U.S. Department of Commerce Volume 92 Number 1 January 1994 U.S. Department of Commerce Ronald H. Brown Secretary National Oceanic and Atmospheric Administration D. James Baker Under Secretary for Oceans and Atmosphere National Marine Fisheries Service Rolland A. Schmitten Assistant Administrator for Fisheries Scientific Editor Dr. Ronald W. Hardy Northwest Fisheries Science Center National Marine Fisheries Service, NOAA 2725 Montlake Boulevard East Seattle, Washington 981 12-2097 Editorial Committee Dr. Andrew E. Dizon National Marine Fisheries Service Dr. Linda L. Jones National Marine Fisheries Service Dr. Richard D. Methot National Marine Fisheries Service Dr. Theodore W. Pietsch University of Washington Dr. Joseph E. Powers National Marine Fisheries Service Dr. Tim D. Smith National Marine Fisheries Service 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, Attn: Chief, Mail List Branch, Mail Stop SSOM, Washington, DC 20402-9373. 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: $24.00 domestic and $30.00 foreign. Cost per single issue: $12.00 domestic and $15.00 foreign. See back page for order form. Managing Editor Sharyn Matriotti National Marine Fisheries Service Scientific Publications Office 7600 Sand Point Way NE, BIN C 1 5700 Seattle, Washington 98 1 1 5-0070 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. 1 103. 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 92 Number 1 January 1994 Fishery Bulletin Biological laboratory/ Je Oceancgraphic Instituuo,, Library B 2 3 1994 Contents Woods Hole, MA 02543 1 Barbieri, Luiz R., Mark E. Chittenden Jr., and Cynthia M. Jones Age, growth, and mortality of Atlantic croaker, Micropogonias undulatus, in the Chesapeake Bay region, with a discussion of apparent geographic changes in population dynamics 1 3 Bigelow, Keith A. Age and growth of the oceanic squid Onychoteuthis borealyaponica in the North Pacific 26 Burke, Vincent J., Stephen J. Morreale, and Edward A. Standora Diet of the Kemp's ridley sea turtle, Lepidochelys kempu, in New York waters 33 Ditty, James G., and Richard F. Shaw Larval development of tripletail, Lobotes sunnamensis (Pisces: Lobotidae), and their spatial and temporal distribution in the northern Gulf of Mexico 46 Ferreira, Beatrice Padovani, and Garry R. Russ Age validation and estimation of growth rate of the coral trout, Plectropomus leopardus. (Lacepede 1 802) from Lizard Island, Northern Great Barrier Reef 58 Gold, John R., and Linda R. Richardson Genetic distinctness of red drum [Sciaenops ocellatus) from Mosquito Lagoon, east-central Florida 67 Incze, Lewis S., and Terri Ainaire Distribution and abundance of copepod naupln and other small (40-300 (im) zooplankton during spring in Shelikof Strait, Alaska Fishery Bulletin 92(1), 1994 79 Jaenicke, Herbert W., and Adrian G. Celewycz Marine distribution and size of juvenile Pacific salmon in Southeast Alaska and northern British Columbia 91 Johnson, Allyn G., William A. Fable Jr., Churchill B. Grimes, Lee Trent, and Javier Vasconcelos Perez Evidence for distinct stocks of king mackerel, Scomberomorus cavalla, in the Gulf of Mexico 102 Milton, David A., Stephen J. M. Blaber, and Nicholas J. F. Rawlinson Reproductive biology and egg production of three species of Clupeidae from Kiribati, tropical central Pacific 122 Perryman, Wayne L., and Morgan S. Lynn Examination of stock and school structure of striped dolphin [Stenella coeruleoalba) in the eastern Pacific from aerial photogrammetry 132 Punsly, Richard G., Patrick K. Tomlinson, and Ashley J. Mullen Potential tuna catches in the eastern Pacific Ocean from schools not associated with dolphins 144 Sinclair, Elizabeth, Thomas Loughlin, and William Pearcy Prey selection by northern fur seals (Callorhinus ursinus) in the eastern Bering Sea 157 Stone, Heath H., and Brian M. Jessop Feeding habits of anadromous alewives, Alosa pseudoharengus, off the Atlantic Coast of Nova Scotia 171 Stoner, Allan W, and Kirsten C. Schwarte Queen conch, Strombus gigas, reproductive stocks in the central Bahamas: distribution and probable sources 180 Wilber, Dara H. The influence of Apalachicola River flows on blue crab, Callinectes sapidus. in north Florida 189 Fargo, Jeff, and Albert V. Tyler Oocyte maturation in Hecate Strait English sole (Pleuronectes vetulus) 203 Polovina, Jeffrey J., and Gary T. Mitchum Spiny lobster recruitment and sea level results of a 1990 forecast 206 List of recent NOAA Technical Reports NMFS Abstract. — Atlantic croaker, Micropogonias undulatus, col- lected from commercial catches in Chesapeake Bay and in Virginia and North Carolina coastal waters during 1988-1991 (n = l,967) were aged from transverse otolith sec- tions. Ages 1-8 were recorded, but eight-year-old fish were rare. Mar- ginal increment analysis showed that for ages 1-7, annuli are formed once a year during the pe- riod April-May. Otolith age read- ings were precise: >99% agree- ment within and between readers. Observed lengths-at-age were highly variable and growth rate decreased after the first year. De- spite the high variability in sizes- at-age, observed lengths for ages 1-7 fit the von Bertalanffy growth model (r2=0.99; n=753) well. No differences in growth were found between sexes. Total annual in- stantaneous mortality (Z) esti- mated from maximum age and from a catch curve of Chesapeake Bay commercial catches ranged from 0.55 to 0.63. Our results do not indicate the existence of a group of larger, older Atlantic croaker in Chesapeake Bay com- pared with more southern waters and suggest that the hypothesis of a basically different population dynamics pattern for this species north and south of Cape Hatteras, North Carolina, should be reevalu- ated. Age, growth, and mortality of Atlantic croaker, Micropogonias undulatus, in the Chesapeake Bay region, with a discussion of apparent geographic changes \n population dynamics* Luiz R. Barbieri College of William and Mary, Virginia Institute of Marine Science Gloucester Point. Virginia 23062 Present address: University of Georgia Marine Institute Sapelo Island. Georgia 31327 Mark E. Chittenden Jr. College of William and Mary, Virginia Institute of Marine Science Gloucester Point, Virginia 23062 Cynthia M. Jones Old Dominion University. Applied Marine Research LaPoratory Norfolk, Virginia 23529 Manuscript accepted 12 August 1993 Fishery Bulletin 92:1-12 (1994) The Atlantic croaker, Micropo- gonias undulatus (Linnaeus), is one of the most abundant inshore demersal fishes along the Atlantic and Gulf of Mexico coasts of the United States (Joseph, 1972). Al- though recent commercial and rec- reational catches have come prima- rily from the South Atlantic Bight and the Gulf of Mexico, Atlantic croaker still support important fisheries along the Mid-Atlantic coast, especially from Maryland to North Carolina (Wilk, 1981). In Chesapeake Bay, they are caught by commercial and recreational fishermen during late spring and early fall migrations and, to a lesser extent, during the summer. In winter, Atlantic croaker leave the Bay to overwinter off the coast of Virginia and North Carolina, where they are caught by otter trawl and gillnet fisheries (Haven, 1959). Little is known about age, growth, and mortality of Atlantic croaker in the Middle Atlantic and Chesapeake Bay regions. Studies based on length frequencies (Ha- ven, 1957; Chao and Musick, 1977) require considerable subjective in- terpretation given the extended spawning period of Atlantic croaker (Morse, 1980; Warlen, 1982; Bar- bieri et al., unpubl. ms.) and the difficulty in distinguishing modal groups at older ages (White and Chittenden, 1977; Jearld, 1983). Al- though scale-ageing has also been used (Welsh and Breder, 1923; Wallace, 1940; Ross 1988), prob- lems in applying this method to Atlantic croaker have been widely reported (Roithmayr, 1965; Joseph, 1972; Barger and Johnson, 1980; Barbieri, 1993). In this study we provide informa- tion on age, growth, and mortality of Atlantic croaker in the Chesa- * Contribution No. 1806 from the College of William and Mary, School of Marine Sci- ence/Virginia Institute of Marine Science, Gloucester Point, Virginia 23062 Fishery Bulletin 92(1). 1994 peake Bay region using a validated otolith-ageing method. We also evaluate the relationship between otolith size and fish size and age, and discuss the implications of using otoliths for ageing Atlantic croaker. Finally, based on current information on growth, and size and age compositions in Chesa- peake Bay, we discuss the hypothesis of White and Chittenden (1977) and Ross (1988) regarding the existence of a basically different population dynam- ics pattern for Atlantic croaker north and south of Cape Hatteras, North Carolina. Methods Atlantic croaker were collected between June 1988 and June 1991 from commercial pound-net, haul- seine, and gillnet fisheries which operate from early spring to early fall in Chesapeake Bay. Local fish processing houses and seafood dealers were con- tacted weekly or fortnightly, and one 22.7-kg (50-lb) box of fish of each available market grade (small, medium, or large) was purchased. Although boxes of fish were not randomly selected, Chittenden (1989) found only minor among-box differences in Atlantic croaker length compositions in pound-net and haul-seine catches. Because nearly all variation in size compositions was captured by the within-box variation, box selection did not present a problem. Since Atlantic croaker migrate from Chesapeake Bay in early fall to overwinter offshore (Haven, 1959), samples for the period November-March were obtained from commercial trawlers which op- erate in Virginia and North Carolina shelf waters. Young of the year (90-114 mm total length, TL) used to validate the first annulus on otoliths were ob- tained from the Virginia Institute of Marine Science juvenile bottom trawl survey. Fish were measured for total length (TL, ±1.0 mm), weighed for total weight (TW, ±1.0 g), sexed, and both sagittal otoliths removed and stored dry. The left otolith was transversely sectioned through the core with the diamond blade of a Buehler low- speed Isomet saw. Sections 350-500 urn thick were mounted on glass slides with Flo-texx clear mount- ing medium and read under a dissecting microscope (6-12x) with transmitted light and bright field, with the exception of samples from the period April-May, when sections were also read with reflected light and dark field to help identify the last annulus. Ages were assigned based on annulus counts; January 1 was taken as an arbitrary average birthdate when fish from one age class were as- signed to the next oldest (Jearld, 1983). Although the average spawning date (average biological birthdate) of Atlantic croaker in the Chesapeake Bay region occurs in September (Barbieri et al., unpubl. ms.), we chose, for ageing purposes, to use January 1 as the average birthdate because annuli are formed during the period April-May (see Age deter- mination below). To assess ageing precision, all otolith sections (n- 1,967) were read twice by each of two readers, and agreement between readings and readers evaluated by percent agreement. All dis- agreements were resolved by a third reading with both readers. Annuli were validated by the marginal increment method (Bagenal and Tesch, 1978). For each age, the translucent margin outside the proximal end of the last annulus was measured along the ventral side of the otolith sulcal groove (Fig. 1). Measurements (±0.02 mm) were taken with an ocular micrometer at 25x. To evaluate growth, observed lengths at ages 1-7 were fit to the von Bertalanffy model (Ricker, 1975) by using nonlinear regression (Marquardt method). Model parameters were the following: Lm, the mean asymptotic length; K, the Brody growth coefficient; and t(), the hypothetical age at which a fish would have zero length (Ricker, 1975). To cor- rect for growth after the time of annulus formation, only data for September, the peak spawning and thus average biological birthdate for Atlantic croaker in the Chesapeake Bay region (Barbieri et al., unpubl. ms.), were used for growth analysis. To evaluate changes in otolith size relative to fish length and age, 30 randomly selected otoliths per age, for ages 1-7 ( 198^100 mm TL), were measured for maximum length (OL, ±0.05 mm) and maximum thickness (OT, ±0.05 mm), and weighed (OW, ± 0.001 g). After sectioning, otoliths were measured for otolith radius (OR, ±0.02 mm), defined as the dis- tance between the center of the core and the otolith outer edge along the ventral side of the sulcal groove (Fig. 1). Relationships between otolith measure- ments and fish TL were evaluated by regression analysis. The effect offish age on these relationships was evaluated by analysis of covariance (ANCOVA). Linear regression was used to determine a length- weight relationship for fish ranging from 152 to 400 mm TL (36.3 to 967.0 g TW). Difference between sexes was tested by ANCOVA. The hypothesis of isometric growth (Ricker, 1975) was tested by t-test. Instantaneous total annual mortality rates, Z, were estimated from maximum age by using Hoenig's pooled regression equation (Hoenig, 1983), by calculating a theoretical total mortality for the entire lifespan following the reasoning of Royce (1972), and by the regression method with a catch curve of combined pound-net, haul-seine, and gillnet. Barbieri et al.: Age, growth, and mortality of Micropogonias undulatus Proximal Ventra Figure 1 Transverse otolith section of an 8-year-old Atlantic croaker caught in Sep- tember 1988 in Chesapeake Bay. Arrows indicate annuli. The translu- cent zone beyond the last annulus represents additional growth after the annulus was formed during April-May. SG =sulcal groove, a = artifact of preparation. Ventral and proximal indicate axes of orientation. data for all recruited ages having five or more fish (Chapman and Robson, 1960). To avoid sampling bias associated with individual gears, we considered the age-frequency distribution obtained from data from combined gears as the best estimate of Atlan- tic croaker age composition in Chesapeake Bay (Ricker, 19751. Commercial trawl collections were not used in this analysis because they had different length compositions than the other gears and could be biased towards small fish. Because in catch curve analysis the age group represented by the apex of the catch curve may or may not be fully recruited to the gears (Everhart and Youngs, 1981), mortal- ity estimates were based on ages 3-7 only. Data from 1988 to 1991 were combined to minimize the effect of variation in year-class strength (Robson and Chapman, 1961). The right tail of the catch curve (Ricker, 1975) was tested for deviation from linear- ity by analysis of variance (ANOVA). Values of Z were converted to total annual mortality rates, A, by using the relationship A = 1 - e -z{ Ricker, 1975). All statistical analyses were performed by using the Statistical Analysis System (SAS, 1988). Rejec- tion of the null hypothesis in statistical tests was based on a=0.05. F-tests in ANCOVA were based on Type III sums of squares (Freund and Littell, 1986). Assumptions of linear models were checked by re- sidual plots as described in Draper and Smith ( 1981). For the OL-TL, OW-TL, and TW-TL relation- ships, and for all ANCOVA and ANOVA analyses, data were log10-transformed to correct for non-lin- earity and heterogeneous variances. For the catch curve analysis, loge-transformed numbers at age were regressed on age. Unless otherwise indicated, back-transformed data and regression equations are presented in the results. Results Age determination Transverse otolith sections of Atlantic croaker show very clear, easily identified marks that can be used for ageing. Typical sections have an opaque core surrounded by a blurred opaque band composed of fine opaque and translucent zones (Fig. 1). This band represents the first annulus. The width of this annulus varies among fish, from a very narrow band that is almost continuous with the core, to a wide, well-defined band clearly separated from the core. Because of this variation in width and proximity to Fishery Bulletin 92(1). 1994 the core, the first annulus is sometimes difficult to identify. Subsequent annuli are represented by eas- ily identified, narrow, opaque bands that alternate with wider translucent bands outside the proximal margin of the first annulus (Fig. 1). Annuli are formed on otoliths once a year in the period April-May. For ages 1-7, mean monthly marginal increment plots show only one minima during the year, indicating that only one annulus is formed each year (Fig. 2). The trough starts abruptly in April, a period when there is, in general, maxi- mum variation in the mean marginal increment, suggesting that some fish have begun to form the annulus while others have not. Lowest marginal increment values occurred in May, the most inten- sive period of annulus formation. Marginal incre- E E c CD E CD CJ c ~co c en ca 18 •' ,6 ft* Age 1 jsJ Age 2 ) I! 1 0 9 rfj 12 rfi "1 17 ^rfl 30 IB i 14 12 6 I r 5 Age 3 12 snfin 39 12 1 9 25 1 T 4 I Age 4 7 1 T | 1 : 23 |* 10 '» 1*1 : . ,. -if II 27 a 10 JFMAMJ JASOND 3 e Age 5 21 13 7 4 * s 4 r5- Age 6 32 11 8 1 1 Age 7 nnn Age 8 H JFMAMJJASOND Months Figure 2 Mean monthly marginal increment for Atlantic croaker ages 1-8 from the Chesa- peake Bay region, 1988-91. Vertical bars are ±1 standard error. Numbers above the bars are sample sizes. Barbien et al Age, growth, and mortality of Micropogonias undulatus ment values progressively rise to a somewhat stable maximum from October through March or April, indicating a period of little or no otolith growth. Because only two age-8 fish were collected, it was not possible to validate annuli beyond age 7. To confirm our interpretation that the blurred opaque band around the otolith core represents the first annulus, (i.e., that fish hatched in the fall form a mark during their first spring), otolith sections of young of the year (94-114 mm) collected during the period March-June were examined. All those col- lected in March-April were developing fine opaque marks around the core, and all those in May-June had an opaque mark already formed (Fig. 3). Otolith age readings were very precise, both within and between readers. Percent agreement was Figure 3 Transverse otolith section of a young-of-the-year Atlan- tic croaker ( 114 mm TL) collected in June 1990 in Chesa- peake Bay. The arrow indicates the outer edge of the first annulus formed during the period April-May. SG=sulcal groove; Ve=ventral; Pr=proximal; a=artifact of preparation. 99.5% for reader 1, 99.3% for reader 2, and 99.2% between readers. In all cases of disagreement, the difference never exceeded 1 year. Only one of the 1,967 left otoliths sectioned was crystallized and could not be read. In that case, the right otolith was read. Difficulty in ageing Atlantic croaker from otolith sections did not increase with increasing age. However, proper identification of the first annulus was very important. All disagreements, independent of age, were due to problems in identifying the first annulus. Otolith size relative to fish size and age Changes in otolith size relative to fish size were not constant along all axes (Fig. 4). Otolith maximum length was the only axis that showed a linear, isometric increase with fish length. Otolith ra- dius, the axis along which annuli were read in transverse sections, showed a non-linear rela- tionship with fish length, and had the small- est r2 of all variables (Fig. 4). The curvilinear relationship suggests that otolith growth rela- tive to fish growth slows down along this axis as fish get bigger. Despite its poor relationship with fish length, otolith radius showed a very strong linear re- lationship with fish age. An ANCOVA model showing length, age, and their interaction ex- plained 97% of the variation in otolith radius (Table 1). All factors in the model were highly significant (P<0.01). Similar models for otolith maximum length, maximum thickness, and weight were also highly significant and had high coefficients of determination (r2>0.85). However, significance for these models was due to fish length only, neither age nor the inter- action factor was significant. Growth Observed lengths varied greatly within ages (Fig. 5). Atlantic croaker showed a rapid in- crease in size during the first year, but annual growth rate greatly decreased during the sec- ond year, remaining comparatively low there- after (Fig. 5). On average, 64% of the cumula- tive total observed growth in length occurred in the first year and 84% was completed after two years. No differences in mean lengths at age were found between sexes (Mest at each age; P>0.05 for all ages). Mean observed total lengths for pooled sexes were 201, 263, 274, 285, 290, 307, 309, and 313 mm, for ages 1-8, respectively. Fishery Bulletin 92(1). 1994 A B 6 OR = -3 90 + 0 03 TL - 0 001 TL2 ^10- OT = -2.73 + 0.04 TL - 0.0004 TL2 r2 = 0 43, P = 0 0001 £ r2 = 0 65; P = 0 0001 Radius (mm) O CO 1.82 kg. Between 1977 and 1982, however, al- though the minimum citation weight was raised to 1.82 kg, 599 citations were issued, including 47 entries for Atlantic croaker >2.27 kg (483-610 mm TL). The largest number of citations occurred in 1979 and 1980, coinciding with Ross's (1988) sam- pling period in North Carolina. Records from the Delaware State Fishing Tournament show the same pattern as that from Virginia. The number of cita- tions was very small during the early 1970's, reached a peak in 1980, and decreased rapidly there- after. Although complete information covering their entire range is not available, state records of Atlan- tic croaker along the east coast of the United States show the same pattern. Records from Georgia to New Jersey were broken during the period 1977-82, indicating that 1) unusually large fish occurred during this period and have not occurred since; and 2) their occurrence was not limited to areas north of North Carolina. In conclusion, recent size and age composition data do not indicate the existence of a group of larger, older Atlantic croaker in the Chesapeake Bay region compared with more southern waters. His- toric information agrees well with our results and indicates that fish >400 mm TL have not repre- sented a large proportion of Atlantic croaker in this area. The abundance of unusually large fish during the period 1977-82 apparently constituted an un- usual event and may reflect passage through the fishery of a few strong year classes that seemingly disappeared after 1982. Similar episodes — the occur- rence of larger fish for a few years — have been pre- viously reported for Atlantic croaker in Chesapeake Bay (Hildebrand and Schroeder, 1928; Massmann and Pacheco, 1960), suggesting the phenomenon happens periodically. An increase in survivorship of Barbien et al.: Age, growth, and mortality of Micropogonias undulatus 1 I early spawned fish, combined with higher mortal- ity of late-spawned fish as a result of low winter temperatures in estuarine nursery areas (Mass- mann and Pacheco, 1960; Joseph, 1972; Warlen and Burke, 1991) could account for an increase in the proportion of larger fish in certain years and explain the episodic occurrence of large Atlantic croaker in this area. Our results for Chesapeake Bay, together with records of large fish south of North Carolina during 1977-82, suggest that the hypothesis of a basically different life history and population dynamics pat- tern for Atlantic croaker north and south of Cape Hatteras, North Carolina, should be reevaluated. However, sampling programs over time describing size and age compositions of Atlantic croaker throughout their range are still necessary to fully evaluate this question. Acknowledgments We would like to thank the Chesapeake Bay com- mercial fishermen and James Owens (VIMS) for helping us obtain samples. Sue Lowerre-Barbieri helped with fish processing and with otolith section- ing and reading. Claude Bain (Virginia Saltwater Fishing Tournament) and Jessie Anglin (Delaware Department of Natural Resources) provided infor- mation on Atlantic croaker recreational citation records. Ronald Hardy, Joe Loesch, Sue Lowerre- Barbieri, Jack Musick, Rogerio Teixeira, and two anonymous reviewers made helpful suggestions to improve the manuscript. Financial support was pro- vided by the College of William and Mary, Virginia Institute of Marine Science, by Old Dominion Uni- versity, Applied Marine Research Laboratory, and by a Wallop/Breaux Program Grant for Sport Fish Res- toration from the U.S. Fish and Wildlife Service through the Virginia Marine Resources Commission, Project No. F-88-R3. Luiz R. Barbieri was partially supported by a scholarship from CNPq, Ministry of Science and Technology, Brazil (process No. 203581/ 86-OC). Literature cited Bagenal, T. B., and F. W. Tesch. 1978. Age and growth. In T. B. Bagenal (ed.). Methods for assessment of fish production in fresh waters, 3rd ed., p. 101-136. Blackwell Scientific Publications, Oxford. Barbieri, L. R. 1993. Life history, population dynamics, and yield- per-recruit modeling of Atlantic croaker, Micropogonias undulatus, in the Chesapeake Bay area. Ph.D. diss., School of Marine Science, Col- lege of William and Mary, VA, 140 p. 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. Barger, L. E., and A. G. Johnson. 1980. An evaluation of marks on hardparts for age determination of Atlantic croaker, spot, sand seatrout, and silver seatrout. NOAATech. Memo, NMFS, SEFC-22. Campana, S. E. 1990. How reliable are growth back-calculations based on otoliths? Can. J. Fish. Aquat. Sci. 47:2219-2227. Casselman, J. M. 1990. Growth and relative size of calcified struc- tures offish. Trans. Am. Fish. Soc. 119:673-688. 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. 75:657-702. Chapman, D. G., and D. S. Robson. 1960. The analysis of a catch curve. Biometrics 16:354-368. Chittenden, M. E., Jr. 1977. Simulations of the effects of fishing on the Atlantic croaker, Micropogon undulatus. Proc. Gulf Carib. Fish. Inst. 29:68-86. 1989. Sources of variation in Chesapeake Bay pound-net and haul-seine catch compositions. N. Am. J. Fish. Mgmt. 5:86-90. Draper, N. R., and H. Smith. 1981. Applied regression analysis, 2nd ed. John Wiley & Sons, NY, 709 p. Everhart, W. H., and W. D. Youngs. 1981. Principles of fishery science, 2nd ed. Cornell University Press, Ithaca, 349 p. Freund, R. J., and R. Littell. 1986. SAS system for linear models, 1986 ed. SAS Institute Incorporated, Cary, NC. Hales, L. S., Jr., and E. J. Reitz. 1992. Historical changes in age and growth of At- lantic croaker, Micropogonias undulatus (Per- ciformes: Sciaenidae). J. Archaeol. Sci. 19:73-99. Haven, D. S. 1957. Distribution, growth, and availability of ju- venile croaker, Micropogon undulatus, in Vir- ginia. Ecology 38:88-97. 1959. Migration of the croaker, Micropogon undulatus. Copeia 1959:25-30. Higgins, E., and J. C. Pearson. 1928. Examination of the summer fisheries of Pamlico and Core sounds, N.C., with special ref- erence to the destruction of undersized fish and the protection of the gray trout, Cynoscion regalis. Rep. U.S. Comm. Fish. 2:29-65. Hildebrand, S. F., and W. C. Schroeder. 1928. The fishes of Chesapeake Bay. Bull. U.S. Bur. Fish. 43:1-388. 12 Fishery Bulletin 92(1), 1994 Hoenig, J. M. 1983. Empirical use of longevity data to estimate mortality rates. Fish. Bull. 82:898-902. Jearld, A., Jr. 1983. Age determination. In L. A. Nielsen and D. L. Johnson (eds.), Fisheries techniques, p. 301- 324. Am. Fish. Soc, Bethesda, MD. Joseph, E. B. 1972. The status of the sciaenid stocks of the mid- dle Atlantic coast. Chesapeake Sci. 13:87-100. Massmann, W. H., and A. L. Pacheco. 1960. Disappearance of young Atlantic croakers from the York River, Virginia. Trans. Am. Fish. Soc, 89:154-159. Morse, W. W. 1980. Maturity, spawning and fecundity of Atlan- tic croaker, Micropogonias undulatus, occurring north of Cape Hatteras, North Carolina. Fish. Bull. 78:190-195. Mosegaard, H., H. Svendang, and K. Taberman. 1988. Uncoupling of somatic and otolith growth rates in Arctic char (Salvelinus alpinus) as an ef- fect of differences in temperature response. Can. J. Fish. Aquat. Sci. 45:1514-1524. Music, J. L., Jr., and J. M. Pafford. 1984. Population dynamics and life history aspects of major marine sportfishes in Georgia's coastal waters. Georgia Dep. Nat. Res., Coast. Res. Div., Contr. Ser. No. 38, Brunswick, 382 p. Pearson, J. C. 1932. Winter trawl fishery off the Virginia and North Carolina coasts. U.S. Bur. Fish., Invest. Rep. No. 10, Washington, 31 p. Ricker, W. E. 1975. Computation and interpretation of biological statistics of fish populations. Bull. Fish. Res. Board Can. 191, 382 p. 1992. Back-calculation offish lengths based on pro- portionality between scale and length increments. Can. J. Fish. Aquat. Sci. 49:1018- 1026. Robson, D. S., and D. G. Chapman. 1961. Catch curves and mortality rates. Trans. Am. Fish. Soc. 90:181-189. Roithmayr, C. M. 1965. Review of industrial bottomfish fishery in northern Gulf of Mexico, 1959-62. Comm. Fish. Rev., U.S. 27:1-6. Ross, J. L. 1991. Assessment of the North Carolina winter trawl fishery, September 1985-April 1988. North Carolina Dep. Environ. Health Nat. Res., Div. Mar. Fish., Spec. Sci. Rep. No. 54, 80 p. Ross, S. W. 1988. Age, growth, and mortality of Atlantic croaker in North Carolina, with comments on population dynamics. Trans. Am. Fish. Soc. 117:461-473. Royce, W. F. 1972. Introduction to the fisheries sciences. Academic Press, NY, 351 p. SAS. 1988. SAS/STAT user's guide. Release 6.03 edition. SAS Institute Inc., Cary, NC, 1028 p. Wallace, D. H. 1940. Sexual development of the croaker, Micropogon undulatus, and distribution of the early stages in Chesapeake Bay. Trans. Am. Fish. Soc. 70:475-482. Warlen, S. M. 1982. Age and growth of larvae and spawning time of Atlantic croaker in North Carolina. Proc. Ann. Conf. S.E. Assoc. Fish and Wildl. Agencies 34:204- 214. Warlen, S. M., and J. S. Burke. 1991. Immigration of larvae of fall/winter spawn- ing marine fishes into a North Carolina estuary. Estuaries 13:453-461. Welsh, W. W., and C. M. Breder. 1923. Contributions to the life histories of Sciaenidae of the eastern United States coast. Bull. U.S. Bur. Fish. 39:141-201. White, M. L., and M. E. Chittenden Jr. 1977. Age determination, reproduction, and popu- lation dynamics of the Atlantic croaker, Micropogonias undulatus. Fish. Bull. 75:109- 123. Wilk, S. J. 1981. The fisheries for Atlantic croaker, spot, and weakfish. In H. Clepper (ed.) Proceedings of the 6th annual marine recreational fisheries sympo- sium, p. 5-27. Sport Fish. Inst., Washington. Wright, P. J. 1991. The influence of metabolic rate on otolith in- crement width in Atlantic salmon parr, Salmo salar L. J. Fish Biol. 38:929-933. Abstract. Statolith micro- structural analysis was applied to 126 specimens of the oceanic bo- real clubhook squid, Onychoteuthis borealijaponica, for estimation of age and growth rates. Specimens were captured from the western, central, and eastern North Pacific between approximately lat. 38° N and 47°N by driftnet fishing, trawling, and jigging in the sum- mers of 1990 and 1991. Results suggest that increments were de- posited at a rate of one per day. Both sexes live approximately one year; males mature at smaller sizes and younger ages than fe- males. Exponential growth models suggest that growth in length was similar for males and females (0.80% ML/day) in the central North Pacific, while growth in weight was higher for females ( 1.90% WT/day) than males (1.40% WT/day). Females in the western North Pacific exhibited faster growth rates than individuals from the central North Pacific. O. borealijaponica were estimated to have hatched year round based on back calculation of statolith incre- ments from the time of capture. Post-recruit individuals exploited in the O. borealijaponica jig fish- ery and Ommastrephes bartramii driftnet fishery typically hatched from late summer to early winter. Age and growth of the oceanic squid Onychoteuthis borealijaponica in the North Pacific Keith A. Bigelow Honolulu Laboratory, Southwest Fisheries Science Center National Marine Fisheries Service. NOAA 2570 Dole Street, Honolulu, HI 96822-2396 The oceanic boreal clubhook squid Onychoteuthis borealijaponica Okada, 1927 is common in subarc- tic waters of the North Pacific. This species ranges from the western coast of the United States and Canada to the eastern coast of Hokkaido, Japan, and the Kurile Islands, but does not occur in the Sea of Okhotsk or Bering Sea (Young, 1972; Murata et al., 1976; Naito et al., 1977a; Fiscus and Mercer, 1982; and Kubodera et al., 1983). Onychoteuthis borealijapon- ica has commercial value through- out its range. Between 1971 and 1979, commercial landings aver- aged 1,171 metric tons (t) per year from a jig fishery in oceanic waters east of Hokkaido, Japan (Okutani and Murata, 1983), and approxi- mately 254 and 2,705 t of O. bor- ealijaponica were caught in 1990 and 1991, respectively, by Japan, Korea, and Taiwan in the Ommas- trephes bartramii highseas driftnet fishery (DiNardo and Kwok, in re- view1). Based on exploratory fish- ing, Fiscus and Mercer (1982) sug- gested that O. borealijaponica could be commercially exploited by a jig fishery from the Gulf of Alaska westward to the Aleutian Islands, and Murata (in Okutani, 1977) in- dicated that the potential fishery yield of O. borealijaponica may be 50,000-200,000 t in an area west of Manuscript accepted 26 July 1993 Fishery Bulletin 92:13-25 (1994) 1 DiNardo, G. T., and W. Kwok. In review. Estimates offish and cephalopod catch in the North Pacific high-seas driftnet fish- eries, 1990-91. long. 152°E and lat. 40-45°N. If a commercial fishery does develop, accurate life-history information is essential for management purposes. The general biology and feeding ecology of Onychoteuthis borealija- ponica have been investigated (Naito et al., 1977b; Okutani and Murata, 1983); however, little in- formation is available on age and growth. Average growth rates have been inferred from length-fre- quency distributions of sequential jigging samples (Murata and Ishii, 1977). This study suggested that the lifespan for boreal clubhook squid is approximately one year; females grow faster and attain a larger size (370 mm mantle length (ML)) than males (270 mm ML). Growth estimates from driftnet studies (Kubodera et al., 1983; Kubodera, 1986) were inconclusive because length-frequency modes were impossible to detect, possibly because of protracted spawning seasons or variable individual growth rates within a population. The accuracy and precision of cephalopod growth estimates have been greatly enhanced through the use of daily increments within sta- toliths (Natsukari et al., 1991). Ageing by counting statolith incre- ments allows the estimation of size at age and may provide informa- tion on individual age and growth rates. Hatchdates can be estimated by back calculation of daily incre- ments. Age and growth estimates derived from statolith analysis 13 14 Fishery Bulletin 92(1), 1994 have been obtained from a variety of neritic squid species (see review by Rodhouse and Hatfield, 1990a). The objectives of this study were to 1) estimate the age and growth of O. borealijaponica from sta- tolith microstructural analysis, 2) determine the periodicity of increment formation, 3) statistically compare appropriate growth models fit to the age- ing data, 4) determine the distribution of back-cal- culated hatching dates of O. borealijaponica and draw inferences about spawning locations, and 5) determine the relationship between age and matu- rity stages. Materials and methods Taxonomic clarifications At least five onychoteuthid species are found in the North Pacific: O. borealijaponica from subarctic waters; an undescribed species occupying the North Pacific transition zone (-29— 40"N, Bigelow, unpubl. data); and three subtropical species of the O. banksii complex (Young and Harman, 1987). Juvenile, sub- adult, and adult O. borealijaponica (69-343 mm ML) were separated from other onychoteuthid species based on the number of tentacular hooks (n =25-29) on each club. Identification of O. borealijaponica paralarvae (11.5 to 35 mm ML) was based on mantle chromatophore patterns (Bigelow, unpubl. data). Data collection Subadults, adults During July-September 1990, O. borealijaponica specimens were collected on vari- ous research cruises in the North Pacific. Most squid specimens were captured by research drift net ( mesh size=48-220 mm stretch mesh) in the western and central North Pacific, but squid jigs were also used to capture specimens from the central and eastern North Pacific (Fig. 1, Table 1). Squid samples were frozen (-20°C) upon capture and returned to the laboratory for analysis. Paralarvae, juveniles From 5 to 24 August 1991, 39 tows with a modified Cobb trawl were made along meridian 179°30'W between 36°56'N and 46°00'N, and along meridian 174°30'W between 39°00'N and 45°00'N. The trawl was dual warp, with a mouth area of approximately 140 m2 when fish- ing and a cod-end liner constructed of 3.2 mm knotless nylon delta mesh ( Wyllie Echeverria et al., 1990; Lenarz et al., 1991). Thirty-one oblique night tows (0-150 m) and eight oblique day tows (0-750 m> were conducted. O. borealijaponica specimens from eight tows (Fig. 1, Table 1) were sorted on board and immediately frozen (-20°C, juveniles) or fixed in 95% ethyl alcohol (paralarvae). Laboratory analysis Dorsal mantle length measurements were made to the nearest millimeter (mm) on thawed specimens. Squids less than 0.5 g were blotted dry and weighed to the nearest 0.001 g, whereas larger specimens were weighed to the nearest 0.1 g. No correction was made for shrinkage of paralarvae from fixation in ethanol, because the species possesses a strong gladius and exhibited minimal shrinkage (<2%). Specimens were sexed and assigned a maturity stage (I: juvenile; II: immature; III: preparatory; IV: maturing; V: mature) based on the appearance and relative size of the gonads and accessory repro- ductive organs (Lipinski, 1979). Statoliths were dis- sected from the specimens and stored in 95% ethyl alcohol. Statolith preparation and microstructural analysis One statolith of the pair was mounted on a microscope slide in Eukitt resin (Calibrated In- struments Inc. 200 Saw Mill Rd., Hawthorne, NY 10532) with the concave side (anterior) facing up. The transparency of paralarval statoliths allowed their examination without further preparation (Fig. 2). The thickening of statoliths from larger squid (>35 mm ML) required that they be ground with fine-grained (1200-grade) carborundum paper and polished with 0.3-|am alumina-silica powder prior to microstructural examination. Increments were counted beginning at the first visible increment outside the nucleus (Fig. 3A), and continued to the margin of the dorsal dome (Fig. 3B). The diameter of the circular nucleus averaged 28.0 urn (SD=2.4 |im, n-37). The precision of increment counts was assessed by using the coefficient of varia- tion (Chang, 1982). Two nonconsecutive blind incre- ment counts were made on each statolith with trans- mitted light at a magnification of 1500x. The mean of the two increment counts was accepted if the co- efficient of variation was <7.0%, otherwise a third count was conducted. With this criteria, two incre- ment counts were acceptable for 115 statoliths, whereas three increment counts were required for 11 statoliths. Hatching dates were computed by subtracting the mean increment count from the date of capture and were pooled into monthly periods. Increment counts were assumed to represent the individuals' age in days, based on the following re- sults (periodicity of increment deposition) which provided support for the hypothesis that one incre- ment is deposited per day. Bigelow Age and growth of Onychoteuthis boreahjaponica 15 Figure 1 Location of stations in the North Pacific sampled for Onychoteuthis boreahjaponica: Western North Pacific 1990 (open circles), central North Pacific 1990 (closed circles), central North Pacific 1991 (closed triangles), and east- ern North Pacific 1990 (open squares). Periodicity of increment deposition Three sub- adult squid caught by jig or trawl in the central North Pacific were placed for two hours in 20 L of seawater containing 250 mg/L oxytetracycline hy- drochloride (OTC). After OTC exposure, squid were maintained in a 20-L tank with flowthrough sea- water under ambient photoperiod and temperature conditions. Freshly captured live saury (Cololabis saira) were introduced as prey, but no feeding was noted or observed. Squids survived up to 61.5 hours in captivity. Statoliths were prepared as above and illuminated with ultraviolet (Fig. 4) and natural light. Under fluorescent light, an ocular marker was aligned with the inner edge of the OTC band. The statolith was then examined under natural light, but increments peripheral to the band were difficult to count. Therefore, to determine the periodicity of increment deposition, statolith growth following OTC exposure was related to the average increment width prior to exposure. The distance from the in- ner edge of the OTC band to the statolith perimeter was divided by the mean width of increments prior to the OTC band. Three estimates of statolith growth after OTC exposure were made, and the average increment width calculated for 15 incre- ments prior to the OTC band. Statistical procedures Mantle length-weight relationships Mantle length-weight regressions were fit to the data by using the model WT(g) = a*ML(mm)b (D Separate ML-weight equations were developed for both sexes, and a single equation was used for squid of unknown sex (<60 mm ML). Fitting of size-at-age data Researchers have used a variety of different models to describe cephalopod growth (e.g., linear, logistic, von Bertalanffy), al- though the rationale for using a given model is usu- ally not stated. Schnute ( 1981 ) proposed a flexible four- parameter model to describe growth which includes most growth models historically used in fisheries re- search as special cases. The model takes the form Y(t): V, +(Vi v ): 1-e-""-'1' -.j U,-r, ) l/i (2) 16 Fishery Bulletin 92(1), 1994 Table 1 Data on samples of Onychoteuthis borealijaponica collected for age analysis Mantle Depth Temperature length Date Lat. Long. Gear (m) CO n (mm) Western North Pacific 24 Jul. 1990 42-00'N 158°58'E Driftnet 0-10 14.9 5 197-316 25 Jul. 1990 43'03'N 158'59'E Driftnet 0-10 15.5 12 203-311 26 Jul. 1990 44'02'N 158a56E Driftnet 0-10 16.1 14 214-343 28 Jul. 1990 44'00'N 160"00'E Driftnet 0-10 16.5 15 206-339 29 Jul. 1990 43'16'N 159'58'E Driftnet 0-10 15.7 8 204-233 06 Aug. 1990 43'30'N 161"02'E Driftnet 0-10 16.0 2 275-288 20 Sep. 1990 44"45'N 160=03'E Driftnet 0-10 15.7 Total 1 57 182 Central North Pacific SAMPLE A 08 Jul. 1990 42'30'N 172°32'W Driftnet 0-8.5 14.8 2 165-195 04 Aug. 1990 46"30'N 152U30W Driftnet 0-8.5 12.1 4 147-180 10 Aug. 1990 46'30'N 157'30'W Driftnet 0-8.5 11.8 7 191-313 12 Aug. 1990 43'29'N 157'27'W Driftnet 0-8.5 14.3 1 343 SAMPLE B 06 Aug.1991 37"59'N 179'28'W Cobb 0-154 11.7-24.1 5 11.5-32 06 Aug.1991 37'55'N 179-26'W Cobb 0-158 11.7-24.1 H 24-35 09 Aug.1991 41°08'N 179"30'W Cobb 0-130 11.0-20.3 1 42 12 Aug.1991 43°12'N 179"30'W Cobb 0-775 3.5-16.4 1 58 12 Aug.1991 43°04'N 179"30'W Cobb 0-156 8.6-15.9 0 69-83 15 Aug.1991 44"59'N 179°27'W Jig 0-5 12.6 7 119-190 18 Aug.1991 45°00'N 174031'W Cobb 0-162 6.8-13.2 4 75-82 20 Aug.1991 43'00'N 174°30'W Cobb 0-142 8.8-16.5 2 72-78 22 Aug.1991 41'14'N 174"29'W Cobb 0-730 5.6-21.1 Total 1 49 66 Eastern North Pacific 18 Aug. 1990 42°47'N 125°25'W Jig II 100 15.1 5 214-251 19 Aug. 1990 44'12'N 124"54'W Jig 0-100 15.9 2 229-236 04 Sep. 1990 44°23'N 124'44'W Jig 0-75 16.4 Total 13 20 218-312 where Y(t) is the estimated length or weight at age t, andy1 and v., represent size at two ages tx and t.„ which are typically the youngest and oldest indi- viduals in the sample. The estimated parameters a and b describe how the model connects y; and y2. Values of a and b and their 95% confidence inter- vals lead to the selection of other submodels. The Schnute model (written in Microsoft Quickbasic) was fit to the size-at-age data (Fig. 5) by nonlinear regression on an IBM-compatable mi- crocomputer. Growth modelling was restricted to individuals from the central North Pacific samples, because of inadequate age representation from the western and eastern North Pacific samples. Paralarval size-at-age estimates were included in the growth models for males and females, because size-at-age results were similar for juvenile (66-83 mm ML) males and females. Model comparison If we assume that the Schnute model exactly predicts the size of an individual, then the residual sum of squares (RSS) of this full model is an estimate of measurement error. To ascertain if a reduced model with fewer parameters (e.g., 2- parameter exponential) adequately describes the data, the RSS's from the reduced model and full model were compared using an F test statistic: ( RSSR - RSSF )/( DFh - DFF ) RSSf/DFF f- with DFR - DFF,DFF degrees of freedom. Bigelow Age and growth of Onychoteuthis borealijaponica 17 Figure 2 Onychoteuthis borealijaponica. Light micrograph of a transverse section of a sta- tolith from a 11.5-mm mantle length paralarva. Duplicate increment counts were 61 and 63. 40 im 40 urn '" Figure 3 Onychoteuthis borealijaponica. Light micrographs of a ground statolith. (A) Increment deposition within early life history. Arrow indicates edge of nucleus. (B) Statolith microstructure within dorsal dome region. Fishery Bulletin 92(1). 1994 Figure 4 Onychoteuthis borealijaponica. UV micrograph of a ground statolith stained with tetracycline. where RSS,, is the RSS from the full (Schnute) model, RSSR is the RSS from the reduced (exponential) model, DFF is the number of degrees of freedom from the full model, and DFR is the number of degrees of freedom from the reduced model (Neter et al., 1985). Differences in the slopes of the ML- weight and size-at-age relationships by sex and geographical location were compared with analysis of covariance (ANCOVA) and F-tests (Sokal and Rolff, 1981). Data were initially In-transformed, and ANCOVA was used to test for differences in slopes of the linearized equations. Elevations of the lin- earized equations were compared with F- tests. Analyses were performed on central North Pacific male and female growth data and western North Pacific female data with the assumption that females in the western North Pacific exhibited a similar type of growth as individuals in the central North Pacific. There were too few individu- als to test for differences in growth rates 300 -i 500 -| 250 - Males 400 - Males 200 - 300 - 150 - 100 - 200 - E E 50 - 100 - X - Z 0 50 100 150 200 250 300 350 400 450 t 0 50 100 150 200 250 300 350 400 450 J o £ 400 -, 1000 n | 350 - Females Females 300 - 750 - 250 - 200 - 500 - 150 - 100 - 250 - 50 - 0 50 100 150 200 250 300 350 400 450 0 50 100 150 200 250 300 350 400 450 AGE (days) AGE (days) Figure 5 Relation between age (determined by number of increments within statoliths) and mantle length (mm) and weight (g) for male and female Onychoteuthis borealijaponica. Western North Pacific 1990 (open circles), central North Pacific 1990 (closed circles), central North Pacific 1991 (closed triangles = juveniles-subadults, open triangles = unknown sex), and eastern North Pacific 1990 (open squares). Bigelow Age and growth of Onychoteuthis borealijaponica of western North Pacific males or eastern North Pa- cific males and females. Results Statolith analysis Statolith microstructural analysis was applied to 131 squid from the western, central, and eastern North Pacific. Five statoliths (3.8%) were broken or poorly sectioned and excluded from further analy- sis. The coefficient of variation about the mean for the aged samples (n = 126) averaged 3.7% based on 2-3 increment counts for each statolith. No obvious trend existed in the coefficient of variation with the increment count or body size. Periodicity of increment formation A fluorescent OTC band was evident in the sta- toliths of the three squid exposed to oxytetracycline. While increments peripheral to the inner edge of the OTC band could not be reliably counted, the rela- tion between the growth of the statolith, rearing period, and the width of increments prior to the OTC band suggested that increments were deposited daily (Table 2). Statolith growth in the dorsal dome region ranged from 1.4 to 4.3 urn over the rearing period (26-61.5 hr). The average number of incre- ments deposited per day after oxytetracycline expo- sure was 1.30 (range 1.08-1.52) for the three squid. Mantle length-weight relationships The ML-weight relationship for paralarval O. borealijaponica from the central North Pacific is represented by the equation WT = 2.484 x 10"5 ML 3015;fl2 = 0.99(n = 36). (4) The ML-weight relationships for juvenile-adult O. borealijaponica from the western, central, and east- ern North Pacific are represented by the following equations: males: WT = 1 .873 x lO^4 ML 2596;J?2 = 0.96(n = 43) (5) females: WT = 3.521 x \0- ML2SX5;R2 = 0.99(n = 68) (6) The slopes of the ML-weight regressions for male and female O. borealijaponica were significantly different (P<0.001). Growth A good relationship existed between the number of increments within statoliths and squid size for in- dividuals in the central North Pacific (Fig. 5). An exponential model (Table 3, Equation 7) was appro- priate to describe the ML-at-age relationship (f-1.82, F=2A9) for females (paralarvae-subadult) in the central North Pacific. A logistic model was ap- propriate to describe the ML-at-age relationship (f=1.85, f=2.93) for males (paralarvae-adult) in the central North Pacific. However, the oldest individual (394 days, 245 mm ML) was a mature male (stage V) which influenced the type of model selected. Omitting that individual resulted in the selection of an exponential model (f=2A9, F=2.55) over a logis- tic model (/"=4.73, F=2.94) to describe paralarval- subadult growth (Table 3, Equation 8). Exponential models were also fit to weight-at-age data for paralarval-subadult males and females (Table 3, Equations 9 and 10). Growth in length (% increase in length per day) was similar for males and females (0.80% ML/day) in the central North Pacific, while growth in weight was faster for females (1.90% WT/day) than males (1.40% WT/day). By using the exponential models, mantle length, weight-at-age, and absolute growth Table 2 Age validation information for Onychoteu this b orealijaponica with oxytetracycline (OTC) tech nique. Width of oxytetracycline band is the distance observed between the fluorescent band and the margin of the statolith. Mean increment width is that of the outer 15 increments formed prior to the OTC band. Width of Estimated Rearing oxytetracycline Mean increment increments No. ML Imm) period (hr) band (pm) width (|im) per day 1 162 26 1 1 1.19 1.08 2 166 61.5 4.3 1.10 1.52 3 175 44 L'K 1.16 1.32 20 Fishery Bulletin 92(1), 1994 Exponential e Pacific Ocean quati ons for growth of ma e and fe Table 3 male Onychoteuthis borea lijaponica from the central North Variable Age interval (d) n Equation r2 Equation no. Length (F) Length (M) Weight (F) Weight (M) 62-376 62-314 62-376 62-314 36 27 36 27 mm = I8.41e000785t mm = I7.17e000798t g = 0.74e00188t g = 2.19e001381 0.97 0.89 0.92 0.82 7 8 9 10 rates (AGR, mm/day or g/day) were predicted for the initial 365 days (Table 4). The slopes of the size-at-age regression equations for females from the western North Pacific were significantly different from those for both central North Pacific males and females (Fig. 6, Table 5). Comparisons of regression slopes between central North Pacific males and females revealed no signifi- cant differences in length or weight-at-age relation- ships (P=0.424, P=0.307). Testing of elevations from the central North Pacific male and female data iden- tified a significant difference (P<0.001, Table 5); therefore, males and females in the central North Pacific grow in length and weight at a similar rate, but females display a significantly greater size at age than males (Table 4). Back-calculated hatching dates Backcalculation of hatching dates demonstrated that O. borealijaponica hatched in all months except March (Fig. 7). The distribution of hatching dates was not necessarily related to spawning intensity, as more subadult squid were available for age analy- sis than paralarvae and juveniles. Subadult and adult squid captured from July to September in the North Pacific had similar hatch dates as samples collected from the western (August-February), cen- tral (July-February), and eastern North Pacific (Au- gust-November). Paralarval and early juvenile squid captured in the central North Pacific during August 1991 were estimated to have hatched be- tween February and June, 1991. Maturity stage-age relationships Maturity stages were closely related to squid size for all three sampling areas; males, however, matured at a smaller size than females (Fig. 8). Females and males recruit to the driftnet fishery after attaining maturity stages III and IV, respectively. No mature females (stage V) were captured by any sampling Table 4 Growth of cen tral North Pacific Onychoteuth is borealijapor ica pred cted by the exponential equations based on statolith analysis Ab solute grow th rates (AGR) are given in mm or g per day. Estimated age Mi les Females Mantle length Weight Mantle length Weight (days) (mm) AGRL (g) AGRW (mm) AGR, (g) AGRW 50 25.6 0.20 4.4 0.06 27.3 0.21 1.9 0.04 75 31.2 0.25 6.1 0.09 33.2 0.26 3.0 0.06 100 38.1 0.31 8 7 0.12 40.3 0.32 4.8 0.09 125 46.5 0.37 12.2 0.17 49.1 0.39 7.7 0.15 150 56.8 0.46 17.3 0.24 59.7 0.47 12.4 0.24 175 69.4 0.56 24.3 0.34 72.7 0.57 19.8 0.38 200 84.7 0.68 34.3 0.48 88.4 0.70 31.7 0.60 225 103.4 0.83 48.5 0.67 107.5 0.85 50.8 0.96 250 126.2 1.01 68.4 0.95 130.8 1.03 81.3 1.54 275 154.1 1.23 96.5 1.34 159.2 1.25 130.0 2.47 300 188.1 151 136.1 1.89 193.7 1.53 ■J IIS II 3.95 325 229.6 1.84 192.1 2.66 235.7 1.86 332.8 6.31 350 280.3 2.24 271.0 3.75 286.7 2.26 532.5 1(1 K) 365 316.0 2.57 337.3 4.66 323.2 2.51 706.9 13.42 Bigelow: Age and growth of Onychoteuthis borealijaponica 8" Males - Central North Pacific E , Females - Centra] North Pacific E 7" Females - Western North Pacific I H 6- , ..-5 ' o .-- — ■" *- z S^' 3 s- ^S UJ ST J jS H «- s' Z ^0< < ^^ 2 3- ^ _c 8 - 7 - — 6 - --"'"'*"/ ' M H 5" = 4- O nj 3 - // * :- c — i - 0 - -1 - / ) 50 100 150 200 250 300 350 400 450 AGE (days) Figure 6 Log-lin ear growth models for male and female Ony- choteut his borealijaponica . method. There was some evidence that males (stage IV-V) and females (stage III— IV) in the western North Pacific were younger than similar stage in- dividuals from the central and eastern North Pacific. Discussion The data presented provide support for the one-in- crement-deposited-per-day hypothesis within the statoliths of Onychoteuthis borealijaponica although further work is required to rigorously test the hy- pothesis. Tetracycline was incorporated into the sta- tolith, but the animals did not feed and survival was not sufficiently long enough (2-3 days) to provide a rigorous test on the rate of increment deposition. Validation of the daily increment hypothesis has come from tetracycline labeled statolith experiments with several neritic squid species (Illex illecebrosus, Dawe et al., 1985, Alloteuthis subulata, Lipinski, 1986, Todarodes pacificus, Nakamura and Sakurai, 1991). Future statolith validation experiments with 20 - Western North Pacific suhadults/adulls 15 - 10 - 5 - ^^ Central North Pacific 0! 15 - m S io- Z 5- 0 - 20 - ^ — subadults/adulls ■1- □- - paralarvae/j uve rules Eastern North Pacific subadulisyadulls 15 - 10 - M 5 - _^Hb JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC MONTH OF HATCHING Figure 7 Estimated hatch-date distributions by month for Ony- choteuthis borealijaponica. Table 5 Comparisons of Onychote uthis borealijaponica growth equations based on analysis of covariance. Slope Elevation (F) (F) P Length-at-age comparisons Central North Pacific, male vs. female 0.647 0.424 1002.4 <0.001 Western North Pacific female vs. central Pacific female 62.9 <0.001 Western North Pacific female vs. central Pacific male 62.2 <0.001 Weight-at-age comparisons Central North Pacific, male vs. female 1.063 0.307 1121.1 <0.001 Western North Pacific female vs. central Pacific female 65.9 <0.001 Western North Pacific female vs. central Pacific male 66.6 <0.001 22 Fishery Bulletin 92(1), 1994 V Males 0 B IV t* 111 • II • 1 , i ' i ■ i Males H3— -B- 100 Females 200 Mil) Females 200 did Kill 500 100 200 300 MANTLE LENGTH (mm) )(i( i 200 300 AGE (days) Figure 8 Relationship between mantle length and number of increments in sta- toliths and male and female maturity stage: western North Pacific (open circles), central North Pacific (closed circles), and eastern North Pacific (open squares). Ranges are represented by horizontal bars. active oceanic squids (e.g., Onychoteuthidae, Ommastrephidae) may require substantial mainte- nance facilities to support long-term survival. Although the rate of increment deposition derived by the statolith marking experiment should be con- sidered preliminary, indirect evidence was obtained to suggest that increments were formed daily. The hypothesis that the lifespan is 1 year (Murata and Ishii, 1977; Naito et al., 1977b) was supported by the present data where only 4 of the 126 individuals aged had more than 365 increments within the sta- tolith. In addition, back-calculated hatch dates (July-February) of post-recruit individuals exploited in the O. borealijaponica jig and Ommastrephes bartramii squid driftnet fishery were consistent with information on spawning (fall-winter) reported in the literature (Murata et al., 1976; Murata and Ishii, 1977; Naito et al., 1977b). This study suggests that spawning for O. borealijaponica occurs year round. While subadult O. borealijaponica are distributed in subarctic waters, evidence from the distribution of paralarvae, juveniles, and sexually mature females suggests that spawning may occur to the south of the subarctic boundary in the North Pacific transi- tion zone (30^42°N, terminology after Roden, 1991). In the central and eastern North Pacific, O. borealijaponica paralarvae and juveniles have been recorded from this study (38°N, 179°30'W°) and the coast of California (~33°N, Young, 1972), respectively. In the western North Pacific, spawning may oc- cur in waters of the Kuroshio Cur- rent and Kuroshio Countercurrent (Murata and Ishii, 1977; Naito et al., 1977a) or between the Kuro- shio and Oyashio fronts. Onycho- teuthid paralarvae have been cap- tured from both the Kuroshio Cur- rent and Kuroshio Countercurrent (Okutani, 1968, 1969, 1975); how- ever, distributional evidence is in- conclusive because of the taxo- nomic uncertainties of the speci- mens captured. Spawning may occur in the transitional area be- tween the Kuroshio and Oyashio fronts, as sexually mature and copulated females have been cap- tured off Hokkaido, Japan (42°30TSJ, 150°40'E and 42°15'N, 144°25'E, Murata et al., 1981). The ML-weight relationships obtained in this study for the western, central, and eastern North Pacific were similar to the values previously given for O. borealijaponica cap- tured off Japan (Murata and Ishii, 1977). Slope val- ues obtained for the ML-weight relationships (males=2.596, females=2.915) were similar to other active oceanic squids having thick muscular mantle walls. Paralarval O. borealijaponica had a higher slope value (3.015) than older males and females, consistent with previous results for loliginid squids and benthic octopods (Forsythe and Van Heukelem, 1987). There is no clear consensus on the type of model which best describes cephalopod growth, although several studies argue against the use of asymptotic models, such as Gompertz or von Bertalanffy (Forsythe and Van Heukelem, 1987; Saville, 1987). Exponential models have been typically used to describe the growth of field caught and laboratory reared paralarval squid (Yang et al., 1986; Balch et al., 1988; Forsythe and Hanlon, 1989; Bigelow, 1992, 1993). For growth estimates derived from statolith analysis, a linear model is frequently used because growth is analyzed over a short segment of the cephalopod's life history, such as post recruitment to a fishery (Rosenberg et al., 1980; Radtke, 1983; Rodhouse and Hatfield, 1990b) or habitat (Jackson and Choat, 1992). Since the Schnute model encompasses a wide range of growth models, it can be used to system- Bigelow Age and growth of Onychoteuthis borealijaponica 23 atically assess the type of growth model which best describes the data. A statistical comparison of sev- eral growth models found that growth in O. borealijaponica from the paralarval to subadult size range could be sufficiently described with an expo- nential model, though there was weak evidence that a logistic model may be sufficient to describe growth in males from the paralarval to adult size range. The most appropriate growth model (exponential or lo- gistic) for the entire life cycle of O. borealijaponica will emerge when sexually mature males and fe- males are aged. Estimated growth rates from this study were higher than estimates derived from length-fre- quency analysis of fisheries data (Murata and Ishii, 1977). Growth estimates based on length-frequency analysis with time often provide evidence of de- creased growth rate, which is usually described by an asymptotic model (Patterson, 1988). Length-fre- quency analysis may be inappropriate for estimat- ing growth in cephalopods (Jackson and Choat, 1992), either because 1) cohorts are difficult to de- tect because spawning occurs throughout the year, 2) variable individual growth rates produce Lee's phenomenon (Ricker, 1975), or 3) samples of a mi- grating population are taken at a point along the migration route, which results in overestimating growth in young squid and underestimating growth in older squid. Growth data presented for O. borealijaponica from the central North Pacific provide a useful compari- son of growth between males and females. The ex- ponential models predict that males and females grow in length at similar rates (0.80% ML/day), but females grow faster in weight (1.90% WT/day) than do males (1.40% WT/day). These rates correspond closely with the average growth rates of similar sized squids from temperate waters (e.g., Illex illecebrosus, O'Dor, 1983; /. argentinus, Rodhouse and Hatfield, 1990b). The most significant advantage of using statolith ageing techniques is the ability to produce indi- vidual rather than population statistics. Using sta- tolith analysis, spatial variations in size at age, growth parameters, and maturity stage at age were observed between O. borealijaponica individuals from the western and central North Pacific. Little is known concerning genetic variation and stock structure of O. borealijaponica in the North Pacific; however, female squid in the western North Pacific were found to grow faster than both male and fe- male squid in the central North Pacific and were younger at maturity stages III and IV than central North Pacific females. Apparent growth rate and maturity stage differences may be related to water temperatures or food availability during the paralarval stage. Forsythe and Hanlon (1989) showed that temperature had a pronounced effect on the increase in length and weight of the squid Loligo forbesi. In their laboratory study, a tempera- ture increase of 1°C increased the growth in length and weight of paralarval squid 0.5% and 2.0% per day, respectively. Subadults in the western Pacific may have hatched in the warm Kuroshio Current or in productive transition waters between the Kuroshio and Oyashio fronts. Paralarvae hatched in the western North Pacific may therefore experience higher temperatures or a greater abundance of prey species, or both, than paralarvae hatched in the central North Pacific, which could explain the ob- served spatial differences in growth. Acknowledgments I gratefully acknowledge the help of the officers and crew of the research vessels Hai Kung, and Shoyu Maru and the help of the officers, crew, and scien- tific field party of the NOAA ship Townsend Crom- well cruise 91-06. 1 would like to thank C. H. Fiscus who kindly provided the statolith samples from the eastern North Pacific and D. R. Kobayashi for assis- tance in fitting the Schnute model. This paper benefitted from comments by G. T. DiNardo, E. E. DeMartini, C. H. Fiscus, and anonymous reviewers. Literature cited Balch, N., A. Sirois, and G. V. Hurley. 1988. Growth increments in statoliths from paralarvae of the ommastrephid squid lllex (Cephalopoda: Teuthoidea). Malacologia 29:103- 112. Bigelow, K. A. 1992. Age and growth in paralarvae of the meso- pelagic squid Abralia trigonura based on daily growth increments in statoliths. Mar. Ecol. Prog. Ser. 82:31-40. 1933. Hatch dates and growth of Ommastrephes bartramii paralarvae from Hawaiian waters as determined from statolith analysis. In T. Okutani (ed.), Proceedings of the international symposium on the recent advances in cephalopod fisheries biology, p. 15-24 University Press, Tokai, Japan. Chang, W. Y. B. 1982. A statistical method for evaluating the repro- ducibility of age determination. Can. J. Fish. Aquat. Sci. 39:1208-1210. Dawe, E. G., R. K. O'Dor, P. H. Odense, and G. V. Hurley. 1985. Validation and application of an ageing tech- 24 Fishery Bulletin 92(1). 1994 nique for short-finned squid (lllex illecebrosus) J. Northwest Atl. Fish. Sci. 6:107-116. Fiscus, C. H., and R. W. Mercer. 1982. Squids taken in surface gillnets in the North Pacific Ocean by the Pacific salmon investigations program 1955-72. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-FNWC-28, 31 p. Forsythe, J. W., and W. F. Van Heukelem. 1987. Growth. In P. R. Boyle (ed.), Cephalopod life cycles, vol. II, p. 135-156. Academic Press, Lon- don. Forsythe, J. W., and R. T. Hanlon. 1989. Growth of the eastern Atlantic squid, Loligo forbesi Steenstrup (Mollusca: Cephalopoda). Aquat. Fish. Manage. 20:1-14. Jackson, G. I)., and J. H. Choat. 1992. Growth in tropical cephalopods: an analysis based on statolith microstructure. Can. J. Fish. Aquat. Sci. 49:218-228. Kubodera, T. 1986. Relationships between abundance of epipe- lagic squids and oceanographical-biological envi- ronments in the surface waters of the subarctic Pa- cific in summer. Int. North Pac. Fish. Comm. Bull. 47:215-228. Kubodera, T., W. G. Pearcy, K. Murakami, T. Kobayashi, J. Nakata, and S. Mishima. 1983. Distribution and abundance of squids caught in surface gillnets in the subarctic Pacific, 1977- 1981. Mem. Fac. Fish. Hokkaido Univ. 30(1/2): 1-49. Lenarz, W. H., R. J. Larson, and S. Ralston. 1991. Depth distributions of late larvae and pelagic juveniles of some fishes of the California current. CalCOFI Rep. 32:41-46. Lipinski, M. 1979. Universal maturity scale for the commer- cially important squids. The results of maturity classification of the lllex illecebrosus population for the years 1973-77. ICNAF Res. Doc. 79/2/38, Serial 5364, 40 p. 1986. Methods for the validation of squid age from statoliths. J. Mar. Biol. Assoc. U.K. 66:505-525. Murata, M., and M. Ishii. 1977. Some information on the ecology of the oce- anic squid, Ommastrephes bartrami (Lesueur), and Onyehoteuthis borealijaponicus Okada, in the Pacific Ocean off Northeastern Japan. Bull. Hokkaido Reg. Fish. Res. Lab 42:1-23. (In Japa- nese; English abstract.) Murata, M., T. Ishii, and H. Araya. 1976. The distribution of the oceanic squids, Ommastrephes bartramii (Lesueur), Onyehoteuthis borealijaponicus Okada, Gonatopsis borealis Sasaki and Todarodes pacificus Steenstrup in the Pacific Ocean off north-eastern Japan. Bull. Hokkaido Reg. Fish. Res. Lab 41:1-29. (In Japa- nese; English abstract.) Murata, M., M. Ishii, and M. Osako. 1981. Some information on copulation of the oce- anic squid Onyehoteuthis borealijaponica Okada. Bull. Jap. Soc. Sci. Fish. 48:351-354. (In Japanese; English abstract.) Naito, M., K. Murakami, T. Kobayashi, N. Nakayama, and J. Ogasawara. 1977a. Distribution and migration of oceanic squids (Ommastrephes bartramii, Onyehoteuthis boreali- japonicus, Berryteuthis magister, and Gonatopsis borealis) in the western subarctic Pacific region. Spec. Vol., Res. Inst. N. Pac. Fish., p. 321-337. (In Japanese; English abstract.) (Engl, transl. by W. G. Van Campen, 1991, 23 p., Trans- lation No. 144; available Honolulu Lab., Southwest Fish. Sci. Cent., Natl. Mar Fish. Serv., NOAA, 2570 Dole St., Honolulu, HI 96822-2396.) Naito, M., K. Murakami, and T. Kobayashi. 1977b. Growth and food habit of oceanic squids (Ommastrephes bartramii, Onyehoteuthis boreali- japonicus, Berryteuthis magister and Gonatopsis borealis) in the western subarctic Pacific region. Spec. Vol., Res. Inst. N. Pac. Fish., p. 339-351. (In Japanese; English abstract.) (Engl, transl. by W. G. Van Campen, 1991, 19 p.. Trans- lation No. 147; available Honolulu Lab., Southwest Fish. Sci. Cent., Natl. Mar Fish. Serv, NOAA, 2570 Dole St., Honolulu, HI 96822-2396.) Nakamura, Y., and Y. Sakurai. 1991. Validation of daily growth increments in sta- toliths of Japanese common squid Todarodes pacificus. Nippon suisan Gakkaishi 57(111:2007- 2011. Natsukari, Y., E. Dawe, and M. Lipinski. 1991. Interpretation of data. Practical procedures of squid ageing using statoliths. A laboratory manual. Section 2. In P. Jereb, S. Ragonese, S. von Boletzky (eds.). Squid age determination us- ing statoliths. Proc. Internat. Workshop, Instituto di Tecnologia della Pesca e del Pescato (ITPP-CNR), Mazara del Vallo, Italy, 9-14 Oct. 1989. N.TR.-I.TP.P. Spec. Publ. 1 Neter, J., W. Wasserman, and M. H. Kutner. 1985. Applied linear statistics, 2nd ed. Richard D. Irwin, Inc., IL, 1127 p. O'Dor, R. K. 1983. lllex illecebrosus. In P. R. Boyle, (ed.), Cephalopod life cycles, vol. I, p. 175-199. Aca- demic Press, London. Okutani, T. 1968. Studies on early life history of decapodan Mollusca-III. Systematics and distribution of lar- vae of decapod cephalopods collected from the sea surface on the Pacific coast of Japan 1960- 1965. Bull. Tokai Reg. Fish. Res. Lab. 55:9-57. 1969. Studies on early life history of decapodan Mollusca- IV. Squid larvae collected by oblique hauls of a larva net from the Pacific coast of east- ern Honshu, during the winter seasons. Bull. Tokai Reg. Fish. Res. Lab. 58:83-96. 1975. Studies on early life history of decapodan Mollusca-V Systematics and distribution of epipe- Bigelow Age and growth of Onychoceuthis boreahjaponica 25 lagic larvae of decapod cephalopods in the south- western waters of Japan during the summer in 1970. Bull. Tokai Reg. Fish. Res. Lab. 83:45-96. 1977. Stock assessment of cephalopod resources fished by Japan. FAO Fish. Tech. Paper 173:1-62. Okutani, T., and M. Murata. 1983. A review of the biology of the oceanic squid Onychoteuthis borealijaponica. Memoirs of the Natl. Mus. of Victoria 44:189-195. Patterson, K. R. 1988. Life history of Patagonian squid Loligo gahi and growth parameter estimates using least- squares fits to linear and von Bertalanffy models. Mar. Ecol. Prog. Ser. 47:65-74. Radtke, R. L. 1983. Chemical and structural characteristics of statoliths from the short-finned squid lllex illecebrosus. Mar. Biol. 76:47-54. Ricker, W. E. 1975. Computation and interpretation of biological statistics of fish populations. Bull. Fish. Res. Board Can. 91:1-382. Roden, G. I. 1991. Subarctic-subtropical transition zone of the North Pacific: large-scale aspects and mesoscale structure. In J. A. Wetherall (ed. ), Biology, ocean- ography, and fisheries of the North Pacific transi- tion zone and subarctic frontal zone, p. 1- 38. NOAATech. Rep. NMFS 105, 111 p. Rodhouse, P. G., and E. M. C. Hatfield. 1990a. Age determinations in squid using statolith growth increments. Fish. Res. (Amst.) 8:323-334. 1990b. Dynamics of growth and maturation in the cephalopod lllex argentinus de Castellanos, 1960 (Teuthoidea: Ommastrephidae). Philos. Trans. R. Soc. Lond. B. Biol. Sci. 329:229-241. Rosenberg, A. A., K. F. Wiborg, and I. M. Bech. 1980. Growth of Todarodes sagittatus (Lamarck) (Cephalopoda, Ommastrephidae) from the north- east 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 P. R. Boyle (ed.), Cephalopod life cycles, vol. II, p. 277-290. Academic Press, London. Schnute, J. 1981. A versatile growth model with statistically stable parameters. Can. J. Fish. Aquat. Sci. 38:1128-1140. Sokal, R. R., and F. J. Rohlf. 1981. Biometry. W.H. Freeman, San Francisco. Wyllie Echeverria, T., W. H. Lenarz, and C. A. Reilly. 1990. Survey of the abundance and distribution of pelagic young-of-the-year rockfishes, Sebastes, off central California. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-SWFSC-147, 125 p. Yang, W. T., R. F. Hixon, P. E. Turk, P. E. Krejci, W. H. Hulet, and R. T. Hanlon. 1986. Growth, behavior, and sexual maturation of the market squid, Loligo opalescens, cultured through the life cycle. Fish. Bull. 84:771-797. Young, R. E. 1972. The systematics and areal distribution of pe- lagic cephalopods from the seas off southern California. Smithsonian Contr. Zool. 97:1-159. Young, R. E., and R. F. Harman. 1987. Descriptions of the larvae of three species of the Onychoteuthis banksii complex from Hawaiian waters. Veliger 29(31:313-321. AbStfclCt. A review of previ- ous studies on Kemp's ridley sea turtle (Lepidochelys kempii) diet was combined with information on the diet of the species in the coastal waters of New York State. Juvenile Kemp's ridleys occupy coastal Long Island, New York waters during the summer and early autumn months. Both fecal and intestinal samples collected between 1985 and 1989 were ana- lyzed to obtain information on the diet of this endangered species. Fecal and intestinal sample analy- sis, as well as information from previous studies, indicated that juvenile Kemp's ridleys primarily consume crabs. Walking crabs of the genera Libinia and Cancer appear to be the primary food sources for the species in New York waters. Diet of the Kemp's ridley sea turtle, Lepidochelys kempii, in New York waters Vincent J. Burke Savannah River Ecology Laboratory and Department of Zoology University of Georgia, Drawer E Aiken. SC 29802 Stephen J. Morreale Center for the Environment, Room 200, Rice Hall Cornell University, Ithaca, NY 14853-5601 v. Edward A. Standora Department of Biology, State University College at Buffalo 1 300 Elmwood Avenue, Buffalo, NY 1 4222 Manuscript accepted 29 September 1993 Fishery Bulletin 92:26-32 (1994) The Kemp's ridley sea turtle, Lepidochelys kempii, was placed on the United States endangered spe- cies list in December 1970 and was listed as one of the twelve most endangered species in the world by the International Union for the Conservation of Nature and Natu- ral Resources in 1986 (Federal Reg- ister, 1989; Marine Turtle Newslet- ter, 1989). Despite a recent in- crease in research on the Kemp's ridley, little attention has been fo- cused on its feeding habits. An un- derstanding of the dietary require- ments and available food resources for the Kemp's ridley is a critical com- ponent in the future management and protection of this species' habitats. While occasional glimpses into the composition of Kemp's ridley diets have been obtained, detailed quantified examinations of the spe- cies' diet have only rarely been undertaken (Table 1). In one of the earliest accounts of the diet of Kemp's ridleys, De Sola and Abrams (1933) dissected "two foot specimens" from the Georgia coast and described the main dietary component as Platyonichus ocel- latus, later renamed the spotted lady crab, Ovalipes stephensonii (Williams, 1984). Two decades later, the first pub- lished record describing the diet of the Kemp's ridley in the Gulf of Mexico was produced (Liner, 1954). In that study, gastrointestinal con- tents of eight L. kempii ranging in size from 3.2 kg to 26.6 kg were examined. All the turtles had con- sumed portunid crabs iCallinectes sp.) and occasional barnacles. Dobie et al. (1961), elaborating on the findings of Liner (1954), re- ported that small molluscs, plant parts, and mud were also contained in the gastrointestinal tracts of two of Liner's turtles. The molluscs in- cluded gastropods (Nassarius sp.) and bivalves of the genera Nuculana, Corbula, and probably Mulinia. In Virginia, Hardy (1962) dis- sected a single specimen and found that the digestive tract contained 95% Callinectes sp. and one swimmerette was identified as that of the blue crab, C. sapidus. Re- search conducted in the waters of Chesapeake Bay, Virginia, by Lutcavage (1981) indicated that three Kemp's ridley carcasses had both blue crabs and Atlantic rock 26 Burke et al.: Diet of Lepidochelys kempii 27 Table 1 Compilation of available diet stud ies of Kemp's ridley sea turtles (Lepidochelys kempii) publ shed from 1933 to 1991. The studies are ord ?red from north to south and east to west. Marquez (1973) is cited from Pritchard and Marquez (1973). Author! s) Location Diet components Life stage Hardy (1962) Chesapeake Bay Blue crabs Juvenile Lutcavage (1981) Chesapeake Bay Blue crabs Juvenile Belmund et al. (1987) Chesapeake Bay Rock and blue crabs Juvenile DeSola and Abrams (1933) Coastal Georgia Crabs [Ovalipes spp.) Not given Carr (1942) Florida Calico crabs Juvenile Liner (1954) Louisiana Blue crabs Juvenile Dobie et al. (1961) Louisiana Crabs, whelks, clams Juvenile Shaver (1991) Texas Various crab species Juvenile Marquez (1973) Tampico, Mexico Crustaceans, fish, molluscs Adult crabs (Cancer irroratus) in their digestive tracts. Recently, Shaver (1991) found that Kemp's ridleys in coastal Texas waters preyed mainly on crabs. The most commonly ingested species was the speckled crab (Arenaeus cribrarius). Many other crab species were recorded by Shaver, including purse crabs (Persephonia sp.), spider crabs (Libinia sp.), and blue crabs (Callinectes sp.). During the past decade, the role of the northeast- ern coast of the United States in the life cycle of Kemp's ridleys has received considerable attention (Carr, 1980; Morreale and Standora, 1990 '; Burke et al., 1991). The northeastern coast includes the New York area which contains over 300 km of shore- line, mainly the coastline of Long Island. Long Is- land has a variety of marine habitats, including the shallow, enclosed waters of the Peconic and south- ern bays, the deeper waters of Long Island Sound, and the Atlantic Ocean (Fig. 1). Each year Kemp's ridleys begin inhabiting the Long Island area dur- ing July (Morreale and Standora, 19892; Morreale and Standora, 19901). To date, all Kemp's ridleys encountered in Long Island have been juveniles (straight-line carapace length from 22 cm to 42 cm x=29.8 cm, SD=3.7 cm [Morreale and Standora, 19892, 19901]). This size class of turtles represents a range of ages from 3 to 7 years (Zug and Kalb, 1989). Between July and early October these young Kemp's ridleys are active within the estuarine wa- ters (Long Island Sound and the Peconic Bays) and the southern bays. Kemp's ridley growth rates as 1 Morreale, S. J., and E. A. Standora. 1990. Occurrence, move- ment and behavior of Kemp's ridley and other sea turtles in New York waters. Annual report to the New York State, Dep. Environmental Conservation, April 1989-April 1990. 2 Morreale, S. J., and E. A. Standora. 1989. Occurrence move- ment and behavior of the Kemp's ridley and other sea turtles in New York waters. Annual report to the New York State, Dep. Environmental Conservation, April 1988-April 1989. high as 25% body weight per month indicate that waters around Long Island, New York, provide abundant food resources for the maintenance and growth of the juvenile turtles (Standora et al., 1989; Burke, 1990). During October the turtles begin moving out of the estuaries and into the ocean. Long distance recaptures of Kemp's ridley, green (Chelonia mydas), and loggerhead (Caretta caretta) sea turtles tagged near Long Island indicate that some turtles emigrate to the southeastern United States (Morreale and Standora, 19892; Burke, 1990; Morreale and Standora, 19901). Kemp's ridleys that do not emigrate by late November are likely to be- come cold-stunned (Burke et al., 1991). Cold-stun- ning, or severe hypothermia, occurs when ambient water temperatures fall below 10°C (Schwartz, 1978). Cold-stunning causes turtles to become tor- pid and buoyant, and eventually results in death. In Long Island, declining water temperatures usu- ally reach 10°C during early December. The cold-stunning phenomenon, other types of strandings, and live captures of sea turtles during commercial fishing operations can be utilized as sources of turtles for dietary studies. The goal of the current study is to provide a quantitative descrip- tion of the diet of Kemp's ridleys in the northeast- ern United States based on gut contents from car- casses, previously preserved dietary samples, and feces from live turtles. Materials and methods The dietary components of the Kemp's ridley were assessed by using two separate approaches. First, fecal samples were collected from live turtles and examined for their constituents. Second, complete gastrointestinal contents were removed from dead turtles and identified. Samples were obtained from 28 Fishery Bulletin 92(1), 1994 Figure 1 The waters from which Kemp's ridley sea turtles, Lepidochelys kempii, were obtained for this study can be di- vided into four habitats: Long Island Sound, where most of the stranded turtles were recovered; the Atlantic Ocean, which was the habitat of two turtles in the study; the southern bays, where one live capture and one boat-hit turtle were recovered; and the Peconic Bay system, where most of the turtles for the fecal analysis and several turtles for the digestive tract analysis wree recovered. turtles encountered in New York waters from 1985 through 1989. Nineteen fecal samples were obtained. Fourteen were collected during 1989, three during 1988, and two during 1987. Of these fecal samples, 17 were obtained from live turtles captured during warmer months (June to October) and two samples were retrieved from revived, cold-stunned turtles in late November. Captured turtles were obtained from lo- cal commercial fishermen who were asked to retain turtles caught incidentally in fishing gear (predomi- nantly pound nets). After the fishermen docked, they called a 24-hour number to reach a biologist, who generally picked up the turtle while the fishermen were still unloading their catch. All noncold-stunned Kemp's ridleys received from commercial fisheries in Long Island were alive and apparently healthy. All turtles were weighed and measured upon re- turn to the laboratory. Each turtle was then allowed to swim freely in an individual 2100-liter tank and was offered either squid or clam meat. Most Kemp's ridleys accepted the food offerings, but many fed only after the food was dangled in front of them for as long as 2-3 hours. Feeding often induced defeca- tion within a relatively short time. Tanks were checked at least three times a day for the appearance of feces. Filter intakes in the tanks were elevated and covered, except for small holes, to insure against sample loss. When feces were ob- served, they were immediately removed and placed in individual sample jars. If a turtle did not defecate within 24 hours of being placed in captivity, it was given an enema of dioctyl sodium sulfosuccinate (Disposaject brand, Pitman-Moore Inc.). If a fecal sample was still not obtained after another 24 hours, the turtle was released. The rate of food passage was examined during this study to insure that samples were not polluted with prey items eaten while the turtles were in the fishermen's nets. Gut passage rates were deter- mined for two Kemp's ridleys by feeding them declawed lobsters (Homarus americanus). Lobster was used as a tracer because it has never been re- ported as a prey item and is consumed relatively readily by the turtles. By monitoring fecal output, the amount of time between ingestion of the lobster Burke et al.: Diet of Lepidochelys kempii 29 and its first appearance in the feces was determined. All fecal samples collected for dietary analysis were immediately placed in preservative. For fecal samples obtained during 1989, animal components were preserved as described by Zinn ( 1984) and al- gae were preserved in Transeau's solution ( 10 parts formalin/30 parts ethanol/60 parts distilled H.20/25 mg CuS04/L). Feces obtained prior to 1989 were pre- served in 10% formalin. Analysis of the fecal samples was conducted in January 1990, after all the samples were collected. The samples were removed from the preservative and air dried for 24 hours on wire mesh in an en- closed hood. The samples were then placed in a U.S. standard number-5 mesh (4 mm) sieve and pieces smaller than 4 mm were separated out by shaking the sample in a Tyler RO-TAP testing sieve shaker for three minutes. Pieces smaller than 4 mm were not identified because of the difficulty of assigning them to a meaningful category. The amount of sample lost because of this constraint was never greater than 5% for any given sample. Each fecal sample was examined under a dissect- ing microscope and each fragment of the sample was identified to the lowest taxon possible. Fragments belonging to the same taxonomic level were grouped. A list of components (e.g., one species of crab is one component) was compiled for each sample and the data were analyzed to determine the percentage of turtles in which each component occurred. Less than 1% of the fragments could not be assigned to a taxo- nomic category. For the 1989 samples only, the relative amount of each dietary component was determined by oven drying each component from each sample for 48 hours at 60°C and weighing it. The dry weights were then used to determine the relative importance of the different dietary components in each turtle's fecal sample. Dry weight analysis was conducted by finding the percentage of each sample weight rep- resented by each component and then determining the mean for that component. This technique of analyzing dry weights as a percentage eliminated over- or under-representation of large or small fe- cal samples. A second method of determining dietary compo- nents was analysis of gastrointestinal contents from stranded, dead turtles. Stranded Kemp's ridleys died from a number of causes: cold-stunning, boat colli- sions, entanglement in a gill net, and natural and unknown causes. Whenever possible, each stranded turtle was weighed, measured (straight-line cara- pace length) and dissected. Following removal, in- testinal contents were placed in 95% ethanol (1985), 10% formalin (1986-1988), or treated in the same manner as the fecal samples (1989). Identification of intestinal tract contents was performed during 1990. All components of each sample were identified to the lowest taxon possible, generally to species. These data were used to determine the percentage of turtles in which the components occurred. Results The food passage rate analysis indicated that lob- ster was retained within the digestive tracts of the two Kemp's ridleys for seven and eight days. Be- cause fecal samples were obtained within 48 hours of receiving a turtle from a fisherman, we believe the possibility of samples having been "contami- nated" by items eaten while the turtles were in the fishermen's nets is minimal. Mean straight-line carapace length for the 19 turtles in the fecal analysis study was 32.3 cm (range=24.7 to 42.7 cm, SD=4.87). Eighteen of the 19 turtles consumed crabs (Fig. 2). Mollusc species were found in 26% of the fecal samples and algae were found in 11%> of the Kemp's ridley feces. Natu- ral and synthetic debris were present in 21% and 11% of the feces respectively. Crab species that were identified included nine- spined spider crabs, Atlantic rock crabs, and lady crabs (Ovalipes ocellatus). Further examination of only the crab portion of the feces revealed that 58% of the turtles had consumed spider crabs, 36% had eaten rock crabs, and 16% had consumed lady crabs. 100 60 w 40 O - n = 19 m '.'.'. i CRAB MOLLUSK ALGAE NATURAL SYNTHETIC DEBRIS DEBRIS Figure 2 Percent occurrence of various prey items identified in the feces of 19 Kemp's ridley sea turtles (Lepidochelys kempii) that were live-captured in Long Island waters. Each bar indicates the percent of turtles in which the prey items occurred. 30 Fishery Bulletin 92(1). 1994 a. z < CRAB MOLLUSK ALGAE NATURAL SYNTHETIC DEBRIS DEBRIS Figure 3 Mean percent of the fecal dry weight of general catergories of Kemp's ridley sea turtles (Lepidochelys kempii) prey items. Each area repre- sents the mean percent of dry weight for that com- ponent of the feces (n=14). Crabs composed the pre- dominant portion of the feces. o \- z 70 - 60 - 50 - n = 18 40 30 - 20 10 - 0 - .-. i CRAB MOLLUSK NATURAL DEBRIS Figure 4 The percent of Kemp's ridley sea turtles (Lepidochelys kempii) from the digestive tract analy- sis that had consumed various types of ingesta. Most of the turtles had consumed crabs. Synthet- ics and algae were not present in the digestive tracts. Included in three fecal samples were crab parts from which the fragments could not be identified to genus. Mollusc species in the samples included blue mussels (Mytilus edulis) and bay scallops (Argopectin irradians). Two Kemp's ridley fecal samples contained mollusc fragments that could not be identified beyond phylum. Algal species in the samples included Sargassum natans, Fucus sp., and Ulua sp. A few turtles had small pieces of the mac- rophyte Zostera marina as well. Natural debris in- cluded such things as pebbles, small rocks, and bird feathers. Synthetic debris included only small pieces of polystyrene and latex. Analysis of fecal components with dry weights (mean of percent per sample) revealed that crabs were the predominant component of all but one of the 14 fecal samples from 1989. The mean percent of crab dry weight for the samples was 80% (Fig. 3). The mean percent dry weight for each crab species revealed that spider crabs composed 60% of the identifiable crab parts. The remainder was com- posed of 22% rock crabs and 18% lady crabs. Thus, most of the Kemp's ridleys had consumed spider crabs, which represented a large portion of the bulk. Although more turtles consumed rock crabs than lady crabs, Kemp's ridleys that consumed lady crabs had feces composed exclusively of them. For the period 1985 througn 1989, 87 dead Kemp's ridleys were recovered from Long Island's waters. Gastrointestinal tracts were removed from 40 of the 87 turtles. Eighteen of the 40 stranded Kemp's rid- leys contained identifiable diet components in the gut. All 18 turtles were juveniles. Mean straight-line carapace length for the 18 stranded turtles was 30.5 cm (range=24.8 cm to 39.7 cm, SD=3.5 cm). Thirteen of the 18 gastrointestinal tracks contained crab parts and seven contained mollusc shells (Fig. 4). The most frequently encountered crabs in the gut content samples were spider crabs and rock crabs. Spider crab fragments were found in five of the 18 samples; rock crabs were found in four of the 18 samples. Lady crabs were found in two of the samples and the blue crab (C. sapidus) was found in the digestive tract of one Kemp's ridley. Two of the turtles had crab parts in their digestive tracts that could not be assigned reliably to any genus. An additional 14 of the 40 Kemp's ridleys that were dissected had completely empty digestive tracts. All of these turtles had stranded from cold- stunning. Upon further review of necropsy data sheets from all of the Kemp's ridleys that had stranded during the study period, but from which samples were not preserved, it was noted that al- most all of the cold-stunned individuals had empty or almost empty gastrointestinal tracts. The remaining eight turtles had been collected in 1985 and 1986, and gut contents were unidentifiable because of improper preservation. These samples had been preserved for as long as five years prior to examination. Burke et al.: Diet of Lepidochelys kempn Discussion The analysis of fecal samples from live turtles and of gut contents from dead specimens strongly sug- gests that crabs are the main dietary component for Kemp's ridleys in New York waters. Crab parts were present in 18 of the 19 turtles from which fecal samples were obtained and were the predominant food item by dry weight analysis. The analysis of fecal material, however, may be biased because it examines only that material which has not been fully digested. This could cause overrepresentation of less digestible components. The gastrointestinal tract results (which are less susceptible to such bias) support the results of the fecal sample analysis. Of the 18 stranded turtles which contained identifiable food items, 13 con- tained crab parts in their guts. Gut contents can potentially be biased because of differential diges- tion. However, from our qualitative observation of the condition of the intestinal contents during dis- section, we believe the components described herein are representative of the diet. One difference between the fecal and intestinal samples was the source of the turtles. Most fecal samples were obtained from turtles captured in the Peconic Bays, but most stranded turtles were recov- ered on beaches adjacent to Long Island Sound. Presumably the dietary samples reflect feeding ac- tivities near the location of capture (or stranding). Thus, the observation of spider and rock crabs as the predominant components in the diets of both live- captured and stranded turtles emphasizes their importance as food items. The dietary components observed during the study may be related to the relative abundance of the prey species in the environment. Of the four species of crab that were identified, the spider crab was both the most frequently encountered fecal com- ponent and the predominant crab identified in the gut contents of dead turtles. During the course of our studies we have noted that the nine-spined spi- der crab was one of the most common crabs in the waters where the turtles occurred. We have observed local commercial fishermen retrieving thousands of spider crabs while hauling in their nets. The Atlan- tic rock crab was also frequently encountered in the feces and gut contents of the turtles. The rock crab is also abundant in many of the areas in which the turtles occur. Not all of the dietary make-up observed in this study can be explained by prey abundance. The green crab (Carcinus maenus) is very common in many of Long Island's estuaries but was not present in any of the turtles examined. This species usually inhabits shallower, rocky intertidal and subtidal habitats (Ropes, 1968; Williams, 1984), and our re- search on turtle behavior indicates that the Kemp's ridleys typically forage in deeper waters (Standora et al., 1990). While we have commonly encountered lady crabs in the waters where turtles forage, this species was represented in only a few samples. Also rare in the samples was the locally and commercially harvested blue crab. Both the lady crab and the blue crab are portunid crabs, capable of swimming very quickly. This characteristic differentiates the portunids from the slower walking crabs, such as the spider and rock crabs. The only molluscs consumed by turtles examined during this study included a few fragments of rela- tively thin-shelled blue mussels (Mytilus edulis) and bay scallops (Argopectin irradians), and entire shells of the small three-lined mud snail (Nassarius trivitattus). These mud snails are scavengers and can be found locally in association with dead fish and crabs (Long Island Shell Club, 1988). Their oc- currence in four turtles, all of which had been cold- stunned, may indicate that the turtles were scaveng- ing during periods of low water temperature. Because sea turtles were obtained from different sources in New York waters, it was possible to ob- tain dietary information on a larger number of Kemp's ridleys. In many of the previous studies presented in Table 1, portunid crabs were indicated as a main dietary component for Kemp's ridleys. Although this crab family was observed in some New York turtles, it was of secondary importance to the walking crabs. In terms of the overall life cycle of Kemp's ridleys, it appears that post-pelagic juveniles exploit the benthic environments of Long Island's estuaries, preying mainly on walking crabs. Data from our ongoing research indicate that sea turtles emigrat- ing from New York inshore waters travel to south- ern coastal areas. Kemp's ridleys exhibiting this behavior may join the more southerly portion of the Atlantic population. Therefore, management plans for Kemp's ridleys should consider factors that af- fect benthic fauna, especially the abundant crab populations in the northeastern region. Such im- pacts could have far-reaching effects on a critical stage in the lives of these endangered sea turtles. Acknowledgments This study was supported by a grant from the Na- tional Marine Fisheries Service under contract num- ber 40AANF902823. We thank Phil Williams for his encouragement and support. Long-term support for 32 Fishery Bulletin 92(1), 1994 sea turtle studies in New York was provided by the N.Y. State Dept. "Conservation's Return a Gift to Wildlife" program. Manuscript preparation was aided by contract DE-AC09-76SROO-819 between the University of Georgia and the U.S. Department of Energy. Workspace was provided by the State University College at Buffalo and the Okeanos Foundation. Turtle collection could not have been accomplished without the help of hundreds of vol- unteers and the commercial fishermen of Long Is- land, New York. We thank Anne Meylan for main- taining intestine samples and records from the years 1985 and 1986. For their efforts in collecting turtles, we thank C. Coogan, P. Logan, S. Sadove, and R. Yellin. The Long Island Shell Club donated mollusc voucher specimens. William Zitek graciously pro- vided necropsy facilities during 1985 and 1986, and veterinary advice during 1989 that allowed us to increase the number of fecal samples obtained. Literature cited Bellmund, S. A., J. A. Musick, R. C. Klinger, R. A. Byles, J. A. Keinath, and D. E. Barnard. 1987. Ecology of sea turtles in Virginia. Spec. Sci. Rep. No. 19, Virginia Institute of Marine Science, Coll. of William and Mary, Gloucester Point, Virginia. Burke, V. J. 1990. Seasonal ecology of Kemp's ridley (Lepidochelys kempi) and loggerhead (Caretta caretta ) sea turtles in the waters of Long Island, New York. Master's thesis, State University of New York, College at Buffalo, NY. Burke, V. J., E. A. Standora, and S. J. Morreale. 1991. Factors affecting strandings of cold-stunned Kemp's ridley and loggerhead sea turtles in Long Island, New York. Copeia 1991:1136-1138. Carr, A. 1942. Notes on sea turtles. Proc. New England Zoological Club 21:1-16. 1980. Some problems of sea turtle ecology. Am. Zool. 20:489-498. DeSola, C. R., and F. Abrams. 1933. Testudinata from southeastern Georgia, in- cluding the Okefinokee swamp. Copeia 1:10-12. Dobie, J. L., L. H. Ogren, and J. F. Fitzpartick Jr. 1961. Food notes and records of the Atlantic ridley turtle (Lepidochelys kempii) from Louisi- ana. Copeia. 1961:109-110. Federal Register. 1989. Endangered and threatened wildlife and plants. 50 CFR 17.11 and 17.12. Hardy, J. D. 1962. Comments on the Atlantic ridley turtle, Lepidochelys olivacea kempi, in the Chesapeake Bay. Chesapeake Science 3:217-220. Liner, E. A. 1954. The herpetofauna of Lafayette, Terrebonne and Vermilion Parishes, Louisiana. Proc. Louisi- ana Academy of Sciences. 17:65-85. Long Island Shell Club. 1988. Seashells of Long Island, New York. Long Island Shell Club, Inc., Long Island, NY. Lutcavage, M. 1981. The status of marine turtles in Chesapeake Bay and Virginia coastal waters. Master's thesis, College of William and Mary, VA. Marine Turtle Newsletter. 1989. IUCN resolution urges maximum size limits, protection of habitat, TED's. Mar. Turtle News- letter. 44:1-3. Pritchard, P. C. H., and R. Marquez. 1973. Kemp's ridley turtle or Atlantic ridley. International Union for the Conservation of Na- ture and Natural Resources Monograph No. 2, Marine Turtle Series. Morges, Switzerland. Ropes, J. W. 1968. The feeding habits of the green crab, Carcinus maenas (L.). Fish. Bull. 67:183-200. Schwartz, F. J. 1978. Behavioral and tolerance responses to cold water temperatures by three species of sea turtles (Reptilia, Cheloniidae) in North Carolina. Florida Marine Research Pub. 33:16-18. Shaver, D. J. 1991. Feeding ecology of wild and head-started Kemp's ridley in South Texas waters. J. Herpetol. 25:327-334. Standora, E. A, S. J. Morreale, E. Estes, R. Thomp- son, and M. Hilburger. 1989. Growth rates of Juvenile Kemp's ridleys and their movement in New York waters. Proceedings of the Ninth Annual Workshop on Sea Turtle Con- servation and Biology, p. 175-177. Standora, E. A., S. J. Morreale, R. D. Thompson, and V. J. Burke. 1990. Telemetric monitoring of diving behavior and movements of juvenile Kemp's ridleys. Pro- ceedings of the Tenth Annual Workshop on Sea Turtle Conservation and Biology, 133 p. Williams, A. B. 1984. Shrimps, lobsters, and crabs of the Atlantic coast of the eastern United States, Maine to Flori- da. Smithsonian Institution Press, Washington, D.C. Zinn, D. J. 1984. Marine mollusks of Cape Cod. Cape Cod Museum of Natural History, Brewster, Massachu- setts, 78 p. Zug, G. R., and H. J. Kalb. 1989. Skeletochronological age estimates for juve- nile Lepidochelys kempi from Atlantic coast of North America. Proceedings of the Ninth Annual Workshop on Sea Turtle. Abstract. The tripletail, Lobotes surinamensis, is the only member of the family Lobotidae in the western Atlantic Ocean, and its life history is poorly under- stood. We describe development of tripletail larvae, clarify the litera- ture on their identification, and discuss their temporal and spatial distribution in the northern Gulf of Mexico. Larval tripletail are characterized by 1) a vaulted, me- dian supraoccipital crest with spines along the leading edge; 2) precocious, heavily pigmented pel- vic fins; and 3) large preopercular spines. In addition, the surface of the frontal and supraoccipital bones have a reticulated pattern of depressions or "waffled" appear- ance. Transition to juvenile stage begins at about 9.0-9.5 mm stan- dard length. Tripletail have three supraneurals, six branchiostegal rays, 11 + 13 vertebrae, 27 dorsal rays (XII, 15), and 14-15 anal rays (III, 11-12). Overall, 75% of trip- letail larvae were found in waters >28.8°C, >30.3 ppt, and at stations >70 m deep. Larval tripletail were collected primarily from July through September and almost exclusively in surface tows. Triple- tail spawn offshore. Juveniles, al- though sporadic, are apparently not uncommon in Gulf of Mexico estuaries during summer. Larval development of tripletail, Lobotes surinamensis (Pisces: Lobotidae), and their spatial and temporal distribution in the northern Gulf of Mexico* James G. Ditty Center for Coastal, Energy, and Environmental Resources Coastal Fisheries Institute, Louisiana State University Baton Rouge, LA 70803 Richard F. Shaw Center for Coastal, Energy, and Environmental Resources Coastal Fisheries Institute, Louisiana State University Baton Rouge, LA 70803 Manuscript accepted 4 October 1993 Fishery Bulletin 92:33-45 (1994) The percoid family Lobotidae is usually considered to comprise two genera with about four species (Nelson, 1984), although Johnson ( 1984) only included Lobotes, ques- tioning the affinity of Datnioides. The tripletail, Lobotes surinamen- sis, is cosmopolitan and found in all warm seas (Fischer, 1978); one adult was recorded as far north as St. Margarets Bay, Nova Scotia (44°37'N, 64°03'W (Gilhen and McAllister, 1985). Lobotes surina- mensis is the only member of the family in the Gulf of Mexico (Gulf) (Hoese and Moore, 1977). Tripletail generally occur along the Gulf coast from April through early Oc- tober (Baughman, 1941) and mi- grate south during fall and winter (Merriner and Foster, 1974). Al- though apparently abundant no- where, adult and juvenile tripletail are not uncommon in bays, sounds, and estuaries along the north-cen- tral Gulf coast during summer (Baughman, 1941; Benson, 1982). Tripletail up to 18.6 kg and 89 cm standard length (SL) have been caught, but most average between 1 and 7 kg (Gudger, 1931; Baugh- man, 1941). Tripletail often are in- cluded as a category in Gulf fishing rodeos (Benson, 1982) because of their reputation as "a bold biter" and strong fighter (Gudger, 1931; Baughman, 1941). Tripletail enter the commercial catch on the east and west coasts of Florida and a few tons are taken annually (Fischer, 1978). The development of tripletail lar- vae and their spatial and temporal distribution is poorly understood. Hardy ( 1978) compiled information on tripletail life history. Uchida et al. (1958) and Konishi (1988) pro- vide limited information and illus- trations of tripletail larvae off Ja- pan; however, Konishi's 5.1-mm larva is misidentified. Johnson (1984) commented on cranial mor- phology. Our objectives were to de- scribe the development of tripletail larvae, to clarify the literature on their identification, and to discuss the spatial and temporal distribu- tion of larval tripletail in the north- ern Gulf of Mexico. * Louisiana State University Coastal Fish- eries Institute Contribution No. LSU- CFI-92-8. 33 34 Fishery Bulletin 92(1). 1994 Materials and methods Tripletail larvae were obtained from museum collec- tions throughout the Gulf of Mexico to determine their spatial and temporal distribution. These in- clude collections from the Southeast Area Monitor- ing and Assessment Program's (SEAMAP) ichthy- oplankton surveys of the Gulf from 1982 through 1986 (SEAMAP 1983-19871); National Marine Fish- eries Service (NMFS, Panama City, Florida) and Louisiana State University (LSU) collections from within riverine and oceanic frontal zones off the Mississippi River delta; and collections made by the Gulf Coast Research Lab (GCRL), Ocean Springs, Mississippi, and by Freeport-McMoRan Inc., New Orleans (Appendix Tables 1 and 2). SEAMAP collections from 1982 to 1986 represent the first time-interval for which a complete set of data were available. Standard ichthyoplankton survey techniques as outlined by Smith and Richardson (1977) were employed in data collection. SEAMAP stations sampled by NMFS vessels were arranged in a systematic grid of about 55-km inter- vals. NMFS vessels primarily sampled waters >10 m deep. Each cooperating state had its own sam- pling grid and primarily sampled their coastal wa- ters. Latitude 26°00'N was the southern boundary of the survey area. Hauls were continuous and made with a 60-cm bongo net (0.333-mm mesh) towed obliquely from within 5 m of the bottom or from a maximum depth of 200 m. A flowmeter was mounted in the mouth of each net to estimate volume of wa- ter filtered. Ship speed was about 0.75 m/sec; net retrieval was 20 m/min. At stations <95 m deep, tow retrieval was modified to extend a minimum of 10 minutes in clear water or 5 minutes in turbid wa- ter. Tows were made during both day and night depending on when the ship occupied the station. Overall, 1,823 bongo-net tows were collected and processed during these years. The SEAMAP effort from 1982 to 1984 also involved the collection and processing of 814 neuston samples taken with an unmetered 1x2 m net (0.947-mm mesh) towed at the surface for 10 minutes at each station. SEAMAP sampling during April and May was primarily be- yond the continental shelf, whereas that during March and from June through December was over or immediately adjacent to the shelf at stations <180 m deep. No samples were taken during January and February Additional information on the temporal and spatial coverage of SEAMAP plankton surveys 1 SEAMAP. 1983-1987. (plankton). ASCII characters. Data for 1982-1986. Fisheries-independent survey data. National Ma- rine Fisheries Service, Southeast Fisheries Center: Gulf States Marine Fish. Comm., Ocean Springs, unpubl. data. is found in Stuntz et al. ( 1985), Thompson and Bane (1986, a and b), Thompson et al. (1988), and Sand- ers et al. (1990). Collections from frontal zones off the Mississippi River delta include 311 surface-towed 1x2 m neus- ton net samples (0.333-mm mesh) made by NMFS. NMFS samples were collected during May, August, September, and December (1986 to 1989), although not all four months were sampled each year (Appen- dix Table 1). We also examined 63 surface-towed 1-m2 Tucker trawl samples (0.363-mm mesh) taken at seven stations during July 1987, and 45 surface- towed multiple opening/closing net and environmen- tal sensing system (MOCNESS) (Wiebe et al., 1976) samples (0.363-mm mesh) collected at five stations during April 1988. These samples were from LSU collections. In addition, we examined 17 samples from stations taken by LSU inside the 100-0m isobath during October 1990. The sampling area during October 1990 extended 140 km west from Southwest Pass of the Mississippi River delta along the inner-to mid-shelf. Samples were collected with a 60— cm bongo net (0.333-mm mesh) towed ob- liquely to the surface from 5 m of the bottom or from a maximum depth of 50 m (Appendix Tables 1 and 2). Museum collections from GCRL and Freeport- McMoRan, Inc. were primarily taken off Mississippi Sound and within the Barataria Bay system of Loui- siana, respectively. Gear type and most environmen- tal data were not available from these two institu- tions (Appendix Table 2). Temperature and salinity data were from the sea surface. Hydrographic data from stations where lar- vae were taken were multiplied by the total num- ber of larvae collected at each station to derive median and mean hydrographic values. This method gives weight to distribution of larvae rather than to distribution of stations. We used percent cumulative frequency for defining the relationship between dis- tribution of larval tripletail and water temperature, salinity, and station depth. Percent frequency indi- cates the range of hydrographic conditions most of- ten associated with occurrences of tripletail larvae. Median, mean, and percent cumulative frequency statistics were calculated (SAS Institute, 1985). An examination of tripletail larvae was made to describe developmental morphology. Body measure- ments were made on 21 tripletail between 2.2 and 23.0 mm SL (Table 1) according to the methods of Hubbs and Lagler (1958) and Richardson and Laroche (1979). Measurements were made to the nearest 0.1 mm with an ocular micrometer in a dis- secting microscope. We follow Leis and Trnski's (1989) criteria for defining length of preopercular spines, body depth, head length, eye diameter, and Ditty and Shaw: Larval development and distribution of Lobotes sunnamensis 35 Table 1 Morphometries of larval triplet ail [Lobotes sunnamensis ) from the northern Gulf of Mexico. Measurements are expres sed as % standard length (SL). Preanal Head Snout Orbit Greatest Upper jaw Prepelvic SL n length lengt h length diameter body depth length distance 2.2-2.4 2 60.5-66.0 29.0-29.5 6.5-7.0 12.5-13.5 25.0-27.5 11.5-14.5 4.0-5.9 3 60.0-70.0 37.5-40.0 7.5-10.0 14.0-14.5 40.0-53.5 20.0-20.0 37.5-55.0 6.0-7.9 4 69.5-79.5 38.0-43.0 6.5-9.5 14.0-16.0 51.0-59.5 15.5-17.5 38.0-57.0 8.0-9.9 4 68.0-77.5 34.5-38.5 5.5-6.5 14.0-15.5 58.0-59.0 14.0-15.5 39.0-48.0 10.0-11.9 2 68.5-74.0 38.0-39.0 6.0-6.5 14.5-15.0 54.0-56.5 14.0-14.5 39.0-40.0 13.0-14.9 2 71.5-72.5 35.5-37.0 6.5-7.0 13.0-14.0 55.0-57.5 13.5-14.0 40.0-44.5 15.0-16.9 2 72.5-77.5 34.5-35.5 6.0-6.5 12.5-13.0 56.5-58.0 12.5-13.0 42.0-47.5 21.0-23.0 2 74.0-76.5 39.5-41.5 7.0-8.0 12.0-13.0 54.5-58.0 13.0-14.0 46.5-52.0 eye diameter/head length ratio. We consider noto- chord length in preflexion and flexion larvae synony- mous with SL in postflexion larvae and report all lengths as SL unless otherwise noted. Specimens were fixed in 10% formalin and later transferred to 70% ethyl alcohol. Representative specimens were illustrated with the aid of a camera lucida. Because of the paucity of material, only two specimens were cleared with trypsin and stained with alizarin to examine head spines. We examined the surface of the occipital and frontal bones with a scanning elec- tron microscope (SEM) after the epithelium was par- tially digested with trypsin. Soft rays of the dorsal and anal fins were counted when their pterygiophores were visible, and spines were counted when present. Results Larval morphometries and pigmentation Ninety-eight larval or juvenile tripletail were exam- ined during this study (Appendix Table 2): 7 were preflexion or flexion (<5.0 mm), 34 were postflexion (5.1 to 9.5 mm), and 57 were transforming or juve- nile (>9.5 mm). Body depth increased rapidly dur- ing preflexion and flexion with depth >50% SL by 5.0 mm. The gut was straight. Larvae had 24 myomeres which became obscured by pigment in postflexion larvae. Preanal length was 60-65% SL in preflexion larvae and increased to 70-75% SL in larvae >5.0 mm. Head length averaged 29% SL dur- ing preflexion and increased to about 40% SL in juveniles. The head became increasingly steep, and the upper profile of the forehead was concave by 20.0 mm. The eye was large and had an orbit diameter usually from 35 to 40% head length ( 12.5 to 15.0% SL) by 4.0 mm. The upper jaw reached about mid- eye. Pelvic fins were precocious, heavily pigmented, and inserted behind the pectoral fins near mid-body, usually about 40-50% SL (Table 1). The pelvic fins extended past the anus by 4.0 mm. Early preflexion larvae of 2.2-2.4 mm were sparsely pigmented; pigment was primarily re- stricted to the head and abdomen. On the head, external pigment was present on the posterior sur- face of the midbrain, posteriorly at the base of the supraoccipital crest, on the nape, and immediately anterior to the cleithral symphysis (Fig. 1). By early flexion (4.0 mm), pigment was added between the fore- and mid-brain and on the preopercle above the dorsal-most preopercular spine (Fig. 1). Pigment occurred at the tip of the upper and lower jaws and at the angle of the preopercle near the base of the angle spine by 5.0 mm. The head became heavily pigmented during postflexion. By 10.0 mm, a band of pigment extended diagonally across the head from the nape to the orbit and from below the orbit to the angle of the preopercle (Fig. 1). The eye was at the apex of this chevron-shaped band of pigment. Two parallel stripes of pigment were present between the orbits by 14.0-15.0 mm, extending from the nares to the anterior margin of the supraoccipital crest. These pigment stripes became better formed as lar- vae developed. On the abdomen, melanophores were distributed dorsally over the air bladder, and dor- sally and ventrally along the visceral mass and hindgut of early larvae (Fig. 1). By early flexion, pigment also was present on the pectoral axilla, posteriorly over the visceral mass and hindgut, and was scattered laterally over the body above the vis- ceral mass. Body pigmentation increased rapidly during early postflexion and extended posteriorly to the caudal peduncle by 6.0 mm (Fig. 1). Blotches or mottled areas of pigment formed over the body by 8.0-9.0 mm, becoming more evident as larvae devel- oped (Fig. 1). Pigment along the ventral midline between the anus and notochord tip was restricted to four to five 36 Fishery Bulletin 92(1), 1994 melanophores in early larvae. By early flexion, only one or two postanal melanophores were present along the ventral midline and these were located on the caudal peduncle and at the posterior margin of the hypural bones (Fig. 1). Pigment was also present on the devel- oping pelvic fins by early flex- ion. Melanophores were distrib- uted over the dorsal and anal spines by 6.0 mm and over the anterior-most dorsal and anal rays by 8.5-9.5 mm. Pigment covered all but the distal tips of the dorsal and anal rays by 15.0 mm. Only the base of the caudal- and pectoral-fin rays were pigmented by 13.0-14.0 mm (Fig. 1) and pigment cov- ered about 50% of the caudal fin in a 23.0-mm larva. Pigment occurred only over the proximal portion of the dorsal-most pec- toral-fin rays in the 23.0-mm larva. Head spination and fin development Tripletail larvae were charac- terized by a vaulted, median supraoccipital crest, which originated above mid-eye, and by numerous spines and ridges on the head. Larvae of 2.2-2.4 mm had five to six spines along the leading edge of the supra- occipital crest and one spine on the posterior edge (Fig. 1). Usu- ally eight spines occurred along the leading edge of the crest by 4.0 mm, giving the crest a ser- rate appearance. Length of the crest and its spines decreased as larvae grew (Fig. 1); and the entire supraoccipital crest was resorbed by 15.0-16.0 mm. The surface of the supraoccipital and frontal bones had a reticu- lated pattern of depressions or "waffled" appearance (Fig. 2). Because so few preflexion larvae were col- lected, we were unable to determine when this char- acter first appeared. A large, laterally projecting Figure 1 Larval development of tripletail iLobotes surinamensis) from the north- ern Gulf of Mexico. (A) 2.2 mm, (B) 4.0 mm, (C) 6.3 mm, (D) 8.5 mm, (E) 10.8 mm, (F) 13.7 mm. All measurements are standard length (SL). supraorbital ridge with a single spine was present above the eye of tripletail larvae by 4.0 mm. Both the supraorbital spine and ridge were resorbed by 19.0 mm. Single, simple spines were present on the Ditty and Shaw: Larval development and distribution of Lobotes surinamensis 37 posttemporal and supraclei- thrum by 4.5 mm; a low, simple ridge occurred along the pterotic at about 5.0 mm (Fig. 1). The posttemporal and supracleithral spines were par- tially covered by epithelium but both they and the pterotic ridge were visible on the larg- est specimen examined. Tripletail larvae developed two series of preopercular spines, one along the outer shelf and the other along the inner shelf. Both outer and in- ner shelves have dorsal and ventral limbs. Three spines oc- curred along the posterior mar- gin of the outer shelf of 2.2-2.4 mm larvae, the longest at its angle (Fig. 1). A fourth spine was forming but was small at 2.2 mm. Fifth and sixth spines were added by 6.0 mm; a sev- enth spine, by 7.0 mm. One to two small additional spines were added as larvae grew. By 15.5 mm, three to five spines were visible along the dorsal margin of the outer preopercular shelf, one at the angle, and usu- ally three along the ventral margin; the anterior-most spine along the ventral margin was short and blunt (Fig. 1). All spines along the outer shelf were present in the largest specimen examined (i.e., 26.0 mm). Along the inner preop- ercular shelf, one spine was present in 2.2-2.4 mm larvae and three to four spines by 5.0 mm (Fig. 1). Spines along the inner shelf were short and blunt and covered by epithe- lium. A spine occurred along the posterior margin of the subopercle by 6.0-6.5 mm, near but dorsal to the angle spine of the outer preopercular shelf. The subopercular spine was resorbed by 20.0 mm. A small, flexible spine was present dor- sally on the opercle by 10.0 mm. This spine was diffi- cult to locate on unstained larvae because it was cov- ered by integument. A continuous median finfold extended posteriorly around the body from the nape to the anus of early larvae. Pelvic fins were precocious and elongate (usually >25% SL) and had a full complement of 38 Fishery Bulletin 92(1), 1994 Figure 2 Scanning electron micrograph of the supraoccipital and frontal bones of a 6.3-mm standard length tripletail, Lobotes surinamensis, from the northern Gulf of Mexico. Magnification: 280x. elements (I, 5) by 5.0 mm (Table 2). We were unable to determine when the pelvic-fin buds formed or flexion began because of a lack of specimens between 2.4 and 4.0 mm. Development of the hypural com- plex (by 4.0 mm) coincided with that of the pterygiophores of the dorsal and anal fins. Anlagen of caudal-fin rays formed obliquely in the caudal finfold. The central-most caudal-fin rays formed first and development proceeded outward from mid-base. Notochord flexion was complete by 5.0 mm. The adult complement of 9+8 principal caudal rays were present by 7.0 mm, as were all procurrent caudal rays by 9.0-9.5 mm. All dorsal- and anal-fin ptery- giophores were present by 4.5-5.0 mm and both dorsal and anal spines developed before their rays in each fin. Dorsal and anal spines began to develop anteriorly and proceeded posteriorly to a full comple- ment of elements in each fin by 6.5 mm. Pectoral rays began to form at 5.5-6.0 mm and a full comple- ment (16 rays) was present by 7.0 mm (Table 2). A Fin ray counts of larval tripletail (Lobotes are in standard length (SL). Table surinamensis 2 i from th e northern Gulf of Mexico. Measurements Size (mm SL) n Dorsal Anal Pectoral Pelvic Caudal 4.0 1 Finbase 4.5 1 II, Anlagen 5.0 1 VII, Anlagen 6.3 1 XII. 15 7.1 1 XII, 15 10.2 1 XII, 15 Finbase I, Anlagen I. Anlagen II, 12 III, 12 III. 11 Anlagen Anlagen 13 16 16 3 I, 5 I, 5 I, 5 I. 5 I, 5 4 + 3 6 + 6 7 + 7 3-9+8-2 4-9+8-4 Ditty and Shaw: Larval development and distribution of Lobotes surinamensis 39 cleared-and-stained 10.2-mm specimen had three supraneurals, six branchiostegal rays, four upper and four lower procurrent caudal rays, 11+13 ver- tebrae, 27 dorsal rays (XII, 15), and 14-15 anal rays (III, 11-12). Scales first appeared at 9.0-9.5 mm and marked the beginning of transition to the juvenile stage. Spatial and temporal distribution Overall, 75% of tripletail larvae in this study (Ap- pendix Table 2) occurred at surface water tempera- tures >28.8°C (median=28.9°C, range=27.6-31.0°C), at salinities >30.3 ppt (median = 31.3 ppt, range=22. 0-36.0 ppt), and at stations >70 m deep (median=205 m, range=l-2707 m) (Figs. 3 and 4). Larvae <5.0 mm were collected only at stations >110 m deep. The two smallest larvae (2.2 and 2.4 mm) were taken on 28 July 1987 in a Tucker trawl sample at a station 110 m deep off Southwest Pass of the Mississippi River (Appendix Table 2). Other life stages were collected throughout the study area (Fig. 5, Appendix Table 2). Tripletail larvae were taken almost exclusively from July through September. Two specimens were collected in neuston nets outside this time period, one taken on 21 May 1983 (7.0 mm) and the other by GCRL on 9 October 1968 (10.2 mm) (Appendix Table 2). Salinity (36.5 ppt) and station depth (2,707 m) for the May specimen were the maximums re- corded for a station where larvae were collected during this study (Appendix Table 2). Larval tripletail were collected primarily near the surface. Only 2 of 528 oblique bongo-net collections between July and September yielded tripletail lar- vae («=6, 6.0-9.0 mm, 18 September 1985). Of 537 total surface net tows taken during this same time period, only 31 tows (5.8%) collected tripletail lar- 67% N = 77 28 29 30 31 TEMPERATURE (C) 22 26 27 28 29 30 31 32 33 34 35 36 SALINITY (PPT) <5 5-50 51-180 >180 DEPTH (M) Figure 3 Summary of hydrographic data from positive catch stations for larval tripletail (Lobotes surinamensis) in the northern Gulf of Mexico. Percent catch is sum of larvae by interval divided by total number of tripletail larvae collected overall. Discrepancies in n (number of larvae), among parameters, are the result of missing hydrographic data. Depth is station depth. 40 Fishery Bulletin 92(1), 1994 ' 1 MS AL \ _, f l \ GA f LA A^, 30- TX A^ -a, ^? ::|^C\V/A \ \ Yt*' ' ' ' ■ \ tovj&i/ ° xK* \ \ wu^'° ' * * *: ' ybr* ' i^-*~~* — *"" * ' ! '. .*}?$.. o \ A . / \ LATITUDE to M V — " V ♦ ♦ + ♦ vr^ . ' A. A .fo FL \ JULY-SEPTEMBER V / 40 M 1B0 M 2-4-| — 95 i i 90 85 80 LONGITUDE Figure 4 Distribution of larval tripletail (Lobotes surinamensis) in the northern Gulf of Mexico. Plus ( + ) signs are stations sampled; open diamonds arc > positive catch stations. Data are for collections between July and September 1966-89. vae («=79) (Appendix Tables 1 and 2). Larvae from GCRL and Freeport-McMoRan collections also oc- curred primarily between July and September, but collection data are not available (e.g., total number of stations sampled and extent of sampling area). Discussion The developmental morphology of tripletail larvae from the Gulf generally agrees with limited infor- mation provided by Uchida et al. ( 1958) and Johnson (1984). Larval tripletail are characterized by Da vaulted, median supraoccipital crest with spines along the leading edge; 2) precocious, heavily pig- mented pelvic fins; and 3) large preopercular spines (Uchida et al., 1958; Johnson, 1984; this study). The supraoccipital crest is resorbed by 15.0-16.0 mm SL in Gulf specimens (this study) and by 17.5 mm TL (probably about 16.0 mm SL) off Japan (Uchida et al., 1958). Johnson (1984) described the surface of the frontal and supraoccipital bones of tripletail larvae as rugose. We would characterize these bones as having a "waffled" appearance rather than an elevated one, as implied by rugose (Fig. 2). Regard- less, this modification is found in relatively few other taxa (Johnson, 1984). Sequence of fin comple- tion in larval tripletail is Pg-Dj-Dg-A-Pj and is unlike the six patterns described by Johnson (1984). The third anal spine is the last dorsal- or anal-fin element to form. The dark band of pigment extend- ing backward from above and below the orbit in 10.0-mm larvae is present at 8.3 mm SL (10.6 mm TL) off Japan (Uchida et al., 1958) and in juveniles and adults (Gudger, 1931; Breder, 1949). We did not find the nasal spine noted by Uchida et al. (1958). The 5.1-mm TL specimen listed as L. surinamensis by Konishi (1988) lacks a supraoccipital crest and precocious pelvics, and it has a small, multi-serrate supraorbital ridge rather than the single supraor- bital spine we found. Thus, we believe that Konishi's 5.1-mm TL specimen is not L. surinamensis. Because tripletail have a cosmopolitan distribu- tion, their larvae may be confused with many taxa. Larval tripletail resemble larvae of caproids, some carangids, cepolids, drepaneids, ephippids, leiog- Ditty and Shaw. Larval development and distribution of Lobotes surinamensis 41 nathids, lethrinids, priacanthids, and Hap- alogenys sp. These taxa generally have a median supraoccipital crest, an elongate spine at the preopercular angle, and about 24 myomeres (except cepolids which have 28+ myomeres). In addition, cepolids are lightly to moderately pigmented and have fewer dorsal spines and more soft dorsal- fin rays than tripletail (Leis and Trnski, 1989). Species of other families may have a median supraoccipital crest during devel- opment, but most have pelvic fins inserted anterior to pectorals. Also, larvae of other percoid families are usually not as deep- bodied and as heavily pigmented as triple- tail by early postflexion, and few possess an elongate preopercular spine and low myomere count. Of the aforementioned taxa, only caproids, carangids, ephippids, and priacanthids occur in the Gulf of Mexico. Larvae of the caproid genus Antigonia are most similar to tripletail but have a serrate frontal crest and lower jaw, a very long and serrate preopercular angle spine, and more than 39 dorsal and 26 anal elements (Tighe and Keene, 1984; Leis and Trnski, 1989). In carangids, the two ante- rior-most anal spines are separated from the third by a distinct gap and most spe- cies have a low, median supraoccipital crest with dorsal serrations; other carangids lack a supraoccipital crest entirely. Some car- angids also have a precocious dorsal fin with elongate anterior spines or rays, or a serrated preopercular angle spine. Drepaneids have pigment on the pectoral fins and multiple barbels along the lower jaw. Both larval drepaneids and ephippids are rotund and have pelvic fins inserted anterior to the pectorals. In addition, the Gulf ephippid Chaetodipterus faber has a supraoccipital crest with a single spine dorsally rather than the vaulted, serrate supraoccipital crest found in tripletail. Atlantic spa- defish also have more anal fin elements (tripletail: A. Ill, 11-12; Atlantic spadefish: A. Ill, 17-18). Lar- val leiognathids and lethrinids have a supraoccipital crest that originates above the anterior margin of the eye and both taxa are lightly pigmented (Leis and Trnski, 1989). Also, lethrinids have higher anal fin counts and serrations along the lower jaw (Leis and Rennis, 1983), and leiognathids have a distinc- tive pattern of pigment ventrally on the tail (Leis and Trnski, 1989). Priacanthids have serrate dorsal, anal, and pelvic spines and other serrate ridges and DEPTH ZONE LENGTH (SL) <5 LZZI 5-50 KX] 51-180 rV^ Figure 5 Distribution of larval tripletail iLobotes surinamensis) in the northern Gulf of Mexico with respect to station depth (m). Length classes are combined as follows: 2 mm = 1.0-2.9 mm, 4 mm = 3.0-4.9 mm, 6 mm = 5.0-6.9 mm, etc. All measurements are standard length (SL). Numbers above bars are number of larvae in each length category. spines on the head that tripletail lack (Johnson, 1984). Hapalogenys sp. larvae are extremely simi- lar to tripletail but Hapalogenys sp. apparently lack pigmented pelvic fins, have a serrate supraorbital ridge, have a lacrimal spine, and have pterotic spines or a ridge (Johnson, 1984). Collections of early larvae (this study) and gravid females (Baughman, 1941; Merriner and Foster, 1974) suggest that tripletail spawn primarily dur- ing summer along both the U. S. Gulf and Atlantic coasts. In the Gulf, spawning begins in May, based on the collection of a 7.0-mm larva, and extends through September with peak spawning during July 42 Fishery Bulletin 92(1). 1994 and August (Appendix Table 2). These findings sup- port Baughman's (1941) observation that eggs in gravid females are largest during July and August and small or absent thereafter. Larvae are collected primarily during August and September off Japan (Uchida et al., 1958). Tripletail spawn offshore. This hypothesis of off- shore spawning is supported by the collection of all larvae <5.0 mm at stations on the outer shelf and in oceanic waters. We found no published informa- tion on larval distribution as related to water tem- perature, salinity, or station depth of capture. Larval and juvenile tripletail are collected prima- rily in surface tows (Uchida et al., 1958; this study). Juveniles are often collected with drifting sea weeds, including Sargassum, and near floating objects (Baughman, 1943; Breder, 1949; Uchida et al., 1958; Dooley, 1972; Benson, 1982) as they float on their side (Gudger, 1931; Breder, 1949). The size at which tripletail become associated with drifting sea weeds is poorly known, but Uchida et al. (1958) collected juveniles between 10.0 and 20.0 mm TL in seaweeds. Adult tripletail occur primarily in gulf waters, but enter passes, inlets, and bays near river mouths (Gudger, 1931; Baughman, 1941). The degree to which tripletail utilize estuaries during their life history is unknown. Juveniles are apparently not uncommon (although they may be sporadic) in Gulf coast estuaries during the summer. We examined eight specimens (14.5-26.0 mm) collected at the surface in waters <3 m deep (Fig. 5). Modde and Ross (1981) collected 236 juvenile tripletail (size range not given) during 1976 in the surf zone of Horn Island, Mississippi, but only one during 1975 and five during 1977. Juveniles also occur in shal- low waters (1-3 m) within the Barataria Bay sys- tem of Louisiana.2 In contrast, juvenile and adult tripletail in the Indian River lagoon off the east coast of Florida occupy areas which average 30—31 ppt. The lagoon typically goes hypersaline, to 40 ppt, during spring when most tripletails first appear in the lagoon. Tripletail have not been observed or captured in extensive collections of oligohaline ar- eas of the St. Lucie River and Sebastian Creek.3 Adult tripletail generally occur along the Gulf coast from April through early October (Baughman, 1941) and are caught in great numbers in Mobile Bay, Alabama, and along the Mississippi coast dur- ing summer (Baughman, 1941). Greatest concentra- tions of adults are found along the northern Gulf from St. Marks, Florida, to the St. Bernard River, 2 Leroy Kennair, Freeport-McMoRan, Inc., New Orleans, LA., pers. commun. 1993. 3 R. Grant Gilmore, Harbor Branch Oceanographic Institution, Fort Pierce, FL, pers. commun. 1993. Texas (Baughman, 1941). Seasonality of adults sug- gests that tripletail migrate south during fall and winter and return in spring (Merriner and Foster, 1974). Tripletail congregate around sea buoys, bea- cons, pilings, and other objects (Gudger, 1931) but have been collected in a wide variety of habitats including rocky and coral reef areas in deeper wa- ter (Baughman, 1941). Acknowledgments This study was supported by the Marine Fisheries Initiative (MARFIN) Program (contract numbers: NA90AA-H-MF111 and NA90AA-H-MF727). We thank SEAMAP and the Gulf States Marine Fish- eries Commission for providing specimens and en- vironmental data, and the Louisiana Board of Re- gents and the LaSER (Louisiana Stimulus for Ex- cellence in Research, contract number 86-LUM(D- 083/13) program for support during July 1987, April 1988, and October 1990 ichthyoplankton cruises. We also thank those who loaned us specimens or pro- vided data: Churchill Grimes, NMFS, Southeast Fisheries Center, Panama City, FL; Wayne Forman and Leroy Kennair, Freeport-McMoRan, New Or- leans, LA.; Stuart Poss, Gulf Coast Research Lab, Ocean Springs, MS. Thanks also to Cathy Grouchy for illustrating larvae, to Joseph S. Cope for computer assistance, and to Laura Younger for providing scan- ning electromicrographs of the frontal and occipital bones. Finally, we thank the reviewers for their com- ments in substantially improving the mansucript. Literature cited Baughman, J. L. 1941. On the occurrence in the Gulf coast waters of the United States of the triple tail, Lobotes surinamensis, with notes on its natural history. Am. Nat. 75:569-579. 1943. Additional notes on the occurrence and natu- ral history of the triple tail, Lobotes surinamensis. Am. Midi. Nat. 29(2):365-370. Benson, N. G. (ed.). 1982. Life history requirements of selected finfish and shellfish in Mississippi Sound and adjacent areas. U.S. Fish and Wildl. Serv., Office Biol. Serv., Washington, D.C., FWS/OBS-81/51, 97 p. Breder, C. M., Jr. 1949. On the behavior of young Lobotes surinam- ensis. Copeia 1949(4):237-242. Dooley, J. K. 1972. Fishes associated with the pelagic sargassum complex, with a discussion of the sargassum community. Contrib. Mar. Sci., Univ. Texas 16:1-32. Ditty and Shaw: Larval development and distribution of Lobotes surinamensis 43 Fischer, W., (ed.). 1978. FAO species identification sheets for fishery purposes. Western Central Atlantic (Fishing Area 31), Vol. 3. FAO, Rome. Gilhen, J., and D. E. McAllister. 1985. The tripletail, Lobotes surinamensis, new to the fish fauna of the Atlantic coast of Nova Scotia and Canada. Can. Field-Natur. 99(1):116-118. Gudger, E. W. 1931. The tripletail, Lobotes surinamensis, its names, occurrence on our coasts and its natural history. Am. Nat. 65: 49-69. Hardy, J. D., Jr. 1978. Development of fishes of the Mid-Atlantic Bight: an atlas of egg, larval and juvenile stages. Vol. Ill: Aphredoderidae through Rachycentridae. U.S. Fish. Wildl. Serv., Biol. Serv. Prog. FWS/OBS- 78/12. Hoese, H. D., and R. H. Moore. 1977. Fishes of the Gulf of Mexico: Texas, Louisi- ana, and adjacent waters. Texas A&M Univ. Press, College Station, 327 p. Hubbs, C. L., and K. F. Lagler. 1958. The fishes of the Great Lakes region. Univ. Mich. Press, Ann Arbor, 213 p. Johnson, G. D. 1984. Percoidei: development and relationships. In H. G. Moser, W. J. Richards, D. M. Cohen, M. P. Fahay, A. W. Kendall Jr., and S. L. Richardson (eds.). Ontogeny and systematics of fishes, p. 464— 498. Am. Soc. Ichthy. Herp., Spec. Publ. No. 1. Konishi, Y. 1988. Lobotidae. In M. Okiyama (ed.), An atlas of the early stage fishes in Japan. Tokai Univ. Press, Tokyo, 1,154 p. (In Japanese.) Leis, J. M., and D. S. Rennis. 1983. The larvae of Indo-Pacific coral reef fishes. Univ. Hawaii Press, Honolulu, 269 p. Leis, J. M., and T. Trnski. 1989. The larvae of Indo-Pacific shorefishes. Univ. Hawaii Press, Honolulu, 371 p. Merriner, J. V., and W. A. Foster. 1974. Life history aspects of the tripletail, Lobotes surinamensis (Chordata-Pisces-Lobotidae), in North Carolina waters. J. Elisha Mitchell Sci. Soc. 90<4):121-124. Modde, T., and S. T. Ross. 1981. Seasonality of fishes occupying a surf zone habitat in the northern Gulf of Mexico. Fish. Bull. 78(41:911-922. Nelson, J. S. 1984. Fishes of the World. 1984, 2nd ed. John Wiley & Sons, NY, 523 p. Richardson, S. L., and W. A. Laroche. 1979. Development and occurrence of larvae and juveniles of the rockfishes Sebastes crameri, Sebastes pinniger, and Sebastes helvomaculatus (Family Scorpaenidae) off Oregon. Fish. Bull. 77(11:1-46. Sanders, N., Jr., T. Van Devender, and P. A. Thompson. 1990. SEAMAP environmental and biological atlas of the Gulf of Mexico, 1986. Gulf States Marine Fish. Comm., Ocean Springs, MS, No. 20, 328 p. SAS Institute, Inc. 1985. SAS user's guide: statistics, 1985 ed. SAS Institute, Cary, NC, 584 p. Smith, P. E., and S. L. Richardson. 1977. Standard techniques for pelagic fish egg and larva surveys. FAO Fish. Tech. Paper No. 175, 100 p. Stuntz, W. E., C. E. Bryan, K. Savastano, R. S. Waller, and P. A. Thompson. 1985. SEAMAP environmental and biologcial atlas of the Gulf of Mexico, 1982. Gulf States Marine Fish. Comm., Ocean Springs, MS, No. 12, 145 p. Thompson, P. A., and N. Bane. 1986a. SEAMAP environmental and biological at- las of the Gulf of Mexico, 1983. Gulf States Marine Fish. Comm., Ocean Springs, MS, No. 13, 179 p. 1986b. SEAMAP environmental and biological at- las of the Gulf of Mexico, 1984. Gulf States Marine Fish. Comm., Ocean Springs, MS, No. 15, 171 p. Thompson, P. A., T. Van Devender, and N. J. Sanders, Jr. 1988. SEAMAP environmental and biological atlas of the Gulf of Mexico, 1985. Gulf States Marine Fish. Comm., Ocean Springs, MS, No. 17, 338 p. Tighe, K. A., and M. J. Keene. 1984. Zeiformes: development and relationships, p. 393-398. In H. G. Moser, W. J. Richards, D. M. Cohen, M. P. Fahay, A. W. Kendall Jr., and S. L. Richardson (eds.), Ontogeny and systematics of fishes. Am. Soc. Ichthy. Herp., Spec. Publ. No. 1. Uchida, K., S. Imai, S. Mito, S. Fujita, M. Ueno, Y. Shojima, T. Senta, M. Tahuka, and Y. Dotsu. 1958. Studies on the eggs, larvae, and juveniles of Japanese fishes. Series 1: Second lab. of fish, biol. Fish. Dep., Fac. Agric, Kyushu Univ., Fukuoka, Japan. Wiebe, P. H., K. H. Burt, S. H. Boyd, and A. W. Morton. 1976. A multiple opening/closing net and environ- mental sensing system for sampling zoo- plankton. J. Mar. Res. 34(3):313-326. 44 Fishery Bulletin 92(1), 1994 Appendix Table 1 Summary of total number of bongo-net/neuston-net stations examined for tripletail larvae (Loboten surina- mensis) in the Gulf of Mexico. Acronyms are as follows: SEAMAP = Southeast Area Monitoring and Assess- ment Program; NMFS = National Marine Fisheries Service, Panama City, Florida; LSU = Louisiana State University. NS means no samples. MAR APR MAY JUN JUL AUG SEP OCT NOV DEC SEAMAP 1982 77J/02 69/68 71/73 102/100 26/24 NS NS 3/8 29/3 NS 1983 15/13 27/27 84/84 55/45 44/42 NS NS 39/26 NS 24/23 1984 23/0 44/0 46/0 55/54 20/26 155/162 NS 24/0 6/0 36/36 1985 29/0 NS NS 85/0 39/0 69/0 20/0 4/0 2/0 24/0 1986 NS 24/0 90/0 57/0 10/0 NS 145/0 43/0 73/0 24/0 TOTAL 144/13 164/95 291/157 354/199 139/92 224/162 165/0 113/34 110/3 108/59 NMFS2 1986 46 1987 68 1988 55 71 36 1989 35 LSU 19873 63 1988'' 45 1990' 17 1 60-cm bongo net, 0.333-mm mesh, oblique-tow from depth. - 1 x 2 m neuston net, 0.947-mm mesh, 10 mill, surface-tow, unmetered. 3 lm2 Tucker trawl, 0.947-mm mesh, 3 min. surface-tow each net, nine net collections per station, 4 lm2 MOCNESS, nine nets of 0.333-mm mesh, 3-min. surface-tow each net, five total stations. ieven total stations. Ditty and Shaw: Larval development and distribution of Lobotes surinamensis 45 Appendix Table 2 Positive catch station data for tripletail (Lobotes surinamensis) larvae from noi •thern Gulf of Mexi co waters. Gear codes are: B=bongo net, N= :Neuston net, T=Tucker trawl, U= unknown. Station Date Gear Latitude Longitude Station depth (m) *C PPT n Length (mm SL) SEAMAP2 1420 5-21-83 N 26*30 88*00 2707 27.6 36.5 7.0 3235 7-17-84 N 28*15 90*30 70 29.4 25.9 8.8 3238 7-17-84 N 28*30 90*30 38 29.4 25.8 7.0 3259 7-22-84 N 29*00 87*00 1251 28.9 32.8 12.3 2511 8-03-84 N 29*00 88*15 1013 27.6 32.4 7.1-18.5 2523 8-03-84 N 29*15 88*30 82 28.0 26.0 7.9 2548 8-05-84 N 29*00 88*45 249 27.6 28.7 16.8 4231 8-05-84 N 29*28 87*00 486 28.9 30.3 16 6.8-13.0 4201 8-01-85 N 28*00 84*52 205 29.6 30.8 10 10.3-15.9 4204 8-01-85 N 28*00 85*02 265 28.8 32.6 5 9.0-16.5 4210 8-02-85 N 28*21 86*00 457 28.8 32.1 4 6.8-10.0 4216 8-03-85 N 28*53 86*16 335 29.1 31.3 2 9.0 4219 8-03-85 N 28*40 86*30 457 28.9 33.6 1 9.9 4320 8-24-85 N 27*38 94*00 455 28.0 — 1 4.0 4326 8-25-85 N 27*40 93*00 265 29.7 36.0 1 7.8 4332 8-26-85 N 27*46 92*00 457 30.0 35.4 1 9.1 4484 9-18-85 B 29*07 89*44 20 27.8 29.5 2 6.0 4490 9-18-85 B 28*37 90*26 27 27.8 32.6 4 6.4-9.0 LSU2 137 7-28-87 T 28*42 89*29 110 29.5 22.0 2 2.2-2.4 145 7-28-87 T 28*35 89*22 182 29.6 32.5 2 5.0 163 7-30-87 T 28 2 t 89*14 640 31.0 33.6 2 6.3 168 7-30-87 T 28*24 89*14 640 31.0 33.6 2 6.3 175 7-30-87 T 28*27 89*16 410 29.8 35.3 2 — 177 7-30-87 T 28*27 89*16 410 29.8 35.3 2 4.5 GCRL5 Station 6 7-13-67 N 29*15 88*11 182 — — 1 12.5 T-108-7- -02 8-25-71 U 29*10 88*45 55 — — 1 8.7 T-108-3- 114 8-27-71 U 29*50 88*05 'J 7 — — 1 11.7 T-208-4- ■01 8-23-72 I' 29*40 88*14 38 — — 2 7.2-7.3 T- 109-6- ■02 9-21-71 V 29*20 88*21 55 — — 1 15.4 T- 109-5- 03 9-22-71 u 29*30 88*24 46 — — 1 8.6 T-209-2- ill 9-15-72 u 30*00 88*14 27 — — 2 7.7-10.7 Station 5 10-09-68 N 29*19 88*14 73 — — 1 10.2 Freeport-McMoRan'' 2 8-24-71 u 29*16 89*57 1 — — 1 14.5 3 8-10-71 u 29*22 89*48 3 — — 2 16.5-18.5 4 8-23-73 u 29*16 89*57 1 — — 1 26.0 5 8-15-66 u 29*16 89 57 3 — — 4 11.5-21.5 NMFS5 53 8-28-88 N 29*00 88*53 149 30.3 27.5 1 10.8 58 8-29-88 N 29*07 88*49 8 J 29.5 29.0 1 13.7 5 9-03-87 N 29*12 88 43 71 29.3 32.8 1 23.0 23 9-25-86 N 28*50 89*05 195 29.4 34.0 2 7.3-13.2 32 9-06-89 N 28*49 89*16 410 29.8 35.3 1 18.7 42 9-26-86 N 29*09 88 40 77 29.3 — 1 8.6 43 9-05-87 N 28*46 89*29 104 29.2 32.1 1 7.5 ; Southeast Area Monitoring and Assessment Program. 2 Louisiana State University, Coastal Fisheries Institute, Baton Rouge. 3 Gulf Coast Research Lab. Ocean Springs, Mississippi. 4 Freeport-McMoRan. Inc., New Orleans, Louisiana. 5 National Marine Fisheries Service, Panama City Lab, Florida. Abstract. — Otoliths were used to determine the age and growth of the coral trout Plectro- pomus leopardus from Lizard Is- land area, Northern Great Barrier Reef, Australia. An alternating pattern of opaque (annulus) and translucent zones was visible in whole and sectioned otoliths. How- ever, compared to sectioned otoliths, whole readings tended to underestimate age of older fish. Otoliths of mark-recaptured fishes treated with tetracycline showed that one annulus is formed per year during the winter and spring. The oldest individual examined was 14 years of age. Schnute's growth formula was used to deter- mine the best model to describe the growth of the coral trout. The von Bertalanffy model for fork length (FL) fitted the data well and the resulting model was Lt= 52.2(1 -e -0.354U + 0.766)). Line-fishing usually does not cap- ture fishes smaller than 25 cm FL, thereby excluding most 0+ and 1+ year old fish and probably the slower growing 2+ year old fish. These first three years of life rep- resent the period of fastest growth, so, if the growth curve is fitted only to the line fishing data, the growth rate of the population is underestimated. Multiple regres- sion was used to predict age from otolith weight and fish length and weight. Otolith weight was the best predictor of age in the linear model and explained as much variation in age as fish size in the von Bertalanffy model. Age validation and estimation of growth rate of the coral trout, Plectropomus leopardus, (Lacepede 1802) from Lizard Island, Northern Great Barrier Reef Beatrice Padovani Ferreira* Garry R. Russ Department of Marine Biology, James Cook University of North Queensland Townsville Q481 1, Australia *Present address: CEPENE-IBAMA. R Samuel Hardman s/n° Tamandare. Pernambuco. Cep. 55578-000. Brazil Manuscript accepted 8 September 1993 Fishery Bulletin 92:46-57 (1994) The coral trouts of the genus Plec- tropomus Oken are members of the serranid subfamily Epinephelinae, commonly known as groupers. These fishes occur in shallow tropi- cal and subtropical seas of the Indo-Pacific region (Randall and Hoese, 1986) where they usually are at the top of food chains and thus play a major role in the struc- ture of coral reef communities (Randall, 1987). Groupers typically represent an important fishery resource throughout the tropical and sub- tropical regions of the world (Ralston, 1987). On the Great Bar- rier Reef, the common coral trout Plectropomus leopardus (Lacepede 1802) is the most abundant species of the genus (Randall and Hoese, 1986) and usually the primary tar- get of recreational and commercial fishermen. The Queensland com- mercial line-fishing fleet takes a total annual catch of about 4,000 metric tons (t) of reef and pelagic species. The coral trout composes the largest single component of this catch (over 30%) with around 1200 t caught annually (Trainor, 1991). The recreational sector of this fish- ery is estimated to catch two to three times the commercial catch of reef fish (Craik, 19891). Worldwide studies on age and growth of Epinephelinae indicate that they are long lived, slow grow- ing, and have relatively low rates of natural mortality (Manooch, 1987). Fishes with these characteristics are susceptible to overfishing. Only by obtaining validated estimates of growth is it possible to determine population dynamics, estimate po- tential yield, monitor the responses of populations to fishing pressure, and properly manage the fishery. Some information on age, growth, and longevity is available for the common coral trout. On the Great Barrier Reef, Goeden (1978) estimated the growth rate of this species at Heron Island from length-frequency data. Mcpherson et al. (1988), determined age and growth of the common coral trout in the Cairns region by counts of annuli in whole otoliths. Loubens (1980) estimated age and growth for P. leopardus from New Cale- donia from counts of annuli in bro- ken and burnt otoliths. The period- icity of formation of annual rings in the latter two studies was verified through observation of marginal 1 Craik, G. J. S. 1989. Management of rec- reational fishing in the Great Barrier Reef Marine Park Tech. Memo. GBRMPA-TM- 23, 35p. 46 Ferreira and Russ: Age-validation and growth rate of Plectropomus leopardus 47 increments in otoliths. Direct validation of age has not yet been attempted for P. leopardus. Fish population models usually require a general description of the growth process by means of an ap- propriate mathematical function. The von Bertalanffy (1938) growth model is the most stud- ied and the most frequently used, since its applica- tion by Beverton and Holt (1957) to the yield-per- recruit problem (Kimura, 1980; Gallucci and Quinn, 1979). Many alternative growth curves have been proposed (see Moreau, 1987) as well as the use of polynomial functions (Chen et al., 1992). In this work, Schnute's (1981) formula was used to find the model that best described the growth of P. leopardus. For several species of fishes, otolith growth has been found to continue with age, independent offish size (Boehlert, 1985; Casselman, 1990; Beckman et al., 1991). Boehlert (1985) suggested the use of otolith weight as a non-subjective, cost-effective methodology for age determination that would de- crease variability among age estimates. The aims of this study were to obtain direct vali- dation of age-at-length information and to find the model that best described the growth of the common coral trout from Lizard Island, Northern Great Bar- rier Reef, Australia. In addition, the relationship between otolith weight, body size, and age of the coral trout was studied to understand the mode of growth of the otolith and to assess the usefulness of otolith dimensions in predicting age. Materials and methods Coral trout (t?=310) were sampled in the Lizard Is- land area (lat. 14° 40' S, long. 154° 28' E) from March 1990 to February 1992. Fishes were caught by rec- reational and commercial fishermen using hook and line (77 = 184) and by recreational spearfishermen (n=94). Individuals smaller than 20-cm total length are usually not vulnerable to line fishing, so they were caught around Lizard Island by scuba divers using fence nets (77=32). Fork length (FL, cm), stan- dard length (SL, cm) and total weight (TW, g) were measured for each fish. FL is defined as the length from the front of the snout to the caudal fork, and SL is defined as the length from the front of the upper lip to the posterior end of the vertebral col- umn. A simple linear regression of the form FL= a + 6*SL was used to describe the relationship be- tween FL and SL. To describe the relationship be- tween FL and TW the variables were logarithmically transformed and the linearized version of the power function TW(g)= a*FL(cm)b was fitted to the data. In the coral trout, the sagittae are the largest of the three pairs of otoliths and were used for read- ings. Sagittae were removed, cleaned, weighed, and stored dry. Left and right sagittae, when intact, were weighed to the nearest milligram. Otoliths were prepared and read as described by Ferreira and Russ (1992). To increase contrast between bands, whole otoliths were burned lightly on a hot plate at 180°C (Christensen, 1964). Both right and left sagitta were read whole under reflected light with a dissecting microscope at 16x magnification. The otoliths, with the concave side up, were placed in a black container filled with immersion oil. Subse- quently, the left sagittae was prepared for reading by embedding in epoxy resin (Spurr, 1969) and sec- tioning transversely through the core with a Buehler Isomet low-speed saw. Sections were mounted on glass slides with Crystal Bond 509 adhesive, ground on 600- and 1200-grade sand paper, polished with 0.3— u alumina micropolish and then examined un- der a dissecting microscope at 40x magnification with reflected light and a black background (Fig. 1). Annuli were counted from the nucleus to the proxi- mal surface of the sagitta along the ventral margin of the sulcus acousticus. Terminology for otolith readings followed defini- tions of Wilson et al. (1987). Two experienced read- ers independently counted opaque zones (annuli) in each whole and sectioned otolith of a random subsample (77 = 136) to assess the precision and ac- curacy of countings obtained by the two methods. The precision of age estimates was calculated with the Index of Average Percent Error (IAPE), (Bea- mish and Fournier, 1981). Results obtained from whole and sectioned otoliths were compared by plot- ting the difference between readings obtained from whole and sectioned otoliths (Section Age- Whole age) against Section Age. The results of this com- parison indicated that whole otolith readings tended to be lower than readings from sectioned otoliths when more than six rings occurred in the otolith. Therefore, remaining otoliths were read whole first and, if the number of rings was higher than six or the whole otolith was considered unreadable, the otolith was sectioned and counts were repeated. The results were accepted and used in the analysis when the counts of the two readers agreed. If the counts differed, the readings were repeated once and if the counts still differed, the fish was excluded from the analysis. Ages were assigned based on annulus counts and knowledge of spawning season. The periodicity of annulus formation was determined with the use of tetracycline labelling. From August 1990 to Febru- ary 1992, 80 fishes were caught in a trapping pro- gram at Lizard Island fringing reeflDavis, 19922), 2 C. Davies. 1992. James Cook University, Townsville, Q4811, Australia, unpubl. data. 48 Fishery Bulletin 92(1), 1994 Figure 1 Whole (A) and sectioned (B) otolith of an 11-year-old coral trout, P. Icopardus, under reflected light with a black background showing alternating pattern of translucent and opaque bands; a = anterior, p = posterior, d = dorsal, v = ventral, di = distal, pr = proxi- mal, ds = dorsal sulcus, vs = ventral sulcus. Scale bar = 1 mm tagged with T-bar anchor tags and injected with tetracycline hydrochloride before being released. The fish were injected in the coelomic cavity under the pelvic fin with a dosage of 50 mg of tetracycline per kg offish (McFarlane and Beamish, 1987), in a con- centration of 50 mg per mL of sterile saline solution. Five fish were recaptured after periods of at least one year at large. Two of those fish were reinjected at the time of recapture and kept in captivity for periods of three to four months. To determine the time of formation of the first annulus, five young of the year were captured with Ferreira and Russ: Age-validation and growth rate of Plectropomus leopardus 49 fence nets. Three of these fishes were injected with tetracycline at the time of capture, and all five fish were kept in captivity for periods of 3 to 17 months. The otoliths of the fishes treated with tetracycline were removed, sectioned, and observed under fluo- rescent light. To determine time of formation of the translucent and opaque zones, the distances be- tween events for which time of occurrence was known (i.e., between two tetracycline bands or be- tween a tetracycline band and the margin of the otolith) were measured on otolith sections and plot- ted against the corresponding time interval. The relative positions of the translucent and opaque zones to these marks were then measured and plot- ted on the same scale. While this method does not provide real distances, it standardizes the measure- ments allowing for comparison between fish of dif- ferent ages. The relation between otolith weight, fish size (length and weight), and age was analyzed. Otolith weight was plotted against FL for each age class separately. A multiple linear regression model was fitted in a step-wise manner to predict age from otolith weight and fish size and to predict otolith weight from age and fish size. The inclusion level for the independent variables was set at P=0.10. The assumptions of normality and homoscedasticity were tested by plotting the residuals from the re- gression models. The growth models were fitted to the data and their coefficients and standard errors estimated by means of standard non-linear optimization methods (Wilkinson, 1989). As the plot of the length-at-age data indicated, some form of asymptotic growth, Schnute's (1981) reformulation of the von Bert- alanffy growth equation for length in which a*0 was fitted to the data: ,-aU-t\) L,=y\h+(y2b-ylb) where Lf is length at age; tl and t2 are ages fixed as 1 and 14 respectively ; yl and y2 are estimated sizes at these ages; and a and b are the parameters which indicate if the appropriate growth curve lies closer to a three or two parameter sub-model. By limiting parameter values, the data were used di- rectly in selecting the appropriate sub-model, namely the generalized von Bertalanffy, Richards, Gompertz, Logistic, or Linear growth models. Sub- sequently, the original von Bertalanffy (1938) -KU -to) To evaluate the effects of gear selectivity (and consequently varying size and age composition) on the estimates of growth parameters, the von Bertalanffy growth equation was fitted first to data collected by line and spear fishing only and then to the same data combined with the fence-net sample composed of younger fish. Results Otolith reading In the coral trout, the sagittae presented a pattern of alternating translucent zones and wide opaque zones (annuli) with no sharp contrast between zones (Fig. 1). The first two annuli were notably wider and less well defined than the subsequent ones in sec- tioned otoliths. Whole sagittae were used to confirm the presence of these first annuli. In whole otoliths, annuli were clearly distinguish- able and easy to count along the dorsal side of the otolith, where up to 12 rings were counted in some otoliths. However, readings from whole otoliths tended to be lower than readings from sectioned otoliths when more than six rings were present, and this tendency increased with the mean number of rings, particularly after ten rings. (Fig. 2 ). Tetra- cycline-labelled otoliths validated the periodicity of annuli in sectioned otoliths, indicating that whole otolith readings tend to underestimate age of > 10- year-old fishes. A comparison between results of )was growth equation for length L( = L^d-e fitted to the data. V is length at age; Lx is the as- ymptotic length, K is the growth coefficient, t is age, and to is the hypothetical age at which length is zero. £, 4 < ,1 5 3 o c 5 2 i of the variability in age of the coral trout (r^O.889, P<0.0001), with fork length accounting for 1.5% (partial r2= 0.015). Otolith weight was a function of age and fish size, as indi- cated by the results of the multiple regression fit- ting. The interaction between age and fork length alone accounted for 89% of the variability (r-^0.892, P<0.0001) Validation of annulus formation All fishes treated with tetracycline displayed clear fluorescent marks in their otoliths (Fig. 4). The re- sults obtained for recaptured and captive fish, rang- ing in age from one to eight years, showed that annuli are formed once per year (Fig. 5). The first annulus is formed in the otoliths of the juvenile coral trout during their first year of life (Fig. 6). The rela- tive positions of the fluorescent bands, in relation to the otolith margin and the translucent and opaque zones (annuli), indicated that the formation of the annulus occurred mainly during winter and early spring (Figs. 5 and 6). Growth model The samples obtained from line-fishing and spear- fishing were selective towards individuals larger than 25 cm FL. Consequently, the 0+ age class was not rep- resented in this sample and the age-1 year class was represented by only four individuals (Fig. 7). The sample collected with fence nets, composed of indi- viduals from the smaller size classes, consisted to- tally of individuals of the 0+ and 1+ year classes (Fig. 7). Table 2 shows the results obtained when fitting the growth model to the data including all age classes and to the data including only age >2+. Table 1 Correlation between otolith weight (mg) and fork length (cm) for each age class of the coral trout P. leopardus. Age r- P< df Age r2 P< df 0 0.826 0.0001 18 8 0.481 0.0001 19 1 0.972 0.0001 10 9 0.405 0.0001 12 2 0.829 0.0001 27 10 0.120 no sig. 8 3 0.747 0.0001 19 11 0.937 0.0001 7 4 0.652 0.0001 18 12 0.526 no sig. 3 5 0.650 0.0001 30 13 0.993 0.05 2 6 0.489 0.0001 43 14 0.049 no sig. 2 7 0.514 0.0001 30 Ferreira and Russ: Age-validation and growth rate of Plectropomus leopardus Figure 4 Sectioned otolith of a recaptured individual coral trout, P. leopardus In" 0057) rescent band. Scale Bar = 0.25 mm showing fluo- When fitting Schnute's model to both sets of data, the value of the parameter b was very close to 1. In the boundary where 6 = 1, the curve was reduced to a three parameter model that corresponds to the von Bertalanffy curve for length (Schnute, 1981). The resulting growth model for all age classes, in the form of a von Bertalanffy model, was L, =52.2 (1-e- 0.354(^ + 0.766)) r = 0.895 (Fig. 8). Table 2 Von Bertalanffy growth parameters V.B. and re- spective standard errors (SE), correlation coeffi- cients (r2) and degrees of freedom (df) for the growth curve fitted to all data and to the data for coral trout, P. Leopardus, >2 year old only. (SE) K (SE) (SE) df V.B. all ages V.B. age >2+ 52.20 (0.768) 0.354 (0.0241 61.29 0.132 (3.483) (0.030) -0.766 (0.097) -4.660 (1.024) 0.895 310 0 622 272 The results obtained when fitting the growth curve to all data and to the data for fish >2+ years old only were quite different (Table 2). From age-2 onwards, the growth rate is much slower than the one esti- mated by using all age classes, as indicated by the growth coefficient K. Consequently, the estimated Lm is larger and the estimated to is a very large, nega- tive value. The resulting growth model was L, =61.29 (l-e-0.132(f + 4.66)) r = 0.622 (Fig. 9). No systematic trend in the residuals was observed (normality test P>0.1) (Figs. 8 and 9). The relation between fork length (FL) and the standard length (SL) was SL = -0.308 + 0.852 * FL, r2 = 0.994, and the relationship between FL and Total Weight (TW) was TW = 0.0079 *FL 3 157 0.967. Discussion While some comparisons between readings of whole and sectioned otoliths have indicated good agree- 52 Fishery Bulletin 92(1), 1994 WSSAWSSAW Tag No AUG 90 0085 r |H FEB 92 NOV 90 age = 5 | H NOV 91 NOV 90 NOV 91 3862 1 ago = 7 | | FEB 92 MAR 90 MAR 91 WW ■ 1 age I ■ Fluorescent | Translucent ^ Opaque YY Winter § Summer ^ Spnng /\ Autumn Figure 5 Diagrammatic representation of otoliths of mark- released-recaptured coral trout, P. leopardus, treated with tetracycline showing relative positions of the fluorescent bands, otolith margin, translucent and opaque zones. Bars represent only the distal part of the radius of the otolith section, measured from the nucleus to the proximal surface of the sagitta along the ventral margin of the sulcus acousticus. The dates on the top of the bars indi- cate time of tetracycline treatment and the dates on the end of the bars indicate time of recapture. ment (Boehlert, 1985; Maceina and Betsill, 1987), others have suggested that reading whole otoliths underestimates true age and that this problem be- comes worse with fish age (Boehlert, 1985; Hoyer et al., 1985). This is mainly due to the fact that in many species, sagittae growth is asymmetrical (Irie, 1960). Growth appears to be linear only up to a cer- tain age or size, after which additions occur mainly on the interior proximal surface, along the sulcus region (Boehlert, 1985; Brothers, 1987; Beamish and McFarlane, 1987). That seems to be the case for the coral trout, as comparison of results of whole and sectioned otoliths indicated that lateral views did not reveal many of the annual growth zones in older individuals. However, whole otoliths require much less time for analysis than sectioned ones and seem to provide more precise readings. Therefore, it is use- ful to know the limit of reliability of whole readings and to define the conditions appropriate for its use. Like the inshore coral trout Plectropomus macula tits (Ferreira and Russ, 1992), the common coral trout P. leopardus is a relatively long-lived, slow-growing species. The results on growth and longevity obtained here differ somewhat from those of previous studies. Goeden (1978), using the Petersen method, identified age cohorts up to age 5+ for P. leopardus. However, the limitations of the use of length-frequency data to estimate age of long- lived fish are well known (Manooch, 1987; Ferreira and Vooren, 1991). Mcpherson et al. (1988), using counts of annuli in whole otoliths, were able to age fish up to seven years old. Longevity was probably underestimated in their study as counts were per- formed only on whole otoliths. More recently. Brown et al. (1992)3 analyzed whole and sectioned otoliths of coral trout from the same area as Mcpherson et al. (1988) and were able to count up to 14 rings. Loubens ( 1980) counted annuli from burnt and bro- ken otoliths and estimated a maximum longevity for 3 Brown, I. W., L. C. Squire, and L. Mikula. 1992. Effect of zon- ing changes on the fish populations of unexploited reefs. Stage 1: pre-opening assessment. Draft interim report to the Great Bar- rier Reef Marine Park Authority, Townsville, Australia, 27 p. SAWSSAWSS age = 2 | Fluorescent | Translucent ^] Opaque yry Winter ^ Sumrr ^ Spnng J\ Auturr Figure 6 Diagrammatic representation of otoliths of young- of-the year coral trout, P. leopardus, kept in captiv- ity, showing relative positions of the fluorescent bands, otolith margin, translucent, and opaque zones. Bars represent the whole radius of the otolith section, measured from the nucleus to the proximal surface of the sagitta along the ventral margin of the sulcus acousticus. The dates on the top of the bars indicate time of tetracycline treatment or cap- ture and the dates on the end of the bars indicate time of death. Ferreira and Russ: Age-validation and growth rate of Plectropomus leopardus 53 so B 2+ years old coral trout, P. leopardus, and plot of residuals. Leaman, 1992). The large variability in size at a given age observed for the coral trout suggests the occurrence of individual variability in growth. The reliability of methods of growth estimation like length-frequency analysis and growth increments from marking-recapture techniques, is greatly af- fected by this kind of variation (Sainsbury, 1980), further enhancing the importance of obtaining vali- dated length-at-age estimates for exploited fish populations. The results of selective mortality are a direct effect of growth variability on the dynam- ics of abundance, and failure to consider the effects Ferreira and Russ: Age-validation and growth rate of Plectropomus leopardus 55 of different growth potentials can result in gross overestimation of optimal fishing levels (Parma and Deriso, 1990). The absence of marked seasonal changes in low latitudes has led to the general belief that tropical fishes do not form annual rings in their calcified structures (Pannella, 1974). Consequently, most of the studies of age determination of tropical fishes have concentrated on examination of daily rings. This technique, however, is time consuming and lim- ited to younger ages (see Longhurst and Pauly, 1987, and Beamish and McFarlane, 1987, for review). The presence of annual marks in otoliths has been vali- dated for an increasing number of species of tropi- cal fishes (Samuel et al., 1987; Fowler, 1990; Fer- reira and Russ, 1992; Lou, 1992) showing the poten- tial of this technique to be used routinely in tropi- cal fishery management. Acknowledgments We would like to thank P. Laycock for his assistance with the otolith readings. Many thanks to Owen Roberts, who kindly gave us access to his commer- cial fishing samples. We thank M. Maida, P. Laycock, C. Davies, M. and L. Pearce, L. Vail, A. Hogget, J. St. John, and D. Zeller for help in the field and in collecting the samples. We are grateful to C. Davies for allowing us to use his trapping and mark-recapture program to validate this study. This work was supported by grants from the Brazilian Ministry of Education (CAPES), Australian Re- search Council (ARC), Fishing Industry Research and Development Council (FIRDC), and the Great Barrier Reef Marine Park Authority (Augmentative). Literature cited Beamish, R. J., and D. A. Fournier. 1981. A method for comparing the precision of a set of age determinations. Can. J. Fish. Aquat. Sci. 38:982-983. Beamish, R. J., and G. A. McFarlane. 1987. Current trends in age determination methodology. In R. C. Summerfelt and G. E. Hall (eds.), Age and growth offish, p. 15-42. Iowa State Univ. Press, Ames. Beckman, D. W., A. L. Stanley, J. H. Render, and C. A Wilson. 1991. Age and growth-rate estimation of sheeps- head Archosargus probatocephalus in Louisiana waters using otoliths. Fish. Bull. 89 (l):l-8. Beverton, R. J. H., and S. J. Holt. 1957. On the dynamics of exploited fish popula- tions. Fish. Invest. Minist. Agric. Fish. Food (G.B.), Ser. 2 (19), 533 p. Boehlert, G. W. 1985. Using objective criteria and multiple regres- sion models for age determination in fishes. Fish. Bull. 83 (2):103-117. Brothers, E. B. 1987. Methodological approaches to the examina- tion of otoliths in aging studies. In R. C. Summerfelt and G. E. Hall (eds.), Age and growth offish, p. 319-330. Iowa State Univ. Press, Ames. Casselman, J. M. 1983. Age and growth assessment of fish from their calcified structures: techniques and tools. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 8, p. 1-17. 1990. Growth and relative size of calcified structures of fish. Trans. Am. Fish. Soc. 119:673-688. Chen, Y., D. A. Jackson, and H. H. Harvey. 1992. A comparison of von Bertalanffy and polyno- mial functions in modelling fish growth data. Can. J. Fish. Aquat. Sci. 49:1228-1235. Christensen, J. M. 1964. Burning of otoliths, a technique for age de- termination of soles and other fish. J. Cons, perm. int. Explor. Mer. 29:73-81. Ferreira, B. P., and C. M. Vooren. 1991. Age, growth and structure of vertebra in the school shark Galeorhinus galeus (Linnaeus, 1758) from Southern Brazil. Fish. Bull. 89 (1):19-31. Ferreira, B. P., and G. R. Russ. 1992. Age, growth and mortality of the inshore coral trout Plectropomus maculatus (Pisces: Serranidae) from the Central Great Barrier Reef, Aus- tralia. Aust. J. Mar. Freshwater Res. 43:1301-1312. Fowler, A. J. 1990. Validation of annual growth increments in the otoliths of a small, tropical coral reef fish. Mar. Ecol. Prog. Ser. 64:25-38. Gallucci, V. F., and T. J. Quinn. 1979. Reparameterizing, fitting and testing a simple growth model. Trans. Am. Fish. Soc. 108:14-25. Goeden, G. B. 1978. A monograph of the coral trout Plectropomus leopardus (Lacepede). Qld. Fish. Serv, Res. Bull. (l):l-42. Hirschhorn, G. 1974. The effect of different age ranges on esti- mated Bertalanffy growth parameters in three fishes and one mollusk of the northeastern Pacific Ocean. In T B. Bagenal (ed.), Ageing offish, p. 13-27. Unwin Bros., Surrey, England. Hoyer, M. V., J. V. Shireman, and M. J. Maceina. 1985. Use of otoliths to determine age and growth of largemouth bass in Florida. Trans. Am. Fis. Soc. 114:307-309. Irie, T. 1960. The growth of the fish otolith. J. Fac. Fish. Anim. Husb. Hiroshima Univ. 3 (l):203-229. 56 Fishery Bulletin 92[ I), 1994 Kimura, D. K. 1980. Likelihood methods for the von Bertalanffy growth curve. Fish. Bull. 77:756-776. Knight, W. 1968. Asymptotic growth: an example of nonsense disguised as mathematics. J. Fish. Res. Board. Can. 25 (61:1303-1307. Longhurst, A. R., and D. Pauly. 1987. Ecology of tropical oceans. Acad. Press, San Diego, 407 p. Lou, D. C. 1992. Validation of annual growth bands in the otolith of tropical parrotfishes (Scarus schlegeli Bleeker). J. Fish. Biol. 41:775-790. Loubens, G. 1980. Biologie de quelques especes de poissons du lagon Neo-Caledonien. Ill: Croissance. Cahiers de l'lndo-pacifique 2:101-153. Maceina, M. J., and R. K. Betsill. 1987. Verification and use of whole otoliths to age white crappie. In R. C. Summerfelt and G. E. Hall (eds.). Age and growth of fish, p. 267- 278. Iowa State Univ. Press, Ames. Manooch III, C. S. 1987. Age and growth of snappers and groupers. In J. J. Polovina and S. Ralston (eds.), Tropical snap- pers and groupers. Biology and Fisheries Manage- ment, p. 329-374. Westview Press, Inc., Boulder. McFarlane, G. A., and R. J. Beamish. 1987. Selection of dosages of oxvtetracycline for age validation studies. Can. J. Fish. Aquat. Sci. 44:905-909. Mcpherson, G., L., Squire and J. O'Brien. 1988. Demersal reef fish project 1984-85: age and growth of four important reef fish species. Fisheries Research Branch Technical Report No. FRB 88/6. Queensland Department of Primary Industries, Australia, 38 p. Miranda, L. E., W. M. Wingo, R. J. Muncy, and T. D. Bates. 1987. Bias in growth estimates derived from fish collected by anglers. In R. C. Summerfelt and G. E. Hall (eds.) Age and growth of fish, p. 211- 220. Iowa State Univ. Press, Ames. Moreau, J. 1987. Mathematical and biological expressions of growth in fishes: recent trends and further developments In R. C. Summerfelt and G. E. Hall (eds.), Age and growth of fish, p. 81- 114. Iowa State Univ. Press, Ames. Mosegaard, H., H. Svedang, and K. Taberman. 1988. Uncoupling of somatic growth rates in Arc- tic char iSalvelinus alpinus) as an effect of differ- ences in temperature response. Can. J. Fish. Aquat. Sci. 45:1514-1524. Mugiya, Y., N. Watabe, J. Yamada, J. M. Dean, D. G. Dunkelberger, and M. Shimizu. 1981. Diurnal rhythm in otolith formation in the goldfish, Carassius auratus. Comp. Biochem. Physiol. 68A:659-662. Mulligan, T. J., and B. M. Leaman. 1992. Length-at-age analysis: can you get what you see? Can. J. Fish. Aquat. Sci. 49:632-643. Pannella, G. 1974. Otolith growth patterns: an aid in age deter- mination in temperate and tropical fishes. In T B. Bagenal (ed.), Ageing offish, p. 28-36. Unwin Bros., Surrey, England. Parma, A., and R. B. Deriso. 1990. Dynamics of age and size composition in a population subject to size-selective mortality: ef- fects of phenotypic variability in growth. Can. J. Fish. Aquat. Sci. 47:274-289. Ralston, S. 1987. Mortality rates of snappers and groupers. In J. J. Polovina and S. Ralston (eds.), Tropical snap- pers and groupers. Biology and Fisheries Manage- ment, p. 375-404. Westview Press, Inc., Boulders. Randall, J. E. 1987. A preliminary synopsis of the groupers (Perciformes: Serranidae: Epinephelinae) of the Indo-Pacific region. In J. J. Polovina and S. Ralston (eds.), Tropical snappers and groupers. Biology and Fisheries Management, p. 89- 187. Westview Press, Inc., Boulders. Randall, J. E., and D. F. Hoese. 1986. Revision of the groupers of the Indo-Pacific genus Plectropomus (Perciformes: Serranidae) (13). Bernice Pauahi Bishop Museum, Honolulu, Hawaii, 31 p. Ricker, W. E. 1969. Effects of size-selective mortality and sam- pling bias on estimates of growth, mortality, pro- duction and yield. J. Fish. Res. Board Can. 26:479-541. Sainsbury, K. J. 1980. Effect of individual variability on the von Bertalanffy growth equation. Can. J. Fish. Aquat. Sci. 37:241-247. Samuel, M., C. P. Mathews, and A. S. Baazeer. 1987. Age and validation of age from otoliths for warm water fishes from the Arabian Gulf. In R. C. Summerfelt and G. E. Hall (eds.), Age and growth of fish, p. 253-266. Iowa State Univ. Press, Ames. Schnute, J. 1981. A versatile growth model with statistically stable parameters. Can. J. Fish. Aquat. Sci. 38:1128-1140. Spurr, A. R. 1969. A low-viscosity epoxy resin embedding me- dium for electron microscopy. J. Ultrastruct. Res. (26):31-34. Trainor, N. 1991. Commercial line fishing. The Queensland Fisherman, March: 17-25. von Bertalanffy, L. 1938. A quantitative theory of organic growth. II: Inquires on growth laws. Hum. Biol. 10:181-213. Ferreira and Russ: Age-validation and growth rate of Plectropomus leopardus 57 Watabe, N., K. Tanaka, J. Yamada, and J. Dean. 1982. Scanning electron microscope observations of the organic matrix in the otoliths of the teleost fish Fundulus heteroclitus (L. land Tilapia nilotica (L.). J. Exp. Mar. Biol. Ecol. 58:127-134. Wilkinson, L. 1989. SYSTAT: the system for statistics. Evanston, IL: SYSTAT, Inc. Wilson, C. A., R. J. Beamish, E. B. Brothers, K. D. Carlander, J. M. Casselman, J. M. Dean, A. Jearld, E. D. Prince, and A. Wild. 1987. Glossary. In R. C. Summerfelt and G. E. Hall (eds.), Age and growth offish, p. 527-530. Iowa State Univ. Press, Ames, Abstract. Red drum, Sciaenops ocellatus, from Mosquito Lagoon, east-central Florida, were examined for variation in products of nine polymorphic nuclear-gene (allozyme) loci and in mitochon- drial (mt)DNA restriction sites. Genetic data from Mosquito La- goon fish were compared to simi- lar data from red drum sampled from the northeastern Gulf of Mexico (Gulf) and the Carolina coast of the southeastern United States. Significant heterogeneity among red drum from the three areas was found in the frequencies of inferred alleles at two to three allozyme loci and in the frequen- cies of six mtDNA haplotypes. Red drum from Mosquito Lagoon were as differentiated genetically from red drum in the northeastern Gulf and Carolina coast as the latter two were from each other. Genetic data are consistent with the hy- pothesis that red drum in Mos- quito Lagoon are self-contained and at least partially isolated from red drum in other U.S. waters. Genetic distinctness of red drum (Sciaenops ocellatus) from Mosquito Lagoon, east-central Florida* John R. Gold Department of Wildlife and Fisheries Science Texas A&M University, College Station. Texas 77843 Linda R. Richardson Department of Wildlife and Fisheries Science Texas A&M University. College Station, Texas 77843 Manuscript accepted 17 August 1993 Fishery Bulletin: 92:58-66 (1994) Over the past five years, our labo- ratory has carried out studies of spatial and temporal genetic varia- tion among red drum (Sciaenops ocellatus) from the northern Gulf of Mexico (Gulf) and the Carolina coast of the southeastern United States (Bohlmeyer and Gold, 1991; Gold and Richardson, 1991; Gold et al., 1993, in press). Red drum cur- rently support important recre- ational fisheries in both the north- ern Gulf and U.S. Atlantic (Mat- lock, 1984; Mercer, 1984), and both fisheries are now regulated to re- duce growth and recruitment over- fishing (Swingle et al., 19841; Goodyear, 19892). Collectively, our genetic data have indicated that red drum in U.S. waters are sub- divided with weakly differentiated subpopulations in the northern Gulf and along the Carolina coast. No genetic heterogeneity has been found among red drum from differ- ent localities within either the northern Gulf or Carolina coast (Gold et al., 1993, in press). The ge- netic data are consistent with sev- eral aspects of red drum biology and life history that suggest red drum dispersal and gene flow among contiguous bays and estuar- ies could be extensive. These in- clude 1) transport of eggs, larvae, or juveniles from spawning locali- ties near the mouths of bays or es- tuaries to adjacent bays or estuar- ies by oceanic currents (Lyczkoski- Schultz et al., 19883), 2) movement of sexually-mature adults from bay or estuarine juvenile nurseries into deeper, offshore waters prior to spawning (Matlock, 1984), and 3) formation of large, offshore schools that can migrate extensively (Overstreet, 1983; Matlock, 1984; Swingle et al., 19841). In this study, data on allozyme and mitochondrial (mt)DNA varia- tion among red drum sampled from Mosquito Lagoon on the east coast of Florida are presented and com- pared to data from previous stud- ies. The goal of the study was to * Contribution No. 24 of the Center for Bio- systematics and Biodiversity, Texas A&M University. 1 Swingle, W., T. Leary, D. 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. Mngmt. Council and Gulf States Mar. Fish. Comm., Lincoln Cntr, Suite 331, 5401 West Kennedy Blvd., Tampa, FL. 2 Goodyear, C. P. 1989. Status of red drum stocks of the Gulf of Mexico: report for 1989. Contrib. CRD 88/89-14, Southeast Fish. Cntr, Miami Lab., Coast. Res. Div., 75 Virginia Beach Drive, Miami, FL. ■' Lyczkowski-Schultz, J., J. P. Steen Jr., and B. H. Comyns. 1988. Early life history of red drum (Sciaenops ocellatus) in the northcentral Gulf of Mexico. Mississippi- Alabama Sea Grant Consortium (Project No. R/LR-12). Gulf Coast Res. Lab., P.O. Box 7000, Ocean Springs, unpubl. ms. 58 Gold and Richardson: Sciaenops ocellatus from Mosquito Lagoon 59 test the hypothesis that red drum from Mos- quito Lagoon and other U.S. waters are geneti- cally homogeneous. Red drum in Mosquito Lagoon are of particular interest because they may represent a self-contained, at least par- tially isolated subpopulation. Evidence for the latter includes documentation within the sys- tem of both post-spawning females and red drum eggs (Murphy and Taylor, 1990; Johnson and Funicelli, 1991). In addition, physical ac- cess to the Atlantic from the lagoon is limited. In brief, Mosquito Lagoon (Fig. 1) is long and narrow (54 km x 4 km) and is separated from the Atlantic by a barrier beach. The lagoon represents the northern part of the Indian River lagoonal system and has two narrow outlets: one, Ponce de Leon Inlet, is a natural pass to the Atlantic located at the northern end of the lagoon; the other, Haulover Canal, is a man-made passageway at the southern end of the lagoon that leads into the Indian River. Access to or from the Atlantic through Ponce de Leon Inlet is restricted because of a series of islands and small passageways in the north- ern part of the lagoon. Access to or from the Atlantic through Haulover Canal (completed in 1929) would only be recent, and the nearest outlet to the Atlantic south from Haulover ca- nal is roughly 90-100 km. We also were inter- ested in studying red drum from Mosquito Lagoon because our earlier work (Gold et al., 1993, in press) did not include red drum from the east coast of Florida, an area of potential importance to tests of hypotheses regarding ge- netic subdivision between red drum from the northern Gulf and the U.S. Atlantic (Gold et al., in press). Finally, adult red drum from Mos- quito Lagoon form a large part of the broodstock used by the Florida Department of Natural Resources (FDNR) to supplement and enhance the red drum fishery in Florida wa- ters. The genetic composition of Mosquito La- goon red drum is thus important to research in stocking hatchery-raised fish. Materials and methods Red drum were collected from Mosquito Lagoon during fall 1988, spring 1990, and spring 1991. Fish were captured with trammel nets. Tissues (heart, spleen, and muscle) were removed and placed in liquid nitrogen for transport to Texas A&M University where they were stored at -80°C. Ages of all but yearling (age zero) individuals (i.e., speci- mens less than 300 mm total length) were deter- Ponce de Leon Inlet 10km Figure 1 Mosquito Lagoon, east-central Florida, showing Ponce de Leon Inlet and Haulover Canal. mined from annuli on otoliths by using methods described in Bumguardner (1991). Individuals sampled in 1988 (41 total) were sur- veyed for variation at nine polymorphic allozyme 60 Fishery Bulletin 92(1), 1994 loci: ACP-2* (acid phosphatase); ADA* (adenosine deaminase); ADH* (alcohol dehydrogenase); sAAT-1* (aspartate aminotransferase); EST-1* (esterase); GPI-B* (glucose phosphate isomerase); and PEPB' , PEPD * , and PEPS' (peptidases). Techniques for ver- tical starch gel electrophoresis, details of grinding and running buffers, starch composition of gels, protein staining, and interpretation of banding pat- terns may be found in Bohlmeyer (1989) and Bohlmeyer and Gold (1991). Designation of allelic variants was based on relative mobility to the most common allele (Allele * 100). All individuals collected (109 total) were assayed for 104 mtDNA restriction sites with 13 restriction enzymes: BamKl, Bell, EcoRV, Hindlll, Ncol, Nsil, PstI, Pvull, Seal, Spel, Stul,Xbal, and Xmnl. Meth- ods used to assay mtDNAs of individual fish may be found in Gold and Richardson (1991). Homology of fragments from single digestions was tested by multiple, side-by-side comparisons. Variant patterns exhibiting only a single band of greater than 15 kb were tested for homology by using double digestions with BawHI as described in Gold and Richardson (1991). Red drum from Mosquito Lagoon were initially subdivided into year classes and tested for hetero- geneity in both allozyme and mtDNA haplotype fre- quencies. Year classes (number of individuals) were 1985 (17), 1986 (25), 1987 (11), 1988 (7), and 1989 (49). No significant heterogeneity (P>0.05) in allozyme or mtDNA haplotype frequencies was found among year classes. Subsequent data analy- ses employed three test groups: 1) red drum from Mosquito Lagoon; 2) red drum from the northeast- ern Gulf; and 3) red drum from the Carolina coast. Data for the latter two were taken from Gold et al. (1993, 1994) and represent red drum from the fol- lowing localities: northeastern Gulf — Apalachicola Bay, Riviera Bay, and Sarasota Bay (west coast of Florida); and Carolina coast — Calibogue Sound, Charleston Bay, and North Inlet (South Carolina), and the Pamlico River and Oregon Inlet (North Carolina). A map showing these localities may be found in Bohlmeyer and Gold ( 1991 ). A summary of allele frequencies at the nine polymorphic allozyme loci and the distribution of mtDNA haplotypes in each test group are given in Appendix Tables 1 and 2, respectively. For allozyme data, tests of Hardy-Weinberg equi- librium expectations and generation of Nei's (1978) unbiased genetic distance were accomplished by using BIOSYS-1 (Swofford and Selander, 1981). Deviations from Hardy-Weinberg expectations were tested by using pooled genotypes and the chi-square statistic with one degree of freedom. Significance testing of allele-frequency differences among test groups was accomplished by using 1) the G-statis- tic (Sokal and Rohlf, 1969) on contingency tables of allele counts and the BIOM-PC program (Rohlf, 1983), and 2) the ^-statistic (DeSalle et al., 1987) on arcsin, square-root transformed allele frequen- cies. For mtDNA data, significance testing of mtDNA-haplotype frequency differences was carried out by using the G- and V-statistics as described above and a Monte Carlo randomization procedure (Roff and Bentzen, 1989). Nucleon diversities and intra- and inter-populational nucleotide sequence diversities were estimated by using equations in Nei and Tajima (1981). Analysis of mtDNA data was facilitated by the Restriction Enzyme Analysis Pack- age (REAP) of McElroy et al. (1992). Significance levels for multiple tests performed simultaneously were adjusted after Cooper (1968). Results No significant deviations from Hardy Weinberg equi- librium expectations at any of the nine polymorphic allozyme loci were found following corrections for multiple tests. Two significant deviations were found in uncorrected tests: at GPI-B* (P=0.015) and PEPS' (P=0.012) in the northeastern Gulf. Both deviations appeared to be due to rare homozygotes for low fre- quency alleles. One new allele (Allele * 110 at EST- l') was found among Mosquito Lagoon fish at a fre- quency of 1.2 percent (Appendix Table 1). Estimates of allozyme variation (Table 1) indicate that red drum from Mosquito Lagoon have fewer Table 1 Allozyme variation in red drum (S iaenops ocel- latus). Mean Mean Mean number of hetero- Test sample alleles/locus zygosity/ group size/locus <± SE) locus' (±SE) Northeastern Gulf of Mexico 246 3.9 + 0.9 0.225 ± 0.076 Mosquito Lagoon, Florida 41 2.9 ± 0.6 0.206 ± 0.081 U.S. Carol id. i Coast 176 3.9 ± 0.9 0.213 ± 0.074 1 Direct-cou nt estimate. Gold and Richardson: Sciaenops ocellatus from Mosquito Lagoon 61 alleles per locus or lower estimates of mean het- erozygosity, or both, than do red drum from the northeastern Gulf and Carolina coast. The differ- ences in genetic variation, however, are non-random across loci. Heterozygosity per locus values among Mosquito Lagoon fish at loci (e.g., ACP-2* , ADA*, ADH*, sAAT-1*, and EST-1*) where alternate alleles occurred at frequencies of five percent or greater were equivalent to values among fish from the northeastern Gulf and Carolina coast (data not shown). Differences in heterozygosity per locus val- ues were observed at loci (e.g., GPI-B *, PEPB* , and PEPD* ) where alleles occurring in a frequency of one to three percent in northeastern Gulf or Carolina coast fish, or both, were not found among Mosquito Lagoon fish (Appendix Table 1). Significant heterogeneity (P<0.05) in allele fre- quencies among test groups was found by using the G-test at ADA* (G=33.92, df=22, P=0.004) and sAAT- 1* (G=13.59, df=6, P=0.036). Additional G-tests were carried out after pooling alleles whose frequency in any sample was less than 10%. Significant hetero- geneity was again found at ADA* (G=9.62, df=4, P=0.048) and also at PEPB* (G=6.86, df=2, PM3.034). Examination of allele frequencies at ADA*, sAAT-1, and PEPB* did not reveal any striking differences among test groups, suggesting that heterogeneity was due to accumulation of small differences in fre- quencies of rare alleles. At ADA*, for example, the frequency of Allele *115 was higher among Mosquito Lagoon fish and lower among Carolina coast fish; whereas the frequencies of Alleles *90 and *85 were higher among northeastern Gulf fish (Appendix Table 1). At sAAT-1* and PEPB*, slight frequency dif- ferences were apparent for Allele * 110 (higher in Mosquito Lagoon fish) and Allele 115 (higher in northeastern Gulf fish and absent from Mosquito Lagoon fish), re- spectively (Appendix Table 1). The observation that G-test heteroge- neity was due to small, cumula- tive frequency differences was cor- roborated by V-tests where no sig- nificant heterogeneity (P>0.05) in allele frequencies was found at any locus following corrections for multiple tests. MtDNA fragment patterns from single digestions with 13 restric- tion enzymes generated 36 com- posite mtDNA haplotypes among fish from Mosquito Lagoon, eleven of which (numbers 114, 134-143) have been found only in Mosquito Lagoon red drum (Appendix Table 2). Estimates of mtDNA variation (Table 2) indi- cated that nucleon diversity (the probability of any two individuals differing in mtDNA haplotype) was highest in red drum from the northeastern Gulf and lowest in red drum from the Carolina coast; whereas intrapopulational nucleotide sequence diversity (the genetic difference between any two individuals) was greatest among Mosquito Lagoon fish. These esti- mates of mtDNA variation are among the highest reported to date for a non-clupeid, marine fish spe- cies (Richardson and Gold, 1993). Highly significant heterogeneity in mtDNA- haplotype frequencies among test groups and be- tween pairwise comparisons of test groups were found in both G-tests and Monte Carlo bootstrapping (Table 3). These results indicate that all three test groups differ significantly from each other. V-tests, carried out on haplotypes found in ten or more individuals (12 haplotypes total), identified six haplotypes (Table 4) that differed significantly among test groups. Genetic distances based on allozymes and mtDNAs (Table 5) indicate that red drum from Mosquito Lagoon are at least as diver- gent genetically from red drum in the northeastern Gulf and Carolina coast as the latter two are from each other. Discussion Tests of heterogeneity clearly indicate that red drum from Mosquito Lagoon differ genetically from red drum in the northeastern Gulf and along the Caro- lina coast and that at least three subpopulations of red drum occur in U.S. waters. That the genetic differences appear more pronounced in mtDNA than Table 2 MtDNA variation in red drum (Sciaenops ocellatus). Nucleotide Number Number sequence Test of of Nucleon diversity group individuals haplotypes diversity (± SD)' Northeastern Gulf of Mexico 247 49 0.947 0.557 ± 0.298 Mosquito Lagoon, Florida 109 36 0.912 0.597 ± 0.321 U.S. Carolina Coast 174 43 0.904 0.560 ± 0.351 1 Values are in percent Standard d eviations are used instead of standard errors be- cause of the large number of pairwise comparisons used to generate mean values. 62 Fishery Bulletin 92(1), 1994 Table 3 Results of tests for heterogeneity in among red drum (Sciaenops ocellatu Mexico, Mosquito Lagoon, Florida, a mtDNA haplotype frequencies s) from the northeastern Gulf of nd the U.S. Carolina coast. Test group Results of G-tests P-value from Monte Carlo randomizations G-score P-value Northeastern Gulf vs. Mosquito Lagoon vs. Carolina Coast 159.5 <0.001' <0.001 Northeastern Gulf vs. Mosquito Lagoon 73.9 <0.0012 <0.00T Northeastern Gulf vs. Carolina Coast 76.2 <0.0013 <0.001 Mosquito Lagoon vs. Carolina Coast 66.2 <0.0014 0.006 Degrees of freedom in G-tests: 48' 182, 193, and 27J. Table 4 Frequency7 of six sign ificantly he terogeneous mtDNA haplotypes of red drum {Sciaenops ocellatus) in the northeast- era Gulf of Mexico, Mosq uito Lagoon, Florida, and the U.S. Carolina coast. Northeastern Mosquito Carolina Probability Haplo- Gulf Lagoon Coast value from type (rc=247) (rc=109) (n = 174) V-test2 8 13.3 23.8 10.3 =0.010 9 7.7 13.8 26.4 <0.001 11 9.3 1.8 7.5 =0.019 12 0.0 7.3 3.4 <0.001 21 4.4 i) i) ii 6 =0.004 29 1 () ii i) 1.7 =0.021 ' Values are in percent. 2 After DeSalle et al. (1987). in (presumed) nuclear-coding genes is not surpris- ing, given that mtDNA is expected to be at least four times more sensitive to population substructuring (Birky et al., 1983; Templeton, 1987). Because pre- vious studies (Gold et al., 1993, in press) found no evidence of genetic heterogeneity among red drum from eleven estuaries or bays in the northern Gulf or among red drum from five estuaries or bays along the Carolina coast, red drum from Mosquito Lagoon are unusual in representing a genetically distinct red drum subpopulation existing within a single bay or estuary. Campton (1992)4 examined red drum from Mosquito Lagoon for allelic variation at several allozyme loci and found genetic homogeneity among red drum from Mosquito Lagoon, the north- ern Gulf, and the Carolina coast. He suggested that our initial study (Bohlmeyer and Gold, 1991) of allozyme variation among northern Gulf and Carolina coast red drum did not account for tem- poral variation among samples within localities. Our subsequent studies (and this one), however, have included temporal sampling of variation in both allozymes and mtDNA and have demonstrated that weak (but significant) genetic heterogeneity exists (Gold et al., 1993, in press). Sampling error associated with specimen procure- ment in varying time and space may account for the different results ob- tained in Campton's (1992)4 study and this one. However, in Campton's (1992)4 study, the total G-statistic, obtained by summing individual G-values and their associated degrees of freedom, was signifi- cant at the 0.01 level. This suggests the existence of spatial or temporal genetic heterogeneity, or both, among the locali- ties sampled. Genetic differentiation of red drum in Mosquito Lagoon is consistent with the hypothesis that red drum in Mosquito Lagoon represent a self-contained, at least partially isolated subpopulation. Three lines of evidence support this hy- pothesis. First, genetic differences be- tween red drum from Mosquito Lagoon and red drum sampled elsewhere involve frequencies of alleles at two or three pu- tative nuclear-gene loci and frequencies of at least six mtDNA haplotypes. Differentiation of several, presumably independent and selectively-neutral, genetic markers suggests a genome-wide effect re- lated to at least partial isolation and reduced gene flow (Wright, 1978; Hartl and Clark, 1989). Second, inferred nuclear-gene alleles present in low fre- quency in red drum sampled outside of Mosquito Campton, D. E. 1992. Gene flow estimation and population struc- ture of red drum iSaaenops ocellatus) in Florida. Final Rep. Coop. Agrmt. No. 14-16-009-1522, U.S. Fish & Wildl. Serv, Natl. Fish. Res. Cntr., 7920 N.W. 71st St., Gainesville, FL. Gold and Richardson: Sciaenops ocellatus from Mosquito Lagoon 63 Table 5 Matrix of Nei's (1978) unbiased genetic distance based on allozymes (upper diagonal) and Nei and Tajima's (1981) corrected interpopulational nucle- otide sequence divergence based on mtDNAs (lower diagonal) among red drum (Sciaenops ocellatus) from the northeastern Gulf of Mexico, Mosquito Lagoon, Florida, and the U.S. Carolina coast. Interpopulational nucleotide sequence di- vergence values are in percent. Northeastern Mosquito Carolina Gulf Lagoon Coast Northeastern Gulf 0.000 0.001 Mosquito Lagoon 0.006 0.002 Carolina Coast 0.006 0.009 Lagoon were not found in red drum from Mosquito Lagoon; whereas one inferred allele and eleven mtDNA haplotypes were unique to red drum from Mosquito Lagoon. The distribution of low frequency nuclear-gene alleles and mtDNA haplotypes is con- sistent with reduced gene flow concomitant with allele-frequency drift expected in isolated subpopu- lations. Finally, both females with ovaries contain- ing postovulatory follicles and spawned red drum eggs have been documented in Mosquito Lagoon (Murphy and Taylor, 1990; Johnson and Funicelli, 1991), clearly indicating that red drum spawn within the system. Assuming red drum in Mosquito Lagoon represent a partially isolated, self-contained subpopulation, one question of interest is how long the subpopula- tion has been semi-isolated. Geological evidence (Mehta and Brooks, 1973, cited from Johnson and Funicelli, 1991) indicates that several tidal inlets once connected Mosquito Lagoon to the Atlantic, the last of which is estimated to have closed about 1,500 years ago. Assuming some variation in the geologi- cal estimate, this date does not differ substantially from an estimate of 2,900 ± 1,550 (SD) years based on 1) a corrected interpopulational nucleotide se- quence divergence (between red drum in Mosquito Lagoon and red drum elsewhere) of 0.0058 ± 0.0031 (SD) percent, and 2) an evolutionary rate for verte- brate mtDNA of 0.01 substitutions/bp/lineage/Myr (Brown et al., 1979; Wilson et al., 1985). Given on- going debates about molecular clocks, the correspon- dence between the two temporal estimates is note- worthy. Because the genetic distinctness of Mosquito La- goon red drum appears to stem largely from physi- cal isolation, the biological reasons for subdivision between red drum in the northern Gulf and those along the Carolina coast remain unknown. Possible reasons for this subdivision could include 1) current patterns between the Gulf and U.S. Atlantic, 2) absence of suitable near-shore habitats along the southeastern coast of Florida, or 3) differences in biogeographic provinces (Gold et al., 1993, in press). Similar genetic discontinuities between U.S. Atlan- tic and Gulf coast fauna have been described by Avise and co-workers (reviewed in Avise, 1992). Their hypothesis is that the concordant phylogeographic patterns provide evidence of simi- lar vicariant histories that are tentatively related to episodic changes in environmental conditions dur- ing the Pleistocene (Avise, 1992). The relative inac- cessibility of Mosquito Lagoon suggests that sam- pling red drum from north or south of Mosquito Lagoon may be more informative for testing hypoth- eses regarding phylogeographic subdivision between the northern Gulf and the U.S. Atlantic. A last point to consider is the use of Mosquito Lagoon red drum as broodstock for stock enhance- ment programs. It could be argued that red drum from Mosquito Lagoon differ genetically from red drum sampled elsewhere (e.g., the northeastern Gulf) and should be used only for stock enhancement at localities where no genetic differences exist. Al- ternatively, it could be argued that the genetic dis- tinctiveness of red drum in Mosquito Lagoon is rela- tively small and possibly inconsequential. This fol- lows from the observation that the documented ge- netic difference between red drum in Mosquito La- goon and red drum sampled elsewhere is consider- ably less than that, on average, among races of man (Cann et al., 1987). One other consideration might be to cross red drum from Mosquito Lagoon with red drum from elsewhere (e.g., the northeastern Gulf) in order to increase performance from potential heterotic effects. Acknowledgments Assistance in procuring red drum specimens from Mosquito Lagoon was provided by J. Burch, J. Camper, B. Denis, C. Furman, M. Murphy, G. Ramos, and D. Roberts. Their assistance is grate- fully acknowledged. Special thanks are extended to C. Amemiya and D. Roberts for providing no-cost lodging during field trips. We also thank B. Colura and B. Bumguardner for carrying out age determi- nations from otoliths, D. Bohlmeyer and C. Furman for assistance in the laboratory, R. Taylor for pro- viding historical information on the construction of Haulover Canal, and M. Murphy for providing criti- 64 Fishery Bulletin 92(1). 1994 cal comments on a draft of the manuscript. Work was supported by the Texas A&M University Sea Grant College Program (grants NA85AA-D-SG128 and NA89AA-D-SG139), by the MARFIN Program of the U.S. Department of Commerce (grants NA89- WC-H-MF025 and NA90AA-H-MF107), and by the Texas Agricultural Experiment Station (Project H- 6703). This paper represents number XI in the se- ries "Genetic Studies in Marine Fishes." Literature cited Avise, J. C. 1992. Molecular population structure and the bio- geographic history of a regional fauna: a case his- tory with lessons for conservation biology. Oikos 63:62-76. Birky Jr., C. W., T. Maruyama, and P. Fuerst. 1983. Mitochondrial DNAs and phylogenetic relationships. In S. K. Dutta (ed.), DNA system- atics, p. 107-137. CRC Press, Boca Raton, FL. Bohlmeyer, D. A. 1989. A protein electrophoretic analysis of popula- tion structure in the red drum (Sciaenops ocellatus). M.S. thesis, Texas A&M University, College Station, TX. Bohlmeyer, D. A., and J. R. Gold. 1991. Genetic studies in marine fishes. II: A pro- tein electrophoretic analysis of population struc- ture in the red drum Sciaenops ocellatus. Mar. Biol. 108:197-206. Brown, W. M., M. George Jr., and A. C. Wilson. 1979. Rapid evolution of animal mitochondrial DNA. Proc. Natl. Academy Sci. (USA) 76:1967- 1971. Bumguardner, B. W. 1991. Marking subadult red drums with oxytetracycline. Trans. Am. Fish. Soc. 120:537-540. Cann, R. L., M. Stoneking, and A. C. Wilson. 1987. Mitochondrial DNA and human evolution. Nature 325:31-36. Cooper, D. W. 1968. The significance level in multiple tests made simultaneously. Heredity 23:614-617. DeSalle, R., A. Templeton, I. Mori, S. Pletscher, and J. S. Johnston. 1987. Temporal and spatial heterogeneity of mtDNA polymorphisms in natural populations of Drosophila mercatorum. Genetics 116:215-233. Gold, J. R., and L. R. Richardson. 1991. Genetic studies in marine fishes. IV: An analysis of population structure in the red drum (Sciaenops ocellatus) using mitochondrial DNA. Fish. Res. 12:213-241. Gold, J. R., L. R. Richardson, C. Furman, and T. L. King. 1993. Mitochondrial DNA differentiation and popu- lation structure in red drum (Sciaenops ocellatus) from the Gulf of Mexico and Atlantic Ocean. Mar. Biol. (In press.) Gold, J. R., T. L. King, L. R. Richardson, D. A. Bohlmeyer, and G. C. Matlock. In press. Genetic studies in marine fishes. VII: Allozyme differentiation within and between red drum (Sciaenops ocellatus) from the Gulf of Mexico and Atlantic Ocean. J. Fish Biol. 116:175-185. Hartl, D. L., and A. G. Clark. 1989. Principles of population genetics, 2nd ed. Sinauer Assoc, Inc., Sunderland, MA. Johnson, D. R., and N. A. Funicelli. 1991. Spawning of the red drum in Mosquito La- goon, east-central Florida. Estuaries 14:74-79. Matlock, G. C. 1984. A basis for the development of a management plan for red drum in Texas. Ph.D. diss., Texas A&M University, College Station, TX. McElroy, D., P. Moran, E. Bermingham, and I. Kornfield. 1992. REAP-The Restriction Enzyme Analysis Package. J. Hered. 83:157-158. Mehta, A. J., and H. K. Brooks. 1973. Mosquito Lagoon barrier beach study. Shore and Beach 41:27-34. Mercer, L. 1984. A biological and fisheries profile of red drum, Sciaenops ocellatus. Spec. Sci. Rep. 41, North Carolina Dep. Nat. Resour. Community Dev, Div. Mar. Fish., Raleigh, NC. Murphy, M. D., and R. G. Taylor. 1990. Reproduction, growth, and mortality of red drum, Sciaenops ocellatus, in Florida. Fish. Bull. 88:531-542. Nei, M. 1978. Estimation of average heterozygosity and ge- netic distance from a small number of indi- viduals. Genetics 89:583-590. Nei, M., and F. Tajima. 1981. DNA polymorphism detectable by restriction endonucleases. Genetics 97:145-163. Overstreet, R. M. 1983. Aspects of the biology of the red drum, Sciaenops ocellatus, in Mississippi. Gulf Res. Rep. (Suppl.) 1:45-68 Richardson, L. R., and J. R. Gold. 1993. Mitochondrial DNA variation in red grouper (Epinephelus morio) and greater amberjack (Seriola dumerili) from the Gulf of Mexico. ICES J. Mar. Sci. 50:53-62. Roff, D. A., and P. Bentzen. 1989. The statistical analysis of mitochondrial poly- morphisms: chi-square and the problem of small samples. Mol. Biol. Evol. 6:539-545. Rohlf, F. J. 1983. BIOM-PC: a package of statistical programs to accompany the text BIOMETRY. W. H. Free- man & Co., San Francisco, CA. Sokal, R. R., and F. J. Rohlf. 1969. Biometry. The principles and practice of sta- Gold and Richardson: Saaenops ocellatus from Mosquito Lagoon 65 tistics in biological research. W. H. Freeman & Co., San Francisco, CA. Swofford, D. L., and R. B. Selander. 1981. BIOSYS-1: a FORTRAN program for the comprehensive analysis of electrophoretic data in population genetics and systematics. J. Hered. 72:281-283. Templeton, A. R. 1987. Genetic systems and evolutionary rates. In K F. S. Campbell and M. F. Day (eds.), Rates of evolu- tion, p. 218-234. Australian Acad. Sci., Canberra. Wilson, A. C, R. L. Cann, S. M. Carr, M. George Jr., U. B. Gyllensten, K. M. Helm-Bychowski, R. G. Higuchi, S. R. Palumbi, E. M. Prager, R. D. Sage, and M. Stoneking. 1985. Mitochondrial DNA and two perspectives on evolutionary genetics. Biol. J. Linnaean Soc. 26:375-400. Wright, S. 1978. Evolution and the genetics of popu- lations. Univ. Chicago Press, Chicago, IL. Appendix Table 1 Allele frequencies at nine polymorphic loci among red drum iSciaenops ocellatus) from the northeastern Gulf of Mexico, Mosquito Lagoon, Florida, and the U.S. Carolina coast. Northeastern Mosquito U.S. Northeastern Mosquito U.S. Locus Gulf of Lagoon, Carolina Locus Gulf of Lagoon, Carolina allele Mexico' Florida coast7 allele Mexico' Florida coast' ACP-2' '125 0.002 0.012 0.000 EST-f '115 0.087 0.073 0.063 '110 0.000 0.012 0.000 '100 0.911 0.915 0.937 '100 0.911 0.915 0.898 (n) (246) (41) (175) "95 0.089 0.073 0.102 in) (246) (41) (176) ADA' '150 0.000 0.012 0.003 GPI-B' '130 0.036 0.024 0.028 '-110 0.004 0.000 0.003 '125 0.315 0.354 0.372 '-100 0.976 1.000 0.971 '118 0.006 0.000 0.003 '-50 0.020 0.000 0.026 '115 0.081 0.122 0.028 in) (247) (41) (176) '113 0.002 0.000 0.003 '110 0.061 0.012 0.060 PEPB' '100 0.443 0.452 0.469 '115 0.022 0.000 0.006 '90 0.010 0.000 0.003 '100 0.974 1.000 0.991 '85 0.024 0.000 0.003 '85 0.004 0.000 0.003 '78 0.000 0.000 0.000 M (247) (41) (176) '75 0.018 0.024 0.028 '65 0.004 0.000 0.000 PEPD' in) (247) (41) (176) '115 0.002 0.012 0.009 '100 0.968 0.988 0.968 ADH' '85 0.030 0.000 0.020 '-100 0.508 0.451 0.566 '75 0.000 0.000 0.003 '-75 0.458 0.525 0.391 in) (247) (41) (176) '-50 0.028 0.012 0.020 '-20 0.006 0.012 0.023 PEPS' in) (246) (41) (175) '105 0.040 0.024 0.023 '100 0.958 0.976 0.977 sAAT-1' '95 0.002 0.000 0.000 '120 '110 0.000 0.134 0.012 0.171 0.017 0.120 in) (247) (41) (176) '100 0.856 0.817 0.854 ' Data are from Gold et al. (in press). '90 0.010 0.000 0.009 in) (242) (41) (175) 66 Fishery Bulletin 92(1), 1994 Appendix Table 2 Distribution of mtDNA haplotypes among red drum iSciaenops ocellatus) from the northeastern Gulf of Mexico, Mosquito Lagoon, Florida, and the U.S. Carolina coast. Composite North- Composite North- mtDNA eastern Mosquito U.S. mtDNA eastern Mosquito U.S. Haplo digestion Gulf of Lagoon, Carolina Haplo- digestion Gulf of Lagoon, Carolina type pattern' Mexico- Florida coast- type pattern' Mexico1' Florida coast- 1 ABAAAAAAAAAAA Ill 4 10 56 AGAAAAAAAAAAA 1 2 ABCCAAAAAAAAA in 6 3 57 AAAAAABAAAEAA — — 1 3 ABBACAAAAAAAA 11 1 10 58 BBAAAFAAAAAAA 3 — — 4 EAAAAABAAAAAA 1 — — 60 FBBAAAAAAACAA — — 1 5 BAAAACBAAAAAA 1 — — 61 AAAAAAAADAAAA — — 1 6 CBAAAAAAAAAAA 2 1 1 62 BBBAAAAAAAAAA — — 1 7 AAABAAAAAAAAA 7 1 — 64 AAAEAABAAAAAA 5 — — 8 AAAAAABAAAAAA 33 26 18 66 BBADAAAAAAAAA — — 1 9 BAAAAAAAAAAAA 19 15 46 68 BBAEAAAAAAAAA 1 — — 10 BBAAAAAAAAAAA 9 2 4 69 AFAAAABAAAAAA 4 — — 11 AAAAAAAAAAAAA 23 2 13 70 ACAAAAAACAAAA 1 — — 12 CBAAAABAAAAAA — 8 6 76 BAAAAAAAABAAA 2 — — 13 ABCAAACAAAAAA 1 — 4 77 AB AAAG FAAAAAA 1 — — 14 BBFAAAAAAABAB — — 4 82 ABAAAAFAAAAAA 4 — — 15 AAAAAABACAAAA — 1 2 89 BIAAAAAAAAAAA — — 1 16 ACAAAAAAAAAAA 6 2 4 90 BAAAAAGAEAAAA — 1 1 18 ABAACAAAAAAAA 5 2 1 91 AAAAAAABAAAAA — — 1 L9 BBAAADAAAAAAA — 2 5 92 ABBAFAAAAACAA — — 1 20 ABBAAAAAAACAA — 3 2 93 AAAFGAAAEAAAA — — 21 BABAAAAAAAAAA 1 1 — 1 94 AAAAAABAAAADA — — 22 BAAAAABAAAAAA 4 1 2 95 BAAAAHAAAAAAC 2 — — 23 AAAABAAAAAAAA 17 6 8 96 BCAAAAAAAAAAA — — 24 AAAAAAAAAAAAC 5 2 1 97 H B AAAAAAAAAAA — — 25 ADCCAAAAAAAAA 2 — 1 98 BAAABAAAAAAAA — — 26 BABABAAAAAAAA 3 1 1 99 B B B AAAAAAAFAA — — 27 AACCAAAAAAAAA — 5 1 100 AAIAAABAAAAAA — — 28 ABAADAAAAAAAA — 2 2 101 ABCCAAFAAAAAA — — 29 AAAAABABAAAAA 10 — 3 106 AAAAIAAAAAAAC — — 31 DBCAAAAAAAAAA 1 — — 107 BAAAABABAAAAA 2 — — 35 AB B AAAAAAAAAA — 2 4 114 AC B AAAAAAAAAA — 1 — 36 ABADAAAAAAAAA 1 — 1 121 ABADAAAADAAAA 1 — — 45 BABAAABAAAAAA 2 — — 134 AAAAGABAAAAAA — — 46 ABEAAAAAAAAAA 1 1 — 135 BBJAADAAAAAAA — — 47 BBAAAFAAEAAAA 2 — — 136 BBADAAABAAAAA — — 48 AAEAAAAAAAAAA 1 — — 137 BBAAAAACAABAA — — VJ CBBAAAAAAAAAA 3 — — 138 BBAAAAAAAABAB — — 50 BBHAAAAAAABAB — — 139 AAACAAAAAAAAA — — 51 ABCAAAAAAAAAA — 1 140 ABACAAAAAAAAA — — 52 BBAAAAACAABAB — — 141 AACABAAAAAAAA — — 53 ABBAAAAAAAFAA 2 — 142 AACAAABAAAAAA — — 54 B.AAEAAAAAAAAA — — 143 ABAACAAAAAAAA — — 55 AHCCAAAAAAAAA — — ' Letters (from left to right! are digestion patterns for: Nco\, Sc/I, Seal, Pvull, Spel, Xbal, Xmnl, HindlU, Stul, BamHl, EcoRV, Pstl, and Nsil Details regarding fragment sizes of individual digestion patterns are available upon request. 2 Data are from Gold et al (1993). AbStfclCt. Microzooplankton retained by a 41-um mesh was sampled along a 50-km transect in the Shelikof Strait between Kodiak Island and the Alaska Pen- insula. We sampled once each year during spring (April-May) 1985- 1989 using Niskin bottles closed at 10-m depth intervals. Sampling was conducted near the time and place of peak hatching of walleye pollock (Theragra chalcogramma) larvae. We examined horizontal and vertical patterns of abundance of potential prey organisms, espe- cially copepod nauplii, and de- scribed these patterns with respect to the oceanography of the Strait. Hydrography, nutrients, chloro- phyll-a and net zooplankton data also were collected and were used to help interpret the microzoo- plankton patterns. Copepod nau- plii composed from 46 to 82% of all organisms in the formalin-pre- served samples. Eggs (3-35%), ro- tifers (up to 14%) and loricate tintinnids (up to 11%) were the next most abundant taxa. The abundance of microzooplankton varied greatly across the Strait and, for copepod nauplii, had maxima associated with the Alaska Coastal Current. A meso- scale feature in the coastal current appeared to influence the distribu- tion of microzooplankton and may affect feeding conditions for larval walleye pollock. Significant differ- ences in abundance of copepod eggs and nauplii were detected between some transects. The inte- grated, 0-60 m depth, across-strait average abundance of copepod nauplii varied from a low of 5.8 x 103 nr2 (sampled in 1985) to a high of 17.6 x 103 nr2 (1987). The maximum concentration found in these same transects varied from 18 to 144 L'1, respectively. Be- tween 60 and 70% of the nauplii sampled were of a size (>125 urn total length) composing approxi- mately 98% of the naupliar diet of larval walleye pollock in spring. Distribution and abundance of copepod nauplii and other small (40-300 jim) zooplankton during spring \n Shelikof Strait, Alaska* Lewis S. Incze Bigelow Laboratory for Ocean Sciences West Boothbay Harbor. ME 04575 Terri Ainaire Bigelow Laboratory for Ocean Sciences West Boothbay Harbor, ME 04575 Manuscript accepted 17 September 1993 Fishery Bulletin 92:67-78 (1994) The high mortality rate of marine fish larvae is attributed to high rates of predation (Moller, 1984; Bailey and Houde, 1989), sensitiv- ity to feeding conditions (Thei- lacker and Watanabe, 1989) and interactions between these factors (Houde, 1987; Purcell and Grover, 1990). The larvae of temperate fishes often occur during spring, when planktonic production is in early stages of its annual cycle and is easily disrupted or delayed by adverse conditions. Also, larvae have small search volumes and generally small energy reserves (Bailey and Houde, 1989). Thus, a spatial or temporal "match" or "mismatch" between the demand for larval food and its availability seems intuitively likely and has been the subject of much research (e.g., Lasker, 1981; Buckley and Lough, 1987; Cushing, 1990). The quest to quantify feeding relation- ships has led to continuing efforts to reduce container effects in ex- perimental studies (Gamble and Fuiman, 1987; McKenzie et al., 1990), to improve the sensitivity of physiological measurements (e.g., Buckley et al., 1990), to understand the small-scale distribution of prey in the field (Owen, 1989), and to understand the role of mixing in enhancing or retarding interactions between predator and prey (Rothschild and Osborne, 1988; Davis et al., 1991). In the ocean, feeding takes place in a complex spatial array of biological and physical conditions. Any study of rate-influencing processes that af- fect larvae must take into account the distribution of these conditions in order to understand effects at the population level. In this paper we examine the springtime community of small zooplankton, primarily copepod nauplii, that may be prey for larval walleye pollock, Theragra chal- cogramma, in Shelikof Strait, Alaska (Fig. 1), and we report on the distribution and abundance of these organisms with respect to oceanographic conditions. A large population of walleye pollock spawns in the Strait in late March and early April, forming dense ag- gregations of planktonic eggs in the deepest part of the sea valley be- tween Kodiak Island and the Alaska Peninsula. Hatching occurs from middle or late April through early May (Kendall et al., 1987; Incze et al., 1989; Yoklavitch and Bailey, 1990). While the eggs re- main mostly below 150 m, larvae * Bigelow Laboratory Contribution No. 93- 006. Fisheries Oceanography Coordinated Investigations Contribution No. 0186. 6 7 6 3 Fishery Bulletin 92(1). 1994 160° 150° 140° . i -60 N 156° 154° 152c 156° 154° Figure 1 Top panel shows location of the study area and a generalized scheme of the surface circulation. Middle and bottom panels show Shelikof Strait and the sampling transect. Stations are numbered con- secutively beginning with 55 near the Kodiak Island shore; only the end and middle stations are labeled. are found primarily in the upper 50 m (Kendall et al., 19931) and have been shown to prey heavily on copepod nauplii during the first several weeks of development (Dagg et al., 1984; Kendall et al., 1987; Canino et al., 1991). The upper water column of Shelikof Strait con- sists of at least three distinct water types (Reed and 1 A. W. Kendall Jr., L. S. Incze, P. B. Ortner. S. R. Cummings, and P. K. Brown. 1993. The vertical distribution of eggs and larvae of walleye pollock in Shelikof Strait, Gulf of Alaska. Sub- mitted to Fish. Bull. Schumacher, 1989). A cold, slightly freshened, tur- bid coastal water band of narrow width (<10 km) remains near the Alaska Peninsula (northern) side of the Strait. This water receives its signature from glacial melt-waters draining into Cook Inlet at the northern end of the Strait and thus varies season- ally in volume. A second water type is encompassed in the Alaska Coastal Current (ACC), part of a baroclinic current running more or less continuously along 1000 km of the Alaskan south coast. The ACC flows from northeast to southwest in a band approxi- mately 20 km wide through the middle portion of the Strait, but it has a highly variable current struc- ture marked by numerous baroclinic instabilities (Mysak et al., 1981; Vastano et al., 1992). In the vertical, the southward flow of the ACC induces an opposite bottom flow of more saline, nutrient rich water that enters the sea valley at the shelf edge south of the study area (Fig. 1; see Reed et al., 1987). A third water type is made up of waters from a mixture of sources, including outer shelf and oceanic intrusions. Most of this water enters from the north and flows the length of the Strait along Kodiak Is- land, but current meter measurements and satellite imagery show that water sometimes enters from the south (Schumacher, 199 12). The work reported here was undertaken as part of a multi-disciplinary program (Fisheries Oceanog- raphy Coordinated Investigations: FOCI) aimed at understanding the influence of environmental fac- tors on the early life history of walleye pollock spawned in the Strait (Schumacher and Kendall, 1991). An extensive grid of sampling stations occu- pied in early May 1985, the first year of the pro- gram, showed that the spring bloom of large diatoms did not occur homogeneously throughout the Strait. Rather, in that year, large diatoms bloomed first in a band which occupied the longitudinal mid-portion of the Strait (Incze, unpubl. observ.). Hydrographic data show that this feature was in the ACC, which had at that time a shallower upper mixed layer than elsewhere in the Strait. It seemed likely, therefore, that conditions affecting the feeding and growth of larval walleye pollock would be subject to dynam- ics of the ACC and would differ across the Strait as well as through time. As part of the research pro- gram, a standard across-strait transect was estab- lished near the southern end of the Strait proper (about halfway up the sea valley: Fig. 1). This transect has been sampled with a CTD (Conductiv- ity, Temperature, Depth) as often as ship and re- search schedules have permitted. Biological sam- 2 J. Schumacher. 1991. Pacific Marine Environmental Labora- tory, Seattle, WA, unpubl. data. Incze and Ainaire: Distribution and abundance of copepod naupln 69 pling begins along this transect near the time of larval hatching each spring and proceeds down-cur- rent (westward) over time. In this paper we report on across-shelf patterns of abundance and vertical distribution of copepod nauplii and other small zoop- lankton from 1985 through 1989 and relate these patterns to hydrographic conditions, chlorophyll concentrations, and distributions of selected taxa of adult female copepods. Materials and methods For convenience, we use the term microzooplankton to refer to small zooplankton captured and pre- served by methods described below. Hydrography, nutrients, and microzooplankton were sampled with a CTD and rosette sampler along a transect of sta- tions across Shelikof Strait, Alaska, during spring from 1985 through 1989 (Fig. 1) (sampling dates are listed in Table 2). Hydrographic (CTD) data were obtained near bottom at 7 stations at 7-km inter- vals and were processed to give 1-m averaged data of salinity, temperature and density. Nutrients were sampled at five or more stations on the transect by removing water samples from 10-L Niskin bottles tripped at standard depths of 10, 20, 30, 50, 75, and 100 m; below this depth we sampled with lower reso- lution, generally at 50-m intervals, plus a sample near bottom. Nutrient concentrations were deter- mined after the cruise by using standard autoanalyzer techniques on frozen samples (Whitledge et al., 19813). Chlorophyll data were obtained from nutrient sampling depths in the up- per 100 m in 1988 and 1989. Analyses were con- ducted on board the vessel following methods of Yentsch and Menzel (1963) as modified by Phinney and Yentsch (1985) with 0.45-|im Millipore HA ac- etate filters. Microzooplankton was sampled from Niskin bottles were tripped at 10-m intervals from 0 to 60 m in 1985 and from 10 to 60 m in other years. We used the same bottles as for nutrient and chloro- phyll samples for those depths which were common to all. The number of stations sampled varied over the years, beginning in 1985 with stations 55, 58, and 61. In 1986 and 1987 we included station 60. In 1988 we sampled all seven stations along the transect, and in 1989 we sampled all except station 57. Niskin bottles were sampled for nutrients and chlorophyll when called for; the remaining contents of the bottles were filtered through small (6 x 18 cm) 3 Whitledge, T. E., S. C. Molloy, C. J. Patton, and C. D. Wirick. 1981. Automated nutrient analyses in seawater. Tech Rep. No. BNL-51398, Brookhaven Natl. Lab., Upton, NY. conical nets made of 41-um mesh nylon netting. Material retained on the netting was flushed into 4— ounce (120 mL) glass jars by using 0.45-um fil- tered seawater and was preserved in a final solu- tion of 5% formalin:seawater. Larger zooplankton was sampled at all seven stations by using 60-cm diameter bongo samplers equipped with 333-um mesh nets and towed in double-oblique fashion from the surface to about 10 m off bottom. From 1986 onward, a 20-cm bongo sampler with 150-um mesh nets was attached to the towing wire 1 m above the larger sampler to try to improve on the sampling of smaller copepods. Properties of each tow were moni- tored by time, wire angle from the towing block, mechanical flowmeters mounted across the mouth of each net, and a bathykymograph attached to the bridle of the large bongo. In the laboratory, each microzooplankton sample was filtered onto a 41-um mesh sieve, stained over- night in Rose Bengal, transferred to a 10-mL scin- tillation vial and examined in approximately 2-mL aliquots. Microzooplankton was analyzed by using a stereo dissecting microscope equipped with an image analysis system consisting of a high-resolu- tion video camera and computer software to make measurements and record data (Incze et al., 1990). The microscopist made identifications, placing each organism into one of thirteen categories (Table 1), and directed the orientation of measurements. Cope- pod nauplii were measured for total length (TL) and maximum width. Total length was the carapace length ("prosome"), plus the abdomen ("urosome") when present. The latter section often was curled beneath the carapace, necessitating measurement along a curved line. We measured the diameter of eggs and only the total body length of all other or- ganisms. In most cases the entire sample was ana- lyzed, but 25% of the original sample sometimes pro- vided adequate counts, which we established as at least 50 nauplii per sample. Subsampling was done by increasing the stored sample volume to 200 mL, dividing as necessary, then recondensing the mate- rial for examination. Subsampling was checked for accuracy by completely analyzing both half-portions from 30 samples. Final counts of microzooplankton were corrected for the subsampling fraction and for differences in the original volume of water filtered and are presented as number of organisms per li- ter. Integrated abundances (No. m~2) were estimated for the upper 60 m of the water column by using a trapezoidal algorithm. Vertical and horizontal patterns of micro- zooplankton distribution were plotted by using an inverse distance gridding technique ("Surfer", Golden Software, Inc., Golden, CO) with a grid size 70 Fishery Bulletin 92(1). 1994 Table 1 (A) Composition of microzooplankton in Shelikof Strait during spring, expressed as a percent of total organisms counted. Hyphens indicate values greater than zero but less than 2%; non-zero val- ues shown are rounded to nearest whole number. Shed ovisacs are from Oithona spp.; "Other" in- cludes infrequent and unidentified organisms. (B) Vertically integrated abundances of organisms are averaged across Shelikof Strait for each year; "All other" refers here to all categories from (A) com- bined except for those specifically listed. A Percent composition ( 'ategory 1985 1986 1987 1988 1989 Copepod nauplii Other nauplii Invertebrate eggs Ovisacs Copepods Euphausiids Rotifers Tinitinnids Larvaceans Polychaetes Echinoderms Foraminifera Other 50 46 54 82 76 25 3 9 35 2 0 7 L3 14 11 0 B Average integrated abundance (1000s m-2) from 0-60 m Copepod nauplii Invertebrate eggs All other Total 5.8 13.9 17.6 9.4 9.6 3.0 10.4 4.6 5.7 13.3 30.0 3.6 0.4 0.6 8.6 1.9 2.6 29.8 11.8 12.8 set at 25 units in both the X and Y directions. The same technique was used for contouring CTD and nutrient data. A subset of contours from all three data types was compared by inspection to the origi- nal input data to look for artifacts caused by the contouring software. Integrated abundances of nau- plii across the Strait were compared for the four years which had late April-early May sampling (1985, '86, '88, '89). Data were taken from those sta- tions (#55, 58, 61) sampled every year in the series and were compared by using a non-parametric two- way analysis of variance (ANOVA) on ranks (also referred to as the Quade test: Conover, 1971). A multiple comparison based on ranks (Conover, 1971) was applied when the ANOVA showed statistically significant differences. We used the estimated abundances of adult fe- male copepods (No. m"2) from the oblique bongo tows to consider possible sources of planktonic eggs and nauplii sampled in our study. Data are from a da- tabase being used to describe spatial and interannual patterns of major zooplankton taxa (FOCI Database, National Marine Fisheries Service, Seattle); subsampling and counting followed stan- dard procedures and are detailed in a series of five reports (e.g., Siefert and Incze, 19914). The relative contribution of each taxon to the standing stock of planktonic copepod eggs and early nauplii was esti- mated by using egg production rates reported in the literature or from unpublished data. This is simplis- tic, because it ignores changes in egg and naupliar concentrations as a function of birth rate, develop- ment time, and mortality, all of which may vary considerably. However, the calculations provide a rough evaluation of potential sources of nauplii in Shelikof Strait. Sizes of eggs and early nauplii (e.g., Nauplius I [NI]) were used when reports were found. We used the following information: Calanus marshallae (eggs 175-185 |im, fecundity 12 eggs d : [Runge, 199051; Calanus pacificus (eggs ca. 160 urn, fecundity 38 eggs d"1 [Runge, 19841; NI ca. 220 Urn CL [Fulton 19721); Metridia pacifica (eggs 150 urn [Runge, 199061; fecundity 2.5 eggs d"1 [Batchelder and Miller, 1989)); Pseudocalanus spp. (eggs ca. 110-130 urn retained in ovisacs [Frost, 1987]; fecundity 4 eggs d"1 [Dagg et al., 1984; Paul et al., 1990]; NI ca. 180 pirn CL [Fulton, 1972]). Jeffry Napp7 and Kenric Osgood8 both have found that Metridia pacifica held in the laboratory may produce eggs at higher rates, and they suggest that the population average at times may be several times greater than the rate given above. Results In this section we designate different transects by the year in which they were sampled but do not mean to imply that the differences necessarily were interannual. We address this distinction in the discussion section. Nitrate concentrations in bottom waters were highest in 1985, 1988, and 1989 (>25 ug-at L"1 com- 4 Siefert, D. L. W., and L. S. Incze. 1991. Zooplankton of Shelikof Strait, Alaska, April and May 1989: data from Fisheries Ocean- ography Coordinated Investigations (FOCI) cruises. Alaska Fish. Sci. Center, NOAA, Seattle, WA, 119 p. 5 J. Runge. 1990. Insti. Maurice Lamontagne, Mont-Joli, Que- bec, Canada, pers. commun. 1990. 6 J. Runge. 1993. Inst. Maurice Lamontagne, Mont-Joli, Quebec, Canada, unpubl. data. 7 Jeffry Napp. Nat. Mar. Fish. Serv., Alaska Fishereis Science Center, Seattle, WA, pers. commun. 1993. 8 Kenric Osgood, Dep. Oceanography, Univ. Washington, Seattle, WA, pers. commun. 1993. Incze and Ainaire: Distribution and abundance of copepod nauplii 71 pared to <20 ug-at L : in the other years); in sur- face waters they were lowest in 1987 (mostly <2 ug- at L"1), followed by 1986 (<4 ug-at L M and 1989 (<5 Ug-at L_1) (Fig. 2). Surface nitrate distributions gen- erally reflected density structure. Isopleths of den- sity (Fig. 2), salinity, and temperature show larger '£■ en CD on >- < 2 IN 3 01 (uj) H)dao IO CN 11 •Q O a «I 1 g 05 Ed Tl Jr U IO e X OS a ^ E k.r a c C/J r ft 05 >- 0 bC J3 C CO C -n 9 C -n a cd i- '<: X -5 o - £ « CD 1- 1^ -c 172 r/l H en o c 0) > 0 a s a 01 •- a a en -j 01 r !* cd +^ >i — *J en C -a o> S ^ T-l CO C S 7- ho 1 c i-l '5 w 0 CO - -r =L a is a 0. rn ^H ~- 11 ^-* P O CL 3 c Hi Cii — 1 -C d H c tj -/. od iC x - tn CL, .2 72 Fishery Bulletin 92(1). 1994 volumes of high density (high salinity) bottom wa- ter in 1985, 1986, and 1989 compared with other years. The upper mixed layer generally was deep- est on the northern end of the transect, near the Alaska Peninsula, with a steeply sloping density gradient near the middle. The exception, in 1988, is discussed later. Averaged across the Strait, the up- per mixed layer was deepest in 1985 and shallow- est in 1986 and 1987. Observations of phytoplankton clogging sampling nets during the cruises showed that the spring bloom of large diatoms occurred latest in 1985. By this approximation, what probably was the major spring bloom in the Strait began after the first week of May in 1985, whereas it already was well under- way when we began sampling in early May 1986 and 1989 and late April 1988. A grid of sampling stations that extended to the northern end of the Strait in 1985 showed that the bloom in that year formed first in a band along the middle of the Strait for virtually its full length of 300 km. Our grid in- terval was not sufficiently fine to resolve the width of the bloom feature, but our findings are consistent with a diameter <25 km. Our samples were dominated numerically by cope- pod nauplii, which composed from 46 to 827c of all organisms sampled along the transect over the five- year period (Table 1), followed in most years by cope- pods eggs, from 3.5 to 35f/r. Of the remaining taxo- nomic categories, only a few ever contributed more than 57c of the total organism count: small copep- ods (including copepodid stages), tintinnids, rotifers, Nauplii Late April - Early May 1.8 x 106 o 1.2 x 106 90 urn long and all nauplii >128 urn. Our data showed a steep decline in frequency of nauplii with length <110 um, between the above esti- mates, and width <50 um, corresponding to the relationship 110/2.2 = 50. Most of the nauplii did not have urosomal segments, so total length and maximum width are equiva- lent to prosome length and width for most of our data. The abundance and size distribution of eggs differed substantially between years (Fig. 8). The greatest number (and smallest median size [ca. 75-um diameter]) of planktonic eggs was present in 1986; the fewest eggs occurred in 1988, when median size was the largest (ca. 165 urn). Abundances of potentially significant contributors to the standing stocks of copepod eggs and nauplii are given in Table 2. Among the taxa of interest, Calanus pacificus had low adult female numbers because most individuals were in copepodid stage 5 (C5) during spring. Other adult female copepods Chl-a (mg m2) 20 177 10 142 7 20- (5 n E i io- I A ^ o £ N> |X t — • Distance (km) 20 30 Figure 4 Mean number of nauplii and total microzooplankton per li- ter in the upper 60 m across the study transect in April 1988 (top panel), viewed looking westward. Numbers at the top of the panel show integrated (0-100 m) chlorophyll-a con- centrations (mg m-2). Temperature (°C) and salinity (g kg-1) are shown in the middle and bottom panels, respectively. Data can be compared with nutrient distributions (Fig. 2), dynamic topography (Fig. 5), and depth distributions of nauplii (Fig. 6). were broadly distributed across the Strait, but the maximum concentration of each taxon occurred in the northern half (among stations 58-61) in all but one instance. The across-Strait patterns of low and high abundances within species were similar from year to year and statistically significant (Spearman rank correlation test, P<0.05). The shift in mesh sizes for Pseudocalanus spp. collections limits the between-transect comparisons that can be made. (Note that there are interspecific differences within 74 Fishery Bulletin 92[ I). 1994 57°30 ;-»T llll Alaska Peninsula 155°30' 154D30 Longitude ( W ) Figure 5 Contours of 0-150 m dynamic height in western Shelikof Strait during April 1988. Solid circles show locations of CTD stations. The study transect is the farthest northeast sec- tion. Open circles denote those transect stations with the highest microzooplankton standing stocks (cf. Figs. 4, 6). A dynamic high (H) and low (L) are labelled; arrows show inferred flow. the genus that prohibit any simple correction for different mesh collections: see Frost, 1987.) Within these limitations, data for 1985 and 1986 (333 pm) were statistically different (Wilcoxon signed rank test, P=0.076), whereas the multi-year comparison for early spring samplings (1986, 1988, 1989: 150 pm mesh) showed no statistically significant differ- ences (Quade test, a= 0.05). Among early spring values, there were no statistically significant differ- ences in abundance of Metridia spp.. Discussion The method of sampling and preservation used in this study under-represented smaller components of the microzooplankton (James, 1991) but was ad- equate to capture the majority of prey items of lar- val walleye pollock based on prey sizes reported from earlier studies of Clarke (1984: Bering Sea), Nishiyama and Hirano (1983, 1985: Bering Sea), Dagg et al. (1984: Bering Sea); and Kendall et al. (1987: Shelikof Strait). For small larvae of 5-10 mm standard length (SL) in those studies, copepod nau- plii composed the majority of items found in larval stomachs. They also made up the bulk of estimated volume or carbon content of prey when these values were calculated (Incze et al., 1984; Nishiyama and Hirano, 1983). The 10-m vertical resolution of our sampling method almost certainly failed to detect the highest concentrations of prey available to larval walleye pollock under some conditions, such as in small patches (Owen, 1989), but prob- ably reflects adequately the average abun- dances found at different depths in the wa- ter column, in different sections across the Strait and in different transects. Size-frequency distributions of sampled nauplii and dimensions of the sampling mesh suggest that there was virtually com- plete retention of nauplii with total length > 125 pm. In most cases these measure- ments were carapace ("prosome") lengths. Unpublished data from stomach content studies (Canino, 19929) show that ca. 98% of the nauplii consumed by larval walleye pollock collected during our cruise in May 1989 had carapace length > 125 pm. Be- tween 60 and 70% of the nauplii in our samples were of this size (Fig. 8). Concentrations and integrated abun- dances of nauplii differed across Shelikof Strait in patterns that appear to be related to circulation features. Our data indicated that standing stocks and maximum concen- trations of copepod nauplii in spring were greatest in the ACC, which is also where greatest chloro- phyll-o concentrations occurred (latter data for 1988, 1989; cf. Figs. 4, 6, 7). The lowest naupliar concen- trations of the early spring samplings occurred in 1985, which had the weakest stratification. In gen- eral, nauplii were most abundant at 20-m depth except in 1988, when maximum concentrations oc- curred at 30-m depth in the deeper mixed perimeter of the anticyclonic feature. The lowest standing stock of nauplii coincided with the latest apparent phytoplankton bloom in 1985, but we cannot deter- mine if lower individual copepod egg production rates or lower standing stocks of copepods were re- sponsible because we lack adequate collections ( 150— pm mesh) of Pseudocalanus spp. in 1985. Alterna- tively, the low naupliar standing stocks could have been due to higher predation, but our data show that springtime populations of predators were gen- erally low and were similar among years. Our data suggest that the distribution of copepod nauplii and some other microzooplankton across 9 M. Canino. 1992. Natl. Mar. Fish. Serv., Alaska Fisheries Sci- ence Center, Seattle, WA, unpubl. data. Incze and Ainaire: Distribution and abundance of copepod nauplii 75 naupui (So. r1) 1985 (R=0-18;CI=2) Distance (km) 1986 (R=1-56;CI=4) 3 y \0 JO ' 1987 (R=1-144;CI=10) 1988 (R=0-26; Cl=2) 0 10 » Figure 6 Contour plots of naupliar concentrations (no. Lr1) across Shelikof Strait during spring. Numbers in parentheses after the year (upper left of each plot) show the range (R) of data and the contour inter- val (CI) used in plotting. Transects are viewed look- ing westward. Shelikof Strait were subject to the influence of baroclinic instabilities. The timing and rotational sense of these instabilities therefore may have a large influence not only on the distribution of wall- eye pollock larvae themselves (Reed et al., 1989; Incze et al., 1990; Vastano et al., 1992), but also on the feeding conditions they experience. For example, the feature sampled in 1988 covered a substantial Chlorophyll - a (ug I1) Distance (km) 20 JO T" Figure 7 Chlorophyll-a distributions across Shelikof Strait, May 1989, looking westward (data may be compared with nutrient and hydrographic structure in Fig. 2 and naupliar concentrations in Fig. 6). portion of the main spawning and hatching area. Although we do not have extended observations of this feature, Vastano et al. (1992) showed that eddy- like features may remain near the hatching area for as long as two weeks, a substantial portion of the hatching period (Yoklavitch and Bailey, 1990). If walleye pollock larvae migrate vertically into the center of a dynamic high after hatching, then the amount of time that passes before they are advected into better feeding conditions (in this case at the periphery of the high ) may be important to early larval feeding and growth. The average integrated abundance of copepod nauplii across the Strait was different for the vari- ous transects. The maximum values that were seen in 1987 probably can be attributed to the compara- tively late sampling of that year. However, among the four years with similar timing of transect sam- pling, there remained statistically significant differ- ences that may have been important to hatching walleye pollock larvae (see Canino et al., 1991, for feeding conditions and larval RNA/DNA ratios). Since hatching takes place over a relatively short time period (Yoklavitch and Bailey, 1990), the phas- ing of hatching and upper layer conditions may play an important role in establishing the larval year class. Unfortunately, we do not know how long the observed conditions persisted in each year relative to the population hatching time or to other require- ments of the early feeding period in larval develop- ment. Advection (Incze et al., 1989) and short-term fluctuations in mesoscale circulation (Vastano et al., 1992) may cause conditions in the Strait to change quickly, requiring more frequent sampling and im- proved techniques to rapidly assess prey distributions. Nauplii that were most abundant in the diet of larval pollock must have come from copepods large enough to be retained by mesh sizes used on the 76 Fishery Bulletin 92(1), 1994 1986 nauplii 500— >> U g 400-| 3 ST300 1*1 200— 100— 0 1986 rn-i-! — -^ 50 100 150 200 260 300 Size ((im) C.D.F. 1.0 1985 0.8 y£^\% 0.6 // nauplii 0.4 - /'" 0.2 - 0.0 1 ys i i i i l 14- 13- 12 11 10- 5 4 3 2 1 0 50 100 160 200 260 300 1988 eggs JL 60 100 160 200 260 300 Size (um) Figure 8 Size-frequency distribution of nauplii and eggs. Graph in upper left shows size frequency of nauplii from 1985. Graph in upper right shows the full range of size distributions of nauplii by comparing the cumulative distribution functions (CDF) for the two extremes, 1985 and 1986. Size distributions of eggs are shown in the two lower graphs for years with the smallest (1986) and largest (1988) eggs. Note changes in frequencies shown on the various ordinates. bongo samplers (Table 2). Based on the average abundance and fecundity (see Methods) of adult fe- male copepods, the approximate contribution of each species to the daily production of NI would be: Pseudocalanus spp., >75% ; Metridia pacifica, 18%; Calanus marshallae, 49r; and Calanus pacificus, <1%. These percentages are useful only for the rela- tive scaling they permit; many factors may influence copepod reproduction rates, and rates of develop- ment and mortality will influence further the total standing stock of nauplii contributed by each spe- cies. These results agree with those of Dagg et al. (1984) with respect to the importance of Pseudo- calanus spp. naupliar production for larval walleye pollock feeding. Our results differ in the greater inferred role of Metridia spp., probably because of the deep waters of the Shelikof sea valley compared with the Bering Sea shelf where Dagg and his co- authors worked. The numerous small nauplii <120 (im that we sampled are from unknown sources. The abundance and fecundity of M. pacifica suggest that they were significant contributors to populations of planktonic eggs and that Calanus marshallae plays a lesser role. A large number of small planktonic eggs <150-um diameter are not accounted for by the adult female copepods retained by our nets. Acknowledgments This research was supported by the U.S. National Oceanic and Atmospheric Administration through the FOCI program. We thank J. Schumacher for providing CTD data, M. Canino for sharing unpub- lished data on larval walleye pollock diet, K. Incze and Ainaire: Distribution and abundance of copepod nauplii 77 Table 2 Abundance (no. m~2) of adult female copepods on a transect across western Shelikof Strait during spring. Data are listed vertically showing mean, (standard deviation) and range. Metridia pacifica is Metridia pacifical M. lucens; unidentified Metridia spp. are not included in this tally. Hyphens indicate absence of data. Taxon and mesh size Year and day 1985 (3 May) 1986 (3 May) 1987 (19 May) 1988 (27 Apr) 1989 (10 May) Pseudocalanus spp. 150 |im — 14,183 (6,523) 6,758-18,994 41,058 (25,527) 6,108-78,976 13,634 (4,128) 7,846-20,316 8,450 (4,026) 2,870-12,563 Pseudocalanus spp. 333 pm 9,119 (4,767) 2,509-16,110 16,232 (8,295) 7,848-30,573 33,098 (19,398) 6,273-51,729 Calanus marshallae 333 pm 130 (146) 0-431 82 (72) 0-211 610 (532) 0-1,343 125 (93) 0-238 618* (786) 0-2,196 Metridia pacifica 333 pm 5,082 (4,128) 68-11,899 3,168 (1,956) 24-6,340 9,537 (5,570) 288-5,715 3,211 (1,626) 288-5,715 2,713 (2,549) 0-6,945 Calanus pacificus 333 pm 15 (27) 0-73 2 (4) 0-9 0 28 (61) 0-164 133 (228) 0-521 McCauley for early work with microzooplankton sorting, D. Siefert for processing net zooplankton samples and our many sea-going colleagues for their help in the field. Our work benefitted from discus- sions with A. Kendall, K. Bailey, J. Schumacher, and J. Runge, and our manuscript from comments by M. Mullin, J. Napp, and an anonymous reviewer. Literature cited Bailey, K. M., and E. D. Houde. 1989. Predation on eggs and larvae of marine fishes and the recruitment problem. Adv. Mar. Biol. 25:1-83. Batchelder, H. P., and C. B. Miller. 1989. Life history and population dynamics of Metridia pacifica: results from simulation modelling. Ecol. Modelling 48:113-136. Buckley, L. J., and R. G. Lough. 1987. Recent growth, biochemical composition and prey field of larval haddock (Melanogrammus aegelfinus) and Atlantic cod {Gadus morhua) on Georges Bank. Can. J. Fish. Aquat. Sci. 44:14—25. Buckley, L. J., A. S. Smigielski, T. A. Halavik, and G. C. Laurence. 1990. Effects of water temperature on size and bio- chemical composition of winter flounder Pseudopleuronectes americanus at hatching and feeding initiation. Fish. Bull. 88:419-428. Canino, M. F., K. M. Bailey, and L. S. Incze. 1991. Temporal and geographic differences in feed- ing and nutritional condition of larval walleye pollock, Theragra chalcogramma, in Shelikof Strait, Gulf of Alaska. Mar. Ecol. Progr. Ser. 79:27-35. Clarke, M. 1984. Feeding behavior of larval walleye pollock, Theragra chalcogramma (Pallas), and food avail- ability to larval pollock in the southeastern Bering Sea. Ph.D. thesis, Univ. California, San Diego, 208 p. Conover, W. J. 1971. Practical Non-Parametric Statistics. John Wiley & Sons, New York, 493 p. Cushing, D. H. 1990. Plankton production and year-class strength in fish populations: an update of the match/mis- match hypothesis. Adv. Mar. Biol. 26:249-293. Dagg, M. J., M. E. Clarke, T. Nishiyama, and S. L. Smith. 1984. Production and standing stock of copepod nauplii, food items for larvae of the walleye pol- lock Theragra chalcogramma in the southeastern Be ring Sea. Mar. Ecol. Prog. Ser. 19:7-16. Davis, C. S., G. R. Flierl, P. H. Wiebe, and P. J. S. Franks. 1991. Micropatchiness, turbulence and recruitment in plankton. J. Mar. Res. 49:109-151. Frost, B. W. 1987. A taxonomy of the marine calanoid copepod genus Pseudocalanus. Can. J. Zool. 67:525-551. Fulton, J. D. 1972. Keys and references to the marine Copepoda of British Columbia. Fish. Res. Board Can. Tech. Rpt. 313, 63 p. 78 Fishery Bulletin 92(1), 1994 Gamble, J. C, and L. A. Fuiman. 1987. Evaluation of in situ enclosures during a study of the importance of starvation to the vul- nerability of herring larvae to a piscine predator. J. Exp. Mar. Bol. Ecol. 113:91-103. Houde, E. D. 1987. Fish early life history dynamics and recruit- ment variability. Am. Fish. Soc. Symp. 2:17-29. Incze, L. S., M. E. Clarke, J. J. Goering, T. Nishiyama and A. J. Paul. 1984. Eggs and larvae of walleye pollock and rela- tionships to the planktonic environment. //; Proc. workshop on walleye pollock and its ecosystem in the eastern Bering Sea, p. 109-159. NOAA Tech. Mem. NMFS F/NWC-62. Incze, L. S., A. W. Kendall Jr., J. D. Schumacher, and R. K. Reed. 1989. Interactions of a mesoscale patch of larval fish (Theragra chalcogramma) with the Alaska Coastal Current. Cont. Shelf Res. 9:269-284. Incze, L. S., P. B. Ortner, and J. D. Schumacher. 1990. Microzooplankton, vertical mixing and advec- tion in a larval fish patch. J. Plankton Res. 12:365-379. James, M. R. 1991. Sampling and preservation methods for the quantitative enumeration of microzooplank- ton. N.Z. J. Mar. Freshwater Res. 25:305-310. Kendall Jr, A. W., 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. 85:499-521. Lasker, R. 1981. The role of a stable ocean in larval fish sur- vival and subsequent recruitment. In R. Lasker, led.) Marine fish larvae: morphology, ecology and relation to fisheries, p. 80-87. Washington Sea Grant Program, Seattle. McKenzie, B. R., W. C. Leggett, and R. H. Peters. 1990. Estimating larval fish ingestion rates: Can laboratory derived values be reliably extrapolated to the wild? Mar. Ecol. Progr. Ser.' 67:209-225. Moller, H. 1984. Reduction of a larval herring population by jellyfish predator. Science 224:621-622. Mysak, L. A., R. D. Muench, and J. D. Schumacher. 1981. Baroclinic instability in a downstream vary- ing channel: Shelikof Strait, Alaska. J. Phys. Oceanogr. 11:950-969. Nishiyama, T., and K. Hirano. 1983. Estimation of zooplankton weight in gut of larval walleye pollock (Theragra chalco- gramma). Bull. Plankton Soc. Japan 30:159-170. 1985. Prey size and weight relations in larval wall- eye pollcok iTheragra chalcogramma). Bull. Plankton Soc. Japan 32:45-59. Owen, R. W. 1989. Microscale patchiness in the larval anchovy environment. Rapp. P.-v. Reun. Cons. int. Explor. Mer. 178: 364-368. Paul, A. J., K. O. Coyle, and D. A. Ziemann. 1990. Variations in egg production rates by Pseudocalanus spp. in a subarctic Alaska Bay during the onset of feeding by larval fish. J. Crust. Biol. 10:648-658. Phinney, D. A., and C. S. Yentsch. 1985. A novel phytoplankton chlorophyll technique: Toward automated analysis. J. Plankton Res. 7:633-642. Purcell, J. E., and J. J. Grover. 1990. Predation and food limitation as causes of mor- tality in larval herring at a spawning ground in British Columbia. Mar. Ecol. Progr. Ser. 59:55-61. Reed, R. K., and J. D. Schumacher. 1989. Transport and physical properties in central Shelikof Strait, Alaska. Cont. Shelf Res. 9:261- 268. Reed, R. K., J. D. Schumacher and L. S. Incze. 1987. Circulation in Shelikof Strait, Alaska. Phys. Oceanogr. 17:1546-1554. Reed, R. K., L. S. Incze, and J. D. Schumacher. 1989. Estimation of the effects of flow on dispersion of larval pollock, Theragra chalcogramma, in Shelikof Strait, Alaska. Can. Spec. Publ. Fish. Aquat. Sci. 108:239-246. Rothschild, B. J., and T. R. Osborne. 1988. Small-scale turbulence and plankton contact rates. J. Plankton Res. 10:465-474. Runge, J. A. 1984. Egg production of the marine, planktonic copepod, Calanus pacificus Brodsky: laboratory observations. J. Exp. Mar. Biol. Ecol. 74:53-66. Schumacher, J. D., and A. W. Kendall Jr. 1991. Some interactions between early life stages of walleye pollock and their environment in the western Gulf of Alaska. CalCOFI Rep. 32:22-40. Theilacker, G. H., and Y. Watanabe. 1989. Midgut cell height defines nutritional status of laboratory raised larval northern anchovy, Engraulis mordax. Fish. Bull. 87:457-469. Vastano, A. C, L. S. Incze, and J. D. Schumacher. 1992. Observation and analysis of fishery pro- cesses: larval pollock at Shelikof Strait, Alaska. Fish. Oceanogr. 1:20-31. Yentsch, C. S., and D. W. Menzel. 1963. A method for the determination of phy- toplankton chlorophyll and phaeophytin-n fluore- scence. Deep Sea Res. 10:221-231. Yoklavich, M. M., and K. Bailey. 1990. Hatching period, growth and survival of young walleye pollock (Theragra chalcogramma) as determined from otolith analysis. Mar. Ecol. Progr. Ser. 64: 13-23. Abstract. Distribution and size during their first summer at sea were determined for juvenile salmon (Oncorhynchus spp.) caught in oceanic waters off north- ern British Columbia and South- east Alaska, and in marine waters within the Alexander Archipelago of Southeast Alaska. More than 10,000 juvenile salmon were caught in 252 purse-seine sets during August 1983, July 1984, and August 1984. Distribution was patchy; juvenile salmon were highly aggregated, rather than dispersed randomly. Distribution and size of pink salmon (O. gorbuscha), sockeye salmon (O. nerha), and chum salmon (O. keta) were similar but differed from coho salmon (O. kisutch). Chinook salmon (O. tshawytscha) were ex- cluded from most analyses because few were caught. Sizes were con- sistent with the concept that juve- nile salmon in more northern and seaward locations had been at sea longer than those in more south- ern and inshore locations. Juvenile salmon migration up the Pacific coast did not peak in abundance off Southeast Alaska until August; movement from inside to outside waters was not complete by the end of August. The migration band of juvenile salmon in outside wa- ters of Southeast Alaska extended beyond the continental shelf to at least 74 km offshore, twice the dis- tance previously reported. Marine distribution and size of juvenile Pacific salmon in Southeast Alaska and northern British Columbia Herbert W. Jaenicke Adrian G. Celewycz Auke Bay Laboratory, Alaska Fisheries Science Center National Marine Fisheries Service. NOAA 1 1305 Glacier Highway. Juneau. Alaska 99801-8626 Manuscript accepted 28 September 1993 Fishery Bulletin 92:79-90 (1994) The general migratory movements of Pacific salmon (Oncorhynchus spp.) during their first year at sea have been described (Hartt and Dell, 1986), but little information is available on the seaward migration of juvenile salmon from the inside waters of Southeast Alaska into the Gulf of Alaska. Salmon moving sea- ward from streams inside South- east Alaska pass first through the complex waterways of the Alexander Archipelago, the "inside waters" of Southeast Alaska. Upon entering the Gulf, these salmon either occupy outer coast inlets or move into exposed outside waters. Salmon entering exposed outside waters either migrate north along the coast or move progressively far- ther offshore (Hartt and Dell, 1986). Determining when and at what size juvenile salmon from Southeast Alaska utilize different habitats during their seaward mi- gration to the Gulf may facilitate understanding the high mortality during their first few months at sea (Parker, 1968; Bax, 1983; Furnell and Brett, 1986). Our goal was to ascertain the distribution and migration of juve- nile Pacific salmon during their first summer at sea after they leave nearshore estuarine habitats. Specific objectives were 1) to deter- mine relative distribution, abun- dance, and size of juvenile salmon in exposed outside waters, in protected waters adjacent to the outer coast, and in the inside waters of Southeast Alaska, and 2) to compare abun- dance and size of juvenile salmon in outside waters of Southeast Alaska and northern British Columbia. Methods Study area and time The study area extended from Lituya Bay, Southeast Alaska, to the northern end of Vancouver Is- land, British Columbia (Fig. 1). Three major habitats were sampled: 1) outside waters (the North Pacific Ocean and Gulf of Alaska adjacent to the outer coast of Southeast Alaska and British Columbia); 2) outer coast inlets (protected waters along the outer coast of Southeast Alaska); and 3) inside waters (marine waters within the Alexander Archipelago). Southeast Alaska was further di- vided at lat. 56°N into a northern and southern region for some analyses. Fishing effort was con- centrated in the northern region of Southeast Alaska (Fig. 1). 79 80 Fishery Bulletin 92(1), 1994 Figure 1 Locations seined in Southeast Alaska and British Columbia in 1983 and 1984. The delineation between northern and southern Southeast Alaska is indicated by the dotted line (running along 56°N lat.). We sampled in Southeast Alaska during three periods: 6 August-3 September 1983 (hereafter des- ignated August 1983), 9-24 July 1984, and 1-30 August 1984. Sampling in British Columbia was conducted 1-6 July 1984. Survey stations in outside waters were located along transects perpendicular to shore (Fig. 1). The nearshore station of each transect was as close to land as net depth and safety permitted. Stations were usually sampled progressively offshore at 5.6 km (3 nautical miles [nmi]) intervals in 1983 and at 9.3 km (5 nmi) intervals in 1984. Sampling gen- erally did not extend beyond 37 km offshore except in Southeast Alaska in August 1984, when transects extended as far as 74 km offshore. Distances are rounded to the nearest 1 km in the text. In large passages in the inside waters, sets were often made along transects near the en- trance to outside waters (Fig. 1). Multiple sets were also made in clusters in the larger inlets. Gear Stations were sampled with table and drum seines as described by Browning (1980). The 28-m NOAA RV John N. Cobb fished a table seine in August 1983 and August 1984; the 24-m FV Bering Sea fished a drum seine in July 1984. Sets were made at predetermined locations without reference to visual or instrument sightings of fish. All sets were round hauls: the net was set in a semi-circle, held open 3—5 minutes, closed, pursed, and retrieved by means of a hydraulic power block (table seine) or a hydraulic roller (drum seine). Only catches from effective seine sets are listed (Table 1). Although the seines differed in size, mesh, and area enclosed, the two nets were assumed to be comparable in their ability to capture juvenile salmon. The table seine was 455 m long; depth tapered from 37 m in the wing to 11 m in the bunt; web sizes (stretch mesh) were 89 mm and 57 mm in the wing, and 25 mm in the bunt. The drum seine was 503 m long, 46 m deep, and had 32-mm mesh in the wing, and 25 mm in the bunt. Depths fished were assumed to be adequate for sampling juvenile pink (O. gorbuscha), chum (O. keta), sockeye (O. nerka), and coho (O. kisutch) salmon, which usually occupy the upper 10 m of the wa- ter (Manzer, 1964; Godfrey et al., 1975; Hartt, 1975). To compensate for the larger surface area enclosed by the drum seine (20,150 m2) compared to the table seine (16,467 m2), drum seine catches (July 1984) were reduced during analyses by 18.3% to standardize the catch per unit of effort (CPUE). This standardization caused the July 1984 catches reported to be sometimes less than the number of fish measured for size that period. Catch processing and analysis The catch was processed aboard ship and in the Auke Bay Laboratory. The number of juvenile salmon captured in each set was counted if the catch was small (i.e., <100 fish) or estimated gravimetri- cally if the catch was large. Up to 100 salmon from each set were preserved in 10% formalin in seawater Jaenicke and Celewycz: Marine distribution and size of juvenile Pacific salmon Table 1 Number of uvenile salmonid s caught by species , period, and hg bitat. All seining occurred in Southeast Alaska (SE AK) except in July 1984 when the outside waters of Briti sh Colum bia (B.C.) were alsc sampl ed. Number of fish caught dumber Period Habitat of sets Pink' Churn1' Sockeye1' CohoJ Chinook5 All species August 1983 Inside waters 54 2,011 385 178 201 3 2,778 Outer coast inlet 27 680 85 0 23 1 789 Outside waters 8 20 2 9 27 0 58 Subtotal 89 2,711 472 187 251 4 3,625 July 1984 Inside waters 18 91 16 17 197 19 340 Outer coast inlet 14 10 2 0 24 0 36 Outside waters B.C. 21 573 189 581 33 5 1,381 SE AK 33 181 34 109 28 1 353 Subtotal 86 855 241 707 282 25 2,110 August 1984 Inside waters 37 1,850 163 23 375 23 2,434 Outer coast inlet 4 0 12 0 3 i) 15 Outside waters <37 km seaward 26 866 152 171 128 5 1,322 >37 km seaward l(i 522 63 119 26 0 730 Subtotal 77 3,238 390 313 532 28 4,501 All Inside waters 109 3,952 564 218 77.3 45 5,552 Outer coast inlet 45 690 99 0 50 1 840 Outside waters 98 2,162 440 989 242 11 3,844 Total 252 6,804 1,103 1,207 1,065 57 10,236 ; Oncorhynchus gorbuscha. 2 0. kela. 3 0. nerka. 4 0. kisutch. 5 0. tshawytsc ha for later species identification and size measure- ments (fork length [FL] to nearest mm). If more than 100 juvenile salmon were captured in a set, the excess fish were released alive. Graphs (Chambers et al., 1983) and exploratory data analysis (Tukey, 1977) were used to present catch data because the data had a nonnormal dis- tribution with values clumped at zero (many seine sets did not capture juvenile salmon). Transforma- tions of catch data were ineffective in making the distribution more symmetrical. Quantile plots (Chambers et al., 1983), which show individual catches from smallest to largest, were used to de- scribe the statistical distribution of catches of each species. Chinook salmon (O. tshawytscha) were ex- cluded from the remaining analyses because few were caught. Morisita's Index of Aggregation (Morisita, 1959; Poole, 1974) was used to test whether each salmon species was randomly dis- persed or aggregated in marine waters of Southeast Alaska. Morisita's index is defined as N £«,(«,-!> i l rc(n-l) N, where N is the number of samples, n{ is the num- ber of individuals in the z'th sample, and n is the total number of individuals in all samples. The sig- nificance of I& is tested with the Ftest described by Poole (1974). Spearman's rho (p) correlation test (Daniel, 1978) was used to measure association be- tween each possible pairing of the four main species caught (pink, chum, sockeye, and coho salmon). For comparisons, catch data were split into cells by 1) species, 2) habitat (outside waters, outer coast inlets, and inside waters), 3) region (northern South- east Alaska, southern Southeast Alaska, and Brit- ish Columbia), and 4) time period (August 1983, July 1984, and August 1984). CPUE was used as an index of abundance; frequency of occurrence (FO) 82 Fishery Bulletin 92(1), 1994 was used as a measure of presence of juvenile salmon. Five null hypotheses were tested during fish length analyses of the four species. The first four hypotheses stated that size of a species did not dif- fer for fish from 1) outside and inside waters, 2) outside waters >37 km offshore and <37 km offshore, 3) northern and southern waters, and 4) July and August of 1984. The alternate hypotheses stated that fish were larger in 1) outside than inside wa- ters, 2) outside waters >37 km offshore than outside waters <37 km offshore, 3) northern than southern waters, and 4) August than July of 1984. The fifth hypothesis stated that length did not differ among species caught within each period. A number of one-tailed, two-sample ^-tests were conducted under null hypotheses 1-4. Only cells that varied in one dimension were directly com- pared. (For example, under the hypothesis that mean sizes of fish from northern and southern wa- ters did not differ, the mean lengths of pink salmon in the inside waters of northern and southern South- east Alaska in August 1983 could be compared be- cause the difference between these two cells was in only one dimension — north versus south. ) Each pos- sible pairwise comparison under one of the hypoth- eses was treated as a separate, single, and indepen- dent test, and all comparisons were equally weighted. No ^-tests could be conducted if one cell had only one fish length. For the overall probability statement, the following statistic was used (Winer, 1971): 22>" where it, lnP. Under the hypothesis that the observed probabili- ties were a random sample from a population of probabilities having a mean of 0.50, the %2 statistic has a sampling distribution which is approximated by the x2 distribution having 2k degrees of freedom, where k is the number of comparisons (Winer, 1971). For size hypothesis 5 (no difference in mean fork length among salmon species), ANOVA was applied by pooling observations for each species from all habitats and regions. In effect, the pooled species length distribution is a weighted sum of the compo- nent distributions represented by the individual samples. Mean lengths of different species were compared separately for each period. If the overall F-test was significant, all possible species compari- sons within a period were tested with two-tailed t- tests. Experimentwise error was controlled at a = 0.05 by adjusting the critical value for each t-test to a = 0.0085, by using the Dunn-Sidak method (Sokal and Rohlf, 1981). Results Total catch Over 10,000 juvenile Pacific salmon were captured in 252 seine sets during the three sampling periods (Table 1). The catch consisted of 66% pink salmon, 11% chum salmon, 12%' sockeye salmon, 10% coho salmon, and 1%> chinook salmon. Pink salmon were the most abundant species (CPUE=27), with 6,804 caught. Chinook salmon were the least abundant species (CPUE=0.23), with only 57 caught. Statistical distribution of catch Catch distribution of juvenile salmon was extremely patchy. None were caught in 22% of the sets; more than half were captured in 5% of the sets. Plotting catch abundance against quantiles illustrated that the underlying statistical distribution for each spe- cies was clustered around zero (Fig. 2). Chinook salmon had the lowest FO in catches ( 12%), followed by sockeye salmon (32%), chum salmon (397/ ), pink salmon (45%), and coho salmon (54%). Coho salmon (median catch=l) was the only species with a me- dian catch >0. Juvenile salmon had highly aggregated distribu- tions. Morisita's Index of Aggregation (I&) was sig- nificantly (P<0.001) greater than 1, indicating all species had aggregated distributions in each habi- tat and for all habitats pooled (Table 2). Species associations Pink, chum, and sockeye salmon catches were closely associated with each other. Catches of pink, chum, and sockeye salmon were positively and sig- nificantly (P<0.05) correlated (Table 3). In contrast, coho salmon abundance was not correlated with that of other salmon (Table 3). Abundance By habitat In Southeast Alaska and British Co- lumbia combined, pink salmon were the most abun- dant species in each habitat (Table 1). The total pink salmon catch exceeded the catch of each of the other species by six times or more. In Southeast Alaska, the CPUE of juvenile pink, chum, coho, and chinook salmon was greater in in- side waters than in outside waters (Fig. 3), whereas sockeye salmon were more abundant in outside waters than inside waters (Fig. 3). For each species, the lowest CPUE and FO were in the outer coast inlets; sockeye salmon were never captured in an outer coast inlet (Fig. 3). The FO of pink, chum, and sockeye salmon was higher in outside than inside waters; the opposite was true for coho salmon (Fig. 3). Jaenicke and Celewycz: Marine distribution and size of juvenile Pacific salmon 83 1500 - ,.i 1000 - Pink talmon 500 - ..---"'" , _J 100 - Chum talmon ...-■"""' ~ 50- £ B O 400 - | . — r^ fe Sockeya talmon a E 200- 3 c ...-•'""* £ O 0 J n ioo - 1 _, J .A 0 Coho talmon 50 - ,- '■ , _J 10 - 5 - Chinook salmon , • . J ° 6 0.2 0.4 I 0.6 08 1 median Fraction of ordered data Figure 2 Quantile plots of abundance of the five species of juvenile Pacific salmon (pink, Oncorhynchus gorbuscha; chum, O. keta; sockeye, 0. nerka; coho, 0. kisutch; chinook, 0. tshawytscha) caught in 252 purse-seine sets in Southeast Alaska in 1983 and 1984 and in British Columbia in 1984. The ranked catches are from the smallest (0) to largest (1) on the X axis. A theoretical normal distribution is in- dicated by the dotted lines. By distance offshore in outside waters Dis- tribution of juvenile salmon varied by distance off- shore. Substantial numbers offish were captured up to the maximum distance fished offshore (74 km, Fig. 4A). At intervals offshore, abundance and pres- ence of each species is shown by the 3RSSH smoothed (Tukey, 1977) natural logarithms (In) of CPUE (Fig. 4B) and smoothed FO (Fig. 4C) respec- tively. Highest In CPUE of pink and chum salmon was near the center of the distance fished offshore (Fig. 4B). The transformed CPUE of sockeye salmon, the least abundant species nearshore (Fig. 4B), was greatest 37-74 km offshore, indicating they may have been abundant beyond 74 km. The In CPUE of coho salmon suggests it was the least abundant species beyond 56 km (Fig. 4B). Table 2 Morisita's Index of Aggregation (J8) and the asso- ciated F-value for seine catches of juvenile pink, chum, sockeye, and coho salmon taken in indi- vidual habitats (inside waters, outer coast inlets, outside waters I and all these habitats pooled in Southeast Alaska in August 1983, July and Au- gust 1984. Dashes indicate no fish captured. Salmon species Habitat h F Pink' Inside waters 20.0 695.7* Outer coast inlet 10.7 153.0* Outside waters 3.6 54.9* All habitats pooled 18.5 474.5* Chum2 Inside waters 13.6 66.6' Outer coast inlet 9.0 18.7* Outside waters 5.4 15.3* All habitats pooled 12.7 47.6* Sockeye3 Inside waters Outer coast inlets Coho-' 11.8 22.7* Outside waters 15.1 23.2 All habitats pooled 9.5 24.2 Inside waters 4.4 25.1 Outer coast inlets 2.9 3.1 Outside waters 7.8 19.5 All habitats pooled 6.2 24.3 F-value is significant for P < 0.001. 7 Oncorhynchus gorbuscha. 2 0. keta 3 O. nerka. 4 O. kisutch. Table 3 Spearman's rank cc rrelati on coefficient (p) test of pair rankings of juvenile salmon species catches taken during 252 separate sets in Southeast Alaska and British Columbia Comparison of Correlation between species of species pair rankings salmon (pi Pink'/Chum2 +0.75* Pink/Sockeye' +0.68* Pink/Coho4 +0.14 Chum/Sockeye +0.55* Chum/Coho +0.13 Sockeye/Coho +0.11 Significant association at P < 0.05 . with rejection criteria adjusted for mu tiple comparisons. 1 Oncorhynchus gorbusc ha. - 0 keta. ■7 0. nerka. ' O kisutch 84 Fishery Bulletin 92(1). 1994 40 30 Outside waters Outer coast inlets Inside waters Pink Coho Chum Sockeye Salmon species Figure 3 Catch per unit of effort (CPUE) and frequency of occurrence of juvenile salmonids (pink, Oncorhynchus gorbuscha; chum, O. keta; sockeye, O. nerka; coho, O. kisutch; in outside waters (77 sets), outer coast inlets (45 sets), and inside waters ( 109 sets) in South- east Alaska in 1983 and 1984 combined. Pink and chum salmon FO was lowest nearshore, then increased and stabilized mid-distance offshore, around 37 km (Fig. 4C). Pink salmon were caught in all sets beyond 37 km and had the highest FO of all species; sockeye salmon FO remained constant 2-74 km offshore. Coho salmon FO was the highest nearshore (2 km) of all species, then the FO stabi- lized at 37 km and beyond (Fig. 4C). By sampling period Abundance of juvenile salmon in Southeast Alaska increased from July (CPUE=11) to August (CPUE=58) 1984 for all species. Summed over all habitats, pink, chum, sockeye, and coho salmon had higher FO's and abundance in August than in July. In outside waters, CPUE of each spe- cies increased two to seven times from July to Au- gust 1984, with juvenile pink salmon showing the largest increase (Fig. 5). In inside wa- ters, CPUE of pink and chum salmon increased 10 and 5 times respectively from July to August, whereas CPUE's of sockeye and coho salmon remained constant (Fig. 5). For all four species, FO increased in outside waters but de- creased in inside waters from July to August 1984 (Fig. 5). The low number of sets (four) made in outer coast inlets of Southeast Alaska in August 1984 precluded seasonal comparisons of CPUE or FO for this habitat. Size Juvenile salmon were larger in outside waters than in inside waters. Thirteen matched pairs of size samples could be compared under the hypothesis that size did not vary between outside and inside waters; the fish were larger in the outside water in all comparisons (Table 4, x2=133.66, df=26, P<0.005) and the null hypothesis was rejected. Juvenile salmon in outside waters were larger farther seaward. Of the eight possible matched pairs of samples compared under the hypothesis that size was not different between outside waters >37 km offshore and <37 km offshore, the juvenile salmon were larger >37 km seaward in all compari- sons (Table 4, x2 = 67.44, df=16, P<0.005). Juvenile salmon in northern waters were larger than those in southern waters. The fish were larger in the northward locations than southward locations in 18 of 23 possible paired size comparisons (Table 4, X2=214.76, df=46, P<0.005). Juvenile salmon were larger in August than in July. Of the matched size samples compared under the hypothesis that size was not different between August and July of 1984, fish in August were larger than in July in 10 of 12 comparisons (Table 4, X2=145.36, df=24, P<0.005). The sizes between the different species of juvenile Pacific salmon differed significantly (P<0.05) (Table 5). Coho salmon juveniles were significantly larger than other species in each sampling period; mean length of coho salmon was always at least 40% greater than in other species, whereas pink, chum, and sockeye salmon were within 9% of each other. Juvenile sockeye salmon were significantly larger Jaenicke and Celewycz: Marine distribution and size of juvenile Pacific salmon than pink salmon in each sampling period and were significantly larger than chum salmon in 1984. In both July and August 1984, pink and chum salmon did not differ in size, and in August 1983 chum and sockeye salmon did not differ in size. Discussion Fish distribution Each species of juvenile salmon was highly aggregated rather than dis- persed randomly. In contrast to our results, Hartt and Dell (1986) seldom observed zero catches and therefore concluded that juvenile salmon in the ocean were evenly dispersed. Several differences between our study and theirs may explain the differing conclu- sions. Seines used by Hartt and Dell were longer than ours and were held open for 30 minutes instead of 3-5 minutes. Our catches may be more of a point estimate or instantaneous pic- ture of fish abundance, whereas their seines were more likely to intercept at least part of a juvenile salmon school. More importantly, Hartt and Dell did not separate juvenile salmon by species when considering their distribution. Species associations Juvenile pink, chum (4) (3) (5) (4) (2) L I ■ 1 1 1 , I ■ i ■ 65 74 and sockeye salmon were generally closely associ- ated with each other in their distribu- tion. The distribution of these species, however, differed from the distribution of coho salmon, a result consistent with the conclu- sions of Hartt and Dell (1986) and Waddell et al. (1989). In the inside waters and outer coast inlets, we found that pink, chum, and sockeye salmon had a lower FO than coho salmon, indicating that those species were more highly aggregated and sparsely distributed than coho salmon. Paszkowski and Olla ( 1985) found that behavior patterns of juvenile coho salmon promoted dispersion, not aggregation. The utilization of similar areas in this study by juvenile pink, chum, and sockeye salmon correlates with the high degree of diet overlap observed between these species; in contrast, juvenile coho salmon showed 28 37 46 Distance offshore (km) Figure 4 Abundance of juvenile salmonids (pink, Oncorhynchus gorbuscha; chum, O. keta; sockeye, O. nerka; coho, O. kisutch ) by distance off- shore in the outside waters of Southeast Alaska in August 1984 (36 sets). Abundance is shown in terms of (A) catch per unit of effort (CPUE), (B) the smoothed natural logarithm of CPUE. and (C) the smoothed frequency of occurrence of the catches; number of sets is in parentheses. All distances are rounded to the nearest kilometer. Actual distance between intervals (except the first) is 9.3 km. little diet overlap with the other species.1 Healey ( 1991) reported that juvenile pink, chum, and sockeye salmon in British Columbia were also aggregated. Migration The migration of juvenile salmon off Southeast Alaska (Hartt and Dell, 1986) consists of two components: 1) fish migrating north from the Pacific Northwest and British Columbia, and 2) fish from Southeast Alaska migrating from inside to outside waters. J. H. Landingham, Auke Bay Laboratory. 11305 Glacier High- way, Juneau, AK 99801-8626, pens, commun. Jan. 1992. 86 Fishery Bulletin 92(1), 1994 1 Outside I waters X ?■ Outer coast inlets Inside waters July 1984 August 1984 UJ Z> 0. o o C ID i_ i- 3 O O O >- o c ID 3 O" 0> 100 Pink Chum Sockeye Coho Pink Chum Sockeye Coho Salmon species Figure 5 Catch per unit of effort (CPUE) and frequency of occurrence of juvenile salmonids (pink, Oncorhynchus gorbuscha; chum, O. keta; sockeye, O. nerka; coho, O. kisutch) in the outside waters (69 sets), outer coast inlets (18 sets), and inside waters (55 sets) in South- east Alaska in July and August 1984. Note change in scale of CPUE from July to August 1984. Juvenile salmon migrations along the Pacific coast in 1984 did not peak off Southeast Alaska until, at earliest, August. In July, CPUE's were much higher in the outside waters of British Columbia than in Southeast Alaska. By August, CPUE of juvenile salmon in outside waters of Southeast Alaska had increased fivefold, and FO had increased for each species. Hartt and Dell (1986) observed that juve- nile salmon abundance peaked in August in outside waters of Southeast Alaska. In Southeast Alaska, juvenile sockeye salmon probably begin their ocean migration to the Gulf of Alaska before juvenile pink and chum salmon, based on two observations from our study. First, the sock- eye salmon did not occur in protected waters along the outer coast of Southeast Alaska like the other species: no sockeye salmon were captured in an outer coast inlet. Second, sockeye salmon was the only species with a higher CPUE in outside waters than in inside waters. This higher abundance outside, coupled with low abundance in inside waters in July and August, is consistent with the conclusion that sockeye salmon commence their ocean migration before pink or chum salmon (Straty, 1981; Healey, 1982). The migration of pink salmon from the inside waters of Southeast Alaska lasts until at least September. Martin (1966) concluded that late July and early August were the peak periods of juvenile pink salmon migration from the inside waters. However, our data show that pink salmon abundance in inside waters increased from July to August and that pink salmon were more abundant in inside waters than outside waters in August, thus indicat- ing that migration out of the inside waters was not complete in August. The seasonal migration of juvenile chum salmon out of Southeast Alaska could not be determined from the abun- dance data of this study. The migration of juvenile pink, chum, and sockeye salmon out of the inside waters in Sep- tember and later has not been studied. The offshore migration of coho salmon in Southeast Alaska is more complex. CPUE and FO of coho salmon in inside waters remained relatively constant for July and August. Coho salmon was the only species with both a higher CPUE and FO in inside wa- ters than in outside waters in August. These data suggest extensive residency in inside waters for a substantial portion of coho salmon juveniles in Southeast Alaska. Other researchers have found that some juvenile coho salmon remain in the east- ern Pacific Ocean inside waters until late fall (Healey, 1984; Hartt and Dell, 1986; Orsi et al., 1987). Winter residency of juvenile coho in inside waters of Southeast Alaska is apparently rare.2 Hartt and Dell ( 1986) and Pearcy and Fisher ( 1990) also found coho salmon offshore as early as May or June; Hartt and Dell ( 1986) noted that juvenile coho salmon migrated seaward earlier than the other salmon species, presumably because of their larger 2 J. A. Orsi, Auke Bay Laboratory, L1305 Glacier Highway, Ju- neau, AK 99801-8628. pers. commun. Jan. 1992. Jaenicke and Celewycz: Marine distribution and size of juvenile Pacific salmon 87 Table 4 Fork length (FL) of juvenile salmonids sampled by period, habitat, north (N) or south (S) region, and dis- tances offshore in outside waters of Southeast Alaska in 1983 and 1984 and outside waters of British Colum- bia (B.C.) in 1984. Values are mean ± standard error, with number of samples in parenthesis. In brackets under the values are the specific paired size comparisons used in the null hypothesis testing of sizes by: northern vs. southern waters (Al, A2, ..., A23); outside vs. inside waters (Bl, B2, ..., B13); August vs. July 1984 (CI, C2, ..., C12); and outside waters >37 km offshore vs. outside waters <37 km offshore (Dl, D2, ..., D8). Dashes indicate no fish caught. Period Habitat (region) FL of salmon (mm) Pink' Chum2 Sockeye-3 Coho" Aug 83 Inside (N) 169 ± 0.8 (890) [All 180 ± 1.8 (199) [A2] 163 ± 2.7 (74) 233 ± 1.8 (136) [A3] Inside (S) 121 ± 1.9 (10) [Al, Bl] 139 ± 4.6 (18) [A2, B2] — 227 ± 11.9 (5) [A3, B3] Outer coast inlet (N) — 166 ± 4.9 (4) [A4] — 221 ± 6.3 (11) [A5] Outer coast inlet (S) 124 ± 0.5 (404) 133 ± 1.5 (76) [A4] — 217 ± 7.9 (11) [A5] Outside (S) 153 ± 3.6 (19) [Bl] 141 ± 13.5 (2) [B2] 152 ± 2.6 (9) 234 ± 3.6 (25) [B3] July 84 Inside (N) 121 ± 1.7 (94) [A6, B4, CI] 112 ± 5.2 (19) [B5, C2] 136 ± 5.9 (20) [B6, C3] 193 ± 2.0 (206) [A7, B7, C4] Inside (S) 132 ± 1.2 (3) [A6, B8] 135 ± 0 (1) — 202 ± 7.8 (3) [A7, B9] Outer coast inlet (N) 105 ± 10.9 (4) 139 + 0 (1) — 177 ± 3.6 (27) Outside (N) 135 ± 0.8 (207) [A8, A9, B4, C5] 133 ± 2.3 (38) [A10, All, B5, C6] 151 ± 2.1 (111) [A12, A13, B6, C7] 220 ± 4.5 (26) [A14, A15, B7, C8] Outside (S) 134 ± 4.6 (10) [A8, A16, B8, C9] 161 ± 18.5 (2) [A10, A17, C10] 157 ± 2.6 (19) [A12, A18, Cll] 224 ± 7.5 (8) [A14, A19, B9, C12] Outside (B.C.) 128 ± 1.0 (126) [A9, A16] 132 ± 1.5 (46) [All, A17] 128 ± 0.9 (197) [A13, A18] 129 ± 10.3 (7) [A15, A19] Aug 84 Inside (N) 143 ± 1.0 (358) [B10, CI] 125 ± 1.2 (118) [Bll, C2] 157 ± 2.1 (18) [B12, C3] 234 ± 1.9 (168) [B13, C4] Outer coast inlet (S) — 132 ± 6.1 (12) — 246 ± 12.2 (3) Outside (N) 144 ± 0.6 (730) [A20, B10, C5] 160 ± 2.0 (93) [A21, Bll, C6] 159 ± 1.5 (75) [A22, B12, C7] 267 ± 5.6 (33) [A23, B13, C8] <37 km 143 ± 0.8 (457) [Dl] 157 ± 2.2 (73) [D2] 156 ± 1.7 (52) [D3] 266 ± 6.5 (28) [D4] >37 km 146 ± 1.0 (273) [Dl] 169 ± 4.1 (20) [D2] 165 ± 2.8 (23) [D3] 274 ± 2.3 (5) [D4] Outside (S) 139 ± 1.0 (373) [A20, C9] 144 ± 2.1 (66) [A21, C10] 149 + 0.9 (141) [A22, Cll] 265 ± 3.3 (37) [A23, C12] <37 km 135 + 1.2 (243) [D5] 144 ± 2.7 (38) [D6] 148 ± 1.0 (103) [D7] 263 ± 3.3 (35) [D8] >37 km 144 ± 1.4 (130) [D5] 145 ± 3.5 (28) [D6] 152 ± 1.5 (38) [D7] 291 ± 15.0 (2) [D8] ; Oncorhynchus gorbuscha. 2 0. keta. 3 O. nerka, 4 O. kisulch. 88 Fishery Bulletin 92(1). 1994 size. An early component of coho salmon juveniles could have moved offshore in June, prior to our sam- pling effort. More extensive sampling from late spring through fall is required to define the timing of migrations of coho salmon in the waters of South- east Alaska. The sizes of juvenile salmon we captured support the findings of Hartt and Dell (1986) that fish in more northern locations have been at sea longer than those in southern locations. Hartt and Dell (1986) observed a general increase in mean length of juvenile salmon from south to north in the out- side waters from Washington to Southeast Alaska. In the coastal waters off Oregon and Washington, larger, presumably older, juvenile coho salmon were found farther north (Pearcy and Fisher, 1988). As- suming they were similar in size on entering the sea, the smaller fish in the southerly locations are recent arrivals from nearby production areas, whereas the larger fish in the northerly locations have been at sea longer and probably migrated from more south- erly production areas (Hartt and Dell, 1986). Our studies also reveal juvenile salmon in Southeast Alaska were larger in the outside waters than in- side waters and farther offshore in the outside wa- ters than closer to shore. The progression of juve- nile salmon migrations over a season may be size- dependent (Healey, 1982, 1984), and certain phases of migration may depend on fish reaching a thresh- old size. According to Hartt and Dell ( 1986), the off- shore migration into the Gulf of Alaska of juvenile Table 5 Comparison of mean fork lengths (FL) of juvenile salmonids caught in the marine waters (all habitats pooled) of Southeast Alaska and north- ern British Columbia in 1983-84. Sample size = n; standard deviation of the size in mm = s. The hypothesis was that there were no size dif- ferences between species during the same period. The rejection crite- ria were adjusted for multiple comparisons so that experimental error did not exceed a = 0.05. Species having the same letter in a column were not significantly different by size. August 1983 July 1984 Salmon species mean FL Immi n mean FL (mm) n Pink' 1,323 Chum2 299 Sockeye'* 83 Coho'' 188 155c 1656 162'' 232° 29 31 23 22 444 108 347 277 130' 129r 138'' 193 1 1 17 20 30 1,461 289 234 241 0 Oncorhynchus gorbuscha. b O. keta. c O. nerka. d O. kisutch. pink, chum, and sockeye salmon does not begin until September or October when fish are 180-230 mm or greater in mean FL. However, our findings show that these species are found offshore earlier (in August) and at a much smaller size (145-170 mm mean FL). Width of migration band Juvenile Pacific salmon typically migrate in nearshore waters during their first few months at sea (Straty, 1981); however, the width of this migra- tion band varies regionally (Straty and Jaenicke, 1984; Hartt and Dell, 1986). Juvenile salmon con- centrated within 37 km of shore along the broad continental shelf (<183 m deep) off Oregon and Washington (Miller et al., 1983; Pearcy and Fisher, 1990). Hartt and Dell (1986) concluded that the band of juvenile salmon was within 37 km of shore off Southeast Alaska where the continental shelf is narrow, but that the band widened in the northern Gulf of Alaska where the shelf is wider. Our results indicate that the coastal band of mi- grating juveniles can be much wider than 37 km and that the offshore migration beyond 37 km may be- gin as early as August. Catches of juvenile salmon 74 km offshore — the maximum distance we fished offshore — and the catch distributions indicate that some juvenile salmon (pink, chum, and sockeye) may have been abundant even farther seaward. Two- thirds of the juvenile salmon captured in outside wa- ters in August 1984 were beyond the continental shelf. The width of the migration band is probably in- fluenced by the Alaska Coastal Current — a dominant feature in the circulation of Gulf of Alaska coastal waters. This freshwater-driven current be- gins along the British Colum- bia coast and flows north then west within 20 km of shore into the Bering Sea (Royer, 1984). The strength of this current is affected by local precipitation, wind, air temperature, and other meteorological condi- tions. Millions of juvenile salmon migrate through the cur- rent every year en route to more oceanic waters. Cooney (1984) theorized that the cur- rent represents a critical early- feeding habitat in the summer and early fall. In modeling the early-ocean limitations of Pa- cific salmon production, Wal- August 1984 mean FL (mmi 142'' 141r 153'' 253° 18 22 12 20 Jaenicke and Celewycz: Marine distribution and size of juvenile Pacific salmon 89 ters et al. (1978) noted that production predictions were critically sensitive to the width of the coastal band within which salmon migrate during their first summer at sea. We recommend additional sampling be conducted from June through September to bet- ter document 1) the width of the coastal band of ju- venile salmon migrations through the summer and 2) the timing of offshore migrations beyond 37 km from the outer coast. Acknowledgments We thank the biologists and technicians who helped in the field and laboratory. We also thank the crew on the NOAA RV John N. Cobb and FV Bering Sea for their cooperation during seining operations. The FV Bering Sea cruise was part of a cooperative coastwide survey from California to Southeast Alaska with W Pearcy, Oregon State University. We especially acknowledge the review of the manuscript by A. Wertheimer. Literature cited Bax, N. J. 1983. Early marine mortality of marked juvenile chum salmon (Oncorhynchus keta) released into Hood Canal, Puget Sound, Washington, in 1980. Can. J. Fish. Aquat. Sci. 40:426-435. Browning, R. J. 1980. Fisheries of the North Pacific: history, spe- cies, gear and processes. Alaska Northwest Publ., Anchorage, 423 p. Chambers, J. M., W. S. Cleveland, B. Kleiner, and P. A. Tukey. 1983. Graphical methods for data analysis. Duxbury Press, Boston, 395 p. Cooney, R. T. 1984. Some thoughts on the Alaska Coastal Cur- rent as a feeding habitat for juvenile salmon. //; W. G. Pearcy (ed.), The influence of ocean condi- tions in the production of salmonids in the North Pacific, p. 256-268. Oregon State Univ. Sea Grant College Program Rep. ORESU-W-83-001. Daniel, W. W. 1978. Applied nonparametric statistics. Houghton Mifflin, Boston, 503 p. Furnell, D. J., and J. R. Brett. 1986. Model of monthly marine growth and natu- ral mortality for Babine Lake sockeye salmon (Oncorhynchus nerka). Can. J. Fish. Aquat. Sci. 43:999-1004. Godfrey, H., K. A. Henry, and S. Machidori. 1975. Distribution and abundance of coho salmon in offshore waters of the North Pacific Ocean. Int. North Pac. Fish. Comm. Bull. 31, 80 p. Hartt, A. C. 1975. Problems in sampling Pacific salmon at sea. In Symposium on evaluation of methods of estimating the abundance and biological attributes of salmon on the high seas, p. 165-231. Int. North Pac. Fish. Comm. Bull. 32. Hartt, A. C, and M. B. Dell. 1986. Early oceanic migrations and growth of ju- venile Pacific salmon and steelhead trout. Int. North Pac. Fish. Comm. Bull. 46:1-105. Healey, M. C. 1982. The distribution and residency of juvenile Pa- cific salmon in the Strait of Georgia, British Colum- bia, in relation to foraging success. In B. R. Melteff and R. A. Neve (eds.), Proceedings of the North Pa- cific aquaculture symposium, p. 61-69. Alaska Sea Grant Rep. 82-2. Univ. Alaska, Fairbanks. 1984. The ecology of juvenile salmon in Georgia Strait, British Columbia. In W. J. McNeil and D. C. Himsworth (eds.), Salmonid ecosystems of the North Pacific, p. 203-229. Oregon State Univ. Press, Corvallis. 1991. Diets and feeding rates of juvenile pink, chum, and sockeye salmon in Hecate Strait, Brit- ish Columbia. Trans. Am. Fish. Soc. 120:303-318. Manzer, J. I. 1964. Preliminary observations on the vertical dis- tribution of Pacific salmon (Genus Oncorhynchus) in the Gulf of Alaska. J. Fish. Res. Board Can. 21:891-903. Martin, J. W. 1966. Early sea life of pink salmon. In W. L. Sheridan (ed.). Proceedings of the 1966 Northeast Pacific pink salmon workshop, p. 111-125. Alaska Dep. Fish Game, Info. Leafl. 87, Juneau. Miller, D. R., J. G. Williams, and C. W. Sims. 1983. Distribution, abundance and growth of juve- nile salmonids off the coast of Oregon and Wash- ington, summer 1980. Fish. Res. 2:1-17. Morisita, M. 1959. Measuring the dispersion of individuals and analysis of the distributional patterns. Mem. Fac. Sci. Kyushu Univ. Ser. E. (Biol) 2:215-235. Orsi, J. A., A. G. Celewycz, D. G. Mortensen, and K. A. Herndon. 1987. Sampling juvenile chinook salmon (Oncorhynchus tshawytscha) and coho salmon (O. kisutch) by small trolling gear in the northern and central regions of southeastern Alaska, 1985. U.S. Dep. Commer., NOAA Tech. Memo. NMFS F/NWC-115, 47 p. 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. Paszkowski, C.A., and B. L. Olla. 1985. Social interactions of coho salmon (Oncorhynchus kisutch ) smolts in seawater. Can. J. Zool. 63:2401-2407. Pearcy, W. G., and J. P. Fisher. 1988. Migrations of coho salmon, Oncorhynchus kisutch, during their first summer in the ocean. Fish. Bull. 86:173-195. 90 Fishery Bulletin 92(1), 1994 1990. Distribution and abundance of juvenile salmonids off Oregon and Washington, 1981- 1985. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 93, 83 p. Poole, R. W. 1974. An introduction to quantitative ecology. McGraw-Hill, NY, 532 p. Royer, T. C. 1984. Annual and interannual variability of tem- perature and salinity in the Gulf of Alaska with emphasis on the coastal waters. In W. G. Pearcy (ed.), The influence of ocean conditions in the pro- duction of salmonids in the North Pacific, p. 244- 255. Oregon State Univ. Sea Grant College Pro- gram Rep. ORESU-W-83-001. Sokal, R. R., and F. J. Rohlf. 1981. Biometry, the principles and practices of sta- tistics in biological research. W. H. Freeman & Co., NY, 859 p. Straty, R. R. 1981. Trans-shelf movements of Pacific salmon. In D. W. Hood and J. A. Calder (eds.), The eastern Bering Sea shelf: oceanography and resources 1:575-595. U.S. Dep. Commer., NOAA, Off. Mar. Pollut. Assess., Juneau, AK. Straty, R. R., and H. W. Jaenicke. 1984. Estuarine influence of salinity, temperature, and food on the behavior, growth, and dynamics of Bristol Bay sockeye salmon. In W. J. McNeil and D. C. Himsworth (eds.), Salmonid ecosystems of the North Pacific, p. 247-265. Oregon State Univ. Press, Corvallis. Tukey, J. W. 1977. Exploratory data analysis. Addison-Wesley Publishing, Reading, MA, 506 p. Waddell, B. J., M. C. Healey, and J. F. T. Morris. 1989. Data analysis of 1986 and 1987 Hecate Strait juvenile salmon surveys. Can. Tech. Rep. Fish. Aquat. Sci. 1719, 76 p. Walters, C. J., R. Hilborn, R. M. Peterman, and M. J. Staley. 1978. Model for examining early ocean limitation of Pacific salmon production. J. Fish. Res. Board Can. 35:1303-1315. Winer, B. J. 1971. Statistical principles in experimental design. McGraw-Hill, NY, 907 p. Abstract. Evidence support- ing a two stock hypothesis for king mackerel, Scomberomorus cavalla, in the Gulf of Mexico was devel- oped principally from the results of electrophoretic patterns of one polymorphic dipeptidase locus and supporting evidence from mark- recapture, charterboat catch, and spawning studies. There are two identifiable stocks of king mackerel in the Gulf of Mexico: a western stock and an eastern stock. The western stock migrates northward along the Mexico-Texas coast during the spring and early summer from its winter grounds in Mexico (Yucatan Peninsula). This stock has a high frequency of the dipeptidase PEPA-2*a allele. The eastern stock migrates at the same time north- ward along the eastern coast of the Gulf of Mexico from its winter grounds in south Florida (Gulf of Mexico and Atlantic coast). This stock has a high frequency of the dipeptidase PEPA-2*b allele. Both stocks migrate simultaneously into the northern Gulf of Mexico and mix at varying degrees in the northern summering grounds (Texas to northwest Florida). Evidence for distinct stocks of king mackerel, Scomberomorus cavalla, in the Gulf of Mexico Allyn G. Johnson William A. Fable Jr. Churchill B. Grimes Lee Trent Southeast Fisheries Science Center National Marine Fisheries Service. NOAA 3500 Delwood Beach Road Panama City. Florida 32408 Javier Vasconcelos Perez Instituto Nacional de la Pesca Mexico City. Mexico Manuscript accepted 17 August 1993 Fishery Bulletin 92:91-101 (1994) The king mackerel, Scomber- omorus cavalla, is a widely distrib- uted, coastal pelagic species in the western Atlantic Ocean. This scombrid is found from the Gulf of Maine to Rio de Janiero, Brazil, in- cluding the Gulf of Mexico and Caribbean Sea (Rivas, 1951; Collette and Nauen, 1983). It is a valuable resource that supports fisheries throughout most of its range (Manooch et al., 1978). The U.S. and Mexico have been major exploiters of king mackerel resources. U.S. commercial land- ings have been reported since 1888. Landings have ranged from 2,213 metric tons (t) (1972) to 4,746 t (1974). U.S. recreational catches are estimated to be two to ten times larger than the commercial catches (Deuel and Clark, 1968; Deuel, 1973; Manooch, 1979; U.S. Dep. Commer., 1984, 1986, 1987). In Mexican waters, commercial land- ings for king mackerel from 1968 to 1988 have ranged from 784 t ( 1968) to 6,133 t (Collins and Trent, 19821). Because king mackerel are pres- ently managed in the southeastern U.S. (represented by more than eight states and two regional fish- ery management council jurisdic- tions) and support both recre- ational and mixed gear commercial fisheries, the identities of compo- nent stocks are important. Current management of king mackerel fish- eries assumes two migratory stocks with overlapping ranges, one in the U.S. Atlantic Ocean and one in the Gulf of Mexico (Gulf of Mexico and South Atlantic Fishery Manage- ment Councils, 1985). This separa- tion is based on mark-recapture results (Sutherland and Fable, 1980; Williams and Godcharles, 19842; Sutter et al., 1991). The concept of a stock is one of the most fundamental to fishery management. A stock is variously defined, ranging from the strict definition of a single interbreeding population to a unit capable of in- 1 L. A. Collins and L. Trent, Natl. Mar. Fish. Serv., Panama City, FL, pers. commun. 1992. 2 Williams. R. O., and M. F. Godcharles. 1984. Completion report, king mackerel tagging and stock assessment. Project 2- 341-R. Fla. Dep. Natl. Resour. Unpubl. Rep., 45 p. 9 1 92 Fishery Bulletin 92(1). 1994 dependent exploitation or management and contain- ing as much of an interbreeding unit or as few re- productively isolated units as possible (Royce, 1972). An additional term that has been used to define the stock concept used in fishery management is "unit stock" which was referred to by Kutkuhn (1981) as "one consisting of randomly interbreeding members whose genetic integrity persists whether they re- main spatially and temporally isolated as a group, or whether they alternately segregate for breeding and otherwise mix freely with members of other unit stocks of the same species." This term is more func- tional for application to many marine resources which have identifiable components but for which reproductive isolation has not been demonstrated. We consider stock and unit stock to be identical with regard to king mackerel resources at the present time. Using Kutkuhn's (1981) definition, this report presents evidence of two stocks of king mackerel existing in the Gulf of Mexico (the Gulf), an east- ern and a western stock which winter off south Florida and off the Yucatan peninsula (Mexico), re- spectively. In the spring these fish migrate along their respective coasts to summer areas in the northern Gulf. The concept of two Gulf of Mexico stocks was first presented by Baughman ( 1941). He based his hypothesis on observations by fishermen of simultaneous migrations along the eastern and western sides of the Gulf. More recently, May (1983)3 reported electrophoretic differences in king mackerel between the eastern and western Gulf. Using more recent tagging data and electrophoretic information, Grimes et al. (1987) reintroduced the hypothesis. Additional evidence for a two-stock hypothesis is the following: 1 Fish movements along the coast, as indicated by mark-recapture studies (Fable et al., 19904). 2 The simultaneous migration along the eastern and western coasts of the Gulf in spring and early summer as detected by analysis of charterboat CPU data (Trent et al., 1987b). 3 The difference in spawning times of king mack- erel in the northern and southern areas of the Gulf (Grimes et al., 1990). 3 May, B. 1983. Genetic variation in king mackerel (Scomberomorus cavalla). Final Rep. Fla. Dep. Natl. Resour. Contract C-14-34, 20 p. 4 Fable, Jr., W.A., J. Vasconcelos P., K. M. Burns, H. R. Osburn, L. Schultz R., and S. Sanchez G. (1990). King mackerel, Scomberomorus cavalla, movements and migrations in the Gulf of Mexico. Natl. Mar. Fish. Serv., Panama City Lab., Panama City, FL (unpubl. ms.l. We report the results from electrophoretic inves- tigations and summarize current information from tagging, migration, and spawning time studies. We also propose a possible mechanism to explain the observed results with regard to the water circula- tion of the area. Methods and materials Samples of muscle tissue, along with fork length (mm) and sex, were collected during 1985 through 1990 from fish obtained in recreational and commer- cial fisheries from North Carolina to Yucatan (Table 1). The samples were frozen as soon as pos- sible in the field and then shipped frozen to the Na- tional Marine Fisheries Service's Panama City Labo- ratory. Muscle tissue (about 10 grams) was excised from each sample and stored in a freezer (in 1985 at -5° to -10°C and from 1986 through 1990 at -100°C). Tissue extracts were prepared by mixing equal volumes of muscle tissue and distilled water and grinding with glass rods to uniform pastes. Extracts were centrifuged at 3,400 rpm (1,000 x G) for five minutes, then supernatants were drawn onto 4 mm x 8 mm filter paper inserts (Whatman 1). Starch gel electrophoretic separation of the ex- tracts was performed following the methods of Kristjansson (1963). Electrophoretic buffers were those of A) Markert and Faulhaber (1965), and B) N-(3-aminopropyl)-morpholine-citrate (pH 6.1) buffer of Clayton and Tretiak (1972). The gel con- sisted of 35 g of starch (Sigma Chemical Co. lots 123F-0591, 35K-0383, and 94F-0536) plus 250 mL of buffer. Amperage during electrophoresis was kept below 50 MA, and voltage varied between 100 and 400 V, depending on the buffer. Temperature was maintained at 2°C by using a refrigerated cooling system (see Aebersold et al., 1987, for description). After electrophoresis, the gels were sliced into four horizontal sections and stained for dipeptidase (EN 3.4.-.-). In 1985 (1,223 fish) and 1988 (879 fish), 27 additional enzymes were examined. Methods fol- lowed May (1983)3 and Aebersold et al. (1987). We conducted statistical analyses using Biosys-1 (Swofford and Selander, 1981) to test for conform- ance to Hardy-Weinberg expectations and spatially related differences in allele frequencies compared to distance and physical feature subdivisions. The Kolmogorov-Smirnov goodness-of-fit test was used for comparing allele distributions by size of fish (100-mm-FL intervals), while the chi-square contin- gency test was used for comparing allele distribu- tions by sex (see Sokol and Rohlf (1981) for proce- dures i. Johnson et al.: Evidence for distinct stocks of Scomberomorus cavalla 93 -o c C m 0 .Q n -r in CO in tr- co as oo CO CO CO oo co c- m o o o co o o co o o o o o o o o o O O O O CO CM t"- in .—• in cm .— i oo CO CO CO CO CO CO CD CTi ^ — in ^ CO CO CM CO r— C i oo ^- -r II en en OO CO CO on CO m oo on c C7S _■ CO ■ — i O CO — O o CO o •3' CM CM © CO „ f h n °£ -rf CO CO ^ CO O CO L r in cm co co co" co co" CM cc -r ■■T. , — i in r- on CO CM CO CO CO CO r— GO -1 £- t- CO r i CO oo O CO O Tf o en t- o o CT) O O co t- co en co oo oo ao oo en en en en en en — co r- oo en co fl oo oo oo oo en .zi en en en en en co t- oo en ao oo co oo en en en en cu co t— co en co m m ^ o — oo oo oo co en — ao co oo cncncncncncn & en en en 3 ,2 94 Fishery Bulletin 92(1). 1994 ■D dj 3 C C o u .15 c o § HO CM C— ■** CO •— < — " t— .— CM ao -CO 00 t-QN CO CM •— < t— ♦-« i— < ■^ to co m O) t- r-( rH t-H Tf — CO — — r- co cm o co co cm •— 1 CO i— < --H CO — — CM CO C- C- -*!• CO Oi M1 rH rH O) (N t^ Oi - o n ^ o m .—I ,-1 .—I .— I lO W O O O O O rH o o o o o o i— I CO CM CO CM O O CM OHHH dodo CO OlOl t CD CM — I CO CO CO © o i-j o o cd cd © © © © ^ •— • uO CO ■^ CO CO o «-< O O O rH O ©' © cd o © r- <— < CO CTJ .— CO © ■— i — , ,-. CO © CO CO in uo co c- ex ao cd cd cd CO CO CO r- co co as as r-- .— < .—i in o ^ -tf CO o o © o .-H O o o — CO CO «-H OO CM •rp CM o o CO CD CO ■— < CO —» to O tJ- , — ' oo OQ CM ■— • 'X en as CO o o ■— * © t~- uo CO CO o oo r- ■*** r - t>- '-- CO ■— * c co co CO ^ o CT> CTJ o OS 00 / CO CO C- CO CO CM U0 CO CO o *-* ©©.-<© CO CO CM CO O ^h o o dodo o r~ o oo CO co ■— < CM CO .— I CO ~ ao •— t CO CO CO CO CO co as oo co oo as co co CO CO — CO an b cd en c cd c ■_■ — 3 CO c CO £ -I CD E *h CO oo CM uo CO * CO ~ / o >.t -I ■■'.■ CO o _ _ o CO — o c o J3 e iO f-i o CO o CD *j c = c CO - fc i° .2 sH 0) s r -c ♦J 0- U, sr a. y-s s T3 C a CO en CO en OS OS © 2 m CD O OO OJ O ctDt^XO co as £ ao ao ao ao ao as ^ ao ao ao as OiCjl y Ol CJ) C31 O) Oi Ol *jO^CT)CUO^ cj uo co r- ao en o ■^2 oo oo oo ao ao 05 Ol Oi OS j; CD Ol OJ C3 CTi I -O CO *^ o .2 co -a « c CJ < (0 u_ ■■■ — : u X C 0 CD 3 CD c CD rf" S ^ B 'J U a. /. J* > 'St Johnson et al.: Evidence for distinct stocks of Scomberomorus cavalla 95 Results Of the 50 loci surveyed in 1985, 30% were variable. In 1988, the 50 loci were again surveyed (879 fish from 10 locations) and 24% of the loci were found to have variants. Variations other than dipeptidase (EN 3.4.-.-) PEPA-2 were found in low frequency (uncommon allele 0.000 to 0.063) in 18 polymorphic systems. Occurrence of these variants differed be- tween locations and years. Electrophoretic variants were found for loci including aspartate aminotrans- ferase (EN 2.6.1.1) sAAT ", acid phosphatase (EN 3.1.3.2) ACP-2*, adenosine deaminase (EN 3.5.4.4) ADA , adenylate kinase (EN 2.7.4.3) AK-1* and AK- 2' , alanine aminotransferase (EN 2.6.1.2) ALAT-1 and ALAT-2', esterase-D (EN 3.1.-.-) ESTD-2' and ESTD-3 , fructose-bis-phosphate aldolase (EN 4.1.2.13) FBALD-2 \ glucose-6-phosphate isomerase (EN 5.3.1.9) GPI-1* and GPI-2\ isocitrate dehydro- genase (NADP+) (EN 1.1.1.42) sIDHP , malic en- zyme (NADP+) (EN 1.1.1.38) ME-2' , mannose-6- phosphate isomerase (EN 5.3.1.8) MPf, dipeptidase (EN 3.4.-.-) PEPA-1 , phosphogluconate dehydroge- nase (EN 1.1.1.44) PGDH*, and phosphoglucomutase (EN5.4.2. 2) PGM-2*. Use of very low-frequency variations for stock identification of king mackerel was impractical, be- cause sufficient sample sizes (numbers of fish) for detection during short time periods (one month or less) were unavailable. Tagging studies (Fable et al., 19904) indicated that discrete geographic population units were not available during the time intervals required to obtain sufficient samples. Only dipepti- dase (glycyl-leucine substrate)5 consistently varied between locations. In 1985 (1,223 fish), 1986 (1,537 fish), 1987 (2,120 fish), 1988 (1,631 fish), 1989(1,502 fish), and 1990 (963 fish), muscle tissues were ex- amined for the dipeptidase variation. This enzyme developed on electropherograms as two zones of activity, and showed the pattern of a two allele ("a and b) polymorphism in the most anodal zone {PEPA-2\ in most collections, as described by May [1983]). We refer to May's 1 and 2 alleles (electro- morphs) as a and *b, respectively (Fig. 1). A third allele (*c) which is anodal of the a allele was found in 1988 and 1989 collections from Veracruz, Mexico to Alabama.6 Only one homozygote (*c*c) and 20 heterozygotes Cc'a) were found from 3,487 fish. 5 Enzyme is also active with valyl-leucine and leucyl-tyrosine as substrates. 6 The genetic nomenclature for this polymorphic system accord- ing to the recommendations of Shaklee, et al. (1990), is dipep- tidase 3A.-APEPA-2') with three variant alleles '//(). '105, and '100. These alleles are represented in this report as *c, *a, and 'b, respectively. PEPA-2 PEPA-1 ■Tit ••!••.— IV'Ii T (D 'a'a a'b Vb (2) I ■ • ^(3) Figure I King mackerel iScomberomorus cavalla) dipeptidase (PEPA-P and PEPA-2*): (1) schematic of gel with 25 samples (PEPA-2* a is 0.700), (2) schematic of en- largement of section of PEPA-2 on gel showing three phenotypes ( a a, ab, b 6), and (3) photo- graph of actual gel section used for schematic (2). Because of the rareness of this allele Cc), it was combined with allele a for analysis. Allele frequencies and phenotypic distributions varied extensively within and between areas from 1985 to 1990 (Table 1). The majority of monthly collections conformed to the Hardy-Weinberg expec- tation; however, many of the yearly collections did not conform. In general, higher *a allele frequencies were found west of Florida than in Florida and along the Atlantic coast. The phenotypic distributions of the dipeptidase polymorphism were not significantly correlated with body length, with few exceptions. When the pheno- typic distribution was compared by 100-mm-FL size intervals for five geographic locations (Atlantic coast, Alabama-Mississippi, Louisiana, east Texas, and south Texas) by year, only seven of the 78 com- parisons were significantly different (Kolmogorov- Smirnov goodness-of-fit test, P<0.05). Four of these 96 Fishery Bulletin 92(1). 1994 deviant collections occurred in the northern Gulf (east Texas and Alabama-Mississippi). The other three (1988— *a*a phenotype on Atlantic coast; 1989- *b*b, and 1990-*a*a phenotypes in northwest Florida) are believed to have resulted from sampling inadequacies (in 1988, only 9 *a*a were collected on the Atlantic coast, and in 1989 northwest Florida had 136 of the 275 *b*b in the <600-mm-FL cell, which represented 167 of the 344 fish; and in 1990, northwest Florida had 12 *a*a of the 17 *a*a in the 900, 1,000, and >1,100 mm cells). When allele distributions were compared by sex at seven locations for each year in which sufficient data were available, eight of the 23 allele compari- sons deviated significantly (chi-square contingency test, P<0.05). Six deviant collections occurred in the northern Gulf (Texas-Mississippi 1985-1989) and were from collections that did not conform to Hardy- Weinberg expectations with regard to their pheno- typic distributions. Two others occurred in Veracruz, Mexico (1988 and 1990). The total allele-sex (1985- 90) comparisons for the seven locations did not de- viate significantly, except for Veracruz, Mexico. Veracruz collections were dominated by small fish (<600 mm FL) of which sex determination was dif- ficult, especially early in the year (Jan. -July) be- cause of undeveloped gonads. Sex could only be de- termined for 68% of the fish tested from this area. The geographic pattern of dipeptidase (PEPA-2*) (1985-90) indicated that western Gulf differed from eastern Gulf and Atlantic coast king mackerel. In all years except 1985, comparison of allele counts (Table 1) of the various geographic groupings of the Gulf varied significantly (P<0.05) both within the Gulf and between the Gulf and the Atlantic coast. On the Atlantic coast (north of Florida vs. Florida), the variation was found not significant (except in 1990). The trend in these comparisons was for ex- cess *a allele in the western Gulf and for excess *b allele in the eastern Gulf and the Atlantic coast. Discussion Comparisons of subdivisions (Table 2) show a con- sistently higher level of PEPA-2*a in western Gulf king mackerel and a deficit of this allele in king mackerel in the eastern Gulf and along the Atlan- tic coast. Electrophoretic data (ours and that of May ( 1983 )3 indicating high dipeptidase PEPA-2* a frequency in the western Gulf and low *a frequency in the east- ern Gulf and along the Atlantic coast supports a two stock hypothesis for king mackerel in the Gulf. Sup- porting information can be obtained from other in- vestigations: mark-recapture (Fable et al., 19904), charterboat catches (Trent et al., 1987b) and spawn- ing date analysis (Grimes et al., 1990). Fish move- ments indicated by mark-recapture are consistent with the two stock hypothesis. The charterboat in- formation provides evidence of simultaneous north- ward migration on both sides of the Gulf, while the spawning date information offers evidence for repro- ductive isolation. The king mackerel dipeptidase (PEPA-2") varia- tion found in 1985-90 was similar to the variation first reported by May (1983)3. His data showed higher dipeptidase *a allele frequencies for Louisiana (0.618) and Texas (0.736) than were found eastward. Temporal variations in the PEPA-2* allele frequen- cies are difficult to interpret without taking into consideration the migratory behavior. The variation was extreme at some locations, giving the impres- sion that the samples were collected from different or mixed schools from different origins. For example, in east Texas (Galveston-Freeport area) (1986), five discrete collections (5 July-28 August) of 27 to 56 fish each (204 total) were sampled. The PEPA-2* a frequencies were 0.933, 0.769, 0.202, 0.839, and 0.037 (in collection order). In other collection peri- ods, variations in frequencies indicated that we had sampled the same school of fish. For example, in Louisiana (1987) three collections 7 days apart (21 Aug.^1 Sept.) were obtained. Their PEPA-2* a fre- quencies were 0.590 (50 fish), 0.580 (50 fish), and 0.594 (48 fish). In view of the extreme variability of PEPA-2* frequencies, numerous deviations from Hardy- Weinberg expectations, and sampling difficul- ties (one or more schools per collection), proper spa- tial subdivision and grouping of collections for test- ing specific hypotheses is arduous. The expanse of the sampling area (Virginia to Yucatan) can be di- vided into various subdivisions representing dis- tance or physical features (Table 2). Examples of subdivisions by distance are the following: Missis- sippi westward vs. Alabama eastward, Alabama to Florida Keys, Florida vs. Atlantic coast, and Florida east vs. Georgia northward. Examples of physical subdivisions are the following: Florida peninsula (Florida east coast versus Florida west coast), east- ern Gulf and Atlantic coast (Alabama to Florida Keys versus Atlantic coast), and northern and west- ern Gulf (Louisiana-Mississippi versus Texas versus Mexican sector of the Gulf) (See also Collard and Ogren, 1990). Caution should be applied to interpreting electro- phoretic results in which variation has not been proven to be of genetic origin by the use of breed- ing analysis (i.e., crossing of phenotypes and analy- Johnson et al.: Evidence for distinct stocks of Scomberomorus cavalla 97 sis of offspring). Deviation from Hardy- Weinberg expectations can result from stock mixing, natural selection, or drift in small populations (Smith, 1990). While we favor the inter- pretation that these king mack- erel data suggest stock mixing, consideration should be given to natural selection as the ul- timate maintenance factor of PEPA-2* frequencies as sug- gested for dipeptidase (PEPA- LT*) and other variations found in Menidia beryllina (Johnson, 1974). Electrophoretic data suggest that two stocks of king mack- erel occur in the Gulf, a west- ern stock with high frequency of the *a allele and an eastern stock with a low frequency of the *a allele. The northern Gulf appears to be a zone of mixing of these two stocks during the summer. Our electrophoretic information does not distin- guish the eastern Gulf fish from those along the Atlantic coast. Historical tagging data showed migration between south Florida and the north and northwest Gulf. Williams and Godcharles (1984)2 (and Sutter et al.'s later analysis (1991) of Williams and Godcharles' data) can be exam- ined in light of the two stock hypothesis. Williams and Godcharles tagged approxi- mately 12,000 king mackerel off south and southeast Florida, primarily in winter months. Forty-nine tags were recovered in the northeast Gulf and another 49 tags were re- turned from the northwest Gulf. Almost all tagged fish were recaptured in the warmer months of the year, supporting the hypothesis of migration from wintering grounds in southeast Florida waters to northern Gulf of Mexico waters Table 2 Comparisions o f geographic groupings of a llele counts of dipeptidase (PEPA- 2*) in king mac terel. {Scomberomorus cava lla), 1985- -90. Location' Year Alleles X2 df P Remarks MS westward vs. AL eastward (distance)2 1985 1,620 297.3417 <0.001 Deficient *b in MS westward 1986 1,676 340.9499 <0.001 Devidient *6 in MS westward 1986 3,976 283.7311 <0.001 Deficient *b in MS westward 1988 2,468 812.6335 <0.001 Excess *b east of Al Deficient *a east of AL 1990 1,926 793.5280 <0.001 Excess *b east of AL Deficient *a east of AL Key West, FL westward vs. Atlantic coast (physical) 1985 2,630 329.0983 <0.001 Excess *a in Gulf 1986 2,662 879.2843 <0.001 Excess *a in Gulf 1987 3,865 271.3356 <0.001 Excess *a in Gulf 1988 3,084 643.4390 <0.001 Excess *b in Atl. coast Deficient *a in Gulf 1989 3,004 657.913 <0.000 Excess *b in Atl. Coast Deficient *a in Atl. Coast 1990 1,926 339.2062 <0.001 Excess *b in Atl. coast Deficient *a in Atl. coast AL to Key West, FL vs. Atlantic coast (distance) 1985 1.518 0.0040 >0.90 1986 1,258 33.1770 <0.001 Excess *a in Gulf 1987 1,550 64.6325 <0.001 Deficient *a in Atl. coast 1988 1,022 10.4639 <0.001 Excess *a in Atl. coast Deficient *a in Gulf 1989 1,406 6,2033 >0.01 Excess *a in Gulf Deficient *a in Atl. Coast 1990 864 22.0855 <0.001 Excess *a in AL to Key West, FL Deficient *a in Atl. coast Within northerr and western Gulf (LA-MS, TX, MX) (physical) 1985 1,110 7.9835 2 >0.01 1986 1,410 135.5281 3 <0.001 Excess *b in LA-MS Excess *a in MX 1987 2,416 71.5602 2 <0.001 Excess *b in LA-MS Excess *a in MX 1988 2.062 40.1994 2 <0.001 Excess *b in LA-MS Deficient *b in TX 1989 1,598 70.2421 2 <0.001 Excess *b in LA-MS Deficient *a in LA-MS 1990 1,062 120.9159 2 <0.001 Excess *b in LA-MS Deficient in *a in LA-MS Deficient in *b in MS Within Atlantic coast (N of FL vs. FL) (distance) 1985 1,008 0.0738 1 >0.70 1986 992 1.8493 1 >0.10 1987 336 0.1133 1 >0.70 1988 616 0.9336 1 >0.30 1990 388 6.0278 1 >0.01 Excess *a in FL ' Abbreviations are used for states: AL=Alabama; Fl ^Florida, LA=Louisiana; MS=Mississippi. TX=Texas; MX=M >XlCO :' In parentheses { ) general ci assification of range subc lvisions. See text 98 Fishery Bulletin 92(1), 1994 in the summer. These authors also tagged fish off North and South Carolina, but none were recovered in the Gulf. According to Fable et al. (1990),4 king mackerel tagged in northwest Florida have been recovered in south Florida. Typically, these are the smallest and youngest tagged in the southeast United States. Sutherland and Fable ( 1980) showed that northeast Gulf fish migrated to south Florida. However, addi- tional tagging (Fable et al., 19904) showed that northeast Gulf fish eventually moved westward to Louisiana, Texas, and Mexico waters when they had been free for a sufficient time and grown to a larger size. Tagging off Louisiana from 1983 to 1985 (Fable et al., 1987) indicated that the northwest Gulf may have year round residental large king mackerel that mix in the warm months with smaller migrants from south Florida and Mexico. Recent tagging data (Fable et al., 19904) from this region have provided additional recoveries from both south Florida and Mexico, strengthening this interpretation. Addi- tional support is provided by the occurrence in Loui- siana of a year-round king mackerel fishery, whereas elsewhere the fishery is seasonal. In contrast to historical reports, recent tagging (Fable et al., 19904) showed movements between Texas and Mexico. Fish tagged in Texas waters mi- grate to both Florida and Mexico. Additionally, fish movements between Texas and eastward (as far as Panama City, FL) were documented. Mark-recapture data (Fable et al., 19904) from tagging in Mexican waters suggest that the states of Campeche and Yucatan are wintering areas for king mackerel in the western Gulf. Fish tagged in warmer months (April-July) in Texas, Tamaulipas, and Veracruz were found in Campeche and Yucatan in the winter. Tagging efforts (Fable et al., 19904) in Veracruz have provided evidence of northward mi- grations to Tamaulipas and Texas in spring and sum- mer, and movement to the Yucatan peninsula in winter. Additional evidence supporting two Gulf stocks can be found in catch-effort data of king mackerel. Although the data are complicated by different fish- ing strategies depending on the type of fishery (rec- reational or commercial) and regulatory closures, detailed analysis of catch data from the southeast- ern United States charterboat fishery indicated that in spring and early summer some stocks of fish si- multaneously migrated northward along the west- ern and eastern coasts of the Gulf (Trent et al., 1987b). They also developed the ". . . idea that part of the population of large fish remains in the Loui- siana area year-round and that the abundance of these fish is greatest during cold months." The fishery for king mackerel in Louisiana is unique among the fisheries in the northern Gulf of Mexico in that it is year-round; elsewhere it takes place mainly from late spring to late fall. The win- ter fishery (commercial hook-and-line) in Louisiana began in 1981-82. Distinctive differences character- ized winter and spring-fall seasons: 1) the smallest fish (both males and females) were caught April to October whereas the largest fish were caught be- tween November and March; 2) females were more abundant in the winter fishery than at other times of the year (Trent et al, 1987a). For two or more populations to maintain separate identities they must be isolated, either physically or reproductively (Hartl, 1980). In the case of Gulf king mackerel, there is evidence for reproductive isola- tion. Grimes et al. (1990) presented a detailed ex- amination of the distribution and occurrence of lar- val and juvenile king mackerel in the Gulf (based on published reports, neuston sampling, and Mexi- can trap net and trawl collections). The spawning season in the northern Gulf (U.S. waters), as indicated by the seasonal occurrence of larvae, is May to Octo- ber. Larval collections off Mexico were sparse and of- fered little information on spawning seasonality. The summer spawning period in the northern Gulf was also indicated by seasonal gonadal devel- opment of king mackerel (Finucane et al., 1986). They reported that reproductive activity occurred from May through September; a few fish were re- productively active as early as April and as late as October. However, spawning dates of January through August for Mexican juveniles estimated from otolith data showed a bimodal distribution, which suggests that spawning seasons in Mexican waters are different from those in the northern Gulf (Grimes et al., 1990). Two of the four collections of juvenile king mack- erel in Mexico used by Grimes et al. ( 1990) had tis- sue samples (Tampico, July 1986, and Playa Norte, Sept. 1986), and we analyzed these samples for PEPA-2* variation. Spawning dates of fish in the Tampico collection ranged from mid-February to mid-April and PEPA-2' a frequency was 0.896. The Playa Norte collection's spawning dates ranged from mid-April to mid-July and PEPA-2* a frequency was 0.600 (Table 1). Water circulation data for the Gulf of Mexico (Salsman and Tolbert, 19637) and information from Trent et al. (1987b), Grimes et al. (1990), Fable et al. 1990,4 along with our data on king mackerel, sug- 7 Salsman, G. G., and W. H. Tolbert. 1963. Surface currents in the northeastern Gulf of Mexico. U.S. Navy Mine Defense Laboratory, Panama City, FL, Res. and Dev. Rep. 209, 43 p. Johnson et al.: Evidence for distinct stocks of Scomberomorus cavalla 99 gest one plausible scenario with regard to king mackerel stocks in the Gulf of Mexico. A western population exists that winters and spawns in the Gulf of Campeche. The Mexican Current serves as an entrainment system for its young. As these young become older and larger, they are able to cross the region of offshore advection and utilize the north- ern Gulf area (Texas to Florida) for summer feed- ing. This stock of fish has a high PEPA-2*a fre- quency and spawns earlier in the year than fish in the northern and eastern Gulf of Mexico. No infor- mation (tagging, electrophoretic, or reproductive) is available on fish of the Yucatan Straits area and the Caribbean Sea to evaluate their relation to the west- ern Gulf of Mexico fish. An eastern population of king mackerel uses the eastern and northern Gulf of Mexico area as entrainment systems for its young and the northern Gulf (Florida-Texas) as summer feeding grounds. The spawning area extends from Texas to northwest Florida between April and Oc- tober; the majority of spawning probably occurs in the northwest Florida-Louisiana area. Tagging stud- ies suggest that this stock uses south Florida and the southeast coast of Florida as its wintering grounds. The Louisiana area is somewhat of an enigma. Tagging studies indicate that the area is used by fish from both sides of the Gulf, fish are in the area year- round, PEPA-2'a frequencies are between the ex- tremes of the east and west Gulf, and tag recover- ies from winter tagging in Louisiana have been from Louisiana and westward, whereas recoveries from summer tagging were both east and west of Louisi- ana. Additionally, Finucane et al. (1986) suggested an earlier distinct peak in gonadal development (May) for Louisiana-Mississippi than in northwest Florida (August) and in Texas (August). The ques- tion still remains: Does the Louisiana area have an independent spawning population that utilizes the northern Gulf currents for its life cycle? The exist- ing evidence (especially tagging) suggests the area is not independent; however, information comes from larger fish. Thus, the area may be occupied by individuals from both sides of the Gulf which may or may not reproduce in the area. Further investi- gation especially on the younger life stages using other methods of analyses may answer this question. Another group (stock) of king mackerel that im- pinges upon the Gulf of Mexico resources (officially recognized by Fishery Management Councils) is the Atlantic Migratory Group. This group has a vary- ing range from Virginia to southwest Florida de- pending on the time of the year (Gulf of Mexico and South Atlantic Fishery Management Councils, 1985). The stock is considered to winter in South Florida and ranges along the Atlantic coast to North Carolina and South Carolina during the summer. The fish probably spawn from May to October with a peak in July (Finucane et al., 1986). These fish are currently regulated as a group with seasonal south- ern boundaries of lat. 25°48'N (the Collier/Monroe County line, FL) from 1 April to 31 October and lat. 29° 25'N (the Volusia/Flagler County line, FL) from 1 November to 31 March. Tagging information sup- ports this separation (Gulf of Mexico and South Atlantic Fishery Management Councils, 1985). PEPA-2' a allele frequencies are generally low (0.00-0.10) along the Atlantic coast as in the east- ern Gulf of Mexico. The higher PEPA-2*a values (>0.10) occasionally encountered may be the result of fish entrapped in water masses coming up the coast from outside the east coast of Florida. This possibility is suggested by the recovery along this coast of drift bottles that were released in the Yucatan Straits area (Salsman and Tolbert, 19637). All these stocks need to be further investigated in order to be elevated to the status of genetic stocks (i.e., completely isolated reproductive populations of the same species). Conclusion Four lines of evidence for a two stock hypothesis for the Gulf of Mexico king mackerel have been pre- sented. The two stock hypothesis states that the Gulf contains a western stock of king mackerel, which winters in Mexico and migrates in spring and early summer to the northern Gulf (Texas-Alabama), and an eastern Gulf stock which winters in south Florida and migrates in spring and early summer to the northern Gulf. The two stocks mix in the northern Gulf during the summer. The four lines of evidence are the following: 1 Dipeptidase (PEPA-2' ) data showing western Gulf fish high in *a allele and eastern fish low in *a allele. 2 Mark-recapture data showing movement along both sides of the Gulf from south to north. 3 Catch data indicating simultaneous migrations northward on each side of the Gulf in early spring and summer. 4 Estimates of spawning dates suggesting pos- sible temporal and spatial differences between the northern and southern Gulf. Acknowledgments Especially helpful in collecting specimens and data were staff members of the following organizations: 100 Fishery Bulletin 92(1). 1994 Florida Department of Natural Resources (Tallahas- see, FL); Gulf Coast Research Laboratory (Ocean Springs, MS); Institute Nacional de la Pesca (Mexico City, Mexico); Louisiana State University (Baton Rouge, LA); Mote Marine Laboratory (Sarasota, FL); North Carolina Division of Marine Fisheries (Morehead City, NO; Savannah State College (Sa- vannah, GA); Texas Parks and Wildlife (Austin, TX); Virginia Institute of Marine Sciences (Gloucester Point, VA); and the various laboratories of the Na- tional Marine Fisheries Service, Southeast Fisher- ies Center (Miami, FL). Special thanks go to B. May, Cornell University (Ithaca, NY) for sharing his ex- perience with king mackerel with us, to K. M. Burns, Mote Marine Laboratory (Sarasota, FL) for coordinating field work and obtaining specimens in Mexico, and to P. Ramsey, Louisiana Technical Uni- versity (Ruston, LA), and to J. Shaklee, Washing- ton State Department of Fisheries (Olympia, WA) for their helpful suggestions and efforts on our behalf. Literature cited Aebersold, P. B., G. A. Winans, D. J. Teel, G. B. Milner, and F. M. Utter. 1987. Manual for starch gel electrophoresis: a method for the detection of genetic variation. U.S. Dep. Commer, NOAA Tech. Rep. NMFS 61, 19 p. Baughman, J. L. 1941. Scombriformes, new, rare, or little known in Texas waters with notes on their natural history or distribution. Texas Acad. Sci. Trans. 24:14-26. Clayton, J. W., and D. N. Tretiak. 1972. Amine-citrate buffers for pH control in starch gel electrophoresis. J. Fish. Res. Board Canada 29:1169-1172. Collard, S. B., and L. H. Ogren. 1990. Dispersal scenarios for pelagic post-hatchling sea turtles. Bull. Mar. Sci. 47:233-243. Collette, B. B., and E. C. 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(2), 137 p. Deuel, D. G. 1973. 1970 salt-water angling survey. U.S. Natl. Mar. Fish. Serv. Curr. Fish. Stat. No. 6200, 54 p. Deuel, D. G., and J. R. Clark. 1968. The 1965 salt-water angling survey. U.S. Fish. Wildl. Serv., Resour. Publ. 67, 51 p. Fable Jr., W. A., L. Trent, G. W. Bane, and S. W. Ellsworth. 1987. Movements of king mackerel, Scomberomorus cavalla, tagged in southeast Loui- siana, 1983-85. Mar. Fish. Rev. 49(21:98-101. Finucane, J. H., L. A. Collins, H. A. Brusher, and C. H. Saloman. 1986. Reproductive biology of king mackerel, Scomberomorus cavalla, from the southeastern United States. Fish. Bull. 84:841-850. Grimes, C. B., A. G. Johnson, and W. A. Fable Jr. 1987. Delineation of king mackerel {Scomberomorus cavalla) groups along the U.S. east coast and in the Gulf of Mexico. In H. E. Kumpf, R. N. Vaught, C. B. Grimes, A. G. Johnson, and E. L. Nakamura (eds.), Proceedings of the stock identification workshop, p. 186-187. Dep. Commer, NOAA Tech. Memo. NMFS-SEFC-199. Grimes, C. B., J. H. Finucane, L. A. Collins, and D. A. Devries. 1990. Young king mackerel, Scomberomorus cav- alla, in the Gulf of Mexico, a summary of the dis- tribution and occurrence of larvae and juveniles, and spawning dates for Mexican juveniles. Bull. Mar. Sci. 46:640-654. Gulf of Mexico and South Atlantic Fishery Man- agement Councils. 1985. Final amendment 1, fishery management plan and environmental impact statement for coastal migratory pelagic resources (mackerels) in the Gulf of Mexico and South Atlantic region, Tampa, FL, var. p. Hartl, D. L. 1980. Principles of population genetics. Sinauer Assoc. Inc., Sunderland, MA, 488 p. Johnson, M. S. 1974. Comparative geographic variation in Menidia. Evolution 28:607-618. Kristjansson, F. K. 1963. Genetic control of two pre-albumins in pigs. Genetics 48:1059-1063. Kutkuhn, J. H. 1981. Stock definition as a necessary basis for co- operative management of Great Lakes fish resources. Can. J. Fish. Aquat. Sci. 38:1476- 1478. Manooch III, C. S. 1979. Recreational and commercial fisheries for king mackerel, Scomberomorus cavalla, in the South Atlantic Bight and Gulf of Mexico, U.S.A. In E. L. Nakamura and H. R. Bullis Jr. (eds.), Proceedings of the mackerel colloquium, p. 33- 41. Gulf States Mar. Fish. Comm., Rep. No. 4. Manooch III, C. S., E. L. Nakamura, and A B. Hall. 1978. Annotated bibliography of four Atlantic scombrids: Scomberomorus brasiliensis, S. cavalla, S. maculatus, and S. regalis. U.S. Dep. Commer, NOAA Tech. Rep. NMFS Circ. 418, 166 p. Markert, C. L., and I. Faulhaber. 1965. Lactate dehydrogenase isozyme patterns of fish. J. Exp. Zool. 159:319-332. Rivas, L. R. 1951. A preliminary review of the western north Atlantic fishes of the family Scombridae. Bull. Mar. Sci. Gulf Caribb. l(3):209-230. Johnson et al.: Evidence for distinct stocks of Scomberomorus cavalla 101 Royce, W. F. 1972. Introduction to the practice of fishery science. Acad. Press, New York, 428 p. Shaklee, J. B., F. W. Allendorf, D. C. Morizot, and G. S. Whitt. 1990. Genetic nomenclature for protein-coding loci in fish. Trans. Am. Fish. Soc. 119:2-15. Smith, P. J. 1990. Protein electrophoresis for identification of Australasian fish stocks. Aust. J. Mar. Freshwa- ter Res. 41:823-833. Sokal, R. R., and F. J. Rohlf. 1981. Biometry. W. H. Freeman and Co., New York, NY, 859 p. Sutherland, D. F, and W. A. Fable Jr. 1980. Results of a king mackerel (Scomberomorus cavalla) and Atlantic Spanish mackerel (Scomberomorus maculatus) migration study, 1975-79. U.S. Dep. Commer., NOAA Tech. Memo., NMFS-SEFC-12, 18 p. Sutter III, F. C., R. O. Williams, and M. F. Godcharles. 1991. Movement patterns and stock affinities of king mackerel in the southeastern United States. Fish. Bull. 89:315-324. Swofford, D. L., and R. B. Selander. 1981. Biosys-1. A computer program for the analy- sis of allelic variation in genetics. Release 1. Uni- versity of Illinois at Urbana, Champaign, IL, 66 p. Trent, L., W. A. Fable Jr., S. J. Russell, G. W. Bane, and B. J. Palko. 1987a. Variation in size and sex ratio of king mack- erel, Scomberomorus cavalla, off Louisiana, 1977- 85. Mar. Fish. Rev. 49(2):91-97. Trent, L., B. J. Palko, M. L. Williams, and H. A. Brusher. 1987b. Abundance of king mackerel, Scomberomorus cavalla, in the southeastern United States based on CPUE data from charterboats 1982-85. Mar. Fish. Rev. 49(2):78- 90. U.S. Department of Commerce. 1984. Fisheries of the United States, 1983. NOAA, NMFS, Curr. Fish. Stats. No.8320, 121 p. 1986. Marine recreational fishery statistics survey, Atlantic and Gulf coasts, 1985. NOAA, NMFS, Curr. Fish. Stats. No. 8327, 130 p. 1987. Fisheries of the United States, 1986. NOAA, NMFS, Curr. Fish. Stats. No.8385, 119 p. Abstract. — The spawning seasonality, fecundity, and daily egg production of three species of short-lived clupeids, the sardine Amblygaster sirm, the herring Herklotsichthys quadrimaculatus, and the sprat Spratelloides delicatulus were examined in Kiribati to assess whether vari- able recruitment was related to egg production. All species were multiple spawners, reproducing throughout the year. Periods of increased spawning activity were not related to seasonal changes in the physical environment. Spawn- ing activity and fish fecundity were related to available energy reserves and, hence, food supply. The batch fecundity of A. sirm and S. delicatulus also varied inversely with hydrated oocyte weight. The maximum reproductive life span of each species was less than nine months and averaged two to three months. Each species had a similar spawning frequency of three to five days, but this varied more in A. sirm and S. delica- tulus. Amblygaster sirm had the highest fecundity and potential lifetime egg production, but the number of eggs produced per ki- logram of fish was highest in the small sprat S. delicatulus. Monthly estimates of the daily egg production of each species varied with the proportion of the population that was spawning. Estimates of egg production showed little similarity to the fre- quency distribution of birthdates back-calculated from length-fre- quency samples. The distribution of back-calculated birthdates con- firmed that fish spawned in all months, but the proportion born each month varied widely from species to species and year to year. The reproductive strategy of these species ensures that successful spawning is likely, and so the level of recruitment is more de- pendent on post-hatching survival rates than on egg production. Reproductive biology and egg production of three species of Clupeidae from Kiribati, tropical central Pacific David A. Milton Stephen J. M. Blaber Nicholas J. F. Rawlinson CSIRO Division of Fisheries. Marine Laboratories, RO. Box 1 20, Cleveland, Queensland 4 1 63, Australia Manuscript accepted 24 September 1993 Fishery Bulletin 92:102-121 (1994) The sprat Spratelloides delicatulus, the herring Herklotsichthys quadri- maculatus, and the sardine Ambly- gaster sirm are the dominant tuna baitfish species in the Republic of Kiribati (Rawlinson et al., 1992). All three species inhabit coral reef lagoons and adjacent waters. Sprats school in shallow water around reefs and adjacent seagrass during the day. Herring also form dense schools in shallow water along the shoreline and among reefs during the day (Williams and Clarke, 1983). Unlike the other species, sardines school near the bottom of the lagoon during the day (Conand, 1988). All species disperse into the mid and upper waters of the lagoon during the night to feed and become available to the com- mercial fishery. A major source of lost fishing time by pole-and-line vessels in Kiribati has been irregular baitfish catches (Maclnnes, 1990). These important tuna baitfish species have shown large seasonal and interannual fluctuations in abun- dance since they were first re- corded during the 1940s (McCar- thy, 19851; Rawlinson et al., 1992). Both A. sirm and H. quadrima- culatus disappear from baitfish catches for variable periods and can be absent for months or years (Kiribati Fisheries Division, 19892). Changes in abundance may be related to variable or irregular re- cruitment, because many clupeoids (especially clupeids and engraulids) have little capacity to compensate for environmental variation during the period of peak spawning and egg production (Cushing, 1967, 1971). Most clupeids, including some tropical species, are multiple spawners (Alheit, 1989). Multiple spawning should be advantageous for short-lived species because it enables them to maintain rela- tively stable population sizes in unpredictable environments (Armstrong and Shelton, 1990). Multiple spawning has been estab- lished for few tropical clupeids (e.g., Sardinella brasiliensis; Isaac- Nahum et al., 1988). Of the three major baitfish species in Kiribati, only S. delicatulus has been shown to be a multiple-spawner (Milton and Blaber, 1991). All three species are subject to high natural mortal- ity in Kiribati (Rawlinson et al., 1992), thus lifetime egg production 1 McCarthy, D. 1985. Fishery dynamics and biology of the major wild baitfish species particulary Spratelloides delicatulus, from Tarawa, Kiribati. Kiribati Fisheries Div., Tarawa, Kiribati, 53 p. 2 Kiribati Fisheries Division. 1989. Fisher- ies Division 1989 Annual Rep., Ministry of Natural Resources Development, Tarawa, Kiribati, 38 p. 102 Milton et al.: Reproductive biology and egg production of three species of Clupeidae 103 may be increased if they spawned multiple batches of eggs. Egg production of multiple spawning species de- pends on reproductive life span, the time between spawnings, and the age structure of the population (Parrish et al., 1986). Batch fecundity of S. delica- tulus varies widely between sites, both within and between countries (Milton et al., 1990). In a short- lived species such as S. delicatulus (<5 months; Milton et al., 1991), reproductive life span may have an important influence on potential lifetime egg production. Batch fecundity of H. quadrimaculatus does not appear to vary throughout its distribution, and ranges from 4,000 to 10,000 eggs (Marichamy, 1971; Hida and Uchiyama, 1977; Williams and Clarke, 1983; Moussac and Poupon, 1986; Conand, 1988). Fish mature at about 90 mm in length at six months of age (Williams and Clarke, 1983), and they sur- vive for at least one year (Milton et al., 1993). Little is known of fecundity and egg production of A. sirm. Fecundity of the species is related to length and weight, with a mean of 20,000 eggs per batch, and individuals probably spawn more than one batch of eggs (Conand, 1988). Temperate clupeids vary widely in life-history parameters (e.g., Clupea spp., Jennings and Beverton, 1991). Food availability and environmen- tal conditions affect the size and number of eggs of Pacific herring (Clupea pallasi) (Hay and Brett, 1988). Results of studies of temperate clupeoids suggest that they do not spawn during periods of high food abundance, but store energy as fat for later reproductive activity (Hunter and Leong, 1981; lies, 1984). There are no similar studies of tropical clupeids. Encrasicholina heterolobus, a tropical engraulid, does not deplete energy reserves in the liver or soma during spawning (Wright, 1990). Fish with higher condition factor (K) also had higher fecundity. Stored energy or fish condition that may influence both spawning frequency and batch fecundity have a marked influence on egg production and, hence, affect subsequent recruitment (Ricker, 1954; Beverton and Holt, 1957). Adult reproductive varia- tion should strongly influence recruitment in short- lived tropical species that have short larval phases and rapid growth. An example is S. delicatulus which, in the Solomon Islands, live a maximum of five months and mature at about two months of age (Milton and Blaber, 1991; Milton et al., 1991). Amblygaster sirm and H. quadrimaculatus live less than two years (Milton et al., 1993) and mature in 6-12 months (Williams and Clarke, 1983; Conand, 1988). In this study, we examined the variability in re- productive biology of the three major baitfishes in Kiribati to determine the influence of adult repro- ductive variability on subsequent recruitment. Our objective was to test the hypothesis that reproduc- tive biology of short-lived clupeids is adapted to maintaining relatively stable population sizes. We determined potential life-time egg production and whether estimated egg production is related to the frequency distribution of back-calculated birthdates. Methods and materials Study areas The Republic of Kiribati covers an area of 3 x 106 km2 in the central Pacific ocean and comprises three main island groups (Gilbert, Phoenix, and Line Is- lands) (see Inset Fig. 1). The Gilbert Island group is the most populated, consisting of 16 coral reef islands. All islands in the group have a typical ocean platform coral reef structure and have been built up by scleractinian corals and coralline algae on a sub- merged mountain (Gilmour and Colman, 19903). Most atolls consist of small islets lying on the east- ern side of a lagoon with an open western side due to the prevailing easterly winds. Most typically have passages between the islets through which water is exchanged. The four study sites (Abaiang, Butaritari, Tarawa, and Abemema) were typical of islands in the Gilbert Island group; all had narrow islets on their south- ern and eastern sides, except Abaiang (Fig. 1). La- goons were mainly shallow (20-30 m deep), often with large areas of intertidal seagrass or sand on their eastern sides. Bottom topography of the deeper parts of the lagoon was generally smooth, with some coral outcrops. Our study sites were similar to those described by Hobson and Chess (1978) in the Marshall Islands. Environmental parameters On each sampling occasion, we measured the time of collection, sea surface temperature (°C), cloud cover (okters), wind direction and speed, and moon phase because these factors may be related to spawning or recruitment (Dalzell, 1985, 1987; Peterman and Bradford, 1987; Milton and Blaber, 1991). For each site, monthly rainfall data for 1989 Gilmour, A. J., and R. Colman. 1990. Report on a consultancy on a pilot environmental study of the outer island development program. Republic of Kiribati. Graduate School of the Environ- ment, Macquarie Univ., Australia, 151 p. 104 Fishery Bulletin 92(1), 1994 173°E I 175°E 177°E Butaritari Abaiang > Tarawa Gilbert Is 2°N- (D .Abemama $><0 o°- <%> <^ 100 kms 1S0°E Gilbert Is ,-, < *.r ^Solomon Australia N ^Tuvalu Fiii % Figure 1 Map of Gilbert Islands, Kiribati showing the four study sites (Butaritari, Abaiang, Tarawa, and Abemama). Inset shows the ter- ritorial boundary of Kiribati, the Gilbert Islands, and their posi- tion in the Pacific. and 1990 were obtained from the Kiribati Govern- ment Meteorological Division. Sampling Fifty to 1,000 Amblygaster sirm, Herklotsichthys quadrimaculatus, and Spratelloides delicatulus were collected monthly at one or more of four sites in Kiribati (Butaritari, Abaiang, Tarawa, and Abemama; Fig. 1) between August 1989 and May 1991. Additional samples of A. sirm and H. quadrimaculatus were collected in November 1988 and January 1989 from Tarawa. Fish were caught by several methods at each site. Most samples were collected from the commercial tuna baitfish catches each month at each site. Supplementary samples were ob- tained by beach-seining (H. quadrimaculatus and S. delicatulus), cast-netting (H. quadrimaculatus) in shallow water during the day, or gill- netting (25- and 38-mm stretched mesh) at night near baitfishing opera- tions. All fish were preserved in 70% ethanol. Reproductive biology Laboratory studies All fish collected from commercial baitfish sampling were measured (standard length in millimetres), and a subsample of 20 to 60 specimens weighed (±0.005 g). Go- nads, otoliths, liver, and viscera were removed and the amount of visible fat subjectively estimated. Both ovaries from the first 20 females of each spe- cies at each site for each month were dried of surface moisture, weighed (±0.001 g) and stored in 4% formalin- seawater for histology. Testes, ovaries of other fish, liver, and the soma were dried at 60°C to a constant weight. Otoliths were used to estimate the age (in days) of each fish by methods outlined in Milton et al. ( 1993). Addi- tional samples offish caught by other methods were treated separately, but in a similar way. We report only on results of studies offish collected from commercial samples unless otherwise stated. For histological preparations, go- nads were embedded in paraffin, sec- tioned at 9 mm, and stained with Ehrlich's haemotoxylin and eosin (McManus and Mowry, 1964). Gonad maturation stages were defined follow- ing Cyrus and Blaber (1984) and Hunter and Goldberg (1980), and were similar to those of Moussac and Poupon (1986) for H. quadrimaculatus from the Seychelles. We staged each gonad accord- ing to the relative numbers of cells at each develop- mental stage (Young et al., 1987; Table 1), and the presence of any post-ovulatory follicles was noted. The percentage of each histological section that cor- responded to each developmental stage was subjec- 2°S- Milton et al.: Reproductive biology and egg production of three species of Clupeidae 105 Table 1 Criteria used for staging female gonads of tropi- cal clupeids stained with haematoxylin and eosin. Stage Histology (1) Immature Chromatin nucleolar stage — prefollicle cells surround each oocyte (2) Developing/resting Perinucleolar stage — uniform staining cytoplasm (3) Maturing Yolk vesicle formation; some non-staining yolk (lipid) (4) Ripe Vitellogenic stage — red- staining yolk; developed chorion (5) Running ripe Globular red-staining yolk; (spawning) oocytes hydrated; develop- ment complete (6) Spent Presence of post-ovulatory follicles; cortical alveoli present and/or atresian of remaining ripe oocytes tively estimated. Post-ovulatory follicles were aged according to stages found in other multiple-spawning clupeoids (Hunter and Goldberg, 1980; Goldberg et al., 1984; Isaac-Nahum et al., 1988). Gonosomatic indices (GSI) were calculated as the ratio of wet gonad weight to somatic weight (total weight minus gonad weight), expressed as a percentage. Similarly, we calculated a hepatosomatic index (HSI) as the ratio of liver dry weight to somatic dry weight (total weight minus en- tire viscera), expressed as a percentage. Length and age at sexual maturity were defined as the minimum size and age at which fish had ripe oocytes (Stage 4), determined by histological exami- nation. Fish that had running-ripe oocytes (Stage 5) were recorded as in spawning condition. We defined the length and age at first spawning as the small- est size where the proportion of running-ripe oocytes in the section exceeded 85% for more than 50% of the fish of that length or age. We chose this crite- rion after examining large numbers of histological sections with running-ripe oocytes. In these sections they always represented more than 85% of the sec- tion area. Our results were similar to that found in other tropical clupeoids (Milton and Blaber, 1991). The reproductive life span of the population of each species at each site each month was determined from the oldest fish (Milton et al., 1993) in each sample minus the age at first spawning. We estimated batch fecundity for each species from fish that had been examined histologically and had oocytes that were starting to hydrate ( ripe-early running ripe; Stages 4-5; Table 1), but we did not examine the fecundity of fish with any empty fol- licles. An advanced modal size group of oocytes could be distinguished in ripe fish. We separated a subsample of between half (A. sirm) and all (S. delicatulus) of the ovary and weighed it. The num- ber of eggs in the advanced mode was counted and the fecundity was estimated by multiplying the number of eggs in the subsample by the ratio of total gonad weight to subsample weight. Fecundity esti- mates were made within three to four days after the ovary was removed from the fish to minimize the potential bias of differential absorption of fixative by oocytes and surrounding somatic tissue. We used hydrated oocytes from fish caught be- tween 2000 and 2330 hours to estimate egg weight. Oocyte weights were estimated from hydrated oo- cytes in ovaries that were almost ready to spawn (late Stage 5; Table 1). We measured oocyte dry weight by counting 10 samples of 10 oocytes from each ovary, drying the oocytes at 50° C to a constant mass and weighing each subsample separately. We scored visceral fat on a five-point scale. If a fish had less than 25% of the intestine covered in fat deposits, it was scored as (1); 25-50%, (2); 50- 75%, (3); and 75-100%, (4). A fish scored (5) when all intestine was covered with fat and deposits were also present around the stomach (Nikolsky, 1963). The proportion of females examined histologically each month that had post-ovulatory follicles (POF; Stage 6) was used to evaluate reproductive season- ality. We determined that these fish had spawned within the previous 15-48 hours, because these structures decompose and cannot be recognised af- ter that time (Hunter and Goldberg, 1980; Clarke, 1989). In samples where no fish had POF's, we used the proportion of fish in the histological subsample whose sections had greater than 85% running-ripe oocytes (Milton and Blaber, 1991). We used this proportion to calculate monthly estimates of mean daily oocyte production and the number of batches of oocytes spawned each month (Parrish et al., 1986). We estimated daily oocyte production (n/kg of adults; egg production index) for samples collected from commercial baitfishing, because these samples were assumed to be most representative of the popu- lation. Our methods were similar to those of Parker (1980, 1985), which have been used to estimate the spawning biomass of a number of multiple spawn- ers (Armstrong et al., 1988; Pauly and Palomeres, 1989; Somerton, 1990). However, our methods dif- fered because we used commercial catch per unit of effort (CPUE) as an index of adult abundance. Egg production index ^(fiPF.SR^^WiYcPUE (1) 106 Fishery Bulletin 92(1). 1994 where f- is the proportion of females in the ith length class, p is the proportion of the sample spawning, F is the fecundity of a fish of that length taken from the fecundity-length regression, SRt is the sex-ratio of the ith length class and Wi is the total weight of fish in the ith sample. CPUE was estimated from the monthly catch returns of the commercial fleet. We chose this method of estimat- ing egg production because S. delicatulus have de- mersal eggs (Leis and Trnski, 1989) and the eggs of A. sirm and H. quadrimaculatus are difficult to sample adequately in the large areas of suitable habitat in each lagoon. For comparison with adult spawning data, we back-calculated the distribution of birthdates offish collected in each length-frequency sample by using the growth equations of Milton et al. (1993). Fre- quencies in each age class were adjusted for mor- tality by using the estimates of Rawlinson et al. (1992). The distribution of birthdates was also back- calculated for H. quadrimaculatus and S. delica- tulus length-frequency samples from previous stud- ies at one site (Tarawa) January 1976 to February 1977 (R. Cross, 19784) and May 1983 to April 1984 (McCarthy, 19851). We used age distribution in these earlier studies and those of the present study to ex- amine seasonal, annual, and site-related differences in the reproductive life span of each species. Statistical analyses Inter- and intra-specific differ- ences in fat index, HSI and K were examined with Fisher's r-tests to account for unequal sample sizes. Seasonal and site-related differences in fecundity (expressed as oocytes per gram) were examined by analysis of covariance with weight as the covariate. Hydrated oocyte weight and reproductive life span were examined by one-way analysis of variance. We examined the relative influence of exogenous and endogenous factors on the fecundity of each species at each site by stepwise regression (Sokal and Rohlf, 1981). We included the following: length, weight, age, sea-surface temperature (°C), wind speed (in knots), moon phase (expressed by fitting a sin/cosin curve to the number of days since the last full moon before the sample was taken divided by the number of days in a lunar month (29.5) (Milton and Blaber, 1991), fish condition (K: weight/length3), fat, and HSIC7r ). We retained only those variables that significantly improved the fit of the model (P<0.05). Because several of these variables were correlated, we did a partial-correlation analysis be- tween these variables and fecundity, and the results 4 Cross, R. 1978. Fisheries research notes. Fisheries Division, Ministry of Commerce and Inductry, Tarawa, Kiribati, 58 p. of the two approaches were compared. If the variable most related to fecundity in the stepwise regression was not the one most related to fecundity in the par- tial-correlation analysis, the stepwise regression model was discarded and no relationship was assumed. In order to estimate egg production (Eq. 1), we estimated the proportion of females in each 5-mm length class from the total sample of each species. The variance of these estimates was calculated by using the normal approximation to the binomial dis- tribution (Walpole, 1974). We assessed whether the monthly percentage of annual egg production was related to the proportion of annual recruitment in the same month by rank-correlations (Conover, 1980). The average age of the potential spawning popu- lation in each sample was compared by a nested analysis of variance with month of sampling nested within year. Significant differences between treat- ments were identified from comparison of the least- squares means of each treatment, as sample sizes differed between cells (Sokal and Rohlf, 1981). Results Environmental parameters Sea-surface temperature in Kiribati varied little throughout the year. During the study period, tem- peratures varied between 29°C and 32°C (Table 2). Rainfall varied along the Gilbert Island group; rain- fall was higher in Butaritari than at the other sites. Some rain fell throughout the study period but was more intense during 1990 at all sites. Rainfall dur- ing 1989 was below the long-term average at all sites and was 16-50% that of 1990. The highest rainfall fell during the north-east monsoon (Decem- ber-April) at all sites. Winds were mostly light, and varied in direction seasonally, blowing from the east during the monsoon, but from the south-south-west for the rest of the year (Table 2). Reproductive biology Maturation The length and age at first maturity of A. sirm varied between sites (Table 3). Ambly- gaster sirm matured younger and smaller in Kiribati than elsewhere. Length and age at first spawning were much greater than the length or age when fish reached sexual maturity, but this size was similar to that of fish from northern Australia (Table 3). Herklotsichthys quadrimaculatus matured and were capable of spawning at 70 mm length and 4 months of age (Table 3). The relative size and age at which fish matured (as a proportion of maximum size and Milton et al.: Reproductive biology and egg production of three species of Clupeidae 107 age) did not differ among fish from the four sites. In Kiribati, S. delicatulus become sexually mature at 40 mm and two months of age and spawn shortly afterwards. Compared to the other species, the length and age at maturity and first spawning varied less among sites (Table 3). The three spe- cies differed in the length and age at sexual maturity and first spawning. However, as a proportion of their maxima, the three species were similiar Cu- test; P>0.1). All matured and spawned at about 70% of maxi- mum size and 50% of maxi- mum age (Table 3). Timing of spawning We iden- tified recent spawning by the presence of post-ovu- latory follicles in the ovaries. In A. sirm, follicles were detected in samples collected between 0100 to 1630 hours, and new post-ovulatory follicles (iden- tified as day-0 [<24 hr]; Hunter and Goldberg, 1980; Goldberg et al., 1984) were observed in fish collected between 0100 and 0510 hours. Female H. quadrimaculatus with post-ovulatory follicles were collected between 2130 and 1630 hours and day-0 follicles were found in samples collected between 2130 to 0300 hours. In female H. quadrimaculatus caught after 0300 hours, follicles could not be dis- tinguished from day-1 type POF's, as the follicles de- generated rapidly. Similarly, we detected post-ovu- latory follicles in female S. delicatulus collected from 2210 to 1930 hours, and follicles of all females col- lected earlier than 0845 hours were identified as day-0. Those in females of the single sample col- lected later in the day ( 1930) were assigned as day-1. Spawning season There was protracted spawning in A. sirm with periods of intense spawning activ- ity (Fig. 2). During both 1989 and 1990, fish spawned August to October and also during May- June in 1990. Condition, fat index, and HSI were less during spawning periods and reached a peak in March-April 1990, i.e., before spawning (Fig. 2). We found less fat deposits in spent fish and the fish were in poorer condition than fish with gonads in other stages of development (P<0.05; Table 4). We noted no significant differences in HSI among fish with gonads at the same stage of development. Herklotsichthys quadrimaculatus spawned throughout the study period: 20 to 50%> of the popu- Table 2 Mean water temperature (°C) , wind speed (kn), clou d cover, and monthly rainfall (mm) at four sites in Kiribati from November 1988 to May 1991. Parameter Butaritari Abaiang Tarawa Abemama Water temperature (°C) 30.2 ± 0.3 30.2 ± 0.4 29.5 ± 0.1 29.9 ± 0.2 Range 28-32 27-33 29-30 29-31 Wind speed (kn) 2.2 ± 0.6 4.2 ± 0.9 5.4 ± 1.2 2.2 ± 0.2 Range 0-7 0-10 1-15 1-5 Prevailing direction East East East East Cloud cover (okters) 2 ± 0.6 5 ± 0.6 3 ± 0.5 1 ± 0.4 Range 0-6 1-7 0-7 0-4 Monthly rainfall (mm) (1945- -88 263 ± 35 181 + 35 165 ± 35 128 ± 33 Range 7-908 0-761 0-824 0-728 Monthly rainfall 1989 (mm) 184 ± 29 42 ± 10 77 ± 23 36 ± 10 Range 51-351 0-108 6-235 3-102 Monthly rainfall 1990 (mm) 404 ± 37 - 298 ± 51 202 ± 31 Range 195-614 - 19-643 93-402 Months sampled 14 12 18 13 lation spawned each month (Fig. 3). Female condi- tion, fat index, and HSI all followed a similar pat- tern during the study but did not appear to be di- rectly related to spawning activity. Fish in spawn- ing condition had the highest HSI, fat, and condi- tion values, but these were only significantly greater than those of spent fish (P<0.05; Table 4). Spratelloides delicatulus spawned almost continu- ously throughout the study period but spawning varied in intensity (Fig. 4). Peak spawning occurred during different periods in each of the years sampled. Female HSI and fat index showed a simi- lar pattern during the study but monthly changes in these parameters or fish condition did not follow the spawning cycle. We found no significant differ- ences in HSI or fat index for females with ovaries in different stages of development (P>0.1; Table 4). Fish condition was lower among spent fish than in ripe or spawning fish (P<0.05; Table 4). Females with ripe ovaries had higher mean HSI, fat, and condition than those in other stages of development, but these differences were not significant (Table 4). Fecundity The relative fecundity of A. sirm and H. quadrimaculatus did not differ among sites or sea- sonally within sites in Kiribati (ANCOVA with weight as covariate; overall P>0.07; Table 5). How- ever, the relative fecundity of H. quadrimaculatus was significantly different between fish from Tarawa and Abemama (<-test; P<0.05). Batch fecundity of both species did not differ among sites in Kiribati. Within their respective species groups, both species had simi- lar batch fecundities to the other species listed, al- though their relative fecundities were lower (Table 5). 