ML, and the progressively colder 1°C isotherms as ML-1, ML-2, and so on. Results The results reported here are from tracks of 1 1 yellow- fin and 4 bigeye tuna. Duration of tracks from release to track termination ranged from 5 hours to 6 days, with an average duration of 31.6 hours. Overall, these tracks encompass 23 complete 24 hour periods of obser- vation. All 4 bigeye and 8 of the 1 1 yellowfm tuna were caught within 500 m of FADs. The remaining three yellowfin tuna tracks were from fish caught on or near the 40-fathom contour around the island of Oahu. Horizontal movements For yellowfin tuna, a clear pattern of horizontal move- ment was apparent. That is, during daylight the fish usually moved within a home range, staying in prox- imity to certain well-defined physical features, such as FADs or the outer reef dropoff. Eight of the 1 1 yellow- fin tuna and 2 of the bigeye tuna showed aspects of this behavior, and this pattern was exhibited by both the FAD-associated and coastline-associated fish. Most of these fish made diurnal movements away from these daytime haunts, with the initial change in location often occurring around sunset. FAD-associated yellowfin tuna These tuna spent daylight hours very close to the FADs, and then left those locations sometime between late afternoon and early nighttirhe to embark on extensive nighttime ex- cursions, returning to the same or another FAD the following day. Five of the eight FAD-associated yellow- fin tuna displayed aspects of this diurnal on-FAD and off-FAD behavior. A sixth FAD-associated fish showed signs of displaying this behavior, but was lost before this could be confirmed. The following is a brief synop- sis of their horizontal movements. Yellowfin tuna YF8404 (51 cm FL) was caught and tagged at V FAD at 0927, and stayed within 100 m of the FAD for 11.3 hours before leaving it a few minutes after sunset (1925, Fig. 2A). The fish then made a nighttime excursion totaling 12 nmi before returning to within 100 m of the FAD by 1312 the next day. During the last few hours of this approach the fish moved very slowly, as if it was drifting in the prevail- ing current. Once at the FAD, the fish again stayed close to the buoy until departing at sunset on a second ' YELLOWFIN TUNA . _ ' ' 8404 \ DAY 1 \ <[7 KAENA POINT ^X A . \ \^ \ ^-- -^ ^\^ RELEASED ^^ -^ X — _ \^ 0927 ^ ^\ ' ' Ij^-::.'' ) ^-V ) / i IT 0 1 1 1 1 I - nmi y 0600 ' 1 I ^ \ YELLOWFIN TUNA - 8404 \ DAY 2 ^r KAENA POINT X. B \ \ ^ ^ N, ^ ^ -^ \ X " \ \ 1312 '°°°V^**\ \ 1 - "-^ \ / V > '*• 1 \ '■^. . C^< " (' 0600 \ x 1 > 1 X 1 1 '\ : 1 0 1 1 1 \, 1 1 1 1 1 ^^ 1- v; nmi 1 1 1 Figure 2 (A) 48-hour track of yellowfin tuna YF8404 tagged and released at 0927 at V FAD, Oahu. Circles = hourly marks when fish is on-FAD; squares = hourly marks when fish is off-FAD. In this and subse- quent figures, solid line = daytime movements, dashed line = night- time movements, and marks on figure axes = 1° divisions. On the first day (A) the fish remained very close to the FAD as it moved in a circular path in the current, departed the FAD after nightfall, and returned at 1312 the following day. The second day's behavior (B) was similar, despite a different pattern of movement by the FAD. nighttime excursion of 11.25 nmi. The maximum dis- tance away from the FAD was 3.0 nmi on the first night and 5.0 nmi on the second. On both excursions, the fish commenced moving back toward the FAD im- mediately after sunrise (Fig. 2A,B). Two hours after fish YF8504 (47 cm FL) was tagged and released at S FAD at 0551, it made a 4-hour, 1-nmi excursion away from the buoy before taking up a posi- 496 Fishery Bulletin 88(3), 1990 - ■ 7 / 1 — NT \ 2052 ^''^Sv ^"^^'^ P° ■ • \ \ • ■ «V FAD 'f / 1 / • YELLOWFIN TUNA ^ ■ 8504 i I ■ ■ R FAD ^^^~\VV. v • ^)^ ^ ■ 0,00--^ . i V ■ ■ 0 1 2 *''VW^ ''*° \ ^-\ ■ nmi V 1.1. V 1 \ Figure 3 Yellowfin tuna YF8504 spent 8 daylight hours on the upcurrent side of S FAD before moving away at night. The next morning it moved to R FAD until disturbed by a school of porpoises, when it immediate- ly took a direct course to the island dropoff, which it patroled for the rest of the day. In this and subsequent figures, squares = hour- ly position marks. tion within 50 m of the FAD for the remaining 8.0 hours of dayHght (Fig. 3). During this time the fish was almost exclusively on the upcurrent side of the bouy. At 2345 the fish left the FAD and moved offshore; by sunrise the next day, it was 3.25 nmi from S FAD. The fish then moved 3.0 nmi in a direct course to the nearest adjacent FAD (R) arriving there at 0910. Once there, it stayed extremely close to the FAD and in the company of a school of other yellowfin tuna that could be seen from the surface. At 1020 a group of porpoises arrived and dove through the school of tuna, which ap- peared to scatter in all directions. The tagged fish dove and swam to the 40-fat.hom contour of the island, where it stayed for the rest of the day. Contact was lost just after sunset on the second night. Fish YF8305 (55 cm FL), after being tagged and released at S FAD at 0705, swam for 4.0 hours on a direct course to V FAD, approximately 10 nmi away from the release point (Fig. 4). The fish spent 95% of the next 5.0 hours moving around on the upcurrent side of this FAD. In the late afternoon, the fish departed the FAD and spent most of the night farther offshore, moving in a 15-nmi loop before heading back toward V FAD the following morning. Just after first light, when the fish had returned to within 2 nmi of the FAD Figure 4 Yellowfin tuna YF8305 departed S FAD immediately after release at 0705 and took a direct course to the next nearest FAD, V (R was off station at the time), where it patroled almost exclusively in the upcurrent area (insert) before making an overnight excursion total- ing 15 nmi. and appeared to be headed back even closer, the transmitter was shed. The tracks of fish YF8406 (62.5 cm FL) and YF8501 (44.0 cm FL) were similar. During the daytime, both fish stayed very close to their respective FADs (for 9 and 12 hours, respectively), and both were lost at sunset in deteriorating sea conditions. Because there was no evidence of transmitter failure, and exhaustive searches at the FAD locations after dark failed to relocate the signal, it is reasonable to assume that both these fish were lost when they made sudden evening departures from the vicinity of their respective FADs. Fish YF8302 (64 cm FL) was caught and tagged at S FAD, but lost 5.0 hours later. When contact was lost, the fish had departed the release site, but appeared to be curving back toward the bouy. Fish YF8502 (74.5 cm FL) was similar to other FAD- associated yellowfin tuna in that, for the first 5.0 hours of the track (0600-1100), it stayed in close proximity (within 100 m) to S FAD. However, it then moved steadily away but was lost 12 nmi southwest of the buoy due to equipment failure after 13.5 hours of track and before any possible long-term trend had become apparent. Fish YF8506 (75.25 cm FL) was caught near S FAD at 0630. Unlike the other FAD-associated yellowfin, it immediately began to move steadily away, maintain- ing a constant southwest course for the next 17.5 hours, at which time the track was terminated (Fig. 5). Holland et al Movements of Thunnus albacares and T obesus near fish-aggregating devices 497 Figure 5 Bigeye BE8401 and yellowfin tuna YF8506 both departed S FAD almost immediately upon release and maintained quite straight azimuths at constant speeds. During darkness, both fish slowed down, and YF8506 became more variable in swimming direction. FAD-associated bigeye tuna Four tracks were made of bigeye tuna caught near FADs. Of these four, one was lost after a few hours and before any diurnal pat- terns became apparent, one moved continuously away from the FAD (Fig. 5), and two displayed diurnal on- FAD and off-FAD behavior similar to that exhibited by the yellowrfin tuna. The following is a synopsis of these two tracks. Bigeye tuna BE8205 (74.5 cm FL) was caught and tagged 200 m from F FAD off Kona, Hawaii, about 1 hour after sunset (Fig. 6). Upon release, it proceeded on a 24.0-nmi overnight excursion which took it away from the FAD, before gradually reapproaching the FAD in the afternoon of the next day. The maximum distance it moved away from the FAD was 5.75 nmi. This track was then terminated after 24.0 h. Fish BE8706 (72.0 cm FL) was caught at first light within 50 m of C FAD located off Kealakekua, Hawaii. The fish remained extremely close to the bouy through- out the day until 1 hour after sunset, when it made a 2-nmi excursion that lasted approximately 3 hours. The fish then returned to the immediate vicinity of the buoy, where it remained for the rest of the night and at least through noon of the following day, when the track was terminated after 30.0 hours. Coastline-associated yellowfin tuna Three tracks were made of yellowfin tuna caught on the reef drop- off on the west coast of Oahu in water between 40 and 50 fathoms (75-95 m) deep. These fish were in the same size range as those caught around the FADs, and tracks spanning 6 days, 36 hours, and 37 hours were obtained, which encompassed a total of six day-night transitions. Two of the three fish displayed offshore excursions Figure 6 Caught and released after sunset at F FAD off the Kona coast of Hawaii, bigeye tuna BE8205 made a 24-nmi overnight loop before returning to the FAD by late afternoon of the following day. every night that they were tracked (a total of five nights). These tracks were as follows. Fish YF8303 (70.0 cm FL) was caught and tagged at 0855 about 0.75 nmi off the leeward coast of Oahu on the 40-fathom dropoff (Fig. 7A). For the next 7 hours, the fish remained on or near the 40-fathom dropoff, before moving slightly further offshore in the late afternoon. During the subsequent 12 hours of darkness, the fish made a 17-nmi offshore excursion before returning to the 40-fathom contour in the sec- ond hour of daylight on the following day. The max- imum distance away from the point at which it re- encountered the dropoff was approximately 8.25 nmi. The fish remained on the reef slope for the remaining daylight hours until contact was lost at 1130, after 26 hours of tracking. Forty-eight hours after contact was lost, a hydro- acoustic search of the 40-fathom contour reestablished contact with the fish, and the track was resumed at 1426 of the third day (Fig. 7B). The fish spent the re- maining daylight hours moving back and forth along the dropoff before moving offshore in late afternoon. The fish spent all night on a 13-nmi excursion before returning to the 40-fathom contour at first light on the next day. The greatest distance from the point of return to the dropoff was approximately 5.25 nmi. The track was suspended at 0730 on the fourth day. The fish was relocated on the 40-fathom isobath on the afternoon of the following (fifth) day at 1642 (Fig. 7C). Again, the fish moved offshore in the late afternoon and spent all night in deeper water on a 7.5 nmi ex- cursion. Maximum distance from the point of return to the dropoff was 3.3 nmi. The time of arrival back at the 40-fathom contour was within 1 minute of the time of arrival on the previous day (Fig. 7B,C). The track was terminated at 0630 on the sixth day because of an impending storm. In summary, this fish was tracked over three day-night-day cycles spanning 6 days and, on each occasion, it made a nighttime off- 498 Fishery Bulletin 88(3), 1990 ';■■■<; I ■ » — I ■ -T ~"^ KAENA POINT \ 1130^^ v^ ■ \ _---"" --^ - ^V\ A ■ »V FAD ""x \ YELLOWFIN TUNA 7 I 8303 7-8 OCTOBER J^°° /v0900 - .^ (L T / 'S F^ ■ • \ / )^" M ->"*' k^. i'^> ^ /' X \ ^ ,^ ■-0 ^ i.-^ ' ,-' /'"osoo \ V^ * «^ ^ V > \ .j ] "- \ \f I Y 0 1 2 1 1 f @ S FAD \ . nmi /' \ 1 1 1 1 1 1 till \ k — . . (. . . \ \ 1 I 1 r ^^^ KAENA P0IN1 0730 ^v . "A\ B -.- *V FAD > \ \ ; \ J - V ^1 . i-2300 .^ [ ir^.--^' \ \^ YELLOWFIN TUNA \ \ 8303 10-11 OCTOBER 1 \ 0 1 2 1 1 1 eS FAD ( I nmi \ ■ . . . • I.I ■ 1 1 Figure 7 Observed over a period spannint; ti days, yellowfin tuna YF8303 moved offshore from the reef dropoff on all three nights that it was tracked and returned to the same section of coastline each morn- ing. Day-1 track (A) commenced at 0900 and terminated at 1130 the following day. Track 2 (B) started 2 days later at 1426 and terminated at 0730 next morning. The third track (C) commenced at 1642 and terminated at 0630 on the morning of the sixth day. FAD R was not on station at the time of these tracks. . • ( ■ ■ ■ '"S^^KAENA POINT \ X,,^^ ®V FAD A 1642 c ■ , YELLOWFIN TUNA 8303 i y^0630 ■ 12-13 OCTOBER /'A f 2100 ^, . \^ : ■ V \ ■ • eS FAD ) 0 1 2 1 1 1 ' \ nmi V 1 1 1 1 , ' ' shore excursion followed by a return to the 40-fathoni contour on the following morning. Fish YF8503 (47 cm FL) was caught on the 50- fathom contour at 1200 and stayed in this vicinity for 7.0 hours before moving offshore in the late afternoon to make a 24-nmi nighttime excursion (Fig. 8). The next morning, the fish returned to within 0.5 nmi of the sec- tion of coastline it had left the previous evening, the maximum distance from the point of return being 9 nmi. The tracking was temporarily suspended at 0930 and resumed at 1150, when the fish was relocated on the 40-fathom contour near Kaena Point. P>om here the fish embarked on a prolonged and constant run around the point and along the 40-50 fathom contour of the north shore of Oahu. At sunset of the second day, the fish again moved offshore. When contact was lost at 0100, the fish was 7.5 nmi offshore. Fish YF8405 (57 cm FL) was caught on the 50- fathom contour and patroled this area for about the next 24 hours. This fish did not move offshore at night. Tracking was suspended at 0715 the second day, but resumed at 1715 the same evening when the fish was Holland et al Movements of Thunnus albacares and 7" obesus near fish-aggregating devices 499 YELLOWFIN TUNA 8505 0 12 3 J 1 1 T — I — I I — I I — I I — » — I I — r 0100 Figure 8 Yellowfin tuna YF8503 moved offshore and made a complete loop during the first night. Tracking was suspended between 0930 and 1150 of the second day. During the second day it swam at a con- stant pace along the reef dropoff bordering the north shore of Oahu before again heading offshore after sunset on the second night. This demonstrates that the diurnal onshore-offshore pattern is not site- specific to the west coast of Oahu. Contact was lost at 0100. relocated on the 40-fathom contour. At times, the fish moved into water less than 30 m deep, but was lost just after simset when the transmitter battery failed. How- ever, this fish was caught two weeks later by a fisher- man trolling at R FAD, 7 nmi away from the initial release point. Combining the FAD-associated and coastline-asso- ciated fish, 7 of the 11 yellowfin tuna demonstrated diurnal behavioral patterns in which daytime haunts were abandoned at night. These 7 fish were observed to make a total of 11 nighttime departures from day- time locations. Of those 11 nighttime excursions, 7 were demonstrated to be completed loops wherein the fish returned to their original starting points. In addi- tion, two FAD-associated bigeye tuna completed two nighttime-initiated loops. For the yellowfin tuna, the average maximum distance away from the following morning's destination was 5.28 nmi (SD = 1.9, N = 7). When the two documented bigeye tuna loops are in- cluded, the average maximum distance is 4.97 nmi (SD = 2.0, N = 9). Swimming speeds Unlike the majority of the fish, one bigeye (BE8401) and one yellowfin tuna (YF8506) almost immediately departed their respective FADs and continued to swim away along straight azimuths for the entire duration of the tracks (Fig. 5). Conse- quently, large sections of these tracks were included in calculations of sustained swimming speeds. Daytime sustained swimming-speed data are summarized in Table 1. The greatest distance traveled in 1 hour was ex- hibited at night by YF8503 which traveled 5.0 nmi. This is equivalent to 2.57 m (4.6 body lengths) per second (Fig. 8). When YF8504 was driven away from R FAD by a porpoise school, it swam the first half-hour towards the 40-fathom contour at a rate of 1.32 m (2.4 body lengths) per second (Fig. 3). Daytime sustained swimming speed Tabie 1 i of selected yellowfin (YF) and bigeye (BE) tunas. Track number Fork length (cm) Segment duration Oi) Average hourly speed ( ± SD) m/s Body lengths (FL)/s YF8506 YF8502 YF8503 YF8305 BE8401 72.25 74.50 .56.00 54.50 57.00 10.0 7.0 6.0 4.0 YF Mean 13.0 Combined mean 0.89 ±0.15 0.95 ±0.10 1.80 ± 0.38 1.32 ± 0.24 1.19 ±0.20 1.28 ±0.14 3.20 ± 0.68 2.40 ±0.40 = 1.24 ±0.41 0.79 ± 0.20 2.01 + 0.96 1.38 ±0.35 = 1.15 ±0.41 1.89 + 0.88 500 Fishery Bulletin 88(3), 1990 ELAPSED TIME (min) 10 20 25 30 35 SURFACE 50 100 E 150 200 Q. m 250 Q 300 350 400 Figure 9 Fine-scale behavior of bigeye tuna BE8401 shown by 45-minute seg- ment of vertical movements transcribed from data tape, including a rapid dive of 230 m in 1 minute, a rate equivalent to 6.7 body lengths/second. Ambient temperature at the maximum depth of 380 m was 9°C. Vertical movements The tape recordings of the piilsed sig^ial from the trans- mitters resulted in detailed, continuous plots of vertical position. Rapid and small-scale changes in depth were discernible (e.g., Fig. 9), as were longer-term diurnal shifts in vertical distribution. Day-night depth distributions In general, yellowfin tuna swam closer to the surface during darkness. For example, a dramatic change in the depth of swimming of fish YF8305 occurred exactly at sunset and lasted until first light the following morning, when it again began to swim deeper (Fig. 10). Combining all the off- shore and off-FAD portions of the yellowfin tuna tracks yielded an average daytime depth of 71.3 m (±42.0 SD, A'^ = 333), whereas the average nighttime depth was 47.3 m ( + 33.1 SD, iV= 444). These daytime and night- time depth distributions (Fig. 11A,B) are significantly different (ANOVA, p = 0.0001). However, this analysis also indicates significant variability among individual tracks. Bigeye tuna swam at significantly greater daytime depths than the yellowfin tuna, and showed even greater shifts between daytime and nighttime distribu- tions. When off-FAD, the predominant bigeye tuna daytime distribution was between 220 and 240 m, whereas the predominant nighttime depth was between 70 and 90 m (Fig. 12A,B). On-FAD versus Off-FAD depth distributions Swim- ming depths of five yellowfin and one bigeye tuna that were on- and off-FAD during daytime hours support the proposition that FADs tend to bring the fish closer to the surface than they would normally be in other off- shore parts of the ocean. These tracks were analyzed to determine if significant changes in depth distribu- tion occurred when the fish were aggregated around the FADs, independant of diurnal influences. When on- FAD, three of the yellowfin tuna were closer to the sur- face than when off-FAD, one showed no change in depth, and one was deeper. Pooling these data for this subset of yellowfin tuna gave a mean on-FAD depth of 59.3 m (±30.7 SD, iV = 225), which is significantly different from the mean off-FAD depth of 85.2 m (±52.0 SD, N=239). DEPTH (m) YELLOWFIN 8305 TIME Figure 10 Plot {24-hour) of vertical movements of yellowfin tuna YF8305. In this and subsequent figures, S = sunset. R = sunrise. Arrows denote the beginning and end of time spent at V FAD. Before and after this period, the fish exhibited the "traveling" behavior of making abrupt shifts between the surface and the top layers of the thermocline. Once at the FAD, the vertical behavior became predominantly sinusoidal. All dives made both on- and off-FAD had a consistent floor around 20°C between 140 and 150 m. The rapidly descending trace at end of plot denotes the tag being shed from the fish. Holland et a\ : Movements of Thunnus albacares and T obesus near fish-aggregating devices 501 35 A B I I I I I I I o o o O) — o CN o n DEPTH INTERVAL (m) Figure 1 1 Yellowfin tuna daytime distribution (A) shows two modes, one at the surface and one between 60 and 90 m, which corresponds to the normal position of the interface between the mixed layer and the top layers of the thermocline. Nighttime distribution (B) reflects the tendency of these fish to swim closer to the surface at night. In this and subsequent histograms, bars = standard error. The tendency of FADs to bring fish closer to the sur- face was even more apparent in the deeper-swimming LJ 25 20 15 _i 10 < I— o LJ Q_ O 0 ^ 25 A If 1 \ t f t + + ■h-M 20-- 15-- 10 5 VJ 0 IMlMMMMM^^i-^4 B t t t t'Mf^ I oooSoooCDoooS r^cn--'^inr^a^^r-)inco5 DEPTH INTERVAL (m) Figure 12 Daytime distribution of bigeye tuna (A) displays a major mode bet- ween 200 and 240 m, whereas nighttime distribution is much shallower (B), with the predominant distribution falKng between 70 and 90 m. bigeye tuna, as was dramatically demonstrated by BE8603 which stayed close to S FAD for 6 hours of daylight before moving away at midday. Figure 13 il- lustrates the rapid change of depth that coincided with BIGEYE TUNA 8603 ^ilSyP^^^te^to^s^Si?^ 'rjjj/ya^'V^^ LiJ Q250: 300^ 350 TIME Figure 13 Plot of the on-FAD and off-FAD movements of bigeye tuna BE8603. Arrow indicates when fish departed R FAD. 502 Fishery Bulletin 88(3), 1990 BIGEYE TUNA 8401 Figure 14 Plot of day-night-day movements of bigeye tuna BE8401. The rapid dive shown in Figure 9 can be seen at 0915 on the first day. Figure 15 Plot of vertical movements of bigeye tuna BE8205 showing extremely regular daytime upward excursions. These excursions consistently terminated in the zone encompassing the interface between the mixed layer and the thermocline. the departure of this fish from the FAD. The subse- quent depth and pattern of swimming were similar to those shown by the other bigeye tuna that departed their respective FADs almost immediately upon release (Figs. 14, 15). Bigeye tuna BE8706, which remained within 200 m of S FAD for 27.0 hours of the 30.0-hour track, displayed a predominant on-FAD daytime depth of between 50 and 60 m, whereas the predominant, pooled, daytime off-FAD depth of the other three bigeye tuna was 230 m. Temperature distribution For the yellowfin tuna, the surface mixed layer (ML) and first degree of the thermocline (ML-1) accounted for 68% of daytime distribution (Fig. 16 A). After sun- set, the mixed layer alone represented over 62% of the distribution of this species, with the upper 2 degrees of the thermocline contributing an additional 26%) of the yellowfin tuna's nighttime distribution (Fig. 16B). Obviously, the warmer nighttime distributions are reflective of the generally shallower depths occupied during the hours of darkness. In the case of the bigeye tuna, daytime distribution was more dependent on absolute temperature than temperature relative to the surface mixed layer. Thus, in daytime 62.5% of the pooled off-FAD distribution occurred between 14° and 17°C (Fig. 17A). On an in- dividual basis, the strong influence of temperature on daytime vertical distribution was demonstrated by fish BE8401 which consistently oscillated between the 15° and 1 7°C isotherms even though these isotherms were changing in depth over the course of the track (Fig. 14). Of this bigeye tuna's daytime movements, 68% Holland et al Movements of Thunnus albacares and 7" obesus near fish-aggregating devices 503 70 60 50-(- UJ 40 ^ 30 < I— O I— Li_ o LJ O cn UJ CL 20 10 0 A 70 60 50 40 30 20 10 + 0 B IIIIIIIII777 2 S S 2 5 2 2 2 2 2 RELATIVE TEMPERATURE INTERVAL 2 2 (°C) Figure 16 Yellowfin tuna temperature distributions relative to the upper mixed layer (ML). (A) day, (B) night. occurred between 14° and 17°C. Similarly, BE8205 spent 70% of daylight hours between the 14° and 16°C isotherms (Fig. 15), and BE8603 spent 76% of off-FAD daylight hours between 14° and 17°C isotherms (Fig. 13). As with the yellowfin, bigeye tuna occupied warmer waters at night (Fig. 17B). The difference between daytime and nighttime temperature distributions was greater for bigeye than for yellowfin tuna because of the much deeper and colder daytime distribution of the bigeye tuna and their comparatively large nocturnal upward shift. Regular, large, upward excursions were a major feature of the daytime swimming behavior of all the bigeye tuna that were tracked in off-FAD situations (Figs. 13, 14, and 15). This behavior occurred regard- less of the type of horizontal movement (e.g., meander- ing versus straight line). These excursions were com- prised of rapid ascents and descents of uniform rate. Ascents usually peaked close to the bottom of the mixed layer and the descending phases usually terminated at the predominant daytime swimming depth. These 50- 40- i ^ UJ 2 30- 1— —I 20 ^ ji^ < O 10- i^ 1— Ll 0- O ^T^, ,^ H : r"! : 1 : 1 : . , \— Z 50- UJ ^ 40^ B UJ Q_ 30- 20- 1 10- ± [VtV....^t^ . c -27 1 - -26 1 ■ -25.1 ■ -24 1 • -23 1 ■ -22 1 ■ -21 1 - -20 1 - -19 1 ■ -18 1- -17 1 ■ -16 1 • -15 1 - -14 1- -13 1- -12 1- -1 1 1 ■ -10 1 ■ 0-9 1- oor^tDin^rocNj.-ocDcOf^toin'j-mrM — — TEMPERATURE INTERVAL (°C) Figure 17 Bigeye tuna temperature distribution. (A) day, (B) night. regular excursions did not occur at night or when on- FAD. For example, these upward excursions were exhibited by fish BE8603 only after it departed S FAD and assumed a deeper off-FAD distribution (Fig. 13). The periodicity and duration of these large excursions were quite regular. For instance, 13 daytime peaks ex- hibited by BE8205 had a peak-to-peak interval of 57.7 minutes (±7.0 SD) and a duration of 12.0 minutes (±2.6 SD). Individual fish of both species tended to adopt con- sistent upper and lower limits to their movements such that, even if they were making frequent upward and downward movements, these were often terminated at consistent depths or temperatures. The most common of these turnaround points was the zone encompass- ing the bottom of the surface mixed layer and the uppermost layers of the thermocline. Thus, in the case of yellowfin YF8305, the top of the thermocline rep- resented the bottom of dives made from the surface and the starting and finishing points of dives made to deeper depths, most of which had a consistent "floor" around the 20°C isotherm at approximately 130 m (Fig. 10). Similarly, in addition to the well-defined tempera- ture confines of their smaller vertical oscillations, most 504 Fishery Bulletin 88(3), 1990 of the bigeye tunas' large upward excursions con- sistently terminated near the interface between the thermocline and the mixed layer (Figs. 13, 14, and 15). Examples of possible fly-glide behavior (Weihs 1973, Carey and 01 sen 1982) were observed in tracks of yellowfin and bigeye tuna, a very consistent example being BE8401, which exhibited "sawtooth" oscOlations with a period of approximately 4 minutes and an amplitude between 35 and 45 m for the entire 11-hour daylight segment of the track. Similar, persistent oscil- lating daytime behavior was exhibited by fish YF8506 as it moved on a direct course away from S FAD (Fig. 18). For instance, between 0930 and 1815 this fish swam on a straight southwesterly course at an ap- parently constant horizontal speed of 1.02 m/second (Fig. 5). Assuming this constant speed, the descending angles during this period averaged 6.06° (±0.7 SD, N = 11), and the climbing angles averaged 9.55° (+1.4 SD, N =11) with 100% of the chmbing angles being greater than the descending angles. Also, of the mea- surable oscillations occurring between 0915 and 1815, 39 out of 46 (85%) displayed longer falling than rising phases, suggesting active upward swimming followed by a comparatively passive downward glide. These oscillations were characterized by constant rates of as- cent and descent, with abrupt changes in direction link- ing the falling and rising phases (Fig. 18). Discussion Using two tie wraps appears to be a satisfactory way of attaching transmitters to tunas. Transmitters at- tached in this way were carried successfully by four captive fish for several weeks, and one of the tracked fish was caught in good health (and still carrying the transmitter) by a fisherman 2 weeks after we termi- nated the track. The similarity of vertical and horizon- tal movements across tracks also suggests minimal alteration of normal behavior. The association of these tuna with a daytime range, whether FAD or reef perimeter, was extremely strong. In the case of FADs, several fish spent many hours within a few meters of the mooring line. Similarly, none of the reef-associated fish made significant offshore movements during daylight hours. In fact, the along- shore movements were remarkable for their fidelity to the outer reef contour. Combining FAD-associated and reef-perimeter fish, 10 of the 15 fish tracked in this study moved within a well-defined home range during daylight hours and two other fish were lost before any diel patterns could be observed. Most of the fish made nocturnal excursions away from their respective day- time habitats. Similar, consistent diurnal behavior has been previously observed in a 44-cm skipjack tuna. ELAPSED TIME (min) 10 15 :0 25 30 35 40 SURFACE 50 E I (- a. in a 100 150 200 250 300 350 ■■ ' ' ' (Sj> ■^ ,/^ y 200 m and in temperatures between 11° and 15°C, where the dissolved oxygen is >1 mL/L. How- ever, the tracking data suggest daytime distribution (220-240 m, 14-17°C) is influenced at least as much by temperature as by depth, and none of the catch-rate studies indicate the dramatic nighttime upward shift that the tracking data reveal. The average sustained swimming speed derived from the straight sections of the current yellowfin tuna tracks was 4.46 Km/hour (range 3.2-6.5 Km/hour, Table 1), which is similar to the 4.27 Km/hour average (range 2.4-7.8 Km/hour) reported by Carey and Olson (1982) for fish having an average fork length 41% greater than those used in our study. This comparison would suggest that absolute traveling (sustainable) swimming speed for this species does not increase with size, at least for fish between 45 and 100 cm FL. This constancy of swimming speed, regardless of increas- ing fish length, is consistent with the model generated by Magnuson (1973, 1978), which predicts that in- creased weight associated with increasing length is 506 Fishery Bulletin 88(3). 1990 offset in this species by the allometric growth of the swim bladder, thereby precluding the need to increase hydrodynamic lift through higher swimming speeds. However, the sustained swimming speeds observed during these tracks (average of 1.24 m/second) are con- siderably higher than the 0.5 to 0.6 m/second predicted by Magnuson (1978) for fish of this size. Obviously, field observations of larger fish would be useful in further testing the Magnuson model. Periods of possible fly-glide behavior were observed in sections of tracks of yeliowfin and bigeye tunas. These oscillations, which in one instance (YF8506) were exhibited for the entire daylight portion of a straight azimuth track, have been hypothesized to result in a saving of energy required for locomotion between two points (Weihs 1973, Magnuson 1978). As predicted by this model, most gliding phases lasted longer than the ascending phases, and the descending and ascending legs were connected by abrupt angles, which maximize energy transfer from the gliding to the flying phase. Thus, using the "tly" angle (fl) of 9.55 and "glide" angle (a) of 6.06 obtained from YF8506, and a swim/glide drag ratio (k) of 1.2 (Magnuson 1978), the equation: Energy saving = 1 tan a sin ft + (tan a ■ cos ft) 1 + 1 sin p\ k sin a results in an energy saving of 9.4% compared with level swimming over the same distance. And, where T is the increased time to travel the same distance using a fly/glide strategy as opposed to level swimming, using the equation sin o -I- sin ft sin {a + P) indicates that this strategy results in only a 0.9% in- crease in time to travel the same distance. The extremely regular, large, upward excursions made by all the off-FAD bigeye tuna during daylight hours may represent behavioral thermoregulation. At the low (14-17°C) ambient temperatures adopted by these fish, their core temperatures possibly drop below some threshold level which requires movement into warmer water to regain optimum body temperatures. If this is the case, it would suggest that there exists a strong motivation for inhabiting the deep cold layers observed during these tracks. A possible motivation would be the opportunity to feed on deepwater fish, squid, and crustaceans which the bigeye tuna then follow into shallower depths at night when these prey organisms migrate toward the surface. Monitoring the core muscle temperatures of bigeye tuna would indicate if, in fact, the large upward excursions are a form of behavioral thermoregulation which, when combined with physiological thermoconservation, allows these fish to exploit an otherwise unreachable resource. Acknowledgments This work was supported by the University of Hawaii Sea Grant College Program (Ultrasonic Telemetry of Horizontal and Vertical Movements of Pelagic Fish Associated with FADs project, MR/R-25) under Institu- tional Grant No. NA85AA-D-SG082 from the NOAA Office of Sea Grant, Department of Commerce; the National Marine Fisheries Service, NOAA; State of Hawaii Department of Planning and Economic Devel- opment; and the Federation of Japan Tuna Fisher- men's Cooperative Association. The help of Lance Asagi, Robert Bourke, Zig Ching, Scott Ferguson, Jeff Koch, and Ruben Yost is also gratefully acknowledged. Citations Bayliff. W.H., and K.N. Holland 1986 Materials and methods fur tagging tuna and billfishes, reciivering the tags, and handling the recapture data. FAG Fish. Tech. Pap. 279, 36 p. Brock. R.E. 1985 F'reliminary study of the feeding habits of pelagic fish around Hawaiian tish aggi-egation devices, or can fish aggrega- tion devices enhance local fish productivity? Bull. Mar. Sci. 37:40-49. Carey, F.G. 1983 Experiments with free swimming fish, /n Brewer, P.G. (ed.). Oceanography, the present and future, p. 58-68. Springer- Verlag, NY. Carey, F.G., and R.J. Olson 1982 Sonic tracking experiments with tunas. ICCAT Collec- tive Volume of Scientific Papers XVII. 2:458-466. Int. Comm. Conserv. Atl. Tuna. Spain. Fedoryako, B.I. 1982 Langmuir circulation as a possible mechanism of forma- tion of fish association around a floating object. Oceanology 22:228-232. Gooding, R.M., and J.J. Magnuson 1967 Ec(j|ogical significance of a drifting object to pelagic fishes. Pac. Sci. 21:486-497. Hanamoto, E. 1976 The swimming layer of bigeye tuna. Bull. Jpn. Soc. Fish. Oceanogr. 29:41-44. [Engl. Transl. No. 21 by T. Otsu, 1977, 7 p.; avail. Honolulu Lab., Southwest Fish. Cent., Natl. Mar. Fish. Serv.. NOAA, Honolulu, HI 96822-2396.] 1987 Effect of oceanographic environment on bigeye tuna distribution. Bull. ,Jpn. Soc. Fish. Oceanogr. 51:20.3-216. Holland et al Movements of Thunnus albscares and T obesus near fish-aggregating devices 507 Hida. T.S. Yuen, H.S.H. 1973 Food of tunas and dolphins (Pisces: Scombridae and Cor- 1970 Behavior of skipjack tuna, Katsuwomai pelamvi, as deter- phaenidae) with emphasis on the distribution and biology of mined by tracking with ultrasonic devices. J. Fish. Res. Board then prey Stolephoms l^urcaneerHEngrauMae). Fish. Bull., Can. 27:2071-2079. U.S. 71:135-143. Holland, K., R. Brill, S. Ferguson, R. Chang, and R. Yost 1985 A small vessel technique for tracking pelagic fish. Mar. Fish. Rev. 47(4):26-.32. Hunter, J.R., and C.T. Mitchell 1967 Field experiments on the attraction of fish to floating objects. J. Cons. Perm. Int. Explor. Mer 31:427-434. Hunter, J.R., A.W. Argue, W.H. Bayliff, A.E. Dizon. A. Fonteneau, D. Goodman, and G.R. Seckel 1986 The dynamics of tuna movements: an evaluation of past and future "research. FAO Fish. Tech. Pap. 277, 78 p. Kihara, Y. 1981 Fishery based on the payao method in the Philippines. Suisan Sekai 30:78-84. [Engl. Transl. No. 76 by T. Otsu, 1982, 12 p.; avail. Honolulu Lab.. Southwest Fish. Cent., Natl. Mar. Fish. Serv., NOAA. Honolulu, HI 96822-2396.] King, J.E., and I.I. Ikehara 1956 Comparative study of the food of bigeye and yellowfin tuna in the central Pacific. Fish. Bull., U.S. 57:61-85. Laurs, R.M., H.S.H. Yuen, and J.H. Johnson 1977 Small scale movements of albacore, Thunnus alalunga, in relation to ocean features as indicated by ultrasonic track- ing and oceanographic sampling. Fish. Bull., U.S. 75:347-355. Magnuson, J.J. 1973 Comparative study of adaptations for continuous swim- ming and hydrostatic equilibrium of scombroid and xiphoid fishes. Fish. Bull., U.S. 71:337-356. 1978 Locomotion by scombroid fishes: Hydromechanics, mor- phology, and behavior. In Hoar, H.S., and D.J. Randall (eds.). Fish physiology, vol. VII, p. 240-313. Academic Press, NY. Matsumoto, W.M., T.K. Kazama, and D.C. Aasted 1981 Anchored fish aggregation devices in Hawaiian waters. Mar. Fish. Rev. 43(9): 1-13. Reintjes, J.W., and J.E. King 1953 Foodof yellowfin tuna in the central Pacific. Fish. Bull.. U.S. 54:91-110. SAS 1985 SAS/STAT guide for personal computer, version 6 edition. SAS Institute Inc., Cary, NC. Saito. S. 1975 On the depth of capture of bigeye tuna by further im- proved vertical longline in the tropical Pacific. Bull. ,lpn. Soc. Sci. Fish. 41:831-841. Shomura, R.S., and W.M. Matsumoto 1982 Structured flotsam as fish aggregation devices. NOAA Tech. Memo. NOAA-TM-NMFS-SWFC-22, Honolulu Lab., Southwest Fish. Cent., Natl. Mar. Fish. Serv.. Honolulu, HI 96822, 9 p. Walker, M.M. 1984 Learned magnetic field discrimination in yellowfin tuna, Thunnus alhacar-es. J. Comp. F'hysiol. A 155:673-679. Weihs, D. 1973 Mechanically efficient swimming techniques for fish with negative bouyancy. J. Mar. Res. 31:194-209. Yonemori, T. 1982 Study of tuna behavior, particularly their swimming depths, by the use of sonic tags. Far Seas Fish. Res. Lab. (ShimizulNewsletter 44:1-5. [Engl. Transl. No. 70 by T. Otsu, 1982, 7 p.; avail. Honolulu Lab., Southwest Fish. Cent., Natl. Mar. Fish. Serv. NOAA, Honolulu. HI 96822-2396.] Abstract.- We studied the food habits of the California sea lion Zalo- phus califoniianus at San Clemente Island, California, from September 1981 through September 1986 using fish otoliths, cephalopod beaks, and other prey remains we recovered from 1476 fecal samples (i.e., scats). We identified 44 types of prey to species and 8 types of prey to genus. Seven types of prey occurred in at least 10% of scats: northern anchovy En- graulis mordax (51.3%); jack mack- erel Tnifhurus symmetriais (24.6%); pelagic red crab Pleuroncodes plani- pes (21.2%); Pacific whiting AfeWwc- cius productus (19.6%); rockfishes, Sehastes spp. (19.0%); market squid Loligo opalescens (16.3%); and black- smith Chromis puwtipinnis (10.7%). We examined trends in the occur- rence of these seven species and of Pacific mackerel Scomber japo7iicus and octopus (Octopus spp.) in scats of California sea lions. We found sig- nificant differences among seasons and years, and season-year differ- ences in the occurrence of northern anchovy, jack mackerel. Pacific whit- ing, and octopus (P< 0.01). Significant effects due to year and season-year interaction were found for pelagic red crab, rockfish, market squid, and blacksmith. Pacific mackerel showed a significant difference only among years. Significant differences in the occurrence of Pacific whiting, north- ern anchovy, jack mackerel. Pacific mackerel, market squid, octopus, and pelagic red crab (P<0.01) were found during pre-El Nifio, El Nifio, and post-El Nino periods. Food Habits of California Sea Lions Zaiophus californianus at San Clemente Island, California, 1981-86 Mark S. Lowry Charles W. Oliver Carolyn Macky Southwest Fisheries Science Center, National Marine Fisheries Service. NOAA PO Box 271. La Jolla. California 92038 Jeannie B. Wexler Southwest Fisheries Science Center, National Marine Fisheries Service, NOAA PO Box 271, La Jolla, California 92038 Present address: Inter-American Tropical Tuna Commission c/o Scripps Institution of Oceanography, La Jolla, California 92037 Manuscript accepted 10 April 1990. Fishery Bulletin, U.S. 88:509-521. Knowledge of California sea lion Za- lophus californianus food habits in the Southern California Bight (SCB, Fig. 1) has come from stomach con- tent analysis by Scheffer and Neff (1948), and Fiscus and Baines (1966); and from the 1978-79 spring and summer study at San Miguel Island from scat (i.e., fecal) analysis by An- tonelis et al. (1984). These studies identified a wide variety of prey. An- tonelis et al. (1984) reported that Pacific whiting, market squid, juve- nile rockfish, and northern anchovy were the four most important prey at San Miguel Island. A 4-year study at the Farallon Islands in northern Cali- fornia (outside the breeding area) found that sea lions switched be- tween Pacific whiting and juvenile rockfish (Bailey and Ainley 1982). From 1981 to 1986, we examined prey consumed by California sea lions using the shoreline located at and south of Mail Point, San Clemente Island (SCI, Fig. 1). San Clemente Island has one of the smallest rook- eries in the SCB (producing approx- imately 666 pups in 1981) as com- pared with San Miguel and San Nicolas islands (producing approx- imately 8255 and 6704 pups, respec- tively, in 1981; DeMaster et al. 1982, Stewart and Yochem 1984, Oliver and Lowry 1987). We analyzed fecal samples (scats) to identify prey com- position and temporal changes in the diet. This period was characterized by an abnormal influx of warm water into the SCB during the 1982-83 El Nino/Southern Oscillation (McGowan 1984). Effects of this El Nino within the SCB became evident in October 1982, when sea-surface temperatures were 1-2°C above normal, and zoo- plankton levels declined (McGowan 1984). Water temperatures returned to normal in October-November 1984 (F. Miller, Inter-Am. Trop. Tuna Comm., La Jolla, CA 92038, pers. commun., Oct. 1986). This warm- water period off California has been referred to as the California El Nino (McGowan 1984, Fiedler etal. 1986). Methods Sample collection We collected fresh and dry scats dur- ing 40 trips to SCI with between-trip intervals ranging from 2 weeks to 3 months. Scats were separated into one of three categories: (1) fresh 509 510 Fishery Bulletin 88(3), 1990 121°W 120"W iirw iis-w in-w 34°N 33°N 32"N San Miguel Is Santa Rosa Is I. OS Angeles San Nicolas Is. SAN CLEMENTE IS. an Diego Unil.eJ.!!^'-''--- Mexico 121°W 120''W 119"W 1 1 8"W Figure 1 Map showing Southern California Bight (SCB) and San Cleniente Island (SCI). scats, (2) dried scats representing one between-trip in- terval, and (3) dried scats representing more than one between-trip interval. Fresh scat samples represented the diet of sea lions within 3 to 4 days of the collec- tion. The second category provided dietary informa- tion from the previous scat collection period to the time represented by fresh scats. The third category repre- sented a much longer period. Each scat was placed in a plastic bag, stored in an air-tight plastic container, and frozen until processed. Scats were soaked in a mild soap solution for 1-5 days, and then rinsed with water through two or three nested sieves. During September 1981-March 1983 we used three sieves (mesh sizes 2.8 mm, 1.5 mm, and 1.00 mm) and during April 1983-September 1986 we used two sieves (mesh sizes 2.8 mm and 0.710 mm). After rins- ing, we collected fish otoliths, cephalopod beaks, shark teeth, and cartilaginous vertebrae. Otoliths, shark teeth, and cartilaginous vertebrae were air dried and stored in gelatin capsules; cephalopod beaks were stored in vials containing 70% ethanol. We noted the presence or absence of fragments of pelagic red crabs Pleuroncodes planipes. Sample analysis Prey items were identified to the lowest taxon possible from comparative specimens and drawings (Iverson and Pinkas 1971, Eschmeyer et al. 1983). We used sagittal otoliths to identify teleost fishes, cephalopod beaks to identify squid and octopus, teeth to identify elasmobranchs, and exoskeletal fragments to identify pelagic red crabs. We counted left and right sagittal otoliths, and upper and lower cephalopod beaks. Lowry et al Food habits of Zaiophus cahfornianus at San Clemente Island. California 511 Table 1 California sea lion Zalophus califomianus scat samples collected seasonally from San Clemente Island, California, 1981-86. Number of scats collected Dry Number of collection 1 between-trip > 1 between-trip Total Year Season trips Fresh interval interval collected 1981 Autumn 3 22 19 7 48 1981-82 Winter 1 0 3 0 3 1982 Spring 4 27 80 52 159 1982 Summer 4 76 67 7 150 1982 Autumn 1 1 53 0 54 1982-83 Winter 2 3 56 0 59 1983 Spring 3 37 73 60 170 1983 Summer 2 6 71 10 87 1983 Autumn 1 22 56 17 95 1983-84 Winter 3 93 51 4 148 1984 Spring 2 81 20 16 117 1984 Summer 3 88 1 0 89 1984 Autumn 2 13 40 0 53 1984-85 Winter 1 0 16 0 16 1985 Spring 1 11 1 0 12 1985 Summer 2 60 0 0 60 1985 Autumn 1 25 0 0 25 1986 Spring 1 44 0 0 44 1986 Summer 2 65 0 0 65 1986 Autumn 1 22 0 0 2^ Total 40 696 607 173 1476 We used indices of occurrence, composition, and number to quantify prey taxon consumed by sea lions. Only scat samples containing otoliths, cephalopod beaks, shark teeth, cartilaginous vertebrae, or pelagic red crab fragments were used in the analysis. The occurrence index, or percent occurrence (PO), is a measure of the percentage of scat samples in which a prey taxon occurred. The composition index, or per- cent prey composition (PPC), is a measure of the percentage of occurrences for each prey taxon from a tally of occurrences from all prey taxa found in a group of samples. The number index, or percent mini- mum number (PMN), measures the percentage of numbers of individual prey taxon. To compute the number index, we used the maximum count of either left or right otoliths, or upper or lower beaks, which represent the minimum number of individual prey taxon. The counts for all prey taxa are summed and the percentage of each prey taxon is determined from the sum of all counts. The number index is, then, the percentage represented by each prey taxon from the sum of all maximum coimts taken of all prey taxa found in a group of samples. Interpretation of the number index is limited because it excludes (1) all prey hard parts found that were broken or digested, and not categorized as either left or right otoliths or upper or lower beaks, and (2) prey that lacked otoliths or beaks (e.g., pelagic red crabs and sharks). However, we think that this index is useful because fish and cephalopods represented the major- ity of prey in scat samples. Percent occurrence indicates prey consumption with- out regard to other prey, and may indicate temporal availability, selectivity, or ease of capture of individual prey. Percent prey composition indicates the relative proportions of prey consumed. Percent minimum num- ber of prey indicates numbers of prey consumed and is an index of the rate of consumption of individual prey. Statistical analysis We used all scats to determine what prey taxa were consumed, and then limited our examination of tem- poral variations in prey consumption to prey that oc- curred in at least 10% of these scats, were found in at least 80% of the seasons, or were occasionally very abundant. Only fresh scats and dried scats represent- ing one between-trip interval were used in the temporal analyses of sea lion diets. We combined the 40 collec- tion trips (Table 1) by season into winter (December- 512 Fishery Bulletin 88(3), 1990 Table 2 Number of prey occurrences n Occur), percent occurrence (PO), percent prey composition (PPC), minimum number of in dividual prey (n Ind.), and percent minimum number (PMN) of prey found in 1309 California sea lion scats collected at San Clemente Island, California, | 1981-86. Prey Occur PO PPC n Ind. PMN Scientific name Common name n Engraulis mordax northern anchovy 672 51.3 24.7 9,398 54.4 Trachurus symmetT-icus jack mackerel 322 24.6 11.8 1,220 7.1 Pleuroncodes planipes' pelagic red crab 277 21.2 10.2 * * — Merluccius productus Pacific whiting 256 19.6 9.4 2,110 12.2 Sebastes spp. rockfish 249 19.0 9.2 1,245 7.2 Loligo opakscens market squid 214 16.3 7.9 1,516 8.8 Chromis punctipinnis * blacksmith 140 10.7 5.1 580 3.4 Scomber japonicus Pacific mackerel 125 9.5 4.6 215 1.2 Octopus spp. octopus 67 5.1 2.5 523 3.0 Onychoteuthis horeaiijaponicus squid 39 2.2 1.1 33 <0.5 Lyopsetta exiiis slender sole 11 0.8 <0.5 19 <0.5 Oxyjulis califomica* senorita 14 1.1 0.5 20 <0.5 Porichthys notatus plainfin midshipman 9 0.7 <0.5 11 <0.5 Ahraliopsis spp.' squid 9 0.7 <0.5 11 <0.5 Gonatus spp. squid 9 0.7 <0.5 11 <0.5 Lycodes cortezianus* bigfin eelpout 7 0.5 <0.5 8 <0.5 Sebastolobus alascanus* shortspine thornyhead 6 0.5 <0.5 7 <0.5 Microstomus pacificus dover sole 6 0.5 <0.5 8 <0.5 Glyptocephalus zachirus rex sole 6 0.5 <0.5 7 <0.5 Icichthys lockingtoni medusafish 5 <0.5 <0.5 7 <0.5 Xeiieretmus ritteri * stripefm poacher 3 <0.5 <0,5 5 <0.5 Citharichthys snrdidus Pacific sanddab 4 <0.5 <0.5 5 <0.5 Atherinops ajfinis* topsmelt 3 <0.5 <0.5 4 <0.5 Scmicossyphus pulcher* sheephead 3 <0.5 <0.5 3 <0.5 Seriphus polittis queenfish 2 <0.5 <0.5 26 <0.5 Cypselurus caiifoniicus ' California flying fish 2 <0.5 <0.5 8 <0.5 Genyonemus lineatus white croaker 2 <0.5 <0.5 6 <0.5 Galeorhinus zyopteruti * soupfin shark 2 <0.5 <0.5 — — Sym bolophorus califomiensL California lanternfish 2 <0.5 <0.5 2 <0.5 Paralabrax clathratus* kelp bass 2 <0.5 <0.5 2 <0.5 Chilara taylori spotted cusk-eel 2 <0.5 <0.5 2 <0.5 Cymmatogaster aggregata shiner surfperch 2 <0.5 <0.5 2 <0.5 Stenobrachius leucopsarns northern lampfish 2 <0.5 <0.5 2 <0.5 Paralabra-i sp.* seabass <0.5 <0.5 3 <0.5 Ocythoe tubereulata * squid <0.5 <0.5 2 <0.5 Girella nigricans opaleye <0.5 <0.5 2 <0.5 Zatembius rosaceus pink surfperch <0.5 <0.5 1 <0.5 Zaniolepis sp.* combfish <0.5 <0.5 1 <0.5 Xenerelmus sp.* poacher <0.5 <0.5 1 <0.5 Prionace glauca * blue shark <0.5 <0.5 — — A rgentina siaiis * Pacific argentine <0.5 <0.5 1 <0.5 Hippoglossina stomata * bigmouth sole <0.5 <0.5 1 <0.5 Parophrys vetuiua english sole <0.5 <0.5 1 <0.5 Xeneretmus triacnnlhtis' bluespotted poacher <0.5 <0.5 1 <0.5 Ceratoscopelus townsendi * dogtooth lampfish <0.5 <0.5 4 <0.5 Citharichthys sp. sanddab <0.5 <0.5 1 <0.5 Careproctus meianu nis * blacktail snailfish <0.5 <0.5 1 <0.5 Medialuna califomiensis * halfmoon <0.5 <0.5 3 <0.5 Synodns lueioceps* California lizardfish <0.5 <0.5 1 <0.5 Octopoteuthis dektron * squid <0.5 <0.5 1 <0.5 Unid. Myctophidae lanternfish 17 1.3 0.6 29 <0.5 Unid. Embiotocidae surfperch <0.5 <0.5 7 <0.5 Unid. Atherinidae silverside 3 <0.5 <0.5 3 <0.5 Unid. Labridae wrasse 3 <0,5 <0.5 8 <0.5 Unid. Cottidae sculpin 2 <0.5 <0.5 2 <0.5 Unid. Zoarcidae eelpout 2 <0.5 <0.5 2 <0.5 Lowry et al Food habits of Zaiophus cahfornianus at San Clemente Island, California 513 Table 2 (continued) Prey Scientific name Common name Unid. Gonatidae Unid. Scombridae squid mackerel or tuna Unid. flatfish Unid. cephalopod Unid. cartilaginous fish Unid. fishes 'Not previously reported as prey of California sea lions. * * Data impossible to compute in fields marked with a dash I : Occur PO PPC 1 1 5 31 2 83 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 2.4 1.1 <0.5 <0.5 6.3 3.1 n Ind. 1 2 6 19 62 PMN <0.5 <0.5 <0.5 <0.5 <0.5 February), spring (March-May), summer (June- Aug- ust), and autumn (September-November). We also categorized our sample into three groups related to the California El Niiio: (1) pre-El Nino (September 1981- August 1982), (2) El Nino (November 1982-September 1984), and (3) post-El Nino (November 1984-September 1986). We tested for seasonal and yearly differences in presence or absence of each prey taxon using two-way ANOVA (Program YD of BMDP-87; Dixon 1985) in the binomial format. Two data sets were identified with full cells in order to meet the requirements of the test: (1) all seasons for the years 1982-85 (called four-season- four-year data set); and (2) three seasons (spring, sum- mer, and autumn) for 1982-86 (called three-season-five- year data set). We identified prey taxa that showed significant season, year, and season-year interaction dif- ferences in their occurrence through time using Browne- Forsythe analysis of variance for unequal variances. For the pre-El Nino, El Nino, and post-El Nino comparisons, we tested for differences in presence or absence of the same nine prey taxa using two-way Browne-Forsythe ANOVA (Dixon 1985) in the binomial format. We also examined temporal differences in the relative number of different prey taxa found per scat (1,2,3,4, 5, or >5). Pearson's chi-square test (Dixon 1985) was used to test for differences among numbers of prey occurring per scat. Results Identification of prey remains In the 1476 scats collected (Table 1), we found the follow- ing: (1) 1455 scats (98.5%) contained fish remains (i.e., bones, otoliths, or eye lenses); (2) 525 scats (35.5%) con- tained cephalopod remains (i.e. , beaks or eye lenses); (3) 1309 scats (88.6%) contained fish otoliths, cephalopod beaks, or other prey remains that were used to identify prey; (4) 1271 scats (86.1%) contained fish otoliths; and (5) 344 scats (23.3%) contained cephalopod beaks. We identified 95.4% of 2647 prey occurrences to species, genus, or family (Table 2). Of the 2,647 prey occurrences, 75.2% were bony fishes, 14.1% were cephalopod, 10.5% were Crustacea (99.3% of Crustacea were pelagic red crab), and 0.2% were cartilaginous fish. We identified 44 prey taxa to species level and 8 to genus level. Sevenprey taxa occurred in > 10% of scats: north- ern anchovy Engraulis mordax (51.3%); jack mackerel Trachurus symmetricus (24.6%); pelagic red crahPleu- roncodes planipes (21.2%); Pacific whiting Merhiccius productus (19.6%); rockfishes, Sebastes spp. (19.0%); market squid Loligo opalescens {16.3%); and blacksmith Chromis punctipinnis (10.7%). Two prey taxa occurred in 5.0-10.0% of the scats: Pacific mackerel Scomber japonicus (9.5%) and octopus. Octopus spp. (5.1%). We identified 17 species and two genera of fish, and two species and one genus of cephalopod not reported previously as prey of California sea lions (Table 2). Nearly half of the scats (46.1%) contained only one prey taxa (Fig. 2). Northern anchovy constituted 56.2% of 603 scat samples with single prey taxa (Fig. 3). Variability in diet through time We found differences in distributions of occurrence (occurrence and composition indices) and numbers of individuals (number index) for nine prey taxa examined for temporal variability during the pre-El Nino. El Nino, and post-El Nino periods (Fig. 4). The distributions of occurrence were significantly different for Pacific whit- ing, northern anchovy, jack mackerel, Pacific mackerel, market squid, octopus and pelagic red crab (P<0.01). Northern anchovy occurred frequently and in the highest numbers of any prey taxa in all three periods. During the pre-El Niiio period, jack mackerel occurred in the scat samples more often than the other prey; and jack mackerel and market squid were found in greater indi- 514 Fisher/ Bulletin 88(3). 1990 46.1% NUMBER OF PREY TAXA PER SCAT 25.1% 15.6% 6.5% 3.7% 3.0% >5 NORTH, ANCHOVY JACK MACKEREL PELAGIC RED CRAB PACIFIC WHITING ROCKFISH MARKET SQUID BLACKSMITH PAC. MACKEREL OCTOPUS OTHERS 20 30 40 PERCENT Figure 3 Percent prey found in 603 California sea lion scats with single prey taxon. Figure 2 Occurrence of single and multiple prey taxa in 1309 California sea lion scats collected at San Clemente Island, California. Figure 4 Occurrence (PO), composition (PPC), and number (PMN) indices of 9 prey taxa found in 1309 California sea lion scats collected at San Clemente Island, California, 1981-86, which are divided into pre-El Nino (September 1981-August 1982), El Nino (November 1982- September 1984), and post-El Nino (November 1984-September 1986) periods. The number index (PMN) could not be determined from fragments of pelagic red crab. NORTH. ANCIIOV'^- O JACK MACK-EREL ^S PELAGIC RED CRAB ^ PACIFIC WHITING E^ ROCKFISH 1 MARK'ET SQUID K BUCKSMITH ^ PAC. MACK'EREL OCTOPUS ^m (PO) NORTH. ANCHOVY JACK MACKEREL PELAGIC RED CRAB PACIFIC WHITING ROCKFISH MARKET SQUID BLACKSMITH PAC. MACKEREL OCTOPUS OTHERS ^^_ (PPC) ^^^ (PMN) 0 ( ) 10 2030405060 70 809 PERCENT 0 3 5 10 15 2025303640456 PERCENT 0 0 10 20 30 40 50 60 70 8 PERCENT NORTH. ANCHOVY' JACK MACKEREL PELAGIC RED CR.\B ;g PACIFIC WHITING 2 ROCKFISH _i MARKET SQUID " BLACKSMITH PAC. MACKEREL OCTOPUS ■■ (PO) NORTH. ANCHOVY JACK MACKEREL PELAGIC RED CRAB PACIFIC WHITING ROCKFISH MARKET SQUID BIACKSMITH PAC. MACKEREL OCTOPUS OTHERS 2 (PPC) I ■ " (I'MN) ■ 1 1 1 D 10 2030405080 70 809 PERCENT 0 3 5 10 15 2025303640455 PERCENT 0 D 10 20 30 40 50 80 70 8 I'KhMENT 0 NORTH. ANCHOVT Q JACK MACKEREL IZ PEUGIC RED CRAB Z PACIFIC WHITING -1 ROCKFISH t MARKET SQUID i7 BLACKSMITH d PAC. MACKEREL '^ OCTOPUS NORTH. ANCHOVY JACK MACKEKEI, PELAGIC RED CRAH PACIFIC WIUTLNG ROCKFISH MARKET SQITD liLACKSMnil PAC. MACKEIvKl, OCTOPUS OTHERS ^1 (PO) ■ 1 1 ■ ■ 1 1 ^ (PMN) 1 1 c ) 10 203040508070809 PERCENT 0 C 6 10 16 20 25 30 35 40 46 5 PEKCFNT 0 3 10 20 30 40 50 60 70 a I'FI'CFNT 0 Lowry et a I Food habits of Zaiophus californianus at San Clemente Island. California 515 O 111 Q. 1981 1982 El Nino 1983 1984 1985 1986 1981 1983 ' 1984 El Nino Figure 5 Seasonal percent occurrence (PO) of 9 prey taxa found in 1304 California sea lion scats. 1981-86. No samples were collected in winter 1986. vidual numbers than other prey taxa, except northern anchovy. During the El Nino period, pelagic red crab were the second most frequently occurring prey in scat samples. During the post-El Nino period, the diet was dominated by northern anchovy. We found yearly and seasonal variability in the rela- tive proportions represented by the three indices for 9 prey taxa examined (Figs. 5-7). Northern anchovy was consistently found in scats throughout the study at varying occurrence, composition, and number index levels. However, during summer and autumn 1984 we noted a drop in the three indices compared with 1982, 1983, 1985, and 1986; and it did not occur as often in 1982 as in 1983, 1985, and 1986. Jack mackerel was the most frequent prey during 1981-82, but after that period it was found infrequently, represented a very small proportion of the diet, and few individuals were found in the scats. An increase in the occurrence of 516 Fishery Bulletin 88(3). 1990 UJ O oc LU Q. 1981 Figure 6 Seasonal percent prey eomp5 Prey Taxa = (= 5.E5IS Q.E5 0.05), therefore the data were pooled and one relationship established: W = 0.03TL-''-*. Asymptotic weight (W^) was 3641.6 g. Condition factor Monthly variation in condition factor for males and females is shown in Figures 1 and 2, respectively. The condition factor was lowest in August for both sexes while it was highest in April for males and in May for females. 1 15 r i O i "0 Ll_ O 5 105 27 51 30 36 26 '\ / \ 1 00 - \ ( ■*■ 095 1 1 1 1 1 1 1 M A M J J A S 0 ^ D Month Figure 1 Mean monthly condition factor (±95% CL) of male croaker in Trinidad waters. Number of fish is given for each month. 125 - 35 24 1 20 ; \ '"' / \ .. 3 ^'/ \ 1 15 85 / \ ^ y\ \ 1 46 C' \ 1 \ >-' 1 10 „ J \ 1 \ U- \ \ 1 1 \ \ 1 o \ \ . »_ '■ \ 1 \ ■ 1 X3 \ \ 1 J Cl ° 105 - \ \ / \ I / V / 100 - \ 0 95 1 1 1 1 1 M A ^ Month \ 5 0 r J D Figure 2 Mean monthly condition factor ( + 95% CL) of female croaker in Trinidad waters. Number of fish is given for each month. Age and growth Mean marginal increments were lowest from May to August, indicating that annulus formation occurred during this period (Fig. 3). Fish length was directly pro- portional to, and highly correlated with, otolith radius (R): TL = 12.0-h32.7R, (?•- = 0.94). Otolith reading showed the presence of six age groups for males and seven for females, while analysis of length-frequency 526 Fishery Bulletin 88(3). 1990 Figure 3 Mean monthly marginal increment ( ± 95% CL) of otoliths of the croaker in Trinidad waters. Number of fish is given for each month. distribution of combined sexes showed six age groups (Fig. 4). There was close correspondence in lengths at respective ages found by the two methods (Table 3). Based on the above, use of otoliths for age determina- tion of the croaker is considered to be valid. A pre- liminary plot of the observed length at age revealed that growth of the croaker could be adequately described by the von Bertalanffy equation. Growth parameters showed that females achieved a greater asymptotic length, but grew at a slower rate than males. In Table 4 these parameters are compared with those found for different populations of croaker in Brazil. Mortality Total mortality rate (Z) which was the slope of the descending part of the length-converted catch curve in Figure 5 was 1.2/year (r- = 0.96, P<0.05). The 95% confidence limits for Z were 0.91 and 1.49. Natural mortality rate (M) was 0.4/year, and fishing mortality rate (F) was 0.8/year. Yield per recruit At the present level of F and t,., Y/R is already at the maximum of 175 g (Fig. 6). Increasing F results in a decrease in Y/R. Increasing t,. to 3 years results in a Y/R of 181 g at the same level of F. At values of M of 0.1, 0.2, and 0.3 the croaker is overexploited at the corresponding levels of F (Fig. 7), while at M of 0.4 and 0.5 it is fully exploited. At low M values the yield curves are domed with peaks at low F values, while at higher M values they are relatively flat. Discussion Several similarities exist in the reproductive biology of the whitemouth croaker throughout its geographical range. Early maturity and year-round spawning have 60 " - 513 55 •"•» 50 45 - ^ ■^ E o " 35 - "^ ^^^ 30 - -— ■ 'H -^ 25 !0 - •A — -—- ' ^ • ' 001 1 10 30 60 90 99 9999 Cum ufotive % Frequency Figure 4 ( 'umulative ]X'rcent;ige length- frequency curve (dotted line) and separation into six age groups (straight lines) for the croaker in Trinidad waters. Arrows indicate inflexion points chosen by eye. Manickchand-Heileman and Kenny Micropogon/as furnien in Trinidad waters 527 Table 3 Mean length at age in Trinidad waters determined from otolith sections and length-frequency distribution of the croaker Otoliths (TL ±SD) Length-frequency distribution (±SD) Age group Male N Female N I 22.0+ 1.92 26 21.6 + 2.09 23 22.3 ±0.35 11 28.3+ 1.85 21 29.4 + 2.71 34 27.3+ 1.65 III 33.6 ±1.49 14 34.2 + 1.89 18 31.6+ 1.65 IV 38.5 + 2.36 29 40.5 + 1.98 30 36.2 ±2.05 V 42.5 + 1.83 10 47.5 ±2.94 12 43.5 + 2.45 VI 45.6 ±2.79 10 51.6 + 2.90 14 49.5 + 1.70 VII — — — 53.5 ± 3.25 11 — — Table 4 von Bertalanffy growth parameters of the croaker in Trinidad and Brazil waters. Area Sex L oo K k Structure Author Brazil (33-29''S) F 69.33 0.149 -2.79 scales Vazzoler 1971 M 89.57 0.076 -4.64 Brazil (29-23°S) F 60.10 0.219 -2.08 scales Vazzoler 1971 M 82.90 0.106 -2.97 Brazil (4°S) F 67.60 0.18 -0.42 scales Rodrigues 1968 M 68.60 0.18 -0.52 Trinidad F 82.90 0.13 -0.13 otoliths This study M 65.30 0.16 -0.16 70 - xxx 60 X \ /I -- 1 0 7 e 50 X X \ \ ^l« 4 0 - \ O X \ ® \ ® 30 - \ 20 - \ ® 10 - ® \ 0 1 1 1 2 4 6 8 10 Reloliv » Age ( Yeors ) Figure 5 Length-converted catch curve of the croaker in Trinidad waters. (®) points used to find Z; (x) ascending part of curve, not used in analysis. 200 r l,= 3 0 ^-^^ in 1; . 10 y'/^'~~ ^^^'^^ 5 150 - /7/^ lj= 20 a. / 1; 100 / 50 / 0 ' p... 1 1 10 20 30 Fishing Mortolil Figure 6 Yield per recruit (g) of croaker in Trinidad waters against fishing mortality for age at first capture of 2, 3, and 4 years. 528 Fishery Bulletin 88(3). 1990 600 500 .[\ o* \ t, 400 q: - \ a. \v Z 300 - ^~--„^ ^^~.,„^^ >- / ^^^^^ M:0 1 200 100 0 " M^ 0 2 M = 0 4 /" ~~~ / ^^ t^ , iO 20 30 Fishing Morlolit Figure 7 Yield per recruit (g) of croaker in Trinidad waters against fishing mortality for natural mortality rates of 0.1. 0.2, 0.3. 0.4. and 0.5 per year. also been recorded in Brazil (Vazzoler 1962, 1971; Castello 1982; Isaac-Nahum and Vazzoler 1983), and in Guyana (Lowe-McConnell 1966). Other workers have been successful in determining the age of tropical and subtropical sciaenids using otoliths (e.g., Bayagbona 1969, Villyamar 1972, Pan- nella 1974, Barger 1985). Annulus formation coincides with both the period of peak spawning and the wet season when salinity decreases due to river run-off. A similar pattern occurs in other species of tropical sciae- nids (e.g., Rao 1966, Le Guen 1971). The annulus in immature croaker is probably formed in the wet season of the year following their year of birth. Use of the Cassie method had two disadvantags: the subjectivity in choosing the inflexion points and the crowding of older age groups into the upper part of the cumulative frequency curve. However, the close agreement of results obtained by both methods validates the use of otoliths for age determination of the croaker. The higher value of A' for males than for females sug- gests a higher mortality rate for the former (Beverton and Holt 1959). This could also account for the presence of only six age groups for males and the predominance of females in the population. The variation in growth between Trinidad and Brazil indicates that the croaker does not present a uniform population in this region. According to Vazzoler (1971) geographical differences in growth characteristics of this species are due to genotypical and ecological fac- tors such as salinity and temperature. Estimation of mortality rates from length-converted catch curves is subject to several assumptions, the most important of which is that of constant recruitment (Ricker 1975). Examination of length-frequency distri- butions indicated that the modal length in the catch was similar for all years of sampling (33.0-34.9 cm), and no particularly strong or weak year class was apparent. Another assumption is that of constant fishing effort, which did not change appreciably during the period of sampling, as indicated by the number of active vessels. Since only fully recruited age groups were used in this analysis, Z was assumed to be constant over these ages. Trawling was considered to be a random method of sampling for older, fully recruited age groups, and the sample was assumed to be representative of that part of the population which was used to compute Z. Vio- lation of these assumptions often results in non-linear right limbs of catch curves (Ricker 1975). However, the linear right limb of the catch curve in this study and the significant regression suggest that the estimate of Z is acceptable. All of the conditions, except that of isometric growth, accompanying the Beverton and Holt model are assumed to exist for the croaker. The assumption of isometric growth, when growth is in fact allometric, leads to incorrect estimates of yield (Paulik and Gales 1964). However, the absolute levels of these estimates are not of as much interest as the differences in yield that result from varying t,. and/or F. The relative error in such differences, when using an incorrect b, tends to be much less than that in the absolute levels (Ricker 1975). Locally, fishing is the major cause of mortality of the croaker. At the present t,. and the level of F the max- imum sustainable Y/R is already being obtained, and any increase in F would result in overexploitation. Increasing t,. to 3 years may not be economically worthwhile since the resulting increase in Y/R is only 6 g. Such a move however, would protect 2-year-old fish that are spawning for the first time. Because of the flattening of the yield curves at high F values, the effect of overexploitation on yield is small. However, the effects of overexploitation on recruitment is un- known, and requires future study. Acknowledgments The authors wish to thank the following: The Univer- sity of the West Indies and the Institute of Marine Affairs, through which this study was carried out; The National Fisheries Company Ltd. and the Caribbean Fisheries Training and Development Institute for as- sistance with obtaining samples; the Seismic Research Unit for assistance with preparation of otolith sections; Manickchand-Heileman and Kenny. Micropogonias furnien in Trinidad waters 529 Dr. Maxwell Sturm for reviewing the manuscript; and three reviewers whose comments served to improve this manuscript. Citations Bagenal, T.B., and E. Braum 1971 Eggs and early life history. In Ricker. W.E. (ed.), Methods for assessment of fish production in fresh waters, p. 166-198. IBP (Int. Biol. Programme) Handb. 3, Blackwell Sci. Publ., Oxford. Barger, L.E. 1985 Age and growth of Atlantic croakers in the northern Gulf of Mexico, based on otolith sections. Trans. Am. Fish. Soc. 114:847-850. Bayagbona, E.O. 1969 Age determination and the von Bertalanffy growth parameters oi Pseudotolithus typus, and P. senegalensis using the burnt otolith technique. In Proc. Symp. Oceanogr. Fish. Res. Trop. Atl., p. 349-359. UNESCO, Paris. Beverton, R.J.H., and S.J. Holt 1957 On the dynamics of exploited fish populations. Fish. Invest. Lond. Ser. II. 19, 533 p. 1959 A review of the lifespans and mortality rates of fish in nature, and their relation to growth and other physiological characteristics. In The lifespan of animals, p. 142-180. CIBA Found. Colloq. on Ageing, Vol. 5. Cassie, R.M. 1954 Some uses of probability paper in the analysis of size fre- quency distributions. Aust. J. Mar. Freshwater Res. 5: 513-.522. Castello, J.P. 1982 Distribution, growth and sexual maturation of Micro- pogonias furmeri juveniles from Patos Lagoon Estuary. Atlantica .5(2):24-25. Fischer, W. 1978 FAO species identification sheets for fishery purposes. Western Central Atlantic (fishing area 31), vol. IV. FAO, Rome (unpaginated). Hile, R. 1936 Age and growth of the cisco Leucichthys nrtedi (Le Sueur), in the lakes of the north-eastern Highlands. Wiscon- sin. Bull. U.S. Bur. Fish. 48:211-317. Holden, M.J., and D.F.S. Raitt 1974 Manual of fisheries science. Part 2. Methods of resource investigation and their application. FAO Fish. Tech. Pap. 115, 214 p. Isaac-Nahum, V.J., and A.E.A. de M. Vazzoler 1983 The reproductive biology of Micropogon funiieri (Des- marest, 1823) (Teleostei, Sciaenidae): 1. Condition factor as an indicator of spawiiing period. Bol. Inst. Oceanogr. S. Paulo 32(l):63-69 [in Portugese, Engl, abstr.]. Le Guen, J.C. 1971 Dynamique des populations de Pseudotolithus (funticulus) elongatus (Bowd. 1825) poissons— Sciaenidae. Cah. ORSTOM Ser. Oceanogr. (9)1:16-84. Lowe-McConnell, R.H. 1966 The sciaenid fishes of British Guiana. Bull. Mar. Sci. 16(l):20-57. Moe, M.A. Jr. 1969 Biology of the red grouper Epinephelus mono (Valen- ciennes) from the eastern Gulf of Mexico. Fla. Dep. Nat. Resourc, Mar. Res. Lab., Prof. Pap. Ser. 10, 95 p. Pannella. G. 1974 Otolith growth patterns: An aid in age determination in temperate and tropical fishes. In Bagenal, T.B. (ed.), Age- ing of fish, p. 28-39. Unwin Bros. Ltd., England. Paulik, G.J., and L.E. Gales 1964 Allometric growth and the Beverton and Hold yield equa- tion. Trans. Am. Fish. Soc. 93:369-387. Pauly, D. 1980 On the interrelationships between natural mortality. growth parameters and mean environmental temperature in 175 fish stocks. J. Cons. Int. Explor. Mer 39(3):195-212. 1984 Some simple methods for the assessment of tropical fish stocks. FAO Fish. Tech. Pap. 234. 52 p. Rao, K.V.S. 1966 Some aspects of the biology of 'Ghol' Pseudosciaena diiu-nnthua (Lacepede). Indian J. Fish. 13 (l-2):251-292. Ricker. W.E. 1975 Computation and interpretation of biological statistics of fish populations. Bull. Fish. Res. Board Can. 191, 382 p. Rodrigues, M.S.S. 1968 Idade e crescimento da cururuca, Micropogon furmeri (Desmarest. 1822), nas Aguas Cearenses. Arq. Estac. Biol. Mar. Univ. Ceara 8(1):7-14 [in Portugese]. Simpson, A.C. 1951 The fecundity of the plaice. Fish. Invest. Lond. Ser. XII 17(5), 27 p. Sparre, P. 1987 Computer programs for fish stock assessment. Length based fish stock assessment for Apple II computers. FAO Fish. Tech. Pap. 101. Suppl. 2, 218 p. Sylvester, C. 1986 Determination of mesh selectivity for demersal fishes in Trinidad, West Indies. Aquatic Sci. Rep., Univ. West Indies, St. Augustine, Trinidad, 144 p. Tanji. S. 1974 Estudo do eontehdo estomacal da pescada-foguete. Macrodon ancylodon (Bloch, 1801) Jordan, Everman and Clark, 1830 e da corvina, Micropogon furnieri (Desmarest, 1822) Jordan, 1844. Bol. Inst. Pesca. Santos 3(2):21-36 |in Portugese). Turner, S.C., C.B. Grimes, and K.W. Able 1983 Growth, mortality, and age/size structure of the fisheries for tilefish, Lopholatilus chamaelonticepn, in the middle Atlantic-southern New England region. Fish. Bull.. U.S. 81: 751-763. Vazzoler, A.E.A. de M. 1970 Micropogon furmerr. Fecundidade e tijio de desova. Bol. Inst. Oceanogr. S. Paulo 18(l):27-32 [in Portugese, Engl, abstr.]. 1971 Diversificagao fisiologiea e morfologica de Micropogon funiieri (Desmarest. 1823) ao sul de Cabo Frio, Brasil. Bol. Inst. Oceanogr. S. Paulo 20(2):l-70 [in Portugese, Engl. abst.]. Vazzoler, G. 1962 Sobre a biologia da corvina da costa sul do Brasil. Bol. Inst. Oceanogr. S. Paulo 12(1):53-102 [in Portugese, Engl, abstr.]. Villyamar, A. 1972 Age determination in fishes of the family Sciaenidae. J. Ichthyol. 13(4):550-561. von Bertalanffy, L. 1938 A quantitative theory of organicc growth. Hum. Biol. 10:181-213. AbStrSCt.— Commercial and rec- reational catches of red drum Sciae- nops ocellatus were sampled from Tampa Bay and Mosquito/upper In- dian River Lagoon, Florida, between August 1981 and March 1983. Males matured at smaller sizes and young- er ages than did females. Males ma- tured when they were 350-799 mm FL (ages 1-3), and females matured when 550-899 mm (ages 3-6). In 1981 and 1982, red drum spawned between August and November, with peak spawning occurring in Septem- ber-October. Histological evidence suggested that spawning occurred in nearshore Gulf of Mexico waters, in passes, and in the estuary. Ages deter- mined from thin sections of otoliths were validated by length-frequency and marginal-increment analyses for fish ages 1-3 and by observations of oxytetracycline-marked fish for ages 5, 12, 16, and 18. Growth was rapid through age 4 or 5 but then slowed markedly. Growth rates of males and females were similar on each coast. Estimates for the von Bertalanffy growth equation parameters K and ifl were not significantly different between coasts; however, L^ was significantly greater on the Atlantic coast. Therefore, predicted lengths of Atlantic fish were greater at all ages. Maximum observed lengths were 980 mm on the Gulf coast and 1110 mm on the Atlantic coast. Max- imum observed ages of sampled fish were 24 years on the Gulf coast and 33 years on the Atlantic coast. The range of 95% confidence intervals for estimates of total annual mortal- ity rate on the Gulf coast was 87-98% for ages 2-4; on the Atlantic coast, the range was 50-76% for ages 2-6. Reproduction, Growth, and Mortality of Red Drum Sciaenops ocellatus in Florida Waters Michael D. Murphy Ronald G. Taylor Florida Marine Research Institute, Department of Natural Resources 100 8th Avenue SE, St Petersburg. Florida 33701 Manuscript accepted 9 March 1990. Fishery Bulletin, U.S. 88:.531-.542. The red drum Sciaenops ocellatus is an estuarine-dependent sciaenid found in nearshore waters (usually <22 m deep) from northern Mexico to Massachusetts (Yokel 1966, Ltix and Mahoney 1969, Ross et al. 1983); it is occasionally found north of Ches- apeake Bay (Yokel 1980). Red drum support important recreational and, until recently, commercial fisheries in most coastal areas of the U.S. south Atlantic and Gulf of Mexico (Mercer 1984). These fisheries have recently undergone strict management in Florida to reduce growth and recruit- ment overfishing (Swingle et al. 1984, Goodyear 1987). Despite the importance of the red drum, little is known about its life history in Florida. While the spawn- ing season and size at maturity have been described for red dnmi in Texas (Pearson 1929, Miles 1951, Matlock 1985) and Mississippi (Overstreet 1983), the spawning season in Florida has been inferred only from larval collections taken along the Gulf coast (Springer and Woodburn 1960, Yokel 1966, Jannke 1971, Peters and McMichael 1987). Age and growth have been described for juvenile and adult red drum in U.S. south Atlan- tic waters (Bearden 1967, Theiling and Loyacano 1976, Music and Paf- ford 1984, Daniel 1988) and in the northern and western Gulf of Mexico (Pearson 1929; Simmons and Breuer 1962; Rohr 1964, 1980; McKee 1980; Doerzbacher et al. 1988). In Florida, larval and juvenile age and growth have been studied in Tampa Bay (Peters and McMichael 1987), but only limited data are available on early growth in other areas (Kilby 1955, Roessler 1967). This paper describes reproduction, age, growth, and mortality for red drum on the Gulf and Atlantic coasts of Florida. Weight-length and length- length relationships are also pre- sented. Methods and materials From August 1981 through March 1983, monthly samples of red drum were collected from commercial and recreational catches in two coastal areas of Florida: (1) Tampa Bay (27° 40'N, 82°35'W) on the Gulf coast, and (2) Mosquito/upper Indian River Lagoon (28°40'N, 80°40'W) on the Atlantic coast. Fish were captured using a variety of gear, including trammel nets, gill nets, hook-and- line, and haul seines. Fork length (FL) in millimeters was measured on all fish in the catch. Each month on both coasts, a random subsample was taken, and up to ten fish per length interval (<300, 400-499, 500-699, 700-899, and >899 mm) were mea- sured for total length (TL) and stan- dard length (SL), weighed for whole weight (W) to the nearest ounce (later converted to grams), and sam- pled for gonads and otoliths (sagit- tae). Gonads were preserved in the field in Davidson's or Zenker's fix- ative (Humason 1972), soaked in water for 24 hours in the laboratory, and then stored in 70% ethanol. A 531 532 Fishery Bulletin 88 (3|. 1990 Table 1 Reproductive classes of red drum gonads of each sex. Class Female Male 1. Immature Few folds of ovigerous lamellae; few or no primary oocytes; Only spermatogonia present; no evidence oogonia predominate; diameter of ovary <3.0 mm. of tubule development. 2. Developing Ovigerous lamellae fill lumen of entire gonad; abundant Early spermatogenesis; few scattered primary oocytes; oogonia present at periphery of lamellae. cysts of primary spermatocytes; periph- or eral tubules differentiating lumen not developed. Resting virgin Lamellae contain abundant primary oocytes; few, if any, oogonia present; absence of atretic oocytes; tunica thickened. 3. Maturing Early vitellogenesis; oogonia, primary oocytes, and oocytes Mid-spermatogenesis; spermatogonia. with yolk vesicle present. cysts of spermatocytes and a few spermatids present along tubules. 4. Mature Late vitellogenesis; oogonia, primary oocytes, and oocytes Late spermatogenesis; few spermato- with yolk vesicles and yolk globules present. gonia; spermatozoa collecting in tubules and central lumen. 5. Gravid Maturation. Oogonia, primary oocytes and oocytes with yolk Reduced number of spermatogenic cysts; vesicles, globules, and migrating nuclei present. efferent ducts filled with spermatozoa; lumen of central duct partially filled with spermatozoa. 6. Spawning Amorphous, hydrated oocytes present; oogonia, primary Efferent ducts filled with spermatozoa; oocytes, and oocytes with yolk vesicles, yolk globules, and spermatozoa in main collecting duct; or migrated nuclei present; fragments of ruptured oocytes distal portions of a few efferent tubules scattered throughout; collapsed, empty follicles present; yolk empty and somewhat thickened; sper- Partially spent remnants usually present in connective tissue and lamellae; tunica greatly thickened. matogonia absent. 7. Spent Oogonia and primary oocytes present in budding ovigerous No spermatogonia or spermatocytes lamellae; ovary with "empty" areas; "plugs" of unshed present; efferent tubules empty; lumen oocytes present; atretic oocytes scattered throughout, usually of central collecting duct with few in association with blood vessels. sperm; testis greatly reduced in size; tunic of previous spawners thickened and convoluted. 8. Recovering Prolific recrudescence of ovigerous lamellae with myriad Network of efferent tubules lined with oogonia and primary oocytes; only gamma and delta atretic spermatogonia; numerous PAS leuco- oocytes present in the lamellae. Tunica thickened and c.Ntes present in central sinus; tunica convoluted. thickened and convoluted. 6-mm-thick sample from the central portion of one lobe of each gonad was embedded in paraffin, sectioned to G-t^m thickness, stained with Mayer's haematoxylin (Humason 1972) and eosin Y, and mounted for micro- scopic examination. Additionally, sections of vitello- genic ovaries fixed in Zenker's fluid were sectioned to S-i^m thickness and stained with Heidenhain's azan (Humason 1972), Eight classes of maturity (Table 1) were distinguished based on the histological criteria of Wallace and Selman (1981) and Hunter and Macewicz (1985) for oogenesis and Grier (1981) and Grier et al. (1987) for spermato- genesis. Recognition of oocyte atresia and rejuvenation was important in determining whether a fish had pre- viously spawned. Atretic oocytes were recognized by degeneration of the zona radiata, nuclear membrane, and yolk globules (Moe 1969. Wallace and Selman 1981, Hunter and Macewicz 1985). "Rejuvenation" is defined as the development of oocytes that were held in an advanced perinuclear stage of development since the previous spawning season. These clutches of oocytes in postspawn, recovering females exhibited a char- acteristic dual cytoplasmic banding (Yamamoto 1956, Howell 1983). Length at maturity was determined by grouping fish by coast and sex into 50-mm intervals, determining the percentage that were mature (> class 4) within each interval, and estimating the length at which 50% of the fish were mature by interpolating between interval midpoints. Oocyte diameters, measured with an ocular micrometer, were used to define the spawning season. Only oocytes whose nuclei were included in the cross section were measured, because these have been shown to represent true oocyte diameters (Foucher and Murphy and Taylor Scisenops ocellatus in Florida waters 533 Beamish 1980). A total of 100 oocytes in a randomly chosen lamella were measured in each female gonad to calculate mean oocyte diameters within subsamples. To further delineate the spawning season and char- acterize variation in egg size, ten of the largest oocytes from each section were also measured (DeMartini and Fountain 1981), and their mean was plotted against the collection date. Otoliths were excised from subsampled fish, cleaned, and stored dry in vials. Sagittae were sectioned in the laboratory using a Beuhler Isomet Low-Speed Saw with diamond wafering blades. A 0.5-mm-thick section, cut through the core of the right sagitta, was mounted on a microscope slide with Coverbond Mounting Media. Each section was examined for age marks using a dissecting scope (32 x ) with reflected light. All sections were independently read once by two individuals; then, when necessary, a common reading was conducted. Determining age by simply counting annuli was not always possible because annulus formation occurs about 2-6 months after the anniversary of the actual hatching date (see "Age Determination and Growth" section). To assign ages accurately, we assumed a biologically realistic hatching date of 1 October (see "Spawning Size, Age, Season, and Location" section; Matlock 1984), and ages were incremented on this date. Observations of otolith sections from oxytetracycline (OTC)-marked fish served to directly validate opaque bands as annuli. Oxytetracycline-injected fish (intra- muscular injection of 25 mg/kg body wt.) were held in outdoor ponds (n = 4) for subsequent recapture and ex- amination after they were free for more than 1 year. All four fish were recaptured after 19 months. Red drum captured by all gear types were included in the growth analyses. The von Bertalanffy (1957) growth equation, FL = L^ (1 - exp [-K(t - /,,)]), was fit to observed age-length data by using the NLIN pro- cedure (Marquardt option) of SAS (Vaughan and Kan- ciruk 1982). Growth equation parameters are defined as follows: L^ = the average fork length (mm) that a fish would achieve if it were allowed to grow indefinite- ly in accordance with the model; K = Brody's growth constant; t^ = the hypothetical age at which a fish would have zero length, and t = age in years (Ricker 1975). The weight-length relationship was described by linear regression of logm -transformed data. Analysis of covariance was used to test for differences between regressions. Age-length keys were applied to length frequencies to separate age groups and develop sample abundance data used to estimate total annual mortality rate (A): Reference to trade names does not imply endorsement by the National Marine Fisheries Service. NOAA. A = 1 - S, where S is the annual survival rate. Age- length keys were developed for each season for com- bined samples captured by trammel net, hook-and-line, and haul-seine. Abundance of each age-group in the samples was then summed over all seasons during the study period. Total annual mortality rates were esti- mated from truncated age data that included all fully recruited age-groups with sample abundance of five or more fish. This procedure eliminated any bias inherent in the collection of large fish (Chapman and Robson 1960). Survival rates were calculated using the modi- fied Heincke method (Seber 1973), Robson and Chap- man (1961) estimates developed for truncated age data, and catch curves. Results and discussion Spawning size, age, season, and location Males matured at smaller sizes (x~-test, jO<0.05) and younger ages than did females on both coasts. Some Gulf coast males were sexually mature after they reached 400 mm, and some on the Atlantic coast were mature after they reached 350 mm (Table 2). Inter- polated lengths at 50% maturity were 529 mm and 511 mm, respectively. Most males were mature at age 1 or 2, and all were mature by age 3 (see "Age Deter- mination and Growth" section). Some Gulf coast females were sexually mature after they reached 600 mm, and some on the Atlantic coast were mature after they reached 550 mm. Lengths at 50% maturity were 825 and 900 mm, respectively. Some females were mature at age 3, and all were mature by age 6. While results of previous investigations suggest that the size and age at which red drum mature may vary over its geographical range (Pearson 1929, Gunter 1950, Miles 1951, Overstreet 1983, Music and Pafford 1984), we show that there are also large maturational differences between sexes. Lengths at 50% maturity were about 300-390 mm smaller for males than females. In Mississippi, males and females began developing when they were 300-549 mm SL (358-629 mm FL); however, only after growing to 700 mm SL (792 mm FL) were more than 50% of the samples of either sex mature (Overstreet 1983). Female red dnim in Texas matured during their fourth or fifth year, at about 750-850 mm TL (71 1-803 mm FL) (Pearson 1929, Miles 1951, Matlock 1985). A 755-mm TL (716 mm FL) male red drum sampled in Georgia had well- developed gonads and was apparently mature (Music and Pafford 1984). Spawning peaked on both coasts of Florida from about September through October. Maximum oocyte diameters and the presence of eggs with yolk vesicles, globules, and migrating nuclei indicated that active 534 Fishery Bulletin 88(3), 1990 Table 2 Percent of s impled red drum in each 50 mm FL length inter- 1 val that were mature (> class 4) on the Gulf and Atlantic coasts | of Florida. Numbers in parentheses are numbers of fish | examined. Gulf Atlantic Fork length (mm) Male Female Male Female 200-249 0 (3) 250-299 0 (7) 0 (4) 0(10) 0 (7) 300-349 0(21) 0(17) 0 (7) 0(17) 350-399 0(18) 0(24) 9(22) 0(22) 400-449 16(19) 0(16) 16(19) 0(32) 450-499 24(21) 0(24) 35(20) 0(28) 500-549 47(19) 0 (6) 56(25) 0(18) 550-599 81(27) 0(24) 70(20) 7(15) 600-649 88(43) 2(53) 85(20) 0(26) 650-699 85(36) 4(28) 82(38) 0(18) 700-749 100(36) 28(36) 93(27) 0(36) 750-799 100 (6) 36(14) 94(17) 7(14) 800-849 100 (3) 50 (2) 100 (6) 7(14) 850-899 100 (1) 100 (3) 100 (4) 0 (1) 900-949 100 (6) 100 (3) 100 (1) 100 (1) 950-999 100 (2) 100 (2) 100 (2) 100 (9) 1000-1049 100 (1) 100 (1) 100 (5) 10.50-1099 100 (2) 1100-1149 100 (1) o O 200 O MEAN • MEAN SON 1981 D 1 J F M A M J 1982 COLLECTION DATE D I J F M I 1983 Figure 1 Mean and mean maximum o(jeyte diameters for red drum sampled from both Atlantic and Gulf coasts of Florida. vitellogenesis occurred in August and September 1981, when the study began on the Gulf and Atlantic coasts, respectively (Fig. 1). Gonads were in spawning condi- tion or were partially spent during September-October 1981. In 1982, active vitellogenesis began in August, and most spawning or partially spent individuals were collected in September. Mean and maximum oocyte diameters declined by mid-November 1981 and by mid- October 1982, indicating cessation of peak spawning (Fig. 1). These results are consistent with the findings of previous studies (Pearson 1929, Mansueti 1960, Yokel 1966, Jannke 1971, Sabins 1973, Peters and McMichael 1987). However, spawning in Tampa Bay may have begun earlier than August in 1982 because larval and juvenile length-frequency distributions and otolith analyses of daily growth indicated that spawning had begun in mid- July (Peters and McMichael 1987). Ap- parently, spawning in mid-July did not occur in a large enough portion of the population to be reflected in our samples. Macroscopic analyses of adult gonads and analyses of gonosomatic indices have suggested that spawning takes place in Texas during October through February (Heffernan 1977) and that spawning off Mis- sissippi takes place during summer (Overstreet 1983). Ovarian histological features provided evidence that spawning occurs in the nearshore Gulf of Mexico, in the vicinity of passes, and within estuaries. Ovaries of females captured in all three areas displayed either (1) advanced oocytes with migrating nuclei or hyaline oocytes released from the follicular layer, indicative of an imminent spawn (DeMartini and Fountain 1981), or (2) atretic bodies and postovulatory follicles (POF), in- dicative of a recent spawn (Yamamoto and Yoshioka 1964, Takita et al. 1983, Hunter and Macewicz 1985). A female captured 4.8 km off the Florida Gulf coast had numerous oocytes with migrating nuclei, and a male from the same site had spermatozoa filling the efferent duct of the testes, both of which suggest im- minent offshore spawning. A large female red drum captured about 7 km within the mouth of Tampa Bay contained ovulated hydrated oocytes, POF, and several atretic oocytes. All of these artifacts indicate that she had recently spawned. Supporting the hypothesis that estuaries are also used as spawning sites is the fact that females {n = 4) that had POF or hyaline oocytes re- leased from the follicle layer were found about 42 km within Tampa Bay, 35 km south of Ponce de Leon In- let in Mosquito Lagoon, and 90 km north of Sebastian Inlet. Based on data from egg and larval collections made just inside passes, most investigators have concluded that red drum spawn principally in nearshore areas close to channels and passes (Pearson 1929, Miles 1950, Murphy and Taylor: Sciaenops ocellatus in Florida waters 535 Simmons and Breuer 1962, Yokel 1966, Jannke 1971, Holt et al. 1985). Our data suggest that spawning may also occur over the nearshore continental shelf and in estuaries. Collections of mature or recently spent fish made in nearshore Gulf waters from just outside the barrier islands to depths of 69.5 m suggest that spawn- ing takes place offshore in the Gulf of Mexico (Christ- mas and Waller 1973; Heffernan 1977; W.A. Fable, NMFS Panama City Lab., Drum seine observers trip report, 18-21 Aug.'l982. Memo OO.Sep.82*002148 to A.C. Jones, NMFS Miami Lab., 4 p.). Peters and McMichael (1987) reported that although most spawn- ing in Tampa Bay occurred close to its mouth, some spawning probably took place in nearshore Gulf waters. Recent egg collections provide evidence that red drum spawn within Mosquito Lagoon (Johnson and Funicelli In press). Our collection of females in Mosquito Lagoon containing POF corroborate these findings and suggest that red drum found equally far within Tampa Bay may also be spawning within the estuary. Moreover, red drum movement was observed during a sonic tracking study in Mosquito Lagoon (Carr and Smith 1977, Carr and Chaney 1976) and little directed movement dur- ing the spawning season was noted. Age determination and growth Ages of red drum from 0 to at least 18 could accurate- ly be determined using otoliths. Red drum sagittae showed clearly delineated, easily interpreted (100% agreement between counts) hyaline and opaque bands. Otoliths with 0-33 opaque bands were examined; they were taken from 1085 red drum (551 Gulf coast and 534 Atlantic coast) with size ranges of 225-1110 mm. Although ages of red drum were validated for only the first 18 years, we assumed, for age and growth anal- ysis, that each opaque band represents an annual mark (annulus). Minimally overlapping monthly length frequencies, a consistent marginal increment minima each year, and observations of oxytetracycline-injected fish supported the hypothesis that the opaque bands are annuli. Modal length progressions representing the 1978, 1979, 1980, and 1981 year-classes on the Gulf coast and 1980 and 1981 year-classes on the Atlantic coast could be fol- lowed in length frequencies (Figs. 2, 3). These length modes could be followed for fish up to 3 years-old and consistently agreed with lengths of fish having cor- responding otolith-determined ages. The periodicity of the mean monthly marginal increment minima each winter (December-March) for fish having one or two opaque bands further suggests that opaque bands were deposited annually (Fig. 4). Opaque bands on fish older than age 3 are also probably annuli. Four adult red drum that were injected and held in outdoor ponds for O 0 UJ a. 50 1980,1 '77^ 1976.3 198^1,0^!^°' Aa"" 979,2 ' ,^ ' , AUG s; T' ' ■■' = 309 ':^°.93o,,/r' SEP B2 1979.3 f^ " '33 IPSO.lT^ 1978,3 JAN 83 N =54 1981 1980.2 1980 1 A 1980.2 979.3 ■^? 19797 ,.^'C>i ..." 200 400 600 800 1000 1200 200 dOO 600 800 1000 1200 FORK LENGTH (mm) Figure 2 Monthly length frequencies fur red drum from the Gulf coast of Florida. No collections were made in April 1982. Horizontal bars show- length ranges for otolith-aged year-classes and are labeled for year- class and age in years. 19 months deposited one opaque band each between release (August 1986) and recapture (March 1988). These fish were 5, 12, 16, and 18 years-old when recaptured. Otoliths have been used to determine the age of red drum (Miles 1951, Rohr 1964, Theiling and Loyacano 1976), but the use of opaque bands as an age deter- minant has never been adequately tested. Off Mis- sissippi, annuli apparently form on otoliths during winter or spring (Rohr 1964). In Texas, where the valid- ity of using scales to determine ages of red drum <4.5 years-old has been established (Matlock et al. 1987), annuli form on scales between February and April in 536 Fishery Bulletin 88 13). 1990 iO ■bt^■ii\ 50 1980-0 N-107 JUL 82 25 1 ' 1979,1 19810 ^ = ^' r\ ' ,"'8'.2 1 ' 1980.1 0 < 1 19792 0 1980.1 ^ = " ,98,Q AUG82 A ""-' ■''8.3 25 ■ /^.^.^, 0 0 tl980,t ^V^ NOV 81 50 1981.0 5,EP82 1 1 N:92 25 / \ 1979,2 25 A 1980.1 / \ "■ -' 1979.2 j \ 197B.3 0 1 ■■ ' 0 50 1980,1 DEC 81 50 OLT82 IVBl.l N^9l 25 > ° O 50 z UJ 1979,2 25 0 SO '""' 1980.2 JV\.-'"" JAN 82 1981.1 NOV 82 '.980,2 ^"'^^ S" ( '19/V, ^ A * 1979.3 UJ ,A, . ^"'S-3 /k _r- . . E " 1981.1 19792 FEB 82 ', ' DEC 82 25 25 \ J2i'.^ 0 50 ,, u \.. . 0 /J V^-^. _ 1980,1 N = 121 25 1980 2 ,n-,n -, ' "• 1978.3 0 50 ^-^^-'X-, 0 ^ . , A— ^ r- ( , , , APB82 19911 fEB83 ^- N-69 N-43 25 1^'.979 2 25 1980.3 A,, r^"" 0 50 -^..-Sv. 0 50 r MAY 82 ,OH, , MAB 83 1980 1 H'7b , — (■ N=96 1 . 1979,2 1980.2 ' ' 1978.3 „^979,3 -■ V__-,_ „)978,4 0 .-^v. ^A^7 0 200 400 500 800 mOO 130C 300 400 600 800 1000 1200 FORK LEN GTH (mm) Figure 3 Monthly length frequencies for red drum from the Atlantic coast of Florida. No collections were made in June 1982. Horizontal bars show length ranges for otolith-aged year-classes and are labeled for year- class and age in years. Matagorda Bay (Wakefield and Colura 1983), during winter in Aransas/Corpus Christi Bay (Pearson 1929), and as early as January in all Texas bays (Matlock et al. 1987). Red drum deposit a first annulus on otoliths and scales during their second winter, when they are about 14-18 months-old (Pearson 1929, Rohr 1964, Theiling and Loyacano 1976, Hysmith et al. 1983, Wakefield and Colura 1983, Matlock 1984, Matlock et al. 1987). Ap- parently, juvenile red drum that measure 40-100 mm during their first winter (Kilby 1955, Peters and McMichael 1987) do not form an annulus; our smallest specimen with an annulus was 379 mm. Red drum grew rapidly until age 4 or 5, and then growth slowed markedly (Table 3; Fig. 5). For each coast, the average observed sizes of fish ages 1-3 were not significantly different between sexes (Student's t- test; p>0.10 in all cases). However, in comparing 1 0 0 0.8 £ 06 t— z Prt-'H \ \ ^.r^"- .H+ 1 -1 00 < 5 tr < 08 * , 1 b. S 06 0 4 02 H-'' ^\ \ • ASONDJFMAMJJASONDJFW 1981 1 1982 1 1983 MONTH Figure 4 Mean monthly marginal increment ( ± 1 SD) for red drum in Florida waters with (a) one and (b) two annuli on otolith sections. Differences between coasts were not significant (ANOVA, p>0.05). therefore pooled data are presented. coasts, the average observed sizes were significantly larger (p<0.001) at ages 1 and 2 on the Atlantic coast than they were on the Gulf coast. After age 2, growth appeared to be slower on the Atlantic coast; however, the variance for age 3 lengths was significantly greater (F-test; p<0.05) on the Atlantic coast. Sample sizes of older age groups were too small to test for statistical differences. Lengths predicted from the von Bertalanffy growth curve agreed with the average observed lengths of red drum on the Gulf and Atlantic coasts (Table 3; Fig. 5). Asymptotic length (L^) was significantly larger (Stu- dent's i-test, p<0.01) on the Atlantic coast than on the Gulf coast, while the Brody's growth constant (A') and age at zero length (^,i) were not significantly different (j:»>0.05). Therefore, predicted lengths were greater at all ages for Atlantic coast red drum than for Gulf coast fish. Our estimates of asymptotic length are generally greater than other reported values: 717 mm TL (680 mm FL), 835 mm TL (789 mm FL), and 803 mm TL (760 mm FL) for Lower Laguna Madre, Mata- gorda, and Galveston Bays in Texas, respectively (Wakefield and Colura 1983); 918 mm TL (865 mm FL) in Texas bays (Doerzbacher et al. 1988); and 950 mm TL (894 mm FL) in Mississippi Sound (Rohr 1980). This suggests that red drum grow larger in Florida or, as Murphy and Taylor. Sciaenops ocell3tus in Florida waters 537 Table 3 Average observed and predicted fork lengths (mm) for red | drum sampled fron Gulf and Atlantic coasts of Florida. Sample sizes are in parentheses. Age Gulf Atlantic Average Average (y) observed Predicted observed Predicted 0 296 (62) 335 (59) 1 436(169) 337 462(192) 373 9 635( 232) 557 655(166) 581 3 736 (65) 696 742 (66) 718 4 845 (5) 784 768 (14) 808 5 877 (1) 839 863 (5) 867 6 874 871 (5) 905 7 896 931 8 910 995 (2) 948 9 975 (1) 919 973 (2) 959 10 925 986 (1) 966 11 928 918 (1) 971 12 925 (1) 930 894 (1) 974 13 932 976 14 933 1028 (2) 977 15 933 954 (1) 978 16 933 960 (1) 979 17 968 (1) 934 979 18 898 (2) 934 979 19 957 (1) 934 975 (1) 979 20 934 980 21 918 (1) 934 980 22 960 (6) 934 1110 (1) 980 23 934 980 24 980 (1) 934 980 25 1038 (3) 980 26 1060 (1) 980 27 980 28 983 (2) 980 29 980 30 980 31 980 32 980 33 1 050 (1) 980 X I- O 0 z lU -J ^ 1000 o • AVERAGE OBSERVED O-O PREDICTED EL - 934.1 ll-eKp[-0.460(AGE-0.029)lf -J I L • AVERAGE OBSERVED O-O PREDICTED mm FL - 979.8 (1-e«p!-0.418l AGE*0 149III 15 20 25 30 35 AGE (YEARS) Figure 5 Average observed ( + 2 SD or range if h = 2; see Table 3) and predicted mean lengths of red drum in Florida waters. Texas (Wakefield and Colura 1983). However, as noted above the differences in len^h-at-age estimates could be attributed to a sampling bias toward smaller near- shore fish. noted by Matlock (1984), that the larger fish which predominately inhabit continental shelf waters were not adequately sampled in Texas and Mississippi. Mat- lock (1984) estimated the von Bertalanffy growth equa- tion for red drum in Texas using data from Pearson (1929) and found what he considered a more reasonable estimate of L^ of 1068 mm TL (1002 mm FL). Length-at-age estimates of red drum in Florida were similar to those reported in past literature for red drum in Texas (Pearson 1929, Miles 1951, Simmons and Breuer 1962; Table 4). However, red drum in Florida appear to grow more rapidly than red drum in Missis- sippi (Rohr 1980), South Carolina (Thieling and Loya- cano 1976), and those in a more recent study from Mortality Ninety-five percent confidence limits for estimates of total annual mortality rate (1 - S) using sample abun- dance data ranged 87-98% for red drum ages 2-4 on the Gulf coast and 50-76% for red drum between ages 2-6 on the Atlantic coast during 1981-83. Red drum were fully recruited by age 2. Sample size of fish older than 4 years on the Gulf coast and 6 years on the Atlan- tic coast fell below the suggested lower limit of five fish to ensure unbiased estimates (Chapman and Robson 1960). Mean estimates of total annual mortality rates were consistently greater for Gulf coast fish (Table 5). Total annual mortality of red drum has been esti- mated for Everglades National Park in Florida and for Texas bay systems. In Everglades National Park, Rago 538 Fishery Bulletin 88(3). 1990 Table 4 Literature accounts of mean fork lengths at age for red drum Fork lengths are in millimeteri- and ivere converted from TLc r SL, if necessary, using the relationships in text. Area, study, and method Age (yr) 1 2 3 4 5 6 7 8 9 10 Texas Pearson (1929) Aransas/Corpus Christi Bays; length frequency 300 530 630 750 840 Miles (1951) Aransas Bay; otoliths 380-423 575 630-684 826 871* 917-940 Simmons and Breuer (1962) Lower coast; tag recapture 322 519 693 Wakefield and Colura (1983) Lower Laguna Madre; scales 290 447 542 Matagorda Bay; scales 255 399 ,526 605 660 Galveston Bay; scales 275 439 547 619 South Carolina Theiling and Loyacano (1976) otolith sections** 429 560 695 782 803 842 803 Mississippi Rohr (1980) otoliths 356 522 638 717 772 810 839 853 867 875 Florida Present study Gulf coast; otolith sections 337 557 696 784 839 874 896 910 919 925 Atlantic coast; otolith sections 373 581 718 808 867 905 931 948 959 966 •A mixture of fish age seven and eight. ** Average size for age group from June th •ough November. Table 5 Pooled mean estimates and 95% confidence intervals of total annual mortality (A) for red drum ages 2-4 on the Gulf coast and ages 2-6 on the Atlantic coast. Method Total annual mortality Gulf Atlantic A 95% CI A 95% CI Seber (1973) Robson and Chapman (1961) Catch curve 0.92 0.91 0.94 0.91-0.94 0.90-0.93 0.87-0.98 0.65 0.66 0.65 0.59-0.70 0.63-0.70 0.50-0.76 and Goodyear (1985) used tag-recapture days-at-large data to estimate a total annual mortality rate of 0.73 during 1984-85. Also based on tag-recapture data, total annual mortality rates in Texas bays were about 68% during 1976-77 (Matlock and Weaver 1979) and 80-87% during 1975-1979 (Green et al. 1985). Matlock (1984) estimated the total annual mortality rate to be 80% for all Texas bays during 1977-1981. Apparent total annual mortality rates in Florida for red drum ages 2-6 are high, especially for a species that may live about 25-35 years. Theoretically, fish with such longevity (assuming constant mortality rate) would have a natural mortality rate of only about 12-18%; this rate allows for survival to the observed maximum ages (Royce 1972, Hoenig 1983). Our much greater estimates of total annual mortality for fish ages 2-6 may reflect one or more of the following: (1) a higher rate of natural mortality for younger fish (i.e., natural mortality is not constant), (2) high fishing- mortality rates within estuaries, and/or (3) emigration from estuaries (sampling area) before age 4 on the Gulf coast or age 6 on the Atlantic coast. In the first two cases, estimates of total mortality are unbiased, al- though component parts (fishing and natural mortal- ity) are not the same. In the third case, total annual mortality would be overestimated because older fish that left the sampling area would be underrepresented in the catch. While data for evaluating whether natural murtal- ity changes with age are not available, tag-recapture data suggest that fishing mortality is high and that emigration does occur. Annual tag-return rates of Murphy and Taylor: Saaenops ocellatus in Florida waters 539 Table 6 Weight-length regressions of red drum sampled from the Gulf and Atlantic coasts of Florida. W = weight (g); FL = fork length (mm). Area N FL a h (mm) (1 SE) (1 SE) r- Gulf 491 Atlantic 484 242-1000 6.1673 x lO"*^ 3.0984 0.993 (0.4301 X 10'') (0.0114) 257-1110 9.3993 X lO'' 3.0275 0.993 (0.6616x10'") (0.0115) 70-80% occurred in southwest Florida during the 1961-65 Schlitz Tagging Programs, and the average tag-return rate throughout the state was 46% (Beau- mariage 1969). More recent annual tag-return rates in Florida were 11-25% (Murphy and Taylor 1985) and 12% (Rago and Goodyear 1985), but if these are ad- justed to account for a probable low tag-reporting rate, e.g., 36% in Texas (Green et al. 1983), then actual an- nual recapture rates in Florida could still be high: 31-69%. Limited data on the emigration of red drum suggest a little exchange from estuarine to nearshore waters. Subadult red dnmi tagging studies within Tam- pa Bay have shown that about 2-6% of recaptured red drum came from outside the Bay (Murphy and Taylor 1985); similar results (1.4%) were reported for the Texas coast (Osburn et al. 1982). Fishing effort and "catchability" of red drum are probably lower in near- shore waters than within the estuary, which could cause the rates of emigration to be underestimated. Weight-length and length-length relations Weight-length regressions (Table 6) were not signifi- cantly different (;)>0.05) between sexes on each coast, Atlantic slope df=l. 480 F = 0.21 elevation df^l. 481 F = 0.34 Gulf slope df=l. 487 F = 1.60 elevation df=l. 488 F = 1.32 although the slopes and elevations were different be- tween coasts with sexes combined. slope elevation df=l, 971 df=l, 972 F = 18.96 F = 27.5. Predicted weights for subadult red drum were similar; e.g., the predicted weight of a 500-mm fish is about 1.4 kg on both coasts. However, adult red drum on the Gulf Table 7 Length-length regressions of red dnmi in Florida. TL = total length (mm); FL = fork length (mm); SL = standard length (mm 1. Sample fork length range for all regressions was | 225- 1110 mm Y = a + bX a b }' A' N (1 SE) (1 SE) ?■- FL TL 1074 23.9383 (0.6704) 0.9162 (0.0011) 0.999 TL FL 1074 -25.2080 (0.7600) 1.0898 (0.0013) 0.999 TL SL 1075 10.3832 (0.8970) 1.1829 (0.0017) 0.998 SL TL 1075 -7.6225 (0.7689) 0.8343 (0.0012) 0.998 FL SL 1075 32.8951 (0.6934) 1.0850 (0.0013) 0.998 SL FL 1075 -29.4619 (0.6738) 0.9202 (0.0011) 0.998 coast were heavier at a given length than they were on the Atlantic coast; e.g., a 900-mm fish weighs 8.8 kg on the Gulf coast and 8.3 kg on the Atlantic coast. Length-length regressions show that total length and standard length increase more rapidly than fork length as fish get larger (Table 7). Total length is about 1% greater at 300 mm FL and 6% greater at 1000 mm FL. Standard length is 18% less than fork length at 300 mm FL but only 12%. less at 1000 mm FL. Summary The following are significant features of the life history of red drum in Florida: rapid growth through age 4 or 5, relatively early sexual maturation (total maturity by age 3-6), a discrete peak in spawning activity during September-October, a life span of up to about 35 years, and spawning grounds located in nearshore waters, in passes and inlets, and inside large estuaries. These characteristics and the apparently high rate of annual mortality for red drum ages 2-6 suggest that only a small portion of the population survives to reach maturity. However, this annual rate of mortality (disap- pearance) has a component of emigration that warrants investigation. Recent management measures enacted by the Gulf of Mexico Fishery Management Council and the Florida Marine Fisheries Commission have limited fishing on the spawning stocks and have reduced fish- ing pressure on the immature fish. 540 Fishery Bulletin 88(3). 1990 Acknowledgments We wish to thank all fishermen who enthusiastically contributed to this study, especially Walter Bell, Cecil Goodrich, and James Burch. Special thanks go to Robert McMichael, Lewis Bullock, and Mark God- charles for assistance with field work. This paper benefited from reviews by Roy Williams, Glenn Par- sons, Behzad Mahmoudi, and Judith Leiby. Age valida- tion work was supported by a MARFIN grant, NA86- WC-H-06136, from the National Oceanic and Atmo- spheric Administration through the National Marine Fisheries Service. Marjorie Myers skillfully typed all drafts of this manuscript. Citations Bearden, CM. 1967 Salt-water impoundments for game fish in South Caro- lina. Prog. Fish-Cult. 29(3):123-128. Beaumariage, D.S. 1969 Returns from the 1965 Schlitz tagging program, including a cumulative analysis of previous results. Fla. Dep. Nat. Resour. Mar. Res. Lab. Tech. Ser. 59, 38 p. Carr, W.E.S., and T. Chaney 1976 Harness for attachment of an ultrasonic transmitter to the red drum, Sciaenops ocellata. Fish. Bull., U.S. 74: 998-1000. Carr, W.E.S., and J.R. Smith 1977 A study of the spawning movements and a tentative spawning site of the red drum. Sciaen-ops ocellata. Unpubl. final rep. to Florida Sea Grant Coll. Prog., Univ. Florida, Gainesville, 31 p. Chapman, D.G., and D.S. Robson 1960 The analysis of a catch curve. Biometrics 16(3):354-368. Christmas, J.Y., and R.S. Waller 1973 Estuarine vertebrates, Mississippi, /m Christmas, J. Y. (ed.). Cooperative Gulf of Mexico estuarine inventory and study, Mississippi, p. 320-343. Gulf Coast Res. Lab., Ocean Springs, MS. Daniel, L.B. Ill 1988 Aspects of the biology of juvenile red drum, Sciaenops ocellatus. and spotted seatrout, Cynoscion twbulosus (Pisces: Sciaenidae). in South Carolina. M.S. thesis. College of Charleston. Charleston, SC. 58 p. DeMartini, E.E.. and R.K. Fountain 1981 Ovarian cycling frequency and batch fecundity in the queenfish, Seniihus politus: attributes representative of serial spawning fishes. Fish. Bull.. U.S. 7!l:!"i47-555. Doerzbacher, J.F., A.W. Green, H.R. Osburn, and G.C. Matlock 1988 A temperature compensated von Bertalanffy growth model for tagged red drum and black drum in Texas bays. Fish. Res. (Amst.) 6:135-1,52. Foucher, R.P., and R.J. Beamish 1980 Pri.iduction of nonviable oocytes by Pacific hake (Merluc- (■i«.s productus). Can. J. Fish. Aquat. Sci. 37:41-48. Goodyear. C.P. 1987 Status of the red drum stocks in the Gulf of Mexico. Con- trib. CRD 86/87-34. Southeast Fish. Cent., Natl. Mar. Fish. Serv., NOAA. Miami, FL 33149, 49 p. Green, A.W., G.C. Matlock, and J.E. Weaver 1983 A method for directly estimating the tag-reporting rate of anglers. Trans. Am. Fish. Soc. 112:412-415. Green, A.W., H.R. Osburn, G.C. Matlock, and H.E. Hegen 1985 Estimated survival rates for immature red drum in north- west Gulf of Mexico bays. Fish. Res. (Amst.) 3:263-277. Grier, H.J. 1981 Cellular organization of the testis and spermatogenesis in fishes. Am. Zool. 21:345-357. Grier, H.J., R.G. Taylor, and R.O. Reese 1987 The mechanism of tubule elongation during testicular recrudesence in the redfish, Snacnops oceUatu,'< (Perciformes). In Kullander, S.O.. and B. Fernholm (eds.), Proc. V Congr. Eur. Ichthyol., Stockholm, 1985. p. 285-291. Swedish Mus. Nat. Hist., Stockholm. Gunter, G. 1950 Correlation between temperature of water and .size of marine fishes on the Atlantic and Gulf coasts of the United States. Copeia 1950:298-303. Heffernan. T.L. 1977 Survey of adult drum {Sciaenopti ocellata). 