108 Fishery Bulletin 92(1), 1994 Table 3 Length and age at sexual maturity and first spawning of Amblygaster sirm, Herklotsichthys quadrimaculatus, and Spratelloides delicatulus from various populations throughout their range. (L t = length at maturity, L&. = length at first spawning, Lmax = maximum size, Tmal = age at maturity, Tf = age at first spawning, maximum age, K = Kiribati, I = India, SI = Solomon Islands). 2 Length at Length at first Age at first maturity (mm) spawning (mm) Age at maturity(d) spawning (d) Species Site «w/*w> a-fip/LmaJ (T IT mat1 max) ^fspITmax) Source' A. sirm Kiribati 110 (0.50) 180 (0.80) 150 (0.29) 330 (0.65) (1) New Caledonia 132 (0.72) — 295 (0.40) — (2) N. Australia 174 (0.79) 193 (0.88) — — (3) Sri Lanka 166 (0.88) — -330 (0.80) — (4) Mean 146 (0.72) — — H. quadrimaculatu 8 Hawaii 80 (0.63) 90 (0.70) 160 (0.53) 190 (0.63) (5) Marshall Is. 90 (0.82) — 190 (0.72) — (6) Fiji 95 (0.78) 98 (0.80) 275 (— ) 294 (— ) (7), (8) Butaritari (K) 65 (0.68) 70 (0.74) 125 (0.50) 135 (0.53) (1) Abaiang (K) 70 (0.74) 70 (0.74) 125 (0.37) 125 (0.37) (1) Tarawa (K) 69 (0.72) 70 (0.73) 138 (0.45) 150 (0.48) (1) Abemama (K) 70 (0.64) 72 (0.65) 140 (0.34) 150 (0.36) (1) New Caledonia 91 (0.64) — 244 ( — ) — (9) Andaman Is. (I) 99 (0.81) 104 (0.85) — — (10) Seychelles 97 (0.71) — 150 (0.30) — (11) Mean 83 (0.72) 82 (0.74) 172 (0.46) 174 (0.47) S. delicatulus Fiji 35 (0.56) 39 (0.63) 52 (0.43) 61 (0.51) (7). (8) Butaritari (K) 40 (0.68) 40 (0.68) 65 (0.51) 68 (0.54) (1) Abaiang (K) 45 (0.75) 53 (0.88) 62 (0.51) 80 (0.64) (1) Tarawa (K) 45 (0.68) 50 (0.76) 77 (0.50) 90 (0.57) (1) Munda (SI) 37 (0.58) 37 (0.58) 72 (0.47) 78 (0.51) (12), (13) Vona Vona (SI) 37 (0.66) 37 (0.66) 68 (0.53) 72 (0.56) (12), (13) Tulagi (SI) 38 (0.60) 38 (0.60) 73 (0.55) 75 (0.57) (12), (13) Maldives 38 (0.69) 40 (0.73) 90 (0.60) 97 (0.65) (13), (14) India 42 (0.71) — — — (15) Mean 40 (0.66) 42 (0.69) 70 (0.51) 78 (0.57) ' Sources: (1) present study. (2) Conand (1991). (3) Okera (1982), (4) Dayaratne and Gjosaeter (1986). (5) Williams and Clarke (1983), (6) Hida and Uchiyama (1977), (7) Lewis et al. (1983), (8) Dalzell et al. (1987), (9) Conand (1988), (10) Marichamy (1971), (11) Moussac anf Poupon (1986), (12) Milton and Blaber 11991), (13) Milton et al. (1991), (14) Milton et al. (1990), (15) Mohan and Kunhikoya (1986). Using stepwise linear regression, we found that fecundity was related to weight in all species (Table 6; Fig. 5). Fecundity of A. sirm was significantly correlated with HSI and fish condition. Fish condi- tion, HSI, and fat index were all correlated with fecundity in H. quadrimaculatus (Table 6). Fecun- dity was significantly correlated with weight and condition at two of the four sites. Although, when data from all sites were combined, weight and fat index were the only significant correlates. Fecundity of S. delicatulus varied widely among sites, both within Kiribati and among countries (Table 5). In Kiribati, relative fecundity was higher at Butaritari than at Abaiang (P<0.05), but differed less than among sites in the Solomon Islands. Fe- cundity did not vary seasonally at any site. Relative fecundity of S. delicatulus was highest in New Caledonia — significantly higher than at all other sites except Butaritari in Kiribati (Table 5). How- ever, the relative fecundity of S. delicatulus was lower than its congeners, S. gracilis and S. lewisi, at sites where they co-occurred (Table 5). We found that the fecundity of S. delicatulus cor- related strongly with fish weight (Fig. 5). The only other factor related to fecundity in S. delicatulus was HSI. There was a significant relationship be- tween fecundity and HSI at Butaritari and Tarawa and when all data were combined. Spawning fish had a higher HSI at Butaritari than at other sites (2.24 ± 0.13 vs. 1.41 ± 0.08; P<0.001). The HSI of male S. delicatulus that had a GSI similar to that of spawning females (>5%) was also Milton et al.: Reproductive biology and egg production of three species of Clupeidae 109 o o 5 4 (5 3 2 1 i~+i ~i — i — i — i — r B J -\ — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i i i 1.5 -, j'f'm'a'm'j'j'a's'o'n'd 1989 j'F m a'm'j'j'a's'o' 1990 Figure 2 Monthly variation (±95% confidence limits) in (A) condition, (B) visceral fat index, (C) hepatosomatic index and (D) proportion spawning of female Amblygaster sirm from Kiribati between January 1989 and October 1990. 5-i £ 4 o 3 H Ny*** V****V 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 4 -i 3 CO O - 1 0 B &*k K/V^V I I I I 1 1 I I I I 1 I I 1 I I I 1 I I I rn n I r~\ r~ * 2-\ £ H #-****^ i i 1 1 i i i i i i i i i i 1 1 i i i i i i i i i i 1 1 i i 100 Figure 3 Monthly variation (±95% confidence limits) in (A) condition, (B) visceral fat index, (C) hepatosomatic index and (D) proportion spawning of female Herklotsichthys quadrimaculatus from Kiribati be- tween November 1988 and April 1991. higher at Butaritari (1.41 ± 0.06; N=57) than at other sites (Abaiang HSI=1.07 ± 0.11; N=7; Tarawa HSI=0.81 ± 0.09; A/=14). The proportion of male S. delicatulus that had GSI greater than 5% was also higher at Butaritari (36%) than at other sites (Abaiang 17.5%; Tarawa 20%). Oocyte weights of A. sirm and S. delicatulus dif- fered significantly from site to site (Table 7). In S. delicatulus, we found the greatest oocyte weight at Abemama and Abaiang — significantly higher than at Butaritari and Tarawa (P<0.01). Oocyte weights in A. sirm were also higher at Abaiang (P<0.001; Table 7). We found no significant differences among sites for oocyte weights of H. quadrimaculatus. Sex ratio The sex-ratio of A. sirm, H. quadrima- culatus, and S. delicatulus changed as fish grew but only among the largest length classes of each spe- cies were there significant deviations from a ratio of 1:1. In all three species, females dominate the largest length classes (Fig. 6). In our samples, we found significantly more female A. sirm and S. delicatulus among fish larger than the length at first spawning (180 and 45 mm respectively). With H. 10 Fishery Bulletin 92(1), 1994 Table 4 Mean hepatosomatic index (HSI: %), visceral fat index (Fat) and condi- tion (K: dry weight/length3) of Amblygaster sirm, Herklotsichthys uadrimaculatus and Spratelloides delicatulus at different stages of gonadal development (SE = standard error ± N = number of females examined). Species Stage HSI ± SE Fat ± SE tf(xlO-6)+SE N A. sirm maturing 0.38 ± 0.06 ripe 0.43 ± 0.04 spawning 0.39 ± 0.06 spent 0.42 ± 0.05 H. quadrimaculatus maturing ripe spawning spent 0.87 ± 0.08 0.96 ± 0.06 1.04 ± 0.07 0.69 ± 0.04 S. delicatulus maturing 1.41 ± 0.19 ripe 1.98 + 0.10 spawning 1.84 ± 0.15 spent 1.46 ± 0.10 3.4 ± 0.6 3.2 ± 0.3 2.6 ± 0.4 1.7 ± 0.2 1.4 ± 0.1 1.6 ± 0.1 1.8 ± 0.1 1.7 ± 0.1 1.3 ± 0.2 1.6 ± 0.1 1 3 • n 1 1.2 ± 0.1 4.05 ± 0.25 4.14 ± 0.07 4.03 ± 0.13 2.77 ± 0.13 ;!ii 16 6 3.89 ± 0.05 45 3.81 + 0.05 127 3.91 ± 0.05 95 3.53 ± 0.06 40 2.36 ± 0.06 2.51 ± 0.04 2.46 ± 0.04 2.28 ± 0.04 15 41 35 55 higher lifetime egg production at all sites than did co-occur- ring S. delicatulus. The num- ber of days between successive spawnings influenced esti- mates of lifetime egg produc- tion. Although longer in A. sirm, the difference was not significant (Table 8). quadrimaculatus, females dominated among fish over 80 mm (Fig. 6). Egg production The number of spawnings per month and the daily egg production of all species generally followed the pattern of the proportion spawning (Fig. 7). We found lower daily egg produc- tion in A. sirm than in the other species. During the period of maximum spawning activity, A. sirm and H. quadrimaculatus spawned up to 20 times per month (Fig. 7), and S. delicatulus spawned daily. Reproductive life span The reproductive life span of A. sirm was significantly longer in Tarawa (60.1 + 15.4 days) than at the other sites during 1989-90 (P<0.01; Table 8). Similarly, we found H. quadrimaculatus had a longer reproductive life span at Abemama (141.8 ± 30.9 days) than at other sites during 1989-91 (P<0.01; Table 8). During the same period, the reproductive life span of S. delicatulus was similar at all sites (57.5 ± 4.6 days). However, the reproductive life span of S. delicatulus at Tarawa varied significantly between years; fish caught during 1990-91 were not as old as those in previous j'ears (P<0.05; Table 8). No corresponding pattern was observed in H. quadrimaculatus from Tarawa. Herklotsichthys quadrimaculatus and S. delicatulus lived significantly longer after maturity than A. sirm (P<0.01). Our estimates of maximum lifetime egg produc- tion of A. sirm were similar at the two sites (Abaiang and Tarawa). Herklotsichthys quadrimaculatus had Recruitment Amblygaster sirm recruited from a single protracted period in Kiribati during 1989 (March to October; Fig. 8). We found a greater pro- portion of survivors had been born between March and July than in all other months except September (P<0.05). There were insufficient data to com- pare monthly egg production with recruitment, but the pe- riod of highest recruitment corresponded with the times of greatest spawning activity. However, this did not appear to be directly related to the absolute number of oocytes produced (Fig. 7). The proportion of H. quadrimaculatus born each month differed over the four years (P<0.05; Fig. 9). In 1976, the greater proportion were born from November to March, while in 1983 over 40% were born during July. Fish caught during 1989-90 showed a different pattern. The highest proportion in 1989 were born in May, whereas in 1990 the high- est proportion were born in January. Over all 4 years' data, December ( 15.4% ) and July ( 13.7% ) had the greatest mean proportion of births (P<0.05), but the July value may be biased by the large value in 1983 (Fig. 10). Where data were comparable, we found no relationship between proportion of annual recruitment and monthly egg production (r.=0.70, P<0.10, N=6 in 1989; rs=-0.15, P>0.5, N=U in 1990). The proportion of S. delicatulus born each month varied considerably among the four years examined (Fig. 10). December had the highest proportion of births in 1976. In 1983, most fish were born between May and August, and a similar pattern was found in 1989. By comparison, the distribution of birthdates was more evenly spread in 1990 (Fig. 10). The months with the largest mean proportion across the four years were May (11.2%), June (14.9%), July (15.8%), and December (11.9%). We found a nega- tive relationship between the proportion of births and egg production in 1990 (rs=-0.58; P<0.05, 7V=10). Milton et al.: Reproductive biology and egg production of three species of Clupeidae I 1 *. 35 | 3.0 o 2.5 o 2.0 i i i i i i i i i i i i i i i % ^vy/** 2- B v fi^^A 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3-j i 1 V 1 1 1 1 1 1 1 "I- 1 — n — r~rn — i — i — l 100 -| 80- r 60 40 20 D n'dIjfMaMj'j'a's'o'n'dIj'fWaMj'j'a's'o'n'dIj'fWaM 1991 1989 1990 Figure 4 Monthly variation (±95% confidence limits) in (A) condition, (B) visceral fat index, (C) hepatosomatic index and (D) proportion spawning of female Spratelloides delicatulus from Kiribati between November 1988 and May 1991. 30000 - A 20000 - F= 180 5W10 r2=0.48 N=32 sb o 10000- ° s' s^° 0 8000-. 50 100 150 3000 - c F = = 699.9W r2=074 1 2000- o° o ?y^^ N=87 O 00Ocf <4/"o o 1000 - „ooe 0 o 0 - Figure 5 The relationship between batch fecundity and fish weight for (A) Amblygaster sirm, (B) Herklotsichthys quadrimaculatus and (C) Spratelloides delicatulus from Kiribati. Discussion The reproductive cycles of A. sirm, H. quadrima- culatus, and S. delicatulus in Kiribati are similar to that reported for temperate multiple-spawning clupeoids (Hunter and Goldberg, 1980; Gil and Lee, 1986; Shelton, 1987; Alheit, 1989). Most studies on multiple spawning clupeoids have been on engraulids; these species spawn many batches of eggs each year and have variable batch fecundity (Alheit, 1989). Our results for H. quadrimaculatus and S. delicatulus from Kiribati agree with previ- ous reproductive studies of these species in tropical areas (McCarthy, 19851; Moussac and Poupon, 1986; Milton and Blaber, 1991). In the tropics, both spe- cies spawn throughout the year, but have periods when spawning activity is greater. In more temper- ate parts of their range, the reproductive season of both H. quadrimaculatus and S. delicatulus is shorter and coincides with increases in water tem- perature in early summer (Williams and Clarke, 1983; Lewis et al., 1983; Conand, 1988). 1 12 Fishery Bulletin 92|l), 1994 Table 5 Mean length (mm), age (y ears), fecundity, relative fecundity (eggs g ') of Amblyg aster sirm, Herklot sichthys quadrimaculatus, and Spratelloides delicatulus and other tropical and subtropical clupeids ( sardines, her- rings, and sprats) (K = Ki ribati, SI = Solomon Island 3, I = India, P.N.G. = Papua New Guinea UK = = United Kingdom, SU = Soviet Un ion, G = Germany ). Length Age Fecundity Rel. fecundity Species Site ± SE ± SE ± SE ± SE N Source Sardines Amblygaster sirm Abaiang (Kl 189 ± 5 0.97 ± 0.03 18789 ± 2757 187.1 ± 25.3 7 (1) Tarawa (K) 194 ± 1 1.04 ± 0.03 20327 ± 1391 192.0 ± 12.0 25 (1) New Caledonia 139-177 0.90-2.2 8000-27780 300.0 ± 16.9 24 (2), (3) Sardinella brasiliensis Brazil 162 ± 2 — 23318 ± 2065 356 ± 37 23 (4) S. marquesensis Marquesas Is. 109 ± 6 — 4150 + 1000 — 6 (5) S. zunasi Korea 75-142 1-3 8800-58800 — 31 (6) Herrings Herklotsich thys uadnmaculatus Hawaii 80-121 — 1155-6296 160-311 46 (7) Marshall Is. 100 + 2 0.59 ± 0.02 4755 ± 380 — 7 (8) Butaritari (K) 75 ± 1 0.45 ± 0.01 1844 ± 108 295.5 ± 12.1 44 (1) Abaiang (K) 75 ± 1 0.45 ± 0.02 1975 ± 133 317.4 ± 19.3 27 (1) Tarawa (K) 76 ± 1 0.44 ± 0.02 2353 ± 110 344.1 ± 10.2 63 (1) Abemama (K) 84 ± 2 0.61 ± 0.04 3008 ± 207 319.1 ± 22.7 33 (1) Andaman Is. (I) 95-115 — 8353 ± - — 19 (9) Seychelles 88-127 — 4500-8000 — 24 (10) Opisthonema libertate Mexico 142 ± 1 — 57125 + 1850 553 ± 14 115 (11) Sprats Spratelloides dehcatul us Butaritari ( K) 52 ± 2 0.27 ± 0.01 1359 ± 143 867 ± 55 1') (1) Abaiang (K) 52 ± 1 0.21 ± 0.02 973 ± 43 667 + 35 7 (1) Tarawa (K) 54 ± 1 0.29 ± 0.01 1255 ± 54 735 ± 25 49 (1) Abemama (K) 41 ± 1 0.20 ± 0.02 524 + 95 702 ± 75 12 (1) Munda (SI) 48 ± 1 0.26 ± 0.01 799 ± 45 554 ± 25 57 (12) Vona Vona (SI) 49 ± 1 0.26 ± 0.01 925 ± 102 717 ± 45 28 (12) Tulagi (SI) 46 ± 1 0.21 ± 0.01 926 ± 93 567 ± 49 28 (12) New Caledonia 45 — 710 883 ± 14 20 (2) India 40 ± 3 — 608 ± 54 — 15 (13) S. gracilis Munda (SI) 50 0.19 514 504 1 (14) Vona Vona (SI) 37 ± 1 0.15 + 0.01 505 ± 51 882 ± 68 13 (12) P.N.G. 53 ± 2 — 2592 ± 313 1690 ± 96 18 (15) Maldives 59 ± 1 0.29 ± 0.02 1998 ± 137 1073 ± 54 33 (12) India 40 + 5 — 790 ± 71 962 ± 53 If. (13) S. lewisi Munda (SI) 44 ± 1 0.18 + 0.01 887 + 20 925 ± 16 219 (14) Vona Vona (SI) 42 ± 1 0.14 ± 0.01 930 ± 51 1032 + 36 62 (14) Tulagi (SI) 49 ± 1 0.28 ± 0.02 1290 ± 84 1230 ± 69 29 (14) Sprattus sprattus Scotland (UK) 108 3 2729 187 64 (16) Baltic Sea (SU) 121 1.9 2174 232 46 (17) North Sea (G) — 2 — 413 — (17) Sources: (li present study, (21 Conand (1988), (3) Conar d (1991), (4i Isaac-Nahum et al. (1988). (5) Nakamura and Wilson (19701, (6) Gil and Lee (19861. (7) Williams and Clarke (1983), (8) Hid a and Uchiyama (1977), (9) Manchamy (1971), (10) Moussac ar d Poupon (1986), (11) Torres-Villegas and Perezgomez (1988), (121 Miltor et al. (1990). (13i Mohan and Kunhikoya (1986), (14) Milton unpubl. data. 1 15i Dalzell (19851, (16) De Silva (1973), (17i Alheit (1988). Although we found A. sirm also had an extended spawning season in Kiribati, the species may not spawn throughout the year. Our result differs from previous studies that found the spawning season lasted two to five months during early summer (Conand, 1991) or the monsoon period (Rosa and Laevastu, 1960; Dayaratne and Gjosaeter, 1986). Neither temperature nor rainfall appear to be the Milton et al.: Reproductive biology and egg production of three species of Clupeidae I 13 proximate stimuli for spawning of A. sirm in Kiribati. Tempera- ture was constant throughout the year and rainfall was higher at all sites in Kiribati between December and April, when spawning activity was lowest. Most spawning activity in this species occurred during the second half of the year when the prevailing wind di- rection changed from east to west, associated with the north- west monsoon that starts at this time (Burgess, 19875). Our limited wind and rainfall data did not indicate that increased spawning activity in A. sirm was related to the shift in weather pattern. Gonad maturation and spawning were also linked to changes in fish liver-weight (HSI), visceral fat, and condi- tion of each species. Either HSI or fat index and condition were all significantly reduced in postspawning fish. Amblygas- ter sirm stores energy in the viscera rather than in the liver. Other multiple-spawning clup- eoids also transfer energy from stored fat to reproductive tis- sue (Dahlberg, 1969; Okera, 1974; Hunter and Leong, 1981). In contrast, spent H. quadrimaculatus and S. delieatulus had reduced HSI, which suggests that the liver is the energy store utilized during repro- duction (Diana and MacKay, 1979; Smith et al., 1990). Energy stored in this organ would be readily available for rapid assimilation; hence, fish could spawn multiple batches of eggs rapidly. Studies of temperate herring, Clupea harengus, have shown that gonad maturation is linked to food availability and fat storage (Linko et al., 1985; Henderson and Almatar, 1989; Rajasilta, 1992). Ovaries of all three species in Kiribati and of S. delieatulus in the Solomon Islands (Milton and Blaber, 1991) vary in a similar way to herring. Milton and Blaber (1991) did not find a direct rela- tion between spawning and prey availability. This suggests that while gonad maturation in these clu- 5 Burgess, S. M. 1987. The climate of western Kiribati. New Zealand Meterological Service, Wellington, NZ. Miscellaneous publ. 188, part 7. Table 6 Stepwise regression of the re lationship between various endogenous factors and fish fecundity from sites in Kiribati. 200 Length (mm) 100 <30 40 50 60 70 80 90 >90 Length (mm) 100 BO 60 - 40 20 - C <30 35 40 45 50 55 60 >60 Length (mm) Figure 6 Ontogenetic change in the proportion female of (A) Amblygaster sirm , (B) Herklotsichthys quad- rimaculatus and (C) Spratelloides delicatulus (±95"7r confidence limits) from Kiribati. peids is probably linked to cycles in prey abundance, fat storage may reduce the effects of short-term fluc- tuations in prey abundance on reproduction. Diel timing of spawning events was similar for all species. We found new post-ovulatory follicles (day- 0) in females collected from 2130 hours onwards with the greatest proportion detected after 0100. This indicates that these species spawn during the early part of the night, probably prior to midnight. Our results are consistent with previous studies that found high densities of A. sirm eggs in the plank- ton after midnight (Delsman, 1926; Lazarus, 1987). Studies of other sardines (Goldberg et al., 1984; Isaac-Nahum et al., 1988; Re et al., 1988) and tropi- . TO Q 0 j'f'm'a'm'j'j'a's'o'n'd 1989 Avr j'f'm'a'm'jjVs'o' 1990 Time o n c i ND J FMAMJ J ASOND J FMAMJ J ASONCJj FMAM 1989 1990 1991 Time t"| i i i i i i i i i i i i n i i i i i n i I i i i ii ND J FMAMJ J ASOND J FMAMJ J ASOND J FMAM 1989 1990 1991 Time Figure 7 Monthly estimates of daily egg production of (A) Amblygaster sirm, (B) Herklotsichthys quadrima- culatus and (C) Spratelloides delicatulus from Kiribati during the study period. cal clupeoids (Clarke, 1987) also showed that spawn- ing peaked before midnight. Length and age at sexual maturity of A. sirm and H. quadrimaculatus in Kiribati differed from those Milton et al.: Reproductive biology and egg production of three species of Clupeidae 1 15 Tab e 8 Mean reproductive life spa n (in days) and days between spawning of Amblygaster sirm, Herklotsichthys quadrimaculatus, and Spra telloides delicatulus from four sites in Kiribati (TV = number of length- frequency samples; No. = number of months examined). Days Max. lifetime Reproductive between egg production Species Site Year life span ± SE Range N spawning Range No. (> 104l A. sirm Abaiang 1989-1990 19.0 + 6.4 0-66 L2 20.0 Tarawa 1989-1990 60.1 ± 15.4 0-127 7 41.6 Abemama 1989-1990 3.2 + 3.1 0-19 6 Overall 1989-1990 26.7 ± 6.8 0-127 25 6.2 ± 2 3 1.5-25.9 10 38/ H. quadrimaculatus Butaritari 1989-1991 47.3 + 15.0 0-201 14 11.9 Abaiang 1989-1991 73.9 + 15.6 0-201 17 12.8 Tarawa 1976/83/89-91 84.1 + 7.8 0-254 64 19.3 Abemama 1989-1991 141.8 + 30.9 0-286 12 27.7 Overall 1989-1991 80.6 ± 8.6 0-286 74 3.1 ± 0.3 1.3-4.7 15 21.1 S. delicatulus Butaritari 1989-1991 53.6 + 4.6 24-74 15 1.9 Abaiang 1989-1991 49.2 ± 5.1 21-80 11 1.5 Tarawa 1989-1991 66.9 + 10.6 0-144 16 3.5 Tarawa 1976 76.6 ± 10.5 45-129 7 3.1 Tarawa 1983/84 84.3 ± 9.5 34-152 16 3.7 all 1989 90.0 + 13.4 53-144 7 3.2 all 1990/91 51.0 ± 4.1 0-109 35 2.5 Overall 1989-1991 57.5 ± 4.6 0-144 42 5.2 ± 1.8 1-30 16 3.2 20 1989 (n = 717) III ■ (%) c Recruitme o I III 0 l.llllllll.l J FMAMJ JASOND Time Figure 8 The proportion of Amblygaster sirm (±95% confidence limits 1 sampled between August 1989 and July 1990 born each month in 1989, backcalculated from length-frequency samples. in other parts of their range (Table 3). We found few differences within Kiribati, but both species became sexually mature and spawned at much shorter body lengths than at other locations. Herklotsichthys quadrimaculatus did not grow as large in Kiribati as elsewhere (Milton et al., 1993). but the propor- tion of maximum size at which this species matured was similar throughout its range. Milton and Blaber ( 1991) found regional differences in length at sexual maturity in other small tropical clupeoids; they sug- gested these differences were consistent with the hypothesis of Longhurst and Pauly (1987) that fish of any species living in cooler water will grow to and mature at a larger size through the interaction of oxygen supply and demand. Our data on H. quadrimaculatus is consistent with this hypothesis — the other studies were all at sites at higher lati- tudes than Kiribati, where the water temperature is lower. Also, the proportion of maximum size at which fish matured was similar at all locations, despite the absolute differences in size at maturity in Kiribati. By comparison, A. sirm matured at a smaller size and grew to a larger size in Kiribati than at other locations (Milton et al., 1993). The proportion of maximum size at which fish matured was also lower than found in previous studies and was less than the proportion common to a wide range of clupeoids (70%; Beverton, 1963). In response to severe fishing pressure, the size and age at sexual maturity of several sardine species have been found to decline (Murphy, 1977). Presumably, this is because any density-dependent effects are reduced during early growth (Beverton and Holt. 1957; Ware, 1980). Amblygaster sirm can have high or variable adult mortality in Kiribati (Rawlinson et al., 1992), favouring early maturation (Stearns and Crandall, 1984). Length at first spawning was a similar proportion of maximum size for the three species and was con- ] 16 Fishery Bulletin 92(1), 1994 50 1976 40 (n = 13131) 30 20 -| 10 0 ■■■■-■■..■II 50 | 1983 40 (n= 1705) ■ 30 ?20 ■ b 10 1 ° ll_«j 5 50 1989 o £40 (n = 4745) 30 20 m 10 ___ll...llll 50 1990 40 (n = 2747) 30 20 ■ : laHlll-l. J FMAMJ JASOND Time Figure 9 The proportion of Herklotsichthys quad- rimaculatus born each month in 1976, 1983, 1989, and 1990, back-calculated from length- frequency samples (95'/ confidence limits of all proportions are all less than 1.5%). 30 1976 • (n = 3002) 20 10 0 ..lllll — ll 30 | 1983 ■ (n = 9198) 20 I- lent (%) o o l...lll. .1 fc 5 30 _ 1989 (n = 2014) 01 20 l| 10 ■ III 0 ■ ■■lllll— 30 -i 1990 (n = 16419) 20 10 o- .llllllllll. J FMAMJ JASOND Time Figure 10 The proportion of Spratelloides delicatulus born each month in 1976, 1983, 1989, and 1990, back-calculated from length-frequency samples I95'i confidence limits of all propor- tions are all less than 1.59f ). sistent with the close relation with maximum size found by Blaxter and Hunter (1982) for other clupeoids. These authors also noted a latitudinal effect; fish from lower latitudes spawned at a smaller proportion of maximum size. Temperate clupeids (especially herrings, Clupea spp.) show a great plasticity in the number and size of eggs produced; many species show seasonal, and inter-annual, as well as geographic, variation in their reproductive outputs (Alheit, 1989; Jennings and Beverton, 1991) reflecting energetic resources and environmental conditions (Hay and Brett, 1988; Henderson and Almatar, 1989). By comparison, the tropical herring, H. quadrimaculatus, spawned throughout the year and showed negligible tempo- ral or spatial variation in fecundity, egg weight, or inter-spawning interval. This indicates that egg production was almost constant throughout the study period and suggests that adult food resources and larval survival are predictable or relatively con- stant (Sibly and Calow, 1983). In comparison to other species, S. delicatulus had a higher relative fecundity that was also correlated with HSI. Females in spawning condition also had a higher HSI at Butaritari. Commercial CPUE was highest at this site (Rawlinson et al., 1992) and S. delicatulus spawned more, smaller eggs than at other sites where relative fecundity was lower. These data suggest that the fecundity of S. delicatulus may be influenced by the amount of energy stored in the liver. This energy store would be important in a small multiple-spawning species; it would enable the fish to continue spawning dur- ing short periods of reduced food supply (Hay and Brett, 1988). The length of the inter-spawning in- terval has been shown experimentally to be related Milton et al.: Reproductive biology and egg production of three species of Clupeidae I / to food supply in other fish species (Townshend and Wootton, 1984). Fish at Butaritari may experience a more predictable environment that enables them to produce more eggs of smaller size than fish in more variable environments. In contrast, A. sirm delayed spawning beyond the size and age at sexual maturity and did not spawn until one year old. As fecundity was related to weight, delayed spawning enabled A. sirm to grow faster than the other species (Milton et al., 1993) and have a higher batch fecundity when spawning started. Murphy (1968) hypothesized that delayed spawning and longer reproductive life span would evolve in response to variable reproductive success. However, Armstrong and Shelton (1990) demon- strated that, even with a short reproductive life span, multiple spawners had a high probability of successful reproduction when subject to random environmental fluctuations over time. Thus, delay- ing spawning would be of adaptive advantage if mortality was low (Roff, 1984) because batch fecun- dity and lifetime egg production would be increased. Our estimates of the reproductive lifespan of A. sirm indicate that this species spawns fewer times in their lifetime than other species and thus would also have less chance of successful spawning than other species. Given that this is the longest-lived of the species examined, our estimate of overall mean lifespan may be biased by the small number of months sampled. Large fish may be under-repre- sented in small catches and may contribute to un- derestimating the reproductive potential of A. sirm. Herklotsichthys quadrimaculatus had a longer reproductive life span and spawned more frequently than did the other species. Reproductive life span varied little among sites (except Abemama) and there was no significant temporal variation, which suggests that survival rates of large adult H. quadrimaculatus are fairly constant in Kiribati. This is reflected in their life-history parameters, which varied little among sites or over time. In con- trast, the frequency distribution of back-calculated birthdates indicated that overall survival was vari- able both between and within years, and was not related to monthly egg production. We have no es- timates of adult abundance during the study period, and so population egg production could not be as- sessed. However, the annual CPUE and abundance of H. quadrimaculatus in the baitfishery were simi- lar in the three years for which both data sets were available (Rawlinson et al., 1992). This suggests that population size was relatively constant during this period. If so, then variation in post-hatching survival probably has an important influence on recruitment in this species (Smith, 1985). The reproductive life span of the smallest species, S. delicatulus, was intermediate between the other species and varied little among sites during 1989 and 1990. Unlike H. quadrimaculatus, the reproduc- tive life span of S. delicatulus varied between years, which suggests that survival rates are not as con- stant or as predictable as those of H. quadrimaculatus. Potential lifetime egg production of each female was only one tenth that of other spe- cies, but, because of the larger number of females, monthly estimates of daily egg production were higher. The distribution of back-calculated birthdates varied between years, but a greater pro- portion of births fell in May-August, irrespective of che pattern of egg production. Annual CPUE of S. delicatulus (Rawlinson et al., 1992) was similar in 1989 and 1990, which suggests that fishing mortal- ity had not contributed to the increased mortality that reduced the reproductive life span in 1990. The reproduction and abundance of S. delicatulus may be more directly influenced by its environment than are the other species. Adult survival is vari- able and low (Tiroba et al., 1990); egg production varies, probably in response to food supply, and sur- vival to recruitment is unpredictable. Yet the poten- tial for successful reproduction with this strategy may still be relatively high (Armstrong and Shelton, 1990). In contrast, H. quadrimaculatus appears to be able to offset environmental variability to produce a relatively constant supply of eggs. The distribution pattern of back-calculated birthdates of each species was not consistent among species. Months when a higher proportion survived differed for each species during all years; months with highest mean survival were not the same for any species. This suggests that the effects of envi- ronmental conditions such as seasonal food avail- ability or favorable physical conditions are not the same for each species. Alternatively, other factors such as predation (Rawlinson et al., 1992) may have greater influence on survival to recruitment. Egg production by S. delicatulus was positively corre- lated to survival rates in 1989 and negatively cor- related in 1990. This seems unrelated to fish abun- dance as catch rates were higher in 1989 than in 1990 (Rawlinson et al., 1992). Large variations in recruitment, reflected in catch rates of the main baitfishes do not appear to be di- rectly linked with variations in egg production. All spawn in the lagoon for most of the year, and dis- tribution of birthdates indicated recruitment in most months. Although the absolute level of recruitment varied throughout the year, multiple spawning re- duces fluctuations in population size due to environ- mental variability and should ensure that relatively 18 Fishery Bulletin 92(1), 1994 stable population sizes are maintained. Earlier stud- ies of A. sirm and H. quadrimaculatus in Tarawa lagoon suggested that these species spend at least part of their life outside the lagoon (R. Cross, 19784; McCarthy, 19851). If this is the case, fluctuations in the relative abundances of these species may be re- lated to migrations; a better understanding of the fac- tors causing large-scale movements is necessary before predicting the potential yield of this fishery. Acknowledgments We thank staff of the Kiribati Fisheries Division for assistance with fieldwork during the project. Peter Crocos and Jock Young and three anonymous re- viewers made constructive comments on earlier drafts. This work formed part of the baitfish re- search project PN 9003 funded by the Australian Centre for International Agricultural Research. Literature cited Alheit, J. 1988. Reproductive biology of sprat {Sprattus sprattus): factors determining annual egg production. J. Cons. int. Explor. Mer 44:162-168. 1989. Comparative spawning biology of anchovies, sardines, and sprats. Rapp. P.-v. Reun. Cons. int. Explor. Mer 191:7-14. Armstrong, M. J., and P. A. Shelton. 1990. Clupeoid life-history styles in variable environments. Envir. Biol. Fish 28:77-85. Armstrong, M. J., P. A. Shelton, I. Hampton, G. Jolly, and Y. Melo. 1988. Egg production estimates of anchovy biom- ass in the southern Benguela system. Calif. Coop. Oceanic Fish. Invest. Rep. 26:3040. Beverton, R. J. H. 1963. Maturation, growth and mortality of clupeid and engraulid stocks in relation to fishing. Rapp. P-V. Reun. Cons. int. Explor. Mer 154:44-67. Beverton, R. J. H., and S. J. Holt. 1957. On the dynamics of exploited fish populations. Fish Invest., Ser. 2, Mar. Fish. GB Minist. Agric. Fish. Food 19:1-533. Blaxter, J. S. H., and J. R. Hunter. 1982. The biology of the clupeoid fishes. Adv. in Mar. Biol. 20:1224. Clarke, T. A. 1987. Fecundity and spawning frequency of the Hawaiian anchovy or Nehu, Encrasicholina purpurea. Fish. Bull. 85:127-138. 1989. Seasonal differences in spawning, egg size, and early development time of the Hawaiian an- chovy or nehu, Encrasicholina purpurea. Fish Bull. 87:593-600. Conand, F. 1988. Biology and ecology of small pelagic fish from the lagoon of New Caledonia usable as bait for tuna fishing. Thesis Studies, ORSTOM, Paris, 239 p. 1991. Biology and phenology of Amblygaster sirm (Clupeidae) in New Caledonia, a sardine of the coral environment. Bull. Mar. Sci. 48:137-149. Conover, W. J. 1980. Practical nonparametric statistics, 2nd ed. J. Wiley & Sons, NY, 493 p. Cushing, D. H. 1967. The grouping of herring populations. J. Mar. Biol. Assoc. U.K. 47:193-208. 1971. The dependence of recruitment on parent stock in different groups of fishes. J. Cons. int. Explor. Mer 33:340-362. Cyrus, D. P., and S. J. M. Blaber. 1984. The reproductive biology of Gerres in Natal estuaries. J. Fish. Biol 24:491-504. Dahlberg, M. D. 1969. Fat cycles and condition factors of two spe- cies of menhaden, Brevoortia (Clupeidae), and natural hybrids from the Indian river of Florida. Am. Midi. Nat. 82:117-126. Dal/.. -II, P. 1985. Some aspects of the reproductive biology of Spratelloides gracilis (Schlegel) in the Ysabel Passage, Papua New Guinea. J. Fish. Biol. 27:229-237. Dalzell, P., S. Sharma, and J. Prakash. 1987. Preliminary estimates of the growth and mortality fo three tuna baitfish species, Herklot- sichthys quadrimaculatus, Spratelloides deli- catulus and Rhabdamia gracilis from Fijian waters. Tuna Billfish Assessment Programme Tech. Rep. 20:1-15. Dayaratne, P., and J. Gjosaeter. 1986. Age and growth of four Sardinella species from Sri Lanka. Fish. Res. 4:1-33. Delsman, H. C. 1926. Eggs and larvae from the Java Sea: 7, The genus Clupea. Treubia 8:218-239. De Silva, S. S. 1973. Aspects of the reproductive biology of the sprat, Sprattus sprattus in inshore waters of the west coast of Scotland. J. Fish Biol. 5:689-705. Diana, J. S., and W. C. MacKay. 1979. Timing and magnitude of energy deposition and loss in the body, liver and gonads of northern pike (Esox lucius). J. Fish. Res. Board. Can. 36:481-487. Gil, J. W., and T. W. Lee. 1986. Reproductive ecology of the scaled sardine, Sardinella zunasi (Family Clupeidae), in Cheonsu Bay of the Yellow sea, Korea. In T Uyeno, R. Arai, T. Taniuchi, and K. Matsuura (eds.), Indo- Pacific fish biology: proceedings of the second in- ternational conference on Indo-Pacific fishes, p. 818-829. Ichthyological Society of Japan, Tokyo. Milton et al.: Reproductive biology and egg production of three species of Clupeidae 1 19 Goldberg, S. R., V. H. Alarcon, and J. Alheit. 1984. Postovulatory follicle histology of the Pacific sardine, Sardinops sagax, from Peru. Fish. Bull. 82:443-445. Hay, D. E., and J. R. Brett. 1988. Maturation and fecundity of Pacific herring (Clupea harengus pallasi): an experimental study with comparisons to natural populations. Can. J. Fish. Aquat. Sci. 45:399^106. Henderson, R. J., and S. M. Almatar. 1989. Seasonal changes in the lipid composition of herring (Clupea harengus) in relation to gonad maturation. J. Mar. Biol. Assoc. U.K. 69:323-334. Hida, T. S., and J. H. Uchiyama. 1977. Biology of the baitfishes Herklotsichthys punctatus and Pranesus pinguis in Majuro, Marshall Islands. In R. S. Shomura, (ed.), Col- lection of tuna baitfish papers. NOAA Technical Report NMFS Circular 408:63-68. Hobson, E. S., and J. R. Chess. 1978. Trophic relationships among fishes and plankton in the lagoon at Enewetak Atoll, Marshall Islands. Fish. Bull. 76:133-153. Hunter, J. R., and S. R. Goldberg. 1980. Spawning incidence and batch fecundity in northern anchovy Engraulis mordax. Fish. Bull. 77:641-652. Hunter, J. R., and R. Leong. 1981. The spawning energetics of female northern anchovy Engraulis mordax. Fish. Bull. 79: 215-230. lies, T. D. 1984. Allocation of resources to gonad and soma in Atlantic herring Clupea harengus. In G. W. Potts and R. J. Wootton (eds.), Fish reproduction: strate- gies and tactics, p. 331-347. Academic Press, London. Isaac-Nahum, V. J., R. de D. Cardoso, G. Servo, and C. L. del B. Rossi-Wongtschowski. 1988. Aspects of the spawning biology of the Bra- zilian sardine, Sardinella brasiliensis, (Clup- eidae). J. Fish Biol. 32:383-396. Jennings, S., and R. J. H. Beverton. 1991. Intraspecific variation in the life history tac- tics of Atlantic herring (Clupea harengus) stocks. ICES J. Mar. Sci. 48:117-125. Lazarus, S. 1987. Studies on the early life history of Sardinella sirm from Vizhinjam, southwest coast of India. Indian J. Fish. 34:28-40. Leis, J. M., and T. Trnski. 1989. The larvae of Indo-Pacific shorefishes. N.S.W. Univ. Press, Kensington, NSW, Australia, 371 p. Lewis, A. D., S. Sharma, J. Prakash, and B. Tikomainiusiladi. 1983. The Fiji baitfishery 1981-82, with notes on the biology of the gold spot herring Herklotsichthys quadrimaculatus (Clupeidae), and the blue sprat Spratelloides delicatulus (Dussumieriidael. Fish. Div. Min. Agr. Fish. Tech. Rep. Fiji No. 6, 35 p. Linko, R. R., J. K. Kaitaranta, and R. Vuorela. 1985. Comparison of the fatty acids in Baltic her- ring and available plankton feed. Comp. Bio- chem. Physiol. 82B:699-705. Longhurst, A. R., and D. Pauly. 1987. Ecology of tropical oceans. Academic Press, San Diego, CA, 407 p. Maclnnes, M. 1990. The status of the tuna baitfishery in Kiribati and its impact on the tuna industry. In S. J. M. Blaber and J. W. Copland (eds.), Tuna baitfish in the Indo-Pacific region, p. 55-59. ACIAR Proc. 30. McManus, J. F. A., and R. W. Mowry. 1964. Staining methods: histological and histochemical. Harper Row, New York, NY, 423 p. Marichamy, R. 1971. Maturity and spawning of the spotted her- ring, Herklotsichthys punctatus from the Andaman sea. Ind. J. Fish." 18:148-155. Milton, D. A., and S. J. M. Blaber. 1991. Maturation, spawning seasonality, and proxi- mate spawning stimuli of six species of tuna baitfish in the Solomon Islands. Fish. Bull. 89:221-237. Milton, D. A., S. J. M. Blaber, G. Tiroba, J. L. Leqata, N. J. F. Rawlinson, and A. Hafiz. 1990. Reproductive biology of Spratelloides delicatulus, S. gracilis and Stolephorus heterolobus from Solomon Islands and Maldives. In S. J. M. Blaber and J. W. Copland (eds.), Tuna baitfish in the Indo-Pacific region, p. 89-99. ACIAR Proc. 30. Milton, D. A., S. J. M. Blaber, and N. J. F. Rawlinson. 1991. Age and growth of three species of tuna baitfish (genus: Spratelloides) in the tropical Indo- Pacific. J. Fish Biol. 39:849-866. 1993. Age and growth of three tropical clupeids from Kiribati, central south Pacific. J. Fish Biol. 43:89-108. Mohan, M., and K. K. Kunhikoya. 1986. Biology of the bait fishes Spratelloides delicatulus and S. japonicus from Minicoy waters. Cent. Mar. Fish. Res. Inst. Bull. 36: 155-164. Moussac, G., and J. C. Poupon. 1986. Growth and reproduction of Herklotsichthys punctatus (Pisces: Clupeidae) from Seychel- les. Cybium 10:31-45. Murphy, G. I. 1968. Pattern in life history and the environment. Am. Nat. 102:391^103. 1977. Clupeoids. In J. A. Gulland (ed.), Fish popu- lation dynamics, p. 283-308. J. Wiley & Sons, New York. Nakamura, E. L., and R. C. Wilson. 1970. The biology of the Marquesan sardine, Sardinella marquesensis. Pac. Sci. 24:359-376. Nikolsky, G. V. 1963. The ecology of fishes. Academic Press, Lon- don, 352 p. 120 Fishery Bulletin 92|1), 1994 Okera, W. 1974. Morphometries, 'condition' and gonad devel- opment of the east African Sardinella gibbosa and Sardinella albella. J. Fish Biol. 6:801-812. 1982. Observations on the maturation condition of some pelagic fishes from northern Australian waters. CSIRO Marine Lab. Rep. 144:1-15. Parker, K. 1980. A direct method for estimating northern an- chovy, Engraulis mordax biomass. Fish. Bull. 78:541-544. 1985. Biomass model for the egg production method. In R. Lasker (ed.), An egg production method for estimating spawning biomass of pelagic fish: application to the northern anchovy, Engraulis mordax, p. 5-6. NOAA Technical Re- port NMFS 36. Parrish, R. H., D. L. Mallicoate, and R. A. Klingbeil. 1986. Age dependent fecundity, number of spawnings per year, sex ratio, and maturation stages in northern anchovy, Engraulis mordax. Fish. Bull. 84:503-517. Pauly, D., and M. L. Palomares. 1989. New estimates of monthly biomass, recruit- ment and related statistics of anchoveta (Engraulis ringens) of Peru (4-14RS), 1953- 1985. In D. Pauly, P. Muck, J. Mendo, and I. Tsukayama (eds.), The Peruvian upwelling ecosys- tem: dynamics and interactions, p. 189- 206. IMARPE and GTZ and ICLARM, Peru. Peterman, R. M., and M. J. Bradford. 1987. Wind speed and mortality rate of a marine fish, the northern anchovy, Engraulis mordax. Science 235:354-356. Rajasilta, M. 1992. Relationship between food, fat, sexual matu- ration, and spawning time of Baltic herring (Clupea harengus membras) in the Archipelago Sea. Can. J. Fish. Aquat. Sci. 49:644-654. Rawlinson, N. J. F., D. A. Milton, and S. J. M. Blaber. 1992. Tuna baitfish and the pole-and-line industry in Kiribati. ACIAR Technical Report No. 24, 90 p. Re, P., A. Farinha, and I. Meneses. 1988. Diel spawning time of sardine, Sardina pilchardus (Teleostei, Clupeidae), off Port- ugal. Inv. Pesq. 52:207-213. Ricker, W. E. 1954. Stock and recruitment. J. Fish. Res. Board Can. 11:559-623. Roff, D. A. 1984. The evolution of life history parameters in teleosts. Can. J. Fish. Aquat. Sci. 41:989-1000. Rosa, H., and T. Laevastu. 1960. Comparison of biological and ecological char- acteristics of sardines and related species — a pre- liminary study. In H. Rosa and G. Murphy (edsj, Proceedings of the world scientific meeting on the biology of sardines and related species, Vol. 2, p. 521-552. FAO, Rome. Shelton, P. A. 1987. Life-history traits displayed by neritic fish in the Benguela current ecosystem. In A. I. L. Payne, J. A. Gulland, and K. H. Brinks (eds.), The Benguela and comparable ecosystems. S. Afr. J. Mar. Sci. 5:235-242. Sibly, R., and P. Calow. 1983. An integrated approach to life-cycle evolution using selective landscapes. J. Theor. Biol. 102:527-536. Smith, P. E. 1985. Year-class strength and survival of 0-group clupeoids. Can. J. Fish. Aquat. Sci. 42 (Suppl. l):69-82. Smith, R. L., A. J. Paul, and J. M. Paul. 1990. Seasonal changes in energy and the energy cost of spawning in Gulf of Alaska Pacific cod. J. Fish Biol. 36:307-316. Sokal, R. R., and F. J. Rohlf. 1981. Biometry, 2nd ed. Freeman and Co., New York, NY, 859 p. Somerton, D. A. 1990. Baitfish stock assessment using the egg pro- duction method: an application on the Hawaiian anchovy or nehu (Encrasicholina purpurea). In S. J. M. Blaber and J. W. Copland (eds.). Tuna baitfish in the Indo-Pacific region, p. 152- 158. ACIAR Proc. 30. Stearns, S. C, and R. E. Crandall. 1984. Plasticity for age and size at sexual maturity: a lifehistory response to unavoidable stress. In (I. W. Potts and R. J. Wootton (eds.), Fish reproduc- tion: strategies and tactics, p. 13-33. Academic Press, London. Tiroba, G., N. J. F. Rawlinson, P. V. Nichols, and J. L. Leqata. 1990. Length-frequency analysis of the major baitfish species in Solomon Islands. In S. J. M. Blaber and J. W. Copland (eds.). Tuna baitfish in the Indo-Pacific region, p. 114-134. ACIAR Proc. 30. Torres- Villegas, J. R., and L. Perezgomez. 1988. Fecundity variation of Opisthonema libertate (Pisces: Clupeidae) from 1983 to 1985 in Bahia Magdalena, Baja California, sur Mexico. Inv. Pesq. 52:193-206. Townshend, T. J., and R. J. Wootton. 1984. Effects of food supply on the reproduction of the convict cichlid, Cichlasoma nigrofasciatum. J. Fish Biol. 24:91-104. Walpole, R. E. 1974. Introduction to statistics, 2nd ed. MacMillan Publ., New York, NY, 340 p. Ware, D. M. 1980. Bioenergetics of stock and recruitment. Can. J. Fish. Aquat. Sci. 37:1012-1024. Williams, V. R., and T. A. Clarke. 1983. Reproduction, growth, and other aspects of the biology of the gold spot herring, Herklotsichthys Milton et al.: Reproductive biology and egg production of three species of Clupeidae 121 quadrimaculatus (Clupeidae), a recent introduction the Indo-Pacific region, p. 83-88. ACIAR Proc. 30. to Hawaii. Fish. Bull. 81:587-597. Young, J. W., S. J. M. Blaber, and R. Rose. Wright, P. J. 1987. Reproductive biology of three species of 1990. The reproductive strategy of Stolephorus midwater fishes associated with the continental heterolobus in the south Java Sea. In S. J. M. slope of eastern Tasmania, Australia. Mar. Biol. Blaber and J. W. Copland (eds.), Tuna baitfish in 95:323-332. Abstract. Determination of stock structure for striped dol- phins (Stenella coeruleoalba) in the eastern Pacific has been prob- lematic, because very few speci- mens have been available for study. We compared length data obtained from vertical aerial pho- tographs of 28 schools of striped dolphins from the northern and southern regions of the eastern tropical Pacific and found no sig- nificant differences in average length for adult animals (> 180cm) or for adult females, defined here as dolphins closely accompanied by a calf. Analyses of back-pro- jected birth dates for dolphins >155cm revealed a broad pulse in reproduction extending from the fall through the spring; however, sample size was inadequate to compare timing of reproduction between the two areas. Striped dolphins measured from aerial photographs were longer on aver- age than those killed incidentally in fishing operations. We found a pattern of segregation by size be- tween schools that is analogous to the separate schools of juveniles and adults that are found in the western Pacific. We hypothesized that the specimen data base may be biased because tuna purse- seine fishermen in the eastern tropical Pacific may selectively set on schools composed of younger, smaller dolphins. Examination of stock and school structure of striped dolphin (Stenella coeruleoalba) in the eastern Pacific from aerial photogrammetry Wayne L. Perryman Morgan S. Lynn Southwest Fisheries Science Center National Marine Fisheries Service. NOAA 8604 La Jolla Shores Drive. La Jolla. Calif 92037 Manuscript accepted 20 September 1993 Fishery Bulletin 92:122-131 (1994) Because striped dolphins, Stenella coeruleoalba, are killed incidentally in purse-seine fishing for yellowfin tuna in the eastern tropical Pacific (ETP), the National Marine Fisher- ies Service (NMFS) is required by the Marine Mammal Protection Act (as amended in 1988) to monitor trends in their abundance (Holt and Sexton, 1989; Wade and Gerrodette, in press). To satisfy this congressional mandate, infor- mation on stock structure is re- quired. The determination of stock structure for striped dolphins in the ETP has been particularly dif- ficult because of the small number of animals killed in the tuna fish- ery and, therefore, small number of specimens available for study (DeMaster et al., 1992). In the ab- sence of morphological, life history, or genetic data to provide evidence of reproductive isolation, stocks of striped dolphins have been identi- fied provisionally based on discontinuities in distribution. With more sighting data from ob- servers aboard fishing vessels and research cruises, the number of proposed stocks has decreased from five or six (Smith, 19791; Holt and Powers, 1982) to one (Dizon et al., in press) pending availability of additional data. For this report, we examined length data to help clarify the issue of stock structure. These data were extracted from vertical aerial pho- tographs collected during line transect surveys and are thus pre- sumably free of any "sampling" bi- ases associated with the fishery. Here, we compare length samples from aerial photographs of animals from the northern and southern stock regions proposed by Perrin et al. (1985) for evidence of differences in average length or timing of re- production. Data were then com- pared with measurements avail- able from specimens killed inciden- tally in purse-seine fishing. We also examined the frequency distribu- tion of lengths within individual schools. These data were used to test for size-age segregation, as reported for dolphins taken in the drive fishery on the Pacific coast of Japan (Miyazaki, 1977; Miyazaki and Nishiwaki, 1978). Methods Length measurements were made on vertical aerial photographs of 28 schools of striped dolphins (Fig. 1). We photographed the schools with a KA-45A military reconnaissance 1 Smith, T. D. (ed). 1979. Report of the sta- tus of porpoise stocks workshop; 27-31 August, La Jolla, California. U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Southwest Fish. Sci. Cent, P.O. Box 271, La Jolla, CA 92038. Admin Rep., L.l-79- 41, 120 p. 122 Perryman and Lynn: Stock and school structure of Stenella coeruleoalba 123 30' 20' 10' 10' 20' 160° W 150° 120° 1 10" 90» 80° W Figure 1 Distribution of schools of striped dolphin, Stenella coeruleoalba, (dark circles), from which data were taken for this report. Boundaries for northern and southern stocks were taken from Perrin et al. 1985. camera mounted below the fuselage of a Hughes 500D helicopter that was launched from the NOAA Ship David Starr Jordan. This photographic sam- pling was part of a long-term research effort con- ducted by NMFS to monitor trends in abundance of dolphin populations in the ETP (Holt and Sexton, 1989; Wade and Gerrodette, in press). The reconnaissance camera was equipped with a very fast, medium focal length lens (152 mm) and a forward image motion compensation system that eliminated the blur normally found in images taken from a low altitude, high-speed platform. We used Kodak Plus-X Aerecon II (thin-base) film, exposed through a medium yellow filter, throughout the experi- ment. This filter significantly reduced the amount of blue light reaching the film, thus enhancing both the contrast and resolution of our photographs. The observer sitting in the right front seat of the helicopter triggered the camera, controlled cycle rate and shutter speed, and adjusted the forward motion compensation system. As each firing pulse was sent to the camera, a data acquisition system recorded the time that the image was captured and an alti- tude reading from the helicopter's radar altimeter. To check for accuracy in our recorded altitude data (A J, we photographed calibration target arrays and compared altitude calculated from measurements of these known distances with recorded altitude (see Perryman and Lynn, 1993). We found a consistent bias in Ar and used the lin- ear regression equation shown below to calculate a corrected altitude (AJ for each photograph used in this report. Ac = (Ar ) 1.013 - 33.755 (r2 = 0.993) . Length determination We reviewed the images of 88 schools of striped dolphins photographed from 1987 through 1990 and selected the images of 28 schools that provided the best combination of image clarity and water pen- etration. From this sample, we selected the photo- graphic pass over each school that captured the larg- est number of dolphins swimming parallel to and 124 Fishery Bulletin 92(1), 1994 very near the surface. Dolphins were not measured if either the rostrum or tail flukes were not clearly visible or if they were surfacing, diving, or jumping, which would make them appear shorter when viewed from above. Because there was from 80 to 90% overlap between adjacent photographs, the same dolphin could often be measured in two to four photographs. If more than one length was available for a dolphin, the largest length was selected, as- suming it was the best determination of true length. This helped to minimize the reduction in apparent length caused by the normal swimming movements of the dolphins (Scott and Perryman, 1991; Perryman and Lynn, 1993). We measured each dolphin from the tip of the rostrum to the trailing edge of the tail flukes (Fig. 2). These points were selected because the fluke notch that is used to determine standard length (Norris, 1961) was very difficult to see in most of the images. For adult specimens, this measurement should exceed standard length by 2-2.5 cm (Chivers, 19932). The measurements were made on sections of the original black and white negatives that we captured with a high-resolution video camera and transferred to a Macintosh Ilci computer. Image enhancement and length measurements were made with the aid of the digital image processing and analysis program, Image (version 1.37), which was developed by the National Institute of Health (W. Rasband, Research Services, Bethesda, Maryland). The length of each dolphin was determined by mul- tiplying its length on the image by the scale of the photograph ( scale= A/lens focal length ). Data analysis Perrin et al. (1985) compared the mean lengths of physiologically adult male and female dolphins from 2 S. Chivers. 1993. Southwest Fisheries Science Center, La Jolla, California 92037. unpubl. data. ■Photo Length- putative geographic stocks of several species to pro- vide supporting morphological evidence for repro- ductive isolation. For our analyses, we used length as the criteria for eliminating the youngest dolphins from our sample. Based on the length data for adult striped dolphins in Perrin et al. (1985) and a review of our length sample, we estimated that the mini- mum length for adult female striped dolphins in the eastern Pacific is about 180 cm. We used this length as our first cut-off point, and tested for differences U-test) between the means of our length samples (<180 cm) from the northern and southern regions (Fig. 1). Since the selection of this value was some- what arbitrary, we repeated the tests on data sets with minimum values of 185 and 190 cm. Based on behavioral arguments reviewed in Per- ryman and Lynn ( 1993), we assumed that the larger dolphin swimming closely alongside a calf was an adult female. Since this determination was based on behavior and not on examination of sexual charac- ters, we qualify the term in quotation marks, "adult female," whenever we are referring to a length sample based on this assumption. A <-test was used to compare the mean lengths of "adult females" from the northern and southern regions. We also per- formed a power analysis to determine what range of differences between means we could expect to detect (probability of type II error < 0.10) for this analysis and the ones described in the paragraph above. Calf birth dates We examined the length data from striped dolphins estimated to be one year old or less for evidence of pulses in reproduction (see Barlow 1 1984], for spot- ted and spinner dolphins; Perryman and Lynn (1993], for common dolphins). Ninety centimeters was used as the best estimate of average length at birth and 155 cm for average length at one year for striped dolphins in the eastern Pacific (Gurevich and Stewart, 1979;i). We as- sumed postnatal growth was linear during the first year and back-projected the birth dates for all dolphins <155 cm in length. Our goal here was not to determine the ex- Photo Length ■Standard Length- Figure 2 Illustration of the difference between points used to determine standard length and length as measured from our vertical photographs. 1 Gurevich, V. S., and B. S. Stewart. 1979. A study of growth and re- production of the striped dolphin iStenella coeruleoalba). U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Southwest Fish. Sci. Cent., P.O. Box 271, La Jolla, CA 92038. Final Rep to NOAA, SWFC Con- tract 03-78-D27-1079, 29 p. Perryman and Lynn: Stock and school structure of Stenella coeruleoalba 125 act date of birth for each dolphin but rather to exam- ine the distribution of birth dates, based on the same assumptions, from the two regions. We used Kupier's modification of Kolmogorov's test for com- parisons of circular distributions (Batschelet, 1965) to compare the calculated distribution of birth dates with a uniform distribution. Comparisons with specimen data We conducted four tests to compare the sample of photogrammetric lengths with data collected from striped dolphins killed incidentally in purse-seine fishing in the ETP (Perrin et al., 1976). The data from specimens included the information published by Perrin et al. ( 1985) and a small set of data from dolphins killed since 1985. T-tests were used to com- pare the mean length of "adult females" with the mean length of adult female specimens and with the mean length of lactating adult female specimens. We also compared the mean (f-test) and shape (Kolmogorov-Smirnov test) of the photogrammet- rically determined length distribution of striped dolphins > 180 cm with data from specimens > 180 cm in length. School structure Examination of the structure of schools of striped dolphins captured in the drive fishery in Japan has revealed a distinct pattern of segregation based on sex, maturity, and length (Miyazaki, 1977, 1984; Miyazaki and Nishiwaki, 1978). Researchers have categorized these schools as adult, juvenile, or mixed depending on the proportion of juvenile dolphins (excluding calves) captured. In these studies, length (<174 cm) or age (<1.5 years) was used as the crite- rion for eliminating nursing calves from the sample; the remainder of the dolphins was determined to be juvenile or adult by direct examination of the go- nads. We examined the length distributions for the pho- tographed schools to see if an analogous pattern of segregation in schools from the eastern Pacific was detectable. We divided our samples into two length categories which we labeled juvenile or adult. The minimum length for the juvenile category was set at 165 cm to eliminate nursing calves as described above. We selected this minimum value because 1) length at birth for striped dolphins from the ETP is apparently about 10 cm shorter than that reported from the western Pacific (Miyazaki, 1977; Gurevich and Stewart, 19793), and we assumed that the dif- ference in the average length at weaning was ap- proximately the same; 2) dolphins larger than 165- 170 cm in length were very rarely found swimming in the characteristic cow/calf configuration we see in our photographs. We selected 195 cm as the upper bound for the juvenile category because this appears to be about the minimum size for adult male striped dolphins that have been killed in the ETP tuna purse-seine fishery (Perrin et al., 1985). This value was keyed to male length data because the studies of school structure from Japan indicated that a disproportion- ate number of the dolphins captured in juvenile schools were males (Miyazaki and Nishiwaki, 1978). Thus dolphins in each school were categorized as juvenile if they were between 165 and 195 cm in length and as adult if they were > 195 cm in length. The goal in this classification scheme was to create one category that would be composed of mostly ju- venile and young adult dolphins and another that would include mostly adult animals. We used chi-square analysis to test the hypoth- esis that the number of dolphins in the two catego- ries in our schools was independent of school. For this analysis, we eliminated schools from which we had measured less than 20% of the school or fewer than 17 dolphins. The second criterion was estab- lished to minimize the number of predicted values in the chi-square analysis that were less than five. Application of these criteria reduced our sample to 21 schools for this test. Because the selection of 195 cm for the cut-off between the two size categories probably includes more adult females in the juve- nile category than males, we decreased the limit to 190 cm and repeated the chi-square test. We also conducted a regression analysis to determine whether the proportion of the measured sample in the juvenile category was related to school size. With the exception of the power analyses and birth date comparison which were done by hand, all tests presented in this report were performed with the program StatView developed by Abacus Con- cepts (Berkeley, CA). Unless noted otherwise, tests were considered significant for P values < 0.05. Results Regional comparisons We compared the average length of striped dolphins from the northern and southern regions and found no significant differences between the samples (Table 1; Fig. 3). In tests for differences in mean lengths of "adult females" (Fig. 4), no differences were found between the regions. Although none of the differences was significant, means of the 126 Fishery Bulletin 92(1), 1994 samples from the northern region were generally a few centimeters smaller than those from the south, a pattern reported by Perrin et al. (1985). This level of difference was less than we could detect given the available sample and the variability of our data Table 1 Results of r-tests for differences between means of length samples from striped dolhin, Stenella coerueoalba, from the northern (Nor) and south- ern (So) regions. Sub-sample (cm) n Nor/So mean (cm) Nor/So P (2-tailed) >180 >185 >190 "Adult females" 160/251 154/484 140/450 19/63 205.1/205.9 206.07207. 7 207.9/209.2 200.2/204.0 0.476 0.138 0.230 0.201 30 25 • 20 • S 15 H o 10 - Northern Region n = 202 I I i I I M tk 80 100 120 140 160 180 200 220 240 260 Length (cm) 80 70 60 50 40 30 20 10 o Southern Region n = 616 ': m n p-i r-f ^=^ 1 40 160 180 200 Length (cm) Figure 3 Distribution of lengths of striped dolphins, Stella coeruleoalba, measured from the northern and southern regions. (Table 2). With this length sample, it appears that we can expect to detect differences between means that differ by at least 4 cm. Table 2 Minimum detectable diffe rences between means for ^-tests for samples from striped dolphins, Stenella coerureoalba, from the northern (Nor) and Southern (So) regions. Beta error set at 0.10. Minimum Variance detectable Sub-sample (cm) Nor/So lvalue difference (cm) >180 164.99 190.11 1.963 4.01 >185 148.23 162.59 1.964 3.82 >190 122.21 141.94 1.964 3.72 "Adult females" 53.61 147.57 1.292 9.63 8 "Adult Females" - n = 19 — Northern Region 7 - b 5 - 4 - 3 - 2 1 n 140 150 160 180 190 200 Length (cm) 210 220 230 240 "Adult Females" n = 63 - Southern Region 140 150 160 170 180 190 200 210 220 230 240 Length (cm) Figure 4 Distribution of lengths of "adult females", defined here as stroped dolphins, Stenella coeruleoalba, closely associated with a calf, measured from the northern and southern regions. Perryman and Lynn: Stock and school structure of Stenella coeruleoalba 127 The sample from the northern region was too small to test for a seasonal pat- tern in reproduction, but the distribution of back-projected births from the south- ern region differed significantly from the uniform distribution (P<0.01; Figs. 5 and 6). Reproduction for striped dolphins from the southern region appears to be broadly pulsed in the fall through spring period. Photogrammetric and specimen data Since significant differences between length samples from the northern and southern regions could not be detected, we pooled length data from the two re- gions in the tests that follow. We found that "adult females" were significantly longer (4.8 cm) on average than adult females from the specimen data base. When the test was repeated by using length data for lactating females from the specimen data base, the two samples no longer differed significantly (Table 3). Striped dolphins > 180 cm in length from the photogrammetric sample were sig- nificantly longer on average than the sample based on the same length crite- ria from specimen data. We also per- formed a Kolmogorov-Smironov test to compare the two distributions (Fig. 7) and found that they differed signifi- cantly (P<0.01). Northern Region 6 - 5 R . R fl.H, . R fl Southern Region 1 m i i! lis. o =1 < Northern and Southern Region ,fl, .BRTOlfc i Figure 5 Distribution of back-projected birth dates for striped dolphins, Stenella coeruleoalba, from the northern and southern regions and for the two regions combined. Table 3 Results of comparisons between means of length data for striped dolphins, Stenella coerueoalba, taken from specimens (spec) and aerial photo- graphs (photo) (f-tests), and the distribution of lengths >180cm (Kolmogorov-Smirnov \k and si test) from these two sources. n Mean (cml P Comparison spec/photo spec/photo (2-tailed) Adult females specimen/photo 50/82 198.2/203.0 0.007 Lactating specimens/ "adult females" 23/82 199.8.203.0 0.202 > 180cm f-test 256/681 199.19/205.73 0.0001 >180cm h and s 256/681 Z=3.378 0.0007 School size and structure We performed a chi-square test to determine whe- ther the number of dolphins in our two size catego- ries were distributed randomly between schools (Fig. 8) and the hypothesis was significantly rejected when the maximum length for the juvenile category was 195 or 190 cm (P<0.001). With a maximum value of 190 cm, four expected values generated by the test were lower than five. When these schools were deleted from the test or lumped with adjacent schools to eliminate these low expected values, the test results remained highly significant. When school size was regressed against propor- tion in the juvenile category, the slope of the regres- sion was not significantly different from zero. Thus, in our sample, the proportion of small dolphins in a school was not related to school size. 128 Fishery Bulletin 92(1). 1994 100% r 90% 80% 70% \ qj 13 ST 60% \ LL 50% 1 > TO 40% 3 3 30% O 20% 10% - 0% ? Birth Months Figure 6 Cumulative distribution of back-projected striped dolphin, Stenella coeruleoalba, birth dates (solid squares) and those predicted by a uni- form distribution of births (open squares). D □ Length (cm) Figure 7 Length-frequency distributions for specimens of striped dolphin Stenuella coeruleoalba, (> 180 cm) taken incidentally in purse-seine fishing in the eastern tropical Pacific and striped dolphins sampled photogrammetrically that are > 180 cm. Samples from northern and southern regions are combined in this figure. Discussion We found no significant differences in our length samples of striped dolphins from the northern and southern regions to support a recommendation that they be managed as separate stocks. This must be tempered by the fact that length differences of a scale not detectable in our sample, i.e. < 4 cm, could exist. The case for two stocks is also weakened by the distribution of sightings of this species from re- cent research vessel surveys (Wade and Gerrodette, in press). These data indicate that, al- though a hiatus in striped dolphin distribution exists in the typically tropical (high temperature, low salinity) inshore habitat centered around lat. 15° N, there appears to be a broad avenue for movement between the northern and southern regions in the upwelling modified habitat east of long. 110° West (Au and Perryman, 1985; Reilly, 1990). When we compared our sample of lengths for "adult females" and dolphins > 180 cm with data from specimens killed incidentally in purse-seine fishing, we found that the means from the photogram- metric sample were significantly larger (by about 3-6 cm). This does not seem unreasonable at first glance because our measure- ments to the trailing edge of the flukes rather than to the fluke notch introduces a positive bias in the photogrammetric data of about 2-2.5 cm. Also, the "adult female" category probably in- cludes only those females who have carried and given birth to a live calf, thus eliminating the younger, presumably smaller, fe- males who are physiologically adult but have not yet had a suc- cessful pregnancy. However, these results for adult females are con- trary to previous comparisons of photographic and specimen data for northern and central common dolphins (Perryman and Lynn, 1993) and eastern spinner dol- phins (Perryman, unpubl. data). Since the photogrammetric data for all of these taxa were collected in the same manner, it seems likely that the difference between the two striped dolphin samples reflects some form of selectivity in either or both sampling systems. The schools of striped dolphins that we photo- graphed showed a pattern of segregation by length Perryman and Lynn: Stock and school structure of Stenella coeruleoalba 129 > O z LU o in rr 35 30 25 20 15 10 5 0 35 30 25 20 15 ■ 10 5 0 35 30 25 20 15 10 5 0 35 30 25 20 15 10 5 0 35 JO 25 20 15 - 10 5 0 35 30 25 20 15 10 5 0 35 30 25 20 ' 15 10 5 0 SCHOOL 1 School Size - 79 No. Measured - 26 J3- M, " i XL SCHOOL 2 School Size - 222 No. Measured = 46 -E3- SCHOOL 3 School Size - 73 No Measured - 30 n SCHOOL 4 School Size - 1 73 No Measured - 16 1 l , P. . -^ SCHOOL 5 School Size - 87 No. Measured = 33 rrm n i IT SCHOOL 6 School Size = 1 S1 No. Measured ■= 42 m* SCHOOL 7 School Size - 25 No Measured = 9 1 " n 100 120 140 160 18 200 220 240 SCHOOL 8 School Size - 56 No Measured - 29 SCHOOL 9 School Size - 86 No. Measured - 54 .rvr^Vt Length cated le LENGTH (CM) Figure frequencies for each school of striped dolphi ngths of dolphins that were included in the SCHOOL 10 School Size = 23 No Measured = 1 7 SCHOOL 11 School Size = 100 No. Measured - 10 SCHOOL 12 School Size - 54 No Measured = 29 la SCHOOL 14 School Size - 46 No Measured « 30 ra ,n, Jfl W SCHOOL 13 School Size- 124 fh .*.m im» & m_ 100 120 140 180 200 220 240 LENGTH (CM) 8 ns, Stenella coeruleoalba. Shaded bars indi juvenile categeory. that is very similar to that reported from the west- ern Pacific (Miyazaki, 1977; Miyazaki and Nishiwaki, 1978). It also appears that the propor- tion of smaller dolphins in our sample of schools is not related to school size. Possibly this segregation is the explanation for differences between specimen and photogrammetric data sets. Tuna fishermen select dolphin schools for encircle- ment based mainly on the amount of tuna associ- ated with the school. Schools of younger/smaller 130 Fishery Bulletin 92(1), 1994 > o z LU ZD o UJ rr LL 35 30 25 20 15 10 5 0 15 30 25 20 15 10 5 0 35 30 25 20 15 10 5 0 35 30 25 20 15 10 5 0 35 30 25 20 15 10 5 0 35 30 25 20 15 10 5 0 35 30 25 20 15 10 5 0 SCHOOL 15 School Size- 125 No Measured ■ 60 SCHOOL 16 School Size - 59 No. Measured = 1 4 ll ■ .■■ I'll SCHOOL 17 School Size. 100 No Measured. 51 P I (71 fT-n MyPI^H 1 In SCHOOL 18 School Size - 95 No Measured - 25 ■itti 1 SCHOOL 19 School Size - 88 No Measuied - 55 it SCHOOL 20 School Size = 10 No Measured • 5 SCHOOL 21 School Size = 30 No Measured - 9 140 160 180 200 220 240 LENGTH (CM) 35 30 25 20 15 10 5 0 35 30 25 20 15 10 5 0 35 30 25 20 15 10 5 0 35 30 25 20 15 10 5 0 35 JO 25 20 15 10 5 0 35 30 25 20 15 10 5 0 35 30 25 20 15 10 5 0 SCHOOL 22 School Size - 40 No Measured - 10 SCHOOL 23 School Size- 175 No Measured - 66 t>.cwx SCHOOL 24 School Size - 58 No Measured - 16 Jl l j SCHOOL 25 School Size - 76 R No. Measured = 36 mi m~ ~ h SCHOOL 26 School Size - 48 No Measured - 30 XI J3L im SCHOOL 27 School Size - 23 No Measured . 9 S i HI SCHOOL 28 School Size - 44 No Measured = 21 JGL 100 120 140 160 180 200 220 240 LENGTH (CM) Figure 8 (Continued) striped dolphins might carry more tuna and be cap- tured more frequently than schools composed of adult animals. If the bond between yellowfin tuna and dolphins is related to size and hydrodynamics as suggested by Edwards i 1992) then it may be that the smaller striped dolphins are hydrodynamically more suitable for this association. Juvenile schools of striped dolphins are made up of animals that are about the same length as schools of spotted or spin- ner dolphins for which the tuna-dolphin association appears to be the strongest. Perryman and Lynn: Stock and school structure of Stenella coeruleoalba 131 Acknowledgments A. E. Dizon, D. P. DeMaster, W. F. Perrin, and two anonymous reviewers read the manuscript and pro- vided very useful suggestions. Valuable assistance and specimen data were provided by S. Chivers. Sev- eral of the photographs for this analysis were taken by J. Gilpatrick and R. Westlake. This work would have not been possible without the field support of the Officers and Crew of the NOAA Ship David Starr Jordan and the pilots and mechanics of NOAA's Aircraft Operations Center. Literature cited Au, D. W. K., and W. L. Perryman. 1985. Dolphin habitats in the eastern tropical Pacific. Fish. Bull. 83:623-643. Batschelet, E. 1965. Statistical methods for the analysis of prob- lems in animal orientation and certain biological rhythms. Am. Inst. Biol. Sci. Monograph, 57 p. Barlow, J. 1984. Reproductive seasonality in pelagic dolphins (Stenella spp.): implications for measuring rates. Rep. Int. Whaling Comm. Spec. Issue 6:191-198. DeMaster, D. P., E. F. Edwards, P. Wade, and J. E. Sisson. 1992. Status of dolphin stocks in the eastern tropi- cal Pacific. In D. R. McCullough and R. H. Barrett (eds.). Wildlife 2001: populations, p. 1038-1050. Elsevier Science Pubis., New York. Dizon, A. E., W. F. Perrin, and P. A. Akin. In press. Stocks of dolphins (Stenella spp. and Del- phinus delphis) in the eastern tropical Pacific: a phylogeographic classification. NOAA Tech. Rep. Edwards, E. F. 1992. Energetics of associated tunas and dolphins in the eastern tropical Pacific Ocean: a basis for the bond. Fish. Bull. 90:678-690. Holt, R. S., and J. E. Powers. 1982. Abundance estimation of dolphin stocks in the eastern tropical Pacific yellowfin tuna fishery deter- mined from aerial and ship surveys to 1979. NOAA Tech. Mem. NMFS SWFSC 23, 95 p. Holt, R. S., and S. N. Sexton. 1989. 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. Fish. Bull. 88:105-111. Miyazaki, N. 1977. School structure of Stenella coeruleoalba. Rep. Int. Whaling Comm. 27:498- 499. 1984. Further analysis of reproduction in the striped dolphin, Stenella coerueoalba, off the Pa- cific coast of Japan. Rep. Int. Whaling Comm. Spec. Issue 6:343-353. Miyazaki, N., and M. Nishiwaki. 1978. School structure of the striped dolphin off the Pacific Coast of Japan. Scientific Reports of the Whales Research Institute. 30:65-115. Norris K. S. (ed.) 1961. Standardized methods for measuring and re- cording data on smaller cetaceans. J. Mammal. 42:471-476. Perrin, W. F., J. M. Coe, and J. R. Zweifel. 1976. Growth and reproduction of the spotted por- poise, Stenella attenuata, in the offshore eastern tropical Pacific. Fish. Bull. 74:229-269. Perrin, W. F., M. D. Scott, G. J. Walker, and V. L. Cass. 1985. Review of the geographical stocks of tropical dolphins (Stenella spp. and Delphinus delphis) in the eastern Pacific. NOAA Tech. Rep. NMFS 28. 28 p. Perryman, W. L, and M. Lynn. 1993. Identification of geographic forms of common dolphin (Delphinus delphis) from aerial photogrammetry. Mar. Mammal Sci. 9:119-137. Reilly, S. B. 1990. Seasonal changes in distribution and habitat differences among dolphins in the eastern tropi- cal Pacific. Mar. Ecol. Prog. Ser. 66:1-11. Scott, M. D., and W. L. Perryman. 1991. Using aerial photogrammetry to study dolphin school structure. In K. Pryor and K. S. Norris, (eds.), Dolphin societies-discoveries and puzzles, p. 227-241. Univ. California Press, Berkeley. Wade, P. R., and T. Gerrodette. In press. Estimates of cetacean abundance and distribution in the eastern tropical Pacific. Rep. Int. Whaling Comm. 43. Abstract. — The eastern Pa- cific purse-seine tuna fishery has historically been very productive, yielding up to 400,000 metric tons (t) per year of primarily yellowfin, Thunnus albacares, and skipjack, Katsuwonus pelamis. However, ef- forts to minimize dolphin (prima- rily spotted dolphin, Stenella attenuata, spinner dolphin, S. longirostris, and common dolphin, Delphinus delphis) mortality inci- dental to tuna seining in the east- ern Pacific ocean have been in- creasing. Therefore, predictions of what the tuna catches will be in the future, if there is a ban or moratorium on catching dolphin- associated tuna, are useful. Based on recruitment levels, age-specific catchability coefficients for yellow- fin tuna caught without dolphins, and average fishing effort ob- served during 1980-88, we pre- dicted that yellowfin catches would be reduced by an average of about 25%. These results were verified by Monte Carlo simula- tions, by using average effort and randomly selected yellowfin re- cruitment and catchability coeffi- cients from 1980 to 1988, which predicted a mean annual decrease of 55,563 t or 24.7% of yellowfin catch. The actual reduction in yel- lowfin catch might be greater be- cause 1) fishing effort will prob- ably decline, 2) the range of the fishery might be reduced to the traditional inshore non-dolphin regions, and 3) yellowfin recruit- ment could be reduced by the change in age structure and popu- lation size likely to result from a moratorium. Because skipjack sel- dom associate with dolphins, redi- rection of fishing effort to schools of tuna not associated with dol- phins would probably result in in- creased skipjack catch rates. How- ever, the magnitude of the in- crease is difficult to estimate, be- cause the population dynamics of skipjack are poorly understood. Finally, this study predicted that the catches in the first years after a moratorium on dolphin sets would not necessarily reflect long- term catches. Potential tuna catches in the eastern Pacific Ocean from schools not associated with dolphins Richard G. Punsly Patrick K. Tomlinson Ashley J. Mullen Inter-American Tropical Tuna Commission 8604 La Jolla Shores Dr. La Jolla. CA 92037 Manuscript accepted 22 July 1993 Fishery Bulletin 92:132-143 (1994) Since the late 1950's, purse-seine fishermen in the eastern Pacific Ocean (EPO), knowing that schools of yellowfin tuna (Thunnus alba- cares) often associate with dolphins (primarily spotted dolphins, Sten- ella attenuata, spinner dolphins, S. longirostris, and common dolphins, Delphinus delphis), have used the dolphins to help locate and capture yellowfin. Dolphins are relatively easy to detect, being larger and closer to the surface than yellowfin. In fact, the most efficient means of catching the 2- and 3-year-old yel- lowfin, which comprise the largest component of the tuna catch in the EPO, is purse-seine fishing for dol- phin associated schools (Punsly and Deriso, 1991). Yellowfin remain associated with dolphins while the net is being set around the dolphin herds. The fishermen attempt to release all of the dolphins from the net; however, incidental mortality sometimes occurs through entang- lement. As a result of increasing public pressure to prevent mortality of dolphins incidental to tuna purse seining, elimination of setting on dolphin-associated tunas is being considered. Therefore, fishermen, biologists, and managers need to know the extent to which tuna catch in the EPO might be reduced by the elimination of sets on dol- phin-associated fish. The objective of this study was to estimate this potential reduction in the catch. No such estimates have been pub- lished previously. Tuna catches could be affected by a ban or moratorium on dolphin sets in six ways: 1 The overall catchability of yel- lowfin by purse seiners could be reduced. 2 The yield per recruit of yellow- fin could decline because non- dolphin-associated yellowfin caught by purse seiners are mostly composed of fish younger than the optimum age of entry (Calkins, 1965; Allen, 1981). 3 The average age of yellowfin and mean biomass may be reduced by fishing on younger age groups. This might not only re- duce the catch in weight, but also reduce the spawning poten- tial and possibly the resulting recruitment. 4 Since the offshore EPO purse- seine fishery is directed prima- rily at dolphin-associated fish (Fig. 1, A and B), a moratorium on setting on dolphin herds could result in a contraction of the range of the fishery into in- shore regions. The number of fish recruited to this new smaller area might be lower than the number recruited to the entire area. Lower effective re- cruitment would also result in lower catches. 5 If a moratorium on catching dol- phin-associated tuna occurs, 132 Punsly et al.: Potential non-dolphin-associated tuna catches in the eastern Pacific Ocean 133 some purse-seine fishermen may decide to move to other oceans or retire, which would reduce total fishing effort and hence the catch. 6 Since skipjack tuna (Katsuwo- nus pelamis), the only other primary target species in the fishery, seldom associate with dolphins, their catch may in- crease if effort remains at 1980-88 levels and is directed only toward tuna schools not associated with dolphins. Because no relation between spawners and recruitment of yel- lowfin has been established (Bayliff, 1992, p. 62), the possible effects of reduced recruitment were not addressed in this study. Also, since the authors cannot predict how many seiners would leave the EPO, or how much the fishery would contract, these two factors were not considered. In other words, this study only at- tempted to estimate how much tuna catches might change due to changes in yellowfin catchability, yield per recruit, total biomass, and age structure. To measure the possible effects of changing the mode of fishing from being directed toward prima- rily dolphin-associated schools of tuna ("dolphin sets," Allen, 1981) to one directed at exclusively free- swimming schools ("school sets") and floating-object-associated schools ("log sets," Greenblatt, 1979), we first estimated what the tuna catches would have been in previous years if dolphin sets had been replaced by non-dolphin sets. Then the estimates were compared with actual catches. Our method used non-dolphin-set catchability coefficients and total effort to estimate what the catches would have been during 1980-88 if there had been a mora- torium on dolphin sets beginning in 1980. Other works in which catches were estimated for alter- Figure 1 (A) Geographic distribution of average yellowfin tuna (Thunnus albacares) catch by purse seiners, during 1980-88, from schools associ- ated with dolphins (Delphinidae). Catches are expressed in metric tons by 2.5-degree quadrangles. (B) Geographic distribution of average yel- lowfin catch by purse seiners, during 1980-88, from schools not associ- ated with dolphins. Catches are expressed in metric tons by 2.5-degree quadrangles. 134 Fishery Bulletin 92(1), 1994 1 -10 Boat-Days CD 11 -50 Boat-Days J\ 51 - 100 Boat-Days 101 -200 Boat-Days 201 - 500 Boat-Days Greater than 500 Boat-Days Figure 2 (A) Geographic distribution of total purse-seiner fishing effort during 1980-88 which lead to dolphin (Delphinidae) sets. Effort levels are ex- pressed in boat-days of fishing by 2.5-degree quadrangles. (B) Geo- graphic distribution of total purse-seiner fishing effort during 1980-88 which lead to non-dolphin sets. Effort levels are expressed in boat-days of fishing by 2.5-degree quadrangles. native catchability coefficients include Holt ( 1958), Jones ( 196 1 ), and Bartoo and Coan (1978). Materials and methods Data The Inter-American Tropical Tuna Commission's (IATTC) logbook and length-frequency data bases were used in this study. The log- book data base, described in Or- ange and Calkins (1981), Punsly (1983; with emphasis on set types), and Punsly (1987; with emphasis on yellowfin catch rates), contains information on the fishing activities of about 90^ of the purse seiners in the EPO. Total catches were estimated by multiplying the logbook catches by the ratio of the sum of the un- loading weights to the sum of the logbook catches. Geographic dis- tributions of the logbook data on catch and effort, during 1980-88, for both dolphin-associated and unassociated schools are shown in Figures 1 and 2. The length-fre- quency data base, described by Hennemuth (1957). Punsly and Deriso (1991), and Tomlinson et al. (1992), has information from samples of about 12-15^ of the catch. Age-specific yellowfin abun- dances from cohort analysis (Pope, 1972; also called sequential computation of stock size in Ricker, 1975; and virtual popula- tion analysis in Gulland. 1965) were taken from Bayliff ( 1990). Data from 1980 to 19S8 were used in this study. Data before 1980 were not used because of the difficulty in modeling the closed seasons for yellowfin (Cole, 1980). Data after 1988 were not used be- cause cohort analysis cannot pro- duce accurate abundance esti- mates for cohorts which have not been in the fishery for a sufficient period of time. Semi-annual age groups used in this study were described in detail Punsly et al.: Potential non-dolphm-associated tuna catches in the eastern Pacific Ocean 135 in Bayliff (1992, p. 52). Monthly age compositions were estimated by combining 1-cm length-interval data into semi-annual age groups by fitting multinormal distributions to the data with the aid of the computer program NORMSEP, (Abramson, 1971), and constraining the fit to the growth param- eters of Wild (1986). "X" and "Y" cohorts were de- fined as those fish reaching 30 cm, which correspond to the approximate age of first recruitment, during the fourth and second quarters of the year, respec- tively. Age groups in our study, 0.5 to 5.5 in 0.5 year increments, correspond to the Y0, XI, Yl ... Y5 co- horts, respectively, in Table 21 of Bayliff (1992). Estimates of fishing effort The total monthly effort by purse seiners was esti- mated as E = f Y l\ om J om om I .' om ' where o, refers to the observed mixture of set types, Y is the yellowfin catch unloaded by purse sein- ers in month (m),yom is the yellowfin catch reported in the IATTC logbooks and f is the effort, in boat- days of fishing, reported in the logbooks. Effort on non- dolphin sets for all purse seiners was estimated by ^nm / , / .lnmcs^omcs/y v„ where fnmcs is the fishing effort which lead to non- dolphin (n) sets by monitored vessels of size (s) from country (c), Yomcs is the total catch of yellowfin from unloadings by size (s) vessels from country (c), and y is the total yellowfin catch by monitored ves- J omcs J "* sels. These estimates were stratified by country and size of vessel because the proportion of dolphin sets is affected by these two factors. Estimation of yellowfin catches if all effort were non-dolphin This method used age-specific, monthly catchability coefficients by fishing mode and allowed the future population structure to be affected by previous catches. First, age-specific catchability coefficients for non- dolphin sets in) in each month (m) were estimated for each semi-annual age group (/'): Qnmj ~ ^nmj y^nm^ mj J i where C are the monthly, total, non-dolphin purse-seine catches (in numbers of fish) of semi- annual age group (j) and Nmj are the age-specific, monthly, average abundances estimated by the co- hort analysis (Bayliff, 1990). Beginning with the population structure in January 1980, obtained from cohort analysis, we estimated what the catch in each month of each semi-annual age group would have been without dolphin sets; i.e., pmj (NmjqnmjEom )/(qnm]Eom + Mj ) where Mj is the age-specific, instantaneous, monthly natural mortality (Bayliff, 1992, p. 52). Yield in weight was estimated by Y =W (i)C , 1 pmj m yJ ' pmj ' where W(j) is the estimated mean weight of age (j) yellowfin in month m caught during 1980-88. The subsequent month's abundance of semi-annual age group (j) was estimated to be Estimates of skipjack catches if all effort were non-dolphin Skipjack are suspected to be mostly transient in the EPO (Joseph and Calkins, 1969), so we assumed that depletion is probably unimportant. Thus, the ratio of the total effort to the non-dolphin effort was used to estimate skipjack catches: Ypm(SJ): Ynm(SJ)Eom/Enn where Y m(SJ) is the potential (p) non-dolphin, skip- jack catch and Ynm(SJ) is the actual non-dolphin-set, skipjack catch. In essence, skipjack catches were estimated to be linear extrapolations of catch rates to higher levels of effort. £*m+Xj mj -iqnm.Eom+M,) except for the months of recruitment (May and January), when NJAN2 and NMAy,3 were set eQual to the historical recruitment previously estimated for that time period by cohort analysis. Yellowfin form the first semi-annual age group (those fish hatched in the middle of the current year) were not included in the analysis because they were not recruited until the next year, when they became semi-annual age group 3. Each January, the semi-annual age groups were graduated as follows: NJANJ+2=N DEC. J iaDEi ' ■ +M, 136 Fishery Bulletin 92(1). 1994 Monte Carlo simulation The age-structure method produced catches specific to the observed time-series of recruitment and age- specific catchability coefficients during 1980- 88. Additional information can be gained by estimating what the trend in catches would be if the recruit- ment and catchability trends were different. In or- der to explore the range of resulting catches which might have occurred under various conditions, a Monte Carlo simulation was used. Paired simula- tions were performed for both the observed mixed- mode fishery and a fishery in which all effort was directed toward non-dolphin-associated tuna. Fre- quency distributions of differences between catches from the two simulated fisheries provide a more comprehensive estimate of future expectations. The simulations used quarterly time steps and 1,000 replicates. At each quarter of each year in each replicate, a year between 1980 and 1988 was ran- domly selected with replacement (i.e., each year could be selected more than once). Pairs of quarterly catchability coefficients (one from the observed mix- ture of fishing modes and one for the non-dolphin sets only) estimated for the corresponding year, were used in the calculations during the time steps. Quar- terly coefficients were calculated with the same equation as that for the monthly coefficients with months replaced by quarters. Quarterly fishing ef- forts were set to the 1980-88 averages. The same average total effort was applied to both the observed and non-dolphin fishing-mode models. Recruitment was simulated to occur in the second and fourth quarter. For each year in each simula- tion, a randomly selected year was chosen. Recruit- ment pairs (X and Y) from the randomly selected year were used for both fishery models. Initial popu- lation sizes and age structures were also set to the 1980-88 averages. One thousand differences between the simulated catches for the mixed- mode and non-dolphin only scenarios were generated for a time series of nine years. The 95% confidence intervals corresponded to the 50th and 950th highest differences from the 1,000 simulations. Because yellowfin usually live for less than 5 years (Fig. 3), results for the last (9th) year were unaffected by the initial age structure. Results Deterministic approach If trends in total effort, recruitment, and non-dol- phin-set catchability coefficients had been the same as during 1980-88, with all effort directed at non- dolphin sets, yellowfin catches (Table 1, column Table 1 Estimated annual tuna (Scombridae) catches by purse seiners in the eastern Pacific ocean, in thousands of metric tons. Year OYF NYF QYF OSJ NSJ OT QT NT 1980 170 129 158 131 155 301 313 284 1981 190 152 146 120 151 310 297 303 1982 134 111 120 99 129 233 249 240 1983 104 96 98 58 73 162 171 169 1984 155 103 125 HI 90 215 215 193 1985 227 132 169 49 99 276 268 231 1986 286 193 168 64 113 350 281 305 1987 285 243 195 62 120 347 314 363 1988 303 266 229 85 123 388 352 389 Mean 206 158 156 81 117 286 274 275 OYF NYF yellowfin tuna iThunnus albacares) - observed mixture of set types. yellowfin tuna - all effort directed at non-dolphin ( Delphimdae) sets, using the observed monthly catchability coefficients for non-dolphin sets. QYF = yellowfin tuna - all effort directed at non-dolphin sets, using the average, observed, quarterly catchability co- efficients for non-dolphin sets. OSJ = skipjack tuna {Katsuwonus pelamis) - observed mixture of set types. skipjack tuna - all effort directed at non-dolphin sets, yellowfin plus skipjack tuna - observed mixture of set types. yellowfin plus skipjack tuna - effort directed at non-dol- phin sets, using quarterly average catchability coeffi- cients. yellowfin plus skipjack tuna - all effort directed at non- dolphin sets, using monthly catchability coefficients. NSJ = OT = QT = NT NYF) were estimated to have averaged 77% of the observed catch (Table 1, column OYF). The range was from 58% in 1985 when dolphin-associated tuna fishing was good to 93% in 1983 when dolphin-as- sociated tuna fishing was poor. The reasons why the ratio of estimated catch without dolphin sets to the observed catch varied annually can be seen in Fig- ures 3-7. For example, the high estimated bio- masses of 1.5-year-old yellowfin in 1988 (Fig. 4), coupled with their high non-dolphin-set catchability coefficients (Fig. 5), produced an estimated catch of 266,000 t for all effort directed at non-dolphin sets, which was almost as high as the 303,000 t catch estimated from the catchability coefficients for the observed mixture of set types (Fig. 6). Catchabilities could have increased in 1988 for a variety of reasons, including the use of deeper nets, the use of "bird radar" (relatively new radar used for detecting birds which commonly have tuna beneath them) or envi- ronmental factors, such as a shoaling of the ther- mocline (Green, 1967). For a given level of effort, catches depended on the age-specific abundances (Figs. 3 and 4) and catchability coefficients (Figs. 5 and 6). Consequently, the estimated catches if all effort were directed at non-dolphin sets approached Punsly et al.: Potential non-dolphin-associated tuna catches in the eastern Pacific Ocean 137 1 - y Vi < Z 21 1 — i — r - 1 — i — i — i t i i i '■rff Rtk ■ \aL ~K_ ; '■■All TK : :-^z TK '■ '■ H~ Ht^ r-| -n-n-^ : -r^TTTTT^_ 1982 1981 1980 Figure 3 Estimated average annual biomasses (t) of yellowfin tuna (Thunnus albacares) by semi-annual age group for the observed mixture of set types. In the left panel, biomasses are summarized by age within year. In the right panel, biomasses are summarized by year within age group. Age refers to the age in years at the middle of the year. -nSl - ■ V) -T. < z rrrihTh-w JH r-TTfl ITr-^- r^TTTl^. 1988 1986 1984 1983 1982 1981 1980 VI < Z 3 -1 1 T 1 1- :rrr^r4~m :rr>^rn HI r^m nrrrrrj I I I I 1 I I - AGE 5.5 AGE 5 . C AGE 4 . 5 AGE 4 . C AGE 3.5 AGE 1 . 0 AGE 0.5 Figure 4 Estimated average annual biomasses (t) of yellowfin tuna (Thunnus albacares) by semi-annual age group for all effort directed at non-dolphin (Delphinidae) sets. In the left panel, biomasses are summarized by age within year. In the right panel, biomasses are summarized by year within age group. Age refers to the age in years at the middle of the year. Fishery Bulletin 92(1), 1994 J 0. JTw-lhTr - 1987 ^-TTH rTfTH^: r^Vn-rr-ITln ^TH^rrT~r-^l ' rT^-r^-fT^l 1980 _□=. rm-vj rThmT> TTT 1:JJlHJJ1l. rT-mll-r n,^ .□L AGE 5 . 5 AGE 5 . 0 AGE 4 . 5 AGE 4 . 0 AGE 3 . 5 Figure 5 Average annual non-dolphin-set yellowfin tuna (Thunnus albacares) catchability coefficients (q in boat-days-^) by semi-annual age group. In the left panel, coefficients are summarized by age within year. In the right panel, coefficients are summarized by year within age group. Annual catchability coefficients are estimated as the mean of the monthly coefficients. Age re- fers to the age in years at the middle of the year. =r£ ^^fn-u -mW^ ■j-ri~H—i 1983 1982 1981 1980 JZL oiEt£^d rffl FT-fTM I rh : -■■i ■ ''I ' I AGE 5 . 5 AGE 5 . C AGE 3 . 0 AGE 2 . 5 AGE 2 . 0 AGE 1 . 5 AGE 1 . 0 AGE 0 . 5 Figure 6 Average annual observed yellowfin tuna (Thunnus albacares) catchability coef- ficients ( rfff Ti ■ n i ^-T-1 — ' — ' h - >•— 1 — i — i ' l ' l ' l ' J l Ctbi . 1988 1987 1986 1985 1984 1983 1982 1981 1980 — i — i — i — i — r- IZI B- I ! T I n^Tr-r^ o. Oil □£l TI - AGE 4 . 0 r^ OZL TTT-rL J~ r— r ~1 1 1 I~~I 1 1 ! T- AGE 5.5 AGE 5.0 AGE 3.5 AGE 3 . 0 AGE 2.5 AGE 2 . 0 AGE 1 . 5 AGE 1.0 AGE 0 . 5 Figure 7 Average annual differences between the observed and non-dolphin (Delphinidae) catchability coefficients (boat-days-1). In the left panel, differences are summa- rized by age within year. In the right panel, differences are summarized by year within age group. Age refers to the age in years at the middle of the year. Nega- tive values (those pointing down) indicate that those non-dolphin-set catchability coefficients were greater than the observed coefficients. the observed levels when the non-dolphin-set catchability coefficients were greater than or equal to the observed overall catchability coefficients (Fig. 7, negative values) for the age groups of the greatest biomass (Figs. 3 and 4). Estimated total yel- lowfin plus skipjack catches, if all effort were di- rected at non-dolphin sets, ranged from 84% during 1985 to 104% in 1983. Estimates (Table 1, column QYF) of what the catches would have been without dolphin sets, us- ing the quarterly average (over years) non-dolphin- set catchability coefficients for 1980-88, indicate that yellowfin catchabilities on non-dolphin sets increased in the late 1980's. Average quarterly catchability coefficients produced noticeably higher catches than the observed non-dolphin-set monthly coefficients in 1983-85 when the observed coeffi- cients on small fish were low. On the other hand, average quarterly catchability coefficients produced lower catches during 1986-88, when the observed non-dolphin-set coefficients were high. Monte Carlo simulation The Monte Carlo simulations (Table 2) predicted that, if total effort, recruitment, and non-dolphin-set catchability coefficients had varied randomly throughout their 1980-88 distributions, and current levels of effort and recruitment had been main- tained, changing to a fishery with all effort directed toward non-dolphin sets would have resulted in an average reduction of 55,563 t (24.7%) of yellowfin catch per year. The 95% confidence interval, based on the 50th and 950th highest simulated differences was 24,000 to 91,000 t (10%-42%). The entire fre- quency distribution of the differences between the two fishing-mode models in the 9th year is shown in Figure 8. Simulated recruitment estimates were selected from the observed values during 1980-88. Thus, average recruitment used in the simulations was higher than the mean actual recruitment to the initial 1980 population structure, which was partly a result of the poor recruitment during 1978 and 1979. Consequently, simulated catches were higher for both the observed mixed mode fishery (229,000 t per year) and the non-dolphin-set only fishery (175,000 t). Yield per recruit Estimated yellowfin catches from both the determin- istic approach (Table 1) and the Monte Carlo simu- 140 Fishery Bulletin 92(1), 1994 ■j. a. Oj IX o o o r-i E c >> u c OJ 3 CT U L- 250 200 150 100 50 50 60 70 80 90 100 110 Yellowfin Catch Retained (%) Figure 8 Frequency distribution of the percent of the yel- lowfin tuna (Thunnus albacares) catch in weight retained in the ninth year of a ban on dolphin (Delphinidae) sets, from 1,000 Monte Carlo simu- lated replicates. The percent of catch retained is calculated as 100x (the catch if all effort were directed at non-dolphin sets) / (the catch from the observed mixture of set types). Table 2 Monte-Carlo simulated annual yellowfin tu na (Th / n n u s albacares) catch es, in thousands of tons, from 1980- -88, quarter y, average catchability coefficien ts. YEAR OYFM OYFL OYFU NYFM NYFL NYFU PCM PCL PCU 1 202 184 229 150 121 186 72 ill S3 2 215 183 247 164 127 205 76 62 89 3 227 183 276 170 129 217 76 59 91 4 229 183 283 175 131 223 71) 68 ss 5 236 188 2H6 181 139 230 76 r,!i 91 6 245 195 302 187 138 241 76 ill 91 7 237 ISO 294 180 138 229 75 59 90 8 231 182 281 176 134 228 76 58 91 9 229 1S1 279 174 131 222 75 58 90 OYFM = mean veilowfin tuna catch *or the observed mix ;ure of se t tvpes. OYFL = OYFM Iowei 959! confidence interval. OYFU = OYFM upper 959! confidence interval NYFM = mean yellowfin tuna catch using total effort and non-dolphin catchability coeffi- cients. NYFL = NYFM lower 95% confidence interval NYFU = NYFM upper 959c confidence interval PCM = mean percent of c atch retained 1 100 x NYFM/OYF: PCX = PCM 1 awer 95% confidence interval. PCU = PCM l pper 95% confidence interval. lations (Table 2) were heavily influenced by the re- cruitment and fishing effort levels used. Recruit- ment in the future may be different from that of past, because of changes in population size, age structure, and environmental factors. Therefore, actual future catches could be different from what we estimated. For these reasons, results in terms of reduction in yield per recruit are of interest. We estimated that the change to non-dolphin sets only would result in the reduction of the yield per recruit of yellowfin from the observed value of 2.8 kg per recruit to 2.1 kg as shown in Figure 9. In addition, effort levels could change in the future, perhaps as a reaction to the moratorium. Therefore, estimates of yield per recruit for various levels of effort might be useful. If effort levels change in the future, the multipliers on the X-axis in Fig. 9 could be used to estimate the potential yellowfin catch. Discussion In order to predict what the tuna catches might be in the future if there were a moratorium on dolphin sets, we estimated what the tuna catches would have been during 1980-88, had there been a mora- torium on dolphin sets beginning in 1980. Using these estimates to predict future catches required the following assumptions: 1 Age-specific, non-dolphin catchability coefficients will be the same in the future as during 1980-88. 2 Fishing effort will remain at 1980-88 levels. 3 The geographic distribution of effort will be the same as during 1980-1988 (Fig. 2, A and B combined). 4 Recruitment will be at 1980-88 levels. 5 Natural mortality will not change in the future. 6 Skipjack abundance will not significantly change. Significant deviations from these assumptions could make our estimates less valid. There- fore, the potential ramifications of deviations from the assump- tions are discussed in detail below. Major changes in the vulner- ability of non-dolphin-associ- Punsly et al.: Potential non-dolphin-associated tuna catches in the eastern Pacific Ocean 141 *: 8 OC 1.5 S a. ■o O 1.0 ated yellowfin to purse seiners could result in significantly different catches than we estimated. Allen and Punsly (1984) showed that both environmental and vessel efficiency factors affect the catchability of yellowfin by purse sein- ers in the EPO. Improvements in vessel efficiency could increase future catchability coefficients; whereas, envi- ronmental factors could produce either higher or lower catchability coefficients than those observed during 1980-88. Environmental factors affecting catchability could conceivably mask the effects of a moratorium on dolphin sets for several years. For example, if a moratorium on dolphin sets had been imposed at the beginning of 1983, the low catch in 1983 would have made it appear that the decline resulted from the moratorium. However, we predicted that a moratorium would have had the smallest effect in 1983 (Table 1). Fish- ermen, biologists, and managers should be aware that catches during the first year after a moratorium starts may not be indicative of long- term averages. However, since 9 years of data were used, our long-term average estimates should only be affected by long-term changes in catchability. An assumption that effort will be lower in the future may be more realistic than our assumption that effort will remain at 1980-88 levels. However, we could not predict the extent to which effort might be reduced because it is affected by ex-vessel tuna prices at canneries all over the world, the prices of other foods, and the cost of fuel. Nevertheless, if we could estimate what the effort reductions would be in the future, the effort multipliers in the the yield- per-recruit estimates in Fig. 9 could still be used. If the fishery contracted into the traditional in- shore school- and log-set areas after a moratorium on dolphin sets, then catches may be lower than we estimated them to be. For example, if the area fished were smaller, and mixing between the fish inside and outside the area were incomplete, then the new fishing area would encompass fewer fish than the total area. Therefore, all of the population sizes of yellowfin used in the equations in the methods sec- tion would be overestimated. Recruitment estimates, which are estimates of the number of 30-cm yellow- fin, would also be overestimated. In addition, if fish- ing effort remained high, but the range contracted, then a gear- competition effect might lower the catch of both yellowfin and skipjack. However, since effort levels are expected to decline after a moratorium, j I l I I I i Observed Mixture of Fishing Modes No Dolphin Sets 1 I I I L 0 0.2 0.4 O.e 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 Effort Multiplier Figure 9 Yield per recruit of yellowfin tuna (Thunnus albacares) (kg) for the observed 1980-88 fishery (solid line) and a fishery with all effort directed at non-dolphin (Delphinidae) sets (dashed line). An effort multiplier of 1.0 refers to the 1980-88 average effort. localized depletion of tuna due to a contracted fish- ery is unlikely. We assumed that yellowfin recruitment in the future would not be affected by the changes in popu- lation size and age structure which might result from re-directing effort toward smaller fish, because a relationship between yellowfin spawning biomass and recruitment has not yet been demonstrated. However, a spawner-recruit relationship for yellow- fin may be discovered in the future, because better estimates of yellowfin fecundity by size offish, sea- son, and area are currently being developed at IATTC. When this work is completed we may be able to predict recruitment levels and their resulting catches more accurately in the future. If future re- cruitment levels could be estimated, the future catches could be derived by multiplying the recruitment esti- mates by the yield per recruit shown in Figure 9. Environmental factors have long been suspected of having significant effects on yellowfin recruit- ment. For example, favorable conditions in the late 1980's may have contributed to the large number of recruits (Bayliff, 1992). In 1987, the number of re- cruits was so large that the effect of a moratorium in 1988 would have been masked by a high catch of 1.5 year old yellowfin, first recruited during 1987. In 1988, the high abundance of 1.5 year old fish (Fig. 4) coupled with their high catchability for non-dol- phin sets (Fig. 5) caused the estimated yellowfin catch if all effort were directed at non-dolphin sets to be almost as high as the estimated actual catch. 142 Fishery Bulletin 92(1). 1994 In order to predict future recruitment, the IATTC is currently studying the relationship between the environment and yellowfin recruitment. If they are successful the yield-per-recruit estimates in Figure 9 could be multiplied by the recruitment estimates to better predict future yellowfin catches. Little is known about the rate of natural mortal- ity of yellowfin. However, there is no reason to be- lieve this rate will change. But, if it does change, a reasonable assumption would be that if natural mortality goes up, catch will go down and vice versa. Little is known about skipjack population dynam- ics. We assumed that local depletion is negligible for skipjack. However, since skipjack are primarily caught in association with floating objects, if the amount of effort per floating object increases as a result of effort being re-directed from dolphin-asso- ciated tunas to floating objects, then the chances of depletion is certainly possible. If this occurs, our estimates of skipjack catch rates will be too high. This effect could be compounded during years in which floating objects are scarce, because the num- ber of sets per floating object would increase. Since the skipjack catches have been increasing in the west- ern Pacific Ocean, their abundance and catch in the eastern Pacific could be lower than our estimates. A moratorium on dolphin sets is likely to result in reduced catchability, yield per recruit, average age, and total biomass of yellowfin. The catch of yellowfin, based on these factors only, was predicted to decline by approximately 55,600 t (25%). On the other hand, skipjack catches could increase, making the reduction in total tuna catches much smaller (4%). The effects of reductions in fishing effort, the range of the fishery, and recruitment were not ana- lyzed in this study because they are currently un- predictable; however, all three would result in an additional decrease in total tuna catches. If better predictions of effort levels and yellowfin recruitment are made, the yield-per-recruit estimates in Figure 9 could be used in conjunction with them to better predict yellowfin catches. The results of our analy- sis indicate that catches in the first years after a moratorium begins may not be indicative of the long- term catches. Fishermen, biologists, and managers should not consider these first-year catches as indices of future catches, because recruitment and catchability vary annually. On the other hand, our estimates of future average catches should be useful unless there are long-term changes in catchability or recruitment. Acknowledgments We would like to thank James Joseph, director of investigations of the Inter- American Tropical Tuna Commission for suggesting the need for this re- search, Richard B. Deriso for his many methodologi- cal suggestions, Alejandro Anganuzzi for his reviews and for sharing his knowledge about about dolphins, and William H. Bayliff for his extensive editorial re- views of this manuscript. Literature cited Abramson, N. J. 1971. Computer programs for fish stock assessment. F.A.O. Fish. Tech. Pap. 101, 154 p. Allen, R. L. 1981. Dolphins and the purse-seine fishery for yel- lowfin tuna. I.A.T.T.C. Int. Rep. 16, 23 p. Allen, R. L., and R. G. Punsly. 1984. Catch rates as indices of abundance of yellow- fin tuna, Thunnus albacares, in the eastern Pacific Ocean. I.A.T.T.C. Bull. 18(4):301-379. Bartoo, N. W., and A. L. Coan. 1978. Changes in yield per recruit of yellowfin tuna (Thunnus albacares) under the ICCAT minimum size regulation. I.C.C.A.T Col. Vol. Sci. Pap. 8(D:120-183. Bayliff, W. H. (ed.) 1990. Annual report of the Inter-American Tropical Tuna Commission 1989, 288 p. 1992. Annual report of the Inter-American Tropical Tuna Commission 1991, 271 p. Calkins, T. P. 1965. Variation in size of yellowfin tuna within in- dividual purse-seine sets. I.A.T.T.C. Bull. 10(8):463-524. Cole, J. S. 1980. Synopsis of biological data on the yellowfin tuna, Thunnus albacares (Bonnaterre, 1788), in the Pacific Ocean. I.A.T.T.C. Special Rep. 2:71-150. Green, R. E. 1967. Relationship of the thermocline to success of purse seining for tuna. Am. Fish. Soc, Trans. 96(2):126-130. Greenblatt, P. R. 1979. Association of tuna with flotsam in the east- ern tropical Pacific. Fish. Bull. 77(11:147-155. Gulland, J. A. 1965. Estimation of mortality rates. Annex to Rep. Arctic Fish. Working Group, Int. Counc. Explor. Sea CM. 1965(3), 9 p. Hennemuth, R. C. 1957. Analysis of methods of sampling to determine the size composition of commercial landings of yel- lowfin tuna (Neothunnus rnacropterus) and skip- jack (Katsuwonus pelamis). I.A.T.T.C. Bull. 2(5):174-243. Punsly et al.. Potential non-dolphin-associated tuna catches in the eastern Pacific Ocean 143 Holt, S. J. 1958. A note on the simple assessment of a pro- posal for mesh regultaion. I. C.N. A. F. Annual Proc. 8(4):82-83. Jones, R. 1961. The assesment of the long term effects of changes in gear selectivity and fishing effort. Mar. Res. Scot. 1961(2), 19 p. Joseph, J., and T. P. Calkins. 1969. Population dynamics of the skipjack tuna (Katsuwonis pelamis) of the eastern Pacific Ocean. I.A.T.T.C. Bull. 13(1), 273 p. Orange, C. J., and T. P. Calkins. 1981. Geographical distribution of yellowfin and skipjack tuna catches in the eastern Pacific Ocean, and fleet and total catch statistics 1975- 1978. I.A.T.T.C. Bull. 18(1):1-120. Pope, J. G. 1972. An investigation of the accuracy of virtual population analysis using cohort analysis. Int. Comm. Northwest Atl. Fish. Res., Bull. 9:65-74. Punsly, R. G. 1983. Estimation of the number of purse-seiner sets on tuna associated with dolphins in the eastern Pacific Ocean during 1959-1980. 18(3):227-299. I.A.T.T.C. Bull. 1987. Estimation of the relative annual abundance of yellowfin tuna, Thunnus albacares, in the east- ern Pacific Ocean during 1970-1985. I.A.T.T.C. Bull. 19(31:263-306. Punsly, R. G., and R. B. Deriso. 1991. Estimation of the abundance of yellowfin tuna, Thunnus albacares, by age groups and re- gions within the eastern Pacific Ocean. I.A.T.T.C. Bull. 20(2):98-131. Ricker, W. E. 1975. Computation and interpretation of biological statistics of fish populations. Fish. Res. Board Can., Bull. 191, 382 p. Tomlinson, P. K., S. Tsuji, and T. P. Calkins. 1992. Length frequency estimation for yellowfin tuna. I.A.T.T.C. Bull. 20(6):357-398. Wild, A. 1986. Growth of yellowfin tuna, Thunnus albacares, in the eastern Pacific Ocean based on otolith increments. I.A.T.T.C. Bull. 18(61:423^182. Abstract. Gastrointestinal tract contents were evaluated from 73 female and juvenile male north- ern fur seals (Callorhinus ursinus) for analysis of their diet in the Bering Sea. Fur seals were col- lected from August to October of 1981, 1982, and 1985. Juvenile walleye pollock (Theragra chalco- gramma) and gonatid squid were the primary prey. Pacific herring (Clupea pallasi) and capelin (Mallotus villosus), considered im- portant fur seal prey in previous reports, were absent from the diet. Prey species and size varied among years and between near- shore and pelagic sample loca- tions. Interannual variation in the importance of pollock in the diet of fur seals was positively related to year-class strength of pollock. Midwater (n=23) and bottom (rc=116) trawls were conducted at the location of fur seal collections to determine availability of fish and squid relative to prey species eaten by fur seals. The species and size composition of prey taken by fur seals was similar to midwater trawl collections, but differed from bottom trawl catches. Contrary to earlier conclusions that northern fur seals are opportunistic in their feeding habits, we conclude that fur seals are size-selective mid- water feeders during the summer and fall in the eastern Bering Sea. Prey selection by northern fur seals (Callorhinus ursinus) in the eastern Bering Sea Elizabeth Sinclair Thomas Loughlin National Marine Mammal Laboratory, Alaska Fisheries Science Center, National Marine Fisheries Service, NOAA 7600 Sand Point Way, N.E., Seattle, Washington 98 11 5 William Pearcy College of Oceanography, Oregon State University Oceanography Administration Building 1 04 Corvallis, Oregon 97331 Manuscript accepted 19 August 1993 Fishery Bulletin: 92:144-156 (1994) The Pribilof Island population (St. George and St. Paul Islands) of northern fur seals (Callorhinus ursinus) represents approximately 75% of the total species breeding population. Between 1975 and 1981, the Pribilof Island population declined from 1.2 million to an es- timated 800,000 animals (York and Hartley, 1981; Fowler, 1985). Abun- dance levels on St. Paul Island ap- pear to have stabilized (York and Kozloff, 1987) at a level 60-70% below estimates of the 1940s and 1950's, and at one-half the esti- mated carrying capacity (Fowler and Siniff, 1992). The number of animals continues to decline on St. George Island (York, 1990). The objectives of this study were to determine the species and size of prey eaten by northern fur seals in the eastern Bering Sea, to compare the seals' present diet with that prior to the population decline, and to examine the seals' consumption of prey relative to prey availability. Previous studies on the feeding habits of northern fur seals in the eastern Bering Sea (Scheffer, 1950a; Wilke and Kenyon, 1952; Wilke and Kenyon, 1957; North Pacific Fur Seal Commission Re- ports 1962, ' 1975, 2 and 19803; Fiscus et al., 1964; Fiscus et al., 1965; Fiscus and Kajimura, 1965) were conducted prior to the 1975- 81 population decline and prior to the 1970s development of a com- mercial walleye pollock (Theragra chalcogramma) fishery in the Bering Sea. Neither the size of fur seal prey, nor fur seal selection of prey relative to real-time availabil- ity have been previously examined in detail. Methods Northern fur seals were collected from 17 to 28 October 1981; from 24 September through 6 October 1 North Pacific Fur Seal Commission Re- port on Investigations from 1958 to 1961: Presented to the North Pacific Fur Seal Commission by the Standing Scientific Committee on 26 November 1962, 183 p. Available: Alaska Fish. Sci. Cent., NOAA, NMFS, 7600 Sand Point Way NE., BinC15700, Seattle, WA 98115-0070. - North Pacific Fur Seal Commission Re- port on Investigations from 1967 through 1972: Issued from the headquarters of the Commission, Washington, D.C., June 1975, 212 p. Available: Alaska Fish. Sci. Cent., NOAA, NMFS, 7600 Sand Point Way NE., BinC15700, Seattle, WA 98115-0070. 3 North Pacific Fur Seal Commission Re- port on Investigations during 1973-76: Issued from the headquarters of the Com- mission, Washington, D.C., February 1980, 197 p. Available: Alaska Fish. Sci. Cent., NOAA, NMFS, 7600 Sand Point Way NE., BinC 15700, Seattle, WA 98115-0070. 144 Sinclair et al.: Prey selection by Callorhinus ursmus 45 59° N • Seal samples (all years) D Marinovich midwater trawls A Diamond midwater trawls Middle Shelf Domain Figure 1 The study area with midwater trawl locations and northern fur seal collection positions. All midwater trawls were conducted in 1985. Seal numbers 1-17 were collected in 1981, 18-40 were collected in 1982, and 41-83 were collected in 1985. 1982; and from 6 to 16 August 1985. Collections were made within 185 km of the Pribilof Islands over the continental shelf, continental slope, and oceanic domain of the eastern Bering Sea (Fig. 1). Seals were shot from a small craft and returned to the NOAA ship Miller Freeman (65-m stern trawler) for examination within 1.5 hours of collec- tion. The esophagus of each seal was checked for food as an indication of regurgitation, and the gas- trointestinal (GI) tract was removed and frozen. Gastrointestinal tract contents were later thawed and gently rinsed through a series of graded sieves (0.71, 1.00 or 1.40, and 4.75 mm in 1981 and 1982; 0.50, 1.00, 1.40, and 4.75 mm in 1985). Fleshy re- mains were preserved in 10% formalin. Fish otoliths and bones were stored dry. Cephalopod rostra and statoliths were preserved in 70% isopropyl alcohol. Prey identification was based on all remains, in- cluding otoliths. Otoliths were not used for fish iden- tification in earlier fur seal diet studies because stomach samples were stored in formalin, which dissolves otoliths. Techniques and references for the identification of prey based on otoliths include Fitch and Brownell (1968), Morrow (1979), Frost and Lowry (1981), and otolith reference collections (see Acknowledgments). References for cephalopod beak and statolith identification include Clarke (1962), Young (1972), Roper and Young (1975), Clarke (1986), and beak and statolith reference collections (see Acknowledgments). A tooth was collected from each fur seal that was shot and ages were derived from direct readings of canine tooth sections follow- 146 Fishery Bulletin 92(1), 1994 ing Scheffer (1950b). In the analysis of data, males and females of all ages were treated as one group because of small sample sizes. The highest number of either upper or lower cephalopod beaks and left or right otoliths was re- corded as the maximum number of each species present. If deterioration made some left and right otoliths of a species indistinguishable, they were counted and the total was divided by 2. The fre- quency of occurrence and number of individuals from each prey taxon was calculated for each seal. The fork length (FL) of pollock and dorsal mantle length (DML) of squid was measured directly when whole prey were present in the stomachs. In the absence of whole prey, body size was estimated by measurement of otoliths and beaks. The maximum length of pollock otoliths and lower rostral length (LRL) of gonatid squid beaks were measured to the nearest 0.05 mm with vernier calipers. Squid DML's were estimated by comparison of LRL measure- ments to the LRL/DML relationship of 51 gonatid squid caught in trawls conducted in the vicinity of seal collections. Walleye pollock fork lengths were estimated by regression against otolith length (Frost and Lowry, 1981). For otoliths measuring: > 10.0mm,(FL) Y = 3. 175X - 9.770 ( R = 0.968) < 10.0mm, (FL) Y = 2.246X- 0.5 10 (/? = 0.981). Walleye pollock ages were estimated from these lengths based on length-age relation described by Smith (1981) and Walline (1983) for walleye pollock from the Bering Sea. Otoliths may dissolve or erode to varying degrees depending on their size and duration in fur seal stomachs. We evaluated the bias introduced in FL estimates due to eroded otoliths by assigning otoliths to four condition categories (excellent, good, fair, and poor) based on amount of wear. After qual- ity categorization, the maximum lengths of otoliths (except those in "poor" condition) were measured for estimation of body length by regression, and length frequencies of each category were determined inde- pendently. Cephalopod beaks are more resistant to digestion than otoliths and were typically identifiable. Beaks with chipped, worn, or broken rostra were rare and were not measured. Cephalopod beaks were identi- fied to species when possible, but most were catego- rized into two groups referred to as Gonatopsis bo- realis-Berryteuthis magister or Gonatus madokai- Gonatus middendorffi . The two individual species within each group can be separated based on their external morphology and statolith structure, but cannot presently be separated based on beak struc- ture alone (Clarke, 1986). Trawl collections of potential seal prey Trawls were conducted throughout the study area from the Miller Freeman between 1900 and 0600 hours within the vicinity of seal collections (Fig. 1). Both bottom and midwater trawls were conducted to provide a relative measure of the availability and size of potential fur seal prey species. Bottom trawls were made at 52-498 m (.v=139 m) depths with an 83/112 Eastern bottom trawl (17-m width, 2.3-m height mouth opening; 3.2-cm codend liner mesh; 360-mesh circumference; 200-mesh depth; 30-m bridle). Thirty-nine bottom trawls were conducted in 1981 (14 October-4 November), 51 in 1982 (24 September-8 October), and 26 in 1985 (5 August- 22 August). Seven 1985 trawls were made beyond maximum recorded dive depths of adult female seals (257 m; Ponganis et al., 1992). They were included in analyses because the species and size offish and squid caught were consistent with those caught by bottom trawl within seal dive depths. Collection and sorting methods and calculation of bottom trawl catch per unit of effort (CPUE) values followed Smith and Bakkala (1982). The total bot- tom trawl catch was randomly split into a sample of about 2500 kg. Individual species of fishes were identified and weighed (wet) and CPUE (no./ha) was estimated based on distance trawled. In 1981 and 1982, cephalopods were classified as squid or octo- pus and discarded. In 1985, all cephalopods were identified, sexed, weighed, and frozen whole. Beaks were extracted and stored in 70% isopropyl alcohol. Sex and age determination and body length mea- surements were made on a subsample of up to 200 walleye pollock from each trawl. Fork lengths were measured to the nearest centimeter. Saccular otoliths were collected for age determination (Smith and Bakkala, 1982) and stored in 70% isopropyl alcohol. Walleye pollock CPUE was calculated by age and body length. For purposes of this study, age- length frequencies for male and female walleye pol- lock were combined for each of the three years. Midwater trawls were made in 1985 with a Dia- mond midwater net (n=8) ( 10-16 fm mouth opening; 3.2-cm codend liner mesh; 354-mesh circumference; and 160-mesh depth with 2-m bridles) and a Marinovich herring trawl (rc = 15) (6.1-m width, 6.1- m height mouth opening; 1-cm codend liner mesh; 150-mesh circumference; and 350-mesh depth with 10-m bridles). Specific trawling positions were cho- sen within the vicinity of northern fur seal collec- tion areas based on the presence of fish or squid as Sinclair et al.: Prey selection by Callorhinus ursinus 147 indicated on 38 kHz echosounders and a chromoscope. Midwater towing depths measured by an attached transducer ranged from 22 to 340 m (x=143 m). All species of fish and cephalopods collected in midwater trawls were identified and counted. The CPUE and frequency of occurrence of each species, LRL and sex of gonatid squid, and walleye pollock frequency of occurrence by age and length were cal- culated separately for each trawl type. Comparison of seal diet and trawl collections The Odds Ratio (Fleiss, 1981) was used to compare prey availability (as determined by midwater and bottom trawls) with selection of prey by fur seals for each sample year: 0 = pV where pi = % of diet comprised by a given prey taxon, ql = % of diet comprised by all other prey taxon, p2 = % of food complex in environment com- prised by a taxa, and q2 = %> of food complex in environment com- prised by all other taxa. Values were calculated for number of each prey species and percent frequency of occurrence among seals, and CPUE values (no./ha) for each trawl type. Values for p2 and ql were also calculated for the trawl types combined in order to provide a compre- hensive description of the water column. The natu- ral log of the calculated Odds Ratio represents ei- ther positive or negative selection- The Odds Ratio was chosen because, unlike other electivity indices, the significance of the distance of calculated values from zero (null hypothesis that prey were consumed non-selectively) can be tested with the Z-statistic (Gabriel, 1978). In order to quantify the degree of overlap in the composition of bottom trawls, midwater trawls and fur seal GI contents, percent similarity (PS) values (Langton, 1982) were calculated: PS =100-0.5^a -b, where a = %> number of a given prey for seals, and b = % number of the same prey for trawls. Results Fur seal diet Eighty-three fur seals were collected. Ten of the 17 GI tracts collected in 1981 were empty and were excluded from the analysis. Of the 73 animals in- cluded in the analysis, 13 were juvenile males, 3 were juvenile females and 57 were adult females. Most fur seals were collected over depths less than 200 m within the outer shelf domain (Fig. 1). Fish represented 89% and cephalopods 11% of prey numbers for all three sample years combined. One-hundred percent of the GI tracts had fish re- mains and 82% of all samples contained walleye pollock. A total of 2,658 walleye pollock otoliths were measured. In all years combined, juvenile walleye pollock (3-20 cm FL) were the most numerous and frequently occurring prey species. Sixty-five percent of prey walleye pollock were from the 0-age group (3-13 cm FL) and 31% were from age group 1 (13- 20 cm FL). Only 4% of prey pollock were from age group 2 (20+ cm FL) and older. Gonatid squids occurred in 36% of the samples, but in comparison with pollock, they were not con- sumed in large numbers (Fig. 2). Gonatus madokai- G. middendorffi and Gonatopsis borealis-Berry- teuthis magister were the second most frequently occurring prey in all years combined. Seventy-nine percent of the 389 beaks measured were from squid 5-12 cm DML. Northern smoothtongue (Leuroglossus schmidti), a bathylagid deepsea smelt, was the second most numerous fish prey overall (Fig. 2) even though it was found only in 1985 (Table 1). Northern smooth- tongue composed a higher percentage of the total number offish than walleye pollock >2 years old for all sample years combined. Atka mackerel (Pleurogrammus monopterygius) composed 23.9% of the 1981 prey sample and was present in five of seven stomachs collected in 1981 that had prey re- mains, but the species was identified from the prey remains of only one other individual among the six collected in the same area in 1982 (Table 1). Although walleye pollock were eaten by fur seals in all 3 years, marked differences in age and body size were found between years (Table 1; Fig. 3). In 1981, the few walleye pollock otoliths found were from fish 3-4 years of age. Fur seal GI tracts con- tained primarily age-0 pollock in 1982 and age-1 pollock in 1985. Exclusion of otoliths that were in fair condition caused a downward shift in modal FL frequencies of 1 to 2 cm, but did not change our es- timation of the age categories of pollock eaten by fur seals. The species of forage fishes and squids consumed by fur seals varied between samples taken on and off the continental shelf (200 m) (Fig. 4). The GI tracts of fur seals collected over oceanic and conti- nental slope regions contained primarily northern smoothtongue and squids, especially Gonatopsis 148 Fishery Bulletin 92f 1). 1994 10 20 Percent number/frequency 30 40 50 60 70 walleye pollock (Theragra chalcogramma) gonatid squid (Gonatidae) Atka mackerel (Pleurogrammus monopterygius) northern smoothtongue (Leuroglossus schmidti) Salmoniformes fOsmendae) Percent of total number of prey, all years Percent frequency of occurrence of prey, all years Figure 2 Percent of total number and frequency of occurrence of primary prey in northern fur seal (Callorhinus ursinus) gastrointestinal tracts for sample years 1981, 1982, and 1985 combined. Species shown include the top three prey from each sample year. borealis-Berryteuthis magister. Seals collected over the continental shelf contained the remains of wall- eye pollock of all ages and squids, especially Gonatus madokai-G. middendorffi. Adult walleye pollock, although rare in stomach contents, were found in greatest frequency in fur seals collected from the outer domain of the continental shelf. Juvenile wall- eye pollock were consumed primarily over the midshelf and outer domain. Atka mackerel was found only in samples collected over the outer shelf domain north of Unimak Island. Comparisons with trawl samples Of the five top-ranked species collected in bottom trawls, only walleye pollock was found in fur seal GI contents (Figs. 2 and 5). Walleye pollock from bottom trawls ranged from 1 to over 12 years of age and had mean body lengths of 38.9 cm (3-4 years old) in 1981, 39.7 cm (4-5 years old) in 1982, and 44 cm (5-6 years old) in 1985 (Fig. 6). All but four of the cephalopods caught in 1985 bottom trawls were Berryteuthis magister ranging from 17.5 to 31.2 cm DML ( x = 2 1.6). As in the seal samples, B. magister was collected in trawls conducted over the outer continental shelf domain along the 200-m contour, or over the continental slope between 200 and 1000 m. Otherwise, the bottom trawl catch for all three years was so dissimilar to the midwater trawl catch (Figs. 5 and 6) and fur seal GI contents (Fig. 2) that electivity computations were not mean- ingful (Odds Ratio=0). Calculation of the Odds Ratio and Z-statistic on 1985 data with midwater and bottom trawl catch combined showed statistically significant positive selection by fur seals for age-0 pollock (P=0.0002), age-1 pollock (P<0.0001), northern smoothtongue (P<0.0001), and gonatid squid (P=0.02). Negative selection for adult walleye pollock was suggested but was not statistically significant (P=0.13). A similarity index of 81% was calculated for spe- cies composition and prey size in the 1985 GI samples and midwater trawls. Fur seals fed on three of the four top-ranked species caught in midwater trawls (Figs. 2 and 5). Midwater trawls and seals caught predominantly juvenile walleye pollock. Gonatid squids (Gonatus madokai, G. middendorffi, and Gonatopsis borealis) had low CPUE values but were second in frequency of occurrence in both fur seal GI tracts and midwater trawls. The modal length of walleye pollock and gonatid squids was 5- 20 cm in both midwater trawl and GI samples in 1985. Few adult walleye pollock and no large squid were collected in midwater trawls or seal GI samples. Seals and midwater trawls caught the same prey species at the same general locations on and off the continental shelf (Fig. 4). As in GI contents, age-0 and age-1 walleye pollock were collected in midwater trawls made on the middle and outer shelf and near the continental slope. Gonatopsis borealis were found on the continental slope and near-slope. Gonatus madokai and G. middendorffi were found throughout the sampling area, but primarily on the outer continental shelf and near-slope sampling areas. Sinclair et al.: Prey selection by Callorhinus ursinus 149 Table 1 Gastrointestinal contents of 73 northern fur seals (Callorh nus ursinus) collected from the Bering Sea in 1981 (n=7), 1982 (n=23), and 1985 (n=43). Tentative identifications are designated as (t). Prey species % number in each year % frequency occurrence 1981 1982 1985 1981 1982 1985 Fish Clupea pallasi — 0.1 — — 4.4 — Osmeridae (t) 8.7 — — 42.9 — — Salmonidae 5.4 — — 42.9 — — Leuroglossus schmidti — — 12.7 — — 9.3 Gadus macrocephalus (t) — — 0.1 — — 7.0 Theragra chalcogramma 54.4 87.3 74.1 100 95.7 72.1 3-5cm fork length — (8.8) (5.7) 5-10cm fork length (4.3) (63.9) (2.3) 10-20cm fork length — — (55.6) >20cm fork length (38.0) (1.4) (1.7) T. chalcogramma (t) — 0.1 0.1 — 8.7 4.7 unidentified Gadidae — — 0.9 — — 20.9 Lycodes sp. 1.1 — 0.5 14.3 — — Pleurogrammus monopterygius 23.9 0.1 — 71.4 4.4 — P. monopterygius (t) — 0.1 — — 4.4 — unidentified percoid 1.1 — — 14.3 — — unidentified fish 5.4 0.4 0.5 14.3 13.0 25.6 Squid Gonatus berryi — — 0.1 — — 2.3 G. pyros — — 0.1 — — 2.3 G. tinro — — 0.1 — — 2.3 G. tinro (t) — — 0.1 — — 2.3 Gonatus madokai-middendorffi — 0.1 4.8 — 4.4 34.9 Gonatus sp. — — 0.1 — — 2.3 Berryteuthis magister — 0.6 — — 8.7 — Gonatopsis borealis-B. magister — 10.2 6.4 — 17.4 20.9 unidentified Gonatidae — — 0.1 — — 7.0 unidentified squid — 1 0 — — 34.8 — Total number prey 92 1638 2189 Total number fish 92 1445 1936 100 100 100 Total number squid 0 193 253 0 52.2 46.5 Discussion The modal size distribution of walleye pollock in GI contents of female and juvenile male fur seals re- flected year-class strength projections of walleye pollock (Fig. 7). Walleye pollock have highly variable recruitment rates (Smith, 1981), and year-class strength varied five-fold between 1977 and 1982 (Bakkala et al., 1987). Population estimates based on bottom trawl and midwater acoustic surveys in the eastern Bering Sea indicated that the 1980 year class (age 1 in 1981) was about half the average year-class size; the 1981 year class (age 0 in 1981) was the weakest observed prior to 1983; and the 1978 year class (age 3 in 1981) was the strongest observed. The 1982 and 1984 year classes were strong and the 1985 year class was considered av- erage (Bakkala et al., 1987). Similarly, walleye pol- lock as prey in 1981 were primarily adults 3 and 4 years of age (from the 1977 and 1978 year class); in 1982, seals ate age-0 pollock exclusively; and in 1985, prey pollock were primarily from the 1984 year class. The concordance of pollock recruitment and fur seal GI content analysis indicates that the variable recruitment of walleye pollock affects prey consumption by northern fur seals. The three basic dive patterns described for adult females in the Bering Sea are shallow, pelagic night- time diving (most commonly to 50—60 m); deep day- and-night diving over the continental shelf (most commonly to 175 m); and some combination of both, including shallow diving over the continental shelf 50 Fishery Bulletin 92(1). 1994 ■ 1981 n.39 ' □ 1982 n-1191 1985 n-1428 Figure 3 Age-length frequencies of walleye pollock (Ther chalcogramma) based on otoliths in northern fur [Callorhinus ursinus) gastrointestinal tracts by year. agra seal and both shallow and deep diving along the conti- nental slope. Dive pattern information is based on time-depth recordings (Gentry et al., 1986; Loughlin et al., 1987; Goebel et al., 1991), radio telemetry (Loughlin et al., 1987), stomach volume estimates (Mead, 1953; Taylor et al., 1955'; Spalding, 1964; Wada, 1971; Kajimura, 1984), and stomach clear- ance studies (Miller, 19785; Bigg, 19816; Bigg and Fawcett, 1985; Murie and Lavigne, 1985). Based on fur seal and trawl collections in this study and on distributional information of prey (Smith, 1981; Dunn, 1983; Kubodera and Jefferts, 1984; Lynde, 1984), shallow diving fur seals over the continental shelf concentrated on juvenile walleye pollock and juvenile gonatid squid (Gonatus madokai-G. middendorffi), while shallow divers off-shelf targeted juvenile gonatid squid (Berryteuthis magister-Gonalopsis borealis) and bathy- lagid smelt. Daytime deep diving over the continental shelf would be advantageous to seals concentrating on prey (i.e., adult wall- eye pollock) that tend to school at depth during daytime hours and disperse as they rise in the water column at night. Adult gonatid squid probably occur in schools at the bottom on the continental shelf and re- main deep along the shelf edge during both day and night. The location and degree of concentration of prey may be closely associ- ated with the hydrography of the foraging region. The hydrography of the foraging re- gion may have the most direct influence on the diving patterns of fur seals. Hydrographic characteristics of the Bering Sea continental shelf, include a two-layered midshelf and a three-layered outer shelf domain that may stratify and concentrate prey by species and age in a vertical plane. Nishiyama et al. (1986) proposed that ver- tical stratification within the eastern Bering Sea shelf serves as a "nursery layer" to con- fine young-of-the-year pollock in the upper 40 m of the water column within the bound- ary region between the upper and lower lay- ers. Copepod nauplii are also concentrated in this area, providing a ready source of food for larval walleye pollock (Bailey et al., 1986). Wada ( 1971) determined that primary foods of fur seals off the Sanriku Coast in Japan consisted of migrating species closely related to boundary regions, especially transition zone regions. The horizontal temperature and salinity structures that occur on either side of frontal regions within our study area (Kinder and Schumacher, 1981 ) may 4 Taylor. F II (' , M. Fuginaga, and F. Wilke. 1955. Distribu- tion and food habits of the fur seals of the North Pacific Ocean Rept. of Coop. Invest by the Govts, of Can., Japan, and the U.S.A. Feb. -July, 86 p Available Alaska Fish. Sci. Cent.. NOAA, NMFS, 7600 Sand Point Waj \K , BinC 15700, Seattle, WA 98115-0070. ' Miller, L. K. 1978. Energetics of the northern fur seal in rela- tion to climate and food resources of the Bering Sea. Final Rep. to U S. Mar. Mamm. Comm. MMC-75/08, 27p. K Bigg. M. A. 1981. Digestion rates of herring (Clupea harengus pallasi i and squid iLali^o npalfsri-nsi in northern fur seals. Submitted to the 24th Annual Meeting of the Standing Sci. Comm., N. Pac. Fur Seal Comm., 6-10 April, Tokyo, Japan. Available: Alaska Fish. Sci. Cent.. 7600 Sand Point Way NE., BmC15700, Seattle, WA 98115-0070. Sinclair et al.: Prey selection by Callorhmus ursinus 151 Percent number 100 90 80 70 60 50 40 30 20 10 10 20 30 40 50 60 70 80 90 100 northern smoothtongue (Leurogkxsus Schmidt) gonatid squid (Gonatopsis borealis/ Berryteu&tts magister) gonatid squid (Gonatus madokairmiddendotlrl) walleye pollock (all ages) (Theragra chalcogramma) Oceanic domain/Continental slope seals midwater trawls Continental shell (<200m) seals midwater trawls Figure 4 Primary species identified in fur seal (Callorhinus ursinus) gastrointestinal tracts and midwater trawls col- lected on and off the eastern Bering Sea continental shelf in 1985. walleye pollock (Themgra chalcogramma) lanternfish (Myaophxlae) northern smoothtongue (Leuroglossus schmidti) gonatid squid (Gonatidae) 40 60 80 100 Q Marinovich midwater trawl ■ Diamond midwater trawl 50 I 100 150 200 walleye pollock (Theragra chalcogramma) yellowfin sole I jronectes asper) iMmnmu (Pleuronectes asper) rock sole (Lepidopsena bilineala) I Pacific cod (Gadus macrocephalus) Bottom trawls ■ 1981 ^ 1982 □ 1985 Mean CPUE values (no. /ha) Figure 5 Catch per unit of effort (CPUE) number/hectare (no. /ha) values for species caught in 1985 midwater trawls and bottom trawls in 1981, 1982, and 1985. 152 Fishery Bulletin 92(1), 1994 400 300- 1 1 200 $ E 100- 1985 Trawls | Midwater □ Bottom N-4140 Jll^l 10 20 25 30 Fork length (cm) J I 45 _L_L 50 2 3 Age (years) 6 7 Figure 6 Age-length frequencies of walleye pollock (Therag chalcogramma) collected in bottom and midwater trawls 19H5. also form boundaries that concentrate prey. Diving depths of 175 m coincide with the depth break of the outer continental shelf. Diving depths of 50-60 m coincide with the depth break of the frontal systems between the midshelf and inner shelf. Previous analyses of fur seal diet in the eastern Bering Sea were based primarily on a sample of 3,530 stomachs collected pelagically in 1960, 1962- 64, 1968, 1973, and 1974 (North Pacific Fur Seal Commission Reports 1962, ' 1975,2 and 19803; Fiscus et al. 1 D64; Fiscus et al. 1965; Fiscus and Kajimura 1965;. i:tviews of the pelagic data cite walleye pol- lock (Kajimura, 1985; Perez and Bigg, 1986), Pacific herring (Clupea pallasi), capelin (Mallotus villosus), Atka mackerel, gonatid squids (Gonatus spp., Berryteuthis magister and Gonatopsis borealis), and intermittently, northern smoothtongue (Kajimura, 1984) as principal fur seal prey in the eastern Bering Sea. Published reports and reviews of fur seal feeding habits prior to the pelagic collections (1892-1950's) also described walleye pollock, cape- lin, gonatid squid, and bathylagid smelt as primary prey in seal spewings or stomachs (Scheffer, 1950a; Wilke and Kenyon, 1952; Wilke and Kenyon, 1957). In terms of prey species composition, the summer diet of female and juvenile male northern fur seals does not appear to have changed dramatically since the turn of the century. Pollock and gonatid squid are still the predominant prey of northern fur seals in the eastern Bering Sea. More subtle changes, such as a decrease in pollock size may have occurred (Smith, 1981; Swartz- man and Haar, 1983) and could play a criti- cal role in foraging success of northern fur seals. Unfortunately, records of prey size in historical fur seal diet studies are incom- plete. It should be noted that Pacific herring and capelin were absent from fur seal di- ets in this study, despite collections in ar- eas where they occurred as important prey in the past. Fluctuation in the population status of Pacific herring and capelin in the Bering Sea has been attributed to the spo- radic and localized nature of their abun- dance (Turner, 1886; Meek, 1916; Favorite et al. 19777; Lowe 19918), overharvesting and displacement by walleye pollock (Wespestad and Barton, 1981; Swartzman and Haar, 1983; Wespestad and Fried, 1983; Bakkala et al., 1987), and/or environ- mental change such as the pronounced warming in the Gulf of Alaska and Bering Sea over the past decade (Royer, 1989). The absence of these previously important prey may be critical to seals during successive years of weak walleye pollock year-class abundance. Fur seals select juvenile walleye pollock as prey despite a wide availability of other prey types within their dive range. Fur seals may select their prey by size and schooling behavior, whether the prey are myctophids in oceanic waters off Japan (Wada, 1971); Pacific herring, capelin, market squid iLoligo opalescens) and Pacific whiting (=Pacific hake, Merluccius productus) in the eastern North Pacific (Kajimura, 1984; Perez and Bigg, 1986); or walleye pollock in the eastern Bering Sea (Kajimura, 1984). The most consistent prey characteristic between feeding studies across the northern fur seal range 7 Favorite, F., T. Laevastu, and R. R. Straty. 1977. Oceanogra- phy of the northeastern Pacific Ocean and eastern Bering Sea, and relations to various living marine resources. NWAFC Proc. Rep. 280p. Alaska Fish. Sci. Cent., NMFS, NOAA, 7600 Sand Point Way NE., Bin C 15700, Seattle, WA 98115-0070, 280p. R Lowe, S. A. 1991. Atka mackerel. In Stock assessment and fish- ery evaluation report for the groundfish resources of the Bering Sea/ Aleutian Islands region as projected for 1992, p. 11-2 to 11-40. North Pacific Fishery Management Council, P.O. Box 103136, Anchorage, AK 99510. Sinclair et al.: Prey selection by Callorhmus ursinus 53 B * C l/i IB C £: O ra E a> = >• c SI o « °- g. o CL 90 - 80 - 70- 60 - 50 - 40- 30 - 20 - 10- 0 - 100 1978 1979 1980 1981 1982 1983 1984 1985 Pollock year class 80 70 - 2 50 ■ 1981 □ 1985 1978 1979 1980 1981 1982 1983 1984 1985 Pollock year class Figure 7 Estimates of walleye pollock iTheragra chalcogramma) year-class strength 1978- 85 (Bakkala et al., 1987), and the relative abundance of specific year classes in northern fur seal gastrointestinal tracts. is size and the tendency to form dense schools. In this sense, a "juvenation" of walleye pollock in the Bering Sea (Swartzman and Haar, 1983) may have provided fur seals with a newly abundant but un- stable resource, due to large fluctuations in the annual year-class strength of walleye pollock and due to potential displacement of other prey species (Pacific herring and capelin). During years of low pollock recruitment, fur seals may switch to other prey such as capelin and Pacific herring, and expe- rience food limitation only if these alternate prey resources have been displaced or depleted. Histori- cal records of northern fur seal diet are inadequate to either support or refute an "alternate prey" ar- gument. However, we suggest that when juvenile walleye pollock are unavailable, such as in our 1981 sampling season, female and juvenile fur seals se- lect other specific prey of the same size and eat adult walleye pollock only if these other preferred prey are not available. During their summer breeding season, northern fur seals consume the most abundant and available fish and squid in the eastern Bering Sea. Walleye pollock make up an estimated 50% of the ground- fish biomass in the eastern Bering Sea and Aleutian Islands area (Walters et al., 1988) and dense aggre- gations of 0-age pollock occur off the Pribilof Islands June through mid-August (Smith, 1981). Kubodera and Jefferts (1984) suggested gonatids are the ma- jor pelagic cephalopod group in the Bering Sea, where large increases in abundances of larval and postlarval gonatid squid occur in early June. Among Bering Sea gonatids, Gonatopsis borealis and Berryteuthis magister are considered to be among the most numerically dominant (Jefferts, 1983; Kubodera and Jefferts, 1984). Selection by northern fur seals of a wide variety of numerically dominant prey species throughout their migratory range has led to the general conclu- sion that they are non-specific, opportunistic feed- ers (Kajimura, 1985). Northern fur seals are flexible in their feeding habits, as indicated by the variation in GI contents of seals collected between California and Alaska. Nonetheless, fur seals concentrate on an average of three primary species within each oceanographic subregion (Perez and Bigg, 1986). In addition, fur seal consumption of walleye pollock, gonatid squid, and bathylagid smelt in the eastern Bering Sea is consistent throughout historical records, despite the wide variety of prey available to fur seals within their diving range. Based on this study, we conclude that female and young male fur seals select juvenile and small-sized fish and squid, despite the availability of larger prey types within their diving range. This study demonstrates that female and young male fur seals are size-selective midwater shelf and mesopelagic feeders, at least during the breeding and haul-out season in the east- ern Bering Sea. Acknowledgments Otolith identifications for the 1981 samples were made by the late J. Fitch. Otolith identifications for 1982 and 1985 were based on the otolith reference collections at the National Marine Mammal Labo- 54 Fishery Bulletin 92(1). 1994 ratory (NMML) and Los Angeles County Museum (LACM). Cephalopod identifications were based on the reference collections of the NMML and Oregon State University (OSU). Voucher specimens of prey material (statoliths, beaks, otoliths, teeth, and bones) are archived at the NMML. Identifications of squid and squid beaks were confirmed by C. Fiscus (NMML, retired), K. Jefferts (OSU), and W. Walker (LACM). Identification of fish otoliths and bones were confirmed by G. Antonelis Jr. (NMML) and J. Dunn (University of Washington) respectively. Voucher samples of juvenile pollock otoliths were confirmed by A. Brown (Alaska Fisheries Science Center [AFSC]), K. Frost (Alaska Department of Fish and Game [ADF&G], and L. Lowry [ADF&G]). Gary Walters (AFSC) helped interpret bottom trawl values, and W Carlson and C. Leap of the AFSC Graphics Unit helped produce the figures. The following individuals contributed to the qual- ity and content of this manuscript: G. Antonelis Jr., J. Baker, L. Fritz, R. Gentry, P. Livingston, and two anonymous reviewers. These data were first pre- sented in part as a Northwest and Alaska Fisher- ies Center Processed Report (Hacker and Antonelis, 19869) and in full as a Masters Thesis from Oregon State University (Sinclair, 1988). Literature cited Bailey, K., R. Francis, and J. Schumacher. 1986. Recent information on the causes of variabil- ity in recruitment of Alaska pollock in the eastern Bering Sea: physical conditions and biological interactions. Int. North Pac. Comm. Bull. 47:155-165. Bakkala, R. G., V. G. Wespestad, and J. J. Traynor. 1987. Walleye pollock. In R. G. Bakkala and J. W. Balsiger (eds.), Condition of groundfish resources of the eastern Bering Sea and Aleutian Islands region in 1986, p. 11-29. U.S. Dep. Commer, NOAATech. Memo. NMFS F/NWC-117. Bigg, M. A., and I. Fawcett. 1985. Two biases in diet determination of northern fur seals (Callorhinus ursinus). In J. R. H. Beddington, R. J. H. Beverton, and D. M. Lavigne (eds.), Marine mammals and fisheries, p. 284- 299. George Allen and Unwin Ltd., London. 9 Hacker, E. S., and G. A. Antonelis Jr. 1986. Pelagic food hab- its of northern fur seals. In T. R. Loughlin and P. Livingston (eds.), Summary of joint research on the diets of northern fur seals and fish in the Bering Sea during 1985, p. 5-22. NWAFC Proc. Rep. 86-19. Alaska Fish. Sci. Cent., NMFS, NOAA, 7600 Sand Point Way NE., BinC15700, Seattle, WA 98115-0070. Clarke, M. R. 1962. The identification of cephalopod "beaks" and the relationship between beak size and total body weight. Bull. Br. Mus. (Nat Hist.) Zool. 8:419- 480. Clarke, M. R. (ed.) 1986. A handbook for the identification of cepha- lopod beaks. Clarendon Press, Oxford, 273 p. Dunn, J. R. 1983. Development and distribution of the young of northern smoothtongue, Leuroglossus schmidti (Bathylagidae), in the Northeast Pacific, with com- ments on the systematics of the genus Leuroglossus Gilbert. Fish. Bull. 81(l):23-40. Fiscus, C. H., and H. Kajimura. 1965. Pelagic fur seal investigations, 1964. U.S. Fish. Wildl. Serv. Spec. Sci. Rep. Fish. No. 522, 47 p. Available: Alaska Fish. Sci. Cent., NOAA, NMFS, 7600 Sand Point Way NE, BINC15700, Seattle, WA 98115-00700. Fiscus, C. H., G. A. Baines, and F. Wilke. 1964. Pelagic fur seal investigations, Alaska wa- ters, 1962. U.S. Fish. Wildl. Serv. Spec. Sci. Rep. Fish. No. 475, 59 p. Available: Alaska Fish. Sci. Cent., NOAA, NMFS, 7600 Sand Point Way NE., BINC15700, Seattle, WA 98115-0070. Fiscus, C. H., G. A. Baines, and H. Kajimura. 1965. Pelagic fur seal investigations, Alaska, 1963. U.S. Fish. Wildl. Serv. Spec. Sci. Rep. Fish. No. 489, 33 p. Available: Alaska Fish. Sci. Cent., NOAA, NMFS, 7600 Sand Point Way NE., BinC15700, Seattle, WA 98115-0070. Fitch, J. E., and R. L. Brownell Jr. 1968. Fish otoliths in cetacean stomachs and their importance in interpreting feeding habits. J. Fish. Res. Board Can. 25(121:2561-2574. Fleiss, J. L. 1981. Statistical methods for rates and proportions, 2nd ed. John Wiley & Sons, 321 p. Fowler, C. W. 1985. An evaluation of the role of entanglement in the population dynamics of northern fur seals on the Pribilof Islands. In R. S. Shomura and H. W. Yoshida (eds.). Proceedings of the workshop on the fate and impact of marine debris; 27-29 Novem- ber 1984, Honolulu, HI., p. 291-307. U.S. Dep. Commer., NOAA Tech. Memo. NOAA-TM-NMFS- SWFC-54. Fowler, C. W., and D. B. Siniff. 1992. Determining population status and the use of biological indices for the management of marine mammals. In D. R. McCullough and R. H. Reginald (eds.), Wildlife 2001: populations, p. 1051-1061. Elsevier Science Publishers. London, England. Frost, K. J., and L. F. Lowry. 1981. Trophic importance of some marine gadids in northern Alaska and their body-otolith size relationships. Fish. Bull. 79(1):187-192. Sinclair et al.: Prey selection by Callorhinus ursinus 55 Gabriel, W. L. 1978. Statistics of selectivity. In S. J. Lipovsky and C. A. Simenstad (eds.), Gutshop '78, p. 62- 66. Wash. Sea Grant Publ. WSG-WO-77-2, Univ. Washington, Seattle, WA. Gentry, R. L., G. L. Kooyman, and M. E. Goebel. 1986. Feeding and diving behavior of northern fur seals. In R. L. Gentry and G. L. Kooyman (eds.), Fur seals: maternal strategies on land and at sea, p. 61-78. Princeton Univ. Press, Princeton, NJ. Goebel, M. E., J. L. Bengtson, R. L. DeLong, R. L. Gentry, and T. R. Loughlin. 1991. Diving patterns and foraging locations of fe- male northern fur seals. Fish. Bull. 89:171-179. Jefferts, K. 1983. Zoography and systematics of cephalopods of the northeastern Pacific Ocean. Ph.D. diss., Or- egon State University, Corvallis, OR, 291 p. Kajimura, H. 1984. Opportunistic feeding of the northern fur seal, Callorhinus ursinus, in the eastern North Pacific Ocean and eastern Bering Sea. U.S. Dep. Commer. NOAA Tech. Rep. NMFS SSRF-779, 49 p. 1985. Opportunistic feeding by the northern fur seal, (Callorhinus ursinus). In J. R. Beddington, R. J. H. Beverton, and D. M. Lavigne (eds.), Ma- rine mammals and fisheries, p. 300-318. George Allen and Unwin Ltd., London. Kinder, T. H., and J. D. Schumacher. 1981. Hydrographic structure over the continental shelf of the southeastern Bering Sea. In D. W. Hood, and J. A. Calder (eds.), The eastern Bering Sea shelf: oceanography and resources, Vol. I, p. 31-52. Office of Mar. Poll. Assessment of NOAA. Kubodera, T., and K. Jefferts. 1984. Distribution and abundance of the early life stages of squid, primarily Gonatidae (Cepha- lopoda, Oegopsida), in the northern North Pacific (Part I and II). Bull. Nat. Sci. Mus. (Tokyo) Ser. A, 10(3 and 41:91-193. Langton R. W. 1982. Diet overlap between Atlantic cod, Gadus morhua, silver hake, Merluccius bilinearis, and fif- teen other northwest Atlantic finfish. Fish. Bull. 80(4):745-759. Loughlin, T. R., J. L. Bengston, and R. L. Merrick. 1987. Characteristics of feeding trips of female northern fur seals. Can. J. Zool. 65(8):2079-2084. Lynde, C. M. 1984. Juvenile and adult walleye pollock of the east- ern Bering Sea. In D. Ito (ed.), Proceedings of the workshop on walleye pollock and its ecosystem in the eastern Bering Sea, p. 43-108. U. S. Dep. Commer., NOAA Tech. Memo NMFS F/NWC-62. Mead Jr., G. W. 1953. The food habits of the North Pacific fur seal in Japanese waters. Ph.D. diss., Stanford Univ., Palo Alto, CA, 138 p. Meek, A. 1916. The migrations of fish. Edward Arnold, Lon- don, 427 p. Morrow, J. E. 1979. Preliminary keys to otoliths of some adult fishes of the Gulf of Alaska, Bering Sea, and Beau- fort Sea. U.S. Dep. Commer. NOAA Tech. Rep. NMFS Circ. 420, 33 p. Murie, D. J., and D. M. Lavigne. 1985. Interpretation of otoliths in stomach content analyses of phocid seals: quantifying fish consumption. Can. J. Zool. 64:1152-1157. Nishiyama, T., K. Hirano, and T. Haryu. 1986. The early life history and feeding habits of larval walleye pollock, Theragra chalcogramma (Pallas), in the southeast Bering Sea. Int. North Pac. Fish. Comm. Bull. 45:177-227. Perez, M. A., and M. A. Bigg. 1986. Diet of northern fur seals, Callorhinus ursinus, off western North America. Fish. Bull. 84(4):957-971. Ponganis, P. J., R. L. Gentry, E. P. Ponganis, and K. V. Ponganis. 1992. Analysis of swim velocities during deep and shallow dives of two northern fur seals, Cal- lorhinus ursinus. Mar. Mammal Sci. 8(l):69-75. Roper, C. F. E., and R. E. Young. 1975. Vertical distribution of pelagic cephalopods. Smithson. Contrib. Zool. 209, 51 p. Royer, T. C. 1989. Upper ocean temperature variability in the northeast Pacific Ocean: is it an indicator of glo- bal warming? J. Geophys. Res. 94(C12):18,175- 18,183. Scheffer, V. B. 1950a. The food of the Alaska fur seal. Transactions of the fifteenth North American Wildlife Conference, p. 410-420. Available: Alaska Fish. Sci. Cent., NOAA, NMFS, 7600 Sand Point Way NE., BinC 15700, Seattle, WA 98115-0070. 1950b. Growth layers on the teeth of Pinnipedia as an indication of age. Science 112:309-311. Sinclair, E. H. 1988. Feeding habits of northern fur seals in the eastern Bering Sea. M.S. thesis, College of Oceanography, Oregon State Univ., 1988, Cor- vallis, OR, 94 p. Smith, G. B. 1981. The biology of walleye pollock. In D. W. Hood and J. A. Calder (eds.), The eastern Bering Sea Shelf: oceanography and resources, Vol. I, p. 527-552, Office of Mar. Poll. Assessment of NOAA, 625 p. Smith, G. B., and R. G. Bakkala. 1982. Demersal fish resources of the Eastern Bering Sea: spring 1976. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-754, 129 p. Spalding, D. J. 1964. Comparative feeding habits of the fur seal, sea lion, and harbour seal on the British Colum- bia Coast. Fish. Res. Board Can. Bull. 146, 52 p. Swartzman, G. L., and R. T. Haar. 1983. Interactions between fur seal populations 156 Fishery Bulletin 92(1), 1994 and fisheries in the Bering Sea. Fish. Bull. 81(1): 121-132. Turner, L. M. 1886. Researches in Alaska. Pt. IV: Fishes. Contributions to the natural history of Alaska: results of investigations made chiefly in the Yukon district and the Aleutian Islands; conducted under the auspices of the Signal Service, U.S. Army, ex- tending from May 1874, to August 1881, p. 87- 113. No. II, Arctic Series of Pubis., issued in con- nection with the Signal Service, U.S. Army Gov. Print. Off., Washington DC. Wada, K. 1971. Food and feeding habit of northern fur seals along the coast of Sanriku. Bull. Tokai Reg. Fish. Res. Lab 64, Jan., p. 1-37. Walline, P. D. 1983. Growth of larval and juvenile walleye pollock related to year-class strength. Ph.D. diss., Univ. Washington, Seattle, WA, 144 p. Walters, G. E., K. Teshima, J. J. Traynor, R. G. Bakkala, J. A. Sassano, K. L. Halliday, W. A. Karp, K. Mito, N. J. Williamson, and D. M. Smith. 1988. Distribution, abundance, and biological char- acteristics of groundfish in the eastern Bering Sea based on results of the U.S. -Japan triennial bot- tom trawl and hydroacoustic surveys during May- September, 1985. U.S. Dep. Commer., NOAA Tech. Memo. NMFS F/NWC-154, 400 p. Wespestad, V. G., and L. H. Barton. 1981. Distribution, migration, and status of Pacific herring. In D. W. Hood and J. A. Calder (eds.), The eastern Bering Sea Shelf: oceanography and resources, Vol. I, p. 509-525. Office of Mar. Poll. Assessment of NOAA, 625 p. Available: Alaska Fish. Sci. Cent., NOAA, NMFS, 7600 Sand Point Way NE., BinC15700, Seattle, WA 98115-0070. Wespestad, V. G., and S. M. Fried. 1983. Review of the biology and abundance trends of Pacific herring (Clupea harengus pallasi). In W. S. Wooster (ed.), From year to year, p. 17- 29. Univ. Washington, Seattle, WA. Sea Grant Prog. Publ. WSG-WO-83-3. Wilke, F., and K. W. Kenyon. 1952. Notes on the food of fur seal, sea-lion, and harbor porpoise. J. Wildl. Manage. 16(3):396- 397. 1957. The food of fur seals in the eastern Bering Sea. J. Wildl. Manage. 21(21:237-238. York, A. E. 1990. Trends in numbers of pups born on St. Paul and St. George Islands 1973-88. In H. Kajimura (ed.), Fur seal investigations, 1987-88, p. 31- 37. U.S. Dep. Commer., NOAA Tech. Memo. NMFS F/NWC-180. York, A. E., and J. R. Hartley. 1981. Pup production following harvest of female northern fur seals. Can. J. Fish. Aquat. Sci. 38:84-90. York, A. E., and P. Kozloff. 1987. On the estimation of numbers of northern fur seal, Callorhinus ursinus, pups born on St. Paul Island, 1980-1986. Fish. Bull. 85(2):367-375. Young R. E. 1972. The systematics and areal distribution of pe- lagic cephalopods from the seas off southern Cali- fornia. Smithson. Contrib. Zool. (97), 159 p. Abstract. — The stomach con- tents of 1,215 anadromous ale- wives collected during winter and summer groundfish research sur- veys (1990-91) off Nova Scotia were examined to 1) describe the diet by season, area, bottom depth (<101 m, 101-200 m, >200 m), time of day and fish size (<151 mm, 151-200 mm, 201-250 mm, >250 mm FL), 2) evaluate diel feeding periodicity, and 3) estimate daily ration. Euphausiids, particu- larly Meganyctiphanes norvegica, were the most important prey and represented more than 82% by vol- ume of total stomach contents sea- sonally and geographically. Contri- butions by other prey groups (hyperiid amphipods, calanoid copepods, crustacean larvae, poly- chaetes, chaetognaths, mysids, pteropods, and fish larvae) were small and varied temporally and spatially. The proportion of eu- phausiids in the diet of alewives from the Scotian Shelf (winter) and Bay of Fundy (summer) tended to increase with increasing depth. Day and night differences in diet composition indicate that alewives may particulate-feed on macrozooplankton when prey vis- ibility is high and filter-feed on microzooplankton when prey vis- ibility is low. Diet composition was relatively homogenous among ale- wife size groups with euphausiids composing most of the total food volume. Alewives of different size groups ate similarly sized M. norvegica, generally the largest M. norvegica available. Diel feeding activity (stomach fullness) peaked at mid-day (summer collections) and mid-afternoon (winter collec- tions); feeding activity was re- duced at night. In all areas, feed- ing activity and the proportion of feeding fish was highest in regions where bottom depths exceeded 200 m. Mean stomach fullness was highest during summer in the Bay of Fundy and during winter on the Scotian Shelf; these regions are seasonally important foraging ar- eas for alewives off Nova Scotia. Daily ration was 1.2% of body weight during winter and 1.9% during summer. Manuscript accepted 17 August 1993 Fishery Bulletin 92:157-170 (1994) Feeding habits of anadromous alewives, Alosa pseudoharengus, off the Atlantic Coast of Nova Scotia Heath H. Stone Department of Fisheries and Oceans, Biological Sciences Branch PO. Box 550, Halifax. Nova Scotia B3J 2S7 CANADA Present address: Biological Station, Department of Fisheries and Oceans St. Andrews, New Brunswick, EOG 2XO CANADA Brian M. Jessop Department of Fisheries and Oceans, Biological Sciences Branch, RO. Box 550, Halifax, Nova Scotia, B3J 2S7 CANADA The anadromous alewife, Alosa pseu- doharengus, is a clupeiform fish whose range extends from New- foundland to North Carolina (Bigelow and Schroeder, 1953). Off Nova Scotia, alewives occur through- out the year in regions characterized by strong tidal mixing and up- welling in the Bay of Fundy-east- ern Gulf of Maine and are abun- dant during spring in the warmer, deeper waters of the central Scotian Shelf and areas of warm slope water intrusion along the Scotian Slope and the edges of Georges Bank (Stone and Jessop, 1992). In the Maritime provinces of Canada and Atlantic coastal United States, alewives and blueback herring, A. aestivalis, are fished commercially during their spring spawning migrations and are often marketed together as gaspereau or river herring. Little is known about the importance of ale- wives as predators in the marine environment or about their feeding habits and food consumption rates. Alewives are generally classified as size-selective, particulate and filter-feeding microphagists and can actively feed on individual zooplankton or passively feed by filtering the water with their gill rakers (Janssen, 1976; Durbin, 1979; James, 1988). Feeding mode depends on prey density, size, and visibility, and on predator size (Janssen, 1976, 1978a, 1978b; Durbin, 1979). The ability to switch feeding modes enables alewives to consume a wide size range of prey in a variety of environmental con- ditions. Size-selective predation by juvenile and nonanadromous fresh- water alewives can shift the species and size composition of zooplank- ton communities towards smaller forms (Brooks and Dodson, 1965; Brooks, 1968; Wells, 1970; Wars- haw, 1972; Vigerstad and Cobb, 1978). No information is available on size-selective predation in the ocean; however, in Minas Basin, a turbid macrotidal estuary, alewives were generally particulate feeders of larger, benthic prey rather than smaller pelagic prey (Stone and Daborn, 1987). Information on the feeding hab- its of anadromous alewives in the ocean is limited to qualitative as- sessments but is better known for freshwater juveniles (Vigerstad and Cobb, 1978; Gregory et al., 1983; Jessop, 1990) and estuarine resident subadults during summer (Stone and Daborn, 1987). Eu- phausiids, calanoid copepods and, to a lesser extent, hyperiid amphi- pods, chaetognaths, mysids, ptero- pods, decapod larvae, and salps 157 58 Fishery Bulletin 92|1|. 1994 have been identified as prey for alewives in conti- nental shelf waters from North Carolina to Nova Scotia (Holland and Yelverton, 1973; Edwards and Bowman, 1979; Neves, 1981; Vinogradov, 1984; Bow- man, 1986). However, none of these studies were comprehensive. We examined the stomach contents of anadromous alewives obtained from winter and summer ground- fish research surveys on the Scotian Shelf, Georges Bank, and in the Bay of Fundy to determine the importance of these regions as foraging areas for these fish. Seasonal, spatial, diel and size-related vari- ability in feeding are examined. Daily ration is esti- mated from information on diel feeding periodicity. Materials and methods Data collection Alewives were collected from seven groundfish re- search surveys conducted by the Canadian Depart- ment of Fisheries and Oceans in three regions (Georges Bank, central Scotian Shelf, and outer Bay of Fundy) during winter (February-March) and summer (July) over a two-year period (1990-91) (Table 1). All surveys used a Western IIA bottom trawl with a 10-mm stretched-mesh liner in the cod end. Thirty-minute tows at each sampling station were conducted throughout the 24-hour day. Up to 40 fish of representative size range from each set were frozen for later analysis. Bottom water tem- perature CO, time of tow deployment, latitude, lon- gitude, and bottom depth (m) were recorded for each set. Stomach content data were grouped by season and sample location: Winter-Fundy, Winter-Shelf, Winter-Georges, and Summer-Fundy (Fig. 1). Stone and Jessop (1992) provide additional details of the survey area and procedures, and seasonal distribu- tion of fish. Fork length (mm), weight (g), sex and species (de- termined by peritoneal colour (Leim and Scott, 1966)) were recorded for each fish. Whole digestive tracts, individually identified, were preserved in 4% buffered formalin. Diet analysis Stomachs were weighed (±0.01 g) and the contents ranked subjectively using a fullness code (0=empty, 1 = 12% full, 2=25% full, 3=50% full, 4=75% full, 5=100% full) and a digestion code ( l=finely digested, nothing recognizable; 2=medium digestion, some recognizable parts; 3=some digested, some undi- gested material; 4=undigested whole animals). The stomach content weight was obtained by subtract- ing the weight of the empty stomach from the total stomach weight. Stomach content weight, as a per- centage offish body weight C#BW), was used as an index of fullness to evaluate feeding activity and estimate daily ration. Stomach contents were iden- tified (to species where possible), enumerated, and the volume of each food type estimated by means of a points system (Swynnerton and Worthington, 1940; Stone and Daborn, 1987). For diet analysis, prey taxa (Table 2) were grouped into nine categories based on taxonomy and ecology: 1) euphausiids (Meganyctiphanes norvegica and some Thysanoessa spp.); 2) hyperiid amphipods (Parathemisto gaudichaudiy, 3) calanoid copepods (Calanus spp., Centrophoges spp. and Metridia spp.); 4) polychaetes (Nereis spp. and unidentifiable spe- cies); 5) fish larvae (Ammodytes dubius and uniden- tifiable species); 6) mysids {Neomysis americana); 1) Table 1 Stomach and fork length statistics, by season and geographic area, tained from groundfish research surveys conducted off Nova Scotia for al (1990 ewives, -1991). Al osa pseudoharengus. ob- Season and area Collection date Number Fork length (mm) 1990 1991 Sets Stomach Stomachs with prey Mean ± SD Range Winter-Fundy 2-10 Feb :i 112 58 201.9 ± 5.38 100-303 Winter-Georges 28 Feb-Mar 7 Feb 16-26 29 438 147 193.6 ± 1.83 118-305 Winter-Shelf 13-19 Mar Mar 15-18 29 489 322 223.8 ± 2.82 95-302 Summer-Fundy 6-10 Jul Jul 05-09 15 176 141 242.6 + 2.48 142-302 Total 82 1,215 668 213.6 ± 1.86 95-305 Stone and Jessop: Feeding habits of Alosa pseudoharengus 159 40°N 70°W Figure 1 Set locations for alewives, Alosa pseudoharengus, obtained from groundfish re- search surveys off the Atlantic coast of Nova Scotia (1990-91) grouped by season and geographic area. Offshore banks are delineated by the 100-m depth contour; the outer edge of the continental shelf is delineated by the 200-m depth contour. chaetognaths; 8) crustacean larvae (furciliae of Thysanoessa spp. and some decapod larvae); and 9) pteropods. The percent frequency of occurrence (%FO), percent of total stomach content number (%N), and percent of total stomach content volume (%V) of prey categories were estimated for stomachs containing recognizable food (digestion code >2). The Index of Relative Importance (IRI=(%N+%V) x <7rFO) was calculated for each prey category (Pinkas et al., 1971) and used for various diet comparisons. Diets were analyzed by season and geographic area (Win- ter-Fundy, Winter-Georges, Winter-Shelf, Suramer- Fundy), as well as by depth range within season and area, to compare food items from shallow regions and offshore banks <<100 m), mid-depths (101-200 m) and the shelf edge or deep basins (>200 m). Diel differences in diet composition (day and night, based on time of gear deployment) were examined for the entire data set. Ontogenetic differences in diet within season/area were examined by grouping fish lengths into four size classes (<151, 151-200, 201- 250 and >250 mm FL), which were assumed suffi- cient for detecting shifts in prey composition. Data from 1990 and 1991 were combined for all compari- sons because the ranks of IRI values for all prey categories between years were highly correlated (Spearman rank correlation coefficient (rs)=0.67; P<0.05; n=9). Predator-prey size analysis Total lengths (±1 mm, tip of rostrum to end of tel- son) of undigested, whole M. norvegica in the stom- achs of 55 alewives (>200 mm FL, since most intact prey occurred only in larger fish) from Winter- Georges, Winter-Shelf, and Summer-Fundy cruises were compared with predator size. Thysanoessa spp. were not measured because of poor condition. Lengths of M. norvegica from Emerald Basin col- lected in June 1991, by Sameoto et al. (1993) using the Bedford Institute of Oceanography Net and Environmental Sensing System (BIGNESS) were compared with euphausiid length frequencies from stomach contents to estimate the proportion of the available size range of M. norvegica consumed by alewives. The BIONESS is not considered to be size- selective for euphausiids (Sameoto et al., 1980). 160 Fishery Bulletin 92(1). 1994 Diel feeding periodicity and daily ration estimate Diel feeding periodicity and daily ration were exam- ined separately for winter (Bay of Fundy, Scotian Shelf, and Georges Bank combined) and summer (Bay of Fundy) collections because of seasonal dif- ferences in photoperiod. Stomach fullness data from tows within each successive 3-hour (winter cruises) and 4-hour (summer cruises) interval were grouped and assigned to the midpoint of the time period. Small sample sizes precluded grouping of summer collections into 3-hour intervals. Daily ration (DR) of alewives during winter and summer and by size class during winter (<151 mm, 151-200 mm, 201-250 mm, >250 mm) was esti- mated in terms of % body weight from the model of Elliott and Persson (1978): c _(*-*■-) a. 1-e ■Rt where the consumption of food (Ct) during the time interval t0 to tt is calculated from the average amount of food in the stomach at time t0 (S„), the amount in the stomach at time tt (S,) and the instan- taneous evacuation rate R. The estimates of Ct cal- culated for each time interval are then summed to give the total daily ration (DR). Feeding is assumed constant within each time interval. R is assumed exponential and temperature dependent (Elliott, 1972), as R = aehT. The slope (6) may be fairly constant for different prey types and fish species (mean=0.115), but the intercept (a) changes with prey type and can be estimated from gastric evacuation experiments (Durbin et al., 1983). Gastric evacuation rate data are unavailable for anadromous alewives; therefore, an intercept (a=0.0406) was obtained from Durbin et al. based on values for a variety of small invertebrates fed to several freshwater and marine fishes. High fat levels in the prey may retard evacu- ation (Durbin et al., 1983) but the principal food item in this study (M. norvegica) has a low lipid content (Ackman et al., 1970). Average bottom tem- peratures for winter (mean=7.16°C) and summer (mean=7.43°C) collections were used to estimate R. Statistical analysis Differences in the rankings of IRI values for prey categories (n =8 1 between three or more groups were tested for significance with the Kendall coefficient of concordance (w) (Siegel, 1956); for paired groups, the Spearman rank correlation coefficient (rj was used (Fritz, 1974). Euphausiids, which consistently ranked highest in importance in all comparisons, were excluded from correlation analysis to reduce bias and emphasize correlations among remaining prey groups. One-way ANOVA was used to examine feeding activity, represented by the index of fullness (arc- sine Vp transformed) by season and geographic area, by depth range within season and geographic area and by diel sampling period (winter and summer collections) and to compare total lengths of eu- phausiid prey. Paired means, adjusted for unequal sample sizes, were compared with the Tukey- Kramer test (Sokal and Rohlf, 1981). The relation between predator fork length and mean prey length was examined by linear regression for alewives with three or more M . norvegica present in their stomachs. Results Alewives examined for stomach contents measured 95 to 305 mm FL (mean=213.6 mm, n=l,215); fish from summer cruises in the Bay of Fundy were larger on average than those from other collections (Table 1). Capture depths ranged from 36 to 269 m, although most (75%) specimens were obtained from regions 101 to 200 m deep. Recognizable prey from over 20 different taxa oc- curred in 55% (668 of 1,215) of stomachs examined (Table 2). Over 95% of the total prey number, vol- ume, and frequency of occurrence were crustaceans (Table 2). Euphausiids were the most prevalent (91% by volume); Meganyctiphan.es norvegica were domi- nant by volume (61%) and furcilia larvae of Thysanoessa spp. were dominant numerically (32%). Other prey, including hyperiid amphipods, calanoid copepods, crustacean larvae, mysids, polychaetes, chaetognaths, pteropods, and fish larvae contributed little and varied temporally and spatially in relative importance. Diet composition by season and area Euphausiids were the most important food of ale- wives during winter and summer for all areas (Fig. 2). During winter, alewives from the outer Bay of Fundy and Georges Bank fed almost exclusively on euphausiids (99% and 95% of total volume, respec- tively). On Georges Bank, small (%V<3) proportions of calanoid copepods, hyperiid amphipods, and ptero- pods were also consumed. Prey diversity was great- est for alewives from the Scotian Shelf; euphausi- ids dominated by volume (82%) but were numeri- Stone and Jessop: Feeding habits of Alosa pseudoharengus 161 Table 2 Prey items found in the stomachs of alewives, Alosa pseudoharengus , collected from groundfi sh research surveys off Nova Scotia, 1990 -91. %FO = percent frequency of occurrence, %N = percent by number, %V = percent by volume. Prey item %FO %N %V Prey item %FO %N %v Crustacea 97.6 95.0 97.3 Decapoda 0.5 0.1 <0.1 Euphausiacea 91.3 72.4 91.0 Zoea 0.2 <0.1 <0.1 Meganyctiphanes norvegica 37.7 29.4 60.9 Megalopa 0.3 0.1 <0.1 Thysanoessa spp 6.9 4.5 6.0 Cirripedia Cypris larvae 0.2 <0.1 <0.1 Thysanoessa spp furcillia Unidentified Euphausiacea 3.7 40.1 32.1 6.3 1.2 23.0 Insecta Hymenoptera 0.5 <0.1 <0.1 Amphipoda 15.9 4.7 4.8 Polychaeta Nereis spp 1.8 1.1 0.1 0.1 0.5 0.4 Hyperiidea Parathemisto gaudichaudi Unidentified Hyperiidea 15.6 9.9 5.7 4.2 3.1 1.0 4.4 2.9 1.5 Unidentified Polychaeta Chaetognatha 0.8 1.1 <0.1 3.6 0.1 <0.5 Gammaridea Caprellidea 0.3 0.2 0.5 <0.1 0.4 <0.1 Hydrozoa 0.3 — <0.1 Gastropoda Pteropoda (Limacina) 5.1 0.8 0.3 Copepoda Calanoidea 8.2 17.4 1.2 Teleost larvae 3.9 0.5 1.4 Centrophages spp Calanus spp 3.1 0.5 1.3 0.7 0.2 <0.1 Ammodytes dubius Unidentified fish larvae 2.7 1.2 0.5 <0.1 1.0 0.4 Metridia spp Unidentified calanoids 2.0 6.9 1.1 14.3 <0.1 0.9 Algae Organic material 1.2 0.6 — 0.2 <0.1 Mysidacea Unidentified remains 6.6 — 0.8 Neomysis americana 0.2 0.4 0.2 Total stomachs with food Total prey number 668 14,752 Cumacea 0.3 <0.1 <0.1 Total prey volume (points) 25,232 cally less than in other areas. Hyperiid amphipods, (Parathemisto gaudichaudi), ranked second in im- portance (%V=10), followed by crustacean larvae (furciliae), calanoid copepods, and fish larvae, (Ammodytes dubius). During summer in the Bay of Fundy, alewives fed heavily on euphausiids (%V=95) but also consumed chaetognaths, mysids, and poly- chaetes (second, third, and fourth in importance). Rankings of IRI values (excluding euphausiids) for Winter-Georges, Winter-Shelf, and Summer-Fundy samples were not significantly correlated (u^O.22, P=0.701), indicating seasonal and geographic differ- ences in the dietary importance of these lesser prey categories. Winter-Fundy samples contained too few prey categories to be analyzed. Diet composition by depth range For Winter-Shelf and Summer-Fundy collections, the proportion of euphausiids in the diet increased with increasing depth (Fig. 3). At bottom depths less than 101 m on the Scotian Shelf, euphausiids com- posed 64% of total volume and 22% of total number; at 101 to 200 m, %V = 83 and %N = 23 and at depths greater than 200 m, %V = 96 and %N = 95. During summer in the Bay of Fundy, euphausiid consump- tion increased with depth such that at less than 101 m, %V = 82 and %N = 35; at 101 to 200 m, %V = 97 and %N = 97; while at depths greater than 200 m, both %V and %N = 100. Other prey categories gen- erally decreased in number with increasing depth as did their relative proportion. For both Winter- Shelf and Summer-Fundy collections, prey diversity and abundance were greatest where bottom depths were less than 101 m. Multiple correlations of IRI values for prey cat- egories (excluding euphausiids) between the three bottom-depth interval groups were not significant (u>=0.54, P=0.12) for Scotian Shelf collections and reflect the decreasing number of prey categories with increasing depth. For Summer-Fundy samples, the Spearman rank correlation of IRI values for the two shallower depth-intervals was not significant (rs=-0.35, P>0.05) and euphausiids were the only prey at depths greater than 200 m. Depth-related differences did not occur in the euphausiid-dominated diet of alewives from the Winter-Fundy and Winter-Georges collections at 162 Fishery Bulletin 92(1), 1994 Winter-Fundy (n = 58) Winter-Georges (n = 147) %N %v 1UU- 60- 20- j^L 20- 60- 100^ I I I I I I 1 1 Winter-Shelf %N %v n = 322) 100- 60- 20- y 20- 60- 100 i I I I I I I I %N %v Summer-Fundy (n - 141) OO-i Legend Eup 60- Cop Pter Amp 20- I CruLar 20- =1 Chaet M Mys 1 Poly 60- 00- 1 1 1 1 1 H- l=J !0% FO t— I = 20% FO Figure 2 Relative importance of prey categories in the diet of alewives, Alosa pseudoharengus, collected from groundfish research surveys off Nova Scotia (1990-91), ranked from highest Index of Relative Im- portance (left to right) by season and area, n = number of stom- achs with prey; '#FO = % frequency of occurrence; %N = % of total prey number; %V = % of total prey volume; Eup = euphausiids; Cop = calanoid copepods; Pter = pteropods; Amp = hyperiid amphipods; CruLar = crustacean larvae; FishLar = fish larvae; Chaet = cha- etognaths; Mys = mysids; Poly = polychaetes. bottom depths exceeding 101 m (no fish were ob- tained at bottom depths less than 101 ml. IRI rankings of" prey categories between depth groups r Georg nk collections were highly correlated • .=0 !9, P fi.01 ' / prey i igorie n present for analysis of Winter-Fundy collections. In both winter and summer, most euphausiids con- med at depths less than 101 m were Thysanoessa pp. v than M. m vegica a) f, r shallower regions < (Table 3). el vari< t t h ou gh i numbers and volumes were ingested during the day (7rN=74, %V=92) than at night (r/rN=16, %V=85) (Fig. 4). IRI values for day and night collections were not significantly correlated (rs=0.26, P>0.05) reflecting the greater consumption of hyperiid am- phipods during the day and copepods, crustacean larvae and fish larvae at night. Diet composition by size class Diet composition was relatively homo- geneous among alewife size groups (<151 mm, 151-200 mm, 201-250 mm, >250 mm) with euphausiids com- posing most of the total food volume (Fig. 5). Multiple correlations of IRI values for prey categories (excluding euphausiids) by fish length group were significant for both the Scotian Shelf (u;=0. 58, P=0.024) and Georges Bank (w=0.65, P=0.011). For Sum- mer-Fundy collections, diets of the two largest size groups were nearly identical; IRI values were not signifi- cantly correlated (rs=0.38, P>0.05) due to slight differences in the rankings of minor prey categories (i.e., amphipods, mysids, polychaetes, chaetognaths). Prey size composition Alewives ingested similar sizes of M. norvegica during winter (Georges Bank, Scotian Shelf) and summer (Bay of Fundy) (Fig. 6). Modal peaks in euphausiid size appeared at 25-27 mm and 30 mm on the Scotian Shelf and at 30-35 mm for Georges Bank and the Bay of Fundy. In com- parison, M. norvegica from Emerald Basin BIONESS collections in June 1991 were bimodally ibut( it 25-27 mm and 34 mm. Euphausiids larger than 29 mm were proportionately less fre- quent than in stomach contents. Mean lengths of M. norvegica consumed by ale- vives varied by season/area group (F._, 7II|=65.5, P<0.001), although differences between means were small (Winter-Georges: mean=32.1±3.13; Winter- Shelf: mean=28.7±3.72; Summer-Fundy: mean ll.2±3.64). The average size of euphausiids con- d did not differ (F, 50=3.31, P =0.075) with ale- wife stum;. I tion -ize i ran: 225-300 mm FL). Stone and Jessop: Feeding habits of Alosa pseudoharengus 163 Feeding activity Feeding activity, as indicated by mean stomach fullness index values, varied by season/geo- graphic area (F3 1910=46.20, P< 0.001). Mean stomach fullness was highest for Summer-Fundy and Winter-Shelf collections and lowest for Winter-Fundy and Winter-Georges collections (Ta- ble 4). The proportion of feeding fish was highest during summer in the Bay of Fundy (80.6%) and lowest during winter on Georges Bank (33.6%). Stomach fullness was significantly higher at bot- tom depths greater than 200 m for all but the Winter-Shelf col- lections, where mean fullness did not differ among depth groups (Table 4). Similarly, the proportion of feeding fish was highest in areas exceeding 200 m deep for all collections. Alewife feeding activity varied throughout the diel period dur- ing winter (F1 1(m=24.97, P< 0.001) and summer (F5 196= 7.98, P<0.001) with maximum full- %N %V Winter-Shelf 100-| < 101 m (n = 49) 60 20- 20- 60 100 Summer-Fundy 100-1 < 101 m (n = 33) Chaot %N %v 60 20- 20" 60 100 Jzj_P" %N %v 100- 101-200 m (n = 262) 60- fl j L 20" 60- 1 1 1 I I 1 i i %N %V 100- 101-200 m 200 m (n = 11) 60- 20- M 20" 60- - 100- I I I I i i i %N %V > 200 m (n = 36) 60- 20- 20-1 60- 1 'III 1 1 1 i = 20% FO i = 20% FO Figure 3 Relative importance of prey categories in the diet of alewives, Alosa pseudoharengus, obtained from groundfish research surveys off Nova Scotia (1990-91), ranked from highest Index of Relative Importance (left to right), by depth range, for Scotian Shelf (winter) and Bay of Fundy (summer) collections. (Symbols as in Fis;. 2). at 7.43°C during summer (Table 5). The winter daily ration of alewives generally decreased from 1.95% BW for fish less than 151 mm FL to 1.13% BW at 151-200 mm FL, 1.19% BW at 201-200 mm FL and 1.00% BW at larger than 250 mm FL. Discussion Our study clearly indicates that alewives off Nova Scotia feed primarily on euphausiids, particularly Meganyctiphanes norvegica; much smaller contribu- tions are made by other prey. Alewives from the 164 Fishery Bulletin 92(1), 1994 Table 3 Mean number of Meganyctiphanes norvegi :a and Thysanoessa spp. in the stomachs of alewives , Alosa pseudoh irengus , by depth interval within season and geogra phic area from groundfi sh researc h surveys off Nova Scotia (1990-91). n = nx. mber of s tomach s with prey. Depth M in i vegica Thysanoessa spp . Season and area (mi Mean t SD n Mean + SD n Winter-Fundy 101-200 11.3 + 7.36 3 5.4 + 0.81 5 >200 10.2 + 2.20 9 2.5 i 1.50 2 Winter-Georges 101-200 5.9 ± 0.81 26 11.8 * 6.55 6 >200 20.3 + 1.68 28 — — — Winter-Shelf <101 — — — 32.0 + 14.63 10 101-200 14.7 + 1.41 89 9.6 ± 3.79 20 >200 5.8 1 3.47 1 — — — Summer-Fundv <101 18.9 + 3.47 14 12.0 ± 2.00 2 101-200 21.9 - 2.67 48 — — — >200 27.5 ± 2.39 31 23.0 — 1 100-1 %N %v 100 Night (n = 290) %N %V Legend Eup 100 100 "i — i — i — r i = 20% FO Figure 4 Relative importance of prey categories in the diet of alewives, Alosa pseudoharengus, obtained from groundfish research surveys off Nova Scotia ( 1990-91), ranked from highest Index of Relative Importance (left to right) for day and night collections (Symbols as in Fig. 2). Atlantic seaboard of the United States consumed relatively fewer euphausiids 137-56% by weight) (Edwards and Bowman, 1979; Vinogradov, 1984) than off Nova Scotia (82-99% by volume). Euphausiids represent a large component of the marine zooplankton community and are abundant in the Bay of Fundy (Kulka et al., 1982; Locke and Corey, 1988), Gulf of Maine (Bigelow, 1926), the deep basins of the Scotian Shelf (Herman et al., 1991) and the outer shelf and shelf slope (Sameoto, 1982). Given their two-year life cycle (Hollingshead and Corey, 1974; Berkes, 1976), the availability and rela- tive abundance of euphausiids is more seasonally stable than for other prey spe- cies (i.e., chaetognaths, hyperiid amphipods, calanoid copepods, mysids), most of which undergo fluctuations in abundance progressing from a spring low to a summer high before declining in fall and win- ter (Evans, 1968; Sherman and Schaner, 1968; Corey, 1988; McLaren et al., 1989). Small seasonal differences in diet composition reflect the op- portunistic foraging behaviour of alewives and the availability of food types from offshore re- gions during winter as com- pared with the Bay of Fundy in summer. During winter, the diet diversity of alewives was greatest on the Scotian Shelf probably because the late win- ter (mid-March) sampling period co- incides with the hatching and occur- rence of the larval forms of Thy- sanoessa spp. (Berkes, 1976) and Ammodytes dubius (Scott, 1972), both of which occurred only in the diet of alewives from the Scotian Shelf. In the Bay of Fundy, alewife consumption of chaetognaths and mysids in the summer reflects their increased abundance and availabil- ity (Sherman and Schaner, 1968; Corey, 1988). The increased proportion of eu- phausiids in the diet of alewives from the Scotian Shelf (winter) and the Bay of Fundy (summer) coin- cides with an increased relative abundance of euphausiids with in- creasing depth. In the Scotian Shelf Basins, M. norvegica occur between 170 m and the bottom with highest concentrations generally below 200 m (Sameoto et al., 1993). In the Bay of Fundy, M. norvegica is most abundant where bottom depths are between 165 and 200 m, while Thysanoessa inermis occur between 95 and 155 m (Kulka et al., 1982). The greater proportion and number of other prey categories at depths less than 101 m on the Scotian Shelf and in the Bay of Fundy likely result from decreased euphausiid abundance (thereby in- creasing the relative contribution of other prey) rather than an absolute increase in the abundance of other zooplankters. Depth-related variation in Stone and Jessop: Feeding habits of Alosa pseudoharengus 165 euphausiid species composition in the diet of alewives from all regions matches differences in the bottom depth preferences of M. norvegicci (>150 m) and Thysanoessa spp. (100- 150 m) (Berkes, 1976; Kulka et al., 1982; Sameoto et al., 1993). Diel differences in the diet of ale- wives may reflect the influence of vary- ing light intensity on prey availability and on their relative success in locat- ing and capturing prey. Consumption of microzooplankters (crustacean larvae, calanoid copepods) was greater at night perhaps because of increased filter- feeding activity (Janssen, 1978b). Con- versely, ingestion of macrozooplankters (euphausiids, hyperiid amphipods) may be highest during the day because visual cues favour a particulate-feeding mode. Large size, darkly pigmented eyes, and a habit of forming large concentra- tions (Mauchline and Fisher, 1969) may make M. norvegica easily detect- able by alewives during daylight whereas at night, photophores along the abdomen of M. norvegica may as- sist detection. Most euphausiid species migrate vertically over the diel period, rising from deep water (150-200 m) towards the surface at dusk, remaining near surface throughout the night, and then migrating to the depths at dawn (Mauch- line, 1984). Alewives also have a diel pattern of vertical migra- tion in the marine environment (Neves, 1981; Stone and Jessop, 1992) and may encoun- ter sufficient light higher in the water column at night to par- ticulate feed on euphausiids. Ontogenetic differences in diet composition were not ap- parent; euphausiids dominated the diet of alewives ranging in length from 95 to 305 mm. Ale- wives switch from feeding pri- marily on microzooplankton to macrozooplankton at some point during their marine de- velopment and like other simi- larly sized clupeids, concentrate their feeding at intermediate trophic levels (James, 1988). Winter- Shelf 100 E O > CD 1) u n. 80 GO 40 20 ■ 35 12 141 134 .;;:;' Winter- Georges 13 65 34 3S Summer- Fundy 66 S3 Legend Eup Cop Ptar Amp CruLar FIshLar Chaet Mys Poly B C C D Predator length group c D Figure 5 Percentage of total volume of prey categories in the diet of ale- wives, Alosa pseudoharengus, for different size classes (mm FL) obtained from groundfish research surveys off Nova Scotia (1990- 91). Euphausiids were the only prey category in Winter-Fundy cruises. A: <151 mm; B: 151-200 mm; C: 201-250 mm; D: >250 mm; Eup = euphausiids; Cop = calanoid copepods; Pter = ptero- pods; Amp = hyperiid amphipods; CruLar = crustacean larvae; FishLar = fish larvae; Chaet = chaetognaths; Mys = mysids; Poly = polychaetes; n = number of stomachs with food. Winter-Georges Winter-Shelf Summer-Fundy BIONESS (n = 257) (n - 269) (n = 178) (n = 785; 25 30 35 Total length (mm) 45 Figure 6 Size frequency distributions of M. norvegica consumed by alewives, Alosa pseudoharengus, obtained from winter (Georges Bank. Scotian Shelf) and summer (Bay of Fundy) groundfish surveys off Nova Scotia (1990-91) and from BIONESS samples in Emerald Basin (Spring, 1991 ). n = sample size. 166 Fishery Bulletin 92(1), 1994 Table 4 Mean stomach fullness index (arcsine Vp transformed) by season and geographic area and by depth interva for al ewives, Alosa pseudoharengus, obtained from groundfish research surveys off Nova Scotia (1990-91) Mean fullness inde x values lacking a letter in com- mon are significantly differe nt (Tukey HSD, P<0.05). n = number of stomachs examined (including empty stomach s). Full ness index (%BW) % with Season and area Depth (m) n food Mean i SD Maximum Winter-Fundy all 112 51.8 2.3z i 0.25 9.9 Winter-Georges all 438 33.6 2.1z ± 0.13 9.9 Winter-Shelf all 489 65.0 3.8y ± 0.10 10.0 Summer-Fundy all 175 80.6 3.9y + 0.21 10.0 Winter-Fundy 101-200 60 28.3 1.5z t 0.29 9.4 >200 52 78.8 3.7y + 0.32 9.9 Winter-Georges <101 7 28.6 2.1z i 0.87 5.9 101-200 376 28.7 1.7z t 0.12 9.9 >200 55 67.3 4.6y * 0.46 9.9 Winter-Shelf <101 92 55.3 3.4z i 0.16 7.4 101-200 385 68.1 3.9z + 0.12 10.0 >200 12 91.7 3.4z + 0.46 8.0 Summer-Fundy <101 48 liS s 3.5z + 0.33 8.8 101-200 87 82.8 3.6z + 0.30 10.0 >200 40 90.1 5.0y t 0.51 9.8 Gilmurray (1980) found mainly microplanktonic prey (e.g., calanoid copepods, cypris larvae, insects) in the diet of alewives less than 80 mm FL obtained from tidal creeks in the upper Bay of Fundy. The shift towards consumption of macrozooplankton likely occurs at fish sizes smaller than those examined in the present study (i.e., <95 mm FL). Diel feeding activity during winter and summer, as indicated by the mean fullness index, reached a maximum near mid-day and is typical of size-selec- tive predators which rely on visual cues (Eggers, 1977). Summer resident subadult alewives in Minas Basin display a similar feeding pattern, although peak feeding occurred later in the afternoon (1500 hours), coincident with the time of high tide when turbidity was lowest and prey visibility highest (Stone, 1985). Summer feeding activity by juvenile anadromous alewives in freshwater also peaks dur- ing the day but ceases or declines overnight ( Jessop, 1990). Nocturnal feeding by alewives was more ap- parent during winter than summer; the significance of this seasonal difference in feeding activity is un- clear. Alewives can and do feed efficiently at night using both particulate (Janssen and Brandt, 1978) and filter-feeding (Janssen, 1978b) modes. Alewives greater than 200 mm FL generally consumed the largest Meganyctiphanes avail- able. Length-frequency distri- butions of M. norvegica, which has a life span of about two years, are typically bimodal (Hollingshead and Corey, 1974; Berkes, 1976). Alewives selec- tively favor larger prey (Brooks and Dodson, 1965; Brooks, 1968; Wells, 1970) and likely use a particulate feeding strat- egy in doing so. Slight seasonal and geographic differences in the average size of M. norvegica ingested likely reflect size differences in euphausiid populations rather than selec- tion by the predator. Daily ration calculations were based on the model of El- liott and Persson (1978) which was originally intended for field samples collected within a given area from the same popu- lation over time. Our stomach fullness data for alewives from the Bay of Fundy, Georges Bank, and the Scotian Shelf covered a wide area geographically and may involve more than one popu- lation. The broad temporal and spatial coverage reduces the effect of day-to-day and regional varia- tions in diet which would arise from more restricted sampling. Calculated daily ration levels for alewives off Nova Scotia were similar to those reported for other teleosts (Fange and Grove, 1979). Lower esti- mates were obtained during winter (1.22% BW at 7.16°C) than for summer ( 1.88% BW at 7.43°C) since temperature is related to metabolic requirements and to the evacuation rate of stomach contents (Durbin et al., 1983). Both estimates are well above maintenance ration levels for temperatures in the 7-8°C range and are sufficient for positive growth (Brett and Groves, 1979). Alewife daily ration de- clined with increasing fish size; small fish, includ- ing marine species such as North Sea cod, Gadus morhua (Daan, 1973), winter flounder, Pseudo- pleuronectes arnericanus (Huebner and Langton, 1982) and silver hake, Merluccius bilinnearis (Durbin et al., 1983), generally consume proportion- ally more food per unit weight than large fish (Windell, 1978). Overall, our estimates of daily Stone and Jessop: Feeding habits of Alosa pseudoharengus 167 Table 5 Mean amount of food (%BW) in the stomachs of alewives, Alosa pseudoharengus , obtained from groundfish surveys off Nova Scotia (1990-91), with estimates of food consumption (C.) and daily ration {DR = XCJ, by season and size c lass, n = number of stomachs examined (including empty stomachs. For winter collections, R = 0.0925, temperature = 7.16°C; 'or summer collections, R = 0.0954, tern perature = 7.43°C. Stomac h contents (%BWl Season (size class) Time period (hr) c, (%BW) DR n Mean + SD (%BW) Winter (all) 2400-0300 84 0.75 + 0.100 1.216 0300-0600 184 0.65 * 0.053 0.098 0600-0900 97 0.23 ± 0.038 -0.308 0900-1200 122 0.52 + 0.079 0.401 1200-1500 96 1.09 i 0.104 0.792 1500-1800 150 0.20 ( 0.032 -0.709 1800-2100 177 0.52 ± 0.043 0.421 2100-2400 189 0.42 i- 0.053 0.033 0.488 Summer (all) 2400-0400 11 0.22 i 0.041 1.880 0400-0800 5 0.58 + 0.198 0.515 0800-1200 61 2.32 ± 0.161 2.316 1200-1600 74 1.32 • 0.190 -0.320 1600-2000 19 0.48 i 0.249 -0.508 2000-2400 5 0.03 + 0.031 -0.357 0.234 Winter 2400-0400 29 0.95 .i 0.233 1.949 (<151 mm FL) 0400-0800 31 1.13 ± 0.146 0.563 0800-1200 4 0.90 + 0.382 0.143 1200-1600 3 1.32 + 0.439 0.842 1600-2000 13 0.42 ± 0.110 -0.593 2000-2400 87 0.55 i 0.103 0.311 Winter 2400-0300 22 1.10 + 0.199 1.126 (151-200 mm FL) 0300-0600 IS 0.61 + 0.095 -0.253 0600-0900 26 0.12 + 0.048 -0.396 0900-1200 29 0.22 i 0.063 0.151 1200-1500 9 0.84 ± 0.532 0.765 1500-1800 119 0.14 t 0.039 -0.565 1800-2100 26 0.74 t 0.153 0.719 2100-2400 25 0.16 * 0.053 -0.457 Winter 2400-0300 30 0.52 ± 0.094 1.189 (201-250 mm FL) 0300-0600 60 0.62 i 0.093 0.256 0600-0900 23 0.27 + 0.062 -0.222 0900-1200 :.] 0.53 * 0.123 0.372 1200-1500 27 0.93 i 0.192 0.600 1500-1800 56 0.31 + 0.065 -0.453 1800-2100 42 0.61 + 0.075 0.434 2100-2400 36 0.50 + 0.093 0.046 0.156 Winter 2400-0300 14 0.24 t 0.052 1.000 O250 mm FL) 0300-0600 34 0.32 + 0.042 0.161 0600-0900 17 0.27 t 0.067 0.029 0900-1200 39 0.68 + 0.173 0.545 1200-1500 57 1.19 ± 0.124 0.772 1500-1800 22 0.11 I 0.032 -0.902 1800-2100 39 0.31 t 0.076 0.257 2100-2400 11 0.50 * 0.141 0.302 168 Fishery Bulletin 92(1], 1994 x •a c 200 m) and temperature, were suitable for M. norvegica (Kulka et al., 1982; Sameoto et al., 1993). Alewives prefer bottom temperatures of 7— 11°C off- shore at mid-depths in spring ( 101-183 m), in shal- lower nearshore waters in summer (46-82 m) and in deeper offshore waters in fall ( 119-192 m) (Stone and Jessop, 1992). During winter, Meganyctiphanes seeks deeper, warmer water rather than the cold upper layers (Bigelow, 1926; Hollingshead and Corey, 1974). While the seasonal pattern of move- ment by alewives (inshore and northward during spring and offshore and southward during fall) is partially linked with spawning migrations, it is apparent that their marine distribution is also in- fluenced by the distribution, availability, and abun- dance of their main prey, M. norvegica. Acknowledgments We thank D. Sameoto and R. Cutting for critically reviewing earlier drafts of the manuscript. We also wish to thank M. Strong, P. Fanning, and J. Martell for collecting the alewives used in our analyses, S. Wilson and J. Tremblay for taxonomic assistance, and D. Ingraham for helping with laboratory work. Literature cited Ackman, R. G., C. A. Eaton, J. C. Sipos, S. N. Hooper, and J. D. Castell. 1970. Lipids and fatty acids of two species of North Atlantic krill (Meganyctiphanes norvegica and Thysanoessa inermis) and their role in the aquatic food web. J. Fish. Res. Board Can. 27:513-533. Berkes, F. 1976. Ecology of euphausiids in the Gulf of St. Lawrence. J. Fish. Res. Board Can. 33:1894- 1905. Bigelow, H. G. 1926. Plankton of the offshore waters of the Gulf of Maine. Bull. U.S. Bur. Fish. 40:1-509. Bigelow, H. B., and W. C. Schroeder. 1953. Fishes of the Gulf of Maine. Bull. U.S. Fish Wildl. Serv. 74, 577 p. Bowman, R. E. 1986. Effect of regurgitation on stomach content data of marine fishes. Env. Biol. Fish. 16:171-181. Stone and Jessop: Feeding habits of Alosa pseudoharengus 169 Brett, J. R., and T. D. D. Groves. 1979. Physiological energetics. In W. S. Hoar, D. J. Randall, and J. R. Brett (eds.), Fish physiology, Vol. 8, p. 280-344. Academic Press, New York. Brooks, J. L. 1968. The effects of prey-size selection by lake planktivores. Syst. Zool. 17:272-291. Brooks, J. L., and S. I. Dodson. 1965. The effect of a marine planktivore on lake plankton illustrates theory of size, competition and predation. Science 150:28-35. Corey, S. 1988. Quantitative distributional patterns and as- pects of the biology of the Mysidacea (Crustacea: Peracarida) in the zooplankton of the Bay of Fundy region. Can. J. Zool. 66:1545-1552. Daan, N. 1973. A quantitative analysis of the food intake of North Sea cod, Gadus morhua. Neth. J. Sea Res. 6:479-517. Durbin, A. G. 1979. Food selection by plankton feeding fishes. In H. Clepper (ed.), Predator-prey systems in fisheries management, p. 203-218. Sport Fish- ing Inst., Washington D.C. Durbin, E. G., A. G. Durbin, R. W. Langton, and R. E. Bowman. 1983. Stomach contents of silver hake, Merluccius bilinnearis, and Atlantic cod, Gadus morhua, and estimation of their daily rations. Fish. Bull. 81:437-454. Edwards, R. L., and R. E. Bowman. 1979. Food consumed by continental shelf fishes. 87-409 In H. Clepper (ed.), Predator-prey sys- tems in fisheries management, p. 387^09. Sport Fishing Inst., Washington D.C. Eggers, D. M. 1977. Factors in interpreting data obtained by diel sampling of fish stomachs. J. Fish. Res. Board Can. 34:290-294. Elliott, J. M. 1972. Rates of gastric evacuation in brown trout, Salmo trutta L. Freshwater Biol. 2:1-18. Elliott, J. M., and L. Persson. 1978. The estimation of daily rates of food con- sumption for fish. J. Anim. Ecol. 47:977-991. Evans, F. 1968. Development and reproduction of Parathemisto gracilipes (Norman) (Amphipoda, Hyperiidea) in the North Sea. Crustaceana 15:101-109. Fange, R., and D. Grove. 1979. Digestion. In W. S. Hoar, D. J. Randall, and J. R. Brett (eds.), Fish physiol., Vol. 8, p. 162- 241. Academic Press, New York. Fritz, E. S. 1974. Total diet comparison in fishes by Spearman rank correlation coefficients. Copeia 1974: 210-214. Gilmurray, M. C. 1980. Occurrence and feeding habits of some juve- nile fish in the southern bight of Minas Basin, Nova Scotia, 1979. Master's thesis, Acadia Univ., Wolfville, Nova Scotia, 100 p. Gregory, R. S., G. S. Brown, and G. R. Daborn. 1983. Food habits of young anadromous alewives (Alosa pseudoharengus) in Lake Ainslie, Nova Scotia. Can. Field Nat. 97:423-426. Herman, A. W., D. D. Sameoto, C. Shunnian, W. R. Mitchell, B. Petrie, and N. Cochrane. 1991. Sources of zooplankton on the Nova Scotia Shelf and their aggregations within deep shelf basins. Cont. Shelf Res. 11:211-238. Holland, B. J., and G. F. Yelverton. 1973. Distribution and biological studies of anadro- mous fishes offshore North Carolina. N.C. Dep. Nat. Res. Spec. Sci. Rep. No. 24, 32 p. Hollingshead, K. W., and S. Corey. 1974. Aspects of the life history of Meganyctiphanes norvegica (M. Sars), Crustacea (Euphausiacea), in Passamaquoddy Bay. Can. J. Zool. 52:495-505. Huebner, J. D., and R. W. Langton. 1982. Rate of gastric evacuation for winter floun- der, Pseudopleuronectes americanus. Can. J. Fish. Aquat. Sci. 39:356-360. James, A. G. 1988. Are clupeid macrophagists herbivorous or omnivorous? A review of the diets of some commer- cially important clupeids. S. Afr. J. Mar. Sci. 7:161-177. Janssen, J. 1976. Feeding modes and prey size selection in the alewife (Alosa pseudoharengus). J. Fish. Res. Board Can. 33:1972-1975. 1978a. Feeding-behaviour repertoire of the alewife, Alosa pseudoharengus, and the ciscoes, Coregonus hoyi and C. artedii. J. Fish. Res. Board Can. 35:249-253. 1978b. Will alewives (Alosa pseudoharengus) feed in the dark? Env. Biol. Fish. 3:239-240. Janssen, J., and S. B. Brandt. 1978. Feeding ecology and vertical migration of adult alewives (Alosa pseudoharengus) in Lake Michigan. Can. J. Fish. Aquat. Sci. 37:177-184. Jessop, B. M. 1990. Diel variation in density, length composition, and feeding activity of juvenile alewife, Alosa pseudoharengus Wilson, and blueback herring, A. aestivalis Mitchell, at near-surface depths in a hydroelectric dam impoundment. J. Fish Biol. 37:813-822. Kulka, D. W., S. Corey, and T. D. Isles. 1982. Community structure and biomass of eu- phausiids in the Bay of Fundy. Can. J. Fish. Aquat. Sci. 39:326-334. Leim, A. H., and and W. B. Scott. 1966. Fishes of the Atlantic coast of Canada. Fish. Res. Board Can. Bull. 155, 485 p. 170 Fishery Bulletin 92(1], 1994 Locke, A., and S. Corey. 1988. Taxonomic composition and distribution of Euphausiacea and Decapoda (Crustacea) in the neuston of the Bay of Fundy, Canada. J. Plank- ton Res. 10:185-198. Mauchline, J. 1984. Euphausiid, stomatopod and leptostracan crustaceans. E. J. Brill and W. Backhuys, Lon- don, 91 p. Mauchline, J., and L. R. Fisher. 1969. The biology of euphausiids. In F. S. Russel and M. Yonge (eds.), Advances in marine biology, Volume 7. Academic Press, London, 454 p. McLaren, I. A., M. J. Tremblay, C. J. Corkett, and J. C. Roff. 1989. Copepod production on the Scotian Shelf based on life-history analyses and laboratory rear- ings. Can. J. Fish. Aquat. Sci. 46:560-583. Neves, R. J. 1981. Offshore distribution of alewife, Alosa pseudoharengus and blueback herring, A. aestivalis, along the Atlantic coast. Fish. Bull. 79:473-485. Pinkas, L., M. S. Oliphant, and I. L. K. Iverson. 1971. Food habits of albacore, bluefin tuna and bo- nito in California waters. Calif. Fish and Game Fish. Bull. 152:1-105. Sameoto, D. D. 1982. Zooplankton and micronekton abundance in accoustic scattering layers on the Nova Scotian slope. Can. J. Fish. Aquat. Sci. 39:760-777. Sameoto, D. D., L. O. Jaroszynski, and W. B. Fraser. 1980. BIONESS, a new design in multiple net zoop- lankton samplers. Can. J. Fish. Aquat. Sci. 37:722-724. Sameoto, D., N. Cochrane, and A. Herman. 1993. Convergence of accoustic, optical, and net- catch estimates of euphausiid abundance: use of artificial light to reduce net-avoidance. Can. J. Fish. Aquat. Sci. 50:334-346. Scott, J. S. 1972. Eggs and larvae of northern sand lance tAmmodytes dubius) from the Scotian Shelf. J. Fish. Res. Board Can. 29:1667-1671. Sherman, K., and E. G. Schaner. 1968. Observations on the distribution and breed- ing of Sagitta elegans (Chaetognatha) in coastal waters of the Gulf of Maine. Limnol. Oceanogr. 13:618-625. Siegel, S. 1956. Nonparametric statistics for the behavioral sciences. McGraw-Hill Book Co., Toronto, 312 p. Sokal, R. R., and F. J. Rohlf. 1981. Biometry, 2nd ed. H. Freeman, San Fran- cisco, 859 p. Steedman, H. F. 1976. General and applied data on formaldehyde fixation and preservation of marine zoo- plankton. In H. F Steedman (ed.), Zooplankton fixation and preservation, p. 103-154. Unesco Press, Paris, 350 p. Stone, H. H. 1985. Composition, morphometric characteristics and feeding ecology of alewives (Alosa pseudoharengus) and blueback herring (A. aestivalis) (Pisces: Clupeidae) in Minas Basin. Master's thesis, Acadia Univ., Wolfville Nova Scotia, 191 p. Stone, H. H., and G. R. Daborn. 1987. Diet of alewives, Alosa pseudoharengus and blueback herring, A. aestivalis (Pisces: Clupeidae) in Minas Basin, Nova Scotia, a turbid macrotidal estuary. Env. Biol. Fish. 19:55-67. Stone, H. H., and B. M. Jessop. 1992. Seasonal distribution of river herring Alosa pseudoharengus and A. aestivalis off the Atlantic coast of Nova Scotia. Fish. Bull. 90:376-389. Swynnerton, G. H., and E. B. Worthington. 1940. Notes on the food of fish in Haweswater (Westmoorland). J. Anim. Ecol. 9:183-187. Vigerstad, T. J., and J. S. Cobb. 1978. Effects of predation by sea-run juvenile ale- wives ( Alosa pseudoharengus) on the zooplankton community at Hamilton Reservoir, Rhode Island. Estuaries 1:36-45. Vinogradov, V. I. 1984. Food of silver hake, red hake and other fishes on Georges Bank and adjacent waters, 1968- 74. NAFO Sci. Counc. Studies 7:87-94. Warshaw, S. J. 1972. Effects of alewives (Alosa pseudoharengus) on the zooplankton of Lake Warskopmic, Con- necticut. Limnol. Oceanogr. 17:816-825. Wells, L. 1970. The effects of alewife predation on zooplankton in Lake Michigan. Limnol. Oceanogr. 15:556-565. Windell, J. T. 1978. Digestion and the daily ration of fishes. In T Bagenal (ed.), Fish production in fresh waters, p. 227-254. Blackwell Scientific Pubis., London, 365 p. Abstract. A survey of queen conch [Strombus gigas) popula- tions near Lee Stocking Island, Exuma Cays, Bahamas, showed that 74% of all adults were on the narrow island shelf adjacent to the Exuma Sound, in 10-18 m of wa- ter. None were found deeper than 25 m, and relatively few adults were found shallower than 10 m. Numbers of juveniles were great- est on the Great Bahama Bank and decreased with increasing depth on the island shelf. No juve- niles were found in shelf regions greater than 15 m in depth. Pat- terns of shell morphology, which were related to growth rates in juveniles, suggest that adults that mature on the Great Bahama Bank rarely move to deep water, and that the most important sources for deep-water stocks are small, nearshore nurseries on the island shelf. The mostly unfished deep-water populations are prob- ably now the primary source of larvae for queen conch in the Exuma Cays. Because virtually all of the conch are within the limits of SCUBA diving, it will be impor- tant to identify and to protect criti- cal nursery habitats for reproduc- tive stocks. Queen conch, Strombus gigas, reproductive stocks in the central Bahamas: distribution and probable sources Allan W. Stoner Kirsten C. Schwarte Caribbean Marine Research Center. 805 E 46th Place Vera Beach, Florida 32963 Manuscript accept 8 September 1993 Fishery Bulletin 92:171-179 (1994) Queen conch (Strombus gigas), once abundant throughout the Car- ibbean region, have been fished to near extinction or to a level at which there is no longer a viable fishery in many localities (Appel- doorn et al., 1987; Berg and Olsen, 1989). This is particularly true in nations where the fishery has been open to SCUBA divers. Stock depletion resulted in at least tem- porary closures of the conch fishery in Bermuda, Florida, Cuba, Bon- aire, and the U.S. Virgin Islands. Regulations including size limits, catch quotas, gear restrictions, and closed areas have been instituted in other countries. This study was conducted in an attempt to understand reasons for the rapid depletion of queen conch populations in the Caribbean re- gion, and to evaluate the signifi- cance of deep-water conch stocks. Several authors have suggested that these deep-water conch, living beyond the normal range of free divers, are the primary source of larvae for shall-water populations and the fishery (Berg and Olsen, 1989; Wicklund et al., 1991; Stoner et al., 1992; Stoner and Sandt, 1992). Therefore, we surveyed the density and age structure of queen conch in the vicinity of Lee Stock- ing Island in the central Bahamas. Differences in shell morphology and growth rate between conch found on Great Bahama Bank and on the windward island shelf adja- cent to Exuma Sound were used as indicators of geographic source for reproductive stocks. The impor- tance of deep-water populations is discussed in terms of fisheries management. Methods and materials Study site An assessment of the adult conch population was conducted between 1989 and 1991 in a 12-km long sec- tion of the Exuma Cays, central Bahamas, adjacent to Lee Stocking Island (Fig. 1). To the west and south of the Cays lies the Great Bahama Bank, a shallow, sand- and seagrass-covered platform that extends to the Tongue of the Ocean. To the east and north is a narrow (1-2 km) island shelf, a steep shelf- break beginning at an approxi- mately 30-m depth, and the deep Exuma Sound. Great Bahama Bank in the re- gion of the study site is character- ized by strong tidal currents that carry oceanic water from Exuma Sound onto the bank through chan- nels between the islands. Approxi- mately 909r of the bank area is less than 3.5 m deep; the remainder is tidal channels with depths to 8 m near the inlets and between the Brigantine Cays. For this study the I 71 172 Fishery Bulletin 92(1). 1994 Figure 1 Map of the survey area near Lee Stocking Island, Exuma Cays, Bahamas. Flood tidal cur- rents (arrows) and the locations of queen conch, Strombus gigas, nursery habitats (cross- hatched) are shown. Dashed lines separate the inner and outer bank regions and delineate the study site. Areas south and west of the Brigantine Cays were not surveyed. bank was divided into an inner section from the Brigantine Cays to a line mid-way between the Brig- antines and Lee Stocking Island, and an outer sec- tion from the mid-line to the cays at the eastern side of the bank (Fig. 1). Each section is approximately 5 km wide. The rationale for this division was that the outer section of the bank is flushed with oceanic water on every tide, while the inner bank is flushed only on extreme tides. Virtually all queen conch nurs- eries in the Exuma Cays are found within the outer 5.0 km of the bank (Stoner et al., in press.) (Fig. 1). The eastern shores of the Exuma Cays are char- acterized primarily by steep aeolianite cliffs and beach rock interspersed with a few high-energy sandy beaches in coves, particularly on Lee Stock- ing Island and Children's Bay Cay. The seagrass Thalassia testudinum is found on shallow, soft-sedi- ment platforms extending a short distance off the sandy beaches. Most of the shallow nearshore, how- ever, is hard-bottom covered with a short turf of the green alga Cladophoropsis sp. The hard-bottom habitat, interspersed with small patches of sand and hard corals, is characteristic to 10-m depth. From 10 to 20 m the bottom comprises mixed hard-bottom and bare sand. Off Lee Stocking Island, corals form a 2-km long steep ledge from 10 to about 18 m, but a gradual slope to 25 m is typical of most of the study area. Patchy sand, coral, and hard-bottom are found between 20 and 30 m. Detailed hydrographic charts are not available for the Exuma Cays; therefore, shelf bathymetry was mapped with 540 electronic depth-sounder points, corrected for tidal state, and positions acquired with Global Positioning System (GPS) from the RV Chal- lenger during summer 1991. GPS positions taken at close intervals along the eastern shores of the is- lands were used as zero-depth data points. Three- dimensional plotting features of Systat 5.0 software were used to provide a bathymetric chart for the shelf region with 0, 2.5, 5, 10, 15, 20, 25 and 30 m contours for depth at mean low tide. Total surface area for each of the seven depth intervals was cal- culated with a digitizing board and SigmaScan 3.9 software. The surface area of the inner and outer bank regions was determined in a similar way with the aid of topographic maps. Stoner and Schwarte: Distribution of Strombus gigas 173 Survey methods The shelf region was surveyed in each of seven depth zones between 2.5 and 30 m (described above) along nine transects (perpendicular from the Cays into Exuma Sound) placed at approximately 1.0-km intervals. At each of the 63 shelf stations, divers swam parallel to the isobaths for a distance mea- sured with a calibrated General Oceanics flow meter equipped with a large propeller for low velocity flows. Calibration was performed by towing the meter repeatedly (n>6) through calm water at the side of a small boat over a pre-measured distance of 100 m. Precision was ±2%. Current velocity on the shelf adjacent to Lee Stocking Island is generally low (<3 cm/sec) and to the northwest, parallel to the isobaths (Smith, 19921). Recognizing the potential effect of current on the calculated distance, each dive included two legs, one up-current and one down- current in parallel lines of equal length separated by approximately 20 m. Two dives were made at most stations for density determinations and shell measurements (described below). For density, all queen conch were counted in an 8-m wide path defined by a line held between two divers. The average swim distance was 380 m, resulting in coverage of just over 3000 m2. Conch density was calculated by using only those conch in the 8-m band. Shell measurements were made for animals outside the 8-m band in areas with low conch densities. Underwater visibility was usually high and the area of bottom searched was actually much larger than the swim path alone. Conse- quently, all conch within approximately 30 m could be collected for measurement. In areas where conch densities were high, one dive was made to collect density data and another to collect only measure- ment data. An attempt was made to measure at least 100 adults from each depth zone, but this was not possible in the 0-5, 5-10, and 25-30 m zones because of low densities in these zones. Statistical differences in density among the survey zones were evaluated with the non-parametric Kruskal-Wallis test (Sokal and Rohlf, 1969) with stations used as replicates (n-9). The shallowest depth zone (0-2.5 m) was limited primarily to sandy coves on the major islands of the survey area. Adult queen conch were few in these areas, and juveniles were distributed unevenly; therefore, the important seagrass areas of the shal- low coves were thoroughly searched. Density mea- sures were not made but all conch encountered were measured (as described below). 1 N. P. Smith, Harbor Branch Oceanography Inst., Fort Pierce. FL, pers. commun. 1992. Sparse distribution of adult conch and the large surface area of the Great Bahama Bank required the use of different survey methods from those applied on the shelf. Because the bank waters are shallow and conch were easily seen, large areas were sur- veyed by towing a diver at the surface in continu- ous lines. The bank region was divided into 95 — 1 x 1 km squares oriented along lines of latitude. Then, in a systematic grid of lines running diago- nally through the squares, every square was crossed at least once during the survey. Additional tows were made in areas already known to have concentrations of adults, i.e., near nurseries previously mapped (Fig. 1; Stoner et al., in press.). Divers were towed a total distance of 126 km. Although water clarity on the bank was not as high as that on the island shelf, the towed diver could usually see at least 2.5 m on either side of the transect line. Surveys were not conducted on a few days when visibility was restricted. While being towed at approximately 50 cm/sec, the diver signaled numbers of adult queen conch to the boat operator, who recorded position. Positions for the ends of all straight line transects were determined with GPS, tow distance was estimated by chart, and conch density was calculated on the basis of the 5-m wide path examined. During the bank survey, 472 adults were gathered and measured. Presence of juveniles on the bank was noted but not quantified in this study. For comparison with shelf sites, a random collection of 322 juvenile conch was made from a nursery west of Lee Stocking Island during August 1991. These conch were measured for shell length. The total number of adult queen conch was esti- mated crudely for each bank and shelf area by ex- trapolating the average density of conch for an in- dividual zone over the total surface area for the same zone. Because variances in the density data were large, confidence intervals for the extrapolated numbers of conch were not calculated. Shell measurements Queen conch reach sexual maturity between 3.5 and 4 years of age, a few months after the shell edge has formed a broadly flared lip (Appeldoorn, 1988). Af- ter the lip flares, queen conch stop growing in length but continue to deposit shell material on the inside of the lip (Egan, 1985; Appeldoorn, 1988). Therefore, with certain limitations, thickness of the shell lip is an indicator of approximate age (Stoner and Sandt, 1992). In this study, shell-lip thickness was mea- sured with calipers in the area of greatest thickness, about two-thirds of the distance posterior from the 174 Fishery Bulletin 92(1), 1994 siphonal groove and 35 mm in from the edge of the shell, according to the methods of Appeldoorn (1988) and Stoner and Sandt (1992). Shell length was measured from the tip of the spire to the end of the siphonal canal in both adults and juveniles. Re- peated measures made by different persons showed that both length and lip thickness measurements were made to ±1 mm. Differences in length-fre- quency and thickness-frequency distributions were tested with the non-parametric Kolmogorov- Smirnov test. Morphological differences between bank and shelf populations were tested with canonical discriminant function analysis from shell length and lip thickness data. This multivariate technique is well suited for differentiating two types where individual charac- teristics do not separate the types. The analysis computes a third variable Z, which is a linear func- tion of both variables (length and thickness, in this case) such that the equation for the new line maxi- mizes the distance between the two types ( Sokal and Rohlf, 1969). The significance of the discriminant function Z was determined with the Hotelling- Lawley trace test statistic (Morrison, 1976). Results of the canonical analysis were then examined to determine what percentage of the individuals were correctly classified according to collection site. Observations were also made on general shell thickness (particularly in juveniles), length of api- cal spines and resultant shell diameter, and num- ber of spines per whorl. None of these characteris- tics were quantified systematically. Shell growth experiment Early observations suggested that shell phenotypes were different between shelf and bank conch. Adults from the shelf appeared to be longer and to have thicker shell lips than those from the bank. Juve- niles from the shelf were more narrow, thin-lipped, and had shorter apical spines than those on the bank (Martin-Mora, 1992). To examine the potential relation between shell morphology and growth rates, juveniles were tagged in two different nursery sites: in the well-studied nursery west of Children's Bay Cay and in seagrass areas off Charlie's Beach in the northeast cove of Lee Stocking Island (Fig. 1). Ju- veniles were individually marked with spaghetti tags (Floy Tag & Manufacturing Co.) tied around the spire and measured to the nearest millimeter with calipers. Charlie's Beach conch between 108 and 150 mm (mean=137 mm, n=281) were measured and released in the last week of August 1990. Children's Bay Cay conch, somewhat smaller than the Charlie's Beach conch (106 to 133 mm, mean=118 mm. n=292), were tagged and released in early Septem- ber 1990. Conch from both populations were remeasured for shell length five months later, at the end of February 1991. Forty-eight conch were recov- ered at Charlie's Beach and 135 were recovered at the Children's Bay Cay site. Daily growth rate was calculated for individuals by dividing increase in length by the number of days between measure- ments. Differences in growth rate between the two sites were evaluated by using the Mann-Whitney U- test. Results Conch densities and abundance Densities of adult queen conch in the survey area were highest between 15 and 20 m depth on the island shelf (Table 1) with nearly 88 conch/ha (Fig. 2). Density was also high between the 10- and 15-m isobaths. In both of these depth zones densi- ties of adults were highly variable, but there was no apparent pattern across transect lines. There was a highly significant difference in the density of adult conch in the survey zones (Kruskal-Wallis test, Hadj =36.195, P<0.001KFig. 2). No conch were found deeper than 25 m, despite an abundance of appar- ently suitable habitat of sand and algae-covered hard-bottom. Adults were most sparsely distributed 50 20 • 10 ■ o c If) c; v Q U C o O 120 ■ 100 ■ 80 60 40 20 0 r+-. Eh Inner Outer Bank •- «- CN ^ Depth in (Meters) Shelf Figure 2 Density of queen conch, Strombus gigas, on the Great Bahama Bank and in six different depth zones of the island shelf hear Lee Stocking Island, Bahamas. Values are ± mean standard error of the Stoner and Schwarte: Distribution of Strombus gigas 175 Table 1 Estimated total number o f adu It queen conch, Strombus gigas, in a 12- km section o fth e Exuma Cays , Bahamas, between Adderly Rocks and Rat Cay. Density (no. /ha) Region Total area (h a) (mean ± SE of mean) Total no. of conch Bank Inner 4,979 0.19 ± 0.14 946 Outer 3,997 3.16 ± 1.69 12,631 Bank total 8,976 13,577 Shelf 0-2.5 m 161 Low — not qualified Negligible 2.5-5 m 198 2.24 ± 1.70 444 5-10 m 465 7.21 ± 4.11 3,353 10-15 m 429 60.1 ± 46.8 25,800 15-20 m 454 87.9 ± 31.5 39,902 20-25 m 320 18.3 ± 9.1 5,843 25-30 m 151 0 ± 0 0 Shelf total 3,687 75,342 Grand Total 12,663 88,919 in the inner section of the Great Bahama Bank with only 0.19 conch/ha (SD=0.15, n=28). Density of adult conch in the outer (seaward) section of the bank was close to the value for the 2.5-5 m depth zone on the shelf. Although not quantified, numbers of adults in the nearshore (0-2.5 m) zone of the shelf were negligible. Few juvenile conch were observed on the shelf between 2.5-and 15-m depth (Fig. 2). A total of 372 juveniles were found in densely aggregated patches on seagrass beds off the eastern beaches of Lee Stocking Island. On the Great Bahama Bank, most juveniles were aggregated in specific nursery loca- tions documented previously (Stoner et al., in press.). None was found deeper than 15 m. The area between 10 and 20 m depth on the is- land shelf was a particularly important habitat for adult queen conch (Table 1). Approximately 74% of all conch in the 12-km long survey area reside in this narrow depth zone. It is also clear that large expanses of shallow bank habitat support a rela- tively small proportion (15.2%) of the adult popula- tion. Mating conch and demersal egg masses were very abundant during summer months at shelf sites deeper than 10 m, but none were observed on the bank. Shell morphology The shelf sites were characterized by large adult queen conch, primarily between 200 and 260 mm (mean=227, SD=23, n=572), whereas most adult conch on Great Bahama Bank were between 170 and 210 mm shell length (mean=187, SD=16, n=472). Pooling all adults measured, there was a highly significant difference in the length-frequency distribu- tion of conch on the shelf and on the bank (Kolmogorov- Smirnov test, P<0.001). The distributions (Fig. 3) show clearly the separation in size of adults between bank and shelf sites, particularly when com- paring nearshore (0-5 m) shelf zones with those from the bank. The distributions show a decrease in shell length be- tween the nearshore shelf and deeper zones, while those be- tween 5 and 25 m are obviously similar. Bank conch had thin shell lips (mean=10, SD=6); conch from nearshore (2.5-5 m) regions of the shelf were intermediate in lip thickness (mean=18, SD=5); and deep-shelf (5-25 m) conch had the thickest shell lips (mean=30, SD=7)(Fig. 4). All three of these groups were significantly different from one another in terms of lip thickness distribution (Kolmogorov- Smirnov tests, P<0.01). There was obvious similar- ity in the thickness distributions of shells in depth zones between 5 and 25 m; therefore, these four depth categories were pooled. Distinctness of the morphs collected on the bank and shelf is further suggested by a plot of shell length and lip thickness for 250 randomly chosen individuals from each of the two regions (Fig. 5). Also, when length and lip thickness data for all 1,029 conch measured in the survey were used in canonical discriminant function analysis, a highly significant separation was found between conch col- lected in the two different regions (Hotelling-Lawley Trace, F= 1,854, P<0.001). Less than 5% of the conch in the survey were not collected in the region pre- dicted by the multivariate equation. Bank conch were small and had thin shell lips, whereas conch from the island shelf were large and had thick shell lips. Results of the analysis, however, do not rule out the possibility that the smallest adult conch from the shelf region, particularly apparent in the 5-10 m depth zone, could be older animals from the bank. Length-frequency distributions of juvenile queen conch were different on the Great Bahama Bank and island shelf (Fig. 6). Both the bank and nearshore 176 Fishery Bulletin 92(1). 1994 40 N - 472 Boik 30 20 10 40 n . m Shelf 0-5 m 30 ■ C 20 j_ O 10 Jl _l 1 - -■■ o « N.57 Shelf 5-10 m D x Q_ 20 o l0 CL v, «o N . 22» Shelf 1 0- 1 5 m O 30 -1 — ' 20 J rcen o o o --^. :, , ,50 Shelf 15-20 m CL) 0- 20 I 10 ^m 40 . N_ loo Shelf 20-25 m JO 20 10 100 120 140 160 180 200 220 240 260 280 300 She!! Length (mm) Figure 3 Length-frequency distributions for adult queen conch, Strombus gigas, on the Great Bahamas Bank and in five different depth zones of the island shelf near Lee Stocking Island, Bahamas. 40 N - 472 Bonk 30 20 A 10 JL to N. j6 Shelf 0-5 m 30 ■ c 10 J_ o 10 Jl 1 , 0 40 . JHH o K,i, Shelf 5-10 m 13 so CL 20 O CL 10 .^4. ^_ 40 N . 229 Shelf 10-15 m o 30 20 c CD (J 10 0 4 0 _jL n. iso Shelf 15-20 m CD 30 CL 20 1 10 ^jjL 40 n . ioo Shelf 20-25 m 30 20 J 10 _jB_ 100 120 140 160 180 200 220 240 260 280 300 Shell Length (mm) Figure 4 Distribution of shell lip-thickness for adult queen conch, Strombus gigas, on the Great Bahama Bank and in five different depth zones of the island shelf near Lee Stocking Island Bahamas. (0-2.5 m) shelf had juveniles less than 100 mm in shell length; however, these were rare in the shelf environment, and few juveniles less than 160 mm were found on the shelf between 2.5- and 15-m depth. None of the juveniles on the bank were near the 227-mm average length of adults on the shelf, but many juveniles collected in deeper water were close to adult size. Other differences were observed in the shells of queen conch from bank and shelf regions. Juvenile conch from the bank differed from shelf juveniles because of thicker shells and longer lateral spines Stoner and Schwarte: Distribution of Strombus gigas 177 w c -XL Q. 50 40 30 20 10 0 ° Bonk • Shelf - ^i t • • v Jn ♦. 400. 10 15 20 25 Shell Length (cm) 30 Figure 5 Scatterplot of shell lip-thickness vs. shell length for adult queen conch, Strombus gigas, collected from the Great Bahama Bank and from the Lee Stocking Island shelf. Two-hundred and fifty randomly chosen points were plotted for each site. (5—6 spines/whorl vs. 7-9 spines/whorl in shelf ju- veniles). Bank conch had a maximum shell diameter between 80 and 90% of shell length at 100 mm length, whereas juveniles from the shelf had diam- eters between 50 and 60% of shell length. These characteristics persisted to adult stages with bank conch having longer spines. The outer whorls of shelf adults, even young individuals, were often nearly smooth. Growth rates Juvenile queen conch on the island shelf at Charlie's Beach grew in length at a rate approximately 2.4 times the rate observed at the Children's Bay Cay site. Conch recovered at Charlie's Beach grew 0.139 mm/day (SD=0.025, n=135). At Children's Bay Cay, mean growth rate was 0.058 mm/day (SD=0.021, n=48). The differences in growth rate between bank and shelf juveniles were highly significant (Mann- Whitney [/-test, P<0.001). Discussion The rapid increase in adult queen conch density at depths greater than 10 m is probably a direct func- tion of fishing, which is limited to free-diving on the bank and shallow nearshore shelf areas around Lee Stocking Island. This conclusion is substantiated by observations of conch depth distribution in other localities. In unfished areas of Islas Los Roques, Venezuela, Weil and Laughlin (1984) found that 60 B0 100 120 140 160 180 200 220 240 260 280 Shell Length (mm) Figure 6 Length-frequency distribution for juve- nile queen conch, Strombus gigas, from the Great Bahama Bank and from two depth zones on the island shelf near Lee Stocking Island, Bahamas. density of queen conch was highest in 4.0 m of wa- ter and density decreased with depth to 18 m. This may represent the natural distribution of queen conch. In comparable 4-m deep habitats not pro- tected from fishing, densities were 5 times less than those in the protected area. Similarly, in the Exuma Land and Sea Park, a 500-km2 fishery reserve 90 km north of Lee Stocking Island, there are large numbers (unquantified) of adult conch at 2-4 m depth, and many of these shallow-water conch have been observed laying eggs (Stoner, pers. observ.); whereas adults are uncommon in shallow water near Lee Stocking Island and spawning has never been observed at less than 5 m depth. Similar to the pattern reported in this study for Lee Stocking Is- land, Torres-Rosado ( 1987) found maximum density of adult queen conch between 10 and 20 m in Puerto Rico, where fishing is heavy in shallower waters. It is recognized that queen conch move to greater depths with age and size (Randall, 1964; Weil and Laughlin, 1984); this has been confirmed in the Lee Stocking Island area by the recovery of individuals that were tagged as juveniles at Charlie's Beach and subsequently found in deeper offshore waters Fishery Bulletin 92(1). 1994 (Stoner, unpubl. data). However, our morphological analyses of conch suggest that very few conch us- ing the bank for a nursery actually reach the off- shore spawning sites. Furthermore, similarities in length frequency and shell morphology between ju- veniles found immediately off the east (windward) side of the Cays on isolated seagrass beds and adults in deep water suggest that the small aggregations of juveniles found on the shelf serve as the primary source for the offshore reproductive stocks. Given that mating and egg-laying are rare on the Great Bahama Bank, it is likely that recruitment to bank nurseries is sustained by deep-water reproductive populations (Wicklund et al., 1991; Stoner et al., 1992; Stoner and Sandt, 1992). Differences in shell morphology between bank and shelf conch are not well understood but appear to be related to growth rate. Alcolado (1976) reported that large, thin shells and short spines in queen conch in Cuba were associated with rapid growth. A similar phenomenon may explain the shell differ- ences observed in this study. Juveniles in the nearshore shelf environment of Charlies' Beach grew rapidly and had the large, thin-shelled, short- spined morphotype typical of the shelf adults. The small, thick-shelled, long-spined conch on the bank had growth rates less than half of those on the shelf. A recent transplant experiment at Lee Stocking Is- land demonstrated that shell form and spination in juvenile conch is an environmentally mediated char- acteristic associated with habitat type and indi- vidual growth rate (Martin-Mora, 1992). The large size of the deep-water reproductive stock may explain high productivity of queen conch in the Exuma Cays. It is likely, however, that abun- dance of conch in the region is now dependent upon the small, isolated pockets of fast-growing juveniles that inhabit the nearshore shelf habitat during the first two or more years of life then recruit to deep- water reproductive populations. Stoner and Sandt (1992) found that the adult population at an 18-m deep site off Lee Stocking Island was relatively stable between 1988 and 1991, but most individu- als were old and thick-lipped. The predominance of old conch in deep water may or may not be a func- tion of low recruitment rates from shallow water in recent years, and the significance of shallow-water spawning to conch abundance is unknown. In an comparison of data from Glazer and Berg (in press.), densities of queen conch in the Exuma Cays are 10 to 100 times higher than those reported for many other localities in the Caribbean region. This may be related to geographic differences in habitat quality, recruitment processes, and fishing methods. The Exuma Cays probably represent a particularly efficient system for retaining conch lar- vae because of unique geographic and oceanographic conditions such as an alongshore current and nu- merous tidal inlets leading to nursery grounds (Stoner et al., in press), but fishing methods can play a large role in the population structure of queen conch. Fishing in the Bahamas is restricted to free- diving and limited diving with surface-supply air for adults with flared shell lips; therefore, conch deeper than 10 m are rarely exploited. Depth distribution of queen conch near Lee Stocking Island suggests that virtually every conch in the Exuma Cays is within the range of SCUBA divers and that popu- lations of S. gigas could be decimated quickly if the fishery were opened to this latter gear. On the other hand, if the source of deep-water conch is shallow- water nurseries, protection of deep-water reproduc- tive stocks only delays the effects of overfishing, and certain nurseries should be protected as well. Analy- sis of larval transport and recruitment processes will be crucial to the sound management of this already threatened commercial species. Acknowledgments This research was supported by a grant from the Undersea Research Program of NOAA, U.S. De- partment of Commerce. We thank P. Bergman, G. Donnly, R. Gomez, J. Lally, D. Mansfield, M. Ray, V. Sandt, D. Wicklund, and E. Wishinski for assis- tance in the field work. R. I. Wicklund prompted us to examine the important juvenile stocks off the windward beaches. The manuscript was improved with helpful criticism by R. Appeldoorn, R. Hardy, E. Martin, M. Ray, and an anonymous reviewer. Literature cited Alcolado, P. M. 1976. Crecimiento, variaciones morfologicas de la concha y algunos datos biologicos del cobo Strombus gigas L. (Mollusca, Mesogas- tropoda). Acad. Ciencias de Cuba, Inst, de Oceanol. No. 34, 36 p. Appeldoorn, R. S. 1988. Age determination, growth, mortality, and age of first reproduction in adult queen conch, Strombus gigas L., off Puerto Rico. Fish. Res. 6:363-378. Appeldoorn, R. S., G. D. Dennis, and O. Monterrosa-Lopez. 1987. Review of shared demersal resources of Puerto Rico and the lesser Antilles region. In R. Mahon (ed.), Report and proceedings of the expert consul- Stoner and Schwarte: Distribution of Strombus gigas 179 tation on shared fishery resources of the lesser Antilles region. FAO Fish. Rep. 383:36-106. Berg Jr., C. J., and D. A. Olsen. 1989. Conservation and management of queen conch (Strombus gigas) fisheries in the Caribbean. In J. F. Caddy (ed.), Marine inverte- brate fisheries: their assessment and management. Wiley & Sons, NY, p. 421-442. Egan, B. D. 1985. Aspects of the reproductive biology of Strombus gigas. M.S. thesis, Univ. British Co- lumbia, Vancouver, Canada, 147 p. Glazer, R. A., and C. J. Berg Jr. In press. Current and future queen conch, Strombus gigas, research in Florida. In R. S. Appeldoorn and B. Rodriguez (eds.), The biology, fisheries, mariculture, and management of the queen conch. Fundacion Cientifica Los Roques, Caracas, Venezuela. Martin-Mora, E. 1992. Developmental plasticity in the shell of the queen conch, Strombus gigas. M.S. thesis, Florida State Univ., Tallahassee, 52 p. Morrison, D. F. 1976. Multivariate statistical methods. McGraw- Hill, New York. Randall, J. E. 1964. Contributions to the biology of the queen conch, Strombus gigas. Bull. Mar. Sci. 14:246-295. Sokal, R. R., and F. J. Rohlf. 1969. Biometry. W. H. Freeman, San Francisco, 776 p. Stoner, A. W., and V. J. Sandt. 1992. Population structure, seasonal movements. and feeding of queen conch, Strombus gigas, in deep-water habitats of the Bahamas. Bull. Mar. Sci. 51:287-300. Stoner, A. W., V. J. Sandt, and I. F. Boidron- Metairon. 1992. Seasonality in reproductive activity and lar- val abundance of the queen conch, Strombus gigas. Fish. Bull. 90:161-170. Stoner, A. W., M. D. Hanisak, N. P. Smith, and R. A. Armstrong. In press. Large-scale distribution of queen conch biology, fisheries, and mariculture: implications for stock enhancement. In R. S. Appeldoorn and B. Rodriguez (eds.), The biology, fisheries, maricul- ture, and management of the queen conch. Fundacion Cientifica Los Roques, Cara- cas, Venezuela. Torres-Rosado, Z. A. 1987. Distribution of two mesogastropods, the queen conch, Strombus gigas Linnaeus, and the milk conch, Strombus costatus Gmelin, in La Parguera, Lajas, Puerto Rico. M.S. thesis, Univ. Puerto Rico, Mayaguez, 37 p. Weil, E., and R. Laughlin. 1984. Biology, population dynamics, and reproduc- tion of the queen conch, Strombus gigas Linne, in the Archipielago de Los Roques National Park. J. Shellfish Res. 4:45-62. Wicklund, R. I., L. J. Hepp, and G. A. Wenz. 1991. Preliminary studies on the early life history of the queen conch, Strombus gigas, in the Exuma Cays, Bahamas. Proc. Gulf Caribb. Fish. Inst. 40:283-298. Abstract. — Regression and time series analyses were used to investigate the relation between Apalachicola River flows and blue crab, Callinectes sapidus, harvests in and around Apalachicola Bay, Florida. Apalachicola River flows in one year were positively corre- lated with Franklin County blue crab landings during the next year (r2=0.32, P<0.001, 1952-90), and the strength of the correlation in- creased when only more recent years were examined (r2=0.49, P=0.001, 1973-90). In this area, blue crabs mature to a harvestable size by one year of age. Apala- chicola River flows were also cor- related with neighboring Wakulla County blue crab landings with a one-year time lag (r2=0.52, P=0.001, n=l7), but were not asso- ciated with blue crab landings for the remaining west coast of Florida. The mean monthly flow from September to May, termed the growout period, was the pa- rameter most highly correlated with the following year's blue crab landings. Of five north Florida riv- ers examined, the Apalachicola River was most highly correlated with Franklin and Wakulla County blue crab landings. Results of this study further document the influence of Apal- achicola River flows on estuarine productivity. The positive relation between flows and blue crab har- vests a year later suggests that low flow conditions in the estuary during the growout period nega- tively affect juveniles. Although the underlying causes of the corre- lations are not known, the effect of inflows on estuarine salinity is one of several possible mechanisms that warrants further investigation. The influence of Apalachicola River flows on blue crab, Callinectes sapidus, in north Florida Dara H. Wilber 1 640 Oak Ridge Road. Vicksburg. MS 39 1 80 Manuscript accepted 20 July 1993 Fishery Bulletin 92:180-188 1 1994) River flow affects many character- istics of estuaries, including salin- ity, turbidity, and nutrient and de- trital concentrations. Changes in flow, therefore, may significantly affect estuarine biota, the extent to which may be inferred by examin- ing historical relations between flow and productivity. Apalachicola Bay, Florida, like many estuaries, is subject to changes in freshwater inflow related to factors such as rainfall and upstream demands for agricultural, municipal, and indus- trial uses. Plans to reallocate fresh- water resources (U.S. Army Corps of Engineers, 19891) have renewed interest in the question of how freshwater inflows are related to productivity in the Apalachicola River and Bay system. This study examined the historical relation- ship between Apalachicola River flows and estuarine productivity. One method of characterizing the importance of freshwater inflow to estuarine productivity is to corre- late historical flow data with the commercial catch (landings) of es- tuarine-dependent species (Fun- icelli, 1984). Commercial landings are used to estimate estuarine pro- ductivity because they are often the only available long-term records from which species abun- dance can be inferred. Long-term records are available for several commercially important species in Apalachicola Bay, including oysters and blue crabs, which have differ- ent trophic requirements and es- tuarine residency patterns. By ex- amining associations between these species and Apalachicola River flows, effects of freshwater delivery upon estuarine productiv- ity can be evaluated. Associations between freshwater inflows and Apalachicola oyster harvests have been previously addressed (Wilber, 1992). The present study examines the influence of Apalachicola River flows on local and regional commer- cial blue crab, Callinectes sapidus, landings. Other north Florida riv- ers were also examined to estimate the relative importance of the Apalachicola River to blue crab landings with respect to these drainages. Blue crabs in the Gulf of Mexico reach a harvestable size within a year of age (Perry, 1984) and com- prise a significant portion of the commercial landings by 18-months of age (Steele, 19922). Blue crabs enter the Apalachicola estuary as megalopae and young juveniles, reaching peak juvenile abundances in the winter (Livingston, 1983). Young crabs concentrate in the less saline portions of the bay, whereas egg-bearing females remain in the higher-salinity gulf waters where they spawn. It has been proposed that adult female blue crabs along the Florida gulf coast migrate to 1 U.S. Army Corps of Engineers, Mobile District. 1989. Draft Post Authorization Change Notification Report for the Real- location of Storage from Hydropower to Water Supply at Lake Lanier, Georgia, 320 p. 2 P. Steele, Florida Marine Research Inst., 108th Ave. SE, St. Petersburg, FL 33701, pers. commun. 1992. 180 Wilber: Influence of Apalachicola River flows on Callinectes sapidus 181 gulf waters near Apalachicola Bay to spawn and that the larvae are distributed to the south by loop currents (Oesterling and Evink, 1977). Evidence supporting this hypothesis was examined in this study. Methods Fisheries data Several aspects of the blue crab fishery may lead to inaccurate fishery representation of adult stock abundance. For example, unreported landings from the recreational fishery and crab bycatch from the shrimp fishery are potential sources of bias in blue crab landings statistics (Perry, 1984). Although these sources of error cannot be controlled, if they are independent of river flow and account for a rela- tively constant proportion of the total landings over time, a valid, although perhaps conservative, rep- resentation of environmental effects on the species can be obtained. The Florida Department of Natural Resources (FDNR) provided monthly landing data for blue crabs from Franklin and Wakulla Counties for 1979- 90, monthly effort data (number of trips) for 1987- 90, annual landing data from Wakulla County from 1973 to 1990 (excluding 1977), and annual landing data from the Florida west coast from 1960-1990. Franklin County annual landing data from 1952 to 1979 were also obtained (Herbert et al., 19883). Sta- tistical analyses (Wilkinson, 1990) were conducted by using the full 39-year Franklin County dataset, as well as a shorter (1973-90) dataset, which al- lowed comparisons between Franklin and Wakulla Counties that were not confounded by differences in time periods. The limited amount of effort data pre- cluded analyses of catch per unit of effort. Flow and rainfall data The Apalachicola River begins at the Florida state line by the confluence of the Chattahoochee and Flint Rivers. Apalachicola flow data were collected at the United States Geological Service gauge at Blountstown, Florida, which is the closest station to the estuary (105 km upstream) with an adequate period of record. This station is not immediately adjacent to the estuary, therefore fresh water from local inputs and storm events are not included. The drainage area downstream from the Blountstown gauge is less than 9% of the total area drained by the Apalachicola-Chattahoochee-Flint River system (Leitman et al., 19834). Parameters examined included the highest and lowest average flows for 7 and 120-consecutive days each year (referred to as the 7- and 120-day maxi- mum and minimum flows). Monthly minimum, mean, and maximum values, and the mean monthly flow during the growout period (September-May) were also examined. By using these flow durations, associations between landings and seasonal high and low flows could be examined, which was not possible when analyses included only mean annual flow. The growout-flow time period was adapted from a similar study correlating blue crab landings in Georgia with river discharges (Rogers et al., 19905). Sufficient historical flow data were also available for the Suwannee, Econfina, St. Marks, and Ochlockonee Rivers (Fig. 1), thus permitting a re- gional analysis of associations between flows and blue crab landings. For each river, the annual one- day minimum, one-day maximum, and annual mean flows were used. One-day high and low flow magni- tudes were used because of their availability and because preliminary analyses which substituted other flow durations (annual minimums and maxi- mums) on the Apalachicola River did not change results considerably. Statistical analyses Blue crab landings and flow data were tested for monthly, seasonal, and inter-annual dependencies through autocorrelations. Data were adjusted to remove dependencies when autocorrelations were significant. If autocorrelations between successive months were present, data were replaced by the difference between each month and the preceding month. If seasonal autocorrelations were present, the effects were removed by dividing each value by a seasonal factor. For instance, if landings exhibited a significant autocorrelation with a 12-month time lag, which reflected a similarity in catches for the same month among years, each monthly value was divided by the month's mean value and replaced by the quotient. Similar analyses were conducted with seasonal (three-month averages) landings and flow data following adjustments to remove significant autocorrelations. Flow data were log10 transformed. 3 Herbert, T. A., and Associates. 1988. The Franklin County Fisheries Options Report, 164 p. 4 Leitman, H. M., J. E. Sohm, and M. A. Franklin. 1983. Wet- land hydrology and tree distribution of the Apalachicola River flood plain, Florida. U.S. Geological Survey Water-Supply Pa- per 2196, 52 p. 5 Rogers, S. G., J. D. Arrendondo, and S. N. Latham. 1990. As- sessment of the effects of the environment on the Georgia blue crab stock. Final Rep. Georgia Dep. Natl. Resources, 69 p. 182 Fishery Bulletin 92(1), 1994 . .A.JlA.£A_M_A_ A- Apalachicola R. O Ocklockonee R. SM- St. Marks R. E- Econfina R. S- Suwannee R. "EST ^..-•'^•° °° Figure 1 Percentage of total Florida west coast blue crab (Callinectes sapidus) landings caught by area (Steele, 1982). The five rivers used in the multivariate regres- sion analyses (Apalachicola, Ochlockonee, St. Marks, Econfina, and Suwannee) are depicted. Autoregressive order 1 (ARIMA) models were con- ducted on the Franklin and Wakulla County blue crab annual data and the residuals from these analyses were correlated with flow. This approach provided statistically rigorous estimates of P-values for the flow/landings relationships that were inde- pendent of any effects resulting from the one-year autocorrelations in landings. Analyses that used the ARIMA residuals and those that used unadjusted blue crab landings data were reported because both methods impart useful information. Correlations that used unadjusted annual blue crab data, i.e., significant autocorrelations were not removed, were biologically relevant because feedback mechanisms inherent to these autocorrelations (such as reproduc- tion and recruitment) may also be associated with flow. Results of analyses that used unadjusted data are also more readily compared to results of other studies. Use of ARIMA models statistically validated the significant relations between blue crab landings and flow data, but may have removed some biologi- cally relevant information. This paper primarily refers to unadjusted regression results. Regression analyses incorporating a one-year time lag between flows and landings were conducted to examine the effects of flow on early blue crab life history stages. Contemporaneous analyses were con- ducted to assess the effect of flow on adults. Univariate and stepwise multivariate regression analyses were conducted to estimate the amount of variability in blue crab landings accounted for by five major rivers on Florida's northern gulf coast. The criterion for admitting a flow variable into the stepwise regression models was an F-statistic greater than 4.0 for its partial correlation with land- ings. Data on blue crab landings for the west coast of Florida were used as a dependent variable in some analyses. To more specifically examine Wilber: Influence of Apalachicola River flows on Callinectes sapidus 183 whether there was evidence that the Apalachicola River affects blue crab landings on a regional basis, Franklin and Wakulla landings were removed from the west coast dataset. Regression analyses were conducted to test whether Apalachicola flows and the remaining west coast landings were significantly related. Results Annual landings Blue crab landings varied nearly 10-fold over the period of record examined in each county (Fig. 2). Significant autocorrelations between consecutive years were present in both Franklin (r2=0.19, P=0.006) and Wakulla (r2=0.37, P=0.016) County landings. Annual flow parameters did not exhibit any significant autocorrelations. Annual Franklin County blue crab landings were most highly correlated with Apalachicola River flows of the previous year and these correlations were positive (Table I). The growout flow with a one-year time lag accounted for the greatest amount of varia- tion in blue crab landings (r2=0.32, P<0.001; Fig. 3A). The regression analysis of ARIMA residuals (autocorrelation in blue crab landings removed) and growout flows of the previous year was also signifi- cant (r2=0.21; P=0.004). Wakulla County landings ANNUAL BLUE CRAB LANDINGS 1.5 O.S 0.0 WAKULLA FRANKLIN —i — i — i — [ — i — i — i — i — | — i — i — i — i — [ — r 1950 1960 1970 1980 YEAR 1990 Figure 2 Annual blue crab (Callinectes sapidus) landings for Franklin (closed squares) and Wakulla (open squares) Counties in millions of kilograms. Table 1 R2 values from regression analyses for Franklin (n=39) and Wakulla (n = 17) Coun ty blue crab (Callinectes sap idus) landings and Apalachicola River flows. All correlations were positive. Flow parameter Franklin Wakulla no lag period 7-day low 0.16* 0.12 120-day low 0.14* 0.18 7-day high 0.04 0.07 120-day high <0.01 <0.01 growout 0.08 0.10 one-year lag 7— day low 0.25** 0.29* 120-day low 0.21** 0.21 7-day high 0.18* 0.17 120-day high 0.21** 0.31* growout 0.32*** 0.52*** * = P < 0.05. ** = P < 0.01. *** = P < 0.001. were significantly correlated only with Apalachicola flows of the previous year, with the growout flow also accounting for the greatest amount of variation in annual blue crab landings (r2=0.52, P=0.001; Fig. 3B). The regression analysis of ARIMA residuals and growout flows one year previous was significant (r2=0.35, P=0.02). The shorter (1973-90) data record for Franklin County landings was more strongly correlated with growout flows with a one-year time lag (r2=0.49, P=0.001; Fig 3C) than was the full 39- year dataset. Monthly and seasonal landings As expected, the monthly Franklin and Wakulla County blue crab landings (1979-90) exhibited sig- nificant autocorrelations for 1- and 12-month time lags. All monthly river flow parameters (minimum, mean, and maximum) also exhibited significant cor- relations between successive months and with 12- month lags. Correlations between monthly landings and flow parameters (without any adjustments for significant autocorrelations) were positive for time lags of 3, 4, and 5 months. Significant negative cor- relations were present for flows that lagged 2-4 months behind landings. Correlations that used landings and flow data with the 1- and 12-month autocorrelation effects removed were not significant for either county. Peak harvests generally occurred between May and September in both counties. There were also no significant correlations between the seasonal (three- 184 Fishery Bulletin 92(1), 1994 ONE-YEAR TIME LAG FRANKLIN COUNTY (1952-1990) ONE-YEAR TIME LAQ WAKULLA COUNTY (1873-1990) 2.0 -i O 1.6 1.0 2 ° 0.6 0.0 r - 0.62 B 300 500 700 800 1100 300 500 700 000 1100 QROWOUT FLOW (lnT/SEC) QROWOUT FLOW (M /SEC) ONE-YEAR TIME LAQ FRANKLIN COUNTY (1973-1990) 1.0 0.8 o j 0.6 I Q < cr o 0.4 - 0.2 - 0.0 T~ 300 500 700 900 QROWOUT FLOW (M^SEC) 1100 Figure 3 Apalachicola River growout (flows m3/sec, mean flow from September through May) plotted against the following year's (A) Franklin County blue crab landings (1952-90), (B) Wakulla County blue crab landings (1973-90 ex- cluding 1977), and (C) Franklin County blue crab landings (1973-90). Flow data were log transformed in the statistical analyses. Wilber: Influence of Apalachicola River flows on Callinectes sapidus 185 Apalachicola 1.00 Ochlockonee 0.59** St. Marks 0.32 Econfina 0.50* Suwannee 0.60** * = P < 0.01. ** = P < 0.001. month average) flow and land- ings data with autocorrelations removed. The timing of peak monthly harvests was not re- lated to the magnitude of the annual harvests. Regional analysis Given the close geographical proximity of the five rivers ( Fig. 1) used in the multiple regres- sion analyses, significant corre- lations between annual flow pa- rameters may be expected among the rivers. Apalachicola River an- nual mean flows, although significantly correlated with other river flows (except the St. Marks), had the lowest correla- tions with the other drainages (Table 2). Significant correlations with blue crab landings were more common for Apalachicola River flows than for any other north Florida river tested (Table 3). Franklin County landings were correlated only with Apalachicola flows, whereas Wakulla County and west coast landings were also correlated with Suwannee and Ochlockonee flows, respectively (Table 3). These significant univariate correlations incorporated a one-year time lag. The Franklin County multivariate re- gression model included Apalachicola and Ochlockonee minimum flows of the previ- ous year (r2=0.45, P<0.001; Table 4). The Wakulla multivariate model accounted for the most variation in blue crab landings (r2=0.64; Table 4) and included Apalachicola mean and Ochlockonee minimum flows of the previous year. The west coast multivariate model with a one-year time lag in- cluded Apalachicola maximum and Ochlockonee minimum and mean flows (r2=0.53; Table 4). The only significant multivariate model that in- cluded parameters with and without time lags was for west coast landings, which used both no-lag Suwannee minimum flows and Apalachicola maximum flows of the previous year (r2=0.49). Analyses that examined associations between Apalachicola River flow and west coast landings with Franklin and Wakulla County landings removed were not significant. Discussion Several consistent results appeared in the correla- tions of annual blue crab landings with Apalachicola Table 2 Pearson correlation matrix of annual mean river flows for all possible combinations of five north Florida rivers. Apalachicola Ochlockonee St. Marks Econfina Suwannee 1.00 0.77* 0.76* 0.93* 1.00 0.64** 0.77** 1.00 0.79* 1.00 Table 3 Univariate correlations between Wakulla, Franklin, an d west coast blue crab (Callinectes sapidus ) landings and th a river flows from five north Florida drainages (Suwannee, Econfina, St. Marks, Ochlockonee, and Apalachicola) wi th a one-year time lag. Signs of the correlations are given in parentheses. Region Correlation r2 P Franklin Apalachicola minimum ( + ) 0.31 0.001 Apalachicola mean ( + ) 0.25 0.004 Apalachicola maximum ( + ) 0.14 0.039 Wakulla Apalachicola minimum (+) 0.29 0.031 Apalachicola mean ( + ) 0.38 0.010 Suwannee minimum ( + ) 0.30 0.028 West Coast Apalachicola mean ( + ) 0.15 0.035 Apalachicola maximum ( + ) 0.26 0.004 Ochlockonee minimum (-) 0.22 0.009 Table 4 Multiple regression results for Franklin, Wakulla, and west coast landings of blue crabs (Callinectes sapidus) with a one-year time lag incorporated into the analyses. The independent variables are the five river drainages listed in Table 3. Listed below are the signs of the correlations in paren- theses, Student's ^-statistics, and associated P- values. Region Variable Franklin Wakulla West Coast Apal. min ( + ) Och. min. (-) Apal. mean ( + ) Och. min. (-) Apal. max. ( + ) Och. mean ( + ) Och. min. (-) 4.38 -2.68 4.57 -3.05 2.99 3.12 -2.61 <0.001 0.012 0.001 0.009 0.006 0.004 0.015 0.45 (I 64 0.53 186 Fishery Bulletin 92(1), 1994 River flows. Statistically significant correlations were positive and primarily restricted to a time lag of one year, indicating higher flows were associated with higher blue crab landings the following year and lower flows with poorer landings the next year. The mean flow during the growout period (Septem- ber through May) of the previous year was the most highly correlated flow parameter with blue crab landings in both counties. A number of explanations are consistent with the observation that more fresh water (within a certain range) was associated with higher blue crab land- ings the following year. Greater freshwater inflows reduce estuarine salinities, thereby increasing the area of suitable habitat in the middle, and perhaps lower, estuary where juvenile blue crabs can forage and develop (Livingston et al., 1976; Perry, 1984). Increases in low salinity habitat may reduce preda- tion by marine species on juvenile blue crabs. Greater freshwater flows may also broaden an estuary's signal to offshore female migrants and/or megalopae, thus increasing the potential recruit- ment population base (Perry and Stuck, 1982; Mense and Wenner, 1989). In addition, higher in- flows carry more detrital and nutrient matter (Mattraw and Elder, 1982), which may either di- rectly or indirectly enhance food availability. In both Franklin and Wakulla counties, flows be- low approximately 600 m3/sec appear more closely related to the following year's landings than higher growout flows, i.e., the regression equation fits the data better at the low end of the flow spectrum (Fig. 3). Several factors may explain this phenomenon. Food availability may limit blue crab production at flows below a certain level but may not be limiting at flows above this level and, therefore, crab produc- tivity is not influenced by further increases in flow. Prey limitation at low flows may also lead to canni- balism, further limiting blue crab population size (Lipcius and Van Engel, 1990). The finding that more recent years produce a stronger correlation between blue crab landings and river flows was also observed in Georgia (Rogers et al., 19905). Total discharges from September to May (growout period) of five Georgia rivers were posi- tively correlated with landings (r2>0.8). Shorter time periods (the most recent 14 and 19 years of land- ings statistics) produced better correlations with flow than the full period of record (37 years). The authors concluded increased fishing pressure in more recent years resulted in only one year class being fished, and, thus, environmental effects were more obvious on a single year class in the shorter dataset. Similarly, that more recent landings for Franklin County were more highly correlated with Apalachicola River flows than landings for the longer 39-year period may reflect a trend toward harvesting a single year class. The significant 1- and 12-month time lags in Franklin and Wakulla County reflect similarities in catches between successive months and a seasonal component, respectively. The 12-month auto- correlation indicates that trends in landings occur at the same time of year (e.g., summer peaks) and should not be confused with an annual auto- correlation, which is indicative of a similarity in harvests between entire years. The positive corre- lations between unadjusted monthly flow and land- ings data correspond to the summer peak in blue crab landings following 3-5 months after the spring peak in flows, and low winter landings following low late-summer and fall flows. The negative correla- tions with 2-4 month time lags reflect fall low flows following peak summer harvests and high spring flows occurring after low winter harvests. The ab- sence of significant correlations between monthly landings and flows, once these data were adjusted to remove seasonal autocorrelations, indicates that residual (non-seasonal) variation in monthly flows is unrelated to the non-seasonal variation in mon- thly landings. Livingston (1991) found a positive contemporane- ous correlation between monthly Apalachicola River flows and blue crab abundances in trawl surveys conducted from 1972 to 1985. This finding corre- sponds to high juvenile abundances during high-flow months. The positive correlation in the present study between monthly flows and blue crab landings 3-5 months later may reflect the maturation of ju- veniles into adults in the summer, and thus the observed time lag in the correlation. The majority of the Apalachicola-Chattahoochee- Flint basin is in Georgia and is subject to different climatic conditions than are the other north Florida rivers examined, which may explain the relatively small correlations between Apalachicola River flows and flows on these other rivers. Georgia rainfall is more strongly correlated with Apalachicola River flows than Florida rainfall (Meeter et al., 1979). A consistent and important finding of the multivari- ate regression analyses was that Apalachicola flows were more highly correlated with Franklin, Wakulla, and Florida west coast landings of the next year than any other river drainage tested. Regressions comparing Apalachicola flows to west coast landings, after Franklin and Wakulla County landings were removed, were not significant, suggesting the influ- ence of the Apalachicola drainage is restricted pri- marily to Franklin and neighboring Wakulla County. Thus, there was no evidence supporting the hypoth- Wilber: Influence of Apalachicola River flows on Callmectes sapidus 187 esis of mass blue crab spawning near Apalachicola Bay and larval transport down the gulf coast of Flor- ida via the loop current (Oesterling and Evink, 1977). Several studies have addressed factors that influ- ence interannual variation in blue crab abundance, primarily concentrating on larval and post-larval recruitment (reviewed in Lipcius and Van Engel, 1990). Lipcius and Van Engel (1990) found high interannual, seasonal, and spatial variation in blue crab abundances in a 17-year fishery-independent dataset collected in the Chesapeake Bay. They ob- served that years with high blue crab abundances appeared to be dominated by the previous year class because peak catches occurred in the summer. Years with low abundances had peak abundances in the fall, suggesting the dominance of the new year class. This observation supports the contention that varia- tion in recruitment plays a major role in determin- ing interannual fluctuations. No interaction between annual abundance and seasonal peak catch was apparent for the Franklin or Wakulla County blue crab landings, which may indicate either the true absence of such a relation, the inadequacies of us- ing fishery statistics, or a difference in growth rates between the two regions that invalidates the use of the same analysis. Interestingly, the fishery-inde- pendent trawl data from the Chesapeake were sig- nificantly (r2=0.33) correlated with the commercial landings data. The influence of physical factors on blue crab abundances has been documented in other areas, such as a positive relationship between blue crab landings and freshwater inflows in Georgia (Rogers et al., 19905), an inverse relation between salinity and juvenile blue crab abundances on the Texas coast (More, 1969), and a positive relation between blue crab productivity and vegetated area in the Gulf of Mexico (Orth and van Montfrans, 1990). The positive correlation between blue crab landings and Apalachicola River flows of the previous year pro- vides additional evidence of the importance of fresh- water inflows to juvenile blue crabs. Apalachicola River flows have a significant impact on estuarine productivity, as indicated by commer- cial harvests of oysters (Wilber, 1992) and blue crabs. Although statistical correlations do not indi- cate the causal mechanisms underlying these asso- ciations, the river's influence on estuarine salinities as a mediating factor is deserving of further exami- nation. Undoubtedly, the Apalachicola River affects estuarine biota via mechanisms other than salinity (Livingston, 1991). Factors such as the transport of nutrients and organic matter, however, are unlikely to result in a significant correlation between low flows and oyster harvests two years later, unless food limitation is only measurably important for newly settled oyster spat. In addition, the majority of nutrient and detrital transport from the river occurs during high flow periods in the spring (Mattraw and Elder, 1982). There was no evidence that above-average flows were associated with either oyster or blue crab productivity. In both fisheries, flows on the low end of the spectrum were most sig- nificantly associated with landings. These signifi- cant correlations were positive and incorporated time lags, suggesting estuarine conditions during low minimum flow periods were not favorable for juveniles of either species. Acknowledgments The careful reviews of R. Hardy, G. Lewis, D. Meeter, R. Lipcius, P. Steele, and R Wilber are grate- fully acknowledged, as well as the technical support of J. Bennett, J. McKenna, G. Miller, and D. Tonsmeire. This work was supported by the North- west Florida Water Management District and the State of Florida's Surface Water Improvement and Management (SWIM) Program. Literature cited Funicelli, N. A. 1984. Assessing and managing effects of reduced freshwater inflow to two Texas estuaries. In V. S. Kennedy (ed. ), The estuary as a filter, p. 435-446. Lipcius, R. N., and W. A. Van Engel. 1990. Blue crab population dynamics in Chesa- peake Bay: variation in abundance (York River, 1972-1988) and stock-recruit functions. Bull. Mar. Sci. 46:180-194. Livingston, R. J. 1983. Resource atlas of the Apalachicola estuary. Florida Sea Grant College Publication No. 55, 64 p. 1991. Historical relationships between research and resource management in the Apalachicola River-estuary. Ecological Applications 1(4):361- 382. Livingston, R. J., G. J. Kobylinski, F. G. Lewis HI, and P. F. Sheridan. 1976. Long-term fluctuations of epibenthic fish and invertebrate populations in Apalachicola Bay, Florida. Fish. Bull. 74(2):311-321. Mattraw, H. C, and J. F. Elder. 1982. Nutrient and detritus transport in the Apalachicola River, Florida. U.S. Geol. Surv. Water-Supply Pap. 2196-C. Meeter, D. A, R. J. Livingston, and G. C. Woodsum. 1979. Long-term climatological cycles and popula- 188 Fishery Bulletin 92(1). 1994 tion changes in a river-dominated estuarine system. In R. J. Livingston (ed.), Ecological pro- cesses in coastal and marine systems. Marine Science 10:315-338. Mense, D. J., and E. L. Wen nor. 1989. Distribution and abundance of early life his- tory stages of the blue crab, Callinectes sapidus, in tidal marsh creeks near Charleston, South Carolina. Estuaries 12:157-168. More, W. R. 1969. A contribution to the biology of the blue crab (Callinectes sapidus Rathbun) in Texas, with a description of the fishery. Texas Parks Wildl. Dep. Tech. Ser. 1:1-31. Oesterling, M. L., and G. L. Evink. 1977. Relationship between Florida's blue crab population and Apalachicola Bay. In R. J. Livingston and E. A. Joyce (eds.), Proceedings of the conference on the Apalachicola drainage sys- tem; 23-24 April 1976, Gainesville, Florida. FL Mar. Res. Pub. 26:101-121. Orth, R. J., and J. van Montfrans. 1990. Utilization of marsh and seagrass habitats by early stages of Callinectes sapidus: a latitudinal perspective. Bull. Mar. Sci. 46:126-144. Perry, H. M. 1984. A profile of the blue crab fishery of the Gulf of Mexico. Gulf State Mar. Fish. Comm. No. 9, 80 p. Perry, H. M., and K. C. Stuck. 1982. The life history of the blue crab in Mississippi with notes on larval distribution: proc. blue crab colloquium; 18-19 October 1979, Biloxi, Mississippi. Gulf States Mar. Fish. Comm. 7:17-22. Steele, P. 1982. A synopsis of the biology of the blue crab Callinectes sapidus Rathbun in Florida: proc. blue crab colloquium; 18-19 October 1979, Biloxi, Mississippi. Gulf States Mar. Fish. Comm. 7:29-35. Wilber, D. H. 1992. Associations between freshwater inflows and oyster productivity in Apalachicola Bay, Florida. Estuarine, Coastal and Shelf Sciences 35:179-190. Wilkinson, L. 1990. SYSTAT: the system for statistics. SYSTAT, Inc. Evanston, IL, 676 p. Oocyte maturation in Hecate Strait English sole [Pleuronectes vetulus) Jeff Fargo Department of Fisheries and Oceans, Pacific Biological Station Biological Sciences Branch. Nanaimo, British Columbia V9R 5V6 Albert V. Tyler School of Fisheries and Oceans University of Alaska, Fairbanks, Alaska 99775 English sole, Pleuronectes vetulus, is an important component of the bottom trawl fishery in Hecate Strait, British Columbia, Canada. It is a small-mouthed flounder that feeds on sedentary inverte- brates associated with sandy sub- strate and is most common at depths of 80-150 m (Hart, 1973). The species is characterized by moderate growth (&=0.22), mortal- ity (M=0.20) and longevity (20 years) (Fargo, 1993). It recruits to the fishery at an age of four years, which is roughly equivalent to the age of sexual maturity (Ketchen, 1956; Tyler et al., 19871). Most of the exploited population is under 12 years of age (30-45 cm in length) (Fargo, 1993). Results from tagging studies (Ketchen, 1956; Fargo et al., 1984) and analysis of landing statistics and age composition data (Fargo, 1993) indicate that a single stock exists in Hecate Strait. Since 1955, abundance for this stock has fluctuated, primarily be- cause of changes in recruitment (Fargo, 1993). Factors influencing recruitment for this stock are poorly understood. Ocean tem- perature and circulation have 1 Tyler, A. V., J. Fargo, R. P. Foucher, and J. B. Lucas. 1987. Studies on the repro- ductive biology of Pacific cod and En- glish sole in Hecate Strait from the cruise of the FR/V W.E. Ricker, Novem- ber 25-29, 1986. Can MS. Rep. Fish. Aquat. Sci. 1937, 43 p. been found to influence spawning time and oocyte maturation for the stock off the Oregon coast (Kruse and Tyler, 1989). These authors postulated that 1) the rate of gonadal development for English sole was inversely related to summer bottom temperatures in the same manner as is somatic growth, and 2) spawning was de- layed by rapid increases in bottom temperature caused by upwelling. In Hecate Strait, where Ekman transport is weak, these tempera- ture changes may be brought about by the fall transition when strong winds from the south cause mixing of the warm surface wa- ters to depths of 150 metres (Dodimead, 19802). Relatively little information exists on spawn- ing time and egg development for the Hecate Strait stock. We inves- tigated oocyte growth and devel- opment to examine the length of the oocyte maturation period and the time and duration of spawn- ing for the English sole stock in Hecate Strait. Materials and methods Samples of English sole ovaries were obtained from research cruises and at ports-of-landing from commercial vessels between November 1987 and November 1990. The fish were caught with bottom trawls at five locations throughout Hecate Strait (PMFC Areas 5C-D, Table 1, Fig. 1). Length-stratified samples were collected to ensure that ovaries were obtained throughout the size range of fish collected. For each collection we attempted to sample fifteen sexually mature fish from each 5-cm length interval over a range of 30-50 cm, though this was not always possible. The minimum size fish (30 cm) from which an ovary was dissected cor- responds to the length at first maturity for this stock (Ketchen, 1956; Tyler et al., 19871). Total length and the condition of matu- rity for each fish sampled was re- corded. The right ovary was then removed and preserved in a buff- ered formalin-saline solution (Foucher et al., 19873). Sampling methods have been described in previous reports (Foucher et al., 19873; Tyler et al., 19871). A list of ovary samples examined is given by sample type and month in Table 1. Preserved ovaries were pre- pared for histological examination by soaking in Davidson's fixative for approximately 24 hours. Sub- sequently, tissue sections were dissected from the anterior por- tion of the ovary (which contained the greatest amount of eggs), em- bedded in paraffin wax, sectioned at 5 u, stained with haematoxylin and counterstained with eosin (Yasutake and Wales, 1983). Oocyte diameter was measured with a light microscope calibrated 2 Dodimead, A. J. 1980. A general review of the oceanography of the Queen Char- lotte Sound-Hecate Strait-Dixon En- trance region. Can. MS. Rep. Fish. Aquat. Sci. 1574, 248 p. 3 Foucher, R. P., J. Fargo, and J.B. Lucas. 1987. Cruise of the FV Nucleus. Janu- ary 5-17, 1987 to Hecate Strait to study reproductive biology of Pacific cod and English sole. Can. MS Rep. Fish. Aquat. Sci. 1941, 25 p. Manuscript accepted 8 October 1993. Fishery Bulletin 92:189-197 (1994) 189 190 Fishery Bulletin 92(1), 1994 Figure 1 Location of trawling grounds in the study area, Hecate Strait, British Columbia, Canada. to the nearest 5 p, or with a projection microscope calibrated to the nearest 4 u. Three hundred oocytes were measured from at least one fish for every cm length interval for each sample (Table 1). Measure- ment of 300 oocytes per fish was necessary to pro- vide complete information on the size composition of developing oocytes. Only oocytes that had been sectioned through the nucleus, close to the center of the oocyte, were measured. Mean diameter was es- timated as the mean of the minimum and maximum diameters for each oocyte (Foucher and Beamish, 1980). For smaller oocytes (10-20 p), precision of the measurement was lower because of distortion of the oocyte by surrounding maturing oocytes (Dunn, 1970). A description of the histological stage of oo- cyte development (Fargo and Sex- ton, 19914) was also recorded. We were unable to obtain oocyte measurements from ovaries col- lected from ripe fish in October and November 1990. These samples were taken from com- mercial vessels at ports of land- ing. Ovaries from these samples had combinations of hydrated and non-hydrated oocytes with many burst cells. These fish had been held in chilled seawater for sev- eral days prior to sampling, prob- ably exacerbating the state of hy- drated oocytes and causing them to burst. Since oocyte diameter data for these samples would have been biased (because most measurable oocytes would not have reached the hydrated state) the slides from these samples were used only to assess the his- tological stage of the oocytes. This problem did not occur with the November 1987 sample collected at sea on a research vessel. Prior to statistical testing of the data, we tested oocyte size distri- butions for normality using the Shapiro-Wilk test. We applied two sample £-tests to test for differ- ences in the mean diameter of previtellogenic and vitellogenic oocytes between months within years and among years. We used linear regression to investigate the relation 1) between fish length and mean oocyte diameter within months and 2) between fish length and mean oocyte diameter at the time of spawning. Results Oocyte development Ovaries were examined from 174 fish (Table 1) caught at five locations in Hecate Strait (Fig. 1). The sampling period encompassed seven different months over three years. Descriptions and micro- 4 Fargo, J., and T. Sexton. 1991. A quide to the ovarian histol- ogy of English sole iParophrys vetulus). Can. MS. Rep. Fish. Aquat. Sci. 2133, 19 p. NOTE Fargo and Tyler: Oocyte maturation in Pleuronectes vetulus 191 graphs of the stages of matura- tion for English sole oocytes have been summarized by Fargo and Sexton (1991).4 Ex- amples of oocyte size distribu- tions for fish of different lengths sampled during the same period, August 1988, are presented in Figure 2. For all sizes of English sole collected, we observed the simultaneous presence of only two modes in the oocyte size distributions. The smaller mode (10-150 u> consisted of previtellogenic oo- cytes and the larger mode (150-500 u) of vitellogenic oo- cytes. No previtellogenic oo- ctyes >150 u were observed. The size modes for previtel- logenic oocytes were similar among fish ranging in size from 33 to 46 cm. The mode for vitellogenic oocytes shifted to the right (increased) with in- creasing fish length. Vitellogenic oocytes in- creased in size from early sum- mer until they became hy- drated prior to spawning in the fall (Fig. 3). We observed no trend in the size composition of previtellogenic oocytes over the same period. As the month of spawning was approached a complete separation between the two modes became appar- ent. The irregular shape of the modal distribution for vitel- logenic oocytes in Figure 3 is caused by combining data for fish of different lengths and de- veloping at different rates. The more normal distribution for this mode during the month of spawning is due to two factors. First, the size range of fish for this sample was smaller than for other samples and, second, egg diameter at the time of spawning was similar for fish of different length. Fargo and Sexton (1991)4 described the events of oocyte maturation for English sole in detail. Briefly, Table 1 A summary of ovary samples examined in the study of oocyte matura- tion in Hecate Strait English sole (Pleu ronectes vetulus) Length class (cm) (No. ovaries Date Sample type Location examined) 7-13 January 1987 Research cruise Two Peaks 30-34 (2) White Rocks 35-39 (6) 40-44 (5) 45-49 (5) 50-54 (4) 55-59 ( 1 ) Total (23) 19 January 1988 Port sample White Rocks 30-34 ( 1 ) 35-39 (1) 40-44 ( 1 ) Total (3) 17 March 1987 Research cruise Horseshoe 30-34 (2) 35-39 (2) 40-44 (3) 45-49 (2) 50-54 (1) Total (10) 16 March 1988 Port sample Horseshoe 30-34 ( 1 ) 35-39 (4) 40-44 (2) 45-49 (1) Total (8) May 5 1988 Port Sample Horseshoe- 30-34 ( 1 ) White Rocks 35-39 (1) 40-44 (2) 45-49 (6) Total (10) 6 June 1987 Research cruise Horseshoe- 30-34 ( 1 i Bonilla 35-39 ( 3 1 40-44 (3) 45-49 (3) 50-54 (3) Total (12) 2 June 1988 Port sample Horseshoe 30-34 ( 1 ) 35-39 (3) 40-44 (3) 45-49 (3) 50-54 (2) Total (12) 27 August 1987 Research cruise Horseshoe 30-34 (1) 34-39 ( 7 ) 40-44 (5) 50-54 1 1 ) Total (14) 192 Fishery Bulletin 92(1). 1994 Table 1 (continued) Length class (cm) (No. ovaries Date Sample type Location examined) 22 August 1988 Port sample Horseshoe Total 30-34 (2) 34-39 (5) 40-44 (8) 45-49 (6) 50-54 (1) (22) 28 August 1990 Port sample Two Peaks Total 35-39 (4) 40-44 (4) 45-49 (4) 50-54 (1) (13) 27 January 1988 Port sample Two Peaks- Butterworth Total 30-34 (4) 35-39 (2) 40-44 (3) 45-49 (1) 50-54 (1) (11) 19 January 1990 Port sample Horseshoe Tota 30-34 (1) 35-39 (6) 40-44 (8) 45-49 (1) (16) 5-6 November 1987 Research cruise Horseshoe- 30-34 (5) Butterworth- 35-39 (7) White Rocks 40-44 (4) 45-49 (2) 50-54 (1) 3 November 1990 Port sample Butterworth Total 30-34 (D 35-39 (ll 40-44 (2) 45-49 (1) 50-54 (1) (6) vitellogenesis occurred when oocytes reached a di- ameter of about 150 p. Vacuolization occurred in oocytes ranging from 180 u to 250 p Deposition of yolk in the outer cortex occurred in oocytes ranging in size from 200 p to 430 p, and hydra ted oocytes ranged in size from 375 (i to 550 p. We began our investigation of the timing and duration of oocyte maturation by examining the size composition and histological stage of oocytes col- lected from fish sampled between January and No- vember. Ovaries examined from 68 of 72 fish col- lected during winter and spring (January 1987-88 and March 1987-88) contained mainly pre- vitellogenic oocytes. The fish examined from the January samples contained previtellogenic oocytes only. Four of 22 fish examined from samples collected during the month of March contained vitellogenic oocytes. Three of these (36-40 cm in length) con- tained vitellogenic oocytes that were hydrated and translucent (405-429 p mean diameter). The fourth fish (46 cm in length) contained oocytes that had recently undergone vitello- genesis (mean diameter=230 p). Vitellogenesis for most fish occurred in the early summer. In May 1988, we observed vitellogenic oocytes in six of nine fish examined, ranging from 40 to 49 cm in length. All of these oocytes were in the early stages of development, prior to vacuolization, with mean diameters ranging from 174 to 263 p. Smaller fish (length range 33-42 cm) con- tained previtellogenic oocytes only . In June (1987, 1988) vi- tellogenic oocytes, ranging in mean diameter from 178 p to 269 p, were present in 23 of 24 fish examined (length range 36-52 cm). Vitellogenic oocytes in one fish of 52 cm were at an advanced stage of development (mean diameter=252 p), with yolk granules formed in the outer cortex. The relation be- tween mean diameter of vitel- logenic oocytes and fish length was not significant for the months of May ( 1988) and June (1987, 1988) (linear regression, P>0.1 for all three, n=6, 11, 12) By August the oocytes in some of the larger fish (45-50 cm) were nearing hydration. Mean diameters for vitellogenic oocytes from fish sampled in August (1987, 1988, 1990) ranged from 226 p to 429 p. There were significant, positive linear relationships between fish length and mean oocyte diameter for all of these samples (Table 2, Fig. 4). The size distributions for previtellogenic and vitellogenic oocytes did not differ significantly (Shapiro- Wilk test, P<0.05) from that of the normal distribution for any of the following cases. There was no significant difference in mean diameter of previtellogenic oocytes for the same months across NOTE Fargo and Tyler: Oocyte maturation in Pleuronectes vetulus 193 33 cm prevHellogenlc 60 110 160 210 260 310 360 410 460 Oocyte diameter (microns) 42 cm prevHellogenlc 60 110 160 210 260 310 360 410 460 Oocyte diameter (microns) 38 cm 10 60 110 160 210 260 310 360 410 460 Oocyte diameter (microns) 46 cm | vltellogenlc prevHellogenlc 10 60 110 160 210 260 310 360 410 460 Oocyte diameter (microns) Figure 2 Oocyte size compositions determined from ovary samples collected from Hecate Strait English sole {Pleuronectes vetulus) in August 1988. the two years (Table 3). However, there were signifi- cant differences in mean diameter for previtellogenic oocytes among months within both years (Table 4). No obvious trend in mean diameter over time was apparent for previtellogenic oocytes. There were sig- nificant differences in the rate of oocyte development between 1987 and 1988 (Table 3). The mean diam- eter of vitellogenic oocytes in June and August of 1987 was significantly larger than for the same months in 1988, suggesting that vitellogenesis oc- curred earlier in 1987 than in 1988. There were also significant differences in the mean diameter of vitellogenic oocytes among months within years (Table 5). The mean diameter of vitellogenic oocytes increased significantly, coinciding with advancing oocyte development, between June-November in 1987 and June-October in 1988. Spawning Ovaries obtained from spawning fish (October 1988, 1990 and November 1987, 1990) were examined to investigate 1) size-dependent spawning and 2) the relation between fish length and egg diameter at the time of spawning. For the October 1988 sample, we observed the presence of vitellogenic oocytes only in fish smaller than 40 cm. The mean diameter of vitellogenic oocytes in these fish ranged from 287 to 408 p. Fish ranging in length from 43 to 52 cm con- tained spent ovaries with previtellogenic oocytes only. Thus, we concluded that the larger fish had spawned prior to the time of the sample collection. In the October 1990 sample, taken two weeks ear- lier than the 1988 sample, some of the fish larger than 40 cm contained hydrated oocytes while oth- ers had spent ovaries with resorbing oocytes, sug- gesting that they were spawning in early October. Oocytes examined from samples collected in Novem- ber (1987, 1990) also indicated that larger fish had spawned previous to this time. Fish larger than 42 cm contained only pre-vitellogenic oocytes and there was no sign of resorbing oocytes. Most smaller fish were in spawning condition during this month. Vitellogenic oocytes were present in fish ranging from 30 to 42 cm. Mean diameter ranged from 373 to 483 p and these oocytes were hydrated and trans- 194 Fishery Bulletin 92(1). 1994 January n=1001 10 60 110 160 210 260 310 360 410 460 Oocyte diameter (micron*) May previtellogenic vltellogenk; 10 60 110 160 210 260 310 360 410 460 Oocyte diameter (microns) June prevttellooenlc 1/ n=676 ^ vttellogenlc Ik 10 60 110 160 210 260 310 360 410 460 Oocyte diameter (microns) 600 — 500 J.400 ?300 V I 200 "- 100 0 August prevttellooenlc 1/ n= )564 vltellogenlc / 10 60 110 160210260310360410460 Oocyte diameter (microns) October (spawning) prevttellogenlc n=377 vttellogenlc A 10 60 110 160 210 260 310 360 410 460 Oocyte diameter (microns) Figure 3 Oocyte size composition determined from ovary samples collected from Hecate Strait English sole iPleuronectes vetulus) during January-October in 1988 (samples combined). lucent. We then combined all the data on mean egg diameter for spawning fish and there was no relationship between mean egg di- ameter at the time of spawning (hydrated and translucent) and fish length (linear regression, P>0.1, rc = 19). Discussion Oocyte development Dunn and Tyler ( 1969) and Dunn (1970) determined the length of time required for oocyte matura- tion in winter flounder iPleuronectes americanus). They observed two size modes of previtellogenic oocytes at any particular time. They documented the rate of increase in size for these modes for three consecutive years and concluded that the oocyte maturation period for this species was three years. We observed only a single mode for both previtellogenic and vitellogenic oocytes in fish sampled during all the months examined in our study. Johnson et al. ( 1991) reported similar results in their study of Puget Sound English sole. If oocytes Table 2 Linear regression statistics for the relationship bet ween vitell ogenic oocyt e mean diameter and fis h length "or Engl sh sole (Pleuronectes vetul us) for the month of August 1987 1988, and 1990. Degrees o f Year freedom F-statistic P Regression equation' r 1987 13 10.72 0.007 Y = 122 + 5.93X 0.687 1988 20 20.93 - n iiiiiii Y = -93 + 9.44X 0.910 1990 12 44.01 . I) 1)001 Y = - 237 + 12. 6X 0.724 ' Y = oocyte mean diameter (a.). X = total length of fish (cm ) produced in year i were spawned in year i+1, we would expect to see two size classes of immature oocytes in year i+1, corresponding to those oocytes that were produced in year i (large immatures) to be spawned in year i+1 and those that were pro- duced in year i+1 (small immatures) to be spawned in year i+2. The fact that there were no significant differences in the mean diameter of previtellogenic oocytes for the same months in consecutive years (1987-88) suggests that the oocyte maturation pe- riod for Hecate Strait English sole is probably one year. NOTE Fargo and Tyler: Oocyte maturation in Pleuronectes vetulus 195 500 450 400 350 300 250 200 150 100 50 0 500 450 T 40° I 350 \ 4 300 | 250 -I 200 § 100 50 0 30 500 450 1" 400 § 350 4 300 I 250 -1 200 -£. 150 <§ 100 50 0 30 Aub-37 ° g ° 6 o o ° o; o provttolloflonlc • vltsllogenic — regression 30 35 40 45 50 Total length (cm) 55 Aug-88 oggooe °86 oo 35 40 45 Total length (cm) 50 55 Aug-90 • • OO OOq nOO 35 40 45 Total length (cm) 50 55 Figure 4 Mean oocyte diameter vs fish length determined from ovary samples collected from Hecate Strait English sole {Pleuronectes vetulus) during the month of August, 1987. 1988, and 1990. We also found no trend in the mean size of previtellogenic oocytes among months within years, contrary to results reported by Dunn and Tyler (1969). One explanation for this is that the recruit- ment of small immature (previtellogenic) oocytes from the germinal epithelium is a continual process for Hecate Strait English sole. Alternatively, there may be a short time period, following spawning for example, during which previtellogenic oocytes recruit and quickly grow to a size of around 80 p. Additional work is needed to resolve these possibilities. Table 3 Results of two sample r-tests of mean diameters of previtellogenic and vitellogenic oocytes for English sole (Pleuronectes vetulus) determined from samples collected during the same month in 1987 and 1988. Month and year n mean diameter (microns) P previtellogenic January 1987 January 1988 6,264 1,001 69.0 69.4 >0.1 March 1987 March 1988 1,603 1,737 59.2 59.8 >0.1 June 1987 June 1988 1,389 1,132 72.4 72.9 >0.1 August 1987 August 1988 1,812 2,071 66.7 65.8 >0.1 vitellogenic June 1988 June 1988 953 1,029 219.1 203.3 <0.0001 August 1987 August 1988 1,774 3,584 362.1 318.1 <0.0001 Spawning In general larger fish produced yolk earlier and spawned earlier than smaller fish. Most of the spawning fish were obtained from samples collected in October and November but there was also evi- dence of spring (March) spawning for smaller fish. Egg size at the time of spawning did not appear to be dependent on fish length. However, there is some evidence from this study to suggest a possible mini- mum size limit for eggs at the time of spawning. That is, the difference in the mean diameter of vitellogenic oocytes between smaller and larger fish decreased over time until there was no apparent difference at the time of spawning. Observations made during this study indicate that atresia was not as prevalent for Hecate Strait English sole as that reported for English sole in Puget Sound by Johnson et al. (1991). Marine fish species show wide variability in the reproductive process, which enables them to miti- gate the uncertain conditions in the marine environ- ment (Murphy, 1968; Roff, 1981). English sole dem- onstrate considerable phenotypic plasticity with regard to spawning. In Hecate Strait the spawning season extends from early fall through the follow- 196 Fishery Bulletin 92|1), 1994 Table 4 Results of two sample t-tests of the mean diameter ( |i ) of previtellogenic oocytes in English sole (Pleuronectes vetulus) among months for samples collected in 1987 and 1988. Year and Month January March June August November 1987 January P<0.0001 P<0.0001 P=0.0004 P<0.0001 (n=6264, 7=69. Out March P<0.0001 P<0.0001 P<0.0001 June P<0.0001 P<0.0001 (/i = 1132, 7=72. 9u) August — P<0.0001 (/i = 2071, 7=65. 8u) October — — (o = 1500, 7=56.9u) Table 5 Results of two sample r-tests of the mean diameter (u) of vitellogenic oocytes in E riglish sole (Pleuronectes vetulus) among months for samples collected in 1987 and 1988. Year and month June August November Year and month May June August October 1987 1988 June — <0.0001 <0.0001 May >0.1 <0.0001 <0.0001 (n= 953, 7=219.1u) (n=191, .7=201. 4u I June — <0.0001 <0.0001 August <0.0001 (/i = 1029. 7=203. 3u) (re=1774, 7=362. lu) August (7i=3584, 7=318. In) <0.0001 <0.0001 November — — — October — — — (n= 488, i=413.7m (71=710, 7=342. lu) ing spring. Johnson et al. (1991) reported a similar spawning period for Puget Sound English sole as did Kruse and Tyler ( 1989) in their study of English sole off the Oregon coast. This reproductive strategy may increase the probability of encountering favorable conditions for larval survival by spreading the re- productive effort over the longest possible time span. Based on our results it is unlikely that cohort-spe- cific spawning occurs as in Pacific herring, Clupea pallasi (Ware and Tanasichuk, 1989), and Norwe- gian Atlantic herring, Clupea harengus (Lambert, 1990). However, in view of the relation between oocyte maturation and fish length and the duration of the spawning period, it is possible that first time NOTE Fargo and Tyler: Oocyte maturation in Pleuronectes vetulus 197 spawners spawn at a different time than the rest of the stock. We can suggest no mechanism to account for this and more data are required to corroborate these results. There is also evidence of interannual variability in oocyte maturation and this process appears to be size-related. Smaller fish matured later and spawned later than larger fish. Our results indicate that the time of peak spawning and the duration of the spawning season are variable from year to year. The results from this study provide baseline infor- mation for an investigation of the recruitment biol- ogy of this stock. Acknowledgments We wish to acknowledge John Bagshaw, Serge Villeneuve, Tammy Laberge, Christina Horvath, Tracy Sexton, and Corinne Kikegawa for their aid in preparing the slides for histological examinations and photography of specimens. Ron Tanasichuk and Doug Hay reviewed the manuscript and provided advice regarding the spawning characteristics for the species. The scientific editor and three anony- mous reviewers provided a number of suggestions which improved the paper. Literature cited Dunn, R. S. 1970. Further evidence for a three year oocyte maturation time in the winter flounder (Pseu- dopleuronectes americanus). J. Fish. Res. Board Canada. 27:957-960 Dunn, R. S., and A. V. Tyler. 1969. Aspects of the anatomy of the winter floun- der (Pseudopleuronectes americanus) with hypoth- eses on oocyte maturation time. J. Fish. Res. Board Canada 26:1943-1947. Fargo, J. 1993. Flatfish. In B. M. Leaman and M. Stocker (ed.), Groundfish stock assessments for the west coast of Canada in 1992 and recommended yield options for 1993. Can. Tech. Rep. Fish. Aquat. Sci. 1919:95-131. Fargo, J, R. P. Foucher, S. C. Schields, and D. Ross. 1984. English sole tagging in Hecate Strait, R/V G.B. REED, June 6-24, 1983. Can. Data Rep. Fish. Aquat. Sci. 427, 49 p. Foucher, R. P., and R. J. Beamish. 1980. Production of nonviable oocytes by Pacific hake (Merluccius productus). Can. J. Fish. Aquat. Sci. 37:41-47. Hart, J. L. 1973. Pacific fishes of Canada. Fish. Res. Board Can. Bull. 180, 740 p. Johnson, L. L, E. Casillas, M. S. Myers, L. D. Rhodes and O. P. Olson. 1991. Patterns of oocyte development and related changes in plasma 17-B estradiol, vitellogenin and plasma chemistry in English sole Parophrys vetulus Girard. J. Exp. Mar. Biol. Ecol. 152: 161-185. Ketchen, K. S. 1956. Factors influencing the survival of the lemon sole (Parophrys vetulus) in Hecate Strait, British Columbia. J. Fish. Res. Board Canada, 13(5): 647-694. Kruse, G. H., and A. V. Tyler. 1989. Exploratory simulation of English Sole (Parophrys vetulus) recruitment mechanisms. Trans. Am. Fish. Soc. 118:101-118. Lambert, T. C. 1990. The effect of population structure on recruit- ment in herring. J. Cons. int. Explor. Mer 47:249-255. Murphy, G. I. 1968. Pattern in life history and the environment. Am. Nat. 102:391-403. Roff, D. A. 1981. Reproductive uncertainty and the evolution of iteroparity: why don't flatfish put all their eggs in one basket? Can. J. Fish. Aquat. Sci. 38:968- 977. Ware, D. M., and R. Tanasichuk. 1989. Biological basis of maturation and spawning waves in Pacific herring (Clupea harengus pallasi). Can. J. Fish. Aquat. Sci. 46(101:1776- 1784. Yasutake, W. T., and J. H. Wales. 1983. Microscopic anatomy of salmonids: an atlas. Fish. Wild. Ser. U.S. Dep. Int. Res. Pub. 150, 189 p. Estimation of weight-length relationships from group measurements William H. Lenarz Tiburon Fisheries Laboratory National Marine Fisheries Service, NOAA 3 1 50 Paradise Drive. Tiburon, CA 94920 Catch sampling provides data that are basic to fisheries re- search and is often an important component of research budgets. Samplers typically select fish ran- domly, measure length, remove ageing structures, and determine sex for each individual. In many schemes for sampling commercial (e.g., Sen, 1986; Tomlinson, 1971) and survey catches (e.g., Gun- derson and Sample, 1980), sample weight is needed to expand the sample results to the total catch. Individual weights are usually not needed to satisfy the main objectives. Often only the aggre- gate weight of the sample is taken to save time, and if at sea, to avoid difficult logistics. While sampling costs are easily justified by program objectives, scientists frequently use the data for addi- tional research. Investigators often use weight- length relations to study possible correlations between condition of fish and environmental factors or population density (e.g., Pat- terson, 1992). A literature search revealed only two previous devel- opments of methods of estimating weight-length relations from samples of individual lengths and aggregate weights (WLRAW). Cammen (1980) used a general nonlinear regression program from the BMDP package (Dixon, 1983) as a WLRAW method. He tested the method with simulated data and compared the results of regression using unweighted ob- servations to using observations weighted by the inverse of sample weights, and with various esti- mates made when individual weights were known. Since the data were simulated, assuming a multiplicative error term, it would have been more appropriate to use the inverse of sample weight squared for weighting. The non- linear method produced good fits to the simulated data, and weighted parameter estimates were closer to the true values than unweighted estimates. Damm ( 1987) developed two non- linear WLRAW methods. One method is a biased approxima- tion, and his report indicated that the other method did not always produce estimates of the param- eters. In this note I describe a new WLRAW method, compare it with Cammen's method, explore error term characteristics, and describe bootstrap estimates of confidence limits of estimates. The methods of Damm (1987) were not studied because his biased approximation method requires as much calcula- tion as my new method and his other method does not always work. Methods The relation between expected weight and length of an indi- vidual fish is usually assumed to be the power equation, E(Wl) = alfl Where V^ = weight of fish i, a - parameter. (1) L, = length of fish i, p = parameter. For the new WLRAW method I modeled the weight-length rela- tionship as W, flK) + £, (2) J i=i where W = L. = e. = T = average weight of fish in sample j, number of fish in sample j, length of fish i in sample j, error term for sample j, 1, ■ • • , T, number of samples. I assumed that error was additive because under field conditions much of the error was due to lim- its to the accuracy in measure- ment of sample weights. Because the dependent variable in Equa- tion 2 was a sample average, its variance should contain a compo- nent which is proportional to the inverse of n . Thus in the new es- timation procedure, I weight each observation by n to stabilize the variance. I made the assumption that, after weighting by sample size, error was random and inde- pendent of,/. The new method treated esti- mation of parameters of (Eq. 2) as a separable least-squares problem (Seber and Wild, 1989). For a trial value of (3 (P'l, y was calculated for each sample, /,=<2X>/», (3) With the new notation, Equation 2 becomes W- = a y ; + e j~ (4) Manuscript accepted 16 August 1993 Fishery Bulletin 92:198-202 (1994) 198 NOTE Lenarz: Estimation of weight-length relations from group measurements 199 I then obtained an estimate of a (a') corresponding to p" by using the standard least squares linear re- gression with zero intercept method. I used a non- linear least squares procedure to obtain the estimate of (3 ( B )■ This procedure was analogous to finding the transformation, Lf, that minimized the sum of squares about the linear regression (Eq. 4). Using this procedure, I estimated brackets for ensuring that the searching range included P with the proce- dure MNBRAK (Press et al., 1989). Then I used the iterative procedure BRENT (Press et al., 1989) to obtain the final estimate. BRENT uses parabolic interpolation to minimize the sum of squares as a function of (3'. Convergence is assumed when the procedure does not change the value of P' more than a tolerance specified by the user. As previously stated, observations were weighted by n to stabilize the variance. I implemented the WLRAW method in double precision using Sun FORTRAN for a Sun SPARC2 work station. Bootstrap approximations of confidence intervals about the line were calculated for the new method. The literature contains a variety of bootstrap meth- ods proposed to approximate confidence intervals (e.g., DiCiccio et al., 1992). I used the nonparamet- eric BC method of Efron (1987) because it often a produces good results and is relatively easy to use. BCa stands for accelerated bias corrected boot- strap confidence intervals. Efron (1987) showed that, in the parametric case, the method is approximately correct if a transformation to a normally distributed variable exists. The transformation does not need to be known and the variance does not need to be con- stant. While the correctness of the BC has not been a mathematically proven for nonparametric cases, such as the WLRAW, Efron (1987) stated, "...empirical re- sults look promising." The BCa confidence limits of an estimate of parameter 8, 0, are IBS(N(z[a]))<6J ;=i (7) where U , 3(A) ef]-e A estimate of 6 whenyth sample has a very small amount of extra weighting (A). If a and z0 are zero, then Equation 7 becomes the percentile method that is the most frequently used bootstrap method in the fisheries literature (e.g., Sigler and Fujioka, 1988). I chose to approximate 90% confidence bands rather than 95% or 99% bands because 90% non- parametric bootstrap intervals tend to perform bet- ter than intervals that cover a wider portion of the distribution (Efron, 1988). Following the advice of Efron, I used 1,000 bootstrap replicates. Cammen (1980) used the general nonlinear re- gression program of BMDP to estimate the param- eters of Equation 2, except that he assumed that the error term is multiplicative and used total sample weight instead of average weight as the dependent variable. The BMDP program uses the Gauss-New- ton algorithm. I used the same algorithm in the nonlinear regression program of the SAS package (SAS Institute Inc., 1989) on a Sun SPARC2 to com- pare parameter estimates and execution times with the new method. Since the correct error model is not known, I also estimated the parameters using no _weighting__and weight set to 1/W ,1/W,2, nlIWr and n] I W", and compared asymptotic stan- dard errors of the parameter estimates. The new es- timation procedure is simpler than the Gauss-New- ton approach because it searches for the least squares by iteratively changing the value of one parameter instead of two. I used data collected on chilipepper rockfish {Sebastes goodei) by a cooperative landing sampling program of the California Department of Fish and 200 Fishery Bulletin 92|1). 1994 Game and National Marine Fisheries Service to examine utility of the WLRAW method. Samplers collected two groups of fish from each sampled land- ing. For each group a container that holds 22.7 kg of fish was filled with fish regardless of species. Then the sampler obtained total group weights to the nearest lb (0.45 kg) for each species and the total length of each fish was measured to the nearest mm. I converted weights to kg. I changed lengths to deci- meters to minimize potential scaling problems in the computations. Before using the WLRAW method, I combined groups within a landing because they may not be independent. I first used data for all months during 1991 from all ports between Morro Bay and Crescent City, California, to develop, test, and time the software. Results of the test runs are described briefly in the Results and Discussion section. More detailed re- sults are presented for a more typical application of the method. Investigators are more likely interested in results from a smaller number of samples taken from more restrictive scales of time and area than from data sets like the one used in the preceding example. I used data for chilipepper rockfish taken during July and August 1991 from Morro Bay to illustrate use of the method. Results and discussion The data from all ports consisted of measurements from 7,687 fish taken in 186 samples. The procedure required 1.6 seconds, compared with 18.8 seconds for the Gauss-Newton method. The Gauss-Newton and new methods produced parameter estimates that were identical to six decimal places. Predicted weights were very close to the results of Phillips (1964), who used data from individually measured fish. Sums of squares plotted against P' indicated that there were no local minima. Residuals were not related to weight, indicating that the additive error assumption is correct. Sometimes transformation of (3' to ln(P') when estimating parameters of power equations avoids problems due to curvature (Rat- kowsky, 1983). Transformation was tried and pa- rameter estimates were identical to the results when P' was not transformed. When P' was transformed, the procedure required more time to complete, so the transformation was not used. Data were available for 583 fish taken from 13 samples taken in Morro Bay, during July and Au- gust 1991. There were no strong trends between the residual and expected weight (Fig. 1). There was a tendency for absolute values of residuals to be nega- tively correlated with the number offish in a sample (Fig. 2A). The tendency was reduced when residu- 004 D 0.02 Residual (kg) 8 o B o D D D D D n -0.04 -0.06 D 04 0.5 06 07 0.8 0.9 1 Expected Weight (kg) Figure 1 Residual of average weight (kg) as a function of expected weight (kg) for chilipepper rockfish (Sebastes goodei) collected in samples taken from Morro Bay during July and August 1991. als were multiplied by sn , as expected under the assumption that variance is proportional to the in- verse of sample size (Fig. 2B). Also, n produced the lowest asymptotic standard errors of the parameter estimates of the six weighting factors explored (Table 1). The results shown in Table 1 and Figures 1 and 2 indicated that the additive error model with weighting by n was appropriate for these data. Bootstrap estimates of standard error using the new method were higher than asymptotic estimates us- ing the Gauss-Newton method. The bootstrap and asymptotic normal confidence intervals were narrow and similar within the range of most observed av- erage weights but diverged when expected weight was greater than 0.75 kg even though individual fish of larger size occurred in many of the samples (Table 2). The bootstrap confidence intervals were skewed at the larger sizes. However, the bootstrap estimates of absolute bias were less than 0.01 kg except they were -0.01 kg for 450-mm fish and -0.02 kg for 500-mm fish. All estimates of the absolute value of a were about 0.015, which indicated that a could have been ignored for this set of data. The new WLRAW method performed well. Good fits to the data were obtained and the residuals agreed with the assumptions. Approximate confi- dence limits indicated that precise estimates of ex- pected weight are obtained with a small number of samples under field conditions for sizes of fish within the range of most observed average weights. The method is fast when used on a work station or on a modern personal computer. The new method is 10 times faster than using the Gauss-Newton ap- NOTE Lenarz: Estimation of weight-length relations from group measurements 201 0.04 A D 0.02 0 •0.02 a D a D n o O D a ■0 04 ■0.08 -0.08 D 1 1 i ■ 40 60 Sample Size 0.2 B a fo, CO * o o n L> D to D ■o i-o.t ir o D o D O -0.2 -0.3 D _l 1 - 1 40 60 Sample Size Figure 2 (A) Residual of average weight (kg) as a function of sample size for chilipepper rockfish (Sebastes goodei ) collected in samples taken from Morro Bay during July and August 1991. (B) Residual multi- plied by yrij as a function of sample size. proach with a standard statistical package. Some of the difference is probably due to the overhead in- volved with using the statistical package. When computationally intensive methods such as bootstrapping are used, time saved by using the new method is significant. The widening confidence limits for expected weights beyond the range of most observed average weights indicated use of expected weights beyond the observed range is extrapolation and should not be done. This also applies to comparison of param- eter estimates from different sets of data. If the range of observed average weights differ much among the data sets, comparison of parameter esti- mates is not meaningful. Estimates of the two pa- Table 1 Estimates of standard errors of parameter esti- mates of weight-length model for chilipepper rock- fish iSebasted goodei) collected from Morro Bay during July and August 1991. The Gauss-New- ton method was used with observations weighted by six factors to estimate the parameters, and the new method with rij as the weighting factor. As- ymptotic standard errors are shown for the Gauss- Newton method and bootstrap standard errors for the new method. Coefficients of variation of the pa- rameter estimates are shown in parentheses. Standard error Weighting factor Gauss-Newton method none 0.0028 (0.30) n, _ 0.0019 (0.21) raj/Wj 0.0020 (0.20) n/W,2 0.0022 (0.20) 1/W, 0.00.30 (0.29) 1/W,: New method 0.0032 (0.28) 0.0046 (0.50) 0.2159 (0.07) 0.1489 (0.05) 0.1528 (0.05) 0.1547 (0.05) 0.2129 (0.07) 0.2069 (0.07) 0.2211 (0.07) Table 2 Expecte d weigh ts for chilipepper roc kfish iSebastes goodei ) collected from Morro Bay dur- ing July and August 1991, and 909t confidence about th 3 line. C onfidence limits were approxi- mated us \>iv, 1 he bootstrap BC (bootstrap ) and the asymptotic normal method 3 ( normal ) . Ex- pected weights were ca culatec from the esti- mated weight- ength relation (0.0091819 Length3 L 758673 , Confidence limits Normal Bootstrap Total Expected length weight Lower Uppei Lower Jpper (dm) (kg) (kg) (kg) (kg) (kg) 3.00 0.30 0.28 0.32 0.28 0.33 3.50 0.49 0.47 0.51 0.47 0.50 3.75 0.61 0.60 0.62 0.60 0.62 4.00 0.75 0.74 0.76 0.73 0.76 4.50 1.09 1.04 1.14 0.99 1.12 5.00 1.52 1.42 1.63 131 1.61 rameters of the weight-length relation are highly correlated even when individuals are weighed and standard linear regression is used (Lenarz, 1974). Thus, regardless of the type of data or statistical 202 Fishery Bulletin 92(1). 1994 procedure, I recommend comparison of weight- length relations among data sets by comparison of expected weights of fish at sizes within the range of observed average weights common to all data sets of interest. The results of this study suggest that an additive error term is more appropriate than a multiplica- tive error term for modeling weight-length relations. Most previous studies have assumed multiplicative error, which is implied when the log-log transforma- tion is used to estimate parameters of the model from individually measured fish by linear regres- sion. The multiplicative error assumption has not been demonstrated correct even when data are available from fish weighed individually. While good fits to data are usually obtained under the multi- plicative assumption, if the assumption is not valid, statistical inferences may be erroneous. Pienaar and Thomson (1969) assumed that the error term was additive for their data and discussed statistical as- pects of the assumption. Further examination of the error term form would be interesting. Copies of the FORTRAN code used in this study are available from the author. Acknowledgments I thank James Bence for considerable statistical advice, particularly on the bootstrap procedure. James Bence, Alec MacCall, and Steve Ralston con- structively reviewed drafts of the note. I also thank David Woodbury for his assistance during an early stage of this study and Dale Roberts for his help with the use of SAS. Literature cited Cammen, L. M. 1980. Estimation of biological power functions from group measurements. Can. J. Fish. Aquat. Sci. 37:716-719. Damm, U. 1987. The estimation of weight at length from the total weight and the length distribution of a sample. ICES CM 1987/D:16, 9 p. DiCiccio, T. J., M. A. Martin, and G. A. Young. 1992. Fast and accurate approximate double boot- strap confidence intervals. Biometrika 79(2): 285-95. Dixon, W. J. 1983. BMDP statistical software. Univ. California Press, Berkeley, 733 p. Efron, B. 1987. Better bootstrap confidence intervals. J. Am. Statist. Assoc. 82(3971:171-185. 1988. Bootstrap confidence intervals: good or bad? Psychol. Bull. 104(2):293-6. Gunderson, D. R., and T. M. Sample. 1980. Distribution and abundance of rockfish off Washington, Oregon, and California during 1977. Mar. Fish. Rev. 42 (3-41:2-16. Lenarz, W. H. 1974. Length-weight relations for five eastern tropical Atlantic scombrids. Fish. Bull. 72:848- 851. Patterson, K. R. 1992. An improved method for studying the condi- tion offish, with an example using Pacific sardine Sardinops sagax (Jenyns). J. Fish Biol. 40: 821-831. Phillips, J. B. 1964. Life history studies on ten species of rockfish (genus Sebastodes). Calif. Dep. Fish Game Fish Bull. 126, 70 p. Pienarr, L. V., and J. A. Thomson. 1969. Allometric weight-length regression model. J. Fish. Res. Board Canada 26:123-131. Press, W. H., B. P. Flannery, S. A. Teukolsky and W. T. Vetterling. 1989. Numerical recipes the art of scientific com- puting (FORTRAN version). Cambridge Univ. Press, Cambridge, 702 p. Ratkowsky, D. A. 1983. Nonlinear regression modeling. Marcel Dekker, NY, 276 p. SAS Institute Inc. 1989. SAS/STAT® User's guide, version 6, 4th edi- tion, Vol. 2. SAS Institute Inc., Cary, NC, 846 p. Seber, G. A., and C. J. Wild. 1989. Nonlinear regression. J. Wiley & Sons, NY, 768 p. Sen, A. R. 1986. Methodological problems in sampling com- mercial rockfish landings. Fish. Bull. 84:409- 421. Sigler, M. F., and J. T. Fujioka. 1988. Evaluation of variability in sablefish, Anoplopoma fimbria, abundance indices in the Gulf of Alaska using the bootstrap method. Fish. Bull. 86:445-452. Tomlinson, P. K. 1971. Some sampling problems in fishery work. Biometrics 27:631-41. Spiny lobster recruitment and sea level: results of a 1 990 forecast Jeffrey J. Polovina Honolulu Laboratory, Southwest Fisheries Science Center National Marine Fisheries Service, NOAA 2570 Dole Street, Honolulu, Hawaii 96822-2396 Joint Institute for Marine and Atmospheric Research (JIMAR) University of Hawaii, Honolulu. Hawaii 96822 Department of Oceanography, School of Ocean and Earth Science and Technology University of Hawaii, Honolulu, Hawaii 96822 Gary T. Mitchum Joint Institute for Marine and Atmospheric Research LIIMAR) University of Hawaii, Honolulu, Hawaii 96822 Department of Oceanography, School of Ocean and Earth Science and Technology University of Hawaii, Honolulu. Hawaii 96822 A relation between recruitment to the fishery and sea level for the spiny lobster Panulirus mar- ginatus, in the Northwestern Ha- waiian Islands, was supported by data from 1985 through 1990 (Polovina and Mitchum, 1992). A forecast of future recruitment was made based on projected sea lev- els (Polovina and Mitchum, 1992). This note updates that forecast with two more years of data. Fishery data from 1985 to 1990 indicated considerable inter- annual variation in recruitment strength of spiny lobster, Pan- ulirus marginatus, between the two principal fishing grounds (Necker Island and Maro Reef), although separated by about 700 km (Fig. 1; Polovina and Mitchum, 1992). Recruitment strength variation between the two fishing areas was measured as the ratio of the commercial landings from Maro Reef divided by the combined commercial land- ings from Necker Island and Maro Reef. A strong correlation was ob- served between this measure of recruitment strength at Maro Reef and the sea level gradient along the Northwestern Hawaiian Islands, advanced by four years (Polovina and Mitchum, 1992). The sea level gradient was mea- sured as the difference in sea level between tide gauges at French Frigate Shoals, southeast of Maro Reef, and Midway Island, northwest of Maro Reef. A high proportion of the commercial landings came from Maro Reef following a steep gradient, while relatively few spiny lobsters were caught at Maro Reef following a flat gradient. The four-year lag is based on the minimum legal har- vest size which, for the spiny lob- ster is about three years old, af- ter benthic settlement. Prior to benthic settlement, the larvae are planktonic for about one year. Since sea level gradient appears to lead recruitment to the fishery by four years, the relation can provide up to a four-year forecast. Based on data through 1990, it was forecast that in 1991 recruit- ment to the fishery at Maro Reef would be weak but would recover in 1992 relative to recruitment at Necker Island (Fig. 2). The 1991 and 1992 fishery data show this forecast correct (Fig. 2), although the fishery for the entire North- western Hawaiian Islands was relatively weak in 1992. Thus, while sea level gradient index does forecast the relative strength of recruitment at Maro Reef, it is not, by itself, an index of absolute recruitment strength. It has been argued that sea level gradient measures the strength of the Subtropical Counter Current, which appears to intersect the Hawaiian ridge as three narrow eastward flowing bands at 20, 24, and 26 degrees north latitude (Polovina and Mitchum, 1992; White and Walker, 1985). Recent studies of P. marginatus larval distribution find a relatively high abundance of late stage larvae consistently present near lat. 26°N, and tracks from Argos drifter buoys drogued at 30 m indicate buoy entrap- ment along lat. 26'N.1 These re- sults provide some additional sup- port to our original hypothesis that a positive relationship exists between the strength of the Sub- tropical Counter Current and lo- cal larval survival, retention, and recruitment to the fishery at Maro Reef (Polovina and Mitchum, 1992). Literature cited Polovina, J. J., and G. T. Mitchum. 1992. Variability in spiny lob- ster Panulirus marginatus re- 1 Polovina, J.J.. and R.B. Moffitt. In re- view. The spatial and temporal distribu- tion of the larvae of the spiny lobster {Panulirus marginatus) in the North- western Hawaiian Islands. Manuscript accepted 11 August 1993 Fishery Bulletin 92:203-205 (1994) 203 204 Fishery Bulletin 92(1), 1994 Figure 1 The Hawaiian Archipelago. o 3 a 3 .s purchase from the Superintendent of Documents. U.S. Government Printing Office, Washington, DC 20402. 206 Fishery Bulletin Guide for Contributors Preparation Title page should include authors' full names and mailing addresses and the senior author's telephone and FAX number. Abstract should not exceed one double-spaced typed page. It should state the main scope of the research but emphasize its conclusions and relevant findings. Because abstracts are circulated by abstracting agencies, it is important that they represent the research clearly and concisely. Text must be typed double-spaced throughout. A brief introduction should portray the broad significance of the paper; the remainder of the paper should be divided into the fol- lowing sections: Materials and methods, Results, Discussion (or Conclusions), and Acknowledgments. Headings within each section must be short, reflect a logical sequence, and follow the rules of multiple subdivision (i.e., there can be no subdivision without at least two items). The en- tire text should be intelligible to inter- disciplinary readers; therefore, all acronyms, abbreviations, and techni- cal terms should be spelled out the first time they are mentioned. The scientific names of species must be written out the first time they are mentioned; subsequent mention of scientific names may be abbreviated. Follow the U.S. Government Printing Office Style Manual, 1984 ed., the CBE Style Manual, 5th ed., and the most current issue of the American Fisheries Society's Common and Scientific Names of Fishes from the United States and Canada for fish nomenclature. Dates should be writ- ten as follows: 11 November 1991. Measurements should be expressed in metric units, e.g., metric tons as (t); if other units of measurement are used, please make this fact explicit to the reader. The numeral one (1) should be typed as a one, not as a lower-case el (1). Text footnotes should be numbered with Arabic numerals and typed on a separate sheet of paper. Footnote all personal communications, un- published data, and unpublished manuscripts with full address of the communicator or author, or, as in the case of unpublished data, where the data are on file. Authors are advised to avoid references to nonstandard (gray) literature, such as internal, project, processed, or administrative reports, wherever possible. Where these references are used, please in- clude whether they are available from NTIS (National Technical Information Service) or from some other public depository. Literature cited comprises published works and those accepted for publica- tion in peer-reviewed literature (in press). Follow the name and year sys- tem for citation format. In the text, cite Smith and Jones (1977) or (Smith and Jones, 1977). If there is a se- quence of citations, list chronological- ly: Smith, 1932; Green, 1947; Smith and Jones, 1985. Abbreviations of serials should conform to abbrevia- tions given in Serial Sources for the Biosis Previews Database. AUTHORS ARE RESPONSIBLE FOR THE AC- CURACY AND COMPLETENESS OF ALL CITATIONS. Tables should not be excessive in size and must be cited in numerical order in the text. Headings should be short but ample enough to allow the table to be intelligible on its own. All unusual symbols must be explained in the table heading. Other incidental comments can be footnoted with italic numerals. Use asterisks for probability in statis- tical data. Because tables are typeset, they need only be submitted typed and formatted. Zeros should precede all decimal points for values less than one. Figures include line illustrations and photographs (or slides) and must be cited in numerical order in the text. Unless photographs are submitted on glossy paper with good contrast, we cannot guarantee a good final printed copy. Figures are to be labeled with author's name and number of figure. Use Times Roman font (upper and lowercase letters) to label within figures. Avoid vertical lettering except for y-axis labels. Zeros should precede all decimal points for values less than one. Figures may be submitted in computer software form along with a laser-printed copy, as photomechanical tranfers (PMT), or as high quality photographic prints. Send only xerox copies of figures to the Scientific Editor; original figures will be re- quested later when the manuscript has been accepted for publication. Copyright law does not cover govern- ment publications; they fall within the public domain. If an author repro- duces any part of a government publi- cation in his work, reference to source is considered correct form. Submission Send printed copies (original and two copies) to the Scientific Editor: Dr. Ronald Hardy, Scientific Editor Northwest Fisheries Science Center, F/NWC3 National Marine Fisheries Service, NOAA 2725 Montlake Boulevard East Seattle, WA 98112-2097 Once the manuscript has been ac- cepted for publication, you will be asked to submit a software copy of your manuscript to the Scientific Editor. The software copy should be submitted in Word Perfect text format (or in standard ASCII text format if Word Perfect is unavailable) and should be placed on a 5.25-inch or 3.5- inch disk that is double-sided, double or high density, and that is compatible with either DOS or Apple Macintosh systems. Copies of published articles and notes are available free of charge to the senior author (50 copies) and to his or her laboratory (50 copies). Addi- tional copies may be purchased in lots of 100 when the author receives page proofs. U.S. Department of Commerce Volume 92 Number 2 April 1994 Fishery Bulletin AP^ ^94 Woods Hole. M/> U.S. Department of Commerce Ronald H. Brown Secretary National Oceanic and Atmospheric Administration D. James Baker Under Secretary for Oceans and Atmosphere National Marine Fisheries Service Rolland A. Schmitten Assistant Administrator for Fisheries =/ l D ![ Scientific Editor Dr. Ronald W. Hardy Northwest Fisheries Science Center National Marine Fisheries Service, NOAA 2725 Montlake Boulevard East Seattle. Washington 981 12-2097 / ,^0Fc-04. \ •%rEs o« * Editorial Committee Dr. Andrew E. Dizon National Marine Fisheries Service Dr. Linda L. Jones National Marine Fisheries Service Dr. Richard D. Methot National Marine Fisheries Service Dr. Theodore W. Pietsch University of Washington Dr. Joseph E. Powers National Marine Fisheries Service Dr. Tim D. Smith National Marine Fisheries Service 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, Attn: Chief, Mail List Branch, Mail Stop SSOM, Washington, DC 20402-9373. 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: $27.00 domestic and $33.75 foreign. Cost per single issue: $13.00 domestic and $16.25 foreign. See back page for order form. Managing Editor Sharyn Matriotti National Marine Fisheries Service Scientific Publications Office 7600 Sand Point Way NE. BIN C 1 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 92 Number 2 April 1994 Fishery Bulletin Contents iii Errata 207 223 236 254 262 275 292 APR 2 8 1994 Woods Hole, MA 02543 Blood, Deborah M., Ann C. Matarese, and Mary M. Yoklavich Embryonic development of walleye pollock, Theragra chalcogramma, from Shelikof Strait, Gulf of Alaksa Brodeur, Richard D., and William C. Rugen Diel vertical distribution of ichthyoplankton in the northern Gulf of Alaska Clark, Malcolm R., and Dianne M. Tracey Changes in a population of orange roughy, Hoplostethus atlanticus, with commercial exploitation on the Challenger Plateau, New Zealand Daniel, Louis B., Ill, and John E. Graves Morphometry and genetic identification of eggs of spring-spawning sciaenids in lower Chesapeake Bay Ditty, James G., Richard F. Shaw, and Joseph S. Cope A re-description of Atlantic spadefish larvae, Chaetodipterus faber (family: Ephippidae), and their distribution, abundance, and seasonal occurrence in the northern Gulf of Mexico Ditty, James G., Richard F. Shaw, Churchill B. Grimes, and Joseph S. Cope Larval development, distribution, and abundance of common dolphin, Coryphaena hippurus, and pompano dolphin, C. equiselis (family: Coryphaenidae), in the northern Gulf of Mexico Kastelle, Craig R., Daniel K. Kimura, Ahmad E. Nevissi, and Donald R. Gunderson Using Pb-2 1 0/Ra-226 diseguilibria for sablefish, Anoplopoma fimbria, age validation Fishery Bulletin 92(2). 1994 302 Moltschaniwskyj, Natalie A., and Peter J. Doherty Distribution and abundance of two juvenile tropical Photololigo species (Cephalopoda: Loliginidae) in the central Great Barrier Reef Lagoon 313 Murie, Debra J., Daryl C. Parkyn, Bruce G. Clapp, and Geoffrey G. Krause Observations on the distribution and activities of rockfish, Sebastes spp., in Saanich Inlet, British Columbia, from the Pisces IV submersible 324 Perrin, William F, Gary D. Schnell, Daniel J. Hough, James W. Gilpatrick Jr., and Jerry V. Kashiwada Reexamination of geographic variation in cranial morphology of the pantropical spotted dolphin, Stenella attenuata, in the eastern Pacific 347 Powell, Eric N., John M. Klinck, Eileen E. Hofmann, and Sammy M. Ray Modeling oyster populations. IV: Rates of mortality, population crashes, and management 374 Prager, Michael H. A suite of extensions to a nonequilibrium surplus-production model 390 Stoner, Allan W., and Megan Davis Experimental outplanting of juvenile queen conch, Strombus gigas. comparison of wild and hatchery-reared stocks 412 Taylor, David M., Paul G. O'Keefe, and Charles Fitzpatrick A snow crab, Chionoecetes opilio (Decapoda, Majidae), fishery collapse in Newfoundland 420 Warlen, Stanley M. Spawning time and recruitment dynamics of larval Atlantic menhaden, Brevoortia tyrannus, into a North Carolina estuary 434 Reilly, Stephen B., and Paul C. Fiedler Interannual variability of dolphin habitats in the eastern uupical Pacific. I: Research vessel surveys, 1986-1990 451 Fiedler, Paul C, and Stephen B. Reilly Interannual variability of dolphin habitats in the eastern tropical Pacific. II: Effects on abundances estimated from tuna vessel sightings, 1 975-1 990 Notes 464 Aurioles-Gamboa, David, Maria Isabel Castro-Gonzalez, and Ricardo Perez-Flores Annual mass strandings of pelagic red crabs, Pleuroncodes planipes (Crustacea: Anomura Galatheidae), in Bahia Magdalena, Baja California Sur, Mexico 471 Ennevor, Bridget C. Mass marking coho salmon, Oncorhynchus kisutch, fry with lanthanum and cerium 474 Hazin, Fabio H. V, Clara E. Boeckman, Elizabeth C. Leal, Rosangela R T. Lessa, Kohei Kihara, and Kazuyuki Otsuka Distribution and relative abundance of the blue shark, Prionace glauca. in the southwestern equatorial Atlantic Ocean Errata (i) Bigelow, Keith A. Age and growth of the oceanic squid Onychoteuthis borealijaponica in the North Pacific Fish. Bull. 92(l):13-25 Figure 5 should read as shown below. — 300 250 200 150 100 50 400 350 300 250 200 150 100 50 0 Males • • - 1 1 1 1 i i 1 1 1 0 50 100 150 200 250 300 350 400 450 Males 400 - 300 - °^D# 200 - ooJr 100 - 0 - A n. ~+t-^ — , — , — , — , 0 50 100 150 200 250 300 350 400 450 Females • V? 1 1 I 1 1 1 1 1 1 X £ S: 1000 -, Females o H .-, » _ <*u 0 50 100 150 200 250 300 350 400 450 AGE (days) 0 50 100 150 200 250 300 350 400 450 AGE (days) Figure 5 Relation between age (determined by number of increments within statoliths) and mantle length (mm) and weight (g) for male and female Onyclwteuthis borealijaponica. Western North Pacific 1990 (open circles), central North Pacific 1990 (closed circles), central North Pacific 1991 (closed triangles = juveniles-subadults, open triangles = unknown sex), and eastern North Pacific 1990 (open squares). (2) Perryman, Wayne L., and Morgan S. Lynn Examination of stock and school structure of striped dolphin {Stenella coeruleoalba) in the eastern Pacific from aerial photogrammetry Fish. Bull. 92(1):122-131 Figures 3, 4, and 7, and Tables 1, 2, and 3 show an incorrectly typeset species name for striped dol- phin. The correct name should read striped dol- phin, Stenella coeruleoalba. The National Marine Fisheries Service (NMFS) does not approve, recommend, or endorse any proprietary product or proprietary material mentioned in this publication. No references shall be made to NMFS, or to this publication fur- nished by NMFS, in any advertising or sales promotion which would indicate or imply that NMFS approves, recommends or endorses any proprietary pro- duct or proprietary material mentioned herein, or which has as its purpose an intent to cause directly or indirectly the advertised product to be used or pur- chased because of this NMFS publication. Abstract. Eggs of walleye pollock, Theragra chalcogramma, from Shelikof Strait, Alaska, were reared at three temperatures: 3.8°, 5.7°, and 7.7°C. Development was divided into 21 stages. A piece- wise regression model with mid- points of each stage describes the relation between time to each stage of development and tempera- ture. Preserved eggs of each stage are described, illustrated, and pho- tographed. Midpoint of hatch was 393 hours at 3.8°C, 303 hours at 5.7°C, and 234 hours at 7.7°C. Mean length of larvae at hatch in- creased linearly with temperature. We compared rate of develop- ment, time to 50% hatch, and mor- phological development with other studies of walleye pollock eggs. Rate of development and time to 50% hatch were similar among populations of eastern North Pa- cific walleye pollock. Western North Pacific walleye pollock re- quired longer incubation times than eastern North Pacific walleye pollock. Morphological develop- ment of Shelikof Strait eggs differs from development of western North Pacific walleye pollock eggs: optic vesicles, myomeres, eye lenses, heart, and otic capsules appear earlier than in Shelikof Strait eggs, and eye pigment ap- pears later. The differences in de- velopment may be exacerbated by the condition of the eggs in which they were examined (e.g. pre- served vs. live). Developmental differences between stocks are dis- cussed with the conclusion that model components for egg mortal- ity and spawning biomass must be based on specimens collected in the area of interest. Embryonic development of walleye pollock, Theragra chalcogramma, from Shelikof Strait, Gulf of Alaska* Deborah M. Blood Ann C. Matarese Mary M. Yoklavich** Alaska Fisheries Science Center, National Marine Fisheries Service. NOAA 7600 Sand Point Way N.E., Seattle, WA 981 15-0070 Walleye pollock, Theragra chalco- gramma, is the most abundant member of the family Gadidae in the subarctic Pacific Ocean and Bering Sea, supporting the largest single-species commercial fishery in the world (Megrey, 1991). In the Gulf of Alaska, Shelikof Strait is the principal spawning area (Kendall and Picquelle, 1990) and has been the site of intensive re- search to understand processes leading to recruitment variability of walleye pollock (Schumacher and Kendall, 1991). Age determination of fertilized eggs is a basis for investigating biotic and abiotic impacts on the earliest life-history stage and thus for understanding interannual variability in walleye pollock re- cruitment. Age of walleye pollock eggs has been crucial to several studies. Egg mortality and spawn- ing biomass are estimated by mod- eling age-specific egg abundance over time (Picquelle and Megrey, 1993; Bates1). Patterns in horizon- tal or vertical distribution and abundance of walleye pollock eggs in the western Gulf of Alaska have been described by grouping devel- opmental stages into broad age groups (Kendall and Kim, 1989; Kendall and Picquelle, 1990). Egg age is an independent vari- able in the models used to estimate egg production and mortality. Therefore, increasing the accuracy in measuring egg ages should im- prove estimates of these values. In past studies, walleye pollock eggs have been incubated in the labora- tory to develop temperature-spe- cific equations that estimate dura- tion of development or age of the eggs, to describe morphological de- velopment, to observe effects of light on egg buoyancy and hatching rate, and to obtain larvae for ex- periments (Table 1). Although these incubation studies provide pertinent data on ontogeny of wall- eye pollock, none can be used with accuracy to determine age of eggs Bates, R. D. 1987. Estimation of egg pro- duction, spawner biomass, and egg mor- tality for walleye pollock, Theragra chalcogramma, in Shelikof Strait from ichthyoplankton surveys during 1981. U.S. Dep. Commer., NOAA, Nat. Mar. Fish. Serv., Northwest Alaska Fish. Cent., 7600 Sand Point Way N.E., Bin C15700, Bldg. 4, Seattle, WA 98115-0070. Proc. Rep. 87-20, 192 p. Manuscript accepted 3 November 1993 Fishery Bulletin 92: 207-222 (1994) * Contribution 0148 of the Fisheries Oceanography Coordinated Investigations, NOAA, Seattle. ** Present address: Southwest Fisheries Science Center, Pacific Fisheries Environmen- tal group. National Marine Fisheries Service, NOAA, P.O. Box 831, Monterey, CA 93942 207 208 Fishery Bulletin 92(2). 1994 Table 1 Summary of Theragra chalcogramma egg incubation stuc ies. Reference Temperature (°C) Source and region Stages' Regression equation Morphological description Illustrations Photographs Gorbunova, 1954 3.4, 8.2 (means) western Pacific Ocean N ii n i • No Yes Yes No Yusa, 1954 6.0-7.0 Ishikan Bay, Japan 272 No Yes Yes Yes Hamai et al., 1971 2.4-2.5, 6.5-6.7, 9.9-10.1, ( means) Funka Bay, Japan 4 Stage-specific equation to predict age (d) at any stage No No No Hamai et al., 1974 5.0 (mean) Funka Bay, Japan 6 No No No No Matarese 1983 unpubl.'' 5.0 N. Gulf of Alaska 21 Stage-specific equation to predict age (h) at any stage'' Nn No No Haynes and Ignell, 1983 2.0, 5.0, 6.0, 8.0, 11.0 Stephens Passage, SE Alaska 7 General equation to predict age (h) at any stage No No No Nakatani and Maeda. 1984 -1.0, 0.0, 2.0, 4.0, 7.0, 10.0, 13.0 Funka Bay, Japan 5 To 507r hatch No No No Paul, 1984, unpubl.1 5.0 N. Gulf of Alaska 21 No No No No Bailey and Stehr, 1986 5.6, 8.5 Puget Sound, Washington None No No No No Olla and Davis, 1993 6.0 Shelikof Strait, Alaska Nunc Nil No No No ; Prior to hatch. 2 Reported as intervals of time. 3 A. C. Matarese, Alaska Fisheries Science Center, National Marine Fisheries Service, 7600 Sand Poin 4 In Bates 1987 (Footnote 1). 1 A J. Paul, University of Alaska Fairbanks, Institute of Marine Science, Seward Marine Center Lab, t Way P.O. B NE., Seattle ox 730, Sewa WA 98115. rd, AK 99664. from Shelikof Strait. Eggs need to be obtained from the study area and incubated at a range of tempera- tures occurring in the area. Categorizing the con- tinuous process of egg development into a large number of stages should increase the precision of the egg-age estimate The first objective of our study was to incubate Shelikof Strait walleye pollock eggs at the mean water temperature for Shelikof Strait, bracketing that temperature to include upper and lower ex- tremes. Egg development times were used to pro- duce a stage duration table and a regression model to estimate egg age based on water temperature and developmental stage. Morphological development is described for 21 developmental stages. These de- scriptions are accompanied by illustrations and photographs to facilitate identification of body struc- tures and stage hallmarks. The second objective was to compare our rates of egg development to other walleye pollock incubation studies. Morphological development is included in this comparison. Blood et al.. Embryonic development of Theragra chalcogramma 209 Methods Incubation Adult walleye pollock were collected with a rope trawl off Cape Kekurnoi (57°42.5'N, 155°16.2'W) in Shelikof Strait, Alaska, on 7 April 1989 from the NOAA research vessel Miller Freeman. Eggs from one female and milt from three or four males were hand stripped into glass petri dishes, gently mixed, and left undisturbed for one minute. Eggs were then rinsed, transferred to 3°C (surface water tempera- ture) seawater in glass jars (3.8 L), and held two hours. Floating eggs with a perivitelline space were assumed to be fertilized (Blaxter, 1969; Alderdice, 1988). Viable eggs were poured into eighteen 0.5-L jars filled with 3°C seawater. Eggs were not counted but apportioned similarly among the jars at a con- centration of about one egg/mL. Six capped jars were held in each of three water bath incubators onboard the Miller Freeman. Initial incubation temperatures were set to include the range of temperatures in the area. Mean water temperature at depths of 150-200 m in Shelikof Strait, where most eggs are found (March-May) (Kendall and Kim, 1989), is 5°C; ex- tremes of 3.6° and 5.9°C have been reported (Reed and Schumacher, 1989). Incubators were sealed to minimize light and movement and placed in sepa- rate refrigerators adjusted to 3°, 5°, and 7 C. One- half of the water in the jars was replaced every day with seawater of the same temperature. Eggs were preserved in phosphate buffered formalin (5%)2 or Stockard's solution3 (Velsen, 1980). Stockard's solu- tion cleared the chorion and darkened embryonic tissue, easing examination of embryonic develop- ment. Phosphate buffered formalin did not darken embryonic tissue as much as Stockard's solution, yielding better definition of some structures like somites and otic capsules. Live, newly hatched lar- vae were measured (standard length in millimeters) and preserved (5% buffered formalin). Detailed exami- nation and morphological description of embryos were completed after eggs were returned to the laboratory. During the first 24 hours after fertilization, eggs were sampled at 2-3 hour intervals. After 24 hours, intervals were increased to about 6 hours. When an interval was less or greater than 6 hours, the sub- sequent sampling time was adjusted to return to the original 6-hour schedule. Data were not recorded for three sampling times late in development because intervals were inadvertently extended to 12 hours (236, 258, and 282 hours). At each interval, 10 to 50 eggs were sampled from one jar per incubator; only one jar was sampled to ensure there would be enough eggs and larvae left to sample near the end of the incubation period. Jars were sampled in rotation throughout the duration of the experiment until no eggs remained. When eggs began to hatch, all jars were checked and newly hatched larvae were removed in addition to eggs scheduled to be sampled. Dead eggs were removed from the designated sample jar at each interval. Water bath temperatures were recorded for every sampling interval. Frequent opening of refrigerators during the initial short sampling intervals increased temperatures in the refrigerators despite thermostat adjustments. Water bath temperatures stabilized after 48 hours to 3.8°, 5.7°, and 7.7°C. Morphological descriptions Eggs were examined with the aid of a dissecting microscope (6-50x magnification) and described ac- cording to a 21-stage scheme adapted from Naplin and Obenchain (1980) (Table 2). Morphological terms follow Trinkaus ( 1951) with one exception: the term "blastodisc," in this paper, includes the germi- nal area from the time of cytoplasm polarization until embryonic shield formation (Markle and Table 2 Stages of embryonic development of Theragra chalcogramma (adapted from Naplin and Obenchain, 1980). Stage Developmental stage 2 50 mL 37% formaldehyde, 4.0 g sodium phosphate monobasic, 6.5 g sodium phosphate dibasic, made up to 1 L with distilled water. 3 50 mL 371 formaldehyde, 40 mL glacial acetic acid, 60 mL glycerin, and 850 mL distilled water. 1 Precell 2 2 cell 3 4 cell t 8 cell 5 16 cell 6 32+ cell 7 Blastodermal cap 8 Early germ ring 9 Germ ring 1/4 down yolk 10 Germ ring 1/2 down yolk 11 Germ ring 3/4 down yolk 12 Late germ ring L3 Early middle (blastopore closure) 14 Middle middle (appearance of pigment 15 Late middle (tail bud thickens) \h Early late (tail bud lifts from yolk) 17 Tail 5/8 around yolk 18 Tail 3/4 around yolk 19 Tail 7/8 around yolk 20 Full circle around yolk 21 Tail 1-1/8 around yolk 210 Fishery Bulletin 92(2). 1994 Waiwood, 1985, in part). Eggs preserved in Stockard's solution were photographed with a Nikon F2 camera fitted with a PB6 200-mm bellows exten- sion and a 24-mm 1:2.8 reverse-mounted lens. This configuration produced a 47x magnification. Re- flected light was supplied by two synchronized flash units. Other photographs (stages 5 and 6) were taken with a single-lens reflex adapter (0.32x) on a Wild M-8 dissecting microscope with transmitted light. At 50x, the phototube and adapter increased magnification to 66x. Analysis Endpoint, midpoint, and duration of stage (in hours) were estimated for eggs incubated at each tempera- ture. For stages 1-20, stage endpoint was deter- mined by the presence of two stages during a sam- pling time; if stages n and n + \ were present, the time at which the eggs were sampled was consid- ered a transition and therefore the endpoint for stage n. If there was no transition, the endpoint for stage n was the midpoint between the last sampling time during which stage-rc eggs were present and the first time stage-« + l eggs were observed. Dura- tion and midpoint of stage n were determined as Duration Stage n = Endpoint Stage (n) - Endpoint Stage in - 1); Midpoint Stage n = Endpoint Stage (n Duration Stage n li Endpoint of stage 21 was the sampling interval when the last embryo had hatched. With the mid- points and time of 50% hatch, a piece-wise least- squares linear regression model (SAS, 1985) was derived to estimate age (hours) of eggs at a specific stage incubated at any temperature within the lim- its of this experiment. Differences in mean lengths of larvae hatched from the three temperature groups were analyzed by a Student-Newman-Keuls test. Lengths of larvae hatching at stages 20 and 21 were analyzed by a two-way analysis of variance (ANOVA) by using stage and temperature. We chose five representative developmental stages and compared time to midpoint of each stage among incubation studies. Comparison with Hamai et al. (1971, 1974) was possible for only three stages. We grouped data on time to 509f hatch into western and eastern North Pacific studies and performed a log- transformed analysis of covariance to test for differ- ences in time to 50% hatch between these two ar- eas with incubation temperature as the covariate. Results Incubation rates Temperatures of the three water baths increased at the beginning of sampling (Fig. 1). Temperature spikes that occurred after 288 hours in the 5.7°C jars and after 396 hours in the 3.8°C jars, were associ- ated with the appearance of large numbers of lar- vae; water baths may have warmed when refrigera- tors were opened frequently to measure larvae. Eggs developed at similar rates among incubation temperatures for the first 36 hours through stage 6 (Fig. 2). After stage 6, at about 36 hours, when tem- peratures had stabilized, development rates began to diverge. Duration of stages 7-21 was variable (Table 3). Usually the duration of a stage was longer at cooler temperatures. However, this was not al- ways the case, and stages 12 and 20 required simi- lar amounts of time regardless of temperature. At all temperatures, hatching began during stage 20; the percentage of eggs hatched by the beginning of stage 21 was 35% at 3.8°, 40% at 5.7°, and 8.1% at 1.1'C Four larvae from the 7.7°C group hatched after 192 hours; another 18 hours elapsed before other larvae hatched at this temperature. These early larvae were not included in this analysis be- cause we assumed that the hiatus in hatching times indicated that early hatching was anomalous, i.e. hatching may have been mechanically induced. Af- ter hatching began, time required for 50% hatch decreased as temperature increased: 48, 36, and 24 hours at 3.8°, 5.7 , and 7.7°C. The elapsed time be- tween hatching of the first and last larvae was 72 hours at 3.8°C and 60 hours at both 5.7 and 7.7°C. Eggs developed normally at 5.7° and 7.7°C; how- ever, curvature of the spine was observed in some late-stage embryos incubated at 3.8C These abnor- mal eggs hatched, but most larvae were not mea- sured because of curvature. Mean length at hatch of all larvae increased with incubation temperature: 4.15 (SD 0.380, n = 100), 4.29 (SD 0.272, rc=192), and 4.55 mm (SD 0.303, rc=84> at 3.8°, 5.7°, and 7.7'C (Fig. 3). Mean lengths of larvae from the three tem- perature groups were significantly different (P<0.05). In addition, larval lengths increased as the hatching period progressed at all temperatures. Length of larvae hatching at stages 20 and 21 was significantly different at all temperatures (P<0.01); larvae that hatched at stage 21 were 9-13% longer than larvae that hatched at stage 20. Blood et al.: Embryonic development of Theragra chalcogramma 21 1 O / J& * **.^*^w**** **•**.*** * ********* *.. ; 3.8 C -+- 5.7 C * 7.7 C I 1 1 1 1 I : 1 1 1 1 1 1 1 1 1 1 0 24 48 72 96 120 144 168 192 216 240 264 288 312 336 360 384 408 Incubation time (h) Figure 1 Temperatures recorded in water baths during incubation of Theragra chalcogramma eggs. o> CO *-■ CO c CD E Q. o CD > CD Q 0 24 48 72 96 120 144 168 192 216 240 264 288 312 336 360 384 408 Incubation time (h) Figure 2 Development of Theragra chalcogramma eggs incubated at 3.8°, 5.7°, and 7.7°C. Points represent occurrence of stages at scheduled sampling times. The piece-wise regression model (SAS, 1985) has two separate components and is discontinuous be- tween stages 6 and 7 (Fig. 4). This type of model was necessary because of the rapid divergence of devel- opmental rates at all temperatures after stage 6; it was not possible to fit one equation to the entire incubation time. The two components are described by the following equations: component 1: stages 1-6 Age = 3.27 - 0.13 (stage) (temperature) + 0.47 (stage2); component 2: stages 7-21 Age = 17.82 + 7.05(stage) - 0.656 (stage) (temperature) + 0.043(stage3) - 0.0032 (temperature) (stage3), 212 Fishery Bulletin 92(2). 1994 Table 3 Endpoint, midpoint, and duration in hours (h) of stage of deve opment of Theragra chalcogramma eggs incu- bated at 3.8°, 5.7°, and 7.7°C 3.8'C 5.7"C 7.7°C Stage Endpoint (h) Midpoint (h) Duration (h) Endpoint (h) Midpoint (h) Duration (h) Endpoint (h) Midpoint (h) Duration (h) 1 4.00 2.00 4.00 4.00 2.00 4.00 3.50 1.75 3.50 2 6.00 5.00 2.00 6.00 5.00 2.00 4.00 3.75 0.50 3 8.00 7.00 2.00 7.00 6.50 1.00 5.00 4.50 1.00 4 10.25 9.12 2.25 9.00 8.00 2.00 7.00 6.00 2.00 5 12.50 11.37 2.25 10.25 9.62 1.25 10.25 8.62 3.25 6 22.50 17.50 10.00 22.50 16.37 12.25 19.50 14.87 9.25 7 64.00 43.25 41.50 51.00 36.75 28.50 40.00 29.75 20.50 8 78.00 71.00 14.00 68.00 59.50 17.00 48.00 44.00 8.00 9 90.00 84.00 12.00 75.00 71.50 7.00 54.00 51.00 6.00 10 105.00 97.50 15.00 87.00 81.00 12.00 57.00 55.50 3.00 11 120.00 112.50 15.00 93.00 90.00 6.00 68.00 62.50 11.00 12 138.00 129.00 18.00 108.00 100.50 15.00 84.00 76.00 16.00 13 153.00 145.50 15.00 114.00 111.00 6.00 87.00 85.50 3.00 14 180.00 166.50 27.00 135.00 124.50 21.00 102.00 94.50 15.00 15 195.00 187.50 15.00 164.00 149.50 29.00 111.00 106.50 9.00 16 219.00 207.00 24.00 174.00 169.00 10.00 117.00 114.00 6.00 17 252.00 235.50 33.00 189.00 181.50 15.00 132.00 124.50 15.00 18 312.00 282.00 60.00 219.00 204.00 30.00 144.00 138.00 12.00 19 336.00 324.00 24.00 258.00 238.50 39.00 180.00 162.00 36.00 20 378.00 357.00 42.00 300.00 279.00 42.00 219.00 199.50 39.00 21 414.00 393.00' 36.00 330.00 303.00' 30.00 270.00 234.00' 51.00 ' 50% hatch. where age of the egg is expressed in hours. The value of R2 is 0.96 for component 1 and 0.99 for component 2. We compared our rates of egg development to other walleye pollock incubation studies in the 5- 7°C range (Table 4). There was a significant differ- ence between regression equations of incubation time to 50% hatch and temperature for western versus eastern North Pacific studies (P<0.01), but the slopes were not different (P=0.18). Based on the 95% confidence interval about the parameter esti- mates, time to 50% hatch of western North Pacific walleye pollock tended to be 1.2 to 1.3 times longer on average than that of the eastern North Pacific fish at a specific temperature. Morphological descriptions Walleye pollock eggs are pelagic and have a smooth, clear chorion and homogeneous yolk. No oil globules are present. Preserved eggs range from 1.2 to 1.8 mm in diameter, although most are 1.35-1.45 mm (Matarese et al., 1989). Appearance of the egg var- ies with type of preservative. There was little or no shrinkage of yolk material in Stockard's solution, whereas yolk of formalin-preserved eggs decreased in volume and the yolk membrane frequently col- lapsed. This effect of formalin preservation was helpful in determining how much of the tail had lifted away from the yolk in late-stage embryos. Development of walleye pollock eggs and embryos, from fertilization to just before hatching, was di- vided into the following 21 stages (Table 2): Precell (stage 1) Cytoplasm at the animal pole forms a blastodisc; bands of cytoplasm extend from below the equator to the blastodisc (Fig. 5A), which is without distinct margins (Fig. 6A). When intact, the yolk membrane almost touches the inner wall of the chorion. The perivitelline space is most vis- ible over the blastodisc. 