1973. Proj. Rep. 1973, Tex. Parks Wildl. Dep. Coastal Fish., Austin, p. 37-66. Hoenig, J.M. 1983 Empirical use of longevity data to estimate mortality rates. Fish. Bull., U.S. 82:898-902. Holt. G.J., S.A. Holt, and C.R. Arnold 1985 Diel periodicity of spawning in sciaenids. Mar. Ecol. Prog. Ser. 27:1-7. Howell. W.H. 1983 Seasonal changes in the ovaries of adult yellowtail flounder, Lunanda ferrtiginea. Fish. Bull.. U.S. 81:341-355. Humason, G.L. 1972 Animal tissue techniques. W.H. Freeman, San Fran- cisco, 641 p. Hunter, J.R., and B.J. Macewicz 1985 Measurement of spawning frequency in multiple spawn- ing fishes. In Lasker, R (ed.). An egg production method for estimating spawning biomass of pelagic fishes: application to the northern anchovy, Engraulis mordax. p. 79-94. NOAA Tech. Rep. NMFS 36. Hysmith, B.T., R.L. Colura, and G.C. Matlock 1983 Effects of stocking rate and food tyjie on growth and sur- vival of fingerling red drum. In Stickney, R.R., and S.P. Meyers (eds.), Proc, Warmwater Fish Culture Workshop, March 1-4, 1982, p. 133-144. World Maricult. Soc. Jannke, T.E. 1971 Abundance of young sciaenid fishes in Everglades Na- tional Park. Florida in relation to season and other variables. Univ. Miami Sea Grant Prog, Sea Grant Tech. Bull. 11, 128 p. Johnson. D.R.. and N.A. Funicelli In press Spawning of red drum in Mosquito Lagoon, east- central Florida. Estuaries. Kilby, J.D. 1955 The fishes of two coastal marsh areas of Florida. Tulane Stud. Zool. 2(8): 176-247. Lux, F.E., and J.V. Mahoney 1969 First record of the channel bass, Sciaenops ocellata (Lin- naeus), in the Gulf of Maine. Copeia 1969:632-633. Mansueti, R.J. 1960 Restriction of very young red drum Sciaenop.<: ocellata. to shallow estuarine waters of Chesapeake Bay during late autumn. Chesapeake Sci. l(3-4):207-210. Murphy and Taylor Sciaenops ocellatus in Florida waters 541 Matlock, G.C. 1984 A basis for the development of a management plan for red drum in Texas. Ph.D. diss., Texas A&M Univ.. College Station, 291 p. 1985 Red drum sex ratio and size at sexual maturity. Manage. Data Ser. 85, Tex. Parks Wildl. Dep., Coastal Fish. Branch, Austin, 7 p. Matlock, G.C, and J.E. Weaver 1979 Fish tagging in Texas bays during November 1975- September 1976. Manage. Data Ser. 1. Tex. Parks Wildl. Dep.. Coastal Fish. Branch. Austin. 136 p. Matlock. G.C, R.L. Colura, A.F. Maciorowski. and L.W. McEachron 1987 Use of on-going tagging programs to validate scale readings. In Summerfelt, R.C., and G.E. Hall (eds.), Age and growth of fish, p. 279-286. Iowa State Univ. Press. Ames. McKee, D.A. 1980 A comparison of the growth rate, standard length-weight relationship and condition factor of red drum Sciaenops ocel- latus (Linnaeus), from an electric generating station cooling lake and the natural environment. M.S. thesis. Corpus Christi State Univ.. Corpus Christi. TX. 53 p. Mercer, L.P. 1984 A biological and fisheries profile of red drum, Sciaenops ocellatus. Spec. Sci. Rep. 41. NC Dep. Nat. Resour. Commu- nity Dev. Div. Mar. Fish.. Raleigh. 89 p. Miles. b.W. 1950 The life histories of the spotted seatrout. Cynoscion nebulosits, and the redfish, Sciaeyiops ocellatus. Annu. Rep. (1949-1950), Tex. Game, Fish and Oyster Comm. Mar. Lab., p. 66-103. Tex. Parks Wildl. Dep., Austin. 1951 The life histories of the spotted seatrout, Cynoscion nebukisnt:, and the redfish. Sciaenops ocellatii.% sexual develop- ment. Annu Rep. (1950-1951). Tex. Game. Fish and Oyster Comm. Mar. Lab. lip. Tex. Parks Wildl. Dep.. Austin. Moe, M.A. Jr. 1969 Biology of the red grouper Epinephelus morio (Valen- ciennes) from the eastern Gulf of Mexico. Fla. Dep. Nat. Resour. Mar. Res. Lab. Prof. Pap. Ser. 10. 95 p. Murphy, M.D., and R.G. Tylor 1985 A tag/recovery study of red drum, Sciaenops ocellatus. on the Gulf coast of Florida— Spring 1984. Unpubl. manuscr., Fla. Mar. Res. Inst., Dep. Nat. Resour., St. Petersburg, 19 p. Music. J.F.. and J.M. Pafford 1984 Population dynamics and life history aspects of major marine sportfishes in Georgia's coastal waters. Ga. Dep. Nat. Resour. Coastal Resour. Div. Cont. Ser. 38. Atlanta, 382 p. Osburn, H.R.. G.C Matlock, and A.W. Green 1982 Red drum (Sciaenops ocellatus) movements in Texas bays. Contrib. Mar. Sci. 25:85-97. Overstreet. R.M. 1983 Aspects of the biology of the red drum. Sciaenops ocel- latus, in Mississippi. Gulf Res. Rep. (Suppl. 1). p. 45-68, Pearson, J.C 1929 Natural history and conservation of the redfish and other commercially important sciaenids on the Texas coast. U.S. Bur. Fish. Bull. 44:129-214. Peters, K.M., and R.H. McMichael Jr. 1987 Early life history of Sciaenops ocellatus (Pisces: Sciae- nidae) in Tampa Bay, Florida. Estuaries 10(2):92-107. Rago, P.J., and CP. Goodyear 1985 Estimation of mortality rates and population size from tagging studies in Everglades National park, 1984-85. U.S. Fish Wildl. Serv.. Gainesville. FL. 32 p. Ricker, W.E. 1975 Computation and interpretation of biological statistics offish populations. Bull. Fish. Res. Board Can. 191, 382 p. Robson, D.S., and D.G. Chapman 1961 Catch curves and mortality rates. Trans. Am. Fish. Soc. 90:181-189. Roessler, M. 1967 Observations on the seasonal occuirence and life histories of fishes in Buttonwood Canal, Everglades National Park, Florida. Ph.D. thesis, Univ. Miami, Coral Gables, FL, 155 p. Rohr, B.A. 1964 Comparison of the growth rings in the scales, otoliths, dorsal rays, and second anal spine as related to growth of red drum. Sciaenops ocellatus. J. Miss. Acad. Sci. 10:208-212. 1980 Use of hard parts to age Gulf of Mexico red drum. In Williams, R.O., J.E. Weaver, and F.A. Kalber (eds.), Proc. Colloq. on biology and management of red drum and seatrout, Oct. 19-20, 1978, p. 15. Sponsored by Gulf States Mar. Fish. Comm., Ocean Springs, MS [abstract]. Ross, J.L., J.S. Pavela, and M.E. Chittenden Jr. 1983 Seasonal occurrence of black drum, Pogonias cromis, and red drum. Sciaenops ocellatus. off Texas. Northeast Gulf Sci. 6(l):67-70. Royce, W.F. 1972 Introduction to fishery sciences. Academic Press, NY, 351 p. Sabins, D.S. 1973 Diel studies of larval and juvenile fishes of the Caminada Pass area. Louisiana. M.S. thesis, Louisiana State Univ., Baton Rouge, 163 p. Seber, G.A.F. 1973 Estimation of animal abundance and related parameters. Griffin, London. 506 p. Simmons, E.G., and J. P. Breuer 1962 A study of redfish. Sciaenops ocellata Linnaeus and black drum. Pogonia.'^ cromis Linnaeus. Publ. Inst. Mar. Sci. Univ. Tex. 8:184-211. Springer, V.G., and K.D. Woodburn 1960 An ecological study of the fishes of the Tampa Bay area. Fla. Board Conserv. Mar. Lab. Prof Pap. Ser. 1. 104 p. Swingle. W.. T. Leary, C Davis, V. Blomo, W. Tatum. M. Murphy, R. Taylor, G. Adkins, T. Mcllwain, and G. Matlock 1984 Fishery profile of red drum. Gulf of Mexico Fish. Manage. Council and Gulf States Mar. Fish. Comm.. Tampa. FL. 164 p. Takita, T., T. Iwamoto, S. Kai, and I. Sogabe 1983 Maturation and s[.)awning of the dragonet. Callionymus enneactis in an aquarium. Jpn. J. Ichthyol. 30(3):221-226. Theiling, D.L., and H.A. Loyacano 1976 Age and growth of red drum from a saltwater marsh im- poundment in South Carolina. Trans. Am. Fish. Soc. 105: 41-44. Vaughan, D.S., and P. Kanciruk 1982 An empirical comparison of estimation procedures for the von Bertalanffy growth equation. J. Cons. Int. Explor. Mer 40:211-219. von Bertalanffy, L. 1957 Quantitative laws in metabolism and growth. Q. Rev. Biol. 32:217-231. Wakefield, CA., and R.L. Colura 1983 Age and growth of red drum in three Texas bay systems. Proc. Tex. Chapter Am. Fish. Soc. 5:77-87. Wallace. R.A., and K. Selman 1981 Cellular and dynamic aspects of oocyte growth in teleosts. Am. Zool. 21:325-343. 542 Fishery Bulletin 88(3). 1990 Yamamoto, K. 1956 Studies on the formation of fish eggs, I. Annual cycle in the development of ovarian eggs in the flounder, Liopsetta nbscura. J. Fac. Sci. Hokkaido Univ. Ser. VI Zool. 12:362-373, Yamamoto, K., and H. Yoshioka 1964 Rhythm of development in the oocyte of the medaka Oryzias latipes. Bull. Fac. Fish. Hokkaido Univ. 15(l):5-20. Yokel, B.J. 1966 A contribution to the biology and distribution of the red drum. Sciaenops ocellata. M.S. thesis, Univ. Miami, Coral Gables, FL, 160 p. 1980 A contribution to the biology and distribution of the red drum, Sciaeiuyps ocellata. In Williams, R.O., J.E. Weaver, and F.A. Kalber (eds.), Proc, Colloq. on biology and management of red drum and sea trout, Oct. 19-20, 1978, p. 5. Sponsored by Gulf States Mar. Fish. Comm., Ocean Springs, MS [abstract]. Abstract.- The sandi'lsh family Trichodontidae is comprised of two species endemic to the boreal Pacific: Arctoscopus japonicus is a common demersal fish of commercial impor- tance in the western area around northern Japan, while Trichodon tri- chodon is found in the eastern waters, from Alaska to California, although there is no fishery for this species. The two species share many similar reproductive features such as large demersal eggs of 3.3-3.5 mm diam- eter, moderate fecundities of 1000- 2000 eggs, and peculiar spawning habits on rocky shores. However, there are striking contrasts in sev- eral reproductive and early-life-his- tory traits: the shape of the egg mass is spherical in A. japonicus versus an irregular shape in T. trichodon; spawning substrate is mostly Sargas- sum spp. versus rock; incubation period is about 2 months versus near- ly 1 year; and, at hatching, T. tricho- don has precocious pectoral and caudal rays as well as preopercular spines. Despite these differences, lar- vae of both species appear in the spring season. These features are discussed in relation to the contrast- ing thermal regimes of the surface waters characteristic of each species. It is suggested that /I. japonicus has a more derived reproductive style than T. trichodon, including a typical pattern of indirect ontogeny. The remarkable seasonal thermal events in Japanese coastal waters may be responsible for the notable abun- dance of ^. japonicus. Contrast \n Reproductive Style Bet>A/een Two Species of Sandfislies (Family Trichodontidae) Muneo Okiyama Ocean Research Institute. University of Tokyo, Minamidai, Nakano, Tokyo 164, Japan 15-1 Manuscript accepted 13 March 1990. Fishery Bulletin. U.S. 88:543-549. The family Trichodontidae is a small family of marine fishes endemic to the boreal Pacific comprised of two species, Arctoscopiis japcmieus (Stein- dachner) from the western Pacific around Japan and Trichodon tricho- don (Tilesius) from the eastern Pacif- ic, ranging from Alaska to California. As indicated by the common name "sandfishes," they typically lie partly buried in the bottom (Nelson 1984). Due to peculiar morphological and ecological features, they have been placed at various times in different groups of fishes such as the Percoi- dei (Matsubara 1963) and Blennioidei (Greenwood et al. 1966). The biology of A. japonicus has been extensively studied because of its special importance to commercial fisheries in northern Japan and Korea (for review, see Ochiai and Tanaka 1986). For T. trichodon, how- ever, information is only briefly docu- mented in the literature (Fitch and Lavenberg 1975, Eschmeyer and Herald 1983), except for Marliave (1981) and Bailey et al. (1983) who reported the early life history in de- tail, and Allen and Smith (1988) who quantitatively illustrated aspects of its zoogeography. These studies and our own ob- servations, mostly in the northern Sea of Japan, enable us to compare life history traits of these fishes and to discuss probable life strate- gies associated with zoogeography. The notable fluctuation of the stocks of A. japonicus is also considered from the standpoint of comparative ecology. Comparison between Arctoscopus Japonicus and Trichodon trictiodon Distribution The family Trichodontidae is restricted in distribution to the coastal boreal regions of the North Pacific between about 35°N and 60°N (Fig. 1). Al- though Schmidt (1950) reported the occurrences of A. japonicus from Akutan Bay and the vicinity of Sitka Island, in the Gulf of Alaska, these records seem questionable (Esch- meyer and Herald 1983). Perhaps these two species have distinct geo- graphica ranges, separated by longi- tude 160-170°E. Arctoscopus japo- nicus occurs over an extensive area on the western side, including the off- shore regions; it is extremely abun- dant in the Sea of Japan and the Pacific coast off Hokkaido, where catches amounted to 3.8 x 10'* tons in 1968 (Ochiai and Tanaka 1986). Tri- chodon trichodon is an eastern Pa- cific form with a wider range from Kamchatka to southeast of San Fran- cisco, California; there is no fishery for this species, but it is fairly com- mon throughout the coastal waters (Hart 1973, Fitch and Lavenberg 1975, Eschmeyer and Herald 1983, Allen and Smith 1988). Depth range is variable from the surface to about 400 meters in both species; however, the main life zones are slightly different for the two spe- cies. Arctoscopus japonieus in the Sea of Japan usually occurs at moderate depths of 200-300 m, with optimum 543 544 Fishery Bulletin 88(3). 1990 -"^^T BERING SEA Arc tosc opus laponicus Trichodon trie hodon EA of/ ::*C-5;i*H: NORTH PACIFIC OCEAN 80 N T I I I I I I I I I I I I I I I I I I I I I I I T 130 E 150 E 170 E 170 W 150 W 130 W 110 W Figure I Geographical distribution of two species of sandflshes, family Trichodontidae. (A) Pohang, (B) Akita, (C) Wakkanai, (D) Kholnisk, (E) Nemuro, (F) Attu, (G) Dutch harbor, (H) Kndiak, (I) Sitka, (J) Vancouver. For references, see text. temperature ranges of 2-5°C (Nishimura 1969). This zone is shifted to a shallower depth of 100-200 m on the Pacific side off Hokkaido (Ochiai and Tanaka 1986). During the spawning season, the adults move into water shallower than about 15 m. In contrast, T. trichodon is most frequent on the middle shelf at less than 150 m (Allen and Smith 1988) and spawning is restricted to the intertidal area. Reproductive ecology Both species spawn large demersal eggs in rocky, shallow waters during winter months, but interesting ecological differences are found between them, sum- marized in Table 1. Typically, A.japonicus deposits a spherical egg mass tightly on Sargassum spp. in such a way that supporting stems pass through the axis of the mass. Successful deposition of secure substrata is of prime importance for survival of eggs, because egg capsules that detach from the substrata become stranded ashore and perish. Onset of spawning at fixed times and locations is characteristic of A. japonicus. In the Akita district, for instance, the onset of spawn- ing at major spawning sites rarely fluctuates beyond 1 or 2 weeks from the end of November (Kato 1980). In contrast, Eschmeyer and Herald (1983) reported that the eggs of T. trichodon are laid in a gelatinous mass attached to rocks. One report of a natural spawn- ing was on the wall of a fully exposed surge channel Table 1 Comparison between reproductive characters of Arrtoscopiix japonicus and Trichodon trichodon (after Ochiai and Tanaka 1986; Marliave 1981, unless otherwise indicated below). Characters A. japonicus T. trichodon Spawning season Mostly from the end of November to December Around February Spawning migration Conspicuous Indistinct Spawning site; sul)strate Coastal waters, 0-15 m depth; mostly Sargassum spp. Intertidal waters, 0-1 m depth; rock Shape of egg mass Spherical, 3.3-7 cm in size Irregular Diameter 3.3-3. ,5 mm* 3..52±0.1i) mm of egg Fecundity 600-2300 ~I000 Incubation period .56-78 days at 6°-9°C** 1 year (?) at 8°-13°C Synchronized hatching No* Yes; within 24 hours publ.) (85); this indicates the most frequent •Okiyama (un *'Maekawa (U period. Okiyama Reproductive style of two species of sandfishes 545 Arctoscopus japonicus A , 13.3mm 14.2mm 16.8mm 17.8mm Trichodon trichodon 13.0mm 14.5mm -.^^ 28.2mm 27.0mm Figure 2 Comparison between early developmental stages of Arctoscopus japomcus (left) and Trichodon trichodon (right). Size in mm SL; preserved specimens (B-D after Okiyama 1988; F-J after Marliave 1981). (Marliave 1981). It seems that T. trichodon selected this peculiar spawning location because it offered a refuge from egg predation as well as the high flow velocities necessary for respiration (Marliave 1981). These contrasting reproductive styles should be asso- ciated with the different reproductive guilds proposed by Balon (1981, 1984). Although A. japonicufi fits the "obligatory plant spawners" category, there is no guild relevant to T. trichodon within Balon's category of "nonguarders," suggesting its spawning habits may be unique. Early life history Larvae and juveniles of both species have been de- scribed and illustrated in detail from field and/or labor- atory-reared materials (Marliave 1981, Okiyama 1988). In Figure 2, a series of early developmental stages of these species is reproduced from these references, with additional specimens of A. japonicus included. Here, specimens of similar sizes are selected to facilitate interspecific comparison. The early growth of the two species is also compared in Figure 3, based on the laboratory-rearing results; the comparison reveals that T. trichodon is consistently larger than A. japonicus throughout the 2 months following hatching (Marliave 1981, Maekawa 1985). Despite close resemblances in the general morphology, there are remarkable differ- ences in early developmental patterns between the two species. A few interesting aspects are as follows: (a) Newly hatched larvae of A. japonicus, 13 mm notochord length (NL), are less advanced than those of T. trichodon, 13.0 mm standard length (SL); the body is slender, the notochord tip is straight, and fin ele- ments, except several incipient caudal rays, and teeth are absent. In contrast, T. trichodon possesses many precocious characters, including a flexed notochord with full complement of principal caudal rays, develop- ing pectoral rays and preopercular spines, in addition to a deeper body, larger eyes, and the presence of 546 Fishery Bulletin 88(3). 1990 5 _^^ E E ,0 £ ^' 1 a> o ^-^ • .-^ 25- ^ n ^ •D .-^ n W o y^ (Q ■^ o 15- T ^ .-'-'''^ ■^I ■'■ * 10- 0 10 20 30 40 50 60 Days after hatching Figure 3 Early growth of laboratory- reared specimens of sandfishes in fresh condition. For Arctoscopu^ japonicus, average ( • ) and ranges (line) are shown in TL mm (after Maekawa 1985); for Trichodon trirhodon, individuals (O) are shown in SL mm (after Marliave 1981). teeth on both jaws. It is clear that these precocious characters are mostly concerned with active swimming. Fast swimming, coupled with large sizes, would not only facilitate successful feeding, but would play an effective role in lowering predation pressure. Slight but distinct size and development differences at hatching between the two species, therefore, are among the most important aspects of their early life histories. (b) Orders of fin ray ossification of the two species can be derived from Table 2, although the available size series of T. trichodon (Marliave 1981) is somewhat limited. Except for the precocious caudal and pectoral fins in T. trichodon, all other fin meristics are formed in smaller specimens of A. japonicus. The entire size ranges for fin development are 14.9-19.0 mm in A. japonicus versus 13.0-30.0 mm in T. trichodon. Pos- sible sequence of fin ossification is caudal-pectoral- anal-2nd dorsal- 1st dorsal-pelvic for both species. (c) Larvae of both species are sparsely pigmented and lack extensive melanin on the lateral side of the tail. Pigmentation is generally heavier in T. trichodon than in ^4. japonicus. In particular, earlier development of distinct spots along the anterior dorsal margin of T. trichodon is remarkable because this character first occurs after transformation into juveniles in A. japonicus. (d) According to the conventional definitions of early developmental stage, transformation from larval to juvenile stages occurs at about 20 mm in A. japonicus versus about 30 mm in T. trichodon. However, T. tri- chodon is distinctly more advanced than A. japonicus in everj' early-life-history trait, except formation of fins other than the caudal and pectoral. Table 2 Comparison between fin formation of yl and Trichodon trichodon (mm SL). rctoscopu^ japonicus Characters A. japonicus * T. trichodon" start finish start finish 1st dorsal 16.2 19.0 20.0- 28.0 2nd dorsal 16.2 17.8 20.0 27.0 Anal fin 16.2 17.8 20.0 27.0 Pectoral 14.9 17.8 13.0 17.2 Pelvic 16.8 19.0 27.0 30.0 Caudal 14.9 14.9 13.0 13.0 (principal) from specimens of the ^(1981), slightly modifie larvae (13.0, 13.2 mm Sea of Japan. d on the basis of two SL; preserved) pro- 'Original data "After Marliavf newly hatchec vided iiy him. Major contrasts in their ontogenetic properties may be summarized as follows: A. japonicus has a distinct, but truncated, metamorphosing phase (20-30 mm SL) closely corresponding to the change from pelagic to demersal life (Minami and Tanaka 1985, Okiyama un- publ.); in contrast, T. trichodon has a less distinct meta- morphosing phase (30-50 mm SL) without any eco- logical shifts other than possible decrease of schooling tendencies (Marliave 1981, Bailey etal. 1983), although it starts burying itself in the sand at 55-60 mm SL (J.B. Marliave, Vancouver Aquarium, P.O. Box 3232, Van- couver, B.C., Canada, pers. commun., April 1989). Contrast In life history strategy Notwithstanding the close systematic relationship be- tween A. japonicus and T. trichodon, there are many life history features unique to each species, particularly with respect to the reproductive styles as discussed earlier. The terms and duration of their early develop- mental stages are shown in Figure 4, revealing that the egg stage of T. trichodon is almost four times long- er than that of A. japonicus; however, the time when the larvae of both species appear is restricted to the spring seasons. The mechanism of this synchronized hatching is unknown, but plausible advantages are ap- parent. It is obvious that spawning substrates play an important role in producing these contrasting patterns of development. Perhaps the specialized mode of egg deposition unique to A. japonicus has evolved in close association with the abundant substrates represented by Sarga.'T,, where T, is the travel time to the upstream compart- ment and r, is the fatigue time of the fish. Travel time to the upstream compartment depends on the average velocity of the current across the redds, U. the average fish swimming velocity, V, and the distance, L, it travels across the redd, and is expressed: T, = LI(V-U). Fatigue time also depends on swimming velocity. Laboratory experiments indicate that among adult rainbow trout, the time to fatigue decreases exponen- tially with swimming velocity for velocities above 3 body lengths per second (bl/s). However, at 2 bl/s adult rainbow trout can swim indefinitely (Beamish 1978). Alevins probably have a similar relationship, although (due to smaller size) they can maintain considerably larger swimming velocities on a body length basis, ex- ceeding 30 bl/s (100 cm/s) (Ostrander et al. 1989). The model assumes time to fatigue is infinite when swim- ming velocity is zero and decreases as swimming velocity increases. A simple power function has these properties: T, = k V-" where A' and « are parameters that depend on condi- tion of the environment and the fish. Equating the two time scales, a critical stream velocity is defined: U* = r - L/kV where successful upstream migration requires U 00 Cnucal Stieam Velocily. U* Stream Velocity, U 0 Swimming Velocity, V Figure 2 Critical stream velocity curve. Successful upstream movement re- quires that the critical stream velocity is greater than the stream velocity. Intersection of U and U* defines the allowable swimming velocity. V. required for successful movement into the upstream compartment. exposed to B[a]P became fati^ied faster or swam upstream slower. In the first case, the critical stream- velocity curve of exposed fish would fall below the curve for unexposed fish. In the second case, the lower swimming velocity of the exposed fish would fall out- side the allowable velocity range for successful up- stream migration. The experiments were not designed to resolve such differences. The model provides a qualitative basis for inter- preting the experiments. It differs from standard dose- response models that describe the degree of change as a continuous function of the level of duration of tox- icant exposure (Krewski and Brown 1981). The model allows a possible interplay of behavior and stamina for successful upstream movement, although such a sep- aration was not pursued in the study. Our results indicate that low levels of toxicants may affect important early-life-history behaviors critical to survival. Though the concentration of B[a]P used in this study is not frequently encountered in nature, the value of this study in predicting what could happen follow- ing some type of major accidental spill or discharge should not be overlooked. Given that a single 24-hour pulse of B[a]P elicited significant behavioral changes, future studies could be directed at understanding the potential effects of low-level chronic exposures which may more accurately mimic natural conditions. and D.J. Randall Academic Press, Plenum Publ., NY, Acknowledgments Data presented in this manuscript are included in a dissertation submitted to the College of Ocean and Fishery Science, University of Washington, in partial fulfillment of the requirements for the Ph.D. degree (GKO). The authors thank Kathleen Sabo for excellent technical assistance. This study was supported by a grant (R-811348) from the U.S. Environmental Protec- tion Agency. Citations Baumann, P.C. 1989 PAH, metabolites, and neoplasia in feral fish popula- tions. In Varanasi, U. (ed.). Metabolism of polycyclic aromatic hydrocarbons in the aquatic environment, p. 269-289. CRC Press, Boca Raton, FL. Beamish, F.W.H. 1978 Swimming capacity. In Hoar, W.S., (eds.). Fish physiology, vol VH, p. 101-187. NY. Champ. M.A.. and P.K. Park 1982 Global marine pollution bibliography. 399 p. Chapman, D.W. 1962 Aggressive behavior of juvenile coho salmon as a cause of emigration. J. Fish. Res. Board Can. 19:1047-1080. Fast, D.E. 1987 The behavior of salmonid alevins in response to changes in dissolved oxygen, velocity, and light during incubation. Ph.D. diss., Univ. Wash., Seattle, 1.50 p. Gelboin. H.V. 1980 Benzo(a)pyrene metabolism, activation, and carcino- genesis: role and regulation of mixed-function oxidases and related enzymes. Physiol. Rev. 60(4):1107-1116. Krewski, D., and C. Brown 1981 Carcinogenic risk assessment: A guide to the literature. Biometrics 37:353-366. Mason, J.C, 1976 Some features of coho salmon, Oncorhynchus kisutch, fry emerging from simulated redds and concurrent changes in photobehavior. Fish. Bull., U.S. 74:167-175. Mason, J.C, and D.W. Chapman 1965 Significance of early emergence, en\ironmental rearing capacity, and behavioral ecology of juvenile coho salmon in stream channels. .J. Fish. Res. Board Can. 22:173-189. Ostrander, O.K., M.L. Landolt. and R,M, Kocan 1988 The ontogeny of coho salmon (Oncorhynchus kisutrh) behavior following embryonic exposure to benzo(a)pyrene. Aquat. Toxicol. (Amst.) 13:325-346. 1989 Wliole life history studies of coho salmon {Oncorhynchus kisutch) following embryonic exposure to benzo[a]pyrene. Aquat. Toxicol. (Amst.) i5(2):109-126. Abstract.- Sagittal otoliths of the Antarctic fish Nototheniaps nudi- /ro?(.s (F^amiiy Nototheniidae) contain internal microincrements which are visible by scanning electron micro- scopy. Microincrements were depos- ited on a daily basis, as validated through tetracycline and acetazola- mide marking experiments. This was the first validation of daily microin- crement deposition for any Antarc- tic fish. Daily formation of microin- crements continued throughout the year, including during the winter when daylight periods were short. Counts of daily microincrements in the otoliths of 32 juveniles and adults allowed the determination of age and growth rates. From this sample, a multivariate regression model relat- ing fish age to otolith morphometries and fish size demonstrated that age could be estimated reliably from sag- ittal otolith weight and length. Age was estimated for a large sample of N. nudifrons. which allowed the de- termination of growth and natural mortality. Fish grew slowly (1.5 mm per year), reaching sexual maturity at an age of 4-5 years, with the larg- est fish attaining ages of more than 8 years. Growth and survivorship were similar for males and females. Age and Growth of the Antarctic Fish Nototheniops nudifrons Richard L. Radtke Oceanic Biology, Hawaii Institute of Geophysics University of Hawaii, Honolulu, Hawaii 96822 Thomas F. Hourigan Department of Zoology, University of Hawaii Honolulu, Hawaii 96822 The Antarctic fishes are character- ized by a high degree of endemism and extensive adaptations to their unique environment (DeWitt 1971). Ahhough commercial exploitation of several species has begun, much of their basic biology is still poorly understood (Kock et al. 1985). Of special importance to ecological and fisheries research is determination of the age of fish captured in the field. Growth, mortality, and fecundity can be derived from age data. Ageing studies can also provide basic life- history information, such as popula- tion structure, and changes in popula- tion growth due to environmental perturbations. These age-related data increase our understanding of fish biology and form the basis of popula- tion dynamics models. Antarctic fishes appear to be slow growing and long lived (Shust and Pinskaya 1978, Freytag 1980a, Kock et al. 1985, Radtke et al. 1989). Al- though age and growth of Antarctic fishes have primarily been deter- mined by counting scale annuli (Olsen 1954, 1955; Wohlschlag 1961, 1962; Everson 1970, 1980; Shust and Pin- skaya 1978; and others), the results from this technique are often diffictiit to interpret and especially unreliable for older fishes, as scales may be regenerated or resorbed (Mugiya and Manu.sfript accepted 14 May 1990. Fishfi-v Bulletin, U.S. 88:557-571. 'Contribution Number 2327 of the Hawaii In- stitute of Geophysics. Watabe 1977; Freytag 1980a, 1980b). Therefore, an alternative ageing method, such as otolith increment analysis, is desirable. In contrast to scales, otoliths are neither regener- ated nor resorbed, and are the most precise structure for ageing of fishes (Six and Horton 1977, Campana and Neilson 1985). Otoliths are calcium carbonate struc- tures deposited in the membranous labyrinth of the inner ear of fish. The otoliths of many fishes grow by daily accretion of layered increments (Pan- nella 1971). These increments are visible in sectioned otoliths as concen- tric rings, which can be enumerated to provide estimates of the age of temperate and tropical fishes (Pan- nella 1971, 1974; Campana and Neil- son 1985; Jones 1986). Otoliths of temperate fishes may also form an- nual increments or annuli. Several studies have utilized oto- liths to age Antarctic fishes. Olsen (1955) found that the otoliths of ice- fishes were small and difficult to analyze by light microscopy. Rings in the otoliths of the icefishes Chaeno- cephaluf^ aceratus and Champsoce- phalus gunnari were interpreted to be annuli (Olsen 1955). Subsequent- ly, other researchers used rhythmic patterns in Antarctic fish otoliths for age determinations (Hureau 1966, Everson 1980, Freytag 1980b, North et al. 1980, Mucha 1980, Chojnacki and Palczewski 1981, Kock 1981, 557 558 Fishery Bulletin 88(3), 1990 Sosinski 1981, Kochkin 1982, Burchett et al. 1984). These patterns were interpreted to be annuli; however, no vahdation data were available for the deposition rates of these increments. These presumed annuli, which are often difficult to discern and interpret, may not represent yearly growth increments (Scherbich 1975, Frey tag 1 980b, Radtke and Targett 1984). Townsend (1980) described the presence of micro- structural growth increments in Antarctic fish otoliths. These increments appeared to be analogous to the daily increments found in a host of fish species (see review by Campana and Neilson 1985). Such microincrements were used by Radtke and Targett (1984) to age Noto- thenia larseni, and are the foundation for the present research. However, the daily or annual nature of rhyth- mic patterns in otoliths has not been experimentally validated for any Antarctic fish species. Microincre- ments may generally form in response to changes in daily periods of light and dark (Radtke 1984), sug- gesting that special problems may exist in the forma- tion of daily increments during times of short daylight in the winter months. Management programs for Antarctic fisheries re- quire knowledge of population parameters of both com- mercial and sympatric, non-commercial, species. Noto- theniops nudifrons ( =Notothenia nudifrons) is among the most abundant demersal fishes in many habitats of the Antarctic Peninsula (DeWitt 1971). As such, this species may play a major role in the trophic structure of the Antarctic marine community, both as competitor with and prey for commercial species. We recently reported on the reproductive biology of A'', nudifrons (Hourigan and Radtke 1989). In the pres- ent study, the age and growth of A'^. nudifrons were determined by the examination of microincrements in sagittal otoliths. Microincrement deposition rates dur- ing different seasons were tested under experimental conditions. Correlations of age to otolith morpho- metries allowed ageing of a large sample of fish and an estimation of their natural mortality rates. Methods Collection of fish and initial measurements A total of 32 Nototheniops nudifrons were collected on 28 March and 17 April 1985 by otter trawl in 54-110 m depths off Low Island (63°24'S to 63°27'S; 62°07'W to 62°17'W). An additional 30 fish were collected in the same area in February 1984, and used for prelim- inary otolith validation studies. All individuals were transported live to Palmer Station and placed in tanks with flow-through seawater. In addition to these fish, scuba divers collected three small juveniles near Palmer Station in Arthur Harbor (64°46'S; 64°04'W) in May and July 1985. The fish from the 1985 trawls were analyzed for length-weight relationships and growth. Standard length (SL) and total length (TL) were measured to the nearest millimeter. Whole body weights were measured to the nearest 0.001 g. Otoliths were removed from all fish, cleaned, dried at 60°C, and stored in a desiccator. The gonads were examined to determine sex and re- productive condition (Hourigan and Radtke 1989). The remaining fish were injected intramuscularly with acetazolamide (samples from 1985 only) or tetra- cycline (samples from 1984 and 1985) and kept in 5()0-L tanks with flow-through seawater, at ambient temper- atures and natural photoperiods. Ten fish were kept in each tank, along with 10 Notothenin gibherfrons. Rocks were added to the tanks to provide shelter and nesting sites. Fish in tanks were fed ad libitum with kril! every 2 days. Fish were kept alive for a maximum of 158 days after injection, from April to October, at which time they were sacrificed, weighed, and measured. Otoliths of these fish were removed, cleaned, and stored dry in vials. Otolith structure and age determination from otolith microincrements To determine the relationship of age to otolith size and shape, the length, width, and weight of sagittal otoliths were measured. Otoliths were segi'egated according to their lateral position in the cranium. Left and right sagittae from all fish were scanned using a computer- aided video digitizer, which produced a measure of maximum length (from the rostrum to antirostrum; nomenclature of Hecht 1978) and width (the widest distance in the dorsal-ventral plane). Otoliths were weighed on a microbalance to the nearest 0.01 mg. A sample of 32 fish (from fish collected in 1985) was chosen for age determinations. This sample was com- posed of males and females of a representative size distribution. Left sagittae from these fish were at- tached to scanning electron microscope (SEM) viewing stubs with epoxy, carefully ground down to the central area by hand using a fine sharpening stone, and then polished with 0.3-micron alumina paste. The polished surfaces were etched for 1-20 minutes with 6% ethy- lene diamine tetraacetate (EDTA) with pH adjusted to 8 with NaOH. After etching, the sections were gently washed with water, dried, coated with gold, and viewed by SEM at various magnifications (50-10 000 x). Otoliths were examined for microincrements. A microincrement was defined as an unbroken incre- mental zone with discontinuous zones as boundaries (Radtke and Dean 1982). Sequential etching made it Radtke and Hourigan Age and growth of Nototheniops nudifrons 559 feasible to enumerate all increments. Individual sec- tions were etched for different lengths of time, with 15-20 minute etching times showing the inner incre- ments most clearly, and shorter etching periods reveal- ing the outer increments. Landmark scratches were placed on each section with an insect needle to follow increments uncovered by different etching times. The otolith was etched in steps of 1 minute and viewed with the SEM after each etching to follow the progression of the smallest increments found. Examination of the three-dimensional microstructure of the left sagitta demonstrated that the midtransverse plane was the best plane for routine sectioning. Table 1 Variance component due to within-fish variance in ages of Nototheniops nudifrotis predicted f •om left and right otolith measurements, compared with variance among fish. Mean Variance df Sums of compo- Vari- source squares squares nent ance Percent Between fish 200 341.24 1.71 0.96 95.87 Within fish 200 6.08 0.04 0.04 4.13 Total 400 347.32 1.00 1.00 100.00 Validation of otolith ageing technique Microincrements are generally assumed to be deposited daily (Pannella 1971, 1974; Radtke and Dean 1982; Radtke and Targett 1984). To test this, individual fish were given a single intramuscular injection of either tetracycline hydrochloride or acetazolamide at 0.025 mg/g body weight. Tetracycline is incorporated into the otolith and, viewed under a compound microscope with filtered reflected wavelengths of 700-800 nanometers, appears as a fluorescent band inside the otolith. Tetra- cycline marks are not visible by SEM. Acetazolamide is a carbonic anhydrase inhibitor which temporarily in- hibits the calcification processes in the otolith resulting in a disruption band which can serve as a reference location for SEM study (Mugiya and Muramatsu 1982, Radtke unpubl. data). Left sagittae from the fish injected with tetracycline were embedded in epoxy-casing resin and serially sec- tioned using a low-speed saw. The sections were polished as above, placed in nonfluorescent immersion oil, and examined under reflected ultraviolet light through a compound microscope. Otoliths with aceta- zolamide marks were prepared for SEM analysis as described above. The number of microincrements from the reference mark to the margin of the otolith was compared with the number of days between injection and the day that the fish was sacrificed. Relationship of otolith dimensions and fish age In an effort to determine other predictors of age in A^. nudifrons, the data generated from otolith and fish measurements were applied to the multiple regression model: Age = a + bii'i + 60X2 + ^: '3 -'3 + h,x., In this model, age in years was determined by count- ing microincrements, a = intercept, h = regression coefficients, x = variables. The data were checked for normality, and the multiple regressions were deter- mined in a stepwise fashion with the inclusion level for variables set at p = 0.05. As a measure of the variance associated with the otolith morphometric procedure, the percent of the total variance component due to variance lietween age estimates from the left and right sagittae of the same fish was calculated. Less than 5% of the total variance was due to variance between sagit- tae of the same fish (Table 1). Results A total of 216 Nototheniops nudifrons were collected in trawls off Low Island. Five were caught during six deep trawls in 100-110 m depth, with the remaining fish caught in four shallow trawls at depths of 54-80 m. Fish captured in trawls ranged in size from 70 to 149 mm SL. More females (A^ = 127) were collected than males (N = 74), and the modal size of females (100 mm SL) was slightly larger than that of males (90 mm SL; Fig. 1). Three smaller juveniles (each 45 mm SL) were collected by hand in 10 m depth at Arthur Harbor. Length-weight relationships Length-weight relationships of males and females are shown in Figiu'e 2. Separate regression equations were calculated for males and females using a gonad-free body weight. Weight was approximately a cubic func- tion of length, indicating nearly isometric growth. The length-weight relationship of the fish conformed to the power curve model Weight = o(SL)*. An analysis of covariance indicated that the slopes of the length- weight relationships for the sexes were not significant- ly different (df = 1, 203; F= 2.30, 0.05 »- U3 ^ to ^ u> ,_ to ,_ to ^_ Ol o o *" *" Cs| OJ en (-1 ■ A^ Q) ."o* 5 6- 0 £lWtt<«^ D jHaJKi ■>-' , ^"SK* ° •— 4- cn . tfifq o .^jlpB CO y''^ 2- y y y o 0 40 80 120 160 1 SL (mm) Figure 4 Relationship of sagittal otolith length to fish standard length (SL) for Nototheniops nudifrons. Figure 5 Relationship of sagittal otolith weight to fish standard length (SL) for Nototheniops nudifrons. (Fig. 4), and sagitta weight was related exponentially to fish length (SL) (Fig. 5). Ground, polished, and decalcified sagittae were examined by SEM. A low magnification view of a sagittal otolith (Fig. 6) shows increments occurring concentrically from the center to the edge. Larger incremental patterns, or bands, were observed, but they could not be attributed to any specific environmental cues. Counts of microincre- ments within these bands showed that they were not true annuli. Sagittae exhibited a well-defined central core region or nucleus (Fig. 7a). Within the core region were concentric microincrements separated from the surrounding otolith by a distinct transition zone (Fig. 7b). Surrounding the core area were mineral crystals in a protein matrix which formed microincrements (Fig. 8a and 8b). Counts of microincrements were made beginning at the core transition zone and ending at the outer edge of the otolith. These microincrements varied in width from about 2 f^m (Fig. 8a) to less than 1 (.(m (Fig. 8b). Differences in increment width may reflect periods of differential growth. Validation of the daily nature of microincrements The periodicity of growth increment formation was ex- amined using tetracycline- and acetazolamide-marked specimens which were held in tanks at ambient light and temperature regimes after injection. After injec- tion of tetracycline or acetazolamide. A'', nudifrons displayed some evidence of short-term stress, including cessation of feeding for 2 or 3 days, but only one mor- tality occurred. Subsequently, the fish survived well in captivity, with 80% of all fish surviving the whole winter. Sagittae of 29 fish injected with tetracycline in February and March 1984 and held 4-34 days under natural daylight photoperiods were examined for post- treatment increments. A discrete fluorescent band was discernible in all specimens, and increments between the band and the margin of the otolith were counted (Table 2). The number of increments averaged 2.5 in- crements less than the number of days after injection. This difference corresponded well to the initial period of stress, and did not change with length of the experi- ment (Median test, 0.2>p>0.1; Siegel 1956). Linear regression analysis (Model I, Sokal and Rohlf 1969) in- dicated that the slope (6 = 1.076, 95% CI of 0.756- 1.396) was not significantly different from 1 (Table 2). The 2/-intercept equaled - 3.27 (95% CI of - 7.73-1.19). Sagittae from five fish injected with acetazolamide in April 1985 were examined by SEM. A discrete dis- ruption of increments caused by the interruption of calcium deposition was discernible (Fig. 9). The number of increments from this disruption mark to the otolith margin averaged 9.2 increments less than the number of days after injection (Table 2). Acetazolamide disrupts metabolic processes and its effects can be expected to last longer than the tetracycline's, since the latter does 562 Fishery Bulletin 88(3), 1990 Figure 6 Low-magnification scanning electron micrograph of a sagittal otolith of Notothenwpf! nvdifrorif^, showing concentric incremental patterns. The sagitta is from a lOO-mmSL specimen. not appear to affect the fish's metabolism. Linear re- gression results indicated a slope of 0.962 (95% CI of 0.983-0.941) and a (/-intercept of -5.40 (95% CI of -2.80 to -8.01). The significance of this analysis is suspect because of the small sample size (A'^ = 5). The discrepancy between postinjection days and microin- crements counted increased from 5 to 12 with length of incubation period. However, this apparent increase in the discrepancy represents an actual percentage decrease in error from 50% to less than 10%. The longest incubation period in the acetazolamide experi- ment was 158 days, which was three times longer than the longest tetracycline incubation. This may explain why in the tetracycline experiment we did not observe an increase in the discrepancy between number of in- crements and days after injection. During the experiment, the photoperiod varied from 9:15 hours Light:Dark, to a minimum of 5:19 L:D in June, to 15:9 L:D in October. The microincrements formed during incubation were uniform. Despite con- siderable variation in photoperiod, we found no indica- tion that microincrements cease to form on a daily basis. Age and growth Microincrements were counted in the sagittal otoliths of 32 fishes. Ages ranged from 1.3 years for the small juveniles collected by hand, to 8.5 years for the largest individual collected. Fish length was related to the number of microincrements in the sagitta (Fig. 10). Length (mm) increased linearly with age (years), and was best described by the equation: SL = 14.87(Age) -t- 23.56; (R- = 0.93, /KO.OOOl) No difference in growth was observed between the sexes in our sample (Fig. 10). When separate regres- sions were calculated for males (A^= 11) and females Figure 7 (a) Scanning electron micrograph of a .sagittal otolith of a 1.30-mmSL Notolkeniops nudifrons, showing the characteristic central core area, or nucleus, separated by a well-defined transi- tional zone from the outer zone with microincrements radiating outward, (b) Microincrements from a 100-mmSL fish found within the core area of the otolith. Radtke and Hourigan: Age and growth of Nototheniops nudifrons 563 564 Fishery Bulletin 88(3), 1990 Radtke and Houngan: Age and growth of Nototheniops nudifrons 565 Figure 8 Scanning electron micrographs of microincrements of sagittal otoliths oi Nototheniops nudifro>is, as utilized for daily increment enumeration, (a) Wide and narrow patterns of microincrements from a 144-mmSL fish, (b) Narrow microincrements from a 70-mmSL fish. (A^=18), analysis of covariance indicated that the slopes of the two lines did not differ between the sexes (df=l, 26; F = 0.82, p>0.37). Relationship of fish size and otolith dimensions to fish age Stepwise multiple regression analysis relating fish size (mm), weight (mg), and otolith morphometries (sagit- ta length in mm and weight in ^g) to age (years), re- sulted in the acceptance of the following model: Age (Yrs) = 3.892 + 0.856(sagitta wt) - 2. 206 (sagitta length) -i- 0.041(SL) -0.025 (fish wt) Addition of the remaining variable (sagitta width) did not improve the regression (N = 32, i?- = 0.967, p<0.05; Table 3). The residuals were randomly dis- tributed, which demonstrated that this multiple regres- sion best explained the variability in age. Over 94% of the variance was explained by the regression of the two otolith measurements (weight and length). The vari- ables in the model, such as otolith length, width, and weight, are not strictly independent variables, nor are length and body weight. This deviation from the as- sumptions of the model may result in an artificially in- flated R'^ value; however, this should not decrease their practical value as predictors. Sagitta weight alone accounted for 92% of the variance in the regression, supporting the usefulness of otoliths as age-related structures. The relationship between age and otolith morpho- metries allowed the estimation of age for all the fish in the sample (Fig. 11). The growth curve derived from the multivariate model (Fig. 11) was comparable to that derived from microincrement counts (Fig. 10), and both were linear in form. Individuals in this population of A'^. nudifrons grow at approximately 15 mm per year, which is a slow growth rate. Analysis of covariance of Table 2 Validation of the daily nature of otolith microincrement deposition in Nototheniops n udifrons. Otoliths of fish injected with tetracycline 1 were examined by light microscope, while those injected with acetazolamide were examined by scanning electron microscope. Number of post-treatment Injection Number of replicates Days after injection microincrements counted Tetracycline 1 4 1 Tetracycline 5 5 2,3,3,3,3 Tetracycline 2 6 3,3 Tetracycline 3 7 4.4,4 Tetracycline 2 9 5,7 Acetazolamide 1 9 4 Tetracycline ^ 5 10 7,7,7,8.10 Tetracycline 1 11 6 Tetracycline 2 14 10,11 Tetracycline 2 15 12,12 Tetracycline 3 16 14,15,16 Acetazolamide 1 21 14 Tetracycline 1 24 23 Tetracycline 2 34 33,34 Acetazolamide 3 1.58 144,147.149 Regression analysis Regression coefficient SE «5 df P Tetracycline b = 1.076 0.074 (/■intercept = -3.268 1.0.36 1.027 27 0.4>p>0.2 Acetazolamide b = 0.962 0.0049 (/■intercept = -5.402 0.606 7.755 3 p<0.01 566 Fishery Bulletin 88 13). 1990 J*. Figure 9 Scanning electron micrograph of a sagitta from a 100-mmSL Nototheniops nudifrons marked with acetazolamide. The acetazolamide mark (arrow) is visible as a disruption in the growth increments. 160n o «° O y 120- ay ••"N o ^o" E - oJSf^ E ^ °o 80- xo _J .# Ul ^ SL = 14.87(AGE) + 23.56 „-- N - 32 40- R"' - 0.93 n- ( ) 1000 2000 3000 ' ( Number of Increments ) 2 4 6 8 10 Age (Years) Figure 10 Regression of fish standard length (SL) on age as determined from the number of sagittal otolith microincrements for Notntheniops nudifrons. Table 3 Associated statistics of one-way regression and stepwise multiple-regression model of fish age in years vs size and weight variables for Nototheniops nudifrons. N = = 32. One -way regression Variable F P>F Model R~ 1 Sagitta weight 314.64 0.0001 0.921 2 Fish length (SL) 213.31 0.0001 0.888 3 Fish weight 134.64 0.0001 0.832 4 Sagitta length 103.35 0.0001 0.793 5 Sagitta width 92.43 0.0001 0.774 Multiple regression Variable Partial Model Step entered F P>F R- R- 1 Sagitta weight 53.68 0.0001 0.921 0.921 2 Sagitta length 23.16 0.0001 0.021 0.942 :5 Fish length (SL) 17.43 0.0003 0.019 0.961 4 Fish weight 4.53 0.0437 0.006 0.967 Radtke and Houngan: Age and growth of Norotheniops nudifrons Sb7 160n /» " ^t. ° X .«<>^ » o J^^ ° ° 120- 'e ^^t E ^-' 80- _l /rf> L^ y / / ^ 40- SL = 15.04(Age) + 24.39 N « 194 R^ = 0.89 f\- u~ 1 1 1 1 1 1 1 1 1 c )123456789 Age (Years) Figure 1 1 Regression of fish standard length (SL) on estimated age for Nototheniops nudij'rons. Age was estimated by a multivariate mathematical model based on fish standard length, fish weight, otolith length, and otolith weight. separate growth curves calculated for males and fe- males showed no significant difference (df= 1,177; F = 0.87; p>0.36), indicating that males and females grow at the same rate. Discussion Otolith microincrement deposition and age determination The accuracy of age and growth estimates depends on the assumption that the microincrements viewed were deposited on a daily basis. Daily increments have been foimd in otoliths from many temperate and tropical fish species (see review by Campana and Neilson 1985). The daily nature of microincrement deposition in Notothe- niops nudifrons was validated through the tetracycline and acetazolamide marking experiments. Daily incre- ment deposition occurred even during the shortest days of the Antarctic winter. The formation of microincre- ments may become less regular in some fishes after sexual maturity (Pannella 1980), but our validation study included both immature subadults and mature males and females, which all deposited daily microin- crements. Hourigan and Radtke (1989) found that daily microincrement deposition also occurs in larval A'^. nudifrons, beginning at or around the date of hatching. Together, these studies provide the first validation of daily otolith microincrements for any Antarctic fish. In addition to the microincrements, the otoliths of N. nudifrons contained larger banding patterns. In temperate regions, environmental factors such as tem- perature and food availability show regular and marked seasonal changes, giving rise to clearly identifiable growth periods in fish otoliths. The regular annual for- mation of these seasonal rings in the otoliths of tem- perate fishes provides a successful and widely used age- ing technique. Growth increments laid down in the otoliths of some Antarctic fishes may be annuli (e.g., Harpagifer bispinis antarcticus; Daniels 1983); how- ever, those observed in this study, as well as those in the otoliths of Nototheniops larseni (Radtke and Targett 1984), are not. Nevertheless, such increments have been used, perhaps erroneously, for ageing pur- poses. Compared with temperate zones, Antarctic habitats undergo smaller fluctuations in temperature, which may result in a lack of distinct annual growth increments. Daily microincrements allow the ageing of Antarc- tic species without otolith annuli. Microincrements are the result of a discontinuous zone which is generally formed under changing light conditions, probably at daybreak (Tanaka et al. 1981). The near-constant light and constant dark conditions experienced by Antarc- tic fishes during the summer and winter, respectively, could potentially interfere with daily deposition pat- terns. However, in the present study, daily increment formation continued despite great variation in photo- period. Although areas in otoliths were detected in which increment width decreased (Fig. 8), none of these appeared to indicate a cessation of increment forma- tion. Slight changes in light intensity, which occur in both Antarctic summer and winter, or endogenous rhythms may be responsible for the daily pattern of otolith deposition. This problem is currently being in- vestigated using Antarctic fishes kept in constant dark or constant light, and under different light intensities (Radtke and Hourigan unpubl. data). Use of a multivariate mathematical model relating age to otolith length and weight and fish size is a simpler method of age determination. The preparation and counting of microincrements in fish over 1 year of age is time-consuming and impractical. However, the multiple regression equation calculated from the 32 in- dividuals analyzed with SEM provided an alternative ageing method. Otolith weight (with its consistent rela- tionship to body length) was the best predictor of age, demonstrating the usefulness of otoliths in such a regression. This method should be applicable to other fish species, serving as a quick means to provide growth information. The dimensions of otoliths have been used to estimate the age of temperate fishes (Templeman 568 Fishery Bulletin 88 13), 1990 and Squires 1956, Boehlert 1985, Radtke et al. 1985). Otolith-age equations are species-specific and possibly population-specific. Using otolith morphometries for ageing will, therefore, require developing the equations for individual species. Nevertheless, in conjunction with ageing techniques using microincrement counts, otolith morphometries offer a simpler method of ageing a large sample of individuals than is possible using more labor- ious techniques. The external features of A^. nudifrons otoliths were distinctive compared with those from eight other Ant- arctic fishes examined (Chaenocephahis aceratus, Champsocephalus gunnari, Harpagifer hispinis antarc- ticus, Notothenia angustifrons, N. gibberifrons, N. larseni, N. corriceps, and Trematomus newnesi; Rad- tke, pers. observ.). Species-specific otolith features of Antarctic fishes have been described by Hecht (1987) and these may be useful in identifying Antarctic fishes from cetacean, pinniped, bird, or fish stomachs (Hecht 1987, North et al. 1984). Otolith morphological char- acteristics may also provide information on taxonomic relationships among Antarctic fishes (Hecht 1987), as they have for other fish groups (Hecht 1978, Hecht and Hecht 1978, Morrow 1979). Life history of Nototheniops nudifrons Nototheniops nudifrons is a slow-growing, relatively long-lived fish. Subadults and adults of both sexes were found in shallow-water (< 100 m) benthic habitats along the Antarctic Peninsula, as reported by other re- searchers (Targett 1981, Daniels and Lipps 1978, Kellermann 1986). The absence of individuals smaller than 75 mm SL from trawls probably represented selection by the fishing gear for larger individuals. The growth curves determined by body length vs. microin- crement count, as well as by body length vs. age cal- culated from the multivariate relations of otolith length and weight, were similar and linear in shape between 45 and 149 mm SL. Linear growth was observed even though the data included a large size-range of fish, with individuals near the largest sizes reported for this spe- cies. This type of linear growth, with no marked slow- ing toward an upper asymptote, has not been reported for other Antarctic fishes, although the growth curve oi Harpagifer bispinis antarcticus, another small Ant- arctic fish, is nearly linear (Daniels 1983). Males and females had similar growth rates. This was surprising, since mature males and females follow different strategies. Males and females reach sexual maturity at the same age (4-5 years); however, there- after females invest much more energy into gonadal growth than do males (Hourigan and Radtke 1989). All parental care is provided by the males, which may spend at least 4 months in nest defense of a single clutch of eggs (Hourigan and Radtke 1989). This probably requires increased energy costs, and may limit food intake by constraining the male to foraging near the nest. These two strategies may entail similar energy expenditures, resulting in similar somatic growth rates. Estimation of mortality rates is central to the deter- mination of demographic parameters (Gulland 1955). It is possible to estimate natural mortality (Z) from the data on trawl catches and age for this population of A^. nudifroyis. The most widely used method to estimate mortality is the Beverton and Holt mortality estimator which uses the von Bertalanffy growth curve (Bever- ton and Holt 1956). The age-length data for the 32 fish were fitted to the von Bertalanffy growth curve, and the parameters and their confidence limits are given in Table 4. These values are primarily useful for com- parative purposes, since the actual shape of the growth curve approximated a linear relationship. Natural mor- tality, calculated from the von Bertalanffy parameters and derived from the sample of 32 fish of known ages, yielded a value of Z = 0.68 (Table 4). When Z was calculated using the von Bertalanffy growth param- eters for the larger sample of 212 fish with predicted ages, Z = 0.66. These estimates assume that the fish grow according to a deterministic von Bertalanffy growth curve. Since the present sample was more ac- curately represented by a linear growth curve, this assumption was violated. An independent mortality rate was calculated using the actual age estimates from the otolith morphometries. In this estimate, the mor- tality rate is the slope of the log-survivorship curve (Ricker 1975) and was calculated to be 0.89. Kock et al. (1985) provide a summary of mortality estimates for larger Antarctic fishes of commercial in- terest. Both estimates of instantaneous mortality of A'^. nudifrons were higher than these estimates, which ranged from 0.20 to 0.35. Predation rates may be higher on A'', nudifrons than on larger species. How- ever, Daniels (1983) estimated the mortality of the still smaller Harpagifer bispinis antarcticus to be 0.22 and 0.25 for males and females, respectively (a recalcula- tion of his data for both sexes combined, using the same methods as for A'^. nudifrons, resulted in a value of 0.223). Although the high mortality rates for A^. nudfrons accurately describe our sample, they may be subject to sampling errors. Our estimates were based on a single sample captured on two occasions 1 month apart. Both mortality calculations assume that the sampled population is age-stationary (i.e., the age distribution remains the same from year to year), an assumption which is frequently not met. Mortality estimates will also be biased if certain age classes larger than Lc (length at first capture) are less subject to capture. Our Radtke and Hourigan: Age and growth of Nototheniops nudifrons 569 Table 4 Parameter estimates derived from the Von Bertalanffy growth equation (L^ constrained to 160 mm SL), natural mortality rates, and associated parameters for Nototheniops nudifroiis collected from the Antarctic Peninsula. Parameter SL vs. age in days SL vs. age in years Estimate Confidence limits Estimate Confidence limits K K From otoliths (A^ = 32) 0.0005 0.00049-0.00061 0.200 2.20 -158.10-162.50 0.006 From predicted ages (N = 194) 0.00048 0.00047-0.00050 0.175 -40.848 -94.681-12.986 0.112 0.000-0.223 -0.434-0.445 0.172-0.183 -0.259-0.036 Lr k Z Mortality estimates (Beverton and Holt 1956) All Male Female 101 91 101 113.1 106.6 112.5 0.175 0.175 0.175 0.678 0.599 0.73 From estimated ages where the mortality rate Z is the slope of the log-survivorship curve (Ricker 1975) Z 0.894 0.807 1.26 sampling caught significantly more females than males (x^, p<0.05), and the modal size of males was smaller. This might mean that there are fewer large males in this population. Alternatively, larger males may be less subject to capture in spring, perhaps because of dif- ferent habitat preferences of nesting males. This would skew mortality estimates upwards. The life history of A'^. nudifrons is characterized by slow growth and a relatively long life for a small fish. Age at sexual maturity is late (4-5 years), and fecun- dity is low (Hourigan and Radtke 1989). Eggs take several months to hatch, followed by a pelagic larval stage of several months duration (Kellermann 1986, Hourigan and Radtke 1989). It has become clear that these life-history traits are the norm for Antarctic fishes (Daniels 1983, Kock et al. 1985). They combine to produce species that are vulnerable to overfishing and that recover slowly when stocks are depleted. In- deed, catches of both Notothenia rossii and A^. gib- berifrons on the shelves of certain islands of the Scotia Sea decreased drastically after several years of fishing (Kock et al. 1985). The methods used in the present study may provide more accurate estimates of the ages of Antarctic fishes, and facilitate management of Ant- arctic fisheries. Acknowledgments We would like to thank the staff of Palmer Station, Antarctica, in particular the winter 1985 scientists and support personnel, without whose assistance this re- search would not have been possible. J. Archie, J. Bell, S. Folsom and T. Telecky provided valuable assistance. Editorial assistance was given by P. Lenz and B. Tilley. The manuscript has benefited from the comments of A. Kellermann, P. Lenz, S. Ralston, and two anony- mous reviewers. This research was supported by NSF Grants OCE 84-15968, OCE 88-00686, DPP 85-21017, and DPP 88-16521. Citations Beverton. R.J.H., and S.J. Holt 1956 A review of methods for estimating mortality rates in fish populations, with special reference to source of bias in catch sampling. Rapp. P.-V. R^un. Cons. Perm. Int. Explor. Mer 140:67-83. Boehlert, G.W. 1985 Using objective criteria and multiple regression models for age determination in fishes. Fish. Bull.. U.S. 28:103-117. Burchett, M.S., A. DeVries, and A.J. Briggs 1984 Age determination and growth oi Di^sostwkvs mawsoni (Norman, 1937) (Pisces, Nototheniidae) from McMurdo Sound (Antarctica). Cybium 8:27-31. 570 Fishery Bulletin 88(3). 1990 Campana. S.E., and J.D. Neilson 1985 Microstructure of fish otoliths. Can. J. Fish. Aquat. Sci. 42:1014-1032. Chojnacki. J., and P. Palczewski 1981 Age and growth oi Pseudochaenichthys georgianiis Nor- man 1937 (family Chaenichthyidae) of the South Georgia Region. Pol. Polar Res. 2(3-4):145-157. Daniels, R.A. 1983 Demographic characteristics of an Antarctic plunderfish, Harpiigifer bispinis antarcticws. Mar. Ecol. Prog. Ser. 13: 181-187. Daniels. R.A., and J.H. Lipps 1978 Predation on foraminifera by Antarctic fish. J. Forami- niferal Res. 8:110-113. DeWitt, H.H. 1971 Coastal and deep-water benthic fishes of the Antarc- tic. Antarct. Map Folio Ser. 15. Am. Geophys. Union, 10 p. Everson, I. 1970 The population dynamics and energy budget of Noto- thenia neglectu Nybelin at Signy Island. South Orkney Islands. Br. Antarct. Surv. Bull. 23:25-50. 1980 M^thodes de determination de I'age chez les poissons an- tarctiques. Cyhium 4(4):41-59 [in French. Engl, abstr.]. Freytag, G. 1980a Length, age and growth of Notothfnia rossii marmorata Fischer 1885 in the west Antarctic waters. Arch. Fische- reiwiss. 30:39-66. 1980b Problems in ageing Notnthenin rossii mnrmorata. Cybium (3e serie) 8:43-51. GuIIand. J. A. 1955 Estimation of growth and mortality in c(jmmercial fish populations. Fish. Invest. Ser. II Mar. Fish. G.B. Minist. Agric. Fish. Food 18(9):l-46. Hecht. T. 1978 A descriptive systematic study of the otoliths of the Neopterygean marine fishes of South Africa. Part I. Introduc- tion. Trans. R. Soc. S. Afr. 43:191-197. 1987 A guide to the otoliths of southern ocean fishes. S. Afr. J. Zool. 17:1-87. Hecht, T., and A. Hecht 1978 A descriptive systematic study of the otoliths of the Neopterygean marine fishes of South Africa. Part II. The delimitation of teleost orders, some systematic notes and a pro- visional new phyletic order sequence. Trans. R. Soc. S. Afr. 43:199-218. Hourigan, T.F., and R.L. Radtke 1989 Reproduction of the Antarctic fish Notutheniops nudi- frons. Mar. Biol. 100:277-283. Hureau, J.C. 1966 Biologie de Chaenichthys rh in ore rat us Richardson et probleme du sang incolore de Chaenichthyidae, poissons des mers australes. Bull. Soc. Zool. Fr. 91(4):735-751 (in French, Engl, abstr. (. Jones, C. 1986 Determining age of larval fish with the otolith increment technique. Fish. Bull., U.S. 84:91-103. Kellermann, A. 1986 On the biology of early life stages of Notothenioid fishes (Pisces) off the Antarctic Peninsula. Ber. Polarforsch. 31:1-155 (in German]. Kochkin, P.N. 1982 A study of the age and growth of Cha mpsucephiduii gun- nari Lonnberg (Chaenichthyidae). Vopr. Ikhtiol. 22(3): 390-400. Kock, K.H. 1981 Fischereibiologische Untersuchungen an drei antarkti- schen Fischarten: Champsocephatus gunnari Lonnberg, 1905, Chuenocephalus aceratus (Lonnberg, 1906) and Pseudocfmenich- thys georgimms Norman, 1937 (Notothenioidei, Channichthyi- dae). Mitt. Inst. Seefisch. 32, 226 p. (in German]. Kock, K.H., G. Duhamel. and J.C. Hureau 1985 Biology and status of exploited Antarctic fish stocks: A review. Biomass Sci. Ser. 6, 143 p. Morrow, J.E. 1979 Preliminary keys to otoliths of some adult fishes of the Gulf of Alaska, Bering Sea, and Beaufort Sea. NOAA Tech. Rep. NMFS Circ. 420, 32 p. Mucha, M. 1980 Characteristics of South Georgia icefish (Pseudochaen ich- thys georgianus Norman) from the region of South Georgia Island (Antarctic) in the years 1977-1979. Pol. Polar. Res. 1(4):163-172. Mugiya, Y., and J. Muramatsu 1982 Time-marking methods for scanning electron microscopy in goldfish otoliths. Bull. Jpn. Soc. Sci. Fish. 49:1225-1232. Mugiya, Y., and N. Watabe 1977 Studies of fish scale formation and resorption-II. Effect of estradiol on calcium homeostasis and skeletal tissue resorji- tion in the goldfish, Carassiiis auratus, and the killifish, Fun- dulus h/'terorlitus. Comp. Biochem. Physiol. 57(A):197-202. North, A.W., M.G. White, and M.S. Burchett 1980 Age determination of Antarctic fish. Cybium (3e s^rie) 8:7-11. North, A.W., M.S. Burchett, C.J. Gilbert, and M.G. White 1984 Identification of fish from the southern Ocean by means of otoliths. Br. Antarct. Surv. Bull. 62:83-94. Olsen, S. 1954 South Georgian cod, Notothenia rossii mnrmonitit Fischer. Norsk Hvalfangsttid. 43(7):373-382. 1955 A contribution to the sytematics and biology of Chaenich- thyid fishes from South Georgia. Nytt. Mag. Zool. (Oslo) 3:79-93. Pannella, G. 1971 Fish otoliths: Daily growth layers and [leriodical ()at- terns. Science (Wash. DC) 173:1124-1127. 1974 Otolith growth patterns: An aid in age determination in temperate and tropical fishes. In Bagenal, T.P. (ed.), The age- ing of fish, p. 28-39. Unwin Brothers, Surrey. 1980 Growth patterns in fish sagitta. In Rhoads. D.C., and R.A. Lutz (eds.). Skeletal growth of aquatic organisms: Bio- logical records of environmental change, p. 519-560. Plenum Press, NY. Radtke, R.L. 1984 Cod fish otoliths: Information storage structures. Flv^de- vigen rapp. 1:273-298. Radtke, R.L., and J.M. Dean 1982 Increment formation in the otoliths of embryos, larvae, and juveniles of the mummichog, Fundidus heteroclihis. Vxsh. Bull.. U.S. 80:201-215. Radtke, R.L., and T.E. Targett 1984 Rhythmic structural and chemical patterns in otoliths of the Antarctic fish Notothmiii tnrseni: Their application to age determination. Polar Biol. 3:203-210. Radtke, R.L., M.L. Fine, and J. Bell 1985 Somatic and otolith growth in the oyster toadfish (Op- saiiiis tan L.). J. Exp. Mar. Biol. Ecol. 90:2.59-275. Radtke, R.L., T.E. Targett, A. Kellermann, J.L. Bell, and K.T. Hill 1989 Antarctic fish growth: Profile of Tremntomun newnesi. Mar. Ecol. Prog. Ser. 57:103-117. Radtke and Hourigan: Age and growth of Nototheniops nudifrons 571 Ricker, W.E. 1975 Computation and interpretation of biological statistics of fish populations. Bull. Fish. Res. Board Can. 191:1-382. Scherbich, L.V. 1975 (1n the age reading methods and the onset of maturity in marbled notothenia A'o^o^/ipw.ia rossii marmorata. Vopr. Ikhtiol. 15(1):94-100. Shust, K.V., and I. A. Pinskaya 1978 Age and rate of growth of six species of Notothenid fish (family Nototheniidae). J. Ichthyol. 18(5):743-749. Siegel, S. 1956 Nonparametric statistics for the behavioral sciences. McGraw-Hill, NY. 312 p. Six, L.D., and H.F. Horton 1977 Analysis of age determination methods for yellowtail rockfish, canary rockfish, and black rockfish off Oregon. Fish. Bull.. U.S. 75:40.5-414. Sokal, R.R., and F.J. Rohlf 1969 Biometry. W.H. Freeman, NY, 859 p. Sosinski, J. 1981 Biologia porownawcza Kergulen {Champsocephalus gan- tiari) Zrejonow Antarktyki. Wydaw. Morsk. Inst. Ryb. Stud. Mater. (Sena 3) 48:1-91 [in Polish]. Tanaka, K., Y. Mugiya, and J. Yamada 1981 Effects of photoperiod and feeding on daily growth pat- terns in otoliths of juvenile Tilapia mlotica. Fish. Bull., U.S. 79:459-466. Targett. T.E. 1981 Trophic ecology and structure of coastal Antarctic fish communities. Mar. Ecol. Prog. Ser. 4:243-263. Templeman, W., and H.J. Squires 1956 Relationship of otolith lengths and weights in the had- dock Melanogrammus aeglejtmis (L.) to the rate of growth of the fish. J. Fish. Res. Board Can. 13:467-487. Townsend, D.W. 1980 Microstructural growth increments in some Antarctic fish otoliths. Cybium (.3e serie) 8:17-22. Wohlschlag, D.E. 1961 Growth of an Antarctic fish at freezing temperatures. Copeia 1961:11-18. 1962 Antarctic fish growth and metabolic differences related to sex. Ecology 43:589-.597. Abstract. — a study was con- ducted on shallow banks near Lee Stocking Island, Bahamas, to deter- mine habitat associations of queen conch Stromhus gigas L. within sea- grass meadows of Thalaiisia testudi- num. Transect data showed that conch density and biomass increased directly with increasing macrophyte cover up to an optimal level of moderate-density seagrass (608-864 shoots/m^), after which conch den- sity decreased sharply. Up to the op- timal level, conch density and bio- mass were closely correlated with seagrass and detritus biomass, shoot density, and depth. Seagrass shoot density was the best predictor of conch abundance. Results of habitat choice experiments showed that two different juvenile size classes (75-100 mm and 12.5-150 mm shell length) are proficient in detecting and choos- ing habitats with moderate seagrass density over bare sand, low or high seagrass density. Larger conch ap- pear to prefer habitat with higher shoot density than conch of smaller size classes; adult conch were less specialized in their habitat associa- tion. These results provide data on one of the key environmental vari- ables, seagrass structure, which will be useful in predicting conch distri- butions in the field and for planning conch outplanting. Distribution and Behavior of Queen Concli Strombus gigas Relative to Seagrass Standing Crop Allan W. Stoner Janice M. Waite Caribbean Manne Research Center, 100 E 1 7th Street Riviera Beach, Florida 33404 and Lee Stocking Island, Exuma Cays, Bahamas Manuscript accepted 14 May 1990, Fishery Bulletin, U.S. 88:.573-58.5. Seagrass meadows, common in pro- tected shallow waters, are known to be important sources of food and shelter for numerous fishes and in- vertebrates from high to low lati- tudes (Kikuchi 1980, Ogden 1980, Virnstein et al. 1984). Living sea- grasses are consumed by a few animals such as surgeonfishes, halfbeaks, parrotfishes (Randall 1964, Ogden 1980), urchins (Law- rence 1975), green sea turtles (Ogden 1980), and sirenia (Bertram and Ber- tram 1968); however, most nonpre- daceous fauna of seagrass meadows consume seagrass epiphytes or detritus (Zimmerman et al. 1979, Leber 1983, Fry 1984, Howard 1984, Kitting et al. 1984, Van Montfrans et al. 1984). Sheltering from predation has been shown experimentally for many seagrass inhabitants, par- ticularly crustaceans (Nelson 1979, 1981; Heck and Thoman 1981; Coen et al. 1981; Stoner 1982; Leber 1985; Wilson et al. 1987). The queen conch Strombus gigas L. is a large gastropod mollusc which derives both food and shelter from seagrass beds of the Caribbean region (Bermuda and southeastern Florida to Brazil). After the first year of life, of which little is known, queen conch juveniles emerge from the sedi- ment and feed epibenthically in sea- grass meadows where there are large quantities of algal and detrital foods (Randall 1964). Juvenile queen conch are important grazers of seagrass detritus, and conch densities may be limited by food abundance (Stoner 1989). Recent experiments conducted in the Bahamas have shown that seagrasses reduce predation rates on juvenile conch in nursery areas (Mar- shall and Lipcius, In review). Throughout its geographic range, the queen conch is an important commercial species which has been severely depleted in many areas (Adams 1970, Brownell et al. 1977). For this reason, there is increas- ing interest in improving the man- agement of conch, and seeding of depleted areas with hatchery- reared juveniles has been proposed (Berg 1976, Brownell and Stevely 1981, Laughlin and Weil 1983, Coulston etal. 1987). Successful management of the fishery and/ or outplanted populations will re- quire detailed information on the relationships between queen conch and environmental variables. This study was conducted to examine the role of macrophyte cover in the distribution of juvenile conch in two nursery areas of the Exuma Cays, Bahamas. In addition to sea- grass shoot density and biomass, abundance of detritus, sediment characteristics, and depth were con- sidered for their potential influence on the density, distribution, and biomass of conch. Field experiments were conducted to determine the role of habitat selection in natural conch distributions. 573 574 Fishery Bulletin 88(3), 1990 Table 1 Habitat characteristics of the two seagrass meadow study sites. Units for each are: depth (meters at low water), sediment grain size 1 (0), organics (% dry wt), Thala ssia shoots (ri/m'-), Thalassia and detritus biomass (g dry wt/m-). Values are means ± SD. n = number of measurements from three transects. Stations 1 2 3 4 5 6 7 Children's Bay Cay Depth {n = 3) 1.0±0.1 1.2±0.1 1.8±0.2 2.7 + 0.3 3.4 + 0.3 — — Sediment grain size {n = 3) 1.50±0.11 1.38 + 0.31 1.50±0.07 1.83±0.59 2.44±0.51 — — Organics (n = 3) 2.44 + 0.20 2.35 ±0.25 2.61 ±0.06 2.98±0.44 3.88 ±0.45 - - July Thaldssia shoots (n = 6) 0.0±0.0 64. 0± 16.0 304 ±48.0 576 ±48.0 704 ±80.0 — — Thahu-isia biomass (m = 6) 0.0±0.0 3.5±1.1 16.0±4.7 48.0 ±4.8 65.6±14.4 — — Detritus (n = 6) 0.0±0.0 0.80 + 0.80 3.68 ±1.9 .30.9 ±18.6 104±44.8 - - February Thahi.tsia shoots (« = 6) 0.0 ±0.0 67. 2± 19.2 304 ±72.0 736 ± 208 864 ±160 — — Thalassia biomass (« = 6) O.OiO.O 8.6±9.9 13.9±4.6 56.0±20.8 80.0 ±19.2 — — Detritus (?? = 6) 0.0±0.0 0.64 ±0.80 0.64 ±1.28 Shark Rock 48.0 ±36.8 125±96.0 " Depth (n = 3) 1.6 + 0.1 1.7 + 0 2.0±0 2.5 + 0.1 2.8±0.1 3.3 + 0.3 3.2 ±0.1 Sediment grain size {n = 3) 1.58±0.06 1.68±0.21 1.96±0.18 2.00±0.13 2.20 ±0.35 2.57±0.01 2.53 ±0.10 Organics (n = 3) 2.44 + 0.03 2.33 ±0.22 2.34±0.12 2.70±0.11 3.21 ±0.22 3.57 ±0.46 2.85 ±0.47 July Thatnxsia shoots (n = 6) 0.0 + 0.0 72. 0± 12.8 256±44.8 528 ±46.4 669 ±56.0 741 ±88.0 728 ±123 Thaln.isia biomass (h = 6) 0.0±0.0 5.28±1.4 20.3±4.6 54.4 ±8.0 62.4 ±14.4 83.2 ±1.6 97.6±20.8 Detritus (w = 6) 0.0±0.0 0.48 ±0.64 3.2±2.7 13.8±13.1 64.0 ±22.4 96.0 ±40.0 176±41.6 February Thaliissia shoots {n = 6) 0.0±0.0 68. 8± 19.2 292 + 81.6 452 ±54.4 608 ±101 816±217 764 ±128 Thahissia biomass (n = 6) 0.0±0.0 6.9±4.8 36. 8± 17.6 68.8 ±33.6 78.4 ±14.4 80.0±9.6 115 ±36.8 Detritus (» = 6) 0.0 + 0.0 8.0 + 1.1 1.3±1.9 14.7±15.4 157 ±35.2 84.8 + 4.8 309 ±174 Site description This study was conducted in the Exuma Cays, Baha- mas, at two sites characterized by turtlegrass Thalas- sia testudinum Konig and known juvenile queen conch populations (Wicklund et al. 1988). The first site (CBC) was located 1.5 km west of Children's Bay Cay (23° 44.3'N, 76°04.5'W). The second site (SR) was located just southwest of Shark Rock (23°45.0'N, 76°07.5'W), a rock outcrop off the southern end of Norman's Pond Cay. There is a gradient in depth at both sites from a shallow sand bar (< 1 .0 m) to a depth of approximately 3.5 m, accompanied by a gradient in seagrass and detritus biomass (Table 1). Tidal range in the area is approximately 1 m and both sites are subject to strong reversing tidal currents, sometimes exceeding 50 cm/ second. Conch density was highest (~2.0/m-), near the popu- lation centers, decreasing in density towards their peripheries. At each site three transects were made, each separated by approximately 50 m. Divers were towed by a small boat around each site to place buoys on the edges of each population. Transects were then positioned through the population centers, parallel to the depth gradient. Transect lengths were an average of 61 m at CBC and 142 m at SR. Stations marked by stakes were placed along each transect (5 for CBC and 7 for SR) from bare sand to the highest density sea- grass. Thalassia shoot density was used to set all sta- tions of equivalent number in similar habitats for both sites. Station 1 of each transect was placed on bare sand, with no seagrass or detritus. Stations 2, 3, 4, and 5 were placed in the following approximate shoot den- sities, respectively: 70, 300, 550, and 700 shoots/m^. Stations 6 and 7 at SR were characterized primarily by increasing accumulated detritus and higher seagrass biomass and not increasing shoot density. Methods and materials Field measurements Counts were made in July 1988 and February 1989 to test for seasonal differences in conch distribution. In February at CBC we found that the center of the population had moved north of its July location, and Stoner and Waite Habitat associations of Strombus gigss within seagrass meadows 575 SO transect 1 (the southernmost transect) was moved between transects 2 and 3 for the February samphng. Measurements, including station depth, conch density and shell lengths, macrophyte shoot density and biomass, and amount of macroscopic detritus, were taken at each station along the transects within a cir- cle of 2.5 m radius around each stake. Sediment grain- size and organic content were measured in July 1988. Depths were measured at low water using an elec- tronic depth sounder on a 17-foot Boston Whaler, ad- justing for the depth of the transducer below the sur- face. Using scuba, conch were gathered within each 2.5-m circle at each station, counted, and measured for shell length using large calipers. A separate collection of 80 conch ranging from 17 to 196 mm in siphonal length was made to generate a length- weight curve. After freezing, the conch were extracted from their shells, washed to remove feces, blotted to remove excess water, and weighed. The following equation describes the significant correlation (r2 = 0.987, F = 5233.4, /XO.OOl) between shell length and wet weight: logio (wet weight) = 3.403 x logi„ (length) - 5.569 The equation was used to convert length-frequency data to biomass values. Two replicate samples of macrophytes and macro- scopic detritus (mostly senescent seagrass blades and debris) were collected at each station by haphazardly placing a quadrat with 25-cm sides within each circle. The number of shoots within the quadrat were counted and then collected along with the macrophyte detritus into nylon bags (3.0-mm mesh). Thalassia blades (using only aboveground parts), and macrodetritus were separated in the laboratory and dried at 80°C for ap- proximately 24 hours to constant weight. Dry-weight biomass was determined for the individual components. A core of 40 mm diameter, with a penetration depth of 5 cm, was taken for determination of sediment grain- size distribution and organic content. A subsample of approximately 100 g wet weight was dried at 80°C to constant mass and incinerated at 550°C for 4 hours to determine sediment organic content. Organic content was quantified as the percent difference between dry weight and ash-free dry weight. Another sediment sub- sample of approximately 50 g was used to determine sediment grain-size. After washing to remove salts and to extract the silt-clay fraction, sand-sized particles were analyzed using standard Ro-Tap procedures. Silt- Reference to trade names does not imjily endorsement by tlie National Marine Fisheries Service, NOAA. clay fractions were analyzed using standard pipet pro- cedures (Folk 1966). Product moment statistics were generated for mean grain size and sortedness (McBride 1971). Habitat preference experiments Experimental manipulations of plots of seagrasses were made at the Shark Rock site to test for habitat preference in conch of various size classes. Cages 3.6 m diameter (10.2 m^) were constructed of black plastic mesh with 20-mm openings. The cages were 25 cm high and supported by 1-m long pieces of reinforcement bar driven into the sediment. The bottom edge of the screen was pushed into the sediment approximately 3 cm to prevent escape of conch. No losses occurred dur- ing the experiments. Four cages were built in locations similar to station 5 (see above), characterized by Thalassia shoot den- sities of 700 shoots/m- (Moderate Density). Two cages were built near stations 3 with approximately 300 shoots/m- (Low Density), and two cages were built in high biomass habitats similar to station 7 with approx- imately 750 shoots/m- plus a heavy layer of detritus (High Density). Macrophytes in two of the moderate- density cages were manipulated so that one-half of each was reduced to 300 shoots/m^ (Low Density) (Table 2). This was accomplished by placing a rope across the cage through the center and parallel to the direction of flood tide current. (Manipulations were oriented with the current because juvenile conch at the study site are known to move either up or down current during migration.) Then shoots were systematically pulled from one-half of the cage to achieve the desired shoot density. The two cages for this Moderate/Low treat- ment were manipulated to provide mirror images of one another. The other two moderate-density plots were manipulated in a similar fashion to provide Moderate/Sand treatments, where all seagrass shoots were removed from one side of each cage. Replicate Low/Sand treatments were constructed by manipu- lating the two low-density plots. High/Moderate treat- ments were set up by manipulating the high-density plots. About 80% of the loose detritus (comprised most- ly of senescent seagrass blades) was removed from half of the high-density plots. Final shoot densities in the high- and moderate-density areas were similar (656 and 672 shoots/m^, respectively), and removal of detritus yielded a macrophyte environment similar to the moderate-density areas (Table 2). The manipulations required maintenance such as removing new growth and accumulated detritus. Except for the removal of detritus which accumulated along the edges of the cages, maintenance of the plots was performed be- tween the short habitat preference runs. 576 Fishery Bulletin 88(3). 1990 Table 2 Summary of biological conditions presented in experiments :)f habitat preference ir queen conch. There were no significant differences between plots in any of the four treatment levels (ANOVA,p>0.05). Seagrass Seagrass shoots biomass Detritus Treatment n (nlm') (g dry wt/m^) (g dry wt/m- Sand 8 0±0 0±0 0±0 Low Density 8 272 + 32 27 ±13 3.0 ±3.5 Moderate Density 12 672 ±80 72 ±20 58 ±37 High Density 4 656 + 64 98±11 269 ±66 Between 27 September and 28 October 1988 three conch size-classes were tested for their preferences for habitat type. The classes included animals between 75 and 100 mm shell length (~1 year old), conch between 125 and 150 mm (2 and 3 years old), and adults with the fully developed flare of the shell lip (~4 years or more). All were collected from the seagrass beds aroimd the experimental enclosiu'es. To provide a conch density similar to that in nature, 10 juveniles of a size- class were added to each enclosure (1.0 conch/m-) after removal of all large invertebrates such as urchins, hermit crabs, and conch. For runs with the adults, 8 animals were placed in an enclosure. Preliminary experiments showed that the distribu- tion of conch in both Moderate/Low and Moderate/Sand treatments did not change significantly between 1 and 5 days; therefore, all runs were terminated at 3 days. At the end of a run, the numbers of conch found on each side of the cage were recorded, and the animals were removed. Because the animals touching the screen may have been moving around the periphery of the enclosure, they were not included in the counts. Three or four runs were made for each size-class and treatment, except those in the High/Moderate en- closures where two runs were made in each cage. No significant differences were found in the distribution of animals in a size-class in any of the pairs of cages providing similar combinations of macrophytes (Fisher's Exact Test, p>0.05); therefore, the paired cages were considered to be replicates. Results Station characteristics Average depths for the five stations at CBC increased from station 1 (1.0 m) to station 5 (3.4 m). Average sta- tion depths for SR increased from .station 1 (1.6 m) to station 6 (3.3 m) and then decreased slightly at station 7 (3.2 m). Thalassia shoot density, biomass, and detritus biomass all increased with increasing station number and depth, with a few exceptions (see Table 1). At CBC shoot density ranged from 0.0 to 864 shoots/m-, Thalassia biomass from 0.0 to 80 g dry wt/m^, and detritus biomass from 0.0 to 125 g dry wt/m^. At SR shoot density ranged from 0.0 to 765 shoots/m-, Thalassia biomass from 0.0 to 115 g dry wt/m^, and detritus biomass from 0.0 to 309 g dry wt/m^. Sediment grain-size at the CBC transects was rela- tively constant (1.38-1.83 0) between stations 1 and 4, but finer at station 5 (2.44 0). Sediment organic content increased slightly over the transects from low- biomass to high-biomass stations. At the SR transects mean grain diameter decreased (increasing 0) from stations 1-6, while stations 6 and 7 had nearly equal grain size (2.57 and 2.53 0). Organic content was high- est at SR station 6, but values ranged only from 2.33 to 3.50% of dry sediment weight. Children's Bay Cay and the Shark Rock site showed highly significant correlations among all of the vari- ables: water depth, seagrass shoot density, above- ground seagrass biomass, and standing crop of macro- detritus (Table 3). Weakest correlations (r = 0.586 to 0.631) occurred between detritus and the other vari- ables at Shark Rock in February, when abundance of detritus was low. Correlations between macrophyte and depth characteristics remained highly significant when all sites, stations, and dates were included in regression models (Table 3) suggesting that the rela- tionships are relatively universal in the study area. The highest correlation occurred between seagrass shoot density and water depth (?• = 0.930), and weakest cor- relations were found between detritus biomass and the other variables. Concin density patterns At CBC there was an increase in conch density in both July and February along the transects from station 1-5, except for a small decrease in February from sta- tion 4 to station 5 (Fig. 1 ). Densities in July were similar Stoner and U/aite Habitat associations of Strombus gigas within seagrass meadows 577 Table 3 Pearson correlation matrices show ng the relationships among depth and macrophyte variables at the seagrass meadow study sites. The upper and lower sections of the first two matrices are for July and February collections, respectively. The bottom matrix is for all stations and dates combined. Shoots | = density of T. testudinum shoots Thalassia = biomass of T. testudinum: Detritus = biomass of macrodetritus. Depth Shoots Thalassia Detritus Children's Bay Cay Depth — 0.966** 0.938** 0.911** Shoots 0.968** — 0.961** 0.817** Thalassia 0.873** 0.880** — 0.839** Detritus 0.728** 0.7.52** Shark Rock 0.826** — Depth — 0.969** 0.964** 0.839** Shoots 0.950** — 0.954** 0.769** Thalassia 0.877** 0.891** _ 0.889** Detritus 0.6.31* 0..586* Combined sites and dates 0.623* — Depth — Shoots 0.930** — Thalassia 0.871** 0.887** _ Detritus 0.648** 0.628** 0.726** *p<0.05: *• p<0.01 CHILDREN'S BAY CAY 50 45 40 ib 20 25 20 15 10 5 0 50 45 40 35 30 25 20 15 10 5 0 cn Numbers FEBRUARY ^3 Biomass 1 ih 2 3 4 STATIONS CZl Numbers JULY [ SSS Biomass r-l P^ ;; rinfll r^c^ 1 '' _.2 1 1 * 1100 1000 900 800 700 500 500 o O z o I 400 300 200 100 0 o > \ CO 1100 ^ 1000 r^ m 900 ^ aoo -^ 700 2. 600 s 500 -^ 400 300 200 100 0 Figure I Number of conch and conch biomass sampled within a 2.5-m radius circle at Children's Bay Cay, July 1988 and February 1989, at each station. Values are means + SD. to those in February at stations 1, 2, and 3. Stations 4 and 5 had densities almost twice as high in July as those in February, At SR, the general pattern was an increase in conch density from station 1-5 (except a small decrease in July at station 4) (Fig, 2). In July, the density then decreased sharply at stations 6 and 7. In February, the density increased further at sta- tion 6 and then decreased at station 7. More pronounc- ed than at CBC, the densities at SR in February at sta- tions 3, 4, and 5 were lower than the July densities. After log transformation of the data to improve heterogeneity in the variance, a multi-way ANOVA was used to test for differences and interactions in conch density between sites, dates, transects, and sta- tions (Table 4). Transects were examined as blocks, and hence there were no interaction terms with transect. Since SR had 7 stations as opposed to 5 at CBC, sta- tions 6 and 7 at SR were not used in this analysis where CBC and SR were compared. There were no significant interaction terms or block (transect) effects in the ANOVA (p>0.05); therefore, the effects of date, site, and station may be interpreted directly. It is clear from Figures 1 and 2 that conch densities were higher in July than in February at both CBC and SR at most stations. Significant site effects appear to be related to the low numbers of conch at SR, particularly in February and at station 1-3. Sta- tion effects occurred because of the increase in conch 578 Fishery Bulletin 88(3). 1990 SHARK ROCK 50 45 4-0 35 30 25 20 15 10 5 0 50 45 40 35 30 25 20 15 10 5 0 ' ' Numbers r^NN Biomass JULY 2400 2200 2000 1800 1600 1400 1200 o 1000 ° 800 600 400 200 0 o I o > in > CD Numbers FEBRUARY E9 Biomoss A^ f\^ i ^i^ 2 3 4 5 STATIONS 2400 -D 2200 S 2000 - ■1800 J 1600 2. ■1400 s 1200 -^ 1000 800 600 400 200 0 Figure 2 Number of conch and conch biomass sampled within a 2.5-m radius circle at Shark Rock, July 1988 and February 1989, at each station. Values are means + SD. density between stations 1 and 5 at all sites and dates, except SR station 4. A one-way ANOVA and a Newman-Keuls multiple range test were used to test for differences in log transformations of conch density among stations 1-5 combining al! sites and dates (CBC-July, SR-July, CBC-February, SR- February). There was a significant difference between stations (F = 22.02, /j<0.0001). Conch densities increased with increasing station number, ranging from 0.42 (±1.38) to 22.4 (±15.8). Newman-Keuls tests showed that the logs of conch den- sity at stations 1 and 2 were significantly different from all other stations, though not from one another, as were stations 4 and 5. Station 3 was significantly different from all other stations (Newman-Keuls, /><0.05). Conch biomass patterns As with conch density, there was an increase in conch biomass at CBC, from stations 1 to 5. The pattern was strong for both July and February, although the actual biomass values were higher in July at stations 4 and 5 than in February. At SR the biomass also increased from stations 1 to 5 (with a slight decrease at station Results of multi-way in seagrass meadows sects were examined Table 4 analysis of variance for conch density Data were log-transformed and tran- as blocks. Source df MS F P Date 1 •2.503 22.755 < 0.0001 Site 1 0.753 6.846 0.011 Transect 2 0.032 0.289 0.750 Station 6 2.400 21.820 <0.0001 Date ♦ Site 1 1.047 3.058 0.086 Date * Station 6 0.539 1.574 0.171 Site * Station 6 0.160 0.468 0.829 Date • Site ' Station 6 0.228 0.642 0.696 Error .56 0.355 Results of multi-way in seagrass meadows sects were examined Table 5 analysis of variance for conch biomass Data were log-transformed and tran- as blocks. Source df MS F P Date 1 9.349 15.564 <0.0001 Site 1 3.025 5.036 0.028 Transect 2 0.155 0.258 0.773 Station 6 11.511 19.162 <0.0001 Date * Site 1 3.8.38 2.372 0.129 Date • Station 6 1.868 1.154 0.343 Site * Station 6 1.614 0.997 0.436 Date * Site * Station 6 1.032 0.610 0.722 Error 56 1.692 4 in July). In July, biomass decreased sharply at sta- tion 6 and 7, while in February the biomass continued to increase at station 6 and decreased at station 7. In July stations 3, 4, and 5 had much higher biomass values than in February. After log transformations of the biomass data, a multi-way ANOVA was used again to test for differ- ences or interactions between dates, sites, transects, and stations in conch biomass with transects examined as blocks (Table 5). Where CBC and SR were com- pared, stations 6 and 7 at SR were not used. There were no significant interaction terms and no block effects on the ANOVA (p>0.05), therefore the effects of date, site, and station are interpreted direct- ly. As mentioned above, conch biomass at both CBC and SR was higher in July than in February at most stations. Site differences appear to be due to low biomass values in February at SR, especially at stations 1-3. Station differences clearly resulted from the in- crease of conch biomass with the increase in station number, at all sites and dates, except SR station 4. Stoner and Waite: Habitat associations of Strombus gigas within seagrass meadows 579 Table 6 Pearson correlation coefficients for re ationships between measures of conch abundance and characteristics of seagrass habitat for different months and sites. CBC = Children's Bay Cay site; SR = Shark Rock site. At the Shark Rock site, analyses were made with 5 and 7 stations because | of decreasing numbers of conch at stations 6 and 7 (see text). CBC CBC SR SR (5 stations) (7 stations) July Feb. July Feb. July Feb. Conch density Depth 0.868** 0.868** 0.813** 0.837** 0.168 0.650** Shoots 0.860** 0.851** 0.831** 0.787** 0.314 0.726** Thalassia 0.910** 0.753** 0.779** 0.803** 0.141 0.495* Detritus 0.608* 0.753** 0.701** Conch biomass 0.853** 0.141 0.495* Depth 0.774** 0.892** 0.850** 0.801 •• 0.216 0.530* Shoots 0.874** 0.831** 0.863** 0.741** 0.423 0.634** Thalassia 0.886** 0.781** 0.836** 0.799** 0.270 0.353 Detritus 0.680** 0.597* 0.827** 0.780** 0.004 0.244 Combined sites and dates (5 stations) Conch density Conch biomass Depth 0.603** 0.492** Shoots 0.636** 0.513** Thalassia 0.518** 0.438** Detritus 0.330* 0.266* *p<0.05; ** )<0.01. A one-way ANOVA and a Newman-Keuls multiple range test were used to test for differences in log transformations of conch biomass among stations for all date-site samples. There was a significant difference in the log of conch biomass across stations. Conch biomass increased from 13.4 g wet weight (±44.5) at station 1 to 851 g wet weight ( ± 766) at station 5. The log of conch biomass at stations 1 and 2 were not dif- ferent from one another but were each significantly dif- ferent from stations 3, 4, and 5. Stations 3 and 5 were each significantly different from all stations but sta- tion 4 (Newman-Keuls, jo<0.05). Relationships between conch and the independent variables There were highly significant correlations between both density and biomass of conch in the field and the depth and macrophyte characteristics of the habitat (Table 6). Tested by individual date at Children's Bay Cay, correlations between conch density and depth, shoot density, and seagrass biomass were high and relatively similar (r values between 0.753 and 0.910), while correlations with detritus were significant but lower. Similar patterns held for conch biomass. At Shark Rock in July, conch numbers and biomass were highest at station 5 (Fig. 2) and did not increase over the entire range of seagrass shoot density or biomass. Therefore, linear regressions of conch density and biomass with depth and macrophyte characteristics did not yield high or significant correlations in most cases when data from all seven stations were included (Table 6). Highly significant correlations were found using the first five stations; all of the habitat characteristics yielded correlation coefficients between 0.701 and 0.863 (Table 6). In regression models including data from all dates, sites, and stations 1-5, all of the environmental variables yielded significant correlations with conch density and biomass (Table 6). Highest correlations oc- curred between seagrass shoot density and both conch density and conch biomass. Lowest correlation coeffi- cients occurred with detritus. Two stepwise multiple regressions were run to deter- mine the best multiple regression models for conch den- sity and conch biomass using all of the data for stations 1-5 (all dates and sites combined). Alpha to enter and remove from the models was set at 0.150. The regres- sion model for conch density included first shoot den- sity followed by detritus standing crop, and yielded a 580 Fishery Bulletin 88(3), 1990 CHILDREN'S BAY CAY JULY FEBRUARY ,15 Station 3 Station 3 30 n=26 30 n=22 25 25 20 20 15 15 10 ID to 5 0 40 ^ ^ >, 5 f-1 rr >i ,Vi Station 4 ,— , ^■i station 4 > 30 n=90 3D n=50 Q Z 25 H 25 b- 20 20 O 15 15 1—1 q: UJ 10 10 " S -J "h 5 rT ^ z .: 1 ^5 Stotion 5 .IS Station 5 30 n=97 30 n=38 25 25 20 20 15 - 15 10 5 0 n 10 5 r-Thr n r-r ABCDEFGHIJK LM 0 E F G H 1 J K L M SIZE CLASS Figure 3 Number of individuals per sample (2..5-m radius circle) within each conch size-class for each station at Children's Bay Cay, July 1988 and February 1989. Size-classes are as follows: A = 40-49 mm, B = 50-59 mm, C = 60-69 mm. D = 70-79 mm. E = 80-89 mm, F = 90-99 mm, G = 100-109 mm, H = 110-119 mm, I = 120-129 mm, J = 130-139 mm, K = 140-149 mm, L = 1.50-159 mm. M = 160-169 mm, N = 170-179 mm, 0 = 180-189 mm, P = 190-199 mm, and Q = 200-209 mm. multiple correlation coefficient of 0.664 (;>< 0.001). The stepwise multiple regression model for conch biomass included only shoot density and provided a multiple regression coefficient of 0.513 (7J<0.001). No other variables contributed significantly to the correlation. Results of the regressions and the Newman-Keuls tests for conch density across stations indicate that sta- tions 3-5 provide a suitable habitat for conch larger than 70 mm SL with habitat characteristics ranging as follows: 256-864 mean shoots/m-, 13.9-80.6 mean g dry wt of ThalcL-isialrn", and 0,64-157.12 mean g dry wt of detritus/m-. In addition, the correlation between conch and macrophyte characteristics are linear only up to station 5 (608-864 mean shoots/m-, 62.4-80.6 mean g dry wt of Thuldssia/m-, 64.0-157.1 mean g dry wt of detritus/m-) which appears to be the near- optimal habitat for juvenile conch larger than 70 mm. Conch length-frequency patterns The Kolmogorov-Smirnov two-sample test was used to test for significant differences in the length-frequency distribution of conch at different stations. Only stations SHARK ROCK JULY Stotion 3 n=7t n r In station 4 n=58 r-r -V, 1 ^-ru Stotion 5 n n=117 "U rl rh n n-r-r-n-Th r-n-J h II I I ru-n DE FGH I J KLMNOPQ DEFGHI JKLMNO Figure 4 Number of individuals per sample (2.5-m radius circle) within each conch size-class for each station at Shark Rock, July 1988 and February 1989. For size-classes see Figure 3. with at least 10 conch were used in the analysis (CBC stations 3-5 in both July and February; SR stations 3-7 in July and stations 5-6 in February). In July, CBC stations 3, 4, and 5 all had significant- ly different conch length-frequencies (stations 3 and 4, p<0.0001, stations 3 and b, p = 0.015, stations 4 and 5, p<0.0001). Mean lengths varied only from 102 to 111 mm, and the statistical differences were a result of subtle differences in the shape of the distributions (Fig. 3). In February at CBC, the length-frequency distribution of conch at station 3 was different signif- icantly from that at station 4 {p = 0.004), and the distribution at station 4 was different significantly from that at station 5 (/)< 0.0001). The length-frequencies at stations 3 and 5 were not different (p = 0.062); the mean values and shapes of the distribution were similar (Fig. 3). In July, conch length-frequency patterns at SR sta- tions 3, 4, and 5 were not different (stations 3 and 4, p = 0.220; stations 4 and 5, /) = 0.260), and had means Stoner and Waite Habitat associations of Strombus gigas within seagrass meadows 581 O § Q_ O Cl U- O o UJ Q_ LOW/SAND 100 BO T 60 r 40 20 n rn 100 80 60 40 20 100 BO 60 40 20 0 L S L S L S MODERATE/SAND MS MS MS MODERATE/LOW ■ r 1 1 • _ ^ r-i r I ML ML ML HIGH/MODERATE • h r-i H M H M H M 75-100 125-150 Adults SIZE CLASS (mm) Figure 5 Percentage of conch in three size-classes found within a haliitat type from a choice of two tyjjes. Choices included combina- tions of bare sand, low and high seagrass biomass. Table 7 Results of G-tests on habitat choice by three size-classes of conch presented with paired habitats. Size-class Total no. of conch (mm) No. of runs in the habitats Gadj P Low/Sand treatments 75-100 6 37/6 24.86 <0.01 125-150 6 28/9 10.24 <0.01 Adults 6 6/5 Moderate/Sand treatments 0.09 >0.05 75-100 6 53/7 42.99 <0.01 125-150 8 61/10 36.41 <0.01 Adults 6 33/0 Moderate/Low treatments <0.01 75-100 B 44/17 12.27 <0.01 125-150 6 38/22 9.72 <0.01 Adults 6 31/5 High/Moderate treatments 20.89 <0.01 75-100 4 2/23 20.72 <0.01 125-150 4 8/25 9.19 <0.01 Adults 4 4/3 0.14 >0.05 between 129 and 136 mm. Distributions at stations 6 and 7 were not different {j) = 0.749) (Fig. 4). This pat- tern of size-frequency distribution showed that, in July, conch were significantly larger at the stations with higher macrophyte biomass (stations 6 and 7), while smaller conch were found at sites with moderate-to- low seagrass biomass (stations 3, 4, and 5). In Febru- ary, the two SR stations with more than 10 conch- stations 5 and 6— had significantly different length- frequencies (;?< 0.0001) (Fig. 4). Again, the large conch were associated with high seagrass biomass. Habitat preference experiments Results of the habitat preference experiments showed that conch were proficient in detecting and choosing habitats with different macrophyte characteristics (Fig. 5). Of the conch in all size classes, 90% were associated with the moderate density plots as opposed to sand habitat (10%). Habitat selectivity was less strong in the other habitat pairs, but it is clear that plots with sea- grass present were selected over bare-sand habitats, and moderate-density seagrass was selected over either high- or low-density seagrass. Heterogeneity G tests showed that the results of runs within all of the individual habitat preference tests were homogeneous (p<0.05) except one (Low vs. Sand; 125-150 mm; Gh= 14.37, p>0.05). Standard G tests were used with the pooled data to test the null hypothesis that animals were distributed equally over 582 Fishery Bulletin 88(3). 1990 the two sides of the enclosures. All treatments and size- classes resulted in highly significant differences (p<0.01), except for two cases (Table 7). Adults in the Low/Sand treatment and adults in the High/Moderate treatment showed no significant habitat preference (p>0.05). In these two experiments adult conch were highly motile, and more than 50% of the individuals were found traveling around the walls of the en- closures. Because only animals not touching the walls were included in the analysis, the numbers for analysis were small and highly variable (Table 7, Fig. 5). Where a high-preference habitat (Moderate Density) was paired with a low-preference habitat (Low Density or Sand), adult conch were not found against the walls of the enclosures in such high numbers (28%) and habitat selection was highly significant (Table 7). Discussion Many marine organisms are known to prefer seagrass beds over bare sand. For example, amphipods, tanaida- ceans, decapods, and fishes have all been found to be more abundant in seagrass beds than on bare sand (Heck and Thoman 1981, Holt et al. 1983, Stoner 1983, Heck et al. 1989). Also, abundance of animals within seagrass beds appears to be influenced strongly by the local amount of macrophjte structure. Numbers of both decapod and peracarid crustaceans have been cor- related with seagrass biomass (Heck and Orth 1980, Stoner 1980a, Lewis 1984). A direct linear relationship between seagrass biomass and harpacticoid abundance and diversity was attributed to an increase in habitable space, nutritional resources, and reduced levels of predation (Hicks 1980). Communities of epibenthic fish on the banks of Florida Bay were associated with areas of high seagrass biomass and with accumulations of seagrass detritus (Sogard et al. 1987). Detritus was thought to provide a rich food source as well as a refuge from predators. The general association of queen conch with turtle- grass Thalassia testudinum is well known (Randall 1964, Hesse 1979, Weil and Laughlin 1984), but quan- titative relationships between the mollusc and seagrass beds are reported for the first time in this study. The distributional patterns of queen conch near Lee Stock- ing Island were similar to patterns found for other taxa, with few individuals associated with bare sand and low-density seagrass. As was true for other spe- cies, juvenile conch densities increased with macro- phyte biomass and shoot density over a wide range; however, optimal levels of macrophyte cover were found, beyond which only larger individuals were associated. The observed association of juvenile queen conch with moderate amounts of macrophyte cover (shoot density, biomass, etc.) probably results from a combination of mechanisms: (1) a simple inability of the animals to maneuver into or through heavy stands of seagrass and detritus, (2) active habitat choice for specific seagrass cover, and (3) differential survivor- ship associated with different macrophyte densities, especially in the smaller size-classes. The upper limit of seagrass density with which juve- nile queen conch are associated is probably set by their locomotory abilities. Randall (1964) speculated that thick stands of seagrass obstruct the movements of small conch. High shoot density is also responsible for heavy accumulations of detritus and soft sediments which may impede locomotion. Given that queen conch propel their heavy shells with thtaists of a pointed oper- culum on the end of a muscular foot, locomotion is most efficient on a firm substratum. The direct relationship between conch density and seagrass density below the optimal level is most likely linked to more complex interactions. It is known that abundance of food can be limiting for juvenile conch in the field at natural density (Stoner 1989). The primary sources of food for juvenile conch in nursery habitats are macrodetritus and algal epiphytes (Stoner and Waite, In review); both of these food items increase with seagrass density. Blade and detritus productiv- ity increase with shoot density, and seagrasses have the effect of increasing the entrapment of fine sedi- ments and detritus (den Hartog 1967, Orth 1977, this study), useful only up to the point of impeded locomo- tion. Algae consumed by conch in the seagrass nursery areas are primarily those growing as epiphytes on Thalasnia blades. Production of these epiphytes can be as high as 50% of the seagrass production (den Har- tog 1979), and increases with seagrass shoot density. The increase in density of conch with seagrass shoot density, therefore, may be a function of habitat choice by individuals for areas with greater abundance of foods and/or a function of food limitation in the population. The relationship between conch density and seagrass shoot density can also be a response to predation. This response could be direct, whereby the survivorship of conch increases with increasing seagrass structure, or indirect, where habitat preference for high-structure habitats has evolved as a response to predators. In either case, the role of seagrass biomass in protecting prey species is known for a wide variety of predator- prey combinations (Coen et al. 1981, Stoner 1982, Leber 1985, Heck and Wilson 1987), and experiments with crustaceans have shown that prey species are pro- ficient in choosing high-density seagrass (Stoner 1980b, Coen et al. 1981, Bell and Westoby 1986). The fact that small conch were more proficient in selecting habitats of particular seagrass biomass or shoot density than Stoner and Waite Habitat associations of Strombus gigas within seagrass meadows 583 large individuals suggests that there is strong selec- tive pressure for habitat choice in small conch, probably via size-specific mortality. Susceptibilty of juveniles to predation is known to decrease with increasing conch size (Appeldoorn 1984). Whether the observed distributional patterns are ultimately a response to foods and/or predators can- not be determined from data presented here; however, new experiments using artificial structures and pred- ator manipulations would be useful. In any case, strong habitat preferences suggest that evolved behavioral mechanisms are influential, and the evolution of habitat choice in conch is probably related to both procurring sufficient foods and avoiding predators. Significant effects of date in the density and biomass of conch in the seagrass beds was related to the normal seasonality of reproduction and recruitment in conch of the Exuma Cays. The 1 -i- year-class (~40-100 mm), spawned during the previous summer (1987), was most abundant in July. In the Exuma Cays, spawning occurs between April and October (Stoner et al. In review), with greatest recruitment to the benthos probably oc- curring between August and October. Juvenile conch spend most of their first year buried in the sediment; therefore, only the 1987 and older cohorts were surveyed in July 1988 and Febnaary 1989. In February, the 1988 cohort had not yet emerged from the sediment and, therefore, the 1 + year-class was not found in that survey. Information on the settlement and distribution of juvenile conch in their first year of life will be par- ticularly useful in elucidating mechanisms of distribu- tion and abundance. Conch less than 40 mm have been found in very shallow, unvegetated subtidal sediments, moving to deeper waters or seagrass beds with age (Weil and Laughlin 1984; Stoner and Sandt, unpubl. data). Highest abundance of 1-year-old conch in sparse seagrass close to sand shoals, therefore, could be a result of ontogenetic migration. The relationship between juvenile conch and easily measured characteristics of the macrophyte community should prove useful in predicting conch distributions and in planning stock-enhancement programs involv- ing the seeding of hatchery-reared conch. The lack of statistical interaction among the factors site, station, and date suggests that the strong relationships be- tween conch and seagrass shoot density (and other macrophyte characteristics) are relatively universal. It must be pointed out, however, that the surveys were conducted within areas known to be long-term conch nursery sites, and there are extensive areas of sea- grasses on the Exuma Bank near Lee Stocking Island which appear to have similar habitat characteristics but no resident conch populations. Recent transplant ex- periments have shown that many of the habitats with appropriate seagrass, sediment, and detritus char- acteristics do not provide good growth or survivorship in juvenile conch (Stoner and Sandt, In press). Clear- ly, there are unmeasured variables in the seagrass meadows which are important aspects of habitat qual- ity for conch. Habitats without natural conch popula- tions, which produced good growth and survival rates, may be recruitment limited. Such limitation may be a function of larval supply or ontogenetic movements in the species. This study demonstrates that there is a strong relationship between conch density and biomass and the amount of seagrass structure in the habitat; however, general seagrass characteristics cannot be used as predictors of conch distribution independent of other variables such as food quality, presence of predators, hydrographic considerations, and recruit- ment processes. Acknowledgments This research was supported by a grant from the National Undersea Research Program, NOAA, U.S. Department of Commerce. We thank P. Bergman, L. Cox, B. 011a, K. McCarthy, V. Sandt, C. Tanner, R. Wicklund, and E. Wishinski for assistance in the field and for discussion of the field data. R. Appel- doorn, L. Marshall, L. Riggs, C. Ryer, V. Sandt, and R. Wicklund provided thoughtful criticism of the manuscript. Citations Adams, J.E. 1970 Conch fishing industry of Union Island, Grenadines. West Indies. J. Trop. Sci. 12:279-288. Appeldoorn. R.S. 1984 The effect of size on mortality of small, juvenile conchs (Strombus (ligag Linne and S. costatus Gmelin). J. Shellfish Res. 4:37-4;i Bell. J.D.. and M. Westoby 1986 Abundance of macrofauna in dense seagrass is due to habitat preference, not predation. Oecologia (Berl.) 68: 205-209. Berg, C.J. Jr. 1976 Growth of the queen conch, Strombus (figaa. with a discus- sion of the practicality of its mariculture. Mar. Biol. (Berl.) 34:191-199. Bertram, G.C.L.. and C.K.R. Bertram 1968 Bionomics of dugongs and manatees. Nature (Lond.) 218:423-426. Brownell, W.N.. and J.M. Stevely 1981 The biology, fisheries, and management of the queen conch. Strombus yigas. Mar. Fish. Rev. 43(7): 1-12. Brownell, W.N., C.J. Berg Jr., and K.C. Haines 1977 Fisheries and aquaculture of the conch Strombus gigas in the Caribbean. FAO Fish. Rep, 206:59-69. 584 Fishery Bulletin 88(3), 1990 Coen, L.D.. K.L. Heck Jr., and L.G. Abele 1981 Expenments on competition and predaion among shrimps of seagrass meadows. Ecology 62:1484-1493. Coulston. M.L., R.W. Berey, A.C. Dempsey, and P. Odum 1987 Assessment of the queen conch (Si)-ombus yigas) popula- tion and predation studies of hatchery reared juveniles in Salt River Canyon, St. Croix, U.S. Virgin Islands. Proc. Gulf Caribb. Fish. Inst. 38:294-306. den Hartog, C. 1967 The structural aspect in the ecology of seagrass com- munities. Helgol. Wiss. Meeresunters. 15:648-659. 1979 Seagrasses and seagrass ecosystems, an appraisal of the research approach. Aquat. Bot. 7:105-117. Folk. R.L. 1966 A review of grain-size parameters. Sedimentology 6: 73-93. Fry. B. 1984 '■'C/'-C ratios and the trophic importance of algae in Fhnda Syringodiurn filiforrne seagrass meadows. Mar. Biol. (Berl.) 79:11-20. Heck, K.L. Jr., and R.J. Orth 1980 Seagrass habitats: The roles of habitat complexity, com- petition and predation in structuring associated fish and motile macroinvertebrate assemblages. In Kennedy, V.S. (ed.), Estuarine perspectives, p. 449-464. Acad. Press, NY. Heck, K.L. Jr., and T.A. Thoman 1981 Experiments on predator-prey interactions in vegetated aquatic habitats. J. Exp. Mar. Biol. Ecol. .53:12.5-134. Heck, K.L. Jr., and K.A, Wilson 1987 Predation rates on decapod crustaceans in latitudinally separated seagrass communities: A study of spatial and tem- poral variation using tethering techniques. J. Exp. Mar. Biol. Ecol. 107:87-100. Heck, K.L. Jr., K.W. Able, M.P. Fahay, and C.T. Roman 1989 Fishes and decapod crustaceans of Cape Cod eelgrass meadows: Species composition, seasonal abundance patterns and compaison with unvegetated substrates. Estuaries 12: 59-65. Hesse, K.O, 1979 Movement and migration of the queen conch. Stromhiix gigas, in the Turks and Caicos Islands. Bull. Mar. Sci. 19: ■303-311. Hicks, G.R.F. 1980 Structure of phytal harpacticoid copepod assemblages and the influence of habitat complexitiy and turbidity. J. Exp. Mar. Biol. Ecol. 44:157-192. Holt, S.A., C.L. Kitting, and C.R. Arnold 1983 Distribution of young red drums among different seagrass meadows. Trans. Am. Fish. Soc. 112:267-271. Howard, R,K. 1984 The trophic ecology of caridean shrimps in an eelgrass community. Aquat. Bot. 18:155-174. Kikuchi, T. 1980 Faunal relationships in temperate seagrass beds. In Phillips, R.C., and C.P. McRoy (eds.). Handbook of seagrass biology: An ecosystem perspective, p. 153-172. Garland Press, NY. Kitting, C.L,, B, Fry. and M.D. Morgan 1984 Detection of inconspicuous epiphytic algae supporting food webs in seagrass meadows. Oecologia (Berl.) 62:145-149. Laughlin, R.A., and E. Weil M. 1983. Queen conch mariculture and restoration in the Ar- chipielago de Los Rogues: Preliminary results. Proc. Gulf Caribb. Fish. Inst. 35:64-72. Lawrence, J.M. 197,5 On the relationships between marine plants and sea urchins. Oceanogr. Mar. Biol. Annu. Rev. 13:213-286. Leber, K,M. Ill 1983 Feeding ecology of decapod crustaceans and the influence of vegetation on foraging success in a subtropical seagrass meadow. Ph.D. thesis, Florida State Univ., Tallahassee, 166 p. 1985 The influence of predatory decapods, refuge, and micro- habitat selection on seagrass communities. Ecology 66: 1951-1964. Lewis, F.G. Ill 1984 Distribution of macrobenthic crustaceans associated with Thalassia, Halodule, and bare sand substrata. Mar. Ecol. Prog. Ser. 19:101-113. Marshall, L.S. Jr., and R.N. Lipcius In review Density-dependent mortality of juvenile queen conch and predator aggregation. Mar. Ecol. Prog. Ser. McBride, E.F. 1971 Mathematical treatment of size distribution data. In Carver, R.E. (ed.). Procedures in sedimentary petrology, p. 109-127. Wiley, NY. Nelson. W.G, 1979 Experimental studies of selective predation on amphi- pods: Consequences for amphipod distribution and abundance. .J. Exp. Mar. Biol. Ecol. 38:225-245. 1981 Experimental studies of decapod and fish predation on seagrass macrobenthos of the Indian River, Florida. Mar. Ecol. Prog. .Ser. 5:141-149. Ogden, J.C. 1980 Faunal relationships in Caribbean seagrass beds. In Phillips, R.C., and C.P. McRoy (eds.). Handbook of seagrass biology: An ecosystem perspective, p. 173-198. Garland Press, NY. Orth, R,J. 1977 The importance of sediment stability in seagrass com- munities. In Coull, B.C. (ed.). Ecology of marine botany, p. 281-300. Univ. South Carolina Press, Columbia. Randall, J.E. 1964 Contributions to the biology of the queen conch, Strom- bus gigas. Bull. Mar. Sci. Gulf Caribb. 14:246-295. Sogard, S,M., G.V,N, Powell, and J.G, Holmquist 1987 Epibenthic fish communities on Florida Bay banks: Rela- tions with physical parameters and seagrass cover. Mar. Ecol. Prog. Ser. 40:25-39. Stoner, A,W, 1980a The role of seagrass biomass in the organization of ben- thic macrofaunal assemblages. Bull. Mar. Sci. 30:537-551. 1980b Perception and choice of substratum by epifaunal am- phipods associated with seagrasses. Mar. Ecol. Prog. Ser. 3:105-111. 1982 The influence of benthic macrophytes on the foraging behavior of pinfish Lagodon rhamhoidi'it (Linnaeus). ,1. Exp. Mar. Biol. Ecol. 58:271-284. 1983 Distril)utional ecology of amphipods and tanaidaceans associated with three seagrass species. J. Crust. Biol. 3: ,505-518. 1989 Density dependent growth and the grazing effects of juvenile queen conch {Stromhus gigas L.) in a tropical seagrass meadow. .1. Exp. Mar. Biol. Ecol. 130:119-133. Stoner, A.W„ and V,J, Sandt In press Transplanting as a test procedure before large-scale outplanting of juvenile queen conch. Proc. Gulf Caribb. Fish. Inst. Stoner and Waite: Habitat associations of Strombus gigas within seagrass meadows 585 Stoner, A.W.. and J.M. Waite In review Trophic biology of queen conch in nursery habitats: diets and primary food sources in seagrass meadows. Mar. Ecol. Prog. Ser. Stoner, A.W., I.F. Boidron-Metairon. and V.J. Sandt In review Reproductive activity of queen conch, Strombus gigas, and the abundance of veligers. Bull. Mar. Sci. Van Montfrans, J., R.L. Wetzel, and R.J. Orth 1984 Epiphyte-grazer relationships in seagrass meadows: Con- sequences for seagrass growth and predation. Estuaries 7:289-309. Virnstein, R.W., W.G. Nelson, F.G. Lewis III, and R.K. Howard 1984 Latitudinal patterns in seagrass epifauna: Do patterns exist and can they be explained. Estuaries 7:310-330. Weil M., E., and R. Laughlin G. 1984 Biology, population dynamics, and reproduction of the queen conch, Strombus gigas Linne, in the Archipielago de Los Rogues National Park. J. Shellfish Res. 4:45-62. Wicklund, R.I., L.J. Hepp, and G.A. Wenz 1988 Preliminary studies on the early life history of the queen conch, Strombus gigas. in the Exuma Cays, Bahamas. NURP Res. Rep. 88-4:347-363. Natl. Undersea Res. Prog., NOAA, Roekville, MD 20852. Wilson, K.A., K.L. Heck Jr., and K.W. Able 1987 Juvenile blue crabs, Callinectes sapidiis, survival: An evaluation of eelgrass, Zosterd marina, as refuge. Fish. Bull, U.S. 85:53-58. Zimmerman, R., R. Gibson, and J. Harrington 1979 Herbivory and detritivory among gammaridean am- phipods from a Florida seagrass community. Mar. Biol. (Berl.) 54:41-47. Abstract.— Seventy-two species of hippolytid shrimp (Decapoda: Cnis- tacea: Hippolytidae) have been re- ported from the eastern Pacific (the Aleutians to Cape Horn). There are more species to the north of Magda- lena Bay, Baja California, than in tropical or temperate waters to the south. Most species oi Euabis, Leb- beus, Heptacarpus, and Spirontoca- n's live along the temperate coasts of North America; species of Thor, Trachycaris, and Lysmata are most common in warm-temperate to trop- ical areas, and species of Hippolyte range from cool to tropical regions. In the eastern Pacific, species oiNau- ticaris and Leontocaris have been found only in the Southern Hemi- sphere. An annotated key to the spe- cies, with geogi-aphic and depth range and selected synonyms, is provided. Key to the Hippolytid Slirimp of the Eastern Pacific Ocean Mary K. Wicksten Department of Biology, Texas A&M University College Station, Texas 77843-3258 Manuscript accepted 22 March UntO. Fishery Bulletin, U.S. 88:587-.598. Members of the family Hippolytidae are among the most abundant and di- verse carideans of the eastern Pacif- ic. Hippolytids are eaten by birds and fishes, and serve as cleaners for fishes such as moray eels (Chace 1937, Lim- baugh 1961, Hobson and Chess 1976, Reynolds 1977). They are common in- habitants of tidepools, kelp beds, sea- gi-ass flats, and other nearshore areas, but also range into the deeper parts of the continental shelf and slope. The gi-eatest diversity of species of hippolytids in the eastern Pacific is in boreal to temperate waters of the northern hemisphere (Table 1). Forty- seven species live from the Aleutian Islands to Baja California, Mexico. Among these, Lebbeus vicinus has the greatest range, from the Aleu- tians to the Gulf of California. Hip- polyte californiensis lives from the coast of Alaska to the Gtilf of Califor- nia. Eighteen other species range from the Aleutians or the mainland of Alaska south to southern Califor- nia, USA, or Baja California, Mexico. Twenty-one species range from the Bering Sea or the Aleutians to Bri- tish Columbia, Washington, or Ore- gon. Two species have been reported only from Puget Sound, and four species range from the Bering Sea or the Aleutians to northern or central California. An additional six species range from central California to Baja California or the Gulf of California, Mexico. In the Northern Hemisphere, spe- cies of Spirontocaris, Heptacarpus, Eualus, and Lebbeus are diverse. Species of the first two genera are not found in the Southern Hemi- sphere, while species of the latter are more diverse in the northern than in the southern hemispheres. Although no species of these genera are found off both North and South America, Lebbeus washingtonianus and L. bi- dentatns, and Eualus pusiolus and E. dozei are very similar and may repre- sent pairs of disjunct sibling species. Tropical areas of the eastern Pacif- ic are poor in hippolytid species. Lys- mata californica ranges from central California to the Galapagos, while Latreutes antiborealis ranges from the Gulf of California to Chile. Thor algicola and L. antiborealis are sib- ling species of T. ynanningi and L. pannilus of the Caribbean and warm- temperate western Atlantic. Hippo- lyte :ostericola, Trachycaris restric- tus, and Lysmata intertnedia occur both in the eastern Pacific and in the Caribbean and tropical Atlantic. Thor spinosus and Lysmata trisetacea range from the eastern Pacific to the Indo-West Pacific, and Thor ajnboi- nensis is circumtropical. Only Hippo- lyte williamsi, Heptacarpus yald- wyni, and Lysmata galapagensis are endemic and distinct. The hippolytid fauna of temperate South America is depauperate com- pared with that of North America. Except for Leontocaris pacificus, the only representative of this genus in the eastern Pacific, all of the species belong to genera present in tropical or temperate areas of the northern hemisphere. Four species oi Lebbeus live on the continental slopes. Hippolytids of the eastern Pacific inhabit a wide range of depths and habitats. Species of Spirontocaris, Lebbeus, and Eualus prefer habitats from the lowest intertidal zone to the 587 588 Fishery Bulletin 88(3), 1990 Table 1 Ranges of Hippolytid species in the eastern Pacific. Bering Sea, Aleutian Island or Alaskan Mainland to Southern California, USA, or Baja California, Mexico Eualus aviniis Heptaearpus breinrostris E. barbatus H. camtschaticus E. berkeleyorum H. carina iris E. biunguis H. deeorus E. fabricii H. flexxis E. lineatus H, herdmani E. macrophthaimus H. kincaidi E. pusiolus H. iittoralis E. suck-Uyi H. moseri E. townsendi H. paludicola Hippolyte caiiforniensis H. piigettensis H. clarki H. sitchensis Lebbeus brandti H. stimpsoni L. catalepsis H. stylus L. grandimamis H. taylori L. groenlandicus H. tenuissimus L. polaris H. tridens L. possjeticus Spirontocaris arruata L. schrenki S. holmesi L. vicinus S. lamellicomis L. washingtonianus S. ochotensis L. zebra S. prionota S. sica S. snydfri S. truncata Central California to Baja California Heptaca-rpus brachydactyiux Lebbe^is lagunae H. franciscanus Lysmata califomica H. fusrimamlatus (south to Galapagos) H. pal pat or H. pictus Gulf of California to Northern Peru Hepfacarpus yaldwyni Lysmata galapagensis Hippolyte will ia nisi L. intermedia H. zostericola L. trisetacea Latreutes nntiborealis Thor algicola (^outh to Chile) T. amboine.nsis T. spinosvs: Trachycaris restrict us Peru-Chile Lebbeits bidentatux Eualus dozei L. carinatus Leontocaris pacificus L. sciippsi Lysmata porteri L. splmdiduf continental slopes. Species oi Lebbeus are the deepest eastern Pacific hippolytids, living as deep as 2824 m. Except for Heptaearpus yaldwyni, species oi Hepfacar- pus usually range from the upper intertidal zone to the continental shelf. Species of Hippolyte live in shallow water with sea grasses or beds of algae. Species of Lysmata live in tidepools or in cracks and caves among coral or rocks from the intertidal zone to shallow sub- tidal areas. Thor amboiensis tends to associate with subtidal cnidarians. Thor algicola, Thor spinosus, Tra- chycaris restrictus, and Latreutes antiborealis live among algae or on subtidal bottoms of sand, rock, coral, or rubble. Leontocaris pacificus lives on the continen- tal slope. Classification of the hippolytid shrimp is in flux. Ac- cording to Williams (1984), species of the family are characterized by having the first two pairs of legs chelate, the first pair not much stronger than the rest, the carpus of the second pair of legs subdivided, the eyes well developed and not covered by the carapace, and the mandibles usually deeply cleft. A recent cladistic analysis by Christoffersen (1988) splits the Hippolytidae into at least seven families and divides them among two superfamilies, the Crangonoidea and the Alpheoidea. However, this analysis did not provide familial placement for all of the genera considered to belong to the traditional family Hippolytidae and pro- vided no key to the groups. Reinterpretation of some of the genera of hippolytids, such as Eualus, also is being considered by Christofferson and others. Despite the confusion surrounding the systematic relationships among hippolytids, it seems useful to pro- vide a key to the species. The last comprehensive key to the species of the northeastern Pacific was prepared by Rathbun (1904); the most recent extensive synon- ymy and list of species was included in the work of Holthuis (1947). More recent keys are available for British Columbia (Butler 1980), the Gulf of California (Wicksten 1983), and Peru (Mendez 1981). Illustrations and keys to intertidal species have been provided in ac- counts of the fauna of Puget Sound and adjacent areas (Kozloff 1974), California (Ricketts et al. 198.5; Smith and Carlton 1975; Chace and Abbott 1980) and the Gulf of California (Brusca 1980; Kerstitch 1989). Fourteen new species have been described since 1947 (Butler 1971, 1980; Jensen 1983, 1987; Kobyakova 1967; Wicksten and Butler 1983; Wicksten and Mendez 1982; Wicksten 1984, 1986. 1987a; Zarenkov 1976). The following key includes all hippolytid species reported from the Aleutian Islands (south of the Ber- ing Sea) to Cape Horn, excluding the Straits of Magel- lan. The arrangement in the key is artificial. Geo- graphic and depth ranges are provided. Superscript numbers in parentheses at the ends of couplets refer to notes following the key, where further information on ranges, synonyms, and species of uncertain classi- fication can be found. Wicksten: Key to hippolytid shnmp of eastern Pacific Ocean 589 Key to the Hippolytidae 1 Lateral surface of carapace with many scattered spines. Carpus of second pereopods with 2 segments Trachycaris restrictus (A. Milne-Edwards) Eastern Atlantic from Canary Islands to St. Helena Island, Bermudas, to Brazil, Gulf of California, Panama; 0-100 m (Wicksten 1983). Lateral surface of carapace with at most three spines. Carpus of second pereopods with 3 or more segments 2 2 Carpus of second pereopods with 3 segments. (Often associated with sea grasses or brown algae) 3 Carpus of second pereopods with 4 or more segments. (Size and associations various) 7 3 Rostrum deep, lamellate. Anterior margin of carapace with 10 spines. No hepatic spine Latreutes antiborealis Holthuis Gulf of California to Chile, 4-46 m (Wicksten 1983).<" Rostrum elongate. Anterior margin of carapace with 2 spines. One hepatic spine 4 4 First antennular segment without outer distal spines 5 First antennular segment with one or more outer distal spines 6 5 Tip of rostrum trifid. (Usually found among giant kelp, Macrocystis spp.) Hippolyte clarki Chace Sheep Bay, Alaska to Cedros Island, Baja California, 0-10 m (Wicksten 1983). Tip of rostrum bifid. (Usually found among sea grasses) Hippolyte zostericola (Smith) Massachusetts to Curacao, western Colombia, shallow subtidal (Wicksten 1989). 6 First segment of antennular peduncle with one (rarely 2) spines, rostrum of female reaching nearly as far as tip of scaphocerite Hippolyte californiensis Holmes Sitka, Alaska to Baja California; rare in Gulf of California, 0-10 m, usually among eelgrass (Zostera sp.) (Wicksten 1983). First segment of antennular peduncle with 3 spines, rostrum of female reaching about to distal 1/3 of scaphocerite Hippolyte williamsi Schmitt Gulf of California to Chile, 0-10 m, among rocks and brown algae (Scirgassum sp.) (Wicksten 1983).(2) 7 Carpus of second pereopod with 4 segments. Left second pereopod longer and stronger than right Leontocaris pacificus Zarenkov Off Chile, 600-700 m (Zarenkov 1976). Carpus of second pereopod with 6 or more segments. Left and right second pereopods equal or nearly so 8 8 Carpus of second pereopod with 6 segments. (Rostrum reaching at most to end of eye. Ranging from Gulf of California to Panama) 9 Carpus of second pereopod with 7 or more segments. (Rostrum reaching end of eye or beyond, found north of Gulf of California to Chile) 11 9 Supraorbital spine present. Finely striped, with large bluish-red eyespots on third abdominal tergite Thor spinosus Boone Indo-Pacific to Gulf of California, subtidal (Bruce 1976, Wicksten 1983). Supraorbital spine absent. Translucent, mottled or spotted, without eyespots 10 590 Fishery Bulletin 88(3). 1990 10 Brightly colored with pattern of bold white spots on dark background. Anterolateral margin of carapace angular. (Symbiotic with sea anemones or corals) Thor amboinensis Heller Tropical Atlantic and Indo-Pacific, Cocos Island and Panama. Sub tidal (Chace 1972; Abele and Patton 1976).(3) Camouflaged, translucent to mottled like algae. Anterolateral margin of carapace rounded. (Not symbiotic with sea anemones or corals) Thor algicola Wicksten Gulf of California to Panama, 0-20 m (Wicksten 1987a).(^> 11 Carpus of second pereopod with 7 segments. Flagellum of antennae not as long as body. (Color in life various, often camouflaged) 12 Carpus of second pereopod with more than 7 segments. Flagellum of antenna as long as body. (Color in life often red striped) 69 12 More than 1 supraorbital spine 13 One or no supraorbital spines 20 13 Rostrum with 3-4 supraorbital spines 14 Rostrum with 2 supraorbital spines 15 14 Rostrum deep, subcircular; with 10-26 mostly small dorsal spines on rostrum proper and 3-4 dorsal carapace spines, ventral margin of third pleuron acute or bluntly pointed Spirontocaris prionota (Stimpson) Bering Sea to SW of San Carlos Point, Baja California, 4-163 m (Butler 1980).<5' Rostrum moderately deep, with 1-3 large dorsal spines, truncate apex; pleuron of third somite rounded Spirontocaris truncata Rathbun Strait of Georgia to SW of San Carlos Point, Baja California, 37-92 m (Wicksten 1984). 15 Rostrum with styliform distal projection, no process on inner surface of eyestalk. (Usually found on offshore soft bottoms) 16 Rostrum without styliform distal projection, process present on inner surface of eyestalk. (Found among rocks or soft bottoms) 17 16 Distal projection of rostrum with 1 ventral spine; epipods on first and second pereopods. (Usually living at 150 m or less) Spirontocaris holmesi Holthuis Yes Bay, Alaska to San Diego, California, 24-386 m (Butler 1980).'«' Distal projection of rostrum without ventral spine; epipods on first pereopods only. (Usually living at 150 m and deeper) Spirontocaris sica Rathbun Restoration Bay, British Columbia to between San Benito Islands and Cedros Island, Baja Califor- nia; 88-849 m (Butler 1980, Wicksten 1987b). 17 Pleura of first to third somites ventrally acute or bluntly pointed, dactyls of third to fifth pereo- pods long, slender, with acute tips Spirontocaris lamellicornis (Dana) Bering Sea to Santa Monica Bay, California, 3-192 m (Butler 1980, Standing 1981). Pleura of first to third somites broadly rounded ventrally, dactyls of third to fifth pereopods short and stout with bifid tips 18 18 Body slender to moderately stout, shell thin; epipods on first and second pereopods Spirontocaris snyderi Rathbun Queen Charlotte Islands to San Cristobal Bay, Baja California, 4-141 m (Butler 1980)."' Body moderately stout to stout, shell thick; epipods on first to third pereopods 19 Wicksten Key to hippolytid shrimp of eastern Pacific Ocean 591 19 Rostrum more or less ovoid; posterior dorsal spine on carapace located near middle of carapace Spirontocaris ochotensis (Brandt) Bering Sea to Vancouver Island, Japan, coast of Siberia; 0-247 m (Butler 1980). Rostrum with lower limb broader than upper; posterior dorsal spine of carapace located on pos- terior third of carapace Spirontocaris arcuata Rathbun Arctic to Sea of Japan and Strait of Juan de Fuca, 5-641 m (Butler 1980). 20 (12) One supraorbital spine 21 No supraorbital spine 36 21 Epipods on only first pereopods 22 Epipods on at least first and second pereopods 23 22 Rostrum with 6-8 dorsal spines, stylocerite not reaching end of first segment of antennular peduncle Lebbeus vicinus vicinus (Rathbun) North of Unalaska, Aleutian Islands, 570-750 m (Rathbun 1904). Rostrum with 3-4 dorsal spines, stylocerite reaching end of first segment of antennular peduncle Lebbeus vicinus montereyensis Wicksten and Mendez Monterey Bay to Gulf of California, 954-2824 m (Wicksten and Mendez 1982). 23 Epipods on first and second pereopods 24 Epipods on at least first to third pereopods 28 24 Carapace with 4 large dorsal spines posterior to orbit. (Often associated with sea anemones) . . . Lebbeus grandimanus (Brazhnikov) Bering Sea to Sea of Japan and Puget Sound, Washington, 6-180 m (Butler 1980, Wicksten and Mendez 1982). Carapace with 1-2 dorsal spines posterior to orbit. (Associations various or not known) 25 25 Dactyl of third pereopod simple, without spines Lebbeus scrippsi Wicksten and Mendez Southern Peru to off Arica, Chile, 768-1164 m (Wicksten and Mendez 1982). Dactyl of third pereopod with spines 26 26 First segment of antennular peduncle with 2 anteroexternal spines. Rostrum with 5-9 ventral spines Lebbeus splendidus Wicksten and Mendez Off Lobos de Tierra, Peru, 712-1100 m (Wicksten and Mendez 1982). First segment of antennular peduncle with 1 anteroexternal spine. Rostrum with 4 or less ventral spines 27 27 Rostrum overreaching antennular peduncle, with 3-4 ventral spines Lebbeus polaris (Sabine) Arctic to Sea of Okhotsk, Cape Cod and northern Europe, Aleutian Islands to Alice Arm, British Columbia, 0-930 m (Green and Butler 1988). Rostrum barely reaching end to first segment of antennular peduncle, with 1-no ventral spines Lebbeus brandti (Brazhnikov) Japan and Sea of Okhotsk to Wilson Bay, Alaska, 10-172 m (Wicksten and Mendez 1982). 28 (23) Epipods on first to fourth pereopods Lebbeus carinatus Zarenkov Off Peru and Chile, 1860 m (Zarenkov 1976). Epipods on first to third pereopods 29 592 Fishery Bulletin 88(3). 1990 29 Pleura of first to fifth abdominal segments ending in 1-3 spines. Four large spines on dorsal mid- line of carapace Lebbeus groenlandicus (Fabricius) Arctic to Cape Cod, Sea of Okhotsk, Puget Sound, 11-518 m (Butler 1980). Pleura of at least first to third abdominal segments pointed to rounded, but not ending in 1-3 spines. One to 3 spines on dorsal midline of carapace 30 30 Rostrum reduced to spine on frontal margin of carapace. Three spines on anterior dorsal midline of carapace 31 Rostrum prominent, not reduced to spine. 1-2 spines on anterior dorsal midline of carapace 32 31 Telson with 2 pair dorsal pines. No large, decurved spines below articular knobs of posterolateral margin of fourth and fifth abdominal somites Lebbeus catalepsis Jensen Strait of Juan de Fuca, low intertidal zone (Jensen 1987). Telson with 3 pair dorsal spines. Large, decurved spines present below articular knobs of postero- lateral margin of fourth and fifth abdominal somites Lebbeus lagunae (Schmitt) Pacific Grove, California to Cedros Island, Baja California, 0-74 m (Wicksten 1978, 1988). 32 Antennular peduncle extending to near middle of scaphocerite. Small subtidal species 33 Antennular peduncle extending nearly to end of scaphocerite. Large species of continental slopes ... 35 33 Dorsal surface of second abdominal segment without transverse furrow and fold. Rostrum not overreaching eye Lebbeus zebra (Leim) East coast of Canada, Bering Sea, Kamchatka, British Columbia, off Santa Rosa Island, California, subtidal- 140 m (Wicksten and Mendez 1982). Dorsal surface of second abdominal segment with transverse furrow and fold. Rostrum overreach- ing eye 34 34 Rostrum short, not reaching end of first segment of antennular peduncle, with 2-5 dorsal spines and 1 ventral spine Lebbeus schrencki (Brazhnikov) Pribilof Islands to Sea of Japan to Passage Island, Strait of Georgia; 12-183 m (Butler 1980, Jensen and Armstrong 1987). Rostrum reaching end of scaphocerite, with 5-7 dorsal spines and 3-4 ventral spines Lebbeus possjeticus Kobyakova Possjet Bay, USSR, Bering Island and off San Nicolas Island, California; 3-57 m (Wicksten and Mendez 1982). 35 First segment of antennular peduncle with 1 spine, bi- or trifurcated. Northern hemisphere . . . Lebbeus washingtonianus (Rathbun) Anthony Island, Queen Charlotte Islands to off San Clemente Island, California; 820-1808 m (Butler 1980). First segment of antennular peduncle with 3 spines. Southern hemisphere Lebbeus bidentatus Zarenkov Peru and Chile, 1680 m (Zarenkov 1976). 36 (20) Third maxilliped with exopod 37 Third maxilliped without exopod 47 37 No epipod on any pereopod 38 Epipod on at least on first pereopod 40 Wicksten Key to hippolytid shrimp of eastern Pacific Ocean 593 38 Rostrum deep, shorter than carapace, eyes large Eualus macrophthalmus (Rathbun) Unalaska to Point Sur, California, 110-1163 m (Butler 1980). Rostrum slender, as long as or longer than carapace, eyes smaller 39 39 Posterior margin of third to sixth abdominal segments armed with median dorsal spine Eualus barbatus (Rathbun) Pribilof Islands to Santa Monica Bay, California, 82-507 m (Butler 1980, Wicksten 1984). Posterior margin of third to sixth abdominal segments unarmed Eualus biunguis (Rathbun) Bering Sea to Sea of Japan and Oregon, 90-2090 m (Butler 1980). 40 (37) Epipods on all 3 pereopods 41 Epipods on first and sometimes second pereopods 45 41 Rostrum with dorsal margin markedly convex over eye, 7-14 spines Eualus avinus (Rathbun) Pribilof Islands to Oregon, 46-642 m (Butler 1980). Rostrum nearly straight over eye, fewer (2-9) spines 42 42 Rostrum not reaching second segment of antennular peduncle 43 Rostrum reaching second segment of antennular peduncle 44 43 Rostrum with 2-5 dorsal teeth, telson with 3-4 pair dorsolateral spines. First pereopod slender, merus about 5 x as long as wide Eualus pusiolus (Kr0yer) Bering Sea to British Columbia and Sea of Japan, Gulf of St. Lawrence to Cape Cod, northern Europe to Spain, 0-1381 m (Butler 1980). Rostrum with 1-3 dorsal teeth, telson with 3 pair dorsolateral spines. First pereopod stout, merus about 1.6 X as long as wide Eualus dozei (A. Milne-Edwards) Chile-Cape Horn, 13-300 m (Holthuis 1952). 44 Pleuron of fourth abdominal segment with ventral spine Eualus lineatus Wicksten and Butler Sitka, Alaska to Gulf of California, 18-232 m (Wicksten and Butler 1983, Wicksten 1988).i^' Pleuron of fourth abdominal segment without spine Eualus berkeleyorum Butler Pribilof Islands to off Trinidad Harbor, California, 46-384 m (Butler 1980, Wicksten 1984, Jensen and Armstrong 1987). 45 (40) Normal size epipods on first and second pereopods Eualus toumsendi (Rathbun) Pribilof Islands to Sea of Japan and Puget Sound, 38-630 m (Butler 1980). Normal or reduced epipods on first pereopods, second pereopods with reduced epipods if present .... 46 46 Most of distal part of rostrum lacking spines on dorsal margin Eualus fabricii (Kr0yer) Bering Sea to Sea of Japan and British Columbia, northwest Atlantic Ocean south to Massachusetts Bay, 4-255 m (Butler 1980). Distal half of rostrum with dorsal spines Eualus suckleyi (Stimpson) Bering Sea to Sea of Okhotsk and Grays Harbor, Washington, 11-1025 m (Butler 1980). 47 (36) Ventral margin of fourth pleuron without spine 48 Ventral margin of fourth pleuron with spine 59 48 Epipod absent on third maxilliped Heptacarpus tenuissimus Holmes Bird Island, Alaska to Santa Catalina Island, California, 2-137 m (Butler 1980). Epipod present on third maxilliped 49 594 Fishery Bulletin 88(3), 1990 49 Epipods present on first and second or first to third pereopods 50 Epipods absent from all pereopods 52 50 Epipods present on first to third pereopods Heptacarpus carinatus Holmes Dixon Harbor, Alaska, to Point Loma, California, 0-27 m (Butler 1980). Epipods present on first and second pereopods 51 51 Posterior margin of third abdominal segment produced, forming rounded carina. Rostrum at least as long as carapace, with 5-8 ventral teeth Heptacarpus flexus Rathbun Bering Sea to Farallon Islands, California, 37-172 m (Schmitt 1921).*"" Posterior margin of third abdominal segment not produced. Rostrum shorter, not reaching end of scaphocerite, with 1 ventral tooth Heptacarpus herdmani (Walker) "Puget Sound" (Wicksten and Butler 1983). 52 (49) Pterygostomian spine absent 53 Pterygostomian spine present 55 53 Rostrum shorter than carapace, distal ventral half convex .... Heptacarpus brachydactylus (Rathbun) Monterey Bay to off Santa Cruz Island, California, 486-763 m (Standing 1981). Rostrum longer than carapace, distal ventral half not convex or only slightly so 54 54 Rostrum overreaching scaphocerite, without dorsal teeth on distal half Heptacarpus stylus (Stimpson) Chichagof Island, Alaska to Puget Sound, 0-439 m (Butler 1980). Rostrum barely exceeding scaphocerite, with dorsal teeth on distal half Heptacarpus franciscanus (Schmitt) San Francisco Bay, California, to Todos Santos Bay, Baja California, 5-14 m (Schmitt 1921, Carvacho and Olson 1984)." 55 (52) Rostrum with tiny lateral spinule on each side, located laterally near base Heptacarpus yaldwyni Wicksten Off Salina Cruz, Mexico, 1052-1145 m (Wicksten 1984). Rostrum without tiny lateral spinule on each side 56 56 Scaphocerite shorter than carapace; sixth abdominal somite longer than telson Heptacarpus decorus (Rathbun) Gabriola Island, Strait of Georgia to San Diego, California, 22-313 m (Butler 1980). Scaphocerite as long as or longer than carapace; sixth abdominal somite shorter than telson 57 57 Three spines on carapace and rostrum; none anterior to eye Heptacarpus tridens (Rathbun) Aleutian Islands to Cape Flattery, Washington, 0-110 m (Butler 1980). Five to six spines on carapace and rostrum, extending anterior to eye 58 58 Dorsal posterior margin of third abdominal somite prominent; third maxilliped exteniling to middle of scaphocerite Heptacarpus camtschaticus (Stimp.son) Bering Sea to Tokyo Bay and Strait of Georj^a, 0-108 m (Butler 1980). Dorsal posterior margin of third abdominal somite flattened, third maxilliped extending almost to end of scaphocerite Heptacarpus kincaidi (Rathbun) East coast of Vancouver Island to San Pedro, California, 10-183 m (Butler 1980). Wicksten: Key to hippolytid shrimp of eastern Pacific Ocean 595 59 (47) Epipod only on first pereopod 60 Epipods on first and second, or first to third pereopods 62 60 Pterygostomian spine absent Heptacarpus littoralis Butler Baranof Island, Alaska to Seattle, Washington, 0-18 m (Butler 1980, Jensen and Armstrong 1987). Pterygostomian spine present 61 61 Outer antennular flagellum extending beyond scaphocerite by about half length of former; lower limb of rostrum broader than upper Heptacarpus moseri (Rathbun) Bering Sea to Destruction Island, Washington; 0-1100 m (Butler 1980). Outer antennular flagellum barely extending beyond scaphocerite; lower limb of rostrum scarcely broader than upper Heptacarpus sitchensis (Brandt) Resurrection Bay, Alaska, to Yaquina Bay, Oregon, 0-12 m (Butler 1980). 62 (59) Epipods on first and second pereopods 63 Epipods on first to third pereopods 65 63 Rostrum barely reaching end of first segment of antennular peduncle Heptacarpus pugettensis Jensen Seattle, Washington to near Morro Bay, California, intertidal (Jensen 1983, Wicksten 1988). Rostrum exceeding end of first segment of antennular peduncle, reaching to end of entire antennular peduncle or beyond 64 64 Rostrum usually reaching to end of scaphocerite but at least to base of antennular flagellum, ventral teeth of rostrum widely spaced Heptacarpus paludicola Holmes Tava Island, Alaska, to San Diego, California, 0-10 m (Butler 1980). Rostrum reaching end of or barely exceeding antennular peduncle, ventral teeth of rostrum crowded Heptacarpus pictus (Stimpson) San Francisco Bay, California, to off Thurloe Head, Baja California, 0-19 m (Wicksten 1988). 65 (62) Dactyls of pereopods 3-5 simple and falcate, rostrum slightly ascending over eye and with dorsal teeth most thickly set in proximal half Heptacarpus stimpsoni Holthuis Sheep Bay, Alaska, to WSW of Punta Abreojos, Baja California, 0-73 m (Butler 1980, Wicksten 1988).'"' Dactyls of pereopods 3-5 bifid and with small spines on flexor margin, rostrum not slightly ascending over eye and with dorsal teeth more widely spaced 66 66 Rostrum not reaching as far as cornea of eye, with series of teeth angled downward from anter- ior margin of carapace to tip Heptacarpus taylori (Stimpson) Queen Charlotte Sound, British Columbia, to Magdalena Bay, Baja California, 0-13 m (Green and Butler 1988). Rostrum exceeding cornea of eye, rostral teeth more widely spaced and not as clearly angled downwai'd 67 67 Spine on distal ventral flexor margin of merus of first pereopod Heptacarpus fuse imacu lata s Wicksten Santa Rosa Island, California, to Guadalupe Island, Mexico, 0-295 m (Wicksten 1986, 1988). No spine on distal ventral flexor margin of merus of first pereopod 68 596 Fishery Bulletin 88(3), 1990 68 Rostrum usually with bifid or trifid tip, exceeding cornea, merus of third and fourth pereopods with 2 spines Heptacarpus palpator (Owen) Monterey Bay, California, to Isla Espiritu Santo, Gulf of California, 0-37 m (Wicksten 1986). Rostrum usually with simple tip, not exceeding cornea, merus of third and fourth pereopods with 1 spine Heptacarpus brevirostris (Dana) Attu, Aleutian Islands, to Santa Cruz, California, 0-128 m (Butler 1980, Wicksten 1986). 69 (11) Dorsolateral antennular flagellum with accessory branch (may be free or partially fused with llagellum) ". '. " 70 Dorsolateral antennular flagellum without accessory branch (may be completely fused with flagellum) 72 70 Rostrum with 12 ventral teeth. (Rostrum with 4 dorsal teeth, 23 segments in carpus of second pereopod) Lysmata trisetacea (Heller) Red Sea to Hawaiian Islands, Gulf of California to Malpelo Island, 0-150 m (Wicksten 1983). Rostrum with 1-5 ventral teeth. (Rostrum with 5-8 dorsal teeth, 17-30 segments in carpus of second pereopods) 71 71 17 Segments in carpus of second pereopods. Rostrum with 1 ventral tooth. In life, with white vertical stripes on abdomen Lysmata galapagensis Schmitt Magdalena Bay, Baja California, and Gulf of California to Galapagos Islands, 0-10 m (Wicksten 1983). Carpus of second pereopod with 23-30 segments. Rostrum with 4-5 ventral teeth. In life, without white vertical stripes on abdomen Lysmata intermedia (Kingsley) Florida Keys to Tobago and Curacao, Azores, Gulf of California to Galapagos Islands and Peru (Wicksten 1983). 72 26-32 segments in carpus of second pereopods. Pterygostomian spine present Lysmata californica (Stimpson) Tomales Bay, California, to Panama, 0-10 m (Wicksten 1983). 21-22 segments in carpus of second pereopods. Pterygostomian spine absent Chile, Juan Fernandez Islands (Holthuis 1952). 0) _J -' 200- c 150- c D S 150- 2 100- 2 100- 50- n = 50 n-947 n-B5 50- j Q . n=H3 n = 42 n=5 1 1 1 1 0 - 1 1 1 PLATFORMS 4 MILE REEF 6 MILE RIEF PUTFORMS 4 MILE REEF 5 MILE REEF 400 350 •2-300 J. SEBASTES MYSTINUS 400 ^ 350 - SEBASTES PAUCISPINIS - -p 300- * ^ ^250 * ♦ * - -§250- ♦ -C JH ^200 ■ 2^200- QJ 0) _l _i c 150 . c 150- D o Q> 0) 2 100 ■ 2 100- 50 n=12a n=1044 n=146 50- n=147 n = 716 n-75 Q ] n _ PL4TF0RMS 4 MILE REEF 5 MILE REEF PUVTFORMS 4 MILE REEF 6 MILE REEF NOTES Love and U/estphal' Fish populations around oil platforms off Santa Barbara, California 603 effort (CPUE), defined as number of fish taken per angler hour, was highest at the 6 Mile Reef, followed by the platforms and the 4 Mile Reef (Fig. 2). At the platforms (Table 3) rockfish made up 8 of the 10 most-frequently-taken species. While olive rockfish Sebastes serranoides, widow rockfish S. entmelas, chub mackerel Scomber japotiicus comprised 55.9% of the total catch, a number of other species (including canary rockfish, Sebastes pinniger, brown rockfish S. auricu- latus, and bocaccio S. paucispinis were also taken in some numbers. AT the 4 Mule Reef, 8 of the top 10 species were rockfish, with blue rockfish Sebastes mystinus, squarespot rockfish S. fiopkinsi, S. serra- noides, and S. paucispinis dominating the catch, mak- ing up 78.8%. Nine of the top 10 species at the 6 Mile Reef were rockfish and S. mystinus, S. entomelas, S. hopkinsi, and S. paudspinis were taken in the largest numbers, forming 82.0% of the total. We compared the rank abundances of the top 10 species at the three sites with the Kendal tau coeffi- cient test (Sokal and Rohlf 1969). Based on this mea- sure, species composition was significantly correlated between natural reefs only, and species arrays from natural reefs were not correlated with that of the plat- form (Table 4). The primary differences between the reefs and platforms were the relative abundances of S. hopkinsi, S. mystinus, starry rockfish S. consteUatus, and flag rockfish S. rubrivinctus at the natural reefs and the scarcity of S. auriculatus and S. pinniger over these reefs. With only a few exceptions, all rockfish taken around the platforms were juveniles (Fig. 3). A major excep- tion was S. hopkinsi, a species in which all individuals takn were mature. In addition, a few adult S. auricu- latus and S. mystinus were caught, as well as one ver- million rockfish S. miniatus. At the 4 Mile Reef, a higher percentage of fish were mature. Again, all S. hopkinsi taken were adults, as were most S. auricu- latus and S. miniatus, many S. mystinus and S. rubri- vinctus, and a few S. serranoides. In general, more adult rockfish were taken over the 6 Mile Reef than at the other two stations. All S. hopkinsi were mature, Figure 3 (bottom facing page and below) Mean size (with 95% confidence interval) of 9 species caught aboard the sportfishmg party vessel Hornet, based on sampling during April 1975-April 1978 at three sites off Santa Barbara. Also included are .50% lengths at maturity from data in Wyllie Echeverria (1987), Love et al. (1990), and Love (unpubl. data). 400-, SEBASTES PINNIGER 400- SEBASTES RUBRIVINCTUS 350- 350- '^300- ^250. "^200- OJ c 1 so- \ -p 300- ^250- ^200- c 150- D il) 2 100- - 2 loo- 50 0 n=191 n=10 n-0 —J 0 - n=12 n=57 n-26 1 { 1 PmlFORMS 4 MILE REEF 6 MILE REEF PLATFORMS 4 MILE REEF 6 MILE REEF ■400 -[ SEBASTES SERRANOIDES 350- 'P300- ^250. g'200- 0) c 150- D OJ 2 100- : ♦ * * 50- 0- n = 831 n=750 n=20 PUTFORMS 4 MILE REEF 6 MILE REEF 604 Fishery Bulletin 88(3), 1990 as were most S. miniatus and S. rubrivinctus and many S. entomelas, S. mystinus, S. paucispinis, and S. serranoides. Discussion During the period sampled, the platforms were impor- tant sites for the Hornet's fishing effort. They were also fished by other sportfishing party vessels from Ventura Harbor, about 30 km to the east, though we are unsure of the total effort from these vessels. In both circumstances, these structures provided the vessel operators with a relatively reliable source of easily caught, though generally small, fish. The ease with which small rockfish were caught made the platforms particularly attractive to the Hornet's operators on half-day runs, when the vessels carried many inexperi- enced fishermen. The platforms were also utilized when fishing was poor at other locations for more desirable species, such as kelp bass Paralabrax clathratus. We have recently (1989) discussed platform fishing with current Santa Barbara party vessel operators. They report that though the platforms are still fished for small rockfish, kelp bass are an equal or more im- portant species during summer months. This is par- ticularly true of the eastern platforms, particularly Houchin, Hogan, and Hope (2 km east of Hogan). Though there are five platforms in this complex, only these three are routinely fished. During the season's peak, as many as 300-400 P. clathratus per trip are caught from around a single platform. Based on tag- ging studies by P. Hart (Dep. Biol. Sci., Univ. Calif., Santa Barbara 93106, pers. commun., Sept. 1988), it is likely that kelp bass move onto the platforms before summer and leave in the fall, traveling both east and west, with reported movements of as much as 32 km in 6 months. In the study conducted between 1975 and 1978, there were significant differences in both species composi- tions and mean sizes between the three sites (though the natural reefs were much more similar with each other). Much of the differences in species composition between the platforms and the natural reefs came from the relative abundance of high-relief substrate-associ- ated rockfish (such as S. consteUatus and S. rubrivinc- tus) over the reefs and their near absence around the platforms. The substrata around these structures are composed of a mixture of drill cuttings and mussels which have broken off the platform pilings. This does not appear to be suitable habitat for many rockfish species. On the other hand, a few substrate-associated spe- cies, such as juvenile 5. auriculatus and S. pinniger, were most abundant around the platforms. Sehastes auriculatus (both juveniles and adults) are most often found associated with low relief or the sand-rock inter- face, habitat most similar to that around the platforms. We do not know what substrata juvenile S. pinniger prefer, though its abundance around the platforms and scarcity over natural reefs imply it favors habitats similar to S. auriculatus. It is unclear why S. mystinus and particularly S. hop- kinsi should be relatively uncommon around the plat- forms, yet abundant on the reefs. Both species are mid- water planktivores, always found in association with high relief. A possibility is that both species take refuge in crevices at night and that these are relatively un- common on the platform pilings and cross members. Differences in depths between the sites were prob- ably responsible for only a few of the observed differ- ences in species abundances and sizes. The absence of P. clathratus from the 6 Mile Reef was undoubtedly due to the reef's relatively extreme depth. The 6 Mile Reef is situated in 74 m, well below the 46 m maximum depth of P. clathratus (Eschmeyer et al. 1983). Love et al (in review) summarizing ontogenetic Sebastes movements, noted that almost all species recruit to waters shallower than adult depth, subsequently mov- ing to deeper water as they mature. If depth were a major factor in our study, we would expect to see the smallest rockfish at the shallowest (4 Mile Reef) sta- tion. However, while both large and small individuals of many rockfish species were taken at 4 Mile Reef, the mean size of most species (except for S. entomelas) was actually larger than at the other two sites. Are the platforms we sampled aiding rockfish pro- duction or do they only act as aggregators? In other words, do rockfish recruit from the plankton to the plat- forms or do they settle out elsewhere and later migrate to these structures? Though our data is not sufficient to answer the question, there is evidence that at least some recruitment takes place around Santa Barbara platforms. In the only detailed study of a California platform, Carlisle et al. (1964) found that both 5. paucispinis and S. serranoides recruited to platform Hazel, located in 30 m of water, about 6 km inshore of our study platforms. We believe it is likely that som.e rockfish species (i.e., S. entomelas, S. mystinus, S. pau- cispinis, S. pinniger, and S. serranoides) recruit to our platforms, particularly in light of the large numbers of small fish, most only 1-2 years old. Smaller and younger fish may be present at these platforms, but their small size precludes their being taken by hook- and-line. There are several reasons that mature rockfish are rare around the platforms. While juveniles of deep- water species (i.e., S. entomelas, S. pinniger, S. pau- cispinis) are able to tolerate the warm shallow waters around these structures, the adults are not (Love et al., NOTES Love and Westphal Fish populations around oil platforms off Santa Barbara, California 605 In review). Their subadults make ontogenetic shifts into deeper water. Secondly, fishing pressure around the platforms probably crops many larger individuals. Lastly, it is possible that artificial habitat in some unknown way discourages adult rockfish. F. Matthews (Dep. Oceans and Fish., Pacific Biol. Stn., Nanaimo, B.C., Canada V9R 5K6, pers. commun.. May 1989) found that only immature rockfish occurred on an ar- tificial reef in Puget Sound. Natural reefs supported both juveniles and adults. As discussed before, most rockfish species (including S. entomelas and S. paucispims) recruit to shallow water, then migrate into deeper zones as they mature. Based on catch rates, we have noted on several occa- sions that large aggregations of these species were present around the platforms for months, only to sud- denly disappear. Hartmann (1987) tagged juvenile S. paucispinis around these platforms and noted recov- eries from as much as 148 km away. In all cases the fish were taken in water considerably deeper than tag- ging depths. Thus, it is quite possible that some rock- fish species settle out around these platforms, live there for a few years, and leave for deeper adult habitats, perhaps at considerable distances from the original recruitment site. Conclusions Platforms were important fishing locations for the sportfishing party vessel surveyed. About 18% of its total fishing time and 17% of the total catch occurred around the platforms. This compares with about 21% of fishing time and 31% of total catch at the nearby natural reefs. Rockfish dominated the catch at both the platforms and at the natural reefs. Confirming what we had first observed, there were significant differ- ences in catches between the platforms and natural reefs. Juvenile rockfish composed most of the platform catch, while mature rockfish were more abundant over reefs. While midwater rockfish species were abundant at both platform and natural reefs, species composi- tions were different, with those benthic rockfish char- acteristics of high-relief substrata absent or rare around the platforms. This work was sponsored by NOAA, Office of Sea Grant, U.S. Department of Commerce, under grant no. 04-7-158-44121 (Project r/f-39). Citations Allen. M.J., and M.D. Moore 1976 Fauna of offshore structures. South. Calif. Coast. Water Res. Proj. Annu. Rep.. Long Beach. CA, p. 179-186. Bascom, W., A.J. Mearns, and M.D. Moore 1976 A biological survey of oil platorms in the Santa Barbara Challen. J. Petrol. Technol. 28:1280-1284. Carlisle, J.G. Jr.. C.H. Turner, and E.E. Ebert 1964 Artificial habitat in the marine en\ironment. Calif. Fish Game, Fish. Bull, 124, 93 p, Eschmever, W.N., E.S. Herald, and H. Hammann 1983 A field guide to Pacific coast fishes of North America. Houghton Mifflin. Boston, 336 p. Hartmann, A.R. 1987 Movement of scorpionfishes (Scorpaenidae: Sebastes and Scorpaena) in the southern California Bight. Calif, Fish Game 73:68-79. Love, M.S., P. Morris, M. McCrae, and R. Collins 1990 Life history aspects of 19 rockfish species (Scorpaenidae: Sebastes) from the southern California Bight, NOAA Tech. Rep. NMFS 87, 38 p. Love, M.S., M. Carr, and L. Haldorson In review Ecology of substrate-associated juvenile and pre- recruit Sehnstes. Environ, Biol, Fish, Sokal. R.R., and F.J. Rohlf 1969 Biometry: The principles and practice of statistics in biological research, W.H, Freeman, San Francisco. 776 p, Wyllie Echeverria, T. 1987 Thirty-four species of California rockfishes: Maturity and seasonality of reproduction. Fish. Bull.. U.S. 85:229-250. Acl37 km) sta- tions. Diets of chum and sockeye salmon were also examined for variation with respect to the different factors when sample sizes were adequate. The relative importance of the interannual, seasonal, and geographic variations seen in the diets of coho and chinook juveniles was tested by comparing the presence and absence of a particular major prey category using a variance test of binominally distributed data (Sne- decor and Cochran 1967) for each factor individually irrespective of the others. When a value exceeded the tabulated 0.05 chi-square percentage, the null hypo- thesis that the diets were similar was rejected. Results Coho salmon General food habits Juvenile coho salmon had a relatively diverse diet with many different prey cate- gories represented in the stomachs examined for all years combined (Appendix Table 1). A high percentage (95.2%) of the stomachs contained food, and the overall stomach fullness (i 3.0) and digestion (x 2.3) codes were high indicating that many of the juvenile coho had fed prior to capture. The primary food groups consumed by juvenile coho salmon were fishes, decapod larvae, amphipods, eu- phausiids, pteropods, and copepods (Appendix Table 1). Larval and juvenile fishes were the most important prey making up 72% of the total weight and 60% of the total IRI. Rockfishes Sebastes spp., northern an- chovies Engraulis mordax, and Pacific sand lance Ammodytes hexapterus were the dominant fish taxa. Although other fish families were represented in the diet, notably the Osmeridae, Cottidae, Hexagrammi- dae, and Pleuronectidae, each of these families made up less than 1% of the total diet by percent IRI. Brodeur and Pearcy Trophic relations of juvenile salmon off Oregon and Washington 621 Table 2 Food habits by percent weight of major food categories for coho salmon, by cruise, off Oregon and Washington. 1980 1981 1982 1983 1984 1985 June May June July Aug. May June Sept. May June Sept. June July Sept. June Prey category Fishes 82.8 75.6 73.3 17.6 65.1 75.6 75.6 83.1 73.2 82.7 81.9 83.6 26.1 61.0 82.1 Euphausiids 4.4 2.0 11.0 ■>o '> 31.7 16.0 13.3 2.9 7.3 0.7 10.5 4.9 15.2 17.5 6.3 Decapods 3.5 8.4 11.9 3.5 0.3 6.9 4.2 0.6 12.8 15.6 0.1 10.9 13.2 5.6 6.3 Hyperiids 0.9 0.7 0.4 3.9 1.2 0.8 3.9 5.7 1.5 0.6 3.9 0.1 3.5 13.4 4.8 Pteropods 0.2 10.1 1.6 51.1 0.3 0.3 0.3 6.7 1.6 <0.1 2.1 — 0.1 0.5 <0.1 Copepods <0.1 0.2 0.3 0.3 0.1 <0.1 <0.1 <0.1 1.3 0.1 <0.1 0.3 0.4 <0.1 <0.1 Insects <0.1 — <0.1 0.1 0.1 0.2 0.1 <0.1 0.3 <0.1 <0.1 <0.1 40.0 <0.1 0.1 Cephalopods 6.9 0.6 0.3 - 1.2 — — 0.9 <0.1 — 1.3 - 0.4 1.0 - Other* 1.1 0.4 0.8 0.8 <0.1 1.8 1.0 <0.1 0.2 <0.1 0.1 0.1 1.2 <0.1 0.4 Predator characteristi cs No. stomachs 77 245 139 115 94 57 124 144 112 125 102 81 75 61 101 No. empty 1 15 13 3 6 1 4 5 1 7 3 7 7 3 3 Mean length 167.9 153.3 173.7 200.4 231.6 151.9 149.6 247.4 164.7 191.5 270.5 175.7 208.9 275.0 188.4 Length 81- 111- 113- 92- 138- 124- 120- 137- 117- 125- 152- 121- 144- 177- 113- range 292 238 314 390 386 188 220 410 421 349 420 247 347 366 262 Mean no. prey 17.8 26.9 45.3 123.2 38.1 9.1 12.1 97.3 35.8 37.2 61.5 20.7 35.3 80.7 35.3 Mean wt. prey 1.46 laetes. 0.57 1.15 1.40 :haetognaths, gammaric 2.28 0.59 0.50 s, isopods. cumaceans. 2.85 cirriper 0.91 1.23 es, mysids. an 5.35 1.16 0.70 0.88 d gelatinous zooplankton. 1.54 * Includes polyc Many invertebrate prey also occurred frequently or were important numerically in the diet of juvenile coho salmon, but were much less important than fish by weight (Appendix Table 1). Principal invertebrate prey were Cancer crab megalopae, the hyperiid amphipods Hyperoche medusarum and Themisto pacifica, the euphausiids Thysanoessa spinifera and Euphausia pacifica, and the pteropod LimaciHa helicina. Numer- ous other species of decapod larvae and amphipods were eaten, as well as copepods and insects, but were of lesser importance. Juvenile Loligo squid were also important prey by weight. Temporal variations Some between-cruise variabil- ity in the weight composition of the major food cate- gories was evident (Table 2). During June 1980, coho salmon consumed mainly fishes, although cephalopods were relatively important compared with other cruises. Except during July, fishes dominated the diet during 1981. Pteropods, mainly L. helicina, comprised over one-half the total weight of prey in July 1981. Fishes (mostly A. hexapterus and E. mordax) and euphausiids (mainly T. spinifera) dominated the diet during 1982. During 1983 and 1984, coho salmon juveniles fed more upon decapod larvae than during most other years (Table 2). Several anomalous prey items of more southerly origin appeared in the diet in 1983 and 1984, including the euphausiid Nyctiphanes simplex, the pteropod Evclio pyrimidnta, and the hyi.ieriid am- phipod Vihilia spp. The greatest number of major prey categories were found during July 1984 when terres- trial insects made up over a third of the biomass found in the coho salmon stomachs. One large lepidopteran, Choristoneura occidevtalis, comprised greater than 37% of the total prey IRI. The diet during June 1985 resembled that of June 1982 at both the specific and general taxonomic levels. Based on the percent of IRI for the major prey cate- gories, juvenile coho salmon diets changed somewhat as the summer progressed (Fig. 3). Fishes were the main prey during May, although decapod larvae and pteropods were important numerically. During June, fishes and decapod larvae were the primary prey, with amphipods and euphausiids of relatively minor impor- tance. By July and August, fish consumption decreased substantially, and pteropods and euphausiids were the major taxa consumed. Fishes were again the dominant food in September, but their importance was based mainly on weight because individual fish in the stom- achs were comparatively larger. In September, ptero- pods, euphausiids, and amphipods were important prey numerically. Geographic variations Diets of juvenile coho salmon in the three geographic areas were similar in that the same major prey categories were represented, despite differences in mean size of salmon among the regions. Fishes, decapod larvae, euphausiids, pteropods, and 622 Fishery Bulletin 88(4|. 1990 MONTH ^m SEPTEMBER (n - 296) L^ JULY-AUGUST (n -268) JUNE (n • 612) MAY (n - 397) FISH EUPH DECA PIER AMPH. COPE. INSE OTHER PREY CATEGORY / AREA OREGON (n - 422) £^ COLUMBIA (n • 841) Wtk VWSHINGTON (n • 310) n^'nl n y^ I »■«! m mi / FISH EUPH. DECA PTER AMPH COPE INSE OTHER PREY CATEGORY Figure 3 Juvenile coho salmon diets by mimth and area for the major prey categories as a percent of total IRI. Sample sizes (no. of fish with food) for each subset are given in the legend. All figures include data collected off Washington and Oregon, 1980-85. amphipods were the dominant prey in each area based on IRI proportions (Fig. 3). The relative proportions of the major prey categories were similar among the regions, although fishes were more important off Washington. Few differences were observed among the three regions even at the lowest taxonomic levels. Engraulis mordax and Sebastes spp. were the dominant fish prey by weight in all three regions. The Washington area differed somewhat from the other regions in that Cancer larvae were the dominant invertebrate taxa numerically, whereas L. helicina, T. spinifera, and//. meduaarum were dominant in the other regions. Chinook saimon General food habits Fishes dominated the diet of juvenile chinook salmon, occurring in over 85% of the stomachs that contained food (n = 795), and accounting for almost 87% of the total IRI for all years combined (Appendix Table 1). The overall mean stomach fullness (2.8) and digestion (2.3) states were similar to that of coho. Also resembling the diet of coho salmon, the main fish prey identified were E. mordax, Sebastes spp., and A. hexapterus, but cottid juveniles (Hemilepidotus spinosus) were also frequently eaten. Pleuronectid and agonid larvae were more common in the diets of juve- nile chinook salmon. Many invertebrate species were also represented in juvenile chinook salmon diets. The relative importance of the various invertebrate taxa was similar to that found for coho salmon. Decapod larvae, euphausiids, and hyperiid amphipods were the dominant inverte- brate groups consumed. The dominant species in these prey categories were C. oregonensis megalopae, T. spinifera, and H. medusarum, respectively (Appendix Table 1). Copepods and mysids were generally more important, and pteropods and insects less important, in comparison with juvenile coho salmon. Temporal variations Between-cruise variability in the consumption of the major prey taxa was less pro- nounced for chinook salmon than for coho salmon (Table 3). Fishes comprised 75% or more of the biomass consumed during every cruise, with the exception of July 1981 when euphausiids and pteropods were also major prey. The only other invertebrate taxa to con- tribute substantially in other cruises were cephalopods (May and June 1981), euphausiids (May 1982 and July 1984) and decapod larvae (May and June 1983). In con- trast to the diet of coho salmon, insects were unimpor- tant during July 1984 (Table 3). Seasonally, the relative IRI proportions of fishes, decapod larvae, and euphausiids were similar for chinook salmon (Fig. 4). The diets contained more major prey categories during July- August when small zooplankton prey (copepods, pteropods, decapod larvae, and hyperiid amphipods) were important numerically. This may be due to the smaller mean size of chinook salmon collected this period; in July-August, there was an influx of subyearling chinook salmon into the sam- pling area. Fishes were again the dominant prey in September, although hyperiid amphipods remained important numerically. Geographic variations The feeding patterns of juve- nile chinook salmon were fairly consistent by major tax- onomic categories among the three regions (Fig. 4). Fishes were the major prey by frequency of occurrence and weight in all three geographic areas, although the dominant species varied somewhat. Off Washington, A. hexapterus, Sebastes spp., and H. spinosus were the main fish prey consumed. Engraulis mordax and //. spinosus were the dominant prey in the Columbia region, whereas E. mordax dominated the diet off Oregon (>65% of the total IRI). Brodeur and Pearcy: Trophic relations of juvenile salmon off Oregon and Washington 623 Table 3 Food habits by percent weight of majoi food categories for Chinook sa! mon, by cruise , off Oregon and Washington. 1980 1981 1982 1983 1984 1985 June May June July Aug. May June Sept. May June Sept. June July Sept. June Prey category Fishes 92.7 84.5 75.4 38.6 97.3 81.9 94.7 91.9 89.0 80.4 98,2 89.5 75.0 88,5 87,8 Euphausiids 0.1 4.1 0.3 32.7 0.6 13.2 0.4 0.6 1.1 0.1 <0,1 0.1 10.0 6,8 6,3 Decapods 0.4 2.8 8.8 6.9 0.3 3.9 3.1 <0.1 9.4 13.8 0,4 5.2 6.2 2,3 4.9 Hyperiids — <0.1 <0.1 3.3 0.5 0.6 <0.1 3.6 <0.1 0.4 0,5 0,2 2.5 1,5 0.3 Pteropods — 0.1 <0.1 10.6 0.4 0.1 <0.1 — 0.1 <0.1 — — — <0.1 — Copepods 0.4 <0.1 <0.1 0.4 0.5 <0.1 <0.1 — <0,1 0.6 <0,1 2,9 5.8 — <0.1 Insects — <0.1 0.8 0.2 <0.1 — — 0.6 — — <0,1 — — 0,6 <0.1 Cephalopods — 7.7 14.1 — — 0.9 — 3.1 0,1 3.6 0,4 1,2 — 0,1 0,3 OtIuT* (i.2 - 0.2 6.9 0.3 <0.1 <0.1 - <0,1 - 0.5 0,2 0.6 0,1 0,5 Predator characteristics No. stomaclis 14 63 32 53 32 121 112 10 95 27 89 60 21 37 74 No. empty 0 7 2 6 2 5 2 0 3 2 1 11 •1 3 3 Mean length 232.5 181.8 196.5 152.9 165.0 218.9 203.7 322,9 192,6 205,7 220.3 182,2 147.2 215,1 209,9 Length 189- 126- 110- 87- 120- 123- 119- 134- 118- 124- 129- 105- 109- 138- 101- range 283 290 331 347 347 400 350 435 396 287 325 370 251 412 354 Mean no. prey 14.6 27.8 19.0 17.0 37.4 10.8 9.4 12,0 17.1 16,0 13.7 21,9 17.4 12,3 19.2 Mean wt. prey 1.85 laetes. 1.23 chaetogr 1.47 aths. c 0.44 amniari 0.37 is, isop 2.44 1.50 ids, cumaceans. 1.14 cirriper 0.99 0.78 es. mysids, an 3.22 0.98 1 gelatinous zor 0.27 0.75 plankton. 0.85 * Includes polyc / MONTH ^M SEPTEMBER (n • 132) iZD JULY-AUGUST (n • 98) JUNE (n . 300) MAY (n • 265) "./ ^/ FISH EUPH DECA. PTER AMPH COPE CEPH OTHER PREY CATEGORY AREA cm OREGON (n - IS-l) ^^ COLUMBIA (n - 478) ^M WASHINGTON (n • 163) ^3^ ■/- / FISH EUPH DECA PTER. AMPH. COPE. CEPH. OTHER PREY CATEGORY Figure 4 Juvenile chinook salmon diets by month and area for the major prey categories as a percent of total IRI, Sample sizes (no, of fish with food) for each subset are given in the legend. All figures include data collected off Washington and Oregon. 1980-85, There were few consistent patterns observed among the invertebrate taxa consumed between the different areas (Fig. 4). Hyperiid amphipods were rarely found in chinook salmon stomachs collected off Washington as opposed to the other regions. The large numbers of decapod larvae eaten off the Columbia River were mainly C. oregonensis and C. magister larvae. Chum salmon General food habits Of the 109 chum salmon stom- achs examined, lUl (92.6%) contained food; however, overall mean stomach fullness (2.4) and digestion (2.1) were low. The diet was dominated by zooplanktonic crustaceans, particularly euphausiids, calanoid cope- pods, and hyperiid amphipods (Appendix Table 2). Euphausiids (mostly juveniles', pacifirn and T. spini- fera) accounted for over 54% of the total weight and 47% of the total IRI for all cruises combined. A tax- onomically diverse array of hyperiids, copepods. and decapod larvae were also consumed. These taxa fre- quently occurred in the stomachs and were important numerically, but were of lesser importance gravi- metrically. Chaetognaths and larvae and juveniles of several fish species were the dominant non-crustacean prey (Appendix Table 2). Temporal variations Chum salmon diets varied con- siderably during the 5 years examined (Table 4). Some 624 Fishery Bulletin 88(4), 1990 Table 4 Summarv of fo jd haljits by percent weight of majoi prey categories for juvenile chu m salmon by year, off Oregon and | Washington. 1981 1982 1983 1984 1985 Prev category Fishes 1.1 - 15.4 57.2 24.1 Euphausiids 78.3 60.7 20.,5 13.7 33.0 Decapods <0.1 S.6 1.0 — 2.0 Hyperiids 4.S 7.8 1.0 1.5 0.7 Pteropods 0.8 <0.1 — - — Copepods 6.0 21.4 1.0 3.8 36.6 (.'haetognaths l.'J 1.4 60. 1 17.5 2.2 Other* - <0.1 0.9 6.1 1.3 Predator characteristics No. stomachs 24 33 l.S 8 29 No. empty 2 1 0 1 4 Mean length 190. (i 161.6 116.2 165.4 132.2 Length 108- 97- 103- 126- 115- range 237 223 133 214 155 Mean no. prey 47.4 8.0 78.1 20.1 83.4 Mean wt. prey 0.69 laetes, ot 0.04 0.23 0.19 0.22 tracods, mysids, insects, and * Includes polyc larvaceans. Table 5 Total number of prey taxa. Shannon Weave ■ niche breadth (H' ), maximun niche value (H'„,,^), and evenness (E) for each | salmon species o ff Oregon and Wash ington. for all samples combined. Total number Species of prey taxa H' H'max E Coho salmon 157 4.55 7.29 0.62 Chinook salmon 136 5.06 7.09 0.71 Chum salmon 51 4.35 5.67 0.77 Sockeye salmon 36 2.74 5.17 0..53 of this may be due to the months and areas sampled in different years. Euphausiids were the main food item by weight during 1981 and 1982. Although euphausiids were important during the later years, their contribu- tion to the diet was less compared with chaetognaths, larval fishes, and calanoid copepods. Sockeye salmon General food habits The diet of juvenile sockeye salmon was similar to that of chum salmon. Juvenile euphausiids, calanoid copepods, chaetognaths, and fish larvae were consumed by the small number of fish ex- amined (Appendix Table 2). Although a large number of prey taxa were identified, only a few were impor- tant. Among these were the euphausiid T. ^pinifcra CD 50 .i COHO 1 * 4 0 30 ■ 1 1 20; 1 1 0 00 - 1 1 : ; . JUN MAYJUNJULAUQ MAYJUNSEP MAYJUNSEP JUN 1980 1981 1982 1983 CRUISE JUN 1985 I I 1- O cr CO n JUN MAYJUNJULAUQ MAYJUNSEP MAYJUNSEP JUNJULSEP 1980 1981 1982 1983 1984 CRUISE JUN 1985 Figure 5 Shannon-Weaver niche breadth (H ) valui'S fur juvenile coho and chinook salmon for each cruise period. and larval osmerid fishes. Chaetognaths were found in substantial numbers but were well digested and not identifiable to species. The relatively small sample size of sockeye salmon stomachs precluded a detailed analysis of diet variability. All salmon species IMiche breadth Niclie bi-eadths, maximum possible niche breadths, and evenness values are given for each salmon species in Table 5. Chinook salmon had the highest overall diversity of prey taxa (H' = 5.06), which was consistent with the high numbers of prey taxa found per stomach, and a high evenness ratio. Coho and chum salmon also consumed a diverse array of prey taxa, although coho stomachs frequently contained small numbers of prey items and showed low evenness overall. Chum salmon had a high diversity (H' = 4.35), despite a substantial amount of unidentified and digested prey, and had the highest evenness (0.77) of all salmon species. The overall prey diversity of sockeye salmon was quite low (H' = 2.74), wliich may be due in part to the small sample size and advanced state of digestion (mean digestion code = 2.0) of prey which prevented identification to lower taxonomic levels. Brodeur and Pearcy Trophic relations of juvenile salmon off Oregon and Washington 625 Table 6 Diet overlap by weight among all salmon species off Oregon and Washington, for all cruises combined. Species O 40 Coho Chinook Chum Sockeye Coho — Chinook 54 — Chum 26 16 Sockeye 25 24 33 Jun May Jun Jul Aug May Jun Sep May JunSep Jun Jul Sep Jun 1980 1981 1982 1983 1984 1985 CRUISE Figure 6 Diet overlap between juvenile coho and chinook salmon for each cruise period. 1980-85. Table 7 Diet ovei lap Ijy weight lit tween juvenile coho and chinook 1 salmon tV r collections off 0 regon and Washington from which | 10 stomachs of each species were ex amined. A.lso shown are the intraspecific overlap values for the same species at adja- cent stations collected on the same day. Distance PSl offshore — Date Location (km) Both Coho Chinook 1982 June 1 45°00' 124°05- 6.1 57 1 44°4r 124°24' 18.0 69 7 47°00' 124''25' 18.5 53 8 46°41' 124°18' 17,4 47 18 40 8 46°4r 124°29' 31.8 49 10 46°30' 124°25' 27.6 45 38 31 10 46*'30' 124°18' 18.3 33 1983 Sept. 20 45°40' 124°03' 8.9 91 22 45°20' 124°0r 3.7 62 59 61 99 45° 17' 124°or 3.7 68 1984 June 8 47°00' 124°25' 18.3 5 1985 June 18 46°iy i24°ir 9.0 56 25 47°40' i24°5:r 37.3 15 Prey diversity varied greatly among cruises for juve- nile coho (Fig. 5). Generally lower values occurred dur- ing 1981, especially in July, when pteropods were very important in the diet. Coho salmon had higher diver- sity values in 1983 and 1984 than in 1981. Chinook salmon showed many of the same interannual and seasonal trends in prey diversity as coho salmon, ex- cept that prey diversity was not appreciably lower in 1981 (Fig. 5). Dietary overlap Diet overlap by weight among the salmon species for all cruises combined was generally low, with none of the six species pairs showing signifi- cant (>60%) overlap (Table 6). The diets of juvenile coho and chinook salmon were most similar (PSI = 54%), reflecting their common foraging on many of the same euphausiid, larval decapod, and fish species. Chum salmon diets showed the least similarity to the other species, which may be a function of the poor digestive state of the stomach contents of this species. To examine the finer-scale variability in diet similar- ity between coho and chinook salmon juveniles, diet overlap was calculated for each cruise month (Fig. 6). With the exception of 1983, which showed the highest overall similarity, monthly overlap values were highly variable within years containing more than one cruise. No consistent patterns were observed between years for the same months. A wide range of diet overlap values was observed for the 13 collections from which 10 stomachs of coho and chinook were analyzed (Table 7). Overlap was highest in the collections from September 1983, due mainly to the common utilization of E. mordax and several hyperiid amphipod species. Intermediate overlaps were generally observed during June 1982, resulting main- ly from consumption of the same euphausiid species by both predators. There appeared to be no relationship between diet overlap and the inshore-offshore location of the collection. Diet overlap between juvenile coho and chinook at a particular station was generally higher than intraspecific overlap for either species at adjacent stations (Table 7). Analysis of dietary variations The contingency table analysis for [iresence or absence of the four most- commonly-occurring major prey categories showed that there were generally highly significant variations (P< 0.001) in diets of coho and chinook juveniles by 626 Fishery Bulletin 88(4), 1990 Table 8 Results of the chi square analysis analyzing each factor in- dependently of the other two for major prey categories of juvenile coho and chinook salmon off Oregon and Washing- ton. Degrees of freedom are in parentheses below each factor. Yeart Month Area Prey category (3) (3) (2) Coho salmon Fishes 75.32*** 83.38*** 51.54*** Euphausiids 31.19*** 106.38*** 10.71** Decapods 39.48*** 38.42*** 7.80* Amphipods 74.61*** 53.21*** 2.79 n.s. Chinook salmon Fishes 68.25*** 65.95*** 2.51 n.s. Euphausiids 33.03*** 3.15 n.s. 15.48*** Decapods 46.92*** 49.15*** 5.96 n.s. Amphipods 14.36** 60.71*** .01, ***P«0.001, n.s. F>0.{ 13.15** 5 *P<0.05, **P<0 T Tested for 1981- -S4 only. year and month (Table 8). The only comparison which showed non-significant variation by year or month was the monthly variation in euphausiid occurrences for juvenile chinook salmon. Three prey categories did not show significant variation (P>0.05) when analyzed by area. Comparisons by area showed less significant variation than by year and month for both coho and chinook salmon juveniles for most of the major prey categories (Table 8). Examination of inshore-offshore variations in major prey composition for four cruises (two each for coho and chinook salmon) showed generally few significant variations by occurrence for the dominant prey cate- gories (Table 9). Although the species and life-history- stage composition of the prey was different in inshore and offshore collections (i.e., more Sebnstes and En- graulis larvae offshore, and more iuvenile Atnmodyteti, Clupea, and Hemilepidotus inshore), all cross-shelf variations in total fish occurrences were not significant (P>U. 05). The most significant differences were for decapod larvae, and were due to higher occurrences oi Cancer spp. megalopae in chinook salmon stomachs inshore in May 1982 and lower occurrences in coho salmon stomachs inshore in June 1984. Discussion Overall food habits This study represents the first detailed description of the diets of several species of sympatric juvenile salmon Table 9 Results of the chi-square analysis analyzing inshore-offshore variations for major prey categories of juvenile coho and chinook salmon off Oregon and Washington. All significances are at 2 degrees of freedom. Prey category Coho salmon Chinook salmon May 81 June 84 May 82 June 85 Fishes Euphausiids Decapods Amphipods 3.86 n.s. 3.14 n.s. 3.83 n.s. 3.14 n.s. 5.82 n.s. 2.66 n.s. 15.07*** 15.88*** 1.34 n.s. 1.40 n.s. 1.82 n.s. 9.76** 9.00* 1.82 n.s. 3.23 n.s. 0.41 n.s. *P<0.05, **P<0.01, ***P<0.001, n.s. P>0.05 in coastal marine waters off Washington and Oregon. Despite generally low diet overlaps at the lowest tax- onomic levels, there were some similarities in the major prey groups consumed by the salmon species. Juvenile chinook salmon were primarily piscivorous, consuming a variety of larval and juvenile fishes. The diet of coho salmon consisted of both fishes and large zooplanktonic crustaceans, such as euphausiids, crab megalopae, and hyperiid amphipods. Chum and sockeye salmon diets were more diverse than the diets of coho and chinook salmon, with fishes, small crustaceans (euphausiid fur- cilia and juveniles, crab larvae, and copepods), and chaetognaths being important prey. Our findings are consistent with what is known of the marine food habits of juvenile salmon off Wash- ington and Oregon and British Columbia. Juvenile chinook salmon tend to be more piscivorous than juvenile coho for the same-sized predator (Healey 1978, 1980; Peterson et al. 1982; Emmett et al. 1986). Coho collected dm'ing this study, however, consumed a larger overall mean length of fish prey relative to predator length (Brodeur In press). Juvenile northern anchovy, Pacific sand lance, and rockfishes were the dominant fish species eaten off Washington and Oregon (Peter- son et al. 1982, Emmett et al. 1986), and herring and Pacific sand lance were the main fish species consumed off British Columbia (Healey 1978, 1980). Many of these prey fish species tend to be heavily pigmented and are often associated with the neustonic layer in coastal waters (Brodeur et al. 1987b, Shenker 1988, Brodeur 1989). Macrozooplankton, such as euphausiids, hyperiid am- phipods, and crab larvae, are also readily consumed by these juvenile salmon. These macrozooplankton prey may be easily detected due to their large size or darkly- pigmented eyes (Peterson et al. 1982) and occur in large aggregations near the surface (Brodeur et al. 1987b, Shenker 1988). Terrestrial insects, which may be blown Brodeur and Pearcy Trophic relations of juvenile salmon off Oregon and Washington 627 to sea in large numbers during favorable meteorolog- ical conditions, can also contribute substantially to the diets of juvenile coho salmon and other salmonids (Brodeur 1989). Fishes were much less important in the diets of juve- nile chum and sockeye salmon; however, the mean and maximum size of the individuals of these species ex- amined in this study were smaller than those of coho and chinook salmon. Juvenile euphausiids were a major component in the diet of chum and sockeye salmon in this study. Peterson et al. (1982) found juvenile eu- phausiids were a major prey of chum juveniles off Oregon, whereas copepods, larvaceans, and hyperiid amphipods were most important off British Columbia (Manzer 1969; Healey 1978, 1980). Both chum and sockeye consumed a greater number of chaetognaths than coho and chinook, and the importance of these and other soft-bodied prey may be greatly underestimated in most studies because they are probably digested very rapidly in salmon stomachs (Black and Low 1983). Patterns in dietary variability The pronounced interannual differences in the diets of most species were expected, even when the collections from same months and areas were examined. Ocean- ographic conditions varied greatly among the years, with both relatively strong (1982 and 1985), weak (1983 and 1984), and highly variable (1981) upwelling occur- ring during the study period (Fisher and Pearcy 1988). In addition, a strong El Nifio event dominated ocean conditions in coastal waters of the northeast Pacific during the summer months of 1983 and 1984, greatly affecting primary and secondary productivity and fish production (Mysak 1986, Pearcy and Schoener 1987). Fish prey, as a proportion of the total diet by weight, was generally invariant for the same months among the different years. However, the species composition alternated between coastal and offshore taxa (as iden- tified by Richardson and Pearcy [1977]) depending on the prevalent hydrogi'apliic regime in the various years. Several coastal taxa (Ammodytes hexapterus, Clupea harengus pallasi, Hemilepidotus spinosus, and Osmeri- dae) were more prevalent during strong upwelling years, whereas offshore taxa (Engraulis mordax, Sebastes spp., Ronquilusjordani, and pleuronectid lar- vae) were eaten more frequently during poor upwell- ing years. Although ichthyoplankton collections were not made during every year of the stomach sampling, one limited study lends support to our diet observa- tions. In a series of plankton tows along one transect off the central Oregon coast in 1983, Brodeur et al. (1985) found high abundances of offshore fish taxa at inshore stations compared with past studies. The lar- vae of the northern anchovy were unusually abundant that year, and juvenile anchovy made up the majority of the diet-by-weight of juvenile coho and chinook salmon during September 1983. Osmerid larvae, gen- erally the dominant larvae inshore off Oregon (Richard- son and Pearcy 1977), were found in low abundance in both the plankton collections and fish stomachs dur- ing 1983. Several invertebrate taxa showed substantial inter- annual variation. The pteropod L. helicina was one of the most important prey consumed in 1981, but was relatively unimportant in other years. The dominant inshore euphausiid T. spimfera was extremely abun- dant in the stomachs following periods of active upwell- ing, but was rarely consumed during the El Niiio of 1983 and early 1984 (Brodeur 1986). Several species of decapod larvae were present in greater numbers in the diets of all salmon species during 1983 and early 1984. Many other El Nino-related anomalies that were observed in the diet of coho salmon during 1983 were described by Pearcy et al. (1985). The diets of coho salmon during 1984 showed above-average abundances of terrestrial insects, which presumably were blown offshore by anomalous winds during the summer of 1984 (Brodeur 1989). Strong seasonal variations in feeding habits of juve- nile coho and chinook salmon were evident in the years that had multiple cruises over the 5-month sampling period. Some of the variability may have been due to the seasonal increase in the mean size of the salmon, which allowed a greater size range of prey to be con- sumed later in the summer (Brodeur In press). How- ever, much of the diet variation may have been due to seasonal variations in the abundance of meroplanktonic prey (e.g., decapod and fish larvae). The timing and duration of spawning, larval development, and settle- ment to benthic juvenile habitat are relatively fixed for most meroplanktonic species (Lough 1975, Richardson and Pearcy 1977, Parrish et al. 1981) such that their seasonal occurrence in the plankton and the diets of juvenile salmon are relatively predictable. Exceptions may occur during anomalous years such as during an El Nino (Bailey and Incze 1985, Brodeur et al. 1985). Several of the major holoplanktonic taxa (e.g. , euphau- siids, hyperiids, and pteropods) also showed a consis- tent seasonal succession of developmental stages, but many species were present throughout the summer period (Brodeur 1990). It was not surprising that geographic (north-south) variations in the diet composition of most salmon species were not as substantial as temporal variations. because most of the prey species are distributed throughout the range of latitudes we sampled. Similar results were found for adult salmonids and nonsal- monid species by Brodeur et al. (1987a). However, oceanographic regimes may be quite variable between 628 Fishery Bulletin 88(4), 1990 regions (Brodeur and Pearcy 1986), which could result in different feeding conditions for juvenile salmonids within each region. The Columbia River region may be quite different hydrographically and biologically from the upwelling regions off Washington and Oregon due to the presence of a warm, low-salinity plume extend- ing over much of the Columbia River area during the summer. Oceanic species (e.g., Sebastes spp., E. paci- fica) rarely occurred in the stomachs collected from the Columbia region, whereas £■. niordax, a species whose northern subpopulation spawns within the relatively stable Columbia River plume (Richardson 1980), was generally well represented in salmon stomachs col- lected in this region. The relatively minor cross-shelf variations in the major prey taxa consumed by coho and chinook were not expected, considering that many studies have found substantial variations in cross-shelf species distribu- tions (Peterson and Miller 1976, Richardson and Pearcy 1977, Richardson et al. 1980, Shenker 1988). However, the frequency of occurrence of major prey may not be a representative measure of the diet from a particular area. The larvae of many meroplanktonic taxa (e.g., Sebastes spp., Hemilepidotus spp., and Cancer spp.) are generally found offshore and progressively migrate or are transported inshore as they grow prior to settling to an inshore benthic habitat as juveniles (Richardson et al. 1980, Shenker 1988). Euphausiids consumed in our inshore study area were mainly T. spinifer-a, whereas E. pacifica were consumed in the offshore area. These euphausiid species have little overlap in distribution (Hebard 1966). Hyperiid amphipods were also represented by an inshore species, Hyperoche medusarum, and an offshore species, T. pacifica (Lorz and Pearcy 1975). These and other cross-shelf species differences were not detectable when analyzing stom- ach contents at higher taxonomic levels. However, species-level distinctions may be irrelevant to a forag- ing predator, if the size, energy content, and behavior of both prey species are similar. The high prey-diversity and generally high niche- breadth values agree with previous studies which indicate that many of these salmon species are not specialists in their oceanic feeding modes, but rather consume any available prey within the proper size- range. Many of the same geographic, interannual, and seasonal patterns found in the feeding habits of the adult salmonids and pelagic nonsalmonid species (Brodeur et al. 1987a) were found in our study. These similar patterns suggest that the zooplankton and ichthyoplankton population cycles, which are intricately coupled to seasonal production cycles, may be impor- tant determinants of the feeding ecology of these salmon species. The ability to switch to alternate prey. when preferred prey are limiting, may be an important factor in the marine survival of salmon. The high overlaps between coho and chinook juve- niles seen for some craises or individual collections may signify that some competition for prey may be occur- ring. This interaction could be particularly acute since coho and chinook exhibit a high degree of spatial overlap in their distributions (Pearcy and Fisher In press). This would be conceivable only if prey resources were limiting. Because of the highly opportunistic feeding mode of most salmonids and the substantial heterogeneity in the physical and biological environment, a large-scale study over extended time-periods may be necessary to adequately describe the feeding dynamics of juvenile salmon. Major departures from the long-term mean oceanographic conditions, as exemplified by an El Nino event, can strongly affect the feeding ecology of many pelagic planktivores. Fulton and LeBrasseur (1985) have hypothesized that a northward shifting of the Subarctic Boundary from its normal position intersec- ting the coast off Oregon or northern California to well above Vancouver Island, British Columbia, as occurs in El Niiio conditions, may expose salmon and other pelagic predators to a novel suite of available prey, with a corresponding downward shift in prey size. Grover and 011a (1987) found that a smaller mean size of copepod was consumed by larval sah\eT\shAnoplopo)fin fimbria, during the El Nino year of 1983 than during 1980, a year of relatively normal oceanographic conditions. Our study also demonstrated anomalies in species composition in the diet during 1983 and early 1984 compared with other years; relatively large northern euphausiids were replaced by much smaller decapod larvae (i.e., Graspidae, Porcellanidae, and Pinnotheri- dae) and euphausiids (Nycti phones simplex) of southern origin (Brodeur 1986). Similar interannual shifts were seen in the fish prey-size spectrum consumed by coho and chinook salmon juveniles with generally smaller prey consumed during 1983 and 1984 (Brodeur In press). Consumption of smaller prey must be balanced by consumption of a greater number of prey of equi- valent caloric content in order to maintain the similar growth rates seen for coho salmon during the early summers of 1983 and 1984 as non-Nino years (Fisher and Pearcy 1988). Unless prey are more aggregated during El Nino years, smaller prey would require a substantial increase in time and energy spent forag- ing, relative to time spent avoiding predators. This in- creased foraging time, at the expense of predator avoidance, may have led to the low coho and chinook salmon survival in the ocean during the El Nifio years (Johnson 1988). Brodeur and Pearcy" Trophic relations of juvenile salmon off Oregon and Washiington 629 Acknowledgments We thank A. Chung, J. Fisher, H. Lorz, K. Sandoy, J. Shenker, W. Wakefield, and tlie captains and crew of the research vessels for their assistance in collec- ting and processing the salmon used in our analyses. Special thanks go to R. Emmett of the Northwest Alaska Fisheries Center Hammond Laboratory for lending his expertise in programming and data anal- ysis. R. Emmett, G. McCabe, C. Simenstad, and an anonymous reviewer provided helpful comments on the manuscript. Initial funding for the sample collections and stomach processing was provided by the Oregon State Univer- sity Sea Grant College Program (Grant No. NA 81 AA- D-00086) and the Northwest and Alaska Fisheries Center Seattle Laboratory (NA-85-ABH00025). Fund- ing for data analysis and writing was supplemented by a grant from the National Coastal Research and Devel- opment Institute, Newport, Oregon. Citations Bailev, K.M., and L.S. Incze 198.5 El Nino and the eai'ly life history and recruitnient of fishes in temperate marine waters. In Wooster, W.S., and D.L. Fluharty (eds.), El Nino North: Nino effects in the eastern subarctic Pacific Ocean, p. 143-165. Wash. Sea Grant Puhl. WSG-WO 85-3, Univ. Wash., Seattle. Ba.\, N.J. 1983 Early marine mortality of marked juvenile chum salmon {Onrorhy>irhi(!< keto) released into Hood Canal, Puget Sound. Washington, in 1980. Can. J. Fish. Aquat. Sci. 40:426-435. Black, E.A., and C.J. Low 1983 Ctenophores in salmon diets. Trans. Am. Fish. Soc. 112:728. Brodeur, R.D. 1986 Northward displacement of the euphausiid Nyctiphanes tiimplex Hansen to Oregon and Washington waters following the El Nino event of 1982-83. J. Crust. Biol. 6:686-692. 1989 Neustonic feeding by juvenile salmonids in coastal waters of the Northeast Pacific. Can. J.^Zool. 67:1995-2007. 1990 Abundance and distribution patterns of zooplankton along the Oregon and southern Washington coasts during the sum- mer of 1981. Fish. Res. Inst. Publ. FRI-UW-9003. Univ. Wash., Seattle. 33 p. In press Ontogenetic variations in the type and size of prey consumed by juvenile coho, Oricoyiiynchus ki.'iutch, and chinook, (>. tKliiiiryl^rha, salmon. Environ. Biol. Fishes. Brodeur, R.D., and W.G. Pearcv 1986 Distribution and relative abundance of pelagic non- salmonid nekton off Oregon and Washington. 1979-84. NOAA Tech. Rep. NMFS 46, 85 p. 1987 Diel feeding chronology, gastric evacuation and estimated daily ration of juvenile coho salmon, Oncorhynchus kisiilrli (Walbaum), in the coastal marine environment. .J. Fish. Biol. 31:465-477, Brodeur. R.D., D.M. Gadomski. W.G. Pearcv, H.P. Batchelder, and C.B. Miller 1985 Abundance and distribution of ichthyoplankton in the upwelling zone off Oregon during anomalous El Nino condi- tions. Estuarine Coastal Shelf Sci. 21:365-378. Brodeur, R.D., H.V. Lorz. and W.G. Pearcy 1987a Food habits and dietary variability of pelagic nekton off Oregon and Washington, 1979-1984. NOAA Tech. Rep. NMFS .57. 32 p, Brodeur, R.D., W.G. Pearcy, B.C. Mundy, and R.W. Wisseman 1987b The neustonic fauna in coastal waters of the Northeast Pacific: Abundance, distribution, and utilization by juvenile salmonids. Oreg. Sea Grant Coll. Prog. Publ. ORESU-T-87 -001., Oreg. State Univ., Corvallis. 61 p. Emmett, R.L.. D.R. Miller, and T.H. Blahm 1986 Food of juvenile chinook, Oncorhynchus tshawytsclin, and coho. 0. kisutch, salmon off the northern Oregon and Southern Washington coasts, May-September 1980. Calif. Fish Game 72:138-46. Fisher, J. P., and W.G. Pearcy 1985 Studies of juvenile salmonids off the Oregon and Washington coast, 1985. Oreg. Sea Grant Coll. Prog. Publ. ORESU-T-85-004, Oreg. State Univ., Corvallis, 31 p. 1988 Growth of juvenile coho salmon {Oncorhyvchus kiKutch) in the ocean off Oregon and Washington, USA in years of differing coastal u|)welling. Can. J. Fish. Aquat, Sci, 45: 1036-1044, Fulton, J.D., and R.J. LeBrasseur 1985 Interannual shifting of the Subarctic Boundary and some of the biotic effects on juvenile salmonids. hi Wooster. W.S., and D.L. Fluharty (eds.). El Nino North: Nino effects in the eastern subarctic Pacific Ocean, p. 237-247. Wash. Sea Grant Publ. WSG-WO 85-3. Univ. Wash.. Seattle. Furnell. D.J., and J.R. Brett 1986 Model of monthly marine growth and mortality for Babine Lake sockeye salmon {Oncorhynchus nerka). Can. J. Fish. Aquat. ,Sci, 4.'-l:y99-l(l()4, Grover. J.J., and B.L. 011a 1987 Effects of an El Nino event on the food habits of larval sablefish, Anoplopoma fimbria, off Oregon and Washington. Fish. Bull., U.S. 85:71-79. Gunsolus, R.T. 1978 The status of Oregon coho and recommendations for managing the production, harvest, and escapement of wild and hatchery-reared stocks. Oreg. Dep. Fish. Wildl., Fish Div. Proc. Rep.. 59 p. Healey, M.C. 1978 The distribution, abundance, and feeding habits of juve- nile Pacific salmon in Georgia Strait. British Columbia. Can. Fish, Mar, Serv. Tech. Rep. 788, 49 p. 1980 The ecology of juvenile salmon in Georgia Strait, British Columbia. In McNeil, W..J., and D.C. Himsworth (eds.), Salmonid ecosystems of the North Pacific, p. 203-229. Oreg. State Univ. Press, Corvallis. Hebard. J.F. 1966 Distribution of Euphausiacea and Copepoda in relation to oeeanographic conditions. Ph.D. thesis. Oreg. State Univ., Corvallis. 85 p. Johnson. S.L. 1988 The effects of the 1983 El Nino on Oregon's coho (0,i- curhynchus kisutch) and chinook (0. tshttwytsrha) salmon. Fish. Res. (Amst.) 6:105-123. Linton, L.R.. R.W. Davies, and F.J. Wrona 1981 Resource utilization indices: An assessment. J. .'^nim. Ecol. 50:283-292. 630 Fishery Bulletin 88(4). 1990 Lorz. H.V., and W.G. Pearcy 1975 Distribution of hyperiid amphipods off the Oregon coast. J. Fish. Res. Board Can. 32:1442-1447. Lough, R.G. 1975 Dynamics of crah larvae (Anomura, Brachyura) off the central Oregon coast. 1969-1971. Ph.D. thesis, Oreg. State Univ., Corvallis, 229 p. Mathews, S.B., and R. Buckley 1976 Marine mortality of Puget Sound coho salmon (Onrarhyn- chits kisutch). J. Fi.sh. Res. Board Can. 33:1677-1684. Manzer, J.I. 1969 Stomach contents of juvenile Pacific salmon in Chatham Sound and adjacent waters. J. Fish. Res. Board Can. 26: 2219-2223. McGie, A. 1984 Commentary: Evidence for density dependence among coho salmon stocks in the Oregon Production Index Area. In Pearcy, W.G. (ed.). The influence of ocean conditions on the production of salmonids in the North Pacific, p. 37-49. Oreg. Sea Grant Coll. Prog. ORESU-W-83-001, Oreg. State Univ., Corvallis. Mysak, L.A. 1986 El Nino, interannual variability and fisheries in the north- east Pacific Ocean. Can. .J. Fish. Aquat. Sci. 43:464-497. Nickelson, T.E. 1986 Influences of upwelling, ocean temperature, and smolt abundance on marine survival of coho salmon (Oncorhynchus kisutch) in the Oregon Production area. Can. J. P'ish. Aquat. Sci. 43:.527-535. Parker, R.R. 1968 Marine mortality schedules of pink salmon of the Bella Coola River, central British Columbia. J. Fish. Res. Board Can. 25:757-794. 1971 Size selective predation among juvenile salmonid fishes in a British Columbia inlet. J. Fish. Res. Board Can. 28: 1,503-1510. Parrish, R.H., C.S. Nelson, and A, Bakun 1981 Transport mechanisms ami reprmluctive success of fishes in the California Current. Biol. Oceanogr. 1:175-203. Pearcy, W.G. 1984 Where do all the coho go? The biology of juvenile salmon off the coasts of Oregon and Washington. In Pearcy, W.G. (ed.). The influence of ocean conditions on the production of salmonids in the North Pacific, p. 50-60. Oreg. Sea Grant Coll. Prog. Publ. ORESU-W-83-001, Oreg. State Univ.. Corvallis. 1988 Factors affecting survival of coho salmon off Oregon and Washington. In McNeil, W.J. (ed.). Salmon production, management, and allocation, p. 67-73. Oreg. State Univ. Press. Corvallis. Pearcy, W.G., and J. P. Fisher 1988 Migrations of coho salmon, Onrnrhipirhus kii^utrh, dur- ing their first summer in the ocean. Fish. Bull.. U.S. 86: 173-195. In press Distribution and abundance of juvenile salmonids off Oregon and Washington. 1981-1985. NOAA Tech. Rep. NMFS 93. Pearcy, W.G., and A. Schoener 1987 Changes in the marine biota coincident with the 1982- 1983 El Nino in the northeastern subarctic Pacific Ocean. J. Geophys. Res. 92:14417-14428, Pearcy, W., J. Fisher, R. Brodeur, and S. .Johnson 1985 Effects of the 1983 El Nino on coastal nekton off Oregon and Washington. In Wooster. W.S.. and D.L. Fluharty (eds.). El Nino North: Nino effects in the eastern subarctic Pacific Ocean, p. 188-204. Wash. Sea Grant Publ. WSG-WO 85-3. Univ. Wash.. Seattle. Peterman, R.M. 1982 Nonlinear relation between smolts and adults in Babine Lake sockeye salmon (Oncorhynchus nerka) and implications for other salmon populations. Can. J. Fish. Aquat. Sci. 39: 904-913. Peterman, R.M., and R.D. Routledge 1983 E.\(ierimental management of Oregon coho salmon (On- cuiiiynchus kisutch): Designing for yield of information. Can. J. Fish. Aquat. Sci. 40:1212-1223. Peterson, W.T.. and C.B. Miller 1976 Zociplankton along the c(jntinental shelf off Newport, Oregon, 1969-1972: Distribution, abundance, seasonal cycle, and year-to-year variations. Oreg. Sea Grant Coll. Publ. ORESU-T-76-002, Oreg. State Univ., Corvallis. Ill p. Peterson, W.T., R.D, Brodeur, and W.G. Pearcy 1982 Food habits of juvenile salmon in the Oregon coastal zone, ,lune 1979. Fish. Bull., U.S. 80:841-851. Petraitis. P.S, 1979 Likelihood measures of niche breadth and overlap. Ecology 60:703-710. Pielou, E.C. 1977 Mathematical ecohigy. .lohn Wiley, NY, 385 p. 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. Richardson, S.L. 1980 Sjiawning biomass and early life of northern anchovy, Kiiiiraiilh monla.r, in the northern sub-population off Oregon and Washington. Fish. Bull.. U.S. 78:855-877. 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. Richardson, S.L., J.L. Laroche. and M.D. Richardson 1980 Larval fish assemblages and associati(ms in the North- east Pacific Ocean along the Oregon coast, winter-spring 1972-1975. Estuarine Coastal Mar. Sci. 11:671-699. Shenker, J.M. 1988 Oceancigraphic associations of neustonic larval and juve- nile fishes and Dungeness crabs off Oregon, 1984. Fish. Bull., U.S. 86:299-317. Snedecor, G.W., and W.G, Cochran 1967 Statistical methods. 6th ed. Iowa State l.'niv. Press, Ames, 593 p. Wallace, R,K. 1981 An assessment of diet overlaii indexes. Trans. Am. Fish. Soc. 110:72-76. Brodeur and Pearcy Trophic relations of juvenile salmon off Oregon and Washington 631 Append ix Table 1 Percent frequency occurrence (F), percent numbei (N). percent weight (W) and perceni inde.x of relative importance (IRI) of food | items in juvenile coho and chinook salmon stomachs taken off Oregon and Washington, for all years combined. Numbers in paren- | theses refer to summaries for major taxonomic groupings. Prey taxa Coho salmon Chinook salmon F N W IRI F N W IRI Cnidaria Velelta vetella 0.3 <0.1 <0.1 <0.1 — — — — Ctenopliora Unidentified <0.1 <0.1 <0.1 <0.1 0.5 <0.1 <0.1 <0.1 Siphcinophora L'nidentified 1.0 <0.1 <0.1 <0.1 0.5 <0.1 <0.1 <0.1 Annelida Tom opteris septen t rion a I in <0.1 <0.1 <0.1 <0.1 — — — — Tomopferis sp. 0.6 0.1 <0.1 <0.1 — — — — Pi'bigobia sp. 0.1 <0.1 <0.1 <0.1 — — — — Nereidae — — — — 0.1 <0.1 0.1 <0.1 Unidentified 0.5 <0.1 <0.1 <0.1 0.1 0.2 <0.1 <0.1 Mollusca Gastropoda (18.4) (13.4) (5.9) (4.1) (1.3) (0.3) Limacina helicina 16.5 13.0 5.5 10.4 4.0 1.3 0.3 0.3 Euclio pyramidata 1.5 0.3 0.5 <0.1 — — — — Clio limacina 0.1 0.1 <0.1 <0.1 — — — — Pteropoda unidentified 0.4 <0.1 <0.1 <0.1 0.2 <0.1 <0.1 <0.1 Cephalopoda (2.0) (0.1) (0.9) (2.0) (0.4) (1.5) Loligo opalescens 1.0 0.1 0.8 <0.1 0.7 0.4 1.1 <0.1 Abraliopsis sp. 0.1 <0.1 <0.1 <0.1 — — — — Gonatidae 0.1 <0.1 <0.1 <0.1 0.1 <0.1 <0.1 <0.1 Octopiift dojleini — — — — 0.2 <0.1 0.1 <0.1 Unidentified 0.8 <0.1 0.1 <0.1 0.9 <0.1 0.2 <0.1 Arthropoda ('(i|)epoda (9.6) (2.1) (0.2) (8.5) (2.6) (0.2) (itiiirfnia princeps — — — — 0.1 <0.1 <0.1 <0.1 Ncocalanus cristatus 3.2 0.5 0.1 <0.1 3.8 0.7 0.1 0.1 Neocalanus plumchrus 0.5 <0.1 <0.1 <0.1 0.2 <0.1 <0.1 <0.1 Calanus marshallae 0.5 <0.1 <0.1 <0.1 0.2 0.1 <0.1 <0.1 Calamis pacificus 1.7 0.5 <0.1 <0.1 0.4 1.2 <0.1 0.1 Calanus spp. copepodites 0.6 0.1 <0.1 <0.1 0.1 <0.1 <0.1 <0.1 Eucalanus bungii 0.3 0.6 <0.1 <0.1 0.9 0.2 <0.1 <0.1 Metndia pacifica 0.2 <0.1 <0.1 , --), and CWT juvenile fish caught in the ocean (A, -•■-). Returning jacks Fork length was measured and scale samples taken from the preferred area of 64 and 99 coded-wire tagged (CWT) jacks (precocious males) within 2 or 3 days of their return in 1983 and 1985, respectively, to the Anadromous Inc. facility on Coos Bay. Scales had also been taken earlier from a subsample of each release group shortly before their release as smolts. Fork length at time of ocean entrance was backcalculated for each returning jack using the relationship between scale radius and fork length at the time of release for that group and other groups released in the same month. (Tag groups were grouped by month of release to obtain adequate numbers for the prerelease scale radius-fish length relationships). Ocean entrance was detected on the scale as an abrupt change in circulus spacing (see Fisher and Pearcy, 1988). Since these fish were released and returned to a site only 8 km from the ocean, their period of growth in the ocean should be very similar to the time between their release and return. Growth rate of each fish while in the ocean was estimated by (FL^ - FLi)/d, where FL-j is length on return to the hatchery, FL; is the backcalculated length at time of ocean entry, and d is the days between release and return. Scale growth rate was estimated as (SRt„t - SRi)/d, where SRI is the scale radius to the ocean entrance mark, and SR,,,, is the total scale radius. Mean circulus spacing during the ocean growth period was calculated as the distance between the first and last ocean circulus divided by « - 1, where n = the number of circuli laid down during ocean growth. Fisher and Pearcy Scale circulus spacing in Oncorhynchus kisutch 639 X oo o a: o o 50- o 40- 30- 20- o o o 10- 0- U¥ o 1 — i 1 1 — — 1 1 — — 1 — 10 20 30 40 50 60 TOTAL FISH GROWTH (mm) 70 80 Table 2 Rate of circulus formation (RCF, circuli per day) vs. fish growth rate (GR, mm/d): Correlation coefficients (r), prob- ability that correlation = 0.0 (jt) ), and geometric mean regres- sion (GM) for age-0 fish held in saltwater tanks (A), CWT jacks (age 1.0) returning in 1983 (B) and 1985 (C), and CWT juvenile fish (both age 0.0 and 1.0) caught in the ocean (D). Group GM regression A 80 0.89 <0.01 RCF = 0.18 -(GR) -I- 0.03 B 64 0.58 <0.01 RCF = 0.10 (GR) -I- 0.03 C 99 0.57 <0.01 RCF = 0.10 (GR) -I- 0.00 D .34 0.84 <0.01 RCF = 0.12 (GR) -I- 0.02 Figure 2 Scale growth vs. fish growth scattergram and GM regression line for individually marked age-0 smolts held in saltwater tanks. Juvenile coho collected in the ocean Growth rates in the ocean were estimated for CWT or spray-marked juvenile coho released (a) very near the ocean in Yaquina and Coos Bays and caught 60 or more days later in the ocean in August or September (15 fish), and (b) in the Columbia River (19 fish) and sampled during downstream migration near the ocean (at rkm 75, Dawley et al. 1985) and caught in the ocean 60 days or more after the median fish passed rkm 75, or released below rkm 75 and caught at least 60 days later in the ocean. For fish released in Yaquina or Coos Bays growth rate was estimated by (FL^. - FLi)/d where FLj is the length at ocean entrance backcalcu- lated from scales using a regression of FL on scale radius from a sample of fish taken at the time of release, FL? is length at capture in the ocean, and d is days between release and recapture. For Columbia River fish released in 1981, 1982, 1983, and 1984 length at ocean entrance was backcalculated using a regres- sion of FL on scale radius derived from fish collected at rkm 75 in 1982 and 1983, combined. The number of days in the ocean was estimated from the date the median fish in each group passed rkm 75 and the mean downstream migration rates for each group. Estimated date of ocean entry for each group ranged from 3 to 11 days following passage of the median fish at rkm 75. For CWT Colimibia River fish caught in 1984, when no sampling occurred at rkm 75, average downstream migration rates for tag groups from the same hatch- eries in 1981-1983 were used in estimating ocean en- trance date and days in the ocean (d). Scale growth rate for juvenile coho caught in the ocean was estimated the same way as for jacks. Results Fork length was positvely and significantly correlated with scale radius (SR) in all groups (Table 1, Fig. 1). The relationship (geometric mean regression, Ricker 1973) for each group of fish appeared linear. However, SR was smaller relative to FL for jacks returning in 1985 (diamonds in Figure 1) than for jacks returning in 1983 (squares) or for CWT juvenile coho salmon col- lected in the ocean (solid triangles). In addition, almost all FL-SR data points for returning yearling jacks and juveniles caught in the ocean were above the extrap- olated regression line for the smaller subyearling fish held in saltwater tanks (Fig. 1). Thus, the relationships between FL and SR varied between age or size groups (small subyearling fish vs. larger yearling jacks and juveniles caught in the ocean) and between years (1983 vs. 1985 Anadromous Inc. jacks). The relationship be- tween scale growth and fish growth appeared linear for the subyearling coho smolts held in saltwater tanks (Fig. 2). Rate of circulus formation was positively and signif- icantly (r = 0.57 - 0.89, jD<0.01) correlated with fish growth rate for all groups (Table 2, Fig. 3). However the slope of the relationship was lower for ocean-caught juveniles and jacks returning in 1983 and 1985 (slopes of GM regression = 0.12, 0.10, 0.10, respectively) than for the subyearling fish held in saltwater tanks (slope = 0.17) At similar fish growth rates, rates of circulus formation were generally higher for the subyearling fish held in saltwater tanks than for juveniles caught in the ocean or for returning jacks (Fig. 3). The spacing between circuli was positively and significantly correlated with fish growth rate for all groups but jacks returning in 1983 (Table 3). The cor- relation was strong (r = 0.80) for age 0.0 fish held in saltwater tanks, but much weaker for jacks returning in 1985 (r = 0.24). However, the relationship between circulus spacing and fish growth rate for subyearling 640 Fisheiy Bulletin 88(4), 1990 0.5 1,0 1.5 2.0 2.5 GROWTH RATE (mm/day) 3.0 Figure 3 Rate of circulus formation vs. fish growth rate scattergrams and GM regression lines for age-0 smolts held in saltwater tanks (O, — ), for CWT jacks returning to Coos Bay in 1983 (D. ) and 1985 (C>, --), and for CWT juvenile fish caught in the ocean (A, ). fish held in saltwater tanks appeared to be nonlinear (see footnote. Table 3). In general, circulus spacing vs. growth rate values in the three groups that had reared in the ocean were fairly close to the extrapolated linear regression line for the smaller fish held in saltwater tanks, although the variability about this regression line was very large (Fig. 4). Mean growth rates of jacks returning in 1985 and juveniles caught in the ocean (1.49 mm/d and 1.53 mm/d, respectively) were signif- icantly greater (i -tests, jo<0.01) than the mean growth rate of jacks returning in 1983 (1.17 mm/d). Mean cir- culus spacings of the two faster growing groups (3.86 and 4.04 mm at 88 x for jacks returning in 1985 and for juveniles caught in the ocean, respectively) were also both greater (;)<0.01) than mean circulus spacing of the slower-growing jacks returning in 1983 (3.61). In each group of fish some of the large variability in the relationship between circulus spacing or circulus deposition rate and fish growth rate was caused by variability in the relationship between fish length and scale radius. To remove this component of variation, we compared circulus spacing and circulus deposition rates with scale growth rates (Tables 4 and 5, respec- tively). Correlation coefficients for the relationships of circulus spacing and deposition rate with scale growth rate were considerably higher than for the correspond- ing relationships with fish growth rate, especially for jacks returning to Coos Bay in 1983 and 1985. (Com- pare Tables 2 with 4, and 3 with 5.) Circulus spacing was positively and significantly correlated with scale growth rate in all groups, whereas the relationship between circulus spacing and fish growth rate was Table 3 Circulus spacing (CSP, mm at 88 x ) vs. fish growth rate (GR, mm/d): Correlation coefficients (r). probability that correla- tion coefficient = 0.0 (/)) and geometric mean regression (GM) forage-0 fish held in saltwater tanks (A). CWT jacks (age 1.0) returning in 1983 (B) and 1985 (C), and CWT juvenile fish (both ages 0.0 and 1.0) caught in the ocean (D). Group n ?■ P GM regression A* 80 0.80 <0.01 CSP = 1.80 (GR) + 1.31 B 64 0.08 N.S. - C 99 0.24 <0.05 CSP = 2.06 (GR) + 0.71 I) 34 0.52 <0.01 CSP = 1.21 (GR) + 2.24 ' For this group of fish the correlation cuefficieiit for ,-i tliinl- order relationship, CSP = 0.91 + 5.35(GR) - 6.44(GR)- + 3.04 (GR)l r = 0.83, was higher than for the linear relationship. 5.0 4.0 o 3.0-- Q- (/I _l 3 U o 2.0 * /^ * D A -^ . •!* Vj>>- ■ oo °*oai8T