2 cells (stage 2) The first cell division of the blastodisc is in the horizontal plane. Cell material may not be equally divided (Figs. 5B and 6B). 4 cells (stage 3) The second cleavage is perpen- dicular to, and in the same plane as, the first. Cells are roughly equal in size and form a square (Figs. 5C and 6C). 8 cells (stage 4) The third cleavage is perpen- dicular to the second cleavage (parallel to first cleav- age). Each cell divides in half in the horizontal Blood et al.: Embryonic development of Theragra chalcogramma 213 E E O) c 9 C CO 9 s 5.5 5.0 4.5 4.0 3.5 "T O 7./C D 6.7°C A 3.8° C 3.0 204 248 292 336 380 424 Total incubation time (h) Figure 3 Mean hatch lengths at each sampling interval during the hatching period of Theragra chalcogramma larvae incubated at 3.8°, 5.7°, and 7.7°C. Stage of de- velopment at hatch is also shown. Shaded circle indicates overall mean length for each temperature. Vertical bars are standard deviations; numbers indicate sample size. plane. Cells form a rectangle with the four cells in the center smaller than those at the corners of the rectangle (Figs. 5D and 6D). 16 cells (stage 5) The fourth cleavage is perpen- dicular to the third; this is the last stage in which cell division is restricted to the horizontal plane. Most eggs have a square or rectangular block of cells with four cells on each side; all cells are in contact with yolk through this stage (Figs. 5E and 6E). 32 cells (stage 6) Initially, the single layer of cells has a flat, irregular square or rectangular shape. Cell division continues in horizontal and vertical planes, transforming the blastodisc into a hollow cap of cells on the yolk resembling a rasp- berry (Figs. 5F and 6F). Cells increase in number but the size of the blastodisc remains constant. The perivitelline space widens between yolk and chorion. Blastodermal cap (stage 7) The blastodisc progresses through two steps: at first, cell size de- creases from continued cleavage; cell material ap- pears granular and the blastodisc resembles a flat- tened dome on the yolk surface. Then, the base of the cell mass sinks below the yolk surface; the periblast extends beyond the equator of the blasto- disc, giving the appearance of a "flying saucer" in lateral view (Figs. 5G and 6G). Early germ ring (stage 8) The center of the blasto- disc flattens and the periph- ery (germ ring) thickens in preparation to overgrow the yolk (epiboly). The blasto- coel, visible on one side of the blastodisc, appears grainy and pale (Fig. 5H). The margin between blasto- coel and blastodisc is indis- tinct (Fig. 6H). Germ ring 1/4 around yolk (stage 9) The blasto- disc, now the embryonic shield, expands as the germ ring begins to overgrow the yolk. The margin of the fu- ture anterior end of the em- bryo is slightly curved and sharply defined. Cell mate- rial covering the blastocoel appears less grainy than in the previous stage. After preservation, this thin cellu- lar layer appears concave in lateral view. The germ ring margin is thin and flattened, extending 1/4 around yolk (Figs. 51 and 61). Germ ring 1/2 around yolk (stage 10) The germ ring envelopes half the yolk and the anterior margin of the embryonic shield is sharply curved and thick (Figs. 5J and 6J). The beginning of neu- ral development is visible; a neural keel extends from the anterior margin of the embryonic shield to 2/3 its length (Fig. 5K). Germ ring 3/4 around yolk (stage 11) Head and upper body region begin to differentiate but no distinct brain lobes are apparent. Optic vesicles develop. Prospective head and body mesoderm out- lines the hour-glass shape of the developing embryo (Fig. 7A). The notochord is visible ventrally. The germ ring has progressed 3/4 down the yolk (Fig. 6K). Late germ ring (stage 12) Myomere differen- tiation begins; separate myomeres are not visible. The midbody expands dorsoventrally; prospective head and body mesoderm forms a narrow outline of the embryo. The blastopore is open and the germ ring envelopes more than 7/8 of the yolk (Figs. 7B and 8A). Early middle stage (stage 13) The blastopore is closed. The notochord and 7-12 incomplete myomeres are visible. Tail margin is indistinct and 214 Fishery Bulletin 92(2). 1994 a> c o a. ■o E o 0) E i- flat; the medial portion of the tail bud is thicker (Figs. 7C and 8B). The body of the embryo appears flattened. Although not distinguish- able in preserved specimens, Kupffer's vesicle is visible in the live egg. Middle middle stage (stage 14) Embryos have 14-16 myomeres. Differen- tiation begins in eyes and mid- and hindbrain. Fore- brain very small and under- developed. The tail bud mar- gin is defined but still flat- tened (Fig. 8C). The entire length of the body is thicker. Small melanophores are scattered along the dorsum between the hindbrain and 4/5 of body length (Fig. 7D). Late middle stage (stage 15) About 20-25 myomeres are visible. Eye lenses are formed. The liver appears as a slight bulge in the body wall, and the gut area is de- lineated. The tail bud is thick and appears lifted from the yolk surface with the margin attached (Fig. 7E). Pigment is darker than in the previous stage and dendritic, extending from midbrain to tip of tail bud and confined mostly to the dorsum. Nares, mid- and hindbrain, and pectoral bud anlagen are visible dorsally (Fig. 8D). 450 400 350 300 250 200 150 100 50 0 n — i — i — i — i — i — i — i — i — i — i r n — i — i — i — i — i — i — r -B-- 3.a°c -e- ■ 9.7 C ▲ 7.7 °C I I I I I I I I L J I L 1 2 3 S 0 7 8 0 10 11 12 13 14 18 Ifl 17 18 10 20 BOM Developmental stage Figure 4 Time (h) to midpoint of stage of Theragra chalcogramma eggs incubated at 3.8°, 5.7°, and 7.7°C. Fitted lines are results of regression model; symbols are observed values. Early late stage (stage 16) Heart tissue begins to expand when the embryo has about 24-36 myo- meres. Forebrain differentiates from midbrain. The tail bud lifts from the yolk surface (Figs. 7F and 8E) and pigment forms two parallel rows dorsoposteriorly. Tail 5/8 around yolk (stage 17) The embryo has 27-36 myomeres. More of the tail lifts from the Table 4 Comparison of time (h) to estimated midpoint of five developmental milestones of Theragra chalcogramma embryos incubated at 5-7°C. Eastern North Pacific incubation stuc ies Western North Pacific incubation studies Matarese, Paul, 1984, Haynes and This Hamai et al., Yusa, Nakatani and Hamai et al., 1983, unpubl. unpubl. - Ignell, 1983' study 1971* 1954 Maeda, 19845 1974 Stage 15.0-C) 15.0 C) 16.0 C) (5.7'C) (6.5-6.7'C) (6.0-7.0 Cl (7.0"C) (5.0 Cl Blastodermal cap 37.5 39.5 35 36.8 28.5 31 Blastopore closure 114 UK 105 108 100 102 Its 139 Tail 3/4 211.7 217 li,l 204 234 192 Tail full circle 274.5 264 250 279 250 270 216 330 Vi<, hatch 349 320 285 303 345 288+ 298 411 A C Matarese. Alaska Fisheries Science Center, National Marine Fisheries Service, 7600 Sand Point Way N.E., Seattle, WA 98115. 2 A .1 Paul, University of Alaska Fairbanks. Institue of Marine Science, Seward Marine Center Lab, P.O. Box 730, Seward, AK 99664. Values except for hatch estimated from Table 3 in Haynes and Ignell (1983). 50% hatch from Table 7. 1 Values except fur hatch estimated from Fig. 3 in Hamai et al. (1971). ' Values except for hatch estimated from Fig 5 in Nakatani and Maeda (1984). Blood et al.: Embryonic development of Theragra chalcogramma 215 cytoplasm - blastodtsc perivitelline space E periblast germ ring blastocoel neural keel Figure 5 Illustrations of preserved Theragra chalcogramma eggs. (Al Stage 1 (precell); (B) Stage 2 (2 cell); (C) Stage 3 (4 cell); (D) Stage 4 (8 cell); (E) Stage 5 (16 cell); (F) Stage 6 (32 cell); (G) Stage 7 (blastodermal cap); (H) Stage 8 (early germ ring); (I) Stage 9 (germ ring 1/4, lateral view); (J) Stage 10 (germ ring 1/2, lateral view); (K) Stage 10 (dorsal view). 216 Fishery Bulletin 92(2), 1994 E II Figure 6 Photographs of preserved Theragra chalcogramma eggs. (A) Stage 1 (precell); (Bi Stage 2 (2 cell); (C) Stage 3 (4 cell); (D) Stage 4 (8 cell); (E) Stage 5 (16 cell); (F) Stage 6 (32 cell); (G) Stage 7 (blastodermal cap); (H) Stage 8 (early germ ring); (I) Stage 9 (germ ring 1/4); (J) Stage 10 (germ ring 1/2); (K) Stage 11 (germ ring 3/4). Blood et al.: Embryonic development of Theragra chalcogramma 217 optrc vesicle prospective head and body mesoderm optic vesicle myomeres blastopore lens E otic capsule H hatching gland: hatching glands J K L Figure 7 Illustrations of preserved Theragra chalcogramma eggs. (A) Stage 11 (germ ring 3/4); (B) Stage 12 (blastopore almost closed); (C) Stage 13 (early middle); (D) Stage 14 (middle middle); (E) Stage 15 (late middle); (F) Stage 16 (early late); (G) Stage 17 (tail 5/8 circle); (H) Stage 18 (tail 3/4 circle); (I) Stage 19 (tail 7/8 circle); (J) Stage 20 (tail full circle, lateral view); (K) Stage 20 (dorsal view); (L) Stage 21 (tail 1-1/8 circle). 218 Fishery Bulletin 92(2). 1994 Figure 8 Photographs of preserved Thcragra chaleogramma eggs. (A) Stage 12 (blastopore almost closed); (B) Stage 13 (early middle); (C) Stage 14 (middle middle); (D) Stage 15 (late middle); (E) Stage 16 (early late); (F) Stage 17 (tail 5/8 circle); (G) Stage 18 (tail 3/4 circle); (H) Stage 19 (tail 7/8 circle); (I) Stage 20 (tail full circle); (J) Stage 21 (tail 1-1/8 circle). Blood et al.: Embryonic development of Theragra chalcogramma 219 3.5 mm SL Figure 9 Illustration of preserved Theragra chalcogramma yolk-sac larva (Matarese et al., 1989). yolk surface (Fig. 8F). The dorsal finfold is formed on the posterior 1/3 of the body and pigment on the head extends at least to the posterior margin of the eye (Fig. 7G). The liver is prominent and the heart is beating in the live egg. Tail 3/4 around yolk (stage 18) The embryo has 36-41 myomeres. The tip of the tail is tapered and curves away from the longitudinal axis of the embryo (Fig. 7H). The dorsal finfold extends to midbody and pectoral fin buds are prominent. Otic capsules are formed. Large stellate melanophores are scattered over the dorsum, extending just to the midlateral surface; posterior to the anus, two rows of melanophores are seen dorsally and a few are found along the ventral midline (Fig. 8G). The tip of the tail is unpigmented. Tail 7/8 around yolk (stage 19) When the embryo has 44-48 myomeres, the dorsal finfold ex- tends anteriorly 2/3 body length, inserting just pos- terior to the pectoral fin buds and centered over the liver (Figs. 71 and 8H). Pigment on the head extends to the middle of the eye. At midbody, pigment is scat- tered on either side of the dorsal midline, extend- ing to just above the lateral midline. Postanal pigment migrates toward the dorsal and ventral midlines. Tail full circle around yolk (stage 20) The embryo has 48—49 myomeres and the pancreas is visible adjacent to the liver (Fig. 7J). The embryo now encircles the yolk and the tail tip may reach from near the snout to as far back as the posterior margin of the eye (Fig. 81). Hatching glands, simi- lar to those of other teleosts (Yamagami, 1988), are discernible on the surface of the snout and may extend over the dorsal surface of the eye (Figs. 7 J and 7K). The posterior portion of the eye is pig- mented. Postanal pigment migrates and begins to form the postanal bars found in yolk-sac larvae (Matarese et al., 1989) (Fig. 9). Tail 1 1/8 times around yolk (stage 21) The embryo has 49-50 myomeres and the tail tip elon- gates, extending beyond the posterior margin of the eye (Fig. 8J). The urinary bladder is visible posterior to the anus (1/3 body length; not shown on figure) and the dorsal finfold extends to mid- brain. Head pigment extends to the anterior margin of the eye (Fig. 7L). The dorsal half of the eye is pigmented. Most body pigment coalesces to three ar- eas: dorsally, on gut; a bar at 1/2 body length; and a bar at 3/4 body length. In the postanal bars, most pigment is along dorsal and ventral midlines; some pig- ment extends onto the lateral body. Pig- ment is scattered on the preanal body. Discussion Time from first hatch to 50% hatch was inversely related to temperature. Hatch times reflected the effects of temperature described by Yamagami (1988), who demonstrated that the hatching enzyme secreted by the embryo solubilizes the chorion more rapidly at higher temperatures. The first larvae to hatch were stage 20. Early hatching may have been an artifact of rearing conditions. However, hatching glands were present at this stage, which, with the appearance of eye pigment, may correspond to a level of development that would enable these larvae to survive. Early hatching may occur naturally with some frequency. Within batches of walleye pollock larvae from Puget Sound that had been incubated in the laboratory, larvae hatching early grew to an equivalent size as larvae hatching later (larvae hatched on day 1 were the same length at day 3 as larvae hatched on day 3). Those early hatched lar- vae also began to feed at the same time as larvae hatched later.4 Rate of development and time to 50% hatch were similar among studies of walleye pollock from the eastern North Pacific, specifically the Gulf of Alaska (Matarese, unpubl. data; Haynes and Ignell, 1983; and this study; Paul5). From data on time (days) to 50% hatch for all temperatures reported in all in- cubation studies (Fig. 10), incubation times of west- ern North Pacific walleye pollock are longer than eastern North Pacific walleye pollock. This finding appears to conflict with Haynes and Ignell's (1983) comparison with Yusa's (1954) study in which they report similar rates of development 4 Olla, B. Mark O. Hatfield Marine Science Center, Oregon State University, 2030 Marine Science Drive, Newport, OR 97365- 5297. Pers. commun. 18 August 1992. s Paul, A. J. University of Alaska Fairbanks, Institute of Marine Science, Seward Marine Center Lab, P.O. Box 730, Seward, AK 99664. Unpubl. data. 220 Fishery Bulletin 92(2). 1994 for eastern and western stocks. However, their com- parison was made with midpoints of stages calcu- lated from a regression model instead of observed midpoints. Also, Yusa (1954) reported a temperature range of 6— 7"C instead of a mean; our interpreta- tion of Haynes and Ignell's (1983) classification and calculation of Yusa's (1954) data suggests incubation temperatures were always above 6.5°C (see their Table 6 and our Table 4). Finally, Haynes and Ignell (1983) monitored midpoints of stages more closely than midpoint of hatch and did not specifically re- fer to 50% hatch.6 We assumed the values reported as observed midpoints of hatch (their Table 7) were close to 50% hatch. Yusa's (1954) study could not be compared with ours with regard to time to 50% hatch. Time of hatch is often a result of how eggs are treated during incubation and may vary with differ- ent batches.7 However, walleye pollock eggs from Japanese waters are larger than those from the Gulf of Alaska (mean=1.4— 1.6 mm and 1.3—1.4 mm, re- spectively; Bailey and Stehr, 1986). At similar tem- peratures, larger eggs take longer to develop (Pepin, 1991). The difference in incubation time emphasizes the need to collect data from fish specific to the area of interest. This will reduce the sources of variation in develop- ment time for laboratory- reared eggs; failure to identify and improve these sources would compromise the useful- ness of models predicting egg age based on water tem- perature. Development is a continuous process. The sampling intervals and arbitrary designation of stage endpoints break develop- ment into subjective units. Us- ing the 21-stage scheme, we did not see a clear decrease in each stage duration with an increase in temperature. However, this will not affect the usefulness of our results. When stages are grouped to encompass a greater degree of morphological devel- opment, as in Haynes and Ignell (1983) and Picquelle and Megrey (1993), de- velopment time is inversely related to temperature. A greater number of stages within a group increases the accuracy of prediction of egg age. A large num- ber of stages also allows others greater flexibility in grouping those stages. Our regression model predicts temperature-spe- cific development time for purposes of computing rates of egg production and egg mortality. There is no biological basis upon which the regression is predicated because stages that are assigned to the eggs are arbitrary; stages are ordinal data that are based on morphological criteria without consider- ation for development time. An alternative method to estimate development time from temperature is to fit a separate regression for each stage. The dis- advantage of this alternative method is that many parameters are fitted with few data points. Two studies describing morphological develop- ment, Gorbunova (1954) and Yusa (1954), have been published. Gorbunova ( 1954) was not comparable to our study. We compared our descriptions of morpho- logical development with Yusa (1954). We assigned stages to descriptions of hourly morphological devel- 6 Haynes, E., National Marine Fisher- ies Service, Auke Bay Laboratory, 11305 Glacier Highway, Juneau, AK 99801-8626. Pers. commun. April 1991. 7 Paul, A. J., University of Alaska Fairbanks, Institute of Marine Sci- ence, Box 730, Seward, AK 99664. Pers. commun. 17 March 1992. O -*- CO o CO Q 30 25 20 15 10 -&- Matarese 1 983 (see caption) —8- Paul 1 984 (see caption) — This study -B- Haynes and Ignell 1983 — Hamaietal. 1974 -A- Hamai et al. 1971 O Nakatani and Maeda 1 984 Western North Pacific ncubation Studies Eastern North Pacific Incubation Studies 4 6 8 Incubation temperature (°C) 10 12 Figure 10 Days to 50% hatch for Theragra chalcogramma eggs at various temperatures of incubation. (A. C. Matarese, unpubl. data, Alaska Fisheries Science Cen- ter, National Marine Fisheries Service, 7600 Sand Point Way N.E., Seattle, WA 98115. A. J. Paul, unpubl. data, University of Alaska Fairbanks, Insti- tute of Marine Science, Seward Marine Center Lab, P.O. Box 730, Seward, AK 99664.) Blood et al.: Embryonic development of Theragra chalcogramma 221 opment of walleye pollock embryos incubated at 6.0— 7.0°C (Yusa, 1954) for comparison with morphologi- cal characteristics of eggs reared at 5.7°C in this study We used hallmarks of each stage (e.g. num- ber of cells, germ ring advancement, number of myomeres, tail growth around yolk) to distribute Yusa's (1954) descriptions into 21 stages. Yusa's (1954) descriptions were similar to ours up to stage 11. Beginning with stage 11, Yusa (1954) described the development of some structures occurring one or more stages earlier than this study: myomeres and nares were sighted one stage earlier; brain dif- ferentiation and eye lenses, two stages earlier; the heart, three stages earlier; and the otic capsules, five stages earlier (Table 5). Otoliths sighted by Yusa ( 1954) were not visible in our specimens. Conversely, eye pigment was observed in our study one stage earlier than that observed by Yusa (1954). Other structures appeared at the same stage in each study: optic vesicles, Kupffer's vesicle, liver, gut, and pec- toral-fin anlagen. Also, after stage 13, similar num- bers of myomeres were visible at like stages in both studies as was the beating of the heart. Differences between the two studies may be the result of egg condition when examined: Yusa (1954) described live eggs, whereas most of our descriptions were of preserved eggs. Formalin preservation may obscure myomeres or destroy structures such as embryonic otoliths (McMahon and Tash, 1979). Stockard's solution darkens embryonic tissue and obscures fine details. Also, morphological develop- ment may differ between western and eastern North Pacific walleye pollock, further emphasizing the need to restrict data collection to specific areas of interest to increase accuracy of interpretation. Acknowledgments We thank the following people whose combined ef- forts helped us accomplish our research and produce this paper: William Rugen assisted with shipboard experiments; Kevin Bailey, Gail Theilacker, and Steve Porter helped us interpret the morphology of late-stage eggs and yolk-sac larvae; Trish Brown provided statistical analyses; Morgan Busby photo- graphed the eggs; and Beverly Vinter illustrated eggs and helped interpret many morphological struc- tures. We thank Art Kendall, A. J. Paul, Kevin Bailey, Susan Picquelle, and Bori 011a for prelimi- nary reviews of the manuscript. Gail Theilacker and Richard Brodeur helped refine later versions. We also thank the members of FOCI who assisted with field collections. Literature cited Alderdice, D. F. 1988. Osmotic and ionic regulation in teleost eggs and larvae. In W. S. Hoar and D. J. Randall (eds.), Fish physiology, Vol. XI, Part A, p. 163-251. Academic Press, Inc., San Diego. Table 5 Descri (1954) ptions of morphological development of Theragra c and this study. halcogramma embryos at comparable stages by Yusa Stage Yusa (6.0-7.CTC) This study (5.7"C) 11 medullary plate and optic vesicles visible optic vesicles visible 12 5-7 myomeres; 3 sections of brain visible myomeres begin to differentiate 13 9-13 myomeres; heart, otic capsules, otoliths, eye lenses, and Kupffer's vesicle visible 7-12 myomeres; Kupffer's vesicle visible 14 16-17 myomeres; nares and pigment along dorsum visible 14-16 myomeres; pigment along dorsum visible; mid- and hindbrain differentiation 15 18-30 myomeres; liver, gut, and pectoral anlagen visible; 3 sections of brain formed 20-25 myomeres; eye lens, nares, pectoral anlagen, liver, and gut visible 16 35 myomeres 24-36 myomeres; heart visible; 3 sections of brain formed 17 37 myomeres; heart beating 27-36 myomeres; heart beating 18 40 myomeres 36-41 myomeres; otic capsules visible 19 44-48 myomeres 20 48-49 myomeres; eye pigment appears 21 eye pigment appears 222 Fishery Bulletin 92(2). 1994 Bailey, K. M., and C. L. Stehr. 1986. Laboratory studies on the early life history of the walleye pollock, Theragra chalcogramma (Pallas). J. Exp. Mar. Biol. Ecol. 99:233-246. Blaxter, J. H. S. 1969. Development: eggs and larvae. In W. S. Hoar and D. J. Randall (eds.), Fish physiology, Vol. Ill, p. 177-252. Academic Press, New York. Gorbunova, N. N. 1954. Reproduction and development of the wall- eye pollock, Theragra chalcogramma (Pallas). Tr. Inst. Okeanol. Akad. Nauk. SSSR 11:132-195. (In Russian, transl. by S. Pearson, 1972, Natl. Mar. Mammal Lab., NMFS, 7600 Sand Point Way N.E., Seattle, WA 98115-0070.) Hamai, I., K. Kyushin, and T. Kinoshita. 1971. Effect of temperature on the body form and mortality in the development and early larval stages of the Alaska pollock, Theragra chalco- gramma (Pallas). Bull. Fac. Fish. Hokkaido Univ. 22:11-29. Hamai, I., K. Kyushin, and T. Kinoshita. 1974. On the early larval growth, survival, and variation of body form in the walleye pollock, Theragra chalcogramma (Pallas), in rearing ex- periment feeding the different diets. Bull. Fac. Fish. Hokkaido Univ. 25:20-35. Haynes, E. B., and S. E. Ignell. 1983. Effect of temperature on rate of embryonic development of walleye pollock, Theragra chalcogramma. Fish. Bull. 81:890-894. Kendall, A. W., Jr., and S. Kim. 1989. Buoyancy of walleye pollock (Theragra chalcogramma) eggs in relation to water proper- ties and movement in Shelikof Strait, Gulf of Alaska. In R. J. Beamish and G. A. McFarlane (eds.), Effects of ocean variability on recruitment and an evaluation of parameters used in stock assessment models, p. 169-180. Can. Spec. Publ. Fish. Aquat. Sci. 108. Kendall, A. W., Jr., and S. J. Picquelle. 1990. Egg and larval distributions of walleye pol- lock Theragra chalcogramma in Shelikof Strait, Gulf of Alaska. Fish. Bull. 88:133-154. Markle, D. F., and K. G. Waiwood. 1985. Fertilization failure in gadids: aspects of its measurement. J. Northw. Atl. Fish. Sci. 6:87-93. Matarese, A. C., A. W. Kendall Jr., D. M. Blood, and B. M. Vinter. 1989. Laboratory guide to early life history stages of Northeast Pacific fishes. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 80, 652 p. McMahon, T. E., and J. C. Tash. 1979. Effects of formalin (buffered and unbuffered) and hydrochloric acid on fish otoliths. Copeia 1979:155-156. Megrey, B. A. 1991. Population dynamics and management of walleye pollock (Theragra chalcogramma) in the Gulf of Alaska, 1976-1986. Fish. Res. 11:321- 354. Nakatani, T., and T. Maeda. 1984. Thermal effect on the development of wall- eye pollock eggs and their upward speed to the surface. Bull. Jpn. Soc. Sci. Fish. 50:937-942. Naplin, N. A., and C. L. Obenchain. 1980. A description of eggs and larvae of the snake eel, Pisodonophis cruentifer (Ophichthidae). Bull. Mar. Sci. 30:413-423. Olla, B. L., and M. W. Davis. 1993. The influence of light on egg buoyancy and hatching rate of the walleye pollock, Theragra chalcogramma (Pallas). J. Fish. Biol. 42:693-698. Pepin, P. 1991. Effect of temperature and size on develop- ment, mortality, and survival rates of the pelagic early life history stages of marine fish. Can. J. Fish. Aquat. Sci. 48:503-518. Picquelle, S. J., and B. A. Megrey. 1993. A preliminary spawning biomass estimate of walleye pollock, Theragra chalcogramma, in the Shelikof Strait, Alaska, based on the annual egg production method. Bull. Mar. Sci. 53:728-749. Reed, R. K., and J. D. Schumacher. 1989. Transport and physical properties in central Shelikof Strait, Alaska. Cont. Shelf Res. 9:261-268. SAS Institute, Inc. 1985. SAS® user's guide: basics, version-5 edition. SAS Institute, Inc., Cary, NC, 1290 p. Schumacher, J. D., and A. W. Kendall Jr. 1991. Some interactions between young walleye pollock and their environment in the western Gulf of Alaska. Calif. Coop. Oceanic Fish. Invest. Rep. 32:22-40. Trinkaus, J. P. 1951. A study of the mechanism of epiboly in the egg of Fundulus heteroclitus. J. Exp. Zool. 118:269-319. Velsen, F. P. J. 1980. Embryonic development in eggs of sockeye salmon, Oncorhynchus nerka. Can. Spec. Pub. Fish. Aquat. Sci. 49, 19 p. Yamagami, K. 1988. Mechanisms of hatching in fish. In W. S. Hoar and D. J. Randall (eds.), Fish physiology, Vol. XI, Part A, p. 447^99. Academic Press, Inc., San Diego. Yusa, T. 1954. On the normal development of the fish, Theragra chalcogramma (Pallas), Alaska pollock. Bull. Hokkaido Reg. Fish. Res. Lab. 10:1-15. Abstract. — The diel vertical distribution patterns of several abundant ichthyoplankton taxa were examined from depth-strati- fied tows off Kodiak Island in the western Gulf of Alaska during 1986 and 1987. Most larvae were found in the upper 45 m of the water column throughout the diel period but were concentrated in higher densities near the surface (0-15 m) in daylight hours and at greater depths at night. Four of the five dominant taxa examined in detail showed significantly greater weighted mean depths during the night than during the day. This pattern was the opposite to that previously reported for the numerically dominant taxa (Ther- agra chalcogramma) in this area. Since there was no clear relation between the diel vertical distribu- tion of these taxa and the vertical distribution of water temperature and density or copepod nauplii prey, we hypothesize that this re- verse migration is either a strat- egy to minimize spatial overlap with predators that follow a nor- mal diel migration pattern or one to optimize light levels for feeding. Diel vertical distribution of ichthyoplankton in the northern Gulf of Alaska* Richard D. Brodeur William C. Rugen Alaska Fisheries Science Center, National Marine Fisheries Service, NOAA 7600 Sand Point Way NE, Seattle, WA 98 1 1 5 Planktonic eggs and larvae of ma- rine fishes exist in three dimen- sions in the open ocean. Unfortu- nately, traditional ichthyoplankton surveys, which use non-closing sampling gear, provide information only on two dimensions, integrat- ing the vertical indirectly into the horizontal dimensions. It is well known that vertical current shear can be substantial over short dis- tances and that light, temperature, hydrostatic pressure and food show much stronger gradients in the ver- tical relative to the horizontal di- mensions in the water column (Laprise and Dodson, 1993). Thus, a larva can often change not only its geographic position, but also its immediate environment by altering its vertical position in the water column. Diel vertical migration is well documented for larval, juvenile, and adult life history stages of ma- rine fishes (see review by Neilson and Perry, 1990). The adaptive sig- nificance of these migrations is presently in dispute, but it has been attributed to position mainte- nance, bioenergetic optimization, thermoregulation, and predator avoidance (Kerfoot, 1985; Lampert, 1989). In addition, the degree of migration and amplitude of depths over which a species vertically mi- grates often changes during ontoge- netic development (Brewer and Kleppel, 1986; de Lafontaine and Gascon, 1989). Knowledge of vertical distribu- tion patterns of marine fish larvae is crucial not only in understand- ing ecological processes but also has practical implications in the assessment of abundance. Sam- pling just the upper depths of a species range can lead to substan- tial underestimates of abundance, whereas sampling the entire water column for surface-dwelling taxa may waste limited ship time. De- spite the importance of the larval phase in recruitment of marine fishes, relatively little is known about larval vertical distribution patterns off the continental shelf in the North Pacific Ocean. With the exception of walleye pollock, Ther- agra chalcogramma, which has been fairly well studied through much of its geographic range (Kamba, 1977; Kendall et al., 1987; Pritchett and Haldorson, 1989; Kendall et al.1), the only compre- hensive studies on vertical distri- bution of coastal ichthyoplankton in the northeast Pacific Ocean are from the California Current region (Ahlstrom, 1959; Boehlert et al., 1985; Brewer and Kleppel, 1986; Lenarz et al., 1991). This paper presents information on the verti- cal distribution of five abundant ichthyoplankton taxa (other than walleye pollock) collected in the Manuscript accepted 18 October 1993 Fishery Bulletin 92:223-235 (1994) "Contribution No. 0181 of the Fisheries Oceanography Coordinated Investigations. 223 224 Fishery Bulletin 92|2), 1994 coastal waters of Alaska during spring and examines diel differences in these patterns in relation to en- vironmental and biotic factors. Materials and methods Samples examined were collected from two cruises of the NOAA ship Miller Freeman in the area south- west of Kodiak Island in the north- ern Gulf of Alaska (Fig. 1). During May 1986 and 1987, 22 depth-strati- fied tows were made with a 1-m2 Multiple Opening/Closing Net and Environmental Sensing System (MOCNESS) (Wiebe et al., 1976) equipped with 153-um mesh. The net was towed obliquely and nets were opened sequentially at the de- sired depth strata. The primary pur- pose of the sampling was to collect information on the vertical distribu- tion of walleye pollock larvae, which are generally found in the upper 50 m (Kendall et al.1), and their prey. Therefore, the emphasis during the sampling was on the upper part of the water column. The nets sampled the following nominal depths: 0-15, 15-30, 30-45, 45-60, 60-80, 80-100, and >100 m. Maximum sampling depth varied (range 150-252 m) depending on the depth of the water column at a particular station. There were eight depth strata sampled at most sta- tions but the cutoff depth between the seventh and eighth net was variable. Therefore, we pooled the catches from these two nets into a single depth stra- tum (>100 m) for analysis. The actual sampling depths are given in Table 1. More complete station and catch information is given in Siefert et al.2,3 -i 1 1 1 1 " 1- cX S j / { Kodiak i y Island A) / + 4 / FJ -57 00 <^y <-, ;• c \ / -£> +5 > C^> +9 ' Sutwik 1. I , /?Ck. r ~ ' (/ Trinity Is. +7 ' / i f Semidi Is. ^ / 0-6 > N + - " Chirikof 1. -56 00 ;- 4 / 1 V / ' , s * 00 TL , . 1 . 157 00W 156 00 1 ' (■ 1 55 00 1 54 Figure 1 Location of MOCNESS sampling series in Shelikof Strait used to de- termine vertical distribution of larvae in 1986 and 1987. 1 Kendall, A. W., Jr., L. S. Incze, P. B. Ortner, S. R. Cummings, and P. K. Brown. In review. The vertical distribution of eggs and larvae of walleye pollock {Theragra chalcogramma i in Shelikof Strait, Gulf of Alaska. Submitted to Fish. Bull. 2 Siefert, D. L. W., L. S. Incze, and P. B. Ortner. 1988. Vertical distribution of zooplankton, including ichthyoplankton, in Shelikof Strait, Alaska: data from Fisheries Oceanography Coordinated Investigations (FOCI) cruise in May 1986. NWAFC Processed Rep. 88-28, 232 p. 1 Siefert, D. L. W., L. S. Incze, and P. B. Ortner. 1990. Vertical distribution of zooplankton, including ichthyoplankton, in Shelikof Strait, Alaska: data from Fisheries Oceanography Coordinated Investigations (FOCI) cruise in May 1987. NWAFC Processed Rep. 90-05, 129 p. The 22 tows were grouped into five collection se- ries (Table 1) based upon date and location of sam- pling (see Kendall et al.1) and included two complete diel series. The first diel series (Series 4) attempted to sample the same body of water over a four day period during 1986 by following a radar-tracked drifter drogued at 35 m (Incze et al., 1990). The second diel series (Series 9) sampled the same loca- tion on three successive days during 1987. Other collections (Series 5, 6, and 7) were taken at vari- ous times of the day but in the same general area as these two series (Fig. 1, Table 1). Retrieved nets were thoroughly washed and con- tents were preserved in 5% buffered formalin. Samples were sorted to the lowest possible taxon and life history stage at the Polish Sorting Center in Szczecin, Poland. The volume filtered was esti- mated from a mechanical flowmeter mounted on the MOCNESS frame and abundances were converted to number per 1000 m3. Up to 50 preserved larvae of each taxon from each net were measured to the nearest 0. 1 mm standard length. Net depth, temperature, and Brodeur and Rugen: Vertical distribution of ichthyoplankton in the northern Gulf of Alaska 225 Table 1 Station and tow data for collection subset used in the diel series ordered by time of day. Net depths (m) Bottom Local Time Series Tow Year Date depth (m) time period 1 2 3 4 5 6 7 4 5 1986 10 May 293 0745 Dawn 2- -15 15- -30 29- 45 46- -58 59 -78 79- -99 101- -229 4 1 1986 8 May 220 0746 Dawn 1- 15 15- -31 35- 45 45- -61 61- -80 80- -100 101- -200 9 4 1987 23 May 201 0747 Dawn 0- 15 15- -30 30- -45 45- -60 60- -80 80- -100 100- -150 9 8 1987 24 May 190 0800 Dawn 0- 15 15 30 30 -45 45 ■60 60- -80 80- -100 100- -150 6 2 1986 L5 May 223 0940 Day 2- -14 15- 29 30 44 45 -60 61 -80 80- -100 100- -175 5 1 1986 13 May 210 1010 Day 2- 15 16 -30 30- -45 45- -61 61 -80 80 -100 101- -152 4 6 1986 111 May 296 1341 Day 2- 14 30- 45 77- -99 99- -214 5 2 1986 13 May 210 1351 Day 2- -14 15 -29 30- 4 4 45- -59 59- -78 80- -99 100- -176 4 2 1986 8 May 227 1356 Day 2- -14 15- -30 30- -47 79- -99 99- -200 9 5 1987 23 May 179 1422 Day 0- 15 15- -30 30- -45 45 -60 60- 80 80- -100 100- -150 9 9 1987 24 May 191) 1543 Day 0- 15 15 -30 30- -45 45 -60 60 -80 80- -100 100- -150 7 2 1986 18 May 123 1911 Dusk 2 15 15 30 30 45 45 -59 60 -79 so -100 4 3 1986 9 May 235 2006 Dusk 2- -13 13- 28 29- -45 46 59 60- -79 100- -200 4 7 1986 11 May 293 2011 Dusk 3- -14 14 -30 31- 44 43- -59 60- -78 78 -97 98- -252 9 6 1987 24 May 179 2107 Dusk 0- -15 15 -30 30- -45 45 -60 60 -80 80- -100 100- -150 9 Hi 1987 25 May 196 2122 Dusk 0- -15 15- -30 30- -45 45 -60 60 -80 80- -100 100- -150 5 3 1986 14 May 210 2200 Night 2- -14 14- -30 30 45 46 -60 60 -80 81- -101 100- -163 7 1 1986 17 May 126 2416 Night 1 15 16 30 31- 45 45- -59 60 -80 80 -99 4 4 1986 9 May 242 0135 Night 2- -15 15- 30 30- -58 59- -80 80- -100 100- -202 9 11 1987 25 May 198 0218 Night 0- -15 15 30 30- 45 45 -60 60 -80 80- -100 100- -150 9 7 1987 24 May 195 0219 Night 0- 15 15 30 30- 45 45- 60 60- -80 80- -100 100- -150 6 1 1986 15 May 229 0353 Night 2- -15 15- 29 30 -44 45- -60 60 80 80 -100 100- -172 salinity were measured continuously in real time dur- ing the tow and stored for later analysis. To examine diel variations in density and size of larvae with depth, collections from the 22 tows were grouped into one of four time periods (hours): dawn (0530-0830), day (0830-1830), dusk (1830-2130), and night (2130-0530). Diel-depth variation in den- sity of eggs and larvae at each depth was examined by using a two-way ANOVA on log-transformed data. The log (X+l) transformation was used to achieve homogeneous variances (Bartlett's Test, all P>0.05). In addition, a weighted mean depth of occurrence of eggs or larvae of the dominant species for each time interval was calculated as follows: IIXa,, A- m where n( = number of tows in time interval t , N = number of larvae in net j in tow i in '.it time interval t, D = midpoint depth of net j in tow i in time interval t with a variance equal to Var(Dt) = ( "' ] w=i ; 2>2(Z)a-A>2, K-i) & where N(/ = number of larvae in tow i in time interval t. Differences in the weighted mean depths over the four time periods were tested with ANOVA, and Tukey multiple-comparison tests were conducted when significant differences were observed. Untransformed larval lengths for the three most abundant species were entered as dependent vari- ables in two-way ANOVAs, with time of day and depth as factors. Results Species composition Eggs and larvae of species other than walleye pol- lock were found in 134 of the 145 samples collected 226 Fishery Bulletin 92(2). 1994 during the 1986 and 1987 cruises. Flathead sole (Hippoglossoides elassodon) eggs were the only pe- lagic eggs other than walleye pollock collected and were found in 28.4% of the samples. This species had a mean density of 62.99 eggs/1000 m3 (SD=179.66) and comprised 74.9% of the total egg abundance in the 22 tows. A total of 33 larval taxa were identified but only a few taxa occurred in more than 10% of the samples (Table 2). Larvae other than walleye pollock oc- curred in 92.4% of the collections but made up only 26.3% of the overall total abundance of larvae (to- tal mean density=143.61 larvae/1000 m3; SD=257.03). Larvae of three taxa (H. elassodon , Am- modytes hexapterus, and Bathymaster spp.4) were found at sufficient densities to enable examination of their vertical abundance and length distribution patterns in detail for the four time periods. Two other species (Gadus macrocephalus and Pleuro- nectes bilineatus) were found at relatively high den- sities during day and night but at low densities during the twilight periods; hence, these taxa were examined only for day-night differences. Vertical distribution The distribution of//, elassodon eggs showed little variation in weighted mean depth by time of day (F=3.10, P>0.05); the highest abundances were 4 Larvae of three Bathymaster species known to occur in the study area are presently not identifiable to species. Based on the abundance and distribution patterns of the adults, most of the larvae present in our collections are probably B. signatus. (A. Matarese, Alaska Fisheries Science Center, Seattle, WA 98115. Pers. commun. 1992). Table 2 Summary of all larvae including walleye pollock collected in the 1986-87 vertical distribution study. Percent Mean Length occurrence density range Scientific name Common name (n=145) (no./1000m:i) (mm) Osmerus mordax rainbow smelt 0.69 0.25 21 Leuroglossus schmidti northern smoothtongue 0.69 0.02 9-15 Stenobraehius leueopsarus northern lampfish 4.14 3.05 4-7 Protomyctophum thompsoni bigeye lanternfish 0.69 0.04 10 Myctophidae unidentified myctophid 0.69 0.06 3 Gadus macrocephalus Pacific cod 15.86 27.51 3-11 Theragra chalcogramma walleye pollock 93.79 402.71 3-8 Gadidae unidentified gadid 2.76 0.36 4 Sebastes spp. unidentified rockfish 1.38 0.17 4-5 Hexagrammos decagrammus kelp greenling 1.38 0.05 8-11 Dasycottus setiger spinyhead sculpin 0.69 0.07 8 Gymnocanthus spp. unidentified sculpin 0.69 0.07 7-8 Hemilepidotus hemilepidotus red Irish lord 1.38 0.13 11-13 Icelinus spp. unidentified sculpin 7.59 0.65 4-5 Malacocottus zonurus darkfin sculpin 0.69 0.05 6-7 Radulinus asprellus slim sculpin 1.38 0.10 4-5 Ruscarius meanyi Puget Sound sculpin 0.69 0.04 4 Agonidae unidentified poacher 10.34 0.84 5-10 Nectoliparis pelagicus tadpole sculpin 0.69 0.07 4-8 Cyclopteridae unidentified snailfish 2.07 0.19 4-5 Bathymaster spp. unidentified ronquil 13.10 30.67 4-7 Anoplarchus spp. unidentified prickleback 2.07 0.26 8-10 Lumpenella longirostris longsnout prickleback 1.38 0.05 10-11 Lumpenus maculatus daubed shanny 6.21 0.64 12-23 Poroclinus rothrocki whitebarred prickleback 4.83 0.97 10-15 Cryptacanthodes aleutensis dwarf wrymouth 1.38 0.19 14 Pholis spp. unidentified gunnel 0.69 0.04 13-17 Zaprora silenus prowfish 3.45 0.40 12-14 Ammodytes hexapterus Pacific sand lance 40.69 12.76 6-19 Hippoglossoides elassodon flathead sole 17.93 59.81 4-19 Pleuronectes bilineatus rock sole 13.10 3.71 3-10 Pleuronectes vetulus English sole 0.69 0.13 8 Psettichthys melanostictus sand sole 0.69 0.14 4-5 Pleuronectidae unidentified flounder 0.69 0.10 4 Brodeur and Rugen: Vertical distribution of ichthyoplankton in the northern Gulf of Alaska 227 found in the surface layer (0-15 m) during all four time periods (Fig. 2). Although there were signifi- cant (P=0.005) differences in density by depth strata, neither the diel density differences alone (P=0.838) nor the interaction between time and depth (P=0.996) was significant. The majority of larvae, excluding pollock larvae, from all collections combined were collected from the upper three depth strata (Fig. 3). The maximum density overall occurred at the second depth stra- tum ( 15-30 m), below which larval density declined with depth. However, this overall vertical distribu- tion pattern was apparently confounded by higher larval densities found during the night when the larvae were mainly caught in the 15-30 m stratum; Dawn Day ro CD c Q. 100 - 0-16 15-30 30-45 45-60 60-80 80-100 > 100- } 100 200 300 No/1000 m3 0 100 200 300 400 No/1000 m3 Dusk Night -15 ■30 •45 0 15 30 45 -60 - I 60 -80 - I 80- 100 100 - 15-30 30-45 } 45-60 60-80 ^ 80-100 ) > 100 100 200 300 400 NO./1000 m' 0 100 200 300 400 No/1000 m3 Figure 2 Diel vertical distribution of Hippoglossoides elassodon eggs. Bars are mean abundances per 1000 m3 at each depth interval and error bars are ± one standard deviation about the mean abundance. during the other three time periods the highest den- sities were in surface waters (Fig. 4). The weighted mean depth of larvae overall was significantly (P<0.05) greater at night than during the other three time periods (Table 3) and the interaction between time and depth was marginally significant (P=0.05; Table 4), suggesting that there were diel differences in overall larval depth distribution. Four of the five most abundant larval taxa showed the greatest weighted mean depths (Table 3) and the lowest surface densities (Fig. 4) at night. This gen- eral pattern was also evident in the two time peri- ods examined for the fifth species, G. macro- cephalus, but the diel differences were not signifi- cant (Table 3). Only A. hexapterus and G. macro- cephalus showed significant diel dif- ferences in larval density, with high- est densities occurring at night (Table 4). None of the dominant taxa, however, showed a significant interaction between time and depth strata. Length distributions The distribution of larval lengths by time of day and depth showed no consistent pattern among the three most abundant species (Fig. 5). Al- though time and time-depth interac- tions were significant (all P<0.03) factors in explaining the variation in mean length of H. elassodon and Bathymaster spp., none of the fac- tors was significant for A. hexap- terus. Examining only the strata where more than two lengths were available, we found that the small- est larvae of both Bathymaster spp. and A. hexapterus were caught in the surface stratum at night but in deeper strata during daylight hours (Fig. 5). However, H. elassodon showed an increase in mean length with depth during daylight hours and the reverse pattern at night (Fig. 5). Hippoglossoides elassodon was the only taxon to show a signifi- cant difference in length distribu- tions between night and day collec- tions (Kolmogorov-Smirnov Test; Z=3.881; P=0.001). Although the lack of larger larvae in daytime col- lections might suggest some daytime gear avoidance by this species 228 Fishery Bulletin 92(2). 1994 (Fig. 6), there were few small larvae caught at night, which cannot be explained by gear avoidance. Since the majority (>95%) of these lengths were from lar- 0-15 I I r I 15-30 I i E — 30-45 N I n | 45-60 3 .c cj 60-80 - a 80-100 > 100 0 100 200 300 400 500 600 NO./1000 mJ Figure 3 Vertical distribution of all larvae ex- cluding walleye pollock (Theragra chalcogramma) combined over all time periods. Bars are mean abundances per 1000 m3 at each depth interval and error bars are ± one standard deviation about the mean abundance. vae collected from the same location (Series 9), sam- pling variability cannot be invoked as an explana- tion for this pattern. Discussion Our results indicate that the vast majority (>99%) of pelagic eggs and larvae (excluding walleye pol- lock) are distributed in the upper 100 m of the wa- ter column during the spring months. Therefore, sampling to this depth should be sufficient to char- acterize the horizontal distribution patterns of these species. Of the common taxa we examined, all but H. elassodon have demersal eggs (Matarese et al., 1989). The transit time to surface waters following hatching from demersal eggs is apparently of such short duration that even newly hatched larvae were rarely collected below 100 m. However, this does not appear to be the case for walleye pollock, which spawn at depths greater than 200 m in Shelikof Strait, with mean depths of eggs and yolk-sac lar- vae generally greater than 100 m (Kendall and Kim, 1989; Kendall et al.1). The diel vertical distribution pattern that we ob- served for several taxa is not the pattern typically observed for most ichthyoplankton and for zooplank- ton in general. The more common pattern, termed a 'Type F migration (Neilson and Perry, 1990), in- volves a nocturnal ascent into surface waters and is undertaken by larvae of a diversity offish species. Table 3 Weighted mean depths (m) and standard deviations of the mean depths (in parentheses) for each taxon and for all larvae excluding walleye pollock by time of day and overall depth for all times combined. Also given are the results of the ANOVAs testing for diel differences in weighted mean depth and the significant (P< 0.05) Tukey multiple-comparison tests between time periods. Dawn Day Dusk Night Overall F-value Tukey test All Larvae (excluding walleye pollock) 16.59(2.72) 17.46(1.52) 15.45(2.25) 25.74 (1.52) 21.75 (1.87) 33. 17**' Night>Day=Dawn=Dusk Hippoglossoides elassodon 14.82(0.63) 16.94(2.73) 10.80 (0.42) 20.10 (0.05) 18.06 (1.33) 3] 89 Nigh t > Day = Da wn> Dusk Ammodytes hexapterus 31.21(11.45) 27.67 (4.59) 22.67 (2.38) 37.51 (4.45) 32.85 (3.39) 6 05" ' Night>Dusk = Day Bathymaster spp. 8.21 (0.11) 11.25 (0.08) 11.28 (1.08) 37.95 i2.84) 18.12 (6.05) 441.48*** Nigh t>Dusk= Day > Dawn Pleuronectes bilineatus 19.75 (1.83) 30.73 (1.63) 25.47 (4.64) 128.36*** Night>Day Gadus macrocephalus 20.36(10.35) 24.92 (0.12) 22.12 (6.56) 1.14 n.s. P<0.001, " P<0.01; n.s. P>0.05. Brodeur and Rugen: Vertical distribution of ichthyoplankton in the northern Gulf of Alaska 229 However, the reverse pattern ('Type II' migration), although less frequently documented, has been ob- served for larvae of several fish species, including many of the taxa we examined. For example, Boehlert et al. (1985) observed larval G. macrocephalus at lower depths at night than dur- ing the day off the Oregon coast. Walline5 found that Bathymaster spp. in the Bering Sea generally mi- grated downward at night. Larvae of A. hexapterus collected in bays around Kodiak Island were concen- trated from 10 to 30 m during the day but were found at lower depths at night (Rogers et al.6), and larvae of a congener (A. personatus) collected off Ja- pan also exhibited reverse migration (Yamashita et al., 1985). Rogers et al.6 and Pritchett and Haldor- Table 4 Results of two-way ANOVAs testing for differences in density of larvae by depth and time of day. Sum of Mean Source of variation df squares square F-ratio P-value All larvae (excluding walley B pollock) Time 3 14.60 4.86 9.85 0.00 Depth 6 51.63 8.61 17.40 0.00 Time x depth 18 14.11 0.78 1.58 0.05 Error 4868 2406.73 0.49 Hippoglossoides elassodon Time 3 4.97 1.66 0.48 0.69 Depth 6 93.59 15.60 4.54 0.00 Time x depth 18 15.49 0.86 0.25 0.99 Error 116 398.59 3.44 Ammodytes hexapterus Time 3 55.08 18.36 11.70 0.00 Depth 6 94.13 15.69 9.99 0.00 Time x depth 18 34.27 1.90 1.21 0.26 Error 116 182.03 1.57 Bathymaster spp. Time 3 8.84 2.95 1.24 0.30 Depth 6 44.33 7.39 3.10 0.01 Time x depth 18 21.35 1.19 0.49 0.96 Error 116 276.73 2.39 Pleuronectes bilineatus Time 1 1.68 1.68 1.35 0.25 Depth 6 19.01 3.17 2.55 0.03 Time x depth 6 3.47 0.58 0.47 0.83 Error 69 85.78 1.24 Gadus macrocephalus Time 1 9.12 9.12 4.09 0.05 Depth 6 22.46 3.74 1.68 0.14 Time x depth 6 7.06 1.18 0.53 0.79 Error 69 153.81 2.23 son (1989) found that rock sole (P. bilineatus), as well as larvae of several other taxa, showed reverse diel migrations during the spring. We believe that sampling bias could not have re- sulted in the observed reverse distributions. Eggs of H. elassodon, as expected, showed no differences by time of day in our study and walleye pollock larvae in these same collections exhibited a normal diel mi- gration pattern (Type I), occurring mainly in the 30- 45 m range during daytime and above 30 m at night (Kendall et al.1; see also Kendall et al., 1987). Net avoidance, although suggested by the higher night catches overall as well as the larger mean size of larvae collected at night, is not a plausible expla- nation for the observed diel pattern. Light-aided daytime avoidance would be ex- pected to influence the catch of lar- vae in the surface strata more than those in deeper strata, thus leading to underestimates of near-surface daytime abundances and the mag- nitude of reverse migration. The prevalence of the reverse diel migration pattern in our study suggests an adaptive role for this behavior. Temperature gradients are relatively minor (<1°C) over the upper 50-60 m where most of the migration oc- curs (Fig. 7), and the majority of the larvae appear to be above the seasonal thermocline at all times of the day. Thus, we see no possi- bility of temperature-mediated energetic advantage related to migration at any time of the day. Similarly, observed density gradi- ents are not pronounced (<0.5 ot units) within this surface layer (Fig. 7; Kendall et al.1) and there appears to be no physical mecha- nism that would aggregate either 5 Walline, P. D. 1981. Hatching dates of walleye pollock (Theragra ehalco- gramma) and vertical distribution of ichthyoplankton from the eastern Bering Sea, June-July 1979. NWAFC Processed Rep. 81-05, 22 p. 6 Rogers, D. E., D. J. Rabin, B. J. Rogers, K. J. Garrison, and M. E. Wangerin. 1979. Seasonal composition and food web relationships of marine organisms in the nearshore zone of Kodiak Island including ichthyoplankton, mero- plankton (shellfish), zooplankton and fish. Univ. Washington, Fish. Res. Inst. Rep. FRI-UW-7925, 291 p. 230 Fishery Bulletin 92(2). 1994 DAWN DAY DUSK NIGHT All non-Pollock larvae 0-15 15-30 30-45 1 45-60 60-80 J 80-100 > 100 3- i 40 80 120 0 50 100 150 200 0 40 80 120 Hippoglossoides elassodon 0-15 15-30 30-45 45-60 60-80 80-100 > 100 ' • 0 50 100 150 200 0 100 200 300 0 40 80 120 160 Bathymaster spp. 0-15 I I *—* 15-30 t 30-45 >m^ 45-60 CO — 60-80 J~ 80-100 *- > 100 } I 4. 200 400 600 800 3- i 0 10 20 30 40 50 100 200 300 0 40 80 120 160 0 20 40 60 80 Ammodytes hexapterus C 0-15 I I 15-30 +- 30-45 © 45"6° p| 60-80 80-100 > 100 | I 4. Gadus macrocephalus Pleuronectes bilineatus * T 3- 10 20 30 0 10 20 30 40 50 0 20 40 60 80 100 10 20 30 40 3- 12 8 10 8 12 5 10 15 20 0 10 20 30 40 50 Density (no. /1 000m3) Figure 4 Diel changes in the vertical distribution of all larvae (excluding walleye pollock), and Hippoglossoides elassodon, Bathymaster spp., Ammodytes hexapterus, Gadus macrocephalus, and Pleuronectes bilineatus larvae. Bars are mean abundances per 1000 m3 at each depth interval and error bars are ± one standard deviation about the mean abundance. Brodeur and Rugen: Vertical distribution of ichthyoplankton in the northern Gulf of Alaska 231 DAWN DAY DUSK NIGHT 0-15 15-30 30-45 Hippoglossoides ^ 45-60 elassodon £ 60-80 80-100 CO » 100 ~-, -^ ^^ • • co < 5 6 7 8 9 10 <■ 56789 10 456789 10 456789 10 > 0-15 1- 15-30 ® 30-45 Bathymaster tT 45-60 SPP- •- 60-80 80-100 -*- > 100 ♦ . — . — , ^^ Q. 15 6/ i 5 6 r t 5 6 7 4 5 6 7 0) 0-15 Q 15-30 Ammodytes 30-45 hexapterus 45.60 60-80 80-100 ' 100 • * ^_ — — • , . , • 4 8 12 16 20 4 8 12 16 20 4 8 12 16 20 4 8 12 16 20 Length (mm) Figure 5 Diel vertical distribution of larval lengths of Hippoglossoides elassodon, Bathymaster spp., and Ammodytes hexapterus. Circles are mean length at each depth interval and error bars are ± one standard deviation about the mean length. The plus signs indicate actual lengths measured when less than three lengths were available from a particular depth stratum. 0.30 0.25 0.20 c o £ 0.15 a. o £ 0.10 0.05 0.00 ! UDay □ Night 2 6 7 8 Length (mm) 10 Figure 6 Day versus night proportional length distribu- tions of Hippoglossoides elassodon larvae. larvae or their prey at certain depths or inhibit them from migrating to different depths. The fact that walleye pollock larvae, which are the dominant fish larvae in this area representing 70- 80% of the larvae present in Shelikof Strait in the spring (Rugen7; this study), show a normal migra- tion pattern (Kendall et al., 1987) suggests one po- tential explanation for reverse migration patterns of other larvae. If other larvae feed on the same microzooplankton prey as larval walleye pollock and these prey resources were limiting, then the pres- ence of these other larvae in surface waters at dif- ferent times of the day than those of walleye pol- lock would reduce competition with the numerically dominant taxon. Copepod nauplii, an important 7 Rugen, W. C. 1990. Spatial and temporal distribution of lar- val fish in the Western Gulf of Alaska, with emphasis on the period of peak abundance of walleye pollock tTheragra chalcogramma) larvae. NWAFC Processed Rep. 90-01, 162 p. 232 Fishery Bulletin 92(2), 1994 Water temperature and density DAWN NIGHT DAY DUSK Sigma-t 2S2 253 254 26.5 266 25.2 25 3 254 25 5 25 6 25 1 25 2 253 25 4 25 5 25.2 25.3 254 25 5 25 6 E 20 -— - -C •*—• a. a> 60 Q BO sigma-t temp 3 0 4 0 6 0 60 30 4 0 5 0 6 0 30 4 0 5 0 6 0 3.0 4 0 5 0 6 0 Temperature (°C) Figure 7 Diel vertical profiles of temperature and density (ot). Data are means of at least four casts within each time interval and were collected at 1 m depth intervals. component of the diet of many larval fishes includ- ing walleye pollock (Kendall et al., 1987), were the most abundant microzooplankton category found in Shelikof Strait, mostly in the upper 30 m during May 1986 and 1987 (Incze and Ainaire8). During diel Series 4, copepod nauplii had overall mean depths between 20 and 34 m but showed no obvious diel pattern in depth distribution (Kendall et al.1). Al- though feeding at a different time of day from wall- eye pollock might reduce interference competition (i.e. behavioral interactions) with the dominant spe- cies, it is highly unlikely, based on typical larval fish and copepod naupliar densities, that prey resources could ever be depleted by larval fish (Cushing, 1983; MacKenzie et al., 1990). Moreover, if food were lim- iting, then it would be advantageous for all larvae to stay in the layer of maximum food concentration throughout the diel period to maximize total intake. Thus, we do not see a trophic benefit accruing from a reverse migration pattern for these larvae. If feeding by these larvae is periodic and depen- dent on some minimum light level, then the verti- cal distribution pattern can be partially explained by larval feeding response. Assuming light levels were limiting feeding at depths below 30 m, then it would be necessary for larvae to ascend to a shal- lower depth during the daytime when light is at a maximum. Following the cessation of feeding at dusk, larvae would be expected to become inactive and passively sink to deeper levels at night. Such a mechanism has been postulated for Japanese sand lance (A. personatus) by Yamashita et al. ( 1985) who demonstrated a nocturnal cessation of feeding in this species. Although we lack data on the diel feed- ing chronology of any of the taxa examined here, it is possible that feeding occurs mainly in the crep- uscular periods, with a temporary cessation of in- gestion occuring during midday as observed in the field for larval walleye pollock (Canino and Bailey9). The shallowest mean depth occurs at either dawn or dusk for the three common species that were ex- amined over the four time periods with slightly greater depths occurring during midday. If larvae were not feeding during the middle of the day, it would be advantageous to cease swimming alto- gether and sink through the water column to avoid being sensed by mechanoreceptive or visual preda- tors. Following a particular isolume would produce a similar daytime pattern but could not account for the deeper distribution at night that we observed. Larval walleye pollock in the laboratory have been shown to avoid high light levels (Olla and Davis, 1990) but they also require relatively low light lev- els to initiate feeding (Paul, 1983). Unfortunately, we have no data available on the light levels neces- sary for feeding in the taxa we examined with which we can evaluate this hypothesis. 8 Incze, L. S., and T. Ainaire. In review. Zooplankton of Shelikof Strait, Alaska. I. Micro-zooplankton prey of larval pollock, Theragra chalcogramma. Submitted to Fish. Bull. 9 Canino, M. F., and K. M. Bailey. In review. Gastric evacuation of walleye pollock, Theragra chalcogramma (Pallas), larvae in response to feeding. Submitted to Journal of Fish Biology. Brodeur and Rugen: Vertical distribution of ichthyoplankton in the northern Gulf of Alaska 233 A potential disadvantage to a diurnal ascent is increased susceptibility to visually feeding planktivorous fishes. However, acoustic and trawl survey data suggest that epipelagic fish predators are rare during the spring in this area and the majority of the nekton biomass is found in midwater or near the bottom (Brodeur et al., 1991), well be- low the depth of most larvae. On the other hand, euphausiids, which are possibly the major inverte- brate predator on walleye pollock yolk-sac larvae, undergo a nocturnal ascent to surface water and descend to greater depths during the day in Shelikof Strait (Bailey et al., 1993). If euphausiids were also predators on non-pollock larvae and feed only in the surface layer above the nightime depths of these lar- vae, then a distinct advantage would be conferred upon individuals adopting a reverse diel migration pattern, as has been postulated for copepods (Ohman et al., 1983; Ohman, 1990). Based on field and experi- mental results, it has become increasingly apparent that predators can alter the diel vertical distribution patterns of invertebrate prey (Ohman et al., 1983; Gliwicz, 1986; Bollens and Frost, 1989; Levy, 1990; Neill, 1990; Frost and Bollens, 1992), but evidence for this effect on larval fish as prey is presently lacking. Although a variation in depth by time of day was apparent for all species and consistent among spe- cies, it was not substantial enough to be statistically significant in all cases (e.g. G . macrocephalus). This may be due in part to the lack of resolution of our sampling intervals. The smallest average migration that we could detect is -15 m; thus, diel vertical migrations less than that were not likely to be de- tected. Although a daily ambit of 30 m is not excep- tional for larger larvae, it may be excessive for newly hatched individuals. For a study specifically exam- ining the diel vertical distribution of the species considered here, we recommend sampling with a multiple net system every 5 m over the upper 40 m of the water column. Some bias may have also re- sulted from combining tows from different years, weeks, or geographic areas into our four time peri- ods, which was necessitated by the relatively low oc- currence rate and densities of these taxa. However, the remarkably strong and consistent diel differ- ences among the different taxa, despite this intro- duced sampling variability, lend credence to our findings. If there was differential migration by size classes of larvae, this condition might also obscure some of our results. The vertical distribution of larval lengths of the dominant species did not show any consistent patterns by time of day. The mean length by depth varied significantly for H. elassodon; smaller larvae were found at greater depths during the daytime and at the surface at night. This can- not be explained by visual gear avoidance alone since the nighttime pattern would then be expected to be random rather than exhibit the increasing mean length with depth that we observed. A possible explanation for this pattern might be that larger larvae may migrate a greater distance than smaller larvae, a pattern frequently observed in other fish larvae (Neilson and Perry, 1990). It is also possible that the migration of different size classes is asyn- chronous (Pearre, 1979). However, the available size ranges of the dominant species in our data was not extensive enough to examine diel migration patterns of different size classes. Moreover, caution should be exercised in examining larval length data in mul- tiple net systems. Since larvae shrink upon death (Theilacker, 1980; Hay, 1981) and the likelihood of death may be related to time in net, we may assume that larvae caught in the first (deepest) net may have undergone more shrinkage than those in the last (surface) net. In conclusion, this study shows that all the com- mon larvae exhibit a reverse vertical migration pat- tern, opposite to that of the overall dominant spe- cies, walleye pollock. In Auke Bay, an inland embayment in Southeast Alaska (58°22' N) on the eastern side of the Gulf of Alaska, Haldorson et al. ( 1993) found a Type I migration for the numerically dominant osmerid larvae in their sampling and a Type II migration for the five next most abundant taxa {T. chalcogramma, H. elassodon, P. bilineatus, Leuroglossus schmidti, and Agonidae). These au- thors attribute this diel-depth distribution pattern to temperature preferences by each species, al- though their vertical temperature gradients were more pronounced than what we observed in our study. Since most abiotic variables (other than light intensity) and food resources varied little over the depths through which much of the migration oc- curred in Shelikof Strait, we hypothesize that the reverse migration pattern that we documented was either a predator-avoidance mechanism or else an optimization of light levels for feeding. The preva- lence of reverse migration in this and other studies suggests that it may be more common than previ- ously suspected, especially in higher latitude ecosys- tems, and the factors contributing to this phenom- enon merit further investigation. Acknowledgments The MOCNESS tow collections used in this study were made available by Lew Incze (Bigelow Labo- ratory) and Peter Ortner (Atlantic Oceanographic 234 Fishery Bulletin 92(2). 1994 and Meteorological Laboratories, NOAA). Field as- sistance was provided by Shailer Cummings (AOML) and the crew of the NOAA ship Miller Free- man. Susan Picquelle and Patricia Brown (Alaska Fisheries Science Center) assisted in statistical analysis. Art Kendall, Gary Stauffer, JeffNapp, Bori Olla, Susan Sogard, and Michael Davis (AFSC), R. Ian Perry (Pacific Biological Station), Steven Bollens (Woods Hole Oceanographic Institution), and two anonymous reviewers provided valuable comments on earlier versions of the manuscript. Literature cited Ahlstrom, E. H. 1959. Vertical distribution of pelagic fish eggs and larvae off California and Baja California. U.S. Fish. Wild. Serv., Fish. Bull. 60:107-146. Bailey, K. M., R. D. Brodeur, N. Merati, and M. M. Yoklavich. 1993. Predation on walleye pollock (Theragra chalcogramma) eggs and yolk-sac larvae by pelagic crustacean invertebrates in the western Gulf of Alaska. Fish. Oceanog. 2:30-39. Boehlert, G. W., D. M. Gadomski, and B. C. Mundy. 1985. Vertical distribution of ichthyoplankton off the Oregon coast in spring and summer months. Fish. Bull. 83:611-621. Bollens, S. M., and B. W. Frost. 1989. Predator-induced diel vertical migration in a planktonic copepod. J. Plankton Res. 11:1047- 1065. Brewer, G. D., and G. S. Kleppel. 1986. Diel vertical distribution of fish larvae and their prey in nearshore waters of southern California. Mar. Ecol. Prog. Ser. 27:217-226. Brodeur, R. D., K. M. Bailey, and S. Kim. 1991. Cannibalism on eggs by walleye pollock Theragra chalcogramma in Shelikof Strait, Gulf of Alaska. Mar. Ecol. Prog. Ser. 71:207-218. Cushing, D. H. 1983. Are fish larvae too dilute to affect the den- sity of their food organisms? J. Plankton Res. 847-854. de Lafontaine, Y., and D. Gascon. 1989. Ontogenetic variation in the vertical distri- bution of eggs and larvae of Atlantic mackerel (Scomber scombrus). Rapp. P.-v. Reun. Cons. int. Explor. Mer 191:137-145. Frost, B. W., and S. M. Bollens. 1992. Variability of diel vertical migration in the marine planktonic copepod Pseudocalan us newmani in relation to its predators. Can. J. Fish. Aquat. Sci. 49:1137-1141. Gliwicz, M. Z. 1986. Predation and the evolution of vertical migra- tion in zooplankton. Nature 320:746-748. Haldorson, L., M. Prichett, A. J. Paul, and D. Ziemann. 1993. Vertical distribution and migration of fish larvae in a Northeast Pacific bay. Mar. Ecol. Prog. Ser. 101:67-80. 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 178:395-^100. Incze, L. S., P. B. Ortner, and J. D. Schumacher. 1990. Microzooplankton, vertical mixing and advec- tion in a larval fish patch. J. Plankton Res. 12:365-379. 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., and S. Kim. 1989. Buoyancy of walleye pollock (Theragra chalcogramma) eggs in relation to water proper- ties and movement in Shelikof Strait, Gulf of Alaska. In R. J. Beamish and G. A. McFarlane (eds.), Effects of ocean variability on recruitment and evaluation of parameters used in stock assess- ment models, p. 169-180. Can. Spec. Pub. Fish. Aquat. Sci. 108. 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, in Sheli- kof Strait, Gulf of Alaska. Fish. Bull. 85:499-521. Kerfoot, W. C. 1985. Adaptive value of vertical migration: com- ments on the predation hypothesis and some alternatives. Contrib. Mar. Sci. 27:91-113. Lampert, W. H. 1989. The adaptive significance of diel vertical migration of zooplankton. Funct. Ecol. 3:21-27. Laprise, R., and J. J. Dodson. 1993. Nature of environmental variability experi- enced by benthic and pelagic animals in the St. Lawrence Estuary, Canada. Mar. Ecol. Prog. Ser. 94:129-139. Lenarz, W. H., R. J. Larson, and S. Ralston. 1991. Depth distributions of late larvae and pelagic juveniles of some fishes of the California Current. Calif. Coop. Oceanic Fish. Invest. Rep. 32:41-46. Levy, D. A. 1990. Reciprocal diel vertical migration behavior in planktivores and zooplankton in British Columbia lakes. Can. J. Fish. Aquat. Sci. 47:1755-1764. MacKenzie, B. R., W. C. Leggett, and R. H. Peters. 1990. Estimating larval fish ingestion rates: can laboratory derived values be reliably extrapolated to the field? Mar. Ecol. Prog. Ser. 67:209-225. Brodeur and Rugen: Vertical distribution of ichthyoplankton in the northern Gulf of Alaska 235 Matarese, A. C, A. W. Kendall Jr., D. M. Blood, and B. M. Vinter. 1989. Laboratory guide to early life history stages of Northeast Pacific fishes. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 80, 652 p. Neill, W. E. 1990. Induced vertical migration in copepods as a defence against invertebrate predation. Nature 345:524-526. Neilson, J. D., and R. I. Perry. 1990. Diel vertical migrations of marine fishes: an obligate or facultative process? Adv. Mar. Biol. 26:115-168. Ohman, M. D. 1990. The demographic benefits of diel vertical migra- tion by zooplankton. Ecol. Monogr. 60:257-281. Ohman, M. D., B. W. Frost, and E. B. Cohen. 1983. Reverse diel vertical migration: an escape from invertebrate predators. Science 220:1404- 1407. OUa, B. L., and M. W. Davis. 1990. Effects of physical factors on the vertical dis- tribution of larval walleye pollock Theragra chalco- gramma under controlled laboratory conditions. Mar. Ecol. Prog. Ser. 63:105-112. Paul, A. J. 1983. Light, temperature, nauplii concentrations, and prey capture by first feeding pollock larvae Theragra chalcogramma. Mar. Ecol. Prog. Ser. 13:175-179. Pearre, S. 1979. Problems in the detection and interpretation of vertical migration. J. Plankton Res. 1:29^14. Pritchett, M., and L. Haldorson. 1989. Depth distribution and vertical migration of larval walleye pollock (Theragra chalcogramma). In Proceedings of the international symposium on the biology and management of walleye pollock, November 14-16, 1988, Anchorage, Alaska, p. 173- 183. Alaska Sea Grant Rep. 89-1, Univ. Alaska, Fairbanks. Theilacker, G. H. 1980. Changes in body measurements of larval northern anchovy, Engraulis mordax, and other fishes due to handling and preservation. Fish. Bull. 78:685-692. Wiebe, P. D., K. H. Burt, S. H. Boyd, and A. W. Morton. 1976. A multiple opening/closing net and environ- mental sensing system for zooplankton. J. Mar. Res. 34:313-326. Yamashita, Y., D. Kitagawa, and T. Aoyama. 1985. Diel vertical migration and feeding rhythm of the larvae of the Japanese sand-eel Ammodytes personatus. Bull. Japan. Soc. Sci. Fish. 51:1-5. Abstract. — The commercial fishery for orange roughy on the Challenger Plateau developed in 1981, increased markedly through- out the mid-1980s, and then de- clined rapidly by 1990. Data from research trawl surveys and com- mercial fishing returns over the period are examined, and changes in the population are described. The distribution of orange roughy changed over the period examined; there was a contraction of the areas of high density and apparent fishing-out of aggrega- tions on relatively fiat bottom. Ag- gregations are now largely con- fined to pinnacles. Biomass of or- ange roughy, measured by bottom trawl survey indices and commer- cial catch per unit of effort, de- clined substantially and is cur- rently estimated to be about 20% of virgin levels. Most other inci- dental species in the trawl surveys have also declined in abundance, and there are no indications of 'species replacement.' Data on size, reproductive stage, size at maturity, and feeding have also been examined. Size structure of the population has not changed over time. Time of spawning (July) and the pattern of gonad develop- ment have been consistent over the years. Diet composition has also remained similar; dominant prey groups are natant decapod crustaceans and small fish. It is suggested that biological changes have not been apparent because orange roughy are a long- lived, slow-growing species, with low productivity. There could be a long response time to fishing pres- sure, yet orange roughy popula- tions can be quickly reduced to low levels by commercial fishing. Changes in a population of orange roughy, Hoplostethus atlanticus, with commercial exploitation on the Challenger Plateau, New Zealand Malcolm R. Clark Dianne M. Tracey MAF Fisheries Greta Point. PO. Box 297 295 Evans Bay Parade, Wellington, New Zealand Manuscript accepted 4 November L993 Fishery Bulletin 92:236-253 (1994) Orange roughy (Hoplostethus atlanticus Collett) has a worldwide distribution on the continental slope at depths of 700 to 1,500 m. However, it is fished commercially only off New Zealand, Australia, and in the northeastern Atlantic Ocean. The New Zealand fishery is the most established, having started in 1978; the others date from 1988 and 1991, respectively. Orange roughy is one of the most valuable commercial species in New Zealand waters, with annual landings of 40-50,000 metric tons (t) and export earnings of NZ $100- 150 million (Robertson, 1991). The New Zealand fishery for or- ange roughy occurs in a number of areas (Fig. 1), including the Chal- lenger Plateau, a broad submarine plateau off the west coast of New Zealand. The commercial fishery on the Plateau developed in late 1981 and rapidly expanded into one of the most important orange roughy fisheries in New Zealand waters, with annual catches up to approxi- mately 16,000 t (Table 1). The fish- ery operates primarily during win- ter (June-August), when the fish form large spawning aggregations at depths of 850-900 m (Clark, 1991a). The fishery has been managed by a Total Allowable Catch (TAC) sys- tem since 1982. Tracey et al. (1990) and Clark (1991a) discussed details of this management regime. Ini- tially, catches were limited to 7,000 t by the TAC for all areas of the New Zealand Exclusive Economic Zone (EEZ), outside the established fishing grounds on the east coast. A TAC of 4,950 t was set for 1983- 84 and 1984-85 (October-Septem- ber fishing year) on the Challenger Plateau and west coast of the South Island. This was raised to 6,190 t specifically for the Chal- lenger fishery in 1985-86 based on biomass estimates from a trawl survey in winter 1984. The quota was raised to 10,000 t in 1986-87, and further to 12,000 t in 1987-88, in order to assess the effects of heavier fishing on the population dynamics of orange roughy ("adap- tive management"). In the follow- ing fishing year only 8,200 t of quota were allocated, the rest with- held because of signs that the or- ange roughy population was declin- ing rapidly. During 1989-90 the TAC was reduced to 2,500 t after new stock assessments showed the population was overexploited and had declined to low levels (Clark and Francis, 1990). The TAC was further reduced for 1990-91 to pro- mote rebuilding of the population. These changes occurred against a background of increasing infor- 236 Clark and Tracey: Population changes of Hoplostethus atlanticus on the Challenger Plateau 237 /175°W * EEZ boundary Figure 1 Map of New Zealand and offshore waters of the Exclusive Economic Zone (EEZ), showing the location of the Challenger Plateau and other major fishing areas for orange roughy (Hoplostethus atlanticus). Table 1 Reported catch es (t) of orange roughy (Hop- lostethus atlanti cus) from the Challenger Plateau (ORH 7A and outside EEZ) (from Clark 1992; to- tal estimated catch includes allowance for re- search survey catches an d a correction for 15-30% under-estimation of true catch in reported catch figures because of bu rst trawls. fish dis- cards, and incorrect official conversion factor). Total Total Fishing year reported estimated (Oct-Sept) catch catch TAC 1980/81 33 43 1981/82 4,248 5,522 1982/83 11,839 15,409 1983/84 9,527 12,514 4,950 1984/85 5,117 6,707 4,950 1985/86 7,753 10,251 6,190 1986/87 11,492 15,750 10,000 1987/88 12,181 15,830 12,000 1988/89 10,241 12,627 12,000' 1989/90 4,309 5,171 2,500 1990/91 1,357 1,560 1,900 ' 8,219 t allocated. mation on orange roughy. It has only recently been realized that orange roughy is very slow-growing and long-lived. Mace et al. (1990) recorded a growth rate of about three cm per year for the first four years of life (validated ages), and an estimated age at maturity of 24 years and maximum age over 50 years. They estimated natural mor- tality to be low (less than 0-1 yr-1) and concluded that sustainable yields of orange roughy would be relatively low and show slow recov- ery from over-fishing. Recent esti- mates of the maximum age of or- ange roughy from Australian wa- ters approach 150 years (Fenton et al., 1991). Quotas for orange roughy har- vest from New Zealand have been reduced in recent years on the ba- sis of information which suggests much lower productivity than origi- nally assumed. However, the Chal- lenger Plateau population had al- ready declined markedly and pro- vides some insight into the effects of heavy fishing pressure on orange roughy population dynamics. There is an extensive literature on general re- sponses offish populations to exploitation, covering lake ecosystems (e.g. Regier and Loftus, 1972; Spangler et al., 1977), coral reef fisheries (e.g. Russ and Alcala, 1989), and relatively shallow-water marine environments (e.g. Hempel, 1978; Pauly, 1979; Grosslein et al. 1980). There have been few studies on deep-water or long-lived species such as orange roughy. The closest is probably Pacific ocean perch (Sebastes alutus) which is found at depths to 600 m in the North Pacific Ocean and has a maxi- mum age of 90 years (e.g. Gunderson, 1977; Lea- man, 1991). There are a number of general population re- sponses to exploitation, which include 1 Decline in abundance of fished species. 2 Contraction of distribution or areas of high density. 3 Change in age structure or size structure, or both, with fewer old, large fish and the population dominated by new recruits. 4 Increase in growth rate of individuals, with a decrease in age for a given length. 5 Lower age at maturity or size at maturity, or both. 238 Fishery Bulletin 92(2), 1994 6 Possible change in species composition over time ('species replacement'). Such responses are often observed in short-lived, fast-growing species (e.g. Pauly, 1979; Grosslein et al., 1980). Some have also been noted with Sebastes alutus (Gunderson, 1977; Leaman, 1991), but it is not clear whether these changes would occur in such a long-lived species as orange roughy, or over what time period such changes would become apparent. Orange roughy on the Challenger Plateau have been exploited for only 10 years, and hence it seems un- likely that marked changes in biological character- istics could occur over such a relatively short time period in relation to the longevity of the species. Therefore, we might expect to observe changes in biomass and distribution, as well as age and size struc- ture of the population, but not changes in growth rate or reproductive potential. In this paper, we summarize some of the available data on distribution, abundance, and biology of or- ange roughy on the Challenger Plateau, primarily over the period 1984-90. This period covers the early years of the developing fishery, to maximum levels of exploitation, and subsequent decline of the population. We describe the reduction in size and distribution of the stock and investigate associated changes in size structure, aspects of reproduction, and feeding. Methods Research trawl surveys Trawl surveys have been carried out in the winter (June-July) of each year from 1984 to 1990. The vessel used, area covered, intensity of trawling, and survey design differed between years, and all are not directly comparable (Table 2). Surveys from 1987 to 1989 were treated as fully comparable, but only se- lected data have been used from other surveys: distribution from 1984 and 1990, and biology (size, reproductive, and feeding data) from 1984, 1985, 1986, and 1990. The general survey design was two-phase stratified random (af- ter Francis, 1984). The survey area was divided into a number of strata based on depth and certain bottom features (e.g. pinnacles). General stratification is shown in Figure 2. The depth range covered was 800 to 1,200 m. New, random station positions within strata were selected each year, except in strata 10 and 11 on pinnacles where a random tow direction was ad- justed to avoid untrawlable ground, and these trawls were repeated each year. A similar net de- sign and gear set up was used for each survey. Tow length was standardized where possible at 1.5 nau- tical miles (nmi). Trawling speed was 3.0-3.5 knots. Biomass indices were calculated by the area swept method as described by Francis ( 1981). Biomass and its standard error were calculated from the follow- ing formulae: and B=^(X,a,)/cb SB = J^sfaf/c2b2 where B is biomass (t), X is the mean catch rate (kg-km-1) in stratum i, ai is the area of stratum i (km2), b is the width swept by the gear (defined as doorspread (m) by MAF Fisheries), c is the catch- ability coefficient (an estimate of the proportion of fish available to be caught by the net), SB is the standard error of the biomass, s( is the standard error of Xt The catchability coefficient was assigned a value of 0.27, which represents the wingend spread di- vided by the doorspread, because orange roughy form schools which are not believed to be herded substantially by doors or sweeps.1 Approximate 95% confidence limits (CD were calculated as CL = B±2SB. 1 Orange Roughy Working Group, MAF Fisheries, Greta Point, P.O. Box 297, Wellington, New Zealand, pers. commun. 1991. Table 2 Trawl surveys carried out on the Cha lenger Plateau for orange roughy (Hoplostethus a 'In nticus). Area Number Vessel Year Date (km2) of trawls Survey design Arrow 1984 3/7-18/7 11,956 118 2 phase SRTS' Arrow 1985 4/7-20/7 209 16 1 phase SRTS Arrow 1986 4/7-17/7 94 10 transect grid Amaltal Explorer 1987 18/6-13/7 8,270 129 2 phase SRTS Amaltal Explorer 1988 4/7-24/7 8,270 85 2 phase SRTS Amaltal Explorer 1989 8/7-30/7 8,270 160 2 phase SRTS Will Watch 1990 7/7-29/7 8,270 141 2 phase SRTS ' Stratified random tra wl survey Clark and Tracey: Population changes of Hoplostethus atlanticus on the Challenger Plateau 239 EEZ boundary Figure 2 The Challenger Plateau survey area, showing bathymetry (depth contour in m) and details of survey stratification. ment of the fishery through to maximum exploitation. It is felt that the fishery was not constrained much by the TAC over this time. CPUE in winter months from 1983 to 1991. This in- cludes data from 1990 and 1991, following a substantial reduction in TAC and effort. CPUE in non-winter months from 1983 to 1991. Trawl survey indices from 1987 to 1989. These sur- veys covered the same area, had the same design, and used the same vessel. Trawl survey indices from 1984 and 1987 to 1990. This series incorporated data from a smaller area surveyed in 1984 and from the 1990 survey, both of which used a different ves- sel from 1987 to 89. The coefficient of variation (CV) is a measure of the precision of the biomass estimate, and was cal- culated by CV = 5B/Bxl00. Stock reduction analysis A stock reduction technique was used to estimate virgin biomass based on the method of Francis (1990, 1992). This incorporated a complete catch history for the stock, a time series of abundance indices, and life history parameters used in a deter- ministic age-structured population model (see Clark, 1992). The latter were the von Bertalanffy growth parameters (Lm=39.5 cm, &=0.059yr_1, t0 = -0.3 yr), natural mortality=0.04yr_1, weight-length parameters (a=0.0963, 6=2.68), age at maturity (24 yr), age at entry to the fishery (24 yr), and Beverton- Holt recruitment steepness of 0.75. Five sets of abundance indices were used from trawl surveys between 1984 and 1990, and commer- cial catch per unit of effort (CPUE) data (unstand- ardized mean catch per tow by monthly groupings): 1 CPUE in winter months (June-September) from 1983 to 1989. This covered the period of develop- The maximum likelihood method was used to estimate virgin biomass. Ninety-five percent confidence intervals were estimated by using bootstrapping techniques with the coefficient of variation fixed at 20%. The best estimate of virgin biomass was then used in an age-structured model (detailed in Francis, 1992) to es- timate current biomass. Biological data Standard procedure during trawl surveys was to take a random sample of about 200 fish from each tow. These were measured (standard length rounded down to the nearest whole cm [standard MAF Fish- eries procedure]) and sexed. Twenty of these fish were randomly selected, their otoliths extracted, and more detailed data collected: standard length (rounded down to the nearest whole mm), weight (rounded down to the nearest gm), sex, stage of gonad maturity (see below), gonad weight (rounded down to the nearest gm), fullness of stomach, state of digestion of contents, and stomach contents (to species level where possible). Size Length-frequency distributions have been con- structed to represent the total population where possible. In the years 1984 and 1987-90, data have 240 Fishery Bulletin 92(2). 1994 been scaled by percentage sampled to represent each catch and further scaled by stratum biomass to ap- proximate the population. Samples in 1985 and 1986 were scaled to represent solely the catch, as survey design was inadequate for biomass estimation. Length-frequency data are difficult to compare statistically and, for the purposes of this study, have not been attempted. However, to enable a general comparison, a single distribution was constructed combining length-frequency data from all years weighted by the number of tows each year. This distribution is plotted together with those from each year separately. Mean size by sex was calculated separately for three main regions of spawning within the survey area (strata 1, 4; 10; 9, 11) as it was unlikely these areas had been fished equally (see later 'Commer- cial Fishery' section). The sample sizes used in cal- culating the standard error were number of tows, not number of fish. Orange roughy can associate in size groups; between-tow variance was greater than within-'tow variance. Variance is represented by ±2.0 standard errors for all years except 1986, when ±2.2 standard errors was arbitrarily used because there were only 10 trawls. Reproduction Macroscopic staging of reproductive condition followed Pankhurst et al. (1987): Stage Female Male 1 2 3 Immature/resting Early maturation Maturation Immature/resting Early maturation Maturation 5 6 Ripe Running ripe Spent Ripe/running ripe Spent Relative frequency of gonad stages was examined. Analyses were based on the samples taken. They were not scaled in any way, as there were no appar- ent differences between the length frequencies of the samples and the distribution of the total population. Only data from females are presented, as their macroscopic gonad stages can be determined more accurately than those from males. Size at maturity was determined from samples taken over the total survey area using a 'probit analysis' approach (after Pearson and Hartley, 1976). It was assumed that length at maturity is normally distributed in the population. The regres- sion part of the analysis was repeated 10 times to ensure convergence of the estimate.2 A standard lin- 2 Francis, C, MAF Fisheries, pers. commun. 1991 ear regression analysis was carried out on results to investigate trends over time by using the SAS sta- tistical package (SAS, 1988). Feeding Data on frequency of occurrence were available from all surveys. Frequency of occurrence was defined as the number of stomachs in which a food item occurs, expressed as a proportion of the total number of stomachs containing food. Only stomachs with part-full or full classifications, and with fresh or partly digested contents, were included in analyses. Commercial fishing data Data on the catch and position of each tow and the start and finish times have been collected since 1980. However, catch and effort information is dif- ficult to standardize and interpret for orange roughy. Fish can be highly aggregated at various times of the year, and 'windows' or escape panels in the net are frequently used to reduce catch size and mini- mise damage to nets. Fishing performance varies with experience of skipper and crew, and technol- ogy has advanced considerably in recent years (in particular, development of Global Positioning Sys- tem navigation, which enabled improved accuracy when fishing pinnacles). Fishing logbooks often do not have accurate information on length of tow on the bottom. Fishing for orange roughy on the Chal- lenger Plateau occurs on a variety of bottom terrain: on flat bottom, in troughs and steep slope, and on the tops and sides of pinnacles. In each case, the effective fishing time and fishing technique differ greatly, and they are almost separate types of fish- eries. In order to gain an indication of trends in catch rates, data were examined on the basis of mean catch per tow for two size classes of vessel (20— 60 m, generally domestic fresh fish boats; and 60- 90 m, domestic factory trawlers). Catch per unit of effort (CPUE) values were similar for both classes, and so data here are combined. Monthly data were amalgamated into two time periods: first, 'winter' (June, July, August) which covers the spawning period; second, 'out of season' (all other months). This division represents two distinct phases of or- ange roughy distribution, as well as differences in the mode of fishing (Clark, 1992). The former period is characterized by the formation of relatively stable, dense aggregations of fish, whereas in the latter period the orange roughy are more dispersed and widely distributed (Clark, 1991a). Fishing in win- ter generally involves shorter tows, often with smaller nets, than does out-of-season fishing. Clark and Tracey: Population changes of Hoplostethus atlanticus on the Challenger Plateau 24 1 In the following text, three colloquial area names have been used. These are given below with specific strata numbers (see Fig. 2): Central Flat Pinnacles Westpac Bank strata 1, 4 stratum 10 strata 9, 11 Results Distribution Trawl surveys The distribution of orange roughy in the survey area changed substantially between years (Fig. 3). In 1984 high catch rates were ob- served across much of the Central Flat area. (No trawls were made on the Pinnacles although heavy marks were observed on the echosounder; the sur- vey did not cover the Westpac Bank area.) In 1987 fish were still widely distributed in the Central Flat; -36' 1984 >48' •40°S Catch rate kg.knrr1 100 j-s^ii -12' ^P 10 000 12' 24' 36' 48' 168° E 12' 24' -36' 1987 >36' 1988 -48

■48'----.. [®j -40°S -40°S r-'iS -12' ■ 12' 12' 24' 36' 48' 168° E 12' 24' 1 2' 24' 36' 48' 168° E 12' 24' 36' 1989 - 3IV 1990 ■48';<3; .---. - € ''--. S; ■40°S r-<5/ ■40°S ■0 ■12' - 12' 1 2' 24' 36' 48' 168° E 12' 24' 12' 24' 36' 48' 168° E 12' 24' Figure 3 Contours of trawl survey catch rates (kg-km l) of orange roughy (Hoplostethus atlanticus) in 1984 and 1987-90. there were two main schools and further concentra- tions around the Pinnacles and the Westpac Bank. In 1988 there was a marked contraction in the area of high catch rates; a single small aggregation was observed on the Central Flat, and by 1989 there were no aggregations in the Central Flat region. High catch rates still occurred on the Pinnacles and Westpac Bank in 1989, and these actually increased in 1990, after the TAC and fishing effort were greatly reduced. Commercial fishery The commercial fishery has been centered mainly inside the EEZ, targeting ag- gregations of orange roughy on the Central Flat and Pinnacles. Distribution of effort (number of tows) and catch between these two areas has changed over time (Table 3). In the period 1982-87, over 80% of the catch from the two areas was taken from the Central Flat with over 75% of the number of tows. In 1988 there was a marked increase in the propor- tion of catch and effort on the Pinnacles, and a corre- sponding reduction on the Cen- tral Flat. This shift continued in 1989 and 1990, during which the Pinnacles accounted for 65-70% of the catch. These changes reflect the change in distribution observed in the re- search trawl surveys. Relative abundance Trawl surveys Biomass indi- ces (estimates of relative bio- mass) from trawl surveys in 1987, 1988, and 1989 are given in Table 4. The indices indicate a marked decline in biomass over the period. The distribution of biomass among strata changed over the years 1987-90 (Table 5). In 1987 and 1988 over 60% of the bio- mass was in the Central Flat area, but only 30% in 1989 and 1990. Over this period, there was an increase in the proportion on the Pinnacles, especially between 1989 and 1990. Biomass levels in the surrounding areas have fluc- tuated but were particularly high in 1989. The proportion of biomass on the Westpac Bank has remained compara- tively constant. 242 Fishery Bulletin 92(2), 1994 Commercial fishery Mean catch per tow for all New Zealand vessels in the fishery from 1983 to 1991 is given in Table 6. Catch rates in winter, when the fish are aggregated for spawning, are generally higher than in other months. Although aggregations occur at other times, presumably for feeding, they are not as large or as stable as in winter. Catch rates in both periods declined steadily from 1983 to 1989 to between about 15% and 20% of original levels. The trend is slightly different in the two periods; winter catch rates declined more sharply to 1988, whereas in the other months the largest decrease was between 1983 and 1984. Catch rates increased in 1990, following a reduction of the TAC, when there were less vessels and fewer trawls on the grounds. Individual trawl catch rates for orange roughy can be highly variable, consisting of 'hits' and 'misses.' Therefore it is not useful to describe the variance around these mean catch rates, beyond commenting that there is wide variation. It should be stressed that the changes in catch rates presented here may give an indication of changes in stock size but should be treated with caution. Difficulties in interpreta- tion of such data for orange roughy are described in the 'Methods' section, and the form of relationship between mean catch per tow and stock abundance is uncertain. Stock reduction results Abundance indices used in, and estimates of virgin biomass from, the stock reduction analyses are given in Table 7. Point estimates of BQ range from 95,000 t to 278,000 t. The best fits of data to the model (those with the lowest CV) are from the winter CPUE se- ries. Results from trawl survey data have higher CVs but confirm that an estimate of the order of 100,000 t is reasonable. The 1987-89 trawl survey series gave the lowest virgin biomass estimate. It was not considered reli- able because there were only three indices, and high Table 3 Distribution of commercial catch (% of total catch taken tn the two areas) and effort (% of number of tows) for orange roughy (Hoplostethus atlan- ticus) for the Central Flat and Pinnacles ( winter period June to August). Year Central Flat Pinnae les % catch % tows % catch % tows 1982 97.2 95.6 2.8 4.4 1983 97.0 94.7 3.0 5.3 1984 95.2 93.6 4.8 6.4 1985 87.7 78.0 12.3 22.0 1986 87.3 83.2 12.7 16.8 1987 84.4 77.3 15.6 22.7 1988 52.9 56.0 47.1 44.0 1989 34.0 43.3 66.0 56.7 1990 30.2 45.0 69.8 55.0 Biomass indices tethus atlanticus) from 1987 to 1989 Table 4 (t) of orange roughy (Hoplos- from trawl surveys, conducted . (CV = coefficient of variation.) Year Biomass (t) CV 1987 1988 1989 78,661 30,946 11,746 26 27 11 fishing mortality rates were required to support the catch history. A maximum F of 1.0 is regarded as realistic for orange roughy (Francis et al., 1992). This constrains the virgin biomass to a minimum value of 94,000 t. The estimate from non-winter CPUE is compara- tively high. It has a large CV and is based on rela- tively low numbers of trawls (because most fishing effort is in winter). Such a biomass level would also Comparison of biomass estimates (t) o] 'orange Table 5 roughy (Hopl ostethus atlanticus) by region from 1987 to 1990. 1987 1988 1989 1990 Region Biomass (t) '; Biomass (t) ', Biomass (tl ', Biomass 0.85 mm) based on the morphological criteria of Joseph et al. ( 1964). Rear- ing chambers were returned to the laboratory and held for 3 to 14 days. In these, larvae were periodically sac- 256 Fishery Bulletin 92(2). 1994 rificed and preserved in 5-8% buffered formalin. Iden- tifications of preserved sciaenid larvae from pigment characters were based on Ditty (1989). Sciaenid eggs collected in the same area during spring 1991, 1992, and 1993 were sorted from fresh plankton samples. To avoid contamination by the morphologically similar eggs of the cynoglossid Symphurus plagiusa and the soleid Trinectes macula- tus that contain several oil globules and are abundant in lower Chesapeake Bay during the spring, all eggs with >3 oil globules were omitted from the samples. Although eggs of most spring-spawning sciaenids gen- erally possess three or fewer oil globules (usually two) those of Menticirrhus saxatilis may contain from 1 to 16 oil globules (Johnson, 1978). After sorting, eggs were measured, placed in scintillation vials with 26 ppt seawater, and frozen at -70°C for genetic analy- sis. Individual eggs were thawed and remeasured prior to homogenization to assess shrinkage. Sciaenid eggs were genetically typed by compar- ing mtDNA restriction fragment patterns of indi- vidual eggs with those of known adults. To obtain patterns of known adults, mature female sciaenids (B.chrysoura, C. nebulosus, C. regalis, M. saxatilis, and P. cromis) were collected by pound net, trawls, and hook and line in April and May 1990 and 1991. Ovarian tissue was excised and frozen at -70°C. MtDNA was purified from ovarian tissue by cesium chloride equilibrium density gradient ultracentrifu- gation following the protocols of Lansman et al. (1981). To determine a restriction enzyme that un- ambiguously identified the different sciaenid spe- cies, aliquots of mtDNA were individually digested with the following restriction enzymes: Apal, Aval, Banl, Banll, Hindlll used according to manu- facturer's instructions. The resulting fragments were separated electrophoretically on 1.0% agarose mini-gels run at 5 V/cm for four hours and visual- ized with ethidium bromide. MtDNA-enriched genomic DNA was isolated from individual eggs following the protocols of Graves et al. (1990). Entire DNA samples were digested with a single discriminating restriction endonuclease, separated electrophoretically, and transferred to a nylon filter (Southern transfer) following standard protocols (Sambrook et al., 1989). Filters were hy- bridized with highly purified black drum mtDNA, nick-translated with biotin-7-dATP, washed, blocked and visualized following the methods of Graves et al. (1990). Results A total of 10,803 sciaenid eggs was sorted from samples collected in 1990 and 1991. Outside egg diameter of all specimens ranged from 0.650 to 1.12 mm. Successive blind readings of samples of 75 to 100 eggs were used to assess measurement error. No differences were found in the size-frequency distri- butions indicating good agreement within the 0.025- mm size classes (two-sample <-test, P<0.05, n=79). Qualitative analysis of culture experiments using the two egg types of Joseph et al. (1964) revealed the presence of three species. Cultures containing eggs designated Type I (<0.80 mm) resulted in lar- vae of B. chrysoura, whereas cultures of eggs desig- nated Type II (>0.85 mm) resulted in larvae of C. regalis and P. cromis. Analysis of preserved ichthyoplankton samples from 1990 and 1991 revealed the presence of larvae of B. chrysoura, C. regalis, and P. cromis. No early life history stages of other sciaenids were identified; however, yolk-sac larvae could not be identified to species. Because rearing studies and analysis of field-caught plankton samples revealed the presence of more than two species, we could not rely on the criteria of Joseph et al. (1964) for specific identifi- cation. We therefore examined weekly frequency of occurrence of all sciaenid eggs during 1990 (Fig. 1) and 1991 (Fig. 2). Based on temporal occurrence and size frequency we identified three modes. The larg- est eggs (>0.975 mm), Type C, were most abundant during the period 23 April through 9 May. Type-C eggs declined in abundance throughout May in both years. Mid-sized eggs (0.850-0.950 mm), designated Type B, generally appeared later than Types A and C. Type-B eggs did not exceed 5% of the total fre- quency of sciaenid eggs until 15 May 1990 and 9 May 1991. Type-B eggs increased in abundance from mid-May until the end of sampling. The smallest eggs (<0.850 mm), designated Type A, co-occurred with Type-C eggs; however, they did not exceed 5% of the total sciaenid eggs until 8 May 1990 and 9 May 1991. In 1990, Type-A eggs peaked in abun- dance on 15 May and gradually declined through- out the sampling period. In 1991, Type-A eggs were most abundant during the last sample on 28 May. To test the hypothesis that eggs designated Types A, B, and C were separate species assemblages, the mtDNA restriction fragment patterns of known adult sciaenids were compared with those of fresh egg samples separated into Types A, B, and C. Re- striction fragment length polymorphism analysis of mtDNA, purified from adult B. chrysoura, C. nebulosus, C. regalis, Menticirrhus saxatilis, and P. cromis, revealed species-specific restriction fragment patterns for each of the five enzymes. Of the five en- zymes, Hindlll showed the greatest differences be- tween species, facilitating visualization with the Southern blotting procedure (Table 2). Daniel and Graves: Morphometry and genetic identification of sciaenid eggs 257 > o z LU o LU cc LU o a: LU a. 23 April 1990 60 15 May 1990 20 I I Type A ■ Type B M Type C 25 20 15 10 0.65 0.75 0 85 0 95 1 05 115 -W III 0.65 0.75 085 0.95 1.05 1.15 1 May 1990 24 May 1990 30 20 -, f , p , Jl fi ■■■■■ , , , i \ \ * -B-l -B — 10 0.65 0 75 0 85 0.95 8 May 1990 IHI, II p II p . ,l,i, J.llJp , 0.65 0 75 0 85 0 95 1 05 1 15 ?0 10 n „n ■ ■lit] r ll 0 65 0 75 0 85 0 95 1.05 1.15 31 May 1990 .-JL JL Jl f" 0 J_ WU 0.65 0.75 0.85 0 95 105 1 15 OUTSIDE EGG DIAMETER (mm) Figure 1 Frequency distributions of outside egg diameters of sciaenid eggs collected over six weeks during spring 1990 in lower Chesapeake Bay. mm and larger (n=32), all pos- sessed the restriction fragment pattern diagnostic for P. cromis. Discussion A total of 62 eggs, representing all sciaenid egg size classes collected in lower Chesapeake Bay, was identified with diagnostic Hi n dill restriction frag- ment patterns. Bairdiella chrysoura, C. regalis, and P. cromis were the only species of sciaenids identi- fied; no other restriction fragment patterns were observed. Genetic identification of eggs designated Type A (<0.850 mm, n=12) resulted in 11 individu- als of B. chrysoura and one specimen (0.825-mm OED size class) of C. regalis (Fig. 3). Cynoscion regalis composed the majority of type-B eggs (0.850- 0.975 mm, rc = 18) analyzed, but seven of the 10 larg- est type-B eggs (0.975-mm OED size class) were identified as black drum. Type-C eggs, those 1.00 Identifications of eggs of sci- aenids are often based on pub- lished diameter distributions or hatching experiments, or both. Results of hatching experiments and genetic analysis in this study indicate that samples of eggs of a single size class may represent the products of two or more species. For example, eggs designated Type I (<0.80 mm) and identified as silver perch by Joseph et al. (1964) were shown with genetic analysis to contain eggs of both weakfish and silver perch. Similarly, eggs designated Type II (>0.85 mm) and identi- fied as black drum by Joseph et al. (1964) were shown with rear- ing and genetic analysis to con- tain eggs of both weakfish and black drum. During the present study, nei- ther hatching experiments nor genetic analysis identified eggs as black drum that were smaller than 0.975 mm OED. While tem- porally limited, the results of this study suggest that the range in size for eggs of black drum (0.975-1.125 mm) in lower Chesapeake Bay may be more restricted than those previously reported. The ranges of egg diameter overlapped for silver perch and weakfish. Eggs genetically identified as silver perch ranged in size from 0.650 to 0.825 mm, in agreement with previously reported size ranges for silver perch in the northwestern Gulf of Mexico (0.59-0.82 mm, Holt et al., 1988) and Chesapeake Bay (0.625-0.775 mm, Joseph et al., 1964). Although Holt et al. (1988) identified eggs of silver perch as small as 0.590 mm, no sciaenid eggs smaller than 0.650 mm OED were collected in the present study. Sizes of eggs genetically identified as weakfish were found to range from 0.825 to 0.975 mm in diameter. These values are comparable with those reported by Wisner ( 1965, 0.84-0.96 mm) but are narrower than 258 Fishery Bulletin 92(2), 1994 30 20 10 I I Type A ■ TypeB W TypeC o z LU o cc LU o a: LU Q. 20 15 the range (0.68-1.18 mm) given by Merriman and Sclar (1952) for Block Island Sound, New York. While the range in sizes for silver perch and weakfish re- ported in this study agree with past research, overlaps in these ranges preclude the sole use of egg size for identification. Neither Joseph et al. (1964), Olney (1983), nor the present study identified eggs of C. nebu- losus or M. saxatilis in samples collected in lower Chesapeake Bay. Fable et al. (1978) described laboratory-spawned eggs of C. nebulosus from a single female and reported a mean diameter of 0.77 mm (range 0.70-0.85 mm). Although based upon a limited sample size, Fable et al.'s data indicate that eggs of C. nebu- losus could be confused with eggs of B. chrysoura; however, no eggs in our limited sample of this size range (n=12) were ge- netically identified as C. nebu- losus. A possible explanation for the lack of eggs of C. nebulosus in the present study may be the tendency for adults to spawn in or around vegetated areas (Brown, 1981). The absence of eggs of Menticirrhus spp. in this genetic analysis may be ex- plained by our exclusion of eggs with greater than three oil glob- ules. Additionally, Menticirrhus saxatilis reportedly spawns off front beaches and possibly off- shore (deSylva et al., 1962); con- sequently, circulation in the bay may prevent eggs of this species from entering the survey area or they may be transported to areas that were not sampled in our study. The identification of species-specific restriction fragment patterns for spring-spawning sciaenids is based on the assumption that there is limited in- traspecific variation of the diagnostic restriction fragment patterns. Recent studies of the population genetics of spotted seatrout, black drum, and weak- fish (Graves et al., 1992; Gold et al., 1993) indicate that these species exhibit low intraspecific mtDNA variability. Furthermore, no variation of the Hi n dill fragment pattern was found in a survey of mtDNA 23 APR 1991 19 MAY 1991 20 10 I I n# II n I, IWr- 0 65 0.75 0.85 0.95 1.05 1.15 29 APR 1991 0.65 0 75 085 0 95 1.05 1.15 22 MAY 1991 30 25 10 JjlLl. I }l, .1^1.. 0 65 0 75 0 85 0 95 1 05 1 15 9 May 1991 065 0.75 085 0.95 1 05 1.15 28 MAY 1991 n l || , PP Ulfl.U, ll I -I 20 15 I I p I :, ,<}■,', =m^h a- 0 65 0 75 0 85 0 95 1.05 1.15 065 0.75 0.85 0.95 1.05 1.15 OUTSIDE EGG DIAMETER (mm) Figure 2 Frequency distributions of outside egg diameters of sciaenid eggs collected over six weeks during spring 1991 in lower Chesapeake Bay. isolated from 25 adult B. chrysoura (L. Daniel, unpubl. data). Consequently, the common restriction fragment patterns used to distinguish species in this study were deemed suitable for use in identifications. Variability in egg-size distributions with changing salinity and over the spawning season were not exam- ined in this study. Consequently, exact size groupings may only be applicable to the particular salinity re- gime (19-25 ppt) that we sampled. However, samples were taken throughout peak spawning for black drum and silver perch and may encompass the ranges that occur for these species in lower Chesapeake Bay. Results of our genetic analysis suggest that iden- tifications of eggs of spring-spawning Sciaenidae in Daniel and Graves: Morphometry and genetic identification of sciaenid eggs 259 Table 2 Common fragment sizes produced by restriction endonuclease {Hindlll) digestion of mtDNA purified from ovarian tissue of spring spawning sciaenids. Species Fragment sizes (Kbi Bairdiella chrysoura 5.0 Cynoscion nebulosus 8.5 Cynoscion regalis 5.6 Menticirrhus saxatilis 5.4 Pogonias cromis 3.3 3.9 4.5 4.3 3 2 2.9 2.8 1.9 3.8' 4.1 2.9 2.4 2.0 2.7 2.5 1.7 1.9 2.1 1.7 1.3 1.8 1.3 1.0 ' J. Gold, Texas A&M, College Station, TX, pers. commun. 1993. 07 0.7250750775 08 08250.850875 09 09250950975 1 1025105107511 1125 Outside Egg Diameter (mm) B. chrysoura | | C. regalis P. cromis Figure 3 Size distributions of all eggs morphologically typed as sciaenids and identi- fied using genetic techniques. wise, measures of spawning stock biomass will be similarly over-es- timated, results that could signifi- cantly impact management deci- sions. Comparable biases in esti- mates of egg production and spawning stock biomass of weak- fish could result from egg mis- identifications. However, the more protracted spawning season and greater area of spawning for weakfish in Chesapeake Bay (Olney, 1983) would make these impacts much less severe. Biochemical techniques are an important tool for the fur- ther study of eggs of sciae- nids. Genetic analysis has the potential to produce reliable results and permit the stor- age of samples for later analy- sis. Additional studies are needed to survey genetic identifications over the entire spawning season and area to determine if egg sizes change over time or are influenced by seasonal changes in hydrog- raphy or by age structure of the spawning stock. Finally, the use of genetic techniques, coupled with an extensive ex- amination of morphology could lead to the delineation of other characters that may be useful in separating the eggs of these species. Acknowledgments lower Chesapeake Bay based on OED are subject to error. These findings are particularly timely in light of the increased use of fishery-independent assess- ments of stock size that require precise estimates of egg abundance (egg production method). Because eggs of black drum and weakfish are spatio-tempo- rally coincident and OEDs overlap, estimates of egg production by black drum in lower Chesapeake Bay may be over-estimated by 50% or greater if identi- fication criteria are based solely on egg size. Like- J. McGovern, M. Cavaluzzi, J. Field, C. Baldwin, and K. Kavanaugh kindly assisted with sample collection. P. Crewe patiently processed plankton samples at sea and in the laboratory, and J. McDowell provided valuable technical assistance with mtDNA analyses. J. Olney helped in the de- sign and implementation of the sampling program and provided comments on the manuscript. Addi- tional reviews were provided by J. Musick, J. Cowan, and E. Heist. This study was funded in part by the Virginia Marine Resources Commission un- der U.S. Fish and Wildlife Contract F-95-R. 260 Fishery Bulletin 92(2). 1994 Literature cited Brown, N. J. 1981. Reproductive biology and recreational fishery for spotted seatrout, Cynoscion nebulosus, in the Chesapeake Bay area. M.A. thesis, College of William and Mary, Williamsburg, 119 p. Comyns, B. H., J. Lyczkowski-Shultz, D. L. Nieland, and C. A. Wilson. 1991. Reproduction of red drum, Sciaenops ocellatus, in the northcentral Gulf of Mexico: sea- sonality and spawner biomass. In R. D. Hoyt (ed.), Larval fish recruitment and research in the Americas: proceedings of the thirteenth annual larval fish conference, p. 17-26. Dep. Commer., NOAA Tech. Rep. NMFS 95. Cowan, J. H., R. S. Birdsong, E. D. Houde, J. S Priest, W. C. Sharp, G. B. Mateja. 1992. Enclosure experiments on survival and growth of black drum eggs and larvae in lower Chesapeake Bay. Estuaries 15(3):392-402. deSylva, D. P., F. A. Kalber Jr., and C. N. Shuster. 1962. Fishes and ecological conditions in the shore zone of the Delaware River estuary, with notes on other species collected in deeper water. Univ. Del. Mar. Lab. Inf. Ser. Publ. 51, 164 p. Ditty, J. G. 1989. Separating early larvae of sciaenids from the western North Atlantic: a review and comparison of larvae off Louisiana and Atlantic coast of the U.S. Bull. Mar. Sci. 44(3):1083-1105. Fable, W. A., Jr., T. D. Williams, and C. R. Arnold. 1978. Description of reared eggs and young larvae of the spotted seatrout Cynoscion nebulosus. Fish. Bull. 76 (l):65-72. Gold, J. R., L. R. Richardson, C. Furman, and T. L. King. 1993. Mitochondrial DNA differentiation and popu- lation structure in red drum (Sciaenops ocellatus) from the Gulf of Mexico and Atlantic Ocean. Mar. Biol. 116(2):175-185. Graves, J. E., M. A. Simovich, and K. M. Schaefer. 1988. Electrophoretic identification of early juve- nile yellowfin tuna, Thunnus albacares. Fish. Bull. 86(4):835-838. Graves, J. E., M. J. Curtis, P. A. Oeth, and R. S. Waples. 1990. Biochemical genetics of southern California basses of the genus Paralabrax: specific identifi- cation of fresh and ethanol-preserved eggs and early larvae. Fish. Bull. 88:59- 66. Graves, J. E., J. R. McDowell, and M. L. Jones. 1992. A genetic analysis of weakfish, Cynoscion regalis, stock structure along the mid-Atlantic coast. Fish. Bull. 90:469-475. Holt, G. J., S. A. Holt, and C. R. Arnold. 1985. Diel periodicity of spawning in sciaenids Mar. Ecol. Prog. Ser. 27:1-7. Holt, S. A., G. J Holt, and L. Young-Abel. 1988. A procedure for identifying sciaenid eggs. Contr. Mar. Sci. 30:99-108. Johnson, G. D. 1978. Development of fishes of the mid-Atlantic bight. Vol. rV: Carangidae through Ephippidae. Fish Wildl. Serv., U.S. Dep. Interior, 314 p. Joseph, E. B., W. H. Massman, and J. J. Norcross. 1964. The pelagic eggs and early larval stages of the black drum from Chesapeake Bay. Copeia 1964:425-434. Lansman, R. A., J. C. Avise, C. F. Aquadro, J. F. Shapira, and S. W. Daniel. 1981. The use of restriction endonucleases to mea- sure mtDNA sequence relatedness in natural populations. Ill: techniques and potential appli- cations. J. Mol. Evol. 17:214-226. Lippson, A. J., and R. L. Moran. 1974. Manual for identification of early develop- mental stages of fishes of the Potomac River estuary. Maryland Dep. Natl. Resources, Power Plant Sitting Prog. Rep. PPSP-MP-13, 282 p. Merriman, D. and R. C. Sclar. 1952. The pelagic fish eggs and larvae of Block Is- land Sound. Bull. Bingham. Oceanogr. Collect. Yale Univ. 13(3): 156-2 19. Morgan, R. P. 1975. Distinguishing larval white perch and striped bass by electrophoresis. Chesapeake Sci. 16:68-70. Olney, J. E. 1983. Eggs and early larvae of the bay anchovy, Anchoa mitchilli, and the weakfish, Cynoscion regalis, in lower Chesapeake Bay with notes on associated ichthyoplankton. Estuaries 6(l):20-35. Olney, J. E., and E. D. Houde. 1993. Evaluation and use of in situ silhouette pho- tography in studies of estuarine zooplankton. Bull. Mar. Sci. 52(2):845-872. Olson, R. R., J. A. Runstadler, and T. D. Kocher. 1991. Whose larvae? Nature 351:357-358. Pearson, J. G. 1929. Natural history and conservation of the red drum and other commercial sciaenids on the Texas coast. U.S. Bur. Fish. Bull. 44:129-214. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, NY. Saucier, M. H., and D. M. Baltz. 1992. Spawning site selection by spotted seatrout, Cynoscion nebulosus, and black drum, Pogonias cromis, in Louisiana. Environ. Biol. Fishes 36:257-272. Saucier, M. H., D. M. Baltz, and W. A. Roumillat. 1992. Hydrophone identification of spawning sites of spotted seatrout, Cynoscion nebulosus (Osteichthys: Sciaenidae) near Charleston, South Carolina. N.E. Gulf Sci. 12(2):141-145. Silberman, J. D., and P. J. Walsh. 1992. Species identification of spiny lobster phyllosome larvae via ribosomal DNA analysis. Mol. Mar. Biol. Biotech. l(3):195-205. Daniel and Graves: Morphometry and genetic identification of sciaenid eggs 261 Smith, P. J., and P. G. Benson. 1980. Electrophoretic identification of larval and 0- group flounders (Rhombosolea spp.) from Wellington Harbor, N.Z. J. Mar. Freshwater Res. 14:401^104. Welsh, W. W, and C. M. Breeder Jr. 1923. Contribution to life histories of Sciaenidae of the eastern United States coast. Bull. 39:141-201. U.S. Bur. Fish. Wisner, B. W. 1965. McClane's standard fishing encyclopedia. Holt, Rhinehart and Winston, Inc., New York, NY, 1057 p. Abstract. — The Atlantic spa- defish (Chaetodipterus faber) is the only member of the family Ephippidae in the western Atlan- tic Ocean and its life history is poorly understood. We redescribe Atlantic spadefish larvae, discuss their relationship to known larvae of other ephippid genera, and dis- cuss the distribution, abundance, and seasonal occurrence of Atlan- tic spadefish in the northern Gulf of Mexico. Larval Atlantic spade- fish are characterized by a small, peak-like, median supraoccipital crest with a single, dorsally di- rected spine; large preopercle spines, numerous serrate ridges, and other spines on the head; a deep, robust body which becomes laterally compressed; heavy body pigmentation; and early develop- ment of specialized spinous scales or "prescales" (at about 5.5-mm standard length [SL]). Transition to juvenile stage begins about 8.0- 8.5 mm SL. Developmental mor- phology and head spination of At- lantic spadefish is similar to that of Pacific spadefish, Chaetodipter- us zonatus. Sequence of fin com- pletion is pelvics — dorsal and anal soft rays — dorsal spines- pectorals. Overall, >85% of Atlan- tic spadefish larvae were found in waters >28.0°C and between 26.7 and 31.3 ppt. Larvae occur prima- rily in coastal waters, except near the Mississippi River delta, an area with a narrow shelf and rap- idly increasing water depths. Delta waters may offer additional habitat suitable to Atlantic spade- fish larvae because of lower salini- ties. Larvae are primarily collected between June and August and in the north-central Gulf of Mexico. Larval Atlantic spadefish are ap- parently rare in the eastern Gulf off Florida. Catch rates near the Mississippi River delta during August were higher than else- where in the north-central Gulf and suggest a possible association with riverine frontal areas which requires further study. A re-description of Atlantic spadefish larvae, Chaetodipterus faber (family: Ephippidae), and their distribution, abundance, and seasonal occurrence in the northern Gulf of Mexico James G. Ditty Richard F. Shaw Joseph S. Cope Coastal Fisheries Institute, Center for Coastal, Energy, and Environmental Resources Louisiana State University, Baton Rouge, LA 70803 The percoid family Ephippidae is usually considered to comprise five genera and 17 species (Nelson, 1984). The Atlantic spadefish (Chae- todipterus faber) is the only mem- ber of this family in the western Atlantic Ocean. Rare north of Chesapeake Bay, Atlantic spadefish inhabit coastal waters which ex- tend southward to Brazil (Johnson, 1978). Historically, Atlantic spade- fish represented a relatively minor portion of recreational fisheries. Nevertheless, fishing tournaments are currently being used to stimu- late interest in their fisheries (Schmied and Burgess, 1987). Ryder (1887) described eggs and yolk-sac larvae of Atlantic spade- fish, but Johnson (1978) questioned the identity of these specimens. Larvae >2.5 mm standard length (SL) are described and illustrated by Hildebrand and Cable (1938), but this study is insufficient to ex- amine important developmental details and is based on the static rather than dynamic approach to larval description (Berry and Richards, 1973). Finucane et al.1 illustrated 5.1- and 6.4-mm SL At- lantic spadefish. Johnson (1984) commented on cranial morphology and provided insight on the value of larval characters in resolving the relations among ephippids and their relation to other families. Aspects of juvenile and adult life history are discussed for Atlantic spadefish from South Carolina wa- ters (Hayse, 1990), but the distri- bution, abundance, and seasonal occurrence of Atlantic spadefish larvae are poorly understood. Our objectives are to redescribe the de- velopment of Atlantic spadefish lar- vae, discuss their relation to known larvae of other ephippid genera, and to describe the distribution, abundance, and seasonal occurrence of Atlantic spadefish larvae in the northern Gulf of Mexico (Gulf). Materials and methods The distribution, abundance, and seasonal occurrence of larval Atian- Finucane. J. H., L. A. Collins, L. E. Barger, and J. D. McEachran. 1979. Ichthyoplankton/mackerel eggs and lar- vae. Environmental studies of the south Texas outer continental shelf, 1977. Final Rep. to Bur. Land Manage., Wash., DC, Southeast Fish. Cent., Natl. Mar. Fish. Serv. NOAA, Galveston, TX 77550, 504 p. Manuscript accepted 19 August 1993 Fishery Bulletin 92:262-274 (1994) Contribution No. LSU-CFI-92-7 of Louisiana State University Coastal Fisheries Institute. 262 Ditty et al.: A redescription of Chaetodipterus faber larvae 263 tic spadefish were determined from collections taken primarily during Southeast Area Monitoring and Assessment Program (SEAMAP) ichthyoplankton surveys of the Gulf between 1982 and 1986 (SEAMAP2). These years represent the first time interval for which a complete set of data were cur- rently available. Latitude 24°30' N was the south- ern boundary of our study area in the eastern Gulf, a cutoff which approximates the continental shelf break off the southern tip of Florida. Latitude 26°00' N was the southern boundary of the central and western Gulf. These coordinates approximate the U.S. Exclusive Economic Zone (EEZ (/Fishery Con- servation Zone (FCZ). Standard ichthyoplankton survey techniques as outlined by Smith and Richardson (1977) were em- ployed in data collection. Stations sampled by Na- tional Marine Fisheries Service (NMFS) vessels were arranged in a systematic grid of about 55-km intervals. NMFS vessels primarily sampled waters >10 m deep. Each cooperating state had its own sampling grid and primarily sampled their coastal waters. Hauls were continuous and made with a 60- cm bongo net (0.333-mm mesh) towed obliquely from within 5 m of the bottom or from a maximum depth of 200 m. A flowmeter was mounted in the mouth of each net to estimate volume of water fil- tered. Ship speed was about 0.75 m/sec; net retrieval was 20 m/min. At stations <95 m deep, tow retrieval was modified to extend a minimum of 10 minutes in clear water or 5 minutes in turbid water. Tows were made during both day and night depending on when the ship occupied the station. Overall, 1,823 2 SEAMAP. 1983-1987. (plankton). ASCII characters. Data for 1982-1986. Fisheries-independent survey data/National Ma- rine Fisheries Service, Southeast Fisheries Center: Gulf States Marine Fisheries Commission, Ocean Springs, MS (producer). bongo net tows were made between 1982 and 1986. The SEAMAP effort between 1982 and 1984 also involved the collection and processing of 814 neus- ton samples taken with an unmetered 1x2 m net (0.947-mm mesh) towed at the surface for 10 min- utes at each station. SEAMAP sampling during April and May was conducted primarily off the con- tinental shelf; sampling during March, and from June through December, was conducted primarily over the shelf at stations <180 m deep. Additional information on the spatial and temporal coverage of SEAMAP plankton surveys is found in Stuntz et al. (1985), Thompson and Bane (1986, a and b), Thomp- son et al. (1988), and Sanders et al. (1990). Atlantic spadefish larvae were also obtained from surface- towed 1x2 m neuston net collections (0.947-mm mesh, 71 samples) made by the National Marine Fisheries Service (NMFS, Panama City, Florida) during August 1988. These NMFS collections were associated with riverine/oceanic frontal zones off the Mississippi River delta. Frontal zones near the delta were not sampled during either June or July. A detailed examination of Atlantic spadefish lar- vae was made to describe developmental morphol- ogy. Body measurements were made on 21 Atlantic spadefish larvae between 1.9 and 12.5 mm (Table 1) and follow Hubbs and Lagler (1958) and Richardson and Laroche (1979). Measurements were made to the nearest 0.1 mm with an ocular micrometer in a dissecting microscope. We follow Leis and Trnski's (1989) criteria for defining length of preopercular spines, body depth, head length, eye diameter, and the eye diameter/head length ratio. We consider notochord length in preflexion and flexion larvae synonymous with SL in postflexion larvae and re- port all lengths as SL unless otherwise noted. Speci- mens were field-fixed in 10% formalin and later transferred to 70% ethyl alcohol. Terminology for Table 1 Morphometries of larval Atlan surements are expressed as % tic spadefish (Chaetodipterus faber) from the northern Gulf standard length (SL) and rounded to the nearest whole num of M ber sxico. Mea- SL N Preanal length Head length Snout length Orbit diameter Body depth pectoral Prepelvic distance 1.8-2.9 3 42-55 21-31 3-8 13-15 34-44 — 3.0-4.9 3 54-65 35-43 5-7 15-17 50-60 30-36 5.0-6.9 1 60-61 30-42 5-6 15-18 56-63 30-35 7.0-8.9 4 61-64 35-39 7-9 11 55-64 27-37 9.0-10.9 4 59-61 34-35 6-8 13-14 60-65 27-34 11.0-11.9 2 54-56 34-35 7-9 11 60-65 27 12.5 1 60 36 K 1 l 68 36 264 Fishery Bulletin 92|2). 1994 location of head spines followed Gregory (1933). One larva was cleared with trypsin then stained with alizarin in each millimeter (mm) length interval to examine small serrate ridges around the orbit (i.e. circumorbital bones), and spines and ridges on the head. We examined spines on the occipital and fron- tal bones with a scanning electron microscope (SEM), and specialized spinous scales with a com- pound microscope. Fin rays were counted when first segmented and spines when present. Representative specimens were illustrated with the aid of a cam- era lucida. Estimates of larval density (number of larvae/ 100m3 of water) and catch (number of larvae/10 tow) were calculated by month. Months were combined across years because not all months were sampled every year (Appendix Table). Densities for stations where larvae were collected (i.e. positive catch sta- tions) were calculated by dividing sum of larvae collected in bongo net tows by total positive catch station volume of water filtered (VWF) and multi- plying the result by 100. In addition, an overall (i.e. grand) density estimate was calculated by dividing sum of larvae by total VWF for all stations sampled that month and multiplying the result by 100. Over- all density more closely reflects the density of lar- vae throughout the area by including the total vol- ume of water filtered in calculations. Estimates of larval catch in neuston nets were calculated by di- viding sum of larvae by number of positive catch neuston stations or by total number of neuston sta- tions sampled and multiplying the result by 10. Estimates of larval density and catch included sta- tions at long. >88°00' W because only one Atlantic spadefish larva was collected east of Mobile Bay, Alabama. Similarly, estimates were calculated only for June through August because May and Septem- ber had but one positive catch station each. Temperature and salinity data were gathered from the sea surface. Positive catch station hydro- graphic data were multiplied by total number of larvae collected at each station to obtain a monthly median and mean. Hydrographic data were also combined across months to obtain an overall (i.e. grand) median and mean. This method gives weight to distribution of larvae rather than to distribution of stations. We used a percent cumulative frequency of >85%> for defining the relation between distribu- tion of Atlantic spadefish larvae and water tempera- ture, salinity, and station depth. Percent frequency indicates the range of hydrographic conditions most often associated with occurrences of larvae. Proc Univariate was used to calculate median, mean, and percent cumulative frequency statistics (SAS Insti- tute, 1985). Results Morphometries and pigmentation Early larvae were rotund and deep-bodied; body depth was >50% SL by 3.5 mm and >60% by 9 mm (Table 1). Atlantic spadefish became increasingly deep-bodied and laterally compressed after noto- chord flexion. There were 24 myomeres but these became obscured by pigment in postflexion larvae. The head was large and averaged about 35% SL in larvae >3.0 mm. Head profile became steep and in- creasingly deeper than long. The mouth was termi- nal and the upper jaw reached to about mid-eye. Eyes were round and large, ranging from 36 to 43% of head length in larvae >3.5 mm (i.e. about 14-15% SL). The gut was tightly coiled in a single loop and the anus was slightly beyond mid-body (usually 55-60% SL). Pigment was largely restricted to the anterior-half of the body in early preflexion larvae of Atlantic spadefish. On the head of a 1.8-mm larva, external pigment was scattered over the mid- and hindbrain, nape, opercle, branchiostegal membrane, and along the isthmus and quadrate. Internally, pigment was present along and above the anterior portion of the notochord, and a single median patch was observed on the roof of the mouth. On the abdomen, there was a patch of pigment on the visceral mass immediately anterior to and below the pectoral-fin base. In ad- dition, melanophores were scattered over the pecto- ral fin base and its finfold and were distributed lat- erally over the visceral mass and hindgut. A row of about 20-25 small, closely spaced melanophores were visible along the ventral midline of the tail in early larvae. Number of melanophores along the ventral midline of the tail decreased as larvae grew. Melanophores on the nape, opercle, pectoral-fin base, and visceral mass formed a "swath" of pigment over the anterior 55-60% of the body by 2.5-3.0 mm (Fig. 1). By 3.0—3.5 mm, internal melanophores were visible anteriorly on the forebrain and laterally on the midbrain above the eye. Melanophores were also scattered both internally and externally over the hindbrain both anterior to and posterior to the base of the supraoccipital crest. By early postflexion (i.e. 5.0 mm), the head and abdomen were densely pig- mented but the posterior portion of the body was sparsely pigmented. Pigmentation increased on the posterior-half of the body as larvae grew, and by 10.0 mm the entire body was pigmented (Fig. 1). Consoli- dation of pigment into bands began on the head of Atlantic spadefish larvae with one band visible above the eye by 10.0-11.0 mm. This band of pig- ment was enclosed by indefinite, pale crossbars. The Ditty et al.: A redescription of Chaetodipterus faber larvae 265 anterior pale crossbar was situated above the middle of the eye and the posterior crossbar was behind the eye, extending mid-way down the preopercle. Lar- vae <12.5 mm had only one band of pigment (Fig. 1). The pelvics were the first fins to have pigment. Pelvic fin buds were pigmented by 4.0 mm; the pelvics were densely pigmented thereafter. Pigment Figure 1 Larval development of Atlantic spadefish, Chaetodipterus faber, from the northern Gulf of Mexico. (All. 8 mm; (Bl 3.5 mm; (C) 5.0 mm; (D) 7.0 mm; (E) 11.6 mm. All measurements are standard length (SL). appeared on the pectoral fin along the proximal portion of the rays at about 4.0-4.5 mm. Melano- phores were lightly scattered over the pectoral fin in the largest specimen examined (Fig. 1). Melano- phores were scattered over the membrane covering the anterior-most dorsal spines by about 6.0 mm and the anal spines by about 8.0 mm. Melanophores were added along the dorsal and anal fins as larvae developed, cov- ering the proximal-third of each soft ray in the largest specimen examined. Pigment was present along the proximal portion of the central rays of the caudal fin by 11.0 mm (Fig. 1). Head and body spination Atlantic spadefish larvae develop two series of preopercular spines, one along the posterior margin of the outer shelf and the other along the inner shelf. Both the outer and inner shelf have dorsal and ventral limbs. Three pre- opercular spines were present along the outer shelf of a 1.8-mm larva, the largest of which was present at its preopercular angle (Fig. 1). A fourth and a fifth spine were added by 3.5 mm, one dor- sal and one ventral to the angle of the preopercle. A sixth preop- ercular spine, smaller than the others and often difficult to locate, was present by 5.0 mm. This sixth spine was the anterior-most spine along the ventral limb of the ex- terior shelf and was resorbed by 11.0-12.0 mm in some specimens. One larva we examined had seven preopercular spines along the outer shelf but most had two spines along the dorsal limb, one at the angle, and three along the ventral limb (Fig. 2). Spines along the outer shelf were simple. Two to three spines were also present along the inner shelf of the preopercle by 3.5 mm. Number of spines along the inner shelf in- creased as larvae grew, resulting in a serrate margin (Fig. 2). A small, poorly developed opercle 266 Fishery Bulletin 92|2). 1994 spine was forming by 5.0 mm and was difficult to locate on larvae not cleared and stained. A spine also was present along the poste- rior margin of the interopercle near its junction with the subopercle by 6.0 mm (Fig. 2). The interopercular spine often was hidden by the large spine at the preopercular angle but was more easily located as the preopercular angle spine regressed. Atlantic spadefish larvae have numerous spines and ridges scat- tered over the head. A thickened ridge was visible dorsally along the supraoccipital of 2.0-mm lar- vae. This thickened ridge became a small, peak-like, median supra- occipital crest with a single, dor- sally directed spine by 2.5 mm. The supraoccipital spine began to regress by 5.0 mm and was re- sorbed by 10.0-10.5 mm. A su- praorbital ridge was present by 3.5 mm. This ridge became ser- rate by 4.0 mm. Small serrate ridges were visible along the dor- sal margin of both the lacrimal and jugal bones (i.e. first and sec- ond suborbitals; Gregory, 1933) and third suborbital bone by 5.0 mm. Spines or spinous ridges were also visible along the fourth and fifth suborbitals, dermo- sphenotic (i.e. sixth suborbital), posttemporal, pterotic, tabular, and supracleithral bones by 6.0 mm. The ventral margin of the jugal bone near the posterior margin of the maxil- lary had a single, ventrally directed spine by 7.0 mm (Fig. 2). Individual spines were also scattered over the frontal and occipital bones of young Atlantic spadefish. The bases of these spines were covered by integument so that only a portion of each spine was visible (Fig. 3). All head spines and spinous ridges were present in the largest specimen exam- ined (12.5 mm) but were difficult to locate on lar- vae not cleared and stained because of heavy body pigment. Teeth in Atlantic spadefish were placed in an in- ner and outer band. Teeth first appeared in a single band on the premaxillary and anteriorly on the dentary at about 2.5 mm. Teeth were pointed and closely spaced. A second band of teeth formed along Figure 1 (Continued) the upper and lower jaws by 4.0 mm; the outer band was slightly larger than the inner band. Teeth were added along the upper and lower jaws as larvae de- veloped (Figs. 1 and 2). Specialized spinous scales or "pre-scales" began to develop at about 5.5 mm. Pre-scales were character- ized by a single, elevated, posteriorly directed spine that was positioned near the center of the scale. Pre- scales developed first on the head and later ap- peared anteriorly along the lateral midline. Pre- scales were added outward toward the dorsal and ventral midlines and proceeded in a posterior direc- tion, covering the body by 10.0 mm. The first bones to ossify were the preopercular spines, supraoccipital crest, premaxillary, dentary, and cleithrum. Three predorsal bones (i.e. supra- Ditty et al.: A redescription of Chaetodipterus faber larvae 267 SUPRAOCCIRTAL SUPRAORBITAL POSTTEMPORAL TABULAR SUPRACLEITHRAL OPERCLE CIRCUMORBITALS INTEROPERCLE INNER PREOPERCLE OUTER PREOPERCLE Figure 2 Location of head spines on a 7.0-mm SL larva of Atlantic spadefish, Chaetodipterus faber, from the northern Gulf of Mexico. neurals) were ossifying by 6.0 mm. The anteriormost precaudal vertebrae and dorsal- and anal-fin pterygiophores ossified first; ossification proceeded posteriorly. All caudal bones were ossifying by 8.0 mm. Six branchiostegal rays and 10+14 vertebrae were present in all cleared and stained specimens. Fin development A continuous median finfold extended around the body from the nape to the anus of early larvae. Fin ray anlagen began forming obliquely downward in the caudal finfold during flexion (usually 3.5-4.5 mm). Caudal-fin ray development proceeded out- ward from mid-base as the hypural complex shifted to a terminal position, with the adult complement of 9+8 principal rays attained at about 6.0 mm (Table 2). Development of the dorsal- and anal-fin bases coincided with notochord flexion. Both fin bases and their ray anlagen began to differentiate near mid-fin; development proceeded outward from mid-fin. All dorsal and anal soft rays were present by about 7.0 mm. Soft dorsal and anal fin ray complements were present before their spines (Table 2); dorsal and anal spines developed in an anterior Figure 3 Scanning electromicrograph of the frontal and occipital spines of a 7.0-mm SL Atlantic spade- fish, Chaetodipterus faber. Epithelium was partially digested with trypsin to enhance visibility of frontal and occipital spines. Magnification: 140x. 268 Fishery Bulletin 92(2). 1994 to posterior direction. Pelvic fins were precocious and heavily pig- mented. Pelvic buds were visible by 4.0 mm; pelvics had a full complement of elements (I, 5) by 6.0 mm. Pectoral rays began to develop by 5.0 mm and a full complement (17) was present by 8.0 mm. Sequence of fin comple- tion was pelvics - soft dorsal and anal rays - dorsal spines - pecto- rals. A full complement of elements in all fins by 8.0-8.5 mm marked the beginning of transition to the juvenile stage (Table 2). Table 2 Fin-ray counts of larval Atlantic spade fish (Chaetodipterus faber) from the northern Gulf of Mexico Length (mm SD' Dorsal Anal Pectoral Pelvic Caudal 4.3 III, Anlagen 8 Anlagen Anlagen 0-7+7-0 5.0 III, 14 11 7 4 0-6+6-0 6.1 VII, 24 II, 17 13 I, 5 3-9+8-3 7.0 VII, 23 II, 18 1(1 I, 5 4-9+8-5 8.3 IX, 21 III, 17 17 I, 5 4-9+8-4 9.3 IX, 21 III, 18 17 I, 5 5-9+8-4 10.0 VIII, 23 III, 18 17 I, 5 5-9+8-5 ; One larva of each length. Temporal and spatial distribution Alantic spadefish larvae were col- lected from May through Septem- ber primarily in the north-central Gulf. Larvae were usually col- lected between June and August, density being highest during June and catch highest during August (Table 3). Larval Atlantic spade- fish were especially abundant near the Mississippi River delta during August 1988, when 19 of 72 neuston tows (26%) associated with riverine frontal zones col- lected larvae. During August 1984, however, <5% of neuston tows (rc = 162) from other areas of the north-central and western Gulf not associated with the delta captured larvae. Only one Atlan- tic spadefish larva was collected east of Mobile Bay, Alabama (long. 88°00' W). This 4.0-mm specimen was found off Apalachicola Bay (Florida) during August 1984 at a station 13 m deep (Fig. 4). Salin- ity at this station (34.2 ppt) was the highest re- corded with a positive catch during the study. The largest specimen collected in surface-towed nets was 12.5 mm; this observation may indicate that larvae move out of surface waters by this size. Overall, >85% of Atlantic spadefish larvae were collected in surface waters >28.0°C (median: 28.TC, mean: 28.7°C, range: 25.0°-32.2<,C), at salinities be- tween 26.7 and 31.3 ppt (median: 28.8 ppt, mean: 28.4 ppt, range: 11.8-34.2 ppt), and at station depths <238 m (median: 83 m, mean: 139 m, range: 9-470 m) Table 3 Density (number of larvae/100 m3) and catch (number of larvae/10 neus- ton tows) of Atlantic spadefish larvae (Chaetodipterus faber) from the northern Gulf of Mexico. Months are combined across years (1982-1986, and August 1988). Not all months were sampled each year. Numbers in parentheses are positive catch stations over total stations sampled by month. Monthly density estimates were calculated by dividing sum of larvae by either sum of volume water filtered (VWF) overall, or sum of positive station VWF. Monthly catch estimates were calculated by di- viding sum of larvae by number of stations sampled overall or by num- ber of positive catch stations. Gear June July August Bongo Overall density Positive density Neuston Overall catch Positive catch 0.3' 6.2 (19/341) 4.0 42.6 (19/201) <0.l2-3 1.3 (4/134) 0.4 13.3 (3/92) <0.\4S 1.5 (4/221) 17.0 131.6 (32/248) ' Total VWF - 43,730 m3, positive catch station VWF - 1,799 m3, number of larvae col- lected was 111. 2 0.02/100 m3. 3 Total VWF - 22,207 m3, positive catch station VWF - 381 m3, number of larvae collected was 5. 4 0.03/100 m3. 5 Total VWF - 35,174 m3, positive catch station VWF - 796 m3. number of larvae collected was 12. (Fig. 5). However, distribution of larvae versus sta- tion depth was strongly influenced by two very large neuston-net collections of 192 and 64 larvae during August 1985 which represented 40% of all larval Atlantic spadefish taken. These two stations were located in waters near the shelf edge, 50 and 75 km east of the Mississippi River delta (28.1°C, 30.1 ppt, 235 m deep; 27.9°C, 28.1 ppt, 238 m deep, respec- tively). Other stations had 27 or fewer larvae. Dis- tribution of larvae versus station depth without the two large collections shifted median station depth Ditty et al.: A redescription of Chaetodipterus faber larvae 269 shoreward from 83 to 26 m; larvae may, therefore, primarily inhabit coastal waters. This shoreward LONGITUDE Figure 4 Distribution of Atlantic spadefish larvae (Chaetodipterus faber) in the northern Gulf of Mexico by month. Months are combined across years ( 1982-1986, and August 1988). Not all months sampled each year. Plus ( + ) signs are total stations sampled and squares are positive catch stations. Distribution of stations are for both bongo and neuston net tows. shift in median station depth was reinforced by dis- tribution of larvae in bongo net tows and by distri- bution of larvae during June and July (Fig. 4, Table 4). About 86% of all Atlantic spade- fish larvae collected in bongo net tows (rc=128) were from waters <25 m deep. In addition, distribution of larvae during June and July was shoreward of that during August. Simi- larly, 51% of all stations where larvae were collected (i.e. 41 of 81) were inside 25 m; 64% were inside 50 m. Only 14% of positive catch stations were located beyond the 100 m isobath; most of these stations were near the Mississippi River delta, an area with a nar- row shelf and rapidly increasing water depths. Discussion Our observations on the morphological devel- opment of Atlantic spadefish larvae generally agree with Hildebrand and Cable (1938). These authors, however, do not discuss pig- ment on the roof of the mouth. The presence of a single, median patch of pigment on the roof of the mouth is helpful in identifying early Atlantic spadefish larvae before the supraoccipital crest is clearly visible. Hildebrand and Cable (1938) do not discuss small spines or ridges along the circumorbital bones (i.e. supraorbital, suborbitals, and dermosphenotic) or tabular bone (Fig. 2) but do illustrate serrate ridges above the eye and in the pterotic region (Hildebrand and Cable, 1938, their Figs. 26 and 27). Spination on the circumorbital bones has generally been found only in those larval percoids with cranial or- namentation (Johnson, 1984). Most of these larval percoids also have other specializa- tions, such as spinous scales and an elongate spine at the angle of the preopercle, among other characters (Johnson, 1984). Neither Hildebrand and Cable (1938) nor Johnson (1984) mention the supracleithral spines we found on Atlantic spadefish larvae (Fig. 2) and in larvae of Pacific spadefish, Chaetodipterus zonatus (Martinez-Pecero et al., 1990). The "short, hair-like spines on the upper surface of the head" noted by Hildebrand and Cable (1938) on 9.0-mm Atlantic spadefish larvae may be the same spines we found scattered over the frontal and occipital bones (Fig. 3). These frontal and occipital spines are difficult to see under a dissecting microscope because 270 Fishery Bulletin 92(2). 1994 • larvae (6) = 123 • larvae (N) - 478 GEAR TYPE Hi BONGO tXS NtuSTON 25 27 26 29 30 31 32 Temperature (°C) 12 IS 16 17 19 20 23 24 25 26 27 2B 29 X 31 32 33 34 Salinity (ppt) Depth (m) Figure 5 Summary of positive catch station hydrographic data for larval Atlantic spadefish (Chaetodipterus faber) from the northern Gulf of Mexico. Percent catch is sum of larvae by interval and gear divided by total number of At- lantic spadefish larvae collected overall. Discrepancies in number of larvae by month among parameters are the result of missing hydrographic data. Table 4 Summary of hydrographic data by month for Atlantic spadefish [Chaetodipterus faber) larvae from the northern Gulf of Mexico. Data are from the surface and for positive catch bongo and neuston net stations only. Station hydrographic data are multiplied by total number of larvae collected at each station to obtain monthly mean and median values.' W is the number of larvae used to obtain mean and median values. Discrepancies in W by month among parameters resulted from missing hydrographic data. Water temperature CO Sal nity (ppt) Station depth (m ) N Mean Median Range N Mean Median Range N Mean Median Range June July August 160 9 433 29.0 29.4 28.1 29.3 29.8 28.6 25.0-30.5 29.3-30.5 27.6-32.2 143 9 433 27.6 27.6 28.8 27.6 27.6 29.4 12.1-33.9 25.4-28.6 11.8-34.2 192 9 433 17.3 27.3 194 16 21 235 9-90 16-70 11-470 This method gives weight to distribution of larvae rather than distribution of stations. they are largely covered by integument. The supraoccipital crest was resorbed by about 10.0-10.5 mm in Gulf larvae but still present on a 11.5-mm specimen from the U. S. Atlantic coast (Hildebrand and Cable, 1938). The identity of Ryder's (1887) yolk-sac Atlantic spadefish larvae is uncertain (Johnson, 1978). Ryder's 3.5-mm and 4.0-mm larvae lack a supra- occipital crest and preopercular spines, both of which Hildebrand and Cable (1938) and we found by 2.5 mm in Atlantic spadefish larvae. Ryder's 4.0-mm larva also has an oil globule in the yolk sac and the gut does not have the single, tightly coiled loop we found in preflexion Atlantic spadefish. Nei- ther Hildebrand and Cable (1938) nor we found an oil globule in Atlantic spadefish larvae of 2.0 mm or Ditty et al.: A redescription of Chaetodipterus faber larvae 271 2.5 mm, respectively. Differences between Ryder's and our study do not support identification of Ryder's larvae <4.0 mm as Atlantic spadefish even if we allow for specimen shrinkage (also noted by Johnson, 1978) and for slower development times due to cooler waters of Chesapeake Bay during the summer when Atlantic spadefish spawn. Johnson (1984) characterized the sequence of fin completion in larval Atlantic spadefish as pattern A: dorsal and anal soft rays - spinous dorsal - pelvics - pectorals. We cleared and stained seven larvae and found the sequence of fin completion more closely resembles Johnson's (1984) pattern F with all ele- ments of the pelvic fin present before dorsal and anal soft rays. This difference in fin completion pat- tern, however, may be due to differences in how we and Johnson interpreted spine formation and fin completion. We counted rays when first segmented and spines when present; Johnson may have counted pterygiophores. Pattern F is found in Hapa- logenys, Monodactylidae, and Pempherididae (Johnson, 1984). Larvae of Atlantic spadefish are characterized by early development of specialized spinous scales or "prescales" (at about 5.5 mm, this study) that even- tually transform into adult ctenoid scales. Spinous larval scales are present to about 15.0 mm (Johnson, 1984). Ctenoid scales are well developed by 18.0 mm (Hildebrand and Cable, 1938). Developmental morphology and head spination of Atlantic spadefish is generally similar to that of Pacific spadefish ( Martinez-Pecero et al., 1990). Both species are deep-bodied (usually 55-60% SL) and preanal length is about 60% SL. Pigmentation and standard length at which fins develop also are simi- lar; a full complement of rays is present in all fins by 8.0-9.0 mm in both species (Hildebrand and Cable, 1938; Martinez-Pecero et al., 1990; this study). However, consolidation of pigment into lat- eral bands, resorption of the supraoccipital crest, and the beginning of transition to the juvenile stage occur earlier in Pacific spadefish than in Atlantic spadefish. Larvae of ephippids from the Indo-Pacific region differ from Chaetodipterus from the western Atlantic and Pacific Oceans in extent of head spination (Leis and Trnski, 1989; Martinez-Pecero et al., 1990; this study). Larvae of Platax from the Indo-Pacific have a median supraoccipital crest with a serrate leading edge (Leis and Trnski, 1989) but do not have the circumorbital series of spinous ridges, nor spines on the jugal, tabular, pterotic, or supracleithral bones found in Chaetodipterus (Martinez-Pecero et al., 1990; this study). Head spination in Ephippus larvae from the Indo-Pacific is similar to that of Chaetodipterus and these two genera are probably more closely related than either is to Platax. Other species-specific head spination found in Chaetodipterus larvae from the western Atlantic and Pacific Oceans, and in Ephippus orbis, Platax batavianus, and three Platax species from the Indo-Pacific region include a posttemporal spine which may be reduced to a ridge in some species, a supraorbital ridge that varies in size among species, and one or two subopercular spines (Leis and Trnski, 1989; Martinez-Pecero et al., 1990; this study). Early larvae of Atlantic spadefish could be con- fused with priacanthids, lobotids, some carangids and stromateoids, the wreckfish — Polyprion amer- icanus, and Menticirrhus spp. because of similari- ties in head spination or in body pigmentation. Priacanthids have an elongate, serrate, median supraoccipital crest that extends posteriorly over the mid- and hindbrain; serrations along the lower jaw and frontal bone; and the angle preopercular spine is elongate and serrate as is the pelvic spine. Trip- letail, Lobotes surinamensis, have a vaulted, serrate supraoccipital crest in early larvae, the pelvics are inserted behind the pectoral fins, and have fewer anal fin elements than Atlantic spadefish (Atlantic spadefish: A. Ill, 17-18, tripletail: A. Ill, 11-12). In carangids, the two anteriormost anal spines are separated from the third by a distinct gap and most species have a low, median supraoccipital crest that has serrations along the dorsal edge; other carangids lack a supraoccipital crest entirely. Some carangids also have a precocious dorsal fin with anterior spines or rays elongate, or with serrations along the angle preopercular spine. Some stromateoids (e.g. Ariommus spp., Nomeus gronovii) resemble Atlan- tic spadefish in early body pigmentation, body shape, and by having precocious pelvics, but stromateoids lack a median supraoccipital crest, a large preopercular angle spine, and all but Hyperoglyphe have >30 myomeres. Polyprion americanus larvae have a small, peak-like median supraoccipital crest, but with serrations along the leading edge, and lack a serrate pterotic ridge and spines on the tabular bone (Johnson, 1984). Wreckfish also have 27 myomeres, fewer dorsal (22- 24) and anal fin (11-13) elements, and the mouth is larger than in Atlantic spadefish. Larval Atlantic spadefish differ from early larvae of Menticirrhus spp. by lack of both preopercular spines and the median supraoccipital crest in the latter. We recently examined specimens reported by Dawson (1971) as larval black driftfish, Hypero- glyphe bythites. These specimens had a supra- occipital crest, pterotic ridge, spine on the inter- opercle, other head spination, and a pigmentation pattern identical to Atlantic spadefish. Vertebral, 272 Fishery Bulletin 92(2). 1994 dorsal, and anal fin counts overlap between black driftfish and Atlantic spadefish, but teeth are found in a single band on the dentary in black driftfish (Ginsburg, 1954) and in two bands in Atlantic spa- defish (Hildebrand and Cable, 1938; this study). Dawson's 5.7-7.9 mm specimens had teeth in two bands along the dentary. Because it is unlikely that black driftfish larvae have the same suite of char- acters as Atlantic spadefish, Dawson's specimens should be assigned to Atlantic spadefish. Atlantic spadefish spawn from May through Sep- tember based on seasonal abundance of Atlantic spadefish larvae in the northern Gulf; peak spawn- ing occurs between June and August (Ditty et al., 1988; this study). Density estimates were highest during June in this study (Table 3), during July in a previous study of coastal waters off central Loui- siana (Ditty, 1986), and during July and August off Mississippi Sound (Stuck and Perry, 1982). Neuston net collections were greatest during August (Table 3). Gonad maturity data off South Carolina support peak spawning of Atlantic spadefish during summer (Hayse, 1990). Spatial distribution data indicate that Atlantic spadefish larvae are apparently rare in the eastern Gulf. Only one larva was collected east of Mobile Bay (Alabama) during this study, and one larva by Houde et al.3 in a survey of Gulf waters off Florida. In addition, distribution of both larvae and station depths where larvae were collected indicates that Atlantic spadefish occur primarily in coastal waters (Ditty and Truesdale, 1984; this study), except near the Mississippi River delta where waters may offer additional habitat suitable to larvae because of lower salinities. The relatively high number of posi- tive stations (26%) near the delta during August 1988 sampling of frontal zones suggests that fron- tal zones may concentrate larvae. Frontal zone wa- ters may also provide a richer environment for feed- ing and growth of larvae because of higher phy- toplankton and zooplankton biomass (Govoni et al., 1989; Grimes and Finucane, 1991). However, Powell et al. (1990) were unable to demonstrate consis- tently that larvae have a nutritional advantage when associated with the Mississippi River plume. A possible association of Atlantic spadefish larvae with riverine frontal areas requires further study. In conclusion, understanding the biology, life his- tory, and relations of Atlantic spadefish requires a knowledge of the morphology, distribution, and ecol- ogy of their larvae. Larval characters (e.g. degree of 3 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. Univ. Miami Report BLM Con- tract No. AA550-CT7-28, Miami, FL 33149, 546 p. head spination) may also provide insight into the interrelationships among the Ephippidae and their relationship to other families. The potential use of larval characters in defining these relationships, however, cannot be clearly understood until larval development within the family is more fully docu- mented (Watson and Walker, 1992). Acknowledgments This study was supported by the Marine Fisheries Initiative (MARFIN) Program (contract numbers: NA90AA-H-MF111 and NA90AA-H-MF727). The authors would like to thank the Southeast Area Monitoring and Assessment Program (SEAMAP) and Gulf States Marine Fisheries Commission for providing specimens and environmental data; Churchill Grimes (NMFS, Panama City Lab, Florida) for access to neuston net collections off the Mississippi River delta during August 1988; and John Lamkin (NMFS, Pascagoula, MS) for provid- ing specimens of the reported black driftfish for examination. We also thank Laura Younger for pro- viding scanning electromicrographs of the head spines of Atlantic spadefish larvae. Finally, we thank the reviewers for their comments in substantially improving the manuscript. Jack Javech (NMFS, Miami, FL) illustrated the larvae. Literature cited Berry, F. H., and W. J. Richards. 1973. Characters useful to the study of larval fishes. Mid. Atl. Coast. Fish. Cent. Tech. Pap. 1:48-65. Dawson, C. E. 1971. Notes on juvenile black driftfish, Hypero- glyphe bythites, from the northern Gulf of Mexico. Copeia 1971(41:732-735. Ditty, J. G. 1986. Ichthyoplankton in neritic waters of the northern Gulf of Mexico off Louisiana: composi- tion, relative abundance, and seasonality. Fish. Bull. 84(4):935-946. Ditty, J. G., and F. M. Truesdale. 1984. Ichthyoplankton surveys of nearshore Gulf waters off Louisiana: January-February and July 1976. Assoc. Southeastern Biol. Bull. 31(2):55-56. Ditty, J. G., G. G. Zieske, and R. F. Shaw. 1988. Seasonality and depth distribution of larval fishes in the northern Gulf of Mexico above 26"00'N latitude. Fish. Bull. 86(4):811-823. Ginsburg, I. 1954. Four new fishes and one little known species from the east coast of the United States, includ- Ditty et al.: A redescription of Chaetodipterus faber larvae 273 ing the Gulf of Mexico. J. Wash. Acad. Sci. 44:256-264. Govoni, J. J., D. E. Hoss, and D. R. Colby. 1989. The spatial distribution of larval fishes about the Mississippi River plume. Limnol. Oceanogr. 34(1): 178-187. Gregory, W. K. 1933. Fish skulls: a study of the evolution of natu- ral mechanisms. Trans. Am. Philos. Soc. 23:75- 481. Grimes, C. B., and J. H. Finucane. 1991. Spatial distribution and abundance of larval and juvenile fish, chlorophyll and macrozoo- plankton around the Mississippi River discharge plume, and the role of the plume in fish recruit- ment. Mar. Ecol. Prog. Ser. 75:109-119. Hayse, J. W. 1990. Feeding habits, age, growth, and reproduc- tion of Atlantic spadefish Chaetodipterus faber (Pisces: Ephippidae) in South Carolina. Fish. Bull. 88(l):67-83. 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. 48(24):505-642. Hubbs, C. L., and K. F. Lagler. 1958. The fishes of the Great Lakes region. Univ. Mich. Press, Ann Arbor, 213 p. Johnson, G. D. 1978. Development of fishes of the mid-Atlantic Bight, an atlas of egg, larval, and juvenile stages. Vol. IV: Carangidae through Ephippidae. U.S. Fish. Wildl. Serv., Biol. Serv. Prog. FWS/OBS-78/ 12, 314 p. 1984. Percoidei: development and relationships. In H. G. Moser, W. J. Richards, D. M. Cohen, M. P. Fahay, A. W. Kendall Jr., and S. L. Richardson (eds.), Ontogeny and systematics of fishes, p. 464- 498. Am. Soc. Ichthy. Herp., Spec. Publ. No. 1. Leis, J. M., and T. Trnski. 1989. The larvae of Indo-Pacific shorefishes. Univ. Hawaii Press, Honolulu, 371 p. Martinez-Pecero, R., E. Matus-Nivon, R. Ramirez- Sevilla, D. E. Hernandez-Ceballos, and M. Contreras-Olguin. 1990. Huevo, larva y juvenil del peluquero Chaeto- dipterus zonatus (Girard) (Pisces: Ephip- pididae). Rev. Biol. Trop. 38(l):71-78. (In Spanish.) Nelson, J. S. 1984. Fishes of the world, 2nd ed. John Wiley & Sons, NY, 523 p. Powell, A. B., A. J. Chester, J. J. Govoni, and S. M. Warlen. 1990. Nutritional condition of spot larvae associ- ated with the Mississippi River plume. Trans. Am. Fish. Soc. 119:957-965. Richardson, S. L., and W. A. Laroche. 1979. Development and occurrence of larvae and juveniles of the rockfishes Sebastes crameri, Sebastes pinniger, and Sebastes helvomaculatus (family Scorpaenidae) off Oregon. Fish. Bull. 77(1):1^6. Ryder, J. A. 1887. On the development of osseous fishes, includ- ing marine and freshwater forms. Rep. U.S. Fish. Comm, Part 13, 1885 (1887):489-604. Sanders, N., Jr., T. Van Devender, and P. A. Thompson. 1990. SEAMAP environmental and biological atlas of the Gulf of Mexico, 1986. Gulf States Marine Fish. Comm., Ocean Springs, MS, No. 20, 328 p. SAS Institute, Inc. 1985. SAS User's Guide: statistics, 1985 ed. SAS Institute, Cary, NC, 584 p. Schmied, R. L., and E. E. Burgess. 1987. Marine recreational fisheries in the south- eastern United States: an overview. Mar. Fish. Rev. 49(2): 1-7. Smith, P. E., and S. L. Richardson. 1977. Standard techniques for pelagic fish egg and larva surveys. FAO Fish. Tech. Paper No. 175, 100 p. Stuck, K. C, and H. M. Perry. 1982. Ichthyoplankton community structure in Mississippi coastal waters. In Fishery monitor- ing and assessment completion report, 1 January 1977 to 31 December 1981, p. VI-I-1 thru VI-I- 53. Gulf Coast Res. Lab. (Ocean Springs, MS), Proj. No. 2-296-R. Stuntz, W. E., C. E. Bryan, K. Savastano, R. S. Waller, and P. A. Thompson. 1985. SEAMAP environmental and biological atlas of the Gulf of Mexico, 1982. Gulf States Marine Fish. Comm., Ocean Springs, MS, No. 12, 145 p. Thompson, P. A., and N. Bane. 1986a. SEAMAP environmental and biological atlas of the Gulf of Mexico, 1983. Gulf States Marine Fish. Comm., Ocean Springs, MS, No. 13, 179 p. 1986b. SEAMAP environmental and biological atlas of the Gulf of Mexico, 1984. Gulf States Marine Fish. Comm., Ocean Springs, MS, No. 15, 171 p. Thompson, P. A., T. Van Devender, and N. J. Sanders Jr. 1988. SEAMAP environmental and biological atlas of the Gulf of Mexico, 1985. Gulf States Marine Fish. Comm., Ocean Springs, MS, No. 17, 338 p. Watson, W, and H. J. Walker Jr. 1992. Larval development of sargo (Anisotremus davidsonii) and salema (Xenistius californiensis) (Pisces: Haemulidae) from the Southern Califor- nia Bight. Bull. Mar. Sci. 51(3):360-406. 274 Fishery Bulletin 92(2). 1994 Appendix Table Summary of total number of bongo net/neuston net stations examined for Atlantic spadefish larvae (Chaetodipterus faber) in the Gulf of Mexico. Acronyms are as follows: SEAMAP - Southeast Area Monitoring and Assessment Program; NMFS - National Marine Fisheries Service, Panama City, Florida. NS means no samples. Mar Apr May Jun Jul Aug Sep Oct Nov Dec SEAMAP 1982 77 VO2 69/68 71/73 102/100 26/24 NS NS 3/8 29/3 NS 1983 15/13 27/27 84/84 55/45 44/42 NS NS 39/26 NS 24/23 1984 23/0 44/0 46/0 55/54 20/26 155/162 NS 24/0 6/0 36/36 1985 29/0 NS NS 85/0 39/0 69/0 20/0 4/0 2/0 24/0 1986 NS 24/0 90/0 57/0 10/0 NS 145/0 43/0 73/0 24/0 Total 144/13 164/95 291/157 354/199 139/92 224/162 165/0 113/34 110/3 108/59 NMFS 1988 55 Tl 36 ' 60-cm bongo net, 0.333-mm mesh, oblique-tow from depth. 2 1 x 2 m neuston net, 0.947-mm mesh, 10 min. surface-tow, unmetered. Abstract. — Dolphinfishes are highly prized commercial and rec- reational species of worldwide dis- tribution in tropical and subtropi- cal seas, but the development and distribution of their larvae are poorly understood. Common dol- phin eggs hatch in about 38 hours at 25°C based on a predictive re- lationship among egg diameter, water temperature, and develop- ment time. Morphometries are generally greater in pompano dol- phin than in common dolphin. Pompano dolphin are deeper-bod- ied and have a larger eye by 9 mm, and a larger mouth and longer pre-anal length by about 13 mm. Differences in pigment along the caudal peduncle and its finfold separate common dolphin from pompano dolphin <4. 0—4.5 mm SL; common dolphin lack pigment in these areas. Number of spines along the outer shelf of the pre- opercle also separate species al- though preopercle spines are often difficult to count on larvae not cleared and stained; common dol- phin have four spines along the outer preopercular shelf and pom- pano dolphin have five. Pigmented pelvic fins and bands of pigment laterally on both the body and me- dian fins of common dolphin are diagnostic for separating species >8 mm SL; pompano dolphin lack these characters. Both common dolphin and pompano dolphin lar- vae usually are found at >24°C, >33 ppt, and beyond the 50 m isobath. Preflexion larvae (<7.0- 7.5 mm SL) were primarily col- lected in oceanic waters. Both spe- cies may spawn year-round, at least in the southern part of the survey area. Larval common dol- phin are significantly more abun- dant than pompano dolphin. Larval development, distribution, and abundance of common dolphin, Coryphaena hippurus, and pompano dolphin, C. equiselis (family: Coryphaenidae), in the northern Gulf of Mexico* James G. Ditty Richard F. Shaw Coastal Fisheries Institute. Center for Coastal, Energy, and Environmental Resources Louisiana State University, Baton Rouge, LA. 70803 Churchill B. Grimes Southeast Fisheries Science Center, National Marine Fisheries Service, NOAA Panama City Laboratory, 3500 Delwood Beach Road Panama City, FL 32408 Joseph S. Cope Coastal Fisheries Institute, Center for Coastal . Energy, and Environmental Resources Louisiana State University, Baton Rouge, LA 70803 The dolphinfishes, Coryphaena hip- purus (common dolphin) and C. equiselis (pompano dolphin), are distributed worldwide in tropical and subtropical seas (Briggs, 1960). Highly prized as food, these fishes are important recreational and commercial species, but relatively little is known about their early life stages. Gibbs and Collette (1959) reviewed spawning and adult sea- sonal distribution for the western North Atlantic Ocean, and Palko et al. ( 1982) compiled dolphinfish bio- logical data. Aoki and Ueyanagi (1989) discussed larval and early juvenile distribution for the eastern Pacific, and similar information is available for the western Pacific and Indian oceans (Shcherbachev, 1973). Preliminary distribution maps are available for the Gulf of Mexico (Gulf), but associated envi- ronmental data are not included (Richards et al., 1984; Kelley et al., 1986). Embryonic development is described for common dolphin (Mito, 1960; Hassler and Rainville, 1975; Hagood and Rothwell1) and osteological development for both species (Potthoff, 1980), but de- scriptive larval morphology is pri- marily limited to sizes >13 mm SL (Gibbs and Collette, 1959; Shcher- bachev, 1973). Okiyama (1988) and Aoki and Ueyanagi (1989) provide information on developmental mor- phology of Pacific specimens <13 mm SL, but their illustrations are 1 Hagood, R. W., and G. N. Rothwell. 1979. Sea Grant interim project report — 1979. Aquaculture in tropical ocean — Cory- phaena sp. Oceanic Inst., Makapuu Point, Waimanalo, HI 96795. Manuscript accepted: 20 October 1993 Fishery Bulletin 92:275-291 (1994). Contribution No. LSU-CFI-92-5 of Louisiana State University Coastal Fisheries Institute. 275 276 Fishery Bulletin 92(2). 1994 insufficient to examine important details, and Okiyama's study is a general overview of exisitng information. The utility of early life stages of Coryphaena in examining previous phylogenetic hypotheses and evolutionary interrelationships of echeneoids (i.e. Coryphaenidae-Rachycentridae- Echeneididae) is discussed by Johnson (1984). Our objectives are 1) to describe and compare early lar- val development of common dolphin and pompano dolphin using the dynamic approach to larval de- scription (Berry and Richards, 1973) and 2) to de- scribe the spatial and temporal distribution and abundance of early life stages of dolphinfishes in the northern Gulf. Materials and methods Seasonal occurrence, distribution, and abundance of dolphinfish larvae were determined primarily from 814 neuston net collections taken during Southeast Area Monitoring and Assessment Program (SEAMAP) ichthyoplankton surveys of the Gulf be- tween 1982 and 1984 (1982-276 stations, 1983-260, 1984-278). These years represent the first time in- terval for which a complete set of data was currently available. SEAMAP collections were made with an unmetered 1x2 m net (0.947-mm mesh) towed at the surface for 10 minutes at each station. The SEAMAP effort also involved the collection and pro- cessing of about 1,819 bongo net stations between 1982 and 1986 (1982-384 stations, 1983-288, 1984- 409, 1985-272, and 1986-466) (SEAMAP 1983- 1987)2. Bongo nets (60-cm net, 0.333-mm mesh) were towed obliquely to the surface from within 5 m of the bottom or from a maximum depth of 200 m. Sampling during April and May was primarily beyond the continental shelf, and that during March and from June to November was primarily over the shelf at stations <180 m depth. No samples were taken during January and February. Tows were made during both day and night depending on when the ship occupied the station. Latitude 24°30'N was the southern boundary of our survey area in the eastern Gulf and latitude 26°00'N the southern boundary of the central and western Gulf (Appen- dix Fig. 1). These coordinates approximate the U.S. Exclusive Economic Zone (EEZ)/Fishery Conserva- tion Zone (FCZ). Additional information on tempo- ral and spatial coverage of SEAMAP plankton sur- veys are found in Stuntz et al. ( 1985), Thompson and 2 SEAMAP. 1983-1987. (plankton i. ASCII characters. Data for 1982-1986. Fisheries-independent survey data/National Ma- rine Fisheries Service. Southeast Fisheries Center: Gulf States Marine Fisheries Commission, Ocean Springs, MS (producer!. Bane (1986, a and b), Thompson et al. (1988), and Sanders et al. (1990). Ichthyoplankton collections were also examined from riverine/oceanic frontal zones off the Missis- sippi River delta. These collections were from sur- face-towed 1x2 m neuston nets (0.947-mm mesh, 10- min. tows, sample rc=311) and were obtained from the National Marine Fisheries Service (NMFS), Panama City, Florida (i.e. May 1988 [55 neuston samples]; August 1988 [71]; September 1986 [46], 1987 [68], and 1989 [35]; and December 1988 [36]). A detailed examination of dolphinfish larvae was made to describe developmental morphology. We examined 25 common dolphin and 19 pompano dol- phin larvae between 3.5 and 15.0 mm SL for differ- ences in pigmentation, developmental morphology, and head spination, but only cursorily discuss fin development because of a thorough review of these structures by Potthoff ( 1980). Body measurements were made to the nearest 0.1 mm with a dissecting scope and ocular micrometer following Hubbs and Lagler (1958) and Richardson and Laroche (1979). We follow Leis and Trnski's (1989) criteria for de- fining length of preopercular spines, body depth, head length, and eye diameter. We consider noto- chord length in preflexion and flexion larvae synony- mous with standard length (SL) in postflexion lar- vae and report all lengths as SL unless otherwise noted. Specimens were field-fixed in 10% formalin and later transferred to 70% ethyl alcohol. We used a compound scope to examine origin and location of epithelial spicules and the maxillary spine. Juve- niles are those >25 mm, when specimens usually have developed a full complement of rays in all fins and scales (Johnson, 1984). Representative speci- mens were illustrated with a camera lucida (Figs. 1 and 2). Only three pompano dolphin <4 mm were collected and these were in too poor a condition to illustrate. Estimates of larval catch (number of larvae/neus- ton tow) were calculated for each station. Mean catch estimates by month and season were calcu- lated by dividing the sum of larvae (by species) by the total number of stations sampled within each category (month, season, etc.) and multiplying the result by 10 (number of larvae/10 neuston tows). Mean catch more closely reflects the abundance of larvae throughout the area by including total sam- pling effort in calculations. Catch was combined by month and by season across years. Seasons were defined as follows: spring=March to May; sum- mer=June to August; and fall=September to Novem- ber (Appendix Fig. 2). Nonparametric tests were used to evaluate diel, seasonal, and overall differences in catch of common Ditty et. al: Larval development, distribution, and abundance of Coryphaena hippurus and C. equiselis 277 Figure 1 Larval development of common dolphin (Coryphaena hippurus) from the north- ern Gulf of Mexico. (A) 3.5 mm, (B) 5.0 mm, (C) 7.1 mm, (D) 9.5 mm, (E) 11.0 mm, (F) 14.0 mm. All measurements are in standard length (SL). 278 Fishery Bulletin 92(2). 1994 Figure 2 Larval development of pompano dolphin (Coryphaena equiselis) from the north- ern Gulf of Mexico. (A) 4.7 mm, (B) 5.5 mm, (C) 7.5 mm, (D) 9.7 mm, (E) 11.5 mm, (F) 15.0 mm. All measurements are in standard length (SL). Ditty et. al: Larval development, distribution, and abundance of Coryphaena hippurus and C. equiselis 279 dolphin and pompano dolphin. Only those stations where either of the two species were present were included in analyses. A Kruskal-Wallis test was used to detect differences among groups (a=0.05) and a Tukey-type test to determine if mean differences were significant (Zar, 1984; SAS Institute, 1985). Dolphinfish >25 mm were excluded from analyses. Temperature and salinity data were taken from the sea surface only. Hydrographic data were mul- tiplied by number of larvae caught (by species) at each station to obtain an overall median and mean. This method gives weight to distribution of larvae rather than distribution of stations. We used a per- cent cumulative frequency of >75% for determining the relationship between distribution of dolphinfish larvae and surface water temperature, salinity, and station depth. Percent frequency indicates the range of hydrographic conditions most often associated with occurrences of larvae. Proc Univariate was used to calculate median, mean, and percent cumu- lative frequency statistics (SAS Institute, 1985). We divided the continental shelf into approximately equal geographic areas (i.e. into sq. km.) based on depth and designated the inner shelf as <50 m deep, outer shelf waters as those from 50 to 180 m deep, and oceanic waters as those beyond the continental shelf (i.e. >180 m). Results Morphology A continuous median finfold extended posteriorly along the body of early larvae of both species from the posterior midbrain to the cleithral symphysis. Remnants of the finfold were visible ventrally along the hindgut (i.e. preanal finfold) at least through 15 mm. Minute epithelial spicules covered the body of each species by 4 mm and were best observed on the head and larval finfold. Spicules were more easily observed as larvae grew. Yolk-sac larvae <3.5 mm of common dolphin and pompano dolphin had unpigmented eyes. Preflexion larvae (<7.0-7.5 mm) of both species were elongate, with body depth usu- ally <20% SL. The body became relatively deeper during flexion (about 7.5-9.0 mm) and pompano dolphin were deeper-bodied than were common dol- phin by early postflexion (Table 1). The head was moderately long (i.e. between 20 and 33% SL) in both species and the snout was short and blunt. The eyes were round and larger in pompano dolphin than in common dolphin by early postflexion (Table 1). The mouth was large and oblique; upper jaw length usually ranged from 42 to 45% of head length in postflexion dolphinfish of both species. Pompano dolphin have a larger mouth than do common dol- phin by 13 mm (Table 1). The foregut was partially convoluted and had a half-twist in preflexion larvae of both species and a single loop in larger larvae; the hindgut was straight. By 13 mm, however, preanal length was generally greater in pompano dolphin than in common dolphin (Table 1). Preanal length usually ranged from 60 to 65% SL during preflexion, but decreased thereafter to 55-60% SL in both spe- cies. The pelvic fins were moderately long ( about 15- 18% SL; Table 1) and extend to the tips of the pec- torals by 12 mm. Myomeres were obscured by heavy Table 1 Morphometries ofl arval common dolphin (Coryphaena hippurus ) and pompano dolphin (C. equiselis ) from the northern Gulf of Mexico. Measurements are expre ssed as % standard len gth (SL). SL Preanal Head Snout Orbit Upper jaw Body depth Prepelvic Pelvic (mm) N length length length diameter length cleithrum distance length C. hippurus 3.5-4.9 5 57.0-65.0 23.0-25.0 4.0-6.0 9.0-10.0 11.0-12.5 16.5-18.5 — — 5.0-6.9 4 61.0-65.0 24.0-27.0 5.0-6.0 8.0-9.5 9.0-13.0 17.0-22.0 31.0-34.0 bud 7.0-8.9 4 60.0-63.0 23.0-27.0 5.0-6.0 7.0-9.0 9.5-12.0 16.0-20.0 27.0-32.5 bud-3.0 9.0-10.9 3 56.0-59.0 25.0-28.0 5.5-6.0 9.5-11.0 12.0-13.0 19.0-21.0 27.0-30.0 4.5-11.0 11.0-12.9 5 54.0-57.0 24.0-27.0 5.0-6.0 10.0-11.0 11.0-13.0 20.0-23.0 27.0-28.0 11.0-15.0 13.0-14.9 1 54.0-56.5 25.0-28.0 4.0-5.0 11.0-11.5 12.0-12.5 21.0-22.5 26.0-30.0 17.0-18.5 C. equiselis 3.7-4.9 3 60.0-65.0 23.0-27.5 5.0-6.0 9.5-10.0 11.5-13.0 19.0-21.0 — — 5.0-6.1 4 60.0-62.0 23.0-24.0 5.0-6.0 8.5-10.0 10.0-13.0 16.0-20.0 — — 7.5-8.9 2 46.0-47.0 19.0-22.0 4.0-4.5 8.0-9.5 10.0-12.0 16.0-22.0 22.0-24.0 bud-6.0 9.0-10.9 3 56.0-60.0 27.0-30.0 4.5-5.0 12.0-12.0 12.0-15.0 25.0-29.0 30.0-35.0 8.5-14.0 11.0-12.9 4 55.0-60.0 25.0-30.0 4.0-5.0 12.0-14.0 12.0-14.0 27.0-29.0 31.0-35.0 13.0-15.0 13.0-15.0 3 55.0-60.0 27.0-30.0 4.0-5.0 13.0-13.0 13.0-15.0 25.5-28.0 28.0-34.0 16.0-18.5 280 Fishery Bulletin 92(2). 1994 pigmentation and were difficult to count on dolphin- fish larvae; however, a 5.5-mm pompano dolphin had 33 myomeres and a partially cleared 11-mm common dolphin had 30 vertebrae. Only two pompano dol- phin between 6.1 and 9.7 mm were collected (7.5 and 8.5 mm) and morphometries for these larvae were considerably smaller than for the other specimens (Table 1). Pigmentation Dolphinfish were heavily pigmented at all sizes, ex- cept the caudal peduncle and its finfold in early preflexion larvae of common dolphin which was unpigmented (Fig. 1). In common dolphin <4 mm, the length of the unpigmented portion of the cau- dal peduncle was 15-20% SL. By 4.5-5.0 mm, how- ever, pigment was present along the caudal peduncle and on the caudal finfold (Fig. 1). Early preflexion pompano dolphin <4 mm had a row of melanophores along the caudal peduncle (both dorsally and ven- trally) and pigment was scattered throughout the caudal finfold (Fig. 2). On the head, pigment was scattered externally over the premaxilla, snout, and fore-, mid-, and hind-brain of early larvae of each species. Pigment also was present along the dentary, lower jaw, isthmus, branchiostegal rays, and on the roof of the mouth. On the visceral mass, melano- phores were scattered over the foregut and anus of early preflexion larvae of both species but the hind- gut was sparsely pigmented laterally (Figs. 1 and 2). Gut pigmentation increased with length. Verti- cal bands of pigment first formed along the dorsal and anal fins of common dolphin at about 8 mm. These bands of pigment subsequently extended across the body; 12 to 13 poorly formed bands were visible by 10 mm. Vertical bands became more dis- tinct as larvae grew (Fig. 1). Bands of pigment do not form in pompano dolphin, but this species does have a row of enlarged melanophores along the body dorso- and ventro-laterally (adjacent to the dorsal and anal fin bases) by 7.5 mm, which was not present in common dolphin (Figs. 1 and 2). Pectoral buds were present on early larvae of each species. Pigment was scattered over the pectoral axilla and was heavier on pompano dolphin than on common dolphin of similar length. The proximal portion of the upper pectoral rays of common dol- phin was pigmented by 14-15 mm; no pigment was present on the pectoral rays of pompano dolphin. Dorsal- and anal-fin bases were thickening by 5 mm in pompano dolphin and by 6 mm in common dol- phin; the anal-fin base developed slightly before that of the dorsal base. Both fin bases and their ray an- lagen developed in a posterior to anterior direction. Pelvic-fin buds of common dolphin were present by 6.5 mm and pigmented by 7.5 mm. No pompano dol- phin between 6.1 and 7.5 mm were examined, but the pelvic buds were present by 7.5 mm. The pelvic rays of pompano dolphin remained unpigmented at all sizes. Pigment occurred on the developing cau- dal rays of each species by early flexion. By 10 mm, all but the distal tips of the caudal rays were pig- mented in common dolphin; only about the proximal third of each caudal ray was pigmented in pompano dolphin. Differences in caudal-fin pigmentation were more pronounced as larvae grew (Figs. 1 and 2). Head and body spination Dolphinfish larvae developed two series of pre- opercle spines, one series along the posterior mar- gin of the inner shelf and the other along the outer shelf. Number and location of spines along the outer shelf of the preopercle separate larval common dol- phin from pompano dolphin. Two spines were present along the margin of both the inner and outer preopercular shelves of 4-mm common dolphin, the largest spines occurring on either side of the angle of the preopercle (Figs. 1 and 2). A third spine was added along both the inner and outer shelf by 7 mm; a fourth spine was added along the outer preopercle by 10.0-10.5 mm. A total of three spines occurred along the inner and four spines along the outer shelf of the preopercle of larval common dolphin (Fig. 1). Arrangement of preopercle spines in larval pompano dolphin <4 mm was similar to that in common dol- phin except three rather than two spines were vis- ible along the outer preopercular shelf. A third spine was added along the inner shelf by 7 mm and a fourth and fifth spine along the outer shelf by 9 mm. A total of three spines occurred along the inner and five spines along the outer preopercular shelf of lar- val pompano dolphin (Fig. 2). Number and place- ment of preopercle spines were consistent through at least 15 mm in both species. All preopercle spines were simple (Figs. 1 and 2). Dolphinfish have several spines and ridges on the head. The pterotic area was swollen in both species by 5 mm and a laterally directed spine was present along the supraorbital ridge of each frontal bone of 6-mm pompano dolphin and 7-mm common dolphin (Figs. 1 and 2). The supraorbital ridge of each spe- cies usually had a single spine, but some pompano dolphin had two or three spines along the ridge. The swollen pterotics and supraorbital spine were best observed when specimens were viewed dorsally; both features were well developed by 7.5-8.0 mm. The frontal bone was notably thicker above the eye of pompano dolphin, but the supraorbital ridge was less well developed in pompano dolphin than in com- mon dolphin by 9.5 mm. The supraorbital spine(s) Ditty et. al: Larval development, distribution, and abundance of Coryphaena hippurus and C. equiselis 28! of pompano dolphin were regressing by 11-12 mm. A small spine was present anteriorly along the maxilla of each species by 5 mm (Figs. 1 and 2). The maxillary spine (difficult to locate because of its position and size) pointed dorso-laterally and was slightly better developed in pompano dolphin than in common dolphin of similar size. A posttemporal spine was present in both species by 9 mm and was most easily observed when specimens were viewed dorsally. The anterior portion of the lacrimal bone was prominent in dolphinfish larvae; the lacrimal was more pronounced in pompano than in common dolphin by late flexion (Figs. ID and 2D). Minute teeth were present anteriorly on the up- per and lower jaws of each species by 3.8 mm. Num- ber and size of teeth increased with SL. A pair of canine-like teeth were present in 10-mm pompano dolphin and 11-mm common dolphin. Spatial and temporal distribution Larval dolphinfish were collected during all months sampled, but small larvae of both species were found primarily during warm months. Preflexion larvae of common dolphin occurred mainly from April through November. One common dolphin larva (7.0 mm) was also collected during December (21, 1983), at a sta- tion due south of Caminada Pass, Louisiana (23.5°C, station depth: 531 m). Larval pompano dolphin were collected from March through October; larvae <10 mm were collected through late September. Only one pompano dolphin larva (5 mm) was collected during March (13, 1982; water temperature: 18°C), at a bongo-net station 29 m deep off Caminada Pass, Louisiana. Two pompano dolphin larvae (18.3 and 22.5 mm) were collected during October (14 and 17, 1983), but they were probably spawned during late September. Larvae of common dolphin and pompano dolphin were collected primarily at water temperatures >24°C (90% of larvae) and salinities >33 ppt (>75%) (Table 2, Fig. 3). The pompano dolphin collected during March was the only larva of either species taken at <21°C. Based on water temperatures when common dolphin larvae usually occurred (>24°C) and using Pauly and Pullin's ( 1988) relationship between egg diameter and water temperature to predict de- velopment time in other marine fishes, we estimate a common dolphin egg of 1.4 mean-mm diameter would hatch in about 38 hours at 25°C and 26 hours at 30°C (Table 3). Few common dolphin larvae and no pompano dolphin were collected at <25 ppt (Table 2; Fig. 3). Larval dolphinfish of both species were widely distributed in neritic and oceanic waters of the Gulf and most were collected near the surface. Over 90% of common dolphin and about 80% of pompano dol- phin occurred over the outer continental shelf and in oceanic waters; preflexion larvae were usually taken in oceanic waters (stations >180 m deep) (Ap- pendix Fig. 3). Overall, larval common dolphin were significantly more abundant than pompano dolphin (Kruskal-Wallis, P<0.0001, df= 362; Table 4). Lar- val common dolphin were also collected at more sta- tions than were pompano dolphin (15.0% versus 5.1%> of all stations sampled, respectively; Table 4). Only 3.1% of oblique bongo-net samples (1982-86, rc=1819) took common dolphin larvae (no. larvae=83, length=6.5 mm, range=3.2-21.8 mm) and <0.01% captured pompano dolphin (no. larvae = 10, x length=4.6 mm, range=4.0-8.7 mm). Differences in catch of common dolphin and pom- pano dolphin, respectively, were not significant among seasons or between day and night. About 25% of spring and 18% of fall neuston stations col- lected larval common dolphin, but <9% of those sta- tions sampled during summer (Table 4). Larval pom- pano dolphin were collected at 7% of spring neus- ton stations, 2% of summer stations, and 8% of fall stations (Table 4). Only two neuston tows collected >13 larvae of either species; these two tows ac- Table 2 Summary of hydrographic data for common dolph in {Coryphaena hippurus) and pompano dolphin (C. equiselis) larvae collected in the northern G ulf of Mexico. Data are from the sea surface only; median values are ob- tained from the di. stribution of larvae versus th e hydrographic parameter. Bon go and neuston net data are combined. 'N' is number of larvae used in obtaining median val ues. Discrepancies in W result from missing values in the hydrographic data. Salinity (ppt) Water temperature CO Station depth (m) N Median Range N Median Range N Median Range C. hippurus 537 34.0 18.7-37.8 590 28.0 21.4-32.0 599 195 11-3475 C. equiselis 80 35.1 25.0-37.8 94 27.6 18.0-30.4 94 195 11-3325 282 Fishery Bulletin 92(2). 1994 Corvphaena hioourus 50 50 N= 590 <0 40 X O 30 30 h- < o H z LU O 20- 20 DC 1X1 Q- 1 | 10 0 ..mil 10 N= 537 <21 22 23 24 25 26 27 25 29 30 31 32 19 21 22 2« 25 27 2B 29 30 31 32 33 34 35 36 37 38 <51 51-180 >180 Temperature (°C) Salinity (ppt) Station Depth (m) Corvnhaena eauiselis N= 94 I O so- 30 < ...lllllll .L.llllll. N= 80 18 21 22 23 24 25 26 27 2fl 29 30 Temperature ( C) 25 27 30 31 32 33 34 35 36 37 38 1B0 Station Depth (m) Salinity (ppt) Figure 3 Summary of hydrographic data for larval common dolphin (Coryphaena hippurus) and pompano dolphin (C. equiselis) in the northern Gulf of Mexico. Data are from both bongo and neuston net tows. Hydrographic values are rounded to the nearest whole number. N = number of larvae. Discrepancies in 'TV among parameters are the result of missing hydrographic data. Ditty et. al: Larval development, distribution, and abundance of Coryphaena hippurus and C equiselis 283 Table 3 Egg development time and hatching length (total length:TL) of common dolphin (Coryphaena hippurus). Author Egg diameter °c Hatching Study location Time (hr) TL (mm) Mito, 1960 1.28-1.62 21-29 48-60 3.95 Japan Hassler and Rainville, 1975 1.3' 27' — 3.02 Atlantic Hagood and Rothwell (see Footnote 1) 1.353 26 263 40 383 — Hawaii Hawaii Soichi, 1978 1.4-1.65 24-25 60 3.8-4.92 Japan Uchiyama et al., 1986 — 24-25 48-50 4.0-4.6 Hawaii Lamadrid-Rose and Boehlert, 1988 1.52-1.66 26 54 4. 3-5. 4* Hawaii This study 1.45 20 25 30 58s 38s 26s — Gulf of Mexico 1 Mean. 2 One-day-old larva. 3 C. equiselis. 4 Standard length. 5 Mean egg diameter and predic :ed hatching times. Table 4 Mean catch (no. larvae/10 neuston tows) of common dolphin (Coryphaena hippurus) and pompano dolphin (C. equiselis) larvae in the northern Gulf of Mexico by month. Collections for 1982-1984 are throughout the Gulf and those from 1986 to 1989 are primarily around the Mississippi River delta. Mean catch is calculated over all stations sampled by month; months are combined across years. Grand mean catch per 10 tows is calcu- lated by dividing total number of larvae collected by all stations sampled. Numbers in parenthesis are posi- tive catch stations over total stations sampled. Tax a .V March April Mai June July August September October November Grand Total C. hippurus 517 0.0(0/13) 3.4 122/100) 7.6(58/221) 2.9(29/208) 0.5(3/92) 0.5(13/248) 13.4(29/163) 3.9(4/33) 16.7(2/3) 4.8(160/1081) C. equiselis 85 0.8(1/13) 1.2(10/100) 0.9(15/221) 1.0(8/208) 0.2(2/92) <0.1 (1/248) 1.8(16/163) 0.6(2/33) 0.0(0/3) 0.8(54/1081) Number of larvae. counted for about 40% of all common dolphin lar- vae taken. Both collections occurred off the Missis- sippi River delta, one during September 1986 (n=161, 195 m station depth ) and the other during May 1988 (n-52, 63 m station depth). Discussion Early preflexion larvae (<4.0-4.5 mm) of pompano dolphin are separated from those of common dolphin by having melanophores along the caudal peduncle and scattered throughout the caudal finfold (Figs. 1 and 2). Number and placement of spines along the outer shelf of the preopercle also separate species (Table 5). Separation of dolphinfishes is particularly difficult between 4.5 and 8.0 mm because preopercle spines are often difficult to count on larvae not cleared and stained. At >8 mm, common dolphin are more easily separated from pompano dolphin by having pigment on the developing pelvic fins and bands of pigment laterally on the body and median fins (Figs. 1 and 2; Table 5). Differences in caudal- fin pigmentation also separate species by early 284 Fishery Bulletin 92(2), 1994 Table 5 Characters helpful in separating larvae of common dolphin (Coryphaena hippurus) from pompano dolphin (C. equiselis). Pigment Outer preopercle spines Meristics Species Caudal peduncle Pelvic fins Vertical bands Number of vertebrae Dorsal fin rays C. hippurus C. equiselis Absent Present' Present2 Absent Present2' 3 Absent 4 r. 30-31 33-44 58-66 52-59 ' At <4.0^1.5 mm SL. 2 At about 8.0 mm SL. 3 Laterally on body and median fins. postflexion (Figs. 1 and 2). In general, our findings agree with those of Aoki and Ueyanagi (1989). Lack of pelvic-fin pigment in pompano dolphin is diagnos- tic for separating the two species when common dolphin lose lateral banding through preservation or specimen deterioration. Although the 8.5-mm com- mon dolphin larva illustrated in Johnson (1984) lacks pelvic pigment, this specimen has bands of pigment laterally on both the body and median fins. Number of myomeres and dorsal-fin rays separate juvenile and adult common dolphin from pompano dolphin (30 or 31 vertebrae and 58-66 [x=61] dor- sal rays in common dolphin; 33 or 34 vertebrae and 52-59 [x=55] dorsal rays in pompano dolphin; Collette et al., 1969; Potthoff, 1980). Great care must be taken when counting the most anterior dorsal- fin elements (Gibbs and Collette, 1959), however, be- cause anterior dorsal rays are short and develop late (Potthoff, 1980). Myomeres are difficult to count without clearing and staining larvae because dolphinfish are heavily pigmented. Early larval development of common dolphin and pompano dolphin from the Gulf is similar to that in the western Pacific Ocean (Aoki and Ueyanagi, 1989). Developmental milestones (e.g. initial formation of dorsal- and anal-fin bases, yolk-sac absorption, and lateral body banding) occur at similar sizes in common dolphin from both the Gulf and western Pacific Ocean. We found yolk-sac absorption in common dolphin com- plete by about 3.7 mm, as did Aoki and Ueyanagi (1989). Off Japan, however, common dolphin do not complete yolk-sac absorption until about 6 mm TL (Okiyama, 1988). Aoki and Ueyanagi (1989) did not discuss either maxillary or posttemporal spines or the epithelial spicules noted during this study. Morphometries are generally greater in pompano dolphin than in common dolphin from the gulf by early postflexion (Table 1). Differences in mean mor- phometric ratios (expressed as % SL) between spe- cies from the Pacific Ocean are significant (Student's £-test, a=0.05) for larvae 5-10 mm; relative growth of all body parts measured (except preanal length) were greater in pompano dolphin than in common dolphin (Aoki and Ueyanagi, 1989). Distribution of dolphinfish larvae (Table 2, Fig. 3), juveniles, and adults is apparently limited by the 20°C isotherm (Gibbs and Collette, 1959). We found larval dolphinfish of both species primarily at tem- peratures >24°C and salinities >33 ppt, as did Fahay (1975), Powles (1981), and Aoki and Ueyanagi ( 1989). On the basis of water temperatures between 25 and 30°C (those when common dolphin larvae primarily occur), we estimate a common dolphin egg would hatch between 26 and 38 hours. Incubation time at 25°C predicted for common dolphin eggs from the Gulf was similar to that of Hagood and Rothwell1 at 26°C, but less than incubation times predicted by other studies (Table 3). Location of dolphinfish spawning is poorly docu- mented. We believe that spawning occurs in oceanic waters based on the collection of preflexion larvae of both species at stations primarily beyond the con- tinental shelf (Appendix Fig. 3). In addition, >80% of larvae of each species (Fig. 3) and 85% of stations where larvae occurred were over or beyond the outer continental shelf (Appendix Fig. 2). These findings support information from along the Atlantic coast of the southeastern U. S. that dolphinfish larvae are most abundant near or beyond the 180 m depth contour (Powles, 1981). In the Gulf, larvae of both common and pompano dolphin were collected over a similar median (Table 2), mean, and range of sta- tion depths (mean: 815 m for common dolphin and 782 m for pompano dolphin based on our weighted method of calculating these statistics). This similar- ity between species in distribution of larvae is rein- forced by the average depth of stations where lar- vae were captured. Average station depth of capture was 1198 m for common dolphin (n=216 stations) and 1042 m for pompano dolphin (n=64 stations). Ditty et. af: Larval development, distribution, and abundance of Coryphaena hippurusand C. equiselis 285 Other studies suggest that common dolphin in the tropical Atlantic (Gibbs and Collette, 1959) and Pacific (Aoki and Ueyanagi, 1989) spawn closer to shore than do pompano dolphin. In the Pacific, mid- oceanic occurrences of common dolphin larvae are lim- ited to waters near islands (Aoki and Ueyanagi, 1989). Overall, larval common dolphin are significantly more abundant than pompano dolphin in the north- ern Gulf (Table 4; Appendix Fig. 2) and along the southeastern United States (Fahay, 1975; Powles, 1981). Larvae of both common dolphin and pompano dolphin were particularly abundant around the Mississippi River delta. Higher larval dolphinfish abundances near the delta may reflect the generally higher abundance of fish larvae in the delta area (Ditty, 1986; Govoni et al., 1989; Grimes and Finucane, 1991) as compared to the open Gulf (Richards et al., 1989), or may reflect greater inten- sity of neuston sampling near the delta rather than the actual distribution of spawning adults. In the Pacific and Indian Oceans, larval pompano dolphin are more abundant than common dolphin (Shcherbachev, 1973; Aoki and Ueyanagi, 1989). Dolphinfish may spawn year-round in the Gulf, at least in the southern part of the study area where seasonal water temperatures remain above about 24°C. Estimated spawning dates based on collection of preflexion common dolphin support spawning in the northern Gulf from at least April to December (Fig. 4). Peak spawning of common dolphin occurs during spring and early fall based on higher catches of larvae during these seasons, although differences among seasons are not significant. Along the Atlan- tic coast, eggs have been collected during July and August in the Gulf Stream (Hassler and Rainville, 1975) and larvae and early juveniles year-round along the southeastern United States (Fahay, 1975; Powles and Stender, 1976) and tropical Atlantic (Gibbs and Collette, 1959). Ripe female common dolphin occur in the Gulf Stream off Cape Hatteras (North Carolina) from at least May through July (Schuck, 1951; Rose, 1966), and in the Florida Cur- rent from November to July (March spawning peak, Beardsley, 1967). Pompano dolphin spawn in the Gulf from spring through at least early fall (Fig. 4; Gibbs and Collette, 1959). If larval pompano dolphin growth rates are similar to those for common dol- phin (about 1 mm/day, Hassler and Rainville, 1975; Uchiyama et al., 1986), the two mid-October col- lected pompano dolphin larvae (18.3 and 22.5 mm) were spawned during late September. Pompano dol- phin spawn year-round in tropical mid-Atlantic and South Atlantic Bight waters based on collection of larvae and juvenile length-frequency data (Potthoff, 1971; Fahay, 1975). We found no significant diel differences in catch of larvae for either species as did Fahay (1975). Eld- ridge et al. (1977), however, found both common dolphin and pompano dolphin significantly more abundant at night, and that catch of larval common dolphin increased with concentration of Sargassum. Larval common dolphin <10 mm are more common in subsurface (i.e. depths of 20-30 m) than in sur- face tows during both day and night (Aoki and Ueyanagi, 1989). Larval pompano dolphin <10 mm are more frequently collected in subsurface tows during the day only; larvae >10 mm are more com- mon near the surface during the night (Aoki and Ueyanagi, 1989). New information on the larval morphology of pom- pano dolphin from this study corroborates Johnson's (1984) hypothesis of a relationship between Cory- phaenidae and Rachycentridae rather than that previously hypothesized between Rachycentridae and Echeneididae. Larvae of dolphinfishes and co- bia share similar patterns of head spination: later- ally swollen pterotics; a single, simple spine on the supraorbital ridge of each frontal bone (except in C. equiselis which may have multiple spines along the ridge); a small posttemporal spine; and both dolphinfish and cobia have 3 or 4 spines along the inner shelf and 4 or 5 spines along the outer shelf of the preopercle with the largest spines on either side of the preopercular angle (Johnson, 1984; Ditty and Shaw, 1992; this study). Dolphinfishes have a small maxillary spine that cobia lack (Ditty and Shaw, 1992; this study), but no spine on the supra- clei thrum found in cobia (Dawson, 1971; Ditty and Shaw, 1992; this study). Echeneis lack head spines. Larval dolphinfishes and cobia also lack large hooked teeth anteriorly on the dentary found in Echeneis (Johnson, 1984; Leis and Trnski, 1989). Dolphinfishes differ from cobia by lacking dorsal and anal spines and by having more vertebrae (30-34 in dolphinfishes versus 25 in cobia). Dolphinfishes also have 50+ soft dorsal rays, whereas cobia have 27- 33 (Ditty and Shaw, 1992). Acknowledgments This study was supported by the Marine Fisheries Initiative (MARFIN) Program (contract numbers: NA90AA-H-MF111 and NA90AA-H-MF727). The authors thank the Southeast Area Monitoring and Assessment Program (SEAMAP) and Gulf States Marine Fisheries Commission for providing speci- mens and environmental data. Bruce Mundy (NMFS, Honolulu Laboratory, Hawaii) provided a 286 Fishery Bulletin 92(2). 1994 Coryphaena hiciourus SUMMER N= 109 FALL N= 237 0 2 4 6 B 10 1! 14 16 16 20 23 24 26 SPRING I o < o en 0 2 4 6 6 10 . 12 14 16 16 20 22 24 26 Corvphaona eauiselis 0 2 4 6 8 10 12 14 16 16 20 22 24 26 0 2 4 6 6 10 12 14 16 16 20 22 24 26 0 2 4 6 B 10 12 14 16 16 20 22 24 26 STANDARD LENGTH (mm SL) 0 2 4 6 6 10 12 14 16 18 20 22 24 26 Figure 4 Length-frequency distribution of larval common dolphin (Coryphaena hippurus) and pompano dolphin (C. equiselis) in the northern Gulf of Mexico by season. Catches are from both bongo and neuston net tows; seasons are com- bined across years. Spring: March-May, Summer: June-August, Fall: September-November. Length categories are combined by 2-mm intervals; >0.5-mm increments are rounded to the nearest whole number. N = number of larvae. copy of the English translation of Aoki and Ueyanagi (1989). We also thank Cathy Grouchy for illustrat- ing the larvae. Literature cited Aoki, M., and S. Ueyanagi. 1989. Larval morphology and distribution of the dolphin-fishes, Coryphaena hippurus and C. equiselis (Coryphaenidae). Tokai Univ. Proc, Oceanogr. Sec, No. 28, p. 157-174. [In Japanese.] Beardsley, G. L. 1967. Age, growth, and reproduction of the dolphin, Coryphaena hippurus, in the straits of Florida. Copeia 1967:441-451. Berry, F. H., and W. J. Richards. 1973. Characters useful to the study of larval fishes. Mid. Atl. Coast. Fish. Cent. Tech. Pap. 1:48-65. Briggs, J. C. 1960. Fishes of worldwide (circumtropical) distribution. Copeia 1960(3):171-180. Collette, B. B., R. H. Gibbs, and G. E. Clipper. 1969. Vertebral numbers and identification of the two species of dolphin iCoryphaena). Copeia 1969(3):630-631. Dawson, C. E. 1971. Occurrence and description of prejuvenile and early juvenile Gulf of Mexico cobia, Rachy- centron canadum. Copeia ( 11:65-71. Ditty, J. G. 1986. Ichthyoplankton in neritic waters of the northern Gulf of Mexico off Louisiana: composi- tion, relative abundance, and seasonality. Fish. Bull. 84: 935-946. Ditty, J. G., and R. F. Shaw. 1992. Larval development, distribution, and ecol- ogy of cobia, Rachycentron canadum (Family: Ditty et. al: Larval development, distribution, and abundance of Coryphaena hippurus and C. equiselis 287 Rachycentridae), in the northern Gulf of Mexico. Fish. Bull. 90:668-677. Eldridge, P. J., F. H. Berry, and M. C. Miller III. 1977. Test results of the Boothbay neuston net re- lated to net length, diurnal period, and other variables. S. Car. Mar. Resour. Cent. Tech. Rep. No. 18, 22 p. Fahay, M. P. 1975. An annotated list of larval and juvenile fishes captured with surface-towed meter net in the South Atlantic Bight during four RV Dolphin cruises between May 1967 and February 1968. NOAA Tech. Rep., NMFS SSRF-685, 39 p. Gibbs, R. H., Jr., and B. B. Collette. 1959. On the identification, distribution, and biol- ogy of the dolphins, Coryphaena hippurus and C. equiselis. Bull. Mar. Sci. 9(2):117-152. Govoni, J. J., D. E. Hoss, and D. R. Colby. 1989. The spatial distribution of larval fishes about the Mississippi River plume. Limnol. Oceanogr. 34(1):178-187. Grimes, C. B., and J. H. Finucane. 1991. Spatial distribution and abundance of larval and juvenile fish, chlorophyll and macrozoo- plankton around the Mississippi River discharge plume, and the role of the plume in fish recruitment. Mar. Ecol. Prog. Ser. 75:109-119. Hassler, W. W., and R. P. Rainville. 1975. Techniques for hatching and rearing dolphin, Coryphaena hippurus, through larval and juvenile stages. Univ. N. Carolina Sea Grant Prog., UNC- SG-75-31, 17 p. Hubbs, C. L., and K. F. Lagler. 1958. The fishes of the Great Lakes region. Univ. Mich. Press, Ann Arbor, 213 p. Johnson, G. D. 1984. Percoidei: development and relationships, p. 464-498. In H. G. Moser, W. J. Richards, D. M. Cohen, M. P. Fahay, A. W. Kendall Jr., and S. L. Richardson (eds.), Ontogeny and systematics of fishes, p. 464-498. Am. Soc. Ichthy. Herp., Spec. Publ. No. 1. Kelley, S., T. Potthoff, W. J. Richards, L. Ejsymont, and J. V. Gartner. 1986. SEAMAP 1983— Ichthyoplankton. Larval distribution and abundance of Engraulidae, Carangidae, Clupeidae, Lutjanidae, Serranidae, Sciaenidae, Coryphaenidae, Istiophoridae, Xiphiidae, and Scombridae in the Gulf of Mexico. NOAA Tech. Mem., NMFS-SEFC-167, no pagination. Lamadrid-Rose, Y., and G. W. Boehlert. 1988. Effects of cold shock on egg, larval, and ju- venile stages of tropical fishes: potential impacts of ocean thermal energy conversion. Mar. Environ. Res. 25:175-193. Leis, J. M., and T. Trnski. 1989. The larvae of Indo-Pacific shorefishes. Univ. Hawaii Press, Honolulu, 371 p. Mito, S. 1960. Egg development and hatched larvae of the common dolphin-fish Coryphaena hippurus Linne. Bull. Jap. Soc. Sci. Fish. 26:223-226. Okiyama, M. 1988. An atlas of the early stage fishes in Japan. Tokai Univ. Press, Tokyo, p. 481-483. Palko, B. J., G. L. Beardsley, and W. J. Richards. 1982. Synopsis of the biological data on dolphin- fishes, Coryphaena hippurus Linnaeus and Coryphaena equiselis Linnaeus. NOAA Tech. Rep., NMFS Circ. 443, 28 p. Pauly, D., and R. S. V. Pullin. 1988. Hatching time in spherical, pelagic, marine fish eggs in response to temperature and egg size. Environ. Biol. Fish. 22(4):261-271. Potthoff, T. 1971. Observations on two species of dolphin {Coryphaena) from the tropical mid-Atlantic. Fish. Bull. 69:877-879. 1980. Development and structure of fins and fin supports in dolphin fishes Coryphaena hippurus and Coryphaena equiselis (Coryphaenidae). Fish. Bull. 78:277-312. Powles, H. 1981. Distribution and movements of neustonic young of estuarine dependent (Mugil spp., Pomatomus saltatrix) and estuarine independent (Coryphaena spp.) fishes off the southeastern United States. Rapp. P.-v. Reun. Cons. int. Explor. Mer 178:207-209. Powles, H., and B. W. Stender. 1976. Observations on composition, seasonality and distribution of ichthyoplankton from MARMAP cruises in the South Atlantic Bight in 1973. South Car. Mar. Res. Cent. Tech. Rep. Ser. No. 11, 47 p. Richardson, S. L., and W. A. Laroche. 1979. Development and occurrence of larvae and juveniles of the rockfishes Sebastes crameri, Sebastes pinniger, and Sebastes helvomaculatus (Family Scorpaenidae) off Oregon. Fish. Bull. 77:1-46. Richards, W. J., T. Potthoff, S. Kelley, M. F. McGowan, L. Ejsymont, J. H. Power, and R. M. Olvera L. 1984. SEAMAP 1982— Ichthyoplankton. Larval distribution and abundance of Engraulidae, Carangidae, Clupeidae, Lutjanidae, Serranidae, Coryphaenidae, Istiophoridae, Xiphiidae, and Scombridae in the Gulf of Mexico. NOAA Tech. Mem., NMFS-SEFC-144, no pagination. Richards, W. J., T. Leming, M. F. McGowan, J. T. Lamkin, and S. Kelley-Fraga. 1989. Distribution of fish larvae in relation to hy- drographic features of the Loop Current boundary in the Gulf of Mexico. Rapp. P. v. Reun. Cons. int. Explor. Mer 191:161-176. 288 Fishery Bulletin 92(2). 1994 Rose, C. D. 1966. The biology and catch distribution of the dolphin Coryphaena hippurus (Linnaeus) in North Carolina waters. Ph.D. diss., N. Carolina State Univ., Raleigh, 153 p. Sanders, N., Jr., T. Van Devender, and P. A. Thompson. 1990. SEAMAP environmental and biological atlas of the Gulf of Mexico, 1986. Gulf States Marine Fish. Comm., Ocean Springs, MS, No. 20, 328 p. SAS Institute, Inc. 1985. SAS User's Guide: Statistics, 1985 ed. SAS Institute, Cary, NC, 584 p. Schuck, H. A. 1951. Notes on the dolphin (Coryphaena hippurus) in North Carolina waters. Copeia 1951(1): 35-39. Shcherbachev, Y. N. 1973. The biology and distribution of the dolphins (Pisces, Coryphaenidae). J. Ichthyol. 13:182-191. Soichi, M. 1978. Spawning behavior of the dolphin, Coryphaena hippurus, in the aquarium and its eggs and larvae. Jap. J. Ichthyol. 24:290- 294. [In Jap., Engl, summ.] Stuntz, W. E., C. E. Bryan, K. Savastano, R. S. Waller, and P. A Thompson. 1985. SEAMAP environmental and biological atlas of the Gulf of Mexico, 1982. Gulf States Marine Fish. Comm., Ocean Springs, MS, No. 12, 145 p. Thompson, P. A, and N. Bane. 1986a. SEAMAP environmental and biological atlas of the Gulf of Mexico, 1983. Gulf States Marine Fish. Comm., Ocean Springs, MS, No. 13, 179 p. 1986b. SEAMAP environmental and biological atlas of the Gulf of Mexico, 1984. Gulf States Marine Fish. Comm., Ocean Springs, MS, No. 15, 171 p. Thompson, P. A, T. Van Devender, and N. J. Sanders Jr. 1988. SEAMAP environmental and biological atlas of the Gulf of Mexico, 1985. Gulf States Marine Fish. Comm., Ocean Springs, MS, No. 17, 338 p. Uchiyama, J. II., R. K. Burch, and S. A. Kraul Jr. 1986. Growth of dolphins, Coryphaena hippurus and C. equiselis, in Hawaiian waters as deter- mined by daily increment on otoliths. Fish. Bull. 84:186-191. Zar, J. H. 1984. Biostatistical analysis. Second ed., Prentice- Hall, Inc., NJ, 718 p. Ditty et. al: Larval development, distribution, and abundance of Coryphaena hippurus and C. equiselis 289 3 "5 i 7 MS Al \___GA .10 TX /_ LA ~ ,,*•*> 53*^* •^»^* s. tC "•♦*# i^is^ i * * *V '," * * * ; »* \y% • °i\ * * * ^— ^* * * * 1 • . . . to. f&*?-- ♦ \* ""In* a% * ■ 'k riT *S**\ A A (AM A AM ♦ \ * V * / ^H V A s^^t i — W 7*V* * * * "j*1* A t 9 *k »<^*\. * /ti. FL \ 2G SUMMER ^~ 1B0M ?< i i B0NG0=0 LONGITUDE NEUST0N=+ B0TH=» Appendix Figure 1 Distribution of sampling by season for the northern Gulf of Mexico. Sea- sons are combined across years. Spring (March-May), summer (June-Au- gust), fall (September-November). 290 Fishery Bulletin 92(2). 1994 Ld O 3 ) i ms| r* L GA ( LA j 30 TX »5[?0 • D [TV ^N 28 I ■ □ ■ m | n* D D * n * A ' 1 i m a a * D * D D * I 0«Q*a«0D0 \ FL \ 26 D N ! * 1 1 * SPRING 1 1 * ■ m a am * D \ a a *a, 40^ 24 D§a i 180 U LONGITUDE cohi=u B0TH=* C0E0=« Appendix Figure 2 Distribution of common dolphin iCoryphaena hippurus) and pompano dolphin (C. equiselis) larvae in the northern Gulf of Mexico by season. Seasons are combined across years. Catches are from both bongo and neuston net tows. Spring (March-May), summer (June-August), fall (September-November). Ditty et. al: Larval development, distribution, and abundance of Coryphaena hippurusand C. equiselis 291 coHi=n LONGITUDE COEQ=« 80TH=* Appendix Figure 3 Distribution of early larval (<7.0 mm) common dolphin (Coryphaena hippurus [COHI]) and pompano dolphin (C. equiselis [COEQ]) in the north- ern Gulf of Mexico. Both bongo and neuston net tows are used in determin- ing early larval distribution. Data are combined across seasons and years. Abstract. Age determination of sablefish (Anoplopoma fimbria) is typically done by counting growth zones on the burnt cross- section of the otolith. The break- and-burn method of age determi- nation is difficult to apply to sable- fish. Therefore, we applied a rela- tively new method of fish age vali- dation, using the disequilibrium of Pb-210/Ra-226 in the otoliths. This method of validation complements previous methods which used oxytetracycline (OTC) marking to validate incremental growth in sablefish otoliths. The Pb-210/Ra- 226 disequilibria generally con- firmed the ageing criteria used to interpret the otolith's burnt cross- section. Using Pb-210/Ra-226 disequilibria for sablefish, Anoplopoma fimbria, age validation Craig R. Kastelle Daniel K. Kimura Alaska Fisheries Science Center, National Marine Fisheries Service, NOAA 7600 Sand Point Way NE, Seattle, WA 981 15-0070 Ahmad E. Nevissi Donald R. Gunderson School of Fisheries, WH-10, University of Washington Seattle, WA 98195 Manuscript accepted 27 October 1993 Fishery Bulletin 92:292-301 (1994) Sablefish (Anoplopoma fimbria) is an important commercial species distributed continuously along the North Pacific Rim from California to northern Japan. On central California's continental shelf, the spawning of sablefish takes place from October to February at depths of over 823 m (Hunter et al., 1989). Both eggs and larvae have been collected at depths of over 400 m in April off British Columbia. After hatching, postlarval sablefish move into the surface waters where on- shore or offshore transport may take place (Mason et al., 1983). Postlarvae have been found off- shore, but as juveniles they are usually seen inshore (Bracken, 1983; Mason et al., 1983). The ju- veniles are believed to reside in- shore for several years and then move to deeper offshore waters as they near maturity and join the spawning population (Bracken, 1983; Mason et al., 1983). Mature sablefish are typically caught at 700 m (Mason et al., 1983). Funk and Bracken1 note that the growth of young fish is fast compared to very slow growth of mature fish. An abrupt slowing of growth coincides with the onset of sexual maturity. Mean fork length increases with depth and the length at 50% matu- rity for females is 60 cm (Hunter et al., 1989). At the Alaska Fisheries Science Center (AFSC) sablefish ages are determined by counting growth zones (assumed annular) seen on the distal surface of the otolith, or more frequently in the burnt dor- sal-ventral cross-section (break- and-burn method, Beamish et al., 1983). Even though ageing criteria have been established for sablefish by using the break-and-burn method (Beamish and Chilton, 1982; Beamish et al., 1983), vari- ability between individual sablefish in the morphology of their otoliths and the appearance of growth zones makes this method difficult to apply. Between-reader variabil- ity in sablefish ages are far greater than for any other species routinely aged at the AFSC (Kimura and Lyons, 1991). Sablefish age validation has been examined by using oxytetracycline (OTC) studies, mark and recapture of known age fish, tagging studies, length-frequency analysis of young fish, and daily growth zone counts in juvenile sablefish otoliths (Beamish 'Funk, F., and B. E. Bracken. 1983. Growth of sablefish in southeastern Alaska. AK Dept. Fish and Game Info. Leaflet No. 223, 40 p. 292 Kastelle et al.: Anoplopoma Fimbria age validation 293 and Chilton, 1982; Beamish et al., 1983). The age range considered by Beamish and Chilton (1982) was from 0 to 43 years, but their study had few individuals from the upper end of the age range. The OTC method can validate only incremental growth zones after injection, leaving interpretation of ear- lier growth in any one fish questionable. Ideally, the age range of OTC-injected fish spans all ages. Younger OTC-injected fish can then be used to in- fer that incremental growth zones seen prior to an OTC injection on older fish are annuli. With these limitations, the procedures used by Beamish and Chilton (1982) and Beamish et al. (1983) confirmed that ages counted on the burnt cross-section were accurate. Our goal was to use radiometric dating techniques to validate the break-and-burn ageing criteria used at the AFSC for sablefish aged to approximately 35 years. This validation used the measured ratio of Pb-210/Ra-226 in the otoliths to provide estimates of total age, thereby complementing previous OTC validation work which only confirmed incremental growth. In the first application of radiometric ageing to fishes, Bennett et al. ( 1982) used the ratio of Pb-210/ Ra-226 to validate ages up to 80 years for splitnose rockfish (Sebastes diploproa). More recently, these radioisotopes were used by Campana et al. (1990) and Fenton et al. (1990, 1991) in age validation and longevity studies in a variety of fish species. Addi- tional radioisotope pairs such as Th-228/Ra-228 have also been used to age fish (Smith et al., 1991). The isotopes Ra-226 and Pb-210 are part of the naturally occurring decay chain of U-238: L/-238 > Ra-226 > Rn-222- 4.5xl093T KSOOvr Pb-210 3.8 d Pb-210- ->Po-210 >P6-206 22.3 vr 138 d (1) where the dashed lines indicate short-lived interme- diary nuclides that are not shown. Both Ra-226 and Pb-210 are found naturally in seawater. Ra-226 is a calcium (Ca) analogue which accompanies Ca through the food chain and is de- posited in fish tissue, particularly calcified struc- tures, along with Ca (Swanson, 1985; Porntep- kasemsan and Nevissi, 1990). The otoliths of teleo- sts consist of an acellular organic protein matrix mineralized with aragonite, a form of calcium car- bonate in which the radioisotopes are deposited (Mugiya, 1977; Campana and Neilson, 1985). Pb-210 is also accumulated by the biota through the food chain (Shannon et al., 1970; Heyraund and Cherry, 1979). In fish, Pb-210 is preferentially deposited in the bone or liver2 (Swanson, 1985). Its initial activ- ity in the otolith must be measured (later as R*) for the application of radiometric ageing. When Ra-226 is deposited in otoliths, like Ca it remains immobile, and a disequilibrium is created between Ra-226 and all of its progeny. With time, the activity of shorter-lived daughter products like Pb-210 will increase. In the pair of radioisotopes used here, Ra-226 and Pb-210, the difference be- tween their half-lives is great (Eq. 1). Therefore, after about 100 years in a closed system the activ- ity of both Ra-226 and Pb-210 will become equal, establishing a so called "state of secular equilibrium" (Faure, 1986). A chronometer is started when Ra- 226 is first deposited in the otolith, and the activity ratio Pb-210/Ra-226 is a function of the time elapsed since deposition. Radiometric dating applied to fish otoliths relies on three basic assumptions (Faure, 1986; Smith et al., 1991): 1 The otoliths are closed with respect to the loss or gain of any radioisotopes in the decay chain. 2 The initial activity ratio of Pb-210/Ra-226 in the otoliths should be much smaller than one, ideally close to zero, and known or measured. 3 The specific activity [disintegrations per minute per gram (dpm/g)] of the radioisotopes in the material incorporated into the otoliths must be constant. These assumptions will be considered in detail later in the "Discussion" section. But first, it is im- portant to consider Assumption 3 because it explains why we did not use whole otoliths. Assumption 3 is the most problematic of the three assumptions when applied to whole otoliths because it also requires assuming a mass-growth rate for the otoliths (Smith et al., 1991). Campana et al. (1990) used otolith cores in their application of radiometric dating to fish. When using otolith cores and individual mea- surements of Ra-226 for each sample being radio- metrically aged, Assumption 3 becomes unnecessary. However, if measurements of Ra-226 are averaged over the different samples being radiometrically aged, Assumption 3 requires that the different core samples have the same activity levels. Our study followed the procedures of Campana et al. (1990) and used otoliths cores. But unlike Campana et al. ( 1990), we used individual Ra-226 measurements for each sample being aged so that Assumption 3 was unnecessary. Noshkin, V. E., K. M. Wong, R. J. Eagle, T. A. Jokela, and J. A. Brunk. 1988. Radionuclide concentrations in fish and inver- tebrates from Bikini Atoll. Lawrence Livermore National Labo- ratory, Livermore, Ca., UCRL-53846, 53 p. 294 Fishery Bulletin 92(2). 1994 Materials and methods Otolith collection Sablefish were collected along the Eastern Pacific Continental Shelf during two research cruises by the AFSC. The first collection was made from the re- search vessel Alaska on 8 August 1986 and con- tained 61 fish captured at 135 m in Morro Bay, California. The second collection was made from the vessel American Viking between 23 September and 31 October 1986 and contained 423 fish captured at depths of 246-1426 m off the California coast be- tween 32°23'N and 42°23'N. Both sagittal otoliths were removed at sea. The sacculus membranes were removed and otoliths were stored in 50% ethanol. Ages were estimated by the first author (Reader 1) from one otolith per fish applying the break-and- burn method, and the other otolith was used in the radiometric study. The criteria used to count annuli for this study were similar to those typically used by experienced age readers at the AFSC and Tiburon Laboratories of the National Marine Fisheries Ser- vice (NMFS), and were the same used by Beamish and Chilton (1982) and Beamish et al. (1983). By using these criteria, an age range of up to ±5 years for older fish was often possible. The oldest age con- sistent with these criteria was often assigned as the most probable age. When a fish was aged as 1 year old the otolith surface was usually adequate and no break-and-burn was done. The otoliths were subse- quently pooled into four age categories (1 year, 9— 11 year, 14—23 year, and 24-34 year), on the basis of Reader- 1 ages. For comparison, a subsample of otoliths initially aged >14 years (re=186) was read by a second expe- rienced sablefish age reader at the AFSC (Reader 2). Additionally, all fish initially aged >14 years (rc=266) were read by an experienced sablefish age reader at the Pacific Biological Station (PBS), Cana- dian Department of Fisheries and Oceans (Reader 3). In the radiometric dating procedures, we used that part of the otolith which was deposited during the first year of life (i.e. the first year core). For age category 1, whole otoliths representing the first 15 to 18 months of growth were used. These otoliths were those classified as age 1 by Reader 1 and were intact and unburnt. For other age categories, the first year cores were isolated by grinding away ex- cess otolith material. The grinding was done with a Buehler metallurgical polishing machine equipped with Buehler wet and dry #600 or #900 paper. Otolith material representing the first year was readily identifiable in older fish from the distal sur- face of the ground otolith, and from the broken-and- burnt section, viewed with a dissecting microscope (25x) as a guide. Removal of material was done slowly, with frequent viewing of the otolith during the grinding process (Kastelle, 1991). The position of the first annulus on the otolith has been con- firmed by several authors in age validation studies of young sablefish (Beamish and Chilton, 1982; Beamish et al., 1983; McFarlane and Beamish, 1983). Average measurements of the core dimensions were not used as an aid in the grinding process. Instead, grinding was completed when the contours of the first year were approximated. Small inaccu- racies in the grinding were inconsequential because samples were pooled into four categories based on age ranges. All otolith cores were cleaned with an ultrasonic cleaner in distilled and deionized water for a mini- mum of 30 seconds. Any soft tissue remaining on the otolith after collection was visibly broken down by the ultrasonic cleaner. The goal of the cleaning was to remove any contamination from soft tissue or grinding paper. After cleaning, the otoliths were stored again in a fresh 50% ethanol solution prior to analysis for radioisotopes. Approximately one gram of material was neces- sary for radioisotope activity to be measurable above background levels, which meant that 83 to 141 otoliths were used for each age category. To increase the weight of category 1 ( 1 year olds), an additional unburnt half of the aged otolith from some speci- mens was included with the intact otoliths (83 speci- mens total: 83 whole otoliths plus 65 half otoliths from some of the same fish). Activity measurements The methods employed in the chemical separation and counting techniques for Ra-226 and Pb-210 were similar to those used by Bennett et al. (1982) and are detailed in Kastelle (1991). In general, the activity of Pb-210 in the otoliths was determined by counting the alpha decays of Po- 210 (the granddaughter-proxy of Pb-210 with which it is in secular equilibrium, Eq. 1) by using a yield tracer of Po-209. Reagent blanks, with and without yield tracers, were processed with each age category as follows: category 1, one blank without yield tracer; categories 2 and 3 processed simultaneously, one blank with a yield tracer and one without; cat- egory 4, one blank with a yield tracer and one with- out. The counting time for each sample or reagent blank was approximately three weeks. The activity of Ra-226 was determined by count- ing alpha decays of Rn-222 (the daughter-proxy of Ra-226 with which it grows into secular equilibrium, Kastelle et al.: Anoplopoma fimbria age validation 295 Eq. 1) in a Lucas cell (Lucas, 1957). The samples were stored in a Rn-222 de-emanation flask for a minimum of three weeks prior to counting. During storage, Rn-222 reached secular equilibrium with Ra-226. Reagent blanks were processed with each sample. Lucas cell (Rn-222) counting times were 4,000 minutes for each sample or reagent blank. Lucas cell and electronic backgrounds (counting time 1,000 minutes) were measured multiple times before and after any sample or reagent blank. The counting efficiency (i.e. the percentage of decays that was detected) of the de-emanation technique, Lucas cells, and associated electronics was determined to be 53.75% with the use of a Ra-226 standard solu- tion supplied by the U.S. Environmental Protection Agency. Activities of Ra-226 and associated errors were calculated by using the methods of Sarmiento et al. (1976). Details concerning incorporation of blank and background measurements into the activ- ity calculation for Po-210 and Ra-226 are given in Kastelle (1991). All sources of error were propagated through the calculations to estimate errors (stan- dard deviation or SD) for the Ra-226 and Pb-210 measurements. Data analysis ANOVA with contrasts was used to test whether break-and-burn otolith ages from the different age readers were significantly different within each age category. Statistical analyses of activity measure- ments from samples, backgrounds, and reagent blanks were conducted by using Z, t, and a likeli- hood ratio x2 test as described below. To test if the Ra-226 (or Pb-210) activity from each age category was statistically different, the likelihood ratio test was employed. Assuming X; is distributed as Ni/u^crf), where the &f are known, the like- lihood ratio x2 test for H0 :/Jl = ... =/J„. is I>, (if/of which is distributed as y2 , . HereX, is the measured Ra-226 (or Pb-210) activity for age category /, /} = £(X,/, and of is the variance for age category i. If of's are underestimated, then the %2 would be inflated. Z tests were carried out between the reagent blanks plus background, and background alone, for Po-209 and Po-210 measurements; a i-test was performed between the mean Rn-222 reagent blank plus back- ground and the background alone. Two sets of estimated Pb-210/Ra-226 ratios were calculated: one by using the mean Ra-226 activity from the four categories and the second by using the Ra-226 activity measured for each category. The delta method was used to determine the variance of the ratios (Seber, 1982). The measured ratio of Pb-210/Ra-226 in the otoliths can be used to predict a radiometric age from the curve: (2) where Aj is the activity for Ra-226, A2 and X2 are the activity and decay constant respectively (^.=ln (2)/half-life, for Pb-210, t is time (i.e. age), and R* is the initial ratio of Pb-210/Ra-226 (Fig. 1, see also Kastelle [1991]). The initial ratio was estimated by solving Equation 2 for R* and applying the activity ratio from age category 1 (R*=-0.034). The negative value for R* is due to measurement error. Therefore, we assume R* - 0 in Figure 1. The actual fish ages were predicted from the radiometric ages by sub- tracting the time between collection and analysis (4.5 yr) from the radiometric ages in Figure 1. Al- ternatively, it is possible to calculate the radiomet- ric ages by correcting the Pb-210 activity estimates to the time of otolith collection. These adjusted ra- diometric ages were then compared with ages read from the burnt otolith cross-section by the three readers. For each of the readers, a linear regression line was fit through the origin and compared with the 45° line of agreement by using a £-test. 0.7 n (0 £ 0.05). The background activity for Po-209 and Po-210 ranged up to 6.413X10-4 ±1.472xl0"4cpm (counts per minute) and 5.600x10^ ±9.75xl0-5, respectively (Table 2). Therefore, the unadjusted sample data (including background counts) were reduced by the appropri- ate background only; no adjustment was made for the reagent blank. Specific activity of Pb-210 at the time of separation from Ra-226 ranged from 0.037 ±0.007 dpm/g in category 1 to 0.265 ±0.041 dpm/g in category 4 (Table 3). The Pb-210 activity levels were significantly different among the four age cat- egories (P<0.001). The average background count (n=66) for the Rn- 222 system was 8.500xl0"3 ±5.27x10^ cpm. The mean reagent blank plus background for Rn-222 was not significantly greater than the background alone (P=0.077). Mean Ra-226 activity (n=4) was 0.414 ±0.050 dpm/g (Table 3). Ra-226 values from the four age categories were significantly different (P<0.001). Therefore, the Ra-226 measurements were specific to each age category. The adjusted radiometric ages (from category spe- cific Ra-226 measurements) for age categories 1 to 4 respectively were -0.09, 5.21, 17.83, and 22.66 years. Although the adjusted radiometric ages were consistently younger than the burnt cross-section ages from Reader 1, there was general agreement between the two methods (Fig. 2). A r>test between the slope of a line fit through the origin vs. the 45° line did not show a significant difference for any of the age readers (P>0.05). Discussion Principal findings The principal result of this study is that the radio- metric ages generally confirmed the burnt otolith ageing criteria that are used to age sablefish by U.S. and Canadian age readers. A factor which facilitated this confirmation is that sablefish otoliths appar- ently accumulate higher levels of radioisotopes than do other fish species that have been previously stud- ied. We found the specific activities of Pb-210 and Ra-226 in sablefish otoliths (Table 3) to be a full order of magnitude greater than values reported in other species (Bennett et al., 1982; Campana et al., 1990; Fenton et al., 1991). These large differences in activity levels may be explained by biological and environmental considerations. Radium-226 is incor- Table 1 For each age category of sablefish. Anoplopoma fimbria, comparison of average esti mated age, number of samples read age range by each age reader, and radiometric age are shown. The time between collection and analysis (4.5 yr) was subtracted from the radiometric age at the time of analysis to mak e the ages comparable. First Reader: Second Reader: Canada (PBS): Radiometric age: Age average age (yr), average age (yr), average age (yr), average age (yr). category n read, range (yrl n read, range (yr) n read, range (yr) n pooled, range (yr) 1 1, rc=83, (1) — — -0.09, n=83 (-0.89, 0.74) 2 9.79, n=130, (9, 11) — — 5.21, n = 130 (3.40, 7.12) 3 18.91, n = 141. 22.07, n=101, 15.58, n = 139, 17.83, n=141 (14, 23) (14, 42) (8, 27) (13.05, 23.441 4 28.55, n = 127, 31.01, n=85. 23.75, n = 127 22.66, n = 127 (24, 34) (17, 49) (11, 48) (16.50, 30.30) ' Based on measured Pb-210/Ra-226 ratio and error with individual Ra-226 measurements and Eq. 2. Kastelle et al.: Anoplopoma fimbria age validation 297 Table 2 Comparison of background (Bk) and blank (BL) activity reported in the literature for tracer spikes, Po-210 in equilibrium with Pb-210, and Ra-226. NR = not reported Tracer spike (Po-209 or Po-208) Po-210 Ra-226 This study Bk (cpm) 1.867 x 10^ ± 5.63 x 10~5 to 2.353 x lO^1 ± 6.53 x 10"5 to 8.500 x 10"3 ± 5.27 x 10"4 6.413 x 10"4 ± 1.472 x 10"4 5.600 x 104 ± 9.75 x 10"5 Bl (dpm) Not significant Bennett et al. (1982) Bk (cpm) NR Bl (dpm) NR Campana et al. (1990) Bk (cpm) NR Bl (dpm) NR Fenton et al. (1990) Bk (cpm) NR' Bl (dpm) NR Fenton et al. (1991) Bk (cpm) 6.94 x 10"4 Bl (dpm) NR Not significant 6.94 x 10"4 Not measurable above Bk 6.94 x 10"4 5 x 10-4 NR' Not significant 1.4 x 10-2 to 2.8 x 10-2 3.9 x 10"2 NR NR 1.03 x 10-2 + 2.7 x 10"3 1.25 x 10"2 ± 4.3 x 10" 6.94 x 10"4 1.4 x 10"3to 3.5 x 10" 1.03 x 10"2 ± 2.7 x 10"3 2.55 x 10"2 ± 2.3 x 10~3 ' Not reported specifically for this radioisotope, but this author suggested a lower value here than reported by Bennett et al. (1982). Table 3 For each age category of sablefish, Anoplopoma fimbria, number of specimens, average estimated age and estimated age range (from Reader 1), specific activity (dpm/gram) of Ra-226 and Pb-210, and the ratios Pb- 210/Ra-226 and the ratios Pb-210/mean Ra-226, at the time of analysis 4.5 years after collection are shown. Errors are ±1 SD and are rounded up in the last significant digit for presentation. Age category Number pooled Average age (yr) Age range (yr) Ra-226 (dpm/g) Pb-210 (dpm/g) Pb-210 Ra-226' Pb-210 Ra-2262 1 83 1 1 0.288 ± 0.012 0.037 ± 0.007 0.128 ± 0.023 0.089 ± 0.019 2 130 9.79 9-11 0.517 ± 0.021 0.135 ± 0.022 0.260 ± 0.043 0.325 ± 0.065 3 141 18.91 14-23 0.386 ± 0.017 0.193 ± 0.030 0.500 ± 0.080 0.467 ± 0.092 4 127 28.55 24-34 0.465 ± 0.019 0.265 ± 0.041 0.570 ± 0.091 0.640 ± 0.126 ' Activity ratios 2 Activity ratios calculated with Ra-226 measured from each group, calculated with mean Ra-226: 0.414 ± 0.050. porated into the biota or food chain through a vari- ety of sources, such as water during the osmoregu- lation process, from food (Porntepkasemsan and Nevissi, 1990) such as phytoplankton (Shannon and Cherry, 1971) and zooplankton (Evans et al., 1938), and from contact with sediments (Swanson, 1985). The geographic area the species inhabits, its verti- cal distribution in the water column, and the effi- ciency of transfer of Ra-226 through its food chain could play a role in the uptake of Ra-226 (Cochran, 298 Fishery Bulletin 92(2), 1994 35n 30 m < 25 o ■c 20 0) E o 15 "D CO DC 10 3 5 V, 3_ 0 CD < CD E g CO cr T5 CD w 3 < CD D) < CD E g 1 ac "D a> fa 3 < /*. i- o- ,1, —i — i — i — i — i — i — i — i — i 10 15 20 25 30 35 40 45 50 35- Burnt Cross-section Age (yr), Reader 1 30- * 25- ,'' 20- / 4 15- / 3 10- / 5- / 0- -5- : — i — i — i — i — i — i — i — i — i — i — i Burnt Cross-section Age (yr), Reader 2 35 -| 30- 25- / 20- / 4 15- -' ' 3 ,' 10- / 5- / 0- ■5i ■ r i i i i T 1 1 1 1 1 -5 0 5 10 15 20 25 30 35 40 45 50 Burnt Cross-section Age (yr), Reader 3 Figure 2 Comparison of burnt otolith cross-section ages (from Readers 1, 2, and 3) with adjusted radiomet- ric ages (4.5 years were subtracted for the time between capture and analysis I. Range in adjusted radiometric ages is 1 standard deviation, and the range in burnt otolith cross-section ages are the maximum and minimum in each age category (see Table 1 ). Ranges of burnt cross-section ages for Readers 2 and 3 are wider than for Reader 1 be- cause age categories were defined by Reader 1 ages. 1982; Swanson, 1985; Fenton et al., 1990). Addition- ally, feeding rate, metabolic rate, and calcium depo- sition rate could all affect the specific activities found in otoliths. Sablefish are one of the fastest- growing epipelagic juvenile fishes (Shenker and 011a, 1986; Kendall and Matarese, 1987). Therefore, the higher Ra-226 and Pb-210 activity levels seen in sablefish otolith cores could be related to rapid uptake of Ra-226. Assumptions for radiometric dating We discussed three assumptions which must be sat- isfied for the radioisotope ageing of sablefish to be valid. Assumption 1, that otolith cores are closed with respect to loss or gain of any radioisotopes in the decay chain, has not been tested. Considering the decay chain containing Ra-226 (Eq. 1), Rn-222 is a source of concern. For Ra-226 to decay to Pb- 210, it must first become Rn-222 (half-life=3.82 days) which is a noble gas. It is conceivable that Rn- 222 could migrate in the otolith. A loss of Rn-222 would lead to underestimation of the true age. How- ever, the calcium carbonate crystalline structure of the otolith probably prevents radioisotopes from migrating.3 Welden (1984) measured Pb-210 activity in the calcified cartilage of vertebral centrum from sharks, applying a procedure similar to the dating of sedi- ments. The calcified cartilage allowed the Pb-210 to migrate in the vertebra so the assumption of a closed system did not hold true. Previous research by Goreau and Goreau (1960), Moore et al. (1973), Dodge and Thomson (1974), and Veeh and Burnett (1982) confirmed that calcium carbonate in coral acted as a chemically closed system. In Rn-222 dif- fusion experiments, Moore et al. (1973) could not detect migrations in coral. For our study, Assumption 2 required that in the core of otoliths, initial ratios of the two radioisotopes be measured or known, and ideally be near zero. More Ra-226 than Pb-210 may be encountered by fish since Ra-226 is not in equilibrium with Pb-210 in seawater (Bacon et al., 1976). The environmen- tal residence time of Ra-226 in seawater has been reported to be as high as 950 years (Szabo, 1967) and as low as 0.7 to 5.5 years (Shannon and Cherry, 1971). The environmental residence time of Pb-210 in surface waters has been reported to be as high 3 At the "CSIRO, International workshop on otolith chemistry" 2-6 March 1992, in Hobart, Tasmania, Australia the possibil- ity of Rn-222 migration was discussed with some enthusiasm. G. Fenton suggested that the otolith is relatively impermeable, and that Rn-222 migration is not a major problem Kastelle et al.: Anoplopoma fimbria age validation 299 as 3.5 years (Shannon et al., 1970) and as low as 1.4 years (Bacon et al., 1976). Pb-210 is incorporated into particulate matter whereby it is removed from the water column and deposited into sediments. To correct for Pb-210 incorporation, we measured the initial ratio, R*, with young fish (1 yr olds). By ana- lyzing very young splitnose rockfish, Bennett et al. (1982) found that the initial ratio of Pb-210/Ra-226 was between 0.1 and 0.2. Similar values for the ini- tial ratio were also found in redfish (Sebastes mentella) (Campana et al., 1990), orange roughy (Hoplostethus atlanticus) (Fenton et al., 1991), and blue grenadier (Macruronus novaezelandiae) (Fenton et al., 1990). These results are comparable with the R*=0 we used. The initial ratio cannot be estimated from older age categories because their true ages are uncertain. Therefore, R*=0 for older age categories was assumed. Assumption 3 states that the specific activity of Ra-226 incorporated into the otolith core be constant over the time span that the otoliths are receiving the radioisotopes. Assumption 3 is not required if only the otolith core is used and individual Ra-226 mea- surements are made for each age category. In sable- fish the core of 1-year-old fish was appropriate be- cause it is large, and there is a strong possibility that fish migration prior to maturity might cause a change in Ra-226 uptake. The likelihood ratio x2 test suggested that Ra-226 measurements differed among age categories. Therefore, ratios calculated by using individual Ra-226 measurements were pre- ferred over those calculated with the Ra-226 mean. Campana et al. (1990), using cores composed of the first 5 years of growth, and a mean of 5 whole otolith samples (13 g per sample), found no significant dif- ference in Ra-226 activity between the cores and whole otoliths. This suggested a constant rate of Ra- 226 uptake. The differences we found between the Ra-226 activity measured in the four age categories were considerable (Table 3). Because the factors controlling Ra-226 uptake are complex and not well understood, the observed differences may well be real. The use of whole otoliths requires modeling the mass growth rate of the otolith over the life of the fish and making assumptions concerning the uptake of Ra-226 and Pb-210 each year. Bennett et al. (1982) modeled the otolith's mass growth rate and assumed that the uptake of Ra-226 over the life span of the fish was proportional to otolith size (i.e. they assumed a constant rate of uptake). In whole otoliths from blue grenadier, the rate of uptake of Ra-226 changed over the life of the fish (Fenton et al., 1990). Therefore, the assumption of a constant rate of uptake of Ra-226 was violated. In whole otoliths from orange roughy, the Ra-226 specific activity increased with age (Fenton et al., 1991). Therefore, Fenton et al. (1991) used two Ra-226 averages: one for young age categories and a second for old age categories. The conservative approach is to use the otolith core and individual category Ra-226 measurements making Assumption 3 unnecessary. Also, by using cores the closed system considered under Assump- tions 1 and 2 is reduced. The behavior of otolith material deposited later in a fish's life need not be considered. The different approaches (otolith core, whole otolith, larger multi-gram samples, or any av- eraging of different age ranges) have trade-offs which should be evaluated in light of a species' biology. Sources of error in ageing methodologies Differences between the three sets of burnt cross- section ages are explainable. First, storage of bro- ken-and-burnt otoliths in ethanol between readings may cause a fading of growth patterns. Second, variations in application of the otolith interpretation criteria of Beamish and Chilton (1982) and Beamish et al. (1983) could also lead to differences. Even with the high specific activity (compared with other species) found in sablefish otoliths, the generally low activity levels of Po-210 and Ra-226 found in otoliths makes it important to carefully evaluate reagent blanks and background activities. Therefore, activity levels for reagent blank and background measurements reported in four previous studies (Bennett et al., 1982; Campana et al., 1990; and Fenton et al., 1990, 1991) were compared with those found here (Table 2). Considering the magni- tude of errors, the Po-210 background we found was similar to other reported values (Table 2). The lit- erature showed a much greater range of activities measured in the reagent blanks. Like Bennett et al. (1982), we found a nonsignificant activity level in the Po-210 reagent blank. The Po-210 reagent blank of 0.0103 ±0.0027 dpm reported by Fenton et al. (1990, 1991) seems high compared with the other findings. Also, we were the only study to report a nonsignificant Ra-226 reagent blank activity. Except for the Ra-226 reagent blank result, our background and reagent blank measurements for both Ra-226 and Pb-210 are in the same range as those reported in previous applications of radiometric ageing to fish. The Ra-226 (or equivalent Rn-222) activity levels were low and difficult to measure. The signal to background ratio can be increased by using a greater sample weight. Fenton et al. (1991) argued that measuring Ra-226 by using barium co-precipi- tation with alpha spectroscopy (Sill, 1987) could also produce a better signal to background ratio. How- 300 Fishery Bulletin 92(2). 1994 ever, our measurements of Ra-226 background were similar to those of Fenton et al. (1991). The low activity of Ra-226 and Pb-210 in the otoliths also requires that care be taken to avoid contamination. The otoliths were stored in 50% ethanol at sea. The grade of the ethanol was not ultra-pure which could introduce contaminating radioisotopes. Ethanol may also leach out some of the radioisotopes during storage. Some previous fish age-validation studies using Pb-210/Ra-226 (Fenton et al., 1990) stored the otoliths dry with desiccated adhering tissues. The dry tissue proved very diffi- cult to remove and introduced Po-210 contamina- tion. We relied on ultrasonic cleaning to break up any soft hydrated tissue after which the otoliths were rinsed thoroughly. This appeared to remove any organic material on the surface of the otolith. Contamination by Po-210 in the adherent tissues would increase the estimated radiometric age. In our study, the results from age category 1 indicate that this was not the case. Also, the grinding paper used could also introduce contamination. Conclusions The goal of this study was to validate the break-and- burn otolith ageing criteria used for sablefish at the AFSC. Although usually lower, the radiometric ages were within two standard deviations of the break- and-burn ages (Table 1; Fig. 2). If a range of pos- sible ages occurred in the break-and-burn method, the older age was usually chosen. Since radiomet- ric ages were consistently lower than burnt cross- section ages (from Reader 1), this method was prob- ably not the best way to interpret the broken-and- burnt otolith. Possible migration of Rn-222 from the otolith, as discussed earlier, could also explain the difference. Nevertheless, the break-and-burn method of age determination for sablefish has been validated by this study for fish up to 34 years old. This is the maximum age regularly seen in commercial or re- search catches. We have shown that the break-and- burn ageing criteria applied to sablefish otoliths pro- duces, on average, ages similar to radiometric ages. Acknowledgments We thank age readers from the Pacific Biological Station, Canadian Department of Fisheries and Oceans, and the Alaska Fisheries Science Center, NMFS, for providing sablefish ages. We thank the Scientific Editor and three anonymous reviewers for comments which led to significant improvements in the paper. Literature cited Bacon, M. P., D. W. Spencer, and P. G. Brewer. 1976. 210Pb/226Ra and ""Po/^Pb disequilibria in seawater and suspended particulate matter. Earth Planet. Sci. Lett. 32:277-296. Beamish, R. J., and D. E. Chilton. 1982. Preliminary evaluation of a method to deter- mine the age of sablefish (Anoplopoma fim- bria). Can. J. Fish. Aquat. Sci. 39:277-287. Beamish, R. J., G. A McFarlane, and D. E. Chilton. 1983. Use of oxytetracycline and other methods to validate a method of age determination for sablefish. In Proceedings of the second Lowell Wakefield fisheries symposium, Anchorage, AK, p. 95-116. Alaska Sea Grant Report 83-8. Bennett, J. T., G. W. Boehlert, and K. K. Turekian. 1982. Confirmation of longevity in (Sebastes diploproa) (Pisces: Scorpaenidae) from 210Pb/226Ra measurements in otoliths. Mar. Bio. 71:209-215. Bracken, B. E. 1983. Sablefish migration in the Gulf of Alaska based on tag recoveries. In Proceedings of the second Lowell Wakefield fisheries symposium, Anchorage, AK, p. 185-190. Alaska Sea Grant Report 83-8. Campana, S. E., and J. D. Neilson. 1985. Microstructure of fish otoliths. Can. J. Fish. Aquat. Sci. 42:1014-1032. Campana, S. E., K. C. T. Zwanenburg, and J. N. Smith. 1990. 210Pb/226Ra determination of longevity in redfish. Can. J. Fish. Aquat. Sci. 47:163-165. Cochran, J. K. 1982. The oceanic chemistry of the U- and Th-se- ries nuclides. In M. Ivanovich and R. S. Harmon (eds.), Uranium series disequilibrium: applications to environmental problems, p. 384-430. Oxford Univ. Press, NY. Dodge, R. E., and J. Thomson. 1974. The natural radiochemical and growth records in contemporary hermatypic corals from the Atlantic and Caribbean. Earth Planet. Sci. Lett. 23:313-322. Evans, R. D., A. F. Kip, and E. G. Moberg. 1938. The radium and radon content of Pacific Ocean water, life, and sediments. Am. J. Sci. 36:241-259. Faure, G. 1986. Principles of isotope geology. John Wiley and Sons, New York, NY, 589 p. Fenton, G. E., D. A. Ritz, and S. A. Short. 1990. 210Pb/226Ra disequilibria in otoliths of blue gTenadier, Macruronus novaezelandiae: problems associated with radiometric ageing. Aust. J. Mar. Freshwater Res. 41:467-473. Fenton, G. E., S. A Short., and D. A. Ritz. 1991. Age determination of orange roughy, Hoplostethus atlanticus (Pisces: Trachichthyidae) Kastelle et al.: Anoplopoma fimbria age validation 301 using 210Pb:226Ra disequilibria. Mar. Bio. 109:197-202. Goreau, T. F., and N. I. Goreau. 1960. The physiology of skeleton formation in cor- als. IV: On isotopic equilibrium exchanges of cal- cium between corallum and environment in living and dead reef-building corals. Biol. Bull. Mar. Biol. Lab., Woods Hole 119:416-427. Heyraund, M., and R. D. Cherry. 1979. Polonium-210 and lead-210 in marine food chains. Mar. Biol. 52:227-236. Hunter, J. R., B. J. Macewicz, and C. A. Kimbrell. 1989. Fecundity and other aspects of the reproduc- tion of sablefish (Anoplopoma fimbria) in central California waters. CalCOFI Rep. 30:61-72. Kastelle, C. R. 1991. Radioisotope age validation and an estimate of natural mortality from the gonad to somatic weight index for sablefish (Anoplopoma fimbria). M.S. thesis, Univ. Washington, School of Fisheries, 87 p. Kendall, A. W., and A. C. Matarese. 1987. Biology of eggs, larvae, and epipelagic juve- niles of sablefish (Anoplopoma fimbria) in relation to their potential use in management. Mar. Fish. Rev. 49(1): 1-13. Kimura, D. K., and J. J. Lyons. 1991. Between-reader bias and variability in the age-determination process. Fish. Bull. 89:53-60. Lucas, H. F. 1957. Improved low-level alpha scintillation counter for radon. Rev. Sci. Instruments 28(9): 680-683. McFarlane, G. A., and R. J. Beamish. 1983. Preliminary observations on the juvenile bi- ology of sablefish (Anoplopoma fimbria) in waters off the west coast of Canada. In Proceedings of the second Lowell Wakefield fisheries symposium, Anchorage, AK, p. 119-135. Alaska Sea Grant Report 83-8. 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. Moore, W. S., S. Krishnaswami, and S. G. Bhat. 1973. Radiometric determinations of coral growth rates. Bull. Mar. Sci. 23(2):157-176. Mugiya, Y. 1977. Effect of acetazolamide on the otolith growth of goldfish. Bull. Jpn. Soc. Sci. Fish. 43(9):1053-1058. Porntepkasemsan, B., and A. E. Nevissi. 1990. Mechanism of radium-226 transfer from sedi- ments and water to marine fishes. Geochem. J. 24:223-228. Sarmiento, J. L., D. E. Hammond, and W. S. Broecker. 1976. The calculation of the statistical counting error for 222Rn scintillation counting. Earth Planet. Sci. Lett. 32: 351-356. Seber, G. A. F. 1982. The estimation of animal abundance and related parameters. Macmillan, New York, NY, 654 p. Shannon, L. V., and R. D. Cherry. 1971. Radium-226 in marine phytoplankton. Earth Planet. Sci. Lett. 11:339-343. Shannon, L. V., R. D. Cherry, and M. J. Orren. 1970. Polonium-210 and lead-210 in the marine environment. Geochim. Cosmochim. Acta 34:701-711. Shenker, J. M., and B. L. Olla. 1986. Laboratory feeding and growth of juvenile sablefish, Anoplopoma fimbria. Can. J. Fish. Aquat. Sci. 43:930-937. Sill, C. w. 1987. Determination of Radium-226 in ores, nuclear wastes and environmental samples by high-resolu- tion alpha spectrometry. Nucl. Chem. Waste Man- agement 7:239-256. Smith, J. M., R. Nelson, and S. E. Campana. 1991. The use of Pb-210/Ra-226 and Th-228/Ra-228 dis-equilibria in the ageing of otoliths of marine fish. In P. J. Kershaw and D. S. Woodhead (eds.). Radionuclides in the study of marine processes, p 350-359. Elsevier Applied Science, England. Swanson, S. M. 1985. Food-chain transfer of U-series radionuclides in a northern Saskatchewan aquatic system. Health Phys. 49(5):747-770. Szabo, B. J. 1967. Radium content in plankton and sea water in the Bahamas. Geochim. Cosmochim. Acta 31:1321-1331. Veeh, H. H., and W. C. Burnett. 1982. Carbonate and phosphate sediments. In M.Ivanovich and R. S. Harmon (eds.), Uranium series disequilibrium: applications to environmen- tal problems, p. 459^180. Oxford Univ. Press, NY. Welden, B. A. 1984. Radiometric verification of age determination in elasmobranch fishes. M.S. thesis, Calif. State Univ., Hayward, Moss Landing Marine Labs, 108 p. Abstract. — This study quanti- fied the temporal and spatial abundance of juveniles of two Photololigo species on the conti- nental shelf off Townsville, Austra- lia with the use of light-traps. The two Photololigo species (A and B) showed very distinct and separate spatial distribution patterns. Photololigo sp. A was found close to the coast and was the smaller and more abundant of the two spe- cies. This species was most abun- dant in surface waters, although larger individuals were generally caught deeper. There was no evi- dence of vertical movements dur- ing the night. The presence of small and large juvenile Photololigo sp. A during summer and winter months suggests spawning and recruitment occur throughout the year. In contrast, Photololigo sp. B was caught pre- dominantly offshore. All sizes of Photololigo sp. B were caught both near the benthos and at the sur- face in the mid-lagoon, but farther offshore juveniles were deeper and larger. The presence of small juve- nile squid of both species through- out the summer suggests that these species spawn for an ex- tended period during the summer. This study demonstrates that light-traps are an effective way of sampling small cephalopods. Distribution and abundance of two juvenile tropical Photololigo species (Cephalopoda: Loliginidae) in the central Great Barrier Reef Lagoon Natalie A. Moltschaniwskyj Department of Marine Biology, James Cook University of North Queensland Townsville, Queensland 481 I, Australia. Peter J. Doherty Australian Institute of Marine Science, PMB 3 Townsville, Queensland 4810, Australia Manuscript accepted 8 November 199.3 Fishery Bulletin 92: 302-312 (1994) The current poor state of knowl- edge about processes important in squid population dynamics is mainly due to limited information about the juvenile phase (Voss, 1983; Boyle, 1990). Life-history characteristics have largely been derived from information about the adult phase. Our limited informa- tion about young squid is demon- strated in attempting to define the life-history phases (Young and Harman, 1988). Jackson and Choat ( 1992) suggest, given the compara- tively short life time of tropical squid (<250 days), that a propor- tionally long period of the life cycle is spent as small individuals. In the case of Loligo chinensis, with a summer life time of 120 days, indi- viduals less than 60 days old (<50- mm mantle length) have not been studied. Hence, for almost half the life history of most squid there is not even the most basic informa- tion. Temporal and spatial abun- dance patterns of juvenile squid will provide a basis for understand- ing the processes of mortality, growth, and recruitment. However, such information has traditionally been difficult to obtain because of problems in capturing and identify- ing a sufficient size range of juvenile cephalopods (Vecchione, 1987). To examine the ecology of juve- nile squid it is necessary to use techniques that catch a size range of individuals, hatchlings to juve- niles, in good condition. Pelagic squid produce either benthic or pelagic eggs and have a planktonic juvenile phase (Boletzky, 1977). Juvenile squid are alert, mobile organisms that easily avoid capture by towed nets (Vecchione, 1987). The use of a combination of differ- ent towed nets to sample an area enables the collection of a wider size range of juvenile squid (Rod- house et al., 1992). However, it is difficult to obtain replicates needed to provide density estimates from towed nets. In this study we have employed an alternative technique based on light-attraction that is effective in sampling pelagic juve- nile fishes. Automated light-traps (Doherty, 1987) can overcome the problems of net avoidance and en- able sampling at discrete depths in the water column. The ability to sample concurrently within an area ensures that estimates of variabil- ity in abundance are not con- founded by time. This technique also collects live material in good condition, which can facilitate taxo- nomic identification. However, sampling an unknown volume of 302 Moltschaniwskyj and Doherty: Distribution and abundance of juvenile Photololigo 303 water by individual traps requires cautious interpre- tation of abundance estimates (Choat et al., 1993). There are four species of loliginid squid currently recognized in the Townsville region: Sepioteuthis lessoniana, Loliolus noctiluca, Photololigo sp. B, and Photololigo sp. A.1 There are currently no morpho- logical descriptions of the two Photololigo species, but they can be readily identified by using allozyme electrophoretic techniques (Yeatman and Benzie, in press). Previously both of these species have been referred to as Photololigo (Loligo) chinensis (Jack- son and Choat, 1992; Yeatman and Benzie, in press), but neither correspond to P. chinensis from Thai- land.2 Electrophoretic analysis of a subset of juve- niles collected during three months of the program found that all Photololigo sp. A were found less than 33 km offshore and 90% of the Photololigo sp. B were found 33 km or more offshore.2 Because these species are morphologically identical as juveniles, we assumed that all individuals found at stations less than 33 km offshore were Photololigo sp. A and that Photololigo collected more than 33 km offshore were Photololigo sp. B. Photololigo sp. A (previously known as Loligo chinensis) has been the topic of recent growth studies using statolith aging tech- niques (Jackson and Choat, 1992). This species is a small short-lived neritic squid. Individuals are ap- proximately 60 days old when they appear in the adult population and they can grow to 180 mm in 120 days. Little is known about the early life-his- tory and juvenile distribution patterns of either Photololigo species. The objectives of this study were to describe the spatial and temporal distribution patterns of juvenile Photololigo species across the continental shelf in the Townsville region of the Great Barrier Reef. Materials and methods Sampling design Two major habitat types are found on the continen- tal shelf, off Townsville, Australia. The inshore habi- tat is a 56 km wide soft bottom coastal lagoon rang- ing in depth from 15 m to 40 m. The offshore habi- tat is a complex reef matrix of similar extent, dis- sected by channels ranging from 40 m to 75 m deep at the shelf break. To assess the cross-shelf distri- bution of juvenile squid, four automated light-traps (Doherty, 1987) were deployed at fifteen sampling stations spanning the continental shelf and the 1 C. C. Lu, Museum of Victoria, Australia, pers. commun. 1990. 2 J. Yeatman, James Cook Univ., Australia, unpubl. data 1993. western Coral Sea (Fig. 1). Abundance along this transect was assessed over four months, October to January, during two austral summers, 1990/91 and 1991/92. At each station, the abundance of juvenile squid was determined at two depths by deploying two pairs of light-traps. In each pair, one light-trap was suspended immediately below the surface while the other light-trap was set deeper. In 1990/91, all deep light-traps were suspended 20 m below the surface. In 1991/92, the deep light-traps were sus- pended within 5 m of the bottom to a maximum of 100 m in the Coral Sea. In all deployments, the two pairs of light-traps were released approximately 300 m apart and al- lowed to drift for one hour. Allowing the traps to drift in the water minimized potential problems with differential water movement among stations. The use of drifting light-traps has been shown to be a more effective way of catching pelagic organisms than anchored light-traps in open water (Thorrold, 1992). After one hour, the four light-traps were re- trieved and the entire catch was fixed and preserved in 100% ethanol. Each evening the first light-trap was deployed after 1930 hours (Eastern Standard Time) and the last light-trap retrieved before 0430 hours. Travel time between each station al- lowed only five cross-shelf stations to be sampled per night. Thus, each night's activity concentrated on one of the two continental shelf habitats or the Coral Sea. Each monthly cruise consisted of nine nights during which time each of the 15 stations was sampled three times. However, sea conditions were not always favorable. Sampling effort at each sta- tion is shown in Table 1. It was not logistically possible to sample all sta- tions in each habitat simultaneously. Therefore, time of night is confounded with station position. Hap- hazard selection of the first station sampled each night ensured that no station was consistently sampled at the same time on all nights. Cruises were scheduled to include the new moon because this is the lunar phase when light attraction has proved most effective for fishes and various inver- tebrates (Milicich, 1992). Temperature and salinity profiles of the water column were collected at each station by using a Seabird Conductivity Tempera- ture Device during the 1991/92 summer. Concurrent with the summer cross-shelf sam- pling, light-traps were anchored within 100 m of the southeasterly side (weather-side) of four reefs; Keeper, Helix, Faraday, and Myrmidon, to sample near-reef water (Fig. 1). The use of drifting light- traps near the reefs was not possible. During the summer of 1990/91, four light-traps were anchored at each reef; three immediately below the surface 304 Fishery Bulletin 92(2), 1994 19°00'S QUEENSLAND 0 10 20 30 kilometres 146WE Figure 1 Map of the cross-shelf transect off Townsville, Australia, showing the position of each station along the transect. Station 1 = 19 km from Townsville, 2 = 24 km, 3 = 33 km, 4 = 43 km, 5 = 52 km, 6 = 61 km, 7 = 75 km, 8 = 92 km, 9 = 100 km, 10 = 115 km, 11 = 136 km, 12 = 145 km, 13 = 152 km, 14 = 163 km, and 15 = 172 km. and one at 20 m below the surface. In 1991/92, an extra light-trap was added at 20 m. The anchored light-traps had an automatic timer, enabling the lights to be switched on and off automatically at predetermined periods during the night. Each light- trap on the reef fished for a total of three hours per night; lights came on for one hour at 2200 hours, 2400, and 0300 hours. Light-traps at all reefs were emptied the following day. Squid were identified in the laboratory and the dorsal mantle length recorded for each individual. Individuals were measured within 14 days of pres- ervation in 100% ethanol. A comparison of measure- ments of individuals (ranging in size from 5.3 mm to 29.5 mm) before and 14 days after preservation found that shrinkage was on average 0.5 mm. Abundance patterns of the two Photololigo species during the two summers of sampling were examined by using 'planned comparisons,' where specific pregenerated hypotheses were examined (Day and Quinn, 1989). For each species we were interested in differences in abundance between years, loca- tions, and depths. To examine seasonality of juvenile Photololigo sp. A, the inshore station (19 km) was sampled during the austral winter months of May, June, July, and August 1991. Three sites at this station were sampled with four shallow and four deep (13-m) light-traps. Sites were sampled during the period of the new moon, on five nights in May and three nights in June, July, and August. Densities in sum- mer and winter months were compared by using an unbalanced one-way analysis of variance (ANOVA), with month as the factor analyzed. Values in each light-trap for nights and sites within a month were treated as replicates. To determine whether vertical migration might influence horizontal distribution patterns we exam- ined the size structure of Photololigo sp. A at two depths during the night. On at least one occasion Moltschaniwskyj and Doherty: Distribution and abundance of juvenile Photololigo 305 Table 1 Total sampli ng effort for Photololigo spp . in each month in light-trap hours (and nc mber of nights sampled) at each station during the two summers of sampling Year and Distance (km) from Townsville Total month 19 24 33 43 52 HI 75 92 100 115 136 145 152 163 172 sampled 1990 Oct 8(2) 15(4) 16(4) 12(3) 16(4) 15(4) 4(1) 4(1) 4(1) 4(1) 12(3) 10(3) 10(3) 10(3) 10(3) 150(40) Nov 12(3) 12(3) 12(3) 0(0) 16(3) 12(3) 12(3) 12(3) 12(3) 12(3) 8(2) 8(2) 8(2) 4(1) 4(1) 144(35) Dec 8(2) 8(2) 4(1) 4(1) 4(1) 4(1) 4(1) 3(1) 4(1) 4(1) 4(1) 4(1) 4(1) 4(1) 4(1) 67(17) 1991 Jan 12(3) 12(3) 12(2) 12(2) 12(2) 8(2) 8(2) 8(2) 8(2) 8(2) 8(2) 8(2) 4(2) 8(2) 8(2) 136(32) Oct 12(3) 12(3) 12(3) 12(3) 12(3) 12(3) 8(2) 8(2) 8(2) 8(2) 4(1) 4(1) 4(1) 4(1) 4(1) 124(31) Nov 12(3) 12(3) 11(3) 12(3) 12(3) 12(3) 12(3) 12(3) 12(3) 12(3) 4(1) 4(1) 4(1) 4(1) 4(1) 139(35) Dec 12(3) 10(3) 11(3) 12(3) 12(3) 12(3) 8(2) 8(2) 8(2) 8(2) 4(1) 4(1) 4(1) 4(1) 4(li 121(31) 1992 Jan 12(3) 12(3) 12(3) 11(3) 12(3) 12(3) 8(2) 8(2) 8(2) 8(2) 4(1) 4(1) 4(1) 4(1) 4(ll 123(31) Total 88(22) 93(24) 90(22) 75(18) 96(22) 87(22) 64116) 63(161 64(161 64(16) 48(12) 46(12) 42(12) 42(11) 42(11) 1004(254) in each month of the 1991/92 sampling period the 19- and 24-km stations were sampled both early and late in the night. The samples were separated into early (captured before 2400 hrs) and late (captured after 2400 hrs). By combining data from stations, across nights and months, it was possible to com- pare the size distributions between depths and time of night. A multiway-frequency analysis was used to determine the effect of time of night and depth on the size-frequency distribution. Results Distribution patterns Juvenile Photololigo individuals were predominantly caught within 52 km of the mainland (Fig. 2). The few individuals found farther offshore were in the Magnetic Passage (five individuals) and on the reefs (six individuals). Photololigo species were not found in the Coral Sea. Photololigo sp. A was numerically the most abundant of the two species during both summers (Fig. 2); 856 individuals were caught in 181 hours of light-trapping (4.73 individuals caught per hour), compared with 379 Photololigo sp. B caught in 348 hours of light-trapping ( 1.09 individu- als per hour). Catch per hour of light-trapping was greatest for Photololigo sp. A, especially at the 24- km station. The catch per unit of effort for Photo- loligo sp. B was greater at the 33-km station (Table 2). Overall, Photololigo sp. A juveniles were present in higher numbers at the 24-km station in the sur- face waters (Table 3). This pattern was consistent in both years, but higher numbers were caught in 1991/92 (Table 3), largely because of very high catches in December 1991 (Fig. 2). In comparison, highest numbers of Photololigo sp. B were consis- tently found at the 33-km station and abundance levels tended to decrease farther offshore (Fig. 2). Overall, Photololigo sp. B demonstrated no differ- ence in abundance levels between the two years (Table 4). In contrast to Photololigo sp. A, juvenile Photololigo sp. B was more abundant deeper in the water column (Table 4). Farther offshore, Photololigo sp. B juveniles were present in very low numbers and were caught only in the deep light-traps (Fig. 2). Photololigo sp. A ranged in size from 2.6 to 47.9 mm. The size-frequency distributions at the two depths were not significantly different between the 19-km and 24-km stations (x2=12.28; df=9; P=0.1979) (Fig. 3). There was no systematic change in the size-frequency distribution of Photololigo sp. A during either summer (Fig. 4). A modal shift in the size-frequency distribution in January 1992 suggested that fewer small individuals were avail- able to be caught. However, catches were very low in this month. Photololigo sp. B ranged in size from 3.6 to 61.6 mm (Fig. 3). From the size-frequency distributions it was clear that larger juveniles were found farther offshore and deeper in the water column (Fig. 3). No 306 Fishery Bulletin 92(2). 1994 modal shift in the size-frequency distribution dur- ing the summers was apparent (Fig. 4). However, catches were low in most months. The multiway-frequency analysis established that the size-frequency distribution of juvenile Photololigo sp. A at both depths changed as a func- tion of time of night (Table 5). Small juveniles domi- nated in the surface waters, but larger individuals were generally found closer to the benthos (Fig. 5). During the night, the relative abundance of small individuals decreased at both depths. Close to the UJ 00 0_ 5 I & _J a. UJ a. 01 —I < Q > D Z Li. O or UJ CD z z s 5 5 J SUMMER 1990/91 4 3 2 1 0 20 f ^ 15 10 I 5 ' 0 30 r 25 i 20 15 10 5 ;, ^ ^_ 20 15 10 5 0 s "--i__ 20 D A 30 40 50 60 70 80 90 100110 sp B SUMMER 1991/92 10 8 6 4 7 0 12 10 8 6 4 2 0 60 50 40 30 20 10 0 , OCTOBER , . NOVEMBER I 20 30 40 50 60 70 80 90 100110 sp A sp B DISTANCE OFFSHORE (km) Figure 2 Catches of juvenile Photololigo sp. A (found at 19 and 24 km) and Photololigo sp. B (found at 33 km and greater) from Tbwnsville, Australia. Most values are averages (+ standard error) of six one-hour sets over three nights. See Table 1 for replicates at each station. (Solid lines, deep light-traps; dashed lines, shallow light-traps). Note the variable scale of the Y-axes. benthos an increase in large individuals was evi- dent. There was no discernible pattern of vertical migration; however, combining data across months to increase the number of juveniles in the analysis removed the possibility of detecting vertical migra- tion in any one month. The number of Photololigo sp. A juveniles cap- tured during the winter months was similar to most of the summer monthly catches (Fig. 6); although winter catches never reached levels such as those seen in December 1991 (Table 6). The large num- ber of small juveniles captured over the winter (Fig. 6) indicates that Photololigo sp. A spawns and hatches in both seasons. A simi- lar size range was captured at each sampling during the summer months (Fig. 7). Physical parameters Both temperature and salinity decreased nonlinearly across the lagoon; discontinuities in both variables occurred midway across the Lagoon (Fig. 8). Temperature or salinity discontinuities were detected on at least six out of nine nights between the 33-km station and one or both of the neigh- bouring stations. This suggested that in the lagoon the water mass was heterogenous and may have influenced the distribution pat- terns of juvenile squid. Salinity-temperature profiles of the water column at each station indicated thermoclines were present on some nights (Table 7). A thermocline was defined as a temperature change greater than 0.5°C between surface and bottom water; differences as great as 3°C were detected during January. However, these thermoclines were a temporally and spatially un- stable feature of the water col- umn, possibly due to variable wind conditions and the shallow body of water being sampled. DECEMBER JANUARY Discussion Light-traps have provided a tech- nique by which spatio-temporal distribution patterns of two Pho- Moltschaniwskyj and Doherty. Distribution and abundance of juvenile Photololigo 307 tololigo species can be described. Identification of Photololigo spe- cies using allozyme electrophore- sis suggests that the two species are separated geographically across the Great Barrier Reef Lagoon (Yeatman and Benzie, in press). This separation occurs in a region of the coastal lagoon where temperature-salinity data indicate heterogeneity. High num- bers of juvenile Photololigo sp. A at stations close to the mainland suggests that spawning grounds for this species may be close to the coast, a feature typical for loliginid squid (Mangold, 1987). Furthermore, the presence of small and large individuals dur- ing summer and winter months indicates that spawning, hatch- ing, and recruitment are not sea- sonal events. This characteristic may be more common for tropical species that tend to have shorter lifespans than temperate species (Jackson and Choat, 1992). Large numbers of small juveniles col- lected during the winter may be a function of slower growth dur- ing the winter (Jackson and Choat, 1992). Little is known about Photololigo sp. B adults; however, the presence of juveniles in this region suggests that an adult population does occur in the Townsville region and that spawning occurs throughout the summer. The identification of ju- venile Photololigo was confirmed on a subsample of specimens cap- tured during the summer. Conclu- sions drawn from this study are based upon the assumption that the offshore distribution pattern of the two species was consistent in all other months of sampling. Juvenile squid are not easily sampled with towed nets (Vecchione, 1979; Vecchione and Gaston, 1985; Holme, 1974). They have highly developed sensory and locomotor sys- tems (Boletzky, 1974) and it is likely that these animals are often undersampled because of net avoidance. Choat et al. (1993) have shown that plankton nets select for small larval fish, but larger Table 2 Catch per hour of 1 ight-trapping for each Photololigo species across the Great Barrier Reef Lagoon for eight month s of summer sampling. Photol oligo sp. A at stations 19 km and 24 km and Photololigc sp. B farther offshore. Month Species A Species B Total 19km 24km 33km 43km 52km 61km 1990 Oct 2.38 1.81 0.25 0 (i 0 0.62 Nov 0.75 8.25 5.42 — 0 0 2.88 Dec 1.88 8.00 0.50 (I i) 0.75 2.63 1991 Jan 0.42 1.33 9.42 3.58 1.00 1.13 2.91 Oct 1.67 2.17 4.25 0.25 0.67 0 1.50 Nov 1.83 4.83 2.17 0.42 0.50 0.25 1.67 Dec 10.67 24.42 1.92 0.17 0.50 0 6.28 1992 Jan 0.83 1.75 1.67 0.17 0.08 0 0.75 Total 2.59 6.31 3.30 0.72 0.36 0.17 2.33 Table 3 Planned comparisons of juvenile Photololigo sp. depths, years, and sites. A densities between Contrast df Contrast sums of squares Mean squares F-value P>F Depths 1 Years 1 Sites 1 Residual 177 9.8165 3.7565 8.6892 142.3838 9.8165 3.7565 8.6892 0.8044 12.20 4.67 10.80 0.0006 0.0320 0.0012 Planned comparisons depths and years. Table 4 of juvenile Photololigo sp. B densities between Contrast df Contrast sums of squares Mean squares F-value P>F Depths 1 Years 1 Residual 335 17.0607 0.0438 148.7448 17.0607 0.0438 0.4507 37.85 0.10 0.0001 0.7554 fish are captured from the same water column by using light attraction. Thorrold (1992), as well as this study, showed that light-traps are a useful tech- nique for capturing juvenile squid. However, like most sampling techniques, the light-traps have bi- ases. One problem is that light-traps sample an unknown volume of water. Nonetheless, they have 308 Fishery Bulletin 92(2), 1994 Summer 1990/91 Photololigo sp. A Summer 1991/92 19km [n=171| 24 km [n=395] 43 km ►a ) 43 km |n-43] 40 ^^ [n=srj 20 0 a L _ 25 75 125 175 225 275 325 375 425 »45 " 75 125 175 22 5 27 5 325 375 42 5 »45 MANTLE LENGTH (mm) Figure 3 Size-frequency distribution of Photololigo sp. A and Photololigo sp. B caught at each station (pooled across months) in deep (shaded) and shallow (unshaded) light-traps. Total number of juveniles indicated in brackets. Table 5 Results of the multiway frequency analysis to examine changes in the size distribution of Photololigo sp. A between time of night and depth. Source df Depth Time Depth x Time 92.8 25.57 0.19 0.00 0.00 0.66 been validated as useful devices for monitoring rela- tive abundance patterns in larval supply of pelagic juvenile fish at fixed locations (Milicich et al., 1992). Great care needs to be exercised when interpreting catch rates from different locations because changes in water transparency can bias light-trap efficiency. Similarly, it is not possible to quantitatively compare catches from drifting and anchored light-traps (Thorrold, 1992). This is because the former act as lagrangian drifters and sample photopositive organ- isms from within a constant light pool. In contrast, the moored light-traps experience a variable water flow that may greatly increase the volume of water swept in an hour of sampling. Despite more inten- sive sampling on the reefs, catches of Photololigo were low and we conclude that spawning does not occur near the reefs and that juvenile Photololigo individuals are concentrated in the lagoon. In the Moltschaniwskyj and Doherty: Distribution and abundance of juvenile Photololigo 309 Summer 1990/91 Photololigo sp A Summer 1991/92 Photololigo sp. B Summer 1990/91 Summer 1991/92 October >- O z LU O LU cr LL 10 15 20 25 30 35 40 '40 li. 10 i ^_ December 8 J _L i ml Jl_ -] BO -| 10 -| 0 5 10 15 20 25 30 35 40 >40 0 5 10 15 20 25 30 35 40 >40 0 5 10 15 20 25 30 35 40 >40 DORSAL MANTLE LENGTH (mm) Figure 4 The size-frequency distribution of juvenile Photololigo sp. A and Photololigo sp. B during eight months of sum- mer sampling. Size classes are mid-points of each class. Data are pooled across depths and stations. present study, a gradient of tur- bidity across the shelf makes it possible that inshore catches would underestimate abundance if cor- rected for diminishing light-pools. However, if the error was signifi- cant, it would only exaggerate, not diminish, our observation that ju- venile squid were more abundant within the coastal lagoon. High catches of juvenile squid in the coastal lagoon were at lo- cations where discontinuities were often observed in surface temperature and salinity. Hydrodynamic modelling of this region suggests that the coastal lagoon is often subject to velocity shear (King and Wolanski, 1992). Water in the lagoon typically flows southward under the influence of the poleward East Australian Current, which pushes water onto the outer shelf and through the reef matrix, especially through channels like the Magnetic Passage. Under typical south-easterly wind conditions the shallow body of water trapped against the coast moves in the opposite direction, northwards. The result is a ve- locity shear between the two water masses and a Table 6 Analysis of variance examining differences between densities of Photololigo sp. A at the 19 km station between summer months of 1990/ 91 and 1991/92 and winter months of 1991. Contrast sums Mean Source df of squares squares F-value P>F Month 11 1118.200 101.654 Residual 214 2277.910 10.644 9.55 0.0001 zone of low residual displacement. Modelling stud- ies suggest that the cross-shelf location of this fea- ture, referred to as a separation front (King and Wolanski, 1992), will shift seawards as the wind strength increases and vice versa. This mobility of the frontal region is consistent with the daily and monthly variability of salinity and temperature at the surface indicated by our physical monitoring during the second summer. This low-shear zone is identified as a significant place for aggregation of planktonic organisms. Cross-shelf studies have shown highest abundances of larval reef fishes in a similar location near the 310 Fishery Bulletin 92(2). 1994 center of the Great Barrier Reef Lagoon (Thorrold, in press). These catches included individuals taken from reefs farther offshore, as well as piscivorous larvae of various scombrids from inshore (Thorrold, 1993). It is not clear whether aggregation of these stages is passive, due to hydrodynamics, or the re- sult of attraction to the coastal boundary area by enhanced secondary productivity in this frontal zone EARLY SHALLOW 13 TRAPS LATE SHALLOW 10 TRAPS > a Z HI D O UJ or DORSAL MANTLE LENGTH (mm) Figure 5 Size-frequency distributions of juvenile Photololigo sp. A from the two inshore stations at two sampling depths (pooled across the summer months 1991/92), captured early (before 2400 hr) and late (after 2400 hr) in the night. K 161 O 14 -■~ 12 or uj £ o Z >Z=JO oouj< < d 2 3 o z o -> s->^< H > o Z O o "J < O z Q ~> 1990 1991 1992 Figure 6 Catches of juvenile Photololigo sp. A at the 19-km station over twelve months: during summer 1990/ 91, winter 1991, and summer 1991/92 (Data pooled across depth and nights.) (Thorrold and McKinnon, 1992). This discontinuity may be a mechanism separating the two Photololigo species geographically. The separation of juvenile cephalopod species in the Gulf Stream east of New England is thought to be closely related to meso- scale hydrological features (Vecchione and Roper, 1986). The importance of hydrological features in ag- gregating juvenile squid has been identified in a number of species (Rodhouse and Clarke, 1985; Brunetti and Ivanovic, 1992; Rodhouse et al.„ 1992). This suggests that these areas are ecologi- cally important for juvenile squid. The second way in which shelf-scale hydrodynamics affects the stability of the water column is the intrusion of upwelled waters from the shelf-break driven onto the shelf by variations in the speed and position of the East Australian Current. These cold intru- sions can be tracked into the Great Barrier Reef Lagoon (King and Wolanski, 1992) and the strong ther- mal stratification observed in Janu- ary 1992 was consistent with an in- trusion at this time. A cold bottom layer at 33 km was evident on one night in November, but the inner sta- tions were not stratified. The pres- ence of juvenile Photololigo at most stations in all months, despite a range of physical conditions, suggests juve- nile Photololigo can tolerate substan- tial environmental variation. This tol- erance is consistent with a nonsea- sonal reproductive strategy, which is essential for a species that lives for only four months. During the night there was little evidence of a pronounced vertical migration such as the mass aggregations of juvenile Loligo spp. on the benthos (Vecchione and Gaston, 1985) or the general move- ment to the surface by juvenile L. pealei (Vecchione, 1981). The absence of vertical movement during the night suggests that the observed ontogenetic shift of Photololigo sp. B farther offshore and deeper is real and not a product of location confounded with time of night when sampling occurred. However, as was noticed in the catch-per-unit-of-effort values, both species are caught in relatively low numbers; hence, conclusions based on small differences that are not significantly different are limited. There was a problem with low numbers in all spatial and tem- poral trends described. However, this was a prelimi- nary study with just two hours of sampling at each station per night. More intensive sampling in bound- Moltschaniwskyj and Doherty: Distribution and abundance of juvenile Photololigo 31 1 Table 7 Depth of the thermocline m ) at each station on each night of sampling during the three months of the 1991/92 summer. Depth of thermocline Sampling period and situation Day 1 Day 2 Day 3 October 1991 19 km 14 11 Absent 24 km 13 14 Absent 33 km 10 13 Absent 43 km in Absent Absent 52 km 25 Absent Absent 61 km 31 Absent — November 1991 19 km Absent Absent Absent 24 km Absent Absent Absent 33 km Absent Absent 20 43 km Absent Absent 22 52 km Absent Absent 25 61 km Absent Absent 29 January 1992 19 km 7 7 9 ■2 4 km 9 9 9 33 km 11 13 12 43 km 15 15 18 52 km 2 s 24 27 61 km 31 17 — 60 40 20 0 60 UJ O < 40 1- z UJ 20 o or LU 0 Q. 60 40 20 SUMMER 1990/81 n=47 WINTER 1991 n=149 SUMMER 1991/92 n=182 2 5 7 5 12.5 17 5 22.5 27 5 32.5 37 5 MID-POINTS OF SIZE CLASSES (mm) Figure 7 Size-frequency distributions of juvenile Photololigo sp. A at the 19-km station during the summer and winter months. (Numbers are pooled across months, depth, and nights.) 26 255 O = o, LU AT 27 5 D 1- t£ 265 LU Q. LU 255 H 31 30 29 28 Surfac( the Gr Octobe tivityl 1-Oc» 3-OCt 1 9-Oct 354 352 35 CO 36 > z 356 < 352 358 354 3-Nov 6-Nov 12-Nov 3-Jan 8-Jan n^Jan 3 20 40 60 8 DIS i temperature eat Barrier R< r and Novembx emperature D( 0 ;ta da sef sr ■VII D 20 40 60 f NCE OFFSHOR Figure 8 shed) and sali Lagoon for et .991 and in Ja e failed during 10 E (km nity ich r nuar ;the ) 20 40 60 8 ) solid) profiles light of samp y 1992. The C December cm 0 across ing in onduc- ise. ary waters, both vertical and horizontal, is needed to understand how juvenile squid react to the physical environment. This study has shown that light-traps are useful devices for catching juvenile squid, providing a basis for a more intensive study of the early life-history of squid. Acknowledgments This project was funded by a FIRDC grant to P.J.D. and a Merit Research Grant to N.A.M. We thank the crew of the RV Lady Basten and volunteers who assisted with fleldwork. Iden- tification of squid was assisted by C. C. Lu. We thank Julia Yeatman for access to electro- phoretic data and discussions about species identification. This manuscript was greatly im- proved by the comments of J. H. Choat, A. Lewis, M. McCormick, the Coral Reef Discus- sion Group, and two anonymous reviewers. This research was conducted while N.A.M. was on a Commonwealth Scholarship. 312 Fishery Bulletin 92(2). 1994 Literature cited Boletzky, S. v. 1974. The 'larvae' of Cephalopoda: a review. Thalassia Jugoslavica 10:45-76. 1977. Post-hatching behaviour and mode of life in cephalopods. Symp. Zool. Soc. Lond. 38:557-567. Boyle, P. R. 1990. Cephalopod biology in the fisheries context. Fish. Res. 8:303-321. Brunetti, N. E., and M. L. Ivanovic. 1992. Distribution and abundance of early life stages of squid (Illex argentinus) in the south-west Atlantic. ICES J. mar. Sci. 49:175-183. Choat, J. H., P. J. Doherty, B. A. Kerrigan, and J. M. Leis. 1993. A comparison of towed nets, purse seine, and light-aggregation devices for sampling larvae and pelagic juveniles of coral reef fishes. Fish. Bull. 91:195-209 Day, R. W., and G. P. Quinn. 1989. Comparisons of treatments after an analysis of variance in ecology. Ecol. Monogr. 59:433-463 Doherty, P. J. 1987. Light-traps: selective but useful devices for quantifying the distributions and abundances of larval fishes. Bull. Mar. Sci. 41:423-431. Holme, N. A. 1974. The biology of Loligo forbesi Steenstrup (Mollusca: Cephalopoda) in the Plymouth area. J. mar. biol. Assoc. U.K. 54:481-503. Jackson, G. D., and J. H. Choat. 1992. Growth in tropical cephalopods: an analysis based on statolith microstructure. Can. J. Fish. Aquat. Sci. 49:218-228. King, B., and E. Wolanski. 1992. Coastal dynamics along a rugged coastline. In D. Prandle (ed.), Dynamics and exchanges in es- tuaries and the coastal zone, p. 577-598. Coastal and Estuarine Studies 74, Springer- Verlag. Mangold, K. 1987. Reproduction. In P. R. Boyle (ed.), Cepha- lopod life cycles. Vol 2: Comparative reviews, p. 157-200. Academic Press, NY. Milicich, M. J. 1992. Light-traps: a novel technique for monitoring larval supply and replenishment of coral reef fish populations. Unpubl. Ph.D. thesis, Griffith Uni- versity, 127 p. Milicich, M. J., M. G. Meekan, and P. J. Doherty. 1992. Larval supply: a good predictor of recruit- ment of three species of reef fish (Pomacentridae). Mar. Ecol. Prog. Ser. 86:153-166. Rodhouse, P. G., and M. R. Clarke. 1985. Growth and distribution of young Mesonychoteuthis hamiltoni Robson (Mollusca: Cephalopoda): an Antarctic squid. Vie Milieu 35:223-230. Rodhouse, P. G., C. Symon, and E. M. C. Hatfield. 1992. Early life cycle of cephalopods in relation to the major oceanographic features of the southwest Atlantic Ocean. Mar. Ecol. Prog. Ser. 89:183-195. Thorrold, S. R. 1992. Evaluating the performance of light traps for sampling small fish and squid in open waters of the central Great Barrier Reef lagoon. Mar. Ecol. Prog. Ser. 89:277-285. 1993. Post-larval and juvenile scombrids captured in light traps: preliminary results from the cen- tral Great Barrier Reef lagoon. Bull. Mar. Sci. 52:631-641. In press. Coupling of pre-settlement reef fish dis- tribution and hydrography in the central Great Barrier Reef lagoon. Proc. Seventh International Coral Reef Symposium, 1992. Thorrold, S. R., and A. D. McKinnon. 1992. Biological significance of the coastal bound- ary layer off Townsville, North Queensland. In D. A. Hancock (ed.), Larval biology. Australian Soci- ety for Fish Biology Workshop, Hobart 1991. Bu- reau of Rural Resources Proceedings 15:104-109. Vecchione, M. 1979. Larval development of Illex Steenstrup, 1880, in the northwestern Atlantic, with comments on Illex larval distribution. Proc. Biol. Soc. Wash. 91:1060-1075. 1981. Aspects of the early life history of Loligo pealei (Cephalopoda; Myopsida). J. Shellfish Res. 1:171-180. 1987. Juvenile ecology. In P. R. Boyle (ed.), Cepha- lopod life cycles. Vol 2: Comparative reviews, p. 61-84. Academic Press, NY. Vecchione, M., and G. R. Gaston. 1985. In-situ observations on the small-scale dis- tribution of juvenile squids (Cephalopoda: Loliginidae) on the northwest Florida shelf. Vie Milieu 35:231-235. Vecchione, M., and C. F. E. Roper. 1986. Occurrence of larval Illex illecebrosus and other young cephalopods in the slope water/Gulf Stream interface. Proc. Biol. Soc. Wash. 99:703-708. Voss, G. L. 1983. A review of cephalopod fisheries biology. Mem. Natl. Mus. Victoria 44:229-241. Yeatman, J., and J. A. H. Benzie. In press. Genetic structure and distribution of Photololigo in Australia. Mar. Biol. Young, R. E., and R. F. Harman. 1988. "Larva," "paralarva," and "subadult" in cephalopod terminology. Malacologia 29:201-207. Abstract. — The distribution and activities of rockfish, Sebastes spp., inhabiting depths between 21 and 150 m in the coastal fjord of Saanich Inlet, British Columbia, were assessed by using the Pisces PV submersible. Quillback rock- fish, S. maliger, was the numeri- cally dominant rockfish, attaining a median density of 5.7 fishlOOm"2 between 21 m and 100 m of depth. Copper rockfish, S. caurinus, tiger rockfish, S. nigrocinctus, yellowtail rockfish, S. flavidus, yelloweye rockfish, S. ruberrimus, and green- striped rockfish, S. elongatus, were all observed in consistently low densities (<1 fish-lOOm"2). The greatest densities of rockfish oc- curred over complex habitat of bro- ken rock and boulders. The major- ity (>50%) of rockfish were ob- served either perched on open sub- strate, hovering, or swimming. All rockfish species were observed near quillback rockfish (>75% oc- currence); and quillback, copper, and yellowtail rockfishes were also found in association with conspecifics. Observations on the distribution and activities of rockfish, Sebastes spp., in Saanich Inlet, British Columbia, from the Pisces IV submersible Debra J. Murie* Daryl C. Parkyn Bruce G. Clapp* Geoffrey G. Krause** Department of Biology. University of Victoria Victoria, B.C.. Canada V8W 2Y2 Prior to the advent of submersibles, in situ observations of deep-water demersal fish assemblages were constrained by time-depth limita- tions of SCUBA, which restrict ob- servations offish assemblages pri- marily to depths above 40 m (130 ft) (e.g. Moulton, 1977; Larson, 1980; Hallacher and Roberts, 1985; Richards, 1987; Murie, 1991). Dis- tributional studies of fishes inhab- iting waters deeper than 30-40 m have therefore relied on hook-and- line surveys, box trapping, or net trawling, all of which have known biases and limitations (Westrheim, 1970; Uzmann et al., 1977; Krieger, 1993). The recent availability of small submersibles for research purposes has allowed direct visual assessment of the depth distribu- tion, density, and habitat of a vari- ety of deep-water fish species (Uz- mann et al., 1977; Carlson and Straty, 1981; Richards, 1986; Dennis and Bright 1988; Pearcy et al., 1989; Stein et al., 1992; Krieger, 1993). Rockfish {Sebastes spp.) are im- portant to nearshore recreational and commercial fisheries along the northeastern Pacific coast (Patten, 1973; Richards, 1987). Many in- shore rockfish species are believed to be ecologically and morphologi- cally similar, and are primarily benthic, sedentary fishes (Patten, 1973; Moulton, 1977; Mathews and Barker, 1983; Richards, 1986, 1987; Murie, 1991). Distributions of nearshore rockfish may depend on a variety of factors, including depth, habitat, and the presence of con- and hetero-specifics. Various species are known to segregate bathymetrically (Larson, 1980; Hallacher and Roberts, 1985; Richards, 1986, 1987; Pearcy et al., 1989), reducing or eliminating the potential for competitive interac- tions between otherwise ecologi- cally similar species (Larson, 1980). Using a submersible, it is possible to observe directly the species-spe- cific depth distributions, as well as to estimate each species' numerical abundance or density with depth. Changes in density with depth within a rockfish species may ulti- mately be related to fish size be- cause rockfish size is often posi- Manuscript accepted 18 October 1993 Fishery Bulletin 92: 313-323 (1994) * Present address: Department of Fisheries and Oceans, Pacific Biological Station, Ma- rine Fish Division, Nanaimo, B.C., Canada V9R 5K6 "Present address: Explorations Unlimited, #1-1012 Richardson St., Victoria B C Canada V8W 3C5. 313 314 Fishery Bulletin 92[2), 1994 tively correlated with depth (Westrheim, 1970; Boehlert, 1980; Wilkins, 1980; Richards, 1986). Nearshore rockfish are usually found in close as- sociation with the substrate or vertical relief (e.g. kelp beds) and their density may be dependent on the type of habitat available (e.g. boulder fields, shelter holes, etc.) (Patten, 1973; Moulton, 1977; Richards, 1986, 1987; Pearcy et al., 1989; Stein et al., 1992). Depth and density distributions of near- shore rockfish in British Columbia have been as- sessed primarily by hook-and-line surveys (e.g. Richards et al., 1988) and observations from submersibles are lacking. To date, only one study has assessed the distribution of nearshore rockfish in British Columbia using direct observations from a submersible (Richards and Cass, 1985; Richards, 1986). In their study, the depth and habitat distri- butions of rockfish were surveyed at depths of 21- 140 m in coastal waters of the northeastern Strait of Georgia. In addition to observations on depth and type of habitat frequented by various rockfish spe- cies, however, submersibles also provide a unique opportunity to observe the behavioral activities of the fishes and their associations with conspecifics and heterospecifics. To date, there has been a lack of submersible studies that attempt to quantita- tively assess the in situ activities of rockfish. In the present study, we examined the distribu- tion of rockfish observed from a submersible de- ployed in Saanich Inlet, a coastal fjord in the south- ern Strait of Georgia, British Columbia (Fig. 1). Fjords in British Columbia differ from open coastal areas, such as those surveyed by Richards (1986), in that they typically rely on estuarine-type circu- lation for mixing but have submerged sills which restrict mixing of waters below the sill depth (Thomson, 1981). For fjords of Vancouver Island, B.C., limited circulation results in low dissolved oxygen levels in relatively deep water (Pickard, 1963). In Saanich Inlet, restricted mixing at depth results in oxygen-deficient waters (<2.0 mg-Lr1) be- low 100 m throughout most of the year and inter- mittent, seasonal (usually during January-August) anoxia with production of hydrogen sulfide (Herlinveaux, 1962; Liu, 1989), which is toxic to aerobic organisms (Martin et al., 1981). Various studies of invertebrates in Saanich Inlet have shown that the concentration of dissolved oxygen in the water limits their depth distribution (e.g. Burd, 1983; Mackie and Mills, 1983; Jamieson and Pikitch, 1988; Liu, 1989). Field studies have also demon- strated that the vertical or horizontal distribution of fishes is positively correlated with oxygen concen- tration (reviewed in Kramer, 1987). We therefore speculated that depth distributions for rockfish spe- cies in Saanich Inlet would be relatively shallow when compared to their reported maximum depths in open coastal areas. The present study is the first to examine in situ species composition and density of rockfish in a coastal fiord in British Columbia. In addition, in situ behavioral activities and species associations of nearshore rockfish at depths greater than 30 m in coastal waters of British Columbia have been de- scribed for the first time. Methods The Pisces PV submersible (Department of Fisher- ies and Oceans, Canada) was used to survey rock- fish populations in Saanich Inlet on 9-10 December 1986. A comprehensive description of the Pisces PV submersible was given by Mackie and Mills (1983). The inlet has a steep, rocky slope bottom inter- spersed with sand-shell valleys and is 7.2 km at its Figure 1 Location of transect sites ( • ) in Saanich Inlet, Vancouver Island, British Columbia. Inset: Location of Saanich Inlet in relation to the mainland of B.C. Murie et al. : Distribution and activities of Sebastes spp. 315 widest. Specifics of the oceanographic characteris- tics of Saanich Inlet are detailed elsewhere (Herlinveaux, 1962; Anderson and Devol, 1973). Of relevance is that the basin of the inlet reaches a maximum depth of 234 m, but a submerged sill at 75 m at its mouth in Satellite Channel restricts the renewal of deep-water into the inlet. The resulting oxygen deficiency, anoxia, and production of hydro- gen sulfide is offset in some years when well-oxy- genated, dense bottom-water intrudes over the sill (Herlinveaux, 1962; Anderson and Devol, 1973). We therefore did a hydrocast at depths of 0-225 m to determine the depth of the oxycline in the inlet and whether hydrogen sulfide was present at the time of the surveys. The sampling station was located at lat. 48°37.80'N and long. 123°30.00'W (Liu, 1989). Three sites within Saanich Inlet were surveyed: five transects were traversed at Elbow Point, eight in an area north of McKenzie Bight, and three in an area north of Sheppard Point (Fig. 1). These ar- eas were known from preliminary SCUBA dive sur- veys to have rockfish present in >30 m water depth (Murie et al., pers. obs.). All surveys were conducted during daylight hours (09:30 hours to 16:00 hours) and were also restricted to a depth range between 20 and 150 m because the buoyancy of the Pisces PV is not finely controlled above 20 m, and bottom time restrictions precluded the transects starting at the basin floor (-200 m). Depth at which each transect started therefore varied among sites owing to slope and positioning of the submersible, with starting depths of 95-109 m at Elbow Point, 92-154 m at McKenzie Bight, and 67-74 m at Sheppard Point. At the start of each transect, the Pisces PV sub- merged in open water and on reaching depth the external floodlights were lit and the submersible was manoeuvred horizontally, slowly, toward the cliff face. Once the bottom substrate (cliff) was lo- cated, the submersible began a slow vertical ascent (~5 m-min-1), keeping the viewing ports (port, pilot, and starboard) directed perpendicular to and ap- proximately 3 m from the substrate. Underwater visibility at the time of the surveys was -5—6 m with external illumination. On ascent, an audio-record was made of the spe- cies, time, depth, estimated size (whenever possible), activity, and habitat for each rockfish observed. Each observer (port and starboard) recorded all rockfish encountered within a plane bisecting the pilot's viewport and extending outward at an angle of ap- proximately 45°, corresponding to approximately 3 m of horizontal distance across the substrate (i.e. viewing width). To avoid counting the same fish twice, any rockfish swimming across the path of the submersible or positioned close to the pilot's view- ing area was pointed out to the other observer. Size was visually estimated (±5 cm) by comparing the fish with an externally mounted graduated rod. Rockfish were designated as small (<20 cm total length [TL]) and large (>20 cm TL); large referring to the size at which they enter recreational and com- mercial fisheries (Richards, 1986). Activity of each fish was scored according to whether the fish was perched in the open, positioned in a crevice, occu- pying a shelter hole, hovering off the substrate, or swimming. Habitat was categorized as vertical wall (may have cracks, small crevices, or ledges; score=l), complex (comprising broken rock and boulder fields; score=2), or sand-mud (score=3). Any change in the habitat or slope of the substrate (±10°) was recorded and the depth noted. Rockfish density was estimated for each habitat type over 20-m depth intervals. The total number of fish recorded by both observers within each habi- tat type over a 20-m depth interval was divided by the total area of that habitat type viewed over the depth interval. The area viewed was calculated by multiplying the viewing width of both observers (i.e. 6 m) by the ratio of the change in depth to the sine of the slope. Median densities of small and large fish were calculated for each habitat type and 20-m depth interval, with transects pooled for increased sample size. Density distributions for rockfish were skewed so densities were calculated as medians with 25% and 75% quartiles. Densities of quillback rockfish, S. maliger, among depth intervals and habitats were analyzed by using Kruskal-Wallis tests (SAS, 1985), as this species had an adequate median density (>1 fish- 100m2). Statistical significance was indicated by P < 0.05. Analyses for the other rockfish species (median densities <1 fishlOOm 2) were limited to qualitative comparisons of their depth distributions and numerical abundances. Activity of each species of rockfish was analyzed using percent occurrence, which was calculated by dividing the sum of all individuals observed in each activity by the total number of individuals of the species for which activities were recorded, and mul- tiplying by 100%. Species associations were deter- mined for individual fish of each species by scoring the presence of a conspecific or a heterospecific within 3 m. The sum of the number of individuals which were observed in the presence of a con- or hetero-specific was then expressed as a percentage of the total number of individuals of the species. Individual rockfish with no other rockfish within 3 m were considered to be 'alone' (solitary). 316 Fishery Bulletin 92(2). 1994 Results Physical habitat An area of approximately 10,521 m2 was surveyed from the submersible, of which 38% was wall, 47% complex, and 15% sand-mud habitat. Area of cover- age among habitat types differed with depth (Fig. 2). The area of complex and sand-mud habitat cov- ered in the surveys decreased with depth whereas that of wall habitat increased. Sand-mud habitat was encountered only at depths of <60 m and the median area surveyed among transects was zero. Wall habitat was the only habitat type observed at depths greater than 120 m. The slope of the sub- strate was correlated with depth (Spearman rank correlation: rs=0.37, P<0.001) and habitat (rs =-0.71, P<0.001). This was evident in that wall habitat found primarily in deep water provided vertical or near- vertical relief (-70-90° slope), whereas complex and sand-mud habitats in shallower depths provided a graded substrate (-20-70° slope). The area of each type of habitat surveyed differed among survey sites (Kruskal-Wallis: P=<0.001, 0.03, and <0.001 for wall, complex, and sand-mud habi- tat respectively). Elbow Point and McKenzie Bight had similar habitats whereas Sheppard Point had less wall and complex habitat and more sand-mud 5 DC < 200 - 180 - 160 140 120 100- 80 - 60 - 40 20 - 0 NUMBER OF TRANSECTS: □ COMPLEX D WALL I I J I 1 ?,. i // i 21-40 41-60 61-80 81-100 101-120 121-140 DEPTH (m) 5 2 16 16 16 13 Figure 2 Overall median area of survey coverage for com- plex and wall habitat in relation to depth. Verti- cal bars represent the interquartile ranges - Median area of sand-mud habitat sur- veyed was zero. habitat than the Elbow Point and McKenzie Bight sites. At the time of the surveys, Saanich Inlet was not anoxic although waters below 100 m were deficient in dissolved oxygen (Liu, 1989) (Table 1). Hydrogen sulfide was not present at any depths sampled in the inlet. Temperature and salinity were relatively stable below depths of 100 m. Depth, size, and density distributions Quillback rockfish represented 88% (681/770) of all rockfish sighted and were observed at a median depth of 54 m (Table 2). Density of quillback rock- fish did not differ among depth intervals between 21-100 m (P=0.35) (Fig. 3A). Only three quillback rockfish were seen at depths >100 m. In total, the median size of 460 quillback rockfish was 23 cm (Table 2). Quillback rockfish size was positively cor- related with depth (^=0.23, P<0.001, rc=460) and the density of small and large quillback rockfish varied among depth intervals (P=0.01 and P=0.02 respec- tively) (Fig. 3B). Density of small quillback rockfish was similar to that of large quillback rockfish in the 21^40 m depth interval (P=0.66), but it was less at depth intervals greater than 40 m (all P<0.05) (Fig. 3B). In contrast, the median density of large quill- back rockfish at depth intervals between 41-100 m was greater than their density in the 21-40 m depth interval (Fig. 3B). Tiger rockfish, S. nigrocinctus, copper rockfish, S. caurinus, yellowtail rockfish, S. flavidus, green- striped rockfish, S. elongatus, and yelloweye rock- fish, S. ruberrimus, all had median densities of zero in 21-150 m depths in Saanich Inlet (Table 2). Table 1 Temperature, salinity, and dissolv ed oxygen meas- ured throughout de pths in Saanich Inlet on 12 December 1986. Hydrogen sulfide was not present at any of the depths sampled. Depth Temperature Salinity Dissolved oxygen (m) CO CM (mg-lr1) 0 6.24 26.68 7.01 10 8.70 30.18 5.83 30 8.94 30.31 4.50 50 9.40 30.56 3.80 75 9.19 30.96 2.80 100 9.19 31.25 1.39 125 9.23 — 0.41 150 9.28 31.44 0.50 175 9.27 — 0.87 200 9.29 31.47 1.70 225 9.29 — 1.75 Murie et al.: Distribution and activities of Sebastes spp. 317 Table 2 Numerical abundance, m€ at depths of 21-150 m in dian density, median depth, and median estimated size of Sebastes species observed Saanich Inlet from the Pisces TV submersible. Species Number Density (fishlOOm-2) Depth (m) Size (cm) Median Range Median Range Median Range S. maliger 681 5 (0-31) 54 (21-115) 23 ( 5-41) S. nigrocinctus 28 0 (0- 2) 55 (33- 97) 28 (20-46) S. caurinus 24 0 (0- 4) 44 (21- 65) 25 ( 5-36) S. flavidus 23 0 (0- 7) 49 (41- 65) 35 (20-40) S. elongatus 8 0 (0- 1) 65 (52-114) 18 (15-23) S. ruberrimus 5 0 (0-4) 89 (76-103) 28 (18-46) Unidentified sp. 1 42 — E o o I CO CO z LU Q E o o I CO E CO 2 LU Q 14 • 12 ' 10 ' 8 ■ 6- 4 ■ 2 ■ 0 ■ A \ \ \ \ \ \\\N* \ V k \ \ X V \ \ \ f / . t I I t * t / •/./•, t * * / * \\S»^ S X \ \ \ \N.*\ •» S S \ V '\'>' \'\'\ \ \ \ \ \ \ \ \ \ \' S/\W\ 21-40 41-60 61-80 81-100 B D SMALL □ LARGE '. L. 1 F2 I 21-40 41-60 61-80 DEPTH (m) 81-100 Figure 3 (A) Overall median densities of all quillback rock- fish, Sebastes maliger, over depth; and (B) median densities of small and large quillback rockfish over depth. Vertical bars represent the interquartile ranges (Q025 to Q075). Median density of quillback rockfish in depths >100 m was zero. Tiger rockfish accounted for 4%, and copper rockfish 3%, of the rockfish encountered; both were most abundant in 41-60 m (0.6 fish- 100m"2). Yellowtail rockfish also represented 3% of all rockfish observed with their density reaching a maximum of 6.3-6.6 fish-lOOm-2 in the 41-80 m depth intervals. Greenstriped rockfish (n=8), yelloweye rockfish (n =5), and one unidentified rockfish each represented less than 1% of all observed rockfish. Greenstriped and yelloweye rockfish both occurred in relatively deeper waters (Table 2). The size of tiger rockfish was not correlated with depth (P=0.27, n = 15) and only relatively large fish were seen from the submersible (Table 2). Copper rockfish size was positively correlated with depth (r2=0.36, P=0.02, n=15), and no small copper rock- fish were seen at depths greater than 40 m. Yellow- tail rockfish seen were all relatively large fish (Table 2) and their size was not correlated with depth (P=0.46, n=21). All greenstriped rockfish observed were small (Table 2). Two juvenile yelloweye rockfish ( 18-20 cm TL) were observed at depths greater than 95 m and three subadult and adult yelloweye rockfish (36-46 cm TL) occurred between 80 and 90 m. Habitat distribution Density of quillback rockfish differed among survey sites (P=0.001). Densities observed at the Elbow Point and McKenzie Bight sites were similar (P=0.236), with a pooled median density of 5.7 fish- 100m2. In contrast, the median density of quill- back rockfish at Sheppard Point was zero. Overall, quillback rockfish density was highest in areas of complex habitat (5.8 fish- 100m-2), followed 318 Fishery Bulletin 92(2), 1994 by wall habitat (3.5 fish-lOOm"2) (Fig. 4). Only four quillback rockfish were observed over sand-mud habitat. Quillback rockfish densities, whether in complex or wall habitat, did not differ among depth intervals <100 m (P=0.52 and P-0.64 respectively) (Fig. 4). Tiger, copper, yellowtail, and yelloweye rockfish were observed only over complex or wall habitats, whereas greenstriped rockfish occurred mostly over sand-mud habitat (Table 3). Tiger rockfish tended to occur in both complex and wall habitats, whereas copper, yellowtail, and yelloweye rockfish were seen mostly in complex habitat. Activities Quillback, copper, and greenstriped rockfish did not appear to be attracted to or obviously repelled by the Table 3 Numerical abundance of rockfish (Sebastes spp.) observed in densities of <1 fish- 100m"2 over com- plex, wall, and sand-mud habitat in Saanich Inlet. Species Complex Wall Sand-Mud S. nigrocinctus S. caurinus S. flavidus S. elongatus S. ruberrimus 13 2(1 2] 0 t 15 •1 2 1 1 E o I CO z LU Q 20- 18- 16" 14 - 12- 10 8 6 4 2 0 1 M 1 □ COMPLEX □ WALL 21-40 41-60 61-80 DEPTH ( m ) 81-100 Figure 4 Median densities of quillback rockfish, Sebastes maliger, in complex and wall habitats among depths. Vertical bars represent interquartile ranges (Q025 to Qo7s)- Median densities of quillback rockfish over sand-mud habitat were zero. presence of the submersible and its lights. At times, rockfish actively finned to maintain station after the submersible had produced currents. Any observed movements away from or towards the submersible were always relatively slow and, at times, hovering quillback rockfish would move slightly away from the path of the submersible, stop, and then resume hovering. Tiger and yelloweye rockfish appeared to have a delayed response to the submersible, in that it was possible to observe them prior to their actu- ally moving into a shelter hole or crevice. In con- trast, some spotted ratfish, Hydrolagus colliei, and a sixgill shark, Hexanchus griseus, were obviously attracted to the Pisces. These fish swam back-and- forth around the front of the submersible, repeatedly approaching the viewing ports near the external lights. Activities were determined for a total of 662 quill- back rockfish and the majority were observed hov- ering or perched on substrate in the open (Fig. 5), regardless of depth interval (all P>0.10). Copper rockfish were also observed primarily hovering and perched in the open, but were seen swimming more frequently than quillback rockfish. Both species were observed infrequently in crevices and rarely in shelter holes. Tiger rockfish were also observed most frequently perched in the open or occupying crevices. Yellowtail rockfish were all observed either hover- ing or swimming close to the substrate. All eight greenstriped rockfish were observed perched on the substrate. Of the five yelloweye rockfish observed, one was in a shelter hole, three were in crevices, and one was hovering. Species associations The majority of quillback rockfish (94% occurrence) were observed within 3 m of at least one other quillback rockfish (Fig. 6). Quillback rockfish were almost never observed alone (2%) and were ob- served in the presence of other species relatively infrequently (-20% occurrence or less). Quillback rockfish formed loose conspecific aggregations that were distinctly different from the conspecific schools of yellowtail rockfish observed from the Pisces. When schooling, yellowtail rockfish formed tight groups of fish that orientated and moved to- gether in the same direction, whereas in the ag- gregations of quillback rockfish, individual fish were orientated in various directions and engaged in various activities. Small groups of quillback rockfish (2-5 fish) observed from the submersible were interspersed between larger aggregations of more than 15 fish. Copper rockfish occurred within 3 m of quillback rockfish 92% of the time (Fig. 6), but also tended to occur near other copper rockfish Murie et al.: Distribution and activities of Sebastes spp. 319 (64% occurrence) and tiger rockfish (32%). Copper rockfish, while usually near quillback rockfish, were also ob- served in conspecific aggregations and they were seldom seen alone (4%). Ti- ger rockfish were almost always ob- served near quillback rockfish (96% occurrence) and, to a much lesser ex- tent, near other tiger rockfish (21%) (Fig. 6). The majority of yellowtail rock- fish were observed in proximity to quillback rockfish (96%) and other yel- lowtail rockfish (91%). All five of the yelloweye rockfish seen were near quill- back rockfish. Six greenstriped rockfish were observed near quillback rockfish whereas two were alone. Discussion COPPER (n=24) ui cr DC z> o o o HOLE CREVICE PERCH HOVER SWM HOLE CREVICE PERCH HOVER SWM Quillback rockfish are the numerically dominant rockfish species at depths of 21-100 m in nearshore areas of south- ern British Columbia, based on sub- mersible observations in a fjord (this study) and in relatively open coastal areas (Richards, 1986). Based on both of these submersible studies, the main depth distribution of quillback rockfish is between 41 and 60 m, and their density at this depth is more than eight times greater than that of any other rock- fish species observed at 41-60 m (Fig. 3A; Richards, 1986). As in Saanich Inlet, greenstriped and yelloweye rockfish were also observed in relatively low densities in the northeastern Strait of Georgia (means of <2 fishlOOm-2 in various habitat types) (Richards, 1986), as well as tiger and copper rock- fish (Richards and Cass, 1985). Yellowtail rockfish were not seen during submersible dives in the north- eastern Strait of Georgia (Richards and Cass, 1985). In Saanich Inlet, complex habitat dominated by broken rock and boulder fields appears to be a com- mon feature for the occurrence of the majority of these rockfish species. Based on Pisces surveys in the northeastern Strait of Georgia, Richards (1986) also observed that quillback and yelloweye rockfish were most abundant in complex habitat. Similarly, densities of copper and quillback rockfish were high- est in complex habitat or in areas of highly irregu- lar relief in Saanich Inlet in <40 m depth (Murie, 1991) and in the northern Strait of Georgia in <18 m depth (Richards, 1987). In SCUBA surveys, Matthews (1990) found the greatest densities of large copper and large quillback rockfish on high- relief rocky reefs in Puget Sound, Washington. Ad- ACTIVITY Figure 5 Percent occurrence of behavioral activities for quillback rockfish, Sebastes maliger; copper rockfish, S. caurinus; tiger rockfish, S. nigro- cinctus; and yellowtail rockfish, S. flavidus, observed in Saanich Inlet from the Pisces submersible. ditionally, submersible observations in the vicinity of Heceta Bank, Oregon (Pearcy et al., 1989), sug- gested that tiger, yelloweye, and yellowtail rockfish were most frequently encountered over rock and rubble habitat. The densities of these near-bottom species may be greatest in this type of habitat be- cause of increased protection from predators or in- creased density of prey due to the increase in mi- crohabitat and vertical structure. Stein et al. (1992) observed fish from a submers- ible at Heceta Bank and determined that the occur- rence of fish species was related to specific sub- strates. Given the propensity of quillback rockfish to aggregate over complex habitat, differences in their density among sites surveyed in our study was not surprising. Sheppard Point, which had more sand-mud areas and a shallower slope, was notice- ably different from Elbow Point and McKenzie Bight. It was also the only site where greenstriped rockfish were observed, which is consistent with the apparent habitat distribution of this species (Richards, 1986; Pearcy et al., 1989; Stein et al., 1992). The overlap in the depth ranges, as well as the similarity in the median depths and the occurrence of fish over complex and wall habitats, suggested that quillback, copper, tiger, and yellowtail rockfish do not segregate in Saanich Inlet within the range 320 Fishery Bulletin 92|2), 1994 UJ cc o o o of bathymetry or habitat surveyed from the submersible. Yellowtail rockfish may be segregated spatially from tiger rock- fish and, to some degree, copper and quillback rockfish, because the activities of yellowtail rockfish consistently placed them in the water column near the sub- strate but never in direct contact with the bottom. The appearance of quillback and copper rockfish near to one another (Fig. 6) was consistent with observations from SCUBA dive surveys at Saanich Inlet in 20-40 m. Sympatric aggrega- tions of quillback and copper rockfish over complex habitat in Saanich Inlet can be dense (-25-50 fishlOOm"2) (Murie, 1991). Published information on in situ be- havioral activity, species associations, and density of tiger rockfish is scarce, no doubt in part due to the consistently low densities in which this species is encoun- tered. Tiger rockfish have been observed in low densities, and primarily as only a single fish encountered at any one time, in waters <30 m deep in Puget Sound (Moulton, 1977), in 21-140 m in the northeastern Strait of Georgia (Richards and Cass, 1985), in 64-305 m depths off Oregon (Pearcy et al., 1989), and in <30 m in Saanich Inlet (Murie, 1991). As 21% (6/28) of S. nigrocinctus were observed within 3 m of another tiger rockfish in our study (Fig. 6), this species may not be as 'solitary' as in- dicated by the previous studies. Density of tiger rockfish may be limited by the availability and de- fense of suitably large shelter holes, which the fish retreat into upon approach by a SCUBA diver (Murie, pers. obs.). The depth distribution of rockfish in Saanich In- let, and hence any size or species associations, may be influenced by a number of factors, including a) the physical regime of the inlet; b) the paucity of observations for relatively uncommon species; and c) the actual depth range and total number of transects surveyed with the submersible. The year- round oxygen deficiency in waters >100 m in Saanich Inlet, and the intermittent anoxia that oc- curs in waters of 125-234 m depth during January to August (Liu, 1989), may act to compress or shift the depth distribution of rockfish compared to open coastal waters where relatively deep water is not limiting in dissolved oxygen. In Saanich Inlet, squat lobsters, Munida quadrispina, migrate vertically en masse to avoid decreasing oxygen levels (Burd, QUILLBACK (n=681) 1 COPPER (n=24) 1 11 -2 Wwft. rrrvrrx TIGER (n=28) L m 4T77T. . JZZ3, T YT YE GS A YELLOWTAIL (n=23) I \rrr\n-r-il rrr\ Q C T YT YE GS A SPECIES Figure 6 Percent occurrence of conspecific and heterospecific rockfish in as- sociation with quillback rockfish, Sebastes maliger; copper rockfish, S. caurinus; tiger rockfish, S. nigrocinctus; and yellowtail rockfish, S. flavidus; Species key: Quillback rockfish (Q), copper rockfish (C), tiger rockfish (T), yellowtail rockfish (YT), yelloweye rockfish (YE), greenstriped rockfish (GS), and alone (A). 1983) although catastrophic mortality of spot prawns, Pandalus platyceros, in 85—90 m depth has been attributed to a sudden intrusion of displaced anoxic bottom-water into midwater depths (Jamieson and Pikitch, 1988). For mobile organisms (such as rockfish) decreasing oxygen levels may elicit a behavioral response involving a vertical or horizontal habitat shift (Kramer, 1987). Of three transects in Saanich Inlet that started at 150 m, no rockfish were observed at depths >115 m, and only three quillback rockfish were seen at depths >100 m (note: one of these transects was not used in the overall analysis owing to loss of the audio-track at <80 m). With the exception of copper rockfish, Hart (1973) reports maximum depths of >250 m for the other rockfish species. In addition, quillback, tiger, greenstriped, and yellowtail rockfish have been ob- served from submersibles at depths >120 m (Richards and Cass, 1985; Pearcy et al., 1989). The scarcity of observations on relatively low den- sity species of rockfish, especially in combination with time limitations of the submersible, could also affect our interpretation of rockfish depth distribu- tion. In Saanich Inlet, few rockfish other than quill- Murie et al.: Distribution and activities of Sebastes spp. 321 back rockfish were observed during the submersible dives. The depth range and density of the relatively uncommon species of rockfish may have been im- proved if we had been able to do more transects and, in the instance of maximum depths, by doing deeper transects. Bias in using submersible transects was exemplified by the downward bias (i.e. deeper) in the observed minimum depth ranges for tiger, yellow- tail, and yelloweye rockfish in Saanich Inlet (Table 2). These rockfish species have been observed dur- ing SCUBA dives in Saanich Inlet in water as shal- low as 15-25 m (Murie, pers. obs.). The density es- timate and depth range for copper rockfish was probably also biased because of the reduced ability to maintain fine control of the Pisces buoyancy as it approaches shallower depths (-20 m). We know that copper rockfish in Saanich Inlet occur from near-surface waters (~2 m) and their distribution ex- tends visibly below 40 m depth (Murie, 1991). The density of copper rockfish on rocky reefs in 20-30 m of water, however, can approach 50 fish- 100m 2 (Murie, 1991), far in excess of any density observed for copper rockfish from the submersible (Table 2). In general, however, copper rockfish do occur in shallower water than quillback rockfish (Moulton, 1977; Richards, 1987; Murie, 1991), as was observed from the submersible transects in Saanich Inlet. Another potentially important bias in the use of the Pisces PV to observe densities and activities of rockfish is whether the fish are attracted or notice- ably repelled by the size, noise, and lights of the submersible. Similar to our study, Carlson and Straty (1981) noted that most of the rockfish were neither repelled nor attracted to the submersible while they observed them in southeastern Alaska. In addition, Richards (1986) observed that none of the common fish species seen in the northeastern Strait of Georgia seemed disturbed by the Pisces PV submersible. A notable exception in Carlson and Straty 's (1981) study was large (7-10 kg) yelloweye rockfish that were obviously attracted to the sub- mersible and actually followed it, similar to the rat- fish and sixgill shark in our study. Pearcy et al. (1989) also noted that large schools of yellowtail rockfish were attracted to their submersible and followed it over substantial periods of time and depth; there was no visible evidence, however, of schools of yellowtail rockfish following the Pisces in our study. The occurrence of perching and hovering activi- ties observed for the majority of rockfish in Saanich Inlet from the Pisces was consistent with behavioral activities of quillback and copper rockfish observed with SCUBA in Saanich Inlet (Murie, 1991). Obser- vations from the submersible were limited in this respect because it was impossible to look into all crevices or into shelter holes under rocks for the presence of fish. Tiger and yelloweye rockfish could be seen in shelter holes and crevices but their size could not always be estimated. Although the Pisces approaches shelter holes from below (during its as- cent), fish in deep shelter holes and crevices may not be detected. The presence of fish in crevices and shelter holes was therefore probably underesti- mated. Nevertheless, at present, submersibles and remotely-operated vehicles (ROVs) provide the best means of observing the activities of rockfish occu- pying complex habitat in deep water. Although it is evident from submersible observa- tions that estimates of abundance and activities of rockfish involve a variety of biases, these direct vi- sual assessments can provide quantitative informa- tion on the densities and depth distributions of rock- fish species in habitats that cannot be surveyed ad- equately using bottom trawls. In addition, submersibles allow direct observation of the behav- ioral activities and associations of individual fish in relation to specific habitat types. This type of infor- mation has not been attainable using conventional survey techniques of fisheries. Acknowledgments We are especially grateful to T Fitch and the pilots and crew of the Pisces PV (Institute of Ocean Sci- ences, Department of Fisheries and Oceans, Sidney, B.C. ) who provided the opportunity and expertise to work efficiently with the submersible. We thank Q. Liu for analyzing the hydrocast samples, and J. Gilbert and C. Steinhoff for rockfish observations. Constructive comments by F. Matthews, L. Richards, J. Mclnerney, N. Wilimovsky, R Gregory, V. Tunnicliffe, D. Mitchell, C. Tolman, and two anonymous reviewers improved the manuscript. Financial support was provided by a Postgraduate Scholarship from the Natural Sciences and Engi- neering Research Council of Canada (DJM) and by the Biology Department, University of Victoria. Literature cited Anderson, J. J., and A. H. Devol. 1973. Deep water renewal in Saanich Inlet, an in- termittently anoxic basin. Estuarine and Coastal Mar. Sci. 1:1-10. 322 Fishery Bulletin 92(2). 1994 Boehlert, G. W. 1980. Size composition, age composition, and growth of canary rockfish, Sebastes pinniger, and splitnose rockfish, S. diploproa, from the 1977 rockfish survey. Mar. Fish. Rev. 42: 57-63. Burd, B. J. 1983. The distribution, respiration and gills of a low oxygen tolerant crab, Munida quadrispina (Benedict, 1902) (Galatheidae, Decapoda) in an intermittently anoxic fjord. M.Sc. thesis, Biology Dept. Univ. Victoria, Victoria, B.C., 151 p. Carlson, H. R., and R. R. Straty. 1981. Habitat and nursery grounds of Pacific rock- fish, Sebastes spp., in rocky coastal areas of south- eastern Alaska. Mar. Fish. Rev. 43:13-19. Dennis, G. D., and T. J. Bright. 1988. Reef fish assemblages on hard banks in the northwestern Gulf of Mexico. Bull. Mar. Sci. 43:280-307. Hallacher, L. E., and D. A. Roberts. 1985. Differential utilization of space and food by the inshore rockfishes (Scorpaenidae: Sebastes) of Carmel Bay, California. Env. Biol. Fish. 12:91-110. Hart, J. 1973. Pacific fishes of Canada. Fish. Res. Board Can. Bull. 180:407-443. Herlinveaux, R. 1962. Oceanography of Saanich Inlet in Vancouver Island, British Columbia. J. Fish. Res. Board Can. 19:1-37. Jamieson, G. S., and E. K. Pikitch. 1988. Vertical distribution and mass mortality of prawns, Pandalus platyceros, in Saanich Inlet, British Columbia. Fish. Bull. 86:601-608. Kramer, D. L. 1987. Dissolved oxygen and fish behavior. Env. Biol. Fish. 18:81-92. Krieger, K. 1993. Distribution and abundance of rockfish deter- mined from a submersible and by bottom trawling. Fish. Bull. 91:87-96. Larson, R. J. 1980. Competition, habitat selection, and the bathymetric segregation of two rockfish (Sebastes) species. Ecol. Monogr. 50: 221-239. Liu, Q. 1989. Ecophysiological studies of Orchomenopsis affinis (Holmes) (Lysianassidae, Amphipoda) in an intermittently anoxic fjord. M.Sc. thesis, Biology Dept., Univ. Victoria, Victoria, B.C., 122 p. Mackie, G. O., and C. E. Mills. 1983. Use of the Pisces PV submersible for zoo- plankton studies in coastal waters of British Columbia. Can. J. Fish. Aquat. Sci. 40:763-776. Martin, D. W., P. A. Mayes, and V. W. Rodwell. 1981. Harper's review of biochemistry, 18th ed. Lange Medical Pub., Los Altos, CA, 126 p. Mathews, S. B., and M. W. Barker. 1983. Movements of rockfish (Sebastes) tagged in northern Puget Sound. Fish. Bull. 82: 916-922. Matthews, K. R. 1990. A comparative study of habitat use by young- of-the-year, subadult, and adult rockfishes on four habitat types in central Puget Sound. Fish Bull. 88:223-239. Moulton, L. 1977. An ecological analysis of fishes inhabiting the rocky nearshore regions of northern Puget Sound, Washington. Ph.D. diss., Univ. Washington, Se- attle, WA, 181 p. Murie, D. J. 1991. Comparative ecology and interspecific com- petition between the sympatric congeners Sebastes caurinus (copper rockfish) and S. maliger (quill- back rockfish). Ph.D. diss., Univ. Victoria, Victoria, B.C., 287 p. Patten, B. G. 1973. Biological information on copper rockfish in Puget Sound, Washington. Trans. Am. Fish. Soc. 102:412-416. Pearcy, W. G., D. L. Stein, M. A. Hixon, E. K. Pikitch, W. H. Barss, and R. M. Starr. 1989. Submersible observations of deep-reef fishes of Heceta Bank, Oregon. Fish. Bull. 87:955-965. Pickard, G. L. 1963. Oceanographic characteristics of inlets of Vancouver Island, British Columbia. J. Fish. Res. Board Can. 20:1109-1144. Richards, L. J. 1986. Depth and habitat distributions of three spe- cies of rockfish (Sebastes) in British Columbia: observations from the submersible Pisces PV. Env. Biol. Fish. 17:13-21. 1987. Copper rockfish (Sebastes caurinus) and quillback rockfish (Sebastes maliger) habitat in the Strait of Georgia, British Columbia. Can. J. Zool. 65:3188-3191. Richards, L. J., and A. J. Cass. 1985. Transect counts of rockfish in the Strait of Georgia from the submersible Pisces PV, October and November 1984. Can. Fish. Aquat. Sci. Data Rep. No. 511, 99 p. Richards, L. J., C. M. Hand, and J. R. Candy. 1988. 1988 research catch and effort data on near- shore reef-fishes in British Columbia statistical areas 12 and 13. Can. Fish. Aquat. Sci. Man. Rep. No. 2000, 89 p. SAS (SAS Institute Inc.). 1985. SAS User's guide: statistics, vers. 5. Cary, NC, 956 p. Stein, D. L., B. N. Tissot, M. A. Hixon, and W. Barss. 1992. Fish-habitat associations on a deep reef at the edge of the Oregon continental shelf. Fish. Bull. 90:540-551. Thomson, R. E. 1981. Oceanography of the British Columbia coast. Can. Spec. Publ. Fish. Aquat. Sci. 56, 291 p. Murie et al.: Distribution and activities of Sebastes spp. 323 Uzmann, J. R., R. A. Cooper, R. B. Theroux, and R. L. Wigley. 1977. Synoptic comparison of three sampling tech- niques for estimating abundance and distribution of selected megafauna: submersible vs. camera sled vs. otter trawl. Mar. Fish. Rev. 39:11-19. Westrheim, S.J. 1970. Survey of rockfishes, especially Pacific ocean perch, in the northeastern Pacific Ocean, 1963- 66. J. Fish. Res. Board Can. 27:1781-1809. Wilkins, M.E. 1980. Size composition, age composition, and growth of chilipepper, Sebastes goodei, and bocac- cio, S. paucispinis, from the 1977 rockfish survey. Mar. Fish. Rev. 42:48-53. Abstract. The spotted dol- phin (Stenella attenuata) is found throughout much of the eastern tropical Pacific Ocean. A previous study evaluated morphological variation in skull morphology, but now specimens are available for a greater portion of the range. Also, corrections have been made in data and an assessment has been made evaluating repeatability of character measurements. We reas- sessed geographic variation in 30 cranial features (26 morphometric measures and 4 tooth counts) based on 611 museum specimens. All characters except two tooth counts showed statistically signifi- cant geographic variation, while 21 of the 30 characters exhibited significant sexual dimorphism. Males were larger in most charac- ters; females were larger in some length measurements involving the rostrum and ramus. As in pre- vious analyses, inshore S. attenu- ata were found to be very distinc- tive, so subsequent analyses fo- cused on offshore spotted dolphins from 29 5° latitude-longitude blocks. Mantel tests and matrix correlations for 19 of the 30 fea- tures demonstrated significant "re- gional patterning," whereas 22 of the characters were shown to have "local patterning." Principal-com- ponents, canonical-variates, and cluster (UPGMA and function- point) analyses also were em- ployed to assess geographic varia- tion. In the eastern portion of the range, the subdivision between northern and southern offshore S. attenuata found in the previous investigation was confirmed. In general, blocks to the west (includ- ing one encompassing part of the Hawaiian Islands) were more like the southern blocks than those of the northeast. Morphological pat- terns were similar to those found in a number of environmental vari- ables, particularly water depth, so- lar insolation (January), sea sur- face temperature (January and July), surface salinity, and thermo- cline depth (winter and summer). Present management units are inconsistent with the pattern of cranial variation; spotted dolphins from west of lat. 120°W probably should not be pooled with those to the east, as they show closer affin- ity with the Southern Offshore unit. In addition, the boundary between the Northern and South- ern units should probably be moved north to about lat. 5°N. Manuscript accepted 12 October 1993 Fishery Bulletin 92:324-346 (1994) Reexamination of geographic variation in cranial morphology of the pantropical spotted dolphin, Stenella attenuata, in the eastern Pacific William F. Perrin Southwest Fisheries Science Center. National Marine Fisheries Service, NOAA PO Box 27 I . La Jolla, CA 92038 Gary D. Schnell Daniel J. Hough Oklahoma Biological Survey and Department of Zoology. University of Oklahoma Norman, OK 73019 James W. Gilpatrick Jr. Jerry V Kashiwada Southwest Fisheries Science Center, National Marine Fisheries Service, NO/V\ PO Box 27 I , La Jolla, CA 92038 Spinner and spotted dolphins (Sten- ella longirostris and S. attenuata) have broadly overlapping ranges in the eastern tropical Pacific Ocean (Perrin et al., 1983). Information concerning geographic variation of these species is of both intrinsic scientific and practical interest. Dolphins in the region are killed as a result of purse-seining for yellow- fin tuna (Thunnus albacares; Allen, 1985). The tuna often are found in association with these two dolphin species (or with Delphinus delphis), and fishermen set nets on the dol- phin schools to capture tuna found below the dolphins. Estimates indi- cate that from 1985 to 1990 roughly 53,000 to 129,000 dolphins were killed annually as a result of fishing operations (Hall and Boyer 1987, 1988, 1989, 1990, 1991, 1992). Most recently the annual kill has dropped to approximately 15,000 to 27,000 (Hall and Len- nert1'2). Government regulations in the United States set limitations on U.S. vessels with respect to the extent of dolphin mortality that will be permitted. Dolphins are managed by defining a series of management stocks. Data, such as those on skull morphology, can pro- vide insight into the underlying population subdivision and may be of considerable value in defining geographic boundaries of biolog- ically relevant management stocks (Dizon et al., 1992). Douglas et al. (1992) have pro- vided a detailed assessment of geo- graphic variation in cranial mor- phology of spinner dolphins. For spotted dolphins, the most recent geographic-variation analyses us- ing skull characteristics were by Douglas et al. (1984) and Schnell et Hall, M. A., and C. Lennert. 1992. Esti- mates of incidental mortality of dolphins in the purse-seine fishery for tunas in the eastern Pacific Ocean in 1991. Int. What Commn. meeting doc. SC/44/SM6, 5 p. Hall, M. A., and C. Lennert. 1993 Inciden- tal mortality of dolphins in the eastern Pacific Ocean tuna fishery in 1992. Int. What Commn. meeting doc. SC/45/SM1, 5 p. 324 Perrin et al.: Geographic variation in cranial morphology of Stenella attenuata 325 al. (1986). Numerous additional specimens have become available, particularly from the western portion of the range and Hawaii. The repeatability of 36 skull measures used in previous studies (Dou- glas et al., 1984; Schnell et al., 1986) was appraised, as was done previously by Douglas et al. (1992) for spinner dolphins. Also, some immature specimens had inadvertently been incorporated into the previ- ous spotted dolphin analyses. For these reasons, we have undertaken a reassessment of geographic variation and sexual dimorphism of spotted dolphins from the eastern tropical Pacific. This study pro- vides an opportunity to re-evaluate variation pat- terns previously described and to compare directly patterns of variation found in spotted and spinner dolphins. Materials and methods Overall, data-gathering and assessment procedures outlined by Douglas et al. (1992) were used. We measured 611 adult museum specimens (maturity evaluated on the basis of premaxilla fusion with maxilla at distal end of rostrum; Dailey and Perrin, 1973) of spotted dolphins (Fig. 1). These included 534 of 613 specimens used in earlier studies (Dou- glas et al., 1984; Schnell et al., 1986; 79 specimens previously used had been incorrectly aged or had inadequate locality data) along with 77 new specimens. As was done with spinner dolphins (Douglas et al., 1992), the first specimen set was measured by M. E. Douglas and the new specimens by W. F Perrin. In addition, Perrin remeasured 81 specimens of spinner and spotted dolphins measured by Douglas to determine whether measurements were repeat- able. Initially, 36 characters were evaluated (illus- trations and character definitions given in Schnell et al., 1985). Comparisons of measurements taken on the same specimens by the two investigators in- dicated that 6 of the original 36 measurements (i.e. width of left premaxillary [at midline of nares], width of right premaxillary [at midline of nares], separation of pterygoids, length of left tympanic cavity, length of right tympanic cavity, and width at pterygobasioocipital sutures) should be deleted, be- cause we were not able consistently to repeat these measurements. For some other measurements, there were differences between investigators, but the dif- ferences were consistent (e.g. one obtained measure- ments that were smaller than those reported by the other). Therefore, we calculated regression equations for each of the remaining characters based on the 81 jointly measured specimens. These regression equations were used to convert the measurements from the rest of the initial specimens to appropri- ate values for inclusion with the measurements taken by Perrin. Through these procedures, we de- veloped a data set of 30 characters (see Table 1) for the 611 specimens. Specimens were not used if, because of damaged parts, we could not obtain most of the 30 measure- ments. Missing values (0.50% of total) for included specimens were estimated by linear regression3 onto the character that explained the greatest proportion of the variance for the variable being considered. Animals were assigned to 5° latitude-longitude blocks and each geographic block assigned a numeri- cal code (see Fig. 1). These codes were modified slightly from those employed by Schnell et al. ( 1986) to accommodate new specimens from more westerly blocks. We had specimens available from 41 blocks, 8 of which were represented by only a single speci- men and 4 of which were inshore blocks (i.e. con- tained only specimens of the inshore form; Douglas et al., 1984). The 29 blocks that were not inshore blocks and had more than one specimen were used as the basis for most geographic variation analyses. While several of the 29 blocks have relatively small samples, geographic-patterning tests (described be- low) suggested that, in general, sample values are representative of what is expected for these blocks based on their geographic positions. Schnell et al. (1985) showed S. attenuata to be sexually dimorphic for 23 of 36 characters. Because some specimens used in that analysis were removed and new specimens added (see above), we conducted a two-way analysis of variance (ANOVA) for block and sex, based on specimens in the 11 blocks with at least four of each sex (Fig. 1). Correction terms were obtained to adjust measurements of the larger sex downward and the smaller sex upward, thus producing sex-adjusted or "zwitter" measurements (method described in more detail by Schnell et al., 1985). As a result, we were able to combine speci- mens for both sexes in an overall analysis of geo- graphic variation. To assess whether combining specimens from dif- ferent cruise sets within blocks confounded geo- graphic patterns based on blocks, we performed a nested ANOVA for cruise sets within the 12 blocks for which two specimens were obtained from at least two cruise sets. Blocks employed in this analysis (with numbers of cruise sets in parentheses) were the following: 0215 (4), 0216 (3), 0312 (2), 0506 (2), 0507 (2), 0512 (9), 0513 (10), 0515 (2), 0612(6), 0613 (7), 0615 (3), and 0712 (2). 3 "Missing Data Estimator" program by Dennis M. Power. Santa Barbara Mus. Nat. Hist., pers. commun. 1975. 326 Fishery Bulletin 92(2). 1994 09 08 07 Lh 05 04 03 02 01 . Ml oJ — * i< 0° . . -'■12 '1\T! "i '0 ; ^"V* \W §M l\ u ■M 1 ^ $ r . 2 8 \ p^ It cp^. I 1 1 1 7 17 \5 2 S o> # N 7 50 26 1 16 1 18 77 32 3 15 * * • * • • * • o 1 4 1 3 4 6 4 6 1 2 1 16 17 46 50 2 4 12 19 3 w • • • • * • * • • * • * 1 1 1 1 1 2 1 2 / 3 3 3 4 1 1 4 / - Stenella attenuata Males Females 1 - 2 - 1 12 * • . 10 1 11 20 10 '".-!, ■ - # • • ^ i I | 1 01 03 05 07 09 15 1S Figure 1 Known range of Stenella attenuata in eastern tropical Pacific Ocean (based on Perrin et al., 1983), with numbers of males (above) and females (below) available for each 5° latitude-longitude block (total of 611 specimens). Each block is identified by numerical code (numbers on left and bottom mar- gins are combined; e.g. block 0802 encompasses the northern Hawaiian Islands). Asterisks indicate the 11 blocks included in analysis of sexual dimorphism. Initial analyses indicated that specimens in four blocks (0516, 0517, 0518, 0812) represented a well differentiated inshore form (see Douglas et al., 1984); thus, these were not used in subsequent evaluations. The remaining 29 blocks with two or more specimens were used as the basis for analyses of geographic variation; for some evaluations, the eight blocks with single specimens were projected onto axes based on the other 29 blocks. For the 16 blocks marked with a dot in lower-right corner, sufficient numbers of S. attenuata and S. longirostris (i.e. at least two of each species) were available for interspecific comparison of geographic trends. Correlation, ordination, and clustering After conversion to zwitters, characters were stan- dardized (means=0, standard deviations=l). Prod- uct-moment correlations were computed among characters, and associations among characters were summarized by clustering characters (unweighted pair-group method with arithmetic averages; UPGMA). This technique is a type of hierarchical cluster analysis that also was used to summarize average distance coefficients (Sneath and Sokal, 1973) calculated for all pairs of blocks based on stan- dardized data. Cophenetic correlation coefficients indicate the extent to which distances in resulting dendrograms accurately represented original inter- block morphologic distances. Standardized data also were summarized by us- ing a nonhierarchical if -group method (function- point cluster analysis; described in Katz and Rohlf [1973] and Rohlf et al. [1979]). Through use of this technique, blocks are assigned to subgroups at a spec- ified level. A w-parameter value used in function- point clustering was varied. An hierarchical, but not necessarily nonoverlapping, system of clusters was Pernn et al.: Geographic variation in cranial morphology of Stenella attenuata 327 Table 1 Geographic variation an d sexual d morphism in Stenella attenuata evaluated for 30 characters. Character' F-va ue2 MeanJ Correction factor'' Percentage difference5 Block Sex Male Female 1 Condylobasal L. 10.59*** 1.94 397.0 398.2 -0.74 -0.30 2 L. Rostrum (from Base) 6.94*** 13.49*** 236.7 239.5 -1.55 -1.20 3 L. Rostrum (from Pterygoid) 6.88*** 6.84** 278.5 281.0 -1.31 -0.87 4 W. Rostrum (at Base) 14.20*** 25.13*** 83.3 81.8 0.73 1.85 5 W. Rostrum (at 1/4 L.) 12.25*** 51.82*** 57.1 55.2 0.90 3.32 6 W. Rostrum (at 1/2 L.) 13.82*** 47.61*** 42.4 40.9 0.78 3.78 7 W. Premax. (at 1/2 L.) 8.51*** 72.88*** 23.6 22.4 0.60 5.25 8 W. Rostrum (at 3/4 L.) 8.65*** 73.90*** 29.9 28.0 0.96 6.66 9 Preorbital W. 16.92*** 45.92*** 149.5 146.6 1.43 1.97 10 Postorbital W. 22.46*** 42.23*** 167.8 165.1 1.31 1.63 11 Skull W. (at Zygomatic P.) 21.98*** 51.05*** 167.4 164.5 1.46 1.79 12 Skull W. (at Parietalsl 5.73*** 54.11*** 129.6 136.4 1.58 2.34 13 Ht. Braincase 10.61*** 53.87*** 95.9 93.7 1.04 2.36 14 L. Braincase 15.67*** 39.94*** 113.1 111.2 0.95 1.71 15 Max. W. Premax. 8.89*** 4.45* 66.2 65.7 0.27 0.85 16 W. External Nares 7.38*** 0.03 42.5 42.5 0.01 0.02 17 L. Temporal Fossa 22.41*** 16.79*** 70.1 68.4 0.82 2.53 18 W. Temporal Fossa 20.60*** 28.50*** 55.2 53.2 0.93 3.72 19 Orbital L. 2.97** 0.06 47.4 47.4 0.02 0.10 20 L. Antorbital P. 8.33*** 4.30* 36.9 36.4 0.24 1.43 21 W. Internal Nares 9.58*** 12.80*** 47.4 46.6 0.36 1.58 22 L. Up. Toothrow 6.20*** 10.56** 204.3 206.7 -1.29 -1.16 23 No. Teeth (Up. Lf.) 0.96 1.54 41.4 41.2 0.10 0.47 24 No. Teeth (Up. Rt.l 0.98 0.65 41.3 41.1 0.07 0.33 25 No. Teeth (Low. Lf.) 2.89* 0.02 41.0 40.9 -0.01 -0.02 26 No. Teeth (Low. Rt.) 2.82** 1.02 40.8 41.0 -0.08 -0.36 27 L. Low. Toothrow 5.38*** 11.63*** 198.6 201.2 -1.40 -1.28 28 Ht. Ramus 16.13*** 0.12 57.3 57.1 0.38 0.24 29 Tooth W. 12.14*** 19.65*** 3.4 3.3 0.06 3.55 30 L. Ramus 9.02*** 8.86** 335.5 338.3 -1.48 -0.82 ' Abbreviations: Ht. = height; L. = length; Lf. = left; Low. = lower; Max. = maximum; No. = number; P. = process; Premax. = premaxillary; Rt. = right; Up. = upper; W. = width. 2 F-values from main effects two-way analysis of variance (5' block vs. sex) involving 11 blocks 1 *P< 0.05; **P< 0.01; ***P< 0.001). Total of 170 individuals. Degrees of freedom 10 for among-block variation and 1 for between sexes. 3 Unweighted mean for 11 blocks. 4 Added to all individual female measurements and subtracted from all individual male measurements to correct for sexual differences. 5 Difference between sexes (males minus females) multiplied by 100; the resulting value was divided by average of male and female means. obtained by repeating the analysis at different clus- tering levels. Results are displayed in modified sky- line diagrams (Wirth et al., 1966). Using standardized data, we constructed scatter diagrams by projecting blocks onto the first two principal components (Sneath and Sokal, 1973) ex- tracted from a matrix of correlations among the 30 characters. Canonical-variates analysis also was employed to obtain the subset of variables that shows the greatest interblock separation relative to intrablock variation (Program P7M of BMDP statis- tical software; Dixon, 1990). Plots of the first two canonical variables show the maximum separation of blocks in two-dimensional space. Mantel test for geographic patterning A Mantel ( 1967) test was used to assess interlocality variation in each character and determine whether measures are geographically patterned or, alterna- 328 Fishery Bulletin 92(2). 1994 tively, vary spatially at random. The observed as- sociation between sets of character differences and geographic distances was tested relative to its per- mutational variance, and the resulting statistic com- pared against a Student's /-distribution with infinite degrees of freedom. We performed analyses using GEOVAR, a computer-program library for geo- graphic variation analysis (written by David M. Mallis and provided by Robert R. Sokal, State Uni- versity of New York at Stony Brook). Character differences were compared first with actual geographic distances (in nautical miles) be- tween centers of blocks and then with reciprocals of distances. In evaluations of reciprocals, where dis- tances are scaled in a nonlinear manner, longer dis- tances are considered effectively to be equal, and the portion of the scale involving smaller distances is expanded. Thus, use of reciprocals of distances in- creases the power of analyses to reveal geographic patterns that are "local" in nature (i.e. involving closely placed blocks), whereas tests involving nau- tical-mile distances evaluate "regional" trends. Posi- tive associations of character differences and nau- tical-mile distances are indicated by positive /-val- ues, while negative /-values denote such associations when using distance reciprocals. Douglas et al. ( 1992) provided a simplified example to demonstrate use of the Mantel procedure. We also computed matrix correlations (Sneath and Sokal, 1973) between character differences and the associated geographic distances or reciprocals of distances between localities. The statistical signifi- cance of these coefficients cannot be tested in the conventional way, because all pairs of localities were used and these are not statistically independent. However, the resulting values can be used as de- scriptive statistics indicating the degree of associa- tion of difference values. Morphological-environmental covariation We calculated product-moment correlations of block means for morphological characters with environ- mental variables. Data were available for 13 envi- ronmental variables for the eastern tropical Pacific Ocean (Table 2; data sources summarized in Dou- glas et al., 1992). The list of environmental variables used is somewhat different than that employed by Schnell et al. (1986), because data for some of the variables were not available for all blocks in the broader geographic range being considered in the current study. We also used UPGMA to summarize associations among these environmental variables for 51 blocks; since these two dolphin species have broadly overlapping distributions in the eastern tropical Pacific, the blocks used are representative of areas inhabited by S. attenuata. In order to obtain summary variables reflecting overall environmental trends, we conducted a prin- cipal-components analysis of the 13 environmental variables for 51 blocks with specimens of S. longirostris (Douglas et al., 1992) or S. attenuata or both. Individual blocks were projected onto the re- sulting environmental principal components based Table 2 Environmental measurements compiled for each 5° latitude-longitude block.' 1 Sea Current (N., Winter) — Average northern component (in knots) of surface water current in winter. 2 Sea Current (W., Winter) — Average western component (in knots) of surface water current in winter. 3 Water Depth — Average sea depth (in m). 4 Solar Insolation (Jan.) — Average incoming solar radiation for January (in gmcal/cnr). 5 Solar Insolation (Annual) — Average annual incoming solar radiation in gmcal/cm2). 6 Sea Surface Temp. (Jan.) — Average January sea surface temperature (in°C). 7 Sea Surface Temp. (July) — Average July sea surface temperature (in°C). 8 Sea Surface Temp. (Ann. Var. ) — Average annual sea surface temperature variation (in°C). 9 Oxygen Min. Layer (Depth) — Annual mean depth (in m) of absolute oxygen minimum surface with respect to the vertical. 10 Surface Salinity — Average salinity (%<■) of surface sea water. 11 Thermocline Depth (Winter) — Mean depths (in m) to top of thermocline for January, February, and March. 12 Thermocline Depth (Summer) — Mean depths lin m) to top of thermocline for July, August, and September. 13 Surface Dissolved Oxygen — Annual mean dissolved oxygen (mL/L) of surface sea water. ' Data sources listed in Douglas et al (1992: table 2). Abbreviations: Ann. Var. = annual variation; Jan. = January. Min. = minimum; N. = north; Temp. - temperature; W. = west. Perrin et al.: Geographic variation in cranial morphology of Stenella attenuata 329 on standardized data. These environmental compo- nents served as composite environmental variables for comparisons with morphological variables. Matrix correlations and Mantel tests were used to test for local and regional patterning of environ- mental variables. Also, differences between each pair of blocks for a given morphological variable were com- pared with those for an environmental variable. Interspecific comparisons The predominant trends in the data sets for each of S. attenuata and S. longirostris were summarized with principal components and canonical variables. Information is available for 16 blocks from which both offshore S. attenuata and S. longirostris were sampled (Fig. 1). These blocks are representative of the total geographic range investigated in our stud- ies. In order to compare general patterns of varia- tion in the two species, we calculated product-mo- ment correlations, Mantel tests, and matrix corre- lations for individual morphological characters, prin- cipal-component projections, and canonical-variable projections of these 16 blocks. In our analyses, average distances based on mor- phological characters were computed between each pair of localities. To evaluate the extent of similar- ity in geographic patterns, the original distance matrices for each species were modified such that only distances among the 16 localities common to both species were included. These matrices were then compared by using the Mantel test and com- puting the matrix correlation. Results Sexual dimorphism Table 1 includes mean measurements for males and females based on 11 blocks. For two-way ANOVAs assessing geographic block and sex (df=21, rc=461), all but 5 of the 30 were very highly significant (P<0.001). The probability was 0.02 for number of teeth (lower right) and 0.007 for orbital length. Two characters showed no significant variation (upper tooth counts) and one character (number of teeth [lower left]) was close to significant (P<0.06). Sta- tistically significant interactions (P<0.05) between block and sex were found for five of the characters: condylobasal length; width of rostrum (at 1/4 length); width of rostrum (at 1/2 length); width of rostrum (at 3/4 length); and width of internal nares. Interaction denotes that the degree of sexual dimor- phism differs among blocks for these characters. Sexual dimorphism was significant for 22 of the 30 characters (Table 1). Females had longer ros- trums, which is reflected in a number of characters (i.e. 2, 3, 22, 27, 30). In general, males had wider skulls and tended to be larger for nonrostral por- tions of the skull. Percentage differences between sexes are presented in Table 1. The average abso- lute difference (i.e. sign ignored) between the sexes for the 30 characters was 1.78%. For 8 characters where females were larger, the average difference was 0.75%, whereas for 22 characters where males were larger the average difference was 2.16%. The greatest differences were found for width of premax- illary (at 1/2 length) and width of rostrum (at 3/4 length) — 5.25% and 6.66%, respectively. Table 3 shows the results for the nested ANOVA for cruise sets within blocks. Twenty-three of the 30 characters showed highly significant or very highly significant block effects, whereas only three charac- ters (those involving the temporal fossa and length of braincase) reflected highly significant or very highly significant effects for cruise set. Even in those three cases, block effects were more pronounced. Therefore, we conclude that combining cruise sets into blocks did not have an important confounding in- fluence on geographic patterns found among blocks. Correlation, ordination, and clustering Most character pairs had positive correlations. An exception was tooth counts and temporal fossa mea- surements, which tended to have negative correla- tions with skull width measurements. The dendro- gram in Figure 2 summarizes absolute correlations (i.e. sign of correlation ignored) among characters based on 29 blocks to provide an assessment of char- acter covariation. The width of external nares was the character with the least association with other measures. Tooth characters join and are separated from the remaining morphometric characters. Brain- case measures and skull width (at parietals) clus- ter in another relatively distinct group. The remain- ing characters are arranged in two groups. The clus- ter at the top of Figure 2 includes most length mea- surements and height of ramus. Width measure- ments along with length of antorbital process, or- bital length, and length of temporal fossa are in- cluded in the adjoining major cluster (Fig. 2). Table 4 includes character loadings on the first two principal components based on data for 29 blocks. Component I explains 45.0% of the total variance for the 30 characters, whereas component II summarizes an additional 16.8% (cumulative to- tal of 61.8%). Projections of blocks onto the two com- ponents are depicted in Figure 3, and a map (Fig. 4) is included that renders geographic block projec- tions onto the first component. Component I repre- 330 Fishery Bulletin 92(2). 1994 Table 3 Results of nested ANOVA (F- values) for different cruise sets within 12 latitude -longitude blocks of offshore Stenella attenuata. Character' F-va lue Cruise set Block 1 Condylobasal L. 1.28 3.03*** 2 L. Rostrum (from Base) 1.34 3.25*** 3 L. Rostrum (from Pterygoid) 1.65* 4.36*** 4 W. Rostrum (at Base) 1.25 8.86*** 5 W. Rostrum (at 1/4 L.) 1.60* 7.51*** 6 W. Rostrum (at 1/2 L.) 1.12 8.34*** 7 W. Premax. (at 1/2 L.) 1.29 5.25*** 8 W. Rostrum (at 3/4 L.) 1.01 5.39*** 9 Preorbitai W. 1.25 10.87*** 10 Postorbital W. 1.22 11.31*** 11 Skull W. (at Zygomatic P.) 1.27 10.81*** 12 Skull W. (at Parietals) 0.97 1.11 13 Ht. Braincase 1.55* 1.55 14 L. Braincase 1.83** 3.59*** 15 Max. W. Premax. 