-- ° % o-o ° ^ (j> ■ o o --«-o 1 1 1 1 1 0.0 0.5 1.0 1.5 2.0 GROWTH RATE (mm/day) 2.5 Figure 4 Circulus spacing vs. growth rate scattergrams and GM regression lines for age-0 smolts held in saltwater tanks (O, — ), for CWT jacks returning to Coos Bay in 1983 (D, regression n.s.) and 1985 (0, --) and for CWT fish caught in the ocean (▲, -•■-). significant for all groups but jacks returning in 1983. Thus, the correlations of circulus spacing or rate of cir- culus formation with scale growth rate were stronger than the correlations with the underlying fish growth rate. Discussion A positive correlation between rate of circulus forma- tion and fish growth rate appears to be a common feature among fishes. We found in young coho salmon that rate of circulus deposition was positively and Fisher and Pearcy Scale circulus spacing in Oncorhynchus kisutch 641 Table 4 Rate of circulus formation (RCF. circuli/d) vs. scale growth rate (SGR, nim/d at 88 x ): Correlation coefficients (r). prob- ability that correlation coefficient = 0.0 (p) and geometric mean regression (GM) for age-0 fish held in saltwater tanks (A),CWT jacks (age 1.0) returning in 1983 (B) and 1985 (C). and CWT juvenile fish (both age 0.0 and 1.0) caught in the ocean (D). Group r GM regression A 80 B 64 C 99 D 34 0.94 <0.01 RCF = 0.30(SGR) + 0.03 0.84 <0.01 RCF = 0.22(SGR) + 0.03 0.83 <0.01 RCF = 0.20(SGR) + 0.03 0.90 <0.01 RCF = 0.22(SGR) + 0.02 significantly correlated with scale and fish growth rates (Tables 2, 5; Fig. 3). Thus, the faster the fish or scales grew, the more circuli were formed per unit time. Positive correlations between rate of circulus forma- tion and growth rate also have been found for walleye (Glenn and Mathias 1985) and cichlids (Sire 1986). In juvenile walleye, rate of circulus formation ranged from 1.5 circuli/d at high growth rates to 1 circulus every 2-3 weeks at low growth rates. Data presented by Bilton and Robins (1971a) indicated that juvenile sock- eye salmon receiving more food, and presumably grow- ing faster, produced more circuli during a given period than those fish that were fed less. Bilton and Robins (1971b) also found that sockeye salmon formed no cir- culi during periods of starvation. In chum salmon from Olsen Creek, Alaska, the number of circuli and radius to the middle of the first ocean annulus were positive- ly correlated (Helle 1980). Therefore, if time to the middle of the first ocean annulus was constant (which may or may not be true), then fast-growing fish (larger radius) produced more circuli per unit time than slow- growing fish. Positive correlations between circulus spacing and growth rate also have been reported in other species of fish. Bhatia (1932) found that scales from juvenile rainbow trout fed abundantly and growing rapidly, and scales from those fed sparsely and growing slowly had zones of widely spaced and narrowly spaced circuli, respectively, near the scale margin. Bhatia also was able to produce zones of widely and narrowly spaced circuli by alternately changing feeding level. Doyle et al. (1987) and Matricia et al. (1989) found positive correlations between circulus spacing and fish growth rate in tilapia. Sire (1986) found more widely spaced circuli among faster growing than slower growing cichlids. In juvenile walleye, mean spacing of circuli formed during the period of most rapid growth was found to be greater than mean spacing of circuli formed during periods of slower growth (Glenn and Mathias Table 5 Circulus spacing (CSP, mm at 88 x) vs. scale growth rate (SGR. mm/d at 88 x): Correlation coefficients (r), probabil- ity of correlation coefficient = 0.0 (p) and geometric mean regression (GM) for age-0 fish held in saltwater tanks (A), CWT jacks (age 1.0) returning in 1983 (B) and 1985 (C), and CWT juvenile fish (both age 0.0 and 1.0) caught in the ocean (D). Group n r P GM regression A* 80 0.82 <0.01 CSP = 3.07(SGR)-^ 1.29 B 64 0.62 <0.01 CSP = 4.22(SGR)-f 1.39 C 99 0.59 <0.01 CSP = 4.14(SGR)-i- l.,56 D 34 0.56 <0.01 CSP = 2. 26 (SGR) -1- 2.22 * For this group of fish the correlation coefficient for a third order relationship, CSP = 1.02 -i- 6.92(SGR) - 10.37 (SGR)- -f 6.77(SGRy', r = 0.84, was higher than for the linear relationship. 1985). Bilton and Robins (1971a) found a significant positive correlation between feeding level and spacing of circuli in sockeye salmon. Rate of circulus formation and spacing of circuli are probably related to a number of other factors beside growth rate. We found that rate of circulus formation for juvenile coho salmon caught in the ocean and espe- cially for returning jacks were all well below the rates predicted by the regression for the much smaller fish held in saltwater tanks but growing at similar rates (Fig. 3). This suggests that the rate of circulus forma- tion varies with age or size of fish or with environmen- tal conditions, as well as with growth rate. Doyle et al. (1987) found that circuli of tilapia were laid down less frequently as fish grew larger. They also found that the relationship between growrth rate and circulus spac- ing was stronger when a correction was made for the size of fish. Bilton (1975) suggested that rate of circulus deposition was probably a function of a combination of factors such as temperature, food, light, and mater- nal and inherent characteristics. Several studies have addressed the possible effects of water temperature on circulus spacing. Generally they suggest that the effect of temperature, by itself, on circulus spacing is relatively small compared with the effect of feeding level or growth rate. By manip- ulating feeding level, Bhatia (1932) was able to produce zones of widely and narrowly spaced circuli in scales of rainbow trout growing in extremely different water temperatures (4°C and 17°C). Kimura and Sakagawa (1972), working with sardines, found that formation of annuli or checks (bands of narrowly spaced circuli) did not appear to be related to temperature. Barber and Walker (1988) found that in sockeye salmon annulus formation occurred before the coldest months of the 642 Fishery Bulletin 88(4), 1990 year and that during the coldest months widely spaced circuli were formed. Hogman (1968) cited work by Deason and Hile (1947) that indicated that annuii and checks formed in scales of kiyis living at depth in Lake Michigan, despite a very small annual variation in water temperature (2-3°C). Bhatia (1932) cited his earlier work showing that when rainbow trout were fed uniformly throughout the year no periodic zones were formed on their scales and all rings were of nearly the same width despite fluctuations in temperature. How- ever, there is some evidence that in brown trout water temperature during egg and alevin stages prior to scale formation affected subsequent rate of circulus forma- tion (Skurdal and Anderson 1985). Inherent differences between groups of fish may mask the relationship between circulus spacing and growth rate. For example, for coho jacks returning to the Anadromous Inc. facility in Coos Bay in 1985, scales were generally smaller at a given fish length (Fig. 1), and rate of circulus formation was lower at a given growth rate (Fig. 2), than was the case for CWT juveniles caught in the ocean. This resulted in signif- icantly different mean spacing of circuli in these two groups (3.86 vs. 4.04, ^test, p<0.05) despite very similar mean growth rates (1.53 and 1.49 mm/d, ^test, n.s.). Because of differences in the relationships be- tween circulus spacing and growth rate among dif- ferent groups of juvenile coho salmon, inferences about relative growth rates based on scale circulus spacing are probably only valid when made between groups that are similar in age, size, and morphometric char- acteristics (i.e., very similar SR-FL relationships). Data from the group of subyearling fish held in salt- water tanks suggest that the relationship between cir- culus spacing and growth rate may be complicated. A third-order relationship (see footnote. Table 3) gave a better fit to the data than did a simple linear relation- ship. In this group of fish there was little change in cir- culus spacing between growth rates of 0.4 and 0.9 mm/d. More rapid changes in circulus spacing with growth rate occurred both above and below this range (circles, Fig. 4). This result suggests that for coho salmon there may be ranges of growth rates within which circulus spacing is a poor indicator of relative growth rate. We compared mean circulus spacing in the ocean growth zone of scales for unmarked yearling (age 1.0) juvenile coho caught in the ocean in late summer of 1981, 1982, 1983, and 1984 (Fisher and Pearcy 1988). These fish had very similar SR-FL relationships in all years and were of similar size at time of entry into the ocean (based on backcalculated size at ocean entry). We also estimated gi-owth rates of these unmarked juvenile yearling coho salmon between early and late summer (between .June and August or September) from changes Table 6 Rank 01 der of mean scale circulus spacing for unmarked 1 juven le coho collected in the ocean during the late summer, 1!I81- 84 and growth rates of unniai ked juvenile c oho esti- mated from changes in mean length between early and late summer (see Fisher and Pearcy 1988 , their Tables 3 and 4). Mean Est. growth spacmg rate Year (mm at 88 x ) Rank (mm/d) Rank litSl 4.16 1 1.56 2 1982 4.15 2 1.7G 1 1983 3.91 4 1.37 3 1984 3.9.T 3 1.33 t in mean lengths with time (see Fisher and Pearcy 1988, their Table 3). Rank order of mean spacing between circuli and growth rates estimated from shifts in mean FL with time are compared in Table 6. Although the rank orders do not agree in detail, they both suggest higher fish growth rates during the summers of 1981 and 1982 than in 1983 and 1984. Acknowledgments We thank Ron Gowan of Anadromous, Inc. for his cooperation and for supplying data and scale samples from returning jacks. We also thank Ric Brodeur, Alton Chung, Harriet Lorz, Jon Shenker, Jamie Trautman, and Waldo Wakefield for their help with various as- pects of this project. This research was supported by the Northwest and Alaska Fisheries Center (NA-88- ABH-00043 and NA-85-ABH-00014) and the Oregon State University Sea Grant College Program (NA 881AAD-D-00086, R/OFP-17). Citations Barber, W.E.. and R.J. Walker 1988 Circulus spacing and annulus formation: Is there more than meets the eye? The case for sockeye salmon, OhcoWij/w- rhiis tifrka. .J. Fish Biol. 3:237-245. Bhatia, D. 1932 Factors involved in the production of annual zones on the scales of the rainbow trout {Saliiio irideus). II. J. Exp. Biol. 9:6-11. Bilton, H.T. 1975 Factors influencing the formation of scale characters. Int. North Pac. Fish Comm. Bull. 32:102-108. Bilton. H.T., and G.L. Robins 1971a Effects of feeding level on circulus formation on scales of young sockeye salmon {Onmrhynchiis twrka). J. Fish. Res. Board Can. 28:861-868. Fisher and Pearcy Scale circulus spacing in Oncorhynchus kisutch 643 1971b Effects of starvation, feeding, and light period on cir- culus formation on scales of young sockeye salmon (Oncorhyn- rhus ncrkn). J. Fisli. Res. Board Can. 28:1749-1755. Clutter, R.I., and L.E. Whitesel 1956 Collection and interpretation of sockeye salmon scales. Int. Pac. Salmon Fish. Comm. Bull. 9, 159 p. Dawley, E.M., R.D. Ledgerwood, and A.L. Jensen 1985 Beach and purse seine sampling of juvenile salnionids in the Columbia River estuary and ocean plume, 1977-1983. Vol. II. Data on marked fish recoveries. NOAA Tech. Memo. NMFS F/NWC-75, Northwest & Alaska Fish. Cent., Natl. Mar. Fish. Serv., 2725 Montlake Blvd. E., Seattle, WA 98112, 397 p. Deason H.J., and R. Hile 1947 Age and growth of the kiyi, Leucichthys kiyi Koelz, in Lake Michigan. Trans. Am. Fish. Soc. 74(i944):82-142. Doyle, R.W., A.J. Talbot, and R.R. Nicholas 1987 Statistical interrelation of length, growth, and scale cir- culus spacing: Appraisal of a growth rate estimator for fish. Can. J. Fish. Aquat. Sci. 44:1.520-1528. Fisher, J. P., and W.G. Pearcy 1988 Growth of juvenile coho salmon (Oncorhynchus kisutch) in the ocean off Oregon and Washington, USA, in years of differing coastal upwelling. Can. .J. Fish. Aquat. Sci. 45: 1036-1044. Glenn, C.L., and J. A. Mathias 1985 Circuli development on body scales of young pond-reared walleye (Stizostedion vitreum). Can. J. Zool. 63:912-915. Godfrey, H.. K.A. Henry, and S. Machidori 1975 Distribution and aliundance nf cciho salmon in offshore waters of the North Pacific Ocean. Int. North Pac. Fish. Comm. Bull. 31, 80 p. Helle, J.H. 1980 Influence of marine environment on age and size at maturity, growth, and abundance of chum salmon, Oncorhyn- rliiis ketii (Walbaum), from Olsen Creek, Prince William Sound, Alaska. Ph.D. thesis, Oregon State Univ., Corvallis, 118 p, Hogman, W.J. 1968 Annulus formation on scales of four species of coregonids reared under artificial conditions. .J. Fish. Res. Board Can. 25:2111-2122. Kimura. M., and G.T. Sakagawa 1972 t)bser\ations on scale patterns and growtli of the Pacific sardine reared in the laboratory. Fish. Bull., U.S. 70: 1043-1052. Koo, T.S.Y. 1962 Age designation in salmon. /)i Koo, T.S.Y. (ed.). Studies of red salmon, p. 41-48. Univ. Wash. Press, Seattle. Lotrich, V.A., and W.H. Meredith 1974 A technitjue and the effectiveness of various acrylic colors for subcutaneous marking of fish. Trans. Am. Fish. Soc. 103:140-142. Matricia, T., A.J. Talbot, and R.W. Doyle 1989 Instantaneous growth rate of tilapia genotypes in un- disturbed aquaculuture systems 1. Red and gray morphs in Indonesia. Aquaculture 77:295-306. Ricker, W.E. 1973 Linear regressions in fishery research. J. Fish. Res. Board Can. 30:409-434. Skurdal, J,, and R. Anderson 1985 Influence "f temperature on number of circuli of 1st year scales of l>rown trout Snlnw-tnittn. J. Fish Biol. 26:363-366. Sire, J.Y. 1986 Ontogenic development of surface ornamentation in scales o{ Hcmirhromif: himnnilitlux (Cichlidae). .1. Fish Biol. 28:713-724. Abstract.- Optimal harvesting policies were developed for a model of the Washington-Oregon-Califor- nia trawl fisheries for bocaccio Sebas- tes paiwispinus, chilipepper S. goodei, widow rockfish S. entomelas, split- nose rockfish S. diploproa, and short- belly rockfish S. Jordan i. In the simu- lated management system, the annual quota for each species was based on an intended fishing mortality rate (F) and an estimate of stock biomass. Traditional constant F policies were compared with variable F policies in which F was a linear function of either stock biomass (single-species) or the combined biomass of all other species (multispecies). Variable F policies were ineffective for increas- ing total harvest of all species com- bined, when compared with constant F policies. However, both single- species and multispecies variable F policies were found to reduce the variance for total harvest substan- tially with no significant loss in yield. The reduction in variance was rough- ly equal to the reduction that would be achieved by lowering the coeffi- cient of variation (CV) for biomass estimates from .50% to 25% or from 25% to 0%. The reduction in vari- ance generally was greater for the multispecies than for the single- species variable F policy. Compared with the case where CV = 0%, aver- age total yield was reduced slightly for both constant F and variable F policies when CVs for biomass esti- mates were greater than 0%. Multispecies Harvesting Policies for Washington-Oregon-California Rockfish Trawl Fisheries Joseph E. Hightower Tiburon Laboratory, Southwest Fisheries Science Center National Marine Fisheries Service, NOAA 3150 Paradise Drive. Tiburon, Calfiornia 94920 Manuscript accepted 4 .June i;i!K). Fishery Bulletin, U.S. 88:64.5-6.56. In the Washington-Oregon-Califoi'iiia (WOC) trawl fisheries for rockfish {Sebastes spp.), the landings are com- prised of many species, including bocaccio S. paucispinus, chilipepper S. goodei, Pacific ocean perch S. alu- tus, canary rockfish S. pinninger, widow rockfish S. entomelas, and yel- lowtail rockfish S. flavidus (PFMC 1982). There is httle evidence of sub- stantial biological interaction among the adults of these species. Techno- logical interactions (Pope 1979) are significant in the shelf rockfish fish- ery (Pikitch 1987), which accounts for most of the landings of bocaccio, chilipepper, and canary and yellow- tail rockfish. Fishermen have some control over the species composition of the catch in this fishery by vary- ing the areas or depths fished. The other significant rockfish fishery is the midwater trawl fishery for widow rockfish. Catches of other rockfish species are negligible in that fishery. Most currently marketed Seh 22 18.10 33.67 1.04 0.84 4.70 0.34 31.20 40.20 multispp. maxh 3.01 2.86 7.84 •-) 00 20.01 35.59 1.71 1.27 7.07 0.19 107.71 115.41 logh 2.91 2.56 7.35 1 39 19.60 33.66 1.82 1.51 8.72 1.13 110..50 66.74 negx 2 22 2.31 7.08 1 14 19.47 32.09 2.79 1.77 9.08 1.48 95.97 40.77 obtained. In some cases, the reduction was comparable to the reduction that would be achieved by lowering the CVs for biomass estimates from 50% to 25% or from 25% to 0%. The percent reduction typically in- creased as the CVs for biomassestimates increased, although the loss in mean yield also increased some- what (Hightower 1990). Single-species vs. multispecies policies When single-species (5) and multispecies (6) policies were compared, the variance for total catch generally was lower for the multispecies policy, except for some cases for the five- species model or when biomass CVs were 50%. In those cases, the better performance of the single-species policy probably was due to the feed- back mechanism it contained. For example, in the five- species models, the single-species policies for shortbelly and splitnose I'ockfish could adapt to the large changes in the size of those stocks. When the five-species model was rerun using five "stocks" having chilipepper life- history characteristics, the variance for total catch was lower for the multispecies policy. For the single-species policy, the variance of total catch typically was reduced by stabilizing the catch of each individual species (Tables 3-5). For the multi- species policy, reductions in the variance for total catch were sometimes achieved by increasing the variance for some of the less abundant individual species (Tables 3-5). The reduction in variance obtained using policy (6) is consistent with Clark's (1984) prediction that a multispecies system could be managed to reduce the variance of total yield, relative to the variance of the individual yields. For the multispecies policies, the optimal parameter estimates were relatively insensitive to the objective function and to the CV for biomass estimates (High- tower 1990). The slope parameter bj i was less than zero for all cases; consequently, fishing pressure varied inversely with the alnmiiance of the other species. 652 Fishery Bulletin 88(4), 1990 Table 5 Comparison of average rockfish catches and year- 100 variances fo • constant F. single-species varialile F. an( multi species harvesting | policies. CVs for biomass estimates were assumed to be 50%. Harvesting Objective Mean jdeld Var[100] chili- split- short- chili- split- short- Species policy function bocaccio pepper widow nose belly total bocaccio pepper widow nose belly total Two const F ma.xh 3.00 2.94 — — 5.93 3.40 3.63 7.39 logh 3.00 2,93 — — — 5.93 3.38 3.59 — — — 7,33 neg.x 2.99 2.93 - - - 5.92 3.38 3.58 - - - 7.31 variable F maxh 3.02 2.97 — — — 5.96 8.49 12.36 — — — 21.14 logh 2.97 2.91 — — — 5.85 1.84 1.83 — — — 4.06 negx 2.95 2.90 - - - 5.83 1.68 1.69 - - - 3.77 multispp. maxh 3.00 2.94 — — — 5.93 3.40 3.65 — — — 7.48 logh 2.9fi 2.91 — — — 5.87 4.04 3.76 — — — 4.24 negx 2.96 2.90 - - - 5.86 4.06 3.91 - - - 4.09 Three const F maxh 3.01 2.92 7.72 — — 13.65 3.35 4.42 19,06 _ _ 24.74 logh 3.01 2.92 7.72 — — 13.64 3.35 4.37 19,03 — — 24.70 negx 3.01 2.91 7.69 - - 13.61 3.34 4.30 18,90 — - 24.53 variable F maxh 3.02 2.94 7.78 — — 13.70 11.44 6.07 33,98 — — 48.11 logh 3.00 2.92 7.66 — — 13.53 2.17 2.64 10,35 — — 14.49 negx 2.97 2.88 7.58 - - 13.37 1.66 2.02 9,47 - - 12.76 nniltispp. maxh 3.01 2 92 7.72 — — 13.65 3.43 4.42 18,91 — — 24.88 logh 2.97 2.89 7.69 — — 13.55 4.58 5.09 20,90 — — 15.79 negx 2.91 2.84 7.68 - - 13.43 6.01 5.34 20.42 - - 13.29 Five const F maxh 3.02 2.84 7.80 1.96 19.84 35.32 3.15 3.87 15.26 0.47 192.76 219.04 logh 3.02 2.84 7.78 1.96 19.73 35.18 3.14 3.85 15.21 0.47 188.46 214.56 negx 3.02 2.82 7.72 1.95 19.38 34.75 3.16 3.80 14.96 0.47 177.72 203.35 variable F maxh 3.03 2.8(i 7.81 2.18 20.43 36.18 5.26 10.09 120.32 4.15 1356.71 1540.97 l"gh 3.02 2.82 7.78 1.43 18.64 33.56 2.90 1.52 11.61 0.54 64.10 84.87 negx 2.9(i 2.78 7.51 1.52 17.32 31.96 1.53 1,26 7.28 0.42 43.02 55.69 multispp. maxh 3.00 2.84 7.79 1.97 19.91 35.38 3.34 3,88 15.26 0.47 191.10 209.32 logh 2.92 2.«4 7.44 1.57 19.58 34.01 3.58 3,45 18.95 1.68 208.31 135.32 negx 2.08 2.2S 7.05 1.37 19.20 31.84 7.13 4,10 20.26 2.46 173.92 81.64 Particularly in the three- and five-species cases, it appeared that the optimal multispecies strategy was to maintain stable catches for the species accounting for the majority of total yield, such as shortbeily or widow rockfish. Catches of less-abundant species (e.g., bocaccio or splitnose rockfish) varied depending on the abundance of the dominant species. For example, in the five-species case, the annual catch of splitnose rockfish was 0 in 24-30 of 100 years for the logh policies. Errors in biomass estimates Constant F policies were essentially identical when different CVs for biomass estimates were used. This indicates that the current management approach for Pacific Coast groundfish is robust to random, non- autocorrelated errors in estimating biomass. One ob- vious impact in the simulated fishery of introducing errors in estimating biomass was that the variance for total catch increased substantially (Tables 3-.5, P"'ig. 4). This increased variability was due to differences be- tween the intended (F) and actual (F') fishing mortal- ity rate. Such errors are nonlinearly related to esti- mated stock size and can be substantial (Rivard 1981 ). Rivard (1981) argued for higher precision in estimates of stock size in order to reduce the biological risk of applying F'»F or the economic loss if F'«F. In this study, the variation in F due to errors in estimating biomass was much greater than that caused by using the variable F policies. For example, when CV = 0%, Fs for chilipepper in the five-species case ranged from 0.00 to 0.24 when using the multispecies logh model. When CV = 25%, the actual Fs ranged from 0.06 to 0.61 when the intended policy was a constant F of 0.22. Actual Fs ranged from 0.00 to 0.63 when CV = 25% and a multispecies logh policy was used, which suggests that multispecies policies do not introduce substantial additional risk, relative to the risk introduced by esti- mating biomass with error. Hightower Rockfish harvesting policies in Washington-Oregon^California trawl fisheries 653 [piS^ ,.er^ o o O 1 five~spp- Vorionce (total 'lotct-i! Figure 4 Relationship between mean total rockfish catch and the variance of total catch for the two- (bocaccio, chilipepper), three- (bocaccio, chilipepper, widow rockfish), and five-species Osocaccio, chilipepper, widow, splitnose, and shortbelly rockfish) models. Open symbols represent Pareto Frontiers derived by simulating the fishery at 20, 40, 60. 80. and 100% of the constant fishing mortality rate (F) that maximized harvest. Shaded and closed symbols represent the catch:variance relationship for the variable F and multispecies policies, respectively (equations 5-6). Simulation runs were made assuming a 0. 25, or 50% coefficient of variation for annual estimates of stock size. Mean yield was only slightly lower at higher biomass CVs; the decrease was similar for constant F and variable F policies (Tables 3-5). Thus, in this model at least, there was no apparent disadvantage in using the more complex policies, even when biomass was esti- mated imprecisely. This was an interesting result because, although all policies used biomass estimates to calculate catch, the variable F policies also use the estimates to calculate Fs. When biomass CVs were 50%, the variance for total catch was in all cases lower for the single-species (5) than for the multispecies policy (6). Apparently it was important to readjust continuously for the changes in stock size due to previous errors in estimating biomass. Based on limited additional runs using the three-species model, it appears that further gains in stability of total Table 6 .\verage total rockfit h catches and year -100 variances for con- 1 stant F, va •iable F, multispecies, and three-pai ameter har- vesting policies for the three-species model. Biomass Objective CVs Policy function Yield Var[100] 0 (4) maxh 13.75 7.40 (5) logh 13.78 7.. 59 (6) logh 13.70 5.60 a) logh 13.77 6.39 (5) negx 13.61 5.75 (6) negx 13.60 4.08 (") negx 13.60 4.09 25 (4) maxh 13.72 11.58 (5) logh 13.67 9.01 (0) logh 13.67 7.97 (7) logh 13.63 6.99 (5) negx 13.53 7.47 (unc.. Portland, OR. 1986 Ahistory of California rockfish fisheries, /n Proc. Int. Rockfish Symp., Anchorage. Alaska, Oct. 1986, p. 35-41. Lowell Wakefield Fish. Symp. Ser. 5. Alaska Sea Grant, Univ. Alaska. Fairbanks. Mendelssohn. R. 1982 Discount factors and risk aversion in managing random fish populations. Can. .]. Fish. Aquat. Sci. 39:1252-1257. Methot. R.D. 1986 Synthetic estimates of historical abundance and mortality for northern anchovy, Engraulis mordax. Admin. Rep. L.J 86-29. Southwest Fish. Cent., Natl. Mar. Fish. Serv., NOAA, La .Jolla. CA 92038, 85 p. Methot. R.. and J. Hightower 1988 Status of the Washington-Oregon-California sablefish stock in 1988. In Status of the Pacific Coast gi-oundfish fishery through 1988 and recommended acceptable biological catches for 1989, Appendix B. Pac. Fish. Manage. Counc, Portland, OR. Murawski. S.A.. and J.S. Idoine 1989 Yield sustainability under constant-catch policy and stochastic recruitment. Trans. Am. Fish. Soc. 118:349-367. Pikitch. E.K. 1987 Use of a mixed-species yield-per-recruit model to explore the consequences of various management policies for the Oregon flatfish fishery. Can. .J. Fish. Aquat. Sci. 44(Suppl. 2):349-3.59. PFMC (Pacific Fishery Management Council) 1982 Final fishery management plan and supplemental en- vironmental impact statement for the Washington, Oregon, and California groundfish fishery. Pac. Fish. Manage. Counc, Portland. OR. 1986 Status of the Pacific Coast groundfish fishery through 1986 and recommended acceptable biological catches for 1987. Pac. Fish. Manage. Counc, Portland, OR. Pope, J. 1979 Stock assessment in multispecies fisheries, with special reference to the trawl fishery in the Gulf of Thailand. Rep. SCS/DEV/79/19, FAO South China Sea Fish. Dev. Coord. Prog., Manila. Philippines, 106 p. [ISBN 92-8.52-1011-2]. Pope, J.G. 1983 Analogies to the status quo TACs: Their nature and variance. In Doubleday, W.G., and D. Rivard (eds.), Sampling commercial catches of marine fish and invertebrates, p. 99-113. Can. Spec Publ. Fish. Aquat. Sci. 66. Pope, J.G., and D. Gray 1983 An investigation of the relationship between the preci- sion of assessment data and the precision of total allowable catches. In Doubleday, W.G.. and D. Rivard (eds.), Sampling commercial catches of marine fish and invertebrates, p. 151-157. Can. Spec Publ. Fish. Aquat. Sci. 66. Raiffa. H. 1968 Decision analysis. Introductory lectures on choices under uncertainty. Addison-Wesley. Reading. MA, 309 p. Rivard. D. 1981 Catch projections and their relation to sampling error of research surveys. In Doubleday, W.G., and D. Rivard (eds.), Bottom trawl surveys, p. 93-109. Can. Spec Publ. Fish. Aquat. Sci. 58. Ruppert. D.. R.L. Reish, R.B. Deriso. and R.J. Carroll 1984 Optimization using stochastic approximation and Monte Carlo simulation (with application to harvesting of Atlantic menhaden). Biometrics 40:535-545. 1985 A stochastic population model for managing the Atlan- tic menhaden (Brevoortia ti/ninnus) fishery and assessing managerial risks. Can. .J. Fish. Aquat. Sci. 42:1371-1379. 656 Fishery Bulletin 88(4). 1990 Schnute, J. 1985 A general theory for analysis of catch and effort data. Can. J. Fish. Aquat. Sci. 42:414-429. Sen, A.R. 1984 Samphng commercial rockfish landings in California. Tech. Memo. NOAA-TM-NMFS-SWFC-45, Tiburon Lab., Southwest Fish. Cent.. Natl. Mar. Fish. Serv., Tiburon, CA 94920, 95 p. Swartzman. G.L., W.M. Getz, R.C. Francis. R.T. Haar, and K. Rose 1983 A management analysis of the Pacific whiting (Merluc- cius produdus) fishery using an age-structured stochastic recruitment model. Can. J. Fish. Aquat. Sci. 40:524-539. Tagart, J.V. 1988 Status of the yellowtail rockfish stocks in the Inter- national North Pacific Fishery Commission Vancouver and Columbia areas. In Status of the Pacific Coast groundfish fishery through 1988 and recommended acceptable biological catches for 1989, Appendix D. Pac. Fish. Manage. Counc, Portland. OR. Walters, C.J. 1975 Optimal harvesting strategies for salmon in relation to environmental variability and uncertain production parameters. J. Fish. Res. Board Can. 32:1777-1784. Walters, C. 1986 Adaptive management of renewable resources. Macmil- lian, NY, 374 p. Weinberg, K.L., M.E. Wilkins, and T.A. Dark 1984 The 1983 Pacific west coast bottom trawl survey of groundfish resources: Estimates of distribution, abundance, age and length composition. Tech. Memo. NMFS-F/NWC-70, Alaska Fish. Sci. Cent, Natl. Mar. Fish. Serv., NOAA, Seattle. WA 98115-0070, 376 p. Abstract.- Otolith microchem- istry of anadromous and non-anad- dromous salmonids was investigated to determine if there were differ- ences among migratory and non- migratory individuals and if the habitat where vitellogenesis took place would affect the composition of the otolith primordia of the prog- eny. Electron microprobe transects across salmonid otoliths showed that there were large differences in oto- lith Sr/Ca ratios among adult anad- romous and non-anadromous indi- viduals, but there were no detectable differences in Na/Ca, KJCa, and S/Ca ratios. The hypothesis that Sr/Ca ratios in the primordia of the prog- eny of anadromous salmonids would be greater than those in the primor- dia of the progeny of non-anadro- mous individuals because of differ- ences in the composition of ova was tested and confirmed by the results of a controlled experiment. Also, the ova of anadromous Oncorhynchus mykiss were found to contain 5 times more Sr than their non-anadromous conspecifics. On the basis of these data, it was concluded that otolith nucleus Sr/Ca ratios can be used to distinguish the progeny of sympatric anadromous and non-anadromous salmonids. Use of Otolith Microchemistry to Distinguish the Progeny of Sympatric Anadromous and IMon-anadromous Salmonids John M. Kalish Department of Zoology, University of Tasmania GPO Box 252C, Hobart. Tasmania 7001, Australia Present address Fisfieries Researcfi Centre, Ministry of Agriculture and Fisheries P O Box 297, Wellington, New Zealand Manuscript accepted 11 June 1990. Fishery Bulletin, U.S. 88:6.57-666. The ability to differentiate between juvenile anadromous salmonids and their sympatric non-anadromous con- specifics is essential for management of these species. However, stock dis- crimination between sea-run and res- ident freshwater salmonids has been limited to the adults upon their re- turn from the sea to spawn. Further- tnore, for those species where anad- romy is a facultative, and not an obligate behavior, such as brown trout Salmo trutta, rainbow trout Oncorhyn- chus mykiss, Atlantic salmon Salmo salar, cutthroat trout Salmo clarkii, and Arctic char Salvelinus alpinus, it has been impossible to distinguish between the co-occurring forms on the basis of meristic or mor-phometric characters (Nordeng 1983, Jonsson 1985, Neilson et al. 1985). McKern et al. (1974) were able to distinguish be- tween winter and summer races of steelhead trout (anadromous rainbow trout) from rivers in British Colum- bia, Washington, and Oregon on the basis of sagittal otolith nuclear dimen- sions that they believed to be affected by both qualitative and quantative differences in yolk. Rybock et al. (1975) concluded that differences in the size of female resident non-anad- romous rainbow trout and steelhead trout resulted in differences in egg size and, subsequently, the size of the otolith nucleus in the progeny. How- ever, both Neilson et al. (1985) and Currens et al. (1988) found that mea- surements of otolith nuclear dimen- sions were of questionable value in distinguishing juvenile non-anadro- mous and anadromous rainbow trout. Development of ova in anadromous salmonids is virtually complete through vitellogenesis or yolk formation be- fore the fish enter freshwater. On the basis of this information, I hypothe- sized that egg composition would, in some way, reflect the chemical com- position of the seawater environment and that this would ultimately affect the composition of the otoliths of the progeny, particularly the otolith nuclei that are formed in the early stages of development and well be- fore yolk utilization is complete. Yolk is formed through the deposition of a phospholipoprotein-calcium com- plex yolk precursor, vitellogenin, in the developing oocyte (Mommsen and Walsh 1988). Solubility and transport of vitellogenin through the circula- tory system of the female and to the developing ovaries may be dependent on the presence of calcium which has been shown to increase markedly in female salmonids (Bailey 1957, Booke 1964, Elliot et al. 1979) and other fishes (Ogiu-i and Takada 1967, Wood- head 1968) during gonad develop- ment. Vitellogenin has a high affin- ity for calcium due to the significant negatively-charged phosphate compo- nent of the molecule (Hara and Hirai 1978, Hara et al. 1980). The calcium binds to the vitellogenin molecule and, in this complexed form, the vitel- logenin and calcium are deposited in 657 658 Fishery Bulletin 88(4), 1990 the oocyte. Clearly, yolk deposition is a conservative process in that the overall composition of the yolk is more dependent on the genetic programming of the female parent than the environment. However, in many calcium-binding proteins, such as vitellogenin, the calcium moiety of the complexed molecule can be substituted by strontium due to the similar structural features of Ca^ ^ and Sr+ + (Skoryna 1981). The rela- tive degree of this substitution would be largely depen- dent on the relative concentrations of calcium and strontium in the ambient environment. Typical Sr/Ca ratios in marine waters (salinity SS-Am) are 0.0087 (0.09 mM/kg Sr:10.3 mM/kg Ca) (Bruland 1983) and average 0.0019 (0.00068 mM/kg Sr:0.35 mM/kg Ca) in freshwater (Rosenthal et al. 1970). These differences would be expected to affect the Sr/Ca ratio of the yolk and, ultimately, the composition of the developing embryo and its otoliths. In a study of otolith and endolymph composition, Kalish (1989) showed that the quantity of strontium in- corporated into the otolith was directly related to the quantity of strontium present in the endolymph, and that anadromous brown trout collected in an estuary had higher levels of strontium in both the endolymph and otolith than non-anadromous brown trout. There- fore, it seems likely that differences in the elemental constituents of the yolk and embryos of anadromous and non-anadromous salmonids would result in differ- ences in otolith composition. This hypothesis is sup- ported by research that shows little or no exchange of calcium between prehatch salmoniti embryos and the environment (Hayes et al. 1946, Zeitoun et al. 1976). This would be expected since calcium present in the yolk of the developing embryo is probably present in a protein-bound form and is destined for the tissues of the fry. Craik and Harvey (1984) found that the cal- cium composition measured in whole rainbow trout eggs was indistinguishable from the calcium measured in protein precipitate obtained from the egg. This, of course, excludes calcium that would be present in the fluid of the perivitelline space following fertilization (Laale 1980, Alderdice 1988). Strontium would prob- ably behave similarly and there would be minimal loss of any seawater-derived protein-bound ions in the yolk to the freshwater environment where the development of salmonid embryos takes place. In this paper, I discuss variations in the elemental composition of the otoliths of non-anadromous and anadromous salmonids with emphasis on the composi- tion of the sagittal otolith nucleus. Several "life his- tory" transects (scans of elemental composition across an axis of an otolith made with a wavelength-fiispersive electron microprobe) are presented to indicate the variety of forms that these data may take in salmonids with differing life histories. I also present the results of an experiment designed to test the hypothesis that otolith primordia of the progeny of anadromous rain- bow trout contain higher levels of strontium than the otolith primordia of non-anadromous rainbow trout. These data are examined in view of their usefulness in distinguishing the progeny of sympatric non-anad- romous and anadromous salmonids and in investiga- tions of diadromous behavior. Methods Brown trout (non-anadromous Snimo fruffn) and sea trout (anadromous Salmo trutta) were collected by gill- net from the Derwent River and estuary, southeast Tasmania, Australia. Juvenile and adult rainbow trout were obtained from both wild and hatchery stock. Atlantic salmon were obtained from hatcheries. Oto- liths were removed from fresh fish, cleaned, and stored in glass vials. Details of otolith preparation and micro- probe analyses are described below. Ripe ova were obtained from ovulating freshwater rainbow trout and sea-farmed rainbow trout and frozen in plastic bags for later analyses of calcium and stron- tium. Whole rainbow trout eggs were used for the determinations. Groups of 10 eggs from 4 freshwater and 4 sea-farmed trout were used. Preparation of eggs was modified from the methods employed by Craik and Harvey (1984). All solutions were made up with Milli-Q water (Millipore Corp.). Groups of eggs were weighed wet and then oven-dried at 50 °C for 24 hours to esti- mate dry weight. The dried whole eggs were ashed at 500°C for 24 hours and then digested in 2.0 mL Aristar ultrapure hydrochloric acid at 100°C for 2 hours. Samples were separated into two e(iual aliquots for sep- arate calcium and strontium determinations. Sample digests for estimation of calcium were diluted in 10 mM lanthanum chloride to suppress interference due to phosphate binding. Calcium concentrations were deter- mined by flame atomic absorption spectrojjhotometry using an air-acetylene flame. Strontium concentrations in eggs were determined by the method of standard additions using gi'aphite fur- nace atomic absorption spectrophotometry on a Varian AA-1475 spectrophotometer equipped with a GTA-95 graphite furnace and an autosampler. Argon was used as the purging gas in the graphite furnace. Samples were diluted with a 0.25% solution of an ionic deter- gent, Triton X-100, and 20 /jL of diluted sample were injected by autosampler into a walled, pyrolytically coated graphite tube. Furnace conditions were: drying Reference to trade names does not in)()ly endorsoinenl by the National Marine Fisheries .Service, N()A.\. Kalish Otolith microchemistry to distinguish salmonid progeny 659 Electron microprobe analy as standards. Calculation dence level. Table 1 tical data for salmonid otolith analyses. Details of rf the counting precision is discussed in Goldstein counting times, counting precision, and materials used et al. (1981). Precision values are for the 95% confi- Spectrometer crystal Element Peak counting time (sec) Counting precision Standard material Source 1/PET Ca-K„ 10 ~28.000 counts collected 1.0% Calcite USNM* 136321 Jarosewich and Maclntyre (1983) 1/PET K-K„ 30 15% when K/Ca = 2 x IQ-'' Anorthite USNM 137041 Jarosewich et al. (1980) :i/PET S-K, 40 36% when S/Ca = 1 x IQ-" Troilite BMP** Museum Ramdohr and Goresy (1971) 3/TAP Na-K„ 20 8.1';^. when Na/Ca = 2 x IQ-- Anorthite Same as for K above 3/TAP Sr-L, 20 mithsonian Inst., nherra. Australia 9.9%i when Sr/Ca = 3 x lO'' Wash., DC Strontianite USNM 10065 Jarosewich and White (1987) *Natl. Mus. ** Bur. Miner Nat. Hist.. S . Resour.. Cr at 90°C for 60 seconds; ramp ashing from 90 to 700°C for 20 seconds; ashing at 2600°C for 1 second; and atomization, with no gas flow, at 2600°C for 3 seconds. To confirm the hypothesis regarding otolith nucleus composition, eggs were obtained from sea-farmed and freshwater broodstock rainbow trout which originated from a similar hatchery stock (Cressy, Tasmania). These two groups of hatchery fish derive from similar stocks of rainbow trout imported into Tasmania; be- cause of minimal gene flow, inbreeding is relatively great and the diversity of the gene pool is low. Both freshwater and sea-farmed adults had been maintained on jack mackerel Trachurus decUvif^ based pellets throughout the period of egg development. Sea-farmed broodstock were kept at the same hatchery as fresh- water broodstock for 3 weeks before stripping, and all eggs were fertilized with sperm from a male of fresh- water stock. The developing embryos frotn sea-farmed and freshwater broodstock wet'e maintained in sep- arate baskets in the same channel of a recirculating water system where temperature was maintained at 10°C. A natural photoperiod was maintained through- out the e.xperiment. A random sample of 20 freshwater progeny and 20 sea-farmed progeny was taken 20 days after hatching, and these fish were frozen for later removal of otoliths. Sagittae and lapilli were removed from fry, the adhering otolith capsule was removed, and otoliths were rinsed in deionized water. Otoliths were oven-dried at 40°C. The length of sagittae and lapilli was measured with an ocular micrometer. Otoliths from adults were embedded in Araldite D epo.xy resin (Ciba-Geigy) and polymerized in an oven at 40°C for 48 hours. Several transverse sections of approximately 200 /.im thickness were obtained from otoliths, including one section containing the pri- mordium, with a low-speed saw (Struers Accutom) equipped with a diamond cut-off wheel. Sections were affixed to glass slides with epoxy. After drying, sec- tions were ground with a graded series of carborun- dum paper (wet/dry paper 600-1200 grade) until the primordia were reached. The much smaller otoliths from fry were mounted, sulcus-side-up, in Crystalbond 509 (Aremco Products, Inc.), a heat-labile thermoplas- tic polymer (Neilson and Geen 1981), and ground in the sagittal plane to the level of the primordia with 1200- grade wet-dry paper. All otoliths were then gi-ound to the precise level of the [jrimordia on a lapping wheel with 0.25 t^m aluminium paste (Linde A) and then finished with 0.25 f^m diamond paste. Polished speci- mens were ultrasonically cleaned in reagent-grade mineral spirits followed by ultrasonic cleaning in de- ionized water and oven drying at 60°C. Samples were then coated in a high-vacuum evaporator with a 25-nm carbon layer and stored in a dessicator. Otolith elemental analyses were carried out with a Cameca SX-50 wavelength-dispersive electron micro- probe. Analyses were made with a square raster of 10x10 (Liiii. Probe current and accelerating voltage were 10 nA (measured on Cu) and 15 kV, respective- ly. Elements analyzed included Ca, Sr, Na, K, and S. Salmonid otoliths are very susceptible to damage from the electron beam, and it was necessary to utilize a low- beam current and reduced counting times for all anal- yses. In conjunction with these conditions that reduce the total number of X-ray counts from a specimen, it is important to be aware of detection limits and count- ing statistics. A treatment of these subjects can be found in Goldstein et al. (1981) and references therein. 660 Fishery Bulletin 88(4), 1990 Details of the microprobe counting procedures, preci- sion, and standards employed in this study are outlined in Table 1 . Background was measured at offsets both above and below peak position for 50% of the peak counting time. The average counts from these two background measurements were subtracted from the peak. Background measurements were made with each analysis. Corrected X-ray intensity ratios were calcu- lated using the ZAF method (Reed 1975), and the final elemental ratios are normalized atom ratios based on concentrations derived from the standards. Elemental data for otolith life-history transects from adult salmonids were collected with approximately 40 ^m separating each 10 x 10 ^im sample region (50 ^ni separating the center point of each sample). This sam- pling spatial frequency was judged to be adequate because it was desired to collect data relating to major life-history changes only, and not recurring or seasonal events. For a comparison of otolith composition among the progeny of sea-farmed fish and freshwater fish, micro- probe data were collected from five individual primcn'- dia within each sagitta or lapillus. In some lapilli it was necessary to make multiple measurements on individ- ual primordia because five primordia were not always exposed or visible. In addition, five microprobe mea- surements were made on each otolith at points near the otolith edge that were indicative of the otolith composition at a time after the completion of yolk absorption. Results Otolith life-history transects representative of several types of adult salmonids are presented in Figure 1 . Differences in otolith transects among anadromous or sea-farmed fish and non-anadromous individuals are only evident in the Sr/Ca data and, therefore, only Sr/Ca life-history transects are presented. Sea trout, sea-farmed Atlantic salmon, and sea-farmed rainbow trout exposed to the marine environment display a clear increase in otolith Sr/Ca that is coincident with entry into seawater. There is an initial peak, associated with the otolith nucleus, in the Sr/Ca ratio of the sea trout and Atlan- tic salmon, but this peak does not appear in the sea- farmed or freshwater rainbow trout. In Tasmania, freshwater broodstock are used for the production of sea-farmed rainbow trout. The peak in otolith Sr/Ca in the otolith nucleus of some individuals is presumably due to the presence of strontium se(juestered in the egg yolk proteins during the seawater phase of ovarian development of the fishes' female parent. 0 004^ A 0 003- (j 0 002 . 0 001 - < 0 000- ,.A---V'''*''''VA^ 0 500 1000 1500 0 004 B 0 003- m. ..-m. P O 0 002 . -^^•^**-»V OT / 0 001, OOCO- ^*%*%^VvY^ 0 500 1000 1500 0 004 _, 1 c , ^ 0 003 - r^Y O 0 002 . \ / (5) \ / 0 001 . n 000 Sa^'vwX'**^ " ""^ 1 — ■ — i i — . — 1 ■ — ■ 1 — 1 — 1 , — 1 1 — 1 1 0 500 1000 1500 0 004 _ r V ° - %^A 0 003 . \ ic0.05). 2.68 ± 0.26 1.27 ±0.27 *P<0.01; **P« 0.001; * stock progeny (unpaired one-tailed t-test, /= 12.69, P<0.0001). One-way analysis of variance showed that there were no significant differences in mean otoHth primordia Sr/Ca ratios among individuals from either seawater (F=0.88, P = 0.476) or freshwater brood- stock (F = 1.29, P = 0.279). Differences in otolith pri- mordia composition between sea-farmed progeny and freshwater progeny were not significant for Na/Ca (un- paired two-tailed f -tests, t = 1.82, P = 0.076) and S/Ca (^ = 0.58, P = 0.568), but were significant for K/Ca ratios (/ = 3.88, P = 0.0004). Similar results were ob- tained for measurements made in lapilli primordia from five freshwater and five marine progeny. Table 2 also presents the results of comparisons between the com- position of the nucleus and post-yolk absorption regions of the otoliths. A histogram of the frequency distribution of Sr/Ca ratios determined for five primordia in the sagittae of 20 progeny of sea-farmed rainbow trout and 20 prog- eny of freshwater rainbow trout shows that there is little overlap between individual measurements from the two groups (Fig. 2). Only 5% of the measurements from the sea-farmed progeny overlap with 13% of the freshwater progeny measurements. A frequency histo- gram of the mean Sr/Ca ratios obtained from each sagitta nucleus shows that the sea-farmed and fresh- water progeny are clearly separated on the basis of otolith primordia Sr/Ca ratios (Fig. 2). 30- c □ Sea-farm progeny ■ Freshwater progeny M 201 Kl- jfe _, n n n n ^ - nnfln □ S Sea-farm progeny ■ Freshwater progeny JL DD .n,oi JX- (12 0,7 1 : 17 : ; ; ? ; ; 17 4.2 4.7 5.; Otolilh Sr/Ca X 10 ' Figure 2 (Aliove) F'requency distribution of i)idividual otolith Sr/Ca measure- ments made in the primordia of 20 sagittae from Ovrorhi/vcliita iiiykiss fry of sea-farmed Ijroodstoek and 20 sagittae from fry of freshwater broodstock {N = 200), and (below) frequency distribution of mean otolith Sr/Ca ratios from the same 40 sagittae. 662 Fishery Bulletin 88(4), 1990 Measurements of otolith composition made in regions of the sagittae formed after the completion of yolksac absorption and corresponding with a period soon after first feeding corroborated data that showed the influ- ence of yolk on otolith composition. There were no significant differences in Sr/Ca (unpaired two-tailed ^tests, / = 1.30, P = 0.201), S/Ca (t = 2.02, P = 0.051), or K/Ca {t = 1.09, P = 0.282) ratios measured in the later-formed regions of the otoliths from the two groups of fry and only slightly significant differences for Na/Ca ratios {t = 2.70, P = 0.010). Differences be- tween the Sr/Ca ratio of the sagitta edge and primor- dium were highly significant in the sea-farmed brood- stock progeny (paired two-tailed ^-tests, t = 10.38, P<0.0001), and only slightly significant in the progeny of the freshwater broodstock (t = 3.13, P = 0.003). Six electron microprobe transects over a sagitta from a single Oncorhynchus mykiss fry of sea-farmed brood- stock origin were made to show Sr/Ca ratios over the entire otolith including the nucleus and pre- and post- yolk absorption regions (Fig. 3). The outermost points towards the otolith rostrum and posterior show the sagitta Sr/Ca ratios after the completion of yolksac absorption and provide evidence for much reduced otolith Sr/Ca ratios when compared with the otolith material produced during the period of yolksac utiliza- tion. The data also indicate that, in general, high Sr/Ca ratios prevail throughout the region of the nucleus and are not confined to the primordia alone. Measurements of egg calcium and strontium com- position show that ova of the sea-farmed rainbow trout contain higher levels of strontium than their freshwater conspecifics. The mean strontium content was 0.054 ± 0.013 /iM/g of ova (dry weight) for ova that had devel- oped in seawater and 0.010 ± 0.002 jxMlg for ova from freshwater fish, and the difference between these values was significant (unpaired two-tailed /-test, / = 6.60, P = 0.0006). Mean calcium content of yolk was 0.035 ± 0.005 mM/g from sea-farmed broodstock and 0.026 ± 0.004 mM/g from freshwater fish and these values were also significantly different (/ = 3.06, P = 0.022). Figure 3 .Six transect.s of otcilith Sr/Ca nitios made with a wavt- length-dispersive electron microprobe on a single sagitta from an Onrorfufnchiis miikina fry of sea-farmed brood- stock origin. The otolith was ground to the level of the primordia in the sagittal plane. Each column in the histo- grams represents an individual mieroprolie measurement (iV = 120). Outermost points show the sagitta Sr/Ca ratios after the completion of yolk.sac absorption. Sagitta length (435 |jm) s6 70 60 Z 40 ^30 =^ 20 10 0 i ■■■ ■":'>'■ ^Mm 0 100 :(X) m) 400 Otolith distance (JJm) KalisiT Otolith microchemistry to distinguish salmonicl progeny 663 Discussion The results present evidence that the Sr/Ca ratios of otolith primordia formed during embryonic develop- ment are directly influenced by the strontium content of the individual's yolk and that yolk composition is also influenced by the composition of the waters where vitellogenesis took place. These findings are important to the study and management of salmonid species where anadromy is a faculative behavior. They may also be important to studies of other diadromous spe- cies, but it remains to be seen whether the relatively small nutritive contribution of the yolk in these fishes is adequate to influence otolith composition. However, since otolith primordia develop very early in the ontog- eny of fishes (Brothers 1984). there may be detectable differences in otolith strontium content within other species that display facultative diadromy. In many cases, the analysis of otolith primordium composition should make it possible to investigate the relative im- portance of parentage and environment in determin- ing diadromous behavior. The basis for the differences in otolith primordium composition among anadromous and non-anadromous fishes is such that there is probably little, if any, effect of race or population. Both wild and hatchery salmonids involved in this study displayed similar ranges of ele- mental ratios, depending on habitat (freshwater or marine) indicating that different salmonid species may incorporate into their otoliths a similar proportion of the ions present in the endolymph. Furthermore, this indicates that the relationship between endolymph and blood plasma composition is regulated in a similar man- ner in each of these species. Kalish (1989) showed that there was a tendency for both endolymph and otolith Sr/Ca ratios to be similar among individuals of a spe- cies, whereas differences were greatest among species. On the basis of the data presented here and in Kalish (1989), it seems likely that freshwater and marine salmonid otolith Sr content couFd be predicted on the basis of endolymph composition. The results of this study confirm that differences in the composition of freshwater and seawater can be reflected in the composition of fish otoliths. These find- ings are similar to those obtained by Casselman (1982), Radtke et al. (1988). and Kalish (1989). Wavelength- dispersive electron microprobe analyses of strontium content in diadromous fish otoliths can provide infor- mation on the rates of migration to and from the sea, residence times at sea and in freshwater, the age at which migrations take place, and confirmation of data obtained from scales. Confirmation of data obtained from scale reading is important because of the possibility of scale resorption (Bilton 1974) and the equivocal nature of data obtained from scales in the discrimination of adult anadromous and non-anadromous salmonids. Bagenal et al. (1973) used the strontium content of whole scales to distin- guish between brown trout and sea trout, and Moreau and Barbeau (1979) used a similar rationale to dis- criminate anadromous from non-anadromous whitefish Coregonus clupeaformis. However, Castonguay and FitzGerald (1982) found that this method was unreliable for distinguishing between anadromous and non-anad- romous brook char Salveliriusfontinalis, and Gausen and Berg (1988) had similar results when investigat- ing migratory and non-migratory Atlantic salmon. Measurements of otolith Sr may provide a more reli- able means of determining migratory behavior in these species because otoliths are not resorbed during peri- ods of stress (Simkiss 1974, Campana 1983), although Mugiya and Uchimura (1989) present evidence for otolith resorption in goldfish Carassius auratus dur- ing extreme anaerobic stress. Also, otolith Ca and, most likely, Sr are derived from ions taken up by the gills (Simkiss 1974), and any variations in the Sr con- tent of the diet would not be reflected in the composi- tion of the otoliths. This is not the case for scales, and differences in the Sr content of the diet as well as the effect of different levels of discrimination against Sr would be manifested in scale composition. It is important to point out the inadequacy of energy- dispersive (ED) electron microprobe analysis for the determination of Sr at the levels typically found in marine fish otoliths and particularly in the otoliths of freshwater species. Under optimal operating condi- tions, the minimum detectability limit for a wavelength- dispersive (WD) electron microprobe is approximately 10 times lower than that of an ED microprobe for all elements (Geller 1977, Goldstein et al. 1981). General- ly, the minimum detection limits attainable with an ED spectrometer are on the order of 0.10% wt (1000 ppm), and these conditions are frequently not realized due to various factors. The range of Sr concentrations mea- sured in this study was approximately 300-5000 ppm, and all marine fish otoliths investigated to date appear to contain Sr in excess of 1000 ppm (Radtke 1987, Edmonds et al. 1989, Kalish 1989). Theoretically the ED spectrometer should be capable of making all oto- lith Sr measurements in marine fish and some of those from freshwater fish. However, the detection and ac- curate quantitative estimation of Sr creates special problems for the ED microprobe. The X-ray line that is used to detect Sr at 1.806 keV (L^) is affected by the silicon escape peak associated with the large Ca K^ peak produced when measuring specimens that are largely calcium (aragonite fish oto- liths are approximately 38% wt Ca). The silicon escape peak results from the production of Si K^ photons following absorption of relatively high energy (> 1.841 664 Fishery Bulletin 88(4), 1990 keV) photons by the siHcon detector of an ED spec- trometer (Reed and Ware 1972). The energy of the photon producing the escape peak is E^ - Esi, where Ex is the incident photon energy and Egj is the energy of a Si Ko photon (1.739 keV). The silicon escape peak due to Ca K^ would have an energy of 1.951 keV (3.690 keV - 1.739 keV) and an intensity 0.78% that of the Ca K„ peak (Reed 1975). Such a peak, within less than 150 eV of the Sr peak, would create signif- icant errors in the estimation of low levels of Sr. Mea- surements of trace levels of phosphorus with the K„ line (2.013 keV) would be even more vulnerable to errors because of the closer proximity to the Ca K„ derived silicon escape peak. Other errors can result from a silicon internal fluorescence peak (Reed and Ware 1972) and the silicon absorption edge (Goldstein et al. 1981, Statham 1981), both artifacts of the ED spectrometer and associated silicon detector, and of particular interest to the determination of trace levels of Sr. These sources of error are not present when using a WD electron microprobe. There are numerous other potential sources of error in otolith trace-element studies carried out with ED electron microprobes. These errors can occur due to the relatively poor energy resolution of the ED spec- trometer and the resultant inability to discriminate between X-ray lines separated by less than approx- imately 150 ev (Statham 1981). Furthermore, the pres- ence of a greater proportion of continuum background radiation or "bremsstrahlung" in the ED spectrometer, due to decreased energy resolution, results in a five- fold increase in the peak to background ratio in a WD spectrometer over an ED spectrometer (Reed 1975). These factors, and others, make it necessary to view with caution studies with ED microprobes that inves- tigate the distribution of large numbers of trace (<0.5% wt) elements, particularly without reference to criteria for determining detection levels or precision. In an in- vestigation of otolith composition using X-ray fluores- cence spectroscopy (XRF) Cu, Cd, Cr, and V were below the detection limits of the instrument (<3.0 ppm), Ni was found at levels averaging 2.0 ppm, and Fe, Zn, and Ba levels were below 10 ppm in the four species studied (J.M. Kalish, unpubl. data). Edmonds et al. (1989) used inductively coupled plasma atomic- emission spectroscopy (ICP— AES) to study otolith composition of the pink snapper Chrysophrys auratus for stock discrimination, and their data indicate that Mg, Si, and Fe are at levels below the detection limits achievable using either ED or WD microprobe analysis. With the above results in mind it appears that with a WD electron microprobe the only elements that can be reliably quantified in otoliths are Ca, Na, Sr, K, S and, in some cases, CI. However, in the case of CI, the presence of this element in the most frequently used mounting medium, epoxy, makes the quantitative determination of this element difficult. If it is desired to detect CI, an alternative mounting medium should be considered. The utility of ED microprobe analysis to studies of the quantitative composition of fish oto- liths seems to be limited to Ca and, in some rare in- stances, Na and Sr. Although WD microprobe analysis is generally cap- able of detecting the levels of Sr found in freshwater and marine fish otoliths, there are some limitations to the method in studies of anadromous fishes. Most im- portant is the limited spatial resolution of the micro- probe. In this study the probe size was maintained at 10 X 10 ^m to minimize beam damage of the specimen. A spot of this size on the otolith of an adult fish would encompass more than a single day of otolith growth and in slow-growing individuals could encompass a month or more. In slow-growing fish, this would limit the ability to determine the temporal scale of migra- tory events. Determination of the temporal resolution of the microprobe data would be dependent on esti- mates of fish age based on otolith annuli and micro- increment data. In many instances, the utility of otolith Sr data is dependent on the accurate determination of age using the same otolith. The elemental analysis of otolith primordia should provide an objective criterion for assessing if an indi- vidual is the progeny of an anadromous or non-anad- romous female. This information alone, or in combina- tion with data from life-history transects, should make it possible to investigate the relationships between genetics and environment on diadromous behavior and aid in the management of species that display facula- tive diadromy. However, as with any new method, it is important to confirm the validity of these results for the particular species and habitat in question. Acknowledgments I thank Jane Andrew (Inland Fisheries Commission, Tasmania) for rearing the rainbow trout and supply- ing trout ova, and Stephanie Kalish, Peter Davies, and Laurie Cook, also of the Inland Fisheries Commission, for providing salmonids and ova from various localities. Citations Alderdice, D.F. 1988 Osmotic and ionic regulation in teleost eggs and larvae. Ill Hoar. W.S., and D.J. Randall (eds.l, Fish physiology, vol. XI. part A. p. I(i3-li.")l. Academic F'ress. San Diego. Bagenal, T.B.. F.J.H. MacKereth, and J. Heron 1973 The distinction between brown trout and sea trout by the strontium content of their scales. .1. Fish. Biol. .5:.5,55-.'5.57. Kalish' Otolith microchemistry to distinguish salmonid progeny 665 Bailey. R.E. 1957 The effect of estradiol on seiiim calcium, phosphoiTJS. and protein of goldfish. J. Exp. Zool. 136:4.55-469. Bilton, H.T. 1974 Effects of starvation and feeding on circulus formation on scales of young sockeye salmon of four racial origins, and of one race of young kokanee. coho and chinook salmon. In Bagenal, T.B. (ed.). Ageing of fish. p. 40-70. Unwin Bros., London . Booke. H.E. 1964 Blood serum protein and calcium levels in yearling brook trout. Prog. Fish-Cult. 26:107-110. Brothers, E.B. 1984 Otolith studies, /n Moser, H.G., et al. (eds.), Ontogeny and systematics of fishes, p. 50-57. Spec. Publ. 1, Am. Soc. Ichthyol. Herpetol., Allen Press. Lawrence, KS. Bruland, K.W, 1983 Trace elements in sea-water. In Riley, J. P., and R. Chester (eds.). Chemical oceanography, vol. 8, p. 157-220. Academic Press, London. Campana, S.E. 1983 Calcium deposition and otolith check formation during periods of stress in coho salmon, Oncorhynchus kisutch. Comp. Biochem. Physiol. A 75:215-220. Casselman. J.M. 1982 Chemical analyses of the optically different zones in eel otoliths. In Loftus, K.H. (ed.). Proe. 1980 North American Eel Conference, p. 74-82. Ont. Minist. Nat. Resour.. Ont. Fish. Tech. Rep. Ser. 4. Castonguay, M., and G.J. FitzGerald 1982 Critique de la methode de distinction entre poissons anadromes et dulcicoles de la meme espece par la teneur en strontium de leurs ecailles. Can. J. Fish. Aquat. Sci. 39: 1423-1425 [in French, Engl, abstr.]. Craik. J.C.A., and S.M. Harvey 1984 Egg quality in rainbow trout: the relation between egg viability, selected aspects of egg composition, and time of strip- ping. Aquaculture 40:115-134. Currens, K.P.. C.B. Schreck. and H.W. Li 1988 Reexamination of the use of otolith nuclear dimensions to identify juvenile anadromous and nonanadromous rainbow trout, Sabno gairdnen. Fish. Bull., U.S. 86:160-163. Edmonds. J.S.. M.J. Moran. N. Caputi. and M. Merita 1989 Trace element analysis of fish sagittae as an aid to stock identification: pink snapper (C/irj/sop/ij-i/s auratus) in Western Australian waters. Can. .J. Fish. Aquat. Sci. 46:50-54. Elliot. J.A.K., N. Bromage. and C. Whitehead 1979 Effects of oestradiol 17/3 on serum calcium and vitello- genin levels in the rainbow trout. .J. Endocrinol. 83:54-55. Gausen. D.. and O.K. Berg 1988 Strontium levels in scales and vertebrae of wild and hatchery-reared Atlantic salmon, Salmo salar L., smolts. Aquacult. Fish. Manage. 19:299-304. Geller, J.D. 1977 A comparison of MDL limits using energy dispersive and wavelength dispersive detectors. SEM/1977/I. p. 281-287. IIT Research Inst., Chicago. Goldstein. J. I.. D.E. Newbury. P. Echlin. D.C. Joy. C. Fiori. and E. Lifshin 1981 Scanning electron microscopy and x-ray microanalysis. Plenum Press, NY, 673 p. Hara. A., and H. Hirai 1978 Comparative studies on immunochemical properties of female specific serum proteins in rainbow trout (Sahno gaird- twri). Comp. Biochem. Physiol. B 59:339-343. Hara. A.. K. Yamauchi, and H. Hirai 1980 Studies on female-specific serum protein (vitellogenin) and egg yolk protein in Japanese eel {Anguilla japonica). Comp. Biochem. Physiol. B 65:315-320. Hayes. F.R.. D.A. Darcy. and CM. Sullivan 1946 Changes in the inorganic constituents of developing salmon eggs. J. Biol. Chem. 163:621-631. Jarosewich, E.. and LG. Maclntyre 1983 Carbonate reference samples for electron microprobe and scanning electron microscope analyses. J. Sediment. Petrol. 53:677-678. Jarosewich. E.. and J.S. White 1987 Strontianite reference sample for electron microprobe and SEM analyses. J. Sediment. Petrol. 57:762-763. Jarosewich. E.. J. A. Nelen, and J. A. Norberg 1980 Reference samples for electron microprobe analysis. Geostandards Newslett. 4:43-47. Assoc. Natl. Rech. Tech. Int. Working Group, Paris. Jonsson. B. 1985 Life history patterns of freshwater resident and sea-run migrant brown trout in Norway. Trans. Am. Fish. Soc. 114: 182-194. Kalish, J.M. 1989 Otolith microchemistry: Validation of the effects of phys- iology, age and environment on otolith composition. J. Exp. Mar. Biol. Ecol. 132:151-178. Laale, H.W. 1980 The perivitelline space and egg envelopes of bony fishes: A review. Copeia 1980:210-226. McKern. J.L.. H.F. Horton. and K.V. Koski 1974 Development of steelhead trout {Salmo ga irdneri) otoliths and their use for age analysis and for separating summer from winter races and wild from hatchery stocks. J. Fish. Res. Board Can. 31:1420-1426. Mommsen. T.P.. and P.J. Walsh 1988 Vitellogenesis and oocyte assembly. In Hoar, W.S.. and D.J. Randall (eds.). Fish physiology, vol. XI, part A, p. 347-406. Academic Press, San Diego. Moreau. G.. and C. Barbeau 1979 Differenciatiiin de populations anadromes et dulcicoles de Grands Coregones {Congonus chipeafonnis) par la composi- tion minerale de leurs ecailles. J. Fish. Res. Board Can. 36: 1439-1444 [in French, Engl, abstr.]. Mugiya. Y., and T. Uehimura 1989 Otolith resorption induced by anaerobic stress in the goldfish. Oinissius nuratus. J. Fish. Biol. 35:813-818. Neilson. J.D., and G.L. Geen 1981 Method for preparing otoliths for microstructure exam- ination. Prog. Fish-Cult. 43:90-91. Neilson. J.D.. G.L. Geen. and B. Chan 1985 Variability in dimensions of salmonid otolith nuclei: im- plications for stock identification and microstructure interpre- tation. Fish. Bull.. U.S. 83:81-89. Nordeng. H. 1983 Solution to the "char problem" based on Arctic char (Salvelinus aipinus) in Norway. Can. J. Fish. Aquat. Sci. 40: 1372-1387. Oguri. M., and N. Takada 1967 Serum calcium and magnesium levels of goldfish with special reference to gonadal maturation. Bull. Jpn. Soc. Sci. Fish. 33:161-166. Radtke. R.L. 1987 Age and growth information available from the otoliths of the Hawaiian snapper. Pristipomoides filamenlostts. Coral Reefs 6:19-25. 666 Fishety Bulletin 88(4), 1990 Radtke, R.L., R.A. Kinzie III, and S.D. Folsom 1988 Age at recruitment of Hawaiian freshwater gobies. Environ. Biol. Fish. 23:205-213. Ramdohr, P., and A. El Goresy 1971 Einiges iiber den meteoriten von Mundrabilla in West- australien. Chemie der Erde 30:269-285 |in German, Engl, abstr.]. Reed, S.J.B. 1975 Electron microprobe analysis. Cambridge Univ. Press, Cambridge. 400 p. Reed, S.J.B., and N,G. Ware 1972 Escape peaks and internal fluorescence in x-ray spectra recorded with lithium drifted silicon detectors. J. Phys. E Sci. Instrum. 5:582-.585. Rosenthal, H.L., M,M. Eves, and O.A. Cochran 1970 Common strontium of mineralized tissues from marine and sweet water animals. Comp. Biochem. Physiol. 32: 445-450. Rvbock, J.T., H.F. Horton, and J.L. Fessler 1975 Use of otoliths to separate juvenile steelhead trout from juvenile rainbow trout. Fish. Bull., U.S. 73:654-659. Simkiss, K. 1974 Calcium metabolism of fish in relation to ageing, hi Bagenal, T.B. (ed.). Ageing of fish, p. 1-12. Unwin Bros., London. Skoryna, S.C. (editor) 1981 Handliook of stable strontium. Plenum Press. NY, 644 p. Statham, P.J. 1981 X-ray microanalysis with Si (Li) detectors. ,1. Microsc. (Oxf.) 123:1-23. Woodhead, P.M.J. 1968 Seasonal changes in the calcium content of the blood of Arctic cod. J. Mar. Biol. Assoc. U.K. 48:81-1)1. Zeitoun, I.H„ D,E. Ullrey. W.G. Bergen, and W.T, Magee 1976 Mineral metabolism during the ontogenesis of rainbow trout {Satnio gairdiieri). .J. Fish. Res. Board Can. 33: 2587-2.591. Abstract.— Density and biomass estimates from a shrimp trawl and a visual survey are compared. Results indicate that the visual survey gives estimates eight to nine times larger than the trawl survey. However, there are important variations by species. Estimates of fish size from the visual censuses are larger than fish sizes observed in the trawl catch. These results suggest that the catch- ability of fish by shrimp trawls may be lower than usually thought for multispecies tropical stocks. In shrimp trawl fisheries, this may have impor- tant consequences for stock assess- ment and for evaluation of param- eters such as fishing mortality. Trawling should be adequate for qualitative and semiquantitative stock assessment, but will need to be compared with other methods for quantitative studies on multispecific stocks. Comparison Between Fish Bycatch from Slirimp Trawlnet and Visual Censuses in St. Vincent Bay, Mew Caledonia Michel Kulbicki Laurent Wantiez ORSTOM, Centre de Noumea B P A5 Noumea Cedex, New Caledonia Manuscript accepted 29 June 1990. Fishery Bulletin, U.S. 88:667-67.5. Shrimp trawl fisheries are the source of a large fish bycatch in the tropical Indo-Pacific (Aoyama 1973, Gran- tham 1980, Villoso and Hermosa 1982, Chong 1984). One question con- cerning these fisheries is how much damage they cause to fish stocks. To answer this question, it is useful to assess the catchability of fish by shrimp trawls. Many methods exist to assess this parameter. Most use the same gear under different condi- tions, such as double codends, split codends, and repHcate hauls with dif- ferent mesh sizes (Macket 1973, Gul- land 1975). The bias of such methods may be acceptable for single-species populations but has seldom been studied in relation to tropical multi- species stocks. Another approach is the cross-evaluation of trawling and echosurveys (Doubleday 1976, Dines 1982, Dickie et al. 1983), but this method is not well adapted to tropical multispecies stocks. Another way is to correlate the results of two methods, such as with a TV camera and otter trawl (Uzmann et al. 1977), submersible and longlines (Ralston et al. 1986, Richards and Schnute 1986, Grimes et al. 1982), visual census and longlines (Kulbicki 1988) or poisoning and beamtrawl (Gray and Bell 1986). The use of several methods has the advantage of allowing some evalua- tion of the biases of each method. In the present study it was possible to compare the results of visual cen- suses and shrimp trawl bycatches. This work was originally designed to estimate the catchability of the net (q = number of fish caught/number of fish present), but it also provided in- formation on the biology of the fish. Material and methods Location The study was performed during August 1986 on the trawling grounds of the South Bay in St. Vincent Bay, in the southwest lagoon of New Cale- donia. The trawling gi-ounds cover 15 km-, or approximatively one-tenth of the Bay area. A total of eight sta- tions were sampled (Fig. 1), one of which was sampled at night (Station n°6). Visual census A 200-m transect line was laid on the bottom with a buoy set at each end of the line. The transect was divided into 10-m sections. For each section, two divers— one on each side of the line— counted all fish, estimated their length, and estimated the perpen- dicular distance of the fish to the transect. Length was given in 2-cm size classes for fish less than 20 cm, 5-cm size classes for fish between 20 and 50 cm, and 10-cm size classes for fish larger than 50 cm. Only fish less than 1.5 m above the bottom were counted. The distance from the fish to the transect was recorded in 1-m 667 668 Fishery Bulletin 88(4). 1990 classes up to 5 m, and in 2-m classes beyond 5 m. Fish were not recorded beyond 10 m from the transect. Both divers had a good knowledge of the fish fauna and also had a good training in visual censuses. The transects were finished between 30 and 45 minutes before the start of the trawling. For this type of data, many density estimators may be chosen (Burnham et al. 1980). Among these esti- mators, Kulbicki and Duflo (unpubl. data) show that the most robust one for fish density (D) in similar condi- tions is given by: where D = fish density (fish/ha), p = number of species, nj = number of fish of species i, dj = average distance (m) of species i to the transect, and L = length of the transect (200 m). Biomass B (kg/ha) was calculated in a similar fashion: B = (10/2L) 5! (Wi/dj) D = (10J/2L) 5! (n,/(l, 1 = 1 where W; = weight of species i (g). The weights of fish were estimated from length-weight relationships (Wantiez and Kulbicki In press). Kulbicki and Wantiez: Estimates of fish stocks by shrimp trawl and visual survey off New Caledonia 669 Trawling The RV Vauban (24 m long) towed a shrimp trawlnet of the semiballoon, floridian type with a 14-m headrope and a 2-cm mesh codend. The vertical opening was 1.2 m and the distance between the otterboards was 7 m at a speed of 2 knots. Towing speed varied from 2 to 2.5 knots. The hauls were each 800 m in length ( + 80 m). The net was submerged immediately after completion of the visual census at a distance of 400 m from the first buoy of the transect. The track of the haul was along the transect line and the net was retrieved 200 m after the second buoy of the transect line. All fish caught were sorted to species (Rivaton et al. 1990). All fish were counted and weighed. For the calculation of density D (fish/ ha), no catch- ability coefficient was used. Therefore, D = X (n,/S) 1=1 where p = number of species caught, nj = number of fish of species i, and S = surface area of the haul (7 m x 800 m = 5600 m- = 0.56 ha). Biomass (kg/ha) B was estimated in a similar fashion, B = X w,/S) i = l where w, = weight of species i (kg). Results Species A total of 82 species were either caught by trawls or recorded by divers (Table 1). All the common species known to occur in trawls in St. Vincent Bay (Kulbicki and Wantiez 1990) were found in the present study except for the Leiognathidae (ponyfishes). This family, represented by eight species in the bay, is character- ized by large fluctuations in trawl catches depending on season and locality. In the present survey, Lei- ognathus rivulatus which is normally an uncommon species in the catch, was the major Leiognathidae. Table 1 indicates that the trawls caught more species (64) than were seen on the transects (51). This could be due to the larger area covered by the trawls (~5600 m-/trawl versus 700-1000 m- per transect). Most cryptic species, such as Scorpaeniforms, Platycepha- lidae (flatheads), flatfishes, and small Balistidae (trig- gerfishes) (including the filfishes Paramonacanthus japonic2is and Pseudaiutarius nasicornis) were poor- ly represented in the visual transects but caught by the trawls. Conversely, small species such as Apogonidae (cardinalfishes) and Pomacentridae (damselfishes) could be detected on the transects but were too small to be retained by the net. Large, fast-swimming species such as Carangidae, Serranidae, or Scombridae were seen on the transects but evaded the trawls. Gobiidae, which are burrowing species, were also seen during the dives but absent from the trawl catch. A total of 32 species were detected by both methods (Table 1). Among these, only eight species were seen or caught in more than 50% of the samples for both methods [Saurida undosquamis, Synodus hoshinonis, Leiognathus rivulatus, Lethrinus nematacanthus, Upe- neus tragula, Upeneus sp. aff. asymetricus, Pristotis jerdoni, and Cantkigaster compressa). These eight species will subsequently be referred to as "main spe- cies." Except for Synodus hoshinonis and Leiognathus rivulatus, all these fish are among the 15 most frequent species in the trawl catch of St. Vincent Bay (Kulbicki and Wantiez 1990). The number of species per station was statistically larger (t test at a = 0.05) for the trawls than for the transects (Table 2). However, a chi-square test for independent samples (Siegel and Castellan 1988) indicates that the ranking of the stations is not sig- nificantly different (at a = 0.05) between methods. Fisfi size Only fish lengths were used to compare fish size be- tween the two methods. Table 3 indicates that visual censuses always yielded larger lengths than trawls, except in the case of Pristotis jerdo7i i. However, the lengths are significantly different (t test at a = 0.05) for only four species (Table 3), and the two methods give mean lengths which are less than 2.8 cm apart. Due to the fact that fish length was estimated in 2-cm size classes for transects and 0.5-cm size classes for trawls, one should be cautious about the significance of the differences observed. In a review on the problems of visual transects, Harmelin-Vivien et al. (1985) indicate that fish size is usually underestimated by this method. In the present study, if this were the case, it would mean that the larger fish (within a given species) are able to evade the trawl. In our opinion this is not the only reason, and the discrepancy of fish length between transects and trawls is probably also due to an overestimate of length by the divers. 670 Fishery Bulletin 88(4), 1990 Table 1 Species composition of transects and trawls in St. Vincent Bay, New Caledonia. For dives and trawls, numbers indicate at how many stations the species were recorded. Species Dives Trawls Species Dives Trawls Muraenidae Muraenidae spp. Synodontidae Saurida gracilis Snurida nelmlosa Sa urida undosquamis Synodus demiatogertis Sifnodus hoshinonis Trn ch y n oreph a his m yops Carapidae Carai)ux hotnei Fistularidae Fifit K ta ria pet i m ha Syngnathidae Hippocampus hiatrix Dactylopteridae Dactyloptena orienialis Scorpenidae Scorpaenidae sp. Dendrochirus hnifhyptenif: Pterois volitans Synanceiidae Inimicus didactylus Aploactinidae Aploactis asjiera Platycephalidae Platycephalidae spp. Oniyocia spinosa Serranidae Cephalopholis hoenack Epinephelus cyanopodus Epinephelus maculatiis Epinephelus malabaricus Priacanthidae Prtacantltu^ hanirur Apogonidae Afiogon spp. Apogon catalai Apogon ellioti Apogon aureus Apogon frenalus Apogon sp. cf compressus Cheilodipterus qiintifiifUneatus Rhabdamia spp. Carangidae Decapterus russeli Gnathnnodon specioxus Siganidae Siganus ca n n I icuUi t ns Scombridae Sco mberomorus Ciimm erso n i Bothidae A rnoglossus sp. Asterorhoyrdyus inlermedius Engyprosopon grandisquama Grn m rnn tobofh uk potyoplith a I m us Balistidae A bii listes stetla tus Porn monnca n th k,s jn pon inix Psi-udo hit a rius n as ico mis 0 0 0 0 1 0 0 1 0 2 2 3 0 2 2 4 0 7 1 1 1 1 2 1 1 0 3 2 7 3 0 0 0 1 2 1 0 1 0 4 1 0 1 1 0 1 0 1 1 0 1 2 1 0 0 3 1 0 0 1 1 8 1 7 0 3 0 3 1 7 2 4 Leiognathidae Leiognathus rivulatus Semtor ruconms Lutjanidae Lutjanus quinqueiinealus Lutjanus vittus Gerreidae Gerres omhis Haemulidae Diagranuna puiutn Lethrinidae Lethrinus nemalacanthus Lethrinus semicinctus Neniipteridae Ni'm ipterus peroni Scolopsis tern porn I is Mullidae Parupeneus plcurospilos Upeneus sp. Upeneus moluccensis Upeneus traguln Upeneus sp. aff. asymetricus Chaetodontidae Hen loch us aruminat us Pomacentridae Chromis fumea Dascilhis aruanus Neopoinacentrus sp. Allen Pristotis jerdoni Labridae Anampses spp. CheUinus himaculatus Xiphocheilus typus Scaridae Srarus ghobban Mugiloididae Pitrapercis sp. Parapercis cyhndnco Blenniidae Pctroscirtes bri'viceps Callionymidae Synch iropus ni tncus (lobiidae Gobiidae sp. Amblyeteotris sp. Amblygobius sp. Ptereleotris hanae Ostraciidae Tetrasoma gi.bbosus Lnctoria cornulo Tetraodontidae A mblyrhynchotrs hypselogcneion A roth run im m acula t us Arothron st.eUatus Canth igaster compressa Canthigaster valentini Lagocephalus sceleratus Total number of species Species common to lioth methods All .species 1 II (1 1 1 U 4 5 2 II •; 0 1 f) 1 II 0 5 n 3 0 3 3 1 6 3 8 8 2 fl 1 1 51 32 82 i;4 Kulbicki and Wantiez Estimates of fish stocks by shrimp trawl and visual survey off New Caledonia 671 Table 2 Number of species per station according to transects and trawls, St. Vincent Bay, New Caledonia. CI = confidence interval at a = 0.0.5. Station number 1 2 3 4 5 6 7 8 Average CI Transects 13 11 18 Trawl 22 31 25 4 16 18 20 = 7.(33. Limit value = 16 14 15 13.4 9.5-17.2 38 21 27 25.3 19.3-31.2 14.07 (7 df) at a = 0.05 (Siegel and Castellan 1988). Chi-square value for two independent samples Station 6 was performed at night. Table 3 Comparison of fish lengths given by transects am travvli ng for main species, St. Vincent Bay, New Caledonia. Species Trawl Transect F t L n s/\/n L n s/\/n Saurida undosquamis 22.7 108 0.07 23.3 8 1.10 ** 0.50 NS Platycephalus sp. 14.6 12 0.29 15.0 1 — — Leioqnathus rimdatus 8.2 130 0.07 9.2 499 0.41 — 1.30 NS Lethrinus nematacanthtis 12.2 186 0.12 14.3 244 0.08 — 14.9 *♦ Scolopsis tetnporalis 14.0 22 0.58 14.8 4 0.25 « * 1.16 NS Upeneus tragula 12.0 26 0.60 14.8 30 0,55 NS 3.46 ** Upeneus sp. aff. asymetricus 10.9 329 0.07 12.0 46 0.20 — 5. .52 ** Pristotis jerdoni 9.1 46 0.08 8.7 490 0.04 — 4.92 ** Paramonacanthns japoninis 9.0 30 0.13 10.0 1 — - Pseudalutarius nasicomis 11.3 3 0.83 12.6 5 1.12 — 0.78 NS Canthigaster compressa 8.3 217 0.06 8.6 38 0.20 1.62 NS L = length in cm, n = sample size. s/v'n = standard error, F = Fmax test for homogeneity of variance (Sokal and Rohlf 1981) performed if «<30, / = Student test if F not significant, otherwise t ' test for samples of unequal variance (Soka 1 and Rohlf 1981), " = significant at o = 0.01. NS = not significant at o = 0.05. dashes ind cate not calculated. Abundance and biomass Density estimates frcim tfansects were on average 9.7 times larger tlian from trawls, this diffei'ence being highly significant (F test for paired comparison (Sokal and Rohlf 1981) at a = 0.01) (Table 4). Depending on the species, the ratio between the two methods varied from 0.9 to 80 (Table 5). If one considers only the eight most common species mentioned in the previous sec- tion, then the ratio is 7.93. This would indicate a catch- ability of the trawl net of 0.103 for all species and 0.127 for the main species. Species for which the ratio is close to 1 (Synodontidae, Cnnthign^ter eompres:sa, Upeneus spp.) are difficult to detect underwater. This can be due to either their behavior (e.g., Synodontidae are usual- ly tnotionless on the bottom, and at times, half buried in the sand) or to their coloration (Canthigatiter com- pr-essa being well camouflaged among algae, and Upeneus spp. being able to drastically change their coloration to mimic the bottom). In addition. Table 6 indicates that these criptic species are usually found singly or in pairs. Species with a high ratio (Apogonidae, Leiognntluis rivulatus. Pristotis jerdo)ti) are not caught by the net for two main reasons. Either they are too small (Apo- gonidae, most Pristotis jerdoni) or they swim too high above the bottom (Leiognathus rivulatus were usually 0.5-3 m aliove the bottom) for the net to catch them. These fish are also found in small schools (21-36 fish/ sighting) (Table 6). The correlation between the two methods (Table 5) is some indication of the patchiness of the distribution of these species. Thus a high cor- relation such as for the Synodontidae (?' = 0.75) or Leiognatlius rirulntiis (r = 0.92). indicates that these 672 Fishery Bulletin 88(4), 1990 Density estimates (fish /ha) fur all s| (a = 0.05). Table 4 lecies for visual transects ami trawls. St. \' incent Bay. New Caledonia CI = confidence interval Station number Average CI 1 2 3 4 5 6 7 8 Transect Trawl 6020 7190 245 620 paired comparison 1930 8650 183 312 = 15.1. Limit value = 2610 1160 590 920 5. .59 (1. 7 df) at a 1750 197 = 0.05. 4340 380 4200 431 1690-6700 200-662 F test value for Table 5 Comparison of density and hiomass estimates from transects and trawls for main families and species, St. \'incent Bay, New Caledonia. Density Biomass species ratio r ratio r Synodontidae 0.88 0.75* 0.80 0.93** Apogonidae 80 0.66 13 0.78* Leiognathuts rivulatus 33 0.92* * 49 0.97** Lethrinua iii'miifacanthiifi 7.4 -0.03 10 -0.02 Upeneus spp. 2.3 0.63 4 0.57 Pristotis jerdoni 29 0.04 39 0.03 Canthigaster compressa 1.2 0.49 1.3 0.40 All species 9.7 -0.17 9.1 0.72" Main species (see text) 7.9 0.23 6.4 0.34* Ratio = density from transects /density from trawls. r = correlation coefficient, * .significant at o<0.05, ** significant at o<0.01. ' night station (n°6) not included. species are rather uniformaly distributed on the bot- tom. Conversely, species such asLethrirmfi nematacan- thus (r = 0.03) or Pnstotis jerdoni (r = 0.04) have a very patchy distribution. A pratical consequence is that a larger number of stations will be needed to get a good stock assessment for the latter species than for the former. Biomass estimates from transects are 9.1 times larger than those from the trawls (Table 7), this dif- ference being significant (F test for paired comparison (Sokal and Rohlf 1981) at o = 0.05). Although density estimates for all species were not correlated between trawls and transects (r= -0.17), biomass estimates show some correlation between the two methods, if the night station n°6 is excluded (r = 0,72, significant at a = 0.07). The absence of correlation between the two Table 6 Average distance between sightings and number of fish |jer sighting for main families and species, St Vincen Bay, New Caledonia. Average distance Family or between sightings Number of fish | species (m) pel sighting Synodontidae 110 1 Apogonidae 260 35 Leiognathux rivulatus 120 36 LethrinuK iicmatacanthus 28 4 Upeneus spp. 35 1.8 Pristotis jerdoni 80 21 1 'anthigaster r(i))iprrssa 55 1.3 methods in the density estimates may be explained by the large contribution of small fish such as Pomacen- tridae to transect density estimates, whereas these species escape the net. Conversely, biomass estimates are almost unaffected by these species because of their small weight. At the species level the correlations be- tween the two methods for biomass estimates are of the same magnitude as those observed for density estimates. Discussion The comparison of two methods with such different concepts requires some adjustments. Trawls swept approximately 5600 m- and transects covered be- tween 700 and 1000 m- (depending on species) for each station. The trawls did fish over the transect lines, but a perfect match would mean a trawl track with a precision of ± 1 m which is untractable even in shallow water. This is, in our opinion, the main source of differ- ence in the number of species given by each method. Kulbicki and Wantiez- Estimates of fish stocks by shrimp trawl and visual survey off New Caledonia 673 Table 7 Biomass estimates (kg/ ha) for transects and trawls, St. Vincent Bay, New Caledonia. CI = confidence interval at a = 0.05. Station number CI 1 2 3 4 5 6 7 8 Average Transect 225 242 49 Trawl 10 19 8.3 65 160 4.9 11 value is 5,59 (1,7 df) 43 123 34 15 at o = 0.05 (Sok; 74 123 6 13.5 1 and Rohlf 1981). 53-193 5.0-22.0 F test value for paired comparison = 15.0. Limit Because of the low number of stations, this difference in sampled area may also be important to explain the low correlation between the two methods for density or biomass estimates of fish with a patchy distribution. Conversely, comparisons of the density or biomass for fishes with a regular distribution should not be affected. An important result from the present survey is the magnitude of the difference between trawl and transect estimates of density. This difference, which is a factor of 9.7 for all species and a factor of 7.9 for the main species, certainly indicates that shrimp trawls poorly sample the bottom fish fauna in St. Vincent Bay. Visual transects are usually performed as strip transects, where distance of the fish to the transect line is not taken into account (Harmelin-Vivien et al. 1985). This underestimates density, especially if the width chosen is large (Burnham et al. 1980). The method chosen in the present survey is more accurate; however, it is not possible to know by how much one underestimates or overestimates the "real" density. Fish numbers (Har- melin-Vivien et al. 1985) and distances are both under- estimated but play opposite roles in the density estima- tion. Even in the unlikely case of a large overestimate of fish density (let's suppose by 100%) by transects, the catchability of the trawl would still be 0.25 for the main species and 0.21 for all species. Several surveys indicate that trawls may not be as efficient as usually thought. Thus Uzmann et al. (1977) found that estimates of fish density from submersibles were 8.0 times greater than density estimates from otter trawls, this ratio varying between 0.8 and 18 depending on the species. Similarly, Gray and Bell (1986) found that poisoning indicated densities 2.5 to 6.5 times greater than a beam trawl over seagrass beds. However, in some conditions trawling may be more efficient, as shown by Harden Jones et al. (1977) who used accoustical tags to estimate that 44% of the tagged flatfishes Pleuronectes platessa were caught by a Granton otter trawl. Uzmann et al. (1977) had found that only 7% of the flatfishes seen from a submersible were caught by the otter trawl. Serebrov (1986) used submersibles to show that the catchability of an otter trawl varied with the size of the fish, mesh size, and overall catch composition. In the case of one or a few species, catchability of trawls may vary between 0.5 and 1.0. Such values of catchability are widely used, even in multispecies tropical fisheries, for which Pauly (1982) states that a value of 0.5 is "realistic" and Gulland (1979) even sug- gests that a catchability of 1.0 should apply to the eastern Indian Ocean trawl fisheries. In many of these fisheries, shrimp trawls of design similar to the one used in the present experiment are used to get a mixed catch of shrimp and fish (Grantham 1980, Poiner and Harris 1986, Lamboeuf 1987, Sainsbury 1987). Our results indicate a catchability near 0.1 (0.103 for all species and 0.127 for the main species). Such a value may need further testing; but if it proved correct, it would have important implications in the management of these tropical multispecies trawl fisheries. In par- ticular, depending on whether one uses a catchability of 1.0 or 0.1. stock size will increase by a factor of 10. This would also affect the estimated fishing mortality (Pauly 1982) and, as a result, most of the equations used in stock management. What is more important, in our opinion, is that if one applied a catchability of 0.1 in most of these tropical trawl fisheries of the Indo- Pacific, the current models would be unable to explain the long-term decline of the catch seen in these fish- eries. This could mean that trawling induces detrimen- tal changes in fish populations which are beyond the simple removal of fish. In particular, habitat changes may be important, as noted l.iy Poiner and Harris (1986) and Sainsbury (1987). Another question arising from the present work is the value of the catch per unit effort (CPUE) of trawls as an indicator of population abundance. Our results show a poor correlation between abundance estimates from trawls and visual transects at the population level, but good correlations for a few selected species. Nu- merous studies have examined the correlation of CPUE by trawling and acoustic surveys. Most of these indicate good correlations (see Tesler (1977), Olsen et al. (1977), or Thorne (1977a) among others) but important varia- 674 Fishery Bulletin 88(4). 1990 tions may result from fish behavior (Thorne 1977b). However, most of these surveys are performed on a single pelagic or semipelagic species, and the correla- tion between acoustic surveys and CPUE by trawling for multispecies tropical stocks is currently poorly documented. Soetre and Paula e Silva (1979) found from comparison between demersal trawl catch and echointegration on a muitispecific stock off Mozam- bique that the catchability of the otter trawl was 0.3. These authors state that this value is "suspiciously low for demersal species." Guillory et al. (1982) concluded from replicate sampling with otter trawls that this type of gear should be used strictly for qualitative purposes in a multispecies situation. This is in part supported by the present work, which indicates a need for alter- nate methods of stock assessment. Longlines show similar problems to trawling. Thus, correlation between longline CPUE and observed densities from a submer- sible was weak at the multispecies level (Richards and Schnute 1986, Ralston et al. 1986) but satisfactory for some selected species. Similar work involving longline CPUE and observed densities by visual transects in- dicated a good correlation for all species (r = 0.88) but considerable variations depending on species (Kulbicki 1988). Visual census either by diving or from submer- sible might be an alternative, but these methods also have numerous limitations. In particular, they do not allow a quick coverage of large areas and are not usable in turbid waters. If reasonably accurate density esti- mates are required in a multispecies tropical trawl fishery, it is likely that this can be achieved only by a combination of several independent methods. Citations Aoyama, T. 1973 The (iemersal fish .stocks ami fisheries of the South China Sea. South China Sea Development and Cooi-ilinatinK Pro- gramme. IPFC, FAO, Rome. 80 p. Burnham K.P.. D.R. Anderson, and J.L. Laake 1980 E;stimati(in of density from line transect sampling of biological populations. Wild]. Monogr. 72, 202 p. Chong, V.C. 1984 Prawn resource management in Ihe west coast of penin sular Malaysia. Wallaceana (Kuala Lampur) 37:3-6. Dickie. L.M., R.G. Dowd. and P.R. Boudreau 1983 An echo counting and logging system (ECOLOC) for demersal fish size distributions and densities. Can. .1. Fish. Aquat. Sci. 40:.}87-498. Dines, N. 1982 Essai d'application de I'echointegration a la mesure de I'efficacite des ch.aluts. Proc. Council Meeting, ICES-CM- 1982/B:12, 11 p. |in French, Engl. abstr.|. Doubleday. \V.(;. 197(i A pilot stnily of a survey design for a combined acoustic and otter-trawl groundfish survey. ICES CM lit7t;/l^, 30 p. Grantham. G.J. 1980 The prospect for bycatch utilization in the (utlf area. Regional fishery survey and development project. Barhain, Iran, Iraq, Kuwait. Oman. Qatar, Saudi Arabia, United Arab Emirates, FAO, Rome, 43 p. Gray, C.A., and J.D. Bell 1986 Consequences of two common techniques for sampling vagile macrofauna associated with the seagrass Zos/crn capri- corni. Mar. Ecol. Prog. Ser. 28:43-48. Grimes. C.B.. K.W. Able, and S.C. Turner 1982 Direct ol)servation from a submersible ves.sel of oonuiier- cial longlines for tilefish. Trans. Am. Fish. Soc. 111:94-98. Guillory, V., J.E. Roussel, and C. Miller 1982 Appraisal of otter trawl tow lengths and replicate sam- pling. Proc. Annu. Conf. Southeast. Assoc. Fish. Wildl. Agen- cies 34:158-166. Gulland, J. A. 1975 Manual of methods for fisheries resource survey and ap- praisal. Part 5. Objectives and basic methods. FAO Fish. Tech. Pap. 145, 29 p. 1979 Report of the FAO/IOP workshop on the fishery re- sources of the western Indian Ocean south of Equator. IOFC/DEV/79/45, FAO. Rome, 99 p. Harden Jones, F.R., A.R. Margetts, M. Greer Walker, and G.P. Arnold 1977 The efficiency of the granton otter trawl determined by sector-scanning sonar and acoustic tran.sponding tags. Rapp. P.-V. Reun. Cons. Int. Explor, Mer 170:4,')-.S1. Harmelin-Vivien. M.L.. J.G. Harmelin. C. Chauvet. C. Duval. R. (iaizin. P. Lejeune, (J. Barnabe. F. Blanc, R. Chevalier. J. Duclerc. and G. Las.serre 1985 Evaluation visuelle des peuplements de poissons: Methodes et problemes. Rev. Ecol. Terre Vie 40:467-539 |in French]. Kulbicki, M. 1988 Correlation between catch data from bottom longlines and fish censures in the SW lagoon of New Caledonia, hi Choat, J.H. etal. (eds.), Proceedings. 6th International Coral Reef Symp., Townsville, Australia, 8-12 August 1988. Vol. 2, p. .30.5-312. Kulbicki, M.. and L. Wantiez 1990 Variations in the fish catch composition in the Bay of St Vincent, New Caledonia, as determined by experimental trawl- ing. Aust. .1. Mar. Freshwater Res. 41 (1): 121 -144. Lamboeuf, M. 1987 Bangladesh demersal fish resources of the continental .shelf. R/V Anii.^iimlhririi trawling survey results. September 19S4-.Iune 1986. FI:DP/B(;i)/S0/025. FAO. Rome, 79 p. Mackett. D.J. 1973 Manual of methods for fisheries resource survey and ap- praisal. Part 3. Standard methods and techniques for demer- sal fisheries resource surveys. FAO Fi.sh. Tech. Pap. 124, 39 p. Olsen. S., S. Tveite, and D. Chakraborty 1977 Acoustic surveying in tropical waters, ii.ipp. I'. V. Reun. Cc.ns. Int. Kxplor. Mer 170:248-2.52. Pauly. 1). 1982 I'ne selection de methodes simples |iour I'estimation des stocks de poi.ssons tropicau.N. F.-\0 circulaire sur les peches 729, 63 p. FAO. Rome |in French]. Poiner, I.R., and A. Harris 1986 The effect of commercial |)rawn trawling on the demer- sal fish communities of the south eastern Gulf of Carpen- taria. Ill Haines, A.K., G.C. Williams, and D. Coates (eds.). Torres Strait Fisheries Seminar, p. 239 -259. Aust. Gov. Publ. Serv,, Canberra. Kulbicki and Wantiez- Estimates of fish stocks by shrimp trawl and visual survey off New Caledonia 675 Ralston, S.. R.M. Gooding, and G.M. Ludwig 1986 An ecological survey and comparison of bottom fish • resource assessments (submersible versus handline fishing) at Johnston Atoll. Fish. Bull., U.S. 84:141-155. Richards, L.J., and J.T. Schnute 1986 An experimental and statistical approach to the question: Is CPUE an index of abundance? Can. J. Fish. Aquat. Sci. 4.3:1214-1227. Rivaton, J., P. Fourmanoir, P. Bourret, and M. Kulbicki 1990 Checkliste preliminaire des poissons de Nouvelle- Caledonie. Rap. Sci. Tech. 54, ORSTOM Oceanogr.. Noumea, New Caledonia, 170 p. [in French, Engl, abstr.]. Sainsbury, K.J, 1987 Assessment and management of the demersal fishery on the continental shelf of northwestern Australia. In Polovina, J.J., and S. Ralston (eds.), Tropical snappers and groupers: Biology and fisheries management, p. 465-503. Westview Press, London. Serebrov, L.I. 1986 Differential catchability of bottom trawls as revealed by underwater exploration. In Underwater exploration for fishery purposes, p. 21-38. Pinro, Murmansk, LISSR [in Russ.]'. Siegel, S., and N,J. Castellan Jr. 1988 Nonparametric statistics, 2d ed. McGraw-Hill Int. Ed., Stat. Ser., NY, 399 p. Soetre. R., and R. de Paula e Silva 1979 The marine fish resources of Mozambique. Reports on surveys with the R/V Dr Fridtjof Mansen. Servio de Inves- tigaoes Pesqueira, Maputo. Inst. Mar. Res., Bergen, Norway, 179 p. Sokal, R.R., and F.J, Rohlf 1981 Biometry, 2d ed. W.H. Freeman, San Francisco, 859 p. Tesler, V,D, 1977 Measuring of average amplitude of echoes for fish den- sity estimation. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 170:1,52-158. Thome. R,E. 1977a Acoustic assessment of Pacific hake and herring stocks in Puget Sound, Washington and southeastern Alaska. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 170:26.5-278. 1977b A new digital hydroacoustic data processor and some observations on herring in Alaska. ,1. Fish. Res. Board Can. .34:2288-2294. Uzmann, J.R.. R.A. Cooper, R.B. Theroux, and R.L. Wigley 1977 Synoptic comparison of three sampling techniques for estimating abundance and distribution of selected megafauna: Submersible VS camera sled VS otter trawl. Mar. Fish. Rev. ,39(12):11-19. Villoso, E.P,, and G.V, Hermosa Jr. 1982 Demersal trawl fish resources of Samar Sea and Carigara Bay, Philippines. Fish. Res. ,J. Philipp. 7:59-78. Wantiez, L., and M. Kulbicki In press Les peches exploratoires au chalut en baie de Saint- Vincent. Rapp. Sci. Tech. 55, ORSTOM Oceanogr., Noumea, New Caledonia [in French, Engl, abstr. |. Abstract.— Relationships between mussel shell j^rowth and environmen- tal parameters were investigated in the mussel Mytilus edulis at an in- shore location, Avila Beach, and at an offshore location, oil platform Holly (ARCO). Temporal patterns of mussel growth were similar at both locations. Mussel growth rate was related to chlorophyll a concentra- tion at Holly, but not at Avila. Theo- retical estimates of scope for growth (SFG) were made for mussels at each location using published physiologi- cal data. Good agreement was found, with a time lag, between estimated SFG and shell growth. The SFG anal- ysis independently supported the con- clusion that temporal changes in phytoplankton concentration limits mussel growth at HoUy, but suggested that changes in the composition, rather than the concentration, of suspended particulates limits growth at Avila, as reported for mussels in estuarine environments. Food Availability as a Limiting Factor to Mussel Mytilus edulis Growth in California Coastal Waters Henry M. Page Yann O. Ricard Marine Science Institute, University of California Santa Barbara, California 93106 Manuscript accepted 16 May 1990. Fishery Bulletin. U.S. 88:677-686. Temporal variability in growth rate has been extensively documented for many species of filter-feeding marine invertebrates. Growth rates frequent- ly vary "seasonally," with most rapid growth occurring during the spring and summer months. Season is an ambiguous concept, however, which does not satisfactorily describe fac- tors regulating temporal patterns of growth. Ultimately, environmental factors, which vary over time and with location, contribute to variation in growth rates. The growth rate of mussels Mytilus edulis varies in both time and space. Mussel growth rates near Santa Barbara, California, are highest from May through August (Harger 1970; Page and Hubbard 1987), but elevated rates can also oc- cur during the winter months (Page and Hubbard 1987). Physiological and ecological evidence indicates that in many situations worldwide, food availability may be the most impor- tant single factor regulating mussel growth (Seed 1976, Widdows et al. 1979, Incze et al. 1980, Rodhouse et al. 1984). Multiple regression and cor- relation analysis indicated that mus- sel growth rate was associated with phytoplankton abundance, but not water temperature, at an offshore location in the Santa Barbara Chan- nel (Page and Hubbard 1987). Variation in the concentration and composition of phytoplankton and other suspended particulates which CDuId influence mussel growth exists in the open coastal environment. For example, episodic upwelling and high primary productivity characterize the region north of Point Conception, California, relative to the Santa Bar- bara Channel (Owen 1980, Willason et al. 1986). Inshore areas tend to be more productive and to possess high- er total seston concentrations than offshore areas, and phytoplankton concentration varies with depth (Raymont 1980). Little information is available regarding spatial relation- ships between phytoplankton abun- dance and mussel growth in inshore waters. In this study, we used correlation analysis and the "scope for growth" concept to evaluate the potential importance of temporal and spatial variation in food availability to mus- sel growth. The concept of scope for growth (SFG, Warren and Davis 1967), as applied to mussels, has been reviewed by Bayne et al. (1976a) and Widdows (1985a). SFG analysis uses physiological relationships, together with environmental parameters, to estimate the potential production of soft tissue (soma and gonad) by mus- sels from the general energy equa- tion. SFG = A - (R -h U), where SFG = energy available for growth of soft tissue, A = energy absorbed from food, R = respiratory heat loss, and U = energy lost as excreta. Radford and Bayne (cited in Bayne et al. 1976a) and Radford et al. (1981) suc- cessfully used this concept to model 677 678 Fishery Bulletin 88(4), 1990 mussel growth in British waters. SFG analysis may thus prove useful in interpreting relationships between mussel growth and environmental factors in Califor- nia waters. The present study, which continued work reported in Page and Hubbard (1987), (1) compared mussel growth rates in highly productive inshore coastal waters north of Point Conception with rates measured concurrently at offshore Platform Holly (Atlantic Rich- field Company), (2) evaluated the relationship between temporal and spatial variation in growth and measure- ments of potential food availability, and (3) used pub- lished physiological data and the "scope for growth" concept (Bayne et al. 1976a) to provide an independent assessment of the response of mussel growth to envi- ronmental conditions in California waters. Materials and methods Study sites Avila Beach (35°10'N, 120°43'W) is located appro.x- imately 84 km north of Point Conception, California (Fig. 1), in a region characterized by episodic upwell- ing and high primary productivity (Owen 1974, 1980; Lasker et al. 1981). The study site was located at the end of the Unocal pier which extended 0.8 km into semiprotected Avila Bay. Water depth at this location was about 12 m. Mytilus edulis were collected inter- tidally on pier pilings, since subtidal mussels were scarce due possibly to starfish predation (Landenberger 1967). Platform Holly is a 20-year-old oil and gas pi-oduc- tion platform located in 60 m of water, about 3 km off- shore of Goleta, California (34°25'N, 119°52'W; Fig. 1). Mussels colonize the support members of this plat- form from the intertidal zone to depths greater than 18 m, and grow rapidly, achieving 50 mm shell-length in 6-8 months (Page and Hubbard 1987). Physical and biological parameters We collected water samples at a depth of 2 m with a Van Doren bottle every 7-10 days at each location from October 1986 to June 1987. Temperature of samples was measured by hand-held thermometer. Estimates of potential food available to mussels were made from the concentrations of seston, particulate organic mat- ter (POM), chlorophyll a (an estimate of phytoplank- ton biomass: Lorenzen 1970, Hunter and Laws 1981), and particulate organic carbon (POC) in replicate 500-mL water samples. All water samples were pre- filtered through a 300-;.(m nylon mesh. Seston concen- tration was measured using standard methods (Wid- I Morro Bay _Avila Beach Cajifomla Pacific Ocean Poinl Conception Platform Holly Figure 1 Locatidns (if the Avila Beach and I'latfdfiii Holly study sites. dows 1985b). Percent of POM (% POM) witliin the seston was calculated as 100 ■ [POM]/[seston]. Chloro- phyll a concentration was determined using standard fluorometric methods (Parsons et al. 1984) and a Turner Designs Fluorometer. Particulate organic car- bon concentration was determined using a Perkin- Elmer CHN analyzer following the methods of Rod- house et al. (1984). Growth rate of Mytilus edulis Twenty individually numbered mussels (~20 mm shell- length), collected intertidally (~ - 15 cm), were enclosed in a cylindrical vexar plastic cage (12 x 20 cm, 5-mm mesh) monthly at each site from November 1986 to June 1987. Using calipers, shell-lengths of caged mus- sels were measured to the nearest 0.1 mm initially and after 4 weeks. No growth data are available at Avila from January and February 1987, due to lack of small mussels. Cages were submerged at depth (-2m) either by suspension on a weighted line (Holly) or by attach- ment to a pole (Avila). Cages were cleaned as needed to keep fouling to a minimum. Transplant experiments We conducted transplant experiments in Fall 1986 to evaluate the potential influence of mussel stock on growth rate. Forty M. rdulls were collected from Holly on 27 September 1986 and transplanted (2 cages of 20 individuals of ~30 mm shell length) to Avila on 29 September 1986. Mussels of equal size and numlicr were collected from Avila on 29 September 1986 and transplanted to Holly on 4 October 1986. Mussels were covered with a moist cloth in transport and maintained in unfiltered seawater prior to placement in the field. Mention of trade names does not imply emlor.semeiit by the National Marine Fisheries Service. N().\A. Page and Ricard: Food availability and Mytilus edulis growth in California coastal waters 679 Table 1 Physiological parameters and relationships used in SFG analysis u { Mytilus eiJiilis. W = soft tissue dry \vt. (g), L = shell length (mm). [seston] = seston concentration (mg/L), T° = water temperature (°C). Regulatory Physiological parameter factor(s) Comments Relationship Source 1) Clearance rate W Independent of tem- CR = 1.73 ■ W""'* - 0.006 • [seston] a) Thompson 1984 (CR, L/h) [seston] perature (a,b,c). ANCOVA adjusted common slope and mean of intercepts of CR vs. W curves from (a) combined with median slope of CR vs. [seston] rela- tionship for three size classes in (b). if CR<0.1 L/h, then CR = 0.1 L/h b) Widdows et al. 1979 c) Widdows 1978 2) Seston filtered CR Assume 100% reten- SF = CR ■ [seston] (SF, mg/h) [seston] tion efficiency. 3a) Pseudofeces threshold W Shape of pseudofeces T = 3.81 ■ log L- 1.93 d) Foster-Smith 1975 (T, mg/L) vs. food concentra- tion curve of (d) b) Widdows et al. 1979 e) Bayne and Worrall 1980 3b) %Pseudofeces (%Ps) T combined with size- %Ps= 100 -[(86.4 e"-*"''^) [seston] specific threshold values measured by (b). Technique alluded to in (e). ■ [seston] -'"•'■^■"" '■■"''] 4) Seston ingested SF SI = SF-(SF %Ps/100) (SI, mg/h) %Ps .5) Absorption efficiency %POM Model of (f). AE = 0.5 ■ log(%POM) - 0.32 f) Bayne et al. 1979 (AE) (i) POM absorbed SI A = (SI ■ %POM/ 100) ■ (AE/ 100) (A. mg/h) AE 71 Respiratory rate W b value determined VO , = a ■ W' a) Thompson 1984 (VO., mhOJh) rpo from average of all monthly regressions in (a). Temperature effect described by exponential curve fitted to data from (a) and (g). VO^ con- sidered independent of ration at [seston] >2 mg/L. b = 0.782 a = 0.117- (10"""') g) Widdows et al. 1984 8) Scope for growth A SFG = (A • 23. .5) - (V0_, ■ 20.3) (SFC, ,I/h) VO, Growth rates of transplanted individuals were com- pared with that of 40 resident individuals enclosed in cages (2 cages of 20 individuals of ~30 mm shell length) and submerged at the same time. Scope for growth Relevant physiological parameters, regulatory factors, and relationships used in our analysis of SFG are given in Table 1. Environmental parameters required to estimate SFG (water temperature, seston concentra- tion, and %POM) were measured in this study, while physiological relationships were derived from informa- tion in the literature. Absorbed food was estimated from measurements of mussel size, seston concentration, %POM, and from published values for size-specific clearance rate, inges- tion rate and absorption efficiency. Wlienever possible, we used data on clearance and ingestion rates and 680 Fishery Bulletin 1990 absorption efficiences obtained using "natural" POM. Clearance and ingestion rates of particulate material were assumed to be independent of water temperature, but dependent on mussel size and particle concentra- tion (Foster-Smith 1975, Bayne et al. 1976b, Widdows 1978, Widdows et al. 1979). The fraction of filtered seston rejected as pseudofeces (% pseudofeces. Table 1) was determined from ingestion rate and the critical POM concentration (3.0 mg/L for a 20-mm mussel; Widdows et al. 1979). The percent pseudofeces repre- sents the fraction of filtered seston rejected as pseudo- feces. Absorption efficiency was estimated using the model of Bayne et al. (1979; Table 1). The caloric con- tent of the absorbed food was assumed to be 23.5 Joules/mg dry POM (Widdows 1985b). Metabolic expenditures were estimated given mussel size, water temperature, and published data on oxy- gen consumption (Thompson 1984, Widdows et al. 1984; Table 1). Published oxygen consumption mea- surements were converted to an energetic ecjuivalent of 20.3 Joules/mL O2 (Crisp 1971). Energy losses due to excretion are generally minor (as a percent of ab- sorbed ration: 4.3-5.9%, Bayne et al. 1979; 1.7-4.3%, Widdows et al. 1980; 0.4-2.0%, Thompson 1984; 0.5-2.9%, Widdows and Shick 1985) and were ignored here. SFG (Joules/hour) was estimated as (absorbed POM, mg/hour x 23.5 J/mg) - (mL O./hour x 20.3 J/mL 0.>). a 18 DCKD D 17 ■r\\ 9 / h /\ 0 16 ^ K t? v t A \ • 1 0) 15 \A/\n \ A fh9-y»o 3 • • \/ \ \ /» 1 1 V y \ V y A J ' L re 14 * ^ V vA D 9 ^ k / K^ 1 1 0 ^ i r^ 1 / a. 13 *-** \ p/ > / E / \ F / 0 12 - V \ f t- 11. 1 C 1 1 i 1 i 4 ) N D J F M A M J ~^ b. 12 "T ^ 10 \l O) \ \ i \ v \i m rei 8 ■ l\ >- sz 6 Q. 0 1 0 4 \ \ * 1 .c \ V M !? k 0 2 wV^ M h ^0 N D J F M A M J ~!\ 1986 Month 1987 Figure 2 (a) Water temperatures and (b) chlorophyll ti concen- trations at Avila Beach (•) and Platform Holly (O). Results Physical and biological parameters Surface water temperature was significantly lower at Avila than at Holly (/ = 2.46, df = 49, P<0.01, Student's /-test; Fig. 2a). Water temperatures at Avila were up to 3°C cooler than at Holly from October through December 1986. Surface chlorophyll (/ concentration was significant- ly higher at Avila "than at Holly (( = 3.53, df=49, P < 0.001, Student's Mest; Fig. 2b). Chlorophyll a con- centrations were as much as 20 times higher at Avila than at Holly from October through November 1986. Seston concentration was higher at Avila than at Holly (/ = 3.88, df = 46, P<0.001; Fig. 3a). The highest values at Avila (10-30 mg/L) January through Febru- ary coincided with seasonal storms. Seston levels re- mained low at Holly, fluctuating between 1 and 5 mg/L during most of the year. Particulate organic matter concentration was higher at Avila than at Holly (/ = 2.04, df = 40, P<0.05; Fig. 3b). Particulate organic matter concentrations at Avila generally varied between 1 and 5 mg/L, but reached 10 mg/L in January 1987. Particulate organic matter concentrations at Holly ranged from <0.5 mg/L to a high value of 4 mg/L during a phytoplankton bloom in March 1987. Differences in percent POM between the two locations were most evident in January and Feb- ruary 1987, when percent POM was 10-35%) at Avila and 40-70% at Holly (Fig. 3c). Inarticulate organic carbon concentration was sig- nificantly higher at Avila than at Holly (/ = 4.88, df = 38, P< 0.001; Fig. 4). Values at Avila ranged from ~400 ^g C/L to ~1400 ^Jg C/L, with values exceeding 1000 /.(g C/L in late November 1986, and January, April, and June 1987. In contrast, POC concentrations at Holly remained below 500 f^g C/L, except during March- April 1987 when values of 650-900 /.(g C/L were recorded. The peaks in POC concentration at Avila coincided with elevated chlorophyll a concentrations in November, April, and June, but not in January. The peaks in POC concentration at Holly coincided with the elevated chlorophyll n concentrations in March-April. The slopes of linear least-square regression lines of POC concentration versus chlorophyll a concentration at Avila and Holly were significantly different from 0 (P<0.05, ANOVA). There was no significant difference Page and Ricard: Food availability and M\4ilus edulis growth in California coastal waters 681 'N D MAM M A M J 1987 Month Figure 3 (a) Seston concentrations, (b) particulate organic matter (POM) concentrations, and (c) %POM at Avila Beach ( • ) and Platform Holly (O). in the slopes of these lines (P>0.1, ANCOVA; Fig. 5). However, a significant difference in the location of the ^/-intercepts (F = 13.46, df = 1, 34, P<0.001, ANCOVA) indicated that a substantially higher concentration of "background" POC existed at Avila (623 Mg C/L) than at Holly (264 f^g C/L). Covariance among environmental parameters We found weak but significant positive correlations between chlorophyll a concentration and both water temperature and POC concentration at Avila (Table 2). Figure 4 Particulate organic carbon (POC) concentrations at Avila Beach ( • I and Platform Holly (O). 1400 • ^ 1200 Avila ^^^^'^^ o o Q. 1000 800 600 400 • o j^^::^ 0 ^^^^ ^^ ^^Ji o" \^^ 200 ^0, 0 c 12 4 6 8 Chlorophyll a (ng/l) 10 Figure 5 Linear regressions of particulate organic carbon (POC) concentra- tion on chlorophyll « concentrations at Avila Beach (•) and Plat- form Holly (6). Avila; (/ = 69.8x -i- 623. >- = 0.51; Holly: y = 184x + 264. /• = 0.60. In contrast, we found a weak negative correlation between chlorophyll a concentration and water tem- perature at Holly. Chlorophyll a concentration was positively correlated with both POC and POM concen- tration at Holly. There was no correlation between chlorophyll a concentration and total seston or percent POM, or between POM and POC concentration at either site (Table 2). Growth rate Shell growth rates for a mussel of 20 mm shell length over a period of 1 month at Avila and at Holly are given in Figure 6. Growth rate was temporally variable at both locations, with slowest growth rate December- March 1987 (5-7 mm/mo) and most rapid growth 682 Fishery Bulletin 88(4). 1990 Table 2 Correlation coefficients { ;■) between fact ors at ins lore (Avila Beach) and offshore (oil jlatform Holly) ocations calculated from least-squares linear regression analysis. *P<0.05, ••p Beach and oil platform Holly. Mean values eduiis-dt Avila ±1 SD. Treatment Date Initial length (mm) Growth rate (mm/mo) Avila-Holly 8 Nov. 86 29.9 + 2.7 30.4 ± 2.9 8.6 + 3.1 8.0 ± 2.2 Holly 8 Nov. 86 30.0 ± 2.9 29.2 + 3.1 8.1 ± 2.0 7.4 ± 2.0 Holly-Avila 5 Nov. 86 33.3 + 3.4 35.7 ±4.2 7.5 ± 1.9 7.4 ± 1.8 Avila 5 Nov. 86 33.6 ± 2.7 31.6 ±3.4 7.6 ± 1.3 7.8 ± 2.3 10 (mm/mo) o o • 0 ° • S 7 5 • o • £ 6 • o o . o • c ) 2 4 6 8 10 Chlorophyll a (ng/l) Figure 7 Relationship between Mytilus edulis growth rate and chlorophyll a concentration with a time lag of 3 weeks at Avila Beach (•) and Platform Holly (O). October-November 1986 and May- July 1987 (8-9 mm/ mo). There was no effect of mussel stock on growth rate (F = 0.901. P>0.1, 1-way ANOVA; Table 3). Mor- tality of caged mussels was <5.0%. no correlation between growth rate and water tem- perature (P>0.05). Relationship between growth rate and chlorophyll a concentration We calculated correlation coefficients relating the growth rate of M. edulis to chlorophyll a concentra- tion and water temperature for grouped Avila and Holly data integi'ated over the 4-week period of mussel exposure and at time lags of 0-4 weeks. There was no correlation between growth rate and chlorophyll a concentration (/'>0.05, n = 14). However, growth rate correlated v«th chlorophyll a concentration with a time lag of 3 weeks, if the fall and winter values from Avila (when seston concentrations exceeded 5 mg/L) were excluded (r = 0.67, P<0.05, n = 11; Fig. 7). There was Scope for growth A summary of our theoretical scope-for-growth calcu- lations, averaged by irionth, is given in Table 4. Our analysis indicates that, overall, the estimated amount of seston filtered from suspension liy mussels was higher at Avila than at Holly (4.00 vs. 1.91 mg/h, t = 4.03, df = 46, P<0.001). However, mussels at Avila had a significantly higher estimated percent pseudofeces production (48.6% vs. 13.7%, ^ = 6.34, df=46, P< 0.001) than mussels at Hully. Thus despite the higher seston concentrations at Avila, the actual amount of estimated POM assimilated by mussels at each local- ity should have been similar (0.43 mg/h at Holly vs. 0.36 mg/h at Avila, P>0.01). Page and Ricard' Food availability and Mytilus eclulis growth in California coastal waters 683 Table 4 Summary of scope-for-growth calculat ons for a 20-mm shell length (0.05 g) Mytilus ei luli!< averaged by month. POM = = particulate organic matter. Clearance Seston Pseudofeces Seston Absorption POM rate filtered threshold % ingested efficiency absorbed VO. Month (L/h) (mg/h) (mg/L) pseudofeces (mg/h) (%) (mg/h) (mL 0,/h) Avila Beach Nov. 0.45 3.55 3.0 50.62 1.76 42.40 0.23 0.06 Dec. 0.46 3.33 3.0 48.15 1.74 45.38 0.32 0.05 Jan. 0.36 8.03 3.0 74.71 2.26 39.04 0.27 0.04 Feb. 0.42 5.70 3.0 64.59 2.06 27.32 0.31 0.04 Mar. 0.47 3.48 3.0 28.87 2.32 54.94 0.70 0.04 Apr. 0.47 2 22 3.0 23.98 1.72 54.72 0.48 0.04 May 0.46 2.32 3.0 43.21 1.34 54.64 0.43 0.05 June 0.45 3.30 3.0 51.57 1.59 45.88 0.29 0.05 Oil platform Holly Nov. 0.48 2.02 3.0 18.58 1.54 42.49 0.24 0.07 Dec. 0.48 1.06 3.0 28.28 1.53 37.29 0.15 0.07 Jan. 0.48 1.58 3.0 0.00 1.58 48.29 0.31 0.05 Feb. 0.48 1.40 3.0 0.00 1.40 55.01 0.48 0.05 Mar. 0,48 1.92 3.0 8.44 1.72 55.90 0.58 0.04 Apr. 0.48 1.42 3.0 0.00 1.42 54.59 0.44 0.05 May 0.48 2.20 3.0 27.12 1.52 60.19 0.65 0.05 June 0.45 3.91 3.0 55.11 1.76 51.65 0.43 0.05 20 .£= ^ 16 t T SI I o 12 , / If >^ h o 8 / 1 \ 0) Q. O 4 0 f4^ K, ' -\ 1 N D >. F M A M J 1986 Month 1987 Figure 8 Estimates of monthly scupe-for-growth of a 0.05-g dry weight (~20 mm shell length) MiilHus edidift at Avila Beach ( • ) and Platform Holly (O). Mean values ±1SE. Figure 8 shows the monthly SFG of a 0.05 g dry weight M. eduli>i (~20 mm shell length) at Holly and at Avila, calculated using the seston concentration and percent POM data from these localities and the equa- tions in Table 1. There was a significant effect of month, but not location, on scope for growth (2-way ANOVA, Table 5). An a posteriori test for significant Table 5 Results of two-way ANOVA evaluating the influence of month and location of estimated scope for growth oi Mytilus edulis. ***P<0.001. Source Sum of square df Mean square F ratio Location Month Month ■ Location Error 16.09 557.95 190.31 640.51 1 7 7 33 16.09 79.71 27.19 19.41 0.83 4.11*" 1.40 differences among months revealed that the low SFG values of November- February differed significantly from the higher SFG values of March- June (F = 4.11, df=l, 33, P<0.001). To evaluate the relationship between shell growth rate and theoretical SFG, we calculated linear corre- lation coefficients between growth rate (Fig. 6) and scope for growth for the combined Avila and Holly data at time lags of 0-4 weeks. Growth was not correlated with SFG at a time lag of 0 weeks (/• = 0.56, P>0.05). However, growth rate correlated with SFG at time lags of 1-4 weeks with strongest correlations at time lags of 2 and 3 weeks {>■ = 0.70 and 0.75, P<0.05, u = 10; Fig. 9). 684 Fishery Bulletin 88(4). 1990 10 o" 9 E 0 , '^ o ^^^-'^ E 8 ^ 7 a> 2 6 .a % 5 O O 4 ^-'"''^ o 1 C 2 4 6 8 10 12 14 16 18 Scope for growth (J/h) Figure 9 Linear regression of Mytilus i-dutis shell growth rate on scope for growth for combined data from Avila Beach (•) and Platform Holly (O) with a time lag of 3 weeks, y = 0.22x + 4.96, r =0.75. Discussion Spatial variation in phytoplankton biomass may affect the growth and nutritional condition of filter-feeding species along the California coast (larval fish Engraulix niordax, Lasker and Smith 1977, O'Connell 1980; cope- pod Calanus pacificus, Willason et al. 1986; anomuran crab Emerita analnga, Dugan and Wenner 1985). We found that mussel gr'owth increased with chlorophyll a concentration, except during the fall and winter months at Avila. The low growth rates at Avila during fall and winter, despite chlorophyll a concentrations exceeding 4 fig/L, may reflect a leveling-off or decline in particle ingestion rates associated with high seston concentra- tions (Foster-Smith 1975, Widdows et al. 1979). Temporal patterns of growth were similar for mus- sels at Avila and Holly (Fig. 6). These data suggested that ingestion rate was not appreciably higher for- mus- sels at Avila, despite the higher POC concentrations there (generally >600 fig/L). In addition, both the high background POC concentration (623 ^^glL, Fig. 5) and the lack of correlation between chlorophyll a and POM concentrations at Avila indicated the presence of a high concenti'ation of nonphytoplankton particulates (e.g., bacteria, microzooplankton, detritus) which may not support rapid mussel gi'owth. MytUus edulis has shown poor growth when supplied only with nonphytoplank- ton food sources in laboratory experiments (Winter 1974, Williams 1981). The lack of a correlation between growth rate and water temperature for grouped data from both loca- tions is consistent with the view that water tempera- ture is not an important factor influencing mussel growth in California waters (Page and Hubbard 1987). The lack of a stock or genotypic effect on mussel growth rate in our transplant experiments was not sur'prising, as the distance between the study sites was only about 120 km. Mytilus edulis has a planktonic larval stage of about 3 weeks, and typical current velocities of 0.5 km/hour (Chambers Group 1986) would permit larvae to drift as much as 250 km. Mytilus edulis also has the potential to delay metamorphosis and to exist as a pediveliger in the plankton for several days (Bayne 1964), facilitating genetic exchange be- tween spatially separated populations. The transplant experiments also indicated that postsettlement selec- tion (Koehn and Hilbish 1987), which might result in differences in growth rate between locations, was not an important factor in this study. Our field growth-rate data generally conform to predictions from the theoretical SFG analysis. This analysis suggested that mussels at Avila and Holly absorbed similar amounts of POM and had similar SFG because seston concentrations were high at Avila (>4 mg/L) and mussels at this location had a higher rate of pseudofeces production and lower absorption effi- ciency than mussels at Holly. Shell growth rate corre- lated with theoretical SFG after incorporation of a time lag. SFG is a measure of the energy instantaneously available for the gi'owth of soft tissue and shell. In small mussels, shell growth rate is correlated with soft tissue growth (Nielsen 1985). The time lag likely reflects the time required for metabolic conversion of absorbed energy and nutrients into the shell. The relationships between potential food availabil- ity and mussel growth at Avila and Holly agree with gener'al predictions developed pr'imarily tVom physio- logical studies of A/, edulis in British estuarine envi- ronments where the energy available for growth was 1'egT.ilated by the food quality, reflected by the percent POM, rather than by quantity of the seston when seston concentrations exceeded 4-5 mg/L (Bayne and Widdows 1978, Widdows et al. 1979). The growth rate of mussels in California coastal waters, with seston concentrations comparable to those at Avila, is thus likely limited by the quality rather than the quantity of the seston. Acll< igical methods for seawater analysis. Pergamon Press, NY, 173 p. Radford, P.J,, LR. Joint, and A.R. Hiley 1981 Simulation models of individual production processes. /)/ Longhurst, A.R. (ed.), Analysis of marine ecosystems, p. 677-670. Academic Press. NY. Raymont. J. E.G. 1980 Plankton and productivity in the oceans. Vol. I. Phyto- plankton. 2(1 ed. Pergamon Press. NY, 489 p. Rodhouse, P.G., CM. Roden, G.M. Burnell. M.P. Hensey, T. McMahon, B. Ottway, and T.H. Ryan 1984 Food resource, gametogenesis and growth of Mytilus edulis on the shore and in suspended culture: Killary Harbour, Ireland. ,1. Mar. Biol. Assoc. U.K. 64:513-529. Seed. R. 1976 Ecology. lu Bayne, B.L. (ed.). Marine mussels: Their ecology and physiology, p. 13-66. Cambridge Univ. Press, Cambridge. Thompson, R.J. 1984 The reproductive cycle and physiological ecology of the nuissel Mytilus edulis in a subarctic, non-estuarine environ- ment. Mar. Biol. (Berl.) 79:277-288. Warren, CE., and G.E. Davis 1967 Laboratory studies on the feeding bioenergetics and growth of fishes. In Gerking, S.D. (ed.). The biological basis of freshwater fish production, p. 175-214. Blackwell Sci. Publ.. Oxford. Widdows, J. 1978 Combined effects of body size, food concentration and season on the physiology of Mytilus edulis. J. Mar. Biol. Assoc. U.K. .58:125-142. 1985a Physiological measurements. In Bayne. B.L. (ed.). The effects of stress and pollution on marine animals, p. 3-45. Praeger. NY. 686 Fisdeiy Bulletin 88(4), 1990 1985b Physiological procedures. In BajTie. B.L. (ed.), The ef- fects of stress and pollution on marine animals, p. 161-178. Praeger. NY. Widdows, J., and J.M. Shick 1985 Physiological responses oiMytilus eduiis and Cardium fdiile to aerial exposure. Mar. Biol. (Berl.) 85:217-232. Widdows. J. P., P. Fieth, and CM. Worrall 1979 Relationships between seston. available food and feeding activity in the common mussel Mytiluf: eduiis. Mar. Biol. (Berl.) .50:19.5-2(17. Widdows, J.. D.K. Phelps, and W. Galloway 1980 Measurements of physiological condition of mussels trans- planted along a pollution gradient in Narragansett Bay. Mar. Environ. Res. 4:181-194. Widdows. J., P. Donkin. P.M. Salkeld, J.J. Cleary. D.M. Lowe, S.W. Evans, and P.E. Thomson 1984 Relative miportance of environmental factors m deter- mining physiological differences between two populations of mussels (Miittlin^ eduiis). Mar. Ecol. Prog. Ser. 17:33-47. Willason. S.W., J. Favuzzi, and J.L. Cox 1986 Patchiness and nutritional condition of zooplankton in the California Current. Fish. Bull. U.S. 84:157-176. Williams, P. 1981 Detritus utilization by Mytilus eduU.<. Estuarine Coastal Shelf Sei. 12:739-746. Winter, J.E. 1974 Growth in Mytilus eduiis using different types of food. Ber. Dtsch. Wiss. Komm. Meeresforsch. 23:360-375. Abstract.— significant changes in the hiomass of sandlanee Amrno- (Ij/tea spp. and in the abundance of the copepod Calanus finniarchicus in the southern Gulf of Maine co-oc- curred with a shift in the occurrence and abundance of four species of ba- leen whales in the region. During the years 1982-88 the abundance of sandlanee was negatively correlated to the abundance of C finniarchicus (r, = -0.883, F<0.05). Peak years of abundance for C. finmarchicus during 1982-88 in the study area were the lowest years of abundance for sandlanee. The abundance of C. finmarchicus and sandlanee was at a regional maximum during 1986 and 1988, respectively. The abuniiance of humpljaek and fin whales were marginally corre- lated to each other (r, = 0.3338, P< 0.08). The abundance of humpbacks was negatively correlated with right whales (r, = -0.7753, P<0.001) and sei whales (r,= -0.5507, P<0.01). The patterns of occurrence for right and sei whales were significantly re- lated to each other (r, = 0.6842i P< 0.001). Right and sei whales were common in the region only during 1986, when copepod abundance reached a regional maximum and sandlanee abundance a regional minimum. These patterns of whale occurrence reflect known prey pref- erences, and are therefore expected between the piseiverous humpback and fin whales and the highly plank- tiverous right and sei whales. We hypothesize that the spatial distribution and abundance of baleen whales in the Gulf of Maine can be characterized as a series of ecological responses to human-induced changes in the abundance of planktiverous finfish. Recent Fluctuations in the Abundance of Baleen Whales in the Southern Gulf of Maine in Relation to Changes in Selected Prey p. Michael Payne Marine Mammal and Seabird Studies, Manomet Bird Observatory Box 936. Manomet, Massachusetts 02345 David l\i. Wiley Sharon B. Young Plymouth Marine Mammal Research Center P O Box 3313, Plymouth, Massachusetts 02361 Sharon Pittman Phillip J. Clapham Cetacean Research Program, Center for Coastal Studies PO Box 1036, Provincetown, Massachusetts 02657 Jack W. Jossi Northeast Fisheries Science Center, National Marine Fisheries Service, NOAA South Ferry Road, Narragansett, Rhode Island 02882 During the mid-1970s, a dramatic in- crease in the abundance of sandlanee Amm-odytes spp.* precipitated a sig- nificant change in the abundance and composition of the ichthyofauna in the shelf waters of the northeastern United States (Smith et al. 1978, 1980; Morse 1982). The population explosion of sandlanee coincided with a 50% reduction in total finfish bio- mass in the same region between 1968 and 1975 (Clark and Brown 1977). This decrease was primarily due to the commercial depletion of stocks of herring Clupea harengus and mackerel Scomber scomhrus (An- thony and Waring 1980, Grosslein et al. 1980). A concurrent increase in sandlanee abundance following the depletion of North Sea herring and mackerel stocks led Sherman et al. (1981) to suggest that sandlanee had taken over ecological niches previous- Manuscript accepted 3(1 May I'.tOO. Fishery Bulletin. U.S. 88:687-696. • Two species of this genus, A. americanux and .4. duhiii^, occur in the shelf waters of the northeastern United States (Richards et al. 1963, Reay 1970. Richards and Kendall 1973, Meyer et al. 1979, Richards 1982). ly occupied by these species in many areas of the Northwest Atlantic. Since the mid-1970s sandlanee have become increasingly important in the Gulf of Maine as prey for com- mercial fish (Bowman et al. 1984), seabirds (Powers and Backus 1987), pinnipeds (Payne and Selzer 1989). and baleen whales (Overholtz and Nicolas 1979, Hain et al. 1982, Payne et al. 1986). Sandlanee were the only confirmed prey of the humpback whale Megaptera novaeangliae be- tween 1975 and 1979 (Hain et al. 1982, Mayo 1982) and the only prey significantly correlated with the dis- tribution of humpbacks in the Gulf of Maine between 1978 and 1982 (Payne et al. 1986). Also, fin whales Biiliienoptera physalus, sympatric with humpbacks in this region, have frequently been observed exploiting sandlanee (Overholtz and Nicolas 1979). Two other species of large baleen whales, the northern right whale Eubalaena glacialis and the sei whale Balaenopfcro horealis. also occur in 687 688 Fishery Bulletin 88(4), 1990 the Gulf of Maine (CeTAP 1982, Watkins and Schevill 1982, Kraus 1985, Schevill et al. 1986, Weinrich et al. 1986, Hamilton and Mayo 1990). Both of these species have been characterized principally as "skim feeders" (Nemoto 1959, Kawamura 1974, Nemoto and Kaw^a- mura 1977) which subsist primarily on dense swarms of calanoid copepods, notably Calanusfinmarchicus in the North Atlantic (Mitchefl 1975a, 1975b; Jonsgard and Darling 1977; Mitchell and Chapman 1977; Winn et al. 1986; Wishner et al. 1988; Mayo and Marx 1990), in preference to schooling fish (Watkins and Schevill 1979). This similarity in prey preference led Mitchell (1975a) to hypothesize that competition between these two whale species might have adversely affected the recov- ery of the right whale in the western North Atlantic. However, sandlance are also known to play an impor- tant trophic role as a major predator of copepods, especially C. finmarchicus (Monteleone and Peterson 1986, Meyer et al. 1979). Given their abundance and prey selectivity towards copepods, Kenney et al. (1986) proposed that sandlance represented a more significant competitor to right whales for copepodite prey than sei whales (as suggested by Mitchell 1975a). The existence of some degree of intra- and inter- specific competition between right/sei whales and sand- lance (ecologically similar species) for a common prey is generally accepted, but is difficult to demonstrate. An overlap in preferred prey is at best a qualitative phenomenon, and offers only "soft corroboration" (from Strong et al. 1984) as evidence for competition between these species. Another approach (used by Smith 1968, 1970; Eck and Wells 1987), which we have adopted for the purpose of this study, is to look for a demonstrable shift in the ecological balance between competing species, or the complete replacement of one competing species by another. In this paper we ex- amine the ecological relationships between sandlance and two of its major predators, the humpback and fin whale; between sandlance and its major prey, the copepod Calanusfinmarchicus; and between sandlance and two potential competitors for C. finmarchicus in this region, the right and sei whale. Methods Study area In this paper we have focused on Stellwagen Bank (Fig. 1), a glacial deposit of sand and gravel in the southwest Gulf of Maine that is associated with high levels of biological productivity and a rapidly developing com- mercial whalewatching industry. The margins of the bank are defined by the 40-m isobath with depths as low as 18 m. Since the early 1980s whalewatching vessels have provided intensive survey coverage of cetaceans in the area from June through September of each year. Collection and treatment of data Fisheries data Data on the abundance of sandlance used here were collected on Stellwagen Bank and im- mediately adjacent waters by National Marine Fish- eries Service, Northeast Fisheries Center (NMFS/ NEFC) biologists and technicians during standard- ized NMFS/NEFC spring bottom-trawl surveys. The NMFS/NEFC survey area has been spatially stratified into approximate ecological units based principally on depth (see Grosslein 1969). Stellwagen Bank is a principal bathymetric feature in the Gulf of Maine- NMFS/NEFC stratum #26. The number of sandlance caught per tow in stratum #26 (for each spring survey) was transformed into logarithmic values, and the mean number of sandlance (transformed data) was con- sidered representative of Stellwagen Bank and used in all further analyses, a procedure which follows that described in Payne et al. (1986). The Calanusfinmarchicus data were also collected by NMFS/NEFC personnel. A transect across the Gulf of Maine has been sampled with the Hardy Continuous Plankton Recorder (CPR) since 1961 (Jossi and Smith 1990). Since the early 1980s a desired sampling fre- quency of one transect per month has been generally achieved. Data used in this study came from the 10- nautical-mile section of the transect centered on Stell- wagen Bank. Water passing through the CPR is filtered with bolting silk having mean aperture dimen- sions of 225 X 234 \jl. All large zooplankton (^2 mm) in the sample are identified and enumerated. Counts of smaller (< 2 mm) zooplankton are made from an aliquot (~l/45) of the sample. Counts were converted to num- ber of organisms per 100 m^, and then transformed to log base 10 for subsequent calculations. Further details of this procedure are given in Colebrook (1975) and in Jossi and Smith (1990). All reference to copepod data and abundance used in this study exclusively concern C. finmarchicus. The abundance of sandlance (expressed as the mean of the log-transformed number of sandlance per tow/ year) and of the copepod C. finmarchicus (log-trans- formed number of individuals/100 m-Vyear) were com- pared by first ranking the data by year, 1982-88 (A'' = 7), then calculating Spearman's coefficient of rank correlation (rj (Zar 1984). The data are therefore compared on an ordinal, rather than on an absolute, scale; this is because of the magnitude of the difference in the scale of the data that were brought together for this study. Payne et a\ Abundance fluctuations of baleen whales in the southern Gulf of Maine 689 70°00' W ATLANTIC OCEAN Figure 1 Baleen whale study area in southern Gulf of Maine (outlined) and geographical regions referred to throughout the text. Whale sighting data Data for this study on the oc- currence and distribution of humpback, fin, right, and sei whales were collected by naturalists working aboard commercial whalewatching vessels during the study period 1982-88. Data recorded from these vessels for all cetacean sightings include species, group size, time, location (using LORAN-C), behavior, and the photo- graphically confirmed identity of individual animals. Vessels ran 4-hour cruises to the Stellwagen Bank and Massachusetts Bay areas from two ports: Province- town and Plymouth, Massachusetts. The cruise track of the whalewatching vessels was decided by the cap- tain based upon the greatest number of whale sightings recorded on the previous trip or day; consequently, sampling was neither random nor systematic. None- theless, the intensity of the coverage resulted in ex- tremely comprehensive surveys of the region during most of the study period. The information collected aboard these vessels in this region represents one of the most detailed databases available anywhere on a multispecies group of whales. Data used in this study were taken from the first cruise of each day from each of the two ports and con- sidered a reliable index of relative abundance of whales in the area. The total number of whales observed dur- ing each morning whalewatching trip from each port was summed by species and month, June-September (the period when coverage from whalewatching vessels was effectively continuous). These totals were then divided by the number of trips each month (each trip was considered a unit of effort), resulting in an index of whale abundance, expressed as number of whales/ effort/month for each species between 1982 and 1988. While there are undoubtedly biases associated with the collection of these data, the intensity of coverage is such that this index can be used to accurately monitor relative changes in the local abundance and distribu- tion of whales. Using these sighting data, the temporal co-occur- rence between the four whale species within the study area was tested by ranking the mean number of whales/effort/month (by species), then calculating Spearman's coefficient of rank correlation (rj (Zar 1984) between species for all paired data, 1982-88 (A^ = 28 for each possible pairing). 690 Fishery Bulletin 88(4), 1990 06 S 05 A N D 04 k H N 03 C E / 02 T 0 W 0 1 6 5 C E ^ 0 D w 1 ^ / \ /^r -V'*^^^-»^^^ T"*"^^ \ / ^^^■^--^ / \/ 1982 1983 1984 1986 1986 1987 1988 YEAR Figure 2 Stratified mean number of sandlance (■) per tow (log transforma- tion) on left, and the number of copepods Calanusfiyonarchicxa ( + ) per m^ (log transformation) on right. Note: the scale for sandlance is different from that of copepods. Results Changes in abundance of sandlance and copepods The mean number of sandlance per tow (log transfor- mation) collected during spring bottom-trawl surveys ranged between years (Fig. 2) from a minimum 0.000 (in 1986) to a maximum 4.06(3 per tow (in 1988). Two trends were apparent. The data show a decline in catches from 1982 to the zero value of 1986. Catches increased in 1987, peaking in 1988 to 4.066 sandlance per tow (log value). The number of sandlance caught in 1988 was a regional abundance maximum for spring surveys, 1968-88 (from information presented in Nelson and Ross 1989). The log-transformed values of C. finmarchicuslm^ of water collected in the CPR ranged from 3.270 (in 1984) to 4.988 (in 1986), and generally increased on Stellwagen Bank between 1982 and 1986 (Fig. 2). The increase was most apparent between 1984 and 1986. The number of C. finmarchicus collected in 1986 was one order of magnitude greater than that collected dur- ing any other year of the study. Following 1986 the number of C. finmarchicus declined rapidly to 3.640 (log value) in 1987, then to 3.070 (log value) in 1988 (Fig. 2). The abundance of sandlance on Stellwagen Bank was inversely related to the abundance of C. finmarchicus (r, = -0.883, P<0.05, N=7). The three years when the spring tow values for sandlance were the lowest (1983, 1985, and 1986) were the three peak years for C. finmarchicus (Fig. 2). Conversely, the number of C. finmarchicus recorded on Stellwagen Bank during 1988 was one of the lowest throughout the study period. This is the year when the sandlance reached its maximum peak of abundance. Changes in whale abundance Humpback and fin whales showed a similar trend in abundance (r, = 0.3338, P<0.074, Table 2), although the variation in the observed number of humpbacks between years was much more pronounced than for fin whales. The number of humpbacks/year increased from 3.59 whales/effort in 1982 to 9.00 whales/effort in 1985 (Table 1, Fig. 3). During 1986 humpback abundance declined to the minimum value recorded during the study period of 0.23 whales/effort. A small increase in 1987 was followed by a much larger increase in 1988 to 6.98 whales/effort (Fig. 3). Less dramatic fluctuations between years were evi- dent in the number of fin whales from 1982 to 1988. The abundance of fin whales declined from the max- imum value of 3.71 whales/effort in 1984 to the low values of 1.9 and 1.4 whales/effort recorded in 1986 and 1987, respectively (Fig. 3). Although the numbers of fin whales/effort were lower in 1986-87 than in previous years, they did not change by more than an order of magnitude, as did other whale species. The number of fin whales increased to 3.61 whales/effort in 1988. The strongest positive correlation occurred between right and sei whales (r, = 0.6842, P<0.()001, Table 2). The right whale is a rare but regularly occurring spe- cies whose temporal occurrence in the study area has been generally restricted to the months of late-winter through early-spring (Winn et al. 1986, Brown and Winn 1989, Hamilton and Mayo 1990). This pattern was generally repeated in all years of the study period except 1986. Excluding 1986, there were only 12 right whale sightings from 1982 to 1988, and during most years there were no sightings (Table 1). However, in 1986, 174 right whale sightings (0.95 whales/effort) were recorded, and right whales were observed vir- tually every day throughout the summer (Hamilton and Mayo 1990). Sightings of sei whales on Stellwagen Bank, and in Cape Cod Bay, are very rare in any season (CeTAP 1982, Mayo et al. 1988). We did not observe sei whales in the study area between 1982 and 1985 (Table 1). However, their numbers increased significantly in 1986 (when they were commonly observed) to 0.26 whales/ effort (Table 1, Fig. 3). The occurrence of sei whales in 1986 was followed by an equally dramatic decline dur- ing 1987 (2 sightings, 0.01 sei whales/effort) and 1988 (no sightings). Between 1982 and 1988, the sei whale and right whale were abundant only during 1986 (Fig. 3), when the abundance of C. finmarchicus also reached a recorded regional maximum (Wishner et al. 1988). Payne et al : Abundance fluctuations of baleen whales in the southern Gulf of Maine 691 Table I Total number of whales counted, number of surveys (effort), and number of whales (by species) per survey in southern Gulf of Maine (in parentheses) for each month and year of the study. 1982-88. No. of whales (per survey) No of Year Month surveys Humpback Fin Right Sei 1982 June 27 90(3.33) 45(1.67) 0.00 0.00 July 31 228(7.35) 33(1.06) 0.00 0.00 Aug. 31 54(1.74) 227(7.32) 0.00 0.00 Sept. 27 4.5(1.67) 63(2.33) 4(0.15) 0.00 Total 116 417(3.59) 368(3.17) 4(0.03) 0.00 1983 June 46 229(4.98) 60(1.30) 0.00 0.00 July 43 266(6.19) 192(4.47) 0.00 0.00 Aug. 36 169(4.69) 172(4.78) 0.00 0.00 Sept. 28 96(3.43) 37(1.32) 0.00 0.00 Total 153 760(4.97) 461(3.01) 0.00 0.00 1984 June 43 317(7.37) 26(0.60) 0.00 0.00 July 45 382(8.49) 224(4.98) 0.00 0.00 Aug. 47 261(5.55) 232(4.94) 0.00 0.00 Sept. 33 133(4.03) 141(4.27) 0.00 0.00 Total 168 1093(6.51) 623(3.71) o.OO 0.00 1985 June 40 411(10.30) 143(3,58) 0.00 0.00 July 53 400(7.55) 122(2.30) 0.00 0.00 Aug. 46 654(14.21) 154(3.35) 0.00 0.00 Sept. 31 63(2.03) 24(0.78) 0.00 0.00 Total 170 1528(8.99) 443(2.61) 0.00 0.00 1986 June 44 16(0.36) 62(1.41) 2(0.05) 0.00 July 54 20(0.37) 136(2. .52) 6(0.11) 2(0.04) Aug. 51 4(0.08) 93(1.82) 44(0.86) 19(0.37) Sept. 34 2(0.05) 102(3.00) 122(3. ,58) 26(0.76) Total 183 42(0.23) 344(1.88) 174(0.95) 47(0.26) 1987 June 55 30(0.55) 56(1.02) 2(0.04) 0.00 July 61 36(0.59) 48(0.79) 1(0.02) 2(0.03) Aug. 55 76(1.38) 150(2.73) 0.00 0.00 Sept. 47 12(0.25) 52(1.10) 5(0.11) 0.00 Total 218 154(0.71) 306(1.40) 8(0.04) 2(0.01) 1988 June 55 199(3.62) 253(4.60) 0.00 0.00 July 62 535(8.63) 263(4.24) 0.00 0.00 Aug. 62 455(7.34) 160(2. .58) O.OII 0.00 Sept. 52 424(8.15) 159(3.06) (1.00 0.00 Total 231 1613(6.98) 83.5(3.61) 0.0(1 0.00 Table 2 Spearman Rank C(jrrelation Coefficient s(/- J (upper number) and statistical level of significant differen ce(P) (lower number) between the abundance (whales/effort) of four baleen whale species on Stellwagen Bank, 1982-88. (V = 28 for all paired tests. Humpback Fin Right Sei Humpliacli 1.0000 0.3338 -0.77.53 -0.5507 0.0000 0.0740 0.0001 0.0019 Fin 1.0000 -0.32,50 -0.1461 0.0000 0.0643 0.4248 Kight 1.0000 0.6842 (1.0000 (1.0001 Sei 1.0000 0.(1000 Several significant negative relationships also oc- curred between whale species (Table 2), most notice- ably between humpbacks and right whales (r, = -0.7753, P<0.0001), humpbacks and sei whales (r, = - 0.5507, P<0.0019), and fin whales and right whales (/•, = -0.3250, P<0.0643). This supports an intuitive interpretation of the data shown in Figure 3, notably the obvious phenomenon of low humpback and fin whale abundance in 1986, the only year right and sei whales were observed in significant numbers. 692 Fishery Bulletin 88(4), 1990 1982 1983 1984 1986 1986 198? 1988 RIGHT WHALES- 1982 1983 1984 1985 1986 YEAR Figure 3 Average number of baleen whales per unit effort for each year of the study by species. Note: the scale for humpback and fin whales (upper) is different from that of right and sei whales (lower). Discussion The data presented here strongly suggest that major changes which occurred in the abundance and distribu- tion of humpback, fin, right, and sei whales in the Stellwagen Bank area, 1982-88, were not independent of each other. During 1986 a dramatic decline in the abundance of humpback and fin whales also co-occurred with the unusual occurrence of right and sei whales dur- ing the summer of the same year. During 1987 and 1988 the abundance of baleen whales returned to a pat- tern similar to that observed in years prior to 1986 (right and sei whales absent, humpback and fin whales abundant). Concurrent with changes in whale abun- dance, an increase in copepod abundance on Stellwagen Bank in 1986 correlated with a significant decline in local sandlance abundance. These correlations likely reflect both the similarities and dissimilarities in prey preferences of individual whale species, and the predator-prey relationship be- tween sandlance and copepods. Given the documented preference of both right and sei whales for copepods (Nemoto 1959, Mitchell 1975a, Nemoto and Kawamura 1977, Watkins and Schevill 1979), a strong positive cor- relation between the two species should be expected. A similar correspondence exists between humpback and fin whales, both of which show an apparent pref- erence for schooling fish in this region (Watkins and Schevill 1979, Mayo et al. 1988). Conversely, the strong negative correlations reported here would be expected between the planktiverous right and sei whales and the pisciverous humpback and fin whales. Competition between whale species These results, in support of previous studies (Kenney et al. 1981, Payne et al. 1986), suggest that the recent distribution of humpback whales in the Gulf of Maine has been dependent on the spatial distribution of sandlance in that region. These results follow a trend which began in the mid-1970s and, when taken with the results of previous studies (Overholtz and Nicolas 1979, CeTAP 1982, Hain et al. 1982, Mayo 1982, Mayo et al. 1988), indicate that sandlance has been the most important prey item of humpback whales in the Gulf of Maine since at least 1976. It is also of interest that, while fin whales also close- ly followed the abundance patterns of sandlance, the overall changes in fin whale abundance fluctuated much less than those shown for the humpback whale. Al- though both whale species prey on a variety of taxa (Jonsgaard 1966) and have been characterized as gen- eralists in their dietary preferences (Mitchell 1975a), behavioral differences between the two species may result in the ability of fin whales to exploit prey other than sandlance in the study area. Watkins and Schevill (1979) noted clear differences between the feeding behavior of humpback and fin whales when exploiting the same prey (schooling fish); fin whales lunge horizontally and humpbacks vertical- ly. Humpbacks are also known to produce "bubble clouds or nets" to potentially concentrate prey (Jurasz and Jurasz 1979, Hain et al. 1982, Hays et al. 1985), a feeding method not observed for fin whales. It has also been suggested that the streamlined form and greater speed of fin whales allow them to exploit more widely separated patches of prey than humpbacks (Brodie 1975, Whitehead and Carlson 1988). Thus fin whales may be more independent of local fluctuations in prey availability, explaining why they were located in the study area in 1986, after humpbacks had aban- doned it. Evidence clearly exists to support the belief that com- petition exists between right and sei whales (Mitchell 1975a, Kawamura 1978). However, Kenney et al. (1986) suggested that competition between right and sei whales in the shelf waters of the northeastern United States is unlikely given their present allopatric distributions (excluding 1986). Perhaps the more im- portant ecological question concerns the reason for the present apparent allopatry. As detailed below, we Payne et al : Abundance fluctuations of baleen whales in the southern Gulf of Maine 693 believe that this recent distributional change has resulted from differing degrees of interspecific com- petition between each of the two whale species and sandlance, moreso than from any competition between the whales themselves. Competition between sandlance and pianktiverous whiale species Historically, and in recent times, right whales have been recorded in the Stellwagen Bank region in late winter and early spring (Allen 1908, Allen 1916, Wat- kins and Schevill 1982, Schevill et al. 1986, Hamilton and Mayo 1990). The species has been virtually absent from the area in summer, when the population's distribution is centered in the Bay of Fundy and on the Scotian Shelf (Kraus et al. 1982, 1986; Mitchell et al. 1986; Winn et al. 1986; Stone et al. 1988; Murison and Gaskin 1989). Therefore, the prolonged residency by a group of right whales in the Stellwagen region throughout summer 1986 may represent the most sig- nificant departure from their usual seasonal pattern of occurrence recorded this century. Is the annual movement of right whales out of the southern Gulf of Maine and into the Bay of Fundy and Scotian shelf areas in early summer related to in- creased activity and competition from pianktiverous fish for calanoid copepods? The question of why right whales generally undertake this northward migration has yet to be resolved. The occurrence of right whales in the Stellwagen Bank area in the summer of 1986 im- plies that, if certain conditions are met, the normal northward movement is not inevitable for at least part of the population. The occurrence of pianktiverous fish (either sandlance, herring or mackerel)— and conse- quently of increased competition through predation on, or disruption of, copepod patches in spring— may be a major factor in directing the seasonal movement of right whales out of the area to seek more acceptable patches elsewhere in the Gulf of Maine. Not only are the size and density of copepod patches important to the feeding energetics of right whales, but also the relative proportions of adult copepods within each patch (Wishner et al. 1988). Wishner et al. (1988) described a copepod surface patch (with associated skim-feeding right whales) which contained a regional maximum of copepodite IVs and Vs, the older, ener- getically richest, developmental stages of copepods (Comita et al. 1966). Although the feeding ecology of right whales is likely more complex than previously thought (Mayo and Marx 1990), these dense aggrega- tions of older, caloric-rich copepods seem to be the required characteristics for energetically successful foraging by right whales (Kenney et al. 1986, Wishner et al. 1988). The principal diet of sandlance >21 mm is also the larger, older developmental stages of C. finmarchicus (Norcross et al. 1961, Scott 1973, Monteleone and Peterson 1986). If the superabundant sandlance are selecting the caloric-rich stages of the copepod popula- tion, then there may be insufficient prey available in the remaining developmental stages (independent of abundance) to provide right whales with the required energy densities (as described by Kenney et al. 1986) to meet the metabolic and reproductive demands of the right whale population. Therefore, during those years when sandlance (or ecologically similar pianktiverous fish species) are abundant, intraspecific competition may be so intense that right whales may only be achiev- ing basal metabolic needs in the southern Gulf of Maine, precipitating the northward movement of right whales immediately following the period of increased sand- lance abundance and activity. The occurrence of consistent numbers of sei whales in 1986 followed many years in which the species had not been recorded from the study area during any season (Mayo et al. 1988). Jonsgaard and Darling (1977) describe "invasion years" where unusually large num- bers of sei whales occur in an area. These years are generally followed by an equally sharp decrease in the number of whales in the same area. Albeit simple, an "invasion year" in response to increased prey abun- dance accurately describes the occurrence of sei whales on Stellwagen Bank during 1986. While sei whales seem to prefer planktonic prey (Watkins and Schevill 1979), they have the ability to forage on fish (Nemoto 1959, Nemoto and Kawamura 1977). Despite this, available distributional evidence suggests that sei whales have not exploited the abun- dant sandlance in the Gulf of Maine to a significant ex- tent. Scant data suggest that sei whales in this region instead subsist primarily on euphausiids and copepods which occur in the waters south of Georges Bank (Ken- ney and Winn 1987). By remaining on the shelf edge and exploiting a wider range of planktonic (and possibly fish) prey, they may be able to minimize the effect of competition from sandlance for copepods, separating them from the more stenophagic right whales. Since the early 1950s sandlance populations have fluctuated on remarkably few occasions. Monteleone et al. (1987) described relatively high sandlance den- sities during 1965-66 (following the depletion of her- ring and mackerel stocks) and 1978-79 (following the larval sandlance explosion in 1975), and exceptionally low densities of sandlance in 1971-74. Meyer et al. (1979) also reported that adult sandlance were not reported on Stellwagen Bank between 1967-76 and that the mean sandlance catch/10 m- in spring 1977 was nine times greater than in spring 1974. Therefore, the two recorded time intervals when sei whales and 694 Fishery Bulletin 88(4), 1990 right whales were most abundant in the study area (April 1970, reported by Watkins and Schevill 1982; and during 1986) were during periods of documented low adult sandlance densities and observed copepod maxima. The March 1975 observations by Watkins and Schevill (1979) of a sei and right whales feeding together in Cape Cod Bay were also made in the inter- val between the reduction of planktiverous herring and mackerel stocks and the subsequent population explo- sion of adult sandlance. These observations further sup- port the theory that planktiverous fish abundance has negatively impacted the occurrence of planktiverous whale species in the region. In this paper, we have examined species interactions between sandlance (as prey and competitor), C.finmar- chicus (as prey), and the abundance of four species of baleen whales in the southern Gulf of Maine. The inter- pretation of these interactions, although largely induc- tive, provide useful management insights into the relationships between planktiverous finfish and the cetacean community. This interpretation also presents an interesting multispecies-management dilemma re- garding finfish and cetacean populations in this region, as well as potential conflicts between single-species re- covery plans and two ecologically different groups of endangered whales, pisciverous and planktiverous. The present abundance of sandlance in the Gulf of Maine followed a long-term depletion of other plank- tiverous fish species through overfishing (Grosslein et al. 1980). Although it is impossible to reconstruct the extent of competition between finfish (herring and mackerel stocks) and planktiverous whales prior to the mid-1960s, the similarities in prey preference and trophic relationships between mackerel, herring and sandlance (Bowman et al. 1984) certainly suggest that competition between herring-mackerel stocks and planktiverous whales could have existed. We hypothe- size that the abundance and distribution of planktiv- erous fishes have played a significant, perhaps critical, role in structuring (as prey and competitor) baleen whale populations in the Gulf of Maine. The present, concentrated distribution of humpbacks in the Gulf of Maine is directly related to the distribu- tion and increased abundance of sandlance in that region. Alternatively, the recovery of the stenophagic right whale (historically and at present) may indeed be inhibited in large regions of the northwest Atlantic by competition for its prey by herring, mackerel, and sandlance. Despite decades of protection Schevill et al. (1986) suggested that the [abundance of the] right whale population [now] which passes Cape Cod is "at worst slightly smaller than it was in the 17th century". The abundance and distribution (and continued recov- ery) of right whales, in contrast to humpback and fin whales, may be contingent upon anomalous disrup- tions in the occurrence of abundant, ecologically equi- valent finfish stocks, which at present are largely human-induced. Acknowledgments The authors would like to thank Stan Tavares, Al Avellar, and Aaron Avellar for supporting whale research and the collection of data aboard their vessels, and for helping establish rational whale-watching guidelines throughout the New England region. We would also like to thank anonymous reviewers for their constructive criticisms, and the following members of the North Atlantic Marine Mammal Association for reviewing previous drafts of this manuscript: Vicki Rowntree, Mason Weinrich, Steve Katona, Peter Tyack, Gordon Waring, Dave Mountain, Fred Wenzel, Bob Kenney, Scott Kraus, and Amy Knowlton. We would also like to thank Howard Winn for his suppor- tive comments regarding this study following its presentation at the Eighth Bienniel Conference on the Biology of Marine Mammals. The authors have been funded by the Manomet Bird Observatory, the National Marine Fisheries Service/Northeast Fisheries Center, Massachusetts Whale Watching Center, Center for Coastal Studies, and l>y private funding. Citations Allen. (i.M. 1916 The whalebone whales of New England. Mem. Boston Soc. Nat. Hist. 8:107-322. Allen, .I.A. 1908 The North Atlantic right whale and it.'; near allies. Bull. Am. Mus. Nat. Hist. 24:277-329. Anthony, V.A., and G. Waring 1980 The assessment and management of the (Jeorges Bank herring fishery. Rapp. P.-V. Reun. Cons. Int. Exi>lor. Mer 177:72-111. Bowman. R., .1. Warzocha, and T. Morris 1984 Trophic relationshijis between Atlantic mackerel and American sandlance. Int. Counc. Explor. Sea (K'ESl. CM I!t84/H:27, 10 p. Brodie, P.F. 197.5 Cetacean energetics, an overview of intraspecific size variation. Ecology ,'36:1.'J2-1()1. Brown. C.W., and H.E. Winn 1989 Relaticinshi|.) between tlie distril)ulion pattern of right whales, Eiihalaena yliiciiilis, and satellite-derived sea surface thermal structure in the (ireat South Channel. Cont. Shelf Res. 9:247-2(iO. CeTAP 1982 A characterization of marine mammals and turtles in the mid- and North Atlantic areas of the U.S. outer continental shelf. P'inal Rep. Cetacean and Turtle Assessment Program, L'niv. Rhode Island. Bur. Land Manage.. Contract AA551- CT8-48. U.S. Dep. Int.. Wash.. DC, 450 p. Payne et al Abundance fluctuations of baleen whales in the southern Gulf of Maine 695 Clark. S.H., and B.E. Brown 1977 Changes in biomass of finfishes and squid from the Gulf of Maine to Cape Hatteras, 1963-1974, as determined from research vessel survey data. Fish. Bull., U.S. 75:1-21. Colebrook. J.M. 1975 The continuous plankton recorder survey: Automatic data processing methods. Bull. Mar. Ecol. 8:123-142. Comita, G.W., S.M. Marshall, and A.P. Orr 1966 On the biology of Ca/(n»;,s/!Hmn(r/ii>HS. XIII. Seasonal change in weight, calorific value and organic matter. J. Mar. Biol. Assoc. U.K. 46:1-17. Eck. G.W., and L. Wells 1987 Recent changes in Lake Michigan's fish community and their probable causes, vnth emphasis on the role of the alewife {Alosn pseudoharenffus). Can. J. Fish. Aquat. Sci. 44:53-60. Grosslein, M.D. 1969 Groundfish survey programs of BCF Woods Hole. Com- mer. Fish. Rev. 31:22-30. Grosslein, M.D.. R.W. Langton, and M.P. Sissenwine 1980 Recent fluctuations in pelagic fish stocks of the northwest Atlantic, Georges Bank region, in relation to species interac- tions. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 177:374-404. Hain. J.H., G.R. Carter, S.D. Kraus, C.A. Mayo, and H.E. Winn 1982 Feeding behavior of the Humpback Whale Megaptera novaeangliae, in the western North Atlantic. Fish. Bull., U.S. 80:259-268. Hamilton, P.K., and C.A. Mayo 1990 Population characteristics of right whales, Eubahwna glacialis, in Cape Cod Bay and Massachusetts Bay, 1978- 1986. In Hammond, P.S., etal. (eds.), Individual recognition and the estimation of cetacean population parameters. Rep. Int. Whaling Comm., Spec. Iss. 12 (in press). Hays, H.E., H. Winn, and R. Petricig 1985 Anomalousfeedingbehaviour of a humpback whale. .J. Mammal. 66:819-821. Jonsgaard, A. 1966 The distribution of Balaenopteridae in the North Atlan- tic Ocean. In Norris, K.S. (ed.), Wliales, dolphins, and por- poises, p. 114-123. Univ. Calif. Press, Berkeley. Jonsgaard, A., and K. Darling 1977 On the biology of the Eastern North Atlantic Sei Whale, Balaenoptera boreaiis Lesson. Rep. Int. Whaling Comm., Spec. Iss. 1:124-129. Jossi, J.W., and D.E. Smith 1990 Continuous plankton records: Atlas of plankton variations in the Gulf of Maine, 1961 through 1988. NO A A Tech. Memo. NMFS-F/NEC, Northeast Fish. Sci. Cent., Natl. Mar. Fish. Serv., NOAA, Narragansett, Rl 02882 (in press). Jurasz, CM., and V.P. Jurasz 1979 Feeding modes of the humpback whale, Megaptera novaeangliae, in southeast Alaska. Sci. Rep. Whales Res. Inst. 31:69-83. Kawamura, A. 1974 Food and feeding ecology in the southern sei whale. Sci. Rep. Whales Res. Inst. 26:25-144. 1978 An interim consideration on a possible interspecific rela- tion in southern baleen whales from the viewpoint of their food habits. Rep. Int. Whaling Comm. 28:411-420. Kenny. R.D., and H.E. Winn 1987 CeUicean biomass densities near submarine canyons C(.)m- pared to adjacent shelf/slope areas. Cont. Shelf Res. 7: 107-114. Kenney. R.D., D.R. Goodale, G.P. Scott, and H.E. Winn 1981 Spatial and temporal distribution of Humpliack Whales in the CETAP study area. Chap. V In A characterization of marine mammals and turtles in the Mid- and North Atlantic areas of the U.S. outer continental shelf. Annu. Rep. 1979, Cetacean and Turtle Assessment Program, Llniv. Rhode Island, Bur. Land Manage. Contract AA551-CT8-48. U.S. Dep. Int., Wash.. DC. Kenney. R.D., M.A.M. Hyman, R.E. Owen. G.P. Scott, and H.E. Winn 1986 Estimation of prey densities required by western North Atlantic right whales. Mar. Mammal Sci. 2:1-13. Kraus, S.D. 1985 A review of the status of right whales (Eubalaena glacialis) in the western North Atlantic with a summary of research and management needs. U.S. Mar. Mammal Comm. Contract MMC-84/08. Wash., DC. Avail. Natl. Tech. Inf. Serv., Rep. PB86-154143, U.S. Dep. Commer., Springfield, VA 22161, 61 p. Kraus, S.D., J.H. Prescott, P.V. Turnbull. and R.R. Reeves 1982 Preliminary notes on the occurrence of the North Atlantic right whale, Eubalaena glacialis, in the Bay of Fundy. Rep. Int. Whaling Comm. 32:407-411. Kraus, S.D.. J.H. Prescott, A.R. Knowlton, and G.S. Stone 1986 Migration and calving of right whales (Eubalaena glacialis) in the western North Atlantic. Rep. Int. Whaling Comm., Spec. Iss. 10:139-144. Mayo, C.A. 1982 Observations of cetaceans: Cape Cod Bay and Southern Stellwagen Bank, Massachusetts 1975-1979. U.S. Mar. Mam- mal Comm. Rep. MMC-80/07, Wash., DC. Avail. Natl. Tech. Inf. Serv., Rep PB82-186263, U.S. Dep. Commer., Springfield, VA 22161, 68 p. Mayo, C.A.. and M.K. Marx 1990 Feeding behaviour of northern right whales, Eubalaena . glacialis, in Cape Cod Bay, and associated zooplankton characteristics. Can. J. Zool. 68 (in press). Mayo, C.A.. O.K. Mattila, S. Pittman, and L. Baraff 1988 Abundance, distribution and habitat use of large whales in Massachusetts Bay and the Great South Channel. North- east Fish. Cent. Contract 50-EANF-6-00059, Natl. Mar. Fish. Serv., NOAA, Woods Hole, MA 02543, 1.52 p. Meyer, T.L., R.A. Cooper, and R.W. Langton 1979 Relative abundance, behavior and food habits of the American sandlance. Ammodytes ameri.canus from the Gulf of Maine. Fish. Bull., U.S. 77:243-253. Mitchell. E. 1975a Trophic relationships and competition for food in the northwest Atlantic. In Proc. Can. Soc. Zool. Annu. Meet., 2-5 June 1974, Fredricton, N.B., p. 123-132. 1975b Preliminary report on Nova Scotia fishery for Sei Whales {Balaenoptera boreaiis). Rep. Int. Whaling Comm. 25:218-225. Mitchell, E.. and D.G. Chapman 1977 Preliminary assessment of stocks of northwest Atlantic sei whales {Balaenoptera boreali,^). Rep. Int. Wlialing Comm., Spec. Iss. 1:117-120. Mitchell, E.D., V.M. Kozicki, and R.R. Reeves 1986 Sightings of right whales, Eubalaena glacialis, on the Scotian Shelf, 1966-1972. Rep. Int. Whaling Comm., Spec. Iss. 10:83-107. Monteleone, D.M., and W.T. Peterson 1986 Feeding ecology of American sandlance. Ammodytes amencanus, larvae from Long Island Sound. Mar. Ecol. Prog. Ser. 30:133-143. 696 Fishery Bulletin 88(4). 1990 Monteleone, D.M., W.T. Peterson, and G.C. Williams 1987 Interaniuial fluctuations in the density of sand lance, .4 w- modytes americaniis, larvae in Long Island Sound. 1951- 1983. Estuaries 10:246-254. Morse, W. 1982 Spawning stock biomass estimates of sandlance. A>n.mo- dytes sp.. off northeastern United States determined from MARMAP Plankton surveys, 1974-1980, Int. Counc. Explor. Sea (ICES). CM 1982/0:59. 11 p. Murison, L.D., and D.E. Gaskin 1989 The distribution of right whales and zooplankton in the Bay of Fundy. Can. J. Zool. 67:1411-1420. Nelson, G.A., and M.R. Ross 1989 The population dynamics of sand lance, {A7nm0dyt.es dubiiis) in the northwest Atlantic: Abundance and distribu- tion. M.S. thesis, Univ. Mass., Dep. Forestry Wildl. Manage., Amherst. 66 p. Nemoto, T. 1959 Food of baleen whales with reference to whale move- ments. Sci. Rep. Whales Res. Inst. 14:149-290. Nemoto, T.. and \. Kavvamura 1977 Characteristics of food habits and ihstnbutions of baleen whales with special reference to the abundance of North Pacific sei/Bryde's whales. Rep. Int. Whaling Comm., Spec. Iss. 1: 80-87^ Norcross, J. J., W.H. Massmann, and E.B. Joseph 1961 Investigations of inner continental shelf waters off lower Chesapeake Bay. Part II. Sandlance larvae. Arinnndyttv amerininuf:. Chesapeake Sci. 2:49-59. Overholtz, W.J., and J.R. Nicolas 1979 Apparent feeding by the fin whale. Bdlaeimpti-rd /i/ij/.sij- bis, and humpback whale, Megaptera novaettngliae, on the American swndhmce. Am modytes americamus, in the Northwest Atlantic. Fish. Bull., U.S. 77:285-287. Payne, P.M., and L.A. Selzer 1989 The distribution, abundance and selected prey of the har- bor seal, Phuca fituliiui ronrolor. in southern New England. Mar. Mammal Sci. 5:173-192. Payne. P.M.. J.R. Nicolas. L. O'Brien, and K.D. Powers 1986 L)istribution of the humpback whale, Meijaptern Home- (tngliiie. on Georges Bank and in the Gulf of Maine in relation to densities of the sand eel, Ammodytes airieriranus. Fish. Bull.. U.S. 84:271-277. Powers, K.D., and E.H. Backus 1987 Estimates of consumption by sealiirds on Georges Bank. In Backus. R.H., and D.W. Bourne (eds.). Georges Bank and its surroundings, p. 372-374. MIT Press, Cambridge, MA. Reay. P.J. 1970 Synopsis of biological data on North /Atlantic sand eels of the genu?, Ammodytes: A. tobiuiins, A. diil)ius. A. omeri- caniis, and A. mnriniis. FAO Fish. Synop. 82, 28 p. Richards, S.W. 1982 A.spects of the biology o{ Ammodytes nmeririnius from the St. Lawrence River to Chesapeake Bay, 1972-75, including a comparison of the Long Island Sound postlarvae with Am- modytes dnhius. J. Northwest Atl. Fish. Sci. 3:93-104. Richards, S.W., and A.W. Kendall 1973 Distribution of sand lance, Ammodytes sp., larvae on the continental shelf from Cape Cod to Cape Hatteras from RV DOLPHIN surveys in 1966. Fish. Bull.. U.S. 71:371-386. Richards, S.W., A. Perlmutter, and D.C. McAneny 1 963 A taxonomic study of the genus .4 m modytes from the east coast of North .•\merica (Teleostei: .4 »( modytes). Copeia 1 963: 358-377. Schevill, W.E.. W.A. Watkins, and K.E. Moore 1986 SVdtus t'{ Eiitialdenii gldcialis off Cupe Cod. Rep. Int. Whaling Comm.. Spec. Iss. 10:79-82. Scott, J.S. 1973 Food and inferred feeding behavior of northern sand lance iAinmadytes dnhius). J. Fish. Res. Board Can. 29: 1667-1671. Sherman, K.C., C. Jones, L. Sullivan, W. Smith, P. Berrein, and L. Ejsymont 1981 Congruent shifts in sand eel abundance in western and eastern north Atlantic ecosystems. Nature (Lond.) 291: 486-489. Smith, S.H. 1968 Species succession and fishery exploitation in the (b'eat Lakes. J. Fish. Res. Board Can. 25:667-693. 1970 Species interactions of the alewife in the Great Lakes. Trans. Am. Fish. Soc. 99:754-765. Smith, W.D., L. Sullivan, and P. Berrien 1978 Fluctuations in production of sand lance larvae in coastal waters off the northeastern United States, 1974 to 1977. Int. Counc. E.xplor. Sea (ICES), CM 197S/L:.30, 8 p. Smith, W.I)., D. McMillan, P. Rosenberg, M. Silverman, and A. Wells 1980 Seasonal and annual changes in the distribution, abun- dance, and species composition of fish eggs and larvae off the northeastern L'nited States, 1977-79. Int. Counc. Explor. Sea (ICES), CM 1980/L:66. Stone, G.S.. S.D. Kraus, J.H. Prescott, and K.W. Hazard 1988 Significant aggregations of the endangered right whale, Eubiiliienii glncitilis, on the continental shelf off Nova Scotia. Can. Field-Nat. 102:471-474. Strong. D.R., D. Simberloff, L.G. Abele, and A.B. Thi-stle 1984 Ecological communities: Conceptual issues and the evi- dence. Princeton Univ. Press, Princeton, 613 p. Watkins, W.A., and W.E. Schevill 1979 Aerial observation of fee300 mm) cutthroat collected during this study were presented by Brodeur et al. (1987a). Results Abundance A total of 163 cutthroat and 134 juvenile steelhead trout were collected during our cruises in 1981-85. These species combined comprised about 3% of our catches of juvenile salmonids (Pearcy and Fisher In press). Cutthroat and steelhead trout were captured from May through August, and occurred in 0-30%i and 2-24%i, respectively, of the seine sets during these months (Table 1). Frequencies of occurrence and catches per set of steelhead were generally highest in May and June and were much lower in July and August (Table 1). Steel- head were absent in September of all years. Abun- dances of cutthroat trout in 1981, the year of most regular sampling, were highest in July and about the same in May, June, and August. Like steelhead, cut- throat trout were absent in September of all years (Table 1). Distributional trends In most years, average catches of cutthroat trout were higher in the region off southern Washington and northern Oregon, near the mouth of the Columbia River (Zone B), than off northern Washington (Zone A) or Oregon south of the Columbia River (Zone C) (Table 2). No obvious latitudinal trends were noted 700 Fishery Bulletin 88(4). 1990 Table I Percent frequency of occurrence (FO), mean catch per set (CPS), and number (n of juvenile cutthroat 1 and steelhead trout caught in purse seine collections, 1981-85 off Oregon and Washington. (S in- dicates total number of purse seine sets by month.) No steelhead or cutthroat trout were collected in September 1982, 1983, or 1984 (n = 38, 51, and 63 sets, respectively) . NS indicates no sets ivere taken during those months. May June July August S = 208 S = 327 S = 130 S = 66 FO CPS n FO CPS n FO CPS n FO CPS n Cutthroat 1981 13 0.29 18 9 0.19 13 30 0.61 41 12 0.21 14 1982 5 0.13 8 14 0.16 9 NS NS NS NS NS NS 1983 9 0.13 7 7 0.14 8 NS NS NS NS NS NS 1984 NS NS NS 6 0.11 7 3 0.05 3 NS NS NS 1985 25 0.46 13* 19 0.28 22 NS NS NS NS NS NS Steelhead 1981 22 0.51 32 15 0.37 25 7 0.07 5 •> 0.02 1 1982 24 0.52 32 4 0.04 2 NS NS NS NS NS NS 1983 7 0.07 4 3 0.03 2 NS NS NS NS NS NS 1984 NS NS NS 8 0.12 8 6 0.08 5 NS NS NS 1985 18 0.29 8* 6 0.13 10 early June sampling off the Columbia NS NS R. mouth. NS NS NS NS * Late May anc Table 2 Average catch per set of cutthroat/steelhead trout in purse seine sets off C )regon and Washington, 1 during May, June, and July combined, 1981 -85 (number of sets in parentheses). NS = no samples. Zone 1981 1982 1983 1984 1985 A Cape Flattery to NS 0.04/0.46 0.05/0.08 0/0.15 0/0.10 Grays Harbor, WA NS (26) (37) (33) (20) B Willapa Bay, WA to 0.58/0.54 0.16/0.31 0.21/0.03 0.32/0.16 0.50/0.15 Nehalem, OR (111) (45) (33) (25) (54) C Tillamook Bay to 0.09/0.02 0.19/0.17 0.14/0.47 0.02/0 0.23/0.23 Cape Blanco, OR (86) (47) (43) (48) (34) D Extended cruise Swiftsure Bank to 0/0 Winter Harbor, B.C. (15) Cape Blanco. OR to 0/0.5 False Cape. CA (8) for steelhead. Steelhead were captured as far south as northern California during the extended cruise of July 1984. The highest average catches of both cutthroat and steelhead trout occurred 37.2-46.3 km offshore (Table 3). Low catches of both species were found in the zone closest to shore (<9.3 km). Catches of cutthroat were high 37.2-46.3 km offshore only in May and June 1985, when juvenile coho salmon were also found far offshore (Fisher and Pearcy 1985). In other years highest cat- ches of cutthroat were 9.4-27.8 km offshore, inshore of maximal steelhead catches. The zone of highest average catches for all years was characterized by a mean sea surface temperature of 13.4°C (standard deviation, SD, 1.4), mean surface salinity of 28.6"/oo (SD 4.3) and secchi depth of 5.4 m (SD 2.2). The low surface salinities in areas where steelhead and cut- throat trout were abundant indicate the influence of freshwater, often the Columbia River plume. Detailed spatial information on catches is given by Pearcy and Fisher (In press) by cruise. Pearcy et al ' Oncorhynchus clarki clarki and O mykiss off Oregon and Washington 701 Table 3 Mean catch per set of cutthroat and st eelhead trout and standard deviation at vai-ying distances from the Oregon/Washington coast. 1981-85, giving equal weight to each cruise (15 cruises sampled inshore of 46 km, except 9 cruises at 46.4 -55.6 km, and 2 cruises at >55.6 km). Distance km <9.3 9.4-18.5 18.6-27.8 27.9-37.1 37.2-46.3 46.4-55.6 offshore: n.mi. <5 5-10 10-15 15-20 20-25 25-30 >55.6 Cutthroat x 0.09 0.25 0.17 0.15 0.33 0.03 0.10 SD 0.09 0.26 0.25 0.28 0.61 0.10 0.14 Steelhead / 0.03 0.09 0.16 0.23 0.52 0.19 0 SD 0.0(1 0.09 0.23 0..36 0.81 0.38 — Table 4 Release and recapture data for tagged cutthroat and steelhead trout off Oregon and Wash mgton. Days Distance Release Capture between N/S of release and release Tagged Recovered Date Location Code Date Location recovery (km) FL (mm) Cutthroat 12 June 85 Yaquina Head Floy 25 July 85 Siuslaw R. 43 72 S 294 — 18 June 85 Cape Disappoint. Floy Fall 85 Umpqua R. - 290 S 295 - Steelhead 27 March- Quinault R., WA 0507.58 9 June 81 Leadbetter Pt., WA 18-74 85 S 172 22 May 81 4 May 81 Clearwater R., ID 102252 21 May 81 Warrenton, OR 17 9 S 206 14 May 82 Quinault R., WA 051043 31 May 82 Cape Lookout, OR 17 269 S 185 4 May 84 Clearwater R., ID 051335 10 June 84 Cape Disappoint. 37 9 N 194 22 June 83 Cape Disappoint. Floy 9 Sept. 83 Jones Beach, Col. R. 79 11 S 445 495 Tag recoveries Two of the cutthroat trout tagged with Floy tags on our cruises in 1985 were recovered by fishermen, one 43 days after release 72 km south of the tagging loca- tion and the other 290 km south of the tagging loca- tion (Table 4). We also caught four hatchery coded-wire tagged steelhead smolts in the ocean 17-74 days after hatchery release. The two released in the Clearwater River, Idaho, were captured close to the mouth of the Columbia River where they entered the ocean. Two fish released in the Quinault River, Washington, were recovered 85 km and 269 km south of where they entered the ocean. Movements to the south by these two steelhead and the cutthroat trout may have been related to advection of surface waters during the upwelling season (Pearcy and Fisher 1988). (.)f the eight maturing steelhead tagged, one (445 mm FL) tagged off Cape Disappointment near the mouth of the Columbia River was recaptured 79 days follow- ing its release at 495 mm FL in the Columbia River. Its growth rate between release and recapture was 0.63 mm/day. Age, lengths, and sex ratios Of the 110 cutthroat trout with readable scales, 32%. 45%, 19%, and 3% migrated to sea for the first time after one, two, three, and four winters in freshwater (ages 1., 2., 3., and 4.)*, respectively (Table 5). This is a younger age distribution than found by Giger (1972) for wild cutthroat trout from coastal Oregon river systems, but similar to that found by Loch and Miller (1988) in the ocean off the Columbia River mouth. Most of the age 1 . cutthroat we caught probably originated from hatcheries (Loch and Miller 1988), including hatcheries on the Columbia River. Most cutthroat were immature or maturing fish on their first ocean migration (age .0). Only eight fish had scales that showed evidence of reduced growth (a * Age designation follows that recommended by Koo (1962), and used by others (Godfrey et al. 1975, Hartt and Dell 1986), where the numbers before and after the decimal point refer to winters spent in freshwater prior to first migration to the ocean and winters spent in the ocean, respectively. Ages 1., 2., 3., etc. designate the fresh- water age of a fish without reference to its ocean age. 702 Fishery Bulletin 88(4), 1990 Freshwater age caught in ocean ' and purse Table 5 number of repeat spawners, as determined from scale analysis, seines off Oregon and Washington. and lengths of cutthroat and steelhead trout Month Cutthroat Steelhead Freshwater age No. of fish Fork length (mm) Freshwater age No. of fish Fork length (mm) X SD spawners X SD May 1. 2. 3. 4. 10 12 7 1 267 258 279 284 46.1 2 38.7 1 27.9 1 0 1. 2. 3. 4. 35 19 7 0 216 185 204 32.5 17.2 17.9 June 1. 2. 3. 4. 13 16 8 2 280 282 314 271 57.4 2 48.0 I 16.1 0 8.5 0 1. 2. 3. 4. 17 7 2 0 209 225 257 28.7 36.2 31.8 July 1. 2. 3. 4. 10 16 5 0 284 299 336 40.9 0 34.8 0 27.2 1 - 0 1. 2. 3. 4. 1 1 0 0 260 246 - August 1. 2. 3. 4. 3 6 1 0 321 316 314 10.1 0 17.9 0 - 0 - 0 1. 2. 3. 4. 0 0 0 0 Total )te. p 701, for 110 explanation 8 of age designation. 89 * See text footnt winter annulus or spawning check) after initial ocean entrance ("repeat spawners," Table 5). SLx of these had apparently made one previous migration from the ocean to freshwater, and two had made two previous trips to freshwater. All of these fish measured 335- 380 mm FL except for one 279 mm FL fish. Half of the repeat spawners had first entered the ocean after one winter in freshwater (age 1.). Seven of the eight prior spawners were collected in May or Jime, probably soon after reentering the ocean. All juvenile steelhead caught in the ocean were in their first ocean summer (age .0). Of the 89 steelhead that had readable scales, 60%, 30%, and 10% were ages 1.0, 2.0 and 3.0, respectively (Table 5). Ten of 11 steelhead caught with clipped adipose fins, probably denoting hatchery origin, were age 1.0. Since most wild steelhead enter the ocean after two or three winters in freshwater and most hatchery steelhead after only one winter (Chapman 1958, Withler 1966, Pauley et al. 1986), the majority of fish we caught were probably of hatchery origin. Length ranges were broad for both cutthroat trout and steelhead during each month (1981-85 combined, Fig. 2). This may have resulted from variable size and age of smolts entering the ocean (Loch 1982, Dawley et al. 1982, Bottom et al. 1984, Ward and Slaney 1988) or variable growth rates. During downstream migra- tion, the length ranges of individual hatchery groups of steelhead smolts can be very broad. For example, Dawley et al. (1986) found length ranges of 150-270 mm FL and 110-240 mm FL for two groups sampled in the upper Columbia River estuary. Back-calculated length at time of ocean entry of cut- throat trout averaged 241 mm FL, but was quite variable (SD 38 mm, n = 101, repeat spawners ex- cluded). Age 3.0 fish tended to be larger at time of first ocean entry than younger fish or age 4.0 fish (average back-calcidated FL at ocean entry = 231, 239, 264, and 244 mm FL for age 1.0, 2.0, 3.0, and 4.0 fish, respec- tively). Our average back-calculated lengths at ocean entry for cutthroat trout caught at sea were larger than those back-calculated by Giger (1972) for wild cutthroat of the same age from the Alsea, Nestucca, and Siuslaw Rivers (239 vs. 210 and 264 vs. 239 mm FL for age 2.0 and 3.0 fish, respectively). However, the mean length at ocean entry estimated by Giger for all fresh- water age groups and for all three river systems com- bined (233 mm FL) by Giger was fairly close to our average (241 mm FL). Pearcy et al.. Oncorhynchus clarki clarki and O mykiss off Oregon and Washington 703 15 10 5 0 15 10 5 0 15 10 5 0 15 10 \ CUTTHROAT LiJ 03 :s FL--^'» August STEELHEAD ^^ May Xn =223 June 287 July g'''^'^ FFT3 120 FORK LENGTH (mm) Figure 2 Size-frequency distributions of cutthroat and steelhead trout, by month, 1981-85 combined. Mean back-calculated FL at time of ocean entry for steelhead was 199 mm (197 mm and 200 mm for fish caught in May and June, respectively) and was quite variable (SD 29 mm, n = 84). These estimated lengths at time of ocean entry were similar to the mean lengths of steelhead smolts caught during downstream migra- tion in the Columbia River (~200 mm, Dawley et al. 1985, their figs. 4-10), and similar to those estimated by Narver (1969, 1974) for steelhead smolts in two British Columbia river systems (182 and 190 mm FL), but were larger than mean lengths of all ages of wild smolts from two Washington streams (156 and 165 m FL; Loch et al. 1988). The mean lengths of both species increased during the summer. This suggests either growth in length dur- ing this time period or higher availability of larger fish later in the summer (Loch 1982, Dawley et al. 1982). Mean lengths of cutthroat trout caught in the ocean in July and August (299 and 318 mm FL, respectively) were similar to mean lengths of age 2. and 3. cutthroat trout (305 and 323 mm FL, respectively) caught in coastal Oregon estuaries from July through September on their initial spawning migrations (Giger 1972). More male than female cutthroat were captured, but fewer male then female steelhead were caught. How- ever, the respective sex ratios of 1.3:1 and 0.7:1 were not statistically different from 1:1 (p>0.1, chi-square test). None of the cutthroat trout examined had en- larged testes or ovaries (>1% of body weight), which is expected since cutthroat spawn in the winter-spring period. Ocean growth Of cutthroat trout caught in the ocean in May, June, July, and August, 65, 83, 90, and 100%, respectively, showed an ocean growth pattern on their scales. Mean estimated ocean growth rates of cutthroat trout (in- cluding fish with and without ocean growth) were 0.47, 0.78, 1.03, and 2.60 mm/day, assuming an ocean entry date of 1 April, 1 May, 10 May, and 31 May, respec- tively. Since the median date of ocean entry of cut- throat trout smolts is sometime in early May (Giger 1972, Loch and Miller 1988, Dawley et al. 1985), growth rate estimates using 1 May or 10 May as an ocean entry date (0.78 and 1.03 mm/day) are probably closest to the true average growth rate of cutthroat trout in the ocean. Ocean growth rates based on the slope of the geometric mean regression (Ricker 1973) of back-calculated ocean growth and Julian date was 1.22 mm/day {n = 101, r = 0.67), fairly close to the mean growth rates of individual fish assuming early May ocean entry dates. Almost all juvenile steelhead with readable scales were caught in May and June. Only 30% and 50%, respectively, of the steelhead caught in these two months showed signs of ocean growth on their scales. This suggests that the steelhead caught in May and June were, on average, in the coastal ocean for less time than the cutthroat trout caught in the same months, either because of later ocean entry or rapid migration of steelhead out of coastal waters. Back- calculated ocean growth rates of juvenile steelhead (in- cluding fish with and without ocean growth) were 0.21, 0.32, and 1.06 mm/day for assumed ocean entry dates of 15 April, 1 May, and 17 May, respectively. Down- stream migration of steelhead smolts in the Columbia River estuary was half completed around the third week in May (Dawley et al. 1985). Consequently, the ocean entry date of 17 May probably gives the best estimate of average growth rate in the ocean (1.06 mm/day). This growth rate was similar to the average 704 Fishery Bulletin 88(4), 1990 growth rate for cutthroat trout using an early May date of ocean entry. The lack of steelhead caught late in the summer precluded estimating mean growth rate from change in back-calculated ocean growth with time. Food habits Cutthroat trout Fishes were by far the dominant prey of juvenile cutthroat trout in terms of frequency, number, and weight. Hexagrammids, scorpaenids, northern anchovy Engraulis mordax, and the brown Irish lord Hemilepidotus spinosus were the dominant fish taxa identified (Appendix Table 1). One uniden- tified Pacific salmon, Oncorhynchus sp. [81 mm total length (TL)], was found in the stomach of a 221 -mm cutthroat trout collected 11.4 km off the mouth of the Columbia River in July 1981 . Fishes made up more than 75% of the biomass consumed during all years, but they decreased in importance in the diet as the summer pro- gressed, so that by late summer several other prey taxa such as euphausiids, hyperiid amphipods, and decapod larvae were important. Prey fishes found in cutthroat stomachs averaged 52.5 mm TL (SD 21.8) and ranged from 21 to 101 mm. No significant relation (n = 46, r = 0.30, p = 0.16) was found between the lengths of cutthroat trout and their fish prey. Steelhead The diet of juvenile steelhead trout was more diverse than that of cutthroat trout. Both arthro- pods and fishes were important prey items. Prey of steelhead trout ranged from small barnacle larvae and copepods to larger juvenile fishes and squids. Euphau- siids, mainly Thysanoessa spinifera or Euphausia pacifica, accounted for over 75% of the total IRI for all years combined (Appendix Table 1). Fishes were more important, however, on a weight basis making up about 60% of the total biomass consumed. Juvenile rockfishes (Sebastes spp.), sandlance Animodytes hex- apterus, brown Irish lord Hemilepidotus spinosus, and greenlings (Hexagrammos spp.) were the dominant fish taxa identified. With the exception of barnacle cypris larvae and hyperiid amphipods which were quite numerous in stomachs collected at several stations, all other prey taxa were relatively unimportant. No rela- tionship between predator and fish prey size was found (n = 38, r = 0.19, p = 0.25). However, steelhead, which had a smaller mean length than cutthroat trout, con- .sumed a smaller mean size (x 35.9 mm TL, SD 16.6 mm) and size range (7-72 mm) of prey fishes than cut- throat trout. The relative proportions by weight of the major prey categories varied substantially among the years sam- pled. Usually, fishes contributed the majority of the weight to the diet, but during the relatively strong upwelling years of 1982 and 1985 (Fisher and Pearcy 1988) euphausiids were more important. The largest number of prey categories occurred in steelhead and cutthroat stomachs during the relatively weak upwell- ing summers of 1981 and 1984. The Percent Similarity Index between cutthroat trout and steelhead was 39% for all cruises combined. Discussion Cutthroat and steelhead trout have evolved contrasting migratory behavior in the ocean. Both species were ab- sent in purse seine catches during September of their first summer in the ocean (Miller et al. 1983, Loch and Miller 1988; this paper), apparently because they migrated out of the coastal ocean. Cutthroat return to estuaries and freshwater streams from mid- to late summer (Giger 1972, Loch 1982), whereas most steel- head migrate far offshore during their first summer in the ocean (Hartt and Dell 1986). Although some aspects of the life histories of cut- throat and steelhead trout are similar, their ecologies in marine waters differ. Steelhead commonly spend 2-3 years in the ocean and attain a large size before initiating their spawning migrations to freshwater. Some, however, return after only one full year in the ocean, and some return, though usually do not spawn, after several months in the ocean (Everest 1973). Ap- parently anadromous cutthroat trout return to fresh- water after only a few months in coastal waters and rarely overwinter at sea. Giger (1972) did not identify any coastal cutthroat that had overwintered in the ocean, based on analyses of circuli patterns from scales of fish returning to coastal rivers of Oregon. Fish that migrated to sea in the spring invariably returned to freshwater in the summer or fall of the same year. Similarly, Armstrong (1971), Jones (1982) and Loch, and Miller (1988) did not report any cutthroat that spent an entire year in the ocean. J. Johnston (Wash. Dep. Wildl., Olympia, pers commun. 8 March 1990) found a few cutthroat trout in Hood Canal, WA, dur- ing February and March, fish that evidently over- wintered in marine waters. Most cutthroat spawn after their return to freshwater, but according to Johnston (1982) a large percentage of the Columbia River, Puget Sound, British Columbia, and Alaska stocks of cut- throat overwinter but do not spawn in freshwater after their first summer in the ocean. Jones (1977) reported that less than 50% of the cutthroat migrating into Petersburg Creek in southeast Alaska were approach- ing sexual maturity. In some instances, coastal cut- throat may migrate downstream but not into the ocean. Smolts from the Cowlitz River in Washington may re- main in the Columbia River estuary for a year (Tipping 1981 as cited by Johnston 1982), and some adults that Pearcy et al Oncorhynchus cisrki cl3rki and O mykiss off Oregon and Washington 705 return to spawn in Sand Creek, OR, and Prairie Creek, CA, had no ocean growth on their scales (Sumner 1972, DeWitt 1954). Tomasson (1978) thought that coastal cutthroat did not migrate in large numbers beyond the Rogue River estuary. Because coastal cutthroat only spend several months in the ocean, it is assumed that they inhabit waters close to the coast and not far from their homestream (Giger 1972). Johnston (1982) observed that anadro- mous cutthroat usually fr-equented waters less than 3 meters in depth in Puget Sound and that migration of tagged wild cutthroat did not extend much beyond 50 km from the home stream. However, cutthroat were caught by Loch and Miller (1988) up to 31.5 km off- shore and within the Columbia River plume between Tillamook Bay, OR, and Willapa Bay, WA, suggesting more extensive offshore movements. Sumner (1972) also thought that cutthroat originating from the Co- lumbia River basin may traverse along the coast within the Columbia River plume. Our data also suggest that some cutthroat undertake substantial movements along the coast (>250 km) and are found as far offshore (37-46 km) as the more oceanic steelhead smolts dur- ing the summer. One fish was caught 66 km offshore. Thus some individuals may reside far offshore along an open coast, as opposed to staying close to shore as they presumably do in protected inlets like Puget Sound (Johnston 1982). Straying of returning sea-run hatchery cutthroat trout is connnon (Bulkley 196(i, Giger 1972, Jones 1977). Giger (1972) reported stray- ing of marked hatchery cutthroat between the Nes- tucca, Siuslaw, and Alsea Rivei's of the central Oregon coast. Most of the strays were recovered in rivers south of the river of release, possibly the result of advection by coastal currents during the early summer period of ocean residence, similar to that seen for eoho salmon (Pearcy and Fisher 1988). Giger (1972) noted a non-random distribution in the number of cutthroat smolts and kelts in seine samples in the Alsea estuary during the spring and "large schools" of sea-run fish in estuaries in the fall. He thought that anadromous cutthroat trout formed schools while at sea. However, we found no evidence of schooling from our limited ocean catches. Most cut- throat were caught singly. To our knowledge, school- ing of cutthroat or steelhead trout in the ocean has not been documented. Steelhead migrate long distances into oceanic waters and are widely distributed in the North Pacific Ocean based on catches of marked and unmarked fish (Sutherland 1973; Pearcy and Masuda 1982, 1987; Okazaki 1983; Hartt and Dell 1986; Light et al. 1988). Miller et al. (1983) found that juvenile steelhead caught in purse seines between Tillamook Bay and Willapa Bay occurred farther offshore than juvenile coho or chinook salmon and migrated out of the coastal sampling area early in the summer. Purse seining and tagging studies by Hartt and Dell (1986) from Cape Flattery, WA, to the Aleutian Islands clearly showed that juvenile steel- head migrate directly offshore rather than along a coastal belt where other juvenile salmonids typically migrate. Recovery of steelhead during their first sum- mer in the ocean in the Gulf of Alaska confirms rapid migrations of some fish into oceanic waters far from land. Pearcy and Masuda (1982) report on a steelhead captured over 1600 km from land only a few months after its initial ocean entry. These conclusions of immediate migrations of steel- head offshore and into subarctic waters of the North Pacific after ocean entry, and residence in oceanic waters during their first winter in the ocean, do not apply to all steelhead. Most steelhead from the Rogue River in Oregon (the "half-pounder" runs) return to freshwater after only a few months following their ini- tial migration to the ocean (Everest 1973), and likely do not migrate very far in the ocean. The recovery of marked Rogue River summer steelhead south of the Rogue River (Everest 1973), and the rarity of marked steelhead originating from streams south of Cape Blanco in waters to the north of Cape Blanco in our catches, suggest that tliese steelhead from the southern extremity of their range may not migrate to the north after ocean entry. This conclusion is supported by the high seas distribution of tagged steelhead. Although 9 tagged steelhead from California and 11 from Oregon streams have been recovered north of 45 °N in the noi'theastern Pacific, only 3 fish with coded-wire tags from California have been recovered at sea of the over 1 million coded-wire tagged steelhead released between 1980 and 1985, and only one was caught north of California (Light et al. i988. Pacific States Marine Fisheries Comm. unpubl.). Possibly southern steelhead reside in the strong upwelling zone off northern Califor- nia and southern Oregon (Bakun 1975). The feeding habits of juvenile steelhead and cutthroat trout in estuaries and in coastal waters are similar. Both species feed intensively on gammarid amphipods and insects in estuaries on their initial migration to the sea, but steelhead smolts also eat the benthic mollusk Corbicula (Loch 1982, McCabe et al. 1983, Bottom et al. 1984). During their early residence in waters along tlie open coast, cutthroat trout feed primarily on fishes (Armstrong 1971, Fresh et al. 1981, Brodeur et al. 1987a, Loch and Miller 1988, this paper). We found that steelhead trout also consumed fish but had a more varied diet than cutthroat trout. Euphausiids and other crustaceans were important in the diet of steelhead trout, especially during the strong upwelling years of our study when euphausiids may have been abundant (Brodeur 1986). Many prey species identified in the 706 Fishery Bulletin 88(4) 1990 diets of the trout species (juvenile roctcfishes, hex- agrammids, anchovy, Cancer spp. megalopae, and in- sects) are commonly found in the neustonic layer (Brodeur et al. 1987b, Shenker 1988) which suggests that these trouts feed in surface waters. The occur- rence of a juvenile salmon in the diet of juvenile cut- throat is noteworthy since larger individuals (>300 mm) of this species were identified as one of the few fish predators on juvenile salmonids among the 20 species of nekton examined from these same purse seine catches (Brodeur et al. 1987a). Dietary overlap of 39% between cutthroat and steel- head trout based on percent similarity was higher than the overlap values between these species of trout and the juveniles of four species of salmon caught in purse seines, with two exceptions. The highest overlap values were between cutthroat trout and juvenile chinook salmon (49%) because of the common utilization of fishes in the diet, and between juvenile steelhead trout and juvenile sockeye salmon (43%) because of the com- mon occurrence of euphausiids. The average growth rate of cutthroat trout in the ocean based on our scale analysis was 0.8 and 1.0 mm/day (based on 1 May and 10 May dates of ocean entry, respectively) and 1.2 mm/day (based on change in back-calculated ocean growth with time). These estimates are similar to the growth rates of about 1.0 mm/day for age 2. wild cutthroat and 0.9 mm/day for hatchery cutthroat, assuming a 3-month ocean resi- dence (Giger 1972) and 0.7-0.8 mm/day for fish after their first 5-6 months in the ocean (Sumner 1962). Johnston (1982) reported an average growth rate of 1.0 mm/day during the time spent at sea for cutthroat trout stocks from the Columbia River, coastal rivers, and Puget Sound rivers. Our average growth rate for steelhead of 1.1 mm/day (based on a 17 May date of ocean entrance) is similar to that for cutthroat trout. It is about the same as the average ocean growth of 1.3 mm/day calculated by Everest (1973), the 0.8 mm/ day estimated by K. Kenaston (Oreg. Dep. Fish Wildl., Corvallis, pers. commun. 16 June 1989) for summer- run steelhead "half-pounders" of the Rogue River, and the 1.5 mm/day estimated for first-year ocean growth of steelhead from Vancouver Island (Hooten et al. 1987). Data on ocean growth of steelhead during their first year in the ocean is also provided by recovery of fish with coded-wire tags. These include growth rates of about 1.0 mm/day for a fish caught in the Gulf of Alaska, based on the mean size of the same tag code of steelhead smolts when caught about 60 days earlier in the Columbia River (Pearcy and Masuda 1982), and growth rates of 0.7, 1.0, and 1.2 mm/day for coded-wire tagged steelhead recovered 198, 186, and 173 days, respectively, after release (Pacific States Marine Fish- eries Comm. unpubl.). Lengths at release were esti- mated from release weight using the length-weight relationships for steelhead smolts given by Everest (1973). The average growth rate of steelhead during their first full year in the ocean was about 1.0 mm/day for fish returning to California streams (calculated from data in Shapovalov and Taft 1954), 0.6 mm/day for fish caught on the high seas (Sutherland 1973), and 0.85 mm/day for fish from the Keogh River, British Colum- bia (Ward and Slaney 1988). These estimates suggest fairly similar growth rates for both cutthroat and steelhead trout during early ocean life. Apparently cut- throat trout are small when they return to spawn because they spend less time in the ocean, not because of an inherently lower growth rate. If coastal cutthroat trout have the potential for a large increase in size during ocean life, why do they curtail marine growth by returning to freshwater each winter rather than remaining in the ocean for several years like most steelhead trout? We present three hypotheses for the adaptive value of this behavior. Early maturation of cutthroat trout, after only a few months in the ocean, may have evolved as a response to low survival (Cole 1954), either in the ocean or fresh- water, especially if postreproductive survival is low after the age at first breeding (Schaffer 1974). Cut- throat trout do not appear to have distinctly lower ocean sui-vival than steelhead, however, based upon the reports in literature. Giger (1972) states that coastal cutthroat exhibited comparatively high rates of survival during summer periods of ocean residence between spawnings. He estimated ocean survival rates of 20-40%! for hatchery cutthroat smolts between release in the spring and return in the fall, and survival rates of 14-39%, 17-35%. and 12-25% between first and second, second and third, and third and fourth spawn- ings, respectively, based on trap or net catches of fish returning to four Oregon coastal streams. Sumner (1962) estimated that 17-50% of coastal cutthroat trout survived between successive spawnings, and Jones (1978) reported marine survival of 17%. Tomasson (1978), on the other hand, noted that 92% of the fish returning to the Rogue River were first migi-ants, sug- gesting low survival of repeat spawners, and Michael (1983, 1989) reported marine survival rates of 2-20% between smolt outmigration and first return to fresh- water for coastal cutthroat trout. Cutthroat survival to an ocean age of 2, the age when many steelhead trout return for their initial spawning (Wit'hler 1966, Shapovalov and Taft 1954, Chapman 1958), is about 5% (assuming 20% survival from smolt to first spawning and 25% survival from first to second spawning, based on Giger's estimates), a value which is close to the average for smolt to adult return foi- steelhead (Bley and Moring 1988), but less than the mean ocean survival of 16% based on maiden-run fish Pearcy et al Oncorhynchus clarki cisrki and O mykiss off Oregon and Washington 707 given by Ward and Slaney (1988). Thus survival rates to ocean age 2 of these two species appear to be roughly similar, as is survival of repeat spawners of cutthroat trout and steelhead trout (Giger 1972, Withler 1966, Ward and Slaney 1988). Lower ocean survival or lower postreproductive survival of cutthroat trout than steel- head may not be a cogent explanation for cutthroat trout spawning at an early age and small size. The second hypothesis is that small size at maturity in cutthroat trout relative to steelhead trout may have evolved as a result of the distance and rigor of spawn- ing migrations. Schaffer and Elson (1975) concluded that the mean age of first spawning of Atlantic salmon increased with the difficulty of upstream migration, as estimated by the distance ascended into freshwater. Coastal cutthroat rarely penetrate inland more than 160 km (Johnston 1982), and hence may not need the swimming performance or energy reserves required for the long and arduous upstream migrations that some steelhead undertake at the time of maturity. Finally, the small size at maturity attained by cut- throat trout may permit utilization of small, shallow tributaries for spawning and rearing where interspe- cific competition with other anadromous salmonids is reduced. Small streams are known to be important spawning and rearing areas of cutthroat trout (DeWitt 1954, Needham and Gard 1959, Lowry 1965, Johnston 1982, Trotter 1989). Sea-run cutthroat trout general- ly spawn in tributaries with a lower velocity and shal- lower depth than steelhead (Hunter 1973). Hartman and Gill (1968) reported that where both anadromous cutthroat trout and steelhead were sympatric, juvenile cutthroat were predominant in headwater tributaries and steelhead in larger river reaches. Cutthroat trout are behaviorally subordinate to steelhead and coho salmon in agonistic encounters (Nilsson and Northcote 1981, Glova 1986, Griffith 1988) and their populations appear to be suppressed by competition from anad- romous salmonids (R. House, Unpubl.). The lack of mor- phological specialization of cutthroat trout to either fast or slow water may be another reason why this species is dominated by coho salmon and steelhead in areas of sympatry (Bisson et al. 1988). Anadromous cutthroat are known to penetrate, spawn, and rear farther into a watershed than steelhead trout (Michael 1983), sometimes above natural falls or log jams (Mitchell 1988; R. House, Unpubl.) that were considered to be barriers to anadromous salmonids (see Michael 1983). Therefore, small size at maturity may be adaptive by alkjwing anadromous cutthroat trout to spawn and rear in numerous small tributaries of coastal streams where other salmonids are absent or less abundant. Acknowledgments We thank B.R. Ward (British Columbia Recreational Fisheries Branch), J. Nicholas and K. Kenaston (Ore- gon Department of Fish and Wildlife), J. Light (Fish- eries Research Institute, University of Washington), J. Dambacher and B. Hicks (Department of Fisheries and Wildlife, Oregon State University), and two anonymous reviewers for helpful comments on the manuscript. This publication is a result of research sponsored by the Northwest and Alaska Fisheries Center (NA-87- ABH-00014 and NA-88-ABH-00043) and the Oregon State University Sea Grant College Program (Grant No. NA-81-AA-b-00086, R/OFP-17). Citations Armstrong, R.H. 1971 Age, food, and migration of searun cutthroat trout, Salw,o (inrki. at Eva Lake, southeastern Alaska. Trans. Am. Fish ."^oc. l(H):.302-306. Bakun, A. 1975 Daily and weekly upwelling indices, west coast of North America. 1967-73. NOAA Tech. Rep. NMFS SSRF-693, 114 p. Bartlett, J.R., P.F. Randerson, R. Williams, and D.M. Ellis 1984 The use of analysis of covariance in the back-calculation of growth in fish. J. Fish. Biol. 24:201-213. Behnke, R.J. 1979 The native trouts of the genus Snlmo of western North America. Rep. to U.S. Fish. Wildl. Serv., Denver, CO. Dep. Fish. Wildl. Biol.. Colorado State Univ., Ft. Collins, 215 p. Bisson, P. A., K. Sullivan, and J.L. Nieben 1988 Channel hydraulics, haliitat use, and body form of juvenile coho salmon, steelhead, and cutthroat trout in streams. Trans. Am. Fish. Soc. 117:262-273. Bley, P.W., and J.R. Moring 1988 Freshwater and ocean survival of Atlantic salmon and steelhead: A synopsis. U.S. Fish Wildl. Serv. Biol. Rep. 88(9), Ft. Collins. CO. 22 p. Bottom, D.L., K.K. Jones, and M.J. Herring 1984 Fishes of the Columbia River estuary. Final Rep., Columbia River Estuary Data Development Prog.. Oreg. Dep. Fish WiML, Portland. 113 p. Brodeur, R.D. 1986 Northward displacement of the euphausiid Nyctiphanef: fiimpler Hansen to Oregon and Washington waters following the El Nino event of 1982-83. J. Crustacean Biol. 6:686-692. Brodeur, R.D., H.V. Lorz, and W.G. Pearcv 1987a Food habits and dietary variability nf )ielagic nekton off Oregon and Washington, 1979-1984. NOAA Tech. Rep. NMFS .57. .32 p. Brodeur, R.D., W.G. Pearcv, B.C. Mundv, and R.W. Wisseman 1987b The neustonic fauna in coastal waters of the northeast Pacific: Abundance, distribution, and utilization by juvenile salmonids. Oregon Sea Grant Coll. Prog. Publ. ORESU- T-87-()01, Oregon State Univ., Corvallis. 61 p. 708 Fishery Bulletin 88(4), 1990 Bulklev. R.V. 1966 Catch of the 1965 tidewater cutthroat sport fishery and notes on the life history of coastal cutthroat in the Siuslaw River, Oregon. Oregon Game Comm., Fish. Res. Rep. 4. Cor- vallis, 29 p. Carl, G.C., W.A. Clemens, and C.C. Lindsey 1959 The fresh-water fishes of British Columbia. British Columbia Prov. Mus., Victoria, 192 p. Carlander, K.D. 1981 Caution on the use of the regression method of back- calculating lengths from scale measurements. Fisheries (Bethesda) 6:2-4. Chapman, D.W. 1958 Studies on the life history of Alsea River steelhead. J. Wildl. Manage. 29:266-279. Clutter, R.I.. and L.E. Whitesel 1956 Collection and interpretation of sockeye salmon scales. Int. Pac. Salmon Fish. Comm. Bull. 9. 159 p. Cole. L.C. 1954 The population conse(|uences of life history phenomena. Q. Rev. Biol. 29:103-137. Dawley, E.M., R.D. Ledgerwood, T.H. Blahm. and A.L. Jensen 1982 Migrational characteristics and survival of juvenile salmonids entering the Columbia River estuary in 1981. Annu. Rep. to Bonneville Power Admin., Contract DE-A179- 81BP30578. Northwest Fish. Cent., Natl. Mar. Fish. Serv., NCAA, 2725 Montlake Blvd. E., Seattle, WA 98112, 43 p. Dawley, E.M., R.D. Ledgerwood, and A. Jensen 1985 Beach and purse seine sampling of juvenile salmonids in the Columbia River estuary and ocean plume, 1977-1983. Volume 1. Procedures, sampling effort, and catch data. NOAA Tech. Memo. NMFS F/NWC-74. Northwest Fish. Cent., Natl. Mar. Fish. .Serv., NOAA, 2725 Montlake Blvd. E., Seattle, WA 98112. 260 p. Dawley, E.M,, R,D. Ledgerwood, T.H. Blahm, C.W. Sims, J.T. Durkin. R.A. Kirn. A.E. Rankis, G.E, Monan. and F.J. Ossiander 1986 Migrational characteristics, biological observations, and relative survival of juvenile salmonids entering the Columbia River estuary, 1966-1983. Final Rep. to Bonneville Power Admin., Contract DE-A179-84BP39652. Northwest Fish. Cent., Natl. Mar. Fish. Serv., NOAA, 2725 Montlake Blvd. E., .Seattle, WA 98112, 256 p. DeWitt, J.W. Jr. 1954 A survey of the coast cutthroat trout, Salmo cla rki da rki Richardson, in California. Calif. Fish Game 40:329-335. Everest. F.H. 1973 Ecology and management of summer steelhead in the Rogue River. Oregon State Game Comm.. Fish. Res. Rep. 7, Corvallis, 48 p. Fisher, J. P., and W.G. Pearcy 1985 Studies of juvenile salmonids off the Oregon and Wash- ington coast, 1985. Oregon Sea Grant Coll. Prog. Publ. ORESU-T-85-004, Oregon State Univ., Corvallis, 31 !>. 1988 Growth of juvenile coho salmon (Oncurhynrhus kitiulrh } in the ocean off Oregon and Washington, USA, in years of dif- fering coastal upwelling. Can. .T. Fish. Ai|uat. Sci. 45: 10.36-1044. Fresh, K.L., R.D. Cardwell, and R,R, Kocms 1981 Food haliits of Pacific salmon, liaitfish, .-ukI their poten- tial competitors and predators in the marine waters of Wash- ington, August 1978 to September 1979. Wash. Dep. Fish. Prog. Rep. 145, Olympia, .58 p. Giger, R.D. 1972 Ecology and management of coastal cutthroat trout in Oreg(}n. Oregon State Game Comm., Fish. Res. Rep. (i, Cor- vallis, 61 p. Glova, G.J. 1986 Interaction for food and space between experimental populations of juvenile coho salmon (Onmrhynchua kisutch) and coastal cutthroat trout {Salmo rlarki riarki) in a laboratory stream. Hydrobiologia 131:155-168. Godfrey, H., K.A. Henry, and S. Machidori 1975 Distribution and abundance of coho salmon in offshore waters of the North Pacific Ocean. Int. N. Pac. F'ish. Comm. Bull. 31, 80 p. Griffith. J.S. 1988 Review of competition between cutthroat trout and other salmonids. Am. Fish. ,Soc. Symp. 4:134-140. Hartman. G.F., and C.A. Gill 1968 Distributions of juvenile steelhead and cutthroat trout {Salmo gairdneri and S. clarki clarki) within streams in southwestern British Columbia. J. Fish Res. Board Can. 25:33-48. Hartt, A.C.. and .M.B. Dell 1986 Early oceanic migrations and growth of juvenile Pacific .salmon and steelhead trout. Int. North Pac. Fish. Comm. Bull. 46, 105 p. Hooten, R.S., B.R. Ward. V.A, Lewynsky, M.G. Livette, and A.R. Facchin 1987 Age and growth of steelhead in Vancouver Island popula- tions. Prov. British Columbia Fish. Tech. Circ. 77, 39 p. House, R. Unpuhl. Comparison of habitat and fish populations in two coastal stream. U.S. Forest Service, Corvallis, OR. 44 p. Hunter, J.W. 1973 A discussion of game fish in the state of Washington as related to water requirements. Wash. Dep. Game, Olympia, 6 June 1990. Fishery Bulletin. U.S. 88:713-718. Commercial trapping for the spiny lobster Panulirus marginatus in the northwestern Hawaiian Islands be- gan in 1976 shortly after research cruises documented the resource in abundance. Two banks, Necker Is- land and Maro Reef, have been most heavily fished since the beginning of the fishery. In 1986 and 1987 re- search cruises repeated the earlier sampling at Necker Island and Maro Reef by trapping at the same sites on each bank with the same gear to document any changes that had oc- curred in density, population, and genetic parameters (Polovina 1989). The repeat sampling documented substantially reduced lobster densi- ties as indicated by catch rates, which in 1986-87 were 37% and 68% of their 1977 levels at Necker Island and Maro Reef, respectively (Polo- vina 1989). A statistically significant (P<0.05) increase in the asymptotic lengths and decreases in the mean lengths at onset of egg production oc- curred between the 1977 and 1986- 87 sampling at both locations (Polo- vina 1989). While changes in these population parameters could be due to density- dependence relationships, some type of genetic change also may have oc- curred from selection induced by the fishing pressure (Nelson and Soule 1987). For example, the fishery har- vests lobsters at about the size the females first begin producing eggs. It is estimated that spiny lobster enter the fishery at 3.1 years of age and that the mean age females begin producing eggs is 2.5 years (Polovina 1989). Thus, heavy fishing pressure may select for female lobsters that sexually mature at a smaller size. Re- duction in the total amount of genetic diversity or changes in the stock structure of the exploited populations are also potential results of intense fishing pressure (Allendorf et al. 1987). These latter types of genetic changes can often be monitored through allozyme analyses (Nelson and Soule 1987), and are the subject of this paper. Spiny lobsters from seven localities in the Hawaiian Archipelago were collected between 1978 and 1980, prior to the expansion of the commer- cial fishery, and analyzed for allo- zyme variation by Shaklee and Sa- mollow (1984). Although overall levels of variation were quite low, they observed polymorphism at 7 of the 46 loci examined. However, no clear pattern of genetic differentia- 713 714 Fishery Bulletin 88(4), 1990 Figure 1 Location of Maro Reef and Necker Island within the Hawaiian Archipelago. tion among populations was detected, and they con- cluded that a single panmictic stock of spiny lobster exists throughout the Hawaiian Ai'chipelago. Our objective was to reexamine allozyme variation at those seven loci to determine if fishing pressure since 1980 affected population structure or genetic variabil- ity. During 1987, new collections of spiny lobsters from two of the same localities in the northwestern Hawaiian Islands (Maro Reef and Necker Island) were made. These two sites represent the mainstay of the commer- cial fishery. They are the most [productive banks, have been fished the longest, and annually receive the greatest fishing pressure. Analysis showed that the observed number of allelic classes and observed heterozygosity have remained essentially unchanged between 1978 and 1987. These data provide no evidence that population bottlenecks have occurred. Fishing pressure since 1980 has not reduced the genetic variability as detected by protein electrophoresis in spiny lobster from either of these two sites in the northwestern Hawaiian Islands. Methods and materials Muscle and digestive gland from 200 specimens from Maro Reef and Necker Island (Fig. 1) were collected, frozen onboard ship, and shipped frozen to the labor- atory for analysis. Upon receipt, tissues were sub- sampled for electrophoresis and stored at -80°C for up to 4 weeks while analysis was completed. Electrophoretic methods generally followed May et al. (1979) and Shaklee and SamoUow (1984) to facil- itate comparison. Samples were analyzed for the pro- ducts of seven loci previously identified as polymorphic: EstS, Esterase, E.G. 3.1. L-; EstD (Umb), Esterase, E.G. 3.1.-.-, resolved with 4-methylumbelliferyl ace- tate; Gpi, glucosephosphate-6-isomerase, E.G. 5.3.1.9; Mpi, mannose-6-phosphate isomerase, E.G. 5.3.1.8; Pgtn. phosphoglucomutase, E.G. 5.4.4.2; Tpepl.2 (Pep-lJ), tripeptidase aminopeptidase, E.G. 3.4.-.-, resolved with L-leucyl-L-leucyl-L-leucine. However, in contrast to the methods of Shaklee and Samollow (1984) who used horizontal and vertical gels, all anal- yses were performed on horizontal starch gels. Tjh'jiI was resolved on an additional buffer, amine-citrate gel and tray buffer, pH 6.9 (Glayton and Tretiak 1972). The electrophoretic data were analyzed using BIOSYS-1 (Swofford and Selander 1981). Allelic classes followed those of Shaklee and Samollow (1984), who pooled alleles into f(fast), m(medium), and s(slow) classes for numerical analyses. Pooling was required in their study and ours because many alleles were too rare to include in the statistical tests. Actual mobilities observed were also recorded. Gomparisons between the 1978-80 and 1987 data sets for each reef were performed using a chi-square contingency analysis based on the number observed in Seeb et a\ : Genetic variation in Panulirus marginatus from the Hawaiian Islands 715 Table 1 Allele frequenciei: foi seven polymorphic oci in the spiny lobster Panulirua mar ginatut;. Data from 1978 -80 are those of Shak ee and Samollow (1984). Alleles are identified by relative mobility classes assigned by Shaklee and Samollow (1984). Actual mobilities ob- 1 served in this study are also given. Numbers of individuals successfully scored are given in parenthesis. Location EstS EstD Gpi f m s f m s f m s ff and date 100 93 100 100 88 66 145 Maro Reef 1978-80 0.003 0.957 (386) 0.040 0.001 0.988 (340) 0.010 0.959 0.039 0.002 (320) — 1987 - 0.968 (174) 0.032 — 1.000 (199) - 0.970 - 0.028 (198) 0.003 Necker Island 1978-80 - 0.971 (666) 0.030 0.011 0.980 (533) 0.008 0.966 0.033 0.001 (535) — 1987 - 0.992 (194) Mpi 0.008 Pgm 1.000 (200) - 0.955 Tpepl 0.002 0.020 (200) Tpep2 0.022 f f+s f m s r s f m s 101 96. 200 100 -33 100 66 100 89 Maro Reef 100 -66 1978-80 0.011 0.989 (418) 0.007 0.984 (345) 0.009 0.725 0.275 (142) - 0.988 (335) 0.012 1987 0.005 0.995 (193) 0.003 0.977 (198) 0.020 0.768 0.232 (179) 0.006 0.975 (180) 0.019 Necker Island 1978-80 0.007 0.993 (649) 0.007 0.976 (558) 0.017 0.785 0.215 (142) 0.001 0.984 (425) 0.015 1987 0.003 0.998 (188) " 0.985 (200) 0.014 0.768 0.232 (144) - 0.980 (176) 0.020 each allelic class. Shaklee and Samollow (1984) found no heterogeneity between year classes for any loci ex- cept EstD. and thus they report pooled frequencies. They provide annual frequency estimates for EstD for 1979 and 1980. We also compared the 1987 data sets from Maro Reef and Necker Island using chi-square contingency analysis. Results Allozyme analysis All loci previously observed to be polymorphic were also found to be variable in the new samples, with the exception oi EstD (Table 1). Allele classes observed by Shaklee and Samollow (1984) and not observed in this study were Est3(f), EstDtJ), and EstD(s). We observed Gpi(H5) in both populations: this allele was not previ- ously reported. Additionally, when analyzing Tpfpl allozymes, we detected a Tpep3(9It) allele in a single in- dividual from Maro Reef which was not observed previously. Isozyme patterns agreed with descriptions provided in the earlier study. There were no indications of non-genetic variation at the esterase or any of the other loci. Genotypic frequencies at all loci fit the expectation of Hardy-Weinberg equilibrium with the exception of EstS in Maro Reef (P = 0.041) and Mpi in both popu- lations. Mpi has an unusual distribution of alleles between the two sexes (Shaklee 1983). Males always carry at least one slow allele, while females rarely have this variant. The common and slow variant alleles were pooled for the statistical analyses, and only the frequency of the fast allele is reported. With pool- ing, a test for fit to Hardy-Weinberg is possible, and no significant deviation from Hardy-Weinberg was observed. 716 Fishery Bulletin 88(4), 1990 Table 2 Average per locus and mean heterozygosity over all loci for spiny lobster Panuiirus marginatus from Maro Reef and Necker Island. Location and date EstS EstD Gpi Mpi' Pgm Tpepl Tpep2 Mean- Maro Reef 1987 Necker Island 1987 Hawaiian Archipelago 1978-80^ 0.061 0.015 0.059 jtribution (Shakl monomorphism lollow (1984). 0.000 0.000 0.032 ee 1983). at the 39 0.059 0.087 0.07(1 additional loci 0.381 0.477 0.374 surveyed by 0.045 0.356 0.030 0.389 0.042 0.370 Shaklee and Samollow (1984). 0.049 0.039 0.026 0.021 0.023 0.021 ' Sex restricted allele di -1987 estimate assumes 'From Shaklee and San Chi-square analyses Contingency chi-square analyses were performed between the two newly sampled collections at each of the polymorphic loci. Because of their low frequencies, all variant alleles at each locus were pooled into one class. Significant heterogeneity between Necker Island and Maro Reef was observed for E»tS (/- = 5.61, df = 1, P<0.020). Contingency chi-square analyses were also per- formed between the data gathered in this study and those of Shaklee and Samollow (1984) at each of the polymorphic loci. Significant heterogeneity was ob- served between the Necker Island 1987 and 1979-80 collections at E^tS (x" = 5.86, df = 1, P<0.025). Shaklee and Samollow (1984) reported fluctuating frequencies over years at the EstD locus and concluded that the 1980 samples represented a cohort with an unusual frequency. Therefore, we tested for hetero- geneity between our data and the 1979 and 1980 EstD data separately. Both Maro Reef and Necker Island were significantly different for the 1980 comparison (Maro Reef, r = 9.71, df = 1, P< 0.005; Necker Island, X- = 15.42, df = 1, P<0.001). The Necker Island 1987 data were significantly different from the 1979 data (r = 6.10, df = 1, P.> c.M»< »i S^^i.i Surt 3 i • 3 I 4 1 £ i I I i } i s 1 1 1 1 1 i i 1 1 f7 57 875- Z 5 ? JO ^ /V 5M/ 50 n.i6 0 I / / 3 5 SAfJ 6/ ZitO 300 pm 1 1 1 ZV2 H:30 5 JO c 1 1 \ 1 )( r > r r 1 f 1 • ^ ^ ^- / ' -»• 1 1 1^ f 1 ^ * ¥e*thft 1 C*l* ic m>4ttnt •(••. cit.r. vt«ib»f ub1.»*1, io i.i.ff^f -.ib offibot* riibio| Set* Ui* ib*o "■ I«»t Codei : Bo vied* •! Slid fo|. u m sola nderi King mackerel 543 1.2 201 1.2 White spotted soapfish 9 0.0 — — Scomberomorus en valla Ryptieus maculatus Ladyfish 29 0.1 3 0.0 Yellowfin tuna 5 0.0 4 0.0 Elops saurus Thunnus albacares Total 44 465 16762 726 Fishery Bulletin 88(4). 1990 Table 5 Nearshore, offshore trolling and offshore bottom fishing CPUE of angler < md charterboa operator (CBO) logbook partici] lants. 1987-88, near oil and gas platforms off Louisiana. 1987 1988 Anglers CBO Anglers CBO Species/group Mean (CV) Mean (CV) Mean (CV) Mean (CV) Nearshore H = 174 n = 109 n = H n = 4 Bluefish 0.24 (327) 0.03 (491) 0.22 (256) 0.36 (132) Cohia 0.03 (460) 0.01 (550) - — 0.01 (200) Hardhead catfish 0.03 (593) - — 0.05 (526) — - Red di'uni 0.10 (330) 0.33 (327) 0.18 (189) 0.56 (98) Sharks 0.04 (509) — — 0.02 (445) — — Silver/sand seatrout 0.55 (374) 0.14 (348) 0.63 (288) 1.40 (200) Spanish mackerel 0.03 (980) — — 0.02 (547) — — Spotted seatrout 3.94 (131) 5.48 (74) 3.33 (138) 0.83 (158) Offshore bottom fishing )( = 362 n = 578 n = 222 w = 2 1 1 Atlantic croaker 0.17 (031) 0.08 (330) 0,11 (452) 0.01 (698) Bluefish 0.30 (510) 0.11 (429) 0.11 (608) 0.01 (549) Blue runner 0.11 (1129) 0.02 (1061) 0.09 (546) — — Cobia 0.08 (360) 0.06 (323) 0.07 (342) 0.04 (258) Dolphin 0.05 (916) 0.02 (1207) 0.01 (1051) 0.01 (1087) Greater amberjack 0.30 (391) 0.25 (272) 0.24 (484) 0.19 (264) Grey triggerfish 0.28 (311) 0.14 (329) 0.39 (289) 0.19 (160) Grouper 0.23 (282) 0.22 (255) 0.21 (265) 0.30 (158) Hardhead catfish 0.11 (641) 0.06 (705) 0.01 (1490) 0.01 (908) King mackerel 0.