1.22 4 01*** 16 W. External Nares 0.95 3.00*** 17 L. Temporal Fossa 2.13*** 7 n;,i*** 18 W. Temporal Fossa 1.87** 9.48*** 19 Orbital L. 1.29 1.24 20 L. Antorbital P. 0.82 5.00*** 21 W. Internal Nares 1.09 3.47*** 22 L. Up. Toothrow 1.33 2.67** 23 No. Teeth (Up. Lf.) 1.23 1.22 24 No. Teeth (Up. Rt.) 1.19 0.78 25 No. Teeth (Low. Lf. ) 1.47* 2.57** 26 No. Teeth (Low. Rt. ) 1.45* 2.57** 27 L. Low. Toothrow 1.63* 2.45** 28 Ht. Ramus 1.09 1.77 29 Tooth W. 1.31 1.45 30 L. Ramus 1.25 2.37 ' ' 1 P < 0.05; **P < 0.01, ***P< 0.001 Abbreviations identified in Footnote 1 of Table 1 Table 4 Principal component loadings for offshore Stenella attenuata involving character means for 29 lati- tude-longitude blocks. Character' Compc nent2 I II 1 Condylobasal L. 0.812 -0.505 2 L. Rostrum (from Base) 0.786 -0.483 3 L. Rostrum (from Pterygoid) 0.795 -0.446 4 W. Rostrum (at Base) 0.867 0.206 5 W. Rostrum (at 1/4 L.) 0.923 0.157 6 W. Rostrum (at 1/2 L.) 0.855 0.308 7 W. Premax. (at 1/2 L.) 0.863 0.161 8 W. Rostrum (at 3/4 L.) 0.868 0.211 9 Preorbitai W. 0.930 0.235 10 Postorbital W. 0.940 0.181 11 Skull W. (at Zygomatic P.) 0.938 0.173 12 Skull W. (at Parietals) 0.740 0.270 13 Ht. Braincase -0.002 0.225 14 L Braincase 0.508 -0.325 15 Max. W. Premax. 0.366 -0.059 16 W. External Nares 0.286 -0.311 17 L. Temporal Fossa -0.544 -0.576 18 W. Temporal Fossa -0.375 -0.541 19 Orbital L. 0.655 -0.018 20 L. Antorbital P. 0.805 0.203 21 W. Internal Nares 0.604 0.000 22 L. Up. Toothrow 0.705 -0.588 23 No. Teeth (Up. Lf. ) -0.137 -0.668 24 No. Teeth (Up. Rt.) -0.112 -0.726 25 No. Teeth (Low. Lf.) -0.301 -0.673 26 No. Teeth (Low. Rt. ) -0.341 -0.635 27 L. Low. Toothrow 0.680 -0.632 28 Ht. Ramus 0.608 0.355 29 Tooth W. -0.129 0.118 30 L. Ramus 0.803 -0.528 ' Abbreviations identified in Footnote 1 of Table 1. 2 Relatively high loadings highlighted in hold as foil iws: 'com- ponent Ii > 10.8 1 . (II) > 10.61. sents general size, with relatively high character loadings (Table 4) for most characters, the excep- tions being tooth characters and the two measure- ments of the temporal fossa. The specimens from blocks in the northeastern portion of the range tend to be small (Figs. 3 and 4), whereas those to the south and southwest typically are larger. The larg- est specimens were found in block 0802, which en- compasses a portion of the Hawaiian Islands. Com- ponent II reflects tooth counts and measurements associated with toothrow length (Table 4). Blocks with relatively high values for these characters are found near the top of Figure 3, whereas those with low values tend to be near the bottom. Blocks with single specimens were not used in the delineation of the principal components but have been projected onto components calculated by using the 29 blocks (Fig. 3). In general, the single-specimen blocks fall close to where one would predict based on their geographic position; some exceptions are ex- pected based simply on expected chance variation. Interblock morphological differences are summa- rized in the phenogram in Figure 5. Two blocks (0312 and 0802) are loosely joined in the most dis- parate cluster. Remaining blocks are divided into two clusters. The one represented at the top of Fig- ure 3 includes the blocks from the south, southwest, and west, whereas the other includes blocks from the northeastern portion of the range. Clusters based on the function-point procedure are summarized in the modified skyline diagram in Figure 6A. The most distinctive block is 0802 (en- Perrin et al.: Geographic variation in cranial morphology of Stenella attenuata 331 0.00 0.25 ■ Correlation 0.50 rC ■5 1 Condylobasal L 30 L. Ramus 3 L Rostrum(frm. Pterygoid) 2 L Rostrum(frm.Base) 22 I. Up. Toothrow 27 L Low Toothrow 28 Ht. Ramus 4 w. Rostrum(at Base) 5W. Rostrum(at 1/4 L) 6W Rostrum(at1/2L) 8 W. Rostrum(at 3/4 L) 7W. Premax(at 1/2L) 9 Preorbital W. 10Postorbital W. 1 1 Skull W.(at Zygomatic P.) 20 L Antorbltal P. 15 Max. W. Premax. 19 Orbital L 21 W. Internal Nares 1 7 L. Temporal Fossa 1 8 W. Temporal Fossa 12 Skull W.(at Panetals) 13 Ht. Braincase 14 L. Braincase 23 No. Teeth(Up.Lf) 24 No. Teeth(Up.Rt) 25 No Teeth(Low Lf ) 26 No Teeth(Low Rt.) 29 Tooth W, 16W. External Nares Figure 2 Correlations among characters based on character means for 29 latitude-longitude blocks. Clustering performed using UPGMA on absolute correlations among characters (i.e. negative signs removed). Cophenetic correlation coefficient is 0.87. 1.0- 0.5 E o O 0.0 -0.5 -1.0 n05io ■ ■ 0314 .0409 ■ 0509 -1.5 -1.0 -0.5 0.0 0.5 Component I 1 0 1 5 2.0 Figure 3 Projections of blocks onto first two principal components based on 30 char- acters. Solid symbols indicate 29 latitude-longitude blocks on which analy- sis was conducted. Open symbols represent blocks with only single specimens projected onto axes generated from 29 blocks with two or more specimens. 332 Fishery Bulletin 92(2). 1994 09 OB 07 OS 05 04 03 02 01 1 tf 1 o° 1! - 1 V ii \j 1 ^ '\ K- irf. I Q 0 U , *, *, *, «, ' ' V V V V Blocks Figure 6 Modified skyline diagrams for 29 latitude-longitude blocks, indicating groups formed using function-point clustering procedures and based on: (A) all 30 characters; (B) five characters that, in combination, best discriminate among blocks (preorbital width, width of temporal fossa, length of temporal fossa, length of rostrum [from pterygoid], length of braincase). For given upvalue (i.e. row), blocks connected in common line are in same cluster. Morphological-environmental covariation Douglas et al. (1992: fig. 9) included a dendrogram summarizing absolute correlations among 13 envi- ronmental variables (listed in Table 2) for blocks having either S. attenuata or S. longirostris or both. These variables were partitioned into five clusters. Sea current (N., winter) is separated by itself, whereas sea current (W., winter) and oxygen mini- mum layer (depth) form a second cluster, which groups with an assemblage of five variables involv- ing surface measures of temperature, oxygen, and salinity (variables 6, 7, 8, 10, and 13). The fourth cluster has the two solar insolation variables (4 and 5), while the fifth includes three measures indicat- ing water and thermocline depths (variables 3, 11, and 12). A principal-components analysis was conducted to obtain variables that would summarize general en- vironmental trends; three components were pre- sented from Douglas et al. (1992: table 6). Highest loadings for environmental variables on principal component I included those for sea surface tempera- tures (variables 6, 7, and 8), particularly July tem- peratures. The correlation with sea surface tempera- ture (annual variable) is negative. The second com- ponent reflected thermocline depth (variable 11 and 12), as well as water depth and surface salinity. The Perrin et al.: Geographic variation in cranial morphology of Stenella attenuata 335 third had relatively high loadings with the two characters involving solar insolation (variables 4 and 5). A more detailed descrip- tion of character associations with the prin- cipal components is available in Douglas et al. (1992). Projection values for environmental princi- pal component I are summarized in Figure 9A for the 29 blocks with larger samples of S. attenuata. It reflects the fact that sea surface temperatures are considerably higher in northern than southern blocks, and that the northern blocks exhibit relatively little an- nual variation in surface temperatures. Block projections on environmental principal com- ponent II are portrayed in Figure 9B, which summarizes the increases in thermocline depth, water depth, and surface salinity as one proceeds west and south. Correlations of morphologic variables, prin- cipal components, and canonical variables with environmental variables and environ- mental principal components are summarized in Table 7. The sea current measures (vari- ables 1 and 2) have virtually no statistical association with morphological characters, while water depth (variable 3) has positive correlations with lengths and widths of the rostrum, as well as principal component I and ca- nonical variable 1 (Table 7). Solar insolation (Jan.), the fourth variable, has larger values in the south; values become smaller to the north. It has significant positive correlations with nine morphologic variables, and negative as- sociations with six others. The negative associations with the two temporal fossa measures are particu- larly strong (width of temporal fossa summarized in Fig. 10A). This environmental variable has rela- tively high correlations with canonical variable 1 and principal component II (Table 7). Not unexpectedly, the fifth variable, solar insola- tion (annual) exhibits high values at the equator. Readings are lower for blocks closer to either pole. It has few significant statistical associations with morphologic characters, although the negative cor- relations with length of braincase and width of tem- poral fossa (Fig. 10A) are relatively high (Table 7). The sixth and seventh environmental measures (surface temperatures in January and July) have negative associations with a number of width mea- surements, as well as with a few length variables (Table 7). They have very strong positive correla- tions with temporal fossa measures. Figure 10 sum- marizes the values for sea surface temperature (July), as well as for the closely associated width of a"'" 0216- "m 0215 pO310 M.5 MhJJ"" Canonical Variable 1 Figure 7 Projections of latitude-longitude blocks onto first two canoni- cal variables based on 30 characters. Solid symbols indicate 25 blocks on which analysis was conducted. Open symbols rep- resent blocks with only single specimens, which were projected onto axes generated from 29 blocks with two or more specimens. temporal fossa (r=0.799; the highest correlation of an environmental and a morphological variable). Sea surface temperature (annual variable), the eighth environmental variable, has significant cor- relations with relatively few morphologic characters (Table 7), although its pattern has affinities with those summarized by principal component II and canonical variable 1. Environmental variable 9, oxy- gen minimum layer (depth), shows very few statis- tically significant correlations with morphological measurements (Table 7). Surface salinity, variable 10, exhibits strong covariation with numerous measurements, particu- larly those involving the anterior portion of the skull (Table 7). It also has high correlations with princi- pal component I (Fig. 4) and canonical variable 1 (Fig. 8B). Salinity, which was depicted in Douglas et al. (1992: fig. 13B) for S. longirostris blocks, shows east-west changes from lower to higher values at a given latitude, as well as a north-to-south trend of increasing values (below 15°N). The eleventh variable, thermocline depth (winter), is summarized in Figure 11B. It has positive corre- lations with 12 morphological measures and a nega- tive correlation with 1 character. The correlation of this environmental variable with skull width (at parietals), shown in Figure 11A, is 0.610. Variable 336 Fishery Bulletin 92(2). 1994 12, thermocline depth (summer), has strong positive correlations with a large number of variables, particu- larly those reflecting measurements in the anterior portion of the skull. Given its covariation with water depth, it is not surprising that ther- mocline depth (winter) has signifi- cant correlations with principal com- ponent I and canonical variable 1. Surface dissolved oxygen (variable 13) has only a few weak statistical associations with morphological char- acters. Environmental principal compo- nent I (Fig. 9A) has a pattern simi- lar to those for sea surface tempera- tures in January and July (variables 6 and 7). The highest correlation (0.733) of this component is with width of temporal fossa (Fig. 10A). The second environmental compo- nent (Fig. 9B) is strongly associated with numerous characters (Table 7), reflecting the general trends from the northeast to the west, southwest, and south. The third component, which is negatively associated with canonical variable 2 (Table 7), has only one strong association with a morphological variable, that being with tooth width (r=-0.680; Table 7). The third environmental component exhibits decreasing values as one moves away from the equator. Tooth width shows an opposite pattern, which is particularly emphasized with the relatively thick teeth in specimens from the Hawaiian Is- lands (block 0802). In Table 8, Mantel ^-values and matrix correlations are provided for associations of environmental vari- ables (including environmental prin- cipal components) with the five morphologic characters selected for inclusion during the canonical-vari- ates analysis. With this approach, covariation patterns are assessed on the basis of difference values between all block pairs. Preorbital width shows a strong association with water depth, the two measures of solar insolation, the sea surface temperatures in January and July, oxygen minimum layer (depth), and surface salinity (Table 8). It also exhibits a pattern that is statistically associated 09 08 07 06 05 04 03 02 01 V O" 1' 0° ll ' : 1 *ri "\jl \f 1 "Sj '\ ^ I a 0 a- J , c» a D 0 □ 0" A <*- e B B . a Q D Q Q ^ ? fl fl fl i A Poslorbltal w nmum yum ^164 0 fl i 1 y rf B B fl \ \ iff - 01 03 05 07 09 11 13 15 17 19 09 08 07 06 05 04 03 02 01 1 0° 10° t: ' i R i »° -^ " V \f 1 ^ ,\ hk >*>'. ^ I a a [0- c» Q D 0 u a vf- B 1 B B B Q fl u a fl 1 B B Canonical Variable 1 298 -^J-067 ^■164 1 1 1 1 i 1 i g Ifl"- ab loi 01 03 05 07 09 11 13 15 17 19 Figure 8 ographic variation in (A) postorbital width and (B) canonical vari le 1. Darkened part of bar indicates value for particular latitude lgitude block. - with all three environmental components. The two measures of the temporal fossa show concordance with patterns for solar insolation (January, as well as annual), all sea surface temperature measures, oxygen minimum layer (depth), surface salinity, and the first environmental principal component. The length of temporal fossa also has a weak statistical Perrin et al.: Geographic variation in cranial morphology of Stenella attenuata 337 association with environmental component III. Table 8 indicates that the pattern for length of ros- trum (from pterygoid) is associ- ated statistically with those for water depth and thermocline depth (summer). This morph- ologic character also is shown to have geographic patterning statis- tically similar to that exhibited by environmental components II and III. For length of braincase, the Mantel tests were significant (but weak) only for sea current (N., winter) and solar insolation (Jan.). The strongest association of length of braincase is with en- vironmental component III; its pattern also is linked statistically to the second environmental com- ponent. Interspecific comparisons The study by Douglas et al. (1992) reported comparable statistical analyses on skulls of S. longiro- stris, a dolphin species that over- laps broadly with S. attenuata in the eastern tropical Pacific. The projections onto the first two prin- cipal components for S. attenuata were evaluated against projec- tions on the two components ob- tained for S. longirostis (for sum- mary information on these compo- nents, see Fig. 3 and Table 4 for offshore S. attenuata, and fig. 3 and table 3 of Douglas et al. [1992] for S. longirostris). A strong correspondence exists be- tween the first principal compo- nents for the two species, as indi- cated by product-moment correlations, Mantel t- tests, and matrix correlations comparing the com- ponent projections. The second principal components for the two studies are not similar (Table 9); they summarize different general trends in variation. A similar interspecific comparison was made of projections of the 16 blocks onto canonical variables (Table 9). The first canonical variable for S. attenuata and that for S. longirostris are virtually identical, reflecting highly concordant geographic patterns for the two species (Table 9). While block projections on the first canonical variables for the Table 6 Association of interlocality character differences with geographic dis- tances (in nautical miles) and the reciprocals of these distances. Results from Mantel tests (t) and matrix correlations (r) for offshore Stenella attenuata. Reciprocal Character' Distance of distance t r t r 1 Condylobasal L. 4 ^g*** 0.460 —4 19*** -0.303 2 L. Rostrum (from Base) 4.34*** 0.449 -4.45*** -0.310 3 L. Rostrum (from Pterygoid) 3.59*** 0.352 -4.10*** -0.276 4 W. Rostrum (at Base) 4.53*** 0.493 -5.07*** -0.365 5 W. Rostrum (at 1/4 L.) 4 jl**+ 0.446 -4.84*** -0.348 6 W. Rostrum (at 1/2 L.) 3.76*** 0.413 —4 73*** -0.343 7 W. Premax. (at 1/2 L.) 1.28 0.134 -2.88** -0.203 8 W. Rostrum (at 3/4 L.) 3.14** 0.347 —4.11*** -0.300 9 Preorbital W. 5.00*** 0.464 -6.83*** -0.445 10 Postorbital W. 4.61*** 0.397 -6.66*** -0.416 11 Skull W. (at Zygomatic P.) 4.22*** 0.380 -6.21*** -0.398 12 Skull W. (at Parietals) 4.81*** 0.583 -4.34* -0.336 13 Ht. Braincase 1.81 0.227 -1.81 -0.143 14 L. Braincase 2.60** 0.339 -1.85 -0.151 15 Max. W. Premax. 1.64 0.159 -3.53*** -0.236 16 W. External Nares 0.78 0.066 -1.38 -0.085 17 L. Temporal Fossa 1.49 0.152 -3.13** -0.216 18 W. Temporal Fossa 0.90 0.097 -2.86** -0.206 19 Orbital L. -0.25 -0.031 -0.35 -0.027 20 L. Antorbital P. 0.89 0.078 -3.65*** -0.229 21 W. Internal Nares 0.44 0.046 -1.09 -0.077 22 L. Up. Toothrow 4.08*** 0.422 -3.93*** -0.274 23 No. Teeth (Up. Lf.) 1.57 0.173 -1.68 -0.123 24 No. Teeth (Up. Rt. 1 3.03** 0.296 -3.68*** -0.247 25 No. Teeth (Low. Lf.) 1.75 0.188 -1.89 -0.135 26 No. Teeth (Low. Rt.) 2.44* 0.283 -2.39* -0.180 27 L. Low. Toothrow 2.99** 0.364 -3.20** -0.249 28 Ht. Ramus 1.74 0.181 -1.69 -0.116 29 Tooth W. 3.36*** 0.398 -3.82*** -0.292 30 L. Ramus 3.02** 0.361 -2.88** -0.221 Component I 4.54*** 0.474 -5.51*** -0.386 Component II 1.42 0.122 -2.75** -0.172 Canonical Variable 1 3.76*** (1 289 -6.69*** -0.393 Canonical Variable 2 1.11 0.143 -1.47 -0.119 * P<0.05; **P<0.01, ***P < 0.001. ' Abbreviations identified in Footnote 1 of Table 1. two species were very similar, the actual characters incorporated into the canonical variables are not the same (see our Table 5 and table 4 of Douglas et al. 1992). Of the five characters entered for each spe- cies, only length of rostrum (from pterygoid) was present in both character sets. However, the first and most important character entered into the analyses — preorbital width for S. attenuata and pos- torbital width for S. longirostris are highly corre- lated (see close association of these two characters in S. attenuata indicated in Fig. 2); because these two characters exhibit very similar variation pat- 338 Fishery Bulletin 92(2). 1994 terns, a canonical-variates analysis typically would not select both for inclusion, since they provide ba- sically the same information for separating blocks. The second canonical variables for the two stud- ies also were compared (Table 9). They showed no statistical association. 09 OB 07 06 05 04 03 02 01 1 0° 1 0° V. ' \ ^ °°rf-^ '-V \ I ^ N "% I e 1 i Llr- o> i 1 i i 1 SL 1 i i 1 i 1 i i 1 D 0 Q A Environmental Principal Componeni I noe2 ^ -100 Q Q □ D tf 0 0 0 I0°- 01 03 05 07 09 11 13 15 17 19 09 08 07 06 05 04 03 02 01 1 ■0° 1 0° i; - : ^ * 0° jJ ' \> V 1 ^ ^ k =fc I - D '0- o> D 0 0 a Q* 10" - e e ■ B 0 a a D D - 0 0 D B Environmental Principal Componeni II ,03 .y-oao "■0 50 Q 0 a a a a a 0 i0"- Ge pr of 01 03 05 07 09 11 13 15 17 '9 Figure 9 ographic variation in environmental variables as summarized in (A incipal component I and (B) principal component II. Darkened par bar indicates value for particular latitude-longitude block. ) t Table 10 includes results of Mantel tests, matrix correlations, and product-moment correlations of in- terspecific comparisons for individual morphological characters. Thirteen of the 30 lvalues for Mantel tests of interlocality differences for the same char- acter in the two species were significant, while 15 of 30 product-moment correlations indicated statistical associations. Nine of 11 characters with positive correlations were width measures. Furthermore, a tenth (length of antorbital process) is essentially a width character as well (for illustra- tion of measurement, see Schnell et al. 1985). The two characters involv- ing upper tooth counts, as well as length of temporal fossa, exhibited significant negative correlations. For S. attenuata, upper tooth counts tend to be higher for the western blocks (but not for the Hawaiian Is- land block), whereas in S. longirostris, higher upper tooth counts are found in the Hawaiian and eastern blocks. The length of temporal fossa is greater in northern localities of S. attenuata (Fig. 10A), whereas the shorter fossae are found in northeastern blocks for S. longir- ostris (see Douglas et al. 1992: fig. 11). Discussion Sexual dimorphism Schnell et al. (1985) conducted the most recent analysis of sexual di- morphism of S. attenuata in cranial morphology. They found statistically significant dimorphism for 23 of 36 characters. Our analyses used many of the same specimens, with some added and some deleted, and 30 of the same characters. For the 30 characters we analyzed, Schnell et al. ( 1985) found the same 22 to have statistical differences between sexes (one statistically significant charac- ter analyzed earlier was not used in our analysis). Results from the two studies on sexual dimorphism are essentially the same. Thus, for S. attenuata, our current findings sim- ply update information in Schnell et Perrin et al.: Geographic variation in cranial morphology of Stenella attenuata 339 Table 7 Product-mome/it correlations of block means for morphological variables and components versus environmental variables and components based on 29 latitude-longitude blocks of offshore Stenella attenuata.1 Environmental Character2 Environmental variable3 component 1 2 3456789 10 11 12 13 I II III 1 Condylobasal L. ++ + ++ +++ +++ 2 L. Rostrum (from Base) ++ ++ +++ +++ 3 L. Rostrum (from Pterygoid) ++ + + +++ ++ 4 W. Rostrum (at Base) ++ +++ ++ +++ +++ 5 W. Rostrum (at 1/4 L.) ++ - +++ + + + - - ++ 6 W. Rostrum (at 1/2 L.) + + + +++ + + 7 W. Premax. (at 1/2 L.) + + _ _ _ _ ++ + 8 W. Rostrum (at 3/4 L.) ++ + +++ + + 9 Preorbital W. +++ + +++ + ++ -- ++ 10 Postorbital W. +++ + +++ + + - - ++ 11 Skull W. (at Zygomatic P.) ++ + +++ ++ + 12 Skull W. (at Parietals) +++ ++ 13 Ht. Braincase ++ 14 L. Braincase - - . ++ 15 Max. W. Premax. + ++ ++ 16 W. External Nares - 17 L. Temporal Fossa +++ +++ - - +++ 18 W. Temporal Fossa +++ +++ +++ 19 Orbital L. ++ 20 L. Antorbital P. ++ + -- ++ + + 21 W. Internal Nares - - - - + ++ - 22 L. Up. Toothrow - + ++ +++ +++ 23 No. Teeth - (Up. Lf.) + + - 24 No. Teeth (Up. Rt.) + 25 No. Teeth (Low. Lf.) + + + 26 No. Teeth (Low. Rt.) + - 27 L. Low. Toothrow + + ++ ++ 28 Ht. Ramus 29 Tooth W. _ 30 L. Ramus + ++ +++ ++ Component I ++ +++ + +++ ++ Component II ++ -- +++ + ++ Canonical Variable 1 +++ +++ + ++ +++ ++ + Canonical Variable 2 - ' Blanks indicate nonsign ficant correlations. Individual symbols refer to significant positive or negative correlations (P<0.05; greater than 0.367). double symbols ndicate highly significant correlations (f<0.01; greater than 0.470), and ti iple symbols represent very highly significant correlations P<0.001; greater than 0.580). 2 Abbreviations identified in Footnote 1 of Table 1. 3 Environmental variable; : (1) Sea Current (N., winter); (2) Sea Current (W., winter); (3) water depth (4) solar insolation (Jan.); (5) solar insolation (annual); (6) sea surface te mp. (Jan.); (7) sea surface temp. (July); (8) sea surface temp, (anr . var.); (9) oxygen min. layer (depth); (10) surface salinity; (11) thermocline depth (winter); (12) thermocline depth (summer); and (13) surface dissolved oxygen. 340 Fishery Bulletin 92(2). 1994 al. (1985) to reflect a modified sample size and a re- duced character set. Douglas et al. (1992) analyzed sexual dimorphism in skull measures for S. longirostris from the east- ern tropical Pacific. They found 15 of the 30 char- acters to be statistically different between sexes. Since S. longirostris samples are somewhat smaller, w. Temporal Fossa 0 Ot 03 OS 07 08 1 B° 0° 0° * R [i t i 1 i 1 1 A OS 1 1 1 1 i 1 i 1 1 ij \ 1 ..-J e y B "« B Sea Surface Temp (Juty) u D 0 D rl J 2 0 D a \ V , 1 K 7<> Figure 1 0 Geographic variation in (A) width of temporal fossa and (B) sea sur- face temperature (July). Darkened part of bar indicates value for par- ticular latitude-longitude block. one might expect fewer significant differences in this species simply due to sample size. Nevertheless, in- spection of the results indicates support for the con- clusion reached by Douglas et al. (1986: 542-543) "that the degree of sexual dimorphism in spotted dolphins is greater than in spinner dolphins." They also pointed out that "the trends are basically the same for both species, suggesting that common behavioral and/or eco- logical factors are influencing sexual dimorphism in these dolphins." Geographic variation From an initial group of specimens, Perrin et al. (1979a) described differ- ences between dolphin skulls avail- able from southern areas and those from more northerly locations. Schnell et al. (1986), based on larger sample sizes, indicated that avail- able information "strongly implies a significant degree of isolation between northern and southern forms." They did not have specimens from west of 125°W and called for additional mate- rial from west of 120°W to help clarify the relationship between southern S. attenuata and other populations, particularly in light of the notation by Perrin et al. ( 1979a) of similarities of specimens from the southern group with those from Ha- waii. In the eastern portion of the range, the subdivision between northern and southern offshore S. attenuata found previously by Perrin et al. (1979a) and Schnell et al. (1986) was confirmed by our analy- ses with a geographically expanded specimen base. In general, blocks to the west (including those from the waters adjacent to Hawaii ) are more like the southern blocks than blocks of the northeast. We found a general concentric pattern of geographic variation (see Fig. 8B), much like that established by Douglas et al. ( 1992) for the broadly overlapping S. longirostris. This also was suggested by Perrin et al. (1985). Reilly (1990) provided some in- sight as to possible reasons why samples of S. attenuata and S. longi- rostris from the south, southwest, Perrin et al.: Geographic variation in cranial morphology of Stenella attenuata 341 4 Hedgepeth, J. B. 1985. Database for dolphin tagging operations in the eastern tropical Pacific, 1969-1978, with discussion of 1978 tagging results. Southwest Fisheries Center Admin. Rep. No. LJ-85-03, 40 p. and west would show close morphologic affinities. He analyzed large-scale dolphin distribution pat- terns and environmental patterns based on re- search-vessel surveys conducted in the eastern tropi- cal Pacific from June through November, comparing his results with those of Au and Perryman (1985). Reply's ( 1990 ) distributional compari- sons between seasons indicated that along 10°N S. attenuata and S. longirostris occur in relatively high density west of 120°N during the summer. Furthermore, they were not in high densities along 4°N between 90 and 120°W, and along 6°N between 88 and 110°W — regions with rela- tively high concentrations of these two species in the winter (Au and Perryman, 1985). Reilly. (1990) indi- cated that "One hypothesis suggested by these complementary changes is an intraregional, seasonal move- ment." Data from mark-recapture ef- forts (Perrin et al., 1979b; Hedgepeth4) are consistent with re- spect to the hypothesized direction of such migrations, although the dis- tances are greater than those sug- gested by the very limited data from these studies. Reilly ( 1990) also noted that the suggested movement pat- terns are at least partially explainable based on seasonal atmospheric and oceanographic changes in the region. Morphological-environmental covariation In the earlier study of S. attenuata, Schnell et al. (1986) assessed envi- ronmental-morphological covariation for a similar, although not identical, set of environmental parameters. Since their investigation was re- stricted largely to eastern blocks, dif- ferent findings with respect to covariation are possible. Schnell et al. (1986) noted that the strongest mor- phological-environmental associa- tions involved solar insolation (Jan.). Sea surface temperatures also co- varied with a number of morphological characteris- tics, as did oxygen minimum layer (depth). The environmental principal components indi- cated that a number of environmental measures have a north-south component (see Fig. 9A), while others (particularly thermocline depth and water Skull W (al Panetals) a 144 5 mm 140 J 1362 J I I D D 0 0 □ D 05 07 15 17 i D° i0° •■ : ^ !^ 1 »' g) *~\> [ 1 ^ p*s V i 20°. A, I Q \ 0 fo- / . c> U y ri D 0" U 0 0 Q . 0 0 D 0 - Q D - B Thermocline Depth (Winter) Q U D 0 tf Q 0 0 ■ i 81 1 05 07 11 13 17 19 Figure 1 1 Geographic variation in (A) skull width (at parietals) and (B) ther- mocline depth (winter). Darkened part of bar indicates value for par- ticular latitude-longitude block. 342 Fishery Bulletin 92(2), 1994 Table 8 Results of Mantel tests (t) and matrix correlations (r) for offshore Stenella attenuata. Comparison of interlocality differences for 13 environmental variables and 3 environmental components against those for five morphological variables selected in canonical-variates analysis. Environmental variable Preorbital width Width of temporal fossa Length of temporal fossa Length of rostrum (from pterygoid) Length of braincase ( r 1 r t r t r ( r 1 Sea Current (N., winter) -1.34 -0.114 2.43* 0.318 -0.20 -0.026 0.67 0.085 2.41* 0.322 2 Sea Current (W.,winter) 0.02 0.001 -1.95 -0.139 -1.68 -0.119 1.94 0.136 0.82 0.059 3 Water Depth 4.12*** 0.259 -0.45 -0.033 -1.08 -0.080 3.24** 0.236 1.17 0.088 4 Solar Insolation (Jan.) 7.38*** 0.466 6.39*** 0.479 6.47*** 0.481 -0.56 -0.041 -2.29* -0.173 5 Solar Insolation (ann.) 7.58*** 0.502 2.85** 0.238 2.91** 0.241 0.68 0.056 -0.84 -0.071 6 Sea Surface Temp. (Jan.) 5.53*** 0.371 5.71*** 0.493 4.70*** 0.401 -1.15 -0.097 -1.43 -0.125 7 Sea Surface Temp. (July) 5.59*** 0.369 6.97*** 0.579 4.08*** 0.335 -0.94 -0.076 -1.30 -0.109 8 Sea Surface Temp, (ann. var.) 1.69 0.114 6.23*** 0.547 2.87** 0.249 -0.74 -0.064 -0.82 -0.073 9 Oxygen Minimum Layer (depth) 4.18*** 0.236 3.60*** 0.191 3.75*** 0.200 -0.56 -0.030 -0.79 -0.042 10 Surface Salinity 9.52*** 0.594 5.31*** 0.386 4.66*** 0.337 0.32 0.023 -1.20 -0.088 11 Thermocline Depth (winter) 0.40 0.026 0.90 0.070 1.88 0.144 0.20 0.016 -0.39 -0.031 12 Thermocline Depth (summer) 1.21 0.085 -0.07 -0.007 0.84 0.079 2.60** 0.240 0.00 0.000 13 Surface Dissolved Oxygen -1.85 -0.147 0.76 0 08!) 1.83 0.211 -0.40 -0.045 -0.01 -0.001 Environmental Component I 3.71*** 0.238 6.75*** 0.476 4.57*** 0.309 -1.69 -0.111 -1.31 -0.104 Environmental Component II 3.77*** 0.364 -0.26 -0.030 0.39 0.041 3.45*** 0.352 2.63** 0.359 Environmental Component III 4.64*** 0.449 0.52 0.059 2.36* 0.252 2.22* 0.227 4.54*** 0.623 *P<0.05; **P<0.01; ***P<0.001. depth) show trends from the east to the west, south- west, and south. Not unexpectedly, a mosaic of varia- tion patterns is present in the suite of morphologic characters we assessed. Some, like width of tempo- ral fossa (Fig. 10A), align closely with environmen- tal variables — such as sea surface temperature (July) (Fig. 10B) — subsumed under environmental component I (Fig. 9A). Others, like skull width (at parietals) (Fig. 11A), display patterns similar to those of thermocline depth (winter) (Fig. 11B) and other environmental measures summarized by en- vironmental component II (Fig. 10B). However, the overall, general morphological trend is reflected best by projections onto the first canonical variable based on morphologic data (Fig. 4), which has a relatively strong negative correlation with environmental com- ponent I and a weaker positive one with environ- mental component II. By adding the more westerly blocks to the analy- sis, environmental-morphological covariation pat- terns that emerged, in some cases, were different from those reported by Schnell et al. (1986). For example, the previous statistically significant corre- lations they found for several morphologic charac- ters with sea current (N., winter) and sea surface temperature (annual variation) were not repeated Perrin et al.: Geographic variation in cranial morphology of Stenella attenuata 343 in our expanded study. Likewise, strong associa- tions with measures of solar insolation were sub- stantially reduced for all but a few characters (i.e. length of braincase, length of temporal fossa, and width of temporal fossa; Table 7). Some charac- ters, such as sea current (W., winter) and oxygen minimum layer (depth), did not have variation patterns in either study that corresponded to those for morphologic variables. Water depth, however, has significant correlations with more characters now that western blocks have been added. At least three environmental measures — sea surface temperature (Jan.), sea surface tem- perature (July) (Fig. 10B), and surface salinity — had relatively strong covariation patterns that stayed relatively constant through the two studies. Availability in the future of bet- ter information on environmental variation and, possibly, on other relevant parameters reflecting environmental heterogeneity will allow researchers to analyze dol- phin-environmental covariation in a more meaningful way. For ex- ample, better environmental data and more comprehensive informa- tion on dolphin feeding ecology could provide a basis for testable predictions concerning why cer- tain morphological characters covary with specific environmen- tal variables. Our admittedly de- scriptive analyses demonstrate some striking cases of dolphin- environmental covariation and, thus, provide initial guidance in terms of possible causal relation- ships that may be examined with greater sophistication by future investigators. Interspecific covariation In this paper we have presented statistical data for trends in geo- graphic covariation of S. longiro- stris and S. attenuata skulls from 16 blocks for which samples of both species were available (brief comments on covariation were included by Douglas et al. 1992). Geographic patterns in 13 of the 30 morphological characters showed statistical correspondence based on Mantel tests, whereas Table 9 Comparisons of principal component and canonical variable projections for offshore Stenella attenuata and S. longirostris based on corresponding data for 16 lati- tude-longitude blocks. Principal component Canonical variable Statistic I (I Product-moment correlation Mantel lvalue Matrix correlation 0.745*** 3.46** 0.528 0.289 0.32 0.039 0.896*** 6.65*** 0.741 -0.328 1.62 0.204 *P<0.05; **P<0.01; ***P<0.001. P-values for Mantel tests based on number of times permutational Z-values less than or equal to observed Z-values (one-tailed test). Table 10 Results of Mantel tests, matrix correlations, and product-moment cor- relations for 16 overlapping latitude-longitude blocks of offshore Stenella attenuata and S. longirostris. Comparison of interlocality dif- ferences for 30 morphological variables. Mantel Matrix Product-moment Character' t-value correlation correlation 1 Condylobasal L. 1.21 0.184 0.256 2 L. Rostrum (from Base) 1.28 0.178 0.159 3 L. Rostrum (from Pterygoid) 0.55 0.074 0.150 4 W. Rostrum (at Base) 4.65*** 0.626 0.837*** 5 W. Rostrum (at 1/4 L.) 4.14*** 0.337 0.693** 6 W. Rostrum (at 1/2 L.) 4.57*** 0.495 0.775*** 7 W. Premax. (at 1/2 L.) -0.49 -0.066 0.351 8 W. Rostrum (at 3/4 L.) 1.01 0.120 0.540* 9 Preorbital W. 5.70*** 0.737 0.865*** 10 Postorbital W. 6.24*** 0.693 0.861*** 11 Skull W. (at Zygomatic P.) 6.37*** 0.732 0.883*** 12 Skull W. (at Parietals) 2.71** 0.446 0.616* 13 Ht. Braincase 0.60 0.093 0.294 14 L. Braincase 2.91* 0.494 0.613* 15 Max. W. Premax. 2.88** 0.318 0.675** 16 W. External Nares -0.64 -0.073 0.224 17 L. Temporal Fossa 0.79 0.088 -0.499* 18 W. Temporal Fossa 0.62 0.077 -0.443 19 Orbital L. 0.92 0.148 0.383 20 L. Antorbital P. 4.10** 0.396 0.658** 21 W. Internal Nares 1.61 0.205 0.613* 22 L. Up. Toothrow 1.19 0.165 0.069 23 No. Teeth (Up. Lf. ) 2.98** 0.501 -0.753*** 24 No. Teeth (Up. Rt.) 2.99** 0.517 -0.691** 25 No. Teeth (Low. Lf.) 0.03 0.004 -0.223 26 No. Teeth (Low. Rt.) 0.66 0.104 -0.428 27 L. Low. Toothrow 0.59 0.090 0.015 28 Ht. Ramus 2.34* 0.298 0.408 29 Tooth W. 0.20 0.038 -0.260 30 L. Ramus 0.70 0.111 0.174 P<0.05; **P<0.01; ***P<0.001. P-values for Mantel tests based on number of times per- mutational Z-values less than or equal to observed Z-values (one-tailed test). Abbreviations identified in Footnote 1 of Table 1. 344 Fishery Bulletin 92(2). 1994 product-moment correlations were significant for 15 characters. Some of these associations were very strong (i.e. correlations as high as 0.883; see Table 10). Given that we are comparing two species, enti- ties that have independent gene pools, the most likely explanation for common positive trends in morphologic covariation is that the two species are being subjected to similar forces of natural selection. However, we found several striking examples where covariation was negative. A significant nega- tive association was found for length of temporal fossa, while the negative correlation for width of temporal fossa was nearly significant statistically. The number of upper teeth (characters 23 and 24) also show significant negative correlations (Table 10). Muscles associated with the feeding apparatus are positioned in the temporal fossa; obviously, tooth numbers could be related to prey types taken. One suspects that the presence of antithetical trends in these particular skull characteristics is a result of competitive interactions involving these two inter- acting species. This may be an example of ecologi- cal character displacement related to differences in feeding and food types taken (Perrin, 1984). Cer- tainly, the two species are found in close association over much of the eastern tropical Pacific (Au and Perryman, 1985; Reilly, 1990); 49% of S. attenuata schools included some S. longirostris, while 73% of the schools of the less common S. longirostris in- cluded S. attenuata (Reilly, 1990). The information available to date indicates that spinner and spotted dolphins may have different feeding habits or preferences in areas of co-occur- rence in the eastern tropical Pacific. Based on analy- sis of stomach contents of spinner and spotted dol- phins captured together in purse-seine hauls in the tuna fishery, Perrin et al. (1973) concluded tenta- tively that while some prey species are taken by both, spinner dolphins in the mixed-species associa- tions specialize in small mesopelagic fishes (mainly myctophids and gonostomatids) and squids, whereas the spotted dolphins consume larger and more epi- pelagic species such as flying fishes, small scom- broids (e.g. Auxis sp.), and larger squids. In addi- tion, state of digestion of the stomach contents in- dicted that the spinner and spotted dolphins had fed at different times of the day. Stomachs of a spotted dolphin from Hawaii, where the two species do not school together, and from two spotted dolphins from the far western portion of the range in the eastern Pacific, where the mixed species associations are less common than in the core area of the tuna fish- ery off Mexico and Central America (Au and Perryman, 1985; fig. 11), contained a large propor- tion of small mesopelagic species like those eaten by the spinner dolphin in the mixed species associa- tions farther to the east. This geographic variation in feeding habits may reflect resource partitioning where the two species associate closely, which in turn may be manifested in morphological character displacement. Genetic subdivision, management units, and implications of cranial variation The results suggest that gene flow is not uniform throughout the range of S. attenuata in the eastern Pacific; the morphological heterogeneity probably reflects genetic subdivision, a conclusion also reached as a result of the earlier study by Schnell et al. (1986). A similar inference was drawn by Dou- glas et al. (1992) based on morphologic studies of S. longirostris. For S. attenuata, we found that 93.3% of the 30 morphologic characters had statistically significant geographic variation, with 60.0% exhib- ited regional patterning and 73.3% local patterning. This geographic partitioning of morphologic variance was demonstrable even with pooling of specimens taken over a number of months and years, a process that would tend to shroud such relationships. The boundaries of management units presently employed (Perrin et al., 1985) are not fully consis- tent with the general pattern of morphologic varia- tion described here. It appears that animals west of about 120°W longitude have greater affinity with those in the Southern Offshore management unit than they do with S. attenuata from the eastern portion of the present Northern Offshore unit, the unit in which they are presently included. Further, the present boundary between the Northern and Southern Offshore units, at 1°S latitude, is probably too far south; a boundary at about 5"N would be more consistent with the general pattern of cranial variation. The proposed Northern Offshore unit bounded by 5°N and 120°W would be nearly congruent with the "conservation zone" suggested for S. longirostris (Perrin et al., 1991). This is to be expected based on the correlated trends of covariation with environ- mental variables; the two species, as they exist in this region, apparently are parts of an endemic fauna uniquely adapted to the far eastern tropical Pacific and, as such, are "evolutionarily significant units" (Dizon et al., 1992). The cranial results are only one line of evidence useful for delineation of management units; others, such as patterns in movements, external morphol- ogy or life-history parameters, also should be taken into consideration. For example, other data may indicate that S. attenuata west of 120°W differ sig- Perrin et al.: Geographic variation in cranial morphology of Stenella attenuata 345 nificantly in some regard from those south of 5°N and should be managed separately (yielding three management units instead of the present two). Cer- tainly, the specimens from the Hawaiian Islands are generally larger than those from the south, south- west, and west with which they have their closest morphologic affinity. Note added in proof Based in part on the results and recommendations in this paper, the manage- ment unit boundaries for the offshore spotted dol- phin in the eastern Pacific have been changed. The boundary of a new "Northeastern stock" runs south along the 120°W meridian to 5°N and then east to the mainland. Offshore spotted dolphins outside this zone to the west and south are now part of a "West- ern/Southern stock." February, 1994. Acknowledgments Support for aspects of this research was received by the University of Oklahoma through Contract 79-ABC-00167 from the U.S. Department of Com- merce, National Oceanographic and Atmospheric Administration, and Purchase Orders 84-ABA- 02177 and 40JGNF0532 from the National Marine Fisheries Service, Southwest Fisheries Science Cen- ter, La Jolla, CA. We thank individuals (listed in Douglas et al., 1992) who provided access to mu- seum specimens. M. E. Douglas measured the ini- tial set of specimens used in this study. Pamela A. Pogorelc assisted with preparation of some of the figures. Paul Fiedler, Steven Reilly, Michael Scott, and an anonymous reviewer provided useful sugges- tions on earlier drafts of this paper. Literature cited Allen, R. A. 1985. Dolphins and the purse-seine fishery for yel- lowfin tuna. In J. R. Beddington, R. J. H. Beverton and D. M. Lavigne (eds.), Marine mam- mals and fisheries, p. 236-252. Allen and Unwin, London, 354 p. Au, D. W. K., and W. L. Perryman. 1985. Dolphin habitats in the eastern tropical Pacific. Fish. Bull. 83:623-643. Dailey, M. D., and W. F. Perrin. 1973. Helminth parasites of porpoises of the genus Stenella in the eastern tropical Pacific, with de- scriptions of two new species: Mastigonema stenellae gen. et sp. n. (Nematoda:Spiruroidea) and Zalophotrema pacificum sp. n. (Trematoda: Digenea). Fish. Bull. 71:455-471. Dixon, W. D. (chief ed.). 1990. BMDP statistical software. Vol. 1. Univ. California Press, Berkeley, CA. Dizon, A. E., C. Lockyer, W. F. Perrin, D. P. DeMaster and J. Sisson. 1992. Rethinking the stock concept: a phylogenetic approach. Conserv. Biol. 6:24-36. Douglas, M. E., G. D. Schnell, D. J. Hough, and W. F. Perrin. 1992. Geographic variation in cranial morphology of spinner dolphins (Stenella longirostris) in the east- ern tropical Pacific Ocean. Fish. Bull. 90:54-76. Douglas, M. E., G. D. Schnell, and D. J. Hough. 1984. Differentiation between inshore and offshore spotted dolphins in the eastern tropical Pacific Ocean. J. Mammal. 65:375-387. 1986. Variation in spinner dolphins (Stenella lon- girostris) from the eastern tropical Pacific Ocean: sexual dimorphism in cranial morphology. J. Mammal. 67:537-544. Hall, M. A, and S. D. Boyer. 1987. Incidental mortality of dolphins in the east- ern tropical Pacific tuna fishery in 1985. Rep. Int. Whal. Commn. 37:361-362. 1988. Incidental mortality of dolphins in the east- ern tropical Pacific tuna fishery in 1986. Rep. Int. Whal. Commn. 38:439-441. 1989. Estimates of incidental mortality of dolphins in the eastern Pacific fishery for tropical tunas in 1987. Rep. Int. Whal. Commn. 39:321-322. 1990. Incidental mortality of dolphins in the tuna purse-seine fishery in the eastern Pacific Ocean dur- ing 1988. Rep. Int. Whal. Commn. 40:461-162. 1991. Incidental mortality of dolphins in the tuna purse-seine fishery in the eastern Pacific Ocean dur- ing 1989. Rep. Int. Whal. Commn. 41:507-509. 1992. Estimates of incidental mortality of dolphins in the purse-seine fishery for tunas in the eastern Pacific Ocean in 1990. Rep. Int. Whal. Commn. 42:529-531. Katz, J. O., and F. J. Rohlf. 1973. Function-point cluster analysis. Syst. Zool. 22:295-301. Mantel, N. 1967. The detection of disease clustering and a generalized regression approach. Cancer Res. 27:209-220. Perrin, W. F. 1984. Patterns of geographical variation in small cetaceans. Acta Zool. Fennica 172:137-140. Perrin, W. F., R. R. Warner, C. H. Fiscus, and D. B. Holts. 1973. Stomach contents of porpoise, Stenella spp., and yellowfin tuna, Thunnus albaeares, in mixed- species aggregations. Fish. Bull. 71:1077-1092. Perrin, W. F., P. A. Sloan, and J. R. Henderson. 1979a. Taxonomic status of the 'southwestern stocks' of spinner dolphin, Stenella longirostris, and spotted dolphin, S. attenuata. Rep. Int. Whal. Commn. 29:175-184. 346 Fishery Bulletin 92(2). 1994 Perrin, W. F., W. E. Evans, and D. B. Holts. 1979b. Movements of pelagic dolphins (Stenella spp.) in the eastern tropical Pacific as indicated by results of tagging, with summary of tagging operations, 1969-1976. U.S. Dep. Commerce, NOAA Tech. Rep. NMFS SSRF-737, 14 p. Perrin, W. F., M. D. Scott, G. J. Walker, F. M. Ralston, and D. W. K. Au. 1983. Distribution of four dolphins (Stenella spp. and Delphinus delphis) in the Eastern Tropical Pacific, with an annotated catalog of data sources. U.S. Dep. Commerce, NOAA Tech. Memo. NMFS, 65 p. Perrin, W. E, M. D. Scott, G. J. Walker, and V. L. Cass. 1985. Review of geographical stocks of tropical dol- phins (Stenella spp. and Delphinus delphis) in the eastern Pacific. U.S. Dep. Commerce, NOAA Tech. Rep. NMFS 28, 28 p. Perrin, W. F., P. A. Akin, and J. V. Kashiwada. 1991. Geographic variation in external morphology of the spinner dolphin, Stenella longirostris, in the eastern Pacific and implications for conservation. Fish. Bull. 89:411^128. Reilly, S. B. 1990. Seasonal changes in distribution and habitat differences among dolphins in the eastern tropi- cal Pacific. Mar. Ecol. Prog. Ser. 66:1-11. Rohlf, F. J., J. Kishpaugh, and D. Kirk. 1979. NT-SYS. Numerical taxonomy system of multivariate statistical programs. State Univ. New York, Stony Brook, New York. Schnell, G. I)., M. E. Douglas, and D. J. Hough. 1985. Sexual dimorphism in spotted dolphins (Stenella attenuata) in the eastern tropical Pacific Ocean. Mar. Mamm. Sci. 1:1-14. 1986. Geographic patterns of variation in offshore spotted dolphins (Stenella attenuata) of the east- ern tropical Pacific Ocean. Mar. Mamm. Sci. 2:186-213. Sneath, P. H. A, and R. R. Sokal. 1973. Numerical taxonomy. W. H. Freeman and Co., San Francisco. Wirth, M., G. F. Estabrook, and D. F. Rogers. 1966. A graph theory model for systematic biology with an example for the Oncidiinae (Orchidaceae). Syst. Zool. 15:59-69. ADStr3Ct. — A time-dependent energy-flow model was used to examine how mortality affects oys- ter populations over the latitudi- nal gradient from Galveston Bay, Texas, to Chesapeake Bay, Vir- ginia. Simulations using different mortality rates showed that mor- tality is required for market-site oysters to be a component of the population's size-frequency distri- bution; otherwise a population of stunted individuals results. As mortality extends into the juvenile sizes, the population's size fre- quency shifts toward the larger sizes. In many cases adults in- crease despite a decrease in over- all population abundance. Simula- tions, in which the timing of mor- tality varied, showed that oyster populations are more susceptible to population declines when mor- tality is restricted to the summer months. Much higher rates of win- ter mortality can be sustained. Comparison of simulations of Galveston Bay and Chesapeake Bay showed that oyster popula- tions are more susceptible to in- tense population declines at higher latitudes. The association of popu- lation declines with disease agents causing summer mortality and the increased frequency of long-term declines at high latitudes result from the basic physiology of the oyster and its population dynam- ics cycle. Accordingly, management decisions on size limits, seasons and densities triggering early clo- sure must differ across the latitu- dinal gradient and in populations experiencing different degrees of summer and winter mortality rela- tive to their recruitment rate. More flexible size limits might be an important management tool. When fishing is the primary cause of mortality, populations should be managed more conservatively in the summer. The latitudinal gra- dient in resistance to mortality requires more conservative man- agement at higher latitudes and different management philoso- phies from those used in the Gulf of Mexico. Modeling oyster populations. IV: Rates of mortality, population crashes, and management* Eric N. Powell Department of Oceanography. Texas A&M University College Station. TX 77843 John M. Klinck Eileen E. Hofmann Center for Coastal Physical Oceanogrphy, Crittenton Hall Old Dominican University. Norfolk. VA 23529 Sammy M. Ray Department of Marine Biology, Texas A&M University College Station, TX 77843 One of the unfortunate character- istics of oyster Crassostrea virgin- ica populations is their susceptibil- ity to periods of heavy mortality, which can extend from a few months to a few years in duration. Oyster population abundances drop precipi- tously during these times and may remain low for extended periods (Schlesselman, 1955; Engle, 1956; Laird, 1961; Engle and Rosenfield, 1962). Why populations decline over several years or crash over shorter periods of time can usually be ex- plained by killing floods (Andrews et al., 1959; Soniat and Brody, 1988; Soniat et al., 1989) or disease epi- zootics (Needier and Logie, 1947; Andrews and Hewatt, 1957; Mackin and Hopkins, 1962) although preda- tors and overfishing have occasion- ally received some credit (Moore and Pope, 1910; Menzel et al., 1957; Quastet al., 1988). A review of the literature shows that declines and crashes in oyster populations have some interesting characteristics (Mackin and Wray, 1950; Mackin et al., 1950; Menzel, 1950, a and b; Menzel and Hop- kins, 1953; Owen, 1953; Gunter, 1955; Mackin and Sparks, 1962; Hofstetter et al., 1965; Copeland and Hoese, 1966; Hofstetter, 1966; Gilmore et al., 1975; and previously cited references): 1 With the exception of killing floods, the times of the year with the most intense mortality are usually restricted to the summer and early fall and to areas of higher salinity. Warm tempera- tures and high salinities pro- mote the growth of the disease- producing organisms Perkinsus marinus and Haplosporidium nelsoni (Ray and Chandler, 1955; Andrews and Hewatt, 1957) and predation by such pests as the oyster drill, Thais haemastoma (Garton and Stickle, 1980; Stickle, 1985). 2 Population crashes or significant declines have been documented throughout the oyster's latitudi- nal range. However, except for permanent changes in salinity, Manuscript accepted 12 October 1993 Fishery Bulletin 92:347-373 (1994) Parts I-III have been published in the Journal of Shellfish Research (Part I in 11:387- 398; Part III in 11:399-416; Part II is in press). 347 348 Fishery Bulletin 92(2). 1994 owing to levee building for instance (Mackin and Hopkins, 1962), population recovery rates appear to be more rapid at lower latitudes (compare Owen, 1953; Hofstetter, 1983; Stanley and Sell- ers, 1986; Mackenzie, 1989). 3 Major population crashes resulting in long-term loss or decline of the C. virginica fishery have oc- curred almost exclusively along the northeast coast of North America. Moreover, significant population declines occurred earlier in the century at higher latitudes (viz. Canada, 1910s, Mid-Atlan- tic area, 1950s; Delaware and Chesapeake Bays, 1980s) (Stanley and Sellers, 1986; Mountford and Reynolds, 1988; Mackenzie, 1989; and others refer- enced previously), although more than one signifi- cant population has declined in some areas. These trends in oyster population dynamics gleaned from the literature are not well documented. Much literature is anecdotal and significant excep- tions do exist. Nevertheless, taken as a whole, these trends suggest two hypotheses: 1) a latitudinal gra- dient in susceptibility to population crashes exists in oyster populations; and 2) as temperature varies both latitudinally and seasonally, temperature, through its effect on oyster physiology (e.g. Koehn and Bayne, 1989), may determine the susceptibility of oyster populations to potentially destabilizing episodes of mortality. In this study, we tested these hypotheses using a population dynamics model. The results of the mod- eling exercise were then used to examine some ba- sic decisions required for fishery management; viz. the timing and length of the fishing seasons and the size limits set for the fishery to obtain a maximum sustainable yield (e.g. Glude, 1966; Hofstetter and Ray, 1988; Young and Martin, 1989). The model Perspective and basic characteristics The oyster population model shown in Figure 1 is designed to investigate the dynamics of the post- settlement phase of the American eastern oyster's, Crassostrea virginica, life from newly settled juve- nile through adulthood. The model consists of a sys- tem of ten coupled ordinary differential equations, with each equation representing a size class of oys- ter; however, the ten size classes are not evenly di- vided across the length or biomass spectrum (Table 1). Size class 1 includes newly settled juveniles (Dupuy et al., 1977). Size class 10 corresponds to oysters that are larger than those normally found in natural popu- lations. The boundaries between size classes 4 and 5, 5 and 6, and 6 and 7 represent size limits that have Table 1 Biomass an d length dimensions of the oyster Crassostrea virginica size classes used in the model. B omass is converted to size using the re- lationshi p given in White et al (1988). Model Biomass Length Size class (g ash-free dry wt) (mm) 1 1.3x10-' - 0.028 0.3 - 25.0 2 0.028 - 0.10 25.0 - 35.0 3 0.10 - 0.39 35.0 - 50.0 4 0.39 - 0.98 50.0 - 63.5 5 0.98 - 1.94 63.5 - 76.0 6 1.95 - 3.53 76.0 - 88.9 7 3.53 - 5.52 88.9 - 100.0 8 5.52 - 7.95 100.0 - 110.0 9 7.95 - 12.93 110.0 - 125.0 10 12.93 - 25.91 125.0 - 150.0 been used or considered for market-size oysters: 2.5 in; 3.0 in and 3.5 in, respectively. We define adults, individuals capable of spawning, as individuals weighing more than 0.65 g ash-free dry weight, about 50 mm in length (Hayes and Menzel, 1981). Therefore, size classes 1 to 3 are juveniles. All calculations were done in terms of energy in calm2. When necessary, oyster energy is converted to oyster biomass by using a caloric conversion of 6100 cal-g dry wt-1 for oysters (Cummins and Wuycheck, 1971) and biomass to an approximate length by using White et al.'s (1988) biomass-length conversion. To calculate any gain, loss, or transfer of energy (or biomass) between size classes, an ad- ditional conversion was made to express the gain, loss or transfer in terms of a specific rate (day-1) which was then multiplied by the caloric quantity in a size class. Transfers between size classes were scaled by the ratio of the average weight of the cur- rent size class (in g dry wt or cal) to that of the size class from which energy was gained or to which en- ergy was lost. This ensured that the total number of individuals in the model was conserved, in the absence of recruitment and mortality. Because, the size classes in the model are not equivalent in dimension, each specific rate for each transfer between size classes was scaled by the ratio between the two size classes: for transfers up : Wj/(Wj+1-Wj) for transfers down : W /(Wj -WJ_l)> where W is the median biomass (in g dry wt) in size class j. For simplicity, we will not include any of these conversions and scaling factors in the equa- tions given subsequently. Powell et al.: Modeling oyster populations 349 Governing equation The change in oyster standing stock with time in each size class (O ) is the result of changes in net production (NPj), taken to be the sum of the produc- tion of somatic (P ) and reproductive (Pr ) tissue, and the addition of individuals from the previous size class or loss to the next largest size class by growth. Following White et al. (1988), net produc- tivity is assumed to be the difference between as- similation (Aj) and respiration (R ), Oyster density/ flowflsld Salinity Temp«ratur« -e ' Particulate Load Food Supply Rltratlon Rata Ingestion //knlmllatkMiS I Efflclancy ) AeslmllarJon Respiration Spawning / Larval \ I Mortality J Rscrultment Figure 1 Schematic diagram of the energy flow model. NP =P +P J gj rj R, (1) Accordingly, dOj dt Pgj + Prj + (gain from.;' - 1) ■ (loss to 7 + 1) (2) for j = 1,10 recognizing Pr = 0 forj = 1,3. Resorption of either gonadal or somatic tissue results in loss of biomass. When NP<0, oysters lose biomass and transfer into the next lower size class. This is an im- portant difference between our size class model and a size class model based on lin- ear dimensions; shell size does not change, however biomass does during periods of negative scope for growth. This is the basis for the use of condition index as a measure of health in oysters (e.g. Newell, 1985; Wright and Hetzel, 1985). To allow for this, equation 2 must be modified as dO , - = Pgj+ Prj + (gain from J - 1) - ( loss to 7' + 1) + (gain from 7 + 1) -(loss to j — 1) (3) for j - 1,10. The last two terms on the right side of Equation 3 represent the individuals losing biomass and, thus, translating down to the next lower size class. The relationships used to parameterize the processes in Equation 3 are described in the following sections. More details and a discussion of the assumptions and support- ing data for the model were presented by Klinck et al. ( 1992), Powell et al. (1992) and Hofmann et al. (1992). Accordingly, the ba- sic oyster size class model is outlined only briefly. However, calculations of spawning size and recruitment, mortality, and the ef- fect of oyster density on feeding are specific to this study and are described in more detail. Feeding and assimilation Ingestion rate depends upon the filtration rate and the ambient food concentration. We adapted Doering and Oviatt's (1986) equa- 350 Fishery Bulletin 92(2), 1994 tion for filtration rate to oysters (Powell et al., 1992) to obtain filtration rate as a function of biomass and temperature (7): and ^•0.96/^0.95 FR, = —!- 1 2.95 ^.=W0 317100.669> (4) (5) where filtration rate (FR) is in mL filtered ind : min-1 and W- is the ash-free dry weight in g for each size class. Equation 4 contains the temperature- dependency described by Loosanoff (1958). Filtration rate was further modified by salinity as described by Loosanoff (1958). Filtration rate de- creases as salinity drops below 7.5%* and ceases at 3.5%*. In mathematical terms: at S > 7.5%* FR„,=FR, "SJ * "V at 3.5 < S < 7.5%c FR„ =FR,(S- 3.5)/ 4.0 (6) at S < 3.5%o FR„=0. where S is ambient salinity and FR- is the filtration rate obtained from Equation 4. The reduction in feeding efficiency at high particu- late loads was included as a reduction in filtration rate according to Loosanoff and Tommers (1948) r = (4.17xl(T4)(1000418x), (7) where x is the total particulate content (inorganic + organic) in g-L-1 and x is the percent reduction in filtration rate. Solving Equation 7 for the percent reduction in filtration rate gives a modified expression for filtra- tion rate of the form: FRV = FRsj 1 0.01 flog II) r + 3.38 0.04 IS (8) Equation 8, if applied to total particulate content (inorganic + organic), limits ingestion rate to ap- proximately the maximum value found by Epifanio and Ewart (1977). Therefore, an additional term to lower ingestion efficiency at high food concentrations was not used. The effect of oyster density on food availability was parameterized from measurements given in Lund (1957) as [k/f0-l]erd + l' (9) where f is the fractional reduction in food, d is oys- ter density expressed as L filtered hr_1m"2, and f0=0.001, an arbitrarily low number conforming to the expectation that food supply is not affected by low oyster density. For the high flow (59 L hr_1) con- ditions given in Lund (1957), k = 0.31 and r = 1.36X10"6. For low flow (12 L hi--1) conditions, k = 0.57 and r = 9.746xl0~7. Food availability at a given oyster density is estimated as ( \-f) times the ambi- ent food concentration. Filtration rate times the ambient available food concentration then gives oyster ingestion. Assimilation is obtained from in- gestion using an assimilation efficiency of 0.75 (Powell et al., 1992). Respiration Oyster respiration as a function of temperature and oyster weight was obtained from Dame (1972) as fl, =(69.7 +12.67) W, 6-1 (10) where Rf is in uL 02 consumed hr 1-g dry wt 1 and b = 0.75. Salinity effects on oyster respiration were param- eterized from data given in Shumway and Koehn ( 1982) by obtaining a ratio (Rr) of respiration at 10%* to respiration at 20%c, R ' , and regressing this ratio against tem- R R 20%r perature. This yielded two equations: at T < 20°C Rr = 0.007T + 2.099; atr>20°C R= 0.09157+ 1.324; (11) which were then used to obtain respiration rate as follows: S > 157« Rtj =Rj\ 10%c5 H 3 T-2 GB M/A-A/S io-7 75% Y >5 H 4 T-2,3,4 GB M/A-A/S 10 ~7 90% Y >5 H 5 T-2 GB M/A-A/S io-7 99% Y >5 H 6 F-5,T-2,3.4 GB M/A-A/S io-7 99.9% Y >5 H 7 T-3 GB M/A-A/S io-7 90% Y :3 H 8 F-6 GB M/A-A/S io-7 99% Y >3 H 9 F-7.T-3 GB M/A-A/S io-7 99.9% Y M H 10 F-8 GB M/A-A/S io-7 99%. Y >1 H 11 F-10e,T-3,4 GB A/S io-7 we; Y >5 H 12 F-10f,T-3,4 GB A/S io-7 99.9% Y >5 H 1.3 T-3, 4 GB A/S io-7 90% Y :3 H 14 F-9.T-3 GB A/S io-7 99.9% Y >3 H 15 T-4 GB A/S io-H 90% Y >3 H 16 F-lOg GB A/S io-H 75% Y >3 H 17 F-lOh GB A/S io-8 50% Y >3 H 18 F-lOa GB A/S io-8 75% Y >5 H 19 F-10b,T-4 GB A/S io-8 90% Y ■f, H 20 F-lOc GB A/S IO"8 99% Y >5 H 21 F-10d,T-4 GB A/S io-8 99.9% Y >5 H 22 T-4 GB M/A-A/S io-8 99.9% Y >5 H 23 T-4 GB M/A-A/S io-8 90% Y •r. H 24 F-ll,13,T-5 ( ; B M/A-A/S io-K 99.9% w :5 H 25 T-5,6 GB M/A-A/S io-8 99% W •r. H 26 T-5 GB M/A-A/S IO8 90% W ■r< H 27 T-5 (ill M/A-A/S io-8 90% s r> II 28 T-5,6 GB M/A-A/S io-8 99% s ■5 H 29 F-12,13,T-5 GB M/A-A/S io-8 99.9% s :5 H 30 T-5,6 GB M/A-A/S io-8 90% s :3 H 31 F- 14, T-5 GB M/A-A/S io-8 99% s :3 H 354 Fishery Bulletin 92(2). 1994 Table 2 (Continued) Table Food Fraction Yearly Season Size class Beginning or time of spawn mortality of suffering density Case figure Bay series recruited rate mortality mortality (day 1) 32 F-14.T-5 (II! M/A-A/S io-« 99% W >3 H 33 T-5,6 GB M/A-A/S io-« 90% W 3 H 34 F-17,18 GB M/A-A/S io-8 99% S >5 M 35 F-17,18 GB M/A-A/S io-K 99% w •5 M 36 F-17,18 GB M/A-A/S io-K 99<7f s ■li M 37 F-17,18 GB M/A-A/S io-K 99% w r, M 38 F-17,18 GB M/A-A/S IO"8 99% s >7 M 39 F-17,18 GB M/A-A/S IO"8 99% W ■7 M 40 T-5 GB A/M-S/O IO"8 99.9% w -r. H 41 F-15,16,T-5 GB A/M-S/O IO"8 99.9% s ■r, H 42 T-5 GB A/M-S/O IO"8 90% w ■■■'. H 43 T-5,6 GB A/M-S/O IO"8 90% s •3 H 44 F-19,20 CB A/M-S/O IO"8 99% s >5 H 45 T-6 CB A/M-S/O io-8 90% s 3 H 46 T-6 CB M/A-A/S io-8 90% s >3 H 47 T-6 CB M/A-A/S io-8 99% s >5 H 48 F-21.T-6 CB M/A-A/S io-8 99%. w :5 H 49 T-6 CB M/A-A/S io-H 90% w >3 H 50 T-6 CB A/M-S/O io-H 90% w >3 H 51 F-20 CB A/M-S/O io-« 99% w :5 H Model solution The model described by Equation 3 was solved nu- merically by using an implicit (Crank-Nicolson) tridiagonal solution technique. The time step for model integration was one day. Simulations were run for six years which is sufficient time for the model solutions to adjust so that trends in popula- tion levels could be identified in the simulations. Results Model initialization The system of equations given by Equation 3 re- quires that an initial oyster population size-fre- quency distribution be specified. The simulations described in the following sections are designed to investigate seasonal and latitudinal mortality effects on oyster population size frequency and stability. Therefore, it proved useful to begin the simulations with a size-frequency distribution representative of a crowded population; that is, one suffering little mortality. In this way, changes in the simulated oyster populations will be the result of mortality only. Also, using the same initial population distri- bution allows for comparison between simulations throughout the entire 6-year simulated time period. The initial oyster size-frequency distribution was obtained from a simulation that was started with 10 individualsm-2 in size-class 7 on 1 January. The food time series for this simulation contained two phytoplankton blooms of two months duration (March/April, August/September) with intervening summer months and winter months as detailed in Figure 2. Dense bivalve populations can deplete the surrounding water column of food (Frechette et al., 1991 ). We used Lund's ( 1957) low flow conditions to simulate the effect of oyster density on food supply. Such conditions might be typical of an enclosed or sheltered reef (Powell et al., 1987). No mortality was allowed in any size class. The time development of the simulated population (Fig. 3A) shows that the mean size of the popula- tion slowly declines from size class 7 to size class 3, as population density increases about 3 orders of magnitude. These trends are characteristic of a crowded population: high population density and reduced adult size. Reproduction continues through- out the simulation (Fig. 3, A and C) with a strong fall spawning pulse (Fig. 3B) occurring in response to the fall phytoplankton bloom (Hofmann et al., 1992). Therefore, food limitation is not sufficient to cap population growth; however, the rate of popu- lation increase has dramatically declined over the 6-year simulation. It is the population size- frequency distribution at the end of the 6-year simu- lation (Fig. 3D) that is used to initialize the mortality simulations described in the following sections. Powell et al.: Modeling oyster populations 355 Effect of continuous mortality The first set of simulations considered the oyster population that would be produced in Galveston Bay, Texas, when continuous mortality (mortality throughout the year) is imposed on size classes 5 and larger. Oyster size class 5 approximates the 2.5 in size limit often desired by the oyster fishery as opposed to the standard size limit of 3 now enforced in most areas. Over this series of simulations, the rate of yearly mortality was varied from 50% to 99.9%, the two extremes being depicted in Figures 4 and 5. For an oyster population with no recruit- ment, these rates would result in a reduction of the GALVESTON BAY i — i — i — i — i — <- 300 600 900 1200 1500 1800 2100 TIME (Days) Number of Individuals SpaWn (kcal) Number of Adults l| J L^_l J_A. 12 18 24 30 36 42 48 54 60 66 72 Julian Month 10000 8000 " 6000 S 4000 2000 12 3 4 5 6 Julian Year 8 910 population by 0.5 and 0.999, respectively, in one year. In our simulations, where recruitment and mortality constantly change population abundance, a 50% mortality rate does not necessarily result in the loss of one-half of the individuals in the popula- tion in one year. Over this series of simulations (Table 3, Figs. 4 and 5), as mortality rate increases from 50% to 99.9%, density declines by about 80% and the size- frequency distribution shifts slightly to lower size classes. Population reproductive effort declines as the number of adults declines, but individual repro- ductive effort increases. At the lower mortality rate, spawning is primarily confined to a single strong pulse in the fall. At the higher mortality rate, spawning effort is distributed between a spring and fall spawning peak; the fall peak is stronger and extends over a longer time (Fig. 4A vs. Fig. 5A). Moreover, spawning is higher in every other year (Figs. 4C and 5C). In the tem- perature time series for Galveston Bay (Fig. Figure 3 Simulated time development and population dis- tribution of a Galveston Bay Crassostrea vir- ginica population with no mortality, allowing the population to approach the carrying capacity of the environment. (A) The number of individuals per size class and reproductive effort per size class. Values are plotted opposite the size class designation, not halfway between; hence all in- dividuals in size class 7 are opposite the grid mark labeled 7 on day 1 of this simulation. Iso- lines, for number of individuals, are the loga- rithms of the number of oysters (log10N). Hence, the zero contour corresponds to one individual. Population concentrations less than this are in- dicated by dashed lines; population concentra- tions greater than this by solid lines. Shading for the amount of reproductive effort (spawn) repre- sents the logarithm of cal (logical) with the darkest shades corresponding to highest values. Contour interval is 0.5 for the number of indi- viduals and 1.0 for reproductive effort. (B) Monthly-averaged values of the number of indi- viduals, the number of adults (j=4, 10), and the monthly reproductive effort in kcal for the 6-year simulation. Values can be converted into joules by multiplying by 4.16y'cal_1; into biomass by us- ing 6100 cal-g dry wt_1; and into the equivalent number of fully developed eggs by 13 ngegg-1 X6.133X10"6 cal-ng-1. (C) The yearly reproductive effort (number of kcal spawned). (D) The final size class distribution in the population at day 2,160. Additional data and explanation in Table 2, case 1. 5000 4000 3000 | 2000 £ 1000 356 Fishery Bulletin 92(2). 1994 .o E z 100000 10000 1000 100 10 50000 40000 " 30000 S 20000 10000 2A), one winter is colder and one summer warmer than the other. As a result, the first year in each pair is characterized by lower reproductive effort as decreased temperatures reduce filtration and inges- tion rate and switch net production to- wards somatic growth. Warmer tempera- tures the second year result in a larger reproductive effort. Within these simulated oyster popula- tions, a complex interaction exists be- tween population density, size frequency, and mortality rate. Increasing mortality removes individuals, thereby increasing the available food supply for the remain- ing individuals. Increased food supply results in increased spawning effort, which then increases population density. This in turn then gives reduced spawning effort. This feedback results in potential population equilibria of different densities and size frequencies for each level of mor- tality (Table 3). Even at 99.9% yearly mortality, however, the population sus- tains itself at a fairly dense level. Of more significance, each population approaches an equilibrium or nearly so, such that recruitment balances mortality over this range of mortality rates. Year-to-year shifts in population size over the 6-year simulation show neither continually strong declines nor increases in popula- tion density for any of the mortality rates. In Figures 6—8, we compare the time-development of oyster populations exposed to similar overall mortality levels, but in which mortality extends into lower size classes than in Figures 4 and 5. In these simulations, mortality was imposed either on all adult sizes and the larger juveniles (Figs. 6 and 7) or on all size classes (Fig. 8). Figures 7 and 5 differ only in the size classes exposed to mortality (5 and larger vs. 3 and larger) as do Figures 6 and 8 (3 and larger vs. 1 and larger). As high (90-99.9%) yearly mortality rates are imposed on smaller oyster size classes (Figs. 6-8), the population becomes more susceptible to significant population declines. For example, a 99.9% yearly mortality rate had little effect when mortality was restricted to size classes 5 and larger (Fig. 5), but results in a population crash if size classes 3 and larger are similarly ex- posed (Fig. 7). Many more individuals die before reproducing in the latter case than in the former. A mortality rate of 99.9% is required for a population crash at size classes 3 and larger (Fig. 7), but only 99' ! at size class 1 and larger (Fig. 8). As mortality Number of Individual* Mortality (no/month) 1 A / — ' <- w » 1 : [•■ 1 ' 1 _^. _L_-^_l_- A ^/^_ l| l\ .k 40000 30000 20000 3 10000 12 18 24 30 36 42 Julian Month 48 54 60 66 72 2 3 4 5 Julian Year 23456789 10 Size Class Figure 4 Simulated time development and population distribution of a Galveston Bay Crassostrea virginica population exposed to a con- tinuous mortality rate of 50% per year on size classes 5 and larger. (A) Monthly averaged values of the number of individuals, the number of adults (j=4, 10), and the monthly reproductive effort in kcal for the 6-year simulation. (B) The yearly reproductive ef- fort (number of kcal spawned). (C) The final size class distribu- tion in the population at day 2,160. Further explanation in Fig- ure 3 and Table 2, case 2. Table 3 A comparison o final densi ty in simulated Crassostrea virgi nica populati ons after 6 years and total reproductive effort in year 6 at various rates of yearly mortality. Additional details in Table 2. Ending Total density reproductive Mortality (day 2160) effort in year 6 Case rate (%) (indm '■*) (kcalm*2) 2 50 50,748 43,158 3 75 16,966 15,093 4 90 33,112 46,426 5 99 12,295 14,211 6 99.9 16.565 19,896 extends into the smaller size classes, the mortality rate that the population can sustain decreases. We note that, although these mortality rates seem high, they are well within the typical range for juvenile survivorship in bivalve communities (e.g. Powell et al., 1984; Cummins et al., 1986). Powell et al.: Modeling oyster populations 357 Furthermore, as mortality extends into lower size classes, the size-frequency distribution shifts to larger sizes (Figs. 6C, 7C, 8C). The effect is signifi- es 100000 Number of Individuals Mortality (no/month) • I. i. JV. A/*Ji 6 12 18 24 h /\ n / i A . 30 36 42 Julian Month H |l i i Jj_i. -H fi i i A 8000 cant because only in cases where mortality is high do oysters grow large enough to reach marketable size for the oyster fishery (size class 6 and larger). Removal of smaller individuals increases the available food supply for the survivors, thereby allowing some to attain market-size. 6000 c 3 4000 ™ 00 2000 48 54 60 66 72 20000 10000- 2 3 4 5 Julian Year 23456789 10 Size Class Figure 5 Simulated time development and population distribution of a Galveston Bay Crassostrea virginica population exposed to a continuous mortality rate of 99.9% per year on size classes 5 and larger. (A) Monthly-averaged values of the number of in- dividuals, the number of adults (j=4, 10), and the monthly re- productive effort in kcal for the 6-year simulation. (B) The yearly reproductive effort (number of kcal spawned). (C) The final size class distribution in the population at day 2,160. Further explanation in Figure 2 and Table 2, case 6. Number ol Individuals Spawn (kcal) 0) 1000(1 Mortality (no/month| - Number ol AdulU ,„„„„ 3 ^V^\^x\ ^-\ _/\ ^-\A^x^ ^ 1000 > f 100 ■ X" "'» ~~?r y- 20000 c S - Nn i i ■ a o 10 - ' i" i \w\ V nJ\ Ar' »: 10000 w Numbe o ) 6 12 18 24 30 36 42 48 54 60 66 7 2U Julian Month B — a ic- 8000 _ ■1 30 - ro IB -o £ 6000 £ ' ■!■ c ■IhI 20 1 5 4000 ■ ■ (■■ , i a ra 2000 r° II ■■■■■■ ■IIII-- 0 1 2 3 4 5 6 1 234 5678910 Julian Year Size Class Effect of food supply Interactions between food supply and mor- tality rate are potentially important in de- termining population density and size-fre- quency distribution. In years in which a spring bloom is reduced or fails to occur (Fig. 2B), the available food spectrum is shifted in time and total food supply for the year is reduced. In Figure 9, we examine the effect of the failure of the spring bloom. Figure 9 can be compared directly with Figure 7, the two differing only in food supply. A failed spring bloom shifts the food spectrum as well as decreasing the total food available over the year. Hofmann et al. (1992) showed that the food supply time series used for Figure 9 results in a strong fall spawning pulse. With an imposed yearly mortality of 99.9% in size classes 3 and larger and no spring bloom, the simulated oyster populations (Fig. 9) are not substantially different from those shown in Figures 6—8. However the simulated oyster population shown in Figure 9 is characterized by a stronger fall spawning pulse, as expected, whereas the previous simulations generally had spawning more evenly distributed over the spawning season. The population still reaches a stable distribution and the size-fre- quency distribution includes individuals in the larger size classes (Fig. 9C). Thus, continuous yearly mortality overrides the effects of varia- tions in the timing of food supply. Figure 6 Simulated time development and population dis- tribution of a Galveston Bay Crassostrea virginica population exposed to a continuous mortality rate of 99% per year restricted to size classes 3 and larger. (A) Monthly-averaged val- ues of the number of individuals, the number of adults (7=4, 10), and the monthly reproductive effort in kcal for the 6-year simulation. (B) The yearly reproductive effort (number of kcal spawned). (C) The final size class distribution in the population at day 2,160. Further information in Figure 3 and Table 2, case 8. 358 Fishery Bulletin 92(2), 1994 However, in this and other simulations, the population density is consistently higher after six years with the lower, more re- stricted food supply associated with the missing spring bloom (e.g. Fig. 7 vs. 9; Table 4). The effect occurs regardless of the size- class distribution of mortality or the mortal- ity rate. The initial surmise that more food should result in higher densities is not con- firmed. Reproductive effort is higher at the higher food supply only in the first year (Fig. 7 vs. 9) and declines more rapidly thereaf- ter as population density declines. Initially this would appear to be counterintuitive; more food should result in higher population densities and greater reproductive effort. However, increased food in the spring in- creases growth rate so that more oysters grow more rapidly into size classes suffering mortality. As a result, the number of adults and population reproductive potential de- clines. This results in a lower population den- sity. The model simulations indicate that oys- ter population abundance is the result of a complicated interplay between the timing of food supply, reproductive effort, and mortality. Lowered recruitment success 30 36 42 Julian Month 10000 8000 B | 6000 ■ I 4000H !■ 2000 ■ „U_ — I 12 3 4 5 6 Julian Year 2345678910 Size Class An additional source of mortality for oyster populations is through decreased survivor- ship of the planktonic larvae (Table 5, Fig. 10). Lower larval survivorship results in decreased recruitment success and lower population densities, as expected. However, loss of the spring bloom en- hances oyster population density as before (Table 5). Nevertheless, a reduction in recruitment success, when combined with mortality on the post-settle- ment population, results in populations that are less resistant to population crashes. For example, a ten- fold reduction in recruitment success in a popula- tion exposed to a 75% mortality rate in size classes 3 and larger produces the effect observed for a mor- tality rate of 99.9% with an order of magnitude higher recruitment success. One additional important concept arises from this series of simulations. Simulations that included high recruitment success and various mortality rates pro- duced final size-frequency distributions similar to those shown in Fig. 10, E and F. Few individuals are found in size classes 5 and larger. The legal size for the oyster fishery is typically size classes 6 and larger. No fishery could exist under these conditions. High population density produces stunted individu- als. A reduction in recruitment success over a range Figure 7 Simulated time development and population distribution of a Galveston Bay Crassostrea virginica population exposed to a continuous mortality rate of 99.9% restricted to size classes 3 and larger. (A) Monthly-averaged values of the number of in- dividuals, the number of adults (j=4, 10), and the monthly re- productive effort in kcal for the 6-year simulation. (B) The yearly reproductive effort (number of kcal spawned). (C) The final size class distribution in the population at day 2,160. Further information in Figure 3 and Table 2, case 9. of mortality rates (Fig. 10, A-D) gives size-frequency distributions shifted towards the larger size classes. In fact, more market-sized animals exist in these populations than in the ones shown in Figure 10, E and F. Shifting mortality to lower size classes results in even more market-size individuals (Fig. 10, G-H). A successful fishery requires some degree of mortality, including juvenile mortality. Effect of seasonal mortality The commercial oyster fishery is typically confined to a winter season. In some cases, a restricted sum- mer season is also allowed. Agents of natural mor- tality, like Perkinsus marinus and Thais haema- stoma, typically extract a greater toll during the summer. The effect of mortality restricted to the summer and to the winter is illustrated in Figures 11 and 12, respectively, and in Table 6. For this se- ries of simulations, we define winter as the months of October through March and summer as the months of April through September. Thus each simulated oyster population has the same number Powell et al.: Modeling oyster populations 359 of days (180) with and without mortality. Regard- less of the mortality rate, when mortality is re- stricted to size classes 5 and larger, populations Number of Individuals — ■ Mortality (no/month) - - Spawn (kcal) Number of Adults 30 36 42 Julian Month 3000 2000 1000 "> 10000 8000 B IlJ 12 3 4 5 Julian Year 2345678910 Size Class Figure 8 Simulated time development and population distribution of a Galveston Bay Crassostrea virginica population exposed to a continuous mortality rate of 99% imposed on all size classes. (A) Monthly averaged values of the number of individuals, the number of adults (j=4, 10), and the monthly reproductive ef- fort in kcal for the 6-year simulation. (B) The yearly reproduc- tive effort (number of kcal spawned). (C) The final size class distribution in the population at day 2,160. Further informa- tion in Figure 3 and Table 2, case 10. 30 36 42 Julian Month 3000 suffer a greater reduction in density when mortal- ity is restricted to the summer (compare Fig. 11B and 12B; Table 6). Summer mortality depresses re- productive effort and depressed reproductive effort, continued over time, results in lower population density Examining the population size-frequency distribution over the year for simulated oys- ter populations suffering winter (Fig. 13, A and B) and summer (Fig. 13, C and D) mor- tality suggests an explanation for the more detrimental effect of summer mortality on population density. Figure 13 shows snap- shots of the population's size-frequency dis- tribution at various times during the year. When mortality is imposed only during the winter, the population size-frequency distri- bution shifts to larger size classes in the summer in response to increased growth rate produced by warmer temperatures. Therefore, during the fall spawning season the population is dominated by the larger size classes that account for much of the reproductive effort. Winter mortality then shifts the population size-frequency distribu- tion back to smaller individuals (Fig. 13B) and the cycle begins again. Hence, winter mortality allows the population to replace, during the next summer and fall, the indi- viduals that are lost. In contrast, restricting mortality to sum- mer months produces a population size-fre- quency distribution that varies little over a year (Fig. 13, C and D). The variation that does occur is a shift towards smaller indi- viduals in the summer. For example, more individuals are found in size classes 6 and 7 in September in populations that experi- - 2000 - 1000 10000 8000 6000 5 4000 2000 B ll.... 12 3 4 5 6 Julian Year 2345678910 Size Class Figure 9 Simulated time development and population dis- tribution of a Galveston Bay Crassostrea vir- ginica population exposed to a continuous mor- tality rate of 99.9% restricted to size classes 3 and larger and in which the food time series con- tained only the fall bloom. The case is compa- rable to Figure 7 in which two blooms occurred. (A) Monthly averaged values of the number of individuals, the number of adults (j=4, 10), and the monthly reproductive effort in kcal for the 6-year simulation. (B) The yearly reproductive ef- fort (number of kcal spawned). (C) The final size class distribution in the population at day 2,160. Further information in Figure 3 and Table 2, case 14. 360 Fishery Bulletin 92(2). 1994 ence winter (Fig. 13B) rather than summer (Fig. 13D) mortality. The shift to smaller individuals in populations with summer mortality results in low- ered reproductive effort. Hence, lost in- dividuals are not replaced in the fall and winter and the population declines. As mortality extends into the juve- nile size classes, the difference in win- ter and summer mortality should de- crease and the seasonal shift in size- frequency as a function of mortality should disappear because a greater fraction of the total mortality occurs in individuals contributing relatively little to the population's spawning potential. This is confirmed by the model (Fig. 14, Table 6). Interestingly, although the seasonal variations in size-fre- quency distributions are muted, changes in size-frequency distri- bution over the year are still greater for populations that expe- rience winter mortality. These populations show a slight shift to smaller size classes in the winter. To examine the effect of varying food supply, we placed the spring and fall blooms one month later in the year (April/May and August/ September) and then compared the time development of oyster populations suffering winter or summer mortality with those pre- viously described when the blooms occurred one month ear- lier (Tables 6 and 7). For popula- tions experiencing winter mortal- ity, delaying the spring and fall blooms by one month (Fig. 2C) does not significantly change the simulated populations from those obtained for the earlier blooms, even when mortality extends to the juvenile size classes (3 and larger). However, for summer mortality, delaying the blooms by one month dramatically improves the population's abil- ity to sustain itself (Fig. 12 vs. Fig. 15; Table 5). Moving the spring and fall blooms one month later in the year produces 1) a strong spring spawning pulse as well as the fall pulse and 2) a shift in the population size-frequency distribution toward the larger size classes, although yearly changes in the size-frequency distribution are still characteristic of summer mortality (Fig. 16 vs. 13). As a result, Table 4 A comparison of final density in simulated Crassostrea virginica populat ons after six years and total reproductive effort in year 6 with and without a spring phytoplankton bloom. Additional details in Table 2. Spring Ending density Total reproductive effort Case bloom? (day 2,160) (ind-m" 2) in year 6 (kcalm-2) 9 Yes 2 12 14 No 248 1,003 7 Yes 2,602 7,787 13 No 2,788 7,398 6 Yes 16,565 19,896 12 No 32,513 36,569 4 Yes 33,112 46,426 11 No 42,758 39,217 Table 5 A comparison of final density in simulated Crassostrea virginica popu- lations after 6 years and total reproducti ve effort in year 6 at various rates of recruitment, with and without a spring phytoplankton bloom. Additional details in Table 2. Total K nding density reproductive Mortality (day 2,160) effort in year Case rate Recruitment (ind-m-2) 6 (kcalm-2) No spring bloom 13 15 90% io7 io-8 2,788 2 7,398 26 11 19 90% io7 IO"8 42,758 1,067 39,217 12,480 12 21 99.9% io-7 io-8 Spring bloom 32,513 11 36,569 236 6 22 99.9% IO"7 IO"8 16,565 1 19,896 40 4 23 9C7f io-7 io-8 33,112 328 46,426 4,935 spawning effort increases under the delayed-bloom condition as fall spawning extends beyond the sum- mer season of mortality. Accordingly, variation in the timing of food supply, under certain circumstances, can be important in the success of an oyster popu- lation, particularly in cases where adult mortality is restricted to the summer months. Size limits for the fishery Three size limits have been used or considered as the legal limit for market-size oysters: 2.5 in, 3.0 in, Powell et al.: Modeling oyster populations 361 ■5 & 3 § o 8. 3 B • to SO- G D TJ > Wh ■o c o 30- & ■ a 20- hbI c _■ s 10- a. Oi II jilL 2468 10 2468 10 Size Class Size Class Figure 1 0 A comparison of the final size-frequency distributions (day 2,160) in simulated Crassostrea virginica populations exposed to a Galveston Bay temperature time series fol- lowing 6 years of recruitment, growth, and mortality un- der varying degrees of recruitment and mortality. In each case, size classes 5 and larger were exposed to continu- ous mortality at a yearly rate of (H) 50%, (A and G) 75%, (B) 90%, (C and E) 99%, (D and F) 99.9%. Recruitment was tenfold higher (or larval mortality tenfold less severe) in E and F. Mortality rates extend down into size classes 3 and 4 in G and H. Further information in Figure 3 and Table 2, cases 11, 12, 16-21. and 3.5 in. These correspond to size classes 5, 6, and 7 in the model. The simulations used to test the effect of these size limits were initialized with a population size-frequency distribution having a component in the larger size classes (Fig. 12). With a yearly mortality rate of 99%, oyster popu- lations increase when mortality is restricted to size class 7 and larger (3.5 in) but decline rapidly if mortality includes size classes 5 and 6 (2.5 in) (Fig. 17). Hence, a change in the legal size limit may have a substantial effect on the fishery and on the oyster population as a whole. Of course, the specific results would vary according to the biomass-to-length conversion used. As the fishing season typically is confined to the winter, we examined the effect of changing size limits when mortality was restricted to the win- ter or to the summer months (Fig. 17). Overall, the same pattern persisted in both seasons. Popu- lations declined more under the smaller size lim- its. However, several significant differences are also observed: 1 Populations in which mortality was restricted to the summer had a stronger spring spawn- ing pulse; most spawning occurred in the mid- summer and early fall in populations suffering only winter mortality. 2 Reproductive effort and population density was consistently higher in populations suffering winter mortality (Fig. 18, C, D, and E), density by a factor of 2 to 4, reproduction by a factor of 2 to 8; increased reproductive effort occurred both because the number of adults increased and because those adults spawned more with the result that reproduction was more than proportionately higher. 3 The size-frequency distribution was shifted toward the smaller size classes in populations having winter mortality (Fig. 18, A and B) but had little impact on the size-frequency distri- bution with summer mortality. Overall, the number of market-size oysters available at the end of the simulation was higher at the larger size limits (Fig. 18F). As a result, a greater potential yield was available to the fish- ery at the larger size limits. One reason for the higher yield available to the fishery at the larger size limit (>3.5 in) is the shift in size-frequency distribution toward larger size classes with adult mortality. A second reason is the protection of a larger portion of the reproductive population. However, if unchecked, the continually growing population in the last set of simulations, where 362 Fishery Bulletin 92(2). 1994 Table 6 A comparison of final density in simulated Crassostrea virginica populations after six years and total repro- ductive effort in year 6 with mortal ty restricted to the winter or the summer season and with the spring and fall phytoplankton blooms early in the year or one month later. Additional detail s in Table 2. Ending densi ty Total reproductive (day 2,160) effort in year 6 Case Season (ind-nr2) Ikcal-m-2 Mortality: >5 Bloom: Early Late Bloom : Early Late 24, 40 Winter 287 220 5,168 4,344 29, 41 Summer 36 500 606 6,904 25 Winter 467 7,365 28 Summer 253 3,586 26 Winter 1,400 4,529 27 Summer Mortality: >3 1,333 13,107 33, 42 Winter 403 595 5,890 8,664 30, 43 Summer 623 692 7,766 8,829 32 Winter 5 111 31 Summer 5 91 Table 7 A comparison of final density in simulated Crassostrea virginica populations after six years and total repro- ductive effort in year 6 with mortality restricted to the winter or the summer season and with the spring and fall phytoplankton blooms early in the year or one month later, in Chesapeake Bay and Galveston Bay. Additional details in Table 2. Case Bay Ending densi (day 2,160) (indm-2) ty Total reproductive effort in year 6 (kcal-m-2) 48 25 Mortality: >5 Winter Chesapeake Galveston Bloom: Early 394 467 Late li oom Early 3,924 7,365 Late 47 28 Summer Chesapeake Galveston Mortality: >3 102 253 644 3,586 46, 45 30. 43 Summer Chesapeake Galveston 72 623 85 692 328 7,766 542 8,829 49, 50 33, 42 Winter Chesapeake Galveston 122 403 M 595 832 5,890 606 8,664 mortality was restricted to size class 7 and larger (>3.5 in), would eventually negate both effects as population density increased. Effect of latitude on population stability In Figure 19 and Table 7, we compare the time-de- velopment of oyster populations under the tempera- ture conditions of Chesapeake Bay with those un- der the temperature conditions of Galveston Bay (Fig. 15). In comparison with the Galveston Bay populations, those in Chesapeake Bay are charac- terized by densities 2 to 5 times lower, reproductive efforts as much as a factor of 10 lower, size-fre- quency distributions considerably shifted toward the large size classes (Figs. 20 and 21), and discrete spo- Powell et al.: Modeling oyster populations 363 GALVESTON BAY ■£ 100000 TO .-f 10000 1000 100 300 600 900 1200 1500 1800 2100 TIME (Days) Number of Individuals Mortality (no/month) 8000 24 30 36 42 Julian Month 15000 « 10000 5000 2 3 4 5 6 Julian Year 1 2345678910 Size Class Figure 1 1 Simulated time development and population distribution of a Galveston Bay Crassostrea virginica population exposed to a continuous mortality rate of 99.9% restricted to size classes 5 and larger and in which mortality occurred only during the win- ter. Compare Figure 12. (A) The number of individuals per size class and reproductive effort per size class. Isolines, for number of individuals, are the logarithms of the number of oysters (log10N). Shading for the amount of reproductive effort (spawn) represents the logarithm of cal (logical). (B) Monthly averaged values of the number of individuals, the number of adults (j=4, 10), and the monthly reproductive effort in kcal for the 6-year simulation. (C) The yearly reproductive effort (number of kcal spawned). (D) The final size class distribution in the population at day 2,160. Further information in Figure 3 and Table 2, case 24. radic spawning pulses typically strongest in midsummer. Like Galveston Bay popu- lations, a shift in the timing of the spring and fall blooms has little effect on the sea- sonal changes in size-frequency distribu- tion (Fig. 21) but considerable effect on the resulting population density in some cases. Populations experiencing winter mortality are more affected by variations in the tim- ing of the food supply than populations ex- periencing summer mortality. Unlike Galveston Bay populations, populations experiencing summer mortality have lower population densities than populations ex- periencing winter mortality only when the blooms occur in March/ April and August/ September. Delaying the blooms by one month results in little variation between populations experiencing summer and win- ter mortality. The most significant factor producing differences between the Galveston Bay and Chesapeake Bay popu- lations is the cooler temperatures that characterize Chesapeake Bay. This results in reduced reproductive effort with more net production going to support somatic tissue growth (Table 7). Discussion The importance of mortality Unlike an oyster population, an oyster fish- ery cannot persist without large adult in- dividuals. One of the consistent messages of this modeling exercise is the require- ment of mortality for the population to produce larger, market-size individuals. Either adult or juvenile mortality will suf- fice, as both juveniles and adults compete for food (Powell et al., 1987). Low rates of mortality result in crowding, food limita- tion, and a stunted population. As mortal- ity extends into the juvenile size classes, and finally into the larval stages (modeled as a reduction in recruitment, reduced re- productive effort, or produced by the colder temperatures of Chesapeake Bay) the population on the average becomes skewed more and more towards the larger adult size classes. Frequently, this proportional shift was sufficient to result in an increase in adult density despite an overall lower population density. An even higher rate of mortality reversed this trend; the popula- 364 Fishery Bulletin 92(2). 1994 tion size-frequency shifted again towards smaller size classes as adult individuals were rapidly removed from the population. Clearly, for a successful fishery, a delicate balance exists between sufficient mortality to permit the fishery to exist and too much mortality which will reduce the harvest- able yield. Food supply is a complicating factor. In- creased food supply will not always result in increased population density or in- creased harvestable yield. The timing of the food supply interacts in subtle ways with the timing and intensity of mortality, sometimes producing higher densities and sometimes lower ones. The simulations show that the effect of variations in food supply is complex; no simple rules apply and a number of feedback mechanisms exist. In one case, for example, lower popu- lation density resulted from increased food supply because increased growth permitted more oysters to enter the size classes that were exposed to mortality, thereby result- ing in a population that declined. In an- other case, a one-month change in the tim- ing of the spring and fall blooms changed population density by a factor of 2 at the same mortality rate. In other cases, little impact occurred in the population despite, for example, the complete failure of the spring bloom. Population stability and population crashes The stability of oyster populations is sen- sitive to several factors, including the tim- ing and intensity of mortality, latitude, and food supply. (We use the term stable in the sense of Underwood [19891 for populations able to recover quickly from perturbation. The terms elasticity and resiliency might also be used. ) Increased mortality reduced population density in every comparison. Oftentimes, a relatively stable equilibrium occurred as recruitment balanced mortal- ity over the long term. In all cases, how- ever, mortality rates sufficient to destabi- lize this equilibrium could be found and a population decline resulted. When mortal- ity extended over a wider range of size classes or affected larval survivorship, population destabilization occurred more easily. In the former case, more oysters GALVESTON BAY -i — i — i — r 300 600 900 1200 1500 1800 TIME (Days) 2100 100000 10000 1000 100 e z B Number ol Individual* Mortality (no/month) Spawn (kcal) Numtxx ol Adults 10 10000 8000 - S 6000 - S 4000 H ll II I I I ■ .'■' ■ .K) ■■ , J ,i. ( i A i I\/^j-*K 4000 3000 - 1000 18 24 30 36 42 48 54 60 66 72 Julian Month c/) 2000 CO 3 D 1 30 ■5 3 20 1 0) en CO c 10 0) o 0) .1 L 3 4 Julian Year 1 2345678910 Size Class Figure 1 2 Simulated time development and population distribution of a Galveston Bay Crassostrea virginica population exposed to a continuous mortality rate of 99.9% restricted to size classes 5 and larger and in which mortality occurred only during the sum- mer. Compare Figure 11. (A) The number of individuals per size class and reproductive effort per size class. Isolines, for number of individuals, are the logarithms of the number of oysters (log1()N). Shading for the amount of reproductive effort (spawn) represents the logarithm of cal (logical). (B) Monthly averaged values of the number of individuals, the number of adults (/=4, 10), and the monthly reproductive effort in kcal for the 6-year simulation. (C) The yearly reproductive effort (number of kcal spawned). (D) The final size class distribution in the population at day 2,160. Further information in Figure 2 and Table 2, case 29. Powell et al.: Modeling oyster populations 365 w \ ■ March B ■ September 3 40 - ■ June B December 2 El August O February 1 30- o i m Percentage o o O 1 If 1 ; i 1 1 1 1 ru Uull. 111 50 - ( ■ March 1 ) ■ September 5 40 ■ S June H December 2 □ August 1 □ February 1 30- o ll 9 a 20 - - i fli 77 Jl S 10- a. ■ ■J II IL- ■ Lull lk_ 1 23456789 10 12 3 4 S 6 7 8 9 10 Size Class Size Class Figure 1 3 A comparison of the changes in size-frequency distribution through the year in the two Crassostrea virginica populations depicted in Figures 11 and 12. (A and B) a population suffering winter mor- tality; (C and D) a population suffering summer mortality. Mortality in both cases was restricted to size classes 5 and larger. •S 30 - o 20 10 B hill ll I ■ I : 11 1 September I December ] February Ik ■d 30 " o 20 ■ March B June E3 August 1 ) B September (3 December D February k 1 23456789 10 Size Class were exposed to mortality. In the latter case, lowered recruitment no longer bal- anced the higher rates of mortality. In cases where mortality was imposed for time periods of less than one year, mortality restricted to the six summer months (April-September) nearly al- ways resulted in decreased population density compared to mortality restricted to the winter months. Rarely did the two yield similar results. Never did summer mortality have a lesser impact. The ef- fect was noted at different latitudes, in populations having mortality restricted to a variety of differing size classes, and in populations varying in larval sur- vivorship. However, adult mortality was required. Extending mortality into the juvenile size classes minimized the effect. Nearly all reports of population crashes in oyster populations result from adult summer mortality, recruitment failure, or floods. Most predators and parasites are most effective in the sum- mer. The series of simulations presented here suggests that the explanation for the importance of adult summer mortal- ity does not necessarily reside in the fact that the most significant agents of adult mortality (except the fishery) operate most effectively in the summer. Al- though this may well be true, the oys- ter itself would appear to be more sus- ceptible to mortality in the summer. That is, a greater chance of population crashes in the summer may be physi- ologically preordained. One potentially important mechanism causing this in- creased susceptibility is the temperature control on the partitioning of somatic tissue and reproductive tissue in the winter, spring, and summer. Fewer in- dividuals are present in the adult size Figure 14 A comparison of the changes in size-fre- quency distribution through the year in simulated Crassostrea virginica populations having size classes 3 and larger exposed to mortality. Compare to Figure 13 where mor- tality was restricted to size classes 5 and larger. (A and B) mortality restricted to the winter; (C and D) mortality restricted to the summer. More information in Figure 3 and Table 2, cases 31 and 32. 366 Fishery Bulletin 92(2), 1994 Galveston Bay 12 18 24 30 36 42 48 54 60 66 72 Julian Month 12000 n 8000 TO D ■o > ■o IV a 12 3 4 5 Julian Year 1 23456789 10 Size Class Figure 1 5 Simulated time development and population distribution of a Galveston Bay Crassostrea virginica population exposed to summer mortality at a yearly rate of 99.9% restricted to size classes 5 and larger and in which the food time series contained blooms in April/May and September/October. Figure 12 contains comparable results in which the food time series contained two blooms one month earlier. For additional information, see Figure 3 and Table 2, case 41. classes in the winter, hence losses are minimized. Juveniles grow rapidly to adulthood in the spring and spawn in the summer. As a result, reproductive effort is higher and population stability is enhanced when mortality is restricted to the winter. One of the interesting observations from the simu- lations is the consistent difference in the seasonal shifts in size-frequency distribution exhibited by populations suffering adult summer or winter mor- tality. Populations impacted most significantly by summer mortality had relatively stable size-fre- quency distributions over the year. Winter mortal- ity produced strong seasonal shifts in the size-fre- quency distribution. The results suggest that sea- sonal shifts in size-frequency distributions might provide a useful measure of the relative importance of summer and winter mortality and of adult mor- tality in oyster populations. For example, the sea- sonal cycle in market-sized individuals on some Galveston Bay reefs (e.g. Figure 2.1 in Quast et al., 1988) is similar to the seasonal shifts observed in simulated populations in which mortality was re- stricted to the winter months, suggesting that the fishery might be an important source of mortality in these populations. Latitudinal gradient in stability Although not conclusive, the literature reviewed earlier suggests a latitudinal gradient may exist in oyster population stability. Populations at higher latitudes may be more susceptible to population crashes. The Galveston Bay and Chesapeake Bay simulations support this possibility. Simulated popu- Powell et al.: Modeling oyster populations 367 4 5 6 7 Size Class 4 5 6 7 Size Class Figure 16 The seasonal changes in size-frequency distribution of the Crassostrea virginica population depicted in Figure 15. Figure 13 gives comparable results for the comparable simulation depicted in Figure 12. lations in Chesapeake Bay were more susceptible to population crashes than those in Galveston Bay. Simulated populations in Galveston Bay consis- tently had higher population densities after 6 years. Reproductive effort was higher because more of the year occurred within the temperature range condu- cive to spawning. Higher reproductive effort bal- anced a larger rate of mortality; hence mortality rates had to be substantially higher in Galveston Bay to effect a population crash. Although not simu- lated, recovery rates should have been faster as well. Like the distinction between winter and summer mortality, this latitudinal gradient in population stability would appear to result from the basic physi- ology of the oyster. The fundamental physiological mechanisms associated with reproduction and the division of net production into somatic and reproduc- tive growth would appear to be responsible. Implications for fisheries management The methods for managing the C. virginica fishery are generally limited to three somewhat intercon- nected decisions: 1) what size limit should be set; 2) what season should be allowed; and 3) what popu- lation density should trigger season closure? The setting of size limits may depend on biological and economic issues. Only biological issues will be con- sidered here. Two aspects of oyster physiology are most important in determining size limits. First, under conditions of crowding and at lower latitudes, oysters fail to grow to large size. The former is due to food-limiting conditions. The latter is due to warmer temperatures resulting in the shunting of net production into reproductive growth (Hofmann et al., in press). A considerable body of data supports food limitation in oyster populations, from aspects of spatial dis- tribution (Powell et al., 1987), to reduced growth in crowded locations (Osman et al., 1989), and the observation of in- creased growth coincident with high mortality (Crosby et al., 1991). A latitu- dinal gradient in size bespeaks of the importance of temperature in determin- ing the degree to which net production is allocated to somatic growth (Hofmann et al., in press). Both phenomena are reproduced by the model. Clearly, in ei- ther case, the setting of size limits as currently done has the effect of artifi- cially reducing yield. If economic consid- erations warrant it, lower size limits should be set in these populations. In crowded conditions, adult mortality might even increase adult size and yield. Second, raising size limits increases population density and, under certain conditions, the resulting increase in reproductive effort can eventually result in an increased number of market-size oysters at the larger size limit. Such conditions are met in popu- lations of relatively low density where oysters of legal size are already abundant. Of importance is the recognition that this condition occurs only in populations suffering a relatively high degree of mortality relative to the recruitment rate. Many other agents of mortality, besides the fishery, are important in oyster populations and these agents generally do not respect legal size limits. The model suggests that raising size limits will only be effec- tive if the fishery is the predominant cause of mor- tality in the population or if other agents of mortal- ity are generally restricted to these same size classes. If all adults are affected, then raising size limits will be ineffective. Besides the setting of size limits, management policy normally includes a restriction of the fishing season. Fishing seasons on public grounds are gen- erally restricted to the winter months. In some cases, certain areas are set aside for a summer sea- son as well. Natural mortality rates are high in oyster populations, generally greater that 70% per year (Mackin, 1959). Oyster populations in the Gulf of Mexico withstand this degree of mortality with- out long-term population declines. In this sense, the populations are stable (other species are stable at much higher mortality rates, e.g. Zonneveld [1991]). Rates of recruitment are sufficient to balance mor- tality over the long term. Nevertheless, population declines do occur (Sindermann, 1968; and others ref- erenced previously) and these have, on occasion, 368 Fishery Bulletin 92(2). 1994 Figure 1 7 Comparison of the time development of simulated Crassostrea virginica popula- tions exposed to mortality in three differ- ing size classes: (A and B), sizes 5 and larger; (C and D) sizes 6 and larger; (E and F) sizes 7 and larger. Cases A, C, and E show the time development under condi- tions where mortality was restricted to the summer. Cases B, D, and F show the time development under conditions where mor- tality was restricted to the winter. Further information in Figure 3 and Table 2, cases 34-39. been blamed on overfishing. Although no adequate data are available, one suspects that the fishery may be a principle source of mortality in the winter, but not in the summer when the various other agents of mortality, such as diseases and predators, are active. Oyster populations are more resis- tant to winter mortality than to sum- mer mortality. The increased likelihood of an intense population decline during the summer observed throughout the oyster's latitudinal range is a product of the basic physiology of the oyster. Simulated oyster populations were most resistant to population declines when mortality was restricted to the winter months under nearly all condi- tions of recruitment, size-class specific mortality and food supply; they were never less resistant. The simulations suggest that oyster populations can withstand substantially higher rates of mortality in the winter than in the summer and, under conditions where fishing is the primary cause of mortal- ity, populations should be managed more conservatively during the sum- mer season. A latitudinal gradient in stability exists in oyster populations. Population declines without short-term recovery are more likely at higher latitudes. The simulations suggest that populations should be more and more sensitive to natural agents of mortality and to management decisions at ever increasing latitudes. In effect, populations in the Gulf of Mexico, by their physiology, can withstand the vagaries of nature and the mistakes of man Galveston Bay 100 - = 10! f^ w^i ■ .V <7f..\. ,..<■■,. ■\.,..J. J\ i«i ft ' 1 i i i i ■•!•■.■■■* - 1500 1000 i -soo 100 - fl I ! n ,fTYr>,,,,-fTr?.>>. ^ 3000 2000 1000 30 36 42 Julian Month much easier than populations on the Mid-Atlantic and northeast coasts of the United States. The evi- dence suggests the need for more conservative oys- ter management at higher latitudes. In effect, the Gulf of Mexico populations and the northeastern Powell et al.: Modeling oyster populations 369 20 A . i i i 1 ■ SeeCk*ss5 D SaBOass6 □ SaBOass7 1. 4 5 6 7 Size Class ■ Sumter Mortality D Winter Mortality 3 4 5 Julian Year E ■ Summer Mortality Q Winter Mortaflty 3 4 5 Julian Year l> ■ ' .Lvl lr- : n D SDEdass6 30 - D SiEaa5s7 20 - 1 10 - l 1 i ^ II J h 1000 _ 800 400 200 1 23456789 10 Size Class I) ■ Summer Mortaflty n Winter Mortality 2 3 4 Julian Year Z 20 - Figure 1 8 A comparison of the final size-fre- quency distributions (day 2,160) (A and B), the yearly reproduc- tive efforts (C, D, and E) and the number of market-size individu- als in the Crassostrea virginica population after 6 years (F) for the simulations depicted in Fig- ure 17. (A) mortality restricted to the summer (Fig. 17, A, C, and E). (B) mortality restricted to the winter (Fig. 17, B, D, and F). (C) Yearly reproductive effort for populations exposed to mortality in sizes 5 and larger (Fig. 17, A and B). (D) Yearly reproductive effort for populations exposed to mortality in sizes 6 and larger (Fig. 17, C and D). (E) Yearly re- productive effort for populations exposed to mortality in sizes 7 and larger (Fig. 17, E and F). (F) The number of market-size indi- viduals in the population after 6 years, restricting the calculation of market-size individuals to the same classes suffering mortality, 5 and larger (sum-5. Fig. 17A; win-5, Fig. 17B), 6 and larger (sum-6, Fig. 17C; win-6, Fig. 17D) and 7 and larger (sum-7, Fig. 17E; win-7, Fig. 17F). populations exist under different physiological con- straints and these constraints demand different management philosophies and decisions. Acknowledgments We thank Elizabeth Wilson-Ormond and Stephanie Boyles for help in data acquisition. This research was supported by an institutional grant NA89-AA- D-SG139 to TAMU by the National Sea Grant Col- lege Program, National Oceanic and Atmospheric Administration (NOAA), U.S. Department of Com- merce, grant DACW64-91-C-0040 from the Army Corps of Engineers, Galveston District Office, and computer funds from the College of Geosciences and Maritime Studies Research Development Fund. Additional computer resources and facilities were provided through the Center for Coastal Physical Oceanography at Old Dominion University. We ap- preciate this support. Literature cited Andrews, J. D., D. Haven, and D. B. Quayle. 1959. Fresh-water kill of oysters {Crassostrea virginica) in James River, Virginia, 1958. Proc. Natl. Shellfish. Assoc. 49:29-49. Andrews, J. D., and W. G. Hewatt. 1957. Oyster mortality studies in Virginia. II: The fungus disease caused by Dermocystidium marinum in oysters in Chesapeake Bay. Ecol. Monogr. 27:1-26. 370 Fishery Bulletin 92(2), 1994 u> 100000 ■o 10000 > c E z 1000 100 10 1 i-— — ir— ID O Jf C 5 TO Q. GO 4000 3000 2000 1000 Bahr, L. M., and W. P. Lanier. 1981. The ecology of intertidal oyster reefs of the South At- lantic coast: a community profile. U.S. Fish Wildl. Serv., Biol. Serv. Program, FWS/OBS-81/15, 105 p. Berg, J. A., and R. I. E. Newell. 1986. Temporal and spatial variations in the composition of seston available to the sus- pension feeder Crassostrea virginica. Estuarine Coastal Shelf Sci. 23:375-386. Butler, P. A. 1949. Gametogenesis in the oyster under conditions of de- pressed salinity. Biol. Bull. (Woods Hole) 96:263-269. Chanley, P. E. 1958. Survival of some juvenile bivalves in water of low salinity. Proc. Natl. Shellfish. Assoc. 48:52-65. Choi, K- S., D. H. Lewis, E. N. Powell, and S. M. Ray. 1993. Quantitative measure- ment of reproductive output in the American oyster, Crasso- strea virginica (Gmelin), using an enzyme-linked immunosor- bent assay (ELISA). Aqua- culture Fish. Manag. 24:299-322. Copeland, B.J., and H.D. Hoese. 1966. Growth and mortality of the American oyster, Crasso- strea virginica, in high salin- ity shallow bays in central Texas Mar. Sci. Univ. Tex. 11:149-158. Crosby, M. P., C. F. Roberts, and P. D. Kenny. 1991. Effects of immersion time and tidal position on in situ growth rates of naturally settled east- ern oysters, Crassostrea virginica (Gmelin, 1791). J. Shellfish Res. 10:95-103. Cummins, K. W., and J. C. Wuycheck. 1971. Caloric equivalents for investigations in eco- logical energetics. Int. Ver. Theor. Angew. Limnol. Verh. 18:1-158. Cummins, H., E. N. Powell, R. J. Stanton Jr., and G. Staff. 1986. The size-frequency distribution in palaeoecology: the effects of taphonomic processes during formation of death assemblages in Texas bays. Palaeontology (Lond.). 29:495-518. Dame, R. F. 1972. The ecological energies of growth, respiration and assimilation in the intertidal American oys- ter Crassostrea virginica. Mar. Biol. (Berl.) 17:243-250. Chesapeake Bay A — — Number of Individual* Mortality (no/month) --- Spawn (kcal) Number of Adi ^ Its -- -rr— — . - i 0 M J, I ~^v(l - - -1— . ._ i. • ._!._. _1 l_~. 1 1 1 1 1 "^ 1 \ j- —Li u_i. ,/ 1 V, 5000 4000 3000 c S to 2000 - 1000 48 54 60 66 72 > 2 3 4 5 Julian Year 1 2 3456789 10 Size Class Figure 19 The time development of a simulated Crassostrea virginica population exposed to a seasonal temperature regime for Chesapeake Bay and sum- mer mortality. Additional information in Figure 3 and Table 2, case 44. Publ. Inst. Dekshenieks, M. M., E. E. Hofmann, and E. N. Powell. In press. Environmental effects on the growth and development of Crassostrea virginica (Gmelin) lar- vae: a modeling study. J. Shellfish Res. Deslous-Paoli, J- M., and M. Heral. 1988. Biochemical composition and energy value of Crassostrea gigas (Thunberg) cultured in the Bay of Marennes-Oleron. Aquat. Living Resour. 1:239-249. Doering, P. H., and C. A. Oviatt. 1986. Application of filtration rate models to field populations of bivalves: an assessment using experimental mesocosms. Mar. Ecol. Prog. Ser. 31:265-275. Dupuy, J. L., N. T. Windson, and C. E. Sutton. 1977. Manual for design and operation of an oys- ter seed hatchery. Virginia Institute of Marine Science, Gloucester Point, VA, 104 p. Engle, J. B. 1947. Commercial aspects of the upper Chesapeake Bay oyster bars in light of the recent oyster mortalities. Proc. Natl. Shellfish. Assoc, for 1946, p. 42-46. Powell et al.: Modeling oyster populations 371 1956. Ten years of study on oyster setting in a seed area in upper Chesapeake Bay. Proc. Natl. Shell- fish.Assoc. 46:88-99. 40 ■5 30 o 20 10 - B ■ March H June E3 August £ 30 en 20 D il i 1 ■ March 9 June G August J" , 1 2 3 4 5 6 7 Size Class Figure 20 The seasonal shift in size-frequency distribution for two Crassostrea virginica populations in Chesapeake Bay. (A and B) exposed to summer mortality; (C and D) exposed to winter mortality. (A and B) coincide with the simulation depicted in Figure 19. (C and D) coincide with Table 2, case 51. »u - § 40 - A ■ March fS June TJ □ August 1 30- o r a 20 " ■ i c « 1 1 . S 10- 4) a nlH 1 ill N ,-rfl, - 1 2 3 4 5 6 7 Size Class 8 9 10 Figure 21 The seasonal shift in size-frequency distribution of a Crassostrea virginica population in Chesapeake Bay exposed to an (A) early spring (March/April) and (B) early fall (August/September) bloom. The complementary case of two later blooms is depicted in Figure 20, C and D. More information can be found in Table 2, case 48. Engle, J. B., and A. Kosenfield. 1962. Progress in oyster mortality studies. Proc. Gulf Carib. Fish. Inst. 15:116-124. Epifanio, C. E., and J. Ewart. 1977. Maximum ration of four algal diets for the oyster Crassostrea virginica Gmelin. Aquaculture 11: 13-29. Frechette, M., D. A. Booth, B. Myrand, and H. Berard. 1991. Variability and transport of or- ganic seston near a mussel aquacul- ture site. ICES Mar. Sci. Symp. 192:24-32. Gallager, S. M., and R. Mann. 1986. Growth and survival of Mercen- aria mercenaria (L.) and Crassostrea virginica (Gmelin) relative to brood- stock conditioning and lipid content of eggs. Aquaculture 56:105-121. Galtsoff, P. S. 1964. The American oyster Cras- sostrea virginica Gmelin. U.S. Fish Wildl. Serv. Fish. Bull. 64:1-480. Galtsoff, P. S., W. A. Chipman Jr., J. B. Engle, and H. N. Calderwood. 1947. Ecological and physiological studies of the effect of sulphate pulp mill wastes on oysters in the York River, Virginia. U.S. Fish Wildl. Serv. Fish. Bull. 51:58-186. Garton, D., and W. B. Stickle. 1980. Effects of salinity and tempera- ture on the predation rate of Thais hae- mastoma on Crassostrea virginica spat. Biol. Bull. (Woods Hole) 158:49-57. Gilmore, G. H., S. M. Ray, and D. V. Aldrich. 1975. Growth and mortality of two groups of oysters (Crassostrea vir- ginica Gmelin), maintained in cooling water at an estuarine electric power generating station. Texas A&M Univ. Proj. Rep. TAMU-SG-75-207, 67 p. Glude, J. B. 1966. Criteria of success and failure in the management of shellfisheries. Trans. Am. Fish. Soc. 95:260-263. Gunter, G. 1955. Mortality of oysters and abun- dance of certain associates as related to salinity. Ecology 36:601-605. Hayes, P. F., and R. W. Menzel. 1981. The reproductive cycle of early setting Crassostrea virginica (Gmelin) in the northern Gulf of Mexico, and its implications for population recruitment. Biol. Bull. (Woods Hole) 160:80-88. 372 Fishery Bulletin 92(2). 1994 Hofmann, E. E., E. N. Powell, J. M. Klinck, and E. A. Wilson. 1992. Modeling oyster populations. Ill: Critical feeding periods, growth and reproduction. J. Shellfish Res. 11:399-416. Hofmann, E. E., J. M. Klinck, E. N. Powell, S. Boyles, and M. Ellis. In press. Modeling oyster populations. II: Adult size and reproductive effort. J. Shellfish Res. Hofstetter, R. P. 1966. Oyster mortality studies along the Texas coast during 1966. Texas Parks Wildl. Dept. Coastal Fish. Proj. Rep., p. 55-68. 1983. Oyster population trends in Galveston Bay 1973-1978. Tex. Parks Wildl. Dept. Management Data Ser. no. 51, 33 p. Hofstetter, R. P., and S. M. Ray. 1988. Managing public oyster reefs: Texas experience. J. Shellfish Res. 7:213. Hofstetter, R. P., T. L. Heffernan, and B. D. King III. 1965. Oyster (Crassostrea virginica) mortality stud- ies along the Texas coast. Tex. Parks Wildl. Dep. Coastal Fish. Proj. Rep., p. 119-131. Kennedy, V. S. 1982. Sex ratios in oysters, emphasizing Crassostrea virginica from Chesapeake Bay, Maryland. Veliger 25:329-338. Klinck, J. M., E. N. Powell, E. E. Hofmann, E. A. Wilson, and S. M. Ray. 1992. Modeling oyster populations: the effect of density and food supply on production. Proc. Adv. Mar. Technol. Conf. 5:85-195. Koehn, R. K., and B. L. Bayne. 1989. Towards a physiological and genetical under- standing of the energetics of the stress response. Biol. J. Linn. Soc. 37:157-171. Laird, M. 1961. Microecological factors in oyster epizootics. Can. J. Zool. 39:449-485. Lee, R. F., and P. B. Heffernan. 1991. Lipids and proteins in eggs of eastern oys- ters (Crassostrea virginica Gmelin, 1791) and northern quahogs (Mercenaria mercenaria Linnaeus, 1758). J. Shellfish Res. 10:203-206. Loosanoff, V. L. 1958. Some aspects of behavior of oysters of differ- ent temperatures. Biol. Bull. (Woods Hole) 114:57-70. Loosanoff, V. L., and F. D. Tommers. 1948. Effect of suspended silt and other substances on rate of feeding of oysters. Science (Wash. D.C. i 107:69-70. Lund, E. J. 1957. A quantitative study of clearance of a turbid medium and feeding by the oyster. Publ. Inst. Mar. Sci. Univ. Tex. 4:296-312. MacKenzie, C. L., Jr. 1989. A guide for enhancing estuarine molluscan shellfisheries. Mar. Fish Rev fill 3 1:1-47. Mackin, J. G. 1959. A method of estimation of mortality rates in oysters. Proc. Natl. Shellfish.Assoc. 50:41-51. Mackin, J. G., and S. H. Hopkins. 1962. Studies on oyster mortality in relation to natural environments and to oil fields in Lou- isiana. Publ. Inst. Mar. Sci. Univ. Tex. 7:1-131. Mackin, J. < ... and A. K. Sparks. 1962. A study of the effect on oysters of crude oil loss from a wild well. Publ. Inst. Mar. Sci. Univ. Tex. 7:230-261. Mackin, J. G., and D. A. Wray. 1950. Report on the second study of mortality of oysters in Barataria Bay, Louisiana, and adjacent areas. Texas A&M Research Foundation Proj. Rep., Proj. 9, 39 p. Mackin, J. G., B. Welch, and C. Kent. 1950. A study of mortality of oysters of the Buras area of Louisiana. Texas A&M Research Founda- tion Proj. Rep., Proj. 9, 41 p. Menzel, R. W. 1950a. Report on oyster studies in Caillou Island oil field, Terrebonne Parish Louisiana. Texas A&M Research Foundation Proj. Rep., Proj. 9, 60 p. 1950b. Report of oyster studies in Lake Felicity and Bayou Bas Bleu, Terrebonne Parish, Louisiana. Texas A&M Research Foundation Proj. Rep., Proj. 9, 46 p. Menzel, R. W., and S. H. Hopkins. 1953. Report on oyster experiments at Bay Sainte Elaine oil field. Texas A&M Research Foundation Proj. Rep., Proj. 9, 170 p. Menzel, R. W., N. C. Hulings, and R. R. Hathaway. 1957. Causes of depletion of oysters in St. Vincent Bar, Apalachicola Bay, Florida. Proc. Natl. Shell- fish. Assoc. 48:66-71. Moore, H. F., and T. E. B. Pope. 1910. Oyster culture experiments and investiga- tions in Louisiana. U.S. Dep. Commerce Labor Bur. Fish. Document No. 731, 52 p. Mountford, K., and R. C. Reynolds. 1988. Potential biological effects of modeled water quality improvements resulting from two pollut- ant reduction scenarios. In Understanding the estuary: advances in Chesapeake Bay research. Proceedings of a conference 29-31 March 1988. Chesapeake Research Consortium Publ. 129, Bal- timore, MD, p. 593-606. Needier, A. W. H., and R. R. Logie. 1947. Serious mortalities in Prince Edward Island oysters caused by a contagious disease. Trans. R. Soc. Canada Ser. Ill Sect. V 41:73-89. Newell, R. I. E. 1985. Physiological effects of the MSX parasite Haplosporidium nelsoni (Haskin, Stauber & Mackin) on the American oyster Crassostrea virginica. J. Shellfish Res. 5:91-95. Osman, R. W., R. B. Whitlatch, and R. N. Zajac. 1989. Effects of resident species on recruitment into a community: larval settlement versus post- Powell et al.: Modeling oyster populations 373 settlement mortality in the oyster Crassotrea virginica. Mar. Ecol. Prog. Ser. 54:61-73. Owen, H. M. 1953. Growth and mortality of oysters in Louisiana. Bull. Mar. Sci. Gulf Caribb. 3:44-54. Powell, E. N., H. Cummins, R. J. Stanton Jr., and G. Staff. 1984. Estimation of the size of molluscan larval settlement using the death assemblage. Estu- arine Coastal Shelf Sci. 18:367-384. Powell, E.N., M.E. White, E.A. Wilson, and S.M. Ray. 1987. Small-scale spatial distribution of oysters (Crassostrea virginica) on oyster reefs. Bull. Mar. Sci. 41:835-855. Powell, E. N., E. E. Hofmann, J. M. Klinck, and S. M. Ray. 1992. Modeling oyster populations. I: A commentary on filtration rate. Is faster always better? J. Shellfish Res. 11:387-398. Prytherch, H. F. 1929. Investigation of the physical conditions con- trolling spawning of oysters and the occurrence, distribution, and setting of oyster larvae in Milford Harbor, Connecticut. Bull. Bur. Fish. 44:429-503. Quast, W. D., M. A. Johns, D. E. Pitts Jr., G. C. Matlock, and J. E. Clark. 1988. Texas oyster fishery management plan source document. Texas Parks Wildl. Dept Coastal Fish. Branch, Austin, TX, 178 p. Ray, S. M., and A. C. Chandler. 1955. Dermocystidium marinum, a parasite of oysters. Exp. Parasitol. 4:172-200. Schlesselman, G. W. 1955. The Gulf coast oyster industry in the U.S. Geogr. Rev. 45:531-541. Shumway, S. E., and R. K. Koehn. 1982. Oxygen consumption in the American oyster Crassostrea virginica. Mar. Ecol. Prog. Ser. 9:59-68. Sindermann, C. J. 1968. Oyster mortalities, with particular reference to Chesapeake Bay and the Atlantic coast of North America. U.S. Fish Wildl. Serv. Bur. Comm. Fish. Spec. Sci. Rep. Fisheries No. 569, 10 p. Soniat, T. M., and S. M. Ray. 1985. Relationships between possible available food and the composition, condition and reproductive state of oysters from Galveston Bay, Texas. Contrib. Mar. Sci. 28:109-121. Soniat, T. M., and M. S. Brody. 1988. Field validation of a habitat suitability index model for the American oyster. Estuaries 11:87-95. Soniat, T. M., S. M. Ray, and L. M. Jeffrey. 1984. Components of the seston and possible avail- able food for oysters in Galveston Bay, Texas. Contrib. Mar. Sci. 27:127-141. Soniat, T. M., L. E. Smith, and M. S. Brody. 1989. Mortality and condition of the American oys- ter in Galveston Bay, Texas. Contrib. Mar. Sci. 31:77-94. Stanley, J. G., and M. A. Sellers. 1986. Species profile: life histories and environmen- tal requirements of coastal fishes and inverte- brates (Gulf of Mexico) — American oyster. U.S. Fish Wildl. Serv. Biol. Rep. 82 (11.64), U.S. Army Corps of Engineers, TR EL-82-4, 25 p. Stickle, W. B. 1985. Effects of environmental factor gradients on scope for growth in several species of carnivorous marine invertebrates. In J. S. Gray and M. E. Christiansen (eds.), Marine biology of polar regions and effects of stress on marine organisms. John Wiley & Sons Ltd., New York, p.601-616. Underwood, A. J. 1989. The analysis of stress in natural populations. Biol. J. Linn. Soc. 37:51-78. White, M. E., E. N. Powell, and S. M. Ray. 1988. Effects of parasitism by the pyramidellid gastropod Boonea impressa on the net productiv- ity of oysters (Crassostrea virginica). Estuarine Coastal Shelf Sci. 26:359-377. Wright, D. A., and E. W. Hetzel. 1985. Use of RNA:DNA ratios as an indicator of nutritional stress in the American oyster Crassostrea virginica. Mar. Ecol. Prog. Ser. 25:199-206. Young, P. C, and R. B. Martin. 1989. The scallop fisheries of Australia and their management. Crit. Rev. Aquat. Sci. 1:615-638. Zonneveld, C. 1991. Estimating death rates from transect counts. Ecol. Entomol. 16:115-121. Abstract. Surplus-produc- tion models, because of their sim- plicity and relatively undemanding data needs, are attractive tools for many stock assessments. This pa- per reviews the logistic production model, starting with the basic dif- ferential equation and continuing with a description of the model de- velopment without the equilibrium assumption. It then describes sev- eral extensions, including "tuning" the model to a biomass index; par- titioning fishing mortality by gear, time, or area; and making projec- tions. Computation of confidence intervals on quantities of interest (e.g. maximum sustainable yield (MSY), effort at MSY, level of stock biomass relative to the optimum level) can be done through boot- strapping, and the bootstrap can also be used to construct nonpara- metric tests of hypotheses about changes in catchability. To fit the model, an algorithm that uses a forward solution of the population equations can be implemented on a small computer. An example of the utility of surplus-production models (illustrating several of these extensions) is given. The ex- ample is loosely based on swordfish (Xiphias gladius) in the North At- lantic Ocean, but is not intended to describe the actual status of that stock. A suite of extensions to a nonequilibrium surplus-production model* Michael H. Prager Miami Laboratory. Southeast Fisheries Science Center National Marine Fisheries Service. NOAA 75 Virginia Beach Drive, Miami, Florida 33149 Cooperative Unit for Fisheries Education and Research Rosenstiel School of Marine and Atmospheric Science University of Miami, 4600 Rickenbacker Causeway, Miami, Florida 33149 Despite the prevalence of age-struc- tured population models, surplus- production models — which gener- ally do not incorporate age struc- ture— remain useful for analysis of fish population dynamics. These models are of particular value when the catch cannot be aged, or cannot be aged precisely, and therefore age- structured models cannot be ap- plied. Surplus-production models are also useful as a complement to age-structured models, providing another view of the data and the fisheries. An especially appealing aspect of production models is their simplicity; from a scientific point of view, this makes exploration of their properties easier; from a manage- ment point of view, it makes their results easier to present and under- stand (Barber, 1988). In this paper, I show that another benefit of these models' simplicity is that model extensions are easily made. Examples of such extensions include modeling several simulta- neous or sequential fisheries on the same stock, "tuning" the model to a biomass index (as is often done in age-structured models; e.g. the CAGEAN model of Deriso et al., 1985; the CAL model of Parrack, 1986; the ADAPT model of Gavaris, 1988), modeling changes in catch- ability or population characteristics (e.g. carrying capacity), and esti- mating missing values of fishing ef- fort. Many of these extensions have not been presented before. The comprehensiveness of a pro- duction model can be further in- creased by introducing another ex- tension: computation of nonpara- metric estimates of variability in the results. These can be obtained by bootstrapping, and can be used both to describe the results more completely and to learn more about the model's behavior under a vari- ety of circumstances. After reviewing the formulation of the simplest surplus-production model (the logistic model), a num- ber of extensions to the model are described. An example, loosely based on swordfish, Xiphias gladius, in the North Atlantic Ocean, is pre- sented to illustrate typical results from the model and the use of many of the extensions. The example, which is not intended to be an assessment of that stock, should not be used to make inferences about stock status. Manuscript accepted 18 October 1993 Fishery Bulletin 92:374-389 (1994) Contribution MIA-92/93-58 of the Miami Laboratory, Southeast Fisheries Science Center, National Marine Fisheries Service. 374 Prager: A nonequilibrium surplus-production model 375 Model formulation and fitting Basic differential equations Surplus-production models characterize a population as an undifferentiated biomass. The number of indi- viduals present or harvested plays no part in these models, nor is age or size structure considered. A quantity termed "surplus production" is used to char- acterize population dynamics at different levels of population size (measured in biomass). Surplus pro- duction is the algebraic sum of three major forces: recruitment, growth, and natural mortality. The ad- jective "surplus" refers to the surplus of recruitment and growth over natural mortality; i.e. the net pro- duction. In this article, surplus production will often be termed simply "production," and the models termed "production models." In the simplest production model, the logistic or Graham-Schaefer (Graham, 1935; Schaefer, 1954, 1957) model, a first-order differential equation de- scribes the rate of change of stock biomass Bt due to production. In the absence of fishing, the population's rate of increase or decrease is assumed to be a func- tion of the current population size only: ^L = rBt-^Bf, dt * K *' (1) where Bt is the population biomass at time t and r and K are parameters. The right side of Equation 1 is simply the start of the Taylor expansion of an ar- bitrary function 0(B) passing through the origin (Lotka, 1924). Equation 1 is written in the parameterization of population ecology, in which K represents the maxi- mum population size, or carrying capacity, and r rep- resents the stock's intrinsic rate of increase (in pro- portion per unit time). In this paper, both are as- sumed constant. Other parameterizations could be used, and indeed a slightly different parameteriza- tion is used for simplicity in the next section. Adding fishing mortality Ft to the model, it be- comes f-^-l* (2) This model, like many fisheries models, is much sim- pler than the real world. In particular, it excludes such factors as environmental variation, interspe- cific effects, or the possible presence of more than one stable regime. Time trajectories of biomass and yield Integration of Equation 2 with respect to time al- lows modeling the biomass and yield through time. Before integration, simplify notation by defining at=r-Ft and fi=r/K to express Equation 2 more sim- ply as dBt dt «tBt-ffi (3) Equation 3 can be conveniently solved for biomass under the assumption that F is constant and that therefore a( is constant. This is a weak assumption, for if Ft varies, time can be divided into short peri- ods of constant or nearly constant F and a solution found for each period. Fitting would then require knowing the catch and effort for each short period. For the period beginning at t = h and ending at time t = h + 8, during which the instantaneous fish- ing mortality rate is Fh, the solution to Equation 3 is B h+5 ahBhe ahS ah+pBh(ea^s-l) when ah * 0, or (4a) B Bh h+5 l+p8Bh when ah = 0. (4b) Equation 4a is the familiar logistic equation. How- ever, if ah = 0 (i.e. if F h = r), Equation 4a is undefined and Equation 4b is used in its place. Modeling the yield during the same period involves another integration with respect to time: J'h+S t=h FhBtdt, (5) where Br the biomass at instant t, is defined by Equa- tions 4a and 4b; Fh is the (constant) instantaneous rate of fishing mortality during the time period; and Yh is the yield taken during the period. Performing the integration in Equation 5, V*=-m PBh(l- „ahS , ■ln(l+5)SBA) when ah *0, or (6a) when ah = 0. (6b) Equation 6a was apparently first given by Pella (1967) (and a similar form developed by Schnute [1977 J); Equations 4b and 6b seem not to have been presented in fishery biology before now. It follows from the definition of F that the esti- mated average biomass during the period is simply Yh I Fh. The surplus production Ph during the time period can be determined by mass balance: B h+S Bh+Yh (7) When yield is equal to surplus production, the popu- lation is in equilibrium. 376 Fishery Bulletin 92(2). 1994 Parameter estimation Parameter estimation for this model can be accom- plished by a number of methods. The method pre- sented here is a slight modification of one originated by Pella (1967), later used by Pella and Tomlinson (1969), and recently termed the "time-series method" by Hilborn and Walters (1992). Although it is not necessary to use equal time periods, the treatment in the balance of this paper assumes, for simplicity, that there are T equal time periods, indexed by x = (1, 2, ..., T), and that a period is one year in duration. The following symbols are used: /; F: population biomass at the start of year x yield in biomass during year x surplus production during year x, fishing effort rate during year x, fishing mortality rate during year x, function of F', a = r - FT. Estimates of the first five of these quantities are rep- resented by BT,YT,PT,fT, andFr. An important additional assumption is that, for all x, F = qf ; in other words, that fishing mortality rate is proportional to fishing effort rate and that the catchability coefficient q is constant. (The as- sumption of constant q is slightly relaxed later.) The data required for fitting are, for each time period x, data on effort fz and the yield Yv where x = II, 2, ..., T\ and T> 4. The parameters to be esti- mated are r and K in Equation l,q, and B,, the bio- mass at the beginning of the first year. The simplest procedure accumulates residuals in yield. To perform the estimation, the following algorithm is used: Al Obtain starting guesses for the four parameters. A2 Beginning with the current estimate of Bv project the population through time according to Equations 4a and 4b. For each year of the projection, com- pute estimated yield from Equations 6a and 6b. A3 Compute the objective function to be minimized. Assuming a multiplicative error structure in yield, this is T y[iog(yr)-iog(yr)f. A4 Monitor the objective function for convergence. If achieved, end. Otherwise, revise the param- eter estimates (using a standard minimization scheme) and continue at step A2. The simplex or "polytope" algorithm (Nelder and Mead, 1965; Press et al., 1986) works well as the minimization scheme in this application. Although not as rapid computationally as some other meth- ods, the simplex algorithm is quite robust to start- ing values and is easily manipulated (by restarts) to avoid local minima (see Press et al., 1986, p. 292). Rivard and Bledsoe (1978) used the Marquardt ( 1963) algorithm successfully for estimation in a similar model. The estimation method just described uses the re- corded effort in each year to estimate yield. Alterna- tively, one could use the recorded yield in each year to estimate the fishing mortality rate (or equivalently, the fishing effort rate). The solutions of Equations 6a and 6b for fishing mortality rate are when ccT * 0, or when aT =0. (8a) (8b) ln[l+)3Br] To use this approach, one must revise the second and third steps of the algorithm to become — A2' Beginning with the current estimate of B,, com- pute the estimated fishing effort for each year by solving Equation 8a or 8b and dividing by q . Project the population to year-end with Equa- tion 4. A3' Compute the objective function to be minimized. Assuming a multiplicative error structure in effort, this is X[log(/-r)-log(/r) r=l This is equivalent to minimizing the sums of squared residuals in the logarithm of catch per unit of effort, i.e. to minimizing £[iog(cT//;)-iog(cT//;) r=l A significant practical advantage of the second approach is that it simplifies the analysis of data with some missing data on effort. During parameter esti- mation, effort is estimated for all years; for years of missing effort, the contribution to the objective func- tion is simply defined to be zero. In contrast, the computations for the first approach are not possible without data on effort for each year. Estimating effort from yield introduces two small practical difficulties. The first difficulty is that Equa- tion 8a is not an explicit solution for effort (because a includes f), so it must be solved iteratively. This is accomplished by putting a starting guess FT into the right-hand side of the equation, solving, and sub- stituting the result repeatedly until convergence is achieved. A logical starting guess is FT = YT/ Br. Prager: A nonequilibrium surplus-production model 377 The second difficulty involves a fundamental dif- ference between predicting yield and predicting ef- fort. For a given starting biomass and effort, one can always compute the corresponding yield. For a given starting biomass, however, there are some yields that can never be obtained, no matter how high the ef- fort. Under these circumstances, the catch equation (6a or 6b) has no solution. Unless a tactic is devised for such cases, it becomes impossible to compute the objective function when they occur, and thus impos- sible to conduct its minimization. A tactic suggested by R. Methot1 as useful in his stock-synthesis model (Methot, 1989, 1990) is to place a constraint on the maximum allowable value of Fx (and consequently of/"T). When an estimate of F reaches this constraint, it is not allowed to increase further, and the quantity [log{Yr)-log(YT)]2 is added to the objective function along with the usual squared residual in effort. This allows computation of a reasonable value of the objec- tive function for such regions of the solution space that may be encountered during optimization. In my expe- rience, however, final estimates have always come from a solution in which yield is always matched exactly. In fitting a linear regression, observation error in the predictor variables causes problems with the pa- rameter estimates, including inconsistency and, in the bivariate linear case, bias towards zero (Thiel, 1971; Kennedy,1979). The problems induced into nonlinear models are less well understood, but are believed to be similar. Schnute ( 1989) has illustrated how the choice of dependent variable in a fisheries model can affect the results substantially. In fisher- ies contexts, yield is usually observed more precisely than fishing effort; for that reason, it seems prefer- able on statistical grounds to use the second ap- proach, estimating effort from yield, rather than es- timating yield from effort. Whichever approach is chosen, the estimation pro- cess results in direct estimates of Bv r, K, and q, which define unique estimates of the stock biomass levels B2, B3, ..., BT and the stock's production dur- ing each period of time. The corresponding estimate of maximum sustainable yield (MSY) under the lo- gistic model is MSY = Kr 1 4 . According to the theory of production modeling, MSY can be attained as a sustainable yield only at one specific stock size; for the logistic model this is BMSY = K/2, estimated by BMSY =K/2. The instantaneous fishing mortality that generates MSY at BMSY is ^msy = r/2 ; the cor- responding rate of fishing effort is fMSY = rl2q, with estimates given by substituting rand q for the un- known true values in these two expressions. The logarithmic objective function assumes mul- tiplicative errors with constant variance. The solu- tion obtained is the maximum-likelihood solution if the transformed residuals are independent, of con- stant variance, and normally distributed (see Seber and Wild, 1989). However, maximum-likelihood methods, while generally desirable, are not neces- sarily robust to outliers, nor do they necessarily have desirable small-sample properties. Use of a robust- regression method (such as least absolute values re- gression) would be an interesting research topic. Another management benchmark An analogue of the management benchmark FQ 1 can be computed for this model (or for any production model). The derivative of equilibrium yield with re- spect to fishing mortality rate for this model is dF (9) 1 Methot, R. Alaska Fisheries Science Center. 7600 Sand Point Way NE, Seattle, WA 98115. Personal commun., 1993. At F = 0, this derivative is equal to K. We define as F0 j for this model as the value of F at which Equa- tion 9 equals 0. 1 K. Substitution into Equation 9 gives the following results: F0 , = 0.45 r, and YQ 1 = 0.2475 rK (where Y0 1 is the equilibrium yield corresponding to FQ j). An equivalent statement is that FQ 1 is 90% of FMSY, and Y01 is 99% of MSY. Punt (1990) used the concept of FQ , for a production model but did not explicitly state these relationships. Penalty for large estimates of 6, Logistic production theory implies that Bx should always be less than K, but the objective functions used here are relatively insensitive to the estimate of By In practice, I have found that the estimate of Bj obtained from some data sets tends to be much larger than the estimate of K. Such results could be eliminated by introducing a fixed constraint into the solution, but I have used another method success- fully: adding a penalty term to the objective function when Bt>K- Including this term, the complete loga- rithmic objective function (assuming residuals in ef- fort) becomes L = ^[log(/-r)-log(/r)]2 + 0[log(B1)-log(A-)]2,(lO) where

K, and 0 = 0, otherwise. While constraining the value of Bj seems logical in accor- dance with the underlying population theory, such constraints can change the estimates of other param- eters, compared to an unconstrained solution. The amount of change can be examined by estimating with and without the penalty term or fixed constraint. 378 Fishery Bulletin 92(2). 1994 Extensions to the model A great strength of the model presented here is the ease with which it can be extended and modified. Such extensions can include, for example, modeling fisheries divided by space, time, or gear type; ana- lyzing data series including some years of no effort, as would occur during a closure; analyzing data se- ries with years of missing or highly uncertain effort data; incorporating changes in catchability within the data series, perhaps after periods of closure or following regulatory changes; and tuning the model to fishery-independent estimates or indices of popu- lation biomass. Missing data Gaps in the effort and yield time series do not present a problem to these dynamic production model analy- ses. Years with no effort (and therefore no catch) can easily be treated by defining the residual to be zero. Although such years do not influence parameter es- timation directly, the time lag during the years of closure carries information that is incorporated in fitting the model, and an estimate of population bio- mass for each missing year is made according to the logistic growth model. In contrast, years of closure contribute no information to production models that assume equilibrium conditions. A slightly more difficult problem is the correct treatment of years in which effort is known to have existed, but for which the data are missing or highly uncertain. In such a case, the framework presented here can be used to estimate, simultaneously with the other parameters, effort levels for a limited num- ber of such years within the series. As in any estima- tion scheme, the total number of estimated param- eters should be kept reasonably small in comparison to the number of years of nonzero data. If residuals are constructed in effort (rather than yield) the esti- mation of missing effort becomes trivial, as a pre- dicted effort is computed for each year during pa- rameter estimation. More than one data series Another simple extension of the basic estimation framework is analysis of stocks fished by two or more different gear types, either in the same years or se- rially For convenience, I refer to these as different fisheries on the same stock. To define the situation more precisely, there are J different fisheries, indexed by j = (1, 2, ..., J). The effort applied by fishery,/ in period T is f-, the catchability coefficient of that fish- ery is o , and the yield in period x is Y z. All q are assumed time— invariant. The total instantaneous fishing mortality in period x is Fr^Qjfjr- (ID Biomass and yield projections can be computed by Equations 4a, 4b, 6a, and 6b as before. The estimated yield from fishery./' in period t is v - -LlLLLv JT F, " (12) where Y is the total yield in period x. During pa- rameter estimation, a residual is obtained for each fishery having nonzero effort in period t. The contri- bution to the objective function for each period is thus composed of a sum of terms, one for each fishery with nonzero effort. In addition, the individual fisheries may carry different statistical weights to reflect vary- ing levels of confidence in the data from each fish- ery. Inverse-variance weighting can be approximated by iteratively examining the mean-squared error (MSE) from each series, weighting, and re-estimat- ing the parameters. Model tuning If an external series of population biomass estimates is available, it can be incorporated into the analysis in a procedure analogous to tuning an age-structured analysis. The external estimates are compared to the population estimates derived within the production model and the residuals incorporated in computa- tion of the objective function. Rivard and Bledsoe (1978) suggested this possibility, but did not pursue the idea, and it has also been described by Hilborn and Walters ( 1992). The external biomass series need not be continuous, but may contain missing values; the series' contribution to the objective function is set to zero for years with missing values. An exter- nal index of biomass can be used similarly, with the cost of estimating one more parameter (the catchability associated with the index). The model formulation involved in tuning the model is similar to that used when fitting more than one fishery. As in that situation, each year's contri- bution to the objective function consists of a sum of terms. Here, the sum includes a term from each fish- ery and a term for each biomass— estimate or index series. For a maximum-likelihood solution, the com- ponents should carry statistical weights inversely proportional to their variances. Varying catchability In many situations, catchability is thought to change relatively suddenly, perhaps because more efficient Prager: A nonequilibrium surplus-production model 379 gear for finding or catching the fish is introduced. In such cases, the formulation represented by Equations 11 and 12 can be used to estimate different catchability coefficients for segments of a single time series. In formulating such a model, the time seg- ments would be treated as separate fisheries, each having nonzero catch and effort data only during its respective time period. Each additional time segment would add one additional parameter to the model. A common concern is determining whether the improvement in fit obtained from a more complex model is statistically significant. A production model with added catchability parameters can be tested against the simpler model (with one estimated q) with a standard F-ratio test. (Here F refers to the F dis- tribution of statistics, not to fishing mortality rate.) The test statistic F* is F* (SSEs-SSEc)/t;1 SSEc/t;2 (13) where SSEs and SSEc are the error sums of squares of the simple and complex models, respectively; vl is the difference in number of estimated parameters between the two models; and v2 is the number of data points less the total number of estimated parameters. The significance probability of F* can be obtained from tables of the F-distribution with vx and v2 de- grees of freedom. As pointed out by a referee, this is equivalent to to a likelihood-ratio test assuming log- normal error structure, which is implicit in using the SSE from log-transformed data. Because of the pos- sibility of specification error, any such significance test must be considered approximate. A nonparametric test of the null hypothesis qx = q2 can also be conducted by examining a bias-cor- rected confidence interval on the ratio of the two catchability coefficients. (Construction of bias-cor- rected confidence intervals is described later.) As an example, assume that the alternative hypothesis is qx * q2. The null hypothesis would be rejected at P<0.05 if a 95% confidence interval on the ratio qx I q2 did not include the value 1.0. Like the F-test, this test is approximate because of the possibility of speci- fication error. In other cases, catchability is thought to vary in more subtle ways than the step function just sug- gested (Paloheimo and Dickie, 1964; Gulland, 1975; MacCall, 1976; Peterman and Steer, 1981; Winters and Wheeler, 1985), and one could incorporate any number of catchability models into the estimation framework. It would be straightforward to model a linear trend (increase or decrease) in catchability with time. This could be parameterized by estimat- ing the first and last years' values of q and generat- ing intermediate years' values by linear interpola- tion, so that only one additional parameter would be estimated. One could also add some form of density- dependent catchability model with a minimal cost in terms of number of parameters estimated; the foundation of such an approach was presented by Fox (1975). However, it might prove difficult to distin- guish varying catchability from trends in biomass itself. If so, the use of external estimates or indices of biomass, as explained above, might be especially valuable. Bootstrap estimates of bias and variability The bootstrap (Efron, 1982; Stine, 1990) is a sample reuse technique often used to estimate sampling vari- ances, confidence intervals, bias, and similar prop- erties of statistics, including parameter estimates. Major advantages of the bootstrap, compared to al- ternative methods (such as those based on the infor- mation matrix), are its flexibility and relative free- dom from distributional assumptions. A minor draw- back is that it demands a great deal of computer time. Bootstrapping is often performed by resampling the original observations. However, in fitting non- equilibrium production models, the order of the catch- effort pairs is as significant as the data themselves. For time-series models (in the broad sense), Efron and Tibshirani (1986) describe a method, used here, that preserves the original time structure of the data. For each bootstrap trial (of which there may be 250 to several thousand), a set of synthetic observations is constructed by combining the ordered predictions from the original fit with residuals chosen at ran- dom (with replacement) from the set of residuals from the original fit. The model is then refit to this set of synthetic observations. The bootstrap can be used to estimate bias in pa- rameter estimates. The median estimation bias Bg in a parameter 6 is estimated by Be -Qm -0, (13a) where 0 is the conventional estimator of 6, and 8m is the median value of 6 obtained from the bootstrap trials (Efron,1982; Efron and Gong, 1983). A bias- corrected estimator 6BC of a parameter 6 can there- fore be given by 0 BC e-B, '6 ■ (13b) It appears that the median bias correction, rather than a mean correction, has been adopted in the bootstrapping literature because a mean correction (which would be expected to produce an "unbiased" estimate in the usual sense) can have extremely high variance (Hinkley, 1978). The resulting problems are 380 Fishery Bulletin 92(2). 1994 avoided by use of a median correction, which is quite resistent to outliers. However, the use of a median correction implies that the estimated bias correction will be nonzero for an estimator that is unbiased (in the usual sense) but arises from a distribution in which the median does not equal the mean. That is, the use of a median bias correction transforms the estimator into a median estimator. Several methods have been developed for comput- ing bias-corrected confidence intervals from the boot- strap (Efron, 1982; Efron, 1987; Noreen, 1989). The most widely used at present appears to be the BC method of Efron (1982). To compute a BC interval, let Mz) be the cumulative distribution function (CDF) of the standard normal distribution and \etN~l be the inverse— normal CDF. Let C be the empirical bootstrap CDF of the parameter 9 ; i.e. C(g) is the proportion of realizations of 9 in the bootstrap distribution that falls below any arbitrary value g. Define the constant z*=N~ C(9) 14) where 9 is the conventional estimator. Then, the (1 - 2a) BC central confidence interval on 9 is de- fined as 0e{c-1[7V(22o+iV-1(a))],C-I[iV(22o-iV-1(a))]}. (15 This method assumes that a transformation exists under which the distribution of 9 becomes normal and homoscedastic. However, the form of the trans- formation need not be known (Efron and Gong, 1983). Kizner (1991) constructed bootstrap confidence in- tervals on production-model results, but he did not state whether bias corrections were used. This discussion of bootstrapping has referred to estimated "parameters" for simplicity, but the method can be used to estimate bias corrections and bias- corrected confidence intervals for any estimated quantity. Such quantities might include estimates of MSY, /"MgY, the population biomass in the final (or any other) year, f01, projections of biomass levels (discussed next), and so forth. Projections Because a production model implicitly includes a recruitment function, it can be used to make projec- tions based on hypothetical catch or effort quotas. As noted above, the historical population biomass trajectory is estimated during parameter estimation. The modeled population can then be projected for- ward in time by using the same population equations (4, 6, 8), and a proposed set of yields or effort rates. If the bootstrap is used following parameter estima- tion, the results of each bootstrap trial can be pro- jected forward. From the results, it is possible to com- pute bias— corrected point estimates and confidence intervals on the projection results. Example: North Atlantic swordfish Many aspects of the production model discussed above are illustrated in this example, which is loosely based on swordfish, Xiphias gladius, in the North Atlantic Ocean. The example comprises two analy- ses, the difference between them being the use of an abundance index for tuning the second analysis. Both the base analysis and the tuned analysis used the same yield and fishing-effort data (Table 1; Fig. 1); the tuned analysis also used a hypothetical index of abundance constructed for this purpose (Table 1; Fig. 1). In both analyses, errors were assumed to occur in fishing effort and to follow a lognormal distribution; in other words, the "second estimation approach" described previously was used. Each analysis in- cluded a projection of five years beyond the actual data; during those five years, it was assumed that a yield of 12,000 metric tons would be taken annually. Each analysis included a bootstrap with 1,000 trials. This example is not intended as, and should not be considered to be, a formal assessment of the sword- fish fishery. Such an assessment would normally in- clude additional information and analysis, including age-structured population models and numerous sen- sitivity analyses. Also, the abundance index used here was developed solely to serve an example, and is not believed to be an accurate reflection of abun- dance over time. The North Atlantic swordfish fishery enjoyed a high catch rate in 1962 and 1963, but it declined in the late 1960s (Fig. 1). The U.S. and Canadian por- tions of the fishery were sharply reduced in the early 1970s because of FDA regulations prohibiting inter- state transportation or importation offish with mer- cury concentrations exceeding the allowable level of 0.5 ppm (Hoey et al., 1989). In 1978, the FDA in- creased the allowable mercury content to 1 ppm, and since then, the catch has increased, but the CPUE has slowly declined (Fig. IB). For the years 1971- 73, early years of the FDA regulations, data are avail- able on catch but not on fishing effort. Results from the two analyses were similar, but they illustrate how tuning can influence the results of a production model. In each analysis, the model fits the effort data reasonably well (Fig. 2); however, because the hypothetical abundance index does not match the observed CPUE well (Fig. IB), the fit in the last years of the tuned model was a compromise Prager: A nonequilibrium surplus-production model 381 between matching the observed effort (Fig. 2B) and matching the index (Fig. 2C). The tuned analysis gave lower estimates of MSY, fMSY, and a less opti- mistic impression of the current level of the stock, as represented by the ratio B1992/BMSY (Table 2). It also estimated that the recent fishing mortality rate, as represented by the ratio FW91/FMSY, was somewhat higher. Estimated median biases from each analysis were small. In the base analysis, no management bench- mark was estimated to have a bias exceeding 1.5% Table 1 Data used in two production model analyses loosely based on swordfish, Xiphias gladius, in the North Atlantic Ocean. Yield and standardized fishing-ef- fort-rate data are from Hoey et al. (1993) with mi- nor revisions. Hypothetical abundance index data are the mean of ages 3 through 5+ in numbers from Scott et al. ( 1992). The index was constructed solely for illustrative purposes, and is designated "hypo- thetical" because it probably is not a good indicator of total-stock biomass. Year Yield (t) Fishing effort rate (106hooks/yr) Hypothetical abundance index 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1076 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 5,342 10,189 11,258 8,652 9,338 9,084 9,137 9,138 9,425 5,198 4,727 6,001 6,301 8,776 6,587 6,352 11,797 11,859 13,527 11,126 12,832 14,423 12,516 14,255 18,278 19,959 19,137 17,008 15,594 13,212 6.45 8.54 24.45 25.30 31.39 28.90 40.11 43.23 38 47 19.22 22.97 21.17 18.14 20.40 40.13 35.44 34.85 40.73 55.10 49.44 59.55 80.75 98.91 97.08 90.46 85.86 69.86 1.000 0.816 DISS 0.483 0.526 0.411 0.377 0.368 0.359 0.352 of the corresponding uncorrected estimate (Table 2). Estimated median biases for the tuned analysis were only slightly higher; with only the estimated bias in /"MSY slightly exceeding 2%. Estimates of median bias in individual model parameters (such as r and K) were slightly higher yet, but only for B1 was bias estimated as higher than about 2.5%. Approximate 80% nonparametric confidence inter- vals computed by Equations 14 and 15 were derived from the bootstrap. These were computed for the in- dividual model parameters, management bench- marks, indicators of stock position, and for each year's relative stock size estimate (Table 2; Fig. 3). A unitless nonparametric measure of the precision of estimates was constructed by dividing the bias-cor- rected 50% confidence interval (interquartile range; not shown here) by the corresponding median bias- corrected estimate. The resulting statistic, the rela- tive interquartile range (RIR) is a nonparametric analog of the coefficient of variation. The RIR was of similar magnitude for both models, and was small- est in MSY and fMSY, the benchmarks that do not depend on q. Estimates of the quantities that depend on q, and that thus involve absolute scaling, exhib- ited relative IQ ranges of about 50% (Table 2). Estimates of relative biomass (Bz scaled to BMSY) and fishing mortality rate (Fz scaled to FMSY) were also similar from the two models (Fig. 3). They show a declining biomass through 1991, with an increase expected thereafter (at the projected harvest rate of 12,000 t/yr, which is less than the MSY estimates). As expected, the precision of estimates during the projection period was less than during the period for which data were available. In summary, this example demonstrates that much more than MSY can be estimated from a production model. Biomass trajectories can be computed easily, as can estimated confidence intervals derived through the bootstrap. If an independent index of abundance is available, the model can be tuned to that index. Another useful feature is that projections can be used to estimate the probable effects of quo- tas or other management measures. Discussion The modeling framework described here is based on the logistic population model. The history of this model was summarized by Kingsland (1982), who pointed out that the model originated in the work of Verhulst (1845) and Robertson (1923), was popular- ized by Pearl and Reed (1920), and was also studied by Lotka (1924). The model was introduced to fish- ery science by Graham (1935) and Schaefer (1954, 382 Fishery Bulletin 92(2). 1994 LU 3 Q_ O 1957). In modeling fish populations, one could just as easily use the exponential yield model of Fox ( 1970) or a model of more flex- ible shape, such as that of Pella and Tomlinson (1969) or its alternative formu- lation by Fletcher ( 1982). (Fletcher's formu- lation lacks the estimated exponent that has been found to complicate estimation [Ricker 1975, p. 326].) Unfortunately, those formu- lations can not supply an analytical formu- lation similar to Equations 6 and 8, which means that numerical integration would have to be used, as in the GENPROD com- puter program of Pella and Tomlinson (1969). Another alternative would be to use a discrete-time model, rather than the con- tinuous-time model presented here. Such models are simpler mathematically, but usually entail assumptions that the growth, recruitment, and catching seasons are brief. The logistic model was used here because it is a simple case, not because using other models would be impractical or inferior. For what types of stocks are the models presented here appropriate? Research is lacking to answer this question definitively, but general comments are possible. One group of fishes for which production mod- els seem to work well is the tropical tunas. These species are characterized by rela- tively fast growth, relatively constant re- cruitment, and reduced annual seasonality in the life processes. Density dependence in growth has been demonstrated in a related species, Scomber japonicus (Prager and MacCall, 1988); such plasticity in growth would allow the compensation inherent in a production model to be expressed in a way beyond recruitment variability. For modeling fish stocks with more seasonality in growth, reproduc- tion, and harvest, a discrete-time production model might prove superior to the continuous-time model presented here. In many fish stocks, recruitment is extremely vari- able. Ordinary production models may not work well when applied to stocks with large recruitment fluc- tuations that are unrelated to population size, espe- cially when the catch-effort series is short. If recruit- ment fluctuations can be linked to external factors (such as variation in rainfall or sea-surface tempera- ture), a production model incorporating these factors might work well (Freon, 1986). It would be simple to modify the logistic model to incorporate an environ- mental factor, perhaps as an influence on r on an annual basis. 120 100 - 20000 1 5000 P 5 25000 1400 1200 1000 800 - 600 400 - 200 Figure 1 Data used to fit production model examples loosely based on swordfish, Xiph ias gladius , in the North Atlantic Ocean. (A) Stan- dardized effort rate (•) and total yield (o). (B) CPUE trajectory (•) computed from data in (A), and index of abundance (°) used to tune the second example. The index, which was used for illustrative purposes only, is not a good measure of total-stock abundance. Other extensions Many other extensions to the production model have been published. An incomplete list includes these: Fox (1975, 1977) presented a logistic production model with mixing of two stocks; Deriso (1980) and Hilborn (1990) demonstrated different methods of fitting production models to age-structured popula- tions (but see also Ludwigand Walters, 1985); Freon (1986) introduced environmental variables into a production model that used the equilibrium assump- tion; Laloe (1989) and Die et al. ( 1990) incorporated fished area into production models; Polovina ( 1989) demonstrated a system of production models in which some parameters are common among models; and Hoenig and Warren (1990) demonstrated Bayesian Prager: A nonequilibrium surplus-production model 383 and empirical Bayes methods for fitting production models. Most of the extensions described by these investigators could be combined with techniques pre- sented here (e.g. tuning, bootstrapping), as required for a particular analysis. 1960 1965 1970 1975 1980 1985 1990 120 I ' ' ' ' ' ' ' ' ' ] 1960 1965 1970 1975 1980 1985 1990 1960 1965 1970 1975 1980 1985 1990 Year Figure 2 Goodness-of-fit of two production model analyses loosely based on swordfish, Xiphias gladius, in the North Atlantic Ocean. These analyses are illustra- tive and are not intended as an assessment of sword- fish. Model 2 differs from Model 1 in being tuned to a hypothetical index of abundance. (A) Observed (o) and estimated ( ) fishing effort rate from Model 1(B) Observed (o) and estimated ( ) effort rate from Model 2. (C) Observed (o) and estimated abun- dance-index from Model 2. Autocorrelation Because catch and effort data are usually autocor- related, the residuals from fitting — whether comput- ed in yield or effort — may also be autocorrelated. A matter of statistical concern is whether a method of fitting that takes the autocorrelation into account (such as one based on time-series analysis sensu Box and Jenkins [1976]) might be more appropriate. Some results relevant to this question were obtained by Ludwig et al. ( 1988) in a study that used two differ- ent objective functions to fit production models to simulated data. The first was a total-least-squares objective function, which did not take autocorrelation into account; the second, an approximate-likelihood objective function, which did. Ludwig et al. (1988) found that the two methods produced very similar es- timates; the authors concluded that the added complex- ity of the approximate-likelihood method was probably not warranted. In addition, the approximate-likelihood method frequently failed to converge from poor start- ing values. This does not mean that autocorrelation should be ignored in all fisheries modeling; however, it was not a major concern in the study cited. Process error The model presented here assumes that the produc- tion of biomass is a deterministic function of the cur- rent biomass; stochasticity occurs only in the obser- vation of catch or effort or in the relation of fishing effort to fishing mortality rate (if effort is being esti- mated from catch). In reality, production is undoubt- edly stochastic to some degree. In recognition of this, fisheries models that explicitly incorporate process error have been developed (e.g. Ludwig et al., 1988; Sullivan, 1992). Because process errors are propa- gated forward in time, it would seem that time- series fisheries models (e.g. production models), should include corrections for process errors, so that the system can be modeled as correctly as possible. Despite the undeniable logic of including process error in fisheries models, there are also negative as- pects, and the practical merit of including process error in fisheries applications remains a topic for research. The theory of models including process er- ror was largely developed in process control (Kalman, 1960), a field in which large data sets are common. Including both observation error and process error in a model generally entails either estimating a large number of nuisance parameters (the process errors), making strong assumptions about the form or value of the process error component, or both. In some cases, the need to estimate additional parameters can make it difficult or impossible to estimate pa- rameters of interest, such as MSY, without additional 384 Fishery Bulletin 92|2). 1994 Table 2 Results of two bootstrapped production model analyses loosely based on swordfish, Xiphias gladius, in the North Atlantic Ocean. The base model used only yield and standardized effort data. The tuned model also used a hypo- thetical index of abundance (Table 1). Each conventional parameter estimate is designated 6, the corresponding bias-corrected estimate is designated 6BC. Nonparametric bias-corrected 80% confidence intervals are derived from the bootstrap; as with most fishery analyses, these are conditional on correct model structure and probably under- estimate true uncertainty (see text). The relative interquartile (IQ) range, a unitless measure of precision, is the 50% confidence interval divided by the median bias-corrected estimate. All results are rounded to three significant digits. Base model Tuned model Quantity estimated 80% 80% Relative lower CL upper CL IQ range 80% 80% Relative lower CL upper CL IQ range Management benchmarks MSY MWSY 'MSY °MSY "W^^MSY *199l'' MSY 13,800 0.257 72.6 53,800 0.932 1.03 13,700 0.259 71.1 53,100 0.929 1.03 Directly estimated parameters r 0.514 0.517 K 108,000 106,000 q 0.00354 0.00363 11,800 0.161 61.7 37,400 0.755 0 751) 0.323 74,800 0.00236 15,100 0.393 82.2 79,700 1.17 1.32 0.785 159,000 0.00541 11.8% 45.3% 14.5% 40.7%- 21.8% 28.3%' 45.3% 40.7% 43.3% 13,400 0.264 68.7 50,900 0.829 1.18 0.528 102,000 0.00384 13,400 0.269 68.3 50,000 0.820 1.18 0.537 100,000 0.00393 11,700 0.169 0.590 33,600 0.650 0.892 14,900 0.432 0.781 71,900 1.01 1.53 0.337 0.865 67,200 144,000 0.00260 0.00612 11.7% 50.9% 14.1% 39.6% 23.6% 29.3% 50.9% 39.6% 45.7%- information or assumptions. (For an example, see Conser et al., 1992, and Prager, 1993). This would not be a serious objection if estimates made by mod- els without process error were known to be severely flawed, but to my knowledge the fisheries literature includes no comprehensive comparisons of equiva- lent models with and without process error. The work by Ludwig et al. (1988) does shed some light on this question, as their simulations and mod- els included both types of error. The authors found that when observation error was ignored (its vari- ance assumed to be zero) during parameter estima- tion, the resulting estimates were biased and resulted in an average loss in harvest value of at least 20%. In contrast, when the relative variance of the pro- cess error component was assumed to be half of its correct value, a substantially smaller loss in harvest value resulted. Unfortunately, Ludwig et al. (1988) did not present results for estimation under the as- sumption that process error was zero. Further re- search into estimation methods for systems with both process error and observation error would allow fish- ery scientists and managers to better balance com- plexity and accuracy in population models. Precision of estimates Production models tend to estimate some quantities much more precisely than others. Hilborn and Wal- ters ( 1992) discuss this phenomenon at some length; the comments here reflect my own experiences. For most stocks, the main biological reference points (MSY, /"MSY) are estimated relatively precisely. How- ever, absolute levels of stock biomass BT and fishing mortality rate Fx are usually estimated much less precisely. This occurs because very few data sets con- tain sufficient information to estimate q well. (The example illustrates this point well — Table 2.) By di- viding biomass and fishing-mortality estimates by estimates of the corresponding biological reference points, the effects of imprecision in estimating q can be removed. The relative levels thus obtained are useful measures in their own right: the relative level of biomass BT I SMSY describes whether a population is above or below the level at which MSY can be ob- tained, and the relative level of fishing mortality rate F / FMaY suggests whether an increase or decrease in fishing effort might provide a higher sustainable yield. When two or more catchability coefficients are es- timated, ratios of catchability coefficients are typi- cally estimated more precisely than the individual values of q. Thus it is possible to compare two differ- ent gears without being able to estimate very pre- cisely the catchability of either one. If a parameter- ization involving K and r is used in fitting, the esti- mates of these quantities are usually quite impre- Prager: A nonequilibrium surplus-production model 385 <0 E o S I i i i i I i i i i I i i i i I i i i i I i i i i I i i i i I i i i 0.5 i | i i i i | i i i i | i i i i | i i i i | i i i i | i i i i | 1965 1970 1975 1980 1985 1990 1995 1965 1970 1975 1980 1985 1990 1995 I ' ' ■ i Li l l i I ' ' ' ■ I ' ' ' ' I ' ' ' ' I ' ' ' ' I I ■ ' ' ■ 1 ' '''I' ' ■ ■ I ' ' i i I cr 1965 1970 1975 1980 1985 1990 1995 Year I I I 1 1 I I I I I I I I I I I I I 1 1 | ! 1965 1970 1975 1980 1985 1990 1995 Year Figure 3 Estimated trajectories of relative biomass and relative fishing mortality rate from two production model analyses (including proposed yields from 1992 through 1996) loosely based on swordfish, Xiphias gladius, in the North Atlantic Ocean. These analyses are illustrative and are not intended as an assessment of sword- fish. "Relative biomass" is the stock biomass divided by BMSy the biomass at which maximum sustainable yield (MSY) can be obtained; "relative fishing mortality rate" is the fishing mortality rate (F) divided by the rate (-P^jgy) that yields MSY when the stock is at SMSY- Production models estimate these relative quantities more precisely than the corresponding absolute quantities. Trajectories ( ) are shown with approximate 80% confidence intervals (— ) from the bootstrap. Model 2 differs from Model 1 in being tuned to a hypotheti- cal index of abundance. Panels (A) and (B), estimates from Model 1; (C) and (D), estimates from Model 2. cise, but because they are correlated, the correspond- ing estimates of MSY and optimum effort can none- theless be quite precise. The estimate of Bp the starting biomass in the first year, is usually quite imprecise even when normal- ized to SMSY. It is also my impression that it can be biased for some data sets, although this does not sig- nificantly affect relative biomass estimates beyond the first few years. I would therefore not recommend using a production model to draw any inferences about the population biomass during the first few (perhaps 2 to 4) years, unless auxiliary information is available. Such information might comprise a bio- mass index (for tuning) or knowledge to support us- ing an assumption of the type B1 = sK, where s is a proportionality constant known a priori. Punt, 1990, provides an example. This indeterminacy in produc- tion modeling is similar to the inability of sequential population (age-structured) analyses to say much about population dynamics in the most recent years unless auxiliary information is used. In practice, it does not seem to degrade the estimates of MSY and optimum effort when a reasonably long time series is used. Validity of bias corrections and confidence intervals Bootstrap confidence intervals are approximations, and bias-correction methods can at times worsen the 386 Fishery Bulletin 92(2). 1994 approximation. DiCiccio and Tibshirani (1987) dem- onstrate an example in which "the BC and BCQ meth- ods seem to pull the percentile interval in the wrong direction and hence the coverage gets worse." (The BCQ method, due to Efron [1987], incorporates a sec- ond-order correction to the BC method.) In that ex- ample, bias correction for the point estimate would also have made it worse. The example presented by DiCiccio and Tibshirani (1987) (estimating the variance of a cor- relation coefficient, true value 0.9, from a data set of 15 observations) seems rather extreme, but it does serve to emphasize that model results, including estimated bias corrections, must not be accepted blindly. Confidence intervals estimated by bootstrap meth- ods entail fewer assumptions than those made by parametric methods, but most likely are still opti- mistic. In a study of an econometrics equation (in- cluding a lagged term) that was fit by generalized least squares with an estimated covariance matrix, Freedman and Peters ( 1984) found the bootstrap es- timates of standard error far superior to those made with asymptotic assumptions. The bootstrap esti- mates were 20% to 30% too low, but estimates from asymptotic formulas were too low by factors of al- most three. One reason for underestimation by the bootstrap was that, due to the effect of fitting, the residuals used for resampling were smaller than the true values of the disturbance term (Freedman and Peters, 1984). A suggested correction is given by Stine, 1990, p. 338. There are other reasons why estimated confidence intervals for fisheries models are likely to be opti- mistic. The time frame encompassed by the data used to fit fisheries models is usually short and does not encompass the full range of environmental variation that can add unexplained variation to observed data. As the time series becomes longer, the random ef- fects of environmental variation tend to become more extreme, making earlier confidence intervals appear overly optimistic (Steele and Henderson, 1984). An- other cause of optimistic confidence intervals is the use of preliminary models (e.g. ANOVA) to construct abundance indices; such models tend to filter the indices and thus reduce apparent variance. There may also be systematic errors in the data (from, e.g. gradual changes in q or gradual or sudden changes in the proportion of the catch reported); these can bias the results, but the confidence intervals include only the effects of variability, not bias from model misspecification. Schenker ( 1985) stated that "boot- strap confidence intervals should be used with cau- tion in complex problems." It is probably appropri- ate to consider estimated confidence intervals from fisheries population models to be, in general, mini- mum estimates. Is there life after death? The concept of maximum sustainable yield was given its epitaph about 15 years ago in a critical review by Larkin (1977). Notwithstanding the title of his pa- per, Larkin's main target was not the concept of MSY itself, but what he called the "religion" of applying MSY dogmatically to every stock. Undoubtedly, one must recognize that MSY is not an immutable quan- tity, and that model results should not be used dog- matically. However, compensation in population dy- namics does give rise to some form of maximum sus- tainable yield. Whether MSY is estimable from the data available for a given stock, and whether it is a useful concept given the stock's dynamics, are rea- sonable questions that, even if answered in the nega- tive, do not invalidate the concept of MSY. In a response to Larkin's (1977) paper, Barber (1988) pointed out that MSY, far from being dead, was still in widespread use. Barber cited the utility of MSY as a formal management objective; its sim- plicity and ability to be understood by the fishing industry, administrators, and managers; and the grounding of the MSY concept in basic ecological theory. He concluded by repeating Holt's ( 1981) sug- gestion that MSY be considered part of a multi-fac- eted management scheme. Shortly following Larkin's ( 1977 ) paper, Sissenwine ( 1978) discussed several shortcomings of MSY as the basis for optimum yield (OY), the "legally mandated immediate objective of marine fisheries management in the coastal waters of the United States beyond the territorial sea of the individual states." In this section, I address those items not discussed earlier. Sissenwine pointed out that it is difficult to estimate q, and that q may vary with population size. This difficulty might be overcome, to some degree, by the methods described earlier for estimating changes in q. More importantly, this problem is not unique to production models. The common use of CPUE series to tune age-structured models also requires strong assumptions about q. Indeed, because an age-struc- tured model generally provides little information about a cohort before it has been substantially fished, its estimate of population biomass in a year close to the present may be more influenced by random varia- tions in q than would a similar estimate from a pro- duction model. Sissenwine (1978) made a number of criticisms of production models fit by equilibrium assumption. The methods described here do not use the equilibrium assumption and are not subject to those problems. Once the assumption is dropped, one is much less likely to get a good, but spurious, fit, when modeling a population whose dynamics are not approximated Prager: A nonequilibrium surplus-production model 387 by the model. This is an excellent reason (but not the only one) to avoid the equilibrium assumption. A final important point raised by Sissenwine ( 1978) is that, because the world is stochastic, one is truly more interested in maximum average yield (MAY) than MSY. Several studies (Doubleday, 1976; May et al., 1978; Sissenwine, 1978) have shown that in gen- eral MAY < MSY; thus harvesting MSY indefinitely would lead to stock collapse. This result does not make production models less useful, but does em- phasize the necessity to use their results in the con- text of other knowledge about the stock and as part of an evolving view of stock dynamics. Fishery as- sessment and management are dynamic processes that must adapt to changing conditions and new knowledge. It is inconceivable that we will ever know enough about any wild stock to establish a manage- ment regime that could be effective into the indefi- nite future. The failure of MSY to be such a regime is no failure at all. Notes added in proof 1 I have recently been made aware of several pro- duction-model applications that were circulated in the Collected Papers of the International Commis- sion on Southeast Atlantic Fisheries (ICSEAF). Per- tinent documents include those by Butterworth et al., 1986; Andrew et al., 1989; and Punt, 1989. 2 Anyone attempting to implement the methods described here should be aware that Equation 6, when solved for F, can be double-valued. Acknowledgments I thank A. Anganuzzi, J. Bence, R, Deriso, A. Fonten- eau, K. Hiramatsu, W. Lenarz, A. MacCall, R. Methot, C. Porch, J. Powers, A. Punt, W. Richards, G. Scott, P. Tomlinson, J. Zweifel, and two anonymous refer- ees for comments on the manuscript or assistance with the techniques described. In addition, sugges- tions from J. Hoenig and V. Restrepo were particu- larly helpful in improving many parts of the work. Portions of this work result from research sponsored by NOAA Office of Sea Grant, U.S. Department of Commerce, under federal Grant No. NA90AA-D- SG045 to the Virginia Graduate Marine Science Con- sortium and the Virginia Sea Grant College Program. Literature cited Barber, W. E. 1988. Maximum sustainable yield lives on. N.Am. J. Fish. Manage. 8:153-157. Box, G. E. P., and G. M. Jenkins. 1976. Time series analysis: forecasting and control, revised ed. Holden-Day, San Francisco. Conser, R., J. M. Porter, and J. J. Hoey. 1992. Casting the Shepherd stock-production model in a statistical framework suitable for swordfish stock assessment and management advice. Int. Comm. Cons. Atl. Tunas (Madrid), Coll. Vol. Sci. Pap. 39:593-599. Deriso, R. B. 1980. Harvesting strategies and parameter estima- tion for an age-structured model. Can. J. Fish. Aquat. Sci. 37:268-282. Deriso, R. B., T. J. Quinn II, and P. R. Neal. 1985. Catch-age analysis with auxiliary informa- tion. Can. J. Fish. Aquat. Sci. 42:815-824. DiCiccio, T., and R. Tibshirani. 1987. Bootstrap confidence intervals and bootstrap approximations. J. Am. Stat. Assn. 82:163-170. Die, D. J., V. R. Restrepo, and W. W. Fox Jr. 1990. Equilibrium production models that incorporate fished area. Trans. Am. Fish. Soc. 119:445-454. Doubleday, W. C. 1976. Environmental fluctuations and fisheries management. Int. Comm. Northw. Atl. Fish., Sel. Pap. 1:141-150. Efron, B. 1982. The jackknife, the bootstrap, and other resampling plans. Society for Industrial and Ap- plied Mathematics, Philadelphia, 92 p. 1987. Better bootstrap confidence intervals. J.Am. Stat. Assn. 82:171-200. Efron, B., and G. Gong. 1983. A leisurely look at the bootstrap, the jackknife, and cross-validation. Am. Statistician 37:36-48. Efron, B., and R. Tibshirani. 1986. Bootstrap methods for standard errors, con- fidence intervals, and other measures of statisti- cal accuracy. Stat. Sci. 1:54-75. Fletcher, R. I. 1982. A class of nonlinear productivity equations from fishery science and a new formulation. Math. Biosci. 61:279-293. Fox, W. W., Jr. 1970. An exponential yield model for optimizing exploited fish populations. Trans. Am. Fish. Soc. 99:80-88. 1975. An overview of production modeling. Int. Comm. Cons. Atl. Tunas (Madrid), Coll. Vol. Sci. Pap. 3:142-156. 1977. Some effects of stock mixing on management decisions. Int. Comm. Whal. Rep. 27:277-279. Freedman, D. A., and S. C. Peters. 1984. Bootstrapping a regression equation: some empirical results. J. Am. Stat. Assn. 79:97-106. Freon, P. 1986. Introduction of environmental variables into global production models. Proc. Int. Symp. Long Term Changes Mar. Fish Pop., Vigo (Spain), p. 481- 528. 388 Fishery Bulletin 92(2), 1994 Gavaris, S. 1988. An adaptive framework for the estimation of population size. Can. Atl. Fish. Sci. Adv. Comm., Res. Doc. 88/29. Graham, M. 1935. Modern theory of exploiting a fishery, and application to North Sea trawling. J. Cons. Int. Explor. Mer 10:264-274. Gulland, J. A. 1975. The stability of fish stocks. J. Cons. Int. Explor. Mer 37:199-204. Hilborn, R. 1990. Estimating the parameters of full age-struc- tured models from catch and abundance data. In L.-L. Low (ed.), Proceedings of the symposium on application of stock assessment techniques to gadids. Int. N. Pac. Fish. Comm. Bull. 50:207-213. Hilborn, R., and C. J. Walters. 1992. Quantitative fisheries stock assessment: choice, dynamics, and uncertainty. Chapman and Hall, NY, 570 p. Hinkley, D. V. 1978. Improving the jackknife with special reference to correlation estimation. Biometrika 65:13-21. Hoenig, J. M., and W. G. Warren. 1990. Bayesian and related approaches to fitting stock production models. ICES CM-1990/D:13, 21 p. Hoey, J. J., R. J. Conser, and A. R. Bertolino. 1989. The western north Atlantic swordfish. In W. J. Chandler (ed.), Audubon wildlife report 1989/ 1990, p. 456-477. Academic Press, San Diego. Hoey, J. J., J. Mejuto, J. Porter, and Y. Uozumi. 1993. A standardized biomass index of abundance for North Atlantic swordfish. Int. Comm. Cons. Atl. Tunas (Madrid), Coll. Vol. Sci. Pap. 40:344-352. Holt, S. J. 1981. Maximum sustainable yield and its application to whaling. In J. Gordon Clark (ed.), Mammals in the seas, vol. 3, p.22-55. FAO Fish. Ser. 5. Kalman, R. E. 1960. A new approach to linear filtering and pre- diction problems. J. Basic Eng. 82:34-45. Kennedy, P. 1979. A guide to econometrics. MIT Press, Cam- bridge, MA, 175 p. Kingsland, S. 1982. The refractory model: the logistic curve and the history of population ecology. Q. Rev. Biol. 57:29-52. Kizner, Z. I. 1991. Bootstrap estimation of the confidence inter- vals of stock and TAC assessments with the use of dynamic surplus production models. NAFO Sci. Coun. Studies 16: 149-152. Laloe, F. 1989. Un modele global avec quantite de biomasse inaccessible dependand de la surface de peche. Application aux donnees de la peche d'albacore [Thunnus albacares) de l'Atlantique Est. Aquat. Liv. Res. 2:231-239. Larkin, P. A. 1977. An epitaph for the concept of maximum sus- tained yield. Trans. Am. Fish. Soc. 106:1-11. Lotka, A. J. 1924. Elements of physical biology. Reprinted 1956 as "Elements of mathematical biology" by Dover Press, NY, 465 p. Ludwig, I)., and C. J. Walters. 1985. Are age-structured models appropriate for catch- effort data? Can. J. Fish. Aquat. Sci. 42:1066-1072. Ludwig, D., C. J. Walters, and J. Cooke. 1988. Comparison of two models and two estima- tion methods for catch and effort data. Nat. Res. Modeling 2:457-498. MacCall, A. D. 1976. Density dependence of catchability coefficient in the California Pacific sardine, Sardinops sagax caerulea, purse seine fishery. Calif. Coop. Ocean. Fish. Invest. Rep. 18:136-148 Marquardt, D. W. 1963. An algorithm for least squares estimation of nonlinear parameters. J. Soc. Ind. Appl. Math. 2:431-441. May, R. M., J. R. Beddington, J. W. Horwood, and J. G. Shepherd. 1978. Exploiting natural populations in an uncer- tain world. Math. Biosci. 42:219-252 Methot, R. M. 1989. Synthetic estimates of historical abundance and mortality for northern anchovy. Am. Fish. Soc. Symp. 6:66-82. 1990. Synthesis model: an adaptable framework for analysis of diverse stock assessment data. In L.- L. Low (ed.), Proceedings of the symposium on ap- plication of stock assessment techniques to gadids, p. 259-277. ICNAF Bull. 50. Nelder, J. A., and R. Mead. 1965. A simplex method for function minimization. Comp. J. 7:308-313. Noreen, E. W. 1989. Computer-intensive methods for testing hypoth- eses: an introduction. John Wiley & Sons, NY, 229 p. Paloheimo, J. E., and L. M. Dickie. 1964. Abundance and fishing success. Rapp. P.-v. Reun. Cons. int. Explor. Mer 155:152-163. Parrack, M. L. 1986. A method of analyzing catches and abundance indices from a fishery. Int. Comm. Cons. Atl. Tu- nas (Madrid), Coll. Vol. Sci. Pap. 24:209-221. Pearl, R., and L. J. Reed. 1920. On the rate of growth of the population of the United States since 1790 and its mathematical rep- resentation. Proc. Nat. Acad. Sci. U.S.A. 6:275-288. Pella, J. J. 1967. A study of methods to estimate the Schaefer model parameters with special reference to the yellowfin tuna fishery in the eastern tropical Pa- cific ocean. Ph.D. diss., Univ. of Washington, Se- attle. 156 p. Prager: A nonequilibrium surplus-production model 389 Pel la, J. J. and P. K. Tomlinson. 1969. A generalized stock production model. Bull. Inter-Am. Trop. Tuna Comm. 13:419-496. Peterman, R. M., and G. Steer. 1981. Relation between sport-fishing catchability coefficients and salmon abundance. Trans. Am. Fish. Soc. 110:585-593. Polovina, J. J. 1989. A system of simultaneous dynamic production and forecast models for multispecies or multiarea applications. Can. J. Fish. Aquat. Sci. 46:961-963. Prager, M. H. 1993. A nonequilibrium production model of sword- fish: data reanalysis and possible further direc- tions. Int. Comm. Cons. Atl. Tunas (Madrid), Coll. Vol. Sci. Pap. 40:433-437. Prager, M. H., and A. D. MacCall. 1988. Revised estimates of historical spawning bio- mass of the Pacific mackerel, Scomber japonicus. Calif. Coop. Ocean. Fish. Invest. Rep. 29:81-90. Press, W. H., B. P. Flannery, S. A. Teukolsky, and W. T. Vetterling. 1986. Numerical recipes: the art of scientific com- puting. Cambridge Univ. Press, Cambridge, 818 p. Punt, A. E. 1990. Is Bj = K an appropriate assumption when ap- plying an observation error production-model esti- mator to catch-effort data? S. Afr. J. Mar. Sci. 9:249-259. Ricker, W. E. 1975. Computation and interpretation of biological statistics of fish populations. Bull. Fish. Res. Board Can. 191, 382 p. Rivard, D., and L. J. Bledsoe. 1978. Parameter estimation for the Pella-Tomlinson stock production model under nonequilibrium conditions. Fish. Bull. 76:523-534. Robertson, T. B. 1923. The chemical basis of growth and senescence. J. B. Lippincott, Philadelphia. Schaefer, M. B. 1954. Some aspects of the dynamics of populations important to the management of the commercial marine fisheries. Bull. Inter-Am. Trop. Tuna Comm. l(2):27-56. 1957. A study of the dynamics of the fishery for yel- lowfin tuna in the eastern tropical Pacific Ocean. Bull. Inter-Am. Trop. Tuna Comm. 2:247-268. Schenker, N. 1985. Qualms about bootstrap confidence intervals. J. Am. Stat. Assn. 80:360-361. Scbnute, J. 1977. Improved estimates from the Schaefer pro- duction model: theoretical considerations. J. Fish. Res. Board Can. 34:583-603. 1989. The influence of statistical error on stock as- sessment: illustrations from Schaefer 's model. In R. J. Beamish and G. A. McFarlane (eds.), Effects of ocean variability on recruitment and an evalua- tion of parameters used in stock assessment mod- els, p. 101-109. Can. Spec. Publ. Fish. Aquat. Sci. 108. Scott, G. P., V. R. Restrepo, and A. Bertolino. 1992. Standardized catch rates for swordfish (Xiphias gladius) from the U.S. longline fleet through 1990. Int. Comm. Cons. Atl. Tunas (Madrid), Coll. Vol. Sci. Pap. 39:554-571. Seber, G. A. F., and C. J. Wild. 1989. Nonlinear regression. John Wiley & Sons, NY, 768 p. Sissenwine, M. P. 1978. Is MSY an adequate foundation for optimum yield? Fisheries 3(61:22-42. Steele, J. H., and E. W. Henderson. 1984. Modeling long-term fluctuations in fish stocks. Science 224:985-987. Stine, R. 1990. An introduction to bootstrap methods: ex- amples and ideas. In J. Fox and J. S. Long (eds. ), Modern methods of data analysis, p. 325- 373. Sage Pubis., Newbury Park, CA, 446 p. Sullivan, P. J. 1992. A Kalman filter approach to catch-at-length analysis. Biometrics 48:237-257. Thiel, H. 1971. Principles of econometrics. John Wiley & Sons, NY, 736 p. Verhulst, P.-F. 1845. Recherches mathematiques sur la loi d'accroissement de la population. Mem. Acad. Roy. Belg. 18:1-38. Winters, G. H., and J. P. Wheeler. 1985. Interaction between stock area, stock abun- dance, and catchability coefficient. Can. J. Fish. Aquat. Sci. 42:989-998. AbStrclCt. Stock enhancement with hatchery-reared juvenile queen conch, Strombus gigas L., has been sug- gested as a means to rehabilitate over- fished populations in Florida and the Caribbean region. A 15-month field ex- periment was conducted in the Baha- mas to compare the survival, growth, morphology, and behavior of hatchery- reared and wild juvenile conch (85-120 mm shell length). Two experimental sites were established. Site CI con- tained a resident conch population whereas few conch occurred naturally at site C2. Survival was higher for wild conch than for hatchery-reared conch. After 7 months, 28% of the original wild conch were recovered compared with only 99c of the hatchery-reared conch. Thin shells, short spines, and low burial frequency in hatchery-reared conch may have caused them to be more vul- nerable to predators. In a tethering ex- periment, about twice as many hatch- ery conch were killed as wild conch, but the difference was not significant inside enclosures. Survivorship was higher at the site with resident juveniles, prob- ably because of density-dependent pro- tection from predation. After a period of high mortality in free-ranging conch during the first two months, tag recov- ery curves for both stock types reached a plateau. Also, near the end of the study, shell characteristics of wild and hatchery conch were identical as was survivorship. Analysis of movement patterns indicated that both stock types moved toward the natural population center. Although survivorship was higher at the site with resident conch, growth rates for both stock types were often lower at this site. Algal foods may have been more abundant at the site without conch because of lower grazing pressure. Athough highest mean daily growth occurred at 1.0 conchm 2, growth rates of conch enclosed at 0.5, 1.0, and 2.5 individuals m"2 were not significantly different in most cases. Growth rates were higher for wild conch than for hatchery conch. In sum- mer, free-ranging and tethered wild conch grew twice as fast as hatchery- reared conch. Success in rehabilitating depleted queen conch populations will require the release of high quality, hatchery-reared juveniles in large num- bers in appropriate habitats. Experimental outplanting of juvenile queen conch, Strombus gigas: comparison of wild and hatchery-reared stocks Allan W. Stoner Megan Davis Caribbean Marine Research Center, 805 E. 46th Place Vera Beach, Fl_ 32963 Manuscript accepted 4 October 1993 Fishery Bulletin 92:390-411 ( 1994) 390 Widespread depletion of natural fishery stocks, particularly in in- shore coastal and estuarine habi- tats, has resulted in increasing in- terest in enhancement and restora- tion of wild populations through releases of hatchery-reared indi- viduals. Among the molluscs, rela- tively sedentary bivalves such as oysters, Crassostrea spp. (Burrell et al., 1981; Goodwin1), clams, Merce- naria mercenaria (Flagg and Malouf, 1983), mussels, Mytilus edulis (Dare and Edwards, 1976), and giant clams, Tridacna spp. (Heslinga et al., 1984; Heslinga and Watson, 1985 ) have been restocked most successfully. Experimental re- seeding of scallops, Argopecten and Patinopecten spp. (Saito, 1984; Ao- yama, 1989; Tettelbach and Wen- czel, 1991) and abalone, Haliotis spp. (Kojima, 1981; Saito, 1984; Uki, 1984; Searcy and Salas, 1985; Tegner and Butler,1985; Tong et al., 1987; Ebert, 1989; Emmett and Jamieson, 1989) also show promise. Queen conch, Strombus gigas, is one of the most important fishery species in the Caribbean region (Brownell and Stevely, 1981; Berg and Olsen, 1989), with an estimated annual value of 30 million U.S. dol- lars (Appeldoorn and Rodriquez, 1993). Heavy fishing for queen conch in shallow water habitats has resulted in a decline of this species throughout most of its biogeogra- phic range (Appeldoorn et al.,1987; Appeldoorn and Rodriguez, 1993), and the U.S. fishery has been closed completely since 1986 (Berg and Olsen, 1989). Mariculture has been suggested as a way to rehabilitate queen conch populations (Berg, 1976; Siddall, 1984a; Davis et al., 1987), and research efforts during the past two decades have made it possible to culture large numbers of juvenile conch for stock enhance- ment (Brownell, 1977; Ballantine and Appeldoorn, 1983; Hensen, 1983; Laughlin and Weil, 1983; Cruz, 1986; Davis et al., 1987; Heyman et al., 1989; Creswell,1993; Davis, 1993). Unfortunately, field outplants of hatchery-reared stock have met with little success because of very high mortality (Appeldoorn and Ballantine, 1983; Laughlin and Weil, 1983; Appeldoorn, 1985; Marshall et al., 1993; Dalton, 1993). Also, little is known about the relative viability of wild and hatchery-reared conch. This study uses a large-scale outplant experiment, together with enclosure and tether experiments, to compare the survival, growth, morphology, and behavior of hatch- ery-reared and wild juvenile conch released into a well-studied nursery Goodwin, W.F. 1981 Use of seed oysters to supplement oyster production in southern North Carolina. Report, North Carolina Division of Marine Fisheries, NCDMF - Project - 2/314-R, 109 p. Stoner and Davis: Outplanting queen conch. Strombus gigas 391 site (Stoner and Sandt, 1991, 1992). This research provides further insight into the potential for using cultured conch in a stock enhancement program, and elucidates possible limitations. Methods and materials Site description During the 15-month period from March 1990 to May 1991, all field outplant, enclosure, and tether experi- ments were carried out at two different sites desig- nated CI and C2 (Fig. 1). The study sites, each a 100-m square area delineated by buoys, were located 0.8 km west of Children's Bay Cay and 5.0 km south- east of the Caribbean Marine Research Center field station on Lee Stocking Island, in the southern Exuma Cays, Bahamas (lat. 23°44.5'N, long. 76°04.4'W) (Fig. 2). A shallow sand bank is to the southwest. The two sites are in a homogeneous seagrass meadow of Thalassia testudinum with mod- erate shoot density (500-700 shootsrrr2) in 3.2 m depth. Tidal currents run northwest (flood) and southeast (ebb) at velocities to 50 cm-sec-1 with a tidal range of approximately 1.0 m. Clear water from the Exuma Sound flows over the sites on flood tides. resulting in high underwater visibility that facili- tated field experiments and recovery of tagged conch. Site CI was established within a well-studied queen conch nursery area that has carried as many as 500,000 individuals in densities between 0.5 and 2.0 conch-nr2 since at least 1984 (Wicklund et al., 1991; Stoner et al. 1993, unpubl. data). Site CI has been the location of numerous investigations on conch mass migration (Stoner et al., 1988; Stoner 1989a), distribution (Stoner and Waite, 1990), and diet (Stoner and Waite, 1991). Site C2 was approximately 0.3 km to the southeast of site CI and had very few juvenile conch (< 0.05 conchm-2). In 1988, small-scale transplants in en- closures showed that young conch survived and grew at nearly identical rates at sites C 1 and C2 despite the absence of wild conch at the latter (Stoner and Sandt, 1992). This suggested that certain unpopulated areas of the extensive seagrass meadows in the Exuma Cays could support outplanted conch stocks. Density estimates were obtained by counting the conch (tagged and untagged) in as many as 20 hap- hazardly placed circles of 4-m radius at each site at five different times during the experiment. The pur- pose of these estimates was to assess the natural population of conch prior to the transplant, to exam- EXPERIMENTS 1990 Mar | April | May | June | July | Auq | Sepl I Oct I Nov | Dec 1991 Jan I Feb I Mar I April I May Free-Range GPI GPII GP III 1 ^ ~^~ N/ \fc -n * * * • * * * • • S1 • 1 4 S2 1* .* S3 • • Enclosures GPI GPII 7 * y — ^9 ° Ml M2 ° GP GPII ^/ M1 M2 Telhers GP I GP II , ^ te i M1 M2 M1 M2 Burial Density o = Macrophyte Analysis s = Survey GP = Growth Penod M = Cumulative Mortality Figure 1 Calendar of free-ranging, enclosure, and tether experiments; showing duration of experiments, data collection points, and time periods for growth and mortality of queen conch, Strombus gigas. The total study period extended from March 1990 to May 1991. See text for descriptions of the measurements and experiments. 392 Fishery Bulletin 92(2), 1994 ine the density of conch at the end of the experiment, and to observe movements by the population. On each date up to 100 conch were measured for shell length (apex to siphonal canal). Seawater temperature for the study site was re- corded with a Ryan Temp Mentor placed on the bot- tom between sites CI and C2. Temperature was re- corded (+ 0.2°C) every 30 minutes, and seven-day mean temperatures were calculated for plotting (Fig. 3). Experimental animals Approximately 6,000 wild and 6,000 hatchery-reared conch were used in the experiments described be- low. Wild conch were collected from the Children's D> — i — 76°06- '76°0A ~\ ■ *""/ Bahamas Florida » v).. ~23^^M w N - Ci h^r^aH^^ * -20« • ~ ~ __■ f .2- ^•,76*^e*i" \ EXUMA \ SOUND 23°46' O— Flood >**▼ CAY Northwest / Northeast Ebb 'ac,a / ^Center fs. Southeast / Southwest \/ C2^ SANDBAR / 100m WINDSOCK CAY | GREAT BAHAMA BANK Bay Cay nursery site (CI); all were between 85 and 120 mm shell length (SL). Hatchery-reared conch were purchased from Tradewind Industries, Ltd. (Caicos Conch Farm) in Providenciales, Turks, and Caicos Islands. These conch originated from 12 egg masses collected near Providenciales in the summer of 1988. The larvae were fed with Caicos Isochrysis and postlarvae with flocculated Chaetoceras gracilis and blended Enteromorpha sp. (Davis et al., 1992). Between December 1988 and March 1989, 50-mm juveniles were transplanted to a protected nursery habitat near the hatchery (Davis and Dalton, 1991). Between 26 and 29 March 1990, 6,000 hatchery- reared conch comparable in size to the 1+ year class conch native to Children's Bay Cay nursery (85-120 mm SLKTable 1) were collected from the grow-out area and held in two 8 x 8 m holding pens. On the morning of 30 March the conch were loaded into 32 large burlap sacks wrapped in plastic bags and transported via cargo plane to Lee Stocking Island. The conch were kept cool and moist during the 7-hour period out of the water. Upon arrival, conch were immediately taken to either site CI or C2. The plastic was removed and the burlap bags were placed on the bottom of the respective sites. On 31 March they were released into two tem- porary pens ( 10 m2) already constructed at each of the study sites. All hatchery- reared conch were tagged and measured over the next 10 days. Wild conch were tagged and placed in temporary pens during a 10-day period prior to the ar- rival of the hatchery-reared conch. All conch were marked with orange spaghetti tags ( Floy Manufacturing Co. ) tied around the spire, and total shell length was measured to the nearest mil- limeter (±1 mm) with calipers. Tags were both letter coded and numbered so that conch type and release site could be identified immediately in the field. CHILDREN'S BAY CAY 23°44- Figure 2 Map of Lee Stocking Island and Children's Bay Cay in the southern Exuma Cays, Bahamas. Study sites CI and C2 are shown as squares, the dotted line around CI represents the approximate boundaries of the natural queen conch, Strombus gigas, population in July 1990. The lower insert shows the geographic zones surveyed during each tag recovery. Free-ranging experiment Hatchery-reared and wild tagged conch were haphazardly released throughout each of the two 100 m x 100 m experi- mental sites (CI and C2) to examine survivorship, growth, morphology, and behavior of free-ranging juveniles be- tween 1 April 1990 and 20 February 1991. The size ranges for hatchery and Stoner and Davis: Outplanting queen conch. Strombus gigas 393 wild conch released at site CI were 80-117 mm SL (mean=102, SD=8, n=2,552) and 85-117 mm SL (mean=100, SD=7, re=2,543), respectively. For site C2, the size ranges for hatchery and wild conch were 80- 117 mm SL (mean=101, SD=8, rc=2,540) and 83-117 mm SL (mean=101, SD=6, n =2,490), respectively. CL E F M A M 1991 Figure 3 Bottom-water temperature recorded between sites CI and C2, during the study period. Seven- day averages for recordings made every 30 min- utes are plotted. Table 1 Density of juvenile queen conch (Strombus gigas) in two field sites (CI and C2) prior to the release of tagged conch (2 Mar. 1990), during the release ex- periment, and at the end of the experiment (21 Feb. 1991). Density was measured by counting conch, including tagged conch, in up to 20 randomly-se- lected 4-m diameter circles ( 50 m2 ) at each site. Mean shell length for up to 100 individuals was measured for conch found in each survey. Values are mean ± SD(n). Date Number of plots Density (mm) Shell length (no./m2) Site CI 2 Mar 90 2 May 90 16 Jul 90 27 Nov 90 21 Feb 91 Site C2 2 Mar 90 2 May 90 16 Jul 90 20 20 20 5 6 20 20 20 0.26 + 0.12 1.27 + 1.01 0.66 + 0.31 0.22 + 0.12 0.19 + 0.19 0.01 + 0.01 0.03 + 0.03 0.62 + 0.74 105+ 14(100) n a 111 + 12(100) 113+ 16(55) 129 + 13 (56) 99+ 14(9) n/a 119 + 10(100) Tag recovery Tag recapture surveys were conducted in June, September, and November 1990, to provide a relative index of survivorship over time and space. Searches encompassed the transplant sites, the zone between the sites, and adjacent seagrass and sand habitats. The search area was sectioned arbitrarily (see insert, Fig. 2), and several divers using snorkel or SCUBA gear drifted repeatedly side by side over the area using the flood tidal current for transport. Conch location was recorded by section, and searches continued until no additional conch were found. Af- ter each survey, all recaptured conch were returned to their original transplant sites (CI or C2). Because very few hatchery-reared conch remained alive in February 1991, collections were made for shell growth and morphology, but the complete survey was not conducted. Tag recovery does not measure absolute survivorship because of potential emigration from the study site and possible inefficiency in finding tagged animals; however, the search effort was intensive, and clear water (usually >10 m horizontal visibility) facilitated the efficiency of the searches. In two blind tests 87 and 92% of 200 uniquely tagged conch were recovered by using standard search procedure (unpubl. data). The surveys were conducted over periods from 7 to 20 days depending upon the num- ber of divers available. Because of known limitations, tag recovery data were used as a relative indicator of survivorship in the two stock types and two study sites. Growth Seasonal growth rates (mmday-1) were determined for three periods: summer (April to Sep- tember 1990), fall (September to November 1990), and winter (November 1990 to February 1991) by comparing shell lengths of individual, tagged conch at the beginning and end of the survey periods. Shell morphology At the beginning (1 April 1990) and end of the experiment (20 February 1991), rep- resentative samples of at least 30 hatchery-reared and 30 wild conch (collected alive) were measured for shell length and width, and shell and tissue weight. Maximum shell width was the distance be- tween the last complete spine formed near the shell aperture and the spine on the opposite side of the shell. Total weight of the shell and soft tissue (live weight) was recorded to the nearest 0.01 g. After freezing and subsequent thawing, the soft tissue of the animal was extracted, lightly blotted, and weighed. Weights of the clean, air-dried shells were also recorded. Behavior Nine times during the study, observations on burial behavior were made for the first 30 hatch- 394 Fishery Bulletin 92|2). 1994 ery-reared and 30 wild conch (tagged individuals) at each outplant site. Burial frequency was quantified as the percentage of conch that had at least part of the shell buried in the sediment, detritus, or algae. Counts for the two sites were pooled for each of the conch types. General observations on locomotory ac- tivity were also recorded. Data from the tag recovery surveys provided in- formation on the movements of free-ranging conch. During each tag recovery the type (wild or hatchery- reared), initial site of transplant (CI or C2), and number of tagged conch found in different regions around the initial release sites (Fig. 2) were recorded. The total number of conch found in each survey was used to calculate the percentage of hatchery or wild conch from site CI or C2 in each area surveyed. Enclosure experiments Experiment I Enclosure experiments were designed to determine the significance of density-dependent growth and survival of hatchery-reared and wild conch in identical habitats. The first 3-month experi- ment was conducted from 7 April to 9 July 1990. At each site (CI and C2) 12 circular pens (30 cm high, 20 m2) without covers were constructed of vinyl coated wire mesh (2.5 x 5.0 cm). Prior to the experi- ment (23 February-12 March) three haphazardly placed 25 x 25 cm quadrants per cage were sampled for Thalassia testudium components to ensure habi- tat similarity among the pens, both within and be- tween stations. In each quadrant, seagrass shoot den- sity was estimated, and all above-ground parts were collected into 3-mm mesh nylon bags. Living blades and detritus were separated in the laboratory, dried at 80°C and weighed. Detritus measurements were made again at the end of the experiment (9-13 July 1990) to test for potential depletion of this impor- tant food source. At each site tagged hatchery and wild conch were placed in pens at three different densities, in two random blocks. Stocking densities, spanning the high range of natural densitites in the wild, were 0.5, 1.0, and 2.5 conch-m-2 (10, 20, 49 conch-pen-1). The size ranges for hatchery-reared and wild conch were 90- 109 mm SL (mean=100, SD=3) and 92-115 mm SL (mean=102, SD=3), respectively. Before stocking the pens with experimental conch, all visible epibenthic predators such as tulip snails, Fasciolaria tulipa, apple murex, Murex pomum, and the giant hermit crab, Petrochirus diogenes, and sea urchins Trip- neustes esculentus were removed. Every two weeks throughout the experiment, dead conch were replaced to ensure constant density; replacements were not used in growth and survivorship measurements. Cumulative mortality was calculated by subtract- ing the number of live conch remaining from the ini- tial loading number. Mortality was examined statis- tically at the midpoint (day 37) and at the end (day 93) of the experiment. Shell length was measured at the start, near the middle (day 37), and at the end (day 93) of the experiment, and growth rates were calculated for the two periods. Experiment II A second 3-month enclosure experi- ment was conducted at sites CI and C2 to compare survival and growth of hatchery-reared and wild conch in the winter (29 November 1990-21 Febru- ary 1991). Enclosures built for experiment I were reused in this experiment after having been clear of conch since July 1990. Four enclosures at each site were stocked with 10 hatchery-reared and 10 wild conch (1.0 conch-m-2) gathered from the surround- ing free-ranging populations. This density was cho- sen because highest mean growth rates frequently occurred at this density in enclosure experiment I. The initial size of the hatchery-reared conch ranged from 104-130 mm SL (mean=118, SD=5) and the wild conch ranged from 109-134 mm SL(mean=122, SD=5). Dead conch were replaced with similar sized free- ranging conch every two weeks. As in the first ex- periment cumulative mortality and growth rates were determined only for the original stock, not the replacements. Mortality was calculated five times throughout the experiment, and analyzed statisti- cally at the midpoint (day 35) and end (day 84) of the experiment. Growth rates were calculated for two growth periods, 29 November 1990 to 3 January 1991 and 3 January to 21 February 1991. Tether experiments Experiment I The first three-month tethering ex- periment was conducted during the summer ( 12 April to 11 July 1990) at sites CI and C2 to examine survivorship, tag effects, and growth rates. The size ranges for hatchery-reared and wild conch were 82- 116 mm SL (mean=100, SD=9) and 89-115 mm SL (mean=101, SD=6), respectively. Each conch was se- cured to a 0.5-m long stainless steel welding rod by a 1 m length of 20-lb test monofilament line that was attached to the shell spire with a clear nylon cable tie. The tether rods were marked with uniquely num- bered tags and pushed 40 cm into the substratum approximately 2 m apart. Conch were tethered in four rows of 20 individuals. Each row contained 10 hatchery-reared conch and 10 wild conch in an al- ternating pattern. For each type of conch, the shell of every second individual was tagged to determine potential tagging effects on conch mortality in the free-ranging experiment. Stoner and Davis: Outplanting queen conch, Strombus gigas 395 Cumulative mortality was examined by using the same procedure as enclosure experiment I. Mortal- ity was calculated three times throughout the experi- ment, and analyzed statistically at the midpoint (day 45) and end (day 90) of the experiment. Growth rates were calculated for two growth periods: 12 April to 27 May and 27 May to 11 July 1990. Experiment II The second 3-month tethering experi- ment was conducted during the winter ( 7 February- 3 May 1991) at site CI. Too few of the original hatch- ery-reared conch remained alive to set up the experi- ment at the second site. Hatchery-reared conch ranged from 100 tol38 mm SL (mean=116, SD=8), and wild conch were 111-133 mm SL (mean=124, SD=5). Tethers were set up as in experiment I with four replicated rows of 20 individuals ( 10 hatchery- reared and 10 wild conch), except the conch them- selves were not tagged. The conch were checked for mortality three times during the experiment and analyzed statistically at days 42 and 84. Because cable ties were secured behind long api- cal spines, escape from tether apparatus would be possible only in the event of failure in the cable tie, monofilament line, or connections. Failure appears to be unlikely because nearly all kills observed in this study were found as empty shells attached to the tether apparatus or as crushed shells within 1 m of the original location. Data Analysis Analysis of variance (ANOVA), following the guide- lines of Day and Quinn (1989), was used extensively in the interpretation of growth and mortality data. The statistical procedures started with full model ANOVA that included all independent effects. When interactions were significant, one- or two-way ANOVAs were performed to examine the effects of site and stock type, the variables most critical in this study. For brevity, non-significant interaction terms in multiple-way ANOVAs are not addressed in the text but are reported in tables. Mortality data were normally examined at the mid-point of individual experiments and at the end. Cochran's test was used to test for homogeneity of variances. Log and arcsine transformations of data were used in some cases to remove heteroscedasticity; these are noted in the text. Where repeated measure- ments were made within one experimental enclosure (i.e. growth rates determined for conch in one pen), mean growth rates in the enclosures were used as replicates rather than individual measurements to eliminate pseudoreplication (Hurlburt, 1984). Analy- sis of covariance (ANCOVA) was used to test for dif- ferences in morphological characteristics ( shell weight, shell diameter, and tissue weight) between hatchery- reared and wild conch. Shell length was the covariate. After release, tagged, free-ranging conch dispersed from the initial 1-ha study sites. Chi-square analy- sis was used to compare dispersion of the two stock types, where the distribution of tagged wild conch was used for the expected frequency in different sur- vey zones. Results Conditions at the outplant sites During the 15-month study period, bottom-water temperature ranged from 24°C in February 1990 to 30°C in late September 1991, then declined rapidly and remained between 24 and 25°C until early May, when temperature rose to 27°C (Fig. 3). Density estimates made 1 month prior to the be- ginning of the free-ranging experiment (2 March 1990) showed that the density of conch at site CI was 16 times higher than at site C2 (Table 1). Be- tween March and May, conch density at site CI in- creased to over 1.2 conchm-2, owing to immigration of the natural population. Transplanted conch from the free-ranging experiment made up 7-15% of the conch in the density estimates; however, on 2 May 1990, transplanted conch accounted for 88% of esti- mated density. In July, there were nearly equal den- sities of conch at sites CI and C2, but in November 1990 and February 1991, densities were close to the original values first observed in March 1990. This may be due to directional changes in movement of conch (towards the northeast) during the winter, which took them away from the transplant sites CI and C2 (see Behavior). As expected, shell length measurements taken during the density surveys show an increase in length over time for wild conch tagged at CI (from mean=105 ± 14 SD in March 1990 to mean=129 ± 13 SD in Feb- ruary 1991 ((Table 1). This represents an overall growth rate of 0.07 mmday"1, similar to that mea- sured in free-ranging tagged conch. No growth rate was calculated for C2, because density surveys yielded low numbers; however, the mean sizes appear to be com- parable to those measured at site CI (Table 1). Handling and tag effects Transporting hatchery-reared conch appeared to have little adverse effect on their subsequent survivorship in the field. Conch were left out of wa- ter for 7 hours, and all remained alive during the 7 days after transport while they were tagged and placed in enclosures for Experiment I. 396 Fishery Bulletin 92(2), 1994 Table 2 Mortality of tagged and untagged queen conch (Strombus gigas) on tethers at the two experimen- tal sites. Ten tagged and ten untagged conch were tethered in each of four replicate blocks at each site. Values are mean percent mortality ± SD (number of dead conch). Mortality Site Tagged Untagged CI ("J 52.5 ± 12.6(21) 40.0 ± 21.6(16) 45.0+ 12.9(18) 52.5 ± 9.6(21) In the first tethering experiment percent mortalities (arcsine-transformed) did not differ among any of the tag and site treatments (Table 2) ( ANOVA, F3 ;2=0.722, P =0.558 for CI; F3 ,2=0.679, P =0.581 for C2). Free-ranging experiment Tag recovery Tag recapture rates for free-ranging juvenile conch were related to both stock type and location (Table 3, Fig. 4). Exhaustive searches in and beyond the study area recovered all visible live conch, and there is no reason to believe that hatchery conch were seen and collected by the divers less often than wild conch. In fact, wild conch had burial rates higher than hatchery conch (see Behavior); therefore, the reverse bias is more likely. In November 1990, approxi- mately 7.5 months after initial re- lease, 206 of hatchery-reared conch were recovered from site CI and 248 from C2, an overall recap- ture of 9% of the original release (Fig. 4). Recoveries of wild conch from sites CI and C2 numbered 542 and 820 conch, respectively, an overall recapture rate of 28%. The highest proportion of loss occurred during the first two months (April and May 1990). After May, recov- ery curves for both hatchery and wild conch leveled off at both sites. Tag recapture was consistently higher for wild conch released at site C2 (34% at experiment end) than for those released at C 1 ( 22% ), despite the presence of large num- 100 • • Wild - C1 (2543) » — ' Wild - C2 (2490) ■o Hotchery - C1 (2552) a Hatchery - C2 (2540) C* cu > o o CD Months Figure 4 Recovery rates for free-ranging hatchery- reared and wild queen conch, Strombus gigas, transplanted to sites CI and C2 in April 1990. Values are percentages of the original conch (April releases) found in June, September, and November surveys. In parentheses are the original numbers of tagged conch. Table 3 Tag recovery summaries for hatchery-reared and wild queen conch (Strombus gigas) released in two study sites near Children's Bay Cay, Ba- hamas, in 1990. Adjustments to the original numbers of conch released in April account for tagged conch taken from the free-ranging study to be used in enclo- sure and tether experiments; these were subtracted from the original number. Site CI C2 Stock type Hatchery Wild Hatchery Wild April to June (64 days) Original Number Released in April Recovered Live % Recovered Live June to September ( 100 days) Adjusted Number Released Recovered Live % Recovered Live September to November (66 days) Adjusted Number Released Recovered Live r7r Recovered Live 2552 2543 2540 2490 480 837 586 1270 18.8 32.9 23.1 51.0 2517 2467 2472 2451 277 582 328 1035 11.0 23.6 13.3 42.2 2502 2452 2457 2436 206 542 248 820 8.2 22.1 10.1 33.7 Stoner and Davis: Outplanting queen conch. Strombus gigas 397 bers of untagged, wild conch at the CI area. Hatch- ery stocks were recovered in about equal proportions at the two sites. Growth Free-ranging wild conch had higher growth rates (log-transformed) than hatchery-reared conch during all three seasons examined (Table 4). During summer, the difference was approximately two times (Fig. 5), but the rates began to converge in the fall. Growth rates were highest during summer and fall, and lowest during winter, following patterns of wa- ter temperature (Fig. 3). Conch grew significantly faster at site C2 than at CI during both summer and winter; site differences were not significant in the fall (Table 4, Fig. 5). Morphology At the beginning of the free-ranging experiment, shells of hatchery-reared conch were significantly lighter than those of wild conch from the Children's Bay Cay nursery site (slopes were homogeneous, P=0.833, P=0.365; ANCOVA: P=92.62, P< 0.001) (Fig. 6A). Lower shell weight in hatchery- reared conch is a function of either thinner shells or differences in shell form compared to wild conch. Regressions of shell width with shell length (Fig. 6B) showed that the spines were, in fact, longer in wild conch than in hatchery-reared stock (slopes were ho- mogeneous, F=l. 76, P= 0.190; ANCOVA: F=73.99,P< 0.001). Regressions of tissue wet weight with shell length show no significant difference in tissue weight Table 4 Results of two-way AN OVAs for growth rates in free- ranging queen conch (Strombus gigas). "Site" refers to the two experimental sites CI and C2. "Stock type" refers to hatchery-reared versus wild conch. Source di MS Period I (April to September 1990—64 days) Site x stock type 1 <0.001 1.205 0.273 Site 1 0.003 22.157 0.001 Stock type 1 0.045 322.407 <0.001 Error 396 <0.001 Period II (September to November 1990—100 days) Site x stock type 1 <0.001 1.332 0.249 Site 1 <0.001 0.418 0.518 Stock type 1 0.004 16.967 <0.001 Error 396 <0.001 Period III (November 1990 to February 1991—66 days) Site x stock type 1 Site 1 Stock type 1 Error 115 <0.001 0.714 0.400 0.001 7.541 0.007 0.001 8.697 0.004 <0.001 between wild and hatchery-reared conch (Fig. 6C), (slopes were homogeneous, F=1.76, P=0.190; ANCOVA: F=3.24, P=0.077). Measurements made on shells of hatchery-reared and wild conch at the end of the experiment in Feb- ruary 1991 show that lines for shell weight and width had converged (Fig. 7, A and B). Shell weights of hatchery-reared conch were still lighter than those of wild conch (slopes were homogeneous, P=0.189, P=0.665; ANCOVA: P=7.44, P=0.008) (Fig. 7A), but the lines were closer than in April 1990 (Fig. 6A). Stock type did not affect the relationship between shell length and shell width in February (Fig. 7B) (slopes were homogeneous, F=2.01, P=0.160; ANCOVA: P=0.957, P=0.33 1 ). Hatchery-reared conch Penod 1 Apnl • Sept 1 990 1 100 100 K M 1 _^H 1 hatchery wild Site C1 hatchery wild Site C2 03 E E 0.10 Penod II Sept- Nov 1990 E E OlOf | a> too ^^_^^ ioo I "I Ml hatchery wild Sited hatchery wild Site C2 Penod III Nov 1990- Feb 1991 hatchery wild Site C1 hatchery wild Site C2 Figure 5 Comparison of growth rates of free-ranging hatch- ery-reared and wild queen conch, Strombus gigas, at sites CI and C2. Growth periods I— III represent summer, fall, and winter, respectively. Values are mean ± SD, with the number of conch measured shown inside the vertical bars. 398 Fishery Bulletin 92(2), 1994 had heavier tissue wet weight than wild conch (slopes were homogeneous, F=0.163, P=0.688; ANCOVA: F=7.12, P= 0.010) (Fig. 7C). This can be explained by examining the ratio between tissue and shell weight. At the beginning of the experiment these ratios for hatchery and wild conch were 0.34 ± 0.04 and 0.22 ± 0.03 (mean ± SD), respectively. This indi- cates that hatchery conch had lighter shells and heavier soft tissue than wild conch. At the end of the experiment ratios for hatchery and wild conch were 0.30 ± 0.04 and 0.25 ± 0.04 (mean ± SD), respectively. The lower ratio for hatchery conch indicates that both the tissue and shell weight were increasing. Hatchery-reared conch that survived 11 months in the field either developed morphological charac- teristics of wild conch, or the survivors had such char- acteristics at release. Because there was little over- lap in regressions of shell width versus shell length at the beginning of the experiment, change in shape is the most plausible explanation for characteristics measured in hatchery-reared conch at the end of the experiment. Presence of short spines on pretran- 120 en s I) in S tn P 20 120 • Wild (n - 30) O Hatchery (n - 30) 0 ° y c o 2 r-6 o SO 90 100 110 120 Shell Length (mm) Figure 6 Shell weights (A), shell widths (B), and tissue weights (C) of hatchery- reared and wild queen conch, Strombus gigas, shown as a function of shell length. Measurements were taken for conch of each stock type collected at the beginning of the free-ranging experiment ( 1 April 1990). • Wild (n-S3) Weight (g) O Hatchery (n-2e) -^-"■'' A . Jti&^ = ISO -^jfBfi? O to 100 110 120 130 140 150 130 E %-• ^ 110 • Wild (n-SJ) O Hatchery (n - 25) # # ^^ ° .8° 8^4^^ U 90 to 100 110 120 130 140 150 • Wild (n-30) # O Hatchery (n - 28) -^ -£ 70 C ->^^ Tissue W( o ^cfi-^r • °J^ref« * yyt> 100 110 120 130 140 150 Shell Length (mm) Figure 7 Shell weights (A), shell widths (B), and tissue weights (C) of hatchery-reared and wild queen conch, Strombus gigas, shown as a function of shell length. Measurements were taken for conch of each stock type collected at the termination of the free-ranging experiment (21 Feb. 1991). Stoner and Davis: Outplanting queen conch, Strombus gigas 399 splant portions of the shells followed by long spines on the outer (newer) portions of the last shell whorl support the hypothesis of changing shell shape. Al- though it is normal for spine length to increase proportionally with shell length in queen conch, the posttransplant increase in hatchery-reared stock was extreme and obviously disproportionate in most shells. Behavior On all nine dates when burial was exam- ined for free-ranging animals, a higher percentage of wild conch were buried than of hatchery-reared individuals (Fig. 8). Pairwise ANOVA of burial fre- quency on arcsine-transformed data from dates as blocks showed that the difference in burial rates be- tween stock types was significant (F t JS=8.51, P=0.011). Hatchery and wild conch showed nearly parallel patterns of burial frequency over time. How- ever, plots of burial frequency should not be inter- preted as seasonal trends, because juvenile conch appear to demonstrate tidal periodicity in locomo- tory activity (pers. observ). Although the patterns were not quantified, it was frequently noted during field observations that hatchery conch were more active than wild conch. While hatchery-reared indi- viduals were almost always moving, wild conch were frequently found nestled motionless beneath algae or detritus. Given the relatively small area of the two outplant sites (1 ha each) tagged conch often dispersed rela- tively far from their original release sites (Fig. 9, A and B). For example, in June 1990 only 15% of the recovered hatchery-reared conch released at site CI CL o Cl o m Apr Moy Jun Jul Aug Sep Oct Nov Dec Jen Feb Date Figure 8 Percentage of queen conch, Strombus gigas, buried during each observation for hatchery-reared and wild conch. Values are based upon observations on 30 haphazardly chosen conch of each stock type on each date. were found in that zone and 35% were found north- west of Cl. Tagged conch tended to move to the north- east and northwest between April and June 1990. By June, hatchery-reared and wild conch initially released at site Cl were widely dispersed and differ- entially distributed (%2= 18.01, df=4, P=0.05). Hatch- ery-reared conch released at C2 tended to disperse more widely than wild conch (i.e., from the south- west sandbar to the north zones); the difference was significant 0.05). At the end of the experiment dry weight of seagrass detritus did not differ among the cages (Table 7)(P23 ^=0.900, P=0.598), and there were no differences in individual cages between the be- ginning and end of the experiment (P7 742=0.090, P=0.764). There is no evidence, therefore, that detri- tus was depleted even at the high density of 2.5 conchirr2. When comparing shoot density between the beginning and end of the experiment, there was no difference (F, /J9=2.95, P=0.088), but biomass of living seagrass did differ between the dates (F, ^2=37.01, P<0.001), probably related to blade growth in the spring season. Experiment II Mortality At the termination (day 85) of enclosure experiment II (29 November 1990-21 February 1991) mortality was obviously higher at site C2 than at CI for both hatch- ery and wild conch (Fig. 12), simi- lar to the results of enclosure ex- periment I (Fig. 10). Midway through the experiment there was no significant mortality difference between site CI and C2; however, the difference was significant by the end of the experiment (Table 8); mortality was higher at site C2. Similar to enclosure experiment I, there were no differences in mor- tality between hatchery and wild conch (Table 8). Growth Growth rates were low in enclosure experiment II (0.01- 0.06 mmd"1) (Fig. 13), paralleling the trend observed in free-ranging conch (Fig. 4) and associated with low winter temperatures (Fig. 3). There were significant site x stock type interactions for period I (Fj /2=5.949, P=0.031) and period II (Fj ;2=5.004, P=0.045) because of differences in growth rate between hatchery and wild conch at site CI (period I: Ft 6=6.48, P=0.044; period II: F; 6=9.747, P=0.021); but not at site C2 (period I: F2 6=0.008, P=0.932; period II: F16=0.2Q7, P=0.665). At site CI wild conch grew approximately twice as fast as hatchery conch. Tether experiments Experiment I Mortality The first tether experiment, conducted from 11 April to 11 July 1990, confirmed that the difference in tag recovery rate between hatchery- reared and wild conch was related to predation (Fig. 14). Hatchery conch were killed at a frequency ap- proximately twice that of wild conch for day 45 and day 88 (Table 9, Fig. 14). Site effects were not sig- nificant at either midpoint or end of the experiment (Table 9, Fig. 14). Growth Growth rates in both hatchery and wild conch on tethers were higher at site C2 than at site CI by the end of the study period ( 11 July 1990) (Fig. 15). This difference also occurred in free-ranging conch (Fig. 5) and enclosure experiment I (Fig. 11). During period I (April and May 1990) there was a significant site x type interaction (Table 10) because wild conch grew faster at site C2 than CI (P/64=49.28, P<0.001), and hatchery conch grew at 402 Fishery Bulletin 92(2), 1994 o o »*- o D O 30 20 □ > o ■ ► • D t> O ■ ► • Wild 0.5/rr>2 Wild 1.0/m2 Wild 2.5/m2 Hotchory 0.5/m2 Hatchery 1.0/m2 Hatchery 2.5/m2 Site C1 ^^ ^-~~m 10 01 ss— • 20 40 60 Figure 1 0 Cumulative mortality curves for enclosure experi- ment I. Hatchery-reared and wild queen conch, Strombus gigas, were held at three different densi- ties at sites CI and C2. Initial numbers of conch in the enclosures were 10, 20 and 49, yielding densi- ties of 0.5, 1.0, and 2.5 conchm'2, respectively. the same rate at both sites (F; ^=0.105, P= 0.747). Wild conch grew significantly faster than hatchery conch at site CI (FJ55=14.54, P<0.001) and C2 CF; 52=69.06, P<0.001) (Fig. 15). In growth period II (June and July 1990) wild conch grew faster than hatchery conch at both sites (Table 10, Fig. 15). Experiment II A second tether experiment con- ducted at site CI from 7 February to 3 May 1991, using wild conch and the few remaining hatchery- reared conch from the free-ranging experiment, re- sulted in mortality curves (Fig. 16) different from the first tether experiment (Fig. 14). Mortality rates did not differ between stock types at either 42 (F;6=0.871, P=0.387) or 84 days (F, e<0.001, P=1.000). Mortalities were identical (65%) at the 84- day termination of the experiment (Fig. 16). Comparison of experiments Mortality rates in hatchery-reared conch were higher Table 6 Results of multi-way ANOVAs for growth rates of queen conch (Strombus gigas) in enclosure experi- ment I. Sources are the same as described in Table 5. Source df MS F P 3-way ANOVA for Period I (7 April to 14 May 1990—37 days) Site x type x density 2 <0.001 0.481 0.630 Site x stock type 1 <0.001 0.893 0.363 Site x density 2 <0.001 2.456 0.128 Stock type x density 2 <0.001 2.284 0.144 Site 1 0.001 28.210 <0.001 Stock type 1 <0.001 16.511 0.002 Density 2 <0.001 4.649 0.032 Error 12 <0.001 3-way ANOVA for Period II (14 May to 9 July 1990—56 days) Site x type x density 2 <0.001 3.546 0.062 Site x stock type 1 0.001 8.722 0.012 Site x density 2 <0.001 1.281 0.313 Stock type x density 2 <0.001 4.261 0.040 Site 1 <0.001 0.778 0.395 Stock type 1 0.0022 3.171 <0.001 Density 2 <0.001 4.787 0.030 Error 12 <0.001 2-way ANOVA for Period II (Site CI Stock type x density 2 0.001 9.513 0.014 Stock type 1 0.002 36.836 0.001 Density 2 <0.001 4.619 0.061 Error 6 <0.001 2-way ANOVA for Period II (Site C2 Stock type x density 2 <0.001 0.015 0.985 Stock type 1 <0.001 1.465 0.272 Density 2 <0.001 1.935 0.225 Error 6 <0.001 than or equal to those of wild conch in all experi- ments and at both study sites (Table 11). Equivalent mortality rates were found in enclosures and in tether experiment II run at the end of the study period. Growth rates were higher in wild conch than in hatchery-reared conch except during the second 5-week period of enclosure experiment I at site C2 and in enclosure experiment II at site C2, when growth rates were equivalent. Site differences in mortality rates were relatively consistent across experiments and stock types (Table 11). Mortality was always lower for both wild and hatchery-reared conch at site CI than at C2, except Stoner and Davis: Outplanting queen conch, Strombus gigas 403 for equivalent mortality rates mea- sured in tether experiment I. Most ex- periments showed that growth rates were lower at site CI than at C2 with certain exceptions ( Table 11 ). A signifi- cantly higher growth rate was found at CI in wild conch during the second enclosure experiment, and equivalent growth rates were found in hatchery- reared conch in the same experiment. Growth rate did not differ between sites during fall in free-ranging conch. 0-20 O 6 0.15 0.10 005 Discussion Importance of seed stock quality Stock enhancement and rehabilitation depend upon the ability of fisheries managers to place viable seed animals in optimal habitats at appropriate times (Stoner, in press). Hatcheries in the Turks and Caicos Islands, Belize, Mexico, and Florida are now produc- ing juvenile queen conch with the ex- pectation that hatchery-reared conch will be seeded into local waters for res- toration of depleted resources. Re- leases of hatchery-reared conch in sev- eral small-scale pilot programs have been relatively unsuccessful in terms of conch survival (Appeldoorn and Ballantine, 1983; Appeldoorn, 1984; Iversen et al., 1986; Coulston et al., 1987; Rathier, 1987; Davis et al., 1992), but it is un- known whether low survivorship was related to char- acteristics of the habitat or the outplanted conch. The only published field comparison of wild and hatch- ery-reared queen conch (Marshall et al.2) showed that hatchery-reared conch may be more vulnerable to predation than are wild conch. Additionally, Jory and Iversen ( 1988) found that hatchery-reared conch may have shells with lower breaking strengths than those of wild conch. The present study shows that poten- tial differences in physiology, behavior, morphology, and survival must all be considered. Differences in growth rate between wild and hatch- ery-reared conch at Children's Bay Cay study sites are surprising given that the hatchery conch had been in a field grow-out enclosure with natural sub- strata and food for 6 months. Several explanations 0.20 0.15 0.10 >> 0.05 TO Period I April - May 1 990 _ 020 015 0.10 1 0.05 E 0.5 1.0 2.5 0.5 1.0 2.5 —' hatchery wild £ Site C1 CO cc 0.5 1.0 2.5 0.5 1.0 2.5 hatchery wild Site C2 Penod II May -July 1990 * 0.20 - 0.15 0.10 0.05 0.5 10 2.5 hatchery 05 1.0 2.5 wild 0.5 1.0 2.5 hatchery 0.5 1.0 2.5 wild Site C1 Site C2 Density (#conch/m2) Figure 1 1 Growth rate comparisons for enclosure experiment I. Hatchery-reared and wild queen conch, Strombus gigas, were held at three different densities at sites CI and C2. Values are mean ± SD for average growth rates in two replicate enclosures. 2 Marshall, L. S., Jr., C. Cox, and R. N. Lipcius. 1992. Survival of wild and hatchery-reared juvenile queen conch in natural habitats. Unpubl. manuscr. Table 7 Seagrass components in 24 cages at the beginning and end of enclosure experiment I. Thalassia biom- ass included all above-ground live blades. Thalassia detritus included senescent and decomposing seagrass blades retained in a 3 mm mesh bag. Val- ues are mean ( ± SD). Seagrass components March 1990 July 1990 Thalassia shoot density (shoots-m-2) Thalassia biomass (g dry wt-m~2) Thalassia detritus (g dry wtm-2) 674.1 (±92.2) 77.2 (±18.5) 649.3 (±79.8) 107.2 (±37.5) 317.9 (±131.7) 324.32 (±25.4) 404 Fishery Bulletin 92 [2), 1994 Table 8 Results from two-way ANOVAs for mortality of queen conch (Strombus gigas ) in enclosure experiment II. Sources are the same as described in Table 4 Source df MS F P Period I (29 November 1990 to 3 January 1991— 35 days) Site x stock type 1 0.563 0.239 0.634 Site l 5.063 2.150 0.168 Stock type 1 0.563 0.239 0.634 Error 12 2.354 Period II (3 January to 21 February 1991—50 days) Site x stock type 1 0.250 0.143 0.712 Site 1 110.250 63.000 <0.001 Stock type 1 0.250 0.143 0.712 Error 12 1.750 Table 9 Results of two-way ANOVAs for mortality of queen conch (Strombus gigas) in tether experiment I. Sources are the same as described in Table 4 Source df MS F P Period I (12 April to 27 May 1990—45 days) Site x stock type 1 1.563 0.926 0.355 Site 1 0.063 0.037 0.851 Stock type 1 14.063 8.333 0.014 Error 12 1.688 Period II (27 May to 11 July 1990 — 45 days) Site x stock type 1 <0.001 <0.001 1.000 Site 1 0.250 0.098 0.759 Stock type 1 12.250 4.820 0.049 Error 12 2.542 Table 1 0 Results of two-way ANOVAs for growth rates of queen conch (Strombus gigas) in tether experiment I. Sources are the same as described in Table 4. Source df MS F P Period I (12 April to 27 May 1990--15 days) Site x stock type 1 0.004 28.152 <0.001 Site 1 0.005 31.072 <0.001 Stock type 1 0.014 85.104 <0.001 Error 107 <0.001 Period II (27 May to 11 July 1990—45 days) Site x stock type 1 <0.001 0.535 0.467 Site 1 0.016 50.410 <0.001 Stock type 1 0.002 7.629 0.007 Error 72 <0.001 may be speculated: 1) hatchery and wild conch were different in their metabolic functions, such as parti- tioning of energy into somatic and shell growth, 2) slow growth in hatchery-reared conch was a suble- thal effect of transport, or 3) poor growth was re- lated to behavioral characteristics of hatchery-reared conch, such as a reduced ability to recognize foods in the new habitat or unusually high motility. Labora- tory experiments by Siddall (1984b) showed that 10-mm juvenile queen conch held at high density had high locomotory activity and associated low growth rates. Seemingly constant motion and lack of burial in our hatchery-reared animals suggest that their metabolic demands may have been high. However, high growth rates in the hatchery conch later in our o o o a? £ o 80 60 • — • Wild - Ct ▲ — A Wild - C2 O — O Hatchery - C1 A — A Hatchery - C2 ^ Days Figure 1 2 Cumulative mortality curves for enclosure experi- ment II. Hatchery-reared and wild queen conch, Strombus gigas, were compared at sites CI and C2. Forty conch of each stock type were held at each site. investigation showed that the problem was not a per- manent characteristic of the stock type, and others have shown that hatchery-reared conch can have normal growth rates in the field (Appeldoorn and Ballantine, 1983; Davis et al., 1992). Nevertheless, as suggested earlier (Stoner and Sandt, 1991, 1992), growth appears to be a very sensitive indicator of a seed animal's physiological performance in a new habitat. A more serious difference occurred in mortality rates. From the first field experiments with hatch- ery-reared queen conch juveniles (Appeldoorn and Ballantine, 1983) it has been clear that small conch are highly susceptible to predation. Recommenda- tions for release size range from 4 cm shell length Stoner and Davis: Outplanting queen conch. Strombus gigas 405 0.10 Period 1 Nov. - Dec 1990 (mm/day) o d O Ul | i i ^^^M hatchery wild hatchery wild £ Site C1 Site C2 CD DC ■g 0.10 o Period II Jan. -Feb 1991. CD 0.05 0 hatchery wild hatchery Wild Site C1 Site C2 Figure 1 3 Growth rate comparisons for enclosure experiment II. Hatchery-reared and wild queen conch, Strombus gigas, were held at 1.0 conchm2 at sites CI and C2. Values are mean + SD for average growth rates in four replicate en- closures. (Berg, 1976) to 10 cm or larger (Jory and Iversen, 1983). Even with the use of 8-12 cm shell length test animals in this investigation, hatchery-reared conch on tethers were killed at a rate nearly twice the rate of wild conch early in the study. Morphological and behavioral differences are probably the most impor- tant factors influencing mortality. Thin shells and short apical spines observed in the hatchery-reared conch would present a smaller, more vulnerable prey to predators. Palmer's (1979) experiments have shown that spination is an important shell charac- teristic for minimizing predation in intertidal gas- tropods. Shell weight and spination are malleable traits in queen conch. Alcolado ( 1976) observed that shell form in the species was related to water depth and habi- tat type, with thin shells and short-spines being as- sociated with rapid growth in shallow water. Envi- ronmental mediation of shell form was tested experi- mentally by Martin-Mora (1992) near Lee Stocking Island. She found that transplanted wild conch took o -w o o -M D -4— ' o 80 60 -• Wild - C1 -» Wild - C2 -O Hatchery - C1 -A Hatchery - C2 80 100 Figure 14 Cumulative mortality curves for tether experi- ment I. Hatchery-reared and wild queen conch, Strombus gigas, were compared at sites CI and C2. Forty conch of each stock type were teth- ered at each site. on the morphology of local conch within several months, and that high shell weight and long spines were associated with slow growth rate. Given the importance of shell quality in molluscan biology, attention has been given to relationships between shell properties and diets, substrata, temperature, salinity, and other physical factors (Wilbur, 1964; Carter, 1980). It is likely, therefore, that culture techniques can be developed to provide seed conch which are less vulnerable to predation. Survivorship of hatchery-reared conch may also have been influenced by their low burial frequency. Wild conch tend to shelter under detritus or al- gae, and remain partially buried and unmoving for long periods of time. This probably provides a certain degree of protection from larger visual predators. Low burial frequency in hatchery- reared conch may be related to the fact that the field grow-out area in the Caicos Islands, where they spent several months before being transplanted to the Exuma Cays, was primarily a hard-bottom environment. Behavioral differences between hatch- ery-reared and wild stocks are rarely documented; however, Schiel and Welden ( 1987) found that hatch- ery-reared red abalone, Haliotis rufescens, did not move to concealed locations as did wild abalone, re- sulting in higher predatory mortality. There are at least three limitations of the pre- sent investigation. One is not knowing whether convergence in the morphology and survivorship of wild and hatchery-reared conch was related to 406 Fishery Bulletin 92(2). 1994 0 15 Period 1 April • May 1 990 0 10 31 -^ 0.05 CO ■D E n " | 22 23 I 35 E. ° hatchery wild hatchery wild ® Site C1 Site C2 CO rr _c 0.15 % o Period II June- July 1990 ° 0.10 0.05 24 i I 14 14 0 hatchery wild hatchery wild Site C1 Site C2 Figure 1 5 Growth rates of hatchery-reared and wild queen conch, Strombus gigas, held on tethers at sites CI and C2 during experiment I. Values are mean ± SD, with the number of conch measured shown inside the vertical bars. adaptation by individuals over the course of the investigation, or is explained simply by differen- tial survivorship (i.e. the most fit hatchery-reared conch survived to the end of the experiments). As discussed earlier, a strong case for adaptation can be made because of the obvious changes in shell morphology over time within individual conch. Second, hatchery-reared conch used in this study were from one season's production in a single hatchery. We know that different hatcheries and different cultures from individual hatcheries can produce conch with different characteristics. For example, Jory and Iversen (1988) found different shell strengths among cultures of queen conch. Because juvenile conch are reared from egg masses collected from the wild, and because both shell morphology and behavior appear to be relatively plastic characteristics in queen conch, we believe that differences shown between hatchery-reared animals and native stocks can be alleviated through modifications in diets, hatchery substrata, and other culture techniques. Field viability must be considered continuously throughout the hatch- ery-rearing process. Third, morphological effects on survival may vary with site because of differ- ences in predator assemblages. For example, at a site where molluscs (such as tulip snails, Fasciolaria tulipa ) are the most important preda- tors, size and escape behavior may be more impor- tant than spine length and shell thickness. More site comparisons and better knowledge of predator-prey relationships are needed. 80 p ^■v • — • Wild - C1 o -5 60 O O Hatchery - C1 S O sy £S 40 // Mortality O ^^"^^# ^^ C ) 20 40 60 80 100 Days Figure 1 6 Cumulative mortality curves for hatchery-reared and wild queen conch, Strombus gigas, held on tethers at site CI during experi- ment II. Forty conch of each stock type were tethered at each site. Importance of stock enhancement sites Site selection for stock enhancement with queen conch is a complex issue and the subject of several earlier papers (Stoner andSandt, 1991, 1992; Stoner etal., 1993; Stoner, in press). It is clear from experi- ments reported here that even carefully chosen locations, such as our non-conch study site C2, may not support juvenile conch over the long term. Conch at this site demonstrated consistently higher growth than conch at the traditional nurs- ery site (CI), but mortality was also higher in both tethered and free-ranging conch. Site differences in mortality could be as- sociated with patterns of predator abun- dance or diversity, or both. Although predators may accumulate where prey density is high, the most likely explana- tion for lower predation rate at site CI Stoner and Davis: Outplanting queen conch, Strombus gigas 407 Table 1 1 Summary of results from free-ranging, enclosure, and tether experiments on mortality and growth of hatchery-reared and wild queen conch (Strombus gigas) at two field sites (CI & C2) . W = wild conch. H = hatchery-reared conch. PI, PII, and PHI refer to different growth periods within the experiment. Signs indicate statistically significant differences (see text). Differences Stock type Site Experiment Site CI Site C2 Mortality differences Free-ranging WH W>H W>H W>H fall (PII) W>H W>H C1H W>H(PI) W = H(PII) C1H W = H W:C1>C2 H:C1 = C2 Tether expt. I W> H W>H C1200 m) fishing area north of the Avalon Peninsula, CPUE declined less and the proportion of newly molted male snow crab remained rela- tively constant during the same period. Coincident with the decline of the Avalon Peninsula fishery was a pronounced drop in mean bottom temperature on the commercial fishing grounds, from -0.6° C to -1.4°C, a phenomenon not observed in Bonavista Bay. This decline in water temperature appears to have been the cause of the fishery col- lapse because temperatures be- came low enough to interrupt the molting cycle of snow crab off the Avalon Peninsula. If the potential impact of the lower water tempera- tures and subsequent long-term cessation of growth and recruit- ment within the snow crab popu- lation had been recognized, the available pool of commercial-sized crab could have been harvested more slowly over a period of years to lessen the disruption of the fishery. A snow crab, Chionoecetes opilio (Decapoda, Majidae), fishery collapse in Newfoundland David M. Taylor Paul G. O'Keefe Charles Fitzpatrick Science Branch, Department of Fisheries and Oceans P O. Box 5667. St. Johns, Newfoundland AIC 5X1 Manuscript accepted 18 October 1993 Fishery Bulletin 92: 412-4 Hi 412 The Newfoundland snow crab, Chionoecetes opilio, fishery began in 1968, and until 1978 was confined to deep water (>220 m) bays and areas within 30 km of the coast. In 1978, fishing effort in offshore ar- eas east of the Avalon Peninsula (Fig. 1 ) increased markedly, result- ing in peak landings in 1981 (Fig. 2) of 8609 t (Taylor and O'Keefe1). The Canadian Atlantic Fisheries Scientific Advisory Committee (CAFSAC) recommended that an- nual exploitation rates for commer- cially harvested snow crab stocks not exceed 50-60% of annual pro- ductivity in order to prevent overexploitation (Anon.2). The Com- mittee adopted this guideline be- cause the fishery targets males only and most males have an opportu- nity to mate at least once before reaching the minimum legal size, thereby ensuring adequate recruit- ment into the fishery and the repro- ductive integrity of the populations. However, should there be a recruit- ment failure, the reproductive po- tential is maintained by sublegal sexually mature males in the popu- lation and females that can produce at least two clutches of viable eggs from one copulation using stored sperm (Paul, 1984). The fishery occurs from April un- til November each year. Typically, catch per unit of effort (CPUE) de- clines throughout the fishing sea- son until July and August when a high level of molting activity results in an increased abundance of newly molted recruits. As fishermen are discouraged by processors from landing these low-yield soft-shelled crabs, new-shelled animals of legal size (>95 mm carapace width (CW)) generally enter the fishery in the following spring. Their recruitment is evident by high CPUE values at the beginning of the next fishing season (Taylor and O'Keefe1). Between 1979 and 1982, exploi- tation rates off the Avalon Penin- sula remained within recommended levels, but beginning in 1982 catch rates declined rapidly until 1984 when fishing became uneconomical (Taylor and O'Keefe3). Other areas in Newfoundland, such as Bona- vista Bay (Fig. 1 ), have consistently had exploitation rates in excess of recommended levels and conse- quently have experienced reduc- tions in catch rates. However, the 1 Taylor, D. M., and P. G. O'Keefe. 1984a. Assessment of Newfoundland snow crab (Chionoecetes opilio) stocks, 1982. Can. Atl. Fish. Sci. Advis. Comm. CAFSAC Res. Doc. 84/13, Dartmouth, Nova Scotia, 35 p. 'Anonymous, 1981. Advice on some invertebrate and marine plant stocks. Can. Atl. Fish. Sci. Advis. Comm. CAFSAC Advisory Document 81/1, Dartmouth, Nova Scotia, 6 p. :< Taylor, D. M., and P. G. O'Keefe. 1986. Analysis of the snow crab, Chionoecetes opilio, fishery in Newfoundland for 1985. Can. Atl. Fish. Sci. Advis. Comm. CAFSAC Res. Doc. 86/57, Dartmouth, Nova Scotia, 24 p. Taylor et al.: Snow crab, Chionoecetes opilio. fishery collapse 413 Figure 1 Offshore Avalon Peninsula and Bonavista Bay, Newfoundland, Chionoecetes opilio fishing grounds. magnitude of the collapse in catch rates and land- ings off the Avalon Peninsula is unprecedented. This paper examines biological data, fishermen's log returns, and temperature records from the Avalon Peninsula in comparison with similar data from Bonavista Bay in an attempt to describe possible reasons for this collapse. Materials and methods A total of 27 spring and fall cruises were conducted on the commercial fishing grounds off the Avalon Pen- insula and in Bonavista Bay ( Fig. 1 ) during 1979-88 to monitor population characteristics and catch rates. Fishing stations were selected randomly and strati- 414 Fishery Bulletin 92(2), 1994 o z o AVALON PENINSULA LANDINGS /X / \ AW.ON PENINSULA EFFORT \ / BdHAVISTA SAY EFFORT YEAR Figure 2 Summary of annual commercial landings and fishing effort for Chionoecetes opilio from offshore Avalon Peninsula and Bonavista Bay, Newfoundland. fied by depth. Japanese-style conical traps baited with approximately two kg of northern shortfin squid, Illex illecebrosus , or with a mixture of squid and At- lantic mackerel, Scomber scombrus, and set in longline fleets of 12 were used to catch crabs. Al- though an attempt was made to duplicate the meth- odology employed by fishermen, space limitations onboard the research vessels restricted fleets of traps to 12 rather than the 50-70 used in commercial fish- ing. Weather permitting, traps were hauled after a 24-hour soak. Crabs were removed from the traps, carapace width (CW) measured to the nearest 1 mm and shell condition determined. Three shell condition classes were used, based upon the following criteria de- scribed by Miller and O'Keefe (1981) and modified by Taylor et al. (1989): Soft Shell (1) Carapace is brightly colored and free of epibiotic growth. Ventrally the crab is off-white to cream in color. Chelae bend, break, or crack with slight pressure. Bright iridescence present on dorsal margin of chelae. Animals within this category are considered as having molted not more than 90 days prior to capture. New/Hard (2) Carapace is duller in color and a num- ber of tube worms are present. Shell is hard and ven- trally may be dark-cream colored and covered with discolored scratches. Chela does not bend or break when moderate thumb pressure is applied. Irides- cence on chelae is reduced in intensity. This category generally applies to animals that have molted up to two years prior to capture. Old/Hard (3) Carapace is dull brown in color and an assortment of calcareous tube worms and bar- nacles are present. The shell, although still hard, may have a slight "leathery" feel. Ventrally, the shell is brownish in color and many dark scratches evident. Iridescence on chelae is faint or absent. This category applies to animals that have not molted for at least two years. Bottom temperatures from depths >170 m were obtained from an oceanographic station (Station 27) near the Avalon Peninsula fishing grounds (Fig. 1). There is no oceanographic station in Bonavista Bay, but bottom temperatures were obtained during eight research cruises by using expendable bathyther- mographs, or reversing thermometers. Taylor et al.: Snow crab, Chionoecetes opilio, fishery collapse 415 CPUE's were derived for research fishing by mul- tiplying the number of commercial crab caught per trap haul by a conversion factor of 0.45 kg/crab (Tay- lor, unpubl. data). Fishermen's logbook catch data were checked against processors' purchase slips. CPUE from logbook data was corrected for the per- centage of sublegal crab in their catch as determined from sampling at processing plants conducted simul- taneously with the research cruises. Total catch re- ported by log/purchase slips was then multiplied by the percentage of legal-sized animals and divided by the reported effort for the same time period to ob- tain a CPUE value comparable to that derived from a coincident research cruise. Results Logbook data Landings and effort data from commer- cial logbooks are summarized in Fig- ure 2. The drop in effort and landings for the Avalon Peninsula in 1980 is the result of a labor dispute and probably does not reflect abundance. Fishery CPUE in both the Avalon and Bonavista Bay areas was at its highest levels dur- ing the spring 1980 at 23.2 kg/trap haul and 11.6 kg/trap haul, respectively (Fig. 3). However, off the Avalon Peninsula, catch rates dropped to 19.9 kg/trap haul during the 1981 spring fishery as land- ings peaked at approximately 8500 t. CPUE declined to 13.8 kg/trap haul in 1982 despite logbook reports that new commercial fishing grounds were being exploited in the offshore areas ( > 100 km from land). This decline continued, reaching 3.7 kg/trap haul in the spring fishery of 1985 (Fig. 3), a drop of 84% from 1980 levels. This decline in CPUE was accompanied by a dramatic reduc- tion in effort falling from 480,000 trap hauls in 1981 to 17,000 in 1985 (Fig. 2). In comparison, the commercial spring fishery in Bonavista Bay, although over- exploited (Taylor and O'Keefe4), has maintained a comparatively stable level of landings and CPUE since 1981 (905- 1805 t and 4.1-to 8.2 kg/trap haul, re- spectively, despite an overall increase in effort from 1980 levels [Fig. 2]). Un- like the Avalon Peninsula fishery, spring catch rates in this area consistently re- flect growth and recruitment into the commercial biomass as newly molted individuals re- cover to commercial acceptability over winter. Research cruise CPUE and shell condition data Logbook-derived commercial CPUEs and research cruise CPUE data with calculated confidence inter- vals are represented in Figures 3 and 4, respectively. Confidence intervals for most offshore Avalon re- search cruise CPUE data are fairly tight, with the exception of those data from a February 1986 cruise. 4 Taylor, D. M., and P. G. O'Keefe. 1987. Analysis of the snow crab (Chionoecetes opilio) fishery in Newfoundland for 1986. Can. Atl. Fish. Sci. Advis. Comm. Dartmouth, Nova Scotia, 26 p. CAFSAC Res. Doc. 87/57, 416 Fishery Bulletin 92(2). 1994 These CPUE values were derived from only four fishing sets that were placed on the commercial fishing grounds opportunistically. On the commercial fishing grounds off the Avalon Peninsula, CPUE's derived from research cruises were nearly identical to those of commer- cial enterprises at approximately 19.5 kg/trap haul in 1979 (spring). However, CPUEs diverged over the years primarily because commercial vessels had a greater fishing range than the re- search vessels. Nevertheless, the sharp drop in CPUE between 1982 and 1983 was reflected in both the commercial and research cruise data. In the commercial fishery, CPUE dropped from 13.8 kg/trap haul in the spring of 1982 to 8.0 kg/ trap haul in the spring of 1983 (Fig. 3). This de- cline in crab abundance was mirrored in research cruise data, which indicate a decline from 9.3 kg/ trap haul to 2.9 kg/trap haul over the same pe- riod (Fig. 4). Research cruise data for this period in Bonavista Bay are not available. Logbook data, however, indicate that CPUE dropped by only 0.2 kg/trap haul (Fig. 4). Data on shell conditions of legal-sized and pre- recruit crabs from research cruises conducted off the Avalon Peninsula demonstrate that the drop in CPUE coincided with a decline in the propor- tion of new-shelled crabs from 52.4% to 18.6% (Fig. 5). In Bonavista Bay, the percentage of new- shelled animals dropped to 68.4% in the fall of 1983 from 97.7% in the spring of 1982. However, the proportion of new-shelled animals quickly rebounded to 97%> in 1984 as opposed to 40.6% off the Avalon Peninsula during the same year. Temperature data From the spring of 1978 through the first half of 1982, mean bottom temperature ranged from -0.3° C to -0.8°C off the Avalon Peninsula (Fig. 5). During the second half of 1982, the beginning of a trend to- wards colder bottom temperatures was evident. Bot- tom temperatures during this period dropped as low as —1.6 C and rarely rose above — 1.0°C the entire pe- riod of mid 1982 to 1986. Two brief periods of warm- ing occurred in both 1983 and 1984 but these peri- ods were short-lived and weak. In 1986 a general warming trend began with an increase from the 1985 low of -1.6°C to around -1.1°C, a trend that has con- tinued to the present. The drop in temperature and the decrease in the proportion of new-shelled crabs appeared to coincide with bottom temperatures declining during April- May of 1982 whereas the percentage of new-shelled 20- 18- AVALON 16- If I | 14- H,2- S io- u 2 8' o 8- 4- ; t i 1 1 1, ! 2- I m ■ z i i o- 20- I I I i i I I I I I I I 18- BONAVISTA 16- 14- E 12- < 2w UJ 2 ■■ o I I 1 i i I 6 ' 4- 2- I 1 I o- 1 1 1 1 1 1 1 1 1 1 ■ | 78 79 80 81 82 83 84 85 86 87 88 89 YEAR Figure 4 Research cruise CPUE's with calculated confidence inter- vals for the Avalon Peninsula and Bonavista Bay, 1979-88. Solid circles represent spring cruises and X's represent fall cruises. A circled 'X' is used to highlight data from the Feb- ruary 1986 cruise during which only 4 stations were fished in a nonrandom manner. crabs dropped from 52.4% in April to 17.0% in Sep- tember. Figure 5 illustrates that between 1982 and 1986 there were two brief periods (1983 and 1984) when mean bottom temperature at Station 27 in- creased slightly and coincidental increases in the proportion of new-shelled animals followed in 1984 and 1985. This delay between increase in water tem- peratures and appearance of new-shelled crabs is consistent with our current understanding of snow crab molting mechanisms (Moriyasu5). The warming trend in July 1984 was short-lived (Fig. 5). During 1985 water temperatures dropped to the lowest level of all the years examined. Shell 5 Moriyasu, M. Dept. Fisheries And Oceans, Gulf Fisheries Centre, Box 5030, Moncton. N.B. E1C 9B6. Personal commun. April, 1987. Taylor et al.: Snow crab, Chionoecetes opilio, fishery collapse 417 z 0 h K 0 CL 0 a. o. 100 ■ BONAVISTA ' 80 • * ' A A~~ * ft 60 ■ * 40 • \ - ' - -' 20 • 0 - YEAR Figure 5 Changes in the percentages of new-shelled Avalon Peninsula and Bonavista Bay Chionoecetes opilio males > 70 mm CW in relation to changes in bottom water temperatures on snow crab commercial fish- ing grounds, 1979-88. Solid circles connected by solid lines represent the proportion of new-shelled crabs and the triangles connected by a broken line represent bottom water temperature. Although there was a significant decline in the proportion of new- shelled crab in 1983 after the water temperature declined (68.4% vs. 97.7%) in the spring of 1982 this value quickly rebounds in 1984 to 97.0% de- spite only a marginal warming of the water (Fig. 5). Discussion condition sampling in February 1986 indicated no molting activity during the summer and fall of 1985 (Fig. 5). A further indication that the impact of these in- creases in new-shelled animals was minimal is il- lustrated by continuing low CPUE's (Figs. 3 and 4). Large proportions of new-shelled animals and higher CPUEs were not observed until the warming trend of 1986 was firmly established. While the decline in temperature between 1982 and 1983 is mirrored in data collected during Bonavista Bay research cruises (Fig. 5), the lowest temperatures encountered in Bonavista Bay are roughly equivalent to normal tem- peratures at Station 27 off Avalon Peninsula. The tight confidence intervals for the Avalon Peninsula research cruise CPUE's (Fig. 4) for 1982-1987 (except February 1986) are indicative of just how severe, widespread, and enduring the resource depletion was in this area. Each research survey fished sta- tions randomly selected and stratified by depth and covered virtually all the approximately 3600 sq. km. of com- mercial snow crab fishing grounds (Taylor et al.6). Had a sustained recov- ery been made by the crab population in any section of the commercial fish- ing grounds, it would have almost cer- tainly been detected, either by our re- search cruises or by commercial crab fishermen. Little is known about environmen- tal factors that affect snow crab molt- ing physiology. Low water tempera- tures may inhibit molting in crabs (Hiatt, 1948; Adelung, 1971; Leffler, 1972; Warner, 1977) and other deca- pods (Travis, 1954; Aiken, 1980;Ennis, 1983). Foyle (1987) determined that snow crab from Cape Breton Island are able to maintain normal physiological func- tions at temperatures much higher than their normal temperature range. At high tem- peratures however and at temperatures below 1°C, reproductive growth and net energy consumption become slightly negative. Snow crab on the north- east coast of Newfoundland live at much lower wa- ter temperatures (<-0.75°C) than do those off the Cape Breton Island and a drop in temperature may result in such a "deficit" in their energy budget that molting physiology is impaired. Taylor, D. M., W. R. Squires,and P. G. O'Keefe. 1983. An alternate methodology for estimating snow crab (Chionoecetes opilio) populations in commercially fished areas. Can.Atl. Fish. Sci.Advis. Comm. CAFSAC Res. Doc. 83/1. Dartmouth, Nova Scotia, 10 p. 418 Fishery Bulletin 92(2). 1994 Aiken (1980) demonstrated that molting is inhib- ited in American lobsters, Homarus americanus, at 5° C if active premolt is not achieved before water temperature drops to that level. A similar physiologi- cal response to declining temperature may have con- tributed to the apparent reduction in molting in the snow crab population off the Avalon Peninsula after temperature dropped in 1982. Molting in Newfoundland snow crab in this area generally occurs during May-August. The sharp de- cline in abundance of new-shelled crab in Septem- ber indicated that many animals that would normally molt during this period apparently failed to do so, possibly as a result of extremely cold water tempera- tures. Both Bonavista Bay and the offshore Avalon are affected by the Labrador Current. In 1982 the cur- rent became wider and deeper than in previous years causing a cooling effect throughout the water column (Akenhead7). Whereas this phenomenon affected the entire east coast of Newfoundland, it was most se- vere near the Avalon Peninsula primarily because of its comparative shallowness ( 174-200 m). Bonavista Bay is 220-486 m deep on the crab grounds and its depth may reduce the cooling effect of the Labrador Current (Akenhead8). This may explain why snow crab molting activity here was not as adversely af- fected as it was on the Avalon crab fishing grounds (Fig. 5). The lack of recruitment into the fishery, between 1982 and 1986, meant that the snow crab resource off the Avalon Peninsula had been in effect "mined" rather than harvested as a renewable resource. The impact of the decline in snow crab abundance was dramatic. This marked decline in landings, from the Avalon Peninsula area, resulted in a substantial drop in employment and earnings from the snow crab fish- ery (Collins9). Evidence of a link between temperature and molt- ing is circumstantial. However, the drop in water temperature followed by a rapid decline in molting activity (Fig. 5), and subsequently CPUE (Fig. 4), makes a compelling argument that the yearly pro- portion of snow crab molting in an area is largely dependent on environmental conditions in the par- ticular case where temperatures are very low and variable. The argument is supported by the observa- tion that twice between 1984 and 1988 water tem- peratures rose and fell, affecting subsequent changes in the proportions of crabs molting off the Avalon Peninsula (Fig. 5). If molting of pre-recruit snow crabs and consequent recruitment into the fishery are affected by changes in water temperature, the impact of these changes should be included in re- source management programs. The existing policy of allowing yearly exploitation rates of 50-60% should be re-examined. As the effects of these environmen- tal changes may be long term, recommended exploi- tation rates could be reduced to prolong a fishery in which recruitment has been interrupted. Assess- ments of the fishery off the Avalon Peninsula (Tay- lor and O'Keefe4) indicate that exploitation rates were <65% between 1979 and 1981. However, owing to the failure of undersized males to molt into the fishery beginning in 1982, each successive year's standing stock was reduced until fishing was no longer economically viable. In contrast, although exploitation rates in Bonavista Bay consistently ex- ceeded 75% (Taylor and O'Keefe10, 198311, 1984b12, and 19874), molting within the population continued to provide sufficient recruitment for a viable fishery even though catch rates declined. With the excep- tion of 1983 when the incidence of new-shelled ani- mals (shell condition 1 and 2) in research cruise catches fell to 68% (Fig. 5), new-shelled animals com- posed in excess of 85% of the catch of legal-sized and immediate pre-recruits. This high level of molting appears to have prevented the precipitous decline in catches experienced in the offshore Avalon Peninsula area. To prevent future declines of such proportions it may be advisable to monitor temperature, catch rates, and crab shell condition more closely on a sea- sonal basis. Efforts should also be made to determine thermal requirements for molting in snow crab. The implications for resource management strategy are simply that, regardless of exploitation levels, changes in temperature likely affect molting and hence re- cruitment to the standing stock to such an extent that assumptions regarding long-term sustainability of annual landings are not justified. ' Akenhead, S. A. 1986. The decline of summer subsurface temperatures on the Grand Bank, at 47°N, 1978-1985. NAFO SCR Doc. 86/25, 8 p. 8 Akenhead. S. A. Institute of Ocean Sciences, Box 6000, Saanick Rd. Sydney, B.C. V81 4B2. Personal commun., April 1987. 9 Collins, J. F. Chief, Economic Analysis Division, Program Coordination & Economics Branch, Northwest Atlantic Fisheries Centre, P.O. Box 5667, St. John's, Newfoundland A1C 5X1. Personal commun. September 1993. 10 Taylor, D. M., and P. G. O'Keefe. 1981. Assessment of snow crab (Chionoecetes opilio) stocks in Newfoundland, 1979. Can. Atl. Fish. Sci. Advis. Comm. CAFSAC Res. Doc. 81/57, Dartmouth, Nova Scotia, 34 p. 11 Taylor, D. M., and P. G. O'Keefe. 1983. Assessment of snow crab (Chionoecetes opilio) stocks, in Newfoundland in 1980. Can. Atl. Fish. Sci. Advis. Comm. CAFSAC Res. Doc. 83/3, Dartmouth, Nova Scotia, 40 p. 12 Taylor, D. M., and P. G. O'Keefe. 1984b. Assessment of Newfoundland snow crab (Chionoecetes opilio) stocks in Newfoundland for 1982. Can. Atl. Fish. Sci. Advis. Comm. CAFSAC Res. Doc. 84/3, Dartmouth, Nova Scotia, 30 p. Taylor et al.: Snow crab. Chionoecetes opilio, fishery collapse 419 Acknowledgments The authors are indebted to the crews of the C. S. S. Shamook and Marinus who provided able support during the research cruises. G. P. Ennis, J. Hoenig, R. Knoechel, D. G. Parsons, and P. Schwinghamer provided helpful comments on the manuscript. P. Collins, H. Mullett and G. King prepared the fig- ures while M. Hynes and J. Lannon typed the manu- script. Literature cited Adelung, D. 1971. Studies on the molting physiology of decapod crustaceans as exemplified by the shore crab {Carcinus maenas). Sci. Mar. Res. at Heligloland 22:66-119. (English translation by Fish. Res. Board Can., Transl. Ser. No. 1877.) Aiken, D. E. 1980. Molting and growth, Chapter 2. In J. S. Cobb and B. F. Phillips (eds. ), The biology and management of lobsters. Vol. 1: Physiology and behavior. Aca- demic Press, 463 p. Ennis, G. P. 1983. Variation in annual growth in two Newfound- land lobster (Homarus americanus) populations in relation to temperature conditions. ICES C.M.1983/K:24, 11 p. Foyle, T. P. 1987. Metabolism and energetics in a cold-water crustacean: the snow crab, Chionoecetes opilio. M.Sc. thesis, Dalhousie Univ., 83 p. Hiatt, R. W. 1948. The biology of the lined shore crab, Pachy- grapsus crassipes Randall. Pacific Sci. 2:135-218. Leffler, C. W. 1972. Some effects of temperature on the growth and metabolic rate of juvenile blue crabs, Callinecetes sapidus, in the laboratory. Mar. Biol. 4:104-110. Miller, R. J., and P. G. O'Keefe. 1981. Seasonal and depth distribution, size, and molt cycle of the spider crabs, Chionoecetes opilio, Hyas araneus, and Hyas coarctatus in a Newfoundland bay. Can. Tech. Rep. Fish. Aquat. Sci. 1003:iv + 18 p. Paul, A. J, 1984. Mating frequency and viability of stored sperm in the tanner crab Chionoecetes bairdi (De- capoda, Majidae). J. Crust. Biol. 4(3):375-381. Taylor, D. M., G. W. Marshall, and P. G. O'Keefe. 1989. Shell hardening in snow crabs Chionoecetes opilio tagged in soft-shelled condition. N. Am. J. Fish. Manage. 9:504-508. Travis, D. F. 1954. The molting cycle of the spiny lobster, Panulirus argus Latreille. I: Molting and growth in laboratory-maintained individuals. Biol. Bull. 107:433-450. Warner, G. F. 1977. The biology of crabs. Van Nostrand Reinhold Company, New York, 202 p. Abstract. — Larval Atlantic menhaden, Brevoortia tyrannus, were collected weekly during their expected recruitment (November- April) to the estuary near Beaufort, North Carolina, over seven con- secutive years beginning 1985-86. The larval density in nighttime quantitative samples was calcu- lated and ages determined from otolith microstructure. Back-calcu- lated birthdates and larval abun- dance data were used to estimate the relative contribution of weekly age cohorts to seasonal recruitment of larvae. Summaries of these data were measures of the spawning dis- tributions. Larvae were recruited to the estuary from mid-November through April, with about 86% col- lected during February-April. In all years, age and size of larvae in- creased linearly throughout re- cruitment until the end of March and then declined. The mean age of recruited larvae over all years was 61 days and the mean stan- dard length was 24.6 mm. Atlantic menhaden spawning season was protracted, lasting 4-6 months. In every spawning season, a dominant birthweek mode in either Decem- ber or January contributed from 25-43% of the total recruits. More than 76% of all spawning occurred in the December- January period. Individual birthweek cohorts re- cruited to the estuary over periods from one week to several months. Cohorts that usually contributed the greatest number of individuals to estuarine recruitment usually recruited over longer periods. At- lantic menhaden have apparently selected a spawning season and lo- cation that ensures transport of larvae across the southeast United States continental shelf and arrival of most larvae during a time when conditions are conducive to optimal survival in the estuary. Spawning time and recruitment dynamics of larval Atlantic menhaden, Brevoortia tyrannus, into a North Carolina estuary Stanley M. Warlen Beaufort Laboratory, Southeast Fisheries Science Center National Marine Fisheries Service, NOAA 101 Pivers Island Road, Beaufort, North Carolina 28516-9722 Manuscript accepted 20 October 1993 Fishery Bulletin 92:420-433 (1994) 420 The Atlantic menhaden, Brevoortia tyrannus, is a commercially impor- tant clupeid that ranges on the east coast of the United States from the Gulf of Maine to the central coast of Florida. Tagging studies have shown that this species makes ex- tensive seasonal, migrations north- ward along the coast in spring and southward in fall and winter (Dryfoos et al., 1973; Nicholson, 1978). Most of the population is thought to overwinter in the area between Cape Hatteras, North Carolina, and northern Florida (Ahrenholz et al., 1987). Spatial and temporal trends in Atlantic menha- den spawning have been suggested by studies on the distribution of eggs and larvae (Reintjes, 1961; Kendall and Reintjes, 1975; Judy and Lewis, 1983) and studies on ovarian maturity and fecundity (Higham and Nicholson, 1964; Lewis et al., 1987). Those studies show that spawning occurs off New England from late spring to early summer and again in early fall, off the mid-Atlantic states in spring and fall and off the southeastern states from October to March. Maxi- mum numbers of menhaden prob- ably spawn in winter in offshore waters south of Cape Hatteras (Reintjes, 1969; Judy and Lewis, 1983) and waters off the North Carolina coast may be one of the major spawning grounds for Atlan- tic menhaden (Higham and Nichol- son, 1964). Plankton collections taken off North and South Carolina suggest that Atlantic menhaden may con- tinuously spawn from late fall to early spring. Collections taken in the vicinity of Beaufort, North Caro- lina, from 1955 to 1961 (unpubl. data, National Marine Fisheries Service, Beaufort Laboratory, cited by Higham and Nicholson, 1964) showed larvae in samples beginning in November or early December and continuously thereafter until mid- April. Subsequent work supported these estimates of the timing of es- tuarine immigration of menhaden larvae. Lewis and Mann (1971) sampled larval menhaden semi- monthly as the fish recruited to the estuary near Beaufort Inlet in the fall/winter 1966-67 and 1967-68 seasons, to estimate relative in- dexes of abundance. Densities of menhaden larvae recruited to estu- aries were reported for North Caro- lina (Hettler and Chester, 1990; Warlen and Burke, 1990) and South Carolina (Allen and Barker, 1990). Examination of otolith micro- structure to count daily growth in- crements has made possible reason- ably accurate estimates of the age and growth of the early life history stages of fishes. Age at estuarine recruitment can be tracked within and among seasons or years. Age at Warlen: Spawning time and recruitment of Brevoortia tyrannus 421 recruitment is also a measure of the total time from offshore spawning to recruitment to the estuary. Valuable new information on the early life his- tory of larvae can be obtained when estimates of recruitment densities are combined with estimates of sea- sonal age composition of the catches. Back-calculated birthdate distribu- tions of larvae in each sample can be multiplied by the catch density to de- termine the relative contribution of birthweek cohorts to the number of immigrants. The purposes of this seven-year study were to document the duration and relative abundance of larval At- lantic menhaden recruitment to the estuary near Beaufort, North Caro- lina; to measure size and estimate ages of recruited larvae; to back-cal- culate birthdate distributions of lar- vae; and to determine the contribu- tion of birthweek cohorts to the num- ber of immigrants for an estimation of temporal spawning within and among sampling years. Methods Larval collection CAPE LOOKOUT Larval Atlantic menhaden were col- lected at a station adjacent to Pivers Island ( Fig. 1 ) as they recruited to the Newport River estuary from the At- lantic Ocean. Sampling, designed to cover the ex- pected recruitment period of mid-November to April (Lewis and Mann, 1971), ran for seven consecutive years beginning with the 1985-86 sampling year. Larvae were sampled from 13 November 1985 to 23 April 1986 with a 60-cm bongo frame with paired 505-|im mesh nets and flow meters pulled by a 6.7-m boat. Tows were made weekly during night- time hours just after mid-flood tide, when the cur- rent was strongest, to reduce potential net avoidance by larvae. Data from the catch of both nets from two consecutive tows, one surface-bottom-surface (double oblique) in 3-5 m of water and one just under the surface (subsurface), were averaged to estimate lar- val density (number larvae/100 m3 water fished) and to provide fish for ageing. The lack of a significant difference (paired comparison r-test, P>0.05) in catch density from subsurface and double oblique tows al- Figure 1 Location of the sampling site for larval Atlantic menhaden, Brevoortia tyrannus, in the lower Newport River estuary near Pivers Island, North Carolina. lowed the paired data to be combined. In the other six sampling years (19 November 1986-30 April 1987, 10 November 1987-4 May 1988, 16 November 1988- 3 May 1989, 15 November 1989-2 May 1990, 14 No- vember 1990-24 April 1991, and 13 November 1991- 6 May 1992) larvae were collected, with a 1 x 2 m neuston net frame fitted with a 945-um mesh net and flow meter fished just under the water surface from a bridge platform over the center of the chan- nel adjacent to Pivers Island. This location was only meters from the site sampled with the bongo nets in 1985-86. As in 1985-86, all samples were collected weekly during nighttime hours at mid-flood tide. Three consecutive sets were made each night in 1986-87 and four each night in subsequent years. Because of the expected seasonal variation in men- haden abundance, sets were between 2 and 16 min- utes long; most were 5-7 minutes long. Volume of 422 Fishery Bulletin 92|2), 1994 water fished by the net ranged from 41 to 805 m3 but most sets filtered 150-350 m3. Ichthyoplankton samples were preserved with 95% ethanol and di- luted so that the final alcohol concentration was >70%. Samples not sorted within 24 hours were rinsed and represerved in 70% ethanol. As for bongo nets, these catches were standardized as the num- ber of Atlantic menhaden/100 m3 of water fished. The mean of the density data from all sets on a given night (bongo nets in 1985—86 and neuston nets in other years) was used as the estimate of density of Atlantic menhaden larvae recruited during the flood tide. Simultaneous larval collections were made with bongo and neuston nets on 17 December 1986 (3 sets) and 18 March 1987 (4 sets) to test for differences in menhaden catch density between gear type. There were no significant differences in density of menha- den caught by gear type (AN OVA, P>0.54) or among sets (ANOVA, P>0.24). Hence, the catch data from bongo and neuston nets are comparable. The mean age or standard length (SL) also did not differ be- tween the two types of nets. Larval ageing Larvae for ageing were subsampled from individual weekly night net sets in proportion to their contri- bution to the total nightly catch. In catches of up to 20 fish, all larvae were used. In catches of >20 fish, subsample sizes were proportional to catch but gen- erally no more than 50 fish were aged per week. The ages of 3,864 larvae were determined. The standard length of each larva to be aged was measured to the nearest 0.1 mm with an ocular mi- crometer. Sagittal otoliths were removed from their surrounding soft tissue, cleaned in distilled water, and placed on a glass microscope slide under a thin layer of Flo-Texx mounting medium. Otoliths were observed with transmitted light on a compound mi- croscope fitted with a television camera. Growth in- crements were counted from images on a video moni- tor at microscope magnifications of 400x or l,000x. One person made dual readings of otoliths from each fish. Readings were averaged if they differed by a count of four or less; if they differed by >5 increments, the fish was excluded from further analyses. Incre- ment counts of about 2% of the aged fish differed by five or more. Estimated age was the number of in- crements counted plus an empirically derived value for the number of days (5) from spawning to first increment formation ( Warlen, 1992). The otolith age- ing technique for Atlantic menhaden larvae has been validated by Maillet and Checkley (1990), who es- tablished that larvae form one otolith growth incre- ment per day. I assumed that the age at initial incre- ment deposition did not vary and that the otolith increment deposition rate was constant within and between sampling seasons. A spawning date (birthdate) was assigned to each ageable larva by using the estimated age of the fish in days to back- calculate from the date of capture. Larvae spawned in a given calendar week were considered in the same calendar birthweek cohort. Estuarine recruitment of birthweek cohorts The percentage contribution of Atlantic menhaden larvae of all birthweek cohorts to the total recruit- ment was measured for each of the seven years from 1985-86 to 1991-92. For each weekly collection, the percentage of larvae spawned in each back- calcu- lated birthweek was determined. Each percentage was then multiplied by the total larval density (num- ber/100 m3) for the week. Based on these results, den- sity estimates were made for larvae from each birthweek cohort recruiting to the estuary on a given collection night. Densities for each birthweek cohort were summed over all collections within a sampling year. The pro- portion that each birthweek cohort contributed was determined by dividing the individual birthweek sums by the total density of all birthweek cohorts for the recruitment year. These computations pro- duced estimates of the relative contribution that each birthweek cohort made to the total recruitment of menhaden larvae into the estuary near Beaufort. Results Larval recruitment abundances Larval Atlantic menhaden recruited to the estuary near Beaufort over a 5 to 5V2 month period from mid- November to the end of April (Fig. 2). Larvae were generally recruited in highest densities during Feb- ruary-April and contained, on average, about 86% of the total estuarine recruitment during those months (Table 1). Recruitment was low in Novem- ber and seldom extended into May. There was a dis- tinct density mode in each of the seven years (Fig. 2). Except for 1988-89, modal density always oc- curred in March or April (Fig. 2, Table 2). Highest densities in a 3-week period (modal week plus ad- joining weeks) in any year contributed 32-89% of the total density of Atlantic menhaden larvae for the year (Table 2). During the recruitment period, most Atlantic men- haden larvae recruited to the estuary in pulses. Mean density varied from zero to about 570 larvae/100 m3 and most samples contained <20 larvae/100 m3 (Figs. 2 and 3). Catch densities varied among sets on any given night and the standard deviation in catch den- sities generally increased with the mean catch den- Warlen: Spawning time and recruitment of Brevoortia tyrannus 423 CO NOV DEC NOV ' DEC 1?0 100 JAN FEB MAR APR NOV DEC JAN FEB MAR APR APR MAY NOV DEC JAN FEB MAR ' APR Collection Date Figure 2 Weekly mean density (larvae/100 m3) of Atlantic menhaden, Brevoortia tyrannus, larvae in collections at Pivers Island, North Carolina, in the lower Newport River estuary during seven consecutive estuarine recruitment years. 424 Fishery Bulletin 92(2), 1994 sity (Fig. 3). Relative abundance of larvae also var- ied among years (Fig. 2). The sum of weekly densi- ties for individual sampling years differed by more than an order of magnitude (Table 2). Lowest relative abun- dances were observed for the last three sampling years. I assumed that larval Atlantic menhaden caught each week were newly recruited to the estuary and that they were in transit past Pivers Island to upper portions of the estuary. These assumptions are sup- ported by the presence of an abundance mode on 19 December 1985 (Fig. 2) and the presence of similar modes from bongo net sampling in the same estuary one-week later about 6 km up-estuary and two-weeks later about 11 km up-estuary (Warlen, unpubl. data). Also, the generally narrow 95% confidence limits in the age of larvae within each collection, along with no increase in the confidence limits through the re- Table 1 Percentage of the sum of the weekly mean densities (number/100 m3) of larval Atlantic menhaden, Brevoortia tyrannus recruited to the estuary near Pivers Island, North Carolina, November- -May of each recruitment year. Month Recruitment year Nov Dec Jan Feb Mar Apr May 1985-86 0.0 9.4 3.9 6.3 68.8 11.6 0.0 1986-87 1.6 11 5.5 15.1 18.5 55.2 0.0 1987-88 <0.1 <0.1 0.5 6.7 20.8 71.9 <0.1 1988-89 <0.1 5.0 14.4 30.9 32.2 17.4 <0.1 1989-90 1 -1 10.7 12.8 16.9 57.0 1.4 0.0 1990-91 0.0 2.1 7.3 10.0 26.1 54.5 0.0 1991-92 0.5 10.4 8.5 13.5 57.5 9.5 0.1 Mean 0.5 6 ii 7 6 14.2 40.1 31.6 <0.1 Table 2 Sum of weekly mean densities ( number/100 m3 ) by year, relative abun dance (related to 1990-91), and peak recruitment density period (highest three consecutive sampling-date catches includi ng the mode) of larval Atlantic- menhaden, Brevoortia tyrannus, collected at Pivers Island, North Carolina, by recruitment year. Peak recruitment density Sum of weekly Recruitment mean Relative Peri od of highest r; of year densities abundance three consecutive catches year 1985-86 554 4.3 Mar 5-Mar 19 59.9 1986-87 288 2.3 Mar 25-Apr 8 57.7 1987-88 1581 12.4 Mar 30-Apr i:i 88.6 1988-89 459 3.6 Jan 19-Feb 1 32.2 1989-90 17.", 1 1 Mar 7-Mar 21 53.4 1990-91 128 1 n Mar 27-Apr in 75.2 1991-92 163 L.3 Mar 4-Mar 17 56.7 cruitment year (Fig. 4), does not suggest an increase in the number of different birthweek cohorts in the lower estuary. In a number of cases, the week follow- ing a peak in density showed relatively low recruit- ment (Fig. 2), a pattern that did not suggest sub- stantial carryover and accumulation of larvae from week to week. While a single sampling location may not reflect patterns of larval estuarine recruitment for all areas south of Cape Hatteras, it does provide a time-series description of relative larval recruitment abundances over several years inside a large inlet near the presumed major fall/winter spawning area. Age and size of larvae In every sampling year the age of larvae increased linearly throughout estuarine recruitment until about the end of March after which the mean age declined (Fig. 4). Linear regressions of the mean es- timated age over time, excluding the end-of-season down trending values, were significant (ANOVA, P<0.001) for each recruitment year. Young larvae were always collected early in each recruitment year (Fig. 4). Virtually all larvae collected in November were less than 40 days old. Larvae collected about late March were 2—b times older than larvae collected early in the recruitment year. Except in a few cases, the within-sample age variation was small and the 95% confidence limits were within ±5 days of the mean age. Larvae recruited to the estuary during peak recruitment were also the older larvae. Peak recruitment densities were in February -April (Fig. 2) and those larvae were older, generally age 60-90 days or older, as in 1987-88 when some were up to age 115 days (Fig. 4). The recruitment-year mean age of larvae varied between 55 and 74 days for the seven years. The mean age of larvae over all years was 61 days. The standard length of larvae also increased significantly (ANOVA, P<0.001) within each re- cruitment year. The mean size of larvae increased to the end of March then decreased slightly to Warlen: Spawning time and recruitment of Brevoortia tyrannus 425 1000.00- E + § 100.00- * + + 1- »s + + + IS CO * *+S+ CO 100°" + *V* +++ _l ** c o .5 i.oo- > CD *s + t/+ Q + «++ + +4++ + i Jw * .+ + + CO "D 0.10- c + + « + 55 0.01- r i i i i v 0.01 0.10 1.00 10.00 100.00 1000.00 Mean Density (Larvae / 100m3) Figure 3 Plot of standard deviation versus mean density (larvae/100 m3) for each collection night for larval Atlantic menhaden, Brevoortia tyrannus, sampled at Pivers Is- land, North Carolina, for seven consecutive recruitment years. the end of the recruitment year (Fig. 5). Larvae re- cruited to the estuary were always >20 mm SL, ex- cept early in the recruitment year (November-De- cember) when the mean size could be as low as 15 mm (Fig. 5). The overall recruitment year mean SL of larvae varied between 23.2 and 25.4 mm for the seven years and the mean SL over all years was 24.6 mm. Spawning time The menhaden spawning season in North Carolina was estimated from the birthdate distributions of larvae that survived to recruit to the estuary. The percentage distribution of spawning by week (Fig. 6) was based on the relative abundance of larvae col- lected at Pivers Island throughout the recruitment year. The spawning season was protracted, lasting four to six months (Fig. 6). Estimated spawning was variable among birthweeks. In every spawning sea- son, there was a dominant birthweek mode that con- tributed from 25 to 43% of the total estuarine re- cruits. Except for 1987-88 (20 December 1987), these modes occurred very close to the new moon phase of the lunar cycle ( 10 January 1986, 3 1 December 1986, 9 December 1988, 28 December 1989, 15 January 1991, and 6 December 1991). Median spawning and modal spawning peaks were always in December or January as were the second and third quartile inter- vals of the distributions, except in 1989-90 when the second quartile interval extended into November (Fig. 6). Over all seven years, an average of more than 76% of all spawning occurred in December- January (Table 3). Earlier season spawning (Octo- ber-November) contributed an average of 16%> of the total, although November spawning alone could sometimes account for about 26% (Table 3). Little spawning occurred in February ( . 7%) and in only two years did the March contribution exceed 1%. The data suggest that within-season spawning fre- quency may be multimodal. There was an indication in 1985-86, 1986-87, 1989-90, 1990-91, and 1991- 92 that a small, early mode may have occurred in October or November (Fig. 6). If there was an early mode in 1988— 89, it may have blended with the mode for later spawned fish. A small, late season mode ( February or March ) may also have been present and was most evident in the 1986-87, 1988-89, 1989- 90, and 1991-92 spawning seasons. Recruitment of birthweek cohorts Individual birthweek cohorts were recruited to the estuary over periods from one week to several months (Fig. 7). Those birthweek cohorts that recruited over the shortest time periods were generally spawned either early (October) or late (February-March) in the season. Birthweek cohorts spawned from mid- November through January usually recruited over 426 Fishery Bulletin 92(2). 1994 Table 3 Percentage monthly spawning of Atlantic menhaden, Brevoortia tyrannus, estimated from the density-weighted, back-calculated birthdate distributions of larvae recruited to the estuary near Pivers Island, North Carolina, during November-April each recruitment year. Recruitment Spawning month year Oct Nov Dec Jan Feb Mar 1985-86 0.4 12.8 17.3 68.3 1.2 0.0 1986-87 4.2 7.3 64.4 21.1 2.9 0.1 1987-88 0 1 5.9 63.4 30.5 0.1 0.0 1988-89 0.6 26.5 49.2 9.1 12.9 1.7 1989-90 6.6 22.4 31.9 36.4 2.5 0.2 1990-91 0.0 8.4 11.6 54.2 25 M 0.0 1991-92 7.6 9.2 62.3 16.5 2.1 2.3 Mean 2.8 13.2 42.9 33.7 6.8 0.6 the longest time periods. However, 70% or more of any cohort were usually recruited in <3 consecutive weeks. In each recruitment year, those birthweek co- horts with the greatest numbers of individuals con- tributed most to the estuarine recruitment mode (Fig. 7). Also, within the larval catch of any given week, from one to 10 birthweek cohorts were represented (Fig. 7). Discussion Relatively large schools of larger-sized Atlantic men- haden, migrating from New England and mid-Atlan- tic states, along with local (North Carolina at least) Atlantic menhaden emigrating from estuaries, spawn off the southeast Atlantic states. The area from Cape Hatteras to about northern Florida is thought to be the major spawning location for this species (Higham and Nicholson, 1964; Reintjes, 1969; Nelson et al., 1977; Judy and Lewis, 1983; Lewis et al., 1987). Al- though the previous studies suggested that Atlantic menhaden spawn in the fall and winter off the south- east Atlantic states, none were able to estimate within season spawning intensity. The present study, which utilized samples collected throughout the re- cruitment year, estimated within season spawning intensity based on survivors entering an estuary. However, without any knowledge of egg production and the survival of cohorts between offshore spawn- ing and estuarine recruitment, it was not possible to know how closely the estimates of the birthdate dis- tributions represented the actual seasonal egg pro- duction. In each year, the Atlantic menhaden spawning sea- son was protracted. This long spawning season might indicate a "bet hedging" strategy (Lambert and Ware, 1984) where eggs are released continuously over the spawning season to ensure some reproduction during the most favor- able periods for survival. Based on the birthdate distributions of larvae that recruit to the estuary, the most favorable period each year appears to be for those fish spawned in a rela- tively short period between early December and mid-to-late January. For all seven years, it was within this major spawning period that the week of peak spawning occurred. Larvae from this major spawning period are probably progeny of the large men- haden schools migrating southward from the New England and mid-At- lantic states. These schools, which contain many fish of spawning age (3+ years), are harvested from about late November to late January during the North Carolina fall purse seine fishery (Smith et al., 1987). Larvae spawned earlier (October-November), may originate from adults inhabiting North Carolina coastal waters and estuaries in the summer or from early fall adult immigrants to North Carolina wa- ters (Wilkens and Lewis, 1971). Larvae spawned late (February-March) may have been spawned further south and immigrated to the estuary late in the sea- son or were the offspring from northward migrating adults in early spring. This late spawned group con- tributed the younger, smaller larvae observed at the end of the recruitment year (Figs. 4 and 5). Spawning locations, and routes and rates of trans- port probably account for the variation in age of lar- vae recruited to the estuary. While the precise loca- tions of menhaden spawning are not known, the gen- eral area is thought to be on the mid- to outer-conti- nental shelf off North Carolina (Checkley et al., 1988; Warlen, 1992). However, accounts of the oc- currence of eggs are limited (Reintjes, 1969; Judy and Lewis, 1983) and no records of actual spawning events exist. Although some early season spawning may occur closer to shore, the largest contribution of recruited larvae to the estuary is later in the season, probably from the warm, plankton rich areas closer to the Gulf Stream. Water temperatures are gener- ally >18°C (Govoni, 1993), even in winter. These warm temperatures are due, in part, to intrusions of warm surface water onto the middle continental shelf (Atkinson, 1985). Frequent upwelling events (Pietrafesa et al., 1985) stimulate primary produc- tivity by providing nutrients (Atkinson, 1985; Yoder, 1985) for phytoplankton growth (Yoder et al., 1983) Warlen: Spawning time and recruitment of Brevoortia tyrannus 427 1 985-86 4 0 201 1 20 60 60 "D w «0 O O) < 20 T3 O 120 w — too (0 LU 80 6 0 4 0 20 1 20 1 00 80 6 0 A 1 986-87 ^r AT *Hs \ No> u. i.i. w» 1 987-8f 1988-89 4 0 20 1 20 1 00 1 989-90 1 990-9 1 ..*' 20 J, 1 20 199 1-92 Collection Date Figure 4 Mean of estimated ages (days) with 95% confidence limits of larval Atlantic menhaden, Brevoortia tyrannus, collected weekly over seven years during their estuarine recruitment at Pivers Island, North Carolina. thus increasing secondary production (Paffenhofer, 1985). This productive area is probably utilized by larval fishes spawned nearby. Atlantic menhaden appear to spawn during peri- ods (daily, seasonal) of lower light intensity. They probably spawn at night, and may have a diel pat- tern of ovulation and spawning that may be linked to the daily cycle of light and darkness as noted for many marine fishes (Bye, 1990). Reintjes ( 1969) ex- amined Atlantic menhaden eggs at sea and concluded that estimated ages of groups of eggs of different stages suggested that "spawning occurred after mid- 428 Fishery Bulletin 92(2). 1994 5 1 5 a> C 10 9 — ' 30 CO CD CO 1 985-86 r N»* 0*0 1 986-87 1 987-88 1 988-89 / • ■■ y*> Ap< HW st. -I «. I 1 0 3 0 1989-90 No* O.c 1 991-92 «(,. . M«| -r. M« . Collection Date Figure 5 Mean standard length (mm) with 95% confidence limits of larval Atlantic menhaden, Brevoortia tyrannus, collected weekly over seven years during their estuarine recruitment at Pivers Island, North Carolina. night but before dawn on three consecutive days." Ferraro (1981) also noted that Atlantic menhaden spawned at night and suggested that night spawn- ing might be a means of reducing predation on spawn- ers and eggs or of avoiding the deleterious effects of ultraviolet irradiation during early embryogenesis. The relatively constant year-to-year spawning mode during December-January, observed in this study, occurs when the daily hours of darkness are maxi- mal, 13V2-14 hours between sunset and sunrise, and when overcast days are more frequent. Peak spawn- ing in each year (except 1987-88) also occurs very Warlen: Spawning time and recruitment of Brevoonia tyrannus 429 t086-ae N=446 _£ziZrZ2zz2. =l=i- 1900-00 n-4 ia J 1: Q> 20 ffi «J o c 448 PT7fZ?73ryJ%4'727: X^ 1967-68 N=593 ii . .fa 7/y -f7-r.ru 1968-69 N=534 4 mar *pq 20 ' >0 20 ' 10 20 ' 10 20 FEB MAR Back-Calculated Spawning Date Figure 6 Birthdate frequency distributions calculated from the relative abundances of larval Atlantic menhaden, Brevoortia tyrannus, collected in the lower estuary at Pivers Island, North Carolina, in each of seven recruitment years. In each distribution, the vertical line is the median value and the box represents the 2nd and 3rd quartile intervals (central 50%). Lines beyond the boxes represent the range of data. close to the new moon phase of the lunar cycle, which would also reduce available light during spawning. Spawning of Atlantic menhaden off the southeast- ern Atlantic states is apparently timed to ensure transport of larvae across the continental shelf and arrival of most of the larvae (about 85% on average) during a time of optimal survival conditions in the estuary. Atlantic menhaden larvae that recruit to the estuary during the peak period (February-April) arrive when the water temperature is rising, prey abundance is high, and predator abundance and es- tuarine resident larval fishes and invertebrate ( cteno- phore) competitor abundances are low (Warlen and Burke, 1990). Some Atlantic menhaden larvae (about 15%) recruited to the estuary early in the season, November— December, before the period of coldest water temperatures in January (Warlen and Burke, 1990). These early recruited larvae may experience cold-related mortality if estuarine water tempera- tures drop to <4°C (Lewis, 1966; Wilkens and Lewis, 1971). However, larval Atlantic menhaden may tol- erate lower temperatures than larvae of other spe- cies (e.g. spot, Leiostomus xanthurus) that recruit over the same period in the Beaufort area (Hoss et al., 1988). In milder winters, earlier recruited lar- vae may survive overwinter and, in the following spring and summer, may be larger than the more abundant Atlantic menhaden larval groups recruited several months later. Ahrenholz et al. ( 1989) observed a multi-modal distribution in lengths of juvenile menhaden collected in the summer in North Caro- lina which may reflect several seasonal abundance groups of immigrating larvae (early, middle, late) from a single spawning season. Variation in egg production, mortality of eggs and larvae, and losses of larvae by advection to other ar- eas of the coast probably account for the observed differences in relative recruitment abundances at the Pivers Island collection site. There was more than an order of magnitude difference between the rela- tive abundances of the most abundant and the least abundant years. Lewis and Mann (1971) also ob- 430 Fishery Bulletin 92(2). 1994 1 <0 1* .'.' 1985-86 OH 15 0 6. H 1 0 4»— 2 2' — .'5 7 6. • 1 IS 1 1 43 0- — • ■ 4 28 7 7. «" 2 I 1 4 1 1. • 7 05.- 30 23 16 9 1 9— • — • 2 0 4» ■ 59 9 NOV DEC JAN FEB MAR APR a N 0 3 — 0 3« 0 9« 1 5» . 0 6« • .C o O CD 0 3- 2 9- 1 1 5 — I I a 25 7- 1 I 2' 13 2- 8 3 • 3 2' • 1 6— • — 0 S-»- 5? ! NOV DEC JAN FEB MAR APR <0 !• 1987-88 <0" <0 !• I 1 3. •— II 6. 4M 3 9. — 5 T •— 9 4- I36>- 25 4 — 12 3. 3 6' 0 2- •- 0 I — <0 !-• NOV DEC JAN FEB MAR APR 2 5 <0 !• J 1 1 1988- BM 0 1 • 05«- 4 1 7 *~ 25 i . -♦- 19 gll 5 6 5 8 - 3 B»- 29 ■ 0 5-»- 22 <0 - 28 -7 4 — 0 Pi ■ — • — .0 4 53 4 NOV DEC JAN FEB MAR APR 23 - 6.8 • 16 9 1990 9 9 1« 6 8» 2 10 0 •— • 26 40 0 —4* 19 - 4 9' — • 12 - 6 •- 5 29 22 " 0 7 — ■ 15 2 1 ' • a 5 0 —• i 1 3 i • 24 1 7 1 7 2 e -- 10 1 6' , — . «■■ , NOV DEC JAN FEB MAR APR Collection Date 28 <0 1» J* At>j&> Cfl&> O&tft* \0\& ^6^J* tOyjP ^^^Jjl* -O*j0>* O^o*6 0*qJ* ^ \S* ^ \J* »*> vO*> I30 CO 30 1 I 20 - " 20 " 1 CO 10 1 . 1 10 1 1 X || .1 0 40 30 J ill- 0 40 30 Jl •^OfcT* -O&oJfr CP 036 06 ttf« aO J& \6 ,1* tP yjP ^Oo?* J&jtft> Ol>o3* 0*tf6 ^\3* v* vl* "ifiiOfi - i 20 _ 20 . 1 M 10 I ll.l 10 -.1 ^DqJ* xPj&> rf>o* O&oJ* ^\Sfr *&Vl* &-aS> ^<&X0* J&X& OPo* 0&©J* ^ ** ^ ^6 1*> *«$ 30 1 1 1 30 I 1 1 1 1 f" 20 - 20 - i ■ X 10 1 III, , . j_^ 10 . 1 ^aOjjI* -O^Jia* 0^03* C*oJ* AD ^afc \b ,1* & JlP ^*>o.1* -Ofto** C*oS* C*olfe ^ \8* ^ \J* tf> VJ0 Canonical Axis 1 Canonical Aids 2 Figure 4 Frequency histograms, transformed to percentages, of the first two environ- mental axis scores in areas where cetaceans were sighted for seven species/stocks. P-value < 0.001. The same was true for the first axis alone; no random permutation had an eigenvalue larger than the observed 0.309, again giving a P-value < 0.001. These results indicate that the prob- ability of a Type-I error is less than 0.1%. (The CCA program, CAN- OCO, provided this test only for the trace and first axis, so no test was done for subsequent axes). The species-environment biplot (Fig. 5A) displays the results for the six variable 'oceanographic' ordina- tion. Fig. 5B shows the ordination with species tolerances, but without the visual distraction of the environ- mental vectors. The first axis sepa- rates common dolphins from all school types containing spotted and spinner dolphins. Positive scores on axis 1 are associated with cooler temperature (r=-0.88, Table 3), a thermocline that is shallower (i.e. smaller Z20, r=-0.70), yet weaker (larger difference in depth between 20° and 15°C isotherms, r =+0.45), denser surface water (higher sigma- t, r=+0.61) and high chlorophyll (r=+0.49). These are characteristics of "cool upwelling" habitat, as found in Equatorial and Peru/California Current surface waters. The dis- tinct placement of common dolphins in the positive region of this axis indicates this is their preferred habitat. Negative scores on axis 1 are associated with warm tempera- ture, a deeper and stronger ther- mocline, and lower chlorophyll, as found in less productive Tropical Surface Water. The placement of all spotted and spinner school types in the negative region indicates that these oceanographic conditions help define their preferred habitats. Site or species scores on axis 2 are uncorrelated with scores on axis 1, by definition. Positive axis-2 scores are associated with a relatively shallow thermocline (r=-0.63, Table 3 ) and high chlorophyll (r=+0.40) as for axis 1, but also with warmer temperatures and lower salinity (lower sigma-t) rather than cool 442 Fishery Bulletin 92(2). 1994 30 S 20 10 - 40 30 10 ii temperature. These are character- istics of "coastal tropical" habitat found along the coast of Central America, where the surface layer is more stratified and upwelling is more intermittent and localized than in the cool upwelling habitat. Whitebelly spinners, alone and with spotted dolphins, had large negative axis-2 scores, while eastern spin- ners with spotted dolphins had posi- tive scores. There was a strong sepa- ration on axis 2 between mixed schools of whitebelly spinners with spotted dolphins, and schools of eastern spinners with spotted dol- phins. Schools of spotted dolphins alone had near-zero axis 2 scores. Striped dolphins loaded near the origin of both axes. The spatial distributions of yearly axis 1 scores are mapped in Figure 6. Areas with positive scores are shaded to allow quick appraisal of changes between years (interpreted below). Also plotted in Figure 6 are sighting localities for spotted and common dolphins. Spotted dolphins occurred mostly in negative areas, common dolphins in positive areas, but with some overlap. Spatial distributions of yearly axis 2 scores are mapped in Figure 7, with positive scores shaded. Whitebelly spinners occurred al- most exclusively in waters with negative axis-2 scores (Figs. 5 and 7). Eastern spinners ranged throughout both positive and nega- tive areas; a modest majority were found in positive areas. They were less closely associated with this axis than whitebelly spinners (Fig. 5) and seemed to be found in the eastern (more coastal) part of the warm tropical habitat defined by negative axis-1 scores. Interannual variability We obtained only a slight increase in the percent of variance explained for the dolphin data (14.7% to 15.1%, Table 5) from addition of categorical variables representing the five sampled years, in addition to the six oceanographic variables. An ordination biplot from this analysis (Fig. 8) shows that the centroid for 1988 (year 3) loads farthest from the origin. Its T 1 1 1 1 1 1 1 1 I ' I I- llL. -I 1 1 1 1 1 1 1 1 1 1 1 1- A OjjJ* j&j&> Kp<&> 0& <&&*> O&tfl* ^ ,3& \t tfb oD ,/jP 40 l — i — l — l — l — l — l — i — i — i — i—i — i — r- lllW - 10 - -I 1 1 1 1 1 1 1 1 1 1 T- _J 1 J 1_ lO^Tft j&j&t oP O&ftT* ^il* ^\1* tP'AP ^■*D^1* -0»c36 OPqZ6 O^tf* ^ \lb ^ \16 1? JP \ii> % 20 O T — I — I — I — i — i — I — l — i — r- .ulllllll 30 i t i i i i—i r i 20 - 10 \ 1 ...1 1 ■ ■ . . Qj&* -0&o3* &<& 0»oJ« *0 ,20 ^,1* 10^0 „^°.Ol* J&X& 0°026 O&tf6 \0 ,j6 \fi \jt tfi j& Figure 4 (Continued) location represents the cooler, more productive con- ditions associated with the 1988 La Nina. An analysis including just years as categories, without oceanographic variables, explained only about 2% of the dolphin variance, but the dominant eigenvalue and trace were both significantly differ- ent from zero (Monte Carlo P-values=0.01 and 0.02, respectively). After extraction of the variance asso- ciated with the six oceanographic variables (by de- fining them as covariables) the ordination was not significant (Monte Carlo P=0.058, Table 5). Reilly and Fiedler: Interannual variability of dolphin habitats 443 Fixed geographic effects Inclusion of latitude and longitude in addition to the six oceanographic variables produced a notable in- crease in dolphin variance explained, from 14.7% to 20.5% (Table 5). The amount of additional influence indicated for fixed geographic effects varied substan- tially among school types. The largest increases were for whitebelly spinners, alone and with spotted dol- phins. No improvement in explaining variance was made for schools of spotted dolphins alone. Group size effects "15POT&ESPIN QWB SPIN SPOT&WBSPtNQ B -t V h This study used encounter rates as an index of abundance. This index does not encompass effects of vary- ing group size. There is some evi- dence for geographic patterns in group size for the dolphin school types studied ( Gerrodette and Wade, 1991), so the analyses reported here were also run with the dependent variables modified as follows. En- counters with schools were weighted by the number of individuals esti- mated to be in the school. The weighted rate was then log-trans- formed. Canonical correspondence analyses run with these modified dependent variables produced essen- tially the same patterns as before, but with a small loss of explanatory power: the cumulative percent of the species variance explained was 14.2%, down from 14.7%. Discussion Species-environment patterns The ordination results were gener- ally consistent with past studies of Figure 5 Ordination results from canonical corre- spondence analysis of cetacean species/ stocks and environmental conditions in the eastern tropical Pacific. (A) Biplot of first two canonical axes and environmen- tal variables. (B) Ordination showing 95% confidence limits for the species loadings. The environmental variables, represented by arrows in 5A, are surface temperature (TEMP), surface salinity (SAL), thermocline depth indexed by 20°C isotherm depth (Z20), thermocline strength, indexed by the difference in depth between the 20°C and 15°C isotherms (ZD), surface water density (SIGMAT), and sur- face chlorophyll, log-transformed (LOGC). These two axes represent 94% of the spe- cies-environment variance, 15% of the to- tal encounter rate variance. 444 Fishery Bulletin 92|2). 1994 ETP cetacean ecology. The first axis contrasts the habitat use of common dolphins with spot- ted and spinner dolphins. The placement of com- mon dolphins into cool upwelling habitat is con- sistent with results reported by Au and Perryman (1985) and Reilly (1990). The place- ment of spotted and spinner dolphins in con- trasting habitat (negative axis-1 values; essen- tially warm tropical water) is also as reported earlier. Consistency with results of Reilly ( 1990) is not surprising, because that study shared data from 1986 and 1987 with this study, but is somewhat reassuring because different analyti- cal techniques were used. The second axis separated eastern spinners from whitebelly spinner dophins. This separa- tion was even clearer between sightings of east- ern spinners mixed with spotted dolphins and whitebelly spinners mixed with spotted dol- phins. The ordination placement of whitebelly spinner dolphins in habitat with a deeper ther- mocline ( negative axis 2 ) follows from their more offshore distribution and the general tendency for the thermocline to become deeper to the west in the ETP. Spotted dolphins alone occurred intermediate to these mixed schools. If this is a general pattern it suggests that the two mixed- school types of spotted and spinner dolphins are utilizing habitats as different as those used by separate species (e.g. common dolphins and spotted dolphins on axis 1). These results are consistent with the hypothesis that the morpho- logical distinctness of the endemic eastern spin- ner dolphin subspecies reflects adaptation to local habitat conditions (Dizonetal., 1991), and the recent finding that spotted dolphins north of the equator and east of long. 120°W, i.e. those available to school with eastern spinner dol- phins, comprise a distinct 'stock' (Dizon et al., in press). The ordination of striped dolphins near the origin of both axes 1 and 2 indicates either that this is near their optimum habitat or that their distribution is unrelated to the environmental patterns represented in the canonical axes. The low "R2" for striped dolphins (Table 4), and their wide- spread spatial distribution (Fig. 1; Reilly, 1990) sup- port the latter interpretation. The species-environment correlations observed were quite high: 0.67 for the first species and envi- ronment axes, 0.42 for the second axes. However, variation extracted by the canonical correspondence analysis accounted for just 15% of the total encoun- ter rate variance. (This was increased to over 20% when fixed geographic effects were considered. ) This CANONICAL AXIS 1 ► Common dolphin ^Spotted dolphin Figure 6 Maps of distribution of canonical axis 1 for 1986-90. Positive areas are shaded. Spotted dolphin, Stenella coeruleoalba, sighting localities are shown as open triangles, common dol- phin, Delphinus delphis, localities as closed circles. A"+" rep resents a sighting day during which neither spotted nor com- mon dolphins were seen. modest explanatory power is in fact fairly good, given the unknown but surely large sampling variability inherent in daily encounter rates, and is consistent with levels of explanatory power in similar CCA analyses of abundance data (e.g. Ter Braak, 1986). Dolphins are very mobile and patchily distributed large predators, and are known to have complex so- cial and behavioral interactions with their own and other species. These characteristics combine to pro- duce highly variable abundance indices. Reilly and Fiedler: Interannual variability of dolphin habitats 445 CANONICAL AXIS 1 1990 120 110 100 90 ftO • Common dolphin ^Spotted dolphin Figure 6 (Continued) Explanatory power for common dolphins was surprisingly high: 36% with the six oceano- graphic variables, and 42% with fixed geogra- phy included. This result indicates that com- mon dolphins have the tightest association with the environmental variables of the seven school types studied here. It also demonstrates the ro- bustness of CCA, considering the bimodal distri- bution of common dolphins on axis 1 (Fig. 4). The notable increase in performance for whitebelly spinner dolphins resulting from con- sideration of fixed geography raises interesting questions. Are they directly responding to some geophysical cue, such as magnetic anomalies (Kirschvink et al., 1986; Klinowska, 1985)? Or, does this result simply reflect orientation to oceanographic features or processes (e.g. prey distribution) not represented in our data? Interannual variability Interannual environmental variability is appar- ent in the geographic distributions of the ca- nonical axis scores, and to a lesser extent in lo- cations of dolphin sightings (Figs. 6 and 7). In 1986, cool upwelling habitat was found along the equator to long. 130°W, north of the equa- tor to about lat. 10°N along the coast of Central America, and off the coast of Baja California. In 1987, cool upwelling habitat south of Baja California did not extend west of 110°W or north of 4°N, except in the Gulf of Panama. The study area was dominated by warm, low-productiv- ity tropical water (negative axis-1 scores). This change was caused by a moderate El Nino event that began in late 1986 and lasted through 1987 (Kousky and Leetmaa, 1989). In 1988, cool up- welling water extended far north of the equa- tor and south of Baja California, considerably reducing the area covered by tropical water. 1988 was a strong anti-El Nino or La Nina year (Leetmaa, 1990; Fiedler et al., 1992). In 1989 and 1990, conditions represented by axis-1 scores returned to a state similar to 1986. Interannual variation along axis 2 was not strongly related to ENSO variability. The area with positive axis-2 scores ("coastal tropical" habitat) was small in 1986 and 1987, increased in 1988 and again in 1989, and showed some diminishment in 1990. Common dolphin distribution was previously observed to show no apparent seasonal changes (Reilly, 1990) but was observed here to change interannually more than the other school types studied, and these changes appear related to 446 Fishery Bulletin 92|2), 1994 ENSO variability. In 1987, with "cool upwelling" con- ditions contracted eastward and southward at the equator as part of that year's El Nino, in the south- ern ETP (south of about 2°S) common dolphins oc- curred only in the far east off South America. In 1988 during the strong La Nina these conditions were well established along the equator to the western extent of the study area, and common dolphins occurred in equatorial waters as far west as 110°W. The maps in Figures 6 and 7 are imprecise repre- sentations of species-environment patterns derived by CCA for two reasons. First, the maps show only presence-absence, while we used an effort-corrected index of abundance (daily encounter rate) in the CCA. Second, the contouring requires some smoothing and interpolation between sites, while the CCA compared abundance indices only to environmental variables measured during the same day, along the same track lines where the cetaceans were sighted. These spe- cies are apparently separating more strongly on a smaller scale than we could effectively represent on the maps. A further consideration is that the scaling of axes for biplot presentation was done by using a method in which the canonical scores (as plotted on the maps here) are rescaled to produce biplot loca- tions (Ter Braak, 1988). The resulting ordination gives an accurate relative placement of species centroids, but does not allow direct projection of centroids or toler- ances onto canonical axis values as mapped in Figures 6 and 7. The small but significant interannual variation in the species data was effectively accounted for by interannual variation in the environment. This was demonstrated by the low eigenvalue (^.=0.02, P=0.06) for interannual differences after extracting variance associated with the six oceanographic variables. This result does not necessarily apply to total population abundances, however, because in the above analy- ses we did not include school size estimates in our species data. Group size effects Inclusion of group-size data in the dolphin abundance index produced ordinations that were very similar to those using simple encounter rates, but with a slight decrease in explanatory power from the envi- ronmental data. We interpret this result to indicate that schools of all sizes occupy approximately the same habitats, and that school size variability within these habitats is not strongly related to the environ- mental variables analyzed here. Applications to dolphin assessments We suggest two approaches to use the results of this study in cetacean abundance and trend monitoring. Table 5 Comparative ordinations from canonica correspondence analyses of seven types of dolphin school in the eastern tropical Pacific, with six different sets ol environmental variables . Set 1 = surface temperature (SST), thermocline depth (Z20) and thermocline strength (ZD). Set 2 = SST, Z20, ZD, surface salinity (SAL), surface chlorophyll (LOGO and surface density (SIGMA-T). Set 3 = Set 2 plus years (1-5) as categorical variables. Set 4 = Set 2 plus latitude and longitude. Set 5 = Set 2 plus both latitude & longitu de and years. Set 6 = years (1-5) as categorical variables. after removing variance associated with all other environmental variables (Set 5). Environmental variable set 1 2 3 4 5 6 Eigenvalue sum 0.384 0.443 0.464 0.622 0.644 0.022 P-value <0.001 <0.001 <0.001 <0.001 <0.001 0.058 Percent variance accounted for total species data 12.8 14.7 15.1 20.5 21.1 0.9 spotted dolphin' 7.8 8.1 8.8 8.8 9.6 0.8 common dolphin1' 32.6 35.5 36.2 41.2 42.2 1.0 spottted and eastern spinner dolphins ' 18.8 22.5 22.8 25.0 25.3 0.3 spotted and whitebelly spinner dolphins1' 9.2 9.7 9.9 16.4 17.3 1.0 eastern spinner dolphin 7.5 8 3 9.5 10.4 11.2 0.9 whitebelly spinner dolphin 3.6 5 1 6.4 20.0 20.3 0.3 striped dolphin"* 1 9 5.9 6.6 12.6 13.8 1.9 ' Stenella attenuata. '-' Delpkinus delpkis. ■? S. longirostris. 4 S. coeruleoalba. Reilly and Fiedler. Interannual variability of dolphin habitats 447 Other, perhaps more sophisticated approaches are possible. We present these only as examples. The most straightforward approach, involving minimal assumptions, would be to post-stratify the data for each year separately, based on the spatial distribu- tion of CCA axis scores and the weighted mean and standard deviation of those scores for the species of interest. This would be done to improve precision of abundance estimates. Populations that have similar means and standard deviations could use common strata. For example, separate strata could be defined by using axis 1 for common and spotted dolphins. CANONICAL AXIS 2 130 120 1 10 100 »0 80 • White-belly Spinner dolphin ^Eastern Spinner dolphin Figure 7 Maps of distribution of canonical axis 2 for 1986-90. Nega- tive areas are shaded. Eastern spinner dolphin, Stenella longirostris, sighting localities are represented by closed circles, whitebelly spinner dolphin, S. longirostris, localities by open triangles. Axis 2 could be used to provide strata for whitebelly spinner dolphins. Because we have probability dis- tributions for the occurrence of these species along the canonical axes, we would not be limited to use just two strata but could use three or four. After the data were stratified based on the species annual habi- tat distributions, standard line transect methodol- ogy would be followed. This is generically similar to the post-stratification approach taken by Anganuzzi and Buckland (1989) to reduce bias in estimates of dolphin abundance from tuna vessel observer data. A second possible approach, aimed at improving accuracy, would quantify the amount of habi- tat available within the study area each year, for each population. The simplest quantifica- tion scheme would define only two strata for each. The cut-point between strata could be the 95% limit of the population's distribution on the axis, or, less conservatively, the appropriate upper or lower quartile. More complex schemes using more than two strata could be developed, as with the post-stratification, based on addi- tional information in the species probability dis- tributions. The amount of any stratum avail- able in a year could be quantified by, say, lightly smoothing and interpolating the CCA site scores (to provide values for all locations) and "sampling" the distribution with the actual cruise tracks for the year. If for example com- mon dolphin habitat was to be defined as axis 1 > [some value], the amount of ocean sampled with axis 1 > [some value] in 1986 could be scaled as 1.0. The amount sampled in subse- quent years could be scaled to the 1986 amount. The result would be a vector of values repre- senting the amount of common dolphin habitat available within the ETP by year. This vector could then be applied to the encounter rate por- tion of the line transect abundance estimate for each year to account for changing availability of common dolphin habitat. If interannual dif- ferences were subsequently observed in the line transect abundance estimates, we could be more confident that they represent real changes in abundance, rather than just apparent changes due to spatial redistribution relative to sam- pling effort following habitat shifts. In a separate study (Fiedler and Reilly, 1994) we applied the CCA ordination approach devel- oped here to investigate interannual variabil- ity in abundance indices for ETP dolphins esti- mated from tuna vessel observer data. We cal- culated annual indices of habitat quality for each dolphin species targeted by the tuna fish- ery, for the years 1975-90, then compared these 448 Fishery Bulletin 92(2), 1994 habitat indices to Anganuzzi et al.'s ( 1991) abun- dance estimates. We used a subset of three en- vironmental variables from those used here, to enable use of existing, large data bases on oceanography of the ETP, to allow computation of environmental axes for years prior to 1986. We found that, for some species, environmen- tal variability does appear to influence abun- dance estimates made from tuna vessel observer data. We are now working on using environmen- tal data to reduce error in dolphin abundance estimates derived from both research vessel and tuna vessel sightings data. Gerrodette et al.1 applied the results of this study in a prelimi- nary attempt to account for movements in and out of the study area when estimating total abundance of dolphins. Acknowledgments The efforts of many people were required to col- lect and process the data used in this analysis. Cetacean sightings were made and recorded by S. Beavers, S. Benson, C. Bisbee, P. Boveng, K. Brownell, S. Buckland, J. Caretta, J. Cotton, A. Dizon, G. Friedrichsen, S. Hill, A. Hohn, W. Irwin, A. Jackson, S. Kruse, C. LeDuc, R. LeDuc, M. Lynn, M. Newcommer, R. Pitman, J. Raffetto, K. Rittmaster, L. Robertson, R. Rowlett, S. Sinclair, D. Skordal, B. Smith, P. Stangl, V. Thayer and M. Webber. Oceanographic data were col- lected by J. Echols, J. Ellingson, J. Fleishman, L. Gearin, L. Lierheimer, B. McDonald, D. Niemer, V. Philbrick, B. Tershey, V. Thayer, G. Thomas and S. Strickland. C. Oliver, R. Holland, A. Jack- son, R. Rasmussen and K. Blum contributed data editing and processing. Holland contrib- uted computer programming for the analyses and produced the figures. We thank these people, plus the officers and crew of the NOAA research vessels David Starr Jordan and McArthur for their contributions. Helpful com- ments on various drafts were made by D. Au, J. Barlow, L. Ballance, D. DeMaster, W. Perrin, W. Perryman, P. Wade and two anonymous re- viewers. Doug DeMaster provided enthusiastic support and encouragement during the course of this study. CANONICAL AXIS 2 1 Gerrodette, T., P. C. Fiedler, and S. B. Reilly. 1991. Including habitat variability in line transect estimation of abundance and trends. NOAA-NMFS Admin. Rep. LJ- 91-37. 1988 10-1 r |» 120 110 '00 1990 130 120 110 100 W 80 • While-belly Spinner dolphin a Eastern Spinner dolphin Figure 7 (Continued) Reilly and Fiedler: Interannual variability of dolphin habitats 449 QSP0T4ESPIN QE SPIN ^_TEMP LOGC Q COMMON 1 i i * "-» 1989 ^S^ 1990 ^^ 1 ' ■' Zf SPOT ^r >Xi987 ^X^STRIPED . '988 \p^^ •* Flxis 1 ^ SAL ^s. J»* SIGUAT 4*120 WB SPIN/~> u o SPOTAWBSPIN Figure 8 Ordination biplot from canonical correspondence analy- sis of seven dolphin school types in relation to six oceano- graphic variables (see Fig. 5 for definitions), with the five sampled years as categories. Mean values for each year are represented by triangles, YEAR 1 = 1986, YEAR 2 = 1987, etc. Literature cited Anganuzzi, A. A., and S. T. Buckland. 1989. Reducing bias in trends in dolphin abundance, estimated from tuna vessel data. Rep. Int. Whal. Comm. 39: 323-334. Anganuzzi, A. A., S. T. Buckland, and K. L. Cattanach. 1991. Relative abundance of dolphins associated with tuna in the eastern tropical Pacific, estimated from tuna vessel sightings for 1988 and 1989. Rep. Int. Whal. Comm. 42:497-506. Au, D.W. K., and W. L. Perryman. 1985. Dolphin habitats in the eastern tropical Pacific. Fish. Bull. 83:623-643. Blackburn, M., R. M. Laurs, R. S. Owen, and B. Zeitschel. 1970. Seasonal and areal changes in standing stocks of phytoplankton, zooplankton and micronekton in the eastern tropical Pacific. Mar. Biol. 7:14—31. Carr, A. 1987. New perspectives on the pelagic stage of sea turtle development. Conserv. Biol. 1(2):103-121. Chavez, F. P., and R. T. Barber. 1987. An estimate of new production in the equato- rial Pacific. Deep-Sea Res. 24:1229-1243. Dizon, A. E., S. O. Southern, and W. F. Perrin. 1991. Molecular analysis of mtDNA types in ex- ploited populations of spinner dolphins (Stenella longirostris). In R. Hoelzel (ed.), Genetic ecology of whales and dolphins. Rep. Int. Whal. Comm., Spec. Issue 13:183-202. Dizon, A. E., W. F. Perrin, and P. A. Akin. In press. Stocks of dolphins (Stenella spp. and Del- phinus delphis) in the eastern tropical Pacific: a phylogeographic classification. NOAA Technical Rep. NMFS. Donguy, J. R., and G. Meyers. 1987. Observed and modelled topography of the 20°C isotherm in the tropical Pacific. Ocean- ologicaActa 10:41-48. Enfield, D. B. 1989. El Nino, past and present. Rev. Geophys. 27:159-187. Evans, W. E. 1975. Distribution, differentiation of populations, and other aspects of the natural history of Delphi- nus delphis Linnaeus in the northeastern Pacific. Ph.D. thesis, Univ. Calif. Los Angeles. Fiedler, P. C. 1992. Seasonal climatologies and variability of east- ern tropical Pacific surface waters. NOAA Tech. Rep. NMFS 109, 65 p. Fiedler, P. C, and S. B. Reilly. 1994. Interannual variability of dolphin habitats in the eastern tropical Pacific. II: Effects on abun- dances estimated from tuna vessel sightings, 1975- 1990. Fish. Bull. 92:451-463. Fiedler, P. C, L. J. Lierheimer, S. B. Reilly, S. N. Sexton, R. S. Holt, and D. P. DeMaster. 1990. Atlas of eastern tropical Pacific oceanographic variability and cetacean sightings, 1986- 1989. NOAA Tech. Memo. NMFS-SWFSC-144. Fiedler, P. C, F. P. Chavez, D. W. Behringer, and S. B. Reilly. 1992. Physical and biological effects of Los Ninos in the eastern tropical Pacific. Deep-Sea Res. 39(21:199-219. Gauch, H. G., Jr. 1982. Multivariate analysis in community ecology. Cambridge Univ. Press. Gerrodette, T., and P. R. Wade. 1991. Monitoring trends in dolphin abundance in the eastern tropical Pacific: analyses of 1989 data. Rep. Int. Whal. Comm. 41:511-515. Haury, L. R., J. A. McGowan, and P. H. Wiebe. 1978. Patterns and processes in the time-space scales of plankton distributions. In J. H. Steele (ed.), Spatial pattern in plankton communities. Plenum Press, New York. Holt, R. S. 1987. Estimating density of dolphin schools in the eastern tropical Pacific Ocean by line transect methods. Fish. Bull. 85:419-434. 450 Fishery Bulletin 92(2). 1994 Holt, R .S., and S. N. Sexton. 1990. 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. Fish. Bull. 88: 105-111. Holt, R. S., T. Gerrodette, and J. B. Cologne. 1987. Research vessel survey design for monitoring dolphin abundance in the eastern tropical Pacific. Fish. Bull. 85:435-446. Hope, A. C. A. 1968. A simplified Monte Carlo significance test procedure. R. Statist. Soc. J. Ser. B 30(3):582-598. King, F. D. 1986. The dependence of primary production in the mixed layer of the eastern tropical Pacific on the vertical transport of nitrate. Deep-Sea Res. 33(6):733-754. Kirschvink, J. L., A. E. Dizon, and J. A. Westphal. 1986. Evidence from strandings for geomagnetic sensitivity in cetaceans. J. Exp. Biol. 120:1-24. Klinowska, M. 1985. Cetacean stranding sites relate to geomag- netic topography. Aquatic Mammals 1:27-32. Kousky, V. E., and A. Leetmaa. 1989. The 1986-87 Pacific warm episode: evolution of oceanic and atmospheric anomaly fields. J. Clim. 2:254-267. Leatherwood, S., R. R. Reeves, W. F. Perrin, and W. E. Evans. 1982. Whales, dolphins and porpoises of the eastern North Pacific and adjacent waters, a guide to their identification. NOAA Tech. Rep. NMFS Circ. 444. Leetmaa, A. 1990. The interplay of El Nino and La Nina. Oceanus 32:30-34. McCreary, J. P., H. S. Lee, and B. B. Enfield. 1989. The response of the coastal ocean to strong offshore winds: with application to circulations in the Gulfs of Tehuantepec and Papagayo. J. Mar. Res. 47:81-109. Perrin, W. F. 1969. Using porpoise to catch tuna. Wld. Fish. 18(61:42-45. 1975. Distribution and differentiation of populations of dolphin of the genus Stenella in the eastern tropi- cal Pacific. J. Fish. Res. Board. Can. 32:1059-1067. 1990. Subspecies of Stenella longirostris (Mammalia: Cetacea: Delphinidae). Proc. Biol. Soc. Wash. 103(2):453-463. Perrin, W. F., M. D. Scott, G. J. Walker, F. M. Ralston, and D. W. K. Au. 1983. Distribution of four dolphins (Stenella spp. and Delphinus delphis) in the eastern tropical Pacific. NOAA Tech. Memo. NMFS-SWFC-38, 65 p. Perrin, W. F., M. D. Scott, G. J. Walker, and V. Cass. 1985. Review of geographical stocks of tropical dol- phins (Stenella spp. and Delphinus delphis) in the eastern tropical Pacific. NOAA Tech. Rep. NMFS 28, 28 p. Pickard, G. L., and W. J. Emery. 1982. Descriptive physical oceanography, 4th (SD enlarged ed. Pergamon Press, Oxford, 249 p. Pimentel, R. A. 1979. Morphometries. The multivariate analysis of biological data. Kendall/Hunt Publ. Co., Dubuque, Iowa, 276 p. Reilly, S. B. 1990. Seasonal changes in distribution and habitat differences among dolphins in the eastern tropical Pacific. Mar. Ecol. Prog. Ser. 66(1-21:1-11. Smith, T. D. 1983. Changes in size of three dolphin (Stenella spp.) populations in the eastern tropical Pacific. Fish. Bull. 81:1-13. Ter Braak, C. J. F. 1985. Correspondence analysis of incidence and abundance data: properties in terms of a unimodal response model. Biometrics 41:859-873. 1986. Canonical correspondence analysis: a new eigenvector technique for multivariate direct gra- dient analysis. Ecology 67(51:1167-1179. 1988. CANOCO— a FORTRAN program for canoni- cal community ordination by [partial] [detrended] [canonical] correspondence analysis, principal com- ponents analysis and redundancy analysis (version 2.1). Tech. Rep. LWA-88-02, Groep Landbouw- wiskunde, Postbus 100,6700 AC Wageningen, The Netherlands. Townsend, C. H. 1935. The distribution of certain whales as shown by logbook records of American whaleships. Zoologica Sci. Contrib. N.Y. Zool. Soc. 19(1): 1-50. Tsuchiya, M. 1974. Variation of surface geostrophic flow in the eastern intertropical Pacific Ocean. Fish. Bull. 72(41:1075-1086. Whittaker, R. H., S. A. Levin, and R. B. Root. 1973. Niche, habitat and ecotope. Am. Nat. 107:321-338. Wyrtki, K. 1964. Upwelling in the Costa Rica Dome. Fish. Bull. 63:355-372. 1966. Oceanography of the eastern equatorial Pacific Ocean. Oceanogr. Mar. Biol. Ann. Rev. 4:33-68. 1967. Circulation and water masses in the eastern equatorial Pacific Ocean. Int. J. Oceanol. Limnol. 1:117-147. Young, T. L., and M. L.Van Woert. 1987. PLOT88 software library. Plotworks, Inc., La Jolla, CA. Abstract. — The results of a canonical correspondence analysis (CCA) of data from research ves- sel surveys of the eastern tropical Pacific were applied to time series of estimated dolphin abundances from tuna vessel sightings. The research vessel survey data con- sisted of daily dolphin school sightings and concurrent environ- mental variables for August-No- vember of 1986 through 1990. Sea- sonal fields of habitat quality for 1975-90 were calculated from his- torical bathythermograph data by using the CCA ordination results. For spotted (Stenella attenuata) and eastern spinner (S. longi- rostris orientalis ) dolphins, annual abundance estimates or inter- annual changes in those estimates are significantly correlated with habitat quality. This effect is at least partly due to expansion of high quality habitat beyond the geographic ranges assumed for the abundance estimate. We discuss ways that environmental data could be used to reduce error in dolphin abundance estimates. Interannual variability of dolphin habitats in the eastern tropical Pacific. II: Effects on abundances estimated from tuna vessel sightings, 1975-1990 Paul C. Fiedler Stephen B. Reilly Southwest Fisheries Science Center, National Marine Fisheries Service. NOAA P O Box 27 I , La Jolla. CA 92038 Manuscript accepted 18 October 1993 Fishery Bulletin 92:451^163 (1994) The eastern tropical Pacific Ocean (ETP) supports a diverse and abun- dant cetacean fauna. By the late 1960s, it had become clear that large numbers of dolphins were being killed in tuna purse seine operations (Perrin, 1969). The dol- phin species affected by the tuna fishery, known as "target species," are spotted dolphin (Stenella atten- uata), the "whitebelly" form and "eastern" subspecies of spinner dol- phin (S. longirostris and S. /. orien- talis, Perrin, 1990), common dol- phin (Delphinus delphis), and striped dolphin (S. coeruleoalba). In 1973, the U.S. government initi- ated a formal program to place ob- servers on purse seiners to monitor dolphin mortality (Smith, 1983). In 1975, the Inter-American Tropical Tuna Commission began putting observers on the international fleet. The tuna vessel observer data from these programs includes sightings of cetaceans, as well as mortality data and biological samples from incidental catches. Time series of target species abundance have been estimated from tuna vessel observer data (Buckland et al., 1992). These are yearly estimates of stocks of spot- ted, spinner, and common dolphins within nominal stock boundaries known as the Status of Porpoise Stocks or SOPS boundaries.1 Buckland et al. (1992) analyzed smoothed time series of these esti- mates and detected significant trends for some stocks. Time series of dolphin abundance estimates are subject to consider- able sampling error plus the effects of environmental variability on abundance and distribution. The effect of environmental factors on abundance estimates must be quantified before such a time series can be properly interpreted in terms of abundance changes or trends. This paper analyzes rela- tionships between abundance esti- mates from tuna vessel observer data and environmental variability. Since little or no environmental data are collected on these ships, we base our analysis on species- environment relationships derived from research vessel data (Reilly and Fiedler, 1994). National Marine Fisheries Ser- vice research vessels have surveyed the ETP several times since 1974 to collect data for abundance esti- mates and supplement the tuna vessel observer data (Holt et al., 1987). The most extensive research vessel observer data were collected 1 Au, D. W. K., W. L. Perryman, and W. F. Perrin. 1979. Dolphin distribution and the relationship to environmental features in the eastern tropical Pacific. NOAA/SWFC Admin. Rep. No. LJ-79-43, 59 p. 451 452 Fishery Bulletin 92(2). 1994 for the Monitoring of Porpoise Stocks (MOPS) pro- gram during August-November in 1986 through 1990 (Wade and Gerrodette, 1992). The MOPS sur- veys were designed to cover the SOPS spotted, spin- ner, and common dolphin stock boundaries. Reilly and Fiedler (1994) showed that sighting rates of dol- phin species or school types on the MOPS surveys were related to concurrently measured environmen- tal variables. We used a robust and efficient multi- variate technique, canonical correspondence analy- sis, to examine relations between spatial distribu- tions of dolphin species and environmental vari- ables. Canonical correspondence analysis (CCA) was developed to relate community composition to known variation in the environment. It is a form of direct gradient analysis that directly estimates or- dination axes as linear combinations of observed environmental variables (Ter Braak, 1986). The advantages of CCA for multivariate species-environ- ment analyses and details of the method are dis- cussed by Reilly and Fiedler (1994). CCA estimates unimodal (Gaussian) responses of species along the ordination axes. In general, the response is observed abundance or probability of occurrence. We assume that a species response (abundance observed at a site in time and space) reflects the suitability of environmental conditions at that site relative to the species' optimal habitat or niche. This suitability, or habitat quality, is defined by the response distribu- tion along the axes. An observed response will also include error caused by behavioral responses to the environment that affect the detectability of schools. In this paper, we analyze relations between abun- dance estimates and habitat quality. It must be stated at the outset that in a 15-year record of popu- lations with growth rates of only -0.02 yr_1 (Reilly and Barlow, 1986), we are able to detect short-term environmental effects on sampling but not long-term effects on population size. Nevertheless, we demon- strate that some of the interannual variability in estimated abundances of ETP dolphins can be ex- plained by environmental factors. Methods We used archived bathythermograph data to quan- tify variability of surface temperature, thermocline depth, and thermocline thickness in the MOPS study area since 1975. These variables were shown to be important in explaining variations in encoun- ter rates in the MOPS surveys (Reilly and Fiedler, 1994). Other important variables (salinity and chlo- rophyll concentration) have not been routinely ob- served with sufficient frequency to be used in this historical analysis. Seasonal fields (gridded values) of surface temperature, thermocline depth, and ther- mocline thickness for the period 1975-90 were de- rived from a bathythermograph data base originally described by Fiedler (1992) and augmented with data from other sources for this study (Table 1). Thermocline depth is defined as the depth of the 20°C isotherm. Thermocline thickness is defined as the difference in depth between the 20°C and 15°C isotherms. Data were objectively gridded by seasons (Decem- ber-February, March-May, June-August, Septem- ber-November from 1975 through 1990) on a 2-de- gree latitude-longitude grid from lat. 20°S to 30°N latitude and from the coast out to long. 160°W. Decorrelation scales, the distances required for a substantial change in surface temperature or ther- mocline depth, have been estimated as 3 degrees latitude and 15 degrees longitude in this region (Sprintall and Meyers, 1991). At each grid point, means of at least 20 observations within up to 4 degrees latitude and 20 degrees longitude were cal- culated. The observations were weighted by the re- Table 1 Numbers of bathythermograph profiles, after screening for errors and replicates, used to define habitat quality in yearly seasonal grids ( 1975-90) and in climatologies (1960-91). NODC = NOAA/ NESDIS/National Oceanographic Data Center CD-ROM NODC-03: Global Ocean Temperature and Salinity Profiles, vol. 2, Pacific Ocean; MOODS = Navy Master Oceanographic Observa- tions Data Set, including non-NODC observations through 1983 obtained from the Naval Oceano- graphic Office through NODC and 1985-90 obser- vations obtained from Steve Pazan, Scripps Insti- tution of Oceanography; SOP = French-American ship-of-opportunity observations obtained from NOAA/ERL/Pacific Marine Environmental Labo- ratory (Kessler, 1990); FSFRL = Japanese Far Seas Fisheries Research Laboratory MBT data obtained from PMEL and from NOS/Ocean Appli- cations Branch (these data will be added to the NODC data set in the near future). 1975-90 1960-91 NODC MOODS SOP FSFRL Total 61,486 10,741 2,859 2,350 77,436 127,365 15,077 11,305 4,744 158,491 Fiedler and Reilly: Interannual variability of dolphin habitats and abundances 453 ciprocal of the distance from the grid point. The range of observations around a grid point was in- creased in increments of 0.4 degrees latitude and 2 degrees longitude to obtain a minimum sample size of 20 for each grid point. Thus, local grid resolution decreases in data-poor regions, generally south of the equator where the maximum distance required was up to 20 degrees longitude. Within the MOPS area, sufficient observations were available within 2 degrees latitude and 10 degrees longitude of 71% of the gridpoints, and within 4 degrees latitude and 20 degrees longitude of 95% of the gridpoints. We converted observations to anomalies (deviations from the seasonal mean) before gridding to reduce the spatial variability of the observations. This mini- mized bias caused by interpolation over or extrapo- lation into large data gaps. Relationships among abundances of dolphin spe- cies and environmental variables were analyzed by using CCA as described in Reilly and Fiedler (1994). Encounter rate, equal to number of schools sighted per unit of sighting effort (trackline distance), was used as a measure of relative abundance. The final abundance estimate also depends on school size and effective track width. However, Reilly and Fiedler (1994) found that weighting encounter rates by es- timated school size in the CCA produced essentially the same species-environment patterns. Therefore, schools of all sizes occupied approximately the same habitats and school size variability within these habi- tats was not related to the environmental variables included in the analysis. CCA was performed on MOPS sightings and en- vironmental data as in Reilly and Fiedler ( 1994), ex- cept that mixed schools of spotted and spinner dol- phins were counted as schools of both species rather than as an additional "species." Also, we used Hill's symmetric scaling of species and site scores (S=-3 in our implementation of CANOCO2). This alterna- tive scaling of the ordination had no qualitative ef- fect on species-environment patterns but seemed to give more reasonable results at the edges of the study area when scores were combined to quantify species habitat distributions. Habitat quality for species i (Ht) at a gridpoint was calculated from the Gaussian responses fit to the two dominant canonical axes by CCA. The re- sponse to each environmental axis was calculated as a normal probability density function: 2 Ter Braak, C. J. F. (1988). CANOCO— a FORTRAN program for canonical community ordination by I partial I [detrendedl [canonical] correspondence analysis, principal components analysis and redundancy analysis (version 2.1). Tech. Rept. LWA-88-02, Groep Landbouwwiskunde, Postbus 100, 6700 AC Wageningen. The Netherlands. HtJ = t-j1 exp (^ -0.5 * ((xj - «y )/ty) J , where, x = the site (gridpoint) score on environmental axis./', utj = the species i score (optimum) on axis j, tu = the tolerance (standard deviation) of species i on axis j. Species scores and tolerances (u- and £••) were output by CANOCO as part of the CCA. Site scores (x ) were calculated as linear combinations, defined by the output canonical axis coefficients, of normal- ized and gridded environmental observations. Habi- tat quality, Ht, was then calculated as the geomet- ric mean of HJ&; each Hy was scaled so that the mean value is 1.0 during 1975-90. Thus, Hi is equal to the abundance expected at a site, based on local environmental conditions, divided by the mean abundance in the study area during 1975-90. Point estimates of annual abundance were pro- vided by Anganuzzi3 for pooled stocks: spotted dol- phins include northern and southern offshore spot- ted dolphins, whitebelly spinner dolphins include northern and southern whitebelly spinner dolphins, and common dolphins include northern, central, and southern common dolphins. No estimates were made for striped dolphins, which were rarely set on by tuna vessels. Results The species-environment biplot (Fig. 1) summarizes the results of the CCA of five species and three en- vironmental variables observed during 1986-90 MOPS research surveys. The eigenvalues of the three canonical axes were 0.296, 0.074, and 0.001. The first two axes explained 99.7% of the species- environment variance accounted for by all three axes. Therefore, the third axis was not used. The first two axes explained 20.5% of the total variance of species encounter rates (Table 2). Positive scores on canonical axis 1 indicate cool surface temperature and a shallow, weak ther- mocline (Table 3). These are characteristics of the productive "cool upwelling" habitat that we identi- fied with the first axis in the complete CCA (seven species and six environmental variables, Reilly and Fiedler, 1994). This habitat is found in the equato- rial and eastern boundary current (Peru and Cali- 3 Anganuzzi, A. Inter-American Tropical Tuna Commission, 8604 La Jolla Shores Drive, La Jolla, CA 92038. Pers. commun. De- cember 1991. 454 Fishery Bulletin 92(2). 1994 0E. SPIN .TEMP CL qCOMMON I ZD Canon ica I 1 QSPOT Q5TRIPED fix is 1 *Z20 WB SP1NQ Figure 1 Ordination biplot of first two canonical axes from CCA of species-environment data from 1986-90 MOPS surveys of the ETP. Points represent species scores (optima) and vectors represent the regression relationships of environmental variables with the canonical axes. TEMP = surface temperature, Z20 = thermocline depth, ZD = thermocline thickness. SPOT = spotted dolphin (Stenella attenuata), COM- MON = common dolphin (Delphinus delphis), E. SPIN = eastern spinner dolphin (S. longirostris orientalis) , W. B. SPIN = whitebelly spinner dolphin (S. longirostris), STRIPED = striped dolphin (S. coeruleoalba). Table 2 Fractions of individual and total species variances explained by canonical correspondence analysis (CCA). AX1 = canonical axis 1, AX2 = canonical axis 2. Dolphin species: spotted = Stenella atten- uata, common = Delphinus delphis, eastern spin- ner = S. longirostris orientalis, whitebelly spin- ner = S. longirostris, striped = S. coeruleoalba. AX1 AX2 AX1+AX2 Spotted 0.191 0.000 0.191 Common 0.321 0.004 0.325 Eastern spinner 0.169 0.089 0.258 Whitebelly spinner 0.014 0.118 0.132 Striped 0.037 0.002 0.039 Total 0.164 0.041 0.205 fornia Currents) waters of the ETP and is also present seasonally in the region of the Costa Rica Dome at 10°N, 90° W (Fiedler, 1992). Positive scores on canonical axis 2 indicate warm surface tempera- ture and a shallow thermocline (Table 3). These are characteristics of the "coastal tropical" habitat of Reilly and Fiedler (1994). This habitat is centered in the warmest tropical surface water of the ETP, along the coast of Mexico south of Baja California. Species responses along the canonical axes (Fig. 2) showed some separation of species habitats, as in the complete CCA (Reilly and Fiedler, 1994). Axis 1 separated common dolphins from spotted and spin- ner dolphins, while axis 2 separated eastern and whitebelly spinner dolphins. The means (optima, Table 4) of a species distribution on the two canoni- cal axes in Figure 2 are equal to the species scores plotted in the ordination biplot (Fig. 1). The validity of the species/environment relation- ships calculated by CCA was confirmed in three ways. First, distributions of climatological H, de- rived from the CCA results and climatological val- ues of environmental variables at each gridpoint in the MOPS area (Fiedler, 1992), were consistent with stock ranges indicated by the SOPS population boundaries, with the exception of whitebelly spin- ner dolphins (Fig. 3). Second, the distributions were similar to patterns in maps of tuna and research vessel sighting records (Perrin et al., 1985), although such maps can give only a rough indication of habi- tat distribution because the sighting or collection frequencies are not standardized by effort. Third, the distributions of H calculated for climatological September-November environmental conditions were significantly correlated with gridded fields of mean (August-November, 1986-90) MOPS encoun- ter rates as follows: spotted dolphin r = 0.52, com- mon dolphin r = 0.45, eastern spinner dolphin r = 0.65, whitebelly spinner dolphin r = 0.36, striped dolphin r = 0.41 (P<0.01 for all relationships). Table 3 Regression/canonical coefficients for standardized environmental variables on two environmental axes (AX1 and AX2). TEMP = surface tempera- ture, Z20 = thermocline (20°C isotherm) depth, ZD = thermocline thickness (difference between 20°C and 15°C isotherm depths). TEMP Z20 ZD AX1 AX2 -0.501 f().439 -0.326 -0.486 +0.111 Fiedler and Reilly: Interannual variability of dolphin habitats and abundances 455 Axis Score - Spotted Dolphin Common Dolphin - Eastern Spinner Dolphin - Whitebelly Spinner Dolphin - Striped Dolphin Figure 2 CCA-derived Gaussian distributions of dolphin species habitat qual- ity, H (from MOPS 1986-90 encounter rates), along the first two canonical axes. The values are scaled so that the mean of Htj is 1.0 over a range of ±2 units (standard deviations) on the canonical axis. Spotted dolphin = Stenella attenuata, common dolphin = Delphinus delphis, eastern spinner dolphin = S. longirostris orientalis, whitebelly spinner dolphin = S. longirostris, striped dolphin = S. coeruleoalba. Spotted dolphin habitat was centered south and southwest of the southern coast of Mexico to about 3°N (Fig. 3). This corresponds to the warm, tropical surface water mass of Wyrtki ( 1966, see Fig. 1 in Reilly and Fiedler, 1994). The Costa Rica Dome (10°N, 90°W) was a notable gap in favorable spot- ted dolphin habitat. Eastern spinner dolphins were even more closely associated with warm tropical surface water. High- est values of H for this stock were found off south- ern Mexico, in the center of the "coastal tropical" habitat defined by canonical axis 1 (Reilly and Fiedler, 1994). Whitebelly spinner dolphins were as- sociated with subtropical surface water (Wyrtki, 1966) to the northwest and southwest of the tropi- cal surface water in the core of the MOPS area. Table 4 Dolphin species scores (standardized) ±tolerances on the first two canonical axes from canonical correspondence analysis (CCA). Dolphin species: spotted = Stenella attenuata, common = Delphi- nus delphis, eastern spinner = S. longirostris orientalis, whitebelly spinner = S. longirostris, striped = S. coeruleoalba. AX1 AX2 Spotted Common Eastern spinner Whitebelly spinner Striped -0.53 ± 0.57 1.68 ± 1.16 -0.84 ± 0.49 -0.37 ± 0.55 0.24 ± 0.69 0.19 ± 0.54 0.28 ± 0.48 0.86 ± 0.71 -1.51 ± 1.19 -0.08 ± 0.55 456 Fishery Bulletin 92(2), 1994 Common dolphin habitat was centered in cool, upwelling-modified water in three regions: off Baja California, along 10°N with a maximum at the Costa Rica Dome, and in the equatorial surface water To no too 90 SO Figure 3 Mean habitat quality (H) for spotted iStenello attenuate!), eastern spinner (S. longirostris orientalis), whitebelly spin- ner (S. longirostris), common (Delphinus delphis), and striped (S. coeruleoalba) dolphins. H was calculated from cli- matological fields of surface temperature, thermocline depth, and thermocline thickness in the MOPS area. Heavy dashed lines delimit SOPS population areas used by Anganuzzi and Buckland (1989) and Anganuzzi et al. (1991) or, for striped dolphins, as defined by Au et al. (Footnote 1.) mass of Wyrtki (1966). These three habitat centers are occupied by the northern, central, and southern stocks of common dolphins (Perrin et al., 1985). The offshore H maximum along 10°N at 120-130°W does not correspond to high encounter rates in the MOPS data, but reflects a shoaling of the coun- tercurrent thermocline ridge at that location (Fiedler, 1992). Striped dolphins are the most widespread and abundant of the target species. The high- est H values tended to be in regions between or offshore of the centers of spotted/eastern spinner dolphin habitat in tropical surface water near the coast of southern Mexico and northern, central, and southern common dolphin habitats off Baja California, near the Costa Rica Dome, and in equatorial water (see also Reilly, 1990). Both seasonal and interannual variability were evident in time series of mean seasonal habitat quality, H (Fig. 4). The strongest interannual signal for all species can be attrib- uted to the El Nino events of 1982-83 and 1986-87. During both events, H increased for spotted and spinner dolphins and decreased for common and striped dolphins. Seasonal vari- ability, indicated by the deviations of the sea- sonal from the smoothed H values, was low for species with large geographic ranges (e.g. striped dolphin) and high for species with more restricted ranges (e.g. eastern spinner dolphin). Seasonal variability of// was as great as interannual vari- ability for eastern spinner dolphins. Annual dolphin abundance estimates, Nt, or interannnual change in abundance estimates, Nl-N/_1, were related to changes in the envi- ronment, //, for spotted and eastern spinner dolphins. Annual spotted dolphin abundance was not significantly correlated with //, but interannual change in abundance was nega- tively correlated with H (Fig. 5, r=-0.65, P=0.01). Calculating year-to-year changes in abundance eliminates multi-year trends in the time series. Buckland et al. (1992) found sig- nificant trends in estimated spotted dolphin abundance which might complicate the relation between annual N and H values. An increase in H for spotted dolphins indicates an expan- sion of favorable habitat to the south of the SOPS population boundary west of 100 W (Fig. 6A). Annual eastern spinner dolphin abundance was negatively correlated with H (r=-0.49, P=0.05). An increase in H for eastern spinner dolphins indicates an expansion of favorable habitat to the west and south of the SOPS population boundary (Fig. 6B). Fiedler and Reilly: Interannual variability of dolphin habitats and abundances 457 Common 100 90 80 70 Figure 3 (Continued) Common and whitebelly spinner dolphins did not show significant linear relationships between N and H (Fig. 7). For all four species, log-transformation of the abundance estimates or lagging N( by up to four seasons did not change the significance levels of the linear relationships. However, common dol- phin abundance appeared to be maximum at H near 1.0 and to decrease at lower or higher H values, except in 1983. Scatterplots of Nt vs. Ht for spotted and whitebelly spinner dolphins suggest similar nonlinear relationships. For common dolphins at low values of//, as in early 1983, very little high-qual- ity or favorable habitat was available in the ETP (Fig. 6C). The only favorable habitat with H>1 was in equatorial water west of the Galapagos. Half of this favorable habitat was outside the SOPS bound- ary. At high values of H, as in early 1985, favorable habitat for the central and southern stocks (along 10°N and the equator, respectively) expanded. Fa- vorable habitat along the equator extended beyond the SOPS boundary. At the same time, favorable habitat for the more abundant northern stock, off Baja California, was reduced. Discussion Estimated abundances of spotted and eastern spin- ner dolphins in the eastern tropical Pacific were correlated with changes in the environment, as de- scribed by the CCA-derived habitat quality index H. The time scale of the changes and the patterns of favorable habitat relative to nominal population boundaries (Fig. 6) suggest that the correlations resulted from a sampling effect, rather than a popu- lation effect. Thus, we have explained biases in an- nual population estimates that result in apparent population changes. For example, Anganuzzi and Buckland (1989) suggested that their low 1983 abundance estimate for spotted dolphins might be explained by dispersal of local concentrations of schools during the strong El Nino. Our results showed that spotted dolphin schools may have moved outside the nominal species range when the "coastal tropical" habitat expanded into equatorial water west of 100°W during this unusual event. Similarly, estimated eastern spinner dolphin abun- dance decreased slightly with increasing H because 458 Fishery Bulletin 92(2). 1994 of the dispersal of schools outside the nominal range used in calculating the abundance estimate. The center of distribution of spotted and eastern spin- ner dolphins is the warm tropical surface water in the core of the ETP. Therefore, the apparent re- sponses of the two populations to environmental variability are similar. The different effects of environmental variability on the habitats of the three stocks of common dol- phins appear to complicate the response of the popu- lation as a whole. Data collected during surveys of the central and northern stocks of common dolphins in fall 1992 and 1993, respectively, will allow us to quantify stock-specific habitats and, perhaps, re- sponses to environmental variability. We detected no effect of environmental variability on the abundance estimates of whitebelly spinner dolphins, but the CCA results inadequately define the geographical % 1.0 - m . j\ A A _£o* y\ i*A. / \ z 0.7 - Spotted Dolphin i i i i i i i i i i i i i i i -0.0 1.3 -i .Eastern Spinner Dolphin A A fr\ m s -0.0 * 1.0 - jmA^Y^ z 0.7 - y i i i i i i i i i i i i V i -0.0 1.3 -| A - 1.5 ^ 1.0 - m^?^ z 0.7 - Whitebelly Spinner Dolphin • i i i i i i i i i i i i i i i -0.0 1.3 -| * 1.0 - i • 1 • \ i * v • • - 1.2 Z 0.7 - Common Dolphin \ / ■ i i i t i i i W i i i i i i i -0.0 1.3 - % 1.0- ^V\f^Pxr^J\^ Striped Dolphin ^ 75 76 77 78 79 80 81 82 83 84 85 86 87 88 88 90 Year Figure 4 Mean SOPS area habitat quality (H) for five dolphin species: sea- sonal values (thin line) and smoothed values (five-season running mean, thick line). Dots are point estimates of species abundance (NxlO"6). Spotted dolphin = Stenella attenuata, common dolphin = Delphinus delphis, eastern spinner dolphin = S. longirostris orientalis, whitebelly spinner dolphin = S. longirostris, striped dol- phin = S. coeruleoalba. Fiedler and Reilly: Interannual variability of dolphin habitats and abundances 459 extent of the habitat of this stock (see Fig. 3). The apparent westward extension of the habitat outside the MOPS area is consistent with the recognition of the "whitebelly" form as a hybrid/intergrade be- tween eastern and pantropical subspecies of spin- ner dolphins (Perrin, 1990). The partial separation of eastern and whitebelly spinner dolphin habitats defined by H is consistent with the management boundary between the two forms proposed by Perrin Spotted Dolphin Figure 5 Abundance-environment relationships for spotted dolphins {Stenella attenuate) in the MOPS area: (Topi Annual abun- dance estimates (Ntxl0~6) vs. annual mean habitat quality (//,). (Bottom) Interannual change in abundance esti- mates (7V(-iVM, x 10~6) vs. mean habitat quality from June of year t-1 through May of year t (Ht _05>. Solid line connects lowess-smoothed values (Wilkinson, 1990). et al. (1991): eastern spinners north of 10°N and east of 125°W, and whitebelly spinners south of 10°N or west of 125°W. Reilly and Fiedler (1994) suggested that species habitats defined by axis scores from CCA could be used to improve the precision and accuracy of abun- dance estimates from research vessel surveys. Pre- cision could be increased by post-stratifying the sighting data based on the spatial distribution of axis scores. Bias could be reduced by using axis scores to quantify the amount of habitat available within a survey area. The present results suggest that this approach could be extended by using spe- cies habitat distributions incorporating environmen- tal variability along more than one canonical axis. For example, a large area of suitable spotted dolphin habitat existed in equatorial water beyond the SOPS population boundary during 1983, apparently caus- ing a serious underestimate of abundance. Gerro- dette et al.4 suggested similar approaches for using fields of H to adjust abundance estimates from MOPS research vessel surveys. We utilized the results of a multi-species CCA for this study. While this approach yields useful infor- mation about community structure, as in the sepa- ration of eastern and whitebelly spinner habitat, it does not retain the maximum amount of informa- tion about any single species for management ap- plications. A similar type of analysis for each indi- vidual species or stock might explain more of the variability in abundance and improve the quantifi- cation of habitat quality defined by Gaussian re- sponses along dominant environmental gradients. In addition, CCA could potentially be used to account for environmental effects on school size and effec- tive trackline width that cause error in dolphin abundance estimates. However, preliminary results of a CCA incorporating school size (Reilly and Fiedler, 1994) showed no meaningful relation be- tween school size and environmental variability. We have only begun to exploit CCA in our work and believe it is a powerful new technique with great potential for quantitative ecological studies of popu- lations of marine mammals and other organisms. For example, environmental variability dominates variations in recruitment of many fish stocks (Longhurst, 1984; Hollowed et al., 1987). Although our time series was not long enough to address popu- lation change, this study demonstrates the potential of CCA to detect environmental effects in multi-stock fisheries studies. 4 Gerrodette, T„ P. C. Fiedler, and S. B. Reilly. 1991. Including habitat variability in line transect estimation of abundance and trends. NMFS/SWFSC Admin. Rep. No. LJ-91-37. 460 Fishery Bulletin 92(2), 1994 SPOTTED DOLPHIN ar-May 1983 s-7^ — -160 -150 -140 -130 -120 -110 -100 -90 -80 -70 ep-Nov 1985 Figure 6A Spotted dolphin (Stenella attenuata) habitat quality (H) calculated for extreme seasons: March-May 1983 and September-November 1985. Acknowledgments We thank the dedicated tuna vessel and research vessel observers, oceanographic technicians, and cooperative ship personnel who contributed to the collection of the sighting and environmental data. Robert Holland assisted in manipulating the large data sets for this study. Doug DeMaster provided continual support and encouragement. Literature cited Anganuzzi, A. A., and S. T. Buckland. 1989. Reducing bias in trends in dolphin relative abundance, estimated from tuna vessel data. Rep. Int. Whal. Comm. 39:323-334. Buckland, S. T., K. L. Cattanach, and A. A. Anganuzzi. 1992. Estimating trends in abundance of dolphins associated with tuna in the eastern tropical Pacific Ocean, using sightings data collected on commer- cial tuna vessels. Fish. Bull. 90:1-12. Fiedler, P. C. 1992. Seasonal climatologies and variability of eastern tropical Pacific surface waters. Dep. Commer., NO AA Tech. Rep. NMFS 109, 65 p. Hollowed, A. B., K. M. Bailey, and W. S. Wooster. 1987. Patterns in recruitment of marine fishes in Fiedler and Reilly: Interannual variability of dolphin habitats and abundances 461 EASTERN SPINNER DOLPHIN -160 -150 -140 -130 -120 -110 -100 -90 -80 -70 -160 -ISO -140 -130 -120 -110 -100 -90 Figure 6B Eastern spinner dolphin (Stenella longirostris orientalis) habitat qual- ity (H) calculated for extreme seasons: March-May 1987 and Decem- ber 1988-February 1989. the northeast Pacific Ocean. Biol. Oceanogr. 5:99-131. Holt, R. S., T. Gerrodette, and A. B. Cologne. 1987. Research vessel survey design for monitoring dolphin abundance in the eastern tropical Pacific. Fish. Bull. 85:435-446. Kessler, W. S. 1990. Observations of long Rossby waves in the northern tropical Pacific. J. Geophys. Res. 95(C4):5183-5217. Longhurst, A. 1984. Heterogeneity in the ocean - implications for fisheries. Rapp. P.-v. Reun. int. Explor. Mer, 185:268-282. Perrin, W. F. 1969. Using porpoise to catch tuna. World Fish. 18(6):42^5. 1990. Subspecies of Stenella longirostris (Mammalia: Cetacea: Delphinidae). Proc. Biol. Soc. Wash. 103:453^63. Perrin, W. F., M. D. Scott, G. J. Walker, and V. L. Cass. 1985. Review of geographical stocks of tropical dol- phins (Stenella spp. and Delphinus delphis) in the eastern Pacific. Dep. Commer., NOAATech. Rep. NMFS 28, 28 p. Perrin, W. F., P. A. Akin, and J. V. Kashiwada. 1991. Geographic variation in external morphology of the spinner dolphin Stenella longirostris in the eastern Pacific and implications for conservation. Fish. Bull. 89:411-428. Reilly, S. B. 1990. Seasonal changes in distribution and habitat 462 Fishery Bulletin 92(2). 1994 COMMON DOLPHIN Mar-May 1983 ■ v>. c-Feb 1985 -160 -ISO -1*0 -1J0 -120 -110 -100 -90 -80 -70 Figure 6C Common dolphin (Delphinus delphis) habitat quality iH) calculated for extreme seasons: March-May 1983 and December 1984-February 1985. differences among dolphins in the eastern tropi- cal Pacific. Mar. Ecol. Prog. Ser. 66:1-11. Reilly, S. B., and J. Barlow. 1986. Rates of increase in dolphin population size. Fish. Bull. 84:527-533. Reilly, S. B., and P. C. Fiedler. 1994. Interannual variability of dolphin habitats in the eastern tropical Pacific. I: Research vessel surveys, 1986-1990. Fish. Bull. 92:434-450. Smith, T. D. 1983. Changes in size of three dolphin (Stenella spp.) populations in the eastern tropical Pacific. Fish. Bull. 81:1-13. Sprintall, J., and G. Meyers. 1991. An optimal XBT sampling network for the eastern Pacific Ocean. J. Geophys. Res. 96(C6): 10,539-10,552. Ter Braak, C. J. F. 1986. Canonical correspondence analysis: a new eigenvector technique for multivariate direct gra- dient analysis. Ecology 67:1167-1179. Wade, P. R., and T. Gerrodette. 1992. Estimates of dolphin abundance in the eastern tropical Pacific: preliminary analysis of five years of data. Rep. Int. Whal. Comm. 42:533-539. Wilkinson, L. 1990. SYGRAPH: The system for graphics. SYSTAT, Inc., Evanston, IL. Wyrtki, K. 1966. Oceanography of the eastern equatorial Pacific Ocean. Oceanogr. Mar. Biol. Ann. Rev. 4:33-68. Fiedler and Reilly: Interannual variability of dolphin habitats and abundances 463 Eastern Spinner Dolphin Whitebelly Spinner Dolphin Common Dolphin Figure 7 Abundance-environment relations for eastern spinner (Stenella longirostris orientalis), whitebelly spinner (S. longirostris), and common (Delphinus delphis) dolphins in the MOPS area, annual abundance estimates WxlO-6) vs. annual mean habitat quality (Ht). Solid line connects lowess-smoothed values (Wilkinson, 1990). Annual mass strandings of pelagic red crabs, Pleuroncodes planipes (Crustacea: Anomura: Galatheidae), in Bahia Magdalena, Baja California Sur, Mexico David Aurioles-Gamboa Centra de Investigationes Biologicas de Baja California Sur Apartado Postal 128, LA Paz B. C. S , Mexico Maria Isabel Castro-Gonzalez Instituto Nacional de la Nutricibn "Salvador Zubir^n" Vasco de Quiroga #15. Col y Del Tlalpan 14000. Mexico D. F Ricardo Perez-Flores Centra de Investigaciones Biologicas de Baja California Sur Apartado Postal 1 28, La Paz B. C S . Mexico. Pelagic red crabs (or langostilla in Spanish), Pleuroncodes planipes, are very abundant galatheid crus- taceans off the west coast of Baja California. Some studies suggest that pelagic red crabs are the most abundant species in the micro- neckton, one of the most important consumers of phytoplankton, and the most common prey item for many marine vertebrates in the area (Boyd, 1962, 1967; Black- burn, 1969;Longhurstetal., 1967; Kato, 1974; Galvan, 1988; Balart and Castro1). Although widely used, the com- mon name (pelagic red crab) de- scribes only the planktonic period (about one year) in the life of the species. Larvae, juveniles, and young adults are planktonic. At 1 Balart-Paez, E., and Castro-Aguirre, J. L. 1992. Habitos alimenticios de la merluza Bajacaliforniana Merluceius angustimanus, en la costa occidental de Baja California Sur, Mexico. Paper pres. at the IX International Symposium of Marine Biology, 1-5 June 1992, La Paz Baja California Sur, Mexico. about 17-20 mm standard cara- pace length (SCL), they become benthic, making occasional move- ments to the surface (mostly at night) in a circadian migration (Boyd, 1967). Once the animals reach 32-34 mm SCL, they are fully benthic as are other galat- heid species (Boyd, 1967; Aurioles- Gamboa, 1992). Pelagic red crabs breed from December through April; the peak of the reproductive season is in February (Boyd, 1962; Kato, 1974; Gomez, 1990). Females about 14- 15 mm SCL, have been found car- rying eggs, but most of the females start to breed when they are about 20 mm SCL (Boyd, 1962; Serrano, 1991). The benthic population per- forms seasonal bathymetric move- ments, at least in the area from lat. 24° to 26° N, in which they dis- perse during winter and spring to occupy the benthos of the conti- nental shelf (0-200 m depth). Af- ter the breeding season, the popu- lation moves to deeper waters ( 100-200 m), and probably invades the continental slope (Aurioles- Gamboa, 1992). Population with- drawal is associated with a rise in bottom temperature above 16°C, and pelagic red crabs remain from June through October in waters 100-200 m deep, where bottom temperature is in the range of 12-16°C (Aurioles-Gamboa, 1992). One of the notable characteris- tics in pelagic red crab life history are mass strandings, which have been reported for Bahia Magda- lena and the California Coast (Glynn, 1961; Boyd, 1962; Kato, 1974; Stewart et al., 1984). The main difference between strand- ings in California and Baja Cali- fornia is the frequency of ocur- rence. Pelagic red crab beachings in California occur during El Nino events, which enable the popula- tion to move northward in warm water currents originating in the south. In contrast, pelagic red crab strandings in Bahia Magdalena are annual, and apparently recur on the same beaches and during the same season of year. In Bahia Magdalena, pelagic red crabs were observed in the upper 50 cm of water of the surf zone before stranding. Onshore winds and receding tides hasten and intensify stranding (Boyd, 1962). Kato (1974) proposed that the presence of pelagic red crabs near shore is primarily due to winds, waves, and currents. On 9 May 1991, one of us (D.A- G.) observed a mass stranding of pelagic red crabs on a beach of Magdalena Island close to the mouth of the Bahia Magdalena on the Pacific coast of Baja Califor- nia (Fig. 1). About 1100 hours, a compact surface swarm of live crabs was seen 1-2 m from the beach. Small groups of crabs were thrown to the beach by waves, and were unable to return to the sea. Manuscnpt accepted 4 October 1993 Fishery Bulletin 92:464-470 ( 1994) 464 NOTE Aurioles-Gamboa et al.: Mass strandings of Pleuroncodes planipes 465 Figure 1 Study area on the west coast of Baja California Sur, Mexico. Dots inside Bahia Magdalena indicate the location of stations sampled. The shaded area is the region where pelagic red crabs, Pleuroncodes planipes, were found as indicated by positive sta- tions and visual observations (from Soli's, 1991). The general appearance of these pelagic red crabs (color and mobility) suggested that the crustaceans were healthy and that the stranding could be con- sidered accidental (Boyd, 1962). We questioned local fishermen about crab strandings and determined that mass strandings 1 ) occur annually, usually in spring, from April through June, and 2 ) are common on Santa Margarita and Magdalena Islands but rarely seen on the peninsular coast. Since strandings coincide with the end of the breed- ing season, we addressed the null hypothesis that stranded pelagic red crabs, particularly females, were in a weakened state because of energy expended in reproduction, as reported for many crustaceans (Hartnoll, 1985). It is known that pelagic red crabs are able to breed twice in a single breeding season and may produce, depending on body size, from 500 to 5000 eggs in each brood ( Serrano, 1991). Based on this fact, debilitation caused by reproductive investment should be more evident in females. For that reason, the chemical composition of pelagic red crabs in Bahia Magdalena was determined for both sexes. We were aware that males were over- represented in samples collected on the con- tinental shelf (Boyd, 1962; Serrano, 1991); thus we also wished to determine if stranded animals were female-biased, as an explana- tion for the unbalanced sex ratio recorded on the continental shelf. Materials and methods To test whether the strandings were due to weakness caused by starvation or malnutri- tion, about 20 kg of live-stranded pelagic red crabs were collected as soon they were stranded. From this sample, a total of 1,150 individuals were sexed and measured for SCL to the nearest 1.0 mm. Sex was deter- mined by the presence or absence of modi- fied pleopods, which males use to fertilize the eggs (Boyd, 1962). Standard carapace length was measured from the antorbital notches of the rostrum to the midpoint of the poste- rior border of the carapace (Kato, 1974). This measurement is usually preferred over total length because it does not vary with shrink- age of the abdomen. Stomach contents were analyzed in order to determine if 1 ) the animals had been feed- ing before stranding, and 2) the number of items and composition was similar to stom- ach contents of pelagic red crabs collected on the continental shelf (Perez and Aurioles-Gamboa2). We examined the stomach contents from a subsample of nine individuals after fixing in formalin (4%), re- moving the cardiac-pyloric stomach from the animal, dissolving its contents in two drops of water and plac- ing them on a smear slide. Perez and Aurioles- Gamboa2 determined that the average stomach composition in a swarm of crabs does not vary sig- nificantly after a sample of six crabs. Food items were identified and counted under the microscope and relative abundance of major groups : Perez, F. R., and D. Aurioles-Gamboa. 1992. Cambios en la alimentacion invierno-verano de la langostilla Pleuroncodes planipes, en la costa oeste de Baja California. Paper pres. at the DC International Symposium of Marine Biology, 1-5 June 1992. La Paz B. C. S. Mexico. 466 Fishery Bulletin 92|2). 1994 of food items was determined. Following Perez and Aurioles-Gamboa2, we recorded some of the phyto and zooplankton components in four major groups: 1) phytoplankton, 2) zooplankton, 3) particulate or- ganic matter (POM), and 4) inorganic matter (small grains of sand, clay or mud). The number of diatoms, crustacean parts, foraminifers, and other components, such as small agglomerations of POM, were counted and their numbers converted to rela- tive frequency. Proximate analyses was based on 200 g 8 and 200 g 9 of sun dried and milled pelagic red crabs (about 4 kgs of fresh crabs). The techniques used were those of the Association of Official Analytical Chem- ists (A.O.A.O, 1984): moisture (7.007), ash (7.009), crude fiber (7.006), crude protein (2.057), ether ex- tract (7.060), carbohydrates (by difference from all other determinations at 100%). This methodology had been used previously to analyze pelagic red crabs sampled from the benthos of the continental shelf (Castro, 1993). Two-sample f-tests (Zar, 1984) were conducted to identify differences in food composition, proximate composition, and mean SCL between stranded pelagic red crabs in Bahia Magdalena and breeding pelagic red crabs collected on the continen- tal shelf in March 1990. Achi-square test for deter- mining a possible deviation of sex ratio was applied for the stranded pelagic red crabs (Zar, 1984). Results The stranded crabs formed a long brilliant red line of several kilometers on the interior coast of the northern part of Bahia Magdalena (Fig. 1). In addi- tion, there were two lines of dried crabs separated by a few meters, higher on the beach and stranded during the previous days. Behavior of pelagic red crabs during stranding Two hours of observations on a surface swarm about 12 m long and 1 m wide were conducted after 1100 hours (11 May 1991) during the receding tide. The swarm was propelled to and from the beach by the waves and was unable to move offshore. When the swarm was pushed toward the beach by the wave action, some animals were thrown onto the sand and exposed as the water receded. During sampling on the beach, the pelagic red crabs moved their legs and actively used their chelae as pincers. This behavior was typical of pelagic red crabs caught in trawls from the continental shelf. Pelagic red crabs had been caught from the shelf in all seasons, but only during mid-summer (when the bottom-surface temperature difference is as great as 17°C) did the crabs show signs of damage as they moved slowly and died rapidly on the deck (Aurioles- Gamboa, unpubl. data). In contrast, crabs were very active in the first minutes after stranding, and moved less frequently later. The crabs were brilliant red, which differentiated them from the lighter color of crabs collected on the continental shelf. Based on their vigor- ous activity, the pelagic red crabs stranded on Magdalena Island (Fig. 1 ) appeared to be in good health. Stomach contents of stranded pelagic red crabs The total number of items and relative frequency of the four major groups found in the stomachs is shown in Figure 2A. For comparison, the results of a typi- cal sample taken on the continental shelf is provided in Figure 2B (Perez, 1992). The minimum number of items counted was 457 and the maximum about 2,266. This range was greater than that found in crabs from the continental shelf (841-1,495 items). However, there was no significant difference in the mean num- ber of items (x=l,027 versus 1,100) between stom- achs from the stranded crabs and those from the shelf (two sample *-test, P>0.05, df=9; r=-0.3082). Food composition (particulate organic matter, zoop- lankton, and phytoplankton) was not different be- tween crabs from the two regions. Inorganic matter (grains of sand, clay, etc.) was more abundant in the stomachs of the pelagic red crabs collected from the shelf (x=365 versus 60; two sample ^-test, P<0.05, df=9; ^=-0.0046). This difference, however, does not account for a significant change in the feeding hab- its of crabs from the two samples. Proximate composition of stranded red crabs In stranded pelagic red crabs, the sexes were not dif- ferent in protein and crude fiber (Table 1), however they differed significantly in lipids and ash contents (two sample t-test, P<0.05, df=9; *=1.870 and 10.012 respectively). Females had higher lipid and lower ash content than males, both by about 1.5%. There were significant differences when the chemi- cal composition of stranded crabs (sexes combined), were compared to crabs from the continental shelf (Table 1 ). Crabs from both areas were similar in their protein content, but differed in lipids and ash. The lipid content in Bahia Magdalena crabs was almost nine times higher than that in shelf crabs U-test, P<0.05, df=16, r=35.664). Crabs from the continen- tal shelf were higher in fiber and ash but lower in carbohydrate content. NOTE Aurioles-Gamboa et al .: Mass strandings of Pleuroncodes planipes 467 It was also noted that during handling of stranded crabs an oily, orange film was left in the containers, a phenomenon not previously observed in crabs col- lected in more than 200 bottom trawls on the conti- nental shelf (Aurioles-Gamboa, 1992). The substance probably contained carotenoids, which in this species have been identified as astaxanthins (Wilkie, 1972). Size and sex of stranded pelagic red crabs Pelagic red crabs collected in Bahia Magdalena were 12-28 mm SCL. Size distribution was similar for 1 23456789 NUMBER OF STOMACHS 2500 1 2 3 4 5 6 7 8 NUMBER OF STOMACHS ORGANIC MATTER m INORG MATTER ^ ZOOPLANKTON □ PHYTOPLANKTON Figure 2 Total number of items and gross stomach composition from pelagic red crabs, Pleuroncodes planipes, collected in Bahia Magdalena (A), and organisms typically found on the benthos of the continental shelf off Baja Califor- nia (B). POM = Particulate Organic Matter, IM = Inor- ganic Matter. Numbers below the x-axis indicate num- ber of stomachs. males and females (Fig. 3); mode and mean were 15.82 and 17.02 mm for females and 17.36 and 17.26 mm for males, respectively. Differences in mean size of males and females were statistically significant (two-sample *-test=2.057,P<0.05,df=l,147). In older organisms, slight sexual dimorphism is evident in males which are slightly larger and heavier with longer and wider chelae (Serrano and Aurioles- Gamboa, 1991). Pelagic red crabs are about 14 mm SCL at the end of the first year of life (about February-March) and grow approximately 1 mm per month (Boyd, 1962). Because the crabs were about 17 mm SCL (mode), they should have been about 14-15 months old. Some individuals were about 24-28 mm SCL (Fig. 3), which according to Boyd (1962) were about 26- 27 months old. The number of females (605) in relation to the number of males (544), deviates significantly from the expected 1:1 proportion (chi-square P<0.001). In contrast, the sex ratio of crabs from the conti- nental shelf is slightly male biased throughout the year (Serrano, 1991). Discussion Debilitated pelagic red crabs as explanation of strandings We rejected the null hypothesis that stranded pe- lagic red crabs represent a debilitated fraction of the population because 1 ) the stranded crabs moved vigorously during and just after stranding, 2) the stomachs of crabs were full and the contents were similar to those collected on the continental shelf, which indicated they were feeding normally, 3) the chemical composition was generally similar to those collected on the continental shelf, and when differ- ent, did not suggest malnutrition, and 4) timing and area of strandings are better explained by physical phenomena (i.e. by accidental stranding due to fun- neling effect, wave action, and receding tide). Our observations support the observations made by Boyd (1962), that waves and receding tide play a major role after the animals enter the surf zone. Two available stranding reports (Boyd, 1962; Jorge Llinas3), indicated that beachings occurred during falling tides. According to local fishermen, the two higher lines of stranded crabs we found on the beach of Isla Magdalena on 9 May 1991, were stranded two mornings before our visit, also during receding tides. 1 LHnas, J. Centro de Investigaciones Biologicas de Baja California Sur. Apdo. Postal 128. La Pax Baja California Sur. Mexico, personal commun. 1993. 468 Fishery Bulletin 92(2), 1994 Differences in pelagic red crab chemical composition The lipid content of the younger stranded pelagic red crabs was dif- ferent from that recorded in older specimens caught on the benthos of the continental shelf (Table 1). Dif- ferences in the proportion of chemi- cal components between young and old individuals have been reported for many decapod species (Herring, 1973; Morris, 1973). The observed differences in lipid concentration and apparent pigmentation between young-adult and older-benthic pe- lagic red crabs cannot be considered abnormal or attributable to un- healthy specimens. There was no evi- dence to support the hypothesis of a debilitated fraction of pelagic red crabs, and strandings can be ex- plained by mere accident. The differences in lipid and ash content between male and female stranded crabs would be attributable to metabolic differences in which fe- males require more lipid to invest in egg production (Hartnoll, 1985). Some females still had eggs attached on the pleopods as evidence of their repro- ductive condition; however, the low numbers of ovigerous females in the sample, suggested that we collected them at the end of the breeding season. Size and sex of pelagic stranded red crabs The stranded individuals we sampled were predominantly 13 to 20 mm SCL, although some larger crabs were also found (Fig. 3). Kato ( 1974) also reported individuals of these two size distributions in a mass stranding in Bahia Magdalena. Photos were available of a mass stranding in May of 1979, in which pelagic red crabs larger than 25 mm SCL (second year of life, Boyd 1962) were very abundant. Therefore, the over-representation of young adults in the 1991 stranding we sampled is not the rule, as both size and age groups were found. The sex ratio (1:1.11) of our sample was significantly female biased (P<0.05, x2=22.97, df=l,147). Benthic samples of crabs in Bahia Magdalena during 1990 (Soli's, 1991), also Table 1 Mean values for proximate chemical analyses of Pleuroncodes planipes samples from Bahia Magdalena and the Continental Sh elf of Baja California. Bahia Magdalena Males Females Significant % % difference n = 5 n = 6 957f conf. * Moisture 4.7 (0.28) 4.73 (0.23) Ash 29.79 (0.36) 28.21 (0.66) + Ether extract 12.79 (0.72) 15.46 (1.39) + Crude fiber 7.91 (0.99) 8.70 (0.65) - Crude protein 41.75 (0.14) 40.64 (0.07) - Carbohydrates 3.06 2.26 - Bahia Magdalena Continental Shelf Both sexes Both sexes % % n= 11 n = 12 Moisture 4.71 (0.20) 4.12 (0.32) Ash 29.04 (0.24) 40.30 (0.03) + Ether extract 14.12 (1.98) 2.75 (0.007) + Crude fiber 8.30 (0.62) 12.83 (0.03) + Crude protein 41.19 (0.22) 38.61 (0.29) - Carbohydrates 2.64 1.39 - * Two-sample Me:- t for means; + significant difference, - no difference (Alph a=0.05). > o z LU D o LU 350 300 250 200 150 100 50 1150 14 n = 724 hk 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 STANDARD CARAPACE LENGTH mm B STRANDED CRABS \^J BENTHIC CRABS Figure 3 Size distribution of pelagic red crabs, Pleuroncodes planipes, stranded in Bahia Magdalena and the typical distribution of benthic crabs on the continental shelf off Baja California. NOTE Aurioles-Gamboa et at: Mass strandings of Pleuroncodes planipes 469 had an overrepresentation of females (1:1.23), sug- gesting that a higher abundance of females in stranded crabs would be a reflection of the sex ratio in the bay. Additional evidence for a female biased sex ratio in coastal waters comes from Serrano ( 1991), who reported higher frequency of females inshore during the breeding season. Similar findings were obtained by Escoto and Orellana (1981) for pelagic red crabs off the Nicaragua coast, and for a closely related species, P. monodon, from Chile (Arana and Culquichicon4). However, several authors have re- ported that the total sex ratio on the continental shelf is male biased (Boyd, 1962; Boyd and Johnson, 1963; Serrano, 1991). Boyd (1962), suggested that the sex- ratio differences on the continental shelf could be due to one or a combination of the following causes: 1) a deviation in primary sex ratio, 2) lower survival rate of females, or 3) the fact that plankton nets may not sample males and females with equal effectiveness. The hypothesis that females die at a higher rate than males is supported by the present data, since females were more abundant where there was a natu- ral cause of mass mortality. It has been mentioned that the breeding season of the species is synchro- nized to the time (winter-spring) when coastal up- welling is more intense. Pelagic red crabs may be more abundant in places where phytoplankton is more plentiful (Blackburn, 1969). Thus, it can be advantageous for pelagic red crab females to move to shore and release larvae in upwelled water. How- ever, by doing so, females are more likely to enter Bahia Magdalena or other coastal lagoons and die in accidental strandings. Literature cited a. o. a c. 1984. Official methods of analyses, 14th ed. Association of Analytical Chemists, Washing- ton D.C. Aurioles-Gamboa, D. 1992. Inshore-offshore movements of pelagic red crabs Pleuroncodes planipes ( Decapoda, Anomura, Galatheidae) off the Pacific coast of Baja Califor- nia Sur, Mexico. Crustaceana 62( 1 ):7 1-84. Blackburn, M. 1969. Conditions related to upwelling which deter- mine distribution of tropical tunas off western Baja California. U. S. Fish Wild. Serv. Fish. Bull. 68:147-176. 4 Arana, E. P., and Z. M. Culquichicon. 1990. Estructura poblacio- nal del langostino Colorado {Pleuroncodes monodon) en la zona centro-sur de Chile. Estudios y Documentos. Final Rep. No. 7/89 to Universidad Catolica de Valparaiso. Facultad de Recursos Naturales, Escuela de Ciencias del Mar, 56 p. Boyd, C. M. 1962. The biology of a marine decapod crustacean, Pleuroncodes planipes Stimpson 1860. Ph.D. the- sis, Univ. Calif. San Diego, 123 p. 1967. Benthic and pelagic habitats of the red crab Pleuroncodes planipes. Pacific Science 21:394- 403. Boyd, C. M., and M. W. Johnson. 1963. Variations in the larval stages of a decapod crustacean, Pleuroncodes planipes, Stimpson (Galatheidae). Biol. BulUWoods Hole) 124:142- 152. Castro, G. M. I. 1993. Procesos tecnologicos aplicados a la langostilla (Pleuroncodes planipes Stimpson) y cambios en su composicion quimica a diferentes latitudes para su aprovechamiento en alimentacion animal. Tesis de Maestria en produccion animal, 82 p. Escoto, R., and F. Orellana. 1981. Segundo crucero de evaluacion del recurso langostino del Paci'fico Nicaraguense. Boletfn tecnico del Instituto Nicaraguense de la Pesca (Diciembre 1980), 42 p. Galvan, M. F. 1988. Composicion y analisis de la dieta del atiin aleta amarilla Thunnus albacares , en el oceano Paci'fico Mexicano durante el periodo 1984- 1985. Tesis de Maestria en Ciencias, CICIMAR, 86 p. Glynn, P. W. 1961. The first recorded mass strandings of pelagic red crabs, Pleuroncodes planipes, at Monterey Bay California since 1959, with notes on their biology. California Fish and Game 47 (1): 97-101. Gomez, G. J. 1990. Variacion de la distribution y abundancia de los estadios planctonicos de Pleuroncodes planipes (Crustacea: Galatheidae) en la costa occidental de Baja California Sur, Mexico (1986). Tesis de Licenciatura en Biologia Marina, Universidad Autonoma de Baja California Sur, Mexico, 86 p. Hartnoll, G. R. 1985. Growth, sexual maturity and reproductive output. In A. M. Wenner (ed.), Crustacean issues 3: factors in adult growth, p. 101-128. A. A. Balkema, Boston, and Rotterdam, Netherlands, Herring, P. J. 1973. Depth distributions of the carotenoid pig- ments and lipids of some oceanic animals. 2: Deca- pod crustaceans. J. Mar. Biol. Assoc. U.K. 53:539- 562. Kato, S. 1974. Development of the pelagic red crab (Gala- theidae, Pleuroncodes planipes) fishery in the east- ern Pacific Ocean. Mar. Fish. Rev. 36 (10):l-9. Longhurst, A. R., C. J. Lorenzen, and W. H. Thomas. 1967. The role of pelagic red crabs in the grazing of phytoplankton off Baja California. Ecology 48:190-200. 470 Fishery Bulletin 92(2), 1994 Morris, R. J. 1973. Relationships between the sex and degree of maturity of marine crustaceans and their lipid compositions. J. Mar. Biol. Assoc. U.K. 53:27-37. Perez, F. R. 1992. Alimentacion de la langostilla Pleuroncodes planipes Stimpson, 1860, durante el periodo reproductive (Marzo, 1990). Tesis de licenciatura en Biologia, Universidad Nacional Autonoma de Mexico, ENEP, Zaragoza, Mexico, D. F., 53 p. Serrano, P. V. 1991. Aspectos reproductivos de la langostilla Pleuroncodes planipes (Crustacea: Decapoda: Gala- theidae). Tesis de Maestria en Ciencias Marinas, CICIMAR, 89 p, La Paz Baja California Sur, Mexico. Serrano, P. V., and D. Aurioles-Gamboa. 1991. Dimorfismo sexual en la langostilla, Pleuroncodes planipes Stimpson, 1860 (Crustacea: Decapoda: Galatheidae). Proc. of San Diego Soci- ety of Natural History 13:1-5. Solis, F. 1991. Composition y distribution espacio-temporal de los macroinvertebrados bentonicos del complejo lagunar Magdalena-Almejas de la costa occidental de Baja California Sur. Tesis de licenciatura en Biologia, Univ. Michoacana de San Nicolas de Hidalgo, Mexico, 96 p. Stewart, B. S, P. M. Yochem, and R. W. Schreiber. 1984. Pelagic red crabs as food for gulls: a possible benefit of El Nino. The Condor 86:341-342. Wilkie, D. W. 1972. The carotenoid pigmentation of Pleuroncodes planipes Stimpson (Crustacea: Decapoda: Gala- theidae). Comp. Biochem. Physiol. 42 (B):731-734. Zar, J. H. 1984. Biostatistical analysis. Prentice-Hall, Engle- wood Cliffs, NJ, 716 p. Mass marking coho salmon, Oncorhynchus kisutch, fry with lanthanum and cerium Bridget C. Ennevor Department of Fisheries and Oceans. 327-555 West Hastings Street Vancouver. British Columbia, Canada V6B 5G3 Most current salmonid tagging programs identify small propor- tions of a population. However, under some circumstances it is desirable to mark entire popula- tions. Chemical marking is a tech- nique that can rapidly mark large numbers offish without individual handling. Marking is accom- plished by exposing the fish to bio- logically rare elements that are subsequently incorporated and retained in certain tissues in which they are not naturally found. Marking of entire hatchery populations could be valuable from a fisheries management perspec- tive for stock identification (hatch- ery versus wild salmon), assess- ment of contribution to fisheries, and evaluation of current tagging and sampling techniques. Ennevor and Beames (1993) have shown that some lanthanide elements (i.e. lanthanum and ce- rium) are suitable for mass mark- ing juvenile coho salmon, Oncorhynchus kisutch. The lan- thanide elements are not absorbed from the gastro-intestinal tract (Kyker, 1961; Ellis, 1968; Luckey and Venugopal, 1977), and there- fore may be introduced through the fishes' rearing water. Because these are bone-seeking elements (Durbin et al., 1956; Jowsey et al., 1958), administered lanthanides are subsequently incorporated into the bony tissues of coho salmon fry and smolts (Ennevor, 1991; Ennevor and Beames, 1993). Analysis of the vertebral column, otoliths, and scales by inductively coupled plasma-mass spectrom- etry (ICP-MS) revealed that ad- ministered lanthanides are present in these bony tissues 10.5 months post-treatment (Ennevor, 1991; Ennevor and Beames, 1993). ICP-MS is capable of detection and quantification of the lanthanide elements at levels as low as 0.01 Ug-g-1 (Longerich et al., 1987; Houk and Thompson, 1988). Trials were performed to deter- mine whether immersion into so- lutions of lanthanide elements would produce recognizable marks on juvenile salmon. These studies were designed 1 ) to investigate dif- ferences in toxicity and uptake between the chloride and acetate forms of lanthanum and cerium, and 2 ) to assess optimal concentra- tions and exposure times for mark- ing coho salmon fry in the extremely soft and slightly acidic water at Capilano River Hatchery, British Columbia. Materials and methods In the following experiments, lan- thanum and cerium were intro- duced into the rearing water of coho salmon fry at Capilano River Hatchery. The river water at this hatchery is slightly acidic and ex- tremely soft (pH=6.5; hardness as CaC03=3.8). Concentrated lan- thanide stock solutions were me- tered into the tanks at a rate of 1 mLmin-1 and the rearing water was set to flow in at a rate of 1 L-min-1. The lanthanide solutions and the rearing water were mixed prior to delivery to the tanks con- taining the fish. Two experiments were con- ducted concurrently with coho salmon fry. One hundred fry, with an average initial weight of 3.2 g, were placed in each 35-L experi- mental tank. Experiment 1 had 4 treatment groups and a control tank where no lanthanide was ad- ministered. The lanthanide treat- ments and elemental concentra- tions were: 50 ugL-1 of LaCl3, CeCl3, La(C2H302)3, or Ce(C2H302)3 continuously for 24 days. Experi- ment 2 involved 7 treatment groups: 50 Ug-L-1 of La(C2H302)3 or Ce(C2H309)3 continuously for a to- tal of 24 treatment days; 100 ugLr1 of La(C2H302)3 or Ce(C2H302)3 on alternate days for a total of 12 treat- ment days over a 24-day period; 150 Ug-L"1 of LafC^O^ or CeCGjHgO^ every third day for a total of 8 treat- ment days over a 24-day period; and a control tank with no lantha- num or cerium. The treatment days consisted of 24 hours of ex- posure. Over the treatment period, equal amounts of lanthanum or cerium were administered to each treatment group of fry at the appro- priate concentration and duration. After completion of the lan- thanide exposures, the fry were provided with untreated river wa- ter for 14 days prior to sampling. Ten fry were randomly sampled from each of the tanks, body weights recorded, and the verte- bral columns were removed and prepared for ICP-MS analysis to determine lanthanide accumula- tion. The majority of flesh was dis- sected away from the bony tissue and any remaining traces of flesh were digested with a 6% sodium hypochlorite solution. The clean Manuscript accepted 29 October 1993 Fishery Bulletin 92:471-473 (1994) 471 472 Fishery Bulletin 92(2), 1994 backbones were oven-dried at 70°C overnight, ground to a pow- der, and a 0.01 g subsample from each fish was used for analysis (Ennevor, 1991; Ennevor and Beames, 1993). The prepared samples were submitted to a com- mercial laboratory in North Vancouver, British Columbia for ICP-MS analyses. Experiments 1 and 2 were ana- lyzed by analysis of variance with SYSTAT statistical software (Wilkinson, 1989) and differences between means were tested at P<0.05 with Tukey's multiple range test. The data were pooled by treatment groups with indi- vidual fish as experimental units. Results Experiment 1 Lanthanum and cerium adminis- tered at 50 ug-L"1 daily for 24 days had no apparent deleterious effect on the fry. They appeared to be healthy and fry weights between treated and non-treated groups did not differ after the 24-day treatments and 14- day rinse were completed. Few mortalities occurred in all groups (Table 1). Analysis of the vertebral columns from the marked fry showed each of the lanthanides to be present in approximately equal amounts. Uptake did not differ between the treatment groups. The average concen- tration of lanthanide in the vertebral columns was 6.1 ng of lanthanum or 6.2 ng of cerium (Table 1). Experiment 2 Throughout the treatments, mortalities were higher in tanks that contained the 150 ug-L"1 treatments of lanthanum or cerium (Table 1). Fewer mortalities were observed in the 100 ug-L-1 treatments; none in the 50 ug-L"1 treatments. However, after the 24-day treatment period and 14-day rinse period were com- pleted, fry weights did not differ between groups treated with lanthanum or cerium and nontreated groups (Table 1). Results of the analyses of the vertebral columns from the marked fry showed a trend of significantly (P<0.05) decreased uptake of lanthanum and cerium with decreased exposure time regardless of concen- Table 1 Percent mortalities during the 24-day treatment period, mean body weights of coho salmon, Oncorhynchus kisutch, fry at time of sampling, mean amounts (ng) ± 1 S.E.M. of lanthanum or cerium in vertebral columns of fry marked in Experiments 1 and 2. Within experiments, mean values shar- ing a similar superscript letter were not significantly different (P<0.05) according to Tukey's Test. Treatment group' Mortalities Mean fry Concentration Duration La or Ce (ng) Element (Ug-L"1) (d) (%) weight2 (g) ± S.E.M. Experiment 1 La* 50 24 2 3.7 7.0° ± 1.2 Ce* 50 24 2 4.2 6.1" ±0.9 La 50 24 (i 3.5 5.6" ± 0.8 Ce 50 24 ii 4.1 6.2° ± 0.8 Control ii 21 1 3.7 O.l't 0.0 Experiment 2 La 50 24 i) 3.5 5.6° ± 0.8 Ce 50 24 0 4.1 6.2° ± 0.8 La 10(1 12 2 2.7 4.6°6 ± 0.5 Ce 100 12 3 3.3 4.4a6 + 0.3 La 150 8 12 1 1 4.0* ± 0.5 Ce 150 8 1 3.7 4.06 ± 0.0 Control 0 24 1 3.7 0.1c± 0.0 1 La* or Ce* represents the chloride forms, LaCl3 or CeCl3, respectively; all other treatments used the acetate forms of La(C2H302>3 or Ce(C2H302)3. 2 Mean weights of treatment groups after 24-day treatment period and 14-day rinse period completed. tration (Table 1). Groups treated with lanthanum or cerium at 50 ug-L"1 daily had the greatest accumu- lation, whereas the groups treated with 150 ug-L"1 every third day had the least. In the groups treated with either element, lanthanum and cerium were accumulated in approximately equal amounts. Discussion Coho salmon fry were successfully marked with lan- thanum or cerium that was administered through the water supply. The lanthanides were detected in the vertebral columns of marked fry, which is con- sistent with previous findings (Ennevor, 1991; Ennevor and Beames, 1993) and with the bone-seek- ing characteristics of the lanthanide elements (Durbin et al., 1956; Jowsey et al., 1958). Ennevor and Beames ( 1993) have shown that lanthanides that are deposited in the vertebral column, otoliths, and scales remain in these tissues for at least 10.5 months after marking. Michibata (1981) also successfully marked medaka, Oryzias latipes, and goldfish, Caras- sius auratus, with samarium, another lanthanide, and these fishes retained detectable amounts of the element in their scales one year after marking. NOTE Ennevor: Marking Oncorhynchus kisutch fry 473 Coho salmon fry exposed to 50 ugL-1 of lantha- num or cerium daily resulted in higher levels of ac- cumulation than fry exposed intermittently to con- centrations of 100 ug-Lr1 or 150 ug-L-1. In tanks with higher concentrations, number of mortalities in- creased as deposition of the elements in the verte- bral column decreased. Therefore, toxicity and accu- mulation may be related to element concentration during treatments rather than accumulated expo- sure. A high concentration of lanthanides may im- pair gill function and prevent further uptake of lan- thanides, as well as essential ions and oxygen (BehrensYamada and Mulligan, 1990). Consequently, marking with a low concentration of lanthanide over an extended period is highly recommended. A potential concern is the ability to detect the lan- thanide mark in the bony tissues of returning adults. Because fish continually accumulate calcium in their bony tissues after marking, the relative amount of lanthanum or cerium will decline gradually as the fish grows (Behrens Yamada et al., 1979; Behrens Yamada and Mulligan, 1982). Marks laid down dur- ing freshwater growth stages will be concentrated in the center portion of bony tissues. A possible solu- tion to this dilution problem is to analyze only the center where the element concentration would be about the same as when marked (Behrens Yamada and Mulligan, 1982). Scales of returning adults may be more suitable for sampling and analysis as they retain higher lanthanide concentrations (Ennevor and Beames, 1993). Also, scales are easier for sam- pling and can be removed for lanthanide determina- tion without sacrificing the fish. These studies demonstrate the successful mark- ing of experimental groups of fry with lanthanum and cerium applied through the water supply. This technique can be adapted to mark large groups of juvenile salmon at hatchery stages quickly and effi- ciently without affecting growth or survival. Mass marking with lanthanides can mark large groups of fish for identification without apparent deleterious effects. In addition, the mark remains in the bony tissues for extended periods of time, and samples of bony tissues (i.e. scales and opercular punches) can be taken from marked fish, without sacrificing the fish, for identification by ICP-MS analyses. Acknowledgments I am grateful to K. R. Pitre, E. A. Perry, and F. K. Sandercock for their helpful comments on the manu- script. The assistance of staff and the facilities ex- tended by Capilano Hatchery during the course of the'-e experiments are appreciated. Thanks also go to R. Brown and his staff at Elemental Research Inc. who performed the ICP-MS analyses. Finally, I wish to acknowledge the continued financial support of the Department of Fisheries, Canada. Literature cited Behrens Yamada, S., and T. J. Mulligan. 1990. Screening of elements for the chemical mark- ing of hatchery salmon. Am. Fish. Soc. Sympo- sium 7:550-561. 1982. Strontium marking of hatchery-reared coho salmon, Oncorhynchus kisutch Walbaum, identifi- cation of adults. J. Fish. Biol. 20:5-9. Behrens Yamada, S., T. J. Mulligan, and S. J. Fairchild. 1979. Stontium marking of hatchery-reared coho salmon (Oncorhynchus kisutch Walbaum). J. Fish. Biol. 14:267-275. Durbin, P. W., M. H. Williams, M. Gee, R. H. Newman, and J. G. Hamilton. 1956. Metabolism of the lanthanons in the rat. Proc. Soc. Exp. Biol. Med. 91:78-85. Ellis, W. C. 1968. Dysprosium as an indigestible marker and its determination by radio-activation analysis. J. Agric. Food Chem. 16:220-224. Ennevor, B.C. 1991. The feasibility of using lanthanide elements to mass mark hatchery-production salmon. M.Sc. the- sis, Univ. British Columbia, Vancouver, B.C., 182 p. Ennevor, B. C, and R. M. Beames. 1993. The use of lanthanide elements to mass mark juvenile salmon. Can. J. Fish. Aquat. Sci. 50:1039-1044. Houk, R. S., and J. J. Thompson. 1988. Inductively coupled plasma mass spec- trometry. Mass Spectrom. Rev. 7:425^461. Jowsey, J., R. E. Rowland, and J. H. Marshall. 1958. The deposition of rare earths in bone. Rad. Res. 8:490-501. Kyker, G. C. 1961. Rare earths. In C. L. Comar and F. Bronner (eds.), Mineral metabolism, Vol. II, Part B, p. 499- 529. Academic Press, New York. Longerich, H. P., B. J. Fryer, D. F. Strong, and C. J. Kantipuly. 1987. Effects of operating conditions on the deter- mination of the rare earth elements by inductively coupled plasma mass spectrometry (ICP-MS). Spectrochim. Acta 42B:75-92. Luckey, T. D., and B. Venugopal. 1977. Lanthanides, the rare-earth metals. In T. D. Luckey and B. Venugopal (eds.), Metal toxicity in mammals, Vol. II. Plenum Press, New York. Michibata, H. 1981. Labeling fish with an activable element through their diet. Can. J. Fish. Aquat. Sci. 38:1281-1282. Wilkinson, L. 1989. SYSTAT: the system for statistics. Systat, Inc., Evanston, IL, 822 p. Distribution and relative abundance of the blue shark, Prionace glauca, in the southwestern equatorial Atlantic Ocean Fa bio H. V. Hazin Department of Marine Science and Technology Tokyo University of Fisheries 5-7, Konan-4. Minato-ku, Tokyo 1 08, Japan The Brazilian Research Council CNPq (Conselho Nacional de Desenvolvimento Cientifico et Technologico) Avenida W3 Norte, Quadra 511. Bloco A., Ed. Birtar II Brasilia. D F, CEP-75000-000, Brazil Clara E. Boeckman Elizabeth C. Leal Universidade Federal Rural de Pernambuco, Rua Dom Manoel de Medeiros s/n. Dois Irmaos, Recife-PE, Brazil Rosangela R T. Lessa Brazilian Research Council, fCNPq) Avenida W3 Norte, Quadra 511, Bloco A , Ed Bittar II Brasilia, D F, CEP-75000-000, Brazil Kohei Kihara Kazuyuki Otsuka Department of Marine Science and Technology Tokyo University of Fisheries 5-7, Konan-4. Minato-ku, Tokyo 1 08, Japan The blue shark, Prionace glauca, is one of the most abundant oce- anic-epipelagic sharks and is prob- ably the widest ranging chon- drichthyian (Compagno, 1984). It is frequently caught by tuna longline fisheries in temperate, subtropical, and tropical waters of the world oceans (Pratt, 1979). Hazin et al. (1990) investigated the distribution and abundance of pelagic sharks caught from 1983 until 1988 by Brazilian longliners in the southwestern equatorial Atlantic. Blue shark and sharks of the genus Carcharhinus were the dominant species, together repre- senting nearly 95% of the shark catches (Hazin et al., 1990). They 474 reported that blue shark abun- dance had a marked seasonal fluc- tuation with the highest catches taking place during the third and fourth quarters of the year and the lowest in the first quarter. The objective of the present study is to further investigate the distribution and relative abundance of the blue shark, Prionace glauca, in the south-western equatorial At- lantic Ocean, including the follow- ing aspects: a) seasonal fluctuation of catch per unit of effort (CPUE) as related to sea surface temperature; b) sex, and size and age composition of blue shark catches; c) vertical dis- tribution as related to the vertical temperature profile. Material and methods This study was based on shark catches during 50 fishing cruises by a commercial tuna longliner, FV Argus, from August to December 1987 and from February 1990 to December 1991. On 325 longline operations during these cruises, a total of 992 blue sharks were caught. The commercial longline consisted of 120 baskets, each with 7 branch lines. The bait was fro- zen Brazilian sardine, Sardinella brasiliensis. Average local time of set, retrieval, and mean soaking time of the longline is shown in Table 1. Further details of longline fishing gear and methods were described in Hazin ( 1986). The fishing ground was located between lat. 2°S and 7°S and long. 32°W and 38°W. Fishing areas were divided into segments of 1° latitude x 1° longitude. The cen- tral positions of the 325 longline sets ( Fig. 1 ) were calculated as the average latitude and longitude of the beginning and end of set and retrieval. Distribution of fishing effort by month is presented in Table 2. Blue shark relative abun- dance was expressed as average catch, in number of fish, per 100 hooks (CPUE). The mean CPUE was calculated as the total catch, in number of fish, divided by the total fishing effort, in 100 hooks. The distribution of blue shark mean CPUE by segments of 1° lati- tude x 1° longitude was observed and its relation to ocean depth analyzed. Ocean depths for each longline set were calculated as the weighted mean of the three clos- est values read from the nautical chart number 50, issued by the Brazilian Navy. The monthly fluctuation of CPUE and sex ratio was analyzed and compared to sea surface tem- perature. An analysis of variance Manuscript accepted 21 October 1993 Fishery Bulletin 92: 474-480 ( 1994) NOTE Hazin et al.: Distribution and relative abundance of Prionace glauca 475 (ANOVA) was performed to determine whether blue shark CPUEs were significantly different among months. The mean CPUE for each month was calcu- lated and all monthly mean CPUEs were then com- Table 1 Average set, retrieval and soaking time (in decimals) with standard deviations, for longline operations of FV Argus from February to December 1990. Local time Standard deviation Set Beginning End 1.61 4.00 0.58 0.53 Retrieval Beginning End 10.31 17.52 1.47 1.68 Soaking time (hours) 11.11 1.54 pared by ANOVA, through one-way classification of data. The months were the only independent vari- able (Table 3). No CPUE data were available from October 1990 and from January and April 1991. Data on the sex of the specimens were not available from January to October 1991. All lengths are reported as fork length (FL), which was measured from the tip of the snout to the fork of the tail. Blue sharks were always measured at the time of landing. Length data were available only for 1990 and 1991. To better understand seasonal varia- tion in CPUE, CPUE data for the various age classes of male specimens from February through Decem- ber 1990 were evaluated. Females were not included because they were present only from February to July. To calculate the age-CPUE distribution, male blue shark lengths were converted to age by Stevens' (1975) growth equation, as follows: L< = 423(l-e-°11,t+1035)) n ft / L^J^Lrf s~* A < ^ V LkJ y \ f ] r y / 3°S 4°S 5°S 6°S 37°W 3 S ° W 33°W 100°W 80°W 60°W 40°W 2fl°W 0° 20°E Figure 1 Location of the fishing ground (hatched area) and central positions of the longline sets made by the FV Argus from August to December 1987 and from February 1990 to December 1991, in the southwestern equatorial Atlantic Ocean. Table 2 Distribution of fishing effort of FV Argus by months, from August December 1991, in the southwestern equatorial Atlantic Ocean. to December 1987 and from February 1990 to Jan. Feb Mar. Apr. May Jun. Jul. Aug. Sep. Oct. Nov. Dec. Total No. of sets 16 12 23 15 27 24 23 32 36 It 39 34 325 % 5 4 7 5 8 7 7 l(i 11 13 12 11 100 476 Fishery Bulletin 92|2). 1994 where Lt = total length, and t - age in years. Fork length data were converted to total length by the regression of Hazin et al. (1991): FL = 11.27 + 0.78 TL where FL - fork length, and TL = total length. Depths of longline hooks were estimated using the equations of Yoshihara (1952, 1954, a and b). Verti- cal distribution of males and females was studied through the relative distribution of mean catches on longline hooks, during February to June and July to December 1990. Differences in mean catch of males and females on longline hooks were evaluated by chi- square analysis (df=6). Sea water temperature from 0 to 300 m was surveyed in 35 DBT (digital bathy- thermograph) profiles: 6 in May 1990, 3 in May 1991, 5 in June 1991, 13 in November 1990, 2 in Novem- ber 1991, and 6 in December 1990. From January to December 1990, sea surface temperature was mea- sured by a mercury thermometer. Results From August to December 1987 and from February 1990 to December 1991, the blue shark mean CPUE by quadrates increased eastward, being particularly high east of long. 35°W (Fig. 2). Of the 325 sets, 260 (nearly 80%) were over bottom depths greater than 1,000 m. In these areas, the mean CPUE of blue shark was 0.50. The remaining 65 sets were in areas with depths shallower than 1,000 m, close to oceanic banks (Aracati, Sirius, and Guara banks), and west of 35°W. In these areas, the mean CPUE of blue shark was only 0.05. The fluctuation of the monthly mean CPUE of blue shark in the area east of 35°W, over ocean depth of 1,000 m, for 1987, 1990 and 1991, were similar (Fig. 3). In 1990 and 1991 the CPUE was low until May, increased during June and July, decreased again in August, increased during September and October, and decreased once more in November and Decem- ber. In 1987 the CPUE was low in August, increased in October and decreased during November and De- cember. Differences in mean CPUE among months were significant (ANOVA; P<0.0001; Table 3). The fluctuation of the monthly mean CPUE of male and female blue sharks was distinct (Fig. 4). CPUE for females was highest during March. From July to December, CPUE for females was low in this fishing ground. CPUE for males, however, was lowest dur- ing March, after which abundance increased and peaked during September and October. The sea sur- face temperature in 1990 was highest in May and lowest in September. During this year, in general, the CPUE of males tended to decrease with an in- crease in the sea surface temperature, whereas the CPUE of females tended to increase. 38 °W 36°W 34°W 3:° IV C=0.00 OO. 75 O0.48 n= 8 n= 12 n= 14 3°S C=0.0 3 C=0.02 C=0.04 C=0.29 C=0. 57 C=0.55 5°S X n= 42 n= 16 n= 25 n= 44 n = 64 N C= 0.110 \n= 8 C=0.04 n= 12 C = 0. r.n n= ID C=0 . 46 n = 4 2 i D.17 n= 4 '•"^SsfS^x ) C=0.40 \ n= 8 Natal ®1 BRAZIL ! \ C=0. 2 0 C=o . 70 1! ! Figure 2 Mean catch per 100 hooks (CPUE) of blue shark, Prionace glauca, in the southwestern equatorial Atlantic Ocean, from August to December 1987 and from February 1990 to December 1991. C= CPUE; n= number of longline sets. Jan . Feb . Mar . Apr . May Jun . Jul . Aug . Sep . Oct . Nov . Dec . Month 1987 1991 Figure 3 Monthly mean catch per 100 hooks (CPUE) of blue shark, Prionace glauca, in the southwestern equato- rial Atlantic Ocean, in the area east of long. 35°W and with ocean depth over 1,000 m, from August to Decem- ber 1987 and from February 1990 to December 1991. NOTE Hazin et al.: Distribution and relative abundance of Prionace glauca 477 Of 810 specimens, 652 (about 80%) were male and 158 were female (about 20%). Overall, the sex ratio (male/female) for the entire period was 4.12:1. The sex ratios for each month are given in Table 4. Females ranged in size from 162 to 226 cm and males from 156 to 250 cm (Fig. 5). Seasonal fluc- tuation in male CPUE during 1990 was different among age groups (Fig. 6). From February to May, ages ranged from 4.5 to 8.5 years, with most (83%) individuals between 6 and 8.5 years old (83%). The CPUE rise in June— July was due to an increase in the same age classes as from February-May. Some older specimens from 9 to 10.5 years also appeared. The CPUE offish younger than 7 years decreased markedly in August. During Septem- ber-October the CPUE of ages 7.5 to 10 increased sharply. In November-December the CPUE of age classes from 7.5 to 8.5, and also 9.5, were reduced, whereas the CPUE of other age classes did not change much. From February to July, younger fish (4.5 to 7.5 years) were more abundant than from August to December, their CPUE being particularly high in June-July From August to December, fish older than 7.5 years had a higher CPUE than in the previous months, with a peak in September-October. The calculated depth of longline hooks during 137 longline fishing operations of 1990 ranged from 87 to 206 m (Table 5). The figures given in this table approximate the actual depths of longline hooks. The most striking feature of the DBT profiles is the pres- Jan . Feb . Mar . Apr . May Jun . Jul . Aug . Sep . Oct . Nov . Dec . Month Male 1987 -*- Male 1990 -t— Male 1991 Female 1987 — *— Female 1990 *- Female 1991 — • — Sea surface temperature Figure 4 Monthly mean sea surface temperature in 1990 and monthly mean catch per 100 hooks (CPUE) of male and female blue shark, Prionace glauca, in the south- western equatorial Atlantic Ocean, from August to December 1987 and from February 1990 to Decem- ber 1991. Table 3 Analysis of variance (ANOVA) and expected mean squares for comparison of monthly mean CPUEs of blue sharks, Prionace glauca, caught by the FV Argus, during 325 longline sets, from August to December 1987 and from February 1990 to December 1991, in the southwestern equatorial Atlantic Ocean. Source of variation Degrees of freedom Sum of squares Mean squares Month Error Total 24 300 324 11.41 30.68 42.09 0.48 0.10 4.65 <0.0001 155-165 176-185 196-205 216-225 236-245 166-175 186-195 206-215 226-235 246-255 Fork Length (cm) n= 104 155-165 176-185 196-205 216-225 236-245 166-175 186-195 206-215 226-235 246-255 Fork Length (cm) n= 374 Figure 5 Length-frequency distribution of male and female blue shark, Prionace glauca, in the southwestern equatorial Atlantic Ocean, from February to Decem- ber 1990 and November to December 1991. Males: n = 374; min. = 156 cm; max. = 250 cm. Females: n = 104; min. = 162 cm; max. = 226 cm. 478 Fishery Bulletin 92|2). 1994 Feb. 2.08 0.48 ence of a steep thermocline during the entire year. The range and mean of sea water temperature at the calculated depths of longline hooks, and the mean depth and temperature at the top and bottom of the thermocline in May-June and November-December are shown in Table 6. All longline hooks were located in or below the thermocline (Fig. 7). During Nov.- Dec. the thermocline was about 30 m deeper than during May-June. Male blue shark catches from February to June were concen- trated among central, deeper hooks, whereas in the second half of the year they were more uni- formly distributed (Fig. 8). Male catch was significantly different among hooks from February to June (chi-square; P<0.005), but not from July to December (0.10 < P<0.25). The difference of male catches, on hooks 3, 4, and 5 between February-June and July-Decem- ber was significant (0.05 < P<0.10). The relative dis- tribution of female blue shark catches along longline hooks from February to July (Fig. 8) was different from that of males; the highest catches took place on hooks 2 and 5. The difference between male and fe- male catches among hooks during February to July was significant (0.01 < P<0.05). The distribution of male and female catch among hooks suggests that males were distributed in shallower waters between July and December than between February and June and that females from February to July had a shal- lower distribution than males. From February to July, CPUE for females was highest in hooks 2 and 5 (130 to 165 m), and CPUE Table 4 Blue shark, Prionace glauca, monthly sex ratio (number of males per female), in the southwestern equatorial Atlantic Ocean. Mar. Apr. May Jun. Jul. Aug. Sep. Oct. Nov. Dec. 1.29 1.79 2.71 2.35 6.00 54.00 25.00 16.00 8.50 Table 5 Mean and range of calculated depths (m) of longline hooks, during 137 longline operations, in the southwestern equatorial Atlantic Ocean, from February to December 1990. Hook no. Mean depth Minimum Maximum 1 and 7 92.66 2 and 6 131.56 3 and 5 164.19 4 179.20 87 96 120 140 145 181 154 206 for males in hooks 3 and 4 (165 to 180 m) (Table 5). The corresponding range of sea water temperature was 13.8° to 20.4°C for females and 12.9° to 17.1°C for males (Table 6). Male catch among hooks from July to December, was much more uniform, suggest- ing a depth range for males from 90 to 180 m (Table 5), or from about 15° to 28°C (Table 6), up to the top of the thermocline (Fig. 7). Discussion The results on horizontal distribution agree with those of Hazin et al. (1990), corroborating the oce- anic character of blue shark (Fig. 2). Abundance was also clearly related to bottom depth. The mean CPUE of blue shark in areas of bottom depths greater than 1,000 m was 10 times higher than in shallower ar- eas. The lower abundance of blue sharks west of 35°W Table 6 Range and mean of sea water temperature ( "C ) at calcula temperature at the top and bottom of the thermocline, as carried out in May-June and November-December 1990 „ed depths (m) of longline hooks 1 to 7 and mean depth and inferred from 35 DBT (digital bathythermograph) surveys and 1991, in the southwestern equatorial Atlantic Ocean. Month Hook number Thermocline Top Bottom 1 and 7 2 and 6 3 and 5 I Depth. Temp. Depth. Temp. May-June Temp, mean 22.6 16.3 14.8 13.2 55 28 120 13.5 Temp, range 18.5-25.7 13.8-20.4 13.1-17.1 12.9-13.3 Nov. -Dec. Mean 27.6 17.0 15.4 14.9 85 27 130 15.5 Range 26.8-29.1 14.3-24.0 14.0-20.1 13.9-18.6 NOTE Hazin et al.: Distribution and relative abundance of Prionace giauca 479 (Fig. 2), therefore, is probably a consequence of ocean depth, because almost all sets performed in this fish- ing ground took place in shallow depths and in the vicinity of oceanic banks. The monthly mean CPUEs of males and females (Fig. 4) show that the higher relative abundance of blue shark during the third and fourth quarters of the year (Hazin et al., 1990) is mostly comprised by males. These results indicate also that males and females are segregated and that their migratory movements are different. The different seasonal fluc- tuation of CPUE for different male age groups indi- cate that male specimens were also segregated by size. The use of Stevens' (1975) growth equation to calculate these ages from fork length may have limi- tations because it is based on data from the North Atlantic Ocean. Nevertheless, Amorim (1992) stud- ied the growth of blue shark in the south-western Atlantic and found a value of k- 0.1126, which ap- proximates Stevens' value of 0.11. Differences in vertical distribution displayed by male and female blue sharks (Fig. 8) indicate that vertical sexual segregation likely occurred in the first half of the year. They also suggest that the depth range of male blue sharks may change seasonally. Acknowledgments We sincerely thank the Norte Pesca S/A, who pro- vided all the data used in this research. We are par- ticularly indebted to Alceu A. Couto and Anibal P. Souza for their valuable help in data collection. We also thank Francis G. Carey of the Woods Hole n= 320 Age (year) ^J Feb. -May | | Jun.-Jul. I I Aug. ~~ Sep. -Oct. | | Nov. -Dec. Figure 6 Mean catch per 10,000 hooks of male blue shark, Prionace giauca, by age classes, from February through December 1990, in the southwestern equa- torial Atlantic Ocean, n = 320. Oceanographic Institution who constructively criti- cized the manuscript, and Sakutaro Yamada for his technical assistance. Financial support was granted by the Ministry of Education of the Japanese Gov- ernment through the Mombusho Scholarship Pro- gram and by the Brazilian Research Council (CNPq). May Temperature (°C) 14 18 22 26 30 November Figure 7 Typical temperature profiles for May and Novem- ber, in the southwestern equatorial Atlantic Ocean, and the calculated mean depths of the hooks 1 to 7 of the longline of the FV Argus. 480 Fishery Bulletin 92(2). 1994 Hook Number Figure 8 Male and female blue shark, Prionace glauca, catch- frequency distribution along the hooks of the longline, in the southwestern equatorial Atlantic Ocean, during February to June and July to December, 1990. Hazin, F. H. V. 1986. Pesca de atuns e afins com embarcacao de pequeno porte no nordeste brasileiro. Graduation thesis, Universidade Federal Rural de Pernambuco, Recife, Brazil, 107 p. Hazin, F. H. V., A. A. Couto, K. Kihara, K. Otsuka, and M. Ishino. 1990. Distribution and abundance of pelagic sharks in the south-western equatorial Atlantic. J. Tokyo Univ. Fish. 77(l):51-64. Pratt, H. L., Jr. 1979. Reproduction of the blue shark, Prionace glauca. Fish. Bull. 77:445-470. Stevens, J. D. 1975. Vertebral rings as a means of age determi- nation in the blue shark (Prionace glauca L. ). J. Mar. Biol. Assoc. U.K. 55:657-665. Yoshihara, T. 1952. Distribution of catches of tuna longline. Ill: Swimming depth. [In Japanese.] Nippon Suisan Gakkaishi, 18(8):187-190. 1954a. Distribution of catches of tuna longline. rV: On the relation between K and psi with a table and a diagram. [In Japanese.] J. Tokyo Univ. Fish. 19(10):1012-1014. 1954b. On the distribution of catches by tuna longline. J. Tokyo Univ. Fish. 41(l):l-26. Literature cited Amorim, A. F. 1992. Estudo da biologia, pesca e reproducao do cacao azul, Prionace glauca L. 1758, capturado no sudeste e sul do Brasil. Ph.D. thesis, Universidade Estadual Paulista, Rio Claro, Sao Paulo, Brazil, 176 p. Compagno, L. J. V. 1984. FAO species catalogue. Vol. 4, Parts 1 and 2: Sharks of the world: an annotated and illustrated catalogue of shark species known to date. FAO Fish. Synop. 125, 655 p. Fishery Bulletin Guide for Contributors Preparation Title page should include authors' full names and mailing addresses and the senior author's telephone and FAX number. Abstract should not exceed one double-spaced typed page. It should state the main scope of the research but emphasize its conclusions and relevant findings. Because abstracts are circulated by abstracting agencies, it is important that they represent the research clearly and concisely. Text must be typed double-spaced throughout. A brief introduction should portray the broad significance of the paper; the remainder of the paper should be divided into the following sections: Materials and methods, Results, Discussion (or Conclusions), and Acknowledgments. Headings within each section must be short, re- flect a logical sequence, and follow the rules of multiple subdivision (i.e. there can be no subdivision without at least two items). The entire text should be intelligible to interdisciplinary readers; therefore, all acronyms, abbreviations, and technical terms should be spelled out the first time they are mentioned. The scientific names of species must be written out the first time they are mentioned; subsequent mention of scientific names may be abbreviated. Follow the U.S. Government Printing Office Style Manual (1984 ed.) and the CBE Style Manual (5th ed.) for editorial style, and the most current issue of the American Fisheries Society's Common and Scientific Names of Fishes from the United States and Canada for fish nomencla- ture. Dates should be written as fol- lows: 11 November 1991. Measure- ments should be expressed in metric units, e.g., metric tons as (t); if other units of measurement are used, please make this fact explicit to the reader. The numeral one (1) should be typed as a one, not as a lower-case el (1). Text footnotes should be numbered with Arabic numerals and typed on a separate sheet of paper. Footnote all personal communications, unpub- lished data, and unpublished manu- scripts with full address of the com- municator or author, or, as in the case of unpublished data, where the data are on file. Authors are advised to avoid references to nonstandard (gray) literature, such as internal, project, processed, or administrative reports, wherever possible. Where these refer- ences are used, please include whether they are available from NTIS (Nation- al Technical Information Service) or from some other public depository. Literature cited comprises published works and those accepted for publica- tion in peer-reviewed literature (in press). Follow the name and year sys- tem for citation format. In the text, cite Smith and Jones (1977) or (Smith and Jones, 1977). If there is a se- quence of citations, list chronological- ly: Smith, 1932; Green, 1947; Smith and Jones, 1985. Abbreviations of serials should conform to abbrevia- tions given in Serial Sources for the BIOSIS Previews Database.. Authors are responsible for the accuracy and , completeness of all citations. Tables should not be excessive in size and must be cited in numerical order in the text. Headings should be short but ample enough to allow the table to be intelligible on its own. All unusual symbols must be explained in the table heading. Other incidental comments may be footnoted with italic numerals. Use asterisks for probability in statis- tical data. Because tables are typeset, they need only be submitted typed and formatted, with double-spaced legends. Zeros should precede all decimal points for values less than one. Figures include line illustrations and photographs (or slides) and must be cited in numerical order in the text. Unless photographs are submitted on glossy paper with good contrast, we cannot guarantee a good final printed copy. Figures are to be labeled with author's name and number of figure. Use Times Roman font (upper and lowercase letters) to label within figures. Avoid vertical lettering except for y-axis labels. Zeros should precede all decimal points for values less than one. Figures may be submitted as com- puter software files (along with laser- printed copies), as photomechanical transfers (PMTs), or as high quality photographic prints. Send only xerox copies of figures to the Scientific Edi- tor; original figures will be requested later when the manuscript has been accepted for publication. Figure leg- ends should explain all symbols and abbreviations and should be double- spaced on a separate page at the end of the manuscript. Copyright law does not cover govern- ment publications; they fall within the public domain. If an author repro- duces any part of a government publi- cation in his work, reference to source is considered correct form. Submission Send printed copies (original and two copies) to the Scientific Editor: Dr. Ronald Hardy, Scientific Editor Northwest Fisheries Science Center, F/NWC3 National Marine Fisheries Service, NOAA 2725 Montlake Boulevard East Seattle, WA 98112-2097 Once the manuscript has been ac- cepted for publication, you will be asked to submit a software copy of your manuscript to the Managing Editor. The software copy should be submitted in WordPerfect text format (or in standard ASCII text format if WordPerfect is unavailable) and should be placed on a 5.25-inch or 3.5- inch disk that is double-sided, double or high density, and that is compatible with either DOS or Apple Macintosh systems. Copies of published articles and notes are available free of charge to the senior author (50 copies) and to his or her laboratory (50 copies). Addition- al copies may be purchased in lots of 100 when the author receives page proofs. / 5333 036 ■TmiLi,„M101 library wh nws x MBL w„rtMa™ BjologfcaJ Laboratory/ ^^_ _ _ Wooat H°"> Oceanographlc InsBtution U.S. Department f^ ■ ■ of Commerce 1 I S ll ^ l^\/ JUL 2 ° 1"4 Volume 92 ' ™ *^ ■ ■ ^* * Number 3 —^ - _ ST Bulletin^ U.S. Department of Commerce Ronald H. Brown Secretary National Oceanic and Atmospheric Administration D. James Baker Under Secretary for Oceans and Atmosphere National Marine Fisheries Service Rolland A. Schmitten Assistant Administrator for Fisheries * X * Scientific Editor Dr. Ronald W. Hardy Northwest Fisheries Science Center National Marine Fisheries Service, NOAA 2725 Montlake Boulevard East Seattle. Washington 981 12-2097 Editorial Committee Dr. Andrew E. Dizon National Marine Fisheries Service Dr. Linda L. Jones National Marine Fisheries Service Dr. Richard D. Methot National Marine Fisheries Service Dr. Theodore W. Pietsch University of Washington Dr. Joseph E. Powers National Marine Fisheries Service Dr. Tim D. Smith National Marine Fisheries Service 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, Attn: Chief, Mail List Branch, Mail Stop SSOM, Washington, DC 20402-9373. 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 peri- odical 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 ap- proved 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: $27.00 domestic and $33.75 foreign. Cost per single issue: $13.00 domestic and $16.25 foreign. See back page for order form. Managing Editor Sharyn Matriotti National Marine Fisheries Service Scientific Publications Office 7600 Sand Point Way NE, BIN C I 5700 Seattle, Washington 98 1 1 5-0070 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 92 Number 3 July 1994 Fishery Bulletin Contents in Errata Articles 481 Castillo, Gonzalo C, Hiram W. Li, and James T. Golden Environmentally induced recruitment variation in petrale sole, Eopsetta jordani 494 Creaser, Edwin R, and Herbert C. Perkins The distribution, food, and age of juvenile bluefish, Pomatomus saltatrix, in Maine 509 David, Andrew W., J. Jeffery Isley, and Churchill B. Grimes Differences between the sagitta, lapillus, and asteriscus in estimating age and growth in juvenile red drum, Sciaenops oceltatus 5 1 6 Dee, Anderson J., and James D. Parrish Reproductive and tropic ecology of the soldierfish Myrlpristis amaena in tropical fisheries 531 Di Giacomo, Edgardo E., and Maria Raquel Perier Reproductive biology of the cockfish, Callorhynchus callorhynchus (Holocephali: Callorhynchidae), in Patagonian waters (Argentina) 540 Kendall, Arthur W., Jr., Lewis S. Incze, Peter B. Ortner, Shailer R. Cummings, and Patricia K. Brown The vertical distribution of eggs and larvae of walleye pollock, Theragra chalcogramma, in Shelikof Strait, Gulf of Alaska 555 Lowerre-Barbieri, Susan K., Mark E. Chittenden Jr., and Cynthia M. Jones A comparison of a validated otolith method to age weakfish, Cynoscion regalis, with the traditional scale method Fishery Bulletin 92(3), 1994 569 McKinnon, Jeff Feeding habits of the dusky dolphin, Lagenorhynchus obscurus, in the coastal waters of central Peru 579 Muter, Franz-Josef, and Brenda L. Norcross Distribution, abundance, and growth of larval walleye pollock, Theragra chalcogramma, in an Alaskan fjord 591 Norcross, Brenda L, and David M. Wyanski Interannual variation in the recruitment pattern and abundance of age-0 summer flounder, Paralichthys dentatus. in Virginia estuaries 599 Sarda, Francisco, Joan E. Cartes, and Walter Norbis Spatio-temporal structure of the deep-water shrimp Aristeus antennatus (Decapoda: Aristeidae) population in the western Mediterranean 608 Steimle, Frank W., Dorothy Jeffress, Stephen A. Fromm, Robert N. Reid, Joseph J. Vitaliano, and Ann Frame Predator-prey relationships of winter flounder, Pleuronectes americanus, in the New York Bight apex 620 Weinberg, Kenneth L. Rockfish assemblages of the middle shelf and upper slope off Oregon and Washington Notes 633 DeLancey, Lawrence B., James E. Jenkins, and J. David Whitaker Results of long-term, seasonal sampling for Penaeus postlarvae at Breach Inlet, South Carolina 641 Forward, Richard B., Jr., William F. Hettler, and Donald E. Hoss Swimbladder deflation in the Atlantic menhaden, Brevoortia tryannus 647 Gilman, Sharon L. An energy budget for northern sand lance, Ammodytes dubius, on Georges Bank, 1977-1986 655 Stanley, Richard D., Bruce M. Leaman, Lewis Haldorson, and Victoria M. O Connell Movements of tagged adult yellowtail rockfish, Sebastes flavidus, off the west coast of North America 664 Xiao, Yongshun, and David C. Ramm A simple generalized model of allometry, with examples of length and weight relationships for 1 4 species of groundfish Errata Thompson, Grant G. Variations on a simple dynamic pool model Fish. Bull. 91(4):718-731 Equation 1 should read as follows: w(a) = wr a-an xo ) Equation 26 should read as follows: Bf(F' b(F',a f M 1+K'/+F (l+F'r Appendix Equation A5 should read as follows: BPR(F')= fwr | -^^o_ JQr ^ ar - a0 j e-Ma+F-Ha-ar)da w\ y (n)kK"k R = \) Marine Biological Laboratory/ Woods Hole OceJnooraphic iXion Library JUL 2 0 1994 Woods Hole, MA ua&u Volume and page numbers for the following cita- tion should be corrected to read as follows: Thompson, G. G. 1992. Management advice from a simple dynamic pool model. Fish. Bull. 90:552-560. The National Marine Fisheries Service (NMFS) does not approve, recom- mend, or endorse any proprietary product or proprietary material mentioned in this publication. No reference shall be made to NMFS, or to this publica- tion furnished by NMFS, in any advertising or sales promotion which would indicate or imply that NMFS approves, recommends, or endorses any proprietary product or proprietary material mentioned herein, or which has as its purpose an intent to cause directly or indirectly the advertised product to be used or purchased because of this NMFS publication. Abstract. — Potential effects of parental stock size and environ- mental factors on year-class strength (YCS) of petrale sole, Eopsettajordani, were investigated in two areas off Oregon and Wash- ington (Pacific States Marine Fish- eries Commission areas 2B: 42°50'N^4°18'N, and 3A: 45°46'N- 47°20'N). Parental egg production indices and YCS were not consis- tently correlated over the period 1970 to 1977. Variation in YCS be- tween 1958 and 1977 was associ- ated with oceanographic conditions from winter to early spring, the pe- riod in which pelagic larval stages are most abundant. A regression model based on indices of offshore Ekman transport from January to March and alongshore transport from December to February ac- counted for nearly 55% of the YCS variation in Area 2B. In Area 3A, the previous two indices plus sea surface temperature from Decem- ber to February explained about 65% of the YCS variation. Inshore advection of eggs and larvae could favor settlement of juveniles into nearshore areas and increase the subsequent recruitment strength of petrale sole. Environmentally induced recruitment variation in petrale sole, Eopsettajordani Gonzalo C. Castillo Hiram W. Li Oregon Cooperative Fishery Research Unit, US Fish and Wildlife Service Nash Hall 1 04, Department of Fisheries and Wildlife, Oregon State University, Corvallis, Oregon 97331 James T. Golden Hatfield Marine Science Center. Oregon Department of Fish and Wildlife Building 3, Newport, Oregon 97365 Manuscript accepted 8 December 1993. Fishery Bulletin 92:481-493 ( 1994). Recruitment fluctuations in fish populations are ascribed to many physicochemical factors and biologi- cal processes, including parental stock size and fishing (Ricker, 1975; Shepherd et al., 1984). However, the importance of various factors to the recruitment of most species of fish is virtually unknown. Although year-class strength (YCS) of many fishes is thought to be determined at the egg and larval stages (Sharp1; Rothschild and Rooth, 1982), it may also be significantly affected during the postlarval stages (Smith, 1981; Sissenwine, 1984). Increasing evidence suggests that oceanographic conditions affect the recruitment of many fishes in the Northeast Pacific Ocean (e.g. Par- rish et al., 1981; Bailey and Incze, 1985; Hollowed et al., 1987; Botsford et al., 1989). A recent hy- pothesis suggests that recruitment of groundfish off the west coast of the United States is related to the timing of the spring transition, a period of major changes in oceano- graphic conditions.2 On the Or- egon-Washington shelf oceano- graphic conditions exhibit strong seasonal patterns (e.g. Huyer et al., 1975; Halpern, 1976; Huyer, 1977; Landry et al., 1989). In winter. alongshore currents are northward at all depths, and cross-shore sur- face currents flow inshore resulting in downwelling. In spring, flow is southward at all depths but stron- ger near the surface. The spring transition usually occurs within a one week period during March or April (Strub et al., 1987; Strub and James, 1988). In summer, a surface coastal current flows southwest- ward and the attendant offshore transport causes upwelling; how- ever, deep flow is northward. In fall, alongshore currents are northward at all depths. Petrale sole, Eopsetta jordani, Pleuronectidae, is a commercially important flatfish of the northeast Pacific Ocean (Ketchen and For- rester, 1966). It is continuously dis- tributed from the Bering Sea (58°N- 152°W) to Baja California (32°26'N- 117 16'W) (Roedel, 1953; Hitz and 1 Sharp, G. D. 1980. Report of the workshop on effects of environmental variation on survival of larval pelagic fishes. In Sharp, G. D. (rapporteur). Report and supporting documentation of the workshop on the effects of environmental variation on the survival of larval pelagic fishes, p. 15-59. Int. Ocean. Comm. Workshop Rep. 28. 2 Lynn, R. J. Southwest Fisheries Science Center, P.O. Box 271, La Jolla, CA 92038. Personal commun., 1991. 481 482 Fishery Bulletin 92(3), 1994 Rathjen, 1965). Recruitment fluctuations of petrale sole appear to be strongly related to environmental factors ( Ketchen and Forrester, 1966). Ketchen(1956) demonstrated a positive correlation between winter sea surface temperature and recruitment of petrale sole off British Columbia from the middle 1940s to the middle 1950s. In the same area, Ketchen and Forrester ( 1966) postulated that warmer sea surface temperatures and onshore transport of pelagic early life stages could favor recruitment of this species. Two central spawning areas of petrale sole, Heceta Bank and Willapa Deep, are located off Oregon and Washington (Fig. 1). Petrale sole spawn from late fall to early spring at depths of about 300-450 m (Cleaver, 1949; Harry, 1959; DiDonato and Pasquale, 1970; Pedersen, 1975). The incubation period of newly fertilized eggs ranges from about 6 to 13 days (Alderdice and Forrester, 1971). The eggs and yolk- sac larvae are stenohaline and stenothermal (Alderdice and Forrester, 1971). Development of pe- lagic eggs and larvae occurs mainly from winter to spring, followed by the presettlement and postsettlement juvenile stages from summer to fall respectively (Fig. 2). Although petrale sole larvae V*3 200 m fv^f *? V ' '^aWJ 48 °N - z 111 o \. ?*r o WILLAPA DEEP*£ WASHINGTON o El 3A .> p =*\ 46" - Columbia River O < Q. 2C 5 f r- < _ _ _ — — ^;~ f UJ X HECETA BANK* | OREGON 44 cr O z 2B ;/? 42° " 130° W 125? | CALIFORNIA „ 120 i Figure 1 Location of Willapa Deep and Heceta Bank spawning grounds of petrale sole, Eopsetta jordani , in Pacific States Marine Fisheries Commission areas 3A and 2B. Spawn- ing areas are located nearly 50 km offshore (modified from Pedersen, 1975). I\\\l ESTIMATED ABUNDANCE PEAK OF EGGS AND LARVAE [ | OBSERVED STAGES OFF OREGON I INFEHRED DURATION OF STAGES OFF OREGON-WASHINGTON K\M C K\M JUVENILE (PHE-SETTLEUENTI [ JUVENILE (POST. SETTLEMENT! D J F WINTER M A M SPRING SON FALL Figure 2 Estimated duration of life stages of petrale sole, Eopsetta jordani, off Oregon and Washington from the beginning of the spawning period to the time in which most age-0 juveniles are found settled in the inner continental shelf. The inferred temporal oc- currence of life stages was based on the reported duration of spawning, incubation period and pres- ence of presettlement and postsettlement juveniles in the Northeast Pacific Ocean (based on Cleaver, 1949; Harry, 1959; Best, 1963; Porter, 1964; Ketchen and Forrester, 1966; Alderdice and Forrester, 1971; Gregory and Jow, 1976; and Pearcy et al., 1977). have been found from 2 to 120 km offshore, Pearcy et al. ( 1977) collected nearly 50% of them 83-120 km offshore. However, postsettlement juveniles have only been found at 18-90 m depth in the in- ner continental shelf (Ketchen and Forrester, 1966; Gregory and Jow, 1976; Pearcy et al., 1977). The recruitment patterns of petrale sole off Or- egon and Washington demonstrated consecutive series of cohorts alternating between below aver- age (weak) YCS and above average (strong) YCS over the base period 1958-77 (Castillo, 1992). Pos- sible causes for such recruitment variations have not yet been studied. Our objectives were to de- termine 1) if spawning biomass of petrale sole is correlated with YCS, 2) if YCS fluctuations are associated with selected environmental factors, and 3) the percentage of YCS variation explained by environmental factors. Data and methods We selected two locations off Oregon and Wash- ington to investigate the effect of environmental factors on YCS of petrale sole: Pacific States Ma- rine Fisheries Commission3 areas 2B (42° 50'N- 44°18'N)and3A(45"46,N-47"20'N)(Fig. 1). Indi- Named Pacific Marine Fisheries Commission ( PMFC I areas until 1990. Castillo et al.: Recruitment variation in Eopsetta jordani 483 ces of YCS for petrale sole were obtained from cohort analyses of numbers of females recruited to six years of age in areas 2B and 3A (Table 1). These YCS indi- ces represent the recruitment strength of year classes hatched from 1958 to 1977. Males were excluded from these YCS indices because of problems of increasing age underestimation in fish over 8 years.4 However, because recruitment variation was similar in males and females of younger age groups, the YCS indices should be representative of both sexes. Potential egg production was used as a proxy for spawning biomass (e.g. Hayman and Tyler, 19801. Egg production was estimated from fecundity and maturity information (Porter, 1964) and from cohort analyses of the parental stock for years 1970 to 1977 ( Castillo, 1992 ). Annual potential egg production was estimated as the sum of the age-specific products of the numbers of females, their fecundity, and their 4 Recent use of the break-and-burn technique for aging otoliths showed that males grow more slowly and lay down less year- marks (annulil on the surface of the otolith than females (William H. Barss, Oregon Department of Fish and Wildlife. Newport, OR 97365, unpubl. data). Table 1 Cohort analyses of numbers of femal e petrale sole, Eopsetta jordani reaching 6 years of age in Pacific States Marine Fisheries Commission areas ZBand 3A( Castillo, 1992 ; after Jones, 1981). Both year-class strength indices were significantly correlate d with Summed-CPUE ndices (P<0.05, Castillo, 1992; af- ter Hayman et a ., 1980). Year-class s trength (thousands offish) Cohort (year) Area 2B Area 3A 1958 208 543 1959 278 544 1960 309 624 1961 385 692 1962 231 333 1963 296 421 1964 240 365 1965 280 547 1966 487 954 1967 352 703 1968 419 737 1969 349 558 1970 426 651 1971 421 570 1972 373 463 1973 356 415 1974 259 326 1975 193 332 1976 157 286 1977 223 424 percent maturity. The number of females was esti- mated from the observed annual sex ratio in the com- mercial landings. Females composed on average 58% of petrale sole landed. Egg production was averaged for fish over age 13 years because of the scarcity of older females. Eight environmental indices available within, or near, areas 2B and 3A were used to investigate pos- sible correlations with petrale sole YCS (Table 2). Ocean transport calculations provided by The Pacific Fisheries Environmental Group (PFEG) were based on Bakun (1973; after Fofonoff5). Sverdrup trans- port was calculated by PFEG by using a finite differ- ence form of equation six in Nelson ( 1977 16. A proxy for salinity was based on observations of water den- sity at constant temperature available at the Colum- bia River estuary (Table 2, Fig. 1). The timing of the spring transition was obtained from Strub and James 5 Fofonoff, N. R 1960. Transport computations for the North Pacific Ocean — 1958. Fish. Res. Board Can. Manuscr. Rep. Oceanogr. and Limnol. No. 80. 6 Mv = k(V x i)/ p where: Mv is the meridional component of the vertically integrated mass transport, Jt(Vxr) is the vertical component of the wind stress curl, and (} is the meridional derivative of the Coriolis parameter. Table 2 Environmental indices used in correlations with year-class strength of petrale sole, Eopsetta jordani. Recruitment areas include Pacific States Marine Fisheries Commission areas 2B and 3A. (Source of data is indicated for each environmental factor. ) Environmental index Recruitment area Sea surface atmospheric pressure' ' 2B, 3A Alongshore coastal transport indices Mean sea level2 Neah Bay (48°22'N-124"38' IV) Crescent City (41'45'N-124"12' W) Northward Ekman transport'1 3 Northward Sverdrup transport'' •' Offshore Ekman transport' 3 Cube of wind speed'' ,f Water properties Sea surface temperature' (43"N-44.9°N), (124"W-124.9'W) (46"N-47.9°N), (124'W-124.9'W) Salinity index Columbia River- Estuary (46"13'N-123°45'W) 2B 3A ' Computed at 45" N-125' W . - Tidal Datum Quality Assurance Section. NOAA, Rockville. Mil 20852. ' Pacific Fisheries Environmental Group. P.O. Box 831. Monterey. CA 93942. 3A 2B 2B 3A 2B 3A 2B 3A 2B 3A 2B 3A 484 Fishery Bulletin 92(3). 1994 (1988). However, for years not available from Strub and James, it was estimated from weekly upwelling indices (at 45°N 125°W, Bakun, 1975; Mason and Bakun, 1986). In the latter case the time of spring transition was assigned to the first week of the year in which the weekly upwelling index became posi- tive and remained positive for at least another week. The spring transition dates reported by Strub and James (1988) were highly correlated with our esti- mates (r=0.80; P<0.01). Because March is usually a month of predominant onshore Ekman transport prior to the spring transition (i.e. negative offshore Ekman transport), we determined whether YCS variation was correlated with mean onshore Ekman transport during March. referred to as original series), unrelated trends in original series often cause spurious correlations and conceal the extent of year-to-year associations (Dickey et al., 1986; Norton7; Cohen et al., 1991). Therefore, first-order differencing was used to evalu- ate the reliability of correlations based on the origi- nal series (Chatfield, 1989; here after referred to as filtered series). By this criterion, a significant corre- lation for original series was deemed reliable only if the attendant correlation for the filtered series had the same correlation sign and a minimum absolute value (I r\ > 0.10). Although this procedure does not account for P-values of filtered series, it provides a more consistent selection of factors potentially asso- ciated with recruitment variation. Analytical methods Spearman's correlation analyses (Tate and Clelland, 1957) were used to account for linear and nonlinear monotonic associations between YCS and indepen- dent variables (i.e. potential egg production, envi- ronmental factors, and timing of the spring transi- tion). Independent variables were lagged to the first year of life of each cohort to determine potential in- fluences on YCS at the time year classes were born. Exploratory Spearman's correlation analyses were used for each season of the year because of the sea- sonality of different early life stages and environ- mental factors (Fig. 2, Appendix A). Because many environmental-YCS relationships showed anomalous correlations during the 1958 El Nino, one of the largest El Nino events in the twen- tieth century (Cannon et al., 1985), this year was not included in Spearman's correlation analyses. However, after the most consistent Spearman's cor- relations for 1959-77 were established, all years from 1958 to 1977 were considered in regression analy- ses. Such regressions consisted of estimated YCS on environmental anomalies. These anomalies were computed as the actual seasonal value of a given fac- tor minus its long-term mean for 1958-77. The use of anomalies as independent variables was justified to reduce multicollinearity effects in polynomial re- gressions (Neter et al., 1989). The Bonferroni correction and the P-value plot of Schweder and Spjotvoll (1982) required individual P-values =0.001 for an overall P-value = 0.05 in mul- tiple comparisons between YCS and environmental factors. Since such P-values would have made it dif- ficult to detect meaningful ecological relationships, significance of correlations was based on individual P-values <0.05. Although statistical significance in our study was based on nonfiltered data (here after Results Potential spawning biomass The effect of spawning biomass on subsequent re- cruitment strength was minimal as indicated by simi- lar variations from 1970 to 1977 in YCS per paren- tal egg and the YCS index (Fig. 3). The initial de- cline in YCS in the 1970s was not linked to a de- ' Norton, J. 1990. Relationship between California Current temperatures and intensity of the Aleutian Low. Southwest Fisheries Science Center. Report of activities. March-April 1990. La Jolla, CA, p. 16-18. 600' AREA 2B 120 500' ~ > » YCS INDEX 100 go UJ ™ sv — - YCS PER * -^ \ PARENTAL EGG 400 K 80 d 1 > * I < » \- ■* EZ o O 300 \ 60 UJ 5 z _ \« X => UJ I «. • < s IT <2 200 *• ^fc 40 0. i 1- "• " " ^" ^ P o CO 100 2 1 e : _l rf 1200 140 fj < EAR-C THOUS, o AREA 3A \ - VCS INDEX 5 * 120 W hj \ •-- YCS PER >. v PARENTAL EGG 100 c/5 °; 800 600 8 LASS EMALE: 60 y «■ 400 40 2| 200 68 70 72 74 76 78 YEAR-CLASS (1970-1977) Figure 3 Comparison between year-class strength indices of petrale sole, Eopsetta jordani, with the attendant number of fish reaching age 6 per parental egg in Pacific States Marine Fisheries Commission areas 2B and 3 A. (Cohorts considered in year-class strength indices were hatched between 1970 and 1977.) Castillo et al.: Recruitment variation in Eopsetta jordani 485 crease in parental-stock size, as negative trends be- tween YCS and potential egg production were ob- served for year classes from 1970 to 1977 (Area 2B: r=-0.81, P<0.05; Area 3A: r=-0.57, P<0.20; Fig. 4A). Examination of the attendant filtered series did not support a density-dependent relationship when the size of the parental stock was large (e.g. Ricker, 1954; Area 2B: r=0.21, Area 3A: r=0.07; Fig. 4B). Environmental-YCS fluctuations (1959-1977) Alongshore transport indices 1 Mean sea level height The strongest northward and southward coastal flow generally coincide with the highest and lowest mean sea levels off Oregon (Huyer et al., 1975). Long-term seasonal sea levels indicated that stronger northward and southward flows from 1959 to 1977 occurred in winter and sum- mer, respectively (Appendix A). Correlations between YCS and mean sea level were highest in winter for both study areas (Fig. 5A). The attendant filtered series showed lower but consistent correlations for win- ter ( Fig. 6A). Thus, recruitment strength of petrale sole seems to be associated with interannual variation in nearshore northward transport during winter. A B N. AREA 2B '0 V. 7 t 100 AREA 2B a 77-76 400 N. ■ 72 50 ENGTH INDEX OF RSH) > o o \. 7« N. • T 7 7 S • N. 7 6. >. 5S STRENGTH o r ■ 71-70 ^^^ ^^« 73-71 ^s^ m 72-71 ^^■75- 7* .74-7, 0 P Q 4 5 6 3 "uo .5 0.0 0.5 1 5 8 O X X t 600 AREA 3A ■ 1 0 CC < LU Q 100 LU AREA 3A ■ 77-76 2 > 400 7 S . . 7 ^"V. 78 • UJ d o u_ •100 ■ 7S 74 .7.-73 ■"■'• ■ 72-71 ""5 6 7 8 9 10 •'U"'1 0 12 3 POTENTIAL EGG PRODUCTION ( X 10 ') FILTERED POTENTIAL EGG PRODUCTION Figure 4 Variation of year-class strength of petrale sole, Eopsetta jordani, in relation to potential parental egg production in Pacific States Marine Fisheries Commis- sion areas 2B and 3A. Comparison is shown for (A) original and (B) filtered series. Cohorts hatched between 1970 and 1977 (years are identified by the last two digits). 2 Northward Ekman transport This index indi- cates the alongshore flow of surface mixed layers driven by wind stress. Unlike sea level height, there was no correlation between northward Ekman trans- port and YCS (Figs. 5B and 6B). The four long-term seasonal means of this index indicated predomi- nantly negative northward transport (i.e. southward transport) of surface waters at offshore areas (45°N- 125°W, Appendix A). 3 Northward Sverdrup transport This index mea- sures alongshore transport over the entire water col- umn by adding geostrophic flow to Ekman transport. For nonwinter seasons, the long-term seasonal means of this index showed southward transport of offshore waters (45°N-125°W, Appendix A). Northward Sverdrup transport and YCS were not consistently correlated (Figs. 5C and 6C). Offshore Ekman transport The long-term seasonal means of this index indicated average offshore trans- port of surface waters and upwelling from spring to summer followed by onshore transport and down- welling from fall to winter (Appendix A). Winter off- shore Ekman transport and YCS showed clear and consistent negative correlations in both study areas ( Figs. 5D and 6D ). Cube of wind speed This in- dex reflects the turbulence transferred to the sea surface by the wind (Niiler and Kraus, 1977). Although cube of wind speed was correlated with YCS for winter and spring (Figs. 5E and 6E ), these correlations were largely explained by onshore Ekman transport during this period. Moreover, correlations for spring vanished in both ar- eas when March was excluded from the analyses. Thus, cube of wind speed seems to be spu- riously correlated with YCS. Sea surface temperature Long- term seasonal means of sea sur- face temperature increased from winter through summer and decreased from fall through winter (Appendix A). Correla- tions between YCS and sea sur- face temperatures for winter and spring showed high positive 486 Fishery Bulletin 92(3). 1994 values only in Area 3A, particularly during winter (Figs. 5F and 6F). Therefore, warmer temperatures may result in increased survival of eggs and/or lar- vae of petrale sole in Area 3A. bia River plume. Although high correlations between filtered YCS and the salinity index were detected for winter and summer, no significant correlations were observed for original series (Figs. 5G and 6G). Nearshore salinity index This index reflects nearshore salinity variations caused by the Colum- O Ui BC c 0 o cr UJ Ll LL Q ZZ < IX CO < cr < LJJ a Original Series MEAN SEA LEV CI CUBE OF WMD SPEED WINTER SPRINGSUMMER FALL WINTER SPRINGSUMMER FALL 4 3 -0-4 4 6 NORTHWARD EUlh TRANSPORT SEA SURFACE TEMPERATURE 06 bl:--_ WINTER SPRINGSUMMER FALL NORTHWARD SVERDRUP TRANSPORT 06 -02 -C4 -0 6 i-r- Kl 't. 2 v, z WINTER SPRINGSUMMER FALL NEWSHORE SALJNrTT IWTXI □ « 4 : -C * -C 6 ^^ WINTER SPRING SWIM EJl FALL WINTER SPRINGSUMMER FALL OFFSHORE EXMAN TRANSPORT SURFACE ATMOSPHERIC PRESSURE _«*M "■fe WINTER SPRINGSUMMER FALL 0 6 0» 02 00 ■OS -04 %:^ &2 WINTER SPRING SUMMER FALL SEASON Figure 5 Spearman's correlations between year-class strength indices of petrale sole, Eopsetta jordani, and seasonal averages of environmental factors lagged to the first year of life (Table 2). Correlation for cohorts 1959 to 1977 are compared for Pacific States Marine Fish- en rs Cnm mission areas 2B (dark bar) and 3A (light ban. Dashed lines correspond to P = 0.05 for indi- vidual correlations. Sea surface atmospheric pressure Nearshore trans- port and water properties such as temperature, sa- linity and density are greatly influenced by the North z o h- < cr EC O U u UJ EC < cr 03 zz < a < 0. CO Filtered Series UEAN SEA LEVEL A 0.4 C.2 B^ v? *2 ^ & U «. CU8E OF WIND SPEED WINTER SPRINGSUNNER FALL NORTHWARD if.MUi TRANSPORT 00 -o : ■04 •0.6 1 w WINTERSPRJNCSUHNER FALL SEA SURFACE TEMPERATURE B WINTER SPRINGSUNNER FALL NORTHWARD SVERDRUP TRANSPORT WINTER SPRINGSUNNER FALL NEARSHORE SALINITY INOEI -o : -0 4 -0.6 ■M* WINTER SPRINCSUHHER FALL OFFSHORE EKHAK TRANSPORT WINTER SPR1N0SUNNER FALL SURFACE ATVOSPHEJUC PRESSURE 5 ■oa -o * I Tod2^ -"feH WINTER SPRINGSUNNER FALL 0 0 4X5 -04 4 6 TT E^-^ WINTER SPRINGSUNNER FALL SEASON Figure 6 Spearman's correlations between filtered year-class strength indices of petrale sole, Eopsetta jordani, and filtered seasonal averages of environmental factors lagged to the first year of life for cohorts from 1959 to 1977 (Table 2). Areas compared are Pacific States Marine Fisheries Commission 2B (dark bar) and 3A (light bar). Significance of correlations was based on original series (see analytical methods). Castillo et al .: Recruitment variation in Eopsetta jordani 487 Pacific high and Aleutian low pressure systems (Huyer, 1983). Large negative correlations between YCS and winter sea surface atmospheric pressure were evident in both areas (Fig. 5H). However, fil- tered series suggested that this index was spuriously correlated with recruitment strength (Fig. 6H). Effects of the spring transition on YCS Correlations between YCS and the week of the spring transition from 1967 to 1977 were not significant (Area 2B: r=0.27, Area 3A: r=0.22; P>0.20; Fig. 7A). Moreover, correlations for attendant filtered series were nega- tive (Area 2B: r=-0.18, Area 3A: r=-0.32; Fig. 7B). However, onshore transport during early spring can affect YCS as recruitment strength was correlated with mean onshore Ekman transport during March in both study areas (Area 2B: r=0.55, Area 3A: r=0.50; P<0.05). Such correlations were also supported by the attendant filtered series (Area 2B: ;-=0.58, Area 3B:r=0.42). Environmental- YCS series Based on exploratory cor- relation analyses for original and filtered series, win- I? 2 o z _ Ei | E § O is su So1 u bOO AREA 2B 400 70 a 68° 71° , 73 72=^^ t^ 69° 300 74 o 77o 200 mn o76 751 AREA 3A □ 68 67 = 70D 750 77 ft 73 74 o 71°, °72 69 o76 o z CO < -J u cr < 8 10 12 14 16 18 SPRING TRANSITION (JULIAN-WEEK) Figure 7 Variation of year-class strength for petrale sole, Eopsetta jordani, in Pacific States Marine Fisheries Commission areas 2B and 3A in relation to the week of the year in which the spring transition occurred. Comparison is shown for (A) original and (B) filtered series. Year classes are considered hatched be- tween 1967 and 1977 and are identified by the last two digits. ter offshore Ekman transport seemed to be the main factor affecting YCS of petrale sole in areas 2B and 3A. In subsequent analyses, the period January- March was chosen for describing the association be- tween YCS and offshore/onshore Ekman transport. This period was selected because of the importance of onshore transport on YCS during March. More- over, correlations between offshore Ekman transport and YCS tended to be higher during January-March (original and filtered series respectively: Area 2B, r=- 0.48 and -0.67; Area 3A: r=-0.52 and -0.65; P<0.05) than December-February (original and filtered se- ries respectively: Area 2B, r=-0.46 and -0.53; Area 3A: r=-0.52 and -0.42; P<0.05). Comparing onshore Ekman transport with YCS, we found that two periods of reduced onshore Ek- man transport ( 1962-65 and 1974-77) coincided with weak year classes of petrale sole (Fig. 8). The years 1958, 1961, and 1968 showed the largest positive anomalies in onshore Ekman transport from 1958 to 1977. However, unlike 1961 and 1968, the 1958 El Nino produced weak YCS in Area 2B and near-aver- age YCS in Area 3A. Other anomalies for indices such as winter and spring sea surface temperature and win- ter sea level height showed some correspondence with YCS anomalies in Area 3A. In Area 2B, only onshore Ekman trans- port and sea level height sug- gested some association with YCS (Figs. 8 and 9). Percentage of YCS variation ex- plained by environmental fac- tors The relationship between YCS and January-March off- shore Ekman transport anoma- lies was best described by sec- ond-order polynomial regres- sions (Fig. 10, Table 3). Associa- tions between YCS and winter sea level height anomalies in areas 2B and 3A, were also de- scribed by second-order poly- nomials (Fig. 11, Table 3). Al- though the association between YCS and winter sea surface tem- perature in Area 3A was better described by a second-order polynomial than by a linear re- gression, only the latter was sig- nificant (Fig. 12, Table 3). Ex- cept for the year 1958, and for the years with large anomalies in offshore Ekman transport AREA 2B 7069 68-67 77-76 73-72 " 72-71 ■ 76-75 " • 71-70 75-74 • . 74-73 ■ 69-68 200 AREA 3A 100 77-76 ™-°9 68-67 ~^^^^^ 75-74 0 ' — ^" 76-75- • 73-72~~~^^^ 100 72-71' 74. 75 71-70 69-68 . 8-6-4-2 0 2 4 6 FILTERED SPRING TRANSITION 488 Fishery Bulletin 92(3). 1994 u; Si 600 500 400 300 200 100 AREA 2B R'= 0.30 P < 0.05 1200 AREA3A 1000 R' = 0.38 P < 0.05 1966 800 600 1961 ' 1958 1968 ^X/ 400 :.\. 150 100 - 50 50 OFFSHORE EKMAN TRANSPORT ANOMALY (METRIC TONS/SECOND PER 100-m COAST) Figure 10 Relations between year-class strength of petrale sole, Eopsetta jordani , in Pacific States Marine Fisheries Commission areas 2B and 3A and mean offshore Ekman transport anomaly off Oregon from January to March. (Regression parameters are shown in Table 3. 1 000- — . — i — . — i — , — . — ^_ _| , , , 1 , , , AREA 3A • 1966 R2 = 0.23 800- P < 005 1 9 6 1^. 600- • • V^-'l 970 1958 400- 200- "" o 2 £ LU ts> X u- K u. !/) O - 2 - 1 0 1 2 SEA SURFACE TEMPERATURE ANOMALY ( °C) Figure 12 Relation between year-class strength of petrale sole, Eopsetta jordani , in Pacific States Marine Fisheries Commission Area 3A and mean win- ter sea surface temperature anomaly off Oregon- Washington (December-February). (Regression parameters are shown in Table 3.) transport during winter (Ketchen and Forrester, 1966). The analyses in Area 3A are also consistent with an association between temperature and sur- 2 600 500 400 AREA 2B R' = 0.386 gi o Z u. uj o DC tn Wo 2§ H 200 100 1200 1000 2 800 600 400 200 AREA 3A = 0.399 0.05 1969 1958 ■0.2 -0.1 0.0 0.1 SEA LEVEL ANOMALY (m) 0.2 Figure 1 1 Relations between year-class strength of petrale sole, Eopsetta jordani, in Pacific States Marine Fisheries Commission areas 2B and 3A and win- ter mean sea level height anomalies. (Regression parameters are shown in Table 3. ) vival of early life stages of petrale sole (Ketchen and Forrester, 1966; Alderdice and Forrester, 1971 ). Con- sidering the discharge of the Columbia River into Area 3A and the stenohaline condition of eggs and larvae in this species, possible salinity- YCS associa- tions may be overridden by cross-shore and along- shore advection and sea temperature. Recruitment strength of petrale sole was correlated between areas 2B and 3A(r=0.82, P<0.01). However, the highest determination coefficient for regression models ofYCS on environmental factors was obtained in Area 3A. Although the proportion of variation in YCS explained by second-order polynomial regres- sions was significant for both offshore Ekman trans- port and sea level height, filtered series suggest that year-to-year variation in YCS were better explained by the former. Unlike Area 3A and off British Co- lumbia (Ketchen and Forrester, 1966), no regression models of YCS on sea surface temperature were sig- nificant or marginally significant in Area 2B. The higher positive temperature-YCS association for Area 3A compared with Area 2B is consistent with tem- perature differences between areas (Appendix A). These observations suggest a latitudinal effect of tem- perature on the recruitment of petrale sole, that is, higher recruitment toward the poleward range of the 490 Fishery Bulletin 92(3), 1994 Table 3 Regressions of year-class strength (YCS) for petrale sole, Eopsetta jordani (thousands of females attaining age 6), on environmental anomalies in Pacific States Marine Fisheries Commission areas 2B and 3A from 1958 to 1977. Environ- mental anomalies are based on the long-term mean 1958-77. SST = sea surface temperature anomaly (Dec.-Feb.), SL = mean sea level height anomaly (Dec.-Feb. ), Me = offshore Ekman transport anomaly ( Jan.-Mar. ). P<0.05 for significant corre- lations (n = 20 years I. In the last regression model, R2 = 0.667 between predicted YCS (eln10%, an additional count was performed and the value that differed by >10% was discarded. The mean of the three remaining counts was recorded. Dates of first ring deposition for juvenile bluefish collected from all sources were backcalculated from the number of growth increments counted. Estimates of swimming speed and swimming distance The distance juvenile bluefish were capable of swim- ming per day was calculated from estimates of the average body length during migration and the rela- tionship between swimming speed and average body length. Average body length was equal to one-half the mean size of fish collected during a 2-week pe- riod immediately following first appearance at a sam- pling site minus 2 mm, the size at hatching (Deuel et al., 1966). The swimming speed for fish such as salmon, cod, and herring is about three times the body length/sec for fish measuring 10-100 mm (Harden Jones, 1968). Juvenile bluefish measuring 155-213 mm TL swam at speeds averaging 94 mm/ sec at 20°C and 180-260 mm/sec when cooler water was added (Olla et al., 1985). We estimated swim- ming speed of juvenile bluefish at no more than 2 body length/sec. The mean size of juvenile bluefish collected from the Marsh River during a 2-week period immediately following their first appearance in 1990 and 1991 was 108 mm FL (n = 17) and 101 mm FL (/i=19). Total growth between hatching and appearance at the col- lection site in 1990 was 106 mm ( 108 mm minus 2 mm [the size at hatching]) and in 1991, 99 mm. The average size of the fish during this migration would be one-half these values or 53 and 50 mm. Swim- ming distances per day during migrations in 1990 and 1991 were 53 mm x 2 (twice the body length) = 106 mm/sec (9.16 km/day) and 50 mm x 2 = 100 mm/ sec (8.64 km/day). Similarly, the mean size of juve- nile bluefish collected from Sagadahoc Bay during August 1990 was 48 mm FL (n=29). The swimming distance per day was therefore 23 mm x 2 = 46 mm/ sec (3.97 km/day). The shortest distances between the northern por- tions of the spawning grounds in the South Atlantic Bight to the collection site in the Marsh River and the Middle Atlantic Bight to the collection site in Sagadahoc Bay were estimated at 1350 km and 425 km respectively. Statistical analysis The relationship between fork length and the num- ber of daily rings for juvenile bluefish captured in Maine during 1990-91 was described by linear re- gressions, whereXis the log10 of fork length and Y is the number of daily rings. Slopes were then compared with ANCOVA. The curvilinear relationship of the same data from Maine (1990-91) and New York (1985-86) was described by quadratic equations, where X is fork length and Y is number of daily rings. The relationship of fork length to growth rate for combined data from Maine was described by a lin- ear regression, where X is the fork length and Y is the rate of growth. Results and discussion Records of juvenile bluefish from the coast of Maine Most records of juvenile bluefish were from south- western Maine (York, Cumberland, Sagadahoc, and Lincoln counties) during the months of July, August, and September (Table 1). Fish varied in size from 37 to 197 mm FL (39-218 mm TL) from both oceanic and estuarine environments. Atlantic puffins (Fratercula arctica ar-ctica ), Arctic terns (Sterna para- dioaea ), and roseate terns (Sterna dougallu dougalli) were reported as feeding on young-of-year bluefish in July near Matinicus Rock and Seal Island (Knox County).3 These observations indicated that juvenile bluefish occur in the offshore waters of Maine prior to their usual appearance inshore during August- September. Clark ( 1973 ) observed that juvenile blue- fish remain where salinities are high when they first arrive, and later, as summer progresses, they pen- etrate into estuaries. Little Kennebec Bay (Table 1, No. 28) is the most northeast location where the cap- ture of juvenile bluefish was confirmed. Three juve- nile bluefish were among 1,000 age-1 Atlantic her- ring, Clupea harengus, tagged at this location on 25 August 1983 (Creaser and Libby, 1986). Young-of-year (YOY) bluefish may be more abun- dant northeast of Boothbay Harbor than suggested by Table 1. The scarcity of information from this area 3 Kress, S. Cornell Univ. Ornithology Lab, Ithaca, NY. Personal commun, January 1991. 498 Fishery Bulletin 92(3). 1994 Table 1 Unpublished records of juven le bluefish, Poma tomas saltatrix, in Maine. Size (mm)1 Date of Oceanic (O) or No. Method of Location N. Lat. E. Long. capture estuanne(E) captured TL FL capture1' Wells Harbor (York Co.) 1 Wells Harbor 43" 19" 70*34' Aug 1991 K 1 68 FN Old Orchard Beach (York Co.) 2 1 Mile off Amusement Pier 43"31" 70*21' Summer 1961-1964 0 — Juvenile HL Scarborough Marsh (Cumberland Co.) 3 Confl. of Dunston, Libby, Nonesuch R. 43'33' 70*20' Summer 1987 E - Juvenile HL Portland Harbor (Cumberland Co.) 4 Union Wharf 43*39'15" 70*15' Sept 1984 0 6 150-200 HL 5 SMVTI Dock 43*39' 70*14'15" Sept 1986 0 — 130-150 HL Royal River (Cumberland Co.) 6 Royal River 43*47'30" 70*10' Summer 1988 E — Juvenile — Merepoint (Cumberland Co.) 7 Merepoint Bay 43'50' 70*00'01" 26 Sept 1991 E 97 150-174 GN Jenny Island, Casco Bay (Cumberland Co.) 8 Jenny Island 43"46' 69*55 16 July 1991 E 1 40 CT New Meadows River (Cumb.-Sagadahoc Co.) 9 Howard Pt, Thomas Bay 43'53'15" 69°53'10" Aug 1988 E 3 70-130 FK Kennebec River (Sagadahoc Co.) 10 Atkins Bay 43'45' 69*48' Summer 1981 E - 80-90 HS 11 Winnegance Bay 43*53' 69"49' Summer 1988-1990 E — 50-150 HL 12 Bath Bridge 43*55' 69"48'30" Summer 1982 E 90 <100 OT 13 Mouth of Androscoggin R. 43"57' 69*53' 5 Aug 1983 E 2 82-86 HS 14 Mouth of Abagadasset R. 44W30" 6951' 17 July 1986 E 5 70-77 HS Mouth of Abagadasset R. 11 Sept 1987 E 2 142-150 HS M mth nl Abagadasset K 3 Aug 1989 E 8 52-76 HS Mouth of Abagadasset R. 3 July 1991 E 6 112-115 HS Mouth of Abagadasset R 18 July 1991 E 2 84-94 HS Sagadahoc Bay (Sagadahoc Co.) 15 Kennebec Pt. 43*45'20" 6945'45" 10-22 Aug 1990 O 29 39-70 37-64 HS Montsweag Bay (Lincoln Co.) 16 Berry Island 43,58'15" 69'40'55" 30 Aug 1972 1. 1 112 HS Berry Island 29 Aug 1973 E 2 132-141 HS Berry Island 8 Sept 1974 E 4 125-140 HS Sheepscot River (Lincoln Co.) 17 Cross River 43 55' 6937' 8 Aug 1991 E 1 115 HS 18 The Eddy 43-59'30" 6939' 9 July 1991 E 3 80-85 HS 19 Marsh River 44*01' 69*35'45" 14 Aug 1986 1. 28 93-121 GN Marsh River 26 Aug 1987 E 6 129-163 GN Marsh River 8-28 Aug 1989 E 102 92-194 GN Marsh River 1 Aug-26 Sept 1990 E 149 89-218 81-197 GN Marsh River 17 July-17 Sept 199 E 60 101-217 193-196 GN 20 Sheepscot Falls 44*02'45" 69*37' Aug 1967 E — 150-200 HL 21 Sheepscot River itown of Sheepscot) 44*03'30" 69 3630' 2 Aug 1989 E 1 140 HL Boothbay Harbor (Lincoln Co.) 22 Lobster Cove 4351' 69 37 30 Aug 1990 0 1 145 131 HL Lobster Cove 11 Aug 1991 O 4 162-192 HL 23 Townsend Gut 13 ."hi 69*39" 5 Sept 1985 0 1 Juvenile HL 24 DMR Dock 43 51145'' 69*3830' 14 Sept 1971 0 5 95-105 89-97 - DMR Dock 25 Aug 1978 O ! 86 DN DMR Dock 4 July 1984 0 3 40-50 HL 25 Foot Bridge 4.3 51' 69 374D Summer 1970-1974 O - •Juvenile.- 3 sizes HS Matinicus (Knox Co.) 26 Matinicus Rock 43 47 68'51'30 18 July 1986 0 1 77 AP Matinicus Rock 5 July 1989 0 2 85-90 AP Matinicus Rock Mid July 1990 0 2 30-40 AT Matinicus Rock 9-17 July 1991 0 14 40-50 AT Creaser and Perkins: Distribution, food, and abundance of Pomatomus saltatnx 499 Table 1 (Continued) Size (mm)' Location N. Lat. E. Long. Date of capture Oceanic (0 lor estuanne 1E1 No. captured TL Method of FL capture- Matinicus Rock 24-30 July 1991 0 Seal Island (Knox Co.) 27 Seal Is. 43"53' 68"45' July 1991 0 Little Kennebec Bay (Washington Co.) 28MarstonPt. 44'39' 67'26'15" 25 Aug 1983 0 50-60 50 100-130 RT AT HW 1 TL = total length; FL = fork length. ; OT = Otter trawl; FN = Fyke net ; HL = Hook and line; - = Unknown; HS = Haul seine , AP = Atlantic puffin, GN = Gill net; AT = Artie tern, DN = Dip net; CT = Common tern , HW = Herring weir; RT = Roseate tern. may result from the lack of scientific fish sampling activity and lack of familiarity with the identifica- tion of juvenile bluefish by the sportfishing public and finfish processors. Clark (1973) attributed high densities of juvenile bluefish in areas such as South Carolina and the New York Bight to sampling activ- ity and availability of records from those areas. Some published information exists on the distribution of juvenile bluefish in southwestern Maine. They have been reported from Casco Bay (Bigelow and Schroe- der, 1953; Wilk, 1977), Boothbay Harbor (Lund, 1961) and Montsweag Bay (Targett and McCleave, 1974). Length-frequency distributions Length-frequency distributions of juvenile bluefish captured in the Marsh River (1990-91), Sagadahoc Bay ( 1990 ), and Merepoint Bay ( 199 1 ) varied between 8 and 20 cm, 3 and 6 cm, and between 14 and 17 cm respectively (Fig. 2, A-D). Length-frequency data from the Marsh River was combined in 1-cm group- ings and presented as one figure per year because fish were collected for only two months. Although the size range recorded during both 1990 and 1991 was similar, larger juveniles were more prevalent dur- ing 1991 (Fig. 2, A and B). The length-frequency com- position of fish captured by beach seine from Sagadahoc Bay on 10 and 22 August 1990 (Fig. 2C) confirms the presence of small juveniles of 3-6 cm FL in Maine waters. Previous reports of juvenile bluefish less than 70 mm from Wells Harbor, Winnegance Bay, the mouth of the Abagadasset River, Boothbay Harbor, Matinicus Rock and Seal Island have been presented in Table 1. The overall length-frequency range (8-20 cm FL) reported for fish captured in the Marsh River during 1 August-26 September 1990 and 17 July- 17 September 1991 was similar to the length-fre- quency range reported from Great South Bay, New York (8-22 cm FL) during 28 May-12 August 1985 and 10 June-20 August 1986 (Nyman and Conover, 1988). Length-frequency distributions generated from gillnet catches (Fig. 2, A, B, D) are biased be- cause panels of individual mesh sizes in gill nets are selective for specific sizes offish. Temperature and salinity data collected during gillnet sets Juvenile bluefish were captured on both incoming (I) and outgoing (O) tides between 1 August and 26 September 1990 and between 17 July and 17 Sep- tember 1991. Specific water temperatures do not appear to be associated with the arrival and depar- ture of juvenile bluefish. Bluefish were first captured at water temperatures of 24.6-27.0°C (1990) and 22.8-24.0°C (1991). None were captured when the temperature dropped below 16.0°C ( 1990) and 19.3°C ( 1991 ). Water temperatures varied between 16.0 and 27.0°C (1990) and between 19.3 and 27.4 C (1991) when bluefish were present. Lund and Maltezos ( 1970) reported that juveniles and adults appear at temperatures of 12-15°C and depart at temperatures of 13-15°C. Oben (1957) re- ported that juveniles appear at temperatures of 18- 24.5°C and depart at 13-15°C. Similarly, Nyman and Conover ( 1988) reported that juveniles appear at 20- 24°C and depart at water temperatures in the "middle-teens." We believe that juvenile bluefish ar- rived at the mouth of the Sheepscot River when the water temperature was relatively low and they were attracted to the Marsh River, 34.3 km upstream, by the warmer water. These fish appeared to remain in the Marsh River until the water temperature dropped to a temperature range that was lower than the wa- ter temperature at the time of their arrival in the sampling area. The salinity of the Marsh River varied between 9.5 and 23.3 ppt (1990) and between 9.6 and 26.5 ppt (1991) when juvenile bluefish were present. 500 Fishery Bulletin 92(3), 1994 Marsh River 1990 N = 149 Mar; N = h River 1991 60 10 11 12 13 14 15 18 Fork Length, cm 12 13 14 15 16 Fork Length, cm Sagadahoc Bay 1990 N = 29 Merepoint Bay 1991 N = 70 8 9 10 11 12 13 14 15 Fork Length, cm 10 11 12 13 14 15 16 17 Fork Length, cm Figure 2 Length-frequency histograms for juvenile bluefish, Pomatomas saltatrix, collected in Maine: (A) Marsh River 1 Aug.-26 Sept. 1990; (B) Marsh River 17 July-17 Sept. 1991; (C) Sagadahoc Bay 10-22 Aug. 1990; and (D) Merepoint Bay 26 Sept. 1991. Smale and Kok (1983) collected juveniles at 12-35 ppt, Clark ( 1973) occasionally found bluefish in very brackish water, and Baird (1873) captured juveniles from the freshwater portion of tidal rivers. Stomach contents Stomach contents of juvenile bluefish (81-200 mm FL) collected from the Marsh River in 1990 and 1991 are presented in Table 2. During 1990, fish were present in 81.4% of stomachs containing food (F), crustaceans in 39.5% (F), and plant material in 4.7% (F). During 1991, crustaceans were slightly more prevalent than fish. Crustaceans were present in 69.0% of stomachs containing food (F), fish in 62.1% (F), and plant material in 5.2% (F). The frequency of occurrence of polychaetes was 3.4% (F) and of insects, 1.7% (Fi. During 1990 and 1991, crustaceans were the pre- dominant prey items in terms of numbers of indi- viduals observed (N) expressed as a percent of the total number of prey items recorded. Crustaceans composed 58.0% ( 1990 ) and 83.4% ( 1991 ) of the total prey items, fish composed 39.9% (1990) and 11.3% (1991), and plant material 2.0% (1990) and 4.5% ( 1991 ). During 1991, 0.5% and 0.3% of the total num- ber of prey items were polychaetes and insects. During 1990 and 1991, fish were the predominant prey in terms of wet or dry weight expressed as a percentage of the total wet weight ( W) or dry weight (Wl) of all prey species recorded. Fish composed 77.4% of total wet weight (W) and 79.9% of total dry- weight (Wl) during 1990 and 61.9% (W) and 67.5%. (Wl) during 1991. Crustaceans composed 22.5% ( W) and 20.0% (Wl) during 1990 and 37.3% (W) and Creaser and Perkins: Distribution, food, and abundance of Pomatomus saltatrix 501 Table 2 The stomach contents of 149 juvenile bluefish, °omatomas saltatrix measuring 81- 197 mm FL ( v = 117 mm), and 60 juveniles measuring 93-200 mm FL ( 1 = 142 mm), collected from the Marsh River, Maine, during 1 August-26 September 1990 an d 17 Julv- -17 September 1991, respectively. Taxon Marsh River 1990 Marsh River 1991 FCJI' Fl(%)2 N(%)3 W(%)4 Wl C7f I5 F(%) Fl(%) N (*)(%) W(%) wirai Class Crustacea Argulus sp. 1.550 1.342 1.024 0.021 0.027 Corophium sp. 1.724 1.667 0.263 (0.6061 0.011 0.009 Crangon septemspmosa 34.884 30.201 55.631 22.351 19.888 63.793 61.667 26.579(61.212) 37.294 31.830 Rhithropanopus harrisii 0.775 0.671 0.341 0.061 0.064 Unidentified Crustacea 1.550 1.342 0.683 0.063 0.031 Unidentified Isopoda 0.775 0.671 0.341 0.029 0.037 Zoea of Majidae sp. 3.448 3.333 56.579 0.006 0.005 Total 39.535 34.228 58.020 22.525 20.046 (i.s 'i,i.; 66.667 83.421(61.8181 37.310 31.844 Class Osteichthyes Alosa pseudoharengus 6.202 5.369 4.096 13.751 13.953 Clupeidae spp. 5.426 4.698 2.730 7.392 8.045 15.517 15.000 3.158(7.273) 27.131 29.792 Fundulus heleroclitus 0.775 0.671 0.683 7.309 9.239 1.724 1.667 0.526(1.2121 6.720 7.366 Memdia menidia 9.302 3.054 5.119 13.229 13.798 8.621 8.333 1.579(3.6361 2.471 2.228 Fish remains 59.690 51.678 27.304 37.739 34.861 36.207 35.000 6.053(13.939) 25.589 28.146 Total 81.395 70.470 39.932 77.421 79.895 62.069 60.000 11.316(26.0611 61.911 67.532 Class Polychaeta Polychaete remains 3.448 3.333 0.526 (1.2121 0.445 0.342 Total 3.448 3.333 0.526 (1.2121 0.445 0.342 Class Insecta Hymanoptera (unident. I 1.724 1.667 0.263 (0.6061 0.009 0.005 Total 1.724 1.667 0.263 (0.6061 0.009 0.005 Plant material Leafy 0.775 0.671 0.341 0.005 0,007 Woody 3.986 3.356 1.706 0.049 0.052 5.172 5.000 4.474(10.303) 0.325 0.278 Total 4.651 4.027 2.048 0.054 0.059 5.172 5.000 4.474 (10.303) 0.325 0.278 Number of stomachs examined 149 60 Number of stomachs containing food ("t containing food 129(86.58) 58(96.671 Number of empty stomachs ($ empty 1 20(13.421 2 (3.33) Mean wet weight/stomach (all stomachs gm 0.651 1.133 Mean dry weight/stomach (all stomachs -'111 0.116 0.183 Mean wet weight/stomach (stomachs with foodl gm 0.752 1.172 Mean dry weight7stomach (stomachs with foodl gm 0.134 0.189 1 F = % frequency of occurrence — the number of si omachs n which a species ( taxon 1 occurred ex Dressed as a ^ of the total number of si omachs containing food. 2 Fl = % frequency of occurrence — the number of s tomachs in which a species (taxon) occurred expressed as a 9t of the total nu mber of stomachs examined. 3 N = % numerical abun iance — the number of ind ividuals of each species (taxon) expressed as a % of the total number of food items i individual prey in stomachs! (*) d esignates th ? % numerical abundance excluding 2 5 zoea of Majidae foun d in the stomachs of two fish. 4 W = *7r weight — wet weight of a species (taxon, prey category! expressed as a % of th e total wet weight o f food items. 5 Wl = ty weight — dry weight of a species (taxon. arey category) expressed as a % of the total dr> weight if food items 31.8% (Wl) during 1991. Plant material accounted Stomach contents of juvenile bluefish collected for 0.1% (W) and 0.1% (Wl ) during 1990 and 0.3% from other areas in Maine during 1990 and 1991 are (W) and 0.3% (Wl) during 1991. Polychaetes made presented in Table 3. Bluefish measuring 150-173 up 0.4% (W) and 0.3% (Wl) during 1991; insects made mm FL were captured from Merepoint Bay on 9 Sep- up 0.01% (W) and 0.01% (Wl). tember 1991. Fish were the predominant prey item 502 Fishery Bulletin 92(3). 1994 for all indices: 80.0% (F), 67.7% (N), 85.0% (W), and 88.6% (Wl). Crustaceans composed 24.0% (F), 32.3% (N), 15.0% (W), and 11.4% (Wl). Bluefish measuring 37-64 mm FL were captured from Sagadahoc Bay on 10 and 22 August 1990. The predominant prey items recorded from these small juveniles were crustaceans (mysids and copepods). The frequency of occurrence of mysids and copepods in stomachs containing food (F) was 10.3% and 96.6%, respectively. Crustaceans were also the predominant prey items in terms of their wet or dry weight, ex- pressed as a percentage of the total wet weight (W) or dry weight (Wl). Crustaceans composed 82.2% of the total wet weight (W) and 80.0% of the total dry weight ( Wl ). Fish remains were a minor constituent of the stomach contents of these small bluefish; they occurred in 6.9% of stomachs containing food (F). Fish also composed 17.7% of the total wet weight (W) and 21.0% of the total dry weight (Wl). No information was available regarding the numbers of individuals observed (N) expressed as a percent of the total num- ber of prey items recorded because copepod remains were so numerous that it was impractical to count them. The diet of juvenile bluefish collected from the Marsh River in 1990 and 1991 and from Merepoint Bay in 1991 is consistent with observations by Breder (1922), Grant (1962), data for bluefish >10 cm re- ported by Smale and Kok (1983), and the 1981 re- sults recorded in Friedland et al. ( 1988). All studies showed that the major portion of the diet (by weight or volume) of juvenile bluefish >8 cm in length con- sisted offish. Although results presented by Lassiter (1962) and Naughton and Saloman (1984) are also consistent with this observation, Lassiter ( 1962 ) did not report bluefish lengths and some of the bluefish in Naughton and Saloman's (1984) data were older than age 1 (e.g. 39.9 cm). Our findings regarding the diet of juvenile blue- fish collected from Sagadahoc Bay (1990) were con- sistent with results presented by Kendall and Naplin (1981) and Smale and Kok (1983) for bluefish <100 mm. Smale and Kok (1983) reported that juvenile bluefish <100 mm feed predominantly on small crus- taceans, and Kendall and Naplin (1981) stated that most of the diet of juvenile bluefish (.r=4.33 mm) consisted of copepods, copepodites, cladocera, and fish eggs. A transition to pisciverous feeding by juvenile bluefish at a size range between 60 and 100 mm has been reported (Nichols, 1913; Greeley, 1939; Oben, 1957; Clark, 1973; Smale and Kok, 1983). The tran- sition to fish as prey had not yet occurred in bluefish measuring 37-64 mm from Sagadahoc Bay (Table 3). On a per cent weight basis, the dominant prey spe- cies of juvenile bluefish (81-200 mm FL) collected from the Marsh River and Merepoint Bay included the mud shrimp Crangon septemspinosa, juvenile alewives, Alosa pseudoharengus , unidentified clu- peids, Atlantic silversides, Menidia menidia, mum- Table 3 The stomach contents of 34 juvenile bluefish, Pomatomas saltatrix, measuring 150-173 mm FL ( v = 156 mm), col- lected from Merepoint Bay on 9 September 1991, and 29 small juveniles, measuring 37-64 mm FL ( v=48 mm), collected from Sagadahoc Bay, Maine, on 10 and 22 August 1990. Taxon Merepoint Bay 1991 Sagadahoc Bay 1990 F(%) Fl'Vi N(%) Wif;i wicji F(%) Fl{%) N(%) \\ Wl(%) Class Crustacea Crangon septemspinosa Mysidacea spp. Copepod remains 24.000 17.647 32.258 14.973 11.400 10.345 96.552 10.345 96.552 - 13.411 i;s74ii 14.360 64.621 Total 24.000 17.647 32.258 14.973 11.400 106.897 106.897 82.157 79.981 Class Osteichthyes Fish remains Clupeidae spp. 68.000 12.000 50.000 8.824 ;-,KII(i.-, 9.677 40.057 44.970 42.353 46.247 6.897 H.SH7 - 17.742 21.019 Total 80.000 "i.sNlM 67.742 85.027 N.ShUll 6 897 (, MIT 17742 21.019 Number of stomachs examined Number of stomachs containing food I'i containing food Number of empty stomachs i'i empty i Mean wet weight/stomach (all stomachs) gms Mean dry weight/stomach (all stomachs i gms Mean wet weight/stomach (stomachs with food) gms Mean dry weightstomach (stomachs with food) gms 34 25(73.53) 9(26.47) n 18288 0.67367 0.12091 29 29(100) 0(0) 0.03057 0.00596 0.03057 0.00596 Creaser and Perkins: Distribution, food, and abundance of Pomatomus saltathx 503 michogs, Fundulus heteroclitus, and unidentified fish remains (Tables 2 and 3). Other authors reported that a significant portion of the juvenile bluefish diet con- sists of shrimp (Linton, 1905; Wilk, 1977; Friedland et al., 1988), clupeids (Hildebrand and Schroeder, 1928; Greeley, 1939; Grant, 1962; Richards, 1976; Naughton and Saloman, 1984), Atlantic silversides (Hildebrand and Schroeder, 1928; Greeley, 1939; Grant, 1962; Lassiter, 1962; Wilk, 1977; Friedland et al., 1988), and cyprinodontids (Greeley, 1939; Grant, 1962; Wilk, 1977). Prey of minor importance included Argulus sp., Corophium sp., Rhithro- panopus harrisii, unidentified Crustacea and isopods, zoea of Majidae sp., polychaete remains, unidenti- fied Hymenoptera, and both leafy and woody plant ma- terial (Table 2). Estuaries in Maine are probably seasonal nursery grounds where high prey densities and warm water temperatures result in rapid growth. During 1990 and 1991, 13.42% and 3.33% of the fish collected in the Marsh River and 26.47% (1991) of the fish collected in Merepoint Bay possessed empty stomachs (Tables 2 and 3). With the excep- tion of the 16% estimate reported by Grant (1962), who also used gill nets, the percent of empty stom- achs reported by several investigators varied between 30 and 87%. Perhaps the low incidence of empty stom- achs reported in our studies resulted from the use of gill nets, which constrict the gills and esophagus thus reducing the possibility of regurgitation. An analysis of daily growth increments on otoliths Growth increments on the otoliths of many species of juvenile fishes, including bluefish, are produced on a daily basis as long as growing conditions are adequate (Nyman and Conover, 1988). Bluefish col- lected from the Marsh River displayed 94-200 (1 = 126) and 97-176 ( .v=134) daily growth incre- ments during 1990 and 1991, respectively. Fish col- lected from Sagadahoc Bay (1990) displayed 58-93 (v=66) daily growth increments, and fish collected from Merepoint Bay (1991) displayed 125-146 ( ,v=134) daily growth increments. First ring deposi- tion occurs approximately 2—4 days after spawning (Nyman and Conover, 1988) so these daily growth increments correspond to a total of approximately 129 days (Marsh River, 1990), 137 days (Marsh River, 1991), 69 days (Sagadahoc Bay, 1990), and 137 days (Merepoint Bay, 1991) of growth after spawning. Spawning occurs somewhere along the Atlantic coast during practically every month of the year (Table 4). Several "seasonal" periods of significant spawning are evident. Major spawning periods oc- cur during the "spring" (March-May) in the South Atlantic Bight and "summer" (May-September) in the Middle Atlantic Bight. A minor "fall-winter" (Kendall and Walford, 1979) and perhaps "summer- fall" (Collins and Stender, 1987) spawning period also occurs in the South Atlantic Bight. It is unlikely that juvenile bluefish captured in the Marsh River during the spring and in Sagadahoc Bay during the summer originated from the major spring spawning in the South Atlantic Bight or the major summer spawning in the Middle Atlantic Bight. Most juvenile bluefish from the Marsh River originated from a spring spawning during March-May (Fig. 3, A and B). Conservative estimates of the time required to swim from the northern portion of the spawning ground in the South Atlantic Bight to the collection site in the Marsh River during 1990 and 1991 would equal 147 days (1350 km + 9.16 km/day) and 156 days (1350 km * 8.64 km/day), respectively. These estimates of swimming time exceed the known ages of juvenile bluefish derived from daily ring counts (129 days, 1990; 137 days, 1991) even before addi- tional time from physical and biological factors are considered. These factors include 1) swimming into the current from Cape Hatteras to Southwestern Maine (Bumpus and Lauzier, 1965), 2) the possibil- ity that swimming speed is less than twice the body length, 3) decreased swimming speed at night (Olla et al., 1985), and 4) feeding behavior which results in nonlinear movement. Juvenile bluefish from Sagadahoc Bay originated from a summer spawning which occurred mainly during June (Fig. 3C). A conservative estimate of the time required to swim from the northern portion of the spawning ground in the Middle Atlantic Bight to the collection site in Sagadahoc Bay during 1990 would equal 107 days (425 km + 3.97 km/day). This estimate greatly exceeds the known age of the fish (69 days). The estimate of swimming time is in- creased further when physical and biological factors that impede swimming are considered. We believe that juvenile bluefish collected from the Marsh River and Sagadahoc Bay may be derived from unknown spawning areas closer to Maine. Lyman ( 1974) stated that "some of their spawning grounds are known but many remain to be discovered along much of the northeast coast." It is also possible that both major spawning areas known to exist in the South and Middle Atlantic Bights, have extended northward. Another possibility is that "larval trans- port mechanisms and spawning periodicities for blue- fish are considerably more complex than previously believed" (Powles, 1981). The relationship between the log10 of the fork length and the number of daily rings for 1990 and 1991 data is shown in Figure 4. The range of fish 504 Fishery Bulletin 92(3). 1994 23 c < — — C s 3 /. * 5 CJ -Q a — c C o / - r. c/3 Cfi _ 6 6 I e 6 = E S 6 6 -*r I CO -<*■ s ■13 fe = ^ m « s .a » ta — cz: ^ t= eg u .~: ca ^0 — ^.2:0-*. o w Ol Xh '■ M § S Jg 'a. ■=> 1= ~ S g 25s g S _; I !§ .2 * S I S cq o o ea n: u 5 rl 3.0 mm SL (n=5) and not detected in fish smaller than 2.8 mmSL(«=3). In specimens that were not dissected, asterisci were first observed in larvae 3.8 mm SL and were present in all larvae larger than 4.0 mm SL. Considerable differences were found in the gen- eral size and shape of the otoliths (Fig. 1 ). All otoliths were spherical or slightly ovoid in larvae <5 mm SL. Sagittae in larvae >5 mm SL began to develop a ros- tral process, and sagittae of 10-mm larvae were oval, laterally compressed, and had developed a prominent rostrum. The general shape of sagittae did not change in fish 10-50 mm SL. Lapilli followed the same ini- tial development, but the posterior margin of these otoliths became scalloped when larvae were approxi- mately 15 mm SL owing to the formation of numer- ous accessory primordia. Initially asterisci were also spherical, but the axis of growth changed, with sub- sequent development resulting in a kidney-shaped appearance by 15 mm SL. Accessory primordia were seldom observed in sagittae and were not observed in asterisci. Lapilli and asterisci were similar in size; however, the diameter of sagittae was approximately three times larger and grew approximately five times faster than asterisci and lapilli (Fig. 2). David et al.: Age and growth in juvenile Sciaenops ocellatus 51 Diameters of asterisci and lapilli were well corre- tion between otolith diameter and standard length lated with standard length, but the strongest rela- was observed in sagittae (Fig. 2). Anterior rostra and 25 um 100 urn D 25 pm 100 pm Figure 1 Red drum, Sciaenops ocellatus, otoliths: (A) hatchery asteriscus 500x; (B) hatchery lapillus 500x; (C) hatchery sagitta 125x; (D) wild asteriscus 500x; (Et wild lapillus 250x; (Fi wild sagitta 125x. 512 Fishery Bulletin 92(3). 1994 otolith primordia were well defined in sagittae, which, along with a uniform axis of growth, resulted in a more consistent axis of measurement. The lapil- lus and asteriscus were more circular in shape than the sagitta and had a less well-defined anterior ros- trum. The lapillus also formed accessory primordia, which were reflected in an irregular perimeter. Al- though accessory primordia were not observed in asterisci, the axis of growth shifted with develop- ment, resulting in diameter measurements through the primordia that did not necessarily reflect the so ; ' •••• •//' 45 : •V; : • jf* * 40 : •Js 35 ■ J/\ ' • \y • 30- •/v • '. •" , yf ' ' Asteriscus 25 " yS " SL = -6.18 + 0.074 (D) 20 " " yC r2=0.90 15 " "v* ' — . ^r ' 1- ■ 1 ' 1 ■ ■ 2C 0 300 400 500 600 700 800 9C 0 «-» 50 - E * • s^* § «- *':» j^ -C 40 " 'r/* ' +* O) ^-*» " ?, 35- s< • ' a> *r • _i y^*n . • 30 " T3 //'' "•' . Lapillus CO 25 - yi\ SL = -2.33 + 0 062 (D) T3 s^ 2 C / r = 088 CO 20 " f *y 05 15- -/* — . ^-t . 1 ■ 1 r 1 ' T ' 1 2 )0 300 400 500 600 700 800 9 )0 50 - • jS ■ •jS • 45- Z* 40 " \y* 35- y^'' 30" y\'''' Sagitta 25 - V-r ' SL = 0 14 + 0 016 (D) yf' r2=0 95 20 - • ~ /< 15 -t— ■ 500 , , r-^T 1 1 1 1 1 l > >— - * r * ' ' ' 1 ' ' ' ' ' 1000 1500 2000 2500 3000 Diameter (urn) Figure 2 Fish length and otolith diameter relationships for each otolith type of red drum Sciaenops ocellatus. Otolith diameter (nm) plotted on standard length (mm) with linear regression lines overlaid i /; =70 ). maximum otolith diameter. These features resulted in greater variability of diameter measurements for asterisci and lapilli than for sagittae. The accuracy and precision of age estimates var- ied among otoliths. Fish used to estimate daily ages were collected 46 days after hatching. Although in- crements were present in all three otoliths, the defi- nition and resolution of the rings differed, especially near the primordia. The mean ring counts, standard deviations, and coefficients of variation are presented in Table 1. All increment widths exceeded the limit of resolution for light microscopy by more than an order of magnitude (Jones and Brothers, 1987; David and Paul, 1989). All ring counts substantially underesti- mated the age of the 46-day-old hatch- ery-reared fish. The asteriscus underes- timated the true age by 6 days, while sagittae and lapilli underestimated age by 21 days and 25 days, respectively. To obtain realistic age estimates using asterisci, ring counts were adjusted be- cause that otolith is not formed at hatch- ing, but age estimates from ring counts in lapilli and sagittae were erroneously low because all rings could not be ob- served and counted because of poor con- trast in the nuclear region. In addition to providing the most accurate age, coef- ficients of variation indicated that the asteriscus was also the most precise in- dicator of true age. Discussion Shrinkage of larvae due to preservation has been found to be significant (Blaxter, 1971; Theilacker and Dorsey, 1980; Hay, 1982; Brothers et al., 1983; and Leak, 1986). We did not adjust length measure- ments for shrinkage due to preservation because most specimens were juveniles (15-50 mm SL), and because we imme- diately fixed specimens in ethanol which has been shown to minimize the problem of shrinkage (Radke, 1989). Daily increment formation has been validated in sagittae by using laboratory- reared red drum that were up to 21 days posthatch (Peters and McMichael, 1987). The 6-day underestimate in age deter- mined from the asteriscus of known-age i hatchery-reared) fish corresponded well with age at otolith formation (6-7 days) David et al.: Age and growth in juvenile Sciaenops ocellatus 513 Table 1 Comparison of ages derived from sagittal sections of each otolith type in known age (46 days), hatchery- reared red drum, Sciaenops ocellatus. Values are reported for mean age, standard deviation, and co- efficient of variation. Otolith N Mean age (d) SD CV Asteriseus Sagitta Lapillus 50 50 50 39.7 25.0 21.0 3.2 4.6 5.5 8.0 18.6 26.3 determined by examination of the age series of lar- vae, indicating that rings were indeed formed daily, and all rings were visible. Therefore, we assumed that rings were also formed daily in the lapillus and that underestimation of age was due to the inability to observe rings of low contrast in the opaque nuclear area rather than to other possibilities that have been reported, i.e. nondaily ring formation due to poor growth in herring, Clupea harengus, and turbot, Scophthalmus maximus (Geffen, 1982), and to ring spacing below the resolution limit of light micros- copy in striped bass, Morone saxatilis, under subop- timal feeding regimes (Jones and Brothers, 1987). Peters and McMichael ( 1987 ) had difficulty distin- guishing the innermost rings in sagittae of some ju- veniles, and they developed an ageing method that did not require counting these rings. This method utilized transverse sections of larval sagittae in which rings were clearly visible to determine the distance from the primordium to the tenth ring. Subsequent juvenile ring counts were initiated at this distance away from the primordium. We also encountered dif- ficulty in detecting all the rings near primordia in sagittae (and lapilli as well) due to the opacity of the nuclear region of the otoliths, but all rings were usu- ally clearly visible in asterisci. Peters and McMichael (1987) made relatively accurate ring counts beyond the tenth ring on sagittae of 21-day-old known-age fish. However, we were not able to duplicate their success using our technique, because our counts underestimated the true age of 46-day-old fish by 21 days with a coefficient of variation of 18.6%. Duplicat- ing their glycerin soaking technique did not improve our ring detection capability. The principle difference between our method and that of Peters and McMichael ( 1987) was the sectioning plane; they used the trans- verse plane, whereas we used the sagittal plane. Because the otoliths were birefringent, they were easily observed within larval fish when illuminated with transmitted light between crossed polaroids. Sagittae and lapilli were easily seen in fish as small as 1.3 mm SL, whereas asterisci were observed in all fish greater than 3.0 mm SL. Comyns et al. ( 1989) detected two rings in sagittae and lapilli of larvae 2 days after fertilization ( 1 day posthatch), demonstrat- ing that sagittae and lapilli are present at hatching; according to their growth curve, fish 3.0 mm SL were approximately 6 days old. Whereas the size at hatch was relatively constant, growth varied considerably with temperature, but the variation was not signifi- cant until the fish attain 4.0 mm SL (Comyns et al., 1989). Changes in growth rates due to water tem- perature are therefore not likely to have a signifi- cant effect on our estimate of age at asteriseus for- mation. Others have reported asteriseus formation at similar ages, e.g. at age 6 days in the Japanese eel, Anguilla japonica (Umezawa et al., 1989). The asteriseus provided the most accurate estimate of age for juvenile red drum because it underesti- mated true age by only 6 days compared with 21 days for the sagitta and 25 days for the lapillus. The asteriseus ages were also the most precise, because the coefficient of variation was only 8.0% compared with 18.6 and 26.3% for the sagitta and lapillus re- spectively. Most of the variance in age estimates for all three otoliths was caused by the inability to re- solve rings near the primordium and at the margin. Rings in the mid portion of asteriseus sections were consistent in shape, increment width, and clarity; a dominant, consistently identifiable, first ring was visible at the edge of the primordium. The sagitta provided the next most accurate and precise estimates. Ages were underestimated by 21 days on average, and rings near the primordia were difficult to distinguish, resulting in a coefficient of variation more than a factor of two higher than that of the asteriseus. Because sagittae were present at hatching and early rings were detectable in smaller and younger fish (Peters and McMichael, 1987; Comyns et al., 1989), better accuracy and precision of ageing using sagittae may be possible with im- provements in preparation and processing of the otolith. However, the Spurr mounting technique (Haake et al., 1982) used by Peters and McMichael ( 1987) and Comyns et al. ( 1989) can require several days for proper dehydration, curing, sectioning, and polishing whereas our polymer method required less than one hour to produce ground and polished slides from whole fish, thus allowing considerably more fish to be processed in comparable time periods. Ages estimated from the lapillus were least accu- rate and precise because of poor clarity of rings near the primordia and because of the formation and fu- sion of numerous accessory primordia. This fusion of accessory primordia resulted in superimposition of rings near the margin and the presence of several planes of growth being visible in the same focal plane. 514 Fishery Bulletin 92(3). 1994 Although additional preparation may have increased resolution of rings near the primordia, poor ring reso- lution caused by accessory primordia would probably not have been improved. The addition of a constant to asterisci ring counts (6 days) adjusted for the time lag between hatching and otolith formation and was not used to compen- sate for uncounted rings. The addition of constants to estimates of age did not affect the rank of the co- efficients of variation. All otolith diameters, especially sagittae, exhibited a strong correlation with fish lengths. This relation can be useful in backcalculating size at age. How- ever, because the ring count did not accurately esti- mate age, accurate ages derived from sagittae can not be associated with back-calculated sizes. The relation between asteriscus diameter and fish length was not as strong as with sagittae, but the improve- ments in accuracy and precision in estimating age would increase the confidence in back-calculated sizes at age. However, because asterisci were not present at hatching, size-at-age information for fish <7 days old could not be backcalculated. Rings formed in the asteriscus were concentric and proportionally spaced throughout the otolith because no accessory primordia were formed, making it straightforward to measure radii or increment widths precisely re- gardless of the chosen axis of measurement. We conclude that the asteriscus is the best struc- ture to use in ageing young red drum >4.0 mm SL. Using our sagittal section technique with juveniles, we found that the asterisci clearly provided superior accuracy and precision in ageing. Previous efforts by Peters and McMichael (1987) to age juvenile red drum using transverse sections of sagittae provided reasonable age estimates; however, because they began counts a standard distance from the primor- dium to allow for uncounted rings ( 10), the accuracy and precision of resulting counts on 21-day-old fish cannot be known in the same sense that we estimated these statistics when all rings were counted on asterisci of 46-day-old fish. The advantage of using asterisci is that clear rings can be seen in sagittal sections, and grinding asterisci in the sagittal plane can be done relatively quickly. Furthermore, all rings were visible in the asteriscus, and only adjustment for age at formation was required to estimate true age. Literature cited Arnold, C. R. 1988. Controlled year-round spawning of red drum Sciaenops ocellatus in captivity. Contrib. Mar. Sci. Supp. to Vol. 50:65-70. 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 chalcogramma (Pallas). J. Exp. Mar. Biol. Ecol. 122:147-161. Blaxter, J. H. S. 1971. Feeding and condition of Clyde herring larvae. Rapp. P.-v. Reun. Cons. Int. Explor. Mer 160:128-136. Brothers, E. B., and W. N. McFarland. 1981. Correlations between otolith microstructure, growth, and life history transitions in newly re- cruited French grunts (Haemulon flavolineatum I. Rapp. P.-v. Reun. Cons. int. Explor. Mer 178: 369-374. Brothers, E. B., E. D. Prince, and D. W. Lee. 1983. Age and growth of young-of-the-year bluefin tuna, Thunnus thynnus, from otolith micro- structure. U.S. Dep. Commer, NOAATech. Rep. NMFS 8:49-59. Campana, S. E., and J. D. Neilson. 1985. Microstructure offish otoliths. Can. J. Fish. Aquat. Sci. 44:1014-1032. Comyns, B. H., J. Lyczkowski-Shultz, C. F. Rakocinski, and J. P. Steen Jr. 1989. Age and growth of red drum larvae in the North-Central Gulf of Mexico. Trans. Am. Fish. Soc. 118:159-167. Comyns, B. H., J. Lyczkowski-Shultz, D. E. Neiland, and C. A. Wilson. 1991. Reproduction of red drum, Sciaenops ocellatus, in the north-central Gulf of Mexico: sea- sonality and spawning biomass. In R. D. Hoyt (ed.), Larval fish recruitment and research in the Americas: proceedings of the thirteen annual lar- val fish conference; Merida, Mexico, May 1989, p. 17-26. NOAA Tech. Rep. NMFS 95. David, A. W., and J. H. Paul. 1989. Enumeration and sizing of aquatic bacteria by use of a silicon-intensified target camera-linked image-analysis system. J. Microbiol. Methods 9(41:257-266. Fuiman, L. A., and D. R. Ottey. 1993. Temperature effects on spontaneous behav- ior of larval and juvenile red drum, Sciaenops ocellatus, and implications for foraging. Fish. Bull. 91:23-35. Geffen, A. J. 1982. Otolith ring deposition in relation to growth rate in herring (Clupea harengus) and turbot (Scophthalmus maxim us) larvae. Mar. Biol. 71:317-326. Haake, P. W., C. A. Wilson, and J. M. Dean. 1982. A technique for the examination of otoliths by SEM with application to larval fishes. /;/ ('. F. Bryan, J. V. Conner, and F. M. Truesdale (eds. ) Pro- ceedings of the fifth annual larval fish conference, p. 12-15. LSU Press, Baton Rouge, LA. Hay, D. E. 1982. Fixation shrinkage of herring larvae: effects David et al.: Age and growth in juvenile Sciaenops ocellatus 515 of salinity, formalin concentration, and other factors. Can. J. Fish. Aquat. Sci. 39:1138-1143. Holt, S. A., C. L. Kitting, and C. R. Arnold. 1983. Distribution of young red drum among differ- ent sea-grass meadows. Trans. Am. Fish. Soc. 112:267-271. Irie, T. 1960. The growth of the fish otolith. J. Fac. Fish. Anim. Husb. Hiroshima Univ. 3( 11:203-221. Jones, C. 1986. Determining age of larval fish with the otolith increment technique. Fish. Bull. 84:91-103. Jones, C, and E. B. Brothers. 1987. Validation of the otolith increment ageing technique for striped bass, Morone saxatilis, lar- vae under suboptimal feeding conditions. Fish. Bull. 85:171-178. Leak, J. C. 1986. The relationship of standard length and otolith diameter in larval bay anchovy, Anchoa mitchilli (Val.l: a shrinkage estimator. J. Exp. Mar. Biol. Ecol. 95:167-172. Mercer, L. P. 1984. A biological and fisheries profile of red drum, Sciaenops ocellatus. N.C. Dep. Nat. Res. Comm. Devel., Div. Mar. Fish., Spec. Sci. Rep. 41, 89 p. Panella, G. 1971. Fish otoliths: daily growth layers and peri- odical patterns. Science (Washington, D.C.) 173:1124-1127. Peters, K. M., and R. H. McMichael Jr. 1987. Early life history of the red drum, Sciaenops ocellatus (Pisces: Sciaenidae), in Tampa Bay, Florida. Estuaries 10:92-107. Radke, R. L. 1989. Larval fish age, growth, and body shrinkage: information available from otoliths. Can. J. Fish. Aquat. Sci. 46:1884-1894. Theilacker, G. H., and K. Dorsey. 1980. Larval fish diversity, a summary of labora- tory and field research. In G. D. Sharp (ed. ), Work- shop on the effects of environmental variation on the survival of larval pelagic fishes, p. 105- 142. FAO, Intergov. Oceanogr. Comm., Workshop Rep. 28. Umezawa, A., K. Tsukamoto, O. Tabeta, and H. Yamakawa. 1989. Daily growth increments in the larval otolith of the Japanese eel, Anguilla japonica. Jpn. J. Ichthyol. 35(4):440-444. Abstract. — Squirrelfish of the genus Myripristis are valued in small-scale fisheries throughout much of the tropics. The life his- tory and species biology of most of these soldierfishes is poorly known. For the brick soldierfish, M. amae- na, in Hawaii and Johnston Atoll, we found that sexual maturity for both sexes was reached between 145 and 160 mm standard length at about six years of age — a large fraction of the apparent maximum size and lifespan. Fecundity was relatively low and increased as the fifth power of body weight. Spawn- ing peaked from about early April to early May, and a secondary peak occurred in September. Myripristis amaena is a nocturnal predator, feeding mostly on meroplankton. especially brachyuran crab mega- lops, hermit crab larvae, and shrimps, but also taking a variety of benthic prey. In pristine fish communities, holocentrids were abundant, quantitatively impor- tant (often dominant) reef preda- tors and prey. Myripristis amaena (and probably other common and important soldierfish) seems to be relatively long lived (at least 14 years), slow growing, and late ma- turing. The populations suffer con- siderable natural predation and depend mainly on the largest and oldest fish for reproduction. Heavy, unregulated fishing of these soldierfish, especially at prerepro- ductive size, may severely reduce populations. Reproductive and trophic ecology of the soldierfish Myripristis amaena in tropical fisheries Anderson J. Dee PO Box 1 154 Hayward. Wisconsin 54843 James D. Parrish* National Biological Survey, Hawaii Cooperative Fishery Research Unit 2538 The Mall. University of Hawaii, Honolulu, Hawaii 96822 Manuscript accepted 30 December 1993. Fishery Bulletin 92:516-530 ( 1994). 516 Soldierfish, Myripristinae, of the squirrelfish family, Holocentridae, occur widely throughout the trop- ics (Greenfield, 1965, 1968, 1974). They are typically abundant and are an important component of com- mercial, recreational, and subsis- tence fisheries in much of the world's tropics. Throughout most of the central and western Pacific Ocean, the brick soldierfish, Myri- pristis amaena (Castlenau), is an important member of this group (Greenfield, 1968). It contributes significantly to fish communities and to fishery catches in shallow reef and rocky habitats. It is par- ticularly important in the recre- ational fishery at Johnston Atoll (JA), where it is typically the spe- cies caught in greatest abundance (Irons et al.1). It is also common in catches throughout the Hawaiian archipelago. Relatively little quantitative in- formation has been published about the life history and biology of spe- cies of the genus Myripristis, and very little is available about M. amaena in particular. Data about diets are available for only a few species of Myripristis; for most of these, sample sizes are small (e.g. 14 specimens for M. amaena; Hobson, 1974). Results on age and growth from our studies of the JA population of M. amaena have been reported (Dee and Radtke, 1989). There have been no thorough pub- lished studies of the reproduction of M. amaena or closely related spe- cies. Because of the wide distribu- tion, considerable abundance, and substantial fishery importance of M. amaena, we undertook to de- scribe more fully its food require- ments and trophic position in the community and to quantify the re- productive characteristics that af- fect the dynamics of its populations. The parameters we determined may also provide reasonable first ap- proximations for similar Myripristis species that are less well studied. The results will contribute to an informed approach to management of species that are now typically un- managed and probably overfished in most localities with even moder- ately dense human populations. The JA population of M. amaena was the major focus of this study for several reasons. Many biological and ecological characteristics of M. amaena (e.g. size, morphology, hab- its, habitat used, fishery value) seem to be representative of a num- To whom correspondence should be sent. Irons, D. K.. R K. Kosaki, and J D. Parrish. 1990. Johnston Atoll resource sur- vey. Final report of Phase Six 1 21 July 1989-20 July 1990). Project rep. to U.S. Armv Engineer Distsrict, Honolulu. HI, 150 p. Dee and Parrish: Reproductive and trophic ecology of Myripnstis amaena 517 ber of common soldierfish species. Johnston Atoll provided a logistically good base for study, where M. amaena was the dominant Myripristis species, with populations not seriously depressed by fishing. The species was plentiful and easily collected at many locations throughout the year, and a good range of sizes was readily available. The importance of M. amaena in the fishery at JA facilitated collection of specimens and fishery data. Materials and methods Specimen collection and handling Most specimens used in all analyses were taken at Johnston Atoll, a remote, coral-rich, oceanic pinnacle about 1250 km SW of Honolulu (Halstead and Bun- ker, 1954; Amerson and Shelton, 1976; Randall et al., 1985; Maragos and Jokiel, 1986). Smaller collec- tions were made from a rich, fringing coral reef tract at Puako in South Kohala on the leeward coast of Hawaii Island (Hayes et al., 1982). Afew specimens were collected from the almost uninhabited North- western Hawaiian Islands (NWHI), primarily from shallow, coralline areas at French Frigate Shoals and Midway ( located about 750 and 2000 km, respectively, northwest of Honolulu). Specimens from all three locations contributed to size-frequency analysis, visual assessment of gonad condition (maturity), and analysis of gut contents. At JA and Puako, gonad weight was taken to deter- mine reproductive season and size at first reproduc- tion; at JA, gonad samples were preserved for histo- logical examination and estimation of fecundity. Most specimens at all locations were collected from shallow waters (<15 m). The major methods of col- lection were spearing, bait casting with a line, and some spot applications of ichthyocide. At JA, we col- lected specimens from several sites inside the lagoon and just outside the barrier reef. Collections and ex- tensive underwater observations were made at fre- quent intervals between February 1984 and Janu- ary 1986. Sampling was less intensive in some months because of constraints on travel to JA. At Puako and in the NWHI, specimens were collected rather widely within coralline habitats. At Puako, about half the specimens were collected as quickly as feasible (May-June 1981), once the species was found to be in reproductive season. Other collections there were distributed throughout the year. In the NWHI, specimens were collected in March, April, May, August, and November. Standard (SL), fork (FL), and total (TL) lengths of all captured specimens were measured to the near- est millimeter, and weights were taken to the near- est 0.1 g. (Appendix A provides functions, fitted by regression, to convert between SL, FL, and TL.) Whole guts and gonads were excised and preserved in 10% buffered formalin for further analysis. Some specimens were frozen whole and stored for some weeks or months before processing. Source of size-frequency data At JA, Puako, and the NWHI, length and weight data from all specimens collected for other purposes were available for size-frequency estimation. In addition, at JA, creel census data were obtained from fisher- men's catches on many days over the period of study. Fish landed by boat and shore fishermen were ex- amined, and each specimen was measured and weighed as above. These data provided a much larger and possibly more representative sample than our collections alone. Feeding The volume of each complete gut specimen was mea- sured by displacement of water before and after the gut was opened and all contents removed. The total volume of contents was determined by difference. All diet items were sorted into systematic categories and identified to the lowest possible taxa. For each prey category, the number of individuals, length, extent of digestion, location in gut, and volume were re- corded. Whole reference specimens were used as an aid in identifying prey items and estimating the origi- nal size of the prey by comparison of the dimensions of recognizable parts. Volume of remains in each prey category was estimated by displacement of water (Wolfert and Miller, 1978).' A measure of overall importance of each prey cat- egory was calculated by using the Index of Relative Importance (IRI), as defined by Pinkas et al. (1971): IRI = frequency % x (numerical ck + volume %), where frequency % = numerical % = and volume % (the number of guts contain- ing prey of one category di- vided by the total number of guts that contained any iden- tifiable prey) x 100; (the number of individuals of one prey category divided by the total number of prey in- dividuals found in all the gutslx 100; (the volume of one prey cat- egory divided by the total volume of all prey found in the guts) x 100. 518 Fishery Bulletin 92(3). 1994 Reproduction Gonads of 430 specimens from JA were visually ex- amined macroscopically and classified (based on their size, color, texture, and morphology) as male, female, immature, or unknown. A subsample of these gonads from specimens collected during probable reproduc- tive and nonreproductive periods was removed, wet weighed to the nearest 0.001 g, and preserved in 709£ isopropyl alcohol for further examination. Gonads selected for histology were prepared and embedded by using Kahles solution (Guyer, 1953, p. 236) with a graded series of ethyl alcohol and butyl alcohol dehydration. Embedded subsamples taken from the anterior, middle, and posterior regions of the selected gonads were then sectioned at 5 and 10 urn, mounted on plain microscope slides, and stained by using Dela- field hematoxylin and eosin Y techniques (Humason, 1979, p. 112, 119-123). Size at first reproduction (SFR) was estimated based on visual examination of gonads and by using the gonadosomatic index (GSI), where: GSI = (gonad wet weight/whole body wet weight) x 100. The GSI was plotted against SL for male and female M. amaena collected from periods during which the species seemed to be reproductively active. A sharp rise in the GSI at some length indicated the SFR. The SFR was also estimated based on histological examination of gonads from specimens collected dur- ing the April 1984 spawning peak. Eggs were exam- ined for size and stage of development from sections of 11 females representing a range of sizes. Staging was based on size, morphology, and staining proper- ties of eggs (Khoo, 1978; Wallace and Selman, 1981). Testes from five males were examined histologically for presence of sperm. Spawning season was estimated by plotting GSI against month of capture for 99 sexually mature males and females (larger than 145 mm SL) collected throughout the sampling period. Histological sections of samples from March, April, May, July, and August 1984 were examined for further validation of repro- ductive periods. We did not identify individual clutches of ova in females. Based on the histology, we identified ova >0.4 mm on the major axis as being in an advanced stage of yolk development and defined this stage as mature (Table 1, See Fig. 3). Our fecundity value is an estimate of the number of such ova in a female specimen. Ovaries from 12 gravid females collected at JA during the January 1986 spawning peak were used to estimate fecundity. Three 0.02-g aliquots were taken from the midsection of each ovary. All mature ova (>0.4 mm on the major axis) were counted with the aid of a binocular dissecting microscope. Fecun- dity, F, was estimated for each female from the formula: F=W1+N2+N3)/3)x(G/A), where Nl,N.2, N3 = the number of mature ova in each aliquot; G = total gonad weight (g); A - weight of a gonad aliquot (0.02 g). For specimens from Puako, the same procedures were followed except that no histological prepara- tion or examination was done, and ripeness was es- timated simply on the basis of visual appearance of Table 1 Stages of development of oocytes in the brick soldierfish, Myripristis amaena (also see Fig 3). Stage Oocyte diameter (mm) Description oogonia and primary growth 0.01-0.04 scant cytoplasm; centrally located nucleus; single large nucleolus; stains dark red perinuclear 0.04-0.056 multiple nucleoli around inner side of nuclear membrane; stains dark purple secondary growth and yolk vesicle 0.056-0.14 clear-staining cortical alveoli begin as ring at cytoplasm periphery, then increase in size and occupy whole cytoplasm yolk granule 0.14-0.40 true vitellogenesis; red-staining yolk granules begin to form around cytoplasm periphery (early vitellogenesis) then increase in size and occupy whole cytoplasm (late vitellogenesis); zona radiata first appears maturation 0.40-0.54 yolk granules fuse into yolk "plates" which stain light blue; fusion begins at center and spreads throughout cytoplasm Dee and Parrish: Reproductive and trophic ecology of Myripnstis amaena 519 the gonad and GSI. The SFR, but not spawning sea- son, was estimated quantitatively by using these measures. Egg sizes were not measured, and fecun- dity was not estimated. Results Feeding Guts from 64 specimens collected at JA at night con- tained identifiable prey items (Table 2). Crab larvae dominated the diet, producing much larger IRI val- ues than any other major systematic group (Table 3). Brachyuran megalops (mostly Portunidael were found in over 90% of the analyzed guts. Despite their relatively small size, they were important in volume (28%) as well as numbers (38%). Hermit crab larvae (Paguroidea) were present in over half the guts, and alpheid shrimp in slightly less than half. Both these groups provided significant fractions of all prey num- bers and volume and had large IRI values; hermit Table 2 Diet ofMyripristis amaena based on 64 specimens fr om Joh nston Atoll, 9 from the Northwestern Hawaiian islands, and 22 from Puako, Hawaii Is land. Results reported by Hobson (1974) based on 14 specimens from Kona, Hawaii Island, are also included for comparison. Table shows the per cent of predator ndividuals that consumed each prey (F), and the percent of all numbers (N) and volume (V) provided by eacl prey. Va ues for the highest systematic levels are printed in bold. (See Table 3 for Indices of Relative Im portance for these highest level groups. ) Numerical percent' Volume jercent' Frequency percent N.W. Puako N.W. Hawaii Island N.W. Hawaii Island Johnston Hawaiian (Hawaii Johnston Hawaii; Johnston Hawaiian in Prey category Atoll Islands Island I Atoll Island i Puako Kona- Atoll Islands Puako Kona1' CrabsJ 60.8 33.3 71.2 50.3 8.4 35.0 75.1 100 11.1 68.2 100.0 Portunidae 37.8 27.6 90.6 Paguroidea (hermit crab) 23.0 0.3 22.7 4 59.4 4.6 Galatheidae 8.5 7.7 4.6 Shrimp 11.8 22.2 11.4 10.8 8.4 >2.0 9.3 46.9 11.1 27.3 28.6 Alpheidae 11.8 3.5 10.8 0.4 46.9 13.6 Palaemonidae 2.0 0.6 13.6 Hippolytidae 3.0 >0.8 13.6 Caridea unidentified 0.6 4 4.6 Shrimp larvae 0.9 4 4.6 Shrimp unidentified 1.5 0.2 4.6 Stomatopods 1.6 11.1 LI 6.7 8.4 6.5 0.1 37.5 11.1 22.7 7.1 Lobsters 0.06 0.9 4 21.5 3.1 13.6 Polychaetes 12.0 0.3 15.3 0.4 0.1 70 3 4.6 7.1 Eunicidae 6.7 11.7 60 5 Nematonereis sp. 6.3 2.8 59.4 Eunice sp. 0.4 8.9 9.4 Ophehidae 5.3 - 3.6 43.8 iPolyophthalmus sp.) Fish 2.4 33.3 l.K 10.5 33.6 5.8 2.9 32 5 22.2 22.7 21.4 Eels, unidentified 2.2 8.5 21.9 Mysids 5.9 0.3 1.3 4 0.3 18.8 4.6 28.6 Amphipods 0.1 5.0 0.1 >0.2 (1.2 9.4 18.2 7.1 Gammaridea 4.7 0.2 0.2 13.6 7.1 Tanaids 5.4 2.1) 50.0 Copepods 1.2 4 (M 9.1 7.1 Isopods D.I 7.1 Cephalopods 0.3 1.5 1.4 4.6 7.1 Octopus sp. 0.3 1.5 4.6 Gastropods 3.5 1.4 0.1 22.7 14.3 Crustacean parts, unidentified 6 " 41.2 25.8 9.7 66.7 77.3 57.1 ' Data for unidentified crustacean parts are exclu ded from the ca culation 2 Data from Hobson (1974); numerical percents were not reportec 3 Larvae (mostly megalops) except those from the Northwestern rlawaiiar Islands that were juvenile m ajids. ■* Data are missing. 5 Calculated from a subset of 31 specimens. h Not countable. 520 Fishery Bulletin 92(3), 1994 Table 3 Summary of the diet of Myripristis aniaena at the highest systematic levels of prey. Results are shown as percents of the summed Index of Relative Impor- tance (IRI). (See Table 2 for details at these and other systematic levels.) Northwestern Puako Johnston Hawaiian (Hawaii Prey Atoll Islands Island) category (rc=64) (71=9) (n=22) Crabs 72.3 18.5 84.8 Shrimp 6.9 13.6 4.3 Stomatopods 2.0 8.6 2.8 Lobsters ' 3.6 Polychaetes 12.4 0.04 Fish 2.7 59.3 2.0 Mysids 1.2 i Amphipods 0.01 1.1 Tanaids 2.4 Cephalopods 0.1 Gastropods 1.3 Crustacean parts, unidentified Few1' Very many1' Very many- 7 Data are missing. 2 Not countable; not included in computation oflRI. crabs were more important by all measures. Two M. amaena specimens contained a total of three lobsters. Fish were moderately important prey by frequency and volume. Many prey specimens appeared to be eels, probably mostly ophichthids. Polychaetes were the most important benthic prey, present in about 70% of all fish and providing over 10% of all prey by numbers and volume. Twenty-two specimens from Puako contained prey identifiable to some level (Table 2). Nearly 70% of these contained crab megalops larvae; a hermit crab was identified in one of these and a galatheid crab, Galathea spinosorostris, in another. Crab larvae were also the major prey in numbers (over 70% ) and vol- ume (35%) and were strongly dominant in the IRI (Table 3.). Shrimp were eaten by the second largest number offish (nearly 30%) and were second in im- portance numerically (over 10%), but minor in vol- ume. They were far less an important prey at Puako than at JA, and were more evenly divided among Alpheidae, Palaemonidae, and Hippolytidae (includ- ing two Saron marmoratus ). Lobsters ( three individu- als) appeared in only three guts ( 14%), but they ac- counted for over 20% of total prey volume. Both the slipper lobster, Scyllarides squammosus, and the Hawaiian clawed lobster, Enoplometopus occi- dentalis, were eaten. No other groups made major- individual contributions to the diet. However, sto- matopods and fish accounted for at least several per- cent of the number of consumers (frequency %) and volume. One octopus appeared in the diet at Puako. Benthic prey included at least a few percent gastro- pods (by the various measures) and a polychaete specimen. Nearly 80% of all fish also contained some quantity of unidentifiable crustacean parts, which accounted for over 25% of the total prey volume. Calculated as a percent of only the total prey identi- fiable to more specific groups, the volume % of each of those groups was considerably higher. Only nine specimens from the NWHI contained prey identifiable to any level (Table 2). The total amount of prey recovered was small. Fish (including at least one pomacentrid), shrimp, juvenile crabs (in- cluding majids), and stomatopods were found in one or two guts each in total numbers of one to three individuals each. Six guts contained unidentifiable crustacean parts, which accounted for over 40% of the total diet volume. Fish were next in volume, and the other three groups contributed little volume. At all three locations, the diet of M. amaena was heavily dominated by small juvenile stages of crus- tacean species that become much larger as adults. These prey included all the major groups of large crustaceans that are dominant in the diets of many demersal species of reef fishes studied in Hawaii (Parrish et al., 1985). Besides the crab and shrimp prey — quantitatively important in the diet of M. amaena at all three locations — stomatopods were eaten at all locations, and were the dominant prey of the Myripristinae generally in the NWHI (Parrish, unpubl. data). Pseudosquilla oculata was a stomato- pod species identified in our M. amaena specimens that also was prominent in the diets of other Myri- pristinae and other demersal reef fishes in our stud- ies in Hawaii. Small peracaridan crustaceans were conspicuous in the diet at JA (Table 2). Mysids and tanaids were present in the gut contents of many fish specimens, and each group contributed a few percent of the to- tal number and volume of prey. Only one individual of each was identified at Puako. Amphipods were somewhat less widely found as diet components at JA and their abundance was trivial. At Puako they were present in about 18% of the fish and made up about W'< of numbers but were trivial in volume. Most of those identified were gammarids. A few isopods were found in JA specimens, and traces of copepods were found in Puako specimens. More peracaridans and other small crustaceans were very likely part of the residue of unidentifiable crustacean parts that appeared widely at all locations. Myripristis amaena showed the expected strongly nocturnal feeding habit. At JA, guts from all 64 speci- Dee and Parrish: Reproductive and trophic ecology of Mynpristis amaena 52! mens taken between 2000 and 0200 hours contained identifiable prey items. From 62 guts collected in the daytime, only four (taken at 0730 hours) contained more than one identifiable food item. At Puako, all 11 guts collected between 2300 and 0930 hours contained identifiable prey items, all of which might reasonably be the result of night or dawn feeding. Eleven specimens taken between 1000 and 1300 hours con- tained some food, much of it only uniden- tified crustacean parts. Sixteen specimens collected between 1100 and 1500 hours contained no identifiable prey. Of 61 speci- mens from the NWHI, only nine (collected between 0800 and 1100 hours) contained identifiable prey, mostly considerably di- gested. The remaining 52 specimens, col- lected between 1000 and 1600 hours, were empty or contained well digested, uniden- tifiable material. Reproduction The GSI of each sex collected during reproductive periods throughout the year from JA and Puako com- bined indicated an increase in the range of 153-156 mm SL for females (Fig. 1) and in the range of 149- 156 mm SL for males (Fig. 2). Based on histological examination of developing oocytes in JA females, five developmental stages were identified (Table 1). 5-i . JA 4/84 D ■ JA 10/85 o o JA 1/86 * 4- X 2 A o -a ro c o O 1 - o Puako 5-6/81 • • • o • o • • • • • • • • • o o a ■ ■ ■ •• • • u t I I I I 1 1 1 1 I I 120 140 160 180 200 220 Male SL (mm) Figure 2 Values of gonadosomatic index and corresponding standard length for male brick soldierfish, Myripristis amaena, collected during reproductive peaks at Johnston Atoll ( JA) and Puako, Hawaii. 5 -, • JA 4/84 ■ JA 10/85 o JA 1/86 ~ 4 - Q o Puako 5-6/81 £ X 0J D TJ D £ 3 - o D 03 O E o co 2 - o • ° TD o o 03 C o • a O • ° 1 - o o OD ° otm m , ,)•■«■ •am* • ■ o - 1 i r i i i i i i i i >0 140 160 180 200 220 Female SL (mm) Figure 1 Values of gonadosomatic index and corresponding standard length for female brick soldierfish, Mynpristis amaena, collected during reproductive peaks at Johnston Atoll ( JA) and Puako, Hawaii. Sexual maturity was defined based on the presence of ovaries with oocytes in the late yolk granule stage for females (>0.4 mm diameter; see Fig. 3) and on the presence of mature sperm in males (Fig. 4). The 11 ovaries and five testes examined histologically indicated that first sexual maturity occurred near 154 mm SL for females and 149 mm SL for males. The results were consistent with those from visual examination of 24 total gonads (males and females) from Puako. When GSI of mature specimens (larger than 145 mm SL) from JA was plotted against month of capture, with data from all months except September, spawning peaks were discernible in January, April, May, and October; the major peak was in April (Fig. 5). Based on visual examina- tion, GSI, and histology, no gravid indi- viduals were recorded among specimens collected during any other month. (Samples were rather small in some months.) The fecundity estimated from the 12 ovaries sampled at JA ranged from 12,400 mature ova for a 156-mm SL fish to 69,200 for a 181-mm SL fish (Appendix B). Re- gressions were performed with fecundity, F, in number of mature eggs as a function of standard length, SL, in millimeters, and alternatively as a function of whole wet body weight, W, in grams. The fit was 522 Fishery Bulletin 92|3). 1994 < Figure 3 Light micrograph of a section of ovary from a brick soldierfish, Myripristis amaena, showing oocyte development stages: (a) primary growth, (b) perinuclear, (c) early yolk vesicle, (d) late yolk vesicle, (e) yolk granule, and (f) maturation. slightly better for power functions than for functions in both cases: inear F= 5.029 x 10-20 ■-.. Figure 4 Light micrograph of a section of testis from a brick soldierfish, Myripristis amc pockets of mature sperm. ia. (S) indicates large 5 -i • JA o Puako o ^ 4 - o £ • X CD • 0 o c 3 - • o u o • E . o p - • • o • • o T3 (C • c o •■ ; • o ° 1- • m o m • o J? .. • • ° ^ • . - * M * t •> f •* t •■ m •» <" ^■•••trtlli at. •• r • U ^ «* ^ ^ ^ 5>* S» ^ e? G* ^ c*« Month Figure 5 Values of gonadosomatic in dex and corresponding dates of capture for brick soldierfish, Myripristis amaena, collected at Johnston Atoll f JA) and Puako, Hawaii. 524 Fishery Bulletin 92(3). 1994 70 / • • / 60 7 • / 50 / m • / Fecundity x 10 o • / ■c 30 o 03 m 20 y^ /* v^* 10 - ^^^ Q Jill! 125 150 175 200 225 Total Body Weight (g) Figure 6 Batch fecundity estimate, F, and corresponding wet body weight, W, for ovaries of 12 brick soldierfish, Myripristis amaena, from Johnston Atoll ( JA). The curve for the functional relationship fitted by regression is shown: F = 1.447 x 10~7 W50038. A large fraction of all food eaten by M. amaena and most other Myripristis species appears to be taken from the water column, often at some distance above the substrate. However, relatively little of what we found or of what has been reported in the diets of these species in Hawaii appears to be holoplanktonic. Among the many small crustacean groups identified in the diets, few copepods were found. A number of the common crustacean prey were from groups such as mysids and amphipods that migrate vertically (of- ten on a diel schedule) within shallow water, and may shelter on or near benthic substrate, within caves, cavities, rubble, or other cover during part of the day. Larval and young juvenile forms of larger benthic Crustacea, such as crabs, lobsters, stomatopods, and some shrimp, may be components of this migrating "semiplankton." Some shrimps may be intermittently sedentary or free swimming as adults. What is known of this semiplankton ( Alldredge and King, 1977; Por- ter and Porter, 1977; Parrish, 1989) and of the diet and feeding of Myripristis suggests that these squirrelfishes are not restricted to either planktonic or benthic feeding, but that they consume these prey groups wherever they are accessible. The dominance of this semiplanktonic, probably vertically migrating, fauna in the diet of these fishes has important implications for their trophic linkage to the surrounding systems. Whatever the spatial and temporal details of their residence in the water column, the dominant "zooplankton" seem charac- teristic of an inshore aggregation, probably tied closely to shallow water. Therefore, these squirrel- fishes depend for their trophic support largely on lo- cal sources of secondary production and possibly even primary nutrients (Parrish, 1989). This trophic ar- rangement is in contrast to the traditional concept of shallow-water planktivorous fishes supported by holoplankton of open ocean origin brought to the coast by prevailing oceanic currents. There are reports of some Myripristis species feed- ing close over the substrate (Brecknock, 1969; ter Kuile, 1989). Hiatt and Strasburg (1960) noted that M. microphthalmus (,=violaceus; Greenfield, 1974) in the Marshall Islands "takes a great variety of crus- taceans which are associated with, or swim near, the coral mounds in which this. ..fish secludes itself." They also found that some of their specimens had eaten tube-dwelling polychaetes — clearly a benthic Dee and Parrish: Reproductive and trophic ecology of Myripristis amaena 525 0 50 100 150 200 250 300 350 400 450 500 Whole wet body weight (g) Figure 7 Size-frequency distribution of population numbers and population egg production of the brick soldierfish, Myripristis amaena, at Johnston Atoll (JA). (A) Distribution of whole wet body weights of all 855 specimens measured from recreational catches and scientific collections combined. (B) Dis- tribution of wet weights of only the reproductively mature portion of the population up to300g(/!=396). (C) Distribution of egg production by wet weight of spawners ( n =396 ) based on the spawning population distribution in (B) and the fecundity vs body weight relationship developed from gonad samples. The fish population is arbitrarily truncated at 300 g for (B) and (C), which excludes 25 specimens distributed very irregularly over 18 weight classes, and with sizes much larger than those used to develop the size- fecundity relationship. prey that was common in adjacent sandy patches. Harmelin-Vivien (1979) reported that polychaetes were the major prey by weight found in M. bowditchae (=murdjan\ Randall and Gueze, 1981) in Madagascar, and she specifically noted that this fish sometimes fed on the bottom. Polychaetes were a minor prey of M. berndti at Oahu (Brecknock, 1969) and were fairly abundant in M. murdjan in the Flores Sea (ter Kuile, 1989). A polychaete was found in a M. amaena specimen from Kona (Hobson, 1974) and in one from our collections at Puako. In our JA col- lections, they were much more common and abundant, producing the second largest IRI value of the major systematic groups (Table 3). The JAtaxa were Seden- taria, which must certainly have been benthic dwell- ers. At Puako, at least five of our 36 specimens of M. berndti and two of our 21 specimens of M. kuntee with prey contained polychaetes. Three specimens of M. berndti contained several polychaete individuals each. Gastropods (presumably benthic species) were re- ported by Hobson (1974) in the diets of M. amaena and M. berndti from Kona. A benthic gastropod was found in the gut of a M. berndti at Puako, and they were fairly common in our M. amaena specimens there. Brecknock (1969) found a few gastropods in guts of M. berndti from Oahu and reported foraging by this species on the bottom of aquaria. One M. berndti specimen from Puako contained part of an arm of a bottom-living ophiuroid. It seems clear that M. amaena and some other Myripristis species eat some fully benthic taxa. It is unlikely that such prey make up a major part of the diet for more than a very few Myripristis species. However, the ability of these squirrelfishes to employ this feeding mode enables them to exploit a greater range of food resources, at least on an opportunistic basis. Again, the trophic source seems to be local. Trophic role of squirrelfishes in tropical communities Holocentrids commonly make up a significant portion of the total natural fish community and are important predators throughout their range. In the uninhabited NWHI, nine species of holocentrids, including M. amaena, made up 4.5% of all individuals in the fish community (Norris and Parrish, 1988). In Tulear, Mada- gascar, three species of holocentrids represented 2.2% of the total fish population (Harmelin-Vivien, 1981). At JA, M. amaena alone provided about 2% of all indi- viduals in the fish community. Together with three other species of holocentrids, it accounted for about 3% of the total fish community there (Dee et al.2). 2 Dee, A. J., D. K. Irons, and J. D. Parrish. 1985. Johnston Atoll re- source survey; a final report of the initial phase ( 19 Jan 1984-20 Jul 1985). Project report to U.S. Army Engineer District, Honolulu, 70 p. 526 Fishery Bulletin 92(3), 1994 The importance of holocentrids as predators has been well documented (Randall, 1967; Hobson, 1974; Vivien and Peyrot-Clausade, 1974; Gladfelter and Johnson, 1983). Gladfelter and Johnson (1983) found that seven species of squirrelfishes made up >99% of the nocturnally active, benthic crustacean-feeding fishes at St. Croix, U.S. Virgin Islands. Randall ( 1967) reported that holocentrids accounted for about 14% by number and 11% by weight of all zooplankton con- sumed as prey by the fish community at St. John, U.S. Virgin Islands. In the NWHI, holocentrids, in- cluding M. amaena, are among the most successful families offish predators. Holocentrids accounted for nearly 40% by number, over 60% by weight, and about 50% by volume of the large crustacean com- munity (crabs, shrimps, stomatopods, and lobsters) taken as prey by the 78 fish species from 28 families that contained large crustacean prey in our NWHI diet studies. Holocentrids also were responsible for about 2.5% of all the individual fish eaten. The frac- tion of the complete food consumption (all prey in the community combined) by this entire fish com- munity that is eaten by holocentrids was about 13- 17% (Parrish, unpubl. data). Holocentrids also are an important element of the community as prey for other fishes. In the NWHI, 4% of all identified fish prey individuals were holocentrids (Norris and Parrish, 1988). In the west- ern Atlantic, Randall (1967) found evidence that seven species of fishes from four families had eaten holocentrids; three species from three families had eaten Myripristinae iMyripristis jacobus). Dragovich ( 1970) also found that postlarval holocentrids (includ- ing Myripristinae) were fairly common prey of skip- jack tuna, Katsuwonus pelamis, and yellowfin tuna, Thunnus albacares, in the western Atlantic. As a widespread and abundant group that is an active predator and vulnerable prey, holocentrids play a major role in the trophic structure of tropical ma- rine ecosystems. Reproduction For specimens from JA, the results from the three independent analyses of gonads (histology, GSI, and visual examination) indicated that sexual maturity of M. amaena occurs at 153-156 mm SL for females and at 149-156 mm SL for males. These results cor- respond closely to SFR estimates from our specimens collected at Puako: 145-160 mm SL, sexes combined (Hayes et al., 1982). Data from both locations are included in Figures 1 and 2. These values of SFR correspond to about 75-80% of hm (as determined by fitting data from length measurements and otolith increment counts to a von Bertalanffy growth model ) and to an age of about six years (Fig. 8 in Dee and Radtke, 1989). Dee and Radtke (1989) aged speci- mens up to nearly 14 years old. Their oldest speci- men (of many available for analysis) was somewhat larger than the Lm derived from the regression, so it seems unlikely that many individuals live much longer. Therefore, the age at first reproduction (AFR) is probably about 40% (or a little less) of the maxi- mum lifespan commonly attained, and some individu- als may reproduce for as many as eight years. The relation between SFR and maximum body size has been investigated by several workers in a num- ber of locations. The only results reported for squirrelfishes are estimates, based on large sample sizes, for two species of Holocentrinae from the Car- ibbean Sea (Wyatt, 1976). For Holocentrus adscen- sionis, FL at sexual maturity was about 175 mm, asymptotic (maximum) FL about 265 mm, and the ratio about 0.66; for Holocentrus rufus, FL at sexual maturity was about 130-135 mm, asymptotic FL about 230 mm, and the ratio about 0.59. Both these species reach considerably larger sizes than M. amaena, and M. amaena has the largest SFR/L^ ra- tio of the three species. The ratios for these squirrelfishes seem to be in the high portion of the range of published values for tropical fishes (Munro, 1974; Loubens, 1980). Myripristis amaena, in par- ticular, matures at an advanced absolute age and at a surprisingly large fraction of its maximum age and size. Spawning of M. amaena at JA seems to occur pri- marily in April-May; a secondary peak probably takes place in late September. All specimens collected during the fall peak showed GSI values above the inactive (off-season ) level, but considerably below the mean value for the spring peak (Fig. 5). Although no collections were possible in September, visual exami- nation and GSI data from collections made throughout October 1985 suggested the late stages of a spawning period that probably peaked in late September. Back calculation using the total number of otolith incre- ments counted for the two smallest individuals aged by scanning electron microscope examination (Dee and Radtke, 1989) indicated that one individual was spawned in late September and the other in early October. A spawning peak also was observed in speci- mens collected in January 1986, but not in January 1985. The 1986 event may have resulted from un- seasonably calm conditions that occurred during that period. Spawning also was recorded during January 1986 for Chaetodon trifascialis, an unusual time of year for that species. Values of the GSI for M. amaena col- lected in January 1986 were generally as high as those of specimens collected during the April spawning peak. Data were not collected at Puako in a way that would permit a comparable assessment of seasonal Dee and Parrish: Reproductive and trophic ecology of Mynpristis amaena 527 distribution of reproduction. Instead, small collec- tions of specimens were made at regular, frequent intervals only long enough to discern a time of ac- tive reproduction. At only that time, a large collec- tion of specimens was made quickly to permit esti- mation of SFR, and seasonal collections were not continued (Hayes et al., 1982). Thus, the Puako re- sults serve only to establish that reproductive devel- opment (e.g. GSI, Fig. 5) is high in May and June. Those months probably represent a peak, but the data do not well define its limits or the pattern for the rest of the year. This early summer high at Puako is contiguous with the late spring high at JA. Many tropical marine species show a collective spring spawning peak and a second peak in fall (Munro et al., 1973; Watson and Leis, 1974; Johannes, 1978; Walsh, 1987). For Hawaiian fishes, the most dominant seasonal spawning pattern, based on numbers of spawning records, is a peak spawn- ing period in about April and May, with a secondary peak in October for some species. Based on numbers of recruitment records, the dominant recruitment period occurs in June and July, and a secondary peak in February and March (Walsh, 1987). Many studies indicate that there can be considerable variability in the timing of recruitment from year to year, and that the timing and intensity may vary at small spa- tial scales (Victor, 1982; Williams, 1983; Sale, 1985; Schroeder, 1985; Walsh, 1987, Doherty, 1991). Larval and newly settled M. amaena were elusive throughout the present study, and few data could be collected regarding recruitment. The youngest speci- men for which we counted short period (apparently daily) increments in otoliths (Dee and Radtke, 1989) showed a discontinuity that probably represented settlement from the plankton. Back calculation based on the number of increments after the assumed settlement mark suggests that the specimen settled in early February. Although our data regarding settlement are minimal, both these and our spawn- ing results are consistent with the above seasonal reproductive periods summarized by Walsh (1987). Walsh suggested that changes in water temperature or photoperiod, or both, are most likely responsible for observed seasonal patterns of spawning and re- cruitment in Hawaiian reef fishes. For M. amaena at JA, there was no indication that water tempera- ture affected the time of spawning. The full annual range of temperature is very small (24.5-26.5 C). Tem- peratures during the reproductive period of January 1986 were among the coldest recorded during the en- tire study, whereas temperatures during spawning in April 1984 were among the highest recorded. Wyatt ( 1976) reported on spawning seasons of two Holocentrinae species in Jamaican waters. He re- corded ripe Holocentrus adscensionis collected in all months except June, and only 2% ( one specimen ) ripe in July. Most spawning occurred in January, Febru- ary, and March, but October was also a peak month. Besides ripe fish, "sexually active" gonads were com- mon (14-37% of all gonads) in September through May. The seasonal pattern was similar for//, rufus; highest peaks of ripe gonads occurred in October (44%) and February (32%). "Sexually active" gonads were found in all months except July. May, June, and July were the months of lowest gonad development. In Bermuda ( 15° farther north in mid-ocean), Winn et al. (1964) reported breeding by both these Holocentrinae species in June, July, and August. Variability in timing of spawning due to factors such as lunar periodicity, water temperature, plankton productivity, photoperiod, currents, and rainfall oc- curs commonly, and spawning time can vary from year to year, even at the same location ( Watson and Leis, 1974; Wyrtki, 1974; Johannes, 1978; Walsh, 1987). The fecundity of M. amaena is relatively low com- pared with many marine species. The most fecund specimen examined contained fewer than 70,000 maturation stage eggs, and the length-fecundity func- tion predicts that a specimen of length LM would con- tain fewer than 100,000 such eggs. Fecundity in- creases sharply with body size; it rises with the fifth power of weight and more than the tenth power of SL. (The sample size used for the regressions was not large, but the values of r2 indicate a reasonably good fit.) These changes with size are much greater than those found in many marine species. These re- sults, together with the results for SFR and the old- est specimen aged, indicate that the species matures slowly. With a relatively long life and steeply increas- ing fecundity, a very large fraction of the reproduc- tive output of the population is provided by old fish. The number of spawnings per year is unknown. There was no clear evidence in ovaries examined under direct light microscopy or histologically of a distinct series of distinguishable groups of ova in a graded-size sequence that might represent serial batches spawned within a season. However, produc- tion of such serial clutches is commonplace in tropi- cal, nearshore fishes. Unless individual females spawn a good many batches (of the full number of maturation stage eggs estimated) within each year, the total or lifetime fecundity of individuals is rela- tively low compared with many common marine spe- cies. For example, an individual maturing at age six and spawning once annually through age 14, accord- ing to the body sizes indicated by our von Bertalanffy expression (Dee and Radtke, 1989) and the fecun- dity indicated by our length-fecundity expression, would produce fewer than 300,000 eggs during such 528 Fishery Bulletin 92(3). 1994 a reproductive life. Clearly, determination of the number of clutches spawned in a reproductive sea- son is an important subject for research on M. amaena. Most types of fisheries tend to produce higher mortality of older age classes. Certainly this is true of the common spear fisheries for squirrelfish in the tropics, where larger specimens are individually se- lected. For a species in which sexual maturity and fecundity are related to age (size) as they are in M. amaena, this means that substantial fishing pres- sure applied to the population can severely reduce egg production (see Fig. 7). The risks of recruitment overfishing are therefore especially great, particu- larly when maturity occurs late enough that fish of prereproductive size are still a desirable catch. The expected trend as fishing pressure on such a stock increases is the appearance of an increasing fraction of prereproductive fish in the catch. At JA, despite low total fishing effort, about 51% of all M. amaena in our creel sample of the fishery were prere- productive; in the small-scale, recreational-subsis- tence fishery at Puako, about 46% were prere- productive in =24). Near many centers of human population, fishing intensity is much greater, and great declines in populations of soldierfish are unof- ficially reported. Clearly the life history of M. amaena, and perhaps soldierfishes in general, cre- ates high vulnerability to conventional, unregulated fishing. It seems essential that fisheries for such spe- cies be managed to conserve the largest, oldest spawners to protect the reproductive potential of the stock. Acknowledgments This research received major funding support from the Department of the Army, U.S. Army Engineer- District, Fort Shafter, Hawaii, under contract DACA83-84-C-0019. The work at Puako was sup- ported by the Hawaii Department of Land and Natu- ral Resources in collaboration with the U.S. Fish and Wildlife Service, and by a grant from NOAA Office of Sea Grant through the University of Hawaii Sea Grant College Program. The work in the Northwest- ern Hawaiian Islands was supported similarly in part by Univ. Hawaii Sea Grant under grants from NOAA Office of Sea Grant, Department of Commerce, by Project NI/R-4 and Institutional Grants NA79AA-D- 00085 and NA81AA-D-00070. This is Sea Grant Pub- lication UNIHI-SEAGRANT-JC-94-15. The Ocean Resources Branch, State of Hawaii, Department of Business, Economic Development and Tourism, con- tributed additional funding. Financial and logistic support also were provided by the Hawaii Coopera- tive Fishery Research Unit, which is jointly sup- ported by the Hawaii Department of Land and Natu- ral Resources, the University of Hawaii, and the National Biological Survey. A number of Fishery Unit research assistants collected much of the data and assisted with the early stages of analysis for Puako and the NWHI. The Defense Nuclear Agency and the staff of its contractor, Holmes and Narver Services Inc., provided invaluable support with facilities, lo- gistics, and technical expertise on Johnston Atoll. The Pacific Islands Office of the U.S. Fish and Wildlife Service provided access to the National Wildlife Ref- uge property and collaborated in the overall program of research at Johnston Atoll. We are grateful for the help and expertise of Jim Howard, Darby Irons, Ben Leorna, Chuck Madenjian, Charley Myers, Tim Tricas, and Brian Tsukimura. The paper was con- siderably improved based on review and comments from T.A. Clarke, E.E. DeMartini, D.W. Greenfield, and E.S. Reese. Susan Monden prepared the figures. Literature cited Alldredge, A. L., and J. M. King. 1977. Distribution, abundance, and substrate pref- erences of demersal reef zooplankton at Lizard Is- land Lagoon, Great Barrier Reef. Mar. Biol. ( Berl. ) 41:317-333. Amerson, A. B., and P. C. Shelton. 1976. The natural history of Johnston Atoll, central Pacific Ocean. Atoll Res. Bull. 192:1-479. Brecknock, E. L. 1969. Some aspects of the ecology of Myripristis murdjan in the Hawaiian Islands. M.S. thesis, Univ. Hawaii, Honolulu, 22 p. Dee, A. J., and R. L. Radtke. 1989. Age and growth of the brick soldierfish, Myripristis amaena. Coral Reefs 8:79-85. Doherty, P. J. 1991. Spatial and temporal patterns in recruit- ment. In P. F. Sale led.), The ecology of fishes on coral reefs. Academic Press, San Diego, 754 p. Dragovich, A. 1970. The food of the skipjack and yellowfin tunas in the Atlantic Ocean. U.S. Fish Wildl. Serv., Fish. Bull. 68:445-460. Gladfelter, W. B., and W. S. Johnson. 1983. Feeding niche separation in a guild of tropi- cal reef fishes (Holocentridae). Ecology 64:552-563. Greenfield, D. W. 1965. Systematics and zoogeography of Myripristis in the eastern tropical Pacific. Calif. Fish Game 51:229-247. 1968. The zoogeography of Myripristis I Pisces: Holocentridae). Syst. Zool. 17:76-87. Dee and Parrish: Reproductive and trophic ecology of Myripristis amaena 529 1974. A revision of the squirrelfish genus Myri- pristis Cuvier (Pisces: Holocentridae). Los Angel. Cty. Mus. Sci. Bull. 19:1-54. Guyer, M. F. 1953. Animal micrology, 5th ed. Univ. Chicago, 327 p. Halstead, B. W., and N. C. Bunker. 1954. A survey of the poisonous fishes of Johnston Island. Zoologica(N.Y.) 39:61-77. Harmelin- Vivien, M. L. 1979. Ichtyofaune des recifs coralliens de Tulear (Madagascar): ecologie et relations trophiques. Annexe D.Sc. these, Univ. Aix-Marseille II, 116 p. [In French.] 1981. Trophic relationships of reef fishes in Tulear (Madagascar). Oceanol. Acta 4:365-374. Hayes, T. A., T. F. Hourigan, S. C. Jazwinski Jr., S. R. Johnson, J. D. Parrish, and D. J. Walsh. 1982. The coastal resources, fisheries and fishery ecology of Puako, West Hawaii. Hawaii Coopera- tive Fishery Research Unit Tech. Rep. 82-1, 159 p. Hiatt, R. W., and D. W. Strasburg. 1960. Ecological relationships of the fish fauna on coral reefs of the Marshall Islands. Ecol. Monogr. 30:65-127. Hobson, E. S. 1972. Activity of Hawaiian reef fishes during the evening and morning transitions between daylight and darkness. Fish. Bull. 70:715-740. 1974. Feeding relationships of teleostean fishes on coral reefs in Kona, Hawaii. Fish. Bull. 72:915-1031. II u mason, G. L. 1979. Animal tissue techniques, 4th ed. Freeman, San Francisco, 661 p. Johannes, R. E. 1978. Reproductive strategies of coastal marine fishes in the tropics. Environ. Biol. Fishes 3:65-84. Khoo, H. K. 1978. The histochemistry and endocrine control of vitellogenesis in goldfish ovaries. Can. J. Zool. 57:617-626. Loubens, G. 1980. Biologie de quelques especes de Poissons du lagon neo-caledonien. II: Sexualite et repro- duction. Cahiers de l'lndo-pacifique 2(11:41-72. [In French.] Maragos, J. E., and P. L. Jokiel. 1986. Reef corals of Johnston Atoll: one of the world's most isolated reefs. Coral Reefs 4:141-150. Munro, J. L. 1974. The biology, ecology, exploitation and manage- ment of Caribbean reef fishes. Part V.m: Summary of biological and ecological data pertaining to Car- ibbean reef fishes. ODA/UWI Fisheries Ecology Re- search Project: 1969-1973. Research report from the Zoology Dept., Univ. West Indies. No. 3. Kingston, Jamaica, 24 p. Munro, J. L., V. C. Gaut, R. Thompson, and P. H. Reeson. 1973. The spawning seasons of Caribbean reef fishes. J. Fish Biol. 5:69-84. Norris, J. E., and J. D. Parrish. 1988. Predator-prey relationships among fishes in pristine coral reef communities. In J. H. Choat et al. (eds.), Proc. sixth int. coral reef symposium, Townsville, Australia, 8-12 Aug. 1988, vol. 2, p.107- 113. Sixth Int. Coral Reef Symp. Exec. Comm. Oda, D. K., and J. D. Parrish. 1982. Ecology of commercial snappers and groupers introduced to Hawaiian reefs. In E. D. Gomez, C. E. Birkeland, R. W. Buddemeier, R. E. Johannes, J. A. Marsh Jr., and R. T. Tsuda (eds.), Proc. fourth int. coral reef symposium, Manila, 18-22 May 1981, vol. 1, p. 59-67. Marine Sciences Center, Univ. Philippines, Quezon City, Philippines. Parrish, J. D. 1989. Fish communities of interacting shallow-wa- ter habitats in tropical oceanic regions. Mar. Ecol. Prog. Ser. 58:143-160. Parrish, J. D., M. W. Callahan, and J. E. Norris. 1985. Fish trophic relationships that structure reef communities. In C. Gabrie and B. Salvat (eds.), Proc. fifth int. coral reef congress, Tahiti, 27 May- 1 June 1985, vol. 4, p. 73-78. Antenne Museum- EPHE, Moorea, French Polynesia. Pinkas, L., M. S. Oliphant, and I. Iverson. 1971. Food habits of albacore, bluefin tuna, and bonito in California waters. Calif. Dep. Fish Game, Fish Bull. 152:1-105. Porter, J. W., and K. G. Porter. 1977. Quantitative sampling of demersal plankton migrating from different coral reef substrates. Limnol. Oceanogr. 22: 553-556. Randall, J. E. 1967. Food habits of reef fishes of the West Indies. Stud. Trop. Oceanogr. (Miami) 5:665-847. Randall, J. E., and P. Gueze. 1981. The holocentrid fishes of the genus Myripristis of the Red Sea, with clarification of the murdjan and hexcigonus complexes. Los Angel. Cty. Mus. Contrib. Sci. 334, 16 p. Randall, J. E., P. S. Lobel, and E. H. Chave. 1985. Annotated checklist of the fishes of Johnston Island. Pac. Sci. 39:24-80. Sale, P. F. 1985. Patterns of recruitment in coral reef fishes. In C. Gabrie and B. Salvat (eds.), Proc. fifth int. coral reef congress, Tahiti, 27 May-1 June 1985, vol. 5, p. 391-396. Antenne Museum-EPHE, Moorea, French Polynesia. Schroeder, R. E. 1985. Recruitment rate patterns of coral reef fishes at Midway lagoon. Northwestern Hawaiian Islands. In C. Gabrie and B. Salvat (eds.), Proc. fifth int. coral reef congress, Tahiti, 27 May-1 June 1985, vol. 5, p. 379-384. Antenne Museum-EPHE, Moorea, French Polynesia. ter Kuile, C. 1989. The forage base of some reef fishes in the Flores Sea with notes on sampling and fishery. Neth. J. Sea Res. 23:171-179. 530 Fishery Bulletin 92(3), 1994 Victor, B. C. 1982. Daily otolith increments and recruitment in two coral-reef wrasses, Thalassoma bifasciatum and Halichoeres bivittatus. Mar. Biol. (Berl.) 71:203-208. Vivien, M. L., and M. Peyrot-Clausade. 1974. A comparative study of the feeding behaviour of three coral reef fishes (Holocentridae), with spe- cial reference to the polychaetes of the reef cryptofauna as prey. In A. M. Cameron, B. M. Campbell, A. B. Cribb, R. Endean, J. S. Jell, 0, A. Jones, P. Mather, and F. H. Talbot (eds.), Proc. sec- ond int. symp. coral reefs, Great Barrier Reef Prov- ince, Australia, 22 June-2 July 1973, vol. 1, p. 179- 192. The Great Barrier Reef Committee, Brisbane. Wallace, R. A., and K. Selman. 1981. Cellular and dynamic aspects of oocyte growth in teleosts. Am. Zool. 21:325-343. Walsh, W. J. 1987. Patterns of recruitment and spawning in Hawaiian reef fishes. Environ. Biol. Fishes 18:257-276. Watson, W., and J. M. Leis. 1974. Ichthyoplankton in Kaneohe Bay, Hawaii. A one year study offish eggs and larvae. Univ. Ha- waii Sea Grant Technical Report TR-75-0 1 , 1 78 p. Williams, D. M. 1983. Daily, monthly and yearly variation in recruit- ment of juvenile coral reef fishes to coral habitats within One Tree Lagoon, Great Barrier Reef. Mar. Biol. (Berl.) 65:245-253. Winn, H. E., J. A. Marshall, and B. Hazlett. 1964. Behavior, diel activities, and stimuli that elicit sound production and reactions to sounds in the longspine squirrelfish. Copeia 1964:413-425. Wolfert, D. R., and T. J. Miller. 1978. Age, growth, and food of northern pike in east- ern Lake Ontario. Trans. Am. Fish. Soc. 107:696- 702. Wyatt, J. R. 1976. The biology, ecology, exploitation and manage- ment of Caribbean reef fishes. Va: The biology, ecol- ogy and bionomics of Caribbean reef fishes: Holo- centridae (squirrelfishes). ODA/UWI Fisheries Ecology Research Project: 1969-1973. Research report from the Zoology Dep., Univ. West Indies. No. 3. Kingston, Jamaica. 41 p. Wyrtki, K. 1974. Sea level and seasonal fluctuations of the equatorial currents in the western Pacific Ocean. J. Physical Oceanogr. 4:91-103. Appendix A Relations hips between standard 1 ength (SL) fork length (FL), and total length (TL ) in millimeters, based on linear regressions for 377 specimens of the brick sole ierfish, Myripristis amaena, from Job. iston Atoll (JA ). SL = 0.9013 FL- 3.22 . r- = 0.991 SL = 0.7811 TL- 3.23 , r2 = 0.992 TL = 1.1493 FL + 0.65 , r- = 0.991 Appendix B Size measured and batch fecundity estimated by counting ova in ah quots from the ovaries of 12 speci- mens of the brick soldierfish, My rip -istis amaena, from Johnston Atoll ( JA). Standard Whole body Estimated body length (mm) wet weight (g) batch fecundity 156 149 12402 166 162 12925 168 167 17077 169 166 17618 175 L89 29505 175 1ST 57200 177 191 47680 179 190 42003 180 189 25301 181 225 69221 187 209 61000 188 211 66719 Abstract. — Callorhynchus cal- lorhynchus Linnaeus ("cockfish" or "pez gallo"), the only holocephalan fish species found in the coastal waters of the southwestern Atlan- tic Ocean, has been caught off Ar- gentina as bycatch of the hake, Merluccius hubbsi, fishery since 1920. Here we describe the mor- phology of its reproductive system and report on several aspects of its reproductive biology. This study is based on survey data and sampling of commercial landings from San Matias Gulf (41-42°S; 64-65°W), conducted from 1984 to 1986. The data suggest that reproduc- tive activities extend nearly throughout the year; mating and spawning occur in spring and early summer, followed by a short period (ca. one month I of gonadal recov- ery. Average size at sexual matu- rity (standard length, measured from the tip of the mouth to the origin of the superior caudal lobe) is 40 cm for males and 49 cm for females. Male gonadal and green gland indices peak asynchronously. During the mating season the green gland forms spermatophores that are transferred to the female at the time of copulation. The cloaca of adult females has a semi- nal receptacle where the mass of spermatophores is stored after copulation. Female gonadal and nidamental gland indices peak syn- chronously. After fertilization the oocytes are encapsulated before spawning. Oocyte diameter in- creases with the size of females up to a maximum of 48 mm. Reproductive biology of the cockfish, Callorhynchus callorhynchus (Holocephali: Callorhynchidae), in Patagonian waters (Argentina) Edgardo E. Di Giacomo Maria Raquel Perier Institute) de Biologia Manna y Pesquera "Alte Storm" CC 1 04, 8520 San Antonio Oeste, Rio Negro, Argentina Manuscript accepted 26 January 1994. Fishery Bulletin 92: 531-539 ( 1994). The cockfish, Callorhynchus callor- hynchus Linnaeus, 1758, is the only holocephalan fish species found in the coastal waters of the southwest- ern Atlantic Ocean (Norman, 1937; Hart, 1946; Menni and Gostonyi, 1982 ). Adults are caught as bycatch of the hake, Merluccius hubbsi, fish- ery that operates off Argentina (Di Giacomo and Perier, 1991). Al- though there is considerable con- cern with regard to harvesting fish species with comparatively low re- productive potential ( such as many elasmobranch and holocephalan species), the reproductive biology and life history of the cockfish are poorly known. In this study we de- scribe the morphology of the repro- ductive system of C. callorhynchus and present information on its re- productive biology. Materials and methods Bottom trawl surveys A survey off the north coast of San Matias Gulf (northern Patagonia; 41-42°S, 64-65°W; Fig. 1) was con- ducted from 15 to 19 October, 1986, aboard the FRV C. Cdnepa. Thirty three 30-minute hauls were made at depths ranging from 20 m to 130 m, by using a 96-foot commercial otter trawl with rectangular doors (Di Giacomo and Perier, 1991). All specimens of C. callorhynchus were processed following laboratory pro- cedures described below. Sampling of the commercial landings The commercial catch of the bottom trawl fishery landed in San Anto- nio Oeste (40°43'04"S, 64°56W) was sampled monthly between Febru- ary 1984 and July 1986. On each sampling date three boxes (N=2Q fish per box) weighing 37 kg each were randomly selected. Samples of the commercial catch were not available during April, September, June, and December because either the fishery was inactive or no cockfish were landed. Laboratory procedures Standard length (SL, distance from the tip of the snout to the origin of the upper caudal lobe, Fig. 2A), to- tal weight and liver weight were obtained from each fish sampled. Specimens were dissected to expose the reproductive system; testes and green glands of males, and ovaries and nidamental glands of females were individually weighed. The number of mature yolk oocytes and immature oocytes, maximum diam- 531 532 Fishery Bulletin 92(3). 1994 eter of the oocytes, number of individuals with egg cases in formation or ready for extrusion, and presence of spermatophores were recorded for females. Length of the myxopterygia (males) was measured from the insertion of the pelvic fin to the distal end (Fig. 2B). Figure 1 (A) Location of the study area in the Argentinean continental shelf; (B) San Matias Gulf. Indices of reproductive activity The annual reproductive cycle was assessed by us- ing the monthly samples of the commercial catch ( 1984-86). The gonadal index (GI), hepatic index (HI), nidamental gland index (NGI, females only ) and green gland index (GGI, males only) were calculated as GI: HI weight of the gonad total weight x 100 weight of the liver total weight x 100 NGI weight of the nidamental gland total weight x 100 _„ weight of the green gland GGI = E x 100 . total weight The duration of the mating season was estimated from the presence of spermatophore masses in the female cloaca and the annual cycle of the green gland index. Regression analysis was performed between the indices of reproductive activity during the periods of maximum activity and SL, and between the diam- eter of mature oocytes and the number of immature oocytes and SL. A Wilcoxon test was used to assess the difference in the average number of mature ova between right and left ovaries of all mature females. Sexual maturity The degree of development of the myxopterygia (male claspers ) relative to SL was used as an indicator of sexual maturity for males (Holden and Raitt, 1975). The size at first maturity in females was determined by examining 1) the percentage of mature females in each 10-mm size class, 2) the maxi- mum diameter of mature ovarian eggs for each size class, and 3) the smallest size class with egg cases in formation. Females with yellow ovarian eggs were considered mature. Figure 2 irements for cockfish, Callorhynchus callorhynchus. (A) Standard length; (B) myxopterygia length in male. Results Morphology of the reproductive system The male reproductive system con- sists of the following paired struc- Di Giacomo and Perier: Reproductive biology of Callorhynchus callorhynchus 533 tures: testes, epididymis (efferent ducts), and defer- ent ducts (Leigh-Sharpe, 1922). In sexually mature individuals, the deferent ducts are differentiated into two fusiform structures called green glands (Fig. 3A). These structures agglutinate the spermatozoa into spermatophores and secrete a gelatinous, green fluid. The reproductive system of the female is composed of the following paired structures: ovaries, oviducts, nidamental or shell glands, uteri, and vaginal open- ings that end in a cloaca (Fig. 3B). Holocephalans are sexually dimorphic. Male sec- ondary sexual structures (frontal tenaculum, and prepelvic and pelvic claspers) are utilized during mating (Fig. 4); the male apparently inserts the mass of spermatophores into the seminal receptacle of the female cloaca. Upon macroscopic examination (Oc- tober), a mass of spermatophores was found at the female genital opening. Spermatozoa are liberated progressively by an unknown mechanism, before fer- tilization. Females extrude fertilized eggs enclosed within a leathery case (Dean, 1906) that acts as protection for the embryo during development. At hatching the fry resembles the adult. Indices of reproductive activity The GI of males reached a maximum in March (3.0) and was low from July to October (range seen in Fig. 5A). The highest mean values of the GGI were found between July and October (Fig. 5B), although the index was also high during January and February. In females, the GI reached its highest values from July to October, indicating low reproductive activity during the rest of the year (Fig. 6A). The NGI (fe- males) has two periods of activity: July to November (highest) and January to June (lowest) (Fig. 6B). The HI for both males and females (Fig. 5C, 6C) did not show significant annual fluctuations; maximum val- ues occurred simultaneously with peaks in the other reproductive indices. Mature and immature females could be differenti- ated based on GI values (range: 0.13 to 4.55); a GI value larger than 1.0 was characteristic of mature females. The presence in March of some females smaller than 50 cm SL with a GI higher than 1.0 and ovaries with translucent oocytes was suggestive of oocyte resorption. Oocyte diameter varied between 10 and 35 mm. testes epididymis ovaries genital opening Figure 3 Reproductive organs of the cockfish, Callorhynchus callorhynchus. (A) Males; (B) females. 534 Fishery Bulletin 92(3). 1994 frontal tenaculum prepelvic claspers lasper Figure 4 Male secondary sexual structures of cockfish, Callor- hynchus callorhynchus. (A) Frontal tenaculum; (B) prepelvic and pelvic claspers. Gonad and green gland indices were highly vari- able within size classes during the period of maxi- mum activity. There was no relationship between SL and GI (r2=0.001;P>0.05; 2V=121), or between SL and GGI (^=0.021; P>0.05;iV=101). Similarly, no relation- ship was found between female SL and GI (r2=0.18; P>0.05; N=58) or NGI (r2=0.012; P>0.05; N=48). Size at sexual maturity Immature and mature males could be identified from the relationship between SL and length of the myxopterygia. The myxopterygia did not exceed the posterior margin of the pelvic fin in immature indi- viduals, whereas in mature individuals it consisted of partially calcified structures that exceeded the pos- terior margin of the pelvic fin. The length offish mea- sured l./V=123) ranged from 26 cm to 55 cm. Length at first maturity of males was estimated to be 40 cm, corresponding to a length of the myxopterygia of 45 mm. No individuals between juvenile and mature stages were observed. Juvenile stages were found in shallow waters (depth range 15-25 m) whereas adults were found in waters deeper than 25 m (Di Giacomo, 1992). Most mature females were larger than 50 cm (SL), whereas all females smaller than 48 cm were immature. Ovarian maturation The percentage of mature oocytes was greatest in August in both ovaries, coinciding with maxima in the indices of reproductive activity. Nevertheless, some mature oocytes were observed throughout the year (Fig. 7A). Mean number of mature oocytes was highest from July to October. There was a signifi- cant difference between the average number of ma- ture oocytes in the right and left ovaries of individual mature females (Wilcoxon test, P<0.05; Fig. 7B). Oocyte diameter increased with female size (r2=0.083;P<0.01;A^=90). Maximum oocyte diameter (48 mm) was recorded for a female of 71 cm SL. The regression of number of immature oocytes on SL showed a similar trend, indicating that average reproductive potential increases with size (?~=0.024;P<0.01; Af=101). Oocytes smaller than 10 mm (diameter) were translucent to lightly whitish. Coloration of oocytes larger than 10 mm changed gradually. Yolk was in- corporated until the oocytes reached complete devel- opment and passed to the nidamental gland. Pro- duction of egg cases began in the nidamental gland, when the oocytes reached maturity. The closed end of the egg case was caudally oriented; the cephalic portion remained open until the case was completely formed. When the egg case was about two-thirds of its final size, ovulation occurred and the egg (with a variable diameter that ranged from 40 to 48 mm ) moved from the ovaries, through the oviduct, to the case. Mating season Spermatophores were found in the vaginal recep- tacles associated with the female cloaca. Mating oc- curred primarily from July to February, when maxima in the GGI were indicative of active produc- tion of spermatophores. The highest percentage of females with spermatophores in the cloaca was also observed during these months (Fig. 8). In October, recently deposited spermatophores in the female cloaca, characterized macroscopically by an intense green color, resembled green gland products. Yellow- ish spermatophores found in March and May (when Di Giacomo and Perier: Reproductive biology of Callorhynchus callorhynchus 535 5 4.5 4 3 5 3 2.5 2 1.5 1 0.5 0 1 A 5 4 5 JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV 0 5 0 B JAN FEB MAR APR MAr JUN JUL AUG SEP OCT NOV 7- c _^ 6 - ~5- E 4- 150 m), which was subdivided into 20-m sampling strata. Later in the season, we subdivided the upper water column (<90 m) into 15-m sampling strata. Samples were fixed in 4-5% formalin and shipped to the Polish Plankton Sorting Center in Szczecin, Poland, where fish eggs and larvae were separated, and walleye pollock eggs and larvae were identified and counted. Stage of development of eggs was de- termined in each tow taken early in the season ac- cording to Blood et al. (1994). When more than 100 eggs were present in a sample, a subsample of 100 was staged. The egg stage data then were compressed into six stage groups as in Kendall and Kim ( 1989). Standard length (SLXtoO.l mm) of the larvae in each sample was measured. A subsample of 50 larvae was measured when more than 50 larvae were present in a sample. Catches of eggs and larvae per depth interval are reported as numbers per 1,000 m3 of water based on volume filtered as determined from digital flowmeter records and net-frame angle. Mean and standard deviations (SD) of numbers per 1,000 m3 for each tow were computed from the weighted aver- age of the numbers per 1,000 m3 from each net within the tow, by using the length of the depth interval for each net as the weight. Estimation of egg and larval mean depth, larval mean length, and their standard deviations is based on cluster sampling where each tow represents a cluster, each net subsamples eggs and larvae from the cluster, and each egg or larva is an element within the cluster ( Equations 8. 1 and 8.2 in Scheaffer et al., 1986). The observation associated 542 Fishery Bulletin 92(3). 1994 with each egg or larva, used to estimate mean depth, is the depth of the net where it was collected. Addi- tionally, observations of lengths of larvae in each net were used to estimate mean lengths. Gaussian sta- tistics, including analysis of variance (ANOVA) and analysis of covariance (ANCOVA), were used to com- pare mean depths of eggs and larvae and stages of eggs, and lengths of larvae with various environmental vari- ables: time of day, temperature, salinity, and depth. Temperature and salinity data from vertical Sea- bird conductivity, temperature, and depth (CTD) casts made in conjunction with the MOCNESS tows in 1986 were processed at the Pacific Marine Envi- ronmental Laboratory, Seattle. CTD data were col- lected with Seabird sensors on the MOCNESS in 1987 and 1988 and were processed at the Atlantic Oceanographic and Meteorological Laboratory, Mi- ami. CTD data were available at 1-m depth inter- vals. In comparing temperature and density (ex- pressed as ot) with depths of occurrence of eggs and larvae among tows, the mean and standard devia- tion of temperature and density were calculated within the depth interval of +/- one standard devia- tion of the mean depth of eggs or larvae. Egg mean depth weighted by abundance was regressed on tem- perature. Meteorological data from the Kodiak air- port for the sampling periods in April and May 1986, 1987, and 1988 were used to model hourly irradi- ance at depth in the area by using an attenuation coefficient of 0.16 (a mean value of 33 measurements [SD=0.103] made aboard ship in Shelikof Strait be- tween 3 and 23 May 199 12). 2 Davis, R. AFSC. Personal commun., April 1992. Data from the 36 tows were grouped into nine "se- ries" that were numbered sequentially based on cal- endar day without regard to year. The first time se- ries included eggs sampled on 12-16 April, and the ninth series included larvae sampled on 23-25 May. Tows within a series were taken with the same ra- tionale; i.e. in a fixed location or following a drogue, with similar depth schemes, and close to the same dates (Table 1). Each series was composed of one to eight tows taken in areas of high egg and/or larval concentrations except series six and seven, which were taken primarily for zooplankton studies (Fig. 1). Series eight, in mid-May, was taken about seven weeks after peak spawning, in the spawning area. Series four and nine had time series sufficient to in- vestigate diel differences in vertical distribution. Tows during these two series were taken at local midnight (midway between sunset and sunrise), dawn (sun 20° above the horizon), local noon, and dusk (sun 20 above the horizon). Results Overall densities Egg densities were highest in the first series of tows, taken between 12 and 16 April, when mean concen- tration in the water column based on all nets in the tows varied from 12,057 to 34,734 eggs/1,000 m3 (SD ranged from 5,827 to 30,619). Egg densities were greater than 1,000/1,000 m3 per tow in series two, and generally decreased as the season progressed (Table 2). Eggs were present during series three to eight through the middle of May, but in reduced con- Table 1 Dates and tow depths for MOCNESS series use d to inves tigate vertical distribu tion of walleye pollock, Theragra chalcogramma, eggs and larvae in the Shelikof Strait region. Series Date No tows Day o "year Depths (ml Net 1 Net 2 Net 3 Net 4 N€ t5 Ne t6 Net 7 Net 8 Start End Mm Max Mm Max Mm Max Min Max Mm Max Mm Max Min Max Min Max 1 12-16 April 1987 3 L02 L06 1 r,o "id 100 100 150 150 1 ?(i 170 190 190 210 210 230 230 250 2 21-27 April 1988 7 111 117 1 60 60 120 120 1 in 140 Kill 160 1SII 180 200 200 220 220 250 3 30 April- 1 May 1988 3 120 121 1 20 20 10 40 60 60 80 Ml 100 100 150 15(1 200 200 255 4 8-11 May 1986 t L28 131 2 L5 15 30 30 45 15 60 60 Ml 80 100 100 1511 150 200 5 13-14 May 1986 3 133 134 2 If, 15 30 30 45 15 60 mi Ml Ml 100 100 15(1 150 180 6 15 May 1986 2 135 135 2 15 15 30 30 45 45 lid 60 Ml Ml 100 100 150 150 17(1 7 17-18 May 1986 2 137 138 1 15 15 30 30 45 45 60 60 Ml 80 lull 8 18 May 1986 1 13S 138 1 15 15 30 30 15 45 60 60 80 80 Kill Km 150 150 195 9 23-25 May 1987 Total tows 8 36 1 13 1 15 0 15 15 :;u ::ii 45 45 lid 60 Ml 80 Kill Kin 1 25 125 150 Kendall et al .: Vertical distribution of eggs and larvae of Theragra chalcogramma 543 Alaska Peninsula =t Cape Kekurnoi >-* , ' +1-1 ^*) / +-2-1.5.6,7.3,8 1-3 + / +2-2.3.4 + +5 t, ++6 4 , Sutwik I. +5 ' +7 ' Semidi Is. . f/ Trinity Is 6 Figure 1 Locations of MOCNESS sampling series ( large numbers tows (small numbers). Tows are shown only when there substantial differences among their locations. centrations, partially because sampling then was southwest of the main spawning area (Fig. 1), and partially because it was after peak spawning. The notable exception was the single tow of series eight taken in the area of maximum spawning. Although it was mid-May, egg densities were relatively high ( 139/1,000 m3), indicating that some spawning had occurred within the previous two weeks. Larvae were abundant in all series except the first when they were absent (Table 2). Mean density of larvae among the tows in series two through nine ranged from 39 to 509/1,000 m3 (SD ranged from 54 to 1,011/1,000 m3). Series were not always in the ex- pected area of maximum concentration of larvae and thus do not necessarily represent the seasonal trends in larval abundance (see Kendall and Picquelle, 1990). Series four followed a surface drifter with a drogue at 30-35 m (Incze et al., 1990), whereas sampling during series nine was at a fixed geographic loca- tion. During the 2.5 days of series four, the buoy moved anticyclonically. Catches in these two series varied considerably; during series four the mean density among tows was 82-285 lar- vae/1,000 m3, compared with 42-482 larvae/ 1,000 m3 during series nine (Table 2). The coef- ficient of variation of density among tows for series four was 0.63 and for series nine it was 0.68, indicating that variability among tows using the two sampling strategies was similar. Overall depth distributions Mean depths of eggs decreased during the sea- son. Multiple comparison tests of mean depths of eggs showed significant differences between series one, two, and three, when eggs were most abundant. Among the 10 tows in the first two series, the observed mean depth of eggs was between 153 and 206 m (Fig. 2). In series three through eight, the observed mean depth was less than 130 m, but the number of eggs was relatively small. The shallower towing schemes of series four through eight may have biased the mean depth of eggs, but the general trend is thought to be real. During the second series, when only newly hatched larvae were present, their observed mean depths of occurrence were from 165 to 212 m among tows (Fig. 3). One standard de- viation of mean depth was 27 to 73 m and gen- were erally increased during the series (Fig. 3). In the third series, when recently hatched larvae dominated, larval mean depths varied from 70 to 106 m (range of SD: 83-91 m). As opposed to series two and three, when larvae were mainly found below 100 m, mean depths of larvae during series four ranged from 24 to 58 m (range of SD=15-71 m)3. Mean depths of occurrence of larvae from series five through seven (13-18 May 1986) varied from 15 to 47 m (range of SD=8-36 m) (Table 3). In series eight, taken in mid-May in the spawning area, the larvae averaged 4.6 mm (SD: 0.18 mm) (Table 2), and their mean depth of occurrence was 21 m(SD= 18 m). Dur- ing series nine in late May, mean depths of larvae among the tows ranged from 15 to 38 m (Table 3) and varied on a diel basis (see below). Larvae in the noon tow on the second day of sampling had a mean depth of 58 m (SD=71 m). This was due to an unusually large catch in the deepest net (607/1,000 m3 [23% of all the lar- vae in the towj at 150-200 m) of larvae with a mean length of 4.71 mm. This appeared to be larger than the overall mean of the larvae collected at this depth during this series (4.35 mm), indicating the catch was not all newly hatched larvae that had not moved to the upper water column. If we discount this net, the mean depth of larvae in this tow was 21.5 m, close to the value in the other tows of the series (Fig. 4). 544 Fishery Bulletin 92(3), 1994 Changes in depth distribution with ontogeny There was considerable variation in the mean depths of occurrence and in the abundance of eggs of differ- ent stages among the 13 tows of series one, two, and Table 2 Mean density (number/1,000 m3) and standard deviation (SD) of density by tow of walleye pollock, Theragra chalcogramma, eggs and 1 irvae, anc mean and standard deviation of larval lengths ( mm SL) from MOCNESS tows. Density (no./l,000 m3) Larval length Series Tow Eggs Larvae (mm SL) Mean SD Mean SD Mean SD 1 1 19,994 14.317.2 1 2 12,057 5,827.0 1 3 34,734 30,618.6 2 1 1,509 833.3 168 271.6 3.8 0.04 2 2 3,543 3,204.0 90 89.5 3.7 0.06 2 3 2,402 1,858.9 89 90.0 3.9 0.10 2 4 1,270 804.1 230 257.8 3.8 0.06 2 5 3,641 2,489.5 322 253.2 4.1 0.04 2 B 3,276 1,696.2 389 296.1 3.6 0.04 2 7 2,333 791.3 341 316.6 3.7 0.07 3 1 411 207.6 341 444.2 1 7 0.08 3 2 671 565.3 208 163.8 1 0 0.09 3 3 779 801.3 304 331.8 4.3 0.07 4 1 52 33,4 82 88.1 5.0 0.09 4 2 77 87.1 105 130.6 I 9 0.08 4 3 98 100.7 183 288.8 4.7 0.11 4 4 124 190.6 285 449.4 5.1 0.09 4 5 170 333.2 152 228.6 5.7 0.11 4 6 204 342.4 148 199.0 5.1 0.09 4 7 59 85.0 232 261.3 5.7 0.07 5 1 26 23.7 423 696.8 5.6 0.09 5 2 HI 50.2 51)9 770.6 5.8 0.08 5 3 24 12.9 114 121.3 5.4 0.09 6 1 5 5.0 355 497 1 6.0 0.08 6 2 7 9.1 419 833.7 6 l 0.10 7 1 14 16.1 39 58.8 6.7 0.18 7 2 5 7.7 53 91.6 6.3 0.10 8 1 139 96.7 109 184.5 4.6 0.18 9 1 42 53.5 7.7 0.50 9 2 475 822.4 : 9 0.13 9 3 181 384.8 8.6 0.16 9 4 208 482.7 0.12 9 5 78 125.1 8 2 0.17 9 6 256 385.5 7.7 0.14 9 7 182 1,011.4 7.5 0.16 9 8 396 639.5 8.5 n 11 three (Table 3). Stage groups one and six were sig- nificantly deeper than stage groups three, four, and five; stage group two was intermediate in depth (Fig. 5) (ANOVA, multiple comparisons test P<0.05). Almost all of the larvae collected deeper than 100 m throughout the study were <5 mm, while the length of larvae in the upper part of the water column appeared to increase later in the season (Fig. 6). In several tows of series four through eight, a bimo- dal depth distribution was evident; most larvae were found in the up- per 60 m, almost no larvae found between 60 and 100 m, and larvae again were present deeper than 100 m. Mean lengths of larvae in the nets of the tows of series nine ranged from 4.8 to 9.8 mm, with an overall mean of 7.8 mm. There was no indication of length strati- fication of larvae within the upper 100 m. The mean lengths of larvae among tows in series nine were rela- tively homogeneous (7.2-8.6 mm). Relationship of depth distribution and hydrography The temperature of the water col- umn measured concurrently with the tows in series one through three increased with depth from about 4.0° to 5.0°C near the sur- face to 5.0 and 5.5 C at 150- 250 m, where most of the eggs oc- curred (there is no hydrographic data from tows 3 and 5 of series two). Temperature at the mean depth of occurrence of eggs varied from 4.7 to 5.4'C among the tows in series one through three (Fig. 7A). Since temperature in- creased with depth, among the tows of series one through three there was a positive linear rela- tionship between mean depth of occurrence of eggs and tempera- ture (P<0.001, r2=0.7619). The re- lationship between the depth dis- tribution of eggs and water den- sity among tows of series one through three was not significant (P=0.632l(Fig. 7B). Kendall et al.: Vertical distribution of eggs and larvae of Theragra chalcogramma 545 50 E 100 I 150 200 250 i i i i i i 4 5 6 7 I I I I I I | I I ! I I I I I I ' ' ' ' I ' I I I I I I I I I < :>c < i i i i i i i i i i i i i i i i i i i i i i i i i i 105 1 1 O 115 1^0 125 Day of year 1 30 135 Figure 2 Eggs of walleye pollock, Theragra chalcogramma. Depth of occurrence by day of year and tow series number. Mean and standard deviation of the mean are shown for each tow. 50 E 100 I 150 200 250 i i i i i i i •". I w- I I I i I I I ^U > I I I I I I ) I I I I I I I I I I I I I I I I I I I I 115 120 125 130 135 UO 145 Day of year Figure 3 Larvae of walleye pollock, Theragra chalcogramma. Depth of occur- rence by day of year and tow series number. Mean and standard de- viation of the mean are shown for each tow. When pollock larvae were first present in April as hatchlings in the lower part of the water column, they experienced temperatures of 4.9° to 5.3C (Fig. 8). When the larvae first reached the upper part of the water column they experienced lower temperatures (~3.6°C during series four), but temperature at the mean depth of larval occurrence increased to about 5.7°C during series nine (Fig. 8). Temperatures during series nine decreased with depth from about 6.2°C at the surface to just above 546 fishery Bulletin 92(3), 1994 Table 3 Means and standard deviations (SD) of depths of walleye group) and larvae from MOCNESS tows. pollock, Theragra chalcogramma, eggs! total and by stage Series Tow Total eggs Eggs by stage group Larvae Group 1 Group 2 Group 3 Group 4 Group 5 Group 6 Mean SD Mean SD Mean SD Mean SD Mean SI) Mean SD Mean SD Mean SD 1 1 198 29.9 209 19.4 207 23.9 194 30.0 182 44.7 173 47.6 197 25.9 1 2 179 44.8 180 31.0 174 45.6 174 51.9 183 51.2 196 51.7 200 44.9 1 3 206 32.1 214 19.2 207 27.4 196 44.0 168 54.2 195 41.0 156 41.2 2 1 L64 50.5 Hil 33.2 158 48.5 156 59.6 155 54.1 158 45.7 199 42.3 212 27.9 2 2 187 33.9 179 22.4 182 27.6 193 35.1 191 36.1 184 36.5 200 35.4 198 49.8 2 3 196 28.9 196 16.9 189 24.1 191 34.7 198 31.1 199 30.1 203 30.1 204 26.6 2 1 186 44.1 177 21.9 181 25.3 163 52.0 140 65.1 177 49.2 210 33.5 195 47.6 2 5 167 43.0 206 14.2 167 36.2 143 28.8 127 38.8 144 39.4 193 34.0 189 49.5 2 6 153 42.0 193 23.2 158 47.1 145 37.3 146 37.4 143 39.7 187 40.6 165 72.7 2 7 169 42.0 184 24.9 M,7 31.1 159 32.6 166 49.8 159 in 1 186 46.7 168 59.7 3 1 68 65.5 1 11 54.6 95 66.9 54 73.3 37 40.0 56 43.6 101 77.0 70 91.1 3 2 92 60.9 133 33.4 68 52.4 84 67.8 76 46.9 92 49.1 77 67.7 811 82.7 3 3 llii 55.7 84 68.8 129 56.1 94 65.5 90 47.6 102 44.8 130 51.7 106 88.9 4 1 59 40.4 25 20.5 4 2 100 61.7 28 .30.1 4 3 S6 52.7 37 48.3 4 4 124 57.9 27 14.5 4 5 126 65.6 24 40.9 4 6 1 29 65.2 58 70.5 4 i 97 64.2 30 25.7 5 1 75 24.7 29 18.6 5 2 111 50.4 33 19.0 5 3 73 50.6 47 36.2 6 1 59 57.6 28 16.5 6 2 39 24.3 25 8.0 7 1 20 11.2 23 12.3 7 2 47 32.7 L5 14.8 8 1 43 47.2 21 17.7 9 1 30 36.4 9 2 38 10.9 9 3 15 15.4 9 4 23 9.6 9 5 31 16.6 9 6 33 14.3 9 7 2.3 8.8 9 8 29 11.3 Kendall et al.: Vertical distribution of eggs and larvae of Theragra chalcogramma 547 0 10 20 -g 30 £ 40 a. o Q 50 60 70 12 24 36 48 60 72 84 96 Time (hours since 0000 8 May) 108 120 MLD Mean larval depth Night Mean nauplii depth Tow 6, less deep net Figure 4 Time series of mixed layer depth, mean depth of copepod nauplii, and mean depth of walleye pollock, Theragra chalcogramma, lar- vae starting at 0000 hours, 8 May GMT. Mean depth of larvae excluding the catch in the deepest net of tow 6 also is plotted (see Results: overall depth distribution). Nighttime is shown by stip- pling near the top border. 30 iii ■ I 2° a 10 © ■a c a> u E E 2 -10 ■ o , § -20 • k 1 1 --«---- Series one \ V 1 Q -30 ' -40 1 , — ■*-- Series two — ••— Series three > 1 2 3 4 5 I Egg stage group Figure 5 Vertical distribution of walleye pollock, Theragra chalcogramma, eggs by developmental stage group from series one, two, and three. Differences in mean depth of each stage group from mean depth of all eggs in a series are plotted. SERIES 2 3 A n i — i s a 7 a 9 DDE10 □ 1 , 9 l|llll|llll|li ii|iiiiiiiMi _i 1 °W Q -lOO m o - E 8 - A - IOC m •*; % 7 O J > DC 2 6 CP IL CP 4 O _ <0 5 X • 0 ° o * 0 1- : o 8 c A * a 4 UI i * / * » _1 z 3 A - < lu . I I I 1 1 ' S , i i i 1 i i i . J .... 1 .... 1 . . 1 15 120 125 130 DAY OF YEAR Figure 6 135 UO 145 Standard lengt is (SL) of walleye pollock, Theragra chalcogramma larvae in tows above and below 100 m by day ol year and tow series number. 5.0"C from about 60 m to the bottom (Fig. 9). There was a gradual thermocline between 10 and 40 m, and most of the larvae were present in or above this fea- ture. Salinity during this series showed a gradual increase with depth from about 31.7 ppt at the sur- face to 32.2 ppt at 140 m. Density increased steadily from a at of 24.85 at the surface to 25.25 at 60 m and 25.45 at 140 m (Fig. 9). 548 Fishery Bulletin 92(3). 1994 a Q Depth - -452 * 122»Temp. 50 3. - 100 3. 3 150 ■ 2. ^ . 2 200 X 100 150 200 4.0 4.5 5.0 5.5 6.0 250 4.0 4.5 5.0 Temperature CO 5.5 Temperature CO 6.0 a B 0 50 3. - 100 3. . 3 150 . 2 .2 -2 .si 200 i :\ 24.5 25.0 25.5 26.0 25.0 25.5 Density (sigma-t) Density (sigma-t) Figure 7 Mean depth of occurrence of walleye pollock, Theragra chalcogramma, eggs in series one, two, and three in relation to temperature (A) and density (B). Series numbers are indicated next to data points. Tem- perature (C ) and density (D I profiles during series one, two, and three. Q 4,00 4 50 Temperature Figure 8 Mean depth of occurrence of walleye pollock, Theragra chalcogramma, larvae by tow in relation to tempera- ture. Series numbers are near appropriate data points. Diel changes in larval depth distribution There was no clear diel pattern in the depth distribution of 5-6 mm larvae in series four ( Fig. 4 ). During the first diel sampling period the mean depths var- ied from 25 to 37 m; the deepest mean depth occurred at dusk. During the sec- ond diel period, night was not sampled and an aberrant catch occurred during the noon sampling period (discussed above). The mean depths varied from 24 to 58 m, if the catch in the deep net of the noon tow is included, and from 21 to 30 m if that catch is excluded. When proportions of larvae at each depth from series nine are examined, a clear pattern of vertical distribution emerges despite differences in overall density among tows. Although the deep- est stratum sampled was 125-150 m, mean depths of occurrence of larvae were between 14.6 and 38.1 m (range of one SD=8.8-36.4 ml, and fewer than V/c of all larvae were collected below 60 m. Within the upper 60 m, the larvae showed evidence of limited diel vertical migrations (Fig. 10). The observed mean depth of larvae was greater at noon than at other times ( 38 m and 33 m ) and shal- lowest at dusk ( 15 m and 23 m). At night the larvae were found somewhat deeper ( 23 m and 29 m ). Like the dusk sampling, the distribution at night appeared deeper during the second day compared with the first. At dawn the larvae were found be- tween the night and noon depths (30 m and 31 m). Although the same pattern of changes in vertical position of the larvae was observed during both 24- hour periods of series nine, differences in observed mean depths between various times of day were con- sistently greater on the first day than on the second (Fig. 10). The average difference in mean depth be- tween sequential time periods on the first day was 13.2 m versus 6.2 m on the second (Table 4). The mean depths at dawn on the two days were within a meter of each other, but at the other sampling times there were differences of 4.8-8.4 m between mean depths on the two days at the same time of day. Dif- ferences were especially pronounced at dusk and night. Examination of standard deviations of mean depth and mean lengths of larvae among tows of se- ries nine showed no consistent pattern of differences with any of the variables under consideration (time of day, depth, day). 26.0 Kendall et at: Vertical distribution of eggs and larvae of Theragra chalcogramma 549 Response of larvae to wind events and prey distributions Increasing winds after 60 hours of sampling during series four prevented further MOCNESS tows, but other observations con- tinued during the storm and documented the deepening of the mixed layer and subsequent changes in microzooplankton distribution (Incze et al., 1990). The path of the drogue, the sequence of CTD data obtained in the area, and limited satellite imagery suggested the presence of an anticyclonic eddy (Incze et al., 1990;Nieman4). The mixed layer depth during the first half of series four was vari- able, but during the last half it deepened, presumably in response to increased winds (Fig. 4). Copepod nauplii of length range 150- 350 p, which has been found to be the size range primarily eaten by 5-6 mm walleye pol- lock larvae (Paul et al., 1991 ), had mean depths between 20 and 34 m during series four (Fig. 4 ). Their observed mean depth increased during the storm, but their densities at some depth within the upper 45 m was always greater than 15 per liter. Excluding the deep net from the noon tow on day two, the mean depths of larvae in series four were at or 5-10 m below the mixed layer depth and the mean depth of 150- 350 urn copepod nauplii. Wind, measured hourly aboard the ship during series nine, in- creased from less than 8 m/second during the first six tows of this series to over 12 m/second by the end of the series. The mixed layer was about 25 m deep at the time of the last tow (night) in series nine as opposed to about 10 m dur- ing the previous seven tows (Fig. 9). The greater mean depths of larvae at dusk and night on the second day of sampling compared with the first day of sampling were possibly caused by increased tur- bulence and deepening of the mixed layer (Fig. 10). Discussion Most walleye pollock eggs in Shelikof Strait developed at depths between 150 and Density (slgma-t) 24.8 24.9 25 25.1 25.2 25.3 0 10 Vv. Oanalty ^^-^ 20 ^> ^^ Temp«f«!ur« ? £30 a a Q 40 - P^-x 50 - f ";-^s 5 5.5 6 6.5 Temperature (°C) — Mean Tow 8 Figure 9 Temperature and density profiles from series nine. Profiles of means of all tows and from tow eight, showing deepening of the mixed layer (see Results: response of larvae to wind events and prey distributions), are plotted. Table 4 Compai •isons ( means and standard deviations [SD]) of walleye pollock. Theragra chalcog ramma, larvae by depth (m), ] ength ( mm SL ) and density (no./l,000 m3) in series nine at four times of day (dawn, noon, dusk and night). Density Tow Time Depth (m) Length imm SLl ino./l,000 m3) Mean SD Mean SD Mean SD 1 dawn 30 36.4 7.7 0.50 42 54 2 noon 38 10.9 7.9 0.13 475 822 3 dusk L5 15.4 8.6 0.16 181 385 4 night 23 9.6 7.2 0.12 208 I S3 5 dawn 31 16.6 8.2 0.17 78 125 6 noon 33 14.3 7.7 0.14 256 386 7 dusk 23 8.8 7.5 0.17 482 1,011 8 night dawn noon dusk night day 1 day 2 29 30.6 36.6 21.4 27.0 26.4 29.0 11.3 8.5 8.06 7.79 7.72 8.09 7.84 7.97 0.11 396 33.6 249.6 228.4 150.7 124.9 156.2 640 4 Nieman, D. R. Rosenstiel School of Marine and Atmospheric Science, Univ. Miami, Miami, Florida 33149. Personal commun., February 1993. 200 m. However, there was considerable variation in the mean depth of eggs among the tows. Mean depth of eggs varied from 153 to 206 m in April (when egg densities were high) and from 20 to 129 m in 550 Fishery Bulletin 92(3), 1994 May. Although many factors probably contributed to the distribution of eggs, their distribution was positively related to temperature, which in- creased with depth. Kendall and Kim ( 1989) developed a model, based on field collections and laboratory experiments, to de- scribe the vertical distribution of walleye pollock eggs from Shelikof Strait in relation to water density. One of the model's assumptions was that the specific gravity of eggs does not vary interannually. Based on this assumption and their observations of the changes in egg buoyancy and depth distribution during develop- ment, eggs would rise to different depths in the middle stages of devel- opment depending on water density in particular years. The vertical dis- tribution might then influence the horizontal distribution of eggs if there was a significant vertical shear in the water column. Like Kendall and Kim (1989), we found middle-stage pollock eggs at shallower depths than early or late stage eggs. Our data, however, suggest that the depth distribution of eggs changes during development regardless of water density. Pollock egg density appears to vary interannually — in the eight tows from which significant numbers of eggs were collected and for which concurrent hydrographic data are available, the mean depth of occurrence varied from 153 to 206 m, and the density varied from 25.31 ot to 25.63 or Among the four years for which Kendall and Kim (1989) present data (1977, 1981, 1985, 1986), density in the middle layer of the water col- umn ( 162-216 m) varied interannually from 25.58 a, to 25.87 ar and modelled depth distribution of middle-stage eggs varied from about 160 to 230 m. While the ranges of density and depth of eggs seen in the present study are similar to those modelled by Kendall and Kim (1989), the proposed relation- ship between depth of eggs and water density is not evident. A relationship might have been seen if a greater range of water densities had been found. The temperature and density of the water in which most of the eggs were found in the present study ( 1987 and 1988) most closely resembled the values reported for 1981 by Kendall and Kim (1989). Ingraham et al. ( 1991 ) compared long-term annual means of water temperature, salinity, and density at 225 m depth in Shelikof Strait with values in in- dividual years when circulation in the Gulf of Alaska was anomalous. They found high values for all three 24 36 48 60 72 Time (hours since 0000 22 May) 84 'J 6 — *— Wind speed Larval mean depth Night Figure 10 Mean depth of walleye pollock, Theragra chalcogramma, larvae and wind speed during series nine. variables in 1985; this was a year when water on the continental slope did not include fresher, colder wa- ters from the eastern Gulf of Alaska owing to reduced westward transport. Of the years considered here (1985-88), only 1985 was characterized by anoma- lous flow conditions that could have produced un- usually warm, dense (>5.4°C, >26.2 a,) bottom water in Shelikof Strait (Ingraham et al., 1991). According to the model in Kendall and Kim (1989), the eggs should have risen closest to the surface in 1985. How- ever, mean depth of eggs in the four tows in 1985 was 220 m, in the five tows in 1986 it was 211 m (Kendall and Kim, 1989), in the three tows of series one in 1987 it was 200 m, and in the seven tows of series two in 1988 it was 176 m. In series three through eight, the mean depth of occurrence of eggs was less than 130 m. Given the low numbers of eggs and the sampling intervals de- signed mainly to sample larvae, these values are not robust. Kendall and Kim ( 1989) also found some eggs with significantly lower density than others in their specific gravity experiments. The data presented here confirm that some eggs have a low specific gravity value, but that these are infrequent and occur pri- marily later in the season after the majority of eggs have hatched. Apparently after hatching, larvae move quickly to the upper part of the water column. Both eggs and larvae in series two had mean depths of occurrence between 153 and 212 m among the seven tows. The mean length of larvae in series two was 3.8 mm, which is within the range of size at hatching (Kendall Kendall et al.: Vertical distribution of eggs and larvae of Theragra chalcogramma 551 et al., 1987). In series three, the mean depths of oc- currence of eggs and larvae decreased to between 68 and 110 m, and the mean length of larvae increased to 4.4 mm, indicating that growth of some pollock larvae had occurred (SD of length was 0.18 in series three as opposed to 0.09 in series two). The standard deviations of depth of larval occurrence in series three (83 to 91 m) were larger than in any other series, suggesting that these larvae were in transition from the deep hatching environment to shallower levels. Although larvae respond positively to light within 24 hours of hatching, their negative geo/barotaxis may enable them to reach the upper layers since in- sufficient light for response penetrates to hatching depths (see Olla and Davis, 1990). The relatively shallow mean depth of eggs in series three also may account for the large variation in larval depth dur- ing that series. Older larvae from eggs at the depths observed during series two (>150 m) could have mixed with larvae hatching from eggs found at the depths observed during series three (<110 m). The larger standard deviation of larval length in series three compared with series two supports this expla- nation. In later series the mean depth of occurrence of the larvae was less than 60 m, and the standard deviation generally was less than 20 m. Once they reach the upper layers, vertical movements increase as larvae develop. No significant diel migrations were noted in series four, as opposed to the pattern seen in series nine. The mean length of larvae in series four was 5.3 mm; in series nine it was 7.9 mm. The larvae sampled by Kendall et al. (1987) in late May were 11.0 mm long and demonstrated a pattern of vertical distribution similar to the larvae collected here in series nine. During series nine, larvae followed a diel (crepus- cular) pattern of vertical movements in which they ranged deepest at noon, shallowest at dusk, and pro- gressively deeper through the following noon. Al- though this pattern was observed on both days, the amplitude of movements were reduced on the sec- ond day. However, the wind had markedly increased by evening of the second day Larvae may have been avoiding the turbulent surface on the second day when their mean depths were deeper. Olla and Davis ( 1990) found that pollock larvae avoid turbulence in the laboratory. The relationships of larval fish feeding, growth, and survival to storms and turbulence have been the subject of numerous studies (e.g. see Sundby and Fossum, 1990; Maillet and Checkley, 1991). Both positive and negative effects have been postulated and observed. Positive effects of increased turbulence include hypothesized enhanced encounter rates be- tween larvae and their prey (Rothschild and Osborn, 1988), and enhanced primary production after mix- ing has ceased owing to infusion of nutrients from below the photic zone. Negative impacts include di- lution of vertically enriched layers of prey to levels below successful feeding thresholds and reduced naupliar production in lower phytoplankton concen- trations (Lasker, 1978). There is evidence that in Shelikof Strait, below-average walleye pollock pro- duction may result if strong wind events occur when larvae are at the first-feeding stage.5 Incze et al. ( 1990 ) found that naupliar concentrations remained above feeding threshold levels during the passage of a storm, but this was a relatively transient phenom- enon. The present study indicates that larvae may avoid upper layer turbulence by moving deeper in the water column. If so, they might experience prey densities or light levels too low for optimal feeding. In a 24-hour study of the vertical distribution of pollock larvae in Auke Bay, Alaska (average depth 60 m), Pritchett and Haldorson (1989) found larvae congregated at 10 m at noon, at 5 m at dawn and afternoon, and at 15-20 m at night. At twilight (0.3 hour before sunrise, 1.5 hours before sunset), larvae were more dispersed, seen mostly at 10 m near sun- rise and at 15 m near sunset. The vertical extent of diel migration increased with larval length. In the present study, depths of occurrence were greater at all times than those reported by Pritchett and Haldorson ( 1989), and noon depths of larger larvae were greater than the night depths. However, in both studies, larvae were found to be deeper at noon than at dawn and dusk, and a relationship between verti- cal migration and larval length was seen. The depth distribution of copepod nauplii in Auke Bay usually centered around 5-10 m (Paul et al., 1991). Inzce et al. ( 1990) reported maximum densities of copepod nauplii in the upper 30 m of Shelikof Strait when pollock larvae are abundant. Since nauplii are the primary prey for pollock larvae, the larvae may well adjust their daytime feeding depths to correspond to those of the nauplii. Alternatively, the greater day- time depth of pollock larvae in Shelikof Strait may be related to the greater depth of light penetration in Shelikof Strait compared with Auke Bay (Zeimann etal., 1990). Light is frequently cited as a factor controlling the depth of occurrence for fish larvae. Larvae of some species follow the common trend of rising toward the surface at night and of remaining deeper during the day (Smith et al., 1978; Kendall and Naplin, 1981; Davis et al., 1990). Other species follow an opposite pattern, ranging deeper at night than by day (Boeh- 5 Bailey, K. M., AFSC, and S. A. Macklin, Pacific Marine Envi- ronmental Laboratory. 7600 Sand Point Way NE., Seattle, Wash- ington 98115-0070. Personal commun., February 1993. 552 Fishery Bulletin 92(3). 1994 lert et al., 1985; Yamashita et al., 1985; Sogard et al., 1987; Davis et al., 1990). Fewer studies have examined the vertical distribution offish larvae dur- ing crepuscular periods. The present study has shown that larger larval pollock range deeper during the day than at night and that, at dawn and dusk, they are present at shallower depths in the water column than at midday. We hypothesize that these changes in vertical position allow pollock larvae to extend the length of their daily feeding period. In the laboratory, first-feeding pollock larvae could not feed at light levels below 0.006 |imol-m_2-s-1 (Paul, 1983). Except at night, light levels are brighter than those at the depths where we found feeding-stage larvae. In studies of behavioral responses of walleye pollock larvae (4-8 mm SL) to light in the labora- tory, Olla and Davis (1990) found reduced activity and orientation in a nonfeeding mode at light levels <0.01 and avoidance of light at levels >13 u.mol-m~2s_1. In the dark, larvae migrated up- ward and remained in the upper part of the cham- I 3 5 7 9 11 13 Sampling Day Figure 1 1 Predicted light levels at depth on days of sampling and mean depths of walleye pollock, Theragra chalcogramma , larvae in MOCNESS tows ( indicated by circles for daytime tows, triangles pointed down for dusk tows, asterisks for night tows, and tri- angles pointed up for dawn tows). Depths of three light levels are plotted: 50, 10 and 0.01 umol-m~2s_1. Light levels above 50 are clear, those 10-50 are light gray, those 0.01-10 are medium gray, and those less than 0.01 |imol-m '2s _1 are dark gray. Light levels are based on incident light at the Kodiak airport during the sampling periods with an extinction coefficient of 0.16. bers, demonstrating negative geotaxis or barotaxis, or both. Light levels between 0.01 and 10 umolirr2-s_1 are estimated to have occurred between about 25 and 60 m during our sampling in Shelikof Strait. The mean depths of feeding larvae were typically in the upper part of this range (Fig. 11). Larvae longer than 7 mm seemed to adjust their vertical position on a diel cycle to stay at light levels similar to those "pre- ferred" in the laboratory. At night these larvae were present at depths where light had been greater than 10 p:mol-nr2-s_1 during the day. The relationship of vertical distribution of larval fish to vertical temperature structure of the water column varies among species (Kendall and Naplin, 1981; Sogard et al., 1987). Hypothetically, there are metabolic advantages to diel descents into cooler waters ( Lampert, 1989). Larvae that stay nearer the surface (at higher temperatures) at night when they are digesting their food may accrue such advantages (Wurtsburgh and Neverman, 1988). However, given the small differences in temperature with depth ob- served here (~1°C), energetic advantages are almost certainly insignificant compared with the ad- vantages of feeding at optimal light levels and at depths of maximum prey abundance. An alternative advantage of residing deeper, and thus at lower light levels during daytime, may be to avoid visual predators. Acknowledgments Many people, mainly from the AFSC, helped collect and process the samples and data upon which this paper is based. We thank them all, particularly those who endured the harsh conditions of Shelikof Strait during the cruises there. David Nieman rendered his expert help in operating the MOCNESS on several cruises. Allen Macklin (PMEL) is thanked for obtaining and analyzing meteo- rological data presented here. Richard Davis (AFSC) graciously let us use irradiance with depth data obtained by his modelling efforts. An earlier draft of the manuscript was helped immeasurably by reviews of many of the re- searchers involved in the FOCI (Fisheries- Oceanography Coordinated Investigations) program: Kevin Bailey, Jim Schumacher, Jeff Napp, Ric Brodeur, Bori Olla, Gary Stauffer, and Susan Picquelle (whom also aided in sta- tistical analyses). Ian Perry (Canada, Depart- ment of Fisheries and Oceans) also provided a very helpful review of the manuscript. Kendall et al.: Vertical distribution of eggs and larvae of Theragra chalcogramma 553 Literature cited Ahlstrom, E. H. 1959. Vertical distribution of pelagic fish eggs and larvae off California and Baja California. Fish. Bull. 60:107-146. Blood, D. B., A. C. Matarese, and M. M. Yoklavich. 1994. Embryonic development of walleye pollock, Theragra chalcogramma, from Shelikof Strait, Gulf of Alaska. Fish. Bull. 92:207-222. Boehlert, G. W., D. M. Gadomski, and B. C. Mundy. 1985. Vertical distribution of ichthyoplankton off the Oregon Coast in spring and summer months. Fish. Bull. 83:611-621. Coombs, S. H., R. K. Pipe, and C. E. Mitchell. 1981. The vertical distribution of eggs and larvae of blue whiting (Micromesistius poutassou) and mackerel (Scomber scombrus) in the eastern North Atlantic and North Sea. Rapp. P.-v. Reun. Cons, int. Explor. Mer 178:188-195. Davis, T. L. O., G. P. Jenkins, and J. W. Young. 1990. Diel patterns of vertical distribution in lar- vae of southern bluefin Thunnus maccoyii and other tuna in the East Indian Ocean. Mar. Ecol. Prog. Ser. 59:63-74. Ellertson, B., P. Solemdal, S. Sundby, S. Tilseth, T. Westgard, and V. Oiestad. 1981. Feeding and vertical distribution of cod larvae in relation to availability of prey organisms. Rapp. P.-v. Reun. Cons. int. Explor. Mer 178:317-319. Enright, J. T. 1977. Diurnal vertical migration: adaptive significance and timing. Part 1: Selective advantage: a metabolic model. Limnol. Ocean. 22:856—872. Hardy, A. C. 1936. Plankton ecology and the hypothesis of ani- mal exclusion. Proc. Linn. Soc. Lond. 148:64-70. Ingraham, W. J., Jr., R. K. Reed, J. D. Schumacher, and S. A. Macklin. 1991. Circulation variability in the Gulf of Alaska. EOS, Trans. Am. Geophys. Soc. 74:257-264. Incze, L. S., J. Gray, J. D. Schumacher, A. W. Kendall Jr., K. M. Bailey, and S. A. Macklin. 1987. Fisheries-Oceanography Coordinated Inves- tigations (FOCI) field operations— 1986. NOAA Data Report ERL PMEL-20, 64 p. Incze, L. S., P. B. Ortner, and J. D. Schumacher. 1990. Microzooplankton, vertical mixing and advec- tion in a larval fish patch. J. Plankton Res. 12:365-379. 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., and N. A. Naplin. 1981. Diel-depth distribution of summer ichthy- oplankton in the Middle Atlantic Bight. Fish. Bull. 79:705-726. Kendall, A. W., Jr., and S. Kim. 1989. Buoyancy of walleye pollock (Theragra chalcogramma) eggs in relation to water proper- ties and movement in Shelikof Strait, Gulf of Alaska. In R. J. Beamish and G. A. McFarlane (eds.l, Effects of ocean variability on recruitment and an evaluation of parameters used in stock as- sessment models, p. 169-180. Can. Spec. Publ. Fish. Aquat. Sci. 108. Kendall, A. W., Jr., and S. J. Picquelle. 1990. Egg and larval distributions of walleye pol- lock Theragra chalcogramma in Shelikof Strait, Gulf of Alaska. Fish Bull. 88:133-154. Kendall, A. W., Jr., M. E. Clarke, M. M. Yoklavich, and G. W. Boehlert. 1987. Distribution, feeding, and growth of larval wall- eye pollock, Theragra chalcogramma, from Shelikof Strait, Gulf of Alaska. Fish. Bull. 85:499-521. Lampert, W. H. 1989. The adaptive significance of diel vertical mi- gration of zooplankton. Funct. Ecol. 3:21-27. Lasker, R. 1978. The relation between oceanographic condi- tions and larval anchovy food in the California Current: identification of factors contributing to recruitment failure. Rapp. P.-v. Reun. Cons. Int. Explor. Mer 173:212-230. Lawrence, L. A., J. Gray, and J. D. Schumacher. 1991. Fisheries-Oceanography Coordinated Inves- tigations field operations — 1987. NOAA Data Rep. ERL PMEL-28, 61 p. Lough, R. G., and D. C. Potter. 1993. Vertical distribution patterns and diel migra- tions of larval and juvenile haddock Melano- grammus aeglefinus and Atlantic cod Gadus morhua on Georges Bank. Fish Bull. 91:281-303. Maillet, G. L., and D. M. Checkley Jr. 1991. Storm-related variation in growth rate of otoliths of larval Atlantic menhaden Brevoortia tyrannus: a time series analysis of biological and physical variables and implications for larva growth and mortality. Mar. Ecol. Prog. Ser. 79:1-16. Munk, P., T. Kiorboe, and V. Christensen. 1989. Vertical distribution of herring, Clupea harengus, larvae in relation to light and prey distribution. Environ. Biol. Fish. 26:87-96. Neilson, J. D., and R. I. Perry. 1990. Diel vertical migrations of marine fishes: an obligate or facultative process? Adv. Mar. Biol. 26:115-168. Norcross, B. L., and R. F. Shaw. 1984. Oceanic and estuarine transport offish eggs and larvae: a review. Trans. Am. Fish. Soc. 113:153-165. Olla, B. L., and M. W. Davis. 1990. Effects of physical factors on the vertical dis- tribution of larval walleye pollock Theragra chalco- gramma under controlled laboratory conditions. Mar. Ecol. Prog. Ser. 63:105-112. 554 Fishery Bulletin 92(3), 1994 Paul, A. J. 1983. Light, temperature, nauplii concentrations, and prey capture by first feeding pollock larvae, Theragra chalcogramma. Mar. Ecol. Prog. Ser. 13:175-179. Paul, A. J., K. O. Coyle, and L. Haldorson. 1991. Interannual variations in copepod nauplii prey of larval fish in an Alaskan Bay. ICES J. Mar. Sci. 48:157-165. Pritchett, M., and L. Haldorson. 1989. Depth distribution and vertical migration of larval walleye pollock (Theragra chalco- gramma). In Proceedings of the international symposium on the biology and management of walleye pollock; 14-16 November 1988, Anchorage, AK, p. 173-183. Alaska Sea Grant Rep. 89-1, Univ. Alaska, Fairbanks. Proctor, P. D. 1989. Fisheries-Oceanography Coordinated Inves- tigations— field operations 1988. NOAA Data Report ERL PMEL-25, 69 p. Rothschild, B. J., and T. R. Osborn. 1988. Small-scale turbulence and plankton contact rates. J. Plankton Res. 10:465-474. Scheaffer, R. L., W. Mendenhall, and L. Ott. 1986. Elementary survey sampling. Duxbury Press, Boston, MA, 324 p. Schumacher, J. D., and A. W. Kendall Jr. 1991. Some interactions between young walleye pollock and their environment in the western Gulf of Alaska. CalCOFI Rep. 32:22-40. Sclafani, M., C. T. Taggart, and K. R. Thompson. 1993. Condition, buoyancy and the distribution of larval fish: implications for vertical migration and retention. J. Plankton Res. 15:413-435. Smith, W. G., J. D. Sibunka, and A. Wells. 1978. Diel movements of larval yellowtail flounder, Limanda ferruginea, determined from discrete depth sampling. Fish. Bull. 76:167-178. Sogard, S. M., D. E. Hoss, and J. J. Govoni. 1987. Density and depth distribution of larval Gulf menhaden, Brevoortia patronus, Atlantic croaker, Micropogonias undulatus, and spot, Leiostomus xanthurus, in the northern Gulf of Mexico. Fish. Bull. 85:601-609. Sundby, S., and P. Fossum. 1990. Feeding conditions of Arcto-Norwegian cod larvae compared with the Rothschild-Osborn theory on small-scale turbulence and plankton contact rates. J. Plankton Res. 12:1153-1162. Weibe, P. D., K. H. Burt, S. H. Boyd, and A. W. Morton. 1976. A multiple opening/closing net and environ- mental sensing system for zooplankton. J. Mar. Res. 34:313-326. Wurtsburgh, W. A., and D. Neverman. 1988. Post-feeding thermotaxis and daily vertical migration in a larval fish. Nature 333:846-848. Yamashita, Y., D. Kitagawa, and T. Aoyama. 1985. Diel vertical migration and feeding rhythm of the larvae of the Japanese sand-eel Ammodytes personatus. Bull. Japan. Soc. Sci. Fish. 51:1-5. Zeiman, D. A., L. D. Conquest, K. W. Fulton- Bennett, and P. K. Bienfange. 1990. Interannual variability in the physical envi- ronment of Auke Bay, Alaska. In D. A. Ziemann and K. W. Fulton-Bennett (eds.), APPRISE-inter- annual variability and fisheries recruitment, p. 99- 128. The Oceanic Institute, Honolulu, HI. Abstract. — Otoliths, scales, dor- sal spines, and pectoral-fin rays were compared to ascertain the best hardpart for determining the age of weakfish, Cynoscion regalis. Each showed concentric marks, which could be interpreted as an- nuli. Sectioned otoliths, however, consistently showed the clearest marks, had 100% agreement be- tween and within readers, and were validated by the marginal in- crement method for ages 1-5. This validated method of ageing weak- fish was then compared with the traditionally used scale method. The scale method was less precise, as demonstrated by lower percent agreement, and generally assigned younger ages for fish older than age 6 (as determined by otoliths). Con- sequently, mean sizes at age based on scales showed no clear signs of an asymptote, whereas those based on otoliths did. Otolith annuli formed in April and May, whereas scale annulus formation was more variable, ranging from April to Au- gust. This extended time of annu- lus formation made scales poorly suited for back calculation. A comparison of a validated otolith method to age weakfish, Cynoscion regalis, with the traditional scale method Susan K. Lowerre-Barbieri Virginia Institute of Marine Science, College of William and Mary Gloucester Point. Virginia 23062 Present address University of Georgia Marine Institute Sapelo Island. Georgia 31327 Mark E. Chittenden Jr. Virginia Institute of Marine Science, College of William and Mary Gloucester Point. Virginia 23062 Cynthia M. Jones Applied Marine Research Laboratory, Old Dominion University Norfolk. Virginia 23529 The weakfish, Cynoscion regalis, is a recreationally and commercially important sciaenid found from east- ern Florida to Massachusetts, and is most abundant from North Caro- lina to New York (Mercer, 1985). Believed to be resident year-round in the Carolinas, they are found far- ther north only seasonally (Bigelow and Schroeder, 1953). In the spring, weakfish migrate northward and inshore to estuarine feeding and spawning grounds; this pattern is reversed in the fall (Wilk, 1979). Most fish are believed to overwinter off North Carolina (Pearson, 1932). Weakfish are found in Chesapeake Bay, roughly from April through No- vember (Pearson, 1941; Massmann et al., 1958), where they support one of the region's most important fisheries (Rothschild et al., 1981). Weakfish age and growth studies have been based almost exclusively on scales (Taylor, 1916; Nesbit, 1954; Perlmutter et al., 1956; Mass- mann, 1963a; Merriner, 1973; Shep- herd and Grimes, 1983). However, problems with this method have been reported: 1 ) small fish may not lay down a first annulus on scales (Welsh and Breder, 1923), 2) older fish have closely spaced annuli that are difficult to interpret (Taylor, 1916; Shepherd, 1988), 3) annuli form over a long time period, April- August, and scales are difficult to interpret during annulus formation (Nesbit, 1954; Massmann, 1963b), 4) the time annuli form varies an- nually and regionally ( Perlmutter et al. , 1956 ), and 5 ) checks (false annuli ) and regenerated scales are common (Merriner, 1973). The scale method of ageing weakfish also has not been conclusively validated by current standards (Beamish and McFarlane, 1983; Brothers, 1983). Perlmutter et al. (1956) and Shepherd and Grimes (1983) both tried to validate annuli on scales by the marginal increment method, however they used pooled age data and did not report the age range. Manuscript accepted 8 November 199.3. Fishery Bulletin 92:555-568 ( 1994). Contribution 1826 from the College of William and Mary. School of Marine Science, Vir- ginia Institute of Marine Science, Gloucester Point. Virginia 23062. 555 556 Fishery Bulletin 92(3). 1994 Although recent studies have shown that for many species the scale method underages older fish at the point where fish growth becomes asymptotic (Beamish and Chilton, 1981; Beamish and McFar- lane, 1983; Barnes and Power, 1984), there has been little evaluation of other weakfish hardparts. Merriner (1973) compared weakfish scales to whole vertebrae and otoliths, and Villoso ( 1989) compared scales to whole otoliths. Both concluded that scales were best. However, Merriner's study was conducted before thin-sectioning of otoliths (Williams and Bedford, 1974; Beamish, 1979; Beamish and Chilton, 1981) and other hardparts became common and Villoso ( 1989) did not consider thin-sectioning. A decline in weakfish landings since 1980, coupled with greater competition between fisheries, caused the Atlantic States Marine Fisheries Commission (ASMFC) to develop a weakfish management plan in 1985 (Mercer, 1985). Since then the ASMFC has issued an updated stock assessment1 and suggested a 25% reduction in coast-wide exploitation rates (Amendment No. 1 of the Weakfish Fishery Manage- ment Plan of the ASMFC). However, it is essential to proper weakfish management that a validated ageing technique be developed and used, as improper ageing can lead to faulty estimates of model param- eters such as age at maturity, growth, longevity and mortality (Beamish and McFarlane, 1983). The objectives of this study were 1) to compare otolith, dorsal-fin spine, and pectoral-fin ray sections with scales in terms of legibility and interpretation of potential annual marks, ease of collection and pro- cessing, and precision, 2) to validate the hardpart demonstrating the greatest clarity by mar- ginal increment analysis for each age group found in the Chesapeake Bay area, and 3) to conduct a more in-depth compari- son of the validated hardpart with scales in terms of precision and accuracy, time of annulus formation, growth estimates, and use in back calculation of body length. Methods Preliminary comparison of hardparts Four hundred weakfish were collected ev- ery other week during April-October in 1989 from three Chesapeake Bay commer- cial pound nets. On each collection day, one 1 Vaughan, D. S., R. J. Seagraves, and K. West. 1991. An assessment of the Atlantic weakfish stock, 1982-1988. Atl. States Mar. Fish. Comm. Spec. Rep. 21. Wash. DC, 29 p. anterior 22.7 kg (50 lb) box of each available grade of weak- fish— small, medium, or large — was bought and all fish within it processed. Fish were measured for to- tal length (TL ±1.0 mm), sexed, and both sagittal otoliths were removed and stored dry. Scales were removed from an area just posterior to the tip of the left pectoral fin, below the lateral line. The left pecto- ral fin and the entire dorsal fin were removed by cut- ting below the base of the rays. Scales and fins were stored in paper envelopes and kept frozen until prepa- ration for ageing. A total of 45 fish, 15 from each grade, were ran- domly selected from the fish collected in 1989 for a preliminary comparison of hardparts. These fish ranged from 244 to 615 mm TL and each of their four hardparts was prepared for reading as described below. The right otolith from each fish was transversely sectioned through the nucleus with a Buehler low- speed Isomet saw. Sections 350-500 urn thick were mounted on glass slides with Flo-Texx clear mount- ing medium and viewed under a dissecting micro- scope at 24x magnification by using transmitted light and bright field, with the exception of samples from the period April-May, when sections were also read with reflected light and dark field to help identify the last annulus. Thin opaque bands, presumed to represent annual marks, were counted along the otolith sulcal groove (Fig. 1). Because opaque bands inhibit light passage, they appeared dark in trans- mitted light (Fig. 2A) and light in reflected light. Scales from each fish were soaked in water until soft, after which they were washed gently with a soft-bristled tooth brush. Three or four clean, unregenerated scales •]■ ".,-.i! posterior proximal distal ventral arm of Ihe sulcal groove thin opaque bands Figure 1 Schematic representation of a transverse section taken through the right sagittal otolith. The ventral arm of the sulcal groove, along which otoliths were measured, is indicated. The whole otolith is positioned as it would be in a weakfish, Cynoscion regulis. Lowerre-Barbieri et al.: A comparison of otolith and scale ageing methods for Cynoscion regalis 557 t-; w en -t_> cO •— 1 ■1-H „ CO o o c- ° 45 o O •S 01 < Oi ^ CO to Ih 4) 5 C S a> J2 s ax to .o 0) 4J "3 -r a. to S ot S-. CO O co -a T3 3 1 c <—■ j_» c S- -c 45 o JD ^ O U CO E S * £ 6 -a rt "3 ii CM Of ll 13 03 03 -w J -O -C Eh •*£ .co 03 "8 "a s - J; £ e .o ^3 CO '-^ +3 5- CO • — +j o £ c to «s 3 c 03 o J3 +* to ure 2 eakfis C to •- to C - 0) c o> o 01 — 1 c gi £ to -a to S 03 o) 3 £ & 0 15 . Ih CO £ «8 c g o «Q > CO — >> |1 T3 £ 5 Of 8 « Of 5e 03 £e 0 CO 03 6 -5 a 12 £ d to "o £ £ s 1^ ^H 03 0 to II J= CO "o" r 01 >> < 6 •fi o J* s *J «3 0 CO ho 15 0 0 £ 0 ya co T3 ■« II si 0) T3 CJ c 15 ~ 01 V -^ to JS M S C be ■~- -^ 03 a cti CO i-o E ■£ S CO eg B. — cd J3 c '£ Of c Of to to C to 03 be G 03 £ c 0 .- to c c 5 -^ o ••"■ to s- '43 c - M O Of Of ". Of Of co to E 558 Fishery Bulletin 92(3). 1994 were then dried, taped to an acetate sheet, inserted between two other blank sheets, and pressed with a Carver laboratory scale press for two minutes at 2,721 kg of pressure at 71°F. Because of the large size of weak- fish scales, scale impressions were read with a stan- dard microfiche reader at 20x. Those scales with po- tential annuli crowded along the scale periphery were also viewed at 48x under a dissecting microscope. Pre- sumed annual marks were identified by standard cri- teria (Bagenal and Tesch, 1978; Shepherd, 1988). One spiny ray from the dorsal fin and one soft ray of the left pectoral fin were prepared from each fish. Rays were serially sectioned by starting at their base and cutting through most of their length at a thick- ness of 400 urn with a Buehler low-speed Isomet saw. Sections were then mounted on microscope slides with Flo-Texx and read under a dissecting microscope with transmitted light and dark field at 64x. Pre- sumed annual marks were counted when they could be identified as individual, opaque bands. Each hardpart was read twice by two separate readers. Readings were done in a randomly selected order, with no knowledge of collection date or fish size. Hardparts were evaluated in terms of clarity of presumed annual marks, ease of collection and pro- cessing, and precision. Precision was measured by average percent agreement within and between read- ers, i.e. percent agreement within readers was cal- culated for each reader separately and then averaged for the two readers and percent agreement between readers was calculated separately for each reading and then averaged for the two readings. Validation of the otolith method Because otoliths were found best for ageing, additional samples were collected for validation. During 1989- 92, 1,928 weakfish were collected from commercial pound-net, haul-seine, and gill-net fisheries in Ches- apeake Bay. During March-November when weakfish are not present in the Chesapeake Bay, fish were col- lected ( n =289 ) from the trawl fishery operating in North Carolina shelf waters north of Cape Hatteras. The marginal increment method was used to vali- date otolith annuli (Brothers, 1983; Casselman, 1987; Hyndes et al., 1992). The translucent margin out- side the proximal end of the last annulus was mea- sured along the ventral side of the otolith sulcal groove (Fig. 1). Measurements were taken with an ocular micrometer to the nearest 0.038 mm (one mi- crometer unit at a total magnification of 24x). Comparison of scales and otoliths To compare the otolith and scale methods in more detail, 155 fish ranging from 140 to 845 mm TL were selected by stratified, random subsampling — strata being otolith-determined ages — from a total of 300 fish collected in 1989 and 1992. Thirty fish were se- lected from each of the age-strata, 1-4. Because older fish were scarce, only 14 age-5, 16 age-6, two age-7, two age-8, and one age- 10 fish were included. Al- though most fish came from Chesapeake Bay com- mercial fisheries, in order to increase the number of older fish, 27 fish were collected in May 1992 at the Delaware Bay Weakfish Sport Fishing Tournament. We collected an additional 20 fish in August 1992 to include fish from each of the summer months for mar- ginal increment and back-calculation analyses. Hardparts were prepared as described for the pre- liminary comparison and read twice by each of two readers. An effort was made to determine annuli on scales based only on physical criteria and not to as- sign annuli based on any preconceived ideas of growth (Casselman, 1983). Reading order was ran- domized and collection date and fish size were un- known. Each reader recorded the number of pre- sumed annuli and a "+" if there was growth beyond the last annulus or a "*" if the last presumed annu- lus was forming or had just formed (Casselman, 1987 ). After all hardparts had been read, we assigned ages using a January 1 birthdate, knowledge of the time of annulus formation, the relative growth of the hardpart margin, and date of capture. Variability within reader, between readers, and between hardparts was analyzed by percent agree- ment. When an individual reader's counts of pre- sumed annuli disagreed, a third reading was made. When readers' ages disagreed, a third reading with both readers present was made to resolve the dis- agreement. To compare time of annulus formation and its vari- ability in scales and otoliths, mean monthly relative marginal increments and their ranges were calcu- lated and plotted (April-October). Relative marginal increments were calculated by dividing the marginal increment by the hardpart radius. All ages were pooled. Additionally, those hardparts which had been designated as having an annulus on the margin ("*") were reviewed and their time of collection recorded. To determine marginal increments and to conduct back-calculation analyses, hardparts were measured by using a Via 100 camera/monitor system with a dissecting microscope at 24x. Otolith radius ( OR) and otolith annular radius (OAR) — the distance from the nucleus to the proximal edge of each annulus — were measured along the ventral arm of the sulcal groove. Scale radius (SR) and scale annular radius (SAR) were measured along the left radius (Ricker, 1992). Marginal growth was measured from outside the last annulus to the hardpart edge. Lowerre-Barbieri et al.: A comparison of otolith and scale ageing methods for Cynosaon regalis 559 To evaluate the applicability of scales and otoliths for back-calculation, it was necessary to first ana- lyze separately their total length to hardpart rela- tionships. Seasonal effects were assessed by compar- ing hardpart size of one age class taken from differ- ent seasons to that predicted by the linear regres- sion of total length on hardpart size for all fish. Only one age class ( age 3 ) was used to remove any confound- ing effects of age. This age class was chosen because it was well-represented throughout the seasons. Back-calculation relationships for both scales and otoliths were based on the "body proportional" hy- pothesis (Francis, 1990) proposed by Whitney and Carlander(1956): L=\g(Si)lg(Sc)]Lc< where g is the total length on hardpart radius func- tion, L; is back-calculated TL at age i, S is the mea- sured hardpart size at annulus i, and S . and L. are the respective hardpart size and total length at cap- ture. Only fish collected in April and May — the be- ginning of the somatic growth season — were used, to remove seasonal effects from the back-calculation equations (Ricker, 1992). Because body-proportional back-calculation is based not just on the relationship of hardpart size to total length but also on the rela- tionship of hardpart size to consecutive annuli, mean annual growth increments were also calculated and compared between scales and otoliths. The tendency for older fish to produce smaller back- calculated lengths at younger ages than observed, known as Lee's phenomenon (Smith, 1983), was evaluated by calculating mean SAR and mean OAR for each age at capture. In this way it was possible to determine if older fish demonstrated slower hardpart growth at younger ages, i.e. true Lee's phe- nomenon (Smale and Taylor, 1987). Data were analyzed by using %2 tests and regres- sion methods available through the Statistical Analy- sis System (SAS 1988). Rejection of the null hypoth- esis in statistical tests was based on a=0.05. Assump- tions of linear models were checked by residual plots as described in Draper and Smith ( 1981). Results Preliminary comparison of hardparts All four hardparts showed concentric marks that were interpreted as annuli (Fig. 2). However, marks on the dorsal spines and pectoral rays were incon- sistent, often blurred or impossible to follow around most of the section and therefore difficult to inter- pret. Presumed annuli on scales were distinctly clearer and more regular than those on dorsal spines and pectoral rays, but they still required some sub- jective interpretation. Presumed annuli on otoliths were exceptionally clear, consistent, and easy to interpret. Typical otolith sections showed an opaque nucleus surrounded by a translucent zone followed by a pat- tern of thin, opaque zones alternating with wide, translucent zones along the sulcal groove (Fig. 2A). In some sections the translucent zone between the nucleus and the first opaque zone was relatively small and made more opaque by a number of fine, circular, opaque bands. However, in all sections the first opaque zone beyond the nucleus was easily iden- tified and considered to be the first annulus. Presumed annuli on scales were harder to iden- tify than those on otoliths but were usually identifi- able as a clear zone in the anterior field, where circuli are either absent or more widely spaced, and by cut- ting over in the lateral fields (Fig. 2D). Checks were most apparent in the anterior field. A clear zone in the anterior field was considered a check if it was not accompanied by distinct cutting over in the lat- eral fields. The first annulus was the hardest to iden- tify. It rarely showed a clear band in the radii zone, although cutting over was sometimes apparent. Its position was based predominantly on the first point at which a large number of secondary radii originated. Presumed annual marks on dorsal spines were fairly clear in some sections but incomplete or blurred in others (Fig. 2C), whereas pectoral-fin ray sections were consistently hard to interpret (Fig. 2B). Pre- sumed annual marks on both these hardparts ap- peared as wide, opaque, semicircular bands alternat- ing with narrow translucent zones. Otoliths showed the greatest precision, with 100% average agreement within and between readers. Scales also had high average agreement: 89% within readers and 80% between readers. Dorsal and pectoral fin sections showed the lowest agreement (Table 1 ) and little confidence was attached to their age assignments. Table 1 Average percent agreement in the prelimi nary com- parison of weakfish, Cynscion regalis, hardpart mark counts within readers, between readers, and with otoliths. Within Between With Hardpart readers readers otoliths Scales 89 80 27 Pectoral rays 59 64 49 Dorsal spines 66 76 46 Otoliths 100 100 560 Fishery Bulletin 92(3). 1994 5-i y 4 Scales / 3- 2 $/ 12 2- a/ 8 i 1 - / 6 2 annuli Pectorals / presumed < IT/ 7 o / 2 W *-• c => o- O O 4 - 3- 2- 1 - Dorsals / 4 \q/ 10 10// 1 0 12 3 4 5 Otolith count Figure 3 Counts of presumed annuli from weak- fish, Cynoscion regalis, scales, pectorals, and dorsals compared with otoliths. The number of fish each point represents is indicated. The 45° line represents 100% agreement. age ' 0 8- age 4 > n 58 101 ft o: 01 ^* 22 I 0 4 - B 0 2- ,^n 0 MJJASONO 32 ^ 7 ril , II ! JFMAMJJASOND E E, c CD E 0 o c To c en b. 03 age 2 I 31 1M age 5 jL^dD FMAMJ JASOND JFMAMJJASOND age 3 a&Q ad I age 6 . n FMAMJ JASOND JFMAMJJASOND Month Figure 4 Mean monthly otolith marginal increments for weakfish, Cynoscion regalis, ages 1-6 from the Chesapeake Bay region, 1989-91. Vertical bars are ± one standard error. Numbers above the bars represent sample size. The number of presumed annual marks on otolith sections agreed poorly with those on other hardparts (Fig. 3). Scale and otolith readings agreed only 27% of the time (Table 1) and scales consistently had one less mark than otoliths (26 out of 45). Pectoral and dorsal rays showed better agreement with otoliths than with scales, 49% and 46% respectively. Validation of the otolith method Opaque bands are laid down on otoliths once a year in the spring. Mean monthly marginal increment plots for ages 1-6 showed only one trough during the year, indicating only one opaque band was formed per year (Fig. 4). A few fish began to lay down annuli in March, as shown by the decrease in mean mar- ginal increment and a relatively high variation in marginal increment size. However, lowest marginal increment values occurred in April and May, indi- cating most fish formed annuli during these months. Greatest otolith growth occurred during the months of June, July, August, and September, as demon- strated by the step-wise increase in mean marginal increments. By October, mean marginal increments reached a fairly stable maximum, indicating little or no otolith growth. This maximum continued until the next March or April, when annuli were again laid down. Because of the scarcity of older fish, it was not pos- sible to validate conclusively fish older than age 5 by separate marginal increment plots. However, there was no evidence that the pattern of annulus forma- tion changed within the weakfish lifespan. Annuli were consistently formed during March-May for fish of different sizes, sexes, and ages (1—6), and otoliths did not form more than one mark per year even though these ages represented various stages in the Lowerre-Barbien et al.: A comparison of otolith and scale ageing methods for Cynoscion regalis 561 fish's life history. Additionally, of the 2,217 otoliths examined (ages 1-10), all those in the process of form- ing or which had just formed annuli were collected in March-May. Thus, we assumed for ages 1-10 that the otolith method provided accurate ages. Comparison of scales and otoliths Scales were consistently more difficult to read than otoliths, and confidence in scale readings was often low. Percent agreement within and between readers was fairly consistent for both hardparts. However, otoliths showed much higher agreement (98-100%), than did scales (78-80%) (Table 2). Although agree- ment between scales and otoliths was fairly high, 79%, agreement decreased with increasing age. Of 32 disagreements, only 6 differed by more than one year (Fig. 5). However, 4 of the 5 fish older than age 6 were underaged by scales and two of the oldest fish, age 10 and 8, were underaged by 3 years. Scales from older fish, if they showed more than 6 annuli, had marks which were severely crowded and fragmented even when viewed at higher magnification (Fig. 6A), whereas otoliths from these same fish showed clear annuli (Fig. 6B). Although the number offish underaged was small, their effect on estimating growth curves would be dramatic. Mean body size at age based on scales, al- though slightly curvilinear, showed no clear indica- tion of an asymptote (Fig. 7A) and thus would not be appropriate for fitting a von Bertalanffy growth curve (Gallucci and Quinn, 1979). In contrast, mean body size at age based on otoliths showed the clear begin- nings of an asymptote (Fig. 7B). Although sex of the fish had no effect on the preci- sion or repeatability of scale readings, it did affect accuracy. Agreement of scale ages among and be- tween readers was quite similar when calculated separately by sex, ranging from 75 to 79.5%. How- ever, agreement between scale and otolith ages, or accuracy, was significantly different for males and females . Time of annulus formation is not the same for scales and otoliths. Both hardparts showed only one trough in their mean monthly marginal increments (Fig. 8). However, otoliths with annuli on their mar- gins were collected only during a discrete time pe- riod, 1 April-1 June, while scales in the process of forming annuli were collected from mid-April to mid- August, although most scales formed annuli in Au- gust. The variable and extended time of scale annu- lus formation is represented by the shallow trough (Brothers, 1983) and the larger standard errors of the scale marginal increment plot, as compared with that of otoliths (Fig. 8). Although total length on hardpart size relation- ships for both scales and otoliths showed linear trends (r2=0.94 and 0.88 respectively, « = 175, P=0.0001), the total length on otolith relationship showed seasonal variation. When a single age class (age 3) was marked by season of collection and plot- ted against the linear relationship predicted by the total sample (Fig. 9), all fish collected in April and May had smaller than predicted otolith radii, whereas fish collected in August and September had larger than predicted radii. Fish collected in June and July were intermediate, although most of their radii were also smaller than predicted. Scales from the same fish did not show similar seasonal trends. Back-calculation equations of total length on hard- part size were calculated only for fish collected at the beginning of the growing season, in April and 10n CD CD ro 0) CO o CO Otolith age Figure 5 Weakfish, Cynoscion regalis, assigned ages from scales and otoliths. The number of fish each point represents is indicated. The 45 line represents 1001?* agreement. 562 Fishery Bulletin 92(3), 1994 B Figure 6 The scale impression (A) and sectioned otolith (B), as seen in transmitted light, from a male, 10-year-old weakfish, Cynoscion regalis, TL=845 mm, collected in mid-May. Arrows indicate marks counted as annuli. May, to remove seasonal effects. Although linear re- gressions were significant for scales (/•-=(). 95, P=0.0001) and otoliths (r2=0.92, P=0.0001), a qua- dratic term improved the model fit and was signifi- cant (P=0.0003 scales, P=0.0001 otoliths) (Fig. 10). Equations were For scales: TL = -151.6 + 160.2 SR - 5.4 SR2 (r2=0.96, n=88, P=0.0001); For otoliths: TL = -220.9 + 543.1 OR- 66.9 OR2 (r=0.94, n=88, P=0.0001). The pattern of mean annual growth increments differed between scales and otoliths. Both scales and otoliths showed their largest growth increment from the focus to the first annulus (Fig. 11 ). However, once fish had reached age 1, the largest otolith annual growth increment occurred between the first and Lowerre-Barbieri et al.: A comparison of otolith and scale ageing methods for Cynoscion regalis 563 E E c .o 900- A Scales 16 16 1 3 I 600- 20 X 26 I 300- 37 I 35 900- B Otoliths 2 14 X 1 2 I 600- 30 X 30 J 300- 30 X n- 30 Age (years) Figure 7 Mean weakfish, Cynoscion regalis, size at age: (A) based on scales and (B) based on otoliths. Vertical bars are ± one standard error. Numbers above the bars represent sample size. ^9 c CD E CD O c Id c E? co E > ■4— ' _ca CD Otoliths 3 12 _L 25 20 27 59 27 Apr May Jun Jul Aug Sep Oct Scales ay 2/ T Apr May Jun Jul Aug Sep Oct Month Figure 8 Mean monthly relative increments for weakfish, Cynoscion regalis, scales and otoliths. Vertical bars are ± one standard error. Numbers above the bars represent sample sizes. second annuli, whereas scales had a very small in- crement between these annuli. The largest scale growth increment after age 1 was between annuli 3 and 4. Neither hardpart showed a consistently de- creasing mean annual growth increment as age in- creased. Although this assumption is often included in scale-reading criteria, it would be inappropriate for weakfish. Back-calculated mean body sizes at age were larger for scales than for otoliths (Table 3). In part, this discrepancy may reflect different times of annulus formation: back-calculated lengths from scales, in general, estimate sizes in August, whereas back-cal- culated lengths from otoliths estimate sizes in April and May. Also, at older ages, back-calculated body sizes at age based on scales would be expected to be larger because of the underageing of older fish by scales. Both scales and otoliths showed smaller back-cal- culated mean body size at age 1 than observed. At later ages, back-calculated TL's from scales were larger than observed, while back-calculated TL's from otoliths showed no consistent trend (Table 3). The cause of the smaller back-calculated TL's at age 1, however, did not appear related to Lee's phenomenon, as there was no consistent trend of smaller age-1 an- nular radii at older ages at capture (Tables 4 and 5). In fact, the largest mean SAR and OAR at age 1 came from 5-year-old fish. However, age-1 OAR's from the oldest fish in the study (>age 6, n=5) were distinctly smaller than those observed in younger fish. 564 Fishery Bulletin 92(3). 1994 5 0 6 0 CO P 7001 Otoliths '<5U Hardpart radius (mm) Figure 9 Total length plotted on hardpart size for age-3 weak- fish, Cynoscion regalis, n=35. Lines represent the linear total length to hardpart regressions calculated from all fish, n = 175. Fish are marked by season of collection: open circles=April-May; shaded triangles=June-July; and black squares=August- September. Discussion Our results indicate that transverse otolith sections are the best method to age weakfish. Sectioned otoliths were characterized by thin opaque bands, considered annuli, interspersed with wider translu- cent zones. This pattern is similar to other sciaenids, such as spotted seatrout, Cynoscion nebulosus (Maceina et al., 1987), Atlantic croaker, Micro- pogonias undulatus (Barbieri et al., 1994), red drum, Sciaenops oeellatus (Murphy and Taylor, 1991), and black drum, Pogonias cromis (Beckman et al., 1990). This pattern should not be confused with the more 1000 Scales . .-^ .♦* * 3 - 500 * ■ * age / *C* m 1 V+ " 3 T 4 / X 5 JL. • 6 E .J^^ ■ A 7 / X 8 E n / O 10 0 5 2.5 4 5 6 5 8 5 _C ■*—> en c © 1000 — Otoliths (0 • • y J? O , ,!j^^ I- , VxX-^« X <$ '***&?* * 500 '•*,* a9e sY* ■ 1 / ' 2 yf , ■ 3 » 4 • XJ*- 6 / O 10 OS 15 2 5 Hardpart radius (mm) Figure 1 0 Weakfish, Cynoscion regalis, total length on hardpart radius regression used for back calcula- tion. Based on fish collected in April and May, n=88. E E c 0) E 0) o c 5 10 o Ol QJ m o co oo Otoliths Scales 63 33 13 090 J^ 0 60 0 30 CQ O n — , ID 3 (D =1 3 3 0-1 1-2 2-3 3-4 4-5 5-6 Between annuli Figure 1 1 Mean annual growth increments of weakfish, Cynoscion regalis, scales and otoliths. Vertical bars are ± one standard error. Numbers above the bars represent sample sizes. Lowerre-Barbieri et al .: A comparison of otolith and scale ageing methods for Cynoscion regalis 565 common otolith pattern found in many temperate fish of thin translucent zones, which are considered an- Table 3 Mean back-calculated weakfish, Cynoscion regalis, total lengths (mm) at age based on scales and otoliths, calculated from a qua dratic body to hardpart regression and observed mean total length at time of annulus formation. Sample size is in parentheses. Observed Observed Age Scales Jul/Aug Otoliths April/May 1 196(1521 240 (7) 162(174) 172(22) 2 305(127) 296(25) 297(144) 260 (2) 3 422 (77) 377 (8) 421 (99) 532(12) 4 564 (42) 514 (51 552 (64) 566(181 5 682 (20) 660 (34) 663 (141 6 733 (4) 711 (14) 741 (16) 7 750 (5) 710 (1) 8 748 (2) 759 (2) 10 845 (1) Table 4 Mean scale annulai radii (SAR for each scale age of weakf sh, Cynoscion rega lis. Age n Scale annulus 1 2 3 4 5 6 1 12 2.59 2 52 2.31 3.20 3 24 2.40 3.42 4.14 4 29 2.38 3.27 4.27 5.56 5 16 2.65 3.44 4.31 5.43 7.15 6 16 2.38 3.25 4.30 5.58 6.64 7.00 7 3 2.11 3.09 3.92 5.65 6.69 7.37 Table 5 Mean otolith ann alar r adii (OAR) for each otolith age of wea kfish, Cynoscion regalis. Age n Otolith annul us 1 2 3 4 5 6 7 8 1 29 0.83 o 45 0.85 1.27 3 35 0.82 1.21 1.56 4 30 0.82 1.20 1.53 1.91 5 14 0.88 1.25 1.58 1.91 2.28 6 16 0.86 1.22 1.54 1.88 2.21 2.52 7 2 0.80 1.18 1.47 1.79 2.16 2.47 2.79 8 2 0.77 1.20 1.56 1.90 2.22 2.47 2.65 2.85 10 1 0.67 1.11 1.52 1.94 2.15 2 32 2.49 2.67 nuli, interspersed with wide opaque zones (Hyndes etal., 1992). Sectioned otoliths were consistently clear and easy to read, as shown by the high precision of repeated age readings. Although it was possible only to vali- date ages 1-5 by separate marginal increment plots, otolith annuli in all ages examined (1-10) were laid down once a year during a discrete time period (April-May). The constancy of annulus deposition at older ages, the lack of severely crowded annuli in older fish, and the similarity between weakfish otoliths and other sciaenid otoliths that have been validated at older ages (Beckman et al., 1990, Murphy and Taylor, 1991; Barbieri et al., 1994) sug- gest that otoliths are a reliable ageing technique for weakfish, although older ages must still be validated. In contrast, we found the scale method of ageing weakfish to be imprecise and apparently inaccurate at older ages. We found that scales form annuli over an extended period, April-August, similar to the re- sults of past studies (Perlmutter et al., 1956; Massmann, 1963b). This protracted period of annu- lus formation made it difficult to assign ages to fish taken in midsummer with moderate growth on the scale margin, as noted by Massmann (1963b). For example, a fish taken in July with a medium mar- ginal increment on its scale could have formed its annuli in early April and have grown since then, or it could have increased its growth increment before forming an annulus in August. Thus, assigning an age to these fish is purely subjective and can lead to ageing errors ± one year, which may explain most of the discrepancies between otolith and scale ages. The long period of annulus formation on scales and the severe crowding of annuli at older ages make it difficult to validate scales by the marginal increment method — as Perlmutter et al. (1956) and Shepherd and Grimes (1983) attempted for pooled age data. Because scale annuli form over a protracted pe- riod, the trough in the marginal increment plot is shallow and the range of marginal growth during other months is large. Additionally, validation by the marginal increment method is not appropri- ate if the hardpart shows severe crowding of an- nuli at older ages, as we found with scales, and has been previously reported (Shepherd, 1988). Shepherd (1988) described annuli in fish older than age 6 or 7 as being crowded and very diffi- cult to detect, which could lead to marginal incre- ments being measured from the last distinguish- able annulus to the edge, rather than from the last real annulus to the edge. This error would inflate marginal increment estimates and there would be no way to detect underaged, older fish in marginal increment plots. 566 Fishery Bulletin 92(3), 1994 The scale method appears to underage older weak- fish. Assuming otolith ages were valid, 4 of the 5 fish in this study older than age 6 were underaged by scales. Although 4 out of 155 fish may seem insig- nificant, the importance of correctly ageing these fish cannot be judged only by the number of discrepan- cies. These fish represent the beginning of an asymp- tote in growth and fish in the asymptotic range are often rare in highly exploited stocks. Obtaining and correctly ageing a few weakfish in this range is criti- cal to correctly estimating the parameters of the von Bertalanffy growth curve. Annulus formation on weakfish otoliths and scales shows different patterns. The formation of otolith annuli over a discrete time period suggests it may be caused by environmental variables. The most com- monly suggested environmental influences on annu- lus formation are temperature, salinity, food, and light (Simkiss, 1974). Weakfish form annuli on their otoliths in April and May, when they migrate from offshore winter grounds to estuarine feeding and spawning grounds. Thus, annulus formation may be linked to their migration into a different environment. Weakfish scales, in contrast, have a more variable time of annulus formation suggesting a cause other than general environmental conditions. Scales may undergo resorption whereas otoliths do not ( Simkiss, 1974 ), and spawning has been linked to scale resorp- tion with a consequent scale mark in salmon and trout (Crichton, 1935). Spawning may also be linked to formation of annuli on weakfish scales (Merriner, 1973). Weakfish mature at age 1 (Merriner, 1976; Shepherd and Grimes, 1984) and are multiple spawn- ers with a protracted spawning period from May through August (Lowerre-Barbieri"). However, indi- vidual spawning periods are asynchronous and vary greatly, especially in time of termination. Spawning activity and annulus formation may be linked in two ways: 1 ) annuli could form on scales early in the spawning season when resources are shifted towards production of reproductive materials — especially the yolking of oocytes, or 2) annuli might form near the end of the season, owing to the cumulative drain of protracted spawning, causing a cessation in growth and thus an annulus. A connection between scale annulus formation and spawning in weakfish would explain the high level of variation in time of annulus formation and the higher accuracy of ages based on scales taken from females, because females usually invest more energy in reproduction. It might also explain the small growth increment between annuli 1 and 2 if one-year-old weakfish begin spawning later 2 Lowerre-Barbieri, S. K. 1993. Reproductive biology of weakfish, Cynosaon regalis, in the Chesapeake Bay region. School of Ma- rine Science, VIMS. College of William and Mary, unpubl. manuscr. in the season than older fish, owing to a threshold size necessary to reach maturity. Our results indicate both scales and otoliths present problems for back-calculation of weakfish. Although scales showed a strong relationship be- tween body and hardpart size and no seasonal dif- ferences in growth, their long and variable time of annulus formation may cause considerable error (Smith, 1983). It is impossible to determine if a fish formed its annuli at the same time each year. Be- cause annuli can form from April to August, incre- ments may represent 8-16 months of growth rather than approximately one year of growth. Additionally, scale annuli are more difficult to distinguish than otolith annuli, making SAR's difficult to measure and somewhat subjective. However, otoliths show sea- sonal change in the body to hardpart relationship, making a season-specific back-calculation equation, such as we developed, inappropriate for fish collected outside of that season. Additionally, comparisons between back-calculated and observed sizes at age were complicated by the weakfish migrational pat- tern, since weakfish age ranges in the Chesapeake Bay vary seasonally — older fish are present only in spring and only occasionally in fall (Joseph, 1972). There was no clear evidence of Lee's phenomenon, as older fish did not consistently show smaller hardpart size at younger ages. The five oldest fish did, however, demonstrate considerably smaller OAR's at age 1 than did their younger counterparts. Nevertheless, these same fish did not demonstrate consistently smaller OAR's at consecutive ages than did younger fish. Thus, the smaller OAR's at age 1, rather than demonstrating Lee's phenomenon, may simply reflect when most fish of those year classes were born, i.e. fish born early in the spawning sea- son would have larger OAR's at age 1 because they had more time to grow before winter, than did fish born later in the season. Previous criticism of back-calculation has focused mainly on the body size to hardpart relationship and its calculation (Campana, 1990; Casselman, 1990; Francis, 1990; Ricker, 1992). However, the validity of back-calculation also depends on the constancy, clarity, and pattern of hardpart growth increments. The different growth increment patterns we found between scales and otoliths demonstrate the need to understand hardpart growth better, how it relates to somatic growth and what causes annulus forma- tion on different hardparts. Future studies of weakfish age and growth should be based on sectioned otoliths because scales appear inaccurate once growth becomes asymptotic. This common failing of the scale method has been reported for many species (Beamish and McFarlane, 1987 ). It Lowerre-Barbieri et al.: A comparison of otolith and scale ageing methods for Cynoscion regalis 567 can result in underestimates of longevity, overesti- mates of mortality, inaccurate growth calculations, and improper modelling and management decisions (Beamish and McFarlane, 1983). Similarly, current estimates of weakfish growth, longevity, and mor- tality may need to be reevaluated, as suggested by our findings that scales underage older fish and have crowded annuli past age 6. The need for this reevalu- ation is underscored by the recording of a 17-year- old, as aged by otoliths, which was previously aged as a 7-year-old by scales (Lowerre-Barbieri3). Acknowledgments We would like to thank the Chesapeake Bay com- mercial fishermen, James Owens, and the people at the Delaware Weakfish Sport Fishing Tournament for helping us obtain samples. Richard Seagraves provided us with information on the Delaware fish- ery as well as otolith samples. Steve Bobko and Donna Kline helped with the processing and origi- nal analysis of the four hardparts. We would like to thank J. M. Casselman, Luiz R. R. Barbieri, and an anonymous reviewer for their helpful suggestions to improve the manuscript. Financial support was pro- vided by the College of William and Mary, Virginia Institute of Marine Science, by Old Dominion Uni- versity, Applied Marine Research Laboratory, and by a Wallop/Breaux Program Grant from the U.S. Fish and Wildlife Service through the Virginia Marine Resources Commission for Sport Fish Restoration, Project No. F-88-R3. Literature cited Bagenal, T. B., and F. W. Tesch. 1978. Age and growth. In T, B. Bagenal (ed.). Methods for assessment offish production in fresh waters, 3rd ed.. p. 101-136. Blackwell Scientific Pubis., Oxford. Barbieri, L. R., M. E. Chittenden Jr., and C. M. Jones. 1994. Age, growth, and mortality of Atlantic croaker, Micropogonias undulatus, in the Chesapeake Bay region, with a discussion of apparent geographic changes in population dynamics. Fish. Bull. 92:1-12. Barnes, M. A., and G. Power. 1984. A comparison of otolith and scale ages for western Labrador lake whitefish (Coregonus clu- peaformis). Env. Biol. Fish. 10:297-299. 1 Lowerre-Barbieri, S. K. 1993. Age and growth of weakfish, Cynoscion regalis, in the Chesapeake Bay region. School of Ma- rine Science, VIMS, College of William and Mary, unpubl. manuscr. Beamish, R. J. 1979. Differences in the age of Pacific hake (Mer- luccius product us ) using whole otoliths and sections of otoliths. J. Fish. Res. Board Can. 36:141-151. Beamish, R. J., and D. E. Chilton. 1981. Preliminary evaluation of a method to deter- mine the age of sablefish iAnoplopoma fimbria). Can. J. Fish. Aquat. Sci. 39:277-287. Beamish, R. J., and G. A. McFarlane. 1983. The forgotten requirement for age validation in fisheries biology. Trans. Am. Fish. Soc. 112:735-743. 1987. Current trends in age determination metho- dology. In R. C. Summerfelt and G. E. Hall (eds.), Age and growth offish, p. 15-42. Iowa State Univ. Press, Ames. Beckman, D. W., A. L. Stanley, J. H. Render, and C. A. Wilson. 1990. Age and growth of black drum in Louisiana waters of the Gulf of Mexico. Trans. Am. Fish. Soc. 119:537-544. Bigelow, H. B., and W. C. Schroeder. 1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv. Fish Bull. 53:1-577. Brothers, E. B. 1983. Summary of round table discussions on age validation. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 8:35-44. Campana, S. E. 1990. How reliable are growth back-calculations based on otoliths? Can. J. Fish. Aquat. Sci. 47:2219-2227. Casselman, J. M. 1983. Age and growth assessment offish from their calcified structures — techniques and tools. U.S. Dep. Commer, NOAA Tech. Rep. NMFS 8:1-17. 1987. Determination of age and growth. In A. H. Weatherley and H. S. Gill (eds.), The biology of fish growth, p. 209-242. Academic Press, London. 1990. Growth and relative size of calcified structures offish. Trans. Am. Fish. Soc. 119:673-688. Criehton, M. I. 1935. Scale resorption in salmon and sea trout. Salm. Fish., Edinb. 4:1-8. Draper, N. R., and H. Smith. 1981. Applied regression analysis, 2nd ed. John Wiley & Sons, New York, 709 p. Francis, R. I. C. C. 1990. Back-calculation of fish length: a critical review. J. Fish Biol. 36:883-902. Gallucci, V. F., and T. J. Quinn II. 1979. Reparameterizing, fitting, and testing a simple growth model. Trans. Am. Fish. Soc. 108:14-25. Hyndes, G. A., N. R. Loneragan, and I. C. Potter. 1992. Influence of sectioning otoliths on marginal increment trends and age and growth estimates for the flathead, Platycephalus speculator. Fish. Bull. 90:276-284. Joseph, E. B. 1972. The status of the sciaenid stocks of the middle- Atlantic coast. Chesapeake Sci. 13:87-100. 568 Fishery Bulletin 92(3). 1994 Maceina, M. J., D. N. Hata, T. L. Linton, and A. M. Landry Jr. 1987. Age and growth analysis of spotted seatrout from Galveston Bay, Texas. Trans. Am. Fish. Soc. 116:54-59. Massmann, W. H. 1963a. Age and size composition of weakfish, Cynoscion regalis, from pound nets in Chesapeake Bay, Virginia, 1954-1958. Chesapeake Sci. 4:43-51. 1963b. Annulus formation on the scales of weak- fish, Cynoscion regalis, of Chesapeake Bay. Chesapeake Sci. 4:54-56. Massmann, W. H., J. P. Whitcomb, and A. L. Pacheco. 1958. Distribution and abundance of gray weakfish in the York River system, Virginia. Trans. N. Am. Wildl. Conf. 23:361-369. Mercer, L. P. 1985. Fishery management plan for the weakfish (Cynoscion regalis) fishery. North Carolina Dep. Nat. Res. Comm. Dev., Div. Mar. Fish., Spec. Sci. Rep. 46, 129 p. Merriner, J. V. 1973. Assessment of the weakfish resource, a sug- gested management plan, and aspects of life his- tory in North Carolina. Ph.D. diss.. North Caro- lina State Univ., Raleigh, NC, 201 p. 1976. Aspects of the reproductive biology of the weakfish, Cynoscion regalis (Sciaenidae), in North Carolina. Fish Bull. 74:18-26. Murphy, M. D., and R. G. Taylor. 1991. Direct validation of ages determined for adult red drums from otolith sections. Trans. Am. Fish. Soc. 120:267-269. Nesbit, R. A. 1954. Weakfish migration in relation to its conservation. U.S. Fish. Wildl. Serv., Spec. Sci. Rep. Fish. 115, 81 p. Pearson, J. C. 1932. Winter trawl fishery off the Virginia and North Carolina coasts. U.S. Bur. Fish. Invest. Rep. 10,31 p. 1941. The young of some marine fishes taken in lower Chesapeake Bay, Virginia, with special ref- erence to the gray sea trout Cynoscion regalis (Bloch and Schneider). U.S. Fish Wildl. Serv. Fish Bull. 50:79-102. Perlmutter, A., W. S. Miller, and J. C. Poole. 1956. The weakfish (Cynoscion regalis ) in New York waters. N.Y Fish Game J. 3:1-43. Ricker, W. E. 1992. Back-calculation of fish lengths based on pro- portionality between scale and length incre- ments. Can. J. Fish. Aquat. Sci. 49:1018-1026. Rothschild, B. J., P. W. Jones, and J. S. Wilson. 1981. Trends in Chesapeake Bay fisheries. Trans. N. Am. Wildl. Conf. 46:284-298. SAS. 1988. SAS/STAT User's Guide, Release 6.03 ed. SAS Institute Inc., Cary, NC, 1029 p. Shepherd, G. R. 1988. Age determination methods for northwest Atlantic species, weakfish Cynoscion regalis. In J. Penttila and L. M. Derry (eds.), Age determina- tion methods for northwest Atlantic species, p. 71- 76. NOAA Tech. Rep. NMFS 72, 135 p. Shepherd, G. R., and C. B. Grimes. 1983. Geographic and historic variations in growth of weakfish, Cynoscion regalis, in the Middle At- lantic Bight. Fish. Bull. 81:803-813. 1984. Reproduction of weakfish, Cynoscion regalis, in the New York bight and evidence for geographi- cally specific life history characteristics. Fish. Bull. 82:501-511. Simkiss, K. 1974. Calcium metabolism of fish in relation to ageing. In T B. Bagenal (ed), The ageing offish, p. 1-12. The Gresham Press, Old Woking, Sur- rey, England. Smale, M. A., and W. W. Taylor. 1987. Sources of back-calculation error in estimat- ing growth of lake whitefish. //; R. C. Summerfelt andG.E. Hall (eds.), Age and growth of fish, p. 189- 202. Iowa State Univ. Press, Ames. Smith, C. L. 1983. Summary of round table discussions on back calculation. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 8:45-47. Taylor, H. F. 1916. The structure and growth of the scales of the squeteague and the pigfish as indicative of life history. Bull. U.S. Bur. Fish. 34:285-330. Villoso, E. P. 1989. Reproductive biology and environmental con- trol of spawning cycle of weakfish, Cynoscion regalis (Bloch and Schneider), in Delaware Bay. Ph.D. diss., Univ. Delaware, Newark, 295 p. Welsh, W. W., and C. M. Breder Jr. 1923. Contributions to life histories of Sciaenidae of the eastern United States coast. Bull. U.S. Bur. Fish. 39:141-201. Whitney, R. R., and K. D. Carlander 1956. Interpretation of body-scale regression for computing body length of fish. J. Wildl. Mgmt. 20(11:21-27. Wilk, S. J. 1979. Biological and fisheries data on weakfish, Cynoscion regalis (Bloch and Schneider). U.S. Dep. Commer., NOAA, NMFS, NEFC, Sandy Hook Laboratory, Tech. Ser. Rep. No. 21, 49 p. Williams, T., and B. C. Bedford. 1974. The use of otoliths for age determination. In T B. Bagenal (ed), The ageing offish, p. 114-123. The Gresham Press, Old Woking, Surrey, England. Abstract. — In this study of the feeding habits of the dusky dolphin, Lagenorhynchus obscurus, stom- ach content samples were collected from dolphins caught by an artisanal fishery operating along the central coast of Peru. Collec- tions were made from three fish- ing ports, Pucusana, Ancon, and Cerro Azul, during the summers and winters of 1985 and 1986. Overall, the anchoveta, Engraulis ringens, the most abundant verte- brate in Peruvian coastal waters, was the principal prey of dusky dolphins with respect to each of four different measures of dietary importance. Anchoveta was also the dominant prey in both seasons of both years, and for all reproduc- tive classes of dusky dolphins. Other prey species commonly found in dolphin stomachs were horse mackerel, Trachurus symmetricus, hake, Merluccius gayi, sardine, Sardinops sagax, Patagonian squid, Loligo gahi, and jumbo fly- ing squid, Dosidicas gigas. Regres- sions of body size on otolith or squid beak dimensions were used to es- timate lengths and weights of an- choveta and some other prey. All prey species averaged less than 30 cm in estimated length and 300 g in weight. Estimated total lengths of anchoveta consumed as prey in- creased between seasons in 1985 and between years, paralleling the lengths of anchoveta taken by the purse-seine fishery. However, esti- mated total lengths of anchoveta eaten by dusky dolphins were con- sistently smaller than lengths of those caught by the fishery. Feeding habits of the dusky dolphin, Lagenorhynchus obscurus, \n the coastal waters of central Peru Jeff McKinnon Department of Zoology, The University of Guelph Guelph, Ontario, Canada NIG 2W1 Present address The Biological Laboratories, Harvard University 1 6 Divinity Ave, Cambridge, MA 02 1 38 Manuscript accepted 11 November 1993. Fishery Bulletin 92:569-578 ( 1994). The dusky dolphin, Lagenorhyn- chus obscurus, is common in the coastal waters of New Zealand, South America, and South Africa, but like most species of small cetacea from the Southern Hemi- sphere, its feeding habits are poorly described (Gaskin, 1982; Goodall and Galeazzi, 1985). The squid, Nototodarus sloanei, and fish have been reported as prey in New Zealand waters (Gaskin, 1972), whereas in Argentina Wiirsig and Wiirsig ( 1980) observed dusky dol- phins feeding on southern anchovy, Engraulis anchoita. Prior to the present study, little was known of the feeding habits or natural history of the dusky dolphin in Peruvian waters, although large numbers of dusky dolphins were being taken in an artisanal fishery ( Read et al., 1988; Van Waerebeek and Reyes, 1990). In Peru, the dusky dolphin is found in the waters of the coastal upwelling system (McKinnon, 1988), which has been extensively studied by those involved in man- aging the system's various fisheries (Pauly and Tsukayama, 1987; Pauly et al., 1989). The cool waters of the coastal upwelling region constitute one of the most productive areas of the world ocean (Ryther, 1969), but oceanographic conditions and the abundance and distribution of fishes can vary greatly within and between years, especially when El Nino's occur (Pauly and Tsukaya- ma, 1987; Pauly et al., 1989). The objective of the present study was to characterize, in terms of both species composition and prey size, the feeding habits of the dusky dol- phin in Peruvian waters. Seasonal and annual variation in feeding habits was also investigated and compared with abundance data for important prey species during the same periods. In addition, potential dietary differences among dolphins of different reproductive states were examined, as feeding habits of lac- tating females differ from those of nonlactating females and males in some small cetaceans (Bernard and Hohn, 1989; Recchia and Read, 1989). Methods Data on feeding habits were ob- tained by analysis of undigested hard parts of prey, specifically fish otoliths and squid beaks, from stom- ach content samples collected through the fishery. Dolphins were usually captured by artisanal fish- ermen in gill nets set from dusk to dawn within the coastal upwelling zone (Read et al., 1988; Van Waere- beek and Reyes, 1990), but six stom- ach samples from a single landing by a purse seiner, in the summer of 1985, were also included. All samples were collected at ports along the central coast of Peru, where the largest dolphin catches occurred during the present 569 570 Fishery Bulletin 92(3), 1994 Figure 1 Map of Peruvian coastline, showing study sites. The inset shows the area in relation to the remainder of South America (Read et al., 1988). study (Read et al., 1988). Pucusana, a small fishing town approximately 50 km south of Lima (Fig. 1) was the principal collecting site. Stomach content samples were also collected from the nearby ports of Cerro Azul and Ancon, about 70 km south and 90 km north of Pucusana, respectively (Fig. 1). Dolphin landings at any given port were highly variable (Read et al., 1988), therefore collecting efforts were concentrated at the port or ports where the most dolphins were being caught. The collecting periods were the aus- tral summers (1 January through 31 March) and winters (1 July through 30 September) of 1985 and 1986. Field procedures Standard length, measured in a straight line along the main axis of the body from the tip of the upper jaw to the notch of the flukes, was taken for each specimen upon being landed. Females were pressed above the nipples and expression of milk, indicating lactation, was noted. When dolphins were eviscer- ated, female reproductive tracts were removed and checked for the presence of a fetus. Ovaries were in- spected for corpora lutea and albicantia, then pre- served in 10% formalin. Collection of stomach contents also began upon evisceration, between approximately 6 and 48 hours after capture. Each of the fore-, main, and pyloric stomachs was separately rinsed through a series of three brass sieves (Treacy and Crawford, 1981; Murie and Lavigne, 1985) of mesh diameters 4.75, 1.40, and 0.425 mm. The sieved contents were then placed in deep, water-filled plastic trays so that any remain- ing flesh could be skimmed off. Fish otoliths and clean squid beaks were retrieved and stored in 5-10% al- cohol; squid beaks with flesh still attached were stored in either 70% alcohol or 10% formalin. The forestomach consistently contained the least digested contents and the greatest volume, so only material from that chamber was later quantified and analyzed (see also Perrin et al., 1973; Bernard and Hohn, 1989). Reproductive status Reproductive status of females was classified after macroscopic examination of gonads and accessory reproductive tissues. Males were classified only as to sex. Females were defined as 1 ) immature, if their ovaries lacked corpora lutea and albicantia; 2) preg- nant, if a fetus was visible in the uterus; 3) lactat- ing; 4) simultaneously pregnant and lactating; or 5) resting, if corpora lutea or albicantia were present but there was neither a fetus visible in the uterus nor evidence of lactation ( Perrin and Donovan, 1 984 ). The "resting" category may have included females with small embryos not detected during field inspec- tions, in addition to individuals actually between reproductive cycles ( Perrin and Donovan, 1984 ). Data were incomplete for several females; therefore they were classified as "unknown females." No informa- tion on gender was available for several additional samples; their sex and reproductive status were clas- sified as "unknown." Identification of prey and calculation of measures of relative importance A reference collection of otoliths from common Peru- vian marine fishes was made from specimens pur- chased at local markets and identified with Chirichigno's ( 1978) key. Otoliths from stomach con- tents were identified by comparison with this collec- tion and collections belonging to the Instituto del Mar del Peru (IMARPE) and to P. Majluf (University of Cambridge). Squid species were identified from their beaks by using published keys (Wolff, 1984; Clarke, 1986), through reference to beaks from local squids identified and supplied by F. Cardoso (Museo de McKinnon: Feeding habits of Lagenorhynchus obscurus 571 Historia Natural de la Universidad Nacional Mayor de San Marcos, Lima), and with the assistance of S. Candela (University of Miami, Florida). Counts and measurements of undigested hard parts of prey allowed calculation of several measures of the relative importance of each prey species. The simplest measure, "percent frequency of occurrence", was defined as 100 multiplied by the number of stom- achs in which a prey species was present/the total number of stomachs in the sample, excluding empty stomachs. "Percent total numbers," was defined as 100 multiplied by the number of individuals of a spe- cies of prey/the sum of individuals for all prey spe- cies (Frost and Lowry, 1980). The number of indi- viduals of each prey species in a sample was esti- mated by dividing the count of its otoliths (for fish) or squid beaks (for squid) by two (Frost and Lowry, 1980). Lengths and weights of consumed fish were esti- mated by using regressions involving fish length and otolith length, or fish weight and either fish length or otolith length. Calculations for anchoveta, Engraulis ringens. Pacific Sardine, Sardinops sagax, hake, Merluccius gayi, and horse mackerel, Track- urus symmetricus, followed Chirinos and Chuman (1968), Samame (1977), McKinnon (1988), and Hawes (1983), respectively, except that length was estimated for T. symmetricus by using L=4.37xW1/3, and weights of E. ringens and S. sagax were esti- mated by using W=0.007xL3 and W=0.015xL3, where W=weight (g) and L=total length (cm).1 For each fish species, ten randomly selected otoliths from each stomach were measured (Murie, 1984). If fewer than ten suitable otoliths were present, all those available were utilized. Only otoliths with minimal degradation were measured. Degradation was apparent from a loss of detail, par- ticularly the loss or reduction of spines and lobulations along the edges of the otoliths (Frost and Lowry, 1986). Squid mantle lengths were estimated for each squid species by using linear regressions of mantle length on rostrum length and squid weights from regressions of loge weight on log rostrum length (Wolff, 1984). Regressions were not available for patagonian squid, Loligo gahi, so regression equa- tions for Loligo opalescens, a closely related species, were used (Wolff, 1984). For each squid species, ten randomly selected beaks were measured from each stomach, unless fewer than ten were present, in which case all were utilized. A mean individual weight (MIW) was calculated for each prey species in each stomach and then used 1 Pauly, D. International Center for Living Aquatic Resources Management, Manila. Personal commun., 1985. in estimating the percent weight contribution of each prey species to the dusky dolphin's diet. The MIW was usually the mean of the regression-estimated weights of individuals of a given prey species in a particular stomach, unless all hard parts were too degraded to permit reliable measurement, in which case an overall MIW, the mean of all regression-esti- mated lengths of that species in all stomachs with measurable hard parts, was employed. For ancho- veta, however, enough measurable otoliths were available from stomach samples to permit statisti- cal analyses by year and season of capture. Overall MIW's for anchoveta were therefore calculated for each group of stomachs within which analyses re- vealed no significant differences (for example, the summer of 1985; see Results). The total weight of each species of squid or fish present in each dolphin stomach was estimated by multiplying the number of individuals present by the appropriate MIW value. The percent weight of each prey species in the dusky dolphin's diet was calcu- lated by using weights summed over all stomachs, as 100 multiplied by the total weight of each prey species/the total weight of all prey present. Species for which regression-estimates of length and weight were not available were excluded from these calculations. The percent gross energy contribution of each spe- cies was defined as 100 multiplied by gross energy of the prey species/summed gross energy of all prey consumed, for all stomach content samples. The gross energy available from a prey species is the caloric density (kcal-g-1)xweight (g) consumed. Caloric den- sity (CD) values were obtained from the literature for each prey species, either directly from bomb-calo- rimetric analyses or indirectly from data on proxi- mate composition, by using CD's for fat, carbohy- drate, and protein of 9.4, 4.15, and 5.65 kcalg-1, re- spectively (Pike and Brown, 1984). By using published data from non-El Nino years only, CD values for anchoveta were calculated as 1.589 kcalg-1 for the summer and 1.548 kcalg-1 for the winter (Lam, 1968). Data were unavailable for S. sagax, but like E. ringens it is a clupeoid, and the reproductive seasons of the two species are similar (Muck et al., 1987; Pauly and Soriano, 1987), there- fore S. sagax was assigned the same seasonal values as E. ringens. A value of 1.244 kcalg1 was calcu- lated for T. symmetricus by using proximate compo- sition values from the related T. trachurus (Sidwell, 1981). Similarly, a value of 1.158 kcalg1 for M. gayi was based on equivalent data from M. productus (Sidwell, 1981). For Loligo gahi, 0.968 kcal-g-1 was obtained from proximate composition data in Croxall and Prince (1982). No published values were avail- able for Dosidicas gigas, so a mean ommastrephid 572 Fishery Bulletin 92(3). 1994 estimate of 0.922 kcalg l was calculated from bomb calorimetric and proximate composition data in Croxall and Prince (1982), Vlieg (1984), and Clarke etal. (1985). Statistical analyses Seasonal and annual variation in the size of con- sumed anchoveta were analyzed with a two-way AN OVA of the mean estimated lengths of the ancho- veta in each dolphin stomach, weighted by the num- ber of otoliths measured (SAS, 1985 and 1987). Be- cause of the unbalanced design, sums of squares and F-values were calculated by using the "Type III sum of squares" (SAS, 1985). Sidak adjusted <-tests were used for comparisons among pairs of means (Sokal and Rohlf, 1981; SAS, 1987). Log-linear analyses were conducted on frequency of occurrence data, also by using SAS (1987), and the resulting G-statistics were tested for significance following Sokal and Rohlf (1981). Data from different locations were pooled for analy- ses of the effects of year and season on dusky dol- phin feeding habits. This was appropriate because the two most distant of the three ports are separated by only about 160 km (Fig. 1 ), and all three ports are found along the central portion of the Peruvian coast. This region is relatively homogeneous and often treated as a single unit in analyses of oceanographic processes and fish populations (e.g. Brainard and McLain, 1987; Pena et al., 1989). It was necessary to pool data from all locations and collection periods to obtain sample sizes sufficient for analyses of prey oc- currence with respect to dolphin reproductive status. Results One hundred and thirty-six stomach samples con- tained recognizable hard parts and were used in sub- sequent analyses (Table 1). Six additional stomachs were empty, containing no recognizable hard parts at the time of their collection. There were no obvious patterns in the years, seasons, or locations in which the empty stomachs occurred, or the reproductive status of the individuals from which they were ob- tained ( Table 1 ). Eight additional samples contained otoliths at the time of their collection, according to field notes, but lacked otoliths when examined in the laboratory several months later. These otoliths may have dissolved during storage; therefore these samples were not included in analyses. Species included in percent weight and percent gross energy calculations — E. ringens, T. symmet- ricus, M. gayi, S. sagax, L. gahi and D. gigcis — repre- sented the vast majority of prey, over 98% of the to- tal number consumed. Other species, anchoa, Anchoa sp., blackruff, Seriolella violacea, a flyingfish, Hirundichthys sp., and deepsea smelt, Leuroglossus urotramus, were found in only trace amounts in stom- ach contents (Table 2). Anchoveta was the most important prey species by all measures of relative importance. It accounted for 92.5% of all dusky dolphin prey items by total num- bers and was present in 97.8% of stomachs. By weight, anchoveta accounted for 83.8% of prey, and by gross energy, 87.3% (Table 2). No other prey species accounted for more than 5.1% of prey by weight, 2.5% by total numbers, or 4.0% by gross energy, or was found in more than 26.5% of the stomachs examined (Table 2). Table 1 Distribution of stomach content samples containing otoliths or squid beaks, or both, collected from dusky dolphins, Lagenorhynchus obscurus, by year, season, and dolphin reproductive status. Numbers in parentheses represent number of empty stomachs containing neither otoliths nor squid beaks. Season Location Reproductive status' Year ImmF's RestF's PregF's LactF's V iknF's Males Unkn's Total 1985 Summer Pucusana 4 (i 0 1 n L0 0 L5 1985 Winter Pucusana () 0 0 1 n (i ii 1 1985 Winter Ancon t 7 0 8(ll 4 19(1) ii 42(21 1986 Summer Pucusana 1 1 3 5 II 10 (' 20 1986 Summer Cerro Azul i) ii 5 1 0 4i 1 i ii 101 1 . 1986 Winter Pucusana 4 5 41 1 i 5 n 12(1) 4 34(21 1986 Winter Cerro Azul 1 2 6 1 0 4(11 0 14(11 lmmF's= immature females: RestF*s=resting females; PregF's=pregnant females, LactF's =1 acta ting females (including simultaneously pregnanl and lactating females; two were collected, one at Ancon in the winter of 1985, one at Pucusana in the summer (if 1986); UnknF's= females of unknown reproductive status; Males=all males; Unkn's=individuals of unknown sex McKinnon: Feeding habits of Lagenorhynchus obscurus 573 By weight, anchoveta accounted for more than half of all prey con- sumed in every collection period and its lowest frequency of occur- rence was 93.3% (Table 3). The percent weight of T. symmetricus in dusky dolphin diets was always low, ranging from 09c in the win- ter of 1985 to a maximum of 14.5% in the summer of 1986, but fre- quency of occurrence was more variable, ranging from 0% to 53.3% (Table 3). In a log-linear analysis, the three-way interac- tion between year, season, and fre- quency of occurrence of T. symmetricus in stomachs was not statistically significant (G=1.91, df=l, P>0.10), nor was the inter- action between season and fre- quency of occurrence significant (G=2.26, df=l, P>0.10). The large increase in frequency of occurrence of this prey from 1985 to 1986, however, resulted in a highly sig- nificant interaction between fre- quency of occurrence and year (G=37.76,df=l,P<0.001). The only other prey found sufficiently often in stomachs for statistical testing of frequency, but not so often that insufficient variation was present (as for anchoveta), wasL. gahi, for which log-linear analysis yielded a significant three-way interaction between frequency of occurrence, year, and season of collection (G=13.44, df=l, P<0.001). Thus neither season nor year exerted a clear, independent effect on con- sumption of L, gahi, although there was considerable variation among collecting periods ( Table 3 ). Reproductive status did not have any obvious effect on dusky dolphin feeding habits (Table 4). The fre- quencies of occurrence of T. symmet- ricus and L. gahi did not differ sig- nificantly between either lactat- ing females and pooled nonlactating mature females (G=0.39, df=l,P>0.5; G=1.06, df=l, P>0.3, respectively for each prey species ) or between lactating females and all other individuals pooled (G=0.29, df=l, P>0.5; G=0.17, df=l, P>0.5, respectively). Anchoveta were con- sumed by both sexes almost without exception (Table 4). Table 2 Relative importance of prey species of the dusky dolphin, Lagenorhyiuhus obscurus, from the coastal waters of central Peru in the summers and win- ters of 1985 and 1986 (n=136 stomachs, 9,137 individual fish and squid). Percent weight and percent gross energy were calculated by using only the six most important prey species. % Frequency % Total % Gross Prey species of occurrence numbers' <7c Weight energy Engrau/is ringens 97.8 92.5 83.8 87.3 Trachurus symmetricus 26.5 2.5 3.5 2.9 Merluccius gayi 8.1 0.6 5.1 4.0 Sardinops sagax 4.4 0.2 2.2 2.3 Loligo gahi 19.1 2.2 3.7 2.4 Dosidicas gigas 11.0 0.4 1.6 1.0 Anchoa sp. 1.5 0.1 — — Seriolella violacea 1.5 0.0 — — Hirundichthys sp. 0.7 0.0 — — Leuroglossus umtramus 0.7 0.5 — — Unknowns2 9.6 1.1 — — Percent total numbers values were sometimes very low, e.g. 0.01% or only 0.09r to an accu- racy of one decimal place, even when a prey species was present in more than one stomach and its percent frequency of occurrence was greater than 1%. There appeared to be at least eight species offish represented among the otoliths which could not be identified. Table 3 Relative dietary contribution of prey species commonly eaten by tl dolphin, Lagenorhynchus obscurus, for each combination of year and e dusky season. Prey species % Total weight {% Frequency of occurence) Summer 1985 (ra=15)] Winter 1985 (rt=43l Summer 1986 Winter 1986 (/)=30) (n=48) E. ringens 76.6 (100.0) 86.4 (97.7) 66.0 (93.3) 91.4 100.0) T. symmetricus 0.2 (6 7 ) 0.0 (0.0**) 14.5 (53.3**) 3.4 (39.6**) M. gayi 22.0 (13.3) 3.1 (9.3) 6.9 (16.7) 0.0 (0.0) S. sagax 1.0 (6.7) 0.4 (2.3) 0.9 (3.3) 4.9 (6.3) L. gahi 0.1 (6.7* : ) 9.6 (34. 9! : i 2.3 (26.7**) 0.2 (4.2**) D. gigas 0.0 (0.0) 0.5 (20.9) 9.4 (16.7) 0.1 (2.1) 1 rc=Number of dusky dolphins from which stomach cont ' : =Significant differences present atP<0.001. ent samples were collected Prey size All dusky dolphin prey species for which lengths and weights could be estimated averaged less than 30 cm (mantle length for the squids, fork length for M. gayi, and total length for all others) and 300 g. The 574 Fishery Bulletin 92(3). 1994 Table 4 Percent composition by weight and frequency of occurrence of prey found in stomach contents of dusky dolphins, Lagenorhynchus obscurus, classified by reproductive status. All collection periods are pooled (summers and win- ters of 1985 and 1986). Reproductive status' n Weight- l '% Frequency of occurrence ) 3. ringens T. symmetricus 74.3 4.6 (92.9) (14.31 82.6 0.9 i kki in (26.7) 85.7 0.7 (100.0) (33.3) ST 1 0.7 (100.0) (22.7) 81.8 7.5 (96.6) (30.5) M. gayi S. sagax L. gahl D. gigas Imm. Fem.'s 14 Rest. Fem.'s 15 Preg. Fem.'s 18 Lact. Fem.'s 22 Males 59 7.3 (14.31 0.0 (0.0) 0.9 (5.6) 8.3 (13.6) 5.9 (8.5) 0.0 (0.01 15.9 (6.7) 3.0 (11.1) 0.6 (4.5) 0.4 (1.7) 12.9 0.8 (21.4) (28.6) 0.6 0.0 (6.7) (0.0) 0.8 8.9 (16.7) (16.7) 3.0 0.4 (22.7) (13.6) 4.0 0.4 (22.0) (6.8) ' Imm. Fem.'s=immature females; Rest. Fem.'s=resting females; Preg. Fem.'s=pregnant females; Lact. Fem.'s=lactating females, including those simultaneously pregnant and lactatmg; Males=all males; individuals of unknown reproductive status omitted. 