(14 (364) 0.05 (335) 0.03 (626) 0.04 (299) Dtlier snapper (l.IO (377) 0.37 (245) 0.05 (321) 0.47 (183) Red drum 0.11 (385) 0.07 (762) 0.06 (462) — - Red snapper 2.83 (171) 2.11 (155) 2.12 (248) 2.53 (148) Sharks 0.06 (556) 0.12 (431) 0.04 (608) 0.05 (304) Silver/sand seatrout 0.82 (388) 0.32 (298) 0.30 (256) 0.01 (693) Spanish mackerel 0.08 (615) 0.03 (1013) 0.02 (887) - - Offshore trolling » = 130 « = 33 ri = 76 n = 1 4 Blue runner 0.34 (308) 0.03 (441) 1.02 (229) 0.03 (374) Ccibia 0.02 (397) O.OS (266) 0.01 (699) 0.02 (296) Crevalle jack 0.07 (398) 0.10 (446) 0.05 (387) — — Dolphin 0.26 (372) 0.71 (236) 0.25 (502) 1.23 (167) tireater amberjack 0.12 (598) 0.09 (377) 0.09 (356) — — King mackerel 0.11 (623) 0.62 (151) 0.25 (232) 0.17 (171) Little tunny 0.43 (222) 0.85 (209) 0.78 (l(i2) 0.21 (241) Other jacks 0.07 (623) 0.08 (400) 0.02 (872) — — Spanish mackerel 0.51 (271) 0.64 (222) 0.84 (227) 0.42 (288) Discussion Private vessel anglers and charterboat operators caught a total of 61227 fish representing over 46 dif- ferent species. The fishing habits and boat character- istics of study participants were as expected. Peak fishing by private vessel anglers and charterboat operators occurred during the summer which agreed with past studies by Ditton and Auyong (1984) and Stanley and Wilson (1989). During other times of the year, cold fronts and other severe forms of weather make fishing off the coast of Louisiana nearly impos- sible or at least uncomfortal)le. Based on number of trips, the most prevalent fishing method was offshore bottom fishing, with nearshore and offshore trolling the next most popular methods for the study participants. Private vessel anglers and charterboat operators utilized the entire range of platform sizes and opera- tional types off the coast of Louisiana. Single well structures, steel template platforms, and mobile semi- submersible drilling platforms were commonly fished, although certain trends in platform size utilization and fishing method were apparent. Nearshore anglers most often fished at the small, single well structures Stanley and Wilson Fish populations around oil and gas structures off Louisiana 727 Table 6 Nearshore and offshore trolling CPUE of logbook participan s around oil and gas platforms in the Delta, Bay, and Cameron Regions | off Louisiana, 1987-88. Delta Bay Cameron 1987 1988 1987 1988 1987 1988 Species/group Mean (CV) Mean (CV) Mean (CV) Mean (CV) Mean (CV) Mean (CV) Nearshore n = 48 H=8 n = 196 H = 46 tt = 37 )i = 12 Bluefish 0.68 (197) 1.28 (64) 0.04 (537) 0.02 (678) 0.14 (324) 0.22 (234) Cobia 0.06 (259) 0.03 (144) 0.01 (947) — - 0.02 (370) — - Hardhead catfish — — — — 0.03 (606) 0.07 (444) — — — - Red drum 0.99 (169) 0.80 (49) 0.02 (660) 0.11 (249) 0.39 (258) 0.07 (186) Sharks 0.01 (495) 0.03 (494) 0.03 (632) 0.03 (439) 0.05 498) — — Silver/sand seatrout 0.25 (326) 0.08 (282) 0.38 (444) 0.85 (249) 0.53 (226) 0.55 (293) Spanish mackerel 0.09 (557) — — 0.02 (608) 0.01 (475) 0.02 (608) 0.07 (346) Spotted seatrout 1.73 (125) 0.48 (70) 5.70 (87) 4.28 (119) 2.02 (97) 1.29 (83) Offshore trolling ri = 78 M = 50 n = 23 H = 13 n = 63 H = 27 Bluerunner 0.46 (260) 0.71 (230) 0.20 (425) 1.116 (234) 0.09 (441) 1.06 (269) Cobia 0.04 (372) 0.02 (592) 0.03 (303) - — 0.03 (369) 0.01 (388) Crevalle jack 0.01 (590) — — 0.07 (268) 0.07 (361) 0.15 (311) 0.10 (263) Dolphin 0.20 (276) 0.48 (363) 0.68 (298) 0.15 (177) 0.40 (312) 0.39 (298) Greater amherjack 0.05 (630) — — 0.14 (315) 0.27 (158) 0.18 (509) 0.12 (.3,50) King mackerel 0.04 (455) 0.05 (.506) 0.07 (209) 0.29 (16.5) 0.49 (157) 0.55 (140) Little tunny 0.69 (20.5) 0.73 (173) 0.14 (215) 0..50 (174) 0.43 (199) 0.72 (174) Other jacks 0.08 (625) 0.03 (707) 0.14 (381) — - 0.05 (498) — - Spanish mackerel 0.25 (363) 0.24 (262) 0.28 (270) 0.46 (302) 1.06 (196) 1.91 (145) Table 7 Offshore bottom fishing CPUE of logbook participants around oil and gas platforms in the Delta, Bay, and Cameron Regions off | Louisiana, 1987-88. Delta Bay Cameron 1987 1988 1987 1988 1987 1988 Species/group Meat 1 (CV) Mean (CV) Mean (CV) Mean (CV) Mean (CV) Mean (CV) « = = 627 n = 280 H = 196 >i = 77 n = 153 n = 75 Atlantic croaker 0.16 (518) 0.10 (485) 0.02 (694) — — 0.04 (393) — — Bluefish 0.23 (592) 0.02 (539) 0.11 (422) 0.25 (437) 0.08 (689) 0.02 (672) Cobia 0.05 (367) 0.05 (,362) 0.14 (310) 0.08 (282) 0.08 (242) 0.05 (273) Greater amberjack 0.30 (331) 0.11 (487) 0.18 (325) 0.15 (235) 0.25 (224) 0.65 (275) Grey triggerfish 0.26 (287) 0.25 (217) 0.07 (571) 0.50 (288) 0.09 (386) 0.23 (391) Grouper 0.26 (238) 0.30 (183) 0.16 (416) 0.20 (283) 0.15 (277) 0.15 (230) King mackerel 0.02 (449) 0.03 (384) 0.07 (198) 0.02 (3.30) 0.14 (224) 0.07 (439) Other snapper 0.40 (240) 0..56 (217) 0.13 (508) 0.55 (329) 0.07 (336) 0.14 (542) Red drum 0.10 (524) 0.02 (876) 0.09 (633) 0.02 (445) 0.05 (338) 0.08 (.356) Red snapper* 2.33 (187)* 2.27 (162)^'' 2.62 (124)"'' 2.72 (247)"'' 2.85 (132)" 2.19 (237)'' Sharks 0.11 (421) 0.05 (428) 0.03 (379) 0.04 (591) 0.08 (596) 0.02 (301) Silver/sand seatrout 0.68 (369) 0.20 (328) 0.17 (.383) 0.15 (,363) 0.12 (460) 0.02 (699) Spanish mackerel 0.02 etter art (1153) not signi 0.01 (1471) 'icanlly differei 0.12 (536) t at the \Vn level. 0.01 (.560) 0.12 (498) 0.02 (442) * Means with the same in shallow water, while offshore bottom fishing and trolling anglers fished much larger steel template platforms in deep water. These results provide evi- dence that participants were maximizing their catch potential, as Stanley (1989) found highest abundances of spotted seatrout near small platforms in shallow water, while highest abundances of red snapper and related species were found near largp platforms (sub- 728 Fishery Bulletin 88(4), 1990 merged surface area 8000-14000 m^) in 70-100 m of water. Charterboat operators had larger vessels than pri- vate vessel anglers due to the business nature of their fishing and larger party size. Using larger vessels, they were able to fish in deeper waters farther offshore than private vessel anglers for both offshore bottom fishing and trolling. Charterboat operators also fished in deeper water while fishing nearshore; however, this is probably not a direct function of boat size, but of past success and preference. A high diversity of fish exist around the oil and gas platforms as evidenced by the reported catch of over 46 different species. The types of fish caught ranged from relatively common and highly desirable species such as spotted seatrout, red snapper, tarpon, blue marlin, king mackerel, and yellowfin tuna to rather rare fishes such as hake, bearded brotula, and squirrel fish. However, catches by angling are selective and biased towards larger, carnivorous individuals due to the gear utilized (Grimes et al. 1982). Therefore, species not susceptible to angling were not represented. Comparison of CPUE estimates between this study and logbook programs from other parts of North America revealed CPUE estimates were generally much higher off the Louisiana coast than from other studies. CPUE estimates from Sztramko (1986) for Lake Erie, Ontario, private vessel anglers and charter- boat operators were approximately 0.15 for the target species of smallmouth bass Micropterus dolomieui and coho salmon Oncorhynchus kisufch. while Casselman et al. (1985) reported CPUE estimates of 0.30 for northern pike Esox lucitis by sportfishing guides on the St. Lawrence River, Ontario. Only CPUE estimates from logbooks maintained by commercial fishermen trolling for Pacific salmon (Jordan and Carter 1987) were as high as CPUE estimates from this study. Catch rates in this study, presented as number offish per angler per hour, are not directly or statistically comparable with those reported by Brusher et al. (1984), Brusher and Palko (1985, 1987), and other earlier studies from the Gulf of Mexico because they used number of fish caught per boat per hour (CPH) as a unit of relative abundance. Since the number of anglers on a sportfishing vessel can be highly variable, catch per vessel hour may also be highly variable. By calculating the number of fish caught per angler hour, catch is broken down to its most standard unit, assum- ing the skill level of anglers is equal, thus eliminating this source of variance. CPH rates by bottom fishing and trolling from this and other studies off the Louisiana coast were much higher than in other regions of the Gulf of Mexico and the southeast Atlantic. However, when mean number of anglers was multiplied by CPUE for charterboat operators from this study and compared with Brusher et al. (1984) and Brusher and Palko (1985, 1987), similar catch rates were noted for the Louisiana coast. Comparison of the catch rates and composition of trolling from this study and others revealed few dif- ferences. Brusher et al. (1984) and Brusher and Palko (1987) found trolling catches of charterboat operators off Louisiana were primarily comprised of dolphin, Spanish mackerel, red drum, little tunny, king mack- erel, and blue runner. We found little change in this trend with the exception of red drum. Based on our results, red drum do not appear to be associated with oil and gas structures. Other logbook programs off the Louisiana coast did not distinguish between areas fished (near oil and gas platforms or otherwise) (Dugas et al. 1979), and consequently red drum catches may have been high in areas not covered by our logbook program, which may explain their absence from our results. When catch rates and composition of our study were compared with those of other logbook programs and a creel census for offshore bottom fishing, major dif- ferences in catch composition were noted. Past studies conducted during 1978 by Dugas et al. (1979), in 1982 by Brusher et al. (1984), and during 1984-85 by Brusher and Palko (1987), found that Atlantic croaker and silver/sand seatrout dominated catches of charter- boat operators while we found that red snapper catch rates were often an order of magnitude greater than for all other species caught. Since many of the same charterboat operators were utilized by both ours and earlier studies, these differences may provide evidence that there has been a shift in the species abundance near oil and gas platforms in offshore waters with little or no change in the composition of the structure (i.e., the number of oil and gas platforms) off the Louisiana coast. Comparison of our nearshore catches around oil and gas platforms with those of creel surveys of Texas bay charterboat operators from 1978 to 1979 (McEach- ron and Matlock 1983) showed that spotted seatrout dominated the catch rate and composition of nearshore Texas fishermen as they did in Louisiana, although spotted seatrout catches off Louisiana were much higher than for Texas sport fishermen. Overall CPUE's of all species between private vessel anglers and charterboat operators were very similar; however, charterboat operator CPUE was generally less variable. Few trends were identified from com- parisons between years or user groups. Only for spotted seatrout was charterboat operator CPUE con- sistently higher than private vessel angler CPUE; otherwise few differences were noted between the two groups. The similarity of the CPUE's indicated that the anglers participating in the logbook program were avid. Stanley and Wilson Fish populations around oil and gas structures off Louisiana 729 knowledgeable fishermen who were as skilled as the professional charterboat operators. The CPUE of the study participants was probably much higher than that for average offshore saltwater fishermen in Louisiana. Since charterboat operators are professional fishermen who make their living by catching fish, and the private vessel anglers partici- pating in the program were equally as skilled, the catch rates by these groups do not reflect average catches. Both groups have a higher success rate than do casual fishermen who rarely catch fish due to improper tech- niques and do not target their effort (Casselman et al. 1985). Consequently, the CPUE's presented are not applicable to all fishermen, only to skilled, dedicated, amateur anglers and charterboat operators. Spotted seatrout and red snapper were the target species of private vessel anglers and charterboat operators, as indicated by their domination of respec- tive CPUE estimates for nearshore and offshore bot- tom fishing and based on past research from a salt- water recreational angling survey (Stanley and Wilson 1989). Casselman et al. (1985) reported that the CPUE of avid recreational fishermen can be used as an index of relative abundance, and if data are collected over a long period of time, changes in CPUE can reflect fluc- tuations in the populations of the target species. Off- shore trolling CPUE was not dominated by a single species, indicating that fishermen were not targeting their effort for any particular species; therefore, CPUE estimates while trolling would not reflect changes in abundance to the same extent as nearshore and off- shore bottom-fishing CPUE estimates. Conclusions The catch rates of private vessel anglers and charter- boat operators around oil and gas platforms in the northern Gulf of Mexico were high, with their effort targeted towards red snapper while offshore bottom fishing, and spotted seatrout while nearshore fishing. Catch rates while trolling were more uniform and not dominated by any one species. Catch rates between the user groups and across the regions were similar, al- though some minor differences were detected. Based on comparisons of catch rates and composi- tion between our results and past studies, a shift in the abundance of certain species has occurred near oil and gas platforms off the Louisiana coast. Studies from 1978 to 1985 found Atlantic croaker and silver/sand seatrout constituted the largest portion of the catch and had the highest catch rates, while we found that red snapper dominated the catch statistics while offshore bottom fishing. Little change in the patterns of species abundance appears to have occurred for species tar- geted while offshore trolling and nearshore fishing around oil and gas platforms, based on comparisons with earlier studies. The physical construction of oil and gas platforms precludes the sampling of the associated sportfish populations using traditional methods (e.g., gillnets, trawls). The success of this logbook program indicates that the collection of CPLIE data over long periods of time may be an effective technique of monitoring fish populations associated with the platforms. Although the data supplied by the logbooks is an index of relative abundance of fish susceptible to angling and is biased towards larger individuals, it provides a valuable source of data which is otherwise difficult to obtain. Acknowledgments The authors thank the 55 anglers and 10 charterboat operators for maintaining logbooks over the study period. Dr. James Geaghan for statistical counseling, and Dr. Linda Jones and two anonymous reviewers for improving the quality of the manuscript. Citations Brusher, H.A.. and B.J. Palko 1985 Charter boat catch and effort from southeastern U.S. waters, 1983. Mar, Fish, Rev. 47(3):.54-6fi. 1987 Results from the 1984 and 1985 charter boat surveys in southeastern U.S. waters and the U,S, Caribbean Sea. Mar. Fish, Rev. 49(2):109-117, Brusher, H.A.. M.L. Williams, L. Trent, and B.J. Palko 1984 Using charter boat catch records for fisheries manage- ment. Mar, Fish, Rev. 46(3):48-.56, Casselman, J.M., M.A. Henderson, and T. Schaner 1985 Fish sampling techniques — natural and observational variability. Contril). No, 85-00, Ontario Minist. Nat, Resourc. Fish. Branch, Res. Sect,, Maple, Ontario. Canada. 52 p, Coleman, E. 1984 Coastal Louisiana: Climate and recreation, Louisiana State Univ. Sea Grant Prog.. Baton Rouge, 24 p. Continental Shelf Associates, Inc. 1982 Study of the effect of oil and gas activities on reef fish populations in the Gulf of Me.xico OCS area. OCS Rep. MMS82-010. U.S, Dep, Int., Minerals Manage. Serv.. Gulf of Me.xico OCS Region. New Orleans. 210 p. Demory, R.L.. and J.T. Golden 1983 Sampling the commercial catch. In Nielson, L.A.. and D,L, Johnson, (eds.), Fisheries techniques, p, 421-430, Am. Fish. Soc. Bethesda. MD. Ditton, R.B., and J. Auyong 1984 Fishing offshore platforms central Gulf of Mexico: An analysis of recreational and commercial fishing use at 164 major offshore petroleum structures. OCS Mon200 of either species were captured. King crabs were measured from the rear of the right eyesocket to the midpoint of the rear margin of the carapace (carapace length, CL). Tanner crabs were measured across the widest portion of the carapace, including spines (carapace width, CW). Shell conditions were coded as 1 = molting, or recently molted with a soft or flexible shell; 2 = hard shelled but molted within recent year, without epifauna or scratches; 3 = oldshell or skipmolt, with scratches, epifauna, scars, or dam- aged dactyls. (B) Vitality coding For most liauls, all measured crabs were also given a vitality code on a 3-point scale: (1) alive and active, i.e., exhibiting spontaneous vigor- ous movement of appendages and attempting to escape or pinch the examiner; (2) moribund (alive but inactive), i.e., exhibiting only slow weak movements, or only in response to prodding; and (3) dead. In practice, death was not easy to determine, as some crabs which initially appeared dead would exhibit spontaneous movements after examination. After the first few tows, a criterion was developed which consisted of (a) looking for spon- taneous movement of ai)pendages, (b) lacking that, at- tempting to stimulate movement by moving and bend- ing appendages, followed by (c) flicking the mouthparts with a finger three times at 5-10 second intervals. The minimum requirement for a vitality code of 2 was any spontaneous movement of the antennules or any other appendage. These criteria were tested by placing 30 crabs which were classified as dead into the live tanks; none recovered. Presence or absence of injuries was not a criterion used in this coding procedure. (C) Injury assessment Approximately 30-40 king crabs per tow (usually the first basket examined of each sex) and occasional Tanner crab were examined for in- juries. Since there was considerable shuffling and repositioning of crabs and baskets during the weighing and subsampling process, this was assumed to be a ran- dom sample, and represents a subset of the crabs ex- amined for vitality in step B above. Examination en- tailed a visual scan of the carapace, sternum, abdomen, and each leg in turn. Bodily injuries were recorded as present or absent. Leg injuries were categorized as being distal to the autotomy or breakage plane, prox- imal to the breakage plane, or recent autotomy, dis- tinguished by the presence of a clear membrane with Stevens- Survival of king and tanner crabs captured incidentally in the Bering Sea 733 no discoloration or scarring. The number of legs with each of these three types of injuries was recorded. (D) Survival or delayed mortality A subset of the injury-assessed crabs (Step C) was placed into one of four tanks of flowing seawater, ~ 1.0 x 1.0 x 0.7 m, and held for 48 hours, after which they were reexamined for delayed mortality, using the same procedures outlined above for vitality. The subset usually consisted of all crabs coded for injures from one haul per day, although towards the end of the experiment several hauls per day were treated in this manner. Crabs which were determined to be dead before placement into the tanks were retained for at least 6 hours for confirma- tion before removal. Crab processing time was recorded for each tow as the interval between the arrival of the codend on deck and the time the crabs were examined and measured. Although examination usually required anywhere from 0.5 to 1.5 hours for an entire sample, the midpoint of that period was used as the mean endpoint of process- ing time for all crabs in a haul. Towing time, i.e., the difference between the time the net was determined to be on bottom and the time haulback started, was obtained from each catcher vessel and divided by 2, in order to approximate the average time each crab spent in the net before delivery to the Svlak. Total captur- ing and processing time (CAPTIME) was calculated as the sum of (towing time)/2 and processing time. Total weight of the catch in each haul was estimated by measuring the height and width of the codend at several intervals along its length, and calculating the volume based on the shape of an oblate cylinder, with density determined by weighing a known volume of fish. Data analysis In order to improve the precision of the survival esti- mates, overall survival was calculated for each species in a stratified manner (Cochran 1963). Vitality codes were used as strata because they provided additional information on the probability of survival, i.e., crabs coded as alive and active were deemed more likely to survive than crabs coded as moribund. For immediate survival estimates only, each crab was weighted by the sampling factor for the tow from which it originated. Estimates of delayed mortality were not weighted, because subsampling was accounted for at the level of immediate mortality. The formulae used were: S„ = 1/N I N,p, (1) where Sst = stratified estimate of population sur- vival rate. V(Sst) = variance of stratified estimate, Nv = number of crabs in stratum (vitality group) V, nv = number sampled from stratum v (placed in live tanks), Pv = proportion of subsample surviving in live tanks, Cjv = 1 - Pv , N = weighted total number of crabs given vitality codes. The effect of CAPTIME on immediate mortality was determined by fitting the data to a logistic curve. Crabs were grouped into hourly intervals of CAPTIME (as described above), and mortality data (weighted by sampling factors) for each interval were fitted using an iterative procedure (FSAS FISHPARM, Saila et al. 1988). The equation fitted was: M = l/{l + exp(-r[X-X5„])} (3) and V(S,t) = 1/N- I N,(N, - n,)p,q,/(n, - 1) (2) where M is expected mortality and 0- 0.6 0.4 D a/ a b'- 0.876 cr / ° o Uj/ 0.2 y^ D i; OBSERVED LT50 • 9 34 hr PREDICTED 0.0 , □ ? — , — , — y, LT50 10 12 14 16 18 20 HOURS B TANNER CRAB Figure 2 Logistic regression of immediate mortality vs. total time spent in captivity. Crabs were grouped into 1-hour intervals before analysis. (A) 6 1 13 observations representing 7487 individuals. (B) 4645 obser- vations representing 16498 individuals. Squares are observed data, solid lines represent predicted values. tr o tr O 2 A KING CRAB □ OBSERVED ° PREDICTED n - LT50 D - R ^- 0.245 D D ^^ D ^^^"^^^ □ "^ a a no ^ LW50 • 21.6 t o D e , , ,— 10 15 20 25 30 35 40 TONS B TANNER CRAB D D 0-8 06 0 D 0.4 ° D D D 0 2 n a D nn n 0 5 to 15 20 25 30 35 40 TONS Figure 3 Logistic regression of immediate mortality vs. total weight of catch. Crabs were groujied into 1-t intervals liefore analysis. Squares are observed data, solid lines represent predicted values, which were incalculable for B. had been tagged, held for observation, and released was recaptured with its tag still attached. Others may have been recaptured as well, but since most of the tags were removed prior to discarding the crabs, they could not have been identified as recaptures. Effects of time in captivity and weight of catch CAPTIME, the total time in captivity (prior to vitality/ injury assessment) ranged from 3 to 17 hours (Fig. 2). The logistic fitting procedure indicated a significant relationship (/■- = 0.878, « = 13, a<0.0()l) between CAPTIME and immediate mortality of king crabs (Fig. 2A); the relationship for Tanner crabs (Fig. 2B) was weaker but still significant (r- = 0.603, >i=ll, o = 0.05). The LT,=,o's (time required for 50% immediate mortality) were 9.3 hours for king crabs and 8.3 hours for Tanner crabs. Haul weights ranged from 4 to 39 t (^ = 20.4 t). The relationship between haul weight and immediate mor- tality for king crabs (Fig. 3A) was poor and non-sig- nificant (r- = 0.245). The relationship for Tanner crabs was so poor (Fig. 3B) that the iterative fitting pro- cedure could not converge on a solution. Shell conditions Soft shell or molting crabs (shell condition 1) accounted for only about 1.1% of male king crabs, and <().\% of females. All other females and most males were hard- shell crabs (condition 2). About 9.4% of males were oldshell crabs (condition 3), and virtually all of these were above 100 mm CL. During the NMFS summer EBS survey in June of 1987 (Stevens et al 1987). 2.8% of males were soft or molting and many others were hardshell but recently molted. Immediate survival odds (ratio of survivors to deaths. Fig. 4) increased signif- icantly (X- = 81.63, p<0.001) with increased shell con- dition, i.e., hardness, but delayed survival was indepen- dent of shell condition (X- = 0.60, />>0.74). Stevens: Survival of king and tanner crabs captured incidentally in the Bering Sea 737 ■g 2.0 0) o ^ .a 0) > 3 1.0 g < 0.5 IMMEDIATE 81.63 " DELAYED 0.60 ns SHELL CONDITION Figure 4 Effects of shell condition on immediate and delayed survival odds of king crabs. Condition coded as 1 = soft, 2 = new hardshell, 3 = old hardshell. Immediate effects are weighted by sample factors, delayed effects are not. *** indicates X- value significant at p<0.001. Sample size indicated above bars. CO 4) a E ^ 0.. KING CRAB TANNER CRAB 61-75 76-90 91-105 106-120 121-135 136-150 CARAPACE SIZE (mm) 3 > O KING CRAB 61-75 76-90 91-105 106-120 121-135 136-150 CARAPACE LENGTH (mm) Figure 5 Effects of size on survival. (A) Immediate survival of king and Tan- ner crabs by 15-mm intervals of carapace length (king crabs) or carapace width (Tanner crabs). Numbers represented by each interval are 226-1520 king crabs and 71-9752 Tanner crabs. (B) Overall survival of king crabs, calculated as the product of immediate and delayed survival estimates. LU CI C2 PAN CHU STD C1 02 PAN CHU STD ^ KING CRAB -^ -N= ■^ TANNER CRAB ^ dp -B- — r 1 0.2 0.4 0.6 IMMEDIATE SURVIVAL Figure 6 Effect of gear type on immediate survival of king and Tanner crabs. Vertical bar represents the mean; outer ends of box represent up- per and lower confidence intervals, defined as 1.96(pq/n)"". All female and most male Tanner crabs were har(i- shell (condition 2). Less than 0.3% of males were old- shell (condition 3), and less than 0.01% were softshell (condition 1). Low incidence of softshell and oldshell Tanner crabs prevented contingency table analysis due to too many empty cells. Size effects Immediate survival of king crabs did not vary much over the range of sizes captured (Fig. 5A). For Tan- ner crab, however, immediate survival increased slight- ly with size. Overall survival of king crabs (Fig. 5B) decreased markedly at sizes above 120 mm CL, most- ly as a result of increased delayed mortality. Delayed mortality and overall survival could not be determined for Tanner crabs because of the limited number of observations. Effects of gear type Since survival was strongly related to CAPTIME, an analysis of variance was performed to determine if CAPTIME differed between the nets used. No signifi- cant differences were found (F = 0.570, a = 0.685, df = 4181), so any differences in survival between nets can be attributed directly to the nets. Immediate survival was least for the standard net (STD), for both species (Fig. 6). Nets C2 and CHU pro- duced the best survival for king crabs, whereas CI and PAN produced better survival for Tanner crabs. Overall, the experimental nets produced no clear dif- ferences in survival from the control nets for either species of crab. Extremely poor survival of Tanner crabs occurred in the standard net, and was probably the result of a biased vitality sample, which included 738 Fishery Bulletin 88(4), 1990 T3 (0 I cc 1 ^H NO INJURY HH BODY ONLY cm LEGS ONLY ^ BODY a LEGS 1004 2276 27 367 24 KING CRABS TANNER CRABS Figure 7 Effects of body and leg injuries on immediate survival odds (ratio of number surviving to number of deaths) of king and Tanner crabs. Sample size indicated above bars. Only injury-assessed crabs were used. only six tows with tliis gear type, three of which had large sampling factors and mostly dead Tanner crabs. Delayed mortality was not compared between nets due to limited data. Effects of injuries A weighted number of 3368 king crabs and 1421 Tan- ner crabs were assessed for injuries. Figure 7 shows the survival odds (ratio of number alive to number dead) within each of four combinations of injury types: (1) none, (2) legs only, (3) body only, and (4) legs and body, calculated only for injury-assessed crabs. The ma- jority of crabs suffered no detectable injuries, and had the best survival odds of 1.8 for king crabs and 4.2 for Tanner crabs. The most common type of injury for all crabs was leg injuries, which resulted in survival odds of 1.14 and 0.63 for king and Tanner crabs, respective- ly. Injuries to the body and combined body and leg in- juries were less frequent, resulting in survival odds of 0.98 and 0.44, respectively, for king crabs, and 0.59 and 0.71 for Tanner crabs. Body injuries were more frequent but less serious among king crabs than Tan- ner crabs and were predominantly broken spines, which are more abundant and prominent on king crabs and thus more likely to be damaged. Body injuries of Tan- Table 3 Contingency table analysis of interactions between delayed mortality, vitality, and row and column factors were independent. Data were taken only from survival exper of crab in each category; numbers in parentheses are calculated expected values or 2x2 tables with 1 degree of freedom, and all are significant at p< 0.005 (**). injuries of king crabs. Null hypothesis was that ment for king crab. Values are observed numbers percent of grand total in row or column. All are Mortality Vitality Row X- value Active Moribund Sum (%) Vitality vs. delayed mortality Lived Died Column sum (%) Injuries vs. delayed mortality Lived Died Column sum (%) Vitality vs. injuries Active Inactive Column sum ('Mi) 124 (64) 10 (70) 1.58 (218) 297 (237) 282 307 (47.9) (52.1) Total Total Total = 589 = .589 = 589 138. 6" df= 1 16.46" df= 1 9.77'* df= 1 1.34 (22.8) 455 (77.2) Injuries Row Uninjured Injured Sum (%) 241 (221) 220 (240) 41 (61) 87 (67) 282 307 (47.9) (52.1) 461 (78.3) 118 (105) 343 (356) 128 (21.7) 16 (29) 112 (99) 134 455 (22.8) (77.2) 461 (78.. 3) 128 (21.7) Stevens Survival of king and tanner crabs captured incidentally in the Bering Sea 739 ner crabs were typically broken or crushed carapaces which were associated with lower survival. In general, injuries were present in a greater proportion of dead crabs than live crabs. Combined body and leg injuries were most fatal for king crabs, whereas leg-only in- juries accounted for the greatest number of dead Tan- ner crabs. The low survival odds for uninjured king crabs suggests that suffocation in the net or bunkers may have contributed strongly to their mortality. Contingency table analysis (Table 3) showed that significant interactions occurred between vitality (VIT) and delayed mortality (DMORT; X- = 138.6), indicat- ing that VIT was an excellent predictor of future sur- vival or mortality for those king crabs held in survival tanks. Similarly, interaction between presence or absence of injuries (INJ) and DMORT (X^ = 16.46) in- dicated that significantly more injured crabs died than were expected to, and that between VIT and INJ (X- = 9.77) suggests that injuries were partly respon- sible for observed vitality levels. Loglinear analysis with DMORT as the dependent variable, and the interactions of DMORT with both VIT and INJ as independent variables (Table 4) indicated that the overall odds of survival were 2.087:1. These odds were increased by a factor of 4.713 for crabs coded as active (VIT = 1), and by a factor of 1.395 for crabs without injuries (INJ = 1). The total delayed survival Table 4 Parameter estimates and survival odds for each factor in the iogit model for king crabs. The dependent factor, DMORT, represents the mean survival odds (ratio of survivors to mor- talities) and has no alternative outcome. Each interaction term has two alternative outcomes; alternative 1 represents cells with equal indices, i.e., values for Vitality and Injuries both equal 1 or both equal 2 (Table 5. lines 1 and 4). and alternative 2 represents cells with unequal indices (Table 5, lines 2 and 3). The survival odds for active crabs with no injuries (Vital- ity and Injuries = 1) is the product of antilogs for DMORT, DMORT X VIT,, and DMORT x INJ,, or 13.73. Survival odds are equal to the antilogs of 2 (coefficient). Factor Coefficient Survival odds Alt. 1 Alt. 2 DMORT DMORT X VIT DMORT X IN,J Sum 0..368 0.775 0.167 0.736 1.550 0.333 2.087 4.713 1.395 0.212 0.717 :.619 13.73 0.317 odds for active, uninjured crabs eciualed 13.73, the product of all three values. Similarly, survival odds for inactive, injured crabs were 0.317, the product of the mean odds (2.087), the odds for inactive crabs (0.212), Table 5 Contingency tables used for calculating survival odds (ratio of survivors to mort Expected odds are calculated as the product of individual parameters from Table ~. odds. Numliers in parentheses are factor codes. (B| Cells combined to show ovei native grouping procedures. Procedure 1 used only the criterion of vitality. Proc placing only active, uninjured crabs in category 1. Procedure 3 used only prese ilities) of king crabs and parameters of Iogit model. (A) Cells of table shown separately, with expected all survival and survival odds calculated using alter- edure 2 simulated the NMFS observer procedure by nee or absence of injuries as the criterion. Cell category Observed Expected Odds Died Proportion surviving Odds Vitality Injuries Lived A Separate cells Active (I) (11 No Yes (1) 110 (2) 14 8 2 0.932 0.875 13.750 7.000 13.727 7.051 Morbid (2) (2) No Yes (1) 131 (2) 27 212 85 0.382 0.241 0.618 0.318 0.618 0.317 B Combined cells 1 VIT only All active All inactive 124 158 10 297 0.925 0.347 12.4 0.532 2 Observer procedure Active, uninjured Inactive or injured 110 172 8 299 0.932 0.365 13.750 0.575 3 INJ only All uninjured M\ injured 241 41 220 87 0.523 0.320 1.095 0.471 740 Fishery Bulletin 88(4), 1990 Table 6 Comparison of effects of body and leg injuries on immediate moi tality of king crabs. (A) Loglinear | analysis of immediate mortality (IMORT) as a function of body and leg injuries. Survival odds are equal to the antilog of 2(Coeff.), and are niultiplicabie. (B) Contingency table of data used for loglinear analysis. Numbers in parentheses are factor codes. Note that expected odds differ from observed odds, | but not significantly, x = 3.084, with df = = 1, and p = 0.08. \ Parameters of logit Factor model Coefficient x2 Antilogs Alt. 1 Alt. 2 IMORT -0.017 -0.034 0.966 IMORT by BODY 0.171 0.341 1.407 0.711 IMORT by LEGS Sum 0.127 0.255 1.290 0.775 0.562 1.755 0.532 B Contingency table. U = Uninjured, I = Injured Observed Expected Cell category Percent Body Legs Alive Dead survival Odds Odds U (1) U (1) 1428 828 0.633 1.725 1.755 I (2) 264 234 0.530 1.130 1.054 1 (2) U (1) 185 190 0.493 0.972 0.886 I (2) 68 1.52 0.309 0.447 0..5.32 Table 7 Results of cont mgency table analyses < f effet ts of types of leg injuries on immediate mortality (IMORT). Null hypothesit tested was independence o f each injury type vs IMORT. Expected ratio of percent live to lead crabs =1.0 foi all cases. Note that t( tals include uninjured as well as injured crab. ns = not significant at p = 0.05. ***sign ificant at p<0.001. Type of Number alive Percent of Number dead Percent of Ratio % dead injury with injury live crabs with injury dead crabs to % live X^ King crabs Autotomy 35 1.8 36 2.6 1.44 2.48 ns Distal 284 14.4 313 22.3 1.55 34.86** Proximal 29 1.5 113 8.1 5.47 88.06*** Total 1967 1401 Tanner crabs Autotomy 47 4.8 123 27.4 5.67 149.85*** Distal 112 11.5 136 30.3 2.63 76.16*** Proximal 11 1.1 35 7.8 6.89 43.63 ••' Total 972 449 and the odds for injured crabs (0.717). Alternatively, the log(ODDS) can be predicted as a linear function of each coefficient by substituting these values into Equa- tion 5, e.g., for active, uninjured crabs, ln(f,/f,„) = i^(l„ + l„.v + lin) = 2(0.368 + 0.77.5 + 0.167) = 2.619 ODDS = exp(2.619) = 1.S.74. Thus a 'good' vitality code of 1 was a better predictor of future survival odds than lack of injuries. Table 5A shows the observed frequencies, percent survival, and survival odds, as well as the survival odds predicted by the logit model, for each combination of factors. Expected odds were calculated as the product of odds for each combination of interaction represented in Table 4. For instance, the expected survival odds for active (VIT = 1), injured (INJ = 2) crabs equalled 7.05, the i)rciduct of odds for overall survival (2.087), active Stevens Survival of king and tanner crabs captured incidentally in the Bering Sea 741 crabs (4.713), and injured crabs (0.717). The likelihood ratio chi-sciuared value, calculated over all categories, was 2.30 with p = 0.317, indicating a good fit between the observed and expected survival odds. In order to compare different methods of crab vitality assessment as tools for predicting future survival, cells of Table 5A were combined. Table 5B shows that using only vitality codes as an indicator of future survival resulted in survival estimates of 0.925 for VIT = 1, and 0.347 for VIT = 2, identical to those shown in Table 2. Simulation of the NMFS observer program procedure resulted in survival estimates which were slightly higher for both categories (0.932 and 0.365, respective- ly), but otherwise very similar. Use of only data on presence or absence of injuries did not provide ade- quate information about future survival. Loglinear analysis was also conducted using imme- diate mortality (IMORT) as the dependent variable and the interactions of IMORT with body injuries (BODY) and leg injuries (LEGS) as the independent variables (Table 6A). The parameter for mean survival odds was 0.966, or about 1:1. Absence of body injuries increased the odds of survival by a factor of 1.407; their presence decreased survival odds by 0.711. Absence of leg in- juries increased survival odds by a factor of 1.290. Coef- ficients for IMORT X BODY and IMORT x LEGS were not significantly different. Expected odds differed slightly but not significantly (x" = 3.084, df = 1, p = 0.08) from observed odds (Table 6B). Evidence of recent leg autotomy was not significant- ly associated with immediate mortality of king crabs (X- = 2.48; Table 7A). However, leg injuries which oc- curred distal to the autotomy plane were significantly (X- = 34.86) associated with IMORT, as were those which occurred proximal to it (X- = 88.06). Proximal injuries had the greatest effect, increasing deaths above the expected proportion by a factor of 5.47. All three types of injuries were significantly associated with IMORT for Tanner crabs (Table 7B). Immediate survival odds, expressed as the ratio of live/dead crabs, decreased significantly with the number of legs injured for both king (X- = 86.5) and Tanner crabs (X- = 225.3; Fig. 8A). Delayed survival odds also decreased significantly with the number of injured legs for king crabs (X- = 17.0), but not Tanner crabs (X- = 4.8), perhaps due to inadequate sample size for the latter. Discussion The results of this study show that, although the ma- jority of crabs died as a result of capture incidental to commercial trawling operations, a significant propor- tion (>20%) can survive the process. The major fac- tors associated with increased crab mortality were shell A IMMEDIATE EFFECTS T3 ni « Q It O 1031 X ^Value 2629 1 , 66.5 ■■■ ^M KINO CRABS 225.3 ••* E5 TANNER CRASS ll 625 M 329 49 ■ 5 12 LEGS INJURED B DELAYED EFFECTS T3 n] V Q It I O cr 65 X Value KINQ CRABS 17.0 •• 4.78 nS ^ TANNER CRABS 52'1 I 84 7 LEGS INJURED Figure 8 Effects of multiple leg injuries on survival odds of l30 weakfish) prevented testing for interaction effects be- tween station and date. However, although not statis- tically significant, there was an apparent pattern: in both years fish were first abundant down estuary, and as the season progressed became more abundant fur- ther up the estuary (Fig. 3). In 1983, 845 fish were aged out of 993 collected, and in 1984, 361 fish were aged out of 571 collected, by the scale circuli method. An additional 98 fish were <14 mm SL (prior to scale formation), and ages were esti- mated by applying a quadratic regi'ession of known age on standard length from laboratory-reared weakfish (Szedlmayer et al. In press). Table 1 Juvenile weakfish CPUE (no. /2-minute tow) from the Chesa- | peake Bay and York River, VA, based on three-1 actor (main effects only) nonparametric ANOVA by rank abundance. Student Newman \euls test (SNK) was used to show specific differences, denot ed by different letters (0.05 level). Source of variation df MS F P>F Year 1 5138.7 16.90 0.000 1 Station 4 389.4 1.28 0.2870 Date 7 1216.4 4.00 0.0012 Error 61 303.7 CPUE Trawls SNK-test Date 16-29 July 1.2 20 C 30 July- 12 Aug 0.8 16 C 13-26 AuK. 5.0 60 A B 27 Aug.-9 Sept 7.1 51 A 10-23 Sept. 8.5 37.5 A 24 Sept-7 Oct. 5.1 52 A 8-21 Oct. 4.4 46 A B 22 Oct.-4 Nov. 1.6 49 B C Year 1983 6.5 152.5 A 1984 3.2 179 B 748 Fishery Bulletin 88(4). 1990 Jul Aug Sep Date Oct Figure 3 Standardized abundance of age-0 weakfish (CPUE = number/ 2-minute tow) by station and date in Chesapeake Bay-York River, VA. Multiple cohorts were apparent in lioth years. The birthdate frequency distributions sliowed three cohorts in 1983, and two cohorts in 1984 (Fig. 4). Cohorts were defined in 1983 as follows: cohort 1 hatched before 16 July, cohort 2 between 16 July and 15 August, and cohort 3 after 15 August. Another cohort may have been present in 1983 before 18 June, but these fish were relatively few in number and subsequently pool- ed with cohort 1. In 1984 two cohorts 4 and 5 were defined as those that hatched before 30 June and those after, respectively (Fig. 4). Weakfish growth rates were significantly different among cohorts within years, and between years (Table 2). Cohorts 1 and 4, both early-season cohorts, had the lowest growth rates and were not significantly dif- ferent from each other. Cohorts 2 and 3 had inter- mediate growth rates, and cohort 5 had the fastest growth rate compared with all other cohorts (Table 2). After separation of individual weakfish into their respective cohorts, a similar pattern of movement by each cohort was apparent: up the estuary with age and season (Figs. 5, 6). For example, the youngest fish in cohort 2 were first most abundant in the lower estuary (stations 2 and 3), and as the season progressed older fish from this cohort were more abundant up the estu- ary at the same location where few individuals of this cohort were collected earlier (station 4 in August vs station 4 in September 1983; Fig. 5). Cohort 4 showed the onlv difference from the above migration pattern. 30 n 15- > o c 0) 3 O a 01 15- 1984 N = 459 May Jun ' Jul Month Aug Sep Figure 4 Birthdate frequency distribution of age-0 weakfish cohorts in Chesa- peake Bay-York River, VA, based on a 3-day moving average. Numbers at the peaks of the distributions refer to cohorts (separated by open and shaded bars). Table 2 ( '.rowth rates (mm/day) of individual age-0 weakfish coliorts. Analysis of covariance of standard length on age with cohort as the covariate (0.05 level). Student Newman Keuls test (SNK-test) was used to show specific differences, denoted by different letters (0.10 level). Source if variat Cohort ion df MS F P>F 4 20178.39 1 597.8 (l.ll Aiie 1 448840.88 35540.4 0.(1 Ajre * ( 'oh( rt 4 1 686.94 133.6 (1.1) P^lrror 1294 12.63 Cohort Growth rate R- N SNK-test 1 0.83 0.9.5 383 A ■? 0.<)(! 0.9() 405 B ■i (i.!m; 0.94 57 B 4 ().S4 0.96 75 A o l.Olt 0.98 384 C Szedlmayer et al Growth of age-0 Cynosc/on regslis in Chesapeake Bay 749 no sample Sta t ion Aug Aug i Aug Sep J HJ^ i Sep llnnn Sep Sep lAk Oct Oct 21 60 21 60 2 Age mI i^ IlLin i J^ T 1 "1 I irn™ \ I J nlkin _E llpl nim ^ 4 dL 1 60 21 60 100 Days Figure 5 Age of weakfish estimated from scale circuli for 1983, versus frequency (number/2-minute tow), by station, date, and cohort in Chesapeake Bay-York River, VA. Cohort 1 = open bars, cohort 2 = solid bars, and cohort 3 = cross-hatched bars. The earliest recruits in cohort 4 were first collected at the upper estuary station (5), then appeared to spread down the estuary in subsequent collections (Fig. 6). An up-the-estuary movement with age pattern was also supported by the mean ages associated with each sta- tion, because weakfish were significantly older as the distance from the Bay mouth station (1) increased (Table 3), and by an increase in total catch-per-unit- effort up the estuary with season (Fig. 3). Different cohorts appeared to segregate habitats. In 1983, cohort 1 was more common further up the estuary compared with other cohorts, while cohort 2 dominated the mid- dle habitats (Fig. 5). Cohorts 3 and 4 were lower in abundance and it is difficult to suggest a pattern, while cohort 5 appeared dominant at all stations after August 1984 (Fig. 6). Discussion To our knowledge this is the first clear identification of multiple age-0 weakfish cohorts, and they showed significantly different growth rates and appeared to partition habitats. Earlier, Massmann (1963) suggested the existence of age-0 multiple cohorts from bimodal length-frequency distributions, but did not examine growth rates or differential habitat use. Except for two early studies (Buchanan-Wollaston and Hodgson 1929, Cooper 1937) reports of multiple cohorts in the juvenile stage of fishes are few (Shlossman and Chittenden 1981, DeVries and Chittenden 1982, Crecco and Savoy 1985, Kumagai et al. 1985, Eckert 1987, Isely et al. 1987, Wicker and Johnson 1987). This may be because of the difficulty of aging juvenile fishes (Geffen 1986, 750 Fishery Bulletin 88(4), 1990 10 0 5^ 5 c o 5 |o U- 5 0 21 I 60 Station Jul Aug Aug Aug Sep Sep J Oct Oct Nov n n rtimn ■ 1 -M— n 21 60 21 60 Age Days .MIMJ. 21 1 1 \ v- 60 100 Figure 6 A^t* of weakfish estimattni from scale circuli for 1984, versus frequency (iiumber/2-niinute tow). Ijy station, date, and cohort in Chesapeake Bay-York River, VA. Cohort 4 = solid bars, cohorts = open bars. Table 3 Comparison of weakfish age by station. pooled jver dates. based on ANOVA (0.05 evel). Student Newman Keuls test (SNK-test) was used to show specific differences, denoted by different letters (0.05 level). Source of variation df MS F F>F Station 4 37126.0 123.6 0.0001 Error 1299 300.3 Station Age A' SNK-test 1 26.1 104 A 2 50.2 248 B 3 57.1 34t; C 4 61.8 279 D 5 GG.S 327 K Essig and Cole 1986, Jenkins 1987, Post and Pranke- vicius 1987), or that fewer studies have examined postmetamorphic life stages because juvenile survival after metamorphosis was considered relatively con- stant compared with larval stages (Gushing 1975, Vic- tor 1986). Constant growth and mortality in postmeta- mori)hic juveniles have lieen questioned; for example. Wicker and Johnson (1987) showed a large increase in the rate of mortality in age-0 largemouth hass Microp- terus ^almiiides when juveniles shift from an inverte- brate to fish diet. Van der Veer and Bergman (1987) suggested that mortality due to predation by shrimp Crangon crangon on newly settled juvenile plaice Pleuronectes platessa may be significant and thus ac- count for the difficulty of predicting year-class abun- dance based on egg and larval surveys. However. Szedlmayer et al Growth of age-0 Cynosaon regalis in Chesapeake Bay 751 studies of American shad Alosa sapidissima suggested that year-class strength is estabHshed before the juve- nile stage (Crecco et al. 1983, Crecco and Savoy 1984, Crecco and Savoy 1985). Consequently, the importance of critical periods during the juvenile stage may be species-specific. Because several age-0 cohorts of weakfish showed variable growth rates and distribu- tion, survival of the juvenile stage of this species should not be assumed to have a constant rate. Different population parameters among cohorts are difficult to relate to salinity and/or temperature dif- ferences observed among stations. First, juvenile weak- fish are transient, as observed over the present study area and in earlier studies (Harmic 1958, Massman 1963, Chao and Musick 1977, Shepherd and Grimes 1983), and until accurate residency times can be esti- mated it may be ineffective to ascribe cohort differ- ences to particular habitat parameters. Second, other factors not measured in the present study, e.g., prey abundance, turbidity, currents, and predation may also be linked to cohort differences. In comparison with other juvenile fish, age-0 weak- fish appear to grow at an average rate. Juvenile growth rates derived from length frequencies for other sciaenid fishes were similar to our estimates for weakfish: C. arenarius (~1 mm/day, Shlossman and Chittenden 1981), C. nothus (0.8-1.3 mm/day, DeVries and Chit- tenden 1982). Shenker and 011a (1986) provide esti- mates of juvenile fish growth rates ranging from a low of 0.26 mm/day {Sehastes melanops) to a high of 4.7 mm/day (Coryphaena hipporus). Other growth rate estimates of juvenile fishes include: 1.5 mm/day for Anoplopoma fimbria (Boehlert and Yoklavich 1985), 1.0-1.3 mm/day for Chanos chanos (Kumagai et al. 1985), and 1.1 mm/day for Alosa sapidissima (Crecco and Savoy 1985). The ecological advantage of extended spawnings that result in multiple cohorts within a single age-0 year- class can be thought of as a "hedged bet" strategy that spreads age-0 production over time to take advantage of a variable environment (Lambert and Ware 1984). However, distinct cohorts within age-0 fish can also result from environmental factors acting on a single spawning effort; for example, through variation in prey availability (Timmons et al. 1980, Keast and Eadie 1985, Wicker and Johnson 1987), or a combination of biotic and abiotic factors (Lambert 1984, Crecco and Savoy 1985). However, previously published informa- tion indicates that the multiple cohorts observed in the present study probably resulted from multiple spawn- ings. Shepherd and Grimes (1984) showed that large weakfish "tiderunners" 55-80 cm enter the Delaware Bay estuary in the spring and spawn. In the summer these were replaced by 25-35 cm gravid weakfish. Har- mic (1958) showed a repeating pattern of multiple spavsming over 3 years, where a peak of egg abundance occurred in mid-June, followed by a conspicuous gap, and another peak in mid-July. In summary, the present study showed that multi- ple cohorts exist within the age-0 year-class of Chesa- peake Bay- York River weakfish. These cohorts showed significant differences in growth rates and appeared to partition habitats. Consequently, population studies directed at predicting year-class strength from juvenile surveys need to consider the potential for age-0 cohort variability. Acknowledgments We thank Steven Weiss and Susan Engels for their help in mounting and reading scales, and Paul Gerdes, Stephen O'Neil, and Robert Siegfried for their help in field collections. Work was completed as part of EPA grant number R810334, to Michael P. Weinstein at the Virginia Institute of Marine Science, College of William and Mary. Citations Boehlert. G.W., and M.M. Yoklavich 1985 Larval and juvenile gi-owth of sablefish, .4 noplopoma fim- bria, as determined from otolith increments. Fish. Bull.. U.S. 83:47.5-481. Buchanan-WoUaston, H.J.. and W.C. Hodgson 1929 A new method of treating frequency curves in fishery statistics, with some results. J. Cons. Int. Explor. Mer 4: 207-22.S. Campana, S.E.. and J.D. Neilson 1985 Microstructure of fish otoliths. Can. J. Fish. Aquat. Sci. 42:1014-1032. 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. 7.5:657-702. Conover, D.D., and M.R. Ross 1982 Patterns in seasonal abundance, growth, and biomass of the Atlantic silverside, Menidia menidia, in a New England estuary. Estuaries 5:275-286. Cooper, G.P. 1937 Food habits, rate of growth, and cannibalism of young largemouth bass in stateoperated rearing ponds in Michigan during 1935. Trans. Am. Fish. Soc. 66:242-266. Crecco, V.A., and T.F. Savoy 1984 Effects of fluctuations in hydrographic conditions on year- class strength of American shad (,4/)ili>poi)w fimbria. Can. J. Fish. Aquat. Sci. 43:930-937. Shepherd. G.R.. and C.B. Grimes 1983 Geographic and historic variations in growth of weakfish. Cynoscion regalis, in the middle Atlantic Bight. Fish. Bull., U.S. 81:803-813. 1984 Reproduction of weakfish. Cynoscion regalis. in the New- York Bight and evidence for geographically specific life history characteristics. Fish. Bull.. U.S. 82:.501-511. Shiossman, P.A., and M.E. Chittenden Jr. 1981 Reproduction, movements, and population dynamics of sand seatrout, Cynoscion arenarius. Fish. Bull., U.S. 79: 649-669. Szedlmayer, S.T. 1988 Early life history of juvenile weakfish, Cynoscion regalis (Bloch and Schneider). Ph.D.. diss.. College of William and Mary. Williamsburg, 118 p. Szedlmayer, S.T., K.W. Able. J. A. Musick, and M.P. Weinstein In press. Are scale cireuli deposited daily in juvenile weakfish. Cynoscton I'egalisl Environ. Biol. Fish. Takita, T., H. Saito, and A. Oiwa 1989 Occurrence of two cohorts in young of the year Nibcn albijlora in the Ariake sound and comparison of their growth and changes in body composition. Bull. Jpn. Soc. Sci. Fish. 55:1149-1156. Timmons. T.J., W.L. Shelton, and W.D. Davies 1980 Differential growth of largemouth bass in West Point Reservoir, Alabama, Georgia. Trans, .^m. Fish. Soc. 109: 176-186. Tonn, W.M., and C.A. Paszkowski 1986 Size-limited predation, winterkill, and the organization of Ihnbra-Pcrcn fish assemblages. Can. J. Fish. Aquat. Sci. 43:194-202. Van der Veer, H.W., and M.J.N. Bergman 1987 Predation by crustaceans on a newly settled 0-group plaice, Pleuronectes platessa population in the western Wad- den Sea. Mar. Ecol. Prog. Ser. 35:203-215. Victor. B.C. 1986 LaiTal settlement and juvenile mortality in a recruitment- limited coral reef fish population. Ecol. Monogr. 56:145-160. Walline. P.D. 1985 Growth of larval walleye pollock related to domains within the SE Bering Sea. Mar.' Ecol. Prog. .Ser. 21:197-203. West. C.J., and P. A. Larkin 1987 Evidence for size-selective mortality of juvenile .sockeye salmon (Oncorhynchus nerka) in Babine Lake, British (.'olum- bia Can. J. Fish. Aquat. Sci. 44:712-721. Wicker, .4.M.. and W.F. Johnson 1987 Relationships among fat content, condition factor, and first-year survival of Florida largemouth bass. Trans. Am. Fish. Soc. 116:264-271. Williams, D. McB. 1983 Daily, monthly and yearly variation in recruitment of a guild of coral reef fishes. Mar. Ecol. Prog. Ser. 10:231-237. Zar, J,H. 1984 Biostatistical analysis. Prentice Hall. Inc., Englewood Cliffs. NJ. 718 p. Abstract.- Over eOOO male snow crabs were tagged during a 6-year period in Conception Bay, Newfound- land, in order to estimate the in- crease in size at the time of molting. Ninety-two animals were recaptured which had usable information on growth increments. Based on the amount of growth, we hypothesized that 20 of these had molted once while the remainder molted twice. Two lines of evidence support this in- terpretation. First, animals in the group presumed to have molted twice were at liberty on average twice as long as those presumed to have molted once. Second, a regres- sion line fitted to data on single- molters predicted the size after two molts in close agreement with a regression line fitted to data on dou- ble-molters. A nonlinear regression model was developed to estimate the parameters of the relationship be- tween post- and pre-molt sizes using the combined data set for single and double molters. The method was also generalized to account for a quadra- tic relationship between post- and pre-molt size. For crabs in the size range 80-110 mm carapace width, the predicted size after molting in mm is equal to 7.398 + 1.038 x pre- molt size. A similar study conducted in Bonavista Bay, Newfoundland, yielded growth information for 18 animals. The molt increments appear similar to those observed from Con- ception Bay. Growth per Molt of Male Snow Crab Chionoecetes opilio from Conception and Bonavista Bays, Newfoundland David M. Taylor John M. Hoenig Science Branch, Department of Fisheries and Oceans PO Box 5667, St John's, Newfoundland AlC 5X1, Canada Manuscript accepted 29 .liine 1990 Fishery Bulletin, U.S. 88:7,53-7(3(1. The snow crab Chionoecetes opilio has supported extensive commercial fisheries on both the Atlantic and Pacific coasts of North America since the 1960s (Elner and Bailey 1986). Snow crabs are also commercially ex- ploited in Japan. Only the males are harvested in North America because females never attain commercially acceptable sizes. Efforts to manage the resources have been hampered by a lack of detailed life-history informa- tion. This is because the crabs are found in deep water (50-700 m) and are thus difficult to study. Information on growth of snow crabs is required for effective man- agement for at least three reasons: (1) for incorporation in a yield-per- recruit model, (2) to forecast the bio- mass available to the fishery from size-specific, pre-season biomass esti- mates, and (3) to interpret size-fre- quency distributions. Although gi'owth can be studied in the laboratory, there is no guarantee that the observed growth will reflect what happens in free-living populations. Consequent- ly, there is a need to estimate gi-owth parameters from field data. There have been some tagging studies of snow crab growth, but the reported results were either of a preliminary nature or were unsatis- factory due to tag retention prob- lems (McBride 1982, Taylor 1982, Bailey and Dufour 1987). Improve- ments in tagging methods (Hurley et al. In press) have made field studies feasible. Growth of Crustacea is often esti- mated by studying two components: the increase in size at the time of molting (molt increment) and the timing of molting (either the inter- molt period or the proportion molting in a given season). Recent work by Moriyasu and Mallet (1986) and O'Halloran and O'Dor (1988) has pro- vided a method for estimating the proportion of the population that will molt in a given time period. In this paper, we concentrate on the prob- lem of estimating the size-specific molt increment of snow crabs in Con- ception and Bonavista Bays, New- foundland, from mark-recapture data consisting of the size at the time of tagging, the size at recapture, and the time at liberty. Conception and Bonavista Bays are deepwater bays on the Northeast coast of Newfoundland (maximum depth 295 m in Conception Bay; 412 m in Bonavista Bay). Commercial fishing grounds for crabs exist at depths exceeding 180 m. The bottom type in these areas is predominantly mud or muddy sand with mean bot- tom temperatures ranging from - 1.3 to 0.5°C. Since the mid 1970s both areas have experienced high levels of commercial crab fishing effort (Taylor and O'Keefe 1987). While aggregations of crabs may be found at depths <180 m, the deeper commercial fishing grounds appear to have the soft substrate necessary for snow crab molting. 753 754 Fishery Bulletin 88(4), 1990 60* 59' 56* 57' 56' 55' 54* 53° 52° 51° 50° 49° 59° 58° 57° 56° 55° 54° 53° 52° 51° 50° Figure 1 Map of Newfoumllarid, Canada, show- ing the location of Conception Bay (lower box) and Bonavista Bay (upper box). Materials and methods From 1979 to 1984, nine research cruises were made to Conception Bay, Newl'oimdland, Canada (47°30'N, 53 °W, Fig. 1), to tag 6296 male snow crabs, size range 56-135 mm carapace width (CW). An additional 2253 male crabs, size range 45-125 mm CW, were tagged during four cruises to Bonavista Bay, Newfoundland (48°50'N, 53°20'W), from 1979 to 1984. Crabs were captured using Japanese-style conical traps baited with squid Ille.r illecebrosu!< and set in longlines of twelve traps at depths ranging from 110 to 285 m. Traps were hauled after soaking for 24 hours. Tagging was conducted in a manner designed to minimize mortality. Traps were spaced approximate- ly 40 m apart, which allowed us to tag crabs from the trap on deck while the other traps remained in the water. Before tagging, we examined each animal and discarded any that appeared injured. Carapace width was measured to the nearest millimeter using vernier calipers. Animals were tagged with a T-bar tag (Floy Reference to trade names does not imply endorsement by the National Marine F'isheries Service. N().'\.-\. Tag Mfg. Co., Inc., Seattle, WA 98105) and immediate- ly released on location. Details of the tagging procedure are described in Hurley et al. (In press). Tags were in- serted in the posterior ecdysial suture (epimeral line) which was made visible by applying gentle upward pressure to the carai)ace. The tagging location was 2-6 mm from the right coxopodite of the last walking" leg. Before releasing the crab, the end of the tag was given a gentle tug. If the tag appeared loose, it was removed and the animal was discarded. To determine measurement errors, one biologist measured 90 animals three times in blind trials. The animals ranged in size from 63 to 124 mm. Recaptured animals were obtained mostly from commercial fishermen. Only animals that were examined by a staff biologist were used in the analysis. Results and discussion Repeated measurements by a biologist of a group of crabs were always within 1 mm of the mean for each animal. Workers studying other species of large Crustacea have suggested that measurements can be Taylor and Hoenig- Growth per molt of Chionoecetes opilio from Newfoundland 755 707 0 5 10 15 20 25 MOLT INCREMENT (mm) Figure 2 Frequency of occurrence of growtli increments for recaptures of Chionoecetes opilio from Conception Bay. precise to within 3 mm (Restrepo 1989, Hunt and Lyons 1986, Little 1972). If measurement errors oc- cur at both the time of tagging and the time of re- capture, then errors as large as 5 or 6 mm might be encountered. Conception Bay We recovered 850 tagged animals from Conception Bay. Of these, 751 had measured increases in size of 3 mm or less; three had increases of 4-5 mm; the re- maining 93 animals had increases of at least 6 mm (Fig. 2). None of the three animals with molt increments of 4 or 5 mm were particularly small (premolt sizes: 92, 96, 99 mm). Consequently, these animals would not be expected to have particularly small increments. Their times at liberty were 325, 187, and 157 days, respec- tively. We assume that the 93 animals with increments of 6 mm or more molted at least once, and none of the other animals molted. The average time at liberty for animals which did not molt is under a year (x 322 days, median 251 days, range 6-1598 days; Fig. 3). The size at recapture of Conception Bay crabs which molted was plotted against the size at tagging, and two distinct linear clouds of points were apparent (Fig. 4). The line indicating a 17% increase in size appears to separate the two clouds nicely. A reasonable working hypothesis is that the lower cloud consists of animals which molted once, while the upper cloud consists of animals which molted twice. One animal in the two-molt group was at liberty for only 29 days and was tagged and recaptured in a hard-shelled condition. Animals re- main soft-slielled for 2-3 months after molting, and soft-shelled animals cannot molt (Taylor et al. 1989). ' 'y- m »: txlPxlf:<]k-;1r;^F:?lF:l. D4fS AT LIBGRTy Figure 3 Histograms of times at liberty for Conception Bay Ch itvioereti's npUio. Upper figiire corresponds to tlie upper cloud of points in Figure 4. i.e.. to animals believed to have molted twice. Middle figure corre- sponds to the lower cloud in Figure 4. Lower figure corresponds to recaptured animals which had not molted while at liberty. For this reason, this animal was eliminated from fur- ther consideration. The times at liberty for 81 of the 92 Concepti(.)n Bay animals which molted ranged from 18 to 1618 days. (Exact times at liberty could not be determined for 1 1 animals.) However, the mean time at liberty for the hypothesized two-molter group was approximately twice as long as for the single-molter group (828 vs. 468 days) and there was little overlap in the distribu- tions of time at liberty for the two groups (Fig. 3). Animals which did not molt were at liberty for an average of 322 days. Although there is little informa- tion in the literature on intermolt periods for snow crabs, it is believed that crabs approaching commer- cial size molt once per year (Robichaud et al. 1989) and 756 Fishery Bulletin 88(4), 1990 140 / !'■" /^ f •s.S.'. >/ • I cc < 0 a ^^ ^ 5 ® .< i X ''- 70 la ?:■ ■ Ml) 120 PRE-MOLT CARAPACE WIDTH (mm) Figure 4 Plot of size at recapture vs. size at tagging for recaptures of Chio- noecetes opilio from Conception Bay. The solid line, which indicates a 17% increase in size while at liberty, seems to separate the data into two clouds of points. Solid symbols represent animals believed to have molted once; open symbols represent animals which molted twice. Bulls-eye symbol represents one animal which was reportedly at liberty for only 29 days and which consequently was e.xcluded from analysis. tliat the largest crabs may skip a year between molts. Thus, the information on time at liberty supports the hypothesis that animals in the lower group of f^igure 4 molted once while those in the upper group molted twice. The data in the lower group in Figure 4 appear to be well described by a linear relationship: Recapture size = a + b (size at tagging) + (1) where e is a random error term. Su[)pose that equa- tion (1) describes the size after one molt. Then the size after two molts would be given by the recursive for- mula (Kurata 19R2): Size after two molts = n + h ((/ -I- /) (size at tagging)) + e. (2) Application of equation (2) to the size at tagging should provide a good prediction of the size at recapture for animals in the upper group if the assumption is cor- rect that equation (1) describes the size after one molt. We fitted lines to the two clouds of points in Figin-e 4 by ordinary least squares (Table 1). The results are consistent with the hypothesis of one and two molts for the two groups. For example, the fitted e(|uation (I) (with parameters estimated from the lower cloud Table 1 Regressions of length at recapture vs. length at tagging for the two groups of Chionoecetes opilio from Conception Bay shown in Figure 4. Lower cloud animals are presumed to have molted once; upper cloud, twice. Parameter estimates « and b pertain to equations (1) and (2) in the text. Attribute Lower cloud Upper cloud Number of observations Adjusted r- Intercept (standard error) Slope (standard error) Interpretation of intercept Interpretation of slojie 20 0.8.5 17.2.50 (8.93:5) 0.941 (0.091) a b 17.250 0.941 72 0.92 12.667 (3.640) 1.104 (0.039) a(l +b) b- 6.179 1.0.51 of points) predicts that an animal 80 mm in size will be 92.53 mm after one molt. Inserting this estimate of 92.53 mm into equation (1) gives a predicted size of 104.32 mm after another molt. In contrast, a linear regression fitted to the uj)per cloud of points predicts that an animal 80 mm in size will be 100.91 mm after two molts, i.e., 3.41 mm smaller than the estimate from equation (1). Similarly, the size after two molts pre- dicted by equation (1) for a 110-mm CW animal is 130.89 mm, whereas the size predicted by the regres- sion for the upper cloud is 134.00 mm, i.e., 3.11 mm larger. Over the range of sizes for which we have data, results fron:i the two equations agree closely. In fact, approximate confidence l)ands for the size after two molts, as determined from the lower regression line, enclose the confidence bands f(.)r the upper regression line over the entire range of the data (Fig. 5; see Ap- pendix A for derivation of approximate confidence hands). We therefore conclude that animals in the lower cloud of points molted once, while those in the upper cloud molted twice. The regressions in Table 1 are based on 20 and 72 animals. Since the sample sizes are small, it would be useful to use the combined molting data from all 92 animals to derive a i)est estimate of the parameters o and /). This can be accomplished by combining equa- tions (1) and (2) into a single equation in the form of a nonlinear regression model. Let Y be the observed size at the time of recovery, X be the size at tagging, and let Z be an indicator variable for whether an animal molted once or twice, i.e.. let 0, animal molted once 1, animal molted twice. Taylor and Hoenig: Growth per molt of Chionoecetes opitio from Newfoundland 757 Figure 5 Recapture data for Chionoecetis tipilm from Conception Bay. Solid symbols represent animals presumed to have molted once: open sym- bols represent animals which molted twice. Solid lines show the 95% confidence band obtained from the linear regi-ession fitted to the open circle data. Dashed lines give an approximate 95% confidence band for the size after two molts based on the regression for the presumed single-molters (solid symbols). (See Appendix A for method of con- structing the confidence band.) The dashed lines enclose the solid lines, indicating that the two predictions of size after two molts are not significantly different. This supports the idea that the two clouds in Figure 3 re|;)resent single- and doulile-molters. Then the growth parameters a and h can be estimated by regressing Y on X and Z in the regression model Y abZ + hX + b{h - 1)ZX -t- e (3) where e is a random error term. Equation (3) reduces to (1) when Z = 0 and to (2) when Z = 1. The model is generalized in Appendix B to account for a quadratic relationship between post- and pre-molt size. The pre- dicted sizes from the fitted equation (3) (Table 2) are very similar to the results obtained using equations (1) and (2) separately. Table 2 Estimates of the parameters a and b in the nonlinear regres- sion model relating size at recapture (Y) of Conception Bay Chionoecetes opilio to size at tagging (X) for animals molting once or twice. The model is: Y = a -t- abZ + bX + b{b - 1)ZX + e. where e is the error term and Z = 0 if the animal molted once, and 1 if the animal molted twice. Parameter Estimate Standard error 7.398 1.038 2.074 0.021 correlation -0.998 Figure 6 Plot of size at recapture vs. size at tagging for recaptures of Chio- noecetes opilio from Conception Bay (circles) and Bonavista Bay (triangles and stars). Solid circles and stars represent animals be- lieved to have molted once: open triangles and circles represent animals which molted twice. Diamond represents an animal from Bonavista Bay at liberty for ~1000 days which was presumed to have molted three or more times. Large circle represents an animal which normally would be assumed to have molted twice: however, it was reportedly at liberty for only 29 days and consequently was excluded from analysis. Regression lines were determined using equation (3) from all data on single and double molters from both bays. Bonavista Bay Molt information was obtained from 18 animals re- captured in Bonavista Bay. Of these, four animals ap- peared to have molted once; 13 animals appeared to have molted twice; and one animal at liberty for ap- proximately 1000 days appeared to have molted three or more times (Fig. (3). This interpretation is supported by the fact that the animals presumed to have molted once were at liberty for an average of 198 days, while the animals presumed to have molted twice were at liberty for an average of 698 days. The molt increments appeared very similar to those from Conception Bay. Since there is not much infor- mation from Bonavista Bay, we computed a regression to predict molt increments for Bonavista Bay crabs using the combined recapture data on single and dou- ble molters from Bonavista and Conception Bays. The predicted size after molting is given by Predicted size = 9.21 + 1.02 (pre-molt size). Other populations The predicted molt increments from equation (3) for Conception Bay crabs (Table 2) were very close to a constant (10.5 - 1 1.7 mm) over the size range of ani- mals we studied (82-113 mm CW). These estimates are 758 Fishery Bulletin 88(4), 1990 consistent with the Hmited information from a tagging study in the Gulf of Alaska. McBride (1982) reported the mean growth increment of six tagged snow crabs at liberty for up to 1 year was 14.7 mm. McBride's animals were somewhat larger than those in our study (x 113.8 mm, range 108-124 mm CW at tagging) and would, on the basis of our regression, be predicted to have slightly larger molt increments. Laboratory studies of snow crab growth are also con- sistent with our findings. O'Halloran (1985) reported an average molt increment of 1 1 .6 mm for eight ani- mals. However, five of the eight animals had their eyestalks ablated bilaterally and these animals died dur- ing ecdysis. Consequently, these estimates are not very reliable. Miller and Watson (1976) and R.J. Miller (Dep. Fish. Oceans, Halifax, Nova Scotia, Canada 83J 2S7, pers. commun., June 1989) observed a mean molt in- crement of 15.5 mm for 18 crabs ranging in pre-molt size from 59.3 to 101.1 mm CW (.r 80.5 mm). G. Hur- ley (Hurley Fish. Consulting, 45 Alderney Dr., Dart- mouth, Nova Scotia, Canada B2Y 2N6, pers. commun., June 1989) observed a mean increment of 13.9 mm for 47 crabs, size range 60.7-83.4 mm CW (.7 66.1 mm). In these studies, crabs were fed ad libitum. In another lal)oratory study, Moriyasu et al. (1987) reported a regression line that indicates that animals in the size range we studied (82-113 mm CW) have predicted molt increments of 14-16 mm CW. Molt increments have also been estimated l)y size- frequency analysis for Japanese populations of snow crabs. Kon et al. (1968) reported estimates for small individuals. Ito (1970) suggestetl that 81 -mm CW in- dividuals molt to 97 mm, then to 111 mm and 121 mm (i.e., molt increments of 16, 14, and 10 mm). Kon (1980) suggested that 80-mm CW individuals molt to 93.4 mm, then to 105.6 and 1 16.7 mm (molt increments of 13.4, 12.2, and 11.1 mm). These estimates are similar to the ones we derived from the tagging data. Robichaud et al. (1989) analyzed size-frequency distributions of snow crabs from the Gulf of St. Lawrence. Their samples consisted of small animals, so their results are not com- pai-able to the results presented here. Effect of possible terminal molt Some workers believe that male snow crabs undergo a "terminal" or final molt which is associated with a change in allometry. Although this idea is controver- sial (see Jamieson and McKone 1988 for reviews), it is worth examining whether the two groups evident in Figure 4 might reasonably be interpreted as those molting to the terminal state (e.g., lower group) and those molting but not to the terminal state (upper group). To accept this hypothesis implies that all of the following are accepted: (1) males do, in fact, undergo a terminal molt; (2) the size increment at the terminal molt is con- siderably different (presumably smaller) than the pen- ultimate molt increment; (3) it is a coincidence that the animals in the upper group were at liberty twice as long as those in the lower group (828 vs. 468 days on average), and that the animals which did not molt were at liberty for less than a year on average {x 322 days); (4) it is a coincidence that the size predicted after two molts by the regression fitted to the lower cloud of points is in close agreement to the size after two molts predicted by the regression fitted to the upper cloud of points; (5) there must be an as yet unidentified reason why a third cloud of points, corresponding to animals which molted twice, is not evident in Figure 4. Even if males undergo a terminal molt, there is no evidence to suggest that the final molt increment is distinctly smaller than the penultimate molt increment. In an aquarium study, Moriyasu et al. (1987) found, for male snow crabs in the size range 60-70 mm CW, that animals molting to the large-clawed state have molt increments about 3 mm smaller than those molting to a small-clawed state. This small difference in molt in- crements is not sufficient to account for the two clouds we identified in the field observations. Ennis et al. (1988) have shown that males can become functionally mature before attaining the morphometry associated with the terminal molt. Hence, the terminal molt is not necessarily associated with a diversion of energy from growth to reproductive processes, and the terminal molt increment need not be small. We conclude that if animals undergo a terminal molt, the molt increment at this time is similar to the penultimate molt inci'e- ment. We do not rule out the possibility that some of the scatter about the regressions through the two clouds of points is due to mixing terminal and non- terminal molt data. In summary, our interpretation of the tagging data is supported by two lines of evidence. Animals pre- sumed to have molted twice were at liberty twice as long on average as animals presumed to have molted once. Also, the predicted size after two molts, as esti- mated from data on animals presumed to have molted once, agrees closely with the predicted size estimated from data on animals presumed to have molted twice. We conclude that male snow crabs in the size range 75-115 mm have molt increments of around 11 mm. Our results are consistent with the limited information ai)OUt snow crab growth in the literature. Taylor and Hoenig Growth per molt of Chionoecetes opilio from Newfoundland 759 Acknowledgments We wish to thank P. O'Keefe, the other technical staff, and the captain and crew of the FRV Shamook for as- sistance with the field work. G. Hurley and R.J. Miller provided original data on laboratory studies. Donald Stansbury helped with the computer graphics, and David Schneider, Russell Millar, Geoff Evans, William Warren, Mikio Moriyasu, and the anonymous reviewers provided helpful comments. Citations Bailey, R.F.J., and R. Dufour 1987 Field use of an injected ferromagnetic tag on the snow crab (ChiiiiKMretes opilio 0. Fab.). J. Cons. Int. Explor. Mer 43:237-244. Elner, R.W., and R.F.J. Bailey 1986 Differential susceptibility of Atlantic snow crab, Chio- noecetes opilio, stocks to management. In Jamieson, G.S., and N. Bourne (eds.), North Pacific workshop on stock assessment and management of invertebrates, p. 33.5-346. Can. Spec. Publ. Fish. Aquat. Sei. 92. Ennis, G.P., R.G. Hooper, and D.M. Taylor 1988 F'unL'tiimal maturity in small male snow crabs {Chionoe- cetes apili,'). Can. .J. Fish. Aquat. Sci. 45:2106-2109. Hunt, J.H., and W.E. Lyons 1986 Factors affecting growth and maturation of spiny lob- sters. Panulirus argus, in the Florida Keys. Can. J. Fish. Aquat. Sci. 43:2243-2247. Hurley. G.V., R.W. Elner. D.M. Taylor, and R.F.J. Bailey In press Evaluation of molt-retainable tags for snow crabs. In Proceedings. International Symposium and Educational Workshop on Fish-marking Techniques. Am. Fish. Soc. Symp. Ser. Ito. J. 1970 Ecological studies on the edible crab, Chionoecetes opilio (0. Fabr.) in the Japan Sea — III. Age and growth carapace width frequencies and carapace hardness. Bull. Jpn. Sea Reg. Fish. Res. Lab. 22:81-116. Jamieson, G.S., and W.D. McKone (editors) 1988 Proceedings of the international workshop on snow crab biology, December 8-10, 1987, Montreal, Quebec. Can. Manuscr. Rep. Fish. Aquat. Sci. 2005, 163 p. Kon. T. 1980 Studies on the life history of the zuwai crab, Chionoecetes opilio (0. Fabricius). Spec. Publ. Ser. 2, Sado Mar. Biol. Stn., Niigata Univ. Kon, T., M. Niwa, and F. Yamakawa 1968 Fisheries biology of the tanner crab — II. On the frequen- cy of molting. Bull. ,Jpn. Soc. Sci. Fi.sh. 34(2):138-142. Kurata. H. 1962 Studies on the age and growth of Crustacea. Bull. Hok- kaido Reg. Fish. Res. Lab. 24. Little, E.J. Jr. 1972 Tagging of spiny lobsters {Ponuliriif: argus) in the Florida Keys, 1967-1969. Fla. Dep. Nat. Resourc. Mar. Res. Lab. Spec. Sci. Rep. 31, 23 p. McBride, J. 1982 Tanner crab tag development and tagging experiments 1978-1982. In Proceedings. International Symposium on the Genus Chionoecetes, p. 383-403. Lowell Wakefield Fish. Symp. Ser., Alaska Sea Grant Rep. 82-10, Univ. Alaska, Fairbanks. Miller, F.J., and J. Watson 1976 Growth per molt and limb regeneration in the spider crab, Chionoecetes opilio. J. Fish. Res. Board Can. 33:1644-1649. Moriyasu, M., and P. Mallet 1986 Molt stage of the snow crab ('h ionoecetes opilio by obser- vation of morphogenesis of setae on the maxilla. .1. Crusta- cean Biol. 6:468-490. Moriyasu, M., G.Y. Conan, P. Mallet. Y.J. Chiasson, and H. Lacroix 1987 Growth at molt, molting season and mating of snow crab [Chionoecetes opilio) in relation to functional and morphometric maturity. Int. Counc. Explor. Sea CM 1987/K: 21, 14 p. O'Halloran, M.J. 1985 Moult cycle changes and the control of moidt in male snow crab, Chionoecetes opilio. M.S. thesis. Dalhousie Univ., Hali- fax, Nova Scotia. 183 p. OHalloran, M.J., and R.K. O'Dor 1988 Molt cycle of male snow crabs, Chionoecetes opilio, from observations of external features, setal changes, and feeding behavior. J. Crustacean Biol. 8:164-176. Restrepo, V.R. 1989 Growth estimates for male stone crabs along the south- west coast of Florida: A synthesis of available data and methods. Trans. Am. Fish. Soc. 118:20-29. Robichaud, D.A., R.F.J. Bailey, and R.W. Elner 1989 Growth and distribution of snow crab. Chionoecetes opilio, in the southeastern Gulf of St. Lawrence. J. Shellfish Res. 8:13-23. Taylor, D.M. 1982. .\ recent development in tagging studies on snow crab. Chionoecetes opilio, in Newfoundland— retention of tags through ecdysis. In Proceedings. International Symposium on the Genus Chionoecetes, p. 405-417. Lowell Wakefield Fish. Symp. Ser.. Alaska Sea Grant Rep, 82-10, Univ. Alaska, Fairbanks. Taylor. D.M., and P.G. OKeefe 1987 Analysis of the snow crab {(.'h ionoecetes opilio) fishery m Newfoundland for 1986. Can. Atl. Fish. Sci. Advis. Comm. CAFSAC Res. Doc. 87/57. Dartmouth. Nova Scotia, 26 p. Taylor, D.M., G.W. Marshall, and P.G. O'Keefe. 1989 Shell hardening in smiw crab tagged in soft-shelled con- dition. N. Am. .J. Fish. Manage. 9:504-508. 760 Fishery Bulletin 88(4), 1990 Appendix A Derivation of Confidence Bands The predicted size after one molt can be obtained by regressing tiie size at recapture on the size at tagging for those animals which molted once (text equation 1). Thus, predicted recapture size = a + bX (A.l) where a and h are parameter estimates and X is the size at tagging. Estimates of the variances and covar- iances (i.e., V(a), V(6), Cov{aJ))) are easily obtained in the standard way. The predicted size after two molts is obtained from A.l as Y = predicted size after two months = a + b{a + bX) = a + ab + 6-X. An approximate (asymptotic) estimate of the variance of the size after two months, V(Y), is found by the Taylor's series or delta method to be V(Y) = dY V(-0 + {^^Y V{b) db . 2 i"i (" da db Cov {ii,li) where the derivatives are evaluated at the parameter estimates. Thus. V(Y) = (1 + b)- Via) + id + 2bx)- Y(b) 2(1 + 6)(o + 26x) Co\'(a.b). An approximate 95% confidence band is thus obtained as Y ± 2 \/V(Y). Appendix B Estimating molt increments for a quadratic model Over the size range of animals we studied, the relation- ship between post- and pre-molt size appeared linear. However, when a wide range of pre-molt sizes is con- sidered, it is common to find a curvilinear relationship which may be modeled satisfactorily by a quadratic equation. The non-linear regression model (equation 3) in the text can be generalized to allow estimation for the quadratic model. Let the size at tagging be denoted by X, and assume the size after one molt is given by size after one molt = « + 6X -i- cX-. (B.l) Then the size after two molts is given by size after 2 molts = n + b(ii + liX + cX-) + c(a + bX + cX~)~. (B.2) As before, let Y be the size at recapture (for animals molting once or twice), and define Z to be an indicator variable for whether an animal molted once or twice, i.e., let Z = 0, animal molted once 1, animal molted twice Then equations (B.l) and (B.2) can be combined in a single non-lineai' regression as Y = a + {ab + r(-r)Z + bX + b(b - 1 + 2r;r)ZX -I- c-X -I- c(b + 2ac + b- - 1)ZX- + 2bc-ZX^ + (■■'ZX-' + e (B.3) where e is the random error term. Equation (B.3) reduces to (B.l) when Z = 0 and to (B.2) when Z = 1. Abstract.- We applied Shep- herd's length composition analysis (SRLCA) to research trawl survey catches of Gulf of Maine northern shrimp Panddlug horealis to test the efficiency of the method in interpret- ing the age structure of a length- freiiuency distribution incorporating significant variation in growth and recruitment rates. We evaluated the performance of the method by com- paring the von Bertalanffy growth parameters provided by SRLCA antl subsequently derived age frequen- cies and instantaneous total mortal- ity rates with previously accepted results based on simple visual inspec- tion of the annual length-frequency distributions. In spite of the variable growth and recruitment rates exhibited by the stock, SRLCA yielded information providing a resolution of the length- frequency data close to a priori as- sumptions, although information ex- ternal to the procedure was needed to select the best interpi'etation from among several locally optimal solu- tions. A Practical Assessment of the Performance of Shepherd's Length Composition Analysis (SRLCA): Application to Gulf of Maine Northern Shrimp Pandalus borealis Survey Data Mark Terceiro Josef S. Idoine Woods Hole Laboratory, Northeast Fisheries Science Center National Manne Fisheries Service, NOAA, Woods Hole, Massachusetts 02543 Manuscript accepted 18 June 1990. Fishery Bulletin, U.S. 88:761-773. The analysis of length-frequency modes was the first method used l)y aquatic biologists to delineate successive co- horts in fish and invertebrate popula- tions. Simple visual inspection of modes was developed first (e.g., the "Peter- sen" method [Petersen 1891]); this relies heavily on the intuition of the scientist for the cori-ect separation of age groups. Later workers assumed a normal distribution underlying the observed length modes, and graph- ical methods using normal probabil- ity paper were used to resolve length distributions to cohorts by the suc- cessive identification and removal of suspected age groups. The method of Cassie (1954) is probably the best known of these graphical procedures, which rely to a large degree on sub- jective decisions of the scientist. Dif- ficulty in defining modes for older age groups, problems in interpreta- tion caused by variable growth rates and recruitment, and an inability to reproduce the interpretation of age groups from one worker to the ne.xt have limited the utility of these sim- ple methods. For most finfish stocks, the inter- pretation of growth intervals on hai'd body parts (e.g., scales, otoliths, spines, and vertebrae) has evolved as the ageing method of choice, replacing length-composition analysis methods. For many taxa, however, the inter- pretation of growth intervals from age structures is difficult, either due to problems in identifying periodic marks or because of the lack of suit- able hard structure. For fast-grow- ing, short-lived tropical finfish, and invertebrates such as lobsters, crabs, shrimps, and squids, the resolution to ages of modes in length-frequency distributions continues to be the pri- maiy method used to estimate growth and age structure of populations. Re- cently developed methods directed to interpreting length-frequency distri- butions generally fall into two cate- gories. The first group treats the problem as one of statistically resolv- ing a mixture of distributions, and usually assumes an underlying nor- mal distribution for the components. The parameters resulting in the best match between the area under the theoretical distribution and the area under the observed length distri- bution are selected by employing chi-square or maximum likelihood methods. These distribution mixture methods require some prior constraint on the number of length modes and the bounds of the parameters to prevent biologically unrealistic re- sults. The lineage of this approach in- cludes the methods of Hasselblad (19(i6), Tomlinson (computei- progi-am 761 762 Fishery Bulletin 88(4), 1990 NORMSEP; 1971), Yong and Skillman (computer pro- gram ENORMSEP; 1975), McNew and Summerfelt (1978), and MacDonald and Pitcher (computer program MIX; 1979). The distribution mixture method of Schnute and Fournier (1980) uses a growth model to impose these constraints. A second category of pro- cedures assumes a specific growth function (usually the von Bertalanffy) and attempts to match predicted length modes to those observed. Among these methods are the ELEFAN I procedure of Pauly and David (1981), and Shepherd's length composition analysis (1987). The Shepherd length composition analysis method (SRLCA) (Shepherd 1987) relies on a goodness-of-fit score function which varies according to the corre- spondence of observed and predicted length-frequency modes for given pairs of von Bertalanffy growth parameters (L,„f and K), thus presumably constraining the indication of optimal parameters to within biolog- ically realistic bounds. SRLCA has fewer subjective input requirements than the distribution mixture methods (Shepherd 1987). Basson et al. (1988) per- formed Monte Carlo tests of SRLCA and noted that although SRLCA provided biased results for simulated data with large variation in !ength-at-age, SRLCA generally performed better than ELEFAN L To test the performance of SRLCA on an observed, potentially difficult-to-interpret data set, as suggested by Shepherd (1987), we applied a version of SRLCA using the von Bertalanffy growth equation to research trawl survey data for Gulf of Maine northern shrimp Pandaliis borealis, and compared our results with ac- cepted interpretations of the data. Currently, simple visual inspection and information on sexual character- istics are used to resolve survey length frequency to age frequency, providing subsequent estimates of relative adult stock abundance, recruitment success, and total instantaneous mortality rates (Z) (Mclnnes 1986, NSTC 1987). Survey results reveal that this stock has experienced variable recruitment and growth dur- ing 1982 to 1988 (NSTC 1984, 1985, 1986, 1987, 1988). Although true ages for northern shrimp are not avail- able for use as "ground truth" in this evaluation of SRLCA performance, we feel our assessment of the method is valuable since it is based on application to real data of the type which in practice might require the use of length-based assessment methodolog}'. Methods Analysis SRLC'A compares the observed length-frequency dis- tribution with that expected from the von Bertalanffy equation for given test pairs of L,„f and K by applica- tion of a continuous, periodic test function of the form: T, = ((sin n {t„„„ - t„„„}] / (n {t,,,., - t„„„}]) X (cos 2n [t|,,„. - ts]) where T, is the value of the function for a given length interval i, t,„ax and tmin are ages at the upper and lower bounds of the interval for a given test set of growth parameters, ti,ar is the average of t,,,^^ and t^jn , and ts is the date of observation, expressed as a fractional part of the age (e.g, annual) cycle (Shepherd 1987). A measure of goodness-of-fit is then used to deter- mine the best fitting set of growth parameters for the observed length-frequency distribution. This measure, the score function S, is given by: S = X T,N,"^ where i indexes the length intervals, T is as indicated previously, and N is the number of animals in each in- terval. Taking the square root of N helps reduce the sensitivity of the score function to unusually large numbers of animals in a given length interval (e.g., in the event of exceptional recruitment; Shepherd 1987). Cumulative scores are large and positive when length modes predicted for a given pair of growth parameters are consistent with observed len.gth-frequency modes, with negative scores indicating inconsistency. Shep- herd (1987) suggested that regions of nearly constant scores within the K by L,,,,- score matrix may provide an indication of the shape of the confidence interval around pairs of parameters (e.g.. Shepherd suggests a region equal to one-half of the maximum score might approximate the 95% confidence interval). Typically, several regions (hereafter called ridges) of high scores will be observed for each length-frequency distribution analyzed, with local maxima in each ridge that provide alternative interpretations of the data. Length-frequency data (kilf of Maine northern shrimp are protandric her- maphrodites, with the females the target of a valuable winter/spring fishery in the western Gulf of Maine (see Mclnnes 1986 for an overview of the fishery). A re- search vessel trawl survey was implemented in 1983 to provide a fisheries-independent source of data for the stock. A stratified random trawl survey is con- ducted annually during late July through mid-August aboard the Northeast Fisheries Center (NEFC) RV Gloria Michelle in the western Gulf of Maine (Fig. 1). Terceiro and Idoine SRLCA application to Pandalus boreslis survey data 763 Figure 1 Western Gulf of Maine region with major bathymetric features and 50-fathom (100-m) isobath. Northern shrimp survey area extends from 68°W longitude to the 50-fathom isobath, except for stratum 2, where the sampled area extends to the 30-fathom (60-m) isobath. Length-frequency data from identical sample strata sets (strata 1, 3, and 5-8) are available for 1984-88 (NSTC 1984, 1985, 1986. 1987, 1988). Trawl gear con- sists of a modified 4-seam conmiercial shrimp trawl with 35 mm (1.4 inch) stretched mesh in the body of the net and 32 inm (1.3 inch) stretched mesh in the ex- tension and codend. with "rockhopper" ground gear to allow sampling over rough bottom (Mclnnes 1986). Samples (2 kg) of each tow are retained for length measurement and sex determination. All shrimp in the sample are measured, with mid-dorsal carapace lengths aggregated by 0.5-mm intervals (nearest 0.5 mm below that measured). A 1-kg subsample is retained for deter- mination of sex and spawning stage. To date, abundance and biomass indices from the RV Gloria Michelle survey (stratified mean number per tow, stratified mean weight [kg] per tow) have proven to be accurate predictors of year-class size and com- mercial fishery performance (total catch and catch-per- unit-effort; NSTC 1988). Coefficients of variation for stratified mean number and weight (kg) per tow, ag- gregated over all sample strata, have averaged (1984- 88) about 13% and 12%, respectively, indicating rela- tively high precision. A priori assumptions Information is available from previous analyses of the Gulf of Maine northern shrimp length-frequency dis- tributions to provide a baseline interpretation for evaluation of SRLCA results. The von Bertalanffy growth equation has been the accepted means of describing the growth of P. horealis (Frechette and Parsons 1983). For the Gulf of Maine stock, growth parameters were expected to be in the range of those previously estimated for this stock. These include (1) parameters derived here by nonlinear least-squares regression analysis of mean carapace length (CL) at age data summarized by Haynes and Wigley (1969; Lj^f = 32 mm, K = 0.46, t„ = -0.12), and (2) parameters derived from NEFC groundfish bottomtrawl survey data for northern shrimp (NEFC unpubl. data in Mclnnes 1986; L,,,,- = 35.2 mm, K = 0.36, t,, = 0.06). In both of these studies, age was estimated by visual in- spection of length-frequency distributions. For Gulf of Maine northern shrimp survey data, a 1 March birthday was assumed (Apollonio et al. 1986), with an average catch date of 1 August (NSTC 1984, 1985, 1986, 1987, 1988); thus t, was set equal to 0.42 for all SRLCA runs in this exercise. The usual presence of four age groups (ages 1-4) in the survey catch was suspected (Mclnnes 1986), with a maximum age of 5 years (Haynes and Wigley 1969, Apollonio et al. 1986), based on modes in length frequencies and sexual characteristics (Allen 1959, McCrary 1971). Northern shrimp in a length range of 13-18 mm CL are generally assumed to be of age-group 1, and mainly immature and mature males, while shrimp 18-22 inm CL are assumed to be age 2 and mostly mature males. Animals in the 22-25 mm CL interval are assumed age 3, usual- ly females with no previous spawning history. Shrimp larger than 25 mm CL are assumed to be females of age-group 4, with possible age-group 5 female shrimp at CL greater than 29 mm. The survey samples north- ern shrimp in a CL range of 10-32 mm (Fig. 2), with animals assumed fully recruited to the gear at about 19-20 mm CL, or age 2 and older (Blott et al. 1983). Previous interpretations of survey results and commer- cial fishery performance suggest that the 1982 and 1987 year-classes (YC) of Gulf of Maine northern shrimp were strong, and the 1983 YC very weak, with the remaining cohorts (1984-86) of about equal strength (NSTC 1987, 1988). Model evaluation Exploratory runs were performed using L,„f values ranging from 20 to 50 mm, in 1-mm steps, and K values ranging from 0.20 to 0.50, in 0.01 steps, encompass- ing a very wide range of values (about ± 50%) around 764 Fishery Bulletin 88(4). 1990 LU Q_ CC LU CQ < LU Q < cc \- C/5 400 200 H 400 200- 400- 200- 1984 1985 1986 1984 YC 1882 YC AQE 2 AQE 4 1987 1988 10 15 20 25 30 35 CARAPACE LENGTH (mm) Figure 2 Length-frequency distributions (stratified mean number per tow) for Pandalvs borealis collected in the western Gulf of Maine during the 1984-88 northern shrimp surveys aboard RV Gloria Michelle. pairs within a given ridge tended to provide similar in- terpretations of the number of modes (assumed age groups) present in the observed length frequency, as expected from earlier testing of the method (Shepherd et al. 1987). Values of the optimum value of to corre- sponding to K and Lmr parameter pairs are the deci- mal fraction part of to, and are indeterminate with respect to the addition or subtraction of any whole number of years (Shepherd 1987). SRLCA results were examined in the context of previously developed parameter estimates and a priori assumptions, and alternative parameters were evalu- ated when the highest scoring values did not agree with prior interpretations of the number and position of modes (assumed age groups) expected. SRLCA was first applied to annual distributions independently, and then to distributions from 1984 to 1988 sequentially pooled in a single run. The annual distributions were analyzed primarily to evaluate inter{)retations provided by SRLCA given the variable patterns of growth and recruitment in the data. Growth parameters and subse- quently resolved age frequencies from the final pooled analysis were used to estimate total mortality for com- parison with rates estimated by visual resolution of the length-frequency data to age. Shepherd (1987) noted that the number of older ages determined by decomposition of the length-frequency distribution according to the von Bertalanffy growth equation is dependent to a large degree on the value of L,nf, and suggested that parameters selected by SRLCA might be most appropriate for subsequent use in other length-based analyses, rather than to slice length frequencies to ages, unless additional data (e.g., knowledge of the expected number of cohorts) are available to select the ridge containing the "correct" parameter pair (see also Shepherd et al. 1987). In this exercise, we elected to proceed with resolution to cohorts, and subsequent age-based mortality estima- tion, both because of the apparently nonequilibrium nature of this northern shrimp population (thus limiting the utility of length-based methods which assume steady-state conditions), and to provide results com- parable with those previously estimated (NSTC 1984, 1985, 1986, 1987, 1988). the growth parameters previously estimated for this stock of P. borealis, in order to adequately explore the high score ridges provided by the SRLCA score func- tion. Preliminary evaluation of the highest scoring parameter values from each ridge, along with several local maxima (higher score than all eight nearest neighbors) within each ridge, indicated that parameter Results Annual length frequencies 1 984 distribution This distribution is characterized by a dominant mode centered at 19.5 mm (Fig. 2). A priori interpretation suggested that the dominant mode at 19-20 mm should be age-group 2 shrimp (a strong 1982 YC), with shrimp 22-24 mm probably age- Terceiro and Idoine: SRLCA application to Pand3lus borealis survey data 765 Table 1 Von Bertalaiiffv growth parameter [jairs ami SRLCA scoi'e (S) for 1984 northern shrimp survey length-frequency distribu- tion: primary and secondary ridge crests. L,„( (mm) K Primary ridge crest 35 0.50 47.0 40 0.43 63.9 45 0.36 64.4 50 0.30 62.7 39 0.50 67.7 41 0.43 65.8 35 0.42 41.4 40 0.30 37.4 36 0.37 40.3 Secondary ridge crest 'High score parameters of ridge crest, on hnundary of ex- plored space. 'High score parameters of ridge crest, nonboundary max- imum. 'Parameters selected for final evaluation. group 3, and animals >25 mm defined as age 4-t- (NSTC 1984). SRLCA scores exhibited a broad, pri- mary (highest scores) ridge of scores within the ex- plored parameter space ranging from a high score on the border of explored parameter space at Linf = 39 min, K = 0.50 to L,nf = 50 mm, K = 0.30, with a non- boundary local maximum at L,„i = 41 mm, K = 0.43 (Table 1, Fig. 3). Based on parameters in this ridge, the length frequency was interpreted as a single domi- nant mode (age-group 1), with the distribution to the right of the mode classified as two older age-groups, for a total of three cohorts present. A secondary ridge of scores, from Li„f=33 mm, K = 0.50 to L|„f=38 mm, K = 0.30, with a nonboundary local maximum at L,„f = 36 mm, K = 0.37 (Table 1, Fig. 3), classified four ages and defined the dominant mode as age 2, but did not interpret the small modes at 22.5 and 25.5 mm in accordance with prior assumptions (assumed age groups 3 and 4, respectively). However, this secondary ridge most closely matched the n priori interpretation, and represented the best performance of SRLCA for the 1984 data (Fig. 4). A tertiary ridge (a spur of the secondary) ranged from a nonboundary maximum at L,„f = 31 mm, K = 0.46 to L,„,-= 34 mm, K = 0.30, and interpreted the length frequency in nearly the same manner as parameters from the secondary ridge. 1 985 distribution Previous vvoi'k suggested that the first two modes of this distribution. 13-18 mm and 20-24 mm, should be interpreted as age groups 1 (1984 YC) and 3 (the strong 1982 YC), thus confronting SRLCA with a "missing cohort" situation (the weak 1984 DISTRIBUTION 1 ^/ Linf 25 ^ i ^^^ -60 60- ^^k UJ CC O -40 O t/J 4^_ \l ^/viNiJc; ^^H L -20 20- '^^^^^w ^f- \ \ \\rH\KR^^^^^^ ^a^^ >^ 45 \^^^^^ r' S ^~~S^i^ Figure 3 Response surface of the SRLCA score function for the noi-thern shrimp survey 1984 length-frequency distribution. p80 X 5 600 o t— 1982 YC O cc UJ Q. 1984 ^"X -60 y Ct / \ 1— ^ 400- / \ A < 2 3 \ / \/^~ -40 ^ z Vy o < UJ 2 200- Ji o (/) -20 ^ > Q UJ ^'^x r „ !< u 1— rrfL-r-TL -0 —1 2 --20 tJ 5 1 1 1 1 1 0 15 20 25 30 3 CARAPACE LENGTH (mm) Figure 4 Northern shrimp survey 1984 length-frequency distribution (histo- gram) and pattern of SRLCA scores (solid curve) for growth param- eters selected for final evaluation (L,„, = 36 mm. K = 0.37). 1983 YC; Fig. 2). Sex determinations further showed that shrimp in the 20-24 mm mode were not the ex- pected first-year females, but mostly mature males that seemed not to have undergone transition during the previous winter (NSTC 1985). Inspection of the posi- tion of this age-3 mode, relative to the modal lengths 766 Fishery Bulletin 88(4), 1990 Table 2 Voii Bertalanffy growth parameter pairs and SKLCA .score | (S) for 1985 northern shr mp survey ength- frequency distribu- tion: primary, secondary, and tertiary ridge crests L,„f (mm) K S Primary ridge crest 35 0.48 85.2- 40 0.34 82.6 45 0.26 78.7 50 0.21 75.2 Secondary ridge crest 42 0.50 (;3.0 46 0.41 69.7 50 0.34 74.1' Tertiary ridge crest 28 0.50 29.0 ' 30 0.37 23.6 32 0.30 22.5 34 0.27 21.4 36 0.23 23.4 38 0.20 23.8 29 of ridge c 0.42 25.3-* ■est, on boundarv nf ex- ' High score parameters- plored space. "' High score parameters of ridge crest, nonbound; I'V max- mium. 'Parameters selected f( r final eva nation . 1985 DISTRIBUTION LInf t-__j^i__^ 35 30 — ' 1__ 25 -80 80- m ^^m 60 ^ DC O o -40 '^ SCORE 8 \ I |j|m|lf ^ '\\\ -20 40- V UMmllii^w »- / ^^B 46 ^ I, ^ ^ — +- Figure 5 Response surface of the SRLCA score function for the northern shrimp survey 1985 length-frequency distribution. of assumed age-3 shrimp in the two subsequent years, suggests that the 1982 YC may also have experienced a slower growth rate between ages 2 and 3 than preceding cohorts. The highest value in the SRLCA primary score ridge was at L„,i- = 35 mm, K = 0.48, with the ridge then continuing to L|„f = 50 mm, K = 0.21. Parameters in this primary ridge classified the first two modes as suc- cessive cohorts, with three age groups total. A second- ary ridge, from L,nf = 42 mm, K = 0.50 to a boundary maximum at Linf = 50 mm, K = 0.34, resolved the distribution to only two age groups, with modal lengths of 17.0 and 26.5 mm for ages 1 and 2. It was necessary to explore a tertiary ridge, with a nonboundary local maximum at L|„, = 29 mm. K = 0.42, to successfully in- terpret the first two length modes as age groups 1 and 3, with shrimp >24.5 mm resolved to two main age- groups (Table 2; Fig. 5, 6). o < UJ Q 1985 1982 YC AGE 3 fin i .' ft/^ Jin Ww^ 10 15 20 25 30 CARAPACE LENGTH (mm) 80 35 h40 ^ O o (/) > _i 2 n h-20 ^ Figure 6 Northern shrimp survey 1985 length-frequency distribution (histo- gram) and pattern of SRLCA scores (solid curve) for growth param- eters selected for final evaluation (L,„, = 29 mm, K = 0.42). 1 986 distribution This distribution is characterized l)y three clearly defined modes at 13-18 mm, 19-22 mm, and >25 mm, assumed to be ages 1, 2, and 4, and a small mode at 23 mm assumed to be either the 1983 cohort at age 3 or slow-growing shrimp from the 1982 cohort at age 4 (NSTC 1986; Fig. 2). Two ridges of com- parable high scores were apparent from an initial SRLCA run, a primary ridge from L,,,)- = 35 mm, K = 0.49 to a boundary maximum at L,,,! = 50 mni, K = 0.22, and a secondary ridge from Linf = 32 mm. K = 0.47 to L,nf = 40 mm," K = 0.23 (Table 3, Fig. 7). The primary ridge identified three modal groups, matching the first observed mode well, but classifying shrimp between 18.5 and 24.5 mm as a single group, with the Terceiro and Idoine: SRLCA application to Pand^lus borealis survey data 767 Table 3 Von Bertalanffy growth parameter pairs ami SRLCA score (S) for 1986 northern shrimp survey length-frequency distrilju- tion: primary and secondary ridge crests. L,„( (mm) K S Primary ridge crest 35 0.49 65.5 40 0.34 74.1 45 0.26 78.0 50 0.22 79.9 ' Secondary ridge crest 32 0.47 65.1 34 0.37 65.6 36 0.31 63.1 38 0.27 59.9 40 0.23 55.7 33 0.42 66.1 -* ' High score parameters of ridge crest, on boundary of ex- plored space. - High score parameters of ridge crest, nonboundary max- imum. 'Parameters selected for final evaluation. last mode interpreted as age-3 shrimp. Tlie secondary ridge high-score parameters classified the first, second, and fourth observed modes fairly well, with the highest scoring parameters (L„,f = 33 mm, K = 0.42) falling in an interval between previously estimated von Ber- talanffy parameters. SRLCA was thus able to match the a priori interpretation of the 1986 length-frequen- cy distribution reasonably well, using parameters from the secondary ridge crest (Fig. 8). 1 987 distribution The first mode of this distribution is fairly pronounced and has been interpreted as age 1 , but the modes following are not well defined (NSTC 1987; Fig. 2). SRLCA indicated best fitting parameter pairs in a primary score ridge ranging from Linf = 35 mm, K = 0.50 to Li„f= 50 mm, K = 0.23, with a max- imum at Ljnf = 40 mm, K = 0.36, and in a secondary ridge ranging from Ljnf = 32 mm, K = 0.50 to L,„f = 40 mm, K = 0.24, with a nonboundary local maximum at L,nf = 33 mm, K = 0.43 (Table 4, Fig. 9). The pri- mary ridge parameters defined three age groups, correctly classifying the first mode, lumping shrimp between 19.0 and 25.0 mm as a second group, and aggregating all shrimp >25 mm as a final group. The secondary ridge parameters resolved the distribution in a similar manner, but by splitting shrimp >23 mm into two groups (Fig. 10). Thus, neither set of param- eters selected by SRLCA successfully resolved shrimp between 19.0 and 25.0 mm to distinct cohorts. 1 988 distribution In the 1988 survey length frequen- 1986 DISTRIBUTION Figure 7 Response surface of the SRLCA score function for the northern shrimp survey 1986 length-frequency distribution. o < 1986 80 -60 o o UJ > 5 10 15 20 25 30 CARAPACE LENGTH (mm) 35 Figure 8 Northern shrimp survey 1986 length-frequency distribution (histo- gram) and pattern of SRLCA scores (solid curve) for growth param- eters selected for final evaluation (L , = 33 mm. K = 0.42). cy, the large mode between 13 and 19 mm was assumed to be age-1 shrimp. A smaller mode of 20-23 mm was interpreted as age 2, with shrimp >25 mm assumed, as usual, to be age-group 4+ (NSTC 1988; Fig. 2). SRLCA provided two ridges of high scoring param- eters, a primar'y ridge extended from a boundary max- 768 Fishery Bulletin 88(4), 1990 Table 4 Von Bertalanfty growtii pai ameter paii s and SRLCA score 1 (S) for 1987 northern slin mp survey length-frequency distribu- | tion: primary and secon dary ridge crests. L,nf (mm) K S Primary ridge crest 35 0.50 55.0 40 0.36 57.1- 45 0.28 56.7 50 0.23 55.9 Secondary ridge crest 32 0.50 42.5' 34 0.39 40.1 36 0.32 37.9 38 0.28 35.5 40 0.24 34.0 of 33 ridge crest, 0.43 41.3-' on boundary of ex- ' High score parameters [ijoi-ed space. - High score parameters of ridge crest, nonbound arv max- imum. •F'arameters selecte o /l98e YC !5 1— _y AGE 1 Tl „ -0 —1 3 " 1 1 I 1 1 ■'" 10 15 20 25 30 35 CARAPACE LENGTH (mm) Figure 10 Northern shrimp survey 19,S7 length-frequency distribution (histo- gram) and pattern of SRLCA scores (solid curve) for growth param- eters selected for final evaluation (L,,,, = 33 mm, K = 0.43). Table 5 \'on Bertalanffy growth parameter pairs and SRLCA score (S) for 1988 northern shrimp sui-vey length-frequency distribu- tion: primary and secondary ridge crests. L,„, (mm) K Primary ridge crest Secondary ridge crest 43 0.50 123.5' 45 o.4(; 123.0- 47 ().41i 122.4 49 0.39 121.9 35 (1.47 93.3- 4(1 (1.34 88.5 45 (1.25 85.0 50 (1.20 85.0 High score parameters of ridge crest, on boundary of ex- jilored space. High score parameters of ridge crest, nonboundary max- imum. 'Parameters si'lected for final exaluatinii. 21.5 mm as a single cohort (age 1), with animals 22 mm an(i larger assigned to age 2. Values from the second- ary ridge provided only a slightly improved interpre- tation, with shrimp 11-19.5 mm classified as age 1. 20-25 mm as age 2, and >25 mm as age 3. Clearly, the modes in this distribution were not sufficiently distinct to allow a reasonable interpretation using SRLCA (Fig. 12). imum at L,,,,- = 43 mm, K = 0.50 to L,,,,- = 49 mm, K = 0.39, while a secondary ridge ranged from L|„f = 35, K = 0.47 to L,„f = 50 mm. K = 0.20 (Table 5, Fig. 11). Values from the primary ridge resolved shrimp 1 1- Pooled distribution: 1984-88 As with the annual disti'ihutioiis, an exploratory run was made with Lj,,!- values ranging from 20-50 mm, in 1-mm steps, and K values of 0.20-0.50, in 0.01 steps. Two broad regions Terceiro and Idoine" SRLCA application to Pandalus borealis survey data 769 1988 DISTRIBUTION Linf '/ l—^L__^°___^t___^ 25 ^^ . /m ^^Ww, -100 m llll^^^»K -'6 ^ »0- M ^^^Ml QC O O -50 ^ 76- \N \ ^ JV\ ^^\M^\$fiwi\t'^A^^ \[f Efi^fflllW ^ -25 60- /' ^^^B^ 'hi^ 2B~ 46 ~~^'^r~'^~~^ w S Figure 1 1 Response surface of the SRLCA score function fur the nnrthern shrimp survey 1988 length-frequency distriliutinn. -120 -j- " 1987 YC AGE 1 O -100 z Q_ 1988 \ /\ r~^ Ld Ct \ / \ / -80 H- ^ «0- \y ^ < 2 Z) Ld -60 q: o z < „ o -40 ^ Ld 2 200- LJ > Q UJ r -20 5 ■ W MUl ]lw„ --20" 5 1 I 1 1 [ 0 15 20 25 30 3 CARAPACE LENGTH (mm) Figure 12 Northern shrimp survey 19SS length-frequency distribution (histo- gram) and pattern of SRLCA scores (solid curve) for growth param- eters selected for final evaluation (L,,,, = 35 mm. K = 0.47). of parameters, with a primary ridge in one region and secondary, tertiary, and quaternary ridges in the other, were evaluated in an attempt to find parameters selected by the SRLCA test function that would inter- pret the 1984-1988 distributions in accordance with a priori assumptions. Table 6 \'on Bertalanffy growth parameter pairs and SRLCA score 1 (S) for pooled 1984-88 northern shrimp survey length-fre- quency distribution: primary, secondary. tertiary. and quater- nary ridge crests. L,„[ (mm) K S Primary ridge crest 43 0.50 244.0 44 0.47 244.8 46 0.43 245.7 48 0.40 246.5- 50 0.37 247.8' Secondary ridge crest 35 0.50 219.2 35 0.48 227.0 36 0.44 230.3 38 0.38 233.7 40 0.34 234.0 42 0.30 234.9- 44 0.27 234.2 46 0.25 233.4 48 0.23 232.5 50 0.21 232.1 Tertiary ridge crest 32 0.50 190.0 32 0.49 191.4- 33 0.43 175.0 34 0.39 160.0 35 0.35 147.3 37 0.29 127.9 40 0.23 108.9 43 0.20 98.3 Quaternary ridge crest 33 0.33 101.5 33 0.32 104.6'* 34 0.29 94.8 35 0.26 88.2 36 0.24 80.7 37 0.22 76.2 38 of ridge crest. 