2 May not total 100 in each row due to rounding. most common prey species, E. ringens, T. symmetricus, and L. gahi, averaged less than 20 cm and 100 g (Table 5). The regression-estimated lengths of consumed anchoveta varied sig- nificantly with year, season, and the interaction between year and sea- son (F=416.06, 62.56, and 35.42, respectively; df=l, 67, P<0.0001; Table 6). In comparisons between pairs of means, anchoveta were found to be significantly larger in the summer and the winter of 1986 than in either the summer or the winter of 1985 (Table 6). In 1985, otolith lengths were significantly different between summer and winter, but not in 1986 (Table 6). Mean total lengths of anchoveta from fishery samples (for each combination of year and season) were positively correlated with mean regression-estimated lengths of anchoveta consumed by dusky dolphins (r=0.98, df=2, P<0.05; Table 6). Mean esti- mates of total length for anchoveta consumed by dusky dolphins were consistently 1.4-2.0 cm smaller than mean total lengths of anchoveta taken by the fishery, however (paired /-test, /=12.4, df=3, P<0.01). Discussion The prey of the dusky dolphin in Peruvian coastal Table 5 Mean lengths (cm) and weights (g), as estimated from regressions on otolith length/radius and squid beak rostrum length, offish and squid species com- monly found in the stomachs of the dusky dolphin, Lagenorhynchus obscurus, landed along the central Peruvian coast in 1985 and 1986. Estimatec mean length (SE) Estimated mean weight (SE) Prey species No of hard pa rts'' (No. of stomachs)'' No. of hard parts4 (No. of stomachs)'' E. ringens' 13.4 (0.14) 17.3 (0.49) 593 (71) 593 (71) S. sagax' 25.1 (0.32) 237.0 (8.67) 4 (2) 4 (2) M. gayi2 25.2 (4.62) 198.0 1104.0) 18 (5) 18 (5) T. symmetricus' 11.5 (1.16) 25.6 18.67) 34 (13) 34 (13) L. gahi" 13.0 (0.27) 33.3 (1.23) 97 (26) 97 (26) D. gigas" 13.4 (1.00) 72.7 (16.41 34 (15) 34 (15) Total length. Fork length. Mantle length. Number of hard parts measured. Number of stomachs from which samples were taken waters can be characterized as schooling, small to medium size, pelagic or semi-pelagic species (M. gayi is usually demersal but sometimes forms large pe- lagic schools [Mejia and Jordan, 1979]). Anchoveta was typical and was unequivocally the most impor- tant prey species in 1985-86. It was most important by all measures of consumption and constituted al- most 90'/t of the dusky dolphin's diet by percent gross McKinnon: Feeding habits of Lagenorhynchus obscurus 575 Table 6 Mean estimated total lengths of anchoveta, Engraulis ringens, consumed by dusky dolphins, Lagenorhynchus obscurus, in central Peruvian coastal waters in 1985 and 1986, compared to mean lengths of anchoveta taken in the purse-seine fishery. Collection source Summer 1985 Winter 1985 Summer 1986 Winter 1986 Stomachs of dusky dolphins Mean estimated total length of anchoveta' (SE) Number of otoliths measured Number of stomachs from which otoliths were collected Purse-seine fishery Mean total length of anchoveta2 11.5° (0.14) 12. 96 (0.09) 14.1' (0.10) 14.3C(0.07) 12 150 103 228 12 18 14 27 13.5 14.6 15.5 16.3 ' Based upon measurements of otoliths in dusky dolphin stomachs. 2 Mean total length of anchoveta taken in the fishery, calculated from data in Pauly and Palomares ( 1989); SE's and /Vs unavailable. ' Means with different letters in the superscript differed significantly (Sidak adjusted (-tests; (>8.41, P<0.001 in all cases) while those with the same letter did not U=1.42, P>0.5). energy, usually considered the best measure of rela- tive prey importance (La vigneetal., 1982). Few data are available on feeding habits for other regions, but in Argentina dusky dolphins also fed on a species of anchovy (Wiirsig and Wiirsig, 1980). There were no consistent seasonal patterns in prey consumption in the present study. Rather, anchoveta was the most important prey species in both seasons of both years, probably owing, in part, to its rela- tively high abundance throughout the study period (Pauly and Palomares, 1989). Consumption of T. symmetricus was more variable and opportunistic. In 1985 T. symmetricus was almost absent from stom- ach samples, but it was a major prey item in 1986 when unusually large numbers of juveniles, similar in size to other important dusky dolphin prey, were observed in the coastal waters of central and north- ern Peru (IMARPE et al., 1986). Other species, par- ticularly L. gahi, varied greatly among collection periods in their importance as prey, but did not ex- hibit consistent seasonal or annual patterns. In examining the effects of season, year, and other variables on diet, only frequency of occurrence was analyzed statistically because analyses of other mea- sures of prey importance involve excessive violations of the assumptions underlying most statistical tests (Recchia and Read, 1989). Percent-weight estimates were used for qualitative comparisons among collec- tion periods and reproductive classes, rather than percent gross energy values, because the latter were very gross approximations. Anchoveta was the main prey of all reproductive classes of dusky dolphins. In contrast, in the eastern tropical Pacific most spotted dolphins, Stcnella attenuate, eat mainly ommastrephid squids whereas lactating females eat principally fish (Bernard and Hohn, 1989). The greater energy and water require- ments of lactating females may force them to feed on fish, which contain more energy and water per unit weight than squid or, alternatively, the presence of a calf may prevent females from feeding at the depths at which squid occur (Bernard and Hohn. 1989). In Peru, lactating dusky dolphins were apparently able to satisfy their energy and water requirements, as did other females, males and juveniles, by feeding on the abundant, high energy anchoveta ( Lam, 1968 ). Like the fishermen, fur seals, and seabirds of the Peruvian coast, dusky dolphins were somewhat op- portunistic in their feeding in respect to the sizes of the anchoveta they preyed upon, taking more of the more abundant size classes (Muck and Pauly, 1987; Majluf, 1989). Anchoveta consumed by dusky dol- phins were consistently 1.4-2.0 cm smaller, however, than those taken in the fishery, perhaps because of the fishery's bias towards larger anchoveta (Palomares et al., 1987). Alternatively, the slight degradation present in some of the otoliths from which anchoveta sizes were estimated may have re- sulted in underestimation of total lengths (Recchia and Read, 1989). The observed discrepancy may also be due to variation in the relationship between otolith size and body size, which can differ among years be- cause of variation ingrowth rates offish (McKinnon, 1988; Reznick et al., 1989; Secor and Dean, 1989; Campana, 1990). Relative consumption of fishes with small otoliths, such as anchoveta, may have been underestimated. Large otoliths and squid beaks are less easily dis- solved by stomach acids than are smaller otoliths (Hawes, 1983; Bigg and Fawcett, 1985; da Silva and 576 Fishery Bulletin 92(3), 1994 Neilson, 1985; Recchia and Read, 1989) and squid beaks are less easily passed through the digestive tract than are otoliths (Hawes, 1983; Bigg and Fawcett, 1985). This potential bias tends to strengthen the principal finding of this study: an- choveta was by far the most important prey species of all reproductive classes of Peruvian dusky dolphins in the summers and winters of 1985-86. Acknowledgments I was supported throughout this study by a scholar- ship from the Natural Sciences and Engineering Research Council (NSERC) of Canada. The research was funded by grants to David Gaskin (University of Guelph, Canada) from the United Nations Envi- ronmental Program and the International Union for the Conservation of Nature, and NSERC. Andy Read, Julio C. Reyes, Koen Van Waerebeek, Linda Lehman, Maria Valle-Riestra, Steve Farnworth, and Mark Chandler assisted with field work. Without Julio Reyes, in particular, the project would have been much less successful. Patricia Majluf, Tony Luscombe, Leonardo Mendizabal Manrique, and the Farnworth, Malasquez, Reyes- Robles, Valle-Riestra, and Yonge families were hos- pitable and helpful in many ways. Jorge Zuzunaga, Marco Espino, and especially Juan Velez (all of IMARPE) provided much useful assistance and in- formation. I thank the Capitania, Empresa Peruana de Servicios Pesqueros, and Ministerio de Pesqueria for allowing us to work in the fish markets. Sue Candela identified many squid beaks, and Franz Cardoso was also instrumental in these iden- tifications. Dan Ryan and O.B. Brian helped with statistics. Kenny Richard assisted with data input and checking. David Gaskin, Ronald Hardy, Linda Jones, Andy Read, and two anonymous reviewers provided many helpful comments on the manuscript. Literature cited Bernard, H. J., and A. A. Hohn. 1989. Differences in feeding habits between preg- nant and lactating spotted dolphins (Stenella attenuata). J. Mamm. 70:211-215. Bigg, M. A., and I. Fawcett. 1985. Two biases in diet determination of northern fur sea\s(Callorhiruis ursinus i. In J. R. Beddington, R. J. H. Beverton, and D. M. Lavigne (eds.). Ma- rine mammals and fisheries, p. 284-291. Allen and Unwin, London, 354 p. Brainard, R. E., and D. R. McLain. 1987. Seasonal and interannual subsurface tem- perature variability off Peru, 1952-1984. In D. Pauly and I. Tsukayama (eds.). The Peruvian an- choveta and its upwelling ecosystem: three decades of change, p. 14-45. ICLARM Studies and Re- views 15. ICLARM, IMARPE, GTZ, Manila, 351 p. Campana, S. E. 1990. How reliable are growth back-calculations based on otoliths. Can. J. Fish. Aquat. Sci. 47:2219-2227. Chirichigno F., N. 1978. Clave para identificar los peces marinos del Peru. Informe Inst. Mar Peru-Callao 44, Bibli- oteca, Instituto del Mar del Peru, P.O. Box 22, Callao, Peru, 387 p. Chirinos de Vildoso, A., and E. Chuman D. 1968. Validez de la lectura de otolitos para determinar la edad de la anchoveta iEngraulis ring- ens). Informe Inst. Mar Peru-Callao 22, Bibli- oteca, Instituto del Mar del Peru, P.O. Box 22, Callao, Peru, 34 p. Clarke, A., M. R. Clarke, L. J. Holmes, and T. D. Waters. 1985. Calorific values and elemental analysis of eleven species of oceanic squids (Mollusca: Cepha- lopoda). J. Mar. Biol. Ass. U.K. 65:983-986. Clarke, M. R. (ed.) 1986. A handbook for the identification of cephalo- pod beaks. Clarendon Press, Oxford, 273 p. Croxall, J. P., and P. A. Prince. 1982. Calorific content of squid (Mollusca: Cepha- lopoda). Br. Antarct. Surv. Bull. 55:27-31. da Silva, J., and J. D. Neilson. 1985. Limitations of using otoliths recovered in scats to estimate prey consumption in seals. Can. J. Fish. Aquat. Sci. 42:1439-1442. Frost, K. J., and L. F. Lowry. 1980. Feeding of ribbon seals tPhoca fasciata ) in the Bering Sea in spring. Can. J. Zool. 58:1601-1607. 1986. Sizes of walleye pollock, Theragra chalco- gramma, consumed by marine mammals in the Bering Sea. Fish. Bull. 84:192-197. Gaskin, D. E. 1972. Whales, dolphins and seals — with special ref- erence to the New Zealand region. St. Martin's Press, New York, 200 p. 1982. The ecology of whales and dolphins. Hein- emann, London, 459 p. Goodall, R. N. P., and A. R. Galeazzi. 1985. A review of the food habits of the small ceta- ceans of the Antarctic and the sub-Antarctic. In W. R. Siegfried, P. R. Condy, and R. M. Laws (eds. ), Antarctic nutrient cycles and food webs, p. 566- 572. Springer-Verlag, Berlin, 700 p. Hawes, S. D. 1983. An evaluation of California sea lion scat samples as indicators of prey importance. Unpubl. M.A. thesis, San Francisco State Univ., 50 p. McKinnon: Feeding habits of Lagenorhynchus obscurus 577 IMARPE, SENAMHI, DHNM, and IGP. 1986. Condiciones bio-oceanograficas y meteor- ologicas frente a la costa Peruana en Enero-Junio 1986. ERFEN Boletin 19:3-11, Comision Perma- nente del Pacifico Sur, Calle 76 No. 9, 88 Apartado Aereo 92292 Bogota, Colombia. Lam, R. 1968. Estudio sobre la variacion del contenido de grasa en la anchoveta Peruana iEngraulis ringens J). Informe Inst. Mar Peni-Callao 24. Biblioteca, Instituto del Mar del Peru, P.O. Box 22, Callao, Peru, 29 p. Lavigne, D. M., W. Barchard, S. Innes, and N. A. Oritsland. 1982. Pinniped bioenergetics. In J. G. Clarke (ed. ), Mammals in the seas, Vol. IV: small cetaceans, seals, sirenians and otters, p. 191-235. FAO Fish- eries Series Number 5, Rome, 531 p. Majluf, P. 1989. Reproductive ecology of South American fur seals in Peru. In D. Pauly, P. Muck, J. Mendo and I. Tsukayama (eds.), The Peruvian upwelling eco- system: dynamics and interactions, p. 332-343. Conference Proceedings 18. ICLARM, IMARPE, GTZ, 438 p. McKinnon, J. S. 1988. Feeding habits of three species of small ceta- ceans from the coastal waters of Peru. Unpubl. M.S. thesis, Univ. Guelph, 94 p. Mejia G., J., and R. Jordan S. 1979. La situacion actual del stock de merluza a setiembre de 1978. Informe Inst. Mar Peru-Callao 57, Biblioteca, Instituto del Mar del Peru, P.O. Box 22, Callao, Peru. 19 p. Muck, P., and D. Pauly. 1987. Monthly anchoveta consumption of guano birds, 1953 to 1982. In D. Pauly and I. Tsukayama (eds. ), The Peruvian anchoveta and its upwelling ecosystem: three decades of change, p. 219-233. ICLARM Stud- ies and Reviews 15. ICLARM, IMARPE, GTZ, 351 p. Muck, P., O. Sandoval de Castillo, and S. Carrasco. 1987. Abundance of sardine, mackerel and horse mackerel eggs and larvae and their relationship to temperature, turbulence and anchoveta biomass off Peru. In D. Pauly and I. Tsukayama (eds.). The Peruvian anchoveta and its upwelling ecosystem: three decades of change, p. 268-275. ICLARM Stud- ies and Reviews 15. ICLARM, IMARPE. GTZ, 351 p. Murie, D. J. 1984. Estimating food consumption of free-living harp seals, Phoca groenlandica (Erxleben 1777). Unpubl. M.S. thesis, Univ. Guelph, 97 p. Murie, D. J., and D. M. Lavigne. 1985. A technique for the recovery of otoliths from stomach contents of piscivorous pinnipeds. J. Wildl. Manage. 49:910-912. Palomares, M. L., P. Muck, J. Mendo, E. Chuman, O. Gomez, and D. Pauly. 1987. Growth of the Peruvian anchoveta iEngraulis ringens), 1953 to 1982. In D. Pauly and I. Tsukayama (eds. ), The Peruvian anchoveta and its upwelling ecosystem: three decades of change, p. 117-141. ICLARM Studies and Reviews 15. ICLARM, IMARPE, GTZ, 351 p. Pauly, D., and M. L. Palomares. 1989. New estimates of monthly biomass, recruit- ment and related statistics of anchoveta iEngraulis ringens) off Peru (4-14°S), 1953-1985. In D. Pauly, P. Muck, J. Mendo, and I. Tsukayama (eds.), The Peruvian upwelling ecosystem: dynamics and interactions, p. 189-206. Conference Proceedings 18. ICLARM, IMARPE, GTZ, 438 p. Pauly, D., and M. Soriano. 1987. Monthly spawning stock and egg production of Peruvian anchoveta iEngraulis ringens), 1953 to 1982. In D. Pauly and I. Tsukayama (eds. ), The Peruvian anchoveta and its upwelling ecosystem: three decades of change, p. 167-178. ICLARM Studies and Reviews 15. ICLARM, IMARPE, GTZ, 351 p. Pauly, D., and I. Tsukayama (eds.). 1987. The Peruvian anchoveta and its upwelling ecosystem: three decades of change. ICLARM Studies and Reviews 15. ICLARM, IMARPE, GTZ, 351 p. Pauly, D., P. Muck, J. Mendo, and I. Tsukayama (eds.). 1989. The Peruvian upwelling ecosystem: dynam- ics and interactions. Conference Proceedings 18. ICLARM, IMARPE, GTZ, 438 p. Pena, N., J. Mendo, and J. Pellon. 1989. Sexual maturity of Peruvian anchoveta iEngraulis ringens), 1961-1987. In D. Pauly, P. Muck, J. Mendo, and I. Tsukayama (eds.), The Pe- ruvian upwelling ecosystem: dynamics and inter- actions, p. 132-142. Conference Proceedings 18. ICLARM, IMARPE, GTZ, 438 p. Perrin, W. F., and G. P. Donovan. 1984. Report of the workshop. Rep. Int. Whaling Comm. Spec. Issue 6:1-24. Perrin, W. F., R. R. Warner, C. H. Fiscus, and D. B. Holts. 1973. Stomach contents of porpoise, Stenella spp., and yellowfin tuna, Thunnus albacares, in mixed species aggregations. Fish. Bull. 71:1077-1092. Pike, R. L., and M. L. Brown. 1984. Nutrition, an integrated approach, 3rd ed. John Wiley & Sons, New York, 1068 p. Read, A. J., K. Van Waerebeek, J. C. Reyes, J. S. McKinnon, and L. C. Lehman. 1988. The exploitation of small cetaceans in coastal Peru. Biol. Cons. 46:53-70. Recchia, C. A., and A. J. Read. 1989. Stomach contents of harbour porpoises, Phocoena phocoena (L.), from the Bay of Fundy. Can. J. Zool. 67:2140-2146. Reznick, D., E. Lindbeck, and H. Bryga. 1989. Slower growth results in larger otoliths: an experimental test with guppies iPoecilia reticulata ). Can. J. Fish. Aquat. Sci. 46:108-112. 578 Fishery Bulletin 92(3), 1994 Ryther, J. H. 1969. Photosynthesis and fish production in the sea. Science 166:72-76. Samame L., M. 1977. Determinacion de la edad y crecimiento de la sardina Sardinops sagax sagax (J). Boletin Inst. Mar Peni-Callao 3:95-112. Biblioteca, Institute del Mar del Peru, P.O. Box 22, Callao, Peru. SAS. 1985. SAS user's guide: statistics, version 5. SAS Institute, Cary, NC, 956 p. 1987. SAS/STAT guide for personal computers, ver- sion 6. SAS Institute, Cary, NC, 378 p. Secor, D. H., and J. M. Dean. 1989. Somatic growth effects on the otolith-fish size relationship in young pond-reared striped bass, Mor- one sa.xatilis. Can. J. Fish. Aquat. Sci. 46:113-121. Sidwell, V. D. 1981. Chemical and nutritional composition of fin- fishes, whales, crustaceans, mollusks and their products. Dep. Commer., NOAA Tech. Memo. NMFS F/SEC-11. Available from National Techni- cal and Information Service, 5285 Port Royal Rd., Springfield, VA 22161, 440 p. Sokal, R. R. and F. J. Rohlf. 1981. Biometry. W. H. Freeman and Company, New York, 859 p. Treacy, S. D., and T. W. Crawford. 1981. Retrieval of otoliths and statoliths from gas- trointestinal contents and scats of marine mammals. J. Wild]. Man. 45:990-993. Van Waerebeek, K., and J. C. Reyes. 1990. Catch of small cetaceans at Pucusana port, central Peru, during 1987. Biol. Cons. 51:15-22. Vlieg, P. 1984. Proximate composition of New Zealand squid species. N. Z. J. Sci. 27:145-150. Wolff, G. A. 1984. Identification and estimation of size from the beaks of 18 species of cephalopods from the Pacific Ocean. Dep. Commer., NOAA Tech. Rep. NMFS 17. Available from National Technical and Infor- mation Service, 5285 Port Royal Rd., Springfield, VA 22161, 50 p. Wiirsig, B., and M. Wiirsig. 1980. Behavior and ecology of the dusky dolphin, Lagenorhynchufs obscurus, in the South At- lantic. Fish. Bull. 77:871-890. Abstract. — This study exam- ines the early life history of a popu- lation of walleye pollock, Theragra chalcogramma (Pallas), that is found in Resurrection Bay, Alaska. Ichthyoplankton samples were taken at six stations in Resurrec- tion Bay during early May and early June 1989 along with hydro- graphic data. Standard lengths of all walleye pollock were measured, and subsamples from two stations were aged by using otolith incre- ments for growth rate and hatch date analysis. Abundances ranged from 60 to 575 larvae m~2 in May and from 0 to 10 larvae irr2 in June with densities of up to 12 larvae m~3 in May. The estimated growth rate was 0.18 mm/day. Back-calcu- lated hatch dates ranged from late March until early May; the median hatch date was 22 April. Compari- sons of abundance and growth rate to values from other habitats indi- cate that this deep fjord provides a suitable habitat for larval walleye pollock. Hydrographic data and lar- val size distribution suggest that advection plays a major role in de- termining the distribution of larvae in the fjord. Distribution, abundance, and growth of larval walleye pollock, Theragra chalcogramma, in an Alaskan fjord Franz-Josef Muter Brenda L. Norcross Institute of Marine Science, School of Fisheries and Ocean Sciences University of Alaska Fairbanks Fairbanks. AK 99775-1080 Manuscript accepted 13 December 1993. Fishery Bulletin 92:579-590 (1994). Fjords have long been recognized as nursery grounds for many commer- cially important fish species (De Silva, 1973; Lie, 1978; Carmo Lopes, 1979). Matthews and Heimdal ( 1980) in their review of food chains in fjords pointed out that many fjords along Scandinavian, Scottish, and North American coasts are highly productive areas. The pro- ductivity of fjords is often enhanced by hydrographic boundary condi- tions or land runoff that can in- crease nutrient levels (Matthews and Heimdal, 1980). Production in fjords may be further enhanced by upwelling conditions at their mouths. This is especially true for the southern coast of Alaska, where the relaxation of easterly winds in summer promotes coastal diver- gence and upwelling (Royer, 1982). Rogers et al. (1987) described the nearshore zone of the Gulf of Alaska as an important spawning or rearing area, or both, for several commer- cially important fish species, includ- ing walleye pollock, Theragra chalco- gramma. However, no work has been done to examine the dynamics of early life history stages of walleye pol- lock or other fishes in Alaskan fiords. We chose walleye pollock for this study because it was more abun- dant than any other species in Res- urrection Bay (Smith et al., 1991) and its development and early life history in other areas of the Gulf of Alaska are well known (Dunn and Matarese, 1987; Kendall et al., 1987; Kim, 1989). Furthermore, it is very important commercially, with annual landings off Alaska ex- ceeding one million metric tons ( Lloyd and Davis, 1989), and the walleye pollock resource shows high fluctua- tions in year-class strength ( Megrey, 1991), which creates a strong incen- tive to determine possible causes. Most of the research on pollock in the Gulf of Alaska has been focused on the Shelikof Strait region (Schu- macher and Kendall, 1991), while other areas along the Gulf, except for Auke Bay in Southeast Alaska (Haldorson et al. 1989, a and b), have received little attention. Al- though the Shelikof Strait spawn- ing area is believed to be the most important in the Gulf of Alaska (Hinckley et al., 1991), substantial pollock spawning occurs in other areas of the Gulf (Muter, 1992; Nor- cross and Frandsen1). Resurrection Bay shares many features with other embayments along the southcentral coast of Alaska and can be considered rep- resentative of the area. This study used growth analysis together with Norcross, B. L., and M. Frandsen. Distri- bution and abundance of larval fishes in Prince William Sound, Alaska, during 1989 after the Exxon Valdez oil spill. EVOS Sym- posium Proceedings. Am. Fish Soc. Sym- posium. In review. 579 580 Fishery Bulletin 92(3). 1994 distribution and abundance data to evaluate the role of this northern Gulf of Alaska fjord in the early life history of walleye pollock. Specifically, the objectives of this study were 1) to determine the distribution and abundance of walleye pollock larvae in a glaci- ated fjord, 2) to quantify growth rates of larvae within this fjord and compare growth rates to literature values from other areas, and 3) to estimate hatch dates of the observed population. Materials and methods Resurrection Bay is a fjord approximately 32 km long and 4-8 km wide, located within the coastal moun- tain range on the Kenai Peninsula on the south- central coast of Alaska (Fig. 1). The fjord's bathym- etry shows an inner basin with a maximum depth of 300 m, separated by a sill from the outer basin. The sill is located about 15 km from the fjord's mouth at the narrowest point, between our sampling stations RES 2.5 and RES 3 (Fig. 1), and rises to a depth of approximately 185 m. The outer basin is slightly shallower (265 m) than the inner basin and has an open connection with the shelf. Six stations were sampled along the fjord axis ( Fig. 1) during two cruises, 1-4 May 1989 and 7-9 June 1989. Ichthyoplankton samples for this study were collected from the RV Little Dipper, a 9-m aluminum boat. Horizontal plankton tows were taken at dis- crete depths by using a 1-m2 Tucker trawl, rigged with two 505-u mesh nets. Because no previous data Figure 1 Map of Alaska and Resurrection Bay including stations sampled were available we took samples throughout the wa- ter column. We tried to obtain at least one sample from each of the following depth strata per station: 0-15 m, 15-30 m, 30-50 m, 50-80 m, 80-150 m, and 150-m to the bottom. Because of weather and time constraints, fewer samples were taken at some sta- tions. Sample depths were initially estimated from wire angle and length of extended wire. Actual depths were recorded with an attached Seabird Seacat con- ducting-temperature-depth (CTD) (SBE 19) profiler and retrieved after completion of the cruise. The nets were rigged to a double tripper which allowed the second net to be opened and closed via a messenger from the surface. The net was towed for five minutes in the direction of tidal flow at a towing speed of 1.5 to 2.5 knots. Only daytime tows were made. Volume filtered during each tow was calculated from a TSK or General Oceanics flowmeter that was attached in a central position to the mouth of the net. Samples used for this analysis were immediately preserved in 509r isopropyl alcohol or 959f ethyl alcohol. The alcohol was renewed for each sample after 24 hours and after 2-3 days. Because differential shrinkage was observed between preservatives, only larvae preserved in isopropyl alcohol were used in size comparisons. A Seabird CTD Profiler was attached to the net during most tows to record conductivity, tempera- ture, and pressure throughout the tow. When no CTD data were recorded, depth was estimated from the wire angle and the length of extended wire. In addi- tion, CTD data were taken at each station and along cross-fjord transects through each station. Because of equip- ment failure, no temperature and salinity data were obtained during the June cruise. Addi- tional CTD-profiles for RES 2.5 and GAK 1 were obtained from a cruise on 6 April of the same year. Samples were sorted in the laboratory to isolate finfish lar- vae. Walleye pollock larvae were identified and measured to standard length (SL). Densities in larvaem ,! were calculated and abundance in larvaem '2 at each station was estimated by integrating larval densities over the water column by using ver- tical distribution profiles. Den- sity was set to zero at the sur- face and was assumed to change in a linear fashion between suc- cessive sampling depths. Be- Muter and Norcross: Distribution, abundance, and growth of Theragra chalcogramma 581 cause no replicate samples were taken, confidence limits could not be calculated. A Student f-test ( two-sample comparison ) or a one- way ANOVA followed by a Tukey multiple compari- son test (multiple samples) was employed to detect differences in mean standard length of larvae among different depths at the same station and among dif- ferent stations. Nonparametric tests were employed in addition to parametric test procedures when the assumptions for parametric tests were violated. The nonparametric tests used were a Kruskal-Wallis test (nonparametric analysis of variance), a Mann- Whitney test (two-sample comparison), and aTukey- type multiple comparison test (Zar, 1984). Differences in larval length among stations were examined by using the most shallow samples from each station (<22 m), thus minimizing potential er- rors resulting from differences in size due to vertical distribution. In addition, pollock lengths from all depths were pooled for each station and the mean of the pooled data was compared between stations. For all between-station comparisons, larval mean SL was corrected for the date of sampling by using growth rates obtained during this study. Ages of larvae were estimated from the number of otolith increments on sagittal otoliths as de- scribed in Kendall et al. (1987). Increments were independently counted a second time by the same reader. Readings were confirmed for a subsample of 20 otoliths by the Alaska Fisheries Science Cen- ter laboratory in Seattle, Washington. Only those independent readings that did not differ by more than one increment (in which case the higher count was used) were used for growth determination. Random subsamples of larvae from two stations, one in the inner basin ( RES 2 ) and one in the outer basin (RES 4) were aged. Only larvae from these stations could be aged because otoliths from all other samples showed signs of erosion. Larval growth rates were determined by fitting linear regression lines to length-at-age data. The linear regression equations describing growth were compared between stations to test for differ- ences in regression coefficients. Slopes and eleva- tions were compared by using Student's /-statis- tic (Zar, 1984). Hatch dates were estimated after correcting for mortality, because older fish in the sample experienced a higher cumulative mortal- ity than larvae hatched closer to the date of sam- pling. Following Yoklavich and Bailey (1990), we created a stepped, size-specific mortality function with rates of 0.1, 0.08, 0.06, 0.03 per day for fish <7, 7.01-10.0, 10.01-15.0 and 15.01-20.0 mm SL, respectively. The range of ages corresponding to each size range was calculated from the growth equation obtained in this study. The hatch date dis- tribution was then estimated by backcalculating numbers of larvae at hatching for each daily cohort with the above mortality rates. Results Hydrography of the fjord On 6 April 1989, temperatures at RES 2.5 (inner basin) and GAK 1 (mouth of fjord) increased with depth from approximately 4 C in the surface layer to almost 6°C below 200 m (Fig. 2). Between April and May 1989 the properties of the water masses inside and outside the fjord changed considerably. In April the upper 100 m were nearly homogenous, but a strong seasonal thermocline had developed between 10 and 20 m in early May. The surface tem- perature in May varied between 5.8CC at RES 2.5 and T C at RES 3 (Fig. 2 ). Temperature profiles in May showed a pronounced minimum of about 3.5°C to 4.5°C near 80 m. While temperatures in April in 6 April 1989 / /—GAK 1 — f-RES 4 100 - \^ f~ RES 2.5 200 \L 300 1-4 May 198^ 1 ■ 3 4 6 6 Temperature (°C) 3 4 6 6 7 Temperature (°C) '/*\ ' RES 25 GAK 1 — 100 \ 200 Y 6 April 1989 \ RES 2.5 RES 1.2.3.4 1'GAK 1 1-4 May 1989 28 29 30 31 32 33 34 26 29 30 31 32 33 34 Salinity (psu) Salinity (psu) Figure 2 Temperature and salinity profiles at six stations in Res- urrection Bay, Alaska, 6 April and 1-4 May 1989. 582 Fishery Bulletin 92(3). 1994 the upper 100 m did not differ by more than 0.5°C between RES 2.5 and GAK 1, the temperature dif- ference in May was almost 1.5°C. The water column was nearly isohaline in April: salinity increased approximately 1.5 psu (practical salinity unit) from surface to bottom. Salinity pro- files in May show a well-developed low salinity sur- face layer at four of the stations (Fig. 2), resulting from river runoff and snow melt. The surface salini- ties were 2 to 4 psu lower than in April. However, at RES 2.5 and GAK 1 the freshwater lens was much less developed than at the other stations. The sur- face layer salinity was above 31 psu and almost iden- tical at both stations. Below the halocline, salinities were very similar at all stations. Distribution and abundance In early May, walleye pollock larvae were caught at all stations and sampled at all depths. A total of 16,950 pollock larvae were collected in 39 tows at depths between 7 and 280 m. Larval densities ranged from 0.03 larvaem-3 (RES 4, 105 m) to a maximum of 11.9 larvaem-3 (RES 4, 26 m). Larvae were gener- ally concentrated in the upper 70 m (Fig. 3). Maxi- mum densities occurred at depths between 18 and 30 m at all stations, except GAK 1, and ranged from 2.2 larvae-m-3 to 11.9 larvae-m-3. Pollock larvae were distributed deeper in the water column outside the sill, at stations RES 4 and GAK 1, than at stations inside the sill (Fig. 3). Between May and June, larval densities decreased by two orders of magnitude and ranged from 0 to 0.4 larvae-m-3 in early June (Fig. 4). In June, a total of 420 walleye pollock larvae were collected in 45 tows at depths between 5 and 250 m. Only tows above 75 m caught pollock larvae. Vertically, the maximum in larval density occurred between 10 m (RES 1) and 58 m (RES 3). The vertical distribution in early June showed no apparent pattern in relation to station location (Fig. 4). Using vertical distribution profiles, we estimated larval abundance at each station. In May, estimated abundances ranged from 60 larvaem-2 at RES 1 to 575 larvaem-2 at GAK 1 (Table 1). Abundances at the outer stations were much higher than in the in- ner basin owing to high larval densities below 50 m at RES 4 and GAK 1. In June abundances ranged from 0.5 larvae-m-2 at RES 1 to 10.3 larvaem-2 at RES 3. The estimated abundances were again higher at the outer stations. The highest abundance was found above the sill, as (V Q ■.ni.i ISO £ RES 1 RES 3 RES 2 RES 25 8 12 RES 4 GAK 1 Abundance (number/m3) Figure 3 Vertical distribution of walleye pollock larvae, Theragra chalcogramma, in Res- urrection Bay, Alaska, 1-4 May 1989. Muter and Norcross: Distribution, abundance, and growth of Theragra chalcogramma 583 Q- Q 100 150 200 150 RES 1 RES 2 RES 25 RES 3 RES 4 , GAK 1 8 12 Abundance (number/m3) Figure 4 Vertical distribution of walleye pollock larvae, Theragra chalcogramma, in Res- urrection Bay, Alaska, 7-9 June 1989. the largest number of larvae was captured at RES 3 at 58 m. Abundance averaged across all stations de- creased from 281 larvaem-2 in early May 1989 to 4.6 larvae-m-2 five weeks later. Larval size distribution Mean SL of larvae differed significantly with depth at all stations in early May (Table 2). Both a t-test and a Mann-Whitney test of differences between means showed highly significant differences between Table 1 Abundance of larva' walleye pollock, Theragra chal- cogramma. in Resurrection Bay in early May and early June 1989. Station Abundance (larvae-m-2) May June RES 1 60 0.5 RES 2 285 4.0 RES 2.5 137 1.8 RES 3 168 10.3 RES 4 461 5.2 GAK1 575 5.8 the shallow and deep samples at RES 1, 2, 2.5, 3, and RES 4 (P<0.01). An ANOVA for station GAK 1 suggested highly significant differences as well (F=42.33; P<0.001). Results from a Tukey HSD test showed significant differences in mean standard length between the samples from 22 m and 65 m at GAK 1 (P=0.01). Differences between any of the re- maining pairs at GAK 1 were not significant. While significant differences in size with depth existed at all stations, the sign of the differences varied between stations. At the two innermost sta- tions (RES 1 and RES 2) larval size decreased with depth, whereas at all other stations the opposite trend was found, i.e. larval size increased with depth, excluding the sample from 100 m at GAK 1. This sample showed a slight decrease in mean SL com- pared to shallower samples, but the difference was not significant. To compare larval size between stations, mean SL was corrected for sample date by using observed growth rates. Since we sampled over a four-day pe- riod, the measured lengths differed because of growth during this period. Thus, standard length was cor- rected for date of sampling by using a growth rate of 0.18 mm/day, the overall growth rate of pollock lar- vae in Resurrection Bay (this study). Table 2 shows 584 Fishery Bulletin 92(3). 1994 mean SL, variance, and corrected mean SL for all samples collected in the upper 100 m. The corrected mean SL will be hereafter referred to as mean SL. A nonparametric ANOVA by ranks showed that mean SL dif- fered significantly between the shallow samples from each sta- tion (Kruskal-Wallis test statis- tic=746.5, P<0.001 ). A Tukey type nonparametric multiple compari- son (Zar, 1984) indicated signifi- cant differences (P<0.05) between the innermost station pair (RES 1 and RES 2) and each of the sta- tions outside RES 2 (RES 2.5, 3, 4, GAK 1 ). Among the outside sta- tions, the only significant differ- ence was found between RES 3 andGAKl(P=0.003). When samples from all depths were pooled and mean SL com- pared between stations, results were very similar. An ANOVA showed a highly significant differ- ence in mean SL between the stations (F=80.00, P<0.001). A Tukey HSD multiple comparison again indicated that significant differences (P<0.05) exist between both of the two innermost stations and any one of the stations outside RES 2. Larvae at stations RES 1 and RES 2 were signifi- cantly larger and older than those at stations out- side RES 2. The observed size differences translate into an age difference of 8.5 days between the aver- age at the two inner stations (RES 1 and RES 2) and that at the outer stations (RES 2.5, 3, 4, and GAK 1 ). Age was calculated by using growth equations ob- tained in this study. The average age of larvae col- lected at stations RES 1 and RES 2 was estimated at 15.1 days. The average age of larvae at the other four stations was estimated to be 6.6 days, relative to 2 May. Thus, the results of size and age comparisons suggest that the stations can be divided into two dis- tinct groups on the basis of larval size. Growth rates Growth rates were determined for larvae collected 1-4 May 1989 at station RES 2 in the inner basin and at station RES 4 in the outer basin. At station RES 2, 62 larvae collected at 7 m on 4 May 1989 were measured and dissected to remove otoliths, of which 54 could be aged. The increment count ranged from 6 to 40 increments for larvae between 5.1 mm Table 2 Range, mean, standard len gth (SL), variance, and mean SL corrected for date of sampling fo r larval walleye pollock, Theragra chalcogr amma, col- lected in early May 1989 and preserved in isopropyl a cohol. Depth Number of Range Mean SL Corrected Station (ml larvae (mm) (mm) Variance mean SL RES 1 13 188 4.81-10.20 7.75 1.15 7.39 39 91 4.49-9.31 6.81 0.85 6.45 RES 2 7 481 4.07-15.03 7 61 2.01 7.25 60 24 4.59-8.40 6.67 1.02 6.31 RES 2.5 19 678 2.65-9.69 5.76 1 17 5.58 90-110 36 5.22-8.93 6.50 0.62 6.32 RES 3 18 919 3.37-10.68 6.27 1.18 6.09 58 95 4.15-9.58 6.91 0.88 6.73 RES 4 18 730 3.78-9.15 5.85 0.86 5.85 66 735 3.58-8.31 6.09 0.50 6.09 GAK1 22 983 2.87-8.81 5.37 0.78 5.55 32 299 3.62-7.88 5.45 0.71 5.63 65 2534 3.80-8.61 5.90 0.40 6.08 100 29 4.31-6.84 5.65 0.37 5.83 and 11.1 mm SL. A linear regression model relating mean SL and increment count yielded a growth rate of 0.18 ± 0.028 mm/day (95% CI)(r2=0.75, Fig. 5), assuming each increment represents growth of one day. From a sample collected at RES 4, at 18 m on 2 May 1989, 38 larvae ranging in length from 5.3 mm to 10.1 mm were aged. The growth rate at this sta- tion was estimated to be 0.19 ± 0.016 mm/day (r2=0.79, Fig. 5). We compared the regression lines of standard length on increment count from RES 2 and RES 4 (Fig. 5) using the ^-statistic according to Zar ( 1984) and found no significant difference between the slopes (r=1.048; 0.205,000 larvaem-2 at some stations in January to 200-400 larvaem-2 in early April (Nakatani, 1988). For the Bering Sea, typical abundance esti- mates range from 10 to 100 larvaem2 distributed over a very large area (Incze et al., 1984). In Auke Bay, Alaska, the observed abundances were much lower with maximum densities of 3-15 larvae-m-2 (Haldorson et al., 1989a). In ichthyoplankton samples taken in Resurrection Bay in the upper 30 m in 1988, maximum densities ranged from 0.8 larvaem-3 at RES 1 to 4.1 larvae-rrr3 at RES 4 (Smith et al., 1991), translating into abun- dances per unit area of 24 larvae-m-2 and 124 larvae-m-2 respectively. How- ever, these abundances may be underestimates, since only the upper 30 m were sampled by Smith et al. ( 1991 ), where- as our study found high abundances below 30 m, par- ticularly in the outer basin of the fjord (Fig. 3). Additional samples were collected in Resurrection Bay in late April and early May 1991. Abundances were similar to those estimated for 1989 (Muter, unpubl. data). The available data from 1988 to 1991 suggest that larval walleye pollock are consis- tently found in Resurrection Bay. The observed abun- dances are close to those re- sulting from the dense spawn- ing aggregations found in Shelikof Strait, Alaska, and Funka Bay, Japan (Kendall and Nakatani, 1992). Since the spatial extent of the spawning area in the vicinity of Resurrection Bay is unknown, total abun- dances cannot be compared at present. Larval size and age distribution Larvae from the shallowest samples at stations RES 1 and RES 2 were significantly larger and older than those at the other stations. Larvae from the shal- lowest tows may not be representative of the popu- lation as a whole because of changes in vertical dis- tribution with age. Thus we also pooled larvae from all tows at each station for between-station compari- sons. Some bias may remain because of inconsisten- cies in the depth sampling regime, but the results were almost identical to those obtained when only shallow samples are used. There is clearly a differ- ence in size and age between larvae at stations RES 1 and RES 2 and larvae at all other stations. This observation is consistent with the hypothesis that larvae are transported into the fjord and accumu- late inside the inner fjord basin. A length-frequency distribution for all larvae collected at each station (Fig. 8) shows a multimodal length distribution and 6 Bates, R. D., and J. Clark. 1983. Ichthyoplankton off Kodiak Island and the Alaskan Peninsula during spring 1981. NWAFC Proc. Rep. 83-89. Northwest and Alaska Fisheries Sci. Center, NMFS, NOAA, Seattle. WA, 105 p. LO- OS- 0.6- 4J 0.4- -H § 0.2- XJ u 5 1-0] c n) 0.8- 4J 0.6 -J 1 0.4- o 0.2- -H U U O LO- CI. 2 0.8- Cm 06- 0.4- 0.2- LO- RES 1 08. 0.6- fli Bfl LO- RES 2 0.8- 0.6- J . 0.4- LO- RES 2.5 08- 06- RES 3 RES 4 GAK 1 Length- capture 2 1 6 8 10 12 2 4 6 8 10 12 Length (cm) Length (cm) Figure 8 frequency distributions of all larval walleye pollock. Theragra chalcogramma. A at six stations in Resurrection Bay, Alaska, 1-4 May 1989. 588 Fishery Bulletin 92(3). 1994 a wide range of measured lengths at both RES 1 and RES 2, whereas at all other stations they show a more narrow, unimodal distribution. This distribution could be the result of several intrusions of surface water and larvae into the inner fjord. Growth The growth rates in Resurrection Bay were close to those reported for larvae from other geographic ar- eas (Table 3), including Shelikof Strait and Auke Bay, which are located in the Gulf of Alaska at latitudes similar to Resurrection Bay. Temperatures in the upper layer in Resurrection Bay were slightly lower in early May 1989 than those observed in Shelikof Strait and Auke Bay at the same time of year (Kendall et al., 1987; Pritchett and Haldorson, 1989; Fig. 2). The low temperatures in the inner basin in May reflect delayed warming of the upper water col- umn relative to the shelf outside the fjord. Thus, it may seem that the fjord in early spring provides less favorable conditions for growth than the shelf, con- sidering the lower temperatures inside the fjord. However, salinities also differ between the shelf and the fjord, resulting in a more pronounced stratifica- tion inside Resurrection Bay. Stratification of the water column will reduce vertical mixing and can result in an earlier onset of phytoplankton and zoop- lankton blooms. In spite of differences in tempera- ture, stratification, and vertical distribution (Kim, 1989; Pritchett and Haldorson, 1989), growth rates are very similar in Shelikof Strait, Auke Bay, and Resurrection Bay We detected no difference in growth rate between stations RES 2 and RES 4 in Resurrection Bay. This result is not surprising, given the proximity of the stations and the similarity in water properties. The growth rates, especially at the outer station, may be biased because only fish from the shallowest samples were aged. Larvae from the upper layer may not ad- equately represent the whole population. More samples would be needed to accurately test for dif- ferences in growth between stations. To test for interannual differences, data from additional years are needed. Differences in growth rates are most com- monly attributed to variations in water temperature and prey concentration. The primary prey of first feeding walleye pollock are copepod nauplii ranging in length from 100 to 300 pm (Kamba, 1977; Clarke, 1978). Smith etal. (1991) found over 20 copepod nau- plii (150-350 pm length) per liter throughout May 1988 in Resurrection Bay with numbers exceeding 100 per liter in mid-May. These prey concentrations are sufficient for successful feeding of larval walleye pollock (Paul, 1983; Haldorson et al., 1989b). Under these conditions growth of larvae in Resurrection Bay is not food limited. Growth rates in Resurrection Bay were also similar to those observed in the laboratory under optimal feeding conditions and at a higher tem- perature (Bailey and Stehr, 1988), further suggest- ing that growth was not food or temperature limited. Many studies have documented the effects of wa- ter temperature on growth of fish larvae (Houde, 1989). Laboratory studies have shown that first-feed- ing walleye pollock larvae reared at 5.5°C are more successful at capturing prey than larvae reared at 3 C (Paul, 1983). Brown and Bailey ( 1992) found geo- graphical differences in growth for juvenile walleye pollock that could be attributed to differences in tem- perature as well as nutrient levels. In our study, tem- peratures in the larval environment ranged from 3.5 to 6.3°C and growth rates fall well within the ob- served range of growth in other habitats. Hatching and spawning Hatch dates in Resurrection Bay fall well within the range of observed hatch dates in other parts of the Table 3 Laboratory and field-estimated growth rates of larval walleye pollock, Theragra chalcogramma. Size range Temperature Growth rate Year Location (mm) range I C) (mm/day) Reference 1981 Shelikof Strait 3-1.3 no data 0.17 Kim and Gunderson ( 1989) 1983 Shelikof Strait 6-15 5.5-7 0.21 Kendall etal. (1987) 1986 Auke Bay 4-13 6-7 0.23 Haldorson etal. (1989a) 1987 Auke Bay 5-11 5.5-7 0.16 Haldorson et al. (1989a) 1988 Auke Bay 5-11 6-7 (i 22 Haldorson et al. (1989a) 1989 Auke Bay 5-12 4-6.5 0.18 Haldorson et al. (1989a) Laboratory -1 — 1 1 9.310.5 0.20 Bailey and Stehr (1986) Laboratory 4-10 8-9 0.18 Bailev and Stehr (1988) Resurrection Bay 3-15 3.5-6.3 0.18 This study Muter and Norcross: Distribution, abundance, and growth of Theragra chalcogramma 589 Table 4 Median hatch dates of larval walleye pollock, Theragra chalco- gramma, in the Gulf of Alaska. Year Location Hatch date Reference 1983 Shelikof Strait 23 April Yoklavich and Bailey ( 19901 1985 Shelikof Strait 23 April Yoklavich and Bailev ( 1990 1 1986 Shelikof Strait 29 April Yoklavich and Bailey ( 1990) 1987 Shelikof Strait 2 May Yoklavich and Bailey ( 1990) 1987 Auke Bav 28 April Haldorson et al. (1989a) 1989 Resurrection Bay 22 April This study Acknowledgments We thank T. Weingartner and T. Royer for the use of unpublished data, B. Holladay and A. J. Paul for review of the manuscript, and A. L. Brown, K. M. Bailey, K. Besser, and L. Haldorson for help with larval otolith age- ing. Two anonymous reviewers provided help- ful suggestions. Funding for this study was provided by Alaska Sea Grant No. NA86AA- D-56041. Gulf of Alaska (Table 4). The median hatch date is remarkably consistent among different parts of the Gulf and among different years which would require a common, underlying mechanism to trigger spawn- ing over such a broad geographical range. The val- ues from Shelikof Strait suggest a trend towards later spawning dates between 1983 and 1987. More data are needed to determine if a similar trend exists in other areas of the Gulf and to identify parameters responsible for the timing of spawning. Conclusions The high abundances and growth rates of larvae in Resurrection Bay indicate that the fjord provides a suitable environment for the successful growth of larval walleye pollock. The hydrography of the re- gion and larval size distributions support the hypoth- esis that larvae recruit to the fjord from outside by advection into the outer basin of Resurrection Bay and across the sill. These observations and the high abundances of pollock larvae in nearby Prince Will- iam Sound during the same year (Norcross and Frandsen1 ) suggest that a large spawning popula- tion of walleye pollock exists in the region and that not all walleye pollock in the northern Gulf of Alaska spawn in Shelikof Strait. Larval walleye pollock are also abundant in the bays of Southeast Alaska (Haldorson et al., 1989, a and b). Thus, it is likely that many embayments along the Gulf of Alaska are utilized by this species. Future work is needed to determine the extent of spawning in the vicinity of Resurrection Bay and Prince William Sound and to test whether the area is consistently used by larval walleye pollock or whether abundances observed in 1989 were unusual. Also, the residence time of larvae in the area is not known. While larval pollock were found in Resurrec- tion Bay in all three years for which data are avail- able, there has been only one report of juvenile wall- eye pollock in the fjord (Feder et al., 1979). Literature cited Bailey, K. M., and C. L. Stehr. 1986. Laboratory studies on the early life history of the walleye pollock, Theragra chalcogramma (Pallas). J. exp. Mar. Biol. Ecol. 99:233-246. 1988. The effects of feeding periodicity and ration on the rate of increment formation in otolith of lar- val walleye pollock, Theragra chalcogramma, Pallas. J. exp. Mar. Biol. Ecol. 122:147-161. Brown, A.L., and K. M. Bailey. 1992. Otolith analysis of juvenile walleye pollock Theragra chalcogramma from the western Gulf of Alaska. Mar. Biol. 112:23-30. Carmo Lopes, P. do. 1979. Eggs and larvae of Maurolicus muelleri (Gono- stomatidae) and other fish eggs and larvae from two fjords in western Norway. Sarsia 64:199-210. Clarke, M. E. 1978. Some aspects of the feeding biology of larval pollock, Theragra chalcogramma (Pallas), in the southeastern Bering Sea. M.S. thesis, ITniv. of Alaska, Fairbanks, Alaska. 44 p. De Silva, S. S. 1973. Abundance, structure, growth and origin of inshore clupeid populations of the west coast of Scotland. J. exp. Mar. Biol. Ecol. 12:119-144. 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. Feder, H. M., A. J. Paul, and J. McDonald. 1979. A preliminary survey of the benthos of Res- urrection Bay and Aialik Bay, Alaska. Sea Grant Report 79-9. Alaska Sea Grant Program, Fair- banks, Alaska. 53 p. Haldorson, L., A. J. Paul, D. Sterritt, and J. Watts. 1989a. Annual and seasonal variation in growth of larval walleye pollock and flathead sole in a south- eastern Alaskan bay. Rapp. P.-v. Reun. Cons. int. Explor. Mer 191:220-225. Haldorson, L., J. Watts, D. Sterritt, and M. Pritehett. 1989b. Seasonal abundance of larval walleye pol- lock in Auke Bay, Alaska, relative to physical fac- tors, primary production, and production of zoo- 590 Fishery Bulletin 92(3), 1994 plankton prey. In Alaska Sea Grant Report 89-1, p. 159-172. Hinckley, S., K. M. Bailey, S. J. Picquelle, J. D. Schumacher, and P. J. Stabeno. 1991. Transport, distribution, and abundance of lar- val and juvenile walleye pollock (Theragra chalco- gramma) in the western Gulf of Alaska. Can. J. Fish. Aquat. Sci. 48:91-98. Houde, E. D. 1989. Comparative growth, mortality, and energet- ics of marine fish larvae: temperature and implied latitudinal effects. Fish. Bull. 87:471-495. Incze, L. S., M. E. Clarke, J. J. Goering, T. Nishiyama, and A. J. Paul. 1984. Eggs and larvae of walleye pollock and relation- ships to the planktonic environment. In D. H. Ito (ed.), Proceedings of the workshop on walleye pollock and its ecosystem in the eastern Bering Sea, p. 109- 159. NOAA Tech. Memo. NMFS F/NWC-62. Kamba, M. 1977. Feeding habits and vertical distribution of walleye pollock, Theragra chalcogramma (Pallas), in early life stage in Uchiura Bay. Res. Inst. N. Pac. Fish., Hokkaido Univ., Spec. Vol., p. 175-197. Kendall, A. W., Jr., and S. J. Picquelle. 1990. Egg and larval distributions of walleye pol- lock Theragra chalcogramma in Shelikof Strait, Gulf of Alaska. Fish. Bull. 88:133-154. Kendall, A. W., Jr., and T. Nakatani. 1992. Comparisons of early-life-history character- istics of walleye pollock Theragra chalcogramma in Shelikof Strait. Gulf of Alaska, and Funka Bay, Hokkaido, Japan. Fish. Bull. 90:129-138. Kendall, A. W., Jr., M. E. Clarke, M. M. Yoklavich, and G. W. Boehlert. 1987. Distribution, feeding and growth of larval wall- eye pollock, Theragra chalcogramma , from Shelikof Strait, Gulf of Alaska. Fish. Bull. 85:499-521. Kim, S. 1989. Early life history of walleye pollock, Theragra chalcogramma, in the Gulf of Alaska. In Alaska Sea Grant Report 89-1, p. 117-139. Kim, S., and D. R. Gunderson. 1989. Cohort dynamics of walleye pollock in Shelikof Strait, Gulf of Alaska, during egg and larval periods. Trans. Am Fish. Soc. 118:264-273. Lie, U. 1978. Eggs and larvae of fish from Lindaspollene. Sarsia63(3):163-167. Lindahl, O., and R. Perissinotto. 1987. Short-term variations in the zooplankton commu- nity related to water exchange processes in the Gull- mar Fjord, Sweden. J. Plankton Res. 9:1113-1132. Lindahl, O., and L. Hernroth. 1988. Large-scale and long-term variations in the zooplankton community of the Gullmar Fjord, Swe- den, in relation to advective processes. Mar Ecol. Prog. Ser. 43:161-171. Lloyd, D. S., and S. K. Davis. 1989. Biological information required for improved management of walleye pollock. In Alaska Sea Grant Report 89-1, p. 9-31. Matthews, J. B. L., and B. R. Heimdal. 1980. Pelagic productivity and food chains in fjord systems. In H. J. Freeland, D. M. Farmer, and C. D. Levings (eds.), Fjord oceanography. Plenum Press, New York, p. 377-397. Megrey, B. A. 1991. Population dynamics and management of walleye pollock Theragra chalcogramma in the Gulf of Alaska, 1976-1986. Fish. Res. 11:321-354. Muter, F. J. 1992. Distribution, abundance, and growth of lar- val walleye pollock (Theragra chalcogramma ) in an Alaskan fjord. M.S. thesis, Univ. Alaska Fair- banks, 91 p. Nakatani, T. 1988. Studies on the early life history of walleye pollock in Funka Bay and vicinity, Hokkaido. Mem. Fac. Fish., Hokkaido Univ. 35:1-46. Olla, B. L., and M. W. Davis. 1990. Effects of physical factors on the vertical dis- tribution of larval walleye pollock Theragra chalco- gramma under controlled laboratory conditions. Mar. Ecol. Prog. Ser. 63:105-112. Paul, A. J. 1983. Light, temperature, nauplii concentrations, and prey capture by first feeding pollock larvae Theragra chalcogramma. Mar. Ecol. Prog. Ser. 13:175-179. Pritchett, M., and L. Haldorson. 1989. Depth distribution and vertical migration of larval walleye pollock ( Theragra chalcogramma ). In Alaska Sea Grant Report 89-1, p. 173-183. Rogers, D. E., B. J. Rogers, and R. J. Rosenthal. 1987. The nearshore fishes. In D. W. Hood and S. T. Zimmerman (eds.), The Gulf of Alaska, physical en- vironment and biological resources. Mineral Manage- ment Service OCS study MMS 86-0095, p. 399-416. Royer, T. C. 1982. Coastal freshwater discharge in the northeast Pacific. J. Geophys. Res. 87:2017-2021. Schumacher, J. D., and A. W. Kendall Jr. 1991. Some interactions between young walleye pollock and their environment in the western Gulf of Alaska. CalCOFI Rep.. Vol. 32, p. 22-40. Smith, R. L., A. J. Paul, and J. M. Paul. 1991. Timing and abundance of herring and other fish larvae in an Alaskan glaciated fjord. In Proc. int. herring symposium Oct. 1990, Anchorage, Alaska. Alaska Sea Grant Report 91-01. Yoklavich, M. M., and K. M. Bailey. 1990. Hatching period, growth and survival of young walleye pollock Theragra chalcogramma as deter- mined from otolith analysis. Mar. Ecol. Prog Ser. 64:13-23. Zar, J. H. 1984. Biostatistical analysis. Prentice-Hall. New Jersey, 718 p. Abstract. — Capture of trans- forming larval and newly settled juvenile (age-0) summer flounder, Paralichthys dentatus, over four years (1986-1989) in the seaside salt marshes of Virginia's Eastern Shore and in the lower Chesapeake Bay verifies Virginia waters as a nursery area. Gear specific for ju- venile flatfish was used and sam- pling was conducted in a broad range of habitats in all months. This study demonstrates a fluctua- tion in the timing of the appearance and magnitude of abundance of age-0 summer flounder in Virginia waters over a four-year sampling period. Age-0 summer flounder ( 11-27 mm TL) began entering the area in October 1986 and were present throughout the winter of 1987. The 1988 and 1989 year classes did not appear until April at larger sizes (22-83 mm TL). Highest catch per unit of effort (CPUE) occurred between April and August and abundance de- clined in the fall. Data indicated that year-class strength declined from 1986 to 1988 and increased slightly in 1989. To monitor year- class strength of age-0 summer flounder, we recommend sampling Virginia estuaries in April, May, and June when both abundance of flounder is high and small-mesh- lined trawl gear is most efficient. Interannual variation in the recruitment pattern and abundance of age-0 summer flounder, Paralichthys dentatus, in Virginia estuaries* Brenda L. Norcross Virginia Institute of Marine Science, School of Marine Science College of William and Mary. Gloucester Point. Virginia 23062 Present address Institute of Marine Science School of Fisheries and Ocean Sciences. University of Alaska Fairbanks Fairbanks. Alaska 99775-1080 David M. Wyanski Virginia Institute of Marine Science, School of Marine Science College of William and Mary, Gloucester Point, Virginia 23062 Present address: South Carolina Marine Resources Research Institute South Carolina Wildlife and Marine Resources Department PO Box 12559. Charleston, South Carolina 29422 Summer flounder, Paralichthys dentatus (Pleuronectiformes: Both- idae), is an important commercial and recreational species along the eastern coast of the United States. It ranges from Nova Scotia (Scott and Scott, 1988 ) to Florida ( Gutherz, 1967) and its center of abundance occurs in the Middle Atlantic Bight (Scarlett, 1981). Though it is known that commercial landings of P. dentatus in the Middle Atlantic Bight fluctuate widely (Wilk et al., 1980), fluctuations in abundance of age-0 summer flounder have not been investigated. Because of the economic importance of summer flounder in Virginia, our first objec- tive was to design a sampling plan based on the early life history of summer flounder to assess the rela- tive yearly abundance of age-0 sum- mer flounder in Virginia waters. This index will provide the fishing industry and fishery managers with knowledge of fluctuations before those fluctuations affect the fishery. A part of designing an effective sam- pling plan was evaluation of appro- priate gear. Therefore, the second objective was to examine the effec- tiveness of sampling gear. Age-0 P. dentatus have been cap- tured in small numbers from Chesa- peake Bay (Orth and Heck, 1980; Weinstein and Brooks, 1983) and the Eastern Shore of Virginia (Richards and Castagna, 1970). Poole ( 1966) hypothesized that Vir- ginia waters and the sounds of North Carolina constitute primary nursery areas for summer flounder, but an insufficient number of speci- mens have been captured to sub- stantiate this hypothesis. Recruit- ment and distribution patterns of age-0 summer flounder have been investigated in estuaries in North Carolina (Powell and Schwartz, 1977; Miller etal., 1984; Burke etal., 1991 1 Manuscript accepted 16 December 1993. Fishery Bulletin 92:591-598 ( 1994). Contribution 1748 from the Virginia Institute of Marine Science. 591 592 Fishery Bulletin 92(3), 1994 and New Jersey (Able et al., 1990); however, the stud- ies in Virginia reporting the capture of age-0 sum- mer flounder were not directed specifically at this species (Richards and Castagna, 1970; Orth and Heck, 1980; Weinstein and Brooks, 1983). Thus the third objective of this study was to assess the region's importance as a nursery area. Methods Sampling sites were located on the seaside (eastern border) and bayside (western border) of Virginia's Eastern Shore and on the western shore of Chesa- peake Bay (Fig. 1 ) because the Chesapeake Bay and Eastern Shore were hypothesized to be prime nurs- 76°00' G. Guinea Marshes O. Occohannock Creek T. Tue Marshes s. Sand Shoal Channel W Wachapreague Channel 1 Supplemental stations sampled April -August 1986, July 1987 and June 1988 Figure 1 Locations of sampling sites for Paralichthys dentatus on Virginia's Eastern Shore and in Chesapeake Bay and its tributaries. Letters indicate sites with regu- lar, fixed station sampling; shading indicates areas in which rivers, (reeks, and nearshore locations were sampled during 1986 preliminary investigations and supplemental investigations in July 1987 and June 1988. ery grounds (Poole, 1966). The eastern border of the Eastern Shore peninsula is an extensive system of barrier islands enclosing salt marshes and shallow bays that are 1-2 m deep at mean high water (MHW). The bays and salt marshes are transected by main channels that are 3-20 m deep at MHW. On the west- ern border of the peninsula, there are shallow creeks, 1-6 m deep at MHW, which extend into upland ar- eas. Fringing and pocket marshes, much less exten- sive than the seaside salt marshes, occur along creek margins. Seagrass beds are present at the mouth of most creeks. The mouth of the York River (Fig. 1 ) is 3.7 km wide; it has extensive shoal areas along its margins and a main channel 18 m deep. Salt marshes, with channels 1-3 m deep, and seagrass beds are present in the shoal areas. Between 1986 and 1989, three different types of 4.9-m semi-balloon otter trawls with 19.1-mm bar mesh in the wings and upper body were used to sample areas 1-11 m in depth. Only bar mesh sizes are noted in this paper. The first unlined trawl, used in 1986 and July 1987, had 6.4-mm mesh in the lower body and codend. We added a 3.2-mm mesh liner to the codend in September to capture the newly settled juveniles. Because ctenophores and jellyfish could clog the mesh, mesh sizes of the unlined trawl were increased to 19.1 mm in the lower body and to 15.9 mm in the codend in August 1987. To compare the sampling efficiency of the lined and unlined trawls, both trawls were towed at each station from Sep- tember 1987 onward. All trawls were fished with a 4.8-mm link tickler chain to increase catches of flat- fish (see Creutzberg et al., 1987). Two 6.1-mm seines were used to sample shallow (<1 ml habitats. A beach seine (6.4-mm mesh) was used in April and May 1986 and a bag seine (3.2-mm mesh) was used from November 1986 until Decem- ber 1988. A 3.2-mm link chain was attached to the leadline of both seines to increase catches of flatfish. Trawling and seining were conducted from April 1986 to August 1989 during daylight hours (Norcross and Hata, 1990). While designing the study from April to August 1986, sampling was conducted at least once in most navigable waters of the Eastern Shore and at the mouth of the York River (Fig. 1). Over the next two years, September 1986-Septem- ber 1988, samples were collected at fixed stations at five sites (Fig. 1): Wachapreague and Sand Shoal Channels, Occohannock Creek, and Guinea and Tue Marshes (also see Wyanski, 1990). At each site, deep (5-11 m) water stations were located in the middle of channels, whereas shallow (<5 m) water stations were situated along channel margins. All stations had sand or fine-grained substrates. Samples were collected semi-monthly from September 1986 through Norcross and Wyanski: Recruitment pattern and abundance of Paralichthys dentatus 593 August 1987, and at monthly intervals thereafter. During expected periods of peak age-0 summer floun- der abundance in 1987 and 1988, additional samples were collected throughout the study area (Fig. 1). Sampling was reduced spatially and further re- duced temporally in 1989. Sampling was eliminated at Occohannock Creek, the site at which the fewest number of summer flounder were captured. Sampling was conducted April through August at the other four sites. Only trawling was continued. We measured the total length (TL) of each sum- mer flounder and used the length-frequency data to identify age-0 individuals. A birthdate of 1 January (Smith et al., 1981) was used when designating year class, although age-0 summer flounder may have been collected the preceding October through Decem- ber. For each gear, data from all sampling efforts were pooled by month and by year class and catch per unit of effort (CPUE) was calculated as the mean num- ber of age-0 summer flounder per 15-m seine haul or 5 minutes of trawl sample. To make sample sizes more similar among the treatment groups (year class) in statistical analyses, the 15-month time period over which a year class was sampled was separated into two time intervals: October-June and September-December. July and August data were not included in analyses because of bias pro- duced by the clogging of meshes. Some data were eliminated from statistical analyses owing to changes in the gear. Only seine data for the 1987 and 1988 year classes were com- pared because a different seine was used in 1986. Unlined trawl data for the 1986 year class were eliminated because the mesh size was smaller than in subsequent years. Because of nonrandom (fixed) station locations and nonindependent samples, nonparametric statistical tests were used to ana- lyze the CPUE data. For each gear, the Mann- Whitney test or Kruskal-Wallis test was used to compare monthly CPUE values among years (Zar, 1984). If the null hypothesis in the Kruskal-Wallis test was rejected, a multiple comparison test ( Dunn, 1964 ) was used to determine which means were significantly different. If P was <0.05, the results were considered significant. Results We were able to identify the age-0 year class for 15 months (October through December of the next year) using length frequencies from all four years of data combined (Fig. 2). Table 1 shows the appli- cation of these monthly size-at-age criteria to iden- tify the age-0 specimens in individual years. Sizes ranged from 11 mm to the largest age-0 specimen of 285 mm. Little to no change in mean size was ob- served from October to May, whereas rapid size changes were apparent from June to September. Though sampling effort and gear varied among years, age-0 summer flounder were caught within Virginia waters in each year of the study. Over the four years of sampling, age-0 P. dentatus were cap- tured each month but not during every month of ev- ery year ( Table 2 ). Summer flounder exhibited a pro- longed period of recruitment to inshore waters as age- 0 specimens were captured in Virginia estuaries from October to May (Table 2, Fig. 2). Newly settled speci- mens (<20 mm) were collected throughout the fall and winter of 1986-87; however, they were not col- lected in the fall and winter of 1987-88 and 1988- 89. When age-0 specimens first appeared in April of 1988 and 1989, they were already >20 mm. The highest CPUE values were reported for April through September (Table 2). Comparisons of CPUE 594 Fishery Bulletin 92(3). 1994 Table 1 Length ranges (mm) of age -0 summer flounder, Paralichthys dentatus, from all sites by year class for 15 months. 1986 1987 1988 1989 Year class Year class Year class Year class Oct 11-27 Nov — 13-19 — — Dee — 14-32 — — Jan — 17-34 — Feb — 17-38 — — Mar — 14-27 — — Apr 26-69 15-48 22-83 36-41 May 22-60 21-80 24-32 17-88 Jun 54 111) 27-160 35-160 35-144 Jul 96-190 68-180 86-180 57-160 Aug 30-220 93-240 115-210 90-210 Sep 96-265 147-275 176-222 — Oct 100-285 170-265 172-245 — Nov 119-218 — — — Dec 131-185 168-209 — — data pooled over the five sampling sites for each of the three gear types showed a general pattern of re- duced age-0 summer flounder abundance in Virginia estuaries between 1986 and 1988; there was a slight increase in 1989, based on trawl data (Table 3). The CPUE of the seine and the unlined trawl decreased an order of magnitude per year from 1986 to 1988. Twice as many summer flounder ( 101 vs. 54) were captured in seven seine hauls in April and May 1986 as in 527 seine hauls over the next two years (Table 3). For October-June data, CPUE in seine hauls was significantly greater in 1987 than in 1988 (Table 4) as no P. dentatus were captured in 1988. Seining, though successful in 1986, did not yield many age-0 flounder in 1987-1988 (Tables 2 and 3), and thus was discontinued. The unlined trawl data revealed no significant dif- ferences in CPUE between years (Table 4). We did not include 1986 unlined trawl data in analyses, but the high CPUE values for this gear type in May and June provided additional evidence that abundance was greater in 1986 compared with 1987-1989. The lined trawl data for October-June revealed significant differences in CPUE among years (Table Table 2 Catch per unit of effort (CPUE) of age -0 summer flounder. Para lichthys dentatus , by year class for 15 months from all sites. Seine CPUE = number of age-0 fl xinder/15 m hau ; trawl CPUE = number of age -0 floun ier/5 min; — = no of sample taken. 1986 vear c ass 1987 year class Hiss yearcl iss 1989 yea r class Monti (1985-86) (1986-87) LIST ss (1988- 89) Ti awl Seine1' Trawl Seine Trawl Seine 'IVawl Seine Lined' Unlined- Lined Unlined Lined Unlined' 1 Lined l'nlined'' Oct 0.03 0 n 0 n 0 0 Nov — — — 0.08 0.36 — 0 n ii ii ii 0 Dec — — — 0.24 0.16 — 0 II ii n 0 0 Jan 0.13 0.14 II n 0 Feb — — — 0.13 0.21 — 0 n 0 — — — Mar — — — 0.29 0.50 — 0 0 II — — — Apr 16.605 — — 0.79 0.75 — II 0.05 II — 0.06 0 Mav 9.00s 13.81 0.33 1.51 — 0 0.13 II — 1.14 0.29 Jun — — 4 08 0.04 3.98 — II 0.06 0.06 — 3.27 2.00 Jul — — 2.01 0.17 3 84 1 :;*-' 0 II 14 0.18 — 1.88 2.00 Aug — — 3.70 0.08 — 0.96' 1) 0.26 0.37 — 0.71 1.28 Sep 0.87 — 0 0.29 0.65' 1) ii 22 0.61 — — — Oct 0.69 — 0 n 12 1 .65' 0 0.35 0.43 — — — Nov 03 0.56 — 0 0 1.00' II 0. 1 3 0 — — — Dec '1 0.31 — 0 n 0.17' (1 0 ii — — — Semi balloon otter trawl (3.2-mm mesh linen Semi balloon otter trawl (6.4-mm meshi. ' Bag seine (3.2-mm mesh). ' Semi balloon otter trawl 15.9-mm mesh ) l-mm mesh!. Norcross and Wyanski: Recruitment pattern and abundance of Paralichthys dentatus 595 4); CPUE was higher in 1987 than in 1988. No other differences were detected. For September-December data, there were no significant differences in CPUE among the 1986, 1987, and 1988 year classes (Table 4). Gear efficiency changed as fish size increased. The unlined trawl with 15.9-mm mesh in the codend pro- duced generally lower CPUE values than the lined trawl during April through June (Table 2). As the age-0 specimens increased in size, the CPUE values for the unlined trawl became higher than those for the lined trawl. in timing of and size at first collection was reported in New Jersey, where age-0 flounder (<50 mm) were collected in the fall and during May but only occa- sionally during the winter months (Able et al., 1990). Thus, appearance of summer flounder in Virginia es- tuaries seems to be more similar to that of New Jer- sey (fall and late spring) rather than to that seen in North Carolina (winter and early spring). The time of first entrance in New Jersey, Virginia, and North Carolina estuaries corresponds with spawning peri- ods of September-December north, and November- Discussion The prolonged time of age-0 sum- mer flounder recruitment to the inshore waters of Virginia is more extended than entry times for North Carolina waters where age-0 P. dentatus enter estuaries from December through April (Deubler, 1958), January through April (Burke et al., 1991), or February through April ( Warlen and Burke, 1990). October through May re- cruitment to Virginia also agrees with reports of transforming lar- vae of P. dentatus (<20 mm TL) entering New Jersey inlets from October through May (Able et al., 1990). Age-0 summer flounder were not collected from October through May during all years of our study. They may appear in the fall or winter but often are not evi- dent until April. Similar variation Table 3 Summary of collection data from all sites by year class for age-0 Paralichthys dentatus: number of 15-m seine hauls, number of age-0 flounder captured, seine catch per unit of effort (CPUE) = number of age-0 flounder/haul, num- ber of trawl tows, total minutes tow time for lined and unlined trawls, num- ber age-0 flounder captured, trawl CPUE = number of flounder/5 min tow. Year class 1986 1987 1988 1989 Seine Number of hauls 46 295 232 Number of flounder 108 54 0 CPUE 1.20 0.18 0.00 Lined Trawl Number of tows 282 739 320 Number of minutes 1410.0 3664.5 1578.5 Number of flounder 192 670 30 CPUE 0.68 0.91 0.10 Unlined Trawl Number of tows 125 206 334 Number of minutes 613.8 1015.5 1657.8 Number of flounder 436 192 33 CIUK 3.55 0.94 0.10 93 426.0 96 1.13 94 467.0 97 1.04 - Table 4 Summary of statistical tests used to compare catch per unit of effort (CPUE) for Paralichthys dentatus between years for various gear and time intervals. H = the null hyp oth jsis; U = Mann -Whitney ( MW) statistic; H = Kruskal- Wallis (KW) statistic; Q = ran ltiple comparison statistic (Dunn, 1964); df = degrees of freedom; * = significant results at 0.05 level of significance. Gear Months tf„ Statistic P Test Seine Oct-Jun 1987=1988 £7=72* <0.001 MW Unlined trawl Oct-Jun 1988=1989 t/=34 >0.20 MW Unlined trawl Sep-Dec 1987=1988 (7=14 0.100.5 KW MC(Di MC(Dl MC(D) Lined trawl Si'p Die 1986=1987= 1988 H=5.734 df=2 0.05 '. O °A 6 UOA a A O O0 * A -0.2 □ -0 .6 -0I4 -0I2 6 o!2 Factor 1 B 0.4 D 0.3 0.2 Factor 2 o p G 0> • •• aD © ffl a 0 oO A »9 * • -0.1 A -0.2 D -0 6 -0^4 -0l2 0 0.2 Factor 1 Figure 3 Plots of scores on the first two correspondence axes (A) and (B) for the entire set of samples of rose shrimp, Aristeus antennatus collected on the upper slope (US) and middle slope (MS) only. F=females; M=males; J=juveniles; circle=spring; square=summer; triangle=autumn; diamond=winter. Black points ( • ) represent canyon samples on upper slope (US). The other population is commercially exploited and variable over the year at shallower depths between 400 and 1,000 m. It is characterized by high abun- dance and by seasonal variations in total number, sex ratio, and depth distribution. The A. antennatus stock, though subjected to con- siderable fishing pressure, has remained at near optimum equilibrium levels (Demestre and Lleonart, 1993). Sarda (1993) and Sarda and Cartes (1993a) attributed this equilibrium to the presence of unexploited biomass at depths below 1,000 m, which annually renews the exploited portion of the stock. Though sampling was localized and did not cover extensive areas of each habitat, the samples from the US and MS nonetheless reflected specific spatio-temporal patterns in population structure: females completely domi- nated the population all year, forming aggregations on the MS in spring and summer. Earlier studies on fishing patterns during the year (Figure 5A, revised after Tobar and Sarda, 1987) have referred to this population migra- tion pattern. Changes in the coefficient of variation of catch rate illustrated in Figure 5B reflect a scattering of the shoals in spring and summer (from April to August) and a more highly aggregated stock structure in autumn and winter. Figure 6 summarizes our conclusions regarding the distribution and movement of A. antennatus. The increase in number of females on the MS in spring and summer co- incides with the period of gonadal rip- ening and fertilization (Relini Orsi and Relini, 1979; Arrobas and Ribeiro- Cascalho, 1987; Sarda and Demestre, 1987; Demestre and Fortuno, 1992) and shortening of the male rostrum (Sarda and Demestre, 1989). These authors also reported that shoals dis- perse after spawning, which occurs mainly from June to September. This might suggest a specific mating area in the MS. After September, shrimps spread out over the slope and subma- rine canyons, leading to a decrease in density and increase in the proportion of males on the slope. Processes linked to the transfer of energy through the slope and submarine canyon systems (Reyss, 1971, 1973; or Rowe, 1971; 604 Fishery Bulletin 92(3). 1994 Koslow and Ota, 1981; Houston and Haedrich, 1984), and feeding habits (Cartes and Sarda, 1989; Cartes, 1991) may also be related to distribution of A. antennatus. Ghidalia and Bourgois (1961) and Bombace (1975) associated certain shrimp species with water masses that were of a characteristic septentrional type, with a low temperature of 12.8C and high salinity (38.1-38.8 ppt). However, it has not been possible to confirm these hydrographic hy- potheses for A. antennatus, because the study by Ghidalia and Bourgois ( 1961 ) reported few catch data of this species. No other studies on related species have established more specific hypotheses. The growth, abundance, parent stock biomass, and recruitment of commercial species are important when calculating parameters directly related to their population dynamics and exploitation (Caddy and number of individuals < » < I I n< n> Median Q3-Q1 MS US LS 3 5 10 9 7 2 163 119 40 321 222 36 Figure 4 Mean number of indi- viduals of rose shrimp, Aristeus antennatus, showing 95% confidence intervals (x2=8.67; P< 0.01): 7;<'=number of samples below median; 'n>'=number of samples above median; Q3-Q1= difference between maximum and mini- mum value. US=upper slope; MS=middle slope; LS=lower slope. Sharp, 1988; Sarda, 1993). Size-frequency data for A. antennatus caught by commercial fishing vessels (Demestre, 1990; Fig. 7) do not show progressions in monthly length-frequency modes. Apparent "nega- tive growth rates" between different months suggest that vessels were following a moving stock and that catches were taken at the most commercially profit- able locations. Procedures for analyzing this type of stock have been considered by Jones (1984), Caddy ( 1982, 1987), and Caddy and Garcia ( 1986) from fish- ery catches. Procedures for treating migratory stocks have been considered primarily by Sousa (1988) in the fish Decapterus russelli. Bias in size frequencies due to migratory effects in D. russelli is similar to £• 2° 1 15 g 10 H 5 8 > 6 Illl'i'i Figure 5 Monthly catch rates (A) and coefficient of variation (B) based on daily catches of rose shrimp, A. antennatus, taken by a vessel during 1984, 1985, and 1986 (redrawn after Tobar and Sarda, 1987). Num- bers are total number of samples by month for all three years combined. Table 3 Results of applying multifactorial nonparametric analysis of variance to the basic data matrix presented in Table 2 for collections of rose shrimp, Aristeus antennatus, with and without lower slope (LS) data. M = males; F = females; J=juveniles, rc=total number; ( + )=significant difference (P<0.05); ( — )=nonsignificant difference. With LS data Without LS data M M Seasonality Depth Interaction Sarda et al.: Spatio-temporal structure of a population of Aristeus antennatus 605 Figure 6 Conceptual model of the spatio-temporal dy- namics of Aristeus antennatus in the study area (see text for more details). MS=middle slope; US=upper slope, and LS=lower slope; circle= spring; square=summer; triangle=autumn; diamond=winter. Jan. Feb. Mar Apr May Jun Jul Aug. Sep. Oct. Nov. Dec J canyon middle slope canyon ) 20 30 40 50 60 CL, mm Figure 7 Monthly size frequencies for rose shrimp, Aristeus antennatus (carapace length) taken in commercial catches from the upper slope (US) and middle slope (MS) (data from Demestre, 1990). that observed for A. antennatus. Size-frequency modes were evident in spring and summer but did not progress in autumn and winter (Fig. 6). Sparre et al. (1989) advised using the annual-return matched sample method to estimate growth parameters in such cases. Failing to take these aspects into account when considering a population exploited by a given fleet in a given area may lead to errors in calculating bio- logical parameters and, consequently, in decision- making for fishery management. Acknowledgments We thank J. B. Company for his assistance in col- lecting the samples and G. Fuster and J. M. Anguita for our technical assistance. We also wish to express their appreciation to all the team members who par- ticipated in the MAR90/757 project, funded by the Spanish Ministry of Education and Science (CICYT). W. Norbis was supported by a doctoral grant from the Agencia Espanola de Cooperacion Internacional del Ministerio de Asuntos Exteriores. R. Sacks prepared the English translation. 606 Fishery Bulletin 92(3). 1994 Literature cited Anonymous. 1992. Working group on Nephrops and Pandalas stocks. ICES. CM. 1992/Assess. 8:1-275. Abello, P., and F. J. Valladares. 1988. Bathyal decapod crustaceans of the Catalan Sea (northwestern Mediterranean). Mesogee 48:47-102. Abello, P., F. J. Valladares, and A. Cast <1 Ion. 1988. Analysis of the structure of decapod crusta- cean assemblages off Catalan coast (North-West Mediterranean). Mar. Biol. 98:39-49. Arrobas, I., and A. Ribeiro-Cascalho. 1987. On the biology and fishery of Aristeus anten- natus (Risso, 1816) in the South Portuguese coast. Inv. Pesq. 51 suppl. 1:233-244. Bas, C. 1965. La gamba rosada (Aristeus antennatus). Publicaciones Tecnicas de la Junta de Estudios de Pesca 5:143-155. Benzecri, J. P. 1980. L' analyse des donnees. 2: L'analyse des correspondences. Ed. Dunod, Paris, 632 p. Bombace, G. 1975. Considerazione sulla distribuzione delle populazione di livello batiale con particulare referi- mento a quelle bentonectoniche. Pubbl. Zool. Napoli, 39, suppl. 1:7-21. Caddy, J. F. 1982. Some considerations relevant to the defini- tion of shared stocks and their allocation between adjacent economic zones. FAO Fish. Circ. 749:1^14. 1987. Size-frequency analysis for Crustacea: moult increment and frequency models for stock assess- ment. Kuwait Bull. Mar. Sci. 9:43-61. Caddy, J. F., and S. Garcia. 1986. Fisheries thematic mapping — a prerequisite for intelligent management and development of fisheries. Oceanogr. Trop. 21(l):31-52. Caddy, J. F., and G. D. Sharp. 1988. Un marco ecologico para la investigacion pesquera. FAO. Doc. Tec. Pesca 283:1-155. Campillo, A., P. Y. Dremiere, B. Liorzou, and J. L. Bigot. 1 990. Observations sur deux crustaces profonds du Golfe du Lyon: Aristeus antennatus (R.) et Nephrops norve- gicus (L.). FAO. Rap. sur les Peches 447:298-313. Cartes, J. E. 1991. Analisis de las comunicdades y estructura trofica de crustaceos decapodos. Ph.D. diss., Universidad de Barcelona, 627 p. In press. Influence of depth and seasonality on the diet of the deep-water aristeid Aristeus antennatus along the slope (between 400-2,300 m ) in the Catalan Sea (western Mediterranean). Mar. Biol. Cartes, J. E., and F. Sarda. 1989. Feeding ecology of the deep-water aristeid crustacean Aristeus antennatus. Mar. Ecol. Prog. Ser. 54:229-238. 1992. Abundance and diversity of decapod crusta- ceans in the deep Catalan Sea (western Mediter- ranean). J. Nat. Hist. U.K. 26:1305-1323. 1993. Zonation of the deep-sea decapod fauna in the Catalan Sea (western Mediterranean). Mar. Ecol. Prog. Ser. 94:27-34. Cartes, J. E., F. Sarda, J. B. Company, and J. Lleonart. 1993. Day-light migrations by deep-sea decapod crustaceans in experimental sampling in the west- ern Mediterranean. J. exp. Mar. Biol. Ecol. 171:63-73. Conover, W. J. 1980. Practical non-parametric statistics, 2nd ed. John Wiley & Sons, New York, 493 p. Demestre, M. 1986. Les diferents comunitats naturals de la Mediterranee. In J. Lleonart (ed.), L'Oceanogra- fia, recursos pesquers de la mar Catalana. Quad. Ecol. Aplicada. Diputacio de Barcelona (9):9-41. 1990. Biologia pesquera de la gamba Aristeus antennatus (Risso, 1816) en el mar Catalan. Ph.D. diss., Univ. Barcelona, 443 p. Demestre, M., and J. M. Fortune 1992. Reproduction of the deep-water shrimp Aris- teus antennatus (Decapoda: Dendrobranchiata). Mar. Ecol. Prog. Ser. 84:41-51. Demestre, M., and J. Lleonart. 1993. The population dynamics of Aristeus anten- natus (Decapoda: Dendrobranchiata) in the north- western Mediterranean. Sci. Mar. 57(2-3):183- 189. Gage, J. D., and P. A. Tyler. 1991. Deep-sea biology: a natural history of organ- isms at the deep-sea floor. Cambridge Univ. Press, London, 503 p. Garcia, S., and L. Le Reste. 1987. Ciclos vitales, dinamica, explotacion y orde- nacion de las poblaciones de camarones peneidos costeros. FAO Doc. Tec. Pesca 203:1-180. Ghidalia, W., and F. Bourgois. 1961. Influence de la temperature et de l'eclaire- ment sur la distribution des crevettes des moyennes et grandes profondeurs. Stud. Rev. Gen. Fish. Count. Medit., FAO 16:1-53. Gooding, R. M. 1984. Trapping surveys for the deep-water caridean shrimps Heterocarpus laevigatus and H. ensifer in the western Hawaiian Islands. Mar. Fish. Rev. 46(2): 18-26. Greenacre, M. J. 1984. Theory and applications of correspondence analysis. Academic Press, London, 364 p. Houston, K. A., and R. L. Haedrich. 1984. Abundance and biomass of macrobenthos in the vicinity of Carson Submarine Canyon, north- west Atlantic Ocean. Mar. Biol. 82:301-305. Hopkins, T. S. 1984. Physics of the sea. In R. Margalefled.), Western Mediterranean. Pergamon Press, Oxford, 363 p. Sarda et al.: Spatio-temporal structure of a population of Ahsteus antennatus 607 Jones, R. 1984. Assessing the effects of changes in exploita- tion using length composition data (with notes on VPA and cohort analysis). FAO Fish. Tech. Pap. 256:0-118. Jones, S. 1969. The prawns fishery resources of India. FAO Fish. Rep. 57(3):725-748. King, M. G. 1981. Increasing interest in the tropical Pacific's deep-water shrimps. Aust. Fish. 40(61:33-41. King, M. G., and A. J. Butler. 1985. Relationship of life-history patterns to depth in deep-water caridean shrimps (Crustacea: Natantia). Mar. Biol. 86:129-138. Koslow, J. A., and A. Ota. 1981. The ecology of vertical migrations in the three common zooplankters in the La Jolla Bight, April- August 1967. Biol. Oceaonogr. 1:107-134. Peres, J. M. 1985. History of the Mediterranean biota and the colonization of the depths. In R. Margalef (ed.). Western Mediterranean. Pergamon Press, New York, p. 198-232. Relini Orsi, L., and J. Relini. 1979. Pesca e riproduzione del gambero rosso Aristeus antennatus (Decapoda, Penaeoidae) nel Mar Ligure. Quad. Civica Staz. Idrobiol. Milano 7:1-39. Relini, G., and L. Orsi Relini. 1987. The decline of red shrimp stocks in the Gulf of Genoa. Inv. Pesq. 51 (supl. 1): 245-260. Reyss, D. 1971. Les canyons sous-marins de la mer Catalane: le rech du Cap et le rech Lacaze-Duthiers. Ill: Les peuplements de macrofaune benthique. Vie et Milieu 22:529-613. 1973. Les canyons sous-marins de la mer Catalane: le rech du Cap et le rech Lacaze-Duthiers. rV: Etude synecologyque des peuplements de macrofaune benthique. Vie et Milieu 23:101-142. Risso, A. 1816. Historie naturelle des Crustaces des environs de Nice, 175 p. Rowe, G. T. 1971. Observations on bottom currents and epiben- thic populations in Hatteras Submarine Canyon. Deep-Sea Res. 18:569-581. Sarda, F. 1993. Bio-ecological aspects of the fisheries of Crus- tacea Decapoda in the western Mediterranean. Aquat. Living Resour. 6:299-305. Sarda, F., and J. E. Cartes. 1993a. Distribution, abundance and selected bio- logical aspects of Aristeus antennatus in deep-wa- ter habitats in NW Mediterranean. B.I.O.S. (Thessaloniky, Greece) 1(11:59-73. 1993b. Relationship between size and depth in de- capod populations on the slope between 900 and 2,200 m in the western Mediterranean. Deep-Sea Res. 4(11-12)2389-2400. Sarda, F., and M. Demestre. 1987. Estudio biologico de la gamba Aristeus anten- natus (Risso, 1816) en el Mar Catalan (NE de Espana). Inv. Pesq. 51 (l):213-232. 1989. Shortening of the rostrum and rostral variabil- ity in Aristeus antennatus (Risso, 1816) (Decapoda: Aristeidae). J. Crust. Biol. 9 (41:570-577. Sarda, F., and P. Martin. 1986. Les pesqueries a Catalunya: evolucio en els liltims decennis. In J. Lleonart (ed.), L'Ocean- ografia, recursos pesquers de la mar Catalana. d'Ecol. Aplicada. Diputacio de Barcelona 91-112. Sarda, R, G. Y. Conan, and X. Fuste. 1993. Selectivity of Norway lobster (Nephrops nor- vegicus) in western Mediterranean. Sci. Mar. 57(2-3):167-174. Sarda, F., C. Bas, and J. Lleonart. In press. A functional morphometry of rose shrimp Aristeus antennatus (Risso, 1816) in western Medi- terranean. Crustaceana. Sousa, M. I. 1988. Sources of bias in growth and mortality esti- mation of migratory pelagic fish stocks, with em- phasis on Decapterus russelli (Carangidae) in Mozambique. FAO Fish. Rep. 389:288-307. Sparre, P., E. Ursin, and S. C. Venema. 1989. Introduction to tropical fish stock assess- ment. FAO Fish. Tech. Paper 306/1, 337 p. Tobar, R., and F. Sarda. 1987. Analisis de la evolution de las capturas de gamba rosada, Aristeus antennatus (Risso, 1916), en los liltimos decenios en Cataluha. Inf. Tec. del Inst. Ciencias del Mar de Barcelona (CSIC) 142:1-20. 1992. Annual and diel light cycle as a predictive factor in deep-water fisheries for the prawn Aristeus antennatus (Risso, 1816). Fish. Res. 15:169-179. Tyler P. A. 1988. Seasonality in the deep sea. Oceanogr. Mar. Biol. Ann. Rev. 26:227-258. Zar, J. H. 1984. Biostatistical Analysis. Prentice-Hall Inter., 718 p. Abstract. — a 39-month study of the effects of cessation of sew- age sludge disposal in the New York Bight apex on the diets of certain fishes and on the benthic macro- faunal community provided an op- portunity to examine predator-prey relationships of winter flounder, Pleuronectes americanus, one of the common predators in the area. Benthic macrofauna and winter flounder were collected monthly and bimonthly, respectively, from July 1986 through September 1989 at three sites in the Bight apex that are variably influenced by sewage sludge. There were limited changes in winter flounder diets and abun- dance of dominant benthic macro- faunal species following cessation of sewage sludge disposal. The com- parison of volumetric contribution of common prey in flounder stom- achs to potential-prey abundance in benthic samples suggested sev- eral relationships. These included evidence of preferential predation on the polychaete Pherusa affinis; this selective preference may be associated with its high caloric con- tent as well as with its average high biomass density. Other com- mon prey, primarily polychaetes but including an anthozoan, were also preyed upon in proportions greater than their abundance in the environment. Some moderately abundant potential prey, such as the small near-surface-dwelling mollusc Nucula proximo and the ribbon worm Cerebratulus lacteus were not commonly preyed upon suggesting they were unavailable as prey or were avoided by winter flounder. Corresponding fluctua- tions in abundances and predation of the pollution-tolerant polychaete Capitella sp. and the pollution-sen- sitive amphipod Unciola irrorata suggested a proportional consump- tion relationship in association with sludge disposal and its cessa- tion. Predator-prey relationships of winter flounder, Pleuronectes americanus, in the New York Bight apex Frank W. Steimle Dorothy Jeffress Stephen A. Fromm Robert N. Reid Joseph J. Vitaliano Ann Frame Sandy Hook Laboratory, Northeast Fisheries Science Center National Marine Fisheries Service. NOAA Highlands. New Jersey 07732 Manuscript accepted 30 December 1993. Fishery Bulletin 92:608-619 (1994). 608 Predator diets provide information on sources of prey, predominant prey types, levels of particular prey use and availability, and prey pref- erence or avoidance, when com- pared with the availability of poten- tial prey in the environment (Lev- ings, 1974; Diehl, 1992). For preda- tory fish such as winter flounder, Pleuronectes americanus, this infor- mation increases our understand- ing of prey selection, based on evi- dence of prey preference or avoid- ance, and how selective predation can affect or be affected by prey pop- ulation dynamics. In studies of aquatic environmen- tal health, benthic macrofaunal di- versity and certain indicator species are often used as response vari- ables. Monitoring predator diets during such studies can aid in de- termining how predation can func- tion as a confounding factor in in- terpreting macrofaunal change as solely the product of altered abiotic factors. Monitoring diets can also indicate how benthic species abun- dance and the overall community structure can be affected by preda- tion. Predation studies can also aid in estimating the effect of benthic macrofaunal changes (natural or anthropogenic) on predator-prey relationships (i.e. loss of a season- ally or energetically important prey) or be used to define potential con- taminant uptake pathways (Cle- ments and Livingston, 1982; Gen- dron, 1987; Schindler, 1987). Studies of the linkage between prey abundance and predation by marine fish are scarce, especially studies based on samples taken over an extended period of time or dur- ing an environmental change. An opportunity to examine predator- prey relationships was provided by the availability of the results of a comprehensive 39-month study of the effects of sewage sludge disposal abatement in the New York Bight apex (the coastal area at the mouth of New York Harbor). This study included monitoring the diets of sev- eral common fishes and large deca- pod crustaceans, the abundance of benthic macrofauna, as well as other biological and environmental variables (Environmental Processes Div., 1988). The winter flounder, Pleuronectes americanus Walbaum (Robins et al., 1991 ) is an abundant demersal fish in the New York Bight apex. It was the third most important contribu- tor (KM) to the total fish biomass collected during the study (Wilk et Steimle et al.: Predator-prey relationships of Pleuronectes amencanus 609 al., 1992). Winter flounder also rely al- most entirely on small benthic macrofauna for food; thus, variability in predation and macrofaunal community structure can have important conse- quences for both the predator and its benthic prey (Kurtz, 1975; Clements and Livingston, 1982; Pihl et al., 1992). This paper compares the diets of win- ter flounder with the abundance of benthic macrofauna at three stations in the New York Bight apex variably af- fected by sewage sludge, for evidence of 1) selective predation (preference or avoidance), 2) variation in the propor- tional consumption of benthic prey over time, and 3) the manner in which preda- tor-prey relationships influenced the macrofauna with the cessation of sew- age sludge disposal. Materials and methods 7 4°00 73°40' Cholera Bank 40°20-- D.plh In Ualot I Winter flounder and benthic macro- fauna collection methods, sample pro- cessing, and primary data analyses for this study have been described in de- tail elsewhere (Reid et al. in press; Steimle, in press). In brief, three stations, R2, NY6, and NYU (Fig. 1), were sampled systematically from July 1986 to September 1989 (Environmental Pro- cesses Div. 1988). These stations represent a gradi- ent of conditions related to sewage sludge disposal, e.g. variable levels of total organic carbon (TOO and chemical sediment contamination, such as chromium (Cr), that changed to some degree after cessation of disposal in December 1987 (Table 1). Station NY6 was the most sludge-affected area, with a markedly altered benthic community characterized by a rela- tively low species richness, low biomass, and high levels of TOC sediment contaminant such as Cr (Table 1). Station R2 was moderately affected by sludge and was biologically enhanced; it showed rela- tively high species richness and macrofaunal biom- ass and moderate levels of sediment contamination compared with NY6 (Table 1). Station NYU was the least affected by sludge disposal, having relatively high species richness but low macrofaunal biomass and low levels of sediment contamination (Table 1). All stations were about 30 m in depth and had simi- lar sediment types (silty-fine sands) and hydro- graphic characteristics (Table 1). Triplicate benthic macrofaunal samples were col- lected monthly with a 0.1-m2 Smith-Mclntyre grab Figure 1 Location of stations and waste disposal sites in the New York Bight apex where winter flounder, Pleuronectes americanus, and benthic macrofauna were collected. sampler and sieved through a 0.5-mm mesh screen. Materials retained were preserved in buffered for- malin and later transferred to 70% ethanol. The samples were sorted and organisms were identified to the lowest possible taxon (usually species), counted, and moist-weighed (Reid et al., in press). Data from a total of 350 benthic macrofaunal samples were available for analysis. Adult winter flounder were collected bimonthly (additional collections were made in August) with a small otter trawl at the same stations and during the same week as the benthic grab sampling. Each periodic trawl collection consisted of six daytime tows of 0.5 km, deployed in an array across the center of the station. Generally, at least 30 fish were collected from each station per collection period. Stomach con- tents were analyzed in the field or laboratory by us- ing semi-quantitative, visual estimates of stomach volume from comparison of stomach boluses with vari- able-diameter, volume-calibrated cylinders. All identi- fiable items in the stomachs were identified to the low- est possible taxon and separated to visually estimate their individual percent contribution to total stomach volume (Langton et al., 1980; Steimle, in press). This method provides reasonable results compared with more labor intensive methods (Hyslop, 1980). 610 Fishery Bulletin 92(3). 1994 Table 1 Habitat characteristics (means) of three sampling sites in the New York Bight apex for the 18 months before (A) and 21 months after (B) sewage sludge disposal cessation; benthic biomass (wet wt) does not include the biomass of two large bivalves, Pitar morrhuanus and Arctica islandica. Stations R2 NYli NY11 Depth (m) Sediment grain size (phi)' TOC (%dry wt)' Cr (ppm dry wt)2 Bottom water' min. dissolved oxygen (mg/L) temp, range CO Benthos biomass (g/m-) species (n )/grab 29 :il 29 3.1 3.2 3.6 3.5 3.1 3.1 0.9 0.9 4.6 2.3 0.3 0.3 37.2 36.7 163.0 96.5 15.8 12.9 4.2 5.7J 4.3 5.8-* -3-17 for all sites5 5.3 <5.54 218.2 30 202.8 33 63.4 19 28.2 33 66.6 41 73.2 45 ' Packer et at, in press. 2 Zdanowicz et al. in press. 3 Arlen. L, A. Draxler. and R. Bruno. Hydrographic observations in the bottom water of the New York Bight at the "12 mile" dumpsite: 1983-1990. Unpubl. manuscr. 4 Dissolved oxygen content of <3 mg/L was recorded in September, 1989. 5 Summer levels in 1988 were 2-3°C lower than the long-term mean. The mean proportional volume (percent of the to- tal volume) of a given prey item in the stomach was compared with the mean proportional biomass of that prey in the benthic macrofauna, and the relation- ship between these proportions was used to deter- mine whether selective predation was evident. Al- though this approach to examining predator-prey as- sociations differs from the traditional use of prey nu- merical abundance in stomachs and the environment, it is nonetheless realistic and useful because 1) prey volume can be a more precise dietary variable com- pared with uncertain enumeration of prey that are easily fragmented and for which only parts are present and 2) volume and biomass are approxi- mately equivalent for most common prey taxa (see below). However, amphipods were also considered numerically because they are usually eaten whole and their exoskeletons are resistant to digestion and thus allow a reasonably accurate assessment of the number of individuals eaten. The stomachs of 3,556 adult winter flounder, 18- 30 cm total length, examined from the three study sites had identifiable food in them. To examine over- all predator-prey relationships, stomach content and benthic data from the entire study period were pooled for each station because there were only minor changes in dominant prey and benthic species asso- ciated with sludge disposal cessation (Reid et al., in press; Steimle, in press). The use of pooled stomach content data to estimate prey preference by a preda- tor population is recommended by some authors, e.g. Rachlin et al. ( 1987). Changes in predator-prey rela- tionships related to cessation of sludge disposal are considered separately. Any seasonal or annual vari- ability was assumed to be distributed equally within the pooled data as there were no gaps in collections. Log transformation of Shorigin's forage ratio index (K=rtlpt), adapted from a numerical approach for prey volume and biomass, was used to estimate prey se- lectivity (Berg, 1979): K = LogI0 irt/pt), where rt = pro- portion of prey in the diet estimated by contribution to stomach volume and pi = proportion of prey in the benthic biomass. Positive K values suggest a degree of selective predation. Near-zero K values suggested that predation is directly proportional to abundance. Nega- tive K values suggest underutilization or avoidance of a potential prey relative to its availability. The use of volume and biomass to calculate K is reasonable because 1 niL or cm:i of prey volume is considered approximately equivalent to 1 g of macro- fauna wet weight (Bowman, 1986). We partially veri- fied this assumption by determining the mean vol- ume to wet weight ratios for a number of individuals for a range of common prey taxa. The ratios for an Steimle et al.: Predator-prey relationships of Pleuronectes americanus 61 1 anthozoan, Ceriantheopsis americanus, a polychaete, Pherusa affinis, and the sand shrimp Crangon septemspinosa varied between 0.96 and 1.00 (n=>26). However, the ratios for calcareous-shelled prey were lower, 0.70 for the bivalve mollusc Nucula proximo (n-50), and 0.77 for the sand dollar Echinarachnius parma (n=30). All fish were collected near mid-day; therefore, the effects of digestion on stomach volume estimates were not considered a major factor. Winter flounder are primarily daytime feeders and all but the most soft- bodied prey should remain identifiable in their stom- achs for several hours (MacDonald et al., 1982). Results Diet spectrum and dominance Forty-nine prey taxa were identified in winter floun- der stomachs, although only about 30 taxa were iden- tified in the stomach contents at any individual sta- tion (Table 2). These are conservative estimates be- cause of some uncertainties in identification caused by digestion. This prey spectrum represents about a quarter of the total available benthic macrofaunal taxa identified at station NY6 ( 119 species) and R2 (133 species), and about a fifth of the 154 species identified at station NYU. Dominant prey, defined as species that composed at least 2% of the total stomach volume of fish from any station, were the polychaetes Pherusa affinis, Asabellides oculata, andNephtys incisa and the tube- dwelling anthozoan Ceriantheopsis americanus. The rhynchocoel Cerebratulus lacteus, juvenile rock crabs Cancer irroratus, and other polychaetes, including Capitel/a sp. and Scoletoma (Lumbrineris) spp., were dominant in the diet at one or two stations (Table 3). Collectively these eight taxa constituted between 76% and 96% of the winter flounder diet by volume at the three stations (Table 3). Predation patterns The log forage ratio (K) indices were positive for sev- eral dominant species. The K index was consistently high (>+0.25) for the polychaetes, P. affinis and A. oculata. High positive K values were also calculated for other prey, but at only one or two stations (Table 3). The K indices were near zero (±0.20) for some dominant species at some stations, such as C. americanus. For other prey or at other stations the K indices were low (<-0.25). This was especially evi- dent for C. americanus, Spio setosa, Glycera sp., and the molluscs (Table 3). Comparison of the contributions of these dominant prey species to flounder diets and to macrofaunal Table 2 List of winter flounder, Pleuronectes americanus. prey identified in stomachs collected from three sta- tions in the New York Bight apex. Prey Station R2 NY6 NYU Algae X X Coelenterates Ceriantheopsis americanus X X X Hydrozoans X X Rhynchocoels X X X Nematodes X Chaetognaths X Bryozoans X Molluscs Nucula proximo X X X Yoldia sp. X Nudibranchia X Ilyanassa trivittata X Polychaetes Nephtys sp. X X X Nephtys incisa X X X Ninoe nigripes X Scoletoma (Lumbrineris) sp. X X X Scoletoma acicularum X Pherusa affinis X X X Asabellides oculata X X X Ampharete acutifrons X Spionidae X X Spio setosa X X Spwphanes bombyx X X X Cirratulus cirratus X Tharyx sp. X Tharyx acutus X X X Capitella sp. X X X Mediomastus ambiseta X Phyllodoce arenae X X Drilonereis sp. X X X Glycera sp. X X X Ophioglycera gigantea X Chone infundibuliformis X Aglaophamus circinata X Nereis succinea X Isopods Edotea triloba X X X Cirolana sp. X Cumaceans Diastylis sp. X X X Amphipods Leptocheirus pinguis X X X Unciola irrorata X X X Monoculodes edwardsi X X Photis pollex X Dyopedos sp. X Mysids X Decapods Cancer sp. X X X Cancer irroratus X X X Crangon septemspinosus X X X Echinoderms Echinarachnius parma X Tunicates Molgula sp. X Salpidae X X 612 Fishery Bulletin 92(3). 1994 Table 3 Comparison of the average contribution of dominant prey in the diet of winter flounder, Pleuronectes americanus (mean % total stomach volume), to that of prey and dom nant benthic infauna (infn) species (mean % tota wet weight biomass) at stations in the New York Bight apex, 1986-1989 Seventy two benthic grab samples (tripl icate samples at 24 collection periods) are i ncluded. K' is the log forage ratio. Species Stations NY6 R2 NYU prey infn (AT') prey infn (A") prey infn I/O Ceriantheopsis amencanus 28.0 18.5 0.18 7.0 6.6 0.03 27.3 6.4 0.63 Pherusa affirm; 25.1 2.7 0.97 76.3 42.8 0.25 48.3 9.7 0.70 Asabellides oculata 9.2 3.8 0.38 3.4 <0.1 >1.83 2.0 <0.1 >1.60 Nephtys incisa 3.5 3.5 0 4.9 4.9 0 2.0 0.7 0.46 Capitella sp. no 11.7 -0.29 0.3 <0.1 >0.78 <0.1 <0.1 0 Cerebratulus acteus 3.0 45.3 -1.15 2.4 2.0 0.08 0.8 2.0 -0.40 Scoletoma sp. 1.5 <0.1 >1.48 2.1 <0.1 >1.62 4.1 3.3 0.09 Cancer irroratus 11.3 2.4 0.67 1.5 4.4 -0.47 1.2 0.3 0.60 Spio setosa <0.1 0.6 <-1.10 <0.1 0.1 <-0.30 0.8 7.5 -0.96 Nucula proximo ii 1 1.9 -1.30 <0.1 0.5 <-1.00 0.1 6.6 -1.70 Glvcera sp. 0.6 0.8 -0.12 <0.1 0.2 <-0.60 0.1 6.5 -1.70 Pitar morrhuanus 0.0 ii 1 0.0 30.7 -oo 0.0 14.7 -oo Arctica islandica 0.0 0.1 -oo 0.0 <0.1 _,-.-, 0.0 28.1 — oo Ensis direclus (III <0.1 -oo 0.0 3.1 _oo 0.0 <0.1 Totals 78.2' 91.8 97.9 95.3 86.6 85.8 Nonempty stomachs (n ) 1405 1628 523 ' Diet residuals at this station were primarily 1 -11'? l unidentifiable nrga lie and inorganic maten al biomass over time illustrates trends in the predator- prey association not evident in the pooled data (Table 3 ). For example, there was an apparent proportional consumption association ( level of prey consumption was closely associated with level of abundance) for the anemone C. americanus at all stations, although the strength of the association varied at times (Fig. 2). However, the proportional contribution of the poly- chaete P. affinis to winter flounder diets was greater than the prey's proportional contribution to total benthic macrofaunal biomass (Table 3, Fig. 3). The contribution of this prey to the winter flounder diet generally paralleled its contribution to macrofaunal biomass at all stations during the study period. Peaks in consumption are often consistent with peaks in proportional contribution to total community bio- mass, especially at stations NY6 and NYU (Fig. 3). For the entire study period, the difference in propor- tional consumption and biomass for P. affinis ranged from 9.3 fold at station NY6 (25.1% stomach volume vs. 2.7% benthic biomass; K=0.Q1) to 1.8 fold at sta- tion R2 (76.3% volume vs. 42.8% biomass; if =0.25); station NY11 had an intermediate difference of 3.7 fold and if =0.70 (Table 3). Some benthic species that were dominant in the overall benthic biomass were seldom identified in the flounder stomachs. For example, the rhynchocoel C. lacteus, which was particularly abundant at station NY6, was not often found in flounder stomachs, de- spite its relative importance in the benthic biomass (Table 3). This underutilization of potential prey was also evident at station NYU, where the polychaetes, Glycera spp. (mostly G. dibranchiata) and Spio setosa, as well as several mollusc species, composed >5% of the mean macrofauna biomass but were never important items in the flounder diet (Table 3). How- ever, a substantial portion of the small, unidentified polychaete fragments found in some winter flounder stomachs may have been S. setosa. The predation- abundance trends over the study period for these and other less common prey are not presented but were similar to the trends presented for C. americanus (Fig 2). In general, there was little evidence of predation on molluscs by winter flounder, despite their some- times high contribution to overall macrofaunal bio- mass. For example, the minute (<5 mm shell width ) Atlantic nut clam, Nucula proxima, a consistent, al- though only moderate component of the infaunal bio- mass, was not commonly found in winter flounder stomachs (Table 3). Steimle et al.: Predator-prey relationships of Pleuronectes americanus 613 Cessation of sewage sludge disposal The change in abundance of two benthic species, the polychaete Capitella sp. and the amphipod U. Station R2 100 90 80 70 60 50 40 30 20 10 0 f^S 7 8 9101112 12 3 4 5 6 7 8 9101112 12 3 4 5 6 7 8 9101112 12 3 4 5 6 ! 86 I 87 I 88 I 89 collection period irrorata, was associated with the cessation of sludge disposal (Reid et al., in press). These changes were also reflected in the occurrence of these two prey in winter flounder stomachs. When sludge disposal ended in 1987, the abundance of the stress- tolerant Capitella sp. declined and so did its occurrence in stomachs at station NY6 (Fig 4). For the more stress-sensitive U. irrorata, predation was also generally associated with abundance at two stations (Fig. 5). At sta- tion NY6, where the effect of sludge disposal was greatest, the abatement of sludge dis- posal was accompanied by a seasonally vary- ing increase in the abundance of this amphi- pod and a corresponding increase in the fre- quency of occurrence in winter flounder stom- achs (Fig. 5). 89 ; stomach volume benthic biomass Station NY6 7 8 9101112 12 3 4 5 6 7 8 9101112 12 3 4 5 6 7 8 9101112 123456789 86 I 87 I 88 I 89 collection period stomach volume benthic biomass Station NY11 7 8 9101112 12 3 4 5 6 7 8 9101112 12 3 4 5 6 7 8 9101112 123456789 86 I 87 I 88 I 89 I collection period stomach volume benthic biomass Discussion Diet spectrum and dominance The spectrum of prey and dominant prey taxa consumed by winter flounder in the New York Bight apex (Table 2) is similar to that re- ported in other winter flounder diet studies (Wells et al., 1973; Hacunda, 1981; Langton and Bowman, 1981; Bharadwaj, 1988; Steimle and Terranova, 1991). Winter flounder diet studies in estuaries and to the north also re- port the dominance of similar prey taxa (Tyler, 1972; Wells et al., 1973; Worobec, 1982). With the possible exception of Capi- tella sp., dominant prey are usually macro- scopic, supporting Keats' (1990) hypothe- sis that winter flounder show a preference for the largest available prey that can be consumed. Predation patterns The results of this study provide evidence that some prey species are consumed prefer- entially, but this was not temporally or spa- Figure 2 Trends in the percent contribution of the burrow- ing anemone Ceriantheopsis americanus to total stomach volume of winter flounder, Pleuronectes americanus, compared to the anemone's percent- age of the total benthic macrofaunal biomass at three sites in the New York Bight apex variably affected by sewage sludge disposal. Disposal gradually abated during 1987 and ceased by De- cember 1987. 614 Fishery Bulletin 92(3). 1994 tially consistent. The parallel temporal relationship of the contribution of the anemone C. americanus to overall benthic biomass and to winter flounder diets Station R2 7 8 9101112 12 3 4 5 6 7 8 9101112 123456789101112 123456789 [ 86 I 87 88 I 89 I collection periods stomach volume benthic biomass Station NY6 100 7 8 9101112 12 3 4 5 6 7 8 9101112 12 3 4 5 6 7 8 9101112 123456789 86 I 87 88 89 collection periods stomach volume benthic biomass Station NY11 7 8 9101112 12 3 4 5 6 7 8 9101112 12 3 4 5 6 7 8 9101112 1 23456789 86 87 88 89 collection periods stomach volume benthic biomass at stations R2 and NY6 (Fig. 2) suggests a propor- tional consumption association. The forage ratio in- dices iK) for this prey are near zero (0.03 to 0.18, Table 3); these suggest that winter flounder neither preferred nor avoided this prey, con- suming it at a level closely related to its avail- ability at these two stations. Although, this relation was more variable at station NY11 (Fig. 2), a moderately high, positive K value (0.63 ) suggests some preference for this prey (Table 3). The proportional consumption differences (Fig. 3) and strong positive K values at each station for P. affinis (Table 3) strongly sug- gest that winter flounder have a preference for this prey. Heavy predation on P. affinis may be energetically advantageous to win- ter flounder because this species has a high caloric equivalence ( x =1.9 Kcal/g wet weight; Steimle and Terranova, 1985). This is about double that of other major prey, which range from 0.9 to 1.1 Kcal/g wet weight, although some amphipod species, such as U. irrorata (consumed to a minor extent), are equally energy rich (Steimle and Terranova, 1985). The energy content of prey has been dis- cussed as an important consideration in op- timum foraging theories ( Mangel and Clark, 1986) and the winter flounder preference for this high energy prey provides support for these theories. It is also possible that prefer- ence for this prey is related to the long-term dominance of P. affinis in the macrofauna of siltier areas of the New York Bight apex, as this relatively productive species has been common in the area and has been preyed upon since at least the mid-1960s (Steimle and Stone, 1973; Steimle, 1985, 1990; Steimle et al., 1990). This extended period of domi- nance may have contributed to winter floun- der becoming experienced predators on this species (and other abundant prey species) and thus maintaining the preference (Gendron, 1987). Figure 3 Trends in the percent contribution of the poly- chaete Pherusa affinis to total stomach volume of winter flounder, Pleuronectes americanus, com- pared to the polychaete's percentage contribution of the total benthic macrofaunal biomass at three sites in the New York Bight apex variably affected by sewage sludge disposal. Steimle et al.; Predator-prey relationships of Pleuronectes amencanus 615 70 60 50 40 30 20 10 0 9 10 11 12 1 86 I 23456789 10 11 12 1234567 87 I 88 I collection periods benthic biomass stomach volume Figure 4 Trends in the percent contribution of the pollu- tion-tolerant polychaete Capitella sp. to total stomach volume of winter flounder, Pleuronectes americanus, compared to the polychaete's percent- age contribution to total benthic macrofaunal bio- mass at NY6, the station most affected by sewage sludge disposal and its cessation, in the New York Bight apex. Figure 5 Trends in the frequency of occurrence of the am- phipod Unciola irrorata in the diet of winter floun- der, Pleuronectes americanus, compared to its nu- merical abundance in the benthic macrofaunal community at two stations in the New York Bight apex variably affected by sewage sludge disposal and its cessation in December 1987. With the exception of P. affinis and per- haps C. irroratus, there is limited evidence of prey preference by winter flounder. This is consistent with the results of many quali- tative winter flounder feeding studies that report a diverse diet, but also with limited evidence of a prey species' preferences (Tyler, 1972; Klein-MacPhee, 1978; Keats, 1990). However, several dominant members of the benthic macrofaunal community at the three study stations were not commonly consumed by winter flounder. This underutilization may be related to prey size, burrowing depth, defense or escape mechanisms, or a variety of other factors (Main, 1985). For example, there was limited predation on C. lacteus, a major contributor to benthic biomass at sta- tion NY6. This nonproportional consumption or possible "avoidance" of this species is evi- dent in the difference between the propor- tional consumption and abundance levels at this station (3.0% stomach volume vs. 45.3% benthic biomass) and the strong negative K ( Table 3); this difference at NY6 was fairly consistent over the study period (Fig. 6). Proportional preda- Station NY6 50 40 30 20 10 #/0.1 sq m 04-' ' I >■ 7 8 91011121 234567 8 91011121 2 3 4 5 6 7 8 91011121 23456 789 I 86 I 87 I 88 I 89 I collection periods % Ireq occurrence # per square meter Station NY11 #/0.1 sq m q i i r ix^ I I I I I I I I I I I I I i I i I I I I I I i xi ^ — i i i i i ; l^ I Q 7 8 9101112 12 3 4 5 6 7 8 9101112 12 3 4 5 6 7 8 9101112 123456789 I 86 I 87 I 88 I 89 collection periods % freq occurrence # per square meter -1.15; tion or slight underutilization of this prey was evi- dent at the other stations where it was less abun- dant and K ranged from +0.08 to -0.40 (Table 3). 616 Fishery Bulletin 92(3), 1994 7 8 9101112 12 3 4 5 6 7 8 9101112 12 3 4 5 6 7 8 9101112 123456789 I 86 I 87 88 I 89 I collection period stomach volume benthic biomass Figure 6 Trends in the percent contribution of the rhynchocoel Cerebratulus lacteus to the total stomach volume of winter floun- der, Pleuronectes americanus, compared to the rhynchocoel's percentage contribution to total benthic macrofaunal biomass at NY6, the station most affected by sewage sludge disposal and its cessation, in the New York Bight apex. The limited predation on C. lacteus could be re- lated to its possession of defensive toxins in its tis- sues or to secretion of mucus that is strongly acidic, both of which can be offensive to potential predators (Kern, 1985; McDermott and Roe, 1985). Paradoxi- cally, rhynchocoels are collected for fish bait in some areas and eaten by other fishes. It is possible, how- ever, that this rhynchocoel's large size (>100 cm in length and >2 cm in width; Gosner, 1971), not its reported toxicity, is responsible for its limited use as prey for the small-mouthed winter flounder. The underutilization of the small bivalve mollusc N. proximo (Table 3) is probably the result of its lack of availability as this species burrows into the sedi- ment. The other molluscs common in the area, such as P. morrhuanus,A. islandica, Ilyanassa trivittata, and Ensis directus, are probably either too large or deeply buried to be suitable prey for winter floun- der. However, N. proximo, as well as these other molluscs, except P. morrhuanus, were noted as prey in the diets of winter flounder elsewhere (Kurtz, 1975; Gilbert and Suchow, 1977; Klein-MacPhee, 1978; MacDonald et al., 1982; Worobec, 1982). The molluscan contribution to winter flounder diets in Narragansett Bay, Rhode Island, was also considered negligible (Bharadwaj, 1988). Cessation of sewage sludge disposal The cessation of sewage sludge disposal was expected to result in a substantial change in abundance of dis- posal-sensitive species in the benthic macrofaunal community. Any macrofaunal changes because of cessation were expected to be reflected in some predation variables (Spies, 1984; Cross et al., 1985; Environmen- tal Processes Division, 1988). However, the only significant changes in the benthic com- munity detected after cessation from 1987 to 1990 were in the overall abundance of a few pollution-tolerant or pollution-sensitive taxa, such as the polychaete Capitella sp. or am- phipods at station NY6, nearest to the former disposal area. The abundance of other domi- nant benthic species at NY6, such as P. affinis, N. incisa, and C. americanus, did not change to any significant degree (Reid et al., in press). The pollution-tolerant Capitella sp. has been a consistent, but variably abundant (hundreds to tens of thousands of individu- als/m2), member of the degraded benthic com- munity at station NY6 since at least the early 1970s (Caracciola and Steimle, 1983). Its density in the macrofauna declined drasti- cally (<100 individuals/m2) after the cessa- tion of sewage sludge disposal (Reid et al., 1991). Predation by winter flounder on this prey is prob- ably related to the response of Capitella sp. to cessa- tion of disposal. There were increases in abundance of contami- nant-sensitive amphipods, especially Photis pollex and U. irrorata, at station NY6 after disposal abate- ment began in 1987. Although contributions of am- phipods to winter flounder diets were generally less than V'/r of total stomach volumes (and thus not in- cluded in Table 3), numerical increases in amphipod abundance in the macrofaunal community coincided with increases in their frequency of occurrence in winter flounder stomachs. The general increase in the numerical abundance of one amphipod, U. irrorata, in benthic samples at NY6, especially after cessation, was accompanied by a corresponding in- crease in their occurrence in winter flounder stom- achs (Fig. 5). However, a similar predation relation- ship for this prey was somewhat evident at station NYU, which was minimally affected by sludge dis- posal and cessation (Fig. 5). At both stations, the fre- quency of occurrence off/, irrorata in the diets closely paralleled seasonal (winter-spring) peaks in abun- dance, including reduced predation during an appar- ently poor recruitment year at NY11 in 1988 (Fig. 5). If the effects of the sludge were more general and included some effect at NYU, the pattern of preda- tion on this species, which increased in abundance after cessation, could be associated with this cessation. Steimle et al.: Predator-prey relationships of Pleuronectes americanus 617 Predation influence on benthic populations and energetics Strong selective predation by winter flounder on cer- tain benthic macrofaunal taxa can influence the population dynamics of these prey and be a factor in interpreting benthic community change relative to disposal abatement. This affect may be evident in the predation-abundance patterns for Capitella sp., although the decline in the abundance of this spe- cies could be a result of its short life span as well as predation. A short life span is not a factor for P. affinis as it lives for up to two years (Steimle et al. 1990). How- ever, since winter flounder are visual predators (Klein-MacPhee, 1978; Bharadwaj, 1988), P. affinis may be at greater risk to predation because it is rela- tively large (up to 7.5 cm in length and 0.5 cm in width) and lives in vertical burrows with its fanned, setaceous cephalic cage (head and bristles) usually exposed. It also actively probes the sediment surface with feeding palps that may attract predator atten- tion (J. Vitaliano, personal observ.; P. Ferri1). The persistent abundance and strong predation on P. affinis by winter flounder (Fig. 3) and other preda- tors at station R2 (Steimle and Terranova, 1991; Steimle, in press) is interesting in that it suggests that this prey must be very productive in this area to sustain heavy predation pressure. This sugges- tion is supported by results of a previous study of the secondary production of this species that found it to be almost twice as productive near R2 as at NY6 and NYU (Steimle et al., 1990). The results of this study are subject to potential biases inherent in most stomach content analyses, such as differential digestion of different prey types (MacDonald et al., 1982). Partial digestion of prey can underestimate a prey's contribution to diets based on stomach volume. Thus, the estimated pro- portional contribution to diets of some soft-bodied prey will be conservative. This study also assumes that the stomach contents of fish represented feed- ing at or very near where the benthic infauna samples were collected (Steimle, in press) and that if a large proportion of the prey was consumed in nearby ar- eas, the benthic community structure did not differ substantially from that at the actual benthic collec- tion site. Some preliminary data from peripheral benthic stations sampled during the study suggest that this assumption was valid, although there were substantial changes in community structure in some parts of the trawling zones (Fromm, personal observ.). Ferri. P., U.S. Environmental Protection Agency. Narragansett, RI 02882. Personal Commun.. 1992. In summary, diets of winter flounder in the New York Bight apex 1) were dominated by a few prey species, typically polychaetes, an anthozoan, and small crustaceans; 2) showed a preference for the energy-rich polychaete Pherusa affinis; 3) suggested there was underutilization or "avoidance" of molluscs and rhynchocoels; and 4) showed that apparent re- sponses of some benthic macrofaunal species (Capitella sp. and U. irrorata) to cessation of sludge disposal and natural fluctuations in abundance were reflected in corresponding changes in their use as prey. Acknowledgments We thank S. Fromm, C. Zetlin, and R. Koch for their contribution to data processing and graphics, and D. Packer, S. Chang, D. Mountain, and others for help- ful comments and suggestions. Literature cited Berg, J. 1979. Discussion of methods of investigating the food of fishes, with reference to a preliminary study of the prey of Gobiusculus flavescens (Gobiidae). Mar. Biol. 50:263-273. Bharadwaj, A. 1988. The feeding ecology of the winter flounder, Pseudopleuroneetes americanus (Walbaum), in Narragansett Bay, Rhode Island. M.S. thesis, Univ. Rhode Island, Kingston, RI, 129 p. Bowman, R. E. 1986. Effects of regurgitation on stomach content data of marine fishes. Environ. Biol. Fish. 16:171- 181. Caracciola, J. V., and F. W. Steimle. 1983. An atlas of the distribution and abundance of dominant benthic invertebrates in the New York Bight apex with reviews of their life histories. NOAATech. Rep. NMFS SSRF-766, 58 p. Clements, W. H., and R. J. Livingston. 1982. Overlap and pollution-induced variability in the feeding habits of filefish (Pisces: Mona- canthidae) from Apalachee Bay, Florida. Copeia 2:331-338. Cross, J. N., J. Roney, and G. S. Kleppel. 1985. Fish food habits along a pollution gradient. Calif. Fish Game 71:28-39. Diehl, S. 1992. Fish predation and benthic community struc- ture: the role of omnivory and habitat complexity. Ecology 73:1646-1661. Environmental Processes Division. 1988. Plan for study-response of habitat and biota of the inner New York Bight to abatement of sew- age sludge dumping. NOAA Tech. Memo. NMFS F/NEC-55, 34 p. 618 Fishery Bulletin 92(3). 1994 Gendron, R. P. 1987. Models and mechanisms of frequency-depen- dent predation. Am. Nat. 130:603-623. Gilbert, W. H., and E. F. Suchow. 1977. Predation by winter flounder {Pseudopleur- onectes americanus) on the siphons of the clam, Tellina agilis. Nautilus 91:16-17. Gosner, K. L. 1971. Guide to identification of marine and estua- rine invertebrates. Cape Hatteras to the Bay of Fundy. Wiley-Interscience, NY, 693 p. Hacunda, J. S. 1981. Trophic relationships among demersal fishes in a coastal area of the Gulf of Maine. Fish. Bull. 79:775-788. Hyslop, E. J. 1980. Stomach contents analysis — a review of meth- ods and their application. J. Fish. Biol. 17:411^423. Keats, D. W. 1990. Food of winter flounder Pseudopleuronectes americanus in a sea urchin dominated community in eastern Newfoundland. Mar. Ecol. Prog. Ser. 60:13-22. Kem, W. R. 1985. Structure and action of nemertean toxins. Am. Zool. 25:99-111. Klein-MacPhee, G. 1978. Synopsis of biological data for the winter flounder, Pseudopleuronectes americanus (Wal- baum). NOAATech. Rep. NMFS Circ. 414, 43 p. Kurtz, R. J. 1975. Stomach content analysis in relation to dif- ferences in growth rate of winter flounder {Pseudo- pleuronectes americanus) from two Long Island Bays. M.S. thesis, Long Island Univ., Greenvale, NY, 60 p. Langton, R. W., and R. E. Bowman. 1981. Food of eight northwest Atlantic pleuronecti- form fishes. NOAA Tech. Rep. SSRF-749, 16 p. Langton, R. W., B. M. North, B. P. Hayden, and R. E. Bowman. 1980. Fish food habits studies — sampling proce- dures and data processing methods utilized by the Northeast Fisheries Center, Woods Hole Labora- tory, USA. Int. Counc. Explor. Sea, C. M. 1980/ L:61, 16 p. Levings, C. D. 1974. Seasonal changes in feeding and particle se- lection by winter flounder {Pseudopleuronectes americanus). Trans. Am. Fish. Soc. 103:828-832. MacDonald, J. S., K. G. Waiwood, and R. H. Green. 1982. Rates of digestion of different prey in Atlan- tic cod (Gadus morhua), ocean pout {Macrozoarces americanus), winter flounder {Pseudopleuronectes americanus) and American plaice {Hippoglossoides platessoides). Can. J. Fish. Aquat. Sci. 39:651- 659. Main, K. L. 1985. The influence of prey identity and size on se- lection of prey by two marine fishes. J. Exp. Mar. Biol. Ecol. 88:145-152. Mangel, M., and C. W. Clark. 1986. Towards a unified foraging theory. Ecology 67:1127-1138. McDermott, J. J., and P. Roe. 1985. Food, feeding behavior and feeding ecology of nemerteans. Am. Zool. 25:113-125. Packer, D., T. Finneran, L. Arlen, R. Koch, S. Fromm, J. Finn, S. A. Fromm, and A. Draxler. In. Press. Fundamental and mass properties of surficial sediments in the inner New York Bight and responses to the abatement of sewage sludge dumping. In A. Studholme, J. E. O'Reilly, and M. C. Ingham (eds.). Effects of the cessation of sew- age sludge dumping at the 12-mile site. NOAA Technical Rep. NMFS. Pihl, L., S. P. Baden, R. J. Diaz, and L. C. Schaffner. 1992. Hypoxia-induced structural changes in the diets of bottom-feeding fish and Crustacea. Mar. Biol. 112:349-361. Rachlin, J. W., A. Pappantoniou, and B. E. Warkentine. 1987. A bias estimator of the environmental re- source base in diet preference studies with fish. J. Freshwater Ecol. 4:23-31. Reid, R., D. Radosh, A. Frame, and S. Fromm. 1991. Benthic macrofauna of the New York Bight, 1979-89. NOAA Tech. Rep. NMFS 103, 50 p. Reid, R., S. Fromm, A. Frame, D. Jeffress, J. Vitaliano, D. Radosh, and J. Finn. In press. Limited responses of benthic macrofauna and selected sewage sludge components to phase- out of sludge dumping in the inner New York Bight. In A. Studholme, J.E. O'Reilly, and M.C. Ingham (eds.), Effects of the cessation of sewage sludge dumping at the 12-mile site. NOAA Tech- nical Rep. NMFS. Robins, C. R., R. M. Bailey, C. E. Bond, J. R. Brooker, E. A. Lachner, R. N. Lea, and W. B. Scott. 1991. Common and scientific names of fishes from the United States and Canada (5th ed. ). Am. Fish. Soc. Spec. Publ. 20. Schindler, D. W. 1987. Detecting ecosystem responses to anthropo- genic stress. Can. J. Fish. Aquat. Sci. 44 (Suppl.l ): 6-25. Spies, R. 1984. Benthic-pelagic coupling in sewage-affected marine ecosystems. Mar. Environ. Res. 13:195- 230. Steimle, F. 1985. Biomass and estimated productivity of the benthic macrofauna in the New York Bight: a stressed coastal area. Estuarine Coastal Shelf Sci. 21:539-554. 1990. Benthic macrofauna and habitat monitoring on the continental shelf of the northeastern United Steimle et al .: Predator-prey relationships of Pleuronectes americanus 619 States. I: Biomass. NOAA Tech. Rep. NMFS 86, 28 p. Steimle, F. W. In press. Effects of sewage sludge disposal abate- ment on winter flounder, red hake, and lobster feed- ing and diets in the New York Bight Apex. In A. Studholme, J. E. O'Reilly, and M. C. Ingham (eds. ), Effects of the cessation of sewage sludge dumping at the 12-mile site. NOAA Technical Rep. NMFS. Steimle, F. W., and R. B. Stone. 1973. Abundance and distribution of inshore benthic fauna off southwestern Long Island, N.Y. NOAA Tech. Rep. NMFS SSRF-673, 50 p. Steimle, F. W., and R. Terranova. 1985. Energy equivalents of marine organisms from the continental shelf of the temperate northwest Atlantic. J. Northwest Atl. Fish. Sci. 6:117-124. 1991. Trophodynamics of select demersal fishes in the New York Bight. NOAA Tech. Memo. F/NEC- 84, 11 p. Steimle, F. W., P. Kinner, S. Howe, and W. Leathern. 1990. Polychaete population dynamics and produc- tion in the New York Bight associated with vari- able levels of sediment contamination. Ophelia 31:105-123. Tyler, A. V. 1972. Food resource division among northern, ma- rine, demersal fishes. J. Fish. Res. Board Can. 29:997-1003. Wells, B. D., D. H. Steele, and A. V. Tyler. 1973. Intertidal feeding of winter flounder {Pseudo- pleuronectes americanus) in the Bay of Fundy. J. Fish. Res. Board Can. 30:1374-1378. Wilk, S., R. Pikanowski, A. Pacheco, D. MacMillan, and L. Stehlik. 1992. Response offish and megainvertebrates of the New York Bight apex to the abatement of sewage sludge dumping. NOAA Tech. Memo. NMFS F/ NEC-90, 78 p. Worobec, M. N. 1982. Field analysis of winter flounder Pseudopleu- ronectes americanus in a coastal pond: abundance, daily ration, and annual consumption. Ph.D. the- sis, Univ. Rhode Island, Kingston, RI, 115 p. Zdanowicz, V. S., S. L. Cunneff, and T. W. Finneran. In press. Reductions in sediment metal contami- nation in the New York Bight apex with the cessa- tion of sewage sludge dumping. In A. Studholme, J. E. O'Reilly, and M. C. Ingham (eds.), Effects of the cessation of sewage sludge dumping at the 12- mile site. NOAA Technical Rep. NMFS. Abstract. — Numerical classifi- cation techniques, recurrent group analysis, and a clustering analysis that uses the Bray-Curtis resem- blance measure were used to iden- tify rockfish (family Scorpaenidae) assemblages in the offshore waters of Oregon and Washington. Catch data from six multispecies ground- fish assessment surveys conducted at three-year intervals (1977-92) by the National Marine Fisheries Service's Alaska Fisheries Science Center revealed three assem- blages. The first, a deep-water as- semblage, consisted of shortspine thornyhead, Sebastolobus alas- canus, Pacific ocean perch, Seba- stes alutus, darkblotched rockfish, S. crameri, and splitnose rockfish, S. diploproa. Redbanded rockfish, S. babcocki, and rougheye rockfish, S. aleutianus, were closely associ- ated with this group. The second assemblage consisted of canary rockfish, S. pinniger, yellowtail rockfish, S. flavidus, and green- striped rockfish, S. elongatus. This group was most abundant in areas over the middle shelf. The third assemblage, closely associated with the second, consisted of sharpchin, S. zacentrus, rosethorn, S. helvo- maculatus, and redstripe, S. proriger, rockfish. While the three assemblages may be of particular interest to ecologists, managers faced with the division of the Sebastes complex management unit into groups that better reflect rockfish cooccurrence may only be able to manage the latter two as- semblages as one shelf-rockfish unit. Rockfish assemblages of the middle shelf and upper slope off Oregon and Washington Kenneth L. Weinberg Alaska Fisheries Science Center National Marine Fisheries Service, NOAA 7600 Sand Point Way N.E., Seattle, WA981 15-0070 Manuscript accepted 4 January 1994. Fishery Bulletin 92:620-632 ( 199 I 620 Over the last three decades, some stocks of the more than 30 species of rockfish (family Scorpaenidae) known to inhabit the offshore wa- ters of Oregon and Washington (Eschmeyer and Herald, 1983) have been the target of intense foreign and domestic fishing pressure in a largely multispecies trawl fishery. In 1982 a groundfish fishery man- agement plan was implemented by the Pacific Fisheries Management Council to address the reductions in groundfish populations, including serious declines of several rockfish species in some areas. This plan was based on an underlying single-species management philosophy. In both the Columbia and the U.S. portion of the Vancouver (US- Vancouver) management areas in- stituted by the International North Pacific Fisheries Commission ( INPFC ), annual harvesting restric- tions have been fashioned for Pacific ocean perch, Sebastes alutus, widow rockfish, S. entomelas, shortbelly rockfish, S. jordani, and thorny- heads, Sebastolobus spp. The re- maining rockfish species have been lumped into a single management unit, the Sebastes complex. In ad- dition to individual trip limits, cur- rent restrictions for this large group involve an overall annual harvest guideline and harvest guidelines for two of its already stressed compo- nents, yellowtail, Sebastes flavidus, and canary, S. pinniger, rockfish (PFMC, 19921). While efforts to pre- vent over-harvesting of yellowtail and canary rockfish continue, the fishing pressure on the minor rock- fish in the management unit has escalated. Species, such as dark- blotched rockfish, S. crameri, are becoming increasingly important to fishermen. This research stems from a con- cern over the long-term effects of the trawl fishery on the condition of the rockfish community as a whole. Current reductions in some stocks along with present bycatch prac- tices may precipitate the need for changes in management policies to conserve these stocks. One possible course of action involves multi- species management, whereby cooc- curring species are managed as a species complex or assemblage. If needed, restrictions could be placed on the fishery for the assemblage when specific components become stressed. An effective assemblage management program requires knowledge of interspecific associa- tions in addition to individual life histories, distribution, and abun- dance patterns. Prior knowledge of offshore rock- fish associations has largely been inferred from shoreside sampling of commercial catches. In recent years several investigations have been 1 Pacific Fishery Management Council l PFMC I. 1992. Status of the Pacific coast groundfish fishery through 1992 and rec- ommended acceptahle biological catches for 1993; stock assessment and fishery evaluation. Pacific Fishery Management Council, Metro Center, Suite 420, 2000 SW First Ave., Portland, OR 97201, 80 p. Weinberg: Rockfish assemblages off Oregon and Washington 621 conducted to identify assemblages using trawl data covering extensive geographic areas. Nagtegaal (1983) studied both annual and seasonal interrela- tionships of some rockfish species off British Colum- bia based on commercial catch data. However, these catch statistics do not allow a full description of the effects of selective harvest on the entire rockfish com- munity, as the landed species are those of highest economic value allowed for harvest at the time and reflect only a portion of the overall rockfish commu- nity exposed to trawling. Rogers and Pikitch (1992) defined several groundfish assemblages based on prediscard data from the commercial trawl fishery off Oregon and Washington. That study included a variety of groundfish families, but only the most abundant rockfish species were considered. Similarly, Gabriel (1982) included a wide variety of ground- fishes in a 1-year assemblage study that utilized 1977 survey data from California to Washington. All of these researchers recognized the value offish assem- blage identification as a tool for fisheries manage- ment and emphasized the need to verify assemblage persistence. Since 1977, the Alaska Fisheries Science Center (AFSC) of the National Marine Fisheries Service (NMFS) has been conducting controlled bottom trawl surveys aimed at assessing and monitoring ground- fish resources off the west coast of the United States. Now that several of these surveys have been per- formed, a unique opportunity exists to monitor per- sistence in fish assemblages. In this paper I describe and summarize these surveys and their rockfish samples and use standard numerical classification tech- niques to identify the major rockfish assemblages. Methods cific hake, Merluccius productus. The survey design was patterned after rockfish distributions, deter- mined by fisheries catch data and the results of a pilot rockfish survey (Gunderson and Nelson2). The 1980 survey was specifically redesigned to better assess the canary and yellowtail rockfish populations, in addition to Pacific hake (Coleman, 1986). Thus, the sampling effort was divided among three depth strata: 55-183 m, 184-220 m, and 221-366 m. The 1983 survey repeated the work conducted in 1980 with the addition of some stations in the northern U.S. -Vancouver area (Weinberg et al., 1984). Based on the results of the previous surveys and in an at- tempt to further reduce the variance of canary and yellowtail rockfish catch rates, station allocation was changed again in 1986 (Coleman, 1988). In that year, sampling was apportioned among four depth strata: 55-91 m, 92-183 m, 184-219 m, and 220-366 m. Almost three times the effort was applied in the U.S.- Vancouver area, most of which was off northern Washington (lat. 48°00'-42J23'N). However, having not been able to improve rockfish estimates signifi- cantly, the 1989 and 1992 AFSC surveys shifted away from rockfish concerns of past surveys and concen- trated on abundance estimation of Pacific hake and young sablefish, Anoplopoma fimbria. Consequently, the high density rockfish strata were abandoned and sampling was allocated within only two depth strata, 55-183 m and 184-366 m (Weinberg et al., 1994). Samples were collected with standardized Nor'eas- tern high-opening rockfish bottom trawls rigged with roller gear. In general the gear's horizontal and ver- tical openings measured 13 and 9 m, respectively. Towing was controlled by fishing along depth con- tours for one-half hour at about three knots. Catches were sorted by species, weighed, and counted. This study utilizes rockfish catch data from six AFSC multispecies groundfish assessment surveys con- ducted triennially over a 16-year period from 1977 to 1992. Only data from bottom trawling in the Co- lumbia and U.S. -Vancouver INPFC areas (43°00'N to the U.S. -Canada border) were examined. Trawl- ing occurred during August and September between the depths of 55 and 366 m. All surveys employed stratified random sampling designs, apportioning towing sites according to various geographic strata and depth intervals. While the overall multispecies assessment goal remained unchanged from one sur- vey to the next, many of the specific objectives did not. Objectives of the 1977 survey included deter- mining the distribution and abundance of several commercially important rockfishes (Gunderson and Sample, 1980) and the on-bottom component of Pa- Assemblage analyses I examined rockfish associations using two tech- niques: recurrent group analysis and cluster analy- sis. These two methods provide somewhat different characterizations of species distribution and cooccurrence and, when used together, can enhance our understanding of rockfish communities. Recurrent group analysis (RGA), a nonhierarchical technique, addresses the question of which rockfish are likely to be caught together, thus reflecting their 2 Gunderson, D. R., and M. O. Nelson. 1977. Preliminary report on an experimental rockfish survey conducted off Monterey, California, and in Queen Charlotte Sound, British Columbia, during August-September 1976. U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Northwest and Alaska Fish. Cent., 2725 Montlake Blvd. E.. Seattle. WA 98112. Unpubl. manuscr., 82 p. 622 Fishery Bulletin 92(3). 1994 spatial distribution patterns (Fager and McGowan, 1963). Species were included in a group based solely on their presence or absence in catches. Fixed groups were defined as the greatest number of members having affinities with one another based on a 40% affinity threshold. However, because fishermen are concerned prima- rily with abundance in terms of biomass, the find- ings of RGA were supplemented by cluster analysis (CA), a method that incorporates sample catch weights into the grouping process. CA calculates re- semblance measures using the Bray-Curtis dissimi- larity coefficient (Bray and Curtis, 1957) and then clusters agglomeratively, using a flexible sort fusion strategy with an assigned clustering coefficient value (beta) equal to -0.25 (Lance and Williams, 1967). The flexible strategy was selected over other clustering methods because of its tendency to reduce the chain- ing effect often seen in dendrograms. Dendrograms display the similarities among species and groups in hierarchical form, permitting greater flexibility in the interpretation of associations than the RGA tech- nique which produces set groups (Clifford and Stephenson, 1975). Relative groupings can be dis- tinguished at varying levels of dissimilarity where the number 0 indicates greatest resemblance. Prior to classifying assemblages, steps were taken to reduce "noise" in the data. First, groundfish spe- cies other than scorpaenids were eliminated, since the objective of the study was to identify rockfish assemblages without the masking effect caused by the presence of other species. Next, the least frequent rockfishes, defined as those taken at fewer than three stations in each of the groundfish surveys, were omit- ted. Knowledge of the path width of our survey gear facilitated the standardization of catch data as catch per unit of effort (CPUE), i.e. kilograms per hectare towed ( kg/ha )\ or roughly equivalent to towing our gear for 0.8 km. Finally, CPUE's were log transformed (log]Q(x+l)) to reduce the influence of high CPUE's (Boesch4). Results Over the six surveys, a total of 1,874 successful hauls were made in the Columbia and U.S. -Vancouver INPFC areas (Fig. 1). Rockfish were present in 79% :1 1 kg/ha=0.1 t/kra2. ' Boesch, D F. 1977. Application of numerical classification in ecological investigations of water pollution. Virginia Inst. Ma- rine Science, Spec. Sci. Rep. 77, EPA-600/3-77-033, 114 p. Environ. Res. Lab., Off. Res. Dev., U.S. Environ. Protection Agency, Corvallis, OR 97330 ( 1,476) of the tows. Sampling effort was greatest in 1986 and considerably lower during 1977, 1989, and 1992. Most of the effort was applied between the depths of 101 and 200 m (Fig. 2). Sampling at greater depths was proportionally higher in 1977 than dur- ing the other surveys. Catch composition and species diversity Thirty three rockfish species (shortspine thornyhead Sebastolobus alascanus and 32 species of Scbastes) were identified. Among these, 20 were commonly caught in 1,468 hauls and included in the assem- blage analysis (Table 1). Catches of rockfish varied widely in size and com- position. Many of the catches were small: 25% had CPUE's under 1.1 kg/ha and 50% had CPUE's under 4.8 kg/ha. In contrast, 8% were greater than 100 kg/ ha while only 1% were greater than 500 kg/ha. Maxi- mum rockfish CPUE's reached 4,126, 564, 1,253, 759, 2,303, and 828 kg/ha during the six respective sur- veys. The average CPUE for each survey was 50, 16, 34, 21, 34, and 27 kg/ha, respectively. On average, abundance levels increased in deeper water, peak- ing in the 151-250 m depth interval (Fig. 3). Species diversity in survey catches depicts the multispecies nature of the rockfish community vul- nerable to the bottom trawl. Eighty three percent of rockfish samples contained more than one species. Of these, approximately 50% contained 2-5 species, 29% contained 6-10 species, and 4% contained 10- 16 species. Of the single-species catches, 78% were under 1 kg/ha. In contrast, the two largest single- species catches (canary rockfish) exceeded 100 kg/ ha. Eighty five percent of the single-species samples were either shortspine thornyhead (7%), canary ( 26%' ), darkblotched (21%), yellowtail ( 18% ), or green- striped, Sebastes elongatus (14%), rockfish. Silver- gray, S. brevispinis, rosethorn, S. helvomaculatus, redbanded, S. babcocki, and yellowmouth, S. reedi, rockfish were never caught alone. Species diversity, like abundance, increased with depth (Fig. 4). Hauls made at the shallowest sam- pling sites (-55 m) demonstrated little variety. As sampling depth increased, nearshore species, such as black, S. melanops, and quillback, S. maliger, rock- fish, were replaced by offshore rockfishes, including juveniles of many species that inhabit even deeper waters as adults. Over the middle-shelf, within the 55-150 m depth interval, up to 13 rockfish species were taken in a single tow. About 22% of the hauls made at these depths contained five or more species (Table 2). Species diversity peaked along the outer- shelf, where the centers of abundance for several species overlapped. In waters 151-250 m deep, catches Weinberg. Rockfish assemblages off Oregon and Washington 623 •«2* $r +* + '/a*'."* >4 * *l *&♦* ♦ ♦•" ♦"* >* ♦♦* fev ♦ +♦+., ***** ♦»* . *+ .*r** »Vi *3& »* •: AY X** * *•.- ***♦ ,t rs'-i 1977 1980 1983 1986 1989 48 47 - 46 - 45 44 43 125 124 125 124 125 124 I25 124 125 124 125 124 123 Longitude Figure 1 The distribution of hauls from the west coast triennial groundfish surveys showing the presence ( + ) or absence (o) of rockfish (Scorpaenidae) in the Columbia and U.S. -Vancouver areas. - id -C :. :.. :: .:_' ~S£ZZ^Z2£ m 2 - ^ 0 - 50 100 150 200 250 300 350 400 Depth (m) Figure 4 Number of rockfish species (Scorpaenidae) per tow (dots) and averaged over 55-100, 101-150, 151- 200, 201-250, 251-300, and 301-366 m depth intervals (squares). Weinberg: Rockfish assemblages off Oregon and Washington 627 Table 3 Major species groups (1-3) of rockfish (Scorpaenidae) determined by recurrent group analysis by survey and for all surveys combined. Depth (m) and catch per unit of effort (CPUE, kg/ha) statistics refer to samples when all species listed in the group cooccurred. The group mean catch is also presented as the percent of the total average rockfish catch in those hauls. See Table 1 for scientific names. Species 1977 1980 His:! 1986 19.H9 1992 All years Shortspine thornyhead Pacific ocean perch Splitnose rockfish Darkblotched rockfish Redbanded rockfish Canary rockfish Yellowtail rockfish Greenstriped rockfish Sharpchin rockfish Rosethorn rockfish Redstripe rockfish 1 1 1 1 1 1 1 1 1 — 1 1 1 1 1 1 1 1 — 1 1 1 2 1 1 — — 1 1 1 2 2 2 2 3 — 2 2 — 2 2 — 2 2 2 2 1 3 3 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 — 3 — — 3 3 Depth and catch statistics Group 1 Group 2 Group 3 1977 Occurrences Mean depth (range) Mean group CPUE (range I Proportion of total CPUE (%) 1980 Occurrences Mean depth (range I Mean group CPUE (range) Proportion of total CPUE (%) 1983 Occurrences Mean depth (range) Mean group CPUE (range) Proportion of total CPUE (%) 1986 Occurrences Mean depth (range) Mean group CPUE (range) Proportion of total CPUE (ri ) 1989 Occurrences Mean depth (range) Mean group CPUE (range) Proportion of total CPUE (%) 1992 Occurrences Mean depth (range) Mean group CPUE (range) Proportion of total CPUE (%) All years Occurrences Mean depth (range) Mean group CPUE (range) Proportion of total CPUE (% ) 41 34 31 255(155-344) 163(104-291) 217(108-315) 48.7 (6.2-496.0) 57.6(1.3-599.4) 11.2(0.2-72.4) 79.8 63.5 9.5 19 37 24 249(150-338) 154(82-274) 195(144-348) 39.7(0.9-155.7) 14.0(0.2-116.7) 11.2(0.1-94.4) 73.6 23.6 13.1 31 80 26 213(154-293) 143(59-251) 196(124-293) 32.2(1.7-126.0) 36.1 (0.6-742.5) 71.1 (0.1-681.8) 67.9 47.1 34.5 31 73 34 233(150-348) 124(59-185) 165 (97-267) 37.3(1.5-417.8) 48.0(0.1-745.3) 9.2(0.3-87.4) 89.2 79.7 15.8 27 22 230(132-353) — 174(112-238) 9.8(0.3-64.8) — 22.7(0.6-92.4) 36.3 — 23.3 9 24 15 230(214-260) 155 (119-219) 167 (113-223) 31.0(7.9-98.8) 25.6(1.3-283.3) 120.8(0.7-533.4) 53.5 76.2 67.3 135 278 89 259(155-366) 141 (59-291) 173(97-293) 45.9(1.7-496.0) 53.3(0.1-1205.2) 26.1 (0. 3-684.7i 80.3 62.8 12.5 628 Fishery Bulletin 92(3), 1994 — 0 1980 210 hauls V - o 0 3 i lz ,1 i: ^ 7 - r 1 IS p— r P-i 20 1 1 1 I m^m^m^w ^^^^w^' ' * r- 1983 1986 341 hauls I IT! 1992 180 hauls 30 dii All Years 14 66 hauls MM ffl] «■*■ 30 WATER TEMPERATURE (*C) Figure 2 Distribution of catch of postlarval penaeid shrimp, Penaeus aztecus and Penaeus setiferus, over the range of water temperature observed at Breach Inlet during regular sampling. ships.2 Numbers of shrimp landed each year were estimated by multiplying landed weight by average grade (average number of shrimp per kg). Results A total of 102,109 Penaeus postlarvae were collected from 7 January 1975 to 3 August 1992 at Breach Inlet. Of the total catch, 68.3% were identified as brown shrimp, P. aztecus, 23.3% as white shrimp, P. setiferus, and 0.8% as pink shrimp, P. duorarum. The remaining 7.6% were identifed as Penaeus spp. The majority of the latter category were tentatively iden- tified as P. aztecus, and were collected primarily from late May through July. Because P. duorarum postlarvae were collected in relatively low numbers and some uncertainty exists with identification, the analysis includes only Paztecus and P. setiferus. The majority off! aztecus postlarvae were collected between February and April, when water temperature was between 10 and 20°C (the catch peaked between 12° and 16°C ). Penaeus setiferus was collected only when temperatures exceeded 20°C; most were taken at tem- peratures between 25 and 30°C (peak abundance oc- curred in June; Fig. 2). Samples off! aztecus averaged several hundred individuals in each year during times of peak ingress. However, catches of f! setiferus were Table 1 Average and range of bottom water temperature and salinity observed at Breach Inlet, South Ca rolinaand number of samples of Penaeus postlarvae collected during regular sampling , 1975-92. Number of Year Temperature CO Salinity (ppt) samples 1975 20.2(9.3-29.4) 28.6(15.0-35.0) 256 1976 19.4(7.0-30.0) 30.7(26.0-35.0) 236 1977 21.4(4.9-32.1) 31.5(21.0-36.0) 244 1978 21.2(5.9-31.5) 29.8(18.0-35.0) 117 1979 18.8(6.7-28.2) 28.8(24.0-35.0) 57 1980 15.3(5.7-25.6) 28.8(20.0-34.0) 34 1981 18.3(10.9-31.0) 32.4(25.0-36.0) 46 1982 18.9(9.5-28.1) 28.5(22.0-32.0) 22 1983 14.9(8.1-26.2) 27.6(18.0-32.0) 33 1984 23.7(8.1-30.5) 29.0 (23.0-34.0) 63 1985 21.8(11.3-30.7) 31.7 (26.0-36.0) 48 1986 20.8(9.1-30.0) 31.1 (27.0-34.0) 36 1987 17.4(8.0-29.2) 32.1(26.0-36.0) 36 1988 19.4(11.2-30.5) 29.1 (21.0-35.0) 38 1989 20.8(10.5-30.6) 32.7(26.0-36.0) 48 1990 21.3 (11.9-30.1) 31.1 (25.0-35.0) 42 1991 20.2(10.4-29.4) 29.4(21.0-35.0) 36 1992 21.3(9.3-30.1) 29.1 (24.0-33.0) 34 2 Low, R. A. 1992. Survey of the South Carolina Shrimp baiting fish- ery, 1991. South Carolina Marine Res. Cent. Data Rep. 9, 29 p. much more variable, averaging zero in some years and several hundred in years of high abundance (Table 2). Consecutive sampling conducted day and night on both ebb and flood tides revealed that both species 636 Fishery Bulletin 92(3), 1994 3.5 1 1 i i i i Q 30 A CO (4 (50) (4 1) (4 ?' (79, co Z 20 LU Q - I , 1 » 1 (80) II O 15 o _l i 1 II Z 1.0 < LU 5 05 i 1 I I - 0.0 DAY NIGHT SURF BOTT FLOOD EBB PHYSICAL FACTORS 3.5 1 III (2 V (24) a 30 co - (2 !> (i 1) B (37) +' 2.5 c ) c ) 2 20 LU - c 3 C ) c ) Q CD 15 o _l - (4_ 0) z 10 < LU 5 0.5 1 1 c ) 0 0 DAY NIGHT SURF BOTT FLOOD EBB PHYSICAL FACTORS Figure 3 Mean density of postlarval penaeid shrimp (A) Penaeus aztecus and (B) Penaeus setiferus grouped by time of day, tide, and location of net at Breach Inlet, South Carolina. Samples were collected over 24-hour periods during times of peak abundance. Log Density= log(( number/1, 000m3) +1). SD=standard de- viation. Numbers in parenthesis=number of samples. Surf=surface, Bott=bottom. were significantly more abundant in samples taken during flood tides than during ebb tides (P<0.001; Fig. 3). Significantly more P. aztecus were collected at night on flood tides ( 3c=228.7/l,000m3) than dur- ing daylight flood tides ( 3c = 147. 4/1, 000m3; P<0.001), whereas no significant difference between catches made during day ( 3c=877.9/l,000m3) versus night (3c =610. 2/1, 000m3) were noted for P. setiferus (P=0.114; Fig. 3). No significant differences in catches were noted for either species when surface and bot- tom collections were compared (P=0.595 for P. aztecus; P=0.270 for P. setiferus; Fig. 3). Significant differences among catches made by regular sampling grouped by lunar phase ( new moon, first quarter, full moon, last quarter) were detected for both P. aztecus (P<0.0001) and P. setiferus (P<0.006). Fewest postlarvae of both species were collected during full moon phases; P. aztecus were more abundant during the last quarter than at other times (Fig. 4), whereas P. setiferus was most abun- dant during the first quarter. Results of correlation analysis revealed a low cor- relation (r2=0.03; P=0.523) between annual indices of abundance of postlarval P. aztecus and subsequent landings. A much higher correlation was obtained for P. setiferus (r2=0.79;P<0.001; Fig. 5). The regres- sion equation for estimated number of P. setiferus landed was Y= 40.4 + 119.0X(y=estimated number landed, X=annual index of postlarval abundance). Discussion Our results are generally similar to observations on Penaeus postlarvae made by others in the southeast- ern United States (including the Gulf of Mexico). Penaeus aztecus is most abundant in springtime at water temperatures comparable to the range of tem- peratures observed in this study (Bearden, 1961; Williams and Deubler, 1968; Allen et al., 1980). It has been postulated that the majority of brown shrimp postlarvae recruited in the spring are pro- duced by spawning from the previous fall (Temple and Fischer, 1967; Aldrich et al., 1968; Anderson, 1970; Whitaker3). Ingress begins after nearshore water temperatures approach 12°C. In contrast, Penaeus setiferus postlarvae are recruited in spring and summer shortly after being spawned in adjacent oceanic waters (Pearson, 1939; Lindner and Ander- son, 1956). When compared to overall catches of P setiferus, P. aztecus have often been collected in greater abundance in South Carolina (Bearden, 1961; Lunz, 1965; Olmi, 1986; Wenner and Beatty, in press) and in Texas (Baxter and Renfro, 1967). Prior studies have shown thatP aztecus postlarvae are often more abundant in nighttime collections and on high (flood) tides (Caillouet et al., 1970; Duronslet et al., 1972), similar to the pattern observed in this study. This is probably due to both increased noctur- nal activity and decreased gear avoidance at night 3 Whitaker, J. D. 1982. A possible mechanism for brown shrimp postlarval recruitment. Paper presented at South Carolina Fish- eries Workers Assoc, annu. meet., Clemson, SC, 24-25 Feb. 1982. Unpubl. manuscr., 3 p. NOTE DeLancey et al.: Seasonal sampling of Penaeus postlarvae 637 Table 2 Mean density ( number pe r 1,000m3), standard deviation, and number of samples collected for brown shrimp. Penaeus aztecus, and white shrimp, Penaeus setiferus. postl irvae dur ng season of peak ingress at Breach I nlet. South Caroling 1975-92. Year P. aztecus P. setiferus Mean SD N Mean SD N 1975 149.9 185.21 98 78.3 265.13 124 1976 119.2 184.83 100 51.7 90.65 104 1977 75.4 105.53 72 0.0 0.00 136 1978 90.1 150.80 40 0.3 1.31 53 1979 86.1 123.80 28 48.5 84.33 25 1980 209.7 144.08 20 2.6 4.17 10 1981 81.1 124.45 26 0.0 0.00 18 1982 — — — 9.4 18.16 10 1983 291.2 417.26 13 3.8 6.19 6 1984 125.4 192.53 10 10.5 36.33 39 1985 68.7 101.65 18 2.7 7.35 24 1986 182.7 177.05 14 4.8 19.55 20 1987 58.3 37.38 13 24.7 55.61 14 1988 69.2 112.20 14 10.8 31.55 23 1989 95.2 83.52 20 671.2 1,627.30 25 1990 336.8 348.82 18 29.5 75.82 21 1991 132.1 170.84 18 351.6 452.27 16 1992 129.7 200.25 12 326.6 1,035.87 19 (Williams and Deubler, 1968; Matthews et al., 1991 ). Olmi ( 1986) in addition to collecting more P. aztecus at night, also collected the majority of P. setiferus postlarvae at night in a tidal creek in South Caro- lina. These findings are in agreement with similar studies in the Gulf of Mexico, where P. setiferus was found to be more abundant near the surface at night in an inlet (Duronslet et al., 1972) and over a tidal flat (Caillouet et al. , 1968 ). Wenner and Beatty ( 1993 ) collected 951 of all penaeid postlarvae at night (in- cluding a collection at a creek adjacent to Breach Inlet). In our study, no difference was detected be- tween day and night catches off! setiferus, possibly because we employed a larger net than in the other studies conducted in South Carolina in a higher- velocity tidal flow. This may have minimized avoid- ance of the gear by postlarvae during daylight. Addi- tional sampling may clarify these observations. Similar to observations made in our study, Olmi ( 1986) generally collected fewer postlarvae during full moon phases in South Carolina. Williams and Deubler (1968) in North Carolina and Allen et al. (1980) in the Florida Keys collected fewer P. duor- arum at night during full moon phases than during new moon phases. Perhaps high light levels at night may delay ingress for several days, causing postlarvae to remain on offshore substrates ( Matthews et al., 199 1 ). -0.5 3.0 <=i 2.5 CO > 2.0 H W NEW MOON FIRST QTR FULL MOON LAST QTR LUNAR PHASE LU Q CD O 1.5 1.0 0.5 < ty o.o -0.5 B ! - (163) " (172) (200) - (149) - - ) o ( - NEW MOON FIRST QTR FULL MOON LAST QTR LUNAR PHASE Figure 4 Mean density of postlarval penaeid shrimp (A) Penaeus aztecus and (B) Penaeus setiferus grouped by lunar phase during months of peak abundance at Breach Inlet. Samples collected near the bottom on flood tide during daylight ( regular sampling). Log Density=log ( ( number/ 1,000 m3) + 1). SD=standard deviation. Qtr= quarter. Numbers in parenthesis=number of samples. Peaks in abundance of P. setiferus postlarvae may be related to prior spawning activity around new or full moon phases, but this relationship is uncertain at present and warrants further investigation. A relationship between recruitment of P. aztecus postlarvae and subsequent commercial landings has been difficult to demonstrate (Williams, 1969; Ford and St. Amant, 1971; Baxter and Sullivan4). Gener- Baxter. K. N„ and L. F. Sullivan. 1986. Forecasting offshore brown shrimp catch from early life history stages. In Proc. shrimp yield workshop. Tex. A&M Univ. Sea Grant Rep. TAMU- SG-86-10:22-36. 638 Fishery Bulletin 92(3). 1994 360 O 32° _i _l 5 280 Y ■ 40 4 • 119.0 (X) ^^ r2-0.79 s^ LANDED ( o o X x jS1^ EC LU CD 160 2 z> Z 120 Q LJJ < 80 CO 40 LU x x \s^ X ^S*^ X X s' /^ X 0 0 « X 0 0.5 1.0 1.5 20 2.5 MAY-AUGUST INDEX OF POSTLARVAL ABUNDANCE Figure 5 Plot of annual indices of abundance of postlarval white shrimp, Penaeus setiferus, versus estimated number of shrimp landed in the fall commercial and recreational fishery, 1975-92. ally, environmental conditions in the nursery area, e.g. spring temperature and salinity (with related factors such as rainfall, river discharge, and meteo- rological conditions; Gaidry and White, 1973; Barrett and Gillespie, 1975; Zein-Elden and Renaud, 1986; Childers et al., 1990) are thought to be important influences on production. Biological factors such as predation and secondary production have also been postulated to influence yield (Hunter and Feller, 1987; Gleason and Wellington, 1988; Minello et al., 1989). Postlarval indices have been used with some success in predictive models that incorporate envi- ronmental variables and indices of juvenile shrimp abundance (Sutter and Christmas, 1983; Baxter et al., 1988). We have been unsuccessful in efforts to produce a model for brown shrimp production using data from consectutive years, although our postlar- val index may be useful in the future. Undoubtedly many factors influence the produc- tion of P. setiferus populations, but recent studies have demonstrated that commercial harvest of P. setiferus can be modeled (Lam et al., 1989) and ap- parent spawner-recruit relationships have been de- scribed in South Carolina (Lam et al., 1989) and in the Gulf of Mexico ( Nance and Nichols, 1988;Gracia, 1991). Our study demonstrates that monitoring off! setiferus postlarvae can be a reliable indicator of harvest. Expanded sampling effort, i.e. more locations and increased numbers of samples, would perhaps yield more statistically significant results than were ob- tained at a single location in our study. These data do, however, represent one of the longest-term stud- ies of postlarval penaeid recruitment to date. In ad- dition to contributing to our overall understanding of penaeid shrimp population dynamics, our baseline monitoring effort may be useful as a management tool for predicting harvest and providing advice on optimal times for flooding coastal impoundments for extensive aquaculture. Acknowledgments Numerous individuals have contributed to this study, notably C. Bearden, C. Boardman, C. Farmer, L. Leseman, and T. Read. M. Clise assisted with data preparation and K. Swanson prepared the map. E. Wenner and B. Stender reviewed the manuscript. Literature cited Aldrich, D. V., C. E. Wood, and K. N. Baxter. 1968. An ecological interpretation of low temperature responses in Penaeus aztecus and P. setiferus postlarvae. Bull. Mar. Sci. 18:61-71. Allen, D. M., J. H. Hudson, and T. J. Costello. 1980. Postlarval shrimp {Penaeus) in the Florida Keys: species, size and seasonal abundance. Bull. Mar. Sci. 30:21-34. Anderson, W. W. 1970. Contributions to the life histories of several penaeid shrimps (Penaeidae) along the South At- NOTE DeLancey et al.: Seasonal sampling of Penaeus postlarvae 639 lantic coast of the United States. U.S. Fish Wild. Serv. Spec. Sci. Rep. 605, 24 p. Barrett, B. B., and M. C. Gillespie. 1975. 1975 environmental conditions relative to shrimp production in coastal Louisiana. La. Wild. Fish. Comm. Tech. Bull. 15, 22 p. Baxter, K. N. 1963. Abundance of postlarval shrimp — one index of future fishing success. Proc. Gulf Caribb. Fish. Inst. 15:79-87. Baxter, K. N., and W. C. Renfro. 1967. Seasonal occurrence and size distribution of postlarval brown and white shrimp near Galveston, Texas, with notes on species identification. Fish. Bull. 66:149-158. Baxter, K. N., C. H. Furr Jr., and E. Scott. 1988. The commercial bait shrimp fishery in Galveston Bay, Texas, 1959-87. Mar. Fish. Rev. 50:20-28. Bearden, C. M. 1961. Notes on postlarvae of commercial shrimp (Penaeus) in South Carolina. Contrib. Bears Bluff Lab. 33, 8 p. Caillouet, C. W., Jr., B. J. Fontenot, Jr., and R. J. Dugas. 1968. Diel fluctuations in catch of postlarval white shrimp, Penaeus setiferus (Linnaeus), with Renfro beam trawl. Bull. Mar. Sci. 18:829-835. Caillouet, C. W., Jr., W. S. Perret, and R. J. Dugas. 1970. Diel fluctuations in catch of postlarval brown shrimp, Penaeus azteeus Ives, with the Renfro beam trawl. Bull. Mar. Sci. 20:721-730. Childers, D. L., J. W. Day, and R. A. Muller. 1990. Relating climatological forcing to coastal wa- ter levels in Louisiana estuaries and the potential importance of El Nino-southern oscillation events. Clim. Res. 1:31-42. Christmas, J. Y., G. Gunter, and P. Musgrave. 1966. Studies of annual abundance of postlarval penaeid shrimp in the estuarine waters of Missis- sippi, as related to subsequent commercial catches. Gulf Res. Rep. 2:177-212. Duronslet, M. J., J. M. Lyon, and F. Marullo. 1972. Vertical distribution of postlarval brown, Penaeus azteeus, and white shrimp, Penaeus setiferus, during immigration through a tidal pass. Trans. Am. Fish. Soc. 101:748-752. Elliot, J. M. 1977. Some methods for the statistical analysis of samples of benthic invertebrates. Freshwater Biol. Assoc. Sci. Pub. 25, 160 p. Ford, T. B., and L. St. Amant. 1971. Management guidelines for predicting brown shrimp, Penaeus azteeus, production in Louisiana. Proc. Gulf Caribb. Fish. Inst. 23:149-161. Gaidry, W. J., Ill, and C. J. White. 1973. Investigations of commercially important penaeid shrimp in Louisiana estuaries. La. Wild. Fish. Comm. Tech. Bull. 8, 154 p. George, M. J. 1962. Preliminary observations of the recruitment of postlarvae and growth of juveniles of the brown shrimp Penaeus azteeus Ives in Barataria Bay. La. Wild. Fish. Comm. Bien. Rep. 9:160-163. Gleason, D. F., and G. M. Wellington. 1988. Food resources of postlarval brown shrimp (Penaeus azteeus) in a Texas salt marsh. Mar. Biol. 97:329-337. Gracia, A. 1991. Spawning stock-recruitment relationships of white shrimp in the southwestern Gulf of Mexico. Trans. Am. Fish. Soc. 120:519-527. Hunter, J., and R. J. Feller. 1987. Immunological dietary analysis of two penaeid shrimp species from a South Carolina tidal creek. J. exp. Mar. Biol. Ecol. 107:61-70. Lam, C. F., J. D. Whitaker, and F. S. Lee. 1989. Model for white shrimp landings for the cen- tral coast of South Carolina. North Am. J. Fish. Manage. 9:12-22. Lindner, M. J., and W. W. Anderson. 1956. Growth, migrations, spawning and size dis- tributions of shrimp, Penaeus setiferus. Fish. Bull. 56:553-645. Loesch, H. 1965. Distribution and growth of penaeid shrimp in Mobile Bay, Alabama. Publ. Inst. Mar. Sci., Univ. Tex. 10:41-58. Matthews, T. R., W. W. Schroeder, and D. E. Stearns. 1991. Endogenous rhythm, light and salinity effects on postlarval brown shrimp Penaeus azteeus Ives recruitment to estuaries. J. exp. Mar. Biol. Ecol. 154:177-189. Minello, T. J., R. J. Zimmerman, and E. X. Martinez. 1989. Mortality of young brown shrimp Penaeus azteeus in estuarine nurseries. Trans. Am. Fish. Soc. 118:693-708. Nance, J. M., and S. Nichols. 1988. Stock assessments for brown, white and pink shrimp in the U.S. Gulf of Mexico, 1960- 1986. NOAA Tech. Memo. SEFC-NMFS-203, 64 p. Olmi, E. J., III. 1986. Factors affecting the abundance of penaeid shrimps in macroplankton samples from open and impounded brackish marsh systems in South Carolina. M.S. thesis. College of Charleston, Charleston, South Carolina, 95 p. Pearson, J. C. 1939. The early life histories of some American Penaeidae, chiefly the commercial shrimp, Penaeus setiferus (Linn.) Fish. Bull. 49:1-73. Ringo, R. D., and G. Zamora Jr. 1968. A penaeid postlarval character of taxonomic value. Bull. Mar. Sci. 18:471-476. Siegel, S. 1956. Nonparametric statistics for the behavioral sciences. McGraw-Hill, New York, 312 p. Sutter, F. C, and J. Y. Christmas. 1983. Multilinear models for the prediction of brown 640 Fishery Bulletin 92(3), 1994 shrimp harvest in Mississippi waters. Gulf Res. Rep. 7:205-210. Temple, R. F., and C. C. Fischer. 1967. Seasonal distribution and relative abundance of planktonic-stage shrimp (Penaeus spp.) in the northwestern Gulf of Mexico, 1961. Fish. Bull. 66:323-334. Wenner, E. L., and H. R. Beatty. 1993. Utilization of shallow estuarine habitats in South Carolina, USA, by postlarval and juvenile stages of Penaeus spp. (Decapoda: Penaeidae). J. Crust. Biol. 13(2):280:295 Williams, A. B. 1959. Spotted and brown shrimp postlarvae (Penaeus ) in North Carolina. Bull. Mar. Sci. 9:281-290. 1969. A ten-year study of meroplankton in North Carolina estuaries: cycles of occurrence among penaeidean shrimps. Chesapeake Sci. 10:36-47. Williams, A. B., and E. E. Deubler. 1968. A ten-year study of meroplankton in North Caro- lina estuaries: assessment of environmental factors and sampling success among bothid flounders and penaeid shrimps. Chesapeake Sci. 9:27^11. Zein-Elden, Z. P., and M. L. Renaud. 1986. Inshore environmental effects on brown shrimp, Penaeus aztecus, and white shrimp, P. setiferus, populations in coastal waters, particularly of Texas. Mar. Fish. Rev. 48 (3):9-19. Swimbladder deflation in the Atlantic menhaden, Brevoortia tyrannus Richard B. Forward Jr. Marine Laboratory, School of the Environment, Duke University Pivers Island, Beaufort, North Carolina 28516-9721 William F. Hettler Donald E. Hoss Beaufort Laboratory, Southeast Fisheries Science Center National Marine Fisheries Service, NOAA Beaufort, North Carolina 28516-9722 Larval clupeoid fishes usually have a pronounced cycle of swimbladder inflation and deflation (Uotani, 1973; Hunter and Sanchez, 1976; Blaxter and Hunter, 1982). Field and laboratory studies of both At- lantic {Brevoortia tyrannus; Hoss et al., 1989) and gulf (Brevoortia pat- ronus; Hoss and Phonlor, 1984) menhaden found that larvae in- flated their swimbladders during the night and deflated them dur- ing the day. Our past studies of Atlantic men- haden larvae found that the cue for inflation is a decrease in light in- tensity at sunset (Hoss et al., 1989; Forward et al., 1993). Inflation oc- curs rapidly and begins within 5 minutes of onset of darkness. The process involves moving to the sur- face, swallowing air into the ali- mentary canal, and moving this air into the swimbladder through the pneumatic duct (Hoss et al., 1989). Menhaden have no connection (pneumatic duct) between the swimbladder and anus as do some clupeiods (Tracy, 1920). Deflation is less studied. It is hypothesized to occur by diffusion of gas from the swimbladder throughout the night and perhaps by active movement of gas to the alimentary canal and then out through the mouth and anus (Hoss and Phonlor, 1984). The present study was under- taken to determine 1) the manner in which swimbladder deflation oc- curs (by diffusion or active gas movement) in Atlantic menhaden larvae, 2 ) the relationship between deflation and light intensity, 3) the time-course for deflation, and 4 ) the presence or absence of an endog- enous rhythm in deflation. Materials and methods Atlantic menhaden, Brevoortia tyrannus, were spawned and reared in the laboratory (Hettler, 1983) on a 12:12 hour light-dark cycle with the dark phase beginning at 1900 hours. Lighting during the light phase was provided by daylight fluorescent tubes at a surface inten- sity of 1.6xl015 photonscm~2-s_1 (400-700 nm) as measured with a scalar irradiance meter with a 4tc collector (Biospherical Instru- ments, Inc.). Our previous study found that swimbladder inflation began when larvae were 10 mm to- tal length (TL) but the percentage with inflated swimbladders was low (Forward et al., 1993). Between 11 and 16 mm TL, swimbladders were inflated during the night and deflated during the day. Above 16 mm TL, most fish always had some gas in their swimbladders. Since the percentage of larvae with de- flated swimbladders varies be- tween day and night for 11-16 mm TL larvae, they were used in the present experiments. Deflation was quantified by de- termining the proportion of larvae with deflated swimbladders. In ad- dition, inflation was quantified by measuring the size of the light-re- fractive bubbles in the swimbladder and alimentary canal to the near- est 0.02 mm under a microscope. It was assumed that bubbles in the alimentary canal were transported either toward or away from the swimbladder and thereby contrib- uted to the swimbladder volume. Gas bubble volume d>) was calcu- lated by using the equation of Hunter and Sanchez (1976): V=(AI 3)7i ab'~, where 6=half the bubble width and o=half the bubble length. Swimbladder volume was the total volume of all bubbles. Since swimbladder volume increases with larval length (Forward et al., 1993), this relationship is pre- sented when volume is considered. Only larvae with inflated swim- bladders were used to calculate the mean volume at each larval length. In contrast, the percentage of lar- vae with deflated swimbladders was calculated for all larvae (11- 16 mm TL), because the previous study found that the proportion of larvae inflating swimbladders did not vary significantly with larval length (Forward et ai., 1993). Four sets of experiments were conducted, all of which began by removing larvae from rearing tanks and by placing them in darkness at the time of the beginning of the dark phase. Our previous study showed that darkness cued initial Manuscript accepted 22 February 1994. Fishery Bulletin 92:641-646 1 1994). 64 1 642 Fishery Bulletin 92(3). 1994 swimbladder inflation at the beginning of the night (Forward et al., 1993). In all experiments each larva was used only once. It was assumed that all test lar- vae could potentially inflate their swimbladders in darkness. However, the maximum percent inflation was around 92% (Fig. 1), which suggests that about 8% of the larvae were developmentally incapable of inflation. This low percentage would not alter the overall result of any experiment. The first experiment measured swimbladder de- flation in larvae kept in continuous darkness with and without access to air. Deflation was defined as the absence of gas bubbles in the swimbladder and alimentary canal. Three hours after the beginning of night (2200 hours), a subsample of the larvae was removed from the dark and measured for total length, presence of gas bubbles in the swimbladder and ali- mentary canal, and for size of the bubbles (standard measurements). There was no evidence that larvae lost or took up gas during the measurement proce- dure. The remaining larvae were separated into two groups. The first group remained in finger bowls ( 19.3 cm diameter) in darkness with access to the air-wa- ter interface and was similarly sampled during the following day at times (0900, 1200, and 1700 hours) that should have occurred during the normal light phase. The bowls ( 19.3 cm diameter) containing the second group of larvae were sealed at 2200 hours, so that larvae did not have an air-water interface. The seal was accomplished by filling the bowl completely / / E 12 / / ^ 1/ p *lf > // / Qlr ^ // / 3h a> 10 "O / / / "O Q / / / /* no °" ZJ / / / / l4h E 80 / / // s y / // CO / / // 6.0- / J // / ~^^~^ / / I no air / />/ "~~~~ 23 h 4.0 / J^ 2.0- /y^ —t i — . 12 13 14 15 Total Length (mm) 16 Figure 2 Swimbladder volume for different size larval Atlantic menhaden, Brevoortia tyrannus, with (air) and without (no air) access to the air-water interface. The times (e.g. 3 hours) indicate the time in darkness after the begin- ning of the dark phase when larvae were sampled. Means are plotted and the average sample sizes for calculating the means in each experiment are as follows: air for 17- 23 hours=19; air for 14 hours=9; air for 3 hours=19; no air for 14 hours=7; no air for 23 hours=6. The asterisks indicate the mean volume was significantly different (P<0.05; Student's /-test) from the mean volume 3 hours after the beginning of the dark phase (air, 3 hours). darkness. When larvae had access to the air-water interface, there was a significant (P<0.05; Z-test) decrease in the percentage of fish with deflated swimbladders between three hours after beginning of the dark phase (23%) and two hours after the time for beginning of the light phase (8%). This signifi- cant decrease remained throughout the time for the light phase. In contrast, when fish had no access to the air-water interface, the percent deflation did not change significantly (Z-test) over these time inter- vals (Fig. 1). Thus, larvae do not sequentially inflate their swimbladders at sunset and then deflate them by gas diffusion or active processes by sunrise. Be- cause the percentage of larvae with a deflated swimbladder decreased through the night when lar- vae had access to the air-water interface, larvae ap- pear to continue to inflate their swimbladders when in darkness. Measurements of swimbladder volume (Fig. 2) support this suggestion. The swimbladder volumes over time in darkness with and without access to the air-water interface (Fig. 2) were compared to volumes after inflation at the beginning of the night phase (air, 3 hours) to indicate volume changes due to bubble ingestion and removal, and gas diffusion. Swimbladder vol- ume increased in darkness when fish had access to the air-water interface and decreased when they lacked access. These changes were apparent after 14 hours in darkness but only become statistically significant (P<0.05; Student's /-test) after 17-23 hours (Fig. 2). Thus, larvae with access to the air-water inter- face continued to actively inflate their swimbladder, whereas swimbladder volume slowly decreased in larvae that lacked this access. Nevertheless, shortly after the beginning of the light phase (Fig. 2; no air, 14 hours), the volumes in larvae without ac- cess to air were not significantly lower (P>0.05; Student's /-test) than levels after inflation at the beginning of the dark phase (air, 3 hours). There- fore, diffusion of gas from the swimbladder played a very small role in normal deflation at sunrise. Relation of swimbladder deflation to light Larvae deflated their swimbladders upon exposure to light. This response was demonstrated in the initial experiment, in which, at the end of the night, larvae were either maintained for three hours in darkness or exposed to white light for this time. The percentage of larvae with a deflated swim- bladder in darkness was 12% (rc=75), whereas the percentage in light was significantly (P<0.001; Z- test) greater at 84% (n=25). Further studies indi- cated percent deflation depended upon light inten- 644 Fishery Bulletin 92(3). 1994 sity (Fig. 3), as the percent deflation increased as light intensity increased. The lowest light intensity to evoke a significant (P<0.05; Z-test) increase in de- flation (threshold) was 9xl012 photonscm~2s_1. 01 Q 80 60 40 jz 20 A<$ /92 *^^ ^^2 ^-"68 Dork 75 27 35 10 10" 10' io1- 10 2,.-lN io- Light Intensity (photons cm" s") Figure 3 The percentage of larval Atlantic menhaden, Brevoortia tyrannus, with deflated swimbladders upon exposure to dif- ferent light levels and darkness. The number below each point is the sample size. The asterisks indicate the lowest light level at which the proportion deflated was significantly (P<0.05; Z-test) greater than the level in darkness. 80 t 60 ill 40 O 20 * T/ 20 40 60 80 Time (mm) 100 120 Figure 4 The percentage of larval Atlantic menhaden, Brevoortia tyrannus, with deflated swimbladders after different times in light. Means and standard errors are plotted. The average number of replicates was four. The asterisk indicates the first time at which the mean percent was significantly (P<0.05; <-test) greater than the initial level (time 0). Timing of swimbladder deflation The time course of swimbladder deflation in response to white light was measured upon transfer from dark- ness to 1.7xl015 photonscm~2-s_1. By producing the maximum rate of light intensity change, it was assumed that the maximum rate of deflation would be evoked. An increase in the percent deflated was evident after 5 minutes and was significantly (P<0.05; f-test) greater than the initial level after 15 minutes (Fig. 4). Micro- scopic examination indicated this rapid defla- tion was accomplished by passing bubbles from the swimbladder into the gut and then out through both the anus and mouth. Endogenous rhythm in swimbladder deflation The percent deflation remained low if larvae were kept in continuous darkness ( Fig. 1 ), which indicates there was no endogenous rhythm in deflation without exposure to light. However, light-cued deflation was not constant over time ( Fig. 5 ). This experiment was conducted twice and involved maintaining larvae in darkness and measuring a subsample after exposure to light for one hour at different times during the solar day. The consistent cycle in both trials was that light-cued deflation was low during the normal dark phase and high during the normal day phase. An interesting observation was made during these experiments. At the last sampling time, larvae had been in darkness for 27 hours, over which time they continued swimbladder inflation. Exposure to light at 2200 hours caused minimal deflation (Fig. 5). The water containing these larvae also had brine shrimp nauplii from the rearing tank. Larvae are visual predators and had, in effect, been starved for 27 hours in darkness. Under these conditions, once exposed to the light, they began to feed. While measuring swimbladder volumes, their digestive tracts were full of nauplii. Thus, larvae can feed with an inflated swimbladder. Discussion Swimbladder deflation by larval Atlantic menhaden is cued by an increase in light intensity. There was no cycle in which larvae inflated their swimbladders at sunset (Forward et al., 1993) and then deflated them gradually over time. In darkness at the end of the night, larvae, with and without air, had inflated swimbladders, and their volumes were not signifi- cantly reduced through the night. Thus, deflation NOTE Forward et al.: Swimbladder deflation in Brevoortia tyrannus 645 A y/32 3o\ 50 / \ £ 30 / \^ ■o / 26" — — ^^ ■o / — — • o / 20 E l0 39 5 CO -o 70- CD B ■4 o / \ S 50- / \ -C / 30\ -*- ^^-^^53 N. 5 ^^^^^ N^ vP 30 ~~* ^\ 0* /43 N. 38 20 10- 20 22 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (h) Figure 5 The percentage of larval Atlantic menhaden, Brevoortia tyrannus, with deflated swimbladders over the solar day. Larvae were maintained in continuous darkness and then exposed to light for one hour. The experiment was repli- cated twice (A and B). The dark bar indicates the time of the normal dark phase of the rearing light-dark cycle. The number under each point is the sample size. through gas diffusion as suggested by Hoss and Phonlor ( 1984) did not occur. If larvae were exposed to light near the time of the beginning of the light phase, deflation began within 5 minutes and was statistically apparent in 15 min- utes. Rapid deflation occurred as the pneumatic duct opened between the gut and swimbladder and gas passed into the alimentary canal, where it was moved to both the mouth or anus for expulsion. The lowest light intensity that evoked deflation was about 1012 photonscm_2-s_1. This threshold is below the lowest light intensity that inhibits inflation at sunset ( 12- 16 mm larvae; 6xl013 photonscm~2s_1; Forward et al., 1993). Thus, larvae appear to be more sensitive to light at sunrise than at sunset. Maximum night- time light intensity from the moon and star light is about 1011 photonscm^s1 (McFarland and Munz, 1975; Lythgoe, 1979). Because this value is below the threshold intensity ( 1012 photonscm^s1 ) for defla- tion, moon and star light probably will not initiate deflation at night. Since surface light levels are about 101' photonscm~2-s_1 at noon (Lythgoe, 1979), an intensity of 1012 photonscm~2s_1 occurs earlier, prob- ably near sunrise. Larvae appear to have an endogenous rhythm in light-cued deflation. If they were maintained in constant darkness, light induced a low per- cent deflation during the night phase and a high percentage during the day phase. This rhythm is the reverse of the inflation rhythm, in which sudden darkness initiates inflation at night but rarely during the day (Forward et al., 1993). The functional significance of the deflation rhythm may be that 1) larvae do not deflate their swimbladder at night in response to any light and 2) they are "prepared" for rapid deflation at sunrise. Field studies suggest Atlantic menhaden lar- vae undergo nocturnal diel vertical migration ( DVM ), in which they remain at moderate depths during the day and occur near the surface at night (Govoni and Pietrafesa, in press). Swim- bladder inflation at sunset would increase buoy- ancy and reduce larval sinking rate (Hoss et al., 1989), which would maintain larvae closer to the surface. The present laboratory study supports a nocturnal DVM pattern by indicating that the percentage of larvae with inflated swimbladder and swimbladder volumes increased through the night, when larvae have access to the air-water interface. These increases are not predicted if larvae inflate their swimbladder only once at sunset and then sink. Thus, there is probably a cycle during the night, in which larvae sink while remaining motionless and then periodically return to the surface for additional gas. This pattern would retain larvae near the surface, which may be useful for transport from the offshore spawning area to the mouth of an estuary (Hoss et al., 1989). A final consideration in the present study is why larvae deflate their swimbladders. Clearly, Atlantic menhaden larvae are adapted for deflation at sun- rise. Their rhythm indicates they are most respon- sive to a light intensity increase at this time, and deflation occurs within 15 minutes. Such a dramatic response suggests deflation has an important func- tional advantage. A fully inflated swimbladder may reduce the speed of movement and, thereby, the effectiveness of prey capture. Larvae feed during the day and use vision to find their prey (Blaxter and Hunter, 1982). Al- though a reduction in capture efficiency is possible, larvae with fully inflated swimbladders can still cap- ture prey as observed in the rhythm experiment. Al- ternatively, an inflated swimbladder may increase detection of menhaden larvae by visual predators. Since larvae are relatively transparent, the differ- ence in refractive index between air and water in- creases the contrast between an inflated swim- 646 Fishery Bulletin 92(3), 1994 bladder and the surrounding water. This increase in visibility could lead to increased predation. Thus, deflation at sunrise may be a predator avoidance response. Acknowledgments This research was funded by the National Oceanic and Atmospheric Administration Coastal Ocean Pro- gram Office through grant CF92-07 to the North Carolina Sea Grant Program. Literature cited Blaxter, J. H. S., and J. R. Hunter. 1982. The biology of clupeoid fishes. Adv. Mar. Biol. 20:1-223. Forward, R. B., Jr., L. M. McKelvey, W. F. Hettler, and D. E. Hoss. 1993. Swimbladder inflation of the Atlantic menha- den Brevoortia tyrannus. Fish. Bull. 91:254—259. Govoni, J. J., and L. J. Pietrafesa. In press. Eulerian views of layered water currents, vertical distribution of some larval fishes and in- ferred advective transport over the continental shelf off North Carolina, USA, in winter. Fish. Oceanography. Hettler, W. F. 1983. Transporting adult and larval gulf menhaden and techniques for spawning in the laboratory. Prog. Fish. Cult. 45:45-48. Hoss, D. E., and G. Phonlor. 1984. Field and laboratory observations on diurnal swim bladder inflation-deflation in larvae of gulf menhaden, Brevoortia patronus. Fish. Bull. 82:513-517. Hoss, D. E., D. M. Checklery Jr., and L. R. Settle. 1989. Diurnal buoyancy changes in larval Atlantic menhaden (Brevoortia tyrannus). Rap. P.-v. Reun. Cons. int. Explor. Mer. 191:105-111. Hunter, J. R., and C. Sanchez. 1976. Diel changes in swim bladder inflation of the larvae of the northern anchovy, Engraulis rnordax. Fish. Bull. 74:847-855. Lythgoe, J. N. 1979. The ecology of vision. Clarendon Press, Ox- ford, 244 p. McFarland, W. N., and F. W. Munz. 1975. The evolution of photopic visual pigments in fish. Vision Res. 15:1071-1080. Munz, F. W. 1958. The photosensitive retinal pigments of fish from relatively turbid coastal waters. J. Gen. Physiol. 42:445-459. Tracy, H. C. 1920. The membranous labyrinth and its relation to the precoelomic diverticulum of the swimbladder in clupeoids. J. Comp. Neurol. 31:219-257. Uotani, L. 1973. Diurnal changes of gas bladder and behavior of postlarval anchovy and other related species. Bull. Jpn. Soc. Sci. Fish. 39:867-876. Walpole, R. E. 1974. Introduction to statistics. Macmillan, New York. An energy budget for northern sand lance, Ammodytes dubius, on Georges Bank, 1977-1986 Sharon L. Gilman University of Rhode Island, Narragansett Bay Campus Narragansett, Rhode Island 02882 The northern sand lance, Ammo- dytes dubius, is a small planktiv- orous fish, classified as a "ubiqui- tous shelf species" (Sherman et al., 1983 ) and is found off the northwest Atlantic coast from North Carolina to Greenland (Nizinski et al., 1990). Sand lance are consumed by many piscivorous marine vertebrates. They have been found in the stom- achs of dogfish, Squalus spp., skates. Raja spp., Atlantic cod, Gadus morhua, haddock, Melano- grammus aeglefinus, pollock, Polla- chius virens, sculpin, Myoxocephalus spp., Atlantic salmon, Salmo salar, various flatfishes, Paralichthys, Limanda, and Pseudopleuronectes, and other fishes (Scott, 1968; Reay, 1970; Meyer et al., 1979; Bowman and Michaels, 1981; Winters, 1981), as well as seabirds (Backus and Bourne, 1987). Humpback whales, Megaptera novaeangliae, have also been observed feeding on sand lance (Payne et al., 1986). Negative correlations have been shown be- tween the abundance of sand lance and right whales, Eubalaena gla- cialis, and it has been suggested that in the northwest Atlantic these two animals may actually compete for their primary food source, the copepod Calanus finmarchicus (Kenney et al., 1986; Payne et al., 1990). Therefore, although the sand lance is not commercially impor- tant, as a plankton feeder and an important prey species, it may ex- ert significant influence over the efficiency of energy transfer from primary to higher trophic levels. Georges Bank was chosen as a study area in which dramatic changes in the northern sand lance population might be examined in terms of the consumption and pro- duction of fish relative to the pro- duction of the region as a whole. This 41,809 km2, 50-m deep plateau (Sherman et al., 1984) is located off the northeast coast of the United States and is a highly productive fishing ground with high annual primary production (350 g carbonm^-y^1 ) owing to the reten- tion of nutrients (Sherman et al., 1984; Backus and Bourne, 1987). Because of its commercial signifi- cance, Georges Bank has been well studied. Energy budgets have been developed for the entire Bank (Cohen et al., 1982; Jones, 1984; Sissenwine et al., 1984) and offer a convenient way to examine the sig- nificance of the consumption and production of an individual species within an important area of the Northeast Shelf ecosystem. Individual energy budgets offish have been developed for many spe- cies (Edwards et al., 1972; Adams, 1976; Kitchell et al., 1977; Kitchell and Breck, 1980; Cho et al., 1982; Kerr, 1982; Diana, 1983; Durbin and Durbin, 1983; Rice and Cochran, 1984; Kerr and Dickie, 1985; Cui and Wooton, 1989). In this study, the energy budget of the northern sand lance was developed from experiments that measured the following parameters: growth, metabolism, feeding and assimila- tion efficiency ( Larimer, 1992), and reproductive production. These pa- rameters were assembled into an annual energy budget based on the daily activity of the fish in the field associated with temperature and food availability. Monthly growth was used to estimate annual ration and the budget was extrapolated to northern sand lance population abundance levels measured on Georges Bank from 1977 to 1986. ' The potential predatory impact of the northern sand lance population on seasonal and annual zooplank- ton productivity on the bank2 (Sher- man et al., 1987) was examined. Finally, the annual production and consumption by these populations were compared with energy budget model values for Georges Bank. Methods Individual energy budget An "average" adult northern sand lance was considered to be age 1+, the dominant age in a population of adults (Nelson, 1990). The average size was 142 mm fork length, 6.02 g wet weight, and 1.40 g dry weight based on the following wet-weight fork length relationship of Larimer (1992): weight = 4.0665 e"1 length361. The annual energy budget for an individual northern sand lance was described by the following equation adapted from Winberg ( 1956): 1 Kane, J. 1992. Macrozooplankton seasonal abundances on Georges Bank. 1977-1986. NOAA, Nat. Mar. Fish. Ser., Northeast Fish. Sci. Center, Narragansett Lab., Narragansett, RI 02882. Unpubl. data. 2 Fogarty, M. 1992. Survey biomass esti- mates for sand lance on Georges Bank. NOAA, Nat. Mar. Fish. Ser., Northeast Fish Sci. Center, Water Street, Wood's Hole, MA 02543. Unpubl. data. Manuscript accepted 16 February 1994 Fishery Bulletin 92:647-654 ( 1994). 647 648 Fishery Bulletin 92(3), 1994 G + R = C-M-W, (1) where the components (in kilocalories) are G = somatic growth; R = reproduction; C = food consumption; M = metabolism; W = fecal loss. Assimilated ration (A) equals consumption minus fecal losses. A small fraction of the assimilated en- ergy is lost through nitrogenous excretion (Brett and Groves, 1979) but was not estimated in this study. The remaining portion of the assimilated energy was assumed available for growth and metabolism. The methods used to estimate each parameter are de- scribed below. Growth The northern sand lance growth rate is 0.87 yr-1 in grams, as determined by back calculation of length at age from otolith increments (Larimer, 1992). Mul- tiplied by 6.02 g, the wet weight of an average fish, this is equivalent to a growth rate of 5.24 gyr-1. Wet weight was converted to dry weight using the fol- lowing equation (Larimer, 1992): dry wt. = 0.309 wet wt. - 0.286 (r2=0.859). The growth rate (in diy grams) was 1.20 g-fish_1-yr_1. Growth in kilocalories was calculated based on a mean caloric content of 6.73 kcaldry gram-1 (Larimer, 1992) and was 8.08 kcal-fisrr'yr-1. Growth was also estimated on a monthly basis. Reay (1972) measured monthly growth in length of age 1+ A. tobianus off the coast of England. These fish and A. dubius are of similar size (range: 84-138 mm, Reay, 1972, versus 85-138 mm, Larimer, 1992) and seasonal temperatures in their habitats are simi- lar (3-19C off England, Reay, 1972; 3.4-14.4°C for Georges Bank, Hopkins and Garfield, 1981). There- fore, I assumed that their monthly growth rates would be similar. Reay ( 1972) found that A. tobianus grows from April to October; however, it spawns from February to March, later than the December to Feb- ruary spawning of A. dubius (Bigelow and Schroeder, 1953; Norcross et al., 1961; Reay, 1970; Colton et al., 1979; Sherman et al., 1984). Nelson and Ross ( 1991) found that gonadal development of A. dubius on Georges Bank was in progress by September. Thus, I assumed that A. dubius weight gain beginning in September is devoted to gonadal rather than to so- matic growth, and therefore, their somatic growing season extends from April to August. The percentage of annual growth occurring dur- ing each month of the growing season was calculated from two years of monthly growth data for A. tobianus as reported by Reay (1972). This monthly average was multiplied by the total annual growth measured for A. dubius (8.08 kcal-yr-1, see above) to determine monthly net growth in caloric content. Reproductive energetics Gonad weight and caloric content were measured to estimate the portion of the northern sand lance an- nual energy budget devoted to reproduction. In De- cember 1990, eight fish judged to be ripe (stage III, of Macer, 1966) were measured (fork length, mm), and wet weighed (g). The gonads were extracted, weighed and dried. The dried gonads were weighed, ground to a powder with mortar and pestle, and their caloric content measured with a Phillipson microbomb calorimeter (Phillipson, 1964). Metabolism Metabolism was estimated for an "average" day in each month by using mean monthly water tempera- ture calculated from averages of the top 40 m on Georges Bank (Hopkins and Garfield, 1981) and the number of hours of daylight at mid-month. I assumed the fish actively feed during half of the daylight hours. Thus, I divided a day into three periods: a nighttime resting period equivalent to the hours of darkness, a feeding period that is assumed to be half of the daylight hours, and a postfeeding period that is the remaining half of the daylight hours. Meta- bolic rates were estimated for each of these periods from the rates measured during similar periods at 6, 12, and 18°C (Larimer, 1992). The lowest tempera- ture that could be maintained in the lab was 6°C so these temperatures were chosen as the best approxi- mation of the annual temperature range on Georges Bank (Backus and Bourne, 1987). Because there was no clear relationship between metabolic rates and temperature evident in the respiration experiments (Larimer, 1992), the 6°C values were used for April, May, and December, the 12 C values were used for June, July, October, and November, and the 18 C values were used for August and September. Assimilation efficiency The efficiency of energy assimilation by sand lance was determined by the monthly temperature on Georges Bank (Backus and Bourne, 1987) and the relationship of assimilation efficiency to temperature found in Larimer ( 1992): AE = 82.41 + 0.764 T, (2) NOTE Gilman: An energy budget for Ammodytes dubius 649 where AE = assimilation efficiency (%); T = temperature ( C ). Ration estimation Annual ration was estimated by summing the meta- bolic requirements, somatic growth requirements, and reproductive requirements, and then by taking assimilation efficiency into account. Because assimi- lation efficiency was found to increase with increas- ing temperature (Larimer, 1992), ration was calcu- lated on a monthly rather than an annual basis. Seasonal water temperatures (and therefore fish activity levels) and food availability on Georges Bank were used to estimate a monthly ration for sand lance based on the energy budget requirements. I assumed that the fish are inactive during January, February, and March. Other species of Ammodytes (A. tobianus, Reay, 1970; A. marinus, Macer, 1966) are known to spend the winter months buried in the sand. This behavior has not been recorded for A. dubius but catches of these fish during the winter months are low (Nelson, 1990) and they have been observed to spend extended periods buried in the sand in the labo- ratory, apparently without feeding3 (personal observ., 1991 ). I assumed that the metabolic requirement for January, February, and March and the annual re- productive requirement were assimilated from May through September when food availability is high and water temperatures are still warm. Monthly gross energy requirements were esti- mated by summing monthly energetic costs and multiplying by the percent of consumed calories lost as waste based on monthly assimilation efficiencies. These were divided by the caloric content of the ra- tion (6.11 ± 0.77 kcalg"1 for Calanus ftnmarchicus; Larimer, 1992) to determine the actual grams of ra- tion required per month. The sum of these monthly estimates is the yearly ration requirement. Population energy budget The energy budget for 'individual adult northern sand lance was extrapolated to the population level by multiplying overall production (growth + repro- duction) and consumption (predicted ration) by the number of individuals estimated to be present on Georges Bank from 1977 through 1986. Northern sand lance population size was estimated from spring sand lance biomass estimates for 1977-86. 2 Mean sand lance weight per tow was divided by mean in- dividual adult fish wet weight (see above) and the average tow volume to estimate the number of indi- :i Halavik, T. NOAA, Nat. Mar. Fish. Ser., Northeast Fish. Sci Center, Narragansett Lab, Narragansett, R.I. 02882. Personal eommun., September 1991. viduals present per unit volume on the Bank. There were no sand lance abundance data for 1979. The energy budget parameters of the population were then compared with estimates of secondary produc- tion on Georges Bank. Macrozooplankton production (including Calanus ftnmarchicus, Pseudocalanus minutus, Centropages species, and Metridia lucens) on Georges Bank was calculated from population estimates measured dur- ing the MARMAP surveys from 1977 to 1986 (Sherman et al., 1987). Zooplankton volumes were reported in Kane.1 These were transformed into an- nual production values following Sherman et al. ( 1987) where volume is converted to biomass using the following equation (Wiebe et al., 1975): loglf|(dry weight) = log1(J( volume + 1.828 )/0.848. A value of 5.25 kcal-g"1 (Laurence, 1976) and a pro- duction-to-biomass ratio (P:B) of 7 (Steele, 1974; Crisp, 1975 ) were used to convert zooplankton biom- ass to production. Annual production was estimated for each year of available zooplankton data ( 1977- 86) and compared with the calculated annual con- sumption by northern sand lance. Results and discussion Predicted ration and individual budget The ratio of production to consumption (P:C) deter- mined from an individual energy budget represents the gross ecological growth efficiency of an animal within a trophic level (Slobodkin, 1960). This ratio was determined from the individual energy budget for the northern sand lance. Monthly growth esti- mates calculated from Reay's ( 1972) data range from 0% from September to March to 36% dry body weight in May (Table 1) and from 0.00 kcal, from Septem- Table 1 Monthly somatic growth of northern sand lance, Ammodytes dubius, on Georges Bank estimated from measurements of A. tobianus growth rates in length (Reay, 1972) and the availability of food on the Bank. Reay measured no net growth September through March. Month r'f Growth (wt) Growth (kcal) April May June July August 0.14 0.36 0.19 0.07 0.24 1.15 2.90 1.50 0.56 1.97 650 Fishery Bulletin 92(3), 1994 Table 2 Monthly energy requirements for northern sand lance, Ammodytes dubius. on Georges Bank. Fish are assumed to be inactive in January, February, and March; therefore metabolic energy requirements for those months ("nonfeed metabolism' ), as well as annual repi •oductive energy requirements were di- vided equally over the months of highest temperature and food avai lability. Metabolic Somatic Reproductive Nonfeed Energy cost growth growth metabolism required Month (kcal) (kcal) (kcal) (kcal) (kcal) January 2.23 0.00 0.00 0.00 0.00 February 2.02 0.00 0.00 0.00 0.00 March 2.23 0.00 0.00 0.00 0.00 April 3.90 1.15 0.00 0.00 5.05 May 4.15 2.90 0.49 1.29 8.83 June 3.33 1.50 0.49 1.30 6.62 July 3.38 0.56 0.49 1.30 5.73 August 3.57 1.97 0.49 1.30 7.33 September 3.33 0.00 0.49 1.29 5.11 October 3.07 0.00 0.00 0.00 3.07 November 2.91 0.00 0.00 0.00 2.91 December 3.44 0.00 0.00 0.00 3.44 Table 3 Gonad weight and energy content of northern sand lance, Ammodytes dubius, from Georges Bank. Males Females Total mean N Fork length (mm) Dry weight (gi Gonad i'< dry body weight) Gonad (kcal/g) 1 132 1.23 7 12717.4 1.22+0.35 14.33 26.36±2.72 6.4 7.12±0.52 8 12817.1 1.23+4.94 24.85+4.94 7.0310.54 ber to March, to 2.90 kcal in May (Table 2). The mean percent of body weight accounted for by the ripe go- nads, 24.85%, is similar to the 20-30% measured for age 1+ A. personatus (Okamoto et al., 1989) and the 25-28% measured for both male and female A. americanus (Smigielski et al., 1984), (Table 3). Esti- mated monthly metabolic requirements range from a low in February of 2.03 kcal to a high in May of 4.15 kcal (Table 4). The annual individual energy requirement was 37.56 kcal (Table 4). Monthly as- similation efficiencies and daily rations are shown in Table 5, and the predicted annual ration is 52.62 kcalfish-1 or 8.60 g-fish-1 (assuming a caloric con- tent of 6.11 kcalg-1 for C. finmarchicus; Larimer, 1992). The energy assimilated from that ingested is 48.09 kcal-fish-1 so the annual individual energy budget in terms of kilocalories is approximated by Table 4 Calculated nonthly respiration requ rements in kilo- calories for northern sar d lance, Ammodytes dubius. on Georges B ank. Fi sh are assume d to be inactive from January to March Hours of Kcal used Month •c light per month January 6.1 9.5 2.23 February 4.3 10.5 2.02 March 3.4 12.0 2.23 April 3.4 13.5 3.90 May 7.7 14.5 4.15 June 10.9 15.5 3.33 July 13.0 15.0 3.38 August 14.4 14.0 3.57 September 14.3 12.5 3.33 October 13.2 11.0 3.07 November 11.0 10.0 2.91 December 8.8 9.0 3.44 Annual total 37.56 (G) + (R) = (AR) -(/?), (3) where growth (G) = 8.08 kcal-yr-1; reproduction (R) = 2.45 kcal-yr-1; assimilated ration (AR) = 48.09 kcal-yr-1; respiration (R) = 37.56 kcal-yr-1. If the budget is converted into percentage of total consumption accounted for by each parameter, the following relationship results: NOTE Gilman: An energy budget for Ammodytes dubius 651 Table 5 Predicted daily ration of adult northern sand lance, Ammodytes dubius, on Georges Bank based on monthly growth (Table 1) and assimilation efficiency for Calanus finmarchicus (Equation 2). Month Assimilation efficiency (%) Required energy (kcal/month) Ration kcal/month) 0.00 0.00 0.00 5.81 9.86 7.23 6.17 7.81 5.45 3.30 3.18 3.81 Ration (g/month) Daily ration (%body wt) January February March April May June July August September October November December Annually 87.07 85.70 85.01 85.01 88.29 90.74 92.34 93.41 93.34 92.49 90.81 89.13 0.00 0.00 0.00 5.05 8.83 6.62 5.73 7.33 5.11 3.07 2.91 3.44 48.09 52.62 0.00 0.00 0.00 0.95 1.61 1.18 1.01 1.28 0.89 0.54 0.52 0.62 8.60 0.00 0.00 0.00 2.26 3.72 2.82 2.33 2.95 2.12 1.24 1.24 1.44 Table 6 Annual consumption of adult north em sand 1 ance, Ammodytes dubius, on Georges Bank, based on individual en- ergy bu dget requirements extrapola ted to population levels from 1977 through 1986. These values are compared to average annual zooplankton productivity for each year. The percent of production consumed by sand lance is shown for each year. Sand land abundance Ann ual consumption Zooplankton production '7c consumed Year (no. per m3) (kcal/m3/yr) (kcal/m3/yr) by sand lance 1977 0.0198 1.03 45.56 2.27 1978 0.0261 1.63 32.07 5.08 1980 0.0782 4.07 21.16 19.24 1981 0.0143 0.74 22.89 3.24 1982 0.0261 1.36 16.25 8.37 1983 0.0091 0.48 12.77 3.76 1984 0.0056 0.29 14.56 1.99 1985 0.0032 0.17 21.58 0.79 1986 0.0081 0.42 18.49 2.27 100C = 15G + 5R + TIM + 9W, (4) where C = consumption; G = growth; R = reproduction; M = metabolic requirement; W = waste (that portion of the predicted ra- tion that is not assimilated based on Equation 2 above). For an individual adult northern sand lance on Georges Bank, total production is 10.53 kcal-yr-1 (growth+reproduction), and total consumption is 52.62 kcalyr_1( Table 5); therefore, ecological effi- ciency is 20.0%. Population energy budget Northern sand lance consumed a significant propor- tion of total annual zooplankton production of Georges Bank from 1977 through 1986. Population abundance of northern sand lance from 1977 through 1986 was negatively correlated with zooplankton abundances during the same period (r2=0.683, P<0.05; Fig. 1). Northern sand lance consumed 0.79-19.24% of the annual zooplankton production from 1977 to 1986 (Table 6; Fig. 2). The trophic efficiency of the northern sand lance is 20%, according to the present energy budget model. Jones's ( 1984 1 Georges Bank energy model found that 652 Fishery Bulletin 92(3). 1994 the primary productivity of the Bank adequately ac- counted for fish production only if a trophic transfer efficiency of greater than 10% was assumed. This may be a valid assumption for the model at the trophic level of the northern sand lance. Zooplankton Sand Lance rO.08 0.06 0.04 0 02 c TO ■D I! n to c 0.00 1977 1979 1981 1983 Year 1985 1987 Figure 1 Average annual copepod abundance of Calanus finmarchicus, Pseudocalanus minutus, Centropages hamatus, C. typicus, Metridia lucens, and of adult northern sand lance, Ammodytes dubius, on Georges Bank from 1977 through 1986. 10 - o zooplankton volume - 120 _ 8 - r e - to o t= 4 - o tr 2 - j\ CD - 100 c ra XJ ~ C CO -80 | I O o -60 ^ E a o -40 o 0 " T T T 1 1 1 1 1 1 1 1 1 tu J FMAMJ J ASOND Month Figure 2 Mean monthly zooplankton biomass of Calanus finmarchicus, Pseudocalanus species, Paracalanus parvus, and Centropages species on Georges Bank (Sherman et al., 1987) and monthly ration required for an individual adult northern sand lance. Ammodytes dubius, based on the individual energy budget. In their budget of Georges Bank bioenergetics, Sissenwine et al. (1984) place sand lance in their "other finfish" compartment (all nonfished species that are vulnerable to fishing gear). From 1973 to 1975 a consumption of 9.3 kcalm~2y_1 was attrib- uted to this category. Sissenwine et al. (1984) sug- gested that the impact of the sand lance population was underestimated in their budget, and it appears from the present study, based on the individual en- ergy budget and population size, that this error was potentially significant. By converting population energetic consumption on Georges Bank (Table 6) to consumption per square meter (assuming a mean depth on the Bank of 50 m; Backus and Bourne, 1987), sand lance consumed from 8.5 to 203.5 kcal- m-2.v-i from 1977 to 1986. This represents nearly all the consumption attributed to the "other finfish" at low northern sand lance abundances and over 20 times the total "other finfish" consumption at high northern sand lance abundances. The results of this study suggest that the budget estimates of annual consumption by exploitable but commercially unde- sirable fishes may need to be revised upward. Acknowledgments Special thanks to Ken Sherman, director of the Narragansett Lab of the National Marine Fisheries Service, for his enthusiasm and assistance in intro- ducing me to his associates, Jack Green and Mike Fogarty, who provided me with the most up to date population data for zooplankton and sand lance on Georges Bank. The Ph.D. dissertation in which this study is detailed is cataloged at the Univ. Rhode Is- land under the name S. Larimer. Literature cited Adams, S. M. 1976. The ecology of eelgrass, Zostera marina I L. I, fish communities. II: Functional analysis. J. exp. Mar. Biol. Ecol. 22:293-311. Backus, R. H., and D. W. Bourne (eds.). 1987. Georges Bank. M.I.T. Press, Cambridge, MA, 593 p. Bigelow, H. B., and W. C. Schroeder. 1953. Fishes of the Gulf of Maine. Fish. Bull. Fish and Wildl. Ser. 53, 577 p. Bowman, R. E., and W. Michaels. 1981. Food habits of seventeen species of northwest Atlantic fish. NOAA Tech. Memo. NMFS-F/NEC-28. Brett, J. R., and T. D. D. Groves. 1979. Physiological energetics. In W. S. Hoar, D. J. Randall, and J. R. Brett (eds.), Fish physiology, Vol. 8. Academic Press, 786 p. NOTE Gilman: An energy budget for Ammodytes dubius 653 Cho, C. Y., S. J. Slinger, and H. S. Bayley. 1982. Bioenergetics of salmonid fishes: energy, in- take, expenditure, and productivity. Comp. Biochem. Physiol. 73B:25-41. Cohen, E. B., M. D. Grosslein, M. P. Sissenwine, F. Steimle, and W. R. Wright. 1982. An energy budget for Georges Bank. Can. Spec. Publ. Fish. Aquat. Sci. 59:95-107. Crisp, D. J. 1975. Secondary productivity in the sea. In D. F. Reichle, J. F. Franklin, and D. W. Goodal (eds.), Productivity of world ecosystems, p. 71-90. Wash- ington, D.C., National Academy of Sciences. Cui, Y., and R. J. Wooton. 1989. Bioenergetics of growth of a cyprinid Phoxinus phoxinus (L.): development and testing of a growth model. J. Fish Biol. 34:47-64. Diana, J. S. 1983. An energy budget for northern pike (Esox lucius). Can. J. Zool. 61:1968-1975. Durbin, E. G., and A. G. Durbin. 1983. Energy and nitrogen budgets for the Atlantic menhaden, Brevoortia tyrannus (Pisces: Clupeidae), a filter feeding planktivore. Fish. Bull. 81:177-199. Edwards, R. R. C, D. M. Finlayson, and J. H. Steele. 1972. An experimental study of oxygen consump- tion, growth, and metabolism of the cod (Gadus morhua L.). J. exp. Mar. Biol. Ecol. 8:299-309. Hopkins, T. S., and N. Garfield III. 1981. Physical origins of Georges Bank water. J. Mar. Res. 39:465-500. Jones, R. 1984. Some observations on energy transfer through the North Sea and Georges Bank food webs. Rapp. P.-v. Reun. Cons. int. Explor. Mer 183:204-217. Kenney, R. D., M. A. M. Hyman, R. E. Owen, G. P. Scott, and H. E. Winn. 1986. Estimation of prey densities required by west- ern North Atlantic right whales. Mar. Mammal Sci. 2:1-13. Kerr, S. R. 1982. Estimating the energy budgets of actively predatory fishes. Can. J. Fish. Aquat. Sci. 39:371- 379. Kerr, S. R., and L. M. Dickie. 1985. Bioenergetics of 0+ Atlantic herring (Clupea harengus harengus). Can. J. Fish. Aquat. Sci. 42:105-110. Kitchell, J. F., and J. E. Breck. 1980. Bioenergetics model and foraging hypothesis for sea lamprey (Petromyzon marinus). Can. J. Fish. Aquat. Sci. 37:2159-2168. Kitchell, J. F., D. J. Stewart, and D. Weininger. 1977. Applications of a bioenergetics model to yel- low perch (Perca flavescens) and walleye (Stizo- stedion vitreum vitreum ). J. Fish. Res. Board Can. 34:1922-1935. Larimer, S. 1992. Aspects of the bioenergetics and ecology of sand lance of Georges Bank. Ph.D. diss., Univ. Rhode Island, 300 p. Laurence, G. C. 1976. Caloric values of some north Atlantic calanoid copepods. Fish. Bull. 741:218-220. Macer, C. T. 1966. Sand eels (Ammodytidae) in the south-west- ern North Sea; their biology and fishery. Fish. Invest. Lond. Ser. 2, 24:1-55. Meyer, T. L., R. A. Cooper, and R. W. Langton. 1979. Relative abundance and food habits of the American sand lance, Ammodytes amencanus, from the Gulf of Maine. Fish. Bull. 77:243-253. Nelson, G. A. 1990. Population biology and dynamics of northern sand lance (Ammodytes dubius) from the Gulf of Maine to the Middle Atlantic Bight region. M.S. thesis, Univ. Mass., 202 p. Nelson, G. A., and M. R. Ross. 1991. Biology and population changes of northern sand lance (Ammodytes dubius) from the Gulf of Maine to the Middle Atlantic Bight. J. Northwest Atl. Fish. Sci. 11:11-27. Nizinski, M. S., B. B. Collette, and B. B. Washington. 1990. Separation of two species of sand lances, Ammodytes americanus and A. dubius, in the west- ern North Atlantic. Fish. Bull. 88:241-255. Norcross, J. J., W. H. Massmann, and E. B. Joseph. 1961. Investigations on Investigations on inner con- tinental shelf waters off lower Chesapeake Bay Part II: Sand lance larvae, Ammodytes ameri- canus. Chesapeake Sci. 2:49-59. Okamoto, H., H. Sato, and K. Shimazi. 1989. Comparison of reproductive cycle between two genetically distinctive groups of sand lance (genus Ammodytes) from northern Hokkaido. Nippon Suisan Gakkaishi 55:1935-1940. Payne, P. M., J. R. Nicolas, L. O'Brien, and K. D. Powers. 1986. The distribution of the humpback whale Megaptera novaeangliae on Georges Bank and in the Gulf of Maine in relation to densities of the sand eel Ammodytes americanus. Fish. Bull. 84:271-277. Payne, P. M., D. N. Wiley, S. B. Young, S. Pittman, P. J. Clapham, and J. W. Jossi. 1990. Recent fluctuations in the abundance of ba- leen whales in the southern Gulf of Maine in rela- tion to changes in selected prey. Fish. Bull. 88:687-696. Phillipson, J. 1964. A miniature bomb calorimeter for small bio- logical samples. Oikos 15:130-139. Reay, P. J. 1970. Synopsis of biological data on North Atlantic sand eels of the genus Ammodytes. F.A.O. Fishe- ries Synopsis 82, 71 p. 1972. The seasonal pattern of otolith growth and its application to back-calculation studies in 654 Fishery Bulletin 92(3), 1994 Ammodytes tobianus L. J. Cons. int. Explor. Mer 34:485-504. Rice, J. A., and P. A. Cochran. 1984. Independent evaluation of a bioenergetics model for largemouth bass. Ecology 65:732-739. Scott, J. S. 1968. Morphometries, distribution, growth, and maturity of offshore sand lance (Ammodytes dubius) on the Nova Scotia banks. J. Fish. Res. Board Can. 25:1775-1785. Sherman, K., R. Lasker, W. Richards, and A. W. Kendall Jr. 1983. Ichthyoplankton and fish recruitment stud- ies in large marine ecosystems. Mar. Fish. Rev. 45:1-25. Sherman, K., W. Smith, W. Morse, M. Berman, J. Green, 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. Serv. 18:1-19. Sherman, K., W. G. Smith, J. R. Green, E. B. Cohen, M. S. Berman, K. A. Marti, and J. R. Goulet. 1987. Zooplankton production and the fisheries of the northeastern shelf. In R. H. Backus and D. W. Bourne (eds.), Georges Bank, p. 268- 282. M.I.T. Press, Cambridge, MA. Sissenwine, M. P., E. B. Cohen, and M. D. Grosslein. 1984. Structure of the Georges Bank ecosystem. Rapp. P.-v. Reun. Cons. int. Explor. Mer 183:243-254. Slobodkin, L. B. 1960. Ecological energy relationships at the popu- lation level. Am. Nat. 94:213-236. Smigielski, A. S., T. A. Halavik, L. J. Buckley, S. M. Drew, and G. C. Laurence. 1984. Spawning, embryo development, and growth of the American sand lance Ammodytes americanus in the laboratory. Mar. Ecol. Prog. Ser. 14:287-292. Steele, J. H. 1974. The structures of marine ecosystems. Harvard Univ. Press, Cambridge, MA, 128 p. Wiebe, P. H., S. Boyd, and J. L. Cox. 1975. Relationships between zooplankton displace- ment volume, wet weight, dry weight, and carbon. Fish. Bull. 73:777-786. Winberg, G. G. 1956. Rate of metabolism and food requirements of fishes. Belorussian State Univ., Minsk. Fish. Res. Board Can. Transl. Ser. 194, p. 1-256. Winters, G. H. 1981. Growth patterns in sand lance, Ammodytes dubius, from the Grand Bank. Can. J. Fish. Aquat. Sci. 38:841-846. Movements of tagged adult yellowtail rockfish, Sebastes flavidus, off the west coast of North America Richard D. Stanley Bruce M. Leaman Pacific Biological Station Department of Fisheries and Oceans, Canada Nanaimo, British Columbia, V9R 5K6 Lewis Haldorson Juneau Center, School of Fisheries and Ocean Sciences, University of Alaska, Fairbanks 11120 Glacier Highway Juneau, Alaska 99801-8677 Victoria M. O'Connell Alaska Department of Fish and Game 304 Lake Street, Room 103, Sitka, Alaska 99835 The Department of Fisheries and Oceans, Canada, and the Univer- sity of Alaska conducted indepen- dent tagging studies on yellowtail rockfish, Sebastes flavidus, in the early 1980's. The Canadian study was designed to validate ageing methodology for rockfishes (Lea- man and Nagtegaal, 1987). The Alaskan study was part of a larger survey of nearshore bottomfish re- sources in southeastern Alaska.1'2 While neither study was designed to quantify the extent of this spe- cies' movement, the recoveries pro- vided new insight into rockfish be- havior and new implications for the management of this species. Tagging studies of shallow dem- ersal (<100 m) species of rockfish, Sebastes spp., have typically indi- cated very limited movement (Table 1), with the exception of a report of a brown rockfish, S. auriculatus, that travelled over 50 km from San Francisco Bay3 Authors have also suggested limited movement for the deeper demersal or "slope" rockfish species, such as Pacific ocean perch, S. alutus, that are found along the continental slope at depths greater than 200 m (Fadeev, 1968; Gunder- son, 1971; Wishard et al., 1980; Leaman and Kabata, 1987). While they appear to make seasonal bathymetric migrations, the avail- able evidence from commercial fish- ing patterns, parasite occurrence, and age/size compositions have led investigators to hypothesize that these species make very limited latitudinal movements along the continental shelf. However, because of decompression and other injuries associated with surfacing from depths of over 200 m, no tagging studies have been performed to test this hypothesis. It is the semi-pelagic species that inhabit the continental shelf ( 100- 200 m ) which appear to exhibit sig- nificant movement. Studies of black rockfish, S. melanops, and imma- ture yellowtail rockfish indicate that at least some individuals move long distances. In northern Puget Sound (Fig. 1), Mathews and Barker ( 1983) tagged 123 black and 153 yellowtail rockfish. Three of eight black rockfish and eight of 10 confirmed yellowtail rockfish recov- eries came from the west coast of Washington at distances up to 400 km from the release site. Because the yellowtail rockfish were all im- mature, the authors proposed an ontogenetic movement offshore in conjunction with reproductive maturation. Similarly, Barss4 re- ported that 12 of 23 recovered im- mature canary rockfish, S. pin- niger, travelled more than 100 km along the Oregon coast. Culver (1987) provided the first evidence of long distance movement of adult or reproductively mature rockfish. He recovered 484 tags from 14,795 black rockfish tagged off Washing- ton and northern Oregon. One fish, which had been tagged off Oregon, was recovered off northern Califor- nia, 555 km south of its release site. More than 12% of the recovered black rockfish moved farther than 80 km. Contrary to those reports, which documented long distance move- 1 Rosenthal, R. J„ L. J. Field, and D. Meyer. 1981. Survey of nearshore bottomfish in the outside waters of southeastern Alaska. Alaska Coastal Research, P. O. Box 368, Langley WA, 98260. Final report to State of Alaska, Dep. Fish Game, Comm. Fish. Div. Juneau, 84 p. 2 Rosenthal. R. J., L. Haldorson, L. J. Field, and V. M. O'Connell. 1982. Inshore and shallow offshore bottomfish resources in the southeastern Gulf of Alaska. Alaska Coastal Research, P. O. Box 368, Langley WA, 98260. Final report to State of Alaska, Dep. Fish Game, Comm. Fish Div. Ju- neau, 166 p. 3 Lenarz, W. Tiburon Laboratory, Nat. Mar. Fish. Serv., CA 94920. Personal commun., March 1993. 1 Barss, B. Marine Science Center, Oregon Dept. Fish, Wildl., Newport, OR 97.365. Personal commun., December 1985. Manuscript accepted 29 November 1993. Fishery Bulletin 92:655-663 (1994). 655 656 Fishery Bulletin 92(3), 1994 ment, studies of blue rockfish, S. mystinus (Miller and Geibel, 1973) and olive rockfish, S. serranoides (Love, 1980), as well as other studies of yellowtail rockfish (Carlson and Haight, 1972; Pearcy, 1992), indicate much more limited movement. In fact, dis- persal from the tagging site was so limited in the latter two studies that the authors hypothesized that Table 1 Reports of rockfish, Sebastes spp. movement. No. of Max. dist. Species recoveries moved, (km) Literature cited Demersal species Black-and-yellow rockfish S. chrysomelas 38 <1 Larson. 1980 Brown rockfish, S. auriculatus 22' <1 Matthews et al.. 1987 IK <2 Hartmann, 1987 1 >50 Lenarz, footnote 3 in text 11 <2 Gowan, 1983 China rockfish. S. nebulosus 13 ■ 1 McElderry, 1979 Copper rockfish, S. caurinus 2 2 Gascon and Miller, 1981 11' ■ 1 Mathews and Barker, 1983 29 < 1 Matthews et al., 1987 16 <3 Hartmann, 1987 75 <3 Gowan, 1983 Gopher rockfish. S. carnatus 49 - 1 Larson, 1980 Quillback rockfish. S. maliger 12 <3 Mathews and Barker, 1983 28' <1 Matthews et al., 1987 Yelloweye rockfish. S. rubernmus 7 ■ 1 Coombs, 1979 3 <1 O'Connell, 1991 Semi-pelagic species Black rockfish, S. melanops <.l <1 McElderry. 1979 to <50 Gowan, 1983 8 400 Mathews and Barker, 1983 I.St 555 Culver, 1987 9 K19 Coombs, 1979 Blue rockfish. S. mystinus 168 24 Miller and Geibel, 1973 98 43 Hartmann, 1987 Bocaccio S. paucispinis liK 148 Hartmann, 1987 Olive rockfish, S. serranoides 435 <33 Hartmann, 1987 Vermilion rockfish. S. mmiatus 1 10 Turner et al.. 19K9 Yellowtail rockfish. S. flavidus 76 <23 Carlson and Haight, 1972 10 144 Mathews and Barker, 1983 25 <4 Pearcy, 1992 ' Repeated dive observations of tagged fish 2 Tagged fish showed strong site fidelity, no numbers provided. yellowtail rockfish have strong homing tendencies and exhibit site fidelity. However, these two studies were conducted over limited spatial and temporal scales. The purpose of this note is to examine the hypoth- esis of limited versus extensive movement of yellow- tail rockfish. We use data from our two studies, which cover a broader time and space coverage than those studies that implied limited move- ment. We present the details of the 42 recaptures, discuss some of the factors which may have influenced the overall likelihood of recapture, and conclude with a comment on the management implications of the results. Methods Methodology used in the Canadian program, which was conducted from 1980 to 1982, has been de- scribed in detail (Shaw et al., 1981). The primary area of tagging was off southwest Vancouver Is- land; additional tagging was per- formed in Queen Charlotte Sound (Figs. 2 and 3). Fish were captured by trawl at depths of 70-80 m over bottom depths of 11 0-1 30 m. Prior to tagging, all fish were anaes- thetized with tricaine methan- esulphonate (MS-222). Fish with hyperinfiated swim bladders were deflated with a hypodermic needle to remove excess gas (Gotshall, 1964 ), measured to the nearest cm (fork length), and tagged with an external Floy anchor tag imbedded in the dorsal musculature between the pterygiophores. Most fish were injected with oxytetracycline (OTC) (50 mg/kg body wt.) (Lea- man and Nagtegaal, 1987). All fish were held for one hour in covered tanks with a continuous flow of seawater. Fish from the last haul of each day were held overnight. Condition after tagging and after release was assessed by using a numerical index based on several categories of injuries (Shaw et al., 1981 ). We did not record sex of re- leased specimens but did examine NOTE Stanley et al.: Movements of tagged Sebastes flavidus 657 0 100 Figure 1 Sites of major yellowtail rockfish, Sebastes flavidus, fishing grounds (hatched areas); stock designations are based on individual or combi- nations of International North Pacific Fishery Commission (INPFC) statistical areas (ie. N. Columbia) and Pacific States Marine Fishery Commission (PSMFC) statistical boundaries within the INPFC ar- eas (ie. 3C-US). two samples of 100 fish sacrificed during the 1980 tagging off the southwest coast of Vancouver Island. In the Alaskan program of 1981 and 1982, fish were caught by hook and line from depths of 40-100 m with commercial jigging machines. Fish were exam- ined for decompression stress and hooking damage. Only those fish with no visible stress symptoms were tagged. Atotal of 397 yellowtail rockfish were tagged as in the Canadian study, but none was injected with OTC or decompressed. All fish were captured and released in July 1982 at two sites in Sitka Sound, Alaska (Fig. 2). All tag recoveries from both programs were ob- tained from commercial fisheries. Data on gender and recovery location were obtained when possible, al- though the latter information was usually limited to statistical area. We calculated the minimum possible distance travelled by assuming the fish travelled a direct course approximating the edge of the conti- nental shelf. Distance was calculated to the border of the statistical area closest to the point of release and rounded down to the nearest 25-km interval. Transit time was calculated as the overall distance divided by the number of days at large. The recovery 658 Fishery Bulletin 92(3). 1994 ratios (number recovered divided by number released) for the two studies were compared with a two-tailed test of binomial proportions (Kalbfleisch, 1976). We used maturity ogives from Tagart ( 1991) to infer the proportion mature for specific lengths. He re- ported 50% maturity at 39.6 cm and 45.4 cm for males and females, re- spectively, for fish from northern Washington. We conducted a linear regression of the natural log of number recov- ered against time (0.5 years— 9.5 years) for the Canadian releases to derive a point estimate of the instan- taneous annual rate of extinction. Results Canadian-tagged fish The Canadian program tagged and released 4,895 fish off central Brit- ish Columbia (B.C.) and 9,557 off southern B.C. and northern Wash- ington waters. Thirty-seven have been recovered, all from commercial catches from Oregon, Washington, and B.C. waters (Figs. 2 and 3; Table 2 ). Of these, 36 were accompanied by reliable information on recapture lo- cation. Twenty-seven of the 36 (75% ), moved less than or equal to 25 km from the release location. However, among the nine fish that travelled more than 25 km, three moved at least 100 km, three others at least 125 km, and one at least 250 km. The farthest displacements of Canadian-tagged fish were one fish that moved from Queen Charlotte Sound to the southwest coast of Vancouver Island (400 km) and one that moved from southwest Vancouver Is- land to southern Oregon (400 km). The most rapid movements from original tagging sites were two re- coveries of Canadian-tagged fish that moved from the north coast of Washington to northern Oregon. One travelled at least 100 km in 73 days (1.37 km/ day) whereas the other travelled at least 125 km in 100 days (1.25 km/day). Average fork length of Canadian-tagged fish was 44.5 cm (23-58 cm). Recovered specimens averaged 45.9 cm (33-54 cm). Among the 21 recoveries for which sex was known, only two were probably im- mature at the time of release, based on their sex and lengths. Among the nine individuals that travelled further than 25 km, we know the sex of seven. Of Figure 2 Number of releases (in hatched areas) and recoveries (in circles) of yel- lowtail rockfish, Sebastes fiavidus, tagged in southeast Alaska and Queen Charlotte Sound. these, only one, a 33-cm male, was likely to have been immature at the time of release. The rate of recoveries from the Canadian study grad- ually declined over time (Table 3). The point estimate of the instantaneous annual rate of extinction was 0.2. However, the 95% confidence range (0.099-0.307) was wide, reflecting the low number of recoveries. Alaskan-tagged fish Of the 397 yellowtail rockfish tagged in Alaska, five have been recovered. All five travelled south to B.C. or to Washington waters over distances of 425-1400 km. The fish were tagged in 1982 but were not re- covered until at least 1987. Time at liberty ranged from 1,827 to 2,842 days. The opportunity for recov- ery of yellowtail rockfish in southeastern Alaska com- mercial fisheries was limited. The total reported com- mercial catch in 1991 was three tons; the sport catch was negligible.5 However, the recovery ratio of Alas- 5 O'Connell, V. Alaska Dept. Fish Game, Sitka, AK 99M:if, Per- sonal commun., February 1991. NOTE Stanley et al.: Movements of tagged Sebastes flavidus 659 Table 2 Recoveries of tagged yellowtail rockfish, Sebastes flavidus, by the Department of Fisheries and Oceans, Canada, and the University of Alaska; Figure 1 indicates Pacific States Marine Fish Commission Areas (PSMFC) " — " indicates that no information was obtained. PSMFC Minimum Length at Length at PSMFC area area of Days at distance release recovery Injuries at Age at of tagging recovery liberty displaced (km) (cm) (km) release recovery (yr) Sex Canadian 3C-US 3B 838 0 44 44 none 15 3C-US 3C-US >3,617 0 44 45 none — M 3C-US 3A-3D' N/A N/A 54 N/A none — — 3C-US 3A 859 125 51 53 none 15 F 3C-US 3C-US 2,577 0 54 N/A none — 3C-US 3C-US 48 0 53 53 none — — 3C-US 3C-US 434 25 46 N/A bleeding1' — — 3C-US 3C-US 55 0 52 54 none — — 3C-US 3C-US 382 0 45 N/A none — — 3C-US 3C-US 2,268 0 45 49 none — M 3C-US 3C-US 18 0 43 48 none — — 3C-US 4A 1,392 25 45 48 none — — 3C-US 3C-US 478 25 47 N/A none — — 3C-CAN 3A 832 125 45 45 loss of scales 13 F 5A+5B 3C-US 1,279 400 42 43 bleeding' 11 M 5A+5B 5A+5B 861 25 45 46 none 13 M 5A+5B 3D 3,215 250 42 43 none — M 5A+5B 5A+5B 2,119 0 44 48 bleeding'' — M 5A+5B 5A+5B 328 25 46 45 none 20 M 5A+5B 5A+5B 3,219 0 48 48 none — M 5A+5B 5A+5B 624 0 51 50 loss of scales 17 M 3C-US 3C-? 1,177 0 49 48 none — — 3C-US 2B 690 400 48 45 none 9 F 3C-US 3C-US 73 0 47 N/A none — — 3C-US 3A 798 100 33 39 none — M 3C-US 3B 1,218 0 4 1 46 none — M 3C-US 3B 186 0 42 41 none — M 3C-US 3A 73 100 46 46 none — — 3C-US 3C-US 1.472 0 47 47 none — M 3C-US 3A 1,994 100 46 N/A none — — 3C-US 3B 739 0 47 46 none 34 M 3C-US 3C-US 877 0 50 51 none 17 F 3C-US 3C-US N/A 25 4 1 39 bruised — — 3C-US 3B 1,231 0 48 N/A none — — 3C-US 3C 2,653 0 42 49 none — F 3C-US 3A 103 125 42 42 none — M 4A 3B 2,438 0 43 46 none — M Alaskan SE Alaska 5A+5B 1,827 700 37 45 — SE Alaska 3B 2,047 1,400 37 45 — — — SE Alaska 5A+5B 2,8425 750 32 N/A — — — SE Alaska 5A+5B 2,7226 625 33 45 — — M SE Alaska 5A+5B 2,7297 N/A 41 50 — — F ' Recovered from Area 3A, 3B, 3C or 3D. 2 Bleeding at tag site. 3 Bleeding at de-gassing site. ■* Bleeding at oxytetracycline injection site. 5 2,842-2,872 days at liberty 6 2,722-3,087 days at liberty. 1 2,729-3,094 days at liberty. 660 Fishery Bulletin 92(3). 1994 yftyis. Figure 3 Number of releases (in hatched areas) and recoveries (in circles) of yellowtail rockfish, Sebastes flavidus, tagged off southern Vancouver Island and northern Washington (one specimen re- leased in PSMFC area 3C-US was recovered in Areas 2B-2C to the south of the areas shown in this figure). kan specimens ( 1.3% ■) was significantly higher (P<0.05 ) than in the Canadian study (0.26% ), even without sig- nificant commercial fisheries in the area of release. The average length of all Alaskan-tagged fish was 37 cm (22-56 cm). The average length of recovered fish was 46 cm (45—50 cm). Based on the lengths, the recovered specimens were probably sexually mature when recov- ered, but probably immature when released. Discussion The recoveries of Alaskan specimens are congruent with earlier work, which indicated that immature yellowtail rockfish can make long distance move- ments (Mathews and Barker, 1983). The Canadian recoveries provide the first evidence that mature yel- lowtail rockfish can also move significantly longer distances than has been previously reported (Carlson and Haight, 1972; Pearcy, 1992). The consistent tendency for individuals who trav- elled away from the release points to be caught far- ther south along the coast probably resulted from the bias in the distribution of the fishing effort. For the Alaskan and central B.C. releases, there were virtu- ally no fisheries north or west of the release areas. It is surprising that none of the fish tagged in northern NOTE Stanley et al.: Movements of tagged Sebastes flavidus 661 Washington were recovered in Queen Charlotte Sound. However, landings of yellowtail rockfish in Queen Charlotte Sound were also low in the years during, and the first three years following, tagging (Table 4). The yellowtail rockfish tagged in the Alaskan study were caught in waters shallower than 30 m and were predominantly immature. A southward, ontogenetic migration to the broader continental shelf off cen- tral B.C. is possible. This would parallel the move- ment of immature yellowtail rockfish from Puget Sound to the outer Washington coast (Mathews and Table 3 Minimum estimate of years at liberty for yellowtail rockfish, Seba stes flavidus, tag recoveries for those with reliable information on recovery date Number of recoveries Time at liberty (yr) Canadian Alaskan <1 8 0 1-2 5 0 2-3 7 0 3-4 5 0 4-5 1 0 5-6 2 2 6-7 2 0 7-8 2 3 8-9 2 0 9+ 1 0 Total 35 5 Table 4 Landings (t) for 1980- -89 and biomass estimates (t) for 1990 of yellowtail rockfish, Sebastes flavidus , by region (Stanley, 1991 Tag art, 1991). Year of Washington- landings SE Alas ka British Columbia' Oregon-California- 1980 704 8.667 1981 426 9,181 1982 1 526 9,184 1983 2 447 9,498 1984 6 353 5,392 1985 5 941 3,449 1986 5 4.458 5,082 1987 8 4.168 5,212 1988 6 4,765 6.193 1989 4 4,298 4,516 1990 Expl oitable biomass unknown 30,000-45,000 20,000-40,000 ' Northern Vancouver stock biom jss partitioned 5G"y to Washington, 50< I B.C. 2 Does not nclude the bycatch in offshore fishing for Pacific hake. Merluccius productus. Barker, 1983 ). However, many of these fish may sim- ply move west towards the deeper water on the outer shelf but remain within southeastern Alaska. The limited commercial fishery in these outside waters prevents resolution of these alternatives, although they are not mutually exclusive. We suggest that the low recovery rate for the Ca- nadian study was caused by short-term post-tagging mortality arising from a combination of the greater depth of capture, the greater stress from trawl cap- ture compared with hook-and-line, and the added handling during anaesthetization, decompression, and OTC injection. The dosage level of OTC may have also contributed to mortality, as studies on other spe- cies have shown increasing mortality with dosages greater than 25 mg/kg body weight (McFarlane and Beamish, 1987). An alternative explanation for low recovery rates in the Canadian study is that the tagged fish were released into very large populations, well in excess of current estimates (Table 4). Thus, landings of 400- 900 t in the early 1980s provided little chance for recovery. However, this same logic should then ap- ply to the likelihood of recapturing Alaskan tags in the B.C. fishery at the same time. Since this recov- ery ratio was much higher, the Alaskan results sup- port the contention of high initial mortality in the Canadian study. The rate of recovery for Canadian tags was low at the outset and gradually declined over the ten years after release. The 95% confidence limits of the point estimate for the instantaneous rate of decline over the ten years (0.099-0.307) is consistent with the range of extinction rates (0.28- 0.54) predicted by combining pub- lished estimates of instantaneous rates for natural mortality, 0.06- 0.12 (Archibald et al., 1981; Tagart, 1991), tag loss for black rockfish, 0.13 (Lai and Culver, 1991), and fishing mortality for yellowtail rockfish off northern Washington, 0.09-0.29 (Tagart, 1991). Despite the commercial impor- tance of this species, the stock boundaries are poorly understood. Assessment biologists have re- sorted to selecting boundaries based on a combination of official statistical areas and on the distri- bution of major fishing grounds. The most recent assessments have assumed four stocks from group- ings of International North Pacific 662 Fishery Bulletin 92(3), 1994 Fisheries Commission statistical areas: Charlotte/ North Vancouver, South Vancouver, North Colum- bia, and Eureka/South Columbia (Stanley, 1991; Tagart, 1991) (Fig. 1). This study does not provide information sufficient to alter the boundaries presently used in stock as- sessments. However, the clear demonstration of movement for immature and mature yellowtail rock- fish should encourage assessment biologists to re- fine their understanding of stock relationships for this species. This understanding can then be applied to examination of the relative impacts of fishing mor- tality and potential movements among stock units. Acknowledgments The Alaskan tagging program could not have been accomplished without the expertise and assistance of L. J. Field, M. LaRiviere, and especially Richard Rosenthal. We also appreciate the review comments of Bill Pearcy and two anonymous reviewers. Literature cited Archibald, C. P., W. Shaw, and B. M. Leaman. 1981. Growth and mortality estimates of rockfishes (Scorpaenidae) from B.C. coastal waters, 1977- 1979. Can. Tech. Rep. Fish. Aquat. Sci. 1048, 57 p. Carlson, R. H., and R. E. Haight. 1972. Evidence for a home site and homing of adult yellowtail rockfish, Sebastes flavidus. J. Fish. Res. Board Can. 29:1011-1014. Coombs, C. 1979. Reef fishes near Depoe Bay, Oregon: move- ment and the recreational fishery. M.A. thesis, Oregon State Univ., Corvallis, 39 p. Culver, B. N. 1987. Results from tagging black rockfish {Sebastes melanops ) off the Washington and northern Oregon coast. In Proceedings of the international rock- fish symposium, p. 231-239. Univ. Alaska Sea Grant Rep. 87-2. Univ. Alaska, Fairbanks, 393 p. Fadeev, N. S. 1968. Migrations of Pacific ocean perch. Proc. Pac. Sci. Res. Inst. Fish. Ocean. 65:170-177. [Fish. Res. Board Can. Transl. Ser. 1447. | Gascon, D., and R. A. Miller. 1981. Colonization by nearshore fish on small arti- ficial reefs in Barkley Sound, British Columbia. Can. J. Zool. 59:1635-1646. Gotshall, D. W. 1964. Increasing tagged rockfish (genus Sebastodes ) survival by deflating the swim bladder. Calif. Dep. Fish Game, Fish. Bull. 50:253-260. Gowan, R. E. 1983. Population dynamics and exploitation rates of rockfish (Sebastes spp.) in central Puget Sound, Washington. Ph.D. thesis, Univ. Wash., Seattle, 90 p. Gunderson, D. R. 1971. Evidence that Pacific ocean perch (Sebastes alutus) in Queen Charlotte Sound form aggrega- tions that have different biological characteristics. J. Fish. Res. Board Can. 29:1061-1070. Hartman, A. R. 1987. Movement of scorpionfishes (Scorpaenidae: Sebastes and Scorpaena) in the southern Califor- nia Bight. Calif. Dep. Fish Game, Fish. Bull. 73(21:68-79. Kalbfleisch, J. G. 1976. Probability and statistical inference. Part 2. Univ. Waterloo, Waterloo, Ontario, 307 p. Lai, H., and B. N. Culver. 1991. Estimation of tag loss rate of black rockfish (Sebastes melanops) off the Washington coast with a review of double tagging models. Wash. Dept. Fish. Tech. Rep. 113, 28 p. Larson, R. J. 1980. Territorial behaviour of the black-and-yellow rockfish and gopher rockfish (Scorpaenidae, Sebastes). Mar. Biol. 58:111-122. Leaman, B. M., and Z. Kabata. 1987. Neobrachiella robusta (Wilson, 1912) (Copepoda:Lernaepodidae) as a tag for identification of stocks of its host, Sebastes alutus (Gilbert, 1890) (PiscesTeleostei). Can. J. Zool. 65:2579-2582. Leaman, B. M., and D. A. Nagtegaal. 1987. Age validation and revised natural mortality rate for yellowtail rockfish. Trans. Am. Fish. Soc. 116:171-175. Love, M. S. 1980. Isolation of olive rockfish, Sebastes serran- oides, populations off southern California. Fish. Bull. 77:975-983. Mathews, S. B., and M. W. Barker. 1983. Movements of rockfish (Sebastes) tagged in northern Puget Sound, Washington. Fish. Bull. 81:916-922. Matthews, K. R., B. S. Miller, and T. P. Quinn. 1987. Movement studies of nearshore demersal rockfishes in Puget Sound, Washington. In Pro- ceedings of the international rockfish symposium, p. 63-72. Univ. Alaska Sea Grant Rep. 87-2. Univ. Alaska, Fairbanks, 393 p. McElderry, H. I. 1979. A comparative study of the movement habits and their relationships to buoyancy compensation in two species of shallow reef rockfish (Pisces: Scorpaenidae). M.Sc. thesis, Univ. Victoria, Victoria, B.C., 168 p. McFarlane, G. A., and R. J. Beamish. 1987. Selection of dosages of oxytetracycline for age validation studies. Can. J. Fish. Aquat. Sci. 44(4):905-909. NOTE Stanley et al.: Movements of tagged Sebastes fiavidus 663 Miller, D. J., and J. J. Geibel. 1973. Summary of blue rockfish and lingcod life his- tories; a reef ecology study; and giant kelp, Macrocy- stis pyrifera, experiments in Monterey Bay, Califor- nia. Calif. Dep. Fish Game, Fish Bull. 158, 137 p. O'Connell, V. M. 1991. A preliminary examination of breakaway tag- ging for demersal rockfishes. Alaska Dep. Fish Game, Comm. Fish. Div., Fish. Res. Bull. 91-06, 8 p. Pearcy, W. G. 1992. Movements of acoustically-tagged yellowtail rockfish Sebastes fiavidus on Heceta Bank, Oregon. Fish. Bull. 90:726-735. Shaw, W., D. A. Nagtegaal, C. P. Archibald, and B. M. Leaman. 1981. Rockfish tagging cruises off southwest Vancouver Island (MV Ocean King) and off north- west Vancouver Island and in Queen Charlotte Sound (MV Blue waters ) during 1980. Can. Data Rep. Fish. Aquat. Sci. 288, 133 p. Stanley, R. D. 1991. Shelf rockfish (silvergray, yellowtail, canary rockfish). In J. Fargo, and B. M. Leaman (eds.), Groundfish stock assessments for the west coast of Canada in 1990 and recommended yield options for 1991, p. 225-276. Can. Tech. Fish. Aquat. Sci. 1778. Tagart, J. V. 1991. Population dynamics of yellowtail rockfish Sebastes fiavidus stocks in the northern Califor- nia to southwest Vancouver Island Region. Ph.D. thesis, Univ. Wash., Seattle, 323 p. Turner, C. H., E. E. Ebert, and R. R. Given. 1969. Man-made reef ecology. Calif. Dep. Fish Game, Fish Bull. 146, 221 p. Wishard, L. N., F. M. Utter, and D. R. Gunderson. 1980. Stock separation of five rockfish species us- ing naturally occurring biochemical markers. Mar. Fish Rev. 42:64-73. A simple generalized model of allometry, with examples of length and weight relationships for 1 4 species of groundfish Yongshun Xiao David C. Ramm Fisheries Division, Department of Primary Industry and Fisheries GPO Box 990. Darwin NT 0801. Australia Allometry is a set of relations be- tween an animal's characteristics and its body size, and is applied in many branches of biological sci- ences including ecology, physiology, and morphology (Peters, 1983; Calder, 1984; Schmidt-Nielsen, 1984; Bookstein et al., 1985; Reiss, 1989). Allometry is represented by the power function, \y = pj*o , where W is a characteristic of an animal (e.g. body weight), L is its body size, and A and X9 are its al- lometric parameters. To determine an allometric relationship for a par- ticular characteristic, the power function is usually, albeit at times inappropriately, double log-trans- formed into a simple linear equation. Y = X,+X,X,, (1) with Y = log(W), X, = log(A), and X3 = log(L), and is then fit to data from different individuals. Use of allometry in this way as- sumes constancy of X; and X,, in Equation 1. While both allometric parameters may be treated ap- proximately as constants in certain applications, the assumption may be violated for a wide variety of bio- logical phenomena because of ge- netic, phenotypic, and/or behav- ioral variability among individual animals. In fact, Mosimann and James ( 1979) have concluded that X.y varies spatially in the Florida red-winged blackbird, Agelaius 664 phoeniceus. Variability in X, is also implied in Reiss' ( 1989) hypothesis thatX., contains phylogenetic infor- mation and is less variable intraspecifically than inter- specifically. Peters (1983) convinc- ingly demonstrated interspecific variation in X9 and computed its mean and standard deviation for metabolic rates scaled to body sizes across many animal taxa. Variabil- ity inX; has not been examined but is certainly implied in the compre- hensive appendices of Peters' ( 1983 ) book on the ecological impli- cations of body size and in Reiss' ( 1989) monograph on the allometry of organismic growth and reproduc- tion. X; may be strongly negatively correlated with X., for length- weight relationships in fish (e.g. Caillouet, 1993). Variability inX; andX9 may have major implications in the widely used allometric equation because it represents a fundamental concept in biology (Peters, 1983). In this paper, we generalize Equation 1 by explicitly incorporating variability in and correlation between, X; and X9, and study the consequences of such variability and correlation in allometric predictions. The gener- alized model is demonstrated by using length and weight relation- ships for 14 species of groundfish of the families Centrolophidae, Haemulidae, Lethrinidae, Lutjan- idae, Nemipteridae, and Synodon- tidae from northern Australian waters. Model Suppose that a joint probability dis- tribution of Xj and X9 conditional on X,j could be formed for a group of animals, with each individual having its own pair of allometric parameters which it retains throughout its life, and that values of pairs of allometric parameters are serially independent. The value of Y for the z'th individual with al- lometric parameter pair (X; , X, ) at X3is y, = x;, + x2,x:l. For a group of animals selected ran- domly from the population, the ex- pected value of Y at X.{ is E\Y\X3] = E[X1 + X2X:1] (2) with variance V[Y\X3] =v\x, +A',A',| (3) = E[Y2\X:l]-E\Y\X3f = E[iXl + X,Xlr]-ElX, +A-.,.Y(|'J Given information on howX; and X., vary, one can develop Equations 2 and 3. X, may closely follow a normal distribution for metabolic rate of animals scaled to body size (Peters, 1983), being strongly nega- tively correlated with X, for length- weight relationships in fish (e.g. Cail- louet, 1993). We will assume below that Xj and X., follow a joint nor- mal distribution, i.e. (X;,X9) ~ N ^j.i14i2\a11,a2\p) with mean p(, and variance of of X., and correlation coefficient p. Under general condi- tions, the sum (or average) of a number of random variables is ap- proximately normally distributed, and such approximation can be quite good even if that number is relatively small. The above assump- Manuscript accepted 25 February 1994. Fishery Bulletin 92: 664-670 ( 1994). NOTE Xiao and Ramm: A simple model of allometry for groundfish 665 tion would be at least approximately valid because both Xt and X.2 can be regarded as the sum (or average) of numerous (e.g. genetic, phenotypic, and behavioral ) random components. Analogous models may be developed for other probability distributions. Under that assumption. Equations 2 and 3 become, respectively, E[Y\X3] = E[X,+X2X3] and \\ -r (Xj + x2X3)e 2" "';-^' 't.-p:. l' dxjdx2 (4) V[Y\X3] = V[X,+X2X3] E[Y2\X3]-E[Y\X3]2 E [( X, + X2X3 )2]-E[X,+ X2X3 f a2 + 2o1o2pX3 + o22X2, (5) Thus variability in, and correlation between, Xt and X2 only affect V[Y\X3].V[Y\ X3] increases linearly with p from (a, - o.,X3 f at p = -1 through o2 + o2X2 at p = 0 to (o, +o2X3)2 at p = 1. It quadratically decreases with o,, o9, and Xo to a minimum T2X3 of a2X^(l-pz)>0 at ct -o2pX3, crjd-p") at a., = -a,p/A3 ana crju-p i at X3=-alpla2, re- spectively, and finally increases unboundedly, under the constraint that ol,o.I, andX3 > 0. However, ifX1 and X0 are both deterministic (ot =0,p = 0), V[Y\X3] = 0. If Xj is random (c? >0) and X0 is deterministic ( 0) and Xj is deterministic (07 = 0,p = 0), V17 I X3] = otX2 . Finally, ifX; andX, are random but independent (a] > 0,ct| > 0 and p = 0)" V[Y \X3\= a2 +a2X2. Data and parameter estimation Data on fish weight at length were collected from Australia's continental shelf in the Timor and Arafura Seas (9-14°S, 127-137°E) from 20 October to 16 De- cember 1990 as part of the Northern Territory De- partment of Primary Industry and Fisheries' pro- gram assessing commercial fish stocks. Of 240 sta- tions allocated randomly within a depth range of 20- 200 m, 199 were successfully sampled with a Frank and Bryce trawl net (headline height, 2.9 m; wing spread, 14.4 m; door spread, 60.1 m) at a speed of 1.54-2.06 rrvs-1. Nearly 48 tonnes of fish' represent- ing about 483 species in 119 families were caught during sampling. A representative subsample of in- dividuals of 14 species, mostly of commercial fish, of the families Centrolophidae, Haemulidae. Lethrin- idae, Lutjanidae, Nemipteridae, and Synodontidae were frozen immediately on board, returned to the laboratory, thawed, sexed, measured (fork length) to the nearest 1 mm, and weighed (wet weight) to the nearest 1 g with an electronic balance (Mettler, PC4000). For each of the 14 species, data on indi- vidual wet weight at length were pooled across all stations and fit to all cases of Equations 4 and 5 for females, males, and mixed sexes. Parameter esti- mates indicated by hats (A) were obtained by linear regression for Equation 1 by using SAS regression procedure (SAS Institute Inc., 1985) and by maxi- mizing the general likelihood function, L = u\2nV[Y\X,] + a2} , = lL J; for all other models by using the simplex algorithm of SYSTAT nonlinear regression procedure (Wilkin- son, 1989). We included a model error term, a2, in the likelihood function to show that, in this case, it is compounded with a2 and is hence equivalent to a2 and a2 + a2 for estimation purposes. For this rea- son, we treated both error components collectively as ' a2 ' during model fitting and result presentation, unless otherwise stated. Results Some statistics offish length and weight data used in this analysis are given in Table 1. We attempted to fit data for mixed sexes (both sexable and unsexable individuals included) and males and fe- males (with unsexable juveniles excluded) of each of 14 species of groundfish to all cases of Equations 4 and 5. However, parameters could be estimated for models with o2 or o2 only; those in models simulta- 666 Fishery Bulletin 92(3), 1994 neously with o? and ol, or simultaneously with a\,c\ and p could not be estimated because of over- parameterization. Estimates of parameters, derived from linear regression of Equation 1 by using least squares method — equivalent to maximizing the like- lihood function n r ,,-i [Y,-E[Y\X3),f L= I~I 2/roj ~e i^f and from maximizing the likelihood function L = n|2/ro-oX 1 = 1 [Y,-EIY\X3] are given in Tables 2 and 3 respectively. Estimates in both tables are very similar between sexes for each species and between species, roughly with a species- wide /j, =-10.89, fi2 =2.99, ot =-0.006638 and 62 =0.014932. Thus, while V[Y\X:i\ can be treated approximately as a constant, as is usually assumed in previous applications, it does change quadratically with XQ. Discussion Peters ( 1983) observed a large amount of variability in most allometric relationships and recognized a need to identify independent variables of general bio- logical interest other than size. The general model presented in this study takes into account both body size and parameter variability among individual ani- mals in allometric predictions. A major problem in allometry is that allometricians are more apt at pro- viding a statistical description of a new data set than at using their data for hypothesis testing (Peters, 1983). This tendency has led to a plethora of only slightly different allometric equations, none of which can be rejected objectively. Our general model or any of its special cases would form a basis for intrataxal or intertaxal generalizations by treating some of those estimates of allometric parameters as intrataxal or intertaxal variations, hence providing a means for a general "house cleaning" in allometry. Incorporating more independent variables in allo- metric modelling may explain more variability in the dependent variable, but it may result in a loss of a basis for comparison between, and manipulation of, allometric equations, such as allometric cancellation (Calder, 1984). The model presented above conforms exactly with conventional allometry and maintains commensuration by its estimated parameter means. Specification of error structures in allometric mod- els is an essential part of allometric modelling. Er- rors for Equation 1 are often assumed to be normally distributed with a constant variance, say a2. Sev- eral other interpretations arise from V[Y"IX3] in that, for estimation purposes, cr2can be interpreted by any combinations of terms on the right-hand side of Equa- tion 5. These and other alternative interpretations may pose problems for some applications. Thus, er- ror structures of an allometric model must be speci- fied cautiously. There was no gain in precision or accuracy in esti- mates of allometric parameters in length and weight relationships of some fishes from considering indi- vidual variability of allometric parameters. Both Equation 1 and Equations 4 and 5 with of or of, alone give an equally adequate description of weight at length data from all 14 species of groundfish con- cerned. Overparameterization occurred in cases of Equations 4 and 5 simultaneously with a1 and o|, or simultaneously with of, oJ , and p, and, as a re- sult, not all parameters could be estimated from our data. The overparameterization lent further support to this conclusion. Also, although a1 and a2 can be estimated separately for each species, they are ei- ther equivalent to model error or take such small values (Tables 2 and 3) that V[yiX2] can be treated effectively as constant. Finally, when interpreting regression results from various cases of the general model, it should be noted that all other variability will be confounded with, and added to, that of allom- etric parameters. Our data sets are of moderate sizes (Table 1) and many others of similar size could be expected to behave similarly. Individual variability of allometric parameters probably has a negligible effect on allometric predictions in length and weight relationships of certain fishes. Thus, our work sup- ports the common use of Equation 1 to model in- traspecific length and weight relationships in those fishes. However, all parameters in Equations 4 and 5 may be estimable simultaneously for length and weight relationships, as well as for other allometric relationships, if larger data sets or higher taxonomic levels, or both, are used. A key assumption in our model is that the inde- pendent characteristic, L, (e.g. length) has little mea- surement error relative to the dependent character- istic, W (e.g. weight). Theoretically, this may not be the case. However, we believe that our model will provide good approximations for many allometrically scaled phenomena, such as length and weight rela- tionships in certain fishes. For other allometric phe- nomena, alternative formulations, such as those of Pienaar and Ricker (1968), Saenger (1989), Seim and Saether ( 1983), and Shoesmith ( 1990) may be useful. V | Y I X3 ] is a function of the independent variable whenever there is individual variability in X9 or in X, andX,,. If this is not taken into account in regres- NOTE Xiao and Ramm: A simple model of allometry for groundfish 667 Table 1 Some statistics of length and weight data for mixed sexes (both sexable and unsexable individuals included). males and females (with unsexable juveniles exc uded) of each of 14 species of groundfish caught in northern Australian waters during 20 October to 16 December 1990. Sex Species n Fork length (mm) Body weight (g) Mean SD Min Max Mean SD Min Max Mixed Diagramma pictum painted sweetlip 413 374.753 135.174 127 610 1,044.94 906.31 27 3,415 Lethnnus fraenatus blue-lined emperor 48 344.562 63.134 201 450 907.77 469.80 165 1,837 Lethnnus lentjan red-spot emperor 334 278.521 43.792 190 430 457.04 234.31 143 1,567 Lutjanus erythropterus scarlet snapper 172 431.105 54.429 255 536 1,269.63 417.65 255 2,373 Lutjanus malabancus saddle-tailed snapper 590 377.398 151.595 86 765 1,170.71 1074.90 13 7,251 Lutjanus sebae red emperor 182 342.346 125.237 94 596 1,144.50 974.22 18 4,736 Lutjanus timorensis Timor snapper 43 415.256 38.608 211 453 1.339.72 271.27 178 1.663 Lutjanus iittus one-band snapper 450 188.364 30.864 98 300 114.65 59.41 15 461 Nemipterus furcosus rosy threadfin-bream 479 164.382 34.187 38 250 95.61 55.97 3 300 Nemipterus hexodon ornate threadfin-bream 479 149.714 28.517 93 230 73.35 44.00 15 252 Pnstipomoides multidens gold-band snapper 293 314.055 117.079 131 585 818.53 882.70 50 3,800 Pristipomotdes typus sharp-tooth snapper 131 207.130 106.140 87 550 302.01 540.41 12 2,705 Psenopsis humerosa black-spot butterfish 254 158.106 14.633 105 195 106.74 32.23 25 202 Saunda mwropectoralis short-finned lizardfish 444 261.218 34.039 110 410 194.26 90.32 12 850 Female Diagramma pictum painted sweetlip 185 405.827 118.834 185 610 1.192.71 847.30 88 3,377 Lethnnus fraenatus blue-lined emperor 32 318.031 48.035 201 445 690.22 313.06 165 1,757 Lethnnus lentjan red-spot emperor 255 265.435 35.665 194 422 389.43 185.90 146 1,567 Lutjanus erythropterus scarlet snapper 78 430.731 43.480 345 536 1,285.32 402.82 627 2,373 Lutjanus malabancus saddle-tailed snapper 193 472.637 90.217 175 716 1.702.28 811.06 89 5,196 Lutjanus sebae red emperor 88 386.159 86.001 197 535 1,357.81 791.64 155 3,176 Lutjanus timorensis Timor snapper 25 414.520 22.417 378 451 1,320.64 207.65 978 1,663 Lutjanus cittus one-band snapper 212 181.835 24.025 120 262 100.01 41.03 29 289 Nemipterus furcosus rosy threadfin-bream 240 161.429 25.781 38 230 85.36 40.47 7 239 Nemipterus hexodon ornate threadfin-bream 270 146.463 23.825 97 208 67.34 33.01 18 176 Pnstipomoides multidens gold-band snapper 98 356.735 117.750 180 585 1.103.23 1,001.25 108 3,800 Pnstipomoides typus sharp-tooth snapper 29 287.034 111.650 135 550 593.48 720.46 42 2.705 Psenopsis humerosa black-spot butterfish 101 167.050 12.046 138 195 126.50 30.13 61 202 Saunda micropectoralis short-finned lizardfish 164 284.860 36.753 197 410 256.20 111.99 71 850 Male Diagramma pictum painted sweetlip 119 448.303 111.902 177 594 1.528.94 917.86 77 3415 Lethnnus fraenatus blue-lined emperor 16 397.625 56.707 216 450 1,342.88 431.39 191 1837 Lethnnus lentjan red-spot emperor 74 325.743 35.281 220 430 698.30 229 21 202 1469 Lutjanus erythropterus scarlet snapper 93 433.312 59.850 258 535 1.267.23 421.10 255 2233 Lutjanus malabancus saddle-tailed snapper 200 149 215 122.859 183 765 1.622.40 1,121.62 105 7251 Lutjanus sebae red emperor 45 423.822 94.510 187 596 1,772.42 1,048.84 124 4736 Lutjanus timorensis Timor snapper IT 128.353 19.193 388 453 1,436.12 183.58 1.021 1613 Lutjanus vittus one-band snapper 22:1 197.858 31.901 128 300 132.84 67.83 32 461 Nemipterus furcosus rosy threadfin-bream 205 178.800 30.351 115 250 120.09 61.04 28 300 Nemipterus hexodon ornate threadfin-bream 125 165.832 32.361 HIT 230 99.21 58.00 20 252 Pnstipomoides multidens gold-band snapper 127 333.276 108.795 141 580 897.81 840.95 60 3,475 Pristipomoides typus sharp-tooth snapper 35 267.314 99.371 114 530 477.31 581.00 27 2,617 Psenopsis humerosa black-spot butterfish 117 153.821 13.416 105 191 97.19 26.08 25 198 Saunda micropectoralis short-finned lizardfish 263 249.433 19.805 186 295 161.19 41.16 63 289 sion analysis, too much weight would be given to observations of the dependent variable in the region with high variances, and the analysis will be overly sensitive to chance events or bias affecting observa- tions in this region of the independent variable. Length and weight relationships in fishes are of- ten required for stock assessment and for intra- and inter-specific comparisons. Although many data are available on weight at length relationships of fishes from New Guinea (Showers, 1993) and New Cale- 668 Fishery Bulletin 92(3). 1994 Table 2 Estimates of mean and standard error of allometric parameters obtained for mixed sexes, males, and females of each of 14 species of grour dfish, caught in northern Australian waters during 20 October to 16 December 1990 by linear regression of Equat ion 1 by using least squares method. D<0.0001 applies to all species for separate sexes. Species' Mixed X, (SE) X, (SE) n-2 F,.n-2 P R2 Diagram ma pictum -11.4249(0.0650) 3.0427 (0.0111) 411 75,363.608 0.0000 0.9946 Lethrinus fraenatus -11.1084(0.2933) 3.0501 (0.0503) 46 3,673.450 0.0001 0.9874 Lethnnus lentjan -10.8678(0.1287) 3.0049(0.0229) 332 17,226.485 0.0001 0.9810 Lutjanus erythropterus -10.2265(0.2323) 2.8569(0.0383) 170 5,550.516 0.0001 0.9701 Lutjanus malabaricus -10.4713(0.0478) 2.8926(0.0082) 588 125,849.921 0.0000 0.9953 Lutjanus sebae -10.7588(0.0752) 2.9931 (0.0130) 180 52,732.028 0.0001 0.9966 Lutjanus timorensis -10.2548 (0.5172) 2.8916(0.0858) 41 1,134.654 0.0001 0.9643 Lutjanus vittus -10.5972(0.09851 2.9136(0.0188) 448 23,905.566 0.0000 0.9816 Nemipterus furcosus -10.6433(0.1163) 2.9552(0.0229) 477 16,672.088 0.0000 0.9721 Nemipterus hexodon -10.8475(0,1277) 3.0010(0.0256) 477 13,778.375 0.0000 0.9665 Pristipomoides multidens -10.4284(0.0629) 2.9192(0.0110) 291 69,881.156 0.0000 0.9958 Pristipomoides typus -10.6474(0.0672) 2.9462(0.0128) 129 52,895.132 0.0001 0.9975 Psenopsis humerosa -11.8119(0.2644) 3.2487(0.0523) 252 3,863.670 0.0001 0.9385 Saurida micropectoralis -12.3581 (0.1948) 3.1560(0.0351) 442 8,106.551 0.0001 0.9482 Species' Female A', (SE) X., (SE) n-2 F 1 1. n-2 R2 Diagramma pictum -11.4854(0.1323) 3.0526(0.0222) 183 18,940.693 0.9904 Lethrinus fraenatus -10.9359(0.4586) 3.0204(0.0797) 30 1,435.635 0.9788 Lethrinus lentjan -11.0141 (0.1823) 3.0314(0.0327) 253 8,591.353 0.9713 Lutjanus erythropterus -11.1443 (0.3965) 3.0123(0.0654) 76 2,120.223 0.9649 Lutjanus malabaricus -10.6937(0.1855) 2.9290(0.0302) 191 9,397.044 (I 9800 Lutjanus sebae -10.9484(0.2014) 3.0256(0.0339) 86 7,943.456 0.9892 Lutjanus timorensis -8.5750(1.6316) 2.6136(0.2708) 23 93.182 0.7934 Lutjanus vittus -10.4418(0.1823) 2.8824(0.0351) 210 6,752.525 0.9697 Nemipterus furcosus -9.0380(0.2490) 2.6379(0.0491) 238 2,888.362 0.9236 Nemipterus hexodon -10.5120(0.1626) 2.9366(0.0327) 268 8,081.145 0.9678 Pristipomoides multidens -10.4544(0.1318) 2.9235(0.0226) 96 16,739.747 0.9942 Pristipomoides typus -10.3553(0.1890) 2.8933(0.0337) 27 7,358.630 0.9962 Psenopsis humerosa -12.0558(0.5391) 3.2969(0.1054) 99 978.916 0.9072 Saurida micropectoralis -12.4764(0.3404) 3.1777 (0.060.3) 162 2,777.965 0.9446 Male A A Species' X, (SE) X, (SE) n-2 F,.n-2 R- Diagramma pictum -11.8373(0.1399) 3.1102(0.0230) L17 18,239.181 0.9936 Lethrinus fraenatus -11.5601 (0.5382) 3.1252(0.0901) 1 1 1,204.055 0.9877 Lethrinus lentjan -11.1870(0.3227) 3.0589(0.0558) 72 3,002.105 0.9763 Lutjanus erythropterus -9.9051 (0.2875) 2.8006(0.0474) 91 3.487.397 0.9743 Lutjanus malabaricus -10.6166(0.1268) 2.9171 (0.0209) 198 19,525.208 0.9899 Lutjanus sebae -11 5487(0.2166) 3.1216(0.0359) 43 7,544.416 0.9942 Lutjanus timorensis -9.4597 (1.8694) 2.7597 (0.3085) 15 80.011 0.8316 Lutjanus vittus -10.5218(0.1447) 2.9007(0.0274) 223 11,186.525 0.9804 Nemipterus furcosus -10.9360(0.1538) 3.0150(0.0297) 203 10.282.691 0.9805 Nemipterus hexodon -10.9499(0.2803) 3.0188(0.0550) 123 3,011.431 0.9604 Pristipomoides multidens -10.3481 (0.1032) 2.9054(0.0179) 125 26.357.314 ii 9952 Pristipomoides typus -10.3289(0.1707) 2.8902(0.0308) 33 8,794.451 0.9961 Psenopsis humerosa -10.6433(0.3927) 3.0174(0.0780) 115 1,495.187 0.9280 Sa u rida m icropectoralis -11.6679(0.4003) 3.0307 (0.0726) 261 1.744.792 0 8694 See Table 1 for common names. NOTE Xiao and Ramm: A simple model of allometry for groundfish 669 Table 3 Estimates of mean and asymptotic standard error (ASE) of a llometric parameters obtained for mixed sexes, males, and females of each of 14 species of groundfish, caught in northern Australian waters during 20 October to 16 December 1990 by fitting Equations 4 and 5 with V\Y\X3] -- = <3%X3 excluding the model error term (cr). Species' Mixed Jj, (ASE) M, (ASE) o, (ASE) Diagramma pictum -11.4010(0.0624) 3.0386(0.0107) 0.015684(0.000536) Lethrinus fraenatus -11.0788(0.2791) 3.0450(0.0480) 0.011364(0.001120) Lethrinus lentjan -10.8528(0.1295) 3.0022(0.0231) 0.011431 (0.000427) Lutjanus erythropterus -10.2207 (0.2246) 2.8559(0.0371) 0.011478(0.000598) Lutjanus malabaricus -10.4315(0.0455) 2.8858(0.0079) 0.015562(0.000445) Lutjanus sebae -10.7324(0.0705) 2.9885(0.0124) 0.013522(0.000691) Lutjanus timorensis -10.3142 (0.4613) 2.9015(0.0766) 0.010599(0.001097) Lutjanus vittus -10.5948(0.0968) 2.9132(0.0186) 0.012502(0.000405) Nemipterus furcosus -10.3566(0.1275) 2.8986(0.0252) 0.028576(0.000918) Nemipterus hexodon -10.8451 (0.1281) 3.0006(0.0257) 0.021340(0.000683) Pristipomoides multidens -10.4311 (0.0626) 2.9196(0.0111) 0.012057 (0.000483) Pristipomoides typus -10.6917(0.0692) 2.9547(0.0134) 0.012299(0.000737) Psenopsis humerosa -11.8293(0.2619) 3.2521(0.0518) 0.015465(0.000673) Saurida micropectoralis -12.3549(0.1919) 3.1555(0.0346) 0.017171(0.000567) Species' Female \L2 (ASE) Ji, (ASE) a, (ASE) Diagramma pictum -11.4694(0.1271) 3.0499(0.0214) 0.016512(0.000844) Lethrinus fraenatus -10.8822(0.4288) 3.0111(0.0747) 0.011972(0.001450) Lethrinus lentjan -10.9932(0.1843) 3.0276(0.0331) 0.011992(0.000515) Lutjanus erythropterus -11.1343(0.3913) 3.0106(0.0646) 0.009438(0.000718) Lutjanus malabaricus -10.6957(0.1742) 2.9293(0.0284) 0.015106(0.000754) Lutjanus sebae -10.9216(0.1938) 3.0211 (0.0328) 0.013024(0.000956) Lutjanus timorensis -8.5443(1.5702) 2.6085(0.2606) 0.011468(0.001567) Lutjanus vittus -10.4305(0.1808) 2.8802(0.0348) 0.012749(0.000602) Nemipterus furcosus -8.0532(0.2559) 2.4435(0.0506) 0.030733 (0.001396) Nemipterus hexodon -10.4947(0.1627) 2.9332(0.0328) 0.017758(0.000753) Pristipomoides multidens -10.4486(0.1296) 2.9225(0.0224) 0.012449(0.000864) Pristipomoides typus -10.3916(0.1853) 2.8998(0.0333) 0.011512(0.001461) Psenopsis humerosa -12.0680(0.5276) 3.2992(0.1032) 0.015033 (0.001037) Saurida micropectoralis -12.4710(0.3370) 3.1768(0.0598) 0.017557 (0.000955) Species' - Male M, (ASE) Vl2 (ASE) a., (ASE) Diagramma pictum -11.7895(0.1296) 3.1023(0.0214) 0.012092(0.000760) Lethrinus fraenatus -11.5461 (0.4630) 3.1229(0.0776) 0.009625(0.001619) Lethrinus lentjan -11.1988(0.3127) 3.0609(0.0541) 0.009055(0.000704) Lutjanus erythropterus -9.9036(0.2774) 2.8003(0.0458) 0.011754(0.000834) Lutjanus malabaricus -10.5854(0.1203) 2.9120(0.0199) 0.014870(0.000728) Lutjanus sebae -11.5462(0.2006) 3.1212(0.0334) 0.009834(0.000989) Lutjanus timorensis -9.4896(1.7559) 2.7646(0.2898) 0.008735(0.001411) Lutjanus vittus -10.5171 (0.1433) 2.8998(0.0272) 0.012371 (0.000566) Nemipterus furcosus -10.9048(0.1524) 3.0089(0.0295) 0.014279(0.000690) Nemipterus hexodon -10.9325(0.2732) 3.0153 (0.0538) 0.023605(0.001481) Pristipomoides multidens -10.3460(0.1004) 2.9051 (0.0175) 0.011230(0.000680) Pristipomoides typus -10.3560(0.1709) 2.8951 (0.0311) 0.010902(0.001254) Psenopsis humerosa -10.6920(0.3878) 3.0271(0.0771) 0.014853(0.000951) Saurida micropectoralis -11.6947(0.3976) 3.0356(0.0721) 0.016898(0.000725) 1 See Table 1 for common names. 670 Fishery Bulletin 92(3). 1994 donia ( Kulbicki et al., 1993 ), systematic data are lack- ing from northern Australian waters. Because our data covered relatively large size ranges of each of the 14 species offish concerned, our estimates of al- lometric parameters and associated relationships will improve stock assessments of major groundfish in northern Australian waters. Acknowledgments We wish to thank J. Robinson of the School of Math- ematics and Statistics of the University of Sydney for his valuable suggestion of the model, R. C. Buckworth and I. Knuckey, both of our Fisheries Division, for their useful comments on the manu- script, and A. Coleman and J. Lloyd also of our Fish- eries Division for their assistance in acquiring the weight at length data. This project was funded partly by the Australian Fisheries Research and Develop- ment Corporation (90/15) and the former Australian Fisheries Service. Literature Cited Bookstein, F. L., B. Chernoff, R. L. Elder, J. M. Humphries, Jr., G. R. Smith and R E. Strauss. 1985. Morphometries in evolutionary biology: the geometry of size and shape change, with examples from fishes. Special Publication 15, The Academy of Natural Sciences of Philadelphia, 277 p. Caillouet, C. W., Jr. 1993. On comparing groups of fishes based on length- weight relationships. NAGA, The ICLARM Quar- terly 16:30-31. Calder, Vf. A. 1984. Size, function, and life history. Harvard Univ. Press, Cambridge, MA, 431 p. Kulbicki, M., G. Mou Tham, P. Thollot, and L. Wantiez. 1993. Length-weight relationships offish from the lagoon of New Caledonia. NAGA, The ICLARM Quarterly 16:26-30. Mosimann, J. E., and F. C. James. 1979. New statistical methods for allometry with application to Florida red-winged blackbirds. Evolution 33:444-459. Peters, R. E. 1983. The ecological implications of body size. Cambridge Univ. Press, Cambridge, 329 p. Pienaar, L. V., and W. E. Ricker. 1968. Estimating mean weight from length statistics. J. Fish. Res. Board Can. 25:2743-2747. Reiss, M. J. 1989. The allometry of growth and reproduction. Cambridge Univ. Press, Cambridge, England, 182 p. Saenger, R. A. 1989. Bivariate normal swimbladder size allometry models and allometric exponents for 38 mesope- lagic swimbladdered fish species commonly found in the North Sargasso Sea. Can. J. Fish. Aquat. Sci. 46:1986-2002. SAS Institute, Inc. 1985. SAS user's guide, Version 5. SAS Institute, Cary, NC, 956 p. Schmidt-Nielsen, K. 1984. Scaling: why is animal size so important? Cambridge Univ. Press, Cambridge, 241 p. Seim, E., and B. Saether 1983. On rethinking allometry: which regression model to use? J. Theor. Biol. 104:161-168. Shoesmith, E. 1990. A comparison of methods for estimating mean fecundity. J. Fish. Biol. 36:73-84. Showers, P. A. T. 1993. Length-weight relationships of five species of the family Sparidae in the Gulf of New Guinea. NAGA, The ICLARM Quarterly 16:32-33. Wilkinson, L. 1989. SYSTAT: The system for statistics. SYSTAT, Inc., Evanston, IL, 638 p. Fishery Bulletin Guide for Contributors Preparation Title page should include authors' full names and mailing addresses and the senior author's telephone and FAX number. Abstract should not exceed one double-spaced typed page. It should state the main scope of the research but emphasize its conclusions and relevant findings. Because abstracts are circulated by abstracting agencies, it is important that they represent the research clearly and concisely. Text must be typed double-spaced throughout. A brief introduction should portray the broad significance of the paper; the remainder of the paper should be divided into the following sections: Materials and methods, Results, Discussion (or Conclusions), and Acknowledgments. Headings within each section must be short, re- flect a logical sequence, and follow the rules of multiple subdivision (i.e. there can be no subdivision without at least two items). The entire text should be intelligible to interdisciplinary readers; therefore, all acronyms, abbreviations, and technical terms should be spelled out the first time they are mentioned. The scientific names of species must be written out the first time they are mentioned; subsequent mention of scientific names may be abbreviated. Follow the U.S. Government Printing Office Style Manual (1984 ed.) and the CBE Style Manual (5th ed.) for editorial style, and the most current issue of the American Fisheries Society's Common and Scientific Names of Fishes from the United States and Canada for fish nomencla- ture. Dates should be written as fol- lows: 11 November 1991. Measure- ments should be expressed in metric units, e.g., metric tons as (t); if other units of measurement are used, please make this fact explicit to the reader. The numeral one (1) should be typed as a one, not as a lower-case el (1). Text footnotes should be numbered with Arabic numerals and typed on a separate sheet of paper. Footnote all personal communications, unpub- lished data, and unpublished manu- scripts with full address of the com- municator or author, or, as in the case of unpublished data, where the data are on file. Authors are advised to avoid references to nonstandard (gray) literature, such as internal, project, processed, or administrative reports, wherever possible. Where these refer- ences are used, please include whether they are available from NTIS (Nation- al Technical Information Service) or from some other public depository. Literature cited comprises published works and those accepted for publica- tion in peer-reviewed literature (in press). Follow the name and year sys- tem for citation format. In the text, cite Smith and Jones (1977) or (Smith and Jones, 1977). If there is a se- quence of citations, list chronological- ly: Smith, 1932; Green, 1947; Smith and Jones, 1985. Abbreviations of serials should conform to abbrevia- tions given in Serial Sources for the BIOSIS Previews Database.. Authors are responsible for the accuracy and completeness of all citations. Tables should not be excessive in size and must be cited in numerical order in the text. Headings should be short but ample enough to allow the table to be intelligible on its own. All unusual symbols must be explained in the table heading. Other incidental comments may be footnoted with italic numerals. Use asterisks for probability in statis- tical data. Because tables are typeset, they need only be submitted typed and formatted, with double-spaced legends. Zeros should precede all decimal points for values less than one. Figures include line illustrations and photographs (or slides) and must be cited in numerical order in the text. Unless photographs are submitted on glossy paper with good contrast, we cannot guarantee a good final printed copy. Figures are to be labeled with author's name and number of figure. Use Times Roman font (upper and lowercase letters) to label within figures. Avoid vertical lettering except for y-axis labels. Zeros should precede all decimal points for values less than one. Figures may be submitted as com- puter software files (along with laser- printed copies), as photomechanical transfers (PMTs), or as high quality photographic prints. Send only xerox copies of figures to the Scientific Edi- tor; original figures will be requested later when the manuscript has been accepted for publication. Figure leg- ends should explain all symbols and abbreviations and should be double- spaced on a separate page at the end of the manuscript. Copyright law does not cover govern- ment publications; they fall within the public domain. If an author repro- duces any part of a government publi- cation in his work, reference to source is considered correct form. Submission Send printed copies (original and two copies) to the Scientific Editor: Dr. Ronald Hardy, Scientific Editor Northwest Fisheries Science Center, F/NWC3 National Marine Fisheries Service, NCAA 2725 Montlake Boulevard East Seattle, WA 98112-2097 Once the manuscript has been ac- cepted for publication, you will be asked to submit a software copy of your manuscript to the Managing Editor. The software copy should be submitted in WordPerfect text format (or in standard ASCII text format if WordPerfect is unavailable) and should be placed on a 5.25-inch or 3.5- inch disk that is double-sided, double or high density, and that is compatible with either DOS or Apple Macintosh systems. Copies of published articles and notes are available free of charge to the senior author (50 copies) and to his or her laboratory (50 copies). Addition- al copies may be purchased in lots of 100 when the author receives page proofs. U.S. Department of Commerce Volume 92 Number 4 October 1994 Fishery Bulletin .Marina Biological Laboratory/ os Hola Oceancgraphic Institution OCT 1 1 1994 Hoto, MA 0*i>43 U.S. Department of Commerce Ronald H. Brown Secretary National Oceanic and Atmospheric Administration D. James Baker Under Secretary for Oceans and Atmosphere National Marine Fisheries Service Rolland A. Schmitten Assistant Administrator for Fisheries Scientific Editor Dr. Ronald W. Hardy Northwest Fisheries Science Center National Marine Fisheries Service. NOAA 2725 Montlake Boulevard East Seattle, Washington 981 12-2097 Editorial Committee Dr. Andrew E. Dizon National Marine Fisheries Service Dr. Linda L. Jones National Marine Fisheries Service Dr. Richard D. Methot National Marine Fisheries Service Dr. Theodore W. Pietsch University of Washington Dr. Joseph E. Powers National Marine Fisheries Service Dr. Tim D. Smith National Marine Fisheries Service 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, Attn: Chief, Mail List Branch, Mail Stop SSOM, Washington, DC 20402-9373. 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 peri- odical 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 ap- proved 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: $27.00 domestic and $33.75 foreign. Cost per single issue: $13.00 domestic and $16.25 foreign. See back page for order form. Managing Editor Sharyn Matriotti National Marine Fisheries Service Scientific Publications Office 7600 Sand Point Way NE, BIN C 1 5700 Seattle. Washington 98 1 1 5-0070 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 92 Number 4 October 1994 Fishery Bulletin Contents Articles 671 Barbieri, Luiz R., Mark E. Chittenden Jr., and Susan K. Lowerre-Barbieri Maturity, spawning, and ovarian cycle of Atlantic croaker, Micropogonias undulatus, in the Chesapeake Bay and adjacent coastal waters 686 Biesiot, Patricia M., Robert E. Caylor, and James S. Franks Biochemical and histological changes during ovarian development of cobia, Rachycentron canadum, from the northern Gulf of Mexico 697 Connaughton, Martin A., and Malcolm H. Taylor Seasonal cycles in the sonic muscles of the weakfish, Cynoscion regalis 704 Doroff, Angela M., and Anthony R. DeGange Sea otter, Enhydra lutris, prey composition and foraging success in the northern Kodiak Archipelago 71 1 Jacobson, Larry D., Nancy C. H. Lo, and J. Thomas Barnes A biomass-based assessment model for northern anchovy, Engraulis mordax 725 Kendall, Arthur W., Jr., and Ann C. Matarese Status of early life history descriptions of marine teleosts 737 Koslow, Julian A., Karl Aiken, Stephanie Auil, and Antoinette Clementson Catch and effort analysis of the reef fisheries of Jamaica and Belize 748 Lehodey, Patrick, Paul Marchal, and Rene Grandperrin Modelling the distribution of alfonsino, Beryx splendens, over the seamounts of New Caledonia Fishery Bulletin 92(4). 1994 760 McCabe, George T, Jr., and Charles A. Tracy Spawning and early life history of white sturgeon, Acipenser transmontanus, in the lower Columbia River 773 Mullin, Keith D., Wayne Hoggard, Carol L. Roden, Ren R. Lohoefener, Carolyn M. Rogers, and Brian Taggart Cetaceans on the upper continental slope in the north-central Gulf of Mexico 787 Parker, Richard O., Jr., Alexander J. Chester, and Russell S. Nelson A video transect method for estimating reef fish abundance, composition, and habitat utilization at Gray's Reef National Marine Sanctuary, Georgia 800 Shimada, Allen M., and Daniel K. Kimura Seasonal movements of Pacific cod, Oadus macrocephalus, in the eastern Bering Sea and adjacent waters based on tag-recapture data 8 1 7 Thresher, Ronald E., Craig H. Proctor, John S. Gunn, and Ian R. Harrowfield An evaluation of electron-probe microanalysis of otoliths for stock delineation and identification of nursery areas in a southern temperate groundfish, Nemadactylus macropterus (Cheilodactylidae) 841 Wilson, Charles A., and David L. Nieland Reproductive biology of red drum, Sciaenops ocellatus, from the neritic waters of the northern Gulf of Mexico 851 Wilson, Christopher D., and Michael R Seki Biology and population characteristics of Squalus mitsukurii from a seamount in the central North Pacific Ocean Notes 865 Erzini, Karim, and Margarida Castro Measures of dispersion as constraints for length-frequency analysis 872 Giorgi, Albert E., David R. Miller, and Benjamin R Sandford Migratory characteristics of juvenile ocean-type Chinook salmon, Oncorhynchus tshawytscha, in John Day Reservoir on the Columbia River 880 Laurenson, Laurie J. B., Ian C. Potter, and Norm G. Hall Comparisons between generalized growth curves for two estuarine populations of the eel tailed catfish Cnidoglanis macrocephalus 890 Raynie, Richard C, and Richard F. Shaw A comparison of larval and postlarval gulf menhaden, Brevoortia patronus, growth rates between an offshore spawning ground and an estuarine nursery 895 Index 908 Errata 909 Awards Abstract. — The reproductive bi- ology of Atlantic croaker, Micropo- gonias undulatus, collected during 1990-91 from commercial catches in Chesapeake Bay and in Virginia and North Carolina coastal waters (n=3,091), was studied by using macroscopic and microscopic gonad staging, the gonadosomatic index, oocyte diameter distributions, and histological analysis. Atlantic croaker are multiple spawners with asynchronous oocyte development and indeterminate fecundity. Mean length at first maturity for males and females was 182 and 173 mm TL, respectively. More than 85% of both sexes were mature by the end of their first year and all were ma- ture by age 2. Spawning extends over a protracted period (July-De- cember), but individual fish appar- ently spawn over a shorter inter- val. Eleven gravid and running- ripe females were collected within the Chesapeake Bay suggesting some spawning occurs in estuarine waters. Monthly sex ratios indi- cated a strong predominance of fe- males during the main period of spawning. A high incidence of atretic, advanced yolked oocytes in spawning females collected through- out the spawning season suggests that a surplus production of yolked oocytes may be part of the repro- ductive strategy of Atlantic croaker. Maturity, spawning, and ovarian cycle of Atlantic croaker, Micropogonias undulatus, in the Chesapeake Bay and adjacent coastal waters* Luiz R. Barbieri** Mark E. Chittenden Jr. Susan K. Lowerre-Barbieri** Virginia Institute of Marine Science. College of William and Mary Gloucester Point, Virginia 23062 The Atlantic croaker, Micropogonias undulatus (Linnaeus), ranges from Cape Cod, Massachusetts, to the Bay of Campeche, Mexico (Welsh and Breder, 1923; Johnson, 1978). Although not common north of New Jersey (Hildebrand and Schroeder, 1928; McHugh, 1981), it is one of the most abundant inshore, demersal species of the Atlantic and Gulf of Mexico coasts of the United States (Joseph, 1972). Despite the large number of stud- ies describing spawning periodicity of Atlantic croaker in the mid-At- lantic and Chesapeake regions (e.g. Hildebrand and Schroeder, 1928; Wallace, 1940; Johnson, 1978; Colton et al., 1979; Morse, 1980; Norcross and Austin, 1988), studies on repro- ductive biology are rare and mostly incomplete. Information on sexual maturity, fecundity, and sex ratios has been reported (Hildebrand and Schroeder, 1928; Wallace, 1940; Morse, 1980). However, speculation on whether or not Atlantic croaker spawn within Chesapeake Bay (Welsh and Breder, 1923; Pearson, 1941; Haven, 1957) has not been investigated; estimates of size at maturity (Wallace, 1940; Morse, 1980) do not agree; estimates of age at maturity (Welsh and Breder, 1923; Wallace, 1940) were based on length frequency and scale ageing, which have been shown to be less accurate than otolith ageing for At- lantic croaker (Joseph, 1972; Barbieri et al., 1994); and available fecundity estimates (Morse, 1980) cannot be used without an evalua- tion of Atlantic croaker's fecundity pattern, i.e. whether they have de- terminate or indeterminate annual fecundity. Traditionally, estimates of fish fecundity have been based on the assumption that the total number of eggs spawned by a female each year (annual fecundity) is fixed prior to the onset of spawning, a condition known as determinate fe- cundity (Hunter et al., 1992). How- ever, recent evidence (Hunter and Goldberg, 1980; Hunter and Mace- wicz, 1985a; Hunter et al., 1985; Horwood and Greer Walker, 1990) indicates that in many temperate and tropical fish annual fecundity cannot be estimated from the stand- ing stock of advanced oocytes be- cause unyolked oocytes continue to be matured and spawned through- Manuscript accepted 9 May 1994. Fishery Bulletin 92:671-685 (1994). 'Contribution No. 1871 from The College of William and Mary, School of Marine Science, Virginia Institute of Marine Science, Gloucester Point, Virginia 23062. "Present address: University of Georgia Marine Institute, Sapelo Island, Georgia 31327. 67! 672 Fishery Bulletin 92|4). 1994 out the spawning season. This condition is called indeterminate fecundity (Hunter et al., 1992). The only way to estimate annual fecundity, therefore, is by estimating batch fecundity — the number of eggs released during each spawning — and multiplying it by spawning frequency — the number of times an average female spawns during the spawning season (Hunter and Macewicz, 1985a; Hunter et al., 1985, 1992). Although the extended spawning season of Atlantic croaker (Wallace, 1940; Colton et al., 1979; Warlen, 1982) suggests it is a multiple spawner with indetermi- nate fecundity, no attempt has been made to evaluate its fecundity pattern. In this study we test the assumption of determinate annual fecundity and describe spawning periodicity, size and age at maturity, sex ratios, ovarian cycle, and oocyte atresia for Atlantic croaker in the Chesapeake Bay and adjacent coastal waters. Methods Four approaches were used to sample Atlantic croaker for this study. In 1990 and 1991 fish were collected from com- mercial poundnet, haul-seine, and gillnet fisheries that operate from late spring to early fall in the lower Chesa- peake Bay (Fig. 1). Local fish process- ing houses and seafood dealers were contacted weekly during April-October 1990 and 1991, and one 22.7 kg (50 lb) box of fish of each available market grade (small, medium, or large) was purchased for processing. Since Atlan- tic croaker migrate out of Chesapeake Bay in midfall to overwinter offshore (Haven, 1959), monthly samples from November to March 1990 and from No- vember through December 1991 were obtained from commercial trawlers op- erating in Virginia and North Carolina shelf waters. In addition to these col- lections, daily samples from a gill net in the lower York River were obtained during the periods August-October 1990 and July-October 1991, except on weekends. In 1991 the net was emptied twice a day: in the early morning (6:00- 8:00 am) and in the evening (5:00-7:00 pm). Time of death was recorded for fish alive at the time the net was emptied. Daily gillnet samples were used to monitor small- scale (less than weekly) changes in Atlantic croaker reproductive condition and to collect hydrated or re- cently spawned females. Finally, collections from the commercial fisheries were supplemented by fish ob- tained from the Virginia Institute of Marine Science (VIMS) juvenile bottom trawl survey. The VIMS trawl survey used a monthly stratified random sampling program in the lower Chesapeake Bay and monthly Figure 1 Map of the Chesapeake Bay and mid-Atlantic region. Black dots in Chesapeake Bay indicate poundnet, haul-seine, or gillnet collection sites. Hatched area off Virginia and North Carolina indicates where otter trawl collections of Atlantic croaker, Micropogonias undulatus, were obtained. Barbieri et al.: Maturity, spawning, and ovarian cycle of Micropogonias undulatus 673 fixed midchannel stations in the York, James, and Rappahannock rivers. Fish were measured for total length (TL, ±1.0 mm), total weight (TW, ±1.0 g), and gonad weight (GW, ±1.0 mm), sexed, and both sagittal otoliths were re- moved and stored dry. The left otolith was sectioned through the core with the diamond blade of a Buehler low-speed Isomet saw. Sections 350-500 urn thick were then mounted on glass slides with Flo-texx clear mounting medium and aged under a dissecting mi- croscope (6-12x) following criteria described in Barbieri et al. ( 1994 ). The gonadosomatic index ( GSI ) was calculated for individual fish as (GW/(TW- GW)100). Females were assigned a macroscopic go- nad maturity stage (Table 1). Males were classified only as sexually mature or immature. Female mac- roscopic stages were verified microscopically by in- specting fresh oocyte samples and histology slides of a randomly selected subsample of ovaries in each maturity stage. Fresh oocytes were removed from one ovary, spread on a microscope slide, and examined under a dissecting microscope (12— 50x). Color pho- tographs were used to permanently record the ap- pearance of fresh oocyte samples. This technique al- lowed fresh oocytes to be compared with histology slides in assessing gonad maturity stage and the Table 1 Description of gonad maturity stages for female Atlantic croaker, Micropogonias undulatus, in the Chesapeake Bay and adjacent coastal waters. Macroscopic appearance refers to fresh ovaries. Gonad stages 3, 4, and 5 are in spawn- ing phase. FOM = final oocyte maturation; POF's = postovulatory follicles. Stage Macroscopic appearance Microscopic appearance ( 1 ) Immature Ovaries very small, light pink in color; translucent. Only primary growth oocytes present; no atresia; ovarian membrane thin. (2) Developing Ovaries ranging from small to medium (< 25% of body cavity); yellow to light orange in color; no opaque (advanced yolked) oocytes present. Only primary growth, cortical alveoli and a few partially yolked oocytes present; no major atresia. (31 Fully developed Ovaries ranging from large (25-50% of body cavity) to very large (-100% of space available in body cavity); creamy yellow to orange in color; opaque oocytes prevalent and easily detected; if partially spent, may have some left-over clear (hydratedl oocytes present at the posterior end of the ovarian lumen. Primary growth to advanced yolked oocytes present; may have some left- over hydrated oocytes from previous spawning; often major atresia of advanced yolked oocytes, but no major atresia of other oocytes. (4) Gravid Ovaries ranging from large (25-50% of body cavity) to very large (-100% of space available in body cavity); creamy yellow to light orange in color; unovulated clear (hydrated) oocytes visible amongst opaque oocytes, giving a speckled appearance; clear oocytes not collected in the ovarian lumen. Primary growth to FOM/hydrated oocytes present; often major atresia of advanced yolked oocytes, but no major atresia of other oocytes; hydrated oocytes are still in follicles (unovulated). (5) Running-ripe Ovaries ranging from large (25-50% of body cavity) to very large (-100% of space available in body cavity); creamy yellow to light orange in color; most clear (hydrated) oocytes are collected in the ovarian lumen (ovulated). Primary growth to FOM/hydrated oocytes present; often major atresia of advanced yolked oocytes, but no major atresia of other oocytes; hydrated oocytes not in follicles (ovulated); may have POF's. (6) Regressing Ovaries ranging from small to medium (<25'7r of body cavity); mustard yellow to light orange in color; more flaccid than previous stages; often contain clear fluid; can detect a few opaque oocytes. Primary growth to advanced yolked oocytes present, but the number of yolked oocytes relative to unyolked oocytes is now much smaller; major atresia of cortical alveoli, partially yolked and advanced yolked oocytes. (7) Resting Ovaries very small; dark orange to reddish in color; no opaque oocytes present; ovarian membrane thickened and more opaque than in immature fish. The majority (>90%) of oocytes are primary growth; may have other oocytes in late stages of atresia; more follicular tissue than immature fish. 674 Fishery Bulletin 92(4), 1994 occurrence and intensity of oocyte atresia. For histo- logical preparation, tissue samples were fixed in 10% neutrally buffered formalin for 24 hours, soaked in water another 24 hours, and stored in 70% ethanol. Samples were embedded in paraffin, sectioned to 5— 6 (im thickness and stained with Harris' Hematoxy- lin and Eosin Y. Histological classification of ovaries (Table 1) was based on the occurrence and relative abundance of five stages of oocyte development (pri- mary growth, cortical alveoli, partially yolked, ad- vanced yolked, and hydrated) and on the occurrence and intensity of alpha (a) atresia. Terminology for stages of oocyte development and ovarian atresia follows Wallace and Selman (1981), Hunter and Macewicz (1985b) and Hunter et al. ( 1992). Fecundity pattern was evaluated through monthly oocyte diameter distributions of fully developed (go- nad stage 3) females collected during the spawning season. Before measurements were taken, oocytes were hydraulically separated from each other and from the ovarian membrane and preserved in 2% a) a uu /■ Males so ll j\ / 1 LK=182 n = 407 r 2 = 0.97 fi- I 1 1 on 200 300 400 50 «^B Females L„= 173 / 1 n = 612 r ' = 0.97 / 1 J. ♦ ** ' ' ' ' 1 ' ' I 'I inn poo 3( 10 400 Length class (mm) Figure 2 Percentage of mature male and female At- lantic croaker, Micropogonias undulatus, by 10-mm total length intervals, with a logis- tic function (continuous line) fitted to the data. Arrows indicate mean length at first maturity (L50). n=sample size. formalin following the method of Lowerre-Barbieri and Barbieri (1993). Oocyte measurements were taken after a preservation period of at least 24 hours. Samples were stirred before oocytes were removed to reduce bias from settling differences caused by oocyte size or density. Oocytes >0.1 mm were mea- sured (±0.02 mm) with an ocular micrometer in a dissecting microscope. Measurements were taken along the median axis of the oocyte parallel to the horizontal micrometer gradations (Macer, 1974; DeMartini and Fountain, 1981). To estimate mean length at first maturity (L50) for males and females, the fraction of mature fish per 10 mm length intervals was fit to the logistic func- tion by nonlinear regression (Marquardt method), by using the statistical software program FISHPARM (Saila et al., 1988). L50 was defined as the smallest length interval in which 50% of the individuals were sexually mature. Females were considered sexually mature if they were in gonad stage 2 (developing) or higher (Table 1). However, to avoid classifying rest- ing (reproductively inactive) or early developing fish as immature, and thus obtaining biased estimates of Lgo, only fish collected in September, when no rest- ing or developing stages were found, were used for this analysis. Monthly sex ratios were tested statistically for sig- nificant deviations from the expected 1:1 ratio with a chi-square test (a=0.05). Results Size and age at maturity Atlantic croaker mature at a small size and early age. Males and females started to mature at 170 and 150 mm, respectively; at lengths greater than these the percentages of mature fish increased rapidly ( Fig. 2). Estimated mean length at first maturity < L50) was 182 mm for males (SE=1.46 mm) and 173 mm for females (SE=1.33 mm). For both sexes, all individu- als >250-260 mm were mature. The percentage of mature fish by age showed a similar pattern of early maturation. More than 85% of both males and females were sexually mature by the end of their first year and all were mature by the end of their second. Spawning Spawning of Atlantic croaker in the Chesapeake Bay and adjacent coastal waters extends over a protracted period. Females in spawning phase (gonad stages: fully developed (3), gravid (4), or running-ripe (5); Table 1 ) were collected from July through December Barbieri et al.: Maturity, spawning, and ovarian cycle of Micropogonias undulatus 675 (Fig. 3). However, the capture of developing females (stage 2) from May through August and regressing females (stage 6) from September through Decem- ber indicates that spawning initiation and cessation were not synchronous among individuals. Although the population spawned over a six-month interval (July-December), individuals apparently spawned over a shorter period. Most females appeared to spawn for 3^1 months as indicated by the large per- centages of fully developed (stage 3) ovaries from August through November. The pattern of gonad development in males pro- vided further evidence of an extended spawning sea- son for Atlantic croaker. Mean and maximum GSI values increased sharply during July and August, and remained relatively high until November or De- cember, depending on the year (Fig. 4). In addition, males with very large testes and free-running milt were common during August-September in collec- tions from all locations and sampling gears, indicat- ing intense male spawning during this period. Spawning of Atlantic croaker occurred in the estu- ary as well as in coastal oceanic waters. Females with hydra ted oocytes (gonad stages 4 and 5), indicative of imminent spawning, were collected in the lower York and James rivers {n=8 in 1990; n=3 in 1991; Fig. 3) and from coastal waters off Virginia and North Carolina (rc = l in 1990; n=3 in 1991; Fig. 3). Collec- tions of spawning fish (gonad stages 3-5) in Chesa- peake Bay during the period July-October («=649 in 1990; n=277 in 1991; Fig. 3) and from offshore waters during November-December in=39 in 1990; n=ll in 1991; Fig. 3) suggest spawning continues offshore and south as Atlantic croaker migrate from the estuary. However, the presence during Septem- ber-October of regressing and resting females in Chesapeake Bay (n=39 in 1990; «=24 in 1991; Fig. 3 ) indicates that some individuals may complete their spawning in estuarine waters. Although gravid and running-ripe females were collected during most of the spawning season (Au- gust-November, Fig. 3), they were present in very low numbers. During both years of sampling only seven gravid and eight running-ripe females were collected. In Chesapeake Bay, despite the large num- ber of poundnet and haul-seine collections (1,422 mature females processed), gravid or running-ripe females were obtained only from gill nets and mainly from collections from the lower James River (six gravid and four running-ripe females). Daily gillnet collections in the lower York River during August- October 1990 and July-October 1991 (456 mature females processed) showed only one running- ripe and one partially spent female, i.e. a fully developed fe- male that had fresh left-over hydrated oocytes in the I 1 2 3 4 5 6 7 Feb 2 3 4 5 8 7 2 3 4 5 8 7 n I Aug 2 3 4 5 8 7 I Sep 2 3 4 5 6 7 Apr 2 3 4 5 L 2 3 4 5 6 7 May Ju 2 3 4 5 6 7 LJ 2 3 4 5 6 7 2 3 4 5 6 7 I It u 234567 234567 Gonad stage Figure 3 Percentage of gonad maturity stages by month for mature female Atlantic croaker, Micro- pogonias undulatus, in the Chesapeake Bay and adjacent coastal waters. Black bars=1990 data; open bars=1991 data. Gonad stages are (2) developing; (3) fully developed; (4) gravid; (5) running-ripe; (6) regressing; and (7) rest- ing. Monthly sample sizes are presented in Table 2. Samples from April to October are from Chesapeake Bay; samples from Novem- ber to March are from coastal waters off Vir- ginia and North Carolina. ovarian lumen indicating recent spawning but that still had a large number of advanced yolked oocytes and could potentially spawn again. Offshore collec- tions during November-December of 1990 and 1991 676 Fishery Bulletin 92(4), 1994 20 10 Males (n = 896) CD X) c ID o o a 03 C o O 20 JMMJSNJMMJSN Females (n = 2.195) 10 No samples JMMJSNJMMJSN - 1990 - — II — -1991 - Figure 4 Monthly mean gonadosomatic index for male and female Atlantic croaker, Micropogonias undulatus, in the Chesapeake Bay and adjacent coastal waters, 1990-91. Vertical bars are ranges. n=sample size. Monthly sample sizes are presented in Table 2. also showed a small percentage of gravid and run- ning-ripe females (Fig. 3). Sex ratios Atlantic croaker showed wide temporal fluctuations in sex ratio. During both years, the frequency of males in samples decreased in June and July at the beginning of the spawning season, reached a mini- mum in the period of September-October, and in- creased again during November-December. Chi- square tests (Table 2) showed highly significant dif- ferences (P<0.01) in sex ratios between months over the periods July-October 1990 and June-October 1991. Table 2 Number of mal es and fe males by mor th and chi- square tests for th e mon thl.y sex ratios of Atlantic croake r, Micropogonias undulatus, in the Chesa- peake Bay and adjacent coastal waters, 1990-91. **=P<0.01. Year Month N umber of Chi-square males females 1990 Jun 107 71 3.64 Jul 185 358 27.80** Aug 132 357 51.74** Sep 40 249 74.91** Oct 33 99 16.50** Nov 56 64 0.22 Dec 41 33 0.37 1991 Jan 22 26 0.04 Feb 27 27 — Mar 25 23 0.04 Apr 36 51 1.29 Mav 98 121 1.10 Jun 52 129 15.96** Jul 44 103 11.84** Aug 21 122 34.96** Sep 16 119 38.99** Oct 9 75 25.61** Nov 15 33 3.37 Dec 32 40 0.44 Oocyte development and spawning pattern Monthly oocyte diameter distributions of fully devel- oped females collected throughout the spawning sea- son showed three main groups of oocytes (Fig. 5). However, oocyte development appears to be asynchro- nous; there is a large degree of overlap and no clearly defined limits between modal groups. Histological analysis showed that the first group, ranging approxi- mately from 0.06 to 0.24 mm diameter, is composed mainly of primary growth and cortical alveolus oo- cytes but may include a few partially yolked oocytes in the beginning stages of yolk deposition ( 0.22-0.24 mm diameter). The second group, ranging approxi- mately from 0.26 to 0.38 mm diameter, is composed of partially yolked oocytes in several stages of yolk deposition. The third group, ranging approximately from 0.40 to 0.60 mm diameter, is formed by advanced yolked oocytes and probably represents the group from which individual spawning batches will be formed. Although frequency distributions of gonad stages and oocyte diameters (Figs. 3 and 5) indicated At- lantic croaker are multiple spawners with indeter- minate fecundity, postovulatory follicles (POF's) were identified only in recently ovulated, running-ripe fe- Barbieri et al.: Maturity, spawning, and ovarian cycle of Micropogonias undulatus 677 July QSI = 12.7 % n = 512 "hftihJliTK October GSI = 9 68 % n = 445 iTTkffMlllti 0.1 0.2 0.3 0.4 16 August GSI = 12 5% n = 616 MfcJk November GSI = 6.8 % n = 448 MU tu 0 1 0.2 0.3 0.4 0.5 September GSI = 7.06 % n = 491 TJlkMllL December GSI = 5 9 % n = 533 0 2 0 3 0.4 0.5 0 2 0 3 0 4 0.5 Oocyte diameter (mm) Figure 5 Monthly oocyte diameter distributions during the spawning sea- son of Atlantic croaker, Micropogonias undulatus, in the Chesa- peake Bay and adjacent coastal waters. Each panel represents one female in the fully developed gonad stage. Samples from July to October are from Chesapeake Bay; samples from November to December are from coastal waters off Virginia and North Caro- lina. GSI=gonadosomatic index; «=number of oocytes measured. males. No POF's were found in fully developed fe- males, even those with left-over hydrated oocytes in the posterior end of the ovarian lumen. As a result, it was usually impossible to distinguish fully devel- oped females spawning for the first time from those that had spawned at least once before (partially spent females). Atresia of advanced yolked oocytes Spawning-phase Atlantic croaker females (gonad stages 3-5; Table 1) showed a high incidence of a atresia of advanced yolked oocytes throughout the spawning season (July-December). However, the exact proportion of atretic oocytes could not be de- termined because of the difficulty in identifying oo- cytes in very early stages of atresia. Some females showed healthy advanced yolked oocytes, atretic advanced yolked oocytes in different stages of degeneration, and atretic follicles (p-, Y"> and 8-stage atresia) in the same ovary. Compared with healthy oocytes (Fig. 6A), early phases of a atresia of advanced yolked oocytes in Atlantic croaker are character- ized by the disintegration of the nucleus, which loses its integrity, becoming amor- phous and slightly basophilic, and by the disintegration of yolk globules, which be- gin to dissolve, forming a continuous, amor- phous mass, especially around the nucleus (Fig. 6B). At this stage, the majority of yolk granules at the periphery of the cytoplasm still maintain their structural integrity, spherical shape, and strong acidophilic staining. At intermediate stages, disintegra- tion of yolk globules progresses towards the peripheral cytoplasm, which by now may have a band of dark, basophilic material ( Fig. 6C ), and the zona radiata begins to de- teriorate. At late stages of a atresia (Fig. 6D), the nucleus has completely disap- peared, the zona radiata has lost its struc- tural integrity, and the cytoplasm has been invaded by phagocytizing granulosa cells. Only portions of dissolved yolk and a few yolk globules remain at this stage. However, atresia will continue until the oocyte is com- pletely resorbed, leaving only the remain- ing follicle. After this phase, a-stage atresia has been completed and follicular atresia begins with the resorption of the remain- ing granulosa and thecal cells. Comparisons of fresh oocyte samples with histology slides confirmed the high inci- dence of a atresia of advanced yolked oocytes in At- lantic croaker. Although the histological method ap- peared more sensitive in detecting earlier stages of atresia (Fig. 7A), the use of fresh oocytes was indis- pensable. Fresh oocytes provided an easy, fast way to assess gonad condition and to identify oocyte atresia. A large proportion of atretic advanced yolked oocytes could be easily identified by clumping and darkening of the yolk granules, formation of a clear zone in the peripheral cytoplasm (Fig. 7B), and at later stages, formation of several light yellow vacu- oles (Fig. 7C). Description of the ovarian cycle A diagrammatic representation of the Atlantic croaker ovarian cycle, based on the temporal distri- 678 Fishery Bulletin 92(4), 1994 ? V m / Q <0 wi^i — : ^ Figure 6 Appearance of advanced yolked oocytes of Atlantic croaker, Micropogonias undulatus. (A) healthy (nonatretic) oocyte; (B) oocytes in early stage of a atresia; (C) oocyte in intermediate stage of a atresia; (D) oocytes in late stage of a atresia. N=nucleus; Zr=zona radiata; Pc=peripheral cytoplasm; La=late stage of a atresia. Bars=0.1 mm. bution of maturity stages and the pattern of oocyte development is presented in Figure 8. The cycle can start either with immature females, which enter the cycle for the first time by reaching sexual maturity, or with adult resting females, which restart the cycle by entering the developing stage at the beginning of each spawning season. After the first batch of ad- vanced yolked oocytes is completed, females, now in the fully developed stage, go through a smaller cycle (spawning phase) that characterizes the pattern of multiple spawning and indeterminate fecundity of Atlantic croaker. During this phase, fully developed Barbieri et al.: Maturity, spawning, and ovarian cycle of Micropogonias undulatus 679 ff£3 Figure 6 (Continued) females cycle through the gravid and running-ripe stages by undergoing the processes of hydration, ovu- lation, and spawning. If spawning has not been com- pleted, left-over advanced yolked oocytes are re- sorbed, a new batch of advanced yolked oocytes is recruited from the group of partially yolked oocytes (redeveloping process), and females are ready to go through the cycle again. If spawning is completed, females will then move to the regressing stage, where, through the process of oocyte atresia, left-over oocytes (cortical alveoli to advanced yolked stage) will be resorbed, after which ovaries return to the rest- ing stage. Discussion Spawning periodicity Our results on spawning periodicity of Atlantic croaker agree with previous reports for the Chesa- peake Bay and mid-Atlantic regions. Prior studies 680 Fishery Bulletin 92(4). 1994 Figure 7 Comparison of the appearance of ot-atretic advanced yolked oocytes of a fully developed Atlantic croaker, Micropogonias undulatus, in a histology slide (A), and in a smear of fresh oocytes under a dissecting scope (B) and (C). Cy=clumping of yolk globules; Pc=peripheral cytoplasm; Va=vacuoles. Bars=0.1 mm. (Welsh and Breder, 1923; Wallace, 1940; Johnson, 1978; Colton et al., 1979; Morse, 1980) describe a protracted spawning season, extending from July/ August through November/December, with peak spawning during September/October. However, re- ports of spawning from September/October through March/April along the South Atlantic Bight (Hilde- brand and Cable, 1930; Bearden, 1964; Warlen, 1982; Lewis and Judy, 1983) indicate that south of Cape Hatteras, North Carolina, spawning seems to start a little later and to continue through early spring, perhaps as a result of the southward late summer/ early fall migration of Atlantic croaker (Hildebrand and Schroeder, 1928; Wallace, 1940; Haven, 1959). Barbien et al.: Maturity, spawning, and ovarian cycle of Micropogonias undulatus 681 Immature Next spawning season I Sexual maturity Resting I Developing \ Resorption First batch of advanced yolked oocytes Fully developed Regressing Redeveloping jy Running-ripe Ovulation ydration = Gravid Spawning phase Figure 8 Diagrammatic representation of the ovarian cycle of Atlantic croaker, Micropogonias undulatus (see text for details). The presence of small juveniles (<20 mm TL) in the York River from August/September through May/ June has prompted suggestions that north of Cape Hat- teras spawning of Atlantic croaker may also continue through spring (Haven, 1957; Chao and Musick, 1977). However, our results confirm previous reports (Wal- lace, 1940; Colton et al., 1979; Morse, 1980) that in the Chesapeake Bay and mid-Atlantic regions spawning is essentially completed by the end of December. Although Welsh and Breder ( 1923) suggested that spawning might take place in large estuaries such as the Delaware and Chesapeake bays, this study represents the first documented report of estuarine spawning for Atlantic croaker. Previous studies have consistently described Atlantic croaker as strict ma- rine spawners whose larval and juvenile stages mi- grate into estuarine nursery areas (Pearson, 1929; Hildebrand and Cable, 1930; Wallace. 1940; Haven, 682 Fishery Bulletin 92(4), 1994 1957; Warlen, 1982; Lewis and Judy, 1983; Setzler- Hamilton, 1987). However, the fact that during both years we found spawning-phase females (stages 3— 5) in Chesapeake Bay from July through October and that regressing and resting females — which probably had completed spawning for the season — were col- lected in the estuary indicates that the role of estu- aries as additional spawning areas for Atlantic croaker may be more important than previously thought. Other sciaenids that were believed to be strict marine spawners have also been reported to spawn occasionally in estuaries (Castello, 1985; Johnson and Funicelli, 1991). However, whether significant spawn- ing of Atlantic croaker occurs in Chesapeake Bay or other estuaries requires further investigation. The fact that spawning-phase Atlantic croaker have not previously been found in Chesapeake Bay can be attributed, at least in part, to their pattern of multiple spawning and indeterminate fecundity. Because in multiple spawning fishes the processes of hydration, ovulation, and spawning usually occur within a matter of hours (Hunter and Macewicz, 1985a; Brown-Peterson et al., 1988), the probability of collecting gravid or running-ripe females is much lower compared with other maturity stages. Addi- tionally, contrary to what happens with total spawn- ers, partially spent ovaries contain oocytes ranging from primary growth to advanced yolked stage, mak- ing the macroscopic identification of post-spawning fish very difficult (Hunter and Macewicz, 1985a). In most cases, we were not able to distinguish macro- scopically between fully developed and partially spent ovaries, and this also may have been a prob- lem with previous studies (e.g. Wallace, 1940). Diel periodicity of spawning could also influence the occurrence of hydrated females in samples from different gears. The thousands of adult Atlantic croaker examined by Haven (1957) and Wallace ( 1940) were collected primarily from Chesapeake Bay commercial pound nets and haul seines, which are usually fished in the predawn or early morning hours (Reid, 1955; Chittenden, 1991). During the rest of the day and through most of the night, fish remain alive in the pound-head or in the seine-bag until the nets can be fished (emptied), usually during slack water, and between 4:00 and 9:00 am. We hypoth- esize that during this period Atlantic croaker spawn within the nets at their usual spawning time of dusk (Holt et al., 1985). Females collected from these nets the following morning would probably show little or no signs of spawning and be identified as "develop- ing" (Wallace, 1940) or fully developed (this study). However, contrary to what happens with pound nets and haul seines, gill nets usually kill the fish within a short time after capture. Females undergoing hy- dration or ovulation, especially those caught a few hours before dusk, would die before they finished spawning, and the presence of hydrated oocytes in the ovaries could be recorded. This may explain why we observed hydrated or recently spent females only in gillnet col- lections. A similar pattern has also been observed for weakfish, Cynoscion regalis, which, like Atlantic croaker, spawn primarily between 6:00 and 9:00 pm.1 Size and age at maturity Our estimates of size and age at maturity are gener- ally below values previously reported for Atlantic croaker in the Chesapeake Bay and mid-Atlantic regions. Disagreement with previous reports can be attributed to three main factors: 1) failure of at least some studies (Wallace, 1940; Morse, 1980) to sample small, young fish from fishery-independent sampling programs; 2) the inclusion of samples collected from a period when resting (reproductively inactive) fish were present in the estimation of the proportion of mature fish by size or age; and 3) disagreement with previous estimates of age at maturity probably re- flects problems with age-determination methods pre- viously used for Atlantic croaker. Because of the dif- ficulty in distinguishing resting and immature go- nads, estimates based on samples pooled over the entire spawning season or during a period when rest- ing fish were present (e.g. Wallace, 1940; Morse, 1980) are probably biased towards larger sizes or older ages. Hunter et al. ( 1992 ) found that estimates of L50 for Dover sole were higher when females were taken during the spawning season than when they were sampled before spawning began. They sug- gested that estimates of length or age at first matu- rity should always be based on samples collected prior to the onset of spawning, when postspawning females with highly regressed ovaries are rare. However, for species like Atlantic croaker, which show individu- ally asynchronous gonadal maturation, sampling before the onset of spawning will not prevent the occurrence of prespawning, resting fish. To avoid this problem we used only fish collected in September, when no resting or developing stages occurred, to estimate size and age at first maturity. Finally, dis- agreement with previous estimates of age at matu- rity probably reflects problems with age-determina- tion methods previously used for Atlantic croaker. The use of length frequencies (Welsh and Breder, 1 Lowerre-Barbieri, S. K., M. E. Chittenden Jr., and L. R. Barbieri. 1993. The multiple spawning pattern of weakfish, Cynoscion regalis, in the Chesapeake Bay and mid-Atlantic Bight, with a discussion of annual fluctuations in reproductive output. Vir- ginia Institute of Marine Science, Gloucester Point, VA 23062. Unpubl. manuscr., 61 p. Barbieri et al.: Maturity, spawning, and ovarian cycle of Micropogonias undulatus 683 1923) requires considerable subjective interpretation given the extended spawning season of Atlantic croaker, the generally asymptotic growth after age 2, and the great overlap in observed sizes at age (Barbieri et al., 1994). Although Welsh and Breder (1923) and Wallace ( 1940) have also used scales, prob- lems in applying this method to Atlantic croaker have been reported (Joseph, 1972). Sex ratios Our results on temporal fluctuations in Atlantic croaker sex ratios agree well with previous reports for the Chesapeake Bay and mid-Atlantic regions (Welsh and Breder, 1923; Wallace, 1940). The pre- dominance of females during the first 3-4 months of spawning may indicate that either males start leav- ing the estuary earlier than females as fish migrate out of Chesapeake Bay or that spawning-phase fe- males are more susceptible to the fishing gears used in Chesapeake Bay (pound nets, haul seines, and gill nets). During both years, the frequency of males de- creased during the first two months of spawning and began increasing again in October/November when the first offshore trawl collections were obtained. Mark-recapture studies are necessary to better evalu- ate the migratory patterns of Atlantic croaker in Chesapeake Bay and the mid-Atlantic region. Atresia of advanced yolked oocytes High levels of atresia typically have been used to identify regressing ovaries, and for many teleosts, have been described as representing a key histologi- cal marker for the cessation of spawning (Hunter and Macewicz, 1985, a and b; Hunter et al., 1986; Dickerson et al., 1992). However, our results with Atlantic croaker indicate that high levels of atresia do not necessarily imply the end of spawning. Al- though we found significant atresia of cortical alveoli and partially yolked oocytes only in regressing ova- ries, indicating it could in fact be used to mark the end of spawning, major atresia of advanced yolked oocytes was observed in actively spawning females throughout the spawning season suggesting it may represent a normal part of the reproductive biology of Atlantic croaker. The fact that hydrated females — either actively spawning or just about to spawn — showed advanced yolked oocytes undergoing atresia suggests that a portion of these oocytes are never matured and spawned. In other words, it appears that a surplus production of advanced yolked oocytes is part of the reproductive strategy of Atlantic croaker. Fully developed females may hydrate and spawn more or less of these oocytes depending, for example, on en- vironmental conditions. Evidence from laboratory studies seems to support this hypothesis. Middaugh and Yoakum (1974) used chorionic gonadotropin to induce laboratory spawn- ing of Atlantic croaker. They found that although the abdomen of females became extremely distended as a result of oocyte hydration, only a limited number of eggs could be stripped from each fish. More re- cently, Trant and Thomas (1988) and Patino and Thomas (1990) evaluated in vitro germinal vesicle breakdown (GVBD, an index of final oocyte matura- tion) in laboratory-spawned Atlantic croaker. They reported that in this species there is always a re- sidual number of "advanced oocytes" which fail to complete GVBD or even enter the morphological maturation process, suggesting that not all oocytes in a spawning batch would be matured and spawned. Conclusion Because of the small number of gravid females col- lected and the fact that POF's could be identified only in running-ripe females, we were not able to esti- mate batch fecundity and spawning frequency for Atlantic croaker. However, our results have shown that 1 ) Atlantic croaker mature at a smaller size and earlier age than previously thought; 2) Atlantic croaker are capable of spawning in the estuary, al- though the magnitude of estuarine spawning is still unclear; 3) they are multiple spawners with inde- terminate fecundity, indicating that the only avail- able estimates of fecundity (Morse, 1980) — those based on the assumption of determinate fecundity — should not be used for management; and 4) the oo- cyte size-frequency method (MacGregor, 1957 ) should not be used to estimate batch fecundity for this spe- cies, because of the high levels of atresia of advanced yolked oocytes observed in spawning females. Future studies on the reproductive biology of Atlantic croaker in Chesapeake Bay and the mid-Atlantic region should concentrate on offshore, preferably fishery- independent, trawl collections to obtain gravid fe- males for batch fecundity estimates following the hydrated oocyte method (Hunter et al., 1985). Rates of deterioration and resorption of POF's must also be evaluated in laboratory-spawned fish to determine if the postovulatory follicle method (Hunter and Macewicz, 1985a) can be used to estimate spawning frequency for this species. Acknowledgments We would like to thank those Chesapeake Bay com- mercial fishermen and James Owens (VIMS) who 684 Fishery Bulletin 92(4). 1994 helped us obtain samples. We especially thank Sonny Williams for his extraordinary help with daily gillnet samples. Histology slides were prepared by Juanita Walker (VIMS). Beverly Macewicz and John Hunter (NMFS, Southwest Fisheries Science Center) helped with histological evaluation of oocyte atresia, identi- fication of postovulatory follicles, and with discus- sions about reproductive strategies in multiple spawning fishes. Louis Daniel, Ronald Hardy, Jack Musick, Mark Patterson, and two anonymous review- ers made helpful suggestions to improve the manu- script. Financial support was provided by the Col- lege of William and Mary, Virginia Institute of Ma- rine Science, and by a Wallop/Breaux Program Grant for Sport Fish Restoration from the U.S. Fish and Wildlife Service through the Virginia Marine Re- sources Commission, Project No. F-88-R3. Luiz R. Barbieri was partially supported by a scholarship from CNPq, Ministry of Science and Technology, Bra- zil (Process No. 20358 1/86-OC). Literature cited Barbieri, L. R., M. E. Chittenden Jr., and C. M. Jones. 1994. Age, growth, and mortality of Atlantic croaker, Micropogonias undulatus, in the Chesapeake Bay region, with a discussion of apparent geographic changes in population dynamics. Fish. Bull. 92:1-12. Brown-Peterson, N., P. Thomas, and C. R. Arnold. 1988. Reproductive biology of the spotted seatrout, Cynoscion nebulosus, in south Texas. Fish. Bull. 86:373-388. Bearden, C. M. 1964. Distribution and abundance of Atlantic croaker, Micropogon undulatus, in South Carolina. Contrib. Bears Bluff Lab., South Carolina 40:1-23. Castello, J. P. 1985. The ecology of consumers from dos Patos La- goon estuary, Brazil. In A. Yanez-Arancibia (ed. ), Fish community ecology in estuaries and coastal lagoons: towards an ecosystem integration, chap- ter 17, p. 383-406. DR(R) UNAM Press, Mexico. 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. 75:657-702. Chittenden, M. E., Jr. 1991. Operational procedures and sampling in the Chesapeake Bay pound-net fishery. Fisheries (Bethesda) 16:22-27. 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 ma- rine fishes, Cape Sable to Cape Hatteras. Fish. Bull. 76:911-915. DeMartini, E. E., and R. K. Fountain. 1981. Ovarian cycling frequency and batch fecun- dity in the queenfish, Seriphus politus: attributes representative of serial spawning fishes. Fish. Bull. 79:547-560. Dickerson, T. L., B. J. Macewicz, and J. R. Hunter. 1992. Spawning frequency and batch fecundity of chub mackerel, Scomber japonicus, during 1985. Calif. Coop. Oceanic Fish. Invest. Rep. 33:130-140. Haven, D. S. 1957. Distribution, growth, and availability of ju- venile croaker, Micropogon undulatus, in Virginia. Ecology 38:88-97. 1959. Migration of the croaker, Micropogon undu- latus. Copeia 1959:25-30. Hildebrand, S. F., and W. C. Schroeder. 1928. The fishes of Chesapeake Bay. Bull. U.S. Bur. Fish. 43:1-388. Hildebrand, S. F., and L. E. Cable. 1930. Development and life history of fourteen te- leostean fishes at Beaufort, N.C. Bull. U.S. Bur. Fish. 46:383-488. Holt, G. J., S. A. Holt, and C. R. Arnold. 1985. Diel periodicity of spawning in sciaenids. Mar. Ecol. Prog. Ser. 27:1-7. Horwood, J. W., and M. Greer Walker. 1990. Determinacy of fecundity in sole (Solea ) from the Bristol Channel. J. Mar. Biol. Assoc. U.K. 70:803-813. Hunter, J. R., and S. R. Goldberg. 1980. Spawning incidence and batch fecundity in northern anchovy, Engraulis mordax. Fish. Bull. 77:641-652. Hunter, J. R., and B. J. Macewicz. 1985a. Measurement of spawning frequency in multiple spawning fishes. In R. Lasker (ed.), An egg production method for estimating spawning biomass of pelagic fish: application to the northern anchovy, Engraulis mordax, p. 79-94. Dep. Commer., NOAATech. Rep. NMFS 36. 1985b. Rates of atresia in the ovary of captive and wild northern anchovy, Engraulis mordax. Fish. Bull. 83:119-136. Hunter, J. R., N. C. H. Lo, and R. J. H. Leong. 1985. Batch fecundity in multiple spawning fishes. In R. Lasker (ed.). An egg production method for estimating spawning biomass of pelagic fishes: application to the northern anchovy, Engraulis mordax, p. 67-77. Dep. Commer., NOAA Tech. Rep. NMFS 36. Hunter, J. R., B. J. Macewicz, and J. R. Sibert. 1986. The spawning frequency of skipjack tuna, Katsuwonus pelamis, from the South Pacific. Fish. Bull. 84:895-903. Hunter, J. R., B. J. Macewicz, N. C. H. Lo, and C. A. Kimbrell. 1992. Fecundity, spawning, and maturity of female Dover sole, Microstomus pacificus, with an evalu- ation of assumptions and precision. Fish. Bull. 90:101-128. Barbieri et al.: Maturity, spawning, and ovarian cycle of Micropogonias undulatus 685 Johnson, G. D. 1978. Micropogonias undulatus (Linnaeus), Atlan- tic croaker. Dep. Interior, U.S. Fish and Wildl. Serv. FWS/OBS-78/12:227-233. Johnson, D. R., and N. A. Funicelli. 1991. Spawning of the red drum in Mosquito La- goon, east-central Florida. Estuaries 14:74-79. Joseph, E. B. 1972. The status of the sciaenid stocks of the middle Atlantic coast. Chesapeake Sci. 13:87-100. Lewis, R. ML, and M. H. Judy. 1983. The occurrence of spot, Leiostomus xanthurus, and Atlantic croaker, Micropogonias undulatus, larvae in Onslow Bay and Newport River estuary. North Carolina. Fish. Bull. 81:405-412. Lowerre-Barbieri, S. K., and L. R. Barbieri. 1993. A new method of oocyte separation and pres- ervation for fish reproduction studies. Fish. Bull. 91:167-170. Macer, C. T. 1974. The reproductive biology of horsemackerel, Trachurus trachurus (L.), in the North Sea and English Channel. J. Fish Biol. 6:415-438. MacGregor, J. S. 1957. Fecundity of the Pacific sardine (Sardinops caerulea). Fish. Bull. 57:427-449. McHugh, J. L. 1981. Marine fisheries of Delaware. Fish. Bull. 79:575-599. Middaugh, D. P., and R. L. Yoakum. 1974. The use of chorionic gonadotropin to induce laboratory spawning of the Atlantic croaker, Micropogonias undulatus, with notes on subse- quent embryonic development. Chesapeake Sci. 15:110-114. Morse, W. W. 1980. Maturity, spawning and fecundity of Atlantic croaker, Micropogonias undulatus, occurring north of Cape Hatteras, North Carolina. Fish. Bull. 78:190-195. Norcross, B. L., and H. M. Austin. 1988. Middle Atlantic Bight meridional wind com- ponent effect on bottom water temperatures and spawning distribution of Atlantic croaker. Cont. Shelf Res. 8:69-88. Patiho, R., and P. Thomas. 1990. Induction of maturation of Atlantic croaker oocytes by 17a, 20(5, 21-trihydroxy-4-pregnen-3- one in vitro: consideration of some biological and experimental variables. J. Exp. Zool. 255:97-109. Pearson, J. C. 1929. Natural history and conservation of the red- fish and other commercial sciaenids on the Texas coast. Bull. Bur. Fish. 44:129-214. 1941. The young of some marine fishes taken in lower Chesapeake Bay, Virginia, with special ref- erence to the grey seatrout Cynoscion regalis (Bloch). Bull. U.S. Fish. Wildl. Serv. 50:79-102. Reid, G. K., Jr. 1955. The pound-net fishery in Virginia. Part 1: History, gear description, and catch. Comm. Fish. Rev. 17(5):1-15. Saila, S. B., C. W. Recksieck, and M. H. Prager. 1988. Basic fisheries science programs. Elsevier, New York, 230 p. SetzlerJJamilton, E. M. 1987. Utilization of Chesapeake Bay by early life history stages of fishes. In S. K. Majumdar, L. W. Hall, and H. M. Austin (eds.), Contaminant prob- lems and management of living Chesapeake Bay resources, p. 63-93. The Pennsylvania Academy of Science, Philadelphia, Pennsylvania. Trant, J. M., and P. Thomas. 1988. Structure-activity relationships of steroids in inducing germinal vesicle breakdown of Atlantic croaker oocytes in vitro. Gen. Comp. Endocrinol. 71:307-317. Wallace, D. H. 1940. Sexual development of the croaker, Micro- pogon undulatus, and distribution of the early stages in Chesapeake Bay. Trans. Am. Fish. Soc. 70:475-482. Wallace, R. A., and K. Selman. 1981. Cellular and dynamic aspects of oocyte growth in teleosts. Am. Zool. 21:325-343. Warlen, S. M. 1982. Age and growth of larvae and spawning time of Atlantic croaker in North Carolina. Proc. Annu. Conf., S.E. Assoc. Fish Wildl. Agencies 34:204-214. Welsh, W. W, and C. M. Breder. 1923. Contributions to life histories of Sciaenidae of the eastern United States coast. Bull. U.S. Bur. Fish. 39:141-201. Abstract. — Female cobia, Ra- chycentron canadum, were sampled on their spawning grounds in the northern Gulf of Mexico to study changes in proximate analysis (pro- tein, lipid, carbohydrate, and ash) of the ovaries during gonadal matu- ration. Four major stages of oocyte development were studied: stage 1, previtellogenesis; stage 2, vitello- genesis; stage 3, final maturation; and stage 4, postovulation. Cobia are multiple spawning fish; there- fore, ovaries engaged in a sequen- tial round of oogenesis were distin- guished as stages 1' and 2'. Protein was the major constituent of cobia ovaries and its contribution re- mained fairly constant (49-55% of the dry weight) throughout all stages of development. Lipid was the second most abundant compo- nent but the levels, ranging from 21 to 41%, changed depending on the stage of ovarian development. Lipid concentration increased from stage 1 through 3 and decreased slightly in stage 4; it was lower in stage- 1 than in stage- 1' ovaries but was the same in stages 2 and 2'. Carbohydrate was the least abun- dant component (3-4%) whereas ash ranked third (6-20%). Most cobia were in prespawning condi- tion ( stages 1-3 ) when they arrived in the northern Gulf of Mexico in April and May; some prespawning fish (stages 1 and 2) were also ob- served in August and September about a month or two before migra- tion to the overwintering grounds normally occurs. Cobia undergoing sequential spawning episodes (stages 1" and 2) were captured from April through August. Gonosomatic indices (GSI) were calculated both for ovarian devel- opmental stage and for month of capture. Mean GSI increased as ovarian development proceeded and decreased during postovula- tion; GSI for month of capture was highest during April and May when the prespawning fish first ap- peared in northern Gulf of Mexico waters. Biochemical and histological changes during ovarian development of cobia, Rachycentron canadum, from the northern Gulf of Mexico Patricia M. Biesiot Robert E. Cay lor Department of Biological Sciences, University of Southern Mississippi Hattiesburg, Mississippi 39406-5018 James S. Franks Gulf Coast Research Laboratory Ocean Springs. Mississippi 39566-7000 Manuscript accepted 9 May 1994. Fishery Bulletin 92:686-696 ( 1994). 686 Cobia, Rachycentron canadum, are large migratory fish with a world- wide distribution in tropical and subtropical seas, except for the Pa- cific coast of North America (Mig- dalski and Fichter, 1983). In the western Atlantic, cobia are found from Massachusetts and Bermuda to Argentina (Briggs, 1958) but are most common in the Gulf of Mexico (Migdalski and Fichter, 1983), rang- ing from Key West along the entire coast to Campeche, Mexico (Daw- son, 1971). Cobia support a popu- lar sport fishery wherever they are present. Total mortality rates for co- bia, including sport and commercial catches plus natural mortality, may be high (Richards, 1967) and it has been questioned whether cobia in the Gulf of Mexico are being exploited at rates beyond which maximum sus- tainable yields can be maintained. Cobia undergo extensive seasonal migrations (Fig. 1), moving from overwintering grounds to distant spawning/feeding grounds during the spring and summer (Briggs, 1958). They are usually absent from the U.S. fishery in more northerly latitudes during fall and winter months (Dawson, 1971 ) and are be- lieved to spend their winters near the Florida Keys.1 During the spring, co- bia move northwest into Gulf waters' or north along the eastern seaboard of the United States (Richards, 1977). Cobia usually enter north-central Gulf waters (Alabama and Missis- sippi) in March or April and begin the return to their wintering grounds in late September.1 Female cobia with ripe ovaries have been collected in the northern Gulf of Mexico from April to May through October.2 Spawning takes place throughout the summer with 1 Franks, J. S., J. T. McBee, and M. T. Allen. 1992. Studies on the seasonal movements and migratory patterns of the cobia, Rach- ycentron canadum, in Mississippi marine waters and adjacent Gulf waters. Gulf Coast Res. Lab., Ocean Springs, MS 39566- 7000. Interim Contract Rep. to Miss. Dept. Wildl., Fish, and Parks/Bur. Mar. Res. and U.S. Fish Wildl. Serv., Atlanta, GA 30303, Project No. F-91, 62 p. 2 Lotz, J. M., R. M. Overstreet, and J. S. Franks. 1991. Reproduction of cobia, Rachycentron canadum, from the north- eastern Gulf of Mexico. In J. S. Franks, T. D. Mcllwain, R. M. Overstreet, J. T. McBee, J. M. Lotz, and G. Meyer, Investigations of the cobia (Rachycentron canadum) in Mississippi marine waters and adjacent Gulf waters. Gulf Coast Res. Lab., Ocean Springs, MS 39564-7000. Final Rep. to Miss. Dept. Wildl., Fish, and Parks/Bur. Mar Res. and U.S. Fish Wildl. Serv, Atlanta, GA 30303, Project No. F-91, p. 2-1 to 2-42. Biesiot et al.: Ovarian development in Rachycentron canadum 687 PATTERNS OF SEASONAL MIGRATIONS presumed winter grounds spring migration (to spawning grounds) fall migration (to winter grounds) Figure 1 Patterns of seasonal migration of cobia, Rachycentron canadum, in the eastern Gulf of Mexico and U.S. Atlantic waters. Cobia are found west of the Mississippi River delta but little is known of their migratory pat- terns there (see Footnote 1 in text). Lines and arrows indicate presumed patterns of migration, not specific migratory routes. mature females releasing eggs at least once but pos- sibly twice or more during the breeding season; the population experiences a spawning peak during late spring or early summer. Fertilized cobia eggs are pe- lagic and egg diameter is between 1.16 and 1.42 mm (Joseph et al., 1964). Although histological changes during development of cobia ovaries have been described previously,2 nothing is known about biochemical changes in the ovary that occur during gonadal maturation. Data on the patterns of change in protein, lipid, and car- bohydrate during oocyte development, and on the subsequent utilization of these reserves by the em- bryos and larvae are important to understanding the early life history of cobia. There is, in fact, very little information about interactions among these three nutrient reserves relative to reproduction despite intensive work on the nutritional requirements of a wide range of fish species. Few studies have considered the relationship of the major biochemical components (protein, lipid, carbohydrate, and ash) throughout the course offish ovarian development. Dawson and Grimm (1980) showed that protein was higher and more constant than lipid during gonadal development of plaice, Pleuronectes platessa; ash was low and carbohydrate was not measured. Other authors have studied only the ripe (prespawning) stage of fish ovaries. Ripe mullet, Mugil cephalus (Lu et al., 1979), and Atlan- tic cod, Gadus morhua (Kjesbu et al. 1991), ovaries also had higher protein than lipid levels and had low ash and carbohydrate. On the other hand, lipid was the major component of ripe anabantid Trichogaster pectoralis ovaries (Hails, 1983). The present study addresses changes in biochemi- cal composition of the cobia ovary throughout the course of gonadal development. Total protein, lipid, carbohydrate, and ash were measured and compared among fish sampled on their spawning grounds in the northern Gulf of Mexico; the different stages of gonadal development were confirmed histologically. In addition, gonosomatic indices (GSI) were calcu- lated on the basis of ovarian developmental stage and month of capture. 688 Fishery Bulletin 92(4), 1994 Table 1 Stages of ovarian development in cobia, Raehycentron canadum. Stage Characteristics 1 Previtellogenesis 2 Vitellogenesis 3 Final maturation 4 Postovulation 1' Sequential previtellogenesis 2' Sequential vitellogenesis Germinal vesicle develops; evaginations appear in nuclear envelope; cortical alveoli form in ooplasm. Lipid vacuoles form; uneven dispersal of protein and lipid yolk. Clearing of lipid around periphery of oocytes; enlarged size; chromosomes condense. Oocytes become distorted and compacted; presence of postovulatory follicles; frothy residual lipid vacuoles. Sequential development of previtellogenic oocytes after a spawning episode; presence of postovulatory follicles and resorbing oocytes in addition to characteristics of stage 1. Sequential development of vitellogenic oocytes after a spawning episode; presence of postovulatory follicles and resorbing oocytes in addition to characteristics of stage 2. Materials and methods Sample collection Cobia examined in this study were collected from coastal waters of Florida, Alabama, Mississippi, Loui- siana, and Texas, mostly through fishing tourna- ments held along the northern Gulf Coast during April through September of 1991 and 1992, although a few fish were caught by project personnel during that same time period. Fish were stored on ice from the time of capture. Immediately after each fish was weighed and measured (total and fork lengths), the ovaries were removed, placed in plastic resealable bags, and stored on ice for 4 to 20 hours until gonad total weights could be recorded and aliquots of the tissue taken. Separate aliquots of each ovary sample were placed in 10% phosphate-buffered formalin and stored at room temperature until the tissues were processed for microscopic examination (see below). Additional aliquots of each ovary were stored at -80°C until the biochemical analyses (see below) were performed. Histology Ovaries were processed according to techniques modi- fied from Humason ( 1979). Tissues were dehydrated in ethyl alcohol and embedded in paraffin by means of a Histomatic automatic tissue processor. The em- bedded tissues were sectioned at 4 or 5 um. Sections were stained with Delafield's hematoxylin and eosin (95% ethyl alcohol) (Humason, 1979). Aspects of fish ovarian development as described by Blaxter (1969), Wallace and Selman (1981), Overstreet (1983, a and b), Guraya (1986), and Mommsen and Walsh ( 1988) were used to determine the stages of development in cobia ovaries. Four cat- egories of development were observed in this study: stage 1, previtellogenesis; stage 2, vitellogenesis; stage 3, final maturation, and stage 4, postovulation (Table 1). Some ovaries appeared to have entered another, sequential round of oocyte maturation. Be- cause we were interested in biochemical differences that might exist between successive clutches of oo- cytes, the following additional categories were stud- ied: stage 1', a sequential previtellogenesis, and stage 2', second (or sequential) vitellogenesis. Biochemistry The frozen tissues were thawed on ice and homog- enized with either a Virtis tissue homogenizer or a hand-held ground glass mortar and pestle. Protein was measured according to Hartree (1972) with bo- vine serum albumin as the standard. Carbohydrate was measured according to Dubois et al. ( 1956 ) with glucose as the standard. Dry weight was determined after drying the samples overnight at 80°C to con- stant weight. The same samples were then com- busted overnight at 500°C to determine ash content. Lipid extraction was performed according to Sasaki and Capuzzo ( 1984) which is a modification of Folch et al. (1957) and Bligh and Dyer (1959); total lipid was measured gravimetrically with a Cahn C-31 microbalance. Calculations and statistics A gonosomatic index (GSI) was calculated as GSI = ovary weigh t/( total fish weight - ovary weight) x 100 (DeVlamingetal., 1982). Nonparametric Kruskal-Wallis analysis of variance by ranks (Zar, 1984) was performed with the SPSS- X2.1 statistical software package in order to test the null hypothesis that there were no significant differ- ences among the means being compared. In cases Biesiot et al.: Ovarian development in Rachycentron canadum 689 where the null hypothesis was rejected (oc<0.05), nonparametric Tukey-type multiple comparisons were performed according to Zar (1984) in order to determine between which of the mean values signifi- cant differences occurred. Results Histology Histological analyses were performed on the gonads of 115 female cobia collected from the northern Gulf of Mexico over the course of two breeding seasons (n=42 in 1991 and n=73 in 1992). Of these fish, 14 were caught in Florida waters, 6 in Alabama, 60 in Mississippi, 26 in Louisiana, and 7 in Texas; loca- tion data could not be obtained for two fish, but they were probably caught in either Mississippi or Loui- siana waters. We observed, as had Lotz et al.,2 that cobia oocyte production appeared to be group syn- chronous as defined by Wallace and Selman (1981), such that each ovary examined contained oocytes at different stages of maturation. However, ovaries could be assigned to specific categories based on the dominant oocyte maturity stage. In stage-1 previtellogenesis, the oocytes were small, compact, and irregularly shaped (Fig. 2A). The previtellogenic stage comprised three substages: a) early previtellogenesis, characterized by small oo- cytes in which the nucleus had swollen to form a large germinal vesicle; b) middle previtellogenesis, char- acterized by nucleoli developing within the nucleus and causing evaginations to form in the nuclear en- velope; and c) late previtellogenesis, characterized by the presence of cortical alveoli. The latter substage marked the beginning of the transition to stage 2. In stage-2 vitellogenesis, the oocytes increased in size as the yolk material increased (Fig. 2B) and formed unevenly dispersed lipid vacuoles. Vitel- logenic oocytes were somewhat more rounded and were not as compacted as previtellogenic oocytes. During stage-3 final maturation, the oocytes were larger and the lipid vacuoles and proteinaceous yolk material had become more evenly dispersed (Fig. 2C). The lipid droplets fused and congregated around the periphery of the oocytes, resulting in a clearing of that region of the cell. Note that although most of the oocytes in Fig. 2C were stage-3 oocytes; some stage-1 oocytes and late stage-2 oocytes were also present. Chromosomes condensed during stage 3 for the initiation of meiosis (Fig. 2D). During stage-4 postovulation, unspent oocytes and postovulatory follicles (POF) were resorbed (Fig. 3A). The oocytes became distorted and compacted, as did the POF. Residual lipid vacuoles were observed in the resorbing oocytes. (A few early previtellogenic oocytes can also be seen in Figure 3A, concurrent with the resorption process. ) A sequential round of ovarian development was observed in some cobia ovaries categorized as stage 1' (Fig. 3B). The presence of resorbing oocytes and POF in ovaries suggested that a prior spawning epi- sode had recently occurred. Early previtellogenic oocytes, resorbing oocytes, and POF were not seen simultaneously in ovaries categorized as stage 1. A sequential vitellogenic stage, stage 2', characterized by oocytes with numerous small lipid vacuoles, was also observed (Fig. 3C). Previtellogenic and resorb- ing oocytes as well as resorbing POF were also present in the ovary during this stage. Timing of ovarian development The stages of cobia ovarian development were tabu- lated according to month of capture for 1991 and 1992 data combined (Table 2). In both April and May, 14- 15% of the ovaries were developing (stages 1 and 2), -60% of the ovaries were ripe and about to be spawned (stage 3), -20% were postspawning (stage 4), and 3—5% had already spawned but were prepar- ing for a sequential spawning episode (stages 1' and 2'). The similarity in ovarian developmental stages for these two months is not surprising because all the April fish were collected in the last week of the month whereas all the May fish were collected in the first week of that month. During July (first and sec- ond weeks of the month), again -15% of the ovaries were developing (stages 1 and 2) but only -30% were ripe (stage 3); 15% were postspawning (stage 4) and over 40% had already spawned at least once and were in the process of developing for a subsequent spawn- ing (stages 1' and 2'). Fewer numbers offish were collected in August ( last week of the month) and Sep- tember (first week of the month) but the predomi- nant stages of ovarian development were dramati- cally different from fish collected earlier in the sea- son. The majority of ovaries (over 80% ) were previtel- logenic or vitellogenic (stages 1, 2, and 2') whereas fewer than 20% were ripe (stage 3) or postspawn (stage 4); no stage-1' ovaries were seen. Gonosomatic index Cobia with ovaries in stages 1, 2, and 3 had increas- ing mean GSI's of 1.1 ± 0.6, 5.0 ± 2.2, and 5.4 ± 2.2, respectively (Fig. 4A). GSI declined to 3.5 ± 1.6 in cobia with stage-4 ovaries (postovulation); the lower GSI reflects the loss of oocytes to spawning. Almost all of the pairwise comparisons were significantly different (Tukey-type multiple comparison test, 690 Fishery Bulletin 92(4). 1994 a<0.05). The exceptions were that stage- 1 ovaries were not significantly different from stage- 1' or stage- 2' ovaries nor were stage- 1' ovaries significantly dif- ferent from stage-2' ovaries. Mean GSI of female cobia was low in both Janu- ary and March (1.1 ± 0.2 for each month) (Fig. 4B). The January fish were caught on the winter grounds in south Florida waters whereas the March fish were early arrivals in Mississippi waters. GSI increased to 5.5 ± 2.9 in April but declined slightly in May to 4.7 ± 2.0; the largest number of cobia enter Missis- sippi waters during these two months.1 Mean GSI continued to decline during July (2.9±1.9) and Au- gust(1.7±1.6). Slightly more than half of the pairwise comparisons showed significant differences (Tukey- type multiple comparison test, oc<0.05). The GSI of January and March fish was significantly different from April, May, and July fish; the GSI of April and May fish was significantly different from July, Au- gust, and September fish; and the July and August fish were significantly different from each other. B Jt¥ -"ASS Figure 2 (A) Stage-1 cobia, Rachycentron canadum, ovary, previtellogenesis. Early previtellogenesis (EP) with large germi- nal vesicle (GV) developing; middle previtellogenesis (MP) characterized by developing nucleoli (N) which cause evaginations to form in the nuclear envelope; late previtellogenesis (LP) characterized by appearance of lipid vacu- oles. (B) Stage-2 ovary, vitellogenesis. Vitellogenic oocytes (st 2) have increased in size and in number of lipid vacuoles. Note nonsynchronous formation of oocytes; early previtellogenic (EP) and middle previtellogenic (MP) stage-1 oocytes also occur. (C) Stage-3 ovary, final maturation. Oocytes (st 3) enter pre-ovulation stage and become more rounded. Lipid vacuoles concentrate around periphery and cause a clearing. Stage-1 (St 1) and stage-2 (St 2) oocytes are also present. (D) Detail of stage-3 ovary. Lipid vacuoles are aggregated around oocyte periphery. Chro- mosomes (C) condense for initiation of meiosis. Stage-1 (St 1 ) and late stage-2 (St 2) oocytes are also present. Scale bars=250 |im. Biesiot et al.: Ovarian development in Rachycentron canadum 691 Biochemistry Biochemical analyses were performed on about one third of the fish sampled for the histological study (rc=43). Protein was the major biochemical compo- nent (Fig. 5A), representing from 49 to 55% of the ovary total dry weight (507.5-550.5 ug/mg dry weight). There were no statistically significant dif- ferences in protein concentration among ovarian de- velopmental stages (Kruskal-Wallis, a>0.05). Lipid concentration ranged from 209.3 to 412.5 ug/ mg dry weight (21-41% dry weight) during ovarian development, increasing from stage 1 through stage 3 (Fig. 5B). The increase was likely due to the for- mation of lipid yolk during oocyte maturation. Lipid concentrations then decreased after ovulation but not to the low level of stage 1, probably reflecting the residual lipid that had not been resorbed during stage 4. The only statistically significant differ- ence in lipid concentration during the course of ovarian development was between stages 1 and 3 (Tukey-type multiple comparison test, a<0.05). Carbohydrate concentration was very low dur- ing all stages of oogenesis in cobia, ranging from 27.2 to 45.2 ug/mg dry weight (3^% dry weight) (Fig. 5C). It decreased from stage 1 to 2, increased from stages 3 through 1', and declined slightly in stage 2'. Almost all of the pairwise comparisons of carbohydrate concentration were significant (Tukey-type multiple comparison test, a<0.05) except that stage 1 was not significantly different from stages 4 and 2' nor was stage 2 significantly different from stage 3. Ash concentration decreased from a high of 196.3 ug/mg dry weight to a low of 55.3 ug/mg dry weight (6-20% dry weight) (Fig. 5D); it increased in stage 4 and stage 1' but declined again in stage 2'. Stage-1 ash concentration was significantly different from stages 3, 4, and 2'; whereas stage 2 was significantly different from stage 3 (Tukey- type multiple comparison test, a<0.05). All of the other pairwise comparisons were not significant. Discussion Protein was the major constituent of cobia ova- ries and its contribution remained fairly constant (49-55%) throughout all stages of gonadal devel- Figure 3 (A) Stage-4 ovary, postovulation. Resorption of unspent stage-3 oocytes (RO) into ovarian tissue (OT); oocytes are distorted and compacted. There is residual lipid in the resorbing oocytes. Stage-1' oocytes (St 1) occur. (B) Stage-1' ovary, second previtellogenesis. Early previ- tellogenic (EP), middle previtellogenic (MP), and late previtellogenic (LP) stage-1' oocytes develop. Resorb- ing oocyte (RO) and resorbing postovulatory follicle (POF) are remnants from stage 4. (C) Stage-2' ovary, second vitellogenesis (early). Formation of second round of vitellogenic oocytes (St 2'). Note resorbing oocytes (RO), postovulatory follicles (POF), and stage-1' oocytes (St 1'). Scale bars=250 urn. 692 Fishery Bulletin 92(4). 1994 Table 2 Percentage of cobia, Rachycenti "on canadum ovaries at each stage of development by month of capture. Stages are as described in Table 1. Data are com- bined from 1991 and 1992. Month of capture Apr May Jul Aug Sep Total Stage (n=21) (n=58) (n=27) (ra=6) (rc=6) (re=118) 1 5.2 3.7 50.0 66.7 9.3 2 14.3 10.3 11.1 16.7 16.7 11.9 3 61.9 60.3 29.6 — 16.7 48.3 4 19.0 20.7 14.8 16.7 — 17.8 1' 4.8 — 29.6 — — 7.6 2' — 3.4 11.1 16.7 — 5.1 opment. We believe that any putative increase in the proteinaceous yolk as oocytes ripened was not de- tectable by the methods used in this study because the follicles were also increasing in size as oocytes matured. That is, protein concentration was rela- tively stable because structural protein (follicles, etc.) contributed far more to the total protein concentra- tion than did yolk proteins. Lipid was the second most abundant component but the levels changed from stage to stage, ranging from 21 to 41%. The fluctuations in lipid concentra- tion during ovarian maturation can be explained by the increasing amount of lipid yolk reserves that are deposited as oocytes mature from stages 1 to 3 fol- lowed by the subsequent loss of ripe oocytes from the ovary after ovulation and spawning. Carbohydrate was the least abundant component (3^i7( ) of cobia ovaries and ash ranked third ( 6-20% ). Boulekbache (1981) noted that the enzymes of car- bohydrate metabolism increased in activity during oogenesis. Carbohydrate concentration, therefore, may be low due to constant catabolism. In the present study, it is not known whether carbohydrate was con- stantly being catabolized and replaced, or whether concentrations were low. In most fish, however, car- bohydrate is not readily available for use until after fertilization occurs (Boulekbache, 1981). The trend in ash concentrations was the inverse of lipid con- centrations; that is, ash concentration declined when the lipid concentration increased and vice versa. Results of biochemical analysis of ripe ovaries from similar studies using other species offish are given in Table 3. Protein was the major component of ripe ova- ries followed by lipid, ash, and carbohydrate for cobia, Rachycentron canadum (this study), striped mullet, Mugil cephalus (Lu et al., 1979), plaice, Pleuronectes platessa < Dawson and Grimm, 1980), and Atlantic cod, Gadus morhua (Kjesbu et al., 1991). The primary dif- o 2 o < r- o O z o u 4 V 2' DEVELOPMENTAL STAGE Jan Mar Apr May Jul Aug Sep MONTH Figure 4 (A) Gonosomatic index (mean ± standard deviation) of cobia, Rachycentron cana- dum, in relation to stage of ovarian devel- opment. (B) Gonosomatic index (mean ± standard deviation) of cobia in relation to month of capture. Numbers above error bars are sample sizes. ferences among the four species of fish were the rela- tive proportions of protein and lipid. Ripe cod ovaries had less than half the amount of lipid than either mul- let or cobia; cobia ovaries had -1.1 times more lipid than mullet ovaries. Only for the anabantid Tricho- gaster pectoralis (Hails, 1983) was lipid the major com- ponent of ripe ovaries. Since lipid is the most efficiently stored energy reserve, supplying 9.5 cal/mg, whereas protein lib- erates 5.7 cal/mg and carbohydrate 4.1 cal/mg (Crisp, 1984), one might expect fish eggs to have large amounts of lipid to supply the energy needed for growth and metabolism during embryogenesis and Biesiot et al.: Ovarian development in Rachycentron canadum 693 700 1 2 3 4 1'/ DEVELOPMENTAL STAGE S a E < □ >- r O et < 1 2 3 4 V 2' DEVELOPMENTAL STAGE 2 3 4 V 2' DEVELOPMENTAL STAGE 2 3 4 1' 2" DEVELOPMENTAL STAGE Figure 5 (A) Mean protein concentration (|ig protein/mg dry weight (DW) ± standard deviation) in developing cobia, Rachycentron canadum, ovaries. (B) Mean lipid concentration (|ig lipid/mg dry weight ± standard deviation). (C) Mean carbohydrate concentration (ug carbohydrate/mg dry weight ± standard deviation); note change in Y-axis scale. (D) Mean ash concentration (|ig ash/mg dry weight ± standard deviation); note change in Y-axis scale. l=stage-l previtellogenesis; 2=stage-2 vitellogenesis; 3=stage-3 final maturation; 4=stage-4 postovulation; l'=stage-l' sequential previtellogenesis; 2'=stage-2' sequential vitellogenesis. Numbers above error bars are sample sizes. subsequent early larval development before first- feeding. The lipid:protein (L:P) ratio of ripe cobia oocytes (not ovaries as reported in the present study) was 1: 0.7 (Caylor, 1992 ), which is similar to both striped bass, Morone saxatilis, eggs (1: 0.6) (Eldridge et al., 1982) and red drum, Sciaenops ocellatus, eggs (1:0.8) (Vetter et al., 1983). Winter flounder, Pseudopleuro- nectes americanus, eggs, however, had a much higher L:P ratio of 1:5.2 (Cetta and Capuzzo, 1982). One factor affecting the storage of biochemical com- ponents is egg size. Many marine fish eggs are rela- tively small and do not have large stores of energetic reserves; these small eggs usually hatch quickly. Co- bia eggs range from 1.16 to 1.42 mm in diameter (Jo- seph et al., 1964). Red drum eggs are 0.86-0.98 mm (Vetter et al., 1983) and striped bass oocytes are 3.3- 3.4 mm after hydration (Eldridge et al., 1981). Win- ter flounder eggs are the smallest, 0.74-0.85 mm di- ameter (Smigielski and Arnold, 1972), yet they are composed of about five times as much protein as lipid. Thus, generalizations about energy reserve storage cannot be made based solely on egg size. Winter flounder eggs are demersal whereas eggs of cobia, red drum, and striped bass are pelagic. De- 694 Fishery Bulletin 92|4). 1994 Table 3 Biochemical composition (% dry weight) of some fish ovaries including cobia, Rachycentron canadum. Species Protein Lipid Carbohydrate Ash Reference Rachycentron canadum (cobia) (ripe ovary) 50.7 41.1 2.7 5.5 present study Mugil cephalus (mullet) (ripe ovary) 59.3 36.0 — 4.7 Luetal., 1979' Pleuronectes platessa (plaice) (ripe ovary) 87.4 8.4 — 3.1 Dawson and Grimm, 19802 (spent ovary) 88.6 3.6 — 7.1 Dawson and Grimm, 19802 Trichogaster pectoralis (anabantid) (ripe ovary) 27.7 72.3 0.16 — Hails, 19831' Gadus morhua (Atlantic cod) (ripe ovary) 77.7 16.5 0.7 5.1 Kjesbu et al., 1991' ' Original data reported as percent wet weight; we converted to percent dry weight. 2 Original data reported as dry weight; we converted to percent dry weight mersal eggs tend to have more protein than lipid (Flachter and Pandian, 1968), which results in nega- tive buoyancy. This could account for the high pro- portion of protein in winter flounder eggs in contrast to the high proportion of lipid in cobia, striped bass, and red drum eggs. Another possible explanation for the two very different patterns of biochemical com- position is that cobia, striped bass, and red drum are warm-temperate species whereas winter flounder is a cold-water species. Cobia (Ditty and Shaw, 1992), striped bass (Harrell et al., 1990), and red drum (Vetter et al., 1983) have short incubation times: 24 hours at 29°C, 48 hours at 18°C, and 22 hours at 23°C, respectively. Winter flounder has a much longer in- cubation time, 11-20 days at 4-6°C (Cetta and Capuzzo, 1982). The difference in incubation times for different species is due in part to the effect of temperature on metabolic rate of the developing embryos. Catabolism of specific endogenous energy stores in fish eggs is known to be related to the tem- perature of incubation. Lipid tends to be consumed in higher quantities at higher temperatures but pro- tein consumption dominates at lower temperatures (Heming and Buddington, 1988). Therefore, it is not surprising to see different patterns of biochemical composition in light of the temperature history dur- ing early development of these different species. In addition to reporting the changes in biochemi- cal composition during cobia ovarian development, we also examined the cyclical variation in ovary size. This was done by means of the gonosomatic index ( GSI ), a commonly used ratio that normalizes gonad size among animals of different size classes in order to assess their reproductive state. The GSI was de- termined for each female cobia sampled in this study and compared both to stage of ovarian development and to month of capture. The majority of cobia landed in April and May had ovaries in stage-3 condition (-60%). This was reflected in the high mean GSI for those months. By July and August fewer cobia, -30% and 0%, respectively, had stage-3 ovaries; this was reflected by the declining mean GSI. In September the increase in cobia with ovaries in prespawning condition was indicated by the slight increase in GSI. It is not clear why there was a greater proportion of stage- 1 and stage-2 ovaries in August and Sep- tember. Possible explanations include 1 ) difficulty in distinguishing stage-2 ovaries from stage-2' ovaries; 2) presence of resident young, small fish that were immature at the beginning of the summer but which grew to maturity late in the season; or 3 ) an influx of older, late-arriving cobia from unknown areas. We believe that a combination of the first two explana- tions is most likely. Some of the late summer/early fall fish with ovaries classified as stage-2 fish may well have spawned a batch of eggs earlier in the sea- son and therefore were actually stage 2'. But after the POF and any unspent stage-3 oocytes are re- sorbed, it is not possible to distinguish between a stage-2 and stage-2' ovary. On the other hand, the late summer and early fall stage- 1 fish were small; fork length was 94.8 ± 5.3 cm and 102.8 ± 7.9 cm in August and September, respectively. Based on cobia growth equations,2 it is highly unlikely that these fish could have spawned the previous year, and they were probably too immature to have spawned ear- lier in the same year. It is not known whether these fish would have spawned in the fall or whether they would have overwintered without further ovarian development. Lotz et al.2 suggested that cobia spawn over some unspecified period of time during the May to Sep- Biesiot et al.: Ovarian development in Rachycentron canadum 695 tember season. Their conclusion was based on the observed nonsynchronous formation of oocytes in the ovaries, considered to be strong evidence of multiple spawning (Hunter et al., 1992). Nonsynchronous development of oocytes was also observed in the present study. Data from this study further suggest that cobia resorb unspent stage-3 oocytes after ovu- lation. This hypothesis is supported both by the bio- chemical data and the histological evidence of re- sidual lipid in stage-4 and stage- 1' ovaries. In summary, we have determined that lipid con- centration, but not protein concentration, changes during cobia ovarian development, presumably as lipid yolk reserves are deposited in the oocytes. Car- bohydrate and ash concentrations also varied dur- ing development, but they were only minor compo- nents of the system. Further research is needed on newly fertilized cobia eggs and developing embryos and larvae in order to answer questions about the patterns and rates of energy reserve utilization dur- ing embryogenesis and during larval development before first feeding in this species. Because cobia eggs and larvae are only rarely found in plankton collec- tions in the Gulf of Mexico, we have initiated studies on the spawning of ripe, field-caught cobia (Caylor et al., in press). Acknowledgments This work was supported in part by the Mississippi- Alabama Sea Grant Consortium (Grant No. NA16RGO155-01, Project No. R/LR-26 awarded to PMB and JSF) and by the U.S. Fish and Wildlife Service, Sportfish Restoration, Atlanta, GA, through the Mississippi Department of Wildlife, Fisheries and Parks/Bureau of Marine Resources (Project #F-91 awarded to JSF). We thank Adam W. Hrincevich for his help collecting samples and Joanne Lyczkowski- Shultz and Robin M. Overstreet for reviewing an earlier version of the manuscript. We also thank two anonymous reviewers and the scientific editor for their helpful comments. Bob Barber kindly provided some of the cobia samples from Texas. Literature cited Blaxter, J. H. S. 1969. Development: eggs and larvae. In W. S. Hoar and D. J. Randall (eds.), Fish physiology, Vol. Ill, p. 179-252. Academic Press, New York. Bligh, E. G., and W. J. Dyer. 1959. A rapid method of total lipid extraction and purification. Can. J. Biochem. Physiol. 37:911-917. Boulekbache, H. 1981. Energy metabolism in fish development. Am. Zool. 21:377-389. Briggs, J. C. 1958. A list of Florida fishes and their distribution. Bull. Fla. State Mus. Biol. Sci. 2:1-318. Caylor, R. E. 1992. Histological and biochemical aspects of the developing ovary of Rachycentron canadum (Lin- naeus). M.S. thesis, Univ. Southern Mississippi, Hattiesburg, MS, 85 p. Caylor, R. E., P. M. Biesiot, and J. S. Franks. In press. Culture of cobia Rachycentron canadum: cryopreservation of sperm and induced spawning. Aquaculture. Cetta, C. M., and J. M. Capuzzo. 1982. Physiological and biochemical aspects of em- bryonic and larval development of the winter floun- der Pseudopleuronectes americanus. Mar. Biol. 71:327-337. Crisp, D. J. 1984. Energy flow measurements. InN.A. Holme and A. D. Mclntyre (eds.), Methods for the study of marine benthos, p. 284-372. Blackwell Scientific Pubis., Oxford. Dawson, A. S., and A. S. Grimm. 1980. Quantitative seasonal changes in the protein, lipid and energy content of the carcass, ovaries and liver of adult female plaice, Pleuronectes platessa L. J. Fish. Biol. 16:493-504. Dawson, C. E. 1971. Occurrence and description of prejuvenile and early juvenile Gulf of Mexico cobia, Rachycentron canadum. Copeia 1971:65-71. DeVlaming, V., G. Grossman, and F. Chapman. 1982. On the use of the gonosomatic index. Comp. Biochem. Physiol. 73A:31-39. Ditty, J. G., and R. F. Shaw. 1992. Larval development, distribution, and ecology of cobia Rachycentron canadum (Family: Rachycen- tridae) in the northern Gulf of Mexico. Fish. Bull. 90:668-677. Dubois, M., K. A. Gilles, J. K. Hamilton, P. A. Rebers, and F. Smith. 1956. Calorimetric method for determination of sug- ars and related substances. Anal. Chem. 28: 350-356. Eldridge, M. B., J. Whipple, and D. Eng. 1981. Endogenous energy sources as factors affect- ing mortality and development in striped bass (Morone saxatilis) eggs and larvae. Rapp. P-V. Reun. Cons. Int. Explor. Mer 178:568-570. Eldridge, M. B., J. A. Whipple, and M. J. Bowers. 1982. Bioenergetics and growth of striped bass, Morone saxatilis, embryos and larvae. Fish. Bull. 80:461^173. Flachter, J., and T. S. Pandian. 1968. Rate and efficiency of yolk utilization in de- veloping eggs of the sole (Solea solea). Helgol. Wiss. Meeresunters. 18:53-60. 696 Fishery Bulletin 92(4). 1994 Folch, J., M. Lees, and G. H. Sloane Stanley. 1957. A simple method for the isolation and purifi- cation of total lipides (sic) from animal tissues. J. Biol. Chem. 226:497-509. Guraya, S. S. 1986. The cell and molecular biology of fish oogen- esis. Karger Press, New York, 223 p. Hails, A. J. 1983. Temporal changes in fat and protein levels in the tropical anabantid Trichogaster pectoralis (Regan). J. Fish. Biol. 22:203-213. Harrell, R. M., J. H. Kerby, and R. V. Minton, eds. 1990. Culture and propagation of striped bass and its hybrids. Striped Bass Committee, Southern Div., Am. Fish. Soc, Bethesda, MD. Hartree, E. F. 1972. Determination of protein: a modification of the Lowry method that gives a linear photometric response. Anal. Biochem. 48:422-427. Heming, T. A., and R. K. Buddington. 1 988. Yolk absorption in embryonic and larval fishes. In W. S. HoarandD. J. Randall (eds.), Fish physiology. Vol. XI, p. 407—446. Academic Press, Inc., New York. Hu mason, G. L. 1979. Animal tissue techniques. W.H. Freeman and Co., San Francisco, 661 p. Hunter, J. R., B. J. Macewicz, N. C. Lo, and C. A. Kimbrell. 1992. Fecundity, spawning, and maturity of female Dover sole Microstomas pacificus, with an evalua- tion of assumptions and precision. Fish. Bull. 90:101-128. Joseph, E. B., J. J. Norcross, and W. H. Massmann. 1964. Spawning of the cobia, Rachycentron eanadum, in the Chesapeake Bay area, with ob- servations of juvenile specimens. Chesapeake Sci. 5:67-71. Kjesbu, O. S., J. Klungsoyr, H. Kryvi, P. R. Witthames, and M. Greer Walker. 1991. Fecundity, atresia, and egg size of captive Atlantic cod iGadus morhua) in relation to proxi- mate body composition. Can. J. Fish. Aquat. Sci. 48:2333-2343. Lu, J. Y., Y. M. Ma, C. Williams, and R. A. Chung. 1979. Fatty and amino acid composition of salted mullet roe. J. Food Sci. 44:676-677. Migdalski, E. C, and G. S. Fichter. 1983. Family Rachycentridae: cobia. In The fresh and salt water fishes of the world, p. 225-226. Crown Publishers, Inc., New York. Mommsen, T. P., and P. J. Walsh. 1988. Vitellogenesis and oocyte assembly. In W. S. Hoar and D. J. Randall (eds.), Fish physiology, Vol. XI, p. 347-406. Academic Press, Inc., New York. Overstreet, R. 1983a. Aspects of the biology of the red drum, Sciae- nops ocellatus, in Mississippi. Gulf Res. Rep. Supplement 1:45-68. 1983b. Aspects of the biology of the spotted seatrout, Cynoscion nebulosus, in Mississippi. Gulf Res. Rep. Supplement 1:1—43. Richards, C. E. 1967. Age, growth and fecundity of the cobia, Rachycentron eanadum, from Chesapeake Bay and adjacent mid-Atlantic waters. Trans. Am. Fish. Soc. 96:343-350. 1977. Cobia (Rachycentron eanadum ) tagging within Chesapeake Bay and updating of growth equations. Chesapeake Sci. 18:310-311. Sasaki, G. C, and J. M. Capuzzo. 1984. Degradation of Artenua lipids under storage. Comp. Biochem. Physiol. 78B:525-531. Smigielski, A. S., and C. R. Arnold. 1972. Separating and incubating winter flounder eggs. Prog. Fish-Cult. 34:113. Vetter, R. D., R. E. Hodson, and C. Arnold. 1983. Energy metabolism in a rapidly developing marine fish egg, the red drum (Sciaenops ocellata i. Can. J. Fish. Aquat. Sci. 40:627-634. Wallace, R. A., and K. Selman. 1981. Cellular and dynamic aspects of oocyte growth in teleosts. Am. Zool. 21:325-343. Zar, J. H. 1984. Biostatistical analysis. Prentice-Hall, Inc., Englewood Cliffs, NJ, p. 138-145. Abstract. — Male weakfish, Cy- noscion regalis, were collected from the southwest portion of Delaware Bay from April through September in 1990 and 1991. Morphometric measurements of the sonic muscles, testis size (gonadosomatic index, or GSI), and plasma androgen concen- trations were recorded to obtain data on the seasonality of sonic muscle condition and its relation- ship with the timing of reproduc- tion in this population. The sonic muscles were bilaterally symmetri- cal and showed no significant sea- sonal differences in length or width across both collecting periods. Sonic muscle thickness did change sig- nificantly across both collecting periods and there was a threefold increase in sonic muscle mass dur- ing the course of each collecting period. GSI and levels of both plasma testosterone and 11- ketotestosterone also varied sig- nificantly across both collecting seasons. Changes in sonic muscle mass followed but lagged one to three weeks behind the rise and fall in plasma androgen levels. Perti- nent models of skeletal muscle hy- pertrophy and atrophy are dis- cussed as is the possibility that in- creased sonic muscle mass during the spawning season may increase the reproductive fitness of male weakfish. Seasonal cycles in the sonic muscles of the weakfish, Cynoscion regalis Martin A. Connaughton College of Marine Studies, University of Delaware 700 Pilottown Road. Lewes. Delaware 19958 Present address UT-Houston Medical School, Laboratory for Neuroendocrmology, PO. Box 20708, Houston, Texas 77225 Malcolm H. Taylor School of Life and Health Sciences and College of Marine Studies University of Delaware, Newark, Delaware 19716 Manuscript accepted 21 April 1994. Fishery Bulletin 92:697-703 ( 1994). Sound production is used by a num- ber of teleost species during aggres- sive or defensive behaviors ( Gray and Winn, 1961; Steinberg et al., 1965; Hawkins and Chapman, 1966; Horch and Salmon, 1973; Hawkins and Rasmussen, 1978), but the greatest volumes of sound produced by teleo- sts are associated with reproduction. A number of marine families includ- ing Batrachoididae (Gray and Winn, 1961; Fish, 1964), Blenniidae (Tavolga, 1958b), Gadidae (Hawkins and Rasmussen, 1978), Gobiidae (Tavolga, 1958a), Triglidae (Moulton, 1956), Sciaenidae (Hildebrand and Schroeder, 1928; Pearson, 1929; Burkenroad, 1931), Serranidae, and Scaridae (Lobel, 1992) produce sound during the spawning season. Cho- ruses of these sounds are generally limited to the season and geographic area in which the species in question spawns (Fish and Cummings, 1972; Fine, 1978; Takemura et al., 1978; Mok and Gilmore, 1983; Saucier and Baltz, 1993). Teleostean sound production is generally accomplished through one of three mechanisms: hydrody- namic sound production via move- ment through the water, stridula- tion of bony body parts, or the use of specialized drumming muscles. The latter two mechanisms are of- ten amplified by sympathetic vibra- tion of the swim bladder, especially in the case of the drumming or sonic muscles. The sonic muscles may be intrinsic or extrinsic to the swim bladder. Intrinsic sonic muscles originate and insert entirely on the swim bladder, and appear as a part of the swim bladder wall. Extrinsic sonic muscles, however, originate on the cranium, pectoral girdle, or lat- eral body wall musculature and in- sert on or near the swim bladder (Tavolga, 1964; Demski et al., 1973). Male sciaenids produce a 'drum- ming' sound through the use of sexually dimorphic, extrinsic sonic muscles. The Atlantic croaker, Micro- pogonias undulatus, is the only member of this family in which the sonic muscles are found in both the male and female (Smith, 1905; Tower, 1908; Fish and Mowbray, 1970; Hill et al., 1987). The drumming behav- iors of sciaenid species are primarily limited to the reproductive periods of these species (Fish and Cummings, 1972; Takemura et al., 1978; Mok and Gilmore, 1983; Saucier and Baltz, 1993). Male drumming in sciaenids is believed to play a role in the spawn- ing behavior of these species (Pearson, 1929; Guest and Lasswell, 1978; Thomas1). 1 D. L. Thomas. 1971. The early life history and ecology of six species of drum (Sciaenidae) in the lower Delaware River, a brackish tidal estuary. Ichthyological Associates, Delaware Progress Rep. 3 (Part III), 247 p. 697 698 Fishery Bulletin 92|4), 1994 The weakfish, Cynoscion regalis, is a sciaenid which spawns in bays and estuaries from North Caro- lina to Long Island, New York, during the spring and early summer (Welsh and Breder, 1923; Mercer, 1983). Merriner (1976) noted a change in the colora- tion of the sonic muscles in male weakfish that par- alleled the changes in testis condition during the course of the year. The purpose of the present study was to determine whether the condition of the sonic muscles of male weakfish changes seasonally. In particular, this study was designed to determine the extent of change, if any, in the morphometries of the sonic muscles over the course of the spawning season and to observe these variations in relation to changes in testis con- dition and plasma androgen levels. Materials and methods Sample collection Male weakfish were sampled near the mouth of the Delaware Bay (lat. 38°50.30'N, long. 07512. 92'W) and roughly 40 km north of this (39°11.98'N, 075°23.20'W). Field collections were made from May through September in 1990 and from April through September in 1991. Specimens were collected by means of anchored or drifting gill nets, hook and line, and otter trawl. Immediately after capture of the fish, blood samples were taken with heparinized syringes from the hemal canal, posterior to the anal fin. The blood was then placed in heparinized microcentrifuge tubes and stored on ice. Samples were centrifuged at 2,000 x g, and the supernatant was removed and frozen at -80°C for determination of plasma testosterone and 11-ketotestosterone levels via radioimmunoassay (RIA). In 1991, blood sampling was preceded by milt collection to determine the number of ripe specimens. Any drumming behavior was also noted. Autopsies of the specimens provided total length (TL), total weight (TW), testis weight, and morpho- metric measurements of the sonic muscles. Testis weight and total weight were used to calculate a gonadosomatic index (GSI=|total testis weight/total weightl x 100). Sonic muscle weight, width (anterior- posterior axis of the muscle), length (dorso-ventral axis of the muscle), and maximal thickness (cross- section of the muscle) were measured for both the right and left sonic muscles. Orientation of sonic muscle width and length measurements is based on the dorso-ventral orientation of the muscle fibers (Ono and Poss, 1982). The sonic muscle-somatic in- dex (SMSI) was calculated as SMSI = {total sonic muscle weight/total weight x 100), and the results were expressed as a percentage of TW. Indices for sonic muscle width (SMWI), length (SMLI) and thick- ness (SMTI) were calculated as the mean of the mea- surement for the right and left sonic muscles/total length x 100 and were expressed as a percentage of TL. The color of the sonic muscles was also noted. Radioimmunoassays Testosterone was measured by direct radioimmu- noassay. Five uL aliquots of serum were placed in 2- mL conical glass tubes (methanol rinsed) and diluted to a total volume of 50 uL with borate buffer. Di- luted samples were incubated at 60°C for one hour to dissociate the steroid from binding proteins. Stan- dard solutions were prepared by dissolving crystal- line testosterone ( Sigma Chemical Co., St. Louis, MO) in absolute ethanol. Working standards (5, 10, 25, 50, 100, 250, 500 pg testosterone 5 uL1) were pre- pared in ethanol, dried to zero volume at 45°C under vacuum, and reconstituted in 50 uL of borate buffer. Standards and sample tubes were incubated with 100 uL (approximately 4,000 cpm) of dilute trace ( 1,2,6,7" 3H testosterone, cat. #NET-370, New England Nuclear Corporation) and 100 uL of reconstituted antiserum (Wein Laboratories, Succasunna, NJ) overnight at room temperature. Total counts were estimated by using vials containing 100 uL of dilute trace and 100 uL of saturated ammonium sulfate. Triplicate standard and serum samples incubated without the addition of antiserum were used to cal- culate nonspecific binding. Bound steroids were pre- cipitated by adding 250 uL of saturated ammonium sulfate to each tube. The vials were centrifuged and 400 uL of the supernatant were removed and placed in counting vials along with 6 mL of scintillation cock- tail. All tubes were shaken for 25 minutes, allowed to sit for at least one hour, and counted for 3-10 min- utes in a liquid scintillation counter. Testosterone measurements were assumed to es- timate total testosterone levels, as the plasma was incubated at 60C for one hour to dissociate any bind- ing proteins in the plasma. The 11-ketotestosterone assay (Woods and Sullivan, 1993 ) measured only free, or unbound, steroid, as the protocol includes tripli- cate ethyl-ether extraction of unbound steroids from the plasma. Cross-reactivities of the testosterone antiserum were >50% for 5a-dihydrotestosterone and 5-1-test- osterone, approximately 18% and 12.5% for 5a- androsten-3p,17P-diol and 5-5-androsten-3p,17P-diol, respectively, and <5% for all other steroids tested (Wein Laboratories, Inc.). Tritiated 11-ketotestoster- one and 11-ketotestosterone antiserum were gifts Connaughton and Taylor: Seasonal cycles in the sonic muscles of Cynoscion regalis 699 from C. V. Sullivan (Dept. Zoology, North Carolina State University, Raleigh, North Carolina). This an- tiserum cross-reacted less than 2% with testoster- one (Hourigan et al., 1991). Extraction efficiencies for the 11-ketotestosterone assay, determined by extraction of samples spiked with a known amount of radiolabelled 11-ketotes- tosterone, were always greater than 85%. RIA par- allelism was determined by measurement of various equivalents of plasma from a single plasma pool. The results were parallel to the standard curve over a range of 1-20 pL plasma for the testosterone assay and 25-250 pL plasma for the 11-ketotestosterone assay (sample sizes for the assays were 5 and 50 pL, respectively). When a range of steroid concentrations was added to plasma pool samples of known hormone concentration, the quantities of spiked steroid recov- ered were not significantly different from the quanti- ties added for either assay. The intra- and inter-assay coefficients of variation were, respectively, 10.3% and 18.5% for the testosterone assay and 7.4% and 17.1% for the 11-ketotestosterone assay. Statistical analyses Specimens with a TL outside a 15-cm range around the sample mean for each year were not included in this analysis. This limited size range was used for two reasons: first, to alleviate the possible effects of allometric growth on the indices used to present the data; and second, to remain within the size range of two- and three-year-old weakfish used by Villoso ( 1989). The calculated indices, such as GSI, were used only to display the data. Statistical analyses on tes- tis weight and sonic muscle morphometric measure- ments were conducted by using analyses of covari- ance with TW or TL of the specimen considered as a covariate. Bilateral comparisons of sonic muscle morphometries were made by using a paired r-test. Statistical analyses of plasma androgen levels were accomplished by using a one-way analysis of vari- ance. The a level for these analyses was 0.05. Results A difference was noted between 1990 and 1991 sur- face water temperatures, which rose more rapidly in 1991 (Fig. 1). Surface temperatures reached 26°C by late May in 1991 but not until late June in 1990. The apparent result of these temperature differences was a two-week difference in the course of events across the collecting period so that 1991 trends be- gan earlier than those in 1990. 30- O 9 » 26- o c5 5 22- Q. E

a of TW in both collecting seasons. After the peak of the spawning season, GSI values decreased rapidly, reaching postspawning lows between 0. 1 and 0.2% of TW in the early fall (Fig. 3A). Plasma androgen levels also varied significantly during both collecting seasons. Total plasma test- osterone titers were somewhat lower in 1990, reach- ing a maximum of only 3.34 ng-mL-1, whereas 1991 values climbed to 5.5 ngmL1 (Fig. 3B). In both years levels decreased to postspawning values of between 1.2 and 1.6 ng-mL1. Unbound plasma 11-keto- testosterone levels followed a trend similar to that noted for total plasma testosterone during both col- lecting periods (Fig. 3C). Maximal levels of unbound 11-ketotestosterone reached 1.1 ng-mL-1 in both years and then decreased to less than 0.1 ng-mL"1 by the fall. Examination of weakfish in 1991 indicated that virtually all the specimens were capable of drum- ming when handled throughout the entire collecting period, regardless of changes in sonic muscle condi- tion. Specimens produced milt throughout most of the study period, although no milt could be obtained before mid-May and none was obtained after mid- August. Discussion The extreme seasonality of drumming activity in sciaenids (Fish and Cummings, 1972; Takemura et al., 1978; Mok and Gilmore, 1983; Saucier and Baltz, a E DO 6- B 5- 4 - 3- 2 ri ■■•$.—€•—-*.....__ 1 - o- i i i i i | i i | '"■ffl 1 Apr 26 May 26 Jun 25 Jul 25 Date Aug 24 Sep 23 Figure 3 (A) Gonadosomatic index (GSI); (B) plasma testoster- one concentration; (C) and unbound plasma 11-keto- testosterone concentration of the weakfish, Cyno- scion regalis, plotted across sampling date for the 1990 and 1991 collecting seasons. Points are means (n=3-10 fish) ± one standard error of the mean. 1993 ) suggests that the condition of the sonic muscles in these species may not remain constant through- out the year. The data presented here indicate that the condition of the sonic muscles of weakfish does change seasonally; there was an approximate three- fold difference in mass between the spawning and pre- or post-spawning periods. As the sonic muscle could not grow beyond its points of attachment (Tower, 1908; Ono and Poss, 1982), which define the length and width of the muscle, seasonal hypertro- phy was expressed as an increase in muscle thick- ness. Seasonal changes in sonic muscle condition have not been documented in other sciaenids; how- Connaughton and Taylor: Seasonal cycles in the sonic muscles of Cynoscion regalis 701 ever, a seasonal increase was noted in the sonic muscle mass of the male haddock, Melanogrammus aeglefinus (Templeman and Hodder, 1958). The sonic muscles of haddock are present in both sexes, but the seasonal increase in volume of the muscles was noted only in the males. Maximal levels of total plasma testosterone ob- served in this study ranged between 3.5 and 5.5 ng-mL-1. Peak testosterone levels of 2.4 ng-mL-1 were noted in the closely related spotted seatrout, Cynoscion nebulosus.2 Similarly, maximal levels of 11-ketotestosterone in the spotted seatrout fell be- tween 8 ng-mLr1 and 10 ng-mL-1. Unbound 11- ketotestosterone levels in the weakfish, presumably expressing only a fraction of the entire plasma pool of this steroid, were roughly one order of magnitude less than maximal levels in spotted seatrout. The similarity of the shapes of the androgen and SMSI curves suggests that plasma androgen levels may play a role in the seasonal cycling of the sonic muscle. Seasonal hypertrophy of the sonic muscles appears to be triggered by increasing plasma andro- gen levels in the spring. Similarly, the increased sonic muscle mass noted during the summer appears to be maintained by high plasma androgen titers. As androgen levels peaked and began to fall, sonic muscle mass continued to increase for a period of one to three weeks, then began to drop off as atro- phy directly followed peak mass. There was no pla- teau in plasma androgen levels, nor was one noted in the plot of changing SMSI. Fine and Pennypacker ( 1986) noted an increase in the mass and a darken- ing in the coloration of the sonic muscles of male and female toadfish after gonadectomy and administra- tion of either testosterone or 11-ketotestosterone. In- jection of testosterone in male anurans can initiate calling behaviors and has been shown to accentuate the sexual dimorphism of the calling apparatus (Obert, 1977; Sassoon and Kelley, 1986). In mammals, increased androgen levels can induce increased muscle protein synthesis and muscle gly- cogen storage, resulting in muscle hypertrophy (Lamb, 1975). Increasing the workload of a muscle can also result in hypertrophy of the muscle. Work- induced hypertrophy can occur in the absence of pi- tuitary growth hormones, insulin, or androgens. In- creased muscle mass in work-induced hypertrophy is the result of increased protein concentrations in 2 P. Thomas, N. J. Brown, and C. R. Arnold. 1982. Seasonal varia- tions of plasma androgens and gonad histology in male spotted seatrout, Cynoscion regalis (Family: Sciaenidae). In C. J. J Rich- ter and H. J. T. Goos (eds.), Proceedings of the international symposium on reproductive physiology of fish, p. 111. Centre for Agricultural Publication and Documentation, Wageningen, Netherlands. the tissue. Much of this new protein is myofibrillar and is believed to result in increased cross-sectional area of the muscle fiber (Goldberg et al., 1975). In- creases in muscle aerobic enzyme activities, mito- chondrial protein concentrations, myoglobin concen- trations, and muscle glycogen storage have been noted in exercise-induced hypertrophy in mammals (Holloszy, 1967; Edgerton et al., 1969; Barnard et al., 1970). It is possible that the hypertrophy experienced by the sonic muscles of weakfish in the spring may involve both of these pathways. Data presented here indicate that elevated plasma androgen levels may have played a role in the seasonal increase in mass noted in these muscles. Increasing androgen levels may play a direct anabolic role in muscle hypertro- phy, or they may cue work-induced hypertrophy by initiating drumming behaviors, or both. Field hydro- phone data from this population collected in 1992 (Connaughton and Taylor, in press) indicate that drumming activity begins approximately 4-6 weeks before maximal sonic muscle mass is reached. The decreasing mass of the sonic muscles of weak- fish in mid- to late-summer may be the result of de- creasing androgen levels and decreased workload. Field recordings of voluntary drumming indicated that this behavior ceased abruptly after the spawn- ing season (Connaughton and Taylor, in press). At- rophy caused by disuse in mammalian systems re- sults in a decrease in fiber cross-sectional area and muscle mass (Desplanches et al., 1987; Musacchia et al., 1988). The decreased use of the sonic muscles after the spawning season might result in atrophy and subsequent weight loss in the sonic muscles. Observations of specimens collected in 1991 sug- gested that while the sonic muscle condition declined throughout the summer and fall, the specimens were still capable of producing sound when handled. If the sonic muscles were capable of producing sound re- gardless of their condition, then the seasonal hyper- trophy of these muscles must play a role other than activation of the muscles. Muscle hypertrophy in mammals can result in more powerful muscle con- tractions by that muscle (Goldberg et al., 1975). An increase in the strength of the sonic muscle contrac- tion might increase the amplitude of the drumming call, allowing the male to be heard at greater dis- tances or at increased intensities at a given distance, or both. Also, potential increases in aerobic capacity and in concentration of mitochondria may increase the stamina of the sonic muscles, permitting calling bouts of longer duration. If male drumming plays a role in weakfish repro- ductive behavior, the condition of the sonic muscles may affect an individual's reproductive success. How- ever, maintenance of peak condition of this other- 702 Fishery Bulletin 92|4). 1994 wise unused muscle throughout the remainder of the year might consume energy that could otherwise be budgeted toward growth, foraging, or predator avoid- ance and thus increase the individual's chances of reproducing again. Seasonal hypertrophy and atro- phy of the sonic muscles ensure peak mass only at the appropriate time. This cycle is presumably driven by an indicator of the proximity of the spawning sea- son, such as day length and correlated temperature changes, operating through changes in plasma an- drogen levels. Acknowledgments This work was supported by the Wallop-Breaux Sport Fishing Act with funds administered through the Delaware Department of Natural Resources and Environmental Control. We would like to thank C. V. Sullivan and his students for the use of their labo- ratory and for their help in determining plasma 11- ketotestosterone levels. The knowledge and experi- ence of Donald Evans was crucial in the collection of specimens. The support of P. Marvil, E. Bjorkstedt, and V. Connaughton was invaluable and greatly ap- preciated. Literature cited Barnard, R. J., V. R. Edgerton, and J. B. Peter. 1970. Effect of exercise on skeletal muscle. I: Bio- chemical and histological properties. J. Appl. Physiol. 28:762-766. Burkenroad, M. D. 1931. Notes on the sound-producing marine fishes of Louisiana. Copeia 1931:20-28. Connaughton, M. A., and M. H. Taylor. In press. Seasonal and daily cycles in sound pro- duction associated with spawning in the weakfish, Cynoscion regalis. Env. Biol. Fish. Demski, L. S., J. W. Gerald, and A. N. Popper. 1973. Central and peripheral mechanisms of teleost sound production. Am. Zool. 13:1141-1167. Desplanches, D., M. H. Mayet, B. Sempore, and R. Flandrois. 1987. Structural and functional responses to pro- longed hindlimb suspension in rat muscle. J. Appl. Physiol. 63:558-563. Edgerton, V. R., L. Gerchman, and R. Carrow. 1969. Histochemical changes in rat skeletal muscle after exercise. Exp. Neurol. 24:110-123. Fine, M. L. 1978. Seasonal and geographical variation of the mating call of the oyster toadfish Opsanus tau. Oecologia 36:45-57. Fine, M. L., and K. R. Pennypacker. 1986. Hormonal basis for sexual dimorphism of the sound producing apparatus of the oyster toadfish. Exp. Neurol. 92:289-298. Fish, M. P. 1964. Biological sources of sustained ambient sea noise. In W. N. Tavolga (ed.), Marine bio-acoustics, Vol. 1, p. 175-194. Pergamon Press, New York. Fish, M. P., and W. H. Mowbray. 1970. Sounds of western North Atlantic fishes. The Johns Hopkins Press, Baltimore, 207 p. Fish, J. F., and W. C. Cummings. 1972. A 50-dB increase in sustained ambient noise from fish (Cynoscion xanthulus). J. Acoust. Soc. Am. 52:1266-1270. Goldberg, A. I., J. D. Etlinger, D. F. Goldspink, and C. Jablecki. 1975. Mechanism of work-induced hypertrophy of skeletal muscle. Med. Sci. Sport. 7:248-261. Gray, G.A., and H. E. Winn. 1961. Reproductive ecology and sound production of the toadfish, Opsanus tau. Ecology 42:274-282. Guest, W. ( '., and J. L. Lasswell. 1978. A note on courtship behavior and sound pro- duction of red drum. Copeia 1978:337-338. Hawkins, A. I )., and C. J. Chapman. 1966. Underwater sounds of the haddock, Melanogrammus aeglefinus. J. Mar. Biol. Assoc. U.K. 46:241-247. Hawkins, A. D., and K. J. Rasmussen. 1978. The calls of gadoid fish. J. Mar. Biol. Assoc. U.K. 58:891-911. Hildebrand, S. F., and W. C. Schroeder. 1928. Fishes of the Chesapeake Bay. Fish. Bull. 43:1-366. Hill, G. L., M. L. Fine, and J. A. Musick. 1987. Ontogeny of the sexually dimorphic sonic muscle in three sciaenid species. Copeia 1987: 708-713. Holloszy, J. O. 1967. Biochemical adaptations in muscle: effects of exercise on mitochondrial oxygen uptake and res- piratory enzyme activity in skeletal muscle. J. Biol. Chem. 242:2278-2282. Horch, K., and M. Salmon. 1973. Adaptations to the acoustic environment by the squirrelfishes, Myripristis violaceus and M. pralinius. Mar. Behav. Physiol. 2:121-139. Hourigan, T. F., M. Nakamura, Y. Nagahama, K. Yamauchi, and E. G. Grau. 1991. Histology, ultrastructure and in vitro steroido- genesis of the testes of two male phenotypes of the protogynous fish, Thalassoma duperrey ( Labridae ). Gen. Comp. Endocrinol. 83:193-217. Lamb, D. R. 1975. Androgens and exercise. Med. Sci. Sport. 7:1-5. Lobel, P. S. 1992. Sounds produced by spawning fishes. Env. Biol. Fish. 33:351-358. Connaughton and Taylor: Seasonal cycles in the sonic muscles of Cynoscion regalis 703 Mercer, L. P. 1983. A biological and fisheries profile of weakfish, Cynoscion regalis. N.C. Dep. Nat. Resour. Comm. Dev., Div. Mar. Fish. Spec. Sci. Rep. 39, 107 p. Merriner, J. V. 1976. Aspects of the reproductive biology of the weakfish, Qynoscion regalis (Sciaenidae), in North Carolina. Fish. Bull. 74:18-26. Mok, H. K., and R. G. Gilmore. 1983. Analysis of sound production in estuarine ag- gregations of Pogonias cromis, Bairdiella chrysoura, and Cynoscion nebulosus (Sciaenidae). Bull. Inst, of Zoology, Acad. Sin. 22: 157-186. Moulton, J. M. 1956. Influencing the calling of sea robins (Prionotus spp.) with sound. Biol. Bull. 111:393-398. Musacchia, X. J., J. M. Steffen, and R. D. Fell. 1988. Disuse atrophy of skeletal muscle: animal models. Exercise Sport Sci. Rev. 16:61-87. Obert, H.-J. 1977. Hormonal influences on calling and reproduc- tive behavior in anurans. In D. H. Taylor and S. I. Guttmann (eds. ), The reproductive biology of amphib- ians, p. 357-366. Plenum Press, New York. Ono, R. D., and S.G. Poss. 1982. Structure and innervation of the swim blad- der musculature in the weakfish, Cynoscion regalis (Teleostei: Sciaenidae). Can. J. Zool. 60:1955- 1967. Pearson, J. C. 1929. Natural history and conservation of redfish and other commercial sciaenids on the Texas coast. Fish. Bull. 44:129-214. Sassoon, I)., and D. B. Kelley. 1986. The sexually dimorphic larynx of Xenopus laevis: development and androgen regulation. Am. J. Anat. 177:457-472. Saucier, M. H., and D. M. Baltz. 1993. Spawning site selection by spotted seatrout, Cynoscion nebulosus, and black drum, Pogonias cromis, in Louisiana. Env. Biol. Fish. 36:257-272. Smith, H. M. 1905. The drumming of the drum-fishes (Sciaen- idae). Science 22:376-378. Steinberg, J. C, W. C. Cummings, B. D. Braby, and J. Y. MacBain (Spires). 1965. Further bio-acoustic studies.off the west coast of North Bimini, Bahamas. Bull. Mar. Sci. 15:942- 963. Takemura, A., T. Takita, and K. Mizue. 1978. Studies on the underwater sound — VII. Under- water calls of the Japanese marine drum fishes (Sciaenidae). Bull. Jpn. Soc. Sci. Fish. 44:121-125. Tavolga, W. N. 1958a. The significance of underwater sounds pro- duced by males of the gobiid fish, Bathygobius soporator. Physiol. Zool. 31:259-271. 1958b. Underwater sounds produced by males of the blenniid fish, Chasmodes bosquianus. Ecology 39:759-760. 1964. Sonic characteristics and mechanisms in ma- rine fishes. In W. N. Tavolga (ed. ), Marine bio-acous- tics, Vol. 1, p. 195-211. Pergamon Press, New York. Te m pi email, W., and V. M. Hodder. 1958. Variation with fish length, sex, stage of sexual maturity, and season in the appearance and vol- ume of the drumming muscles of the swim-blad- der in the haddock, Melanogrammus aeglefinus (L.). J. Fish. Res. Board Can. 15:355-390. Tower, R. W. 1908. The production of sound in the drumfishes, the sea-robin and the toadfish. Ann. NY Acad. Sci. 18:149-180. Villoso, E. P. 1989. Reproductive biology and environmental con- trol of spawning cycle of weakfish, Cynoscion regalis (Bloch and Schneider), in Delaware Bay. Ph.D. diss., Univ. Delaware, Newark, 295 p. Welsh, W. W., and C. M. Breder Jr. 1923. Contributions to life histories of Sciaenidae of the eastern United States coast. Fish. Bull. 39:141-201. Woods, L. C, III, and C. V. Sullivan. 1993. Reproduction of striped bass, Morone saxatilis ( Walbaum ), broodstock: monitoring maturation and hormonal induction of spawning. Aquat. Fish. Manage. 24:213-224. Abstract. — During 1987 and 1988, sea otter, Enhydra lutris, prey composition and foraging suc- cess were studied by observing for- aging otters in the northern Kodiak Archipelago. Study areas differed in the number of years in which they were occupied by sea otters and were categorized as estab- lished (occupied >25 years), inter- mediate (occupied 5-15 years), and frontal (occupied <5 years). Clams were the most frequently identified sea otter prey (57-67%) in all study areas, and of the clams identified to species, Saxidomus giganteus was the most frequently observed. Mussels, Mytilus spp., crabs (primar- ily Telmessus spp.), and green sea urchins, Strongylocentrotus droe- bachiensis, contributed <25% to the total prey within each study area. Adults did not differ in the propor- tion of clams, mussels, or crabs cap- tured as prey among study areas. Adults captured clams with a greater frequency and mussels with lesser frequency than did juvenile sea ot- ters for all study areas combined. Forage success did not differ among study areas for adults nor between adults and juveniles for all study areas combined. Adult sea otters in the established area appear to have compensated for reduced prey size by retrieving more prey items per dive; however, they obtained less clam bio- mass per dive than otters in the in- termediate and frontal areas. Sea otter, Enhydra lutris, prey composition and foraging success \n the northern Kodiak Archipelago Angela M. Doroff U.S. Fish and Wildlife Service, Alaska Fish and Wildlife Research Center 1011 East Tudor Road. Anchorage. Alaska 99503 Present address: U.S. Fish and Wildlife Service. Marine Mammals Management 1011 E Tudor Road. Anchorage, Alaska 99503 Anthony R. DeGange US Fish and Wildlife Service, Alaska Fish and Wildlife Research Center 1011 East Tudor Road, Anchorage, Alaska 99503 Present address U S Fish and Wildlife Service, Ecological Services 1011 East Tudor Road, Anchorage, Alaska 99503 Manuscript accepted .31 March 1994. Fishery Bulletin 92:704-710 ( 1994). 704 The Kodiak Archipelago in south- central Alaska (Fig. 1) supported an abundant sea otter, Enhydra lutris, population prior to their commercial exploitation during the 18th and 19th centuries (Lensink, 1962; Kenyon, 1969). Following this pe- riod of unregulated harvesting of sea otters, which was terminated in 1911 (Kenyon, 1969), an isolated remnant population of sea otters remained at the northern tip of Shuyak Island (Schneider1). During the late 1950s through mid 1980's, episodic range expansion occurred throughout the northern Kodiak Archipelago (Lensink, 1962; Schnei- der1; Simon-Jackson et al.23). In the absence of sea otters, dense populations of clams, crabs, sea ur- chins, and abalones may develop. As sea otters recolonize former habitat, shellfish densities decrease owing to sea otter predation, sometimes in combination with commercial and subsistence shellfish harvest (Garshelis et al., 1986). Sea otters have been implicated in closure of commercial and recreational fisher- ies in California for abalone, Hal- iotis spp. (Estes and VanBlaricom, 1985) and pismo clams, Tiuela stul- torum, (Stephenson, 1977; Miller et al.4). In Alaska, sea otters impacted the recreational and commercial fisheries for Dungeness crab, Can- cer magister, in Prince William Sound (Garshelis, 1983; Garshelis etal., 1986; Kimker5). During 1987-1988 the sea otter range continued to expand near southeastern Afognak Island of the Kodiak Archipelago. The natural recolonization pattern of the archi- pelago provided an opportunity to study the effects of sea otters on Schneider, K. B. 1976. Assessment of the distribution and abundance of sea otters along the Kenai Peninsula. Kamishak Bay and the Kodiak Archipelago. U.S. Dep. Commer., NOAA, OCSEAP Final Rep. 37:527-626. '■ Simon-Jackson, T., D. Taylor, S. Schliebe and M. Vivion. 1985. Sea otter survey, Kodiak Island-1984. U.S. Fish and Wild- life Service, Anchorage, Alaska. Unpubl. rep., 16 p. ; Simon-Jackson, T., M. Vivion, and D. Zwiefelfofer. 1986. Sea otter survey, Kodiak Island-1985. U.S. Fish and Wild- life Service, Anchorage, Alaska. Unpubl. rep., 11 p. Miller, D. J., J. E. Hardwick, and W. A. Dahlstrom. 1975. Pismo clams and sea ot- ters. Calif. Dep Fish and Game. Mar. Re- sources Tech. Rep. 31:1-49. Kimker, A. 1985. A recent history of the Orca Inlet, Prince William Sound Dunge- ness crab fishery with specific reference to sea otter predation. In B. R. Metleffled.), Symposium on Dungeness crab biology and management, p. 231-241. Univ. of Alaska, Alaska Sea Grant Rep. 85-3. Doroff and DeGange: Prey composition and foraging success of Enhydra lutris 705 prey populations (Kvitek et al., 1992) and an opportunity to assess changes in sea otter foraging characteristics (prey composition, forage success, prey size and biomass) as they relate to the duration the habitat had been occupied. We describe the foraging characteristics of sea otters in relation to the length of habitat occupancy along the Kodiak Ar- chipelago. Methods Study area Study areas in the Kodiak Archipelago were chosen in regions that differed in the number of years since sea otters had reoccupied the habitat (Fig. 1). We cat- egorized the areas following Kvitek et al. (1992) as established (occupied for >25 years), intermediate (occupied for 5-15 years), and frontal (occupied for <5 years) based on sea otter surveys (Lensink, 1962; Kenyon, 1969; Schnei- der1; Simon-Jackson23; and interviews with local inhabitants). Established study sites were on southern Shuyak and northern Afognak islands, interme- diate study sites were located between southern Afognak and northern Kodiak islands, and frontal study sites were southeast of Afognak and Raspberry is- lands. Study sites had broad expanses of shallow water (<20 m) with prima- rily sand and gravel sediments support- ing infaunal bivalve assemblages (Kvitek et al., 1992). Foraging observations Observations of foraging sea otters were made from shore with the aid of lOx binoculars and 40-80x tele- scopes (Questar Corp., New Hope, PA). Foraging data were collected by focal animal sampling (Altmann, 1974). Repeated dives were recorded for a focal ani- mal while the animal remained in view and contin- ued to forage (Calkins, 1978). All observations were made on unmarked animals that were within ap- proximately 1 km of shore. Data were collected dur- ing June— October 1987 and during March, June, and September of 1988 during daylight hours and dur- ing various tidal states. Data for each recorded dive included sex and age class of otter, presence of a pup, number of prey items ESTABLISHED INTERMEDIATE Figure 1 Study areas for observations of foraging sea otters, Enhydra lutris, in established (occupied >25 yr), intermediate (occupied 5-15 yr) and frontal (occupied <5 yr) areas during 1987 and 1988 in the Kodiak Archipelago, Alaska. obtained, identification of prey (classified to lowest possible taxon), and categorization of prey size (small <5 cm, medium 5-9 cm, and large >9 cm). Size class of prey was estimated relative to the mean forepaw width ( 4.5 cm ) and mean skull width ( 10 cm) for adult sea otters (Johnson6). Adult otters were classified as male, female, female with pup, or unknown sex. Ju- veniles that were estimated to be <2 years of age were differentiated from adults by their small body size (estimated to be <18 kg) and dark pelage. For- age data on pups still associated with their mother were not collected. Forage dives were classified as successful (prey captured), unsuccessful (no prey captured), or of unknown success (observer could not determine if prey were captured). Johnson, A. M. 1987. Sea otters of Prince William Sound, Alaska. U.S. Fish and Wildlife Service, Alaska Fish and Wildlife Re- search Center, Anchorage, Alaska. Unpubl. rep., 87 p. 706 Fishery Bulletin 92(4). 1994 Data partitioning A forage record was defined as the forage data spe- cific to a focal animal and was used as the sample unit in comparisons of prey composition, forage suc- cess, and the mean number of prey captured per dive. For assessing variation in prey composition and for- age success, only forage records containing >10 for- age dives were used; adults of unknown sex were de- leted in comparisons of sex classes. Sample sizes for juveniles were small and created an unbalanced sample design in 2-way comparisons. Consequently, separate tests were conducted to assess age-class differences. For comparisons of prey composition, we calculated the proportion of dives resulting in the capture of clams, crabs, and mussels for each forage record. Differences in the proportion of prey items captured by adult sea otters were tested among areas. Sample sizes were insufficient to test prey composition dif- ferences among areas for juveniles. Data were pooled from all study areas and the proportion of prey cap- tured was tested by age class. Forage success (the proportion of successful dives) was normalized by an arcsine transformation of the square root. Differences in forage success among study areas and among adult sex classes (male, fe- male, and females with pups) were tested. Sample sizes were insufficient to test for differences among study areas for juveniles. Data were pooled for all juveniles and all adults to test age differences in forage success. Number of prey items captured per dive was cal- culated by dividing the total number of prey captured by the number of forage dives per foraging record and averaging these values by sex class and area. Dives resulting in the capture of mussels (which may be difficult to count) and dives of unknown result were excluded. We assumed mean shell lengths of 4.0, 7.0, and 10.0 cm were representative of small, medium, and large bivalve size classes, then estimated mean wet-tissue mass of Saxidomus giganteus by using the weight-length relationships generated by Kvitek et al. ( 1992). We estimated caloric gain per dive by using caloric values for this genus reported byKenyon(1969). Data analysis Kruskal-Wallis nonparametric ( 1-way) tests were used to assess differences in the proportion of clams, mussels, and crabs captured among study areas by adult sea otters; data were pooled for all study areas and the proportion of clams, mussels, and crabs were tested by age class. Analysis of variance (2-way AN OVA) was used to test 1) dif- ferences in forage success among study areas and adult sex classes, and 2) differences in the mean number of prey captured per forage dive among study areas and adult sex classes. A 1-way ANOVA was used to test differences in the mean number of prey captured per dive among study areas for juvenile sea otters. A Student's t-test was used to test differences in forage success between adult and juvenile sea ot- ters for all study areas combined. For all compari- sons, significance was set at a=0.05. Results Sea otters were observed foraging on clams (57-67%), mussels (19-25%), crabs (2-A%) and green sea ur- chins, Strongylocentrotus droebachiensis (0-3%) (Fig. 2). Clams were identified to species in 23% («=535), 65% (rc=957), and 63% (« = 1,060) of the observations in established, intermediate, and frontal areas, re- spectively. The majority of clams identified were Saxidomus in established (98%), intermediate (89%), and frontal (96%) areas. Other clams identified (<10% per study area) were Tresus capax, Mya spp., Protothaca staminea, and Entodesma macroschisma. Mytilus spp. was the most common mussel observed within the study areas. Crabs were primarily Tel- messus spp.; however, a small number of Cancer mag- ister, were recorded. Other prey which contributed from <1 to 7% of the diet in each study area included Clinocardium spp., Cucumaria fallax, Echiurus echiurus alaskensis, Niicella spp., Octopus spp., Pisa- ster spp., Pycnopodia helianthoides, barnacle (class Crustacea), chiton (class Polyplacophora), tunicate ■ Established (N = 798) □ Intermediate (N = 1694) □ Frontal (N = 1852) M r~r~i JL b£dl Clam Mussel Crab Sea Urchin Other Unidentified Figure 2 Frequency of occurrence of food items obtained by sea ot- ters, Enhydra lutris, as determined by visual observation along the Kodiak Archipelago during 1987 and 1988 in areas of established (>25 yr), intermediate (5-15 yr), and frontal (<5 yr) sea otter forage areas. Doroff and DeGange: Prey composition and foraging success of Enhydra lutris 707 (class Ascidiacea), and kelp (primarily kelp hold-fasts with small unidentified invertebrates attached). Uni- dentified prey constituted 4—6% of prey per area. The proportion of forage dives resulting in the cap- ture of clams, mussels, and crabs did not differ among study areas for adults. For all study areas combined, adult and juvenile sea otters differed in the propor- tion of forage dives capturing clams (x2=13.35, df=l, P<0.001) and mussels (x2=10.40, df=l, P=0.001) but not crabs {%2=3.22, df=l, P=0.07). The median pro- portion of dives resulting in the capture of clams ranged among study areas from 0.62 to 0.85 for adults and from 0.00 to 0.52 for juveniles. Conversely, me- dian values for mussels ranged from 0.00 to 0.93 for juveniles and was zero for adults. Crabs were cap- tured infrequently and the median proportion of dives capturing crabs was zero for both age classes. Forage success did not differ among study areas (F=0.52, df=2, P=0.59) nor among sex classes (P=2.22, df=2, P=0.12) within areas for adults; the interac- tion between sex class and area was not significant (P=0.50, df=4, P=0.74). Mean forage success for all study areas combined was 89% for adults and 90% for juveniles and did not differ significantly (<=-0.59, df=107,P=0.56) (Table 1). Mean number of prey captured per dive by adults in established, intermediate, and frontal areas dif- fered among areas (l.&tl.O, 1.1±0.4, and 1.2±0.8, respectively) (P=3.88, df=2, P=0.02) but not among sex class (P=0.98, df=2, P=0.38); the interaction be- tween sex class and area was not significant (P=1.00, df=4, P=0.41). Juvenile sea otters, did not differ in the mean number of prey captured per dive among study areas (P=0.55, df=2, P=0.59) (Table 1). In the established area, 92% (rc=526) of the clams captured by sea otters were small (<5 cm), and 8% were medium (5-9 cm). In intermediate and frontal areas, however, only 27% (n=943) and 38% («=1,039) of all clams captured were small and the majority were medium sized. The mean caloric content of Saxidomus captured by adult otters per forage dive in established, intermediate, and frontal areas was estimated to be 10 kcal, 21 kcal, and 21 kcal, respec- tively (Table 2). Discussion The composition of the diet was similar for sea ot- ters in the Kodiak Archipelago among forage areas Table 1 Summary of foraging success and mean number of prey items per dive for juvenile and adult sea otters, Enhydra lutris, along the Kodiak Archipelago in established (occupied >25 years), intermediate (occupied 5-15 years), and frontal (occupi ed <5 years) areas. No. of Mean % Mean no. forage No. of successful prey items Study area Age and sex records dives dives per dive ' + SD Established Juvenile 3 30 83 1.0 ±0.5 Adult male 6 59 83 1.3 ±0.8 Adult female " 4 63 96 2.1 ± 1.4 Adult female w/pup 12 136 97 1.8 ±0.9 Adult unknown 9 93 97 1.4 ±0.3 Total 34 381 Intermediate Juvenile L6 223 93 1.0 ± 0.6 Adult male 19 239 89 1.1 ± 0.6 Adult female •1 7 343 78 1.0 ±0.4 Adult female w/pup 28 349 93 1.2 ± 0.4 Adult unknown 8 92 86 1.1 ± 0.4 Total 98 1,246 Frontal Juvenile 13 146 88 1.2 ± 0.6 Adult male 25 296 96 1.4 ± 1.3 Adult female 24 369 si; 1.1 ±0.4 Adult female w/pup 11 272 96 1.3 ±0.7 Adult unknown 4 69 St 1.0 ±0.03 Total 80 1,152 ' Dives resulting n the capture of mussels. Mytilus spp. and dives of unknown result were not used in calculating mear number of prey per dive. 708 Fishery Bulletin 92(4), 1994 Table 2 Frequency and estimated biomass of Saxidomus giganteus retrieved per dive by adult sea otters, Enhydra lutris, in established (occupied for >25 years), intermediate (occupied 5-15 years), and frontal (occupied <5 years) study areas along the Kodiak Archipelago, Alaska, 1987-1988. Study area Size class (mm) Proportion in sample Mean number of prey/dive Estimated wet-tissue weight obtained/dive (g) ' Estimated caloric content (kcal) Established Intermediate Frontal <50 50-90 <50 50-90 >90 <50 50-90 >90 0.83 0.17 0.18 0.71 0.11 0.28 0.62 0.10 I 6 1.1 1.2 16 33 33 10 21 21 Wet-tissue weight=2.14 (10-4 ((shell length I" 7H:r-=0. 86 for Saxidomus giganteus where shell lengths equal 40, 70, and 100 mm representing small, medium, and large size classes, respectively (Kvitek et al.. 1992). irrespective of the number of years the habitat had been occupied by sea otters. Clams, particularly Saxidomus, were the predominant prey identified in all study areas, although 35-77% of the clams were not identified to species. Green sea urchins were absent in the diets of sea otters in established areas but were found, infrequently, in the prey composi- tion in intermediate and frontal areas. Sea urchins were apparently locally abundant in intermediate and frontal areas prior to the initiation of our study (Kvitek et al., 1992; Stanford and Cunningham7). Sea urchin abundance had been reduced to low levels by sea otter predation in other regions of Alaska and in California (Lowry and Pearse, 1973; Estes et al., 1978; Laur et al., 1988; Kvitek et al., 1989) and it is likely that sea otter predation affected urchin popu- lations in the Kodiak Archipelago. Juvenile sea otter diets contained a higher propor- tion of mussels than that of adults. A higher occurrence of mussels in the diet of juveniles than of adult sea otters has also been demonstrated by other studies con- ducted in Alaska (VanBlaricom, 1988; Doroff and Bod- kin, in press; Johnson6). Mussels are an easily obtain- able intertidal prey, and young sea otters may rely on mussels as a food source until they become more profi- cient foragers (Estes et al., 1981; VanBlaricom, 1988). Sea otters at Kodiak were highly successful in se- curing prey, even where prey had been reduced by years of otter predation (Kvitek et al., 1992). There- fore, forage success was not a useful criterion for dis- criminating among study areas that varied in the Stanford. S., and W. Cunningham. Bare Island. Port Bailey, AK 99615. Personal commun., June 1987. duration of sea otter occupancy. For sea otters, for- age success may vary with prey type, hunting tac- tics, or locality (Ostfeld, 1991) and may not be re- lated to prey abundance or biomass (Estes et al., 1981). Ostfeld (1991) suggested, however, that for- age success is a useful means of comparing forage strategies and habitat characteristics for sea otters. The lack of variation in forage success among our study areas may have resulted, in part, from simi- larities in habitat (Kvitek et al., 1992). Kruuk et al. (1990) recommended caution in defining and using the concept of forage success on a per dive basis and suggested that a more meaningful approach would be to examine the biomass captured per unit of effort. We estimated the average biomass and subsequent caloric value captured on a per dive basis for sea ot- ters. Sea otters foraging in habitat occupied an esti- mated 1-15 years obtained approximately twice the biomass of otters foraging in habitat occupied >25 years. This suggests that sea otters foraging in long- occupied habitat may need to compensate for reduced prey size and abundance through increased alloca- tion of time for foraging to meet minimum daily ca- loric requirements (Costa, 1978; Estes et al., 1982; Estes et al., 1986; Garshelis et al., 1986). Biomass and caloric values were similar for intermediate and frontal areas. Possible explanations for the lack of disparity between intermediate and frontal areas are 1) preexisting habitat differences among study ar- eas, 2) resilience of Saxidomus to sea otter preda- tion over the short term (see Kvitek et al., 1988), or 3) an error in the classification of study areas. We made the assumption that observed differences in foraging characteristics resulted primarily from Doroff and DeGange: Prey composition and foraging success of Enhydra lutris 709 sea otter predation. There were likely preexisting differences in the community structure among our study areas that were not assessed, such as the dis- tribution and abundance of bivalve species prior to sea otters re-occupying the study areas. However, we believe that comparisons of study areas are valid given the similarities in habitat and infaunal inver- tebrate assemblages among study areas documented by Kvitek et al. (1992). Saxidomus may appear resilient to sea otter pre- dation pressure over the short term because it is present in high densities in our study areas (Kvitek et al., 1992). Saxidomus was found in higher densities than was any other forage species and it was selected preferentially (based on differences between in situ population of clams and the shells discarded by forag- ing otters) in intermediate and frontal areas (Kvitek et al., 1992). Saxidomus was also the most abundant clam (in situ) in the established area; however, Protothaca was selected preferentially (Kvitek et al., 1992). Protothaca was not identified visually as sea otter prey in the established area; however, only 239c of the clams could be identified to species. We believe the classification of our study areas and those used by Kvitek et al. ( 1992) were correct; how- ever, our methods lacked the refinement needed to distinguish between intermediate and frontal areas. Kvitek et al. ( 1992) was also unable to detect differ- ences between the intermediate and frontal areas by measuring prey size directly from the shells of clams consumed by sea otters. However, there were differ- ences in the size of the in situ population of clams between areas (Kvitek et al., 1992). Newly exploited habitat in our study was represented by an area es- timated to have been occupied 1-4 years by sea ot- ters. Rapid changes may occur within the first year that sea otters occupy unexploited habitat. Garshelis et al. (1986) observed an approximate twofold de- crease in kcal/dive in areas occupied by sea otters <1 year compared with areas occupied 1-2 years. Co- incident with the change in kcal/dive was a shift in prey from crabs to clams between areas studied by Garshelis et al. (1986). In the Kodiak Archipelago, we did not observe differences in mean kcal/dive or changes in prey composition between intermediate and frontal areas. Changes in prey composition, such as the potential removal of green sea urchins from the study area, may have occurred in the frontal area during the first year and were undetected. Adult sea otters in the established area appear to have compensated for reduced prey size by retriev- ing more prey items per dive. However, they still obtained less clam biomass (and subsequently less caloric intake) per dive than otters in the intermedi- ate and frontal areas, suggesting that they may need to forage longer to meet minimum daily caloric needs. Interestingly, juveniles in established areas did not appear to compensate for reduced bivalve prey size by increasing the number of prey captured per dive. Ju- veniles may be less efficient foragers and may compen- sate by increasing their consumption of Mytilus spp., which are an easily obtainable intertidal prey (Estes, 1981; VanBlaricom, 1988; Doroff and Bodkin, in press). Acknowledgments We thank Larry Barnes, John Baird, Elizabeth Belan- toni, Walt Cunningham, Dave Douglas, Dan Monson, Jay Nelson, Doug Sheperd, Annetta Smith, Susan Stanford, and Shelli Vacca for assistance with data collection. Walt Cunningham and Susan Stanford graciously hosted us at their home on Bare Island. Logistic support was provided by Jay Bellinger and staff of the Kodiak National Wildlife Refuge and Al Bayer and the crew of the MV Tiglax. Shelli Vacca and especially Bob Stehn provided computer support. For reviewing drafts of this manuscript, we thank Brenda Ballachey, Edward Bowlby, James Bodkin, Dan Esler, Dave Garshelis, Rikk Kvitek, and three anonymous reviewers. Literature cited Alt matin. J. 1974. Observational study of behavior: sampling methods. Behavior 49:227-267. Calkins, D. G. 1978. Feeding behavior and major prey species of the sea otter, Enhydra lutris, in Montague Strait, Prince William Sound, Alaska. Fish. Bull. 76:125- 131. Costa, D. P. 1978. The ecological energetics, water, and electro- lyte balance of the California sea otter, Enhydra lu- tris. Ph.D. diss., Univ. California, Santa Cruz, 75 p. Doroff, A. M., and J. L. Bodkin. In press. Sea otter foraging behavior and hydro- carbon levels in prey. In T. R. Loughlin (ed.), Im- pacts of the Exxon Valdez oil spill on marine mammals. Academic Press. Estes, J. A., and G. R. VanBlaricom. 1985. Sea otters and shellfisheries. In R. Beverton, J. Beddington, and D. Lavigne (eds.), Conflicts be- tween marine mammals and fisheries, p. 187- 235. Allen and Unwin, London, England. Estes, J. A., N. S. Smith, and J. F. Palmisano. 1978. Sea otter predation and community organi- zation in the western Aleutian Islands, Alaska. Ecology 59:822-833. 710 Fishery Bulletin 92(4), 1994 Estes, J. A., R. J. Jameson, and A. M. Johnson. 1981. Food selection and some foraging tactics of sea otter. In J. A. Chapman and D. Pursley (eds.), The worldwide furbearer conference proceedings, p. 606- 641. Worldwide FurbearersConf., Inc., Frostburg, MD. Estes, J. A., R. J. Jameson, and E. B. Rhode. 1982. Activity and prey selection in the sea otter: influence of population status on community structure. Am. Nat. 120:242-258. Estes, J. A. , K. Underwood, and M. Karmann. 1986. Activity time budgets of sea otters in California. J. Wildl. Manage. 50:626-639. Garshelis, D. L. 1983. Ecology of sea otters in Prince William Sound, Alaska. Ph.D. diss., Univ. Minnesota, MN, 321 p. Garshelis, D. L., J. A. Garshelis, and A. T. Kimker. 1986. Sea otter time budgets and prey relationships in Alaska. J. Wildl. Manage. 50(4):637-647. Kenyon, K. W. 1969. The sea otter in the eastern Pacific Ocean. North Am. Fauna. 68:1-352. Kruuk, H., D. Wansink, and A. Moorhouse. 1990. Feeding patches and diving success of otters, Lutra lutra, in Shetland. Oikos 57:68-72. Kvitek, R. G., A. K. Fukuyama, B. S. Anderson, and B. K. Grimm. 1988. Sea otter foraging on deep-burrowing bivalves in a California coastal lagoon. Mar. Biol. 98:157-167. Kvitek, R. G., D. Shull, D. Canestro, E. C. Bowlby, and B. L. Trout man. 1989. Sea otters and benthic prey communities in Washington State. Mar. Mamm. Sci. 5(3):226-280. Kvitek, R. G., J. S. Oliver, A. R. DeGange, and B. S. Anderson. 1992. Changes in Alaskan soft-bottom prey commu- nities along a gradient in sea otter predation. Ecology 73(2):413-428. Laur, D. R., A. W. Ebeling, and D. A. Coon. 1988. Effects of sea otter foraging on subtidal reef communities off central California. In G. R. VanBlaricom and J. A. Estes (eds.), The commu- nity ecology of sea otters, p. 159-168. Springer- Verlag, Berlin, West Germany. Lensink, C. J. 1962. The history and status of sea otters in Alaska. Ph.D. diss., Purdue Univ., West LaFayette, IN, 188 p. Lowry, L. F., and J. S. Pearse. 1973. Abalones and sea urchins in an area inhab- ited by sea otters. Mar. Biol. 23:213-219. Ostfeld, R. S. 1991. Measuring diving success of otters. Oikos 60:258-260. Stephenson, M. D. 1977. Sea otter predation an Pismo clams in Mon- terey Bay. Calif. Fish Game 63:117-120. VanBlaricom, G. R. 1988. Effects of foraging by sea otters on mussel- dominated intertidal communities. In G. R. Van- Blaricom and J. A. Estes (eds.), The community ecology of sea otters, p. 48-91. Springer- Verlag, Berlin, West Germany. Abstract. — We developed a relatively simple and parsimonious (SMPAR) biomass dynamics model for estimating abundance of north- ern anchovy, Engraulis mordax, off southern California and Baja Cali- fornia, Mexico, during the 1963 to 1991 fishing seasons. The SMPAR model was a compromise between simple surplus production and complex age-structured models. It was designed to give more precise biomass estimates for management of northern anchovy for which there are no age-composition data and only noisy abundance index data. We evaluated consistent bias in biomass and recruitment esti- mates, bias in recruitment esti- mates due to log transformation, and retrospective bias. Simple cor- rections based on bootstrap proce- dures were used to remove consis- tent bias and log transformation bias. Retrospective bias was not a significant problem. Results indi- cate that the SMPAR model esti- mates stock biomass more reliably than recruitment because abun- dance indices for northern anchovy contain little information about interannual recruitment variabil- ity. Asymptotic variance estimates calculated by inverting the Hessian matrix averaged 20% smaller than variances calculated by bootstrap- ping. Outliers in abundance data were the biggest source of uncer- tainty in biomass estimates. Simu- lation results indicate that our ap- proach could be useful in a variety of situations. A biomass-based assessment model for northern anchovy, Engraulis mordax Larry D. Jacobson Nancy C. H. Lo Southwest Fisheries Science Center, National Marine Fisheries Service. NOAA RO. Box 271, La Jolla, California 92038 J. Thomas Barnes California Department of Fish and Game 330 Golden Shore, Suite 50, Long Beach, California 90802 Manuscript accepted 28 April 1994. Fishery Bulletin 92:711-724. Northern anchovy, Engraulis mor- dax, is a small (<18 cm TL), short lived (<8 years) pelagic schooling fish (Baxter, 1967). The central stock of northern anchovy extends from Mexico to central California (lat. 30°-35°N); most of the stock inhabits the Southern California Bight. Spawning occurs all year with a peak between February and April (MacCall and Prager, 1988). The central stock of northern an- chovy is among the world's most thoroughly studied fish stocks. Re- liable estimates of northern an- chovy biomass were not available, however, until the daily egg produc- tion method was used to estimate spawning biomass from 1980 to 1985 (Lasker, 1985). Estimates of long-term trends in biomass were not available until the stock synthe- sis model for northern anchovy was developed (Methot, 1989). The stock synthesis model was used to man- age northern anchovy until 1992 after availability of age-composition data declined.1 As data became limited, variance and bias of biomass estimates2 from the stock synthesis model in- creased. Bias problems included a positive "retrospective" bias in re- cent estimates and a smaller but consistent positive bias in estimates for earlier seasons (Lo et al., 1992). Retrospective bias is a newly rec- ognized but common problem in fish stock assessment work (Sinclair et al., 1991) that makes recent biom- ass estimates too large. Consistent bias (usually positive) is a problem of variable severity in biomass esti- mates from most assessment mod- els including the stock synthesis model (Lo et al., 1992; Bence et al., 1993), derivatives3 of CAGEAN (Der- iso et al., 1985), and virtual pop- ulation analysis or VPA(e.g. Lapointe etal., 1989). Consistent bias, unlike retrospective bias, affects all or most of the biomass estimates from a model. We hypothesized that problems in the stock synthesis model for north- 1 Jacobson, L. D., and N. C. H. Lo. 1992. Spawning biomass of the northern anchovy in 1992. U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Southwest Fish. Sci. Cent., P.O. Box 271, La Jolla, CA 92038. Admin. Rep. LJ-92-24, 71 p. 2 Jacobson, L. D., and N. C. H. Lo. 1991. Spawning biomass of the northern anchovy in 1991. U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Southwest Fish. Sci. Cent, P.O. Box 271, La Jolla, CA 92038. Admin. Rep. LJ-91-19, 53 p. :' Deriso, R. 1993. A report on integrated stock assessment of Pacific sardine, Appen- dix 2. In F. J. Hester, Project report on Pacific sardine {Sardinops sagax ) resource research, 1991/1992 phase III. Living Ma- rine Resources Inc., 11855 Sorrento Val- ley Road, Suite A, San Diego, CA 92121. Final Rep. to Calif. Seafood Council, P.O. Box 91540, Santa Barbara, CA 93190, 118 p. 71 1 712 Fishery Bulletin 92(4), 1994 ern anchovy were exacerbated by use of a compli- cated model with insufficient data. The model we de- veloped for northern anchovy was, therefore, simpler and more parsimonious (SMPAR). The SMPAR model is a biomass dynamic model designed to give more precise biomass estimates for management of northern anchovy. SMPAR is a hy- brid between simple surplus production and complex age-structured models. It resembles a surplus pro- duction model because age-composition data are not used and fishing mortality rates are equal for all age groups. The model is age structured, however, and some rudimentary relationships between age-specific abundance and abundance indices are assumed. As described above, the SMPAR model for north- ern anchovy did not use any age-composition data, although data were available for most fishing sea- sons prior to 1991. We chose to exclude age-composi- tion data from our model because the data are diffi- cult to interpret, require complex modeling ap- proaches, and were not available for recent seasons. In this paper, we describe the SMPAR model and data for northern anchovy. Bias and variance in bio- mass and recruitment estimates are assessed by us- ing bootstrap techniques. Sensitivity analyses show how model assumptions and contradictory trends in the data affect biomass estimates from SMPAR. We use simulation analysis to show how SMPAR would perform under a wide range of fishing mortality and recruitment conditions. Data and methods Fishing seasons were used to aggregate most of our data including landings data and indices of abun- dance (Table 1; Fig. 1). Fishing seasons for northern Table 1 Data for northern anchovy, Eng raulit mordax , by fi shing season (1 July to 30 June): fish spotter data (SPOTTER), historical egg production data (HEP) new egg production index (EPI), sonar data (SONAR), an d daily egg produc- tion method data (DEP) CVder otes coefficient of variation Temperatures are average sea surface temperatures at Scripps Pier, San Diego California during January and February. HEP EPI SPOTTER (eggs (eggs Mexican US Total Fishing I short tons- CV 0.05 m2- CV 0.05 m2- CV DEP CV landings landings landings Temperature season block'1) (%) day"1) {%) day1 1 (Si SONAR (10:itl r;i (103t) (103t) (103t) i C) 1963 0.9 28 4.1 65 0.000 1.795 1.795 15.0 1964 10.5 31 4.0 29 0.000 2.324 2.324 13.3 1965 8.2 32 5 3 3 1 0.000 18.958 18.958 13.8 1966 11.2 30 0.000 42.725 42.725 14.0 1967 6.7 31 0.000 13.470 13.470 14.5 1968 3.9 35 MS 28 0.44 0.000 .1.122 1 33.224 14.3 1969 18.6 :-;ii 0.28 0.000 83.391 83.391 13.6 1970 10 2 32 0.23 0.000 81.854 81.854 13.1 1971 13.8 30 1 7 18 0.82 0.000 55.624 55.624 12.8 1972 10.3 33 1.67 0.000 76.059 76.059 15.0 1973 73.6 29 llll.-, 0.000 116.666 116.666 13.1 1974 42.9 29 19.7 53 3.09 28.088 113.782 141.870 13.2 1975 60.2 2S 35.287 135.573 170.860 13.9 1976 34.7 31 1.98 108.962 104.095 213.057 16.0 1977 33.4 29 2.3 192 0.39 127.229 76.236 203.465 15.7 1978 49.5 34 5.4 48 0.29 195.675 55.966 251.641 13.9 1979 77.2 34 2 7 17 2.2 27 0.60 870 26 157.543 40.091 197.634 15.0 1980 42.0 38 t 1 48 3.0 1 1 0.57 635 22 287.547 65.906 353.453 14.6 1981 22.0 :;s 3.3 11 1 'i 17 0.25 415 26 255.086 53.212 308.298 14.1 1982 27.3 39 3.9 :in 2 2 34 0.53 652 21 156.725 11.003 167.728 15.9 1983 7.3 19 2.9 37 r, s 11 0.57 309 17 66.260 7.507 73.767 15.0 1984 25.7 15 2.6 26 6.9 IS 1.02 521 19 123.359 4.762 128.121 13.8 1985 20.2 12 7 1 32 85.801 6.321 92.122 15.2 1986 8.6 13 5.0 16 116.334 4.783 121.117 15.0 1987 5.7 u; 6.3 33 98.498 5.794 104.292 14.0 1988 30.5 10 1 5 29 86.361 5.795 92.156 13.0 1989 9.7 45 2.3 35 55.647 8.228 63.875 14.4 1990 5.0 55 1 s 23 0.796 10.328 11.124 14.9 1991 12.1 54 119 32 0.134 4.546 4.680 15.3 Jacobson et al.: A biomass-based assessment model for Engraulis mordax 713 1000 T 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 FISHING SEASON Figure 1 Indices of abundance for northern anchovy, Engraulis mordax (log scale). The SPOTTER index is based on fish spotter data, HEP is the historical egg production index, EPI is the new egg production index, SONAR is based on sonar data, and DEP is daily egg production method estimates of spawn- ing biomass. anchovy begin on 1 July, end on 30 June, and are identified by the calendar year on 1 July. We assumed that indices for northern anchovy measured abun- dance during the peak spawning period which is about 15 February. Indices of abundance used to estimate northern anchovy biomass included 1) spawning biomass es- timated by the daily egg production method (DEP, Gunderson, 1993; formerly called the egg production method or EPM, Lasker, 1985), 2) a historical egg production index (HEP, Lo, 1985), 3) our new egg production index (EPI), 4) relative biomass of schooled fish estimated from fish spotter data (SPOT- TER, Lo et al., 1992), and 5) relative biomass of schooled fish estimated from California Department of Fish and Game sonar data (SONAR, Mais, 1974; Methot, 1989). Indices that measured relative abun- dance of northern anchovy (Fig. 1) were not all posi- tively correlated and none of the correlations were statistically significant. Lack of significant correla- tion may have been due to relatively few years of data used to compute some correlations, to impre- cise indices, and to differences among indices in area surveyed. Of particular concern is the lack of signifi- cant positive correlation between the EPI and SPOT- TER indices which are the only relative abundance data available for recent fishing seasons. Daily egg production method (DEP) estimates of spawning biomass during the 1979 to 1984 seasons (Table 1, Fig. 1) measured spawning biomass in met- ric tons (t), rather than in relative units, and are relatively precise (coefficients of variation (CV) less than 27%). Only six DEP observations were available, but the DEP data were important because they helped scale indices of relative abundance for northern anchovy to units of absolute biomass (Bence et al., 1993). SONAR data did not include variances that were required in our model. We used the standard devia- tion (e=0.439) of log-scale residuals for SONAR val- ues from a previous study2 and the relationship CV 1 (1) to obtain a crude estimate (46%) of the arithmetic scale coefficient of variation. The SPOTTER index used in this study (Table 1, Fig. 1) was similar to the one developed by Lo et al. ( 1992 ) except that April to March, rather than Janu- ary to December, annual periods were used to aggre- gate data. The SPOTTER index value for the 1963 fishing season was anomalously low ( Fig. 1 ). Follow- ing Lo et al. ( 1992), we excluded the 1963 value be- cause the data collection program was new in 1963 and the information may not have been reliable. New egg production index (EPI) Our new egg production index (EPI) measures egg production by northern anchovy during the 1979 to 714 Fishery Bulletin 92(4), 1994 1991 fishing seasons in the reduced CalCOFI (Cali- fornia Cooperative Oceanic Fishery Investigation) sampling area surveyed since 1985 (Hewitt, 1988). The HEP index, in contrast, is based on a relatively large grid of CalCOFI sampling stations occupied prior to the 1985 season. As explained below, the EPI makes use of all egg and larva data and is more pre- cise than the modified historical egg production (MHEP) index used for recent seasons by Lo and Methot (1989). Both the HEP and EPI indices of egg production for northern anchovy during the 1979 to 1984 fishing seasons were used so that the model could calibrate the HEP and EPI against each other and against DEP spawning biomass estimates. The EPI for northern anchovy averages values obtained by using the HEP method and a single equa- tion model (SEM) developed by Lo (1986). We refer to the HEP "method" here to distinguish between the HEP index and calculations for recent seasons based on data from the reduced CalCOFI grid. The EPI index was computed as / = W 1EPI.y ''SEM, lSEM. + **HEP,y *HEP,y> (2) where the It are indices of egg production for north- ern anchovy during fishing season y, and the Wf are weights. Weights were derived from squared inverse coefficients of variation: W L_ CV,2. I CV HEP.y (3) where CVs is the coefficient of variation for index s (either SEM or HEP) in fishing season y. Variances for the EPI index were approximated VAR(IEPly) = WiEMy VAR(lSEM,y) + W^EPyVARUHEPy J, (4) where VAR denotes variance. The weighted EPI estimates of egg production for northern anchovy were similar to HEP method and SEM estimates but more precise (Table 2). The im- provement in precision is overestimated, however, because covariance between measurement errors in SEM and HEP values in each year were not included in Equation 4. Model Our model for northern anchovy was based on a for- ward simulation approach (Hilborn and Walters, 1992) like that used in the stock synthesis model (Methot, 1989; Methot, 1990) and CAGEAN (Deriso et al., 1985). The model simulated abundance of northern anchovy during the 1963 to 1991 fishing seasons given a set of parameter estimates, data for catches, and ocean temperatures. Parameters were estimated by maximum likelihood calculations that compared observed abundance indices with values predicted by the simulation model. Catch data for northern anchovy and temperature data were as- sumed to be measured without error; abundance in- dices were assumed to include measurement error. Table 2 Egg production indices for northern anchovy, Engraulis mordax in the re iuced CalCOFI area during peak spawn- ing (15 February): historical egg production method (HEP), sin gle equat ion model (SEM), and the new weighted index (EPI). SE is standard error, CV is coefficient of variation, and "weights" are weights used in computing EPI values. HEP SEN EPI Eggs Eggs Eggs Calendar Fishing (0.05 m"1 CV (0.05 rrr1 CV (0.05 m"1 CV year season day-1 SE ', Weights day-1 SE c; i Weights day-1 SE c; ) 1980 1979 1.49 0.49 33 0.61 3.19 1.30 n 0.39 2.16 0.59 27 1981 1980 2.20 0.37 17 0.64 4.33 0.98 23 0.36 2.96 0.42 14 1982 1981 1.16 3.42 296 0.00 1.90 0.32 17 1.00 1.90 0.32 17 1983 1982 3.18 1.72 54 0.36 1.71 0.69 in (i in 2.24 0.76 34 1984 1983 6.12 3.57 58 0.04 5.74 0.65 11 0.96 5.75 0.64 11 1985 1984 8.13 4.43 55 0.70 3.92 3.30 84 0.30 6.89 3.27 48 1986 1985 8.23 4.48 5-4 0.35 6.53 2.62 40 0.65 7.13 2.32 32 1987 1986 8.98 3.18 35 0.17 4.13 0.67 in 0.83 4.97 0.78 16 1988 1987 6.58 3.41 52 0.40 6.05 2.54 42 0.60 6.26 2.04 33 1989 1988 0.23 0.11 46 0.30 2.01 0.61 30 0.70 1.47 0.43 29 1990 1989 2.35 0.85 HI, 0.91 1.89 2.24 119 0.09 2.31 I) Ml 35 1991 1990 2.13 1.07 50 0.20 1.74 0.44 25 0.80 1.82 0.41 23 1992 1991 0.15 0.04 'J!) 0.61 1.98 II 71 36 0.39 0.86 0.28 32 Jacobson et al.: A biomass-based assessment model for Engraulis mordax 715 Population dynamics Fishing seasons were used as annual time steps, and ages 0 to 4+ were included (age group 4+ includes northern anchovy age 4 and older). Fish were aged in the model at the beginning of each fishing season on 1 July when recruitment of age-0 northern an- chovy was assumed to occur (Methot, 1989). In real- ity, some recruitment of northern anchovy occurs throughout the year (MacCall and Prager, 1988). Therefore, our estimates of recruitment should be regarded as estimates of "effective" recruitment, i.e. biomass of age-0 fish that would have been neces- sary on 1 July to account for the biomass of the co- hort in later years. Numbers of northern anchovy were not included in SMPAR; abundance was measured solely in units of biomass because weight at age for northern an- chovy changes rapidly throughout the year, and de- pends on where samples are taken (Parrish et al., 1985). In addition, weight-at-age data from commer- cial fisheries for northern anchovy were not avail- able for recent fishing seasons. Biomass dynamics were modeled as B. H - fio., (5) where Bav is the biomass of northern anchovy age a (a>0, i.e. excluding new recruits) at the beginning of fishing season y and n is the net instantaneous rate of change for northern anchovy in fishing season y. Random process errors (e.g. variation in growth and natural mortality, Hilborn and Walters, 1992) were captured in the model by recruitment estimates. For modeling purposes, recruitment of northern anchovy in each year was assumed independent of spawning stock size: Bn By (6) where BQ is recruitment ( biomass age-0 fish ) in fish- ing seasony, BQ is mean recruitment during the study period, and 8yis a log-normally distributed error term for fishing season y with mean zero and standard deviation o. Recruitments in each fishing season (S0y) were treated as parameters and estimated by the model. The net instantaneous rate of change for northern anchovy biomass in each fishing season (r; in Eqn. 5) is the sum of rates for fishing mortality growth, and natural mortality: nv =Fv + M-G, (7) where Fy is the fishing mortality rate in fishing sea- son y, M is the natural mortality rate, and G is the growth rate. All rates are defined as positive values. The fishing mortality rate for each fishing season (Fy) was assumed constant over ages but variable over time, whereas rates for natural mortality (M) and growth (G) were assumed constant over ages and time. Fishing mortality rates were calculated by us- ing the "forward solution" algorithm in Sims (1982) and actual catch data (Table 1; Fig. 1). The rate of natural mortality (M) for northern an- chovy was assumed to be 0.8 yr_1, which is reason- able for a fish that seldom exceeds seven years in age (Hoenig, 1983). Methot ( 1989) found that differ- ent levels of natural mortality had only modest ef- fects on biomass estimates for northern anchovy be- cause the estimates were anchored by DEP spawn- ing biomass measurements. Modeling growth as an instantaneous rate (G) is appropriate for northern anchovy because fish grow rapidly throughout the fishing season (Zhang and Sullivan, 1988). By treating growth as an instanta- neous rate, northern anchovy are, in effect, allowed to continue growing in the model until they are caught. The rate for growth used in the SMPAR model for northern anchovy (G=0. 198 yr"1, SE=0.0166) was es- timated by fitting an exponential growth model to mean weight at age data from three sources (Methot, 1989). The exponential growth model was logarith- mically transformed to give \n(Wda) = \n(Wd0) + aG, (8) where Wda is the mean weight of northern anchovy age a in data set d, and Wd 0 is the estimated weight at age 0. The approach assumes that northern an- chovy may differ in initial weight as measured by the W(l0 parameters but experience the same rate of exponential growth (G). Parameter estimates for Equation 8 were obtained by linear regression and standard general linear model techniques (Weisberg, 1980). Residuals were dome-shaped because of the linear approximation to the asymptotic growth pat- tern but the linear regression model explained most of the variation in log-scale size at age (R2=93%). Abundance data Abundance data (EPI, HEP, SONAR, DEP, and SPOTTER abundance indices) were assumed to be measured with log-normally distributed random er- rors. Predicted values for abundance data during each fishing season were calculated in the model as: l,y=Qt^Pt,aBaye-"', (9) 716 Fishery Bulletin 92(4). 1994 where hats ( A) denote estimates, It is the value for abundance index t in fishing season y, Qt scales north- ern anchovy biomass to the units of abundance in- dex t, and pt is the relative contribution of a north- ern anchovy at age a to abundance index t. We as- sumed age-specific selectivity patterns for abundance indices because estimates for most parameters were available outside of the model. This approach gave a more realistic model without increasing the number of parameters estimated. Values of pt were relative measures scaled to the interval [0,1], and the age with maximum relative contribution for abundance index t had p =1.0. Estimates of the scaling parameter for DEP data (QDEp=l) and age-specific parameters (pta) for DEP, HEP, and EPI data were from Methot (1989). Two- year-old northern anchovy are all sexually mature during the peak spawning period (pD£P2+=1.0), whereas the fraction of one-year-olds that are ma- ture (pDEP i ) depends on water temperatures. Matu- rity of age-1 northern anchovy during the peak spawning season was calculated from mean Janu- ary-February sea surface temperatures at Scripps Pier, San Diego, California (Table 1), as described in Methot (1989). Estimates of age-specific egg production for ac- tively spawning northern anchovy (Methot, 1989) were used to estimate the age-specific parameters (Phep.o and Pepia ) for egg production indices. No age- 0 northern anchovy spawn during the peak spawn- ing period but all are actively spawning by age 2. The fraction of actively spawning fish was also cal- culated from mean sea surface temperatures (Methot, 1989). Age-specific parameters for contribution to egg production indices (Phep.o and Pepia) were assumed to be the product of relative egg production and frac- tion active. Relative egg production values were the same as those used by Methot (1989) and originally by Parrish et al. (1986). For simplicity, relative age-specific contributions to indices of schooling biomass (SPOTTER and SO- NAR) for northern anchovy ages 1 and older (Pspotter,i+ and Psonar,i+) were assumed to be 1.0. The contribution of age-0 northern anchovy to the SPOTTER and SONAR indices was estimated as -X^Xa^-^IX, (id Psi'DTTER.O ~ PSONAR.O 10) 1+e" where jt was a parameter estimated by the model. Objective function Parameters in SMPAR were estimated by maximizing a function proportional to the total log-likelihood (Llotal): where Nt is the number of observations for abundance index t, and N is the number of recruitment esti- mates. The kt values are weights that determine how important different types of data are in parameter estimation; they were set to one except during sensi- tivity analyses. Dt is the log-scale standardized re- sidual for abundance index t in fishing season y and R is the log-scale standardized residual for recruit- ment in fishing season y: I) In (/,,,//,,,) £t.y ln(/,v)-ln(/,v) (12) R, -t.) \n(BQy/B0) a a (13) where et is the log-scale standard error for abun- dance type t in fishing season v, and a is the stan- dard deviation for log-scale recruitment deviations (8 in Eqn. 6). Log-scale standard errors for abun- dance data (c ) were calculated from arithmetic scale coefficients of variation by inverting Equation 1. The first term on the right side of Equation 11 gives the log likelihood of abundance indices given param- eters in the model. The second term gives the log likelihood of recruitment estimates. Mean recruit- ment ( B„ in Eqn. 13) is a "nuisance" parameter that was set equal at each iteration to the mean of cur- rent recruitment estimates. The log-scale standard deviation assumed for recruitments (a=0.71) was calculated from stock synthesis model2 recruitment estimates and was higher than the average standard deviation (0.48) for 41 other stocks of clupeoid fishes (Beddington and Cooke, 1983; Myers et al., 1990). The likelihood term for recruitments in Equation 11 is a constraint that penalizes individual recruit- ment estimates that are different from the mean. Larger deviations and smaller a values result in larger penalties. The constraint does not penalize serial correlation so that "runs" of good or bad re- cruitments can be estimated by the model. This was important because northern anchovy recruitments tend to be serially correlated (see below). Jacobson and Lo2 showed that a northern anchovy model without age-composition data or a recruitment Jacobson et ai.: A biomass-based assessment model for Engraulis mordax 717 constraint like that in Equation 11 was overpar- ameterized because recruitments need occur only once every two to three years for the model to match observed and predicted abundance data. Age-com- position data for northern anchovy indicate, however, that some recruitment occurs during every fishing season (Lo and Methot, 1989). We included the re- cruitment constraint and a recruitment parameter for each season to obtain a more realistic model and to constrain the recruitment estimate for the last fish- ing season which was otherwise difficult to estimate. The constraint on recruitment biases recruitment and biomass estimates towards the mean because recruitment estimates will be high in years with poor recruitment and low in years with high recruitment. Parameters in the model were estimated by using the simplex algorithm (Press et al., 1990). Variances and correlations for parameter and biomass esti- mates were calculated by using a parametric boot- strap approach (Efron, 1982) as described in Lo et al. (1992) except that simulated abundance data were generated by assuming log-normal errors with stan- dard deviation equal to the root-mean-squared log- scale residual for each data type (see below). Param- eters for bootstrap runs were estimated as described for the original run by using the original CVs for each abundance index observation. Thus, our boot- strap runs included process error to the extent that it was reflected in the variance of residuals, and in- cluded measurement error to the extent that it was reflected in the original CVs. Two thousand bootstrap iterations were generally used. Asymptotic variance and correlation estimates for parameters were also calculated by inverting a numerical approximation to the Hessian matrix (Bard, 1974; Mittertreiner and Schnute, 1985) because we were interested in compar- ing the asymptotic and bootstrap approaches. Parameters with all feasible values positive were estimated as log-transformed values. The log trans- formation constrains parameters to feasible values on the original scale and improves the statistical characteristics of parameter estimates. Standard errors for log-scale recruitment parameters were transformed to CVs for arithmetic recruitment esti- mates by using Equation 1. Results and discussion Estimates from preliminary runs indicated that availability of age-0 northern anchovy to indices of schooling biomass was close to zero. For final runs, Pspotter.o and PsoNAJt.o were set to zero and not esti- mated even though age-0 fish were assumed to be fully recruited to the fishery. Outliers and residual analysis Standardized residuals (Dt , in Eqn. 12) for most abundance indexes were serially correlated in pre- liminary runs. There were two outliers (DEPI 1983=3.6 and DspoTTER 1979=3.5) identified by a £-test with Bonferronip- values (critical value D. =3. 41 forn=77; Weisberg, 1980). Residual plots for the final run with outliers omitted still indicated some serial correla- tion. All but three biomass estimates for northern anchovy during the 1963 to 1991 fishing seasons in- creased when the two outliers were omitted. The average increase was 24%. CVs for abundance indices and goodness of fit The root-mean-squared residual for each abundance index was calculated to measure how well the SMPAR model fit the data for northern anchovy. Standard deviations were not calculated because degrees of freedom were unknown. Arithmetic CVs implied by the goodness-of-fit statistics were calcu- lated by using Equation 1. For comparison, median CVs for our data (Table 1) were also calculated. Goodness-of-fit statistics and implied CVs (Table 3) indicate that the CVs for our abundance data un- derestimated the true log-scale standard errors. The order of abundance indices ranked by median CVs was, however, the same as when they were ranked by goodness of fit. Thus, CVs used in the model re- flected the relative precision of different types of abundance data for northern anchovy. Consistent bias Percent bias (%BIAS) in biomass and log-scale recruit- ment estimates for northern anchovy was estimated Table 3 Goodness-of-fit and CV statistics for northern an- chovy, Engraulis mordax, abundance indices used in the SMPAR model. Median nominal CVs were cal- culated from arithmetic CV values in Table 1. Root- mean-squared residuals measure goodness of fit to abundance index. Implied CVs are the arithmetic CV values calculated from the goodness-of-fit measures. Median Root- Abundance nominal mean-squared Implied index n CVi'-i residual CVC*) DEP EPI SPOTTER HEP SONAR 6 L2 ■I 7 11 16 Lil 31 34 11 ■Hi 0.19 0.48 0.49 0.50 0.53 19 51 52 53 57 718 Fishery Bulletin 92(4), 1994 %BIAS = _ Eboot E, bfsl (14) E, hest where Eboot is the average of estimates from two thou- sand bootstrap runs and Ebest is the best estimate from the model fit to the original data. We used the correction factor yy-Eboot-Ebest to remove consistent bias (Efron, 1982). The corrected estimate of log re- cruitment in fishing season y, for example, was PY~Jy, where B was the original biased log-scale recruit- ment estimate. Uncorrected log-scale recruitment estimates were biased by amounts ranging from -7% to 7% and arith- metic scale biomass estimates by amounts ranging from -15% to 27% (Fig. 2). Consistent bias was ex- aggerated when uncorrected recruitments were transformed to arithmetic scale (—40% to 43%, Fig. 3). Bias from log-transformed recruitment estimates Arithmetic scale recruitments in each fishing sea- son (B0 ) were calculated B0y=e\^-^ (15) where VAR is a variance estimated by bootstrapping. The term VARifi )/2 adjusts for bias due to transfor- mation of log- normally distributed random variables (Beauchamp and Olson, 1973). Bias due to log trans- formation is in addition to consistent bias estimated by "iy=Eb„„r^besf Tne correction for bias due to log transformation increased northern anchovy recruit- ment estimates by 1% to 22% (average 9%). Corrections for bias due to log transformation make arithmetic recruitment estimates for northern an- chovy easier to interpret because the amount of bias varies among uncorrected recruitment estimates as a function of their variance. Many stock assessment models (Deriso et al., 1985; Methot, 1990) estimate recruitments as log-scale parameters but corrections for bias in arithmetic-scale recruitment estimates are not made. We recommend that bootstrap or other variance estimates be used to correct arithmetic scale recruitment estimates for bias where appropriate. Retrospective bias We evaluated potential for retrospective bias in the SMPAR model by comparing our best biomass esti- mates to estimates from runs that omitted data for recent years. Bias corrections for retrospective analy- sis were based on fifty bootstrap iterations. Results indicated a negligible amount of retrospective bias. Estimates and comparisons Biomass estimates for northern anchovy age 1 and older from the SMPAR (Table 4) and stock synthesis 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 FISHING SEASON Figure 2 Biomass estimates for northern anchovy, Engraulis mordax (age 1+ on 15 February, in thousands of metric tons), during the 1963 to 1991 fish- ing seasons from the SMPAR model and the stock synthesis model used by Lo and Methot (1989). Estimates from the SMPAR model are shown with and without correction for bias. Jacobson et al.: A biomass-based assessment model for Engraulis mordax 719 SIMPLE MODEL [NO BIAS CORRECTION! 1962 1964 1966 1966 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 FISHING SEASON Figure 3 Recruitment estimates for northern anchovy, Engraulis mordax (age-0 fish on 1 July, in thousands of metric tons) for the 1963 to 1991 fishing seasons from the SMPAR model (biomass) and the stock synthesis model (number of recruits) used by Lo and Methot ( 1989). Estimates from the SMPAR model are shown with and without correction for bias. models were generally similar (Fig. 2) except where retrospective bias in previous estimates was a prob- lem. Results from SMPAR indicate that high north- ern anchovy biomass in the early 1970's was due to a single large cohort spawned in the 1971 fishing sea- son (age 1 in 1973) rather than to a series of strong recruitments (Figs. 2 and 3). This difference is due to omission of age-composition data and inclusion of SPOTTER data not available to Lo and Methot ( 1989). Unlike SONAR data, SPOTTER data did not increase significantly until the 1973 fishing season (Fig. 1). Coefficients of variation for northern anchovy 1+ biomass estimates ranged from 14% to 38% and av- eraged 26% (Table 4). Precision of the biomass esti- mate for the most recent fishing season (29%) was better than that from the original stock synthesis model for northern anchovy (CV=40%, Lo et al., 1992). Improvements in data contributed to higher precision, but this result indicates a substantial im- provement as a result of using a simpler, more par- simonious model. Recruitment estimates for northern anchovy from the SMPAR model showed less year-to-year varia- tion than those from the stock synthesis model al- though both sets of estimates indicate that northern anchovy recruitment was low during the 1963 to 1968 fishing seasons and high in the 1972 fishing season (Fig. 3). CVs for arithmetic-scale recruitment esti- mates (11% to 69%, average 41%) from the SMPAR model were about 50% larger on average than CVs for biomass estimates (Table 4). The most important conclusion to be drawn in com- paring recruitment estimates from SMPAR with those from the stock synthesis model for northern anchovy is that disparate recruitment estimates (Fig. 3) resulted in similar biomass estimates (Fig. 2). This suggests that abundance index data for northern anchovy contain relatively little information about recruitment variability. The SMPAR model probably underestimates recruitment variability for northern anchovy because it uses only abundance data and includes a recruitment constraint that biases recruit- ment estimates towards the mean. Comparison of bootstrap and asymptotic variance estimates Asymptotic standard errors for parameters obtained by inverting the Hessian matrix were about 19% smaller, on average, than standard deviations ob- tained by bootstrapping. This result indicates that asymptotic variance estimates for parameters in the SMPAR model were too small. 720 Fishery Bulletin 92(4). 1994 Sensitivity to weights We varied weights (A() in the objective function of the SMPAR model to determine how sensitive recent biomass estimates were to different types of data and to the recruitment constraint. Biomass estimates for age 1+ northern anchovy during the 1991 fishing season and average biomass during the 1985 to 1991 fishing seasons were calculated with the weight for each data type set equal to a range of values while other weights were kept at 1.0. Sensitivity analysis for the weight applied to the recruitment constraint (A6 in Eqn. 11) is equivalent to a sensitivity analysis on the standard deviation for log-scale recruitments (o in Eqn. 13). Doubling the weight (A6=2.0) is the same as reducing the assumed standard deviation by 1/V2=0.71. For simplicity biomass levels in sensi- tivity analyses were not corrected for bias. SMPAR did not completely converge when the weight for the recruitment constraint was set to 0.0 or 10.0. Results indicate that current biomass estimates and management advice are affected by weights on EPI data and the recruitment constraint (Table 5). Estimates of anchovy biomass for 1991 were sensi- tive (changes >10%) to halving (A,=0.5) or doubling (A,=2.0) weights on EPI data and the recruitment constraint, but estimates of average biomass during 1985 to 1991 were not. Biomass estimates for 1991 were sensitive to removing (A(=0.0) EPI data, EPM data, or the recruitment constraint from the model. Average biomass estimates were sensitive to remov- ing EPI, EPM, or SPOTTER data. Simulation analyses We used simulation analyses to determine how well SMPAR would estimate biomass under a wide range of recruitment and fishing mortality rates during the 1985 to 1991 fishing seasons. Calculations were the same as those for bootstrapping except that fishing Table 4 Total, spawning , and recruitment biomass estimates for northern anchovy, Engraulis mordax, during the 1963 to 1991 fishing seasons (corrected for bias). Recruitment estimates are for age-zero northern anchovy at th e beginning of the fishing season (1 July). Total and spawning biomass estimates are for northern anchovy age 1 and older during the peak spawning period ( 15 Feb ruary) of each fishing season. Total Spawning Recruitment Fishing Fishing biomass cv biomass CV biomass CV mortality season (1.000 t) (%) ( 1,000 t) (%) ( 1,000 t) (%) (vrl) 1963 764 38 734 38 275 43 0.002 1964 477 27 434 31 334 1" 0.003 1965 356 24 313 23 627 12 0.023 1966 370 26 311 25 347 42 0.063 1967 293 26 281 26 319 39 0.022 1968 249 25 232 24 1.066 39 0.037 1969 429 26 260 23 364 43 0.111 1970 301 27 232 25 1.348 47 0.074 1971 527 33 195 25 1,008 52 0.045 1972 558 32 547 32 5,941 If. 0.020 1973 2,066 34 638 25 2.667 69 0.034 1974 1,733 31 1,243 29 3,955 64 0.042 1975 1,827 30 1,481 27 1,242 54 0.068 1976 1,216 26 1,215 26 693 19 0.120 1977 742 26 741 26 1.211 is 0.132 1978 639 26 530 24 1,204 36 0.172 1979 619 20 607 20 1.611 3 1 0.121 1980 670 16 630 15 623 35 0.301 1981 415 19 381 17 1.624 22 0.224 1982 589 1 1 587 14 246 32 0.200 1983 341 1 1 339 1 1 2,321 26 0.039 1984 937 21) 609 17 2,372 ;,s 0.062 1985 1,072 32 1,057 32 388 35 0.065 1986 620 27 616 27 280 39 0.125 1987 360 2 1 343 24 1,480 38 0.095 1988 543 27 235 25 261 3M 0.120 1989 317 30 309 30 201 37 0.112 1990 199 :;i 198 2 1 396 39 0.018 1991 223 29 221 29 739 11 0.005 Jacobson et al.: A biomass-based assessment model for Engraulis mordax 721 mortality rates and recruitment levels for 1985 to 1991 were ad- justed. For simulations, fishing mortality rates during the 1985 to 1991 fishing seasons were either low (best-fit estimates from actual data, 0.005 to 0.12 yr"1, Table 4) or high (F=1.0 yr-1). Recruitment levels were either low (one third of best-fit estimates in Table 4), equal to best-fit levels, high (three times best-fit levels), or alternat- ing (three times best fit for 1985, one third of best fit for 1986, and so on). There were eight scenarios in total (two fishing mortality pat- terns combined with four recruit- ment patterns) and results for each scenario are averages based on fifty simulations. Each scenario used the same series of random numbers to facilitate comparisons. Results indicate that the SMPAR model is able to track trends in bio- mass (Fig. 4) under a wide variety of conditions. Uncorrected esti- mates underestimated year to year variability but this problem would be reduced after bias correc- tions were applied. Table 5 Sensitivity of biomass estimates for northern anchovy, Engraulis mordax, to weights (A,) used in SMPAR. Differences between best-fit results and results wi th different weights are shown in metric tons and as a percent- age of the best-fit estimates. Percentages of values are in parentheses. Weight Recruitment a,) EPI EPM HEP SONAR SPOTTER constraint 1991 Biomass 0.0 141 65 -18 2 -4 -55 (46) (21) (-6) (1) (-1) (-18) 0.5 52 -6 -1 7 0 -32 (17) (-2) (0) (2) (0) (-11) 2.0 -50 10 8 -1 2 53 (-16) (3) (3) (0) (1) (17) 10.0 -166 22 22 -31 -17 209 (-54) (7) (7) (-10) (-6) (69) Mean 1985 -91 Biomass 0.0 -51 -137 -28 12 53 35 (-10) (-28) (-6) (2) (11) (7) 0.5 -11 -14 -1 13 24 0 (-2) (-3) (0) (3) (5) (0) 2.0 7 22 11 -8 -35 10 (1) (4) (2) (-2) (-7) (2) 10.0 -76 53 34 -59 -145 53 (-16) (11) (7) (-12) (-30) (11) Age composition in the Mexican fishery Catch curves (Ricker, 1975) for a segment of the Mexican fishery based in Ensenada, Baja California, indicate that age at full recruitment to the Mexican fishery decreased from age 2 to age 0 during 1982 to 1988. 4 Prior to 1983, most year classes reached full recruitment at age 2 but the 1985 and 1986 year classes were fully recruited at about age 1. The trend toward younger recruitment continued with the 1987 and 1988 year classes, which were fully recruited at age 0. The stock synthesis model interpreted the Mexi- can age-composition data as evidence for increased recruitment and biomass of northern anchovy. In retrospect, this interpretation seems incorrect be- cause biomass estimates for anchovy declined steadily after the 1985 fishing season (Fig. 2). De- clines in biomass were not evident at the time, how- 4 Arenas, P., T. Barnes, and L. D. Jacobson. 1994. Fishery and biological data for northern anchovy taken in Mexican waters, 1978-1989. U.S. Dep. Commer., NOAA, Nat. Mar. Fish. Serv., Southwest Fish. Sci. Cent., P.O. Box 271, La Jolla, CA 92038. Admin. Rep. LJ-94-03, 24 p. ever, because only one index of relative abundance for northern anchovy (the modified historical egg production index or MHEP, Lo and Methot, 1989) was available, and because recent biomass estimates are relatively uncertain. Status of the stock Northern anchovy biomass (Fig. 2) and recruitment (Fig. 3) declined after the 1985 fishing season to about the same levels as during the 1963 to 1971 fishing seasons. Northern anchovy have been too scarce off Baja California, Mexico, to support a fishery since the 1990 season (Table 1). Declines in biomass dur- ing recent years were due to low recruitment rather than to high fishing mortality rates because fishing mortality rates were moderate after the 1986 fish- ing season (<0.14 yr-1) and low (<0.03 yr-1) during the 1990 to 1991 fishing seasons (Table 4). The re- cent period of low northern anchovy biomass occurred as Pacific sardine, Sardinops sagax, biomass levels began to increase in the early 1980's and water tem- peratures began to warm (Barnes et al., 1992). We 722 Fishery Bulletin 92(4), 1994 did not attempt, however to identify environmental (Prager and MacCall, 1993) or ecological factors that may have affected northern anchovy abundance in recent fishing seasons. o o o to V) < I O 1.400 1.200 1.000 800 ' 000 400 200 1.400 1.200 1.000 800 BOO 400 200 LOW RECUITMENT. LOW F .400 - -■- - TRUE .200 ' ' AVERAGE 800 «. 800 \ 400 ^^^"i-» 200 *"""-*-=-=» =-*-—-. _ 7 ■ ■ . L— . - - ■ 4i 1986 1987 19 1989 1990 1991 BEST FIT RECRUITMENT. LOW F 1987 1988 19 HIGH RECRUITMENT. LOW F 1990 1991 1986 1987 1989 1990 1991 ALTERNATING RECRUITMENT. LOW F 1.400 1.200 1.000 800 600 400 200 1.400 1.200 1.000 800 600 400 200 1.400 1.200 1.000 ■00 800 400 200 LOW RECRUITMENT. HIGH F 1.400 1.200 1.000 800 600 400 L* • 200 0 1 v ■ ■ — — ~^i ■ ■ i i _ _ _ 1986 1987 1988 1989 1990 1991 BEST FIT RECRUITMENT. HIGH F 1986 1987 1988 19 HIGH RECRUITMENT. HIGH F 1990 1991 ALTERNATING RECRUITMENT, HIGH F FISHING SEASON Figure 4 Northern anchovy, Engraulis mordax, biomass results (age 1+ on 15 February in thousands of metric tons) for the 1985 to 1991 fishing seasons from simulation analyses. Each panel contains results for one recruitment pattern (low, best fit, high or alternating) and one fishing mortality pattern (low or high). The "TRUE" line in each panel is the true biomass assumed in the simulations. The "AVERAGE" line is the mean of biomass estimates from fifty simulations. Estimates from simulation analyses were not corrected for bias. Jacobson et al.: A biomass-based assessment model for Engraulis mordax 723 Conclusion Finding the appropriate compromise between real- istic (but potentially overparameterized and impre- cise) models and parsimonious (but simplistic and potentially biased) models is an important part of a stock assessment research (Ludwig and Walters, 1985, 1989; Hilborn and Walters, 1992). The best choice among models depends on the task, data avail- ability, and complexity of the situation. Our work indicates that models intermediate in complexity between simple surplus production and complex age- structured models can perform well under a wide range of circumstances for estimating the biomass of stocks such as northern anchovy. Acknowledgments W. Garcia (Centro Regional Investigacion de Pes- quera), Pesquera Zapata S.A. de C.V., and LMR Fish- eries Research provided Mexican fishery data. E. Konno and P. Wolf (California Department of Fish and Game) provided U.S. fishery data. A. MacCall, B. Lenarz (Southwest Fisheries Science Center, Tiburon Laboratory), R. Methot (Alaska Fisheries Science Center), R. Parrish (Southwest Fisheries Science Center, Pacific Fisheries Environmental Group), P. Smith, and J. Hunter (Southwest Fisher- ies Science Center), and an anonymous reviewer made suggestions that improved the scientific basis and presentation of results. Literature cited Bard, Y. 1974. Nonlinear parameter estimation. Academic Press, New York, NY, 341 p. Baxter, J. L. 1967. Summary of biological information on the northern anchovy Engraulis mordax Girard. Calif. Coop. Oceanic Fish. Invest. Rep. 11:110-116. Barnes, J. T., L. D. Jacobson, A. D. MacCall, and P. Wolf. 1992. Recent population trends and abundance es- timates for the Pacific sardine (Sardinops sagax). Calif. Coop. Oceanic Fish. Invest. Rep. 33:60-75. Beauchamp, J., and J. Olson. 1973. Corrections for bias in regression estimates after logarithmic transformation. Ecology 54:1403-1407. Beddington, J. R., and J. G Cooke. 1983. The potential yield offish stocks. FAOFish. Tech. Pap. 242. Bence, J. R., A. Gordoa, and J. E. Hightower. 1993. Influence of age-selective surveys on the reli- ability of stock synthesis assessments. Can. J. Fish. Aquat. Sci. 50:827-840. Deriso, R. B., T. J. Quinn II, and P. R. Neal. 1985. Catch-age analysis with auxiliary informa- tion. Can. J. Fish. Aquat. Sci. 42:815-824. Efron, B. 1982. The jackknife, the bootstrap and other re- sampling plans. Society for Industrial and Applied Mathematics, Philadelphia, PA, 92 p. Gunderson, D. R. 1993. Surveys of fisheries resources. John Wiley and Sons, Inc., New York, NY, 248 p. Hewitt, R. P. 1988. Historical review of the oceanographic ap- proach to fishery research. Calif. Coop. Oceanic Fish. Invest. Rep. 29:27-41. Hilborn, R., and C. J. Walters. 1992. Quantitative Fisheries stock assessment. Routledge, Chapman, and Hall Inc., New York, NY, 570 p. Hoenig, J. M. 1983. Empirical use of longevity data to estimate mortality rates. Fish. Bull. 81:898-903. Lapointe, M. F., R. M. Peterman, and A. D. MacCall. 1989. Trends in fishing mortality rate along with errors in natural mortality rate can cause spuri- ous time trends in fish stock abundances estimated by virtual population analysis (VPA). Can. J. Fish. Aquat. Sci. 46:2129-2139. Lasker, R. (ed.) 1985. An egg production method for estimating spawning biomass of pelagic fish: application to the northern anchovy (Engraulis mordax). U.S. Dep. Commer, NOAA Tech. Rep. NMFS 36, 99 p. Lo, N. C. H. 1985. Egg production of the central stock of north- ern anchovy, Engraulis mordax, 1951-82. Fish. Bull. 83:137-150. 1986. Modeling life-stage-specific instantaneous mortality rates, an application to northern anchovy, Engraulis mordax, eggs and larvae. Fish. Bull. 84:395-407. Lo, N. C. H., and R. D. Methot. 1989. Spawning biomass of the northern anchovy in 1988. Calif. Coop. Oceanic Fish. Invest. Rep. 30:18-31. Lo, N. C. H., L. D. Jacobson, and J. L. Squire. 1992. Indices of relative abundance from fish spot- ter data based on delta-lognormal models. Can. J. Fish. Aquat. Sci. 49:2515-2526. Ludwig, D., and C. J. Walters. 1985. Are age-structured models appropriate for catch-effort data. Can. J. Fish. Aquat. Sci. 42:1066-1072. 1989. A robust method for parameter estimation from catch and effort data. Can. J. Fish. Aquat. Sci. 46:137-144. 724 Fishery Bulletin 92|4), 1994 Mac-Call. A. D., and M. H. Prager. 1988. Historical changes in abundance of six fish species off southern California, based on CalCOFI egg and larva samples. Calif. Coop. Oceanic Fish. Invest. Rep. 29:91-101. Mais, K. F. 1974. Pelagic fish surveys in the California current. Cal. Dep. Fish and Game, Fish Bull. 162. Methot, R. D. 1989. Synthetic estimates of historical abundance and mortality for northern anchovy. Am. Fish. Soc. Symp. 6:66-82. 1990. Synthesis model: an adaptable framework for analysis of diverse stock assessment data. Int. N. Pac. Fish. Comm. Bull. 50:259-277. Mittertreiner, A., and J. Schnute. 1985. Simplex: a manual and software package for easy nonlinear parameter estimation and interpre- tation in fishery research. Can. Tech. Rep. Fish. Aquat. Sci. 1384, 90 p. Myers, R. A., W. Blanchard, and K. R. Thompson. 1990. Summary of North Atlantic fish recruitment 1942-1987. Can. Tech. Rep. Fish Aquat. Sci. 1743. Parrish, R. H., D. L. MaUincoate, and K. F. Mais. 1985. Regional variations in the growth and age composition of northern anchovy, Engraulis mordax. Fish. Bull. 83:483-496. Parrish, R. H., D. L. MaUincoate, and R. A Klingbeil. 1986. Age dependent fecundity, number of spawnings per year, sex ratio, and maturation stages in north- ern anchovy, Engraulis mordax. Fish. Bull. 84:503-517. Prager, M. H., and A. D. MacCall. 1993. Detection of contaminant and climate effects on spawning success of three pelagic fish stocks off southern California: northern anchovy Engraulis mordax. Pacific sardine Sardinops sagax, and chub mackerel Scomber japonicus. Fish. Bull. 91: 310- 327. Press, W. H., B. P. Flannery, S A. Teukolsky, and W. T. Vetterling. 1990. Numerical recipes. Cambridge Univ. Press, NY. Ricker, W. E. 1975. Computation and interpretation of biological statistics of fish populations. Fish. Res. Board Can. Bull. 191. Sims, S. E. 1982. Algorithms for solving the catch equation for- ward and backward in time. Can. J. Fish. Aquat. Sci. 39: 197-202. Sinclair, A., D. Gascon, R. O'Boyle, D. Rivard, and S. Gavaris. 1 99 1 . Consistency of some northwest Atlantic ground- fish stock assessments. NAFO Sci. Counc. Stud- ies 16:59-77. Weisberg, S. 1980. Applied linear regression. John Wiley and Sons, NY, 283 p. Zhang, C. I., and P. J. Sullivan. 1988. Biomass-based cohort analysis that incorpo- rates growth. Trans. Am. Fish. Soc. 117:180-189. Abstract. — Use of early life his- tory stages offish in systematic and ecological studies has increased in recent years. It is now recognized that eggs and larvae present a wide array of characters suitable for sys- tematic analysis that are largely independent of adult characters. Fisheries recruitment studies focus on survival of eggs and larvae as the most important factor influenc- ing variations in population abun- dance. A requisite to these studies is detailed information on the ap- pearance offish eggs and larvae in order to identify them in plankton samples. This paper reviews the proportions offish species for which at least illustrations of eggs and larvae, sufficient to permit their identification in plankton samples, have been published worldwide and by geographic region. Factors which may influence differences in proportion of identifiable eggs and larvae by region are discussed. Fac- tors considered important include species diversity, a history of im- portant commercial fisheries, re- search emphasis, and interests of individual scientists in various re- gions. We conclude that although eggs and larvae of most species can now be identified in some regions of the world, there are still gaps in our knowledge that prevent realiz- ing the full potential of ichthyo- plankton studies in systematic and fisheries research. Filling these gaps will require continued tradi- tional morphological research as well as application of biochemical genetic and rearing techniques. Status of early life history descriptions of marine teleosts Arthur W. Kendall Jr. Ann C. Matarese Alaska Fisheries Science Center, National Marine Fisheries Service. NOAA 7600 Sand Point Way NE. Seattle. Washington 98 1 1 5-0070 Manuscript accepted 20 April 1994. Fishery Bulletin 92:725-736. Fish eggs and larvae collected in plankton samples are becoming in- creasingly important to the study of fisheries, oceanography, and sys- tematics (e.g. Moser et al., 1984; Rothschild, 1986). Accurate identi- fication of early life history stages of fish is a requisite for studies in these fields. Several kinds of publi- cations assist researchers in iden- tifying fish eggs and larvae: 1 ) some revisions of primarily oceanic groups include descriptions of lar- vae (e.g. ceratioids [Bertelsen, 1951], bregmacerotids [D'Ancona and Cavinato, 1965], and scope- larchids [Johnson, 1974]); 2) de- tailed descriptions of the develop- ment of a species based on reared or plankton-caught individuals or on a combination of both are com- mon (e.g. Ahlstrom and Ball, 1954; Potthoff et al., 1980; Ditty and Shaw, 1992); and 3) descriptions of several species of a genus or family are also available (e.g. Ahlstrom et al., 1976; Belyanina, 1984; Kendall and Vinter, 1984; Baldwin, 1990). While these detailed accounts are essential for systematic studies, publications that describe species by geographic region rather than by systematic group are more useful for identification of eggs and larvae from plankton samples. Agassiz (1882), Ehrenbaum (1905, 1909), and D'Ancona et al. (1931-56) are early examples of such publications, and several others have been pub- lished recently (e.g. Miller et al., 1979; Auer, 1982; Fahay, 1983; Leis and Rennis, 1983; Wang, 1986; Okiyama, 1988; Leis and Trnski, 1989; Matarese et al., 1989; Olivar and Fortuno, 1991). Besides assist- ing researchers in identifying fish eggs and larvae in plankton collec- tions, these publications facilitate evaluation of descriptive develop- mental information available for fishes of a particular region. In this paper we 1) evaluate the degree to which descriptions of the early life history stages of marine fish are available on the basis of recently published early life history guides, 2) compare our current level of knowledge of fish egg and larva identification by geographic region throughout the world, and 3) dis- cuss potential reasons for regional differences in our level of knowledge. Methods We collected information on the sta- tus of early life history descriptions primarily from recently published guides on development of fishes in marine waters of seven specific re- gions of the world (Fig. 1): North- east Atlantic (Russell, 1976); North- west Atlantic (Mid-Atlantic Bight) (Fahay, 1983; supplemented by Fahay, 1993); Indo-Pacific (Leis and Rennis, 1983; Leis and Trnski, 1989); Japan (Okiyama, 1988; sup- plemented by Ozawa, 1986); Antar- ctica (Kellerman, 1989); Northeast Pacific (Matarese et al., 1989); and Southeast Atlantic (Olivar and Fortuno, 1991). We also used publi- cations that summarized available 725 726 Fishery Bulletin 92(4). 1994 descriptive early life history information for two ad- ditional geographic regions: the Mediterranean Sea (Aboussouan, 1989) and the west central Atlantic (Richards, 1990). Publications summarizing devel- opment of fishes in other major geographic regions (e.g. Southeast Pacific) are not available. We com- piled data from the published guides as well as from the more recent, but restricted, publications on the basis of the illustrations of eggs and larvae they con- tained and then compared our results with those of Richards (1985) who summarized early life history information available at that time based on the work of Moser et al. ( 1984). We employed six early life his- tory stages (egg, yolk sac, preflexion, flexion, post- flexion, and transforming) and considered a particu- lar stage of a species as known if an illustration of that stage had been published. The quality of egg and larval illustrations varies, and we subjectively excluded those that we thought would be inadequate for identifying plankton-collected specimens. Since accurate species lists were not always available in the guides, some subjectivity was involved in deter- mining the number of species present in a region. In cases where the geographic regions covered by early life history guides were more restricted than those con- sidered in regional species lists we used the lists found in the guides. Paxton et al. (1989) was also consulted for the number of species in several regions. Although information was available for eggs and juveniles in some regions, data on larvae were generally used for comparisons because they were more widely available. To assess the impact of particular scientists on the availability of early life history information on fish for a geographic region, we developed a key author index based on references in Moser et al. (1984). Moser et al. (1984) summarized available early life history information for all fishes, so its bibliography provides an indicator of the contributions of individu- als up to about 1982. Key authors were identified as having four or more publications describing the early life history of marine fishes, or as having published a regional compilation of early life history informa- tion. The key author index was calculated by divid- ing the number of publications by key authors by the number of species in a region. Since our purpose here was to indicate the relative amount of research on larval fish taxonomy in various regions, the num- ber of publications was tabulated rather than the number of taxa described. To give a historical per- spective, this index was calculated separately for papers published before and after 1950; 1950 was chosen arbitrarily but coincides roughly with in- creased worldwide harvest of fish following World War II (the world fish catch doubled between 1952 and 1965 [Schaefer and Alverson, 1968]). In the Dis- cussion section we also refer to work done before and after 1900 (as did Ahlstrom and Moser [1981]) to highlight the roots of ichthyoplankton research since the International Council for the Exploration of the Sea was founded in 1898 and since it began field work on fish eggs and larvae in 1901. To investigate the relationship between commer- cial fishing activity and the status of early life his- tory information on a regional basis, we compiled regional commerical catch data (Fig. 1) for 10-year intervals between 1938 and 1988 (Food and Agricul- ture Organization [FAO] 1965, 1974, 1979, 1984, 1991, 1992). The first year that such statistics were available was 1938; so compiling data in 10-year increments should document changes in the regional contribution to the world catch which could be re- lated to changes in early life history information. The reported catches include organisms other than ma- rine fish (e.g. molluscs, crustaceans, and seaweeds), but they are a rather uniform fraction of the total catch, and marine fish represent more than 80% of the total catch. Regional catches in 1938 and 1948 were averaged to represent conditions before 1950, and regional catches in 1958, 1968, 1978, and 1988 were averaged to represent conditions after 1950. The three Antarctic regions (48, 58, 88) used by FAO were combined for our analysis. Catches in the Indian Ocean (51, 57), east central Pacific (77), and west central Pacific (71) were combined to be comparable to the Indo-Pacific region considered by Leis and Rennisl 1983) and Leis and Trnski (1989), although the FAO areas included temperate waters not in- cluded in Leis and Rennis ( 1983) and Leis and Trnski ( 1989). Other FAO regions do not correspond exactly to the regions included in early life history guides (Fig. 1), but at the level of resolution used here such differences are probably insignificant. To indicate families with the greatest need for ad- ditional larval fish taxonomic research, we calculated the ratio of the number of species for which larvae had been described over the number of species present by family for the nine geographic regions for which data were available. It was not possible to cal- culate this ratio with the data in Aboussouan ( 1989) or Okiyama ( 1988). Among those above the median of this ratio, up to 10 per region, families in each region for which this ratio was >0.5, were then ranked on the basis of the number of species they contained. Results On a worldwide basis, Richards ( 1985 ) concluded that 20,423 fish species were included in the material summarized in Moser et al. ( 1984 ). Of these species, Kendall and Matarese: Early life history descriptions of marine teleosts 727 80N -60S Figure 1 FAO fishing areas ( numbers in squares ) and areas included in early life history guides ( shaded, numbers in circles which are keyed to references in Table 2, except for 10 which refers to Ozawa, 1986). Early life history guides in Indo-Pacific area 8 deal with reef and shore fishes in that area, although oceanic areas are also shaded on the map. the eggs of 726+ (4%) were known, and larvae of 1,932+ (10%) were known. At a higher taxonomic level, however, larvae of representative species from about two-thirds of the families of marine fishes are known ( Ahlstrom and Moser, 1981 ). Richards ( 1985 ) did not subdivide the larval stage as we have (yolk- sac, preflexion, flexion, and postflexion), but most of the illustrations in Moser et al. ( 1984) are of flexion and postflexion larvae. Although not specifically in- tended to provide original information for specific identification offish larvae, Moser et al. ( 1984) pre- sented many original illustrations. Regional guides to early life history stages of fishes are now available for nine large areas of the world ocean (Fig. 1). Such guides for both coasts of South America are noticeably lacking, as are guides to most oceanic regions. Some species described in the guides are also found outside the areas specifically covered in these guides; therefore identification of eggs and larvae collected elsewhere is also facilitated by the use of these guides. Some recent guides partially overlap the geographic regions addressed in earlier guides (e.g. Brownell [1979J studied fishes from the Cape of Good Hope, an area that was included in Olivar and Fortuno [1991]; Miller et al. [1979] stud- ied Hawaiian fishes, and their research was included in Leis and Rennis [ 1983] and Leis and Tmski [ 1989]; Ozawa [1986] studied oceanic fishes in the areas in- cluded in the works of Okiyama [1988], Leis and Rennis [1983], and Leis and Trnski [1989]; and Fritzsche [1978], Hardy [1978, a and b], Johnson [1978], Jones et al. [1978], and Martin and Drewry [1978] studied a portion of the Northwest Atlantic that was included in Fahay [1983]). We selected the Northeast Pacific as addressed in Matarese et al. (1989) (i.e. the Pacific Ocean and Bering Sea within 200 nautical miles of the coast between lat. 38°N and 66°N and west to long. 180°) for detailed evaluation of the taxonomic composition and available early life history information of the ichthyofauna (Table 1). Atotal of 627 species of fishes are found in the region, and 592 are thought to spawn in marine waters there (Matarese et al., 1989). These species represent 22 orders and 94 families. The most speciose orders are the Scorpaeniformes (272 species) and Perciformes (140 species). The most speciose families in the Scorpaeniformes are the Cottidae, Cy- clopteridae, and Scorpaenidae, and in the Perciformes, the Zoarcidae. While most of the fishes in the North- east Pacific are oviparous, only about 252 species 728 Fishery Bulletin 92(4). 1994 (43%) of them are expected to spawn pelagic eggs. Among the Scorpaeniformes, only Anoplopoma fim- bria is known to have pelagic eggs. Eggs have been illustrated for only 44 (less than 10%) of the species in the Northeast Pacific, and 8 of these species produce demersal eggs. Eggs are known for 16 of the 31 species of Pleuronectiformes with pelagic eggs (36% of all described eggs). Yolk-sac lar- vae are known for 90 species of fishes, including 32 species of the viviparous scorpaenid genus Sebastes and 18 species of Pleuronectiformes. Preflexion lar- vae are known for 165 species, flexion larvae for 169 species, postflexion larvae for 217 species, and trans- forming larvae for 150 species. Some pelagic juveniles are included in the count for transforming larvae, par- ticularly those of the genus Sebastes. At least one illus- tration of an early life history stage is available for 263 species in the Northeast Pacific (44% of the total). To give an indication of the rate of advances in knowledge of the ichthyofauna and its early life history in the Northeast Pacific, since publication of Matarese et al. ( 1989), one new species has been described (Yabe, 1991 ), and descriptions of the larvae of seven additional spe- cies have become available (Maeda and Amaoka, 1988; Matsui, 1991; Busby and Ambrose, 1993). The percentage of fishes with descriptive informa- tion available on early life history stages varies con- siderably in various regions of the world (Table 2). Compared with the Northeast Pacific where the eggs of 14% of the species with pelagic eggs are known, only 5% of such species in the western central Atlantic have been identified, whereas the eggs of about 70% of such species are known in the Northeast Atlantic. Compared with the Northeast Pacific, where the lar- vae of 44% of the species are known, larvae are known for 34% of the species found in waters around Japan, for 27% of the species in the western central Atlantic, and for only 10% of the species in the Indo-Pacific (Table 2). However, larval illustrations are available for more than half of the species in several geographic regions: Northwest Atlantic, Mediterranean Sea, Southeast At- lantic, Northeast Atlantic, and the Antarctic (Table 2). Based on early life history guides, the number of species for which early life history information is lack- ing varies by family and by region (Table 3). Among the families which have larvae described for fewer Table 1 Taxonomic composit ion and status of early life history descriptions of northeast Pacific fishes based on Matarese et al. (1989). Number of species illustrated Species Larva] stage with at F"ji rnilip*; S peril's present ]p?iQt nnp Order ± . j i ii 1 1 1 t a present Eggs Yolk-sac Preflexion Flexion Postflexion Transforming iitini uiic illustration Notacanthiformes 1 2 0 0 il n ii 0 0 Anguilliformes 8 Id n ii 2 2 ii ii 6 Clupeiformes 2 5 3 2 3 3 3 ■J 3 Salmoniformes 7 50 6 2 7 (i 12 5 12 Stomiiformes ii 23 2 1 1 in 17 'i 18 Aulopiformes 7 10 n 1 4 5 7 5 10 Myctophiformes ■J. Tl n 2 11 in 15 in 15 Gadiformes 5 L9 '_' 5 7 8 in 7 10 Ophidiiformes 3 in (i 0 3 2 1 0 3 Batrachoidiformes 1 1 0 0 ii n n ii 0 Lophiiformes 2 7 0 0 n 2 2 2 2 Gobiesociformes 1 2 0 1 2 1 1 1 2 Beloniformes 1 1 1 1 1 1 1 1 1 Atheriniformes 1 2 ■1 0 2 2 2 0 2 Lampriformes 2 2 1 0 2 0 2 1 2 Beryciformes 5 11 II 0 1 1 2 1 2 Zeiformes 1 1 0 II 0 ii 0 II 0 Gasterosteiformes 3 4 0 0 1 1 1 1 1 Scorpaeniformes 6 272 ■1 16 65 68 79 75 103 Perciformes 26 140 6 11 24 24 36 13 42 Pleuronectiformes 3 32 17 18 25 23 is 16 28 Tetraodontiformes 1 1 ii 0 1 0 1 1 1 Totals 94 627 11 •in 165 169 217 150 263 Percent of total species 7.0 14.4 26.3 27.0 34.6 23.9 41.9 Kendall and Matarese: Early life history descriptions of marine teleosts 729 Table 2 Comparison by geographic region of the number of species with egg and larval illustrations available key author index, and rank of commercial catches. Number Percent Key author Rank of illustrated known index catches Geographic region Species (FAO areas) present Eggs Larvae Eggs Larvae Source <1950 >1950 <1950 >1950 Northeast Pacific 592 263 14 44 Matarese etal., 1989 0.000 0.039 _; 7 (1:67) Japan (Northwest Pacific) 3500 1181 34 Ozawa, 1986; Okiyama, 1988 0.001 0.042 3 1 (2:61) Northeast Atlantic 131 91 108 70 82 Russell, 1976 0.298 0.008 1 3 (3:27) Southeast Atlantic 239 48 141 20 59 Olivar and Fortuno, 1991 0.017 0.0502 8 6 (4:47) Western Central Atlantic 1803 97 486 5 27 Richards, 1990 0.003 0.003 5 9 (5:31) Northwest Atlantic 317 135 222 43 71 Fahay, 19933 0.063 0.146 4 5 (6:21) Antarctic 158 80 51 Kellerman, 1989 0.000 0.044 12 13 (7:48,58,88) Indo-Pacific 3921 394 10 Leis and Rennis, 1983; Leis and Trnski, 1989 0.006 0.007 2 2 (8:51,57,71,77) Mediterranean Sea 569 360 63 Aboussouan, 1989 0.111 0.035 6 10 (9:37) World 20423 726 1932 4 10 Richards, 1985 1 North Pacific not divided into east and west regions. 2 Olivar and coauthors ha ve publish ;d at least nine descriptive papers since 1986 (see Olivar and Fortuno 1991) that i "included here would raise the value to 0.088. 3 M. P. Fahay (NOAA. Sandy Hook Laboratory. Highlar ds, NJ 0773 comm. Sept. 1993) indicated tha t data for the New Jersey area reported here are representative of the Northwest Atlantic. than half of the species present in several regions are Scorpaenidae, Macrouridae, and Bothidae. Be- sides these widely distributed families, large propor- tions of species of families with more restricted ranges are undescribed as larvae, such as the Cot- tidae in the Northeast Pacific. More species in oce- anic families are poorly known as larvae than are indicated in Table 3, because only Ozawa ( 1986) deals exclusively with that fauna. Discussion Factors contributing to variations in the amount of early life history information available among geo- graphic regions include the history of fisheries in the region, the presence of key researchers, and the taxo- nomic diversity and scientific interest in each region. History of the fisheries Generally the geographic regions where larvae of the majority of species are known have had long histo- ries of important fisheries. Studies on fish eggs and larvae were pioneered in the late 1800's by countries that engaged in the fisheries of the Northeast Atlan- tic (see Hempel, 1979; Ahlstrom and Moser, 1981). Work before 1900 consisted mainly of basic biologi- cal studies, sampling eggs and larvae at sea, and rearing eggs and yolk-sac larvae following artificial fertilization for release at sea. Similar studies were initiated concurrently off the east coast of North America. Although during this period the identity of the eggs and larvae of many species was established and knowledge of oceanography of the regions was expanded greatly, these early studies resulted in ill- fated mass releases of young larvae reared in hatch- eries on both sides of the North Atlantic (Shelbourne, 1965). From 1900 to 1950, most early life history studies were conducted on North Atlantic fishes, expanding beyond descriptive work, rearing experiments, and release of young larvae (see Ahlstrom and Moser, 1981). The amount of such research was related to harvest by region. Based on catches in 1938 and 1948, 730 Fishery Bulletin 92(4), 1994 ~ T3 2 CO Q> "± cu QJ C 5 S .2 S •J3 *3 1 S3 M - X I- t> — CO CN -i - 3! 00 CC IC r 3 cZ c o X. *ir "c CO re CO c '1- lO c lO q o tr- in = q in q -5 2 -r -r CO — — ^ ce- re -T — ee 'C cc t> r^ — i c o T3 •a j- to CU » » c "1 .§ ■-§ ia ■* re CN : ] M CN CN N ?! ce CN N CN CN CN >> X 3 i C Z o | of 2; « to x> > h? -a fc- CO ^ ?s _« S CD B « s ■" o CN — CO -a N Ol C CC g - ca cu 0) CJ a J2 'C o s s o co CU Cu ■ - -ib T3 C 3 ho C CO ""^ ca — S ,2 co .Si cu c*- ■£ 'S O P-. Z -r X CN ~. / ■ X c s h 00 a ■_ OJ -9 3a H3 EX. ea B. cu 0 Ox. 3 "> 3 a3 15 a X co 0) 0) ft co GO 2 OS 1h C cu cu -5h O ■£ c .2 2 < CO X t^ CC - ce CN -T s 3 I 3 3 B 0) QJ >> . 1 r~_' XI X C*H _ "3 o c O - !*H 03 Js QJ S Table 3 :han hal c o ■a CO < CO S 0 •a CU m CN B ; CU o T3 C CC IS en o ca - - U 0 N •a h en s -4-0 0J 3 ai -tJ 0/ c ¥ l< -*u CO a 9 . - CC '5 to 1 CO Eh 05 ca ai a; ■- z< 03 « 03 - M r; r-4 £ 3 3 tl 3 ■.■; c3 eu ca ea g t^ 00 °^ i i | "£ :5 'c g « O eS 0 c5 t£ > > c^ b >. ■-T, o &i c_> CO 3 f eg -O GO IS ^* — cu CT3 DO co" c as 0 cu ■d "3 • ■• — _-z .2 ca "-1 CO s- CN ce X : • 3 3 3 ■a g 00 c a : < - o t! — • o CO A X CU CO CU ~z c c T3 3 A X u rr: IS t5 cu -- CO 3 Cj o s — 3 CU s ■a cu CO — - t o. - — X z- z CU CO — 5 V - XI c c cu S ■a c a u - -. -3 /. X ; S5 cu — -3 — — CO 6 3 CU :0 — cu CO IS - 3 — 5 z *- 1 / 1 CJ 1c _£ z u u cu CO 5 IS C - ca c *u - l a CO o 3 >i 5 CJ >. a. o c o. J3 a fe X X 'J •_ '/. uC X O 3 0 ^: < _ u- Kendall and Matarese: Early life history descriptions of marine teleosts 731 the total world catch of all marine species at this time was less than 200 million metric tons ( t). Among regions, the Northeast Atlantic ranked first in catches, followed by the western central Pacific, the Northwest Pacific, the Northwest Atlantic, and the western central Atlantic, which together accounted for over 75% of the total catch (Table 2). With the goal of understanding stock fluctuations of North- east Atlantic fishes, ichthyoplankton studies were conducted largely under the auspices of the Interna- tional Council for the Exploration of the Sea and this effort is reflected in the state of knowledge for that region. About the same time, similar research was beginning in the Northwest Atlantic. Equally pro- ductive research, following work that began in the late 1800's, was being conducted on fishes of the Mediterranean Sea. Early life history studies were also emerging in Japan (where local fisheries have always been vital) with ichthyoplankton surveys be- ginning in 1938. Work that included descriptions of fish eggs and larvae was also being conducted in In- dia. The western central Pacific and Atlantic contrib- uted significantly to the world catch, and in spite of the immense numbers of species in these regions, progress was made in describing larval fishes there. Larvae of many groups of oceanic fishes were de- scribed based largely on collections of the worldwide Dana expeditions and published as Dana Reports. This work resulted partially from interest in the far- reaching early life history of freshwater ee\s,Anguilla spp., which were important to the economy of north- ern Europe. Since 1950, early life history knowledge has in- creased significantly in several regions and is still roughly in proportion to the activity of the fisheries in the regions and countries involved. The total world catch during this period increased to over 800 mil- lion t, and the Northwest Pacific replaced the North- east Atlantic as the most productive region (Table 2). The Southeast Pacific moved from a rank of 11 to 4 among the 14 regions. Earlier, the North Pacific was not divided into east and west portions, but since 1950 the Northeast Pacific was ranked 7, although the catch here was only 15% of that in the North- west Pacific. Since 1950, Russian scientists have documented early life histories of a wide variety of fishes worldwide as a result of the activity of their distant-water fishing and research fleets. During this time, many fish eggs and larvae from the Northwest Pacific, particularly from waters around Japan, were described. Larvae of most of the fishes of the depau- perate Antarctic fauna have been identified as a re- sult of scientific interest and developing international fisheries in the region, although the catches in this region are still insignificant when compared with other regions. Ichthyoplankton studies in the North- east Pacific initially concentrated on a few commer- cially important species (e.g. Sardinops sagax, Scomber japonicus, and Hippoglossus stenolepis). The creation of the California Cooperative Oceanic Fish- eries Investigations (CalCOFI) in 1947, which tar- geted research on the Pacific sardine, Sardinops sagax, was an important step for egg and larval iden- tification in the California Current region of the Northeast Pacific. Although only 3% of marine spe- cies worldwide are expected to spawn in the North- east Pacific, some species there are very abundant and support large commercial fisheries. Except for the speciose genus Sebastes, early life history stages of most commercially important fishes in the region are now well known. Recent catches in the South- east Pacific ranked fourth among FAO regions world- wide; however, little early life history work has been conducted in this region and larvae of most fishes remain undescribed. Key researchers Our use of the bibliography in Moser et al. ( 1984 ) to develop a key author index means that some impor- tant scientists may not have been included. However, based on the key author index we developed, the con- tributions of authors who produced multiple descrip- tions of eggs and larvae of marine fishes have influ- enced the early life history knowledge of particular geographic regions in terms of the proportion of spe- cies whose larvae are known. The key author index for papers produced prior to 1950 is highest (>0.06) in regions where the percent larvae known is also highest (Table 2): Northeast Atlantic (82%), North- west Atlantic (71%), and Mediterranean (63%). Re- searchers on eggs and larvae before 1950 in the Northeast Atlantic included Ehrenbaum, Holt, Mcin- tosh, Schmidt, and Petersen (Table 4). In the North- west Atlantic, early researchers included Agassiz, Breder, Kuntz, and Hildebrand. The Mediterranean has a long history of early life history research where eight key authors including DAncona, Padoa, Sanzo, and Sparta were identified for publications before 1950. Since 1950, the only region with a key author in- dex >0.05 is the Northwest Atlantic, although the index may be inflated because the number of species in the region is probably underestimated. Active re- searchers in the Northwest Atlantic since 1950 in- clude Eldred, Evseenko, Houde, Leiby, and Smith (Table 4). Other regions where key authors have made notable contributions since 1950 are Japan, the Southeast Atlantic, Northeast Pacific, and the Antarctic. The relatively high taxonomic diversity of 732 Fishery Bulletin 92(4), 1994 Table 4 Senior authors of four or more publications with original early life history descriptions, or of regional compi lations of early life history descriptions based on citations in Moser et al. ( 1984). Senior Total no Senior Total no Region author papers <1950 >1950 Region author papers <1950 >1950 Oceanic Belyanina, T. N. 7 7 NE Pacific Ahlstrom, E. H. 8 8 and Bertelsen, E. L3 13 Moser, H. G. 8 8 >1 region Castle, P. H. J. 6 6 Richardson, S. L. 7 7 Dekhnik, T. V. 4 4 Totals 23 0 23 Efremenko, V. N. 4 4 Number of species 592 Ege, V. 5 2 3 Key author index 0.039 0.000 0.039 Gorbunova, N. N. 7 7 Percent larvae known 44 Pertseva- Ostroumova, T. A Rass, T. S. Totals 11 4 61 1 3 11 3 58 Japan Amaoka, K. Dotsu, Y. Fugita, S. Kobayashi, K. Minami, T 1(1 11 14 7 5 10 11 14 7 5 Mediterranean Bertolini, F. 2 2 Mito, S. 20 20 D'Ancona, U. 8 7 1 Okiyama, M. 18 18 Fage, L. 2 2 Shiogaki, M. 13 13 Padoa, E. 12 1 11 Suzuki, K. 9 9 Raffaele, F. 1 1 Takita, T 7 7 Roule, L. 2 2 Uchida, K. 4 2 2 Sanzo, L. 29 28 1 Yusa, T. 9 9 Sparta, A. 27 20 7 Totals 127 2 125 Totals 83 63 20 Number of species 3,500 Number of species 569 Key author index 0.036 0.001 0.036 Key author index 0.146 0.111 0.035 Percent larvae known 37 Percent larvae known 63 Indo-Pacific Delsman, H. C. Fourmanoir, P. 22 5 22 5 NE Atlantic Clark, R. S. 2 2 Jones, S. 9 2 7 Cunningham, J. T 3 3 Leis, J. M. 7 7 Ehrenbaum, E. 5 5 Miller, J. M. 2 2 Holt, E. W. L. 4 4 Robertson, D. A. 6 6 Lebour, M. V. 3 3 Totals 51 24 27 Mcintosh, W. C. 4 4 Number of species 3,921 Petersen, C. G. J. ti 6 Key author index 0.013 0.006 0.007 Russell. F S. 1 1 Percent larvae known 10 Schmidt, J. 12 12 Totals 40 39 1 WC Atlantic Beebe, W. 5 5 Number of species 131 Richards, W. J. 5 5 Key author index 0.305 0.298 0.008 Totals 10 5 5 Percent larvae known 82 Number of species 1,81)3 Key author index 0.006 0.003 0.003 Percent larvae known 27 NW Atlantic Agassiz, A. 2 2 Breder, C. M., Jr. 7 ti 1 SE Atlantic Aboussouan, A. 11 11 Dannevig, A. 1 ] Brownell, C. L. 1 1 Eldred, B. 6 6 Gilchrist, J. D. F 4 1 Evseenko, S. A. 5 5 Totals 16 1 12 Fahay, M. P. 3 3 Number of species 239 Hildebrand, S. F. 3 3 Key author index 0.067 0.017 0.050 Houde, E. D. 7 7 Percent larvae known 59 Kuntz, A. 1 1 Lei by. M. M. 6 6 SW Atlantic de Ciechomski, J. D. 8 8 Smith. D. G. 9 9 Totals 53 16 37 Antarctic Yefremenko, V. N. 7 7 Number of species 317 Number of species 158 Key author index 0.167 0.050 0.117 Key author index 0.044 0.000 0.044 Percent larvae known 64 Percent larvae known 51 Kendall and Matarese: Early life history descriptions of marine teleosts 733 fishes in Japanese waters accounts for its low key author index (0.036), but the Japanese probably have been the most prolific scientists working on early life history descriptions since 1950; 125 papers by Japa- nese key authors were cited in Moser et al. (1984). Even so, Leis (1985) has indicated that the bibliog- raphy of Moser et al. (1984) underrepresents the con- tributions of Japanese scientists. Two early standouts among the Japanese were Uchida and Mito, but re- cent work by Dotsu, Yusa, Amaoka, Okiyama, Ozawa and Fujita has significantly increased the knowledge of the region. Relatively low taxonomic diversity and an abundance of recent work by Olivar, among oth- ers, published since 1986 contributes to a rather high (0.088) key author index for the Southeast Atlantic. In the California Current region the CalCOFI pro- gram stimulated early life history descriptions that resulted in many publications by Ahlstrom and Moser (1981) and their coauthors. The Russians, as a re- sult of the activity of their far-reaching fishing ef- forts since 1950, have contributed to a more com- plete understanding of early life history of eggs and larvae in several regions through the work of key authors such as Belyanina, Deknik, Gorbunova, and Rass. Also, several Russians, such as Pertseva-Ostrou- mova as well as non-Russians such as Bertlesen, Castle, and Ege have published papers on early life histories of oceanic groups. A graph of the cumulative key author index plot- ted against the percent of the number of species oc- curring in various regions whose larvae are known, indicates that in all regions where >60% of the larvae are known, the key author index is >0.1 (Fig. 2). Re- gions with low key author indices (Japan, western cen- tral Atlantic, and Indo-Pacific) also have low percent- ages of known larvae (<40%). However, these regions also have large numbers of species present (Table 2). Scientific interest and taxonomic diversity Richards ( 1985 ) points out that early life histories of commercially important groups (e.g. herrings, salmons [salmonids], tunas, flatfishes [pleuro- nectiforms], and cods) have been the subject of a dis- proportionately greater number of studies and are thus better known than groups that are not the ob- jects of large commercial fisheries. Significant inter- est in the systematics of particular taxa (e.g. cods in the Northeast Pacific) or the recruitment of particu- lar species (e.g. Gadus morhua, Theragra chalco- gramma, Clupea harengus, and Scomber scombrus) can lead to increased knowledge of the general early o c CD co c CD O i— CD Q- 80 60 ^ 40 20- 0 ONEA OSEA ONWA OMED OANT ONEP OJAP OWCA OIPC i i i i i i 0.1 0.1 0.2 0.2 Key author index 0.3 0.3 0.4 Figure 2 Key author index (see text) plotted against percent of species in geographic regions where larvae are known. ANT=Antarctic; IPC=Indo-Pacific; JAP=Japan; MED=Mediterranean; NEA=Northeast Atlantic; NWA=North west Atlantic; NEP=Northeast Pacific; SEA=Southeast Atlantic; WCA=western central Atlantic. 734 Fishery Bulletin 92(4), 1994 life history of fishes for a region. The taxonomic diversity of Indo-Pacific coral reefs and Japanese waters is much greater than that found at higher latitudes, resulting in lower proportions of species with identified larvae in these and other low- latitude regions. Speciose perciform families present at lower latitudes contribute to the difficulty of iden- tifying larvae in these regions. In other regions, taxo- nomic groups that have undergone extensive radia- tion complicate identification of larvae. For example, about 70 species ofSebastes are present in the North- east Pacific and they cannot be identified routinely in plankton samples (Matarese et al., 1989). The lar- vae of some closely related taxa in commercially im- portant groups in other regions have proven very difficult to identify (e.g. tunas [scombrids], some her- rings [clupeids], and North Atlantic cods [gadids]). Conclusions In spite of the relatively large proportions of fishes in some regions for which some early life history stages have been illustrated, identification problems still limit the usefulness of ichthyoplankton studies. For example, in the Northeast Pacific where identi- fication is possible only to family or genus for sev- eral groups (e.g. Sebastes spp., cottids, agonids, and stichaeids), more descriptive work remains to be done. It is ironic that some groups containing some of the world's most important fisheries (e.g. tunas, cods, and herrings) also pose some of the more diffi- cult problems regarding egg and larval identification. Research involving field studies offish eggs and lar- vae in several parts of the world is now concentrat- ing on recruitment dynamics of commercially impor- tant species, whose early stages are well described. Two of the goals of the Ahlstrom Symposium held in 1983 were to accumulate information on fish de- velopment by taxa, and thus stimulate additional research on poorly known groups, and to highlight the potential usefulness of developmental informa- tion in systematic studies. Since the volume based on the Ahlstrom Symposium was published (Moser et al., 1984), many important original descriptive pa- pers have appeared (e.g. Ditty, 1989; Fahay, 1992), but it does not seem that there has been a signifi- cant increase in the number of larvae known. Rather than an increase in original descriptions, the late 1980s saw the publication of several regional guides to fish early life history (see Table 2). While some recent systematic studies have considered larval as well as adult characters (e.g. Cohen, 1989; Baldwin, 1990; Baldwin and Johnson, 1993; Strauss, 1993), there are still unresolved theoretical problems with this approach. Some ichthyologists still do not want to deal with those "unidentifiable pinheads" (Winter- bottom, 1986), and rigorous analysis of developmen- tal, in addition to adult, characters can be a daunt- ing task. According to Johnson (1993): "Almost 10 years after its [i.e. Moser et al., 1984] publication the historical separation between studies of early life history stages and 'mainstream' systematic ichth- yology appears only slightly diminished. Most com- parative osteological and phylogenetic studies of fishes do not incorporate development and thus ig- nore the potential for additional suites of characters and for testing homology." A combination of rearing studies and developing series from plankton samples as well as more inno- vative techniques such as biochemical genetics (Seeb and Kendall, 1991) will be required to fill the gaps in our knowledge on the identification of early de- velopmental stages of marine fishes. Although the value of egg and larval studies are recognized in fish- eries science, their usefulness will probably remain limited without the continued efforts of scientists who often describe early life stages as ancillary but enjoy- able endeavors. Acknowledgments Deborah Blood (AFSC) provided helpful suggestions and ideas that were incorporated into early drafts of the manuscript. Morgan Busby (AFSC) assisted with gathering and interpreting the literature we used. Also, several of our colleagues encouraged us in this exercise and provided helpful reviews of earlier ver- sions of the manuscript: Bill Richards, Southeast Fisheries Science Center, Miami, FL; Geoff Moser, Southwest Fisheries Science Center, La Jolla, CA; and Mike Fahay, Northeast Fisheries Science Cen- ter, Sandy Hook, NJ, who actually commented on two versions of the manuscript. Literature cited Aboussouan, A. 1989. L'identification des larvae de poissons de la mer Mediterranee. Cybium 13:259-262. Agassiz, A. 1882. On the young stages of some osseous fishes. Proc. Am. Acad. Arts Sci. 16 (Pt. 3):271-303. Ahlstrom, E. H., and O. P. Ball. 1954. Description of eggs and larvae of jack mack- erel {Trachurus symmetricus) and distribution and abundance of larvae in 1950 and 1951. U.S. Fish Wildl. Serv., Fish. Bull. 56:209-245. Kendall and Matarese: Early life history descriptions of marine teleosts 735 Ah 1st rom, E. H., and H. G. Moser. 1981. Systematics and development of early life his- tory stages of marine fishes: achievements during the past century, present status and suggestions for the future. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 178:541-546. All 1st rom, E. H., J. L. Butler, and B. Y. Sumida. 1976. Pelagic stromateoid fishes (Pisces, Perci- formes) of the eastern Pacific: kinds, distributions and early life histories and observations on five of these from the northwest Atlantic. Bull. Mar. Sci. 26:285^102. Auer, N. A. (ed.). 1982. Identification of larval fishes of the Great Lakes Basin with emphasis on the Lake Michigan drainage. Great Lakes Fish. Comm. Publ. 82-3, 744 p. Baldwin, C. C. 1990. Morphology of the American Anthiinae (Teleostei: Serranidae), with comments on relation- ships within the subfamily. Copeia 1990:913-955. Baldwin, C. C, and G. D. Johnson. 1993. Phylogeny of the Epinephelinae (Teleostei: Serranidae). Bull. Mar. Sci. 52:240-283. Belyanina, T. N. 1984. Larvae of hatchetfishes of the genus Argyro- pelecus (Sternoptychidae). J. Ichthyol. 24(2):7-20. Bertelsen, E. 1951. The ceratioid fishes: ontogeny, taxonomy, dis- tribution, and biology. Dana-Rep. Carlsberg Found. 39, 276 p. Brownell, C. L. 1979. Stages in the early development of 40 marine fish species with pelagic eggs from the Cape of Good Hope. Ichthyol. Bull J. L. B. Smith Inst. Ichthyol. 40. Busby, M. S., and D. A. Ambrose. 1993. Development of larval and early juvenile pygmy poacher, Xeneretmus trispinosa, and blacktip poacher, Xeneretmus latifrons, (Scorpaeni- formes: Agonidae). Fish. Bull. 91:397^113. Cohen, D. M. (ed.). 1989. Papers on the systematics of gadiform fishes. Nat. Hist. Mus. Los Angeles Co., Sci. Ser. No. 32, 262 p. D'Ancona, U., and G. Cavinato. 1965. The fishes of the family Bregmacer- otidae. Dana-Rep. Carlsberg Found. 64, 92 p. D'Ancona, U., et al. 1931-1956. Uova, larve, e stadi giovanili de Teleostei. Staz. Zool. Napoli. Fauna e Flora de Gol- fo di Napoli. Publ. in 4 parts, 1064 p. Ditty, J. G. 1989. Separating early larvae of sciaenids from the western North Atlantic: a review and comparison of larvae off Louisiana and Atlantic coast of the U.S. Bull. Mar. Sci. 44:1083-1105. Ditty, J. G., and R. F. Shaw. 1992. Larval development, distribution, and ecology of cobia Rachycentron canadum (Family: Rachy- centridae ) in the northern Gulf of Mexico. Fish. Bull. 90:668-677. Ehrenbaum, E. 1905. Eier und larven von fischen des Nordischen planktons. Verlag von Lipsius und Tischer, Kiel und Leipzig 1:1-216. 1909. Eier und larven von fischen des Nordischen planktons. Verlag von Lipsius und Tischer, Kiel und Leipzig 2:217-413. Fahay, M. P. 1983. Guide to the early stages of marine fishes oc- curring in the western North Atlantic Ocean, Cape Hatteras to the southern Scotian Shelf. J. North- west Atl. Fish. Sci. 4, 423 p. 1992. Development and distribution of cusk eel eggs and larvae in the Middle Atlantic Bight with a de- scription of Ophidion robinsi n. sp. (Teleostei: Ophidiidae). Copeia 1992:799-819. 1993. The early life history stages of New Jersey's saltwater fishes: sources of information. Bull. N. J.Acad. Sci. 38:1-16 Food and Agriculture Organization of the United Nations. 1965, 1974, 1979, 1984, 1991, 1992. Yearbook of fisheries statistics, 18, 36, 46, 56, 68, 70. FAO, Rome. Fritzsche, R. A. 1978. Development of fishes of the mid-Atlantic Bight, an atlas of egg, larval, and juvenile stages. Vol. V: Chaetodontidae through Ophidiidae. U. S. Fish. Wildl. Serv., Biol. Prog. FES/OBS-78/12. Hardy, J. D., Jr. 1978a. Development of fishes of the mid-Atlantic Bight, an atlas of egg, larval, and juvenile stages. Vol. II: Anguillidae through Syngnathidae. U. S. Fish. Wildl. Serv., Biol. Prog. FES/OBS-78/12. 1978b. Development of fishes of the mid-Atlantic Bight, an atlas of egg, larval, and juvenile stages. Vol. Ill: Aphredoderidae through Rachycentri- dae. U. S. Fish. Wildl. Serv., Biol. Prog. FES/OBS- 78/12. Hempel, G. 1979. Early life history of marine fish: the egg stage. Univ. Washington Press, Seattle, 70 p. Johnson, G. D. 1978. Development of fishes of the mid-Atlantic Bight, an atlas of egg, larval, and juvenile stages. Vol. IV: Carangidae through Ephippidae. U.S. Fish. Wildl. Serv., Biol. Prog. FES/OBS-78/12. 1993. Percomorph phylogeny: progress and prob- lems. Bull. Mar. Sci. 52:3-28. Johnson, R. K. 1974. A revision of the alepisauroid family Scope- larchidae (Pisces: Myctophiformesl. Fieldiana Zool. 66. Jones, P. W., F. D. Martin, and J. D. Hardy Jr. 1978. Development of fishes of the mid-Atlantic Bight, an atlas of egg, larval, and juvenile stages. Vol. I: Acipenseridae through Ictaluridae. U.S. Fish. Wildl. Serv., Biol. Prog. FES/OBS-78/12. 736 Fishery Bulletin 92(4), 1994 Kellermann, A. (ed.). 1989. Identification key and catalogue of larval Ant- arctic fishes. Biomass Scientific Ser. No. 10., 136 p. Kendall, A. W., Jr., and B. Vinter. 1984. Development of hexagrammids (Pisces, Scor- paeniformes) in the northeastern Pacific Ocean. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 2, 44 p. Leis, J. M. 1985. Review of "The bibliography on the identifi- cation of the eggs, larvae and juveniles of the ma- rine shorefishes of Japan." Am. Fish. Soc. Early Life Hist. Newsletter 6 (31:11-13. Leis, J. M., and D. S. Rennis. 1983. The larvae of Indo-Pacific coral reef fishes. New South Wales Univ. Press, Sydney, 269 p. Leis, J. M., and T. Trnski. 1989. The larvae of Indo-Pacific shorefishes. Univ. Hawaii Press, Honolulu, 371 p. Maeda, K., and K. Amaoka. 1988. Taxonomic study on larvae and juveniles of agonid fishes in Japan. Mem. Fac. Fish. Hokkaido Univ. 35:47-124. Martin, F. D., and G. E. Drewry. 1978. Development of fishes of the mid-Atlantic Bight, an atlas of egg, larval, and juvenile stages. Vol. VI: Stromateidae through Ogcocepha- lidae. U.S. Fish. Wildl. Serv., Biol. Prog. FES/ OBS-78/12. Matarese, A. C, A. W. Kendall Jr., D. M. Blood, and B. M. Vinter. 1989. Laboratory guide to early life history stages of Northeast Pacific fishes. Dep. Commer., NOAA Tech. Rep. NMFS 80, 652 p. Matsui, T. 1991. Description of young of the mesopelagic platytroctids Holtbyrnia latifrons and Sagam- ichthys abei (Pisces, Alepocephaloidea) from the Northeast Pacific Ocean. Fish. Bull. 89:209-219. Miller, J. M., W. Watson, and J. M. Leis. 1979. An atlas of common nearshore marine fish larvae of the Hawaiian Islands. Sea Grant Misc. Rep. Univ. Hawaii Sea Grant MR-80-02, 179 p. Moser, H. G., W. J. Richards, D. M. Cohen, M. P. Fahay, A. W. Kendall Jr., and S. L. Richardson (eds.). 1984. Ontogeny and systematics of fishes. Am. Soc. Ichthyol. Herpetol., Spec. Pub. 1. Okiyama, M. (ed.). 1988. An atlas of the early stage fishes in Japan. Tokai Univ. Press, Tokyo, 1154 p. Olivar, M. P., and J. M. Fortuno. 1991. Guide to ichthyoplankton of the southeast Atlantic (Benguela Current region). Sci. Mar. 55 (1): 1-383. Ozawa, T. 1986. Studies on the oceanic ichthyoplankton in the western North Pacific. Kyushu Univ. Press, 430 p. Paxton, J. R., D. F. Hoese, G. R. Hoese, G. R. Allen, and J. E. Hanley. 1989. Zoological catalogue of Australia. Vol. 7: Pi- sces, Petromyzontidae to Carangidae. Aust. Gov. Pub. Sevr., Canberra, 664 p. Potthoff, T., W. J. Richards, and S. Ueyanagi. 1980. Development of Scombrolabrax heterolepis (Pisces, Scombrolabracidae) and comments on fa- milial relationships. Bull. Mar. Sci. 30:329-357. Richards, W. J. 1985. Status of the identification of the early life stages of fishes. Bull. Mar. Sci. 37:756-760. 1990. List of the fishes of the western central At- lantic and the status of early life history infor- mation. U.S. Dep. Commer., NOAA Tech. Memo NMFS-SEFC-267, 88 p. Rothschild, B. J. 1986. Dynamics of marine fish populations. Har- vard Univ. Press, 277 p. Russell, F. S. 1976. The eggs and planktonic stages of British marine fishes. Academic Press, London, 524 p. Schaefer, M. B., and D. L. Alverson. 1968. World fish potentials. In Fish (new ser. 4), p. 81-85. Univ. Washington Publ. Seeb, L. W., and A. W. Kendall Jr. 1991. Allozyme polymorphisms permit the identifi- cation of larval and juvenile rockfishes of the ge- nus Sebastes. Environ. Biol. Fish. 30:173-190. Shelbourne, J. E. 1965. Rearing marine fish for commercial pur- poses. Calif. Coop. Oceanic Fish. Invest. Rep. 10:53-63. Strauss, R. E. 1993. Relationships among the cottid genera Arte- dius, Clinocottus, and Oligocottus (Teleostei: Scor- paeniformes). Copeia 1993:518-522. Wang, J. C. 1986. Fishes of the Sacramento-San Joaquin estu- ary and adjacent waters, California: a guide to the early life histories. Interagency Ecological Study Program for the Sacramento-San Joaquin Estuary. Tech. Rep. 9 (Available from Ecological Analysts, Inc., 2150 John Glenn Drive, Concord, CA 94520). Winterbottom, R. 1986. Review of "Ontogeny and systematics of fishes." Bull. Mar. Sci. 39:141-143. Yabe, M. 1991. Bolinia euryptera, a new genus and species of sculpin (Scorpaeniformes: Cottidae) from the Bering Sea. Copeia 1991:329-339. Abstract. — The catch and effort of reef fisheries in seven areas of Belize and in six of south Jamaica were intensively surveyed to pro- vide data for area-based surplus- production models (SPM) to man- age these fisheries. Data were nor- malized to area of productive habi- tat. SPM's could not be defined for the Belizean or Jamaican data treated separately because the slopes of the relationships between catch per unit of effort and effort were nonsignificant and positive. This appeared to be due 1 ) to vio- lations of the model's assumptions (catch composition was heteroge- neous because fishermen target spawning aggregations and migra- tory fishes at particular sites) and 2) to possible differences in commu- nity composition among areas (the communities were not at equilib- rium and productivity possibly dif- fered among sites). Other assump- tions had been violated by previ- ous area-based SPM's so that the level of exploitation on the south Jamaican shelf has been seriously underestimated in recent decades. Although a SPM could be defined for the combined Jamaica-Belize data set, we conclude that these models should be used with caution in reef fisheries management be- cause underlying assumptions are likely to be seriously violated. The surveys indicate, however, that lev- els of catch per unit of effort, catch, and effort in the south Jamaican reef fishery are significantly lower than those of 10 years ago. Deple- tion of a wide range of fish groups has apparently led to a decline in the equilibrium productivity of the fishery. Catch and effort analysis of the reef fisheries of Jamaica and Belize Julian A. Koslow CSIRO Division of Fisheries, Marine Laboratories GPO Box 1538, Hobart, Tasmania 7001 Australia Karl Aiken Zoology Department, University of the West Indies Kingston, Jamaica Stephanie Auil Fisheries Unit, Princess Margaret Drive Belize City, Belize Antoinette Clementson Zoology Department, University of the West Indies Kingston, Jamaica Manuscript accepted 5 May 1994. Fishery Bulletin 92:737-747. Tropical coral reef fisheries are typi- cally small scale but highly com- plex, artisanal multispecies fisher- ies. They are often overexploited (Munro, 1983; Koslow et al., 1988; Russ, 1991) but are rarely managed with conventional fishery methods (see Johannes [1978] on traditional management of reef fisheries). Con- ventional fishery models are not particularly suitable for complex, multispecies fisheries, and the req- uisite data, because of the highly decentralized landing and market- ing systems typical of these fisher- ies, is often difficult to obtain. The regions supporting these fisheries often lack the technical and finan- cial resources to manage them, and even if the resources were available, it is arguable whether these small- scale fisheries would justify the ex- penditure that such an exercise would require. However, although the overall yield of these fisheries is modest, they may provide an im- portant source of employment, pro- tein, and foreign exchange earnings for local economies. In a pioneering study based on catch and effort data collected dur- ing a single survey of landing sites that he grouped by coastal parishes, Munro (1978) developed a prelimi- nary surplus production model (SPM) for Jamaican reef fisheries. His approach was attractive because data inputs and analytic require- ments were modest, and the model provided long-term, albeit simple, guidance for optimal fishing levels. However, the area-based SPM assumes that the fish assemblages, their habitats and productivity, and the fishery do not differ signifi- cantly among fishing areas; that the relationship between catch and ef- fort is at equilibrium in each area; and that the fish stocks and effort are contained within the designated fishing areas (Caddy and Garcia, 1982; Nicholson and Hartsuijker, 1982). Nicholson and Hartsuijker ( 1982 ), in particular, pointed out the perils of violating the model's as- sumptions, but area-based SPM's for reef fisheries are being used in- creasingly to obtain first-order ap- proximations of maximum sustain- able yield based upon available catch and effort data (Aiken and Haught- on, 1987; Haughton, 1988; Appel- 737 738 Fishery Bulletin 92(4), 1994 doom and Meyers, in press). Area-based SPM's have not been developed further for management of reef fisheries, although more focussed studies have led to interesting results in freshwater systems (Mar- ten, 1979). Although both are within the Caribbean region, the reef fisheries of Belize and Jamaica contrast markedly. Belizean finfish stocks appear to be lightly to moderately exploited, an assessment not based on quantitative catch and effort data, which have never been systematically collected, but upon the contin- ued availability of prime commercial species (snap- pers [Lutjanidae] and groupers [Serranidae]) that are the basis of an export-oriented fishery. This assess- ment is also based on the country's estimated low consumption of seafood; Belize has a sparse popula- tion (7.8 persons/km2 totalling less than 200,000 per- sons) and has traditionally relied little on seafood. Conch (Strombus gigas) and lobster (Panulirus argus) are the main focus of Belizean commercial fisheries, followed by snapper and grouper, which are fished primarily for export. In contrast, seafood is traditionally an important part of the Jamaican diet; the country is densely populated (216 persons/km2 with a total population of 2,362,000), and its coastal fisheries have been heavily exploited for at least the past several decades (Aiken and Haughton, 1987). Since 1970, catch rates in the reef fisheries have markedly declined (Aiken and Haughton, 1987; Haughton, 1988), and the catch composition has shifted to commercially less valu- able species (Koslow et al., 1988). Snappers, grou- pers, and large parrotfishes (Scaridae) that were abundant off Jamaica in the last century (Gosse, 1851 ) have virtually disappeared from most reef areas. Our objective was to develop a SPM to manage the reef fisheries of Jamaica and Belize. To improve upon previous area-based SPM's, we carried out focussed surveys of catch and effort to better quantify the model in relation to some of its underlying assump- tions. In particular, we assessed the productive area underlying each fishery by estimating the propor- tion of productive reef habitat in different parts of the shelf and by localizing the fishing grounds used, and we quantified annual fishing effort. By survey- ing reef fisheries in these two countries, we hoped to relate catch and effort over a range of exploitation rates and develop a broadly applicable SPM. Methods Field study A two-phase survey was carried out in Belize and along the south coast of Jamaica (Fig. 1, A and B). First, a stratified systematic survey was carried out to determine the numbers of fishermen by region, the types of vessels and gears in use, and the grounds fished, and to obtain general information on effort, catch, and seasonality in catch composition and abun- dance. Validated lists of fishing vessels in the two countries were obtained from the licensing registers of fisheries departments and from surveys of land- ing sites (20 active fishing beaches on the south coast of Jamaica and 10 cooperatives and markets in Belize). The lists were stratified by area and a sample from each area was systematically selected. A ques- tionnaire was administered to the selected fishermen in Belize, but owing to difficulties in locating selected fishermen in Jamaica, a number of fishermen were chosen from those available on the fishing beaches. Based on this survey, the fishing grounds were subdivided into seven areas in Belize and six in Ja- maica. However, spawning aggregations fished in several of the areas in Belize accounted for a signifi- cant proportion of the fish landings and seemed likely to draw fish from nearby areas. In calculating the SPM, data from areas 4 and 5 (east and west Amber- gris Cay) were pooled, as were data from areas 2, 3, and 7 (Fig. IB). In the second phase of the survey, six landing sites in Jamaica and five in Belize were visited to collect data on effort and landings over an annual cycle. Thus one site that was deemed representative was selected from each area, except in Belize City (Gal- lows Pt.) area, where two cooperatives were visited. Sites were visited every two weeks in Belize between July 1990 and August 1991 (except Placencia, which was sampled from March through August 1991) and in Jamaica from February through April and August through November 1991. Sites were monitored for the entire period during which fish were landed. As each vessel landed its catch, overall weights were recorded by family, and fishermen were interviewed to ascertain the gears used, the effort by gear-type, and the areas fished. In Jamaica, there were too many vessels at some sites to monitor all landings. In these instances, total effort and landings statis- tics for the site were estimated by the proportion of vessels actually surveyed: XT = XJF, where XT is the total landings or effort for a site on a particular day, X is the landings or effort recorded, and F is the proportion of vessels surveyed. The effective area of the fishing grounds in each area was estimated. The total area of the shelf was estimated from charts both with a planimeter and by weight, whereby the shelf area was traced from a chart, cut out, weighed, and the weight related to that of a unit area (e.g. 10 km2). The extent of the actual fishing grounds was determined from inter- views conducted during the surveys. In Belize, the Koslow et al.: Catch and effort analysis of reef fisheries 739 ^^""l JAMAICA L*-1^ C\J Scott's Cove | BLACK RIVER / Greenwich | - Town J JKINGSTON I MORANT ^ POINT Farquhars Beach [Old Hartjour Bay [ J^S-^w ^r\ /■• 1 Morant Bay[ 3 .. 0 16 km • Survey Sites Figure 1 Maps of (Al survey sites and the six fishing areas along the south coast of Jamaica and (B) the seven fishing areas in Belize. 740 Fishery Bulletin 92(4), 1994 fishermen noted their fishing grounds on a chart in relation to the cays, and the fishing grounds were assumed to extend to the reef crest. The areas of these grounds were then measured with a planimeter. In Jamaica, four line transects orthogonal to the shoreline from nearshore to the shelf edge were car- ried out in each of three areas: Old Harbour Bay, Farquhars Beach, and Great Bay (Fig. 1A). Transects within each area were approximately 2 nmi (=3.7 km) apart. The mean depth of the south Jamaican shelf is 20 m ( Woodley and Robinson, 1977), and the dropoff is at about 50 m (Nicholson and Hartsuijker, 1982); observations of bottom type were made with a glass- bottomed viewing box deployed over the side of a small vessel. Observations were carried out at 4-km intervals of the dominant substrate material (i.e. sand, grass, coral, or mud). The proportion of shelf represented by each substrate type was estimated from the proportion of stations at which the particu- lar substrate type was dominant. The results of these surveys were compared with historical surveys of the Jamaican shelf (Nicholson and Hartsuijker, 1982). Data analysis Catch and effort data were summed by area, gear type, and species group. Totals were standardized by proportions of the fishing year and of the popula- tion of fishermen surveyed. Fishing effort in a SPM must be expressed in a common unit. Hook-and-line effort (hook h/km2) was selected as the common unit of effort because our surveys indicated that it was the dominant fishing gear in Belize (in terms of inci- dence of use and yield obtained) and the most widely used gear overall. Fishing effort from other gear types (i.e. bottom gill net, trap, weir, and spearing) was standardized to hook-and-line gear by using the weighted mean of the ratio of catch rates from the particular gear to the hook-and-line catch rate within each area. Effort was then summed within each area. Catch and effort for each area were standardized per square kilometer of fishing ground. Log-trans- formation of the catch-per-unit-of-effort (CPUE) data (the Fox [19701 variant of the SPM) did not improve the fit, so it is not presented. However, the relation- ship between CPUE and effort (/") was highly nonlin- ear; therefore, the relationship is presented both without transformation and with effort data log- transformed, which linearized the relationship be- tween CPUE and f. Results Catch and effort Annual catch was estimated to be more than four- fold higher off the south coast of Jamaica (998 tonnes) than off the coast of Belize (240 tonnes) (Table 1). When landings were normalized to the area of pro- ductive fishing ground (Tables 1 and 2) (i.e. the por- tion of shelf estimated to be coral and sea grass), yield per unit of area from Jamaican waters (552 kg/km2) was 39% higher than off Belize (340 kg/km2). How- ever, this difference was not statistically significant in a comparison of mean yield from the different fish- ing areas in the two countries (Kruskal-Wallis [KWJ one-way ANOVA: x2=0.33, « = 13,P>0.2). Table 1 Catch (Y) and fishing effort (/) data summary for sites in Bel ze and Jamaica. See Figure 1 for areas. Area (A) Total Y Prime Y Total f y/A f/A Country Site (km2) (tl (t) (000 hook h) (kg/km2 ) (hook h/km2) Belize 1 312 30 26 22 97 71 2 32 7 7 3 20N 86 3 17 31 24 34 655 720 ■1 H» 16 11 15 1,688 1,505 5 33 8 4 9 250 275 6 231 l,s 17 lf> 79 65 7 11 130 126 52 2,929 1,172 Jamaica 1 46 55 8 286 1,197 (1,209 •J 252 34-4 159 2,038 1,364 8,089 3 607 265 51 1,303 437 2,147 4 652 266 L8 1.448 409 2.222 5 115 37 ■1 265 319 2,307 6 135 31 6 257 223 1,901 Totals Belize 709 240 218 150 340 210 Jamaica 1,807 998 246 5,597 552 3,098 Koslow et al.: Catch and effort analysis of reef fisheries 741 Table 2 The shelf area of the south Jamaican shelf and the proportions repre- sented by coral, seagrass, sand, and mud benthic habitat types. The regions are shown in Figure 1A. Region 1 2 3 4 5 6 Total Total shelf area (km2) 127 331 797 1,390 316 372 3,333 Proportion of coral 0.32 0.48 0.48 0.22 0.32 0.32 0.33 Proportion of seagrass 0.05 0.28 0.28 0.25 0.05 0.05 0.21 Proportion of sand 0.32 0.15 0.15 0.34 0.32 0.32 0.27 Proportion of mud 0.32 0.09 0.09 0.19 0.32 0.32 0.19 The fishing effort and catch rates of the two coun- tries differed considerably. The mean fishing effort per unit area on the Jamaican grounds was fifteenfold higher than off Belize: 3,098 hook h/km2yr (equivalent to 527 trap hauls/km2) in Jamaica and 210 hook h/km2-yr in Belize (KW: x2=9.00, P<0.005). However, catch rates were ninefold higher in Belize: 1.61 kg/hook h compared with 0.18 kg/hook h (equiva- lent to 1.06 kg/trap haul) in Jamaica (KW: x2=9.00, P<0.005). fishing grounds (Fig. 2). Landings of the main species groups were approxi- mately log-normally distributed among regions within each country, especially in Belize, where landings per unit of area generally varied among fishing grounds by two to three orders of mag- nitude. In Jamaica the differences were generally closer to two orders of magni- tude. Thus in Belize, landings per unit of area of lutjanids were highest in ar- eas 4 (west Ambergris Cay) and 7 (Placencia); of serranids in areas 3 and 4 (Gallows Point and west Ambergris Cay); and of haemulids in areas 3 and 5 (Gallows Pt. and east Ambergris Cay). Several of these areas were sites of major spawning aggregations (S. Auil, unpubl. data). The outer atolls, Halfmoon Cay and Turneffe (areas 1 and 6), did not appear to be intensively fished for finfish. In Jamaica, lutjanid and haemulid landings were higher in areas 1 and 2; serranid landings were higher in areas 2 and 3. The catches of low-valued fish were more evenly distributed. Catch composition The composition of the fishery also was substantially different in the two countries (Fig. 2). Prime com- mercial fishes from the Lutjanidae (snappers) and Serranidae (groupers) dominated the Belizean fish- ery, representing 74% and 11% of the catch, respec- tively. In contrast, lutjanids represented 23% of the Jamaican catch and serranids only 2%. Of the land- ings in Jamaica, 62% were of low-value species, fishes in the families Scaridae, Sparidae, Labridae, Mullidae, Holocentridae, and Acanthuridae. Another 14% were haemulids, which composed only 2% of the catch in Belize. When the data were aggregated by area, the differences in catch composition between the countries were all significantly different (Table 3), as were the differences in actual catch for all groups except lutjanids. The catch of serranids was significantly higher in Belize and that of haemulids and 'other' fish was higher in Jamaica. When the data were examined on the basis of individual landings, the num- ber of degrees of freedom was greatly increased. Differences were highly signifi- cant for all groups: landings of serranids and lutjanids were again higher in Belize; landings of haemulids and 'other' fishes were higher in Jamaica (Table 3). Within each country, species compo- sition also varied significantly among Surplus production models Because of the heterogeneity of the fishery, we ex- amined catch-effort relationships for species groups, both individually and combined. The slopes of the relationships of CPUE with effort (f) were nonsig- nificant but were positive in sign for the Jamaican and Belizean reef fisheries considered separately (Table 4). When the data for the two countries were combined, the relationship between CPUE and /"was negative (Table 4, Fig. 3A). A linear relationship, obtained after log-transforming the data on f, was due largely to the substantial difference in f and CPUE between the two countries (Fig. 3B). Based upon these relationships, MSY for the total reef fish- eries was estimated to be 1,046 kg/km2 of productive Table 3 Results of Kruskal-Wallis one-way ANOVA to test for differences in catch composition between Jamaica and Belize for the data shown in Figure 2. (A) test for differences in proportion of catch by fish groups aggregated by area (n = 13); (B): test for differences in landings offish groups aggregated by area (n = 13); (C): test for differences in land- ings by individual landing (n=503). The statistic shown is the x2 value. *P< 0.05; **P< 0.01; ***P< 0.001. NS = not significant. Lutjanidae Serranidae Haemulidae Other A B C 7.37** 2.47 NS 19.42*** 6.61* 5.22* 38.2*** 4.00* 4.00* 52.6*** 7.37** 4.00* 12.2*** 742 Fishery Bulletin 92(4), 1994 habitat (sea grass+coral), with an annual fishing ef- fort (f^ ) to obtain MSY of 3,497 hook h/km2 (Fig. 4A). Current fishing effort in Jamaica and Belize (Table 2) is 89% and 6% off r respectively. MSY for the piscivorous fishes (e.g. Serranidae, Lutjanidae, and Sphyraenidae) was estimated to be 638 kg/km2 of productive habitat, which can be caught at f = 2,200 hook h/km2 (Table 4, Fig. 4B). To maximize catch of piscivorous fishes, present fishing effort in Ja- maica and Belize is 141% and 10% off , respectively. Serranidae B B 1 B2 B3 B4 B6 B7 Jl J2 J3 J4 J5 J6 Fishing Area % 200 Haemulidae < mA B 1 B2 B3 B4 35 B6 B7 J Fishing Area 2 J3 J4 J5 J6 B2 B3 84 85 B6 B7 Jl J2 J3 J4 J5 J6 Fishing Area %, 05 I ■-■■ a M B2 B3 B4 B5B6 87 Jl J2 J3 J4 J 5 J6 Fishing Area Figure 2 Annual landings by species groups for seven fishing areas in Belize ( B 1-B7 ) and six areas off south Jamaica (J1-J6); (A) Lutjanidae; (B) Serranidae; (Cl Haemu- lidae; (D) other fishes; and (E) per cent composition of the catch: fish groups as in A-D. Koslow et al.: Catch and effort analysis of reef fisheries 743 Table 4 Surplus production model based upon Jamaica and Belize fishery data. Results of regressions between catch per unit of effort for all reef fish (Total CPUE) and prime commercial species (Prime CPUE) with fishing effort (/). Maximum sustainable yield (MSY) cannot be estimated for regression models for Belize and Jamaica data sepa- rately because the slopes are positive. Regression models for prime commercial species for Belize and Jamaica separately also had nonsignificant positive slopes and are not shown. R2=% variance explained; P=probability level; f=f at MSY. ' msy ' Model Area R2 (%) P Slope Intercept MSY (kg/km2) f ' msy (hook h/km2) Total CPUE//" Belize 20 0.56 0.00048 1.21 — — Total CPUE//' Jamaica 3 0.79 2.0 10-6 0.17 — — Total CPUE//" Jamaica and Belize 38 0.06 -1.51XH 1.02 1,720 3,357 Prime CPUE// Jamaica and Belize 39 0.07 -1.4KH 0.85 1,253 2,942 Total CPUE/Log(/"> Jamaica and Belize 56 0.02 -0.69 2.74 1,046 3,497 Prime CPUE/Log(fl Jamaica and Belize 54 0.02 -0.67 2.52 638 2,200 1? CPUE- 1 02-0.00015(0 r2 - 0.38. p - 0.06 3 4 5 6 Effort (1.000 hook h/km2 yr) CPUE - 2 74 - 0 69 log (() r2 - 0.56. p < 0 05 Log effort (hook h/km2 yr) Discussion Differences in catch composition between the two countries largely arose from the greater abundance of lutjanids and serranids in Belize. However, there may be several contributing factors. The Belizean fishery largely targets fishes for export, and only a narrow range of species, primarily lutjanids and ser- ranids, are marketable overseas. In Jamaica, there is a large domestic market, in which a wide range of fishes may be sold. Furthermore, the predominant fishing gear in Belize is hook-and-line, which selec- tively catches piscivorous fishes, whereas the pre- dominant gear in Jamaica is the Antillean fish trap, which catches a greater diversity of fishes. Less de- sirable species may be discarded in Belize, whereas in Jamaica, virtually all species are marketed locally. However, lutjanids and serranids are also considered prime commercial species in Jamaica, and local hook- and-line fisheries target lutjanids in particular. Thus if these groups were more abundant, they would rep- resent a greater proportion of the catch. Historical records show that they were formerly caught in quan- tity by fish traps off south Jamaica (Gosse, 1851). Figure 3 The fit of total catch of reef fish per unit of effort (CPUE) in relation to fishing effort (/") for fishing grounds in Belize (B) and Jamaica (J). (A) Fishing effort on an arithmetic scale; (B) fishing effort on a logarithmic scale. 744 Fishery Bulletin 92(4). 1994 Results of the SPM suggest that the Belize fishery is capable of further expansion in most areas but that Jamaican fishing areas are overfished. Current lev- els of effort in Belize seem to be only 10% of the lev- els that would maximize landings of prime commer- cial species. Landings are presently at about half of MSY for this group (Table 5). This is not surprising because many Belizean fishermen report that this fishery is virtually incidental to their lobster fish- ery. In Jamaica, on the other hand, present fishing effort is 41% above the level that would maximize the catch of prime commercial species, but effort is below the level predicted to maximize total fish land- ings. However, the low present catch of prime com- mercial species in Jamaica relative to their appar- ent potential (21% of MSY) is clearly due to the ef- fects of overfishing rather than to under-exploitation. Y = 2.74 f- 0.69 flog (() MSY = 1 05t/km2 Log effort (hook h/km2 yr) 15 Y = 2.52 f - 0.67 f log (f) MSY = 0.64 t/km2 Log effort (hook h/km2 yr) Figure 4 The relationship offish yield (F) to fishing effort (/) for reef fishing grounds in Belize (B) and Jamaica (J). (A) The relationship for total yield of reef fish; (B) the relationship for prime commercial species. The model's predictions must be regarded with caution, however, because of the poor fit of the SPM data. The relationships of CPUE and /"within coun- tries were nonsignificant but positive (Table 4). When the relationship between CPUE and f is non-nega- tive, MSY cannot be estimated: the relationship of yield (Y) with effort (/) continues to increase rather than attain a maximum. Although a negative slope might be obtained if particular data points were re- moved, there was no objective basis for doing this. Thus, when data from Belize and Jamaica were pooled, the negative slope of the regression between CPUE and /"was predominantly based upon the re- lationship between countries. This decreases the ef- fective number of degrees of freedom and diminishes confidence in the estimate of MSY. The estimate may, therefore, serve to establish initial levels of MSY, but if time series of catch and effort are developed in the two countries, the present relationship is likely to be modified and should be reevaluated for each fish- ery and country as data allow. The lack of significant relationships between CPUE and /'within the Jamaican and Belizean reef fisheries may arise from several factors: heterogeneity of the fishery among areas; mixing of fish stocks between fishing areas or migration offish into or out of these areas; and disequilibrium of the fisheries in the dif- ferent areas. All of these factors appear to be present, although their relative importance is unclear. Heterogeneity is apparent from the differences in composition of the catches within countries as well as between them. Heterogeneity was noted when land- ings were classified at familial or broader taxonomic groupings and likely is greater at the species level. Movements of fishes among areas were noted par- ticularly in the Belizean fishery, which is based upon a mix of targeted fishing on spawning aggregations and fishing on the nonspawning, more dispersed phase of the populations. CPUE may be expected to vary between these two phases of the fishery, thereby confounding the use of a spatially based surplus pro- duction model. Separation of these two phases of the fishery is difficult. CPUE is a function of both the degree of aggregation (or behavior) of the fish and of their abundance, which is presumably affected by total f. Therefore, data cannot be used from only one phase of the fishery. Furthermore, the catchability of a particular gear — and hence its impact per unit of /upon fishing mortality — presumably varies be- tween the different phases of a fishery. It may there- fore be necessary to standardize each gear type be- tween different phases of the fishery, as well as to standardize among gears. Jamaican, and perhaps Belizean, reef fisheries may be in a state of flux, which violates the model's as- Koslow et al.: Catch and effort analysis of reef fisheries 745 Table 5 Comparison of present levels of total and prime commercial fish yields (Y) and sustainable yield (MSY) (in tonnes) and fat MSY ( f ) (in thousands of hook h) tion model. fishing effort (/) with maximum predicted by the surplus produc- Present fishery Model Total Y (t) Primer (t) f ( '000 hook h) Total MSY (t) Prime MSY (t) Total fmsy (000 hookh) Prime f , (000 hookh) Belize 241 Jamaica 998 216 247 149 5,598 742 1,890 452 1,153 2,480 6,317 1,560 3,974 sumption of equilibrium. In Jamaica, one index of fishing effort, the number of fishing canoes, ap- pears to have declined by 55% over the past de- cade. In a 1981 survey, 2,137 fishing canoes were recorded along the south coast (Haughton, 1988) but only 963 in the present study. (Because the fishery could be easily censused, the number of canoes on fishing beaches was the primary mea- sure of fishing effort in most previous studies of Jamaican reef fisheries [Munro, 1978, 1983; Haughton, 1988]). Landings offish from the south Jamaican shelf declined 82% during this decade from 5,475 metric tons (t) in 1981 (Haughton, 1988) to 998 t in 1991. The decline in landings and effort resulted in a 60% decline in CPUE from 2.56 to 1.04 t/canoeyr. The decline in fishing ef- fort may be a consequence of falling catch rates. The datum for CPUE in relation to /"for 1991 does not fall along the line defined by the 1968-1981 data for the Jamaican fishery (Fig. 5), possibly be- cause the fishery is not at equilibrium, that is, it has not recovered in response to recently reduced effort. It may be expected that estimates of sustainable yield and effort obtained from the present survey would be significantly lower than previous estimates owing to reduced levels of CPUE, catch, and effort. Munro ( 1978) estimated that MSY for the Jamaican reef fisheries was 4.1 t/km2 and that f was 3.2 canoes/km2 shelf area. These estimates were based primarily on data from the north coast, where the shelf is narrow and much of the substrate is coral, therefore they are probably comparable to our esti- mates based upon the coral and sea-grass fraction of the south Jamaican shelf. Munro's spatially based SPM used data from a 1968 fishery survey. Haughton (1988) developed an SPM for the Jamaica reef fish- ery based upon three fishery surveys of the north and south Jamaican shelves conducted between 1968 and 1981. Differences in the productivity per unit area of the north and south Jamaican shelf were not considered. Haughton estimated MSY for the south 5000 -i 1968 S • 01968S' 4000 - 3000 - i^Tr---l^iN 01973' • ^ ~^_^ 01981S' 1981 S ----— ^___ 2000 - 1000 - «1991 S 1981N"- 0 I I I I I 1 I 1 I 2 0.3 04 05 06 07 08 09 10 11 I 1.2 I 1.3 1 1 1.4 1.5 Fishing intensity (canoes/km^) Figure 5 The relationship of catch per unit of effort for the Jamai- can reef fisheries based upon data from 1968 to 1981 (from Haughton, 1988) and from the present study. Data are shown for reef fisheries off the north (N) and south (S) coasts of Jamaica. ( ' ) indicates data from the southern shelf that have been corrected for the proportion of reef and sea grass habitat. The regression line is based upon the original, uncorrected data from 1968 to 1981. Jamaican fishery to be 2.2 t/km2 of shelf with an/1 of ~1 mechanized canoe/km2. (The units of effort of the Munro and Haughton models are not entirely comparable because most canoes became mechanized after Munro's survey; Haughton standardized his effort data to the mechanized canoe.) If the data for the southern shelf in Haugh ton's model are normal- ized per unit of area of productive habitat (~50% of the total shelf area), so that data from the north and south coast are comparable, the revised estimates of MSY and f ' are 3.1 m/km2 and 1.5 canoes/km2 of productive habitat. However, there is no longer a sig- nificant relationship between CPUE and /"(Fig. 5). Our estimate of MSY for the combined Jamaican and Belizean reef fisheries is 1.0 t/km2 (Table 4). Based upon present effort in Jamaica being 89% of f . (/=3,099 hook h/km2 [Table 2];/ =3,497 hook h/km2 ii l J" msy ' [Table 4]),/" may be estimated to be approximately 0.6 canoes/km2 of productive habitat; the present density of canoes is 963 canoes over a productive shelf 746 Fishery Bulletin 92(4), 1994 area of 1,807 km2 or 0.5 canoes/km2). Thus, present estimates of MSY and f are on the order of 20- 30% of earlier estimates. Declining levels of CPUE, catch, and effort in the south Jamaican fishery and lower estimates of sus- tainable yield and effort all indicate that the pro- ductivity of Jamaican reef fishes significantly de- clined because of overfishing. The species composi- tion of the trap fishery in the 1800's appears to have been broadly similar to that off Belize today (Gosse, 1851). By 1968-71, when the first research surveys on the south Jamaican shelf were carried out, the catch was already dominated by relatively low-value fish: the Haemulidae, Scaridae, and Acanthuridae (Munro, 1983). By 1986, when these surveys were repeated, overall CPUE had declined 33%. Several families across a wide trophic range that represented the bulk of the catch in 1968-71 had declined by more than 50% (haemulids, small serranids, and acan- thurids) or virtually disappeared (large serranids and large scarids) (Koslow et al., 1988). The Holocentridae and Pomacentridae were the only families that in- creased significantly. Thus large segments of the demersal fish community may be depleted on reefs overfished by traps. This is in contrast to reefs ex- ploited by more selective gear, such as spears that target large piscivores, where a range of unfished or lightly fished species may increase because of reduced predation (Bohnsack, 1982). Several factors in addition to overfishing may have contributed to the decline in productivity of Jamai- can reefs. Pollution can be severe in the coastal zone (Goodbody, 1989). There has been extensive reef dam- age from hurricanes in recent decades. Reduced coral production is associated with coral bleaching and coral overgrowth by algae, which may be exacerbated by the decline in herbivorous fishes, as well as by eutrophication. Previous estimates of sustainable yield from the south Jamaican shelf may have been biased upward. An important assumption of the SPM is that the fish- ery is at equilibrium, such that the reported catch and effort are sustainable. The progressive decline of the fishery indicates that previous yields were not sustainable; therefore, estimates of MSY based on those catch and effort data were likely inflated. Despite the progressive decline of the reef fish fauna on the south Jamaican shelf over recent de- cades, fishery assessments based on area-based SPM's indicated that the region was underutilized or only moderately exploited until as late as 1981 (Munro, 1978; Haughton, 1988). These analyses seem to have been confounded by combining data from the northern and southern Jamaican shelves without normalizing for the -50^ lower density of productive habitat on the southern shelf. The level of exploitation of the southern shelf relative to the northern shelf was therefore underestimated by -50% (Fig. 5). At present, the reef fishery on the south coast of Jamaica seems to be at the point of economic self- regulation (Gordon, 1954), such that effort has de- clined over the past decade owing to dramatically declining catch rates as a result of over fishing. In view of the general lack of opportunities in the Ja- maican economy, an unmanaged reef fishery will re- main heavily overfished and its productivity substan- tially reduced. Present landings from the southern shelf (0.5 t/km2 of productive habitat) are approxi- mately half the estimated potential MSY. Our estimate of MSY (0.5 t/km2 of shelf) is at the low end of estimates of maximum yield for reef fish- eries in the Caribbean, which have generally ranged from 0.5 to 1.5 t/km2 (Munro, 1983; FAO, 1985). Glo- bally, estimates of sustainable yield from reef fisher- ies have ranged as high as 20 t/km2, although these higher yields are generally from localized reefs rather than from entire shelf areas (Russ, 1991). Thus there may be a problem of standardization among studies. In conclusion, we had only limited success in de- veloping an area-based SPM for Jamaican and Belizean reef fisheries despite detailed surveys of catch and effort and estimation of the proportion of productive habitat in different areas. The difficul- ties seemed to be attributable to the nonequilibrium condition of the fisheries; the heterogeneous mix of species both within and between the two countries; the diversity of the fisheries that target a variety of spawning, sedentary, and possibly migratory ani- mals; and to possible differences in productivity among sites. Thus the model's assumptions seem too restrictive to permit meaningful analysis of catch and effort data for such complex multispecies fisheries. Violation of the model's assumptions, particularly the nonequilibrium condition of the fishery, seems to have led to serious bias in previous analyses of sustain- able yield and effort for the Jamaican fishery. More generally, these problems indicate that area-based multispecies SPM's should be used cautiously in guid- ing the future development of reef fisheries, unless the model's assumptions can be shown to be reason- ably satisfied. Nonetheless, the changes in catch com- position and the sharp declines in CPUE, yield, and estimated MSY in the Jamaican fishery over the past decade, despite declining fishing effort, are indica- tive of a severely overexploited fishery. Acknowledgments We thank the many people at the University of West Indies, the Belize Fisheries Unit, and the Jamaica Koslow et at.: Catch and effort analysis of reef fisheries 747 Fisheries Division who assisted with the project, particularly I. Goodbody, V. Gillett, A. Kong, and R. Mooyoung. We also thank R. Mahon and M. Haughton who provided useful discussions. The project was supported by a grant from the Interna- tional Centre for Ocean Development of Canada (Project No. 870138). Earlier drafts of this paper were constructively reviewed by R. Johannes, T. Smith, J. Munro, and two anonymous reviewers. V. Mawson provided editorial advice. Literature cited Aiken, K. A., and M. Haughton. 1987. Status of the Jamaica reef fishery and pro- posals for its management. Proc. 38th Annu. Gulf and Caribbean Fisheries Inst. Meeting, p. 469^184. Appeldoorn, R., and S. Meyers. In press. Fisheries resources of Puerto Rico and Hispaniola. FAO Fish. Rep. Bohnsack, J. A. 1982. Effects of piscivorous predator removal on coral reef fish community structure. In G. M. Cailliet and C. A. Simenstad (eds.), Gutshop '81: fish food habits studies. Washington Seagrant Publ., Univ. Washington, Seattle, WA. Caddy, J. F., and S. Garcia. 1982. Production modelling without long data series. FAO Fish. Rep. 278 Suppl.:309-313. FAO. 1985. Review of the state of the world fishery resources. FAO Fish. Circ. 710, Rev. 4, 61 p. Fox, W. W. 1970. An exponential yield model for optimizing exploited fish populations. Trans. Am. Fish. Soc. 99:80-88. Goodbody, I. 1989. Caribbean coastal management study: the Hellshire coast, St. Catherine, Jamaica. Marine Science Unit Research Rep. 2, 176 p. Gordon, H. S. 1954. The economic theory of a common-property resource: the fishery. J. Pol. Econ. 62:124—142. Gosse, P. H. 1851. A naturalist's sojourn in Jamaica. In D. B. Stewart (ed.), Gosse's Jamaica 1844-45, p. 3- 47. Institute of Jamaica Publ., Kingston, Jamaica. Haughton, M. 1988. An analysis of statistical data from the Ja- maican inshore fisheries. In S. Venema, J. Moller- Christensen, and D. Pauly (eds.), Contributions to tropical fisheries biology. FAO Fisheries Rep. 389, p. 443-454. Johannes, R. E. 1978. Traditional marine conservation methods in Oceania and their demise. Annu. Rev. Ecol. Syst. 9:349-364. Koslow, J. A, F. Hanley, and R. Wicklund. 1988. Effects of fishing on reef fish communities at Pedro Bank and Port Royal Cays, Jamaica. Mar. Ecol. Prog. Ser. 43:201-212. Marten, G. G. 1979. Impact of fishing on the inshore fishery of Lake Victoria (East Africa). J. Fish. Board Can. 36:891-900. Munro, J. L. 1978. Actual and potential fish production from the coralline shelves of the Caribbean Sea. FAO Fish- eries Rep. 200:301-321. 1983. Caribbean coral reef fishery resources. ICLARM, Manila, Philippines, 276 p. Nicholson, W., and L. Hartsuijker. 1982. The state of the fisheries resources of the Pedro Bank and south Jamaica shelf. FAO Fish- eries Rep. 278 Suppl.:215-254. Russ, G. R. 1991. Coral reef fisheries: effects and yields. In P. F. Sale (ed. ), The ecology of fishes on coral reefs, p. 601-636. Academic Press, San Diego, CA. Woodley, J. D., and E. Robinson. 1977. Field guidebook to the modern and ancient reefs of Jamaica. Third Int. Symposium on Coral Reefs, 33 p. Abstract. — Commercial and scientific bottom longline catches of alfonsino, Beryx splendens, from seamounts off New Caledonia were sampled to study length-frequency distributions. A total of 14,674 fish were measured. CPUE of Beryx splendens on two seamounts is mod- elled in terms of length and depth. The data show that mean length in- creases with depth; this is well de- scribed by a bivariate normal model that estimates catch for a given sea- mount. In addition, the data show that mean length also varies with the depth of the top of seamounts; this is described by a recursive model that is designed to predict ap- proximate catch for any seamount. The limitations of both models are discussed, particularly with regard to temporal variation. Modelling the distribution of alfonsino, Beryx splendens, over the seamounts of New Caledonia Patrick Lehodey Paul Marchal Rene Grandperrin Centre ORSTOM, BPA5, Noumea. New Caledonia Manuscript accepted 31 March 1994. Fishery Bulletin 92: 748-759. A bottom longline fishery operated on the seamounts of the Exclusive Economic Zone (EEZ) of New Cale- donia from February 1988 to July 1991. ' Three vessels were involved but only one vessel was operated at any given time. The fishing effort, which totalled 4,691,635 hooks, fo- cused on five seamounts (B, C, D, J, and K) whose summits are lo- cated at depths ranging from 500 to 750 m (Fig. 1 ). The target species, alfonsino, Beryx splendens, ac- counted for 92% of the catch by weight. This species has a world- wide distribution, from the equator to the temperate latitudes, and is fished by bottom trawl or longline. Alfonsino generally occupies waters between 200 and 800 m, although it has been caught at depths of only 25 m and as deep as 1,240 m (Busakhin, 1982). Some authors have noted an increase in mean length with depth2 (Yamamoto et al., 1978, Seki and Tagami, 1986), a trend which has been observed in other fishes (Heincke, 1913), par- ticularly some deep-water demersal species3' 4- 5 (Ralston and Williams, 1988). There have been few studies relating the size distribution of alfonsino to depth. The objective of this paper is to describe an ap- proach for estimating the abun- dance of alfonsino by modelling its distribution in terms of fork length and depth of capture. A bivariate normal model describes this distri- bution for a given seamount and a recursive model predicts catch on any seamount. Material and methods Data Alfonsino were captured with long- line gear (Fig. 2). The main line, averaging 4,000 m, was held on the bottom by means of terminal an- chors and regularly spaced heavy sinkers that delimited five equal line sections. During a fishing trip 1 Grandperrin, R., and P. Lehodey. 1993. Etude de la pecherie de poissons profonds dans la zone economique de Nouvelle- Caledonie. Rapport final. Contrat de re- cherche ORSTOM /Territoire de Nouvelle- Caledonie. Noumea: ORSTOM, Conv. Sci. Mer, Biol. Mar. 9, 321 p. 2 Masuzawa, T., Y. Kurata, and K. Onishi. 1975. Results of group study on population of demersal fishes in water from Sagami Bay to the southern Izu Islands — popula- tion ecology of Japanese alfonsin and other demersal fishes. Japan Aquatic Resources Conserv. Assoc. Fish. Res. Paper 28, 105 p. (English translation held at Fisheries Research Centre Library, MAF, P.O. Box 297, Wellington]. :t Brouard, F, and R. Grandperrin. 1985. Deep bottom fishes of the outer reef slope in Vanuatu. South Pacific Commission 17th Regional Technical Meeting on Fish- eries, W P 12, 127 p. 4 Clark, M. R., and K. J. King. 1989. Deep- water fish resources off the North Island, New Zealand: results of a trawl survey, May 1985 to June 1986. N.Z. Fish. Tech. Rep. 11, 55 p. 5 Dalzell, P., and G. L. Preston. 1992. Deep reef slope fishery resources of the South Pacific. South Pacific Comm. Inshore Fish. Res. Project. Tech. Doc. 2, 299 p. 748 Lehodey et al.: Modelling the distribution of Beryx splendens 749 ^¥T pa Ne,«»*0 Depth (m) below surface o'iooo ^^ 1000-2000 ■i 2000 - 3000 ■i >3000 Figure 1 Main seamounts fished for alfonsino, Beryx splendens, by bottom longline in New Caledonia. made by the longliner Humboldt from May to July 1991 over seamounts B, C, D, J, and K,6 the depth profile of the bottom was recorded on an echosounder as the line was set. The position and the depth at the exact time the terminal anchors and intermedi- ate sinkers were thrown overboard were also re- corded. The longliner Humboldt was equipped with a Doppler sonar current indicator which provided current velocity and direction at three selected depths. Data recorded suggest that the current ve- locity rapidly decreased with depth (Fig. 3A) and that horizontal drift was probably minimal.6 On 23 occa- sions over the total of 73 longline sets, the depth of the bottom was recorded at the time the buoy was grabbed at the beginning of retrieval. This depth was compared with the depth of the corresponding ter- minal anchor recorded when the line was set. Depth difference was less than 10 m for 74 % of the paired comparisons (Fig. 3B) which indicates that either the 6 Lehodey, P. 1991. Mission d'observations halieutiques sur le palangrier Humboldt. Campagne de peche du 30 mai au 12 juillet 1991, Noumea. ORSTOM Rapp. Missions, Sci. Mer Biol., Mar. 8, 44 p. drift of the line during sinking was limited or the slope of the bottom was slight. Therefore, despite the lack of a maximum depth recorder to determine the actual depth of the main line (Somerton and Kik- kawa, 1992), it was reasonable to assume that its configuration was similar to the depth profile indi- cated by the echosounder. The estimated depth of the sinkers was used to allo- cate a mean value of depth of capture di = 1/2 (dt +dl+l) to all the fish caught on the same 800-m line section (Fig. 2). Ten meters, which is roughly half the length of the branch lines, was then added to each mean depth of capture dt to correct for bias introduced by the fact that catches may occur at any hook level. Figure 3C gives the depth variation within each sec- tion. Eighty-one percent of the variation in depth is less than 15 m and 929c is within the 0-25 m range. This indicates that in most cases the longline was nearly horizontal with the bottom. Therefore, the allocation of a single depth of capture to all fish caught on the same line section seems reasonable, particularly as the depth of capture data were ag- gregated into 25-m depth classes for analysis. Dur- 750 Fishery Bulletin 92(4). 1994 float 1 70 m 780 m 780 m line Detail of branch I (20 hooks) Figure 2 Longline gear employed from the longliner Humboldt during a set ( 15) on 7 June 1991 on seamount K (lat. 24°43'S; long. 170°06'E) and detail of a branch line. Main line is 4,000 m long and divided into five sections (each section has 840 hooks [42 branch lines x 20 hooks] and is 800 m longhji, depth recorded on the echosounder at time t{ when sinker was thrown overboard; d, =1/2 (dl cated to all fish hooked on section i. + d ,) + 10 = mean depth of capture allo- Figure 3 Measurements taken to assess the deviation between the depth pro- file recorded on the echosounder and the actual configuration of the longline on the bottom: (A) current velocity from the Doppler sonar current indicator recorded at dif- ferent depths during the settings of the longline by the Humboldt (rc=694 current measurements); (B) deviations between the depth of the terminal anchor recorded when the line was set and the depth of the bottom recorded at the time the cor- responding buoy was grabbed at the begining of retrieval (n=23 sets); (C) depth variations within sections recorded during the fish- ing cruise carried out by the longliner Humboldt (n=290 line sections of 800 m each); (D) depth variations (differences between maximum and minimum depths) for the whole line recorded during the fishing cruises carried out by the longliners Hokko Maru and Fukuju Maru (n=287 longlines of 4,000 m). 100 -, A 90 80 ■ 100 m(N = 2281 70 O70m(N = l64) ,_ 60 ■ 50 m (N = 69) 7/ '.mlN 23 ., BE a. 40 30 - 20 1 P ,0 i ll i « a a oJlBJU l ■BUJU-M n •. ■, -i -i 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1 1 1.2 1.3 CURRENT VELOCITY (Vnolsl 60 B 50 - N = 23 40 - I 1- o 30 - ■ a. 20 ■ 10 - II 1 0 ■ ■ ■ M ■ M, 5 10 15 20 25 30 35 DEPTH DEVIATION (m) 0 5 10 15 20 ?& 30 35 40 45 50 SS 60 DEPTH VARIATION WITHIN SECTIONS (m) I) 25 50 75 100 125 150 175 200 225 250 DEPTH VARIATION OVER WHOl-E UNE (ml Lehodey et al.: Modelling the distribution of Beryx splendens 751 ing two other commercial cruises conducted by the longliners Hokko Maru and Fukuju Maru, observers recorded the maximum and minimum depths reached by the longline7- 8 (Fig. 3D). Fork length (FL) was measured on a total of 14,674 alfonsino. During the commercial fishing trips, fish to be measured were randomly sampled from each set. When there were only a few fish, they were all measured. As the samples varied in size between sets, the length-frequency distribution of all the alfonsino caught was estimated by multiplying the number in the sample by the ratio of the total number of alfonsino caught to the total number of alfonsino measured for each longline set. During several sci- entific cruises all of the fish were measured.1 Because 7 Anonymous. 1988. Rapport de la campagne de peche a la palangre profonde dans la zone economique de la Nouvelle- Caledonie. Hokko Maru 107: fevrier-mai 1988. Territoire de Nouvelle-Caledonie, Service Territorial de la Marine Marchande et des Peches Maritimes, 57 p. 8 Laboute, P. 1989. Mission d'observation halieutique sur le palangrier japonais Fukuju Maru du 21 nov. au 12 dec. 1988, Noumea. ORSTOM Rapp. Missions, Sci. Mer Biol. Mar. 2, 15 p. all the depth zones were not sampled in the same way, catch per unit of effort (CPUE=number offish caught per 100,000 hooks) was taken as the abundance index. The data collected during the Humboldt cruise was used to model the distribution of CPUE in terms of length and depth over seamounts. The models were validated with data generated by the scientific cruises and with data collected on board the two ad- ditional commercial boats, Hokko Maru and Fukuju Maru. These commercial data are less precise be- cause only maximum and minimum longline depths were recorded. Preliminary data analysis Table 1 shows the mean fork length by depth zone for each seamount sampled during the Humboldt cruise. This table suggests a significant increase in mean length with depth on each seamount. In order to model this increase, the data must be as repre- sentative as possible of the fish population over its depth range. For this reason, only data recorded for seamounts B and J were used for modelling the in- crease in length with depth. Table 1 Analysis of variance and multiple comparisons of mean fork length (cm) of alfonsino, Beryx splendens, sampled on five seamounts during the Humboldt fishing cruise (sample size in parentheses). Water depth of top (m) Seamount B [502 m] C [560 m] J [630 m] D K [630 ml 1710 m) a* <525 33.27(52) 525-549 34.90 (665) 550-574 37.11(635) 34.72 (405) 0.0001 575-599 38.08 (82) 36.78(1083) 0.0007 600-624 40.39(125) 38.95(74) 35.23 (262) 0.0001 625-649 36.24 (462) 650-674 37.55 (913) 35.73(138) 0.0001 675-699 38.16(115) 35.54(295) 0.0001 700-724 725-749 40.60(205) 750-774 38.21 (309) 775-799 39.90 (308) + 800 40.57 (53) a* 0.0001 0.0001 0.0001 0.4979 0.0001 * If a is less than 0.05 then the hypothesis the multiple comparison test of Tukey- different at the 0.05 level. that the means are the Kramer (in SAS, 1988) same in all classes is The shaded boxes rejected. All individual means were compared pairwise with indicate that the two included means are not significantly 752 Fishery Bulletin 92(4), 1994 CPUE distributions by size class and depth zone over seamounts B and J from the Humboldt are shown in Figure 4. A preliminary examination of the data revealed that they fitted portions of curves con- forming to a normal distribution. It was therefore assumed that for a given seamount the CPUE, in terms of length and depth, was distributed over a surface described by a bivariate normal distribution function delimited by the maximum and minimum of lengths and depths sampled. This assumption is the basis of the first modelling exercise ("bivariate normal model"). Table 1 also shows that, for a given absolute depth, mean length significantly decreases as the depth of the top of the seamount increases. This decrease sug- gests that the length distribution depends both on the absolute depth (in relation to the sea surface) and on the depth of the top of the seamount. Conse- quently, the bivariate normal model constructed for a given seamount may not be applicable to other sea- mounts whose summits lie at different depths. It is therefore necessary to construct a more general model (referred to as the "recursive model") which would predict extrapolated estimates of CPUE over any seamount by taking into account both the abso- lute depth of the water column and the depth of the top of the seamount. Temporal validation of these two models requires data that were not used during model construction but were collected in the same area at different periods. Data collected on board RV Alis and the fishing vessels Hokko Maru and Fukuju Maru were used for model validation. Modelling method Bivariate normal model In the bivariate normal model, CPUE by length and depth are calculated on the basis of a bivariate normal distribution defined by the density function ( 1 1 B(xhxd) = exp< 2(l-p*) */-/*/ '/ 2 no, od Vl-P2 {*i-Vi)(*d-Hd} 2p- tf/o-rf Hi ((xd-Md)) where xl is the length, xd is the depth, p, is the mean length, o, is the standard deviation of length, p , is the mean depth, CTf/is the standard deviation of depth, and p is the regression coefficient of length on depth. Because sampling of the seamounts is limited up- wards by their summit (Ds) and downwards by the maximum depth accessible with the bottom longline (Da ), CPUE distributions will be modelled by a por- tion of the bivariate normal distribution (2) CPUEest (xt ,xd ) = 0 for xd > Da or xd < Ds CPUEest(.r„V = XW#ltxd)for Ds|||#feW' 575 75 ~~^~~~^^ ^^§§§§§£^5* 550 DEPTH (m) 35 ^^^-525 FL (cm) K SEAMOUNT J BIVARIATE NORMAL MODEL CPUE pL 750 700 DEPTH (m) FL (cm) SEAMOUNT B RECURSIVE MODEL CPUE 5000-,''' if 4000 XOoj 2000 J 625 Do I0OO ^^^^^AVlW^^^^' 600 u — O-l^ H^rr \$$§$>>^' 575 25 X \0§£^ 550 DEPTH (m) W <°7, ^ FL (cm) 45 SEAMOUNT J RECURSIVE MOOEL CPUE DEPTH (m) Figure 4 Actual CPUE for alfonsino, Beryx splendens, by fork length (cm) and depth (m), recorded on seamounts B and J during the fishing cruise carried out by the longliner Humboldt, and predicted CPUE from the bivariate normal model and the recursive model. 754 Fishery Bulletin 92(4). 1994 Deplh zones Seamount 0* . j j+l j+2 j+3 Do D, Y , >- P Y , >- p2Y ^ *~ p3 Y D,+l \ \ "~-*- P^l-PJY ^p(l-p)Y^— — >~ p2(l-p)Y X (l-p)Y^ >- p(l-p)Y X > p2(I-p)Y Z , >~ p Z A *- P*Z ^ . X > P3 z D,+2 \ \\ NN^ "(i-")2y \ \\ \S Pd-P)2Y \ \\ X p2(l-p)Z \ V ('-p)2y ^t *- p('-p)2y \ %(I-p)z\ > p2(lp)Z A (l-p)Z, ^ P(1-P)Z X.X> p2(l-p)Z D,+3 \ \S^ Cp>3y \ \Ap(i.p)2z \ Xpd-p)2^ * (l-prz ^ >- pd-prZ Di+4 """■""^ (1-P)3 z Figure 5 Changes in distribution of alfonsino, Beryx splendens, according to the recursive model of two subpopulations Y and Z that migrate from a sea- mount j to deeper seamounts_/+l,7+2,7+3. y=subpopulation distributed on a seamount./ at absolute depth Dt; Z=subpopulation distributed on a seamounty at absolute depth D|+1;p=probability that fish will redistrib- ute according to absolute depth; l-p=probability that fish will redistrib- ute according to the depth of the top of the seamount; *=original sea- mount. group will move down to zone D-+1 with a probability 1-p (Fig. 5). Thus, it is possible to determine the CPUE (X- k) for the population of a depth zone Dr on a seamount j, for a size class k in terms of the subpopulations of zones Di and Dl_l on the higher- level seamount j—1. This is expressed as follows: PXlllM+(l-p)X, 1J L,* (3) Specifically, if p=0, CPUE is distributed solely ac- cording to the depth of the top of the seamount and if p=l, it is distributed solely according to the abso- lute depth. If the parameters X^ p,, op prf, od, p, and p. estimated from a known seamount length-depth distribution are known, it is possible to calculate all the CPUE (Xi]k) values for any seamount j (deeper or shallower), depth zone Di and size class k. The foregoing seven parameters can be estimated by mini- mizing the SSE between the CPUE recorded on one of the best sampled seamounts (B or J) and the CPUE estimated by Equation 3. This estimation is per- formed by a nonlinear regression (SAS, 1988). Results Bivariate normal model Application of a bivariate normal model implies that mean length can be deduced from depth by a linear regression weighted by the CPUE x, = axd + b where a and b are constants). The results of this regression for seamounts B and J show mean length and depth to be significantly correlated ( Table 2). Consequently, the bivariate normal model can be tested for each of these seamounts. The parameters of the bivariate normal model were calculated separately for seamounts B and J (Table 3). The determination coefficient,9 R2, for seamounts B and J respectively equals 0.87 and 0.93. The re- sidual analysis was carried out to test the fit of the model to the data from the Humboldt cruise on sea- mounts B and J. The results show the residuals are fl2 = h- (Y. - Y) 7i (V, with Y = CPUE. Lehodey et al.: Modelling the distribution of Beryx splendens 755 Bivariate normal model: CPUE - Table 2 weighted linear regression of length of alfonsino, Beryx splendens, on depth. No. offish Seamount measured Min. depth Max. depth (m) (ml p a H0: r = 0 a b B 1,557 J 1,957 516 615 0.549 606 761 0.486 < 0.0001 < 0.0001 rejected 0.063 0.037 rejected 1.251 13.122 p = regression coefficient. a = significance probability of the regression under the null hypothesis that the statistic is Hg = null hypothesis i.e. length and depth are independent; if a < 0.05, Ho is rejected, a and b = parameters of the linear regression. zero. satisfactory; in particular, the residuals are centered on zero, are not correlated with the length and depth variables, and have a constant variance (Table 4; Fig. 6). These characteristics indicate a good fit of the bivariate normal model to the data as demonstrated by comparison of actual and predicted CPUE (Fig. 4). Extrapolation of the model to the data not used in the modelling exercise is unsatisfactory because the mean value of the residuals is not centered on zero for the Fukuju Maru data and because the residuals are correlated with the length variable for the RV Alls data and with the depth variable for the Hokko Maru data (Table 4; Fig. 6). This suggests the exist- ence of factors affecting the population's distribution not accounted for by the model. Recursive model The parameters of the recursive model were esti- mated separately for seamounts B and J. The deter- Table 3 Bivariate normal model: predicted parameters for seamounts B and J. SD=Standard deviation. Seamount and number offish measured B 1,557 J 1,957 Parameters Estimation SD Estimation SD X V-i V-d °; P2 3.66xl06 0.85xl06 0.5xl09 4.5xl09 35.22 0.96 12.0 52.0 542.87 14.49 -123.1 1,666.7 5.31 1.01 8.8 9.6 71.07 17.04 272.5 317.1 0.75 0.10 0.9 0.2 A=theoretical cumulative CPUE. ^=mean length (cm). p^=mean depth (m). <7,=standard deviation oflength. <7j=standard deviation of depth. p2=regression coefficient of length on depth. Table 4 Bivariate normal model: results of analysis of residuals (e) for fit control and temporal validation of the model for seamounts B and J. Cruise Seamount No. of fish measured ai ff„ : e = 0 a., H0 : P, = 0 a3 ff„:p2=0 Fit control Temporal validation Humboldt Humboldt Hokko Maru RV Alls Fukuju Maru B J 1! B J 1,557 1,957 2,840 1,688 4,320 0.289 O.OISH 0.262 0.908 0.0002 not rejected not rejected not rejected not rejected rejected 0.153 0.061 0.601 0.016 0.265 not rejected not rejected not rejected rejected not rejected 0.149 0.431 <0.0001 0.391 0.284 not rejected not rejected rejected not rejected not rejected H0 : e = 0 . The mean value of the deviations between estimated and observed CPUE is 0. If a, is <0.05, Hu is rejected. p,= regression coefficient oft on length. p2=regression coefficient of e on depth. //0:p, = 0. If ct,is< 5%, H0 is rejected. H0:p2 = 0. If a3 is < 5%, H0 is rejected. 756 Fishery Bulletin 92(4). 1994 BIVARIATE NORMAL MODEL SEAMOUNTB S 4000 T DC .- . S5 2000 O o 21 iooo 0. 6 t yfi ' 0 1000 2000 3000 4000 ACTUAL CPUE (Wo. fish/ 100.000 hooks) SI AM' UNI II 9 S SI 2 32 40 48 Ft (cm) St AMOUNT B 570 620 DEPTH (m) PREDICTED CPUE No. fish/100.000 hooks) SE AMOUNT J v / •• * v» •* 0 1000 2OO0 3000 4000 ACTUAL CPUE (NoflslvlOO.000 hooks) < g 9 2 • • I I 9 5 tr I -1000 j-1 o" ?• 2000 i— H-h- 630 670 710 750 DEPTH (m) RECURSIVE MODEL 1000 2000 3000 4000 ACTUAL CPUE (No.SSnV100.000 hooks) SEAMOUNTB « 3000 •• RESIDUALS j.Hshv 100.000 hoc i 3 d i - •••••2. •^•p^i •• z inoo 24 32 40 48 FL (cm) SEAMOUNTB y8 3000 t 2000 *• 1000 +J 1000 ■ ■2000- 3000- t-t-t 520 540 560 560 600 620 DEPTH (m) 1000 2000 3000 4000 ACTUAL CPUE {No. lsh/100.000 hooks) 2000 1000- 0- 1000- 2000 -3000 24 SE AMOUNT J ^ 2000t - -• „ 5 1000 1 • • • §8 o in S 1000 1 £ 2000 • - • ' 590 630 670 710 DEPTH (m) 750 Figure 6 Bivariate normal model and recursive model for seamounts B and J: distribution of predicted CPUE of alfonsino, Beryx splendens, in relation to actual CPUE and distributions of residuals in relation to length (cm) and depth (m). Dotted lines delimit the confidence interval at cx=0.05. Table 5 Recursive model: estimated parameters of the dis- tribution of CPUE on the hypothetical original sea- mount calculated from seamount J data. SD = stan- dard deviation. Parameters Estimation SD M 22.86 91.3 V-A -66.17 499.9 a. 5.98 16.5 °A 30.27 99.2 P2 0.87 0.70 P 0.09 0.04 /,, 1.15xl08 2.33x10s Li;=mean length (cm). H,/=mean depth (m), 0/=standard deviation oi Length (^standard deviation of depth. p2=regression coefficient of length on depth. p=probability that fish will redistribute according to absolute depth A„=theoretical cumulative CPUE. mination coefficient, R2, calculated for seamount J equals 0.82, while for seamount B it equals 0.69. Therefore, the parameters estimated for seamount J were used in the model (Table 5). The residuals resulting from fitting the model to data from the Humboldt cruise on seamount J are satisfactory; in particular, they are centered on zero and are not cor- related with the studied variables (Table 6; Fig. 6). These features indicate a good fit of the recursive model to the data as demonstrated by comparison of actual and predicted CPUE for seamount J (Fig. 4). It is interesting to note the low value of p (close to 0.1 as shown in Table 5), which indicates that the seamount top depth parameter has greater impact on the length distribution than does the absolute depth parameter. Spatial validation was carried out for seamount B from the data collected during the Humboldt cruise (Table 6). The residuals are centered on zero and not correlated with the length and depth variables. How- ever, since their variance is not constant (Fig. 6) and Lehodey et al.: Modelling the distribution of Beryx splendens 757 Table 6 Recursive model: results of analysis of residuals (£ ) for fit control and validation of the model for seamounts B and J. No. offish Cruise Seamount measured a, H„ : e = 0 a2 H0 : P, = 0 a3 H0 : p, = 0 Fit control Humboldt J 1,957 0.87 not rejected 0.85 not rejected 0.06 not rejected Spatial validation Humboldt B 1,557 0.13 not rejected 0.06 not rejected 0.99 not rejected Temporal validation Hokko Maru B 2,840 0.15 not rejected 0.76 not rejected <0.0001 rejected RV Alis B 1,688 0.01 rejected 0.005 rejected 0.35 not rejected Fukuju Maru J 4,301 0.007 rejected 0.67 not rejected 0.08 not rejected H0 : £ = 0 . The mean value of the residuals is 0 If a,is<5%. in is rejected. Pj = regression p2 = regression Ho:p, = 0. Ifcc, Ho.p2 = 0. Ifa3 coefficient of e on length, coefficient oft on depth, s <5%, Ha is rejected, is <5%, H0 is rejected. since the standard deviations of the parameters are high (Table 5), spatial extrapolation of the model to seamount B is rather crude as demonstrated by com- parison of actual and predicted CPUE (Fig. 4). Temporal validation was carried out on data from the Fukuju Maru and RV Alis fishing on seamounts B and J (Table 6). It is unsatisfactory because the mean values of the residuals are not centered on zero and the residuals are correlated with the length vari- able for the RV Alis data and with the depth vari- able for the Hokko Maru data (Fig. 6). As with the bivariate normal model, this suggests the existence of factors not accounted for by the model. Discussion Alfonsino length structure variation observed over the seamounts of New Caledonia is similar to that noted in Japan2 and in New Zealand10 (Massey and Horn, 1990) where it was assigned to age-specific migrations. In Japan, it was noted that alfonsino move south as they grow,2 young fish predominate over some seamounts and old fish predominate over other seamounts. In New Caledonia, age segregation over the seamounts is so marked that it has been possible to describe it mathematically. The bivariate normal and recursive models appear to be complementary. The bivariate normal model provides an instantaneous picture of alfonsino popu- lation distribution on a given seamount; it provides good CPUE estimates provided a sufficient amount of length and depth data are available. The recur- sive model takes into account the dynamic nature of the population's distribution as it allows the extrapo- lation of CPUE obtained for one seamount to sea- mounts that were not sampled. It allows preliminary population estimation of unexploited stocks. Depend- ing on current economic parameters, the model might be used to indicate the depths at which fishing is most economic. Once a fishery is operational, more refined data will be available, which will enable the bivariate normal model to be applied and stock man- agement parameters defined for each of the sea- mounts fished. The poor results obtained for the temporal valida- tion could be due to poor precision of the depth data collected from the longliners Hokko Maru and Fukuju Maru. Also, neither of the models incorporate a time factor. The data were collected from cruises carried out in different years and in different seasons. Hence, it is unlikely that conditions remained stable, par- ticularly with regard to exploitation history, repro- ductive behavior, or long-term climatic variations. Fishing methods and strategies were not modified during the fishing period considered. Therefore, the catches are probably representative of the standing stock of alfonsino within the size limits determined by the selectivity of the fishing gear. Since the daily observation window did not change, vertical trophic migrations would seem unlikely to contribute to the observed variability. With regard to sex as a source of variability, although the mean length of females exceeds that of males11 (Kotlyar, 1987; Massey and 10 Horn, P. L., and B. R. Massey. 1989. Biology and abundance of alfonsino and bluenose off the lower east coast. North Island, New Zealand. N. Z. Fish. Tech. Rep. 15, 31 p. Lehodey, P. 1994. Les monts sous-marins de Nouvelle-Caledonie et leurs ressources halieutiques. These de doctorat de l'Universite Francaise du Pacifique, 398 p. 758 Fishery Bulletin 92(4), 1994 Horn, 1990), Humphreys et al. (1984) have shown that sexual dimorphism is not responsible for the existence of different size groups of alfonsino. Marked declines in CPUE are observed in the Southern Hemi- sphere during summer. This season corresponds to the alfonsino breeding period in New Caledonian waters.11 The summer decline in catch rate could be due to breeding migrations drawing the fish to spawning grounds that are different from the fish- ing grounds2 (Chikuni, 1971) or to changes in vul- nerability to the gear owing to seasonal physiologi- cal or behavioral changes (Ricker, 1980). Data used to build the models were collected on board the Humboldt during the winter season. Data used to validate the models were collected on board Fukuju Maru and Hokko Maru at the beginning and end of the warm season and during six scientific cruises, five of which were carried out in summer. This sug- gests that reproductive seasonality might be a fac- tor in the poor temporal validation of the models. Other sources of temporal variation might be re- lated to the environment. The ocean habitat of alfonsino is not affected by continental influences but is subject to hydrological fluctuations affecting the deep-water masses. Some of these influences are of short period such as internal waves and tidal cur- rents (Eriksen, 1985; Roden, 1987), whereas others recur at longer intervals such as seasonal variations in ocean currents and multi-annual hydroclimatic anomalies of the El Nino Southern Oscillation (ENSO) (Delcroix and Henin, 1989). Such fluctua- tions might have an impact on alfonsino stock struc- ture, either at the recruitment stage (survival and dispersal of eggs and larvae) or by modification of the behavior of adults (migrations from one seamount to another). However, it is difficult to demonstrate the effect of these fluctuations on the presence and catchability of fish. It is even more difficult to ex- plain the very large differences in fishery productiv- ity observed between seamounts of identical depth, located only a few dozen miles apart and appearing to have the same hydrological environment. Seafloor topography and bottom type might account for these differences, but other hypotheses can be postulated, some based on the existence of a low-energy hydro- thermalism (Rougerie and Wauthy, 1990) and oth- ers on a hydrological anomaly called "Taylor's col- umn," which could enhance species sedentarity (Royer, 1978; Genin and Boehlert, 1985; Roden, 1987; Dower et al., 1992; Sime-Ngando et al., 1992). Fluc- tuations in intensity of this anomaly, or its disap- pearance, could also be responsible for the variations in productivity observed over time over a given sea- mount (Boehlert and Genin, 1987). These unknown environmental fluctuations cause problems in the interpretation of results from exploratory and com- mercial fishing cruises carried out over seamounts. The data collected at a given location constitute an instant picture of a stock whose abundance is likely to vary, irrespective of fishing effort, as a result of unknown environmental variations. In other words, the fertility of the seamounts could vary quite un- predictably over the history of a fishery. Conse- quently, modelling the distribution of a stock should be confined to a relatively small temporal sampling scale. Conclusion The bivariate normal model and the recursive model provide complementary interpretations of length dis- tribution in terms of depth of alfonsino fished on the seamounts of New Caledonia by the bottom longline fishery. They could be useful for the proper manage- ment of fisheries over seamounts, where stocks are known to be vulnerable ( Sasaki, 1986 ) because of the limited habitat afforded by seamounts and the slow growth rate of deep-water species. However, it would appear that annual or seasonal factors, in particu- lar those which account for recruitment fluctuations and behavioral changes linked to reproduction, will need to be incorporated into the models before they can be generalized. A better understanding of the functioning of the ecosystems concerned would also assist in establishing the limits of generalization, particularly with regard to depth and area inhab- ited by alfonsino. These models could possibly be adapted to other deep-water species such as certain snappers and groupers. Acknowledgments We wish to thank G W. Boehlert and some of the staff members of the Southwest Fisheries Science Center, Honolulu Laboratory, for their helpful com- ments on the draft manuscript and Tim Adams of the South Pacific Commission Fisheries Programme for his editorial comments. Literature cited Boehlert, G. W., and A. Genin. 1987. A review of the effects of seamounts on bio- logical processes. In B. H. Keating, P. Fryer, R. Batiza, and G. W. Boehlert (eds.), Seamounts, is- lands and atolls. Geophysical Monograph. 43:319-334. Lehodey et al.: Modelling the distribution of Beryx splendens 759 Busakhin, S. V. 1982. Systematics and distribution of the family Berycidae (Osteichthyes) in the world ocean. J. Ichthyol. 22(6):1-21. Chikuni, S. 1971. Groundfish on the seamounts in the North Pacific. Bull. Jpn. Soc. Fish. Oceanogr. 19:1-14. [English translation by K. Tatara, 1972, Fish. Res. Board Can., Translation no. 2130, 12 p.] Delcroix, T., and C. Henin. 1989. Mechanisms of subsurface thermal structure and sea surface thermohaline variabilities in the southwestern tropical Pacific during 1975-85. J. Mar. Res. 47: 777-812. Dower, J., H. Freeland, and K. Juniper. 1992. A strong biological response to oceanic flow past Cobb Seamount. Deep-Sea Res. 39 (78): 1139-1145. Eriksen, C. C. 1985. Implications of ocean bottom reflexion for in- ternal wave spectra and mixing. J. Phys. Ocean- ogr. 15:1145-1156. Genin, A., and G. W. Boehlert. 1985. Dynamics of temperature and chlorophyll structures above a seamount: an oceanic experi- ment. J. Mar. Res. 43:907-924. Heincke, F. 1913. Untersuchungen uber die Scholle, General- bericht I. Schollenfischerei und Schonmassregeln. Vorlaufige Kurze Ubersicht tiber die wichtigsten Ergebnisse des Berichts. Rapp. P-Verb. Cons. Int. Explor. Mer 16:1-70. Humphreys, R. L., Jr., D. T. Tagami, and M. P. Seki. 1984. Seamount fishery resource within the south- ern Emperor-northern Hawaiian Ridge area. In • R. W. Grigg and K. Y. Tanoue (eds. ), Proceedings of the symposium on resource investigations in the northwestern Hawaiian Islands, 25-27 May, 1983, Vol. 1, p. 283-327. Kotlyar, A. N. 1987. Age and growth of alfonsino, Beryx splendens. J. Ichthyol. 27 <2):104-111. Massey, B. R., and P. L. Horn. 1990. Growth and age structure of alfonsino {Beryx splendens) from the lower east coast, North Island, New Zealand. N. Z. J. Mar. Freshwater Res. 24 (1):121-136. Ralston, S. V., and H. A. Williams. 1988. Depth distributions, growth, and mortality of deep slope fishes from the Mariana Archipelago. NOAA Tech. Mem. NMFS-SWFC-113, 47 p. Ricker, W. E. 1980. Calcul et interpretation des statistiques biologiques des populations de poissons. Bull. Fish. Res. Board Can. 191 F, 409 p. Roden, G. I. 1987. Effect of seamounts and seamounts chains on ocean circulation and thermohaline structure. In B. H. Keating, P. Fryer, R. Batiza, and G. W. Boeh- lert feds.), Seamounts, islands and atolls. Geo- physical Monograph. 43:335-354. Rougerie, F., and B. Wauthy. 1990. Les atolls oasis. La Recherche 223:832-842. Royer, T. C. 1978. Ocean eddies generated by seamounts in the North Pacific. Science 199:1063-1064. SAS (Statistical Analysis System). 1988. SAS/STAT user's guide, release 6.03 ed. SAS Institute Inc., Cary, NC, 1028 p. Sasaki, T. 1986. Development and present status of Japanese trawl fisheries in the vicinity of seamounts. In R. N. Uchida, S. Hayasi, and G. W. Boehlert (eds.), The Environment and resources of seamounts in the North Pacific. NOAA Tech. Rep. NMFS 43: 21-30. Seki, M. P., and D. T. Tagami. 1986. Review and present status of handline and bottom longline fisheries for alfonsin. In R. N. Uchida, S. Hayasi, and G. W. Boehlert (eds.), Envi- ronment and resources of seamounts in the North Pacific. NOAA Tech. Rep. NMFS 43:31-35. Sime-Ngando, T., K. Juniper, and A. Vesina. 1992. Ciliated protozoan communities over Cobb Seamount: increase in biomass and spatial patchiness. Mar. Ecol. Prog. Ser. 89: 37-51. Somerton, D. A., and B. S. Kikkawa. 1992. Population dynamics of pelagic armorhead Pseudopentaceros wheeleri on Southeast Hancock Seamount. Fish. Bull. 90:756-769. Yamamoto, S., K. Ishii, S. Sasaki, and T. Meguro. 1978. Outlines of fisheries investigation on the Emperor seamounts by the R.V. Hokusei Maru in 1977 and some technical problems. Bull. Jpn. Soc. Fish. Oceanogr. 33:56-64. Abstract. — Spawning and early life history of white sturgeon, Aci- penser transmontanus, were stud- ied in the lower Columbia River downstream from Bonneville Dam from 1988 through 1991. From white sturgeon egg collections, we determined that successful spawn- ing occurred in all four years and that the estimated spawning period each year ranged from 38 to 48 days. The spawning period ex- tended from late April or early May through late June or early July of each year. Spawning occurred pri- marily in the fast-flowing section of the river downstream from Bonneville Dam, at water tempera- tures ranging from 10 to 19C Freshly fertilized white sturgeon eggs were collected at turbidities ranging from 2.2 to 11.5 nephelo- metric turbidity units (ntu), near- bottom velocities ranging from 0.6 to 2.4 m/s, mean water column veloci- ties ranging from 1.0 to 2.8 m/s, and depths ranging from 3 to 23 m. Bottom substrate in the river sec- tion where freshly fertilized eggs were most abundant was primarily cobble and boulder. White sturgeon larvae were collected from river kilometer (rkm) 45 to rkm 232, sug- gesting wide dispersal after hatch- ing. Larvae were collected as far downstream as the upper end of the Columbia River estuary, which is a freshwater environment. Young- of-the-year (YOY) white sturgeon were first captured in late June, less than two months after spawn- ing was estimated to have begun. Growth was rapid during the first summer; YOY white sturgeon reached a minimum mean total length of 176 mm and a minimum mean weight of 30 g by the end of September. Young-of-the-year white sturgeon were more abun- dant in deeper water (mean mini- mum depth >12.5 m) of the lower Columbia River. The results indi- cate that a large area of the lower Columbia River is used by white sturgeon at different life history stages. Spawning and early life history of white sturgeon, Acipenser transmontanus, in the lower Columbia River George T. McCabe Jr. Coastal Zone and Estuanne Studies Division, Northwest Fisheries Science Center National Marine Fisheries Service. NOAA 2725 Montlake Blvd. East. Seattle. Washington 981 12-2097 Charles A. Tracy Columbia River Fisheries Laboratory, Washington Department of Fish and Wildlife RO. Box 999. Battle Ground. Washington 98604 Manuscript accepted 28 February 1994. Fishery Bulletin 92:760-772. White sturgeon, Acipenser trans- montanus, is the largest of all North American sturgeon species and is found along the west coast of North America from the Aleutian Islands, Alaska, to Monterey, California (Scott and Crossman, 1973). Al- though this species is generally anadromous (Scott and Crossman, 1973), some populations in the Co- lumbia River Basin are landlocked because of dam construction or natural barriers (Cochnauer et al., 1985; Beamesderfer et al.1). Historically, white sturgeon were abundant in the Columbia River (Oregon and Washington ) and in the late 1800's supported an intense commercial fishery. Commercial catches peaked in 1892, when more than 2.4 million kg were landed (Craig and Hacker, 1940). After 1892, catches declined, and by 1899 the annual catch was less than 33,250 kg. Annual catches during the early 1900's were less than 104,930 kg (Craig and Hacker, 1940). White sturgeon populations in the Columbia River, particularly the one downstream from Bonneville Dam (the lowest dam at river kilo- meter [rkm] 234), have recovered sufficiently from the overfishing to support important recreational and commercial fisheries. The popula- tion of white sturgeon in the lower Columbia River, which extends from the mouth of the river to Bonneville Dam, is one of the larg- est in the world. From 1984 through 1988, the combined recreational and commercial catch in this area was at least 50,000 fish annually (Wash. Dep. Fisheries and Oregon Dep. Fish and Wildlife2). During 1992, estimated catches of white sturgeon for recreational and com- mercial fisheries were 40,100 and 6,200 fish, respectively (Melcher and King3). Presently, white stur- geon is the principal recreational fish in the Columbia River down- stream from Bonneville Dam (Mel- cher and King3). 1 Beamesderfer, R. C, T. A. Rien. C. A. Fos- ter, and A. L. Ashenfelter. 1990. Report A. In A. A. Nigro (ed. ), Status and habitat re- quirements of white sturgeon populations in the Columbia River downstream from McNary Dam, p. 6-37. Ann. Rep. to Bonneville Power Admin. (Project 86-50) by Oreg. Dep. Fish Wildl., Wash. Dep. Fish.. Natl. Mar. Fish. Serv., and U.S. Fish Wildl. Serv. Avail Bonneville Power Admin., P.O. Box 3621, Portland, OR 97208. 2 Washington Department of Fisheries and Oregon Department of Fish and Wildlife. 1992. Status report — Columbia River fish runs and fisheries, 1938-91, 224 p. Avail. Wash. Dep. Fish., P.O. Box 999, Battle Ground, WA 98604. 1 Melcher, C E., and S. D. King. 1993. The 1992 lower Columbia River and Buoy 10 recreational fisheries, 77 p. Oregon Dep. Fish Wildl., 17330 S.E. Evelyn St., Clackamas, OR 97015. 760 McCabe and Tracy: Spawning and early life history of Acipenser transmontanus 761 Although white sturgeon supports important fish- eries in the Columbia River and other rivers within its range, little is known about the spawning char- acteristics and early life history of this long-lived species. Using larval collections, Stevens and Miller (1970) described the distribution of white or green, A. medirostris, sturgeon larvae, or both, in Cali- fornia's Sacramento-San Joaquin River system, and Kohlhorst ( 1976 ) described sturgeon spawning in the Sacramento River. Parsley et al. (1993) described spawning and rearing habitats of white sturgeon in the Columbia River downstream from McNary Dam; however, important specific information about spawning and early life history of white sturgeon in the Columbia River downstream from Bonneville Dam was not presented. From 1988 through 1991, we studied spawning characteristics and early life history of white stur- geon in the lower Columbia River. Primary goals of the study were 1) to define where and when spawn- ing occurred and 2) to assess the environmental con- ditions at the time of spawning. Additional goals were to determine larval distribution and habitat use by young-of-the-year (YOY) white sturgeon. Methods Egg and larval sampling From 1988 through 1991, white sturgeon eggs and lar- vae were collected in the Columbia River downstream from Bonneville Dam. The collection period varied among years; however, in all years, it extended from at least April through early July. Generally, samples were taken weekly during this period. A D-shaped plankton net was used to collect white sturgeon eggs and larvae. This net was 0.8 m wide at the bottom of the mouth opening, 0.5 m high, and constructed of 7.9-mesh/cm nylon marquisette netting. Depending on water velocity, two to six lead weights (4.5 or 9.1 kg each) were attached to the net frame to hold the net on the river bottom. A digital flow meter (Gen- eral Oceanics Model 2030) was suspended in the mouth of the net to estimate the volume of water sampled. Typically, two plankton nets were fished simultaneously for about 30 minutes from an an- chored 12.2-m research vessel. When water veloci- ties at 0.2 of the total depth were greater than 2 m/s and other adverse sampling conditions were present, only one plankton net was fished, often for one hour. Artificial substrates constructed of latex-coated animal hair also were used to collect white sturgeon eggs (McCabe and Beckman, 1990). Each artificial substrate, which was 76 x 91 cm, was enclosed in an angle-iron frame. The substrate and frame were held in place on the bottom with a three-fluke anchor simi- lar to a grapnel. A buoy line was attached to the an- chor to allow retrieval of the substrate, frame, and anchor. Artificial substrates were generally retrieved and examined weekly for eggs. In 1990 and 1991, a 3.0-m beam trawl was used weekly or biweekly in late June, July, and August to collect white sturgeon larvae and YOY. The estimated fishing width of the trawl was 2.7 m and the height was 0.5 m. A 1.59-mm knotless nylon liner was in- serted into the body of the net. The beam trawl was towed slowly upstream along the bottom for periods ranging from 2 to 20 minutes. White sturgeon eggs and larvae were initially pre- served in an approximately 4% buffered formalde- hyde solution. After the eggs and larvae were pro- cessed in the laboratory, they were transferred to a 20% methanol solution. Processing of the eggs and larvae was done within 60 days after collection. White sturgeon egg or larval sampling was con- ducted at various sites in the lower Columbia River from rkm 29 to 234 (Table 1, Fig. 1). Following ex- ploratory research conducted in 1987, we decided to concentrate egg and larval sampling with station- ary gear (plankton nets and artificial substrates) between rkm 172 and 234. We selected a site at rkm 230, used in previous research for monitoring white sturgeon spawning, for the most frequent egg sam- pling. We call this the index site. In 1988, a 12-hour collection with a plankton net was made at the index site to determine whether catches of white sturgeon eggs and larvae changed during different light conditions. The study began at 1843 hours on 25 May and ended at 0623 hours on 26 May. Normally, one plankton net was fished for one hour during each sampling effort; 11 sampling efforts were made. Young-of-the-year sampling A 7.9-m (headrope length) semiballoon shrimp trawl was used to collect juvenile white sturgeon, includ- ing YOY, from 1988 through 1991. Mesh size in the trawl was 38 mm (stretched) in the body; a 10-mm mesh liner was inserted in the cod end. Trawling ef- forts with the shrimp trawl were normally five min- utes in duration in an upstream direction, beginning when the trawl and the proper amount of cable were deployed, and ending five minutes later. Trawl speed over the bottom was usually 3 to 5 km/hour. In 1990 and 1991, a 3.0-m beam trawl was also used to collect YOY white sturgeon (see Egg and Larval Sampling section). Using a radar range-finder, we estimated the distance fished during each sampling effort. Beam trawl speed over the bottom was usually 1 to 3 km/hour. 762 Fishery Bulletin 92(4), 1994 i ' i ' ' i f\JO 25 km Washington rkm 45 Index site rkm 223 lrkm 230) Figure 1 Location of the white sturgeon, Acipenser transmontanus, study area in the lower Columbia River, 1988-91. Trawling was conducted from late March or early April through September or October of each year. In 1989, a limited amount of sampling was conducted in early November. Sampling stations were selected to determine the range of habitat used by juvenile white sturgeon and extended from rkm 29 to 218 (Table 1). Trawling effort and geographic range of sampling var- ied among years owing to limited personnel and gear (Table 1). In 1988 and 1989, more trawling effort was concentrated in the river upstream from rkm 120. How- ever, in 1990 and 1991, much more trawling was done in the river between rkm 45 and 120 than in previous years. White sturgeon captured in bottom trawls were measured (total length) and weighed (g). On 31 July and 1 August 1990, 14 trawling efforts (7.9-m shrimp trawl) were undertaken from 1155 through 0800 hours at rkm 75 to determine whether catches of juvenile white sturgeon, particularly YOY, increased during hours of darkness. Physical conditions Selected physical parameters were measured in con- junction with biological sampling: minimum and maximum bottom depth ( m ); bottom water tempera- ture (°C); bottom water turbidity (ntu); and water velocities at 0.2 of the total depth, 0.8 of the total depth, and about 0.6 m above the bottom. By aver- aging water velocities measured at 0.2 and 0.8 of the total depth, we calculated a mean water column ve- locity (Buchanan and Somers, 1969). Water veloci- ties were measured only during egg and larval sam- pling. Depth was measured with electronic depth sounders, and velocity with a Gurley current meter attached to a 45.4-kg lead fish. A Van-Dorn water bottle was used to collect water samples just above the bot- tom. The water temperature of each sample was mea- sured immediately after collection, and a subsample of water was removed and placed in a glass bottle. The turbidity of the subsample was determined in the labo- ratory with a Hach Model 2 100A Turbidimeter. Substrate type was determined from bottom samples collected with a 0.1-m2 Van Veen grab sam- pler. In addition, a substrate sample was collected at rkm 230 (index site) by scuba divers. Particle size was defined following the classifications presented in Parsley et al. (1993). McCabe and Tracy: Spawning and early life history of Acipenser transmontanus 763 Table 1 Numbers of sampling efforts for white sturgeon, Acipenser transmontanus, eggs, larvae, and young of the year in the lower Columbia River, 1988-91 . When two plankton nets were fished s multaneously the data were combined and considered as one sampling effort. Location is shown in river kilometers (rkm). Year and location Mar Apr May Jun Jul Aug Sep Oct Total Plankton net 1988 rkm 172-228 0 n 8 17 1 (i ii 0 26 rkm 229-230 1 2 16 5 2 1 0 0 27 rkm 231-233 1 3 i> 1 ii 0 0 0 11 1989 rkm 153-171 0 n 2 0 n n 0 ii 2 rkm 172-228 0 0 IS 16 5 0 0 n 39 rkm 229-230 1 2 4 5 3 1 (1 i) 16 rkm 231-233 0 0 1 (i 0 (1 (1 0 1 1990 rkm 112-171 0 ii i) 1 n II II 0 1 rkm 172-228 0 n 29 10 3 0 0 0 42 rkm 229-230 0 5 5 4 3 1) I) 0 17 rkm 231-233 0 1 2 2 0 0 II 0 5 1991 rkm 193-228 •1 1 13 8 5 0 0 0 27 rkm 229-230 0 4 4 7 6 II 0 0 21 rkm 231-233 0 II 0 II 0 0 II 0 0 Artificial substrate 1988 rkm 197-228 0 II 3 1 0 (1 0 (1 4 rkm 229-230 1) il 5 1 2 1) (1 I) 8 rkm 231-234 (1 n 5 6 0 II 0 II 11 1989 rkm 220-228 0 0 2 3 0 II 0 0 5 rkm 229-230 0 1 0 0 0 (1 0 0 1 rkm 231-234 0 0 12 6 0 II 0 (1 18 1990 rkm 229-230 0 3 5 ii 0 II 0 (1 8 rkm 231-234 II n 9 4 1 II 0 0 15 1991 rkm 229-230 0 2 :■; 3 3 II 0 II 11 rkm 231-234 0 n H 4 0 II 0 0 10 Beam trawl 1990 rkm 29-120 0 ii 0 15 11 5 0 0 31 rkm 121-212 II n 0 2 12 2 0 II 16 1991 rkm 44-120 0 0 1 H 19 4 0 (1 33 rkm 121-218 0 (i 1 12 9 4 0 (1 26 Shrimp trawl 1988 rkm 46-120 4 3 3 6 6 3 1 3 32 rkm 121-211 HI 17 10 :il Hi 56 6 39 218 1989 rkm 38-120 3 3 3 6 is 11 7 30' 81 rkm 121-218 17 24 34 10 50 i;t 25 19 306 1990 rkm 45-120 ii 42 2!) L'l 27 37 13 32 204 rkm 121-212 0 L8 2\ (1 19 5 4 23 90 1991 rkm 45-120 n 33 29 16 15 30 31 n 154 rkm 121-212 0 24 2 20 21 1 25 0 93 ' Includes eight sampling e fforts conducted n \'(i\ rmlirr 764 Fishery Bulletin 92(4), 1994 Data analyses The developmental stages of white sturgeon eggs were determined from descriptions by Beer (1981). Timing of spawning was estimated from developmen- tal stages of eggs and temperature-egg developmen- tal data from Wang et al. (1985 ). Water temperature at the time of egg collection was used in making the estimates, and a daily index of spawning activity was calculated from these estimated spawning dates. The index of spawning activity was treated as a dichoto- mous variable: spawning occurred or did not occur on a particular day. Stepwise regression (Ryan et al., 1985) was used to determine relationships between the abundance of freshly fertilized ( stage 2) white sturgeon eggs col- lected in plankton nets at the index site and physi- cal parameters, including water temperature, tur- bidity, mean water column velocity, near-bottom water velocity, and Bonneville Dam discharge. Stage- 2 eggs were assumed to be approximately three hours old or less (Beer, 1981). Bonneville Dam discharge for these comparisons was estimated by averaging hourly discharge at the time of sampling with dis- charges during the three hours prior to sampling. White sturgeon egg abundance and all physical pa- rameters were tested for normality. Egg abundance (eggs/1,000 m3) and turbidity failed the test for nor- mality; both were normalized by using a logj0 trans- formation. Data collected just prior to, during, and just after the spawning period were used for the re- gression analyses. Transformed egg abundance was also plotted against each of the above physical pa- rameters to investigate the possibility of nonlinear relationships. The plots suggested that the relation- ship between egg abundance and water temperature may be nonlinear; therefore, we used second-degree polynomial (quadratic) regression (Ryan et al., 1985) to examine this relationship. For data analysis, YOY white sturgeon were sepa- rated from older juvenile sturgeon by length. A YOY was defined as being between 25 and 325 mm total length and less than one year old. Sturgeon shorter than 25 mm were considered larvae. A white sturgeon's birth date was assumed to be 1 January, although in reality the birth date was generally later in the year. Results Eggs The number of white sturgeon eggs collected from 1988 through 1991 ranged from 1,404 in 1988 to 2,785 in 1990 (Table 2); however, sampling effort was not equal each year. The percent of white sturgeon eggs collected in plankton nets, as opposed to artificial substrates, also varied annually, ranging from 37% in 1991 to 87% in 1989. Virtually all white sturgeon eggs were collected in the 11-km section of river ex- tending from rkm 223 to 234, immediately down- stream from Bonneville Dam. In both 1990 and 1991, four white sturgeon eggs were collected at rkm 193. In all years, 4% or less of white sturgeon eggs col- lected in plankton nets were infected with fungus, indicating infertile or dead eggs. From the spawning index, which was derived from back calculations by using the developmental stages of all eggs, we estimated that spawning occurred on 38 days in 1988, from 22 April to 22 June, and that 58% of the spawning days were in May (Fig. 2). In 1989, spawning occurred on an estimated 43 days, from 22 April to 2 July, and 53% of the spawning days were in May. In 1990, spawning was estimated to have occurred on at least 48 days, from 23 April to 14 July, and 46% of the spawning days were in May. Finally, for 1991, we estimated that spawning oc- curred on 39 days, from 5 May to 14 July, and 56% of the spawning days were in May. Water temperatures measured at Bonneville Dam and at sampling sites during the spawning period varied annually (Fig. 2). Water temperatures at Bonneville Dam sometimes differed by about 1°C from those at egg collection sites. From 1988 to 1991, white sturgeon spawned at water temperatures rang- ing from 10 to 19°C (Bonneville Dam or sampling site temperatures). Bonneville Dam discharge (mean hourly discharge by day) also varied annually (Fig. 2). The highest daily flows through Bonneville Dam during the sam- pling periods occurred during the spawning periods in 1990 and 1991. Combining data from all years, we concluded that spawning occurred on days with mean discharges ranging from 3,399 to 10,505 mVs. During the 4-year study, stage-2 eggs were col- lected at temperatures from 10 to 18°C, turbidities from 2.2 to 11.5 ntu, near-bottom velocities from 0.6 to 2.4 m/s, mean water column velocities from 1.0 to 2.8 m/s, and depths from 3 to 23 m. White sturgeon spawned primarily in the area upstream from rkm 222. Virtually all stage-2 eggs were collected between rkm 223 and rkm 234 (about 600 m downstream from the spillways at Bonneville Dam). Small numbers of stage-2 eggs were collected at rkm 193— three in 1990 and one in 1991. Exact spawning locations could not be determined because it was not possible to measure the distance that white sturgeon eggs were carried by the river current im- mediately after spawning. In addition, at least some white sturgeon eggs, which adhere to bottom sub- McCabe and Tracy: Spawning and early life history of Aapenser transmontanus 765 strate, were dislodged by water currents and carried downstream. Substrate in the river section, where stage-2 eggs were most abundant, was primarily cobble and boul- der. We are not sure of the composition of the sub- strate near rkm 193; however, there are small rocky islands in the area, and on occasion large amounts of sand were collected in the plankton net. In addi- tion, there is a rocky reef several kilometers upstream from this sampling site. At the index site, stage-2 eggs were collected over a range of environmental conditions from 1988 through 1991 (Table 3). Water temperatures ranged from 10 to 18°C, bottom water turbidities from 2.2 to Table 2 Numbers of white sturgeon, Acipenser transmontanus, eggs and larvae collected in the Colu mbia River downstream from Bonneville Dam, 1988-91. Plankton nets and artificial substrates were used to collect eggs; plankton nets and a 3.0-m beam trawl ( n 1990 and 1991) were used to collect larvae. Area refers to the geographic range (in river kilometers [rkm]) over which eggs or arvae were collected. Fungu 3-infected eggs collected in plankton nets are shown in parentheses and are included in the numbers reported for the nets. A dash ( — ) indicates that no s sampling was conducted. Sampling period Eggs Larvae Area (rkm) Net Substrate Area (rkm) Net Trawl 1988 15-30 Apr 230-231 19 — 0 — 1-15 May 228-233 163(1) 46 228-230 11 — 16-31 May 230-233 405(10) 539 193-231 71 — 1-15 Jun 226-234 112(5) 84 181-230 5 — 16-30 Jun 226-230 20(1) 16 226 3 — 1-15 Jul 0 0 0 — 16-31 Jul 0 0 0 — Total 719(17) 685 90 — 1989 15-30 Apr 230 385 47 ii — 1-15 May 224-234 275(1) 37 174-232 19 — 16-31 May 224-234 703(6) 212 181-230 39 — 1-15 Jun 222-234 640(23) 9 193-230 64 — 16-30 Jun 226-230 13 (3) 0 181-230 13 — 1-15 Jul 226 2 (1) — 0 — 16-31 Jul 0 — 0 — Total 2,018(34) 305 135 1990 15-30 Apr 230-231 386 258 0 — 1-15 May 223-234 904(38) 153 223-232 34 — 16-31 May 193-234 187 (7) 275 181-230 34 — 1-15 Jun 224-234 210(8) 260 112-230 33 — 16-30 Jun 224-230 109(20) 0 45-230 11 12 1-15 Jul 226-234 8 35 67-226 1 25 16-31 Jul 0 0 127-230 9 1 Total 1,804 (73) 9*1 152 38 1991 15-30 Apr 0 0 i) — 1-15 May 224-234 129(1) 589 i) — 16-31 May 193-234 303(3) 265 193-230 28 0 1-15 Jun 226-234 46(7) 205 193-230 17 — 16-30 Jun 224-234 227(1) 164 45-230 15 33 1-15 Jul 193-230 30(2) 50 98-230 :i7 18 16-31 Jul 0 0 0 0 Total 735(14) 1,273 127 51 766 Fishery Bulletin 92(4). 1994 5/16 5/31 6/15 6/30 Date 5/16 5/31 6/15 6/30 Date Discharge 6/15 6/30 7/15 7/30 ./' Temperature I Spawning Index Figure 2 Water temperatures ( °C ) and Bonneville Dam discharges (mean hourly water discharges by day ) from 1 Apri I through 31 July ( 1988-91 ); discharge is shown as m3/s x 1,000. Water temperatures were measured at Bonneville Dam. The spawning index shows the days on which we estimated that white sturgeon spawned. 11.5 ntu, near-bottom water velocities from 1.0 to 2.1 m/s, and mean water column velocities from 1.5 to 2.6 m/s. Estimated Bonneville Dam discharges at the index site during spawning ranged from 3,890 to 9,600 m3/s. In all years, the highest egg catches oc- curred during late April or May. Results from stepwise regression indicated that water temperature and mean water column velocity together were the best predictors of stage-2 egg col- lections at the index site; however, they explained only 27. 1% of the variation in the egg collections. The regression equation was log10 (number of eggs plus one/1,000 m3) = 1.01 - 0.110 (water temperature) + 0.637 (mean water column velocity); F = 7.99 and P = 0.001. Bottom water turbidity was an important predictor in the first three steps of the stepwise re- gression, but dropped out in the fourth and final step. If a high turbidity value and accompanying zero egg catch (18 April 1989; see Table 3) are removed from the stepwise regression, the results change. Exclud- ing the turbidity data for 18 April 1989, stepwise re- gression indicated that water temperature and turbid- ity together were the best predictors of stage-2 egg collections at the index site. The regression equa- tion was logU) (number of eggs plus one/1,000 m3) = 0.937 - 0.0857 (water temperature) + 1.70 (log10 tur- bidity); F=11.65, P=0.000, and r2=35.7%. Second-de- gree polynomial regression indicated a significant relationship between stage-2 egg collections and water temperature (F=8.70,P=0.001, and r2=28.8%). The regression equation was log10 (number of eggs plus one/1,000 m3) = -4.44 + 0.846 (water tem- perature) - 0.0326 (water temperature2). Catches of freshly fertilized white sturgeon eggs during the 12-hour collection at the index site fluc- tuated; catches ranged from 0.0 to 72.2 eggs/1,000 m3 (Table 4). On the basis of collections of freshly fertilized eggs during the 12-hour collection and day- light collections during the 4 years, it appears that adult white sturgeon spawn throughout the 24-hour day. Larvae White sturgeon larvae were collected from rkm 45 to rkm 232 in the lower Columbia River from 1988 through 1991 (Table 2), suggesting wide dispersal McCabe and Tracy: Spawning and early life history of Acipenser transmontanus 767 Table 3 White sturgeon, Acipenser transmontanus, egg (freshly fertilized) catches and accompanying physical measure- ments for the index site near rkm 230 in the lower Columbia River, 1988-91. Water temperatures (on site), turbidities, and bottom velocities were measured just above the bottom; Bonneville Dam flow was the average of the hourly discharges at the time of sampling and during the three hours prior to sampling. Velocity (m/s) Date Temperature CO Turbidity (ntu) Bottom Mean column Dam flow (m3/sx 1,000) No. eggs Eggs/ 1,000 m3 1988 25 Apr 5 May 10 May 18 May 23 May 2 Jun 8 Jun 16 Jun 20 Jun 29 Jun 1990 23 Apr 30 Apr 7 May 14 May 22 May 29 May 5 Jun 11 Jun 18 Jun 25 Jun 2 Jul 9 Jul 18 Jul 1991 30 Apr 6 May 14 May 20 May 29 May 3 Jun 19 Jun 26 Jun 3 Jul 9 Jul 15 Jul 10 11 13 13 14 14 14 16 16 17 11 11 12 12 13 14 15 15 15 17 is 19 20 10 11 12 12 11 14 16 15 18 19 1!) 4.5 6.4 6.0 6.0 5.4 6.0 3.5 3.5 3.2 3.1 3.0 3.5 3.9 3.7 3.0 3.0 3.4 6.5 5.6 3.7 2.2 1 9 2.0 3.6 3.0 4.2 5.3 6.3 11.5 4.2 3.7 4.3 2.8 3.5 1.0 1.4 1.0 1.2 1.2 1.4 1.4 1.1 1.1 1.0 1.1 1.1 1.7 1.8 1.1 1.4 1.7 1.5 2.1 1.7 1.2 1.1 1.1 1 3 1.3 1.4 2 1 1.5 2.1 1.4 ] 1 1 7 1.2 1.0 1 5 1.8 1 6 1.6 2.1 2.3 2.0 1.8 1.5 1.6 1.5 2.2 2.6 2.3 1.7 2.3 2.5 2.4 2.7 2.5 2.1 2.2 1.8 2 1 2.1 2.0 2.6 2.6 2.6 2.5 2.2 2.4 1.6 1 s 3.89 5.14 5.95 6.24 5.59 6.51 5.43 4.50 4.17 3.90 6.61 7.20 7.57 7.03 4.69 5.90 8.32 9.07 9.01 7.52 6.71 5.56 4.25 7.67 6.80 6.81 9.60 8.79 9.10 7.39 7.36 7.38 5.55 5.40 3 24 80 4 1() is 19 (i 9 0 0 225 32 292 0 39 L9 23 0 2 1 i) o 0 2 98 L5 2 11 0 12 9 0 0 1.7 19.5 54.1 2.1 6.4 10.3 12.4 0.0 6.9 0.0 1989 18 Apr 10 12.0 1.5 1.9 6.46 0 0.0 27 Apr 12 10.0 1.5 2.2 7.11 348 219.9 1 May 12 7.5 1.7 2.0 6.96 17 10.3 10 May 13 7.5 — 2.4 7.89 70 40.9 17 May 14 9.2 1.6 2.3 8.30 525 284.5 24 May 14 5.9 1.4 2.2 6.82 18 10.8 1 Jun 14 5.4 1.6 2.2 6.80 1 0.6 8 Jun n; 5.3 2.0 2.6 7.66 99 58.3 15 Jun 17 3.4 1.5 2.2 5.91 o 0.0 21 Jun 17 3.7 1.0 1.9 4.60 0 0.0 28 Jun 18 3.3 1.1 1.6 3.98 0 0.0 6 Jul 19 3.5 1.0 1.6 3.72 i) 0.0 0.0 133.2 14.8 166.9 0.0 25.6 10.0 12.9 0.0 11 0.5 0.0 0.0 0.0 1.2 56.5 14.4 1.4 7.1 0.0 7 5 5.2 0.0 0.0 768 Fishery Bulletin 92(4), 1994 Table 4 Summary of white sturgeon, Acipenser transmon tan us, egg (freshly fertilized) and larval collections during a 12-h study at the index site near rkm 230 in the lowei Columbia River. Sampling was done from 1843 hours on 25 May to 0623 hours an 26 May 1988 with a plankton net. Bonneville Dam flow was the average of hourly discharges at the time of samph ng and during the 3 hours prior to sampling. Eggs Larvae Sampling times Bonneville Dam (h)' flow (m3/s x 1,000) No. No./l,000 m3 No. No./l.OOOm'1 1843-1943 7.24 4 2.8 9 6.4 195 1-205 l2 7.40 3 2.3 (1 0.0 2100-2200 7.00 0 0.0 5 3.3 2206-2306 6.74 108 72.2 7 4.7 2314-0014 6.50 7 4.7 9 6.0 0020-0120 6.51 27 18.2 9 6.1 0128-0228 6.58 34 21.9 3 1.9 0234-0334 6.66 3 2.0 7 4.7 0340-0440 6.72 1 0.7 8 5.3 0445-0545 6.77 32 20.6 1 0.6 0553-0623 6.76 is 22.6 7 8.8 ; Sunset on 25 May was at 2030 hours; sunrise on 26 May was at 0515 hours. 2 Questionable sampling effort; net was damaged. after hatching. River kilometer 45 is located in the upper end of the Columbia River estuary (Fig. 1); however, this section of the estuary is a freshwater environment. Larvae were collected from early May through late July, reflecting a protracted spawning period (Table 2). All white sturgeon larvae in 1988 and 1989, and 71% or more in 1990 and 1991, were collected in plankton nets. In 1988 and 1989, larvae were not collected as far downstream as in 1990 and 1991. Undoubtedly, smaller areas of capture in 1988 and 1989 were due to lack of sampling with the 3.0- m beam trawl in these years. All white sturgeon lar- vae collected in the upper estuary in 1990 and 1991 were collected in the beam trawl. Larvae were col- lected at depths ranging from 4 to 29 m. When the larvae were collected in plankton nets, they were most likely being transported by water currents, be- cause the nets were fished from an anchored boat. Catches of white sturgeon larvae during the 12- hour collection at the index site fluctuated with catches ranging from 0.0 to 8.8 larvae/1,000 m3 (Table 4). Young of the year Annual catches of YOY white sturgeon varied con- siderably, ranging from 11 in 1988 to 273 in 1990 (Table 5). Annual catches shown in Table 5 are not necessarily indicative of YOY abundance in respec- tive years, because sampling gears and schemes were not the same each year. In 1988 and 1989, the 3.0-m beam trawl was not used, whereas in 1990 and 1991 it was used. The beam trawl was more effective at capturing small YOY white sturgeon than was the 7.9-m semiballoon shrimp trawl. Also, in 1990 and 1991, more sampling was conducted in the lower 120 km of the river than in 1988 and 1989. On the basis of sampling from 1988 through 1991, it appears that YOY white sturgeon are primarily using the section of river extending from rkm 45 to 166 (Table 5). Relatively few YOY white sturgeon were collected in the 68 km of river between Bonneville Dam (rkm 234) and rkm 166; small catches were made at rkm 211 in July 1990 and Sep- tember 1991. In 1990 and 1991, YOY white sturgeon were first captured in late June, less than two months after spawning was estimated to have begun. In all four years, YOY white sturgeon appeared to grow well during their first summer; however, monthly mean lengths and weights varied among years (Table 5). During all years, YOY white sturgeon reached a mini- mum mean total length of 176 mm and a minimum mean weight of 30 g by the end of September. No statistical comparisons among years were done be- cause of small sample sizes, the protracted spawn- ing period of white sturgeon, and the fact that YOY white sturgeon were collected throughout the month. The YOY white sturgeon were more abundant in deeper areas of the lower Columbia River, at least during daylight; mean minimum depths during trawling efforts in which YOY were captured were >12.5 m in all years. Mean maximum depths at which YOY white sturgeon were captured were > 15.8 m in all years. Bottom substrate over which YOY white McCabe and Tracy: Spawning and early life history of Aapenser transmontanus 769 Table 5 Summary of young-of-the-year white sturgeon, Acipenser transmontanus, stream from Bonneville Dam, 1988-91. SD = standard deviation. catches in the Columbia River down- Month Capture location (rkm) Number Total length (mm) We ight (g) Mean SD Mean SD 1988 Jul 126 1 86.0 0.0 3.0 0.0 Aug 127-153 2 134.0 41.0 13.0 9.9 Sep 153 2 235.0 35.4 60.5 29.0 Oct 127-162 6 248.3 9.8 68.2 8.9 Total 11 1989 Jul 49-153 17 93.4 25.8 5.0 3.1 Aug 49-153 15 176.7 29.9 31.6 13.5 Sep 46-153 12 224.4 30.4 59.7 18.7 Oct 49-162 56 269.4 23.5 87.4 18.5 Nov' 107-120 11 273.8 17.7 90.4 20.2 Total 111 19902 Jun 45-120 7 32.1 4.3 <1.0 <1.0 Jul 45-211 125 75.6 27.3 3.2 2.8 Aug 50-166 79 123.8 37.5 12.3 10.4 Sep 49-166 14 222.6 28.4 54.4 19.8 Oct 46-166 48 224.4 28.5 51.9 17.4 Total 273 19913 Jun 45-166 27 30.4 4.1 <1.0 <1.0 Jul 45-166 89 55.7 17.8 1.3 1.2 Aug 49-127 55 97.1 27.6 6.1 4.8 Sep 45-211 47 176.4 38.3 29.8 16.2 Total 218 1 Sampling for November was conducted on 1 No ~ Includes samples collected at rkm 75 from 31 J 3 No sampling was done in October 1991. vember 1989. jly to 1 August 1990. sturgeon were found was predominantly sand; how- ever, much of the bottom in the lower Columbia River is composed of sand. In addition, the bottom trawls could not be used in rocky areas. During the 20-hour sampling survey from 31 July to 1 August 1990 (sampled from 1155 through 0800 hours) at rkm 75, 52 YOY white sturgeon were col- lected (Table 6). Over 78% of YOY white sturgeon were collected during hours of darkness, indicating that they were more vulnerable to the trawl at night or that they moved into the sampling area at night. The YOY were collected at depths that ranged from 11 to 15 m. Discussion White sturgeon successfully spawned in the lower Columbia River in all years of the study. All white sturgeon eggs collected downstream from Bonneville Dam were probably released by sturgeon spawning in this area and not by sturgeon spawning in the impoundment created by Bonneville Dam. Although white sturgeon spawn in the impoundment upstream from Bonneville Dam (Miller et al.4), it is unlikely that any of these eggs are carried through Bonneville Dam. In 1990, Miller et al.4 collected white sturgeon eggs between rkm 298 and 308. The locations of white sturgeon egg collections upstream from Bonneville 4 Miller, A. I., P. J. Anders, M. J. Parsley, C. R. Spraque, J. J. Warren, and L. G. Beckman. 1991. Report C. In A. A. Nigro (ed.). Status and habitat requirements of the white sturgeon populations in the Columbia River downstream from McNary Dam, p. 82-144. Ann. Rep. to Bonneville Power Admin. (Project 86-50) by Oreg. Dep. Fish Wildl., Wash. Dep. Fish., Natl. Mar. Fish. Serv, and U.S. Fish Wildl. Serv. Avail. Bonneville Power Admin., P.O. Box 3621, Portland, OR 97208. 770 Fishery Bulletin 92|4), 1994 Dam strongly suggest that eggs are found only in the upper Bonneville Pool, since rkm 298 is about 64 km upstream from the dam. Spawning in the lower Columbia River in 1988- 91 occurred during temperature regimes for success- ful egg incubation. Successful egg incubation for white sturgeon occurs at temperatures between 10 and 18°C; highest survival and uniform hatching occur between 14 and 16°C (Wang et al., 1985). In our study, we estimated that peak spawning occurred at water temperatures of 12 to 14°C. We estimated that some spawning occurred at water temperatures of 18 or 19°C. Survival for these eggs was probably less than for eggs spawned at lower water tempera- tures. Wang et al. (1985) observed that substantial white sturgeon embryo mortalities may occur at water temperatures of 18 to 20°C and that tempera- tures greater than 20°C are clearly lethal. In the Sac- ramento River, Kohlhorst (1976) observed that water temperatures during the white and green sturgeon spawning period ranged from 7.8 to 17.8°C and that peak spawning occurred at about 14.4°C. It should be noted that Kohlhorst's estimates of the spawning pe- riod are based on back calculations of larval ages, rather than on sturgeon eggs. Spawning dates can be more accurately estimated by using eggs rather than larvae. Sampling for white sturgeon larvae was done with gear that sampled along or very near the bottom; therefore, no data were collected regarding vertical distribution of white sturgeon larvae. However, Stevens and Miller (1970) reported that white or green sturgeon larvae, or both, are primarily demersal in the Sacramento-San Joaquin River System. They caught 33 larvae in 16 bottom sampling efforts and only one larva in eight surface and midwater efforts. River currents disperse white sturgeon larvae out of spawning and egg incubation areas. Stevens and Miller (1970) noted a direct relationship between river flow and catches of white or green sturgeon lar- vae, or both, in the Sacramento-San Joaquin Delta. In a laboratory experiment, Brannon et al.5 observed that white sturgeon larvae swam up the water column after hatching. In addition, Brannon et al.5 found that the behavior of white sturgeon larvae was affected by current velocity in laboratory experiments. There was an inverse relationship between water velocity and the amount of time larvae spent in the water column. Dispersal of white sturgeon larvae over a wide area is probably very important in maintaining a stable population of white sturgeon in the lower Columbia River. Wide dispersal allows utilization of more feed- ing areas and rearing habitats by larval and postlar- val white sturgeon and minimizes competition for these limited resources. However, it is also impor- tant that white sturgeon not be carried into saline portions of the Columbia River estuary. Brannon et al.5 found that salinities >16 ppt killed white stur- geon larvae and fry. Food resources for YOY white sturgeon in many of the deeper areas (>12 m) of the lower Columbia River are probably not abundant. Little is known about the diet of YOY white sturgeon in the lower Columbia River; however, limited observations suggest that the amphipod Corophium salmonis is the primary prey (Muir et al., 1988). Densities of C. salmonis in many of the deeper areas probably are low because of un- stable substrates. Corophium salmonis is a tube- builder and requires a more stable substrate to densely populate an area. In 1990, densities of C. salmonis at a deep area ( 19-21 m) at rkm 153 aver- 5 Brannon, E., S. Brewer, A. Setter, M. Miller, F. Utter, and W. Hershberger. 1985. Columbia River white sturgeon (Acipenser transmontanus) early life history and genetics study. Final Rep. to Bonneville Power Admin. iProject 83-316) by Univ. Wash, and Natl. Mar. Fish. Serv., Seattle, 68 p. Avail. Bonneville Power Admin., P.O. Box 3621, Portland, OR 97208. Table 6 Summary of young-of-the-year white sturgeon, Acipenser transmontanus, catches during a 20-hour study at rkm 75 in the lower Columbia River, 31 July-1 August 1990. Sampling was done from 1155 hours on 31 July to 0800 hours on 1 August with a 7.9-m semiballoon shrimp trawl. Depth Length Depth Length Hour' range No. No./ha range (mm) Hour' range No. No./ha range (mm) 1155 13-14 m 0 0 2130 12-14 m 12 39 57-120 1.357 13-14 m 0 0 — 2230 13-15 m 12 38 54-122 1533 10-15 m i) 0 — 0030 12-14 m 15 55 61-141 1702 13-14 m 1 4 79 0218 12-14 m 2 10 S2 86 1830 12-14 m 1 1 108 0527 13 m 0 0 — 1933 11-14 m 6 2\ 79 114 0646 12-13 m 2 in 86-113 2029 12-14 m 1 4 79 0800 12-13 m ii 0 — Sunset on 31 July was at 2047 hours; sunrise on 1 August was at 0555 hours. McCabe and Tracy: Spawning and early life history of Acipenser transmontanus 77! aged less than 105 organisms/m2 in June through September (McCabe and Hinton6). However, in a deep area at rkm 120 that had large numbers of YOY white sturgeon, the density of C. salmonis was relatively high in August 1990 (2,289/m2) but dropped to 433 or- ganisms/m2 in September (McCabe and Hinton6). More research is needed to assess the abundance of benthic organisms in rearing areas of YOY white sturgeon. Although prey abundance may be low in many of the deeper areas of the lower Columbia River, the sub- strate in these areas is probably ideal for efficient feed- ing by YOY white sturgeon. The white sturgeon has a protrusible mouth that is used to suck prey from the bottom. In a laboratory experiment with juvenile Rus- sian sturgeon, Acipenser gueldenstaedti, Sbikin and Bibikov (1988) observed that juveniles (<130 mm) pre- ferred even, sandy bottoms to bottoms with stones or depressions. Juveniles avoided vegetated areas. Apparently YOY white sturgeon are very effective and efficient predators on prey found in the rearing areas, as evidenced by their rapid growth during the summer and early fall. The YOY white sturgeon reached a mean total length of at least 176 mm by the end of September. Rapid growth during the first growing season reduces natural mortality; by the end of summer or fall, YOY white sturgeon in the lower Columbia River probably have few natural predators. Sampling equipment used to collect YOY white sturgeon in the lower Columbia River was limited to two types of bottom trawls that could not be used in shallow littoral areas. Observations made during other studies suggest that YOY white sturgeon do not use shallow littoral areas. No YOY white sturgeon have been collected in intensive beach seining efforts at rkm 75 during the last 15 years.7 Most sampling was done during daylight; limited sampling was done at night. The beach seining location was adjacent to the sampling site where 52 YOY white sturgeon were col- lected during a 20-hour study in 1990. No YOY white sturgeon were collected in backwaters and shoreline 6 McCabe, G. T., Jr., and S. A. Hinton. 1991. Report D. In A. A. NigTO led.), Status and habitat requirements of white sturgeon populations in the Columbia River downstream from McNary Dam, p. 145-180. Ann. Rep. to Bonneville Power Admin. (Project 86-50) by Oreg. Dep. Fish Wildl., Wash. Dep. Fish., Natl. Mar. Fish. Serv., and U.S. Fish Wildl. Serv. Avail. Bonneville Power Admin., P.O. Box 3621, Portland, OR 97208. ' Richard D. Ledgerwood, National Marine Fisheries Service, P.O. Box 155, Hammond, Oregon 97121. Personal commun., 1992. 8 McCabe, G. T., Jr., S. A. Hinton. and R. J. McConnell. 1989. Report D. In A. A. Nigro (ed. ). Status and habitat requirements of white sturgeon populations in the Columbia River down- stream from McNary Dam, p. 167-207. Ann. Rep. to Bonneville Power Admin. (Project 86-50) by Oreg. Dep. Fish Wildl., Wash. Dep. Fish., Natl. Mar. Fish. Serv., and U.S. Fish Wildl. Serv. Avail. Bonneville Power Admin., P.O. Box 3621, Portland, OR 97208. areas during limited beach seining tows in the lower Columbia River in August 1988 (McCabe et al.8). We conclude that white sturgeon spawned success- fully in the lower Columbia River during the period 1988 through 1991. Collection of YOY white sturgeon indicated that recruitment occurred in all years. Acknowledgments We thank personnel from the National Marine Fish- eries Service, Hammond, Oregon, and the Washing- ton Department of Fish and Wildlife, Battle Ground, Washington, who assisted in field sampling or sample analyses. Also, we thank Benjamin Sandford for as- sistance with the statistical analysis and two anony- mous reviewers for their constructive comments. The study was funded primarily by the Bonneville Power Administration. Literature cited Beer, K. E. 1981. Embryonic and larval development of white sturgeon (Acipenser transmontanus). M.S. thesis, Univ. Calif., Davis, 93 p. Buchanan, T. J., and W. P. Somers. 1969. Discharge measurements at gaging stations. Chapter A8, 65 p. Techniques of water-resources investigations of the United States Geological Sur- vey, Book 3, Applications of hydraulics. U.S. Gov. Printing Office, Washington, D.C. Cochnauer, T. G., J. R. Lukens, and F. E. Partridge. 1985. Status of white sturgeon, Acipenser transmontanus, in Idaho. In F. P. Binkowski and S. I. Doroshov (eds.), North American sturgeons: biology and aquaculture potential, p. 127-133. Dr. W. Junk, Dordrecht, Netherlands. Craig, J. A., and R. L. Hacker. 1940. The history and development of the fisheries of the Columbia River. U.S. Dep. Interior, Bur. Fish., Bull. 32( 491:133-216. Kohlhorst, D. W. 1976. Sturgeon spawning in the Sacramento River in 1973, as determined by distribution of larvae. Calif. Fish Game 62:32-40. McCabe, G. T., Jr., and L. G. Beckman. 1990. Use of an artificial substrate to collect white sturgeon eggs. Calif. Fish Game 76:248-250. Muir, W. D., R. L. Emmett, and R. J. McConnell. 1988. Diet of juvenile and subadult white sturgeon in the lower Columbia River and its estuary. Calif. Fish Game 74:49-54. Parsley, M. J., L. G. Beckman, and G. T. McCabe Jr. 1993. Spawning and rearing habitat use by white sturgeons in the Columbia River downstream from McNary Dam. Trans. Am. Fish. Soc. 122:217-227. 772 Fishery Bulletin 92(4). 1994 Ryan, B. F., B. L. Joiner, and T. A. Ryan Jr. 1985. Minitab handbook. PWS-KENT Pub. Co., Boston, MA, 386 p. Sbikin, Yu. N., and N. I. Bibikov. 1988. The reaction of juvenile sturgeons to elements of bottom topography. J. Ichthyol. 28:155-160. Scott, W. B., and E. J. Crossman. 1973. Freshwater fishes of Canada. Bull. Fish. Res. Board Can. 184, 966 p. Stevens, D. E., and L. W. Miller. 1970. Distribution of sturgeon larvae in the Sacra- mento-San Joaquin River system. Calif. Fish Game 56:80-86. Wang, Y. L., F. P. Binkowski, and S. I. Doroshov. 1985. Effect of temperature on early development of white and lake sturgeon, Acipenser transmontanus and A. fulvescens. Environ. Biol. Fishes 14:43-50. Abstract. — Little is known about cetaceans in the oceanic Gulf of Mexico (depths >200 m). From July 1989 to June 1990, we con- ducted aerial surveys in the oceanic north-central Gulf (long. 87.5°W- 90.5°W) with the following objec- tives: 1) to determine which ceta- cean species were present; 2) to document temporal and spatial dis- tribution for each species; and 3) to estimate relative abundance for each species. We surveyed a total of 20,593 transect km and sighted at least 18 species. Of 278 identi- fied herds (6,084 animals), 94% of the herds and 98% of the animals represented seven species or spe- cies groups: Risso's dolphin. Gram- pus griseus (22% of the herds, 13% of the animals); sperm whale, Phy- seter macrocephalus (16%, 1%); bot- tlenose dolphin, Tursiops truncatus 14%, 7%); Atlantic spotted dolphin, Stenella frontalis (13%, 15%); pygmy sperm whale, Kogia bre- viceps, and dwarf sperm whale, Ko- gia simus (12%, 1%); striped dol- phin, Stenella coeruleoalba, spin- ner dolphin, S. longirostris , and clymene dolphin, S. clymene (9%, 34%); and pantropical spotted dol- phin, S. attenuata (8%, 27%). Each of these species or species groups was sighted throughout the area surveyed in at least three seasons. Mean water depths of bottlenose dolphin and Atlantic spotted dol- phin sightings were less than 400 m; mean water depths of Risso's dolphins and pygmy and dwarf sperm whales were between 400- 600 m; and mean water depths of striped, spinner, and clymene dol- phins, sperm whales, and pantropical spotted dolphins were greater than 700 m. Mean herd sizes varied by species and species groups and ranged from 1.9 ani- mals for pygmy and dwarf sperm whales to 87.8 animals for striped, spinner, and clymene dolphins. Cetaceans on the upper continental slope in the north-central Gulf of Mexico Keith D. Mullin Wayne Hoggard Carol L. Roden Ren R. Lohoefener* Carolyn M. Rogers Southeast Fisheries Science Center National Marine Fisheries Service, NOAA PO Drawer 1207. Pascagoula. Mississippi 39568 Brian Taggart NOAA Aircraft Operations Center PO Box 6829. MacDill Air Force Base, Florida 33608 Manuscript accepted 9 May 1994. Fishery Bulletin 92:773-786. The Gulf of Mexico encompasses an area of over 1,500,000 km2 and has an average depth of 1,700 m (Gore, 1992). The continental shelf (depths <200 m) is wide (up to 260 km) in most parts of the northern Gulf (Fig. 1). Directed studies (Fritts et al., 1983; Scott et al.1 ) and opportunistic sightings (Schmidly, 1981; Rade- macher2) have suggested that only the bottlenose dolphin, Thrsiops trun- catus, and the Atlantic spotted dol- phin, Stenella frontalis, are common in most continental shelf waters of the U.S. Gulf. However, there are records (primarily from strandings) of 29 cetacean species from the Gulf (Schmidly, 1981; Perrin et al., 1981; Hersh and Odell, 1986; Perrin et al., 1987; Bonde and O'Shea, 1989; Bar- ron and Jefferson, 1993). Therefore, if species other than the bottlenose dolphin and the Atlantic spotted dol- phin are represented in substantial numbers, their distributions must be primarily oceanic (depths >200 m). Mineral deposits have been mined widely in U.S. Gulf shelf waters west of Mobile, Alabama, and as of 1988, over 4,500 drilling structures have been in use for oil and gas production. Mineral devel- opment on the continental slope (depths 200-2,000 m) in the central and western Gulf has begun and additional exploratory drilling is being planned. Before large-scale exploration, development, and pro- duction can take place, an assess- ment of cetacean diversity, distribu- tion, and abundance is required to satisfy the intent of the U.S. Ma- rine Mammal Protection Act and the U.S. Endangered Species Act. Both acts mandate that federal agencies take appropriate actions to ensure that their activities do not Present address: U.S. Fish and Wildlife Service, 3100 University Boulevard, South Suite 120, Jacksonville, Florida 32216 1 Scott, G. P., D. M. Burn, L. J. Hansen, and R. E. Owen. 1989. Estimates of bottlenose dolphin abundance in the Gulf of Mexico from regional aerial surveys. U.S. Dep. Commer., NOAA, Nat. Mar. Fish. Serv., Southeast Fish. Sci. Cent., Miami Labora- tory, 75 Virginia Beach Drive, Miami, FL 33149. Admin. Rep. CRD-88/89-07, 24 p. 2 Rademacher, K. R. 1991. Opportunistic sightings of cetaceans in the Gulf of Mexico from NOAA Ship Chapman, 1989-90. U.S. Dep. Commer., NOAA, Nat. Mar. Fish. Serv., Southeast Fish. Sci. Cent., Pascagoula Facility, P.O. Drawer 1207, Pascagoula, MS 39568. Unpubl. data. 773 774 Fishery Bulletin 92(4). 1994 contribute to the demise of endangered species or to the depletion of marine mammal populations. To as- sess potential impacts of oil and gas activities on marine mammal populations, it is imperative that we know when, where, and how many marine mam- mals may be vulnerable to such activities. Only limited data from strandings, opportunistic sightings (Schmidly, 1981; Mead3), and aerial sur- veys (Fritts et al., 1983) are currently available to assess these parameters for oceanic cetaceans in the Gulf. In July 1989, the U.S. Minerals Management 3 Mead, J. G. 1992. Marine mammal strandings. National Mu- seum of Natural History, Smithsonian Institution, Washington, D.C. 20560. Unpubl. data. Service and the Southeast Fisheries Science Center (SEFSC) began aerial surveys of cetaceans on the upper continental slope in the north-central Gulf. The objectives of the surveys were 1) to determine which species were present; 2) to document temporal and spatial distribution for each species; and 3) to esti- mate relative abundance for each species. Methods The study was conducted in two phases. Phase 1 was a five-month pilot study carried out from July through November 1989. The primary objective of Phase 1 was to determine which species of cetaceans, 97 58 32039' 24 05 97 58' 24 05' 79°46' 26''23■ 8Z;48' OOV 30°461 30°46' 26^23' 90°34' 87°48' Figure 1 Location of the area surveyed for cetaceans in the north-central Gulf of Mexico (top panel). The locations of the survey blocks for Phase 1 (July-November 1989) are shown in the lower left panel and for Phase 2 (January-June 1990), in the lower right panel. Mullin et al.: Cetaceans of the north-central Gulf of Mexico 775 if any, inhabited the upper continental slope in the north-central Gulf. Since studies elsewhere indicated that cetaceans may concentrate in areas of high sea- floor relief (Hui, 1979; Payne et al., 1986; Kenney and Winn, 1986; Selzer and Payne, 1988), three sur- vey blocks were initially selected on the upper conti- nental slope (Fig. 1, Blocks A1-A3; Table 1): 1) the Upper Mississippi Fan, 2) the Mississippi Canyon, and 3) an area of submarine salt domes. The DeSoto Canyon survey block (Block A4) was added in Sep- tember when additional flight time was available. The Mississippi Canyon survey block was shifted southeast to include deeper waters for the October and November surveys. This was done because sperm whales, Physeter macrocephalus, were sighted near the 1,000-m isobath during September in the Upper Mississippi Fan survey block. The sperm whale is listed as an endangered species under the U.S. En- dangered Species Act (USFWS, 1989), and we were interested in denning its distribution. The results of Phase 1 indicated that a variety of cetaceans were relatively abundant on the upper con- tinental slope; ten species and 171 herds were sighted. Therefore, Phase 2 was implemented and monthly surveys, except for December, were contin- ued to complete a full year period. Phase 2 was conducted from January through June 1990. The study area selected for Phase 2 consisted of seven adjacent blocks 30 minutes wide (48.7 km) that extended from long. 87.5'W to long. 90.5'W (Fig. 1, Blocks B1-B7). The northern border of the study area, except near the Mississippi Canyon, generally followed the 200-m isobath. Each block extended 44.1 km south of its northern border. The Mississippi Canyon was not surveyed because we wanted to focus on oceanic waters. Because of the shape of the canyon, surveys of oceanic waters would have been logistically inefficient. We also be- lieved that the results of Phase 1 established the canyon as important cetacean habitat; eight species were identified in 54 sightings. Aerial surveys were conducted when the sea state was Beaufort 0—4 and visibility was good and were designed to sample blocks A1-A4 at least twice each month during Phase 1, and blocks B1-B7 twice each month during Phase 2. Line transect sampling meth- ods were used (Buckland et al., 1993) although line transect analyses are not presented here. The survey aircraft, a DeHavilland Twin Otter with a large plexiglass bubble window on each side that allowed observers to view an area on both side of the transect line, was flown at an altitude of 750 feet (229 m) and a speed of 110 knots (204 km/h). Transects that were uniformly spaced from a ran- dom starting point were surveyed in each block (Table 1). Transects ran north-south, perpendicular to the bathymetry. One observer was stationed at each bubble window and one at a computer station. Ob- servers rotated every 30 minutes to avoid fatigue. The bubble windows were divided into seven 10' Table 1 Summary of area, water depth, transect length, number of transects, and effort per month for each survey block in the Gulf of Mexico. Transect kilometers of effort per month and block Range of Transect Number Area water length of Block (km2) depths (m) (km) transects' 1989 1990 Jul Aug Sep Oct Nov Jan Feb Mar Apr May Jun Total 934 677 956 634 920 — — — — — — 4,121 440 447 396 — — — — — — — — 1,283 — — 499 167 — — — — — — 666 394 489 412 0 164 — — — — — — 1,459 — — 176 356 535 — — — — — — 1,067 0 179 362 362 269 265 1,437 178 178 223 178 352 717 1,826 — — — — 357 176 631 360 445 449 2,418 231 0 355 357 170 544 1,657 — — — — — 128 0 358 360 361 361 1,568 — — — 312 0 178 273 358 357 1,478 356 0 180 361 354 362 1,613 1,768 1,613 1,940 1,489 1,786 1,562 533 2,287 2,251 2,309 3,055 20,593 Al 2,099 18-1,317 46.3 5 A2a 2,255 29-573 55.5 3 A2b 2,255 134-966 55.5 3 A3 2.640 104-1,152 55.5 :s A4 1,180 66-2,003 59.2 3 Bl 2,160 168-1,792 44.1 4 B2 2,160 139-1,710 44.1 ■1 B3 2,160 163-1,070 44.1 1 B4 2,160 183-1,125 44.1 4 B5 2,160 230-979 44.1 4 B6 2.160 152-933 44.1 4 B7 2.160 176-1,098 44.1 4 Total ' Number of planned transects each time the block was surveyed. 776 Fishery Bulletin 92(4), 1994 sighting intervals corresponding to perpendicular distances from the transect line of 40, 83, 132, 192, 273, 397, and 629 m. Observers searched on and near the transect line and scanned periodically out to 629 m. Sighting cues beyond 629 m were ignored unless the observer was certain it was a cetacean. When cetaceans were encountered, the sighting interval was noted and the herd was circled. Before continuing on the transect, the herd was identified and its size estimated. The identifying characteris- tics of each cetacean species were noted. Data were entered on a computer interfaced with a LORAN-C navigation receiver. Latitude, longitude, and head- ing were automatically recorded with each data record. Cetaceans were identified to the lowest taxonomic level possible from descriptions in field guides by Leatherwood et al. (1976) and Leatherwood and Reeves ( 1983). Our ability to make an identification was dependent on water clarity, sea state, and ani- mal behavior. We were not able to distinguish spe- cies of some genera or groups of species. These groups included 1) the species ofMesoplodon; 2) the melon- headed whale, Peponocephala elect ra, and pygmy killer whale, Feresa attenuata; 3) the dwarf sperm whale, Kogia simus, and pygmy sperm whale, K. breviceps; and 4) the short-finned pilot whale, Globicephala macrorhynchus , and long-finned pilot whale,4 G. melaena. Cuvier's beaked whale, Ziphius cauirostris, and Mesoplodon spp. could not always be distinguished and these sightings were classified as unidentified ziphiids. While we did make positive identifications of striped dolphins, S. coeruloealba, spinner dolphins, S. longirostris, and clymene dol- phins, S. clymene, from photographs, they were usu- ally difficult to distinguish in the field and were grouped together for analyses. In some cases, ani- mals could only be identified as large cetaceans ( greater than about 7 m ) or small cetaceans (less than about 7 m). For species or species groups sighted 20 or more times, the null hypothesis that water depth did not vary among species or species groups was tested with one-way analysis of variance. If the null hypothesis was rejected, Duncan's multiple-range test was used to determine where significant differences in mean water depths occurred. Sighting rates of herds and individuals were used as measures of overall, temporal, and spatial rela- tive abundance. Seasons were defined as summer (June-August), fall (September-November), winter (January-February), and spring (March-May). To summarize spatial relative abundance, the area sur- 4 Only the short-finned pilot whale is known to inhabit the Gulf of Mexico (Sehmidly, 1981). veyed was divided into an eastern zone (Blocks A4, Bl, and B2), a central zone (Blocks Al, A2, B3, B4, and B5), and a western zone (Blocks A3, B6, and B7). All sightings from each season and zone were pooled. For each season and for each zone, the sighting rate of herds (herds/100 transect km) and animals (ani- mals/100 transect km) of each species or species group was calculated. We also compared the relative abundance of indi- viduals of each species or species group from our sur- veys to those from the Gulf stranding database (Mead3). The database of Gulf strandings contained 2,321 records identified to species. Only 516 records (22%) were not those of bottlenose dolphins. To com- pare our results with these data, we excluded bottle- nose dolphins and unidentified cetaceans from both data sets. We used our species or species-group cat- egories and calculated the relative abundance of each within each data set as a percentage of the total num- ber of animals. Results In total, we sighted 320 herds (7,438 animals) and identified 18 species of cetaceans (Table 2); 45 herds (14%) could not be identified. Of the 275 identified herds (6,084 animals), 93.5% of the herds and 97.7% of the animals consisted of seven species or species groups: Risso's dolphins (herds, 22.2%; animals, 12.6%); sperm whales (15.6%, 1.4%); bottlenose dol- phins (14.2%, 7.4%); Atlantic spotted dolphins ( 12.8%, 15.0%); pygmy and dwarf sperm whales (11.6%, 1.0%); striped, spinner, and clymene dolphins (8.7%', 33.8%); and pantropical spotted dolphins (8.4%, 26.5%). Mean herd sizes of species or species groups sighted more than 20 times ranged from 1.9 to 87.8 animals (Table 2). The largest herd consisted of 325 striped, spinner, or clymene dolphins (SSC dolphins). Dol- phins of the genus Stenella had the largest mean herd sizes and the largest ranges of herd sizes. However, the mean herd sizes of pantropical spotted dolphins and SSC dolphins were each about three times that of the Atlantic spotted dolphin. The mean herd sizes of sperm whales and pygmy and dwarf sperm whales were close to two, and they exhibited the smallest ranges of herd sizes. Bottlenose dolphins and Risso's dolphins had similar means and ranges of herd sizes. Mean water depths of species or species groups sighted 20 or more times ranged from 257 to 905 m (Table 2). Differences between these means were sta- tistically significant (Table 3). Mean water depths of pantropical spotted dolphin, sperm whale, and SSC dolphin sightings were the largest (>700 m). Each of Mullin et al.: Cetaceans of the north-central Gulf of Mexico 777 Table 2 Cetaceans sighted, mean herd size (H) and mean water depth (W; n = number of herds) from aerial surveys con- ducted in the Gulf of Mexico from July 1989 to June 1990 (%CV=percent coefficient of variation). Species or species group Herd size (animals) Water depth (meters) n H SE Range W SE Range %CV Risso's dolphin (Grampus griseus) 61 12.8 1.46 1-48 440 25.5 97-1,079 46 Sperm whale (Physeter macrocephalus) 43 2.1 0.30 1-9 877 35.5 199-1,573 27 Bottlenose dolphin (Tursiops truncatus) 39 11.9 2.23 1-60 257 41.0 20-973 100 Atlantic spotted dolphin (Stenella frontalis) 35 26.6 5.15 2-137 367 40.3 91-1,152 65 Pygmy/dwarf sperm whales [Kogia breviceps/simus) 32 1.9 0.20 1-4 544 63.8 96-1,780 65 SSC dolphins1 (S. eoeruleoalba/longirostris/elymene ) 24 87.8 20.44 8-325 712 76.3 93-1,567 53 Pantropical spotted dolphin (S. attenuata) 23 71.8 9.38 7-186 905 76.6 65-1,566 39 Pilot whale (Globicephala sp.) 5 18.2 3.73 5-28 605 71.3 364-781 28 Cuvier's beaked whale [Ziphius cavirostris) 3 1.3 0.33 1-2 1,268 275.1 916-1,810 38 Mesoplodon sp. 1.0 — — 910 — — — Unidentified ziphiids (Mesoplodon/Ziphius) 1.3 0.33 1-2 668 238.2 204-993 62 Pygmy killer/melon-headed whales ( Feresa/Peponoeephala ) 25.0 — — 318 — — — False killer whale (Pseudorca crassidens) 3.0 — — 1,107 — — — Killer whale {Orcinus orca ) 8.0 — — 964 — — — Rough-toothed dolphin (Steno bredanensis) 4.0 — — 933 — — — Fin whale (Balaenoptera physalus) - 1.0 — — 148 — — — Bryde's whale (B. edeni) 1.0 — — 342 — — — Unidentified small cetacean 40 30.1 10.38 1-325 530 68.3 87-1,779 82 Unidentified large cetacean 5 1.8 0.45 1-3 857 288.2 316-1,673 57 1 S. coeruleoalba, S. longirostris and S. clymene were each positi vely identified at least once. these species inhabited a range of water depths greater than 1,300 m. However, most sperm whales inhabited a very narrow range of water depths (Fig. 2). The mean depth of sperm whale sightings had a coefficient of variation of 0.27, the lowest of any spe- cies or species group (Table 3). Bottlenose dolphins, Atlantic spotted dolphins, and Risso's dolphins had the shallowest mean water depths (<450 m) and in- habited a range of water depths less than 1,100 m. Cuvier's beaked whales were only sighted at depths greater than 900 m. Cetaceans were widely distributed between sea- sons (Table 4). Five species or species groups were sighted in every season of the year: sperm whales, bottlenose dolphins, Risso's dolphins, Atlantic spotted dolphins, and SSC dolphins. Pygmy and dwarf sperm whales and pantropical spotted dolphins were sighted in all seasons except winter and Cuvier's beaked whale, 778 Fishery Bulletin 92|4), 1994 Table 3 Duncan's multiple range test of mean water depths ( W) inhabited by cetacean species and species groups in the Gulf of Mexico c uring 1989-90 (ANOVA: F=29.3, P<0.05; species or species groups sighted more than 20 times; n=number of herds). W Duncan Species or species group n (meters) grouping Pantropical spotted dolphin 23 905 A (Stenella attenuata) Sperm whale 43 877 A (Physeter macrocephalus) SSC dolphins 21 712 B (S. coeruleoalbal longirostrislclymenc ) Pygmy/dwarf sperm whales :si> 544 C (Kogia brevicepslsimus) Risso's dolphin 6] 440 C D (Grampus griseus) Atlantic spotted dolphin 36 368 D E (S. frontalis) Bottlenose dolphin 39 257 E (Tursiops truncatus) * Means with the same letter are not signi ficantly different, P>0.05. in all seasons except summer. The number of species or species groups sighted in summer, fall, winter and spring was 12, 10, 6, and 10, respectively. Seasonal sighting rates of all cetacean herds ranged from 0.91 herds/100 km (winter) to 2.01 herds/100 km (fall) and those of all animals sighted ranged from 16.80 animals/ 100 km (summer) to 52.25 animals/100 km (fall). The relative abundance of several species or spe- cies groups varied seasonally (Table 4 ). Sighting rates (herds and animals) of Risso's dolphins showed a dis- tinct peak during spring. During April alone, 30% of the Risso's dolphin herds and 48% of the animals were sighted. Sighting rates of sperm whales and Atlantic spotted dolphins peaked in the fall. Sight- ing rates of bottlenose dolphins and pygmy and dwarf sperm whales were highest during summer and fall. SSC dolphins exhibited similar herd sighting rates in each season, but the animal sighting rate was much lower during summer. SSC dolphin herds av- eraged only about 20 animals in summer but near 100 or more during other seasons. Cetaceans were sighted throughout the area sur- veyed (Fig. 2). Each species or species group sighted 20 or more times had a wide spatial distribution and was sighted in all three zones (Fig. 2, Table 5). Ten species were sighted in the eastern zone, 13 in the central zone, and nine in the western zone. Sighting rates of all herds sighted were generally similar in the eastern and central zones (1.67 and 1.71 herds/ 100 km, respectively) and a little lower in the west- ern zone ( 1.05 herds/100 km). However, because the mean herd size of all cetaceans sighted from the east- ern zone (35.8 animals) was larger than those of the central (19.2 animals) and western (21.5 animals) zones, the sighting rates of animals were more vari- able: 59.5 animals/100 km (eastern), 32.7 animals/100 km (central), and 22.6 animals/100 km (western). Sighting rates of species or species groups sighted 20 or more times varied by zone (Table 5). Except for Atlantic spotted dolphins and pantropical spotted dol- phins, herd sighting rates were lowest in the west- ern zone. Except for pantropical spotted dolphins, the animal sighting rates were also lowest in the west- ern zone. Sighting rates (herd and animal) of Risso's dolphins and SSC dolphins were highest in the east- ern zone. In the central zone, sighting rates of sperm whales, bottlenose dolphins, Atlantic spotted dolphins, and pygmy and dwarf sperm whales were highest; those of pantropical spotted dolphins were lowest. Risso's dolphin sightings in the eastern part of the study area were generally concentrated near the 200- m isobath. Most of the sperm whale sightings (65%) occurred southeast of the Mississippi River delta along the 1,000-m isobath. Of 39 bottlenose dolphin herds sighted, 19 were sighted in waters less than 100 m deep, at the head of the Mississippi Canyon and on the Upper Mississippi Fan. A concentration of pygmy and dwarf sperm whales occurred along the western Mississippi Canyon. Most of the SSC dolphin sightings occurred on the Upper Mississippi Fan and in DeSoto Canyon. Four of the five pilot whale herds sighted were encountered on the Upper Mississippi Fan on 4 November 1989. The only ba- leen whales sighted, one fin whale, Balaenoptera physalus, in the fall and one Bryde's whale, B. edeni, in the summer, were both sighted in waters about 200 m deep in the DeSoto Canyon (Fig. 2). The relative abundance of many species or species groups was different from the Gulf stranding data (Table 6). Compared with the stranding database, the relative abundances of Risso's dolphins, pantropical spotted dolphins, and Atlantic spotted dolphins were greater in our study. The relative abundances of balaenopterid whales, ziphiids, pilot whales, and pygmy and dwarf sperm whales from the stranding data were larger than those observed during our surveys. Discussion This study was the first to focus on cetaceans in the oceanic Gulf of Mexico. We sighted at least 18 of the 29 cetacean species with one or more historical Mullin et al.: Cetaceans of the north-central Gulf of Mexico 779 V B EDEM A 9 PHYSALUS o T T RUN CAT US + S. COERULEOALBA' S. LONQtROSTRIS/ S CLYMENE D S BREOANENSIS & •■•) -I. CAV1ROSTRIS MtsonoDOH sr - UNIO. 2IPHIID - 0. QRISfUS Figure 2 The locations of sightings of each cetacean species or species group in the Gulf of Mexico for the entire study (July 1989-June 1990). records from the Gulf. The first at-sea identifications of Bryde's whale, pygmy and dwarf sperm whales, spinner dolphins, and Cuvier's beaked whales in the Gulf were recorded during this study. Prior to this study, species with five or fewer herd-sighting records (nonstranding) from the Gulf included pantropical spotted dolphin, clymene dolphin, Risso's dolphin, killer whale, false killer whale, rough-toothed dol- phin, fin whale, pygmy killer whale, melon-headed whale, and Mesoplodon spp. (Schmidly, 1981; Jennings, 1982; Fritts et al., 1983; Rademacher2). Sperm whales were hunted commercially in the Gulf until the early 1900's (Townsend, 1935) but were re- cently thought to be rare (Lowery, 1974). However, our data and the 17 Gulf sperm whale sightings re- ported by Collum and Fritts ( 1985) indicate they may be more abundant than previously thought. Species known from the Gulf that could be distinguished from aircraft, but were not identified during our surveys, included the northern right whale, Eubalaena glacialis, the blue whale, B. musculus, the minke whale, B. acutorostrata, the humpback whale, Meg- aptera novaeangliae, and the Fraser's dolphin, Lag- enodelphis hosei. 780 Fishery Bulletin 92(4), 1994 Table 4 Seasonal sighting rates of cetacean herds (herds/100 km) anc individual animals (animals/100 km) in the Gulf of Mexico during 1989-90. The effort in transect km each season is in parentheses (h= =number of herds sighted; a=number of animals sighted). Summer Fall Winter Spring Total (6,436 km) (5,215 km) (2,095 km) (6,847 km) (20,593 km) Herds Animals Herds Animals Herds Animals Herds Animals Herds Animals Species or species group (hi (a) (hi (a) (hi (a) (h) (a) (h) (a) All cetaceans 1.46 16.80 2.01 52.25 0.91 32.12 1.49 42.87 1.55 36.12 (94) (1,081) (105) (2,749) (19) (673) (102) (2,935) (320) (7,438) Risso's dolphin {Grampus griseus) 0.26 2.63 0.06 1.00 0.10 1.53 0.57 7.73 0.29 3.80 (17) (169) (3) (52) (2l (32) (39) (529) (61) (782) Sperm whale (Physeter macrocephalus I 0.11 0.34 0.46 1.04 0.29 0.33 0.09 0.09 0.21 0.43 (7) (22) (24) (54) (6) (7) (6) (6) (43) (89) Bottlenose dolphin {Tursiops truncatus) 0.28 2.41 0.33 4.93 0.10 2.10 0.03 0. 10 0.19 2.25 (18) (155) (17) (257) (2) (44) (2) (7) (39) (463) Atlantic spotted dolphin iStenella frontalis) 0.12 2.41 0.31 11.03 0.10 2.10 0.13 2.31 0.17 4.53 (8) (155) (16) (575) (2) (44) (9) (158) (35) (932) Pygmy/dwarf sperm whales (Kogia brevieeps/simus ) 0.19 0.37 0.23 0.46 0 0 0.12 0.19 0.16 0.30 (12) (24) (12) (24) (8) (13) (32) (61) SSC dolphins (S. coeruleoalbal 0.12 2.42 0.12 11.43 0.14 25.78 0.10 11.89 0.12 10.23 longirostris/clymene) (8) (156) (6) (596) (3) (540) (7) (814) (24) (2,106) Pantropical spotted dolphin (S. attenuata) 0.08 4.29 0.13 10.26 0 0 0.16 12.27 0.11 8.02 (5i (276) (7) (535) (11) (840) (23) (1,651) Pilot whale {Globicephala sp.) 0.02 0.08 0.08 1.65 ii 0 (i 0 0.02 0.44 )ll (5) (4) 1861 (5) (91) ( 'ontinued >>n next page The ecological implications of Gulf stranding records are not clear since there are only a small number of strandings of most species. It is not known whether the stranded animals strayed into the Gulf from their primary ranges or whether they inhab- ited Gulf waters on a regular basis. The number and broad seasonality of sightings during this study of Risso's dolphins, sperm whales, pygmy and dwarf sperm whales, SSC dolphins, and pantropical spot- ted dolphins indicate that they are probably perma- nent residents of the Gulf. How accurately our results reflected the abundance of each species relative to other species is uncertain. Factors that vary among species, such as surface behavior, herd size, and time spent at or near the surface, can affect sighting rates. In our study, wa- ter depth or area, or both (e.g. Mississippi Canyon), affected the distribution of some species. However, our survey effort was not equal seasonally over wa- ter depths or by area and this probably affected at least some of our relative abundance results (Tables 4-5). Forty-nine percent (19/39) of the bottlenose dolphin herds we sighted were encountered during summer and fall at the head of the Mississippi Can- yon and on the Upper Mississippi Fan at depths less than 100 m. (Survey effort at <100 m made up <5% of the total effort, 8.5% of both the summer and fall effort, and 0% of the winter and spring effort.) With- out these sightings, the seasonal sighting rates of bottlenose dolphins were less variable. Also, 28% (9/ 32) of the pygmy and dwarf sperm whale sightings and 26% (9/35) of the Atlantic spotted dolphin sightings occurred in the Mississippi Canyon (Block A2, surveyed during summer and fall) where only Mullin et al.: Cetaceans of the north-central Gulf of Mexico 781 Table 4 (Continued) S jmmer Fall Winter Spr ing Total (6,436 km) (5,215 km) (2,095 km) (6,847 km) (20,593 km) Herds Animals Herds Animals Herds Animals Herds Animals Herds Animals Species or species group (h) (a) (h) (a) (h) la) (h) (a) (h) (a) Cuvier's beaked whale [Ziphius cavirostris) (l 0 0.02 0.02 0.05 0.05 0.01 0.03 0.01 0.02 (1) (1) (1) (1) (1) (2) (3) (4) Mesoplodon sp. 0.02 (1) 0.02 (1) 0 0 (i 0 0 0 <0.01 (1) <0.01 (1) Unidentified ziphiids (Mesoplodon/Ziphius) 0.02 (1) 0.03 (2) 0.02 (1) 0.02 (1) 0 0 0.01 (1) 0.01 (1) 0.01 (3) 0.02 (4) Pygmy killer/melon-headed whales {Feresa/Peponocephala 1 0.02 (1) 0.39 (25) 0 0 ii II 0 0 <0.01 (1) 0.12 (25) False killer whale (Pseudorca crassidens) 0.02 (1) 0.05 (3) 0 0 0 0 0 0 <0.01 (1) 0.01 (3) Killer whale (Orcinus orca) 0 0 0 0 0 1) 0.01 (1) 0.12 (8) <0.01 (1) 0.04 (8) Rough-toothed dolphin {Steno bredanensis) I) 0 0 0 0 0 0.01 (1) 0.06 (4) <0.01 (1) 0.02 (4) Fin whale (Balaenoptera physalus) 0 0 0.02 (1) 0.02 (1) 0 II (1 0 <0.01 (1) <0.01 ID Bryde's whale [B. edeni) 0.02 (1) 0.02 (1) (i 0 0 II 0 0 <0.01 (1) <0.01 (1) Unidentified large cetacean 0 0 0.02 (1) 0.06 (3) 0.05 (1) 0.05 (1) 0.04 (3) 0.07 (5) 0.02 15) 0.04 (9) Unidentified small cetacean 0.20 1.35 0.23 10.81 0.10 0.19 0.19 8.00 0.19 5.84 (13) (87) (12) (564) (2) (4) (13) (548) (40) (1,203) 10% of the total survey effort occurred. The Missis- sippi Canyon region is probably an important ceta- cean habitat. Eight species or species groups were sighted there and when herd sighting rates were calculated for each survey block (Mullin et al.5), it had the highest sighting rate of any block. The region near the 1,000-m isobath on the Upper Mississippi Fan appeared to be an important habi- tat for sperm whales. Most sperm whale herd sightings (72%, 31/43) occurred on the Upper Mis- sissippi Fan (Blocks Al, B3, and B4) but only 40% of the total effort occurred there. Fall may have been a 5 Mullin, K., W. Hoggard, C. Roden, R. Lohoefener, C. Rogers, and B. Taggart. 1991. Cetaceans on the upper continental slope in the north-central Gulf of Mexico. Outer Continental Shelf Study/ MMS 91-0027. U.S. Dep. Interior, Minerals Mgmt. Service, Gulf of Mexico OCS Regional Office, New Orleans, LA, 108 p. time of increased sperm whale abundance on the Upper Mississippi Fan. Of the total effort, 20% oc- curred in fall on the Upper Mississippi Fan and yielded 47% (20/43) of the total sperm whale herd sightings. Of course the same animals could have been seen repeatedly, but even if that were true, it indicates that a very small area (Fig. 2) could be important to some animals for a period of at least several months. Our study was confined to the outer continental shelf and the upper continental slope and did not cover the entire range of water depths that each spe- cies inhabits. However, our results do not conflict with what is generally known about the water depth distribution of each species (Leatherwood and Reeves, 1983 ). While the supposition that only bottle- nose dolphins and Atlantic spotted dolphins inhabit 782 Fishery Bulletin 92(4), 1994 Table 5 Sighting rates of cetacean herds (herds/100 km) and animals (animals/100 km) by zone (Fig Bl and B2; Central=Al, A2, B3, B4 and B5; Western=A3, B6 and B7) in the Gulf of Mexico in kilometers in each zone is in parentheses (/i=number of herds, a=number of animals). '. 2: Eastern = during 1989- blocks A4, -90. Effort Species or species group Eastern (4,330 km) Central (11,713 km) Western (4,550 km) Herds (A) Animals (a) Herds ih) Animals (a) Herds (/!) Animals (a) All cetaceans 1.67 (72) 59.54 (2,578) 1.71 (200) 32.70 (3,830) 1.05 (48) 22.64 (1,030) Risso's dolphin (Grampus griseus) 0.51 (22) 7.39 (320) 0.25 (29) 3.04 (356) 0.22 (10) 2.33 (106) Sperm whale (Physeter macrocephalus) 0.14 (6) 0.18 (8) 0.30 (35) 0.67 (79) 0.04 (2) 0.04 (2) Bottlenose dolphin {Tursiops truncatus) 0.07 (3) 1.39 (60) 0.29 (34) 2.95 (346) 0.04 (2) 1.25 (57) Atlantic spotted dolphin (Stenella frontalis) 0.09 (4) 4.32 (187) 0.20 (23) 5.20 (609) 0.18 (8) 2.99 (136) Pygmy/dwarf sperm whales (Kogia brevicepslsimus) 0.14 (6) 0.30 (13) 0.19 (22) 0.34 (40) 0.09 (4) 0.18 (8) SSC dolphins (S. coeruleoalbal longirostrislclymene ) 0.18 (8) 22.96 (994) 0.11 (13) 8.81 (1,032) 0.07 (3) 1.76 (80) Pantropical spotted dolphin (S. attenuata) 0.18 (8) 9.63 (417) 0.08 (9) 5.43 (636) 0.13 (6) 13.14 (598) Pilot whale {Globicephala sp.) 0 (i 0.03 (4) 0.73 (86) 0.02 (1) 0.11 (5) Cuvier's beaked whale iZiphius cauirostris) 0.07 (3i 0.09 (4) 0 0 n 0 Continuec on next page most continental shelf waters of the U.S. Gulf (Lo- wery, 1974; Scott et al.1; Rademacher2) may be true, we did sight Risso's dolphins, pygmy and dwarf sperm whales, SSC dolphins, pantropical spotted dolphins, and a sperm whale at depths less than 200 m. Fritts et al. (1983) identified a sperm whale and SSC dol- phins on the continental shelf off southern Florida. In general, our results of mean herd size for most species were similar to those reported from the At- lantic and Pacific oceans (Yamada, 1954; Ross, 1978; Leatherwood et al., 1980; Fritts et al., 1983; Leath- erwood and Reeves, 1983; Vidal et al., 1987; Scott and Cordaro, 1987; Kruse, 1989; Wade and Ger- rodette, 1993; CeTAP6). Our mean herd size for sperm whales (2.1 animals) was slightly smaller than the mean of 3.5 animals reported by Collum and Fritts (1985) from Gulf sightings. However, in other areas of the world, mixed-sex herds and bachelor herds of sperm whales range from 20 to 40 whales (Rice, 1989). The maximum herd sizes reported from the Atlan- tic and Pacific oceans were much greater than those observed in this study, where the largest herd was estimated at 325 animals. Maximum herd sizes of Risso's dolphins, bottlenose dolphins, striped dol- 6 CeTAP. 1982. Acharacterization of marine mammals and turtles in the mid- and north-Atlantic areas of the U.S. outer continen- tal shelf. Final Report of the Cetacean and Turtle Assessment Program, BLM Contract AA551-CT8-48, U.S. Dep. Interior, Washington D.C., 450 p. Mullin et al.: Cetaceans of the north-central Gulf of Mexico 783 Table 5 (Continued) Species or species group Eastern (4,330 km) Central (11,713 km) Western (4,550 km) Herds (h) Animals (a) Herds (h) Animals (a) Herds (h) Animals (a) Mesoplodon sp. 0 0 0 0 0.02 (1) 0.02 (1) Unidentified ziphiids (MesoplodonlZiphius) 0.02 (1) 0.02 (1) 0.02 (2) 0.03 (3) 0 0 Pygmy killer/melon-headed whales (FeresalPeponocephala ) 0 0 <0.01 (1) 0.21 (25) 0 0 False killer whale (Pseudorca crassidens) 1) 0 <0.01 (1) 0.03 (3) 0 0 Killer whale (Orcinus orca) II 0 <0.01 (1) 0.07 (8) 0 0 Rough-toothed dolphin (Steno bredanensis) •I 0 <0.01 (1) 0.03 (4) 0 0 Fin whale (Balaenoptera physalus) 0.02 (1) 0.02 (1) 0 0 0 0 Bryde's whale (B. edeni) 0.02 (1) 0.02 (1) 0 ii 0 0 Unidentified large cetacean 0.05 (2) 0.09 (4) 0.02 (2) 0.02 (2) 0.02 (1) 0.07 (3) Unidentified small cetacean 0.16 (7) 13.12 (568) 0.20 (23) 5.13 (601) 0.22 (10) 0.75 (34) phins, and pilot whales in the Atlantic exceeded 350 animals (CeTAP6). Herds of striped dolphins, pantropical spotted dolphins, and spinner dolphins exceeding 1,000 animals in the Pacific are not un- common (Leatherwood and Reeves, 1983). In Monterey Bay, California, Risso's dolphin herds as large as 500 animals have been reported (Kruse, 1989). These differences in maximum herd sizes may be related to how prey or predators, or both, are dis- tributed in these areas (Norris and Dohl, 1980; Wells et al., 1980). We sighted two mixed species herds (Risso's dol- phins and pilot whales; Atlantic spotted and bottle- nose dolphins). Fritts et al. ( 1983) reported only three mixed species herds in the Gulf: two herds of pilot whales and Stenella sp., and one herd of Risso's dol- phins with an unidentified whale. Risso's dolphins are often associated with other oceanic cetaceans (Leatherwood and Reeves, 1983; Kruse, 1989). In the eastern tropical Pacific, spinner and pantropical spot- ted dolphins are commonly found together (Au and Perryman, 1985). Bottlenose dolphins were associ- ated with other species (primarily pilot whales) in 20% of the sightings in the Pacific (Scott and Chivers, 1990). While we sighted only five herds of pilot whales, other species that are commonly in mixed species herds elsewhere (Risso's dolphins, bottlenose dolphins, and pantropical spotted dolphins), ac- counted for 44% of our identified herd sightings. The abundance and distribution of prey or predators, or both, may be factors involved in the formation of mixed species herds (see Scott and Chivers, 1990). There may be differences in these factors in the Gulf of Mexico compared with those in the northwestern Atlantic and Pacific oceans. The abundance of prey species has been demon- strated to be positively correlated with the abundance of several species of cetaceans (e.g. Kenney and Winn, 1986; Payne et al., 1986; Selzer and Payne, 1988). Fish and squid are the primary prey of most odon- 784 Fishery Bulletin 92(4). 1994 tocetes (e.g. Clarke, 1986; Barros and Odell, 1990; Sekiguchi et al., 1992). However, very little is known about the distribution and abundance of potential prey species in Gulf oceanic waters beyond limited information on cephalopods (e.g. Voss, 1971; Voss and Brakoniecki, 1985) and records of the presence of species (e.g. Hoese et al., 1977). Oceanographic features undoubtedly affect the distribution of prey species and, ultimately, cetacean diversity, abundance, and distribution. The Missis- sippi River and its distributary, the Atchafalaya River, enter the Gulf north of the area we surveyed and account for nearly one-half of the total freshwa- ter flow into the entire Gulf. The Loop Current, the major oceanographic feature in the eastern Gulf, carries 25-30 million m3 of water per second into the Gulf. At times, the Loop Current extends as far north Table 6 Comparison of relative abundances of cetaceans deter- mined by aerial counts during 1989-90 and by historical strandings in the Gulf of Mexico. Species or species group Strandings' This study Balaenopterids 9.3% (48) <0.1%(2) Sperm whale (Physeter macrocephalus ) 4.7% (24) 1 r,', (89) Pygmy/dwarf sperm whales (Kogia brevicepslsimus) 15.3% (79) 1.1% (61) Unidentified ziphiids (Mesoplodon/Ziphius I 10.6% (55) 0.2% (9) Risso's dolphin (Grampus griseus) 2.7% (14) 13.6% (782) Atlantic spotted dolphin (Stenella frontalis) 7.0% (36l 16.2% (9321 Pantropical spotted dolph (S. attenuata) n 1.4'; (7) 28.6% (1,651) SSC dolphins' (S. coeruleoalbal longirostrislelymene ) 22.3% (115) 36.5% (2,106) Pilot whale (Globieephala sp.) 16.5% (85i 1.6% (91) Pygmy killer/ melon-headed whales {FeresalPeponoeephala ) 3.1% (16) 0.4% (25) False killer whale (Pseudorca crassidens) 3.5% (18) -or; (3) Killer whale (0 rein us orca) ui/, '■; i) V, (8) Rough-toothed dolphin (Steno bredanensis) 2.9% (15) <0.1% (4) Fraser's dolphin lLagenodelphis hosei) 0.2% ( 1 ) o%. ' See Footnote 3. as the Upper Mississippi Fan or the DeSoto Canyon. As the Loop Current flows onto the continental slope it causes nutrient-rich upwellings (Jones et al., 1973; Weber et al., 1990). All these features interact with the diverse bottom topography of the north-central Gulf, making it a very dynamic area. In 1990, the Southeast Fisheries Science Center began conducting annual cetacean shipboard surveys of the entire oceanic U.S. Gulf of Mexico. Results to date (775 herd sightings) suggest that the compara- tively small maximum herd sizes and single species herds found in this study in the Gulf are accurate (SEFSC7). Comparisons of the ecology of Gulf ceta- ceans with those from other areas should provide an excellent opportunity to understand the physical and biological factors that affect cetacean diversity, dis- tribution, abundance, herd sizes, and associations. Acknowledgments This cooperative study was supported by the U.S. Minerals Management Service (MMS), Gulf of Mexico Region; the NOAA Aircraft Operations Center (AOC); and the SEFSC. R. Avent of the MMS was instrumental in making this project a success. We thank the AOC staff for all their help and especially the Twin Otter pilots: Comdr. R Wehling, Comdr. D. Eilers, Lt. T. Gates, Lt. T. O'Mara, Lt. M. White, and Lt. BT This study could not have been completed without the contributions of SEFSC personnel J. Benigno, A. Shah, M. McDuff, W. Stuntz, T. Henwood, and L. Hansen. We thank J. Mead for providing the Gulf of Mexico stranding data and T Jefferson for reviewing the manuscript. The data for this paper were collected under interagency agreement number 14- 12-000 1- 30398 between the MMS and the SEFSC. Literature cited Au, D. W., and W. L. Perryman. 1985. Dolphin habitats in the eastern tropical Pacific. Fish. Bull. 83:623-643. Barron, G. L., and T. A. Jefferson. 1993. First records of the melon-headed whale (Pep- onocephala electra) from the Gulf of Mexico. Southwest. Nat. 38:82-85. 7 SEFSC. 1990-93. Reports of NOAA ship Oregon II cruises 187, 194, 199, 203, and 204. U.S. Dep. Commer., NOAA, Nat. Mar. Fish. Serv., Southeast Fish. Sci. Cent. Pascagoula Fa- cility, P.O. Drawer 1207, Pascagoula, MS 39568. Mullin et al.: Cetaceans of the north-central Gulf of Mexico 785 Barros, N. B., and D. K. Odell. 1990. Food habits of bottlenose dolphins in the southeastern United States. In S. Leatherwood and R. R. Reeves (eds.), The bottlenose dolphin, p. 309-328. Academic Press, San Diego, CA. Bonde, R. K., and T. J. O'Shea. 1989. Sowerby's beaked whale (Mesoplodon bidens) in the Gulf of Mexico. J. Mammal. 70:447-449. Buckland, S. T., D. R. Anderson, K. P. Burnham, and J. L. Laake. 1993. Distance sampling: estimating abundance of biological populations. Chapman and Hall, 446 p. Clarke, M. R. 1986. Cephalopods in the diet of odontocetes. In M. M. Bryden and R. Harrison (eds.), Research on dolphins, p. 281-321. Oxford Sci. Pub. Collum, L. A., and T. H. Fritts. 1985. Sperm whales (Physeter catodon ) in the Gulf of Mexico. Southwest. Nat. 30:101-104. Fritts, T. H., A. B. Irvine, R. D. Jennings, L. A. Collum, W. Hoffman, and M. A. McGehee. 1983. Turtles, birds, and mammals in the northern Gulf of Mexico and nearby Atlantic waters. U.S. Fish and Wildl. Serv., Div. Biol. Serv., Washington, D.C. FWS/OBS-82/65, 455 p. Gore, R. H. 1992. The Gulf of Mexico: a treasury of resources in the American Mediterranean. Pineapple Press, Sarasota, FL, 384 p. Hersh, S. L., and D. K. Odell. 1986. Mass stranding of Fraser's dolphin, Lagen- odelphis hosei, in the western North Atlantic. Mar. Mamm. Sci. 2:73-76. Hoese, H. D., R. H. Moore, and V. F. Sonnier. 1977. Fishes of the Gulf of Mexico: Texas, Louisi- ana, and adjacent waters. Texas A&M Univ. Press, College Station, 327 p. Hui, C. A. 1979. Undersea topography and distribution of dol- phins of the genus Delphinus in the Southern Cali- fornia Bight. J. Mammal. 60:521-527. Jennings, R. 1982. Pelagic sightings of Risso's dolphin, Grampus griseus, in the Gulf of Mexico and Atlantic Ocean adjacent to Florida. J. Mammal. 63:522-523. Jones, J. I., R. E. Ring, M. O. Rinkel, and R. E. Smith. 1973. A summary of the knowledge of the eastern Gulf of Mexico. The State Univ. System of Florida, Institute of Oceanography, Gainsville, FL. Kenney, R. D., and H. E. Winn. 1986. Cetacean high-use habitats of the northeast United States continental shelf. Fish. Bull. 84:345-357. Kruse, S.L. 1989. Aspects of the biology, ecology and behavior of Risso's dolphin (Grampus griseus) off the California coast. M.S. thesis, Univ. Calif, Santa Cruz, 120 p. Leatherwood, S., D. 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 CIRC-396, 176 p. Leatherwood, S., W. F. Perrin, V. L. Kirby, C. L. Hubbs, and M. Dahlheim. 1980. Distribution and movements of Risso's dol- phin, Grampus griseus, in the eastern North Pacific. Fish. Bull. 77:951-963. Leatherwood, S., and R. R. Reeves. 1983. The Sierra Club handbook of whales and dol- phins. Sierra Club Books, San Francisco, CA, 302 p. Lowery, G. H., Jr. 1974. The mammals of Louisiana and its adjacent waters. Louisiana State Univ. Press, Baton Rouge, LA, 565 p. Norris, K. S., and T. P. Dohl. 1980. The structure and function of cetacean schools. In L. M. Herman (ed.), Cetacean behav- ior: mechanisms and functions, p. 211-261. Wiley, New York, NY. Payne, P. M., J. R. Nicolas, L. O'Brien, and K. D. Powers. 1986. The distribution of the humpback whale, Megap- tera novaeangliae, in Georges Bank and in the Gulf of Maine in relation to densities of the sand eel, Ammodytes americanus. Fish. Bull. 84:271-277. Perrin, W. F., E. D. Mitchell, J. G. Mead, D. K. Caldwell, and P. J. H. van Bree. 1981. Stenella clymene, a rediscovered tropical dol- phin of the Atlantic. J. Mammal. 62:583-598. Perrin, W. F., E. D. Mitchell, J. G. Mead, D. K. Caldwell, M. C. Caldwell, P. J. H. van Bree, and W. H. Dawbin. 1987. Revision of the spotted dolphins, Stenella spp. Mar. Mamm. Sci. 3:99-170. Rice, D. W. 1989. Sperm whale Physeter macrocephalus Linnaeus, 1758. In S. H. Ridgway and R. Harrison (eds.), Handbook of marine mammals, Vol. 4: River dolphins and the larger toothed whales, p. 177- 234. Academic Press, San Diego, CA. Ross, G. J. B. 1978. Records of pygmy and dwarf sperm whales, genus Kogia, from southern Africa, with biological notes and some comparisons. Ann. Cape Prov. Mus. 11:259-327. Schmidly, D. J. 1981. Marine mammals of the southeastern United States coast and the Gulf of Mexico. U.S. Fish Wildl. Ser., Office of Biological Service, Washing- ton, D.C. FWS/OBS-80/41, 163 p. Scott, M. D., and J. G. Cordaro. 1987. Behavioral observations of the dwarf sperm whale, Kogia simus. Mar. Mamm. Sci. 3:353-354. Scott, M. D., and S. J. Chivers. 1990. Distribution and herd structure of bottlenose dolphins in the eastern tropical Pacific. In S. Leatherwood and R. R. Reeves (eds.), The bottle- nose dolphin, p. 387-402. Academic Press, San Diego, CA. 786 Fishery Bulletin 92(4). 1994 Sekiguchi, K., N. T. W. Klages, and P. B. Best. 1992. Comparative analysis of the diets of smaller ondontocete cetaceans along the coast of southern Africa. S. Afr. J. Mar. Sci. 12:843-861. Selzer, L., and P. M. Payne. 1988. The distribution of white-sided (Lageno- rhynchus acutus) and common dolphins (Delphi- nus delphis) vs. environmental features of the con- tinental shelf of the northeastern United States. Mar. Mamm. Sci. 4:141-153. Townsend, C. H. 1935. The distribution of certain whales as shown by logbook records of American whaleships. Zoologica 19:3-50. United States Fish and Wildlife Service (USFWS). 1989. Endangered and threatened wildlife and plants. 50 CFR 17.11 and 17.12, Division of En- dangered Species and Habitat Conservation. U.S. Fish Wildl. Ser. Washington, D.C. 20240. Vidal, O., L. T. Findley, P. J. Turk, and R. E. Boyer. 1987. Recent records of pygmy sperm whales in the Gulf of California. Mar. Mamm. Sci. 3:354-356. Voss, G. L. 1971. The cephalopod resources of the Caribbean sea and adjacent regions. In Symposium on in- vestigations and resources of the Caribbean Sea and adjacent regions, p. 307-323. FAO, Fish. Rep. 71.2. Voss, G. L., and T. F. Brakoniecki. 1985. Squid resources of the Gulf of Mexico and Southeast Atlantic coasts of the United States. In NAFO Scientific Council Studies Number 9, Spe- cial session on squids, p. 27-37. September 1984, Dartmouth, Nova Scotia, Canada. Wade, P. R., and T. Gerrodette. 1993. Estimates of cetacean abundance and distri- bution in the eastern tropical Pacific. Rep. Int. Whaling Comm. 43:477-493. Weber, M., R. T. Townsend, and R. Bierce. 1990. Environmental quality in the Gulf of Mexico: a citizen's guide. Center for Marine Conservation, Washington, D.C, 130 p. Wells, R. S., A. B. Irvine, and M. D. Scott. 1980. The social ecology of inshore odontocetes. In L. M. Herman (ed.), Cetacean behavior: mecha- nisms and functions, p. 263-317. Wiley, New York, NY. Yamada, M. 1954. Some remarks on the pygmy sperm whale, Kogia. Sci. Rep. Whales Res. Inst. 9:37-61. Abstract. — Reef fish communi- ties at Gray's Reef National Marine Sanctuary, Georgia, differed over five different habitat types. Num- bers of species and overall densi- ties were highest on ledge habitat, intermediate on live-bottom (three categories of low relief [<15 cm] rock outcroppings covered by algae and macrofauna), and lowest over sand. On average, abundance over ledges exceeded that over sand bot- toms by a factor of 50. Generally, community composition at sites over ledges and dense live-bottom areas was similar and distinct from sites found over sparse live-bottom and sand. Many species were found in more than one habitat and few individual species could be consid- ered indicators of a single habitat type. A nondestructive, repeatable procedure of randomly dispersed video transects was devised for as- sessing diurnally active fishes. A video transect method for estimating reef fish abundance, composition, and habitat utilization at Gray's Reef National Marine Sanctuary, Georgia Richard O. Parker Jr. Alexander J. Chester Russell S. Nelson Beaufort Laboratory, Southeast Fisheries Science Center National Marine Fisheries Service, NOAA Beaufort Laboratory, Beaufort, North Carolina 28516-9722 Manuscript accepted 22 February 1994. Fishery Bulletin 92:787-799. The geographic and depth distribu- tion of fishes associated with reefs or hard bottom off the southeastern U.S. coast is generally known (Struhsaker, 1969; Huntsman and Manooch, 1978; Miller and Richards, 1980; Powles and Barans, 1980; Wenner, 1983; Chester et al., 1984; Sedberry and Van Dolah, 1984; Parker and Ross, 1986). Most of these studies have demonstrated changes in commu- nity structure associated with dif- ferent depths and water tempera- tures. Although trawl collections over sand have been compared with collections over hard bottom (Wenner, 1983; Sedberry and Van Dolah, 1984), no quantitative in situ studies of the distribution of reef fishes by type of substrate have been published. Gray's Reef National Marine Sanctuary (GRNMS), Georgia, one of 14 Marine Sanctuaries managed by the National Oceanic and Atmo- spheric Administration (NOAA), encompasses nearly 32 km2 at a depth of about 22 m. Compared with surrounding areas, Gray's Reef con- tains extensive, but patchy and dis- continuous, hardbottom of moder- ate relief (up to 2 m). Rock outcrops or "ledges" have formed in a north- west to southeast direction (Fig. 1 ). Ledges are often separated by wide expanses of sand and are subject to weathering, shifting sand, and slumping, which create a complex habitat with caves, burrows, troughs, and overhangs (Hunt, 1974). Sandy areas between the ledges consist of coarse shell with varying amounts of "rock-like" litter (Henry and Van Sant1). Reef fish assemblages are diffi- cult to sample because of the diver- sity and mobility of the fauna and because of the variety of microhabi- tats within complex reef substrates (Russell et al., 1978). The applica- bility and limitations of various techniques for estimating reef fish abundance have been reviewed (Russell et al., 1978; Sale, 1980; Sale and Douglas, 1981; Brock, 1982; DeMartini and Roberts, 1982; Sale and Sharp, 1983; Kimmel, 1985; Sanderson and Solonsky, 1986; Bortone and Kimmel, 1991). Henry, V. J., Jr., and S. B. Van Sant. 1982. Results of reconnaissance mapping of the Gray's Reef National Marine Sanctuary, a report prepared for the Georgia Depart- ment of Natural Resources, Coastal Re- sources Division, Brunswick, GA, under co- operative agreement with the Sanctuary Programs Division of the National Oceanic and Atmospheric Administration (No. NA81AA44-C2098, Amendment 1), 21 p. 787 788 Fishery Bulletin 92(4), 1994 56' 80° 55' 54' 53' —I — 52' 51' I ABUNDANT GROWTH ASSOCIATED WITH EXPOSED ROCK (DENSE LIVE BOTTOM AND LEDGE HABITAT) □ MODERATE GROWTH ASSOCIATED WITH ROCK THINLY COVERED BY SAND SPARSE AND PATCHY GROWTH ASSOCIATED WITH ROCK COVERED BY UP TO 30 CM OF SANO BARE SAND, ROCK ABSENT 31* 25- FISH HAVEN BUOY 24' KILOMETERS 23' 22' Figure 1 Sand, live-bottom, and ledge substrates at Gray's Reef National Ma- rine Sanctuary (GRNMSl (after Hunt, 1974). Techniques include the use of traditional fisheries assessment gear (nets, traps, and hook-and-line), poisons, explosives, and visual observations. The need for repeatable surveys and the constraints of working in a national marine sanctuary necessitated the use of nondestructive survey techniques. Diver observations are the most common method used in studies of reef fish assemblages (Brock, 1954; Bardach, 1959; Hobson, 1972; Chave and Eckert, 1974; Sale, 1975; Jones and Chase, 1975; Jones and Thompson, 1978; Anderson et al., 1981; Ogden and Ebersole, 1981; Sale and Douglas, 1981; Brock, 1982; Kimmel, 1985; Bohnsack and Bannerot, 1986; Parker, 1990). Although a variety of sampling tech- niques have been employed to make quantitative assessments of reef fish abundance, all rely on divers to identify and record fish species observed in a pre- defined area (transect and point counts) or over a period of time (rapid visual assessment techniques). Accuracy of the visual techniques is affected by light levels, water clarity, currents, fish species diversity and densities, substrate complexity, diver familiar- ity with the fishes, and number and size of the sam- pling units. Biases are induced by 1) a tendency to undersample small, cryptic, and nocturnal species (Brock. 1982); 2) identification, counting, and record- ing errors (Brock, 1954; Russell et al., 1978); 3) at- traction and aversion reactions of some species to the divers (Chapman et al., 1974); and 4) species dif- ferences in territory, home range, life history pat- terns, and behavior (Russell et al., 1978). Remote observation techniques, using movie or closed-circuit television cameras deployed from ves- sels or carried by divers, have been used to estimate abundance of reef fish (Smith and Tyler, 1973; Alevizon and Brooks, 1975; Powles and Barans, 1980; Boland et al.2). A major advantage is that a perma- nent record of observed fishes is obtained without destroying the fauna. However, remote systems that are tethered to a surface vessel are severely limited by sea conditions. In addition, camera resolution, light levels, water clarity, depth, and lack of in situ guidance limit the effectiveness of remote observa- tions, and biases are imposed by the attraction or avoidance of some species to the gear. The objectives of this study were 1 ) to develop a nondestructive, repeatable procedure for assessing diurnally active fishes inhabiting Gray's Reef Na- tional Marine Sanctuary, and 2 1 to describe and com- pare fish communities associated with ledge, live- bottom, and sand habitats. - Boland, G., B. Galloway, J. Baker, and G. S. Lcwbel. 1984. Eco- logical effects of energy development on reef fish of the Flower Garden Banks. Final Rep. Contract No. Na80-GA-C-00057. U.S. Natl. Mar. Fish. Serv.. Galveston, TX, 466 p. Parker et al.: Reef fish abundance, composition, and habitat use 789 Methods Research site selection Based on preliminary observations (1-2 May 1985), 30,000 m2 of bottom in GRNMS were divided em- pirically into sand, live-bottom, and ledge habitats (Parker et al.3). For detailed community analyses, the live-bottom area was further divided into three subunits. The habitat classifications and approxi- mate proportional areas within GRNMS (calculated from Hunt, 1974) were the following: Sand: sand or sand and shell bottom with bot- tom relief (<20 cm) provided by sandy swales; occasional (<1%) depressions or burrows (5-25 cm deep) in sand surrounded by algae, macrofauna, and sometimes rocks; approxi- mately 18% of GRNMS. Live-bottom: approximately 1 to 75% of bottom composed of rock outcroppings covered by algal and benthic macrofauna; little or no (<15 cm) relief; sandy areas, 2 to 25 cm deep.underlaid by rock; approximately 58% of GRNMS. a Sparse live-bottom: covers 1 to 25% of the substrate. b Moderate live-bottom: covers 26 to 50% of the substrate. c Dense live-bottom: covers greater than 50% of the substrate. Ledge: distinct rock outcroppings of 15 cm to over 200 cm; associated rock bottoms covered by algal and benthic macrofauna; approximately 24% of GRNMS. Thirty-six potential sampling sites, 12 each over sand, live-bottom, and ledge substrates, were randomly selected from a pool of 76 lo- cations of known habitat type defined by our preliminary work and by existing Georgia Department of Natural Resources data (Nicholson4; Hudson5). Of the 36 sites, 14, 17, 9, and 12, respectively, were randomly se- lected for sampling during four surveys: sum- mer ( 15-22 August) and fall (13-18 Novem- Parker, R. 0., Jr.. R. S. Nelson, and A. J. Chester. 1988. A quantitative approach to the estimation of reef fish abundance and community composition in the Gray's Reef National Marine Sanctuary using SCUBA divers. Final Rep. Contract No. NA84DOC-C2004. U.S. Natl. Ocean Serv., Washington, D.C., 71 p. Nicholson, N. 1982. The Gray's Reef National Marine Sanctuary Visual Reef Fish Censusing Workshop, fi- nal report. Georgia Dep. Nat. Resour., Coastal Resour. Div., Brunswick, 16 p. ' Hudson, J. A. 1984. Summer Internship Report, Valdosta State College, Georgia Dep. Nat. Resour., Coastal Resour. Div., Brunswick, 33 p. ber) 1985, and spring ( 13-18 May) and summer (21- 27 August) 1986. Sampling was stratified by habitat type. As a stratified random design, optimal allocation of effort among habitats normally would be deter- mined by the variance of the population size of a particular species and the size of each habitat. For this study we used species richness as a proxy for variance because the focus was on a multispecies assemblage. Although the relationship between spe- cies richness and total variance is not clearly defined, even approximate estimators of variance usually are adequate for allocating sampling effort (Steel and Torrie, 1960). Because the sample size of our preliminary work on live-bottom habitat was roughly one-third the number of intervals available for sand and ledge habitats (Fig. 2), we extrapolated the live-bottom data (10 species observed in eight sampling intervals) to a hypothetical sample of 25 sampling intervals. On both ledge and sand habitats approximately 71% of the total species observed were encountered after eight sampling intervals ( Fig. 2 ). Assuming the same relationship for live-bottom data, 14 species would have been encountered in 25 sampling intervals. Prior experience by both diving investigators sug- gests this is a reasonable approximation. Sampling effort was allocated among the different reef habi- tats in proportion to the product of the area of a given 35 LEDGE . _ LIVE BOTTOM a a SANO _.-' ^y D 30 LU > S' It CO S a S ° 25 y CO Ul O / a / to O 20 / a / CD £ / IVE N tn - / < / 3 3 3 io - / „- z y < UJ 5 - / a—0, | 1 1 1 1 L 5 10 15 20 25 30 NUMBER OF 5 MINUTE SAMPLING INTERVALS Figure 2 Species vs. time curve from preliminary work, May 1985. 790 Fishery Bulletin 92(4). 1994 habitat times the square root of anticipated species richness for that habitat (adapted from Steel and Torrie, 1960). The preliminary data (Fig. 2) indicate that 19, 5- minute sampling intervals (95 minutes) were suffi- cient to observe at least 95% of the total species re- corded in each habitat during the preliminary work. A maximum of 40 dives, limited to 20 minutes of survey time per dive (four 5-minute intervals), were available for each of the four surveys. Thus, the avail- able sampling time likely was sufficient to obtain a complete record of the noncryptic species present in each habitat. Sampling procedures All dives were completed between 0930 and 1630 hours to take advantage of maximum light levels. At each site a marker buoy was deployed at the start of a transect. One of three dive teams, each consisting of two divers, swam a 15-minute transect with the prevailing current. One diver operated a color video camera with a 51-mm lens; the other towed a sur- face buoy. The video operator waved his hand in front of the camera to signify the beginning and end of each transect. Appearance of the towed buoy released by divers signaled to the boat the beginning of a transect. The camera was held in a rigid forward position about 1 m off the bottom. Fishes in cryptic locations were not recorded by the video camera. At the termination of a transect the camera was turned off, the towed surface buoy was anchored, a stan- dard black and white Secchi disk (30 cm in diam- eter) was used to measure horizontal visibility, and bottom water temperature was recorded. During each transect swim the vessel approached the towed buoy at 5-minute intervals, and the crew recorded the LORAN C coordinates and plotted its position. The plots were used to measure the length of each transect and to calculate distance covered during each 5-minute interval. Transect length and Secchi disk visibility could be used to estimate area sampled. However, after reviewing the initial tapes, we estimated the transect width to be 4 m (2 m on each side of focal center), because small fishes (70 to 150 mm) could be identified with certainty only out to an estimated distance of 2 m. Larger fish were recorded as they came into view. To avoid duplicate counts, only maximum numbers of species that passed by the camera more than once were used. These species were easily identified by the camera operator. Because transect width remained constant, data are reported as number of fish per meter of transect. Generally, two transects were swum at each site, beginning at the same location and heading with the prevailing current. Plots of the transects showed little overlap. Videotaped transects were viewed to estimate abundance of each species seen within each 5-minute interval. Videotapes were projected on a 50-cm color NEC Corporation Television and were analyzed by a single observer. Viewing was in real time with fre- quent pauses, reverses, and repeated counts until the observer obtained the same count of species three times. Date, location, Secchi disk measurement, bot- tom water temperature, and number of each species per type of habitat were recorded on data sheets. Because habitat type often changed during a transect, habitat changes were closely monitored and species were apportioned appropriately. Community analysis Species-specific data were summarized by habitat type. Statistical analysis, data summarization, and graphic representation were accomplished with SAS version 6.03 software system (SAS, 1987). Data were summarized over sites within habitats and surveys, and the effects of survey (4 surveys) and habitat (3 habitat types) on total fish density and number of species observed were tested with two-way ANOVA's. Cluster analysis was used to classify Gray's Reef sampling sites according to the species composition of the fish community. Species that were not found in at least 10% of ledge, live-bottom, or sand sites in any one survey were eliminated. For each survey, relative abundance data (number/m of transect) were arranged in a species-by-site matrix, standardized by dividing each element by the square root of the product of the row total and column total (simulta- neous double standardization), and converted to a site-by-site Canberra Metric dissimilarity matrix (Clifford and Stephenson, 1975). Sites were grouped by means of the "flexible sorting" algorithm of Lance and Williams (1967) and the cluster intensity coeffi- cient was set at -0.25 to approximate the median clustering strategy. Analysis was conducted with SIMCLUST statistical software (Wolfe and Chester, 1991). Results A total of 110 transects covering a distance of 24 km (4.9 km over ledge, 12.7 km over live-bottom, and 6.4 km over sand) were made during the study. Over 92,000 fish, including 66 species and 36 families, were recorded and identified from the videotapes (Table 1). Number of species and density offish ( individuals/m transect) varied significantly among the four surveys Parker et al.: Reef fish abundance, composition, and habitat use 791 Table 1 Species observed at Gray's Reef National Marine Sanctuary (GRNMS) between 12 August 1985 and 27 August 1986 in ledge (L), live-bottom (LB), sand (S), or pelagic (P) habitats. Species indicated by asterisk in column labeled 'Both sites' represent those seen in study site off North Carolina by Parker (1990) and those seen at GRNMS, whereas those species indicated in GRNMS column represent those seen only at GRNMS. Species Habitat LB Both sites GRNMS Orectolobidae Ginglymostoma cirratum, nurse shark Dasyatidae Dasyatis americana , southern stingray Muraenidae Moray, unidentified Ophichthidae Myrophis punctatus, speckled worm eel Clupeidae Brevoortia tyrannus, Atlantic menhaden Sardinella aurita, Spanish sardine Synodontidae Synodus foetens, inshore lizardfish Trachinocephalus myops, snakefish Batrachoididae Opsanus sp., toadfish' Holocentridae Holocentrus ascencionis, squirrelfish Syngnathidae Hippocampus erectus, lined seahorse Mierognathus crinitus, banded pipefish Syngnathus louisianae, chain pipefish Serranidae Centropristis ocyurus, bank sea bass C. philadelphica, rock sea bass C. striata, black sea bass Diplectrum formosum, sand perch Mycteroperca microlepis, gag M. phenax, scamp Serranus subligarius, belted sandfish Grammistidae Rypticus maculatus, whitespotted soapfish Priacanthidae Priacanthus arenatus, bigeye Pristigenys alt a, short bigeye Apogonidae Apogon pseudomaeulatus, twospot cardinalfish Phaeoptyx pigmentaria, dusky cardinalfish Continued on next page 792 Fishery Bulletin 92(4), 1994 Table 1 (Continued) Species Habitat Both sites GRNMS L LB S P Carangidae Caranx bartholomaei, yellow jack * < * * C. ruber, bar jack * ; * * Caranx sp., unidentified jack * * Decapterus punctatus, round scad * * * * * Seriola dumerili, greater amberjack * * * * S. rivoliana, almaco jack i * * Lutjanidae Lutjanus campechanus, red snapper * * * Lutjanus sp., juvenile snapper * Haemulidae Haemulon aurolineatum, tomtate * * * Orthopristis chrysoptera, pigfish * * Sparidae Archosargus probatocephalus, sheepshead i * * Calamus leucosteus, whitebone porgy : * * Diplodus holbrooki, spottail pinfish * * * Pagrus pagrus, red porgy * * * Stenotomus eaprinus, longspine porgy 1 * * * Sciaenidae Equetus acuminatus, high-hat * * * E. lanceolatus, jacknife-fish 1 * * E. umbrosus, cubbyu ■t ■ * Mullidae Mullus auratus, red goatfish * * Ephippidae Chaetodipterus faber, Atlantic spadefish ! * • * Chaetodontidae Chaetodon ocellatus, spotfin butterflyfish ! * C. sedentarius, reef butterflyfish ! ! * C. striatus, banded butterflyfish » * * Pomacanthidae Holocanthus bermudensis, blue angelfish t * Pomacentridae Pomacentrus partitus, bicolor damselfish • 1 * P. variabilis, cocoa damselfish ! * * Sphyraenidae Sphyraena barracuda, great barracuda * * • * Labridae Halichoeres bivittatus, slippery dick * * * * Hemipteronotus novacula, pearly razorfish 1 * * * Tauloga omtis. tautog * * * Continued on next page Parker et al.: Reef fish abundance, composition, and habitat use 793 Table 1 (Continued) Habitat Species LB Both sites GRNMS Scaridae Sparisoma sp., parrotfish Opistognathidae Unidentified jawfish Blenniidae Ophioblennius atlanticus, redlip blenny Parablennius marmoreus, seaweed blenny Unidentified Gobiidae Ioglossus calliurus, blue goby Microgobius carri, Seminole goby Acanthuridae Acanthurus bahianus, ocean surgeon A. chirurgus, doctorfish Scombridae Euthynnus alleteratus, little tunny Scomberomorus maculatus, Spanish mackerel Stromateidae Psenes maculatus, silver driftfish Triglidae Prionotus sp., unidentified searobin Bothidae Unidentified flounder Balistidae Aluterus heudoloti, dotterel filefish A. schoepfi, orange filefish Balistes capriscus, gray triggerfish Monocanthus hispidus, planehead filefish Ostraciidae Laetophrys quadricornis , scrawled cowfish L. triqueter, smooth trunkfish Diodontidae Diodon hystrix, porcupinefish Others fish larval fish juvenile fish Number of taxa 63 62 SI 15 42 UN Opsanus sp. is likely an undescribed offshore form. and three major habitat types (two-way ANOVA, P<0.05, no significant survey x habitat interaction). Numbers of species and overall densities were great- est on ledge habitats, intermediate on live-bottom, and smallest over sand (Figs. 3 and 4, Table 2). Num- bers of species and densities were highest during the summer of 1985, intermediate during the fall of 1985 and summer of 1986, and lowest during the spring of 1986 (Table 2). The lower number of species ob- served in spring of 1986 may be a result of fewer samples having been taken because of inclement weather. Underwater visibility varied from 2.4 to 17.9 794 Fishery Bulletin 92(4). 