0.20 73.3 on boundarv of e.x- ' High score parameters pliired space. ' High score parameters of ridge crest nonboundary ma.x- | mium. •Parameters selected for final evaluation. The primary ridge ranged from L,„f = 43 mm. K = 0.50 to a boundary maximum at L,„f=50 mm. K = 0.37. The highest nonboundary score was at L,,,,- = 48.0 mm, K = 0.40. A secondary ridge ranged from L,„f=35 mm, K = 0.50 to L,„f = 50. K = 0.21, with a nonboundary maximum score at L,,,,- = 42.0, K = 0.30. A tertiary ridge ranged from L,nf = 32 mm, K = 0.50 to Linf = 43 mm, K = 0.20, with a nonboundary max- imum at Li„f = 32 mm, K = 0.49. A quaternary ridge ranged from L,n, = 33 mm, K = 0.33 to Ljnf = 38 mm, K = 0.20, with a nonboundary maximum score at L,nf = 33 mm, K = 0.32 (Table (3," Fig. 13). 770 Fishery Bulletin 88(4), 1990 1984-88 POOLED DISTRIBUTION Linf 1 .1 J ^^_4^^ 30 25 / ^: — ' — -^-^—1^ / ^^ ■A -200 XM- 1 \ \ \\\\ ^m -150 ^^ CC o o -100 <^ KO^ ' \\m\!\\MWreMl 1 i -60 K>0- [\ W mM^mi\w¥»k , 50- { W / \ / \i^-'^^^^ff^^/'''" ■ ' " / / WMivf^^Si&F" ° Unf '° ^ Figure 13 Response surface of the SRLCA score function for the northern shrimp survey 1984-88 length-frequency distribution. Values from the primary ridge tended to treat the assumed first and second observed age-groups, up to 21 mm, as a single mode, with animals greater than 21.5 mm classified as age 2. Values from the secondary ridge accurately interpreted the assumed age-group 1, but lumped the assumed second and third age-groups as a single mode, with shrimp >25 mm classified as age 3. Values from the tertiary ridge also correctly characterized the first age group and failed to resolve the second and third modes to distinct age classes, but split shrimp >24 mm into age-groups 3 and 4. It is in- teresting to note that the two sets of previously derived von Bertalanffy parameters for this stock noted earlier (see Methods: A Priori Assumptions) fell within this tertiary ridge of scores. This indicated that growth rates based on those parameters were too fast to pro- vide an interjjretation of length distributions consistent with visual inspection of the length modes. Quaternary ridge values successfully classified the first-fourth assumed age-groups in line with prior assumptions. These parameters also provided an age- group 5 for CL >27.5 mm. The quaternary ridge non- Ixiundary maximum score parameters (L,„f = 33 mm, K = 0.32) were selected for final SRLCA evaluation of the pooled length frequency, as these values provided the characterization of the pooled distribution closest o CD =) Z Z < < 1984-88 X o o o LJ > _1 o 15 20 25 3D CARAPACE LENGTH (mm) 35 Figure 14 Northern shrimp survey 1;)cH)led length-frequency distribu- tion (Tiistogram) and pattern of SRLCA scores (solid curve) for growth parameters selected for final evaluation (L|„( = 33 mm, K = 0.32). to the assumed age structure (Fig. 14). The growth curve defined by these parameters indicates a slower rate compared with previously estimated growth curves (data from Haynes and Wigley 1969, Mclnnes 1986), a result of the influence, or "bias," of the abun- dant, apparently slower-growing, 1982 cohort on the score function (Fig. 15). We derived annual age frequencies by slicing annual length frequencies to age-groups according to the max- imum-scoring growth parameters provided by the four score ridges of the pooled analysis (Table 6). Since shrimp of age-group 2 and older are assumed fully recruited to the survey trawl gear, total instantaneous mortality estimates (Z) were derived from: In (^ age 2 + for Year^ / X '^>^'^ ^ + f"*" Yearx + 1 ). These values were then compared with jireviously calculated estimates of total mortality from the length- frequency data (NSTC 1985, 1986, "l987, 1988). Table 7 shows the widely divergent age frequencies and subsequent Z estimates provided by the four different pairs of growth parameters, which again illus- trate a necessity for some external source of informa- tion to interpret SRLCA results in a manner consis- tent with a priori assumptions. We note that for the 1987 distribution, the quaternary ridge parameters and NSTC visual inspection sliced the length-frequency distribution at the same length intervals, resulting in identical age frequencies for 1987. As expected, use of the quaternary SRLCA score ridge parameters to Terceiro and Idoine- SRLCA application to Pandalus borealis survey data 771 _^ ^^^ ^O 25- y^y ... ■ ? / / b 20- HW 69 / / T / / ,• •■ Z / / ..••■ SRLCA 15- / / ■■' UJ f ) / ^ ■' n^ / ^■'' < rr 10 H / ^■' < 1 / ■' u // 5- 0- 1 1 I 1 1 0 12 3 4 5 AGE Figure 15 Von Bertalanffy growth curves for Gulf of Maine PtiHt/a/Msftorpalfs. Upper curve (soliti line) is for parameters derived from age-length data in Haynes and Wigley (1969); center curve (large dashed line) is for growth parameters cited in Mclnnes(1986); lower curve (dot- dashed line) is for parameters derived by SRLCA for northern shrimp survey length-frequency data. age the length frequencies provided mortahty esti- mates consistent with those produced by visual inspec- tion of length modes. Discussion We subjected SRLCA to a fairly stern test by attempt- ing to interpret a data set exhibiting variable recruit- ment and growth patterns, and by using a broad ini- tial parameter search space. As noted in the Monte Carlo tests of SRLCA by Basson et al. (1988), these variations in recruitment, and in mean length-at-age (presumed variable growth rate, especially for the abundant 1982 cohort) between cohorts, made inter- pretation of the northern shrimp length-frequency distributions difficult. The shape and proximity of the assumed age-2 and -.3 modes frequently caused SRLCA to interpret these modes as a single age-group, result- ing in highest scoring parameters that provided posi- tively biased estimates of growth rate. This problem was most severe for the annual distributions and per- sisted in the pooled length frequency, although the increased amount of information in the pooled distribu- tion did increase the effectiveness of the SRLCA ap- proach, witli a "correct" interpretation of the data available from the quaternary score ridge. Analysis of Table 7 Age-frequency matrices and instantaneous total mortality rates (Z) for age 2+ northern shrimp estimated by growth parameters from SRLCA high-score parameter ridges for the 1984-88 pooled distribution, compared with estimates using method of visual inspection of length modes (NSTC 1984, 1985. 1986, 1987, 1988). Age 1984 1985 1986 1987 1988 SRLCA Primary score 1 '2150 ridge 887 1489 1091 3192 2 856 2644 1872 1357 1136 3 0 0 0 0 0 i+ 0 0 0 0 0 Zss = z. = - z,, = - z,, = - Secondary score ridge 1 633 660 753 575 2903 2 2001 2083 1504 1257 818 3 373 789 1104 620 604 4+ 0 0 0 0 0 Zsr, = = 1.10 Z, , = 0.96 Zs; = 1.44 Z,, = 1.13 Tertiary score 1 ' 672 ridge 660 755 577 2906 2 18.53 1849 1283 1088 693 3 410 876 857 545 504 4-1- 71 146 465 239 224 Zsr, = = 0.83 Z, , = 0.78 Z,- = 1.20 Z,, = 0.94 Quaternary score ridge 1 193 634 701 539 2564 2 1987 350 845 575 632 3 425 1710 615 663 433 4+ 401 837 1200 671 699 Z.V-, '- = 0.10 Z, , = 0.47 Zs7 = 0.69 Z,, = 0.52 NSTC Visual inspection 1 49 646 710 539 2828 2 2051 337 959 575 614 3 442 1 596 491 663 187 4 + 463 952 1200 671 699 Zsr, = -0.15 Z, = 0.53 Zs, = 0.69 Z,, = 0.77 the pooled data in a truly sequential fashion, after the projection matrix approach of Rosenberg et al. (1986) and Basson et al. (1988), as a supplement to SRLCA might help in alleviating these problems. This exercise demonstrated that the best objective fit obtained by SRLCA does not necessarily provide the best interpretation of the data, as with most of the existing length-frequency distribution analysis meth- ods. Pragmatically, we could not rely on SRLCA to pro- vide a single set (or even region) of growth parameters that yield both the highest parameter score and the "correct" interpretation of the data, unless supple- 772 Fishery Bulletin 88(4), 1990 mented by an external source of information. For SRLCA to be effective, the subjectivity required in the selection of input parameters for use with the family of distribution mixture methods (e.g., MacDonald and Pitcher 1979) must instead be applied to interpretation of the output of the procedure. These problems are similar to those encountered by others in evaluating the ELEFAN I method of Pauly and David (1981). Testing of ELEFAN I has indicated it produces biased results unless the range of growth parameters considered for testing is relatively narrow, and that it is sensitive to increased variation in length- at-age or increased variation in recruitment timing (Rosenberg and Beddington 1987). Recent work by Morgan (1987), supplementing ELEFAN with age- length data, improved the performance of that method by allowing selection of the appropriate parameters from among several locally optimal solutions. We are encouraged, however, that such external in- formation requirements were moderate for SRLCA. For instance, with an expectation of age of the oldest cohort, we would have immediately selected the quater- nary ridge of the pooled analysis as best. Or, had we limited the initial search space for Li„f to within 10% of the CL of the largest animal, and/or 10% of an "average" L,„f given previously derived parameters (e.g., 33 + 2 mm), we again could have proceeded direct- ly to selection of the appropriate score ridge. Even given conditions of variable growth and recruit- ment, we were able to inspect the response surface of the test function and, using a moderate degree of sub- jectivity and a stepwise procedure of evaluation, select growth parameters from alternative high-score ridges that resolved the pooled distribution in a manner con- sistent with previous interpretations, and, more impor- tant, in a more satisfactory manner than previously derived von Bertalanffy growth parameters. Overall we found SRLCA to be a simple and generally effec- tive tool for the estimation of growth parameters, and subsequently age frequencies, from length-frequency distributions for Gulf of Maine northern shrimp. Acknowledgments We thank the ships officers, crew, and scientific staff of the RV Gloria Michelle and the members of the Northern Shrimp Technical Committee for their efforts in collecting the data on which this paper is based. We also acknowledge Dr. Steve Clark, Dr. Wendy Gabriel, the members of the 1989 NAFO Working Group on Progress in Age Determination of Pandalus, and an anonymous reviewer for their valuable comments on the manuscript. Citations .Vllen, J. .A. 1959 On the biology of Pnndiilus biirealit; Kr^yer, with refer- ence to a population off the Northumberlami Coa.st. J. Mar. Biol. .Assoc. U.K. ;58:189-2'20. .\pollonio, S., D.K. Stevenson, and E.E. Dunton Jr. 1986 Effects of temperature on the biology of northern shrimp, Fandahis borealis, in the Gulf of Maine. NOAA Tech. Rep. NMFS 42, 22 p. Basson, M., A. A. Rosenberg, and J.R. Beddinprton 1988 The accuracy an 100 m (Bigelow and Schroeder 1955, Leim and Scott 1966). They are the deepest ranging species described in this study, being found between 162 and 366 m, the survey extreme. They were collected within a relatively restricted temperature range throughout the year, with little or no seasonal changes. Silver hal280 mm for 2-year-olds. If the spotted hake captured during this study follow the size at ages given by Barans (1969), the population present in spring is com- posed primarily of older fish, with young-of-the-year and yearlings migrating into the area during summer and fall. The apparent split spawning season observed during this study may be attributable to dissimilar %RIPE 0 40 (6 5 I — — 25 67 U 0 0 16 ' ^ ' ^11 Spotted Hake 62 -SOMATIC INDEX 03 ro f £ r 9 T o 0 14 J L K H H J J A S 0 N D J F M A M J 1974 1975 Figure 6 Annual cycle of gonad-somatic indices for spotted hake Urophycis regia collected in the New York Bight, June 1974-June 1975, in- cluding monthly means, ranges. 95% confidence intervals, and number examined. In addition, percent ripe females per month is given on the upper axis. Figure 7 Annual cycle of gonad-somatic indices for black sea bass Cmitropristis striata collected in the New York Bight. June 1974-June 1975, in- cluding monthly means, ranges, 95% confidence intervals, and number examined. In addition, percent ripe females per month is given on the upper axis. distributional patterns of specific age groups at the northernmost extreme of the species range. Black sea bass Centropristis striata Mean GSI was highest during July, with ripe females, 184-452 mm, collected from May through July (Fig. 7). Black sea bass are protogynous hermaphrodites with the size at which females become males extremely 780 Fishery Bulletin 88(4), 1990 Figure 8 Annual cycle of gonad-soniatic indices for butterfish Peprilti^ triaran- thua collected in the New York Bight, June 1974-June 1975, including monthly means, ranges, 95% confidence intervals, and number ex- amined. In addition, percent ripe females pjer month is given on the upper a.xis. Figure 9 Annual cycle of gonad-somatic indices for northern searobin Pritimi- tus carolinus collected in the New York Bight, June 1974-June 1975. including monthly means, ranges, 95% confidence intervals, and number examined. In addition, percent ripe females per month is given on the upper axis. variable (Kendall and Mercer 1982). In this study female and male black sea bass were in the range 173-452 and 178-556 mm, respectively. Bigelow and Schroeder (1953) reported spawning from mid-May through June off southern New England and New Jersey. To date, black sea bass eggs have not been reported from plankton collections. However, larvae have been collected from Long Island, New York waters from mid-May to June (Perlmutter 1939) and from offshore New York Bight waters from June through November (Kendall 1977). Black sea bass are common from Cape Cod, Massa- chusetts to Florida (Hildebrand and Schroeder 1928, Bigelow and Schroeder 1953). In the Middle Atlantic, they migrate northward and inshore in spring and off- shore and southward in fall (Kendall and Mercer 1982). Mean temperature and depth of occurrence in July, the month of highest mean GSI, were among the warmest and shallowest of the survey (Table 1). Butterfisli Peprilus triacanthus Highest mean GSI occurred in June 1974 and again in May and June 1975, with ripe females, 124-242 mm, being present throughout the survey area from May through August (Fig. 8). The butterfish is a migratory semi-pelagic species which ranges from Newfoundland, Canada to Florida (Bigelow and Schroeder 1953, Leim and Scott 1966). Colton et al. (1979) reported butter- fish spawning in the Middle Atlantic from May through October with a peak in July or August. Smith et al. (1980) collected eggs off North Carolina during spring and larvae off New Jersey during early summer. Kawahara (1978) reported the spawning time of butter- fish in the Middle Atlantic as April or May through August. Butterfish were collected at a wide range of temperatures and depths during the months of peak GSI (Table 1). Northern searobin Prionotus carolinus Highest mean GSI was in July; however, ripe females, 143-341 mm, occurred from May through September (Fig. 9). According to Richards et al. (1979), based on data from the north shore of Long Island, New York, GSI rose early in May, peaked in late May, and declined gradually through June and July. Colton et al. (1979) reported spawning from Block Island, Rhode Island to Cape Hatteras, North Carolina from May through November. Eggs and larvae of Prionotus spp. are among the dominant ichthyoplankton occurring in the Middle Atlantic from New Jersey to Cape Hatteras, North Carolina during June through October (Smith et al. 1980, Sherman et al. 1984). This survey demonstrated that northern searobin favored warm shallow inshore waters during peak spawning (Table 1); however, from November through April they favored depths >50 m. Striped searobin Prionotus evolans Highest mean GSI occui'red during June and .luly 1974 and again in June 1975, with ripe females, 204-414 mm, present from May through August (Fig. 10). Richards et al. (1979) determined spawning time for striped searobin to be almost the same as northern Wilk et at - GSI annual cycles as indicators of finfish spawning in the New York Bight 781 rOO 75 36 %R1PE 0 0 0 — — — 50 80 20.0- LU 2 15.0 o I- < O10.0 CO < i 5.0 0.0 I I I I Str/ped Searobin 19 34 25 ^^- JJ ASONDJF 1974 M A M J 1975 Ffgure 10 Annual cycle of gonad-somatic indices for striped searobin Prionotus evolans collected in the New York Bight, June 1974-June 1975, in- cluding monthly means, ranges, 95% confidence intervals, and number examined. In addition, percent ripe females per month is given on the upper axis. % RIPE Summer Flounder J J A S O N D 1974 J F M A M J 1975 Figure 1 1 Annual cycle of gonad-somatic indices for summer flounder Para- lichthys dentatus collected in the New York Bight. June 1974-June 1975, including monthly means, ranges, 95% confidence intervals, and number examined. In addition, percent ripe females per month is given on the upper axis. searobin. Perlmutter (1939) collected larvae off Long Island, New York between May and June; and Herman (1963) reported spawning from June through August off Narragansett Bay, Rhode Island. This study indicated temperature preferences dur- ing spawning were somewhat cooler than those of the northern searobin, while depths were similar (Table 1). However, throughout the remainder of the year, striped searobin were collected in comparatively warmer and shallower waters than northern searobin. Summer flounder Paralichthys dentatus Ripe females, 309-716 mm, were collected from Sep- tember through November, and sporadically from February to June (Fig. 11). Although there were several slight peaks in GSI during the spawning season, these probably resulted more from individual sample variability than actual peaks in spawning activity (Fig. 11). Morse (1981), using additional maturity data from the New York Bight, reported spawning from Septem- ber through March with a peak in October. Smith (1973) reported that eggs and larvae began to appear in September off Long Island, New York and southern New England, and occurred farther southward as fall progressed. After December, recently hatched larvae were common only south of Chesapeake Bay (Smith 1973). Summer flounder range between Nova Scotia, Canada and Florida with their center of abundance lying in the Middle Atlantic region (Leim and Scott 1966, Gutherz 1967, Wilk et al. 1980). They migrate to shallow coastal waters and bays in summer, then offshore in fall (Bigelow and Schroeder 1953). At the GSI peak in October, developing females were collected in relatively warm shelf waters (Table 1). Depths of col- lection for summer flounder varied during the year, from a mean of 11 m in July to 92 m in April, reflect- ing their seasonal migration pattern. Fourspot flounder Paralichthys oblongus Mean GSI were highest in June and July 1974 and again in June 1975, with ripe females, 153-419 mm, present from April through September (Fig. 12). Smith et al. (1975) found small larvae off North Carolina and Chesapeake Bay in May and June, and off New Jersey and New York in July and August. Fourspot flounder are widely distributed on the con- tinental shelf between Georges Bank and Cape Hat- teras, North Carolina (Bigelow and Schroeder 1953, Gutherz 1967). In the New York Bight, they inhabit the entire shelf, with deepest occurrences during winter (Ralph 1982). Collections from this survey in- dicate they were deepest during February and March, averaging 90 and 92 m, respectively, moving somewhat inshore during the rest of the year, but always at depths averaging >40 m. They were collected over a wide range of depths and temperatures during the June spawning peak (Table 1). Smith et al. (1975) collected larvae between 6 and 9°C at depths of 35-80 m. 782 Fishery Bulletin 88(4), 1990 Figure 12 Annual cycle of gonad-somatic indices for fourspot flounder Para- lichthys oblongus collected in the New York Bight, June 1974-June 1975, including monthly means, ranges, 95% confidence intervals, and number examined. In addition, percent ripe females per month is given on the upper axis. %RIPE 64 10 4 41 3 0- - 3 U IS 40 17 20-1 Windowpane 165 X 150 167 uj 16- Q z 153 142 o K 12 < 101 2 9 119 o "? 8- 132 217 1 / < I J^ / N 1 z Fk. / \ 61— (" y 8 4- J / L —1 [^ ^1 1 - 1 0- J- -^ - I '■ J J ASONDJFMAMJ 1974 1975 Figure 13 Annual cycle of gonad-.somatic indices for windowpane Scophthal- mi<.s aquos'u.'i, collected in the New York Bight, June 1974-June 1975, including monthly means, ranges, 95% confidence intervals, and number examined. In addition, percent ripe females per month is given on the upper axis. Windowpane Scophthalmus aquosus The GSI plot for windowpane was bimodal, with peaks in May and again in September (Fig. 13). Ripe females, 201-387 mm, were collected from February through October, indicating that windowpane may spawn to some degree in the New York Bight during most of the year. Unfortunately, data for winter are lacking. There was no conclusive evidence that the two peaks repre- sented spawning of either different populations or age classes. In September the mean lengths were 261 and 230 mm for females and males, respectively; in May, windowpane were somewhat more abundant and larger with mean lengths of 285 and 247 mm for females and males, respectively. Split spawning of windowpane has been reported previously in the Middle Atlantic region. In 1952 and 1953, Wheatland (1956) found eggs from May through early July and again from late September through November in Long Island Sound, New York. Temper- atures in the Sound were >20°C from late July through September when eggs were not present. Smith et al. (1975) proposed two patterns of spawning for win- dowpane depending on latitude. First, off New Jersey and New York one prolonged spawning season was hypothesized based on the occurrence of 2-6 mm lar- vae on the inner shelf from June through December 1966. Bottom temperatures on the inner shelf that year ranged between 6 and 17°C during August (Clark et al. 1969). Second, off Virginia and North Carolina spawning was interrupted as indicated by the occur- rence of larvae in April through June and again in November and December 1966. During summer 1966 bottom temperatures ranged between 15 and 25 °C (Clark et al. 1969). Smith et al. (1975) therefore con- cluded that in waters >20°C windowpane stop spawn- ing. Windowpane were collected during the May GSI peak at 5.6-13.2°C and 4-51 m; while during the Sep- tember peak they were collected at higher tempera- tures but at similar depths (Table 1). In July and August, when GSI decreased noticeably, maximum bot- tom temperatures were 20-22 °C at nearshore stations. Based on the results of this survey, coupled with the previously cited historical observations, a split spawn- ing season is probably ty|3ical for windowpane in the New York Bight with continuous spawning occurring only during unusually cold summers. Yeliowtail flounder Limanda ferruginea Highest mean GSI occurred from February to April, with ripe females, 212-422 mm, collected in October and November and February through May (Fig. 14). Howell (1983) observed female yeliowtail off Rhode Island beginning to mature in September, with the peak GSI in April. Colton et al. (1979) reported spawn- ing in the Middle Atlantic from April through August with peaks in May and June. Smith et al. (1975) found the largest concentrations of larvae off New Jersey in April and May. They also reported larvae in abundance off southern New England in May and June. Wilk et a\ : GSI annual cycles as indicators of finfish spawning in the New York Bight 783 0 0 0 8 %RIPE 6 58 8 r 63 t3 0 30 B 24 z Q < 18 2 O CO i 12- < 2 o =1= I w Yellowtall Flounder 1 S, ' f E— J J A S O N D 1974 Figure 14 Annual cycle of gonad-somatic indices for yellowtall {\ounder Li man- da feniiffiyiea collected in the New York Bight, June 1974-June 1975, including monthly means, ranges, 95% confidence intervals, and number examined. In addition, percent ripe females per month is given on the upper axis. Figure 1 5 Annual cycle of gonad-somatic indices for winter flounder Pseudo- pleuroiiectes americanus collected in the New York Bight, June 1974-June 1975, including monthly means, ranges, 95% confidence intervals, and number examined. In addition, percent ripe females per month is given on the upper axis. Yellowtail flounder commonly occur from Nova Sco- tia, Canada to Delaware Bay (Bigelow and Schroeder 1953, Leim and Scott 1966). Smith et al. (1975) con- sidered the population center to be on Georges Bank, with migration along the southern New England shelf eastward in spring and westward in fall (Lux 1963). In the New York Bight, few were caught in summer. In March of this survey, the period of greatest gonad development, they were collected on the inner shelf at 9-68 m and 4.4-7.5°C (Table 1). Winter flounder Pseudopleuronectes americanus Highest mean GSI occurred during January in the estuarine sui'vey area and during Febiiiary in the ocean survey area (Fig. 15). Ripe females, 169-432 mm, were collected from September through April wdth 21% oc- curring in Raritan Bay. GSI declined rapidly after February; however, spawners were likely upriver beyond the range of this survey. Smith et al. (1975) cap- tured larvae at inshore stations between the offing of Delaware Bay and Block Island Sound, Rhode Island from April through June, with spawning beginning in the south and progressing northward. Perlmutter (1947) reported spawning from December through May and concluded that the peak time varied with temper- ature. Pearcy (1962) found evidence of spawning from mid-February through April in the Mystic River estu- ary, Connecticut. Croker (1965) found winter flounder larvae in plankton collections in Sandy Hook Bay, New Jersey from April to June. Although the water temperature at actual spawning sites is unknown, the average bottom temperature in Raritan Bay during January, the month of highest mean GSI, was 5.7°C. Of all the species encountered during this survey, winter flounder were collected in the coldest water. Discussion According to Braum (1978), most temperate zone fishes spawn in one of three generic seasonal patterns: spring, summer, or fall-winter. Based on the current study, alewife and yellowtail flounder are spring spawners; silver and red hake, black sea bass, butterfish, northern and striped searobin, and fourspot flounder are sum- mer to fall spawners, while spotted hake and summer and winter flounder are fall-winter through early spring spawners. Offshore hake and windowpane belong to none of these groups, but appear to have pro- tracted spawning seasons, i.e., these species' GSI were elevated, ovaries were ripe, and eggs and larvae were collected from spring through fall. Offshore hake in- habit the deeper waters of the Bight which are less in- fluenced by changes in light and temperature. In con- trast, windowpane inhabit the shallow, more seasonally influenced, inshore environs. In temperate seas, there are distinct annual cycles of light intensity, temperature, nutrients, and winds 784 Fishery Bulletin 88(4), 1990 which drive cycles of productivity. An abundant source of food must be available for adult fish to produce stored fat for subsequent gamete production (de Vlam- ing 1972). Plankton of the right size, density, and qual- ity must be available to pelagic larvae at the critical time of first feeding. As an example, the spawning of yellowtail flounder corresponds to the spring increase in zooplankton off southern New England (Sherman et al. 1984). The maximum concentration of zooplank- ton in the Middle Atlantic occurs from mid- to late- summer which coincides with the peak concentrations of larval searobins and butterfish as well as numerous other species (Sherman et al. 1984). The fishes that reproduce in the New York Bight during winter have already accumulated fat reserves during the summer and fall. Releasing eggs during winter may give a sur- vival advantage to the larvae since many other fishes, potential competitors and predators, have migrated from the area and there is, therefore, little competi- tion at a crucial time of development. A short spawning season, represented by a steep GSI curve with a single peak, is a common strategy in temperate marine zones. Black sea bass, butterfish, and the searobins had the shortest spawning seasons, with ripe fish present for only 3-5 months. These four species are abundant during summer in the Middle Atlantic (Wilk and Silverman 1976ab, Wilk et al. 1977, Colvocoresses and Musick 1984), but not on the more boreal Georges Bank (Overholtz and Tyler 1985). They were collected at the highest average temperatures during their reproductive peaks (Table 1). Most of the other species described are abundant in both regions and were collected in colder waters. Sherman et al. (1984) describe a ubiquitous or pro- tracted spawning strategy for certain species including silver hake and Urophycis spp., in which eggs are pro- duced for several months and, although mortality is high, some larvae survive when environmental condi- tions and food supply are properly matched. In this study, ripe female offshore silver and spotted hake, as well as windowpane, summer, and winter flounder, were collected for 5 or more months. Eggs and small larvae of these species were observed for many more months in other surveys (Clark et al. 1969; Smith et al. 1975, 1980; Colton et al. 1979). Eggs are typically released from the ovaries of fishes in batches (Bagenal 1978). In species with short repro- ductive periods, these batches are released in the span of a few days with all individual fish spawning at about the same time. In multiple or serial spawners, batches of eggs mature and are shed several times during a long spawning season (Bagenal 1978, Burt et al. 1988). Ex- amples of protracted spawners are silver hake which produce as many as three batches of eggs per spawn- ing season (Fahay 1974) and summer flounder which produce up to six batches (Morse 1981). Summer flounder had the lowest mean GSI of any species col- lected during this survey, apparently because their eggs develop in batches and are released over an ex- tended period. In contrast, the highest mean GSI noted were for yellowtail and winter flounder which produce only one batch of eggs per season (Howell 1983, Bur- ton and Idler 1984). In both species, GSI increased gradually in fall through winter as eggs matured, with highest GSI levels being reached just prior to spring spawning. Similarly, Burton and Idler (1984) reported that GSI of winter flounder in Canadian waters grad- ually rose during the 8 months before spawning takes place in June. A high GSI does not always parallel a high level of fecundity (Bagenal 1978). As an example, larger sum- mer, yellowtail, and winter flounder have exhibited fecundities of over 1 million (Bigelow and Schroeder 1953, Morse 1981, Lux and Livingstone 1982); egg production in each of these species is accomplished, however, on different timetables. Spotted hake and windowpane, both protracted spawners, had bimodal GSI curves. Various sizes of developing and ripe female spotted hake were collected in fall, while a few large ripe females were collected in spring. These findings may rellect behavioral dif- ferences among size or age classes, or perhaps the migratory habits of discrete unit stocks. Windowpane exhibited two GSI peaks in the New York Bight dur- ing 1974-75; however, continuous larval production throughout the summer was noted in an earlier survey of the area (Smith et al. 1975). Windowpane are perm- anent residents in the Bight; that is, no evidence of seasonal migration was found during this survey based on observed changes in abundance, size structure, or distribution of collections. Rather than a result of emigration, the bimodal GSI curve exhibited by win- dowpane may very well be a result of a reproductive tactic in which spawning is curtailed during the year if temperature rises above a preferred level. Some of the most important and prolific commercial fisheries taking place in temperate seas around the world are based on families of fishes, such as gadids (codfishes and hakes) and pleuronectids (flounders), that are typically very fecund and produce relatively small pelagic eggs (Garrod and Horwood 1984). Of the species described in this study, the pleuronectids, yellowtail and winter flounder, and the bothid, sum- mer flounder, have been reported to have extremely high fecundities. Such species are well-suited for com- mercial exploitation since they can recover rapidly from natural or man-induced depletion. The gadids described in this study are all small-bodied, mature early at a relatively small size, and produce many tiny pelagic eggs over a protracted spawning period; however, they Wilk et al ■ GSI annual cycles as indicators of finfish spawning in the New York Bight 785 are not as fecund as larger and more robust gadid species such as Atlantic cod Gadus morhua (Hislop 1984). A variety of reproductive strategies and tactics are used by the species to adapt to seasonal shifts in en- vironmental parameters in the Middle Atlantic in general and the New York Bight specifically. In this way the entire region is continually used for reproduc- tion by a wide range of economically and ecologically important species of resident and migratory finfish. These spawning adaptations contribute to the high diversity and abundance of finfish found in the Middle Atlantic region. Acknowledgments We thank the following members of the National Oceanic and Atmospheric Administration, National Marine Fisheries Service, Northeast Fisheries Center, Sandy Hook Laboratory staff: Ms. Alicia Gruber for scientific publication assistance and the typing of numerous drafts; Ms. Michele Cox for preparation of the figures; and Ms. Anne Studholme, Mr. Wallace Smith, and Mr. Anthony Pacheco for review and editing of the manuscript. Citations Bagenal, T.B. 1978 Aspects offish fecundity. In Gerking, S.D. (ed.). Ecology of freshwater fish production, p. 75-101. John Wiley and Sons, NY. Barans. C.A. 1969 Distribution, growth and behavior of the spotted hake in the Chesapeake Bight. M.S. thesis. College of William and Mary, Williamsburg, .53 p. Bigelow, H.B.. and W.C. Schroeder 1953 Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 53, 577 p. 1955 Occurrence off the Middle and North Atlantic United States of the offshore hake, Merluccius albuhiti (Mitchell) 1818, and of the blue whiting GadMs {Microw.esistiiis) poutas- sou (Risso) 1826. Bull. Mus. Comp. Zool., Harv. Univ. 113: 205-226. Braum, E. 1978 Ecological aspects of the survival of fish eggs, embryos and larvae. In Gerking, S.D. (ed.). Ecology of freshwater fish production, p. 102-131. John Wiley and Sons, NY. Burt, A., D.L. Kramer, K. Nakatsuru. and C. Spry 1988 The tempo of reproduction in Iliiphfvsohrycon pulfhripin- nis (Characidae), with a discussion on the biology of 'multiple spawning' in fishes. Environ. Biol. Fish. 22:15-27. Burton, M.P., and D.R. Idler 1984 The rei>roductive cycle in winter fiounder, Pseudopleuro- nectes am^ricanus (Walbaum). Can. J. Zool. 62:2563-2567. Bye, V.J. 1984 The role of environmental factors in the timing of repro- ductive cycles. In Potts, G.W., and R.J. Wootton (eds.), Fish reproduction: Strategies and tactics, p. 187-205. Academic Press, London. Clark, J.W., W.G. Smith, A.W. Kendall Jr., and M.P. Fahay 1969 Studies of estuarine dependence on Atlantic coastal fishes. Data report I: Northern section. Cape Cod to Cape Lookout. RA' Dolphin cruises 1965-66: zooplankton volumes, mid-winter trawl collections, temperatures and salinities. U.S. Dep. Int., Bur. Sport Fish. Wildl. Tech. Pap. 28. 132 p. Colton, J.B. Jr., W.G. Smith, A.W. Kendall Jr., P.L. Berrien, and M.P. Fahay 1979 Principal spawning areas and times of marine fishes. Cape Sable to Cape Hatteras. Fish. Bull.. U.S. 76:911-915. Colvocoresses, J. A., and J. A. Musick 1984 Species associations and community composition of Mid- dle Atlantic Bight continental shelf demersal fishes. Fish. Bull, U.S. 82:295-313. Croker, R.A. 1965 Planktonic fish eggs and larvae of Sandy Hook estuary. Chesapeake Sci. 6:92-95. Gushing. D.H. 1969 The regularity of the spawTiing season of some fishes. J. Cons. Int. Explor. Mer 33:81-92. de Vlaming, V.L. 1972 Environmental control of teleost reproductive cycles: A brief review. J. Fish. Biol. 4:131-140. de Vlaming, V.L., G. Grossman, and F. Chapman 1982 On the use of the gonosomatic index. Comp. Biochem. Physiol. 73A:31-39. Fahay, M.P. 1974 Occurrence of silver hake. Merluccius bilinearis, eggs and larvae along the Middle Atlantic continental shelf during 1966. Fish. Bull.. U.S. 72:813-834. Garrod, D.J., and J.W. Horwood 1984 Reproductive strategies and the response to exploita- tion. In Potts, G.W., and R.J. Wootton (eds.), Fish reproduc- tion: Strategies and tactics, p. 367-384. Academic Press, London. Gross. M.G. (editor) 1976 Middle Atlantic continental shelf and the New York Bight, special symposia, vol. 2. Am. Soc. Limnol. Oceanogr.. Lawrence, KS, 441 p. Grosslein, M.D. 1969 Groundfish survey program of BCF, Woods Hole. Com- mer. Fish. Rev. 31:22-30. Grosslein, M.D., and T.R. Azarovitz (editors) 1982 Fish distribution. MESA Atlas Monogr. 15, NY Sea Grant Inst., Albany, 182 p. Gutherz, E.J. 1967 Field guide to the flatfishes of the family Bothidae in the western North Atlantic. U.S. Dep. Int., Fish Wildl. Serv., Circ. 263, 47 p. 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. Bull. U.S. Bur. Fish. 43. 388 p. Hislop, J.R.G. 1984 A comparison of the reproductive tactics and strategies of cod, haddock, whiting and Norway pout in the North Sea. In Potts, G.W., and R.J. Wootton (eds.). Fish reproduction: Strategies and tactics, p. 311-329. Academic Press, London. 786 Fishery Bulletin 88(4). 1990 Hoar. W.S. 1969 Reproduction. In Hoar, W.S.. and D.J. Randall (eds.), Fish physiology, vol. 3, p. 1-72. Academic Press, NY. Howell, W.H. 1983 Seasonal changes in the ovaries of adult yellowtail flounder, Limnnda femigiywa. Fish. Bull., U.S. 81:341-355. Kawahara, S. 1978 Age and growth of butterfish. Peprilus triacanthus (Peck), in ICNAF Subarea 5 and Statistical Area 6. Int. Cnmm. Northwest Atl. Fish. Sel. Pap. 3:73-78. Kendall, A.W. Jr. 1977 Biological and fisheries data on black sea bass, Cera- tropristis striata (Linnaeus). Tech. Ser. Rep. 7, Sandy Hook Lab., Northeast Fish. Sci. Cent., Natl. Mar. Fish. Serv., Highlands, NJ 07732, 29 p. Kendall. A.W., and L.P. IVIercer 1982 Black sea bass, Ce7itropristis striata. In Grosslein, M.D., and T.R. Azarovitz (eds.). Fish distribution. MESA Atlas Monogr. 15. p. 82-83. NY Sea Grant Inst., Albany. Kissil, G.W. 1969 Contributions to the life history of the alewife. Alosa pneiidiiharengiis (Wilson), in Connecticut. Ph.D. thesis, Univ. Connecticut, Storrs, HI 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. 1963 Identification of New England yellowtail flounder groups. Fish. Bull., U.S. 63:1-10. Lux. F.E.. and R. Livingstone Jr. 1982 Yellowtail flounder. Limanda frrraginea . In Grosslein. M.D., and T.R. Azarovitz (eds.), Fish distribution. MESA Atlas Monogr. 15, p. 117-119. NY Sea Grant Inst.. Albany. Marak, R.R. 1967 Eggs and early larval stages of the offshore hake, Merluc- rius albidus. Trans. Am. Fish. Soc. 96:227-228. Mayer. G.F. 1982 Ecological stress and the New York Bight: Science and management. Estuarine Res. Fed., Columbia, SC, 715 p. Morse. W.W. 1981 Reproduction of the summer flounder, Panilifhlluis den- tatus (L.). J. Fish Biol. 19:189-203. Musick, J. A. 1969 The comparative biology of two American Atlantic hakes. Urophycis rlui.ts and IJ. tenuis (Pisces; Gadidae). Ph.D. thesis. Harvard Univ.. Cambridge, 150 p. 1974 Seasonal distribution of sibling hakes, Urophi/ris chnss and U. tenuis (Pisces, Gadidae) in New England. Fish. Bull., U.S. 72:481-495. Xikolskii, G.V. 1963 The ecology of fishes. Academic Press, NY, 352 p. Overholtz, W.J., and A.V. Tyler 1985 Long-term responses of the demersal fish assemblages of Georges Bank. Fish. Bull.. U.S. 83:507-520. Pearcy. W.G. 1962 Ecology of an estuarine population of winter flounder Pseudopleuronectes anieriranus (Walbaum). Bull. Bin.gharii Oceanogr. Coll. 18:5-78. Perlmutter. A. 1939 \n ecological survey of young fisli and eggs identified from tow-net collections. In A biological survey of the salt waters of Long Island, 1938, part 2. p. 11-71. NY State Con- serv. Dep.. Suppl. 28th Annu. Rep. .'Mbany. 1947 The blackback flounder and its fishery in New England and New York. Bull. Bingham (.)ceanogr. Coll. 11:1-92. Ralph. D. 1982 Fourspot flounder. Paralii-hthys oblonyus. In Grosslein, M.D., and T.R. Azarovitz (eds.). Fish distribution. MESA Atlas Monogr. 15, p. 113-114. NY Sea Grant Inst.. Albany, NY. Richards. S.W., J.M. Mann, and J. A. Walker 1979 Comparison of spawning seasons, age, growth rates, and food of two sympatric species of searobins, Priormtus carolinus and Prioriotus ei'olans. from Long Island Sound. Estuaries 2:255-208. Sherman. K., W. Smith, W. Morse, M. Herman. J. (ireen. and L. Ejsymont 1984 Spawning strategies of fishes in relation to circulation, phytoplankton production, and pulses in zooplankton off the northeastern United States. Mar. Ecol. Prog. Ser. 18:1-19. Smith. R.A. 1971 The fishes of four low salinity tidal tributaries of the Delaware River estuary. M.S. thesis, Cornell Univ., Ithaca, 304 p. 1973 The distribution of summer fiounder, I'aralichlhys den- tatus. eggs and larvae on the continental shelf between Cape Cod and Cape Lookout, 1965-00. Fish. Bull, U.S. 71:527-.548. Smith, W.G., J.D. Sibunka, and A. Wells 1975 Seasonal distributions of larval flatfishes (Pleuronec- tiformes) on the continental shelf between Cape Cod, Massa- chusetts and Cape Lookout, North Carolina, 1965-0(). NOAA Tech. Rep. NMFS SSRF-691, 68 p. Smith. W.G.. D.G. McMillan. ('. Obenchain. I'. Rosenberg. A. Wells, and M. Silverman 1980 Spawning cycles of marine fishes of northeastern United Stiites based on broadscale surveys of eggs and larvae, 1977-79. Int. Counc. Exjjlor. Sea ICES CM 1980/L:66, 22 p. Wheatland, S.B. 1956 Oceanography of Long Island Sound, 1952-1954. \'II. Pelagic fish eggs and larvae. Bull. Bingh.am. Oce;inogr. Coll. 15:2.34-314. Wilk. S.J.. and M.J. Silverman 1976a Fish and hydm.graphic collections made liy the research vessels Dolphin and Delaware II during 1968-72 from New York to Florida. NOAA Tech. Rep. NMFS SSRF-097, 159 p. 1976b Summer benthic fish fauna of Sandy Hook Bay, New .lersey. NOAA Tech. Rep. NMFS SSRF-698, 10 p. Wilk. S.J. W.W. Morse. D.E. Ralph, and T.R. Azarovitz 1977 Fishes and associated environmental data collecteil in the New York Bight. June 1974-June 1975. NOAA Tech. Rep. NMFS SSRF-710. 53 p, Wilk, S.J.. W.G. Smith. D.E. Ralph, and J. Sibunka 1980 Population structure of summer flounder between New Ynrk and Florida based on linear discriminant analysis. Trans. Am. Fish. Soc. 109:265-271. Wootton, R.J. 1984 Strategies and tactics in fish reproduction. In Potts, G.W., and R.J. Wootton (eds.). Fish reproduction: Strategies and tactics, p. 1-12. Academic Press. London. Abstract. — Three new species of the genus Eptatretus-E. mccon- naugheyi, from off Southern Califor- nia and Northern Baja Cahfornia, and an isolated population in the southern portion of the Gulf of Cali- fornia; E.fritzi. known only from the immediate vicinity of Guadalupe Is- land. Mexico; andE. sinus, confined to the midriff area of the Gulf of California— are described. New data are presented for E. dearii and E. stoutii. A neotype is designated for E. stoutii. Despite the extensive ranges of £■. deani (southeastern Alaska to Gua- dalupe Island, Mexico) and E. stoutii (Vancouver Island, Canada, to Pt. San Pablo, about 80 miles southerly from Cedros Island, Mexico), no ap- preciable differences were found in counts and body proportions within each species. However, the disjunct populations of £". mcconnaugheyi dis- play significant differences (i-'>0.001) in numbers of trunk and total slime pores; in all other characters the two populations are very similar. Sex ratios for four of the species are notably uneven, ranging between (iO% and 74% female, 26% and 40% male. The sex ratio for E. stoutii is essentially even, 49% female and .51% male. Three New Species of Hagfishes, Genus Eptatretus (Cyclostomata, Myxinidae), from the Pacific Coast of North America, with New Data on E. deani and E. stoutii Robert L. Wisner Charmion B. McMillan Marine Bioiogy Research Division. A-002 Scnpps Institution of Oceanography, La Jolla, California 92093 Manuscript accepted 11 June 1990. Fishery Bulletin, U.S. 88:787-804. This study continues our efforts on the genus Eptatretus, a part of the unfinished work of Carl L. Hubbs (deceased 30 June 1979). Two previ- ous studies (McMillan and Wisner 1984, Wisner and McMillan 1988) and this have resulted in descriptions of seven new species of Eptatretus, two from off Central Chile, one each from the Hawaiian and Philippine areas, and three herein. Also, we have described a new species, Nema- myxine kreffti, from off Rio de Plata, Argentina (McMillan and Wisner 1984). In addition, a study on the gemi?,Myxine is progressing, involv- ing at least six new species from off the Atlantic and Pacific coasts of North and South America. The present effort includes the largest amount of study material, about 2300 specimens. Many speci- mens oi E. stoutii and E. deani are not included in the lists of material examined due either to poor preser- vation, to questionable capture data, or to a surfeit of material from an area, particularly off southern Cali- fornia. All specimens from near the extremes of ranges are included, as are all specimens of the three new species. Due to the availability for the first time of large numbers of specimens of £■. stoutii and E. deani. each hav- ing extensive distributions, we at- tempted to delineate an annual and spatial period of deposition (extru- sion) of eggs for each. Capture data for females containing eggs of at least 20 mm length showed that eggs of this size were present in both spe- cies throughout the year throughout all areas of their ranges. The max- imum lengths of eggs were 28.6 mm for E. stoutii and 52.3 for E. deani. Although these lengths may not in- dicate the ultimate lengths of eggs at time of deposition, it is very probable that depositi(jn by both species occurs throughout the year and in all areas of their distributions. Except for two collected in Novem- ber, all specimens of E. fritzi were taken in April; eggs to 33 mm were present. Similarly, specimens of E. sinus were taken within two short periods, 20-22 January and 28 Feb- ruary-2 March eggs to 32 mm were present. The occtu'rence and arrangement of head gi'ooves (lateral lines of authors) in all species treated here agree well with those described and figured by Ayers and Worthington (1907:332- 333, figs. 5, 6) and by McMillan and Wisner (1984:255, fig. 4). Methods and materials Methods of counting and measuring are those recommended by McMillan and Wisner (1984) and Wisner and McMillan (1988). Features used in 787 788 Fishery Bulletin 88(4), 1990 Figure 1 Outline t)f hagfish {Eptatretus) showing regions and features used in measuring and counting: A-H. total length (TL); A-B. prebran- chial length: B-C, branchial length, including gill apertures (GA); D. external opening of pharyngocutaneous duct (PCD): E, ventral finfold; C-F, trunk length; F. origin of cloaca; F-H, tail length; G, caudal finfold. The linear series of dots represents the prebranehial, branchial, trunk, and cloacal-caudal (tail) slime pores. counting and measuring are shown in Figure 1 . Ab- breviations used are given below and identified in Fig- ure 4. Figure 2 Branchial areas of holotypes of (A) Eptatretux mrmnnaugheyi. (B> E.Jrilzi. andCC)^. sinus; (D) neotype of £■. stout ii; (E)E. deani. The dental mus- cle (DM) is the cylindrical mass lying between the rows of gill pouches (GP). Abbreviations ABA Afferent branchial artery; one of the small blood vessels that lead to each gill pouch from VA or its branches (Fig. 4). DM Dental muscle; the firm elongate, cylindrical com- plex of muscles and cartilages that moves the dental plates and sets of cusps during feeding. Posterior portions of DM are shown in Figures 2-4, lying be- tween rows of gill pouches. GA Gill (branchial) aperture; external opening of the efferent duct leading from a gill pouch (Fig. 1). GP Gill pouch; rounded, serially arranged structures along and posterior to the dental muscle (Fig. 4). PCD External opening of the pharyngocutaneous duct; always confluent with the posteriormost left gill aperture in most species of Eptatretus, and much larger than all other apertures (Fig. 1). VA Ventral aorta; the portion between the heart (ventricle) and to where it branches to each side of DM (Fig. 4). *-4 . Jl i.« I Wisner and McMillan- Three new Eptstretus species from North American Pacific coast 789 Figure 3 Mature eggs of holotv^ie oi E. mcconnaugheyi showing )'« situ ar- rangement and connections by anchor filaments (terminal hooks). Otlier features used are: Preocular length From center of left eyespot to center of rostrum. Barbel length From crease at mesial base to tip of barbel. Collection data and deposition of specimens are listed for each species. All specimens were taken on bottom, with a trap or trawl, except for two of £■. stoutii taken by scuba divers near San Diego. Institutions which have provided study material, or in which type specimens are deposited, are SIO, CAS, LACM, MCZ, USNM, OSUO, BCPM, and UW. These symbolic institutional codes follow the usages in Leviton et al. 1985. We introduce length of barbel as a systematic char- acter. Barbels have not been used in the taxonomy of hagfishes except for the presence, absence, or degree of pigmentation between bases and tips. No tabular comparison of lengths has previously appeared in print, perhaps due to the small numbers of specimens avail- able. However, in this study a total of 2613 measure- ments of length of first nasal barbel, left side, provided material for comparison. Data in Table 6, first barbel length for 25-mm increments of total length, show sig- nificant differences in lengths within the five species treated. Key to species of Eptatretus of the Pacific coast of North America la Prebranchial length usually less than branchial length, very rarely equal to or even slightly greater. Tail length always less than branchial length E. mcconnaugheyi n. sp. lb Prebranchial and tail lengths always greater than branchial length 2 Figure 4 Sketch of branchial area, ventral view, of a hagfish (Eptatretus) delineating Areas I, II, and III. referred to in counts of gill pouches for each Area given in Table 8. (1) DM, posterior portion; (2) left branch of VA; (3) gill pouch; (4) point of branching of VA; (5) VA; (6) ventricle. 2a Prebranchial slime pores 7(4-10). Ventral finfold weakly developed, often vestigial or absent. Barbels small, not robust. Color dark purplish brown E. deani 2b Prebranchial slime pores 12-17. Ventral finfold variably well developed to absent. Barbels variably large to small. Color purplish to light or dark reddish brown 3 3a All barbels large, robust. Third barbel 49%(42-59%) of preocular length. Pre- branchial slime pores 12-13(10-15). Bran- chial apertures 11(10-12). Ventral finfold very weakly developed, usually absent. Color dark purplish brown E. fritzi n. sp. 3b All barbels small, not robust. Third barbel 34%(31-37%) of preocular length. Prebranchial slime pores 13-17. Ventral finfold either prominent, vestigal or ab- sent. Branchial apertures 10-14. Color dark brown to light reddish brown 4 790 Fishery Bulletin 88(4), 1990 4a Branchial apertures 12(10-14). Ventral finfold prominent, with wide pale margin. Color light reddish gray or light brown. Extensive piebaidness common E. stoutii 4b Branchial apertures 10(9-12). Ventral fin- fold vestigial or absent. Color dark red- dish brown, not piebald E. sinus n. sp. Systematlcs Eptatretus mcconnaugheyi new species It is not feasible to attempt a meaningful synonymy for this new species, as it and E. stoutii are very similar and occur sympatrically off southern California. Ep- tatretus stoutii has long been extensively used in bio- chemical and physiological research. However, none of this research material has been examined by us, and no identification credited to competent ichthyologists. As£'. mcconnaugheyi is not previously known, it is very likely to have formed a part of this research material, particularly from southern California. Holotype SI069-231E, female, 482 mm TL, taken at 32°32'0"N, 117°21'07"W, in a trap on bottom at 148 m, 11-12 April 1969. Paratypes SI069-228B, 14(185-400 mm TL), taken at32°05.9'N, 117°04.4'W, 177 m, 10-11 March 1969; SI071-114, 7(360-440 mm TL), taken at 28°21.0'N, 115°43.0'W, 689 m, 25-26 May 1971; SI069-231E, 7(200-448 mm TL), taken with the holotype; SI068- 126, 5(355-409 mm TL), taken at 25°49'N, 110°51'W, 351 m, 25-26 January 1968; CAS 63203, 7(210-400 mm TL), taken at 32°31.8'N, 117°21.6'W, 145 m, 29 September 1972; LACM 44409-1, 6(195-430 mm TL), taken at 32°31.8'N, 117°21.1'W, 145 m, 29 Septem- ber 1972; USNM 29630, 7(245-380 mm TL), taken at 32°31.8'N, 117°21.1'W, 145 m, 29 September 1972. Additional material SI056-9, 1(340), 42 m; SI068- 109, 1(390), 415 m; SI068-11, 1(230), 177 m; SI068- 124, 1(249), 205 m; SI068-125. 19(239-347), 278 m; SI068-126, 5(355-409), 351 m; SI068-127, 3(315-440), 238 m; SI069-179, 2(235-376), 109 m; SI069-181, 2 (141-312), 183 m; SI069-225C, 1(392), 371 m; SI069- 228B, 14(195-406), 99 m; SI069-228C, 1(433), 166 m; SI069-231C. 1(372), 139 m. Distribution Eptutrefiis inrconnaugheyi appears to consist of two disjunct populations: one from Santa Monica Bay, California, to the Cedros and San Benito Islands, Mexico, and one apparently restricted to the lower portion of the Gulf of California. These two pop- ulations appear to differ significantly (P>0.001) in numbers of trunk and total slime pores, the higher counts occurring in the southern California population. Depths of capture range between 42 and 384 m off southern California and between 177 and 415 m in the Gulf of California. Collecting efforts between the Cedros and San Benito Islands and the mouth of the Gulf have failed to take the species. Diagnosis Prebranchial length usually less than bran- chial length, rarely equal to or very rarely even slight- ly longer. Body motlerately robust, deepest at midhody, its width about half its depth, increasingly laterally compressed toward tail. Tail length 13-16% of TL, its depth about half its length. Ventral finfold usually prominent, often low, with pale margin. Caudal finfold prominent, wide, the margin thin and pale. Three fused cusps (multicusp) on anterior, two fused cusps on pos- terior, set of cusps. Etymology We take great pleasure in dedicating this new species to Ronald R. McConnaughey, friend and superb marine technician, who was instrumental in the development of capture gear and party to the collect- ing of many of the thousands of specimens of species examined by us. Description Counts (Tables 2-8) and body propor- tions (Table 1) are given and compared with similai' data for the other four species treated here. Body moderately robust, deepest at midbody, the width about half the depth, increasingly laterally compressed toward tail. Tail spatulate, its length about twice its depth, its ventral margin slojMng downward below le\'el of body. A tliin, deep caudal finfold, with a narrow pale margin, extends from cloaca around end to about over cloaca dorsally. Ventral finfold low, usually entirely un- pigmented. Head at eyespots about as deep as wide, narrowing to rostrum. Width of rostrum about half the width of head at eyespots. Eyespots rather prominent, of moderate size, the margins well defined. Oral sur- face short, sloping posteriorly at about 45° angle from the vertical. Barbels small, the distal thirds unpig- mented. First barbel only slightly shorter than second, often equal to or slightly longer, averaging 64% (52- 79%) of length of third, the second barbel averaging 70%(51-84%) of length of third. Length of first nasal barbel, left side, is given in Table 6 and compared with that of other species treated. Head grooves Grooves present before and behind eyespots on each side of dorsal midline. Three to six grooves, each side, before eyespots lie in longitudinal rows, and two to four ventrad in transverse rows. One to eight grooves may occur behind eyespots. Most lie transversely, with none to three longitudinally, at sides. As many as 24 grooves may occur on one specimen. Wisner and McMillan: Three new Eptatretus species from North American Pacific coast 791 Color a dark, reddish brown in life without pale spots or areas. Barbels pale on distal thirds. Length of DM 27%(21-39%) of TL. DM width 68% (57-90%) of its depth. VA variable in length, averag- ing 11% (5-22%) of DM length. DM to VA 14% (5-26%) of DM length, and 1.4% (0.5-9%) of VA length. Number of GP in positions relative to DM and VA, Areas I, II, III, are given in Table 7, defined in Figure 4, and com- pared with similar data for the other four species discussed herein. Afferent duct of last GP, left side, always confluent with PCD. Higher numbers of GA, left vs. right sides, may differ between the two populations (Table 8). Off southern California, the higher count is about equal for both sides at 4 and 5 each. In the Gulf population, the higher count is always on the left side. Eggs The holotype contains 18 mature eggs, the largest 25.2 x 9.0 mm. The largest egg found in any female is 26 x 8 mm. The smallest female with eggs, most in round or slightly ovoid stages, is 267 mm TL. It may be that females of the Gulf population mature at a smaller size than do those from southern Califor- nia. One, 267 mm TL, has eggs to 15.8 mm; one, 352 mm TL, has eggs to 20.8 mm; and one, 32 mm TL, has eggs to 17.6 mm. These figures contrast sharply with those of the southern California population, in which eggs ranging between 15 and 29 mm occur only in females of 400-508 mm TL. Perhaps these differences are an artifact of sampling or of developmental stages relative to both time and size, or to the relatively small number of specimens available (97). All mature eggs of the holotype have fully developed anchor filaments (terminal hooks) with most free of en- capsulating membrane (Fig. 3), and apparently ready for extrusion. This condition was rarely observed among the nearly 2400 specimens of all species of Ep- tatretus examined by us in the course of our studies. The eggs are linked by filaments in rows in a fashion similar to that shown by Dean (1989) and Jensen (1966). The sex ratio is unbalanced. Of a total of 58 speci- mens for which sex could be reliably determined, 64% were female and 36% male. Differentiation All body proportions and most counts for the two disjunct populations are very similar. However, significant differences (P>0.001) occur in numbers of trunk and total slime pores (Tables 3, 4), despite considerable overlap in values of thousandths of total length. Discussion This species and E. stoutii appear to be closely related and occur sympatrically off southern California. However, the two are readily separable in that the preliranchial length of E. inccoiinaugheyi is less, rarely even slightly longer, than the branchial length, and always greater in E. stoutii. Also, the num- bers of prebranchial slime pores differ notably, 8-9 (6-11) in E. mcconnaugheyi and 13(10-16) in E. stoutii (Table 2). Eptatretus fritzi new species Holotype SI066-26, male, 550 mm TL, taken at 28°51'N, 118°14'W, in a trap on bottom at 512 m, 3-4 April 1966. Paratypes SI066-26, 30(280-585 mm TL), taken with the holotype; CAS 63201, 15(325-540 mm TL), taken with holotype; LACM 44407-1, 15(375-495 mm TL), taken with the holotype; USNM 296318, 15(350- 570 mm TL), taken with the holotype; SI063-177, 20 (324-522 mm TL), taken at 28°52'N. 118°14'W, 2743 m, 24 April 1963; SI066-22, 36(232-521 mm TL), taken at 29°06'N, 118°17'W, 402 m, 12 April 1966; SI066-23, 195(286-556 mm TL), taken at 28°54'N, 118°13'W, 444 m, 2-3 April 1966; SI066-36, 2(354-405 mm TL), taken at 29°30'N, 117°17'W, 512 m, 6-7 April 1966; SIO67-60, 120(281-498 mm TL), taken at 29°09'N, 118°16'W, 832 m, 26-27 April 1966; SI068-664, 2 (425-535 mm TL), 183 m, 15-16 November 1968; SI072-294, 4(207-541 mm TL), taken at 29°10'N, 118°16'W, 256 m, 12-13 April 1970. Distribution Known only from the immediate vicinity of Guadalupe Island, Mexico. Depths of capture range between 182 and 2743 meters. Diagnosis All barbels notably larger and more robust than on any other species ot Eptatretus known to us (Table 6). Ventral f infold absent or vestigial, without pale margin. Prebranchial and tail lengths each gi'eater than branchial length. Color a dark purplish-brown. Prebranchial slime pores 12(10-15). Three fused cusps (multicusp) on anterior set of cusps, two fused cusps on posterior set. Etymology We take great pleasure in dedicating this species to Frithjof (Fritz) Ohre, friend, willing, eager, and industrious volunteer on many of the expeditions on which all species treated here were taken, particu- larly those to Guadalupe Island. Description Counts (Tables 2-8) and body propor- tions (Table 1) are given and compared with similar data for the other four species treated here. Body robust, deeper than wide, deepest at midbody, increas- ingly laterally compressed toward tail. Tail spatulate, its depth about half its length, its ventral outline not sloping downward from cloaca. Caudal finfold variably 792 Fishery Bulletin 88(4), 1990 thin to thick, beginning at cloaca and extending around tail to about over origin of cloaca, without pale margin. Ventral finfold usually absent, present in only 28 of 341 specimens (8.2%). When present it is intermittently and weakly developed along its length. It is thick, heavy, and without pale margin. Head at eyespots about as deep as wide, narrowing to rostiiim. Eyespots prominent, the margins irregular. Oral surface sloping posteriorly at about a 35° angle from the vertical. All barbels are notably longer and thicker in proportion to total length than on any other myxinid known to us (Table 6). First and second barbels about equal in length, the third averaging 49% (42- 50%) of preocular length. First barbel as long or longer than width of nasal orifice when manually flattened (124%(100-203%)). First barbels average less than 50% of this width in all other species of Eptatretus. Head grooves Grooves always present behind eye- spots but occasionally absent before. Those behind average 5(3-10) in number and are arranged mostly in transverse rows with an occasional few grooves lying ventrad and longitudinally. Those before eye- spots average 3(1-6) in number and lie longitudinally. Grooves behind eyespots originate very near the dor- sal midline but do not cross. Those before eyespots are notably more distant from the midline. Color Color a dark puiple to black in most specimens, a chocolate brown in others, rarely showing pale spots. Notes by the senior author, party to the first and most subsequent captures, state, "a dark color." The choco- late color may be an artifact of preservation. GA, slime pore margins, and ventral finfold are without pale margins. DM short, moderately robust, its length 27%(21- 34%) of TL. DM width i3%(ll-15%) of its length, its depth 73%(58-88%) of its width. VA length variable, 16% (8-42%) of DM length. DM to VA also variable, 15% (4-32%) of DM length and 90% (24-206%) of VA length. Numbers of GP in positions relative to DM and VA, Areas I, II, III, are given in Table 7, defined in Figure 4, and compared with similar data for the other four species treated here. Afferent duct of last GP, left side, always confluent with PCD. Variation occurs in numbers of GA between left and right sides, the higher number always on the left side, 81 vs. 0 instances (Table 8). Eggs The largest egg, 33.8 x 8.5 mm, occurred among 14 large eggs in a 386-mm TL female. No more than 16 and as few as 10 almost fully developed eggs were found in any female. No egg had free anchor fila- ments, but occasionally they were visible through the encapsulating membranes. The sex ratio is unbalanced. Of a total of 358 speci- mens for which sex could be reliably determined, 60% are female and 40% male. Discussion This species and E. deani occur sympa- trically near Guadalupe Island, Mexico. Both were taken in the same trap on two occasions. They are similar in coloration, being very dark brown to pur-plish- black, but E.fritzi seems much less prone to piebald- ness than E. deani as very few pale spots were noted. Primary differences are that the barbels of E. J'ritzi are much longer and more robust than those of £■. de.ani (Table 6), and that it has more prebranchial pores, 12 (10-15) vs. 7(4-10). Also, E.fritzi usually has no ven- tral finfold (313 of 341 specimens, 91.8%), whereas in E. deani this finfold is usually present (863 of 892 specimens, 96.6'7o), although rather weakly developed. Eptatretus sinus new species Holotype SIO68-108, female 307 mm TL. taken at 25°49N, 110°44'W, in a trap on bottom at 70S m. 22-23 January 1968. Paratypes SIO68-108, 31(230-425 mm TL), taken with the holotype; SI068-94, 29(129-346 mm TL), taken at 29°20'N, 113°10'W, 263 and 283 m, 20 Jan- uary 1968; SIO68-100, 60(267-464 mm TL). taken at 29°00N, 113°25'W, 467 m, 20-21 January 1968; CAS 63202, 15(270-400 mm TL), taken at 25°49'N, 110° 44'W, 708 m, 22-23 January 1968; LACM 44408-1, 14 (270-400 mm TL), taken at 25°49'N, 110°44'W, 708 m, 22-23 January 1968; USNM 296319, 15(275-420 mm TL), taken at 25°49'N, 110°44'W, 708 m, 22-23 January 1968. Additional material SI068-97, 7(223-308), 881 m; SI068-98, 17(276-338), 881 m; SI068-99, 11(263-420), 668 m; SIO68-101, 16(264-355), 198 m; SIO69-201, 1 (190), 1454 m; SIO69-203, 11(381-630), 759 m; SI069- 206, 54(363-430), 768 m; SIO69-207, 57(259-404), 475 m. Distribution Known only from the midriff area of the Gulf of California, Mexico, between about 28° and 30°N latitudes. Depths of capture range between 198 and 1330 meters. Diagnosis Prel)ranchia! length greater than branchial length. Tail length usually greater than branchial length, occasionally equal to or less. Ventral finfold low, unpigmented, occasionally absent. Barbels usually un- pigmented, except rarely at bases. Three fused cusps (multicusp) on anterior, two fused cusps on posterior, sets of cusps. Wisner and McMillan: Three new Eptatretus species from North American Pacific coast 793 Etymology From the Latin "sinus," meaning gulf, by reason of its apparent restriction to tlie midriff area of the Gulf of California, Mexico. Description Counts (Tables 2-8) and body propor- tions (Table 1) are given and compared with similar data for the other four species treated here. Body robust, deepest at midbody, deeper than wide, increasingly laterally compressed toward tail. Tail spatulate, its depth about half its length, its ventral outline not sloping downward from cloaca. Caudal fin- fold thickened ventrally. thinner around tail, ending dorsaily over cloacal origin. Ventral f infold usually present, low, unpigmented. It is absent on 10% of 363 specimens. Head at eyespots as deep as wide, the oral surface sloping posteriorly at about 35° from the vertical. Eye- spots prominent, the margins well defined. Barbels short, unpigmented except rarely at bases. Second barbel usually longer than first but occasionally shorter or of equal length. First and second barbels respective- ly 60% and 64% of third barbel. Length of first nasal barbel is given in Table 6 and compared with that of other species treated. Head grooves Grooves present before and behind eyespots, eacli side. One to five grooves occur behind eyespots, most lying transversely, but a few lying longitudinally at side of head. One to four grooves lie before eyespots in longitudinal rows. Occasionally no grooves occur before eyespots. Color Color reddish-brown, without pale markings or spotting. GA, PCD, and an occasional slime pore have narrow pale margins. Length of DM 24%(21-29%) of TL. DM width 14% (12-17%)) of its length, its depth 65%(54-86%) of its width. DM to VA 16%(15-26%) of DM length. VA length 25%i( 17-39%.) of DM length. VA length variable, averaging 66%(22-127%) of distance from DM to VA. Numbers of GP in position relative to DM and VA, Areas L IL IIL are given in Table 7, defined in Figure 4, and compared with similar data for the other four species treated here. Afferent duct of last GA, left side, always confluent with PCD. Variation in numbers of GA occurs between left and right sides, with higher numbers predominating on the left, 65(15%) vs. 8(2%,), A^ 444 (Table 8). Eggs The holotype, 370 mm TL, contains many small eggs, to about 4 mm. The largest egg found (among 21 large ones) was 32 x 7.5 mm in a female of 371 mm TL. The most eggs found was 37 (18 x 5 mm) in a female of 420 mm TL. Eggs (28) to 21 mm were found in a female of 372 mm TL, and to 23 mm in one of 335 mm TL. Apparently, E. sinus matures at a short total length, as a male of 130 mm and a female of 142 mm were noted. These are surprisingly small sizes, for in other species, these total lengths are distinctly juvenile, and the sex determinable only under very high mag- nification, if at all. Encapsulated anchor filaments are visible on eggs of 20 mm length. The sex ratio of £". sinus is unbalanced. Of a total 424 specimens for which sex was reliably determined, 65% are female, 35% male. Eptatretus deani (Evermann and Goldsborough 1907) Synonymy Polistotrema deani Evermann and Goldsborough 1907: 221-222, 225-226, fig. 1 (original description; com- pared with Polistotrema stoutii; Albatross Station 4235, Spacious Bay, Cleveland Peninsula, 130-193 fathoms, and Albatross Station 4238, Nose Point, Behm Canal, southeastern Alaska, 229-231 fathoms); Wilimovsky 1954a: 3, and 1954b: 281 (SE Alaska to California), 1958: 18 (compared with P. stoutii; SE Alaska to California); Barham et al. 1967: 780, fig. 7 (shown associated with Anoplopoma fimbria. Trieste Dive 105; 1060 m in San Diego Trough) [presumptive]. Polistotrema curtiss-jam.esi Townsend and Nichols, 1925: 4-5, 18, 20, fig. 1 (original description: com- pared with P. stoutii; south of Monterey to Santa Barbara L: 440-585 fathoms; Albatross Stations 5695, 5696, 5697, 5698; type locality Station 5697, 585 fathoms). Heptatretus deani Regan, 1912: 534-535 (comparisons; synonymy; references; diagnosis; Alaska). Bdellostomu deani. Adam and Strahan 1963; 6 (11-12 gills; length to 620 mm; Alaska). Eptatretus deani. Day and Pearcy 1968: 2668, 2771 (listed; about 1 200 meters off Oregon); Bourne and McAllister 1969: 3246-3248, fig. 1 (first British Co- lumbia records, 54°47'N, 130°58'30"W at 90-250 fathoms: 48°53'N, 126°40'W at 373 m: description); Kukowski 1972: 7, 24 (Monterey Bay; references); Miller and Lea 1972: 32 (Cedros and Guadalupe Is., Baja California to SE Alaska: length to 20 inches; depth 1560-3500 feet; common in deep water, uni- form purplish black); Fitch 1973; 815 (off Eureka. California, 565-585 m); Hart 1973: 10, 16-18, fig. 1 (first description for Canada; comparisons in key; description; geographical and bathymetric distribu- tions; references); Gates and Frey 1974: 60 (ver- nacular, California): Smith and Hessler 1974; 72-73 (14 miles off San Diego; depth 1230 m; respiration rate measured in situ; a dominant species in San Diego Trough); Jespersen 1975: 189-190, 195, 197, 794 Fishery Bulletin 88(4), 1990 fig. 15 (spermiogenesis; southern California off San Clemente I. at about 1280 m; off Los Coronados I. at about 1097 m; Theisen 1976: 167-173, figs. 2. 4. 7. 10 (niicrostructure of olfactory system; southern California; 1097 m. Holotype USNM 57820, sex not determined, "21" inches (about 535 mm TL), taken near Yes Bay, Behm Canal, Cleveland Peninsula, southeastern Alaska, east- erly of Ketchikan, about 55°55'N, 131°55'W, by dredge at 419-423 m (Albatross Station 4238), 8 July 1903. Paratypes USNM 61162, two: one a female, in rot- ting condition, taken at Albatross Station 4229 and 4235, Behm Canal, southeastern Alaska, Nose Point, near the holotype, by dredge at 419-423 m. 16 July 1903. (The type specimens were examined by Carl L. Hubbs at USNM, 25 February 1972.) Additional material SI059-47, 6(359-458), 823 m; SIO60-540, 1(484), 1920 m; SI063-177, 20(286-457), 2743 m; SI063-868, 21(345-498), 107 m; SI064-281, 1(451), 587 m; SI064-496, 5(374-455), 475 m; SI064- 497, 26(335-445), 512 m; SI064-498, 37(291-441), 616 m; SI064-499, 38(294-473), 960 m; SI065-499, 54 (321-521), 933 m; SI065-465, 19(370-554), 1302 m; 81066-27, 2(148-455), 1252 m; SI066-36, 130(233- 509), 512 m: SI066-535, 21(377-453), 1013 m; SI066- 537, 1(417), 1244 m; SI066-551, 2(390-401), 1156 m; SI067-64, 20(380-510), 980 m; SIO67-109, 52(312- 523). 1024 m; SIO67-110, 1(369), 1445 m; SI067-118, 51(288-467), 732 m; SI068-427, 5(396-470). 1840 m; SI068-428, 6(320-481), 800 m; SI069-225F, 38(342- 429), 633 m; SI069-227B, 26(288-495), 744 m; SI069- 288D, 56(182-423), 609 m; SIO71-103. 4(401-484), 1373 m; SI071-7, 42(280-484), 1244 m. The following collections, with incomplete data were taken between British Columbia and northern Califor- nia: OSUO 3869, 3(267-450), OSUO 3870, 2(385-415), OSUO 3871, 2(404-447), BCPM 46-15, 1(366), BCPM 72-9, 4(402-475), BCPM 72-10, 1(460), CAS 10877, 1(390), CAS 26633, 2(318-365), CAS 20404, 1(395), Albatross Station 3126, 1(317), 378 m; UW 00379, 1(440), 155 m; UW 18148, 1(391), 595 m; UW 18149, 1(383), 1098 m; UW 18151, 1(405), 1190 m; UW 18152, 7(370-445), 1190 m; UW 18153, 2(402-466), 932 m; UW 18154, 1(354), 183 m; UW 18155, 6(376-433), 1190 m; UW 18157, 5(400-440), 732 m; UW 18158. 12(342-459), 824 m; UW 18159, 2(343-375), 732 m; UW 19095, 3(421-440), UW 19127. 2(414-444), UW 19206, 1(320), UW 19310, 6(382-479). Distribution Southeastern Alaska to near Guadalupe Island, Mexico. Despite this extensive range of about 2600 miles, no significant variation in counts or mea- surements with latitude was demonstrated. Diagnosis Prebranchial slime pores 7(4-10). Pre- branchial length usually greater than branchial length, rarely equal to or less. Tail length variably equal to, or slightly less or greater than branchial length. Color a purplish-black, with occasional pale spots and areas. Anterior part of head often pale. Three fused cusps (multicusp) on anterior set of cusps, two fused cusps on posterior set. Description Counts (Tables 2-8) and body propor- tions (Table 1) are given and compared with similar data for the other four species treated here. Body robust, deeper than wide at midbody, progres- sively laterally compressed toward tail. Prebranchial slime pores 7(4-10). Branchial apertures 11(10-12). Prebranchial length usually greater than branchial length, rarely equal to or less. Tail length greater than branchial length in 45%, equal-to in 30%, and less-than in 26% of 441 specimens. Ventral finfold weakly developed, occasionally absent, without pale margin. Tail spatulate, its ventral surface sloping downward at a slight angle from the ventral body line. Head wider than deep. First barbels usually shorter than second, occasionally equal to or slightly longer; the first about 58% of length of third barbel, the third about 70% of preocular length. Eyespots prominent, variable in size, usually with irregoilar margins. Length of first nasal barbel is given in Table 6 and compared with that of other species treated. Head grooves Of 104 specimens selected for mucus- free condition, 18 (17%i) had no grooves. Of those hav- ing grooves before eyespots, 45% had them on both sides, about 6% had them on either the right or left side only. Grooves occur on both sides beliind eyespots in 80'7o, on the left only in 8%i, and on the right side only in 12% of the 104 specimens. Color Color usually a purplish-black or very dark brown. Specimens with a brownish base color show variable amounts of pale spots and patches (piebald- ness) that range from slight to considerable. An occa- sional specimen has a pale head from rostrum to first few prebranchial slime pores. Most heads are pale near tip of rostrum. On pale heads, the barbels usually are unpigmented. Ventral and caudal finfolds without pale margin except when spotted in piebald specimens. Gill apertures and slime pores with narrow pale margins. Cloacal margins with rather wide pale areas. Variation occurs in numbers of GA l)etween left and right sides with higher numbers predominating on the left, 125 (11%) vs. 21 (2%) on the right side in 1145 counts (Table 8). Wisner and McMillan Three new Eptatretus species from North American Pacific coast 795 DM length 24% (20-26%) of TL. Width of DM 15% (13-18%) of its length, its depth 64% (54-80%) of its width. DM to VA 9% (0-15%) of DM length. VA length 31% (23-36%) of DM length. Numbers of GP in posi- tions relative to DM and VA, Areas I, II, III, are given in Table 7, defined in Figure 4, and compared with similar data for the other four species treated herein. Afferent duct of last GA, left side, always confluent with PCD. Eggs Mature eggs large, usually distinctly curved. The largest egg found measured 52.3 x 10.5 mm and was among 14 large eggs in a 500-mm TL female. Ap- parently very few eggs mature at one time from the many hundreds of tiny, round to slightly ovoid eggs normally present. No more than 15 nor fewer than 8 large eggs (30 mm and longer) were present in any female. Anchor filaments were visible on all eggs of 35 mm length or more, but were still encapsulated. The sex ratio oiE. deani is notably unbalanced. Of a total of 480 specimens, 74% were female, 26% male. The smallest female noted was 249 mm TL, the smallest male 255 mm TL. Eptatretus stoutii (Lockington 1 878) Synonymy The following synonymy includes only those references arbitrarily considered by us to have taxonomic, behavioral, or distributional value and not mere usage of the name. About 120 references using the name stouti or stoutii pertain only to biochemical and/or physiological studies that serve no apparent taxonomic purpose and are not included. Bdellostoma stoutii Lockington, 1878:792-793 (original description; compared with B. polytrema; mouth of Eel River, Humboldt County, California); Jordan and Gilbert 1883:6 (brief description, after Lockington; coast of California; not rare). Bdellostoma stouti, Dean 1899:223-276 (embryonic development of larvae and eggs). BdeUostom.a dombeyi, Jordan and Gilbert 1883:57 (in part; characters and feeding habits; California); Worthington 1905:625-663 (description, behavior in aquaria; habitat in Monterey Bay; counts of gill open- ings; feeding; eggs); Ayers and Worthington 1907: 327-336 (figures and description, lateral lines found only on head). Honiea stoutii, Dean 1903:295-298 (partial albinism in anterior portion), 1904: 6. 7. 16, 18, 20 (seasonal spawning; largest are females; California). Heptatretus stouti, Regan 1912:534-535 (synonymy; comparisons; diagnosis; California). Eptatretus stouti, Starks and Morris 1907:161 (com- mon in deep water off San Diego; north to Cape Flat- tery; abundant in Monterey Bay); Strahan 1963:22 (behavior and attitudes in aquaria); Linthicum 1971: 17-22 (immunological study; tying itself in knots to remove slime; slime production, teeth); Kukowski 1972:24, 37, 43 (Monterey Bay and Elkhorn Slough; references), 1973:7, 19, 22-23 (Monterey Bay; occur- rence data; recurrent groupings); Anonymous 1973: 68 (observed at camera station at 306 m, SW of White's Point, southern California); Allen 1976:95 (only species not taken in otter trawl or by setline in Santa Monica Bay, southern California; has been so collected elsewhere); Downing et al. 1981:326-328 (description and figures of mucus thread cells in slime glands); Fernholm 1985:113-122 (in part; no ciliated cells or innervation or electroreceptory capacity found in lateral lines). Polistotrema stouti, Clemens and Wilby 1946:9 (char- acters; white margin on ventral finfold; reaches 25 inches; species attacked; attachment of eggs; W coast, Vancouver I.). Polistotrema stoutii, Fowler 1908:461 (Pacific Grove on Monterey Bay, California; gill opening counts for 8 specimens); Clemens and Wilby 1961:18, 49-50 (description; history on British Columbia west coast; eggs; southern California to SE Alaska; in part); Mclnerney and Evans 1970:966-968 (habitat char- acteristics; Mayne Bay in Barkley Sound, Vancouver I.); Sasaki 1972:283 (Queen Charlotte Sound, Van- couver I.; rare). Dodecatrema stoutii, Fowler 1947:3 (proposed new name to replace Bdellostoma). Eptatretus stoutii, Jensen 1959:798 (albino and piebald variants; ratio to normal; habitat; abundance in southern California; submarine canyons tributary to San Diego Trough; 210 fathoms), 1966:82-90 (habits; description and figures of the four hearts; functions); Taylor 1967:181-187 (chromosome numbers: prob- ably supernumeraries; Alaska to Baja California; habitat; 60-1800 feet); Day and Pearcy 1968:2668 (depth distribution off Oregon); Bourne and McAlis- ter 1969:3248 (compared with E. deani); Miller and Lea 1972:32 (Pt. San Pablo. Baja CaHfornia, to SE Alaska: length to 25 inches; depth range 30-2400 feet; description); Hart 1973:16, 18-19, 53 (compari- sons in key; description; common; geographical and bathymetric distribution in and off Barkley Sound; probably quite generally distributed in British Co- lumbia; Alaska records not confirmed in recent re- views; references; egg cases); Smith and Hessler 1974:72-73 {in situ respiration rate significantly greater than that oi E. deani); Anonymous 1974:50 (observed only during day at baited camera off Palos Verdes, California); Fernholm 1975:201-203 (struc- ture and deposition of eggs compared with that of E. btirgeri; Bdellostoma dombeyi of Worthington referrable to E. stoutii); Jespersen 1975:189-198 796 Fishery Bulletin 88(4), 1990 (spermiogenesis; offshore in 45.7 and 53.0 m); Knaggs et al. 1975:56-57 (taken off central Baja California; southern range extended from Cedros I. to Pt. San Pablo); McMillan and Wisner 1984:255 (lateral line described and figured); Wisner and McMillan 1988:231 (the term "head grooves" sug- gested to replace that for lateral lines. Weotype SI068-426, male, 530 mm TL, taken about 2 miles southwest of the whistler buoy at entrance to Humboldt Bay, California, in a trap on bottom at 38-44 m, on 27 August 1963. Additional material, northernmost collections SIO 68-426, 34(291-527), taken with the neotype. BCPM 60-1, 2(474-580), taken at about 49°33'N, 126°38'W, outside Nootka Bay, Vancouver I.; no other data; BCPM50-3, 2(192-420), taken off west coast of Van- couver I.; no other data; BCPM98-2, 1(506). taken off west coast of Vancouver I.; no other data; UW-15869, 4(144-400), taken at 48°51.8'N, 124°35.8'W, in Strait of Juan de Fuca, in a trawl; depth not recorded; 27 March 1956; UW-18161, 6(437-512), taken "8 mi out" from Depoe Bay, Oregon at 79 m; method of capture not recorded, 30 August 1964; CAS 19191, 1(403), taken at Holmes Harbor, Puget Sound; no other data; OSU-OTB-180, 1(435), taken at 44°43.5'N, 124° 18.1' W, in a bottom trawl at 96 m, 24 July 1967. Twelve collections in the California Academy of Sci- ences, comprising one or two specimens each, and with very incomplete or absent capture data, were taken between Cape Mendocino and San Diego, California (CAS 6832, 7353, 11115, 12751, 12881, 20290, 26491, "Ace. 1954:XI:23," lU 1006, 1069, 1086, 1613 (the lU collections are now at CAS). Similarly, three collections (MCZ 28802, 28803, and 32782), totaling five speci- mens, with very limited data, were taken in Monterey Bay, California. Material was taken in 1963 by David Jensen, then of Scripps, in collaboration with the University of Washington Atomic Energy Program. Most specimens were sacrificed to that program. A total of si.x collec- tions were made between Cape Disappointment, at mouth of Columbia River, off Depoe Bay, Oregon, off Humboldt Bay, and Fort Bragg, California. Numbers of specimens ranged between 6 and 152, and were taken at depths of 38-132 m in bottom traps. More than 40 collections, comprising nearly 1500 specimens, taken in the southern California area (Pt. Conception to Islas Todos Santos, Baja California, Mex- ico) were used for certain counts but are not listed among the study material. In addition, hundreds of specimens in the Scripps Marine Vertebrate Collection taken off San Diego were examined for certain char- acters only and are not listed. However, all collections from outside the Southern California area are listed. Southernmost collections SI059-92, 1 (345), 28°23' N, 115°21'W, 280 m; SI062-91, 3(164-207), 28°11'N, 115°23W; SI064-951, 1(363), 31°08'N, 116°35'W, 137 m; SI071-114, 30(300-465), 28°21'N, 115°43'W, 384 m; SI071-121, 7(283-384), 28°18'N, 115°29'W, 311- 330 m; SI071-126, 83(195-460), 30°22'N, 116°07'W, 201 m; SI073-373, 2(320-395), 28°50'N, 114°48'W, 92-95 m. Distribution Nootka Bay, west side of Vancouver Is- land, British Columbia, Canada, to Pt. San Pablo, Baja California (about 27°14'N, 114°30'W). Although E. sfoutii has been reported from southeastern Alaska, we have found no valid records of capture north of Nootka Bay (about 49°33'N, 126°38'W). Despite this extensive range of about 2200 miles, no significant variation in counts or measurements with latitude was demonstrated. Carl L. Hubbs, at USNM, 28 June 1972 (personal notes), examined collections from southeastern Alaska, labeled as E. stoNtii. and found them to be either mis- identified oi' with the capture data indicating localities far from Alaska, as follows: USNM 53963, listed as from southeastern Alaska, but with coordinates indi- cating the vicinity of Monterey, California; USNM 53964, listed as from Alaska, but with coordinates given as 33°08'N, 118°40'W, the vicinity of Santa Bar- bara, California. Hubbs recorded this specimen as E. dcdni "without (juestion;" USNM 73737, Albatross Station 3077, east of Sitka, Alaska, labeled as E. Htautii but referrable to E. deani. based on the skin being jet lilack and eggs to 38 mm length— a length much greater than found in E. stout ii (28 mm); in addition, a collec- tion. UW 02738, four, 332-390 mm TL, was recorded simply as "S.E. Alaska; International P^ish Commis- sion, 1931." Depths of capture range between 16 m off the San (_)nofre Nuclear Power Plant, near San Clemente, southern California, and 633 m at 31 °47'N, 116°50'W, south of Ensenada, Baja California, Mexico. Latitude may be of limited significance in depths of capture (habitat), as the farthest-north least depth is 44 m, off Humboldt Bay, and the farthest south the 16 m capture off the San Onofre Nuclear Power Plant. The species has been taken by scuba divers between 18 and 21 m near La Jolla. California. Diagnosis Body slender, only slightly deeper than wide. Prebranchial length greater than branchial length. Branchial apertures 12(10-14). Ventral finfold prominent, the distal margin pale. Eyespots small, prominent, the margins well defined. Color a light brown with pale spotting and small blotches common. Wisner and McMillan Three new Eptatretus species from North American Pacific coast 797 Three fused cusps (multicusp) on upper and two fused cusps on lower sets of cusps. Description Counts (Tables 2-8) and body propor- tions (Table 1) are given and compared with similar data for the other foui' species treated herein. Body slender, only slightly deeper than wide, pro- gressively laterally flattened toward tail. Tail spatu- late, its ventral profile not slanting downward, but lying on a straight line with ventral profile of body, its depth 43% (30-56%) of its length. A thin, well devel- oped, pale-margined finfold extends from cloaca around tip of tail, ending about over cloaca. Tail length usual- ly greater than branchial length, but is equal-to in 12%i and less-than in 8% of 421 specimens. Ventral finfold prominent, with pale distal margin. Head at eye spots about as deep as wide, narrowing toward rostrum. Eye- spots small, with distinct margins. Barbels small, usual- ly without pale tips. Second barbel usually longer than first, the first shorter than second in 41%, about equal to in 51%, and longer than second in 7% of 315 speci- mens. First barbel 51% (41-84%), the second 64% (49-98%) of third barbel. Length of first nasal barbel is given in Table 6 and compared with that of other species treated. Variation occurs in numbers of GA, left and right sides, with higher numbers predominating on the left, 167 (11%) vs. 56 (4%), N-1554 (Table 8). Head grooves Grooves present above and below eye- spots. Two to four (rarely five) grooves lie above eye- spots in lines nearly parallel to the longitudinal axis of the body. Grooves below eyespots are arranged in two groups; 2-5 grooves lie transversely, and from 0-3 lie alongside laterad to and nearly parallel with the longi- tudinal axis. No grooves cross the dorsal midline. Color Color ranges between a light to purplish-gray to light brown, with occasional pinkish overtones. Pale spots and blotches (piebaldness) are common. Complete albinism and piebaldness was reported by Jensen (1959:798, figs. 1, 2). Dean (1903:295-296, fig. 1) de- scriped and figured a specimen with large pale areas on head and anterior ventral surface. DM short, moderately robust, its length 25% (22- 28%) of TL, its width 12% (11-13%) of DM length, its depth 74% (64-77%) of its width. Distance from tip of DM to branching of VA 14% (8-20%) of DM length, and that distance is 42% (25-72%) of VA length. VA length 32% (22-40%) of DM length. Numbers of GP in positions relative to DM and VA, Areas I, II, III, are given in Table 7, defined in Figure 4, and compared with similar data for the other four species treated here. Afferent duct of last GA, left side, always con- fluent with PCD. Eggs The largest egg found measured 28.6 x 7.5 mm, in a female of 435 mm TL. Well developed eggs with anchor filaments are visible but encapsulated. The number of almost fully developed eggs (20 mm or longer) varies from 11 (23 x 7 mm) in a 330-mm TL female to 48 (20 x 6 mm) in one of 515 mm TL. Neither specimen had been opened prior to our examination. Sex was determinable in a female of 179 mm TL (the eggs tiny) and in a male of 200 mm TL. The sex ratio of the study material is essentially equal. Of a total of 870 specimens for which sex was reliably determined, 49% were female, 51% male. This ratio contrasts notably with other species described herein, in which females dominated by 60-74%. The maximum size of each sex appears to be equal, the largest male measuring 550 mm TL, the largest female 515 mm TL. Four large females ranged between 491 and 500 mm TL, the next largest 475 mm TL. Discussion Lockington (1878:792-793) very briefly described BdeUndornd stoutii as, "Eleven gill openings on each side; ten teeth in the anterior and nine in the posterior series. I5V2" long. Eel River, Humboldt County." There is no doubt that the specimen represented a hagfish, but Lockinglon's subsequent statement indi- cates that perhaps he confused it with a species of lam- prey, possibly Lampetra tridentata, as he stated, "It is rather singular that this fish, which is abundant in Eel River and is sold for food, and also occurs in this harbor, should have hitherto escaped notice. I believe it to be the only species of its genus hitherto found on the Pacific coast of North America; and it differs from BdeUostmna polytrerna, a species which occurs along the coast of Chile, both in the number of its gill openings and that of the teeth, B. polytrema having fourteen of the former and twelve of the latter in each series." There can be no serious question as to the pertinence of the name stoutii to this species, although the diag- nosis is extremely short and Lockington ascribed the species primarily to Eel River, indicating it as abun- dant there. But he added, "also occurs in this harbor," meaning, presumably, San Francisco Bay, since there is no harbor at the mouth of Eel River, but only an estuarine embayment. Perhaps the specimen was sent from near the mouth of Eel River, almost certainly from central or north- ern California. We maintain the name stoutii because it is the only other species of the genus from the area fitting the description. Eptatretus deani is a deep puiplish-black in color, not light brown. It generally oc- curs at much greater depths, and it is unlikely that a fisherman would have taken E. deani in the 1870s because of this greater depth of habitat. Unfortunate- ly, Lockington did not state the numbers of prebran- 798 Fishery Bulletin 88(4), 1990 chial pores, which, aside from color, is an important character separating the two species. Lockington did not designate his specimen as a holotype or state that it was deposited anywhere. If deposited, the most lii^ely place would have been the California Academy of Sciences. W.E. Eschmeyer (Calif. Acad. Sci., San Francisco, pers. commun.) has informed us that Lockington's specimen is not now at the Academy, and stated that it and any record of it may have been destroyed in the San Francisco earth- quake and fire. Therefore, in view of the possible confusion of E. sfuutii with E. mcconnnugheyi, especially off southern California, we designate as neotype of Eptatretus stoutii, SI068-426, a male, 530 mm TL, taken near the whistler buoy at entrance of Humboldt Bay, Califor- nia. This locality is about 8 miles northerly from the mouth of Eel River, the type locality of Lockington's Bdellosfoma stoutii. Discussion The relationship of the five species treated above super- ficially appear to be close. We are aware of the spec- ulative and debatable nature of intraspecific variation, and that any proposed new species may be considered a variant of a similar-appearing, sympatric known species. Thus, it may be argued that E. deani and E. fritzi are mere variants, due primarily to their very dark coloration, much darker than in any of the other species treated, and due perhaps to their sympatric occurrence in the presently restricted range of E. fritzi near Gua- dalupe Island, Mexico. However, the virtually non- overlapping numbers of prebranchial slime pores, 7(4-10) in E. deani vs. 12-13(10-15) in E. fritzi (Table 2), alone are adequate for specific separation. In addi- tion, the very large barbels oi E. fritzi do not occur in E. dean i (Table 6) of sympatric occurrence or in any portion of the nearly 2600 miles of its total range. The superficial similarity of £". stoutii and the sym- patric E. mcconnaughtyi also invites speculation as to being mere variants. However, the notable and non-overlapping difference in respective prebranchial lengths alone distinguishes between the two species (Table 1). Also, the numbers of prebranchial slime pores differ significantly, with minor overlap (Table 2). Similary, E. stoutii and E. sinus are alike in most counts and proportions, but the very prominent ven- tral finfold of E. stoutii does not occur in E. sinus, where it is absent or vestigial. Also, when fresh material of each is compared, the much darker reddish- brown of E. sinus contrasts sharply with the much lighter gray-reddish brown of E. stoutii. All the above principal characters differentiating the five species are constant, often non-overlapping, and show little or no variation within species. Acknowledgments We are most grateful to Carl L. Hubbs for instigating a review of the Myxinidae of the world, and for his efforts (liegun in 1965) in amassing a voluminous study material. Without his early efforts, this and our pre- vious studies would not have been possible. We are most grateful to the various assistants and graduate students working under Hubbs' supervision for their efforts in counting and measuring and general accum- ulation of data used herein. Alan J. Stover and Ronald R. McConnaughey deserve our great thanks for devel- opment of gear, and their many hours at sea, involved in the capture of specimens. Especial thanks are due to David Jensen for contributing data uu thousands of E. stoutii captured during his investigations on the unique cardiovascular system of hagfishes. Also, we thank Richard H. Rosenblatt for critically reading the manuscript. Citations Anonymous 1973 Baited camera observations of demersal fi.sli. South. Calif. Coastal Water Res. Proj. TM 207, El Segundo. p, l-l(i. 1974 Observations with a baited movie camera. In South. Calif. Coastal Water Res. Proj.. Annu. Rep. 1974. Long Beach, p. 4.5-.''i2. Adam, H., and K. Strahan 1963 Notes on the habitat, aquarium maintenance, and experi- mental use of hagfishes. In Brodal, A., and R. Fange (eds.). The biology of Myxine. Universitets forlaget, Oslo, 588 p. Allen. M.J. 1976 .Mternative meth(ids I'oi- assessing fish jiopulatinns. //( Coastal Water Res. Proj.. Annu. Rep. Ut7.'i. Long Beach, p, 9.5-98, Avers, H., and J. Worthington 1907 The skin and end organs of the trigeminus and lateralis nerves i){ Bdcltiistnma dombeyi. Am. ,J. Anat. 7:.S27-.H37. Barham, E.G.. N.J. Ayer, and R.E. Boyce 1967 Macrobenthos of the San Diego Trough; photograpliic census and observations from the bathyscape Tru-xlc. Deepi Sea Res. 14:773-784. Bourne. N., and D.E. McAllister 1969 The black hagfish, Eptiitrelu^ deani. from Briti.sh Colum- bia. .1, Fish. Res, Board Can, 26:324t;-3248, Clemens, W.A., and G.V. Wilby 1946 P^ishesof the Pacific coast of Canada, Fish, Res, Board Can.. Bull. 68. 368 p. 1961 Fishes of the Pacific coast of Canadri. Fish. Res, Board Can,. Bull. 68 (2d ed.). 443 p. Day, D.S., and W.G. Pearcy 1968 Species association of benthic fishes on the continental shelf and slope off Oregon. J. P'ish. Res. Board Can. 2.5: 2665-2675. Wisner and McMillan: Three new Ept3tretus species from North American Pacific coast 799 Dean, B. 1899 On the embryology of Bdellostoma ritouti. A general ac- count of myxinoid development from the egg and segmenta- tion to hatching. Festschrift zum Siebenzigsten Gerburstag von Carl von Kruppfen. Verlag. F. Fischer, Jena:221-276. 1903 Albinism, partial albinism and polychromism in hag- fishes. Am. Nat. 37(437):2%-298. Downing. S.W., R.H. Spitzer. W,L. Sale, J.S. Downing, L.J. Saidel, and E.A. Koch 1981 Threads in the hagfish (Eptntretu.': fitutilii) slime gland thread cells: organization, biochemical features and length. Science (Wash., DC) 212(4492):326-328. Evermann. B.W., and E.L. Goldsborough 1907 The fishes of Alaska. Bull. U.S. Bur. Fish. 26 (for 1906): 219-360. Fernholm, B. 1975 Ovulation and eggs of the hagfish Eptntretiis hurgeri. Acta Zool. (Stockholm) .55:199-204. 1985 The lateral line system of cyclostomes. hi Fornian, R.E.. A. Gorbman, J.M. Dodd, and R. Olsson (eds.). Evolutionary biology of primitive fishes, p. 113-122. Plenum Press, New York and London. NATO AST Ser., Ser. A: Life Sciences, 463 p. Fitch, J,E, 1973 The ta.xonomic status of genus Astrotheca and clarifica- tion of the distribution of Buthygnnii^ jicntacdnthuf: (Pisces; Agonidae). Copeia 1973:815. Fowler, H.W. 1908 Notes on lancelets and lampreys. Proc. Acad. Nat. Sci. Phila. 59:461-466. 1947 New taxonomic names for fish-like vertebrates. Notulae Naturae, Acad. Sci., Phila. 147:1-16. Gates, D.E., and H.W, Frey 1974 Designated common names of certain marine organisms of California. Calif. Dep. Fish Game, Fish Bull. 161:55-90. Hart, J,L, 1973 Pacific fishes of Canada. Fish. Res. Board Can. Bull. 180, 740 p. Jensen, D. 1959 Albinism in the California hagfish. Eptatrftu.-i aloutii. Science (Wash, DC) 130(3378):798. 196G The hagfish. Sci. Am. 214(2):82-90. Jespersen, A. 1975 Fine structure of spermiogenesis in eastern Pacific species of hagfish (Myxinidae). Acta Zool. (Stockholm) 56: 189-198. Jordan, D.S., and C,H. Gilbert 1883 Synopsis of the fishes of North America. Bull. U.S. Natl. Mus. 16, 1018 p. Knaggs, E,H,, J.S. Sunada. and R.N. Lea 1975 Notes on some fishes collected off the outer coast of Baja California. Calif Fish Game 61:56-59. Kukowski, G. 1972 A checklist of the fishes of the Monterey Bay area, in- cluding Elkhorn Slough, the San Lorenzo, Pajaro, and Salinas Rivers. Tech. Publ. 72-2 (Annu. Rep., Pt. 2, July 1972), Moss Landing Mar. Lab., Moss Landing, CA, 69 p. 1973 Results of the Sea Grant fishes sampling program f:>r the 1971-1972 season. Tech. Publ. 73-6 (Annu. Rep., Pt. 6, 1973), Moss Landing Mar. Lab.. Moss Landing. CA. 48 p. Lcviton, A.E., R.H. Gibbs Jr., E. Heal, and G.E. Dawson 1985 Standards in herpetology and ichthyology: Part 1. Stan- dard symbolic codes for institutional resource collections in herpetology and ichthyology. Copeia 3:802-832. Linthicum. D.S. 1971 Immunity of the hagfish. Sea Frontiers 17(l):17-22. Lockington, W,N, 1878 Walks around San Francisco. No. Ill ■ Lake Honda and Seal Rock. Am. Nat. 12:786-793. Mclnerney, J,E.. and D,0, Evans 1970 Habitat characteristics of the Pacific hagfish, Poiisto- iiTiwi stoiitii. J. Fish. Res. Board Can. 27:966-968. McMillan, C.B., and R.L. Wisner 1984 Three new species of seven-gilled hagfishes (Myxinidae. Eptiitretitf:) from the Pacific Ocean. Proc. Calif. Acad. Sci. 42(16):249-267. 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. Regan, C,T, 1912 A synopsis of the myxinoids of the genus Heptatretus or Bdellostomo. Ann. Mag, Nat. Hist, (ser, 8) 9:534-536. Sasaki. T, 1972 Demersal fishes collected in the southeastern shelf waters of Alaska. Bull. Fac. Fish., Hokkaido Univ. 22(4):281-289. Smith, K.L., and R. Hessler 1974 Respiration in benthopelagic fishes: in situ measiu-ements at 12.30 meters. Sciences (NY) 184(H132):72-73. Starks, E.G., and E.L, Morris 1907 The marine fishes of southern California. Univ. Calif Publ. Zool. 3(11):159-251. Strahan, R, 1963 The behaviour of Myxinoids. Acta Zool. (Stockholm) 44: 73-102. Taylor. K,M, 1967 The chromosomes of some lower chordates. Chromo- soma (Bert.) 21:181-188. Thiessen, B, 1976 The olfactory system in the Pacific hagfishes Eptatretiis stoutii amd Eptatretus deani. and Myxine circifrons. Acta Zo(.l. (Stockholm) 57(3):167-173. Townsend, C,H,, and J,T. Nichols 1925 Deep sea fishes of the 'Albatross' lower California ex- pedition. Bull. Am. Mus. Nat. Hist. 52:1-20. Wilimovsky. N.J. 1954a Provisional list of the fishes of Alaska. Tech. Rep. V. Contract N60NR-251 36 NR 307 204. Nat. Hist. Mus.. Stan- ford Univ.. 35 p. 1954b List of the fishes of Alaska. Stanford Ichthyol. Bull. 4(5):279-294. 1958 Provisional keys to the fishes of Alaska. Fish. Res. Lab.. U.S. Fish Wildl. Serv., Juneau, 113 p. Wisner, R,L,, and C,B. McMillan 1988 A new species of hagfish, genus Efitatretus (Cyclosto- mata, Myxinidae), from the Pacific Ocean near Valparaiso, Chile, with new data on E. hischoffii and E. polytrema. Trans. San Diego Soc. Nat. Hist. 21(14):227-244. Worthington. J. 1905 Contributions to our knowledge of the Myxinoids. Am. Nat. 39:625-663. 800 Fishery Bulletin 88(4), 1990 Table I Averages and ranges (in thousandths of total length) of selected body propt rtions for five species of hagfishes (genus Epiatretus) from the Pacific coast of North A Tierica. Values for type specimens are given first and followed in parentheses by ranges in values for | all material. Species Average (range of values) E. mcconnaugheyi E. fritzi E. sinus E. deani E. stoutii A'' (Size range in mm) 91 (147-470) 233 (207-592) 458 (129-481) 460 (130-523) 520 (179-468) Preocular length 62 (45-74) 54 (48-82) .52 (41-82) • 61 (42-89) 53 (40-77) Prebranchial length 153 (147-181) 198 (176-245) 226 (197-278) 175 (144-204) 240 (187-253) Branchial length 200 (155-216) 140 (115-158) 113 (92-172) 155 (127-182) 138 (115-142) Trunk length 532 (482-557) 499 (462-556) .501 (4.50-540) 517 (480-555) .509 (470-535) Tail length 130 (122-161) 158 (132-181) 151 (102-174) 151 (126-192) 134 (104-178) Tail depth 72 (60-88) 65 (58-92) 78 (48-90) 80 (52-103) 55 (45-83) Body depth with finfcild 91 (61-98) No finfold 67 (49-104)* 78 (47-105) 62 (50-97) Body depth without finfold 89 (55-90) 63 (53-102) .58 (46-101)* 72 (45-105) 51 (41-90) Body depth at cloaca 53 (47-71) 56 (45-72) 61 (39-86) * 63 (38-85) .53 (38-79) *330 specimens measured. Table 2 Numbers o 'gill pouches, prebranchial. liranchial. and t ail slime pores for five species of hagfish es(gfc nus Eptn reliis)f ■om the Pacific coast of North America. Valuer , left side, for type specimens are indicated by an asterisk. Gill pouch es 9 10 11 12 13 14 N X 6 E. mceonnc ugheyi 28 67* 2 97 12.73 0.487 E. fritzi 31 261* 64 356 11.09 0.508 E. siniis 77 248* 103 7 435 10.09 0.685 E. deani 21 536 43 600 11.04 0.325 E. stoutii 2 173 691* 63 2 \r.', 1 11.88 0..507 Prebranchial slime pores 4 5 6 7 8 9 10 11 12 13 14 15 16 17 N X d E. mceonnc ugheyi 1 5 27 50* 11 2 96 8.74 0.845 E. fritzi 3 40* 123 118 31 3 318 12.45 0.898 E. sinus 3 27 85 142* 117 44 14 3 435 13.26 1.231 E. deani 1 7 90 275 197 23 7 600 7.26 0.854 E. stoutii 9 29 142 366 315* 59 7 920 13.27 0.946 Branchial slime pores 8 9 10 11 12 13 14 15 16 N X 6 E. mccunnnugheyi 15 52 16* 5 2 O 92 12.27 1.012 E. fritzi 14 233 105* 4 556 10.28 0.549 E. sinus 72 246* 107 10 435 9.13 0.698 E. deani 16 425 125 30 4 600 10.30 0.636 E. stoutii 3 151 587 80* 1 1 823 1(1.91 0.551 Tail slime pores 7 8 9 10 11 12 13 14 15 N X 6 E. mccoymaugheyt 2 21 44* 22 8 97 11.13 0.915 E. fritzi 2 19 87 149 80* 18 1 356 11.96 0.999 E. sinus 2 28 130 168 76* 22 7 2 435 9.90 1 .088 E. deani 7 128 224 188 47 5 1 600 11.27 0.9.58 E. stoutii 4 64 239* 357 125 22 2 813 10.75 0.935 Wisner and McMillan Three new Eptstietus species from North American Pacific coast 801 Table 3 Numbers of trunk slime po res for five species of hagfishes (gen js Epta trelu)- ) from the Pacific coast of North A merica. Values, left | side, for type specimens are indicated by an isterisk. Trunk slime pores 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 N X 6 E. mcconnaugheyi Southern California 8 7 6 13 13* 8 3 2 60 46.03 1.871 Gulf of California 1 2 3 4 11 11 3 2 37 43.08 1.566 E. fritzi 2 3 15 38 85 112 67* 24 9 1 356 44.77 1.413 E. sinufi 1 4 10 35 64 102 86* 65 40 12 7 7 1 1 435 41.67 1.943 E. deani 1 2 6 38 87 137 152 99 51 22 5 600 44.70 1.601 E. stoutii 1 8 38 96 154 196* 154 110 41 13 5 9 1 819 44.12 1.726 Table 4 Numbers of total slime pores for five species of hagfishes (genus Eptat retus) from the Pacific coast of North America. Values, left side, for type specimens are indicated by an asterisk. E. mcconnaugheyi Total slime pores Southern California Gulf of California E. fritzi E. sinus E. deani E. stoutii 66 1 67 2 2 2 68 9 5 (;9 11 16 70 1 16 26 71 2 39 48 1 72 2 5 57 100 1 73 2 8 53 111 1 74 5 11 1 63 98 11 75 3 3 6 60* 78 26 76 2 1 14 59 63 61 77 11 1 38 21 34 107 78 12 2 50 16 16 142 79 9 1 87 12 2 141 80 6* 71 9 1 116* 81 2 46* 4 100 82 4 23 3 60 83 1 10 29 84 1 8 12 85 2 7 86 3 87 - 88 1 N 60 37 3.56 435 600 819 X 77.80 73.43 79.36 73.98 75.91 79.03 6 2.650 2.444 1.909 2.796 2.186 2.280 802 Fishery Bulletin 88(4). 1990 Table 5 Numbers of anterior and posterior unicusps, and total cusps of North America. Values, left side, for type specimens are for five indicated -spec by ies of hagfishes an asterisk. (genus Ep aire fus) from the Pacific ' coast Anterior unicusps 6 7 8 9 10 N X 6 E. mcconnaugheyi 25 151* 15 191 7.95 0.455 E. fritzi 73* 266 32 1 372 7.90 0..531 E. sinus 62 453* 336 21 872 7.36 0.649 E. deani 3 62 604 308 11 988 8.27 0.601 E. stoutii 16 524 390* 11 1 942 7.42 0.555 Posterior unicusps 6 7 8 9 N X 6 E. mcconnaiighpyi 40* 143 8 191 7.83 0.472 E. fritzi 37 270* 65 372 8.08 0.518 E. sinus 11 310* 491 60 872 7.69 0.614 E. deani 233 680 75 988 7.84 0.535 E. stoutii 4 347 570* 21 942 7.65 0.,531 Total cusps 34 35 36 37 38 39 40 41 42 43 44 45 46 N X 6 E. mcconnaugheyi 4 3 11* 14 48 9 4 o 95 41.60 1.364 E. fritzi 3 5 29* 22 75 25 17 7 3 186 41.92 1.527 E. sinus 1 3 18 13 83* 46 92 39 105 16 15 1 4 436 40.1(1 2.089 E. deani 1 5 12 55 59 160 68 25 14 3 402 41.86 1.433 E. stoutii 3 4 84 66 126 76 93* 13 5 1 1 471 4(1.16 L.-.s:; Numbers, averages of hagfishes, genus Table 6 and ranges of lengths (in mm) of first nasal barbel, left side, for 25 Eptatretus, from the Pacific coast of North America. mm increments c f total lei gth of five species Average A'(Range) E. mcconnaugheyi E. fritzi E. sinus E. dea ni E. stoutii 200 -225 3.0 3(2.6-3.5) 6.5 K - ) 2.2 6(1.5-2.6) 3.3 16(2.4- 4.5) 2.3 20(2.0-3.(1) 226 -250 3.3 5(2.9-3.8) 5.9 4(5.6-6.2) 2.6 7(2.0-3.9) 3.5 33(2.6- 4.5) 2.8 16(1.9-3.2) 251 -275 3.9 8(3.1-4.6) 6.1 3(5.7-6.4) 2.8 25(2.3-3.7) 3.6 31(2.5- 4.5) 3.0 52(2.0-1.(1) 276 -300 4.1 12(3.2-4.9) 6.9 4(6.3-7.2) 2.9 50(2.2-4.3) 4.2 30(3.3- 5.3) 3.3 70(2.2-4.5) 3(11 -325 4.2 13(3.6-5.4) 6.3 8(5.0-6.9) 3.1 71(2.0-4.4) 3.8 39(3.2- 5.2) 3.5 77(2.4-4,6) 326 -350 4.5 10(3.8-5.0) 6.8 24(6.5-9.6) 3.4 80(2.1-4.4) 4.1 42(2.5- 5.5) 3.7 110(2.5-5.2) 351 -375 4.9 11(4.1-6.3) 7.4 40(5.6-9.1) 3.1 72(2.7-5.6) 4.4 103(2.8- 6.0) 3.7 127(2.5-5.1) .376 -400 4.9 8(3.9-5.6) 7.8 49(6.2-8.9) 4.0 35(3.0-5.2) 4.6 202(3.0- 6.6) 4.1 148(2.6-6.0) 401 -425 6.1 6(5.2-7.0) 7.7 71((;.5-1().2) 4.2 17(3.5-5.4) 5.(1 235(3.1- 6.9) 4.2 1(19(2.9-5.3) Wisner and McMillan: Three new Eptatretus species from North American Pacific coast 803 Table 6 (continued) Average iV(Range) E. mcconnaugheyi E. fritzi E. sinus E. deani E. stoutii 426- -450 6.0 6(5.0-7.1) 8.2 50(7.4-10.2) 3.9 8(2.9-5.8) 5.1 174(3.1-7.0) 4.3 65(2.8-6.1) 451- -475 6.5 2(6.5-6.5) 9.8 42(7.6-10.8) 3.8 2(3.2-4.4) 5.3 68(3.8-6.6) 5.3 33(3.7-5.9) 476- -500 6.8 K - ) 8.9 21(7.2-10.3) 4.1 6(3.8-4.4) 5.4 28(4.0-6.5) 4.7 11(3.9-6.0) .501- -525 6.0 2(5.8-6.1) 10.4 9(9.1-11.3) 5.2 10(4.1-5.9) 5.0 13(3.8-6.9) 526 -550 10.9 7(10.3-11.7) 4.7 K - ) 4.4 8(4.3-5.4) 551- -575 10.4 2(9.7-11.3) 4.6 K - ) 6.1 K - ) 576- -600 10.2 2(9.5-10.8) 6.8 K - ) .Y S7 273 379 1013 861 Table 7 Gill jxiuehes, both side ^. in positions ■elative t 1 Areas I, II, an 1 1 1 1 of the branchial regions of five species ot hagfishes (genus Epti trrti/s] from the Pacific C( ast of North America. Areas ai'e identifi ed in Figure 4. Values, left side for type specimens are indicate( 1 liv an asterisk. Gill pouches 0 1 2 3 4 5 6 7 8 9 10 N X 6 Area I E. mcconnaugheyi o 13 23 45* 34* 9 126 7.98 1.137 E. fritzi 9 49 34* 8 100 4.41 0.763 E. villus 2 61* 136 103 23 11 5 341 2.40 1 .067 E. deaiii 1 46 62 11 120 4.69 0.643 E. stoutii 7 61 75* 49 19 1 212 3.07 1.023 Area II E. inccoiinfiiitjhf'iii 5 13 42* 30* 23 13 126 2.73 1.275 E. fritzi 2 7 12 43* 31 5 100 3.09 1.040 E. sinus 7 20 85 151 65* 11 1 341 2.84 0.994 E. deani 30 44 43 3 120 1.16 0.827 E. stoutii 19 38 68 65* 13 9 212 2.20 1.197 Area III E. mcriiinuiugheiji 4 25* ,so 16 126 1.86 0.663 E. fritzi 1,1 42* 35 13 1 100 3.55 0.865 E. sinus 7 11 13 65 125* 101 17 ■) 341 4.97 1.229 E. deani 3 16 71 28 •> 120 5.08 0.726 E. stoutii 1 3 17* 71 87 32 1 212 6.6(1 (1.93:'. 