1994 m but did not affect identification and counts, since it exceeded 2 m (see Methods section). Species composition differed over five different habitat types (Table 3). Nearly three times as many species were identified from ledge habitats (63) than tS LEDGE (S3 LIVE-BOTTOM □ SAND SUMMER 1963 FALL 1983 SPRING 1986 SUMMER 1986 Figure 3 Number of species observed at Gray's Reef National Ma- rine Sanctuary by survey and habitat. S3 LEDGE ESI LM-BOTTOM □ SAND 1 SUMMER 1985 PAI_L 1985 SPRING 1996 SUMMER 1986 Figure 4 Density offish (estimated means) at Gray's Reef National Marine Sanctuary by survey and habitat. Number offish/ m over sand for fall 1985, spring 1986, and summer 1986 was 0.02. from sand habitats (22), and over one-third as many species were seen on ledge as were seen on either dense (46) or moderate (46) live-bottom. More spe- cies were recorded over dense and moderate live-bot- tom than over sparse live-bottom (33). Mean rela- tive abundances also were related to habitat; high- est values were found for ledge habitat, progres- sively declining values for the three live-bottom habitats, and lowest values for sand. On average, abundances over ledges exceeded those over sand bottoms by a factor of 50. Cluster analyses for each of the four surveys (Fig. 5) indicate clear distinctions in community composition among habitats. Generally, sites over ledges and dense live-bottom areas were classi- fied similarly and were distinct from sites found over sparse live-bottom and sand. Classification of moderate live-bottom sites was more variable. Many species were present in more than one habi- tat (Tables 1 and 3), and few individual species could be considered indicators of a single habitat type. The following species were present at more than half the respective habitat sites. Ledges were characterized by black sea bass,6 belted sandfish, gag, scamp, sand perch, round scad, tomtate, sheepshead, spottail pinfish, longspine porgy, cubbyu, Atlantic spadefish, slippery dick, doctorfish, and planehead filefish. Dense live-bot- tom was characterized by black sea bass, belted sandfish, tomtate, longspine porgy, and slippery dick. Moderate live-bottom had black sea bass, belted sandfish, round scad, longspine porgy, and slippery dick. Sparse live-bottom had black sea bass, round scad, longspine porgy, and slippery dick. Sand habitats were relatively depauperate but were best characterized by the presence of pearly razorfish. Discussion Comparison of the fauna of GRNMS with that of other reefs off the southeastern U.S. coast sug- gests a high level of variability among reef com- munities. The fish species composition at GRNMS differs considerably from that of an intensely stud- ied reef in 30 m of water, 44 km south of Beaufort Inlet, North Carolina (Parker, 1990). Of 113 spe- cies observed by divers at the two reefs, only 42 (37%) were common to both (Table 1). Twenty-eight species were unique to GRNMS (Table 1) and 43 species were seen only at the North Carolina site. Although more effort was expended at GRNMS Scientific names of fishes in this study are listed in Tahle 1. Parker et al.: Reef fish abundance, composition, and habitat use 795 (97 transects over 21 hours vs. 51 point counts over 17 hours), 15 more species were observed off North Carolina. The major difference appears to be that more temperate species usually associ- ated with inshore environments (i.e. inshore lizardfish, toadfish, rock sea bass, pigfish, pearly razorfish, and Spanish mackerel) were present at GRNMS, whereas more tropical species (i.e. red grouper, harlequin bass, wrasse bass, white grunt, knobbed porgy, and queen angelfish) were seen at the North Carolina location. The warm waters of the Gulf Stream provide a mechanism for recruitment and survival of many tropical species (Briggs, 1974). GRNMS is 12 km closer to shore and 8 m shal- lower than the North Carolina site. More importantly, although the position of the Gulf Stream varies across the con- tinental shelf, it generally follows the 200-m isobath which is much farther offshore from GRNMS ( 105 km) than from the North Caro- lina site (40 km). The diversity of species collected is partly a reflection of the sampling method. Our ob- servations on species abundance agree only partially with results obtained by trawling. The 10 most abundant and common species observed in this study (Table 3 ) included four (tomtate, black sea bass, cubbyu, and longspine porgy) of the most abundant spe- cies caught by trawling over hardbottom similar to GRNMS off the southeastern U.S. coast (Wenner, 1983; Sedberry and Van Dolah, 1984; Table 4). Size, form, and behav- ior of three of the other six species may pre- clude their capture by trawls. Two of the three most abundant species (round scad and slippery dick) are small and fusiform and can pass through the meshes of most trawls. Slip- pery dick and belted sandfish usually live close to the bottom where they are protected from trawls by the substrate. Round scad have been seen swimming freely in and out of the mouth of trawls towed up to 3.5 knots (Workman7). A major source of unmeasured error in many visual assessments is observer error in sighting, identifying, counting, and recording. In a prior study of ledge fishes at GRNMS, 10 divers operating in pairs per- Table 2 Mean number of species and density (number/m), standard errors, and nu mber of site-habitat combinations (n ) for each cruise at Gray's Reef National Marine Sanctuary, Georgia. Cruise Habitat a Species Density Mean SE Mean SE 1 Ledge 6 25.83 2.40 19.90 4.09 Live-bottom 12 13.42 1.47 5.77 1.51 Sand 6 5.17 1.01 0.85 0.49 2 Ledge 7 21.86 2.54 18.87 7.54 Live-bottom 13 11.46 1.37 2.00 0.77 Sand 7 2.71 0.57 0.02 0.01 3 Ledge 3 14.67 0.33 4.35 2.12 Live-bottom 6 6.33 0.95 1.26 1.09 Sand 5 3.20 1.20 0.02 0.01 4 Ledge 6 18.17 1.66 8.82 2.82 Live-bottom 11 9.09 0.94 2.57 0.96 Sand 3 2.67 0.33 0.02 0.01 Workman, I. NOAA, NMFS, Mississippi Laboratory, Pascagoula, MS 39567. Personal commun., January 1994. 0.0 2 3 SUMMER 1985 am m r^TH nl rl M lH 354S43433444S54551 1 1223222231 1 1 H DC < AUTUMN 1985 ,4-iChTi rT ^rtlnrt i 1 -- . rrh-i 55524334 1 -LEDGE 2 - DENSE LIVE-BOTTOM 3 - MODERATE LIVE-BOTTOM 4 - SPARSE LIVE-BOTTOM 5 - SAND Figure 5 Dendrograms of cluster analyses of sites referenced by habitat type for each of four surveys conducted in Gray's Reef National Marine Sanctuary. Note that dissimilarity axes differ in scale. 796 Fishery Bulletin 92|4). 1994 Table 3 Numbers of fish per meter transect and number of sites present in parentheses ) by habitat at Gray's Reef Na- tional Marine Sanctuary, August 1985-August 1986. Dense Moderate Sparse Ledge live-bottom live-bottom live-bottom Sand Species (n=22) (n- =22) (n = =22) (« = 23) . In choosing between transect and point sampling, we considered the particular conditions at GRNMS. When properly applied, the precision of both proce- dures can be high (Keast and Harker, 1976; Sale and Douglas, 1981; DeMartini and Roberts, 1982; Bohnsack and Bannerot, 1986; Witzig, 1988). Lim- ited visibility at GRNMS was thought to bias point counts for some species. Bohnsack and Bannerot ( 1986) found that point samples with a radius of 2 m or less underestimated abundances of 11 of 15 spe- cies observed. In contrast, Parker ( 1990) found that during low visibility some species of reef fish (e.g. gag, black sea bass, and white grunt) concentrate under and around ledges. Extrapolating density of these fish in a small visible area to the total popula- tion over an entire reef that consists mostly of low profile (<1 m) rock outcroppings sparsely inhabited by fishes grossly overestimates their abundance. Off- bottom tidal currents, frequently in excess of 20 cm/s at GRNMS, make it impossible for the vessel to re- main stationary for the 5 to 10 minutes necessary to conduct enumerations. For these reasons we devel- oped a random transect technique that allowed us to swim with the prevailing current, covering 86 to 544 m per transect. Because visibilities at GRNMS can be consistently less than 5 m, this technique allowed us to sample large areas with minimum underwater time. The technique is a consistent, repeatable pro- cedure for assessing the noncryptic, diurnally active fishes at GRNMS. Acknowledgments This work is the result of research sponsored by the U.S. Department of Commerce, National Oceanic and Atmospheric Administration, National Ocean Ser- vice, Office of Ocean and Coastal Resource Manage- ment, Sanctuary Programs Division, under contract NA84DOC-C2004. We sincerely thank the Georgia Department of Natural Resources, Coastal Resources Division, and in particular the captain of the RV George T. Bagby, the late David Ansley, and the crew, Henry Ansley, Byron Kroscavage, and Tom and Matt Dougherty, for the vessel, for diver support, and for suggestions on survey methods. Special thanks is extended to Dave Colby of this laboratory and two anonymous reviewers for their suggestions and re- views of the final drafts of the manuscript. Finally, we thank Beverly W Harvey for typing and quick turnaround of the many drafts of this paper. Literature cited Alevizon, W. S., and M. G. Brooks. 1975. The comparative structure of two western Atlantic reef fish assemblages. Bull. Mar. Sci. 25:482-490. Anderson, G. R V., A H. Ehrlich, P. R. Ehrlich, J. D. Roughgarden, B. C. Russell, and F. H. Talbot. 1981. The community structure of coral reef fishes. Am. Nat. 117:476-495. Bardach, J. E. 1959. The summer standing crop offish on a shal- low Bermuda reef. Limnol. Oceanogr. 4:77-85. 798 Fishery Bulletin 92|4). 1994 Bohnsack, J. A., and S. P. Bannerot. 1986. A stationary visual census technique for quan- titatively assessing community structure of coral reef fishes. Dep. Commer., NOAA Tech. Rep. NMFS 41, 15 p. Bortone, S. A., and J. J. Kimmel. 1991. Environmental assessment and monitoring of artificial habitats. In W. Seaman Jr. and L. M. Sprague (eds.), Artificial habitat for marine and fresh- water fisheries, p. 177-236. Academic Press, New York. Briggs, John C. 1974. Marine zoogeography. McGraw-Hill, New York, 475 p. 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. F. 1954. A preliminary report on a method of estimat- ing reef fish populations. J. Wildl. Manage. 18:299-308. Chapman, C, A. Johnstone, J. Dunn, and D. Creasey. 1974. Reactions offish to sound generated by diver's open-circuit underwater breathing apparatus. Mar. Biol. 27:357-366. Chave, E. H., and D. B. Eckert. 1974. Ecological aspects of the distribution of fishes at Fanning Island. Pac. Sci. 28:297-317. Chester, A. J., G. R. Huntsman, P. A. Tester, and C. S. Manooch HI. 1984. South Atlantic Bight reef fish communities as presented in hook-and-line catches. Bull. Mar. Sci. 34:267-279. Clifford, H. T., and W. Stephenson. 1975. An introduction to numerical classifica- tion. Academic Press, New York. 229 p. DeMartini, E. E., and D. Roberts. 1982. An empirical test of biases in the rapid vi- sual technique for species-time censuses of reef fish assemblages. Mar. Biol. 70:129-134. Ebeling, A. W., R. J. Larson, W. S. Alevizon, and R. N. Bray. 1980. Annual variability of reef-fish assemblages in kelp forests off Santa Barbara, California. Fish. Bull. 78:361-377. Hobson, E. S. 1972. Activity of Hawaiian reef fishes during the evening and morning transitions between daylight and darkness. Fish. Bull. 70:715-740. Hunt, J. L., Jr. 1974. The geology and origin of Gray's Reef, Geor- gia continental shelf. M.S. thesis, Univ. Georgia, Athens, 83 p. Huntsman, G. R., and C. S. Manooch III. 1978. Coastal pelagic and reef fishes in the South Atlantic Bight. In H. Clepper (ed.), Marine recre- ational fisheries, Vol. 3, p. 97-106. Sport Fish. Inst., Washington, D.C. Jones, R. S., and J. A. Chase. 1975. Community structure and distribution of fishes in an enclosed high island lagoon in Guam. Micronesia 11:127-148. Jones, R. S., and M. J. Thompson. 1978. Comparison of Florida reef fish assemblages using a rapid visual survey technique. Bull. Mar. Sci. 28:159-172. Keast, A., and J. Harker. 1976. Strip counts as a means of determining den- sities and habitat utilization patterns in lake fishes. Environ. Biol. Fishes 1:181-189. Kimmel, J. 1985. A new species-time method for visual assess- ment of fishes and its comparison with established methods. Environ. Biol. Fishes 12:23-32. Lance, G. N., and W. T. Williams. 1967. A general theory of classificatory sorting strat- egies. I: Hierarchical systems. Comput. J. 9:373- 380. Miller, G. C, and W. J. Richards. 1980. Reef fish habitat, faunal assemblages, and factors determining distributions in the South At- lantic Bight. Proc. Gulf Caribb. Fish. Inst. 32: 1 14- 130. Ogden, J. C, and J. P. Ebersole. 1981. Scale and community structure of coral reef fishes: a long term study of a large tropical artifi- cial reef. Mar. Ecol. Prog. Ser. 4:97-103. Parker, R. O., Jr. 1990. Tagging studies and diver observations offish populations on live-bottom reefs of the U.S. south- eastern coast. Bull. Mar. Sci. 46:749-760. Parker, R. O., Jr., and S. W. Ross. 1986. Observing reef fishes from submersibles off North Carolina. Northeast Gulf Sci. 8:31-49. Powles, H., and C. Barans. 1980. Groundfish monitoring in sponge-coral areas off the southeastern United States. Mar. Fish. Rev. 42(51:21-35. Russell, B. C, F. H. Talbot, G. R. V. Anderson, and B. Goldman. 1978. Collection and sampling of reef fishes. //iD. R. Stoddart and R. E. Johannes (eds.), Coral reefs: research methods, p. 329-345. UNESCO, Paris. Sale, P. F. 1975. Patterns of use in a guild of territorial reef fishes. Mar. Biol. 29:89-97. 1980. The ecology of fishes on coral reefs. Oceanogr. Mar. Biol. Annu. Rev. 18:367-421. Sale, P. F., and W. A. Douglas. 1981. Precision and accuracy of visual census tech- nique for fishes assemblages on coral patch reefs. Environ. Biol. Fishes 6:333-339. Sale, P. F., and B. J. Sharp. 1983. Correction for bias in visual transect censuses of coral reef fishes. Coral Reefs 2:37-42. Sanderson, S. L., and A. C. Solonsky. 1986. Comparison of a rapid visual and a strip transect technique for censusing reef fish assem- blages. Bull. Mar. Sci. 39:119-129. Parker et al.: Reef fish abundance, composition, and habitat use 799 SAS (SAS Institute, Inc.). 1987. Changes and enhancements to base SAS. Soft- ware for personal computers, release 6.03. SAS Tech. Rep. Cary, NC, 171 p. Sedberry, G. R., and R. F. Van Dolah. 1984. Demersal fish assemblages associated with hard bottom habitat in the South Atlantic Bight of the U.S.A. Environ. Biol. Fishes 11:241-258. Smith, C. L., and J. C. Tyler. 1973. Population ecology of a Bahamian supra- benthic shore fish assemblage. Am. Mus. Novit. 2528, 38 p. Steel, R. G. D., and J. H. Torrie. 1960. Principles and procedures of statistics. Mc- Graw-Hill, New York, 581 p. Struhsaker, P. 1969. Demersal fish resources: composition, distri- bution, and commercial potential of the continen- tal shelf stocks off southeastern United States. Fish. Ind. Res. 4:261-300. Wenner, C. A. 1983. Species associations and day-night variabil- ity of trawl-caught fishes from the inshore sponge- coral habitat, South Atlantic Bight. Fish. Bull. 81:537-552. Witzig, J. F. 1988. Visual assessment of reeffish communi- ties. Ph.D. thesis, NC State Univ., Raleigh, 161 p. Wolfe, N. A., and A. J. Chester. 1991. SUMCLUST: a cluster analysis program for ecological data. Am. Stat. 45:158. Abstract. —Approximately 12,396 Pacific cod, Gadus maerocephalus, were tagged and released from fish- ery research vessels in the eastern Bering Sea and adjacent waters between 1982 and 1990. Recapture data from 373 tags recovered through the first quarter of 1992 re- vealed a strong seasonal component in fish movement between summer and winter areas. Prespawning fish were tagged throughout their sum- mer distribution, primarily over the inner and middle shelf (<30-100 m depths), and recaptured on the outer shelf ( > 100-200 m ) and upper conti- nental slope (>200 m) in subsequent quarters. Recoveries from the win- ter quarter (January-March) showed the most directed movement, when Pacific cod aggregated in ma- jor spawning areas between Unalas- ka and Unimak islands in the east- ern Aleutian Islands, seaward of the Pribilof Islands along the shelf edge in the eastern Bering Sea, and near the Shumagin Islands in the west- ern Gulf of Alaska. By early summer, a hypothesized postspawning dis- persal was observed from these over- wintering areas, when tagged Pacific cod moved from deep off-shelf waters to shallower depths on the eastern Bering Sea shelf. The importance of seasonal migration was examined statistically by contingency table analysis, which indicated that sea- son of recovery affected area of re- covery more than either the season or area of tagging. Seasonal move- ments were further quantified by modeling the population dynamics of tagged individuals, which allowed estimation of the seasonal distribu- tion in the eastern Bering Sea popu- lation. These estimated seasonal dis- tributions compare well with the sea- sonal distribution of catches from the commercial fisheries. This analysis of tag-recapture data suggests a single winter spawning population in the eastern Bering Sea, nearby waters of the Aleutian Islands, and western Gulf of Alaska waters be- tween longitude 157°W and 170°.W Seasonal movements of Pacific cod, Gadus maerocephalus, in the eastern Bering Sea and adjacent waters based on tag-recapture data Allen M. Shimada Alaska Fisheries Science Center National Marine Fisheries Service. NOAA 7600 Sand Point Way NE. Seattle, Washington 98 1 I 5-0070 Present address. Office of Research and Environmental Information National Marine Fisheries Service, NOAA 1335 East-West Highway. Silver Spring, Maryland 20910 Daniel K. Kimura Alaska Fisheries Science Center National Marine Fisheries Service, NOAA 7600 Sand Point Way NE, Seattle, Washington 98 1 1 5-0070 Manuscript accepted 6 April 1994. Fishery Bulletin 92:800-816. 800 Pacific cod, Gadus maerocephalus, are widely distributed throughout the North Pacific Ocean. Off North America, Pacific cod range from California north through British Columbia and southeast Alaska, into the Bering Sea, and west along the Aleutian Islands (Bakkalaetal., 1984). Pacific cod are found epiben- thically over the Bering Sea shelf and slope. The Bering Sea repre- sents the center of greatest regional abundance, although Pacific cod are also abundant in the neighboring Aleutian Islands and Gulf of Alaska waters (OCSEAP, 1987). Pacific cod catches rank third among the east- ern Bering Sea groundfish resources following walleye pollock, Theragra chalcogramma, and yellowfin sole, Pleuronectes asper (Low, 1991 ). Beginning in the early 1980's, Pacific cod catches increased sub- stantially above the 10,000 to 50,000 metric tons (t) landed annu- ally between 1958 and 1979. East- ern Bering Sea and Aleutian Islands catches over the last decade have averaged 132,500 t; a historic high of 198,000 t was taken in 1988 (Thompson, 1994). Much of this growth was due to the recruitment of exceptionally strong 1977-78 year classes, in combination with greater fishing effort from joint- venture (i.e. U.S. fishing vessels de- livering catches to foreign proces- sors at sea) and new domestic groundfish fisheries (Bakkala, 1984; Shimada, 1985). More recently, the eastern Bering Sea and Aleutian Is- lands Pacific cod fishery generated a catch of 177,300 t, valued at $90 million in 1991 (NMFS, 1992). These developments in resource availability and in expanding fish- ery exploitation patterns provided the impetus for new studies into the biology of Pacific cod in Alaskan waters. In September 1982, a pilot tag- ging experiment for Pacific cod and walleye pollock in the eastern Bering Sea was conducted by Na- tional Marine Fisheries Service (NMFS) scientists from the Alaska Fisheries Science Center (AFSC). The initial objective of this experi- Shimada and Kimura: Seasonal movements of Gadus macrocephalus 801 ment was to evaluate the feasibility of a tag- recapture program for Pacific cod and walleye pollock (Shimada1). During the second year, as tag-recovery information accumulated, field efforts focused exclu- sively on Pacific cod because few tagged walleye pol- lock were recovered. Growth-increment data obtained from these tag returns have been analyzed by Kimura et al. (1993). This paper presents new information on the seasonal movements and long-range migra- tion of Pacific cod in the eastern Bering Sea and ad- jacent waters. Materials and methods Between 1982 and 1990, Pacific cod were success- fully captured and released from AFSC-chartered fishing vessels engaged in summer bottom trawl sur- veys off Alaska. Pacific cod were tagged throughout their eastern Bering Sea distribution (Fig. 1). This ef- fort was augmented by tag releases from cooperating Japanese, Korean, and U.S. research vessels operat- ing in the Aleutian Islands and Gulf of Alaska (Fig. 1 ). Capture gear included bottom trawls, pots, and hook-and-line. For the bottom trawl, predetermined stations were sampled each year across the eastern Bering Sea shelf (Bakkala, 1993). Thirty-minute trawl hauls and biological samplings were performed at each station. Occasional opportunistic hauls of 10 to 30-minute durations were made to obtain addi- tional tag releases. On retrieval of the trawl net, Pacific cod were taken from the unprocessed portion of the catch and placed in on-deck holding tanks sup- plied with running sea water. After a recovery pe- riod (typically 1-2 h), fish were removed with a dip net and examined for visible signs of injury or stress. Those not seriously harmed during capture were placed in a padded cradle, tagged, and measured for fork length to the nearest 0.5 cm. General condition was noted, and fish were quickly returned to the sea. Two tag types were used in this study: 3.5-inch anchor tags and 8-inch lock-on spaghetti tags.2 Both types were constructed from international orange #20 vinyl tubing and labeled with a tag serial number and return address. The majority of releases were made with the spaghetti tag (69%). This tag was applied through the dorsal musculature, behind the head and anterior to the first dorsal fin, with a hol- low needle applicator and secured by interlocking 1 Shimada, A. M. 1982. Cruise results, NOAA ship Chapman, Cruise CH-82-06, 7 p. Available: Alaska Fish. Sci. Cent., NOAA, NMFS, 7600 Sand Point Way NE„ Bin C15700, Seattle, WA 98115-0070. 2 Floy FD68BC anchor tag and FT-4C lock-on spaghetti tag. plastic terminals. Anchor tags were inserted between individual fin rays at the base of the first or second dorsal fin. Pacific cod were tagged across the entire size range available to the capture gears (Fig. 2); priority was placed on the release of fish less than 55 cm (i.e. younger than about age 5 yr; conversion from length to age in this paper is based on growth data in Kimura and Lyons [1990]). Pacific cod smaller than about 42 cm (i.e. age 3 yr or younger) were not re- cruited to the commercial fisheries and were of par- ticular interest as the preexploited population com- ponent; however, owing to availability larger fish made up the majority of tagged fish. Data recorded at release included date, haul or set number, gear type, depth fished, bottom water temperature, re- lease position, fork length, and general fish health. Tag information from Pacific cod recovered by com- mercial trawl, longline, and pot fisheries (Table 1) through the first quarter of 1992 were used in this paper. The Bering Sea groundfish fishery was domi- nated by foreign fishing until 1987, when harvest allocations shifted to joint- venture and domestic fish- eries. By early 1991, the Pacific cod fishery had evolved into an exclusively U.S. enterprise. Mirror- ing this transition in fleet involvement, tag recover- ies initially came from foreign and joint- venture trawl and foreign longline fisheries. More recent tag re- turns have come from the domestic trawl and longline fisheries. Tag recovery reports provided information regarding capture date, catch location, and body length. Some tag returns also included capture method, depth fished, sex, body weight, maturity, and a collection of scales or otoliths. Tag recovery data were analyzed by three comple- mentary methods: 1) mapping, which described the location of release or recovery or the movement of fish from the area of tagging to the area of recovery; 2) multiway contingency table analysis (Fienberg, 1977), which was used to analyze the strength of re- lationships between the season and area of tagging and the season and area of recovery; and 3 ) direct popu- lation dynamics modeling of the tagged population. Three primary areas of interest (Fig. 3) were de- fined for use in the mapping, contingency table analy- sis, and the population dynamics model: 1) the inner shelf from depths <30 m to 100 m [Area 1]; 2) the outer shelf between >100 m and 200 m and incorporat- ing the upper continental slope at depths greater than 200 m [Area 2]; and 3) winter spawning grounds sur- rounding Unimak Pass and adjacent waters [Area 3]. To emphasize the main features in the data, release-recovery positions were plotted individually (Fig. 4) and as mean-movement vectors based on tags released over blocks of 2° latitude x 5° longitude (Fig. 802 Fishery Bulletin 92(4), 1994 Vi ■ 1 7*^01 "^'■i ■ :ndividual tag releases v_, ... ..i ...... . i i i — i 1 1 [ r rr^oo-* 16BT30'* 9 WW :ndividual tag recoveries ATIO'W 159 WW I SOW* OW*. Figure 1 Jittered plots showing locations of individual Pacific cod, Gadus macrocephalus, tag releases and recoveries (r=373), 1982-92. "Jittering" is accomplished by adding small amounts of random noise to the data to avoid overplotting. Shimada and Kimura: Seasonal movements of Gadus macrocephalus 803 s => o z o 40 60 80 SIZE AT RELEASE (CM) 100 120 a. asm 20 100 120 SIZE AT RECOVER i Figure 2 Length-frequency histograms of Pacific cod at time of release and recovery, 1982-92. 5) and aggregated by recovery season (winter= Jan- Mar, spring=Apr-Jun, summer=Jul-Sept, and fall=Oct-Dec; this definition of seasons is used throughout the paper). For these plots each 2° x 5° block is associated with just one arrow. The modeling of the tagged population was per- formed similarly to that done by Hilborn ( 1990) and by Heifetz and Fujioka ( 1991 ). However, rather than describing movement through time as a Markov pro- cess, we only estimated seasonal distribution, a much easier task. Our model assumed that the quarterly natural mortality rate (M) was the same in all sea- sons and areas, and that the instantaneous fishing mortality rate (F) varied by season but not by area. Years were assumed to be homogeneous (i.e. we as- sumed no year effects). Modeling the tagged popula- tion allowed the estimation of seasonal distribution by taking into account the actual time and area of tagging and recovery. The model assumed that time (i) is divided into seasons (four per year over all years, so that the sub- script i runs from 1 to 4 x nyr, where nyr=10, the number of years being modeled) and that there are three areas (j) being modeled. Table 1 Tabulations of tag releases and tag recoveries (top) by areas described in Figure 3 (Area l=the eastern Bering Sea shelf, Area 2=the outer eastern Bering Sea shelf and upper slope. Area 3= the main spawn- ing area) for Pacific cod, Gadus macrocephalus, in the eastern Bering Sea. Percentage release and re- coveries by fishing gear (bottom I. Release Recovery Area Number Percent Number Percent 1 2 3 Other Total 4,967 1,303 3,048 3,078 12,396 Release 40.1 10.5 24.6 24.8 91 107 158 17 373 24.4 28.7 42.4 4.5 Recovery Gear Percent Gear Percent Trawl Hook and line Pot Unknown 97.2 0.6 2.2 0.0 trawl longline pot unknown 62 24 2 12 804 Fishery Bulletin 92(4), 1994 T3 C CO 5 o o CO ^ \ \ \ "^i < o in 15 c CO ^-r ^--»/ viA* ^ *■•* •1 CD >. ^\ * ^S w" ■ J 3 i' \ ^F%S^J ■* - '• < CO k\~ \ <-?/?> 'MaR -6 Ll. >> ^o A. '^fe. vs ■■ o o a vj V.'y'l)^ '': ll _J cu be \\ ^ "• 3 O c o \ S_ ^) < ^ \ \ S ^ v\ ( \ ) ( (/) b a. m \N-A--^?\ S ■"* — c? CO < r\ \\\Vv c \ : in S _i xy . \\ %\&, o "~ !m < ^~o s- ' x'\ s4 E c2 /^ CD \ <,_ W fi '. '■ tD 13 < Y %r* , -J " ' 5 CD «£* S, . ' Z CO 3 ^1 ■ : < CL '• ■. CO r o ; - Art ■ '< CO CO cu u CO \ 'i :a\£ CO $S I : 1 i ^ LjJ .ce S-S CO T3 x V^V - S ™ ■^ ' uo _p.tr in c ure 3 ea boui ^if^ r\i ■•■£ £ ■ ■■$*■ U \\Vl D W m - V 5 < / m QJ E CL '• f°" .V ■ G 'E 3 • _ CO !r. ^ "-" LJ / o T""^V c / ,o E ** CO 4> ■ — / £ / " o < /•.■-. o ■ CM >> fO LJ V) "' t/1 cu ■£■ < - p a> / .. . *- E , ■R.- : E ■o s CO o ■ °o- - ■ S /•■£'' • ■ . . o .' '^. 'O '.o £1 %/ ■■■$>■:'■■ / M "°- .CN c S o c/l / ,••-•■/ 'ft b 'El cu CU ■ J= ■4-3 / ".•*: C / a • / v 'is /.. i.-i o / ^ ■A CO / f'/ •: ^ '• ■'/ ," / 1 LlJ 1 < bE "cu cu P ■ / <0~ " ■ •"•'/' o * ^ •^ CO '■■/ ^■/ ■■- b C .S CU P *i CO '■:':■ "-> ° /'■ a the eas on dyn i i A i i o in o in c^ <*. ■" P F - p o CO CM T. !_£, - ^ ^ 'a i£> iTl u- in co g- dpruuoi Shimada and Kimura: Seasonal movements of Gadus macrocephalus 805 s o {wlj o m \H Bl tn ' VR th 5 < CO UJ i4*$> \\ : in < 4 ^v* z " f&\ m Q Ld ,cf.y *^ol O D O Vnf' °~ ^ < h- v9 U5 O I / \ U_ J / / * //// s b PQ •1 \ I 1^ b p o z z ■z. z :z o in o un o P 5" ^ ? P (Nl Ol l£) r-~> r- in IT) iD u-i spnvioi < i.i i (X < Q LJ CC UJ > O o UJ (X I CO CM < UJ DC < Q UJ O o < h- x en < ui [X < Q UJ O O < X 00 CO 0) CU ft fe ° c 12 co 'a; to T5 £| .S o "O 2 be -" cm c co -ft o bo u bE O CO £ 8 "cO s T3 CO ■5 "co C 3 5 o > -s o C II CQ £ ~ o 5 co S ,? «rl « S> O &D o bC CU CO T3 -O O O CO CO CL, CU "co "a 3 3 T3 -a ;g ;s -5 -3 c g o o c c CU CU e a cu a> > > o o S 6 806 Fishery Bulletin 92(4). 1994 Z> ""3 cm o. •-I o z a a o Q ■! :• i >- n •i •i ~3 I) LlJ II a . i i a: i,i m o 1 > o Si QJ u 01 X 3 £ o -a o to es 03 T3 cu 3 C CD E CU > c E CD bi C3 h CD < CO c UJ in ii iii c/i I - . j Z) CO b o spninol 3 C -w o O. en « es o CD CO 5 I- u -•-, O c£ r SI aJ m *a3 w 1-. be S.S ° h '-.■' i <*- -o o-g ■O ,--. U CD tC > 1 8 CO v PL, I-. 5 *S "O O ° C < j r ■ I -a o> o S3 £ g. 5 a o g o to O. it, CO ?» — QJ *- CO CD CD 'CJ be 3 CO ■" > 5 < .2 Shimada and Kimura: Seasonal movements of Gadus macrocephalus 807 Let ru = N, Ne = y sij = "y = Ps(i) = the number of fish tagged in time period i in areaj, the recoveries of tagged fish in time period i in areaj, the number of tags at the beginning of time period i in area j, after tags have been redis- tributed according to estimates of their sea- sonal distribution, the number of tags at the end of time period i in areaj, s, = exp(-M - Fsll) ) the survival of tagged fish in time period i and areaj, Ul = Fsil)(l.0-exp(-M-FsU)))/(M + FsU)) the exploitation rate of tagged fish in time period i and areaj, and the areal distribution by seasonal time period. Here, the subscript s(i) refers to the season corre- sponding to the time period subscript i. Therefore s(i) could be w=winter, sp=spring, sw=summer, or /"=fall. Each seasonal area-distribution vector (pw for example) is a vector containing one element for each of the three areas. There are 13 parameters to be estimated in this model: the seasonal distribution vectors (pw, psp, psu, pf); the seasonal instantaneous rates of fishing mor- tality (Fw, Fs , Fsu, Ff); and the seasonal instanta- neous natural mortality rate (M). The seasonal area-distribution vectors each contain only two pa- rameters to be estimated because they are probabil- ity distributions that must sum to one. The model is tied together by three simple equations: N^Nfc + Refii, r^Nfc + Rfr/2, Nb f \ V J ) Pjisii+m • Here, pjlsU+l)] represents the estimated proportion of tags in areaj in season s (i+1). Note that -^s, and u,/2 are the estimated survival and exploitation rates, respectively, over half a season. Following Hilborn (1990) and Heifetz and Fujioka ( 199 1 ), the model was fit by using maximum likelihood and by assuming recoveries were distributed as Pois- son random variables. That is, the parameters were estimated by minimizing minus the log-likelihood: -L = ]T rtJ - rtJ log(/v ) + const . f \ I* V •> / The probabilities in (pw, psp, psu, pf) were modeled as expressions similar to exploitation rates following the method of Heifetz and Fujioka (1991). Parameters were estimated on the logarithmic scale and coeffi- cients of variation were estimated by using the in- verse Hessian of the minus log-likelihood and the delta method. Our seasonal population dynamics model of the tagged population was applied to the areas described in Figure 3. For modeling purposes, we used only fish tagged and recovered in these three areas. Thus, 9,318 releases, and 353 recoveries, V o were available for analysis. A problem associated with our population dynam- ics model is the strong assumption that the Fs(i) are constant across areas. An attempt was made to use existing commercial trawl and longline data to de- termine recovery effort, but these data varied too much in their seasonal coverage, and gears and ar- eas were confounded. Nevertheless, the population dynamics model pro- vides evidence that our tag data are representative of the entire eastern Bering Sea population. In the following sections we present evidence that our tag- ging study probably suffered from significant tag loss, tag mortality, or under-reporting of tag recoveries. By comparing the estimated seasonal distribution of the tagged population with the distribution from com- mercial catches, we can verify that the tagged popu- lation and untagged population were distributed similarly. Since catch distribution should reflect abundance in a heavily fished region such as the eastern Bering Sea, we interpret this as meaning that the behavior of the tagged Pacific cod population largely reflected similar patterns in the entire Bering Sea popu- lation. Commercial catch statistics were taken from the Alaska and Pacific Northwest Historical Groundfish Database (Berger3), from which we calculated the ar- eal trawl and longline catch distribution (for numbers), by season, for the 1982-92 study period. Results Approximately 12,396 tagged fish were released be- tween 1982 and 1990 (Table 1; Fig. 1). A total of 373 3 Berger, J. Alaska Fish. Sci. Cent., Seattle, WA 98115-0070. Per- sonal commun., January 1993. 808 Fishery Bulletin 92(4). 1994 tag recaptures with useful information were reported to the AFSC through March 1992 for a recovery rate of 3.0%. Although tag returns occurred over a broad area and time period, this rate is much lower than the 24-26% recovery rate reported by Canadian and U.S. researchers working off British Columbia, Canada, and in Puget Sound, Washington (Westrheim, 1984; Karp, 1982). In our study, research trawl-caught fish accounted for the majority of tag releases (97%), and commercially fished bottom trawls and longlines accounted for the majority of tag recoveries (Table 1). Thompson ( 1994) estimated that the exploitation rate for Pacific cod in the east- ern Bering Sea during 1981-1992 averaged 9-11% annually. This suggests additional tag loss, tagging mortality, or under-reporting of rates which sum to about 2/3 in some combination. Fish lengths at release were between 25 and 118 cm, representing fish as young as age 2 yr as well as very large, mature fish. The distribution of lengths at recovery corresponds well with the overall tag-release size frequency but is shifted to the right due to growth and gear selectivity (Fig. 2). In the commercial fisheries, Pacific cod are first recruited at about 40 cm or age 3 yr. They become available to different gear types (i.e. initially to bot- tom trawls, then to longline gear) at progressively older ages and larger mean size (Shimada, 1985). Most tag recaptures were of commercially recruited, sexually mature fish, older than age 5 yr, and larger than about 60 cm, as defined by Teshima ( 1985). More than 75% of all tag releases were from U.S. survey vessels in the eastern Bering Sea (Table 1; Fig. 1 ). Cooperative foreign research vessels operat- ing in the Aleutian Islands and Gulf of Alaska were responsible for the remaining 25%>. Twenty-four per- cent of recoveries were made over the inner shelf, and 29% over the outer shelf and upper slope (Table 1; Fig. 1 ). Of particular note was the high concentra- tion of tag recoveries (>42%) in Unimak Pass and its surrounding waters (including the adjacent western Gulf of Alaska) during the winter months (Fig. 1). Only 17 tags (<5%) were recovered from outside the three primary study areas (i.e. from the outer Aleu- tian Islands and the central Gulf of Alaska). Mapping seasonal movements Tagged Pacific cod exhibited marked spatial and tem- poral displacement from their point of initial release (Fig. 4). Individual movements generally conformed to seasonal shifts in centers of Pacific cod abundance (Ketchen, 1961; Bakkala, 1984) and to the corre- sponding movement of fishing effort ( Shimada, 1985 ). We attribute the observed pattern in tagged fish movements to hypothesized migratory shifts between perennial summer (feeding) and winter (spawning) areas (Moiseev, 1952, 1953; Ketchen, 1961). This is most easily seen in the vector movements of indi- vidual fish into and out of the main spawning area, Area 3. These data were grouped in two ways to show 1) the origin offish released in all areas and recov- ered within Area 3 [Fig. 4A]; and 2) the outward re- coveries of fish tagged within Area 3 [Fig. 4B]. Al- though the movement into and out of Area 3 is clear, the movement into the spawning areas seems to oc- cur in two stages: 1) movement off the inner shelf [Area 1| into slope areas [Area 2] [Fig. 4C|; and 2) subsequent movement into spawning areas in Areas 2 and 3 [Fig. 4D]. This shift is counterbalanced by spring and summer recaptures on the inner shelf [Fig. 4B]. The annual cycle of Pacific cod migration appears to begin in late September, when tagged fish move off the eastern Bering Sea shelf and seaward to the 200 m shelf break. By fall, tags were recovered, pri- marily along the outer shelf edge. In winter, Pacific cod converged in large spawning masses over rela- tively small areas. Major aggregations were usually encountered between Unalaska and Unimak islands on the Bering Sea side of Unimak Pass. Other recur- ring centers of abundance were located southwest of the Pribilof Islands along the shelf edge and near the Shumagin Island group in the western Gulf of Alaska (Fig. 1). Following the spawning season, tagged Pacific cod dispersed from these overwintering areas and were recaptured farther inshore in concert with seasonal warming of the inner shelf environment. For ex- ample, fish tagged in areas of deep off-shelf waters adjoining Unimak Pass, close to the time of winter spawning, were recaptured progressively over the shelf (and especially north of the Alaska Peninsula following the 30-m isobath ) beginning in late spring. Tagged Pacific cod also moved to the northwest outer shelf (100-200 m) during the spring quarter. By sum- mer, the feeding range was well established back in central Bristol Bay (30-50 m) and the outer shelf. This distribution persisted until late fall and the beginning of the next yearly cycle. The seasonal nature of Pacific cod movement is most easily seen in the average vector movement of fish tagged within a particular 2° latitude and 5 lon- gitude rectangle and recovered during a specific sea- son (Fig. 5). From these maps, the off-shelf move- ment is clearly visible in fall, and movement to the spawning ground is clearly visible in winter. How- ever, spring and summer maps show relatively little directed movement, because during these time peri- Shimada and Kimura: Seasonal movements of Gadus macrocephalus 809 ods, Pacific cod have presumably returned to the feed- ing grounds on which they were originally tagged. The absence of any definitive within-season pattern is further illustrated in Figure 6. Pacific cod released and recaptured during the same three-month period (i.e. within the same season, perhaps in different years) showed only random movement and little di- rectional bearing. This is in marked contrast to the strong interseasonal movements traced between fall-winter and winter-spring quarter tag recaptures. Multiway contingency analysis Results from a multiway contingency table analysis (Table 2) indicated that month of recovery most strongly influenced the area of recovery. Although area of tagging also had a significant affect on the area of recovery, the month of tagging was seen to have only a relatively small affect on the area of re- covery. Thus season of recovery appeared as the strongest correlate to area of recovery, which further supports our finding of strong seasonal migrations in Pacific cod. Seasonal-areal population dynamics modeling and catch distribution A histogram of residuals from predicted tag recover- ies indicates that the 13-parameter population dy- namics model of the tagged population fit the data quite well (Fig. 7). The parameter estimates from the tagging model (Table 3) show strong movement from Table 2 Multiway contingency table analysis of tag recovery data for Pacific cod, Gadus macrocephalus, in the eastern Bering Sea. The model examined all two- factor interactions (i.e. no factor deletions). Each two- factor interaction was tested for significance by de- leting it from the model. Factor 1 = season of release (4 levels), 2 = area of release (3 levels), 3 = season of recovery (4 levels), 4 = area of recovery (3 levels). Factor Likelihood Test of deleted ratio stat. df P interaction none 66.150 96 0.0087 not applicable [12] 139.382 102 0.9917 (Z=8.786)' [13] 75.892 105 0.0145 (Z=0.291) [14] 79.206 102 0.0459 (Z=1.793l [23] 118.317 102 0.8713 (Z=6.898) [24] 209.201 100 1.0000 (Z=14.269)*** 134] 245.890 102 1.0000 *** ' Interaction large but fixed by design. Z-statistic is a standard normalization of the hierarchical chi-square test. *** Interactions were significant; a=0.0001. Table 3 Estimates of parameters (and coefficients of varia- tion measured as proportions) for the population dynamics model for tagged Pacific cod, Gadus macro- cephalus, in the eastern Bering Sea. Areas are shown in Figure 3. 1 Quarterly instantaneous natural mortality rate: M= 0.235(0.109). 2 Seasonal instantaneous fishing mortality rates: F =0.01389(0.149), F =0.0082(0.176). Fsu =0.0075(0.152), Ff =0.0087(0.141) 3 Seasonal distribution over (Area 1, Area 2, Area 3): a pu = [0.0246 (0.570), 0.1884 (0.188), 0.7870 (0.047)] b psp = [0.4849 (0.148), 0.3937 (0.175), 0.1214 (0.371)] c psu = [0.5830 (0.122), 0.2330 (0.255), 0.1840 (0.318)] d pf = [0.1103 (0.304), 0.5167 (0.105), 0.3730 (0.143)] the major spawning area (Area 3) in spring, further movement onto the shelf (Area 1) in summer, move- ment off the shelf in fall, and movement to the spawn- ing areas in winter. Furthermore, the model results confirm strong seasonal movement between areas in a manner consistent with our previously described pattern of seasonal Pacific cod movements. The apparent high annual instantaneous natural mortality rate ( M = 0.96 ) and low annual fishing mortality rate ( F = 0.038 ) of the tagged population can be noted. This is probably due to tag loss, tag- ging mortality, or the under-reporting of tag recov- eries as previously discussed. Multiplying releases by 1/3, or multiplying recoveries by 3 would lower M and increase F nearer to expected levels ( M = 0.87, F = 0. 1 1 ). However, the model fit and es- timated seasonal distribution of the tagged popula- tion were not affected by this scaling of observed re- leases or recoveries. Therefore, the population dy- namics model estimates of seasonal distribution ap- pear to be robust to tag loss, tag mortality, or under-reporting of recoveries. The estimated areal distributions (within seasons) of the tagged population largely reflect the areal dis- tribution of tag recoveries. Also, although no catch data were used in the population dynamics model, the results coincide well with the estimated seasonal distribution of the commercial catches (Table 4). The main difference appears to be that the hypothesized fall season off-shelf movement, and subsequent move- ment into the winter spawning area (Fig. 4, C and D), is more pronounced in the commercial catch data 810 Fishery Bulletin 92(4), 1994 $ ':■■ %.., b o O < o in V^ 6 Cv3^ in • D * ' ■ 1 o . " • ^3 c/i < < Li_ o _l . • • . O CO H CO 0) >> +-< c i < L _ . ' a c/> < \ ( 5. m ov ^ D . t E • cr D O- b -T : in CO u CO _l *~ Q. < %■;] : ~ 1 A4 D C o CO CB CD CO _^^j^i ^$k / P .in CD 0 CO CO CD c ^~~^*-» 1 tt'T^l \, \V V / s«i •^- ~"\ J 1/ b^ -a o-^ . CD _ ^ ' ' • So;-' ]/• 3f ' c' in *U o1 •O i- ^l c )\% ^^y^ljX .2' - m c *- _J tu > 3 S 2 V '•O u. -a c CO 0> ^ ' z "a> "^s/ •" :"C\ £ . "~---^Qr: XI o m^^--^ 0) bo in ■5r . "a! • £ o in A '^ CO -o o o lj iy CN . O .CN in ui E .-■ • -: CO . . . ■' < J —•2 .'*"' ' '■ FT b r- (X 3 '' " j S. jr 6. ■ . "2 > \ ■> *i V 1 d * ■' on ■;.'•.! z 'ftfV '. 3 / # a- c o -m c a> 6 CD > O e a _l / -.' D.' < / .••«si:4;-. o .-:.""'"v':-- ...\. o C c CO .■•■■■v--v-*^ ■ i ■I i I i i i b un b in b P S* £1 ? P CN Oi •^ r^ — j-i U 1 spn|i»on in in Shimada and Kimura: Seasonal movements of Gadus macrocephalus Time (quarters) Illl 1 l.lllll.l 1 . . Residual Figure 7 From the fit of the population dynamics model to the tagged Pacific cod population: graph of expected versus observed tag recoveries by quarter (top) and histogram of residuals by quar- ter and area (bottom). (Hollowed and Low4) than the tagging data. There- fore, we conclude that observed movements in the tagged population generally reflect the population movements of Pacific cod in the eastern Bering Sea. Emigration and immigration relative to the study area We provide direct evidence that Pacific cod migrate from the eastern Bering Sea into the Gulf of Alaska. Of the 95 winter recoveries made in Area 3, 21 of these occurred in the Gulf of Alaska (Fig. 8; note that Fig. 8 includes 30 Gulf of Alaska recoveries from all seasons). Longline vessels operating in the winter quarter between Sanak Island and Shumagin Bank were the main source of returns. These data suggest that 22% of fish found in Area 3 in winter may mi- grate into the Gulf of Alaska. Multiplying this figure by the population dynamics estimate of 78% (which is the total eastern Bering Sea winter population within Area 3), or by the 68% of the winter commer- cial catch that is taken in Area 3, suggests that 15 to 4 Hollowed, A. B., and L. L. Low. 1986. Productive gadoid fishing grounds based on species assemblage analysis. In M. A. Alton (ed.), A workshop on comparative biology, assessment, and man- agement of gadoids from the North Pacific and Atlantic oceans. Proceedings: part II, p. 681-712. Available: Alaska Fish. Sci. Cent., NOAA, NMFS, 7600 Sand Point Way NE., Bin C15700. Seattle, WA 98115-0070. Table 4 Distribution of catch (numbers) by season and area (Area l=the eastern Bering Sea shelf, Area 2=the outer eastern Bering Sea shelf and upper slope, Area 3= the main spawning area), from 1982 to 1992, for Pacific cod, Gadus macrocephalus, in the eastern Bering Sea. Estimated from the Alaska and Pacific Northwest Historical Groundfish Database (J. Ber- ger, Alaska Fish. Sci. Cent., Seattle, WA 98 115-0070, Personal commun., January 1993) and from the rela- tive magnitude of trawl and longline catches. Season Area 1 Area 2 Area 3 Winter Spring Summer Fall 0.085 0.496 0.703 0.176 0.235 0.393 0.276 0.651 0.680 0.111 0.021 0.173 17% of the total population in the eastern Bering Sea may migrate into the Gulf of Alaska during winter. A number of individual longer-range migrations tie together the Pacific cod population from 150° W to 180° W longitude (Fig. 8). We note with interest the recovery of two Bering Sea tags in the central Gulf of Alaska near Cape Chiniak on Kodiak Island, after 103 and 334 days. Other tagged Bering Sea emigrants have been recaptured on the North 812 Fishery Bulletin 92|4). 1994 o i/l o ld b o f< ■* o CN Ol (O ro «- T I i T , i , t , i ,T , T S - I ■ o ; i ■ i in to C n y. 4) ' Q. * < ~ Vs s 00 < E - 4 o z- « \V o < 13 c o - *^.\ o Li_ •ffl 9) " ^s' ' \ ' ^ O 0) < w _* S — if) ^\« — - c b r ' o c ~ < -• . ^\ \ '"^ ° . c E ', cy> 3 - O m - < "7"r3 \ \4 X / ^ M CO es 53 - tf° ^7T~~* - \^,-f ''•AT ^M> - o o -~ ^/b&C' / u2\ CD +3 P. o c 1/) •v "t^yl /I -^ JUftp c' tft re 8 gSea - ^l r \ i^r VTA ■ o „ tD O ■- _l Figu Berin ;i a to s , / — 0) \\\M "£. c s l> :o E. W-UA 3 , M CD - ...-■-• .-*'• A . ■ • o HI in 5: Ld r.i o in- Vv'^^X ' ^ CO C8 aj ■ E o -;z EH LU .-- o \ ■ o P. . \ ^ trt .O csi \ d cm ■\ <3CL - , B o in (ft < to / . 0> / / / CL • -a o i/i "Hi / "\ E 5 o ■$ ,-f ■ E 7 ' v\o ' b fS J o / ElP /; G -- - ■ £ J/-- o / CM CO :-■ ^ O 3 ' i i A1 1 ' i ' I 'A ' ' i b .M O lo b ? 5" '.; > ? 9 CN Ol ID K> — to if if) 3pnv>Dn in in Shimada and Kimura: Seasonal movements of Gadus macrocephalus 813 Pacific side of Akutan Pass and Unimak Pass in the Aleutian Islands. Additionally, two fish tagged in the major spawning area (Area 3) were recaptured to the west in Seguam Pass within 250 days. In a striking occurrence of immigration to the Bering Sea, a pair of Pacific cod (65 cm fork length) tagged in Tanaga Pass near Adak Island were recaptured on the outer northwest shelf (above 57°N) after 3 and 5 years at liberty (Fig. 8). Although substantial numbers of Pacific cod were tagged along the Aleutian Islands west of 170°W, in- cluding to about 174°E (Fig. 1), few recoveries have been made. These releases came from a single 1986 summer trawl survey and tagged fish were in poor condition because commercial foreign fishing opera- tions were employed. Discussion Our analysis identified a seasonal circuit that we attribute to annual migrations for spawning and feed- ing; it also provided preliminary indications of emi- gration from the eastern Bering Sea. The former is described in terms of three eastern Bering Sea ar- eas. The latter ties together more expansive distances as defined by regional geography or fishery manage- ment boundaries, or both (OCSEAP, 1987). Although the majority of tagged Pacific cod exhibited the sea- sonal character of short-term cross-shelf movements, a small number of tagged individuals provided em- pirical evidence for much longer transits. We recognize that emigration and immigration probably occur with respect to the main study area (i.e. Areas 1-3). However, the locations of tag releases and numbers of tag recoveries received to date make it difficult either to quantify the amount of emigra- tion from the eastern Bering Sea, or to conclude with certainty that return immigration to the eastern Bering Sea is significant. At this time, we believe there is some eastern Bering Sea immigration from the surrounding regions. However, considerable un- certainty exists because so few Pacific cod were tagged outside the study region, most importantly in the central and eastern Gulf of Alaska (Fig. 1). Also, we have some evidence that the western Aleu- tian Islands stock(s) may be fairly independent of the eastern Bering Sea, but this evidence is far from conclusive. Despite these conjectures, statistics such as "dis- tance traveled" and "rate traveled" versus "time at liberty" (Fig. 9) generally support our seasonal move- ment model. Observed "distance traveled" already is maximized within the first year of freedom. Simi- larly, observed "rate traveled" is maximized within the first year at liberty. This behavior is consistent E ■o > CO 1— Q> O c 03 a o o o o ro o o o D a D D D D □ ci D D S D a d □ _D n □ □ □ ~ D D |§mp 11 □ DO a ff □ D HDd D □ a D a^iLi n ° D [ftP D □cJ-f§§deL OjOD □ D ;flkB a nn* U WbB) fDi° DO C a a a D 1@ a D CV D D 500 1000 1500 Time at liberty (days) 2000 co I T3 a> S3 m ■ 1 1 1500 2000 500 1000 Time at liberty (days) Figure 9 Scatter plots showing (left) the relationship between distance traveled (in nautical miles) and time at liberty and (right) the relationship between rate of travel and time at liberty for Pacific cod. 814 Fishery Bulletin 92(4). 1994 with seasonal migrations within a closed system. Other movements occur against a backdrop which is dominated by regular seasonal movements. Movements of Pacific cod may be better understood in the context of the general life history, population dynamics and physical environment requirements of this species. Bakkala (1984) examined distribution patterns based on an analysis of research survey and commercial fishery catch per unit of effort and size composition data. He described a gradual shift over the southeast shelf with time, corresponding to a progression in cohort ages and the influence of year-class abundance. A tendency towards the off- shore environment was noted from coastal waters to the outer shelf and slope edge. This was based on an areal transition stemming from ontogenetic develop- ment in younger age ( 1-3 yr) to older age (4+ yr) groups. Further, during years of higher than average abun- dance, the population range was much more extensive than that in years of low abundance (Bakkala, 1984). From Russian trawl surveys, Stepanenko5 de- scribed winter concentrations along the upper slope at depths between 400 and 545 m. Prespawning and spawning aggregations were consistently found northwest of Unimak Island, in the Pribilof Islands sector, and along the northern slope on either side of the U.S. -Russia Convention Line. The most signifi- cant spawning aggregations occurred in the vicinity of Unimak Pass along the outer shelf edge. The literature indicates that preferred water tem- peratures (0 to 10°C) are the primary factor for de- termining centers of Pacific cod abundance. Towards its southern range off British Columbia, Ketchen (1961) reported highest catch rates at bottom tem- peratures of between 6 and 9°C. Off Russia, Moiseev (1953) noted that spawning Asian Pacific cod pre- ferred 80-290 m depths and water temperatures be- tween 0 and 2 to 3°C; optimal summer temperatures were between 0.2 and 4.5°C. Hirschberger and Smith ( 1983 ) reported water temperatures around 5.4°C for spawning Gulf of Alaska Pacific cod at 150-250 m. In the eastern Bering Sea, high winter concentra- tions of Pacific cod coincide with warmer water ( mean 4.0°C) found year-round in depths off Unimak Pass and the upper slope ( Kihara, 1982, a and b; Bakkala, 1984). Bottom temperatures on the shelf drop from the 0.2 to 4.5°C range in summer to below 0°C in winter (Schumacher and Reed, 1983). Thus at the high latitudes of the Bering Sea, the stimulus for 5 Stepanenko, M. A. 1989. Condition of stocks, interannual vari- ability of catch per unit of effort and fishing of cod in the east- ern part of the Bering Sea. Pacific Research Institute of Fisher- ies and Oceanography (TINRO), Vladivostok, USSR. Document submitted to the US-USSR bilateral meetings, November 1989, 35 p. Available: Alaska Fish. Sci. Cent., NOAA, NMFS, 7600 Sand Point Way NE., Bin C15700, Seattle, WA 98115-0070. offshore migration appears to be avoidance of the intense cooling of inshore waters that accompany advancing ice formation from the Bering Strait in favor of warmer temperatures at depth. The spring feeding migration, shoreward, is most likely timed to the warming of the coastal shelf environment and a return to summer norms (Bakkala, 1984). Interestingly, at lower latitudes, seasonal migra- tions are reversed. At the southernmost edge of its range, off Korea, Japan, and in Puget Sound, Wash- ington (Karp, 1982; Mishima, 1984; Zhang, 1984), Pacific cod migrate to deep offshore waters during summer months to avoid excessively heated (>10°C coastal waters. A returning inshore spawning migra- tion occurs each winter. Moiseev (1953) noted very limited along-shore migrations in Russian waters. However, active sea- sonal migrations between coastal shallows and off- shore depths perpendicular to the shoreline (mainly in response to inshore and offshore temperature shifts) were found. He further hypothesized that the potential for stock intermingling was reduced because of limited along-shore movement. Winter offshore move- ments were observed throughout the northern range of Pacific cod. Local abundance centers were always in the direction of the preferred temperature regime in response to the pronounced cooling of their onsheff en- vironment. Westrheim ( 1984) noted that Pacific cod off Vancouver Island, British Columbia, exhibited the same bathymetric seasonal movements as Pacific cod in Alaska but very limited along-shore movements. In all of the cases described above, seasonal mi- grations of Pacific cod appear to be triggered by the desire to avoid temperature extremes that accom- pany the changing seasons. In the eastern Bering Sea, movements represent necessarily long-range migrations across the Bering Sea shelf which on av- erage is 300 nmi wide. For southern coastal stocks, the same result can be achieved with much shorter offshore migrations to depth. It is likely that inter- regional along-shore migrations seldom are found because they are unnecessary for achieving the pre- ferred temperature regime (Moiseev, 1952). Rose ( 1993) found a similar pattern in Atlantic cod, Gadus morhua, based on hydroacoustic surveys. He attributed seasonal movement to springtime feeding migrations, which shifted Atlantic cod from offshore winter spawning grounds shoreward. Migration path- ways were facilitated by stable bottom temperature regimes (2— 3°C) associated with trenches on the north- eastern Newfoundland shelf. How large-scale stock structure is affected by mi- grations motivated by preferred temperatures is unclear. Grant et al. (1987) screened Pacific cod ge- netic samples from throughout their range. Two ge- Shimada and Kimura: Seasonal movements of Gadus macrocephalus 815 netically distinct stocks were detected: a western North Pacific Ocean (Asian) group, and an eastern North Pacific group which included the Bering Sea, Aleutians Islands, and Gulf of Alaska regions. There were virtually no regional genetic differences among any of the North American samples. These authors were unable to identify where the effective northern boundary between the western and eastern groups lies, though the western Bering Sea was considered most likely. Grant et al. (1987) attributed this lack of genetic differentiation to gene flow between vari- ous subareas and regions on either side of the Pa- cific. Grant et al. (1987) were puzzled that most of the literature on Pacific cod pointed toward locally isolated stocks (Moiseev, 1953; Svetovidov, 1948; Ketchen, 1961; Wilimovsky et al., 1967). Our tagging study confirms sufficient migration to explain Grant et al.'s findings of genetic homogeneity in Pacific cod over broad areas of the North Pacific. We have confirmed from tagging that Pacific cod migration occurs between the Bering Sea and Aleu- tian Islands. Because of the experimental design, the majority of tag returns demonstrate emigration from the Bering Sea but have not shown conclusively that immigration to the Bering Sea takes place. Even so, this study shows that significant exchange may oc- cur within the open ocean populations of Pacific cod off Alaska. Whether Bering Sea Pacific cod have a reciprocal exchange to the wider Gulf of Alaska be- yond the nearby waters of major Aleutian passes re- mains an open question. Lack of data precludes any further statement or quantification of exchanges be- tween these regions. Further elucidation must await additional tagging results, particularly from the east- ern Gulf of Alaska and western Aleutian Islands. Acknowledgments We thank R. G. Bakkala, J. T. Fujioka, D. A. Som- erton, G. G. Thompson, V. G. Wespestad, and H. H. Zenger for reviewing our manuscript. Numerous in- dividuals in the domestic and foreign commercial fisheries, the Alaska Department of Fish and Game, and the National Marine Fisheries Service Fishery Observer Program contributed by returning tag-recapture information to the Alaska Fisheries Science Center. We thank Jack Turnock for develop- ing the coastal mapping program in the S Language which greatly facilitated our analysis. Literature cited Bakkala, R. G. 1984. Pacific cod of the eastern Bering Sea. Int. N. Pac. Fish. Coram. Bull. 42:157-179. 1993. Structure and historical changes in the groundfish complex of the eastern Bering Sea. U.S. Dep. Commer., NOAATech. Rep. NMFS 114, 91 p. Bakkala, R. G., S. Mishima, S. J. Westrheim, C. I. Zhang, and E. S. Brown. 1984. Distribution of Pacific cod in the North Pacific Ocean. Int. N. Pac. Fish. Coram. Bull. 42:111-115. Fienberg, S. E. 1977. The analysis of cross-classified categorical data. The MIT Press, Cambridge, MA, 151 p. Grant, W. S., C. I. Zhang, T. Kobayashi, and G. Stahl. 1987. Lack of genetic stock discretion in Pacific cod (Gadus macrocephalus). Can. J. Fish. Aquat. Sci. 44:490-498. Heifetz, J., and J. Fujioka. 1991. Movement dynamics of tagged sablefish in the northeastern Pacific. Fish. Res. 11:355-374. Hilborn, R. 1990. Determination of fish movement patterns from tag recoveries using maximum likelihood estimators. Can. J. Fish. Aquat. Sci. 47:635-643. Hirschberger, W. A., and G. B. Smith. 1983. Spawning of twelve groundfish species in the Alaska and Pacific coast regions, 1975-1981. U.S. Dep. Commer., NOAA Tech. Memo. NMFS F/ NWC-44, 50 p. Karp, W. A. 1982. Biology and management of Pacific cod (Gadus macrocephalus) in Port Townsend, Washington. Ph.D. diss., Univ. Washington, Seattle, WA, 119 p. Ketchen, K. S. 1961. Observations on the ecology of the Pacific cod (Gadus macrocephalus) in Canadian Waters. J. Fish. Res. Board Can. 18:513-558. Kihara, K. 1982a. Fluctuations of the water temperature and the salinity in the eastern Bering Sea. Bull. Jpn. Soc. Sci. Fish. 48(12):1685-1688. 1982b. Influences of the marine environment upon the formations of the benthic species community in the eastern Bering Sea. Bull. Jpn. Soc. Sci. Fish. 49(11:49-54. Kimura, D. K., and J. J. Lyons. 1990. Choosing a structure for the production age- ing of Pacific cod (Gadus macrocephalus). Int. N. Pac. Fish. Comm. Bull. 50:9-23. Kimura, D. K., A. M. Shimada, and S. A. Lowe. 1993. Estimating sablefish (Anoplopoma fimbria) and Pacific cod (Gadus macrocephalus) von Bertalanffy growth parameters using tag-recapture data. Fish. Bull. 91:271-280. Low, L. L. (ed.). 1991. Status of living marine resources off Alaska as assessed in 1991. U.S. Dep. Commer., NOAA Tech. Memo. NMFS F/NWC-221, 95 p. Mishima, S. 1984. Stock assessment and biological aspects of Pacific cod (Gadus macrocephalus Tilesius) in Japa- 816 Fishery Bulletin 92[4). 1994 nese waters. Int. N. Pac. Fish. Comm. Bull. 42:116-129. Moiseev, P. A. 1952. Some characteristics of the distribution of bottom and demersal fishes of far eastern seas. Izv. Tikhookean. Nauch-Issled. Inst. Morsk. Rybn. Khoz. Okeanogr. 37. [Transl. Fish. Res. Board Can., Transl. Ser. 94.] 1953. Cod and flounder of the far eastern seas. Izv. Tikhookean. Nauch-Issled. Inst. Morsk. Rybn. Khoz. Okeanogr. 40. [Transl. Fish. Res. Board Can., Transl. Ser. 119.] NMFS (National Marine Fisheries Service). 1992. Our living oceans: report on the status of U.S. living marine resources, 1992. NOAA Tech. Memo. NMFS-F/SPO-2, 148 p. OCSEAP (Outer Continental Shelf Environmental Assessment Program). 1987. Marine Fisheries: resources and environ- ments. In D. W. Hood and S. T Zimmerman (eds.), The Gulf of Alaska — physical environment and bio- logical resources, p. 417-458. U.S. Government Printing Office, Washington, D.C. 20402. Rose, G.A. 1993. Cod spawning on a migration highway in the north-west Atlantic. Nature 366:458-461. Schumacher, J. D., and R. K. Reed. 1983. Interannual variability in the abiotic environ- ment of the Bering Sea and Gulf of Alaska. In W.S. Wooster (ed. ), From year to year, p. 111-133. Univ. Washington, Coll. Ocean Fish. Sci., Wash. Sea Grant Program, Sea Grant publ. WSG-WO 83-8. Shimada, A. M. 1985. A study of eastern Bering Sea Pacific cod: re- source assessment, fishery development and man- agement under variable recruitment. M.M.A. the- sis, Univ. Washington, Seattle, WA, 212 p. Svetovidov, A. N. 1948. Gadiformes. Acad. Nauk. SSSR, Zool. Inst., Fauna SSSR, Ryby 9(4), N.S. 34, 304 p. [Transl. Israel Program Sci. Transl., 1962. Available from U.S. Dep. Commer. Natl. Tech. Inf. Serv., Spring- field, VA, as OTS 63-11071.1 Teshima, K. 1985. Maturation of Pacific cod in the eastern Bering Sea. Bull. Jpn. Soc. Sci. Fish. 51(11:29-31. Thompson, G. G. 1994. Pacific cod. //* L. L. Low (ed.), Status of liv- ing marine resources off Alaska, 1993, p. 14-16. U.S. Dep. Commer., NOAA Tech. Memo. NMFS- AFSC-27, 110 p. Westrheim, S. J. 1984. Migrations of Pacific cod (Gadus macro- cephalus) in British Columbia and nearby waters. Bull. Int. N. Pac. Fish. Comm. 42:214-222. Wilimovsky, N. J., A. Peden, and J. Peppar. 1967. Systematics of six demersal fishes of the North Pacific Ocean. Fish. Res. Board Can. Tech. Rep. 34:1-95. Zhang, C. I. 1984. Pacific cod of South Korean waters. Bull. Int. N. Pac. Fish. Comm. 42:116-129. Abstract. — The chemistry of calcified tissues has been suggested as a source of useful information on the population structure and environmental histories of fishes. We evaluated this possibility by examining in detail regional and ontogenetic variability in the chemical composition of sagittae of juveniles and adults of the temper- ate marine groundfish Nemadac- tylus macropterus . Six elements (in order of decreasing abundance, Ca, Na, Sr, K, S, and CI) were consis- tently detected in the sagittae at concentrations greater than 200 ppm; all exhibited levels of indi- vidual, ontogenetic, and regional variability well in excess of their respective scales of measurement error. Comparisons of juveniles and adults from different sites indicate that composition of the otolith is most alike in fish from adjacent sites, that most juveniles are simi- lar to adults collected from the same site, and that the differences in composition that characterize sites are manifest during most, if not all, of the fish's ontogeny. These results are consistent with the hy- potheses that otolith composition reflects population structure and that these populations are largely self-recruiting. However, the re- sults also suggest that the chemi- cal composition of otoliths is much less sensitive to environmental con- ditions than previously thought. Rather, it appears that regional dif- ferences in composition either have a genetic basis or are set by envi- ronmental influences early in life and are then maintained throughout subsequent life history. An evaluation of electron-probe microanalysis of otoliths for stock delineation and identification of nursery areas in a southern temperate groundfish, Nemadactylus macropterus (Cheilodactylidae) Ronald E. Thresher Craig H. Proctor John S. Gunn CSIRO Division of Fisheries POBox 1538 Hobart, Tasmania Australia 7001 Ian R. Harrowfield CSIRO Division of Mineral Products ROBox 124 Port Melbourne, Victoria Australia 3207 Manuscript accepted 11 April 1994. Fishery Bulletin 92:817-840. Knowledge of geographic structure is fundamental to understanding the dynamics of marine fish popu- lations (e.g. Sinclair, 1987). None- theless, even for the small number of species thus far investigated, there remains considerable uncer- tainty regarding population struc- ture. This is due to the lack of a widely applicable, direct means of mapping how far and in what direc- tions larvae disperse. A variety of indirect techniques have been ap- plied to the problem, e.g. modelling larval advection from oceano- graphic features, analyzing parasite loads, mapping phenotypic charac- ters, and locating and enumerating discrete spawning areas. However, all are limited in scope and in the strength of the inferences that can be drawn from them. For this rea- son, the geographic structure of marine populations is usually in- ferred from genetic studies (e.g. Avise et al., 1987; Waples and Ros- enblatt, 1987; Smith et al., 1990). But genetic techniques are also far from ideal for this task: they will not detect differences in the face of even low levels of larval or adult mixing among populations (Hartl and Clark, 1989); they cannot directly measure rates of individual exchange among sites or, usually, specify the origin of individuals; and results from even a successful study can be am- biguous, i.e. a genetic difference between sites suggests little dis- persal, but the lack of any difference is largely uninformative. Such dif- ficulties have prompted continuing research into alternative, and per- haps more definitive, techniques for evaluating population structure. One alternative technique is the analysis of the chemical composi- tion of calcified structures. As early 817 818 Fishery Bulletin 92(4). 1994 as 1967 (Fisheries Agency of Japan, 1967), prelimi- nary studies suggested that the quantitative analy- sis of the microconstituents and trace elements in otoliths, vertebrae, and scales could provide infor- mation on population structure and the movements of individual fish. This suggestion was based on two assumptions and a hypothesis. The assumptions were that 1) the calcified structures are not suscep- tible to dissolution or resorption and 2) the growth of these tissues continues throughout life. If these assumptions are correct, calcified structures are per- manent records of the influence of endogenous and exogenous factors on their calcium-protein matrices. The hypothesis is that differences in the environ- ments to which fish in each population are exposed affect the incorporation of elements in calcified struc- tures, which results in chemical compositions spe- cific to each population. An extensive fisheries lit- erature supports the assumptions for otoliths, if not perhaps for scales and vertebrae (e.g. Sauer and Watabe, 1989). The working hypothesis also appears reasonable, given an extensive literature on inver- tebrates that relates differences in the composition of, for example, mollusc shells, and coral skeletons to a range of environmental and physiological condi- tions (Thompson and Livingston, 1970; Weber, 1973; Houck et al., 1977; Buchardt and Fritz, 1978; Smith et al., 1979; Rosenberg, 1980; Schneider and Smith, 1982). Since 1967, several studies have investigated whether the composition of calcified structures indi- cates stock or subpopulation identity in fishes (e.g. Klokov and Frolenko, 1970; Calaprice, 1971, 1985; Calaprice et al., 1971, 1975; Bagenal et al., 1973; Gauldie and Nathan, 1977; Behrens Yamada et al., 1987; Lapi and Mulligan, 1981; Mulligan et al., 1983, 1987; Edmonds et al., 1989; Calaprice1), using a va- riety of analytical techniques (see reviews by Coutant, 1990; Gunn et al., 1992). The results have been mixed. In part, this is because most techniques used required a relatively large amount of material for analysis. Otoliths or bones from many individu- als often had to be pooled to reach the minimum sample mass required. Because individual and onto- genetic variability could not be addressed, it has been difficult to assess the potential of the approach. In 1987, we began experiments with a view to us- ing fine-scale, ontogenetic variation in the composi- tion offish otoliths as an indicator of movement and migration patterns. The results of the first step — an investigation of the operating characteristics of probe microanalyzers as they affect data quality and the development of reliable techniques for 'life his- 1 Calaprice, J. R. 1983. X-ray fluorescence study of stock varia- tion in bluefin tuna. Status report submitted to NMFS, Miami, March 1983, 60 p. tory scans' across otoliths — are reported in Gunn et al. (1992) and Sie and Thresher (1992). In this pa- per, we evaluate the extent to which otolith composi- tion in a test species varies ontogenetically, among individuals within sites, and among sites, in order to assess whether such variation is of sufficient mag- nitude for, and contributes to, resolving population structure in the species. The species chosen for study was the jackass morwong, Nemadactylus macropterus (Cheilo- dactylidae), a moderate-sized (maximum about 70 cm standard length), bottom-dwelling fish common on the middle and outer continental shelf off south- ern Australia, New Zealand, South Africa, and the Pacific coast of South America (Robertson, 1978). The species was chosen for two reasons. First, the popu- lation structure of the species in Australian waters is contentious. On the one hand, regional declines in catch rates suggest localized stocks, which is consis- tent with work in New Zealand, where the species has three geographically discrete populations (Gauldie and Nathan, 1977; Robertson, 1978) . On the other hand, a small amount of tagging data for adults (Smith, 1989), allozyme data for specimens collected in southeast Australia (Richardson, 1982), and recent allozyme and mitochondrial DNA analy- ses for the entire Australian range (Elliott and Ward, 1994; Grewe et al., in press) suggest a single, broadly distributed population (Smith, 1989; Tilzey et al., 1990). This interpretation also appears to be consis- tent with the early life history of the species: N. macropterus spawns along the middle continental shelf, has a planktonic duration of 9-12 months, and has a morphologically specialized late-stage larva ('paper fish') that is neustonic and generally caught offshore of the continental shelf (Vooren, 1972). This combination is taken to imply high rates of local mix- ing during the larval stage. Nemadactylus macropterus was also chosen be- cause of uncertainty about the location of its nurs- ery areas in Australia. To date, the only place in Aus- tralia where large numbers of juveniles have been found is the shallow bays and inlets of southeastern Tasmania. As a result, it has been suggested that this area is a critical habitat supporting the entire Australian population (Tilzey et al., 1990). Given continuing coastal development in this area, if the hypothesis is correct, conservation measures need to be developed and implemented to ensure the contin- ued viability of the fishery. Analysis of otolith composition could help resolve both questions. With regard to population structure, we hypothesized that if there are discrete spawning populations in Australian waters, then the composi- tion of the central, first-forming portion of the otolith Thresher et al.: Otolith analysis of Nemadactylus macropterus 819 would differ geographically. With regard to the num- ber and location of nursery areas, we further hypoth- esized that if there is only one nursery area, then the composition of that part of the otolith deposited during residence in the nursery ground would be similar for all adults, irrespective of where they were caught, and would match that of juveniles caught in southeastern Tasmania. Methods Collection details for juvenile and adult N. macro- pterus are provided in Figure 1 and Table 1. Recently settled (0+) juveniles were collected by hand-lining and trawling at six sites off Tasmania and southern Victoria. Adult specimens were also obtained at six sites, from commercial and scientific trawls on the southeastern Australian continental shelf. To mini- mize possible effects of interannual variation in otolith composition, we minimized the number of year classes in the sample by using only adults in the size range of 30-35 cm fork length. Otolith macrostruc- ture and published length-at-age keys for the spe- cies (Smith, 1982) suggest that the specimens were a mixture of the 1980 to 1984 year classes and that year- class distributions overlapped broadly among sites. The juveniles were from the 1987 and 1988 year classes. All specimens were frozen at -20°C shortly after collection and remained frozen (up to 30 days) until the otoliths were removed. After extraction, each otolith was cleaned of adhering tissue with fine for- ceps and a soft-bristled brush in millepore-filtered distilled water. They were then dried in an oven at 40- 45°C for at least 6 hours, after which they were stored in polyurethane capsules in a desiccating cabinet. Procedures for embedding, sectioning, and prepar- ing otoliths for probe microanalysis are detailed in Gunn et al. (1992). Only sagittae were used in this study, because of their larger size. Prior to embed- ding, a scaled diagram of the distal surface of each otolith was made in order to guide subsequent sec- tioning. The otolith was then fixed upright on its ventral edge to the base of an embedding mold with a drop of Araldite. The mold was then filled with a harder-setting resin. After hardening, the otolith was sectioned with a diamond-edged saw blade (350 urn thick) on a rotary saw. Grinding to the plane of the primordium was done by hand with 2400-grade sili- con carbide wet/dry paper. Final polishing was done by using progressively finer grades of diamond paste (6-3 pm) and aluminum oxide powder (Linde B) on a lapping machine. After polishing, the section was ultrasonically cleaned and stored in a moisture-free environment. Prior to probe microanalysis, the sec- tion was heated on a hot-plate at 80°C for 10 min- Adults W. Tas. NSW E. Tas 39 45 Juveniles t v Phillip Island 0 % fi. r TASMANIA 7 "* -v 1* X Vp»_A. Maria jipAV Island West Tas X 1** -J y Nutgrove/ Cygnet Derwenl 143° 149° Figure 1 Source locations of samples of adult and juvenile Nemadactylus macropterus examined by elec- tron probe microanalysis. Sample details are provided in Table 1. Adult sites: GAB=Great Aus- tralian Bight, W. Vict.=western Victoria; E. Vict.=eastern Victoria; W. Tas.=western Tasmania; E. Tas.=eastern Tasmania; and NSW=New South Wales. Nutgrove and Derwent are two in- shore sites sampled for juveniles, both located in the Storm Bay (SE Tasmanian) estuary. Phillip Island is an inshore site off Victoria. 820 Fishery Bulletin 92(4), 1994 utes to remove any residual moisture, coated with a 250-300 A (measured by color on brass) coat of car- bon with a sputter coater, and then stored under vacuum until insertion into the probe. The procedures used to analyze otolith composi- tion are detailed in Gunn et al. (1992). Damage to the specimens under the electron beam is inevitable. The amount of damage, and hence quality of the data, is proportional to beam-power density (i.e. beam cur- rent x accelerating voltage/ target area). In Gunn et al. (1992), we concluded that beam power densities greater than 3.0 pW pm~2 resulted in unacceptable levels of specimen damage, data precision, and ac- curacy. Hence, data for the current study were ac- quired by using the following beam conditions: 25 nA current, 15 kV accelerating voltage, a 14 pm di- ameter 'defocused' beam (and hence a beam power density of 2.44 pW pnr2' and a total acquisition time of 3 minutes, 42 seconds per point. Comparisons of parallel scan lines (see Fig. 6) included some analy- ses at a 6 pm beam diameter, 5.5 nA current, and 15 kV accelerating voltage; despite the small beam di- ameter, beam power density for this series (2.92 pW pm-2) is within our 'safe' limit. The electron probe microanalyzer used was a Cameca Camebax fitted with three wave-length dispersive detectors. The concentrations (weight-fractions) of Na (sodium), K (potassium), Ca (calcium), S (sulphur), and CI (chlo- rine) were calculated based on the count rates mea- sured for their respective Ka lines, and for Sr (stron- Table 1 Collection details for Nemadactylus macropterus adults and juveniles analyzed in this study. Date Collection Sample Size range Location collected method size (cm FL) Adults NSW 12 Jan 90 Trawl 11 36.0-38.0 (Lat.34°40'S Long.l51°10'E) W. Victoria 30 Aug 87 Trawl 6 34.6-37.1 10% (105 ppm) by weight; Na, Sr, K, S, and CI constitute the 'micro-constituents,' which occur in mean con- Figure 2 Diagram of section of a sagitta of an adult Nemadactylus macro- pterus, as prepared for electron probe microanalysis, depicting pro- grammed transects used to analyze points along the main growth axis from the primordium (p) to the otolith's posterior edge (pe). 822 Fishery Bulletin 92(4), 1994 Table 2 Minimum detection limits (MDL), mean and ranges of estimated concentrations, and measurement error for Nemadactylus macropterus, of the six elements that could be assayed reliably with a WD electron probe. The values are based on a random subset of our data (n = 478 points), including numerous individuals and positions along the scanned axis of points analyzed. Concentrations are given in ppm (by weight), except for Ca which is in percent of the target mass. Note that values below the minimum detection limit are effectively zero. 'Minimum significant differ- ence' is based on comparison of 'replicate' points in parallel life history scans (see text). CI = Confidence interval. Element MDL Minimum significant difference Mean (range) concentrations Measurement error (absolute, %) Mean (range) 99% CI Ca Sr Na K S CI 311 159 136 149 157 38.8% (35.3-44.5) 2240(1430-3860) 3331 (2680-4240) 729 (280-1630) 421 (220-1220) 255 (0-1230) ±157(7%) ±122(3.7%) ± 72 ( 10%) ± 76(18%) ± 72(28%) 210(12-964) 160(10-450) 77(0-320) 80 (4-247) 48(10-210) 331 235 118 121 73 10 g> CD s -Q 10 E Q. a. c o I 10 c 0) o c o o 0 1 Ca Sr S Hg Se Cd Mn Ni Br Na K CI Ba Cu Fe Zn Pb Figure 3 Mean (solid circle) and ranges (vertical line) of concen- trations of the elements detected in the sagittae of Nemadactylus macropterus by means of electron probe (for Ca, Na, Sr, K, S, and CI) and proton probe (for mercury (Hg), barium (Ba), selenium (Se), copper (Cu), cadmium (Cd), iron (Fe), manganese (Mn), zinc (Zn), nickel (Ni), lead (Pb), and bromine (Br)) microanaly- ses. For proton probe methodology and results, see Sie and Thresher (1992). Data are a compilation of >500 points across numerous individuals and positions along the growth axis. The minimum detection limit (MDL) for each element is indicated by the irregular horizon- tal line and is based on the standard output of the re- spective probe microanalyzers at our standard operat- ing conditions (see Gunnetal., 1992; Sie and Thresher, 1992). The minimum concentration of CI is below the detection limit of the electron-probe microanalyzer but could not be determined with the more sensitive pro- ton-probe microanalyzer because of our operating con- ditions (see Sie and Thresher, 1992). centrations of 100-5000 ppm; and a variety of 'trace elements' (e.g. iron, copper, and bromine) occur at concentrations <10 ppm. Only the micro-constituents and Ca can be measured accurately by WD-EPMA. Absolute ranges of concentrations, measurement er- ror (absolute and 95% confidence intervals), and minimum detection limits (MDL's) for each of these elements are given in Table 2. Measurement error is inversely correlated with mean concentration, rang- ing from 3.7% in Na to 28% in CI. Of the six ele- ments measured, only CI occurred occasionally at less than its respective MDL (157 ppm). Although the microanalyzer reports values less than the MDL, these values were considered noise and set equal to zero for analyses of population structure. Life history scans for three fish chosen randomly from the data set (Fig. 4) illustrate several points typical of our data. First, all six elements vary onto- genetically in concentration well in excess of the uncertainty associated with measurement. Second, concentrations at any given position correlate strongly with those at neighboring points; for all six elements, autocorrelations are highly significant at scales <100 urn (e.g. Fig. 5), which suggests that in N. macropterus this is the typical scale of ontoge- netic variability in composition. Comparison of 'small spot-closely spaced' analyses (6 urn beam diameter spaced at 8 urn intervals) with 'large spot-widely spaced' analyses ( 14 urn diameter at 16 urn spacing) suggests that a sampling scale finer than our stan- dard analysis reveals few, if any, major variations in otolith composition that would not be detected at the coarser sampling scale (Fig. 6). Third, absolute vari- ability is highest for Sr, which can vary within speci- mens over half an order of magnitude. However, rela- tive variability is as high in S and CI; coefficients of Thresher et at.: Otolith analysis of Nemadactylus macropterus 823 #278 #366 #384 431 40 37 60001 ___ 350' Q. c o ra 1000 o U 1000 500 80 160 0 Point Number (25um steps) Figure 4 Cross-otolith (primordium to posterior margin) variation in the measured concentrations of the six elements (Ca, Na, Sr, K, S, and CD that could be reliably measured by WD electron-probe microanalysis in Nemadactylus macropterus sagittae. Concentrations are reported in units of ppm by weight for all elements other than Ca, which is reported as percent by weight. Data are adjusted to standards and cleaned of effects of surface pitting and cracks but are otherwise unfiltered. Specimens were chosen randomly from the adult data set and appear to be represen- tative of the variation observed. Figure 5 Auto-correlation analysis of the unfiltered Sr data for the three adult Nemadactylus macropterus of Figure 3, at lags ( intervals between points correlated ) of 0 to 10 points. The correlations become insignificant (P>0.05) at lags of 4 to 6 points, varying slightly among specimens. Points are spaced 25 um center-to-center. 824 Fishery Bulletin 92(4), 1994 £: 3300 o O 10 Point number Figure 6 Comparison of parallel fine-scale (open circle) and coarse-scale (solid circle) transects of the outer margin of an adult Nemadactylus macropterus sagitta. Distance between the parallel scans was approximately 20 urn. Fine-scale analysis was done with a 6-um diameter beam at 8-um intervals, center to center, and with a beam-power density of 2.92 uW unr2; coarse-scale analysis was done with a 14-um beam at 16-um intervals and with a beam-power den- sity of 2.44 uW urn-2. Concentrations are reported in units of ppm by weight for all elements other than Ca, which is reported as percent by weight. The large disparity between the two analyses in Na and Ca concentrations is an expected consequence of the different beam-power densities used in the two runs (see Gunn et al., 1992). variation in the three specimens depicted range up to 45.4 for Sr, 43.4 for CI, and 31.3 for S (as opposed to, at the other extreme, 1.8 for Ca). Fourth, ontoge- netic patterns in the variation are often consistent across specimens. All N. macropterus that we have analyzed, for example, show steep gradients in Sr levels in the region immediately around the primor- dium. Similarly consistent, though less pronounced, patterns are evident in Na and Ca. The quality of these data were assessed by com- paring life history scans from the left and right otoliths from the same fish. The quality of the match within each otolith pair differs markedly (Fig. 7). The comparisons suggest two principal sources of error. First, there is consistent evidence of the difficulty of tracking identical growth trajectories even within a pair of otoliths from the same individual. In all three pairs, the match between left and right otoliths de- teriorates as the otolith margin is approached. We attribute this to the decline in the growth rate of the otolith with age, a corresponding compression of on- togenetic variability and, therefore, a larger effect of errors in tracking through the growth axis on the apparent ontogenetic pattern of composition. Slight differences in the shape of the otoliths also give rise to differences in the length of each section, and hence the spacing of scan points relative to the distance along the growth axis. Most of the left and right dif- ferences in specimens #304 and #312 appear to re- sult from these tracking errors; that is, the same ontogenetic patterns and mean concentrations are generally evident but variously expanded or com- pressed along the growth axis. Second, in four of the six elements examined (Na, K, S, and CD, mean con- centrations occasionally differ between left and right otoliths over relatively large portions of the otoliths. This second source of error is difficult to assess. The mismatch is most evident for CI in #339 and S in #312, where the scale of the mismatch greatly ex- ceeds machine-induced measurement error. The pat- tern of the mismatch varies widely and inconsistently among the samples: for example, CI levels match well in #304, match intermittently and poorly in #312, and differ markedly near the margin of #339, whereas S matches well in #304 and #339 but very poorly in #312. Comparisons of parallel life history scans across a single otolith (e.g. Fig. 6) suggest that dif- ferences between otolith pairs of the magnitude ob- served cannot easily be attributed to either measure- ment error or slight differences between otoliths in the position of the scan line relative to the main growth axis. We conclude, therefore, that the differ- Thresher et al.: Otolith analysis of Nemadactylus macropterus 825 #304 #312 #339 6000 E Q. C o c Q) o c o O 140 0 140 Point number Figure 7 Comparisons of life history transects of the left and right sagittae of three adult Nemadactylus macropterus based on unfiltered data for four representative elements (Sr, Na, S, and CD. The dashed horizontal line in the CI plots is the minimum detection limit for the element. The comparison is based on identical analyti- cal conditions and an attempt to track as closely as possible the same growth axis in the two otoliths in each pair. ences between otolith pairs are real, appear to be more common in some elements than in others (e.g. evident in CI, but not in Sr), and may be more com- mon near the otolith margin than closer to the pri- mordium. The data are limited but clearly indicate that slight differences in elemental concentrations, particularly CI and S, should be used with caution for stock delineation. Evaluation of stock structure We hypothesized that stocks of N. macropterus would differ in either or both spawning grounds or times and that these differences would result in diagnos- tic patterns of composition in the first forming part of the otolith. However, in practice, compositional data for the primordium itself proved of low quality because of the specimen preparation required for EPMA and the incremental structure of otoliths. In most specimens, minute cracks or a pit several mi- crons in diameter developed at the primordium dur- ing preparation, the latter because of the 'plucking' of the primordium from the otolith center during polishing. Our previous work (Gunn et al., 1992) in- dicated that topographic irregularities degrade EPMA data because of unpredictable patterns of x- ray absorption. A comparison of data for point 1 (on and immediately around the primordium) and point 2 (25 urn from the primordium) for the adults sup- ports this conclusion: for all elements except CI the variance in estimated concentrations is 36 (Na) to 270 % (S) higher for point 1 than for point 2 (Fig. 8). Nevertheless, for all elements, concentrations at point 1 are significantly correlated with those at point 826 Fishery Bulletin 92(4), 1994 4000 2000 C\j 2000 c o Q_ 42 32 32 .*#, '<}*' Na r2= 0.21 4000 Sr r2= 0.33 * • • * . • * .,• .1 ... ' • mnn • 1000 r= 0.09 4000 1 000 ,1000, 200' 4000 200 1000 Ca r2= 0.18 r = 0.25 3001 42 300 1200 1000 1200 Point 1 Figure 8 Scattergrams of the relationship between concentrations of the six detected elements (Ca, Na, Sr, K, S, and CI) as measured on the otolith primordium (point 1) and as measured at a point 25 pm from the primordium (point 2) along the long growth axis of 64 adult Nemadactylus macropterus. Concentrations are reported in units of ppm by weight for all elements other than Ca, which is reported as percent by weight. Correlations between the two estimated concentrations are significant for all elements. The slope of the regressions differ significantly from 1.0 for all elements except CI, although for several elements (Na, K, and Ca) the trend line is very close to a slope of 1 and the calculated regression shifted mainly because of relatively few very low values at point 1. The latter are expected conse- quences of pitting and cracking on the primordium. Dashed lines in the CI scattergram indicate minimum detection limits. 2 (r2 values from 0.09 (S) to 0.85 (CD). Slopes for re- gressions between point 1 and point 2 data are typi- cally not equal to 1, which for most elements prob- ably reflects the highly proteinaceous nature of the primordium. Nonetheless, we conclude that differ- ences among specimens evident in the primordium are also evident adjacent to the primordium, where they can be measured more precisely. The distributions of mean concentrations of the six elements for the 68 adult N. macropterus analyzed are depicted in Figure 9. Three of the six (K, S, and CD are significantly skewed to higher concentrations. Ca also shows evidence of a weak skew, and Sr evi- dence of a weak bimodality. Mean concentrations of four of the six elements differ significantly among sites (Fig. 10), the exceptions being Sr and S. The differences are manifest in both the point 2 and point 2-6 data and are of similar pattern and comparable magnitude in both data sets. For most elements, er- ror bars are smaller in the filtered data, which pre- sumably reflects the reduced effect of random mea- surement errors. Differences among sites are great- est for CI: the mean values for three sites (eastern and western Tasmania and the Great Australian Bight) do not differ significantly from the minimum detection limit (MDL), whereas means for the Victo- rian and New South Wales (NSW) samples are well above the MDL and do not differ significantly from each other. There are suggestions of a similar, though less pronounced grouping of sites in Na and K (con- centrations in the Victorian and NSW fish higher than in those from Tasmania and the Bight) and Ca ( lower concentrations in Victorian and NSW adults ). The grouping of sites was examined further by plot- ting Na/Cl and Sr/Ca ratios for specimens from each of the six sites (Fig. 11 ). These ratios were chosen on the basis of a preliminary survey of the data as likely to separate sites. As expected, the scatter of points Thresher et al.: Otolith analysis of Nemadactylus macropterus 827 J[ h Na x = 3300 PN>0.5 20 Sr x = 2230 PN>01 15 ■ I K x = 730 p < 0.001 30 2700 4200 1600 3100 55 Ca \f\ R">°'1 / \ / ^ JO A s x = 450 p <0.01 25 _ Cl x = 360 p < 0.001 c 3 cr 0 ? 1900 O 3 Q. 20 | Cffl^ c BI 38% 46% 300 1050 0 Concentration 1500 Figure 9 Frequency distributions of concentrations of the six detected elements (Ca, Na, Sr, K, S, and Cl) at point 2 for 64 adult Nemadactylus macropterus pooled across sites in SE Australia, /^probability the distribution is normal. Concentrations are reported in units of ppm by weight for all elements other than Ca, which is reported as percent by weight. for individual fish is greater for the single-point data than for the filtered data, but the patterns of regional groupings and the relationships between the concen- trations of elements are much the same for the two data sets. Victorian and NSW samples appear to group based on both Na/Cl and Sr/Ca ratios. The two Tasmanian samples group with specimens from the Bight on the basis of Na/Cl ratios but appear to dif- fer based on mean Sr/Ca ratios. Linear discriminant function analysis (LDFA), for a 3-group discrimination, produces similar results for both single-point and filtered data. For both data sets, the three groups are statistically separable but at a relatively low rate of successful classifications. For the single-point data, 66% of individuals were accurately classified into their three respective 'source populations'; for the filtered data, the suc- cess rate increases to 78%. The relatively poor sepa- ration is due, in part, to an overlap of the three groups in discriminant space and, in part, to a few individu- als located in discriminant space well outside the areas defined by most individuals collected at the same place and time. For the filtered data, discriminant analysis devel- oped five discriminant functions to classify the six sites. However, only the first three are significant (P<0.05), and of these there is a large difference be- tween the first two functions (both at P<0.01) and the third (P=0.02). Examination of the canonical load- ings indicates that discriminant function 1 is corre- lated with Na, K, and Cl concentrations, and hence represents mainly the initial separation of sites along the Na/Cl axis indicated in Figure 11. The second discriminant function loads heavily only on Sr, whereas the third is mainly a K residual from the first discriminant function. Step-down procedures, in which sites are sequentially pooled, raises the con- tribution of Ca to the second discriminant function, identifying it with the Sr/Ca axis in Figure 11. The nature of the site separations is indicated in Figure 13. Function 1 separates the two Tasmanian and the Bight ( GAB ) samples from the two Victorian and the NSW samples; function 2 distinguishes weakly be- tween the GAB sample and the remainder; and func- tion 3 separates the east and west coast Tasmanian samples. The remaining two functions do not clearly distinguish among any sites. The primacy of the first three functions remains in a step-down procedure, as the sites are sequentially pooled based on their degree of overlap. The final step, at which all func- 828 Fishery Bulletin 92(4), 1994 Point 2 Points 2 - 6 3900 Na 3000 3000 Sr 1800 900 H =21 72 P< 0.001 H = 23 91 P< 0.001 H = 4 82 N.S H = 9 42 P = 0 05 c o C-> C O o 500 ■1 I Ca 38 560 H = P< 16 30 D.005 I H= 17.01 P < 0.005 I I H = 1 1 .89 P<005 H = ! 6 13 MS I 160 H= 1 45 NS. H = 1 60 N.S CI H = 40.07 P <0 001 I \ H P = 42.10 < 0 001 I } ETas WTas VICT NSW GAB ETas WTas VICT NSW GAB Source location of adults Figure 1 0 Differences among the six adult Nemadactylus macropterus sam- pling sites in the estimated concentrations (means and standard errors) of the six elements detected by EPMA (Ca, Na, Sr, K, S, and CI), based on point 2 only and on the mean of points 2-6 inclu- sive for each individual examined. Concentrations are reported in units of ppm by weight for all elements other than Ca, which is reported as percent by weight. The two Victorian samples were pooled for this comparison. Site labels are defined in Figure 1 and Table 1. Thresher et al.: Otolith analysis of Nemadactylus macropterus 829 tions contribute significantly (P<0.001) to the dis- crimination, is at the level of three groups and two discriminant functions (i.e. Fig. 12); the third func- tion, separating the two Tasmanian samples, is not quite significant in the final step (P=0.057). Post-hoc analyses (Steffe's P-test) of the discriminant func- tions indicate that the samples from the Victorian and NSW sites do not differ significantly and consis- tently in any of the three functions, the GAB sample differs from all other sites (which do not differ sig- nificantly) in function 2, and none of the sites differs significantly in function 3, though the western Tas- manian sample nearly differs significantly from the other sites. We draw three general conclusions from these analyses. First, there are significant differences among samples from different sites in terms of the composition of the primordial region of their otoliths. Second, analyses of the primordium itself, of a point 25 urn from the primordium, and of the mean value for the region between 25 and 125 urn from the primordium produce similar results, indicating that distributional differences of adults are manifest through at least the first 125 urn of otolith growth. And third, on the basis of common patterns of composition, the sites pool into three groups: one com- posed of the NSW and the two Victo- rian samples, a second consisting uniquely of the Bight sample, and a third consisting of the two Tasmanian samples. In both cases where sites are pooled, the pooled sites are geographi- cally contiguous and nearest neighbors. Site-specific differences and similarities in ontogenetic variation in composition That differences in composition among sites can be discerned at points as far out as 125 urn from the primordium and to an apparent age of 45-55 days post- hatching suggest that delineation among samples is not a function of con- ditions specific to the spawning sites. Any environmental differences must encompass at least several weeks, per- haps months, of larval development. To assess the ontogenetic patterning of these chemical differences, we com- pared the concentrations of apparent key elements for specimens from the three pooled areas (NSW/Victoria, the ci Ca 381 1400 Bight, and both Tasmanian sites) at several points along their respective long growth axes. Five-point filtered data were assessed at four positions: points 2-6, 6-10, 36^0, and 80-84. The first position is immediately adjacent to the primordium; the second immediately exterior to the first (and presumably encompassing the second 2—3 months of planktonic larval development); the third we estimate to corre- spond approximately to the age when the prejuven- iles recruit to the nursery areas; and the fourth, out- ermost position, is the farthest along the growth axis at which we had data for all specimens (the number of points depended upon the length of the axis) and, we estimate, corresponds to otolith deposition at an age of 2-3 years. The results of the comparison (Fig. 14) lead to three conclusions. First, the mean pattern of ontogenetic change in composition is very similar for samples from all three pooled sets of sites, e.g. Na and Sr concentrations decline between points 2 and 6 and Pt 2 only Pts. 2-6 : . ■° '.' CI Na .o • S * s . 9 *° Sr 3400 14°0 Figure 1 1 Scattergrams of the relationships between Na and CI and between Sr and Ca concentrations among the 64 adult Nemadactylus macropterus examined, grouped by site, based on concentrations measured at point 2 only and the mean concentration for points 2-6, inclusive. Concen- trations are reported in units of ppm by weight for all elements other than Ca, which is reported as percent by weight. Sites are described in Figure 1 and Table 1. Site key: E. Tas.=solid square; W. Tas.=solid circle; NSW=open square; E. Vict.=open circle; GAB= plus sign; and W. Vict.=open diamond. 830 Fishery Bulletin 92(4), 1994 then increase towards the otolith margin in samples from all three sites. Second, the pattern of ontoge- netic variation in concentrations differs among ele- ments. And third, similar mean differences are evi- dent among sites irrespective of where the analysis was done in the otolith. NSW/Victorian specimens, for example, at all stages of their life histories to an apparent age of at least 2-3 years tend to have CI levels higher than those of fish collected elsewhere. As a result, discriminant analyses based on mean concentrations at points 36 and 80 result in site de- lineations virtually identical to those derived from concentrations measured near the primordium. Evaluation of signatures specific to nursery areas and the links between nursery areas and adult groups The links between nursery areas and spatial compo- nents of the adult populations can be assessed in two 0 Pt. 2 only ^a^^ NSW/Vict O Oj o / Tas ~~-*S GAB g c 0 Pts. 2-6 SO ° c & ° V "\NSW/Vict D — . c ^v o o o) ^"-^ — * 0 o ~^"~^ Distrib sites d detern functio concen detect arounc among analys lambd; lambd. 4 0 Function 2 Figure 1 2 ution of adults from each of the six samplei >ee Fig. 11 for key) in two-function space a lined by a three-group linear discriminan n analysis based on point 2 data and the meat trations for points 2-6 inclusive of the si. ;d elements for each individual. Outline each group were drawn by eye. Difference groups are significant at P<0.001 in botl >es (for point 2 only, Fl2 120=6.21, Wilk' »=0.35; for filtered data, F12 i20=8.72, Wilk' i=0.29i. I 1 3 t 1 { 3 3 1 3 3 complementary ways: 1) by determining the source affinities (e.g. spawning site) of juveniles collected in each nursery area and 2) by developing a specific P«0001 • • • • • i • • 1 1 i ! I ! i • i 1 > i • P«0.001 P<0.05 NS NS *> **> $u ^ ty&, Q.}/, Source location Figure 13 Distribution among sites (see Fig. 1) of values of each of the five discriminant functions defined by the initial (six-site) discriminant function analysis for adult Nemadactylus macropterus, based on mean concentrations for points 2-6, in- clusive. Horizontal bars indicate sites that pool together based on post-hoc analysis (Steffe's F- test). In the case of Function 3, the overall ANOVA is just significant at P<0.05, but pair- wise comparisons among sites indicate no single or set of sites that differs consistently and sig- nificantly from the others. NS=not significant. Thresher et al.: Otolith analysis of Nemadactylus macropterus 831 signature for each nursery area (based on otolith material deposited during residence in the nursery areas) and using these to classify adults collected in different regions. In essence, the former assesses how juveniles from each of the putative populations are distributed among nursery areas, whereas the lat- ter assesses the contribution of each nursery area to adults collected at each site. With specific regard to N. macropterus, if SE Tasmania is the sole nursery area for the species in Australia, then we would ex- pect that 1) the complete range of chemical patterns documented in the adults, at all sites sampled, would be seen in juveniles collected in the single, common nursery area, and 2) adults collected at all sample sites would have a nursery area 'fingerprint' similar to that of the juveniles collected. TAS. NSW/VICT. GAB 3600 2800 3000 Na c o c Q) U C O O 1600 42 (%) 34 800 Sr Ca CI U 2 6 36 80 2 6 36 80 2 6 36 80 Point analyzed For the first analysis, the three-site discriminant functions developed from the adults were used as a training set to classify each of 116 recently settled N. macropterus. The data for the juveniles were ac- quired in the same way as for the adults. Analysis is based on the mean values for points 2-6 from the primordium. Most juveniles were collected in SE Tasmania; a small number were also collected at Phillip Island, Victoria (Table 1). Most juveniles examined fell within or close to the areas in discriminant function space defined by the adult groups (Fig. 15); only one, with an exception- ally high value on the function 2 axis, did not match the characteristics of at least one of the three adult groups. Moreover, most juveniles classified with the adults collected in the same area. Of the 106 juveniles caught in Tasmania, all but 25 classified with the Tas- manian adult samples, and of these, 13 classified am- biguously, with a probability >25% of being Tasmanian. Overall, only 7% of the Tasmanian-collected juveniles had a probability of <10% of classifying with the Tas- manian-caught adults (Fig. 16). Samples from the five Tasmanian sites were distributed similarly in two-func- tion space (Fig. 15), though the variance was conspicu- ously higher at one site (Cygnet). The pattern was similar for juveniles collected off Victoria (Phillip Island), although sample sizes were too small to draw strong inferences. Of the 10 indi- viduals examined, six classified with the NSW/Vic- torian adults, three classified with the Tasmanian- caught adults (at probabilities ranging from 72 to 85%), and one classified with the Bight-caught adults (atP=63%). The probability that the Victorian juve- niles classify with the NSW/Victorian adults is mark- edly bimodal (Fig. 16) with peaks at >95% and be- tween 5-10%. That is, most individuals had either a very high or very low probability of classifying with the local adults. A similar pattern may also be the Figure 14 Comparisons of mean concentrations of four elements (Ca, Na, Sr, and CI ) at four points along the life history scan of each adult Nemadactylus macropterus pooled by groups identified by linear discriminant function analysis ( LDFA). For each individual and point, data were calculated as the mean of the 5-point moving average, beginning at the point indicated (i.e. 2=mean of points 2-6, inclusive; 6=mean of points 6-10, inclusive, etc.). K and S are not depicted as the former generally varies similarly to Na, whereas the latter did not differ significantly among groups. Horizon- tal dashed line in CI plot indicates minimum detection limit. Vertical lines about each mean indicate one stan- dard error of the mean. Concentrations are reported in units of ppm by weight for all elements other than Ca, which is reported as percent by weight. 832 Fishery Bulletin 92(4), 1994 Maria Isl. \^_" ^^Vt, 1 . J ^C Function 2 Figure 1 5 Distribution of individual juvenile Nemadactylus macropterus from each of the six sampling sites (see Fig. 1 and Table 1) in two-func- tion space as defined by the three-group LDFAof the adults, based on mean concentrations of the six detected elements at points 2-6, inclusive. Circles enclose areas in two-function space defined by the three adult groups (see Fig. 12). case among the Tasmanian-caught juveniles. There is an indication of modes at either end of the probability spectrum and possibly a third mode centered near 35%. The second analysis of the link between nursery areas and the adult population requires analysis of that portion of the otolith deposited while the indi- viduals were in the nursery areas. The otolith of the smallest juvenile we found had a longest growth axis (posterior to primordium) 500 urn in length; several other small fish had similar growth axes in the range of 570-650 urn. Therefore, we examined a standard region approximately 600-800 urn posterior to the primordium along the main growth axis as otolith deposited early in the nursery area stage of develop- ment. Specifically, we used as the datum of interest for Figure 16 Distribution of probabilities that juvenile Nemadactylus macropterus caught in each of the two regions sampled < Tasmania and Victoria I group with adult samples collected in the same regions as defined by the three group LDFA of the adults. 35 20 IS n E Tasmania Victoria 0.5 Probability juveniles group with adults from region where caught Thresher et al.: Otolith analysis of Nemadactylus macropterus 833 each specimen the mean composition of points 35—39, inclusive, in a standard life history scan (680 to 780 urn from the pri- mordium). All juveniles used in these analy- ses had otoliths at least this large. Discriminant analysis of the juveniles indicated highly significant differences among all six areas sampled. The weak- est discriminator (the fifth root of the dis- criminant analysis) was significant at P<0.01. The preliminary conclusion then is that there are signatures specific to each nursery-area that could be sought in the adult population. Further analyses of the data, however, indicated this conclusion was premature. Specifically, if there are nursery-area-spe- cific environmental signals in the otolith, then we would expect them to be manifest ontogenetically in either or both of two ways. First, we would expect that at the end of the larval period (approximately points 25-30), the mean concentrations of various elements would diverge among sites, reflecting the specific environment at each (i.e. the nursery area 'fingerprint'). Second, we would also expect that among individuals, concentrations of these same elements would converge within sites, re- flecting recruitment into a common envi- ronment. Again, this convergence should occur at approximately points 25—30. For the second prediction, we analyzed in detail one site (Cygnet) for which the sample size of juveniles was large enough that we could reduce possible variability due to differences in date of recruitment. This was done by examining juveniles caught on the same day and falling within a narrow size range (7-11 cm SL). For most elements (all but Sr), neither prediction is supported by the data (Figs. 17 and 18). Although variance is high at all points, there is little or no indication of either divergence among sites (in the case of mean concentrations) or convergence among individuals (in the case of variation within the single site) at or near points 25-30 for any element other than Sr. For Sr, however, both predictions appear to be borne out. Mean concentrations overlap broadly among juveniles from all sites during the larval stage but diverge significantly among sites at about point 25. Among individuals, juveniles at Cygnet appear to converge on two different postrecruitment Sr tra- jectories, also beginning at about point 25. The avail- able evidence suggests that concentrations of ele- ments other than Sr are largely unaffected by the Derwent « Nutgrove ° Maria l.«W. Tas. 'Phillip l.°Cygnet Point number Figure 1 7 Group mean concentrations of each of the six detected elements (Ca, Na, Sr, K, S, and CI) for the interior-most 40 points analyzed along the long growth axis of sagittae of juvenile Nemadactylus macropterus caught at each of the six nursery areas sampled (see Fig. 1 and Table 1). Point 1 is on the primordium. Data for each individual were filtered through a five-point moving average be- fore the group mean was calculated. transition from the larval to the juvenile stages and that the apparent discrimination among nursery- areas is the manifestation of differences among in- dividuals already evident in their larval stages. We tested this conclusion by reanalyzing for "nurs- ery-area-specific signals" using data for points 2-6 (early larval life) rather than for points 35-39 (early juvenile stage). In general, the results were similar to those obtained with points 35-39, with good dis- crimination among most nursery grounds and a com- parable level of overall site separation (Wilk's lam- bda=0.20 for points 2-6 vs. 0.22 for points 35-39, P<0.001 in both cases). However, the accuracy of cor- rectly assigning juveniles to nursery areas was less in the point 2-6 analysis (51% vs. 82%), which re- flects the divergence of Sr concentrations in the nurs- ery areas and its increased importance as a discrimi- 834 Fishery Bulletin 92(4). 1994 Point number Figure 18 Individual life history trajectories (interior-most 40 points) for each of the six elements (Ca, Na, Sr, K, S, and CI) detected in the sagittae of juvenile Nemadactylus macropterus collected at Cygnet (SE Tasmania). To minimize potentially confounding effects of varia- tion in date of settlement, the comparison was restricted to juve- niles collected on the same day (19 December 1987) and in the same size range (7-11 cm standard length). The data for each in- dividual have been filtered through a five-point moving average. nator. Two discriminant functions (2 and 3) load onto Sr at r>0.5 in the original analysis, whereas Sr does not achieve this load level for any function in the point 2-6 analysis. Reflecting this, mean Sr concen- trations differ among nursery areas at P<0.001 (F5no=6.21) in the original analysis, but at only P<6.02 CF5iU0=2.82) for the point 2-6 analysis. By comparison, differences among nursery areas for the other five elements are significant at similar levels for the two analyses. Discussion Effects of data quality on stock delineation Electron-probe microanalysis with WD-spectrom- eters revealed extensive variability in the concen- trations of six elements in N. macropterus otoliths. Some of this variability is induced by the inherent, small-scale compositional heterogeneity of otoliths and some is noise that reflects the limits of detect- ability and precision of the electron probe. However, comparisons of life history scans along similar growth axes of left and right otolith pairs indicate signifi- cant ontogenetic variability for all elements. For most elements, there is also evidence of geographic vari- ability in composition. The extent to which this ontogenetic and geo- graphic variability can be used to detect differences in either life histories or population structure criti- cally depends on the scale of the life history or popu- lation 'signal' relative to analytical 'noise.' In that regard, data quality varies widely among elements. Two identifiable sources of this 'error' are the effects Thresher et al.: Otolith analysis of Nemadactylus macropterus 835 of beam conditions, which differ among elements (e.g. Na analysis was more sensitive to effects of pitting than was analysis of Sr) (see Gunn et al., 1992), and the low precision of estimates for elements at low mean concentrations (e.g. CI and S). The effects of these factors can be estimated from standard formu- lae and empirically by comparing 'replicate' analy- ses on the same otolith. In practice, true replication is impossible, owing to the effects of beam damage and small-scale heterogeneity in composition, but it can be approximated by comparing points in two parallel life history scans. Our comparison (Fig. 6) is also a worst-case scenario in that it also includes the effects of different beam-power densities and point spacing in the two scans, which can be expected to have a marked effect on the estimated concentrations of some elements, such as Ca. Nonetheless, for the five elements other than Ca, differences between 'rep- licate points' are still on the order of the theoretical analytical precision (Table 2) and suggest a conser- vative difference criterion between point analyses that ranges from 331 ppm for Sr to 73 ppm for CI (Table 2). Ontogenetic comparisons also require that otolith sections be accurately duplicated among individuals. Our test of this accuracy — a comparison of left and right otoliths from single individuals — leads to three conclusions. First, despite our best efforts we could not guarantee duplication of the life history track between otoliths. Pairwise comparisons suggest vari- able compression and expansion of the ontogenetic signal between pairs, which presumably reflects slight differences between otoliths in the beam path relative to the main growth axis. The accuracy of duplication was generally high, but also differed among individuals and declined as distance from the primordium increased. Second, nonetheless, the over- all pattern of peaks and troughs in the pairs of otoliths compared was generally quite similar. As a result, the principal ontogenetic patterns in, for ex- ample, Sr concentrations would be reflected in both otoliths, but examination of one alone could lead to erroneous conclusions about the life history stage at which a particular change in concentration occurred. The variability in life history scans induced by dif- ferential compression renders statistical comparisons of ontogenetic patterns extremely difficult and liable to subjective interpretation. In theory, these difficul- ties could be overcome by calibrating ontogenetic changes in composition against real age — as opposed to distance along the growth axis — but difficulties in resolving the ages of larger individuals are likely to make this approach problematic for most species. Third, for at least some elements, mean concen- trations and ontogenetic patterns appear to differ between otoliths even within the same individual. In two of the three pairs examined, mean CI and S concentrations differed significantly between otoliths over relatively large sections of the main growth axis and at levels well above measurement error. This asymmetry was so surprising that we repolished and reanalyzed one pair of otoliths (specimen #312) to confirm the results; the second series of data were virtually identical to the first. The implication is that otolith pairs do not encode life history information in the same way. As yet, sample sizes for this com- parison are much too small to assess the generality of mismatches and the scale of the problem, but the available information suggests treating with caution data obtained from single point analyses in otoliths. Another potential source of methodological errors is specimen contamination. J. Calaprice (in press), for example, discounted CI as a stock discriminator in his studies on Atlantic bluefin tuna, Thunnus thynnus, because the element is widely present in the laboratory environment and easily transferred during specimen preparation and handling. Given this, the dependence of our site separation on varia- tion in CI concentrations is of concern. Although con- tamination is a critical issue (particularly at the sub- ppm level), several of our observations are not im- mediately consistent with the contamination hypoth- esis. First, CI does not vary independently; its con- centrations in otoliths covaries among specimens with Na and K. If CI concentrations are principally contaminants, the same contamination must affect Na and K concentrations, which is unlikely. Second, samples collected in the same region but at different times, different places, and with different gear types (e.g. juveniles and adults from the Tasmanian sites collected by hook-and-line and trawling) exhibit simi- lar concentrations of CI, suggesting that observed variability is not a consequence of the way individu- als are caught and handled. And third, the order in which the specimens were prepared and analyzed was randomized to check for systematic error; none was detected. Evidence for regional variation in otolith composition Several previous studies, using probe microanalysis (e.g. Radtke, 1989; Kalish, 1990) and whole otolith analysis (e.g. Gaudie et al., 1986; Edmonds et al., 1991 ), have demonstrated that otoliths vary in com- position ontogenetically and regionally. Our data permit a detailed evaluation of the interaction be- tween these components of variability. However, be- cause we could not collect and analyze otoliths of lar- vae from known spawning areas our results only test 836 Fishery Bulletin 92(4), 1994 indirectly the potential of the technique to resolve spawning stock structure. Specifically, we sought evidence of regionally different patterns in otolith composition that might reflect stock structure. In that regard, concentrations measured near the primordium differed significantly for four of six ele- ments among adults from the six sites. In all four cases, the range of mean values among sites exceeded an empirically derived 'minimum significant differ- ence' by at least 50% (Table 2). Furthermore, the pattern of differences among sites appeared to be regionally based: the two Tasmanian samples pooled together in the discriminant analysis, as did the geo- graphically contiguous NSW and Victorian samples. Such a grouping of sites could imply any of four dif- ferent mechanisms: 1) all sites differ, and the group- ing is a statistical artefact of the small number of sites and individuals sampled; 2) regional differences result from retrospective changes in otolith chemis- try in response to the latest conditions encountered by each adult, and adjacent sites pool because their environmental characteristics are more similar than those of widely separated sites; 3) the sites pool be- cause each regional set derives uniquely from a com- mon spawning ground or spawning population; and 4) each set is derived from a number of spawning grounds or populations that have similar chemical fingerprints, within which individuals mix widely and the boundaries of which are set by constraints on adult or larval mixing. The possibility that the regional groupings are an artefact is difficult to evaluate without knowing the range of chemical fingerprints possible and their like- lihood of occurrence. Assuming three chemical phe- notypes randomly distributed among six individu- als (=sites), then the probability that at least two adjacent sites will have identical characteristics is extremely high. However, given the number of pos- sible permutations, the probability that all pooling of sites will be only among nearest neighbors is less than 0.01. Therefore, we reject the hypothesis that the apparent regional groupings are a statistical artefact. We also think it unlikely that the groupings (and similarity offish within sites) are the result of retro- spective modification of otolith chemistry. It is a consistent assumption of otolith-based aging stud- ies that otolith structure is not modified after depo- sition. A similar assumption underlies chemical stud- ies, although there are no experimental data to verify the point (as opposed to studies on scales, the chemi- cal compositions of which are modifiable retrospec- tively, e.g. Sauer and Watabe, 1989). In fact, it is likely that at least some water- and alcohol-soluble compounds are transported into or out of otoliths during preservation. However, our data are not con- sistent with such retrospective modification of the micro-constituents. Nemadactylus macropterus col- lected at the same site and time show little evidence of convergence on a common marginal composition. This implies that recent environmental history has little or no effect on the composition of the otolith margin and presumably even less on the interior. Where a common marginal composition was evident, as in Sr levels among juveniles collected in the same area, it appears to be related to an environmental effect during deposi- tion rather than to retrospective modification. Distinguishing between the other two hypoth- eses— a single spawning ground for each regional phenotype or multiple spawning grounds with region- ally restricted mixing — is not possible without addi- tional information. As noted, information on the re- productive biology of Australian N. macropterus is sparse. Smith (1989) found running-ripe individu- als in autumn (February-March) off NSW; we found large numbers of relatively young larvae present along the east, but not the west, coast of Tasmania (for sampling sites and protocol, see Thresher et al., 1989), and several unpublished reports indicate simi- lar larvae off Victoria and South Australia (in the Bight). These scattered observations suggest that N. macropterus spawn at a number of sites along the southeastern Australian coast and certainly spawn in each of the three regional groupings of sites iden- tified by otolith chemical analysis. But sampling is not yet detailed enough to determine whether there are discrete spawning areas, or whether spawning occurs in a continuous band of activity all along the coast. Genetic data provide little additional informa- tion. Richardson's (1982) samples were drawn from Tasmanian and NSW/Victorian sites and hence ap- pear to bracket two otolith-based regional groupings but indicate no significant genetic differences across this range. This result has recently been confirmed by Elliott and Ward ( 1994) for allozymes and Grewe et al. (1994) for mitochondrial DNA. The lack of genetic differentiation in southeastern Australian N. macropterus populations is consistent with our observations of apparent examples of lar- val mixing. Probe microanalysis of otoliths of juve- niles from Victorian and Tasmanian coastal habitats indicated that most are similar in composition to those of adults collected at the same sites, which suggests regional, self-recruiting populations. How- ever, the distribution of the probabilities that each juvenile originated in the region where it was col- lected was conspicuously bimodal. Four out of ten juveniles caught off Victoria had chemical phenotypes more typical of Tasmanian (3) or Bight (1) origin, whereas 8 of 106 Tasmanian-caught juveniles clas- sified mainly with the NSW/Victorian adult sample. Thresher et al.: Otolith analysis of Nemadactylus macropterus 837 Although the data are obviously preliminary, these mismatched individuals could be direct evidence of an exchange of individuals among populations dur- ing the larval stage. The apparent exchange rate varies depending upon the criterion selected and may well differ with site. Even a conservative estimate for the Tasmanian samples (i.e. defining migrants as individuals with a probability >90% of not being derived from the Tasmanian adults) nonetheless sug- gests an exchange rate of about 7-8%, which is high enough to prevent genetic divergence among samples from the NSW, Victorian, and Tasmanian sites (Elliott and Ward, 1994). Determinants of otolith composition The working hypothesis underlying our approach is that otolith composition is largely determined by environmental factors that presumably differ at rela- tively fine space scales. The data to support this en- vironmental sensitivity, however, are not abundant and to a large extent are drawn from the inverte- brate literature (e.g. Rosenberg, 1980; Schneider and Smith, 1982). Studies on teleosts are ambiguous. To date, all reported effects have involved Sr, which has been reported as sensitive to changes in salinity (Radtke et al., 1988; Kalish, 1990; Secor, 1992) and temperature (Radtke et al., 1990; Townsend et al., 1992; however, see Kalish, 1989; and Gallahar and Kingsford, 1992). There are two reasons to suspect that most of the elements detected in our study are less responsive to the environment than is widely assumed. First, most are physiologically important and their concen- trations tightly regulated in plasma and hence pre- sumably in endolymph (Kalish, 1991). For example, an expectation that relatively slight changes in sa- linity significantly affect the incorporation of Na and CI in otoliths is unrealistic in an animal with well- developed osmoregulatory mechanisms. Of the six elements detected, only Sr is likely to be relatively unaffected by such physiological controls, though it is presumably affected by many of the same factors that constrain variation in Ca concentrations and may well be subject to a suite of other physiological constraints (see Kalish, 1991). Second, our data are not consistent with a strong and direct effect of the environment on composition. Two observations are particularly relevant: 1) settle- ment into nursery areas had no apparent effect on otolith chemistry, other than a slight effect on Sr, and 2) differences among regional groupings are manifest from the primordium to nearly the otolith margin and hence were apparently unaffected by life history stage, irrespective of habitat occupied. Re- garding the transition to the nursery areas, for ele- ments other than Sr there was no indication of con- vergence on a common chemical phenotype by indi- viduals in a given nursery area, nor evidence of di- vergence among nursery areas in response to local conditions. This suggests that the concentrations of 5 of the 6 elements we measured do not vary in re- sponse to environmental conditions in the nursery areas in any direct way. The nursery areas sampled ranged from mid-shelf to shallow coastal embay- ments and differed markedly in temperature and salinity histories, water-column chemistry, depth, substratum, turbidity, and in invertebrate composi- tion (and hence presumably in the diets of the juve- niles). The apparent lack of an impact of any of these on otolith composition suggests their effects at the >100 ppm level are weak or indirect (or both), except possibly for effects on Sr. Similarly, the consistency of regional differences in concentrations through life suggests these differences are largely unaffected by changes in habitats that range from high seas nekton to coastal embayments. Although the concentrations of several elements ( Sr, Na, K, and S ) clearly vary on- togenetically in otoliths, this variation is superimposed on, and apparently separate from, whatever determines regional differences in composition. The causes of the regional 'base' differences in composition are not clear. There are several broad possibilities: • The chemical phenotype is modified retrospec- tively, based on the adult habitat or sample prepa- ration; for reasons discussed above, we think this mechanism unlikely; • Life cycles for each region are closed within areas of a uniquely diagnostic environment. This seems unlikely given the diversity of habitats occupied by the species during its life history, but cannot be rejected until the factors that affect otolith compo- sition are determined; • The base composition is ontogenetically set by en- vironmental influences early in the larva's life, and then maintained, although overlayed by ontoge- netic modification, throughout its subsequent life and environmental history; • The base composition is determined genetically. The information currently available is not suffi- cient to discriminate between a 'locked phenotypic effect' (#3) and a genetic hypothesis (#4). A key da- tum that would permit such discrimination is a mea- sure of year-class effects on otolith composition. Re- gional differences in otolith composition that vary among year classes argue against a genetic basis and for an environmental influence early in larval devel- 838 Fishery Bulletin 92(4). 1994 opment. Our current data are much too sparse for any statistically powerful test of year-class differ- ences, but preliminary results suggest only small differences among years for most sites. This is con- sistent with the similar classifications of adults and juveniles for both the NSW/Victorian and Tasmanian regional groupings; although the adult and juvenile samples differ in mean birth date by five years ( adults from the 1980-84 year classes, juveniles from the 1987 and 1988 year classes), the samples overlap broadly in the concentrations of the regionally diag- nostic elements. We tentatively conclude that the regionally diag- nostic 'base' concentrations of most measured ele- ments probably have a genetic basis. This conclu- sion conflicts with both genetic analyses of the spe- cies in Australia (Richardson, 1982; Elliott and Ward, 1994; Grewe et al., 1994), which indicate no regional differences and with our preliminary, conservative estimate of larval mixing among regional groupings. At this point, the data are not adequate to resolve this contradiction. Its resolution, however, critically affects the way compositional data obtained from electron-probe microanalysis are used for stock de- lineation. If the regional differences are primarily genetically determined, then year-class effects are likely to be relatively unimportant. This simplifies analysis but also implies that the usefulness of the approach depends on the extent and pattern of ge- netic differentiation among populations. However, if regional differences in 'base' concentrations are pri- marily determined by environmental effects, perhaps via a 'locked phenotype' mechanism of some kind, then variability among year classes could be a criti- cal covariate in an analysis of population structure. In that case, electron-probe microanalysis is likely to be useful wherever significant environmental dif- ferences between spawning grounds are known or suspected. Acknowledgments We thank N. Barrett, C. Bobbi, T Carter, A. Gronell, A. Jordan, S. Sie, and G. Suter for assistance in labo- ratory and field studies, J. Kalish, N. Manning, and T Rees for advice on specimen preparation, K. Haskard and M. Cameron for advice on statistical analyses, V. Mawson, D. Mills, D. Secor, R. Ward, and two anonymous referees for comments on the manuscript, and C. McRae and R Rummel for assis- tance in microprobe analysis and discussion of re- sults. This study was funded in part by grants 1987/ 15 and 1989/30 from the Australian Fishing Indus- try Research and Development Committee. Literature cited Ancey, M., F. Bastenaire, and R. Tixier. 1978. Applications of statistical methods in micro- analysis. In F. Maurice, L.Meny, and R. Tixier (eds.). Microanalysis and scanning electron microscopy, p. 319-343. Proc. Summer School St. Martin-d'Heres, 1978. Les Editions de Physique, Orsay, France. Avise, J. C, C. A. Reeb, and N. C. Saunders. 1987. Geographic population structure and species differences in mitochondrial DNAof mouthbrooding marine catfishes ( Ariidae) and demersal spawning toadfishes (Batrachoididae). Evolution 41:991- 1002. 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. J. Fish. Biol. 5:555-557. Behrens Yamada, S., T. J. Mulligan, and D. Fournier. 1987. Role of environment and stock on the elemen- tal composition of sockeye salmon iOncorhynchus nerka ) vertebrae. Can. J. Fish. Aquat. Sci. 44:1206- 1212. Buchardt, B., and P. Fritz. 1978. Strontium uptake in shell aragonite from the freshwater gastropod Limnaea stagnalis. Science 199:291-292. Calaprice, J. R. 1971. X-ray spectrometric and multivariate analyses of sockeye salmon (Oncorhynchus nerka ) from differ- ent regions. J. Fish. Res. Board Can. 28:369-377. 1985. Chemical variability and stock variation in northern Atlantic bluefin tuna. Collect. Vol. Sci. Pap. ICCAT/ Reel. Doc. Sci. CICTA/ Collecc. Doc. Cient. CICAA 24:222-254. In press. Applications of X-ray flourescence spec- trometry to stock delineation: a historical per- spective. In R. E. Thresher, J. Ianelli, D. Mills, and C. Proctor (eds.), International workshop on skeletal chemistry of fishes, abstracts and edited transcripts. NOAA Tech. Memo. NMFS. Calaprice, J. R., H. M. McSheffrey, and L. A. Lapi. 1971. Radioisotope X-ray fluorescence spectrometry in aquatic biology: a review. J. Fish. Res. Board Can. 28: 1583-1594. Calaprice, J. R., L. A. Lapi, and L. J. Carlsen. 1975. Stock identification using X-ray spectrometry and multivariate techniques. Bull. INPFC 32:81- 101. Cameron, M. In press. Statistical methods for classification of otoliths. In R. E. Thresher, J. Ianelli, D. Mills, and C. Proctor (eds.), International workshop on skeletal chemistry of fishes, abstracts and edited transcripts. NOAA Tech. Memo. NMFS. Coutant, C. C. 1990. Microchemical analysis offish hard parts for reconstructing habitat use: practice and promise. Am. Fish. Soc. Symp. 7:574-580. Thresher et al.: Otolith analysis of Nemadactylus macropterus 839 Edmonds, J. S., M. J. Moran, N. Caputi, and M. Morita. 1989. Trace element analysis offish sagittae as an aid to stock identification: pink snapper (Chrysophrys auratus) in western Australian waters. Can. J. Fish. Aquat. Sci. 46: 50-54. Edmonds, J. S., N. Caputi, and M. Morita. 1991. Stock discrimination by trace-element analy- sis of otoliths of orange roughy (Hoplostethus atlanticus), a deep-water marine teleost. Aust. J. Mar. Freshwater Res. 42:383-389. Elliot, N. G., and R. D. Ward. 1994. Enzyme variation in jackass morwong, Nemadactylus macropterus (Schneider, 1801) (Teleostei: Cheilodactylidae), from Australian and New Zealand waters. Aust. J. Mar. Freshwater Res. 45:51-67. Fisheries Agency of Japan. 1967. Study on identification of the Pacific salmon by neutron activation analysis. Survey and Re- search Division, Fisheries Agency of Japan (cited in Mulligan et al. 1983). Gallahar, N. K., and M. J. Kingsford. 1992. Patterns of increment width and strontium: calcium ratios in otoliths of juvenile rock blackfish, Girella elevata (M. ). J. Fish. Biol. 41:749-763. Gauldie, R. W., D. A. Fournier, D. E. Dunlop, and G. Coote. 1986. Atomic emission and proton microprobe stud- ies of the ion content of otoliths of chinook salmon aimed at recovering the temperature life history of individuals. Comp. Biochem. Physiol. 84A:607-615. Gauldie, R. W., and A. Nathan. 1977. Iron content of the otoliths of tarakihi (Tele- ostei: Cheilodactylidae). N. Z. J. Mar. Freshwa- ter Res. 11:179-191. Grewe, P. M., A. J. Smolenski, and R. D. Ward. 1994. Mitochondrial DNA variation in jackass mor- wong, Nemadactylus macropterus (Teleostei: Cheilodactylidae) from Australian and New Zea- land waters. Can. J. Fish Aquat. Sci. 51(51:1101- 1109. Gunn, J. S., I. R. Harrowfield, C. H. Proctor, and R. E. Thresher. 1992. Electron probe microanalysis offish otoliths — evaluation of techniques for studying age and stock discrimination. J. Exp. Mar. Biol. Ecol. 158:1-36. Hartl, D. L., and A. G. Clark (eds.) 1989. Principles of population genetics, 2nd ed. Sinauer Associates Inc., Sunderland, MA, 682 p. Houck, J. E., R. W. Buddemeier, S. V. Smith, and P. J. Jokiel. 1977. The response of coral growth and skeletal strontium to light intensity and water temper- ature. In D. L. Taylor (ed.), Proc. third int. coral reef symp. Miami, Vol. 2:425-431. Kalish, J. M. 1989. Otolith microchemistry: validation of the ef- fects of physiology, age and environment on otolith composition. J. Exp. Mar. Biol. Ecol. 132:151-178. 1990. Use of otolith microchemistry to distinguish the progeny of sympatric anadromous and non- anadromous salmonids. Fish. Bull. 88:657-666. 1991. Determinants of otolith chemistry: seasonal variation in the composition of blood plasma, en- dolymph and otoliths of bearded rock cod Pseudo- phycis barbatus. Mar. Ecol. Prog. Ser. 74:137-159. Klecka, W. R. 1980. Discriminant analysis. Sage, Beverly Hills, CA, 70 p. Klokov, V. K., and L. A. Frolenko. 1970. Elementary chemical composition of scales of pink salmon. Izv. Tikhookean. Nauchno-Issled. Instit. Rybn. Khoz. Okeanogr. 71:159-168. Lapi, L. A., and T. J. Mulligan. 1981. Salmon stock identification using a microana- lytic technique to measure elements present in the freshwater growth region of scales. Can. J. Fish. Aquat. Sci. 38:744-751. Mulligan, T. J., L. Lapi, R. Kieser, S. B. Yamada, and D. L. Duewer. 1983. Salmon stock identification based on elemen- tal composition of vertebrae. Can. J. Fish. Aquat. Sci. 40:215-229. Mulligan, T. J., F. D. Martin, R. A. Smucker, and D. A. Wright. 1987. A method of stock identification based on the elemental composition of striped bass Morone saxatilis (Walbaum) otoliths. J. Exp. Mar. Biol. Ecol. 114:241-248. Pichou, J. L., and F. Pichoir. 1984. A new model for quantitative x-ray micro- analysis. La Recherche Aerospatiale, 3:167-192. Radtke, R. L. 1989. Strontium-calcium concentration ratios in fish otoliths as environmental indicators. Comp. Biochem. Physiol. 92A: 189-193. Radtke, R. L., R. A. Kinze, and S. D. Folsom. 1988. Age at recruitment of Hawaiian freshwater gobies. Environ. Biol. Fishes. 23:205-213. Radtke, R. L., D. W. Townsend, S. D. Folsom, and M. A. Morrison. 1990. Strontiumxalcium concentration ratios in otoliths of herring larvae as indicators of environ- mental histories. Environ. Biol. Fish. 27:51-61. Richardson, B. J. 1982. Geographical distribution of electrophoreti- cally detected protein variation in Australian com- mercial fishes. II: Jackass morwong, Cheilodactylus macropterus Bloch and Schneider. Aust. J. Mar. Freshwater Res. 33:927-931. Robertson, D. A. 1978. Spawning of tarakihi (Pisces: Cheilodacty- lidae ) in New Zealand waters. N.Z.J. Mar. Fresh- water Res 12:277-286. Rosenberg, G. D. 1980. An ontogenetic approach to the environmental significance of bivalve shell chemistry. In D. C. Rhoads and R. A. Lutz (eds.), Skeletal growth of aquatic organisms, p. 133-168. Plenum Press, NY. 840 Fishery Bulletin 92(4), 1994 Sauer, G. R., and N. Watabe. 1989. Temporal and metal-specific patterns in the ac- cumulation of heavy metals by the scales oiFundulus heteroclitus. Aquat. Toxicol. 14:233-248. Schneider, R. C, S. V. Smith. 1982. Skeletal Sr content and density in Porites spp. in relation to environmental factors. Mar. Biol. 66:121-131. Secor, D. H. 1992. Application of otolith microchemistry analysis to investigate anadromy in Chesapeake Bay striped bass Morone saxatilis. Fish. Bull. 90:798-806. Sie, S. H., and R. E. Thresher. 1992. Micro-PIXE analysis offish otoliths: method- ology and evaluation of first results for stock discrimination. Internat. J. PDCE 2:357-380. Sinclair, M. 1987. Marine populations: an essay on population regulation and speciation. Univ. Wash. Press, Seattle, WA, 252 p. Smith, D. C. 1982. Age and growth of jackass morwong (Nerna- dactylus macropterus Bloch & Schneider) in east- ern Australian waters. Aust. J. Mar. Freshwater Res. 33:245-253. 1989. The fisheries biology of jackass morwong in eastern Australian waters. Ph.D. thesis, Univ. New South Wales, Australia. Smith, P. J., A. Jamieson, and A. J. Birley. 1990. Electrophoretic studies and the stock concept in marine teleosts. J. Cons. Int. Explor. Mer 47:231-245. Smith, S. V., R. C. Buddemeier, R. C. Redalje, and J. E. Houck. 1979. Strontium-calcium thermometry in coral skeletons. Science 204:404-407. Sokal, R. R., and F. J. Rohlf. 1981. Biometry. Freeman, San Francisco, CA, 859 p. Thompson, G., and H. D. Livingston. 1970. Strontium and uranium concentrations in aragonite precipitated by some modern corals. Earth Planet. Sci. Lett. 8:439-442. Thresher, R. E., B. D. Bruce, D. M. Furlani, and J. S. Gunn. 1989. Distribution, advection, and growth of larvae of the southern temperate gadoid, Macruronus novaezelandiae (Teleostei: Merlucciidae), in Austra- lian coastal waters. Fish. Bull. 87:29-48. Tilzey, R. D. J., M. Zann-Schuster, N. L. Klaer, and M. J. Williams. 1990. The south east trawl fishery: biological synop- ses and catch distributions for seven major commer- cial fish species. Bureau Rural Resour. Bull. 6, 80 p. Townsend, D. W., R. L. Radtke, S. Corwin, and D. A. Libby. 1992. Strontiumxalcium ratios in juvenile Atlantic herring Clupea harengus L. otoliths as a function of water temperature. J. Exp. Mar. Biol. Ecol. 160:131-140. Vooren, C. M. 1972. Post-larvae and juveniles of tarakihi, (Tele- ostei: Cheilodactylidae) in New Zealand. N. Z. J. Mar. Freshwater Res. 6:601-618. Waples, R. S., and R. H. Rosenblatt. 1987. Patterns of larval drift in southern Califor- nia marine shore fishes inferred from allozyme data. Fish. Bull. 85:1-11. Weber, N. J. 1973. Incorporation of strontium into reef coral skel- etal carbonate. Geochim. Cosmochim. Acta 41: 2173-2190. Abstract. — The reproductive bi- ology of red drum, Sciaenops ocel- latus, in the northern Gulf of Mexico is described from examina- tion of 3,351 specimens sampled from March 1986 through Septem- ber 1992. The sex ratio of the spawning population, as manifest in purse seine collections, was es- sentially 1:1. Gonosomatic indices and ovarian histology demon- strated an 8-9 week spawning sea- son from mid August to early Octo- ber. Both sexes achieved >50% maturity at age 4; however, at 50% maturity males were somewhat smaller than females (660-670 mm vs. 690-700 mm, 3.4-3.5 kg vs. 4.0- 4.1 kg). Simultaneous observations of oocytes in all stages of matura- tion throughout the spawning sea- sons confirmed group-synchronous oocyte maturation and multiple batch spawning. Batch fecundity of 51 females (age 3-33 yr) ranged from 0.16 million to 3.27 million ova per batch (mean=1.54 million ova) and was positively correlated with fork length, gonad-free body weight, eviscerated body weight, and age. Seasonal spawning fre- quencies estimated from the pro- portion of mature females with postovulatory follicles varied widely from once every 3 days to once every 80 days. More plausible spawning frequencies (2-4 d) were obtained if proportions of females exhibiting oocyte yolk coalescence and oocyte hydration, indicative of imminent spawning, were included in estimates of this variable. Reproductive biology of red drum, Sciaenops ocellatus, from the neritic waters of the northern Gulf of Mexico* Charles A. Wilson David L. Nieland Coastal Fisheries Institute, CCEER Wetland Resources Building, Louisiana State University Baton Rouge, LA 70803-7503 Manuscript accepted 18 March 1994. Fishery Bulletin 92:841-850. Red drum, Sciaenops ocellatus (fam- ily: Sciaenidae), has been a prime target species for both recreational and commercial fishermen in the northern Gulf of Mexico. As conflicts over allocation of a purportedly de- clining population escalated in the mid 1980s, management of the red drum offshore spawning stock be- came an imperative. However, ba- sic to the formulation of any man- agement strategy is the need for sound biological information, in- cluding the various aspects of the species' reproductive biology. Most of the literature on red drum repro- duction (see Murphy and Taylor [1990] for a review) has been de- rived from studies of juveniles and itinerant adults in estuarine wa- ters. Little is known of the repro- ductive biology of adult "bull" red drum that assemble into large schools in the northern Gulf and constitute the spawning stock for the species in this area. Overstreet1 first reported on vari- ous aspects of the biology, including reproduction, of schooling red drum based on specimens gathered from the purse-seine fishery for the spe- cies. Fitzhugh et al. ( 1988) added to this body of knowledge by describ- ing ovarian development in speci- mens similarly taken from the purse-seine fishery. They further provided the first documentation of feral red drum as group synchro- nous, batch spawners (Wallace and Selman, 1981) in which clutches of newly matured ova are spawned periodically throughout the spawn- ing season. Thus all previous esti- mates of red drum fecundity in the wild, based on numbers of vitello- genic oocytes present in the ovary, were rendered invalid. The present study is a continua- tion and an expansion of the work begun by Fitzhugh et al. ( 1988) and was undertaken in conjunction with two studies of red drum age and growth (Beckman et al., 1988; Wil- son et al.2). Our specific objectives were 1) to verify the sex ratio of the spawning population; 2) to ascer- tain the duration of the spawning season; 3) to determine age, length, and weight at sexual maturity; and ♦Contribution LSU-CFI-93-4 of the Coastal Fisheries Institute, CCEER, Louisiana State University. 1 Overstreet, R. M. 1983. Aspects of the bi- ology of the red drum, Sciaenops ocellatus, in Mississippi. Gulf Res. Rep. (Suppl. 1), p. 45-68. Gulf Coast Res. Lab., Ocean Springs, MS. 2 Wilson, C. A., D. L. Nieland, and A. L. Stanley. 1993. Variation of year-class struc- ture and annual reproductive output of red drum Sciaenops ocellatus and black drum Pogonias cromis from the northern Gulf of Mexico. Final Report for 1991-1992 to U. S. Department of Commerce, Marine Fish- eries Initiative (MARFIN) Program, NA90AA-H-MF724. LSU-CFI-93-3. 37 p. and 16 figs. 841 842 Fishery Bulletin 92(4). 1994 4) to estimate batch fecundity and spawning fre- quency of red drum from the neritic waters of the northern Gulf of Mexico. Methods and materials Red drum were sampled in the neritic waters of the northern Gulf of Mexico (Mobile Bay, Alabama west- ward to Galveston Bay, Texas) from March 1986 through September 1992, a period spanning seven spawning seasons. Various aspects of the reproduc- tion of those specimens taken from March 1986 through November 1986 have been previously re- ported by Fitzhugh et al. (1988). The availability of red drum during this study was sporadic and gener- ally limited by federal and state restrictions on both the commercial and recreational fisheries in the Gulf of Mexico. Although the vast majority of specimens were taken by purse seine, the strategies used to lo- cate red drum schools for capture varied among years. Prior to the closure of the Exclusive Economic Zone to all red drum harvest in 1986, specimens came from the commercial purse-seine fishery. Red drum col- lected from July 1986 through 1988 were taken con- currently with a National Marine Fisheries Service tag-recapture investigation. Their methodology simu- lated a commercial purse-seine fishery and used spot- ter airplanes to locate schools at the surface. Since August 1989 most red drum were taken incidentally in the directed purse-seine harvest of blue runner, Caranx chrysos, detected visually at the surface. The above samples were supplemented with speci- mens from sportfishing tournaments, from inciden- tal catches of vessels targeting snappers ( Lutjanidae), and from gillnet and haul-seine catches. These sources were sampled as circumstances allowed to permit tracking of ovarian development and gonosomatic indices during those months when speci- mens taken by purse seine were not available. Protocols for the collection of morphometric data (fork length [FL] in mm, total weight [TW] in kg, eviscerated body weight [BW] in kg), processing of ovaries for histological examination, and enumera- tion of oocyte maturation stages from histological slides (Wallace and Selman, 1981; Fitzhugh et al., 1988 ) are given in Nieland and Wilson ( 1993). Histo- logical slides were also scanned for the presence of yolk coalescence, and for postovulatory follicles and atretic follicles. Ages of individuals were estimated from sagittal otoliths as described in Beckman et al. (1988). Their methodology assumed a biologically reasonable hatching date of 1 October; however, for our purposes age estimates were calculated with 1 August as the arbitrary red drum hatching date. This modification allows all members of a cohort to be assigned the same integer year age. Sexual maturity of females captured during the spawning season was defined as the progression of oocyte maturation to vitellogenesis (Brown-Peterson et al., 1988; Nieland and Wilson, 1993). Milt flow from the central lumen of the testes produced by gentle squeezing indicated sexual maturity in males simi- larly taken during the spawning season (Pearson, 1929; Brown-Peterson et al., 1988; Murphy and Tay- lor, 1990; Nieland and Wilson, 1993). Only those red drum females captured by purse seine were included in calculations of spawning fre- quency and batch fecundity. Batch fecundity was estimated gravimetrically from fresh ovarian weights for 51 females exhibiting overt macroscopic and mi- croscopic hydration of oocytes with the hydrated oo- cyte method (Hunter and Goldberg, 1980; Hunter et al., 1985). Seasonal spawning frequencies were esti- mated with two different methods after examination of 572 ovaries collected during the spawning seasons. The postovulatory follicle method (Hunter and Goldberg, 1980; Hunter and Macewicz, 1985; Hunter et al., 1985; Brown-Peterson et al., 1988; Nieland and Wilson, 1993) uses the number of mature females with postovulatory follicles to determine a spawn- ing fraction or that proportion of the female spawn- ing population that spawned the previous day. The inverse of the spawning fraction, the spawning fre- quency, is the average number of days over which each reproductively active female will spawn once. The spawning frequency estimates of Fitzhugh et al. (1993), referred to as the "time-calibrated" method, are based on a time-course of final oocyte matura- tion for black drum, Pogonias cromis, and another sciaenid species, the spotted seatrout, Cynoscion nebulosus, (Brown-Peterson et al., 1988). This meth- odology calculates proportions of day-0 females (im- minent spawners evidenced by oocyte yolk coales- cence or hydration) and day-1 spawners (previous spawners evidenced by postovulatory follicles) in the female spawning population. The average of the pro- portions of day-0 and day-1 females yields a spawn- ing fraction which is inverted to produce spawning frequency as defined above. Note that females spawn- ing on consecutive days will be classified as both day- 0 and day-1 individuals. Also those females evidenc- ing oocyte atresia states 2 (atresia of >509r of vitel- logenic oocytes) and 3 (atresia of 100% of vitellogenic oocytes) (Fitzhugh et al., 1993) were not included in either estimate of spawning frequency. Both condi- tions, usually encountered at the end of the spawn- ing season, indicate a zero probability of future spawning and an effective exit from the spawning population (Hunter and Macewicz, 1985). Wilson and Nieland: Reproductive biology of Sciaenops ocellatus 843 The temporal persistence of postovulatory follicles in red drum ovaries was investigated in two captive red drum that were induced to spawn by means of photoperiod manipulation and gonadotropin injection (Nieland, Wilson, and Thomas, unpubl. data). Postovu- latory follicles at 16-hour postspawning showed defi- nite signs of degeneration, yet were recognizable as such and resembled those seen in many wild caught females. However, postovulatory follicles at 24-hour postspawning were extremely degenerate and had assumed an aspect much like that of an atretic fol- licle. Such a condition was rare in feral specimens; thus, all identifiable red drum postovulatory follicles were assumed to be less than 24 hours old. Post- ovulatory follicles seen in histological material were classified as early, late, or very late based on their degree of degeneration. Relative investment of energy to reproduction of red drum was assessed with gonosomatic indices (GSI) cal- culated as GSI = (gonad weight/BW) x 100 (Nieland and Wilson, 1993). Calculations of mean monthly GSI exclude immature individuals of both sexes (Wilk et al., 1990; Nieland and Wilson, 1993). Because the in- crease in ovary mass, which occurs concomitantly with oocyte hydration, does not reflect energy to be expended in reproduction, females with hydrating oocytes were also excluded from calculations of mean monthly GSI. The Statistical Analysis System (SAS Institute Inc., 1985) was used for analysis of variance (ANOVA), maxi- mum-likelihood analysis (PROBIT), and linear regres- sion (GLM). Significance level for statistical analyses was 0.05 unless indicated otherwise. Results A total of 3,351 red drum ( 1,585 males, 1,765 females) were sampled for reproductive analysis. Of these, both the intact gonad weight and BW necessary for calculation of GSI were available for 2,859 mature and 341 immature specimens. Data on ovarian his- tology were compiled for 1,379 mature females and 123 immature females. Age at time of capture for 3,316 red drum for which otoliths were available ranged from 1 to 36 years for males and from 1 to 39 years for females. Proportions of the younger age classes were particularly high during our 1992 sam- pling when 327 of 504 individuals were age 6 or less. Total weight and FL ranges among all specimens were 0.7-19.2 kg and 399-1,115 mm, respectively. Sex ratio Sex ratios for red drum were highly variable among source and gear categories and between mature and immature individuals within these categories (Table 1). Among all specimens and all mature specimens, females were predominant; however, immature males exceeded immature females in number by two to one. Females also outnumbered males among all mature individuals taken by sportfishing, among mature individuals caught incidentally with lutjanids, and among those captured with haul seine. Conversely, males were more common among all purse-seine specimens, all immature specimens, and immature specimens taken by all methods except for sportfishing. Sex ratios were not statistically differ- ent from 1:1 for all mature red drum captured with purse seines, for all taken incidentally with lutjanids, and for those caught in gill nets. Seasonality Gonosomatic indices and ovarian histological data indicated a potential 8—9 week red drum spawning season beginning in mid August and extending into October. Minimal GSI values for mature individuals were found from January to July, averaging 0.26 for males and 0.81 for females during these months (Fig. 1). Abrupt escalation of male and female GSI in Au- gust signalled potential, if not certain, spawning ac- tivity. Maximum GSI values were achieved in Sep- tember followed by a return to near minimum levels Table 1 Number of specimens sampled by sex and sex ratios for all, mature, and immature red drum collected in the northern Gulf of Mexico from March 1986 through September 1992 by source 3r gear of cap- ture. No immature individua s were taken by haul seine or gill net. Sex ratio Source or gear Females Males (Female:Male) All sources and gears 1,766 1,585 1:0.90 Mature 1,642 1,364 1:0.83 Immature 124 221 0.56:1 All purse seine 1,247 1.362 0.92:1 Mature 1,152 1,212 0.95:1" Immature 95 150 0.63:1 All sportfishing 383 102 1:0.27 Mature 369 87 1:0.24 Immature 14 15 0.93:1" All incidental with lutjanids 54 69 0.78:1" Mature 39 13 1:0.33 Immature L5 56 0.27:1 Haul seine 43 11 1:0.26 Gill net 39 41 0.95:1" Not significantly different from 1:1 Ichi-square test.df=l,P<0.05). 844 Fishery Bulletin 92|4), 1994 MM I JSNJMMJSNJMMJSNJMMJSNJMMJSNJMMJSN 986 | 1987 | 1988 | 1989 | 1990 | 1991 JMMJS | 1992 | Month/Year Figure 1 Mean monthly gonosomatic indices for mature male (n=l,306) and female (rc = l,553) red drum from the north- ern Gulf of Mexico. Sample sizes for months during which red drum were available varied from 1 to 110 for males and from 1 to 183 for females. Absence of vertical bar for either sex indicates that no specimens were sampled dur- ing the month and year indicated below. Narrow vertical bars at tops of GSI bars indicate +1 standard deviation (SD) from mean value (plotted only for those months when SD>0.5). in October. Our data, collected over more than six years, demonstrated a single annual GSI maximum (ranging from 4 to 8) in each sex. Red drum ovarian tissues undergo an annual cycle of oocyte maturation and recrudescence coincident with the female GSI cycle. Over the entire study duration, oogonia and primary growth oocytes were present in varying numbers in all ovary samples but were virtually ubiquitous from January through June (Fig. 2). Recrudescence and maturation of primary growth oocytes to the cortical alveoli and vitellogenic stages, indicative of preparation for spawning, was seen in July. Maximum numbers of maturing vitellogenic and mature hydrated oocytes were found during August and September. Declines in numbers of cortical alveolar and vitellogenic oocytes and con- comitant increases in numbers of primary growth oocytes occurred in October suggesting the cessation of spawning activity at this time. Although we have few data from November and December, primary growth oocytes are assumed to constitute nearly 100% of the oocyte population during these months. Among all spawning seasons, the onset of spawn- ing activity, evidenced by the first observation of yolk coalescence (late vitellogenesis) or postovulatory fol- licles in ovarian histological samples, ranged from 14 August to 18 August. Atretic states 2 and 3, indicative of cessation of spawning, became in- creasingly common from late September through early October. When we were able to extend our sampling in 1987 and 1988, 100% atresia of vitellogenic oocytes was realized by the end of October indicating completion of spawning at this time. Age, length, and weight at maturity The onset of sexual maturity in both male and fe- male red drum in the northern Gulf of Mexico is variable with respect to age. Estimates of percent maturity at age are comparable and increase at much the same rate for both sexes, the only major discrepancies occurring at age 2 and age 5 (Table 2). Greater than 50% maturity is achieved in both sexes at age 4. All males and all females are ma- ture at age 5 and age 6, respectively. Sexual maturity in red drum is similarly unre- lated to size of the individual (Table 2). Fork length and TW minima among mature female red drum were 598 mm (age 4, 4.18 kg) and 3.43 kg (age 3, 675 mm). Maximum-likelihood analysis (PROBIT analysis, Murphy and Taylor, 1989) of 10 mm in- crements and of 0.10 kg increments indicated 50% maturity is achieved at 690-700 mm and 4.00- 4.10 kg. All females greater than 810 mm and 6.10 kg were mature. Male red drum from the northern Gulf of Mexico mature at somewhat lesser length and weight than do females; however, sex-specific percent maturities at size become roughly equivalent at 750-700 mm and 5.00-5.49 kg (Table 2). Fork length and TW minima for mature males were 593 mm (age 5, 2.56 kg) and 2.35 kg (age 2, 615 mm). Fifty percent matu- rity (maximum-likelihood analysis as above) occurred at 660-670 mm and 3.40-3.50 kg and all males greater than 810 mm and 5.40 kg had matured. We observed no instances of decreased or arrested gonadal development among older individuals. Red drum of both sexes appear to be fully capable of re- productive activity from the onset of maturity until death. Batch fecundity Fitzhugh et al. (1988) reported significant ovarian location effects in their estimates of red drum batch fecundity. To test the precision of our estimates within individuals, replicate samples (30-60 mg) of ovarian tissue from each of six ovarian regions (anterior, medial, and posterior of both right and left lobes) were removed from six hydrated females captured Wilson and Nieland: Reproductive biology of Saaenops ocellatus 845 26 September 1988. Numbers of hydrated oocytes per gram of ovarian tissue were calculated for each sample (n=72). Nested ANOVAshowed significant variation only among individuals; no location ef- fects were demonstrated (Table 3). All individual batch fecundities herein are means of estimates made from three different randomly selected re- gions as defined above. Batch fecundity estimates were generated for a combined sample of 51 red drum captured by purse seine during the 1986 (previously reported by Fitzhugh et al. (1988)), 1987, 1988, 1989, and 1991 spawning seasons. All displayed overt macroscopic and microscopic manifestations of oocyte hydra- tion throughout the length and diameter of the ovarian lobes and were captured during the late afternoon or early evening hours (1600-1900 h). No sufficiently hydrated females were encountered in 1990 and 1992. Age, FL, batch fecundity ranges, and numbers of hydrated specimens examined by year of cap- ture are given in Table 4. Regression analyses of batch fecundity against FL (r2=0.58), gonad-free body weight (r2=0.46), age in year (r2=0.43), and BW (r2=0.43) are of reasonable predictive value (Fig. 3). Significant positive relations (P>0.0001) were indicated between batch fecundity and nontransformed values of the four independent variables. The relatively low r2 values for the re- gressions appear to result from individual varia- MJSDMJSDMJSDMJSDM 1986 | 1987 | 1988 | 1989 | Month/Year Figure 2 Monthly mean percent occurrence of primary growth (dia- monds), cortical alveoli (crosses), and vitellogenic (squares) oocytes in ovaries of red drum from the northern Gulf of Mexico. Total sample size is 1,379; for months during which female red drum were available, sample size ranged from 1 to 160. Dashed lines span months during which speci- mens were unavailable. Ranges of standard deviations (SD) during August-October were 3.4-18.3 for primary growth oocytes, 0.0-3.4 for cortical alveoli oocytes, and 2.2- 13.1 for vitellogenic oocytes. Table 2 Percent maturity and numbers sampled (in parentheses) of female and male red drum at age, fork length, and total weight. Specimens included are those taken by all gears during August through October of 1986-1991 and August and September 1992. Total sample sizes are 1,262 females and 1,137 males. Class Female Male Class Female Male Age (years) Fork length I mm) 1 0(0) 0(0) 750-799 95(129) 97(178i 2 0(8) 13(24) 800-849 99(216) 99(280) 3 28(81) 30(148) >850 100(764) 100(391) 4 5 >6 71 (75) 88 (68) 100(1,011) 73(88) 100(77) 100(787) Total weight (kg) <3.00 3.00-3.49 0(45) 8(24) 13(96) 35(54) 3.50-3.99 33(18) 60(40) Fork length (mm) 4.00-4.49 75(28) 84(31) <550 0(7) 0(15) 4.50-4.99 83 (23) 90 (52) 550-599 8(13) 8(25) 5.00-5.49 94 (33) 97 (39) 600-649 0(26) 22 (68) 5.50-5.99 95(60) 100 (55) 650-699 24(42) 48 (82) 6.00-6.49 98 (59) 100 (79) 700-749 82(65) 91 (98) >6.50 100 (963) 100(678) 846 Fishery Bulletin 92(4). 1994 Table 3 Nested analysis of variance on numbers of hydrated oo- cytes per gram ovary weight among ovarian regions (six total-three per lobe), between ovarian lobes (right and left), and among six female red drum captured 26 September 1988 from the northern Gulf of Mexico. MS=mean square. Source of Variation df MS F-value P>F-value Individuals Lobes Ovarian regions Error Total 5 6 24 36 72 42,580,790 242,773 2,290,678 4,266,679 49,380,920 69.86 0.33 0.78 0.0001 0.9155 0.7318 Table 4 Age, fork length (FL), and batch fecundity (BF ranges for red drum Sciaenops ocellatus from the northern Gulf of Mexico by year of capture, n =number of specimens. Age range FL range BF range Year n (yr) (mm) (ova x 106) 1986 s 6-21 800-964 0.75-2.54 1987 2 20-33 933-1005 1.65-1.67 1988 6 9-30 820-950 1.87-3.22 1989 23 3-24 697-999 0.16-3.27 1990 i) — — — 1991 12 5-25 760-924 0.57-3.13 1992 0 — — — Total ;M 3-33 697-1005 0.16-3.27 Table 5 Comparison of red drum seasonal spawning frequencies (SF, expressed as average days between successive spawnings) estimated with the postovulatory follicle (POF) method of Hunter and Macewicz ( 1985 ) and the time-cali- brated (TO method of Fitzhugh et al. (1993). Day-0 fe- males are those evidencing yolk coalescence or hydration of oocytes; day-1 females are those with postovulatory fol- licles from previous day's spawning. All specimens collected by purse seine from 13 August through 8 October. POF method TC method Year Mature females Females with POF SF Day-0 Day-1 females females SF tion within classes. Exclusion from the regression analyses of those specimens captured during Oc- tober, based on the possibility of declining output toward the end of the spawning season as sug- gested by Fitzhugh et al. (1988), produced r2 val- ues ranging from 0.23 to 0.39. Spawning frequency Red drum spawning frequencies estimated with the postovulatory follicle method were highly vari- able among years (Table 5) ranging from one spawning event every 2.8 days in 1986 to one spawn every 80.0 days in 1991. A total of 65 of 572 sexually mature females captured during the 1986-1992 spawning seasons evidenced postovu- latory follicles for a seven season average spawn- ing frequency of 8.8 days. Spawning frequencies calculated with the time-calibrated method (Table 5) showed less variation and gave more plausible estimates. Except for the 1992 spawning season, when sampling was limited to three dates during the spawning season (28 August, 3, 12 Septem- ber), spawning frequencies of one spawn every 2- 4 days were predicted. Discussion Aspects of red drum reproductive biology in the Gulf of Mexico have been variously inferred from visual observation of gonadal development, from larval and juvenile abundances and lengths, and from histological documentation of ovarian devel- opment. Given both the disparities and subjectiv- ity inherent among, and even within, these meth- odologies and the expanse of the Gulf of Mexico, it is not surprising that published accounts of red drum reproduction vary widely and, perhaps, geo- graphically. We agree with West (1990) that his- tological methods produce the most accurate and most reliable results in assessing ovarian devel- opment and predicting reproductive variables. We also stress that our findings should not be bm;idly applied to red drum populations throughout the Gulf of Mexico. The sex ratio of the 2,364 mature individuals taken by purse seine is undoubtedly most reflec- tive of the offshore spawning stock of red drum in the northern Gulf of Mexico. Given that most of our specimens taken by purse seine were captured either just prior to or during the spawning season and considering the substantial sample size, our data establish a 1:1 sex ratio in schools of pre- spawning and spawning red drum. This supports Wilson and Nieland: Reproductive biology of Sciaenops ocellatus 847 BF = 8,099 X FL - 5,241,504 4 4 r 2 _ 0.58, N = 51 ra > o 3- ' '} + 3 o (0 m + c o 2- M.m 2 — jr + + F ■ m^S ■■ ■ u. 1 m X 1 CD ^x ** K m X 0 0 I 1 1 1 1 6 rs 750 825 FL 900 (mm) 975 1,050 BF = 251,794 x GFBW - 172,855 rJ = 0.46, N = 51 ■ ■ + + * + ^^^ + X + * X 3 4 5 6 7 8 9 10 11 12 13 14 GFBW (kg) BF = 267,361 x BW - 219,064 4- r2 = 0.43, N = 51 CO > o ■ m + + ^_ 3 '+ ^ o in m ^^ c j^^ o I 2 v> + + il- 1 J^% ■ ea ^0^t~ m ■^ m x x^ -: BF = 77,784 x Age + 669,908 4 r2 = 0.43, N = 51 3 ■ . ■ + - ■ -*•«''*■ 2 m - 1 n ^-"* • X x • x 3 4 5 6 7 8 9 10 11 12 13 14 BW (kg) 0 2 4 6 8 10 12 1416182022242628303234 Age (years) Sept 17-18 + Sept 26-27 x Oct 4-8 Figure 3 Regressions of eviscerated body weight (BW), fork length (FL), gonad-free body weight (GFBW), and age versus estimated batch fecundity for red drum from the northern Gulf of Mexico. Squares denote specimens captured on 17 September 1986 and 18 September 1991, pluses on 26 September 1988 and 27 September 1987, and crosses on 4 October 1989 and 8 October 1986. n=total sample size. and further validates the use of a 1:1 ratio by Comyns et al. ( 1991) in their estimation of red drum spawner biomass in the north-central Gulf of Mexico. For immature red drum, males significantly out- numbered females across all source and gear catego- ries except among the few specimens randomly en- countered at sportfishing tournaments. This numeri- cal dominance of immature male red drum in off- shore waters may indicate a predisposition for emi- gration from estuarine habitats at younger age than females which is reflected in the somewhat lesser age and size at maturity seen in males. Previous accounts of red drum seasonality in the northern Gulf of Mexico have relied on inferences drawn from postspawning capture of larvae and ju- veniles and, to a lesser extent, from visual assess- ment of ovaries and testes. The red drum spawning season has been variously estimated with these methodologies as September to December (Boothby and Avault, 1971), August to November (Sabins, 1973), and from early September to early October (Comyns et al., 1991). However, our delineation of a mid-August to early October red drum spawning sea- son is in accord with other studies that used histo- logical techniques. Within our study area, analyses of red drum oocyte maturation by Overstreet1 in Mississippi and by Fitzhugh et al. (1988) in Louisi- ana both demonstrated that red drum spawning is initiated in August and continues into October. Murphy and Taylor ( 1990) found spawning red drum from August to mid-November 1981 and from Au- gust to mid-October 1982 in the Tampa Bay, Florida area. The concordance among these estimates, drawn objectively and directly from histological data, dem- 848 Fishery Bulletin 92(4), 1994 onstrates the effectiveness of this technique in as- sessing onset and duration of spawning seasons. Determining the combination of environmental factors which are the impetus for an August-October red drum spawning season is beyond the scope of this study. However, temperature data gathered at a weather buoy located approximately 28 km south of Biloxi, MS (lat. 30.1°N, long. 88.8°W) indicate that seasonal mean sea surface temperatures ranged from 27.3° to 28.8°C and that daily mean sea surface tem- peratures varied from 23° to 31°C during periods of active spawning (National Climatic Data Center, Asheville, North Carolina). Similar temperatures have proven to be optimal in the spawning, hatch- ing, and rearing of red drum in the laboratory (Arnold et al., 1977; Roberts et al., 1978; Holt et al., 1981; Arnold, 1988; Henderson-Arzapalo, 1992). Estimates of age and size at maturity for red drum in the Gulf of Mexico, based largely on visual assess- ment of gonadal development, show extensive varia- tion. In his study of red drum in Texas waters, Pearson ( 1929) perhaps originated the long held and widely applied belief that few red drum of either sex mature either before age 5 or before attaining 10 lb (4.5 kg) and 700 mm. Among red drum populations in Texas waters, maturity has been reported at 425 mm (Gunter, 1950), 625 mm (Miles3), age 4 and 29.5 inches (750 mm) (Miles4), and age 3 to 5 (Holt et al., 1981). For red drum off Mississippi, Overstreet1 pro- vided only a tabular compilation of the relation be- tween standard length (SL) and gonad maturity stages. These data were interpreted by Murphy and Taylor ( 1990) to show 50% maturity in both sexes at about 700 mm SL. Murphy and Taylor also presented maturity schedules, which were based on histology of ovaries and testes and gross appearance of each, for red drum in Florida waters. They found fifty-per- cent maturity of males at 529 mm FL and all males mature at age 3; among females 50% and 100% matu- rity occurred at 825 mm FL and 6 years, respectively. Given the maturity data cited above and that of the present study, one might infer the existence of geographical variation in maturity schedules among red drum populations in the Gulf of Mexico. We de- cline to discount this possibility. However, we sug- gest that differences in methods of maturity assess- ment and disparate definitions of maturity, especially in females, confound comparisons. West (1990) re- 3 Miles, D. W. 1950. The life histories of the spotted sea trout, Cvnosewn nebulosus, and the redfish, Sciaenops ocellat us. Annu. Rep. (1949-1950), Tex. Game and Fish Comm. Mar. Lab., p 66-103. Tex. Parks and Wildl. Dept., Austin. 4 Miles, D. W. 1951. The life histories of the sea-trout, Cynoscion nebulosus, and the redfish, Sciaenops ocellatus: sexual develop- ment. Annu. Rep. 1 1950-1951 1, Tex. Game and Fish Comm. Mar. Lab., 11 p., 2 figs., and 3 tables. Tex. Parks and Wildl. Dept., Austin. viewed methods of assessing ovarian development in fishes and concluded that histology, though less efficient in both cost and time, is less subjective than, and preferable to, other methodologies. For the purposes of the present study, we defined maturity in male red drum as the flow of milt from the central lumen of the testis during the August- October spawning season. The use of this subjective definition may account for some of the discrepancy in male maturity schedules between our study and that of Murphy and Taylor (1990). However, for assessment of the maturity schedule of female red drum in the northern Gulf we employed a histologically objective benchmark definition: the presence of vitellogenic oocytes in the ovaries of in- dividuals captured during the spawning season. Murphy and Taylor ( 1990) considered as mature only those females of class 4 (late vitellogenesis) or greater among their eight female reproductive classes in es- timating an 825 mm FL at 50% maturity. This ne- cessitated the categorization of out-of-season females (their class 2) and of in-season females evidencing early vitellogenesis (their class 3) as immature. The former would not have been included in our analysis of female maturity; the latter would have been clas- sified as mature under our definition which precludes judgments between early and late vitellogenesis. A cautious re-interpretation of the tabular data in Overstreet1 would yield greater than 50% maturity of females at 550-699 mm SL rather than the >700 mm SL as stated by Murphy and Taylor (1990). Applying our definition of maturity to Murphy and Taylor's data would perhaps produce a length at 50% maturity more in line with our estimate of 690-700 mm FL. Group-synchronous maturation of oocytes (Wallace and Selman, 1981) and multiple, or batch, spawning has been demonstrated in several species of sciaenid fishes, including red drum (Fitzhugh et al., 1988). Among these are queenfish, Seriphus politus (DeMartini and Fountain, 1981); black croaker, Cheil- otrema saturnam (Goldberg, 1981); white croaker, Genyonemus lineatus (Love et al., 1984); spotted seatrout (Brown-Peterson et al., 1988); and black drum (Fitzhugh et al., 1993; Nieland and Wilson, 1993). For these and other such species, the stand- ing crop of oocytes of some arbitrary size or of vitellogenic oocytes gives little indication of the individual's seasonal fecundity. Rather fecundity is indeterminate and is the result of clutches of oocytes matured and spawned periodically over the length of the spawning season. Thus any estimate of sea- sonal fecundity must consider the length of the spawning season, the number of ova released in each spawning event (batch fecundity), and the periodic- ity of these spawning events (spawning frequency). Wilson and Nieland: Reproductive biology of Sciaenops ocellatus 849 Prior to the confirmation of batch spawning in fe- ral red drum by Fitzhugh et al. ( 1988), fecundity of wild caught specimens had been variously estimated as 0.5-3.5 million ova per season (Pearson, 1929; Holt et al., 1981; Miles5). Much greater potential fecun- dities (up to 94.5 million), based on volumetric and gravimetric estimates of oocytes available for spawn- ing, were presented by Overstreet1. This potential for an immense seasonal reproductive output in wild red drum has been demonstrated in the laboratory where specimens have been manipulated to produce repeatedly a few hundred thousand to millions of ova per spawning event (Arnold et al., 1977; Roberts et al., 1978; Anonymous, 1979; Arnold, 1988). Other than the batch fecundity estimates for feral red drum presented by Fitzhugh et al. (1988) and those herein, only one other estimate has appeared in the litera- ture. Comyns et al. (1991), from our data for Sep- tember of 1986, 1987, and 1988, used a mean batch fecundity of 2.128 million ova in their computations of red drum spawner biomass in the north-central Gulf of Mexico. Our seasonal estimates of spawning frequency are the first to be presented for red drum in the wild. Those seasonal frequencies (3-5 d) calculated for 1986, 1987, and 1989 with the postovulatory method and those (2-4 d) calculated for 1986-1991 with the time-calibrated method are believed to be most rep- resentative of the spawning population as similar spawning frequencies have been observed in the labo- ratory (Arnold et al., 1977; Arnold, 1988). However, spawning frequency is likely not constant over the course of the spawning season. Within-season spawn- ing peaks coinciding with the new and full moon have been postulated by Peters and McMichael ( 1987) and Comyns et al. (1991) based on larval abundances. The irregularity of our sampling precluded our in- vestigation of this phenomenon. Given an 8—9 week spawning season, a mean batch fecundity of 1.54 million ova, and a spawning fre- quency of 2— A days, an average red drum female could be expected to spawn some 20^10 million ova per season. Among sciaenid species, this estimate of an- nual fecundity is exceeded only by that of the black drum, a species of similar size which has an annual fecundity of 35-45 million ova (Nieland and Wilson, 1993). Females of both species are potentially long- lived (30-35 yr) and, thus, might produce up to a billion of ova during their lifetimes. Acknowledgments The authors gratefully acknowledge Daniel Beck- man, Gary Fitzhugh, Brigitte Nieland, Robert Parker, Louise Stanley, Bruce Thompson, a host of graduate students and student workers, and person- nel of the National Marine Fisheries Service for their assistance in field sampling. Phillip Horn, Ralph Horn, Harlon Pearce, and numerous fishing tourna- ment organizations generously allowed us access to and use of their facilities. We especially thank Paul Morse, Jim Reahard, Jimmy Reahard, and the crews of the fishing vessels Captain Grumpy and Mistake for their help and hospitality. Permits for possession and transport of red drum, logistical and moral sup- port were provided by the National Marine Fisher- ies Service, Louisiana Department of Wildlife and Fisheries, and Mississippi Bureau of Marine Re- sources. Major sources of funding for this research were the U.S. Department of Commerce Marine Fish- eries Initiative (MARFIN) Program and the Louisi- ana Sea Grant College Program, a part of the Na- tional Sea Grant College Program administered by the National Oceanic and Atmospheric Administra- tion, U. S. Department of Commerce. The Louisiana Sea Grant College Program is also supported by the State of Louisiana. Literature cited Anonymous. 1979. Redfish spawn produces 11.4 million fry. Texas Parks Wildl. 37(71:15. Arnold, C. R. 1988. Controlled year-round spawning of red drum Sciaenops ocellatus in captivity. Contrib. Mar. Sci., Suppl. to vol. 30:65-70. Arnold, C. R., W. H. Bailey, T. D. Williams, A. Johnson, and J. L. Lasswell. 1977. Laboratory spawning and larval rearing of red drum and southern flounder. Proc. Annu. Conf. S.E. Fish and Wildlife Agencies 31:437-440. Beckman, D. W., C. A. Wilson, and A. L. Stanley. 1988. Age and growth of red drum, Sciaenops ocellatus, from offshore waters of the northern Gulf of Mexico. Fish. Bull. 87:17-28. Boothby, R. N., and J. W. Avault Jr. 1971. Food habits, length-weight relationship, and condition factor of the red drum (Sciaenops ocellatus) in southeastern Louisiana. Trans. Am. Fish. Soc. 100:290-295. Brown-Peterson, N., P. Thomas, and C. R. Arnold. 1988. Reproductive biology of the spotted seatrout, Cynoscion nebulosus, in South Texas. Fish. Bull. 86:373-388. Comyns, B. H., J. Lyczkowski-Schultz, D. L. Nieland, and C. A. Wilson. 1991. Reproduction of red drum, Sciaenops ocellatus, in the northcentral Gulf of Mexico: sea- sonality and spawner biomass. In R. D. Hoyt (ed.), 850 Fishery Bulletin 92(4). 1994 Larval fish recruitment and research in the Ameri- cas, p. 17-26. Dep. Commer., NOAA Tech. Rep. NMFS 95. DeMartini, E. E., and R. K. Fountain. 1981. Ovarian cycling frequency and batch fecun- dity in the queenfish, Seriphus politus: attributes representative of serial spawning fishes. Fish. Bull. 79:547-560. Fitzhugh, G. R., T. G. Snider III, and B. A. Thompson. 1988. Measurement of ovarian development in red drum {Sciaenops ocellatus) from offshore stocks. Contrib. Mar. Sci., Suppl. to vol. 30:79-83. Fitzhugh, G. R., B. A. Thompson, and T. G. Snider IH. 1993. Ovarian development, fecundity, and spawn- ing frequency of black drum Pogonias cromis in Louisiana. Fish. Bull. 91:244-253. Goldberg, S. R. 1981. Seasonal spawning cycle of the black croaker, Cheilotrema saturnum (Sciaenidae). Fish. Bull. 79:561-562. Gunter, G. 1950. Correlation between temperature of water and size of marine fishes on the Atlantic and Gulf coasts of the United States. Copeia 1950:298-304. Henderson-Arzapalo, A. 1992. Red drum aquaculture. Rev. Aquat. Sci. 6:479-491. Holt, J., A. G. Johnson, C. R Arnold, W. A. Fable Jr., and T. D. Williams. 1981. Description of eggs and larvae of laboratory reared red drum, Sciaenops ocellata. Copeia 1981:751-756. Hunter, J. R., and S. R. Goldberg. 1980. Spawning incidence and batch fecundity in northern anchovy, Engraulis mordax. Fish. Bull. 77:641-652. Hunter, J. R., and B. J. Macewicz. 1985. Measurement of spawning frequency in mul- tiple spawning fishes. In R. L. Lasker (ed.), An egg production method for estimating spawning biomass of pelagic fish: application to the northern anchovy, Engraulis mordax, p. 79-94. Dep. Commer., NOAA Tech. Rep. NMFS 36. Hunter, J. R., N. C. H. Lo, and R. J. H. Leong. 1985. Batch fecundity in multiple spawning fishes. In R. L. Lasker (ed.), An egg production method for estimating spawning biomass of pelagic fish: application to the northern anchovy, Engraulis mordax, p. 67-77. Dep. Commer., NOAA Tech. Rep. NMFS 36. Love, M. S., G. E. McGowen, W. Westphal, R. J. Lavenberg, and L. Martin. 1984. Aspects of the life history and fishery of the white croaker, Genyonemus lineatus (Sciaenidae), off California. Fish. Bull. 82:179-198. Murphy, M. D., and R. G. Taylor. 1989. Reproduction and growth of black drum, Pogonias cromis, in northeast Florida. Northeast Gulf Sci. 10:127-137. 1990. Reproduction, growth, and mortality of red drum, Sciaenops ocellatus, in Florida waters. Fish. Bull. 88:531-542. Nieland, D. L., and C. A. Wilson. 1993. Reproductive biology and annual variation of reproductive variables of black drum in the north- ern Gulf of Mexico. Trans. Am. Fish. Soc. 122:318- 327. Pearson, J. C. 1929. Natural history and conservation of redfish and other commercial 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 (Pi- sces: Sciaenidae) in Tampa Bay, Florida. Es- tuaries 10:92-107. Roberts, D. E., Jr., B. V. Harpster, and G. E. Henderson. 1978. Conditioning and induced spawning of the red drum (Sciaenops ocellata ) under varied conditions of photoperiod and temperature. Proc. World Maricult. Soc. 9:311-332. 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. SAS Institute. 1985. SAS user's guide: statistics, version 5 ed. SAS Institute Inc., Cary, NC, 956 p. Wallace, R. A., and K. Selman. 1981. Cellular and dynamic aspects of oocyte growth in teleosts. Am. Zool. 21:325-343. West, G. 1 990. Methods of assessing ovarian development in fishes: a review. Aust. J. Mar. Freshwater Res. 41:199-222. Wilk, S. J., W. W. Morse, and L. L. Stehlik. 1990. Annual cycles of gonad-somatic indices as in- dicators of spawning activity for selected species of finfish collected from the New York Bight. Fish. Bull. 88:775-786. Abstract. — Little information exists on the biology of the demer- sal shark, Squalus mitsukurii . Re- cently, large numbers of this spe- cies were taken incidentally during research surveys conducted at Southeast Hancock Seamount in the central North Pacific Ocean. The information collected during 1985 to 1988 from these surveys is used to describe the life history, depth distribution, and biology of S. mitsukurii. Bathymetric distributional pat- terns of female and male S. mitsu- kurii differed slightly, although bottom longline catches revealed a depth distribution extending from the summit (260 m) to 740 m for both sexes. Males generally were found deeper than females. In ad- dition, the size of males generally increased with depth whereas no apparent trend was observed for females. Reproductive parameters for both sexes are presented. Males tended to reach maturity at smaller sizes than did females. Gravid fe- males had broods of up to six uter- ine embryos. Length of young close to parturition was 21-26 cm. Tentative estimates of age and growth were made from dorsal spine increment counts. Maximum ages were 27 years for females and 18 years for males. Females exhib- ited more rapid growth than males after about age 9. The diet of S. mitsukurii in- cluded both benthic and mesope- lagic prey. Fishes, cephalopods, and crustaceans were the major compo- nents of the diet. Comparison of the biological characteristics suggest that this species is probably typical of other slow-growing, low fecund members of the genus Squalus. The 50% de- cline in catch rates observed dur- ing this study suggests that the number of S. mitsukurii on the sea- mount declined dramatically, pos- sibly as a result of overfishing. Biology and population characteristics of Squalus mitsukurii from a seamount in the central North Pacific Ocean Christopher D. Wilson Alaska Fisheries Science Center. National Marine Fisheries Service. NOAA 7600 Sand Point Way NE, Seattle. Washington 98 1 I 5 Michael R Seki Honolulu Laboratory, Southwest Fisheries Science Center National Marine Fisheries Service. NOAA 2570 Dole Street. Honolulu. Hawaii 96822-2396 Manuscript accepted 23 February 1994. Fishery Bulletin 92:851-864. The discovery of large stocks of the pelagic armorhead, Pseudopenta- ceros wheeleri, on seamounts of the southern Emperor-northern Hawai- ian Ridge by Soviet fishermen in 1967 signaled the inception of a large, intense foreign trawl fishery for this species during the early 1970s (Uchida and Tagami, 1984). During 1967-75, for example, nearly one million metric tons of pelagic armorhead were taken from this area by Japanese and Russian trawlers (Boehlert and Genin, 1987). By the mid-1970's, catch rates of pelagic armorhead had de- clined dramatically, and commercial fishing for the species effectively ceased by 1984. The National Ma- rine Fisheries Service (NMFS) ini- tiated research stock-surveys of the area in 1985, and in August 1986 a six-year fishing moratorium was enacted (NMFS, 1986). Results from the NMFS surveys provided information to describe the population dynamics of P. wheeleri (Somerton and Kikkawa, 1992). However, the population biology of several other fish species which were caught incidentally in large numbers during the NMFS surveys was largely unknown. One example was the demersal shark, Squalus mitsukurii, which represented the largest bycatch in the NMFS sur- vey data (Somerton1). Little is known about the life his- tory and population dynamics of S. mitsukurii (Compagno, 1984) in contrast to the more cosmopolitan congener S. aeanthias which has been extensively studied (Com- pagno, 1984; Ketchen, 1986). Squa- lus mitsukurii is broadly distributed in the Pacific and Indian Oceans (Compagno, 1984; Parin, 1987). Specimens resembling S. mitsu- kurii have also been taken in the Atlantic although their taxonomic status is unclear (Compagno, 1984). Squalus mitsukurii is known to in- habit the waters around various is- lands and seamounts in addition to coastal waters (Parin, 1987; Tan- iuchi et al., 1993). Litvinov (1990) reported on several aspects of the biology of S. mitsukurii from 117 specimens taken from the Sala-y- Gomez Seamounts in the Southeast Pacific. Off southeast Africa, Bass et al. (1976) presented limited in- formation on the life history of S. mitsukurii, which had earlier been identified as S. blainvillei (Bass et 1 Somerton, D. National Marine Fisheries Service, Seattle, WA 98115. Personal commun., 1992. 851 852 Fishery Bulletin 92(4). 1994 al., 1986). Whether populations of S. mitsukurii from different geographic areas exhibit the high variabil- ity in life history characteristics reported for S. acanthias (Ketchen, 1972, 1986; Compagno, 1984; Nammack et al., 1985) is unclear. The aim of the present study was to provide information on the bi- ology and population characteristics of S. mitsukurii using a larger number of specimens than had previ- ously been available, and thus elucidate its role as a member of the unique fauna associated with sea- mounts in the central North Pacific Ocean. Methods Squalus mitsukurii specimens were collected aboard the NOAA ship Townsend Cromwell during nine cruises to Southeast (SE) Hancock Seamount from January 1985 to November 1988 (Table 1). The sea- mount is located on the northern Hawaiian Ridge in the central North Pacific Ocean at lat. 29°48'N and long. 179°04'E (Fig. 1). It has a circular, flat-topped summit with an area of about 4.5 km2 at a depth of 260 m. The seamount flanks have an average slope of 0.22 and reach bottom depths of 5,200 m about 22 km away from the summit (Brainard, 1986). Water temperatures generally are 13°-15°C at 260 m depth (i.e. summit) and decrease to 4°-6°C by 750 m (Brainard, 1986). Most S. mitsukurii were collected with bottom longline gear (Somerton and Kikkawa, 1992; Shiota2), although a few specimens were also taken by bottom trawl, handline, vertical longline, and one set with a 25-mm square mesh bottom gill net (Table 1). Only data from hook-caught specimens were used (except for the food habits portion of the study) to avoid any bias due to differences in gear selectivity. Sets were conducted primarily during daylight hours, between the summit depth and 744 m. All specimens of S. mitsukurii from a set, or a ran- dom subsample from large catches, were processed. Fish were sexed and weighed to the nearest 10 g. 2 Shiota, P. M. 1987. A comparison of bottom longline and deep- sea handline for sampling bottom fishes in the Hawaiian archi- pelago. Honolulu Lab., Southwest Fish. Cent. Natl. Mar. Fish. Cent. Admin. Rep. H-87-5, 18 p. 35°N- 30° 25°- 20°- % SE Hancock Seamount CD %. Colahan Smt a » Kure Atoll , % Midway Island **•>»* PACIFIC OCEAN 179-03'E OS •S., •5 Main Hawaiian Islands 175°E 180° 175°W 1 m 165° 160° 155° Figure 1 Location of SE Hancock Seamount. Wilson and Seki: Biology and population characteristics of Squalus mitsukuni 853 Table 1 Catch by gear type, effort by bottom longline gear, and catch per unit of effort (CPUE; bottom longline catch per total number of hooks minus those with pelagic armorhead, Pseudopentaceros wheeleri) for Squalus mitsukurii from research operations at the SE Hancock Seamount during 1985-88. A dash indicates a particular gear type was not used on a cruise. Date Catch (no. I Effort (no. hooks) Bottom trawl Handline Vertical longline Gill net Bottom longline Total Total minus armorhead CPUE Winter 1985(1/30-2/25) 1 66 — — 335 2,001 1,607 0.208 Summer 1985 (6/19-7/16) 0 10 56 521 2,659 1,995 0.261 Summer 1986 (8/11-9/9) 8 — 42 506 6,568 4,208 0.120 Fall 1986(10/31-11/12) 0 — — — 339 2,783 1,944 0.174 Spring 1987(4/11-28) 20 5 — — 218 3,218 2,243 0.097 Summer 1987(8/8-25) 19 — — — 228 4,182 3,646 0.063 Winter 1988(1/12-30) 4 — — — 253 3,546 2,842 0.089 Summer 1988 (7/13-8/21) — — — — 182 5,500 4,171 0.044 Fall 1988(10/26-11/8) — — — — 219 4,067 2,921 0.075 Total length (TL) was determined to the nearest 1 mm by placing the shark on its side and measuring from the anterior tip of the snout to the posterior edge of the upper lobe of the caudal fin in the "natu- ral" upright position. For some specimens, body length (STL) was measured from the anterior tip of the snout to the posterior edge of the upper lobe of the caudal fin after the lobe was depressed to a posi- tion in line with the body axis; fork length (FL) was also measured from the snout to the fork or middle of the caudal fin. Simple linear regression relation- ships among the three length measurements were determined to facilitate comparisons with other pub- lished studies. The relationships are TL= 1.08 x LF + 1.67 (^=0.99, rc=342); TL = 0.95 x STL + 0.24 (^=0.99, /i=212); STL=1.01 x FL + 6.31 (rM).91, rc=463). A nonlinear estimation procedure (Wilkinson, 1988) was used to fit individual length and weight (WT; in grams) data to the relationship, WT = a x TLb. Weight-length relationships between sexes were evaluated with analysis of covariance (ANCOVA) on the log-transformed data. Catch per unit of effort (CPUE) for S. mitsukurii was determined by using only specimens taken by bottom longline for several reasons. The bottom trawl, handline, and gillnet gear did not have fish- ing effort comparable with the bottom longline, and the former data were too limited to construct indi- vidual time series for each gear. Furthermore, the vertical longline gear was used to define the vertical distribution of pelagic armorhead above the summit; therefore only a few hooks of uncertain number were on the summit and available to S. mitsukurii. Finally, specimens from gear other than bottom longline accounted for only 8% of the total numbers caught. Therefore, it was unlikely that the exclusion of these data from the CPUE calculations would sig- nificantly bias our results. Fishing effort on S. mitsukurii was expressed in number of hooks set, minus hooks that caught pelagic armorhead. Pelagic armorhead respond more quickly to the bottom longline than do S. mitsukurii (Somerton3). Thus, excluding those hooks occupied by pelagic armorhead, although not removing all bias from the species-gear interaction, likely improved estimates of fishing ef- fort on S. mitsukurii over the unadjusted value. An estimate of the initial exploitable biomass of S. mitsukurii was made with the Leslie model (Ricker, 1975) in which a linear function was fitted to CPUE and cumulative catch data from all cruises. The func- tion has a slope equal to catchability (q) and inter- cept equal to the product of q and the initial exploit- able biomass. The model assumes that changes in CPUE over time are due to fishing and that other sources of losses and additions to the population are relatively minor or in balance (e.g. natural mortal- ity and recruitment). The 95% confidence limit on the abundance estimate was calculated by using the method of Polovina ( 1986). First and second dorsal spines were collected dur- ing the summer of 1986 for ageing S. mitsukurii with procedures outlined by McFarlane and Beamish (1987). All age results were based on the second 3 Somerton, D. National Marine Fisheries Service, Seattle, WA 98115. Personal commun., 1992. 854 Fishery Bulletin 92(4), 1994 dorsal spine which was considered most suitable for ageing (also see Litvinov, 1990). All ridges on the surface of the mantle covering the spine were counted by a single reader with a low-power dissecting scope and methods described by Ketchen (1975) and Beamish and McFarlane (1985). No ridges were grouped and counted as a single "annulus" as was done in several earlier studies (e.g. Holden and Mead- ows, 1962). Loss of increments due to abrasion of the tip of the spine has been reported for other species of Squalus (e.g. Ketchen, 1975), but this did not occur in the present study. For example, when a worn spine was observed, the worn area was confined to the re- gion of the spine tip having a diameter of <3 mm. However, unworn spines having spine base diameters of <3 mm had not yet formed any increments (the spine base diameter of a late-term uterine embryo was 2.0 mm). Validation of the annual nature of spine increment formation (following the methods described by McFarlane and Beamish [1986, 1987]) was not pos- sible. Therefore increment counts from spines re- ported here must be considered as tentative esti- mates of age. Nonetheless, validation studies con- ducted on S. acanthias have verified that spine in- crements do represent annual marks (Beamish and McFarlane, 1985; Tucker, 1985). Individual length-at-age data for each sex were fit- ted with a nonlinear estimation procedure (Wilkin- son, 1988) to the von Bertalanffy growth model (Ricker, 1975): L, =L„(l-e<-*"-'",)), where L, is length at age, L^ is asymptotic length, k is the growth coefficient, and r0 is the theoretical age when L( = 0. Reproductive data were first collected from shark specimens during the summer of 1986. For females, counts were made of mature ovarian eggs (greater than about 2.5 cm diameter), candled embryos (i.e. gelatinous uterine capsules containing embryos in early stages of development), and embryos free in the uteri. Females possessing any or all of these re- productive products were considered sexually ma- ture. Sex and TL of uterine embryos were also re- corded. Ketchen ( 1972) determined that a period of rapid increase in clasper length indicates the onset of sexual maturity in male S. acanthias. A similar allometric growth phase between clasper length and TL was observed for S. mitsukurii. Thus, the right clasper length was measured to the nearest 1 mm from the body juncture to the clasper tip. To determine the size at 50% sexual maturity for both sexes, data were fitted to the logistic function by using an iteratively weighted (i.e. inverse of vari- ance) nonlinear estimation procedure (Wilkinson, 1988) and evaluated at 50% (Somerton, 1980). The logistic equation is defined as Y=l/(l+AeBX), where Y is the proportion of animals sexually ma- ture, X is the midpoint of a length class, and A and B are parameters defining the curve. Before the logis- tic equation was fitted to length data for males, the data were transformed to approximate the logistic pattern of growth. For example, clasper length (CD plotted against fish TL (not shown) produced a roughly sigmoid curve, although at large and small fish sizes, clasper length continues to increase with size of the fish. To flatten the ends of the curve as required by the logistic equation, clasper length was expressed as a proportion offish TL and normalized to values between 1 and 0 (i.e. with a=CL/TL, the quantity Y'=(a-amin)/(amax-amin) was plotted against fish TL). To obtain information on feeding habits, the stom- ach contents from 251 S. mitsukurii caught on longlines and 42 fish caught in gill nets during the summer of 1986 were examined. Longline-caught fish ranged from 20.6 to 79.5 cm TL ( mean length L =50.4 cm) and gillnet-caught fish ranged from 49.5 to 78.9 cm TL ( L =65.5 cm). Samples from the bottom longline and gillnet gear were compared to examine the potential feeding bias that might occur if attrac- tion to baited longlines varied as a function of stom- ach fullness for S. mitsukurii. For most of the fish, stomachs were extracted upon capture and preserved in 10% formalin until exami- nation in the laboratory. However, if time permitted, stomachs were examined at sea for the presence of food items; empty stomachs were noted and discarded (those with food items were saved). In the labora- tory, samples were sorted, counted, and identified to the lowest possible taxon. Food items were sorted to taxa, blotted dry, and weighed to the nearest 0.1 g. Stomach contents data were analyzed for frequency of occurrence, numerical abundance, and gravimet- ric proportions of prey items to quantitatively de- scribe the diet and feeding habits of S. mitsukurii (Hyslop, 1980). Prey items that were attributed to the presence of the research vessel (i.e. bait, galley refuse, or processed fish offal) were not included in the analyses; stomachs containing only those items were considered empty. To examine diel feeding be- havior, samples were grouped by time of capture: 0600-1200 (n=39), 1200-1800 (ra=86), 1800-2400 (n=64 ), and 0000-0600 ( n =62 ). Chi-square ( y? ) analy- ses of 2 x 2 contingency tables were used to test for Wilson and Seki: Biology and population characteristics of Squalus mitsukurii 855 differences in stomach fullness (defined simply in terms of presence versus ab- sence of food) between time blocks. Sun- rise was at about 0545 h and sunset at 1930 h. The x2 test was similarly used to identify differences in stomach full- ness between longline- and gillnet- caught fish from a single time block (1800-2400 h). Results Abundance A total of 1,392 female, 1,539 male, and 7 unsexed, hook-caught S. mitsukurii were collected. CPUE was highest in 1985 and declined by more than 50% by spring 1987 (Table 1). The CPUE data plotted as a function of cumula- tive catch (C) appeared generally lin- ear with a negative slope (Fig. 2); coef- ficients of the fitted model were CPUE = -6.87 x 10"5 x C + 0.2498 (r2=0.76). The model estimate of the initial exploit- able population was 3,641 fish (±1,954). Based on this estimate, about 80% (±55%) of the initial population of S. mitsukurii had been removed by fishing. Both sexes of S. mitsukurii were caught over the full depth range sampled. The median depth of occur- rence for males was usually greater than that of females (Table 2). Size Female S. mitsukurii were on average longer and reached larger maximum sizes than males (Fig. 3). Maximum lengths recorded were 91 cm for fe- males and 82 cm for males. Differences were also detected in length-weight data between female and male S. mitsukurii (ANCOVA, P<0.05). Non- linear fits of length-weight growth curves between sexes diverged at about the size at 50% maturity for male sharks (see Reproduction). Length-weight parameter estimates were a = 1.7 18 x 10"2, 6=2.687 for males and a=3.773 x 103, 6=3.089 for females. The size of male sharks increased with depth. For example, for each cruise and sex, fish were divided into either a shallow or deep subgroup, depending on whether they were caught above or below the median depth of occurrence for that group on that cruise. Estimates of the median length for males from the deep group were Table 2 Maximum and median depth (m) of occurrence for female and male Squalus mitsukurii , median length (cm) for each sex above and below median depth of occurrence and statistical significance between me- dian depths or lengths (Mann- Whitney U test, *P<0.05, **P<0.001). Median length Maximum Median above/below Sampling period Sex depth depth median depth Winter 1985 F 443 302** 66.0/61.0 M 459 327 59.9/58.7 Summer 1985 F 744 309* 58.4/50.0 M 744 327 52.6/58.5* Summer 1986 F 556 260* 41.8/45.6 M 454 269 43.5/50.9** Fall 1986 F 468 291* 54.4/51.2 M 494 298 52.1/54.1 Spring 1987 F 384 293 48.9/47.9 M 518 269 46.8/54.2 Summer 1987 F 483 283 43.2/45.4* M 459 283 40.5/46.2** Winter 1988 F 512 272 50.3/56.0* M 446 276 52.2/58.6* Summer 1988 F 569 411 50.2/52.3 M 569 448 46.0/55.2** Fall 1988 F 430 274 56.0/50.9 M 382 278 50.0/52.2 0.30 0.25 0.20 0.15 0.10 0.05 0.00 500 1000 1500 2000 2500 3000 Cumulative Catch (number) Figure 2 Catch per unit of effort (CPUE) for each sampling period (filled circles) and predicted CPUE based on the Leslie model (solid line), and plotted as a function of cumulative catch for Squalus mitsukurii from SE Hancock Seamount during 1985-88. 856 Fishery Bulletin 92(4). 1994 females males 20 10 0 10 0 10 *- 20 0.05). No well-defined parturition season was detected either from a large increase in catches of small (i.e. 21-26 cm), free-living fish (Fig. 3) or by the absence of large, near-term embryos in females (Fig. 6). However, the smallest uterine embryos were found in fall and winter and were absent in spring and summer (Fig. 6), which is indicative of a reasonably well-defined seasonal production cycle for young. Specimens as small as 21 cm were captured from longline gear in summer 1986. This confirms that 21-26 cm long embryos would be close to parturition. The only significant correla- tion between numbers of "large," "small," or total uterine embryos, candled embryos, or large ova- rian eggs was a positive association between num- bers of uterine embryos and large ovarian eggs (Pearson's r=0.62, P<0.05). Size at sexual maturity for male S. mitsukurii was estimated indirectly from 812 specimens. The estimated size at 50% sexual maturity was 48 cm (Fig. 5). This corresponds to a tentative age of about 4 years based on the von Bertalanffy growth equation. Feeding Overall, 101 of the 293 (34.5%) stomachs exam- ined contained prey. Stomach contents averaged 2.0 prey items (SD=1.4) and weighed 4.5 g (SD=9.3 g). Fishes, cephalopods, and crustaceans were the Age (years) Figure 4 Fitted von Bertalanffy growth curves and mean length- at-spine increment counts for female (dashed line, filled circles) and male (solid line, open circles) Squalus mitsukurii. Vertical lines represent 95% confidence inter- vals. The von Bertalanffy growth parameters are defined in the text. 1 .0 2? 0.8 O 2 0.6 c o T 0.4 o Q O £ 0.2 0.0 7r„ o o 1 .,-*+-'- 1 /. • 30 40 50 60 70 80 90 Total Length (cm) Figure 5 Fitted logistic function and proportion mature at each size- class for female (filled circles) and male (open circles) Squalus mitsukurii. Size class intervals are 2 cm. Points representing the three largest size classes for males ap- pear as outliers but were not eliminated because few ob- servations comprised these points ( ;? =6, 1 , 1 ). Thus we were uncertain whether they accurately represent the relation- ship between clasper and total length for large fish (which we had insufficient data to define) or whether the points were a result of measurement error. Regardless of whether these "extreme" values were removed or left in the data set. they had relatively little influence in fitting the func- tion (see text). 858 Fishery Bulletin 92(4), 1994 Table 3 Mean length, standard deviation (SD), and number of gravid Squalus mitsukurii specimens as a funct on of number ol large ovarian eggs and candled and uterine embryos. Number of eggs or candled or uterine embryos Maternal statistics 1 2 3 t 5 6 Large ovarian eggs Mean length 68.0 71.1 70.1 72.3 74.4 75.1 SD — 2.49 3.97 2.96 3.18 2.91 Number 1 4 33 35 18 4 Candled embryos Mean length 70.5 74.6 72.0 74.0 76.1 76.2 SD — 5.30 4.03 2.17 2.58 — Number 1 2 14 17 5 1 Uterine embryos Mean length 73.1 68.1 72.4 74.4 76.1 79.5 SD 3.76 1.63 3.89 2.24 2.09 — Number 4 2 21 18 11 1 > 'J c □ CD 4 0 2 0 20 ■1 0 fall 88 '-J1 -^ 20 c of the prey organisms, but only 3.6% of the total aggre- gate weight. In particular, euphausiids and the lophogastrid mysid Gnathophausia longi- spina, were commonly found. Cephalopods, predominantly digestion-resistant squid beaks and eye lenses, were present in 45.1% of the stomachs, representing 26.4% of the total prey items and 19.9% of the prey weight. Remaining prey items included pelagic tuni- cates, coelenterates, and unidentified remains. Gut fullness was evaluated as a function of diel feeding activity and gear type. For longline data, no significant difference was found between the two daytime (x~=0.001, P>0.1) or two nighttime blocks (x2=3.822, P>0.1). However, when longline data from daytime blocks were pooled and compared with pooled nighttime blocks, the proportion of empty stomachs was significantly greater during the night ( 87% ) than during the day (46% ; x2=26.250, P<0.001 ). Longline-caught fish had a significantly greater pro- portion of empty stomachs than did fish caught by bottom gill net ', co, and Lx. In the case of the CV, relative length at age seems to be the most important variable and the growth parameter vari- ables were not selected for the models with five or less variables. Models with more than five indepen- dent variables are not shown as there was little fur- ther improvement in the amount of variation explained. The influence of growth parameters can be seen in three-dimensional smoothed plots of the SD against relative length and Lx (Fig. 1), the SD against rela- tive length and (J)' (Fig. 2), and the SD against rela- tive length and co (Fig. 3). Magnitude of the variabil- ity of mean length at age generally increases with L„, <(>', and co. In contrast, no growth-parameter-re- lated trends were found in plots involving CV. For example, in the plot of CV against relative length and K (Fig. 4), relative variability consistently de- creased with size for all values of K. Coefficent of variation and variation in CV decreased with increased relative length in all regressions for Table 1 Examples of multiple linear regression models with the SD as the independent variable («=3,050). 0' is the growth performance index, Lx is mean length-at-age, L;2 is the square of Lt, RL is relative length (LJLJ), A95 is the age corresponding to 0.95Lm and co is the Gallucci and Quinn (1979) growth parameter. MSE = the mean square error. Model MSE R2 SD = -7.341 + 3.642(|>' 1.99 0.62 SD = -4.769 + 2.415' + 0.022L,. 1.66 0.68 SD = -2.739 + 1.9520' + 0.028L - \A19RL 1.59 0.07 SD = -3.532 - 0.049A95 + 0.092L - 4.280#L - 0.0002L 2 1.49 0.72 SD = 2.771 - 0.037A95 + 0.085L, - 3.922RL - 0.0002L,2 + 0.023w 1.47 0.72 Table 2 Examples of multiple linear regression models with the Vas the independent variable (n=3,050l at age, RL is relative length (LJLji, A is age, A95 is the age corresponding to 0.95L_, co is the G (1979) growth parameter, Lx is the von Bertalanffy growth parameter. MSE = the mean square L is mean allucci anc error. length Quinn Model MSE R2 V = -6.314 + 0.0478L, 243.62 0.58 V = -2.798 + 0.576L, - 1.516A 224.42 0.61 V = 2.582 + 0.421L - 19.33RL + 0.464co 216.45 0.62 V = 1.198 + 0.472L - 16.80i?L + 0.379co - 0.032A2 213.94 0.63 V= 7.587 + 0.434L, - 11.16RL - 0.025A2 - 0.383A95 + 0.098L, 211.12 0.63 Table 3 Examples of multiple linear regression models with the CV as the independent variable (rc=3,050). RL is relative length (LJLJs, A is age, A95 is the age corresponding to 0.95Ln, AA^ is age divided by Aq5, and AMAXA is age divided by the maximum observed age. MSE = the mean square error. Model MSE R2 CV = 16.38 - 12.96/cL 16.50 0.31 CV = 22.09 - 17.36KL - 0.171A95 13.95 0.42 CV = 23.39 - 21.32AL - 0.173A95 + 3.614AA95 13.56 0.44 CV = 22.97 - 19.91 RL - 2.34AMAX4 + 4.34AA95 - 0.155A95 13.41 0.44 CV = 23.55 - 20.53RL - 2A11AMAXA + 3.49AA95 - 0.179Ag5 + 0.11A 13.36 0.45 NOTE Erzini and Castro: Measures of dispersion for length-frequency analysis 867 SD 1 0 - 8 - 6 - J / / / l~lf~j~7-~-L/ //"/-/—/ //^S>0\/ ~7~ ' 5 ° 2 /_/ / ////// III 1 l~f~lr-~J^ / //"--/- / /7 X 120 0 //////~/T~~r~r~~f~~l-~L-LJ / ///~~tC l/r-~l 1 / 9 ° 73^ ~7^^^^^-^Z2vv7///V-~Z^o^w;/ / 60 Figure 1 Smoothed surface graph of the standard deviation (SD) of mean length at age as a function of relative length (RL) and the asymptotic maxi- mum length Lx. SD RL Figure 2 Smoothed surface graph of the standard deviation (SD) of mean length at age as a function of relative length (RL) and the growth perfor- mance index <$>'. 868 Fishery Bulletin 92(4), 1994 SD 2 0 - 1 5 ■ 1 0 5 x^c^^^^^^7~ZZZ^/-5?v ^\ /**r~/~ — 7\ /C \^ — /\ /n. fQ^\j£^~~~v£^^ 7"~ ' ° / N. v/*""/"/ /""vC yK.^s/^0?^O!^^C^V /» » "^AX^yOSQy/" « rl i 0 i — -^_v u) 0 Figure 3 Smoothed surface graph of the standard deviation (SD) of mean length at age as a function of relative length and the parameter co. (Gallucci and Quinn [1979] growth parameter.) cv 35 - 30 - 25 - 2 0 - 1 o - / ^^^^^^^^^^~^^~?'^~y~ ' ° / ^*^^^^>^ / ° a 5 - / /o.« 0 - 0 /• RL 0 2 ~ ~~~r~-~^_ / K 0 0 Figure 4 Smoothed surface graph of the coefficient of variation (CV) of mean length at age as a function of relative length (RL) and K, the growth rate. NOTE Erzini and Castro: Measures of dispersion for length-frequency analysis 869 Table 4 Regressions of the coefficient of variation (CV) against relative length (RL) for data grouped by the growth parameter. K Intercept Slope n MSE R2 P a 0.05-0.099 13.39 -12.51 326 11.49 0.31 0.0001 b 0.10-0.149 15.51 -14.04 611 9.80 0.41 0.0001 c 0.15-0.199 20.05 -17.50 602 16.10 0.44 0.0001 d 0.20-0.249 20.15 -19.22 220 11.64 0.48 0.0001 e 0.25-0.299 21.53 -19.13 214 12.32 0.48 0.0001 f 0.30-0.349 22.65 -20.69 260 11.05 0.58 0.0001 g 0.35-0.399 19.27 -15.87 221 10.70 0.34 0.0001 h 0.40-0.449 23.09 -20.44 181 5.13 0.61 0.0001 i 0.45-0.549 22.02 -16.14 124 13.43 0.37 0.0001 j > = 0.55 23.40 -19.33 220 11.59 0.43 0.0001 data grouped by the growth parameter K (Fig. 5, A-J). With the exception of group- ings for K<0.15 (Fig. 5, A and B), which have smaller intercept and slope values, the re- gressions are similar (Fig. 5, C-J). The re- gression line and the 95% confidence inter- vals are also shown and the associated sta- tistics are given in Table 4. The slopes of the regressions are all significantly different from0(P<0.001). Discussion A number of LFA methods, especially those that estimate parameters by maximum like- lihood methods, allow constraints on mea- sures of dispersion. For example, the simplex method of Kumar and Adams (1977) incor- porates linear constraints on the standard deviations (SD's) of normal components. The SD's can be equal or fixed and the coefficient of variation (CV) can be fixed or constant in the program MIX (Macdonald and Pitcher, 1979; Macdonald and Green, 1985). The SD's can be linear functions of mean length or of age in the Schnute and Fournier (1980) method. MULTIFAN (Otter Software, 1988) allows age-dependent or length-dependent trends in SD's. A common CV between 0.01 and 0.5 for all lengths at age or SD's that increase linearly with mean length was pro- posed for LFA constraints by Liu etal. (1989) Figure 5 The coefficient of relative variation as a function of relative length for data grouped by K. RL is relative length (LJLx, length-at-age divided by Lj. The interval classes and the regression sta- tistics are given in table 4. Parallel lines are 95% confidence intervals. c o CO > C o it (D o O 0 0 0 2 0 4 06 01 10 12 0 0 0 : 0 4 06 OS 10 12 00 02 04 06 Oi 10 12 00 02 04 06 0B 10 12 ) 0 0 2 0 2 0 0 0 2 0 i RL RL 870 Fishery Bulletin 92(4). 1994 In addition to these methods which allow specific constraints, some iterative methods require starting or initial values for some parameters, such as num- ber of components, corresponding mean lengths at age, proportion in each age class and SD's of the com- ponent distributions (e.g. Akamine, 1982, 1984, 1985). Our results can be used to select appropriate con- straints and starting values for measures of disper- sion for LFA methods. We have shown that the mag- nitudes of SD and V are dependent to a large extent on life history parameters. Therefore, if the LFA user has estimates of growth parameters, the multiple linear regression models in Table 1 can be used to estimate the SD for the species and size in question. However, in most cases the objective of LFA is to estimate growth parameters, which are therefore not available for input into the predictive models. In this case, the CV may be more useful as a constraint. While the magnitudes of SD and V of mean length at age are related to characteristics of each species, rela- tive variability in length at age (CV) is similar in species that differ greatly in life history parameters. Furthermore, while there are no consistent age- and size-dependent trends in absolute measures of vari- ability, relative variability decreases in a predictable manner in almost all cases. This was confirmed in a previous investigation of the shapes, magnitude, and age and size dependence of length-at-age distributions of marine fishes (Erzini, 1994). Analysis of 415 individual data sets showed that in 97% of the data sets the CV was nega- tively related to relative length at age, and the slope was significant (P<0.05) in 53% of the sets. CV val- ues were similar for all species. A negative relation- ship between CV and size and decreasing variation with size are to be expected because changes in vari- ability with growth are typically of smaller magni- tude than changes in size with growth. In contrast, although there was no dominant size- dependent or age-dependent trend for the SD, the most common pattern was that of increasing vari- ability to a maximum at an intermediate age or size. This trend for increasing variability to a maximum at an intermediate size is illustrated in Figure 1, where the SD is plotted against relative length and asymptotic maximum length (LJ. It is particularly evident for species with large Ln values. In conclusion, we believe that the practical impli- cations for LFA are that these empirically derived relationships between measures of dispersion, size, age, and life history parameters can be used to se- lect starting values and to impose constraints on measures of dispersion corresponding to particular lengths at age. This is useful as there are no well established rules or guidelines for this process, which consequently has been highly subjective and depen- dent on each LFA user. The choice of model depends on the availability of the data for the independent variables of the mod- els. In the absence of any such data, the simplest model of the CV as a function of relative length can be used. As a preliminary step, length-frequency dis- tributions should be examined and the number of possible component distributions and modes that may represent mean lengths at age identified visu- ally. An estimate of Ln obtained from the literature or on the basis of the maximum observed size can be used to convert lengths to relative lengths. The esti- mated CV values and their corresponding confidence intervals for these modes can then be estimated with the models presented in this study. One possible ap- proach is to use the estimated CVs as starting values and the confidence intervals as lower and upper con- straints. Such a strategy would provide realistic start- ing values, reasonably narrow constraints, and would improve the often arbitrary choices which are made. Acknowledgments We would like to thank the Scientific Editor and an anonymous reviewer for their suggestions which greatly improved the manuscript. Literature cited Akamine, T. 1982. A BASIC program to analyze the polymodal fre- quency distributions into normal distributions. Bull. Jpn. Sea Fish. Res. Lab. 33:163-166. 1984. The BASIC program to analyze polymodal frequency distributions into normal distributions with Marquardt's method. Bull. Jpn. Sea Fish. Res. Lab. 34:53-60. 1985. Consideration of the BASIC programs to ana- lyze the polymodal frequency distribution into nor- mal distributions. Bull. Jpn. Sea Fish. Res. Lab. 35:129-160. Basson, M, A. A. Rosenberg, and J. R. Beddington. 1988. The accuracy and reliability of two new meth- ods for estimating growth parameters from length- frequency data. J. Cons. Int. Explor. Mer 44:277- 285. Castro, M., and K. Erzini. 1987. Comparison of two length frequency based methods used to obtain growth and mortality pa- rameters using simulated samples with varying recruitment patterns. Fish. Bull. 86:645-653. Erzini, K. 1990. Sample size and grouping of data for length frequency analysis. Fish. Res. 9:355-366. NOTE Erzini and Castro: Measures of dispersion for length-frequency analysis 871 1991. A compilation of data on variability in length- at-age of marine fishes. Fish. Stock Assess. CRSP Working Paper 77, 100 p. 1994. An empirical study of variability in length- at-age of marine fishes. J. Appl. Ichthyol. 10:17^41. Gallucci, V. F., and T. J. Quinn. 1979. Reparametrizing, fitting, and testing a simple growth model. Trans. Am. Fish. Soc. 108:14-25. Kumar, K. 1 )., and S. M. Adams. 1977. Estimation of age structure offish populations from length frequency data. In W. Van Winkle (ed.), Assessing the effects of power-plant induced mortality on fish populations, p. 256-281. Perma- gon Press, Oxford. Liu, Q., T. Pitcher, and M. Al-Hossaini. 1989. Ageing with fisheries length-frequency data, using information about growth. J. Fish Biol. 35 (Suppl. A):169-177. Longhurst, A. R., and D. Pauly. 1987. Ecology of tropical oceans. Academic Press, Inc. Macdonald, P. D. M. 1987. Analysis of length frequency distributions. In R. C. Summerfelt (ed. ), Age and growth offish. Iowa State Univ. Press, p. 371-395. Macdonald, P. D. M., and T. J. Pitcher. 1979. Age groups from size-frequency data: a ver- satile and efficient method of analyzing distribu- tion mixtures. J. Fish. Res. Board Canada 36:987- 1001. Macdonald, P. D. M., and P. E. J. Green. 1985. User's guide to program MIX: an interactive program for fitting mixtures of distributions. Ichthus Data Systems, 28 p. Morgan, G.R. 1987. Incorporating age data into length-based stock assessment methods. In D. Pauly and G. R. Mor- gan (ed.), Length-based methods in fishery re- search: proceedings of the international conference on the theory and application of length-based meth- ods for stock assessment; 11-16 February 1985, Mazzara del Vallo, Sicily, Italy. ICLARM Conf. Proc. 13:137-146. Neter, J., W. Wasserman, and M. H. Kutner. 1983. Applied linear regression models. Richard D. Irwin, Inc., 547 p. Otter Software. 1988. MULTIFAN, user's guide and reference manual. Nanaimo, B.C. Canada. SAS Institute Inc. 1985. SAS® user's guide: statistics. SAS Institute Inc., 956 p. Schnute, J., and D. Fournier. 1980. A new approach to length-frequency analy- sis: growth structure. Can. J. Fish. Aquat. Sci. 37:1337-1351. Migratory characteristics of juvenile ocean-type Chinook salmon, Oncorhynchus tshawytscha, in John Day Reservoir on the Columbia River Albert E. Giorgi Don Chapman Consultants. Inc.. 7981 Redmond. Washington 98052 I 68th Ave NE David R. Miller Benjamin R Sandford Northwest Fisheries Science Center, National Marine Fisheries Service, NOAA 2725 Montlake Blvd. E.. Seattle, Washington 981 12 Both stream-type and ocean-type chinook salmon, Oncorhynchus tshawytscha, are found in the Co- lumbia River system. Ocean-type chinook salmon migrate seaward and enter seawater as subyearlings or zero-age juveniles within a year of emergence, whereas stream-type fish reside in fresh water at least one full year before migrating (Healey, 1991). Yearling stream- type chinook salmon migrate through the mainstem Columbia River and its largest tributary, the Snake River, during the spring months (Raymond, 1979). In con- trast, zero-age ocean-type chinook salmon migrate during the sum- mer, but their migration can extend into autumn. Information regard- ing the migratory behavior of ocean-type chinook salmon in the impounded reaches of the Colum- bia River is limited. Early research showed that even during high-flow years, large numbers of zero-age ocean-type chinook salmon re- mained in John Day Reservoir on the Columbia River for a protracted time compared with stream-type chinook salmon (Raymond et al.1; Sims et al.2). 872 Hydroelectric development has been identified as an important fac- tor that has contributed to de- creased salmon and steelhead (On- corhynchus spp.) production in the Columbia River Basin (Raymond, 1979, 1988; Williams, 1989). Direct mortality of downstream migrant juvenile salmonids is associated with passage through the turbines, spillways, and juvenile bypass sys- tems at dams. Apart from direct mortality, a number of studies have indicated that the creation of im- poundments, altered flows result- ing from electric power demand, and irrigation withdrawals as a result of dam construction have slowed the seaward migration of juvenile salmonids (Raymond, 1969, 1979; Ebel and Raymond, 1976). In an effort to lessen deleterious effects associated with hydroelec- tric dam construction, fisheries managers have developed water management strategies to augment instream flows to provide improved passage conditions for juvenile salmonids during their seaward migration (Northwest Power Plan- ning Council, 1987). Rationale sup- porting these actions is based largely on data by Sims and Ossi- ander3 which described the migra- tory characteristics of juvenile stream-type chinook salmon and steelhead, O. mykiss, within the Snake River and in portions of the Columbia River. They found that increased instream flow volumes during the spring reduced smolt travel time through the hydroelec- tric complex and increased smolt survival. Similar data for ocean- type chinook salmon that migrate during the summer as zero-age ju- veniles are not available. Berggren and Filardo ( 1993 ) sug- gested that increased water veloc- ity increased migration speed for ocean-type chinook salmon and led to increased survival by reducing exposure time to predatory fish and to increasing summer water tem- peratures. There is ample evidence that predatory fish, principally northern squawfish, Ptychocheilus oregonensis, are abundant and con- sume large numbers of juvenile salmonids particularly during the summer in John Day Reservoir (Rieman et al., 1991; Vigg et al., 1991). However, the relationships between flows, migration rate, and survival are uncertain. 1 Raymond, H. L., C. W. Sims, R. C. Johnsen, and W. W. Bentley. 1975. Effects of power peaking operations on juvenile salmon and steelhead trout migrations, 1974. Northwest Fish. Sci. Cent., Natl. Mar. Fish. Serv., Seattle, WA98112-2097. Report to U.S. Army Corps of Engineers, 46 p. 2 Sims, C. W.. R. C. Johnsen. and W. W. Bentley. 1976. Effects of power peaking operations on juvenile salmon and steel- head trout migrations, 1975. Northwest Fish. Sci. Cent., Natl. Mar. Fish. Serv., Seattle, WA 98112-2097. Report to U.S. Army Corps of Engineers, 36 p. 3 Sims, C. W„ and F J. Ossiander. 1981. Migrations of juvenile chinook salmon and steelhead trout in the Snake River from 1973 to 1979: a research summary. North- west Fish. Sci. Cent., Natl. Mar. Fish. Serv., Seattle, WA 98112-2097. Report to U.S. Army Corps of Engineers, 31 p. Manuscript accepted 31 March 1994. Fishery Bulletin 92:872-879. NOTE Giorgi et al.: Migratory characteristics of juvenile Oncorhynchus tshawytscha 873 Developing water management strategies to ben- efit the juvenile stages of ocean-type chinook salmon has become an important issue in the Pacific North- west; however, basic information describing migra- tory characteristics is required before such strate- gies can be designed. We undertook the present in- vestigation to describe the migratory characteristics of ocean-type chinook salmon in John Day Reservoir, a major impoundment on the Columbia River. This paper describes the movement and residence time of zero-age ocean-type chinook salmon within the res- ervoir and examines the relationship between mi- gration time through the reservoir and key environ- mental variables. Study area John Day Dam is a hydroelectric project on the Co- lumbia River at river kilometer (rkm) 345, approxi- mately 200 km east of Portland, Oregon (Fig. 1). The project was constructed and is operated by the U.S. Army Corps of Engineers (COE). The John Day Res- ervoir is the largest impoundment on the river, ex- tending 122 km upstream to the tailrace of McNary Dam, located approximately 52 km downstream from the confluence of the Columbia and Snake rivers. The width of the reservoir ranges from 0.8 to 4.2 km, and its mid-pool depth extends to 48 m. The dam is ap- proximately 1 km in length and is currently fitted with 16 turbines. Methods Migrant zero-age chinook salmon entering the juve- nile fish sampling facilities at McNary Dam were collected from mid-June through August 1981 through 1983. The fish were predominantly a mix- ture of fall and summer races (named for the time of adult returns) from the Columbia River and some small portion of fall races from the Snake River. The yearling chinook migration peaks during May at McNary Dam but can extend from April into June (FPC4). By mid-June more than 95% of the yearlings have passed the dam. During late June some year- lings remain mixed with the zero-age migrants. To minimize the inclusion of the larger yearlings in our experimental groups we used fish less than 110-mm fork length during June. Each week, up to three groups of fish were freeze branded with a unique mark (Mighell, 1969). All fish bearing the same brand were released into the tailrace below McNary Dam at 2100 h on their respective release dates to con- tinue their downstream migration. 4 Fish Passage Center. 1992. Fish Passage Center 1991 Annual Report. Columbia Basin Fish and Wildlife Authority, Portland. OR, 52 p. Purse seine transects rkm 373 :t locatic imbia R rkm 467 rkm 439 rkm 453 \^_ rkm 430 X-^Cr15^ "^\ McNary rkm 407/^=r~4^_/X ^ Dam rkm 389 j^ Y p f ^ rkm 468 G0' t N Washington Oregon i rkm rkm 348^/ John Day -^v Dam ^< rkm 345 357 nse< Cob Purse-seine tra Day Reservoir, 0 10 20 km Figure I ns for sampling zero-age chinook salmon in John ver. 874 Fishery Bulletin 92(4). 1994 Some of the freeze-branded fish were subsequently recovered downstream at John Day Dam. An airlift pump (Sims et al.5) was used to extract fish from the gatewells at Turbine Unit 3; however, it was unknown what proportion of recovered fish represented those passing into the turbine intake. Reliable estimates of that proportion are not available. Each day, col- lected fish were examined and brands enumerated. To provide a relative measure of daily passage at John Day Dam, the daily catch was expanded in propor- tion to the daily total river flow that was discharged through the sampled turbine unit. That proportion varied with prevailing spill volumes and the num- ber of turbine units that were operating. Some wa- ter was also discharged through the navigation locks and fish ladders, but the amount was small, typi- cally less than 1% of the total river discharge (Sims et al.6). The expanded daily catch was referred to as the passage index and was a relative measure of the number offish passing the entire dam. The calcula- tion of the passage index assumed 1) that the pro- portion offish passing the dam through the spillway was equal to the proportion of water spilled, and 2) that the proportion offish entering the gatewells from the turbine intake was relatively constant. For each branded group, we constructed a distri- bution of daily passage indices. The median migra- tion time for each group was estimated as the elapsed time between the known release date at McNary Dam and the date of median passage index distribution at John Day Dam. In addition, we estimated the pas- sage index for the entire population passing John Day Dam each week. Additionally, to characterize the movement pat- terns within John Day Reservoir, we freeze-branded, released, and subsequently recaptured zero-age chinook salmon at fixed cross-sectional transects lo- cated along the length of the reservoir (Fig. 1). We sampled fish with a 305 x 11 m purse seine ( 12-mm stretched mesh, knotless web throughout) aboard an 11-m power-block seiner. At each transect, a seine set was made as close to each shore as possible, al- lowing a minimum depth of 5 m for the seiner; the skiff would extend the net toward shore. A third set was executed at midreservoir. Sampling continued 5 Sims, C. W., J. G. Williams. D. A. Faurot, R. C. Johnsen, and D A. Brege. 1981. Migrational characteristics of juvenile salmon and steelhead in the Columbia River Basin and related pas- sage research at John Day Dam, Vols. I and II. Northwest Fish. Sci. Cent., Natl. Mar. Fish. Serv., Seattle, WA 98112-2097 Re- port to U.S. Army Corps of Engineers, 61 p. 8 Sims. C. W., A. E. Giorgi, R. C. Johnsen, and D. A. Brege. 1983. Migrational characteristics of juvenile salmon and steelhead in the Columbia River Basin — 1982. Northwest Fish. Sci. Cent., Natl. Mar. Fish. Serv., Seattle, WA 98112-2097. Report to U.S. Army Corp of Engineers, 35 p. throughout the summer and autumn until late No- vember each year. Sampling extended from the fore- bay at John Day Dam ( rkm 348 ) to the McNary Dam tailrace (rkm 467). We initially established and sampled nine transects spanning the length of the reservoir (Fig. 1). However, catches were so small at the three locations farthest upstream that we dis- continued sampling those sites halfway through the 1981 sampling period. We cycled through all transects approximately every other week. All fish were anesthetized with MS-222, counted, and exam- ined for marks. Unmarked fish were freeze branded, a subsample was measured for fork length, and af- ter processing, all fish were allowed to recover from the anesthetic and were released. To examine the effects of several key variables on migration rate from McNary to John Day Dam, we used correlation and regression techniques, analyz- ing each year separately and pooled together. The dependent variable was the median migration time (travel time) for each release group. The indepen- dent variables included release date, water tempera- ture, and inverse river-flow volume. We used the in- verse of volume, based on the hypothesis that fish would most likely respond to water velocity (water velocity is the river-flow volume divided by the cross- sectional area) and that fish travel time is related to water particle travel time, which is functionally in- versely related to water velocity. Water temperature and flow were represented by a daily average over the 10-day period following the release date of each marked group. Water temperature and flow data were acquired from the COE. All data were originally reported by Giorgi et al.7 Results Migration timing and migrant size Each year, there was a minor peak in abundance of zero-age chinook salmon passing McNary Dam near the beginning of July and a major peak at the end of July (Fig. 2). In 1982 and 1983, the migration times for the zero-age chinook salmon populations passing John Day Dam were nearly identical. In 1982, 90% of the outmigrants had passed John Day Dam by the week ending 4 September and in 1983, by 26 August. In 1981, the passage distribution was somewhat dis- similar to those of 1982 and 1983; however, the 90th percentile of passage occurred during the week end- Giorgi. A. E., D. R. Miller, and B. R Sandford. 1990. Migratory behavior and adult contribution of summer outmigrating subvearling chinook salmon in John Day Reservoir. Northwest Fish. Sci. Cent., Natl. Mar. Fish. Serv., Seattle, WA 98112-2097. Report to Bonneville Power Administration, 68 p. NOTE Giorgi et al.: Migratory characteristics of juvenile Oncorhynchus tshawytscha 875 £ 250- 200 150 - a 100 - Figure 2 Weekly passage indices of zero-age chinook salmon pass- ing John Day Dam, Columbia River. The dashed, solid, and dotted lines represent 1981, 1982, and 1983, respectively. ing 22 August, the same general time frame as in the following years. Migrant size increased steadily over the course of the sampling period in all years. The mean size ranged from approximately 90 mm in early June to near 145 mm by the end of August. By mid-Decem- ber, the length of fish passing John Day Dam aver- aged approximately 170 mm. The smallest mi- grants observed passing either McNary or John Day Dam were 55 to 60 mm during the month of June each year. The largest migrants approached 225 mm in December. The fish sampled at all three sites (McNary Dam, John Day Dam, and within the reservoir) displayed the same size distributions during the sampling periods. Environmental conditions Each year the trend in river flow was similar: dis- charge consistently decreased during the summer. However, the absolute flow volumes varied con- siderably over the three years of study (Tables 1- 3). The greatest differences in flow volumes were observed each year from mid-June to mid-July. Beyond that period, flow volumes were nearly the same from year to year. The highest flow year oc- curred in 1982 when discharge levels averaged up to 11.11 x 103 cubic meters per second during the last week of June. Overall, the lowest flows oc- curred in 1983. For example, during the last week of June, river discharge volumes were nearly half the maximum level observed in 1982, averaging between 5.91 x 103 and 6.87 x 103 cubic meters per second. Water temperatures were similar among years and displayed the same tendency to increase throughout the summer (Tables 1-3). Early in the summer, water Table 1 Summary of 1981 brand release and recovery data from groups of zero-age chinook salmon marked and released at McNary Dam and recaptured at John Day Dam. Travel time is the number of days required to traverse the reser- voir from McNary Dam tailrace to John Day Dam. The percentiles were calculated from the passage indices. Number offish Flow2 m3sec_1 Travel time (days) Release Passage Temperature1' Percentiles date Released Recovered index' (xlO3) CO in 50 90 6/15 3,325 28 437 9.76 15 16 18 23 6/18 \y.r, t 44 667 9.25 15 14 16 26 6/24 3,458 37 554 7.49 L6 2 10 26 6/29 6,286 38 591 7.16 n; 5 7 17 7/10 10,115 79 840 6.36 L6 10 L9 33 7/16 10,143 65 628 5.94 17 L3 21 37 7/22 10,012 50 526 5.66 L8 5 1 1 22 7/29 12,310 til 624 5.43 L9 7 9 50 8/03 2,512 11 105 5.06 L9 5 6 18 8/10 2,663 15 113 4.66 L9 11 17 98 8/13 2,545 12 81 4.33 20 8 26 126 8/17 2,547 10 63 4.13 21 4 is 24 8/20 2,536 22 145 3.87 L'l 7 19 81 8/26 1,577 6 35 3.56 21 5 13 33 ' The passage index is calculated daily as the ratio of the number recovered to the sampling effort and summed over days. Sampling effort was the average proportion of the total river flow discharged through Turbine Unit 3 during the 10-hour period 2000-0600 h. 2 The average river-flow volume and water temperature over the 10-day period following release of the marked group. 876 Fishery Bulletin 92|4). 1994 Table 2 Summary of 1982 brand release and recovery data from groups of zero-age chinook salmon marked and released at McNary Dam and recaptured at John Day Dam. Travel time is the number of days required to traverse the reser- voir from McNary Dam tailrace to John Day Dam. The percentiles were calculated from the passage in dices. Number offish Flow1' Travel time (days) Release Passage m3sec" Temperature2 Percentiles date Releasee Recovered index' (xlO3) CO 10 50 90 6/24 2,396 7 148 11.11 16 6 9 46 6/26 3,235 17 346 10.92 16 5 13 27 6/29 2,690 9 136 10.44 L6 12 22 92 7/13 3,035 15 181 6.96 18 3 Hi 87 7/15 4,323 i:i 143 6.42 18 7 18 78 7/17 4,012 17 219 6.82 18 6 13 25 7/20 5,001 L6 172 5.80 19 7 17 71 7/22 2,012 1!) 168 5.54 IS 11 31 78 7/27 3,262 33 299 5.46 20 8 19 59 7/29 4,500 11 368 5.43 20 8 24 71 8/03 1,007 7 63 5.37 20 5 34 90 8/05 2,383 29 253 5.10 'JO 7 24 78 8/10 3.000 32 J 5 9 4.52 20 5 12 76 8/13 2,571 :il 247 4.16 20 9 46 68 8/17 3,450 46 321 4.02 20 L2 41 76 8/20 3,005 31 231 3.39 21 7 39 62 8/24 1,467 22 160 3.34 2] 6 35 59 8/27 3,581 35 246 3.17 2] L2 31 46 8/31 1,589 Hi 133 3.70 21 9 23 59 9/03 4,541 Hi 125 3.79 20 9 45 98 ' The passage index is calcu lated daily as the ratio of the number recovered to the sam pling effort and summed over days Sampling effort wa , the average proportion of the total river flow dischar ged through Turbine Unit 3 during the 10-hour period 2000 -0600 h. 2 The average river-flow vo ume and water tempei ature over the 10-day period following release of the marke i group. Table 3 Summary of 1983 brand release and recovery data from groups of zero-age chinook salmon marked and released at McNary Dam and recaptured at John Day Dam. Travel time is the number of days required to traverse the reser- voir from McNary Dam tailrace to John Day Dam. The percentiles were calculated from the passage indices. Number offish Flow2 m3sec_1 Travel time (days) Release Passage Temperature- Percentiles date Released Recovered index' (xlO3) CO L0 50 90 6/16 4,839 11 601 6.87 13 5 11 30 6/23 5,196 2?, 327 5.91 1 1 15 19 26 7/01 5,010 28 421 5.54 16 s L5 19 7/08 1.9HK 35 557 5.60 H, 9 12 24 7/13 5,005 20 333 6.14 16 3 7 23 7/15 5,014 42 i;l-7 5.97 16 4 7 24 7/20 5,019 60 700 6.00 17 7 19 53 7/23 5,009 62 596 5.80 L8 7 29 50 7/27 4,659 11 374 5.71 L8 L2 25 9,S 7/29 5,939 71 621 5.46 18 9 29 83 8/05 4,657 60 499 4.84 19 6 24 115 8/12 4,850 M9 304 4.67 20 6 28 101 8/19 4,878 47 363 4.10 2] t 23 73 8/26 5,641 54 417 3.59 19 5 15 84 9/02 1,855 17 127 3.40 L8 (i 9 59 ; The passage index is calculated daily as the ratio of the number recovered to the sampling effort and summed over days. Sampling effort was the average proportion of the total river flow discharged through Turbine Unit 3 during the 10 hour period li'MHi Otiliil h 2 The average river-flow volume and water temperature over the 10-day period following release of the marked group. NOTE Giorgi et al.: Migratory characteristics of juvenile Oncorhynchus tshawytscha 877 temperatures ranged from 13 to 16°C, then increased steadily during the summer and peaked near 21°C by the end of August. Overall, 1982 was character- ized by slightly higher water temperatures than the other two years. Fish travel time from McNary Dam to John Day Dam For the three study years, a total of 49 freeze-branded groups were released to estimate fish travel time through the reservoir. The number of fish released in each group ranged from 1,007 to 12,310 (Tables 1-3). The estimated median travel time through John Day Reservoir for freeze-branded groups ranged from 6 to 26 days in 1981, 9 to 46 days in 1982, and 7 to 29 days in 1983 (Tables 1-3). Overall, the estimated median travel times were longest in 1982. All individual groups exhibited protracted passage distributions at John Day Dam. The elapsed time between the 10th and 90th percentile of the recap- ture distributions typically exceeded several weeks (Tables 1-3). The fastest moving fish, those repre- sented by the 10th percentile, traversed the reser- voir in 2 to 16 days. The slowest moving fish, those represented by the 90th percentile, took 17 to 126 days to migrate through the reservoir. The linear regression analyses, treating each year separately and pooling all years, did not identify a single model that was applicable to all years. Trans- formation of predictor variables did not improve the model. In fact, for each year, different sets of vari- ables were included in the model constructed by the stepwise procedure. In 1981, the variability in travel time could not be explained by any predictor (Table 4), and none of the predictor variables entered the model. In 1982, only one predictor, release date, was entered into the model. In 1983, two variables, release date and water temperature, were entered Table 4 Regression models derived from stepwise multiple regression. The modelling procedure was applied to median zero-age chinook salmon travel times pre- sented in Tables 1-3. Average water temperature, inverse average flow, and Julian release date were used in the model selection process. Year Model R2 1981 No variables were entered into the model 0.00 1982 Travel time = -53.02 + 0.37 ( release date ) 0.47 1983 Travel time = -1.16 + 5.20 (temperature) -0.34 (release date) 0.46 Combined Travel time = -22.83 + 2.36 (temperature) 0.24 into the model. For the three years combined, only water temperature entered into the model. In all years, strong correlations were observed among the three pre- dictor variables, with r-values ranging from 0.64 to 0.98. Intrareservoir movement Upstream movement offish after branding was regu- larly observed in the reservoir (Table 5). Detailed recapture histories for individual fish were reported in Sims and Miller,8 and Miller and Sims.910 In 1981, 1982, and 1983, the percentages of marked fish that were recaptured at or upstream from the transect of release were 67, 63, and 60%, respectively (Table 5). In each year, upstream movement was observed more frequently than stationary or downstream move- ment. Upstream movements were often pronounced, ranging from 9 to 82 km. Over the three years of study, the duration of the observed upstream move- ments ranged from 6 to 104 days. These observations indicated that the population at large was not con- sistently displaced downstream: rather, a large seg- ment was engaged in pronounced upstream move- ment, or was stationary for extended periods. 8 Sims, C. W., and D. R. Miller. 1982. Effects of flow on the mi- gratory behavior and survival of juvenile fall and summer chinook salmon in John Day Reservoir. Northwest Fish. Sci. Cent., Natl. Mar. Fish. Serv., Seattle, WA 98112-2097. Report to Bonneville Power Administration, 22 p. 9 Miller, D. R., and C. W. Sims. 1983. Effects of flow on the mi- gratory behavior and survival of juvenile fall and summer chinook salmon in John Day Reservoir. Northwest Fish. Sci. Cent., Natl. Mar. Fish. Serv, Seattle, WA 98112-2097. Report to Bonneville Power Administration, 25 p. 10 Miller, D. R., and C. W. Sims. 1984. Effects of flow on the mi- gratory behavior and survival of juvenile fall and summer chinook salmon in John Day Reservoir. Northwest Fish. Sci. Cent., Natl. Mar. Fish. Serv, Seattle, WA 98112-2097. Report to Bonneville Power Administration, 23 p. Table 5 Purse-seine recoveries of marked zero-age chinook salmon that were previously marked and released at various John Day Reservoir sampling transects, 1981-83. 1981 1982 1983 Number of release groups 34 44 32 Number of fish released 14,273 13,126 22,206 Number offish recaptured at transects in reservoir 63 41 111 Proportion recaptured at release site 0.11 0.12 0.16 upstream transects 0.56 0.51 0.44 downstream transects 0.33 0.37 0.40 Upstream recaptures excursion length; range (km) 10-80 16-82 9-82 excursion duration; range (d) 6-75 8-104 6-79 878 Fishery Bulletin 92(4), 1994 Discussion Our analyses indicated that no consistent set of pre- dictors (water temperature, release date, or flow) could explain the travel time of zero-age ocean-type chinook salmon through John Day Reservoir. The predictors for travel time changed each year. The stepwise regression procedure failed to find any sta- tistically significant variables to explain results in 1981, and flow was not a statistically significant pre- dictor in any year. However, strong correlations among all predictor variables suggested that flow was nearly equally as likely a predictor as water tem- perature in 1983 and in the combined years, or as release date in 1982 and 1983. Release date was included as a predictor variable to provide a generic measure to characterize time- based changes in fish development, such as size or physiological changes that progress over the course of the migration period. Since release date entered the model in two of the three years, this suggested that some time-based biological process may have been important. However, the strong correlations among predictor variables in each year limited the utility of such multivariable regression analyses for identifying the importance of any particular variable. Furthermore, in examining bivariate correlations we found no consistent relationships between migration time and any predictor variable. Other measures of migratory behavior should be considered when characterizing the migratory dy- namics of a population. One such measure we con- sidered involved describing the directional intrareservoir movement of fish. We observed that within the body of the reservoir, zero-age ocean-type chinook salmon did not exhibit consistent down- stream movement indicative of a continual, directed seaward migration. The majority of fish that were marked and released at transects throughout the reservoir were recaptured at or upstream from the site of release. This indicated that the population was not consistently displaced downstream passively via current. Based on laboratory observations of coho salmon, O. kisutch, Smith (1982) suggested that smolts in the Columbia River may be oriented mostly head-first upstream during outmigration, thus drift- ing downstream tail-first while being swept seaward. Our results indicate that zero-age chinook salmon do not fit this conceptual model. The protracted reservoir-residence times apparent in our data are not necessarily peculiar to Columbia River stocks. Reimers (1973) studied fall chinook salmon in the Sixes River, Oregon, and suggested the optimum size at ocean entry is about 130 mm for that stock. He noted that this length was attained by juveniles that remained in fresh or estuarine wa- ters for extended periods of time, suggesting that extended freshwater residence is beneficial to zero- age fall chinook salmon. Extended residence of zero- age chinook salmon was observed in the Columbia River during the late 1950s (Mains and Smith, 1964), and even prior to dam construction (Rich, 1922). The absence of a strong relationship between the migration rate and water velocity (flow) for ocean- type chinook salmon contrasts with evidence link- ing travel time to flow (Sims and Ossiander3; Sims et al.6), or developmental (smoltification) state, or both (Giorgi, 1990; Berggren and Filardo, 1993; Beeman et al.11) for migratory yearling stream-type chinook salmon. The effects of smolt development on migratory be- havior of zero-age fish are not clear. Zaugg (1982) cited a number of examples that suggested smolt development might be an important process govern- ing migratory behavior of zero-age fall chinook salmon. In contrast, investigations conducted in the Rogue River, Oregon, indicated smolt development was not a requirement for downstream migration in ocean-type juveniles and its importance in affecting the rate of migration was not apparent (Ewing et al., 1980). Although the regression analysis in our investigation used a surrogate variable that may reflect smoltification-related effects (release date), its adequacy in representing such effects has not been verified. Future investigations should include direct assessments of effects associated with developmen- tal processes, such as sodium and potassium ion lev- els and gill ATPase levels, as well as migrant size. Berggren and Filardo (1993) also examined the relationship between travel time and a host of pre- dictor variables for zero-age chinook salmon in John Day Reservoir. Their analysis included a subset of our data, as well as similar releases that were ex- ecuted in 1986-88 (Harmon et al.12). In their multi- variable approach, data were pooled across years. The variables in the final multiple regression model in- cluded release date, inverse flow, and an index of the absolute change in flow. The bivariate relationship between smolt travel time and inverse flow had an associated r2 value of 0.28. In contrast to our results, they concluded that increased flows reduced travel time of zero-age chinook salmon. 11 Beeman, J. W., D. Rondorf, J. Faler, M. Free, and P. Haner. 1990. Assessment of smolt condition for travel time analysis. U.S. Fish Wild. Serv., Cook, WA 98605. Report to Bonneville Power Administration, 71 p. 12 Harmon, J. R., G. M. Matthews, D. L. Park, and T. E. Ruehle. 1989. Evaluation of transportation of juvenile salmonids and related research on the Columbia and Snake Rivers, 1988. Northwest Fish. Sci. Cent., Natl. Mar. Fish. Serv., Seattle, WA 98112-2097. Report to U.S. Army Corp of Engineers, 1 1 p. NOTE Giorgi et al.: Migratory characteristics of juvenile Oncorhynchus tshawytscha 879 The variability in travel-time estimates observed in this study may in part have resulted from limited sampling capability at John Day Dam, since only 0.3 to 1.3% of any marked group was recovered at that site. There were four groups from which less than 10 recaptures were observed (Tables 1 and 2). However, the travel-time estimates of Berggren and Filardo ( 1993) displayed generally the same range of values, even though sampling effort at John Day Dam was increased during the latter years on which their analyses were based. Another confounding factor is that our investiga- tion treated the entire composite population of zero- age juveniles. For future studies, we suggest study- ing individual stocks of fish to describe any unique migratory characteristics that may be stock-specific. Acknowledgments We are grateful to Carl Sims and Dennis Umphries who participated in various phases of the study. John Williams and Kurt Gores reviewed the manuscript and provided useful comments. This research was supported in part by the Department of Energy and the Bonneville Power Administration. Literature cited Berggren, T. J., and M. J. Filardo. 1993. An analysis of variables influencing the mi- gration of juvenile salmonids in the Columbia River Basin. N. Am. J. Fish. Manage. 13(l):48-63. Ebel, W. J., and H. L. Raymond. 1976. Effects of atmospheric gas saturation on salmon and steelhead trout of the Snake and Co- lumbia rivers. Mar. Fish. Rev. 38(7):1-14. Ewing, R. D., C. A. Fustish, S. L. Johnson, and H. J. Pribble. 1980. Seaward migration of juvenile chinook salmon without elevated gill (Na + K)-ATPase activities. Trans. Am. Fish. Soc. 109:349-356. Giorgi, A. E. 1990. Biological manipulation of migratory behav- ior: the use of advanced photoperiod to accelerate smoltification in yearling chinook salmon. In D. L. Park (convener). Status and future of spring chinook salmon in the Columbia River Basin — con- servation and enhancement; proceedings of the spring chinook workshop, 8-9 November 1989, p. 108-114. NOAATech. Memo. NMFS F/NWC-187, Northwest Fish. Sci. Cent., Natl. Mar. Fish. Serv., Seattle, WA 98112-2097. Healey, M. C. 1991. Life history of chinook salmon (Oncorhynchus tshawytscha). In C. Groot, and L. Margolis (eds.), Pacific salmon life histories. UBC Press, Van- couver, Canada, 564 p. Mains, E., and J. Smith. 1964. The distribution, size, time and current pref- erences of seaward migrant chinook salmon in the Columbia and Snake Rivers. Wash. Dep. Fish., Fish. Res. Pap. 2(3):5-43. Mighell, J. L. 1969. Rapid cold-branding of salmon and trout with liquid nitrogen. J. Fish. Res. Board Can. 26:2765- 2769. Northwest Power Planning Council. 1987. Columbia Basin fish and wildlife pro- gram. Northwest Power Planning Council, Port- land, OR 97204, 246 p. Raymond, H. L. 1969. Effect of John Day Reservoir on the migra- tion rate of juvenile chinook salmon in the Colum- bia River. Trans. Am. Fish. Soc. 98(3):513-514. 1979. Effects of dams and impoundments on migra- tions of juvenile chinook salmon and steelhead from the Snake River, 1966 to 1975. Trans. Am. Fish. Soc. 108:529-565. 1988. Effects of hydroelectric development and fish- eries enhancement on spring and summer chinook salmon and steelhead in the Columbia River Basin. N. Am. J. Fish. Manage. 8(1): 1-24. Reimers, P. E. 1973. The length of residence of fall chinook salmon in Sixes River. Oreg. Fish Comm. Res. Rep. 4:2-43. Rich, W. 1922. Early history and seaward migration of chinook salmon in the Columbia and Sacramento Rivers. U.S. Bur. Fish., Bull. 37:1-74. Rieman, B. E., R. C. Beamesderfer, S. Vigg, and T. P. Poe. 1991. Estimated loss of juvenile salmonids to pre- dation by northern squawfish, walleyes, and small- mouth bass in John Day Reservoir, Columbia River. Trans. Am. Fish. Soc. 120:448-458. Smith, L. S. 1982. Decreased swimming performance a neces- sary component of the smolt migration in salmon in the Columbia River. Aquaculture 28:153-161. Vigg, S., T. P. Poe, L. A. Prendergast, and H. C. Hansel. 1991. Rates of consumption of juvenile salmonids and alternative prey fish by northern squawfish, walleyes, smallmouth bass, and channel catfish in John Day Reservoir, Columbia River. Trans. Am. Fish. Soc. 120:421-438. Williams, J. G. 1989. Snake River spring and summer chinook salmon: Can they be saved? Regul. Rivers Res. & Manage. 4:17-26. Zaugg, W. S. 1982. Relationships between smolt indices and mi- gration in controlled and natural environments. In E. L. Brannon and E. O. Salo (eds.), Proceed- ings of the salmon and trout migratory behavior symposium, Seattle, 1981. School of Fish., Univ. Wash., Seattle, WA 98195, p. 173-183. Comparisons between generalized growth curves for two estuarine populations of the eel tailed catfish Cnidoglanis macrocephalus Laurie J. B. Laurenson School of Biological and Environmental Sciences. Murdoch University Murdoch, Western Australia, 6 1 50, Australia Ian C. Potter* School of Biological and Environmental Sciences, Murdoch University, Murdoch. Western Australia. 6 1 50. Australia Norm G. Hall Western Australian Marine Research Laboratories Perth. Western Australia 6020, Australia The eel tailed catfishes (Plotosidae) are distributed throughout the Indo-west Pacific region and com- prise approximately 30 species. Just over half of these species are found in Australian waters (Hoese and Hanley, 1989). The estuarine catfish or cobbler, Cnidoglanis macrocephalus Glinther, is one of three plotosid species that are found in the marine and estuarine waters of the southwestern region of Australia (Kowarsky, 1976; Hutchins and Swainston, 1986). Cnidoglanis macrocephalus can complete its life cycle in estuaries as well as in coastal marine waters (Laurenson et al., 1993a), suggest- ing that the populations of this spe- cies in each of the different estuar- ies represent separate demes, a view supported by the results of electrophoretic studies (Ayvazian et al., 1994). Cnidoglanis macrocephalus is the most valuable of several teleo- sts fished commercially in Western Australian estuaries (Lenanton and Potter, 1987). While the perma- nently open Swan and Peel-Harvey estuaries on the southwestern coast of Western Australia were previ- ously the main contributors to the fishery for this species (Laurenson et al., 1992), this role has now been assumed by Wilson Inlet on the southern coast of the state (Lauren- son, 1992; Laurenson et al., 1993b). In contrast to the Swan and Peel- Harvey estuaries, Wilson Inlet is seasonally closed and, because of its more southerly location, does not reach as high a temperature in the summer (c.f. Loneragan et al., 1989; Potter etal., 1993). Fish are commonly aged by counting the number of annuli on hard structures, such as scales, otoliths, vertebrae, or spines (e.g. Beamish and McFarlane, 1983; Casselman, 1987). However, prior to carrying out such counts, it is important to validate that each of the sequential growth zones is formed annually (e.g. Beamish and McFarlane, 1983; Beckman et al., 1989; Collins et al., 1989; Hyndes et al., 1992). Although Nel et al. ( 1985) showed that the translucent zones in the asterisci of C. macro- cephalus from the Swan Estuary tended to be formed annually, their results were based on pooled data for all fish and, thus, did not verify that this applied equally to each of the sequential translucent zones. Moreover, since the data for males and females were pooled, it was not possible to determine whether the growth rates of the two sexes in this system were the same. A variety of different forms of growth equations can be calculated from 1) the lengths at given ages and 2) back calculations of body length at each annulus, using the relationship between body length and otolith radius. Both calcula- tions use a predetermined "birth date" for the species. The effective- ness of using length-at-age data relies on obtaining representative samples of all age classes. Back calculations are particularly useful when certain age classes have not been sampled effectively but may produce biased estimates of the lengths of younger fish, i.e. Lee's phenomenon (Ricker, 1975). Fur- thermore, the lack of independence of the multiple measures for lengths at annulus formation ob- tained for a single fish by this method may introduce a statistical bias. The aims of our study were 1) to validate that each of the sequential translucent growth zones on otoliths of C. macrocephalus in Wil- son Inlet and the Swan Estuary correspond to an annulus and 2) to construct growth curves for each sex in both populations, using both lengths offish at age of capture and back-calculated lengths. These curves were then used to compare a) growth between sexes within each estuary, b) growth between estuaries, and c) growth calculated using lengths at age and back-cal- culated lengths. * Send reprint requests and correspondence to the second author. Manuscript accepted 11 April 1994. Fishery Bulletin 92:880-889. 880 NOTE Laurenson et al.: Growth curves of two estuarine populations of Cnidoglanis macrocephalus 881 Materials and methods Collection of fish Juvenile and adult Cnidoglanis macrocephalus were collected by seining at eight sites, gillnetting at nine sites, and otter trawling at six sites located through- out the basin of Wilson Inlet between September 1987 and April 1989 (see Fig. 1 in Potter et al. [1993] for location of this estuary and the sampling sites). Some of the sampling by each method was carried out monthly, while the rest was undertaken bimonthly (see Potter et al., 1993). The seine was 41.5 m long (stretched mesh = 51 mm in wings and 9.5 mm in pocket), while the gill net consisted of six 30-m con- tiguous panels, each with a different stretched mesh size, i.e. 38, 51, 63, 76, 89, or 102 mm. The stretched mesh in the wings and codend of the otter trawl were 51 and 25 mm, respectively. Seine netting and otter trawling were carried out during the day, while gill- netting was undertaken overnight. A small number of larval and post-larval C. macrocephalus were also collected in night-time plankton tows (Neira and Potter, 1992) and from their nests by dip net (Lauren- son et al., 1993a). Sampling in the Swan Estuary employed winged funnel traps between August 1982 and April 1983 (see Nel et al., 1985). Fish were also taken in a seine and otter trawl similar to those used in Wilson Inlet and with gill nets containing panels with the same mesh sizes as those employed in Wilson Inlet, but with additional panels of 13- and 25-mm mesh. Validation of translucent zones as annuli and otolith measurements The first 10 males and 10 females of C. macro- cephalus caught in each panel of the gill nets at each site in Wilson Inlet on each sampling occasion, to- gether with all fish caught in seine nets, were kept for ageing. All fish caught in otter trawls, except for a small number that were retained for tagging ex- periments, were also used for ageing. The total length and wet weight of each fish were recorded to the near- est 1 mm and 0.1 g, respectively. Each C. macro- cephalus was sexed, except in the case of smaller fish (- l-exp(-a(712-T1)) Vfc Case 2: a* 0,6 = 0 L, = y1 exp log ' y.,) l-exp(-a(t-T,)) Vi ) 1- exp (-a (To-!;)) Case 3: a = 0,6*0 L, y, +(y2-y1)T * i2 1 j \h Case 4 : a = 0,6 = 0 Lt = y, exp I..- y3 t-T, T9-T, When a >0 and 6 = 1, the generalized growth curve is equivalent to the traditional form of the von Bertalanffy growth curve, with a = k. The resultant form of the generalized growth equation was deter- mined by the parameters a and 6 that resulted in the minimum sum of squared deviations. Data were fitted by using a nonlinear least squares method, employing the nonlinear (NLIN) procedure of SAS (Ihnen and Goodnight, 1987). All back calculations and curve fittings were carried out separately for each sex in both populations. Juveniles, for which the sex could not be determined, were included in calculating growth curves of both sexes from length- at-age data. Calculations of all curves assumed a birth date of 1 December in Wilson Inlet and 1 No- vember in Swan Estuary (Laurenson et al., 1993a). Each growth curve, fitted by using the traditional form of the von Bertalanffy growth equation, was compared with the corresponding generalized growth curve by using a likelihood ratio test, an approach adopted with several other fish species (Kimura, 1980; Kirkwood, 1983; Cerrato, 1990; Hampton, 1991; Buxton, 1993). The generalized growth curves of both sexes in Wilson Inlet and Swan Estuary based on lengths at age and back calculated lengths, were compared by using the same likelihood ratio test, which involved determining the improvement of fit obtained by using the two separate curves, rather than a common curve. This involved 1) comparing the curve for males with that for females in each sys- tem, using first lengths at age and then back-calcu- lated lengths; 2) comparing the curves for each sex in Wilson Inlet with that for the corresponding sex in Swan Estuary, using first lengths at age and then back-calculated lengths; and 3) comparing the curves calculated from lengths at age with those obtained from back-calculated lengths, first for males in each system and then for females in each system. Results Mean monthly percentages of otoliths from Wilson Inlet with a peripheral translucent zone and one, two, or three inner translucent zones followed similar seasonal trends (Fig. 1). The percentage of such otoliths rose sharply in early spring and fell to close to zero in the late spring or early summer where they NOTE Laurenson et al.: Growth curves of two estuarine populations of Cnidoglanis macrocephalus 883 1 Translucent Zone 19 Aug Oct Dec Feb Apr Jun Aug Oct Dec Feb Apr 1987 1988 1989 Figure 1 The percentage of lapilli of the eel tailed catfish Cnidoglanis macrocephalus from Wilson Inlet pos- sessing a clearly defined peripheral translucent zone in each month. Data are presented separately for otoliths in which there are one to five or more inner translucent zones. Black rectangles on the x-axis represent summer and winter months, white rect- angles the autumn and spring months. MHI 1 Translucent Zone 2 Translucent Zones 5 «e 120 3 Translucent Zones 1 7 3 1 **\i >4 Translucent Zones 2 1 Aug Oct Dec Feb Apr Jun Aug Oct Dec Feb Apr 1982 1983 1984 Figure 2 Mean relative marginal increments for asterisci otoliths of the eel tailed catfish, Cnidoglanis macrocephalus from the Swan Estuary. Recalculated from the data used by Nel et al. (1985). Data are presented separately for otoliths in which there are one to four or more translucent zones. Stan- dard errors are given when sample size was >3. Black rectangles on the x-axis represent summer and winter months, white rect- angles the autumn and spring months. remained through the following summer, autumn, and winter months. Although data for otoliths with both four and five or more translucent zones were less abundant, they followed a similar trend (Fig. 1). The mean marginal increment on otoliths with one, two, and three translucent zones from Swan Estuary fell to a mini- mum in the spring and rose progressively during the ensuing summer and early autumn, before levelling off in the late autumn and winter (Fig. 2). While the number of otoliths with four or more translucent zones was small, the trend shown by the marginal increment on these otoliths is similar. A cubic polynomial equation, using logarithm (natural) transformed data, provided the best description of the relationship between otolith radius and total fish length in both Wilson Inlet and Swan Estuary, when lapilli and asterisci otoliths were used, respectively (Fig. 3). The equations were as follows: Wilson Inlet Males: Females: y = 5.700 + 1.388* y = 5.708 + 1.374* 0.191a:2 - 0.315.x-3 {R2 = 0.935, P < 0.001, n = 462) 0.235.V2 - 0.317*3 (R2 = 0.926, P < 0.001, n = 876) 884 Fishery Bulletin 92(4), 1994 Swan Estuary Males: Females: y = 6.152 + 0.95 lx - 0.875*2 - 0.400*3 (R2 = 0.931, P < 0.001, n = 499) y = 6.174 + 1.171* - 0.479*2 - 0.230*3 (R2 = 0.931, P < 0.001, n = 568) Examination of the otoliths suggests that the curvi- linearity at the upper end of these relationships (Fig. 3) is due to the otoliths of larger fish tending to thicken rather than lengthen. The lengths offish of a given age class were highly variable (Figs 4 and 5). For example, the lengths of female fish that were about four years old in Wilson Inlet ranged from 478 to 631 mm and those that were about three years old in the Swan Estuary ranged from 351 to 591 mm. The predicted lengths offish, derived from generalized growth curves, were greater when lengths at age rather than back-calculated lengths were used for fish of ages 1 and 2 (Table 1). The high values for R2 for the generalized growth curves, derived from both lengths at age and back- calculated lengths, show that these curves fit the data well (Table 2). The oldest male and female C. macrocephalus caught in Wilson Inlet were 123/4 years old (718 mm, 1885 g) and 93A years old (670 mm, 1738 g), respectively. The corresponding values for fish from the Swan Estuary were 5 years (582 mm, 1142 g) and 63A years (683 mm, 1880 g), respectively. The use of common curves in the cases of both lengths at age and back-calculated lengths for each of the two sexes in each system accounted for 89 to 94% of the observed variance. By assuming that a difference exists between the growth curves of the two sexes in each system and with each of the two methods, the fit was improved by only 0.003% for back calculated data for the Swan Estuary and 0.3% for length at age data for Wilson Inlet. Applying likelihood ratio tests, the length-at-age growth curves for males and females differed signifi- cantly in both the Wilson Inlet (P<0.001) and Swan Estuary populations (P<0.05). Back-calculated growth curves calculated for the two sexes also dif- fered significantly (P<0.001) in Wilson Inlet but not in Swan Estuary. The use of a common curve for each sex by using both lengths at age and back-calculated lengths for Wilson Inlet Males Swan Estuary Males 0.5 (III (IS In Otolith Radius (mm) 2.0 1.5 III (I 5 (1 0 0 5 In Otolith Radius (mmi Figure 3 Relationships between the natural logarithms of total length (x) and lapilli ra- dius (y) of the eel tailed catfish Cnidoglanis macrocephalus from Wilson Inlet and the Swan Estuary. Broken lines represent the best fit for a linear regres- sion, the solid lines the best fit for the cubic polynomial equation. NOTE Laurenson et al.: Growth curves of two estuarine populations of Cnidoglanis macrocephalus 885 Lengths at Age Males 4 6 Age (years) Back Calculated Lengths Males Females Age (years) Figure 4 Growth curves obtained from lengths at age and back-calcu- lated lengths of the eel tailed catfish Cnidoglanis macrocephalus from Wilson Inlet with the method of Schnute (1981). Mean ±1 standard error of the mean of back-calculated lengths at each age are given. Figure 5 Growth curves obtained from lengths at age and back-calcu- lated length data for the eel tailed catfish Cnidoglanis macrocepha- lus from the Swan Estuary with the method of Schnute (1981). Mean ±1 standard error of the mean of back-calculated lengths at each age are given. Lengths at Age Back Calculated Lengths 800 r Males 800 Males E | e s 3 e H 600 400 0 ( 600 400 200 *^-*-±^ T 2 4 6 8 10 ) 2 4 6 8 10 !00 - Females 800 Females I § S 2 o H MM) 400 200 ( 600 400 200 0 ( X^""^ ) 2 4 6 8 10 2 4 6 8 10 Age (years) Age (years) 886 Fishery Bulletin 92(4), 1994 Table 1 The total lengths mm) at sequential ages of the eel tailed catfish Cnidogl anis macrocephalos in Wilsor Inlet and Swan Estuary, predicted from generalized growth curves (Schnute, 1981 ) calculated from lengths at age (LAA) and back-calculated le ngths (BCD. Wilson Inlet Swan Estuary Male Female Male Female Male Female Male Female Age LAA LAA BCL BCL LAA LAA BCL mi, 1 203 180 156 158 239 225 185 184 2 335 324 293 298 356 353 324 323 3 436 449 436 448 447 456 436 440 4 513 541 543 556 525 538 527 531 5 573 603 605 613 594 603 601 598 6 619 643 636 638 654 647 7 655 668 650 648 694 8 682 684 657 652 9 703 693 660 653 10 720 699 661 Table 2 The parameters of the generalized growth curves fitted to lengths at age an d back-ca culated 1 ?ngths for the eel tailed catfish Cnidoglanis macrocephalus in Wilson Inlet and S wan Estuary, y and y., are lengths ( mm ) at reference ages 1 and 4 and a and b are the parameters of th e growth equation. Location ■vi y2 a b R2 /; Wilson Inlet Lengths at age Female 180 541 0.51 0.20 0.90 916 Male 203 513 0.26 1.04 0.92 502 Back-calculated lengths Female 158 556 0.96 -0.99 0.94 2354 Male 156 543 0.82 -0.71 0.96 1102 Swan Estuary Length at age Female 225 537 0.25 0.90 0.91 517 Male 239 525 0.02 1.75 0.85 447 Back-calculated lengths Female 183 530 0.37 0.53 0.90 615 Male 184 527 0.20 1.04 0.85 426 the populations in the two systems accounted for 90 to 94% of the observed variance. The additional vari- ance explained by assuming a difference between the growth curves for each sex in each system improved the fit to the four data sets by 0.3 to 0.6%. The growth curves estimated for males from lengths at age and from back-calculated lengths in Wilson Inlet differed significantly from those estimated for males in Swan Estuary using the corresponding types of data; the same was true for females (P<0.001 ). The percentage of the variance explained by the common curves derived from lengths at age and back- calculated lengths for each sex in each system ranged from 81% for males in the Swan Estuary to 94% for both males and females in Wilson Inlet. The percent- age of the variance explained by assuming that the growth curves determined from lengths at age and back-calculated lengths are different was improved by 0.8 and 0.2% respectively for males and females from Wilson Inlet and by 4.2 and 1.6% respectively for males and females from the Swan Estuary. The length at age and back-calculated growth curves for males in Wilson Inlet and Swan Estuary differed sig- nificantly; the same applied for females (P<0.001). Discussion The present study of the lapilli of C. macrocephalus in Wilson Inlet is the first to demonstrate in a plotosid that each of the otolith's first four translucent zones, and probably all other translucent zones, are formed annually. Furthermore, re-analysis of the data of Nel et al. ( 1985) has shown that this also applies to the NOTE Laurenson et al : Growth curves of two estuanne populations of Cnidoglanis macrocephaius 887 asterisci in C. macrocephaius from the Swan Estu- ary. The importance of confirming that each succes- sive translucent zone is formed annually is demon- strated by the results obtained by Hyndes et al. (1992) for whole sagittae of Platycephat 'us specula- tor in Wilson Inlet. In that species, mean monthly marginal increments showed a very clear seasonal trend when individual marginal increments on all unsectioned otoliths were pooled, irrespective of the number of translucent zones. However, they did not show conspicuous trends when the data for the mar- ginal increments on unsectioned otoliths with two, three, four, and five or more translucent zones were each plotted separately. In other words, when mar- ginal increment data for all otoliths were pooled, the pronounced seasonality exhibited by the mean mar- ginal increments on otoliths with one translucent zone of P. speculator had an overwhelming influence on the data set. The von Bertalanffy growth curve did not suffi- ciently describe the growth of C. macrocephaius from Wilson Inlet; the lengths were consistently greater than the mean length at ages 7 and above and showed increasing divergence with age. This was far less of a problem in Swan Estuary where older fish were less abundant. The generalized growth curve pro- vided better fits to the data than the von Bertanlanffy curve for males and females in both systems, when both lengths at age and back-calculated lengths were used. Furthermore, likelihood ratio tests showed that this improvement was significant in three of the four cases for the population in Wilson Inlet. Such im- provement is consistent with the observation that when there is an acceleration of growth early in life, the von Bertalanffy growth curve does not provide as adequate a fit as the Schnute, Gompertz, or Richard's curves (Schnute, 1981; Campana and Jones, 1992). While the presence among younger fish of smaller back-calculated lengths than mean lengths at age (Table 1 ) would be consistent with Lee's phenomenon (Ricker, 1975), it could also have been brought about by the low numbers of younger fish in the samples. The fits of the common curves constructed for each sex in Wilson Inlet from lengths at age and back- calculated lengths were improved by only 0.2% for females and 0.8% for males when separate curves were used. However, this was not the case for fish in the Swan Estuary, where the sum of squares was improved by 1.6% for females and by as much as 4.2% for the males. The differences in improvement in fit in the two systems probably reflects the fact that, while the 0+ age class in the Swan Estuary was caught in greater numbers, it tended to be represented in samples by the larger members of this age class. The improvement of fit obtained by using separate growth curves was small, both in comparisons be- tween males and females in Wilson Inlet and Swan Estuary and in comparisons between corresponding sexes in the two systems. This applied to curves con- structed both from lengths at age and back-calcu- lated lengths. In none of these cases was the sum of squares improved by more than 0.6%. However, al- though the differences between the curves for each sex in each system and for the corresponding sexes in the two populations were small, and even though the lengths varied considerably at a given age, the curves were still statistically different with a likeli- hood ratio test (usually P<0.001). These differences probably reflect the influence of the large number of data points used to construct the growth curves. The small magnitude of the differences between these growth curves is demonstrated by the fact that at age 4, the lengths of males and females in Wilson Inlet and the Swan Estuary, predicted from the gen- eralized growth curve, generally differed by less than 3%, irrespective of whether the curve was constructed from lengths at age or back-calculated lengths. Thus, although there were usually highly statistically sig- nificant differences between curves, the actual dif- ferences between the curves for the two sexes in each population and between the corresponding sexes in those populations are almost certainly of limited bio- logical significance. In conclusion, the growth of C. macrocephaius in Wilson Inlet was similar to that in the Swan Estu- ary. This similarity occurred despite the fact that water temperatures in the latter system were over 5°C higher in the summer (c.f. Loneragan and Pot- ter, 1990; Potter et al., 1993). Wilson Inlet is eutrophic and therefore more productive (Lukatelich et al., 1987) and consequently contains a greater abundance of the large deposit-feeding benthic invertebrates1 that make a major contribution to the diet of C. macrocephaius (Nel et al., 1985; Laurenson, 1992). Therefore the similarity between the growth rate of C. macrocephaius in Wilson Inlet and the Swan Es- tuary may reflect a compensation for lower water temperatures by greater prey abundance. Acknowledgments Our thanks are expressed to F. Baronie, D. Gaughan, P. Geijsel, P. Humphries, G. Hyndes, and F. Neira for assistance with sampling. Gratitude is also ex- 1 Platoll, M. School of Biological and Environmental Sciences, Murdoch Univ., Murdoch, Western Australia 6150, Australia. Personal commun., 1991. 888 Fishery Bulletin 92(4), 1994 pressed to the several commercial fishermen, and particularly B. Miller, O. Mcintosh, and C. Smith, who provided fish and access to their log books. R. I. C. C. Francis provided a valuable input into the ap- proaches adopted in this study. We appreciate the constructive comments made on the manuscript by G. Hyndes, M. Platell, W. Fletcher, B. Wise, by the two referees, and the Scientific Editor. Literature cited Ayvazian, D. G., M. S. Johnson, and D. J. McGlashan. 1994. High levels of genetic subdivision of marine and estuarine populations of the estuarine catfish Cnidoglanis macrocephalus (Plotosidae) in south- western Australia. Mar. Biol. 118:25-31. Beamish, R. J., and G. A. McFarlane. 1983. The forgotten requirement for age validation in fisheries biology. Trans. Am. Fish. Soc. 112:735-743. Beckman, D. W., C. A. Wilson, and A. L. Stanley. 1989. Age and growth of red drum, Sciaenops ocellatus, from offshore waters of the northern Gulf of Mexico. Fish. Bull. 87:17-27. Beckman, D. W., A. L. Stanley, J. H. Render, and C. A. Wilson. 1990. Age and growth of black drum in Louisiana waters of the Gulf of Mexico. Trans. Am. Fish. Soc. 119:537-544. Buxton, C. D. 1993. Life-history changes in exploited reef fishes on the east coast of South Africa. Environ. Biol. Fishes 36:47-63. Campana, S. E., and C. M. Jones. 1992. Analysis of otolith microstructure data. In D. K. Stevenson and S. E. Campana (eds.), Otolith microstructure examination and analysis. Can. J. Fish. Aquat. Sci. 117:73-100. Casselman, J. M. 1987. Determination of age and growth. In A. H. Weatherley and H. S. Gill (eds. ), The biology offish growth, p. 209-242. Academic Press, London. Cerrato, R. M. 1990. Interpretable statistical tests for growth com- parisons using parameters in the von Bertalanffy equation. Can. J. Fish. Aquat. Sci. 47:1416-1426. Collins, M. E., D. J. Schmidt, W. C. Waltz, and J. L. Pickney. 1989. Age and growth of king mackerel. Scorn beromorus cavalla, from the Atlantic coast of the United States. Fish. Bull. 87:49-61. Crozier, W. W 1989. Age and growth of angler-fish Lophius piscatorius L. in the north Irish Sea. Fish. Res. i Amst. 17:267-278. Francis, R. I. C. C. 1990. Back calculations of fish lengths: a critical review. J. Fish. Biol. 36:883-902. Hampton, J. 1991. Estimation of southern bluefin tuna Thunnus maccoyii growth parameters from tagging data, using von Bertalanffy models incorporating indi- vidual variation. Fish. Bull. 89:577-590. Hoese, D. F., and J. E. Hanley. 1989. Plotosidae. In J. R. Paxton, D. F. Hoese, G. R. Allen, and J. E. Hanley (eds.), Zoological cata- logue of Australia. Vol 7: Pisces: Petromyzontidae to Carangidae, p. 222-226. Australian Govern- ment Publishing Service, Canberra. Hutchins, B., and R. Swainston. 1986. Sea fishes of southern Australia. Swainston Publishing, Perth. Hyndes, G. A., N. R. Loneragan, and I. C. Potter. 1992. Influence of sectioning otoliths on marginal increment trends and age and growth estimates for the flathead, Platycephalus speculator. Fish. Bull. 90:276-284. Ihnen, L.A., and J.H. Goodnight. 1987. The NLIN procedure. In R. C. Luginbuhl, S. D. Scholtzhauer, J. C. Parker and K. P. Ingraham (eds.), SAS/STAT guides for personal computers, Version 6 ed. SAS Institute, Cary, North Carolina. Kimura, D. K. 1980. Likelihood methods for the von Bertalanffy growth curve. Fish. Bull. 77:765-776. Kirkwood, G. P. 1983. Estimation of von Bertalanffy growth curve parameters using length increment and age-length data. Can. J. Fish. Aquat. Sci. 40:1405-1411. Kowarsky, J. 1976. Clarification of the name and distribution of the plotosid catfish Cnidoglanis macrocephalus. Copeia 1976:593-594. Laurenson, L. J. B. 1992. Biology and commercial exploitation of the estuarine catfish, Cnidoglanis macrocephalus (Valenciennes), in south western Australia with emphasis on the seasonally closed Wilson Inlet. Ph.D. diss., Murdoch Univ., Perth, Western Australia. Laurenson, L. J. B., F. J. Neira, and I. C. Potter. 1993a. Reproductive biology and larval morphology of the marine plotosid Cnidoglanis macrocephalus (Teleostei) in a seasonally closed Australian estuary. Hydrobiologia 268:179-192. Laurenson, L. J. B., I. C. Potter, R. C. J. Lenanton, and N. G. Hall. 1993b. The significance of size at sexual maturity, mesh size and closed fishing waters to the commer- cial fishery for the catfish Cnidoglanis macro- cephalus in Australian estuaries. J. Appl. Ichthyol. 9:210-221. Lenanton, R. C. J., and I. C. Potter. 1987. Contribution of estuaries to commercial fisher- ies in temperate Western Australia and the concept of estuarine dependence. Estuaries 10:28-35. Loneragan, N. R., and I. C. Potter. 1990. Factors influencing community structure and NOTE Laurenson et al.: Growth curves of two estuarine populations of Cnidoglanis macrocephalus 889 distribution of different life-cycle categories of fishes in shallow waters of a large Australian estuary. Mar. Biol. (Berl.) 106:25-37. Loneragan, N. R., I. C. Potter, and R. C. J. Lenanton. 1989. Influence of site, season and year on contri- butions made by marine, estuarine, diadromous and freshwater species to the fish fauna of a tem- perate Australian estuary. Mar. Biol. (Berl.) 103:461-479. Lukatelich, R. J., N. J. Schofield, and A. J. McComb. 1987. Nutrient loading and macrophyte growth in Wilson Inlet, a bar-built southwestern Australian estuary. Estuarine Coastal Shelf Sci. 24:141-165. Maceina, M. J., D. N. Hata, T. L. Linton, and A. M. Landrey Jr. 1987. Age and growth analysis of spotted sea trout from Galveston Bay, Texas. Trans. Am. Fish. Soc. 116:54-59. Neira, F. J., and I. C. Potter. 1992. The ichthyoplankton of a seasonally closed estuary in temperate Australia. Does an extended period of opening influence species composi- tion? J. Fish Biol. 41:935-953. Nel, S. A., I. C. Potter, and N. R. Loneragan. 1985. The biology of the catfish Cnidoglanis macrocephalus (Plotosidae) in an Australian estuary. Estuarine Coastal Shelf Sci. 21:895-909. Potter, I. C, G. A. Hyndes, and F. M. Baronie. 1993. The fish fauna of a seasonally closed Austra- lian estuary. Is the prevalence of estuarine-spawn- ing species high? Mar. Biol. (Berl.) 116:19-30. Ricker, W. E. 1975. Computation and interpretation of biological statistics of fish populations. Fish. Res. Board Can., Bull. 191, 382 p. Schnute, J. 1981. A versatile growth model with statistically stable parameters. Can. J. Fish. Aquat. Sci. 38:1128-1140. A comparison of larval and postlarval gulf menhaden, Brevoortia patronus, growth rates between an offshore spawning ground and an estuarine nursery Richard C. Raynie* Richard F. Shaw Coastal Fisheries Institute Louisiana State University Baton Rouge, Louisiana 70803 The fishery for gulf menhaden, Brevoortia patronus, was the larg- est by weight in the United States from 1963 through 1988 and has had a significant impact on the economy of the northern Gulf of Mexico coast. The species is also an ecologically important prey item for a number of commercially and recreationally important species (Lassuy, 1983). Gulf menhaden early life history has been reviewed by a number of authors (Lassuy, 1983; Deegan, 1985; Powell and Phonlor, 1986; Shaw et al., 1988; Christmas and Waller1). Adults spawn in offshore and coastal waters in depths rang- ing from 11 to 128 m; peak spawn- ing occurs between the 10- and 60- m isobaths (Shaw et al., 1988). Most spawning generally occurs between October and March; peak spawning occurs in December (Fore, 1970; Shaw et al., 1985; Christmas and Waller1). Once spawned, gulf menhaden eggs are pelagic and hatch within about two days. The offshore larval drift pe- riod may last from 4 to 10 weeks (Deegan and Thompson, 1987; Shaw et al., 1988). Peak immigra- tion through tidal passes into es- tuarine nurseries generally occurs 890 between December and March (Suttkus, 1956; Lassuy, 1983). In the estuary, larval gulf men- haden move into bayous and other low salinity areas at the onset of transformation into juveniles (Fore and Baxter, 1972; Simoneaux, 1979; Deegan, 1990; Raynie and Shaw, in press). Estuarine resi- dence is typical during summer months. As juveniles grow larger, they tend to move downstream to higher salinity waters and from late summer to winter many emi- grate to open coastal waters (Dee- gan, 1990). Daily otolith increment forma- tion has been validated and is esti- mated to begin at the onset of ex- ogenous feeding, which occurs about three days after hatching (Warlen, 1988). Growth rate esti- mates based on larval gulf menha- den otolith analyses have been made from larvae collected off the Mississippi River Delta, Florida and Texas (Warlen, 1988) and from young of the year collected within estuarine waters of Louisiana (Deegan and Thompson, 1987). A comparison between growth rates of gulf menhaden captured from continental shelf and adjacent es- tuarine waters during the same time period has not been done. The purpose of this paper is to examine growth rates of larval and postlar- val gulf menhaden from offshore and estuarine habitats and relate the results to metamorphosis. Materials and methods Sampling procedure Gulf menhaden larvae and post- larvae were collected at two sta- tions in the northern Gulf of Mexico (6 and 32 km from shore) on 23 January 1990 and from three loca- tions (Lower Bay, Mosquito Island, and Big Carencro Bayou) within the adjacent estuary, Fourleague Bay, Louisiana, on 24-25 January 1990 (see Raynie and Shaw, in press). Larvae were collected off- shore with a 60-cm bongo frame equipped with a 505-um mesh net fitted with a flow meter (General Oceanics Model no. 2030). Within Fourleague Bay, collections were made with a bow-mounted plank- ton push net of the same diameter and mesh size as that used offshore. One three-minute collection was taken at each station offshore and at each station each day within Fourleague Bay. Plankton push nets have been shown to be effec- tive at collecting larval fish (Miller, 1973; Raynie and Shaw, in press) and juvenile fish (Herke, 1969; Kriete and Loesch, 1980). The use of this gear in this highly turbid and shallow estuarine system (mean depth=1.5 m; Teague et al., 1988) minimizes net avoidance. Address correspondence to Louisiana De- partment of Natural Resources, Coastal Restoration Division, RO. Box 94396, Ba- ton Rouge, Louisiana 70804-9396 1 Christmas, J. Y., and R. S. Waller. 1975. Location and time of menhaden spawn- ing in the Gulf of Mexico. CCRL/NMFS Contract Rep. 03-4-042-24. Manuscript accepted 16 March 1994. Fishery Bulletin 92:890-894. NOTE Raynie and Shaw: Comparison of larval and postlarval growth rates of Brevoortia patronus 891 Samples were initially preserved with 95% etha- nol, stored in ice, and later preserved with a 70% ethanol solution in the lab. Temperature and salin- ity were measured with a Beckman Portable Elec- tronic Salinometer (Model No. RS5-3). Laboratory analysis Notochord lengths (preflexion, NL) or standard lengths (SL) were measured to the nearest 0.1 mm with an ocular micrometer under a dissecting micro- scope. Sagittal otoliths were then removed from a random subsample (^=111) of larvae under a dissect- ing microscope with polarized light. Otoliths were air-dried and mounted on a glass microscope slide with S/P Accu-mount 60. Otoliths were sufficiently thin and rings sufficiently spaced to allow for optical sectioning (focusing to the plane of maximum clar- ity) under a compound microscope (400x or l,000x) to make total increment counts and otolith diameter measurements. Increments were independently counted by each author and averaged. Spawning dates were back-calculated for each larva by subtracting the estimated age from date of capture (i.e. capture date -[ring count + 5 days for egg incubation and yolk-sac absorption]) (Warlen, 1988). It was assumed that there were no differences in the age at first increment deposition (5 days) among larvae. Statistical analysis Age and growth data from each environment (offshore and es- tuarine) and the combined data were fit to the Laird version of the Gompertz growth equation (Laird et al., 1965) by means of nonlinear least squares regres- sion techniques (SAS Institute, Inc., 1985): Lt=L0e \ I, where Lt = standard length of larvae at day f ; L0 = initial length; K = A0/a;A0 = age-spe- cific growth rate at L0; and a = the exponential decline in the age-specific growth rate. Be- cause five days were added to otolith counts to attain age es- timates (2 days incubation + 3 ic, 10 K 5 days between hatching and exogenous feeding and first increment formation), two days were subtracted from the age estimates, so that the Y-intercept would approximate the hatching length. The length at hatch- ing has been observed from laboratory data (2.6-3.0 mm NL; Hettler, 1984) and estimated from field data (2.4 mm NL; Warlen, 1988). With these data, we fixed the hatching length at 3.0 mm NL in our models. Average daily growth rate was estimated by Average daily growth = ( standard length - 3.0 mm ) days posthatcb (after Deegan and Thompson, 1987). Results and discussion The mean surface water temperature offshore at the time of capture was 17.8°C (range 17.5-18.0°C) and the mean salinity was 31.0 ppt (range 29.0-33.0 ppt). Within Fourleague Bay, the mean temperature was 19.0°C (range 17.5-20. 1°C) and salinities ranged from 2.7 to 7.3 ppt, with the exception of our 24 January 1990 Lower Bay collection when the salinity was 23.9 ppt. According to age estimates, most of the larvae col- lected offshore were spawned within one week be- tween 27 December 1989 and 3 January 1990. Vir- tually all larvae collected within Fourleague Bay were spawned between mid-November and 24 De- cember 1989, with a peak between 17 and 24 De- cember 1989 (Fig. 1). One 85-day-old larva was col- □ Offshore, n =36 ■ Fourleague Bay, n=75 IjiLlIijjiiII ,n , ri n Estimated Spawning Date Figure 1 Percent frequency of birth (spawning) dates estimated from otolith increment data for gulf menhaden collected offshore and within Fourleague Bay, Louisi- ana, between 23 and 25 January 1990. rc=number of larvae collected from each environment. 892 Fishery Bulletin 92(4), 1994 lected from Mosquito Island and was estimated to have been spawned on 1 November 1989. Growth of larval fish (and other vertebrates) typi- cally proceeds through a series of consecutive inter- vals (thresholds) which characterize ontogeny. Peri- ods of rapid growth are generally followed by peri- ods of slower development during which complex structures prepare for the next series of changes (Balon, 1984). The Laird-Gompertz equation has been used to describe larval fish growth when the length- age plots are nonlinear and upper asymptotes are apparent (Zweifel and Lasker, 1976; Methot and Kramer, 1979; Laroche et al., 1982; Warlen and Chester, 1985; Warlen, 1988). This model was used to estimate the age-specific growth rate and the ex- ponential decline in the age-specific growth rate as larval gulf menhaden approach metamorphosis to juveniles. Growth rates were estimated from larvae between 5.8 and 16 mm SL collected offshore and larvae be- tween 17 and 24 mm SL collected within Fourleague Bay. The average daily growth of larvae collected off- shore (0.44 mm/day) was greater than within Fourleague Bay (0.12 mm/day). The average daily growth rate of larval gulf menhaden from the com- bined data was 0.25 mm/day. Postlarval gulf menhaden are estimated to be 15- 25 mm SL (Shaw et al., 1988) upon entering the es- tuary and begin transformation to the juvenile stage around 20 mm SL. Transformation is complete at about 30 mm (Suttkus, 1956; Hettler, 1984). Between 20 and 30 mm SL, however, growth characteristics change (Fig. 2): mouth parts and gill rakers are modi- fied and the body begins to thicken and take on the deep-bodied characteristics of juveniles and adults (Suttkus, 1956). During this threshold, postlarval (prejuvenile) gulf menhaden growth in weight is dis- proportionately greater than growth in length (Deegan and Thompson, 1987). The period of slowed growth in length just before and during juvenile transformation is followed by a dramatic increase in growth rate (Springer and Woodburn, 1960; Deegan and Thompson, 1987). Av- erage daily growth of gulf menhaden between 18 and 82 mm SL reportedly ranges from 0.20 to 0.48 mm/ day within Fourleague Bay (Deegan and Thompson, 1987). Our estimate of average daily growth rate for postlarvae (17-24 mm SL) within Fourleague Bay was expectedly lower (0.12 mm/day), since our lar- vae were approaching or were in the process of trans- formation. Our estimate of average daily growth from offshore is similar to Warlen's (1988) growth estimates for larval gulf menhaden collected off Southwest Pass, Louisiana (0.28-0.42 mm/day). Some marine larvae have been shown to grow faster at higher tempera- tures (Laurence et al., 1981); however, this has not been demonstrated for gulf menhaden (Warlen, 1988). During the winter, surface water temperatures are generally warmer offshore than within Fourleague Bay; however, this difference is gener- ally minimal (Raynie and Shaw, in press). Our tem- perature data are insufficient (and may be atypical of the average conditions) to evaluate the relation- ship between growth and temperature. The differ- ence in growth rates between environments, how- ever, is most likely the result of ontogeny. A 71% de- crease in growth rate between larval and juvenile 30 40 50 60 Age (days posthatch) Figure 2 Laird-Gompertz growth models for larval gulf menhaden collected from offshore (n=36) and Fourleague Bay (n=75), Louisiana, between 23 and 25 January 1990. n=number of larvae from each environment. NOTE Raynie and Shaw: Comparison of larval and postlarval growth rates of Brevoortia patronus 893 stages based on developmental history alone has been shown for Atlantic herring, Clupea harengus, and a 96% decrease in growth rate has been shown for bay anchovy, Anchoa mitchilli (Houde, 1987). These two clupeiform species have high larval growth rates and relatively long metamorphosis intervals (Houde, 1987) as does gulf menhaden. Lewis et al. (1972) also related varying growth in length to growth in weight through larval, prejuvenile, and juvenile stages of Atlantic menhaden, Brevoortia tyrannus. Therefore, physiological and morphological changes occurring between larval and juvenile stages may be more im- portant in the regulation of the shape of growth curves (both length and weight) than variability in exogenous factors. Acknowledgments This study was submitted as part of a M.S. thesis to the Department of Oceanography and Coastal Sci- ences, Louisiana State University, by the senior au- thor. We would like to thank J. W. Day Jr., E. B. Moser, and C. A. Wilson for their help and advise. J. H. Power deserves special thanks for his advise and assistance with the age and growth analysis, as does J. G. Ditty for assistance in the laboratory. Finan- cial support for this project was provided by the Loui- siana Sea Grant College Program, a part of the Na- tional Sea Grant College Program. Literature cited Balon, E. K. 1984. Reflections on some decisive events in the early life of fishes. Trans. Am. Fish. Soc. 113:178- 185. Deegan, L. A. 1985. The population ecology and nutrient trans- port of gulf menhaden in Fourleague Bay, Louisi- ana. Ph.D. diss., Louisiana State Univ., Baton Rouge, 134 p. 1990. Effects of estuarine environmental conditions on population dynamics of young-of-the-year gulf menhaden. Mar. Ecol.-Prog. Ser. 68:195-205. Deegan, L. A., and B. A. Thompson. 1987. Growth rate and early life history of young- of-the-year gulf menhaden as determined from otoliths. Trans. Am. Fish. Soc. 116:663-667. Fore, P. L. 1970. Oceanic distribution of the eggs and larvae of the gulf menhaden. In Report of the Bureau of Commercial Fisheries Biology Lab, Beaufort, North Carolina, for fiscal year ending June 30, 1968, p. 11-13. U.S. Fish Wildl. Serv. Circ. 341. Fore, P. L., and K. N. Baxter. 1972. Diel fluctuations in the catch of larval gulf men- haden, Brevoortia patronus, at Galveston Entrance, Texas. Trans. Am. Fish. Soc. 101(4):729-732. Herke, W. H. 1969. A boat-mounted surface push-trawl for sam- pling juveniles in tidal marshes. Prog. Fish-Cult. 31(31:177-179. Hettler, W. F. 1984. Description of eggs, larvae, and early juve- niles of gulf menhaden, Brevoortia patronus, and comparisons with Atlantic menhaden, B. tyrannus, and yellowfin menhaden, B. smithi. Fish. Bull. 82:85-95. Houde, E. D. 1987. Fish early life dynamics and recruitment variability. Am. Fish. Soc. Symp. 2:17-29. Kriete, W. H., Jr., and J. G. Loesch. 1980. Design and relative efficiency of a bow- mounted pushnet for sampling juvenile pelagic fishes. Trans. Am. Fish. Soc. 109:649-652. Laird, A. K., S. A. Tyler, and A. D. Barton. 1965. Dynamics of normal growth. Growth 29:233-248. Laroche, J. L., S. L. Richardson, and A. A. Rosenberg. 1982. Age and growth of a pleuronectid, Parophrys vetulus, during the pelagic larval period in Oregon coastal waters. Fish. Bull. 80:93-104. Lassuy, D. R. 1983. Species profiles: life histories and environ- mental requirements (Gulf of Mexico) — gulf menhaden. U.S. Fish and Wildl. Serv., Div. Biol. Serv. FWS/OBS-82/11.2 and U.S. Army Corps of Engineers, TR EL-82-4, 13 p. Laurence, C. G., A. S. Smigielski, T. A. Halavik, and B. R. Burns. 1981. Implications of direct competition between larval cod iGadus morhua) and haddock (Melano- grammus aeglefinus) in laboratory growth and sur- vival studies at different food densities. Rapp. P- V. Reun. Cons. Int. Explor. Mer 178:304-311. Lewis, R. M., E. P. H. Wilkens, and H. R. Gordy. 1972. A description of young Atlantic menhaden, Brevoortia tyrannus, in the White Oak River estu- ary, North Carolina. Fish. Bull. 70:115-118. Methot, R. D., Jr., and D. Kramer. 1979. Growth of northern anchovy, Engraulis mordax, larvae in the sea. Fish. Bull. 77:413-423. Miller, J. M. 1973. A quantitative push-net system for transect studies of larval fish and macrozooplankton. Limnol. Oceanogr. 18(1):175-178. Powell, A. B., and G. Phonlor. 1986. Early life history of Atlantic menhaden, Brevoortia tyrannus, and gulf menhaden, B. patronus. Fish. Bull. 84:991-994. Raynie, R. C, and R. F. Shaw. In press. Ichthyoplankton abundance along a re- cruitment corridor from offshore spawning to es- 894 Fishery Bulletin 92(4). 1994 tuarine nursery ground. Estuarine Coastal Shelf Sci. SAS Institute, Inc. 1985. SAS user's guide: statistics, 1985 ed. Cary, NC, SAS Institute Inc., 584 p. Shaw, R. F., B. D. Rogers, J. H. Cowan Jr., and T. L. Tillman. 1985. Distribution and density of Brevoortia patronus (gulf menhaden) eggs and larvae in the continental shelf off western Louisiana. Bull. Mar. Sci. 36:96-103. Shaw, R. F., B. D. Rogers, J. H. Cowan Jr., and W. H. Herke. 1988. Ocean-estuarine coupling of ichthyoplankton and nekton in the northern Gulf of Mexico. Am. Fish. Soc. Symp. 3:77-89. Simoneaux, L. F. 1979. The distribution of menhaden, genus Brevoortia, with respect to salinity, in the upper drainage basin of Barataria Bay, Louisiana. M.S. thesis, Louisiana State Univ., Baton Rouge, 96 p. Springer, V. G., and K. D. Woodburn. 1960. An ecological study of the fishes of the Tampa Bay area. Florida Board of Conservation Marine Laboratory Professional Paper Series 1, 104 p. Suttkus, R. D. 1956. Early life history of the gulf menhaden, Brevoortia patronus, in Louisiana. Trans. N. Am. Wildl. Conf. 21:390-407. Teague, K. G., C. J. Madden, and J. W. Day Jr. 1988. Sediment-water oxygen and nutrient fluxes in a river-dominated estuary. Estuaries 11:1-9. Warlen, S. M. 1988. Age and growth of larval gulf menhaden, Brevoortia patronus, in the northern Gulf of Mexico. Fish. Bull. 86:77-90. Warlen, S. M., and A. J. Chester. 1985. Age, growth, and distribution of larval spot, Leiostomus xanthurus, off North Carolina. Fish. Bull. 83:587-599. Zweifel, J. R., and R. Lasker. 1976. Prehatch and posthatch growth of fishes — a general model. Fish. Bull. 74:609-621. Fishery Bulletin Index Volume 92 (1-4), 1994 List of Titles 92(1) 1 Age, growth, and mortality of Atlantic croaker, Micropogonias undulatus, in the Chesapeake Bay region, with a discussion of apparent geographic changes in population dynamics, by Luiz R. Barbieri, Mark E. Chittenden Jr., and Cynthia M. Jones 13 Age and growth of the oceanic squid Onychoteuthis borealijaponica in the North Pacific, by Keith A. Bigelow 26 Diet of the Kemp's ridley sea turtle, Lepidochelys kempii, in New York waters, by Vincent J. Burke, Stephen J. Morreale, and Edward A. Standora 33 Larval development of tripletail, Lobotes surinamensis (Pisces: Lobotidae), and their spatial and temporal distribution in the northern Gulf of Mexico, by James G. Ditty and Richard F. Shaw 46 Age validation and estimation of growth rate of the coral trout, Plectropomus leopardus, (Lacepede 1802) from Lizard Island, Northern Great Barrier Reef, by Beatrice Padovani Ferreira and Garry R. Russ 58 Genetic distinctness of red drum (Sciaenops ocellatus) from Mosquito Lagoon, east-central Florida, by John R. Gold and Linda R. Richardson 67 Distribution and abundance of copepod nauplii and other small (40-300 pm) zooplankton during spring in Shelikof Strait, Alaska, by Lewis S. Incze and Terri Ainaire 79 Marine distribution and size of juvenile Pacific salmon in Southeast Alaska and northern British Columbia, by Herbert W. Jaenicke and Adrian G. Celewycz 91 Evidence for distinct stocks of king mackerel, Scomberomorus cavalla, in the Gulf of Mexico, by Allyn G. Johnson, William A. Fable Jr., Churchill B. Grimes, Lee Trent, and Javier Vasconcelos Perez 102 Reproductive biology and egg production of three species of Clupeidae from Kiribati, tropical central Pacific, by David A. Milton, Stephen J. M. Blaber and Nicholas J. F. Rawlinson 122 Examination of stock and school structure of striped dolphin (Stenella coeruleoalba) in the eastern Pacific from aerial photogrammetry, by Wayne L. Perryman and Morgan S. Lynn 132 Potential tuna catches in the eastern Pacific Ocean from schools not associated with dolphins, by Richard G. Punsly, Patrick K. Tomlinson, and Ashley J. Mullen 144 Prey selection by northern fur seals (Callorhinus ursinus) in the eastern Bering Sea, by Elizabeth Sinclair, Thomas Loughlin, and William Pearcy 157 Feeding habits of anadromous alewives, Alosa pseudoharengus , off the Atlantic Coast of Nova Scotia, by Heath H. Stone and Brian M. Jessop 171 Queen conch, Strombus gigas, reproductive stocks in the central Bahamas: distribution and probable sources, by Allan W. Stoner and Kirsten C. Schwarte 180 The influence of Apalachicola River flows on blue crab, Callinectes sapidus, in north Florida, by Dara H. Wilber 189 Oocyte maturation in Hecate Strait English sole (Pleuronectes vetulus), by Jeff Fargo and Albert V. Tyler 198 Estimation of weight-length relationships from group measurements, by William H. Lenarz 203 Spiny lobster recruitment and sea level: results of a 1990 forecast, by Jeffrey J. Polovina and Gary T. Mitchum 92(2) 207 Embryonic development of walleye pollock, Theragra chalcogramma, from Shelikof Strait, Gulf of Alaska, by Deborah M. Blood, Ann C. Matarese, and Mary M. Yoklavich 223 Diel vertical distribution of ichthyoplankton in the northern Gulf of Alaska, by Richard D. Brodeur and William C. Rugen 236 Changes in a population of orange roughy, Hoplostethus atlanticus, with commercial exploitation on the Challenger Plateau, New Zealand, by Malcolm R. Clark and Dianne M. Tracey 254 Morphometric and genetic identification of eggs of spring-spawning sciaenids in lower Chesapeake Bay, by Louis B. Daniel III and John E. Graves 262 A re-description of Atlantic spadefish larvae, Chaetodipterus faber (family: Ephippidae), and their distribution, abundance, and seasonal occurrence in the northern Gulf of Mexico, by James G. Ditty, Richard F. Shaw, and Joseph S. Cope 275 Larval development, distribution, and abundance of common dolphin, Coryphaena hippurus, and pompano dolphin, C. equiselis (family: Coryphaenidae), in the northern Gulf of Mexico, by James G. Ditty, Richard F. Shaw, Churchill B. Grimes, and Joseph S. Cope 292 Using Pb-210/Ra-226 disequilibria for sablefish, Anoplopoma fimbria, age validation, by Craig R. Kastelle, Daniel K. Kimura, Ahmad E. Nevissi, and Donald R. Gunderson 895 896 INDEX: TITLES Fishery Bulletin 92(1-4), 1994 302 Distribution and abundance of two juvenile tropical Photololigo species (Cephalopoda: Loliginidae) in the central Great Barrier Reef Lagoon, by Natalie A. Moltschaniwskyj and Peter J. Doherty 313 Observations on the distribution and activities of rockfish, Sebastes spp., in Saanich Inlet, British Columbia, from the Pisces TV submersible, by Debra J. Murie, Daryl C. Parkyn, Bruce G. Clapp, and Geoffrey G. Krause 324 Reexamination of geographic variation in cranial morphology of the pantropical spotted dolphin, Stenella attenuata, in the eastern Pacific, by William F. Perrin, Gary D. Schnell, Daniel J. Hough, James W. Gilpatrick Jr., and Jerry V. Kashiwada 347 Modeling oyster populations. IV: Rates of mortality, population crashes, and management, by Eric N. Powell, John M. Klinck, Eileen E. Hofmann, and Sammy M. Ray 374 A suite of extensions to a nonequilibrium surplus- production model, by Michael H. Prager 390 Experimental outplanting of juvenile queen conch, Strombus gigas: comparison of wild and hatchery- reared stocks, by Allan W. Stoner and Megan Davis 412 A snow crab, Chionoecetes opilio (Decapoda, Majidae), fishery collapse in Newfoundland, by David M. Taylor, Paul G. O'Keefe, and Charles Fitzpatrick 420 Spawning time and recruitment dynamics of larval Atlantic menhaden, Brevoortia tyrannus, into a North Carolina estuary, by Stanley M. Warlen 434 Interannual variability of dolphin habitats in the eastern tropical Pacific. I: Research vessel surveys, 1986-1990, by Stephen B. Reilly and Paul C. Fiedler 451 Interannual variability of dolphin habitats in the eastern tropical Pacific. II: Effects on abundances estimated from tuna vessel sightings, 1975-1990, by Paul C. Fiedler and Stephen B. Reilly 464 Annual mass strandings of pelagic red crabs, Pleuroncodes planipes (Crustacea: Anomura: Galatheidae), in Bahia Magdalena, Baja California Sur, Mexico, by David Aurioles-Gamboa, Maria Isabel Castro-Gonzalez, and Ricardo Perez-Flores 471 Mass marking coho salmon, Oncorhynchus kisutch, fry with lanthanum and cerium, by Bridget C. Ennevor 474 Distribution and relative abundance of the blue shark, Prionace glauca, in the southwestern equatorial Atlantic Ocean, by Fabio H. V. Hazin, Clara E. Boeckman, Elizabeth C. Leal, Rosangela P. T. Lessa, Kohei Kihara, and Kazuyuki Otsuka 92(3) 481 Environmentally induced recruitment variation in petrale sole, Eopsetta jordani, by Gonzalo C. Castillo, Hiram W. Li, and James T. Golden 494 The distribution, food, and age of juvenile bluefish, Pomatomus saltatrix, in Maine, by Edwin P. Creaser and Herbert C. Perkins 509 Differences between the sagitta, lapillus, and asteriscus in estimating age and growth in juvenile red drum, Sciaenops ocellatus, by Andrew W. David, J. Jeffery Isely, and Churchill B. Grimes 516 Reproductive and trophic ecology of the soldierfish Myripristis amaena in tropical fisheries, by Anderson J. Dee and James D. Parrish 531 Reproductive biology of the cockfish, Callorhynchus callorhynchus (Holocephali: Callorhynchidae), in Patagonian waters (Argentina), by Edgardo E. Di Giacomo and Maria Raquel Perier 540 The vertical distribution of eggs and larvae of walleye pollock, Theragra chalcogramma, in Shelikof Strait, Gulf of Alaska, by Arthur W. Kendall Jr., Lewis S. Incze, Peter B. Ortner, Shailer R. Cummings, and Patricia K. Brown 555 A comparison of a validated otolith method to age weakfish, Cynoscion regalis, with the traditional scale method, by Susan K. Lowerre-Barbieri, Mark E. Chittenden Jr., and Cynthia M. Jones 569 Feeding habits of the dusky dolphin, Lagenorhynchus obscurus, in the coastal waters of central Peru, by Jeff McKinnon 579 Distribution, abundance, and growth of larval walleye pollock, Theragra chalcogramma, in an Alaskan fjord, by Franz -Josef Muter and Brenda L. Norcross 591 Interannual variation in the recruitment pattern and abundance of age-0 summer flounder, Paralichthys dentatus, in Virginia estuaries, by Brenda L. Norcross and David M. Wyanski 599 Spatio-temporal structure of the deep-water shrimp Aristeus antennatus (Decapoda: Aristeidae) population in the western Mediterranean, by Francisco Sarda, Joan E. Cartes, and Walter Norbis 608 Predator-prey relationships of winter flounder, Pleuronectes amerieanus, in the New York Bight apex, by Frank W. Steimle, Dorothy Jeffress, Stephen A. Fromm, Robert N. Reid, Joseph J. Vitaliano, and Ann Frame INDEX: TITLES Fishery Bulletin 92(1-4). 1994 897 620 Rockfish assemblages of the middle shelf and upper slope off Oregon and Washington, by Kenneth L. Weinberg 633 Results of long-term, seasonal sampling for Penaeus postlarvae at Breach Inlet, South Carolina, by Lawrence B. DeLancey, James E. Jenkins, and J. David Whitaker 641 Swimbladder deflation in the Atlantic menhaden, Brevoortia tyrannus, by Richard B. Forward Jr., William F. Hettler, and Donald E. Hoss 647 An energy budget for northern sand lance, Ammodytes dubius, on Georges Bank, 1977-1986, by Sharon L. Gilman 665 Movements of tagged adult yellowtail rockfish, Sebastes flavidus, off the west coast of North America, by Richard D. Stanley, Bruce M. Leaman, Lewis Haldorson, and Victoria M. O'Connell 664 A simple generalized model of allometry, with examples of length and weight relationships for 14 species of groundfish, by Yongshun Xiao and David C. Ramm 92(4) 671 Maturity, spawning, and ovarian cycle of Atlantic croaker, Micropogonias undulatus, in the Chesapeake Bay and adjacent coastal waters, by Luiz R. Barbieri, Mark E. Chittenden Jr. , and Susan K. Lowerre-Barbieri 686 Biochemical and histological changes during ovarian development of cobia, Rachycentron canadum, from the northern Gulf of Mexico, by Patricia M. Biesiot, Robert E. Caylor, and James S. Franks 697 Seasonal cycles in the sonic muscles of the weakfish, Cynoscion regalis, by Martin A. Connaughton and Malcolm H. Taylor 704 Sea otter, Enhydra lutris, prey composition and foraging success in the northern Kodiak Archipelago, by Angela M. Doroff and Anthony R. DeGange 711 A biomass-based assessment model for northern anchovy, Engraulis mordax, by Larry D. Jacobson, Nancy C. H. Lo, and J. Thomas Barnes 725 Status of early life history descriptions of marine teleosts, by Arthur W. Kendall Jr. and Ann C. Matarese 737 Catch and effort analysis of the reef fisheries of Jamaica and Belize, by Julian A. Koslow, Karl Aiken, Stephanie Auil, and Antoinette Clementson 748 Modelling the distribution of alfonsino, Beryx splendens, over the seamounts of New Caledonia, by Patrick Lehodey, Paul Marchal, and Rene Grandperrin 760 Spawning and early life history of white sturgeon, Acipenser transmontanus, in the lower Columbia River, by George T. McCabe Jr. and Charles A. Tracy 773 Cetaceans on the upper continental slope in the north-central Gulf of Mexico, by Keith D. Mullin, Wayne Hoggard, Carol L. Roden, Ren R. Lohoefener, Carolyn M. Rogers, and Brian Taggart 787 A video transect method for estimating reef fish abundance, composition, and habitat utilization at Gray's Reef National Marine Sanctuary, Georgia, by Richard O. Parker Jr., Alexander J. Chester, and Russell S. Nelson 800 Seasonal movements of Pacific cod, Gadus macrocephalus, in the eastern Bering Sea and adjacent waters based on tag-recapture data, by Allen M. Shimada and Daniel K. Kimura 817 An evaluation of electron-probe microanalysis of otoliths for stock delineation and identification of nursery areas in a southern temperate groundfish, Nemadactylus macropterus (Cheilodactylidae), by Ronald E. Thresher, Craig H. Proctor, John S. Gunn, and Ian R. Harrowfield 841 Reproductive biology of red drum, Sciaenops ocellatus, from the neritic waters of the northern Gulf of Mexico, by Charles A. Wilson and David L. Nieland 851 Biology and population characteristics of Squalus mitsukurii from a seamount in the central North Pacific Ocean, by Christopher D. Wilson and Michael P. Seki 865 Measures of dispersion as constraints for length- frequency analysis, by Karim Erzini and Margarida Castro 872 Migratory characteristics of juvenile ocean-type chinook salmon, Oncorhynchus tshawytscha, in John Day Reservoir on the Columbia River, by Albert E. Giorgi, David R. Miller, and Benjamin P. Sandford 880 Comparisons between generalized growth curves for two estuarine populations of the eel tailed catfish Cnidoglanis macrocephalus, by Laurie J. B. Laurenson, Ian C. Potter, and Norm G. Hall 890 A comparison of larval and postlarval gulf menhaden, Brevoortia patronus, growth rates between an offshore spawning ground and an estuarine nursery, by Richard C. Raynie and Richard F. Shaw Fishery Bulletin Index Volume 92 (1-4), 1994 List of Authors Aiken, Karl 737 Ainaire, Terri 67 Auil, Stephanie 737 Aurioles-Gamboa, David 464 Barbieri, Luiz R. 1, 671 Barnes, J. Thomas 711 Biesiot, Patricia M. 686 Bigelow, Keith A. 13 Blaber, Stephen J. M. 102 Blood, Deborah M. 207 Boeckman, Clara E. 474 Brodeur, Richard D. 223 Brown, Patricia K. 540 Burke, Vincent J. 26 Cartes, Joan E. 599 Castillo, Gonzalo C. 481 Castro, Margarida 865 Castro-Gonzalez, Maria Isabel 464 Caylor, Robert E. 686 Celewycz, Adrian G. 79 Chester, Alexander J. 787 Chittenden, Mark E., Jr. 1,555,671 Clapp, Bruce G. 313 Clark, Malcolm R. 236 Clementson, Antoinette 737 Connaughton, Martin A. 697 Cope, Joseph S. 262,275 Creaser, Edwin P. 494 Cummings, Shailer R. 540 Daniel, Louis B. Ill 254 David, Andrew W. 509 Davis, Megan 390 Dee, Anderson J. 516 DeGange, Anthony R. 704 DeLancey, Lawrence B. 633 Di Giacomo, Edgardo E. 531 Ditty, James G. 33, 262, 275 Doherty, Peter J. 302 Doroff, Angela M. 704 Ennevor, Bridget C. 471 Erzini, Karim 865 Fable, William A., Jr. 91 Fargo, Jeff 189 Ferreira, Beatrice Padovani 46 Fiedler, Paul C. 434,451 Fitzpatrick, Charles 412 Forward, Richard B., Jr. 641 Frame, Ann 608 Franks, James S. 686 Fromm, Stephen A. 608 Gilman, Sharon L. 647 Gilpatrick, James W., Jr. 324 Giorgi, Albert E. 872 Gold, John R. 58 Golden, James T. 481 Grandperrin, Rene 748 Graves, John E. 254 Grimes, Churchill B. 91, 275, 509 Gunderson, Donald R. 292 Gunn, John S. 817 Haldorson, Lewis 655 Hall, Norm G. 880 Harrowfield, Ian R. 817 Hazin, Fabio H. V. 474 Hettler, William F. 641 Hofmann, Eileen E. 347 Hoggard, Wayne 773 Hoss, Donald E. 641 Hough, Daniel J. 324 Incze, Lewis S. 67, 540 Isely, J. Jeffery 509 Jacobson, Larry D. 711 Jaenicke, Herbert W. 79 Jeffress, Dorothy 608 Jenkins, James E. 633 Jessop, Brian M. 157 Johnson, Allyn G. 91 Jones, Cynthia M. 1,555 Kashiwada, Jerry V. 324 Kastelle, Craig R. 292 Kendall, Arthur W., Jr. 540, 725 Kihara, Kohei 474 Kimura, Daniel K. 292, 800 Klinck, John M. 347 Koslow, Julian A. 737 Krause, Geoffrey G. 313 Laurenson, Laurie J. B. 880 Leal, Elizabeth C. 474 Leaman, Bruce M. 655 Lehodey, Patrick 748 Lenarz, William H. 198 Lessa, Rosangela P. T. 474 Li, Hiram W. 481 Lo, Nancy C.H. 711 Ix>hoefener, Ren R. 773 Loughlin, Thomas 144 Lowerre-Barbieri, Susan K. 555,671 Lynn, Morgan S. 122 Marchal, Paul 748 Matarese, Ann C. 207, 725 McCabe, George T, Jr. 760 McKinnon, Jeff 569 Miller, David R. 872 Milton, David A. 102 Mitchum, Gary T. 203 Moltschaniwskyj, Natalie A. 302 Morreale, Stephen J. 26 Mullen, Ashley J. 132 Mullin, Keith D. 773 Murie, Debra J. 313 Muter, Franz-Josef 579 Nelson, Russell S. 787 Nevissi, Ahmad E. 292 Nieland, David L. 841 Norbis, Walter 599 Norcross, Brenda L. 579, 591 O'Connell, Victoria M. 655 O'Keefe, Paul G. 412 Ortner, Peter B. 540 Otsuka, Kazuyuki 474 Parker, Richard O. Jr. 787 Parkyn, Daryl C. 313 Parrish, James D. 516 Pearcy, William 144 Perez, Javier Vasconcelos 91 Perez-Flores, Ricardo 464 Perier, Maria Raquel 531 Perkins, Herbert C. 494 Perrin, William F. 324 Ferryman, Wayne L. 122 Polovina, Jeffrey J. 203 Potter, Ian C. 880 Powell, Eric N. 347 Prager, Michael H. 374 Proctor, Craig H. 817 Punsly, Richard G 132 Ramm, David C. 664 Rawlinson, Nicholas J. F. 102 Ray, Sammy M. 347 Raynie, Richard C. 890 Reid, Robert N. 608 Reilly, Stephen B. 434, 451 Richardson, Linda R. 58 Roden, Carol L. 77.3 Rogers, Carolyn M. 773 Rugen, William C. 223 Russ, Garry R. 46 Sandford, Benjamin P. 872 Sarda, Francisco 599 Schnell.GaryD. 324 Schwarte, Kirsten C. 171 Seki, Michael P. 851 Shaw, Richard F. 33, 262, 275, 890 Shimada, Allen M. 800 Sinclair, Elizabeth 144 Standora, Edward A. 26 Stanley, Richard I). 655 Steimle, Frank W. 608 Stone, Heath H. 157 Stoner, Allan W. 171,390 Taggart, Brian 773 Taylor, David M. 412 Taylor, Malcolm H. 697 Thresher, Ronald E. 817 898 INDEX. AUTHORS Fishery Bulletin 92( 1-4). 1994 899 Tomlinson, Patrick K. 132 Tracey, Dianne M. 236 Tracy, Charles A. 760 Trent, Lee 91 Tyler, Albert V. 189 Vitaliano, Joseph J. 608 Warier., Stanley M. 420 Weinberg, Kenneth L. 608 Whitaker, J. David 633 Wilber, Dara H. 180 Wilson, Charles A. 841 Wilson, Christopher D. 851 Wyanski, David M. 591 Xiao, Yongshun 664 Yoklavich, Mary M. 207 Fishery Bulletin Index Volume 92 (1-4), 1994 List of Subjects Abundance — see also Population studies anchovy, northern 711 cetaceans 773 conch, queen 171 copepod nauplii 67 crab, snow 412 dolphin common 451 eastern spinner 451 eastern tropical Pacific 451 spotted 451 eggs clupeids, tropical 102 crab, snow 412 roughy, orange 236 flounder, summer 591 ichthyoplankton, Gulf of Alaska 223 larvae dolphin [fish] common 275 pompano 275 menhaden, Atlantic 420 reef fish 787 sole, petrale 481 spadefish, Atlantic 262 macrofauna, benthic 608 menhaden, Atlantic, larvae 420 oyster, eastern 347 pollock, walleye, larvae 579 reef fish 787 rockfish, submersible observations 313 roughy, orange 236 salmon, Pacific 79 shark blue 474 demersal 851 shrimp, postlarval 633 soldierfish, brick 516 squid, tropical 302 squirrelfish 516 zooplankton 67 Acipenseridae — see Sturgeon, white 760 Acipenser transmontanus — see Sturgeon, white 760 Aerial survey cetaceans in the Gulf of Mexico 773 dolphin, striped 122 Age at sexual maturity clupeids, tropical 102 conch, queen 171 soldierfish, brick 516 squirrelfish 516 Age determination bluefish, juvenile 494 catfish, eel tailed 880 comparison of hard parts 555 conch, queen 171 Age determination (continued) coral trout 46 croaker, Atlantic 1 drum, red 509 menhaden, Atlantic, larvae 420 otoliths bluefish, juvenile 494 catfish, eel tailed 880 comparison of otolith types 509 coral trout 46 drum, red 509 menhaden, Atlantic, larvae 420 pollock, walleye, larvae 579 radiometric 292 sablefish 292 pollock, walleye, larvae 579 radiometric 292 sablefish 292 shark, demersal 851 spines, shark 851 statoliths, clubhook squid 13 weakfish 555 Age-size estimation conch, queen 173 croaker, Atlantic 1 menhaden, Atlantic, larvae 420 method of 865 pollock, walleye, larvae 579 squid, clubhook 13 Age validation catfish, eel tailed 880 coral trout 46 radiometric 292 sablefish 292 weakfish 555 Alaska pollock, walleye, larvae 540, 579 Alaska, southeastern juvenile salmon 79 sea otter 704 Alewife, anadromous 157 Alfonsino, length-frequency distributions 748 Allometry 664 Allozymes 58 Alosa pseudoharengus — see Alewife, anadromous 157 Amblygaster sirm — see Herring 102 Ammodytidae — see Sand lance, northern 647 Ammodytes dubius — see Sand lance, northern 647 Anchovy anchoveta, prey of dusky dolphin 569 northern, assessment model 711 Anoplopoma fimbria — see Sablefish 292 Anoplopomatidae — see Sablefish 292 Argentina cockfish 531 Aristeidae — see Shrimp, deep-water 599 Aristeus antennatus — see Shrimp, deep-water 599 Assemblages, rockfish 620 Atlantic Bight, south bluefish, juvenile 494 Atlantic croaker 1,671 Atlantic Ocean alewife, anadromous 157 bluefish, juvenile 494 cockfish 531 conch, queen 171 crab, snow 412 croaker, Atlantic 1,671 drum, red 58 menhaden, Atlantic, larvae 420 reef fish 787 sciaenids 254 shark, blue 474 shrimp, postlarvae 633 swordfish 374 Australia catfish, eel tailed 880 Great Barrier Reef 46 groundfish length and weight relationships 664 morwong, jackass 817 squid, tropical 302 Bairdiella ehrysoura — see Perch, silver 254 Baitfishes, tropical Pacific 102 Bahama Islands conch, queen 171,390 Behavior crab, pelagic red 464 diel feeding, alewife 157 diel vertical migration ichthyoplankton 223 pollock, walleye, larvae 540 rockfish 313 Belize 737 Bering Sea cod, Pacific 800 northern fur seals 144 von Bert alanffy growth 1,46 coral trout 46 croaker, Atlantic 1 Beryx splendens — see Alfonsino 748 Bias in biomass estimation 711 Biochemical analysis cobia, ovaries 686 crab, pelagic red 464 Biological indicators 608 Biological rhythm menhaden, Atlantic 641 Bivalve oyster population modeling 347 prey of sea otters 704 Bluefish, juvenile 494 Bothidae — see Flounder, summer Brevoortia patronus — see Menhaden, gulf tyrannus — see Menhaden, Atlantic 900 INDEX: SUBJECTS Fishery Bulletin 92( 1-4). 1994 901 Butterfish, black-spot 664 Bycateh 132 California Baja anchovy, northern 711 crab, pelagic red 464 Bight, southern anchovy, northern 711 Callinectes Bapidus — see Crab, blue 180 Callorhinus ursinus — see Seal, northern fur 144 Callorhynchidae — see Cockfish 531 Callorhynchus callorhynchus — see Cockfish 531 Canada British Columbia rockfish 313 salmon 79 sole, English 189 Ne wf oundl and crab, snow 412 Nova Scotia alewife, anadromous 157 Capelin, in northern fur seal diet 144 Carangidae — see Mackerel, jack 569 Carcharhinidae — see Shark, blue 474 Caribbean conch, queen 171, 390 reef fish 737 Carolina coast drum, red 58 menhaden, Atlantic, larvae 420 shrimp, postlarval 633 Catch estimation — see also Population studies roughy, orange 236 tuna, eastern Pacific spp. 132 Catch-per-unit-of-effort clupeids, tropical 102 crab, snow 412 reef fish 737 shark, blue 474 squid, tropical 302 Catch rates crab, blue 180 roughy, orange 236 Catfish, eel tailed 880 Cephalopoda — see Squid Cetaceans, abundance in the Gulf of Mexico 773; see also Dolphin; Whale Chaetodipterus faber — see Spadefish, Atlantic 262 Chemical marking 471 Chesapeake Bay croaker, Atlantic 1,671 flounder, summer 591 oyster, eastern 347 sciaenids, eggs 254 Chilipepper 198 Chionoecetes opilio — see Crab, snow 412 Chinook salmon, ocean-type 872 Clams, prey of sea otters 704 Classification — see Taxonomy Clupea pallasi — see Herring, Pacific Clupeidae — see Herring; Menhaden; Sardine, Pacific; Sprat 102 Cnidoglanis macrocephalus — see Catfish, eel tailed 880 Cobia 686 Cockfish 531 Cod, Pacific 800 Columbia River chinook salmon, ocean-type 872 sturgeon, white 760 Community structure, reef fish 787 Conch, queen 171, 390 Contaminant effects in winter flounder 608 Copepods, nauplii 67 Coral reef fishes 737,787 Coral trout 46 Coryphaeruj equiselis — see Dolphin [fish], common 275 hippurus — see Dolphin [fish], pompano 275 Coryphaenidae — see Dolphin [fish] 275 Crab blue 180 pelagic red 464 snow 412 Telmessus, prey of sea otters 704 Cranial morphology of dolphin 324 Crassostrea virginica — see Oyster, eastern 347 Croaker, Atlantic 1, 671 Cynoscion nebulosus — see Seatrout, spotted 254 regalis — see Weakfish 254, 555, 697 Daily ration 157 Deep sea 599 Delphinidae — see Dolphin Delphinus delphis — see Dolphin, common Density copepods, nauplii 67 ichthyoplankton 223 menhaden, Atlantic, larvae 420 reef fishes 787 rockfish 313 copper 313 greenstriped 313 quillback 313 tiger 313 yelloweye 313 yellowtail 313 roughy, orange 236 zooplankton 67 Depth distribution alfonsino 748 conch, queen 171 copepod, nauplii 67 ichthyoplankton 223 pollock, walleye, eggs 540 rockfish assemblages 620 species 313 yelloweye 313 shark, blue 474 shark, demersal 851 Depth distribution (continued) shrimp 599 squid, tropical 302 zooplankton 67 Development dolphin [fish], larvae common 275 pompano 275 pollock, walleye, eggs 540 spadefish, Atlantic, larvae 262 tripletail, larvae 33 Diagram ma pictum — see Grunt 664 Diet alewife, anadromous 157 bluefish, juvenile 494 dolphin, dusky 569 flounder, winter 608 seal, northern fur 144 sea otter 704 sea turtle, Kemp's ridley 26 shark, demersal 851 soldierfish, brick 516 squirrelfish 516 Diel variation alewife, diet 157 Dispersion, length at age 865 Distribution bluefish, juvenile 494 cetaceans 773 cod, Pacific 800 conch, queen 171 contraction of 236 copepods, nauplii 67 dolphin common 434, 451 spinner 434, 451 spotted 434, 451 striped 434, 451 dolphin [fish] larvae common 275 pompano 275 ichthyoplankton, Gulf of Alaska 223 pollock, walleye 579 eggs and larvae 540 rockfish, assemblages 620 copper 313 greenstriped 313 quillback 313 tiger 313 yelloweye 313 yellowtail 313,620,655 roughy, orange 236 salmon, juveniles 79 shark, blue 474 shrimp 599 spadefish, Atlantic, larvae 262 squid, tropical 302 tripletail 33 Dolphin Atlantic spotted 773 bottlenose 773 clymene 773 common 132,434,451 dusky 569 eastern tropical Pacific spp. 132, 434, 451 902 INDEX: SUBJECTS Fishery Bulletin 92| 1 -4), 1994 Dolphin (continued) Risso's 773 spinner 132, 324, 434, 451, 773 eastern 132 spotted 132,324,434,451,773 striped 122,434,451,773 Dolphin [fish] common 275 pompano 275 Dosidicas gigas — see Squid, jumbo Hying 569 Driftnets 13 Drum black 254 red 58, 509, 841 Dumping ocean 608 sewage sludge 608 Early-life-history studies dolphin [fish] common 275 pompano 275 drum, red 23 flounder, summer 591 menhaden, Atlantic 420 pollock, walleye 207, 223, 540, 579 spadefish, Atlantic 262 status of 725 sturgeon, white 760 tripletail 33 Egg studies clupeids, tropical 102 cobia 686 drum, red 841 menhaden, Atlantic 420 pollock, walleye 207, 540 sciaenids 254 soldierfish, brick 516 sole, English 189 squirrelfish 516 status of descriptions 725 sturgeon, white 760 Embryonic development of walleye pollock 207 Embryos — see Larval studies Emperor blue-lined 664 red 664 red-spot 664 Endangered species 26 Energetics cobia, eggs 686 oyster, eastern 347 sand lance, northern 647 Engraulidae — see Anchovy Engraulis mordax — see Anchovy, northern 711 ringens — see Anchovy, anchoveta 569 Enhancement, queen conch 390 Enhydra lutris — see Otter, sea 704 Environmental effects crab, blue 180 crab, snow 412 clupeids, tropical 102 Environmental effects (continued) dolphin 434,451 common 434, 451 spinner 434, 451 spotted 434,451 striped 434, 451 flounder, winter, diet 608 ichthyoplankton, distribution 223 lobster, recruitment 203 shark, blue 474 sole, petrale 481 squid, distribution 302 zooplankton 67 Environmental-morphological covariation 324 Eopsetta jordani — see Sole, petrale 481 Ephippidae — see Spadefish, Atlantic 262 Estuarine organisms catfish, eel tailed 880 crab, blue 180 croaker, Atlantic 1,671 flounder summer 591 winter 608 menhaden Atlantic 420 gulf 890 morwong, jackass 817 oyster, eastern 347 Euphausiidae — see Meganyctiplianes norvegica 157 Fecundity clupeids, tropical 102 drum, red 841 soldierfish, brick 516 squirrelfish 516 Feeding — see Food habits Fishery coral reef fishes 737 crab, blue 180 crab, snow 412 lobster, Hawaiian spiny 203 recruitment overfishing 516 roughy, orange 236 shrimp 59(1 small scale 516 squid, clubhook 13 tuna, eastern tropical spp. 132 Fishery interactions tuna-dolphin 132 Fishery management coral reef fishes 737 dolphin 434,451 common 434, 451 spinner 434, 451 spotted 434,451 striped 434,451 enhancement, queen conch 390 models 718 age-structured 711 surplus production 374,711 Graham-Schaefer 374 multispecies 620 Fishery management (continued) roughy, orange 236 size limits, eastern oyster 347 stock-recruitment, shrimp 633 tuna, eastern Pacific spp. 132 weight-length relationships 198 Fishery reserves conch, queen 171 Fishes, coral reef 516, 737 Fjords 579 Flatfishes 591,608 Florida conch, queen 390 crab, blue 180 drum, red 58 Flounder summer, recruitment in 591 winter, predator-prey relationships of 608 Food habits alewife, anadromous 157 bluefish, juvenile 494 crab, pelagic red 464 dolphin, dusky 569 flounder, winter 608 seal, northern fur 144 sea otter 704 sea turtle, Kemp's ridley 26 shark, demersal 851 soldierfish, brick 516 squirrelfish 516 Freshwater inflow, effect on blue crab 180 Gadidae — see Cod; Hake; Pollock, walleye Gadus maerocephalus — see Cod, Pacific 800 Galatheidae — see Crab, pelagic red 464 Galveston Bay 347 Genetic studies drum, red 58 mackerel, king 91 sciaenids, eggs 254 species identification, sciaenids 254 stock identification, king mackerel 91 Geographic variation anadromous alewife, diet 157 croaker, Atlantic 1 dolphin, pantropical spotted, morphology 324 drum, red 58 mackerel, king 91 orange roughy, abundance 236 oyster, eastern, mortality 347 red drum, genetic structure 58 salmon, distribution 79 Georges Bank sand lance, northern 647 Gonatidae — see Squid Grampus griseus — see Dolphin, Kisso's 773 Growth — see also Age-size estimation catfish, eel tailed 880 conch, queen 171, 390 coral trout 46 INDEX: SUBJECTS Fishery Bulletin 92(1-4). 1994 903 Growth (continued) croaker, Atlantic 1 drum, red 509 menhaden, gulf 890 pollock, walleye, larvae 579 shark, demersal 851 squid, clubhook 13 sturgeon, white 760 weakfish 555 Gulf of Alaska copepods, nauplii 67 ichthyoplankton 223 pollock, walleye embryonic development 207 eggs and larvae 540 zooplankton 67 Gulf of Mexico cetaceans 773 cobia 686 dolphin (fish], larvae common 275 pompano 275 drum, red 58,841 mackerel, king 91 menhaden, gulf 890 oyster, eastern 347 spadefish, Atlantic, larvae 262 tripletail, larvae 33 Habitat conch, queen 171 dolphin 434, 451 common 434, 451 spinner 434, 451 spotted 434, 451 striped 434, 451 reef fish 787 rockfish 313 Haemulidae — see Grunt 664 Hake, Merluccius gayi, prey of dusky dolphin 569 Harvest refugia, conch, queen 171 Hatchery/wild, queen conch 390 Hawaiian Islands dolphin, spotted 324 lobster, spiny 203 soldierfish, brick 516 squirrelfish 516 Hawaiian Islands fishery lobster, spiny 203 Herklotsichthys quadrimaculatus — see Herring 102 Herd size, cetacean 773 Herring 102 Amblygaster sirm 102 Herklotsichthys quadrimaculatus 102 menhaden, Atlantic 420 menhaden, gulf 890 Pacific, in northern fur seal diet 144 Spratelloides delicatulus 102 Hippoglossoides elassodon — see Sole, flathead 223 Holocentridae — see Soldierfish and Squirrelfish 516 Holocephali 531 Hoplostethus atlanticus — see Roughy, orange 236 Hormones, weakfish 697 Hydrography 67, 180, 203 crab, blue 180 dolphin 434, 451 common 434, 451 spinner 434, 451 spotted 434, 451 striped 434, 451 lobster, spiny 203 pollock, walleye, larvae 579 sole, petrale 481 zooplankton 67 Ichthyoplankton 33, 223, 725 Identification eggs 725 larvae 725 dolphin, striped 122 king mackerel, stock 91 sciaenid eggs, genetic and morphological 254 spadefish, Atlantic, larvae 262 tripletail, larvae 33 Impact assessment sewage disposal 608 Interannual variation dolphin abundance 451 habitat 434 flounder, summer, recruitment 59 lobster, spiny, recruitment 203 northern fur seal, diet 144 Jamaica 737 Juvenile studies bluefish 494 chinook salmon, ocean-type 872 conch, queen 390 drum, red 509 salmon 79 sea turtle, Kemp's ridley 46 squid, tropical 302 Katsuwonus pelamis — see Tuna, skipjack 132 Kingfish, northern 254 Kogia breviceps — see Whale, pygmy sperm 773 simus — see Whale, dwarf sperm 773 Lagenorhynchus obscurus — see Dolphin, dusky 569 Larval studies dolphin [fish] common 275 pompano 275 menhaden Atlantic 420,641 gulf 890 pollock, walleye 207, 540, 579 sole, petrale 481 spadefish, Atlantic 262 status of descriptions 725 Larval studies (continued) sturgeon, white 760 swimbladder deflation 641 tripletail, development and distribution in Gulf of Mexico 33 Latitudinal variation oyster, eastern, mortality 347 red drum, genetic structure 58 salmon, distribution 79 Length-based sampling 122 Length-frequency analysis as measure of dispersion 865 soldierfish, brick 516 squid, tropical 302 squirrelfish 516 Length studies — see also Age-size estimation alfonsino 748 catfish, eel tailed 880 Length-weight relationships 198, 664 Lepidochelys kempii — see Sea turtle, Kemp's ridley 26 Lethrinidae — see Emperor Life history pollock, walleye 207 shark, demersal 851 soldierfish, brick 516 squirrelfish 516 Light swimbladder deflation, effects on 641 Light traps 302 Lizardfish, short-finned 664 Lobotes Surinam crisis — see Tripletail Lobotidae — see Tripletail Lobster, Hawaiian spiny 203 Loliginidae — see Squid Loligo chinensis — see Photololigo chinensis 302 gayi — see Squid, Patagonian 569 Lutjanidae — see Snapper Lutjanus — see Snapper 664 erythropterus — see Snapper, scarlet 664 malabaricus — see Snapper, saddle-tailed 664 sebae — see Snapper, red emperor 664 timorensis — see Snapper, Timor 664 vittus — see Snapper, one-band 664 Mackerel jack, prey of dusky dolphin 569 king 91 Macrofauna, benthic 608 Maine, bluefish 494 Majidae — see Crab, snow 412 Mallotus villosus — see Capelin 144 Management — see Fishery management Mass marking 471 Mass stranding, pelagic red crab 464 Mathematical methods estimation of weight-length relationships 198, 664 numerical classification 620 simulations 347,711 Maximum sustainable yield 374 904 INDEX: SUBJECTS Fishery Bulletin 92| 1-4), 1994 Mediterranean 599 Meganyctiphanes norvegica, in diet of anadromous alewives 157 Menhaden Atlantic 420,641 gulf 890 Menticirrhus saxatilis — see Kingfish, northern Merlucciidae — see Hake 569 Merluccius gayi, prey of dusky dolphin 569 Metamorphosis menhaden, gulf 890 Methods ageing comparison of hard parts 555 comparison of otolith types 509 length-based 122 light trap, sampling 302 radiometric ageing 292 Mexico anchovy, northern 711 crab, pelagic red 464 Microchemistry, otolith 817 Micropogonias undulatus — see Croaker, Atlantic 671 Microzooplankton 67 Mid-Atlantic Bight menhaden, Atlantic 420 Migration — see Movements Mitochondrial DNA analysis drum, red 58 sciaenids, eggs 254 Models alfonsino, depth distribution 748 allometry 664 length at age 865 energetic 347 growth Schnute 13 von Bertalanffy 1,46 length-depth 748 stock assessment age-structured 711 surplus production 374,711,737 Graham-Schaefer 374 Molting, snow crab 412 Morphology cranial variation, pantropical spotted dolphin 324 dolphin [fish], larvae common 275 pompano 275 pollock, walleye, eggs 207 queen conch, shell 171, 390 reproductive 531 sciaenids, eggs 254 spadefish, Atlantic, larvae 262 spotted dolphin, cranial 324 tripletail, larvae 33 Morphometries dolphin [fish], larvae common 275 pompano 275 dolphin, spotted, cranial 324 spadefish, Atlantic, larvae 262 Morphometries (contiued) tripletail, larvae 33 weakfish, sonic muscles 697 Mortality crab pelagic red 464 snow 412 croaker, Atlantic 1 due to marking 471 oyster, eastern 347 queen conch, hatchery/wild 390 roughy, orange 236 Morwong, jackass 817 Movements chinook salmon, ocean-type 872 cod, Pacific 800 conch, queen 171 copepods, nauplii 67 flounder, summer 591 mackerel, king 91 menhaden, Atlantic, larvae 420 pollock, walleye, eggs 540 rockfish, yellowtail 655 salmon, juveniles 79 shrimp, postlarval 633 sole, petrale 481 zooplankton 67 Muscle, sonic 697 Mussel, prey of sea otters 704 Mustelidae — see Otter, sea 704 Myripristis amaena — see Soldierfish 516 Mytilus — see Mussel 704 Nemipteridae — see Threadfin bream 664 Nem ipterus furcosus — see Threadfin bream, rosy 664 hexodon — see Threadfin bream, ornate 664 New Caledonia 748 New York Bight 26,608 predator-prey relationships of winter flounder 608 sea turtle, Kemp's ridley 26 sewage disposal 608 New Zealand, orange roughy 236 Ocean dumping 608 Ommastrephes bartramii — see Squid, oceanic 13 Ommastrephidae — see Squid, oceanic 13 Oncorhynchus gorbuscha — see Salmon, pink Oncorhynchus keta — see Salmon, chum Oncorhynchus kisutch — see Salmon, coho Oncorhynchus nerka — see Salmon, sockeye Oncorhynchus species — see Salmon, Pacific Oncorhynchus tshawytscha — see Salmon, chinook Onychoteuthidae — see Squid, oceanic Onychoteuthis borealijaponica — see Squid, oceanic; Squid, boreal clubhook 13 Oocytes of English sole 189 Oregon rockfish, assemblages 620 sole, petrale 481 Osmeridae — see Capelin 144 Otoliths ageing bluefish, juvenile 494 catfish, eel tailed 880 croaker, Atlantic 1 comparison of 509 drum, red 509 menhaden, Atlantic, larvae 420 pollock, walleye, larvae 579 radiometric 292 sablefish 292 weakfish 555 annuli validation 46 coral trout 46 croaker, Atlantic 1 drum, red 509 growth 46 microchemistry 817 pollock, walleye, larvae 579 squid, tropical 13 Otter, sea 704 Outplanting, queen conch 390 Oyster, eastern, population modeling 347 Pacific Ocean alfonsino 748 anchovy, northern 711 crab, pelagic red 464 dolphin 434,451 common 434, 451 spinner 324, 434, 451 spotted 324,434,451 striped 122,324,434,451 lobster, Hawaiian spiny 203 rockfish assemblages 620 yellowtail 655 roughy, orange 76 salmon, juveniles 79 seamount 851 shark, demersal 851 sole English 189 petrale 481 Pacific Ocean, North salmon, Pacific 79 shark, demersal 851 squid, clubhook 13 Pacific Ocean, tropical clupeids, tropical 102 dolphin 434,451 common 434,451 spinner 324,434,451 spotted 324,434,451 striped 122,324,434,451 soldierfish, brick 516 squirrelfish 516 tuna 132 Pandalidae — see Shrimp Panuliridae — see Lobster, spiny INDEX: SUBJECTS Fishery Bulletin 92( 1-4], 1994 905 Panulirus marginatus — see Lobster, Hawaiian spiny Paralichthys dentatus — see Flounder, summer 591 Penaeidae — see Shrimp 633 Penaeus aztecus — see Shrimp, brown 633 uorarum — see Shrimp, pink 633 setiferus — see Shrimp, white 633 Perch, silver 254 Peru, dusky dolphin feeding 569 Photography, dolphin school size 122 Photololigo chinensis 302 Photololigo species juvenile distribution and abundance 302 Physeteridae — see Whale, sperm 760 Physeter macrocephalus — see Whale, sperm 760 Physiology, eastern oyster 347 Pinnipedia — see Seal, northern fur 144 Plectropomus leopardus — see Coral trout Pleuroncodes planipes — see Crab, pelagic red 464 Pleuronectes americanus — see Flounder, winter 608 vetulus — see Sole, English 189 Pleuronectidae — see Flounder; Sole Plotosidae — see Catfish, eel tailed 880 Pogonias cromis — see Drum, black 254 Pollock, walleye 207, 540, 579 Pollution sewage disposal 608 Pomatomidae — see Bluefish 494 Pomatomus saltatrix — see Bluefish 494 Population dynamics crab, snow 412 croaker, Atlantic 1 length at age model 865 oyster, eastern 347 roughy, orange 236 shark, demersal 851 Population studies catfish, eel tailed 880 conch, queen 171 crab, snow 412 croaker, Atlantic 1 dolphin, eastern tropical Pacific spp. 122,434,451 egg production 102 drum, red 58 mackerel, king 91 morwong, jackass 817 oyster, eastern 347 roughy, orange 236 shrimp 599 sole, petrale 481 tuna, eastern Pacific spp. 132 Portunidae — see Crab, blue 180 Prawn (see also shrimp) 599 Predation — see also Mortality rates flounder, winter 608 seal, northern fur 144 Predator-prey relationships anadromous alewife-euphausiid 157 northern fur seal-pollock 144 Predator-prey relationships (continued) sea otter-bivalves 704 winter flounder-benthic macrofauna 608 Prionace glauca — see Shark, blue Pristipomoides multidens — see Snapper, gold-band 664 typus — see Snapper, sharp-tooth 664 Proximate analysis cobia, ovaries 686 crab, pelagic red 464 Psenopsis humerosa — see Butterfish, black-spot 664 Pseudopleuronectes americanus — see Flounder, winter 608 Purse seine tuna, eastern Pacific spp. 132 Rachycentridae — see Cobia 686 Rachycentron eanadum — see Cobia 686 Radiometric ageing 292 Recruitment clupeids, tropical 102 conch, queen 171 flounder, summer 591 herring 102 menhaden, Atlantic 420 lobster, Hawaiian spiny 203 overfishing 516 sardine 102 shrimp, penaeid 633 shrimp 599 sole, petrale 481 sprat 102 Redfish 58 Reef fishes community composition 787 soldierfish, brick 516 squirrelfish 516 Reproduction conch, queen 171 dolphin, striped 122 shark, demersal 851 Reproductive biology clupeids, tropical 102 cobia 686 cockfish 531 croaker, Atlantic 671 drum, red 841 herring 102 lobster, Hawaiian spiny 203 sardine 102 shark, demersal 851 soldierfish, brick 516 sole, English 189 sprat 102 squirrelfish 516 weakfish 697 Reservoir 872 Reverse migration 223 Rockfish 313,620,655 assemblages 620 canary 620 chilipepper 198 copper 313 darkblotched 620 greenstriped 313, 620 Rockfish (continued) Pacific ocean perch 620 quillback 313 redbanded 620 redstripe 620 rosethorn 620 rougheye 620 sharpchin 620 shortbelly 620 shortspine thornyhead 620 splitnose 620 tiger 313 widow 620 yelloweye 313 yellowtail 313,620,655 Roughy, orange 236 Sablefish 292 Salmon chinook 79, 872 chum 79 coho 79,471 Pacific 79 pink 79 sockeye 79 Salmonidae — see Salmon, chinook; Salmon, chum; Salmon, coho; Salmon, Pacific; Salmon, pink; Salmon, sockeye Sampling design weight-length relationships 198 Sanctuary, marine 787 Sand lance, northern 647 Sardine Amblygaster sirm, reproductive biology 102 Pacific, prey of dusky dolphin 569 Sardinops sagax — see Sardine, Pacific 569 Saurida micropectoralis — see Lizardfish, short-finned 664 Saxidomus giganteus, prey of sea otters 704 Scales, ageing weakfish 555 School structure, dolphin 122 Sciaenidae 254; see also Croaker 1, 671; Drum 58, 509, 841; Weakfish 555, 697 Sciaenops ocellatus — see Drum, red Seomberomorus cavalla — see Mackerel, king Scombridae — see Mackerel; Tuna Scorpaenidae — see Rockfish Seal, northern fur 144 Seamounts alfonsino 748 shark, demersal 851 Seasonal studies anadromous alewife, diet 157 blue shark, catch 474 lobster, Hawaiian spiny 203 northern fur seal, diet 144 oyster, eastern, mortality 347 Pacific cod, movements 800 shrimp, distribution 599 906 INDEX SUBJECTS Fishery Bulletin 92( 1-4), 1994 Seasonal studies (continued) spadefish, Atlantic, larvae 262 tropical clupeids, abundance 102 weakfish, sonic muscles 697 Seatrout, spotted 254 Sea turtle, Kemp's ridley diet in New York waters 26 Sea urchin green, prey of sea otters 704 Sebastes — see Rockfish babcocki — see Rockfish, redbanded 620 caurinus — see Rockfish, copper 313 crameri — see Rockfish, darkblotched 620 elongatus — see Rockfish, greenstriped 313, 620 entomelas — see Rockfish, widow 620 diploproa — see Rockfish, splitnose 620 flavidus — see Rockfish, yellowtail 313 goodci — see Rockfish, chilipepper 198 helvomaculatus — see Rockfish, rosethorn 620 jordani — see Rockfish, shortbelly 620 maliger — see Rockfish, quillback 313 nigrocinctus — see Rockfish, tiger 313 pinniger — see Rockfish, canary 620 proriger — see Rockfish, redstripe 620 ruberrimus — see Rockfish, yelloweye 313 zacentrus — see Rockfish, sharpchin 620 Sebastolobus alascanus — see Rockfish, shortspine thornyhead 620 Selectivity anadromous alewife, diet 157 northern fur seal, diet 144 Serranidae — see Coral trout Sexual dimorphism dolphin, spotted 324 Sexual maturity — see also Reproductive Biology clupeids, tropical 102 cockfish 531 croaker, Atlantic 671 drum, red 841 shark, demersal 851 soldierfish, brick 517 squid, clubhook 13 squirrel fish 516 Sewage disposal 608 Shark blue 474 demersal 851 Shrimp brown 633 deep-water 599 penaeid 633 Size estimation — see Age-size estimation Size limits, eastern oyster 347 Size segregation, dolphins 122 Size-selectivity anadromous alewife, diet 157 northern fur seal, diet 144 Snapper gold-band 664 one-band 664 red 664 Snapper (continued) saddle-tailed 664 scarlet 664 sharp-tooth 664 Timor 664 Soldierfish, brick 516 Sole English 189 flathead 223 petrale 481 South America 569 Spadefish, Atlantic 262 Spain 599 Spawning — see also Reproductive Biology bluefish, juvenile 494 clupeids, tropical 102 cobia 686 cockfish 531 croaker, Atlantic 671 drum, red 841 mackerel, king 91 menhaden, Atlantic 420 pollock, walleye 579 shark, demersal 851 soldierfish, brick 516 sole, English 189 squid, tropical 302 squirrelfish 516 sturgeon, white 760 Spawning biomass clupeids, tropical 102 Species association cetaceans. Gulf of Mexico 773 ichthyoplankton, Gulf of Alaska 223 reef fish 787 rockfish 313,620 salmon 79 tuna-dolphin 132 zooplankton, Gulf of Alaska 67 Species identification 254 Spines, ageing 851 Sprat 102 Spratelloides delicatulus — see Sprat 102 Squid clubhook 13 gonatids in northern fur seal diet 144 jumbo flying, prey of dusky dolphin 569 oceanic, age and growth 13 Patagonian, prey of dusky dolphin 569 Squirrelfish 516 Stenella attenuata — see Dolphin, spotted 324, 760 clymetie — see Dolphin, clymene 760 coentleoalba — see Dolphin, striped 122, 773 frontalis — see Dolphin, Atlantic spotted 773 longirostris — see Dolphin, spinner 773 Stock assessment age-structured models, modified 711 anchovy, northern 711 dolphin, eastern tropical Pacific spp. 122,434,451 dolphin, striped 122 Stock assessment (continued) reef fish 737 stock reduction analysis 236 surplus production models 737 extensions to 374 Graham-Schaefer 374 modified 711 tuna, eastern Pacific spp. 132 Stock enhancement, queen conch 390 Stock identification dolphin spotted 324 striped 122 drum, red 58 mackerel, king 91 Stock structure cod, Pacific 800 dolphin, striped 122 morwong, jackass 817 Stranding, pelagic red crab 464 Stress crab, pelagic red 464 oyster, eastern 347 Stromateidae — see Butterfish, black-spot 664 Strombidae — see Conch, queen Strombus gigas — see Conch, queen Strongylocentridae — see Sea urchin, green 704 Strongyloccntrotus droebaehiensis — see Sea urchin, green 704 Sturgeon, white, spawning and early life history 760 Submersible surveys of rockfish 313 Surplus production models 374 Survey, aerial dolphin, striped 122 Survival — see Mortality rates Swimbladder deflation 641 Swordfish 374 Synodontidae — see Lizardfish 664 Systematics, egg and larval 725 Tagging studies cod, Pacific 800 rockfish, yellowtail 655 Taxonomy dolphin [fish] common 275 pompano 275 spadefish, Atlantic 262 tripletail 33 Tclmessus, prey of sea otters 704 Temperature and abundance of crab, snow 412 sole, petrale 481 shark, blue 474 and development of pollock, walleye, eggs 207 and distribution of dolphin (fish], larvae common 275 pompano 275 ichthyoplankton pollock, walleye, eggs 540 INDEX: SUBJECTS Fishery Bulletin 92( 1-4), 1994 907 Temperature and distribution of (continued) petrale 481 spadefish, Atlantic, larvae 262 zooplankton, Gulf of Alaska 67 Temporal variation alfonsino 748 distribution 236, 748 Temporal variation (continued) reproduction 236 roughy, orange 236 shark, demersal 851 size at maturity 236 spawning time 236 Theragra chalcogramma — see Pollock, walleye Thunnus albacares — see Tuna, yellowfin Trachichthyidae — see Roughy, orange 236 Trachurus symmetricus — see Mackerel, jack 569 Transport, larval menhaden, Atlantic 420 sole, petrale 481 Tripletail 33 Tuna and dolphin abundance 132 skipjack tuna 132 yellowfin tuna 132 Tursiops truncatus — see Dolphin, bottlenose 773 Turtle, sea, Kemp's ridley 33 Vertical distribution ichthyoplankton 223 pollock, walleye, eggs 540 shrimp 599 squid, tropical 302 Video, survey 787 Virginia croaker, Atlantic 1, 671 flounder, summer 591 oyster, eastern 347 Virginia (continued) sciaenids, eggs 254 Vitellogenesis 189 Washington rockfish, assemblages 620 sole, petrale 481 Weakfish 555,697 Weight-based sampling 198 Weight-length relationships 198, 664 Whale dwarf sperm, in Gulf of Mexico 773 pygmy sperm, in Gulf of Mexico 773 sperm, in Gulf of Mexico 773 Yield per recruit tuna, eastern Pacific spp. 132 Zooplankton, Gulf of Alaska 67 Errata Erratum: Fish. Bull. 91:310-327. Erratum: Fish. Bull. 92:374-389. Prager, M. H., and A. D. MacCall. 1993. Prager, M. H. 1994. Detection of contaminant and climate effects on A suite of extensions to a nonequilibrium surplus- spawning success of three pelagic fish stocks off production model, southern California: Northern anchovy Engraulis mordax, Pacific sardine Sardinops sagax, and chub Correction: On page 376, in Equation 8a, a term con- mackerel scomber japonicus. taining "-1" was misplaced. The correct equation is Correction: On page 319, in Table 5, the R2 statistic for the "combined" model of Pacific sardine is given as 0.47. The correct figure is 0.83, as stated on page 322. r In r- x '/3B,{e"'-V -.' " a, 91 18 Publication Awards, 1992 National Marine Fisheries Service, NOAA The Publications Advisory Committee of the National Marine Fisheries Service is pleased to announce the awards for best publications authored by NMFS scientists and published in the Fishery Bulletin for 1 992 and in Marine Fisher- ies Review for 1992. Eligible papers are nominated by the Fisheries Science Centers and Regional Offices and are judged by the NMFS Editorial Board. Only articles that significantly contribute to the understanding and knowl- edge of NMFS-related studies are eligible. We offer congratulations to the fol- lowing authors for their outstanding efforts. Fishery Bulletin, 1 992 Outstanding Publication Elizabeth F. Edwards Energetics of associated tunas and dolphins in the eastern tropical Pacific Ocean: a basis for the bond. Fish. Bull. 90:678-690. Elizabeth Edwards is with the Southwest Fisheries Sci- ence Center, La Jolla, California. Marine Fisheries Review, 1 992 Outstanding Publication Robin S. Waples Pacific salmon, Oncorhynchus spp., and the defi- nition of "species" under the Endangered Spe- cies Act. Mar. Fish. Rev. 53(3)11-22. Robin Waples is with the Northwest Fisheries Science Center, Seattle, Washington. Honorable Mention Jeffrey J. Polovina Variability in spiny lobster Panulirus maginatus recruitment and sea level in the Northwestern Hawaiian Islands. Fish. Bull. 90:483-493. Jef- frey Polovina is with the Southwest Fisheries Science Center, Honolulu, Hawaii. Honorable Mention Dean W. Ahrenholz Population biology and life history of the North American menhadens, Brevoortia spp. Mar. Fish. Rev. 53(4):3-19. Dean Ahrenholz is with the Southeast Fisheries Science Center, Beau- fort, North Carolina. 909 U S Postal Service STATEMENT OF OWNERSHIP, MANAGEMENT AND CIRCULATION Required by 39 U S. C 3685) 1A. Title of Publication Fishery Bulletin IB PUBLICATION NO 3 6 2 Date o* Filing 8-29-94 3. Frequency of Issue Quarterly 3A No of Issues Fublished Annually 4 36 Annual Subscription Price $27.00 domestic $33.75 foreign 4 Complete Mailing Address ot Known Ollice of Publication (Street, City. County. State ami ZIP-t 4 Code) (Not printers} NMFS Scientific Publications Office 7600 Sand Point Way NE, BIN C15700, Seattle, WA 98115-0070 5 Complete Mailing Address ol the Headquarters of General Business Offices of the Publisher INot printer) National Marine Fisheries Service, NOAA/DOC 1315 East-West Highway, Silver Spring, MD 20910 6 Full Names and Complete Mailrng Address of Publisher, Editor, and Managing Editoi (This i MUST NOT be blank) Publisher (Name ami Complete Mailing Address) Willis Hobart, NMFS Scientific Publications Office 7600 Sand Point Way NE , Seattle, WA 98115-0070 Editor (Name and Complete Mailing Address) Sharyn Matriotti, NMFS Scientific Publications Office 7600 Sand Point Way NE, Seattle, WA 98115-0070 Managing Editor (Name and Complete Mailing Address) Same as above 7. Owner (If owned by a corporation, us name and address must be stated and also immediately thereunder the names and addresses of stockholders owning or holding I percent or more of total amount of stock. If not owned by a corporation, the names and addresses of the individual owners must be given If owned by a partnership or other unincorporated firm, its name and address, as welt as that of each individual must be given If the publication is published hy a nonprofit organization, its name and address must be stated) (Item must be completed.) Complete Mailing Address II T fi . Depf . of Cnmmprrp 14th St., NW, Washington DC 202 30 Known Bondholders. Mortgagees, and Other Security Holders Owning or Holding 1 Percent or More of Total Amount of Bonds. Mortgages or Other Securities (If there are none, so state) Complete Mailing Address 9. For Completion by Nonprofit Organizations Authorized To Mail at Special Rates (DMM Section 423.12 only) The purpose, function, and nonprofit Status of this organization and the exempt status for Fedeial income tan purposes (Check • !:■! Has Not Changed During Preceding 12 Months □ (21 Has Changed During Preceding I 2 Months (If changed, publisher must submit explanation of change with this statement ) Extent and Nature of Circulation (See instructions on reverse side) Average No Copies Each Issue During Preceding 12 Months Actual No Copies of Single Issue Published Nearest to Filing Date A. Total No Copies (Net Press Run) 2025 2012 B Paid and/or Requested Circulation 1 Sales through dealers and carriers, street vendors and counter sales Handled by U.S. Sup:, of Documents, Washington, DC 2Q24f 2. Mail Subscription (Paid and/or requested) 2025 2012 C. Total Paid and/or Requested Circulation (Si™ or I0BI and I0B2) I, llll 600 D Free Distribution by Mail, Carrier or Other Means Samples. Complimentary, end Other Free Copies 1405 1386 E- Total Distribution