804 Fishery Bulletin 1990 Numbers and percent of occurrence sides of five species of hagfishes (ge Table 8 of higher numbers of branchial a nus Ejjtat ret us) from the Pacific pertures on Coast of Nc left and right jrth America. Species N Higher on left side right side E, mceonnaugheyi Southern California (tulf of California 61 37 4 5 (6%) (19%) 5 (8%) 0 E. jritzi 316 81 (26%) 0 E. sinus 446 63 (14%) 7 (2%) E. denni 1145 125 (11%) 21 (2%) E. stinitii 1554 167 (11%) 56 (4%) Totals 3559 445 (12.5% 89 (2.5%) Age and Growth of Bluefish Pomatomus saltatrix from the Morthern Gulf of Mexico and U.S. South Atlantic Coast Lyman E. Barger Panama City Laboratory, Southeast Fisheries Center National Marine Fisheries Service, NOAA 3500 Delwood Beach Road, Panama City, Florida 32408 The liluefish Pomatomus snltntrix is a migratory coastal pelagic fish generally found in temperate and warm continental shelf waters of all oceans (Briggs 1960). The species occurs along the east coast of the United States and the coast of the Gulf of Mexico. Important commer- cial and recreational fisheries for bluefish exist throughout the U.S. range (Wilk 1977). Age and growth studies have been conducted on bluefish from U.S. Atlantic waters (Hamer 1959, Backus 1962, Lassiter 1962, Rich- ards 1976, Wilk 1977), but not from the Gulf of Mexico. The primary purposes of this study were to eval- uate and use the best of several bony structures to estimate the age and determine growth of bluefish from the northern Gulf of Mexico (hereafter referred to as Gulf) and from the U.S. Atlantic coast. Methods Bluefish from the northern Gulf gill- net fishery were sampled monthly from 1978 through 1982 (Feb-Nov) and in the Atlantic along the south- ern U.S. east coast in 1980 and 1981 (Jan-July). These samples were augmented by catches from the seine and hook-and-line fisheries. The Gulf samples were collected from the coastal waters off north- west Florida and Louisiana, while Atlantic samples were taken along the coast from South Carolina to Florida. Fork length (FL) to the nearest millimeter (mm), weight (W) to the nearest gram (g), and sex were recorded from 1 190 Gulf and 842 Atlantic bluefish, and one or both otoliths (sagittae) were removed, wiped clean, and stored dry in vials. A subsample of 100 fish represent- ing the entire size range of bluefish caught in the Gulf during May and June 1978 was selected for com- parison among ageing structures. From these 100 fish, in addition to otoliths, the tenth vertebra anterior to the hypural plate was removed and scales were taken from the left side under the pectoral fin. Verte- brae were cleaned and air-dried. Both vertebrae and scales were stored dry in envelopes. Otoliths were placed in glycerol, sulcus acousticus down, in a black dish and were examined under re- flected light using a binocular-dis- secting microscope with an ocular micrometer. The most legible oto- lith from each fish was examined for age marks. The second otolith from 25 of the bluefish was sectioned to allow interior examination. Tlie otoliths were embedded in Lakeside 70C thermoplastic cement and, to include the locus, 2 or 3 thin sec- tions (0.15 mm) were cut along the Reference to trade names does not imply en- dorsement by the National Marine Fisheries Service, NOAA. transverse plane, using a Buehler isomet slow-speed saw. The cement was dissolved with isopropyl alco- hol, and then the otolith sections were mounted on glass slides and examined on a black background in the manner of whole otoliths. On whole or sectioned otoliths an opaque zone (mark) preceded by a translucent (hyaline) zone (Fig. la) was assumed to be an age mark. Measurements (OR) were made along the longitudinal axis of the rostrum from the focus to the distal edges of the marks and of the otolith. A determination of the edge character (opaque or translu- cent) was made, and marks were counted. A mark was not consid- ered complete, so not counted or measured, unless the portion of the otolith distal to it was translucent. The rostrum was selected for ex- amination because the postrostrum had an uneven edge which proved difficult to measure. Vertebrae were stained with crys- tal violet following the technique of Johnson (1979). To facilitate the ob- servation of age marks, the verte- brae were cut in half, anterior- posteriorly, with a Dremel saw. Both halves of the vertebrae were examined under a binocular-dissect- ing microscope using reflected light. The most legible posterior centrum was used. An age mark on the vertebral cone surface was counted if a prom- inent concentric ridge preceded by a depression was observed (Fig. lb). Measurements were made from the vertex (focus) of the centrum to the distal edge of each mark as well as to the terminal edge of the vertebra. Scales were cleaned in a weak solution of water and liquid deter- gent, then mounted between two glass slides. Scales from 15 blue- fish were also impressed on plastic slides with a cold roller press. Both Manuscript accepted Hi Ma.v 1990. Fishery Bulletin. U.S. 88:80.5-809. 805 806 Fishery Bulletin 88(4), 1990 Figure 1 otolith (a), vertt'bral section (b), and scale (c) from a L! + year-oM l)luefish, 375 mm FL, captured in May. Distal edges of age marks are indicated by Roman numerals. F = focus; T = translucent edge. slides and impressions were viewed on an Eherliach scale reader. The most legible unregenerated scale was examined for age marks. On scales and scale impressions an age mark was con- sidered to be a band of widely spaced circuli. usually witli broken circuli in the anterior field and/or ana- Figure 2 Percent of otoliths with opaque edges liy month from (iulf of Mex- ico bluefish, 1978-82 (»). and U.S. South Atlantic coast, 1981-82 ( + ). stomosis (crossing over) in the lateral field, followed by a series of closely spaced circuli (Fig, ]c). Measure- ments were made on the projected image from the focus along a radius in the center of the anterior field to the distal edge of the marks and scale edge. Two investigators each made one reading of a set of otoliths, vertebrae, and scales from 100 bluefish. All examinations were made without reference to fish length or interpretations of the other investigator. Computer analyses and plots were accomplished using SAS (Ver. 6,03) software, Backcalculations of length-at-age were accomplished with a program writ- ten by the author using means weighted by multi])lica- tion of number of samples. Gulf and Atlantic bluefish were analyzed separately. Separate analyses by sex in- cluded regressions of fork length and natural log of fork length on otolith length. Least-squares regressions were used to backcalculate length-at-age for males and females separately (Ricker 1975). Multiple regres- sion of sex and age was run to determine difference in length-at-age by sex. Von Hertalanffy theoretical growth curves were calculated using weighted-mean backcalculated fork lengths. The growth etiuation (von Bertalanffy 1938, 1957) was the following: L(i -k(t- ...)) where 1, = length at age. 'oo = asymptotic length, k = growth coefficient, and t|, = time when length would theoretically be zero. NOTE Barger Age and growth of Pomatomus saltatn\ from U S South Atlantic and Gulf coasts 807 Table 1 Backcalculated weighted mean fork length 3 of bluefish. Mean backcalculated length-at-age Estimated Mean length (mm age group Number at capture (mm) 1 2 3 4 5 6 7 8 From northern Gulf of Mexico, 1978-82 1 389 364 322 2 69 441 284 400 3 30 640 307 415 488 4 27 719 285 430 515 575 5 50 745 272 411 510 572 622 6 29 766 283 424 517 579 628 673 7 11 766 295 423 512 576 627 668 708 8 6 767 275 423 530 598 650 695 728 766 Weighted mean 308 413 509 576 627 675 715 766 Annual increment 105 96 67 51 48 40 51 Number 611 222 153 123 96 46 17 6 From U.S. Atlantic coast, 1980 -81 1 .38;) 382 299 2 161 394 275 362 3 26 446 265 350 412 4 12 536 261 342 421 473 Weighted mean 290 361 415 473 Annual increment 69 56 58 Number .=^88 199 38 12 Results and discussion Agreement between investigators in enumeration of marks was highest (92%) with whole otoliths. Lower agreements were attained for scale impressions (67%), vertebrae (33%), and scales (24%). In comparison of whole otoliths with cross-sections taken from the other otolith of a pair, investigators agreed 70% of the time. The close spacing of marks in the sections caused more disagreement in mark enumeration than did the wider sjiacing in the rostrum of the whole otolith. Also, frac- tures sometimes occurred in preparation of thin sec- tions, making enumeration of marks on cross-sections of otoliths a less viable option. In addition, 30 otoliths were examined and no difference was found in mark counts on either the rostrum or postrostrum of the otolith or between pairs of otoliths. To use otoliths, or any structure, for age determina- tion, the deposition of regular detectable age marks is essential. Because samples were obtained from catches of the fishery, no accepted method of direct validation could be employed. However, indirect evidence was established by correlation of the ol)served mark forma- tion at the distal edge of the r'ostrum with month. Despite the lack of samples for all months, the results suggested that opaque marks were formed annually in late winter or early spring around March and April in the Gulf and Atlantic samples, respectively (Fig. 2). Gulf bluefish show an unexpected flattening of the curve in June and July. The reason is not known, but a likely hy]3othesis is a stress-induced check from en- vironmental causes. An alternate hypothesis could be a multiple spawning. However, there are no reports of a summer spawn of bluefish in the Gulf. Backcalcula- tion of length at the time of mark formation is depen- dent on the relationship between the size of the age- ing structure and fish length. Improved fit of the otolith radii (OR) to length relationship occurred when natural log transformation was used for fork length. The equa- tion for Gulf bluefish was LOG(FL) = 4.200 + 0.389 x OR (/•- = 0.86). The equation for Atlantic bluefish was LOG(FL) = 4.822 -i- 0.248 x OR (r- = 0.61). Sexes were pooled for analysis because no significant difference (a>0.10) was found between mean backcalculated fork lengths of sex at age. Studies of Hamer (1959) in the New York Bight, Lassiter (1962) off North Carolina, and Richards (1976) off Long Island also showed no appreciable difference in growth bet- ween sexes. The length-weight equation for Gulf bluefish was W = - 10.02 x FL-^" and was W = -9.18 X FL2 "?' for Atlantic bluefish. Backcalculated lengths-at-age from Gulf and Atlan- tic bluefish were similar to the respective lengths at Fishery Bulletin 88(4), 1990 capture (Table 1). The differences were attributed to growth after mark formation. Gulf bluefish were con- sistently greater in length for age compared with Atlantic bluefish. The difference gradually increased through age-4, the ma.ximum age of the available Atlan- tic samples. The Gulf samples included large fish (>600 mm FL) in relatively large numbers (>30% of the sample) only in May and June 1978. During other years, the large fish made up only 4% or less of the Gulf samples. Bluefish of this size are not commonly found in the Gulf. Fable et al. (1981) reported that these fish were decidedly larger than Gulf bluefish observed in 1973 and 1977. Atlantic samples included fish estimated to be no older than age 4 and FL less than 600 mm. A general comparison with some previous bluefish studies was made (Fig. 3). All other studies used scales to age bluefish and did not report validation. Scale studies were used, as no similar otolith studies could be obtained. Wilk (1977) analyzed scales from bluefish sampled along the Atlantic coast. His results showed slower gi-owth to age-1 than either Gulf or Atlantic samples from this study. By age-2 a larger rate of growth had converged his reported lengths to just short of the Gulf samples from this study, a relationship maintained until age-7 at which point small sample sizes may have contributed to errors. Reported lengths- at-age from scales of Long Island Sound bluefish (Richards 1976) were similar to those of this study. North Carolina spring-spawned bluefish aged by scales (Lassiter 1962) fall within range of Gulf bluefish from this study, but have shorter lengths at each age. The von Bertalanffy (1938, 1957) theoretical growth parameters derived from this study are: Gulf: k = 0.180, 1^ = 944, t„ = -1.033 Atlantic: k = 0.096, 1^ = 1,019, t,, = -2.493. Growth coefficient (k) for Gulf bluefish is within the range of those reported by Lassiter (1962) for North Carolina spring-spawned (0.103) and summer-spawned (0.342) and by Manooch (1979) for Gulf and Atlantic maximum age-8 (0.230) and maximum age-9 (0.340) bluefish. The high growth coefficient of Gulf bluefish reflects the relatively rapid initial growth. The large bluefish taken in the summer of 1978 may influence this coefficient. Otoliths appear to be better than either scales or vertebrae for ageing Gulf bluefish. Based on the per- cent of opaque edge occurrence and the close fit with other studies, it appears that age marks on the otoliths of Gulf bluefish are formed annually. Direct validation of age, which was not possible in this study, should be included in future studies. The cause of a higher per- cent of otoliths with opaque edges in the months of f ; ' I U A U D ft f, [ Figure 3 Meai) length iif bluefish from the ( nJf of Me.xico and the U.S. Athmlic coast by age in years from this and other published studies. North- ern Gulf of Mexico, this study (»); U.S. South Atlantic coast, this study ( + ); Long Island Sound. NY (Richards 1976), (0); New York Bight (Hamer 1959). (O); Atlantic coast (Wilk 1977), (D); North Carolina (Lassiter 1962), (A). June and July in Gulf bluefish should also be a point of further investigation. Acknowledgments I wish to thank Drs. Charles S. Manooch III of the National Marine Fisheries Service, Stephen Bortone of the University of West Florida, and Michael J. Van Den Avyle of the University of Georgia for their critical reviews of the original manuscript. I would like to thank Drs. Allyn G. Johnson and J. Jeffery Isely of the National Marine Fisheries Service for assistance in reading structures and guidance in statistical analysis. Citations Backus, R.A. 1962 Age in a small sample ul' hluel'ish {PditKiloniiia salliilrix Linneaus). Breviora l.'i9:l-4. Briggs, J.C. 1960 Fishes of worldwide (circumtropical) distriliution. Copeia 1960:171-180. Fable. W.A. Jr.. H.A. Brusher. L. Trent, and J. P'innegan Jr. 1981 Fos.sible temperature effects on charter Ixi.-it catclies of king mackerel and other coastal pelagic species in northwest Florida. Mar. Fish. Rev. 43(8):21-26. Hamer, P.E. 1959 Age and growth studies of the bluefish (Pomaloynus saltatrii Linnaeus) of the New York Bight. M.S. thesis, Rutgers Univ.. New Brunswick. N.I, 27 p. NOTE Barger' Age and growth of Pomatomus saltatnx from US South Atlantic and Gulf coasts 809 Johnson, A.G. 1979 A simple method for staining the centra of teleost vertebrae. Northeast Gulf Sci. 3(2):113-115. Lassiter, R.R. 1962 Life history aspects of the bluefish, Pomatomus saltatrix (Linneaus). from the coast of North Carolina. M.S. thesis, North Carolina State College, Raleigh. 10.3 p. Manooch, C.S. Ill 1979 Recreational and commercial fisheries for tcing mackerel, Scomberomorus cavalla, in the South Atlantic Bight and Gulf of Mexico, U.S.A. hi Nakamura, E.L., and H.R. Bullis Jr. (eds.), Proceedings: Colloquium on the Spanish and king mackerel resources in the Gulf of Mexico, p. 33-41. Gulf States Mar. Fish. Comm. 4. Richards, S.W. 1976 Age, growth and food of bluefish {Pomatomus saltatrix) from east-central Long Island Sound from July through November 1975. Trans. Am. Fish. Soc. 105:523-525. Ricker, W.E. 1975 Computation and interpretation of biological statistics of fish populations. Bull. Fish. Res. Board Can. 191, 382 p. von Bertalanffv, L. 1938 A quantitative theory of organic growth (inquiries on growth laws. II). Hum. Biol. 10(2):181-213. 1957 Quantitative laws in metabolism and growth. Q. Rev. Biol. 32(3):217-231. Wilk, S.J. 1977 Biological and fisheries data on liluefi.sh, Pomatomus sa/fa(nx (Linneaus). Tech. Ser. Rep. 11, Sandy Hook Lab., Northeast Fish. Cent., Natl. Mar. Fish. Serv., NOAA, Highlands, NJ 07732, 56 p. Food Habits of Larval Sablefish Anoplopoma fimbria from the Bering Sea Jill J. Grover College of Oceanography, Oregon State University Hatfield Marine Science Center, Newport, Oregon 97365 Bori L. Olla Cooperative Institute for Marine Resources Studies, Alaska Fisheries Science Center National Marine Fisheries Service, NOAA, Hatfield Marine Science Center Newport, Oregon 97365 Sablefish Anoplopoma fimbria spawn at depths exceeding 400 m (Mason et al. 1983) and their larvae ascend to surface waters soon after hatching (Kendall and Matarese 1987). While in the neuston, they are subjected to highly variable oceanographic conditions. A com- parison of the diet of sablefish lar- vae collected off Oregon and Wash- ington between years that differed markedly in oceanographic condi- tions (1980 and 1983) revealed a difference in the size of copepods that were ingested (Grover and Olla 1987). Clearly diet can vary be- tween years in one location in re- sponse to oceanographic conditions, yet it is unclear how larval diet might vary between distant geo- graphical areas of differing ocean- ographic conditions. The aim of the present work was to examine the food habits of larval sablefish that were collected in the Bering Sea in 1979 and compare these findings with those from an earlier study of specimens collected off Oregon and Washington (Grover and Olla 1986, 1987). Methods Sablefish larvae were collected dur- - 1 1 St Matthew I Nunivak I, "^^^y ■" • Pnbilof Is. • • • Dutch Harbor / <&^ ^1,— °- 1 1 1 1 60° N - 58° - 56° - 54° I76°W 172° 168° Figure 1 Location of collec- tion stations in the I64°W eastern Bering Sea. ing 1979 as part of an ichthyoplank- ton survey in the Eastern Bering Sea (Walline 1981). Samples were collected using a modified Sameoto neuston sampler, 0.3 x 0.5 m with 0.505-mm mesh, towed at the sur- face for 10 minutes at 2-3 knots. Surfacewater temperatures ranged from about 6°C to just over 8°C. Collections were made from the RV Miller Freeman between 17 June and 10 July 1979. All specimens (A'^= 127) from 5 stations (Fig. 1) where larval den- sities were at least 1/10 m- surface area were examined. This repre- sents more than 40% of the total number of sablefish larvae that were collected in 1979. Larvae were preserved in 10% formalin and were switched into 5% formalin after sorting. Standard length (SL) of each larva was mea- sured, the digestive tract was re- moved, and contents from the en- tire digestive tract were identified. Diet was analyzed in terms of numerical percentage composition (%A''), percent frequency of occur- rence (%F0), and volumetric per- centage composition (%VOL). Prey volumes were calculated from prey dimensions (Grover and Olla 1987). An index of relative importance (IRI = (%A^ + %VOL) X %F0) (Pin- kas et al. 1971) was used for a more comprehensive assessment of prey importance. Data were examined by larval size group (12-15 and 16-23 mm SL). Prey was categorized as diatoms, copepod eggs, copepod nauplii, Oi- thona similis, Pseudocalanus sp. adults and copepodites, Acartia longiremis, Acartia spp., unidenti- fied copepods, amphipods, and all other prey items. Prey widths that were used to ex- amine prey-size selection were ob- tained only from items that were not badly broken or flattened. Over 2900 prey items were measured. Manuscript accepted 13 July 1990. Fishery Bulletin. U.S. 88:8i 1-814. Fishery Bulletin 88(4). 1990 Median widths of diatoms were the smallest (0.01-0.10 mm), followed by copepod eggs, nauplii, and 0. similis (0.11-0.20 mm), Pseudocalanus copepodites, A. longi- remis, Acartia spp., unidentified copepods, and other prey (0.21-0.30 mm), and Pseudocalanus sp. adults and amphipods (0.31-0.40 mm). Results A single prey-size class was most often ingested by all larvae (Fig. 2A). While both large and small larvae con- sumed small prey, the ingestion of larger prey by large larvae resulted in the two curves being significantly different (P<0.01, Kolmogorov-Smirnov test; Conover 1980). The shared mode of the two curves was attrib- uted to large larvae ingesting a number of small prey items, particularly copepod eggs (0.11-0.20 mm) (Table 1). When we eliminated copepod eggs from the anal- yses, assuming that they were taken incidentally in the course of ingesting adult female Pseudocalanus sp., the prey-size distribution of small larvae did not change noticeably (Fig. 2B). But the removal of copepod eggs from the diet of large larvae shifted the prey-size mode towards larger prey. Copepod nauplii were the primary prey of small lar- vae (Table 1). Pseudocalanus sp. also comprised a major portion of the diet, with adults and copepodites nearly equal in importance. Unidentified copepods made a small contribution, while Acariia spp., Oithona similis, and copepod eggs were rather insignificant in the diet of small larvae. Copepods were the primary prey of large larvae (Table 1), with Pseudocalanus sp. adults predominant over copepodites and all other species. A large number of copepod eggs were ingested, while amphipods, un- identified copepods, Acartia longiremis, and copepod nauplii made progressively smaller contributions to the diet. The relatively large size of amphipods likely con- tributed to their being ingested only by large larvae. The mean number of prey items ingested by small larvae was 26.9, compared with 34.1 by large larvae. The incidence of feeding was 100%. Discussion The pattern of prey-size selection observed in the Ber- ing Sea contrasts with the pattern seen for larvae col- lected off the Oregon and Washington coasts in 1980 (Grover and 011a 1986). In each study functional breaks in diet (e.g., the shift away from dependence on cope- pod nauplii) were used to define larval size groups. The Oregon- Washington collections were categorized in- to three size groups: 8.2-12.5, 12.6-20.5, and 20.6-28.5 o Q_ O O o 80 70 ~ 60- 50- 40- 30- 20 10 0 16 12 23 mm SL 1 5 mm SL 001- oil- OPI- OID 0 20 OJO OBI- 0 90 PREY WIDTH (mm) Figure 2 Size of prey selected by two size classes of larval sablefish (12-15 and 16-2.3 mm SL) in the Bering Sea in 1979: (A) including copepod eggs, (B) omitting copepod eggs. mm SL. Each size group favored a different peak prey size, with the smallest larvae ingesting the smallest prey, and the largest larvae ingesting the largest prey (Grover and 011a 1986). However, peak prey sizes for the two size groups from the Bering Sea were the same as long as copepod eggs were included in the analyses. Their removal shifted the prey-size utilization curve of large larvae in the Bering Sea towards larger prey. Copepod eggs had a greater impact on the diet in the Bering Sea than off Oregon due to a difference in the biology of the dominant copepod species ingested in the two regions. The dominant copepod in the diet of large larvae collected off Oregon in 1980 was Paracalanus parvus (Grover and OUa 1987), a species that broad- casts its eggs into the sea (Checkley 1980). Pseudo- calanus sp., the dominant copepod in the diet of larvae collected from the Bering Sea, carries its eggs (Cor- kett and McLaren 1978) which could thus be ingested incidental to the adults. Nevertheless, as female cope- pods with eggs were not observed in sablefish guts, and as other copepods that broadcast their eggs (such as all Acartia spp.) contributed to the diet, it remains a possibility that some eggs were ingested independently NOTE Graver and Olla Food habits of larval Anoplopoma fimbria from the Bering Sea 813 Table I Composition of the diet of larval sablefish in the Bering Sea in terms of the Index of Relative Impo •tance (IRI) and its components: 1 Numerical percent composition (%7V), frequency of occurrence (%F0), and volumetric percent composition (%VOL), by size class. Prey Larval size class 12-15 mm 16- -23 mm <7oN %VOL %F0 %IRI WoN %VOL %F0 %IRI Diatoms 0.2 <0.1 4.5 <0.1 3.3 <0.1 30.8 0.8 Copepod eggs 2.6 0.3 22.7 0.4 20.2 O.S 74.4 11.2 Copepod nauplii 74.1 41.6 100.0 77.9 12.9 3.0 43.6 5.0 Oithona tiimilis 2.4 2.5 27.3 0.9 5.0 2.2 38.5 2.0 Pseudocalanus sp. adults 4.2 23.6 39.8 7.4 23.7 55.7 92.3 .52.7 Pseudornlanus copepodites 9.6 17.8 50.0 9.2 9.4 8.2 74.4 9.4 A ca rt ill longi rem is 0.:^ 0.8 1.1 <0.1 8.4 8.6 41.0 5.0 Acartia sp. 2.4 4.7 19.3 0.9 2.1 1.7 .38.5 1.1 Unident. copepods 3.7 7.4 40.9 3.1 7.5 6.3 56.4 5.6 Amphipods 7.2 13.2 48.7 7.1 Other 0.5 1.2 10.2 0.1 0.3 0.3 7.7 <0.1 of females. However, since the number of copepod eggs in tlie diet appeared to be positively correlated with the number of adult Psei((localanuf> sp. ingested, we conclude that most copepod eggs were probably at- tached to adult female Pseudocalanus sp. when they were ingested. The mechanical action of ingestion most likely liberated the eggs from their delicate egg sacs (see Grover 1990). These data suggest that copepod eggs may be important in the diets of larval sablefish and other species in the Bering Sea. In contrast to the larval diet off Oregon and Wash- ington, copepods >2 mm were inconsequential in the diet of larvae in the Bering Sea. While copepod species with southern affinities were absent from the diet of sablefish larvae collected in the Bering Sea, species with northern affinities, i.e., Pseudocalanus sp. and A. longiremis, occurred more frequently in the diet of lar- vae collected in the Bering Sea than off Oregon (Grover and Olla 1987). Euphausiid larvae, appendicularians, and pteropods were absent from the diet of sablefish larvae collected in the Bering Sea, while they made a noticeable contribution to the diet of sablefish larvae collected off Oregon. Specific differences in prey items could be expected when comparing geogi'aphically separated populations; however underlying principles of prey-size selection are assumed to be universal in nature (Hunter 1981). We looked beyond the initial data plots and found a bio- logical explanation as to why the size of prey consumed by large larvae in the Bering Sea differed from the pat- tern observed off Oregon (Grover and Olla 1986). Our data revealed that copepod eggs have such a confound- ing influence on prey-size distributions that the mode of their ingestion (i.e., free-floating vs. attached to adults) must be recognized in order to adequately in- terpret prey-size selection patterns. Acknowledgments We wish to thank C.B. Miller for sharing his knowledge of copepod biology, and Art Kendall for his continued interest in the early life history of this species. This work was supported by the Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, Contract Nos. NA-85-ABH-00025 and NA-89-ABH-00039. Citations Corkett. C.J., and I. A. McLaren 1978 The hiohgy- of Pspiiiiocnlarius. Adv. IVIar. Biol. 15:1-231. Checkley, D.M. Jr. 1980 The egg production of a marine planktonic copepod in relation to its food supply: Laboratory studies. Limnol. Oceanogr. 25:430-446. Conover, W.J. 1980 Practical nonparametric statistics. 2d ed. John Wiley and Sons, NY, 493 p. Grover, J.J. 1990 Feeding ecology of late-larval and early-juvenile walleye pollock Theragra chalcogrammn from the Gulf of Alaska in 1987. Fish. Bull., U.S. 88:46.3-470. Grover. J.J., and B.L. Olla 1986 Morphological evidence for starvation and prey size selection of sea-caught larval sablefish, Anuplopumafimhna. Fi.sh. Bull., U.S. 84:484-489. 814 Fishery Bulletin 88(4), 1990 1987 Effects of an El Nino event on the food habits of larval sablefish, Anoplopuma fimbria, off Oregon and Washington. Fish. Bull.. U.S. 85:71-79. Hunter, J.R. 1981 Feeding ecology and predation of marine fish larvae. In Lasker, R. (ed.), Marine fish larvae: Morphology, ecology, and relation to fisheries, p. 33-77. Sea Grant Prog., Univ. Wash. Press. Seattle. Kendall, A.W. Jr., and A.C. Matarese 1987 Biology of eggs, larvae, and epipelagic juveniles of sable- fish, Anoplopoma fimbria, in relation to their potential use in management. Mar. Fish. Rev. 49(1):1-13. Mason, J.C, R.J. Beamish, and G.A. McFarlane 1983 Sexual maturity, fecundity, spawning, and early life history of sablefish (Anoplopoma fimbria) off the Pacific coast of Canada. Can. J. Fish. Aquat. Sci. 40:2126-2134. Pinkas, L., M.S. Oliphant, and I.L. Iverson 1971 Food habits of albacore, bluefin tuna, and bonito in California waters. Calif. Dep. Fish Game Fish Bull. 1.52. 105 p. Walline, P.D. 1981 Distribution of ichthyoplankton in the eastern Bering Sea during June and July 1979. Proc. Rep. 81-03. Northwest and Alaska Fish. Cent., Natl. Mar. Fish. Serv., NOAA, Seattle, WA 98112, 32 p. Movement of Tagged Lingcod Ophiodon elongatus at Meah Bay, Washington Thomas H. Jagielo Washington Department of Fisheries 7600 Sand Point Way NE, Seattle, Washington 981 15 Lingcod Ophiodon elongatus is an important component of Wasliing- ton's coastal commercial and spoi't fisheries (Jagielo 1989). Female lingcod comprise most of the off- shore commercial landings, grow faster than males, and attain a larger size. Males predominate in the near- shore sport catch, are typically smaller than females, and perform a nest guarding role during repro- duction. These characteristics of the coastal lingcod population suggest the need for an integrated stock assessment which addresses males and females separately and takes the geographic distribution and mi- gratory behavior into account. This paper reports the movement of lingcod tagged nearshore in the western Strait of Juan de Fuca near Neah Bay, Washington, during 1986-89. Results are presented for tags recovered through 1989 and analyzed by sex and size for extent and direction of movement. I also evaluated tag retention of a wire spaghetti tag applied on the preop- ercular plate as an alternative to strap, anchor, or dart tags used pre- viously for lingcod tagging (Chat- win 1958, Forrester 1973, Cass et al. 1983, Mathews and LaRiviere 1987). Methods From 1986 to 1989, 3478 lingcod were tagged using a chartered com- mercial vessel trolling with 6-10 jigs from a hydraulic gurdy. Each year fishing occurred from mid- March to mid-April in advance of the sportfishery opening on 15 April. Only fish not injured by cap- ture were tagged and released. All tagged fish were measured to the nearest millimeter, and sex was determined by noting the presence of anal papillae in males. The tagging area was within 3 miles of the shoreline in the vicin- ity of Neah Bay, Washington, and extended from the Sekiu River to Makah Bay (Fig. 1). In 1986, the tagging effort was distributed even- ly among areas NB-1, NB-2, and NB-3. Area NB-4 was added in 1987, and effort was distributed evenly among the four areas from 1987 to 1989. Most of the tagging occurred at depths between 15-25 m. In 1986 two tag types were re- leased: 481 fish were tagged with a large dart tag (Floy FT-1) applied dorsally, and 487 fish were tagged with a wire spaghetti tag (Floy FT-4) twist-tied to the preopercuiar plate. Fish were alternately tagged with one of the two tag types and were released back into the popula- tion. Recaptures from 1986 to 1989 for the two tag types are shown in Table 1. In 1987 and 1988, 207 fish were double-tagged with the spag- hetti tag to evaluate tag shedding. Through September of 1989, 20 of the double-tagged fish were recov- ered, all with both tags in place. Since tag shedding appeared to be negligible for the spaghetti tag, only fish released with the spaghetti tag (2997 lingcod) were analyzed for movement trends in this paper. Recapture information, including the date and location of captiu-e. was obtained both by direct inter- views with fishermen and by volun- tary returns submitted by fisher- men. A $10 reward was paid for the return of tags, which was available directly on landing at Neah Bay. Migratory and nonmigratory lingcod were defined as fish recap- tured at distances greater than or less than 8.1 km (5 miles) from the tagging location, respectively. This reference distance was selected to enable comparisons with previous tagging studies. Chi-square contin- gency-table analysis was used for comparing release length-frequency distributions of migratory and non- migratory recoveries and migra- tional tendency by sex. A chi-square goodness-of-fit test was used to test the null hypothesis that the release length distribution of recaptured lingcod was the same as the release length disti'ibution of all tagged lingcod. Length-frequency distribu- tions were grouped into 5-cm inter- vals and pooled at the tails so that no expected cell frequency was < 1.0 and no more than 20% of the ex- pected cell frequencies was <5.0 (Zar 1974, p. 50). One-way analysis of variance was used to test the null hypothesis that the mean time at liberty was the same for fish that had migrated different distances. Results Through September 1989, 393 (13.1%) tagged lingcod were recap- tured (Table 2). The percent recap- tured for each release group rang- ed from 9.96%, in 1989 to 18.89%. in 1986. The lower recovery rate for the 1989 release is probably a re- flection of fewer recovery years as compared with releases from 1986 to 1988. The length distribution of tagged lingcod by sex are shown in Figure 2a. Of all tagged lingcod, 99% were sexed; 84% of this sample were Manuscript accepted 2;t .June 1990. Fishery Bulletin. U.S. 88:815-820. il6 Fishery Bulletin 88(4), 1990 Figure 1 Lingcod tagging area at Neah Bay. Washington. Rates of recapture of tagged Table lingcod for 1 two types of tags released in 1986. Tag type No. tagged No. tags -eturned by fishermen % returned 1986 1987 1988 1989 Total Large dart (Floy FT-1) Spaghetti (Floy FT-4) 481 487 41 47 25 29 7 13 1 3 74 92 1.5.38 18.89 Table 2 Number of lingcod tagged. 1986-89. and ret aiitured th ■ough 1989 by year of recapture. No. % Year tagged tagged 1986 1987 1988 1989 Total recaptured 1986 487 47 29 13 3 92 18.89 1987 564 0 36 20 15 71 12.59 1988 982 0 0 79 55 134 13.64 1989 964 0 0 0 96 96 9.96 Total 2997 47 65 112 169 393 13.11 NOTE Jagielo Movement of Ophiodon elongatus at Neah Bay. Washington o c CO O" CD > 0.4 0.3 0.2 0.1 0, 0.4 0.3 0.2 0.1 0, 0.4 0.3 0.2 0.1 0 Female n = 443, 'X = 60.68 Male n = 2525, X = 58.45 20 80 100 120 All rel. n = 2993, X = 58.66 All rec. n = 391, X = 60.01 0 40 60 80 100 120 < 8.1 km n = 292, 7 = 59.59 >= 8 1 km n = 70, "X = 61 .93 20 40 60 80 100 120 Length (cm) Figure 2 Length-frequency distributions of tagged lingcod. (A) Known male and female lingcod tagged; (B) release length distribution of all lingcod released and all lingcod recovered; (C) release length distribu- tion of all lingcod recovered <8.1 km from release location and all lingcod recovered >8.1 km from release location. males. The average size of tagged males (58.45 cm) was less than the average size of females (60.68 cm). Based on the length-at-maturity values of 46 cm for males and 76 cm for females (Mathews and LaRiviere 1987), approximately 95% of the tagged males were mature, and approximately 10% of the tagged females were mature at the time of tagging. The average size of all tagged fish recaptured (60.01 cm) was greater than the average size of all tagged fish released (58.66 cm) (Fig. 2b). The null hypothesis that the length distribution of all recaptured lingcod is the same as that of all tagged lingcod was rejected (x" = 14.91 with 7 df; P = 0.0371). The average size of fish considered migratory (61.93 cm) was greater than those considered nonmigratory (59.59 cm) (Fig. 2c). The null hypothesis that migratory fish had the same length distribution as nonmigratory fish was rejected (x" = 51.42 with 5 df; P<0.00001). Of the 363 lingcod with known tagging and recap- ture locations, 70 (19.3%) were recaptured >8.1 km (5 miles) from the tagging location and were considered migratory, while the remaining 293 were recaptured within 8.1 km of the tagging location and were con- sidered nonmigratory (Table 3). Of those that migrated, 46 were recaptured from 8.1 to 50 km from the tag- ging site, and 24 were recaptured >50 km from the tag- ging site. Relative to the tagging location, most of the migra- tory recaptures were westward and out of the Strait of Juan de Fuca as opposed to eastward and inside the Strait. Of the 70 lingcod judged to be migratory, 54 were recaptured west and north/south of the tagging location, while only 16 were recaptured east and north/south of the tagging location (Table 3). The null hypothesis that male and female tagged lingcod were as likely to be recaptured east and north/south as op- posed to west and north/south) of the tagging location was rejected (r = 20.62 with 1 df; P<0.00001). Recap- tures came from as far north as Queen Charlotte Sound (241 miles), as far south as Cape Falcon (149 miles), and as far east in the Strait of Juan de Fuca as Cres- cent Bay (36 miles). Most of the migratory fish (25) were recovered on trawl grounds off the Washington coast and included those in the vicinity of Cape Flat- tery (5). the Cape Flattery Spit (9), Umitilla/La Push (4), Destruction Island (4), and Cape Elizabeth (3). Table 3 Distribution of recoveries of tagged lingcod Ijy distance and dn-ection of migration. 8.1-50 km >50 km Recaptured No. % No. tagged <8.1 km East West Total East West Total 2997 293 15 31 46 1 23 24 393* 13.11 •Includes 30 recaptures with unknown recapture location. Fishery Bulletin 88(4), 1990 Table 4 Distribution of recaptured iingcod by sex and distance of migration. Sex Total released Number recaptured Unknown distance Recaptured <8.1 km 8.1 -50 km >50km Total % Male 2526 245 42 24 23 334 13.22 Female 443 47 4 0 6 57 12.87 Unknown 28 1 0 0 1 Total 2997 293 411 24 30 393 13.11 Thirteen were recovered on Canadian trawl grounds including La Perouse Bank (8) and Swiftsure Bank (4). Twenty of the migratory recaptures occurred within the tagging area between Makah Bay and the Sekiu River (Fig. 1), and 12 recaptures were made east of the study area in the Strait of Juan de Fuca. Of fish with known sex and recapture location, a higher proportion of male Iingcod were migratory as compared with female Iingcod. Of 311 male recaptures, 66 (21.2 %) were judged to be migratory, while 4 of 51 female recaptures (7.8%) were considered migratory (Table 4). The null hypothesis that male and female recaptures were equally likely to be migratory was re- jected (r = 4.20 withl df; P = 0.0402 with Yates correction). The time span between tagging and recapture for all recaptured Iingcod averaged 237.6 days and ranged from 6 to 1197 days (Table 5). The null hypothesis that the average time span between tagging and recapture was the same for fish recaptured at rlifferent distances was not rejected (F^sm = 1-50; P = 0.2240). Discussion This study gives qualitative evidence that a portion of the nearshore Iingcod population in the vicinity of Neah Bay is vulnerable to the offshore trawl fishery; how- ever, fishery tag recapture data, unadjusted for dif- ferential fishing effort, are inadequate to make quan- titative statements about the net mixing rates of fish between areas. Clearly, from a harvest management perspective, it is important to know whether offshore stocks of Iingcod contribute to nearshore recruitment, or vice versa, since valued fisheries operate in both areas. Previous studies in the Strait of Juan de Fuca and Strait of Georgia have reported variable Iingcod move- ment and indicate some exchange between inside and outside waters. Hart (1943) observed that fish tagged in the vicinity of the Strait of Juan de Fuca and Sey- mour Narrows made more extensive migrations than Table 5 Time span between date of tagging and date of ret apture of tagged Iingcod with known date of recapture. Distance between No, Time span (days) release and recapture recaptured Mean Range <8.1 km 293 224.91 18-1197 8.1-.50 km 46 251.85 36-1122 >.S0 km 24 306.63 78-770 Total 393* 237.61 6-1197 * Includes 30 fish with unkniiwii ilistun ce traveliM . fish tagged in other adjacent inside waters. Of 1993 fish released during 1939-43, 209 were recovered of which 34 (16%) traveled distances >8.1 km (5 miles). For 342 recoveries from tags released in the Strait of Georgia during 1943-54, 73 (21.3'Fn) were recaptured within 1.6 km (1 mile) and 32 (9.3%) were recaptured >8.1 km from the point of release. Of those recaptured >8.1 km, the average time at liberty was 3 years and the net movement was northwesterly within the Strait (Chatwin 1956). Mathews and LaRiviere (1987) re- ported that of 1692 Iingcod tagged during 1976-81 in the eastern Strait of Juan de Fuca and in the vicinity of the San Juan Islands, 74 (49.6%) of 149 fish re- covered moved >8.1 km and were judged to be migra- tory. Most recaptures were south or west of the tag- ging site; the predominant pattern of movement was south and west through the Strait of Juan de Fuca. Fish tagged in the eastern Strait of Juan de P\ica migrated more than fish tagged in the San Juan Islands. Five recaptures were reported from the Pacific Ocean. The longest movement to the northeast was to Porlier Pass, British Columbia; the longest movement to the south- west was off Newport, Oregon. Previous offshore tagging studies have reported some movement between the various offshore fishing banks. Reeves (1966) reported that of 437 tagged on La Perouse (Forty Mile) Bank in June of 1960, 284 NOTE Jagielo Movement of Ophiodon elongatus at Neah Bay, Washington !I9 were recovered of which 74% were captured in the area of release, 21% had uncertain recovery location, and 5% were recaptured away from Forty Mile Bank (as far as Cape Flattery to the south and Ucluelet-Barkley Sound to the north.) The majority of all recaptures (82%) occurred within a 6-week period following the release because of an intensive trawl fishery for lingcod in the vicinity of tagging on Forty Mile Bank. Forrester (1973) reported the release of 2000 tagged lingcod on the Lennard Island trawling grounds in September of 1964; 535 were recovered with known locations of which 92 (17.2%i) moved from the tagging site. Most of the fish recovered away from the tagging site were recaptured on Big Bank (southern La Perouse Bank) to the south. Fish were recaptured from as far as Cape Russell to the north and as far as Cape Flattery to the south. Most recaptures occurred in the summer months of the year following tagging (A.J. Cass, Pac. Biol. Stn., Dep. Fish. Oceans, Nanaimo, BC, Canada, pers. com- mun.). Jack Robinson (Oreg. Dep. Fish. Wildl., New- port, OR, pers. commun.) reported the tagging of 3800 lingcod in offshore waters near Newport, Oregon in July of 1978. Within 17 months, approximately 10% of the tagged fish were recovered of which approx- imately 9% were recovered away from the area of tag- ging. Mathews and LaRiviere (1987) reported the results of H. Horton from 522 lingcod tagged off Depot Bay on the central Oregon coast during June 1978- January 1982. Of 19 recaptures reported through 1985, 10 had not moved significantly and 9 (47%) had migrated more than 10 km. Of those that migrated, 2 went a distance of more than 100 km. Chatwin (1956) reported evidence of homing behavior in Strait of Georgia lingcod. Of 14 lingcod captured at Entrance Island and transported 9.7 km (6 miles) to Departure Bay (Hart 1943), 4 were subsequently recap- tured at Entrance Island, and one at Newcastle Island (between Entrance Island and Departure Bay) within 2 years of release. Buckley et al. (1984) reported evi- dence of homing behavior in Strait of Juan de Fuca lingcod. Of 187 adult lingcod transferred from the eastern Strait of Juan de Fuca to Pulali Point in Hood Canal in 1978, 9 recoveries all were recaptured at distances >8.1 km from the release site. Of the 9, 7 were recaptured northward in the direction of their original capture site. My results at Neah Bay show more lingcod migratory behavior than most of the previous studies, but less than that reported in the eastern Strait of Juan de Fuca (Mathews and LaRiviere 1987). I found 70 of 363 recap- tures (19%) to be migratory. Of the 70 that migrated, 24 (34%) moved in excess of 50 km. Mathews and LaRiviere (1987) reported 74 of 149 recaptures (50%) to be migratory, of which 13 (18'^i) moved over 50 km. The difference in percent migratory could lie due to differences in exploitation rates. The Neah Bay tag- ging was conducted in March and April, and most tags were recaptured in the spring and summer months immediately following tagging. Most of the tags came from the intense sportfishery operating in the vicinity of Neah Bay, which may have removed potential mi- grants. Most of those that moved over 50 km escaped the sportfisheries and were recaptured on trawl gi'ounds offshore. In the eastern Strait of Juan de Fuca, tagging was conducted through May and most of the recaptures occurred the year following tagging by Canadian trawlers on Constance Bank (Mathews and LaRiviere 1987). These fish were probably not subject to the same recreational fishing pressure in the vicinity of tagging, and have had a greater opportunity to mi- grate. The difference in the relative proportion of fish moving > 50 km could be due to the distances from re- lease sites to trawling sites; Constance bank is 18 km west of Middle Bank, where much of the eastern Juan de Fuca tagging occurred, while most of the coastal trawling occurs over 50 km from the Neah Bay study area. The predominance of male fish tagged at Neah Bay can be explained by the different bathymetric distribu- tion of the two sexes. Others have reported that lingcod are distributed by depth according to sex and size; larger fish (mainly females) inhabit deep banks or reefs, while smaller fish (typically males) inhabit the shallower reefs nearshore (Chatwin 1956, Forrester 1973, Miller and Geibel 1973, Cass et al. 1984). Mathews and LaRiviere (1987) noted a similarly skewed sex ratio for fish tagged nearshore in the eastern Strait of Juan de Fuca. While the results of this and previous nearshore tag- ging studies give evidence of nearshore to offshore movement, a coherent pattern is not evident and a reliable working model of coastal lingcod migration is not yet available. Migratory recaptures from the pres- ent study were tyj^ically larger at the time of release than nonmigratory fish, suggesting a size threshold for movement; however, Mathews and LaRiviere (1987) failed to show a relationship between size at release and migratory tendency, and Hart (1943) concluded that large lingcod move less than small lingcod and that "some but not more than 5% of lingcod are more or less migratory during each year." Since the tagging at Neah Bay occurred nearshore in a narrow depth range (15-25 m) where the relative abundance of females is low, the effect of sex and size on lingcod movement reported here is likely biased with reference to the population as a whole. This depth-related bias may explain the discrepancies between this and other studies with regard to lingcod movement. Some level of female movement from offshore to nearshore areas for spawning is implied by the high 820 Fishery Bulletin 88(4), 1990 relative abundance of females at depth, though the na- ture and extent of such spawning migrations is poorly understood. It is unclear, for instance, whether females spawning nearshore reside nearby at depth and make vertical seasonal migrations to spawn, or whether fe- male lingcod migrate seasonally from the deep offshore trawling grounds to spawn nearshore. Such movements have been widely accepted as fact, though tagging ex- periments to date have failed to confirm a seasonal mass spawning migration (Miller and Geibel 1973). An appropriate tagging study design to model coastal lingcod movement will require tagging in both near- shore and offshore areas and the estimation of both the probability of survival as well as the probability of cap- ture across the time-area strata. To separate movement from survival, a minimum of three samples is needed. Iwao (1963) and Arnason (1972) gave models under this scenario, but only for the case with multiple recaptures. These approaches are not applicable to most fisheries tagging studies in which individuals are recaptured once by a commercial or recreational fishery and recovered dead. Potentially more applicable to fisheries tagging studies, Schwarz (1988) and Schwarz and Ar- nason (1990) have extended the traditional exploitation- based models of Brownie et al. (1985) to include tag recoveries over both time and space, and Hilborn (1990) recently provided a general framework for analysis of movement and mortality which incorporates a popula- tion dynamics and movement model using a maximum- likelihood minimization approach. In conclusion, this study gives evidence that a compo- nent of the lingcod population at Neah Bay is exposed to fishing mortality from the offshore trawling fleet. Research is needed to yield quantitative estimates of lingcod mixing rates, stratified by sex and size, be- tween the nearshore and offshore fishing grounds. These mixing rates will be essential to establish mor- tality rates by sex and age for the population as a whole, to clarify the collective impact of the nearshore and offshore fisheries on the coastal lingcod population. Acknowledgments This research was supported by the Washington De- partment of Fisheries and the Federal Aid in Sport Fish Restoration (Wallop-Breaux) Program. Special thanks are due to Donald "Bud" Carleson, skipper of the commercial troUer P«m/ Revere, for his knowledge of lingcod habits and help in capturing lingcod for tagging. Citations .\rnason, A.N. 1972 Parameter estimates fVom mark-recapture ex[)eriments on two populations subject to migration and death. Res. Popul. Ecol. (Kyoto) 13:97-113. Brownie C, D.R. Anderson, K.P. Burnham. and D.S. Robson 1985 Statistical inference from band recovery data— A hand- book. 2d ed. U.S. Fish Wildl. Serv. Resourc. Publ, 1,%, 305 p. Buckley. R., G. Hueckel, B. Benson, S. Quinnell, and M. Canfield 1984 Enhancement research on lingcod (Ophiadoti ehmgatus) in Puget Sound. Wash. De|.. Fish. Prog. Rep. 216, 93 p. Cass, A.J., E. Cameron, and 1. Barber 1983 Lingcod tagging study off Southwest Vancouver Island, M/V Pacifir Euijie - .July 14-27. 1982. Can. Data Rep. Fish. A(iuat. Sci. 40r,. 84 p. Cass, A.J., R.J. Beamish, and M.S. Smith 1984 Study of the biology of lingcod off the west coast of Van- couver Island, MA^ Arctic Hamester, November 22-December 2. 1977. Can. Data Rep. Fish. Aquat Sci. 461. 73 p. Chatwin, B.M. 1956 Further results from tagging e.\periments on lingcod. Fish Res. Board Can.. Prog. Rep. Pac. Coast Stn. 107:19-21. 1958 Morality rates and estimates of theoretical yield in rela- tion to minimum commercial size of lingcod (Ophiuilori elongatus) from the Strait of Georgia, British Columbia. .J. Fish. Res. Board Can. 15:831-849. Forrester, C.R. 1973 The lingcod {Oph union cloniinlux) in waters off Western Canada. Fish. Res. Board Can. Manuscr. Rep. Ser. 1266. Hart, J.L. 1943 Migration of lingcod. Fish Res. Board Can., Prog. Rep. Pac. Coast Stn. 57:3-7. Hilborn, R. 1990 Determination of fish movement patterns from tag recoveries using ma.ximum likelihood estimators. Can. J. Fish. Aquat. Sci. 47:635-643. Iwao, S. 1963 One method for estimating the rate of population inter- change between two areas. Res. Popul. Ecol. (Kyoto) 5:44-50. Jagielo. T.J. 1989 Washington groundfish fisheries and associated investiga- tions in 1988. Washington contribution; thirtieth annual meeting of the Canada/United States (Jroundfish committee, June 6-8, 1989. Nanaimo, British C'olumbia. Mathews. S.B., and M. LaRiviere 1987 Movement of tagged lingcod, Ophiodon elonyatus, in the Pacific Northwest. Fish. Bull.. U.S. 85:153-159. Miller, D.J., and J.J. Geibel 1973 Summary of blue rockfish and lingcod life histories; a reef ecology study; and giant kelp. Murrocyfitia pyrifera, experi- ments in Monterey Bay, California. Calif. Dep. Fish Game, Fish Bull. 1.58:1-137. Reeves, J.E. 1966 An estimate of survival, mortality, and of the number of lingcod (Ophiodon elongatus) off the southwest coast of Van- couver Island, British Columbia. Wash. Dep. Fish., Fish. Res. Pap. 2(24):55-66. Schwarz, C.J. 1988 Post-release stratification and migration models in band- recovery and capture-recapture models. Ph.D. diss., Univ. Manitoba, 384 p. Schwarz, C.J., and A.N. Arnason 1990 Use of tag-recovery information in migration and move- ment studies. Proc, American Fisheries Society workshop on tagging (In press). Zar, J.H. 1974 Biostatistical analysis. Prentice-Hall Biol. Sci. Ser., Englewood Cliffs, NJ, 620 p. Captive Tunas in a Tropical Marine Research Laboratory: Growth of Late-larval and Early-juvenile Black Skipjack Euthynnus lineatus Robert J. Olson Inter-American Tropical Tuna Commission c/o Scripps Institution of Oceanography, La Jolla, California 92093 Vernon P. Scholey Inter-Amencan Tropical Tuna Commission, Achotines Laboratory Las Tablas. Los Santos Province, Republic of Panama Little is known about the biology of tunas during larval and early-juve- nile stages because they are rela- tively inaccessible to scientists. In the eastern Pacific Ocean fishermen seldom catch juveniles of less than about 30 cm in length. Concurrent laboratory and field studies of tuna growth and mortality are important to gain insight into the recruitment process. These considerations moti- vated the Inter-American Tropical Tuna Commission (lATTC) to estab- lish a research center at Achotines 7»20 Achotines Bay Fraites del Norte Frailes del Sur, ecis' 30' 25' 7»20' Figure 1 Location of the study site (cross- hatched) south of Achotines Bay, Panama. The lATTC's Achotines Laboratory is situ- ated on the east side of the Bay. Bay in the Republic of Panama, a site located near tuna spawning grounds (Fig. 1). Black skipjack tuna Euthynnus lineatus are not commercially im- portant; however, their similarity to other tunas makes them valuable subjects of study. Their distribution is limited to tropical and subtropical regions of the eastern Pacific Ocean (Collette and Nauen 1983), with two stray specimens reported from the Hawaiian Islands (Matsumoto 1976). For many years, E. lineatus was thought to inhabit only coastal waters and waters around islands (Calkins and Klawe 1963, Yoshida 1979). However, recent fishing rec- ords show that black skipjack also occur in oceanic habitat (Schaefer 1987, fig. 1; Bayliff 1988a, fig. 62). Clemens (1956) reported the only previous information on rearing and growth rates of early-juvenile black skipjack. Some information was reported by Peterson (1983:54) on the growth of 32-51 cm fork length (FL) black skipjack in the field, based on tagging studies and length-frequency modal progression analysis. Houde and Richards (1969) described the growth of larval little tunny E. alletteratus reared in the laboratory from planktonic eggs. The larvae were fed unspecified quantities of mostly copepod nauplii and copepodites, and grew from less than 3.0 mm total length at hatching to almost 8.5 mm in 18 days, a rate of about 0.3 mm/day. No data are available on larval growth of kawakawa E. affinis (Yoshida 1979). The purpose of this note is to report the establishment of the Achotines Laboratory in Panama, to describe sampling results and rearing procedures of late-larval and early-juvenile black skipjack tuna, and to report growth experi- ments on captive black skipjack. Manuscript accepted 23 May 1990. Fishery Bulletin. U.S. 88:821-828. 821 822 Fishery Bulletin 88(4), 1990 Laboratory studies of growth are important in light of evidence in other fishes that variability in growth rates in the wild can affect early-life-stage duration, poten- tially resulting in large changes in larval survival and subsequent recruitment (Houde 1987). The Achotlnes Laboratory and study area The lATTC's Achotines Laboratory is located on a broad south-facing headland on the southern tip of the Azuero Peninsula in the Los Santos Province of the Republic of Panama (Fig. 1 ). This site is located at the northwestern part of the Panama Bight, which extends from Panama to Ecuador (Forsbergh 1969). The Pana- ma Bight is a region with large seasonal variations in atmospheric and oceanic characteristics, in part influ- enced by the seasonal position of the intertropical con- vergence zone (ITCZ). Sea-surface temperature data (Fig. 2) clearly show an upwelling cycle driven by the Caribbean trade winds when the ITCZ is displaced to the south, t.\"[Dically beginning in December and lasting at least through April. The continental shelf is narrow, 2-30 km in width, off the Azuei'o Peninsula, l)ut widens abruptly on either side of the Peninsula. Thus, oceanic habitat, thought to be a reijuirement for tuna spawn- ing, occurs close to shore. Scombrid larvae and early juveniles are routinely captured as close as 9 km from Achotines Bay. The following late-larval and early- juvenile scoml>rids have been captured and held in the laboratory with varying degrees of success: black skip- jack tuna, yellowfin and/or bigeye tuna {Thioinus nlbd- cnres and/or T. ohofusy, frigate and/or luillet tuna (Auxis thnzard and/or A. rochei)-. sierra Scotnhero- morus sierra, chub mackerel Scomber japonicus, and Indo-Pacific bonito Sorda orientalis. Materials and methods Larval and early-juvenile fishes were attracted to a bright underwater light at night (nightlighting) in the vicinity of the 100- and 20')-m isobaths (Fig. 1) during 'Early -juvenile T. nlhacarex and T. ohesiia cannot he distinguished on the basis of meristic. morphological, pigmentation (Matsumoto et al. 1972). or osteological characters (Potthuff 1974). However, an electrophoretic distinction between yellowfin and bigeye adults pro- vides a means of separating the larvae and earlv juveniles (Graves et al. 1988). -Larval and early-juvenile Auxit: thnzard and A. rochei have been distinguished by minor differences in pigmentation and body depth, but identifications are ambiguous (Uchida 1981). dill raker counts can be used for identification of juveniles >2."> mmSL, Imt gill rakers are too tiny and difficult to count in smaller specimens (Uchida 1981). Figure 2 Sea-surface temperatures taken while nightlighting in the sampling areas (Fig. 1) off the Achotines Laboratory. October 1986-April 1988. A 24-volt DC 300-watt light was lowered from a drifting boat to a depth of about 14 m, left for about 5 minutes, and slowly raised to a depth of 1-2 m. Late-larval and early-juvenile'' black skipjack tuna approaching the light were collected by dipnet and quickly placed into 61 x 76 cm polyethylene bags with rounded corners containing aerated sea- water. A water conditioner, Fritz-guard, was added to the water to minimize damage to the fish from mucous loss caused by abrasion. Time from capture to arrival at the laboratory ranged between 1 and 3 hours. Sea- surface temperature, weather, and sea conditions were recorded. The fish which did not survive nightlighting and transfer procedures were measured (standard length, SL, to nearest 0.1 mm) and weighed (round wet weight to nearest 0.001 g) soon after capture. At the laboratory, the captive fish were placed in 1.2-m diameter circular fiberglass tanks containing 0.3 m" of aerated seawater with a weak current. Water 'Late larvae are defined as the "postflexion larvae" and "transfor- mation larvae." and early juveniles as the "pelagic or special juveniles" of Kendall et al. (1984). Reference to trade names does not imply endorsement by the National Marine Fisheries Service. N().'\;\. NOTE Olson and Scholey Growth of late-larval and early-juvenile Euthynnus /meatus 823 from Achotines Bay was pumped through a sand filter containing #20 silica sand and added to the tanks at a rate sufficient to exchange the water several times per day. Wild zooplankton, predominantly copepods, caught with a 505-Mm mesh net were provided as food in den- sities of approximately 50-150/L two or three times per day until the black skipjack grew to about 13-17 mmSL. Then, whole Poec(7/o latipixHd and Poeciliopsis turrubarensis fry or chopped fish of several species were provided four to six times per day. The black skip- jack were fed until satiated. Maximum rations were provided throughout the growth experiments. When the fish reached about 35-45 mmSL, they were trans- ferred into 3.0-, 4.6-, or 6.4-m diameter plastic-lined pools containing 2.3, 11.4, or 24.6 nr' of water. Water temperature in the laboratory aquaria was recorded to the nearest 0.1 °C several times per day with a mercury thermometer. Salinity was measured to the nearest l"/"u (readability) several times per month with an optical salinometer. Daylight illumina- tion was supplemented by fluorescent lights over the containers. At night, low levels of indirect fluorescent lighting were maintained to prevent the fish from col- liding with the container walls. Because black skipjack are delicate, initial lengths of those held for growth experiments were estimated by visually comparing them with a ruler held near the tank and adjusting for the magnifying effect of the water. Lengths of fish that had recently died or were near death in the tanks soon after capture were estimated in the same way, and subsecjuently measured (SL to nearest mm) after retrieval from the tanks to improve estimating skills and to provide a measure of estima- tion error. The maximum error measured was used to calculate the largest potential errors in growth rates that could have resulted. Weights were calculated from lengths based on a weight-length relationship derived from 184 black skip- jack measured (SL to nearest 0.1 mm) and weighed (round wet weight to nearest 0.001 g) soon after cap- ture. Length and weight measurements were con- verted to natural logarithms (In), and regression parameters Ina and h were estimated by the method of least scjuares. \nW = Ina + bh^SL, (1) where W = weight in g, and SL = standard length in mm. Validity of the assumptions of linear regression were tested using residual analyses (Draper and Smith 1981). Standard lengths and weights at the end of the growth experiments were measured to the nearest 0.1 mm and 0.1 g, respectively. The experiments were I m L. O to a. n = 103 "1 1 10 15 STANDARD LENGTH AT CAPTURE (MM) Figure 3 Standard leiigth.s of late-larval and early-juvenile black skipjack tuna measured soon after capture by nightlighting in the waters south of the Achotines Laboratory, October 1987-April 1988. terminated when the fish died from some accidental or unknown cause or were near death. The average daily growth rate of each fish over the duration of captivity was expressed as growth in SL (mm) during captivity divided by days in captivity. The average daily weight increase of each fish was calcu- lated using final weights of the fish and the estimated SL at capture converted to weight. Results Late-larval and early-juvenile scombrids have been captured in all months of the year, but not in every month since the inception of routine sampling in 1984. During the period of October 1986-April 1988, 212 late-larval or early-juvenile black skipjack tuna were captured and transferred to the laboratory. Speci- mens which did not survive live transfer procedures ranged between 7.1 and 18.4 mmSL soon after capture (Fig. 3). A total of 79 (37%) survived handling and liv- ed longer than 48 hours. Sampling dates, numbers caught, and other information are given in Table 1. Young black skipjack were caught during 16 of the 18 months in which sampling took place during this period. No nightlighting took place during January 1987 due to poor weather conditions. The greatest catches were made during November and December 1986, December 1987, and January 1988. Sampling fre- quency was governed by weather conditions, boat availability, and other factors. At times, sampling was terminated prematurely due to poor weather, which resulted in no catch. For this and other reasons, the 824 Fishery Bulletin 88(4), 1990 Table 1 Sampling information, numbers of late-larval and/or early -juvenile black ski jjack which survived at least 48 hours after transfer to laboratory aquaria and standard lengths of those which did not survive nightlighting and transfer procedures. Accurate measurements were not obtained from all those which did not survive. Sea- surface temp. No. Standard length at capture No. sur- Sea- surface temp. No. Standard length at capture No. sur- Date (°C) caught Mean Range SD n viving Date (°C) caught Mean Range SD n viving 1986 1987 (continued) 8 Oct. 28.5 7 15.8 13.6-17.0 1.279 5 2 14 July 28.7 3 — — — — 2 9 Oct. 28.3 0 _ _ _ — — 23 July 28.4 5 12.8 11.5-14.0 1.770 2 0 U Oct. 28.0 0 _ _ _ - — 10 Aug. 28.6 0 — - - — - 28 Oct. 28.4 17 11.4 9.4-15.3 1.631 16 1 11 Aug. 28.9 0 — — — — — 29 Oct. 28.4 10 11.8 10.2-12.8 1.124 5 5 18 Aug. 28.4 0 — — — — - 31 Oct. 28.4 0 _ _ _ - — 20 Aug. 28.8 4 - — — — 4 5 Nov. 28.4 0 _ _ — - — 27 Aug. 28.6 0 — — — — — 8 Nov. 28.4 0 _ _ _ — — 31 Aug. 28.5 0 — — — - - 19 Nov. 28.0 11 9.3 7.7-11.1 1.(147 8 o 15 Sep. 28.5 3 10.5 — — 1 o 20 Nov. 28.0 o _ _ _ — 1 17 Sep. 28.8 3 16.4 14.4-18.4 2.830 •> 1 26 Nov. 28.2 9 _ _ _ — 4 23 Sep. 29.0 6 8.1 — - 1 1 27 Nov. 28.4 4 _ _ _ — 0 2 Oct. 28.4 7 12.7 11.2-13.7 1.162 4 3 4 Dec. 28.1 3 11.5 9.4-13.2 1.930 3 0 15 Oct. 28.9 0 — — — — — 18 Dec. 27.6 0 10.5 10.0-11.0 0.707 o 0 23 Oct. 28.7 1 — — — — 1 22 Dec. 27.5 1 _ _ _ — 0 26 Oct. 28.5 1 - — — - 0 29 Dec. 27.5 0 _ _ _ — — 9 Nov. 28.6 4 11.5 10.0-13.7 1.565 4 0 1987 6 Feb. 16 Feb. 17 Feb. 18 Feb. 9 Mar. 12 Mar. 16 Nov. 28.2 3 — — — — 3 25.5 25.5 25.3 27.4 24.5 0 0 4 0 0 0 9.8 7.7-11.9 2.970 o 2 17 Nov. 23 Nov. 9 Dec. 15 Dec. 16 Dec. 22 Dec. 28.5 28.0 28.4 28.5 28.6 28.5 1 0 37 8 4 4 8.6 10.3 8.0 7.1-9.5 0.728 9.5-11.7 1.217 7.9-8.0 0.071 24 3 2 1 5 5 4 2 25 Mar. 27.5 't 11.9 - - 1 1 1988 26 Mar. 27.1 0 _ _ _ — — 1 Jan. 28.0 0 — — — — — 30 Mar. 26.2 0 _ _ _ — — 13 Jan. 27.6 (1 — — — - — 3 Afir. — 0 _ _ _ — — 18 Jan. 27.0 0 - - - - — 8 Apr. 27.0 {) _ _ _ — — 20 Jan. 27.3 36 11.3 9.0-14.4 1.717 13 23 9 Apr. 26.4 0 _ _ _ — — 25 Jan. 27.3 0 — — — — — 1 May 28.4 0 — — — — — 10 Feb. 26.0 1 12.2 — — I 0 4 May 29.0 2 9.4 1 1 8 Mar. 23.4 0 — — — — — 6 May 29.0 0 _ _ _ — — 10 Mar. 24.4 ■> 10.9 - — 1 1 1 .lune 28.9 0 _ _ _ — — 16 Mar. 23.9 0 — — — — — 3 June 28.5 0 _ _ _ — — 30 Mar. 23.5 0 — — — — — 4 June 28.5 0 _ _ _ — — 5 Apr. 25.5 0 — — — - - 25 June 29.7 0 _ _ _ — — 11 Apr. 27.1 1 9.2 — — 1 (1 26 June 29.3 0 _ _ _ — — 13 Apr. 26.4 1 — — — - (1 2 July 29.5 0 _ _ _ — — 26 Apr. 26.6 0 — — — — — 7 July 29.0 3 12.0 1 2 28 Apr. 27.3 0 — — — - - 9 .luly 28.8 0 T..l:,| 212 103 79 capture frequency of black skipjack in Tal)ie 1 is not meant to reflect the true spawning frequency of the afhilts. Thirty-nine specimens were held for laboratory growth experiments. Their estimated lengths and weights at capture are shown in Figure 4. The larger fish in the catch (Fig. 3) survived capture and transfer in greater proportions than the smaller ones (Fig. 4). The captive fish survived an average of 36 days in the laboratory; most (64%) died in 30 days or less. Three fish survived in excess of 130 days. The longest-lived black skipjack grew in captivity for 167 days, and attained an SL of 259 mm and weight of 336 g. It was sacrificed when it ceased feeding due to eye infec- NOTE Olson and Scholey Growth of late-larval and early-juvenile Eu:hynnus hneatus 825 12 10 n^39 1 - 8 - - 6 - - 4 I in 2 8 10 12 14 16 18 ESTIMATED STANDARD LENGTH AT CAPTURE (K*i) 0.024 0-040 0.056 0.072 0.088 ESTIMATED WEIGHT AT CAPTURE (G) Figure 4 Standard lengths estimated in tlie lalxiratory and weiglits calculated fr(ini a weight-length regression (Fig. 6) for 39 black skipjack tuna held for growth experiments. tions. Water temperature and salinity in the laboratory aquaria ranged from 23.7 to 29. TC and 29 to 34"A.o, respectively, over all the experiments. Growth in length Captive black skipjack fed ml libitum grew in a curvi- linear relationship of SL with time (Fig. 5a). The data were not fitted to derive a predictive growth equation because they are inadequate for that purpose. The fish were fed to satiation, but the rations were not measured. The highest rates of growth in length were attained during the first month in captivity (Fig. 5b). After about 4 weeks, the fish had progressively lower aver- age growth rates. Black skipjack from experiments that terminated during the first 15 days of captivity grew at extremely variable rates, from 1.0 to 4.8 mm/day. After 15 days, there was a significant nega- tive correlation between growth rate and days in cap- tivity (r = -0.877, n = 23, P«0.001). All the fish that survived between 15 and 50 days grew rapidly, 3.2-4.8 mm/day. After about 50 days in captivity average growth rates declined drastically. Errors in estimating lengths of newly-captured fish that had recently died or were near death in the tanks soon after capture ranged up to ± 4 mm, although most a n-39 b n=39 T " s s S I 200 - I C3 O 3 Z Z ^ 150 - UJ cc K z 5 o " 50 ■ 0 d 1 o °0 30 60 90 120 150 IG 30 60 90 120 150 leo 450 c ■n*37 d n-37 400 4 5 300 " 3 I 250 I CD O ^ 200 _ * p z 150 ■ I 100 O 1 (r 50 ■r'' °0 30 60 90 120 150 160 DAYS IN 0 30 CAPTIVITY 60 90 120 150 ISO Figure 5 Standard lengths (a), growth rates in standard length averaged over the duration of captivity (b), weights (c), and growth rates in weight averaged over the duration of captivity (d) of juvenile black skip- jack tuna at the end of laboratory growth experiments versus days in captivity. Two fish held 1 1 days were not weighed. were less. Maximum potential errors in growth rates were calculated based on the assumption that ± 4-mm errors were made when estimating capture lengths of all the experimental fish. In general, the importance of measurement error diminished with increasing time in captivity. Potential errors ranged from only ±0.8 to a high of ±53.3% for one fish held the least time of all experimental fish (7.5 days). The next greatest error estimates were -26.1 and -i-23.8%, and the means were ±8.8%. Potential errors for the majority (85%) of the samples were ± 13.5% or less, and less than 10% for all fish held in captivity more than 15 days (n = 2b). Growth in weight Growth in weight appeared to follow a sigmoidal rela- tionship with time in captivity (Fig. 5c). The rate of weight gain during the first 30 days was low, increas- ing dramatically between about 30 and 70 days, and then tapering off after 70 days. While the greatest and most variable growth rates in length took place during the first 30 days, average rates of weight gain during that period were less than 826 Fishery Bulletin 88(4). 1990 I- g 0 08 W= 0.000031 SL' n = 184 10 15 STANDARD LENGTH (MM) Figure 6 Weights (W) and standard lengths (SL) cif 184 late-larval and early- juvenile black skipjack tuna measured soon after capture. The power function of the log-log fitted regression line and 95% predic- tion belts of weight on length are included. The r- of the log-log fit was 0.927, the 95% confidence limits of the regression coefficient were 2.6641-2.8917 g/mni, and the standard error of the estimate was 0.1790. those for fish held longer (Fig. 5d). The greatest aver- age growth rates in weight (3.9-4.0 g/day) were re- corded for fish surviving 60-73 days {n = 3). After that time, average daily weight increases declined dras- tically (ri = 4). Weight-length relationship Capture weights of the live black skipjack held for growth experiments (Fig. 4) were estimated using a weight-length regression equation based on fresh mea- surements and weights of 184 other black skipjack ranging from 5.6 to 19.7 mmSL and 0.004 to 0.132 g (Fig. 6). The power function of the fitted regression is lnU'= -10.3806 -H 2. 77791aSL. The standard er- rors of parameters Ina and h are 0.1326 and 0.0577, respectively. Tests for normality of the residuals (Filliben 1975), independence or lack of autocorrelation of the residuals (Durbin-Watson statistic), and a constant variance (homoscedasticity) of the residuals (Wesolowski 1976) failed to indicate any violation of the assumptions of linear regression. Effect of temperature Sea-surface temperatures during each sampling trip are shown in Figure 2. The data show a gradual decline in temperatures commencing in November or Decem- ber, reaching lows of about 23-25 °C in March, followed by a gradual warming to maximum stable temperatures of about 28-29°C from April or May until October or November. Black skipjack spanning the entire size range encountered (7.1-18.4 mmSL) were taken when the temperature ranged between 28.0 and 29.0°C. Only smaller individuals (7.7-12.2 mmSL) were caught when sea-surface temperatures were below 27.0°C, but the sample size was low (n = 4). Neither the growth rates nor final sizes attained in captivity were significantly correlated with mean water temperatures in the aquaria over the course of the ex- periments. However, the final lengths (r= -0.559, P< 0.001) and weights attained (r = -0.572, P< 0.001), average daily growth in weight (/• = -0.503, O.OOK P<0.002), and the number of days in captivity (r = -0.568, P< 0.001) were negatively correlated with minimum temperatures recorded in the laboratory. No such relationships were observed with the maximum temperatures recorded. Discussion Except for the black skipjack held in shipboard aquaria by Clemens (1956), these experiments are the first in whicli late-larval or early-juvenile tunas were collected in the wild and reared in captivity for experimental growth studies. Similar experiments on other scom- l)rids are being conducted at the Achotines Laboratory. To our knowledge, 167 days is the longest time any scombrid has been held in captivity beginning at an early life stage. Harada et al. (1973) held .4((.r('.s- tapei- nosomn ( = A. rochei) larvae for 52 days, Harada et al. (1974) reared Sarda orientalh larvae up to 99 days, and Harada et al. (1980) grew Thunnus nlbacares larvae for a maximum of 38 days. These larvae were all hatched from artificially-fertilized eggs obtained from ripe females, and fed unspecified rations of zoo- plankton and fish larvae. Previous to our studies, young E. lineatus were reared on only one occasion. Clemens (1956) collected black skipjack by dipnet at night, and held 10 individuals averaging about 27 mmSL for up to 12 days in shipboard aquaria. Average growth in length was 3.1-3.6 mm/day. The paper does not state at what water temperature the fish were held nor how much food was provided during the experiment. The growth rates measured in Harada et al.'s (1973, 1974, 1980) experiments were 3.0, 3.0, and 1.3 mm/days, respectively. The growth rates reported by Clemens NOTE Olson and Scholey Growth of late-larval and early-juvenile Euthynnus line^tus 827 (1956) and Harada et al. (1973, 1974, 1980) are within the range of growth rates reported here, although we obtained greater rates too, up to 4.8 mm/day. Clemens (1956) stated that his results probably approximate minimum growth rates in nature. An undesirable aspect of our study was the neces- sity of estimating, rather than measuring, lengths at the time of capture. This may have contributed to the large variance in the data reported for the first 15 days (Fig. 5b). Estimation errors were minimal, however, because the second author's ability to estimate lengths of live fish was continually refined by estimating, then measuring the fish that died in the tanks soon after cap- ture. Except for a few fish held for short times, max- imum estimation errors of ± 4 mm translated to low potential error in growth rates. Despite unavoidable estimation errors, this study reveals some interesting aspects of growth of young tunas. Black skipjack are capable of growing at high but variable rates when food is plentiful. The growth rates we measured might be considered upper limits for this species; feeding to satiation in the wild would likely be detrimental to survival because laboratory observations suggest that the added weight of food in the gut inhibits mobility. The variability in rates mea- sured during the early part of the experiments may be due in part to the stress of captivity. However, high, variable growth rates suggest a large scope for gi'owth (Brett 1979), an advantageous characteristic for fish that spend their early life stages in the epipelagic zone where predation risk is great. A large scope for growth permits rapid gi'owth when food is abundant, providing young tunas an earlier transition to piscivorous feeding and early formation of schools. The decline in growth rates with increasing time in captivity may be related in part to inadequate-sized rearing containers (Thei- lacker 1980). Further laboratory growth experiments should concentrate on maintaining constant, controlled rations. Growth rates of late juvenile and/or adult black skip- jack from the commercial catch in the eastern Pacific Ocean are much lower, as expected, than those re- ported here. Tagging and length-frequency modal progression analyses provided rates that agree with each other (Peterson 1983:54-55). Black skipjack measuring 32-45 cmFL grew about 0.36 mm/day. Larger fish, 45-51 cm, gi-ew more slowly, 0.22 mm/day. The upper seven data points in Figure 5a, correspond- ing to the largest fish (224-272 mmSL) which were held for the longest times (59.5-167.5 days), appear to represent a slower growth stanza than a previous stanza <60 days. A straight line fitted to these seven points yielded a significant slope of 0.28 mm/day, com- parable to the natural growth rates reported for 32- to 51 -cm black skipjack (Peterson 1983:54-55). jack tuna Katsuwonus pelamis, another primitive species of the tribe Thunnini, measuring 37.5-42.5 cmFL when tagged and released east of 100°W longi- tude grew an average of 1.22 mm/day in 31-180 days at liberty (Bayliff 1988b). Much additional work on growth of tunas under con- trolled conditions, as well as growth studies in nature, is needed to understand how physical and biotic vari- ability in the ocean affects growth rates and life-stage duration. By virtue of black skipjack's biological sim- ilarities to commercially important tunas, we believe that such studies utilizing black skipjack tuna will yield important information that would be applicable to other tunas. Acknowledgments We wish to thank the staff members of the Achotines Laboratory for their hard work and long hours. R.R. Lauth and R.C. Jope provided special assistance with weight-length and laboratory temperature data. We acknowledge the efforts of G. Schumann, B.M. Chat- win, and J.M. lanelli for much of the early development of the Achotines Laboratory. The Panamanian Minis- terio de Comercio e Industria, Departamento de Recur- sos Marinos, has provided continued support of the Achotines Laboratory. Valuable reviews were provided by W.H. Bayliff, W.L. Klawe, R.R. Lauth, D. Mar- gulies, K.M. Schaefer, and two anonymous reviewers. Citations Bayliff, W.H. 1988a (editor) Annual report of the Inter-American Tropical Tuna Commission, 1987. Inter-Am. Trop. Tuna Comm., 222 p. 1988b Growtli of skipjack, Katsuworms pelamis, and yellowfin. Thunnus albacares, tunas in the eastern Pacific Ocean, as estimated from tagging data. Inter-Am. Trop. Tuna Comm. Bull. 19:307-,385. Brett. J.R. 1979 Environmental factors and growth. /« Hoar, W.S., D.J. Randall, and .J.R. Brett (eds.), Fish physiology. Vol. VIII. Bioenergeties and growth, p. 599-67,5. Acad. Press, NY. Calkins, T.P.. and W.L. Klawe 1963 Synopsis of biological data on black skipjack, Euthynnus llueatus Kishinouye 1920. FAO Fish. Rep. 6:130-146. Clemens, H.B. 1956 Rearing larval soombrid fishes in shipboard aquaria. Calif. Fish Game 42:69-79. Collette, B.B., and C.E. Nauen 1983 F.-\<) species catalogue. Vol. 2. Scombrids of the world. An annotated and illustrated catalogue of tunas, mackerels, bonitos. and related species known to date. FAO Fish. SjTiop. 125. Vol. 2. 137 p. Draper, N.R.. and H. Smith 1981 Applied regression analysis. Wiley, NY. 709 p. 828 Fishery Bulletin 88(4), 1990 Filliben, J.J. Yoshida, H.O. 1975 The probability plot correlation coefficient test for 1979 Synopsisof biological data on tunas of the genus £"«?/!!/"- normality. Technometrics 17:111-117. nus. NOAA Tech. Rep. NMFS Circ. 429, .57 p. (FAO Fish. Forsbergh. E.D. Synop. 122). 1969 On the climatology, oceanography and fisheries of the Panama Bight. Inter- Am. Trop. Tuna Comm. Bull. 14:45-385. Graves, J.E., M.A. Simovich, and K.M. Schaefer 1988 Electrophoretic identification of early juvenile yellowfin tuna, Thunmis albacares. Fish. Bull., U.S. 86:835-838. Harada. T., O. Murata, and H. Furutani 1973 On the artificial fertilization and rearing of larvae in marusoda, Auj:is tapeinosomn. Kinki Univ. Fac. Agric. Bull. 6:113-11(;. Harada, T., O. Murata. and S. Miyashita 1974 On the artificial fertilization and rearing of larvae in bonito. Kinki Univ. Fac. Agric. Bull. 7:1-4. Harada. T., O. Murata. and S. Oda 1980 Rearing of and morphological changes in larvae and juveniles of yellowfin tuna. Kinki Univ. Fac. Agric. Bull. 13:33-36. [Engl, transl. no. 51 by T. Otsu, 1980, 8 p., avail. Honolulu Lab., Southwest Fish. Cent., Natl. Mar. Fish. Serv., NOAA, H(molulu, HI 96822-2396). Houde, E.D. 1987 Fish early life dynamics and recruitment variability. Am. Fish. Soc." Symp. "2:17-29. Houde. E.D., and W.J. Richards 1969 Rearing larval tunas in the laboratory. Commer. Fish. Rev. 31(12):32-34. Kendall. A.W. Jr.. E.H. Ahlstrom. and H.G. Moser 1984 Early life history stages of fishes and their characters. In Moser, H.G., et al. (eds.), Ontogeny and .systematics of fishes, p. 11-22. Spec. Publ. 1, Am. Soc. Ichthyol. Herpetol., ,\llen Press. Lawrence. KS. Matsumoto, W.M. 1976 Second record of black skipjack, Euthynnus lineatuf:, from the Hawaiian Islands. Fish. Bull., U.S. 74:207. Matsumoto, W.M., E.H. ,\hlstrom. S. Jones, W,L. Klawe, W.J. Richards, and S. Ueyanagi 1972 On the clarification of larval tuna identification particular- ly in the genus Thunnus. Fish. Bull., U.S. 70:1-12. Peterson, C.L. (editor) 1983 .^nnual report of the Inter- American Tropical Tuna Com- mission, 1982. Inter-Am. Trop. Tuna Comm., 294 p. Potthoff, T. 1974 Osteological development and variation in young tunas, genus Thiinnuf: (Pisces, Scombridae), from the Atlantic Ocean. Fish. Bull., U.S. 72:563-.588. Schaefer, K.M. 1987 Reproductive biology of black skipjack. Eiilhytinus lineatus, an eastern Pacific tuna. Inter-Am. Trop. Tuna Comm. Bull. 19:165-260. Theilacker. G.H. 1980 Rearing container size affects mor])hology and nutritional condition of larval jack mackerel, Triirhiinis xyniniftricun. Fish. Bull.. U.S. 78:789-791. Uchida. R.N. 1981 Synopsis of biological data on frigate tuna, Auxis thazard, and bullet tuna, A. rochei. NOAA Tech. Rep. NMFS Circ. 436, 63 p. (FAO Fish. Synop. 124). Wesolowski, G.O. 1976 Multiple regression and analysis of variance: An intnxluc- tion for computer users in management and economics. Wiley, NY. 292 p. Length-weight, Age and GroNA/th, and Landings Observations for Sheepshead Archosargus probatocephalus from North Carolina Frank J. Schwartz Institute of Marine Sciences, University of North Carolina Morehead City, North Carolina 28557 Sheepshead Archosargus jn-obato- cephahis range from Nova Scotia to Brazil (Robins and Ray 1986). Three subspecies have been recognized, based on the number and size of body bars: A. p. probatocephalus ranges along the Atlantic coast of the United States; A. p. oviceps, from St. Mark's, Florida to the Campeche Banks; and A. p. aries, from Belize to Babia de Sepetiba, Brazil (Caldwell 1958, 1965). Little is known concerning length-weight or age and growth relationships for this common edible fish. This paper presents the age and growth and length-weight relation- ships, reviews changes in historical catch for sheepshead from North Carolina, and resolves discrepancies in the literature concerning max- imum length and weight of this species (Hildebrand and Schroeder 1928, Hildebrand and Cable 1938, Bigelow and Schroeder 1953, Rob- ins and Ray 1986). Methods Most specimens were captured in 12-m otter trawls, 91-m gillnets, or by spear. Most specimens were ob- tained from two sites where sheeps- head are presently common in North Carolina: the Masonboro and Carolina Beach Inlets-Cape Fear River area of New Hanover and Brunswick counties, and Radio Island Jetty, Carteret County, ex- tending from Beaufort Inlet in the Atlantic Ocean to just south of Shackleford Banks and Cape Lookout Jetty. The largest speci- mens and the state record-sized fish were caught in the Carolina Beach area by hook-and-line fishermen as part of the state's fishing citation progi'am. Specimens were measiu-ed to the nearest millimeter standard (SL) and total length (TL) and weighed to the nearest gram, ex- cept for the tournament fish, which represented the upper end of the length-weight curve and were weighed to the nearest 114 g. Con- version from total length to stan- dard length, for study fish larger than 100 mmSL, was possible by the formula SL = 0.817TL, N= 240; for fish smaller than 100 mmSL the conversion formula was SL = 0.780 TL. The latter conversion was deter- mined by utilizing data for 412 yoiuig specimens (6-48 mm) measured by Hildebrand at Beaufort, North Carolina in 1914 (Hildebrand's field notes, Inst. Mar. Sci., Morehead City). Length-weight and standard length-scale radius relationships were calculated using log-log for- mulas where log(y) = a + b log (x), where x is either standard length or scale radius, measured from the focus to the outer lateral edge of the scale, and y equals weight. Scales were removed for ageing from just below the spinous/soft-ray dorsal-fin junction and the area above the lateral line. Scales were read at lOK magnification using a Baush and Lomb dissecting micro- scope. No validation of annulus for- mation was attempted considering the diverse availability of speci- mens. Therefore, the values I report are only presumed ages. Results The length-weight relationship for 282 sheepshead, measuring 9-591 mmSL (723 mniTL) and weighing 0.042-8370 g (18 lb, 7 oz), was cal- culated by the formula log (y) = - 4.5287 + 3.0446 log (x), r = 0.9929 (Fig. 1). A linear relationship between stan- dard length and scale radius was described for 68 fish measuring 31-525 mmSL by the formula log (y) = 0.8801 -H 0.820 log (x), r = 0.9789. Too few scales from speci- mens smaller than 30 mm were available for inclusion in the rela- tionship calculation. Scales of fish 17-400 mmSL or those to age 4 were easy to read. Scales of older and larger specimens were difficult to read as the focus often became opaque and thickened, thereby obscuring the first two annuli. Backcalculations of age and size from 50 of the best scales suggested seven age-classes (Table 1); how- ever, these did not agree with a simple length-frequency plot where eight age classes seemed to prevail (Fig. 2). Also the maximum back- calculated size was 482 mm, where- as the largest fish studied was 525 mmSL. This discrepancy strength- ened the observation that perhaps one or two annulae were obscured on scales of specimens larger than 400 mmSL, which were probably older than 8 years of age. Reference to trade names does not imply en- dorsement by the National Marine Fisheries Service, NOAA. Manuscript accepted 29 June 1990. Fishery Bulletin, t'.S. 88:829-832. 829 830 Fishery Bulletin 88(4), 1990 9988 95J4 9oec 8626 erT2 6156 i902 o- 5448 I 4994 i 4540 4086 3632 3178 2724 2270 1816 K'-A 5267 . 3 0446 Log t: N- 282 r - 0 9929 20 *:■ 6C 9C: (OC i2C' >40 I6C' 160 200 22C' 240 260 280 5CO 320 MO 360 380 400 420 4 Stondard Length (mm ) 3 460 460 500 520 540 560 580 600 620 640 660 Figure 1 Length- weight relationship for 282 sheeps- head captured in North Carolina. NC = largest North Carolina specimen; LA = world record Louisiana sheepshead. Table I Backcalculation of standard lengths at p ■esuined age for | sheepshead from North Carolina. Age N Yof Y 1 2 3 4 5 6 7 8 2 67 91 30 73 156 189 1 74 142 218 237 1 lOti 158 205 248 295 7 100 177 242 316 349 381 7 81 135 242 274 325 366 387 1 95 140 225 322 351 372 411 432 1 97 138 222 266 355 397 422 466 482 Weighted mean 50 79 152 207 289 336 Incre 375 nient 394 449 482 73 55 82 47 39 19 55 33 20- 19- n Z 10- H r^ / w fW^ U\[ C > 100 20O 500 400 500 600 STANDARD LENGTH Imm) Figure 2 Length-frequency histogram, in 10-mni units, of the 282 sheepshead captured in North Carolina. Discussion Only Ogburn (1984) has examined the length-weight of sheepshead from Masonboro Inlet Jetty, New Han- over County, NC. Her specimens measured 56-340 mmSL and weighed 4.3-1535 g, N = 45; data on five extremely small specimens, 56-63 mmSL, were ex- cluded from the length-weight relationship. However, recalculation of this relationship, using all data, yield- ed the equation log (y) = -4.6927 -t- 3.1309 log (x), /• = 0.9829. Superimposing that recalculation on Figure 1 indicated good agreement between her data and that reported here. Mook (1977) noted scales on 10-12 mm specimens. Johnson (1978) noted no scales for specimens of 12-mm lengths, but did depict them on a 17-mm specimen. North Carolina specimens smaller than 16 mm pos- sessed no scales but their outlines were present on 11-12 mmSL specimens. A variety of structures have been used to age fishes (Summerfelt and Hall 1987). The reason only scales were used to age sheepshead was that they were the only structures consistently available during this study, as the fish were obtained from many sources or could not be kept for age determination by vertebra, otoliths, etc. The largest sheepshead from nearby South Carolina were 513 mmSL (641,4 mmTL) weighing 6015,5 g, and 505 mmSL (625 mmTL) weighing 4900 g (D. Hammond and E. Wenner, S.C. Wild!. Mar. Res. Dep., Charleston, pers. commun., Jan. 1990). These data also fall within the length-weight curve plotted for North Carolina sheepshead (Fig. 1). NOTE Schwartz: Length-weight, age-growth, and landings of Archosargus probatocephalus 831 o o o I/) M 1 / I I I ,\ X- 1885 1890 I90O 1910 1920 1930 1940 1950 i960 1970 YEAR O O o I Figure 3 North Carolina catch landings for sheepshead, 1989. Dashed lines indicate missing year data. 1887- Like Hildebrand and Cable (1938) for Beaufort, NC, and Springer and Woodburn (1960) for Gulf of Mexico sheepshead, the smallest North Carolina sheepshead were also captured between May and October. Note the peculiar hiatus in the length frequency and length- weight curve (Figs. 1, 2) for sheepshead 90-150 mmSL. Absence of specimens within these size ranges may have been caused by their shifting from a seagrass habitat to piling, jetty, and other hard substrates preferred by larger young and adults (Hildebrand and Cable 1938, Johnson 1978). A search of major museums and taxidermist records has failed to uncover sheepshead that attain the size and weight (91 cm and 9-13.5 kg) reported in the literature (Hildebrand and Schroeder 1928, Hildebrand and Cable 1938, Bigelow and Schroeder 1953, Robins and Ray 1986). Even if 91 cm was an accurate total- length measurement, conversion to SL would make that specimen 743 mm, a size far larger than even the officially recognized world-record specimen from Loui- siana (probably A. p. oviceps), which was 596 mmSL and 730 mmTL, and weighed 9648 g (21 lb, 4 oz, and plotted in Figure 1) and would fall far outside the length-weight data depicted in Fig. 1. This is strong evidence that sheepshead do not attain the large sizes mentioned in the literature. In essence, future litera- ture should be emended to note that the maximum sizes of sheepshead in North Carolina, to date, are 591 mmSL, 723 mmTL, weighing 8370 g, and elsewhere 596 mmSL, 730 mmTL, weighing 9248 g. All larger sizes reported should remain or be considered erroneous. A dramatic shift in the commercial landings of sheepshead (mostly caught by haul seine) has occurred in North Carolina between" 1887 and 1989 (Fig. 3) (Chestnut and Davis 1975; Goode 1884; NC Div. Mar. Fish, statistical data, Morehead City). Yarrow noted (Smith 1907) that sheepshead were very abundant in 1871. Commercial catches prior to 1900 remained over 61500 kg. Catches between 1918 and 1981 remained low (27000-47000 kg, lowest in 1970, 675 kg). Only since 1981 has a recent surge been noted in the land- ings (Fig. 3), mostly in Carteret County. Hildebrand and Cable (1938) and Johnson (1978) noted that larval and juvenile sheepshead are usually found associated with seagrasses which they depend 832 Fishery Bulletin 88(4), 1990 upon for shelter and food. Whether the early and re- cent landings can be correlated with seagrass abun- dance remains unknown, for no early records of sea- grass abundance exist prior to and following the wasting disease of the 1930s (Orth and Moore 1981, Short et al. 1987). To date, the beds have seemingly not increased (G. Thayer, NMFS Beaufort Lab., pers. commun., Jan. 1990). It would have been beneficial to know whether the early-life-history stages that depend on vegetation for food and protection were or are in- creasing in relation to seagrass incidence and just how dependent they are on that habitat for their growth and survival. Acknowledgments Thanks are extended to Dr. D. Lindquist, UNC-Wil- mington, for access to specimens collected by Ogburn; to G. Ogburn for comments on her length-weight calculations; to K. Hartel of the Museum of Compara- tive Zoology' (Harvard) for a search of possible speci- mens reported by Bigelow; and to taxidermists Cher- finski and Swanson of NC, and Bill Alden Reese Tax- idermy, FL, for a search of their records for data on extremely large sheepshead. The International Game Fish Association, Fort Lauderdale, FL provided infor- mation on the world-record recognized sheepshead from Louisiana. C.R. Robins, Univ. Miami-RSMAS reviewed the manuscript and commented on large specimens. C. Manooch, NMFS Beaufort Lab., also reviewed the manuscript and provided helpful com- ments on age interpretation. D. Tootle and R. Finer of N.C. Division of Marine Fisheries, Morehead City, provided catch landings for years 1962-89, while Susanne Guthrie supplied size data for fish caught in North Carolina tournaments. G. Thayer, NOAA- NMFS, Beaufort, NC gave freely of his historical knowledge about seagrass abundances in NC. R. Barnes and H. Page of IMS produced the figures. B. Bright typed the manuscript. Citations Bigelow, H.B.. and W.C. Schroeder 1953 Fishes of the Gulf of Mexico. Fish. Bull. 74, 577 p. Caldwell, D.K. 1958 Notes on the barred pattern in the sheepshead, Ar- choyaryus jjrobatocephalus and .4. oince}>s. Q. J. Fla. Acad. Sci. 21(2):1,38-144. 1965 Systematics and variation in the sparid fish A rckvsaryus probalocephalus. Bull. South. Calif. Acad. Sci. 64(2):89-100. Chestnut, A.F.. and H. Davis 1975 Synopsis of marine fishery. UNC-SG-75-12, UNC Sea Grant Coll. Prog.. NC State Univ.. Raleigh, 425 p. Goode, G.B. (editor) 1884 The fisheries and fishery industries of the United States. Sec. 1 plates. U.S. Comm. Fish Fish.. Wash. D.C., 277 p. Hildebrand, S.F., and L.E. Cable 1938 Further notes on the development and life history of some teleosts at Beaufort, NC. Bull. U.S. Bur. Fish. 48(24):505-642. Hildebrand, S.F., and W.C. Schroeder 1928 Fishes of Chesapeake Bay. Bull. U.S. Bur. Fish. 43(1): 1-366. Johnson, G.D. 1978 Development of fishes of the Mid- Atlantic Bight, an atlas of eggs, larval, and juvenile stages. U.S. Fish Wildl. Serv. 4, Biol. Surv. Proj. FWS/OBS-78-12. 311 p. Mook, D. 1977 Larval and osteological development of the sheeps- head, Archosaryiis prohatorcjihalu.': (Pisces:Sparidae). ('opeia 1977(1):126-135. Ogburn, G. 1984 Feeding-ecology and the role of algae in the diet of .sheeps- head ArcliDsaryus probatoccphaluft (Pisces; Sparidae) on two North Carolina jetties. M.S. thesis, Univ. NC-Wilmington, 72 p. Orth, R.J., and K.A. Moore 1981 Submerged aquatic vegetation of the Chesapeake Bay: Past, present, and future. Trans. 46th N. .'Xm. Wildl. Nat. Resour. Conf., p. 271-283. Robins. C.R., and G.C. Ray 1986 A field guide to Atlantic coast fi.shes of North America. Houghton Miftlin, Boston, 354 p. Short, F.T., L.K. Muehlstein. and D. Porter 1987 Eelgrass wasting disease: cause and recurrence of a marine epidemic. Biol. Bull. (Woods Hole) 173(3):557-562. Smith. H.M. 1907 ThefishesofNorthCarolina. Vol. 2. N.C. Geol. Econ. Surv., Raleigh, 453 p. Springer. V.G.. and K.D. Woodburn 1960 An ecological study of the fishes of Tampa Bay area. Fla. Board Conserv. Mar. Res. Lab. Prof. Pap. Ser. 1. 104 p. Summerfelt, R.C.. and G.E. Hall 1987 Age and growth of fish. Iowa State Univ. Press, Ames, .544 p. Fishery Bulletin Index Volume 88 (1-4), 1990 "A comparative study of the megalopal stages of Cancer oregonensis Dana and C. productus Randall (Decapoda: Brachyura: Cancridae) from the northeastern Pacific," by Gregory A. DeBrosse. Adam J. Baldinger, and Patsy A. McLaughlin 39 "A comparative study of habitat use of young-of-the-year, subadult, and adult rockfishes on four habitat types in central Puget Sound." by Kathleen R. Matthews 223 "A fishery-dependent based study of fish species composition and associated catch rates around oil and gas struc- tures off Louisiana," by David R. Stanley and Charles A. Wilson 719 "A practical assessment of the perfor- mance of Shepherd's length composi- tion analysis (SRLCA): Application to Gulf of Maine northern shrimp (Pmi- dalus borealis) survey data," by Mark Terceiro and Josef S. Idoine 761 "A seasonal autoregressive model of the anchovy Engrauiis encrasicoltis in the eastern Mediterranean," by Konstan- tinos L Stergiou 411 Able, Kenneth W., R. Edmond Matheson, Wallace W. Morse, Michael P. Fahay, and Gary Shepherd, "Patterns of summer flounder Paralichthys den- tatiis early life history in the Mid- Atlantic Bight and New Jersey estuaries" 1 Able. Kenneth W.— see Barshaw and Able "Abnormal development and growth reductions of pollock Theragra rhulcngromma exposed to water- soluble fractions of oil," by Mark G. Carls and Stanley D. Rice 29 Abramson, Norman J.— see Smith and Abramson Abundance estimates— see Population studies Africa, southeast coast fishery shrimp, penaeid studies 21 "Age and growth of bluefish. Pomatum its !i panna—see Sand dollar Echinodermata Middle Atlantic Bight 187 "Effects of starvation on the frequency of formation and width of growth in- crements in sagittae of laboratory- reared Atlantic menhaden Bremortia tyrannus larvae," by Gary L. Maillet and David M. Checkley Jr. 155 "Effects of water temperature on size and biochemical composition of winter llounder Pseudopleuronectes americamis at hatching and feeding initiation," by Lawrence J. Buckley, Alphonse S. Smigielski, Thomas A. Halavik, and Geoffrey C. Laurence 419 "Egg and larval distributions of walleye pollock Theragra chalcogramma in Shelikof Strait, Gulf of Alaska," by Arthur W. Kendall Jr. and Susan J. Picquelle 133 Egg predation lobster, spiny 279 Eggleston, David B,, and Eleanor A. Boch- enek, "Stomach contents and parasite infestation of school bluefin tuna Thuniuis thiintiiift collected from the Middle Atlantic Bight, Virginia" 389 Eggs— see Embryos Embryos— see also Larval studies pollock, walleye 471 Emigration— see Behavior studies Engraulis encrasicotus—see Anch(.)vy Eptatr-etus—see Hagfish "Estimating relative survival rate for two groups of larval fishes from field data: Do older larvae survive better than young?," by John M. Hoenig, Piefre Pepin, and William D. Lawing 485 Euthyrinus lineatus—see Tuna, black skipjack "Evidence of survival value related to burying behavior in queen conch Strombus gigas," by Edwin S. Iver- son, Scott P. Bannerol, and Darryl E. Jory 383 Evolution— see Biogeographic studies F'ahay, Michael P.— see Able et al. "Feeding ecology of late-larval and early- juvenile walleye pollock Thenigiv chalcogramma from the Gulf of Alaska in 1987," by .1111 .1. Grover 463 "Feeding habits, age, growth, and repro- duction of Atlantic spadefish Chaefii- dipterus fahe)' {Viscei^: Ephippidae) in South Carolina," by John W. Hayse 67 Feeding— see Food habits Fish aggregating devices tuna 493 Fish, Antarctic, 557 Fish disease studies pollock 29 Fish interaction rockfish trawl harvests 645 salmon. Pacific 620 Fish oil snapper. New Zealand 201 P^ish populations rocky subtidal zone 403 Fisher, Jeffrey P.— see Ostrander et al. Fisher, Joseph P., and William G. Pearcy, "Distribution and residence times of juvenile fall and spring chinook salmon in Coos Bay, Oregon" 51 Fisher, Joseph P., and William G. Pearcy, "Spacing of scale circuli versus growth rate in young coho salmon" 637 Fisher, Joseph P.— see Pearcy et al. Fishery management lingcod 815 rockfish trawl harvests 645 Fishing interaction with marine mammals 344, 347 Fishing, commercial fleet productivity 85 lingcod 815 Fishing industry— see also Foreign fishing productivity measurement 85 Fishing, recreational logliook and creel surveys 719 Fishing, sport lingcod 815 marlin, Pacific blue 397 southern California Bight 599 Floldevigen Biological Station 192 Flounder 777 summer 1 windowpane 429 Fluorochrome dye methods 611 "Food availability as a limiting factor to mussel Mylilus ediiUs growth in California coastal waters," by Henry M. Page and Yann 0. Ricard 677 Food habits hake, Argentine 173 herring. Pacific 193 marine mammals 355 menhaden, Atlantic 158 mus.se I 677 pollock, walleye 463 sablefish 811 salmon. Pacific 617 sea lions, California 509 spadefish, Atlantic 68 trout, anadromous 704 tuna, bluefin 389 whale, baleen 687 "Food habits of California sea lions Zalu- pliiis catiforruanus at San Clemente Island. California, 1981-86," by Mark S. Lowry, Charles W. Oliver, Carolyn Macky, and Jeannie B. Wexler 509 "Food habits of larval sablefish Anoplo- poma fimbria from the Bering Sea," by Jill J. Grover and Bori L. 011a 811 Forbes, Anthony T.— see Benfield et al. Foreign fishing interaction with marine mammals 347 Frequency— see Population studies Freshwater salmonids— see Salmonids Gauldie, Robert W., and Zophie Czochan- ska, "Hyperostosic bones from the New Zealand snapper Chrysophrys auratus (Sparidae)" 201 Gender determination lingcod 815 porpoise, Dall's 119 Genetic studies bass, California 59 lobster, spiny 713 salmonids 657 "Genetic variation in highly exploited spiny lobster Panuiirus marginatus populations from the Hawaiian Ar- chipelago," by Lisa W. Seeb, James E. Seeb, and Jeffrey J. Polovina 713 Georges Bank summer flounder fishery 1 Gerrior, Patricia— see Waring et al. "Gonad morphology, histology, and sper- matogenesis in South Pacific albacore tuna Thunnuft alalnnga (Scombridae)," by Frank J. Ratty, R. Michael Laurs, and Raymond M. Kelly 207 Graves, John E., Michelle J. Curtis, Paul A. Oeth, and Robin S. Waples, "Bio- chemical genetics of southern Califor- nia basses of the genus Pantlabrax: Specific identification of fresh and ethanol-preserved individual eggs and early larvae," 59 Grover, Jill J., "Feeding ecology of late- larval and early-juvenile walleye pollock Theragra chalcogramma from the Gulf of Alaska in 1987" 463 Grover, Jill J., and Bori L. 011a, "Food habits of larval sablefish Anoplopoma fimbria from the Bering Sea" 811 "Growth and emigration of Penaeus iiidi- cus H. Milne-Edwards (Crustacea: Decapoda:Penaeidae) in the St. Lucia Estuary, southern Africa," by Mark C. Benfield, Jaap R. Bosschieter, and Anthony T. Forbes 21 "Growth per molt of male snow crab Chimioecetes vpilio from Conception and Bonavista Bays, Newfoundland," by David M. Taylor and .lohn M. Hoenig 753 Growth rates— see also Age-size estimation armorhead, pelagic 219 bluefish 805 crab, snow 753 croaker, whitemouth 523 dolphin, bottlenose 292 836 INDEX Fishery Bulletin 88(1-4), 1990 Growth rates (continued) drum, red 531 fish, Antarctic 557 herring, Pacific 191 salmon, chinook 57 coho 637 sand dollar 179 tuna 821 vveakfish 745 Growth studies bluefish 805 menhaden, Atlantic 155 mussel 678 sheepshead 829 shrimp, penaeid 21 squid, loliginid 113 trout, anadromous 703 tuna, black skipjack 821 Gulf of Alaska fishery pollock, walleye, 464 473 walleye studies 135 Gulf of California, northern fishery porpoise, vaquita 341 Gulf of Maine fishery northern shrimp mathematical analysis 763 rocky subtidal zone habitat 403 whale, baleen 689 (Julf of Mexico fishery liluefish 805 threadfin. Atlantic 440 tuna larvae 607 Habitat effects 599 conch, queen 575 lobster, American 415 mussel 680 oil and gas structures 719 rocky subtidal zone 403 salmon, chinook 257 salmonids 657 weakfish 747 Habitat studies rockfish 223 Hagfish 787 Hake 95, 777 Argentine 167 Halavik, Thomas A.— see Buckley et al. Hampton, .John, and Geoffrey P. Kirk- wood, "Tag shedding by southern bluefin tuna Thumius nutrcDi/ii" 313 Hancock Seamounts fishery armorhead, pelagic studies 217 Hawaiian fishery fish aggregating devices 493 lob.sler, spiny 714 Hawvermale. Mary P.— see .Seipt et al. Hayse. .John W.. "Feeding habits, age, growth, and reproduction of Atlantic spadefish Chactodipterus /nher- (Pisces: Ephippidae) in South Carolina" 67 Hematology flounder, windowpane 429 Herrick, Samuel F. Jr., and Dale Squires, "On measuring fishing fleet produc- tivity: Development and demonstra- tion of an analytical framework" 85 Herring. Pacific 191 Hightower, Joseph E.. "Multispecies harvesting policies for Washington- Oregon-California rockfish trawl fisheries" 645 Hinckley, Sarah, "Variation of egg size of walleye pollock Therayra chalcu- granuna with a preliminary examina- tion of the effect of egg size on larval size" 471 Hippolytidae—see Shrimp Ho, Ju-shey, "Phylogeny and bio- geography of hakes {Meriiicciux: Teleostei): A cladistic analysis" 95 Hoenig, John M., Pierre Pepin, and William D. Lawing, "Estimating relative survival rate for two groups of larval fishes from field data: Do older larvae survive better than young?" 485 Hoenig. John M.— see Taylor and Hoenig Holland. Kim. Richard Brill, and Ran- dolph K.C. Chang, "Horizontal and vertical movements of Pacific blue marlin captured and released using sportfishing gear" 397 Holland, Kim N., Richard W. Brill, and Randolph K.C. Chang. "Horizontal and vertical movements of yellowfin and bigeye tuna associated with fish aggregating devices" 493 Holt. Rennie S.. and Stephanie N. Sex- ton. "Monitoring trends in dolphin abundance in the eastern tropical Pacific using research vessels over a long sampling period: Analyses of 1986 data, the first year" 105 Hoinnrus aynenaniua—see Lobster, American "Horizontal and vertical movements of Pacific blue marlin captured and released using sportfishing gear," by Kim Holland, Richard Brill, and Ran- raiion. iH ntime and iiJJrcii must be stated and also immediately thereunder the namei and addresses of stockholders owning or holding ! percent or more of total amouni of nock Ij not otned bv a iorporalian, the names and addresses of the indnidual owners rnusi be given If owned by a partnership or other unincorporated firm, its name and address, OS well as thai of each individual must be given If the pubtuaiion is published by a nonprofit organiztiiion, us name and address musi be siaied I litem must be lompleted ) Complete Mailing AddrsBS U.S. Dept. of Commerce 14th St., NW, Washington, DC 20230 Known Bondholders. Mongagees, end Other Security Holders Owning or Holding 1 Percent or More of Total Amount of Bonds, Mortgages or Other Securities lU there are none, so siaiel Compldte Mailing Address 9 For Complei.on by Nonprofit Ofganiietions Authoriied To Mail at Special Rates (DMM Section 42J 12 only) The purpose, function, and nonprofit status of this Ofganizetion and the exempt Status for Federal income tan purposes iCheck one) a Has Not Changed During Preceding 12 Months □ Has Changed During Preceding 12 Months (It ihanged, publisher must submit explanation of ihan^r wuh ihit siaiemeni > Extent and Nature of Circulation (See inslructioni on reverse side) Averege No Copies Each Issue During Preceding 12 Months Actual No Copies of Single Issue Published Nearest to Filing Date A Total No Copies (Set f're-.s Run) 2109 2081 B Paid and/Of Requested Circulation handled by U.S. Supt. 1 Sales thiough dealers and earners, street vendors and counter sales of Documents, Washingtor , DC 20240 2 Mail Subscription (Paid and ''or requested) C Total Paid and'or Requested Circulation iSum or lOBI and IOB2l D. Free Distribution by Mail, Carrier or Other Means Samples. Complimentary, and Other Free Copies E. Total Distribution (Sum of C and D) F Copies Not Distributed 1 Office use, leH over, unaccounted, spoiled after printing 2 Return from News Aganta G . TOTAL 'Sum of E, Fl and 2— should equal net press run shown in A) I certify that the Btatements made by me above are correct and complete Sioneture and Title of Edlfor^^-Pnbllshar, Business Manager, '\Jh/nh\^ S Form 3526. Ftb IVSV / on reverse) i9l Fisheiy Bulletin Guidelines for Contributors Form of manuscript The document should be in the following sequence: Title Page, Abstract (not re- quired for Note), Ifext, Acknowledgments, Citations, Text footnotes, Appendices, Tables, Figure legends, and Figures. Title page should include authors' full names and mailing addresses and the senior author's telephone and FAX numbers. Abstract Not to exceed one double- spaced typed page. Should include a sentence or two explaining to the gen- eral reader why the research was under- taken and why the results should be viewed as important. Abstract should convey the main point of the paper and outline the results or conclusions. No footnotes or references. Text A brief introduction should por- tray the broad significance of the paper. The entire text should be intelligible to readers from different disciplines. All technical terms should be defined, as well as all abbreviations, acronyms, and symbols in text, equations, or formulae. Abbreviate units of measure only when used with numerals or in tables and figures to conserve space. Measure- ments should be expressed in metric units, with other equivalent units given in parentheses. Follow the U.S. Govern- ment Printing Office Style Manual, 1984 ed., and the CBE Style Manual, 5th ed. Fishery and invertebrate no- menclature should follow the American Fisheries Society Special Publication 12 (for fishes), 16 (for moUusks), and i7(for decapod crustaceans). Text footnotes should be numbered in Arabic numerals and typed on a sep- arate sheet from the text. Footnotes are not used for reference material or per- sonal communications, but rather to explain or define terms in the text and for contribution numbers on the title page. Informal sources Personal communi- cations, unpublished data, and untitled manuscripts in preparation are noted parenthetically in the text (full name. affiliation, brief address including zip code, and month and year when appro- priate). Acknowledgments Gather all ac- knowledgments into a brief statement at the end of the text. Give credit only for exceptional contributions and not to those whose contributions are part of their normal duties. Citations All titled sources should be listed in the Citations section, including unpublished and processed material. In text, cite as Smith and Jones (1977) or (Smith and Jones 1977); if more than one citation, list chronologically (Smith 1936, Jones 1975, Doe In press). All sources cited in the text should be listed alphabetically by the senior authors' surnames under the heading CITA- TIONS. Abbreviations of periodicals and serials should conform to Serial Sources for the BIOSIS Data Base™. Indicate whether sources are in a langu- age other than English. For informal literature, include address of author or publisher. Authors are responsible for the accuracy of all citations. Tables should supplement, not dupli- cate, the text. Each table should be numbered and cited consecutively, with headings short but amply descriptive so that the reader need not refer to the text. For values less than one, zeros should precede all decimal points. In- dicate units of measure in column head- ings; do not deviate from the unit of measure within a column. Table foot- notes should be noted consecutively in Roman letters across the page from left to right and then down. Since all tables are typeset, they need not be submitted camera-ready. Figures Photographs and line draw- ings should be of professional quality- clear and concise— and reducible to 42 picas for full-page width or to 20 picas for a single-column width, and to a max- imum 55 picas high. All graphic ele- ments in illustrations must be propor- tioned to insure legibility when reduced to fit the page format. Line weight and lettering should be sharp and even. Let- tering should be upper and lower case, and vertical lettering should be avoided- whenever possible (except for vertical, y, axis). Zeros should precede all deci- mal points for values less than one. Re- productions of line artwork are accepted in the form of high-quality photographic prints from negatives or photomechan- ical transfer (PMT). Halftones should be sharply focused with good contrast. Micron rules should be inserted on elec- tron micrographs, even when magnifi- cation is included in the figure legend. There should be clear distinction be- tween identifying letters (press-on or overlay) and background of photograph. Label each figure in pencil on the back. Send only xerox copies of figures to the Scientific Editor; originals or photo- graphic prints will be requested later when the manuscript is accepted for publication. Copyright Government publications are in the public domain, i.e., they are not protected by copyright. Submission of manuscript Disks Authors are encouraged to re- tain manuscripts on word-processing storage media (diskettes, floppy disks) and submit a double-spaced hardcopy run from the storage media. Submit disks as MS-DOS "print" or "non- document" files (often called "ASCII files"). If a disk cannot be converted to an ASCII file, the author should indicate on the disk the source computer and software language along with the file name. Either 5V4-inch or 3V2-inch disks from IBM-compatible or Apple/Macin- tosh systems (non-graphics only) can be submitted, double-sided/double-density or high-density, limiting each file to 300 kilobytes. All 8-inch word-processing disks (e.g., Wang or NBI) must be con- verted onto 5V4- or 3V2-inch MS-DOS print disks. Send original hardcopy and two dupli- cated copies to: Dr. Linda L. Jones, Scientific Editor National Marine Mammal Laboratory F/AKC3 National Marine Fisheries Service, NOAA 7600 Sand Point Way NE Seattle, WA 98115-0070 Copies of published articles and notes The senior author and his/her organiza- tion each receive 50 separates free-of- charge. Additional copies may be pur- chased in lots of 100. MBL WHOI LIBRARY H nuK P